Y. Yang et al. / Journal of Catalysis 365 (2018) 43–54
45
as shown in Fig. S1. We observed a similar behavior for the amine-
functionalized SBA-15 silica materials.
ing successful incorporation of organic moieties into the silica
framework. An additional peak at -16 ppm assignable to
ASiAOASiMe3 functionalities on P-SBA15 was observed for sam-
ples passivated with chlorotrimethylsilane prior to reaction with
LiPPh2. The subsequent immobilization of RhCl(PPh3)3 on the
surface-functionalized SBA-15 causes marginal changes in the
29Si CP-MAS spectra for the Tm and Qn groups, as shown in
Fig. 1B, C, and D (spectra b, d, f). It is unclear why T-groups for
the P-SBA-15 and Rh-P-SBA-15 (Fig. 1) are so weak, but all other
characterizations and reactivity studies are consistent with RhCl
(PPh3)3 grafted to phosphine.
Functionalization and immobilization of Wilkinson’s complex
caused a reduction in the surface area, total pore volume, and
mean pore size compared with pure SBA-15, as listed in Table S1.
The low-angle XRD pattern (Fig. S2) for all samples showed three
well-resolved peaks in the region of 0.6–2° indexed to (1 1 0), (2
0 0), and (2 1 1) reflections of hexagonal mesoporous arrays and
a significant decrease in their reflection intensities compared with
pure SBA-15. A positive shift in peak position for the P-SBA-15 and
Rh-P-SBA-15 samples was observed relative to the parent SBA-15
due to the increased thickness of the pore wall, demonstrating
the organic functional groups and rhodium complex were grafted
predominantly onto the internal surface of the pore. The HR-TEM
images in Fig. S3 clearly demonstrate the mesoporous channels
were preserved upon functionalization and immobilization of
Wilkinson’s complex and no metallic Rh nanoparticles were
formed.
In order to further characterize differences in SBA-15 before and
after functionalization and immobilization, we conducted 13C and
29Si solid-state NMR experiments. Fig. 1 shows the 29Si CP-MAS
NMR spectra for pure SBA-15, N-SBA-15, 2N-SBA-15, and P-SBA-
15, respectively. As shown in Fig. 1A, two signals around À101
and À110 ppm for the pure SBA-15, characteristic of Q3 and Q4 sil-
icon sites of the SiO4-substructures (Qn = Si(OSi)n(OH)4-n, n = 2–4)
are present. The structural changes of the silica after the function-
alization are visible in Fig. 1B, C, and D (spectra a, c, e). An addi-
tional set of peaks between À50 and À70 ppm, assignable to Tm-
site groups (Tm = RSi(OSi)m(OH)3-m, m = 1–3) are present, indicat-
Fig. 2 summarizes the 13C CP-MAS NMR spectra for the SBA-15
upon organic functionalization and the corresponding catalysts
after subsequent immobilization of RhCl(PPh3)3. Peaks correspond-
ing to the pure organic functional linkers (Fig. 2A, spectrum a, 2B,
d; and 2C, g) appear in the 13C solid-state NMR spectra of the
organically functionalized SBA-15 samples and their respective
immobilized Rh complexes, indicating the successful grafting and
structural retention of the organic linkers on the silica surface.
Using N-SBA-15 as an example, the appearance of peaks at d =
10.8, 18.6, 27.9, 45.4, and 58.6 ppm, corresponding to ASiCH2,
AOCH2CH3, ACH2CHCH2, ACH2NH2, and AOCH2CH3, respectively,
demonstrates the successful grafting of 3-aminopropyl linkers to
the SBA-15 surface through the condensation reaction between
AOH and ASiOCH2CH3 (Fig. 2A, spectrum b). After subsequent
immobilization of RhCl(PPh3)3, no significant differences in the
13C CP-MAS NMR spectra were found among the catalysts with
the exception of the appearance of a broad peak around d = 128–
132 ppm, assignable to the phenyl group of triphenylphosphine
ligands coordinated to rhodium.
Notably, in the case of Rh-P-SBA-15 (Fig. 2C, spectrum k), the
signal intensity of the phenyl group ranging from 128 to 132
ppm significantly increases; while the same signal for Rh-2N-
SBA-15 (Fig. 2B, spectrum f) is virtually absent in the 13C CP-MAS
spectra. The most likely reason for this observation is due to the
complete replacement of the PPh3 ligand with the secondary-
amine groups on the silica surface. Collectively, these results
demonstrate the organic functional linkers and RhCl(PPh3)3 were
successfully immobilized onto the surface of SBA-15. The results
thus far demonstrate differences among grafted samples, but do
not provide significant insight into the local structures of the indi-
vidual immobilized Rh centers.
2.2. Local structure of the immobilized Rh complex on SBA-15
Various attempts towards the covalent immobilization of
Wilkinson’s complex, RhCl(PPh3)3, on a variety of solid supports
have been undertaken and the resulting catalysts demonstrate
high activity and stability during catalysis [27–34], but insuffi-
cient attention has been paid to the structure of the heteroge-
nized catalyst. We deem this a critical element in the synthesis
of supported molecular catalysts because knowledge of the local
structure allows for the elucidation of structure-function relation-
ships between activity/selectivity and the structure of the Rh cen-
ter. To address this shortcoming, we applied
a series of
techniques in concert to obtain a better understanding of the
chemical environment of surface-supported Rh complexes includ-
ing 31P CP-MAS NMR, 2D 31P{1H} HETCOR NMR, XPS, and Rh K-
edge EXAFS.
2.2.1. Solid-State 31P NMR characterization of immobilized Rh
catalysts
Fig. 3 shows the 31P CP-MAS NMR spectra for the as-prepared
catalysts, Rh-N-SBA-15, Rh-2N-SBA-15, and Rh-P-SBA-15, respec-
tively. The spectrum of Wilkinson’s complex upon immobilization
on the surface-functionalized SBA-15 differs significantly from the
Fig. 1. 29Si CP-MAS NMR spectra for (A) SBA-15; (B) a: N-SBA-15, b: Rh-N-SBA-15;
(C) c: 2N-SBA-15, d: Rh-2N-SBA-15; (D) e: P-SBA-15 passivated with ClSiMe3, f:
Rh-P-SBA-15 passivated with ClSiMe3.