Proton-enhanced Nuclear Induction Spectroscopy

Abstract

A method is described for the detection of the natural magnetic and/or electric quadrupole reasonance frequencies of isotopically rare or chemically dilute nuclei in the presence of at least one abundant nuclear spin species. The free induction decay of the dilute nuclei is directly detected after an applied r.f. field at the Larmor frequency of the dilute nuclei is removed. A high resolution spectrum of the dilute nuclei is obtained by applying an r.f. field at the Larmor frequency of the abundant spin system during the detection interval.

Description

This invention was made in the course of work supported in part by a grant from the National Institutes of Health and a grant from the National Science Foundation.

This invention relates to a procedure for the enhanced detection of the natural magnetic dipole and/or electric quadrupole resonance frequencies, the Larmor frequencies, of isotopically rare or chemically dilute nuclei (to be called S-spins) in the presence of at least one abundant nuclear spin species (to be called I-spins). The method is applicable to solids or solid-like material (polymer suspensions, for example) in which the S-spins are spin-spin coupled to I-spins possessing, in turn, a homonuclear spin-spin coupling.

The technique employs established ideas of double resonance in a novel fashion to obtain an enhanced nuclear induction signal of the rare (S) spin system. The signal-to-noise advantage of this method facilitates the rapid acquisition of NMR (nuclear magnetic resonance) data, making feasible the study of very dilute systems or extremely small samples. The method is general and applicable to the study of both broadline and high resolution magnetic resonance phenomenon. The "high resolution"NMR is accomplished by strong radio frequency (r.f.) irradiation at the Larmor frequency of the I-spins during the recording of S-spin induction signal. The irradiation removes I-spin dipolar interactions as they affect the S-spin signal. The normal broadline signal is obtained in the absence of the I-spin decoupling field. The "high resolution" version enhances the usefulness of the technique as a chemical tool since the strong I-S dipolar interactions typically obscure the information of chemical interest, i.e., chemical shielding parameters.

The enhancement method exploits the abundant spins I (N.sub.I is the abundance, .gamma..sub.I is the gyromagnetic ratio) as a giant reservoir of nuclear polarization to be transferred to the rare S-spins (N.sub.S is the abundance, .gamma..sub.S is the gyromagnetic ratio) bit by bit in a time short compared to any spin-lattice relaxation times of the I system. The I-spin polarization is obtained by placing the sample in the strong laboratory field, H.sub.o, and waiting for a time on the order of the spin-lattice relaxation time. The acquired S-spin polarization is observed as a transient induction decay which is recorded and stored for later processing. The Fourier transform of this time domain transient will yield the "normal" absorption (frequency) spectrum characteristic of the spin system. The gain in power signal-to-noise of this technique over conventional Fourier transform NMR goes as

(N.sub.I /N.sub.s)(.gamma..sub.I /.gamma..sub.S).sup.2

Although the method has been discussed in terms of the magnetic resonance of nuclei, the electric quadrupole analog is just as feasible.

Some pertinent references to the prior art are as follows: S.R. Hartmann and E.L. Hahn, Phys.Rev., 128, 2042 (1962); F.Lurie and C.P. Slichter, Phys.Rev., 133, A1108 (1964); L.R. Sarles and R.M. Cotts, Phys, Rev., 111, 853 (1958); N. Bloembergen and P.P. Sorokin, Phys.Rev., 110, 865 (1958); M.Goldman and A.Landesman, Phys.Rev., 132= 610 (1963); A.Abragam and W.G.Proctor, C.R.Acad. Sci., 246, 1258 (1958); M.Mehring, A.Pines, W-K.Rhim and J.S. Waugh, J.Chem. Phys., 54, 3239 (1971); A.Abragam and W.G. Proctor, Phys.Rev., 109, 1441 (1958).

