Method for two field nuclear magnetic resonance measurements

Abstract

A method for carrying out two-field nuclear magnetic resonance (=2FNMR) measurements involves preparing a sample (9a) in a first working volume (5) of a highly homogeneous magnetic field with a first field strength; transferring the sample (9a) to a second working volume (7) with a magnetic field having lower homogeneity and having a second field strength, wherein the first field strength is at least 2 Tesla larger than the second field strength; manipulating the sample (9a) at the second working volume (7) by applying a sequence of radio-frequency (=RF) and/or field gradient pulses; transferring the sample (9a) back to the first working volume (5); and detecting an NMR signal of the sample (9a) in the first working volume (5). The method allows for NMR experiments with which more and/or improved quality information about an investigated sample can be obtained.

Claims

1. A method for carrying out two-field nuclear magnetic resonance (=2FNMR) measurements on a sample, the method comprising the steps of: a) preparing the sample during a preparation period in a first working volume of a first magnetic field with a first field strength and having a homogeneity better than 0.01 ppm; b) transferring, during a first transfer period, the sample to a second working volume with a second magnetic field having a homogeneity of 100 ppm or better or in a range between 2 ppm and 100 ppm, and having a second field strength of 0.05 Tesla or larger or of 0.1 Tesla or larger, wherein the first field strength is at least 2 Tesla larger than the second field strength; c) manipulating, during a manipulation period, the sample at the second working volume by applying a sequence of radio-frequency (=RF) and/or field gradient pulses, wherein the sequence is insensitive to or minimizes effects of an inhomogeneity of the second magnetic field in the second working volume; d) transferring the sample back to the first working volume during a second transfer period; and e) detecting an NMR signal of the sample in the first working volume.

2. The method of claim 1, wherein in step c), the RF and/or field gradient pulse sequence is insensitive to or minimizes the effects of the inhomogeneity of the magnetic field in the second working volume by exciting and/or filtering nuclear spin transitions with energy differences independent of local variations of a strength of the magnetic field in the second working volume.

3. The method of claim 1, wherein in step c), the RF and/or field gradient pulse sequence excites zero- or multiple-quantum coherences of nuclear spins of the sample in the second working volume.

4. The method of claim 1, wherein the steps a) through e) are repeated with the sample several times, for obtaining a multidimensional NMR measurement which displays at least one dimension based on a development and/or manipulation of nuclear spins of the sample in the second working volume and at least one dimension based on a development and/or manipulation of nuclear spins of the sample in the first working volume.

5. The method of claim 4, wherein, in each step c), at least one time parameter of the sequence of RF and/or field gradient pulses is set to a different value.

6. The method of claim 1, wherein the 2FNMR measurement includes a two-field zero-quantum NMR spectroscopy experiment, with building-up and preparing polarization of nuclei of a specific type during step a), with a sequence of RF pulses, delays and field gradient pulses generating and filtering zero-quantum coherences of a specific type during step c), and with detecting single-quantum coherences during step e).

7. The method of claim 6, wherein the nuclei of a specific type comprises 13C nuclei.

8. The method of claim 1, wherein a total duration of the first and second transfer periods and the manipulation period is less than a spin-lattice relaxation time of nuclear spins of the sample manipulated in the second working volume during step c).

9. The method of claim 1, wherein the first and the second transfer period each have a duration of less than 100 ms.

10. The method of claim 1, wherein a sequence of RF and/or field gradient pulses is applied to the sample in the first working volume during step a).

11. The method of claim 1, wherein step e) includes applying a single or a sequence of RF and/or field gradient pulses to the sample in the first working volume before starting NMR signal acquisition.

12. The method of claim 1, wherein step c) is triggered by position sensors sensing a position of the sample in the second working volume and/or that step e) is triggered by position sensors sensing a position of the sample in the first working volume.

13. The method of claim 1, wherein NMR probes at the first and second working volume are synchronized.

14. The method of claim 1, wherein the preparation period of step a) is longer than a spin relaxation time in the first working volume of the nuclear spins of the sample manipulated during step c).

15. The method of claim 1, wherein, during step a) and/or step c) and/or step e), RF and/or field gradient pulses are applied to different types of nuclei of the sample.

