EP1853935A2

EP1853935A2 - Method for enhancing the nmr signal of a liquid solution using the long-range dipolar field

Info

Publication number: EP1853935A2
Application number: EP06709116A
Authority: EP
Inventors: Hervé DESVAUX; Lionel Dubois
Original assignee: Commissariat a lEnergie Atomique CEA
Current assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2005-01-27
Filing date: 2006-01-18
Publication date: 2007-11-14
Also published as: FR2881226A1; WO2006079702A2; WO2006079702A3; FR2881226B1; US7564243B2; US20080204014A1; JP2008528991A

Abstract

The invention relates to an NMR analysis system and method for obtaining an NMR signal from a liquid solution (1) enhanced in relation to the value that it would have on the basis of the thermodynamic equilibrium. It has been discovered in a surprising manner that if the spins of the cores (9) of a highly-polarised polarisation source (3) and the spins of the cores (7) of the liquid solution (1) are irradiated such that a polarisation transfer, such as a Hartmann-Hahn transfer, of the spins of the cores (9) of the source (3) to the spins of the cores (7) of the liquid solution (1) is carried out by means of a coherent coupling resulting from the dipolar field created by the cores of the source, an NMR signal of the liquid solution is obtained, said signal being significantly enhanced in relation to the value it would have on the basis of the thermodynamic equilibrium.

Description

Method for increasing the NMR signal of a liquid solution

Field of the invention

The invention relates to a system and method for NMR analysis, for obtaining an NMR signal of an increased liquid solution with respect to that which it would have starting from the thermodynamic equilibrium.

Prior art

Nuclear magnetic resonance "NMR" is a spectroscopic technique that allows, especially in the liquid phase, a fine analysis of the material. Its spectral resolution capability makes it possible, for example, to assign a resonance line to a nucleus or group of nuclei of a molecule. We are thus able to determine or verify the chemical constitution of the medium. In practice, this great fineness of resolution results from the low energies involved, since typically the waves inducing the transitions are in a frequency range from a few kilohertz to several hundred megahertz. The energy ΔE of these transitions is deduced from the Hamiltonian Zeeman which corresponds to the interaction between the applied magnetic field BQ and the nuclear magnetic moment whose gyromagnetic ratio is γ: where Ti is Planck's constant. In practice, the resonance frequency is equal to the precession frequency of the nuclear magnetic moments in the magnetic field BQ; this pulsation is ω = γBQ. By abuse of language and according to the uses in the field, we will speak of frequency for ω. In the same way, the gyromagnetic factor γ being the proportionality factor between the nuclear magnetic moment and the nuclear spin, we shall speak indifferently of spin or moment magnetic. At a few parts per million, the resonance frequency of an isotope is a constant characterized by γ.

At present, the magnitude of the permanent magnetic field B ₀ created by the magnets implies that the energies involved are less than about 10 ^-5 J. When this energy is compared with thermal energy at room temperature (about 3.10 -2 ⁾ J), it is deduced that nuclear magnetic resonance is not sensitive spectroscopy. Considering to simplify a spin core Vz and n _a and n β the states of spin + Vz and -Vz respectively, a simple calculation shows that in the regime of high temperatures, at the thermodynamic equilibrium the polarization P is: _p . ⁿ a - * β ₌ ≠ ^β Q ₍₂₎ n _a + n β 2kT with k the Boltzmann constant and T the temperature.

In NMR, the signal is proportional to the global magnetization M which corresponds to the sum of the contributions of each nuclear magnetic moment:

M ₊ n ^ _≠ J "" ⁺ "> ^{) rVB} ° (3)

2 ° ^β AkI

This magnetization is always weak at thermodynamic equilibrium and therefore the signal is weak, which implies, for example, to use samples of high concentrations.

Any method that makes it possible to increase the magnetization improves the sensitivity of the NMR thus allows a saving of time or an improvement of the precision of the measurements or opens the way to studies on samples of smaller sizes. Conventionally, the most exploited solution is to increase the value of the magnetic field BQ. This solution is however technologically limited. On the one hand, the value of the critical fields of the known superconducting materials limits the amplitude of the field. This The limit is currently around 3OT while the 21T magnets are already on the market. On the other hand the use of resistive magnets suffers from problems of spatial homogeneity and temporal stability. Hybrid technologies (superconducting magnet + resistive magnet) are being validated, however the problems of spatial homogeneity and temporal stability due to the resistive component will remain. These limits prohibit high resolution NMR applications for chemical analysis in liquids with this type of magnets.

Another solution, when the concentration and not the amount of sample is the limit of sensitivity, is to increase the number (n + n) of nuclear magnetic moments by increasing the useful volume in the detection NMR probe.

An alternative approach is to address this signal problem from the point of view of noise. The reduction of the electronic noise, in particular produced by the probe and the amplifier of the detected signal, makes possible the detection of the NMR signal. The most successful solution today is to lower the temperature of the coil to about 2OK and that of the preamplifier to about 77K.

Another solution which has proved to be rich in possibilities since it hardly depends on technical constraints is to excite and detect only the spins with the greatest gyromagnetic ratio. From the point of view of non-radioactive nuclei, it is the most abundant isotope of hydrogen, the proton ¹ H. To access a higher γ, we must consider the unpaired electrons. But the frequency resolution in Electronic Paramagnetic Resonance is not that of NMR. Thus, this solution consists in exploiting the long coherence durations of an excited system of nuclear spins and the interactions between these magnetic moments to transfer the polarization from one type of nucleus to another. There are in practice three different approaches to perform this transfer, two exploit a bilinear interaction between the two types of spins whose temporal mean is non-zero, it is therefore coherent transfers, the last implements the cross relaxation, that is to say the random interactions of zero time average, and thus a transfer of incoherent type. The latter approach is mainly known as the Overhauser nuclear effect (nOe), where, at least for its first implementations, the principle of the transfer consisted in saturating by radiofrequency field irradiation a species of spin and to be observed on the other variations of intensity. It is thus possible, by saturating the protons ¹ H, to increase the polarization of the ¹³ C by a factor of 4 with respect to its thermodynamic value.

In the category of coherent transfers, the first approach is to excite a first species of spin A, and to let the magnetization of the latter spin freely in the plane perpendicular to the magnetic field BQ. Due to the existence of a bilinear interaction between A and the other spin X species, antiphase two-spin coherences of type A _y X ₂ will be created. Two pulses make it possible to transform them into two-spin transverse coherences on the X-ring of type A _z X _y ; the latter, under the effect of the same bilinear interaction, can be detected. The magnetization detected on X, originating from the initial polarization of A by the polarization transfer, produces the same factor 4 in the case of a spinning pair ¹³ C- ¹ H. The transfer described is of the INEPT type, others implementations still exist based on the concept of multiple excitations and free evolutions under the effect of bilinear interactions. The other approach to perform a coherent transfer is to simultaneously apply two continuous radiofrequency field irradiations or composites on A and X and whose amplitude has been carefully selected. In the Hartmann-Hahn conditions, which are specified below and which fix the frequency and the amplitude of the radiofrequency field irradiations, a transfer takes place which differs from the previous one because it characterized by a phase transfer, i.e., directly from A _x to X _x . The gain in polarization is the same [see documents 1, 2 and 3].

Finally there are two types of bilinear interactions involving two different spins; scalar coupling that requires the existence of chemical bonds between A and X and the dipole coupling that takes place through space. The latter depends on the orientation of the internuclear vector with respect to BQ and j / ~, where r is the distance between the two spins

/ r nuclear. For distances less than about 30 ° and in the liquid phase, the Brownian motion makes the orientation of this vector random and this coupling can not contribute to a coherent evolution, but simply to an incoherent evolution in the form of dipole relaxation. This point differentiates the NMR from the liquid of the solid NMR for which the dipolar coupling leads to coherent transfers. It is therefore admitted that in the liquid phase there exists only one type of bilinear interactions between different spins: the scalar coupling across the links.

