Combining heteronuclear correlation NMR with spin-diffusion to detect relayed Cl – H – H and N – H – H proximities in molecular solids

Analysis of short-to-intermediate range intermolecular interactions offers a great way of characterizing the solid- state organization of small molecules and materials. This can be achieved by two-dimensional (2D) homo- and heteronuclear correlation NMR spectroscopy, for example, by carrying out experiments at high magnetic fields in conjunction with fast magic-angle spinning (MAS) techniques. But, detecting 2D peaks for heteronuclear dipolar coupled spin pairs separated by greater than 3 Å is not always straightforward, particularly when low-gamma quadrupolar nuclei are involved. Here, we present a 2D correlation NMR experiment that combines the ad-vantages of heteronuclear-multiple quantum coherence (HMQC) and proton-based spin-diffusion (SD) pulse sequences using radio-frequency-driven-recouping (RFDR) to probe inter and intramolecular 1 H-X (X = 14 N, 35 Cl) interactions. This experiment can be used to acquire 2D 1 H{X}-HMQC filtered 1 H – 1 H correlation as well as 2D 1 H-X HMQC spectra. Powder forms of dopamine ⋅ HCl and L -histidine ⋅ HCl ⋅ H 2 O are characterized at high fields (21.1 T and 18.8 T) with fast MAS (60 kHz) using the 2D HMQC-SD-RFDR approach. Solid-state NMR results are complemented with NMR crystallography analyses using the gauge-including projector augmented wave (GIPAW) approach. For histidine ⋅ HCl ⋅ H 2 O, 2D peaks associated with 14 N – 1 H – 1 H and 35 Cl – 1 H – 1 H distances of up to 4.4 and 3.9 Å have been detected. This is further corroborated by the observation of 2D peaks corresponding to 14 N – 1 H – 1 H and 35 Cl – 1 H – 1 H distances of up to 4.2 and 3.7 Å in dopamine ⋅ HCl, indicating the suitability of the HMQC-SD-RFDR experiments for detecting medium-range proximities in molecular solids.


Introduction
Solid-state (ss)NMR spectroscopy is widely used to characterize the local structures and dynamics of molecular solids today. Site selectivity, natural isotopic abundance of several NMR active nuclei, and availability of a large pool of one-(1D) and two-dimensional (2D) pulse sequences to carry out homo and heteronuclear correlation experiments are among the unique capabilities of ssNMR spectroscopy for characterizing the solid form. Applications have been presented in the areas of optoelectronics, catalysis, biomacromolecules, pharmaceutical formulations, energy harvesting and storage materials [1][2][3][4][5][6][7]. In addition, ssNMR, in conjunction with diffraction-based techniques and modelling approaches (also referred to as NMR crystallography), is increasingly applied to characterize the intermolecular interactions and to determine the three-dimensional structures of (semi)crystalline and amorphous solids [8][9][10][11][12][13][14][15][16][17][18].
To take full advantage of NMR analysis of molecular solids, it is important to probe packing interactions in the sub-nanometer to nanometer range, or even greater than a nanometer distance. Gaining access to the intermediate-range (1-10 nm) length scale is particularly important to fill the gap between the structural analysis enabled by the short-range probes (e.g., conventional magnetic resonance spectroscopy probes local structures at sub-nanometer distances) and the long-range techniques such as X-ray scattering and electron microscopy that offer morphological and structural information at tens to hundreds of nanometers regimes. For example, it has been feasible to probe intermediateto-long range distances via spin-diffusion experiments by making use of spin magnetization exchange between nuclei at high natural abundance such as 1 H and 19 F due to the strong dipolar interactions between them, or between isotopically enriched nuclei [43][44][45][46][47][48][49][50][51]. Several structural elucidation studies benefit from the analysis of such intermediate-range to long-range structures, including catalytic surfaces, reactive interfaces in stacked thin films in optoelectronic devices, bio-macromolecules, polymers, functional supramolecular assemblies, and pharmaceutical formulations [44,[52][53][54][55][56][57][58][59][60][61][62][63][64]. Notably, analysis of spin-diffusion build-up curves in conjunction with computational modelling has been used to determine three-dimensional structures of small to moderately sized molecules [47,48,[65][66][67][68]. Since these molecular systems contain both spin-half and quadrupolar nuclei, detecting 2D NMR peaks for heteronuclear 1 H-X spin-pairs involving low-gamma quadrupolar nuclei (X = 14 N, 35 Cl), probing long-range proximities of up to 5 Å or beyond is difficult due to the weak heteronuclear dipolar interactions. Thus, a combination of features involving a 2D HMQC-type experiment and 1 H-1 H spin diffusion to probe the short-to-medium range interactions is expected to facilitate structure elucidation in organic solids and materials.
