NMR and Magnet Technology Development
The struggle for greater sensitivity continues to be the primary motivation for NMR technology development, followed closely by the need to improve spectral resolution. Higher magnetic fields leads to increased signal strength in proportion to B_o² and a decrease in signal averaging time in proportion to B_o³ . Moreover, resolution for homogeneously broadened lines improves linearly with B _o and under a variety of specific circumstances it can improve much more dramatically. Such effects are observed, for example, when the linewidth is dominated by second-order quadrupolar broadening or when linewidth dominated relaxation mechanisms become less efficient at higher fields. Continuous improvement in magnet and probe technology, and in preamplifiers has led to steady improvements in sensitivity. In the decades since 1970, sensitivity as measured on standard samples in state-of-the- art spectrometers of the day, has improved steadily from 10:1 to more than 2000:1 today in conventional probes (Figure 3). This reflects both changes in field strength as well as major instrumental improvements. Recent development of high quality factor (Q) probes show the potential for spectacular improvements in sensitivity above and beyond the present state as do recent developments with microcoils. In addition, polarization transfer methods can lead to even more dramatic sensitivity enhancements for limited applications.

While much progress in the development of high sensitivity NMR probes has been made in the past, there continues to be much untapped potential. Several high temperature superconducting (HTS) prototype probes have been built with very high Q values that demonstrate a five fold improvement in sensitivity for non-conducting samples. However, triple resonance and pulsed field gradient capabilities have not been incorporated into these probes, thereby greatly limiting their range of application. Cryogenic probes using low resistivity metals may have the benefits of the HTS probes without these limitations. Moreover, high Q probes reduce the demands for high sample/ coil filling factors, thereby allowing for reduced sample exposure to electric fields and reduced eddy currents both in the sample and in the coil. GaAs technology is leading to ultra-low noise preamplifiers and to very fast Q switching devices so that both high and low Q probe configurations can be optimally utilized. It is clear that there is much potential for gaining sensitivity through the development of instrumentation in addition increasing magnetic field strengths.

Figure 3. NMR sensitivity as measured with a standard sample in a 5mm diameter probe has increased steadily over the past three decades. The development of a cryogenic probe (*) is providing a significant jump in sensitivity.

Recent advances in NMR microcoil probe technology are particularly attractive for high field applications. Using such coils, fundamental sensitivity per unit sample mass can be increased by up to two orders of magnitude compared to conventional larger-volume probes, giving limits of detection of a few picomoles of material. Excellent resolution can be achieved for biological solution-state preparations. In addition, the small volumes required by these probes considerably relax spatial magnetic field homogeneity requirements, which is important for some of the new magnet technologies described below.

Polarization transfer relies on a source of high polarization that can be transferred to the nuclear spins of interest, sometimes in multiple steps. Specifically, electrons in some form are often used to polarize ¹H nuclei, and the polarization is subquently transferred to the spins of interest - ¹³C, ¹⁵N, etc. Examples of this phenomena include optically excited triplets which are used in photosynthetic reaction centers, xenon, and semiconductor quantum wells. In addition, dynamic nuclear polarization (DNP) methods have been used for many years to produce polarized targets for nuclear scattering experiments by irradiating the electron paramagnetic resonance (EPR) spectrum with microwaves. The microwave radiation induces electron-nuclear spin flips and, via at least three mechanisms, results in significant polarizations. Enhancements of signal strengths of approximately 200 have been achieved with DNP in frozen aqueous solution. This means that experiments that would take approximately 100 years can be performed in a day.

The recent progress in magnet technology has been impressive. The transition from NbTi superconducting wire technology to the more difficult Nb₃Sn technology required almost 10 years. The development of multifilamentary wires, high temperature heat treatment, and persistent joints required great efforts and today this technology is enabling the development of the first 900 MHz (21 T) magnet systems. Above 21 T the current carrying capability degrades so severely that this field strength represents an upper limit for this composite. A new and exciting composite, Nb₃Al, represents an incremental improvement over Nb₃Sn and should have the potential to permit the development of 1 GHz magnet systems. The development of Nb₃Al multifilamentary composites is being actively pursued in Japan but, unfortunately, there is no parallel program in the United States. A fundamental requirement of any national effort in high field NMR magnets must also include a focused materials program targeted at materials critical to both magnet and probe technologies.

