nuclear magnetic resonance (NMR) spectroscopy.
1. Background
Over the past fifty years nuclear magnetic resonance spectroscopy, commonly referred to as nmr, has become the preeminent technique for determining the structure of organic compounds. Of all the spectroscopic methods, it is the only one for which a complete analysis and interpretation of the entire spectrum is normally expected. Although larger amounts of sample are needed than for mass spectroscopy, nmr is non-destructive, and with modern instruments good data may be obtained from samples weighing less than a milligram. To be successful in using nmr as an analytical tool, it is necessary to understand the physical principles on which the methods are based.
The nuclei of many elemental isotopes have a characteristic spin (I). Some nuclei have integral spins (e.g. I = 1, 2, 3 ....), some have fractional spins (e.g. I = 1/2, 3/2, 5/2 ....), and a few have no spin, I = 0 (e.g. 12C, 16O, 32S, ....). Isotopes of particular interest and use to organic chemists are 1H, 13C, 19F and 31P, all of which have I = 1/2. Since the analysis of this spin state is fairly straightforeward, our discussion of nmr will be limited to these and other I = 1/2 nuclei.
The following features lead to the nmr phenomenon:
1. A spinning charge generates a magnetic field, as shown by the animation on the right. The resulting spin-magnet has a magnetic moment (μ) proportional to the spin.
2. In the presence of an external magnetic field (B0), two spin states exist, +1/2 and -1/2. The magnetic moment of the lower energy +1/2 state is alligned with the external field, but that of the higher energy -1/2 spin state is opposed to the external field. Note that the arrow representing the external field points North.
3. The difference in energy between the two spin states is dependent on the external magnetic field strength, and is always very small. The following diagram illustrates that the two spin states have the same energy when the external field is zero, but diverge as the field increases. At a field equal to Bx a formula for the energy difference is given (remember I = 1/2 and μ is the magnetic moment of the nucleus in the field). Strong magnetic fields are necessary for nmr spectroscopy. The international unit for magnetic flux is the tesla (T). The earth's magnetic field is not constant, but is approximately 10-4 T at ground level. Modern nmr spectrometers use powerful magnets having fields of 1 to 20 T. Even with these high fields, the energy difference between the two spin states is less than 0.1 cal/mole. To put this in perspective, recall that infrared transitions involve 1 to 10 kcal/mole and electronic transitions are nearly 100 time greater. For nmr purposes, this small energy difference (ΔE) is usually given as a frequency in units of MHz (106 Hz), ranging from 20 to 900 Mz, depending on the magnetic field strength and the specific nucleus being studied. Irradiation of a sample with radio frequency (rf) energy corresponding exactly to the spin state separation of a specific set of nuclei will cause excitation of those nuclei in the +1/2 state to the higher -1/2 spin state. Note that this electromagnetic radiation falls in the radio and television broadcast spectrum. Nmr spectroscopy is therefore the energetically mildest probe used to examine the structure of molecules. The nucleus of a hydrogen atom (the proton) has a magnetic moment μ = 2.7927, and has been studied more than any other nucleus. The previous diagram may be changed to display energy differences for the proton spin states (as frequencies) by mouse clicking anywhere within it.
4. For spin 1/2 nuclei the energy difference between the two spin states at a given magnetic field strength will be proportional to their magnetic moments. For the four common nuclei noted above the magnetic moments are: 1H μ = 2.7927, 19F μ = 2.6273, 31P μ = 1.1305 & 13C μ = 0.7022. The following diagram gives the approximate frequencies that correspond to the spin state energy separations for each of these nuclei in an external magnetic field of 2.34 T. The formula in the colored box shows the direct correlation of frequency (energy difference) with magnetic moment (h = Planck's constant).