Here we combine HMQC and SD using RFDR pulse sequence elements so as to probe 35  complemented by periodic density functional theory calculations using the gauge-including projector augmented wave (GIPAW-DFT) approach. These results suggest that the proposed method has wider relevance in investigating the medium-range interatomic proximities between spinhalf and low-gamma quadrupolar nuclei in molecular solids.  14 N pulse lengths, dipolar recoupling periods (τ rcpl ), and spin diffusion times using RFDR (τ RFDR ) were empirically optimized using 1D versions of the HMQC and HMQC-SD-RFDR pulse sequence shown in Fig. 2. 1D 1 H MAS NMR spectra of L-histidine⋅HCl⋅H 2 O and dop-amine⋅HCl were acquired with 128 co-added transients. 1D 35 Cl spectra of L-histidine⋅HCl⋅H 2 O and dopamine⋅HCl⋅H 2 O were acquired with 512 and 4096 co-added transients using a Quadrupolar Carr-Purcell-Meiboom-Gill (QCPMG) sequence [69]. All 1 H chemical shifts are calibrated with respect to neat TMS using adamantane as an external reference ( 1 H resonance, 1.85 ppm) [70]. 14 N shifts were referenced to a saturated NH 4 Cl aqueous solution at − 352.9 ppm, corresponding to the primary reference, liquid CH 3 NO 2 (0 ppm) [71]. To compare with the  HMQC-SD-RFDR spectra. The sequence shown in (b) can be used to acquire 1 H {X}-filtered 1 H-1 H SD correlations spectra as well as 1 H-X HMQC-SD-RFDR correlation spectra, depending on the placing of the t 1 evolution period. In both cases, SR4 2 1 pulses are applied for the recoupling of 1 H-X (X = 35 Cl, 14 N) dipolar interactions, and in (b), an RFDR block is added to enhance the spin diffusion process leading to a 1 H-detected 2D experiment. The evolution during t 1 and the 1 H RFDR pulses during spin diffusion are rotor-synchronized with respect to the sample spinning.

