Heteronuclear cross-relaxation effect modulated by the dynamics of N-functional groups in the solid state under 15 N DP-MAS DNP Journal of Magnetic Resonance

In a typical magic-angle spinning (MAS) dynamic nuclear polarization (DNP) nuclear magnetic resonance (NMR) experiment, several mechanisms are simultaneously involved when transferring much larger polarization of electron spins to NMR active nuclei of interest. Recently, speciﬁc cross-relaxation enhancement by active motions under DNP (SCREAM-DNP) [Daube et al. JACS 2016] has been reported as one of these mechanisms. Thereby 13 C enhancement with inverted sign was observed in a direct polar- ization (DP) MAS DNP experiment, caused by reorientation dynamics of methyl that was not frozen out at 100 K. Here, we report on the spontaneous polarization transfer from hyperpolarized 1 H to both primary amine and ammonium nitrogens, resulting in an additional positive signal enhancement in the 15 N NMR spectra during 15 N DP-MAS DNP. The cross-relaxation induced signal enhancement (CRE) for 15 N is of opposite sign compared to that observed for 13 C due to the negative sign of the gyromagnetic ratio of 15 N. The inﬂuence on CRE efﬁciency caused by variation of the radical solution composition and by tem- perature was also investigated. (cid:1) 2020 The Authors. Published by Elsevier Inc. ThisisanopenaccessarticleundertheCCBYlicense(http:// creativecommons.org/licenses/by/4.0/).


Introduction
Dynamic Nuclear Polarization (DNP) is an emerging magnetic resonance technique to enhance the sensitivity of high-field magic-angle spinning (MAS) nuclear magnetic resonance (NMR) [1][2][3][4]. DNP enables signal enhancement, defined as the spin polarization achieved by DNP relative to the thermal equilibrium polarization of the respective nucleus, up to several orders of magnitude by transferring the large electron-spin polarization of paramagnetic polarizing agents to target NMR nuclei using microwave (MW) irradiation at specific frequencies. Alleviating the low sensitivity, which is the key challenge of traditional NMR spectroscopy, DNP has been successfully applied not only to the study of various biological samples [5][6][7] but also to solid samples in materials science [8][9][10][11][12][13][14]. However, despite substantial advances in DNP methodology [15,16] and instrumentation [17][18][19], NMR sensitiv-ity is still a limiting factor in many applications [3]. Therefore, it is important to understand the different polarization transfer mechanisms occurring simultaneously during DNP in order to enable a systematic optimization of these experiments.
Molecular dynamics affecting the longitudinal relaxation time constant, T 1 , which play a key role in the polarization transfer processes, are often neglected in MAS DNP measurements that are performed typically at 100 K, because most motional modes are effectively frozen out. However, some exceptional dynamics with small activation energies [20], such as the fast rotational motion of the methyl group [21], may show fast dynamical properties even at cryogenic temperatures. Recently, it has been reported that methyl group reorientation dynamics under DNP conditions can cause heteronuclear cross-relaxation mediated polarization transfer in the solid state [21][22][23][24]. Thereby, mobile methyl protons are hyperpolarized by DNP ( Fig. 1(a) and (b), pink arrow). Their magnetization is then spontaneously transferred to dipolarly coupled carbon atoms by cross-relaxation ( Fig. 1(a) and (b), red arrow), resulting in a negatively enhanced 13  polarization (DP) MAS NMR experiment. The transferred polarization subsequently spreads to other carbons through the 13 C dipolar network by 13 C-13 C spin diffusion ( Fig. 1(b), green arrow). This cross-relaxation process is identical to that underlying the Nuclear Overhauser Effect (NOE) [25]: saturation of proton spins with onresonance radio frequency (rf) pulses results in a positive signal enhancement in the 13 C NMR spectra even in solids if the dynamics occur on a favorable time scale [26]. Under DNP conditions, when proton spins are hyperpolarized instead of saturated, double quantum cross-relaxation between 1 H and 13 C spins results in a negative signal enhancement, which is opposite in sign to typical NOE enhancement in 13 C NMR.
