Polymer and silica supported tridentate schiff base vanadium catalysts for the asymmetric oxidation of ethyl mandelate - Activity, stability and recyclability

Article abstract of DOI:10.1002/adsc.200800486

Homogeneous tridentate Schiff base vanadium catalysts derived from salicylaldehydes and tert-leucinol or tert-leucine are known to be excellent catalysts for the asymmetric oxidation of α-hydroxy esters including ethyl mandelate. Herein, new analogous supported, semi-soluble and insoluble catalysts are synthesized and their activities relative to the homogeneous catalyst are reported. The new catalysts are characterized by ¹H and ¹³C NMR spectroscopy, mass spectrometry (EI, ESI, FAB), X-ray crystallography, elemental analysis, gel permeation chromatography (GPC), Fourier transform infrared (FT-IR) spectroscopy, and nitrogen physisorption. The effects of support material, synthesis procedure, and reaction solvent are examined to probe the utility of these catalysts. Linear poly(styrene) supported catalysts are partially soluble under the reaction conditions, and it is shown that the soluble species contribute significantly to the catalytic reactivity. Insoluble catalysts based on the same vanadyl complexes supported on cross-linked poly(styrene) resin or mesoporous silica allow for catalyst recovery and recycle, showing equivalent selectivities over multiple reaction cycles. The mesoporous silica supported catalyst exhibits greater selectivity than the analogous homogeneous and polymer supported catalysts. Rigorous recycle studies show a loss of activity in each recycle, which is attributed to the decomposition of some portion of the vanadyl complexes in each cycle.

Full text of DOI:10.1002/adsc.200800486

FULL PAPERS
DOI: 10.1002/adsc.200800486
Polymer and Silica Supported Tridentate Schiff Base Vanadium
Catalysts for the Asymmetric Oxidation of Ethyl Mandelate –
Activity, Stability and Recyclability
Rebecca A. Shiels,a Krishnan Venkatasubbaiah,a and Christopher W. Jonesa,*
a
School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, GA
30332, USA
Fax : (+1)-404-894-2866; phone: (+1)-404-385-1683; e-mail: cjones@chbe.gatech.edu
Received: August 4, 2008; Revised: October 22, 2008; Published online: November 19, 2008
Abstract: Homogeneous tridentate Schiff base vana- tions, and it is shown that the soluble species contrib-
dium catalysts derived from salicylaldehydes and ute significantly to the catalytic reactivity. Insoluble
tert-leucinol or tert-leucine are known to be excellent catalysts based on the same vanadyl complexes sup-
catalysts for the asymmetric oxidation of a-hydroxy ported on cross-linked polyACTHNUTRGNE(UNG styrene) resin or meso-
esters including ethyl mandelate. Herein, new analo- porous silica allow for catalyst recovery and recycle,
gous supported, semi-soluble and insoluble catalysts showing equivalent selectivities over multiple reac-
are synthesized and their activities relative to the ho- tion cycles. The mesoporous silica supported catalyst
mogeneous catalyst are reported. The new catalysts exhibits greater selectivity than the analogous homo-
1
are characterized by H and 13C NMR spectroscopy, geneous and polymer supported catalysts. Rigorous
mass spectrometry (EI, ESI, FAB), X-ray crystallog- recycle studies show a loss of activity in each recycle,
raphy, elemental analysis, gel permeation chromatog- which is attributed to the decomposition of some
raphy (GPC), Fourier transform infrared (FT-IR) portion of the vanadyl complexes in each cycle.
spectroscopy, and nitrogen physisorption. The effects
of support material, synthesis procedure, and reac-
tion solvent are examined to probe the utility of Keywords: immobilization; kinetic resolution; oxida-
these catalysts. Linear poly
ACHTUNGTNER(NUNG styrene) supported cata- tion; polymer-supported catalyst; silica supported-
lysts are partially soluble under the reaction condi- catalyst
Introduction
esters by the selective oxidation of only one enantio-
mer (Scheme 1).[11–13] This reaction is particularly at-
The ability to synthesize enantiomerically pure a-hy- tractive not only due to the high enantioselectivity it
droxy esters is of significant interest because these imparts but also due to the use of molecular oxygen
compounds are useful building blocks for chiral syn- at atmospheric pressure as the primary oxidant. Mo-
theses in the pharmaceutical and fine chemical indus- lecular oxygen is highly abundant, low in cost, and
tries.[1] A great deal of research has been devoted to non-toxic, which make it more attractive than other
discovering catalytic systems that can efficiently gen- oxidants, particularly inorganic oxidants that can be
erate enantiomerically pure a-hydroxy esters and expensive and potentially hazardous to dispose of
other secondary alcohols,[2–4] and an important subset after use.
of these systems exploit the kinetic resolution of race-
Previous research efforts in our group have investi-
mic substrates through the appropriate selection of gated the immobilization of enantioselective homoge-
catalyst and reaction conditions.[5–10] One such system neous catalysts for improved catalyst performance,
that effectively utilizes kinetic resolution to yield
enantiomerically pure alcohols from racemic mixtures
involves the application of vanadium complexes with
Schiff base ligands for selective oxidation with molec-
ular oxygen. Specifically, tridentate Schiff base vana-
dium catalysts derived from salicylaldehydes and tert-
leucinol or tert-leucine have been shown to promote
the oxidative kinetic resolution of racemic a-hydroxy ester.
Scheme 1. Oxidative kinetic resolution of an a-hydroxy
Adv. Synth. Catal. 2008, 350, 2823 – 2834
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2823

