NEW NMR TECHNIQUES ARE CLARIFYING THE STRUCTURE OF SOLIDS
THE PHENOMENON KNOWN AS NUCLEAR MAGNETIC RESONANCE (NMR) takes place at a dimly lit crossroads of science where the ordered grid of geometry, the metaph ysical maze of quantum mechanics, and the concrete highway of materials science all come together.
In that paradoxical place, scientists who work with NMR may find-like Al ice's White Queen-that they have to think about six impossible things before bre akfast. Members of LBL chemist Alex Pines' group, for example, are currently thi nking about conundrums like the following:
Alex Pines began thinking about some of these conundrums more than 20 years a go, when he was a graduate student of chemist John W. Waugh at the Massachusetts Institute of Technology. A paper based on his thesis described the transfer of polarization from hydrogen nuclei (protons) to carbon-13, thus making possible t he first high-resolution NMR study of carbon-13. Today, that germ of an idea is the basis of a promising new possibility for studying the surfaces of solids.
Another idea from Pines' student days-making time appear to run backward s, in a complex system with many interacting particles, by forcing the spins of nuclei into an earlier configuration-is proving useful in a new technique known as ''multiple pulse zero-field" NMR spectroscopy. And his lifelong love of mathe matical puzzles and oddities turns out to have a lot of bearing on the study of glass and other amorphous materials.
Pines' interests may diverge in many directions, but his focus is res ults-oriented. Over the past decade, Pines' team, associated with LBL's Material s Sciences Division and the UC Berkeley Department of Chemistry, has become a le ading center in the research, development, and use of NMR techniques. Among the group's pioneering advances in NMR spectroscopy are multiple-quantum NMR, zero-f ield NMR, and double-rotation NMR. All three techniques have greatly extended th e applications of NMR to new materials.
Now, Pines' team-currently consisting of about 20 researchers, includ ing postdoctoral fellows, foreign visitors, and graduate students-is hard at wo rk on a number of new advances in NMR and the application of all the techniques to new problems.
There are three broad areas of activity under way. First and foremost (because it provides the foundation for all the other work) is pure science-ex periments aimed at the understanding of such phenomena as the nature of the int eraction between radiation and matter, chaotic behavior, nonlinear dynamics, top ology and geometry, and dynamical symmetry. These studies are usually divorced f rom specific techniques but may have important implications for all of them.
Following closely in the trail blazed by the pure science experiments co mes the development of new methods and technologies. There are currently a numbe r of new techniques under development-including dynamic-angle spinning (DAS) and double- rotation (DOR) NMR; surface-enhanced NMR; multiple-pulse zero-field NMR ; NMR imaging of flow and turbulence; SQUID (Superconducting QUantum Interferenc e Device) NMR; and two-dimensional NMR correlation methods.
Finally, there is the application of the various techniques, old and new , to the study of materials (see sidebar, page 18).
Nuclear magnetic resonance is phenomenon that occurs when matter interac ts with a magnetic field. It is based on the fact that the spinning protons (hyd rogen nuclei) in matter oscillate like tiny gyroscopes when they are trapped ins ide a magnetic field.
In NMR spectroscopy, a sample is placed in a magnetic field, which force s the spins of the nuclei into alignment. The sample is then bombarded with radi o-wave pulses. As the nuclei absorb energy from a pulse, they topple out of alig nment with the magnetic field, and as they lose the absorbed energy they line up again. By measuring the specific radio frequencies that are emitted by the nucl ei, and the rate at which realignment occurs, scientists can gain detailed infor mation on the atomic structure and motion of the sample.
But magnetic and electrical interactions between atoms can blur the spec trum associated with the spinning nuclei. For a long time, the application of co nventional NMR spectroscopy was limited to liquids (where the rapid, random moti on of the molecules averages out blurring effects) and to certain crystalline ma terials.
