Removal of ruthenium using a silica gel supported reagent

Article abstract of DOI:10.1021/ol402339e

A solid-supported isocyanide ligand was developed to destroy active metathesis catalysts and to remove ruthenium byproducts from metathesis reactions. This method was able to significantly reduce the concentration of residual ruthenium from the organic products of several alkene and ene-yne metathesis reactions, under a variety of different conditions.

Full text of DOI:10.1021/ol402339e

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Removal of Ruthenium Using
a Silica Gel Supported Reagent
Jonathan M. French, Caley A. Caras, and Steven T. Diver*
Department of Chemistry, University at Buffalo, The State University of New York,
Amherst, New York 14260, United States
diver@buffalo.edu
Received August 19, 2013
ABSTRACT
A solid-supported isocyanide ligand was developed to destroy active metathesis catalysts and to remove ruthenium byproducts from metathesis
reactions. This method was able to significantly reduce the concentration of residual ruthenium from the organic products of several alkene and
eneÀyne metathesis reactions, under a variety of different conditions.
Metathesis promoted by ruthenium carbenes is a robust
method for creating carbonÀcarbon double bonds in
molecules with high densities of organic functionality. As
a result, catalysts Ru1ÀRu3 have found applications in a
broad range of chemistry from total synthesis1a,b to the
synthesis of pharmaceutically relevant compounds on the
industrial scale.2 Although metatheses are fast and efficient,
the removal of ruthenium during purification is a difficult
problem. The removal of residual ruthenium is necessary, as it
can potentially lead to decomposition or isomerization of the
desired product.3 The use of transition metal catalysis for the
preparation of pharmaceuticals must be weighed against the
need for a cost-effective removal method and a quality control
assay. Herein we disclose a user-friendly filtration method for
the rapid and efficient removal of ruthenium from a metathesis
reaction. A solid-supported isocyanide reagent 2destroys ruthe-
nium carbene activity through a well-defined Buchner reac-
tion and is applicable to a variety of Grubbs catalysts currently
used to promote alkene and eneÀyne metathesis (Scheme 1).
Currently there are several different methods used for
the removal of ruthenium after the reaction is complete.4
These methods are based on two different approaches:
(1) treatingthe catalystwithanadditive once the reaction is
complete; (2) the use of a tailored catalyst with a specially
designed chemical handle to assist catalyst removal.4,5 The
addition of quenching agents that facilitate the removal of
ruthenium catalysts and ruthenium byproducts is attrac-
tive because they can be used with any second generation
Grubbs catalyst, thereby offering a general solution to a
cleanup protocol.
Scheme 1. Supported Reagent 2 and Commonly Used Grubbs
Catalysts
€
(1) (a) Furstner, A. Chem. Commun. 2011, 47, 6505–6511. (b) Nicolaou,
K. C.; Bulger, P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44, 4490–4527.
(2) Yee, N. K.; Farina, V.; Houpis, I. N.; Haddad, N.; Frutos, R. P.;
Gallou, F.; Wang, X.-j.; Wei, X.; Simpson, R. D.; Feng, X.; Fuchs, V.;
Xu, Y.; Tan, J.; Zhang, L.; Xu, J.; Smith-Keenan, L. L.; Vitous, J.;
Ridges, M. D.; Spinelli, E. M.; Johnson, M.; Donsbach, K.; Nicola, T.;
Brenner, M.; Winter, E.; Kreye, P.; Samstag, W. J. Org. Chem. 2006, 71,
7133–7145.
(3) (a) Schmidt, B. Eur. J. Org. Chem. 2004, 2004, 1865–1880.
(b) Hong, S. H.; Sanders, D. P.; Lee, C. W.; Grubbs, R. H. J. Am.
Chem. Soc. 2005, 127, 17160–17161.