2. Proton NMR Spectroscopy
This important and well-established application of nuclear magnetic resonance will serve to illustrate some of the novel aspects of this method. To begin with, the nmr spectrometer must be tuned to a specific nucleus, in this case the proton. The actual procedure for obtaining the spectrum varies, but the simplest is referred to as the continuous wave (CW) method. A typical CW-spectrometer is shown in the following diagram. A solution of the sample in a uniform 5 mm glass tube is oriented between the poles of a powerful magnet, and is spun to average any magnetic field variations, as well as tube imperfections. Radio frequency radiation of appropriate energy is broadcast into the sample from an antenna coil (colored red). A receiver coil surrounds the sample tube, and emission of absorbed rf energy is monitored by dedicated electronic devices and a computer. An nmr spectrum is acquired by varying or sweeping the magnetic field over a small range while observing the rf signal from the sample. An equally effective technique is to vary the frequency of the rf radiation while holding the external field constant.
For a description of the pulse Fourier transform technique, preferred by most spectroscopists over the older CW method, Click Here.
As an example, consider a sample of water in a 2.3487 T external magnetic field, irradiated by 100 MHz radiation. If the magnetic field is smoothly increased to 2.3488 T, the hydrogen nuclei of the water molecules will at some point absorb rf energy and a resonance signal will appear. An animation showing this may be activated by clicking the Show Field Sweep button. The field sweep will be repeated three times, and the resulting resonance trace is colored red. For visibility, the water proton signal displayed in the animation is much broader than it would be in an actual experiment.
Since protons all have the same magnetic moment, we might expect all hydrogen atoms to give resonance signals at the same field / frequency values. Fortunately for chemistry applications, this is not true. By clicking the Show Different Protons button under the diagram, a number of representative proton signals will be displayed over the same magnetic field range. It is not possible, of course, to examine isolated protons in the spectrometer described above; but from independent measurement and calculation it has been determined that a naked proton would resonate at a lower field strength than the nuclei of covalently bonded hydrogens. With the exception of water, chloroform and sulfuric acid, which are examined as liquids, all the other compounds are measured as gases.
Why should the proton nuclei in different compounds behave differently in the nmr experiment ?
The answer to this question lies with the electron(s) surrounding the proton in covalent compounds and ions. Since electrons are charged particles, they move in response to the external magnetic field (Bo) so as to generate a secondary field that opposes the much stronger applied field. This secondary field shields the nucleus from the applied field, so Bo must be increased in order to achieve resonance (absorption of rf energy). As illustrated in the drawing on the right, Bo must be increased to compensate for the induced shielding field. In the upper diagram, those compounds that give resonance signals at the higher field side of the diagram (CH4, HCl, HBr and HI) have proton nuclei that are more shielded than those on the lower field (left) side of the diagram.
The magnetic field range displayed in the above diagram is very small compared with the actual field strength (only about 0.0042%). It is customary to refer to small increments such as this in units of parts per million (ppm). The difference between 2.3487 T and 2.3488 T is therefore about 42 ppm. Instead of designating a range of nmr signals in terms of magnetic field differences (as above), it is more common to use a frequency scale, even though the spectrometer may operate by sweeping the magnetic field. Using this terminology, we would find that at 2.34 T the proton signals shown above extend over a 4,200 Hz range (for a 100 MHz rf frequency, 42 ppm is 4,200 Hz). Most organic compounds exhibit proton resonances that fall within a 12 ppm range (the shaded area), and it is therefore necessary to use very sensitive and precise spectrometers to resolve structurally distinct sets of hydrogen atoms within this narrow range. In this respect it might be noted that the detection of a part-per-million difference is equivalent to detecting a 1 millimeter difference in distances of 1 kilometer.
I was born 1933 in Winterthur, Switzerland, where our ancestors resided at least since the 15th century. We lived in a home built in 1898 by my grandfather, a merchant. My father, Robert Ernst, was teaching as an architect at the technical high school of our city. I had the great luck to grow up, together with two sisters, in a town that combined in a unique way artistic and industrious activities.Invaluable art collections and a small but first rank symphony orchestra carry the fame of Winterthur far across the borders of Switzerland. On the other hand, industries producing heavy machinery, like Diesel motors and railway engines, provided the commercial basis of prosperity.
