Dissolution DNP-NMR spectroscopy using galvinoxyl as a polarizing agent

Abstract

The goal of this work was to test feasibility of using galvinoxyl (2,6-di-tert-butyl-α-(3,5- di-tert-butyl-4-oxo-2,5-cyclohexadien-1-ylidene)-p-tolyloxy) as a polarizing agent for dissolution dynamic nuclear polarization (DNP) NMR spectroscopy. We have found that galvinoxyl is reasonably soluble in ethyl acetate, chloroform, or acetone and the solutions formed good glasses when mixed together or with other solvents such as dimethyl sulfoxide. W-band electron spin resonance (ESR) measurements revealed that galvinoxyl has an ESR linewidth D intermediate between that of carbon-centered free radical trityl OX063 and the nitroxide-based 4-oxo-TEMPO, thus the DNP with galvinoxyl for nuclei with low gyromagnetic ratio γ such as ¹³C and ¹⁵N is expected to proceed predominantly via the thermal mixing process. The optimum radical concentration that would afford the highest ¹³C nuclear polarization (approximately 6% for [1-¹³C]ethyl acetate) at 3.35 T and 1.4 K was found to be around 40 mM. After dissolution, large liquid-state NMR enhancements were achieved for a number of ¹³C and ¹⁵N compounds with long spin-lattice relaxation time T₁. In addition, the hydrophobic galvinoxyl free radical can be easily filtered out from the dissolution liquid when water is used as the solvent. These results indicate that galvinoxyl can be considered as an easily available free radical polarizing agent for routine dissolution DNP-NMR spectroscopy.

1. INTRODUCTION

Nuclear magnetic resonance (NMR) is one of the most widely used analytical and structural elucidation techniques in chemistry and materials science. It is also the underlying principle of magnetic resonance imaging (MRI), an important clinical imaging modality. NMR, however, is a relatively insensitive technique compared to other spectroscopic methods such as electron spin resonance (ESR), optical, or mass spectroscopy [1]. This inherent insensitivity emanates from the relatively small magnetic moments of the nuclei leading to low nuclear polarization P, a parameter which determines the strength of the NMR signal [2,3].

The problem of low sensitivity in liquid-state NMR spectroscopy has been recently addressed by the invention of the dissolution dynamic nuclear polarization (DNP) method [4]. DNP, a technique used in the production of polarized targets for nuclear and particle physics experiments since the 1960s, amplifies the NMR signal via the microwave-induced transfer of the high electron thermal polarization to the target nuclear spins at low temperature and moderate magnetic field [5–7]. The incorporation of a dissolution device in the DNP polarizer that rapidly dissolves the frozen polarized samples (e.g. compounds containing NMR-active nuclei dissolved in glass formers and doped with paramagnetic electrons) using pressurized superheated water or other solvents has allowed the production of highly polarized solutions at room temperature [4]. As a result, the liquid state NMR signal is enhanced several-thousand-fold thereby facilitating the high sensitivity liquid-state MR spectroscopy and imaging of nuclei with low gyromagnetic ratio γ, in particular, ¹³C and ¹⁵N, in chemistry [8] and biomedical applications [9–11].

One of the crucial parameters to achieve high nuclear polarization levels in DNP is the choice of free radicals used as polarizing agents [12,13]. The ESR properties of the free radical (e.g. the ESR linewidth D) play a central role in the DNP polarization transfer process [5,12,13]. Currently, the commonly used free radical polarizing agents for dissolution DNP-NMR include the carbon-centered trityl OX063 radical as well as the nitroxide-based TEMPO and very recently the carbon-centered BDPA [14] free radical. Other free radicals that have been considered for dissolution DNP-NMR spectroscopy include the perchlorinated trityls, trityl biradicals, and other derivatives [15–17]. In this work, we have investigated the feasibility of a well-known stable free radical, 2,6-di-tert-butyl-α-(3,5-di-tert-butyl-4-oxo-2,5- cyclohexadien-1-ylidene)-p-tolyloxy or galvinoxyl as a polarizing agent for dissolution DNP. Galvinoxyl (see Fig. 1a), first synthesized by Galvin Coppinger in 1957 [18], has been mainly used as scavenging free radical to test the potency of antioxidants in food and pharmaceuticals [19,20]. This free radical has never been tried for dissolution DNP-NMR spectroscopy primarily because it is insoluble in water. However, we have found that galvinoxyl is reasonably soluble in ethyl acetate, chloroform, or acetone and these solvents form a good glassing matrix when mixed together or with other solvents such as dimethyl sulfoxide. The goal of this research was to investigate the utility of galvinoxyl as a polarizing agent for routine dissolution DNP-NMR spectroscopy in chemistry.