(4) Vougioukalakis, G. C. Chem.;Eur. J. 2012, 18, 8868–8880. See
Table 1 for a comparison of ppm levels of ruthenium obtained using each
method.
r
10.1021/ol402339e
XXXX American Chemical Society

Although several purification protocols exist, the iso-
cyanide-triggered Buchner reaction is advantageous over
other methods. Many of the existing methods require long
treatment periods and toxic metals or require expensive
reagents. Examples of these include tris(hydroxymethyl)-
phosphine,6 addition of Pb(OAc)4,7 excess DMSO, or
triphenylphosphine oxide,8 or hydrogen peroxide,9 as well
as the addition of activated carbon followed by column
chromatography.10 Several of these methods require long
treatment times (12 to 24 h) and operate by an undefined
oxidative mechanism. A mesoporous silicate system was
developed which was also found to aid in the removal
of ruthenium.11 The advantage of this system is that it
works relatively quickly, but the material is expensive and
tedious to prepare. Finally, a method developed previously
in our group relies on the addition of KO2CCH2NC to
the reaction mixture.12 The isocyanide ligand is unique
because it rapidly destroys the carbene and causes the
ruthenium to become highly polar. Isocyanide coordina-
tion initiates the insertion of the carbene into the N-hetero-
cyclic carbene (NHC) ligand via a Buchner reaction,13
effectively “turning off” metathesis activity. Second, once
the isocyanide is coordinated to the metal center, it pro-
duces a very polar complex which can be easily removed
using standard column chromatography. The isocyanide-
promoted Buchner reaction was previously shown to
be effective for a wide range of Grubbs catalysts. In this
study, we sought a simple filtration-based method using an
isocyanide-modified silica gel.
the silica gel was subsequently determined by titration.15
Typically the isocyanide was titrated to be between 1.1 and
1.6 mmol/g of silica gel.
The isocyanide-grafted silica gel was able to remove the
Grubbs catalyst Ru2 and the GrubbsÀHoveyda catalyst
Ru3. These catalysts are very efficient at promoting the
ring-closing metathesis (RCM) of diethyl diallylmalonate.
As a qualitative test, Ru2 was exposed to the modified silica
gel 2. After stirring for 30 min, the color of the solution
changed from red to yellow. The silica gel was then removed
and the solvent was collected. To the solution was added
diethyl diallylmalonate; after 2 h no reaction was observed by
TLC or 1H NMR. Identical results were obtained with Ru3
(see Supporting Information (SI) for full experimental details).
The synthesized isocyanide-modified silica gel 2 was
characterized by a distinctive IR band for the isocyanide
byanalogy topreviously reportedprocedures. Graftingthe
isocyanide monomer 1 onto silica gel employed a literature
procedure.14 The polymer-bound isocyanide 2 (Scheme 1)
was detected with reflectance-based IR microscopy. An
absorption signal at 2147 cmÀ1 was found, corresponding
to the surface-bound isocyanide (Figure 1). This value was
in good agreement with the absorption band found for
the monomer, 2150 cmÀ1, and in agreement with pre-
viously published results.14 Loading of the isocyanide on
Figure 1. (A) Optical image of a single particle of silica gel 2 with
aperture area (70 μm  70 μm) marked. (B) Focal plane array
(FPA) image of 2 integrated to show relative abundance of
isocyanide (2135À2147 cmÀ1). (C) Full FT-IR spectrum of 2;
the isocyanide peak (2147 cmÀ1) is noted. (D) Isocyanide region
for single beads of silica gel 2, quench product of silica 2-Ru2,
and the quench product of silica 2-Ru3.
(5) Some recent approaches: (a) Allen, D. P.; Van Wingerden, M. M.;
Grubbs, R. H. Org. Lett. 2009, 11, 1261–1264. (b) Hobbs, C.; Yang
Y.-C.; Ling, J.; Nicola, S.; Su, H.-L.; Bazzi, H. S.; Bergbreiter, D. E. Org.
