Immobilization of Carbonylnickel Complexes: A Solid-State NMR Study

Article abstract of DOI:10.1021/ic950756c

Silica is commonly modified with bifunctional phosphines like PPh₂CH₂CH₂Si(OEt)₃ prior to immobilization of catalysts. Here, besides PPh₂(CH₂)₃Si(OEt)₃, ligands that are more stable toward oxidation, namely PPh₂(C₆H₄)-Si(OEt)₃, PPh₂(C₆H₄)SiMe₂(OEt), and PPh₂(CH₂)₄OH, are applied. The di- and tricarbonylnickel complexes of these ligands are synthesized, characterized by ⁶¹Ni, ³¹P, ¹³C, and ¹H NMR and IR spectroscopy, and immobilized on silica. Alternatively, the corresponding phosphine-modified silicas are treated with Ni(CO)₄. The obtained materials are investigated by ³¹P and ¹³C solid-state NMR spectroscopy. Di- and tricarbonylnickel species can be distinguished by their different chemical shift anisotropies and also by their IR data. While the first route allows immobilization of both di- and tricarbonylnickel complexes, reaction of surface-bound phosphines with Ni(CO)₄ gives the tricarbonyl species exclusively. Among the different ligands, PPh₂(CH₂)₄OH has optimal qualities: It gives high surface coverages without oxidation or cross-linking, and therefore narrow NMR lines, and it allows ¹³C CP/MAS NMR spectroscopy.

Full text of DOI:10.1021/ic950756c

1814
Inorg. Chem. 1996, 35, 1814-1819
Immobilization of Carbonylnickel Complexes: A Solid-State NMR Study
K. D. Behringer and J. Blu1mel*
Anorganisch-chemisches Institut der TU Mu¨nchen, Lichtenbergstrasse 4, 85747 Garching, Germany
ReceiVed June 15, 1995X
Silica is commonly modified with bifunctional phosphines like PPh2CH2CH2Si(OEt)3 prior to immobilization of
catalysts. Here, besides PPh2(CH2)3Si(OEt)3, ligands that are more stable toward oxidation, namely PPh2(C6H4)-
Si(OEt)3, PPh2(C6H4)SiMe2(OEt), and PPh2(CH2)4OH, are applied. The di- and tricarbonylnickel complexes of
these ligands are synthesized, characterized by 61Ni, 31P, 13C, and 1H NMR and IR spectroscopy, and immobilized
on silica. Alternatively, the corresponding phosphine-modified silicas are treated with Ni(CO)4. The obtained
materials are investigated by 31P and 13C solid-state NMR spectroscopy. Di- and tricarbonylnickel species can
be distinguished by their different chemical shift anisotropies and also by their IR data. While the first route
allows immobilization of both di- and tricarbonylnickel complexes, reaction of surface-bound phosphines with
Ni(CO)4 gives the tricarbonyl species exclusively. Among the different ligands, PPh2(CH2)4OH has optimal
qualities: It gives high surface coverages without oxidation or cross-linking, and therefore narrow NMR lines,
and it allows 13C CP/MAS NMR spectroscopy.
1. Introduction
developed. The sol-gel process for instance provides the option
of a tailored silica support material with a high loading of
catalyst. It has been successfully applied for example by
Lindner14,15 and Schubert.16,17 However, this procedure is only
feasible for very stable transition metal complexes, and not all
of the catalyst is later easily accessible for the substrate. The
steric and chemical environment of the metal nuclei is not very
well defined. On the other hand, using commercial silica offers
the advantage that even sensitive complexes can be immobilized
and all the active catalyst centers on the surface are accessible
by reagents. However, care has to be taken to avoid oxidative
side reactions of the bifunctional phosphines11 during the
immobilization step. Since triarylphosphines are less sensitive
toward oxidation, here we use, besides the conventional
PPh2(CH2)3Si(OEt)3, phosphines of this kind, namely PPh2-
(C6H4)Si(OEt)3 and PPh2(C6H4)SiMe2(OEt). The latter ligand
excludes the complicating effect of cross-linking.18-22 In order
to avoid oxidation, as well as cross-linking, and addition of the
ethoxysilyl functions to surface siloxane groups,18 PPh2(CH2)4-
OH11 is included in our study.
Immobilized catalysts are of growing interest, because they
can, in principle, combine all the advantages of heterogeneous
and homogeneous catalysis:1-4 They can easily be removed
from the reaction mixture, while the reactions themselves are
homogeneous, thus providing high selectivity and yield. Hereby,
inorganic oxides often serve as support materials, because they
are inexpensive and inert against most chemicals and elevated
temperatures. The most common support is silica. It is neutral,
and its surface properties are already well explored.5-8
Among several analytical tools,9 the most powerful method
to characterize surface-bound species is multinuclear solid-state
NMR spectroscopy.1,5,10-12 Since bifunctional phosphines11 are
often used as linkers, the spin 1/2 nucleus of choice is 31P with
its great natural abundance (100%) and large chemical shift
dispersion. The quality of the spectra can be additionally
improved by magic angle spinning (MAS)5,12 and cross-
polarization (CP)13 techniques.
For the immobilization of transition metal catalysts using
bifunctional phosphine ligands, several strategies have been
* Corresponding author. Tel.: (01149)89/3209-3108. FAX: (01149)-
89/3209-3762.
X Abstract published in AdVance ACS Abstracts, March 1, 1996.
(1) Hartley, F. R. Supported Metal Complexes; D. Reidel Publishing
Co.: Dordrecht, Holland, 1985; see also literature cited therein.
(2) Deschler, U.; Kleinschmit, P.; Panster, P. Angew. Chem., Int. Ed. Engl.
1986, 25, 236.
(3) Finch, W. C.; Gillespie, R. D.; Hedden, D.; Marks, T. J. J. Am. Chem.
Soc. 1990, 112, 6221.
(4) Alexiev, V. D.; Clayden, N. J.; Cook, S. L.; Dobson, C. M.; Evans,
J.; Smith, D. J. J. Chem. Soc., Chem. Commun. 1986, 938.
(5) Engelhardt, G.; Michel, D. High-Resolution Solid-State NMR of
Silicates and Zeolites; John Wiley: Chichester, U.K., 1987; see also
literature cited therein.
