$$\rightleftharpoonup{xx}$$
$$\longleftharp{xx}$$,
$$\longrightharp{xx}$$,
Firstly, preparing silica 700 as described in section 2 of the protocol, the fumed silica should be mixed with enough deionized water to make it compact, left overnight in the oven at 120 °C, and then, loaded to a quartz reactor (Figure 3). Dehydroxylated silica SiO2-700 was obtained by heating the silica gradually to 700 °C under a dynamic vacuum, FTIR spectrum for SiO2-700 in (Figure 1) shows the characteristic isolated silanol of SiO2-700.
The grafting of metal complex to the silica was achieved by a protonolysis reaction of the amido ligands (NMe2) by surface silanols. For the grafting of organometallic complexes onto highly dehydroxylated silica surface SiO2-700 (Figure 4), dehydroxylated silica was loaded with complex Hf(NMe2)4 into a double Schlenk inside a glovebox. The double Schlenk was moved out of the glovebox for the reaction to occur. The grafting reaction lasted around 3 h. Three washing cycles were then carried out by filtration; more solvent was transferred by distillation when needed. Finally, the solvent was distilled and then removed, with all volatiles, using an interceptor (solvent trap). All solids were dried by connecting the double Schlenk to the HVL (Figure 4 and Figure 5).
Elemental analysis using inductively coupled plasma (ICP) (using the EPA 3052 method for digestion) and carbon, hydrogen, nitrogen, and sulfur analyzer (CHNS) accompanied with FTIR was first verified for the determination of grafting stoichiometry. Elemental analysis showed that N/M = 3.9 (theory = 3) and C/M = 7.1 (theory = 6) for the grafted material (see Table 1); for 2, the N/M and C/M ratios were 2.5 (theory = 2) and 4.6 (theory = 4), respectively. Thus, the carbon and nitrogen contents were declined. For the preparation of the samples for the FTIR measurement, using the designed FTIR cell, see Figure 6. The first sample is the dehydroxylated SiO2-700, which displayed a characteristic peak for isolated silanol. After grafting the complex on dehydroxylated silica, this characteristic peak almost completely disappeared, and new peaks appeared in the alkyl region at 2,776-2,970 cm-1 and 1,422-1,465 cm1. After a heat treatment of prepared material at 200 °C for 1 h, its infrared (IR) spectrum showed a new peak for the imido fragment at 1,595 cm-1.
SSNMR experiments were carried out for a deeper understanding of the surface structure. Samples for SSNMR (Figure 7) were loaded in a specific rotor (32.1 µL). The 1H SSNMR spectrum of the grafted material displayed broad peaks at 2.2 and 2.7 ppm for alkyl groups bonded to the nitrogen ligands. The breadth of the resonances was expected in 1H SSNMR (in contrast to liquid NMR) and was also associated with less mobile surface species (Figure 7B).
The 13C cross-polarized magic angle spinning (CP-MAS) spectrum displayed signals that were less broadened but had a low sensitivity (alike liquid NMR). The spectrum of the grafted material revealed two overlapping peaks at 37 ppm -N-(CH3)2 and at 46 ppm, attributed to the nonequivalent methyl group in -N-(CH3) alongside a low-intensity peak at 81 ppm. A heteronuclear correlation spectroscopy (HETCOR) experiment was carried out to show the correlation between proton and carbon directly bonded to each other. The 81 ppm signal was correlated with the proton peak at 2.7 ppm in the HETCOR spectrum7. As previously reported with zirconiaaziridene, this peak represented the methylene (CH2) group in a metallaaziridine cycle39.
A multiple-quantum experiment makes it possible to visualize the correlation between proximal protons. Double-quantum (DQ) NMR indicates the sum of the two single quantum NMR frequencies to provide an autocorrelation peak located where ω1 = 2ω2. Similarly, for triple-quantum (TQ) NMR impart from the sum of three protons single quantum frequencies where ω1 = 3ω2. CH2 and CH3 afford characteristic autocorrelation peaks in the double- and triple-quantum dimensions, respectively. DQ and TQ proton SSNMR experiments were performed with the grafted material. For the strongest autocorrelation peak observed for the signal at 2.2 ppm in both the DQ and TQ spectra (appearing at 4.4 ppm and 6.6 ppm in DQ and TQ, respectively), revealing -CH3 protons, see Figure 7B. The second overlapping peak for the proton at 2.7 ppm showed an autocorrelation only in the double quantum (DQ) spectrum; thus, it confirms the presence of a methylene group (-CH2-) in the grafted species.
15N SSNMR experiments using the DNP-SENS technique (Figure 7B) were carried out to characterize the nitrogen atoms coordinated to the metal center. The spectrum obtained for the grafted material displayed two peaks, around 7 and 32 ppm. On the basis of its relative intensity, the intense signal downfield at 32 ppm was assigned to the nitrogen nuclei of the (η2-NMeCH2) and (-NMe2) functionalities. The weak upfield-shifted peak at 7 ppm was attributed to an NH(CH3)2 moiety; upon heat treatment, -HNMe2 remains on the surface.
For the imido metal fragment in catalyst 2, generated after heat treatment, one broad peak appeared at 2.2 ppm, and the weak peaks at 1.2 and 0.7 ppm in the 1H NMR spectrum attributed to some minor impurities, dimethylamine moieties. For 13C CP-MAS, NMR spectrum displays two peaks at 37 and 48 ppm. Interestingly, the peak for (-CH2-) in metallaaziridine disappeared (Figure 7C). Additionally, we inferred from the multiple quantum experiments that the 1H peak appearing at 2.2 ppm represented the (-CH3) protons. For the 15N SSNMR spectrum of the imido metal fragment in 2 (Figure 7C), an additional downfield-shifted peak at 113 ppm along with a peak at 34 ppm became much less intense after the heat treatment of the grafted material. The 113 ppm peak was assigned to the new fragment generated from the hafnium imido moieties.
In a glovebox, imine substrates with a catalyst were loaded either in an ampule tube or a sealed vial with toluene, and Figure 8 shows the reaction of imine metathesis with three imine compounds, namely N-(4-phenylbenzylidene)benzylamine, N-(4-fluorobenzylidene)-4-fluoroaniline, and N-benzylidenetert-butylamine. Mass spectra for products analyzed by GC-MS (Figure 8) are a" (1-([1,1'-biphenyl]-4-yl)-N-(tert-butyl)methanimine), c' (N,1-diphenylmethanimine), a' (1-([1,1'-biphenyl]-4-yl)-N-(3-fluorophenyl)methanimine) and b' (1-(4-fluorophenyl)-N-phenylmethanimine).
Imine metathesis is a combination of two imine substrates mixed with the catalyst to produce a new two imine substrates after characterizing by GC-MS. To calculate the conversion percentage as shown in (Table 2), use the following formulas.



