In 1996 the DOE formed the Materials Microcharacterization Collaboratory (MMC) to bring together the four DOE-sponsored electron microscope user facilities in one collaboratory. The MMC will bring the microanalysis and microcharacterization tools that are available in national centers to geographically dispersed researchers working in industries, universities, and Government laboratories. It will enable these remote users to share on-line the instrumentation, knowledge and expertise available at the individual facilities making up the collaboratory.

LBNL first demonstrated on-line remote control of a high-voltage TEM from Kansas City in 1995 in a joint project by LBNL computer scientists and NCEM microscopists. The microscope chosen was a 1.5MeV Kratos EM-1500 that is used for in-situ electron microscopy. In the demonstration, a specimen of an Al/Pb alloy was heated and observed, with remote-operator control of heating rate, microscope focus, and stage movement (translation and tilt).

To introduce engineering undergraduates to microstructural characterization in the scanning electron microscope, The Department of Materials Science and Engineering (MS&E) at The University of Michigan has developed the Teaching SEM. This instrument, based on a Philips XL30 SEM and located in the teaching laboratories, may be controlled from a lecture theater or classroom virtually anywhere on the University campus via a combination of the campus computing network and the campus cable television network. The use of the campus cable TV network is a limitation. Many areas of the University have Internet connections, but no cable connections. A number of local schools have expressed an interest in incorporating use of this system into their classes. The current work seeks to eliminate the cable TV network and send both the control and video signals over the Internet.

The integration of nano-science and technology concepts into upper-division high school and lower-division college curricula will require innovative educational approaches that will help students understand the structures and properties of matter on a scale below 100 nanometers, i.e., the nanoscale. This Interactive Nano-Visualization in Science and Engineering Education (IN-VSEE) project will create a consortium of university and industry researchers, community college and high school science faculty, computer scientists and museum educators with a common vision of creating an interactive World Wide Web (WWW) site to develop a new educational thrust based on remote operation of advanced microscopes and nanofabrication tools coupled to powerful surface characterization methods. The centerpiece of this project is the web-based operation of the revolutionary scanning probe microscope (SPM), which has evolved rapidly into a relatively simple, yet powerful, technique capable of imaging and manipulating materials at resolutions down to the atomic scale.

The continuing remarkable progress of computer technology is dramatically changing how electron optical instrumentation is being used. Instead of concentrating on taking just one spectrum or one image at a time, it now has become possible to design complete experiments aimed at solving the problems at hand more rapidly and more directly.

We have designed a PC-based acquisition system that aims to innovate and simplify both TEM and STEM operation. The first advance this system represents is integrated data acquisition from all detectors commonly found on electron microscopes: TV and slow-scan CCD cameras, different flavors of STEM detectors, and EDX, CL, Auger and EELS spectrometers. The new system allows experiments combining microscope control with the acquisition capabilities of all these detectors. The second advance is the implementation of “live” processing. Most common processing tasks can be executed on-the-fly, even at video-rate acquisition speeds.

One of the most difficult problems on the way to sub-Ångstrom resolution at a medium accelerating voltage of about 200 kV, the correction of the spherical aberration, has been solved. Already more than fifty years ago, Scherzer showed that the two major axial aberrations, the chromatic and the spherical aberration, can be corrected by a suitable combination of electrostatic and/or magnetic multipole elements. Due to the availability of high-coherent electron sources with a small relative energy width DE/E the necessity of correcting the chromatic aberration can be avoided.

Our solution for spherical aberration correction follows a suggestion by Rose. It is based on two electromagnetic hexapoles and four additional lenses. The principle by which correction is achieved is based on the fact that the primary aberrations of second order of the first hexapole are compensated by the second hexapole element.

There are three main algorithms for autotuning a transmission electron microscope (TEM), (1) image variance method, (2) tilt-induced image shift (TIS) method, and (3) automated diffractogram analysis method (ADA). In practice, each method has its own advantages and limitations. Therefore it is necessary to obtain a comprehensive understanding of each in order to be able to select the most appropriate one(s).

There are four key elements in any TEM autotuning setup, (1) microscope control by an external computer, for example via the standard RS232 serial port; (2) proper image acquisition device, for example a TV or a slow-scan CCD camera; (3) appropriate software for image analysis, processing, and optimization; (4) a fast computer that is the central control unit in the setup and dictates all the activities and operations.

Acquisition of large dynamic range information is important for optimizing the performance of SEM’s, TEM’s, and other microscopies. The majority of digital imaging performed in desktop publishing and other consumer markets today is limited to 8-bit resolution, and these technologies are often modified for use in electron microscopy (EM). The limitations of these techniques are clearly evident in many EM situations despite the fact that human visual system is not equipped to differentiate more than 2^5 gray levels. The most common case is when a user seeks to view information hidden in both the image's dark levels and in the brightest levels. If multiple images at different detection levels are required despite digital processing then the image information is being limited by the digitization resolution.

