|
|
|
| Six Sorcerers & One Apprentice |
| |
This is the story of six top scientists who were kind enough to guide
me, as a complete newcomer, in learning the tricky business of
unstructured research. I had been unable to find any accepted
do-it-yourself manual that could be of any practical use to someone
setting out on a research career. I read a book by Bright Wilson, but it
did not offer any deep insight into the philosophy I was looking
for. The arcane art of the pure researcher must be passed down from
established experts - the giants in their field. Luck was on my side,
for I found some remarkable men at the top of their game; what follows
is an account of how, in their different ways, each one gave me a
much-needed helping hand. How I came to work with them is largely
covered in the story "As Luck Would Have It"
in another section of this website. There it will be clear that I had
little choice in the matter, apart from recognising true leaders when I
found them. So here is my personal list of six great sorcerers: Rex
Richards, Anatole Abragam, Robert Pound, Ionel Solomon, Martin Packard,
and Wes Anderson.
Sir Rex Richards FRS - Fellow and Tutor in chemistry at Lincoln College, Oxford in the 1950s.
Rex Richards was one of the very first to appreciate that the newly
discovered phenomenon called nuclear magnetic resonance could have
unexpected applications in chemistry. Against all the odds, he set out
to build his own NMR spectrometer from scratch, despite the perceived
wisdom that heavy magnets and masses of electronics had no place in a
chemistry laboratory. Rex was an ideal tutor for undergraduates, a
charismatic lecturer in physical chemistry, and an excellent research
supervisor, so my fate as an NMR spectroscopist was sealed at an early
age.
War surplus radar equipment featured heavily in this enterprising
radiofrequency project, and I was lucky enough to have followed a radar
course in the RAF. Indeed 90% of my doctoral research was devoted to
building electronics for the excitation, detection and measurement of
the resonant frequencies of the arcane nuclei lithium, cobalt, gallium,
indium and thallium. Much of this work was purely exploratory, but it
was Rex who suggested that the NMR spectra of aqueous solutions of
cobalt-III complexes would have important implications for Ramsey's
theory of the chemical shift, because the cobalt atom has a low-lying
electronic state that dominates the calculation of the nuclear
shielding, and which also gives rise to the colour of these complexes.
This interpretation was confirmed by measuring the NMR frequencies and
comparing them with the visible/ultraviolet wavelengths for
symmetrical cobalt complexes. The result was a straight-line graph that
confirmed that the reciprocal of the appropriate electronic energy gap
was the dominant parameter determining the cobalt chemical shifts. It
seemed that NMR did have some useful applications in chemistry after
all.
I learned a great deal by simply observing Rex in action - teaching by
osmosis. Rex clearly understood the importance of a "hands off" style of
supervising research. A neophyte student was allowed the chance to
shrug off the doctrine (inherited from the undergraduate courses) that
there is always a correct, prescribed way to attack a scientific
problem. For the very first time, one could use one's own initiative and
perhaps discover new things for oneself. Trial and error, with a large
dose of the latter. I suppose it was inevitable that, many years later, I
adopted a similar laissez-faire approach to supervising my own
students. As an analogy, I am reminded of an almost surreal episode
recounted by our youngest daughter Louise. One day, on a visit to
Magdalen College, she was passing the old squash courts when "it started
to rain ducklings". Together with another girl, Louise tried to catch
these fluffy little balls before they hit the hard ground. The
ivy-covered walls of the squash courts must have been over 15 feet high,
but Mama duck had decided that the time had come to ease the ducklings
out of the nest so that they could find their own way to the river
Cherwell and learn to fend for themselves. Darwin would have been very
pleased. I found that my own research students seemed to fall neatly
into two distinct groups (a) highly motivated ones who would have been
seriously hindered by any micro-management, and (b) others who were not
really interested and who would soon graduate to careers outside of
chemistry. In this aspect the Oxford Part II (research) year has proved
to be a valuable exercise to help young students make a choice about
their future career.