(6) Sindorf, D. W.; Maciel, G. E. J. Phys. Chem. 1982, 86, 5208.
(7) Rocha, J.; Welch, M. D.; Klinowski, J. J. Am. Chem. Soc. 1991, 113,
7100.
(8) Iler, R. K. The Chemistry of Silica; John Wiley: New York, 1979.
(9) Niemantsverdriet, J. W. Spectroscopy in Catalysis; VCH: Weinheim,
Germany, 1995.
(10) Bell, A. T., Pines, A., Eds. NMR Techniques in Catalysis; Marcel
Dekker: New York, 1994.
(11) Blu¨mel, J. Inorg. Chem. 1994, 33, 5050.
(12) Fyfe, C. A. Solid-State NMR for Chemists; CFC Press: Guelph,
Canada, 1983; see also literature cited therein.
(13) Pines, A.; Gibby, M. G.; Waugh, J. S. J. Chem. Phys. 1973, 59, 569.
These ligands are successfully applied for the synthesis and
immobilization of di- and tricarbonylnickel phosphine com-
plexes. Nickel catalysts are of interest for academic as well as
(14) Lindner, E.; Schreiber, R.; Kemmler, M.; Schneller, T.; Mayer, H. A.
Chem. Mater. 1995, 7, 951.
(15) Lindner, E.; Kemmler, M.; Mayer, H. A. Chem. Ber. 1992, 125, 2385.
(16) Schubert, U.; Egger, C.; Rose, K.; Alt, C. J. Mol. Catal. 1989, 55,
330.
(17) Breitscheidel, B.; Zieder, J.; Schubert, U. Chem. Mater. 1991, 3, 559.
(18) Blu¨mel, J. J. Am. Chem. Soc. 1995, 117, 2112.
(19) Sindorf, D. W.; Maciel, G. E. J. Am. Chem. Soc. 1983, 105, 3767.
(20) Gilpin, R. K.; Gangoda, M. E. J. Chromatogr. Sci. 1990, 28, 277.
(21) Bayer, E.; Albert, K.; Reiners, J.; Nieder, M.; Mu¨ller, D. J. Chro-
matogr.1985, 264, 197.
(22) Liu, D. K.; Wrighton, M. S.; McKay, D. R.; Maciel, G. E. Inorg.
Chem. 1984, 23, 212.
0020-1669/96/1335-1814$12.00/0 © 1996 American Chemical Society

Immobilization of Carbonylnickel Complexes
Inorganic Chemistry, Vol. 35, No. 7, 1996 1815
1
Table 1. 61Ni and 31P Chemical Shifts, Line Widths ν1/2(61Ni), Coupling Constants J(61Ni-31P), and ν(CO) Values for Di- and
Tricarbonylnickel Complexes at Room Temperature (298 K)
61
ν
( Ni)
1J(61Ni-31P)
1/2
compounda
no.
δ(61Ni)b
[Hz]c
[Hz]d
δ(31P)f
ν(CO) [cm-1
]
(CO)3Ni[PPh2(C6H4)SiMe2OEt]
(CO)3Ni[PPh2(C6H4)Si(OEt)3]
(CO)3Ni[PPh2(CH2)3Si(OEt)3]
(CO)3Ni[PPh2(CH2)4OH]
(CO)2Ni[PPh2(C6H4)SiMe2OEt]2
(CO)2Ni[PPh2(C6H4)Si(OEt)3]2
(CO)2Ni[PPh2(CH2)3Si(OEt)3
5
6
7
8
9
10
11
24.2
22.9
8.0
26
50
180
150
1050
1150
800
234
241
218
188
e
e
e
32.1
32.1
23.6
24.5
33.7
33.8
24.9
2068/1990
2066/1993
2066/1988
2068/1993
2001/1940
2001/1942
2002/1939
7.0
93.0
87.3
31.6
a Et ) ethyl, Me ) methyl, Ph ) phenyl, (C6H4) ) para-disubstituted phenyl. b The chemical shifts of the complexes, dissolved in THF, are
measured with respect to external liquid Ni(CO)4 as the reference (δ ) 0) using the method described earlier.29 c Half-widths ν1/2 ( 5 Hz. d 1J(61Ni-
31P) ( 1 Hz. e Could not be obtained because of the great line width. f Nickel complexes dissolved in acetone-d6; chemical shifts are measured with
respect to external 85% H3PO4(aq).
commercial reasons.23,24 In particular, carbonylnickel phosphine
complexes catalyze a variety of reactions,23-25 and therefore
early attempts at their immobilization on different supports were
undertaken.26-28 However, IR spectroscopy was the only
analytical tool at that time and all research efforts were obviously
halted in their beginnings.
In this contribution, we show that the concept described
earlier,11 together with a proper choice of bifunctional phosphine
linkers, can fruitfully be applied to the immobilization of nickel
carbonyl complexes. Hereby, solid-state NMR serves as a
powerful method for characterizing the surface-attached com-
pounds. For example, early IR results of Basset,28 with a
different ligand, could be confirmed. Using a method described
elsewhere,29 it is also possible to measure the 61Ni NMR spectra
of the catalysts in solution.
Figure 1. 75.5 MHz 13C CP/MAS NMR spectrum of surface-bound
phosphine PPh2(CH2)4-O-{SiO2} (4i). Asterisks denote rotational
sidebands of the signal of the phenyl groups; the arrows mark the signals
of the butyl chain carbons. For NMR details, see the Experimental
Section.
2. Results
mmol/g of SiO2 is the highest obtainable, and additionally no
OEt signals are present to overlap the desired 13C signals of
the immobilized compounds. Therefore, the 13C CP/MAS NMR
spectrum of 4i can easily be recorded even at a low spinning
rate, and all 13C signals of the butyl chain are visible (Figure
1): The signals of the phenyl groups at 129 ppm are not
resolved. The chemical shift values of 63.3, 34.1, 28.2, and
20.8 ppm for the alkyl chain correspond very well to those of
the free phosphine of 61.48, 33.77, 27.52, and 22.07 ppm (for
assignment see Experimental Section).
2.2. Carbonylnickel Phosphine Complexes. Di- and tri-
carbonylnickel complexes can be synthesized in quantitative
yield by reaction of Ni(CO)4 with the corresponding phosphines,
using the literature procedure (Scheme1; Experimental Section).
