Synthesis, reaction, and recycle of light fluorous Grubbs-Hoveyda catalysts for alkene metathesis

Article abstract of DOI:10.1021/jo048001n

(Chemical Equation Presented) Light fluorous versions of first- and second-generation Grubbs-Hoveyda metathesis catalysts are introduced. These exhibit the expected reactivity profile, are readily recovered from reaction mixtures by fluorous solid-phase e

Full text of DOI:10.1021/jo048001n

Synthesis, Reaction, and Recycle of Light Fluorous
Grubbs-Hoveyda Catalysts for Alkene Metathesis†
Masato Matsugi and Dennis P. Curran*
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260
curran@pitt.edu
Received November 10, 2004
Light fluorous versions of first- and second-generation Grubbs-Hoveyda metathesis catalysts are
introduced. These exhibit the expected reactivity profile, are readily recovered from reaction mixtures
by fluorous solid-phase extraction, and can be routinely reused five or more times. The catalysts
can be used in a stand alone fashion, or supported on fluorous silica gel.
Introduction
Alkene metathesis is among the most powerful and
popular methods for carbon-carbon bond formation
today, and an assortment of metathesis catalysts are now
available (Figure 1).5 Among these, first and second
generation Grubbs-Hoveyda catalysts 16 and 27 and
related molecules8 are especially useful because of their
scope and stability.9 The high cost of these catalysts
encourages recovery and reuse, even on a small scale.
Organic molecules bearing small fluorous tags (C6F13,
C8F17) are called light fluorous molecules.1 Light fluorous
reagents,2 scavengers, and catalysts are especially con-
venient since they typically induce reactions of organic
substrates under the same conditions as their nonfluo-
rous relatives, but are reliably removed from crude
reaction products by fluorous solid-phase extraction
(fspe).3 Only a handful of light fluorous catalysts, includ-
ing palladium,4a,b platinum,4c and nickel4d complexes,
have been described to date.
An assortment of polymer- and ionic liquid-supported
metathesis catalysts have been introduced to facilitate
separation in metathesis reactions.10 Recently, Yao has
(4) (a) Curran, D. P.; Fischer, K.; Moura-Letts, G. Synlett 2004,
1379-1382. (b) Cavazzini, M.; Pozzi, G.; Quici, S.; Maillard, D.; Sinou,
D. Chem. Commun. 2001, 1220-1221. (c) Zhang, Q.; Luo, Z.; Curran,
D. P. J. Org. Chem. 2000, 65, 8866-8873. (d) Croxtall, B.; Hope, E.
G.; Stuart, A. M. Chem. Commun. 2003, 2430-2431.
† Dedicated to Dr. Alfred Bader in celebration of his 80th birthday.
(1) Curran, D. P.; Luo, Z. J. Am. Chem. Soc. 1999, 121, 9069-9072.
(2) Representative fluorous reagents and scavengers: (a) Dobbs, A.
P.; McGregor-Johnson, C. Tetrahedron Lett. 2002, 43, 2807-2810. (b)
Dandapani, S.; Curran, D. P. Tetrahedron 2002, 58, 3855-3864. (c)
Zhang, W.; Chen, C. H.-T.; Lu, Y.; Nagashima, T. Org. Lett. 2004, 6,
1473-1476. (d) Lindsley, C. W.; Zhao, Z.; Leister, W. H. Tetrahedron
Lett. 2002, 43, 4225-4228. (e) Lindsley, C. W.; Zhao, Z. J.; Leister, W.
H.; Strauss, K. A. Tetrahedron Lett. 2002, 43, 6319-6323. (f) Lindsley,
C. W.; Zhao, Z.; Newton, R. C.; Leister, W. H.; Strauss, K. A.
Tetrahedron Lett. 2002, 43, 4467-4470. (g) Zhang, W.; Curran, D. P.;
Chen, C. H.-T. Tetrahedron 2002, 58, 3871-3875. (h) Zhang, W.; Chen,
C. H.-T.; Nagashima, T. Tetrahedron Lett. 2003, 44, 2065-2068.
(3) Fluorous solid-phase extractions use silica gel with a fluorocarbon
bonded phase: (a) Curran, D. P. Synlett 2001, 1488-1496. (b) Zhang,
W. Tetrahedron 2003, 59, 4475-4489. (c) Curran, D. P. Separations
with Fluorous Silica Gel and Related Materials. In Handbook of
Fluorous Chemistry; Gladysz, J., Horva´th, I., Curran, D. P., Eds.;
Wiley-VCH: Weinheim, Germany, 2004; pp 101-127. (d) Curran, D.
