Recently a new, non-optical polarisation method has been discovered in ³He [1]. It appears as a very attractive alternative to established laser-based hyperpolarisation schemes (spin exchange optical pumping, SEOP, or metastability exchange optical pumping, MEOP). For instance, application to ³He nuclear magnetic resonance (NMR) magnetometry with a relative precision of 10-12 has been demonstrated.

A very simple experimental set-up has been used for demonstration in [1]: a sealed glass cell containing the gas sample; a small wire coil, attached to it, for discharge excitation; a transmit/receive wire coil for ³He NMR.

Polarisation of Atoms in a Magnetised Plasma (PAMP)
The new method simply requires operation in high magnetic field and excitation of the gas by a strong radiofrequency discharge. As pragmatically indicated by the acronym, ³He nuclear polarisation is achieved in a regime where the mean free path of the free electrons is much larger than their gyration radius in the rf gas discharge. Noteworthily, the measured nuclear spin polarisation exceeds the thermal equilibrium spin polarisation of the free electrons present in the discharge.
A series of NMR measurements has been performed at Johannes Gutenberg University (Mainz, Germany) on spherical ³He gas samples in 4.7 T, for different cell volumes, filling pressures, or rf excitation levels (with a discharge frequency of about 100 MHz). Typically, 1 − 10 % nuclear polarisation is obtained at room temperature in 8 – 19 mm diam. cells filled with 1 – 15 mbar ³He gas, possibly including samples with suboptimal gas purity. Further tests, performed in a 1.5 – 11 T range, and from room temperature to ca. 10 K, suggest that PAMP is not only potentially efficient but, also, fairly robust.
The scenario tentatively proposed in [1] involves: 1/ hyperfine hyperpolarisation of metastable ³He atoms, through strong atomic alignment (due to anisotropic RF-driven electronic bombardment) and alignment-to-orientation conversion (AOC) in the excited 2³P state via radiative decay at 1083 nm; 2/ polarisation build-up in the pool of ³He ground-state atoms, through metastability exchange collisions.

Work on PAMP
– In-depth investigation of PAMP at high magnetic field is included in the ANR-funded project HELPING (2021-2025, LKB and LDSRM/IRAMIS/CEA Saclay). Combination of NMR and optical spectroscopic studies will allow a detailed analysis of the PAMP process. As for MEOP, theoretical and experimental investigations will be jointly performed. The project targets experiments in He gas discharges at room temperature and 7 T, as well as PAMP tests in a variety of operating conditions.
– Further investigation of PAMP is underway at moderately low magnetic field (0.1T) in Paris at LKB.
– NMR measurements of PAMP-induced nuclear orientation are performed in Kazan (MRS Lab) on ³He gas samples in 3.66 T in a single spherical cell [2] or in a dual cell (spherical measurement cell connected by a capillary glass tube to a cylindrical hyperpolarisation cell) [3], with a discharge frequency of about 27 MHz.

[1] Nuclear hyperpolarization of ³He by magnetized plasmas, Maul A., Blümler P., Nacher P.-J., Otten E., Tastevin G., and Heil W., Phys. Rev. A (2018) 98, 063405 [12 pages]
[2] A. Makarchenko et al, Instrum Exp Tech 64:911–916 (2021).
[3] A. Makarchenko et al, Phys. Rev. A 106, 023101 (2022) – ArXive:2202.10307

