Click for next page ( 18


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 17
PART ii

OCR for page 17

OCR for page 17
A Lifetime of Experience in the Growth of Modern Instrumentation for Organic Chemistry By John D. Roberts, Professor Emeritus of Chemistry, California Institute of Technology Dr. Roberts (Ph.D., UCLA, 1944) began his career as a National Research Council fellow and instructor at Harvard University. He joined Caltech as a professor of organic chemistry in 1953 and became chair of Chemistry and Chemical Engineering (1963-1968) and vice president, provost, and dean of the faculty. During his tenure at Caltech, he was a friend and colleague of Arnold O. Beckman. Dr. Roberts is a member of the National Academy of Sciences. He has received the Welch Award, the National Medal of Science (1990), and the ACS Arthur C. Cope Award. Dr. Roberts has authored more than 500 research publications, including 10 books. His current research involves applications of nuclear magnetic resonance spectroscopy to physical organic chemistry. minted DU visible-ultraviolet spectrophotometer in undergraduate research. A rnold Beckman created new ways of analysis that truly revolutionized how chem- ical, biochemical, and medical research are done. Near the beginning of this revo- lution, I used a Beckman pH meter at UCLA in 1938 and subsequently a newly In 1938, organic chemistry was characterizing its products just as for the previous 100 years. For solids: melting points, elemental analysis, and molecular weights. For liquids: boiling points, den- sities, and refractive indices. Indeed, a Zeiss refractometer was our only instrument for characterizing liquids. Later, at MIT, a DU spectrometer served me well, but it was not widely applicable to the compounds I was study- ing. Infrared was better, and MIT's spectroscopist had a Perkin-Elmer single-beam infrared spectrometer, which was more applicable but difficult to use. Here, "simplify, innovate, automate" produced double-beam infrared spectrometers, first marketed by Baird Associates and later by Beckman and Perkin-Elmer. Infrared then took over most organic characterization by storm. However, only 3 to 4 years later, nuclear magnetic resonance (NMR) stirred up its own hurricane of interest. What was different? Both infrared and NMR provide Early refractometer, circa 1920. spectral regions characteristic of particular structural ele- Courtesy of Richard A. Paselk, ments, but NMR offers more of them. The phenomenon Humboldt State University. INSTRUMENTATION FOR A BETTER TOMORROW 19

OCR for page 17
of spin-spin splitting allows determination of FIGURE 5 40-MHz iron magnet, probe, and sample (1954). local structural environments of many kinds of functional groups. Quantitative analysis by infrared requires reference samples and cali- bration, while for NMR, the relative integrated intensities of spectral peaks of NMR spectra can be quite accurate. NMR has the advantage, except for two particulars. First, Arnold infrared absorptions are easier to understand than excitation and detection of nuclear mag- netic states. Second is the cost differential, with NMR being more expensive by a factor of Beckman 10 to perhaps 500 or so on the high end. Cost might seem to depress NMR sales, but chem- istry departments in research universities normally have between $5 million and $15 mil- created new lion or more invested in NMR equipment. Fifty years following the first commercial NMR spectrometers, innovation seems not to be ways of analysis slowing but speeding up. We now trace the evolution of NMR instrumentation from the era of throwing switches, turning knobs, and pressing buttons to computer-driven black that truly boxes with no user-serviceable parts inside. revolutionized How does NMR work? First, know that most NMR spectra are taken of hydrogen nuclei (protons). Hydrogen is generally abundant in organic compounds, and protons are the how chemical, best NMR-active nuclei to observe. Consequently, the first commercial NMR spectrome- ters were primarily focused on proton spectra. biochemical, When an atomic nucleus in a magnetic field is exposed to photons that have an and medical energy corresponding to the difference in energy between two possible orientations research of its magnetic moment, it will res- onate--that is, its magnetic moment will are done. rapidly change orientation, in the process first absorbing energy and then radiating it. Only a finite number of different ori- entations are possible for the magnetic moments of any such nucleus in a mag- FIGURE 4 First published proton NMR spectrum of ethyl alcohol. netic field, each orientation having its own characteristic energy. This process occurs at a very precise frequency, = B0, where is the nuclear constant for the nuclei undergoing absorption of energy and B0 is the strength of the magnetic field at the nucleus. The common way to detect the absorption of energy is with a receiver coil tuned to the frequency, . 20 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 17
One-of-a-kind NMR spectrometers have been constructed in several laboratories, but our con- cern here are the commercial instruments sparked by Felix Bloch and his co-workers at Stanford in conjunction with Varian Associates. Bloch shared a Nobel prize with Edward Purcell (Harvard) for condensed-state NMR. His instrument was geared to liquid samples and well suited for com- mercial development. Initial customers were FIGURE 6 40-MHz console, magnet, and power supply (1954). chemical and petroleum laboratories. DuPont was sufficiently impressed by a 1951 proton spectrum of ethyl alcohol (see Figure 4) to advance $10,000 to Varian to facilitate completion of its first commercial spectrometer. The spectrum shown in Figure 4 is crude but informative in showing three peaks, the area under which is in a ratio of 1:2:3 as suits the structure HO-CH2-CH3. Clearly, the OH proton resonance peak is on the left, the CH2 resonances in the middle, and the CH3 resonances on the right. No other physical procedure is so simple and clear in confirming the structure of a liquid molecule. An early commercial NMR had an electromagnet weighing about 1,500 lbs., water-cooled with 12-in. pole faces, operating at 9,400 gauss (Figure 5). The sample was contained in a 5-mm glass tube surrounded by oscillator and receiver coils at right angles to one another. The console (Figure 6) has the 40-MHz oscillator and receiver controls (driven by knobs, dials, and buttons, not by a computer); to the left is the power supply for the magnet. Atop the console is a Hewlett-Packard audio oscillator--its first product of the instrument revolution. The more detailed alcohol spectrum (Figure 7) illustrates three important things NMR does for organic chemists. First, it shows the same three groups of protons as in Figure 4 separated in frequency by chemical shifts. Chemical-shift differences result largely from differences in diamagnetic shielding of the protons by nearby valence electrons. Chemical shifts are reasonably predictable and useful in structural analyses. FIGURE 7 Proton spectrum of ethanol taken in 1955 with our first generation spectrometer. INSTRUMENTATION FOR A BETTER TOMORROW 21

