Catalytically active, recyclable polymeric titanocene disks: A batch-flow reactor

Article abstract of DOI:10.1016/j.tetlet.2005.10.075

Dichlorotitanocene bound within porous polystyrene disks catalyzes the coupling of vinylmagnesium chloride and chlorosilanes to form 1,4-bis(silyl)-2-butenes. A simple batch-flow reactor permits catalyst reuse by repeated addition of fresh reagents and de

Full text of DOI:10.1016/j.tetlet.2005.10.075

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 8869–8871
Catalytically active, recyclable polymeric titanocene disks:
a batch-flow reactor
Patrick E. Berget and Neil E. Schore*
Department of Chemistry, University of California, Davis, One Shields Avenue, Davis CA 95616, USA
Received 9 September 2005; revised 12 October 2005; accepted 18 October 2005
Available online 2 November 2005
Abstract—Dichlorotitanocene bound within porous polystyrene disks catalyzes the coupling of vinylmagnesium chloride and chloro-
silanes to form 1,4-bis(silyl)-2-butenes. A simple batch-flow reactor permits catalyst reuse by repeated addition of fresh reagents and
decantation of products.
Ó 2005 Elsevier Ltd. All rights reserved.
The strategy of affixing a soluble catalyst to an insoluble
polymer support has considerable potential, including
at the start of this work, we applied our earlier experi-
ence with zirconocene systems to the synthesis of
bis(p-vinylbenzylcyclopentadienyl)dichlorotitanocene
1
the very attractive possibility of use in a continuous flow
2
5
manner. Flow-through processes utilizing monolithic
(1) (Scheme 1).
materials such as polymer rods and disks have been
3
described. In addition, flow reactors have been developed
As with zirconocenes, the hydrolytic sensitivity of tita-
nocenes precludes the use of suspension-polymerization
methods. We thus prepared Sherrington-style titano-
for microwave-promoted, photochemical, enzyme-cata-
lyzed, and metal-catalyzed syntheses.
4
3
d
cene-crosslinked polymer rods using the methodolo-
gies we first used for the synthesis of polymer-bound
zirconocene. Titanocene 1 was dissolved into a solution
of styrene and 1,2-dichlorobenzene using gentle warm-
ing. The mixture was then transferred into a silanized
Pyrex test tube containing 2 wt % benzoyl peroxide—
initial polymerization attempts had given materials
that adhered to the tube interiors, thus necessitating
pretreatment with dichlorodimethylsilane. The capped
tube was placed in an 85 °C oil bath; argon was bubbled
through the solution for 1h to ensure good initial
We have developed a procedure to incorporate dichloro-
titanocene into a polymer matrix, specifically at the
crosslink position, and have found it to be active in a
catalytic batch-flow-through carbon–carbon bond-
forming reaction system.
Positioning transition metals at crosslinks provides the
opportunity for robust metal binding through chelation
by multiple polymer-linked ligand moieties. Although
no titanocene-derived crosslinking reagents were known
Cl
i, ii
Cl
iii, iv
Ti
Cl
1
4
Scheme 1. Reagents and conditions: (i) NaCp, NaI, THF; (ii) aq HCl; (iii) NaH, THF, 0 °C; (iv) 0.5 equiv TiCl , THF.
*
0
Corresponding author. Tel.: +1530 752 6263; fax: +1530 754 76 10 ; e-mail: neschore@ucdavis.edu
040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.10.075