Fig. 1.

(a) Structure of galvinoxyl free radical. (b) Field-swept ESR spectrum of 40 mM galvinoxyl in 1:1 (v/v) CHCl₃:DMSO glassing matrix measured at 100 K in a W-band ESR spectrometer.

2. EXPERIMENTAL METHODS

Materials

The galvinoxyl and 4-oxo-TEMPO (4-oxo-2,2,6,6-tetramethylpiperidine-1-oxyl) free radicals, ¹³C (ethyl acetate, urea, methyl pyruvate, sodium pyruvate, diethyl-1,3-acetone-dicarboxylate) and ¹⁵N (nitrobenzene, acetonitrile, and pyridine) compounds, glassing agents, and solvents used in this work were obtained from Sigma-Aldrich (St. Louis, Missouri) and were used without further purification. The free radical trityl OX063 (tris{8-carboxyl-2,2,6,6-benzo(1,2-d:5-d)-bis(1,3)dithiole-4-yl}methyl sodium salt) was obtained from Oxford Instruments (Tubney Woods, U.K.).

ESR measurements

The ESR measurements of galvinoxyl samples at W-band (3.35 T, 5 K base temperature) were done in a commercial Bruker E680 ESR spectrometer (Bruker Biospin, Billerica, Massachusetts) at the National High Magnetic Field Laboratory (NHMFL) in Tallahassee, Florida. The galvinoxyl samples (40 mM in 1:1 (v/v) CHCl₃:DMSO) for ESR were loaded into 0.55 mm OD quartz capillary tubes. For comparison, the ESR spectra of trityl OX063 (15 mM in 1:1 (v/v) glycerol:water) and 4-oxo-TEMPO (15 mM in 1:1 (v/v) glycerol:water) were also measured in the W-band at 100 K (see the Supplementary Information). The W-band ESR spectra at 100K were measured by echo-detected field sweep by using a two-pulse echo 90°-τ-180°.

Dynamic nuclear polarization of compounds

The samples were polarized in the HyperSense commercial polarizer (Oxford Instruments, Tubney Woods, UK) at 3.35 T and 1.4 K with a 100 mW microwave source as described previously [12]. The ¹³C microwave DNP spectra and polarization buildup curves were done with a built-in automated program in the HyperSense NMR spectrometer.

Hyperpolarized ¹³C compounds

a) [1-¹³C]ethyl acetate. A 100 µL solution containing 3.37 mg galvinoxyl in CHCl₃ was prepared and then mixed with an equal volume of [1-¹³C]ethyl acetate. The final concentration of galvinoxyl in the sample was 40 mM. 50 µL aliquots were polarized in the HyperSense for 1 hour. 4 mL methanol was used as the dissolution solvent. b) [¹³C]urea. A 200 µL solution contained 25 mg [¹³C]urea solution was prepared in 1:1 (v/v) DMSO:CHCl₃ and then doped with 3.4 mg galvinoxyl. 40 µL aliquots were polarized in the HyperSense at 3.35 T and 1.4 K for 2 hours. 4 mL water was used as the dissolution solvent. c)¹³C-natural abundance methyl pyruvate. A 200 µL solution containing 6.75 mg galvinoxyl in CHCl₃ was prepared and then mixed with equal volume of methyl pyruvate. 100 µL aliquots were polarized in the HyperSense for 2 hours. 4 mL methanol was used as the dissolution solvent. d)¹³C-natural abundance diethyl-1,3-acetone-dicarboxylate. A 200 µL solution containing 6.75 mg galvinoxyl in CHCl₃ was prepared and then mixed with equal volume of diethyl-1,3-acetone-dicarboxylate. 100 µL aliquots were polarized in the HyperSense polarizer for 2 hours. 4 mL methanol was used as the dissolution solvent.