Solid-state NMR spectroscopy
alternative IUPAC reference of liquid NH 3 at − 50 • C as used in protein NMR, it is necessary to add 379.5 ppm [71]. 35 Cl experimental shifts are referenced to 0.1 M NaCl solution in D 2 O based on the IUPAC recommendation (see Table A1 in Ref. [72] 35 Cl spectra were acquired with 80 t 1 FIDs, each with 128 co-added transients, corresponding to an experimental time of ~8.5 h each. HMQC-SD-RFDR pulse sequence: To begin with, an HMQC pulse sequence ( Fig. 2a) is used for the excitation and reconversion of HMQC coherences, whereby the interval between the two 14 N pulses are rotor synchronized with sample spinning, i.e., t 1 = nτ r , where n is an integer. During this sequence, the SR4 pulse sequence was applied during the excitation and reconversion periods to decouple the 1 H-1 H homonuclear dipolar interactions and recouple the 1 H-14 N heteronuclear dipolar interactions. To this sequence, a spin-diffusion filter is added with two 90 • pulses separated by a mixing delay, also referred to as a spin diffusion delay. Since 1 H-1 H spin diffusion is suppressed when experiments are carried out at fast sample spinning (here 60 kHz MAS), radio-frequencydriven recoupling (RFDR) [74] pulses ( Fig. 2b) are added within the spin diffusion filter to enhance the spin diffusion process. This leads to a 1 H detected 2D pulse sequence that can be applied to acquire 2D 1 H { 14 N}-HMQC filtered 1 H-1 H NOESY-like spectra (by placing the t 1 evolution between the HMQC block and the spin diffusion block, as depicted in gray color) as well as 2D 1 H-14 N relay HMQC spectra. To compare the spectral filtering efficiency, spectra acquired using the 1D versions of the HMQC and HMQC-SD-RFDR experiments are compared with a single-pulse experiment in Fig. 3 and in a later section (all spectra are acquired at identical experimental conditions). The sensitivity of the conventional 1D 1 H{ 14 N/ 35 Cl} HMQC experiment is about 5%, and the 1 H{ 14 N/ 35 Cl}-HMQC-SD-RFDR experiment is about 1%, as compared to the single-pulse 1 H MAS experiment. Of particular note, the sensitivity of the proposed experiment depends on the relative strengths of both homo and heteronuclear dipolar couplings. For example, in both samples, the efficiency for NH 3 protons is less due to the rotational dynamics of the NH 3 group as compared to the aromatic protons. 2D HMQC-SD-RFDR experiments were acquired under identical conditions to those of conventional HMQC experiments, except the RFDR-based spin-diffusion mixing times are added as indicated in the appropriate figure captions.
Quantum mechanical calculations using a periodic-DFT approach. The molecular coordinates obtained from previously published crystal structures of L-histidine.HCl.H 2 O (CCDC code: HISTCMO1) [75] and dopamine⋅HCl (CCDC codes for the HCl salt and zwitterionic forms are DOPAMN and TIRZAX) [76] were taken to prepare input files for the quantum mechanical calculations. The geometry optimizations of crystal structures were carried out with periodic density functional theory (DFT) calculations. For all the geometry optimized structures, the NMR chemical shifts were computed using the gauge including projected augmented wave (GIPAW) method as described by Pickard and Mauri [77,78]. All periodic DFT calculations were performed using the CASTEP 19.11 code [79]. Each self-consistent field (SCF) loop was performed until the energy was converged to within 3.67 × 10 − 8 Hartrees. For all calculations, the generalized density approximation DFT functional PBE with the Tkatchenko-Scheffler (TS) dispersion correction scheme (DFT-D method) was applied with ultrasoft pseudopotentials [80,81]. The maximum plane wave cut-off energy was 23.15 Hartrees. The Broyden-Fletcher-Goldfarb-Shanno (BFGS) optimization algorithm was used, and a Monkhorst− Pack grid of minimum sample spacing 0.07 HMQC-filtered 1 H spectrum recorded using τ rcpl of 320 μs and also with τ RFDR = 1000 μs, and a 1 H{ 35 Cl} HMQC-filtered 1 H spectrum (800 MHz, 60.24 kHz MAS) recorded using τ rcpl = 265.6 μs and also with τ RFDR = 265.6 μs.
× 2π Å − 1 was applied to sample the Brillouin zone. The positions of the atoms were allowed to vary within the unit cell until the average forces, energies, and displacements remaining were below 3.6749 × 10 − 7 Hartree/Å, 0.0011025 Hartrees, and 0.001 Å, respectively. The interatomic distances for histidine.HCl.H 2 O and dopamine⋅HCl molecules stated in this study correspond to the DFT optimized structures.