In retrospect I realize that in the background Rex had been quietly
guiding my career. In a sense he was a father figure to me, for I had
lost my own father in 1940. Rex was instrumental in arranging my next
position at a renowned magnetic resonance laboratory in France, in the
belief that immersion in an environment of physicists would be a good
idea. Better than I, he understood what my future métier should be. In a sense we are both physiciens manqués, rather than chemists. So in the summer of 1957 I went to work as a stagiaire at Saclay. As a wise precaution I took some evening classes in Oxford in colloquial French.
Professor Anatole Abragam, Chef, Service de la Physique des Solides et de Resonance Magnétique, Centre d'Etudes Nucléaires de Saclay, France.
The incredible intellect of Anatole Abragam was evident to everyone who
met him. His comprehensive notes on magnetic resonance were already the
stuff of legend. At that time, they took the form of two paperback
volumes, written in French, and no longer obtainable in bookshops owing
to the enormous demand. He was busy converting this material into the
his new book "Principles of Magnetic Resonance" published in English in
1961 by the Oxford University Press. Naturally the focus was largely on
the physical aspects of magnetic resonance, for the great surge in
chemical applications had yet to take place. This immense work soon
became the "Bible" for magnetic resonance aficionados. Many ideas for
new experiments can now be traced back to almost passing remarks in that
book, but one has to admit, it was never an easy read. Anatole had
spent a year in the Clarendon Laboratory, and he held Oxford in high
esteem. Perhaps that is one reason that he took the trouble to find me a
slot in his group. To visitors he liked to joke that he kept me there
simply to calibrate his English, but I am pretty sure that he was making
a play on the word étalon (a calibration standard) because this also translates as stallion.
For someone like myself, trained as a chemist, the sudden immersion in a
cold bath of pure physics at Saclay felt like another relentless
application of Darwin's prescription for survival of the fittest. The
physicists at Saclay were trained in the rigorous mathematical formalism
characteristic of the French higher education system (particularly l'Ecole Polytechnique and l'Ecole Normale Supérieur).
This stood in sharp contrast to the Oxford emphasis on a practical
"nuts and bolts" approach to research. One is reminded of the
unashamedly practical steam engine invented in Britain, whereas
the key theories of thermodynamics that flowed from analysis of heat
engines were developed in France. Here is one example of this cultural
divide. Anatole Abragam was amazed to see a young Oxford chemist using a
lathe to fashion an NMR probe - cutting a screw thread on the inside of
a hollow Perspex cylinder with the intention of maximizing the filling
factor of the radiofrequency coil. By contrast, the French physicists
tended to delegate any such mechanical work to the support technicians,
tacitly accepting any limitations of this second-hand solution. A
chemist seeks pictorial descriptions of physical phenomena whenever
possible; the physicist looks for mathematical rigour. Sometimes I found
my colleagues dangerously overconfident in their blackboard
calculations; once they were quite surprised when I challenged a
computation that concluded that the concentration of a particular
component in solution was 210 Molar!
At that time the Overhauser effect was a fairly new concept; indeed at
first the acknowledged experts in magnetic resonance found it very hard
to accept Overhauser's predictions. As I was trying to understand this
rather surprising phenomenon, a Saclay colleague told me "Oh, you just
have to write down the equations". This is of course true, but I believe
that simple every-day ideas can offer valuable insight into the
Overhauser effect (see for example the ski analogy set out in "A Handbook of NMR"
page 143). A mathematical equation is just a shorthand representation
of reality; we should always bear in mind that every variable carries a
physical meaning. My own experience has been that physical intuition is
often more productive than mathematical formalism for finding new
experiments in spin choreography. On the other hand, our Oxford group
did come across one marked exception to this belief. During a routine
density matrix calculation, Tom Mareci observed that the conversion of
double-quantum coherence into an observable NMR signal is best carried
out by setting the pulse flip angle to 135° (instead of the traditional
90°) since this allows the determination of the signs of the
double-quantum frequencies. My two years at Saclay certainly taught me
the importance of the rigorous theoretical approach, but I still harbour
the suspicion that scientific breakthroughs are more often triggered by
insights gleaned from pictorial "hand waving" visualizations. Of course
intuition can sometimes mislead, but then this quickly focuses one's
attention on the unexpected counter-intuitive finding, and the
resolution of this conflict can be rewarding.