Hyperpolarized ¹⁵N compounds

a) ¹⁵N-enriched nitrobenzene. A 100 µL solution containing 3.37 mg galvinoxyl in CHCl₃ was prepared and then mixed with an equal volume of ¹⁵N-nitrobenzene. 50 µL aliquots were polarized in the HyperSense at 3.35 T and 1.4 K for 3 hours. 4 mL methanol was used as the dissolution solvent. b)¹⁵N-natural abundance acetonitrile. A 200 µL solution containing 6.75 mg galvinoxyl in CHCl₃ was prepared and then mixed with equal volume of acetonitrile. 100 µL aliquots were polarized in the HyperSense at 3.35 T and 1.4 K for 3 hours. c)¹⁵N-natural abundance pyridine. A 200 µL solution containing 6.75 mg galvinoxyl in CHCl₃ was prepared and then mixed with equal volume of pyridine. 100 µL aliquots were polarized in the HyperSense polarizer at 3.35 T and 1.4 K for 3 hours. 4 mL water was used as the dissolution solvent.

NMR signal enhancement ε and spin-lattice T₁ relaxation measurements in the liquid-state

The liquid-state NMR signal enhancement ε and T₁ measurements were all done at 298 K in a 9.4 T Varian VNMRS high resolution spectrometer (Agilent Technologies, Santa Clara, California). The dissolution liquid from the polarizer was collected directly in a 10-mm NMR tube inside a 9.4 T magnet via a Teflon tube with a dissolution transfer time of 8 s. Thus, the NMR enhancements ε were measured 8 s after dissolution of the hyperpolarized ¹³C or ¹⁵N compounds. The liquid-state ε was calculated by taking the ratio of the integrated NMR area of the sample in the hyperpolarized state over the thermal NMR signal [12]. In the case of ¹⁵N natural abundance compounds, neat ¹⁵N natural abundance nitrobenzene was used as the thermal ¹⁵N NMR signal for enhancement calculations. The NMR enhancement measurements were done in triplicate (N=3) and the averages were calculated along with the standard deviation. For liquid-state T₁ measurements, 1 mL of the 4 mL dissolution liquid collected in the beaker was pipetted into a 10-mm NMR tube. The sample tube was placed higher than normal to make sure that the hyperpolarized liquid is enclosed by the rf coil. The decay of the hyperpolarized NMR signal was monitored by applying a small flip angle θ_flip every time interval TR. The T₁ values were calculated using an equation [21,22] which accounts for the loss of nuclear magnetization due to relaxation and rf excitation (see also the Supplementary Information). The analyses of the NMR data were done using the software ACD Labs version 12 (Advanced Chemistry Development Inc., Toronto, Canada) and Igor Pro version 6 (Wavemetrics Inc., Portland, Oregon).

Filtration and UV-Vis spectroscopy

The hydrophobic galvinoxyl free radical naturally precipitated out in the dissolution liquid when water is used as the solvent. In separate experiments, the free radical was rapidly removed from the hyperpolarized solution by simple mechanical filtration using a 0.2-micron syringe filter. The absence of galvinoxyl in the filtered aqueous dissolution liquid was confirmed by UVVis spectroscopy (see the Supplementary Information) using a Shimadzu UV-3600 spectrophotometer (Shimadzu Scientific Instruments, Japan).