Since the 31P NMR chemical shifts of di- and tricarbonylnickel
complexes are very similar (Table 1), 61Ni NMR spectroscopy29
is a useful analytical tool to control the success of the
syntheses: As shown earlier,29 tricarbonylnickel phosphine
complexes give doublets with small line widths as 61Ni NMR
signals, due to coupling of 61Ni with one 31P nucleus. Dicar-
bonyl complexes, on the other hand, lead to broad 61Ni
resonances, whose triplet character (coupling to two 31P nuclei)
can generally hardly be predicted (Figure 2).
2.1. Bifunctional Phosphine Ligands. All the applied
phosphine linkers, namely PPh2(C6H4)Si(OEt)3 (1), PPh2-
(C6H4)SiMe2OEt (2), PPh2(CH2)3Si(OEt)3 (3), and PPh2(CH2)4-
OH (4), can be prepared easily in a few steps and in good yields
by procedures given in the literature (see Experimental Section).
They are grafted with high coverage to the surface of dried silica
by a method described earlier (ref 11 and Experimental Section).
In the case of 1, 2, and 4, temperatures as high as 90 °C can be
applied for the immobilization. After drying, all the materials
are reasonably air stable and are only slowly transformed into
the phosphine oxides by oxygen in the course of days (3, 4) or
weeks (1, 2). 31P CP/MAS NMR spectroscopy shows that the
obtained phosphinated silica are free of side products. The
signals of surface-immobilized 1 (1i) and 2 (2i) resemble those
of immobilized 3 (3i) and 4 (4i).11 Their signal shifts correspond
to the solution chemical shifts very well (see Experimental
Section). The 13C CP/MAS NMR spectra show in the case of
1i, 2i, and 3i, besides the dominant signals of the aryl groups,
huge resonances for OEt groups. The latter most often prevent
the detection of the signals of the alkyl chain groups. The
surface-attached ligand 4i has optimal qualities for solid-state
13C NMR measurements: The surface coverage of about 0.4
All the 61Ni NMR and IR data for the molecular complexes
5-11 are given in Table 1. The 61Ni chemical shifts, coupling
constants, and line widths of 5, 6 and 9, 10 correspond very
well to the shifts of (CO)3Ni(PPh3) and (CO)2Ni(PPh3)2,29
respectively. The 61Ni NMR data of 7, 8, and 11 are in accord
with those of the corresponding butyl- and ethyldiphenylphos-
phine nickel complexes.29 The IR data agree with those of
similar di- and tricarbonylnickel complexes very well.25,27,28
2.3. Immobilization of the Carbonylnickel Phosphine
Complexes. All the molecular nickel complexes 5-11 can be
immobilized on silica under the same mild reaction conditions
(23) Jolly, P. W.; Wilke, G. The Organic Chemistry of Nickel; Academic
Press: New York, 1985; Vol. II.
(24) Pignolet, L. H., Ed. Homogeneous Catalysis with Metal Phosphine
Complexes; Plenum Press: New York, London, 1983.
(25) Gmelin Handbuch der Anorganischen Chemie, Erga¨nzungswerk zur
8 Auflage; Springer: Berlin, 1975; Vol. 16, Part 1, p 133.
(26) Evans, G. O.; Pittman, C. U.; McMillan, R.; Beach, R. T.; Jones, R.
J. Organomet. Chem. 1974, 67, 296.
(27) Beck, W.; Ho¨fer, R.; Erbe, J.; Menzel, H.; Nagel, U.; Platzen, G. Z.
Naturforsch. 1974, 29B, 567.
(28) Smith, A. K.; Basset, J. M.; Maitlis, P. M. J. Mol. Catal. 1977, 2,
223.
(29) Behringer, K. D.; Blu¨mel, J. Magn. Reson. Chem. 1995, 33, 729.

1816 Inorganic Chemistry, Vol. 35, No. 7, 1996
Behringer and Blu¨mel
Table 2. MAS NMR and IR Data for Surface-Attached Metal
Complexes
ν1/2
(31P)
ν(CO)
compounda
no. δ(31P)b [Hz]c
[cm-1
]
(CO)3Ni[PPh2(C6H4)SiMe2O-{SiO2}] 5i
31.3
31.6
22.4
23.5
31.3
330 2069/1999
400 2068/1998
540 2068/1999
400 2069/2000
745 2001/1939
610 2005/1943
830 2003/1939
(CO)3Ni[PPh2(C6H4)Si{O-SiO2}3]
(CO)3Ni[PPh2(CH2)3Si{O-SiO2}3]
(CO)3Ni[PPh2(CH2)4O-{SiO2}]
6i
7i
8i
(CO)2Ni[PPh2(C6H4)SiMe2O-{SiO2}]2 9i
(CO)2Ni[PPh2(C6H4)Si{O-SiO2}3]2
(CO)2Ni[PPh2(CH2)3Si{O-SiO2}3]2
10i 32.3
11i 22.9
Figure 2. 26.6 MHz 61Ni NMR spectrum of (CO)2Ni[PPh2(C6H4)Si-
(OEt)3]2 (10) in THF. t denotes the doublet signal of traces of the
tricarbonylnickel complex 6.
a {SiO2} denotes the silica support. b External reference: 85%
H3PO4(aq). ν1/2 ( 5 Hz.
c
from the di- to the tricarbonyl species. The 31P CP/MAS spectra
of 9i-11i show an unsymmetric overall appearance (Figure 3A),
whereas the spectra of 5i-8i are nearly totally symmetric (Figure
3B/C) (the use of “symmetric” or “unsymmetric” in this context
should not be confused here or below with the symmetry or
asymmetry of the 31P shielding tensor12). The CSA values have
been determined by a nonspinning NMR experiment for two
representative cases: σ11, σ22, and σ33 for 7i are 98, 24, and
-32 ppm, respectively; the corresponding σ values for 11i are
85, 34, and -39 ppm, respectively.
The residual line width is greater for immobilized dicarbonyl
than for tricarbonyl complexes (Table 2). The narrowest signal
can be observed for 8i.
The ν(CO) values do not change significantly when the free
nickel complexes (Table 1) are bound to silica (Table 2).