However, still in the liquid phase, and when we consider the dipolar contributions from the distant nuclei, the dipolar interaction no longer vanishes due to the Brownian motion. The resulting dipolar contribution is however weak because of the distance, it is nevertheless non-zero since all the spins of the sample contribute. In practice, the effect of this contribution is detectable only if the sample is highly concentrated and the polarization, therefore the magnetic field BQ, is important. The dipole coupling felt by the protons of water and created by these same protons is typically of the order of one Hz in a magnetic field of the order of 1OT. Because of this contribution, after excitation of the magnetization and free evolution, two-spin coherences involving two different molecules are created and can be detected. Being a long-range effect, nothing requires that both The molecules are in the same container, and in practice Warren and colleagues have observed coherences involving two molecules belonging to different containers [see document 4]. All of these effects are known as the effect of the dipolar field in the liquid phase.

In addition, various solutions have been proposed for preparing nuclear spin systems whose populations of the + V2 and -V_> states differ greatly from their thermodynamic values. The lifespan of these highly polarized systems depends on many internal and external parameters but can reach several days to several months. This gave rise to the hope that the use of these highly polarized systems could become an alternative to the low nuclear polarization specific to NMR.

To create this strong nuclear polarization, the concept of polarization transfer is used, and a more strongly polarized system must be used. The weakness of the coupling between the nuclear spins and the other degrees of freedom of the system limits the number of solutions.

A first solution is to transfer the polarization from highly polarized electron spins. The gyromagnetic ratio of a free electron is in fact about 600 times higher than that of the proton. Dynamic Nuclear Polarization (DNP) exploits this method. It requires the presence of paramagnetic species and requires that irradiation affecting the resonance of unpaired electrons be applied. Its implementation for high magnetic fields encounters the difficulty of creating high power microwave irradiation, the problem of the penetration of these microwaves into the material and the relative weakness of the electronic polarization at the temperature. room. It is therefore a method that is mainly interesting at low temperatures. Moreover, the spectral resolution observed in NMR often suffers from the presence of the centers. paramagnetic. However, a solution has recently been proposed which consists in polarizing a solid sample at low temperature (below 4K) by adding radicals and then heating it very quickly by dissolving it in a liquid in order to study its NMR signal. liquid at room temperature.

A second solution consists in using the degrees of freedom of rotation of a molecule in gas phase and their quantification. The use of parahydrogen is an application that allows the signal to increase when the molecule of H ₂ in its para state is added to a carbon-carbon double bond.

A third solution is based on optical pumping to create directly (pumping by metastability of helium 3) or indirectly (pumping by spin exchange) a significant polarization of the nuclei. In practice, strong polarizations (greater than 0.01) can only be achieved if the longitudinal relaxation is very inefficient. This leads to the use of rare gases of ^sp h spins, helium 3 or xenon 129.

In the latter cases (pre-polarization of the sample by DNP, parahydrogen and rare gas polarized by laser), the polarization of the samples thus obtained can be kept for several minutes to several hours and thus the sample moved to a place of measurement . The sensitivity of the NMR experiments using a sample pre-polarized by DNP, parahydrogen, helium 3 or laser-polarized xenon is thus significantly improved (by a factor greater than 10 ³ ). On the other hand, only these isotopes are polarized, but the spins of the nuclei not directly involved in the polarization process do not see their polarization directly affected. As a result, the ability to chemically analyze any molecule does not benefit from a gain in sensitivity.

Currently, in liquid phase, the only known approach to use the strong polarization of these polarized gases in order to increase the NMR sensitivity of the proton or any other nucleus is the Spin Polarization Induced Nuclear Overhauser Effect (SPINOE) method described, for example, in US Pat. No. 6,426,058. In this approach, the magnetization transfer is caused by crossed dipolar relaxation. between dissolved xenon and spatially close hydrogen nuclei. Implementing relaxation, this process is incoherent in nature.

Qualitatively, the physical functioning of this process is as follows. The xenon polarized by optical pumping is dissolved in the solution to be studied. As the magnetic field BQ is defined, this is equivalent to saying using equation (2) that the spin xenon temperature T _Xe is in absolute value very low. Initially, the thermodynamic equilibrium polarization of protons defines their spin temperature as being equal to the temperature of the sample. The fact that xenon and protons are close (some Angstroms) induces the existence of a dipolar coupling term between these two magnetizations. The temporal average of this coupling is in fact zero in the liquid phase because of the Brownian movement. However, in a time-dependent second-order perturbation theory, this dipolar coupling induces the existence of a cross-relaxation between these magnetizations which depends on λ / ₆ where r is the distance between the two nuclear spins. The very character

cold nuclear spins xenon will therefore induce a change in the proton spin temperature and thus a change in its polarization, and therefore potentially a signal gain. The lowering capacity of the proton spin temperature will therefore depend on the xenon-proton cross relaxation rate, the magnetization of the xenon dissolved relative to that of the protons and the direct relaxation rate of xenon and that of the protons. These last two actually define the ability of xenon and proton to maintain or acquire a spin temperature different from that of the sample. However, this approach requires very strong xenon magnetization and is only effective in two cases.

The first concerns the case where the longitudinal relaxation of the proton is very slow. In this case, despite the weakness of the xenon-proton cross relaxation rate, the final magnetization of the proton may be very different from that at thermodynamic equilibrium. This is the case, for example, in the case of deuterated benzene.

The second concerns the case where xenon has a particular affinity for the molecule. Under these conditions, some xenon - proton distances may be short and the affinity of xenon for this site large enough for efficient transfers to be observed. This type of case occurs for molecules with a hydrophobic cavity. The maximum signal gain observed so far

(factor 1.6 in a field of 11.7T) corresponds to a proton nuclear polarization of 6.4 10 ^-5 .

The major drawback of this SPINOE method is the fact that the polarization transfer between xenon and proton can only work when the nuclei are separated by extremely small distances of the order of 3 to 6 Angstroms because of the dependence velocity of cross relaxation in V _b . Such conditions are so restrictive

they make extremely difficult any industrial application.

Finally, and in order to be complete, there are two other approaches for solid-state NMR applications aimed at transferring the polarization of xenon to other nuclei. In both cases the physical process of transfer involves xenon adsorbed on the surface of the solid and the nuclei of this same surface. Both use static dipole interaction (non-zero time mean). The process consists of condensing polarized xenon on the surface of the solid to be studied and then carrying out a Hartmann-Hahn type polarization transfer between the xenon and the solid surface cores [see documents 5 and 6] or via a thermal mixture. The latter consists in reducing the magnetic field BQ to a sufficiently low value so that the dipolar coupling between the xenon and the nuclei of the solid is greater than the difference in resonant frequency between the two types of nuclei. In both cases the transfer only directly affects the nuclei of the surface of the solid.

Object and summary of the invention

The invention thus proposes producing a system and a method for NMR analysis of a liquid solution comprising non-zero nuclear spin nuclei, which overcomes the drawbacks of the prior art.

This object is achieved in that the inventors have surprisingly found that if the spins of the nuclei of a source of polarization and said spins of the nuclei of said liquid solution are irradiated so as to carry out a polarization transfer. of the Hartmann-Hahn type of the spins of the nuclei of said source at the spin of the nuclei of said liquid solution by means of a coherent coupling resulting from the dipole field created by the nuclei of the source, then an NMR signal of said liquid solution is obtained which is increased significantly in relation to the value it would have from the thermodynamic equilibrium.

This dipolar field is a long-range field which is advantageously obtained by using a high core concentration of the polarization source (typically greater than 0.01 mol.l ^-1 ) and / or a strong nuclear polarization of these same nuclei (typically greater than 0.01).

Thus, the average dipolar field created by the polarization source makes it possible to have a coherent coupling exerted at distances of several millimeters, against the 3 to 6 Angstroms of the incoherent coupling of the prior art, between the spins of said nuclei. of the liquid solution and those of the polarization source. Advantageously, said polarization source can be separated from said liquid solution by a distance, representative of the effective range of said dipolar field, which is equal to or greater than 1 nanometer and which can be between 3 nanometers and 10 millimeters. The implementation of this coupling makes it possible, in a liquid medium, to transfer the polarization of the source to said nuclei of the liquid solution, without requiring any particular affinity between the atoms or molecules of the polarization source and those of said solution. The gain in polarization makes it possible to increase the NMR signal of a sample of the liquid solution substantially, thus allowing a fine analysis of this sample.

