Magnetic Shielding of Parahydrogen Hyperpolarization Experiments for the Masses

Graphical Abstract

We report on the inexpensive ($80) custom electronics for efficient robust and fast degaussing of mu-metal shields for use in parahydrogen induced polarization experiments. The degaussing process is fully automated and can be performed in ~2 seconds resulting in residual magnetic field of <20 nT. The utility of this device is demonstrated for ¹⁵N hyperpolarization of [¹⁵N₃]metronidazole below 1 microtesla. The reported electronics will be of practical use to those working in the field of parahydrogen utilizing both hydrogenative approaches for magnetic field cycling and non-hydrogenative approaches employing sub-microtesla magnetic fields to create level anti-crossing for polarization transfer from parahydrogen-derived nascent protons to the heteronucleus.

Introduction

NMR techniques have inherent low sensitivity due to low nuclear spin polarization (P) or the degree of alignment of nuclear spins with applied magnetic field. However, P can be transiently enhanced by several orders of magnitude via NMR hyperpolarization techniques.^[1–5] Corresponding gains in signal-to-noise ratio (SNR) of hyperpolarized (HP) dilute biologically compatible compounds enable new applications: most notably, in vivo tracking of biological processes.^[6–9] Parahydrogen Induced Polarization (PHIP) allows for efficient and low-cost hyperpolarization of nuclear spins using parahydrogen (p-H₂).^{[10, 11]} PHIP was pioneered in 1986,^[10] and the utility of p-H₂-induced hyperpolarized (HP) compounds for in vivo imaging applications was demonstrated in 2001.^[6] Despite substantial developments over the past decade,^[12–14] this hyperpolarization technique is certainly lagging behind compared to more established Spin Exchange Optical Pumping (SEOP)^[15] and dissolution Dynamic Nuclear Polarization (d-DNP),^{[2, 4]} which have numerous completed and pending clinical trials.^[16–18] It is our opinion that successful PHIP clinical translation requires robust hyperpolarization equipment support similar to that of SEOP and d-DNP techniques. The work reported here is aimed to enable the PHIP community with an inexpensive tool to improve the robustness and precision of PHIP hyperpolarization equipment.

In PHIP, once the symmetry of nascent parahydrogen-derived protons is broken, the spin order can be transferred to X-nucleus (anything but protons, e.g., ¹³C or ¹⁵N) via spin-spin couplings through the use of nanotesla magnetic fields.^[19–21] The process of polarization transfer requires precise manipulation of static magnetic fields in the range of 10–1000 nT, where spin level anti-crossings (LAC) are created between nascent p-H₂-derived protons and to-be-hyperpolarized nucleus.^[22]

In hydrogenative PHIP, p-H₂ is first added to the unsaturated precursor in a pairwise manner. Next, the magnetic field is decreased below B_LAC (ca. 100 nT). Finally, the adiabatic passage via B_LAC enables efficient polarization transfer to X-nucleus, Figure 1b.^[23] This approach has been called Magnetic Field Cycling (MFC)^[19] or Sweeping (MFS).^[24] P_13C of up to 21% has been obtained^[19] and P_13C of 5–10% has been demonstrated for [1-¹³C]pyruvate,^[25–27] the leading ¹³C HP contrast agent currently under evaluation in many clinical trials.

Figure 1.

Schematics of SABRE-SHEATH and PHIP MFC hyperpolarization techniques.

The Non-hydrogenative PHIP variant called Signal Amplification by Reversible Exchange (SABRE) employs the process of simultaneous chemical exchange of p-H₂ and to-be-hyperpolarized molecule on a metal center.^[28–31] The polarization transfer complex (PTC) containing p-H₂-derived hydrides and substrate containing X-nucleus (¹⁵N,^{[32] 31}P,^{[33] 13}C,^{[21, 34, 35]} etc.) is transiently formed. The efficient polarization transfer from p-H₂-derived hydrides to X-nucleus is achieved via LAC creation by matching i) the spin-spin coupling between hydride proton and X-nucleus, ii) difference of gyromagnetic ratios between hydride protons and X-nucleus, iii) and applied static magnetic field B_LAC. This condition is fulfilled at magnetic fields in the range of 0.2–1 μT,^{[20, 36]}, and this approach was called SABRE-SHEATH (SABRE in SHield Enables Alignment Transfer to Heteronuclei). B_LAC is applied throughout the SABRE-SHEATH process (Figure 1a) although alternative pulsed schemes have been developed too.^[37] P_15N over 50%^[38–40] have been demonstrated using SABRE-SHEATH for FDA-approved drug metronidazole, which can be potentially employed for hypoxia sensing applications. ¹³C polarization exceeding 1% has been demonstrated for ¹³C-labeled pyruvate.^{[41, 42]}

The Earth’s field is approximately two orders of magnitude stronger than B_LAC for PHIP and SABRE polarization of X-nucleus. In practice, multi-layered mu-metal shields represent a convenient and cost-effective approach (ca. $1k–5k) to attenuate the Earth’s field nominally by three orders of magnitude. As a result, the desired nanotesla magnetic field can be obtained. However, mu-metal shields are susceptible to magnetization and can retain a substantial degree of residual magnetization: sometimes in excess of 1,000 nT. While a small (0.1–0.2 μT) residual field can be simply treated as an offset in SABRE-SHEATH or PHIP MFC, the B_LAC calibration is required to account for residual static field of the magnetic shield. This is highly inconvenient at best, and it makes the daily equipment operation non-trivial.

