Seeing Nanoscapes A family of tools for seeing and manipulating atoms and molecules is moving out of the lab and onto the factory floor.
By Ivan Amato

(FORTUNE Magazine) – One summer day in 1985, not that long before he shared the Nobel Prize in Physics for a device he had invented five years earlier for seeing atoms, known as a scanning tunneling microscope, IBM scientist Gerd Binnig was lying on a sofa in a Palo Alto apartment staring up at a cheap, stucco-like ceiling. As he lounged there, he imagined each bump on the ceiling as an atom. What followed was an "Aha!" moment that continues to reverberate through the scientific and business worlds.

Within weeks Binnig, then 38 and with an infant daughter, worked with a colleague to build a prototype for what has become one of the most versatile scientific tools ever created: the atomic force microscope, or AFM. A few years later the first commercially available version went on sale. AFMs have since helped propel the science and engineering of nanotechnology to what many now view as a latter-day industrial revolution.

Scientists and engineers today are using AFMs to view just about everything they can lay their hands on--biological cells, chemical catalysts, computer chips, even sunscreen agents and textiles. All that is driving a potentially huge new industry. The National Science Foundation projects that nanotechnology, which refers both to the know-how to exploit the properties of atomic-and molecular-scale structures and to the products made from them, will constitute a $1-trillion-a-year megamarket by 2015. By comparison, semiconductor sales last year amounted to $166 billion.

One trillion dollars is an enormous number, yet experts believe that nanotech is destined to transform everything from medicine and electronics to transportation and pollution control. "Eventually all of industry is in its path," says Ed Braun, chairman and CEO of Veeco Instruments in Woodbury, N.Y., the world leader in AFMs. At Pacific Nanotechnology, a smaller AFM maker in Santa Clara, Calif., vice president Paul West says, "We have customers using our AFMs to look at rubber tires and paper and razor blades."

Plenty of companies are investing increasing amounts in nanotechnology R&D, but the foundation for this growth is being laid by multibillion-dollar government-sponsored nanotechnology projects around the world, including a $3.7 billion, four-year program in the U.S. They are designed in part to underwrite purchases of AFMs and other big-ticket tools by government and academic research laboratories. At the same time, a potentially vast new crop of customers in industry sectors ranging from aerospace to information technology have begun buying AFMs to help them improve and control manufacturing processes. All told, the AFM market now approaches $200 million. BCC, a technology market analysis firm in Norwalk, Conn., estimates that demand will grow at an average annual rate of more than 22% through 2008. Supporters believe that examining products on atomic and molecular scales--the nanoscale--will reveal technical clues that ultimately will give them a competitive edge. It might be possible to make a tougher, longer-lasting tire, for example, by discovering some nanoscale surface property in the carbon black used to toughen the rubber.

The roots of nanotechnology are often traced to a seminal lecture in 1959 by Nobel physicist Richard Feynman on the future of miniaturization in manufacturing. But only in the 1980s did the field start to take on an air of genuine possibility, when tools like the AFM gave scientists unprecedented access to both viewing and manipulating the tiny atomic and molecular building blocks that make up all physical things. A single strand of DNA spans about two nanometers, or two-billionths of a meter, while a red blood cell measures about 5,000 nanometers and a strand of hair about 75,000 nanometers.

Nanotechnology is already here-and-now in the data-reading heads of computer hard drives. They rely on a property called giant magnetoresistance, which makes the head supersensitive to the tiny magnetic signals associated with each bit of data. To create the heads, manufacturers must deposit a dozen or more exquisitely thin layers--some only a nanometer or so thick--of different materials in a precise sequence. Another current application is computer chips, whose transistors are made with features that measure 90 nanometers across, roughly one-hundredth the span of the smallest visible dust speck. In the future, claim some visionaries, are cell-repairing nanomachines and pocket-sized data-storage devices that will be capable of carrying the text of every book ever written.

AFMs are to nanotechnology what microscopes have always been to biology--they extend researchers' vision to previously unseeable dimensions. Instead of relying on light and lenses, however, AFMs generate images using what amounts to a tiny, stylus-like probe. It assembles a picture of a molecular or even atomic landscape by using what is effectively a scaled-down sense of touch. Sitting in the lab, most AFMs resemble ordinary tabletop microscopes, though they often are in specialized housings that isolate them from acoustic and other mechanical vibrations that would make seeing atoms impossible. Rather than peer through an eyepiece, users watch a screen as a computer makes visual sense of the signals coming from the AFM's ultrasharp tip.

Just as carpenters are well advised to measure carefully before they cut, nanotechnologists need to precisely measure their sub-Lilliputian parts and products to make sure they're actually producing what they've set out to make. "For manufacturing to exploit nanotechnology, you need measurement tools that can work on those scales," says Joseph Stroscio of the National Institute of Standards and Technology in Gaithersburg, Md. That's where he has built one of the world's most advanced scanning tunneling microscopes, a kind of device that set the stage for AFMs. Not for sale, Stroscio's device serves as a test-bed for pushing nanoscale measurement and manipulation techniques forward so that industry will be prepared when it needs those abilities.