I soon became interested in both sides. Playing the violoncello brought me into numerous chamber and church music ensembles, and stimulated my interest in musical composition that I tried extensively while in high school. At the age of 13, I found in the attic a case filled with chemicals, remainders of an uncle who died in 1923 and was, as a metallurgical engineer, interested in chemistry and photography. I became almost immediately fascinated by the possibilities of trying out all conceivable reactions with them, some leading to explosions, others to unbearable poisoning of the air in our house, frightening my parents. However, I survived and started to read all chemistry books that I could get a hand on, first some 19th century books from our home library that did not provide much reliable information, and then I emptied the rather extensive city library. Soon, I knew that I would become a chemist, rather than a composer. I wanted to understand the secrets behind my chemical experiments and behind the processes in nature.
Thus, after finishing high school, I started with high expectations and enthusiasm to study chemistry at the famous Swiss Federal Institute of Technology in Zurich (ETH-Z). I was rapidly disappointed by the state of chemistry in the early fifties as it was taught at ETH-Z; we students had to memorize incountable facts that even the professors did not understand. A good memory not impeccable logic was on demand. The physical chemistry lectures did not reveal much insight either, they were limited just to classical thermodynamics. Thus, I had to continue, similar as in high school, to gain some decent chemical knowledge by reading. A book from which I learned a lot at that time was "Theoretical Chemistry" by S. Glasstone. It revealed to me the fundamentals of quantum mechanics, spectroscopy, statistical mechanics, and statistical thermodynamics, subjects that were never even mentioned in lectures, except in a voluntary and very excellent lecture course given by the young enthusiastic Professor Hans H.Günthard who had studied chemistry and physics in parallel.
It was clear to me, after my diploma as a "Diplomierter Ingenieur Chemiker" and some extensive military service, I had to start a PhD thesis in the laboratory of Professor Günthard. Fortunately, he accepted me and associated me with a young most brilliant scientist Hans Primas, who never went through any formal studies but nevertheless acquired rapidly whatever he needed for his work that was then concerned with high resolution nuclear magnetic resonance (NMR), a field in its infancy at that time. Much of his and also my time was spent on designing and building advanced electronic equipment for improved NMR spectrometers. In parallel, we developed the theoretical background for the experiments we had in mind as well as for the optimum performance of the instruments. Signal-to-noise ratio calculations and optimizations were daily routine as NMR suffers from a disappointingly low sensitivity that severely limits its applications. Hans Primas developed and analyzed field modulation techniques, constructed a field frequency lock system, and contributed a new design of shaped pole caps for the electromagnet that was supposed to deliver an extremely homogeneous magnetic field. These developments led to two types of spectrometers that were adopted by Trüb-Täuber, a Swiss electronics company, and sold all over Europe. Later in 1965, Trüb-Täuber was dissolved, and the NMR spectroscopy section led to the foundation of Spectrospin AG that is, together with Bruker Analytische Messtechnik, nowadays the world leading producer of NMR spectrometers.
My own work dealt with the construction of high sensitivity radio frequency preamplifiers and in particular high sensitivity probe assemblies, initially for a 25 MHz, later for a 75 MHz proton resonance spectrometer. On the theoretical side, I was concerned with stochastic resonance. The goal set by Hans Primas was the usage of random noise for the excitation of nuclear magnetic resonance, following the famous concepts of Norbert Wiener for the stochastic testing of non-linear systems. The theoretical treatment was based on a Volterra functional expansion using orthogonal stochastic polynomials. I tried in particular to design a scheme of homonuclear broadband decoupling to simplify proton resonance spectra. By applying a stochastic sequence with a shaped power spectral density that has a hole at the observation frequency, all extraneous protons should be decoupled without perturbing the observed proton spin. The theoretical diffculties were mainly concerned with the computation of the response to nonwhite noise. Experiments were not attempted at that time, we did not believe in the usefulness of the concept anyway, and I finished my thesis in 1962 with a feeling like an artist balancing on a high rope without any interested spectators.