P. A User’s Guide to Light Fluorous Chemistry. In Handbook of
Fluorous Chemistry; Gladysz, J., Horva´th, I., Curran, D. P., Eds.;
Wiley-VCH: Weinheim, Germany, 2004; pp 128-155.
(5) (a) Grubbs, R. H. Handbook of metathesis; Wiley-VCH: Wein-
heim, Germany, 2003. (b) Furstner, A. Angew. Chem., Int. Ed. 2000,
39, 3012-3043. (c) Furstner, A. Alkene metathesis in organic synthesis;
Springer: New York, 1998.
(6) Kingsbury, J. S.; Harrity, J. P. A.; Bonitatebus, P. J.; Hoveyda,
A. H. J. Am. Chem. Soc. 1999, 121, 791-799.
(7) Garber, S. B.; Kingsbury, J. S.; Gray, B. L.; Hoveyda, A. H. J.
Am. Chem. Soc. 2000, 122, 8168-8179.
(8) Leading reference: Zaja, M.; Connon, S. J.; Dunne, A. M.; Rivard,
M.; Buschmann, N.; Jiricek, J.; Blechert, S. Tetrahedron 2003, 59,
6545-6558.
(9) (a) Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J.
Am. Chem. Soc. 1992, 114, 3974-3975. (b) Nguyen, S. T.; Grubbs, R.
H.; Ziller, J. W. J. Am. Chem. Soc. 1993, 115, 9858-9859. (c) Wu, Z.;
Nguyen, S. T.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1995,
117, 5503-5511. (d) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs,
R. H. Angew. Chem., Int. Ed. Engl. 1995, 34, 2039-2041. (e) Schwab,
P.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996, 118, 100-110.
10.1021/jo048001n CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/09/2005
1636
J. Org. Chem. 2005, 70, 1636-1642

Light Fluorous Grubbs-Hoveyda Catalysts
SCHEME 1a
FIGURE 1. Grubbs-Hoveyda (GH) metathesis catalysts.
reported a heavy fluorous Grubbs-Hoveyda catalyst 3.11
The catalyst is linked to a fluorous polyacrylate and can
be used in conjunction with a fluorous solvent for fluorous
biphasic catalysis. We report herein complementary light
fluorous Grubbs-Hoveyda (f-GH) catalysts 4a,b and 5.
These induce alkene metathesis reactions under the same
conditions as their nonfluorous parents but can be readily
separated from crude reaction products by fluorous solid-
phase extraction (fspe) when added to reaction mixtures
in pure form or by filtration when added in supported
form on fluorous silica gel. The catalysts are reasonably
stable to both the reaction and separation conditions, and
can be recovered and reused repeatedly.
a Reagents and conditions: (a) 2.2 equiv of C8F17I, 1 equiv of
Me3Al, 5 mol % of Pd(PPh3)4, CH2Cl2, rt, 64% (recrys.); (b) 1.5 equiv
of Bu3SnH, 0.2 equiv of AlBN, benzene, reflux, 81%; (c) 4 equiv of
BBr3S(CH3)2, dichloroethane, reflux, 88%; (d) 1.5 equiv of NaH, 2
i
equiv of PrI, THF/DMF, rt, 94%; (e) 1.1 equiv of Br2, 0.04 equiv
of AcOH, CH2Cl2, rt, 94%; (f) 50 mol % of Pd(PPh3)4, 3 equiv of
tributylvinylstannane, toulene, reflux, 64%; (g) 1.1 equiv of
Grubbs-I, 1.25 equiv of CuCl, CH2Cl2, rt, 71%.
Second-generation f-GH catalyst 5 differs from 4a by
having the N-heterocyclic carbene (NHC) ligand17 in place
of the phosphine, and it has an ethylene spacer rather
than a propylene spacer. The ethylene spacer was used
because ligand precursor 9b became commercially avail-
able during the course of this work.18 Trans-metathesis
of 9b with standard second-generation Grubbs catalyst
(Grubbs-II) gave f-GH catalyst 5 as green crystals, mp
136.5-137.5 °C, after column chromatography and re-
crystallization from CH2Cl2/pentane.
