Nuclear magnetic resonance (NMR) is one of the most powerful tools in a synthetic or structural chemist's arsenal. By probing the magnetic properties of spin-active nuclei, chemists can learn more from an NMR spectrum than almost any other analytical technique. The application of NMR to chemistry is different from pH, UV-vis, and IR spectroscopy in that instrumentation was produced before applicable chemical problems were known. The Beckman Model G, Beckman DU, and PE 12 and 21 spectrometers were all instruments designed to answer specific questions asked by chemists of the time, but Varian began development of its line of NMR spectrometers well before commercially-viable applications were known. The result was an unusually short gestation period from the discovery of nuclear magnetic resonance to the introduction of the first widely successful instrument, the Varian A-60.

The first magnetic resonance experiments were something of a physicist's parlor trick. In 1946, Purcell, Torrey, and Pound at Harvard (Purcell et al., 1946) and Bloch, Hansen, and Packard at Stanford (Bloch et al., 1946) independently reported a way to "listen in" on the magnetic fields of spinning atomic nuclei: nuclear magnetic resonance or NMR. Purcell's group discovered nuclear magnetic resonance in paraffin and Bloch's group found it in water. Widespread application of NMR was never intended outside of the nuclear physics community, but a few chemists were quick to see the potential of this new technique and start honing it to become the powerful analytical tool it is today.

The first applications of NMR spectroscopy used hand-built instruments to investigate the magnetic moments and vibrational frequencies of molecules. Finding the magnetic moment of a nucleus was difficult because different electronic environments around each equivalent magnetic nuclei shifted the peak from its true value. Annoyed physicists dubbed the phenomenon the "chemical shift," not realizing the analytical potential of their discovery. Several other characteristics were soon discovered that led to a better understanding of NMR data, notably chemical shielding in 1950 and spin-spin interactions in 1952 (Waugh, 1993). When the three peaks of ethanol, shown in Figure 1, were discovered by Purcell's group in 1951, chemists began to understand that the "annoying" features of an NMR spectrum, from a physicist's point of view, gave an ideal combination of "enough information to be useful and not so much to be an embarrassment of riches" (Waugh, 1993). But even before the most chemists understood what NMR stood for, scientists at Varian Associates pushing the technique forward.

Russell Varian, a physicist, and his brother Sigurd, a pilot and engineer, founded Varian Associates in 1949 to continue to develop and produce klystron tubes. Russell Varian had invented the klystron tube in 1937 as a way to produce high frequency microwave signals that could be used to detect aircraft and other objects at long distances. This invention led to the development of radar, and heavy use of klystron tubes in radar equipment during World War II established the Varian name and gave the brothers enough funding to begin their own company. (Pudrier and Moynihan, 1999)

Russell Varian had received his education at Stanford, and while there had become friends with William Hansen, later a member of Bloch's group that would pioneer NMR spectroscopy in liquids. Hansen and the two brothers began to save money for their own company, but their plans were halted until after World War II and Bloch and Purcell's discoveries. Because of connections with Hansen and others at Stanford (Varian klystron tubes had been developed largely at Stanford), Russell Varian was familiar with the Bloch group's work and perceived some of its future significance. In 1946, Russell Varian convinced Bloch and Hansen to patent their instrument and give "Varian Associates" an exclusive right to develop its commercial applications. From the start of Varian Associates in 1949, the company had two distinct foci: klystrons for military and commercial use and NMR devices. (Shoolery, 1993)

Varian Associates' unique product line worked well for a company nurturing a completely new technology from discovery to commercial application. With revenue from its klystron tubes, Varian Associates was free to devote money and manpower to the development of NMR without worrying about immediate profit. Working closely with Bloch's group at Stanford, the company began NMR development by creating a 12-inch electromagnet for its applied research. Development was funded by offering the magnet and other innovations for sale commercially. Varian Associates' policy of commercially offering its advances served the company well, and generated enough publicity that Humble Oil of Texas ordered the first commercial NMR spectrometer in 1952, before even Varian Associates had a finished instrument. That commercial NMR customers existed in 1952 may not seem unusual now, but given how little theoretical or practical explanation of NMR spectra existed at the time made this early sale remarkable.

The first instrument to be designed and built at Varian Associates' Applications Laboratory, or Applab, was the HR-30. The HR-30 used the same type of 12-inch electromagnet as the Humble oil instrument, and was the state-of-the-art in NMR spectroscopy. Although the HR-30 was an "operator's nightmare" (Shoolery, 1993), the results it produced began to bring nuclear magnetic resonance spectroscopy into view as a possible, if not yet viable, tool for chemists. Jim Shoolery, one of Varian Associates' early proponents of NMR spectroscopy, compared the operation of the HR-30 to another instrument of the time: the dual beam infrared spectrometer:

The IR spectrometer was very stable and gave a highly reproducible spectrumÉ. Most IR instruments would fit on a table, required no special facilities or services, were reasonably easy to operate, and worked with dilute samples.