Beyond 1 GHz new technology is required. The current carrying capacity of many high temperature superconductors does not decrease significantly even at field strengths as high as 30 T and, therefore, magnet development groups are universally looking to this technology for the development of fields above 23 T. Although magnets of modest field strength (approximately 4 T) have been fabricated from HTS materials, much more development is needed to overcome the engineering hurdles required for fabrication of high quality, long-length material and for the production of high quality joints. At present, a 3 T HTS coil is under construction at the National High Magnetic Field Laboratory that will be tested in a 20 T background field to achieve 23 T. This is part of a phased program for the development of a 25 T (1066 MHz) persistent high resolution superconducting magnet in which the NbTi/Nb₃ Sn widebore 900 MHz magnet presently under construction at the NHMFL will be fitted with an HTS insert to achieve the frequency goal. Higher fields are also possible, but require additional long- term development of the conductors and compatible insulating systems.

A parallel activity is exploring even higher fields that could be achieved through the development of resistive or hybrid (part superconducting, part30 resistive) magnets. A prototype 25 T resistive magnet is operational at the NHMFL. Since the magnet is continuously powered (20 MW), temporal stability of the magnet is a major engineering challenge. The power supply has been extensively filtered, and the inlet cooling water temperature for the magnet is now controlled to ± 0.3°C (at a flow rate of 6000 L/min). Without a flux stabilization coil or field/frequency lock unit in the magnet ± 5 ppm is achieved, and with these correction devices ±0.5 ppm has been demonstrated. The spatial field homogeneity in this 52 mm bore magnet is currently being corrected with shim coils towards a goal of 1 ppm over a 1 cm diameter spherical volume. NMR magnets at still higher fields could be developed along with exploring hybrid magnet technology.

Superconducting magnets are limited in field by the maximum current density the wire can carry at high field, but have the advantage of negligible power consumption. Resistive magnets show opposite characteristics: high power consumption, but virtually unlimited current carrying capability. The two technologies are therefore complementary and can be used together to form a hybrid magnet consisting of a superconducting outer magnet with a large room temperature bore ("outsert") that houses the resistive magnet ("insert"). In spite of their technological complexity, hybrid magnets have proven to be reliable research instruments in the physics community around the world.

The advantages of hybrid magnets can also be applied to the field of NMR research. Two modes of operation are suggested: the development of a 1.5 GHz hybrid using a modest 20 MW insert and a 2.0 GHz hybrid that pushes the hybrid magnet design to its present technological limits with a 40 MW insert. The 1.5 GHz magnet would consist of a 600 mm bore, 10 T superconducting magnet wound from NbTi conductor. The 20 MW, 25 T resistive insert would have four independent coils optimized to generate maximum field and homogeneity. It is proposed here that the superconducting outsert be connected in series with the resistive insert greatly reducing the current ripple as a result of coupling to the much higher inductive load generated by the outsert. Calculations indicate a frequency dependent ripple reduction of 10 at 10 Hz and 100 at 100 Hz. Spatial homogeneity better than 1 ppm can be anticipated. With the lessons learned in achieving 1 ppm homogeneity in the resistive magnets, spatial homogeneity to match the temporal stability can reasonably be anticipated. A 2.0 GHz magnet would be based on the 45 T hybrid currently being constructed at the NHMFL. This magnet has a 600 mm bore 14 T outsert. A 33 T, 40 MW resistive insert with modest homogeneity and stability could be developed to yield a 47 T (2.0 GHz) magnet for NMR experiments.

Hybrid and resistive magnets will provide the first opportunities to study NMR phenomena at very high fields. Not only will many unique experiments be possible for answering scientific challenges in materials science and structural biology using solid- state NMR, but these magnets will also afford opportunities to test NMR theories, to measure relaxation properties, and to separate Zeeman from quadrupolar interactions. These results will not only advance the scientific arenas, but will provide justification for increased efforts to develop persistent, high field (greater than 25 T), high performance superconducting magnets.

It is envisioned that the NMRC will generate a cost effective environment to mobilize distributed human resources and to leverage existing infrastructure in response to scientific, technological and educational challenges ahead. One of the biggest beneficiaries of this connectivity will be in the area of technology development where the communities can nearly instananeously contribute and share technological advances.31