One-Pot, selective hydroformylation, condensation, and hydrogenation processes by a sol-gel entrapped rhodium complex, an immobilized base, and an ionic liquid

Article abstract of DOI:10.1002/ejoc.200801258

We report on a one-pot, multistep process in which styrene derivatives are selectively hydroformylated to give branched aldehydes, which in turn, are condensed with reactive methylene compounds (malononitrile, ethyl cyanoacetate), and then hydrogenaled. T

Full text of DOI:10.1002/ejoc.200801258

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200801258
One-Pot, Selective Hydroformylation, Condensation, and Hydrogenation
Processes by a Sol–Gel Entrapped Rhodium Complex, an Immobilized Base,
and an Ionic Liquid
Khalil Hamza,[a] Herbert Schumann,*[b] and Jochanan Blum*[a]
Keywords: Hydroformylation / Hydrogenation / Ionic liquids / Rhodium
We report on a one-pot, multistep process in which styrene
derivatives are selectively hydroformylated to give branched
aldehydes, which in turn, are condensed with reactive meth-
ylene compounds (malononitrile, ethyl cyanoacetate), and
then hydrogenated. The process takes place at 80 °C under
20.7 bar H2 and 20.7 bar CO in the presence of [Rh(cod)Cl]2
and Na[Ph2P-3-(C6H4SO3)], which have been co-entrapped
within ionic liquid confined silica sol–gel together with a
separately encaged base. The catalyst can be reused at least
four times, but the base has to be renewed after each run.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
base, which may be the ionic liquid itself, it is possible to
perform a one-pot, multistep process that consists of the
selective hydroformylation of styrenes, a base-promoted
condensation of the resulting aldehydes with active methyl-
ene compounds, and controlled hydrogenation of the unsat-
urated products so formed (Scheme 1).
Introduction
In the course of our studies on the application of the sol–
gel technology to organic synthesis,[1] we have demonstrated
that chemicals encaged within different ceramic matrices do
not interact, even if they are of opposing nature, although
they tend to react with each other in their non-heterogen-
ized form.[2,3] In a recent study, we found that dichlo-
robis[(1,2,5,6-η)-1,5-cyclooctadiene]dirhodium, [Rh(cod)-
Cl]2, that had been entrapped within an ionic liquid con-
fined silica sol–gel promotes the hydroformylation of sty-
rene derivatives in a highly selective manner to give mainly
branched aldehydes.[4] In this paper, we show that by com-
bining the aforementioned catalytic system with a suitable
Results and Discussion
The reaction of a mixture of 1 mmol styrene (1, R1 = H),
1 mmol propanedinitrile (malononitrile, 3, R1 = H, R2 =
R3 = CN), 20.7 bar H2, and 20.7 bar CO in the presence of
2.04ϫ10–2 mmol
[Rh(cod)Cl]2
co-entrapped
with
4.01ϫ10–2 mmol sodium 3-(diphenylphosphanyl)benzene-
sulfonate, ionic liquid confined silica sol–gel (vide infra),[4]
and 10 mL of each of THF and 1,2-dichloroethane af-
forded, after 14 h at 80 °C, a mixture of 86% (2-phenylpro-
pyl)propanedinitrile (5, R1 = H, R2 = R3 = CN),[5] 7% (2-
phenylpropylidene)propanedinitrile (4, R1 = H, R2 = R3 =
CN),[6] 2% (3-phenylpropylidene)propanedinitrile (8),[7]
and traces (Ͻ0.1%) of its precursor 7.[8] The latter two
products result from the linear aldehyde 6 formed in small
quantities during hydroformylation.[4] After removal of the
products and addition of a new portion of the reagents and
a base (vide infra), the used catalyst could be employed in
a second run.
Scheme 1.