€
Lett. 2011, 13, 3904–3907. (c) Monge-Marcet, A.; Pleixats, R.; Cattoen,
The shift in the stretching frequency of the isocyanide
was consistent with metal binding. The IR frequencies of
the catalysts Ru2 and Ru3 deactivated by silica gel 2 were
compared to the solution-based results of the isocyanide
monomer 1 and metal-bound isocyanides (Table 1). The
previous study conducted in our lab demonstrated that
more than one isocyanide will coordinate to the metal
center to make an 18-electron complex.13 Catalyst deacti-
vation with silica gel 2 resulted in two different isocyanide
absorption bands indicating thatmore thanone isocyanide
may be coordinated to the metal center (Table 1, entries 3
and 5; Figure 1d). The IR frequencies of the isocyanides in
the silica 2-based and solution-based coordinated products
of Ru2 and Ru3 were in close agreement with each other
(Table 1, entry 3 vs 4; entry 5 vs 6).
X.; Wong Chi Man, M. J. Mol. Cat. A 2012, 357, 59–66. (d) Review:
Clavier, H.; Grela, K.; Kirschning, A.; Mauduit, M.; Nolan, S. P.
Angew. Chem., Int. Ed. 2007, 46, 6786–6801.
(6) Maynard, H. D.; Grubbs, R. H. Tetrahedron Lett. 1999, 40, 4137–
4140.
(7) Paquette, L. A.; Schloss, J. D.; Efremov, I.; Fabris, F.; Gallou, F.;
ꢀ
Mendez-Andino, J.; Yang, J. Org. Lett. 2000, 2, 1259–1261.
(8) Ahn, Y. M.; Yang, K.; Georg, G. I. Org. Lett. 2001, 3, 1411–1413.
(9) Knight, D. W.; Morgan, I. R.; Proctor, A. J. Tetrahedron Lett.
2010, 51, 638–640.
(10) Cho, J. H.; Kim, B. M. Org. Lett. 2003, 5, 531–533.
(11) McEleney, K.; Allen, D. P.; Holliday, A. E.; Crudden, C. M.
Org. Lett. 2006, 8, 2663–2666.
(12) Galan, B.; Kalbarczyk, K.; Szczepankiewicz, S.; Keister, J. B.;
Diver, S. T. Org. Lett. 2007, 9, 1203–1206.
(13) Galan, B. R.; Pitak, M.; Gembicky, M.; Keister, J. B.; Diver,
S. T. J. Am. Chem. Soc. 2009, 131, 6822–6832.
(14) Berry, M.; Champaneria, R. K.; Howell, J. A. S. J. Mol. Catal.
1986, 37, 243–252.
(15) Arshady, R.; Ugi, I. Talanta 1984, 31, 842–843.
B
Org. Lett., Vol. XX, No. XX, XXXX

Table 1. Observed Isocyanide IR Frequencies
Table 2. Optimization of Removal of Ruthenium from a Standard
Metathesis Reaction
entry
reagent
IR frequency (cmÀ1
)
1
2
3
4
5
6
monomer 1
2150
silica grafted isocyanide 2
silica quenched isocyanide, 2-Ru2
product of 1-Ru2
2147
2145, 2172
2143, 2184
2139, 2172
2132, 2174
silica quenched isocyanide, 2-Ru3
product of 1-Ru3
The removal of residual ruthenium was evaluated in
a standard assay, the RCM of diethyl diallylmalonate
(Table 2). The residual ruthenium was determined by in-
ductively coupled plasma mass spectrometry (ICP-MS).16
As a control, untreated silica gel only partially removed
ruthenium (Table 2, entry 1). The solution treatment with
soluble isocyanide KO2CCH2NC and passage through a
silica gel plug (our previous method) gave a 10-fold reduc-
tion in ruthenium (Table 2, entry 2). Next, addition of 30
equiv and 60 equiv of isocyanide-grafted silica gel 2 was
tested (Table 2, entries 3 and 4).17 60equiv of 2 worked best
but required at least 30 min; a longer reaction time was
found to be unnecessary (Table 2, entries 5À7). During the
first 30 min, a color change from red/brown to a light
yellow was noted. Entry 8 demonstrated that doubling the
amount of 2 to 1 g (120 equiv) further reduced the residual
concentration of ruthenium. To evaluate the limits of this
method, a reaction was run with an increased catalyst
loading (Table 2, entry 9). This method was also found to
work with the Ru3 catalyst with similar efficiency at 2.5
and 10mol % catalystloadings(Table2, entries10and 11).