Rebecca A. Shiels et al.
FULL PAPERS
with the key parameters by which performance is sired catalyst functionality. Since previous research in
judged being productivity, selectivity, ease of recovery, our group utilized the second method to synthesize
and recyclability. This research predominantly focused polymeric salen catalysts for the hydrolytic kinetic
on the immobilization of cobalt salen catalysts for ap- resolution of epoxides with much success, we elected
plication in the hydrolytic kinetic resolution (HKR) to use this approach first in this work.[14–16]
of epoxides, and a number of successful polymer sup-
The monomer bearing both the desired ligand func-
ported catalysts have resulted from these efforts.[14–16] tionality and a styryl group capable of being polymer-
Since the tridentate Schiff base vanadium catalysts ized (2) was produced using a synthetic procedure
used in the oxidative kinetic resolution reaction re- similar to those reported in the literature (Scheme 2).
semble “half” salen functionalities, we set out to The formation of 2 was confirmed by 1H NMR,
apply the knowledge gained from the immobilization 13C NMR, mass spectrometry (ESI) and elemental
of the cobalt salen HKR catalysts to the development analysis. To verify and prove the coordination mode
of immobilized vanadium oxidative kinetic resolution of 2 with vanadium, we prepared the vanadyl complex
catalysts. Even though the activities of the homogene- of 2 (metalated-2) using vanadium oxytriisopropoxide
ous catalysts are well-established, there are, to the [VOACTHNUTRGNEG(NU O-i-Pr)3]. Complex formation was confirmed by
best of our knowledge, no reports of analogous immo- FAB mass spectrometry as well as single crystal X-ray
bilized catalysts derived from salicylaldehydes and crystallography. The molecular structure from X-ray
tert-leucinol for the oxidative resolution reaction. crystallography revealed that the vanadium is in a
There are reports of supported manganese salen cata- pentacoordinated environment with a distorted tet-
À
lysts for the oxidative kinetic resolution of secondary ragonal pyramid geometry. The V O bond distances
alcohols, although these reports utilize diacetoxyiodo- are comparable with literature precedents and de-
benzene as the oxidant instead of molecular crease in the order of phenolate oxygen [1.895(5) ꢁ],
oxygen.[10,17,18] Also, in one case, the enantiomeric ex- alcoholato oxygen [1.839(6) ꢁ], ethanolato oxygen
cesses reported were very low.[17] Therefore, we [1.756(6) ꢁ] and V=O oxygen [1.602(6) ꢁ].[19] The
À
sought to develop active and selective supported cata- V1 N1 distance is 2.079(5) ꢁ, which is well within
lysts for this reaction that are straight-forward to syn- the range of reported literature values of similar com-
thesize and that utilize molecular oxygen instead of plexes. The ORTEP diagram and metric parameters
an organic oxidant. Since the removal of the residual of metalated-2 are given in Figure 1.
metal from the product is important due to economic,
Linear polymers of 2 were prepared by the AIBN
environmental, and product quality issues, the ability initiated free radical homopolymerization of the
to have a recoverable and recyclable catalyst that can styryl monomer (2a) or copolymerization of the styryl
be easily removed from the product mixture is very monomer with styrene (2b) (Scheme 3). 1H NMR
attractive. We report here our description of the im- spectroscopy of the polymers exhibited the typical
mobilization of these vanadium catalysts on semi- broadening features that are expected for macromole-
soluble polymers, on insoluble polymers and on SBA- cules (Figure 2). Also, the peaks associated with the
15 mesoporous silica, and we demonstrate their utility vinyl protons in the monomer spectrum (d=5.0–6.0)
in the oxidative kinetic resolution of ethyl mandelate. were absent in the polymer spectra, which indicated
Particular attention is paid to a rigorous study of cata- that residual monomer was removed during the isola-
lyst recycle.
tion of the polymers. Using gel-permeation chroma-
tography (GPC) in THF with poly(styrene) standards,
AHCTUNGTRENNUNG
the number average molecular weight (Mn) of the ho-
mopolymer was determined to be 5,330 with a poly-
dispersity index (PDI) of 2.16. This corresponds to an
Results and Discussion
Polymer Catalyst Syntheses and Characterization (2a’,
2b’ and 2c’)
In general, there are two strategies for developing im-
mobilized polymer catalysts. The first strategy in-
volves grafting the desired catalyst functionality onto
an insoluble polymeric support such as a resin. This is
an efficient way to create a heterogeneous catalyst,
but if the procedure is performed in a step-wise
manner or if the surface reaction is not quantitative,
many types of species can exist on the resin surface.
In order to circumvent this issue, a second strategy
for immobilization is employed which involves the
direct polymerization of monomers containing the de- Scheme 2. Synthesis procedure for catalyst ligands.
2824
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2823 – 2834

Polymer and Silica Supported Tridentate Schiff Base Vanadium Catalysts
FULL PAPERS
Figure 1. Molecular structure of metalated 2 with thermal el-
lipsoids at the 30% probability level. Hydrogen atoms
except the one attached to C8 are omitted for clarity. Select-
À
ed interatomic distances (ꢁ) and angles (8): distances V1
À
À
À
O1 1.895(5), V1 O2 1.602(6), V1 O3 1.756(6), V1 O4
Figure 2. 1H NMR spectra of styryl functionalized monomer
2, homopolymer 2a, and copolymer 2b.
À
À
À
À
1.839(6), V1 N1 2.079(5); angles O1 V1 O4 140.9(3), O2
V1 O4 108.6(3), O2 V1 O3 103.9(3), O3 V1 N1 160.1(3)
À
À
À
À
À
average of 14 repeat units in the polymer chain. The
The metalated polymers were characterized by
Mn of the copolymer was determined to be 8,439 with Fourier transform infrared spectroscopy (FT-IR)
a PDI of 2.01. This corresponds to an average of 11 (Figure 3). The spectra for homopolymer 2a’ and co-
catalyst repeat units with a 2:5 ligand to styrene ratio polymer 2b’ closely resemble one another, and both
1
(determined from H NMR). These data are summar- exhibit the expected peaks that indicate the presence
ized in Table 1.
The homo- and copolymers were metalated using Particularly noteworthy is the strong C=N stretch lo-
VO
(O-i-Pr)3 to form catalysts 2a’ and 2b’. The vanadi- cated at 1617 cmÀ1 which is indicative of the coordina-
of the immobilized catalyst functionality (Table 2).
ACHTUNGTRENNUNG
um content of the metalated polymers was deter- tion of the imine nitrogen with the vanadium. This
mined by elemental analysis. In the case of homopoly- stretch would be expected to appear around
mer 2a’, the vanadium loading was 10.3 wt% 1630 cmÀ1 if the ligand was not coordinated with the
(2.02 mmol/g), which corresponds to roughly 87% of vanadium.[20] Also, the V=O stretch is seen at
the Schiff base ligand being metalated. For copolymer 986 cmÀ1 in the homopolymer and 984 cmÀ1 in the co-
2b’, the vanadium loading was 6.0 wt% polymer, and these values are in line with reported
(1.18 mmol/g), which corresponds to roughly 83% of literature values.[13,19–21]
the Schiff base ligand being metalated.
Scheme 3. Synthesis of linear poly
ACHTUNGTRENNUNG
Adv. Synth. Catal. 2008, 350, 2823 – 2834
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2825

Rebecca A. Shiels et al.
FULL PAPERS
Table 1. Polymeric ligand characterization.
Cross-linked polymer 2c was prepared by the co-
polymerization of 2 and divinylbenzene (Scheme 4).
Metalation of 2c with VOACHTNUTRGENNG(U O-i-Pr)3 in ethanol at 608C
Polymer
Mn[a]
PDI[a]
m,n actual[b]
produced the resin catalyst 2c’. Elemental analysis in-
dicated a vanadium loading of 5.3 wt% (1.04 mmol/g),
which corresponds to roughly 68% of the Schiff base
ligand being metalated. The FT-IR spectrum of 2c’
shows a strong C=N stretch at 1620 cmÀ1, indicating
that the vanadium is coordinated with the imine nitro-
gen, and a V=O stretch at 987 cmÀ1, which is consis-
2a
2b
5,330
8,439
2.16
2.01
14, 0
11, 28
[a]
From GPC in THF with polyACTHNUTRGNEUGN(styrene) standards.
[b]
Calculated from Mn and 1H NMR; m=repeat units of
catalyst ligand, n=repeat units of styrene.
Table 2. FT-IR data (KBr disk) for metalated polymeric catalysts 2a’, 2b’, 2c’, and silica catalyst 3a.
2a’ n [cmÀ1]
2b’ n [cmÀ1]
2c’ n [cmÀ1]
3a n [cmÀ1]
À
C H aromatic and aliphatic
stretches
C=N
3020, 2953, 2908,
3055, 3021, 2948, 2908,
3017, 2955, 2914,
2956, 2871,
2845
1617
2867
1617
986
2862
1617
984
2865
1620
987
V=O
955
Figure 3. FT-IR spectra of resin catalyst 2c’, copolymer catalyst 2b’, homopolymer catalyst 2a’, and silica catalyst 3a
Scheme 4. Synthesis of cross-linked resin catalyst 2c’.
2826
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2823 – 2834

