Apparatus and method for imaging metallic or partially metallic bodies by magnetic resonance; application of this method to the imaging of electrochemical cells

Abstract

A magnetic resonance imaging (MRI) apparatus (1), arranged to perform imaging of a body (a) containing at least one metallic part, comprising means (b'') for generating a polarizing magnetic field intended to be applied to said body (a), radio frequency (RF) means (g) for exciting said body (a), detection means cooperating with said antenna means to deliver a magnetic resonance imaging (MRI) signal, and means for processing said MRI signal so as to deliver information characteristic of the state of said body (a). The body (a) is subjected to a very weak field of less than 10 mT, and the detection means comprise a pickup coil (h) magnetically coupled with the polarizing means (b'') and the radio frequency means (g) and operating as a flux concentrator, and a SQUID (superconducting quantum interference device) detector disposed downstream of said pickup coil (h) via a transformer.See Figure 3.

Description

Title of the invention: Apparatus and method for imaging metallic or partially metallic bodies by magnetic resonance, application of this method to the imaging of electrochemical cells Field of the invention

[0001] The present document describes an apparatus for imaging metallic or partially metallic bodies by magnetic resonance. It also relates to a method for imaging metallic or partially metallic bodies implemented in this apparatus, and its application to the imaging of electrochemical cells, in particular Li-ion cells.

[0002] Current state of the art

[0003] Magnetic resonance imaging (MRI) is based on the principle of nuclear magnetic resonance (NMR). This phenomenon exploits the resonance between two energy levels that occurs when a quantum magnetic moment (a spin) is subjected to an external magnetic field. This spin can be that of an electron or an atomic nucleus. In the case of medical NMR, the nucleus of interest is in most cases hydrogen (which has only one proton) present in water, fat or tissues of the human body.

[0004] The NMR (or MRI) experiment is done in a few precise steps. First, the sample (in imaging a Li-ion cell or any metallic component, this would be the component itself), is placed in a very homogeneous static magnetic field Bo called the bias field, produced by a large superconducting coil. In most commercial devices, this magnetic field varies from 1.5 T to more than 10 T. When the protons in the sample are placed in a magnetic field Bo, they precede around this field with a frequency cüo given by û)q = yB 0, where y is called the gyromagnetic ratio.

[0005] Then, the protons are excited using a radiofrequency signal of the same frequency co = ct>o as the proton precession. This excitation causes the protons to come out of their precessional motion, after a while they slowly decay back to their original precessional state while emitting a signal that is detected using an antenna. Measuring this signal is a way to gather information about the composition of the body's interior, and ultimately to reconstruct an MRI image.

[0006] The overall principle of the MRI experiment can be summarized as follows: The body part to be imaged is placed in a static homogeneous magnetic field Bo- - When placed in this field, the protons precede with a frequency cu0 = yBo- - Using an excitation antenna, we send a pulse of intensity B j and frequency eu = cu0 tuned to the precession of the protons. - The protons are sent to a higher energy state and decay back to their original state while sending a signal of intensity B, and frequency eu o. More importantly, this signal will decay with the typical times Ti, corresponding to a recovery of the longitudinal magnetization in the direction of B 0, and r2, corresponding to a loss of spin-spin coherence in the sample. These typical time constants are characteristics of the precise state at each point in the sample and are used to produce contrast in an MRI image. In the case of medical MRI, they provide information about the type of tissue at a particular point (water, fat, muscle, etc.).

[0007] To construct an MRI image, it is necessary to determine where the signal detected by the MRI antenna comes from. For this, gradient coils are used. In addition to the permanent homogeneous field Bo, a small additional field <5 B(x, y, z) is added, which varies linearly along the three directions of space.

[0008] Typically, in an MRI device implementing a magnetic field of 5 T, an additional field ôB of the order of 100 mT.m ' is observed. As the resonance frequency of the nuclei is proportional to the local magnetic field, the gradient fields make it possible to spatially encode the resonance frequency and the phase of the signal, thus making it possible to form 3D images.