The objects and featuresoof the invention will become more apparent from the following detailed description when read in conjunction with the following figures:

FIGS. 1(a), 1(b), and 1(c) show sequences of radio frequency magnetic fields at the Larmor frequency of the I-spins for cooling the I-spins by spin locking, spin-locking followed by adiabatic demagnetization in the rotating frame, and pulse preparation of the ordered state, respectively.

FIGS. 2(a) and 2(b) show pulse sequences for the I-spins and S-spins irradiation respectively, which provides a broadline undecoupled spectrum for the S-spins.

FIGS. 2(b) and 2(c ) show pulse sequences for the S-spins and I-spins irradiation respectively, which speeds up cross polarization of the S-spins and provides a broadline spectrum for the S-spins.

FIGS. 3(a) and 3(b) show pulse sequences for the I-spins and S-spins radio frequency fields, respectively, which provides a high resolution spectrum of the S-spins.

FIG. 4(a) shows the broadline spectrum obtained by Fourier transformation of a single free induction decay of natural abundance .sup.13 C in adamantane.

FIG. 4(b) shows the broadline spectrum of (a), but with the enhancement provided by 50 single shot contacts of the cross polarization method.

FIG. 4(c) shows the spectrum with decoupling and cross polarization. It is the result of one proton polarization and 14 I-S contacts.

FIG. 5 is a block diagram of an apparatus suitable to the invention.

The Invention

The method involves a sequence of steps, some of which are familiar, standard NMR techniques combined in an innovative manner and coupled with an original mode of detection. This section will outline the steps of the method and provide several of many alternative means of performing the invention.

Step 1: The I-spin system (abundant) is allowed to acquire a large degree of polarization. This may be accomplished by placing the sample of material, for example, an organic compound with abundant protons .sup.1 H and dilute in .sup.13 C in a strong magnetic field and waiting for several spin-lattice relaxation times of the protons.

Step 2: The I-spin system is strongly cooled in the spin temperature sense in a frame of reference rotating at the Larmor frequency, .omega..sub.o, of the I-spins. (.omega..sub.o = .gamma..sub.I H.sub.o, where .omega..sub.o is the angular frequency of precession, .gamma..sub.I is the gyromagentic ratio of the I nuclear spins, and H.sub.o is the magnitude of the static laboratory field.) This higher ordered, non-equilibrium state will serve as a source of magnetic order made available to the S-spin system (dilute) by means of Step 3.

This cooling process may be effected by one of several traditional means, i.e., spin-locking, spin-locking followed by adiabatic demagnetization in the rotating frame (ADRF), or by a pulsed preparation of the ordered state. FIGS. 1(a), 1(b) and 1(c) show the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins for these three methods.

a. Spin-locking may be performed by applying a 90.degree.-pulse on the I-spins followed by 90.degree. phase shift in the radio frequency (r.f.) carrier and continuous irradiation with the latter phase. The r.f. excitations are provided by a coil whose axis is perpendicular to the Zeeman field. While the r.f. carrier produces an oscillating magnetic field in the coil, it is only one of the circularly polarized components which results in the magnetic transition of importance. The net effect of spin-locking is to take a magnetization, originally parallel to the laboratory Zeeman field, and lock it along a relatively weaker field, H.sub.1 (I), rotating at the Larmor frequency. This ordered spin state appears, in the rotating frame, to have acquired a temperature (T.sub.RF) equal to [H.sub.1 (I)/H.sub.o ]T.sub.L for the case H.sub.1 (I)>>H.sub.L, where T.sub.L was the original spin-temperature of the I system in the laboratory frame and H.sub.L is a measure of the strength of the local dipolar field. For typical values of H.sub.1 (I) and H.sub.o, T.sub.RF can be extremely low, .about.10.sup.-.sup.3 T.sub.L. Adiabatic fast passage stopping on resonance would accomplish the same as spin-locking.

b. Spin-locking followed by adiabatic demagnetization is another means of preparing the desired ordered state and has the particular advantage that there is no continuous irradiation of the I-spins. It is accomplished by a gradual shutting-off of the r.f. carrier in a time long compared to the precession period of the I-spins in the field H.sub.L, i.e., 1/.gamma..sub.1 H.sub.L. FIG. 1(b) demonstrates this technique.