16. The method of claim 15, wherein nuclear spins of the different nuclei are manipulated at different working volumes.

17. The method of claim 1, wherein, during step a) and/or step c) and/or step e), RF and/or field gradient pulses are applied only to nuclei of the sample which have an electric quadrupole moment of zero.

18. The method of claim 17, wherein the nuclei of the sample are 1 H, 13C and/or 15N.

19. The method of claim 1, wherein during step c), a pulse sequence performs a transfer of polarization in networks of nuclei through scalar coupling interactions or a transfer of polarization by isotropic mixing.

Description

BRIEF DESCRIPTION OF THE DRAWING

(1) FIG. 1 shows a pulse sequence used to calibrate I=1 H, 13C and 15N pulse lengths at low field;

(2) FIG. 2 shows a 1-Dimensional 1H spectrum measured at 0.33 T on a sample of 1 M glycine in D.sub.2O;

(3) FIG. 3 shows a pulse sequence of a two field heteronuclear single quantum coherence experiment (2F-HSQC), in accordance with the invention;

(4) FIG. 4 shows a HSQC 1H/13C spectrum of 1 M Glycine (13C and 15N labelled), in accordance with the invention;

(5) FIG. 5 shows a two-field zero-quantum NMR spectroscopy pulse sequence, in accordance with the invention;

(6) FIG. 6 shows two-field zero-quantum NMR spectroscopy 2D 13C.sub.HF/13C.sub.LF spectra of an 1 M glycine sample, in accordance with the invention;

(7) FIG. 7 shows highly folded two-field zero-quantum NMR spectroscopy 2D 13C.sub.HF/13C.sub.LF spectra of an 1 M glycine sample, in accordance with the invention, with (a) 1D 13C.sub.LF spectrum extracted at (13C.sub.HF)=172.267 ppm; and (b) 2D two-field zero-quantum NMR spectroscopy spectrum, with spectral window in the indirect dimension reduced to 10 ppm; and

(8) FIG. 8 a schematic cross section of an inventive apparatus for carrying out a 2FNMR measurements on a sample, in accordance with the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENT

(9) The present invention relates to high-resolution two-field nuclear magnetic resonance spectroscopy.

(10) Nuclear magnetic resonance (NMR) is a unique spectroscopic technique to characterize the properties of matter in exquisite detail at atomic resolution. Over decades of methodological developments, many NMR experiments have been proposed to probe physical and chemical properties at a particular magnetic field and observe the NMR signals at a different field. Here, the inventors introduce a two-field NMR spectrometer that permits the manipulation of spin systems and the observation of signals at two magnetic fields. Coupled with a fast pneumatic shuttling system, the two-field spectrometer permits the acquisition of two-dimensional spectra where the indirect and direct dimensions are acquired at different magnetic fields. In spite of the moderate field inhomogeneity of the low-field center, the inventors introduce a zero-quantum 2D experiment that offers high-resolution in both chemical shift dimensions.

(11) The ability of nuclear magnetic resonance (=NMR) to probe the chemical and physical properties of matter at atomic resolution makes it a ubiquitous spectroscopic tool in molecular chemistry, material science, structural biology and medicine. Such achievements were made possible by the enhanced resolution offered by the introduction of two-dimensional NMR and subsequent multi-dimensional experiments. Magnetic-field dependent properties have been probed with many techniques that explore two (or more) magnetic fields over the course of a single experiment. Fast field-cycling relaxometry, zero-field NMR and dynamic nuclear polarization, among others take great benefit from such schemes. However, no experiment has been proposed so far where chemical shifts are recorded at two different fields. The availability of very high magnetic fields (1 GHz and above) challenges the usual rule that higher fields are necessarily better for all spin systems (e.g. carbonyl carbon-13 relaxation due to the chemical shift anisotropy becomes a major limitation at such fields). Most multi-dimensional NMR experiments include chemical shift evolution delays that differ in the optimal magnetic field at which they should be recorded. Such optimal magnetic fields can be extremely high (e.g. to reduce effects of second order quadrupolar couplings), but some optimal magnetic fields are nowadays accessible (i.e. close to 14.1 T for carbonyl single-quantum evolution in a large protein or 1 GHz for TROSY).