[0009] During an MRI sequence, these gradient fields are rapidly changed on time scales of a few milliseconds to selectively excite certain parts of the sample. The pattern that is chosen to apply the gradients and excitation pulses constitutes an MRI sequence.

[0010] Different sequences allow different contrasts to be displayed and different things to be seen. In medical MRI, there is a wide variety of sequences: some are suitable for seeing blood vessels (time-of-flight sequences), others are more suitable for functional imaging (BOLD for "Blood Oxygenation Level Dependent"), while more common T-weighted or T2-weighted sequences can discriminate between fat, organic tissue, tumors, etc.

[0011] It should be noted that MRI sequences function as image constructs in reciprocal Fourier space (or k-space) where different MRI sequences respond to different acquisition trajectories.

[0012] We will now explain why it is necessary to implement low-field MRI for imaging metal samples.

[0013] For over forty years of existence, innovation in MRI has consistently pushed toward higher fields. A higher field means that more nuclei in the sample are polarized, resulting in an increased signal-to-noise ratio. Furthermore, because the detection frequency is proportional to the magnetic field, it allows working at high frequencies, in a range where typical inductive detection antennas are inherently more sensitive. In short, a high field means better resolution and faster acquisition times. However, this strategy has obvious drawbacks. Commercial high-field machines require heavy and expensive superconducting magnets that are cooled to 4 Kelvin with liquid helium, and require expensive maintenance. A typical 7.5 T scanner costs nearly $1 million upfront plus a few hundred thousand dollars of maintenance per year.The MRI machine must be installed in a dedicated, magnetically shielded room with a reinforced floor. This makes MRI technology a luxury imaging procedure compared to ultrasound or CT scans.

[0014] A less obvious problem with high-field MRI is the inability to image in the presence of metal, let alone image a metal sample. This problem is actually twofold: - having a sample with metallic and non-metallic parts means that there will be huge local differences in magnetic susceptibilities. This causes magnetic field gradient artifacts that destroy the linearity of the gradient field diagram used to perform MRI. - High frequency magnetic fields cannot penetrate into most of a metal sample due to the skin effect of electrical conductors.

[0015] When a propagating electromagnetic field reaches a conductor / dielectric interface, eddy currents are created inside the inductor. These currents in turn generate another magnetic field that offsets the incipient magnetic field. This causes the total magnetic field to effectively die out upon entering the bulk of the conductive sample. Skin depth <5 is a measure of how deep the magnetic field can penetrate inside the metallic sample:

[0016] where a) is the pulsation, q is the DC resistivity, is the vacuum permeability and r is the relative permeability of the sample.

[0017] In the case of proton MRI, we have H / (2 ji): 42 MHz.T while for lithium this number is a little lower at yL / (2 :16.5 MHz.T In all cases, This means that in traditional high-field MRI, the excitation frequencies used are in the order of 30 MHz or higher. For this reason, the skin depth in most metals is usually less than 100 ^m, making MRI of metallic samples impossible. The table below compiles the typical resistivities and skin depths associated with typical metals used in the composition or packaging of Li-ion cells. Material Skin depth vs. frequency (mm) 5 kHz 50 kHz 50 MHz Copper 0.93 0.30 0.01 Stainless steel 6.3 2.0 0.06 Lithium 2.2 0.69 0.02

[0018] Traditional high-field machines operate with fields ranging from 1 T to over 10 T. If lithium is considered as a reference, this covers frequencies from 25 MHz to over 170 MHz. In the table above, we observe that this corresponds to skin depths of a few tenths of a micron, which therefore makes high-field MRI impossible for metallic samples.

[0019] It is therefore necessary to lower the magnetic field to work at lower frequencies. This would also mean lighter, inexpensive machines that do not require heavy maintenance or magnetic shielding, and that would be immune to susceptibility artifacts. However, if we simply take a traditional machine and lower the field, we will have a much lower signal-to-noise ratio or SNR (for "Signal Noise Ratio"), therefore degraded image quality and long acquisition times.