Ruthenium from Ru1 was removed (Table 2, entry 12), but not
as well as the case observed with second generation catalysts.
Isocyanides react with Ru1 by a different mechanism.18
When combined with silica gel purification, the ruthe-
nium was removed to ca. 1 ppm. With the simple RCM
reaction, purification of the cyclopentene product is trivial
since the only byproduct is ethylene gas. More commonly,
column chromatography is needed to purify the desired
product. Entry 13 combines this method with column
chromatography. The reaction was stirred with 60 equiv
2 for 30min and then directly concentrated, and dry loaded
on a pre-equilibrated 1 Â 10 cm bed of silica gel. The
product was eluted with 10% ethyl acetate in hexanes to
afford the desired cyclopentene product in 93% yield. This
result demonstrates that, in combination with column
chromatography, silica gel 2 can remove over 99.99% of
the original ruthenium, leaving behind 1.2 ppm Ru in the
final sample.
a Concentration expressed as μg [Ru]/5 mg of crude, evaporated
organic sample. b Using 500 mg of untreated silica gel. c This reaction
was subsequently purified over 5.2 g (1 cm  10 cm) of silica gel.
Next we sought to evaluate this method for catalysts
under more typical reaction conditions (Table 3). Higher
temperatures are often used to promote difficult ring
closing and cross-metathesis reactions. Elevated reaction
temperatures were screened because they can lead to
catalyst decomposition and olefin isomerization.3b In al-
kene metathesis, a ruthenium methylidene is the propagat-
ing species which has been shown to thermally decompose
to a dinuclear hydride species.19 We wanted to test for
“clean-up” of ill-defined decomposition products, not just
active ruthenium carbenes. Depending on the application
and the temperature used, ruthenium carbenes may no
longer be present at the end of a metathesis reaction.
Entries 1 and 2 demonstrated that this method worked
equally well for high temperature reactions. It was also
noticed that, during treatment with 2, the color change
from red (Ru2), or green (Ru3), to yellow took mere
minutes. To be sure the effectiveness of this treatment
(16) Residual Ru is commonly expressed as μg [Ru]/5 mg of the
evaporated crude organic sample. We have included the ppm levels of
Ru in the organic sample, where 0.005 μg of Ru in a 5 mg sample is equal
to 1 ppm Ru. Lower ppm values are actually recorded in the ICP-MS
quantitative assay, but these correspond to diluted samples and are not
included here. See SI for further details.
(17) The silica gel tended to cling to the walls of the glass vessel at
lower loadings. This resulted in less silica gel in contact with the
homogeneous catalyst.
(18) Galan, B. R.; Pitak, M.; Keister, J. B.; Diver, S. T. Organome-
tallics 2008, 27, 3630–3632.
(19) Hong, S. H.; Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2004,
126, 7414–7415.
Org. Lett., Vol. XX, No. XX, XXXX
C

was not a product of the temperature but due to silica gel 2;
two additional experiments were run. The reaction was
performed at 80 °C with a rt quench (Table 3, entry 3) and
at rt with a quench performed at 80 °C (Table 3, entry 4).
Similar values of residual ruthenium were obtained, in-
dicating that the temperature of the quenching procedure
is not a critical variable.