Up to now, the cross-relaxation effect in DP-MAS DNP NMR has only been reported for 13 C nuclei. Here, it is investigated whether such an effect can also be confirmed for 15 N by providing experimental evidence for cross-relaxation mediated enhancement (CRE) during 15 N DP-MAS DNP using benzyl ammonium and glutamine as model compounds (Fig. 2).

Theoretical background
The cross-relaxation mediated polarization transfer can be modelled by starting with the Solomon equations [22,29] for two dipole-coupled spins I, S. To empirically include solid-state DNP, the equations of motion for a closed two-spin system can be amended by a term that supplies non-equilibrium polarization from the outside. This corresponds to the coupling of the twospin system to an external ''DNP bath", with the hypothetical steady-state polarization enhancement values e I 0 and e S 0 and corresponding enhancement rates q I DNP and q S DNP for the I and the S spin, respectively (for details see electronic Supporting Information). e I 0 and e S 0 correspond to the DNP enhancement factors that would be achieved without coupling between I and S and with infinitely slow relaxation towards thermal equilibrium. For such a system, the total steady-state S spin polarization enhancement (e tot ) including cross-relaxation under DNP is For the systems studied here, e tot is the 15 N enhancement factor, or 15 N spin polarization relative to its thermal equilibrium, and e I is the actual steady-state enhancement factor of 1 H. c I and c S are the gyromagnetic ratios of 1 H and 15 N, respectively. r IS is the 1 H-15 N cross-relaxation rate and q S is the 15 N longitudinal relaxation rate without coupling to the DNP bath. The first term on the right-hand side of this equation presents thermal equilibrium polarization (TP), the second term constitutes 15 N direct enhancement (DDE) achieved without the influence of cross-relaxation, and the third term represents the contribution from cross-relaxation effect (DCRE or DNOE). Notice the similarity of this third contribution with the enhancement in liquid-state Overhauser DNP, where the factor r IS =ðq S þ q DNP S Þ has been identified as the product of a coupling factor between the two spins and a leakage factor caused by relaxation back to thermal equilibrium [30].
The equation for the polarization enhancement may be simplified if it can be assumed that the DNP enhancement rate is much higher than the spin-lattice relaxation rate, i.e. q DNP S ) q S , which often applies at cryogenic temperatures. Then we get e tot % e 0 If, on the other hand, 15 N direct enhancement (DDE) on mechanisms other than cross-relaxation mediated from the I spin is very low, q DNP S % 0, then the same expression as by Daube et al. is obtained [22], The gyromagnetic ratio of 15 N (negative sign; c N = À27.116 Â 1 0 6 radÁs À1 ÁT À1 ) has an opposite sign compared to 13  (e I = 0) would induce a net negative enhancement (e tot < e 0 S ), a phenomenon that is traditionally called hetNOE [25,31].
In this paper, nomenclatures and symbols as shown in Table 1 are used to explain the spontaneous cross-relaxation induced

Experimental
All chemicals were analytical grade. The 15 N labeled benzyl ammonium sample was prepared by mixing 15 N labeled benzyl amine (98 atom % 15 N, Sigma Aldrich) and bis(trifluoromethylsulfo nyl)imide (!95.0%, Sigma Aldrich) in a 1:3, w/w ratio in an aprotic solvent consisting of a mixture of tetrachloroethane/DMSO d 6 / DMSO in a ratio of 67/30/3 vol%. As polarizaing agent, TEKPol [32] (Cortecnet) radical was added in a final concentration of 10 mM. Since the TEKPol radical is unstable in an acidic environment and at room temperature as can be seen in Fig. S1 in the supporting information, the TEKPol radical was added just before loading the sample into the MAS rotor and cooling down. The nitrogen spectrum exhibits a quartet splitting pattern in the 15 N solution NMR spectrum (Fig. S2) without proton decoupling, indicating that the N-functional group exists as the primary ammonium ion.