3. RESULTS AND DISCUSSION

3.1 ESR measurements

As mentioned previously, the ESR properties of the free radical polarizing agent play an important role in achieving the maximum nuclear polarization of the target nuclei in DNP [5–7, 12, 13]. In this regard, we have performed ESR measurement on the optimum DNP concentration of galvinoxyl in the W-band which operates at approximately the same magnetic field as the HyperSense polarizer (3.35 T). Fig. 1b shows the ESR spectrum of galvinoxyl (40 mM) in 1:1 (v/v) CHCl₃:DMSO glassing matrix at 100 K in the W-band. The shape of the W-band ESR spectrum at 100 K is expected to be very similar to that at 1.4 K where DNP is performed. The principal values of the g-tensor of galvinoxyl measured at 100 K in the W-band are g_xx=2.00625, g_yy=2.00445, and g_zz=2.00237. As seen from Fig. 1b, it is difficult to determine the ESR linewidth due to the inhomogeneously broadened lineshape, which is dominated by g-anisotropy arising mainly from the spin-orbit interaction with the oxygen atom [23]. In this light, we define the linewidth D in this work as the distance from base to base of the ESR spectrum. Based on this criterion, the ESR D of galvinoxyl (approximately 250 MHz) is estimated to be intermediate between the carbon-centered trityl OX063 [24] (115 MHz; see the Supplementary Information) and the nitroxide-based 4-oxo-TEMPO ESR linewidths [25] (D=465 MHz; see the Supplementary Information).

3.2 DNP of galvinoxyl-doped [1-¹³C]ethyl acetate in the solid state

Mounting evidence [12, 13, 26, 27] have shown that DNP of low-γ nuclei such as ¹³C, ¹⁵N, and ⁸⁹Y with trityl OX063 and TEMPO at moderate fields (e.g. 3.35 T) and at temperatures close to 1 K operate predominantly via the thermal mixing mechanism. Since the ESR linewidth of galvinoxyl is significantly larger than the Larmor frequencies of ¹³C (35.96 MHz) or ¹⁵N (14.45 MHz) in the HyperSense polarizer, the expected DNP mechanism for such low-γ nuclei is the thermal mixing process [5–7, 12, 13]. As described by the spin temperature or the Borghini model of DNP [28], off-center microwave irradiation by a certain optimum offset frequency Δ (upfield or downfield from the center of the ESR lineshape) will lead to minimum spin temperature of the electron dipolar system, which consequently cools the nuclear Zeeman system [5–7, 12, 13]. Fig. 2a shows the location of the optimum polarization peaks in the ¹³C microwave DNP spectrum (plot of ¹³C NMR signal enhancements as a function of microwave irradiation frequency) of 1:1 (v/v) [1-¹³C]ethyl acetate:CHCl₃ sample doped with 40 mM galvinoxyl. The separation between the positive P(+) and negative P(−) polarization peaks is about 150 MHz. In comparison, the ¹³C microwave DNP spectrum of a [1-¹³C]pyruvate sample doped with the carbon-centered trityl OX063 free radical also shown in Fig. 2a has a separation |P(+)-P(−)| that is roughly half that of galvinoxyl. These values reflect the size of their ESR linewidths. A ¹³C polarization close to 10 % was obtained for 1.4 M [1-¹³C]sodium pyruvate sample in 1:1 (v/v) glycerol:water glassing matrix doped with 15 mM trityl OX063 at 3.35 T and 1.4 K. On the other hand, galvinoxyl (D=250 MHz) has a narrower ESR spectrum than 4-oxo TEMPO (D=465 MHz; see W band ESR data in the Supplementary Information). ¹³C polarization values in the range of 5–10% were reported elsewhere [25, 26, 29–31] for ¹³C samples doped with TEMPO radicals, whereas under similar DNP conditions we have obtained 5.9±0.3% ¹³C polarization for [1-¹³C]ethyl acetate sample doped with galvinoxyl as shown in Fig. 2. It should be noted that parameters such as the type of glassing matrix, lattice temperature, microwave power, and other factors that influence DNP performance [12, 13, 25] may be different in some polarizers, which may preclude direct comparison of polarization values obtained in different instruments. However, DNP of ¹³C sodium pyruvate sample doped with 4-oxo-TEMPO performed in the same polarizer yielded ~5.5% [30], thus, we can say that galvinoxyl DNP performance is at least comparable with the results obtained with the nitroxide-based 4-oxo-TEMPO radical.

