Atomic Imaging Turns 50

Atomic-resolution imaging. The description sounds so absolute, so ultimate, so technically demanding. Yet nowadays, images that resolve individual atoms are so common. Journals, conference lectures, and textbooks are filled with them. Even newspapers have displayed atomic-resolution images.

But it wasn't always that way. From the time of the ancient Greeks, natural philosophers wondered about the composition of matter and tried to observe its building blocks. Some 500 years ago, scientists began using magnifying glasses—and later on, optical microscopes—to reveal structures that were much too tiny to observe with the naked eye. By the early 1950s, far smaller structures—but not atoms—were imaged with electron microscopes, which at that time were able to resolve nanometer-sized features. Shortly thereafter, just 50 years ago last month, the first ever atomic-resolution images were recorded with a field ion microscope.

On Oct. 11, 1955, Pennsylvania State University physics professor Erwin W. Müller and Kanwar Bahadur, who at the time was a Ph.D. student working with Mller, made history by being the first people to image individual atoms. The scientists were using a relatively simple and inexpensive instrument, and with it they directly observed individual tungsten atoms at the tip of a sharply pointed tungsten specimen.

Long before scanning tunneling microscopy (STM) and atomic force microscopy (AFM) became popular atomic-resolution methods for analyzing materials, and even before transmission electron microscopy (TEM) was shown capable of imaging individual atoms, Mller and his students were advancing field ion microscopy (FIM) toward its ultimate resolution.

The landmark studies conducted by the Penn State team in 1955 were published the following year (Phys. Rev. 1956, 102, 624 and J. Appl. Phys. 1956, 27, 474). It wasn't until some 15 years later that University of Chicago physics professor Albert V. Crewe and coworkers demonstrated atomic resolution with a scanning transmission electron microscope (Science 1970, 168, 1338). STM, a relative latecomer to the field, joined ranks with its atomic-resolution predecessors in the 1980s when Gerd Binnig and Heinrich Rohrer of IBM's Zurich Research Laboratory invented the third type of microscope able to discern individual atoms (Phys. Rev. Lett. 1983, 50, 120).

The run-up to the 1955 milestone starts with Müller's invention in 1936 of the field emission microscope, which predates the field ion microscope by more than 10 years. In the field emission instrument, a specimen in the form of a sharp tip or needle is maintained under high vacuum and subjected to a large negative voltage. Applying a few thousand volts to a specimen with a tip radius of approximately 50 nm results in an electric field at the specimen tip of roughly 10 V per nm. Fields of that magnitude are strong enough to cause electrons to be emitted from the tip via quantum mechanical tunneling, a process that had been predicted by pioneers of quantum theory in the 1920s. As the field-emitted electrons emerge from the specimen tip, they are accelerated toward a nearby phosphor screen, where they project a magnified image of the sample surface.

Müller determined that the lateral resolution of the new microscope was on the order of 2 nm. The device was used to examine diffusion and rearrangement of surface layers, and it outperformed other microscopes available in the 1930s and '40s but did not achieve atomic resolution. Nowadays, high-brightness electron sources based on the field emission process are widely used in electron microscopes and flat-panel display technologies.

A key advance in the resolving power of high-field microscopes came several years later when Müller used gas atoms—not electrons—to image the surface from which they emerged. Starting with a setup quite similar to the field emission microscope, Müller reversed the polarity of the sample, making it positive, and admitted an imaging gas such as hydrogen to the evacuated glass apparatus.

The imaging gas is the messenger that carries the information about the positions of the atoms in the sample lattice, explains David N. Seidman, a professor of materials science and engineering at Northwestern University and a longtime user of field-ionization methods.

Müller invented the field ion microscope in 1951 at the Kaiser Wilhelm Institute in Berlin; he is shown holding it on the cover of this issue of C&EN. He reasoned that the field ion microscope would surpass the field emission instrument in resolution because the imaging gas molecules would be confined to very small lateral motions at the surface. There, they would be ionized by the strong applied electric field and repelled toward a phosphor screen for imaging. Shortly after reporting in Zeitschrift fr Physik in 1951 on the initial results obtained with his new instrument, Müller immigrated to the U.S., where he took a position as a physics professor at Penn State and continued to work to improve FIM's resolution.