c. The pulsed preparation of the dipolar state (J.Jeener and P.Brockaert, Phys.Rev., 157, 232 (1967)) is an additional means of effecting the cooling, but is less efficient than (a) or (b), i.e., a degree of the initial spin order is lost (about .about. 50 percent). This method does have the dual advantage of no continuous I-spin irradiation without the need for ADRF apparatus. As shown in FIG. 1(c), two pulses are required: the first, a 90.degree. pulse, and the second a 45.degree. pulse. The second follows the first by a time on the order of T.sub.2 and is phase shifted by 90.degree. as in FIG. 1(c).

These are but three of numerous means of cooling the I-spin system.

In the figures 90 degrees indicates the length of duration of the X pulse, i.e., the time it takes the I-spin magnetization to nutate about the X direction in the rotating frame from the Z direction along the magnetic field to the X-Y plane perpendicular to the Z direction. The Y portion of the r.f. excitation corresponds to a 90.degree. phase shift in the r.f. carrier and is the Y direction in the rotating frame.

Step 3: After having cooled the I-spin system by some suitable method, the nuclear species (S) of interest is brought into thermal contact with the I-spin system. This is accomplished by turning on a strong r.f. field at the Larmor frequency of the S-spins. Establishing this thermal contact between the two spin reservoirs results in an ordering of the S-spin system with the corresponding growth of S magnetization along the direction of the rotating component, H.sub.1 (S) of the r.f. field. As with the r.f. fields applied to the I-spin system, these excitations are achieved from a coil transverse to the laboratory static magnetic field. Once the S-spin magnetization is acquired via the I-spin system, one of several modes of operation can be used depending upon what information is desired.

a. The normal, undecoupled free induction decay of the S-spins is observed by abruptly (in a time short compared to T.sub.2) turning off H.sub.1 (S). The transient, precessing magnetization, M.sub.S, induces a detectable voltage in the same coil which generated H.sub.1 (S). In the absence of a strong decoupling field on the I-spins, the signal is the time domain equivalent of the broadline (frequency) spectrum obtained by more traditional means.

Multiple contacts between the S-spin and I-spin systems can be made with the intent of improving the signal-to-noise ratio. After each I-S contact the nuclear induction signal is recorded and stored, preferably, in digital form for averaging with later accumulations. The contacts continue until the I-spin reservoir heats up and further accumulation begins to degrade the signal-to-noise ratio. In this manner it is possible to make many I-S contacts with one I-spin polarization giving a substantial enhancement in signal-to-noise.

FIG. 2 shows the sequence of events pertaining to an undecoupled, multiple contact version of the method. The cooling process is accomplished by spin-locking followed by adiabatic demagnetization in the rotating frame, FIG. 2(a). Pulsing-on an r.f. field H.sub.1 (S) at the Larmor frequency of the S-spins is sufficient to establish thermal contact between the I- and S-spin reservoirs. This is shown in FIG. 2(b): The signal is acquired during the sampling window which shows the nuclear signal as it might appear at the output of the phase detector. To speed up the cross polarization, it may be advantageous to turn on an additional r.f. field at Larmor frequency of the I-spins. As demonstrated in FIG. 2(c), the turning on and off of H.sub.1 (I) is done adiabatically so as not to destroy the magnetic order of the I-spins.

b) The greatest utility of the method for chemical or biological application will perhaps be the "high resolution"adaptation of the single or multiple contact procedure. By "high resolution" is meant the removal of the static dipole-dipole interactions of the I-spin system as they affect the dynamical behavior of the S-spins. The suppression of these interactions is brought about by application of intense r.f. irradiation at I-spin Larmor frequency during the observation of the S-spin free induction decay.

