30 years of moving atoms: How scanning probe microscopes revolutionized nanoscience

Thirty years ago this week, scientists turned science fiction into science. By picking up one atom at a time and using a handful to draw the letters I, B, and M on a solid surface, researchers at IBM’s Almaden Research Center in San Jose, California, showed that controlling matter at the atomic scale wasn’t a futuristic dream; it was real, and it was here, and it was now.

For 22 h over Nov. 9 and 10, 1989, Donald M. Eigler and Erhard K. Schweizer used the ultrasharp tip of a scanning tunneling microscope (STM) to pick up 35 xenon atoms and spell IBM in 5 nm tall letters on a cold nickel surface (shown). By the end of the experiment, the IBM team recorded a microscope image of the masterpiece—published several months later (Nature 1990, DOI: 10.1038/344524a0)—that created a scientific sensation and was displayed in newspapers around the globe.

“The experiment showed that the STM could do more than make images of the atoms on a surface; it could also be used to modify the surface atom by atom,” says Christopher P. Toumey, a specialist in nanomaterials at the University of South Carolina. Writing those letters boosted “the credibility of nanotechnology,” Toumey adds, because at the time, scientists were still skeptical of the discipline and the scientific value of the STM itself.

The IBM work didn’t convince just scientists of the technique’s worth, says longtime STM innovator Wilson Ho of the University of California, Irvine. “Anyone on the street could appreciate the IBM image,” he says. “It didn’t take much explaining to make people say, ‘Wow, it’s amazing that you can actually do these things with atoms.’ ”

The feat was made possible by a lot of foundational science, according to Eigler. In the early 1970s, Russell Young, a scientist at the US agency now known as the National Institute of Standards and Technology, took a key step toward manipulating atoms by developing an imaging instrument known as the topographiner. By bringing an electron-emitting tip to within a few nanometers of a solid surface and then switching on an electric current while scanning, the device caused the surface to emit electrons and photons. Young detected these species and converted their emission patterns into surface images with nanometer resolution.

In the early 1980s, Gerd K. Binnig and Heinrich Rohrer, staff scientists at IBM’s Zurich Research Laboratory, took the next big step when they invented the STM. By bringing the scanning probe tip to within less than 1 nm of a surface and applying a voltage, they caused electrons to tunnel between the tip and the surface. Because the tunneling current depends sensitively on the distance between the tip and the surface, the STM generated images with subnanometer resolution.

Just a few years later, when Eigler was using an STM to do atomic-scale imaging, he found that sometimes he “could not get these pesky little atoms to hold still.” He realized that the troubling behavior occurred when he brought the tip to within a few atomic diameters of the atoms. It wasn’t long before he figured out how to control that behavior and deliberately move atoms where he wanted.

These discoveries gave birth to the iconic IBM image, but they also gave birth to innovative science. Within a few years, many research groups got the hang of imaging and manipulating atoms. Before long, scientists were using the STM to drive single-molecule chemistry, to control electrons and quantum phenomena, to build atomic-scale circuitry and data-storage devices, and in some cases, to simply delight the public and shine light on the wonders of science.

Here, we celebrate several of those innovations, highlighting standout images made possible by atomic manipulation.

The quantum corral (1993)

The particle-in-a-box problem is well known among chemistry students taking quantum mechanics. It describes the wavelike behavior of an electron trapped inside an imaginary box. In 1993, Eigler and IBM colleagues Michael F. Crommie and Christopher P. Lutz physically constructed their own box—or in this case, a ring. They used an STM to position 48 iron atoms on a cold copper surface in a ring roughly 71 Å in diameter. The atomically engineered quantum corral confined copper’s surface electrons to the interior of the ring. The now-famous image published at that time (shown) depicts concentric waves—a rippled interference pattern—caused by surface electrons being reflected from the ring of atoms (Science 1993, DOI: 10.1126/science.262.5131.218).

The demonstration deepened understanding of the surface phenomena that underpin atomic-scale data storage and computation, and notably, it converted the well-known particle-in-a-box problem from an abstract concept to an eye-catching illustration.

“Every student who studies introductory quantum mechanics has to solve this classic textbook problem. But it’s abstract—there was never an image,” Ho says. The quantum corral shows what an electron standing wave inside a textbook box looks like, he says.

Good (molecular) vibrations (1998)

By the late 1990s, many researchers had published studies that exploited the STM’s ability to reproducibly manipulate and image individual atoms and molecules. Taking an important step forward, Ho showed that the instrument could also be used to record the vibrational signatures of individual molecules adsorbed on a surface, thereby turning the STM into a chemically specific microspectrometer.

In one key study, Ho and coworkers—at that time based at Cornell University—deposited acetylene on a copper surface, then used a homemade STM to move and image the molecules and to measure molecular vibrations via inelastic electron tunneling spectra (Science 1998, DOI: 10.1126/science.280.5370.1732).

To demonstrate the chemical resolution of this vibrational method, Ho and coworkers juxtaposed an acetylene molecule (C₂H₂) next to a deuterated analog (C₂D₂) on a surface. They published images (shown) pinpointing the exact positions of those two molecules.

Then they rescanned the area twice, once in a mode that selectively excites the C–H stretching vibrational mode of acetylene—causing C₂H₂ to stand out prominently and C₂D₂ to become invisible—and then in a way that excites the C–D stretching mode—causing the opposite effect. Because of the imaging conditions, each molecule appears as a single peak or valley—the atoms are not resolved.

