Scanning the surface: from gold atoms to benzene molecules, the scanning tunneling microscope probes the intricate structure of surfaces.

Scanning the Surface

"The surface was invented by the devil," said the late physicist Wolfgang Pauli in expressing his frustration with the complexity of surfaces.

Unlike atoms inside a solid, which lie nestled within cocoons of congenial companions, surface atoms stand guard at the frontier between the solid and the rest of the world. They reside in a radically different environment, which means a solid's surface properties differ considerably from those of its interior. That difference has long thwarted attempts at building a precise theoretical and empirical picture of what happens at surfaces.

In the scanning tunneling microscope, researchers now have a surprisingly simple but versatile tool for probing the complexities of surface structures, atom by atom. Already important for investigating the atomic and electronic structure of semiconductor surfaces, the technique reveals contaminants and flaws in the surface structure of materials such as silicon -- information that may be valuable for designing faster integrated circuits. Now scanning tunneling microscopy is destined to play an equally important role in studies of the molecular and chemical properties of a wide range of solid surfaces. Surface structure largely determines many crucial properties of materials, including their reactivity, resistance to corrosion and electronic behavior.

"We can observe which surface atoms have reacted and which have not, and determine how the products of the reaction are distributed on the surface," says Phaedon Avouris of the IBM Thomas J. Watson Research Center in Yorktown Heights, N.Y. "This unique capability brings new insight to our understanding of surface chemistry."

The brief time between its invention and its present widespread use demonstrates the microscope's impact. In 1981, the scanning tunneling microscope was a one-of-a-kind scientific instrument. Five years later, its inventors, Gerd K. Binnig and Heinrich Rohrer of IBM's Zurich (Switzerland) Research Laboratory, received the Nobel Prize for their work (SN: 10/25/86, p.262).

Today, more than 100 laboratories worldwide have either built their own versions of the instrument or purchased them from commercial sources. It's difficult to pick up an issue of a journal such as PHYSICAL REVIEW LETTERS without seeing some mention of scanning tunneling microscopy.

In a scanning tunneling microscope, an extremely sharp metal needle, ideally terminating in a single atom, is brought within a few angstroms of the sample's surface. This distance is small enough for electrons to leak, or tunnel, across the gap between sample and needle and generate an electrical current. As the gat between the tip and the sample increases, the current decreases. A scanning mechanism pulls the needle over the sample's surface, constantly adjusting the tip's height to keep the current constant.

The needle's bobbing journey produces a sketch of the surface's microscopic contours -- a kind of topographic map in which the hills and valleys represent arrays of atoms. The instrument is so sensitive to the separation between tip and surface that it can detect differences in height equivalent to one-hundredth of an atomic diameter. The horizontal resolution depends on the needle's sharpness.

Data from a set of scans initially appear as rows of contour lines. Computer processing of that information converts the lines into filled surfaces, rendered in shades of gray as if illuminated by some ethereal light source. Sometimes adding color emphasizes important details or produces an elevation map in which different colors represent different heights.

Interpreting the images is still an art. The microscope doesn't identity what atoms are present. Moreover, the tunneling current observed depends on a number of factors, including the nature of the surface, the needle's geometry and composition, and the electronic properties of both the surface and anything that sits atop it. In addition, the needle itself may sometimes modify the surface, gouging it by pushing aside atoms or molecules.

The microscope works best with substances that readily conduct electrons. For this reason, much early work focused on materials such as silicon, gallium arsenide and graphite (SN: 9/6/86, p. 149). On the other hand, many biological materials transfer electrons poorly and can't be examined in a straightforward manner using the scanning tunneling microscope.

When researchers started making images of graphite and gold surfaces in air, the absence of contaminating molecules that were bound to settle on exposed surfaces puzzled them. These molecular visitors seemed invisible to the scanning tunneling microscope.