(12) Here, by way of example a two-field NMR spectrometer is introduced, in which a low magnetic field center, including a homogeneous magnetic field, a triple-resonance probe with single-axis gradient is installed in the stray field of a commercial 14.1 T (600 MHz) spectrometer. Transfer between the two fields is possible with the use of a pneumatic sample shuttle system. In the following, this system is briefly described, and by way of example a series of two-field NMR experiments is introduced, including a two-field heteronuclear single-quantum coherence (2F-HSQC) experiment as well as a two-field zero-quantum/single quantum correlation experiment (two-field zero-quantum NMR spectroscopy) that provides high-resolution spectra in both dimensions.

(13) Two-Field NMR Spectrometer

(14) The two-field NMR spectrometer with which the experimental results described below were obtained is designed from a commercial NMR spectrometer (here operating at 14.1 T, 600 MHz proton resonance frequency) augmented by a series of accessories: a ferroshim system designed to obtain an homogeneous magnetic field (at 0.33 T) in the stray field of the 14.1 T superconducting magnet; a low-field triple resonance probe equipped with a single-axis gradient; a system to generate and amplify radiofrequency pulses at low frequency; and a pneumatic sample shuttle used to ensure fast transfer between the two magnetic centers. Here follows a brief description of each element.

(15) Ferroshims:

(16) The ferroshim system includes pieces of iron placed around the second magnetic center. This design used is based on U.S. Pat. No. 8,154,292 B2, but is refined with the use of ferroshim foils and a dedicated temperature regulation. A set of room temperature shim coils increases the homogeneity of the magnetic field in the second magnetic center. The distribution of magnetic fields in the active volume of the low-field center can be evaluated from a proton spectrum: full width at half height is usually close to 10 ppm (see below).

(17) Low-Field Probe:

(18) This triple resonance probe consists of two coils: the inner coil has a three winding saddle coil geometry and is optimized for the proton frequency (14 MHz). The outer coil has a seven winding saddle coil geometry and is optimized for carbon-13 (3.52 MHz) and nitrogen-15 (1.42 MHz). In addition, this probe is equipped with a coil for z-axis pulse field gradients.

(19) Low-Field Spectrometer:

(20) Current Advance NMR spectrometers can operate down to a frequency of 6 MHz. A new low-frequency signal-generating unit was developed to enable the generation of radiofrequency (RF) pulses with defined frequencies, phases and amplitude profiles below 6 MHz. Two low-frequency 200 W amplifiers are used for RF amplification. The six-channel spectrometer allows for the combination of up to three channels at both high and low field (e.g. .sup.1H, .sup.13C, .sup.15N) or four channels at high field (.sup.1H, .sup.13C, .sup.2H, .sup.15N) and two channels at low field (e.g. .sup.1H, .sup.13C or .sup.1H, .sup.15N).

(21) Sample Shuttle:

(22) The sample shuttle is based on a pneumatic transport mechanism described by Charlier C, et al. (2013) Nanosecond Time Scale Motions in Proteins Revealed by High-Resolution NMR Relaxometry. J Am Chem Soc 135(49): 18665-18672. A shuttle guide (i.e. the tube that contains the moving sample shuttle) was designed to fit tightly between the high-field and the low-field probes.

(23) Radiofrequency Pulses at Low Field.

(24) Nutation experiments have been performed for each nucleus (.sup.1H, .sup.13C, .sup.15N) in order to calibrate the amplitude of radiofrequency pulses at low field and characterize the homogeneity of RF fields. For enhanced sensitivity, the polarization builds up during the recovery delay at high field, the sample shuttle is then moved to the low field position (B.sub.low=0.33 T) and a single pulse is applied. Coherences are suppressed by the strong gradient during the transfer to the high-field spot (B.sub.high=14.1 T). The acquisition follows a 90 pulse at high field.

(25) FIG. 1 shows a pulse sequence used to calibrate I=.sup.1H, .sup.13C and .sup.15N pulse lengths at low field. The first filled rectangle represents the variable pulse at low field (LF) with a nutation angle , and the second one is a 90 pulse at high field (HF). .sub.rec was between 2 and 10 s, depending on the nucleus. .sub.H-L was 100 ms, and to .sub.L-H=100 ms. .sub.s=300 ms and .sub.s=200 ms are stabilization delays.

(26) These experiments have been performed on two samples: 4 M .sup.13C and .sup.15N labelled urea in 90% D.sub.2O and 1 M .sup.13C and .sup.15N labelled urea in 100% D.sub.2O. Results of these calibrations are shown in Table 1. Nutation profiles are available for each nucleus in the supplementary information. High power pulses, with amplitudes ranging from 12.5 to 25 kHz make highly broadband pulses readily accessible at low magnetic fields.