Advances in metastability exchange optical pumping (MEOP) at high laser powers, but also at high gas pressures and high magnetic field strengths, has provided strong motivation for revisiting our understanding of this powerful technique and of its potential limitations. A comprehensive model has been developed for improved description of the combined effects of OP, ME, and relaxation, and of the detailed MEOP features observed over a broad range of operating conditions. This improved tool is used to explain the excellent photon efficiency of OP obtained at all field strengths. It also provides a new quantitative basis for analysis of the angular momentum budget and investigation of the recently discovered OP-enhanced polarisation losses. Simple linearised models developed in the 1970s have been sufficient to account for the MEOP experiments performed at that time, with low pump powers (He lamps) and low polarisations (around 10%). When laser sources are used, intrinsic non-linearities cannot be overlooked: the optical transition rates scale with the local light intensity but pump absorption depends on the involved sublevel populations that, in turn, significantly vary with both polarisation and light excitation. The main complexity however originates from the dispersion of optical transition rates due to the velocity dependence of the atomic Doppler shifts which, for instance, introduces correlations between atomic orientations and velocities for spectrally narrow light excitation. A detailed model for MEOP without restriction on the pump light intensity nor on the nuclear polarisation has been proposed in the 1980s [1]. This model accurately describes the effects of both ME and OP and accounts for relaxation by introducing a phenomenological mixing rate between magnetic sublevels in each energy state. However, the model introduces a crude dichotomy in the thermal velocity distribution, between a class of “effectively pumped” atoms (at resonance with the laser, coupled to the pumping light) and a class of “un-pumped” atoms (out of resonance, not excited by the laser light), in the 2³S state. This model only describes low field MEOP. It has been initially established for pump lasers with a few spectral modes and used for input parameters corresponding to operation at low pressure (millibar range) where MEOP was known to have the highest efficiency. This model has then be used to discuss the first OP experiments with a broadband laser [2], partially extended to specific other conditions such as OP of ³He-⁴He mixtures [3] or OP at high magnetic field in the Dehmelt regime (full collisional redistribution in the 2³P state) [4], [5], [6].
Recently, we have developed a more comprehensive model of MEOP [7]. It is based on similar rate equations for the populations of all Zeeman sublevels but it is suitable for arbitrary magnetic field, for pure ³He gas as well as for isotopic gas mixtures. It still includes a rather coarse description of the Maxwell distribution but introduces two broad velocity classes (“strongly pumped” atoms, in the centre of this distribution, and “weakly pumped” atoms, in the wings of the distribution). It therefore provides a unified frame that is much more adequate for description of MEOP with modern, broadband fibre lasers, for which the contribution of atoms in the second class becomes increasingly significant at high light powers. The self-consistent numerical computation of the local light intensity and solving of the coupled rate equations for the various populations (6 for 2³S state, 18 for the 2³P state) have also been improved so as to now take into account the spatial inhomogeneity introduced by the OP process, due to radial distributions of light intensity (for the incident laser beam) and of number density (for the metastable atoms). *** The improved model has a limited number of free parameters and can be used for quantitative predictions of MEOP efficiency in gas polarisers (polarisations, growth rates, and magnetisations, with or without gas flow in OP cells) when reliable input values are available.
Finally, a new approach has been proposed to describe the net growth rate of ³He nuclear polarisation in the ground state. It is based on a global angular momentum budget that combines the photon efficiency of the 2³S-2³P OP cycles and a total polarisation loss that only involves 2³S and ground state relaxation rates. This model has been satisfactorily validated at low magnetic field by extensive dedicated measurements. *** This global budget approach is useful to understand the origin of the excellent efficiency of MEOP at all field strengths. It allows a quantitative assessment of the additional losses induced by the pump laser light, that appears to drastically limit the nuclear polarisations reached in all experiments. References:

• [1] P.J. Nacher and M. Leduc, J. Physique (Paris) 46 (1985) 2057–2073, “Optical pumping in ³He with a laser” – Open access archive HAL.
• [2] M. Leduc, P.-J. Nacher, G. Tastevin, and E. Courtade, Hyperfine Interactions 127 (2000) 443–449,”Kinetics of helium-3 laser optical pumping”.
• [3] C. Larat, PhD thesis, 1991: “Laser LNA de puissance, application au pompage optique de l’hélium-3 et des mélanges hélium-3/hélium-4”.
• [4] E. Courtade, F. Marion, P.-J. Nacher, G. Tastevin, K. Kiersnowski, and T. Dohnalik Eur. Phys. J. D : Atom., Molec., Opt. and Plasma Phys. 21 (2002) 25–55, “Magnetic field effects on the 1083 nm atomic line of helium”.
• [5] P.-J. Nacher, E. Courtade, M. Abboud, A. Sinatra, G. Tastevin, and T. Dohnalik Acta Physica Polonica B 33 (2002) 2225–2236, “Optical pumping of helium-3 at high pressure and magnetic field”.
• [6] M. Abboud, A. Sinatra, X. Maitre, G. Tastevin, and P.-J. Nacher Europhysics Letters 68 (2004) 480–486, “High nuclear polarization of helium-3 at low and high pressure by metastability exchange optical pumping at 1.5 Tesla”.
• [7] M. Batz, P.-J. Nacher, and G. Tastevin, J. Phys. : Conf. Series 294 (2010) 012002, 21 p. “Fundamentals of metastability exchange optical pumping in helium” – Open access

Systematic studies of 2³S-2³P₀ pumping below 30 mT have been performed to understand the current limitations of ³He MEOP (M. Batz, PhD thesis, 2011). 30mT OP experiment