[0020] One possible solution is to use SQUIDs (superconducting quantum interference devices), which are ultrasensitive magnetometers that allow the magnetic field of the NMR experiment to be lowered while maintaining sufficient SNR to produce images with good resolution and contrast in a reasonable time. With this technology, an MRI image can be made with fields ranging from 100 pT to a few mT. Taking lithium as a reference, this means NMR frequencies ranging from 1.6 kHz to about 20 kHz. Looking at the table, this means skin depths of a few mm to 1 cm depending on the metal, which allows imaging of metal samples.

[0021] On this subject, we can cite the work of Alexej Jerschow's team at New York University. To circumvent the problems related to high frequencies and the skin effect, this team designed a technique called reverse MRI (ioMRI for "IntraOperative Magnetic Resonance Imaging") [6] [7]. The technique is well detailed in the very recent book by Haber-Pohlmeier, Bliimich and Ciobanu

[12] . Rather than making a direct image of a Li-ion cell, the idea is to make an indirect measurement of the local magnetic susceptibility of the anode and the cathode using an MRI of the cell's surrounding environment. More precisely, the cell is placed in a support which contains a measuring medium, generally doped water, surrounding the Li-ion cell. Then, a permanent Bo field is applied to the cell.The cell in turn generates a secondary field Bs proportional to the local susceptibility of the cell components, this secondary field being mapped with a special MRI sequence of the surrounding doped water. Reference can be made to Illot et al. 2018 [7], [Fig.l] for a schematic of the arrangement of this device.

[0022] With an accurate inoperable measurement of this secondary field, it is possible to deduce the overall magnetic susceptibility of the anode and cathode. In turn, this overall susceptibility is shown in [Fig.2] of Ilott et al. 2018 [7] as being related to the state of charge SOC (for “State Of Charge”). This figure shows the measured overall magnetic susceptibility as a function of the state of charge.

[0023] The sensitivity of the anode or cathode (in ppm) increases by about 50% from a fully discharged state to a 250 mAh charge. The magnetic field map (in ppm) is given with reference to the fully charged cell. The cell used is produced by the Rochester Institute of Technology (RIT). Figure 4 of document [7] also demonstrates that magnetic field maps can be used to classify cells into different defect categories: folded, small craps, for example. In this figure, the magnetic field maps are given in ppm relative to the field value of one of the non-defective cells. Also shown are the mean and standard deviation of the field map taken over the entire photo.The MRI sequence used is initially a simple FLASH sequence, then specific point imaging with Ti enhancement was implemented to treat magnetic susceptibility artifacts.

[0024] While ioMRI demonstrates that MRI can indeed be used effectively for the diagnosis of SOC and SOH (State of Health), it suffers from a major drawback: it only shows an overall map of the state of the anode or cathode, but cannot produce a spatially resolved map of the local SOC and SOH of each electrode.

[0025] Furthermore, until now, SQUID MRI has not seen a breakthrough due to the lack of available SNR at low field. To increase the SNR, most attempts have used a technique called prepolarization: first, a high pulsed field of about 100 mT is applied to increase the Boltzmann polarization of the sample, which increases linearly with the field. Then, the field is rapidly lowered and the MRI acquisition is performed in a field of interest, usually around 100 ^T.

[0026] This allows taking advantage of both the moderately high bias at 100 mT and the advantages associated with the ultra-low field, mainly improved contrasts and reduced field homogeneity constraints. However, pre-biasing requires heavy machinery for generating pulse fields and complex shielding to counteract noise due to eddy currents. More importantly, the rapid switching of magnetic fields would make this technique unviable on metallic samples.