(27) TABLE-US-00001 TABLE 1 Channel Power 90 degree pulse H-1 0.5 W 10.2 s C-13 100 W 11.1 s N-15 50 W 20 s

(28) Table 1 describes the calibration of radiofrequency pulses at 0.33 T with polarization and acquisition at high field.

(29) FIG. 2 shows an 1D .sup.1H spectrum measured at 0.33 T on the sample of 1 M glycine in D.sub.2O.

(30) FIG. 2 displays a typical 1D proton spectrum measured at low field, with polarization at high field. The magnetic field homogeneity is not sufficient to obtain proton high-resolution spectra. However, it is sufficient to ensure that broadband radiofrequency pulses can be used on all nuclei.

(31) Two-Field-HSQC.

(32) The two-field heteronuclear single quantum coherence experiment (2F-HSQC) was first run to demonstrate the principle of two-field NMR spectroscopy. The original HSQC experiment was modified to provide a correlation between carbon-13 nuclei at low field and protons at high field. The pulse program is shown in FIG. 3. The shuttle transfer occurs before the first 90 pulse on carbon-13 in the INEPT sequence, with the polarization stored as a C.sub.zH.sub.z term. An antiphase coherence on carbon evolves under the chemical shift, magnetic fields inhomogenities and homonuclear couplings at low fields, while the scalar coupling with protons is refocused by a 180 pulse on the proton channel. Polarization is stored again as a C.sub.zH.sub.z term for the transfer back to high fields for detection on the high-field proton channel.

(33) FIG. 3 shows a pulse sequence of the 2F-HSQC (.sup.1H.sub.HF, .sup.13C.sub.LF) experiment. All narrow (filled) and open (wide) rectangles represent 90 and 180 pulses, respectively. Pulse phases are along the x-axis of the rotating frame unless otherwise mentioned. .sub.a=1/(4J.sub.CH)=1.7 ms; is chosen to obtain a perfect echo on the carbon-13 at low field for the first increment of .sub.1; .sub.H-L=100 ms and .sub.L-H=100 ms. The stabilization delay .sub.s=200 ms allows for convection currents and vibrations to settle. Composite pulse decoupling was performed with a WALTZ-16 scheme and RF amplitudes of 1 kHz on the carbon channel. The phase cycles were: .sub.1={x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x}; .sub.2={x, x}; .sub.3={x, x, x, x}; .sub.4={x, x, x, x, x, x, x, x}; .sub.5={x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x}, .sub.acq={x, x, x, x, y, y, y, y}.

(34) The 2F-HSQC experiment was performed on a 1 M carbon-13 and nitrogen-15 labelled sample of glycine. This 2F-HSQC spectrum is shown on FIG. 4. As in the proton one-dimensional spectrum (see FIG. 2), the line width is dominated by magnetic field inhomogeneities. However, the doublet due to the one-bond scalar coupling of the alpha carbon-13 of glycine with the carboxyl carbon-13 is partially resolved. This coupling is 55 Hz, or 16 ppm for carbon-13 at 0.33 T.

(35) FIG. 4 shows a HSQC .sup.1H.sub.HF/.sup.13C.sub.LF spectrum of 1 M Glycine (.sup.13C and .sup.15N labelled).

(36) Two-Field Zero-Quantum NMR Spectroscopy.

(37) In order to reduce the contribution of magnetic field inhomogeneities to the line width at low field, a correlation experiment between single-quantum coherences at high field and zero-quantum coherences at low field was implemented. Zero- and multiple-quantum coherences have been used for many years to obtain high-resolution spectra in inhomogeneous magnetic fields. Here, the inventors introduce a two-field zero-quantum NMR spectroscopy sequence, which was adapted from a refocused INADEQUATE experiment.