All-optical measurements have been carried out to monitor OP dynamics and quantitatively assess the efficiency of the transfer of angular momentum from the light to the atoms. Absorption-based measurements yield accurate absolute values of ³He nuclear polarisation and allow monitoring of MEOP dynamics with good signal-to-noise ratios. The transmitted pump light power is simultaneously recorded to monitor absorption of OP light by the atoms. The 2³S-2³P transition at 1083 nm was used for detection of the laser-induced orientation of ³He in the rf discharge. The method has been extended to operation in the presence of a strong OP laser beam. The changes induced by the OP light in the distribution of 2³S sublevel populations have been measured and taken into account to accurately infer the nuclear polarisation. The experimental data have also been used for quantitative tests of the numerical predictions of our improved MEOP model [1]. Time recordings of two probe signals, each proportional to population of a selected 2³S sublevel. Sharp discontinuites are observed when pump is turned on/off in this dedicated experiment. Polarisation-dependent jumps are large at low pressure and strong pump power.

A detailed comparison has been made between the measured polarisation build-up rates (number of polarised atoms per unit time) and the measured OP light absorption rates (number of absorbed pump photons).

– Measurements performed at null polarisation provide accurate relaxation-free data that allow quantitative tests the improved MEOP model and provide access to photon efficiencies.

– Measurements performed during polarisation build-up provide direct access to polarisation losses in the system, obtained as the net differences between the incoming and the stored amounts of angular momentum. The loss rates have been found to systematically exceed (by up to two orders of magnitude) the relaxation rates measured in the rf plasmas when optical pumping is stopped.

Example of measured (symbols) and expected (line) steady-state polarisations at 2.6mbar (C9 pump, 1 mT). At fixed decay time in the plasma without OP (850 s), the plateau must be attributed to additional losses during OP. Extra relaxation times decrease from 550 s at 0.2 W to 25 s at 5 W pump power to account for these data.

To quantitatively account for the faster-than-expected decrease of polarisation growth rates and the lower-than-expected values of steady-state polarisations, one must introduce in the model an additional relaxation rate that consistently turns out to be proportional to the absorbed laser power.

Example of experimental data demonstrating a linear increase of the additional polarisation loss rate with the absorbed pump power, at fixed plasma conditions.

A similar behaviour is actually observed in our MEOP experiments for all the investigated ranges of gas pressures or discharge conditions and at all magnetic field strengths. Search for the physical origin of the ubiquitous light-enhanced relaxation is under way. To this aim, new dedicated optical tools are developed and implemented to suitably probe plasma changes during MEOP. Processes potentially associated with loss rates that vary linearly with absorbed pump light power include:

– Radiation trapping: Absorption of the 1083 nm fluorescence light that is randomly re-emitted by the 2³P atoms excited by the OP beam necessarily contributes to depolarisation. Two complementary series of investigations have been launched to quantitatively assess this effect in usual MEOP conditions. Our experimental findings at low gas pressures seem yet inconsistent with some of the expected features.

– Collisions with metastable He dimers: Weakly bound dimers may be obtained in He plasmas by association of excited and ground state atoms in 3-body collisions. They have a nearly 100 times larger cross-section for formation from atoms in the excited 2³P state than in the metastable 2³S state. An experiment has been set up to simultaneously record MEOP dynamics and probe He₂* number density (B. Glowacz, PhD thesis, 2011). To this aim, a tunable blue laser has been developped for detection of the metastable He₂* molecules and absorption measurements at 465 nm [2]. As expected, the abundance of these weakly bound dimers (obtained in He plasmas by association of excited and ground state atoms in 3-body collisions) is found to be laser-enhanced during OP. However, preliminary investigations around 30 mbar indicate that the observed increase in He₂* density is significant but limited, if compared to the measured increase of ³He polarisation loss rate. Furthermore, at low pressure (where most extensive MEOP studies have been performed) He₂* molecules are necessarily much less abundant.

– Excitation transfer between excited 2³P and ground state atoms: In contrast with metastability exchange (excitation transfer with between two L=0 atoms), this collisional process may induce changes in nuclear spin orientation. An experiment has been set up to probe collisional redistribution in the 2³P state, assess the contribution of 2³P-ground state collisions, and infer the value of the cross-section for excitation transfer (A. Dia, PhD thesis, 2021).   References:

First evidence of the benefits of operation in a high magnetic field for metastability exchange optical pumping (MEOP) of ³He has been obtained at 0.1 T [1], [2]. Nuclear polarisations obtained in pure ³He gas at 32 mbar are, for instance, twice larger than those obtained at 1 mT [3]. Example of a MEOP-based system (left) operated inside a 0.1T whole-body MRI scanner at IR4M in Orsay (right). Tens of millibars of pure ³He gas are laser polarised in the cylindrical cell, extracted by our dedicated peristaltic compressor, and transferred into the plastic bag that will be handed to the subject for inhalation.