FIG. 3 shows one of many possible solutions to the experimental problem. The I-spins, FIG. 3(a), as before are spin-locked in the usual fashion. Subsequently, the r.f. field is reduced adiabatically to allow rapid thermal mixing between the two spin systems and then raised again to its original strength to decouple the I-spins from the S-spins during the free induction decay of the S-spins. After sufficient signal has been acquired, the r.f. field is shut off adiabatically with the intent of allowing the probe to cool somewhat before the next contact. Typically, one might need a larger H.sub.1 (I) to decouple the I-spins than mix them efficiently with the S-spins. The signal, FIG. 3(b) recorded during the decoupling process, free of I-spin dipolar coupling, might contain the interesting chemical information which might otherwise be obscured, i.e., chemical shifts of the S-spin system. The signal from each contact might be stored in an averager and later fed to a computer for Fourier transformation to obtain the more readily analyzable frequency spectrum.

In the event that the spin-lattice relaxation time in the rotating frame is too short to allow multiple contacts, the single contact version will still yield a substantial signal-to-noise enhancement. For a single contact the polarization enhancement over the usual equilibrium value equals .gamma..sub.I /.gamma..sub.S. For I being protons and S being .sup.13 C or .sup.15 N this factor is .about.4 and .about.10, respectively. Even for the single contact method large signal-to-noise increases are feasible.

c. The S-spin magnetization obtained in a single I-S contact can be utilized also for spin-lattice relaxation studies for signal-to-noise enhancements. The acquired S magnetization is restored anti-parallel to the Zeeman magnetic field by some suitable method like adiabatic fact passage or 90.degree. pulse shifted in phase by 90.degree. from the original S-spin r.f. field. The decay of this magnetization back to its equilibrium value is monitored in the well-known fashion, i.e., by applying a 90.degree. pulse sometime later and recording the transient signal.

An example of the application of method (a) and (b) is shown in FIG. 4. FIG. 4(a) is the broadline spectrum of natural abundance C.sup.13 in adamantane obtained by Fourier transformation of a single free induction decay on a fully equilibrated sample. FIG. 4(b) is the broadline spectra obtained by 50 I-S single contacts (Method a). FIG. 4(c) is the high resolution, multiple contact (Method b) experiment and the result of a single proton polarization. The two lines show clearly the resolution of the two types of carbon lines characteristic of adamantane. The spectra of FIG. 4(a) and 4(c) were acquired in a nearly equivalent amount of time and demonstrate dramatically the power of the invention for chemical inquiry.

The effect of Steps 2 and 3 is to transfer nuclear polarization from the I-spins to the S-spins. The rotating frame, double resonance process is not the only means of accomplishing this. A potentially useful method would be the "solid effect". R.F. irradiation at the sum and differences of the two spins (.omega..sub.s + .omega..sub.I and .omega..sub.I - .omega..sub.s) would provide efficient transfer of I-spin polarization to the S system. The resulting S-spin polarization would be parallel to the laboratory field. A 90.degree. pulse on the S system would provide the nuclear induction signal. Similarly, the rotating frame analog of the "solid effect" experiment is feasible. To effect the transfer of polarization, the I-spins may be "spin-locked" with an r.f. field of amplitude H.sub.1 (I) which is followed with field H.sub.1 (S) on the S-spins and an audio frequency magnetic field parallel to the laboratory Zeeman field with a frequency equal to .gamma..sub.I H.sub.1 (I) + .gamma..sub.s H.sub.1 (S) or .gamma..sub.I H.sub.1 (I) - .gamma..sub.s H.sub.1 (S). The S-spin magnetization is along H.sub.1 (S) and may be observed in the usual manner by abruptly shutting off H.sub.1 (S).

An additional means of transferring I-spin polarization is to allow thermal mixing of the I and S system in weak laboratory magnetic fields. The sample is first brought to equilibrium with lattice in a large Zeeman field. The sample is removed from the large magnetic field to a weak field and returned again to the original field in a time short compared to the spin-lattice relaxation times but long enough to allow mixing of the I-S systems in the weak field. The S-spin free induction decay is acquired in the large laboratory Zeeman field upon applying a 90.degree. pulse to the S-spins.