Polymer and Silica Supported Tridentate Schiff Base Vanadium Catalysts
FULL PAPERS
Scheme 5. Synthesis of silica supported vanadium Schiff base catalyst 3a.
tent with our results above for the other polymer cat- Oxidative Kinetic Resolution using Metalated-1 as
alysts.
Catalyst
To provide a benchmark for our supported catalysts,
we conducted experiments with both the homogene-
Silica Supported Catalyst Synthesis and
Characterization (3a)
ous catalyst (metalated-1) and unligated VO
ACHTUNGTNER(NUNG O-i-Pr)3
to determine their reactivities toward the oxidation of
A trialkoxysilane functionalized catalyst analogue was ethyl mandelate. The reactivity of metalated-1 was
synthesized as described in the Experimental Section tested in dry acetone, dry acetonitrile, and dry di-
(Scheme 5). This analogue was then grafted onto chloromethane as described in the Experimental Sec-
SBA-15 silica, and the resulting supported catalyst 3a tion for Method 1. The results showed acetone to be
was characterized. FT-IR was used to determine the the best solvent for the reaction, providing >99%
nature of the organic species on the silica surface conversion of S-ethyl mandelate and 99% enantio-
(Figure 3). The spectrum shows the anticipated peaks meric excess (ee) of R-ethyl mandelate in less than
which suggest that the catalyst functionality is sup- 4 h (Figure 4) (total conversion S+R: 54%). Acetoni-
ported on the silica surface (Table 2). Of particular in- trile was also a suitable solvent, providing >99% con-
terest are the C=N stretch at 1617 cmÀ1, indicating the version of S-ethyl mandelate and 99% ee of R-ethyl
vanadium is coordinated with the imine nitrogen, and mandelate in 5 h (total conversion S+R: 50%). How-
the V=O stretch at 955 cmÀ1. These values are in line ever, the resolution was significantly slower in di-
with those reported for similar materials.[20]
chloromethane, reaching only 85% conversion of S-
Nitrogen physisorption provided additional evi- ethyl mandelate and 79% ee of R-ethyl mandelate
dence that the catalyst was immobilized in the meso- after 7 h. This trend is similar to that reported by
pores of the silica. The SBA-15 prior to functionaliza- Toste, although the reaction times found in this report
tion as well as after functionalization exhibits a type are different than those for their system.[11]
IV isotherm with hysteresis, which is indicative of its
mesoporous nature. The bare SBA-15 had a BJH
average adsorption pore diameter of 65 ꢁ, whereas
the vanadium functionalized material had a BJH aver-
age adsorption pore diameter of 62 ꢁ. Also, the BET
surface area for the SBA-15 silica material decreased
from 858 m2/g before functionalization to 325 m2/g
after functionalization. This is typical of what is ob-
served when grafted organometallic or organic species
block access to the micropores that are known to
exist in many samples of SBA-15, and thus much or
all of the microporous surface area is no longer in-
cluded in the total surface area calculation.[22–25]
Figure 4. Enantiomeric excess of R-ethyl mandelate vs. time
for homogeneous catalyst 1 in various solvents.
Adv. Synth. Catal. 2008, 350, 2823 – 2834
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2827

Rebecca A. Shiels et al.
FULL PAPERS
Since the homogeneous catalyst (metalated-1) is 61% conversion of S-ethyl mandelate and 42% ee of
metalated in situ prior to reaction, the catalytic activi- R-ethyl mandelate. We attribute this behavior to the
ty of VO
this compound plays a part in the oxidation of ethyl pared to the catalyst prepared via metalation and iso-
mandelate if it is not fully ligated by the Schiff base lation. Specifically, after addition of VO(O-i-Pr)3 to
ACHTUNGTRENNUNG(O-i-Pr)3 alone was probed to determine if higher solubility of the in situ metalated catalyst com-
AHCTUNGTRENNUNG
ligand. Test reactions were conducted in both dry ace- the solution of ligand 2a dissolved in acetone, the
tone and dry acetonitrile and were allowed to run for products remained fully soluble initially, and no solid
23 h. At the end of this period, the reaction in ace- precipitation was observed for approximately the first
tone yielded a total ethyl mandelate conversion of hour of the reaction. After that time, some solid poly-
only 7% with no increase in ee. The reaction in aceto- meric material began to precipitate from solution.
nitrile gave a total ethyl mandelate conversion of 5% However, when the premetalated complex, 2a’, was
with no increase in ee. These results suggest that any added to acetone, solid polymeric material was always
catalytic contribution from free VO
ACHTUNGTERN(NUNG O-i-Pr)3 in solu- observed in the reaction flask, and complete dissolu-
tion can be neglected relative to the contribution tion was never achieved.
from the metalated ligand.
Copolymer catalyst 2b’ was also studied in the oxi-
dative kinetic resolution of ethyl mandelate. Method
1 (in situ metalation) was examined first. In this case,
it was observed that a portion of metalated polymer
precipitated out of solution during the initial oxygen
pre-treatment prior to reaction. Also, as the polymer
Oxidative Kinetic Resolution using Polymeric
Catalysts 2a’ and 2b’
Since acetone was the most effective solvent for the precipitated, it formed large agglomerates which
oxidative kinetic resolution of ethyl mandelate with broke into smaller particles as the reaction proceeded.
the homogeneous catalyst, all reactions using the Nonetheless, copolymer catalyst 2b’ successfully com-
polymeric catalysts were conducted in acetone. To in- pleted the resolution, taking 9.5 h to reach 99% con-
vestigate the activity of the homo- and copolymers, version of S-ethyl mandelate with 98% ee of R-ethyl
two catalyst metalation methods were employed: in mandelate (total conversion S+R: 58%) (Figure 6).
situ metalation (Method 1) and metalation and isola-
When Method 2 (metalation and isolation) was
tion prior to reaction (Method 2).
used with copolymer catalyst 2b’, the catalyst was
First, homopolymer catalyst 2a’ was assessed using able to reach 99% conversion of S-ethyl mandelate
Method 1 and Method 2. Significant differences in re- and 95% ee of R-ethyl mandelate in 8 h (total conver-
action rates were observed between the two catalyst sion S+R: 61%), even though complete dissolution
metalation methods (Figure 5). The homopolymer of the polymeric material was not observed. This co-
catalyst that was metalated and isolated prior to use polymer catalyst gives roughly comparable reaction
(Method 2) was significantly less active than the ho- rates regardless of metalation method.
mopolymer catalyst that was metalated in situ
Since these catalysts were not fully soluble in the
(Method 1). For instance, after 9 h of reaction, the reaction media, we sought to assess whether the
Method 1 reaction had proceeded to 94% conversion active species were associated with soluble polymeric
of S-ethyl mandelate and 88% ee of R-ethyl mande- species or precipitated polymer species. To probe this,
late, whereas the Method 2 reaction had only reached
Figure 5. Conversion of S-ethyl mandelate vs. time for ho-
mopolymer catalyst 2a with in situ metalation (Method 1)
and metalation and isolation (Method 2).
Figure 6. Conversion of S-ethyl mandelate vs. time for copo-
lymer catalyst 2b with in situ metalation (Method 1) and
metalation and isolation (Method 2).
2828
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2823 – 2834