Most of the experiments with first generation f-GH
catalyst 4a were complete when styrene 9b became
available to us. However, to show that propylene- and
ethylene-spacer catalysts were analogous, we also syn-
thesized 4b by cross metathesis from 9b and Grubbs-I
catalyst as above. Complex 4b was isolated in 92% yield
as a brown solid, mp 154.5-155.5 °C.
Complexes 4a,b and 5 were air stable over several days
either as solids or dissolved in solvent in an NMR tube,
so no special handling or storage precautions were taken.
However, 4a and 5 slowly decomposed over several days
when heated in CDCl3 at 80 °C (estimated half-lives: 4a,
∼20 h; 5, ∼8 days). The better thermal stability of 5 is
consistent with its better recovery in the reuse experi-
ments described below. The catalysts were not signifi-
cantly soluble in fluorous solvents such as FC-72 (per-
Results and Discussion
Syntheses of the first- and second-generation f-GH
catalysts are summarized in Scheme 1. For first-genera-
tion catalyst 4a, addition of the fluorous tag to 1-allyl-
4-methoxybenzene 6 was accomplished by standard atom
transfer addition12 and reduction13 reactions to provide
7. Demethylation14 and isoproylation provided 8. This
was brominated15 and the bromide was coupled with
vinyl tributyltin16 to provide ligand precursor 9a. Trans-
metathesis7 with the standard first-generation Grubbs
catalyst (Grubbs-I) provided the f-GH catalyst 4a as
brown crystals (mp 159.0-160.0 °C) after recrystalliza-
tion from CH2Cl2/pentane. The crystal structure of this
complex was solved, and ORTEP diagrams and full data
are provided in the Supporting Information.
(10) (a) Nguyen, S. T.; Trnka, T. M. In Handbook of Metathesis;
Grubbs, R. H., Ed.; Wiley-VCH: Weinheim, Germany, 2003; Vol. 1,
pp 61-94. (b) Harned, A. M.; Probst, D. A.; Hanson, P. R. In Handbook
of Metathesis; Grubbs, R. H., Ed.; Wiley-VCH: Weinheim, Germany,
2003; Vol. 2, pp 361-402.
(11) Yao, Q.; Zhang, Y. J. Am. Chem. Soc. 2004, 126, 74-75.
(12) Maruoka, K.; Sano, H.; Fukutani, Y.; Yamamoto, H. Chem. Lett.
1985, 1689-1692.
(13) Aimetti, J. A.; Hamanaka, E. S.; Johnson, D. A.; Kellogg, M. S.
Tetrahedron Lett. 1979, 20, 4631-4634.
(17) Connon, S. J.; Blechert, S. Angew. Chem., Int. Ed. 2003, 42,
1900-1923.
(18) (a) Ligand 9b and FluoroFlash silica gel and TLC plates were
purchased from Fluorous Technologies, Inc., www.fluorous.com. Fluo-
roFlash silica gel has a perflourooctylethyl silyl (C8F17CH2CH2Si)
bonded phase. (b) DPC holds an equity interest in this company.
(14) Williard, P. G.; Fryhle, C. B. Tetrahedron Lett. 1980, 21, 3731-
3734.
(15) Audic, N.; Clavier, H.; Mauduit, M.; Guillemin, J.-C. J. Am.
Chem. Soc. 2003, 125, 9248-9249.
(16) Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508-524.
J. Org. Chem, Vol. 70, No. 5, 2005 1637

Matsugi and Curran
TABLE 1. Reuse of f-GH Catalysts 4a,b and 5 in the
Metathesis of 10 to 11
entry cat. cyc 10 mg cat mg 11,% (mg) cat., % (mg)
1
2
4a
4a
4a
4a
4a
4a
4a
4b
5
1
2
3
4
5
6
7
1
1
2
3
4
5
196
170
90
41
36
19
16
11
6
99 (180)
99 (155)
98 (78)
99 (67)
99 (46)
99 (25)
98 (18)
99 (107)
96 (256)
98 (217)
98 (187)
94 (134)
97 (121)
88 (36)
72 (26)
84 (16)
69 (11)
73 (8.0)
73 (4.4)
80 (3.2)
84 (21)
91 (58)
91 (48)
85 (39)
89 (31)
90 (27)
FIGURE 2. TLC behavior of 1, 4a, and 5 on FluoroFlash
3
silica.