Initially, Müller misunderstood the process by which ions are formed via field ionization, according to Seidman and other FIM aficionados. The mechanism was worked out in the early 1950s by University of Chicago professors Robert Gomer and Mark G. Inghram, who showed that the gas atoms (helium and neon are most commonly used) are ionized a few angstroms above the sample surface. Seidman explains that the process forms narrow streams of He⁺ or Ne⁺ ions that emanate from individual lattice atoms.

In hindsight, advancing from the first fully functioning field ion microscope to obtaining atomic-resolution images didn't entail great strides, at least not in terms of technical requirements. Yet four years went by before the Penn State team made the leap.

One of the key requirements was cooling the tip to cryogenic temperatures for imaging. As fate would have it, the cooling experiment had been carried out twice—;unsuccessfully. Allan J. Melmed, who was another Ph.D. student working with Mller at the time, recalls that Mller eventually concluded, on the basis of those experiments and theoretical arguments, that cooling the tip would not improve the resolution. But Bahadur thought otherwise and was persistent.

Another requirement was avoiding the standard high-temperature treatments used for preparing metal samples. The way Melmed understands it now, after a career's worth of research experience at the National Institute of Standards & Technology (NIST), Gaithersburg, Md., and later at Johns Hopkins University, the conventional practice of heating a metal sample to high temperature to drive away impurities and produce a pristine surface worked against Mller and Bahadur.

They didn't realize that they were thwarting themselves every time they heated the specimen to clean it, because doing so made the tip a little blunter, Melmed says. The process, which is still widely used in many surface science experiments that do not depend on sharp samples, does indeed clean the specimen. But heating also evaporates metal atoms from the tip and thereby blunts it. As a result, in the early days of FIM, the Penn State group kept preparing tips that were too dull for atomic-resolution imaging, which typically requires a radius of less than 50 nm.

On that day 50 years ago, Bahadur must have made a really sharp tip that remained sharp enough to do atomic-resolution imaging even after heating, Melmed remarks. He explains further that part of the success was due to a final preparation step in which Bahadur briefly ramped up the potential above the field ionization voltage—sweeping the sample in early FIM jargon—to complete the cleaning process.

Later on, according to Melmed, Mller discovered that sweeping the sample causes field evaporation of metal atoms, which is sublimation induced by a strong electric field. But when Bahadur swept his tungsten sample on that historic day, he did not yet understand precisely how the procedure improved his sample. But it did so, just the same.

Oct. 11, 1955, is a day long to be remembered by those of us who were in the laboratory, Melmed says. As he recalls the events, Bahadur began the experiment and then went to summon Mller to observe the results with him. Melmed says he and fellow Ph.D. student Russell D. Young watched Mller step into Bahadur's lab; they waited outside the door quite anxiously, wondering what his reaction would be.

Melmed points out that, in those days, microscopes didn't have image intensifiers, so you had to wait about 20 minutes for your eyes to adapt to the dark to see the phosphor screen, and the waiting further heightened the suspense. As Bahadur told Melmed later, when Müller saw the image of the tungsten lattice atoms on the microscope screen, the professor declared, This is it!—meaning here was the ultimate microscopy result he had been striving to achieve. Then, Melmed recalls, Müller emerged from the lab muttering in German, Atoms, ja, atoms!

Over the years, Melmed has confirmed with Bahadur that their recollections of the 1955 events are in agreement. The former labmates discussed them yet again this past June at Penn State, where a conference was held to commemorate the 50th anniversary of atomic imaging and to recognize Müller's scientific contributions.

Looking back to his graduate school days, Melmed says, At the time, we thought Müller would soon be awarded a Nobel Prize. After all, his microscope had made it possible to see atoms.