(38) The two-field zero-quantum NMR spectroscopy pulse sequence, shown in FIG. 5, corresponds closely to the gradient-selected refocused INADEQUATE experiment and only differences will be discussed. Most of the pulse sequence takes place at low field, in order to reduce off-resonance effects of carbon-13 pulses. A band-selective 180 pulse is applied to invert carbonyl/carboxyl carbon-13 polarization at the start of the sequence at high field in order to maximize the polarization of the zero-quantum coherence. The shuttle transfer to low field occurs with carbon-13 longitudinal polarizations. A bipolar pulsed-field gradient is applied at low field after the ti chemical-shift labelling period to select zero-quantum coherences and populations. After the refocusing element, a 90 pulse on carbon-13 brings back polarizations along the z-axis for storage during the shuttle transfer to high fields. At high field, a carbon-13 90 pulse is followed by the signal detection. Proton composite pulse decoupling is applied during the whole experiment except when the sample is outside of either probe (i.e. during shuttle transfers) for both decoupling and carbon-13 polarization enhancement by nuclear Overhauser effects. A single 180 pulse is applied on the nitrogen-15 channel during the .sub.1 delay. Frequency sign discrimination in the indirect dimension was performed with a TPPI scheme.

(39) FIG. 5 shows the two-field zero-quantum NMR spectroscopy pulse sequence. Narrow black and wide open rectangles represent 90 and 180 pulses, respectively. Pulses are applied along the x-axis of the rotating frame unless otherwise indicated. The bell-shaped pulse on the high-field carbon-13 channel is a 44 ms rectangular inversion pulse that leaves carboxyl carbon-13 unperturbed. The delay .sub.a=J.sub.CC=1.7 ms. Composite pulse decoupling was performed with a WALTZ-16 scheme and RF amplitudes of 1 kHz on both the HF and LF proton channels. Gradient pulses were applied along the z-axis and had durations of 1 ms and amplitudes: G.sub.1=20% and G.sub.2=0% of the maximum amplitude. In the 2F-INADEQUATE variant, G.sub.2=40% of the maximum amplitude, and the initial band-selective inversion pulse was omitted. The phase cycle is: .sub.1=x, x; .sub.2=x; .sub.3=x, x, x, x; .sub.4=x, x, x, x, x, x, x, x; .sub.acq=x, x. Frequency sign discrimination in the indirect dimension was performed with a TPPI scheme applied on .sub.2.

(40) FIG. 6 shows the two-field zero-quantum NMR spectroscopy 2D .sup.13C.sub.HF/.sup.13C.sub.LF spectra of a 1M glycine sample.

(41) The two-field zero-quantum NMR spectroscopy experiment was performed on a 1 M glycine (.sup.13C and .sup.15N labelled) sample. Results are show in FIG. 6 and FIG. 7. Due to the evolution on the zero quantum coherence, the effects of magnetic field inhomogeneities at low field are suppressed in the indirect dimension. A narrow signal can be observed in FIG. 6 at 158 ppm, that is 132 ppm from the maximum frequency (which is at 280 ppm) in the indirect dimension. Zero-quantum coherences are immune to the TPPI shift of frequencies in the indirect dimension, so that resonance frequencies should be measured from the edge and not the center of the spectrum. This zero-quantum chemical shift is then equal to the expected zero-quantum chemical shift of 132 ppm. The broad spectral window used in FIG. 6 precludes the observation of the natural line width in a reasonable experimental time. A two-field zero-quantum NMR spectroscopy spectrum, highly folded in the indirect dimension (with a 10 ppm spectral window), was recorded and is shown on FIG. 7. The line width is dramatically reduced in the indirect dimension, with a full width at half height equal to 0.17 ppm, or 0.6 Hz. This is about two orders of magnitude narrower than single quantum lines, measured in simple one-dimensional spectra (FIG. 2) or in two-dimensional correlations (FIG. 4).

(42) FIG. 7 shows highly folded two-field zero-quantum NMR spectroscopy 2D .sup.13C.sub.HF/.sup.13C.sub.LF spectra of a 1M glycine sample. (a) One dimensional .sup.13C.sub.LF spectrum extracted from the 2D two-field zero-quantum NMR spectrum at (.sup.13C.sub.HF)=172.267 ppm. (b) 2D two-field zero-quantum NMR spectrum, the spectral window in the indirect dimension was reduced to 10 ppm in order to reach the full natural resolution of the experiment. No apodization function was used in the indirect dimension.