At that time the key project in Abragam's group was to use all kinds of
clever tricks to align the nuclei in a solid sample at very low
temperature in order to prepare a polarized target for the particle
physicists to bombard. After many years of really hard work they
gradually achieved a degree of polarization approaching 100%, a tour de force that perhaps never received the full recognition that it deserved. Here lies the dichotomy - should one seek for the quick eureka
moment, or work towards an important long-term goal? Watson and Crick
managed the enviable trick of successfully combining both. Abragam
himself commented that although there were many brilliant Frenchmen in
scientific research, it appeared to be the Anglo-Saxons (Americans and
British) who took the lion's share of Nobel prizes. An unfortunate
obsession with the purely theoretical approach perhaps?
Some years after I left Saclay, Abragam asked me to translate his book "Réflexions d'un Physicien"
into English. His own command of English was excellent, so I could only
assume that he was too busy with more important matters to undertake
the translation himself. The book in question consisted of a compilation
of short articles and some letters that he had written at various
times. He told me that he would leave the details entirely to me, but he
omitted to say that one or two sections had originally been written in
English and later translated into French, leaving open the possibility
that Anatole could have a laugh by comparing the English originals with
my translations. This was my first foray into preparing an actual book,
and it served as a useful dry-run when I set out to write a book of my
own.
Robert Pound, Professor of Physics, Harvard University, USA
By a fortunate turn of fate, Robert Pound was taking a sabbatical at
Saclay during the academic year 1957-58, and he took me under his wing.
Bob had been part of the prestigious Radiation Laboratory during the war
and he retained an active interest in all things electronic,
particularly those related to radar. Among magnetic resonance
spectroscopists he was perhaps best known at that time for his "Pound
spectrometer", a regenerative oscillator/detector that allowed one to
scan large radiofrequency ranges in search of weak "unknown" NMR
frequencies, without the need for high-gain amplification. Yet
physically the device was very simple - a mere double-triode vacuum
tube. I had built a copy of this "Pound box" for my doctoral research at
Oxford. Naturally I was delighted at the idea of working with one of
the real pioneers of NMR, a key figure in the Harvard research group
(Purcell, Torrey and Pound) that shared a Nobel Prize in physics in 1952
with the Stanford group (Bloch, Hansen and Packard).
Realizing that I was interested in high-resolution NMR applied to
chemistry, Bob suggested I should work on a project involving two of his
untried ideas. The first was a novel concept to employ weak modulation
sidebands (rather than the main radiofrequency transmitter) to detect
the NMR signals. He suggested that in this manner we should be able to
use the very simple Pound spectrometer as the detector. Although this
marginal oscillator device operates at such a high radiofrequency level
that normally it would completely saturate high-resolution NMR signals,
the use of modulation sidebands of low modulation index would circumvent
this drawback. The Pound-Watkins spectrometer was versatile because it
had a wide operating frequency range and required no high-gain
intermediate-frequency amplifiers. Incidentally we worked from some
primitive photocopies of Watkins' doctoral thesis, kept in a vice
because otherwise the pages curled up and became unreadable.
I built the radiofrequency parts of this new high-resolution
spectrometer from scratch, including that notoriously tricky item - the
probe. A radiofrequency probe is very sensitive to field distortion from
trace inclusions of paramagnetic material, such as tiny particles of
iron (one prescription was to boil in hydrochloric acid all the plastics
that had been machined). Furthermore the magnet that I borrowed was not
intended for high-resolution work, so the available field uniformity
was unknown. Consequently I was disappointed (but not terribly
surprised) to discover that all the resonance lines that I recorded were
broad (about 200 Hz) whereas we were aiming at a resolution of a
fraction of 1 Hz. Once I had eliminated all the practical problems of
materials that made up the probe, I began to wonder whether the
regenerative feature of the Pound spectrometer might be the culprit,
enhancing the broadening by radiation damping. Now the key paper on
radiation damping was that of Bloembergen and Pound, so it was with some
trepidation that I suggested to Bob that this might prove to be the
Achilles heel of his device. By this time, Bob had returned to Harvard
and was incredibly busy on a brilliant new experiment to measure
Einstein's gravitational red shift (Pound and Rebka). However, after a
flurry of transatlantic letters I eventually convinced him that
radiation damping was indeed the problem, so I replaced the Pound
spectrometer with a conventional transmitter and all was well again.