[0027] The aim of the present invention is precisely to remedy this drawback by proposing an MRI imaging device for a metallic component which can provide a spatially resolved map of it. Statement of the invention

[0028] This objective is achieved with a magnetic resonance imaging (MRI) device, arranged to perform imaging of a body containing at least one metal part, comprising: - means for generating a polarizing magnetic field intended to be applied to said body, - RF radiofrequency means for exciting said body, - detection means cooperating with said radiofrequency means, to deliver a magnetic resonance imaging (MRI) signal, - means for processing said MRI signal so as to deliver information characteristic of the state of said body.

[0029] According to the invention, the body is subjected to a very weak field of less than 10 mT, and the detection means comprise a pickup coil magnetically coupled with the polarization means and the radiofrequency means and operating as a flux concentrator, and a SQUID detector (superconducting quantum interference device) arranged downstream of said pickup coil via a transformer.

[0030] In the following, body means any object, device, component, structure or all or part of a human or animal body.

[0031] In a preferred version of the invention, the polarization means, the radiofrequency means and the pickup coil are enclosed in a shielded chamber.

[0032] The SQUID detector and at least part of the processing means are advantageously arranged in a cryostat.

[0033] The polarization means may comprise a gradient coil, and the pickup coil may have a volumetric gradiometric geometry.

[0034] The pickup coil may have a surface geometry, in particular a second-order surface gradiometric geometry.

[0035] In a first example of application of the invention, the capture coil is arranged to receive an electrochemical cell, in particular a Li-ion cell.

[0036] In a second example of application of the invention, the pickup coil is arranged to receive an electronic chip or an electronic component.

[0037] In a third example of application of the invention, the capture coil is arranged to receive an industrial mechanical structure.

[0038] In a fourth example of application of the invention, the MRI imaging apparatus is adapted to image a structure of a power plant.

[0039] According to another aspect of the invention, there is provided a method for imaging by magnetic resonance a body containing at least one metal part, comprising the following steps: - generate a polarizing magnetic field intended to be applied to said body (a), - exciting said body (a) by radiofrequency (RF) waves, - detect an MRI signal of the response of said body to radiofrequency excitation, - process the said MRI signal thus detected, so as to deliver information characteristics of the state of said body (a),

[0040] According to the invention, the body is subjected to a very weak field of less than 10 mT, and the detection step comprises a generation of an induced signal captured by concentration of flux in a capture coil, and an application of this signal thus captured to a SQUID detector.

[0041] The MRI imaging method according to the invention can be advantageously implemented for the characterization of an electrochemical cell, and arranged to provide a map of the electrochemical cell, representative of the state of charge (SOC) of this electrochemical cell and / or representative of its state of health (SOH).

[0042] The MRI imaging method according to the invention can be arranged to provide a map of the electrochemical cell, representative of the state of health (SOH) of this electrochemical cell.

[0043] Thus, the ultra-low field magnetic resonance imaging device of metallic components according to the invention, the system uses detection based on a low-temperature superconducting interference device (SQUID). Using SQUID detection, we work on fields ranging from the 50 (earth field) range to a few mT. By using ultra-low fields, which correspond to very low NMR frequencies, we are able to perform imaging of metallic samples, in particular electrochemical cells.

[0044] The MRI imaging apparatus according to the invention is designed for the spatially resolved diagnosis of the state of metallic components, and in particular of the state of charge (SOC) and state of health (SOH) of electrochemical cells.

[0045] The objective is to develop inexpensive and portable MRI devices that do not require any precautions to operate and that can image metal samples. To achieve this, the operating magnetic field is lowered from the 1.5 T used in traditional medical machines to less than 1 mT. Since a resistive magnet is used, the working field can easily be adjusted between 100 and a few mT.

[0046] The amount of available signal decreases linearly with the magnetic field Bo , which means a factor of at least 1000 in our case, compared to high-field MRI. To counter the signal loss, an ultrasensitive antenna is used that relies on SQUID detection. SQUIDs are very sensitive magnetometers made from a loop of superconducting material intercepted by two Josephson junctions. They have a very wide bandwidth and can detect signals from DC to 100 MHz with a flat frequency response. Superconductors must be cooled to cryogenic temperatures 4 K to operate, which is a problem for the heavy superconducting magnets used in traditional MRI. In the case of SQUIDs, this is much less restrictive, as they are mounted on very small chips of a few centimeters and therefore require only very light cryogenic machines to operate.

