Section A1: Powder XRD Characterization of CPO-27-Mg Samples
The identity and purity of CPO-27-Mg samples were confirmed by powder XRD. PXRD patterns (Figure 2-A1) were recorded on a Rigaku diffractometer equipped with a graphite monochromator using Co Kα radiation (λ = 1.7902 Å). Diffraction data were collected from 5° to 65° in 2θ at a step size of 0.02°.
Figure 2-A1: Powder XRD patterns of CPO-27-Mg samples.
The guest contents of CPO-27-Mg samples were measured by thermogravimetric analysis (TGA). The samples were heated under N2 atmosphere on a Mettler Toledo
TGA/DTA851e instrument from 25 to 500 °C at a constant heating rate of 10 °C/min. The samples were also characterized by 13C SSNMR experiments to confirm the inclusion of guest species (Figure 2-A2).
Section A2: 13C Solid-State NMR Characterization of CPO-27-Mg Samples
Figure 2-A2:13C MAS spectra of CPO-27-Mg samples. 1H→13C CP: contact time = 10 ms. * indicates spinning sidebands. ◊: residual CH3OH.
Magic-angle spinning (MAS) 13C SSNMR spectra were collected on a Varian Infinity Plus 400 WB spectrometer operating at 100.5 MHz at a magnetic field of 9.4 T using a 4 mm triple-tuned T3 MAS probe with a spinning speed of 8 kHz. For 13C one- pulse experiments with 1H decoupling (1pda), a 30° pulse (1.5 µs) was used. The radio- frequency field for 1H decoupling was approximately 40 kHz, and the pulse delay of 13C was 30 s. In the case of 1H→13C cross-polarization (CP) experiments, the Hartmann- Hahn matching conditions were calibrated on solid admantane, which is also a secondary
reference for 13C chemical shift (δiso = 37.8 ppm for the methylene signal). Proton 90° pulse width was 5.3 µs, and a proton pulse delay of 10 s was used.
Three peaks were observed in the 13C MAS spectra of the CPO-27-Mg samples. Two of the three chemically inequivalent carbon atoms of benzene rings have very similar chemical shifts, giving rise to an overlapping peak at 127 ppm, which could not be resolved under the experimental conditions we used. The adsorbed acetone and acetonitrile were also confirmed.
Section A3: Additional 25Mg Static Spectra of CPO-27-Mg Samples
Due to the spectrometer-time limit, 25Mg spectra in Figure 2-2 and 2-4 were all recoded with 16384 scans. To ensure all the spectral features were captured, we selected two (dehydrated and 0.6H2O/Mg CPO-27-Mg) samples and obtained their spectra with a better S/N ratio by acquiring them with significantly more scans (65536 and 49152 scans for dehydrated and 0.6H2O/Mg CPO-27-Mg, respectively). The resulting spectra are compared to those obtained with 16384 scans in Figure 2-A3. As one can see clearly, no additional feature is identified.
Figure 2-A3:25Mg static spectra of selected CPO-27-Mg samples at 21.1 T. Section A4: Theoretical Calculations of 25Mg EFG Parameters
The calculated 25Mg EFG parameters of as-made and dehydrated CPO-27-Mg were listed in Table 2-A1. Although the DFT calculation overestimated the CQ of as-
made CPO-27-Mg, the results do confirm the predication that the CQ(25Mg) in dehydrated sample is significantly larger than that of the as-made sample.
Table 2-A1: Calculated 25Mg EFG tensors of as-made and dehydrated CPO-27-Mg.
Sample |CQ| (MHz) ηQ δiso (ppm)
As-made 9.11 0.46 0.54
Dehydrated 14.10 0.32 6.50
Figure 2-A4: Two types of distortion used in the DFT calculations to describe the effect of bond angle. Arrows indicate the change of guest locations.
The acetone was chosen to represent the organics loaded inside the CPO-27-Mg samples. The effect of acetone orientations was probed by systematically: (1) varying the O1–Mg–O4 angle from 92.8° while keeping the Mg–O distance at 2.14 Å (the values taken from the structure of as-made CPO-27-Mg). See the distortion 1 in Figure 2-A4; (2) altering the C1–O1–Mg–O4 dihedral angle from 49.6° in the as-made sample. See the distortion 2 in Figure 2-A4. The results are shown in Figures 5-A5. In general, the effect of the Mg–O bond length on NMR parameters is much larger than that of the bond angle. Therefore, the observed distribution of the CQ(25Mg) is likely due to the distribution of the Mg–O distances.