[a] Institute of Chemistry, The Hebrew University,
Jerusalem 91904, Israel
Fax: +972-2-6513832
E-mail: jblum@chem.ch.huji.ac.il
[b] Institut für Chemie, Technische Universität Berlin,
10623 Berlin, Germany
The observed selectivity of two steps in the combined
process are remarkable: (a) the hydroformylation takes
place only on the double bond of the styrene, but not on
Fax: +49-30-31422168
E-mail: schumann@chem.tu-berlin.de
1502
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1502–1505

A Sol–Gel Entrapped Rh Complex in One-Pot Reactions
that of intermediate 4 (R1 = H, R2 = R3 = CN), and (b) un- recycle the system in further runs met with difficulty be-
der our reaction conditions, hydrogenation occurs only on cause of the increase in the voluminal mass of the ceramic
the internal double bond of 4 rather than on the terminal support of the base.
one of 1. Moreover, despite the fact that hydrogenation re-
In a similar manner to the transformation of 1 (R1 = H)
actions accompany most of the hydroformylation pro- to 5 (R1 = H, R2 = R3 = CN), we were able to convert the
cesses,[9] no ethylbenzene is formed as long as the pressure substituted styrenes 1 (R1 = Cl and OMe) into the corre-
of H2 and CO is roughly the same (vide infra).
sponding chloro- and methoxy dinitriles in ca. 90 and 57%
While several selective hydroformylation reactions have overall yield, respectively. We were also able to perform the
already successfully been performed with rhodium catalysts one-pot, multistep process with ethyl cyanoacetate. Some
and imidazolium salts,[10,11] the multistep, one-pot reaction typical results are summarized in Table 1.
outlined in Scheme 1 takes place efficiently only when the
rhodium catalyst is entrapped within the ceramic support
and the ionic liquid is chemically bound to the sol–gel back-
bone. Our choice for the ionic liquid was, as in the previous
Table 1. The results of some one-pot, selective hydroformylation,
condensation, and hydrogenation processes described in
Scheme 1.[a]
Entry
R1
R2
R3
Overall yield (5) [%][b]
study,
1-butyl-3-[3-(trimethoxysilyl)propyl]imidazolium
chloride (9) together with Na[Ph2P(3-C6H4SO3)].[4] Imid-
azolium derivatives that cannot be bound to the backbone
of the support, such as the commercially available 1,3-di-
tert-butylimidazolium tetrafluoroborate (10) and 1,3-
bis(2,4,6-trimethylphenyl)imidazolium chloride (11), proved
to be difficult to recycle.
1
2
3
4
H
Cl
CH3O
H
CN
CN
CN
CN
CN
CN
CN
83Ϯ3
87Ϯ7
57
CO2C2H5 75Ϯ4
[a] Reaction conditions: 1 mmol 1, 1 mmol 3, sol–gel material pre-
pared from 11 mmol prehydrolyzed Si(OMe)4, 0.37 mmol 9,
0.02 mmol [Rh(cod)Cl]2, 0.04 mmol Na[Ph2P-3-(C6H4SO3)],
1.8 mmol 12 entrapped within a non-modified sol–gel; 10 mL THF
and 5 mL 1,2-dichloroethane; 20.7 bar H2 and 20.7 bar CO; 80 °C;
14 h. [b] In the first run.
While the best solvents for the stereoselective hydrofor-
mylation are hydrocarbons and 1,2-dichloroethane,[4] the
preferred medium for the combined hydroformylation, con-
densation, and hydrogenation process was a 1:1 mixture of
1,2-dichloroethane and THF.
A series of comparative experiments revealed the sensi-
tivity of the process described in Scheme 1 towards the
physical nature of the various reaction components. Under
homogeneous conditions, i.e. when both the rhodium cata-
lyst and base, as well as the ionic liquid, are not entrapped
in the sol–gel, the multistep process does not take place at
all. Entrapment of either the base or the rhodium complex
alone causes partial decomposition of the catalyst, and,
consequently, only a low yield of the final product is
formed. The results of these comparative experiments are
listed in Table 2.