(43) A two-field NMR spectrometer is presented by way of example, where one magnetic field center is generated by a commercial superconducting magnet, while the second magnetic center is an engineered plateau of magnetic field in the stray field of the superconducting magnet. Both magnetic centers are equipped with triple-resonance and single-axis gradient probes. A six-channel NMR console is used for radiofrequency pulse generation and fast motion between the two magnetic centers is achieved with a pneumatic shuttle device. Two-field pulse sequence programming is conveniently integrated so that most high-resolution techniques can be adapted on this system. Pulse sequences are introduced where RF pulses and evolutions under chemical shifts are performed at either of the two fields. In particular a 2F-HSQC sequence allows for the measurement of correlation spectra where the direct and indirect dimensions are recorded at two distinct magnetic fields. In the two-field zero-quantum NMR spectroscopy experiment, a zero-quantum chemical shift evolution takes place at low field and suppresses the effects of the moderate magnetic field inhomogeneity of the low-field center. The resulting two-field NMR spectra display high resolution in both dimensions. This opens the way to the development of a new library of pulse sequences designed, in particular, to observe chemical shift evolutions at the most relevant frequency for each different nucleus.

(44) FIG. 8 shows an apparatus 1 for carrying out an inventive 2FNMR method.

(45) The apparatus 1 comprises a cryostat 2 containing a magnet configuration 3. The magnet configuration 3 comprises superconducting coil sections (not shown) cooled by liquid helium in a tank 4.

(46) The magnet configuration 3 generates a magnetic field in a first working volume 5 inside a room temperature bore 6 of the cryostat 2; the first working volume 5 surrounds the magnetic center of the magnet configuration 3. The magnetic field within the first working volume 5 has a homogeneity better (lower) than 0.01 ppm, for example 0.005 ppm. The homogeneity is typically achieved by means of a suitable design of the magnet configuration 3 and a shim coil system (not shown). In the example shown, a first field strength of the magnetic field (B.sub.0 field), which is oriented parallel to the magnetic axis A, is about 14 Tesla in the first working volume 5.

(47) The stray field of the magnet configuration 3 further generates a magnetic field at a second working volume 7, also within the room temperature bore 6. The stray field alone would exhibit a significant magnetic field gradient along the magnetic axis A within the second working volume 7. In order to obtain a magnetic field within the second working volume 7 with a homogeneity of 100 ppm or better (lower), for example of 50 ppm, the apparatus 1 comprises a device 8 for homogenizing the magnetic field, which is here designed as a system of ferroshims located in the room temperature bore 6. The ferroshims surround the second working volume 7 here. In the example shown, a second field strength of the magnetic field (B.sub.0 field), which is again oriented parallel to the magnetic axis A, is about 0.3 Tesla in the second working volume 7.

(48) A sample carrier 9 carrying a sample 9a to be investigated may be transported between the first working volume 5 and the second working volume 7 by means of a for transporting the sample 9a. The device 10 for transporting the sample 9a here comprises a straight tubular guide 11 in which the sample carrier 9 may move, and further two pieces of auxiliary tubing 12a, 12b for applying gas pressure at the opposing ends of the tubular guide 11. By means pressure differences between the ends of the tubular guide 11, the sample carrier 9 may be accelerated within the tubular guide 11. The gas pressure is controlled via shuttle controller 13. By means of position sensors 20, an arrival of the sample carrier 9 at a desired working volume 5, 7 may be verified.

(49) The first working volume 5 is surrounded by a first NMR probe 14, and the second working volume 7 is surrounded by a second NMR probe 15. In the shown example, the first NMR probe 14 is connected to a first NMR signal excitation and detection hardware 16, and the second NMR probe 15 is connected to a second signal excitation and detection hardware 17. Hardware 16 and hardware 17 are connected to an NMR timing device 18. Here, each of the two NMR probes 14, 15 may be used for both sending RF and/or gradient pulses into the sample 9a at their respective working volume 5, 7, and for detecting RF signals form the sample 9a in their respective working volume 5, 7. Hardware 16, hardware 17 and the NMR timing device 18 belong to an NMR console 19.

(50) By means of the apparatus 1, a sample 9a may be located, manipulated with RF and field gradient pulses, and measured at both working volumes 5, 7. The working volumes 5, 7 have magnetic field strengths with a significantly different absolute value, namely with a difference of at least 2 Tesla, to allow a physically different behavior of the sample 9a at the different working volumes 5, 7. The device 10 for transferring the sample 9a allows a quick change between the working volumes 5, 7, in particular such that a typical NMR signal decay during a change is negligible (such as with less than 10% signal loss during a change).