2.4. Reaction of Ni(CO)4 with Phosphinated Silica. After
the surface-bound phosphines 1i-4i are treated with Ni(CO)4
(Scheme 1), the obtained 31P CP/MAS NMR spectra of the
materials are identical with those of 5i-8i (Figure 3B/C). This
result is independent of the applied Ni:P ratio, provided a
stoichiometric amount of Ni(CO)4 is added. No resonances
indicative of oxidized species, free phosphines, or dicarbon-
ylnickel complexes can be detected. Therefore, reaction of
surface-bound phosphines with Ni(CO)4 gives the corresponding
immobilized tricarbonylnickel complexes exclusively. This
result is additionally supported by the IR data (Tables 1, 2) of
the materials.
2.5. Reaction of (CO)2Ni(PPh3)2 with Phosphinated Silica.
When the immobilized phosphines 1i-4i are treated repeatedly
with fresh phosphine-free (CO)2Ni(PPh3)2, the resulting 31P CP/
MAS NMR spectra are identical with those of 9i-11i (Figure
3A). Again, no resonances indicative of other than the
immobilized dicarbonylnickel phosphine complexes are visible.
Identical spectra are obtained, however, when an excess of
(CO)2Ni(PPh3)2 is applied to phosphinated silica only once.
Figure 3. 121.5 MHz 31P CP/MAS NMR spectra of (A) (CO)2Ni-
[PPh2(CH2)3Si{O-SiO2}3]2 (11i), (B) (CO)3Ni[PPh2(CH2)4O-{SiO2}]
(8i), and (C) (CO)3Ni[PPh2(CH2)3Si{O-SiO2}3] (7i). Spinning sidebands
are denoted by asterisks; for NMR details, see the Experimental Section.
The dots designate the signal of trace amounts of an oxidic impurity.
3. Discussion
that are applied for grafting the ligands on the surface
(Experimental Section and ref 11), yielding the surface-attached
species 5i-11i. The success of the procedure is again best
controlled by 31P CP/MAS NMR spectroscopy. The δ(31P)
values of 5-11 in solution (Table 1) correspond very well to
the chemical shifts of 5i-11i in the solid state (Table 2).
Typical 31P CP/MAS NMR spectra of immobilized di- and
tricarbonylnickel complexes are displayed in Figure 3. No
signals of uncomplexed phosphines are visible. Spectrum 3A
demonstrates that even trace amounts of oxidic impurities show
up in the 31P CP/MAS NMR spectra very well. While the δ-
(31P) values of di- and tricarbonyl complexes are very similar,
the overall shapes of the signals are different: Their chemical
shift anisotropy (CSA)5,12 and thus their pattern of rotational
sidebands, denoted by asterisks in the spectra, change on going
3.1. Bifunctional Phosphine Ligands. The above results
show, that on the basis of ref 11, a variety of bifunctional triaryl-
and diarylalkylphosphines can be grafted onto a silica surface
without side reactions of any kind. While 31P CP/MAS NMR
spectroscopy is successfully applied to all the immobilized
phosphines, 13C CP/MAS NMR is somewhat hampered by the
signals of ethoxy groups in the case of 1i, 2i, and 3i. These
signals could stem from residual ethoxy groups that have not
undergone a condensation with surface silanol groups. The
presence of OEt groups in the case of 2i, however, suggests
that an addition of ethoxysilyl groups to surface siloxane groups
also takes place. This type of surface reaction has been
demonstrated to be even dominant in an earlier study of the
reactions of ethoxysilanes with silica.18 With respect to solid-
state 13C NMR spectroscopy, ligand 4 is superior to the others,

Immobilization of Carbonylnickel Complexes
Inorganic Chemistry, Vol. 35, No. 7, 1996 1817
Scheme 1. Generally Valid Immobilization Reactions for
width increases dramatically on going from the tri- to the
dicarbonylnickel species (Table 1). We interpret this as the
result of a less symmetric electronic environment around the
quadrupolar 61Ni nucleus, and hence a greater electric field
gradient qzz.29 A comparison of the ν1/2(61Ni) values of
compounds 5, 6 and 9, 10 with the corresponding values of
(CO)3NiPPh3 (16 Hz)29 and (CO)2Ni(PPh3)2 (750 Hz)29 shows
Di- and Tricarbonylnickel Complexesa
29
that the greater electric field gradient qzz might also play a
role here. The line widths increase with the number of ethoxy
groups present. The ν1/2 values are greater for compounds with
Si(OEt)3 groups than for those with SiMe2(OEt) groups. One
might suggest that the former is a more effective anchor than
the latter, leading to a longer correlation time due to the greater
steric demand. Comparison of the δ(61Ni) values of 5, 6 and
9, 10 with those of (CO)3NiPPh3 and (CO)2Ni(PPh3)2, respec-
tively, shows, that the ethoxysilane function has only little
influence on the chemical shift.
Regarding the greater line widths ν1/2(61Ni) of compound 8
as compared to 5, one might suppose an intramolecular Ni-O
bond and therewith increased distortion. This is for example
observed in the case of ether phosphine complexes of ruthe-
nium.14,32 However, then the δ(61Ni) and ν1/2(61Ni) values should
be very different from the ones of the oxygen-free model
compound PPh2Et (-0.1 ppm, 131 Hz),29 which is obviously
not the case. Additionally, the ν(CO) values of 5 and 8 are
very similar (Table 1).
Unfortunately, due to the NMR properties of the 61Ni nucleus,
measurements in the solid state are at present not possible.29
3.3. Immobilization of the Carbonylnickel Phosphine
Complexes. The results described under section 2.3 have
demonstrated that all the di- and tricarbonylnickel complexes
with different types of bifunctional ligands can be immobilized
without side reactions. Since there are no resonances of free
phosphines visible in the 31P CP/MAS NMR spectra (Figure
3), we conclude that no ligand dissociation reactions occur
during the immobilization procedure. The spectra can also be
recorded with smaller rotors and higher spinning speeds, in order
to remove all rotational sidebands. However, much information
would be lost, since the CSA can be estimated from the
intensities of the spinning sidebands.12 Here, for example, it is
possible to judge from the overall appearance of the spectra,
recorded with a moderate spinning speed (4 kHz), whether a
di- or tricarbonyl species is present. All di- and tricarbon-
ylnickel species display similar 31P signal shapes, respectively.