With the misplaced excitement of youth I was thinking of writing a brief
communication to warn about radiation damping in regenerative
oscillators, when a casual visitor to Saclay pointed out that Lösche's
group in Leipzig had already analyzed this problem in an East German
scientific journal that we had all overlooked. Take home message for
beginners: if all else fails, read the literature.
Bob's second idea involved a super-regenerative oscillator, essentially a
pulse-modulated Pound box with the feedback turned up very high. He had
already used this machine to detect aircraft by radar. Consider first
of all the condition where the oscillator is quenched by the gating
pulse. Then, once the gate is slowly opened, the high degree of
regeneration induces a rapidly growing oscillation where the phase is
triggered by random circuit noise. After the exponential build-up of
these noisy oscillations, the gating pulse quickly quenches them, and
the cycle is repeated. The radiofrequency spectrum of this sequence of
incoherent pulses spans a broad frequency band. However if there is a
returning radar echo stronger than the circuit noise, this triggers the
subsequent oscillations in phase, and the oscillator becomes coherent.
Essentially the device has "locked on" to the radar echo.
An analogous regime is established if the super-regenerative oscillator
is used to excite an NMR signal, for this would trigger the conversion
from incoherent to coherent mode. So I set out to build this device.
Here the cultural divide showed up once again. Conventional wisdom
suggests that the Q-factor of an NMR coil should be as high as possible,
for this improves the signal-to-noise ratio. On the other hand, a high
Q-factor in a super-regenerative oscillator tends to "pull" the desired
NMR frequency. When I asked Bob what kind of compromise should be
adopted, he embarked on pages and pages of differential equations - a
very long calculation that never did converge to a useful conclusion. So
I simply went ahead with the Q-factor I already had. In the NMR
application, the wideband incoherent oscillator excites a free
precession signal from any point within that broad frequency range. If
this signal exceeds the circuit noise, the subsequent oscillations
become coherent, synchronized with the NMR Larmor frequency. When the
main magnetic field drifts, the oscillator frequency follows,
maintaining the appropriate NMR resonance condition, thus compensating
the magnet instability to a very high degree (with errors less than
0.5%). Fed to a second (high-resolution) NMR probe, this radiofrequency
signal keeps the second field/frequency ratio essentially constant so
that excellent stability is achieved. We probably had the most stable
high-resolution spectrometer at that time, permitting very slow scanning
rates of the order of 0.2 Hz per second, such that the usual transient
sweep "wiggles" were not observed. However the Saclay physicists were
not at all interested in high-resolution experiments ("When the chemists
arrive, it is time to move into another field") and my equipment was
soon banished to a basement oubliette when I left.
Ionel Solomon, Polytechnician and Saclay physicist.
Ionel does not fit into my rather glib generalisation about the French
physicists. Although he graduated from the same rigorous Gallic
educational system, he also had a natural gift for visualizing the
behaviour of nuclear spins in a simple pictorial manner. For this reason
he was an ideal teacher and supervisor, and helped me enormously,
particularly after Bob Pound had returned to Harvard. In one instance he
very gently led me to discover for myself why my radiofrequency
amplifiers were misbehaving. The problem was solved by teaching me about
the concept of a "grid stopper", a trick that prevented positive
feedback.
Ionel had written the first analysis of the effect of chemical exchange
on NMR spectra - formalized in the famous "Solomon equations" (I. Solomon, Phys. Rev. 99, 559, 1955).
As one practical example, they describe the complex chemical exchange
behaviour of hydrogen fluoride, with and without small traces of water.