To date, about 20 companies have sprung up to manufacture AFMs. Veeco is the giant among them. It sold some $120 million worth of AFMs last year, accounting for roughly two-thirds of the world market. Most sales have been to scientists doing basic research, but at least 10% of the company's units have been going to quality-control specialists, process engineers, and others who work on factory floors in industries that include semiconductors, data storage, and display technologies. A typical mid-range AFM research system sells for between $100,000 and $200,000, while newer factory-floor systems range up to $2 million each. Of an estimated total of 10,000 AFMs, Veeco claims responsibility for more than 7,000.

Veeco also makes equipment for depositing and analyzing nanoscale layers of various materials on wafers for making computer chips, light-emitting diodes, and other devices. Last year it lost $9.7 million, but its AFM business always has been profitable, Braun says. He notes that his company's AFM sales have quadrupled from $30 million in 1998, and he expects them to reach $400 million within five years. Propelled in part by its AFM success, Veeco stock has shot up about a third in the past 12 months.

Demand for AFMs is driven by government spending and research investments by companies like the Dows and DuPonts of the world, whose business plans are rife with nanotechnology. A still largely untapped yet vast market for the AFM industry, say Braun and others, is manufacturers that need tools to see and manipulate atoms and molecules to monitor and control nanomanufacturing processes. Intel, Samsung, IBM, and Toshiba have been buying AFMs to measure critical nanoscale features of raw silicon wafers and their microcircuitry.

The size of defects that can wreck chips keeps shrinking; that means the level of wafer perfection has to go up. One vision of the chipmaking future is to have AFMs on the factory floor, so that every wafer goes through an AFM system to check the surface quality. Of the 80 to 100 AFM systems that Veeco ships per quarter, ten to 15 of their new, high-end, online models go to semiconductor companies.

The principle behind all those AFMs derives from Binnig's stucco-staring moment in 1985. That spark of innovation had actually been kindled about five years earlier, when Binnig, along with Heinrich Rohrer, a senior colleague at IBM's Zurich Research Laboratory, bowled over the science world with their invention of the scanning tunneling microscope, the precursor to the AFM. Like no instrument before it, the scanning tunneling microscope enabled scientists to visualize the world all the way down to its very atoms. So scientifically powerful was it that it took a mere five years for Binnig and Rohrer to win a Nobel Prize for its invention.

But tunneling microscopes had an Achilles' heel that nagged Binnig: They could examine only samples that were electrically conductive. If the sample wasn't a metal or a semiconductor like silicon, it couldn't be seen. The reason is that the microscopes work by means of a so-called tunneling current--a trickle of electrons that flows, via quantum-mechanical processes, through the tiny space between the microscope's supersharp tip and the underlying sample. As the tip scans back and forth over the sample's nanoscape like a blind man wielding a walking stick, the tunneling current fluctuates. It increases when the tip is close to the conductive surface and decreases when it is farther way. A computer tracks the changes and recasts the raw electrical data into a map of the sample's atomic landscapes. Image-processing programs then massage these maps into often arrestingly beautiful steps, corrugations, and egg-carton moguls reminiscent of fantastic topographies seen in National Geographic photos.

For the scientific community, the images were portals to discovery. "Anytime you can create images on that scale and look at things at those dimensions, it is as new a world as Mars is," says Michael Myrick, associate director of science at the University of South Carolina's Nanocenter in Columbia. What Binnig and many other scientists craved in the mid-1980s was a similar instrument that would work on insulators, biological cells, and pretty much anything else. In that way the new luxury of visualizing atoms--which physicists were doing in their studies of silicon wafers, superconductors, metals, and other conductive solids--could be extended to the rest of the scientific community.

With that in mind, Binnig envisioned a stylus dragging up and down over the stucco bumps, physically deflecting as it parried with the surface's ups and downs. He knew instantly that if such an instrument could be built, it could render an atomic-scale image of just about any surface, not just electrically conductive ones. In a flash, he had figured out the basic design of the atomic force microscope.

In a matter of weeks Binnig teamed up with his IBM colleague Christoph Gerber and Stanford University collaborator Calvin Quate to work out the new tool's operation, and Gerber built the first prototype. As word of the new tool began to spread in the scientific community, many researchers hungry for nanoscale visualization began trying to make AFMs of their own. By 1987 two of those scientists had founded Digital Instruments; two years after that they introduced the first commercial AFM, called the Nanoscope.

Whether homemade or purchased, the first AFMs were far from easy to operate. "In those days the AFM was not a tool you used in your research; it was your research," says Myrick. "You had to be an artist to use one." Myrick learned the hard way: When he first got his hands on an AFM in 1989, he tried creating images of DNA molecules with enough clarity to directly read their genetic codes. The experiment failed because the instruments were too crude to see subtle differences between similar molecules.

Although in theory the first AFMs could be used to image atoms, in fact only imaging of larger molecule-scale objects was possible. It took until the early 1990s and a series of incremental developments before the instruments could produce atomic-resolution images. They included improvements in tip design, motion control for the probe, and reduction in vibration from the building and even from sounds in the laboratory. That's about when sales at Digital Instruments began to take off and when a number of other companies began to enter the business.