OCR for page 17
Then, the three groups of protons are resolved FIGURE 9 Varian A-60 (60 MHz) spectrometer, the first real hands-on into multiple peaks, seen in Figure 7, each result- NMR instrument for everyone. Sitting at the spectrometer is Edward ing from spin-spin splittings. Without explaining L. Ginzton, former chairman and CEO of Varian, Inc.; standing left to right is Tim Kingston; Wesley A. Anderson, NMR spectroscopist from the complexities, such splittings tell much about Varian, Inc., who was involved with Fourier-transform NMR; Andy what other magnetic nuclei are close to a nucleus Baker; George Schulke; John Moran; James N. Shoolery, director of of interest, usually one to three connecting the Varian NMR Applications Laboratory, who worked closely with potential customers to show how Varian spectrometers could be bonds away. used in their own applications; and Bob Gang. Photo taken in 1961. Last, note that the OH hydrogen of alcohol in Figure 7 shows no splitting. This is an addi- tional example of NMR's unusual powers, here in connection with reaction rates. If the intermolecular exchange of the OH protons is fast, a single OH resonance will result. With purified alcohol, the OH line becomes a triplet, which indicates that intermolecular exchange occurs less than once every 0.01 seconds. Research on such exchange processes requires good temperature control not available on the early spectrometers. Here, I turned instrument developer and designed a vacuum- jacketed, temperature-controlled probe (Figure 8), which, in improved form, is standard on almost all modern NMR instruments. "Simplify, innovate," was achieved by Varian Associates' A-60 spec- trometer (Figure 9), the first hands-on, easy-to-operate NMR machine. Its major flaw was vacuum-tube electronics. Trouble meant changing and rechanging tubes until order was restored. Despite this, the A-60 was immensely successful and allowed any interested organic chemist to use NMR. The A-60 spectral charts were standard.Varian published two volumes of sample proton spectra, which were very useful, and that was great low-key advertising. At that point in history, one might have concluded that the NMR spectrometer problem had been solved--it only needed modern elec- tronics, and a perfect hands-on NMR machine would emerge. However, new vistas opened up. One was C spectra. The C nucle- 13 12 us has no magnetic moment and no NMR signals. However, 13C is an important nucleus for chemical work, but it has a low abundance in nature, 1.3 percent, and a nuclear moment 1/4 that of protons. Consequently, it gives weak NMR signals at the natural abundance FIGURE 8 Vacuum-jacketed, temperature-controlled NMR probe. level. Acetic acid (Figure 10) gave the first 13 C spectrum at that level. 22 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 17
CO2H The CO2H carbon peak is on the left and the CH3 carbon is CH3 on the right, split into four by the three attached protons. Such spectra, even if noisy and poorly resolved, whetted the interest of chemists in C. 13 Hz 0 850 1,700 However, the very weak 13C signals required improve- FIGURE 1013C NMR spectrum of acetic acid. ments in stability and enhancement by repetition and averaging. The first spectrometer combining ultrastability and time averaging for C looked differ- 13 ent from an A-60. It incorporated a Hewlett-Packard digital-frequency sweep oscillator and was called the DFS-60 (Figure 11). With it, useful spectra could be taken on quite large molecules, such as cholesterol. Spectra of 15N at its low natural abundance (0.3 percent) and a magnetic moment 1/10 that of protons were a greater challenge. The first 15N spectrum taken at natural abundance concentration was the single resonance of liquid hydrazine taken in the DFS spectrometer at 6 MHz. Why not use abundant N nuclei? They give NMR spectra, but the signals are 14 too broad to be useful. Routine 15 N spectra required three major improvements: First, commercial development of superconducting magnets with fields 5 to 15 times stronger were needed to achieve greater magnetization of the N nuclei. 15 Pulse FT NMR was the next giant step in NMR technology, where nuclei are flipped to upper magnetic states by powerful, very short pulses. A short enough pulse (microseconds) excites protons with very different chemical shifts. The decay of the magnetization induced in the sample is recorded digitally and usually takes FIGURE 11 The DFS-60 NMR spectrometer, the first routine13C a few seconds for protons at room instrument, used time averaging. temperature. The result is a free- Photo taken in 1966. INSTRUMENTATION FOR A BETTER TOMORROW 23