Digitization is the process of converting a signal from the analog form that originates from our microscope detectors to a digital form that is suitable for processing within computers.

Color and Color Perception

Accurate color reproduction is an essential component in the effective use of digital imaging techniques in light microscopy. The color reproduction process begins with an understanding that color is the result of three key elements; light, the illuminated object, and the observation method. When light strikes an object, wavelengths may be reflected, absorbed or transmitted. Additionally, the observed colors are strongly influenced by the intensity of the illumination and its spectral content. Colors we think of as “white” can vary significantly in their spectral distribution, e.g., skylight is actually a bluish white while tungsten bulbs produce a yellowish white.

Light waves that reach the eye stimulate a complex process that is not yet fully understood. Within the retina, three different types of cones respond to color hues and brightness while rods sense only brightness.

Image analysis is the process of quantifying some aspect of an image—its particle size distribution, for example. Manual methods were in use long before computers made image analysis much faster and more reproducible. Linear measurements of diameter, point counting to measure volume fraction, and intercept counting to determine grain size have been used for over 100 years. Automatic image analysis (AIA), however, can make more measurements, and even calculate derived measurements, such as aspect ratio or circularity. AIA of a specimen or micrograph, of course, is only as good as the contrast mechanism used, so the imaging signal must be chosen carefully to reveal to the computer what is to be measured. Obtaining sufficient contrast is often the limiting task.

Once an imaging signal is chosen and the digital resolution set, the computer can analyze the image. The feature descriptors can be generic:

Accurate characterization of electron diffraction patterns can be tedious, which encourages development of computer assisted tools and methods. We developed a spot measurement tool to characterize rapidly arrays of diffraction spots that are characteristic of a single crystal, and to measure precisely the d-spacing values for individual spots. The spot tool determines these vectors for averaged measurements from many spots in a digital image of the diffraction pattern. Previously we developed automated methods for spot pattern analysis. Why make an operator-assisted tool? This tool is faster than either our automated or entirely manual methods, and it allows assessment of the quality of the data at the beginning of the analysis.

It is easy to see regular patterns of spots in a zone axis diffraction pattern (Fig. 1.1). An operator guides the initial calculations by adjusting an array of circles to approximately cover the spots in the pattern to be analyzed.

Mathematical representation of particle orientation and shape can show quantitative differences in particle morphology and microstructures of ceramic green bodies with apparently equiaxed powders. In this work, fourier descriptors and moment invariants were used to quantify the shape of ceramic particles and identify their orientation. Edges constitute a significant portion of the information contained in these images, and hence edge detection forms an important issue in these image processing algorithms.

A Field Emission Scanning Electron Microscope (FESEM, Leo 982) was used to collect images of various ceramic green bodies in order to assess the ceramic particle shape and orientation. The images were collected and displayed as rectangular pixels on the microscope monitor which were then stored as square pixels (square-pixel-interpolation). Application of various edge detection algorithms on the stored images show strong bias to edges in the direction perpendicular to the scanning direction.

In designing automated systems for interpretation of micrographs, it is often the goal to separate and discern various objects within the image data. After segmentation, measurements of the objects, called features, can then be calculated and used for process or statistical analysis. Simple segmentation schemes based on single threshold operations, often lack the sophistication to deal with intricate or subtle details of the image data, or require user intervention in the threshold selection process. This talk will present and discuss advanced techniques which adapt to the data, and which can operate autonomously without human supervision.

Segmentation is a process which transforms pixel based image data into symbolic descriptors representing groups of pixel elements. These descriptors are called objects and take the form of lines, regions, polygons, points, windows of interests, or other unique representations. Most laboratory image analysis software packages utilize segmentation schemes based on the principle of intensity analysis.

Micro-graphs acquired using optical or electron microscopes reveal information on a local scale. When a sample is not homogeneous at the scale of the image, it may be necessary to view the sample at a magnification lower than that available by the instrument. Montages of micro-graphs can be constructed through the following steps: acquisition of individual images, stitching individual images into a single image (a montage), and producing the final exhibit(s). Performing these steps has traditionally required careful stage and illumination control during acquisition, dodging during printing, and hours with scissors and glue-pot during paste-up. Because this procedure is labor intensive, the creation of montages was limited to only the most important or critical situations. Digital microscopy, image processing software, and Web-enabled multi-media applications have reduced the barriers for creation of montages and provided for new display modes, thereby stimulating a revival of the use of montages as an enhanced presentation mode for technical data.

Images obtained from an analog SEM are traditionally viewed and recorded from a cathode-ray tube (CRT). Many laboratories use instant film (e.g. Polaroid #52, #55 instant film) to justify image quality and obtain permanent image quickly. Digital imaging provides an alternative approach for image acquisition and recording. One major advantage of digital SEM is image averaging that allows one to improve the signal-to-noise ratio (SNR) from a noisy quick-scan image to reduce charging. SEM signal yield is proportional to incident beam intensity, image acquisition time or duration of beam interaction with specimen (dwell time). The higher beam intensity, or longer the dwell time, the more signal generated. However, for high-resolution SEM imaging, the beam dose and dwell time are limited by drafting, radiation damage, and contamination. Therefore high-resolution biological SEM images invariably have poor SNR.