The residual line widths of the immobilized dicarbonylnickel
complexes are always greater than those of the corresponding
tricarbonylnickel complexes. This might reflect the reduced
mobility12 and diminished ordering in the surroundings of 31P
of the chelate-type bound dicarbonyl- in contrast to the
tricarbonylnickel complexes.
The smallest line width is obtained for 8i, and we tentatively
suggest that the local environment is most ordered here due to
a “brush” type configuration31 of the unbranched ligands on
the surface.
While the ν(CO) values are different for di- and tricarbon-
ylnickel complexes, the IR data do not change significantly when
the molecular nickel complexes (Table 1) are bound to the silica
surface (Table 2). Therefore, we conclude that the electronic
environment of the complexes is not changed by the im-
mobilization.
a Ligand 2 is depicted as an example. Details of the reaction
conditions are given in the Experimental Section. {SiO2} denotes the
silica support.
since there are no overlapping signals of OEt groups after
immobilization. Additionally, the high surface coverage is
advantageous for the sensitivity-limited 13C CP/MAS NMR
method. We interpret the higher coverage of 4i as compared
to 1i-3i as the result of denser packing of the ligands on the
surface due to the minor steric demand of the long unbranched
alkyl chain. In comparison, studies of the esterification of silica
surfaces with alcohols8,30 have shown that the surface coverages
are densest for primary, unbranched alcohols like 1-butanol.30
Therefore, although in the meantime a plethora of methods have
been developed for modifying silica,31 the esterification via the
hydroxyl group of an alcohol still presents a very easy
alternative. The water liberated by the condensation reaction
of 4 with surface silanol groups can easily be removed in vacuo,
while there is at present no way to selectively remove residual
or surface-bound ethoxy groups in the case of 1-3. Addition-
ally, oxidative side reactions due to the combined action of
ethoxy groups and silica surface do not take place.11 The
rotational sidebands of the 13C phenyl signal (Figure 1) could
be removed entirely by faster spinning rates (>10 kHz), if
necessary.
3.2. Carbonylnickel Phosphine Complexes. The results
described under section 2.2 have shown that 61Ni NMR
spectroscopy can routinely be used in order to check the purity
of the obtained compounds. As in the case of other nickel
complexes with monofunctional phosphine ligands,29 the line
3.4. Reaction of Ni(CO)4 with Phosphinated Silica. Reac-
tion of Ni(CO)4 with surface-immobilized phosphines gives
selectively the silica-bound tricarbonylnickel complexes (Scheme
1). This has already been anticipated by Basset, using IR
spectroscopy.28 Our solid-state NMR study shows that this
(30) Ballard, C. C.; Broge, E. C.; Iler, K. R.; St. John, D. S.; McWhorter,
J. R. J. Phys. Chem. 1961, 65, 20.
(31) Scott, R. P. W. Silica Gel and Bonded Phases; John Wiley & Sons:
New York, 1993.
(32) Lindner, E.; Kemmler, M.; Mayer, H. A.; Wegner, P. J. Am. Chem.
Soc. 1994, 116, 348.

1818 Inorganic Chemistry, Vol. 35, No. 7, 1996
Behringer and Blu¨mel
5.1. Liquid-State NMR Spectroscopy. All the spectra were
recorded on a Bruker MSL 300 NMR spectrometer, equipped with a
multinuclear 10 mm broad-band probehead. The spectra were recorded
with acetone-d6 as the solvent, if not stated otherwise. The measured
reaction behavior is general and independent of the type of
phosphine and the reaction conditions. Additionally, no signals
of uncomplexed phosphines are detectable in the 31P CP/MAS
NMR spectra. This means, that all the surface-bound phos-
phinessand therewith later the nickel complexessare easily
accessible to reagents, a great advantage of the surface-
immobilization of catalysts described here in contrast to the sol-
gel process.
1
nuclei were H (300.1 MHz), 13C (75.5 MHz), 29Si (59.6 MHz), 31P
(121.5 MHz), and 61Ni (26.6 MHz). 31P and 61Ni NMR data for all
nickel complexes are given in Tables 1 and 2. Hexamethyldisiloxane
(δ(29Si) ) 6.8733 ), 85% H3PO4(aq) (δ(31P) ) 0), and Ni(CO)4 (δ-
1
(61Ni) ) 029) served as external chemical shift standards. For H and
13C measurements, the solvents were used as internal standards and
the shifts were recalculated with respect to TMS. The 61Ni NMR
spectra were recorded using a solenoid probehead and special sample
tubes described earlier.29
3.5. Reaction of (CO)2Ni(PPh3)2 with Phosphinated Silica.
Phosphine ligand exchange reactions are a very common method
for immobilizing catalysts.1 However, there is, in principle, the
possibility of 1- and 2-fold exchange.1 The latter should be
more favorable with respect to the disadvantageous effect of
“bleeding” during catalysis. The exchange rate can be deter-
mined by the ratio of excess complexed to free PPh3 in the
solution after the exchange process. However, the 31P CP/MAS
NMR spectra of the obtained modified silica did not change
with the phosphine exchange rate. Therefore, we conclude that,
for example, the surface species (CO)2Ni(PPh3)[PPh2(C6H4)-
SiMe2O-{SiO2}] cannot be distinguished from 9i by solid-state
31P NMR spectroscopy. This can also be anticipated from the
similarity of the 31P NMR signal shapes of all immobilized
dicarbonylnickel complexes, irrespective of the type of phos-
phine applied. The IR data are also not sensitive to the degree
of ligand exchange. Since ligand exchange is a tedious and
costly procedure because of catalyst wasting, we prefer the
immobilization strategy displayed in Scheme 1 for studies of
catalyst activity. Then the interpretation of data is not hampered
by ignorance of the nature of the surface-bound catalyst.
5.2. Solid-State NMR Spectroscopy. All the spectra were recorded
on a Bruker MSL 300 NMR spectrometer, equipped with a 7 mm broad-
band double-bearing MAS probehead and ZrO2 rotors. The modified
silica was loosely filled into the rotors under nitrogen. Cross
polarization (CP)13 and magic angle spinning (MAS)5,12 with a rotational
speed of 4 kHz were applied for all the spectra shown. The isotropic
lines were determined by variation of the spinning speed, if not obvious.