4
76
5
52
fluorohexanes), but instead dissolved in common organic
solvents including diethyl ether, THF, CH2Cl2, and
CH3CN.
6
28
7
21
4
8
121
300
249
215
161
140
25
64
53
46
35
30
9
To evaluate prospects for fspe separation, we analyzed
the TLC behavior of the standard and fluorous GH
catalysts under typical “fluorophobic” conditions (to model
the first stage of a fluorous spe) and “fluorophilic”
conditions (to model the second stage) on commercial
FluoroFlash fluorous silica gel.18 The results of these
informative experiments are summarized pictorially in
Figure 2.
When the FluoroFlash TLC plate was eluted with
100% acetonitrile, nonfluorous catalyst 1 moved with the
solvent front, as expected, while f-GH catalyst 4a was
well retained. In contrast, f-GH catalyst 5 exhibited a
long streak, suggesting decomposition. In 100% metha-
nol, both f-GH catalysts 4a and 5 provided well-resolved
spots, but the retention factors (Rf) were too high for
filtration-based separation. In 10/1 MeOH/H2O, the Rf
factors of both complexes decreased significantly and they
only barely moved from the baseline. In the fluorophilic
solvent THF, both complexes exhibited clean spots near
the solvent front. A similar behavior was observed in
ether (not shown).
10
11
12
13
5
5
5
5
mg, >99%) followed by 18 mL of ether to provide a
fluorous fraction containing recovered catalyst 4 (36 mg,
88%).
The recovered catalyst exhibited a 1H NMR spectrum
similar to the starting catalyst, and it was used directly
in the second cycle with appropriate adjustment of all
other components to maintain the same ratio (5 mol %
of catalyst). This process was continued through seven
cycles with the weights and yields of product and catalyst
shown in Table 1, entries 1-7. The yield of the product
was uniformly high, while the yield of recovered catalyst
ranged from 69% to 88%. At the end of the seventh cycle,
about 10% of the original catalyst was recovered. While
the recovered complex is no longer pristine (see the
1
Supporting Information for the H NMR spectrum), we
believe that it is still active. Overall, the initial 41 mg of
catalyst (0.04 mmol) was used to metathesize 633 mg
(about 2.5 mmol) of 10. No effort for further optimization
of the loading or recovery was attempted.
The first-generation catalyst with the ethylene spacer
4b was tested in a single experiment (entry 8) and gave
results very comparable to those of 4a. Based on this and
the structural similarity, we suggest that 4a and 4b can
be used interchangeably.
The second-generation catalyst 5 behaved similarly,
and a series of five metathesis cycles with this catalyst
are summarized in Table 1, entries 9-13.20 Optimized
fspe conditions required somewhat more fluorous silica
gel (50× weight of the catalyst) compared to 4a. The
fluorophobic elution was conducted with 80% MeOH/H2O
and THF was used for the fluorophilic elution. Yields of
product 11 were again high (94-98%) in each cycle, while
These results suggest that complex 4a can be retained
in a fluorophobic pass with water-free acetonitrile while
aqueous methanol is preferred for 5. Both complexes
should be eluted in a fluorophilic pass with ether or THF.
These experiments show how simple fluorous TLC ex-
periments can be used to quickly identify suitable condi-
tions for solid-phase extractions.
The reaction, separation, and reuse features of f-GH
catalysts 4a and 5 were studied in detail in the conver-
sions of N,N-diallyl-p-toluenesulfonamide 10 to N-p-tosyl-
2,5-dihydro-1H-pyrrole 11, and the results of this series
of experiments are summarized in Table 1. In a typical
experiment with the first-generation catalyst 4a, 196 mg
(0.8 mmol) of 10 was heated with 41 mg (5 mol %) of 4a
in dichloromethane at reflux for 2 h.19 The cooled mixture
was loaded to 1.2 g of fluorous silica gel (30× weight of
the catalyst), which was eluted with 6 mL of acetonitrile
to provide an organic fraction containing product 11 (180
(20) Tosyl amide 10 (300 mg, 1.20 mmol) and 5 (64.2 mg, 0.060
mmol, 0.050 equiv) and dichloromethane (24 mL, 0.05 M) were placed
in a flask under an argon atmosphere and the mixture was refluxed
for 2 h at 55 °C. After removal of the volatile components by
evaporation, the brownish mixture was submitted to separation by fspe.
