Solid-State 77Se NMR of Organoselenium Compounds through Cross Polarization Magic Angle Spinning (CPMAS) Method

Characterization of selenium states by 77Se NMR is quite important to provide vital information for mechanism studies in organoselenium-catalyzed reactions. With the development of heterogeneous polymer-supported organoselenium catalysts, the solid state 77Se NMR comes to the spotlight. It is necessary to figure out an advanced protocol that provides good quality spectra within limited time because solid state 77Se NMR measurements are always time consuming due to the long relaxation time and the relatively low sensitivity. Studies on small molecules and several novel polymer-supported organoselenium materials in this article showed that cross polarization (CP) method with the assistance of magic angle spinning (MAS) was more efficient to get high quality spectra than the methods by using single pulse (SP) or high power 1H decoupling (HPHD) combined with MAS. These results lead to a good understanding of the effect of the molecular structure, the heteronuclear coupling, the long-range ordering of the solid (crystal or amorphous), and the symmetry of 77Se on quality of their spectra.

etc [39][40][41][42][43] . The method offers a beneficial solution to achieve higher SNR based on the applications of polarization transfer from the abundant proton ( 1 H) spin to low abundance nuclei i.e. 77 Se in the present paper via heteronuclear dipole-dipole interactions. The improvements are originated from enhanced sensitivity of the Se and faster longitudinal relaxation time (T 1 ) of 1 H compared to 77 Se leading to shorter recycle delay time (D 1 ) between every acquisition (≥5T 1 ) [44][45][46] . Experiments through traditional method of SP and a single pulse sequence with special high power 1 H decoupling (HPHD) were also performed for comparison with the CP method. The CP method has been successfully applied to determine the state of the low-content selenium in several novel polymers. The novelty of the present paper is that the correlations of intensity of 77 Se NMR signal with the solid state and molecular structure of polymers were demonstrated for the first time. Herein, we wish to report our findings.

Results
Solid-state properties of the samples were initially investigated. Figure 1 presents the powder X-ray diffraction (PXRD) patterns of the model compounds and polymer materials containing 77 Se. PXRD pattern of H 2 SeO 3 1 is completely different from that of PhSeO 2 H 2, confirming their different crystal structures due to the presence of a benzene ring 47 . In contrast, all the polymer materials exhibit a broad reflection, indicating their amorphous structures 48 .
To optimize the experimental parameters for different pulse sequences, H 2 SeO 3 1 was first measured and used as a secondary external reference standard. Figure 2 presents the 77 Se NMR spectra obtained by the different  pulse sequencesas described in experimental section: single pulse (SP), high power 1 H decoupling (HPHD) and cross polarization (CP). To well illustrate the efficiency of the different methods, the intensity of the isotropic center-bands was normalized. A significant spinning side band manifold separated by the rotation frequency as a result of the chemical shift anisotropy (CSA) was observed in Fig. 2 left. The position of the isotropic center-bands (*) was determined by acquiring an additional spectrum at a different spinning speed of 8 kHz, which was enlarged in Fig. 2 right. The isotropic band remained at the same position for different speeds, while spinning side bands were changed (dashed lines, Fig. 2). Application of both HPHD and CP could enhance relative intensity of the resonance peak and narrow the lines (Fig. 2). An asymmetric isotropic band was observed for HPHD and CP, which might be related to defect in H 2 SeO 3 crystal 49 . The chemical shift obtained by SP sequence was chosen to be the reference standard (1288.1 ppm). Full Width at Half Maxima (FWHM) and the SNR of the isotropic band were obtained and listed in Table 1. The SNR was improved from 13 for SP to 28 for HPHD and to 45 for CP, and the corresponding FWHM decreased from 101 Hz to 45 Hz and to 51 Hz. The existence of abundant 1 H would broaden the line of 77 Se spectra due to heteronuclear dipolar coupling between 1 H and 77 Se, which could be removed by 1 H decoupling. Hence, the decoupling could improve SNR significantly and reduce FWHM, which explained the improvement for both HPHD and CP because of both of them including a decoupling pulse sequence. Moreover, the improvement for the CP was also related to the efficiency of polarization transfer between 1 H and 77 Se. It should be mentioned that CP needed less time to achieve comparable results as HPHD due to the shorter D 1 (Table 1).