The contact times were 5 ms (13C) and 1 ms (31P), and the relaxation
delays were 4 s for all surface-immobilized compounds. In the case
of 31P CP/MAS NMR spectra, 500-1000 transients gave satisfactory
signal to noise ratios. Simple high power decoupling with long
relaxation delays gave identical spectra but required much more time
for measurement. The applied 31P CP/MAS NMR parameters give
optimal S/N ratios also for uncomplexed surface-bound phosphines and
P(V) side products.11 The 13C CP/MAS NMR spectrum (Figure 1)
required 10 000 scans. All spectra were recorded at room temperature
(298 K). The 31P CP/MAS NMR spectra were referenced with respect
to 85% H3PO4(aq) by setting the 31P NMR peak of solid NH4H2PO4 to
+0.81 ppm. The 13C CP/MAS NMR spectra were referenced with
respect to external adamantane.34 The Hartmann-Hahn match was
also set by using solid NH4H2PO4 and adamantane. For the exponential
multiplication, a line-broadening factor of 100 Hz was applied.
5.3. Preparation of Bifunctional Phosphines. 5.3.1. PPh2(C6H4)-
Si(OEt)3 (1). Compound 1 was synthesized in 90% yield via
(p-bromophenyl)diphenylphosphine35 by the procedure given in ref 36.
Only Si(OEt)4 was used instead of Me2Si(OEt)2. Spectral data: mass
spectrum (70 eV) m/e (relative intensity) 297 (11.4), 262 (20.7), 255
(3.3), 240 (11.1), 255 (4.4), 200 (27.2), 195 (38.6), 183 (38.0), 162
(33.1), 149 (100.0), 139 (28.1), 119 (18.8), 77 (5.8), 51 (6.0); 1H NMR
δ 7.72-7.12 (m, 14H, aryl H), 3.87 (q, J ) 7.0 Hz, 6H, OCH2), 1.20
(t, J ) 7.0 Hz, 9H, CH3); 13C NMR δ 138.25-129.4 (signals
overlapping, aryl C), 59.26 (OCH2), 18.55 (CH3); 31P NMR δ -4.53;
29Si NMR δ -54.93.
5.3.2. PPh2(C6H4)SiMe2(OEt) (2). Compound 2 was synthesized
via (p-bromophenyl)diphenylphosphine35 by the procedure given in ref
36. The 1H NMR data are in agreement with the literature data.36 Data
not given in ref 36: mass spectrum (70 eV) m/e (relative intensity)
364 (67.3, M+), 349 (4.4), 305 (12.3), 262 (100.0), 183 (82.3), 152
(8.9), 108 (34.9), 75 (7.2), 57 (6.1); 13C NMR δ 138.16 (1J(31P-13C)
) 12.0 Hz, Cipso), 137.94 (1J(31P-13C) ) 12.0 Hz, Cipso), 134.49-129.46
(overlapping, aryl C), 59.04 (OCH2), 18.79 (CH2CH3), -1.60 (SiCH3);
31P NMR δ -4.50; 29Si NMR δ 4.77.
5.3.3. PPh2(CH2)3Si(OEt)3 (3). Compound 3 was prepared from
Ph2PLi37 and (3-chloropropyl)triethoxysilane according tho the proce-
dure given by Capka.38 Since the NMR data do not seem to be literature
known yet, they are given here. The signal assignments were made
on the basis of the data of PPh3 and (nBu)3P, given in ref 39: 1H NMR
(CDCl3) δ 6.90-6.77 (m, 10H, aryl H), 3.27 (q, 6H, J ) 6.6 Hz, OCH2)
1.62 (t, J ) 7.3 Hz, 2H, PCH2), 1.09 (m, 2H, CH2CH2CH2), 0.67 (t, J
4. Conclusion
In summary, we have shown that, on the basis of earlier
studies,11,18 various bifunctional phosphines and their di- and
tricarbonylnickel complexes can selectively be immobilized on
silica without side reactions. While the free nickel complexes
can be investigated in solution by 61Ni NMR spectroscopy,29
the surface-bound species are best studied by 13C and 31P solid-
state NMR spectroscopy. The nickel complexes are easily
accessible to reagents and strongly bound to the support, as can
be demonstrated by the absence of 31P resonances of free
phosphines in the 31P CP/MAS NMR spectra. Immobilized di-
and tricarbonylnickel complexes can be clearly distinguished
by their 31P CP/MAS NMR spectra, and also by IR spectros-
copy. An early IR study by Basset,28 using a different ligand,
could be confirmed. With regard to studying catalytic processes
with the immobilized nickel complexes, ligand 4 is superior to
the others, because its surface-attached nickel derivatives give
narrow 31P NMR lines and above all 13C CP/MAS NMR
spectroscopy is possible without overlapping ethoxy signals.
Studies to compare the catalytic activities of the different
components in solution and on silica are presently underway.
5. Experimental Section
All the experiments were performed under inert gas by Schlenk
techniques. Solvents were dried by standard methods, and oxygen was
removed. The identity of molecular compounds was checked by
elemental analysis (all values were satisfactory within a range of (0.5%
for carbon and hydrogen), mass spectrometry, and their solution-state
NMR spectra. All the applied compounds are commercially available,
if not described below. All the experiments were carried out with
Merck silica 40 (specific surface area 750 m2/g; average pore size 40
Å; particle size 0.063-0.2 mm) that was dried at 600 °C in vacuo
(10-2 Pa for 12 h) in order to condense surface silanol groups.8 The
IR data were obtained after mixing an appropriate amount of free nickel
complex or supported complex with KBr and pressing the mixture into
a disk. Usually, less than 100 scans gave a satisfactory S/N ratio.
(33) Harris, R. K.; Mann, B. E. NMR and the Periodic Table; Academic
Press: New York, 1978; p 366.
(34) Hayashi, S.; Hayamizu, K. Bull. Chem. Soc. Jpn. 1991, 64, 685.
(35) Ravindar, V.; Hemling, H.; Schumann, H.; Blum, J. Synth. Com-
mun. 1992, 22, 841.
(36) Takuma, Y.; Imaki, N. J. Organomet. Chem. 1989, 379, 27.