Systematic investigation of MEOP efficiency at high magnetic field has been undertaken, both theoretically and experimentally.  A detailed study of OP and ME processes has been carried out to take into account the deep changes in atom level structure when, due to the increased field strength, the Zeeman effect becomes significant compared to the fine and hyperfine interactions [4]. It has lead to the development of an absorption-based method for ³He polarisation measurement and of an improved MEOP model that are both valid at arbitrary magnetic field strength. Water-cooled resistive magnet built at LKB to investigate ³He MEOP up to 0.1T Optical access to the OP cell is also designed for transverse probe beams.

 Experimental studies of high-field MEOP have started at 1.5 T in a whole-body clinical scanner used for ³He MRI, in collaboration with U2R2M [5], [6], [7]. MEOP set-up with pump and probe optics (left) designed for systematic studies at Kremlin-Bicêtre hospital in the former CIERM 1.5T MRI scanner (right).  Work has been continued in collaboration with the group of Pr T. Dohnalik in Cracow. MEOP has been investigated at 0.45, 0.9, 1.5 and 2T [8]. MEOP setup (left) and superconducting magnet (right) used for the extensive 0.45 – 2T study in Krakow University. Optical fibres are used in order to let the lasers sources sit at low field, in a remote part of the room.  Complementary MEOP experiments have been carried out inside a 4.7T MRI system implemented in the Institute for Nuclear Physics in Cracow. [9] A similar experimental setup has been used. The reduced width of the whole MEOP system matches the smaller bore size of this superconducting magnet.

—>A strong body of consistent experimental data has been collected. At low pressure (1 mbar), high nuclear polarisations are obtained at all field strengths. At high pressure, much higher polarisations are obtained at high field than at low field. At low field (mT) nuclear polarisations do not exceed a few percents above 10 mbar (cyan diamonds). Much better results are systematically at higher fields (other symbols): polarisation reach up to 55% (resp.30%) around 100 mbar (resp. 270 mbar) at 2T, for instance. Published in Eur. Phys. J. D (2013), 67:200 – Open access article.

Our pioneering studies of the angular momentum budget for high field MEOP, performed at 1.5T (M. Abboud, PhD thesis, 2005) show that photon efficiency remains high (1 or 2 polarised atoms per absorbed photon, depending on the 2³S-2³P line component used for OP) and that light-enhanced polarisation losses are about an order of magnitude smaller than at low field, which certainly explains the systematically larger nuclear polarisation measured in the experiments. Investigation of the field dependence of the measured additional loss rates may thus provide clues about the involved physical processes. Even better results may be achievable with optimised operating conditions, in particular to increase the polarisation build-up rates (higher collision rate only partly compensates for lower amount of polarisation transferred by ME, per collision, at large hyperfine decoupling). Operation at high magnetic fields already provides the largest magnetisations ever obtained with MEOP in pure ³He gas. It also makes requirements for compression stages in polarized gas production systems much less demanding. A prototype high-field gas polariser [10] has been built in Krakow for operation in a 1.5T clinical scanner.

A quantitative study of high-field MEOP in He gas discharges is planned in the ANR-funded HELPING project (2021-2025). Theoretical and experimental investigations will be jointly performed. They target laser-enhanced losses in room temperature ³He gas at 7 T and extension of the analysis of high-field MEOP dynamics to a broad range of operating conditions.

References:

• [1] E. Courtade, F. Marion, P.-J. Nacher, G. Tastevin, T. Dohnalik, K. Kiersnowski, Hyperf. Interact. 127 (2000) 451-455, “Spectroscopy of the helium 2³S-2³P transition above 0.01Tesla – Application to optical pumping studies” – Open access archive HAL
• [2] E. Courtade, PhD thesis, Paris (2001): “Pompage optique de l’hélium dans des conditions non-standard”
• [3] Nacher P.-J.et al., Acta Phys. Polon. B 33 (2002) 2225-2236: “Optical pumping of ³He at high pressure and magnetic field” – Open access archive HAL”.
• [4] Courtade E., et al., Eur. Phys. J. D 21 (2002) 25-55, “Magnetic field effects on the 1083 nm atomic line of helium” – Open access archive HAL.
• [5] M. Abboud, A. Sinatra, X. Maitre, G. Tastevin, P.-J. Nacher Europhys. Lett. 68 (2004) 480–486: “High nuclear polarization of ³He at low and high pressure by metastability exchange optical pumping at 1.5 Tesla” – Open access archive HAL
• [6] M. Abboud, A. Sinatra, G. Tastevin, P.-J. Nacher, and X. Maitre Laser Physics 15 (2005) 475-479: “Metastability exchange optical pumping of ³He at high pressures and 1.5T: comparison of two optical pumping transitions” – Open access archive HAL.
• [7] M. Abboud, PhD thesis (Paris, 2005): “Pompage optique de l’hélium-3 à forte pression dans un champ magnétique de 1.5 Tesla”
• [8] A. Nikiel et al, Eur. Phys. J. Special Topics 144 (2007) 255−263, “Metastability exchange optical pumping of ³He at high pressure and high magnetic field for medical applications”.
• [9] A. Nikiel-Osuchowska, G. Collier, B. Głowacz, T. Pałasz, Z. Olejniczak, W.P. Węglarz, G. Tastevin, P.-J. Nacher, and T. Dohnalik, Eur. Phys. J. D (2013) 67:200, “Metastability exchange optical pumping of 3He gas up to hundreds of millibars at 4.7 Tesla.” – Open access
• [10] G. Collier, PhD thesis (Krakow, 2011)