The HR-30 had none of these virtues. Its 5400-lb electromagnet required cooling water,É electrical service, [and] a substantial floor. Although it could reproduce the general features of a spectrum, É finer details [varied] with the instability of the magnetic field. Sensitivity was so poor that only [pure]Éor concentrated liquids would produce signals above noise, andÉ range ofÉspectral dispersion severely limited the interpretability and usefulness of the spectra. (Shoolery, 1993)

The first task in making nuclear magnetic resonance common was to give the technique an easily recognizable name and standard methods for data presentation. Infrared spectroscopy had become popularly known as IR, so nuclear magnetic resonance was dubbed NMR, simultaneously capitalizing on the success of IR and avoiding the foreboding connotations of "nuclear." Early NMR spectra were recorded by Polaroid photograph, but this method was soon abandoned in favor of pre-calibrated strip chart recorders, again similar to IR instruments of the time. The example of IR spectra was abandoned in deciding that NMR spectra should have positive peaks rather than the deflection of IR spectra; this change supposedly "symbolized the vitality of the new technique." (Shoolery, 1993)

The first problem solved by NMR spectroscopy was submitted by E. J. Corey, then of the University of Illinois. Corey's sample contained one of two products, and attempts to differentiate between the two structures by UV or NMR had been inconclusive. By integrating the NMR spectrum, one of the two structures was shown clearly correct. This problem was used in the "This is NMR at Work" series for Analytical Chemistry, and the technical portion of the advertisement is reproduced in Figure 2. (Shoolery, 1993).

Varian continued to improve its NMR instrumentation by introducing the HR-40 NMR spectrometer in 1955. The HR-40, operating at a 40 MHz Larmor frequency and 9600 G field strength, introduced one fundamental improvement at its introduction and two more over the course of its life. The first advancement introduced by the HR-40 was sample spinning. A report from Bloch's group at Stanford had shown that the linewidth of NMR peaks dramatically decreased with sample spinning. The reason for this was thought to be that sample spinning produced a special case of averaging such that all nuclei the same distance from the axis of spinning experienced the same magnetic field. This "motional" averaging greatly decreased the width of NMR peaks, as shown by Figure 3, a spectrum of ethyl acetate with and without spinning. Furthermore, because all nuclei the same distance from the axis of spinning experienced the same magnetic field, the operator only had to look along one dimension, the axis of spinning, to find the minimum field gradient required to take an NMR spectrum.

The third and final improvement introduced over the lifetime of the HR-40 was a magnetic field stabilization loop. Magnetic field instability was the problem underlying all others; despite sample spinning and electronic shimming, spectra exhibited poor resolution, irreproducible baselines, and rapid drift without a stable magnetic field. The problem was solved by attaching a sensitive induction coil over one of the magnetic poles. If the magnetic field varied, a current was induced in the coil, which passed by a complicated mechanism to an amplifier. The amplifier produced a current proportional and opposite to the disturbance in the magnetic field, thereby improving the field homogeneity from ~1 in 10 million to 1 in 1 billion, an order of magnitude better than was required. Jim Shoolery likens the difference to going from "looking through binoculars at a row of birds on a telephone wire from the back of a truck bouncing on a rutted road" to "the truck had stopped and the binoculars were mounted on a tripod. [One] couldÉcount all the birdsÉ[and] even tell if their feathers were ruffled" (1993). The stabilization loop was named the "Super Stabilizer" and introduced as an accessory for all new and existing spectrometers (Shoolery, 1993). With the introduction of the Super Stabilizer, NMR spectra were finally reproducible, drift could be eliminated, and spectra taken slowly enough to allow maximum resolution. An example of the difference these improvements made in NMR spectra is dramatically illustrated by the low and high-resolution spectra of ethanol shown in Figures 4a and 4b.

The first machine to take full advantage of spinning, electronic shimming, and feedback stabilization at its inception was the HR-60, introduced in 1958 with a Larmor frequency of 60 MHz and field strength of 14, 096 G. The HR-60 was able to solve a much larger range of problems than its predecessors and was a small-scale commercial success. However, its mammoth size (the magnet still weighed 5400 lb) and price kept it out of range of most chemists. The HR-60 was soon followed by the HR-100 in 1959. The HR-100 used a smaller magnetic pole gap to produce a field strength of 23,490 G and operating frequency of 100 MHz, the highest resolution commercially until superconducting NMR instruments were introduced in 1962. (Shoolery, 1993)

As the tremendous potential of NMR techniques became evident, Varian Associates perceived the market potential for a low-cost, proton-only NMR spectrometer. Development began in earnest in 1957 for such a spectrometer, code-named "ASP" for the timeframe: As Soon as Possible. The new instrument was to employ a 6-inch electromagnet with a resolving power and sensitivity of the HR-60. A 6-inch magnet size was chosen to make the entire instrument fit into two secretary-sized consoles, one for the magnet and most electronics and the other for the controls, a flatbed recorder, and the remaining electronics. ASP was to be simple enough to be used by an organic chemist or graduate student, have a mean time between failures of one year, and be priced comparably to good infrared instruments of the time. After much deliberation, the name A-60 was chosen for the instrument; "A" designated the instrument's analytical intent and "60" was the operating frequency and planned year of introduction.