8870
P.E.Berget, N.E.Schore / Tetrahedron Letters 46 (2005) 8869–8871
mixing. For polymerization of a mixture containing
.5 mol % 1, 2–3 days was required for the process to
go to completion (as judged to be the case when the
polymer rod had completely pulled away from the test
tube walls). For a 5.0 mol % crosslinked polymer 7–
version of the process has negligible impact (<5%),
was going to be a problem (Eq. 2). Although formation
of the desired 1,4-disilylbutene was observed, the re-
duced overall rate of this polymer-supported catalyzed
process allowed vinylsilane formation to compete.
2
8
days was necessary. Upon complete polymerization
the test tubes were carefully broken and the rods
removed intact. The deep red rods, which have a rub-
bery consistency, were then sliced with a razor into
bright red disks 1–3 mm thick. Since the sole crosslink-
ing monomer is titanocene 1, the degree of crosslinking
is the same as the titanium loading. To ensure that the
disks were free from unreacted monomer they were sub-
jected to Soxhlet extraction with hot THF. Evaporation
of the extracts under reduced pressure showed only 1,2-
dichlorobenzene and trace amounts of styrene by NMR.
There was no evidence of any soluble titanocene-derived
species.
After extensive experimentation we found that a proce-
dure involving addition of excess Grignard in portions
provided the best results.
polymer-bound 1,
o
THF, 0 C, 6 hr,
MgBr
PhMe SiCl
SiPhMe2
SiPhMe2
2
ð2Þ
Me PhSi
2
+
We designed a simple flow reactor from a dual chamber
Schlenk recrystallizing tube and charged one chamber
with several 2.5% crosslinked disks which were then
swollen with THF and cooled to 0 °C. Vinylmagnesium
bromide (8 equiv relative to polymer-bound Ti) was
added, followed after 10 min by 10 equiv of chloro-
dimethylphenylsilane. After 2 and 4 h two additional
Watabe et al. have reported that multicomponent cou-
pling of two molecules of vinylmagnesium bromide and
two of various chlorosilanes in the presence of catalytic
Cp TiCl gives 1,4-disilyl-2-butenes, essentially bis(allyl)-
silanes (2, Eq. 1). This rapid, high-yield reaction
appeared to be an interesting and attractive one to study
in order to evaluate recyclable disk methodology.
2
2
6
8
6
equiv aliquots of Grignard were added. After the total
h of reaction time the supernatant was decanted
through the frit into the second chamber, from which
it was removed for analysis maintaining an argon atmo-
sphere. Fresh reagents were added to the side containing
the disks and the sequence repeated through several
runs, giving the results in Table 1 for two separate but
similarly prepared batches of disks. In each run carried
out as described above, conversion of chlorosilane to the
mixture of silylated products was 100%. Through three
runs (i.e., an initial run plus two recyclings) yields of 2
averaged better than 70%, the remainder being vinyl-
silane. In addition, the 87/13 (or better) E/Z ratio was
an improvement over the solution-phase 74/26 result.
Despite these encouraging results, catalytic reactivity
dropped precipitously after the third run, and vinyl-
silane formation predominated.
cat Cp TiCl ,
THF, 0 C, 10 min
2
2
o
MgBr
2
+ 2 R SiCl
3
ð1Þ
R Si
3
SiR3
2
6
The mechanism proposed for this process is illustrated
in Scheme 2. Keys to the success of the process include
efficient conversion of divinyltitanocene into a titana-
cyclopentene, which proceeds via the butadiene p-com-
plex, and a fortuitous selectivity for chlorosilanes to
react more rapidly with allylic than with vinylic Grig-
nard reagents.
Initial experiments revealed that the competing uncatal-
yzed conversion of vinylmagnesium bromide and chlo-
rosilane into a vinylsilane, which in the solution phase
We also carried out a set of runs using disks loaded with
5.0% of the crosslinking monomer to determine whether
the higher loading would improve catalytic reactivity
Cp Ti(h -C H )
Cp Ti
2
4
4
6
2
2
CH =CHLi
2
CH =CHMgBr
Cp TiCl
Cp Ti(CH=CH )
2 2
2
2
2
2
BrMg
SiR3
Cp Ti
2
R SiCl
3
CH =CHMgBr
MgBr
2
Cp Ti
2
R SiCl
3
R Si
3
SiR3
SiR3
6
Scheme 2. Mechanism proposed by Watabe et al.

