Hyperpolarized water through dissolution dynamic nuclear polarization with UV-generated radicals

Abstract

In recent years, hyperpolarization of water protons via dissolution Dynamic Nuclear Polarization (dDNP) has attracted increasing interest in the magnetic resonance community. Hyperpolarized water may provide an alternative to Gd-based contrast agents for angiographic and perfusion Magnetic Resonance Imaging (MRI) examinations, and it may report on chemical and biochemical reactions and proton exchange while perfoming Nuclear Magnetic Resonance (NMR) investigations. However, hyperpolarizing water protons is challenging. The main reason is the presence of radicals, required to create the hyperpolarized nuclear spin state. Indeed, the radicals will also be the main source of relaxation during the dissolution and transfer to the NMR or MRI system. In this work, we report water magnetizations otherwise requiring a field of 10,000 T at room temperature on a sample of pure water, by employing dDNP via UV-generated, labile radicals. We demonstrate the potential of our methodology by acquiring a ¹⁵N spectrum from natural abundance urea with a single scan, after spontaneous magnetization transfer from water protons to nitrogen nuclei.

Fig. 1. dDNP setup.

a Schematic representation of the dDNP set up composed of the 6.7 T polarizer, the magnetic tunnel, and the 9.4 T NMR magnet. The vertical component of the magnetic field measured along the sample pathway described by the gray arrow in panel a is reported with magnetic tunnel in panel b and without magnetic tunnel in panel c. In a, b and c, the purple hexagon and blue circle are markers to help guide the eye. d Schematic representation of the magnetic tunnel profile. e Custom fluid path (CFP) dissolution system.

Fig. 2. X-band ESR measurements.

a Normalized X-band ESR spectra after 10 min UV-light irradiation at 77 K and b radical generation time evolution of PYR_sample (blue circles) and 2CPYR_sample (orange circles). Data points and error bars are the average and standard deviation of repeated measurements from distinct samples (n = 3), respectively. The blue and orange curves were obtained by fitting the data to a mono-exponential function. The time constant resulting from the fits are reported in the inset. Error on fits was below 5%. X-band ESR spectra and radical generation time evolution of 2CPYRd_sample can be found in Supplementary Fig. 1.

Fig. 3. Solid-state LOD-ESR and DNP measurements.

a LOD-ESR spectrum and b ¹H DNP microwave sweep spectra measured at 6.7 T and 1.15 K without microwave modulation are reported for PYR_sample (blue), 2CPYR_sample (orange), and TEMPOL_sample (yellow). The zero crossings in b have been corrected to coincide with the center of gravity (first moment of ESR spectrum) in a.

Fig. 4. ¹H DNP sweep spectra and T_1e measurements.

¹H DNP microwaves sweep spectra measured at 6.7 T and 1.15 K with (color) and without (gray) microwave frequency modulation for PYR_sample (a), 2CPYR_sample (b), and TEMPOL_sample (c). d T_1e measurements using the LOD-ESR probe for PYR_sample (blue), 2CPYR_sample (orange), and TEMPOL_sample (yellow). The experimental data was fitted (smooth curves) to the expression $S=A\left(\exp \left(-t/T_{1e}\right)-\exp \left(-t/\tau \right)\right)$, where τ represents the pickup coil time constant and A a proportionality factor. Error on fit was below 5%.

Fig. 5. Solid-state DNP build-up and liquid-state relaxation measurements.

Solid-state ¹H polarization build-up comparison at 6.7 T and 1.15 K between PYR_sample (blue circles), 2CPYR_sample (orange circles) and TEMPOL_sample (yellow circles) (a) and between 2CPYR_sample (orange circles) and 2CPYRd_sample (red circles) (c). Liquid-state relaxation comparison at 9.4 T and 313 K after DT transfer between PYR_sample (blue circles), 2CPYR_sample (orange circles) and TEMPOL_sample (yellow circles) (b) and between 2CPYR_sample (orange circles) and 2CPYRd_sample (red circles) (d). Panel e and f show the effect of increasing ¹H nuclei concentration in the liquid state. Data points and error bars are the average and standard deviation of repeated measurements from distinct samples (n = 3), respectively. All curves were obtained by fitting the data to a mono-exponential function. The different time constants resulting from the fits are reported in the insets. Error on fits was below 5%.

Fig. 6. Polarization of heteronuclei measurements.

a ¹⁵N enhancement as a function of time after injection of hyperpolarized water sample (60 beads of 2CPYR_sample, D transfer) into a 10 mm NMR tube containing 500 μL of 400 mM [¹³C,¹⁵N₂]urea, acquired at 9.4 T and 313 K. The inset shows the comparison between the DNP and thermal equilibrium ¹⁵N spectra. b DNP enhanced (SNR = 80) and thermal equilibrium natural abundance ¹⁵N spectra of 500 μL of 400 mM urea acquired with a single 90° scan 40 s after the injection of the hyperpolarized water sample (60 beads of 2CPYR_sample, D transfer). In a and b, the ¹⁵N resonances at 76.2 and 75.7 ppm are due to the ¹⁵N-¹³C coupling (20 Hz coupling).