Modem technical imaging devices acquire digital image data at a precision level that often exceeds that of available display technologies so that parts of the data information may not be revealed in the displayed image. Data resolution and display resolution are closely related to image information and digital image processing provides several useful tools for compensation of such display deficiencies.

Typical precision data may be captured by a CCD camera, a scanning microscope or flatbed digitizer from photographic film and may exceed a size of 1,000×1,000 data points and a data point depth of 12 bit. Digital data has two general characteristics that deal with spatial resolution and intensity resolution. The distance represented between data points determines the spatial data resolution. Equally important is the bit-resolution which describes the maximum number of intensity levels possible in each point.

The Electron Microscope Unit at the University has a good selection of microscopes: two tungsten SEMs, two FESEMs, a microprobe analysis system, a 125 kY TEM, a 200 kV TEM, an AFM, an FIB miller, a Zeiss Photomicroscope and a Leica Macroscope. We service around 300 clients’ projects a year in every field of experimental science and engineering, logging over 8,000 hrs of beam time annually.

But funding constraints have always kept us short staffed and our laboratory has been working towards complete digital image capture for the past ten years to enhance our productivity. The perceived benefits of digitisation for us are:

• Photographic processing of negatives eliminated.

• Archiving of (bulky) photographic negatives eliminated.

• Need for special darkroom & graphics skills eliminated

• Responsibility for archiving and indexing images devolved to individual users

• Responsibility for image processing devolved to individual users.

• Rapid turnaround of images.

• Rapid sharing of results.

The Atom-Probe Field Ion Microscope was introduced in 1967 at the 14th Field Emission Symposium in Gaithersburg, Maryland. The Atom-Probe was, and remains, the only instrument capable of determining “the nature of one single atom seen on a metal surface and selected from neighboring atoms at the discretion of the observer”. The development of the Atom-Probe is a story that highlights Erwin Muller's strong and sometimes volatile personality. It is a story of an instrument that one NSF proposal reviewer called “impossible” because “single atoms could not be detected”. It is also the story of the Field Emission Laboratory at Penn State in the late 1960s and the contributions of two superb technicians, Gerald Fowler and Brooks McLane, and two graduate students, Douglas Barofsky and John Panitz. The anecdotes from this time are colorful and reflect Erwin's pedigree as Gustav Hertz's student in the Berlin of the 1930s.

A possible description of the ideal microscope would be an instrument which was able to reconstruct, with atomic resolution and in 3 dimensions, both the position and the chemical identity of atoms in a material. The 3-dimensional atom probe (3DAP) is the technique which comes closest to this goal.

The position-sensitive atom probe (PoSAP) was the first 3DAP. In the PoSAP, the high magnification of the field-ion microscope is combined with the time-of-flight mass spectroscopy of the atom probe, and position-sensitive detection based on a wedge-and-strip anode, Fig.1. This combination allows the chemical identity and the original surface position to be determined for single atoms removed from a field-ion specimen by pulsed field evaporation. Continued field evaporation and analysis builds up a 3D image of the distribution of all the atomic species originally present in the material, Fig. 2.

Today material science requires the use of increasingly powerful tools in material analysis. The last twenty years have witnessed the development of a number of analytical techniques. However, among these techniques, only a few allow observation and analysis of materials at the nanometer level ; one can mention scanning transmission electron microscopy coupled with electron energy loss spectrometry (STEM-EELS), SIMS and atom-probe techniques. In the STEM combined with electron energy loss spectrometry the spatial resolution can be better than one nanometer on a wide scanning area. However the in-depth spatial resolution is limited by the specimen thickness and therefore strongly depends on samples preparation. On the contrary, In SIMS techniques the in-depth spatial resolution can be as good as one nanometer but the lateral resolution is set by the ion beam diameter.

In a three dimensional atom probe, the identity and spatial coordinates of the atoms field evaporated from the specimen are determined. Their identity is calculated from the flight time from the specimen to the single atom detector. The x and y coordinates of the atom in the specimen are determined from the coordinates of its impact position on the position-sensitive detector and the z coordinate is determined from its position in the evaporation sequence. These data may then be reconstructed to visualize and quantify the distribution of all the elements in the specimen. Several types of position-sensitive detectors have been used including a wedge-and-strip detector (position-sensitive atom probe), a 10 by 10 array of anodes (tomographic atom probe), and a gateable CCD camera (optical atom probe). The wedge-and strip and the CCD camera detectors both suffer from the limitation that if more than one atom strikes the detector on a field evaporation pulse then the impact positions cannot be determined in many cases.