In order to carry out this Hartmann-Hahn type polarization transfer, a first radiofrequency field irradiation, the frequency of which is close (difference typically less than 10OkHz), of the resonance frequency in the magnetic field BQ of said nuclei of the liquid solution is applied at the same time as a second radiofrequency field irradiation, whose frequency is close (typically less than 10OkHz difference) of the resonant frequency of the nuclei of the polarization source. The shape of these two irradiations can be of constant amplitude, frequency and phase (continuous irradiation) as well as amplitudes, frequencies and variable phases during the duration of the irradiation (composite irradiation).

The description of a Hartmann-hahn type polarization transfer is generally made in the double rotating reference frame, ie the one for which the two radio frequency fields are independent of time. In the absence of interactions between the spins, the nuclear magnetization then describes a precession perpendicular to the axis of the effective field in this frame rotating at a frequency that depends on the amplitude of the irradiation. In order to make the Hartmann-Hahn type transfer possible, it is preferable that, in this frame of reference, the frequency difference of preœssion of the spins of the polarization source and the spins of said nuclei of the liquid solution due to the effect of the two radiofrequency field irradiations is less than ten times the amplitude of the dipolar field created by the nuclei of the source of polarization and felt by the spins of the nuclei of said solution. The conditions are optimal when this same difference is smaller than the amplitude of this average dipolar coupling. In addition it is desirable to use a highly concentrated polarization source and highly polarized because the dipolar field created is more intense so the previous conditions are less drastic.

The factor for increasing the polarization of the nuclei of said liquid solution with respect to the thermodynamic value is all the more important as the nuclear polarization rate of the nuclei of the source is large and the nuclei of the source are much more numerous. than those of said solution.

The nuclei present in the liquid solution may be any nucleus having a nonzero nuclear spin, in particular the isotope 1 of the hydrogen nuclei (proton) or the isotope 13 of the carbon nuclei. However, it is favorable to carry out the polarization transfer towards nuclei of the solution having a high gyromagnetic ratio since the dipolar coupling between the nuclei of the source and the nuclei of the solution is linearly proportional to the gyromagnetic ratio of the nuclei of the solution. It is still possible anyway after this first transfer to perform a second transfer between these nuclei of said solution and other nuclei scalarily coupled to this, for example perform a ¹ H transfer to the ¹³ C. However, note that the factor The increase in polarization obtained by the process described in the present invention is all the stronger as the initial polarization of said nuclei is low, ie their gyromagnetic ratio is low. Any efficient method capable of creating a highly polarized and / or highly concentrated nuclear spin system can be used to prepare the polarization source. This can in particular be obtained by addition of para-hydrogen on a carbon-carbon double bond, or preliminary polarization of a sample by the low-temperature solid phase dynamic nuclear polarization method followed by the dissolution of the sample in a liquid . The polarization source may also consist of a rare gas polarized by optical pumping and selected from all the following gases: xenon, helium, neon, krypton, and mixtures thereof. Xenon 129 and helium 3 being the only two rare gases of spin Vi it is desirable to use them since the polarization rates accessible by optical pumping exceed 10%. In addition, the high solubility of xenon in certain organic solvents or the capacity to liquefy it is favorable to obtain a very high concentration of xenon, which is capable of creating a strong dipolar field. Moreover, the high gyromagnetic ratio of helium 3 (2.75 times that of xenon) and the very high polarization rates obtained (up to 80%) are extremely favorable for the use of this atom in the gas phase. Indeed, both the growth rate of the polarization of said nuclei of the solution and the dipolar coupling between these nuclei and the nuclei of the source are proportional to the gyromagnetic ratio of the nucleus of the polarization source. Also, it is all the more favorable to use a polarization source using a type of core with a very large gyromagnetic ratio. The nuclei of the polarization source may be dissolved in said liquid solution or, on the contrary, physically separated from this same liquid solution. In the latter case the distance separating the polarization source from the liquid solution must be minimal so that the dipole field created by the polarization source is maximum. It is preferable that this minimum distance is less than ten millimeters. The state The source of the polarization source in this compartment separate from the said liquid solution may be arbitrary, the nuclei of the source may be dissolved in a liquid, or in the form of a gas or even condensed into a liquid solution. However, it is favorable to have a high concentration and an effective renewal of the polarization source after applying the polarization transfer method described here, to use a compressed gas phase.

The two compartments respectively intended to contain the liquid solution and the polarization source may be separated by a distance equal to or greater than 1 nanometer and, preferably, between 3 nanometers and 10 millimeters.

The invention also relates to a system for performing an NMR analysis of a liquid solution comprising nuclei with non-zero nuclear spins, the system comprising a magnet surrounding: a source of polarization,

polarization transfer means adapted to irradiate the spins of nuclei of said source of polarization and said spins of the nuclei of said liquid solution, so as to perform a Hartmann-Hahn type polarization transfer of the spins of the nuclei of said source to spinning the nuclei of said liquid solution by means of a coherent coupling resulting from the dipole field created by the nuclei of the source, to obtain an NMR signal of said increased liquid solution relative to its value at the thermodynamic equilibrium. The polarization means comprise:

a first irradiation means adapted to subject the spins of the nuclei of said liquid solution to a first irradiation of radio-frequency fields of the continuous or composite type, and a second irradiation means adapted to subject the spins of the nuclei of said polarization source to a second irradiation of continuous or composite radio frequency fields.

According to one embodiment, the system comprises, inside said magnet, a tube containing said source of polarization dissolved in said liquid solution.

According to another embodiment, the system comprises, inside said magnet, a container comprising two adjacent and independent compartments of each other, which are respectively intended to contain said source of polarization and said liquid solution.

According to yet another embodiment, the system comprises, inside said magnet, a container comprising two internal and external compartments, said outer compartment being intended to contain said source of polarization and surrounding said internal compartment which is intended to contain said liquid solution.

Brief description of the drawings

Other characteristics and advantages of the present invention will emerge more clearly on reading the description of particular embodiments, made with reference to the appended drawings and in which:

FIG. 1 very schematically illustrates a polarization system for increasing the polarization of the nuclei of a liquid solution with respect to its thermodynamic value, according to the invention;

FIGS. 2A and 2B show very schematically two examples of polarization systems of FIG. 1 with a polarization source physically separated from the liquid solution;

FIG. 3 very schematically illustrates one embodiment of the system of FIG. 1;

FIG. 4 shows a pulse sequence affecting the spins of the nuclei of the polarization source and the liquid solution used. showing the existence of the coupling between these two types of nuclei, according to the invention;

FIG. 5 illustrates the variation of the resonance frequency of the protons of the solution as a function of the number of pulses applied to the xenon cores, in order to determine the amplitude of the dipolar coupling according to the invention;

FIG. 6 illustrates the variation as a function of the number of pulses applied to xenon cores, the proton resonance frequencies contained in two distinct compartments, only one of which contains polarized xenon, according to the invention;

FIG. 7 illustrates a sequence of pulses used to observe the polarization transfer of xenon to protons of said solution, according to the invention; and

FIG. 8 shows spectra obtained after polarization transfer from xenon to protons using negative and positive xenon polarization, according to the invention.

Detailed description of embodiments

In accordance with the invention, FIG. 1 very schematically illustrates a polarization system for increasing the nuclear polarization of the nuclei of a liquid solution 1 with respect to its thermodynamic value. This results in an increase of the NMR signal. This system comprises a source of polarization 3, a sample of the liquid solution 1 to be studied and a means for transferring the polarization 5. Note that the liquid solution 1 must have nuclei 7 having a non-zero nuclear spin, for example nuclei hydrogen isotope ¹ H ₇ carbon ¹³ C or others.