(37) Luther, G. W.; Beyerle, G. Inorg. Synth. 1977, 17, 186.
(38) Capka, M. Synth. React. Inorg. Met.-Org. Chem. 1977, 7, 347.
(39) Kalinowski, H.-O.; Berger, S.; Braun, S. 13C NMR Spektroskopie;
Georg Thieme Verlag: Stuttgart, Germany, 1978.

Immobilization of Carbonylnickel Complexes
Inorganic Chemistry, Vol. 35, No. 7, 1996 1819
) 6.6 Hz, 9H, CH3), 0.32 (t, J ) 8.0 Hz, 2H, SiCH2); 13C NMR (CDCl3)
δ {1J(13C-1H)} 138.79 {-} (1J(31P-13C) ) 12.1 Hz, Cipso), 132.53
{161.6} (2J(31P-13C) ) 19.3 Hz, Cortho), 128.18 {164.1} (Cpara), 128.15
{159.2} (3J(31P-13C) ) 7.2 Hz, Cmeta), 58.12 {142.3} (OCH2), 31.35
{127.9} (1J(31P-13C) ) 12.1 Hz, PCH2), 19.40 {127.9} (2J(31P-13C)
) 19.3 Hz, CH2CH2CH2), 18.09 {125.5} (CH3), 12.08 {120.6} (3J(31P-
13C) ) 12.1 Hz, SiCH2); 31P NMR (CDCl3) δ -16.3.
aryl H), 3.68 (q, J ) 7.0 Hz, 4H, OCH2), 1.14 (t, J ) 7.0 Hz, 6H,
CH2CH3), 0.34 (s, 12H, SiCH3); 13C NMR δ {1J(13C-1H)} 200.35 {-}
(CO), 140 {-} (aryl C), 139.20 {-} (d, 1J(31P-13C) ) 37.0 Hz, Cipso),
137.72 (d, 1J(31P-13C) ) 30.1 Hz, Cipso), 138.10-129.02 (overlapping,
aryl C), 59.07 {141.0} (OCH2), 18.83 {124.5} (CH2CH3), -1.55
{119.1} (SiCH3).
5.4.6. (CO)2Ni[PPh2(C6H4)Si(OEt)3]2 (10). Mass spectrum (70 eV)
m/e (relative intensity) 882 (7.9), 772 (11.8), 610 (17.0), 587 (49.3),
424 (100.0), 262 (12.1); 1H NMR δ 7.59-7.10 (m, 28H, aryl H), 3.86
(q, J ) 6.9 Hz, 12H, OCH2), 1.22 (t, J ) 6.9 Hz, 18H, CH3); 13C
NMR δ {1J(13C-1H)} 200.35 {-} (CO), 140.57-128.97 (overlapping,
aryl C), 59.29 {142.9} (CH2), 18.58 {125.7} (CH3).
5.3.4. PPh2(CH2)4OH (4). Compound 4 was synthesized from
PPh2Li37 and THF using the procedure given by Garner.40 Since the
NMR data do not seem to be literature known yet, they are reported
here. Signal assignments are made as above (section 5.3.3). Addition-
1
ally, selective H decoupling was used: 1H NMR (CDCl3) δ 7.05-
5.4.7. (CO)2Ni[PPh2(CH2)3Si(OEt)3]2 (11). 1H NMR δ 7.49-7.21
(m, 20H, aryl H), 3.63 (q, J ) 7.0 Hz, 12H, OCH2), 1.70 (m, 4H,
PCH2), 1.35 (m, 4H, CH2CH2CH2), 1.12 (t, J ) 7.0 Hz, 18H, CH3),
0.40 (t, J ) 7.5 Hz, 4H, SiCH2); 13C NMR δ 201.47 (CO), 133-129
6.88 (m, 8H, Hortho, Hmeta), 6.80 (t, J ) 8.0 Hz, 2H, Hpara), 3.25 (s,
broad, 1H, OH), 3.17 (t, J ) 6.6 Hz, 2H, CH2OH), 1.69 (t, J ) 8.0 Hz,
2H, PCH2), 1.30 (quint, J ) 7.8 Hz, 2H, CH2CH2OH), 1.16 (quint, J
) 7.8 Hz, 2H, PCH2CH2); 13C NMR (CDCl3) δ {1J(13C-1H)} 138.54
{-} (1J(31P-13C) ) 13.2 Hz, Cipso), 132.30 {160.6} (2J(31P-13C) )
18.3 Hz, Cortho), 128.06 {160.6} (3J(31P-13C) ) 7.1 Hz, Cmeta), 127.85
{160.6} (Cpara), 61.48 {140.0} (CH2OH), 33.77 {123.9} (1J(31P-13C)
) 13.2 Hz, PCH2), 27.52 {128.5} (2J(31P-13C) ) 11.2 Hz, PCH2CH2),
22.07 {121.6} (3J(31P-13C) ) 16.2 Hz, CH2CH2OH); 31P NMR (THF)
δ -16.1.
5.4. Synthesis of Nickel Complexes. All the di- and tricarbon-
ylnickel complexes were prepared by reaction of Ni(CO)4 with the
appropriate amount of the corresponding phosphines at room temper-
ature according to the literature.25 All the 61Ni and 31P NMR data are
given in Table 1. The compounds are colorless or slightly cream
colored. The mass spectra generally do not give the M+ peak, due to
facile ligand dissociation.
1
(m, aryl C), 58.56 (OCH2), 33.83 (d, J(31P-13C) ) 21.1 Hz, PCH2),
19.44 (d, 2J(31P-13C) ) 7.0 Hz, CH2CH2CH2), 18.55 (CH3), 12.73 (d,
3J(31P-13C) ) 15.2 Hz, SiCH2).
5.5. Immobilization of Phosphines and Nickel Complexes. All
the phosphines and their di- and tricarbonylnickel complexes are
immobilized using the procedure given in ref 11: SiO2 (1 g) is
suspended in ca. 10 mL of toluene. A solution of 0.4 g of phosphine
or nickel complex in ca. 10 mL of toluene is added under rigorous
stirring. Then the mixture is stirred at room temperature for 12 h.