This was the first use of "I" and "S" to designate two coupled spins (in
adopting this nomenclature we unwittingly render homage to Ionel
Solomon). Incidentally Ionel was a visiting scientist at Harvard at the
time of this work, and benefitted from advice about the chemistry of
hydrogen fluoride from Rex Richards, who was also visiting Harvard at
that time. It was a small NMR world in the 1950s.
I was fortunate to work in the same room as Ionel and was able to
observe him in action every day. On one occasion, in an Oxfordesque
manner, Ionel was playing with the NMR spectrometer, and quite by chance
discovered a completely new physical phenomenon (his "penicillin
moment"). While applying pulses to the radiofrequency crystal in the
Varian spectrometer, he observed a strange new response that had the
appearance of a new kind of echo. Other mortals might have dismissed
this unexpected effect as an instrumental glitch (of which there were
many), but Ionel persisted, deduced exactly what was happening, and was
able to maximize the new response by adjusting the pulse length by
trial-and-error. This was the first rotary echo, analogous to the Hahn
spin echo, but attributable to refocusing in the spatial inhomogeneity
of the radiofrequency field (I. Solomon, Phys. Rev. Letters 2, 301, 1959).
The optimum condition corresponds to a 180° radiofrequency phase shift
that exactly reverses the sense of rotation of the spin isochromats,
bringing them back into focus. Ionel was then able to generate a
sequence of multiple echoes (more than a thousand) demonstrating that,
in contrast to conventional spin echo experiments, errors in the 180°
pulses were not cumulative. This proved to be a new way to study spin
relaxation in a liquid sample.
From my own narrow parochial perspective, it seems a pity that Ionel
moved out of magnetic resonance soon afterwards. I realize now that this
is often the destiny of research scientists. They may change to an
entirely new field - Wes Anderson left magnetic resonance and went on to
develop an important ultrasound imaging device. Others move into
administration and are forgotten by the NMR community. The revolutionary
invention of magnetic resonance imaging might have tempted me to change
fields in 1973, but at the time I was starting a new appointment at
Oxford, so I stayed in NMR spectroscopy. I consider myself very lucky to
have worked in the same field for almost sixty years. However, today
many more people recognize the term MRI than have ever heard of NMR.
Dr Martin Packard, head of Varian Instrument Division.
When I arrived at Varian in November 1961 I was enormously impressed by
the extremely warm welcome from everyone in the group. The whole NMR
field was so new that any enthusiast from abroad was accepted,
apparently without question. I had been there only a few days when
Martin Packard handed me the keys to a car on indefinite loan, a
kindness that would have been unheard of in Europe. The car was in fact a
rather elderly Ford; the accepted American jargon would have dismissed
it as mere "transportation" but it served our family very well indeed. A
physics graduate of Stanford, Martin was one of the original pioneers
of NMR. He may well have been the very first person to actually see
an NMR signal (but see below). Following Felix Bloch's suggestion,
Martin spent all day vainly searching for the first proton signal from
water by slowly varying the magnetic field. It must have seemed like
looking for a needle in a haystack because their calculations of the
actual magnetic field strength were rather unreliable (in fact the field
was too high). It was only after he finally decided to abandon the
search for the day and shut down the magnet, that he saw a fleeting
water signal flash across the oscilloscope (F. Bloch, W. W. Hansen and M. E. Packard, Phys. Rev. 69, 127 1946).
An independent group at Harvard had also observed an NMR signal (E. M. Purcell, H. C. Torrey and R. V. Pound, Phys. Rev. 69, 37, 1946).
Their report was submitted earlier (December 24, 1945) than that of the
Stanford group (January 29, 1946), and it was more comprehensive, but
the Nobel committee clearly believed that the East coast and West coast
developments were independent and virtually contemporary, and the prize
was divided between Bloch and Purcell. Had the Harvard group been judged
to have absolute priority, thus scooping the Stanford group, Bob Pound
might well have secured a share in the prize. His later brilliant
determination of the gravitation red shift might also have attracted the
attention of the Nobel committee, but instead they gave the prize to
Mössbauer who had discovered the recoilless emission of gamma rays on
which Bob's experiment relied. Two near misses in a single lifetime?