We speculate that under the aforementioned conditions
the imidazole derivative serves both as a carbene ligand of
the rhodium complex (and is responsible for stereoselective
hydroformylation) and acts as a “base” that promotes the
condensation of the aldehyde with malononitrile. Thus, a
part of it is consumed during the first run of the multistep
process, and a new supplement of base is required for the
second run. The best results were obtained when the base
was the sol–gel bound 1,5,7-triazabicyclo[4.4.0]dec-5-ene
modified
with
(3-glycidoxypropyl)trimethoxysilane
(12),[2d,12] although the addition of a new portion of sol–
gel containing 9 to the reaction mixture in advanced runs
proved to be quite satisfactory.
Table 2. Dependence of the multistep, one-pot process of 1 (R1 =
H) and 3 (R2 = R3 = CN) on the immobilization of the catalyst
and the base.[a]
Entry
Rhodium
catalyst[b]
Guanidine
base[b]
Yield (products) [%][c]
2
4
5
8
1
2
3
4
–
+
–
–
–
+
+
9.0
1.8
2.6
2.2
–
0.7
–
8.3
–
1.2
In fact, there is an advantage in using the sol–gel en-
trapped guanidine base already in the first run. This in-
creases the yield of 5 (R1 = H, R2 = R3 = CN) from 86 to
91% at the expense of the unsaturated dinitrile 4 (R1 = H,
R2 = R3 = CN), which under such conditions hardly ap-
pears among the final products. Typical yields of the satu-
rated dinitrile in the second, third, and fourth runs were 86,
trace 17.1 1.3
trace 91 3.4
+
[a] Reaction conditions: the amounts of the reaction components,
the temperature and the time as in Table 1. The yields in this series
of experiments were determined by GC. [b] The signs “+” and
“–” refer to the sol–gel entrapped and non-entrapped materials,
respectively. [c] The yields of 6 and 7 (R1 = H, R2 = R3 = CN)
were in each case Ͻ0.2%. The amounts of recovered 1 (R1 = H)
80, and 82%, respectively, after 14 h at 80 °C. Attempts to have been omitted.
Eur. J. Org. Chem. 2009, 1502–1505
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1503

K. Hamza, H. Schumann, J. Blum
SHORT COMMUNICATION
Gelest Silanes and Silicons. Di-µ-chlorobis[(1,2,5,6-η)-1,5-cyclooc-
Whereas an equal pressure of 13.8–20.7 bar of H2 and
CO does not lead to the formation of detectable quantities
of ethylbenzene derivatives, an increase in the H2 pressure
at the expense of the CO pressure causes substantial re-
duction of the double bond of 1. On the other hand, excess
CO leads to an increase in the amount of the unsaturated
dinitriles 4 and 7. The results of some representative experi-
ments are listed in Table 3.
tadiene]dirhodium,[16]
1-butyl-3-[3-(trimethoxysilyl)propyl]imid-
azolium chloride,[4] sodium 3-(diphenylphosphanyl)benzenesul-
fonate,[17] (2-phenylpropylidene)propanedinitrile,[6] (3-phenylpro-
pylidene)propanedinitrile,[7] (2-phenylpropyl)propanedinitrile,[5] (3-
phenylpropyl)propanedinitrile,[7] and ethyl α-cyano-γ-methylben-
zene butanoate[18] were prepared according to literature procedures.
The immobilization of the catalyst was performed essentially as
described previously.[4] The following analytical instruments were
used: Bruker Vector 22, FTIR spectrometer, Bruker DRX-400
NMR instruments, Hewlett Packard model Agilent 4890 D gas
chromatograph, Hewlet Packard model 4989A mass spectrometer
equipped with an HP gas chromatograph model 5890 series II, and
a Q-TOF-II (Micromass, UK) spectrometer that enabled direct in-
jection by nanoelectrospray through a glass capillary at 1200 V. A
Perkin–Elmer model ELAN DRC II instrument was used for in-
ductively coupled plasma (ICP) measurements. Transmission elec-
tron microscopy was carried out with a Scanning Transmission
Electron Microscope (STEM) Tecani G2 F20 (FEI Company,
USA) operated at 200 kV and equipped with EDAX-EDS for iden-
tification of the elemental composition. Initial powders were dis-
persed in ethanol and dropped onto a standard 400 mesh carbon-
coated copper TEM gird.
