Norman Ramsey

Contributing to other Sciences by Lucky Accidents

Tuesday, 28 June 2005
15:00 - 17:00 CEST


Magnetic resonance originated in one limited field of research, but subsequently made important contributions to many other scientific and applied fields. In this talk I shall first briefly discuss the scientific basis of magnetic resonance and the first successful experiments. I will then discuss in turn the subsequent discoveries and inventions that led to contributions of magnetic resonance to so many other disciplines.

I.I. Rabi invented the molecular beam magnetic resonance method in 1972, just two months after I began working for my Ph.D. with him at Columbia University. Immediately, six of us in his laboratory dropped all other activities to adapt two of our apparatuses to his new invention. Magnetic resonance depends on the facts that electrons and most nuclei are spinning like tops with angular momentum J and are magnetized like a compass needles with magnetic moments.Just as a tilted top in the earth’s gravity field will precess about a vertical axis, the J of a nucleus or electron in a magnetic field B will precess about that field with a frequency 0. Rabi’s proposed experiment was to apply an oscillatory magnetic field at frequency and to vary in hopes of finding a resonant change in the orientation of when at resonance = 0. Initially we tried this with 7Li in LiCl and with molecular H2. The experiment with 7Li worked as we had hoped and gave the hoped for changes of state at the resonance frequency = 0 , from which we could calculate. In the next few years many nuclear magnetic moments were measured in this way. In subsequent years many nuclear magnetic moments were measured in this way. At the time of these measurements we knew that our nuclear magnetic moment measurements had to be corrected for magnetic shielding from the circulation of the electrons in the molecule induced by the external magnetic field. W. Lamb calculated this correction for single atoms but not for molecules. After WWII I developed a method for calculating this magnetic shielding correction for polyatomic molecules and the shielding corrections had the interesting property that the corrections were different for the same atom in different locations in the same molecule. At about the same time this chemical shift was independently discovered experimentally by N. Bloembergen , Bloch and others. In contrast to the success of the first 7Li measurements, the experiment with H2 was initially very disappointing.

Instead of the expected sharp resonance we obtained a broad pattern looking like noise. However when we greatly reduced the strength of the oscillatory field we found that we had six resonance frequencies due to the proton magnetic moment also interacting with the magnetic field from the other proton and from the electric charges of the rotating molecule. In other words we were observing In other words we were observing the radio-frequency spectrum of the molecule. This extended the contributions of magnetic resonance to the fields of molecular structure and chemistry. When we studied the radio-frequency spectrum of D2 we hit another surprise. The separation of the spectral lines in D2 were greater than in H2 even though the spin- spin and the rotational magnetic fields should be much less. We finally interpreted this as due the deuterium nucleus having a quadrupole moment (being ellipsoidal in shape) which gave rise to a spin dependent electrical interaction. The existence of the quadrupole moment, in turn, implied the existence of a new elementary particle force called a tensor force.

The magnetic resonance method was then extended to atomic physics to measure accurately the magnetic interaction between the atom and electron which in atomic spectroscopy was called the hyperfine separation frequency, hfs. Since hfs is dependent only on interactions internal to the atom, it could be used as a standard for frequency and time, but such atomic clocks would then have been no more accurate than clocks based on vibrating crystals. Since the magnetic resonance methods then in use had the great disadvantage that for high accuracy the oscillatory field region should be as long as possible and the wave length of the radiation should be as short as possible, but these two requirements were mutually incompatible since the oscillatory field region had to be less than one-half a wave length long to avoid cancellations by opposite phases. In 1949 I invented the separated oscillatory fields method which overcomes this problem by confining the phase coherent oscillatory fields to two separate regions, each of which is shorter than one half a wave length, but with their separation being many wave lengths apart, so the resonance width is correspondingly narrow. This greatly increased the accuracy of the frequency measurements so that atomic clocks were more stable than any other time or frequency standard.

In addition it had the advantage over all previous standard of being determined by internal atomic properties which by the quantum mechanical identity principle were the same everywhere. As a result the international definitions of both the second and the meter are now based on magnetic resonance atomic Cs clocks and atomic clocks are extensively used for many purposes and in many different disciplines. If these were counted as different disciplines to which magnetic resonance contributed, the list would be very long, including radio-astronomy, metrology, precision navigation in outer space and on the earth, the Global Positioning System (GPS), tests of relativity, tests of the constancy of fundamental constants, etc.

Shortly after WWII, E.M. Purcell, F. Bloch and their associates invented methods, called NMR, for detecting magnetic resonance by the effect of the resonance absorption on the oscillator inducing the resonance transition, which enabled them to study liquid, gaseous and solid samples. The molecules in these substances were subject to frequent collisions which averaged out most of the nuclear interactions in molecules. This collision narrowing had the advantage of producing strong narrow resonances, but the disadvantage of losing most molecular information. However, the relaxation time T1 for the nuclear spin system reaching thermal equilibrium with the sample and T2 for the precessing nuclear spin system to lose phase coherency can be measured. T1, T2, the observed chemical shifts and the electron coupled nuclear spin-spin interaction provide sufficient information to be a guide in determining the location of atoms in a molecule and information about the surroundings of an atom in a molecule. Improved procedures for measuring the relevant NMR parameters have been developed, such as the spin-echo techniques of E. Hahn for measuring relaxation times and the pulse and Fourier transform methods developed by R.R. Ernst and others for obtaining magnetic resonance spectra. As a result, NMR has become an immensely valuable tool for chemical analysis.

Although NMR provided valuable information about the materials being studied the observations suffered by not being localized in the material. P.C. Lauterbur, P. Mansfield and R.V. Damadian developed different methods for using inhomogeneous magnet fields to localize the NMR signals in the sample, leading to the present beautifully detailed magnetic resonance images (MRI) A typical modern MRI apparatus includes a very strong and stable homogeneous magnetic field of approximately 20,000 gauss (2 tesla), usually produced by a large super-conducting magnet. In addition there are three variable gradient magnets at 180 to 270 gauss each of which can produce an inhomogeneous magnetic field such that there are magnetic resonances only along a line so the observed signal is the result of all of these resonances. When the gradient field is changed the results are obtained for a different line. When a different gradient magnet is used the resonance lines lie in different planes.

These results are then analyzed in a computer using Fourier transforms and programs similar to those used in X-ray CAT scans to give two dimensional pictures in arbitrary planes and in remarkably fine detail. Even the early MRI scans were of great value in biological and medical research and treatments, but the technology and applications of MRI’s have been and are being greatly improved and extended by the development of even sharper images, of images showing blood circulation, of injectable contrast materials, of functional MRI (fMRI) which shows which portion of the brain responds to different stimuli, etc.

The contributions of magnetic resonance to so many fields have already been so great that there is every reason to expect important new developments and applications will come in the future.

Related Laureates