Results and discussion
L.histidine⋅HCl⋅H 2 O: The solid-state form of L-histidine⋅HCl⋅H 2 O, and protonated and tautomeric structures of histidine have been well characterized by 1 H, 13 C, 14 N/ 15 N, and 35 Cl MAS NMR spectroscopy in previous studies [36,38,40,82,83]. Here, we used L-histidine⋅HCl⋅H 2 O as a control sample to test the proposed HMQC-SD-RFDR experiment. double-quantum-single-quantum (DQ-SQ) correlation NMR spectrum together with the DQ peak assignments corresponding to 1 H-1 H distances between 3 Å and 4.5 Å is presented in Fig. S1 (Supporting Information). For the spectra acquired under identical experimental conditions, spectral simplification and sensitivity of the 1 H signals achieved by 1 H{ 14 N} and 1 H{ 35 Cl} HMQC filters with respect to the single-pulse experiment are examined and compared (Fig. 3, colored shaded regions). The 1D 1 H{ 14 N}-HMQC filtered spectra acquired with 320 μs of SR4 recoupling time exhibited signals corresponding to the directly bonded 1 H-14 N sites, i.e., H1, H5, and H7 peaks are detected, while the signals originating from all other protons are poorly detected (or undetected). Although the addition of a spin diffusion delay leads to the exchange of magnetization from 1 H{ 14 N} sites to the neighboring 1 H sites, the proton spin-diffusion (SD) is less efficient when experiments are carried out under fast MAS [84][85][86]. Therefore, the RFDR pulses have been used within the spin-diffusion filter to enhance the 1 H-1 H spin-diffusion process. The addition of a SD-RFDR block (1000 μs) to the 1D HMQC experiment with the same 320 μs of SR4 recoupling time enables the weak intensity signals (H2, H3, H4, and Hw) to be detected with much higher intensities, although the intensities of H5 and H7 peaks are reduced, yielding signal uniformity like that of single-pulse experiment. A similar trend is observed for the 1 H{ 35 Cl} HMQC filtered spectra of L-histidine⋅H 2 O⋅HCl, whereby the spectra acquired with 265.6 μs of SR4 recoupling time before and after the addition of RFDR (265.6 μs) showed different 1 H peak intensities. In the latter, the H5 and H2-H4 sites exhibited signal enhancements, as indicated in Fig. 3. A striking advantage of the SD-RFDR block is that it enables the 1 H signals corresponding to the dipolar coupled 1 H-1 H pairs in the vicinity of 14 N and 35 Cl sites to be detected, demonstrating the sensitivity enhancements for specific 14   correlation spectra of L-histidine.HCl.H 2 O acquired with different mixing times in the 1-100 ms range are presented in Supporting Information (Fig. S2), whereby the short mixing time of 10 ms was insufficient to exchange the magnetization between the neighboring sites, and 100 ms enabled magnetization exchange between different sites. The advantage of using RFDR blocks to enhance spin-diffusion process in 2D experiments acquired under fast MAS is demonstrated in Fig. S3. By adding an HMQC filter prior to the 1 H-1 H SD-RFDR pulse sequence, a 2D 1 H{ 14 N} HMQC-filtered 1 H-1 H SD spectrum can be acquired. In a HMQC-filtered 2D 1 H-1 H SD-RFDR spectrum acquired with a mixing time of 16.6 μs, i. e., one rotor period (Fig. S3c), the diagonal peaks associated with NH 3 , H5, and H7 protons are observed. By comparison, the RFDR mixing time of 531.2 μs enabled the 2D correlation peaks between all the sites to be observed (Fig. S3d), confirming that the RFDR enhances the spindiffusion process. It is to be noted that similar experiments can be carried out using a 35 Cl{ 1 H} HMQC filter to achieve spectral simplification in HCl salts of small organic molecules.