A short column was packed with fluorous silica gel (3.2 g) using aq
80% MeOH as the solvent. The crude reaction mixture was then loaded
onto this column and eluted with 16.1 mL of aq 80% MeOH and
successively 9.7 mL of THF. The evaporation of the 80% MeOH fraction
and the THF fraction by vacuum centrifuge gave RCM product 11 in
96% yield (256 mg) and catalyst 5 in 91% yield (58.3 mg), respectively.
(19) Tosyl amide 10 (1.00 g, 3.98 mmol) and 4a (211 mg, 0.199 mmol,
0.050 equiv) and dichloromethane (79.6 mL, 0.05 M) were placed in a
flask under an argon atmosphere and the mixture was refluxed for 2
h at 55 °C. After removal of the volatile components by evaporation,
the brownish mixture was submitted to separation by fspe. A short
column was packed with fluorous silica gel (6.3 g) using acetonitrile
as the solvent. The crude reaction mixture was then loaded onto this
column and eluted with 31.5 mL of acetonitrile to give RCM product
11 in 98% yield (865 mg). The next elution of diethyl ether (94.5 mL)
afforded the catalyst 4a in 87% yield (184 mg).
1638 J. Org. Chem., Vol. 70, No. 5, 2005

Light Fluorous Grubbs-Hoveyda Catalysts
the recovery of catalyst 5 was better, ranging from 85%
to 91%. At the end of five cycles, 915 mg (4.1 mmol) of
product 11 was produced starting from 64 mg (0.06 mmol)
of 5, and 42% of the original catalyst mass was recovered.
The catalyst recovered after cycle five is not as pure as
lyst, but deactivation of the catalyst was observed in the
third cycle.14 We conducted three cycles of metathesis of
this substrate without problem, and recovered 70% of the
original catalyst mass.
Catalyst 5 was more reactive as expected, and all
substrates were consumed in 2 h. However, only the
products in entries 1-3 were reasonably pure as assayed
by 1H NMR spectroscopy. The target products from
entries 4 and 5 were indeed the major components, but
these samples also had significant impurities that may
result from ring opening metathesis or polymerization.
These products were not further purified. Recovery of the
catalyst 5 was satisfactory in entries 1-4 (76-89%), but
lower (62%) in entry 5.
1
the starting complex (see H NMR spectra in the Sup-
porting Information), but we expect that it is still active.
And if desired, it could be repurified by chromatography
or crystallization prior to reuse.
1
The H NMR spectra of the products of Table 1 are
clean and ligand resonances cannot be detected in either
these or the 19F NMR spectra (see the Supporting
Information). However, the crude products are typically
light tan. This, coupled with the observation that the
recovered yield of the catalyst is never quantitative,
suggests that the product may still contain small amounts
of ruthenium. To obtain more information on residual
ruthenium, we metathesized 1 g of substrate 10 with 211
mg (5 mol %) of f-GH catalyst 4a under the standard
conditions. After fspe, we recovered 965 mg (98%) of 11
along with 184 mg (87%) of the catalyst 4a.
Next, we tested the reactivity of 4a and 5 in a
representative cross-metathesis reaction between 4-phenyl-
1-butene and benzyl acrylate under the standard condi-
tions with the usual fspe separation (eq 1). First-
Crude product 11 was light tan in color, though no
resonances for the ligand could be detected in either its
1H or 19F NMR spectrum. Spiking the product with 0.3
mol % of benzotrifluoride (an amount equivalent to 6%
of the original amount of catalyst) showed that the CF3
group of the spike was readily detected, so the fluorine
content of the sample must be well below this. Elemental
analysis of the crude product showed that it contained
0.15% ruthenium, which corresponds to about 6% of the
original ruthenium added.21 Because the corresponding
6% of the ligand is absent, we presume that most of this
is ruthenium that has been released from the fluorous
component of the ligand and is therefore not retained on
the spe. Portions of the crude product were further
purified by recrystallization (to give a very faint tan solid)
and flash chromatography (to give a white solid). These
products exhibited <0.05% ruthenium (the limit of
delectability) in elemental analysis. These levels were
deemed satisfactory and further trace analysis was not
conducted. A similar preparative experiment with 5 gave
qualitatively similar results.22
The scope of the new catalysts was briefly probed by
conducting ring closing metathesis with five substrates
under the standard conditions, and the results of these
experiments are summarized in Table 2. All five sub-
strates cyclized smoothly and in good yield with f-GH
catalyst 4a, though the cyclization of acrylate in entry 3
required 3 days. Yields of products were uniformly good
(84-99%), and the catalyst was recovered in 76-92%
yields. Metathesis of substrate in entry 5 was described
with an ionic liquid-supported first-generation GH cata-
generation catalyst 4a provided the homodimer 12 in 82%
yield alongside recovered benzyl acrylate (100%) and
catalyst 4a (91%). Second-generation catalyst 5 provided
cross-coupled product 13 in 88% yield and homodimer
1
12 was not detected by H NMR spectroscopy. Catalyst
5 was recovered in 63% yield. These results are in line
with expectations from the nonfluorous catalysts6,7,23 and
provide additional evidence that the fluorous catalysts
will exhibit reactivity profiles that can be readily antici-
pated from results in the standard series.