Not content with his first discovery, Martin can also lay claim to be
the first person in the world to observe a high-resolution NMR spectrum,
the famous three lines of ethanol. It was an Indian chemist, Srinivas
Dharmatti at Stanford, who had pointed out to the physicists that almost
any organic compound should show a proton spectrum of several distinct
lines, provided that the magnetic field was sufficiently uniform. For
this purpose Jim Arnold built a permanent magnet with resolution
approaching one part in 108 (see "The Dawn of NMR" in another
section of this website). Martin then ran the spectrum of ethanol, and
chemistry was changed forever (J. T. Arnold, S. S. Dharmatti and M. E. Packard, J. Chem. Phys. 19, 507, 1951).
In fact that group later resolved additional splittings on the three
chemically shifted responses, but whenever they had a chemist visiting
the laboratory, they intentionally degraded the effective resolution to
disguise this effect because it represented a rather messy complication
to what was otherwise a beautifully simple concept - one resonance for
every distinct proton site. One step at a time.
Although Martin Packard was not directly involved in my research at
Varian, his influence was enormous. I suppose one could say that I
learned mainly by observing someone who had already "been there, done
that". As head of Instrument Division, Martin had the foresight to allow
his young scientists to pursue their own interests without
interference. Fortunately these enthusiasts were naturally disposed
towards investigations that advanced the methodology of magnetic
resonance, either to improve Varian instruments directly, or to generate
valuable publicity for the company by writing innovative scientific
papers in the field. There was little need to assign practical goals;
largely unstructured research was producing useful results anyway. As
one specific example, our attempts to understand the theory of double
resonance led to the development of high-stability NMR spectrometers
based on the internal field/frequency lock. This extraordinary freedom
of action was possible because Varian was a small emerging company that
had only just accumulated a critical mass of young talent. Scientists
from abroad sometimes assumed that Varian was one of the many California
universities. It was only later, when the company grew much larger,
that complacency and stagnation set in. Then top scientists began to
leave, and a flurry of young masters of business administration (with no
knowledge of magnetic resonance) began to impose the conventional
American management philosophy.
Weston Anderson, Physicist at Varian
When Felix Bloch was appointed director of CERN in Geneva, he took his
two research students, Jim Arnold and Wes Anderson, with him to
Switzerland and told them to write up their thesis work for publication.
These became the classic papers on high-resolution NMR. Wes Anderson's
thesis was written up in the Physical Review, 102, 151, 1956.
Compared with our own modest experiments at Oxford in 1956 on cobalt
chemical shifts, this classic served as a real wake-up call,
demonstrating for the first time the full beauty and versatility of
high-resolution NMR of protons. Martin Packard's famous three-lines
spectrum of ethanol turned out to be merely the tip of an enormous
iceberg, because Wes was able to record well-resolved proton spectra of
quite simple organic compounds that showed many chemically-shifted
lines, together with their associated fine structure. Physicists had
ventured into chemistry, with the result that this new spectroscopy
eventually revolutionized the study of molecular structure.
At that time a spectrometer could be thought of as an essentially
neutral device. There was a direct one-to-one relationship between the sample and its spectrum;
the operator could not interfere (except for trivial changes in
pressure or temperature). NMR spectra recorded in Paris or Palo Alto
were essentially identical. Wes took the first steps along a new path
where spectra could be manipulated by the hand of man. This
innovation can be regarded as the forerunner of all the present
adventures in spin gymnastics that have made the field so productive.
This deliberate intervention involved decoupling protons from protons,
demonstrating which pairs of groups were related through the scalar
coupling. From this modest beginning flowed more sophisticated
manipulations like spin tickling, determination of signs of coupling
constants, the measurement of the rates of chemical exchange, and the
nuclear Overhauser effect. Few doctoral theses have been so influential
in opening up new fields.