Next, 2D 1 H-14 N HMQC spectra were acquired without and with SD-RFDR pulses (Fig. 4)  respectively. These distances are depicted in the periodic DFT optimized crystal structure of L-histidine⋅HCl⋅H 2 O shown in Fig. 4c. Further increase in the recoupling time up to 800 μs (spectra not shown) yields a poor signal to noise ratio and no additional long-range N-H correlations were detected, suggesting that the conventional HMQC experiment is unable to probe through-space proximities with N-H exceeding 3 Å (dashed boxes in Fig. 4b) [87]. This is because of the dipolar-based HMQC experiment utilizing heteronuclear dipolar interactions (D NH ), which are an order magnitude weaker than the homonuclear proton-proton dipolar couplings (D HH ). For example, a through-space dipolar coupled 14 N-1 H pair separated by ~4.43 Å distance exhibits a dipolar coupling of ~100 Hz, which is too weak to produce a detectable 2D peak in the conventional HMQC spectrum. For the same 14 (Fig. 4f), suggesting that the correlations between nitrogen atoms and remote protons can be observed using this approach. It is noteworthy that 3D 1 H-1 H-14 N or 1 H-1 H- 19 F correlation experiments provide identical information [36,88,89], however, the 2D HMQC-SD-RFDR experiment presented in this study requires shorter acquisition times than the 3D experiments to provide identical information, which is an added advantage. The proposed HMQC-SD-RFDR approach has been applied to examine relayed 35 Cl⋯ 1 H-1 H proximities in L-histidine.HCl.H 2 O. A 1D 35 Cl NMR spectrum acquired with a QCPMG sequence is presented in Fig. S4 (Supporting Information) together with a lineshape fitting analysis to obtain the isotropic chemical shifts and quadrupolar coupling constants, as listed in Table S1. The narrow quadrupolar lineshape is due to the weak quadrupolar interaction associated with the chloride anions. Fig. 5 compares 2D 1 H-35 Cl D-HMQC spectra recorded without and with SD-RFDR pulses. In the conventional 1 H- 35 Cl HMQC spectra of L-histidine.HCl.H 2 O acquired with a short recoupling time of 265.6 μs (Fig. 5a), the 2D correlation peaks associated with closely proximate and dipolar coupled 1 H- 35 Cl are observed for the Cl atoms and H1, H6 and H8, and Hw protons. The inter and intramolecular Cl⋯H distances associated with Cl⋯H1, Cl⋯H6, Cl⋯H7, Cl⋯H8, and Cl … Hw are 2.20, 2.44, 2.94, 2.75 and 2.16 Å, respectively (Fig. 5c). Similar 2D peaks are observed for a 2D HMQC spectrum acquired with a recoupling time of 531.2 μs, shown in Fig. 5b; however, a weak intensity peak corresponding to Cl⋯H5 (3.86 Å) appears. This indeed corroborates that the detection of the 2D 35 Cl-1 H peak for a Cl⋯H distance closer to 4 Å is difficult to achieve in a conventional 2D HMQC experiment due to the weaker D Cl-H . By combining heteronuclear 35 Cl⋯ 1 H and homonuclear 1 H-1 H spin diffusion process, these weak intensity peaks are detected with much higher intensities, as shown in the HMQC-SD spectra presented in Fig. 5d and e. In addition, the enhanced uniformity of 2D peak intensities is apparent from these spectra, which indicates the addition of RFDR enables the detection of relayed 35 Cl-1 H-1 H peaks with an interatomic 35 Cl-1 H distance exceeding 3.8 Å (Fig. 5f).
Dopamine⋅HCl. Dopamine is a catecholamine derived from L-tyrosine, which has key biological significance as a neurotransmitter in the central nervous system and as a hormone in vesicles of the adrenal medulla that influences and controls the heartbeat rate, oxidative stress, and blood pressure [90,91]. The impaired dopamine metabolism and the absence of dopamine-containing neurons are associated with brain disorders such as Parkinson's disease [92]. In addition, dopamine is a food supplement (L-Dopa) that can be used as a direct precursor to increase the dopamine level in the body. In the presence of oxygen and alkaline/acidic environments, dopamine self-assembles into polydopamine (pDA) [93]. Care was taken to prevent the self-polymerization of dopamine by storing the powder form in a glass desiccator. To ensure that the dopamine characterized in this study is in monomeric form, liquid-state 1D 1 H and 2D 1 H-1 H and 1 H-13 C correlation NMR spectra were acquired and analyzed (Supporting Information, Figs. S5-S7): the narrow 1 H signals indicate the monomeric form of dopamine rather than the polymeric form. In the solid state, dopamine exists in different crystalline polymorphs such as neutral and zwitterionic forms, for which crystal structures have been previously solved [76,94]. Here, the solid-state form of dopamine⋅HCl is characterized by an NMR crystallography approach that combines solid-state NMR and periodic DFT calculations of previously published crystal structures. To further ensure that the dopamine preserved its molecular form, the GIPAW-DFT calculated chemical shifts for 1 H SQ, and DQ correlation peaks for all 1 H-1 H pairs (within 3 Å) in molecular and zwitterionic forms are overlaid on the experimental DQ-SQ peaks (Supporting Information, Fig. S8). A good agreement between the experimental 1 H SQ and DQ chemical shifts and the GIPAW-DFT calculated 1 H NMR chemical shifts is observed for the molecular form, but not for the zwitterionic form.