[0047] Since SQUIDs are very small, typically loops of a few in diameter, they are often used in conjunction with an antenna that functions as a flux concentrator. Such detectors can achieve field sensitivities of order j for fields in the kHz range. REFERENCES

[0048] 1. Superconducting quantum interference device instruments and applications. RL Fagaly et al., Review of Scientific Instruments 77 101101 (2006). 2. Electrochemical Impedance Spectroscopy (EIS): Principles, Construction, and Biosens- ing Applications. HS Magar et al., Sensors (Basel) 19 6578 (2021). 3. SQUID-Based Magnetic Resonance Imaging at Ultra-Low Field Using the Backprojection Method. Q. Guo et al., Concepts in Magnetic Resonance part. B, Magnetic resonance engineering 8882329 (2020). 4. Progress toward a déployable SQUID-based ultra-low field MRI System for anatom- ical imaging. M. A. Espy et ail., IEEE transactions on applied superconductivity, 25 no. 3, 1601705 (2015). 5. MRI of the human brain at 130 microtesla. B. Inglis et ail., PNAS 110 no. 48, 19194-19201 (2013). 6. Real-time 3D imaging of microstructure growth in battery cells using indirect MRI.

[0049] A.J. Ilott et ail., PNAS 113 (39) 10779-10784 (2016). 1. Rechargeable lithium-ion cell State of charge and defect détection by in-situ inside- out magnetic résonance imaging. A. J. Ilott et ail., Nature Communications 9 1776 (2018). 2. Distortion-free inside-out imaging for rapid diagnostics of rechargeable Li-ion cells.

[0050] K. Romanenko et ail., PNAS 116 (38) 18783-18789 (2019). 1. Diagnosing current distributions in batteries with magnetic résonance imaging. M. Mohammadi et ail., J. Mag. Res. 309 106601 (2019). 2. Sensitive magnetometry reveals inhomogeneities in charge storage and weak tran- sient internai currents in Li-ion cells. Y. Hu et ail., PNAS 117 (20) 10667-10672 (2020). 3. Observation of memory effects associated with dégradation of rechargeable lithium- ion cells using ultrafast surface-scan magnetic résonance imaging. K. Romanenko et ail., J. Mater. Chem. A 9 21078-21084 (2021). 4. S. Haber-Pohlmeier et ail., Magnetic Résonance Microscopy : Instrumentation and applications in Engineering, Life Science, and Energy Research, chapters 17-18, Wiley-VCH, 2022. ISBN : 9783527827251. 5. Three-dimensional characterization of electrodeposited lithium micro structure s us- ing synchrotron X-ray phase contrast imaging. D. S. Eastwood et ail., Chem. Co- munn. 51 266-268 (2015). 6. Three-dimensional high resolution X-ray imaging and quantification of lithium ion battery mesocarbon microbead anodes. F. Tariq et ail., J. of Power Sources 248 1014-1020 (2014). 7. Characterization of the 3-dimensional microstructure of a graphite négative elec- trode from a Li-ion battery. P. R. Shearing et ail., Eletrochem. Comunn. 12-3 374-377 (2010). 8. Multi Length Scale Micro structural Investigations of a Commercially Available Li- Ion Battery Electrode. P. R. Shearing et ail., J. Electrochem. Soc. 159-7, 1023-1027 (2012). 9. Local Tortuosity Inhomogeneities in a Lithium Battery Composite Electrode. D. Kehrwaldet ail., J. Electrochem. Soc. 158-12 1393 (2011). 10. In-operando high-speed tomography of lithium-ion batteries during thermal run- away. D. P. Finegan et ail., Nat. Comm. 6 6924 (2015). 11. Non-invasive battery analysis via micro-computed tomography. E. L. Ballard et ail., US7902518B2 (2008). 12. In-situ battery diagnosis method using electrochemical impédance spectroscopy. R. Mingant et ail., US8849598B2 (2010). 13. Method for determining an aging condition of a battery cell by means of impédance spectroscopy. J. Ziegler et ail., US20120019253A1 (2009). 14. Methods and apparatus for battery testing. J. A. Tinnemeyer et ail., US20110074432A1 (2009). 15. Electrochemical impédance spectroscopy in battery management System. CN107076801A (2015). 16. A kind of EIS method for fast measuring of lithium ion battery. CN106970266A (2016). 17. https: / / www.ieco.fi / index.php?k=10909 18. https: / / www.shicryogenics.com / product / rp-082B2S-4K-pulse-tub e- cryocooler- sériés / 19. https: / / www.pure-devices.com / index.php / products / products-research / products-search readerl.html Description of figures