Finally, we conducted a series of experiments with a
silica-supported catalyst to show that f-GH catalysts are
compatible with this mode of delivery and removal.
Recently, Gladysz, Bannwarth, and others have devel-
oped procedures to isolate supported fluorous compounds
from reactions,24 and Teflon and fluorous silica have been
used as supports to date. In some procedures, the
supported catalyst or reactant is added directly to the
reaction mixture, while in others a soluble catalyst or
reactant is used and the support is added after the
reaction. This later approach is one of the standard ones
for spe loading.25
(21) A control experiment with standard catalyst 1 (14 mg) and 10
(119 mg) provided 109 mg of crude product 11 (>100%). This was a
dark brown solid and was contaminated with 0.56% ruthenium,
corresponding to 25% of the original amount of ruthenium added.
Surprisingly, 10 mg (71%) of the original catalyst was recovered from
the fluorous spe. The TLC results (Figure 2) suggest that the catalyst
should not be retained, so this may be due to precipitation during spe
loading or elution.
(22) The crude product 11 was light tan and contained 0.46% Ru
by elemental analysis (19% of original catalyst charge; however, there
may be an error in this analysis since 91% of the original catalyst mass
was recovered). The chromatographed product was a white powder
containing <0.09% Ru while the crystallized product was a very pale
tan powder containing <0.05% Ru.
(23) (a) Chaterjee, A. K.; Choi, T.-L.; Sanders, D. P.; Grubbs, R. H.
J. Am. Chem. Soc. 2003, 125, 11, 360-11, 370. (b) Fu¨rstner, A.;
Ackermann L.; Gabor, B.; Goddard, R.; Lehmann, C. W.; Mynott, R.;
Stelzer, F.; Thiel, O. R. Chem. Eur. J. 2001, 7, 3236-3253.
(24) (a) Schwinn, D.; Glatz, H.; Bannwarth, W. Helv. Chim. Acta
2003, 86, 188-195. (b) Tzschucke, C. C.; Markert, C.; Glatz, H.;
Bannwarth, W. Angew. Chem., Int. Ed. 2002, 41, 4500-4503. (c)
Wende, M.; Meier, R.; Gladysz, J. A. J. Am. Chem. Soc. 2001, 123,
11490-11491. (d) Wende, M.; Gladysz, J. A. J. Am. Chem. Soc. 2003,
125, 5861-5872. (e) Jenkins, P. M.; Steele, A. M.; Tsang, S. C. Catal.
Commun. 2003, 4, 45-50.
(25) See Fluoruos Technolgies, Inc. Product Application Note “Fluo-
rous Solid-Phase Extraction”: http://fluorous.com/download.html.
J. Org. Chem, Vol. 70, No. 5, 2005 1639

Matsugi and Curran
TABLE 2. Examples of RCM with f-GH Catalysts 4a and
5
TABLE 3. Examples of RCM with Silica-Supported f-GH
Catalyst 5
a Yes: trace signals detected in the 19F NMR, estimated less
than 0.4% compared to the product. No: no clear signals in the
19F NMR spectrum. b At 85% conversion of 10. c Contains 0.4%
signal relative to the product
a Contains impurities. b Three cycles with the same original
catalyst charge. c After flash chromatographic purification.