Any polarization source which comprises nuclei 9 having identical highly polarized (P greater than 0.01) and highly concentrated (concentration c greater than 0.01 mol L ^-1 ) spins can be envisaged. Thus, this polarization source 3 may consist of a rare gas polarized by optical pumping chosen from the following set of gases: xenon (Xe), helium (He), neon (Ne), krypton (Kr), or their mixtures. Advantageously, the polarization source 3 comprises xenon (Xe 129) or helium (He 3) polarized by laser or hyperpolarized which are the only two rare gases of spin ¹ Zz, therefore for which polarizations greater than 1% can easily be obtained. It will be noted that the polarization source 3 can also be derived from an addition of parahydrogen on a carbon-carbon double bond, or by prior polarization of a sample via the low temperature nuclear dynamic polarization process.

For reasons of simplicity in the description, we will consider hereinafter a polarization source 3 comprising xenon "129" polarized by laser 9 (hereinafter referred to as xenon) and a liquid solution 1 comprising isotope cores "1" hydrogen 7 (i.e., ¹ H protons). In the following, in order to describe mathematically the physical processes involved, we consider a system consisting of N xenon nuclei I * and a proton S.

Thus, for any proton 7 of solution 1, there is a dipolar interaction 11 between each xenon Ϋ 9 and this proton 7. The Hamiltonian which describes this interaction is:

where μo is the magnetic permeability of the vacuum (μo = 4π) O ^-7 S. L), f {θ, φ) is a function that depends on the orientation (θ, φ) of the vector

internuclear I * S with respect to the axis of the magnetic field B ₀ .

However, because of the liquid phase Brownian motion, the orientation of the internuclear vector varies. This has the consequence of zeroing this dipolar interaction. Also this interaction only contributes to through relaxation (effect on the width of the resonance line). However, for the Brownian motion to be effective in averaging this interaction, the internuclear vector between xenon 9 and proton 7 must be able to take all directions in space. The relative diffusion of xenon 9 and proton-bearing molecule 7 ( ¹ H) during the acquisition time thus defines a sphere around proton 7 within which the average of the intermolecular dipolar interaction is zero. There therefore remains a contribution of the distant xenon atoms. The dipolar contribution from a distant xenon depends on the distance r between the xenon 9 and the proton as a decrease in \ \ ₃ r

The contribution felt by the proton S is in fact the sum of all the individual contributions of each xenon:

In practice, as the concentration of xenon is important, one can take a continuous model and replace the discrete sum by integration over the entire volume of the sample; this compensates for the decay in U /, of the dipole contribution. Two types of

phenomena can then support. In the first, we are interested only in the proton and so we can factor in the equation (5), the contributions of each xenon and finally take the average value of I ^A. Everything returns, given the proton S, to the existence of an additional magnetic field contribution parallel to B ₀ . This additional magnetic field B is created by the set of xenon atoms and is felt by the proton. The other type of phenomenon appears when one simultaneously excites protons and xenons. In this case, the same equation (5) shows that there is a coupling between the two spin systems. The existence of this medium dipole coupling (Equation 5) implies the existence of a coupling between the spin system This coupling is now coherent since it directly influences, for example, the resonant frequency of the protons 7 across the B-field and not mainly their linewidth. In fact, this dipolar contribution will be all the more important as the amount of xenon 9 will be important and the polarization of this xenon 9 will be important. This phenomenon is described as the dipolar field created by the whole of the polarization source 3 on the protons 9. This dipole field depends on the geometry of the container of the polarization source 3. The value of this magnetic field created by this magnetization Xenon 9 via dipolar interaction is:

B = - ^ M _Xe (6)

where, Mχ _e is the magnetization of xenon 9 per unit volume. It is expressed by:

M _Xe = 1000c5VP- ^ (7)

with the concentration of xenon 9 in mol. L ^"1 , w the number of Avogadro and ~ Y ₂ ≤ ξ ≤ \ a factor which depends on the geometry of the sample, and which has for example been totally tabulated for samples of ellipsoidal shape.

Under the polarization and concentration conditions described in the invention, the magnetization associated with this xenon 9 is very important, several times that of the protons 7 of the water placed in a magnetic field of 14T at room temperature. Field B (Equation 6) can indeed be of the order of some 10 ^"7 T.

However, as the Hamiltonian associated with this dipolar coupling (Equation 5) between the spin systems of the polarization source 3 (xenons) and the liquid solution 1 (protons) commutes with the Hamiltonian Zeeman, it is necessary to use a sequence of pulses so that, because of this coupling, the magnetization of the protons 7 and that of the xenon 9 are very intimately related and a polarization transfer can take place. This is achieved in the invention by means of a Hartmann-Hahn coherence transfer [see document I].

Thus, according to the invention, a Hartmann-Hahn type polarization transfer is carried out between the polarization source 3 (xenon) and the nuclei 7 (protons) of the liquid solution 1 via the coherent coupling resulting from the long-distance dipole field. created by the source of polarization 3 and felt by the protons.

Indeed, the polarization transfer means 5 comprises a first irradiation means (for example obtained by means of a first coil) and a second irradiation means (for example obtained by means of a second coil). The first irradiation means generates a first radiation 13 continuously or composite RF field whose amplitude is ωf / χ _s and the frequency ω * / 2π, more commonly referred ωξ is close to the resonant frequency ω _s spins of the cores 7 in the magnetic field BQ and the second irradiation means generates a second continuous or composite radiofrequency field irradiation whose amplitude is Iγ, and whose frequency llπ, more commonly referred to as is close to that of the resonance ω _λ of the spins of the cores 9 of the polarization source 3 in the magnetic field B ₀ . It is better that the differences

CO * - CO _< and COn - CO remain respectively less than 10 rads and preferable to be less than 10f and 10ω, ^J , respectively. Thus, the polarization transfer is achieved by applying the first continuous irradiation 13 or composite radiofrequency fields on the spin of the nuclei 7 of the liquid solution 1 and the second continuous irradiation 15 or composite of radiofrequency fields on the spins of the nuclei 9 of the polarization source 3.

To ensure that the dipolar interaction does not switch with

Zeeman, it is advantageous for the first and second irradiations 13 and 15 to have a difference in amplitude ω between them; less than ten times the amplitude of the dipole coupling δ between the spins of the nuclei of the polarization source 3 and those of the liquid solution 1:

d = - γ _H B = ^ -1 OOOξcW V _7x JiYn (8)

By way of example, the first and second irradiations 13 and 15 may have between them a difference in amplitude less than once the amplitude of the dipolar field created by the polarization source 3 and felt by the spins of the nuclei 7 of the liquid solution 1.

In fact, continuous irradiations 13, 15 or series of composite pulses affecting protons 7 of the liquid solution 1 and xenons 9 of the polarization source 3 cause the effective fields in the double reference frame to be sensed by the xenon 9 and the proton 7 (reference rotating to for proton 7 and turning at ωl for xenon 9) coincide at best; the scale being the dipole coupling δ between the xenon 9 and the proton 7.

Thus, according to these conditions, the dipolar interaction no longer commutes with these Zeeman Hamiltonians:

U (9) in the double reference rotating with OZ axes aligned with the effective fields. The system then evolves under the effect of dipole coupling δ.

Moreover, since the polarization source 3 has a very large number of xenon atoms 9, a statistical process appears which makes the spin temperatures of xenon 9 and protons 7 re-coupled in the rotating frame will converge to the same value. It follows a cooling of the proton spins system 7 and therefore an increase in the polarization of the latter and their associated magnetization therefore a signal gain.