Table 3. Dependence of the amount of ethylbenzene and the unsat-
urated dinitriles 4 and 7 on the H2 and CO pressures.[a]
Entry Initial H2 Initial CO Yield
Yield (4 + 7) {R1
pressure
[bar]
pressure
[bar]
(C6H5C2H5) = H, R2 = R3 =
[%]
CN} [%]
1
2
3
4
5
13.8
13.8
20.7
27.6
34.5
27.6
34.5
13.8
13.8
13.8
–
–
7
18
27
17
32
1.2
0.5
–
[a] Reaction conditions: except for the pressure of the gases, the
same conditions as in Table 1 were employed. The yields in this
series of experiments were determined by GC.
Finally, it is notable that although hydroformylation[13,14]
and hydrogenation of alkenes[15] by transition-metal com-
plexes have been shown to often involve metallic nanopar-
ticles as reaction intermediates, thorough TEM measure-
ments were unable to detect the presence of Rh0 nanopar-
ticles in our one-pot processes.
General Procedure for the One-Pot, Multistep Process: Typically, a
100 mL glass-lined mini-autoclave equipped with a mechanical stir-
rer and a sampling device was charged with the styrene derivative
1 (1 mmol: 104 mg for R1 = H; 138.5 mg for R1 = Cl; 134 mg for
R1 = CH3O), nitrile 3 (1 mmol: 66 mg for R2 = R3 = CN; 113 mg
for R2
= = CO2C2H5), the heterogenized catalyst
CN, R3
{containing [Rh(cod)Cl]2 (10 mg, 0.02 mmol), Na[Ph2P-3-
C6H4SO3] (14.6 mg, 0.04 mmol) within the ceramic material from
Si(OMe)4 (5 mL) and ionic liquid 9 (260 mg, 0.8 mmol)}, together
with the guanidine base 12 (675 mg, 1.8 mmol) within a separate
ionic liquid-free sol–gel matrix from Si(OC2H5)4 (4.5 mL), 1,2-
dichloroethane (10 mL), and THF (10 mL).[4] The autoclave was
sealed, perched with N2, and then pressurized with H2 and CO
(usually 20.7 bar of each of the gases). The reaction mixture was
Conclusions
This study reveals that the co-entrapment of [Rh(cod)-
Cl]2 and sulfonated triphenylphosphane within a silica sol–
gel matrix confined with 1-butyl-3-[3-(trimethoxysilyl)pro-
pyl]imidazolium chloride, forms a ceramic material that cat- stirred and heated at 80 °C for 14 h. After cooling to 0 °C, the
excessive gases were released, and the remaining mixture filtered.
The solid was washed with CH2Cl2 (2ϫ15 mL). The filtrate was
concentrated and analyzed by GC, mass spectrometry, and NMR
spectroscopy and, when possible, compared with authentic samples.
The solid was refluxed with CH2Cl2 (20 mL) and sonicated with
the same solvent for 15 min. The combined washings were concen-
trated and subjected to ICP analysis. The solids were dried at room
temperature for 5 h and then mixed with a fresh portion of 12
(1.8 mmol) entrapped within a non-modified sol–gel before use in
a second catalytic run.