The power of the A-60 was demonstrated during Shoolery's first trial of the prototype A-60:

I put a sample in the instrument, adjusted the resolution, and ran a spectrum on the precalibrated chart. It was perfect. But could the A-60 reproduce a spectrum with the fingerprint quality of the IR instrument? I moved the pen back to the start and restarted the scan. I was momentarily distracted, and when I looked back I saw only one line on the chart. "Why didn't the second scan run properly?" I asked. The answer came back, "It did!" Amazed and almost incredulous, I returned the pen three more times. It laid down five identical spectra with a single trace showing on the paper! At that instant, I knew the field of organic chemistry would never be the same again. (1993)

The A-60, introduced at the Pittsburgh Conference in 1961, helped establish NMR spectroscopy as a standard tool for organic chemists. A catalog of 700 NMR spectra was included with each A-60 and made available to others at cost. This catalog proved very useful to organic chemists as a tool for learning how to interpret and use NMR spectroscopy. Varian's spectral catalog also settled a long-standing debate as to whether the TMS reference peak should be on the left, with the scale increasing from left to right, or on the right, with the scale increasing right to left. Factions argued on either side of the debate, but Varian's compilation of 700 spectra with TMS on the right firmly established that method it as the standard way to report NMR spectra. Over 1000 A-60 NMR spectrometers were produced over its lifetime, of which more than 120 were sold in the first year (Shoolery, 1993).

Sales of the A-60 were a testament to the instrument's popularity, usefulness, and reliability. NMR spectroscopy soon became a tool used by almost every organic chemist, and the A-60 was the NMR spectrometer of choice in most labs for many years. As can be expected from a relatively new technology, NMR spectroscopy has changed considerably since its widespread acceptance, making the speed and reliability of the A-60 less impressive than at its introduction. An illustration if this is this amusing note taped to an archived A-60 at the Chemical Heritage Foundation in Philadelphia:

WARNING: THIS MACHINE SUBJECT TO BREAKDOWNS DURING PERIODS OF CRITICAL NEEDS.

A special circuit in the machine called a "CRISIS DETECTOR" senses the operator's emotional state in terms of how desperate he or she is to use the machine. The "CRISIS DETECTOR" then creates a malfunction proportional to the desperation of the operator. Threatening the machine with violence only aggravates the situation. Likewise, attempts to use any other machine may cause it to malfunction too--they belong to the same union.

KEEP COOL AND SAY NICE THINGS TO THE MACHINE: NOTHING ELSE SEEMS TO WORK. (Author Unknown, viewed 2001)

While NMR instruments are still finicky, several advancements have made modern NMR data much more valuable. Superconducting magnets were introduced in NMR instruments by Varian in 1962, as mentioned above, but a greater advancement was the introduction of pulsed-field Fourier transform (FT-NMR) spectroscopy in the late 1960s. Perhaps because of their extensive use of continuous wave instruments, Varian was late to adopt the now-ubiquitous FT-NMR technique. Competitors, notably Bruker but also JEOL, Nicolet (later absorbed by GE), and IBM instruments (later absorbed by Bruker) were quick to take advantage Varian Associates' hesitation and take over the FT-NMR market (Waugh, 1993). In the 1970s, polarization transfer experiments were developed that would lead to powerful 2-dimensional techniques such as COSY and NOESY (Shoolery, 1993). As 2-dimensional techniques were developed further and field strengths continued to escalate, solving protein structures by NMR become a reality, something few people would have dreamed of in the 1950s.

References

Author Unknown. Note on Varian A-60 NMR spectrometer archived at the Chemical Heritage Foundation, Philadelphia, PA. Viewed January 2001.

Freeman, R. "The Fourier transform revolution in NMR spectroscopy." Anal. Chem. 1993, 65(17), 743A-753A.

Roberts, J. D. Nuclear Magnetic Resonance: Applications to Organic Chemistry. McGraw-Hill: New York, 1959.

Shoolery, J.N. "NMR spectroscopy in the beginning." Anal. Chem. 1993, 65(17), 731A-741A.

Varian Associates. "Instruction manual: Model A-60 analytical NMR spectrometer system" (instrument manual). Varian Associates Instrument Division. 1961-1965.

Waugh, J.S. "NMR spectroscopy in solids: A historical perspective." Anal. Chem. 1993, 65(17), 725A-729A.