In fact, the conditions defined here correspond to an extension of Hartmann-Hahn's conditions [see document I]. These are indeed: ωl = ω ^} and = ωf. They correspond to the cases of a continuous irradiation. Thus, the invention consists in carrying out the polarization transfer from xenon to protons using the dipole coupling that exists between them. For this, it is necessary to make sure to obtain an effective Hamiltonian in the double rotating reference which is the dipolar coupling. As in fact we are interested in the type I _Z S _{Z part} of this interaction (Equation 5), the dipolar recoupling conditions are in fact similar to those required to perform a heteronuclear Hartmann-Hahn type magnetization transfer, where it is the interactions through the chemical bonds that ensure the coupling between the two baths. This last case has been the subject of numerous studies [see document 3]. The major differences between these two types of transfer are the amplitude of the coupling, which is much lower here than the scalar coupling encountered in an NH or CH pair (greater than 90 Hz), and its highly non-local character, since all xenons must to be considered. By the similarity of the form of the coupling Hamiltonian, all Hartmann-Hahn transfer sequences using composite irradiations are likely to be used.

WALTZ-16, MLEV-16, WALTZ-8, DIPSI-2 sequences. DIPSI-2- ⁺ -, MLEV-17, SHR-I, MGS-I, MGS-2 or even based on adiabatic pulses can thus be used effectively [see document 3]. However, it should be noted that all the composite or multi-component pulses for which there is no effective field are less effective than those with effective field. This results from the fact that in the first case the contribution of the dipolar energy will not be negligible, that the effect of the imperfections in the pulses to create the effective Hamiltonian will be very quickly not negligible compared to δ and that finally for a heteronuclear coupling, a Isotropic transfer is 50% slower than effective field transfer [see document 3]. The major advantage of using these multiple pulse sequences is their tolerance to the effects due to the difference between the frequency of the irradiation 13, (respectively 15, aξ) and the resonance frequency of the spin I ₁ ω ¹ (respectively 9, ω ^s ). This offers the ability to simultaneously polarize all the protons of a molecule.

Thus, the principle according to the invention consists in using a highly polarized and highly concentrated source 3 in order to have a large dipolar coupling between the spins of the cores 7 and 9. The latter makes it possible, thanks to irradiations under Hartmann conditions. Hahn, to lower the spin temperature of the irradiated nuclei 7. Under these conditions it is obvious that this method can be applied to any type of nuclear spins. The major difference according to the isotope used will be the efficiency of the transfer and thus the increase of the signal obtained. Indeed, the dipole coupling δ (Equation 8) is proportional to the gyromagnetic ratio of the core 7 to which the transfer will be made (for example the proton). Consequently, this coupling is, for example, four times lower if the transfer is made to ¹³ C rather than to ¹ H. Similarly, the use of paramagnetic systems or quadrupole nuclei is possible.

Still according to the expression of δ, it appears clearly that at concentration c and polarization P in identical polarized nuclear system, the gyromagnetic ratio of this polarized isotope becomes it is essential to grow δ and thus to facilitate the obtaining of Hartmann-Hahn conditions, to increase the transfer speed, and finally to increase the signal gain. Thus, helium ³ He is of particular interest since its gyromagnetic ratio is 2.75 times greater than that of xenon "129".

Finally, note that in this process nothing imposes that the source of polarization 3 (for example xenon) and the liquid solution 1 (comprising for example the proton) are intimately mixed, since what induces the transfer is the coupling between the two spin systems caused by the strong magnetization of the source that creates the middle dipolar field, which itself is felt by the spins of the liquid solution 1. We can therefore consider a compartmentalization of the polarization source 3 and the liquid solution 1 to study.

Indeed, the transfer principle is to use the dipolar field created by the source 3 of polarized nuclear spins to create a coupling between the spin system of the source 3 and that of another species 7 (contained in the liquid solution ) and succeed in transferring the polarization. In practice, this dipolar field does not exist only within the volume containing the source of polarization 3 but extends outside the simple fact of the laws of magnetostatics. Therefore the polarized nuclear spin system 7 (xenon 129, helium 3 or others) can be in a compartment different from the sample containing the liquid solution 1 to be studied.

Indeed, FIGS. 2A and 2B show very schematically that the polarization source 3 can be physically separated from the liquid solution 1. Advantageously, the distance separating the polarization source 3 from the liquid solution 1 can correspond to a few nanometers and preferably less than ten millimeters. However, because of this separation, the dipole field δ will decrease, the factor ξ (Equation 8) tending to zero for the two samples as a function of the inverse of the distance separating them to the power three. So, it is advantageous in the current state of the art to optimize the geometry and not to exceed a separation distance of ten millimeters.

Thus, the polarization system according to the invention may comprise a container 20 comprising a first compartment 23 containing the source 3 of polarization and a second compartment 25 containing the liquid solution 1. For example, the first and second compartments 23, 25 can be two disjoint capillaries, a geometry that allowed, for example, the observation of multiple intermolecular coherences obtained by the effect of the dipolar field by WS Warren et al [see document 4].

For example, FIG. 2B shows that the first compartment 23 can be arranged around the second compartment 25. By way of example, a sample of the liquid solution 1 can be placed in a small central tube and the polarization source 3 placed in a second tube surrounding the first tube.

In the following, in order to expose the signal gain of a proton in a liquid solution, we consider a model composed of N xenon atoms I * and a proton S. Thus, at thermal equilibrium, the matrix density is given by:

^th Tr (exp - /? _L H) 2 ^{N + 1} 2 ^{N + 1} 1, Y ^LJ ^ ^s J

Tr (A) corresponds to the trace of the operator A, β _L = ft / kT is the inverse temperature of the network, and where a weak polarization has been admitted at this temperature T. H corresponds to the Hamiltonian Zeeman in the field B ₀ considered: n = ω _} Σl ^k _z + ω _s S ₂ (11) k

Under these conditions, the xenon signal after a 90 ° pulse along an axis Oy (which transforms σ, _h into σ _th ) is proportional to:

When xenon is polarized by optical pumping, the xenon spin density matrix is: exp - /? PoiHχ _e ^σ pol = - 2TrCeXp - ^ ₀ , H _Xe )

ClS)

We therefore deduce the signal of polarized xenon:

With σ ' _pol the density matrix after the 90 ° pulse following Oy. By noting K _Xe the measured signal gain factor, the spin inverse temperature of the polarized xenon can be deduced:

It will be noted that during the Hartmann-Hahn transfer and for a short time (typically at most a factor of 0.3) in front of the relaxation times along a spin-lock spin-lock field (Ti _p ), the energy of the spin system is constant. However, because of the dipolar coupling, the xenon and proton spin systems or baths will tend towards the same temperature in the double rotating reference system. Assuming that the irradiations are applied to the resonance

^Û > Ô - ^Û > I û> o - û> _s = 0, in this reference, PHamiltonien Zeeman

H = ₁ ¹ Y ¹ l * + û> fs is assumed to be large in the dipolar interaction, so that the contribution of the dipolar energy is negligible. Thus, the energy is given by the following formula:

(H) ≈ + fflfS _z ) l = Trf σ _/ (û> _i ^I ΣlJ + UJfS ₂ ) I (16) with σι and σ _f the density matrices at the beginning and at the end of the irradiation. Using equation (13) and noting /? _po i of the inverse temperature xenon spins defined with respect to the amplitude of the radiofrequency field ( ), we deduce the energy at the initial moment:

5 (H) = Tr, K ¹ Σi | _{+ ω} fS _z ) - | t _a nh ^ _- . (I ₇ )

The value of /? P _O ι is calculated below considering the conservation of energy:

At the end of the spin-lock, the calculation of the energy indicates from the model 10 of the spin temperature that:

(H) (19) where we used the Hartmann-Hahn condition ω] = ωf and where β ' _f is the inverse spin temperature in rotating fields in steady state. We deduce for the inverse spin temperature:

Consequently, in the laboratory reference system, when radiofrequency irradiations are stopped and thus the evolutions of the xenon and proton magnetizations are decorrelated: _{tanh (21)}

We thus deduce the gain K _H as the expected proton signal:

The previous calculation is for a proton and N xenons. By immediate extension in the case of a dilute proton system 7, N can be considered as the ratio between the xenon and proton concentrations. According to Equation (22), the transfer of polarization will be effective if the magnetization of xenon, through its number of atoms N (compared to the number of proton) and its polarization rate K _Xe , is very high. to that of the proton. The polarizations reached for xenon reach 70%, ie K _Xe ≈ 50000 for a magnetic field BQ of 11.7T so the expected signal gain for the protons could reach K _H ≈ 12000 when the ratio N between the number of xenon atoms and the number of proton atoms is large in front of 1. The Equation (22) immediately shows that under these same conditions, K _H becomes of the order of 6000 if we have as many xenons as protons and tends to 0 if there are more protons than xenons. The proposed solution is therefore particularly useful for low concentration samples. Obviously using helium "3" instead of xenon "129", one would benefit from a higher polarization rate (up to 80%) and especially from a higher gyromagnetic ratio, the gain in signal s' would find even more. Note, however, that for industrial realizations for which the polarized gas is dissolved in the liquid solution, the low solubility of helium compared to that of xenon impairs the obtaining of a very strong volume magnetization. This restriction disappears in the case where the transfer is made from polarized gas present in a separate compartment from that containing the liquid solution (see Figure 2A).