Optical pumping (OP) is a very efficient technique, introduced by A. Kastler (Nobel Prize 1966), that achieves control of the internal state of the atoms through interaction with a resonant light beam carrying angular momentum. If relaxation processes are weak, in the presence of a magnetic guiding field the net result of the repeated OP cycles (selective absorption and random re-emission of light) is an imbalanced distribution of the atoms among the involved magnetic sublevels, i.e., creation of spin orientation. Helium-3 can be hyperpolarised by two established optical pumping methods, developed in the 1960s for nuclear physics experiments and constantly revisited for improved efficiency:  spin exchange optical pumping (SEOP) [1], where alkali metal atoms (such as rubidium) are optically pumped and further transfer of angular momentum to the ³He nuclei occurs through magnetic interactions during binary alkali-helium collisions.  metastability exchange optical pumping (MEOP) [2], where optical pumping for ³He metastable atoms (He*) induces both electronic and nuclear orientation thanks to strong hyperfine coupling in this energy level (direct electron-nucleus magnetic interaction, within the He atom) and further transfer of angular momentum from He* atoms to ground state He atoms through metastability exchange (ME) collisions (a binary exchange of electronic excitation with no loss of nuclear orientation). Each method has its own series of advantages and limitations. SEOP can be applied to other noble gases, directly operates at high pressure (bars), but it is a slow pumping process for ³He (several hours). MEOP exclusively applies to helium, involves pure He gas and best operates at low pressure (millibar), but it is a much faster pumping process (seconds). Left: schematic view of an OP setup. Right: atomic levels of ³He involved in MEOP. Metastability exchange optical pumping involves the 2³S-2³P transition at 1083 nm wavelength for OP. The metastable 2³S state is usually populated by electronic collisions in a weak rf plasma discharge and the relative abundance of He* atoms is low (a few ppm). In standard conditions, MEOP is performed at low magnetic field (mT). Polarisations vary with OP conditions (rf discharge, OP line, cell geometry, light beam features) but they consistently exhibit a maximum around 0.5 mbar and a sharp decrease when the operating pressure increases. A few tens of milliwatts of circularly polarised 1083nm laser light are generally sufficient to overcome plasma-induced relaxation and large nuclear polarisations can be obtained under optimal conditions (typically 50% with 50 mW and up to up 80-90% with a few watts in pure ³He). However build-up rates increase if more light power is absorbed, which is often desirable for applications of hyperpolarised ³He. In the metastable state OP transitions (a few 10⁶ s⁻¹ per W/cm²) and ME (3.8 10⁶ s⁻¹/mbar at room temperature) are fast processes. ME tends to establish a distribution of 2³S populations in spin-temperature equilibrium with the ground state nuclear polarisation. The OP competes with ME and relaxation and the induced redistribution of atoms among 2³S sublevels provides the driving term for ³He polarisation build-up. Polarisation grows much more slowly in the ground state (through ME between minority optically-pumped He* and majority He atoms) and uniformly (thanks to atomic diffusion). When OP is stopped, nuclear polarisation decays irreversibly towards its (negligible) equilibrium value on a time scale that is, in practice, limited by wall relaxation in the cell and varies from a few seconds to thousands of seconds in the rf plasma (depending on discharge intensity). MEOP also efficiently operates in ³He-⁴He isotopic mixtures [3], [4], [5]. Higher nuclear polarisations and faster pumping can be achieved, which may be of interest for applications where ⁴He is needed (e.g., low temperature studies of polarised He mixtures) or at least not a nuisance (e.g., neutron spin filters). Optical measurement of nuclear polarisation The degree of circular polarisation of the fluorescence light emitted by the plasma discharge can be measured for a selected He line in the visible spectrum. Atoms in the light-emitting excited state retain part of the nuclear polarisation of the ground state and their radiative decay light is partially circularly polarised. Measurements are conveniently performed on the red (668 nm) or orange (588 nm) line using a static optical polarimeter [6]. However, this technique is inefficient at high gas pressure (above a few mbar, due to depolarising collisions) or at high magnetic field (above 20 mT, due to hyperfine decoupling in the excited states). It also requires absolute calibration as a function of gas pressure and ³He concentration in the case of isotopic mixtures. Absorption measurements performed with a 1083 nm probe laser provide a way to measure relative populations of the 2³S₁ metastable sublevels and to infer the nuclear polarisation [7]. This method yields absolute values of the ³He nuclear polarisation and is valid at all magnetic field strengths. It has been implemented for all our investigations of high-field MEOP. It is used for dual on-line monitoring of polarisation buid-up the twin OP cells of our upgraded gas polariser. Various systematic effects have been studied in detail [8].