The Apparatus

A diagram of a double resonance spectrometer capable of performing the enhancement experiments is given in FIG. 5.

The basic operation of the instrument requires inter alia the logic generator which provides the logic levels for opening and closing mixer gates and pulses to trigger the sampling circuits and signal averager.

The spectrometer provides two stable frequencies sources 51 and 52: one, 51, at the Larmor frequency of the I-spins and the second, 52, for S-spins. These frequencies may be derived from crystal oscillators suitably thermostated for time stability. .omega..sub.I is split by a power divider 53 into two channels. One channel passes through a delay line 54 to give the r.f. carrier a 90-degree phase shift relative to the second channel. As mentioned above the two r.f. phases can be used for the spin-locking process. Both channels of .omega..sub.I encounter mixer gates 55, 56 which are slaved either directly (for pulse output) or indirectly (for adiabatic variation of r.f. amplitude) through the ADRF circuit to the logic generator 58. RF passing through the mixer gates 55, 56 goes through a channel combiner 59 and is amplified strongly by a power transmitter 60. When .omega..sub.I reaches the probe 61, it will have acquired sufficient intensity to decouple or mix the I-spins.

Like the r.f. for the I-spins, .omega..sub.s is split into two channels at power divider 62. One channel of .omega..sub.s provides a reference signal to the phase detector 63. The delay line 64 enables the experimeter to pick the particular reference phase he desires. As with .omega..sub.I, .omega..sub.s encounters a mixer gate which again is under the control of the logic generator 58. When .omega..sub.s is pulsed on, it gets amplified on its way to the probe by a second high power transmitter 66. The duplexer 67 protects the receiver 68 from the intense r.f. irradiation of the S-spin transmitter 66. During the reception mode the duplexer switches (in the absence of large .omega..sub.s) to enable detection of signal voltage from the probe. The low-noise receiver 68 amplifies the weak signal and transmits it to the phase detector 63. The NMR signal coming from the phase detector is in the audio-frequency range. It is filtered, sampled discretely in time, and converted to digital form in the A/D converter 69. Each point in time is stored in its corresponding memory channel of the signal averager 70 to be averaged with subsequent accumulations. After a signal has been acquired with the requisite signal-to-noise ratio, it is transferred to a computer 71 for Fourier transformation to provide the more easily interpretable frequency spectrum .

Claims

What is claimed is:

1. A method for improving the signal-to-noise ratio and resolution of the NMR spectrum of a dilute nuclear spin system, the S-spins, in a solid in the presence of a more concentrated nuclear spin system, the I-spins comprising the steps of

1. transfering nuclear polarization from the I-spins to the S-spins and,

2. then applying radiofrequency fields in the neighborhood of the I-spin Larmor frequency to reduce the broadening of the S-spin resonance which would otherwise be caused by said I-spins, and

3. detecting the nuclear induction radiation signal of the S-spins obtained from the S-spin polarization acquired via the I-spins.

2. The method of claim 1 in which step 1 is accomplished by an adaptation of rotating frame, double-resonance procedures, comprising

polarizing said sample ofI-spins of I-spins spin-lattice processes, cooling the I-spins in their rotating frame,

establishing contact between I and S spin system by turning on an r.f. field, H.sub.1 (S), near the Larmor frequency of S-spins, thereby acquiring said S-spin polarization along H.sub.1 (S).

3. The method of claim 1 in which step 1 is accomplished by "solid effect" techniques, comprising

polarizing said system of I-spins by spin-lattice processes,

irradiating the sample with an r.f. field at the sum or difference of the I-spins and S-spins Larmor frequencies, thereby acquiring said S-spins polarization along the laboratory Zeeman field.