We initially dissolved f-GH complex 5 in CH2Cl2, added
30× weight of fluorous silica gel, and then evaporated to
dryness. The resulting green, free-flowing supported
complex was then suspended in 80% MeOH/water. The
powder remained green while the solution stayed clear;
little or no reaction occurred when substrates were added
followed by heating. As expected, the complex prefers the
fluorous silica to the polar, hydrophilic solvent. However,
the complex is readily extracted back from the fluorous
silica by washing with the standard reaction solvent
dichloromethane. Thus, simply by switching solvents, the
complex can be driven onto or extracted off of the fluorous
silica gel.
This material may be of sufficient purity for many uses;
however, to upgrade it even further, we poured the
MeOH/water solution through a small pad (60 mg) of
additional fluorous silica gel. Although the solution
appeared clear, a faint green line was visible at the top
of the silica pad after filtration. Product 11 (24 mg, 97%)
was obtained after evaporation of the solvent, and it was
fluorine free according to the standard 19F NMR analysis.
The green silica gel from the initial filtration was washed
with THF to provide the recovered catalyst 5 in 83% yield
after evaporation.
Table 3 summarizes the results of several experiments
conducted to probe the features of supported catalyst 5.
In these experiments, the slurry resulting from addition
of MeOH/water was simply filtered and evaporated; it
was not passed through an additional pad of fluorous
silica. Diethyl diallyl malonate was subjected to three
cycles of metathesis using the same supported catalyst
charge (entries 1-3). Unlike the experiments in Table
2, we did not adjust the amount of precursor to compen-
sate for lost catalyst (because the supported catalyst is
a material and we do not know how much catalyst was
lost); we simply used the same amount of precursor in
each of the three cycles. Cycles 1 and 2 proceeded
smoothly in the allotted time (1 h), while only 85%
conversion was observed in cycle 3. Presumably this is a
result of the gradual erosion of the catalyst, as in Table
2. A fourth cycle (entry 4) was conducted for 3 h, and
this gave complete conversion to 11, which was isolated
in 94% yield.
Equation 2 summarizes a very simple procedure that
we developed to capitalize on this two-state, solvent-
dependent behavior. Silica-supported 5 (180 mg of silica
containing 6 mg of 5, 5 mol %) was added to 10 (28 mg)
in CH2Cl2 (2.2 mL). The resulting slurry was refluxed
for 5 h, then the mixture was cooled and the solvent was
evaporated to yield a dry green powder. Methanol/water
(80%, 3 mL) was added and the slurry was filtered with
a standard suction filter. The liquid phase containing the
product was clear to the naked eye, yet it still contains a
trace (<1%, see below) of the originally added catalyst.
1640 J. Org. Chem., Vol. 70, No. 5, 2005

Light Fluorous Grubbs-Hoveyda Catalysts
All of the crude products contained tiny, but clearly
visible resonances in the 19F NMR spectra, which we
estimate to account for less than 8% of the original
amount of catalyst 5 that was added (in other words, less
than 0.4 mol % compared to the product). This estimate
is based on an accurate quantitation of the product of
entry 4 by adding a BTF standard; this contained 0.4%
of the original product. While the 19F NMR assay quanti-
ties total fluorine content and does not provide structural
information, we speculate that most of the trace fluorine-
containing material is the catalyst because of the green
band observed in filtration (see above).
added initially on fluorous silica gel) or by fluorous solid-
phase extraction (when used in a standard solution-phase
reaction). Only a small amount of fluorous silica gel is
needed (30-50× weight of the catalyst), and catalysts
4a,b are especially convenient because they can be
retained under water-free conditions with acetonitrile.
While most of the work on first-generation catalysts has
been conducted with propylene-spaced complex 4a, the
commercial availability of the styrene precursor 9b of
ethylene-spaced complex 4b and the similarities between
4a and 4b suggest that 4b is presently the catalyst of
choice in this series.
Fresh fluorous silica-supported catalyst charges were
used to metathesize the substrates in entries 5-7, and
yields were uniformly excellent. Purities were also sat-
isfactory, with very small ligand peaks (again possibly
from the catalyst) being detected in the products from
entries 5 and 7. The 19F NMR from entry 6 was free from
resonances. The colors of all the products were very pale
tan and we judged these products to be more faintly
colored than typical products from crude fspes (Table 2).
The very small amounts of fluorine-containing residues
in these samples are acceptable for many applications;
indeed, it is only the high sensitivity of the assay that
allows us to detect these otherwise spectroscopically
silent contaminants. If better quality product is needed,
it can be filtered through a small pad of fluorous silica
gel to remove the last traces of fluorous residues.