However, it is important to note that the previous calculation is a thermodynamic calculation which defines the rate of increase of the signal of a proton for a long time in front of the transfer time but short (factor at most 0.3) vis-à- screw relaxation times Tχ _p in the rotating reference frame. In fact, for a model without interaction, the time of Xenon relaxation in the rotating reference system must be of the order of the longitudinal relaxation time Ti of xenon and therefore of the order of several minutes. On the other hand, between the proton T _lp and the inverse of the dipolar coupling 1 / δ, it is impossible to decide as to their relative values. The maximum gain in proton signal will therefore be reached only if the polarization transfer rate characterized by the dipole coupling δ is large vis-à-vis the relaxation along the effective field 1 / TV Finally the rate of increase of signal K _{H /} observed depends on many factors or constraints, such as the quality of the Hartmann-Hahn condition tuning, the inhomogeneity of the radio frequency fields, the accuracy in the irradiation frequencies, and the type of continuous irradiation or composites used to let the dipolar coupling act.

From the point of view of the implementation of this approach, by its character of preparation of the proton polarization, it is general and can replace, advantageously as soon as K _H is greater than 1, the delay of return to thermodynamic equilibrium which precedes any sequence of pulses. In practice, it can be combined with all the approaches capable of increasing the polarization and therefore the signal of a diluted spin system. It is thus possible to combine a return to the thermodynamic equilibrium followed by a polarization transfer by the method described here, in order to also benefit from the thermal polarization. One can even after this transfer of polarization complete by a second transfer using the scalar coupling between nuclei of the same molecule to polarize them. The advantage of such an approach if a system ¹⁵ NH example we consider is the better transfer efficiency in the case of xenon to the proton _(N γ / γ _H = 0.1) resulting from a larger δ , and the good efficiency of INEPT transfers from proton to nitrogen 15. Finally, it should be noted that the use of composite sequences for the Hartmann-Hahn transfer opens up the possibility of improving the polarization of all the protons simultaneously. From an industrial implementation point of view, as the transfer of polarization on the one hand and relaxation on the other hand will induce a decrease in the polarization of xenon, it is to be expected to renew between each transfer a part of the source of polarization to reach a state of regime, thus opening the door to the use of this method combined with any liquid NMR sequences with one, two, three, or more dimensions.

In the current state of the art, such a transfer is possible only if the dipole coupling δ between the nuclei of the polarization source and those of the solution is greater than 0.1 Hz. However, it is preferable that it be greater than IHz.

Reference is now made to FIGS. 3 to 8 which illustrate an example according to the invention for increasing an NMR signal of a liquid solution with respect to that obtained by starting from the thermodynamic equilibrium by using strongly polarized xenon.

The experimental choices that have been made to obtain the results presented here, are based on the technical constraints of the equipment used, intended to bias a small amount of xenon 129. Indeed, according to the invention, it is it is preferable that the two radio frequency fields 13, 15 applied to the proton 7 and xenon 9 channels are identical to better than d so that the transfer can take place efficiently. As a result, the ability to implement this polarization transfer method depends on the importance of dipole coupling d. These constraints are particularly severe since, given the equipment used, the irradiations 13 and 15 are created by two separate coils and therefore under conditions in which the effects of the inhomogeneity of radio frequency fields were maximum. Moreover, a Hartmann-Hahn type transfer had never been described using a scalar coupling between two species of spins as weak [see document 3]. As In comparison, the scalar couplings through the connections ensuring in the liquid phase that type of transfer between different isotope spins are generally greater than 90 Hz. It is therefore essential to maximize the dipole coupling δ, that is to say the polarization of the nuclei of the source and their concentration. Indeed, the choice of the polarization source is fixed by the apparatus and one can not, by the choice of another nucleus, modify the gyromagnetic ratio γ \. The choices we made therefore consisted in maximizing the cP product by increasing the amount of xenon in the optical pumping cell (but not too much because it impairs P), using xenon enriched in isotope 129, minimizing the volume of the sample and using a solvent for which the solubility of xenon is very large.

In fact, in order to achieve the conditions required to enable Hartmann-Hahn transfer polarization transfer, isotope enriched xenon 129 (> 96%) is preferably used and conditions ensuring high production of polarized gas. By way of example, partial pressures greater than 40 torr of xenon can be used during optical pumping to maximize the amount of polarized xenon and thus the magnetization of xenon. FIG. 3 illustrates very schematically that, after optical pumping 31, the xenon 9 is condensed in a tube 33 immersed in a solenoid 35 cooled with liquid nitrogen. The xenon 9 contained in this tube 33 is then reheated in the leakage field of the superconducting magnet 36 of the NMR spectrometer. It is then condensed in a tube 37 containing the liquid solution 1 to be studied, this tube 37 having been degassed beforehand. To maximize the magnetization per unit volume which in fact defines the magnitude of the dipolar coupling between the two spin systems, it is preferable to use a shortened thick-wall NMR tube, for example a tube 37 having an outer diameter. of 5mm, a wall thickness of 1.4mm, and a total length of 150mm. Initially, the tube 37 contains a solution of 3,3-diethoxy-1-propyne dissolved in deuterated cyclohexane (99%) prepared at room temperature. Typically the pressure inside this tube 37 just after the addition of the polarization source 3 (that is to say xenon 9 gas) is of the order of eight atmospheres. The tube 37 is then strongly agitated in order to dissolve the xenon 9, and then put back inside the magnet 36 of the NMR spectrometer.

By way of example, the NMR spectrometer may be of the Bruker DRX500 type equipped with a Bruker broadband reverse probe. In this case, channel ¹ and channel H isotope xenon 129, xenon hereinafter, using two different coils 41 and 43 respectively. There is therefore no correlation in the inhomogeneities of radiofrequency fields created by these two coils 41 and 43. The first coil generates the first irradiation 13 on the proton cores 7 and the second coil generates the second irradiation 15 on the nuclei. xenon 9.

It will be noted that the presence of the strong magnetization per unit volume of the laser-polarized xenon 9 induces a frequency shift (Equation 8) of the proton 7 in this tube 37 (considered as an infinite cylinder) given by the following equation:

d = - ^ \ 000c! NPγ _Xe hr _{} i} = + 41,83c? } iz (23)

6

According to this example, therefore, an offset of 6.3 Hz is expected for a xenon concentration of lmol.L ^"1 and a polarization of 15%.

Accordingly, FIG. 4 shows a pulse sequence 50 using proton channel H and xenon channel 53. This sequence 50 is used to measure the dipolar coupling between xenon 9 and protons 7. Thus, a series of n proton spectra 57 is acquired. Each spectrum 57 is acquired after excitation of the magnetization of the protons by a pulse 56 on the channel 51 proton. After the acquisition of each spectrum a pulse 55 of angle θ is applied to the xenon channel 53 in order to reduce the magnetization of xenon by a factor cos θ.

After adding xenon 9, put the 37 NMR tube in the magnet 36 of the NMR spectrometer, optimized the spatial homogeneity of the magnet by the resonance of the deuterium, any blocking of field was suppressed so that the modification of the Xenon magnetization 9 is not compensated.