The saturated 15  DNP experiments were performed on a Bruker (Karlsruhe, Germany) wide-bore Avance III HD 600 MHz spectrometer equipped with a triple resonance TCI ( 1 H, 13 C, 15 N) cryoprobe connected to a 395 GHz gyrotron with 60 mA of beam current as a continuous microwave source. Experiments were conducted at 100 K for all samples and additionally at 140 K for the glutamine sample in glycerol/H 2 O (60/40 vol%) to confirm the temperature dependence of the cross-relaxation. For all experiments, excitation 90°pulses of 6.0 and 2.8 ms duration, corresponding to an rf field strength of 42 kHz and 89 kHz, were applied for 15 N and 1 H, respectively. SPINAL-64 decoupling was applied during acquisition with a 1 H rf field of ca. 90 kHz. The number of scans was 8 at a MAS spinning frequency of 9 kHz. The pulse sequence [24] in Fig. S3 was used to identify the intimately linked the CRE effect and the NOE, depending on the presence or absence of the 1 H saturation pulse train. A presaturation pulse-train with 16 90°pulses separated by 3 ms was applied to both 1 H and 15 N to destroy any transverse magnetization left. All direct DP spectra were measured using a single 90°p ulse excitation of 15 N without 1 H saturation pulses. For the 1 H saturation experiment (DP sat ), 180°1H saturation pulses with a pulse interval of 500 ms (Fig. S3, d21) were used to prevent further 1 H polarization build-up. All DP and DP sat NMR experiments were performed using 6 variable polarization delays ranging from 10 s to 600 s.

Results and discussion
To estimate the CRE effect and the related NOE in the 15 N DP-MAS DNP spectrum of the benzyl ammonium cation, 15 N DP measurements were conducted with and without MW irradiation as well as with and without 1 H saturation pulse trains as a function of polarization time. Comparisons of the different data sets according to different properties are shown in Fig. 3 and Fig. S4. As can be seen in Fig. 3(a) and S4(a), the signal intensities in the spectra and the build-up curves are drastically affected by 1 H saturation pulse trains, which means that the CRE effect and the NOE play a significant role at 100 K. If the primary ammonium group of benzyl ammonium were not mobile at 100 K, a CRE effect and a NOE would not be expected, and the same intensity and build-up curves would be observable in Fig. 3(a,b) and S4(a,b), regardless if a 1 H saturation pulse was applied or not. Likewise, 1 H saturation also leads to a reduction of the signal intensity in the absence of DNP hyperpolarization, as shown in Fig. 3(b) and S4(b), which is a reflection of the NOE in the 15 N spectrum. Comparing the effects observed in Fig. 3(a) and (b), the CRE effect contributes more strongly to the signal intensity of the 15 N spectrum than the NOE and the DE in this sample. From Fig. 3(c) and (d), we can determine the contribution of the CRE in the 15 N direct hyperpolarization effect (DP). Fig. 3(c) shows a larger difference than Fig. 3(d), which means that the CRE effect dominates the positive enhancement of the 15 N signal in 15 N DP-MAS DNP.
The full observable 15 N signal enhancement factor (e tot ) in benzyl ammonium is strongly influenced by two different effects, the CRE (Fig. 1 a and b, red arrows), and the 15 N DE (Fig. 1 a and b, blue and green arrows). The 15 N MAS NMR spectrum recorded with 1 H saturation and without MW, i.e. without any DNP effect, shows the smallest signal intensity due to the presence of the NOE which contributes a negative signal enhancement, while the highest signal intensity is found when both the 15 N DE and the CRE effect contribute, which is achieved without 1 H saturation and with MW irradiation (Figs 3, 4 and S4). From these results, it can be seen that ammonium ion rotational mobility, which enables CRE in 15 N DP-MAS DNP NMR, contributes significantly to the overall 15 N polarization even at 100 K (Fig. 4, orange dashed line). As opposed to 13 C, due to the negative sign of the 15 N gyromagnetic ratio, an additional positive enhancement is found for the 15 N polarization.