Figure 1: Dehydroxylation. Formation of isolated silanols by dehydroxylation to produce silica dehydroxylated at 700 °C (SiO2-700). Please click here to view a larger version of this figure.

Figure 2: General reaction scheme of the imine metathesis mechanism7. Please click here to view a larger version of this figure.

Figure 3: Dehydroxylation of silica. The reactor was inserted into the oven and connected to a high vacuum line (HVL). The photograph shows the actual setup. Please click here to view a larger version of this figure.

Figure 4: Grafting in a double Schlenk. Schematic representation of the grafting process. The photograph shows the actual setup. Please click here to view a larger version of this figure.

Figure 5: Handling grafting in a double Schlenk with a high vacuum line (HVL). First, the solvent transfer procedure was performed, followed by the washing procedure, the solvent removal, and drying the material. Please click here to view a larger version of this figure.

Figure 6: Fourier-transform infrared spectroscopy (FTIR) measurement. (A) FTIR disc pellet preparation. (B) FTIR spectra of the support, SiO2-700 have a characteristic signal for isolated silanols observed at 3,747 cm-1 and silica undertone between 1,400-2000 cm-1. For the grafted complex, strong new signals appeared in the regions of 2,800-3,000 cm-1 and 1,400-1,500 cm-1. They represent the alkyl groups. After heat treatment, spectrum shows a new signal appeared at 1,595 cm-1 for the imido group7. Please click here to view a larger version of this figure.

Figure 7: NMR data from the grafted material and the catalyst. (A) Preparing a solid-state NMR sample. (B) (B-1) 1D 13C cross-polarized magic angle spinning (CP-MAS) with 2D 1H-13C heteronuclear correlation (HETCOR( NMR spectra of the grafted complex before treatment. (B-2) 1D 1H NMR spectrum with 1H-1H double-quantum (DQ) and 1H-1H triple quantum (TQ) spectra of the grafted complex. (B-3) 15N dynamic nuclear polarization surface enhanced the NMR spectroscopy (DNP-SENS) spectrum and (inset 1) the proposed structure of the grafted surface hafnium complex [(≡Si-O-)Hf(η2-MeNCH2)(η1-NMe2)(η1-HNMe2)]. (C) (C-1) 1D 13C CP-MAS with 2D 1H-13C HETCOR NMR spectra of the grafted complex after heat treatment to generate an imido fragment. (C-2) 1D 1H NMR spectrum with 1H-1H DQ and 1H-1H TQ spectra of the grafted complex after heat treatment which is the catalyst 2. (C-3) 15N DNP-SENS spectrum, and (inset 2) proposed a structure of the surface hafnium complex catalyst [(≡Si-O-)Hf(=NMe)(η1-NMe2)]7. Please click here to view a larger version of this figure.

Figure 8: Imine metathesis catalysis and mass spectral data of products. Three imine compounds were tested: N-(4-phenylbenzylidene)benzylamine (a), N-(4-fluorobenzylidene)-4-fluoroaniline (b), and N-benzylidene-tert-butylamine (c). Gas chromatography-mass spectrometry (GC-MS) was used to analyze the products obtained, namely 1-([1,1'-biphenyl]-4-yl)-N-(tert-butyl)methanimine (a"), N,1-diphenylmethanimine (c'), 1-([1,1'-biphenyl]-4-yl)-N-(3-fluorophenyl)methanimine (a'), and 1-(4-fluorophenyl)-N-phenylmethanimine (b')7. Please click here to view a larger version of this figure.
| Metal | %Metal | M/Silanol | %C | %N | C/N | N/M | C/M |
| 1 | Hf | 4.94 | 0.92 | 2.49 | 1.5 | 2 | 3.9 | 7.1 |
| 2 | Hf | 4.48 | 0.81 | 1.4 | 0.8 | 1.8 | 2.5 | 4.6 |
Table 1: Elemental analysis7.
| Catalyst | Time | substrate conversion % | Substrate conversion % |
| 2 | 1 | a (54) | c (54) |
| no catalyst | 1 | a (11) | c (11) |
| 2 | 6 | a (50) | b (55) |
| no catalyst | 6 | a (25) | b (20) |
| 2 | 4 | a (36) | b (30) |
Table 2: Catalysis Conversion7.