"Catalyst-on-a-tape" - Teflon: A new delivery and recovery method for homogeneous fluorous catalysts

Article abstract of DOI:10.1002/anie.200500237

(Chemical Equation Presented) Nothing sticks to Teflon? Going one better than "soap on a rope", a recyclable fluorous rhodium hydrosilylation catalyst is described that is released from Teflon tape into dibutyl ether upon warming, and readsorbed upon cool

Full text of DOI:10.1002/anie.200500237

Angewandte
Chemie
Catalyst Recycling
“Catalyst-on-a-Tape”—Teflon: A New Delivery
and Recovery Method for Homogeneous Fluorous
Catalysts**
Long V. Dinh and J. A. Gladysz*
There are intense efforts to develop new or improved
strategies for recovering homogeneous molecular catalysts.[1]
Many of these approaches are bi- or multiphasic in nature,
and one particularly innovative method involves catalysts that
are derivatized with (CH2)m(CF2)nÀ1CF3 ((CH2)mRfn) moieties
or fluorous “ponytails”.[2,3] The initial applications of this
now-familiar technique utilized perfluoroalkane solvents for
recovery. However, the relatively high costs of such fluorous
solvents, and other factors, pose limitations for industrial
applications.
Second-generation methods eliminated the fluorous sol-
vent by exploiting the temperature-dependent solubilities of
fluorous catalysts in common organic solvents.[4,5] Appropri-
ately designed fluorous molecules are soluble only at elevated
temperatures, and essentially insoluble at low temperatures.
Such a thermomorphic character allows homogeneous catal-
ysis at the high-temperature limit, and catalyst recovery by a
simple liquid/solid-phase separation at the low-temperature
limit. The presence of an insoluble fluorous support can be
beneficial, for example, when small quantities are involved.
Initial efforts have involved Teflon shavings[4] and fluorous
silica gel.[6–8]
We sought to expand the range of catalysts and fluorous
supports that could be applied in such methods, and report
herein a surprisingly effective procedure involving common
commercial Teflon tape. This approach provides not only a
convenient means of catalyst recovery, but also of catalyst
delivery. From an engineering standpoint, tapes offer a
variety of unique attributes, and it is reasonable to extrapolate
that meshes and/or reactor parts, such as liners or stirring
assemblies, could be fabricated with similar properties.
As a test reaction, we selected the ketone hydrosilylation
shown in Figure 1, which in earlier work was catalyzed by the
fluorous rhodium complexes 1-Rf6 and 1-Rf8 in a liquid/liquid
biphase system comprising an organic and a fluorous sol-
vent.[9] The red-orange compounds 1-Rf6 and 1-Rf8 have very
little or no solubility in organic solvents at room temper-
ature.[10] However, their solubilities increase markedly with
[*] Dr. L. V. Dinh, Prof. Dr. J. A. Gladysz
Institut fꢀr Organische Chemie
Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Henkestrasse 42, 91054 Erlangen (Germany)
Fax: (+49)9131-85-26865
E-mail: gladysz@organik.uni-erlangen.de
[**] We thank the Bundesministerium fꢀr Bildung und Forschung
(BMBF) for support, Prof. Dr. A. Behr and Dr. R. Roll (University of
Dortmund) for rhodium measurements, and Prof. Dr. D. Curran
(University of Pittsburgh) for helpful discussions.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 4095 –4097
DOI: 10.1002/anie.200500237
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4095

Communications
Figure 1. Recycling of a thermomorphic fluorous rhodium hydrosilylation cata-
lyst by liquid/solid phase separation.
temperature. Several features make this catalyst system a
particularly challenging test for recovery by precipitation. For
example, a variety of rest states are possible (e.g., various
{Rh(H)(SiR3)} or {Rh(OR’)(SiR3)} species), each with unique
solubility properties.[11] Also, the first cycle exhibits an
induction period,[9] indicating some fundamental alteration
of the catalyst precursor.
Figure 2. Recycling of a thermomorphic fluorous rhodium hydrosilyla-
tion catalyst using Teflon tape.
The solvent dibutyl ether was selected due to its extended
liquid range (b.p. 1428C). A solution of cyclohexanone (2;
0.53m), PhMe2SiH (1.2 equiv), and a GC standard was treated
with 1-Rf6 (1 mol%), and warmed to 658C, achieving
homogeneous conditions.[12] After 8 h, the mixture was
cooled to room temperature. GC analysis showed a 98%
yield of the silyl ether 3. A brown-red catalyst residue slowly
precipitated over several hours, and no color remained in the
product-containing supernatant solution. To speed up this
process, samples were transferred to a freezer (À308C); the
supernatant solutions were then removed by syringe. The
residue was washed twice with cold dibutyl ether and charged
with fresh reactants. The cycle was repeated three additional
times, each giving a 98% yield of 3.
Supports were sought that would facilitate the recovery of
lower catalyst loadings. Thus, a similar sequence was con-
ducted with 1-Rf6 (0.15 mol%), but in the presence of Teflon
tape (thickness/width 0.0075/12 mm). For this system, homo-
geneous conditions could be achieved at 558C. Photographs
of a typical sequence are shown in Figure 2. The white tape
became lightly colored during the reaction, and orange-red
when the sample was cooled. Enthalpic interactions between
perfluoroalkyl segments are very small. Hence, it is note-
worthy that the catalyst rest-state phase separates onto the
tape, as opposed to giving a second solid phase. Interestingly,
the Teflon stirring bar remained white.
GC analysis as above indicated a 97% yield of 3. In an
identical reaction the concentrations of 2 and 3 were assayed
at 15 min intervals over four cycles. Data are summarized in
Figure 3. As in earlier work,[9] the first cycle exhibited an
induction period. Retention of activity was excellent in the
second and third cycles, but there was substantial loss in the
fourth. In view of the similar rates of the first three cycles, this
loss of activity can be attributed to catalyst deactivation, as
Figure 3. Reaction profile for four cycles under the conditions given in
~
^
*
Figure 2: & Cycle 1 Cycle 2 Cycle 3 Cycle 4. Top: decrease (%) in
2 with time, bottom: increase (%) in 3 with time. TON=turnover
number.
opposed to leaching or some intrinsic problem with the
recycling method.
Leaching was probed in two ways. First, the dibutyl ether
supernatants from the first three cycles of the experiments
shown in Figure 3 were combined, a fluorine-containing
standard added, and a 19F NMR spectrum recorded. The
À
chemical shifts of the (CF2)5CF3 signals are independent of
the remote substituent, and integration indicated a “total
pony-tail leaching” corresponding to 11.4% of the phosphine
4096 ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 4095 –4097