In Fig. 6 and 1D 1 H{ 14 N} and 1 H{ 35 Cl} HMQC spectra acquired with and without SD-RFDR blocks are compared with a single-pulse 1 H spectrum. The broad peak centered at ~7.6 ppm is due to an overlapped contribution from the NH 2 and aromatic protons. In the aliphatic region, the chemical shifts at 3.4 ppm and 5.9 ppm correspond to a CH 2 group next to the aromatic ring, and one of these protons is hydrogen-bonded to the oxygen atoms of the adjacent dopamine molecules. The peak at 4.9 ppm corresponds to the CH 2 group next to the NH 3 group. The peak integral ratio of 7:4 associated with the signals at 6.5-9.5 ppm and 2.5-6.5 ppm further confirms the peak assignments. While all 1 H peaks are observed in the conventional 1 H spectrum, the 1 H{ 14 N}-HMQC filtered 1D spectrum acquired with 132.8 μs of 1 H-14 N dipolar recoupling time using the SR4 sequence yields the 1 H signals of NH 3 peaks, and the 1 H signals of CH 2 protons are suppressed. For the same SR4 recoupling time, the addition of RFDR pulses (531.2 μs) leads to a 1 H NMR lineshape similar to that of the single-pulse experiment, i.e., the 1 H signal intensities in the 2.5-6.5 ppm range are enhanced. This trend is also evidenced in the 1 H{ 35 Cl} HMQC filtered experiments (τ rcpl = 265.6 μs) with and without RFDR blocks, whereby the signals in the range of 2-7 ppm were only detected when an RDFR duration of 531.2 μs was added. Furthermore, 2D 1 H-1 H spin diffusion NMR spectra We extend this analysis to probe the 1 H- 35 Cl proximities in dop-amine⋅HCl. A 1D 35 Cl NMR spectrum acquired with a QCPMG sequence is presented in Fig. S10 (Supporting Information) together with a lineshape fitting analysis. The isotropic chemical shift, quadrupolar coupling constant, and asymmetry parameter are listed in Table S2. In 1 H- 35 Cl HMQC spectra of dopamine⋅HCl acquired with and without SD-RFDR blocks (Fig. 8), short and long-range correlations are observed.