A pureorganic compound, PhSeO 2 H 2, was used to justify the efficiency of the three pulse sequences, and the spectra of 2 obtained were displayed in Fig. 3. As shown in Fig. 3 left, more spinning side bands were observed compared to H 2 SeO 3 1. The isotropic center-band was determined as marked by the asterisk (*), and its chemical shift was 1127.4 ppm with respect to H 2 SeO 3 ( Fig. 3 right). Similar to H 2 SeO 3 , HPHD and CP were able to enhance the relative intensity of the signal and narrow the line. The SNR was improved from 32 (SP) to 45 (HPHD) and to 35 (CP), and the FWHM was reduced from 109 Hz (SP) to 82 Hz (HPHD) and to 85 Hz (CP, Table 1). It was interesting that there was less enhancement of CP than that for H 2 SeO 3 . For H 2 SeO 3 , relatively strong hydrogen bond interactions induced the formation of well-ordered crystal and shortened the distance between Se atom and proton, whereas the presence of benzene ring in 2 distorted the hydrogen bond leading to total different crystal structure ( Fig. 1) and consequently enlarged the distance between Se and proton 50 . These led to a weaker dipolar-dipolar interaction between 77 Se and 1 H spins for 2, which explained the weaker enhancement of CP since the interaction was critical for cross polarization transfer efficiency. Moreover, low symmetry 77 Se environments would have a large anisotropic shielding, i.e.large CSA, which led to a large number of spinning side bands 51 . Thus, the more spinning side bands for 2 might be attributed to lower symmetry 77 Se environments compared to 1 as indicated from their molecular structures (Figs. 2 and 3).
The molar content of Se in polymer 6 (Scheme 1) was determined to be 1.46 mmol/g by an acid-base titration with NaOH. If all the benzenes were selenized, the Se content should be 4.33 mmol/g. Hence, only around 1/3 of benzenes in the polymer were substituted (Scheme 1). Accordingly, similar low-content of Se could be found in polymer 4 (-Se-CH 3 ), which lead to low sensitivity of 77 Se in NMR measurement. Figure 4 presents the spectra of polymer 4 obtained by the three methods. Several strong and wide spinning side bands could be observed in Fig. 4 left. The isotropic band (*) was determined by a faster speed acquisition (13 kHz), and its chemical shift was around 198 ppm ( Fig. 4 right), close to values determined by solution NMR 36,37 . In contrast to compounds 1 and 2 (Figs 2 and 3), only CP was able to effectively improve the relative intensity of the signals with SNR changing from 4 to 13 ( Fig. 4 and Table 1). No obvious change happened to the FWHM value for HPHD and CP compared to SP, which might indicate weaker heteronuclear coupling. Thus, the improvement by using CP was originated to shorter relaxation time of 1 H than 77 Se and the cross polarization transfer efficiency. As indicated from FWHM of the isotropic band became much broader than those of compounds 1 and 2, i.e. around 1600 Hz. The broadening of the 77 Se peak might be attributed to chemical shift distribution 52 . The chemical shift distribution might be originated from the change in solid-state from more ordered crystalline to amorphous structures 38,47,48 . This explanation can be well supported by PXRD results (Fig. 1). In order to determine the state of Se in different polymers, and gain effect of different substituted functional groups on the spectrum quality, polymer 5 (-SeBr) was measured by SSNMR with the same experimental parameters applied for polymer 4. Both 79 Br and 81 Br are half-integer (I = 3/2) spins with large quadrupole moments (3.3 × 10 −25 and 2.8 × 10 −25 cm, respectively), which were so large that yieldedextensive line broadening 53 . As a result, the sensitivity of Se in this polymer was too weak to give a reliable result under the same conditions as polymer 4 (Data not shown). As for the polymer 6 (-SeO 3 H, Fig. 5), a well-defined single isotropic band and rather weak spinning side bands were observed. The isotropic band was determined to be 1023 ppm in line with reports ( Fig. 5 left, *) 47 . Compared with polymer 4 (ca. 1600 Hz), the width of the isotropic bands became narrow (ca. 700 Hz, Table 1). The weak spinning side band might manifest higher symmetry 77 Se environments in polymer 6 than that in polymer 4, leading to lower CSA 51 . Thus, the relative intensity of the isotropic band was much stronger than polymer 4showing higher SNReven for SP of 11 (Table 1). Although the polymers 4 and 6 were both in amorphous state (Fig. 1), the different line widths observed indicateddifferent motional properties and chemical shift distributions of Se atomas the substituted group changed from -Se-CH 3 to -SeO 3 H 36, 37 . In addition, similar to polymer 4, only CP could significantly enhance the relative intensity of signals (Figure 5left), and correspondingly improved the SNR values from 11 to 21 (Table 1).