Polymer and Silica Supported Tridentate Schiff Base Vanadium Catalysts
FULL PAPERS
Table 3. Reaction rate data for catalysts 1, 2a’, 2b’, 2c’ and 3a.
Catalyst
Metalation Method
Solvent
Initial Rate[a]
Time [h]
Total Conversion [%]
ee [%]
s[b]
1
1
1
1
1
1
1
2
1
2
2
2
2
2
acetone
5.7
3.9
2.0
2.1
0.6
3.1
3.5
5
6
10.75
9
9
54
50
48
57
35
58
61
63
48
48
49
98
98
71
96
42
97
97
95
68
96
93
50
458
16
24
12
23
17
12
13
acetonitrile
dichloromethane
acetone
acetone
acetone
acetone
acetone
acetone
acetone
2a’
2a’
2b’
2b’
2c’
2c’
3a
3a
2.5
8
1.8[c]
1.2[d]
1.3
10.5
10.5
11
11
204
211
acetonitrile
1.3
[a]
mol S-ethyl mandelate/(mol Vhour).
[b]
[c]
[d]
s=[ln
First cycle.
(1Àconv.)
N
U
ACHTUNGTRENN(UNG 1+ee)].
Fourth cycle
Table 4. Kinetic resolution of a-hydroxy esters.
copolymer 2b’ was metalated using Method 1 and
used in a series of reactions. After complete conver-
sion of the S-ethyl mandelate in reaction 1, the solid
polymer was removed via filtration and a fresh batch
of reactants was added to the filtrate and monitored
for additional reaction. The reaction in the second,
solid-free reaction proceeded at a rate similar to but
slightly less than the first reaction, suggesting the re-
activity from the first reaction is associated with, at
least in part, soluble polymer catalyst.
The initial rates and selectivity factors for the cata-
lysts are tabulated in Table 3. We note that all the
rates differ by less than a factor of 10. Furthermore,
the selectivity factors for the polymer immobilized
catalysts (2a’, 2b’, 2c’) are similar to the those of the
parent homogeneous catalyst, 1, in acetone. Surpris-
ingly, the selectivity factor for the catalyst supported
Entry R1
R2
Time Conversion[a] ee[a]
s[b]
[h]
[%]
[%]
1
2
3
4
Ph
Ph
Ph
OEt
OMe
OCH2Ph 18
10.5
9
63
52
56
54
95
77
91
90
12
14
19
23
Me2CH OCHMe2 87
[a]
[b]
Determined by chiral GC or chiral HPLC.
s=[ln (1Àee)]/[ln(1Àconv)(1+ee)].
(1Àconv)
G
E
G
ACHTUNGTRENNUNG
on mesoporous silica (3a) was significantly higher, Oxidative Kinetic Resolution and Recycling Studies
suggesting that the nature of the support (or linker) using Catalysts 2c’ and 3a
can play an important role in the catalyst selectivity.
To probe the scope of the catalyst 2c’ in asymmetric Solid, completely insoluble catalysts were evaluated
oxidation, reactions were performed with four addi- in these reactions in an effort to develop easily recy-
tional substrates. The results are presented in Table 4. clable systems. First, the cross-linked resin catalyst 2c’
The alkyl substrate (entry 4) resolved with good selec- was evaluated for the kinetic resolution of ethyl man-
tivity followed by substrates bearing a-phenyls. delate. In comparison with the semi-soluble homo-
Within the substrates bearing a-phenyls, the benzyl- and copolymers 2a’ and 2b’, the insoluble resin 2c’
substituted species (entry 3) resolved better than the gave a similar reaction rate. For instance, copolymer
ethyl and methyl derivatives.
2b’ reaches 99% conversion of S-ethyl mandelate in
To assess whether oxygen solubility in the liquid 9 h, whereas the resin 2c’ converts 94% of the S enan-
media might limit the initial rates, two experiments tiomer in the same duration. Silica catalyst 3a was
were run using metalated-1 as catalyst with different also tested in the oxidative kinetic resolution of ethyl
oxygen pressures. After 60 min, the conversion run- mandelate in both dry acetone and dry acetonitrile.
ning under standard conditions with the reaction For the reaction in acetone, the resolution was com-
head-space purged with pure oxygen was comparable plete in 11 h, reaching 99% conversion of S-ethyl
to the conversion when running under 45 psig of mandelate and 98% ee of R-ethyl mandelate (total
oxygen pressure (~51 vs. 57%), suggesting that conversion S+R: 49%). The reaction in acetonitrile
oxygen solubility is not a dominating rate-controlling proceeded at the same rate, also taking 11 h to com-
step under the standard reaction conditions.
plete the resolution.
Adv. Synth. Catal. 2008, 350, 2823 – 2834
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2829