Scheme 2. Evaluation of the Isocyanide 2 in Additional
Metathesis Reactionsa
Table 3. Evaluation of Temperature as a Variable
temp treatment concn
concn
Ru
entry RuX (°C) temp (°C) [Ru]a [Ru]/ppm removed/%
1
2
3
4
Ru2
Ru3
Ru3
Ru3
80
80
80
rt
80
80
rt
0.59
0.03
0.02
0.03
118
6
99.01
99.96
99.96
99.95
4
80
6
a Concentration expressed as μg [Ru]/5 mg of crude, evaporated
organic sample.
a Isolated yield, and concentration of ruthenium in μg Ru/5 mg of
sample, and ppm Ru in the organic sample. Method A: Filtration,
column chromatography, and concentration of the reaction mixture
prior to ICP-MS analysis. Method B: Filtration and concentration of the
reaction mixture, and ICP-MS analysis.
Finally the effectiveness of this method was evaluated in
more challenging metathesis applications (Scheme 2). The
synthesis of trisubstituted alkenes requires more forc-
ing conditions and a more active catalyst.20 With higher
temperatures, metathesis reactions that can form Ru=
CH2 intermediates and reactions done in the presence of
alcohols result in greater decomposition of ruthenium
carbenes.3,19 The sterically permissive complex Ru4 was
evaluatedfor thisRCM (eq3). Two methods were used: (a)
filtration to remove 2 followed by purification with silica
gel; (b) filtration of 2 through a fritted funnel. Both
samples provided low levels of ruthenium in varying
chemical yields. Next an eneÀyne metathesis reaction with
an alkenol was evaluated using catalysts Ru2 and Ru3
(eq 4). This reaction is also known to cause catalyst
decomposition resulting in the formation of ruthenium
hydrides.21 Even with several different ruthenium species
present at the end of this reaction, quenching with silica gel
2 was effective at reducing the ruthenium content in the
diene product.22 Finally, the recently reported catalyst
Ru523 was effectively quenched after the Z-selective homo-
dimerization of methyl 10-undecenoate (eq 5). Catalyst
Ru5 is also known to have a unique mechanism of
decomposition, different from the earlier generations of
ruthenium catalysts.24 Purification with silica gel 2 signifi-
cantly reduced the levels of ruthenium.
In conclusion, isocyanide-grafted silica gel 2 was found
to be an effective reagent for the removal of ruthenium
from metathesis reactions. Metathesis activity was com-
pletely stopped in typical applications with commonly
used second generation Grubbs metathesis catalysts. This
method can be used as a stand alone procedure simply
requiring filtration or can be used in combination with
column chromatography to further reduce the levels of
residual Ru. Further studies of this grafted silica gel with
other metal catalysts are currently underway.
Acknowledgment. This work was partially supported
by the NSF (CHE-1012839) and by a Kapoor Graduate
Fellowship awarded to J.M.F. The authors would like to
thank Professor F. V. Bright (SUNY-Buffalo) for use of
his IR microscope. We also thank Dr. J. C. Vinci (SUNY-
Buffalo) for helpful discussions and Dr. Richard Pederson
(Materia) for generously providing the Grubbs catalysts
used in this study.
(20) Stewart, I. C.; Ung, T.; Pletnev, A. A.; Berlin, J. M.; Grubbs,
R. H.; Schrodi, Y. Org. Lett. 2007, 9, 1589–1592.
(21) (a) Clark, D. A.; Clark, J. R.; Diver, S. T. Org. Lett. 2008, 10,
2055–2058. (b) Clark, J. R.; Diver, S. T. Org. Lett. 2011, 13, 2896–2899.
(22) A longer reaction time (5 mol % Ru2, 18 h, eq 4) results in greater
decomposition of the catalyst, but the silica gel 2 is still an effective
purification procedure. Method A resulted in 0.058 μg/5 mg sample, or
11.7 ppm residual Ru. See SI.
(23) Keitz, B. K.; Endo, K.; Patel, P. R.; Herbert, M. B.; Grubbs,
R. H. J. Am. Chem. Soc. 2011, 134, 693–699.
(24) Herbert, M. B.; Lan, Y.; Keitz, B. K.; Liu, P.; Endo, K.; Day, M.;
Supporting Information Available. Experimental pro-
cedures characterization data for compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Houk, K. N.; Grubbs, R. H. J. Am. Chem. Soc. 2012, 134, 7861–7866.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX