When Dan Rugar first heard about an idea for a new microscope that could peer beneath the surface of molecules and pick out individual atoms, he was skeptical. Scientists had dreamed of having such a device to help them unravel the complex structures of proteins, spot the defects in semiconductors, and solve a thousand other mysteries. But so far nobody had come up with a way of building a microscope powerful enough to produce a three-dimensional image showing the precise location of each and every atom--without destroying or changing the structure of the material.
Rugar knew the problem as well as anybody. As a physicist at IBM, he had helped develop the atomic force microscope (AFM), which uses a tiny mechanical cantilever to feel individual atoms on the surface of a sample--or more precisely, to feel the electrostatic repulsion exerted by the atoms’ electrons. Indeed, the proposal Rugar was listening to that day in 1991 sounded remarkably familiar.
In the new microscope, as in the AFM, a cantilever would be moved across a sample like a microscopic phonograph needle. But instead of responding only to surface atoms, this cantilever would pick up the far feebler magnetic force exerted by resonating atoms below the surface. That technique is known as magnetic resonance imaging, or MRI; hospitals already use it every day, in a much cruder form, to image internal organs.
After Rugar ran a few calculations of his own, his skepticism melted away. It might indeed be possible, he realized, to build a magnetic resonance force microscope, or MRFM--a device that would combine the atom- scale resolution of an atomic force microscope with the three-dimensional imaging capability of an MRI scanner. Rugar and his colleagues at IBM’s Almaden Research Center in San Jose, California, began working on the idea right away.
Today a prototype sits in their lab--a very early prototype. It can measure a force that is one-millionth the strength of that measured by an AFM. In so doing it can see beneath the surface of a sample and detect features far smaller than an MRI scanner can. But it can’t see nearly as small as Rugar wants. We’re not at atoms yet, he says. That’s our goal. And if you go through the physics of how it works, it looks as though it may be possible.
The inspiration for the MRFM, and the man sitting in Rugar’s office that day four years ago, was John Sidles, a medical physicist in the department of orthopedics at the University of Washington in Seattle. The payoff that Sidles was seeking was an understanding of the molecular mechanisms at work in such intractable diseases as bone cancer. But the desire to see atoms is pretty widespread in science and engineering today. Biologists would like to plumb the intricate folded structure of proteins, whose shapes are crucial to their myriad functions in our bodies. Semiconductor manufacturers would love to have a way to pinpoint not only atom-size defects but also the atoms of boron or phosphorus that are used to dope a silicon chip with the desired electrical properties. The list of possible applications for a three-dimensional atom-seeing microscope is long.
Conventional microscopes aren’t up to any of these jobs, for obvious reasons. Light microscopes can see bacteria, and electron microscopes can see viruses, but neither can see atoms; the diameter of a hydrogen nucleus is just one angstrom, or one ten-billionth of a meter. And the various scanning probe microscopes, of which the AFM is one, don’t see in three dimensions; they see only surface atoms. Besides, they can’t distinguish one type of atom from another, which you have to do if you want to figure out the structure of a molecule.
To be sure, researchers can work out the structure of some materials, such as proteins, by bombarding them with X-rays and observing the scattered rays, a technique known as X-ray crystallography. But that technique works only on crystals, and not all proteins can be crystallized. Another technique, called nuclear magnetic resonance spectroscopy, can only see the structure of small molecules. Says Sidles: There are no truly general-purpose techniques for studying molecular structure at the angstrom level.
Sidles thought that the principle of magnetic resonance imaging held the key to penetrating below surface atoms. I published a couple of theory articles, and then I trotted around looking for experimentalists to do experiments, he says. Perhaps because of his experience on the AFM, Rugar was the only one to take Sidles up on the idea. He enlisted his colleagues Nino Yannoni, an expert in magnetic resonance spectroscopy, and physicist Othmar Züger to develop a prototype.
Sidles’s scheme starts with the basic principle of MRI, which exploits the magnetic moment of protons and neutrons--that is, their tendency to act like tiny bar magnets. When protons and neutrons occur in pairs, each tends to cancel the other’s magnetic moment. But when a nucleus has an odd number of protons or neutrons, the leftover particle imparts an overall magnetic moment to the nucleus. Since hydrogen, for instance, has only one proton in its nucleus, hydrogen atoms placed in a strong magnetic field will tend to align themselves, by virtue of their magnetic moments, with the field. If you then bombard the atoms with a radio wave at a certain frequency, the spinning hydrogen nuclei resonate: they tip over and begin to wobble like tiny tops. If you point your finger up, that’s the way the nuclei are lined up to begin with, explains Rugar. Then if you tilt your finger over, say 30 degrees, and spin it around in a little circle, keeping it at 30 degrees but changing the direction that it’s pointed, that’s the wobbling motion.
When the radio wave is turned off, the nuclei return to their upright state, and as they do so, they emit a weak but detectable radio signal. By measuring the intensity of this signal, scientists can deduce the concentration of hydrogen in a sample. If the sample happens to be a hospital patient lying in the huge magnetic coil of an MRI scanner, the varying concentrations of hydrogen throughout the body produce signals that form an image of the internal organs.