Fig. 2.

(a) ¹³C microwave DNP spectrum of 1:1 (v/v) [1-¹³C]ethyl acetate:CHCl₃ doped with 40 mM galvinoxyl (solid circles) at 3.35 T and 1.4 K. The up [P(+)=94.10 GHz] and down [P(−)=94.255 GHz] arrows represent the positive and negative polarization peaks, respectively. The ¹³C microwave DNP spectrum of 1.4 M [1-¹³C]sodium pyruvate in 1:1 (v/v) glycerol:water doped with 15 mM trityl OX063 (open squares) is also shown for comparison. (b) Polarization buildup time constants τ_bu taken at different microwave irradiation frequencies. The error bars are due to fits with a mono-exponential buildup equation. (c) ¹³C polarization buildup curves of 1:1 (v/v) [1-¹³C]ethyl acetate:CHCl₃ doped with different concentrations of galvinoxyl. These data were taken at 3.35 T and 1.4 K using a 100 mW microwave source. The dashed lines are fits to a mono-exponential buildup equation. (d) The corresponding polarization buildup time constants extracted by fitting the data in (c). The error bars are due to fits with a mono-exponential buildup equation.

Another note is that although trityl OX063 is expected to be a better polarizing agent in DNP of low-γ nuclei such as ¹³C via thermal mixing, it provides low DNP enhancements for high-γ nuclei such as protons using the same polarizer setup as shown in a previous work [32]. Thus, there are some cases in which free radicals with large D are advantageous. For instance, TEMPO was used to simultaneously polarize ¹H and ¹³C spins in ¹H-¹³C DNP cross polarization (CP) transfer experiments [31, 33, 34] where ¹H, owing to its larger γ, yielded higher polarization and consequently led to significant improvements of ¹³C DNP polarization after transfer at shorter microwave irradiation times. Our present DNP hardware is currently not equipped for such experiments, but it should be noted that the addition of galvinoxyl to the list of polarizing agents for dissolution DNP could offer such potential advantages.

Fig. 2b displays the polarization buildup time constants τ_bu across the ¹³C microwave DNP spectrum. The τ_bu values were extracted by fitting the corresponding polarization buildup curves at each microwave frequency with a singleexponential buildup equation P(t)=P₀[1-exp(-t/τ_bu)] where P₀ is the maximum polarization achieved and t is the microwave irradiation time. It is interesting to note that τ_bu remains roughly constant in between the polarization peaks P(+) and P(−), while it increases as the microwave irradiation frequency is set outside this range. The large values of τ_bu further out in the DNP spectrum was also previously reported in TEMPO-doped ¹³C samples at 94 GHz and 140 GHz [25] where the experimental plot of DNP spectrum is narrower than the ones predicted by the Borghini model. The increase in τ_bu’s on the wings of the DNP spectrum is attributed to the partial saturation of the ESR lineshape as the irradiation frequency is set further away from the center of the ESR lineshape [35]. Fig. 2c shows the polarization buildup curves of 1:1 (v/v) [1-¹³C]ethyl acetate:CHCl₃ doped with different concentrations of galvinoxyl free radical. The maximum ¹³C polarization approaching 6% is reached at a concentration of 40 mM. The faster ¹³C polarization buildup (shorter τ_bu) with increasing galvinoxyl concentration shown in Fig. 2d is attributed to more paramagnetic centers polarizing the ¹³C nuclear spins.

