Scanning Microscopes Bring Imaging To The Atomic Scale

In 1990, scientists at IBM Corp.'s Almaden Research Center in San Jose, Calif., used the minute tungsten probe of a scanning tunneling microscope to move atoms of xenon around on a nickel surface, ultimately spelling out the letters I-B-M (Nature, 344:524-6, 1990). This remarkable feat of engineering opened up an entirely new realm of applications for this microscope, which was designed to image atomic-sized specimens, not push them around. Since the first transmission electron microscope was sold in 1935, microscopes that use electrons--rather than light waves--to image objects have brought into focus levels of detail that were previously unimaginable. The scanning tunneling microscope (STM) is one of the more recent developments in this field, and using STM to write with atoms is one of its more innovative applications. But scientists are using electron microscopy in other new ways, as well. These include constructing three- dimensional images of minerals and some biological molecules, creating and analyzing polymers and superconductors, and even watching--on the atomic level--complex biological processes in action. The electron microscope is essentially a light microscope that uses beams of high-energy electrons in place of light waves to illuminate the specimen. Using electrons with wavelengths as small as 0.1 nanometer, rather than longer-wavelength visible light, removes the limits on resolution imposed by diffraction in optical microscopy. The transmission electron microscope (TEM) directs a beam of electrons at the sample. These electrons pass through the sample--which must be cut into very thin sections-- like light waves in an optical microscope. Whereas the resolution of a light microscope is typically greater than 250 nanometers, TEM can resolve objects in the 3- to 5-nanometer range. The scanning electron microscope (SEM) scans a beam of electrons back and forth over the surface of a sample. The detector then collects the backscattered electrons emitted from the specimen and builds up a three-dimensional image. Specialized electron microscopes that use high-energy electrons (which have a very short wavelength) have enabled researchers to achieve near-atomic resolution (1 to 3 angstroms, or 0.1 to 0.3 nanometer) for certain kinds of samples, like thin crystalline films. Transmission and scanning electron microscopes are both widely used in applications ranging from determining the composition of superconducting substances to imaging blood cells flowing through arteries. At the Lawrence Berkeley Laboratory in California, for example, a research team led by biophysicist Ken Downing has obtained what they believe to be the first three-dimensional images of an atomic structure. They used a combination of electron microscopy and computer simulation to create a three- dimensional image of staurolite, a mineral found in the Swiss Alps. Their image included even the hard-to-see oxygen atoms, usually obscured by heavier metal atoms. Downing and his collaborators hope that materials scientists will be able to use this technique to determine the atomic structure of superconducting materials, and for other high-resolution structure studies in which the position of critical atoms like oxygen must be pinpointed. SEM is commonly used to look at the surfaces of many different types of materials, from cells to microchips. "Resolution of 2 nanometers can be achieved using SEM with a field emission gun [as the source of electrons]," says Karleen Davies, product manager for SEM at Carl Zeiss Inc. in Thornwood, N.Y. Zeiss' customers use the instrument to examine bulk properties of materials--for example, to look at impurities in the raw materials that go into making steel or cement, or to study the integrity and smoothness of the internal structures of an artificial heart. SEM has advanced beyond simple imaging. "SEM is a platform for a lot of other analytical sciences," says Davies. Scientists can use SEM to perform certain chemical analyses and measure particle size or porosity. SEM has been used extensively in microelectronic manufacturing to monitor production quality of integrated circuit elements on microchips. "The circuitry must be flawless, especially as the size of the chips shrinks," says former IBM systems engineer Kevin Ahern. One drawback to SEM is that the sample must be coated with metal before it is imaged. "A major challenge in SEM is to eliminate the need to spend lots of time doing sample preparation," says Zeiss' Davies. Furthermore, samples must be imaged in a vacuum, so that the electrons don't interact with air molecules instead of the specimen. Finally, the electrons bombarding the sample may contaminate or even destroy it, complicating the use of the SEM device for biological applications and in the microelectronics industry, in which integrity of the circuit elements is critical. As microelectronic circuitry shrinks to sizes below SEM's limits of resolution, scientists and engineers will have to turn to more powerful instruments for quality control of these devices. One such instrument is the scanning probe microscope, which relies on a tiny probe less than a nanometer in size, placed extremely close to the surface of the specimen. This new family of microscopes, which includes the scanning tunneling microscope, overcomes many of the limitations of electron microscopy. STM does not rely on an external source of radiation to illuminate the sample and produce an image, as do electron and optical microscopy. Instead of shooting atoms at a sample, it uses electrons already present on the probe and on the surface of the sample, measuring the "tunneling current" generated as these electrons shuttle back and forth. The basic principles behind scanning probe microscopy have been known since 1956, when J.A. O'Keefe, then with the U.S. Army Mapping Service, proposed that to overcome the diffraction limit imposed by lenses in optical microscopy, light should be sent through a tiny aperture to illuminate only the part of the specimen that was directly in front of the hole (Journal of the Optical Society of America, 46:359). In O'Keefe's model, resolution would be limited by the size of the aperture, and not by the wavelength of the light. In principle, this instrument would be capable of creating images showing details smaller than the one-half wavelength limit imposed on optical microscopes. At the time, however, the technology to precisely position and move the probe he conceived did not exist. The breakthrough that led to implementation of O'Keefe's technology was the advent of piezo- electrics--ceramics that show minute changes when the electrical potential of the material is altered. In 1981, Gerd Binnig and Heinrich Rohrer of the IBM Zurich Research Laboratory combined O'Keefe's ideas with piezoelectric controls and developed STM, earning them a Nobel Prize in 1986. Piezoelectric micropositioning controls maneuver the tungsten tip into place only a nanometer or two from the sample. A small voltage applied to the tip causes the electrons to "tunnel" across the