Under these conditions 1 (R1 = H) and 3 (R2 = R3 = CN) afforded
up to 158 mg (86%) (2-phenylpropyl)propanedinitrile (5, R1 = H,
R2 = R3 = CN), which proved to be identical with an authentic
sample.[5] Likewise, 1 (R1 = H) and 3 (R2 = CN, R3 = CO2C2H5)
alyzes in one pot: (i) the hydroformylation of styrenes to
give in high selectivity branched aldehydes; (ii) promotes
the condensation of the latter products with active methyl-
ene compounds; (iii) the hydrogenation of the internal
double bonds of the condensation products. The unsatu-
rated condensation products are not further hydroformyl-
ated. We still do not have a sound explanation for the pref-
erential hydrogenation of the sterically hindered internal
double bonds over that of the terminal one of the styrene
derivatives, as well as over the easily reducible aldehyde
groups. As noted above, the ionic liquid which is assumed
to act as a carbene ligand of the rhodium complex can pro-
mote, in the absence of an additional base, the condensation
process. Thus, while the rhodium catalyst is perfectly recy- gave up to 180 mg (78%) ethyl 2-cyano-(2-phenylpropyl)acetate [5,
R1 = H, R2 = CN, R3 = (O2C2H5)].[18] [2-(4-Chlorophenylpropyl)]-
clable it is necessary to renew a part of the base after each
run of the multistep process.
propanedinitrile (5, R1 = H, R2 = R3 = CN) was obtained in 87–
94% yield as a pale yellow viscous oil. 1H NMR (400 MHz,
CDCl3): δ = 1.35 (d, J = 7 Hz, 3 H), 2.18 (m, 1 H), 2.37 (m, 1 H),
3.03 (m, 1 H), 3.32 (dd, J1 = J2 = 5 Hz, 1 H), 7.25 (ABq, JAB
=
Experimental Section
8 Hz, 4 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 20.92, 21.64,
36.9, 38.42, 112.40, 128.70, 129.58, 133.35, 140.59 ppm. MSEI: m/z
(%) = 218/220 (5) [M+·], 139/141 (100) [ClC6H4CHCH3+], 103 (31)
[C8H7+], 77 (29) [C6H5+], 65 (5) [C3HN2+]. C12H11ClN2 (218.676):
calcd. C 65.91, H 5.07, N 12.81; found C 65.93, H 5.23, N 12.42.
General: The various styrenes, propanedinitrile, ethyl acetoacetate,
2- and 3-phenylpropanecarboxaldehyde, 2-phenylpropanol, and 1-
butylimidazole were purchased from Sigma–Aldrich. (3-Chloropro-
pyl)trimethoxysilane and tetramethoxysilane were obtained from
1504
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1502–1505

A Sol–Gel Entrapped Rh Complex in One-Pot Reactions
[2-(4-Methoxyphenylpropyl)]propanedinitrile (5, R1 = OCH3, R2 =
R3 = CN) was obtained in 57% yield as a pale yellow viscous oil.
1H NMR (400 MHz, CDCl3): δ = 1.35 (d, J = 7 Hz, 3 H), 2.14 (m,
1 H), 2.36 (m, 1 H), 2.98 (m, 1 H), 3.30 (dd, J1 = J2 = 5 Hz, 1 H),
3.80 (s, 3 H), 7.01 (ABq, JAB = 9 Hz, 4 H) ppm. 13C NMR
(100 MHz, CDCl3): δ = 20.91, 21.90, 36.72, 39.00, 55.23, 112.49,
114.60, 127.82, 128.98, 159.01 ppm. MSEI: m/z (%) = 214 (8) [M+·],
135 (100) [CH3OC6H4CHCH3+], 121 (4) [C8H9O+], 107 (9)
[C7H7O+], 91 (14) [C7H7+], 77 (8) [C6H5+], 65 (6) [C3HN2+].
C13H14N2O (214.263): calcd. C 72.87, H 6.58, N 13.07; found C
72.67, H 6.54, N 12.85.
Price, M. Kobaslija, D. T. McQuade, Org. Biomol. Chem. 2005,
3, 2899–2906 and references cited therein.
[4] K. Hamza, J. Blum, Eur. J. Org. Chem. 2007, 4706–4710.
[5] M. Ohashi, K. Nakatani, H. Maeda, K. Mizuno, Org. Lett.
2008, 10, 2741–2743.