OCR for page 17
induction decay (FID) (Figure 12). The frequencies in this decay can be extracted with the Fourier transform (FT). For multiple frequencies, the decay is complex, but FT delivers resonance positions and their relative intensities. Computation of an FT for even a few digitized points is difficult, because many sin and cos values are required. Finally, fast computers were needed before this innova- tion could be implemented for NMR. FIGURE 12 A multifrequency free induction decay (FID) and its Fourier transform. These elements together led to a useful spectrometer for N at the natural abundance level 15 with a superconducting magnet. An N spectrum of the amide nitrogens of vitamin B12 is 15 shown in Figure 13. The peaks go downward, because the constant of N is negative. 15 Another giant step was massaging FIDs with pulses or continuous radiation to change the decaying magnetizations. An example, a two-dimensional correlation spectroscopy (COSY) plot (Figure 14) of nonexchanging ethyl alcohol, shows the chemical shifts and which of the hydrogens split each other. Spots on the diagonal show chemical shifts. The off- diagonal spots denote splittings, if any. Protons separated by three bonds split each other's resonances, but HO and CH3 groups separated by five bonds do not split each other. With complicated molecules, COSY is a very important tool for analyzing splittings. Currently, probably 500 or more programs, like COSY, are available for massaging FIDs to improve the information content of NMR spectra. What next? A relentless war to increase sensitivity by reducing electronic noise in the pickup coils. One way is to cool receiver coils and preamps close to liquid helium temperatures without cooling the sample. Both leading NMR purveyors, Varian and Bruker, offer this complex and expensive option, where the helium used for cooling is recycled. Signal-to-noise is improved by a factor of 3 or 4. An obvious way to increase the sensitivity of NMR detection, which advantageously spreads peaks separated by chemical shifts farther apart, is to use higher magnetic fields. Commercial NMR started with iron magnets at 30 MHz, but now almost all FIGURE 14 Two-dimensional correlation spectroscopy (COSY) plot of ethanol. 24 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 17
commercial spectrometers use superconducting mag- nets. The current highest field is a 900-MHz Bruker H2NCOCH2CH2 CH3 CH3 CH2CONH2 H2NCOCH2 self-shielded magnet. This is a super technological CH2CH2CONH2 accomplishment that provides a 30-fold (compared CH3 N R N CH3 Co+ with 50 years ago) enhancement in field strength. N N CH3 H2NCOCH2 Where do we go from here? One innovative approach OH COCH2CH2CH3 CH3 CH2CH2CONH2 is the extraordinary BOOMERANG spectrometer NH CH2 developed at Caltech by Daniel P. Weitekamp. Figure CH-CH3 N CH3 O O 15 shows a proton NMR prototype about the size of a P O N O HO CH3 coffee mug, operating at 7,000 gauss (27 MHz) with FIGURE 13 Vitamin B12 (left) O 15 a 2.6-mm sample. It is a force-detection NMR HOCH2 and the N spectra of its amide group (right). (FDNMR) spectrometer. How does it work? The magnets are unusually shaped, but it has the customary excitation coil. The detection sys- tem is different. Basically, magnets detect the changes of force on excitation and decay of the magnetic nuclei in the sample. The changes in force are transmitted mechanically to a vibrating silicon plate, the resulting picometer changes in vibration amplitude are detected by an optical interferometer, and the FID is turned into a spectrum by FTs. Where do we need this new detection scheme? It is the preferred method for obtaining spectra of very small samples. The changes in signal-to-noise ratio plotted on a log scale for calculated inductive-coil detection as a function of sample size compared with changes in signal-to-noise ratio for force detection show the latter is much better for very small samples. Weitekamp's goal is to be able to obtain an NMR spectrum for a single molecule. How can the FDNMR spectrometers be made smaller? Put them on a chip! Such chips are being developed at Caltech's Jet Propulsion Laboratory, and we may eventually be able to send NMR spectrometers to Mars! It is fair to conclude that NMR is very much alive, and even after 50 years of commercial development, it is not yet mature. Still a teenager with great promise ahead! FIGURE 15 Cross section of a force-detection NMR (FDNMR) magnet on a chip. INSTRUMENTATION FOR A BETTER TOMORROW 25