For someone who had struggled with solid-state NMR, and with NMR of the
"other" nuclei, it seemed clear to me that high-resolution spectroscopy
of protons was the way ahead. Whereas it had taken me many months to
determine a single inter-proton distance by recording a Pake doublet
from a polycrystalline sample of potassium amide, a high-resolution
proton spectrum could be obtained in a matter of minutes and promised a
wealth of important molecular data. The idea that even more structural
information could be gleaned from double irradiation experiments was an
intriguing possibility. One could argue that double-irradiation
experiments were just the first step in the "manipulation" revolution. A
decade later the invention of two-dimensional spectroscopy by Jean
Jeener suggested a brilliant new scheme for recording the kind of
information that had formerly been painstakingly recorded by double
irradiation.
For these reasons I decided to join Varian in late 1961 and work under
Wes' guidance. Immediately the pace of research accelerated. Only three
days after my arrival I was asked to give an impromptu evening seminar
at Wes' home attended by a handful of Varian enthusiasts. There I showed
some slides of spin decoupling experiments carried out previously in
England. Harry Weaver, a physicist in the audience, commented that
certain features in these spectra seemed to show an effect sometimes
given the jocular name "spin tickling". In fact there was no general
agreement about what the term actually meant, and nothing had been
published about the concept. Incidentally this was an unusual case of an
experiment acquiring a name even before it had been properly
implemented in practice. Later I realized that any new spin manipulation
should always be given catchy name or acronym in order to replace an
unwieldy description along the lines "You remember that pulse sequence where proton polarization is transferred to carbon spins in order to enhance sensitivity?"
Like the early explorers of our planet, we claimed the privilege of
naming each new discovery, and thereby establishing priority.
Out of pure curiosity Wes and I decided to look into spin tickling in
detail. The basic observations turned out to be relatively
straightforward - any NMR transition that shares an energy level with
the tickled transition splits into a doublet, the splitting being
proportional to the intensity of the irradiation field. In the published
article (J. Chem. Phys. 37, 2053, 1962) we bowed to political correctness and never mentioned the actual word tickling;
this necessitated some awkward circumlocution. What did take us quite
by surprise was the observation that the new doublets were sometimes
well resolved and sometimes poorly resolved. This did not seem to be a
mere instrumental glitch. It was Wes who found the explanation. Tickling
involves quantum-mechanical mixing of an allowed transition with a
connected forbidden transition (akin to Fermi resonance in infrared
spectroscopy). There are two kinds of connected transitions, which David
Whiffen named regressive and progressive. When the
tickled resonance line is regressively-connected to the observed
resonance line, mixing involves a (forbidden) zero-quantum transition.
Because the latter is insensitive to the spatial inhomogeneity of the
magnet, this situation gives a particularly well-resolved doublet. In
contrast, in the progressively connected case, mixing involves a
(forbidden) double-quantum transition, which is doubly sensitive to
magnet inhomogeneity, so that tickling doublet is poorly resolved. How
fortunate to be working with a physicist!
The tickling condition is very sensitive to the position of the
irradiation frequency with respect to the centre of the chosen resonance
line. If there is a slight offset, the connected doublet becomes
asymmetric, one component getting stronger and the other weaker (at very
large offsets the weaker line becomes a purely forbidden transition).
Careful adjustment of the tickling frequency to give a symmetrical
doublet offers a precise method for measuring line frequencies in a
high-resolution proton spectrum, with accuracies approaching ±0.01 Hz.
However highly accurate frequencies were not a major concern to most
organic chemists at that time, for they were mainly interested in
chemical shifts, where an accuracy of ±1 Hz was more than adequate for
the purpose. Chemists were busy exploring the exciting (and seemingly
endless) applications in organic chemistry, and were not really
interested in spin tickling.
At that time (1962) spectrometers necessarily operated at quite low
magnetic fields (typically equivalent to a proton frequency of only 60
MHz) and there was therefore considerable interest in the computer
analysis of strongly-coupled proton spectra. The first step in such
programs was the assignment of the observed transitions to an
energy-level diagram, something that tickling achieved very effectively.
Once this assignment was complete, the iterative fitting program for
determining shifts and coupling constants converged reliably and
rapidly. Later, as magnets began to be constructed with higher fields,
strongly-coupled proton spectra became less common, and computer
analysis consequently less important. With hindsight I would now have to
admit that tickling experiments aroused only minor interest at that
time. Tickling appeared to be an interesting exercise in spin physics,
but not terribly productive.