References:

• [1] T. G. Walker J. Phys. : Conf. Series 294 (2010) 012001 “Fundamentals of Spin-Exchange Optical Pumping” – Open access
  
• [2] M. Batz, P.-J. Nacher, and G. Tastevin, J. Phys. : Conf. Series 294 (2010) 012002, 21 p. “Fundamentals of metastability exchange optical pumping in helium” – Open access
  
• [3] E. Stoltz, M. Meyerhoff, N. Bigelow, M. Leduc, P. -J. Nacher and G. Tastevin Applied Physics B 63 (1996) 629-633: “High nuclear polarization in ³He and ³He-⁴He gas mixtures by optical pumping with a laser diode”.
• [4] Christian Larat, PhD thesis (Paris, 1991): “Laser LNA de puissance, application au pompage optique de l’³He et des mélanges ³He-⁴He”
• [5] E. Courtade, F. Marion, P.J. Nacher, G. Tastevin, T. Dohnalik, K. Kiersnowski, Hyperf. Interact. 127 (2000) 451-455
• [6] Stoltz E., et al., Applied Physics B 63 (1996) 635-640 “Polarization analysis of the light emitted by an optically pumped ³He gas”.
• [7] Bigelow N.P., Nacher P.J. and Leduc M., J. Physique II 2 (1992) 2159-2179: “Accurate optical measurement of nuclear polarization in optically pumped ³He gas” – Open access archive HAL.
• [8] C. Talbot, M. Batz, P.-J. Nacher, and G. Tastevin, J. Phys: Conf. series 294 (2010) 012008, 12 p. “An accurate optical technique for measuring the nuclear polarisation of He-3 gas” – Open access

In the 1980s the LKB has been strongly involved in the early development of tunable 1083 nm solid-state laser devices for ³He OP, with cw output powers that rapidly went up from a few milliwats to several watts. This has opened the way to detailed numerical and experimental investigations of MEOP dynamics [1]. The move from He lamps to dedicated lasers has also boosted up the use of MEOP techniques for a variety of applications.

In the early 1990s monochromatic laser diodes have become commercially available at 1083 nm wavelength. They are suited for He magnetometry. They are also very useful for the absorption-based measurements of ³He polarisation and of metastable atom density in He plasmas.

Simultaneously Yb-doped silica fibers have emerged as convenient and powerful gain media in the near infrared range. We have demonstrated that ³He MEOP can be successfully performed with a fibre-based MOPA system [2] that combines a 1083 nm laser diode, used as master oscillator, and a Yb-fibre power amplifier. These MOPA devices have quickly become commercially available. They are well suited for laser cooling and trapping of metastable helium atoms, for instance. More powerful systems have been developed, e.g., for LIDAR applications.

Monochromatic excitation of the Doppler-broadened He line (1.9GHz fwhm for ³He at 300K) is however far from optimal for MEOP [1]. We have shown that fast frequency modulation applied to single-mode lasers [3] is a way to effectively address more velocity classes within the Doppler spectrum, hence to obtain significantly higher light absorption and better OP. Modulation of the driving current has thus been implemented in commercial MOPA devices designed for ³He MEOP. More recently, our group has shown that a non-periodic modulation scheme involving white noise modulation of laser frequency can be used for enhanced efficiency of ³He optical pumping at 1083 nm [4].

Dedicated Yb-doped fibre lasers at 1083 nm [5] have been developed for MEOP application at the turn of the century and are now commercially available. They deliver several watts, are tunable over the whole He fine and hyperfine structure spectrum (of order 100 GHz, depending on the operating magnetic field), and have a line width matched to the helium Doppler width.