To determine the CRE effect in 15 N DP-MAS DNP NMR spectra in a protic solvent, glutamine having one ammonium and one amide group (N a and N e ) with well-separated chemical shifts was used in a popular protic radical solution with a protonation degree of~10% i.e., glycerol-d 8 /D 2 O/H 2 O (60/30/10 vol%) [34]. As seen in Fig. 5(a), 1H saturation has no impact on the signal intensity, which means the CRE effect and the NOE found in the previous benzyl ammo-  nium sample are not significant in glutamine with this radical solution composition. This could be caused by two reasons: Either, the 1 H-15 N CRE effect induced by mobile protons under DNP conditions may be reduced because H/D exchange is expected for amine protons in protic solvent containing deuterium on the one hand (Figs. S5 and S6). Further, the amine dynamics of glutamine at 100 K could be insufficient for double quantum cross-relaxation between 1 H and 15 N on the other hand. In order to investigate this, the experiments were repeated with increasing 1 H concentration in the radical solution. As shown in Fig. 5(a-c), an increasing signal intensity ratio (SIR) between the data with and without proton saturation can be identified at higher 1 H concentration. This indicates that the protonation ratio of the amine of glutamine, which is altered by H/D exchange, can affect the efficiency of the 1 H-15 N CRE effect.
In general, deuterium in a radical solution is known to be beneficial for improving the diffusion of polarization from the sites of initial electron-nucleus polarization transfer to nuclei far away from the radical center and to increase the polarization transfer efficiency from protons to heteronuclei through relaxation-prolongating effects [35][36][37]. While some protons in the radical solution are needed to spread the polarization in the bulk, too many or too little of them can weaken the enhancement. In NMR of biological samples, the radical solution composition used for the spectrum shown in Fig. 5(a) has generally lead to the best performance [38]. However, when comparing the intensities of the plots in Fig. 5(a-c), the 15 N NMR signal intensity for glutamine increases with higher 1 H concentration. The relative concentrations of the free proton in the radical solvents have a H/D (molar/molar) ratio of 16:84, 64:36, and 100:0. We further determined the DNP enhancement factor (e 15N CP ) of 15 N CPMAS (Fig. S6, Table S1) and with this the 1 H polarization. A change in the protonation degree from 64% to 100% results in a decrease of the 1 H polarization by a factor of 0.7, while in the DP experiment the additional signal by the CRE effect increases by a factor of about 2. Therefore, under the tested experimental conditions, for glutamine the 1 H-15 N CRE effect modulated by the amine protons appears to be more significant for the signal enhancement than the relaxation-prolongating effect of the deuterium.
To confirm the difference in CRE by the dynamic properties of the two nitrogen groups (N a and N e ) in glutamine, the temperature was increased to 140 K. In Fig. 5(c) and (d), the fully protonated sample is compared at different temperatures. The rate of change of the enhancement factor for N a is more sensitive than that of N e . The structure of glutamine at neutral pH contains zwitterionic forms consisting of an a-amino group (N a ) in the protonated -NH 3 + form and a carboxylic acid group in the deprotonated -COO À form, and a simple amide (N e ) side chain. As can be seen in Fig. S7, the amide (N e ) side chain is stabilized due to the partial-double bond, resulting in slower dynamics. Therefore, the amide (N e ) side chain, which may have a smaller change in dynamics when the temperature is increased, experiences a smaller CRE effect on the spectrum.

Conclusions
We have shown evidence for the CRE effect, resulting in additional positive signal enhancement for the 15 N NMR spectrum, by the spontaneous polarization transfer from hyperpolarized 1 H to 15 N during 15 N DP-MAS DNP. A larger CRE effect can be achieved in an aprotic solvent due to the prevention of H/D exchange between solvent and substrate, as well as a more rapid reorientation dynamics of N-functional groups caused by weaker interactions.
Similar to a previous application [24] of a specific CRE by active motion under DNP (SCREAM-DNP) through the introduction of a 13 CH 3 labeled functional group as a probe into biomolecular systems, we expect that further surface signal enhancement using this effect could lead to more efficient and selective DNP-Surface enhanced NMR spectroscopy (DNP-SENS) in materials science. For example, surface structure information of nitrogen functionalized materials [39,40], which are prosperous as heterogeneous catalysis and energy materials, is critical for their further development. The surface signal can be selectively enhanced more strongly during DNP-SENS by binding molecules carrying rotatable nitrogen functional groups to the surface of the material. This would allow distinguishing clearly between surface and bulk signals due to a specific 1 H-15 N CRE effect.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.