Investigators speculated that perhaps such molecules diffuse along the surface too rapidly to be caught by the microscope. Alternatively, the microscope's needle may physically push molecules aside, or the molecules themselves don't participate in the electron tunneling process. Further studies showed all three mechanisms at work, and researchers soon developed methods for circumventing such problems and capturing images of otherwise elusive visitors.

One recent success was the detection of benzene and carbon monoxide molecules -- both highly mobile and electrically insulating -- on a clean metal surface. Robert J. Wilson and his colleagues at the IBM Almaden Research Center in San Jose, Calif., got their images by depositing a layer of closely packed benzene and carbon monoxide molecules on a carefully smoothed rhodium surface. The carbon monoxide molecules help wedge the benzene molecules in place to keep them from moving along the surface and blurring the image. Having two different types of molecules packed together in the same array allowed the researchers to study for the first time how strongly different molecules show up in a scanning tunneling microscope image under the same experimental conditions.

Images of the structure revealed individual benzene molecules as threefold ring-like features, confirming the traditional picture of benzene molecules as rings of six carbon atoms. The researchers also observed the movement of individual benzene molecules from site to site within disordered regions. Though barely visible, carbon monoxide molecules appeared as very small protrusions between benzene molecules within ordered regions.

Wilson and his colleagues also made the first tunneling-microscope images showing the internal structure of an isolated molecule on a surface to illustrate this technique's potential for observing individual molecules. Earlier experiments had produced detailed images of molecules only in closely packed arrays. In this case, the researchers looked at a distinctively shaped molecule called copper phthalocyanine, a blue pigment, "mounted" atop a metal surface.

Images of copper phthalocyanine molecules deposited on silicon or graphite surfaces are blurred, indicating that weak bonding or chemical interactions between the molecules and the surface allow the microscope's needle to shift the molecules. In contrast, microscope images of a copper surface clearly reveal the molecule's fourfold symmetry--a pattern resembling a four-leaf clover. Moreover, although the phthalocyamine molecules lie flat, they prefer to rest among the neatly ordered copper atoms in one of two particular orientations. Wilson and his co-workers report their observations in the Jan. 9 PHYSICAL REVIEW LETTERS.

Tycho Sleator and Robert Tycko of AT&T Bell Laboratories in Murray Hill, N.J., were among the first the study in detail individual organic molecules at a crystal surface. They worked with the electrically conducting organic molecule tetrathiafulvalene tetracyanoquinodimethane (TTF-TCNQ). At low magnification, the researchers saw a terraced surface with steps about 10 angstroms high. Each step appeared to correspond to the addition or subtraction of a single molecular layer. Under a higher magnification, they detected parallel rows of objects, each consisting of a large ball, about 3 angstroms in diameter, with two smaller balls on either side. The researchers interpreted these structures to be the TCNQ parts of the molecules, in which the large ball is a ring of six carbon atoms and the smaller balls are nitrogen atoms.

High-temperature superconductors haven't escaped the probing gaze of the scanning tunneling microscope as scientists seek to demystify their unusual characteristics. Michael D. Kirk and his colleagues at Stanford University studied a superconducting concoction consisting of bismuth, strontium, calcium, copper and oxygen. They found evidence for a complex crystal structure that may play an important role in determining the material's superconducting properties. Their results appear in the Dec. 23 SCIENCE.

Normally, crystals have a regular structure consisting of identical groupings of atoms, called unit cells, that stack together like building blocks. X-ray scattering experiments show that in the bismuth superconductor, the unit cells do not fit together precisely.

By obtaining tunneling-microscope images of the superconductor surface, the researchers discovered that every ninth or tenth row of bismuth atoms is missing from the regularly repeating pattern of atoms in the material. The absence of the bismuth causes nearby oxygen and copper atoms to shift out of their usual positions, distorting some of the unit cells. Because superconducting properties are very sensitive to the positions of copper and oxygen atoms, such distortions may affect the material's superconductivity. Whether the "defect"