3.3 Hyperpolarized ¹³C compounds in the liquid-state

Dissolution DNP has been very successful in amplifying the NMR signals of ¹³C (spin-1/2, γ=10.705 MHz/T) spins for NMR spectroscopy and imaging where several thousand-fold signal enhancements are routinely achieved using trityl OX063 as the polarizing agent [4, 9–11]. Here we have demonstrated that large ¹³C NMR signal enhancements could also be obtained with galvinoxyl free radical. Fig. 3 shows the hyperpolarized (HP) and thermal (TH) ¹³C NMR signals of four model compounds that contain at least one ¹³C located in carbonyl or carboxyl functionality. Such carbon nuclei are termed as “long T₁ carbons” since their spin-lattice relaxation time is usually very long (over 30 s) due to the absence of directly-bonded H. Although long T₁ values are considered a nuisance in conventional NMR, they are very useful in experiments with hyperpolarized compounds since the hyperpolarized state decays with T₁ relaxation. The HP spectrum of ¹³C enriched [1-¹³C]ethyl acetate was measured 8 s after the dissolution and transfer of the frozen polarized sample (50 µL 1:1 [1-¹³C]ethyl acetate:CHCl₃ doped with 40 mM galvinoxyl) with 4 mL methanol. Owing to the relatively long T₁ (45 s) of ¹³C in the C1 position of ethyl acetate, a substantial NMR signal enhancement ε=5500-fold in the liquid state was obtained relative to the ¹³C TH NMR signal. Fig. 3b shows that the hydrophobic galvinoxyl free radical could also be used in the dissolution DNP of water-soluble compounds such as [¹³C]urea where the ¹³C NMR signal was enhanced 2800-fold in the liquid state. In this case, water was used as the dissolution solvent, and the water-insoluble galvinoxyl free radical naturally precipitated out from the dissolution liquid and could easily be removed by a simple filtration using a 0.2-micron syringe filter. UV-Vis spectroscopy confirmed the absence of the free radical in the solution after filtration (see results in the Supplementary Information). Fig. 3c and Fig. 3d illustrate that hyperpolarization with galvinoxyl is not limited to ¹³C-labeled compounds but could facilitate the structural characterization of compounds containing ¹³C at natural abundance (1.11 %). The HP and TH NMR spectra of unlabeled methyl pyruvate and diethyl-1,3-acetone-dicarboxylate are shown in Fig. 4b and Fig. 4c, respectively, where the ¹³C spins present in the carbonyl and carboxyl groups were enhanced about 3000–4000-fold. Galvinoxyl, however, was not compatible with [1-¹³C]pyruvic acid; it appears to react with pyruvic acid as indicated by a color change from green to red.

Fig. 3.

(a) Hyperpolarized (HP; θ_flip=0.5°) and thermal (TH; θ_flip=90°, 1 scan) ¹³C NMR spectra of [1-¹³C]ethyl acetate (64 mM in MeOH at 298 K and 9.4 T). The average (N=3) ¹³C NMR signal enhancement ε is 5500±800 and the liquid-state T₁ is 45 s. (b) HP (θ_flip=2°) and TH (θ_flip=90°, 1 scan) NMR spectra of [¹³C]urea (23 mM in H₂O at 298 K and 9.4 T); average ε=2800±450 (T₁=45 s). (c) HP (θ_flip=90°) and TH (θ_flip=90°, 16 scans) NMR spectra of ¹³C natural abundance methyl pyruvate (138 mM in MeOH at 298 K and 9.4 T); average C1 ε=4300±930 (T₁=42 s), C3 ε=4100±710 (T₁=44 s). (d) HP (θ_flip=90°) and TH (θ_flip=90°, 16 scans) NMR spectra of ¹³C natural abundance diethyl-1,3-acetone-dicarboxylate (68 mM in MeOH at 298 K and 9.4 T); average C1 ε=3400±820 (T₁=32 s), C3 ε=3700±760 (T₁=34 s).

Fig. 4.