[6] D. Kruger, A. E. Sopchik, C. A. Kingsbury, J. Org. Chem.
1984, 49, 778–788.
[7] U. Holzgrabe, J. Reinhardt, G. Zlotos, Arch. Pharm. 1994, 327,
515–523.
[8] K. Yamashita, T. Tanaka, M. Hayashi, Tetrahedron 2005, 61,
7981–7985.
[9] See e.g. Carbonylation: Direct Synthesis of Carbonyl Com-
pounds (Eds.: H. M. Colquhoun, D. J. Thompson, N. V.
Twigg), Plenum Press, New York, 1991.
Acknowledgments
[10]
V. C ésar, S. Bellemin-Loponnazi, L. H. Gade, Chem. Soc. Rev.
2004, 33, 619–636.
[11]
a) A. C. Chen, L. Ren, a. Decken, C. M. Crudden, Organome-
tallics 2000, 19, 3459–3461; b) M. Haumann, A. Riisager,
Chem. Rev. 2008, 108, 1474–1497.
We thank the Israel Science Foundation (grant no. 269/06) and the
Exchange Program between the Hebrew University of Jerusalem
and the Technische Universität, Berlin for financial support of this
study.
[12] a) Y. V. Subba Rao, D. E. De Vos, P. A. Jacobs, Angew. Chem.
Int. Ed. Engl. 1997, 36, 2661–2663; b) N. Hüsing, U. Schubert,
R. Mezei, P. Fratzl, B. Riegal, W. Kiefer, D. Kohler, W. Moder,
Chem. Mater. 1999, 11, 451–457.
[13] F. Wen, H. Bönnemann, J. Jiang, D. Lu, Y. Wang, Z. Jin, Appl.
Organomet. Chem. 2005, 19, 81–89.
[1] a) J. Blum, D. Avnir, H. Schumann, Chemtech 1999, 29, 32–
38; b) J. Blum, D. Avnir in Handbook of Sol–Gel Science and
Technology (Ed.: S. Sakka), Kluwer Academic Pub., Boston,
2005, vol. 3, ch. 24.
[14] A. J. Bruss, M. A. Gelesky, G. Machado, J. Dupont, J. Mol.
Catal. A 2006, 252, 212–218.
[2] a) F. Gelman, J. Blum, D. Avnir, J. Am. Chem. Soc. 2000, 122,
11999–12000; b) F. Gelman, J. Blum, D. Avnir, Angew. Chem. [15] See e.g. A. Roucous, A. Nowicki, K. Philippot in Nanoparticles
Int. Ed. 2001, 40, 3647–3649; c) F. Gelman, J. Blum, D. Avnir,
J. Am. Chem. Soc. 2002, 124, 14460–14463; d) F. Gelman, J.
Blum, D. Avnir, J. Sol–Gel Sci. Technol. 2003, 26, 43–46; e) F.
Gelman, J. Blum, D. Avnir, New J. Chem. 2003, 27, 205–207;
f) Y. Levin, K. Hamza, R. Abu-Reziq, J. Blum, Eur. J. Org.
Chem. 2006, 1396–1399; g) K. Hamza, J. Blum, Tetrahedron
Lett. 2007, 48, 293–295.
and Catalysis (Ed.: D. Astruc), Wiley-VCH, Weinheim, 2008,
pp. 349–388.
[16] G. Giordano, R. H. Crabtree, Inorg. Synth. 1990, 28, 88–90.
[17] S. Arland, J. Chatt, N. R. Davis, A. A. Williams, J. Chem. Soc.
1958, 276–288.
[18] T. Mizagaki, Y. Miyauchi, M. Murata, K. Ebitani, K. Kaneda,
Chem. Lett. 2005, 34, 286–287.
[3] For further examples of one-pot, multistep reactions conducted
in other laboratories, see: S. J. Broadwater, S. L. Roth, K. E.
Received: December 18, 2008
Published Online: March 4, 2009
Eur. J. Org. Chem. 2009, 1502–1505
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1505