4. The method of claim 1 in which step 1 is accomplished by a rotating frame version of the "solid effect" technique, comprising

polarizing said sample I-spins by spin-lattice processes,

cooling the I-spins in their rotating frame by spin-locking said I-spins along an r.f. field, H.sub.1 (I),

simultaneously applying an r.f. field, H.sub.1 (S), to the S-spins and an audio frequency magnetic field parallel to the laboratory Zeeman field direction with a frequency of .gamma..sub.I H.sub.1 (I) + .gamma..sub.s H.sub.1 (S) or .gamma..sub.I H.sub.1 (I) - .gamma..sub.s H.sub.1 (S), thereby acquiring said S-spins polarization along H.sub.1 (S).

5. The method of claim 1 in which step 1 is accomplished by mixing the I and S spin systems in weak magnetic fields, comprising

polarizing said sample of I-spins by spin-lattice processes in a large laboratory magnetic field,

removing the sample to a weak magnetic field thereby enabling I-S mixing,

returning the sample to the large magnetic field thereby acquiring S-spins polarization along the laboratory Zeeman field.

6. The method of claim 1 in which step 2 is accomplished by continuous wave irradiation of the sample near the I-spin Larmor frequency.

7. The method of claim 1 in which step 2 is accomplished by pulsed irradiation near the I-spin Larmor frequency comprising

a train of 180.degree. pulses spaced in time of the order of the shorter of the I-spin or S-spin spin-spin relaxation times.

8. The method of claim 1 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

9. The method of claim 2 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

10. The method of claim 3 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

11. The method of claim 4 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

12. The method of claim 5 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

13. The method of claim 6 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

14. The method of claim 7 comprising, in addition,

acquiring said S-spin polarization a plurality of times,

detecting said S-spins free induction decay after each polarization to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal from a single I-spin polarization.

15. The method of claim 1 comprising, in addition,

Fourier transforming of said averaged S-spins signal to provide the frequency spectrum of the S-spins.

16. The method of claim 1 comprising, in addition,

applying said I-spins field excitation a plurality of times corresponding to the number of S-spins field excitations,

said I-spins excitation being gradually applied to avoid destroying the I-spin order and of substantially full magnitude before said S-spin excitation is removed,

said I-spins excitations being gradually reduced to substantially zero amplitude after said free induction decay signal is detected and before said S-spin excitation is applied again, the amplitude of said I-spins excitation being sufficient to remove the dipolar coupling of the I-spins.

17. The method of claim 2 comprising, in addition,

applying said I-spins field excitation a plurality of times corresponding to the number of S-spins field excitations,

said I-spins excitation being gradually applied to avoid destroying the I-spin order and of substantially full magnitude before said S-spin excitation is removed,

said I-spins excitations being gradually reduced to substantially zero amplitude after said free induction decay signal is detected and before said S-spin excitation is applied again, the amplitude of said I-spins excitation being sufficient to remove the dipolar coupling of the I-spins.

18. The method of claim 1 comprising, in addition,

maintaining said I-spin r.f. magnetic field during the time said S-spin r.f. magnetic field is applied, and

adiabatically reducing said I-spin field to zero before said S-spin field is abruptly turned off.

19. The method of claim 1 comprising, in addition,

maintaining said I-spin r.f. magnetic field during the time said S-spin r.f. magnetic field is applied, and

adiabatically reducing said I-spin field to zero before said S-spin field is abruptly turned off.

20. The method of claim 1 comprising, in addition,

adiabatically reducing the amplitude of the I-spins r.f. magnetic field when said S-spins magnetic field excitation is applied,

adiabatically increasing the amplitude of the I-spins r.f. magnetic field prior to turning off said S-spins field excitation,

maintaining said increased amplitude of said I-spins r.f. magnetic field during the time said S-spins free induction decay signal is being detected in order to remove the dipolar coupling of the I-spins,

adiabatically reducing said I-spins r.f. magnetic field to zero after said S-spins signals are not being detected,

applying said S-spins field excitation a plurality of times,

the amplitude of said I-spins field excitation during said plurality of S-spin excitation fields being at the amplitudes recited previously during and between said S-spin excitation fields with adiabatic changes in said amplitudes of the I-spin fields.