[0051] - [Fig.l] is a schematic view of a SQUID detection system, implemented implemented in an MRI imaging device according to the invention; - [Fig.2] is an example of the geometry of a flow concentrator used in a SQUID detection system implemented in an MRI imaging device according to the invention; - [Fig.3] is a schematic view of a metal component MRI imaging apparatus according to the invention. Detailed description

[0052] A SQUID detection system 10 comprises, with reference to [Fig.l], a flux concentrator 2 comprising an excitation coil c and a capture coil b surrounding a metallic component a such as a Li-ion electrochemical cell, a shielded transformer 3 connected at the output to an input inductance L;, of a SQUID 4 whose output signal is amplified by a 5 LNA amplifier (for “Low Noise Amplifier”) coupled to a 6 FLL phase-locked loop (for “Flux Locking Loop”) controlling an Lf power supply coil.

[0053] The SQUID can be made of niobium, magnesium diboride (MgB) or any other superconducting material with a medium critical temperature, or based on high temperature copper oxide.

[0054] The metallic component emits a signal which is captured by the pickup coil b which acts as a flux concentrator, and sent via the two terminals 20,21 - on which an impedance matching capacitor Ca is connected in parallel -, to the SQUID 4 via the input coil L The shielded transformer L b L2 is used for impedance matching and ground isolation.

[0055] The excitation coil c is connected to an MRI console d. The signal output of the SQUID is amplified via the low noise amplifier LNA 5, while the operating stability is ensured by the flux lock loop FLL 6.

[0056] In this example the pickup coil is a saddle coil but gradiometric antennas are often preferred, like the one shown in [Fig.2].

[0057] In most SQUID MRI experiments, the pickup coil is a second-order surface gradiometer, as described in Fagaly et al. [1].

[0058] In the present invention, it is preferred to use volumetric geometries, which make it possible to capture more signal while remaining robust to noise. To maintain a gradiometric configuration in order to remain robust to noise emitted by distant sources, the antenna is made up of two saddle coils wired in series with opposite currents. This particular geometry was the subject of a patent application filed on April 12, 2022 in the name of the same Applicant.

[0059] The pickup coil 2' comprises a first saddle-shaped coil c' surrounding a second saddle-shaped coil b'. The first coil c' comprises a first longitudinal conductor cl, a first end conductor c2, a second longitudinal conductor c3, a second end conductor c4, a third longitudinal conductor c5, a third end conductor c6 and a fourth longitudinal conductor c7 connected to a first output conductor 21 to the transformer 3.

[0060] The second coil b', concentric with the first coil c', comprises a first longitudinal conductor b1, a first end conductor b2, a second longitudinal conductor b3, a second end conductor b4, a third longitudinal conductor b5, a third end conductor b6, a fourth longitudinal conductor b7 connected to a second output conductor 20 towards the transformer 3.

[0061] We will now describe, with reference to [Fig. 3], an MRI imaging device 1 of a metallic component, such as a Li-ion electrochemical cell.