Then, a series of one-dimensional spectra on the proton channel according to the pulse sequence 50 is recorded by performing between each acquisition a pulse 55 of an angle θ (<90 °) on the xenon channel 53. It will be noted that a variation of the resonance frequency of each of the proton peaks (variation identical for all) is obtained along the series of spectra. In parallel, a variation of the intensity of the proton signal 57 is observed. It results from the joint action of a recovery time between each short proton acquisition and the variable SPINOE effect following the reduction of the xenon magnetization. Thanks to this effect we can know absolutely the sign of the polarization of xenon.

Indeed, Figure 5 illustrates the variation of the resonant frequency of the protons as a function of the number of pulses applied to the xenon. The determined dipolar coupling is equal to -3.05 ± 0.05, respectively

Hz for a negative xenon bias 61 and 4.32 ± 0.10 Hz for a positive bias 63.

From the curve adjusted in (δ cos "^{" 1} θ + B) with δ the initial frequency offset due to the dipole coupling, θ the xenon pulse angle, n the proton spectrum number and B the frequency limit value proton resonance, we can deduce an estimate of the product cP (see equation 23).

By comparing the signal of the laser-polarized xenon with that of xenon at thermodynamic equilibrium, it can be deduced that according to this example the polarization of the dissolved xenon is 9.6%. By way of example and to illustrate the existence of a dipole field created by a polarization source present in a compartment and felt by protons of another compartment, a sealed cylindrical capillary of 1.1 mm external diameter containing mainly deuterated benzene was introduced into the previous tube. This is the configuration of Figure 2B, with the capillary corresponding to the compartment 25. The outer compartment 23 contained a solution, whose solvent was perdeuterated cyclohexane and 0.7% chloroform (purity 99%, grade analysis of the company Merck). The impurities present in the trace state make it possible to have different fine lines with shorter ¹ H relaxation times. The same experimental procedure as before was used (addition of negatively polarized xenon at 6.4%, stirring, homogeneity adjustment, stopping of the B ₀ field blocking, and launching of the sequence 50). The map 70 of FIG. 6 corresponds to a subset of the two-dimensional spectrum obtained. The horizontal dimension corresponds to the dimension of the resonance frequency of the proton, the vertical dimension corresponds to the succession of spectra. This map is represented as a series of contour lines. Peaks 71, 73 and 74 clearly show distinct behavior along the vertical dimension, when compared to a constant frequency illustrated by the two dotted lines 75 and 76. The trace of the left peak 71 sees its frequency increased. during the experiment, while the trace of the right doublet 73 and 74 sees its frequency decrease. Similarly, the intensity of the peak 71 increases along the experiment since the decrease in the magnetization of the polarized xenon induces a lower SPINOE effect, hence an increase in the proton signal due to T _Xe <0. The determination of the frequency maximum of these peaks for each spectrum and the adjustment according to the preceding procedure leads to a dipole coupling δ of 5.4 ± 0.3 Hz for the peaks (peak type 71) corresponding to the compartment 23 in which the xenon is dissolved and of the order of -1.2 ± 0.5 Hz for the peaks (type peaks 73 and 74) of the internal compartment 25. A rapid magnetostatic calculation shows that the magnetic field created by the xenon polarized in the inner cylinder, (compartment 23, the one in which there is no xenon), must be of opposite sign to the magnetic field in the outer cylinder (compartment 25).

With a dipole coupling between xenons and protons of the order of 2 to 4 Hz and the use of two separate coils 41 and 43 to create the two irradiations 13 and 15, a Hartmann-Hahn coherence transfer can not be effectively realized. only if the amplitude difference of radio frequency fields ω ^ -ωf is typically less than this value.

As the radiofrequency fields are known to be inhomogeneous, and in the case where is used a conventional probe for which the two proton and xenon channels correspond to different coils 41 and 43, thus without correlation of radio frequency field inhomogeneities, it is not not conducive to using intense radio frequency fields for which the inhomogeneities would be much higher than the dipolar coupling. In this case, it is preferable to use the simplest Hartmann-Hahn transfer sequence with continuous low power irradiation.

Indeed, Figure 7 illustrates a pulse sequence used to observe the polarization transfer from xenon 9 to protons 7 by means of medium dipole coupling. The spin-lock 85 on the xenon channel 53 is comprised between two pulses 87 and 89 of 90 ° and of opposite phases (x and -x) intended to put the magnetization of the xenon 9 in the transverse plane initially and along the field static at the end of the mixing time T _m . The spin-lock 85 is applied simultaneously on the channel 51 of the proton and the channel 53 of the xenon with exactly the same amplitude of radiofrequency field ω ^ -ωf = 0. The proton signal 88 is detected after the spin-lock 85.

In addition, the simple 180 ° inversion of the irradiation phase in the middle of the irradiation time is added to partially compensate for the inhomogeneity effects of radiofrequency fields. In this example, it is preferable to use low radiofrequency field amplitudes so that the inhomogeneity is of the order of the measured dipolar coupling. The measurement of the amplitude of the radiofrequency xenon (respectively proton) field consists, in a known manner, in nutation of the xenon magnetization (respectively proton) followed by a purge gradient and a read pulse. To avoid any artefact during the measurement, the irradiation of increasing duration is applied simultaneously on the two channels 51 and 53 of the proton and xenon. Thus, one can determine the conditions of Hartmann-Hahn, = ωf, better than IHz near. For example, it is possible to use a radio frequency field whose inhomogeneity of field (width at half height) is of the order of 1.8 Hz, and whose field amplitude / 2π is 46.87Hz for the maximum probability.

Using the sequence of FIG. 7, with irradiation applied during, T _m = 800 ms, the proton spectra 91 and 93 shown in FIG. 8 are obtained with xenon polarizations of opposite sign.

Indeed, FIG. 8 shows a first spectrum 90 corresponding to the thermal signal of a proton, a second spectrum 91 obtained with a negative bias of xenon with respect to the thermodynamic polarization and after using the excitation method described above (see FIG. 7) and a third spectrum 93 obtained with a positive polarization of xenon with respect to the thermodynamic polarization and the use of the excitation method described above (see FIG. 7). The proton signals obtained 91 and 93 correspond respectively to 0.40 times and -1.85 times the thermal signal 90 of this same proton. Thanks to the magnetization of xenon measured before the acquisition of these spectra, it is possible to estimate a xenon-proton dipole coupling δ equal to 2.9 Hz and -2.1 Hz respectively.

Alternatively, it is conceivable to replace spin-lock 85 with a sequence of composite pulses, the irradiations being able to be provoked by means of probes with double or triple agreement on the same coil, that is to say that the coils 41 and 43 would be identical. A double ¹ H agreement, ¹²⁹ Xe would ensure a very good correlation of radio frequency fields since the same coil would create the field of excitation.

It will be appreciated that the implementation comprising two 90 ° xenon pulses which frame irradiations under Hartmann-Hahn conditions can easily be included in the recoupling sequence. The immediate extension of the adiabatic demagnetization methods in the rotating frame for which there is an adiabatic rotation of the xenon magnetization, followed by an adiabatic decrease in the amplitude of the radiofrequency field (in order to fulfill the conditions of Hartmann-Hahn thanks to another irradiation on the proton), then a return of the xenon magnetization following B ₀ by a new adiabatic rotation is quite feasible. Note finally that one can also benefit to increase the proton signal of the magnetization of the latter to thermodynamic equilibrium (Equation 3) via an impulse on the proton channel preceding the spin-lock 85. The phase of this impulse obviously depends on of the sign of the xenon polarization. References of cited documents:

1. R. R. Hartmann and E. L. Hahn, Nuclear double resonance in the rotating frame. Phys. Rev., (1962) 128 2042-2053. 2. M. G. Schwendinger, J. Quant, J. Schleucher, SJ. Glaser, and C. Griesinger, Broadband heteronuclear Hartmann-Hahn sequences. J. Magn. Reson. A, (1994) 111 115-120.