I thus decided to leave the university forever and tried to find an industrial job in the United States. Among numerous offers, I decided for Varian Associates in Palo Alto where famous scientists, like Weston A. Anderson, Ray Freeman, Jim Hyde, Martin Packard, and Harry Weaver, were working along similar lines as we in Zürich but with a clear commercial goal in mind. This attracted my interest, hoping to find some motivation for my own work. And indeed, I was extremely lucky. Weston Anderson was on his way to invent Fourier transform spectroscopy in order to improve the sensitivity of NMR by parallel data acquisition. After his involvement in the development of a cute mechanical device, the "wheel of fortune", to generate and detect several frequencies in parallel, he proposed to me in 1964 to try a pulse excitation experiment that indeed led to Fourier transform (FT) NMR as we know it today. The first successful experiments were done in summer 1964 while Weston Anderson was abroad on an extensive business trip. In this work I could take advantage in an optimum way of my knowledge in system theory gained during my studies with Primas and Günthard. The response to our invention was however meager. The paper that described our achievements was rejected twice by the Journal of Chemical Physics to be finally accepted and published in the Review of Scientific Instruments. Varian also resisted to build a spectrometer that incorporated the novel Fourier transform concept. It took many years before in the competitive company Bruker Analytische Messtechnik Tony Keller and his coworkers demonstrated in 1969 for the first time a commercial FT NMR spectrometer to the great amazement of Varian that had the patent rights on the invention.
Still at Varian, I was further extending my earlier work on stochastic resonance with the introduction of heteronuclear broadband decoupling by noise irradiation, the "noise decoupling" that led to a rapid development in carbon-13 spectroscopy. It has been replaced later by the much more effcient multiple pulse schemes of Malcolm H. Levitt and Ray Freeman using composite pulses.
Of major importance for the success of more advanced experiments and measurement techniques in NMR was the availability of small laboratory computers that could be hooked up directly to the spectrometer. During my last years at Varian (1966-68), we developed numerous computer applications in spectroscopy for automated experiments and improved data processing.
In 1968 I returned, after an extensive trip through Asia, to Switzerland. A brief visit to Nepal started my insatiable love for Asian art. My main interest is directed towards Tibetan scroll paintings, the so-called thangkas, a unique and most exciting form of religious art with its own strict rules and nevertheless incorporating an incredible exuberance of creativity.
Back in Switzerland, I had a chance to take over the lead of the NMR research group at the Laboratorium für Physikalische Chemie of ETH-Z after Professor Primas turned his interests more towards theoretical chemistry. Despite an initial lack of suitable instrumentation, I continued to work on methodological improvements of time-domain NMR with repetitive pulse experiments and Fourier double resonance. In addition, we performed the first pulsed time-domain chemically-induced dynamic nuclear polarization (CIDNP) experiments. We developed at that time also stochastic resonance as an alternative to pulse FT spectroscopy employing binary pseudo-random noise sequences for broadband excitation, correlating input and output noise. Similar work was done simultaneously by Prof. Reinhold Kaiser at the University of New Brunswick.
The next fortunate event occurred in 1971 when my first graduate student, Thomas Baumann, visited the Ampere Summer School in Basko Polje, Yugoslavia, where Professor Jean Jeener proposed a simple two-pulse sequence that produces, after two-dimensional Fourier transformation, a two-dimensional (2D) spectrum. In the course of time, we recognized the importance and universality of his proposal. In my group, Enrico Bartholdi performed at first some analytical calculations to explore the features of 2D experiments. Finally in the summer of 1974, we tried our first experiments in desperate need of results to be presented at the VIth International Conference on Magnetic Resonance in Biological Systems, Kandersteg, 1974.
At the same time, it occurred to me that the 2D spectroscopy principle could also be applied to NMR imaging, previously proposed by Paul Lauterbur. This led then to the invention of Fourier imaging on which the at present most frequently used spin-warp imaging technique relies. First experiments were done by Anil Kumar and Dieter Welti.
From then on, the development of multi-dimensional spectroscopy went very fast, inside and outside of our research group. Prof. John S. Waugh extended it for applications to solid state resonance, and the research group of Prof. Ray Freeman, particularly Geoffrey Bodenhausen, contributed some of the first heteronuclear experiments. We started 1976 an intense collaboration, lasting for 10 years, with Professor Kurt Wüthrich of ETH-Z to develop applications of 2D spectroscopy in molecular biology. He and his research group have been responsible for most essential innovations that enabled the determination of the three-dimensional structure of biomolecules in solution.