Tickling offers a good example of the dichotomy between
curiosity-inspired and applied research. The former is often given the
adjective "blue-sky" after the work of the nineteenth century scientist
John Tyndall, who set out to show (erroneously as it turned out) that
the colour of the sky arises from scattering by dust particles or water
droplets in the atmosphere. The alternative category of "focused"
research works towards solving a specific problem, such as a search for a
synthesis of a natural product like quinine. Both kinds of endeavour
are important, and both are useful in their different ways. But we
should never assume that "pure" research does not have direct practical
uses. In order to study spin tickling we had first to stabilize the
field/frequency ratio to permit experiments where the irradiation
"tickling" frequency could be held fixed, while the field was held
constant and the rest of the spectrum was scanned by sweeping the
observation frequency. For this purpose we borrowed an idea suggested by
Hans Primas at the Eidgenössiche Technische Hochshule in Zürich
for using an error signal derived from the internal reference compound
tetramethylsilane (TMS). There was already a Varian stabilization system
that used a mirror galvanometer and two sensitive photocells to provide
a correction current to counteract magnetic field drift. However this
could not handle very slow drifts of the resonance condition because it
lacked an absolute reference point. The insertion of a TMS error signal
provided this crucial locking information. Our new stabilization scheme
proved to be so successful that it became the basis of the next
generation of Varian spectrometers (HA60 and HA100). A project born out
of pure curiosity turned out to have widespread practical usefulness
after all.
In parallel with this work Wes and I spent some time investigating the
details of spin decoupling, calculating the so-called "magic curves"
that map the way the structure of a proton spin multiplet changes when
its coupling partner is irradiated at various decoupler offsets.
Eventually these diagrams evolved into three-dimensional "stacked plots"
of intensity as a function of the observed frequencies in one
dimension, versus the decoupler frequency in the second dimension. This
form of representation anticipated the later explosion of
three-dimensional spectra engendered by Jeener's concept of
two-dimensional spectroscopy. The first decoupling magic curve was
actually published much later in the Journal of Magnetic Resonance 26, 133 1977.
Not long after this work we decided that my apprenticeship was
essentially complete, and Wes and I moved into separate projects, the
most important of which was carried out by Wes, and is described below.
Wes loved practical gadgets and built various devices at home in his
spare time. Telescopes and seismographs were two of his favourites. One
evening while we were having dinner with several other guests at Wes'
house, he asked if we would like to see his latest home-made seismograph
in the garage. I don't know how he contrived to do this, but at the
very moment that we entered the garage, the pen recorder started to go
absolutely crazy. It turned out later that it was recording a major
earthquake in Alaska. Since boats in San Francisco Bay were also being
visibly shaken by the quake, the extreme sensitivity of Wes' device was
not strictly necessary on that occasion. But it was an impressive party
trick nonetheless.
Wes' gadget that had the greatest impact of all was one that was never
actually tested in practice. The details of this story are set out in
section 9.1 Multichannel Excitation in "When Chemists first 'Discovered'
NMR" on this website. Wes was the first to appreciate that the accepted
frequency-sweep slow-passage regime, used by all high-resolution NMR
practitioners at the time, sadly lacked sensitivity. He predicted that
simultaneous multichannel irradiation of the entire spectrum would
improve sensitivity by a large factor, roughly equal to the square root
of the number of independent channels. So he set out to build his
"Prayer Wheel", a gadget designed to create a comb of equally-spaced
modulation sidebands. It was never used in anger because Wes soon
realized that pulse excitation followed by Fourier transformation of the
transient NMR signal would achieve the same predicted order of
magnitude improvement in sensitivity, and would do so far more
effectively. The resulting paper by Ernst and Anderson (Rev. Sci. Instr. 37, 93, 1966) changed NMR forever. The actual Prayer Wheel went to the Smithsonian Museum in Washington, DC.
I remain greatly indebted to these six gifted sorcerers who eased my
transition from formal book-learning to the exciting arena of scientific
research.
|
| |
|
|
|