Rebecca A. Shiels et al.
FULL PAPERS
Table 5. Recycle data for the oxidation of ethyl mandelate
with catalyst 2c’ after 16 h.
the elemental analysis of the fresh catalyst 2c’ and the
used catalyst after the fourth cycle. Elemental analysis
indicated a 50% loss in nitrogen and vanadium but
only a 1% loss in carbon. From these results we hy-
pothesize that imine cleavage leads to the loss of va-
nadium which results in decreased catalyst activity. It
should be noted that a similar mode of deactivation
of supported salen catalysts was also suggested in our
recent work on the hydrolytic kinetic resolution of ep-
oxides.[26] Filtration tests using the catalyst 2c’ suggest
that the reaction does not proceed in the absence of
Cycle Loading
(mol%)
Conversion of S-ethyl ee [%] of R-
mandelate [%]
ethyl mandelate
1
2
3
4
5.5
5.5
5.5
5.5
99
99
>99
>99
96
96
>99
>99
Recycling experiments were done using the insolu- the solid catalyst (Figure 8).[27]
ble catalysts 2c’ and 3a by recovering the catalysts, We also performed recycling experiments using
washing, and reusing them in subsequent reactions silica supported catalyst 3a. In this case, a gradual de-
with fresh reagents. The conversion of S-ethyl mande- crease in activity in each cycle was again observed.
late after 16 h was 99% using 2c’ over four cycles Elemental analysis of the catalyst after the third cycle
(Table 5). We note that standard practice in the litera- showed a decrease in the nitrogen and carbon percen-
ture describing supported catalysts would suggest a tages but surprisingly, no decrease in the vanadium
conclusion of “catalysts can be recycled without loss loading. We propose that after complex degradation
of activity” would be reasonable. However, this would due to imine cleavage (or other mechanisms) the re-
be an improper conclusion, as it is based on measur- leased vanadyl ions may be captured by the silica sup-
ing reaction yields at long times. In contrast, measure- port. This would explain the observed loss of activity
ments of the reaction kinetics clearly suggest there is without loss of vanadium from the silica catalyst and
a gradual decrease in reaction rate upon recycle. A is consistent with the known stability of vanadia/silica
comparison of the kinetic profiles of the S enantiomer materials.[28,29] To support this possibility, we stirred
conversion in the first and fourth reaction cycles using SBA-15 with vanadium oxytriisopropoxide in acetone
catalyst 2c’ is given in Figure 7. After 10.5 h, the con- for 18 h, filtered the solid material, washed it with
version in the first cycle was 98% of the S enantio- acetone, and dried it under vacuum. The vanadium
mer, whereas the conversion was only 83% after the content (1.65 mmol/g or >65% of the added vanadi-
same time in the fourth cycle. The initial rate calculat- um) of this material suggested that SBA-15 will effec-
ed for the first reaction cycle was 1.8 hÀ1, whereas the tively capture leached vanadyl ions from decomposed
initial rate calculated for the fourth reaction cycle was complexes.
1.2 hÀ1. These results clearly suggest that there is a de-
We note that it is quite uncommon for researchers
crease in activity with each cycle even though the to study recyclability of supported coordination com-
final conversions are similar at long times. To probe plex catalysts in the manner described in this section.
the reason for the decrease in activity, we compared Rigorously analyzing for catalyst decomposition and
studying kinetics upon recycle is not routinely prac-
ticed. Rather, many researchers simply recycle cata-
lysts and measure yields only at very long times. Such
Figure 7. Recycle experiments for the conversion of S-ethyl
mandelate vs. time with catalyst 2c’, comparison of the first
Figure 8. Filtration test using catalyst 2c’: the reaction was
and fourth reaction cycle.
filtered after 3 h to probe the heterogeneity of the reaction.
2830
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2823 – 2834

Polymer and Silica Supported Tridentate Schiff Base Vanadium Catalysts
FULL PAPERS
an approach effectively hides complex deactivation mesoporous silica supported catalyst (3a) proved to
and leads to the false conclusion that the catalysts are be more selective than the other supported and ho-
fully stable and recyclable. The decomposition of the mogeneous catalysts, suggesting that the nature of the
imine-containing ligand here and in our related work supported catalyst (or linker) can influence the reac-
on Co-salen catalysts suggest that the instability of li- tion selectivity. The insoluble catalysts (2c’ and 3a)
gands that contain this imine functional group may were active for the kinetic resolution of ethyl mande-
limit the number of recycles that can be achieved in late and were recycled with similar yields after long
many cases.
reaction times for 4 cycles. If only these long time
yields were reported, it could be suggested that the
catalysts were recyclable with little loss in activity and
equal selectivity. However, unlike most studies, recy-
cle kinetics and catalyst decomposition were probed
in this work, and they clearly demonstrated there was
On the Mechanism of the Oxidative Kinetic
Resolution
In our previous work on supported Co-salen catalysts reduced catalyst activity in each run. Elemental analy-
for the hydrolytic kinetic resolution of epoxides, the sis suggested cleavage of the imine functionality in
reaction rates could be substantially enhanced over the ligand and subsequent decomposition of the com-
the rate using the parent small molecule complex by plex is a likely cause of the decreased catalyst activity
using properly designed supported catalysts. This is upon recycle.
due to the importance of a hypothesized cooperative
bimetallic slow step in the catalytic cycle, whereby
two Co-salen complexes simultaneously interact with
Experimental Section
the epoxide reactant. As Toste and Chen have both
hypothesized a mechanism for the oxidative kinetic
resolution of a-hydroxy esters that includes a step
where two vanadium complexes interact, we sought to
see if there was any influence of catalyst structure on
General Remarks
3,5-Di-tert-butyl-2-hydroxybenzaldehyde (Aldrich), (S)-tert-
leucinol
(Aldrich),
3-isocyanatopropyltrimethoxysilane
(O-i-
rate.[11,12] The initial rate data in Table 3 suggest that (Gelest), and vanadium(V) oxytriisopropoxide [VO
ACHTUNGTRENNUNG
Pr)3] (Aldrich) were used as received. 2,2’-Azo-bis(isobutyr-
onitrile) (AIBN) (Aldrich) was dried under vacuum and
stored in a nitrogen dry box. Styrene was distilled and
stored in a nitrogen dry box. Divinylbenzene (Alfa Aesar)
was treated with 20% sodium hydroxide to remove the sta-
bilizer, washed with water, and dried over sodium sulfate
prior to use. Dichloromethane and acetonitrile were dried
over calcium hydride and distilled; chlorobenzene was dried
with 4 ꢁ molecular sieves and distilled under vacuum; ace-
tone (BDH) was dried over anhydrous calcium sulfate and
all the catalysts studied here, including the parent
small molecule complex, the semi-soluble polymers
and the insoluble solid catalysts all catalyze the reac-
tion with roughly similar rates. In addition, no en-
hancement in rates is seen using the supported cata-
lysts (which force catalytic sites to be in close proxim-
ity) compared to the small molecule, parent complex
at lower catalyst loadings. Hence, in this work there
appears to be no evidence for the presence of a bi-
metallic step that is kinetically significant in this reac- distilled. Toluene and tetrahydrofuran (THF) were used
after purification and drying via a packed bed solvent
tion, although this does not preclude the occurrence
of a kinetically insignificant bimetallic step.
system made of dual alumina columns.[30]
1
Solution H and 13C NMR spectroscopies were conducted
on a Varian Mercury Vx 400 spectrometer. FT-IR spectros-
copy was conducted on a Bruker Equinox 55 using disks
made from solid catalyst samples dispersed in dry potassium
bromide (KBr). The samples were analyzed using 1024
scans with a resolution of 4 cmÀ1.
For single crystal X-ray diffraction, a suitable crystal of
metalated-2 was coated with Paratone N oil, suspended in a
Conclusions
We have developed four new supported tridentate
Schiff base vanadium catalysts for the oxidative kinet-
ic resolution of ethyl mandelate: a semi-soluble ho-
mopolymer catalyst (2a’), a semi-soluble copolymer small fiber loop, and placed in a cooled nitrogen gas stream
at 173 K on a Bruker D8 APEX II CCD sealed tube diffrac-
tometer with graphite monochromated MoKa (0.71073 ꢁ)
radiation. Data were measured using a series of combina-
tions of phi and omega scans with 10 s frame exposures and
0.58 frame widths. Data collection, indexing and initial cell
refinements were carried out using APEX II software.[31]
Frame integration and final cell refinements were done
using SAINT software.[32] The final cell parameters were de-
termined from least-squares refinement on 3626 reflections.
The structure was solved using direct methods and differ-
catalyst (2b’), an insoluble cross-linked resin catalyst
(2c’) and an insoluble silica supported catalyst (3a).
The ligands and metalated catalysts have been charac-
terized by a number of quantitative and qualitative
techniques to ensure that the desired functionalities
are present in the final supported catalysts. These cat-
alysts are all active for oxidative kinetic resolution of
ethyl mandelate, and all catalysts displayed roughly
similar activity, with the supported catalysts somewhat
slower than the parent homogeneous complex. The ence Fourier techniques (SHELXTL, V6.12).[33] Hydrogen
Adv. Synth. Catal. 2008, 350, 2823 – 2834
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2831