Many of Pines' contributions, over the past decade, have been aimed at o vercoming these limitations and extending the use of NMR to new areas and new ma terials- particularly solids and amorphous materials. Now, an approach involving surface-enhanced NMR is leading to an extension of the possibilities of NMR in the analysis of the surfaces of solids.
''Surfaces are important," says Pines, ''because lots of chemistry takes place there. But NMR has not proven useful in the study of surfaces, except in systems with very large surface areas. There are two main reasons-one absolute, one relative."
The ''absolute" reason is the fact that surfaces constitute only a tiny fraction of any material's volume. "NMR is a relatively low- sensitivity techniq ue," notes Pines, "so when it's combined with the small percentage of surface, y ou don't get much signal."
Added to this ''absolute" problem is the ''relative" one that even if an NMR signal from the surface is detectable, it is overwhelmed by the signal comi ng from the bulk. And some phenomena-catalysis, for example-are strictly surface phenomena, scarcely involving the bulk material.
The way around both of these problems may lie in the fusion of two ideas. One is the optical-pumping technique developed by William Happer (now director of D OE's Office of Energy Research) and his co- workers at Princeton University. The other is the idea of transferring spin from one kind of particle to another-a p henomenon first demonstrated in Pines' ''cross-polarization" work with carbon-13 at MIT almost 20 years ago. In this case, cross-polarization-based on the pione ering work of Elwin Hahn and Sven Hartmann at Berkeley-is used to transfer spin from polarized light to a gas and then to the surface of the material of interes t.
Happer developed a two-step, optical-pumping process that begins with ci rcularly polarized light (photons) from a laser and transfers the angular moment um of the photons to the nuclear spin polarization of xenon gas, resulting in a highly polarized gas. Adapting Happer's technique, Pines' team then brings the g as into contact with the surface of interest. The gas is adsorbed to the surface , forming a thin film of solidified xenon that yields high-intensity, characteri stic NMR signals.
Says Pines, ''The next-and crucial- step will be to see if we can transf er the spin one more time-from the adsorbed film of xenon to the protons of the surface we're interested in. The real pay-off, of course, will be to transfer th e polarization and show that the NMR spectrum of the surface is now different fr om the spectrum of the bulk. If we can do that, we will have a way to distinguis h surface from bulk, and focus in on what's happening on the surface."
Another major thread of research in Pines' group is aimed at improving t he ability of NMR spectroscopy to be applied to solid materials of various types .
Zero-field NMR was pioneered in Pines' lab in the early 1980s. This tech nique allowed NMR to be applied to amorphous substances like glass, and polycrys talline ones, like zeolite catalysts. In conventional NMR spectroscopy these mat erials yield a broad, relatively featureless spectrum that contains little or no structural information. That happens because each 'crystallite,' or individual grain, of the substance has its own random orientation with respect to the magne tic field.
By means of zero-field NMR, in which orientation to an external field pl ays no role, one obtains sharp, crystal-like spectra, even in orientationally di sordered materials.
Now, an extension of the technique, known as multiple pulse zero-field ( MPZF) NMR spectroscopy, is under development in Pines' lab. The new technique in volves manipulating the spinning particles in such a way that their spins-and, i ndirectly, the direction of time-seem to run backwards. While the sample is m th e ''zero-field" condition (that is, out of the strong external field) a much wea ker magnetic field is rapidly pulsed on and off. This pulse affects the internal dipole field and causes the spinning protons to precess-like tops winding down- in such a way that they reverse direction.
When the spectrum is finally measured in high field, the ''memory" of th e time-reversed state causes a dramatic reduction of the number of lines in the spectrum, permitting scientists to see details of structure that were previously hidden.
Two other important new techniques- double rotation (DOR) and dynamic -angle spinning (DAS)-use ideas rooted in pure mathematics to extend NMR spectro scopy to a whole new class of important solid materials.
Like zero-field NMR, both DOR and DAS seek to ''make a solid behave more like a liquid"-that is, to give solid materials the symmetry with respect to an exte rnal magnetic field that liquids have by virtue of the rapid, random motion of t heir molecules.