Sidles wanted to see atoms, though, not organs, and to do that he proposed taking advantage of a little bit of MRI physics that the medical scanners don’t exploit. The frequency of the radio wave that will cause hydrogen nuclei in a sample to resonate isn’t fixed; it depends on the strength of the magnetic field that’s bathing the sample. In a medical MRI scanner, that doesn’t matter. The powerful magnetic coils create a uniform field throughout the patient’s body, and when the frequency of the radio waves is tuned to the appropriate level, hydrogen nuclei throughout the body start resonating. Sidles’s idea for an atom-seeing MRFM, though, was to use a magnetic field that wasn’t spatially uniform. That way only those few nuclei would resonate that happened to be in the part of the sample where the magnetic field and the radio waves were in harmony. In principle you could detect a signal from a single nucleus.
In the design that Rugar has since developed, the magnetic field comes from a tiny magnet on the tip of a silicon cantilever much like the one in an atomic force microscope. The field grows weaker the farther away a sample gets from the tip. At any given frequency of the radio wave only those nuclei in a thin slice of the sample, at a precise distance from the tip, will resonate.
The trick then is to get the nuclei to emit signals the microscope can detect. Whereas MRI scanners do this by turning the field on and off and letting the nuclei return to their upright positions, Rugar hit upon an easier method. By subtly varying the frequency of the radio wave, he found, the nuclei in the resonating slice can be made to do flip-flops, reversing their magnetic moments by 180 degrees. When this happens, the atoms go from exerting an attractive force on the cantilever’s tip to exerting a repulsive one and back again. It’s like taking two bar magnets, says Rugar. If you put them together, they might attract each other, but then if you flip them around, they repel. If you flip them back, they attract.
As the nuclei continue to flip back and forth, the cantilever vibrates. To measure this movement precisely, Rugar’s team uses a laser beam that travels through an optical fiber, bounces off the back of the cantilever, and goes back up the fiber. By looking at the amount of laser light that is reflected, the researchers can detect movements of less than an angstrom. As the cantilever scans horizontally over the sample, its vibrations reveal the presence of resonating hydrogen nuclei hidden beneath the surface. And as the cantilever is moved toward and away from the sample, different slices of the sample come into resonance. A computer program then adds all the slices together, building a composite, three- dimensional picture of all the hydrogen nuclei in the sample.
The final step is to locate the atoms of other elements. In principle that’s straightforward. Since each element resonates at a distinct radio frequency, Rugar’s idea is simply to repeat the process at different frequencies. By making a composite of all these measurements, researchers would be able to piece together a three-dimensional image of a sample. Rugar estimates that the microscope would be able to look as deep as about a millimeter below the surface.
All this, however, lies in the future. So far Rugar and his colleagues have developed only a rough prototype that demonstrates the instrument’s basic principles. It consists of a cantilever, a magnet, and a radio-frequency coil sealed in a vacuum chamber, which is in turn immersed in a bath of liquid helium to keep the apparatus at close to absolute zero. The fact is that everything that is at a temperature above absolute zero is vibrating, explains Rugar. Even when there is no magnetic resonance signal, the cantilever will be continuously vibrating in a noisy sort of way. Lowering the temperature makes the noise smaller, so that we can see smaller signals.
The researchers have tested their apparatus on, among other things, a tiny blob of ammonium nitrate--a material chosen because the abundance of hydrogen atoms makes it convenient to work with. At present, the microscope can’t detect the resonance of a single atom; it can detect only the force exerted on the cantilever by a trillion or so atoms at once. Making the microscope able to detect a single atom would be equivalent to improving its resolution by a factor of 10,000, from one-millionth or two- millionths of a meter down to one angstrom. Right now the resolution of the MRFM does not even equal that of the best light microscope.
One problem is that Rugar and his colleagues haven’t yet succeeded in building a crucial part of their design. The tiny magnet attached to the tip of the cantilever is still under development. Instead the prototype MRFM uses an awkward arrangement in which a separate magnet sets up the magnetic field and the sample is attached to the cantilever itself. Once the magnet is completed and put in place on the cantilever’s tip, says Rugar, and the sample is mounted on a fixed slide, resolution should improve.
In addition, Rugar and his colleagues are developing thinner, more flexible cantilevers that bend and vibrate more readily in response to minuscule atomic forces. The instrument’s cantilever is now a mere 900 angstroms thick, but Rugar’s lab has already made one that is 200 angstroms thick. Such a cantilever is so soft that, were it a spring, a paper clip would cause it to deflect one kilometer, says Rugar. Even so, he is now planning to make another one that is half that thick.
Finally, he and his colleagues plan to improve resolution by finding a way to place the magnetic tip much closer to the sample. The strength of the magnetic field falls off most steeply right around the tip, so by bringing the tip as close as possible to the sample, the researchers will be able to generate resonance and detect nuclei in as thin a slice of the sample as possible. The researchers hope ultimately to position a 300- angstrom-wide magnetic tip within 50 angstroms of the sample.
In short, many things need to happen before detecting single buried atoms becomes possible. Rugar isn’t the only one trying to build an MRFM; Sidles, whose idea it was in the first place, has now joined him in a friendly rivalry. Doing MRFM experiments is a little bit like sailing a ship along a rocky and uncharted coastline in a dense fog, Sidles muses. If you are the first ship sailing in these waters, all too often you learn about a rock by going aground on it. Rugar issues the same caution with less metaphor and more optimism. We’re in a very early stage of developing this technique, and it is possible that we will not be able to achieve our ultimate goal, he says. Yet there seems to be no law of physics that says we can’t do it.