[0062] The MRI imaging apparatus 1 according to the invention comprises a flux concentrator 2'' arranged to receive a metallic component a, for example a chemical cell. The coil b produces a magnetic field B 0 which polarizes the nuclei of interest in the metallic component a, creating energy levels separated by an energy hcu0-

[0063] The 2” flux concentrator also includes a gradient coil h used to produce the spatial resolution in the MRI image. This coil h is powered via a pair of conductors 8c by a stabilized current source c which makes it possible to deliver a very regular current of the order, depending on the precise geometry of the coil, of a few A to a few tens of A.

[0064] The gradient coil b” is powered by a gradient amplifier d, for example the XPA-175-350 model from IECO in Finland

[25] . To keep the field very stable and homogeneous in the field of view, the field is continuously monitored with a fluxgate magnetometer probe e. Low-frequency fluctuations in the field are measured and a command is sent in real time to the current source, to keep the field variations within a typical drift of 50 ppm for the duration of the MRI acquisition.

[0065] The radiofrequency amplifier f powers the radiofrequency coil g, which may for example be of birdcage geometry, to send a pulse B1 tuned to the magnetic resonance frequency cu0- The signal is then picked up by the detection system described in [Fig.2].

[0066] The pickup coil h sends the signal to the SQUID and the rest of the readout chain i inside a 4K cryostat j, cooled by a cryogenic cooler k, for example the Sumitomo model RP-082B2S

[26] .

[0067] The output signal is then processed at room temperature by an analog-to-digital conversion module and an MRI console 1, for example the drive-1 model from Pure Devices

[27] which interprets the time signal as an MRI image displayed on the screen m. The system of bias, excitation and detection coils and the metallic component a are all enclosed in a shielded chamber n composed of a metallic grid acting as a Faraday cage, protecting signals in the 10 kHz and above, and a layer of magnetic materials such as ferrites or mu-metal which protects the system from very low frequency noise below a few kHz.

[0068] The present invention implements an antenna which makes it possible to increase the sensitivity even further, approaching 0.1 VHz. This antenna has a geometry with both gradiometric and volumetric which gathers the most magnetic flux of the sample while remaining very robust to noise. Document FR3117218 in the name of the present applicant describes the use of this antenna in the context of NMR and MRI. Thanks to this antenna and the increase in the measurement field from the usual 100 / / T to approximately 1 mT, it is possible to increase the SNR sufficiently to perform an MRI acquisition without prepolarization and sufficient resolution and contrast for in-operando imaging of Li-ion cells.

[0069] The ultra-low field MRI method according to the invention is particularly suitable for producing a direct MRI image of a single Li-ion cell or a cell assembly. Compared to the ioMRI technique described previously, the direct MRI method makes it possible to have an image of the interior of the cell, with contrasts showing temporally resolved maps of the local SOC and SOH. Here are some possible strategies: - Li MRI sequences, to map the local lithiation state of the cathode, providing spatially resolved insights into the local SOC state of charge; [7] - 3C MRI sequences to study the SOC in the anode; [1] - susceptibility-weighted MRI sequences could provide access to a spatially resolved map of magnetic susceptibility, which, as shown by Illot et al. [5], is related to local SOC and SOH.

[0070] Approximate quantitative estimates of the important parameters of the device are given below: - Acquisition time around 5 minutes for a fully resolved image, but it is possible to perform rapid sequences with less information in a few seconds. - Spatial resolution in the millimeter range, most likely an in-plane resolution in the order of 1.0 to 2.5 mm. - Volume of interest (field of view) up to 50 cm, possibly more. This allows imaging of a single cell or an assembly of cells with a single acquisition. - Portable device on wheels, the size of a small washing machine and weighing about 100 kg.

[0071] Possible applications - Evaluation of the state of charge SOC and the state of health SOH of electrochemical cells or full batteries, consisting of cell assemblies, by providing a spatially resolved map of the SOC and SOH in operando. - Quality control of electrochemical cells. - Quality control and monitoring of chips and electronic components. - Detection of defects in metal structures in the following industries: automotive, construction, energy, defense and aerospace, transport infrastructure. - Medical MRI of patients with metal implants.