References:

MEOP-based gas production A major issue for a wide dissemination of ³He MRI is access to hyperpolarised gas. Metastability-exchange optical pumping (MEOP) provides both fast production rates and high nuclear polarisations, which makes the production systems based on this technique best suited for MRI applications. However MEOP operates best at low pressure (millibar) and compression of the polarised gas is needed to allow NMR and imaging with high signal-to-noise ratio. Two routes with complementary advantages can be pursued: Centralised production of very high grade hyperpolarised gas and transportation to the MRI centres. This requires a large facility for massive production and storage in dedicated containers to avoid prohibitively fast decays of ³He nuclear polarisation. On-site production of gas with compact polarisers implemented next to the MRI scanners. The delivered gas may thus be produced with a lower grade (i.e., optically pumped to a moderate ³He polarisation or highly polarised and subsequently diluted in some neutral gas, like nitrogen) and problems with long term storage and transportation are avoided. The main advantage of the on-site production scheme is flexibility for the end user: polarised gas can be repeatedly produced with an adjustable rate, whenever requested for each series of experiments. A compact ³He gas polariser for on-site production We have developed a helium polarising system for MRI applications, designed to combine efficiency, ease of use, and compactness. The key element of this ³He polariser is a patented polarisation-preserving peristaltic compressor that allows production of hyperpolarised gas next to or inside the MRI scanner. The first prototype built at LKB has been typically able to produce ³He gas with 40% nuclear polarisation for a yield of 0.15 litre/hour. Another stand-alone system, derived from this prototype system but significantly more compact (designed to fit in the trunk of a car), has been built for a series of joint LKB- U2R2M lung MRI experiments carried out near Paris. Hospital production unit (size: 1.1 x 0.64 x 0.64 m³) routinely used for ³He MRI since 2003 The U2R2M (Paris 11-CNRS, UMR 8081) has then kept this dedicated polariser running for several years: at CIERM (Centre Inter-Etablissement en Résonance Magnétique), a clinical MRI research facility formerly implemented at Kremlin-Bicêtre Hospital (until Sept. 2011), for ³He lung MRI at 1.5T in the U2R2M physics laboratory (Paris Sud university campus, Orsay) for methodological studies of ³He MRI at 0.1T. The “U2R2M” research lab has now evolved to become IR4M and the CIERM has moved to SHFJ (Service Hospitalier Frédéric Joliot, CEA / I2BM / SHFJ – Centre Hospitalier d’Orsay), where the gas polariser is still operated. Polariser upgrade The in-house compact system used for development and test, as well as gas production for the very low field experiments performed at LKB, has been upgraded for higher throughput and better efficiency. The footprint remains unchanged but the coils, cells, and optics have been modified. Light is more efficiently absorbed in the larger OP volume (3.8L, 2 long cylinders in series). The system is now equipped with a new optical polarisation measurement setup [1], for improved diagnosis and run-time monitoring. Without gas flow the ³He nuclear polarisation reaches 70% at 1.2 mbar gas pressure in the OP cells. Current yield of compressed gas meets the expected 3-fold improvement: 0.5 L/hour with 50% ³He polarisation. Gas polarisation system before (left) and after (right) upgrade. A similar upgrade of the polariser design is being performed for the new gas polarising system built and implemented at SHFJ hospital (Orsay) by IR4M for lung MRI. Other MEOP-based gas polarisers Applications of polarised ³He are indeed not limited to human lung MRI, and include small-animal lung MRI, polarised targets for nuclear physics experiments, neutron spin filters, diffusion studies in porous materials, search for permanent electric dipole moments violating parity or symmetry rules. Centralised production has been successfully established by the Mainz group. Liters of pure ³He gas are produced with nuclear polarisation in the 70% range and pressure around 2 and 3 bars, then delivered to the end users by car, train or plane [2], [3]. This group actually pioneered the development of dedicated systems for massive production of hyperpolarised ³He gas for a variety of applications, including the most demanding ones (neutron spin filters and nuclear targets). Their very sophisticated optimised piston compressors lead to so far unrivalled performance in terms of output pressure, gas polarisation (in the high 80%), and production rate. The LKB on-line production system for MRI applications has been duplicated through bilateral collaborations or dissemination projects, such as PHIL or PHeLINet (see , e.g., the low field polariser built in Krakow). A few other compression schemes have also tested or used for MEOP-based on-line gas production. Operation with a modified diaphragm pump, for instance, has been tried at NIST [4]). Several other groups have developed on-line production units using dedicated piston compressors to obtain polarised ³He gas for their research activities (at ILL [5] and, with a more compact design, at Mainz University [6]). In parallel several SEOP-based systems have been developped for polarisation of ³He. Production rates with this OP technique are likely to increase in a near future, due to recent breakthroughs (e.g., faster SEOP demonstrated with Rb-K mixtures).