3. SJ. Glaze and JJ. Homonuclear and heteronuclear Hartmann-Hahn transfer in isotropic liquids, in Adv. Magn. Opt. Reson., W.S. Warren, Editor. (1996), Academy Press Inc .: San Diego, p. 59-252. 4. W. Richter, S. Lee, W. S. Warren, and Q. He, Imaging with intermolecular multiple-quantum coherence in solution nuclear magnetic resonance. Science, (1995) 267-654-657.

5. H. W. Long, H. C. Gaede, J. Shore, L. Reven, CR. Bowers, J. Kritzenberger, T. Pietrass, A. Pines, P. Tang, and J. A. Reimer, High-field cross polarization NMR from laser-polarized xenon to a polymer surface. J. Am. Chem. Soc., (1993) 115 8491-

8492.

6. J. Smith, LJ. Smith, K. Knagge, E. Mac Namara, and D. Raftery, Hyperpolarized xenon-mediated cross-polarization to material surfaces observed at room temperature and above. J. Am. Chem. Soc., (2001) 123 2927-2928.

Claims

A method for performing an NMR analysis of a liquid solution (1) comprising nuclei (7) with non-zero nuclear spins, characterized in that the spins of the cores (9) are irradiated with a source of polarization ( 3) and said spins of the cores (7) of said liquid solution (1), so as to perform a Hartmann-Hahn type polarization transfer of the spins of the nuclei (9) of said source (3) to the nuclei spins (7). ) of said liquid solution (1) by means of a coherent coupling resulting from the dipolar field created by the nuclei of the source, to obtain an NMR signal of said liquid solution (1) increased compared to its value at the thermodynamic equilibrium .

2. Method according to claim 1, characterized in that obtaining said dipolar field results from a concentration greater than 0.01 mol. L ^"1 and / or a polarization greater than 0.01 of the cores (9) of said polarization source (3).

3. Method according to claim 1 or 2, characterized in that one simultaneously subjects the spins of the cores (7) of said liquid solution

(1) and the spin of the cores (9) of said polarization source (3) respectively at a first and a second radiofrequency field irradiation (13 and 15) which are independently of continuous or composite type, for obtaining said polarization transfer.

4. Method according to claim 3, characterized in that said first and second irradiations respectively have substantially equal frequencies, within 200 kHz, at resonant frequencies of the spin of the cores (7) of said liquid solution (1) and spins. cores (9) of said polarization source (3).

5. Method according to claim 3 or 4, characterized in that the values of the spin precession frequencies of the cores (7 and 9) respectively resulting from said first and second irradiations (13, 15) have between them a difference which is less than ten times the amplitude, expressed in frequency unit, of said dipolar field.

6. Method according to claim 5, characterized in that the values of the spin precession frequencies of the cores (7 and 9) respectively resulting from said first and second irradiations (13, 15) have between them a difference which is less than amplitude, expressed in frequency unit, of said dipolar field.

7. Method according to one of claims 3 to 6, characterized in that one implements said transfer of polarization by composite or multi-pulse sequences of said irradiations, for which there is an effective magnetic field.

8. Method according to any one of the preceding claims, characterized in that said source of polarization (3) is obtained as desired:

by the addition of para-hydrogen on a carbon-carbon double bond,

by preliminary polarization of a sample by the low-temperature solid phase dynamic nuclear polarization method followed by dissolution of said sample in a liquid, or

by optical pumping at least one polarized rare gas which is selected from the group consisting of xenon, helium, neon and krypton isotopes.

9. Method according to claim 8, characterized in that said source of polarization (3) is obtained by optical pumping of at least one polarized rare gas which is selected from the group consisting of isotopes of xenon, helium, neon and krypton.

10. The method of claim 9, characterized in that said or at least one (s) said (s) rare gas (s) is xenon 129 or helium 3.

11. Method according to any one of the preceding claims, characterized in that said polarization source (3) is dissolved in said liquid solution (1).

12. Method according to one of claims 9 to 11, characterized in that said or at least one (s) said (s) rare gas (s) is the isotope 129 of xenon.

13. Method according to one of claims 1 to 12, characterized in that said polarization source (3) is physically separated from said liquid solution (1).

14. Method according to any one of the preceding claims, characterized in that said polarization source (3) is in the gas phase.

15. Method according to one of claims 13 or 14, characterized in that said source of polarization (3) and said liquid solution (1) are respectively contained in two adjacent compartments and independent of each other.

16. Method according to one of claims 13 to 15, characterized in that said liquid solution (1) is contained in an internal compartment which is surrounded by an outer compartment containing said polarization source (3).

17. Method according to one of claims 13 to 16, characterized in that said polarization source is separated from said liquid solution by a distance representative of the effective range of said dipolar field which is equal to or greater than 1 nanometer.

18. The method of claim 17, characterized in that said polarization source is separated from said liquid solution by a distance representative of the range of said dipolar field which is between 3 nanometers and 10 millimeters.

19. Method according to one of claims 13 to 18, characterized in that said or at least one (s) said (s) rare gas (s) is the isotope 3 of helium.

20. Method according to one of claims 1 to 19, characterized in that the cores (7) of said liquid solution (1) comprise hydrogen nuclei.

21. Method according to one of claims 1 to 19, characterized in that the cores (7) of said liquid solution (1) comprise carbon cores 13.

22. System for performing an NMR analysis of a liquid solution (1) comprising non-zero nuclear spin nuclei (7), characterized in that it comprises a magnet (36) surrounding: a source of polarization (3) , polarization transfer means (5) adapted to irradiate the core spins (9) of said polarization source (3) and said spins of the cores (7) of said liquid solution (1), so as to carry out a transfer of Hartmann-Hahn type polarization of the spins of the nuclei (9) of said source (3) at the spin of the nuclei (7) of said liquid solution (1) by means of a coherent coupling resulting from the dipolar field created by the nuclei of the source, to obtain an NMR signal of said liquid solution (1) increased with respect to its value at thermodynamic equilibrium.

23. System according to claim 22, characterized in that said polarization transfer means (5) comprise:

first irradiation means adapted to subject the spins of the cores of said liquid solution to a first irradiation of continuous or composite radio frequency fields, and

a second irradiation means (45) adapted to subject the spins of the cores (9) of said polarization source (3) to a second irradiation (15) of radio frequency fields of the continuous or composite type.

24. The system of claim 22 or 23, characterized in that obtaining said dipolar field results from a concentration greater than 0.01 mol. L ^"1 and / or a polarization greater than 0.01 of the cores (9) of said polarization source (3).

25. System according to one of claims 22 to 24, characterized in that said source of polarization (3) is obtained as desired:

by the addition of para-hydrogen on a carbon-carbon double bond,

by preliminary polarization of a sample by the low-temperature solid phase dynamic nuclear polarization method followed by dissolution of said sample in a liquid, or by optical pumping at least one polarized rare gas which is selected from the group consisting of xenon, helium, neon and krypton isotopes.

26. System according to claim 25, characterized in that said source of polarization (3) is obtained by optical pumping of at least one polarized rare gas which is selected from the group consisting of isotopes of xenon, helium, neon and krypton.

27. System according to one of claims 22 to 26, characterized in that it comprises, inside said magnet (36), a tube (37) containing said source of polarization (3) dissolved in said liquid solution (1) .

28. System according to one of claims 22 to 26, characterized in that it comprises, inside said magnet (36), a container (20) comprising two compartments (23 and 25) adjacent and independent one of the other, which are respectively intended to contain said polarization source (3) and said liquid solution (1).

29. System according to claim 28, characterized in that the polarization source (3) is in the gas phase.

30. System according to one of claims 22 to 26, characterized in that it comprises, inside said magnet (36), a container (20) comprising two inner and outer compartments (23 and 25), said outer compartment ( 23) being intended to contain said polarization source (3) and surrounding said internal compartment which is intended to contain said liquid solution (1).

31. System according to one of claims 28 to 30, characterized in that said two compartments are separated by a distance representative of the range of said dipolar field which is between 3 nanometers and 10 millimeters.