“Nowadays, scanning probe studies are almost entirely about the spectroscopy,” says Andreas Heinrich of the Center for Quantum Nanoscience at Ewha Womans University in Seoul, South Korea. He emphasizes what an important milestone Ho’s work was in the STM landscape. “You still see an occasional high-resolution image, but it’s almost like a footnote” compared with the valuable information coming from the spectroscopy, Heinrich says.

Seeing is believing (2009)

In the early days of chemistry, researchers could indirectly figure out what molecules looked like by measuring X-ray diffraction patterns and then calculating the molecules’ structures. Around 2009, Leo Gross of IBM Research–Zurich and coworkers used STMs to directly generate such images. Aside from wowing chemists, the studies, which trace molecules’ shapes and the positions of their atoms, give us ways to probe electronic and other properties of isolated compounds. And they give researchers a new technique for benchmarking the results of other structural methods.

In one striking recent case, Gross and colleagues used an STM with a CO molecule attached to its tip to not just image molecules but also to mediate the steps of a chemical reaction known as a retro-Bergman cyclization. Then they used an atomic force microscope to image the reactants, intermediates, and product (shown; Nat. Chem. 2016, DOI: 10.1038/nchem.2438).

Specifically, by delivering a pulse from an STM, they broke the C–Br bonds in a tricyclic compound (9,10-dibromoanthracene), forming mono- and diradical intermediates. Then they zapped the system again, converting the diradical to a bicyclic diyne. In a subsequent step, the group reversed the reaction by pulsing the diyne, which regenerated the diradical and provided an up-close look at the reactive intermediate.

Atoms on film (2013)

If scientists toil for hours on end with state-of-the-art equipment, they usually do it to make scientific discoveries, not movies. But in 2013, Susanne Baumann, Ileana Rau, Lutz, and Heinrich, all at IBM at the time, used their STM to make an animated video, A Boy and His Atom. The feature, which holds a world record for being the tiniest stop-motion film, portrays a cartoon character happily playing with a ball (shown). The team created the frames individually by painstakingly positioning CO molecules on a copper surface (only the oxygen atoms are visible) and imaging each configuration.

Although the project was tied to data-storage research, “the movie itself wasn’t a breakthrough,” Lutz says. Nor was it the goal. “We did it purely for outreach—to attract young people to science,” Heinrich says.

With an inviting, gentle tone, the video, which has nearly 10 million views on YouTube, gives nonscientists a sense of the atomic scale. It points out that each dot is a single atom magnified more than 100 million times, then lets viewers watch the little atom boy dance and play.

Bit by terabit (2016)

One of nanotechnology’s ultimate quests is a practical, rewritable data-storage system in which individual atoms serve as bits of information. In 2016 at Delft University of Technology, A. F. “Sander” Otte and coworkers took a major step toward that goal by designing a memory system based on atoms and holes, which are vacancies between atoms.

Starting with a chlorine-covered copper surface, the team used an automated STM procedure—akin to playing with a sliding-tile puzzle—to push chlorine atoms into predetermined lattice positions on the crystalline surface. The manipulations created atom-vacancy pairs that serve as the ones and zeros of digital bits. By patterning the surface in this way, the researchers stored 1 kilobyte of data that could be read and rewritten automatically. The pattern formed letters and spaces (shown), spelling a portion of physicist Richard P. Feynman’s famed 1959 talk, “There’s Plenty of Room at the Bottom.” The memory system stores about 500 terabits per square inch—a data density that’s roughly 500 times as great as that of commercial hard disks (Nat. Nanotechnol. 2016, DOI: 10.1038/nnano.2016.131).

Otte stresses that the atom-at-a-time data-storage device is a lab demo, not a prototype. “Given STM’s extreme vacuum and temperature requirements and its slow operating speed, I cannot foresee a way in which it could ever become an economically viable manufacturing tool,” he says. The knowledge gained from these experiments, however, might lead to new engineered materials that make a technological impact, he says.

Giving electrons a bump (2018)

More than 40 years ago, scientists proposed building molecule-sized electronic circuits. Last year, researchers in Canada demonstrated an automated STM-based patterning system that makes such circuits from individual silicon atoms functioning as electronic bits. Led by Taleana Huff and Robert A. Wolkow of the University of Alberta, the team used the method to make logic gates, the basic building blocks of digital circuitry (Nat. Electron. 2018, DOI: 10.1038/s41928-018-0180-3).

To build their system, the researchers started with an industry-standard hydrogen-capped silicon wafer. A voltage pulse from an STM tip removes a hydrogen atom from the wafer to expose a dangling bond—an electron vacancy—on one silicon atom. The computer-controlled system repeats the procedure, forming pairs of dangling bonds on adjacent atoms (shown here as depressions). Then, it deposits a single electron (red) into each pair. To switch the electron’s position from one atom of the pair to the other, the team adds to the surface yet another electron, which bumps the first electron to the next dangling bond through repulsive forces. Placing the electron on one atom of the pair or the other represents the one and zero of digital logic. Using electrostatics instead of conventional current to drive logic operations eliminates most heat-generating steps, leading to a fast computing system that draws very little power.

“The scanning tunneling microscope continues to amaze me even after 25 years in this field,” Heinrich says. Not only can scientists use it to place atoms where they want, but they can also now use it to manipulate electrons. Ho also continues to feel that excitement decades after he began developing innovative ways to use scanning probes. He points out that the scientific literature is filled with ongoing STM advances related to quantum computing, tip-enhanced molecular spectroscopy, and new imaging methods based on magnetic single-molecule sensors.

“It always surprises me what it can do,” Ho says. “With STM, there’s always something new.”