P.E.Berget, N.E.Schore / Tetrahedron Letters 46 (2005) 8869–8871
8871
Table 1. Preparation of 1,4-bis(dimethylphenylsilyl)-2-butene, 2, using
18 h for I , the initially dark disks lightened to an amber
2
2.5% crosslinked disks
color rather than the expected red for a titanocene di-
halide, and they were found to be completely inactive.
These experiments suggest that neither divinyltitanocene
a
Run
%Yield
E/Z
Batch 1
nor any other simple Cp TiR is present at the end of a
1
2
3
4
5
82
2
1
88/1
2
2
catalytic run. Another complicating issue is the forma-
tion over the course of the catalysis of several equiva-
lents of inorganic salts that may impede reactivity.
Procedures designed to extract and/or coordinate and
remove these salts from the disks have been carried
out without reactivity being restored. Experiments are
currently underway to ascertain directly what is the nat-
ure of the Ti-containing species and on that basis to
design a practical regeneration procedure. Our results
will be reported in due course.
72
79
29
13
88/12
87/13
87/13
90/10
Batch 2
1
2
3
4
80
89/1
74
62
22
91/9
89/11
88/12
a
Remainder dimethylphenylvinylsilane.
despite the increased crosslinking. Using the same
procedure as that described above, we found that 5%
of crosslinked disks could be reused through four or five
runs, but gave rather lower yields (50–60%) of 2. Again,
catalytic reactivity dropped precipitously thereafter.
Acknowledgements
We thank the National Science Foundation, for finan-
cial support (Grant CHE-0313888).
References and notes
To further evaluate these results, we compared the total
number of moles of product produced through all of the
runs with the number of moles of polymer-bound catal-
yst, obtaining overall turnover numbers. The two
batches of 2.5% crosslinked disks gave aggregate
turnover numbers of 39.2 and 40.5, through five and
four runs, respectively. For comparison, using 5.0%
crosslinked disks a total turnover number of 17.9
was obtained. The analogous solution-phase process is
1
. Some reviews on polymer supported catalysis: (a) Cernia,
E. M.; Graziani, M. J.Appl.Polym.Sci. 1974, 18, 2725; (b)
Chauvin, Y.; Commereuc, D.; Dawans, F. Prog.Polym.
Sci. 1977, 5, 95; (c) Akelah, A.; Sherrington, D. C. Chem.
Rev. 1981, 81, 557; (d) Lindner, E.; Schneller, T.; Auer, F.;
Mayer, H. A. Angew.Chem,. Int.Ed.
1999, 38, 2155; (e)
Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.;
Leach, A. G.; Longbottom, D. A.; Nesi, M.; Scott, J. S.;
Storer, R. I.; Taylor, S. J. J.Chem.Soc.Perkin Trans.1
6
reported to give a turnover number of 53.6.
2
000, 3815; (f) McNamara, C. A.; Dixon, M. J.; Bradley,
M. Chem.Rev. 2002, 102, 3275; (g) Fan, Q.-H.; Li, Y.-M.;
Chan, A. S. C. Chem.Rev. 2002, 102, 3385.
The detrimental effect of higher crosslinking implies
that, despite the higher loading, the rate of catalysis
using the 5% crosslinked resin is reduced compared with
the 2.5%, allowing the solution-phase vinylsilane-pro-
ducing side reaction to compete more effectively. Our
observations do not directly address the reason behind
the reduced effectiveness of the more highly-crosslinked
resin, but both reduced rates of encounter between sub-
strates and catalyst as well as changes in chemical envi-
ronment of the resin interior with increased crosslinking
may contribute. It is also not clear what causes deactiva-
tion of the catalyst after several cycles of use. To address
the latter we set out to investigate whether or not the
disks could be regenerated.
2
3
. Jas, G.; Kirschning, A. Chem.Eur.J. 2003, 9, 5708.
. (a) Svec, F.; Fr e´ chet, J. M. Anal.Chem. 1992, 64, 820; (b)
Viklund, C.; Svec, F.; Fr e´ chet, J. M.; Irgum, K. Chem.
Mater. 1996, 8, 744; (c) Peters, E. C.; Svec, F.; Fr e´ chet,
J. M. Adv.Mater. 1999, 11, 1169; (d) Hird, N.; Hughes, I.;
Hunter, D.; Morrison, M. G. H. T.; Sherrington, D. C.;
Stevenson, L. Tetrahedron 1999, 55, 9575; (e) Tripp, J. A.;
Stein, J. A.; Svec, F.; Fr e´ chet, J. M. Org.Lett. 2000, 2, 195;
(f) Tripp, J. A.; Svec, F.; Fr e´ chet, J. M. J.Comb.Chem.
2
001, 3, 216; (g) Monolithic Materials: Preparation, Prop-
erties, and Applications; J.Chromatography Library ; Svec,
F., Tennikova, T. B., Deyl, Z., Eds.; Elsevier: Amsterdam,
Boston, 2003; Vol. 67.
4
. (a) Comer, E.; Organ, M. G. J.Am.Chem.Soc. 2005, 127,
8
160; (b) Bagley, M. C.; Jenkins, R. L.; Lubinu, M. C.;
7
Based upon work of F u¨ rstner and Hupperts as well as
Mason, C.; Wood, R. J.Org.Chem. 2005, 70, 7003; (c)
Hook, B. D. A.; Dohle, W.; Hirst, P. R.; Pickworth, M.;
Berry, M. B.; Booker-Milburn, K. I. J.Org.Chem. 2005,
70, 7558; (d) Engle, S.; Vyazmensky, M.; Berkovich, D.;
Barak, Z.; Merchuk, J.; Chipman, D. M. Biotech.Bioeng.
8
our own in a related system, we first examined regener-
ation methods in the solution phase. The proposed
mechanism (Scheme 2) suggested that divinyltitanocene
was the likely Ti-containing species at the end of a cycle.
We therefore tested the efficacy of several reagents at
converting divinyltitanocene back to titanocene di-
halide. We found that I , SOCl , and ClCOCOCl in
THF all gave quantitative conversion to Cp TiX
X = Cl or I) in under 2 h at 25 °C. In attempting to
adopt these methods to recycle our titanocene-function-
alized disks, we exposed samples of the latter to THF
solutions of each of these reagents at 25 °C. After a per-
iod of time that ranged from 1.5 h for ClCOCOCl to
2
005, 89, 733; (e) Greenway, G. M.; Haswell, S. J.; Morgan,
D. O.; Skelton, V.; Styring, P. Sensors Actuators 2000, 63B,
53; (f) Solodenko, W.; Wen, H.; Leue, S.; Stuhlmann, F.;
1
2
2
Sourkouni-Argirusi, G.; Jas, G.; Sch o¨ nfeld, H.; Kunz, U.;
Kirschning, A. Eur.J.Org.Chem. 2004, 3601.
. Hok, S.; Vassilian, J.; Schore, N. E. Org.Lett. 2002, 4, 2365.
. Watabe, H.; Terao, J.; Kambe, N. Org.Lett. 2001, 3, 1733.
. F u¨ rstner, A.; Hupperts, A. J.Am.Chem.Soc.
2
2
(
5
6
7
1995, 117,
4468.
8. Berget, P.; Schore, N. E. Unpublished results.