Here, we report on simple circuitry to perform fast (2 seconds) automated degaussing procedure of mu-metal shields of up to 27 in. (~0.69 meter) of height with reproducible residual magnetization less than 20 nT. The degaussing process applies alternating electric current with adiabatic decay achieved by thermistors. We employ the process of ¹⁵N SABRE-SHEATH hyperpolarization of [¹⁵N₃]metronidazole as a test bed to demonstrate the utility of the reported electronics circuit.

Materials and Methods

The circuit (Figure 2a) employs 120–240 VAC power mains as a source of alternating current. Three thermistors (9 Ω each) are placed in parallel to boost the current carrying capacity of the circuit during the charging phase. When the switch is activated, the “charging” phase is initiated. During this phase, the current increases causing the thermistors to warm up, thus resulting in an increase of their resistance by ~4–5 orders of magnitude. As a result, in a few hundred milliseconds, the current begins to decay adiabatically due to increasing resistance of the thermistors. The current decay is completed in less than 2 seconds, i.e. the thermistors ensure the mains power is no longer supplied after this relatively short period of time. In practice, we employed a push-button switch, which is un-pressed after approximately 1 second. A diode is added in series with resistors to ensure the flow of power from the mains to the degaussing inductor is unidirectional. The diode output is connected to capacitor (470 μF) and degaussing inductor, see Figure 2. In practice, we have successfully tested a wide range of degaussing inductors made in house (40.4 mH) and also supplied by the vendor for the following shields ZG-203 (4.8 mH), ZG-206 (18.1 mH), ZG-209 (105.8 mH), Magnetic Shield Corp., Bensenville, IL USA. We have also added a LED indicator to inform the user that the circuit was truly energized (light ON) and that the process was completed (dimming light). An ~0.7 kΩ resistor is added in series to protect the LED and to also modulate the decay rate of the electromagnetic field in the LC circuit. The complete list of part numbers, manufacturers’, vendors’ information, and technical drawings can be found in Supporting Information (SI). The cost of construction was $73 – note that some components were purchased in bulk.

Figure 2.

a) Schematics of automated degaussing circuit; b) photograph of the assembled circuit in plastic electric enclosure.

¹⁵N SABRE-SHEATH experiments were performed as described previously.^[43–45] Briefly, Ir-IMes pre-catalyst^[46] and [¹⁵N₃]metronidazole were dissolved in CD₃OD. The prepared 0.6 mL solution contained approximately 2 mM pre-catalyst and 40 mM [¹⁵N₃]metronidazole. The solution was placed in an economy 5 mm NMR tube jacketed with 0.25 in. (~6.35 mm) OD Teflon extension. The solution was then purged with ultra-high purity argon gas for approximately 2 minutes before connecting it to our p-H₂ bubbling setup via Teflon extension described in Figure 3 and Figure S1. Once the tube was connected to the manifold, the catalyst was activated for approximately 1 h using 20 standard cubic centimetres per minute (sccm) flow rate of p-H₂ (~98%^[47]) at 100 PSI (~690 kPa) overpressure. After catalyst activation, the formation of the polarization transfer complex (PTC) allows for efficient polarization transfer of nuclear spin polarization from p-H₂-derived hydrides to ¹⁵N nuclei in [¹⁵N₃]metronidazole. The details of spin-relayed polarization transfer to all three ¹⁵N sites are thoroughly reviewed elsewhere.^[43–45] For SABRE-SHEATH hyperpolarization, we have employed 70 sccm flow rate for p-H2 bubbling and ZG-203 shield equipped with degaussing solenoid coil. The coil was connected to the degaussing circuit shown in Figure 2, and degaussing was performed using 120 VAC mains. The degaussing circuit was then disconnected and the RF solenoid coil was connected to 5 VDC power supply and current attenuation resistor bank. A residual field of less than 20 nT was measured repeatedly by a three-axis fluxgate magnetometer (Bartington Instruments, Oxford, U.K.) with 10 nT resolution.

Figure 3.

Schematic of experimental setup.