A comparison with the same spectrum range for polymer 4 and 6 was presented in Fig. 6. No detectable signal assigned to polymer 4 (a) could be observed in the spectrum of polymer 6 (b), indicating fully transformation  from polymer 4 to polymer 6, which manifested that SSNMR could characterize the Se state in polymers even under low content.

Conclusion
The Se state in amorphous polymers even under low content was able to be efficiently determined by SSNMR through CP methods. It was found that low symmetry of molecular structure and less long range ordering structures of the solid would broaden the signal and enlarge the spinning side bands and conseqeuntly reduce SNR. These results would provide preliminary information for further investigation about the characterization of solidified selenium-containing materials from both molecular level perspective and macrostructure perspective.   Detailed procedures for the preparation of polymer 5 from 4. 4.0 g of polymer 4 was immersed in chloroform (60 mL) under N 2 overnight. A solution of 6 mmol of Br 2 in CHCl 3 (20 mL) was added and stirred for 30 min cooled with ice-water.The resin was then washed with anhydrous EtOH and removed into a flask containing 60 mL of anhydrous EtOH. The color turned red after heating at 70 °C for 2 h and after filtration, the polymer was washed with EtOH (10 mL, twice) and CH 2 Cl 2 (10 mL, twice) subsequently and then dried at 65 °C under vacuum for 24 h to produce the polymer 5 (4.3 g).
Detailed procedures for the preparation of polymer 6 from 5. 0.71 g of polymer 5 was immersed in 60 mL CHCl 3 under N 2 overnight. 10 mmol of H 2 O 2 (30w/w %) was added at 0 °C. The color of the polymer turned white gradually after stirring at 25 °C for 2 h. After filtration, the resin was washed with EtOH (10 mL, twice) and CH 2 Cl 2 (10 mL, twice) subsequently and then dried at 65 °C under vacuum to produce the pure polymer 6 (0.6849 g). Acid-base titration with NaOH indicated that the content of Se in polymer 6 was 1.46 mmol/g. Solid-state 77 Se NMR. Solid-state 77 Se NMR was performed on Bruker Avance III spectrometers operating magnetic field strengths of 9.4 T, corresponding to Larmor frequencies at 76.3 MHz for 77 Se. A Bruker 4 mm double resonance HX MAS probe was used. H 2 SeO 3 was used as a secondary external reference standard for PhSeO 2 H and polymers, and its isotropic chemical shift (δ) was calibrated to be 1288.1 ppm with respect to dimethylselenide (0.0 ppm). Three pulse sequences were applied (Fig. 8): 1, a single pulse (SP). 2, a single pulse with high power 1 H decoupling (HPHD), and the composite pulse 1 H decoupling program is spinal64.The decoupling power was set to be 66.72 W with the pulse length of 9.0 μs. 3, cross polarization (CP) transfer from 1 H spin using optimized contact pulse durations of 4.5 ms or 5.0 ms (ramped for 1 H), and two-pulse phase modulation (TPPM) 1 H decoupling during acquisition (spinal64). The decoupling power was set to be 66.72 W with the pulse length of 8.4 μs. 90° radio frequency (rf) pulse length for 1 H excitation was 4.2 μs. In order to shorten recycle delay time (D 1 ) between each acquisition, a 30°rf pulse of 1.54 μs (D 1 , ca. 0.1T 1 ) was applied instead of the normally used 90° rf (D 1 , ca. 5T 1 ) for 77 Se excitation in SP and HPHD pulse sequences. The applied D 1 are listed in Table 1. The acquisition time for all of the three pulse sequences was set to be 4.56 ms. Number of scans was varied from 16 to 17408 depending on the intensity of the samples. The data was acquired with spinning speed from 8 kHz to 13 kHz. The experiment durations of compound H 2 SeO 3 and PhSeO 2 H are 12~15 min and 10~15 h, respectively, and the duration for the polymers is 24 h.