gap, creating a current directly related to the distance between the two surfaces. Since the probe is scanned back and forth across the specimen, changes in the tunneling current are measured, accentuating the bumps and valleys on the surface of the specimen. In STM, resolution is not dependent upon the wavelength of the radiation source, but instead is based on the size of the probe, which is only a single atom in diameter. This gives STM a resolution of less than 0.01 nanometer--well within the subatomic range. While both SEM and STM are used to analyze surfaces, says Davies, SEM is used to detect scratches on a surface, for example, while STM can image the very arrangement of the molecules that compose it. Development of STM by Rohrer and Binnig created a surge of research activity on scanning near-field technology, leading to the introduction of new types of microscopes of the same basic architecture. Variations on the scanning probe theme include the atomic force microscope (AFM), with a probe that measures forces near the surface of a sample; the scanning thermal microscope, which detects changes in temperature; and the laser force microscope (LFM), which discerns minute changes in resonance frequency of a vibrating probe. And while most of these microscopes are still in development, there are plenty of projected applications, including measuring electrical potentials inside of living cells, or nondestructive imaging of biological molecules in their native state. And, of course, one of STM's most novel applications is IBM's experiment in using it to manipulate atoms. "Being able to move single atoms around means that we can potentially manufacture computer components on the atomic level," says Ahern. "This will contribute to the further miniaturization of computer chips and circuitry, which will increase the speed and power of computers." At present, the key application of scanning probe microscopes is in the area of surface science. One of the major objectives of research involving STM is to learn more about the superconducting properties of substances like silicon, a component of integrated circuits, by examining the atomic arrangement on surfaces of single crystals. IBM researcher H. Kumar Wickramasinghe and colleagues developed the laser force microscope for the specific purpose of examining microelectronic devices without any potentially damaging surface contact by the probe. Wickramasinghe uses LFM to scrutinize the delicate circuitry of computer chips. The imaging power of STM has led many researchers to try their hand at imaging biological materials, like DNA or protein. However, STM relies on the tunneling current generated between the tungsten probe and an electrical conducting sample to form an image; hence, nonconducting samples, and even some semiconductors, must be made conductive before they can be imaged. Samples are often coated with a thin layer of metal, but this means one cannot examine biological materials in their native state. To overcome this obstacle, biological samples are placed on conducting surfaces, or immersed in electrolyte solutions, to provide the necessary electrical conduction. Scientists at the University of Minnesota in St. Paul, for example, placed samples of DNA on a graphite surface and were able to obtain images of DNA that closely matched the appearance of computer-generated models based on X-ray diffraction data (Nature, 339:484-86, 1989). An IBM research team has imaged molecules of DNA and protein complexes by absorbing them onto a salt-coated conducting film (Science, 243:1708-11, 1989). These constraints on imaging nonconducting materials led to the 1985 development of the atomic force microscope by IBM Zurich's Binnig and Christof Gerber, along with Calvin Quate of Stanford University. AFM measures attractive or repulsive forces that are generated between the probe and atoms in the sample. "Each atom in the sample exerts a different force on the tip, which corresponds to distance," explains Michael Cook, general sales manager of Burleigh Instruments Inc. in Fishers, N.Y. Since tunneling electrons are not used, it is possible to image nonconducting samples with the same atomic resolution seen with STM. By submerging the probe and the sample in water, a research team headed by Quate and Paul Hansma of the University of California, Santa Barbara, has been able to create images of biological substances, and may have even succeeded in capturing the first rudimentary images of a biological process in action at the molecular level--the polymerization of fibrin into a fibrin clot--as it occurred (H.K. Wickramasinghe, Scientific American, 261[5]:98-105, 1989). Researchers at Lawrence Livermore National Laboratory in California have used both STM and AFM to image DNA derived both naturally and synthetically. Senior biochemist Rodney Balhorn and colleagues have even obtained STM images of individual bases by placing the bases in a dilute solution on a graphite surface, where they formed a crystalline-like lattice. "The images are not well-defined, but we can distinguish between adenine [a purine base] and thymine [a pyrimidine base]," says Balhorn. Conceptually, this technique may someday be useful as a rapid way to sequence DNA, although, says Bal-horn, "it's not that simple." Biological specimens such as DNA vibrate when they are imaged in air, and water condenses on their surfaces, making it next to impossible to distinguish individual bases in a long piece of DNA. Better methods of preparing and imaging large DNAs will have to be developed before scanning probe microscopy can evolve into a gene mapper's tool. Some scientists, however, consider it nearly impossible to judge whether these images are truly representative of biological structures, or are mere artifacts or focused "noise." According to Jeff Travis, electron microscopy specialist in the biology department at the State University of New York, Albany, scanning probe microscopes have "not yet been proven to be useful for imaging biological specimens." Bal-horn agrees that this skepticism is well founded. "In many cases, imaging artifacts can look like DNA molecules," he says. To overcome this problem, Balhorn and other researchers construct synthetic DNAs of known length and then label them to ensure that the image formed is truly the DNA sample. Balhorn, for example, makes short DNA molecules with single-stranded ends, which he says can be seen in the STM image. Scanning probe microscopes have already proved their worth in the physical sciences, in which they are used for surface studies and on atomic-level analyses of organic molecules. And although resolution perhaps has not yet reached the atomic level for biological specimens, experiments by Balhorn, Hansma, and others clearly indicate the potential of the scanning probe microscope in the biological sciences, the next domain waiting to be explored at the atomic scale. Holly Ahern teaches cellular and molecular biology at the State University of New York, Albany.

Scanning Microscopes Bring Imaging To The Atomic Scale

Interested in reading more?

Become a Member of