Near-unity nuclear polarization with an open-source 129Xe hyperpolarizer for NMR and MRI

Abstract

The exquisite NMR spectral sensitivity and negligible reactivity of hyperpolarized xenon-129 (HP(129)Xe) make it attractive for a number of magnetic resonance applications; moreover, HP(129)Xe embodies an alternative to rare and nonrenewable (3)He. However, the ability to reliably and inexpensively produce large quantities of HP(129)Xe with sufficiently high (129)Xe nuclear spin polarization (P(Xe)) remains a significant challenge--particularly at high Xe densities. We present results from our "open-source" large-scale (∼1 L/h) (129)Xe polarizer for clinical, preclinical, and materials NMR and MRI research. Automated and composed mostly of off-the-shelf components, this "hyperpolarizer" is designed to be readily implementable in other laboratories. The device runs with high resonant photon flux (up to 200 W at the Rb D1 line) in the xenon-rich regime (up to 1,800 torr Xe in 500 cc) in either single-batch or stopped-flow mode, negating in part the usual requirement of Xe cryocollection. Excellent agreement is observed among four independent methods used to measure spin polarization. In-cell P(Xe) values of ∼90%, ∼57%, ∼50%, and ∼30% have been measured for Xe loadings of ∼300, ∼500, ∼760, and ∼1,570 torr, respectively. P(Xe) values of ∼41% and ∼28% (with ∼760 and ∼1,545 torr Xe loadings) have been measured after transfer to Tedlar bags and transport to a clinical 3 T scanner for MR imaging, including demonstration of lung MRI with a healthy human subject. Long "in-bag" (129)Xe polarization decay times have been measured (T1 ∼38 min and ∼5.9 h at ∼1.5 mT and 3 T, respectively)--more than sufficient for a variety of applications.

Keywords: hyperpolarization; laser-polarized xenon; lung imaging; optical pumping.

Fig. 1.

The XeNA polarizer. (A) Schematic of the polarizer’s key components (liquid N₂ dewar not shown for clarity). The optical path (shown in B) is represented by (“λ/4”). For the gas cylinders, “N” and “E” designate Xe with naturally abundant ¹²⁹Xe and isotopically enriched ¹²⁹Xe, respectively.

Fig. 2.

In situ monitoring of Rb electron spin polarization. Near-IR spectra of the pump laser transmitted through the cell at room temperature (“cold cell,” black lines); 57 °C and B₀ = 5.26 mT (blue lines); 57 °C and B₀ = 0 mT (green lines); 65 °C and B₀ = 5.26 mT (red dashes); and 65 °C and B₀ = 0 mT (orange dashes). Three scans for each condition are shown; error bars were determined from SDs of spectral integral values. (Insets) The effects of depletion pumping of the Rb vapor on the transmitted laser intensity (assuming σ⁺ circularly polarized light). When B₀ is “off” (Left), optical pumping is inefficient, resulting in near-equal populations of the ground electronic spin states (m_J = +1/2 and −1/2) and a high density of gas-phase absorbers. An applied magnetic field along the quantization axis (provided by the polarized pump laser) results in more efficient optical pumping and unequal ground state populations, corresponding to high electronic spin polarization; the reduced number of atoms in the m_J = −1/2 ground state gives rise to increased laser transmission (Right).

Fig. 3.

Determination of P_Xe at low and intermediate magnetic fields. (A) HP¹²⁹Xe NMR spectrum from the SEOP cell containing 1,545 torr Xe (and 455 torr N₂) acquired at 5.26 mT (1 scan; 59-μs rf pulse; 30° tipping angle—i.e., τ_30° = 59 μs). (B) Reference NMR spectrum from water ¹H spins (111 M), doped with 5 mM copper sulfate, thermally polarized at 1.46 mT [170,000 scans; τ_30° = 16 μs; repetition time (TR), 0.3 s; P_H = 5.0 × 10⁻⁹]. (Inset) ¹H free induction decay for the spectrum in the main figure; SNR was used to calculate a relative error bar of 5.7% for P_Xe values. (C) HP¹²⁹Xe NMR spectrum recorded at 47.5 mT (1 scan; τ_30° = 18 μs) from a 52-mL phantom following transfer of some of the gas (originally 300 torr Xe, 1,700 torr N₂) to the phantom. (D) Reference ¹³C NMR spectrum from 170 mmol of sodium 1-¹³C acetate, 14 g in 50 mL of D₂O, also measured at 47.5 mT (256 scans; τ_90° = 54 μs; TR = 200 s; P_C = 4.1 × 10⁻⁸).

Fig. 4.

High-field ¹²⁹Xe measurements. T₁ decay of HP¹²⁹Xe NMR signals from Xe gas transferred to Tedlar bags and kept at 3 T throughout the decay (blue triangles) or stored at ∼1.5 mT (red circles) or ∼0 mT (white squares) and rapidly transferred to/from 3 T for acquisition. Loss from rf pulsing was negligible (tipping angle, <1°). (Inset) (Left) Spectrum from HP¹²⁹Xe in a Tedlar bag containing ∼800 cc of gas, 38% Xe by volume following transport to a 3-T clinical MRI (1 scan; 1.4° tipping angle). Gas mixture was nearly identical to main figure. (Right) Reference spectrum from thermally polarized ¹²⁹Xe spins in the 3-L spherical phantom containing a Xe/O₂ mixture (32 scans; 90° rf pulses).

Fig. 5.

Performance of the XeNA polarizer. (A) P_Xe values measured at 5.26 mT, 47.5 mT, and/or 3 T, plotted versus Xe pressure before SEOP. Labels “b.t.” and “a.t.” refer to Xe gas remaining in the SEOP cell before and after some was transferred to another container, respectively. Error bars are determined from uncertainties in the spectral integrals obtained from thermally polarized reference samples. The 725-torr value was obtained with 82%-enriched ¹²⁹Xe. (Inset) Two-dimensional fast low-angle shot ¹²⁹Xe projection image of HPXe in a Tedlar bag following transport to the 3-T MRI [field of view (FOV), 40 cm; slice thickness, 1.7 cm; matrix size, 80 × 80; echo time (TE)/repetition time (TR), 2.6/5.1 ms; tipping angle, ∼4°; SNR, ∼40:1]. (B) Two slices selected from a 3D ¹²⁹Xe GRE chest image from a healthy human volunteer following inhalation of HPXe from an 800-cc Tedlar bag (filled via expansion from a 500-cc cell containing 1,300 torr HPXe (86%-enriched ¹²⁹Xe) and 700 torr N₂). Subject performed two respiration cycles (total lung capacity to functional residual capacity), inhaled from bag, and then took a small gulp of air (to help push HPXe out of the trachea); TE/TR, 1.12/11 ms (specific absorption rate-limited); tipping angle, 6°; 80 × 80 × 14; acquisition time, 4.5 s; FOV, 320 × 320 × 196 mm³; 2 × 2 × 14 mm digital resolution after zero-filling (SNR ∼8–15). In-cell P_Xe was 25%, reduced from typical values due to the use of a SEOP cell that had to be thermally regenerated following partial oxidation (during installation).