Overhauser dynamic nuclear polarization (ODNP)-enhanced two-dimensional proton NMR spectroscopy at low magnetic fields

Abstract The majority of low-field Overhauser dynamic nuclear polarization (ODNP) experiments reported so far have been 1D NMR experiments to study molecular dynamics and in particular hydration dynamics. In this work, we demonstrate the application of ODNP-enhanced 2D J-resolved (JRES) spectroscopy to improve spectral resolution beyond the limit imposed by the line broadening introduced by the paramagnetic polarizing agent. Using this approach, we are able to separate the overlapping multiplets of ethyl crotonate into a second dimension and clearly identify each chemical site individually. Crucial to these experiments is interleaved spectral referencing, a method introduced to compensate for temperature-induced field drifts over the course of the NMR acquisition. This method does not require additional hardware such as a field-frequency lock, which is especially challenging when designing compact systems.


ODNP Spectrometer
The ODNP spectrometer used for experiments is shown in Figure S1. Key components of the spectrometer are numbered in the figure: (1) Bridge12 microwave power source (MPS) (2) Kea spectrometer (3) RF amplifier (4) shim and B0 power supplies (5) incubator (6) SABR magnet (7) ODNP resonator (8) remote tuning box. The compact system occupies a 24" x 40" x 32" 5 space. Further reduction in space can be achieved by reducing the size of the shim power supplies.

Linewidth Measurements 10
The NMR spectrum of water with 200 µM TEMPOL sample is shown in Figure S2. The water peak was referenced to 4.8 ppm.
The water spectrum has a full width half max (FWHM) linewidth of 2.3 Hz. (0.16 ppm). The enhancement at 33 dBm microwave power was determined to be -19. The broad baseline features indicate higher order inhomogeneities, which would require additional higher order shims to correct. Lower concentrations of polarizing agent produce much narrower lines but result in lower enhancements. 15 Figure S1: ODNP Spectrometer used for measurements. Refer to text for numbered components.
The linewidth for the ethyl crotonate spectra were obtained by fitting the methyl protons to the sum of four Voigt distributions. The Lorentzian linewidth ( ) and the Gaussian linewidth ( ) for all peaks were assumed to be the same. The 20 amplitude and position of each peak were free parameters. The experimental data and fits are shown in Figure S3.
Although the true number of methyl peaks is 5, the two overlapping peaks at 1.7 ppm are close in chemical shift value and are 25 fit with a single Voigt distribution. The JRES data were processed using a Lorentz-Gauss transformation because it removes the long tail of the Lorentzian distribution.

Ethyl Crotonate Phase Correction:
In addition to compensating for any magnetic field drift, a phase correction was also applied to the 2D JRES spectrum. The phase was determined by applying a zeroth order phase correction to the interleaved reference spectra. The phase for each 30 JRES slice was then corrected with the corresponding phase for the 1D spectra. The phase varies approximately linearly over the course of the experiment as shown in Figure S4. If not corrected, the phase drift would result in a frequency offset in the indirect dimension. We attribute the phase drift to acquisition delay which will result in a linear phase roll in frequency domain as the field drifts. Additionally, instrumental and environmental instabilities (e.g. temperature fluctuations in the room) can cause long term phase drift. All these effects can be corrected for by interleaved spectral referencing. 35

ODNP Enhanced JRES of Ethanol
As the second example we performed ODNP-enhanced JRES spectroscopy using a sample of neat ethanol with 10 mM TEMPOL. Analogous to the ethyl crotonate JRES spectrum, the ethanol JRES spectrum was corrected for field drift and phase 40 drift. The interleaved 1D proton spectra used for the ethanol field drift correction are shown in Figure S5 (a). The field drift was typical of a 30 minute experiment at about 60 ppm. The 4 th degree polynomial fit to the field drift is shown in Figure S5 (b) and the residual of the fit is shown in Figure S5 (c). The JRES spectrum for ethanol with 10 mM TEMPOL is shown in Figure S6 Figure S6 (b). As expected, we observe a triplet for the protons of the methyl group (A protons) at a chemical shift of 1.2 ppm and a quartet for the methylene protons at 3.5 ppm (B protons). A broad resonance is observed at a chemical shift of 5.1 ppm, which is assigned to the proton of the hydroxyl group (C protons). This peak is 50 significantly broadened due to the exchange rate of the OH protons. In the absence of exchange, this peak would be much narrower. In addition, the quartet observed at 5.1 ppm would be observed as a doublet of quartet. However, due to the proton exchange this multiplet is observed as a broadened quartet.
Features of strong coupling between the methyl and methylene groups can be clearly distinguished in the 2D JRES 55 spectrum and are located at a chemical shift of 2-3 ppm and around a J-coupling of ±15 Hz (see Figure S6 (a), dashed box).
The broad features at 4-5 ppm and a J-coupling of ±10 Hz can be attributed to strong coupling peaks of the methylene protons and the proton of the hydroxy group. Individual slices for the different proton sites and the resulting coupling patterns are shown in Figure S6 (b). Overall, all expected features can be observed in these slices. The skyline projection and 1D proton spectrum for ethanol with 10 mM TEMPOL are shown in Figure S7. The 1D proton spectra are broadened by the viscosity of the solution and polarizing agent, however, it is still possible to resolve the Jcouplings in the 1D proton spectrum. The CH3 triplet is visible at 1.2 ppm. Similarly to the JRES slice, the CH2 group is a 65 broadened quartet at 3.5 ppm. The OH peak is visible as a broad peak at 5.1 ppm. The skyline projection is not quantitative.
The CH2 and OH peaks are visible, but significantly reduced in amplitude. A quantitative analysis could be obtained by integration of the peaks. Integration would allow for accurate measurement of enhancements for each chemical site.

Additional Window Functions
To process the JRES data, a Lorentz-Gauss transformation with equal Lorentzian and Gaussian linewidths was used. Such an 75 apodization function will have minimal effect on the linewidth, but removes the long tail from the Lorentzian distribution. linewidth parameters. The apodization function for each spectrum is shown in Figure S8 (a). The corresponding spectrum is shown in Figure S8 (b). As can be seen from the spectra, there is minimal change in linewidth, however, the lines will begin to broaden as larger apodization linewidths are used. In Figure S8 (c-d), we use a matched filter for the Lorentzian linewidth of 7 Hz and vary the Gaussian linewidth. In this case, improved spectral resolution can be achieved, but only at the expense of signal to noise. 85 In addition, we show the effect of using a Hamming window for apodization in Figure S9. The ethyl crotonate linewidth obtained by using a Hamming window was 6.7 Hz (0.45 ppm). 90 Figure S9: (a) Comparison of ethyl crotonate with 10 mM TEMPONE linewidth using a 1 Hz Lorentzian apodization function (orange) and Hamming window (green). (b) Fit of methyl peaks for ethyl crotonate spectrum with Hamming window applied using the sum of 4 Voigt distributions.