21. The method of claim 2 comprising, in addition,

adiabatically reducing the amplitude of the I-spins r.f. magnetic field when said S-spins magnetic field excitation is applied,

adiabatically increasing the amplitude of the I-spins r.f. magnetic field prior to turning off said S-spins field excitation,

maintaining said increased amplitude of said I-spins r.f. magnetic field during the time said S-spins free induction decay signal is being detected in order to remove the dipolar coupling of the I-spins,

adiabatically reducing said I-spins r.f. magnetic field to zero after said S-spins signals are not being detected,

applying said S-spins field excitation a plurality of times,

the amplitude of said I-spins field excitation during said plurality of S-spin excitation fields being at the amplitudes recited previously during and between said S-spin excitation fields with adiabatic changes in said amplitudes of the I-spin fields.

22. The method of claim 20 wherein said I-spin magnetic field amplitude and the S-spin magnetic field amplitude during the time said S-spins field is applied satisfy the Hartmann-Hahn condition, .gamma..sub.I H.sub.1 (I) = .gamma..sub.s H.sub.1 (S), for a time sufficient to most rapidly establish spin thermodynamic equilibrium.

23. The method of claim 21 wherein said I-spin magnetic field amplitude and the S-spin magnetic field amplitude during the time said S-spins field is applied satisfy the Hartmann-Hahn condition, .gamma..sub.I H.sub.1 (I) = .gamma..sub.s H.sub.1 (S), for a time sufficient to most rapidly establish spin thermodynamic equilibrium.

24. The method of claim 1 comprising, in addition,

applying the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins by spin-locking by applying a 90.degree. pulse followed by a 90.degree. phase shift in the frequency of the radio frequency carrier which is continuously applied,

applying and removing said S-spins excitation a plurality of times,

detecting said S-spins free induction decay field after each removal to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal.

25. The method of claim 2 comprising, in addition,

applying the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins by spin-locking by applying a 90.degree. pulse followed by a 90-degree phase shift in the frequency of the radio frequency carrier which is continuously applied, which

applying and removing said S-spins excitation a plurality of times,

detecting said S-spins free induction decay field after each removal to provide a plurality of signals,

averaging the S-spin signals to provide an averaged S-spin nuclear induction signal.

26. The method of claim 1 in which the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins comprises

spin-locking by applying a 90.degree. pulse followed by a 90.degree. phase shift in the frequency carrier.

27. The method of claim 2 in which the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins comprises

spin-locking by applying a 90.degree. pulse followed by a 90.degree. phase shift in the frequency carrier.

28. The method of claim 1 in which the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins comprises

spin-locking followed by adiabatic demagnetization by applying a 90.degree. pulse followed by a 90.degree.phase shift in the radio frequency carrier followed by a gradual shutting off of the carrier in a time long compared to 1/.gamma..sub.1 H.sub.1.

29. The method of claim 2 in which the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins comprises

spin-locking followed by adiabatic demagnetization by applying a 90.degree. pulse followed by a 90.degree. phase shift in the radio frequency carrier followed by a gradual shutting off of the carrier in a time long compared to 1/.gamma..sub.1 H.sub.1.

30. The method of claim 1 in which the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins comprises pulsed preparation of the dipolar state by

applying a first 90.degree. pulse followed by a 45.degree.pulse,

said 45.degree. pulse follows the 90.degree. pulse by a time on the order of T.sub.2 and is phase shifted by 90.degree..

31. The method of claim 2 in which the sequence of radio frequency magnetic fields at the Larmor frequency of the I-spins comprises pulsed preparation of the dipolar state by applying a first 90.degree. pulse followed by a 45.degree.pulse, said 45.degree. pulse follows the 90.degree. pulse by a time on the order of T.sub.2 and is phase shifted by 90 degrees.