John Panitz had the same thought several years later. In 1968, while working as a Ph.D. student with Müller, Panitz was part of a team that invented the atom-probe field ion microscope—an atomic-resolution microscope that revealed the chemical identity of individual atoms. The group, which included Panitz, Müller, electronics specialist S. Brooks McLane, and senior lab technician Gerald A. Fowler, coupled a field ion microscope to a specially designed time-of-flight mass spectrometer and demonstrated that imaged atoms could be picked off at will and identified—one at a time—by using pulsed field evaporation.

Shortly thereafter, Müller presented results from the new atom probe at a symposium at NIST, which at the time was the National Bureau of Standards. There was great excitement because people in the field knew the atom probe was coming, Panitz recalls. After Mller gave the talk, he received a standing ovation. It was just amazing. All of us were convinced he was going to win a Nobel Prize.

But it didn't happen. Müller died in 1977, and in 1986 the Nobel Prize in Physics was awarded to Binnig and Rohrer for their STM work and to Ernst Ruska of the Fritz Haber Institute in Berlin for designing the first electron microscope.

In the years following the introduction of the first atom probe, later-generation instruments were built and shown to be unique in their ability to uncover atomic-scale details such as lattice defects, grain boundaries, chemical impurities, and other critical phenomena in metallurgy, materials science, and surface science.

Panitz, who now serves as a consultant in the field after a career at Sandia National Laboratories and the University of New Mexico, notes that many types of materials including metals, alloys, and semiconductors have been subjected to single-atom scrutiny using field-ionization methods.

Despite its tremendous resolving power, the atom probe, like FIM, has mainly been considered a scientific curiosity, Panitz comments, not a one-size-fits-all type of tool with wide-reaching appeal. He says part of the reason the instrument never gained broad popularity is that, until now, there has never been a commercial atom probe that could collect, analyze, and process data quickly and conveniently for a wide range of specialized materials of interest to industry.

But according to Panitz, the outlook for atom probes may be changing: Researchers who study layered structures and other complex materials are discovering that the newest instruments can indeed record vast amounts of data quickly and provide detailed structure images that can be manipulated in three dimensions.

One such instrument, known as the local-electrode atom probe (LEAP), has just been commercialized by Imago Scientific Instruments, Madison, Wis. Imago's founder, Thomas F. Kelly, explains that the instrument's name comes from the design, which makes use of a small electrode that can be scanned close to a sample to extract atoms from sharp points on its surface via field evaporation. Pulsing the electrode at 200,000 Hz rapidly evaporates surface atoms one at a time, thereby exposing the internal three-dimensional atomic structure of the specimen, which is then reconstructed tomographically.

Seidman, who is director of Northwestern's Center for Atom-Probe Tomography, notes that his newly installed LEAP is able to collect data some 700 times faster than conventional modern atom probes. He adds that the instrument enables atomic-resolution analysis from the subnanometer- to the mesoscopic-length scales. There is no equivalent instrument available with these unique capabilities, he asserts.

Recently, the group led by Lincoln J. Lauhon, an assistant materials science professor at Northwestern, teamed up with Seidman's group and others to probe semiconductor nanowires by using LEAP. They demonstrated that individual gold catalyst atoms can be pinpointed in an InAs nanowire and that the composition of the Au-InAs interface can be mapped in 3-D with 3- resolution (Nano Lett., published online Oct. 18, dx.doi.org/10.1021/nl051602p).

Lauhon comments that the performance of future nanoelectronics devices is likely to depend strongly on the structure and composition of buried interfaces, which the atom probe is well-suited to analyzing. If we know the composition of the interface, we can adjust the nanowire synthesis to alter the interface, and hence the device properties, in a controllable way, Lauhon says.

Back in the earliest days of FIM, had Mller and his students gazed in a crystal ball and seen a future in which atomic-resolution images were recorded with ease and practically taken for granted, they probably would have shaken their heads in disbelief. What a difference 50 years makes.