DOR and DAS make use of complex spinning techniques to achieve the symme try of the icosahedron-a geometric figure with 20 faces and 12 vertices that rep resents an appropriate approximation to the ideal symmetry of the sphere.
While the problem with NMR spectroscopy of solids is the broadening of spectr al lines due to the fact that each atom has its own individual orientation to th e external magnetic field, in a liquid symmetry is spherical: The tumbling of th e atoms causes all the orientations to average out. But there is no way to make a ''lump" of solid matter spin so as to achieve spherical symmetry.
It turns out, however, that for nuclei that are themselves spherical, cu bic (fourfold) symmetry is a close enough approximation to the sphere to yield s atisfactorily narrow NMR lines. Cubic symmetry can be attained through a techniq ue known as magic-angle spinning, originally developed by Raymond Andrew and Irv ing Lowe in the 1950s.
In magic-angle spinning, a sample is rapidly rotated around an axis that is i nclined at a carefully chosen ''magic" angle (the angle between the body diagona l and the face of a cube, or 54.74 degrees), with respect to the external magnet ic field. This spinning motion approximates the natural tumbling that would occu r in a liquid.
The technique has proven extremely useful for those materials in which n uclei are spherical. In carbon-13 and silicon-29, for example, resolution is imp roved 100 times or more.
But in nuclei that are not spherical (also called quadrupolar nuclei), c ubic symmetry is not enough. Such nuclei have electric as well as magnetic momen ts, which complicate the picture. In these materials (which include such importa nt species as oxygen-17, sodium- 23, and aluminum 27), magic-angle spinning impr oves resolution only three or four times-usually not enough to permit the differ entiation of atomic sites.
We realized that the logical next step was to go to higher symmetry," says Pines, "but we also knew that such symmetry could never be achieved by rot ating the sample about a single axis." It became clear that one must rotate the sample around two axes, which can be carried out in two ways: double- rotation ( DOR) NMR and dynamic-angle spinning (DAS), based on icosahedral symmetry.
In DOR, the sample is spun around two axes at the same time, like a top within a top. The result is two ''magic angles"- the conventional one of 54.74 d egrees plus a new magic angle of 30.56 degrees.
In DAS, the sample is spun sequentially about two different axes, inclin ed to the magnetic field at angles of 37.38 and 79.19 degrees; the sample is rot ated about the first axis, reoriented, and rotated about the second axis. A comp uter program then analyzes the data, providing a high-resolution spectrum.
Based as they are on basic mathematical principles of icosahedral symmetry, D AS and DOR may prove particularly valuable in the study of local order in glass and other disordered materials which are thought to involve ''frustrated" icosah edral structures. In theory, such structures shouldn't exist at all because icos ahedrons, with their five-fold symmetry, can never uniformly fill a three-dimens ional space-hence the ''frustration."
There is also considerable controversy about the structure of quasicryst als which involve icosahedral structures. X-ray diffraction (until now the only available technique) shows only the overall, long-range order of the whole syste m. Dynamic-angle spinning (DAS), on the other hand, can reveal areas of local or der, and thus may soon provide answers to some important questions.
Pines switches easily between the ivory tower and the real world. He is particularly proud of his group's accomplishments in the area of technical innov ation and technology transfer. He notes that two companies have already licensed systems based on DOR and DAS-just as instruments based on the earlier mass-pola rization and multiple- quantum techniques long ago entered the marketplace. DOR was honored last year with an ''R&D-100" award, presented by the publication Res earch and Development for outstanding achievement in technology.
The most valuable form of technology transfer, though, may be a human on e-the steady stream of former students and postdocs from Pines' team who enter i ndustry. They are as likely to be found at AT&T Bell Laboratories, IBM, Exxon, M onsanto, Du Pont, Squibb, GE, or Lockheed as they are at UC Berkeley, Yale, MIT, Princeton, Caltech, or the Weizmann Institute.
-JUDITH GOLDHABER