Rebecca A. Shiels et al.
FULL PAPERS
atoms were placed in their expected chemical positions
using the HFIX command and were refined as riding atoms.
All non-hydrogen atoms were refined anisotropically. Crys-
tallographic data for the structure of metalated-2 has been
deposited with the Cambridge Crystallographic Data Center
as supplementary publication no. CCDC 695255. These data
can be obtained free of charge from The Cambridge Crystal-
lographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif or on application to CCDC, 12 Union Road,
Cambridge CB21EZ, UK (fax: (+44) 1223-336-033; email:
deposit@ccdc.cam.ac.uk).
Nitrogen physisorption measurements were taken on a
Micromeritics ASAP 2010 at 77 K. Prior to analysis, the as-
synthesized SBA-15 sample was degassed with heating at
1508C under vacuum overnight, and the functionalized
SBA-15 sample was degassed with heating at 508C under
vacuum overnight. Gas chromatography was used to moni-
tor the reaction progress and was conducted on a Shimadzu
GC 17 A equipped with a flame-ionization detector and a
Chiraldex g-TA column (length=40 m, inner diameter=
0.25 mm, and film thickness=0.25 mm). The following tem-
perature profile was used: heat to 908C, ramp to 1408C at a
rate of 68CminÀ1, and hold for 11 min.
7.37 (d, J=2.12 Hz, 1H), 7.48 (m, 4H), 7.58 (d, J=2.12 Hz,
1H), 8.42 (s, 1H), 13.94 (s, 1H); 13C NMR (101 MHz,
CDCl3): d=27.4, 29.6, 33.5, 35.3, 62.8, 81.6, 113.8, 118.8,
126.90, 126.95, 128.4, 128.7, 130.8, 136.2, 136.7, 138.1, 140.7,
160.4, 167.0; anal. calcd. (%) for C25H33NO2: C 79.11, H
8.76, N 3.69; found: C 79.12, H 8.42, N 3.63; accurate mass
calcd.: 379.2511; found: 379.2505.
For the crystal structure analysis, VOACTHNUTRGEN(UNG O-i-Pr)3 (0.09 g,
0.40 mmol) was added to a solution of 2 (0.17 g, 0.44 mmol)
in methanol (5 mL). The reaction mixture was stirred for
4 h. The resulting dark brown solution was reduced to 2 mL
and kept in the freezer to form brown colored crystals.
Yield: 0.15 g. X-ray quality crystals were grown by recrystal-
lization of metalated-2 from oxygenated ethanol.
Synthesis of Schiff Base Homopolymer Ligand (2a)
The styryl functionalized Schiff base monomer (2) was poly-
merized to form the homopolymer ligand (2a) (Scheme 3).
In a typical preparation, monomer (0.25 g, 0.66 mmol) was
added to a 15-mL pressure tube reaction vessel in a nitrogen
dry box. Then, approximately 3 mL of dry, degassed chloro-
benzene were added along with AIBN (11 mg, 0.067 mmol).
The pressure tube was sealed and placed in an 808C oil bath
and allowed to stir for 48 h. The reaction mixture was then
cooled to room temperature and precipitated into cold hex-
anes. The precipitated polymer was filtered from the solu-
tion, washed with hexanes, and dried under vacuum to give
2a as a yellow solid; yield: 76%; Mn: 5,330; PDI: 2.16.
Synthesis of Schiff Base Ligand (1)
The homogeneous Schiff base ligand (1) was synthesized ac-
cording to literature procedures (Scheme 2).[11,34] 3,5-Di-tert-
butyl-2-hydroxybenzaldehyde (4 g, 17 mmol) was dissolved
in anhydrous methanol. Then, sodium sulfate (12.2 g) and
(S)-tert-leucinol (2 g, 17 mmol) were added, and the mixture
was stirred at room temperature for 4 h. The solution was
filtered through Celite, the solvent was removed, and the
product was dried under vacuum. The synthesis produced 1
as a yellow solid which was used without further purifica-
tion; yield: 98%. 1H NMR (400 MHz, CDCl3): d=0.96 (s,
9H), 1.30 (s, 9H), 1.44 (s, 9H), 2.91 (dd, J=9.48 and
2.72 Hz, 1H), 3.73 (m, 1H), 3.90 (m, 1H), 7.13 (d, J=
2.36 Hz, 1H), 7.39 (d, J=2.32 Hz, 1H), 8.35 (s, 1H), 13.60
(s, 1H); 13C NMR (101 MHz, CDCl3): d=27.4, 29.7, 31.7,
33.5, 34.4, 35.3, 62.8, 81.7, 117.9, 126.4, 127.4, 137.0, 140.4,
158.4, 167.4; anal. calcd. (%) for C21H35NO2: C 75.63, H
10.58, N 4.20; found: C 75.39, H 10.64, N 4.24; accurate
mass calcd.: 333.2668; found: 333.2656.
Synthesis of Schiff Base Copolymer Ligand (2b)
The styryl functionalized Schiff base monomer (2) was copo-
lymerized with styrene to form the copolymer ligand (2b)
(Scheme 3). In a typical preparation, one equivalent of
styryl Schiff base monomer (152 mg, 0.4 mmol) and four
equivalents of styrene (167 mg, 1.6 mmol) were added to a
15-mL pressure tube reaction vessel in a nitrogen dry box.
Then, approximately 3 mL of dry, degassed chlorobenzene
were added along with AIBN (8.2 mg, 0.05 mmol). The pres-
sure tube was sealed and placed in an 808C oil bath and al-
lowed to stir for 48 h. The reaction mixture was then cooled
to room temperature and precipitated into cold hexanes.
The precipitated polymer was filtered from the solution,
washed with hexanes, and dried under vacuum to give 2b as
a yellow solid; yield 63%; Mn: 8,439; PDI: 2.01.
Synthesis of Styryl Functionalized Schiff Base
Monomer (2) and Metalated-2
Synthesis of Resin Supported Catalyst (2c’)
The styryl functionalized Schiff base monomer was prepared
similarly to the homogeneous ligand 1 (Scheme 2). 3-tert-
Butyl-2-hydroxy-5-(4’-vinylphenyl)benzaldehyde was synthe-
sized according to published literature procedures.[14,35] This
styryl aldehyde (1.5 g, 5.35 mmol) was dissolved in anhy-
drous methanol. Then, sodium sulfate (4 g) and (S)-tert-leu-
cinol (0.627 g, 5.35 mmol) were added, and the mixture was
stirred at room temperature for 4 h. The solution was fil-
tered through Celite, the solvent was removed, and the
product was dried under vacuum. The synthesis produced 2
as a yellow solid which was used without further purifica-
tion; yield: 98%. 1H NMR (400 MHz, CDCl3): d=0.98 (s,
9H), 1.48 (s, 9H), 2.95 (dd, J=9.44 and 2.60 Hz, 1H), 3.76
(m, 1H), 3.96 (m, 1H), 5.25 (d, J=10.88 Hz, 1H), 5.77 (d,
J=17.60 Hz, 1H), 6.74 (dd, J=17.60 and 10.88 Hz, 1H),
The styryl functionalized Schiff base monomer (2) was
cross-linked with divinylbenzene to form the resin 2c. In a
typical preparation, monomer 2 (0.50 g, 1.3 mmol) was dis-
solved in chlorobenzene (2.5 mL). Divinylbenzene (80%,
mixture of isomers, 0.34 g, 2.6 mmol) was then added along
with AIBN (0.04 g, 0.25 mmol), and the solution was purged
with argon for 30 min. The solution was then stirred in an
oil bath at 808C for 48 h. The cross-linked polymer 2c ob-
tained was filtered and washed with copious amounts of di-
chloromethane, toluene and methanol and subsequently
dried under vacuum; yield: 0.81 g.
VOACTHUNTRGNE(UGN O-i-Pr)3 (172 mg, 0.70 mmol) was added to a suspen-
sion of resin 2c (0.5 g) in ethanol and stirred in an oil bath
at 608C for 24 h. The resulting metalated resin was filtered
2832
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2823 – 2834