References:

• [1] C.Talbot, M. Batz, G. Tastevin, and P.J. Nacher, J. Phys: Conf. Series 294 (2011) 012008 (12 pages): “An accurate optical technique for measuring the nuclear polarisation of ³He gas” – Open access
• [2] E.J.R. van Beek, J. Schmiedeskamp, J.M.Wild, et al., Eur. Radiol. 13 (2003) 2583-2586 “Hyperpolarized 3-helium MR imaging of the lungs: testing the concept of a central production facility”
• [3] F. Thien et al, Respirology 13(2008) 599–602, “Feasibility of functional magnetic resonance lung imaging in Australia with long distance transport of hyperpolarized helium from Germany”
• [4] T.R. Gentile, et al., J. Res. Nat. INST Stan.106 (2001) 709-729 ”Compressing spin-polarized ³He with a modified diaphragm pump”- Open access
• [5] K.A. Andersen et al., Physica B, 356 (2005)103-108 “First results from Tyrex, the new polarized-He-3 filling station at ILL”
• [6] C. Mrozik et al., J. Phys.: Conf. Series 294 (2011) 012007 “Construction of a compact ³He polarizing facility” – Open access

Challenge 1 : Explain laser-enhanced losses during MEOP

• “Experimental study of redistribution and exchange collisions in the 2³P level of helium”  (A. Dia, PhD thesis, 2021).

Accurate characterization of the collisional redistribution of atoms has been achieved by laser absorption spectroscopy on the 2³P-3³S transition at low magnetic field (mT) at low gas pressure, using a double-resonance ladder scheme and two monochromatic laser beams. A tunable laser diode at 1083 nm (pump) was operated at fixed wavelength to selectively excite  2³S atoms and populate some 2³P Zeeman magnetic sublevel(s). A tunable laser diode at 706.5 nm was used to scan the entire 2³P-3³S absorption spectrum and probes the steady-state atomic distribution among the accessible 2³P sublevels.

– Spectroscopic findings: Computation of expected 2³P-3³S line positions and strengths for ³He and ⁴He isotopes and comparisons with spectra obtained in pure gas samples and isotopic gas mixtures allowed quantitative measurements of the 3³S state hyperfine structure of ³He and of the ³He-⁴He isotopic line shift. With suitable assignments and line shape adjustments, contributions of the various line components could be analysed. Typically, line areas yielded access to atomic number densities in the probed Zeeman sublevels while line shapes reflected the distributions of atomic velocities within each sublevel.

– Rate equation models: The probed lines exhibited pure Lorentzian, pure Gaussian, or mixed profiles (superposition of a Lorentzian peak and of a broader Gaussian pedestal), as a result of velocity selective optical pumping (VSOP) by the 1083 nm laser and redistribution of atoms by one/a few collisions. Phenomenological descriptions of the latter were developped to account for relevant processes (collisions with no/little or large velocity changes,  with or without changes in magnetic sublevel and/or in energy level). Constraints on transfer rates between sublevels could be derived from the experimental data. 

– Excitation transfer cross section :

Qualitative evidence of excitation transfer between isotopes has been obtained in ³He-⁴He gas mixtures by, e.g., tuning the pump laser to a 1083 nm even isotope line and probing the odd isotope lines at 706.5 nm (or vice versa) – A. Dia et al, SFP conference, 2019: “Spectroscopic study of collisions in the 2³P state of ³He and ⁴He in low pressure gas discharges”)

Quantitative comparison of line areas in 706.5 nm obtained in samples with distinct total gas pressures and isotopic ratios provides a estimate for the excitation transfer cross-section. Interestingly, this result value would quantitatively explains the measured values of  laser-induced loss rates in MEOP if one assumes full collisional mixing in the 2³P state and a coupling rate between 2³P and ground state nuclear polarisations determined by the excitation transfer collision rate.

• Prospects

Further measurements of collisional rates are needed, in isotopic mixtures as well as in pure ³He gas. A reliable measurement of the excitation transfer cross section is needed. It might be compared with recent ab initio computations [Vrinceanu et al, 2010].

Impact on low-field MEOP dynamics should be established. This is expected to open the way to more accurate MEOP models and, hopefully, to clues for overcoming current limits of the overall efficiency of hyperpolarisation of ³He  nuclei. Obviously, extension to high magnetic field strength should follow.

Challenge 2: Hyperpolarisation at high magnetic field  (HELPING project, 2021-2025)

• High field MEOP

• PAMP