Figure 2-A5: The plot of calculated CQ(25Mg) as a function of: (a) O1–Mg–O4 angle from 92.8° and (b) C1–O1–Mg–O4 angle from 49.6°.
Chapter 3
3
Resolving Multiple Non-Equivalent Metal Sites in
Magnesium-Containing Metal–Organic Frameworks by
Natural Abundance
25Mg Solid-State NMR
Spectroscopy
†3.1 Introduction
Metal–organic frameworks (MOFs) are a group of novel inorganic–organic hybrid porous materials. Because of their many unique properties including rich structural diversity, large surface area, tunable porosity, high thermal stability and selective adsorption, MOFs are suitable for a broad range of applications, in particular for gas separation and storage.1 In recent years, incorporating Mg2+ into MOFs has drawn much attention since it is inexpensive, nontoxic, and especially, has low atomic weight.2 Structural characterization is very important for these MOFs. Unfortunately, due to the difficulty in obtaining suitable single crystals for X-ray diffraction, many MOFs’ structures were determined from more limited powder XRD data. In such cases an unambiguous structure solution requires additional information from complementary techniques such as solid-state NMR (SSNMR) spectroscopy. The number of non- equivalent metal centers is usually determined by crystal symmetry. Therefore, it is desirable to directly determine the number of non-equivalent Mg sites by 25Mg SSNMR spectroscopy. However, although recent work3-17 has demonstrated that 25Mg SSNMR spectroscopy can be employed as a powerful tool to characterize Mg-containing minerals, organometallics, and biomolecules, directly probing the local Mg structure in MOF-based materials by natural abundance 25Mg SSNMR spectroscopy is still rare.18 This is due to several reasons: 1) the low intrinsic sensitivity arising from the unfavorable 25Mg (I = 5/2) nuclear properties, such as a small gyromagnetic ratio (γ) of -1.639 × 107 rad·s-1·T-1 and a relatively low natural abundance of 10.0%,19 2) a relatively large quadrupole moment of
†
A version of this chapter (except the data of the acetone sample) has been published elsewhere: Xu, J.; Terskikh, V. V.; Huang, Y. Chem. Eur. J. 2013, 19, 4432-4436. Reproduced by the permission of John Wiley and Sons.
199.4 mb,20 3) a very low 25Mg concentration due to low densities of MOFs. For example, the number of 25Mg atoms per nm3 in microporous α-Mg3(HCOO)6 (a MOF examined in this work) is only 0.73 compared to the number for the dense MgO at 5.3. Furthermore, 25
Mg has a narrow chemical shift range, which makes it challenging to differentiate multiple Mg sites with very similar local environments. Nevertheless, several recent studies have demonstrated that the low-sensitivity associated with low-γ unreceptive quadrupolar nuclei in MOFs can be alleviated by performing NMR experiments at high magnetic fields.3-6,18,21
Herein, using microporous α-Mg3(HCOO)6, a representative Mg-containing MOF as an example, we demonstrate that multiple (four) non-equivalent Mg sites with very similar local structures can be directly differentiated by 25Mg natural abundance two- dimensional triple-quantum magic-angle spinning (3QMAS)22 at a magnetic field of 21.1 T in combination with theoretical calculations using the density functional theory (DFT) gauge including projector augmented wave (GIPAW) method.23,24, 3a-d
Microporous α-Mg3(HCOO)6 is an important commercialized MOF (trade name Basosive M050).25-27 It can be facilely prepared on a large scale under solvent-free conditions using low-cost and nontoxic starting materials such as MgO and formic acid. It displays permanent porosity up to 400 °C after desolvation (activation) as well as exceptional stability in many solvents. Microporous α-Mg3(HCOO)6 has been demonstrated to have great potential in gas separation such as separating C2H2 from CO2 and CO from H2.28-30 Another reason for choosing this MOF is because good single crystal structures25 are available, against which the viability of NMR approach can be checked.
In this work, we examined natural abundance 25Mg SSNMR spectra of four samples of microporous α-Mg3(HCOO)6: as-made (i.e., containing solvent DMF molecules), activated (empty framework), acetone- and benzene-loaded phases (hereafter referred to as DMF, activated, acetone and benzene samples, respectively).