(a) Hyperpolarized (HP; θ_flip=5°) and thermal (TH; θ_flip=90°, 1 scan) NMR signals of ¹⁵N-enriched nitrobenzene (61 mM in MeOH) at 298 K and 9.4 T. The average ¹⁵N NMR signal enhancement (N=3) is 4800±540 and the liquid-state T₁ is 135 s. (b) HP (θ_flip=90°) and TH (θ_flip=90°, 1 scan) NMR spectra of ¹⁵N natural abundance acetonitrile (239 mM in MeOH) at 298 K and 9.4 T.; average ε=4200±730 (T₁=110 s). (c) HP (θ_flip=90°) and TH (θ_flip=90°, 1 scan) NMR spectra of ¹⁵N natural abundance pyridine (155 mM in H₂O) at 298 K and 9.4 T.; average ε=3300±770 (T₁=58 s). The coupling of ¹⁵N to protons is also visible [J(¹H-¹⁵N)=9.7 Hz]. The ¹⁵N NMR spectra were referenced to nitrobenzene (¹⁵N δ=372 ppm with respect to anhydrous ammonia).

3.4 Hyperpolarized ¹⁵N compounds in the liquid-state

In addition to ¹³C compounds, we have also polarized a number of ¹⁵N (spin-1/2, γ=4.3143 MHz/T) substrates using galvinoxyl as the DNP polarizing agent. It is well established in previous works [12, 13, 27, 36] that in the thermal mixing regime, the locations of the optimum polarization peaks should be approximately the same for all low-γ nuclei because the shape of the microwave DNP spectrum depends mainly on the properties of the free radical polarizing agent. Thus, the ¹⁵N compounds were polarized at the optimum positive polarization peak (94.10 GHz) determined for ¹³C compounds. The ¹⁵N NMR spectra in the liquid-state were referenced to nitrobenzene (¹⁵N δ=372 ppm with respect to anhydrous ammonia) [37]. Fig. 4 shows the HP and TH ¹⁵N NMR signals of a) ¹⁵N-enriched nitrobenzene b) ¹⁵N natural abundance acetonitrile, and c) ¹⁵N natural abundance pyridine measured at 298 K in a 9.4 T high resolution magnet. The T₁ of ¹⁵N is generally longer than that of ¹³C owing to its lower gyromagnetic ratio. Despite the low number of naturally abundant ¹⁵N spins (natural abundance is only 0.36 %), we were able to detect the ¹⁵N NMR spectra of acetonitrile and pyridine with 1 scan as shown in Fig. 4b and Fig. 4c. The thermal equilibrium NMR signal of neat ¹⁵N natural abundance nitrobenzene was used for the liquid-state NMR enhancement calculations of ¹⁵N natural abundance compounds. The average ¹⁵N NMR signal enhancement ε of acetonitrile was 4200-fold with a liquid-state T₁ of 110 s while the natural abundance pyridine sample yielded an average enhancement of 3300-fold. The T₁ of ¹⁵N pyridine was slightly shorter (58 s) due to the proximity of ¹⁵N to ¹H nuclei in the molecule.

4. CONCLUSION

In summary, we have demonstrated that the commercially available galvinoxyl free radical can be used as a polarizing agent for dissolution DNP-NMR characterization of a variety of compounds. The optimum concentration of galvinoxyl for DNP at 3.35 T and 1.4 K was found to be around 40 mM. Several-thousand-fold NMR signal enhancements were achieved in the liquid-state for ¹³C- and ¹⁵N-enriched compounds as well as for naturally abundant samples. Dissolution DNP with galvinoxyl is currently restricted to the glassing solvents in which this radical is soluble (e.g. CHCl₃, ethyl acetate, acetone, or DMSO). The water-insoluble galvinoxyl can readily be filtered off from the dissolution liquid by mechanical filtration when water is used as the dissolution solvent. Certain optimization steps such as Gd³⁺ doping [12,30], solvent and sample deuteration [26,38], and lower operating temperatures of the polarizer [39] may be explored in the future to further increase the NMR signal enhancements of DNP samples doped with galvinoxyl.

Highlights.

• Galvinoxyl was tested in ¹³C and ¹⁵N dissolution dynamic nuclear polarization (DNP) NMR.

• Several-thousand-fold ¹³C and ¹⁵N NMR signal enhancements were achieved.

• The water-insoluble galvinoxyl can be filtered off from aqueous dissolution liquids.