When RFDR pulses of duration 1024 μs are applied, 2D peaks between 35 Cl and CH 2 groups are also observed, as evident in Fig. 8b. The nearest internuclear 35 Cl-CH 2 (adjacent to NH 2 ) and 35 Cl-CH 2 (adjacent to the aromatic ring) distances are 3.5 and 3.7 Å, respectively. Although the 2D HMQC-SD-RFDR approach enables the relayed correlation peaks to be detected, the measurement of interatomic distances from 2D volume integrals is less straightforward. In order to gain a quantitative insight into the distance constraints, a series of 2D experiments varying the heteronuclear recoupling time and spin diffusion mixing time would need to be acquired and the signal intensity buildup would need to be analyzed. Further improvements to the proposed HMQC-SD-RFDR approach can be sought, for example, using the TRAPDOR-HMQC (TRAPDOR -transfer of population in double resonance) approach [42,95], by detecting the double-quantum (DQ) satellite transitions of quadrupolar nuclei. It has been shown that the DQ version is less sensitive to MAS instabilities and also benefits from a slightly higher resolution since the DQ chemical shift evolves at a frequency twice as large as the SQ chemical shifts. It has also been demonstrated that frequency-swept 14 N overtone CW-RESPDOR (continuous wave rotational-echo saturation-pulse double-resonance) experiments efficiently probe 1 H-14 N proximities with much longer internuclear distances than for the HMQC experiment [30]. The use of SD-RFDR block with CW-RESPDOR is expected to further extend the limit of detection of 2D peaks. Acceleration in the acquisition times can also be achieved at fast MAS by frequency selective (FS)-HMQC experiments, whereby selective excitation and detection of 1 H leads to a reduction in the inter-scan delays [41]. Notably, the proposed pulse sequence and its variants can be applied to detect 1 H-1 H and 1 H-14 N/ 35 Cl proximities in order to facilitate the structure elucidation of weak intermolecular interactions in organic molecules, reactive surfaces and interfaces of functional materials. It is to be noted that the SD transfer at fast spinning (>60 kHz MAS) may also be influenced by the chemical shift offset difference particularly when there is a small range of chemical-shift differences in the 1 H MAS spectra. It can be compensated by the addition of a spin lock, although the short T 1ρ of protons allows only short mixing times to be applied and long mixing times often reduce sensitivity. Reverse amplitude-modulated mixed rotational and rotary-resonance (AM-MIR-ROR) sequences can be tailored to compensate for offsets in the relevant range of chemical-shift differences, and these sequences also offer relatively high sensitivity compared to other sequences [84]. In addition, the reverse AM-MIRROR sequence requires lower rf fields than RFDR, and the optimum rf frequencies do not depend on the MAS frequency but only on the recoupled chemical-shift range.

Conclusions
This study presents a 2D HMQC-SD-RFDR experiment for probing the 1 H-1 H and 1 H-X (here, X = 14 N and 35 Cl) proximities in small organic molecules. A detailed analysis of 1D and 2D HMQC-SD-RFDR spectra is presented for L-histidine⋅HCl⋅H 2 O and dopamine⋅HCl. The 1D version of the SD-HMQC sequence leads to spectral simplification by detecting 1 H { 14 N} and 1 H{ 35 Cl} peaks for protons that are either J-coupled or dipolar coupled with 14 N or 35 Cl sites. By introducing an SD-RFDR block, 1 H peaks associated with 14 N-1 H-1 H and 35 Cl-1 H-1 H sites can be selectively excited and detected through relay transfer. Therefore, the spectral simplification in 1D spectra can be achieved by tuning the heteronuclear recoupling durations within the HMQC block as well as the spin diffusion time in the RFDR blocks. This can be simply achieved through the one-dimensional version of the proposed experiment. More elaboratively, the proposed experiments can be used to acquire 2D 14 Nfiltered 1 H-1 H spin-diffusion spectra and 1 H-14 N or 35 Cl-1 H correlation NMR spectra. Specifically, the 2D 1 H-14 N D-HMQC-SD-RFDR correlation experiment demonstrates a strong ability for observing remote 14 N-1 H correlations for which the N-H distances exceed 4 Å, which is difficult to achieve with the conventional 1 H-14 N HMQC experiment. The proposed approach particularly benefits from using higher magnetic field strengths under fast MAS conditions.

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.

Data availability
Data will be made available on request. μs and τ RFDR of 304 μs? In both cases, data were acquired with 128 coadded transients and 80 t 1 FIDs, corresponding to a total experimental time of 8.5 h each. The base contour levels are at 3% of the maximum peak height. A schematic representation of crystal packing in periodic DFT geometry optimized structures illustrating the short-and long-range interactions probed by the HMQC and HMQC-SD-RFDR approaches. Specifically, spheres represent the 1 H- 35 Cl proximities within a distance of (a) 2.2 Å and (b) 3.7 Å.