Polymer and Silica Supported Tridentate Schiff Base Vanadium Catalysts
FULL PAPERS
and washed with copious amounts of ethanol and acetone to
flask, followed by 3.5 mL of solvent. Once the ligand was
completely dissolved, VO(O-i-Pr)3 (17.5 mL, 0.075 mmol, 5
afford 2c’; yield: 0.59 g; IR (KBr): n=3017, 2955, 2914,
ACHTUNGTRENNUNG
2865, 1620,1544, 1436, 1269, 1173, 1068, 987, 896, 825 cmÀ1
anal. found: C 75.44, H 7.29, N 2.14, V 5.3.
;
mol%) was added to the flask. The catalyst mixture was al-
lowed to stir under an oxygen blanket for 15 min. At the
completion of this time, a solution of ethyl mandelate
(270.3 mg, 1.5 mmol) and hexamethylbenzene (12.5 mg as
an internal standard) in 4 mL of solvent was added to the
reaction flask, and 100 mL samples were taken periodically
to monitor the extent of the reaction. An oxygen blanket
was maintained in the flask throughout the course of the re-
action.
Synthesis of Hydroxy Functionalized Ligand (3)
3-tert-Butyl-5-chloromethyl-2-hydroxybenzaldehyde was syn-
thesized according to published literature procedures.[36]
This compound was then converted to 3-tert-butyl-2-hy-
droxy-5-(hydroxymethyl)benzaldehyde.[37] The 5-hydroxy-
methyl aldehyde (2.08 g, 10 mmol) was dissolved in anhy-
drous methanol. Then, sodium sulfate (6 g) and (S)-tert-leu-
cinol (1.17 g, 10 mmol) were added, and the mixture was
stirred at room temperature for 4 h (Scheme 2). The solu-
tion was filtered through Celite, the solvent was removed,
and the product was dried. The product was purified via
column chromatography (1:1 hexanes: ethyl acetate) and
then recrystallized from methanol to produce 3 as a pure,
Catalytic Reactions with Pre-Metalated Ligands
(Method 2)
In the second method, the selected ligand was metalated
and isolated prior to reaction. For the polymer catalysts 2a
and 2b, the polymerized ligand first was dissolved in dry
acetone. Then VO
ACHTUNGTRNEN(UNG O-i-Pr)3 (0.9 equivalents) was added, and
1
the solution was stirred under an oxygen blanket for 11=
pale yellow solid; yield: 79%; H NMR (400 MHz, CDCl3):
2
d=0.96 (s, 9H), 1.42 (s, 9H), 1.55 (s, 2H), 2.92 (dd, J=9.48
and 2.76 Hz, 1H), 3.75 (m, 1H), 3.92 (m, 1H), 4.60 (d, J=
2.72 Hz, 2H), 7.16 (d, J=1.84 Hz, 1H), 7.31 (d, J=1.92 Hz,
1H), 8.34 (s, 1H), 13.85 (s, 1H); 13C NMR (101 MHz,
CDCl3): d=27.3, 29.6, 33.5, 35.1, 62.7, 65.6, 81.6, 118.5,
129.1, 129.3, 130.3, 138.0, 160.4, 166.8; anal. calcd. (%) for
C18H29NO3: C 70.32, H 9.51, N 4.56; found: C 70.26, H 9.59,
N 4.51; accurate mass calcd.: 307.2147; found: 307.2116.
hours. The solvent was removed. The metalated polymers
were then dried under vacuum overnight. Cross-linked cata-
lyst 2c’ was metalated as described above and always tested
using Method 2. Silica catalyst 3a was always tested with
Method 2 and was never metalated after the ligand was
anchored to the silica surface.
For the oxidative kinetic resolution, the selected metalat-
ed catalyst (2a’, 2b’, 2c’ or 3a) was added to the reaction
flask at a 5 (or) 5.5 mol% vanadium loading (determined
from elemental analysis) with 3.5 mL of solvent, and it was
allowed to stir under an oxygen blanket for 15 min. At the
completion of this time, a solution of ethyl mandelate
(270.3 mg, 1.5 mmol) and hexamethylbenzene (12.5 mg as
an internal standard) in 4 mL of solvent was added to the
reaction flask, and 100 mL samples were taken periodically
to monitor the extent of the reaction. An oxygen blanket
was maintained in the flask throughout the course of the re-
action.
Synthesis of SBA-15 Supported Catalyst (3a)
The SBA-15 silica support material was synthesized and cal-
cined according to literature procedures, dried under
vacuum at 2008C for 3 h, and then stored in a nitrogen dry
box prior to use.[22,38] The hydroxy functionalized ligand (3)
(0.25 g, 0.81 mmol) was first reacted with VOACTHNURGTNEUNG(O-i-Pr)3
(188 mg, 0.77 mmol) by stirring the compounds in dry ace-
tone under an oxygen blanket for 30 min. The acetone was
removed under vacuum, and the metal complex was re-dis-
solved in dried tetrahydrofuran. Then, 3-isocyanatopropyl-
trimethoxysilane (183 mg, 0.89 mmol) was added to the solu-
tion in a nitrogen drybox along with triethylamine (90 mg,
0.89 mmol), and the solution was stirred at reflux conditions
overnight (Scheme 5).[39,40] The solvent was again removed,
and the compound was dried under vacuum. The silane
functionalized catalyst was dissolved in a small amount of
toluene and added to a solution of calcined SBA-15 (0.45 g)
in toluene in a nitrogen drybox. The slurry was stirred at
reflux conditions overnight. The functionalized silica catalyst
was filtered in the drybox and washed with copious amounts
of toluene, hexanes, and acetone until each solvent ran
clear. Catalyst 3a was dried overnight under vacuum to give
a light brown/maroon powder. The procedure yielded ap-
proximately 0.4 g of compound 3a. Elemental analysis indi-
cated a vanadium loading on the silica of 2 wt%.
Acknowledgements
This work was supported by the DOE BES through Cataly-
sis Science contract DE-FG02-03ER15459. We also thank Dr.
Kenneth I. Hardcastle of Emory University for acquisition of
the X-ray data.
References
[1] G. M. Coppola, H. F. Schuster, a-Hydroxy Acids in
Enantioselective Synthesis VCH, Weinheim, Germany,
1997.
[2] J. B. Arterburn, Tetrahedron 2001, 52, 9765.
[3] M. J. Schultz, M. S. Sigman, Tetrahedron 2006, 52, 8227.
[4] R. Irie, T. Katsuki, Chem. Rec. 2004, 4, 96.
[5] D. E. J. E. Robinson, S. D. Bull, Tetrahedron: Asymme-
try 2003, 13, 1407.
Catalytic Reactions with in situ Metalation
(Method 1)
Oxidative kinetic resolution reactions were preformed via
two methods. In the first method, the ligand was metalated
in situ just prior to the reaction. The selected ligand (1, 2a,
or 2b) (0.0825 mmol, 5.5 mol%) was added to the reaction
[6] W. Sun, H. Wang, C. Xia, J. Li, P. Zhao, Angew. Chem.
Int. Ed. 2003, 42, 1042.
Adv. Synth. Catal. 2008, 350, 2823 – 2834
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2833