We conducted a series of control experiments with
diallyl malonate to better understand the supported
procedure, and the results of these experiments are
summarized in the Supporting Information. Briefly,
attempts to use the standard Grubbs-Hoveyda catalyst
2 under the procedure outlined above with fluorous silica
gel, reverse phase silica gel, or no silica gel at all were
uniformly unsuccessful. The standard catalyst 2 simply
dissolved when the 80/20 MeOH/water was added, and
it ended up contaminating the product. Little or no
catalyst was retained on the support or left in the flask
(when support was absent).
In contrast, while the use of f-GH catalyst 5 supported
on standard silica gel gave inferior results, its use
supported on reverse phase silica gel or unsupported gave
comparable results to the use of fluorous-supported
catalyst. In the unsupported procedure, the catalyst was
simply precipitated by addition of 80/20 MeOH/water, but
its recovery was difficult since it clung to the flask and
the stir bar. These results suggest that the prime factor
behind the success of the separation in the solvent-switch
procedure is the insolubility of 5 in 80/20 MeOH/water.
This is not surprising, since the same insolubility drives
the related fspe. Because of the mutually selective nature
of fluorous interactions, we recommend fluorous silica gel
as the material of first choice for this procedure, though
other materials or even no material at all can potentially
be used.
Standard Grubbs-Hoveyda catalysts can often be
removed from crude reaction products by regular silica
gel chromatography,6,7 but this method is probably not
as reliable or general as the fluorous spe; some organic
molecules will have similar Rf values to standard cata-
lysts on regular silica gel while essentially no organic
molecules will be retained on fluorous spe. Furthermore,
we find that both fluorous and nonfluorous Grubbs-
Hoveyda catalysts tend to streak to some extent on
standard silica gel. This suggests that purification of
products that elute after the complexes can be a problem
in standard silica gel chromatography. In contrast, on
fluorous silica gel, no product ever elutes after the
complex. This “one size fits all” aspect of the fluorous spe
is attractive.
The recoveries of the catalyst are good (typically 75-
95%) but not quantitative, at least in part because the
catalysts are not fully stable to the reaction conditions.
Small amounts of ruthenium are divorced from the
fluorous carbene ligand during the reaction and leach into
the product. The levels of contamination are tolerable for
many types of multistep syntheses and can be reduced
by chromatography, crystallization, or other means. On
the positive side, because the fluorous spe targets a
ligand and not the metal, the material recovered from
the spe is largely active catalyst. In addition to reusing
the catalyst, we routinely reused the fluorous silica gel
after washing with THF and reconditioning with the
fluorophobic solvent.
Assorted resin-bound,9 heavy fluorous10 and ionic
liquid-supported14 Grubbs catalysts have been reported,
and these catalysts are effectively materials. Analyses
of these materials are typically based on performance.
This is convenient when the material is reused to
repeatedly conduct the same reaction, but has detractions
when the material is used to conduct different reactions.
For example, degraded catalyst products or left over
products or reactants from a prior reaction may remain
in the active catalyst material and contaminate subse-
quent reactions. In contrast, the light fluorous catalysts
are molecules, and can be analyzed and characterized as
such. The purity of the recovered catalyst is readily
assessed by NMR spectroscopy, melting point, and other
means. If this is deemed unacceptable, then the catalyst
can be recrystallized or otherwise purified. Standard
Grubbs-Hoveyda catalysts are expensive even in small
quantities, so the ability to recover and reuse the catalyst
can result in significant cost savings.
Conclusions
New fluorous Grubbs-Hoveyda catalysts 4a,b and 5
show typical features of light fluorous catalysts; they
react under similar conditions and show similar reactiv-
ity profiles to the nonfluorous analogues, but they are
readily separated and recovered either by filtration (when
Acknowledgment. We thank the National Insti-
tutes of Health for funding this work. We thank Dr.
J. Org. Chem, Vol. 70, No. 5, 2005 1641

Matsugi and Curran
data for the crystal structure of 4a, and copies of NMR spectra
of typical products after spe. This material is available free of
charge via the Internet at http://pubs.acs.org.
Steven J. Geib for solving the crystal structure of 4a,
and we thank Dr. Mark Layton for helpful discussions.
Supporting Information Available: Procedures for reac-
tions and spe purifications, spectroscopic data for products,
JO048001N
1642 J. Org. Chem., Vol. 70, No. 5, 2005