What has made the AFM a darling among nanoresearchers is its chameleon nature. By changing the design of the tip and the patterns of scanning motions, a single instrument can be made into multiple instruments (collectively known as scanning probe microscopes, or SPMs) for mapping many physical properties--including magnetism, frictional traits, and electrical charges--at nanoscales. It's akin to the way you might use all of your senses to see how colors, shapes, aromas, tastes, wetness, dryness, heat, texture, and other features are distributed on your dinner table.

Consider the most basic AFM operation, known as contact mode. Somewhat like sweeping a finger over Braille writing, the technique discerns the nanoscapes of an underlying sample. Buy a probe whose tip is coated with a magnetic material, however, and you have yourself a magnetic force microscope that maps out a surface's magnetic field, molecular patch by molecular patch. That's handy if you're in the data-storage industry and you want to take a close look at some surface flaws that have been ruining your disks. Substitute a metal-coated silicon tip, and your AFM becomes an electrostatic force microscope in which the probe deflects in response to electric fields. That's good for mapping locations in a chip's microcircuitry that could be prone to failure.

"The last count I saw, there were 20 or so different [scanning modes]," says Clayton Teague. He directs the National Nanotechnology Coordinating Office, which serves as the hub for government-sponsored nanotechnology programs sprinkled throughout agencies such as the National Science Foundation and the Department of Energy. He adds, "As the critical dimensions of various technologies decrease to the tens of nanometers range, the need to bring SPMs in to be the workhorses of inspection is going to go up."

Even as Veeco and other AFM companies strive to push their instruments into manufacturing settings, they are expanding the ways the tools can engage the nanoworld. AFMs now are used as conduits not just for "feeling" atomic and molecular landscapes but for sculpting them and for writing on them. The tools can serve as fingers for making nanoscale structures, perhaps quantum dots for holding bits of data in tinier spaces, or for "feeling" when a candidate drug molecule docks into a specific cellular receptor associated with cancer or another disease.

Veeco, for example, now offers AFMs packaged with a nanowriting technology developed by Chad Mirkin, a Northwestern University researcher who formed a company called NanoInk in 2001. With modified probes and specialized molecular inks, AFM systems also become tools for literally writing on the nanoscale or, more likely, for spot applications of materials to repair the tiny lines on the photo masks used to make chips.

In another innovation of as yet undetermined value, Pacific Nanotechnology offers a force feedback system made by the Swiss company NanoFeel. With this add-on, users can move the AFM's tip or nanoscale objects using a hand-held interface that converts hand motions in three dimensions into corresponding nanometer-scale motions of the AFM tip. In turn, the forces "felt" by the probe are magnified and fed back into the user's hand. That offers the user, in effect, the ability to feel the very molecular landscape he or she is imaging or manipulating.

With an eye on the as yet untapped market in bioscience and biotechnology, a decade-old company in Tempe, Ariz., called Molecular Imaging unveiled in February its new PicoTrec systems. Co-founder Stuart Lindsay, a professor at Arizona State University, says the AFM-based systems are capable of recognizing specific molecules and mapping their locations, such as receptors on a cell's surface where drug molecules might be able to do their medicine. Molecular Imaging's key innovation is a means of attaching particular molecules, such as antibodies, to the AFM tip, which bends when the molecule interacts with target molecules in the sample. Lindsay envisions a vast market for his AFMs in the thousands of biology laboratories where optical and electron microscopes have long been the primary tools for visualizing cells.

At the National Institute of Standards and Technology, Joe Stroscio dotes on his one-of-a-kind microscope. It's part of an impossibly complex-looking assemblage of steel vacuum chambers, high-voltage boxes, thick glass portals, and adjustment dials, all housed in a vaultlike room and resting on a huge concrete slab that is mechanically isolated from the rest of the building to reduce vibrations. Using it, he and his colleagues have been developing a technique they call automatic atom assembly, or AAA. It's robotic manufacturing writ small: A scanning probe, programmed with a general set of instructions for moving atoms around, nudges atoms, which are peppered onto a surface, into any physically sensible configuration that Stroscio might want to create. It's difficult to figure how such a technique might be put to profitable use anytime soon, but Stroscio did recently extract potential publicity from the gadget by spelling out "CNN" in cobalt atoms for a cable network film crew.

For the moment, Stroscio notes, automatic atomic assembly can work only with a limited number of atoms in just two dimensions, under a vacuum more extreme than that of outer space and at temperatures that would be cold on Pluto. With those provisos, he notes, "you can make a circle, a square--you name it." They are perhaps the kinds of nanostructures that in the future might be assembled by the gazillions or more into, say, Library of Congress chips. Even with its limitations, the fact that automatic atomic assembly works at all makes it at least as promising as Gerd Binnig's stucco daydream was in 1995.

"I would have never thought you could do this," remarks Stroscio, admiring a monitor as an AAA run creates a zigzag pattern of cobalt atoms. "It's equivalent to reaching up and moving planets around with your hand." In nanoscale, of course.