[0072] Of course, the invention is not limited to the examples which have just been described. and many other embodiments can be envisaged without departing from the scope of the invention.

Claims

1.

2. Claims Apparatus for imaging a metallic or partially metallic body by magnetic resonance (MRI) (1), comprising: - means (b”) for generating a polarizing magnetic field intended to be applied to said body (a), - RF radiofrequency means (g) for exciting said body (a), - detection means (10) cooperating with said radiofrequency means (g), to deliver a magnetic resonance imaging (MRI) signal, said detection means comprising a pickup coil (h) acting as an antenna and magnetically coupled with the polarization means (b”) and the radiofrequency means (g), and a SQUID (superconducting quantum interference device) detector (4) arranged in a cryostat, downstream of said pickup coil (h), - means (1) for processing said MRI signal so as to deliver information characteristic of the state of said body (a), said body (a) being subjected to a very weak excitation field of less than 10 mT and at a frequency providing for the metal constituting said body a skin thickness of a few mm to 1 cm, characterized in that: - the excitation field is between 100 ^T and a few mT, - the pickup coil (h) has a volumetric gradiometric geometry in which the antenna consists of two saddle coils wired in series with opposite currents, and acts as a flux concentrator, the polarization means (b”), the radiofrequency means (g) and the pickup coil (h) are enclosed in a shielded chamber (n), said pickup coil (h) being connected via terminals (20,21) and a transformer (3) to said SQUID detector. MRI imaging device (1) according to the preceding claim, characterized in that at least part of the processing means (i) is arranged in the cryostat (j).

3. MRI imaging apparatus (1) according to any one of the preceding claims, characterized in that the polarization means (b”) comprise a gradient coil.

4. MRI imaging apparatus according to any one of the preceding claims, characterized in that the pickup coil (h) is arranged to surround a Li-ion electrochemical cell (a), and in that it is arranged to map the local lithiation state of the cathode of said cell

5. An MRI imaging apparatus according to any preceding claim, characterized in that the pickup coil is arranged to surround an electronic chip or electronic component.

6. An MRI imaging apparatus according to any preceding claim, characterized in that the pickup coil is arranged to surround an industrial mechanical structure.

7. MRI imaging apparatus according to any one of the preceding claims, characterized in that it is adapted to image a structure of a power plant.

8. Method for imaging by magnetic resonance a metallic or partially metallic body (a), comprising the following steps: - generating a polarization magnetic field intended to be applied to said body (a), - exciting said body (a) by radiofrequency (RF) waves, - detecting an MRI signal of response of said body to the radiofrequency excitation, comprising a capture of a flux applied to said body in magnetic coupling with the polarization magnetic field and the radiofrequency excitation field, and a SQUID (superconducting quantum interference device) detection (4), - processing said MRI signal thus detected, so as to deliver information characteristic of the state of said body (a), said body being subjected to a very weak field of less than 10 mT and at a frequency providing for the metal constituting said body a skin thickness of a few mm to 1 cm,characterized in that: - the detection step implements a capture coil (h) having a volumetric gradiometric geometry acting as a flow concentrator,

9.

10.

11. - the polarization (b”), radiofrequency (g) and capture (h) steps being carried out in a shielded chamber (n). MRI imaging method according to the preceding claim, characterized in that the metallic or partially metallic component to be imaged is a Li-Ion electrochemical cell (a) and in that it is designed to map the lithiation state of the cathode of said cell. MRI imaging method according to the preceding claim, characterized in that it is arranged to provide a spatially resolved in-operand mapping of the Li-Ion electrochemical cell (a), representative of the state of charge (SOC) of said electrochemical cell (a). MRI imaging method according to one of the two preceding claims, characterized in that it is arranged to provide a spatially resolved in-operand mapping of the electrochemical cell (a), representative of the state of health (SOH) of said electrochemical cell (a).

Images