Rebecca A. Shiels et al.
FULL PAPERS
[7] Y. Nishibayashi, A. Yanauchi, G. Onodera, S. Uemura,
J. Org. Chem. 2003, 68, 5875.
[8] D. R. Jensen, J. S. Pugsley, M. S. Sigman, J. Am. Chem.
[24] J. M. Richardson, C. W. Jones, J. Catal. 2007, 251, 80.
[25] K. Miyazawa, S. Inagaki, Chem. Commun. 2000, 2121.
[26] C. S. Gill, K. Venkatasubbaiah, N. T. S. Phan, M. Weck,
C. W. Jones, Chem. Eur. J. 2008, 14, 7306.
Soc. 2001, 123, 7475.
[9] E. M. Ferreira, B. M. Stoltz, J. Am. Chem. Soc. 2001,
123, 7725.
[27] R. A. Sheldon, M. Wallau, I. Arends, U. Schuchardt,
Acc. Chem. Res. 1998, 31, 485.
[10] W. Sun, X. Wu, C. Xia, Helv. Chim. Acta 2007, 90, 623.
[11] A. T. Radosevich, C. Musich, F. D. Toste, J. Am. Chem.
Soc. 2005, 127, 1090.
[12] S.-S. Weng, M.-W. Shen, J.-Q. Kao, Y. S. Munot, C.-T.
Chen, Proc. Natl. Acad. Sci. USA 2006, 103, 3522.
[13] C.-T. Chen, S. Bettigeri, S.-S. Weng, V. D. Pawar, Y.-H.
Lin, C.-Y. Liu, W.-Z. Lee, J. Org. Chem. 2007, 72, 8175.
[14] X. Zheng, C. W. Jones, M. Weck, Chem. Eur. J. 2006,
12, 576.
[28] P. VanderVoort, K. Possemiers, E. F. Vansant, J. Chem.
Soc. Faraday Trans. 1996, 92, 843.
[29] M. Baltes, P. Van Der Voort, O. Collart, E. F. Vansant,
J. Por. Mater. 1998, 5, 317.
[30] A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K.
Rosen, F. J. Timmers, Organometallics 1996, 15, 1518.
[31] APEX II, Bruker AXS, Inc., Analytical X-ray Systems,
5465 East Cheryl Parkway, Madison WI 53711–5373.
[32] SAINT Version 6.45 A, Bruker AXS, Inc., Analytical
X-ray Systems, 5465 East Cheryl Parkway, Madison WI
53711–5373.
[15] X. Zheng, C. W. Jones, M. Weck, J. Am. Chem. Soc.
2007, 125, 1105.
[16] X. Zheng, C. W. Jones, M. Weck, Adv. Synth. Catal.
[33] SHELXTL V6.12, Bruker AXS, Inc., Analytical X-ray
Systems, 5465 East Cheryl Parkway, Madison WI
53711–55373.
[34] G. Liu, D. A. Cogan, J. A. Ellman, J. Am. Chem. Soc.
1997, 119, 9913.
2008, 350, 255.
[17] M. L. Kantam, T. Ramani, L. Chakrapani, B. M. Chou-
dary, J. Mol. Catal. A: Chem. 2007, 274, 11.
[18] R. I. Kureshy, I. Ahmad, K. Pathak, N.-U. H. Khan,
S. H. R. Abdi, J. K. Prathap, R. V. Jasra, Chirality 2007,
19, 352.
[35] H. Sellner, J. K. Karjalainen, D. Seebach, Chem. Eur. J.
2001, 7, 2873.
[19] J. Hartung, S. Drees, M. Greb, P. Schmidt, I. Svoboda,
H. Fuess, A. Murso, D. Stalke, Eur. J. Org. Chem. 2003,
2388.
[36] L. Canali, E. Cowan, H. Deleuze, C. L. Gibson, D. C.
Sherrington, J. Chem. Soc. Perkin Trans. 1 2000, 2055.
[37] M. Holbach, X. Zheng, C. Burd, C. W. Jones, M. Weck,
J. Org. Chem. 2006, 71, 2903.
[20] M. R. Maurya, U. Kumar, I. Correia, P. Adao, J. C.
Pessoa, Eur. J. Inorg.Chem. 2008, 577.
[38] D. Zhao, Q. Huo, J. Feng, B. F. Chmelka, G. D. Stucky,
[21] R. J. H. Clark, The Chemistry of Titanium and Vanadi-
um, Elsevier Publishing Company, Amsterdam, 1968.
[22] R. A. Shiels, C. W. Jones, J. Mol. Catal. A: Chem. 2007,
261, 160.
J. Am. Chem. Soc. 1998, 120, 6024.
[39] N. Bellec, F. Lerouge, O. Jeannin, G. Cerveau, R. J. P.
Corriu, D. Lorcy, J. Organomet. Chem. 2006, 691, 5774.
[40] L. Chen, Y. Cui, G. Qian, M. Wang, Dyes and Pigments
2007, 73, 338.
[23] J. C. Hicks, C. W. Jones, Langmuir 2006, 22, 2676.
2834
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2823 – 2834