Higher temperatures should be avoided, because they do not increase
the surface coverage substantially but lead to side products11 and
decomposition of the nickel complexes with formation of Ni0. Finally,
the supernatant solution is decanted, and the silica is washed three times
with 50 mL portions of toluene and dried in vacuo (10-2 Pa) for about
5 h. Their resulting 31P CP/MAS NMR data are given in Table 2 and
ref 11. The δ(31P) values of immobilized 1i and 2i are -4.3 and -5.7
ppm, respectively.
The surface coverages were not determined by carbon elemental
analysis, because there is always a variable amount of residual and
surface-bound ethoxy groups or adsorbed solvent present (ref 18 and
section 3.1). A more reliable and reproducible method is treating a
known amount of silica with an excess of phosphine or nickel complex
and weighing the residual phosphine or nickel complex after removal
of the solvents of the supernatant solution. Typical values (in
parentheses) for the surface coverages (mmol/g of SiO2): 1i (0.2), 2i
(0.3), 3i (0.3), 4i (0.4), 5i (0.3), 6i (0.2), 7i (0.3), 8i (0.4), 9i (0.1), 10i
(0.1), 11i (0.2).
5.6. Reaction of Phosphinated Silica with Ni(CO)4. Ni(CO)4 was
condensed into a Schlenk flask, weighed, and diluted with toluene. The
appropriate amount of solution was transferred into a dropping funnel
and added slowly to a suspension of the respective phosphinated silica
in toluene under rigorous stirring. Irrespective of the amount of Ni-
(CO)4 or the dropping rate or dilution, the tricarbonylnickel complexes
were always the result exclusively.
5.7. Reaction of (CO)2Ni(PPh3)2 with Phosphinated Silica. A
suspension of phosphinated silica in toluene was stirred with a solution
of (CO)2Ni(PPh3)2 in toluene. The ligand exchange was followed by
31P NMR of the solution. The liquid phase was removed several times
and replaced by PPh3-free solutions of (CO)2Ni(PPh3)2. The 31P CP/
MAS NMR spectra of the obtained material did not change with the
degree of ligand exchange.
5.4.1. (CO)3Ni[PPh2(C6H4)SiMe2OEt] (5). Mass spectrum (70eV)
m/e (relative intensity) 364 (20.2), 262 (100.0), 183 (20.0), 154 (6.8),
1
108 (4.0); H NMR δ 7.79-7.19 (m, 14H, aryl H), 3.70 (q, J ) 7.1
Hz, 2H, CH2CH3), 1.13 (t, J ) 7.1 Hz, 3H, CH2CH3), 0.36 (s, 6H,
SiCH3); 13C NMR δ {1J(13C-1H)} 196.84 (CO), 138.22-129.75
(overlapping, aryl C), 59.11 {139.1} (OCH2), 18.77 {125.7} (CH2CH3),
-1.63 {118.9} (SiCH3).
5.4.2. (CO)3Ni[PPh2(C6H4)Si(OEt)3] (6). Mass spectrum (70eV)
m/e (relative intensity) 424 (53.1), 262 (97.2), 183 (100.0), 154 (31.5),
1
108 (28.6); H NMR δ 7.80-7.21 (m, 14H, aryl H), 3.90 (q, J ) 6.8
Hz, 6H, OCH2), 1.22 (t, J ) 6.8 Hz, 9H, CH3); 13C NMR δ {1J(13C-
1H)} 196.72 {-} (CO), 138.71-129.55 (m, aryl C), 59.31 {142.9}
(CH2), 18.54 {125.7} (CH3).
5.4.3. (CO)3Ni[PPh2(CH2)3Si(OEt)3] (7). 1H NMR δ 7.21-7.08
(m, 10H, aryl H), 3.38 (q, J ) 6.7 Hz, 6H, OCH2), 2.05 (t, J ) 7.5 Hz,
2H, PCH2), 1.24 (quint, J ) 7.5 Hz, 2H, CH2CH2CH2), 0.39 (t, J )
7.5 Hz, 2H, SiCH2); 13C NMR δ {1J(13C-1H)} 197.16 {-} (CO),
137.05 {-} (d, 1J(31P-13C) ) 31.8 Hz, Cipso), 132.57 {160.2} (d,
2J(31P-13C) ) 12.4 Hz, Cortho), 130.43 {161.6} (Cpara), 129.30 {161.6}
(d, 3J(31P-13C) ) 8.3 Hz, Cmeta), 58.65 {142.2} (OCH2), 33.87 {129.8}
1
2
(d, J(31P-13C) ) 20.7 Hz, PCH2), 19.41 {125.7} (d, J(31P-13C) )
5.5 Hz, CH2CH2CH2), 18.58 {125.7} (CH3), 12.74 {117.4} (d, 3J(31P-
13C) ) 15.2 Hz, SiCH2).
5.4.4. (CO)3Ni[PPh2(CH2)4OH] (8). 1H NMR δ 7.80-7.19 (m,
10H, aryl H), 3.60 (t, J ) 7.0 Hz, 2H, CH2OH), 3.41 (s, broad, 1H,
OH), 2.42 (t, J ) 6.9 Hz, 2H, PCH2), 1.63 (m, 2H, CH2CH2OH), 0.94
(m, 2H, PCH2CH2); 13C NMR δ 197.32 (CO), 132.9-128.5 (overlap-
ping, aryl C), 61.73 (CH2OH), 34.80 (d, 1J(31P-13C) ) 14.0 Hz, PCH2),
30.96 (d, 2J(31P-13C) ) 21.1 Hz, PCH2CH2), 22.29 (d, 3J(31P-13C) )
7.0 Hz, CH2CH2OH).
Acknowledgment. We thank the Deutsche Forschungsge-
meinschaft (DFG), the Fonds der Chemischen Industrie, the
Leonhard Lorenz Foundation and the Pinguin Foundation for
financial support.
5.4.5. (CO)2Ni[PPh2(C6H4)SiMe2OEt]2 (9). Mass spectrum (70
eV) m/e (relative intensity) 754 (2.1), 713 (2.9), 651 (7.8), 568 (13.6),
1
467 (66.8), 365 (100.0), 262 (19.0); H NMR δ 7.68-7.15 (m, 28H,
(40) Garner, A. Y.; Tedeschi, A. A. J. Am. Chem. Soc. 1962, 84, 4734.
IC950756C