For ¹⁵N SABRE-SHEATH experiments, p-H₂ was bubbled in the shield at B_LAC (created by the RF solenoid inside the shield) for ~1 min at room temperature, 70 sccm and 100 PSI (690 kPa) overpressure. Next, p-H₂ flow was ceased via opening the bypass valve, and the sample was quickly transferred for ¹⁵N detection in 1.4 T bench-top NMR spectrometer (Nanalysis, Canada). The total delay from p-H₂ cessation to ¹⁵N NMR acquisition was less than 5 seconds. ¹⁵N signal enhancement and polarization levels were computed by employing external signal reference (12.4 M [¹⁵N]pyridine, Figure 4b) as described in detail previously.^[43]

Figure 4.

a) Simultaneous chemical exchange of p-H₂ and [¹⁵N₃]metronidazole (MNZ) on activated Ir-IMes catalyst enables spontaneous polarization transfer from p-H₂-derived hydrides to ¹⁵N₃ nucleus of [¹⁵N₃]metronidazole followed by the spin-relayed spontaneous polarization transfer from ¹⁵N₃ to ¹⁵N₁ to ¹⁵NO₂ sites; b) ¹⁵N NMR spectroscopy of [¹⁵N₃]metronidazole hyperpolarized via SABRE-SHEATH and thermally polarized [¹⁵N]pyridine employed as a signal reference using 1.4 T bench-top NMR spectrometer; c) Magnetic field dependence of ¹⁵N signal of HP [¹⁵N₃]metronidazole on the residual magnetic field inside the shield after magnetic shield degaussing to less than 20 nT residual magnetic field.

Results and Discussion

The described above circuit was successfully employed for degaussing three different kinds of magnetic shields (see Methods section) ranging in size from 3” (~7.6 cm) inner diameter (ID) and 9” (~22.9 cm) in height to 9” (~22.9 cm) ID and 27” (~0.69 m) in height using four different degaussing inductor coil configurations. These shields cover a wide range of scenarios for PHIP experiments and certainly provide sufficient volume for clinical-scale production of HP contrast agents.^[48] In all cases, the mu-metal shields were degaussed to ≤ 20 nT residual magnetic field, even when the shields were strongly magnetized (>2 μT residual field) prior to degaussing. The degaussing procedure was reproducible – test-retest reproducibility with back-to-back degaussing events (N≥3) spaced by at least 20 mins with all three magnetic shields.

Note that the thermistors warm up during the degaussing process, and therefore require a sufficiently long (20 mins or more) cooling time. In cases, when such long waiting delays is not acceptable, we employed to alternatives: (i) construction of replica circuits and (ii) the use of air cooling to reduce the circuit recovery time to less than 5 min.^[49] For example, we have employed one degaussing circuit with continuous air cooling to repeatedly degauss the 27-in. (~0.69-meter) shield after in-situ ¹³C NMR detection performed at B₀ = 7.8 mT.^[49] In this study, MFC was employed to produce ¹³C-hyperpolarized ethyl [1-¹³C]acetate via hydrogenative PHIP approach. We therefore anticipate the circuit recovery time not to be an issue in most envisioned applications.

The utility of the reported circuit to degauss the magnetic shield to less than 20 nT is demonstrated here for application in ¹⁵N SABRE-SHEATH studies. [¹⁵N₃]metronidazole was successfully hyperpolarized using spin-relayed SABRE-SHEATH,^[44] Figure 4a—Figure 4b shows ¹⁵N spectrum of HP [¹⁵N₃]metronidazole recorded using 1.4 T bench-top NMR spectrometer. The ¹⁵N signal dependence on the applied magnetic field (by the solenoid coil placed inside the shield, Figure 3b), clearly shows a maximum at ~0.6 μT (Figure 4c), and ¹⁵N signal reduction to nearly zero, when no additional field was applied (i.e., the residual in-shield field was < 20 nT). This result is in sharp contrast with the previously published study performed without a precise magnetometer and the reported degaussing circuitry: as a result, the residual magnetic field was likely 0.2 μT resulting in two artifacts: (i) apparent maximum of ¹⁵N polarization at ~0.4 μT (vs. ~0.6 μT in Figure 4c) applied via solenoid magnetic field, and (ii) ¹⁵N signal phase shift (and thus the null point) at ~0.2 μT.^[44] Such a residual magnetic field retained by the shield is therefore highly detrimental to SABRE-SHEATH experiments, because the actual “dialled” magnetic field of the in-shield solenoid (Figure 3) is added on top of the residual magnetic field of the shield, and may lead to systematic experimental biases. The described circuit in this Application paper mitigates these experimental challenges by conveniently fast and robust elimination of the residual in-shield field to below 20 nT.

Conclusion

In summary, we reported on robust and inexpensive ($73 cost of parts) circuitry for fast (~2 s) degaussing of mu-metal magnetic shields for their application with hyperpolarization techniques based on parahydrogen (PHIP and SABRE). Less than 20 nT residual field was reproducibly achieved with mu-metal shields of various sizes. The utility of degaussing was demonstrated by measuring the magnetic field profile in SABRE-SHEATH experiments with ¹⁵N hyperpolarization studies of [¹⁵N₃]metronidazole. The simple circuitry presented here may be of practical use for those working and using PHIP and/or SABRE hyperpolarization techniques.