Angewandte
Chemie
ligands of 1-Rf6. Since the active and resting states of the
catalyst probably involve only two phosphine ligands, some
extraction is not surprising. Second, the supernatant from the
first cycle was analyzed for rhodium by AAS-ICP (atom
absorption spectrometry with inductively coupled plasma).
An average of three determinations indicated leaching
corresponding to 0.57% of the original charge. The second
cycle gave a value of 5.3%.
One procedural refinement would be to precoat the
catalyst on the Teflon tape. When coating is uniform, this
would allow low loadings to be delivered by length as opposed
to mass measurements. Accordingly, two 50 ꢀ 12 ꢀ 0.0075 mm
strips of tape were added to a solution of 1-Rf6 (0.013 g,
0.0039 mmol) in CF3C6F11 (1.0 mL). The solvent was removed
under an inert gas stream to give a yellowish catalyst-coated
tape. This tape was applied in a three-cycle sequence and gave
yield results similar to those in Figure 3. Photographs are
supplied in the Supporting Information.
When catalysts are recycled as solid residues, it is
important to exclude impurities that may “piggyback”, such
as metal particles, as the active species. This situation was
probed in two ways. First, the tape was removed after a first
cycle, rinsed with cold dibutyl ether, and transferred to a new
vessel. A second charge of 2 and dibutyl ether was added, but
not the PhMe2SiH. The sample was warmed to 558C, the now-
off-white tape “fished out”, and PhMe2SiH added. The rate
profile was similar to the first cycle (ca. 20% slower at higher
conversions), consistent with predominant homogeneous
catalysis by desorbed fluorous species. Second, the second
cycle of a sequence was conducted in the presence of
elemental mercury (Hg:Rh 500:1), which inhibits catalysis
by metal particles.[13] However, the rate profile was the same
as a sequence in the absence of mercury (see Supporting
Information).
Hydrosilylations of 2-octanone, acetophenone, and ben-
zophenone to the corresponding ethers have been conducted
under identical conditions. As summarized in the Supporting
Information, similar results are obtained, albeit with up to
20% activity loss in the third cycle. The basis for this
somewhat diminished performance is under study and further
details will be provided in a full paper.
introduction of low loadings without recourse to a balance or
the tedious preparation of standard solutions. Finally, the
slightly different solubility profiles of 1-Rf6 and 1-Rf8[12] nicely
illustrate how physical properties of fluorous compounds can
be engineered to optimally fit a particular application or
reaction condition.
Received: January 21, 2005
Published online: May 18, 2005
Keywords: fluorine · homogeneous catalysis · hydrosilylation ·
.
rhodium · Teflon
[1] Chem. Rev. 2002, 102, Special Issue (pp. 3215–3892) on “Recov-
erable Catalysts and Reagents” (J. A. Gladysz, issue Editor).
[2] I. T. Horvꢁth, Acc. Chem. Res. 1998, 31, 641.
[3] Handbook of Fluorous Chemistry (Eds.: J. A. Gladysz, D. P.
Curran, I. T. Horvꢁth), Wiley-VCH, Weinheim, 2004.
[4] a) M. Wende, R. Meier, J. A. Gladysz, J. Am. Chem. Soc. 2001,
123, 11490; b) M. Wende, J. A. Gladysz, J. Am. Chem. Soc. 2003,
125, 5861.
[5] a) K. Ishihara, S. Kondo, H. Yamamoto, Synlett 2001, 1371; b) K.
Ishihara, A. Hasegawa, H. Yamamoto, Synlett 2002, 1299; c) K.
Mikami, Y. Mikami, H. Matsuzawa, Y. Matsumoto, J. Nishikido,
F. Yamamoto, H. Nakajima, Tetrahedron 2002, 58, 4015; d) J.
Otera, Acc. Chem. Res. 2004, 37, 288, and references therein;
e) G. Maayan, R. H. Fish, R. Neumann, Org. Lett. 2003, 5, 3547.
[6] a) C. C. Tzschucke, C. Markert, H. Glatz, W. Bannwarth, Angew.
Chem. 2002, 114, 4678; Angew. Chem. Int. Ed. 2002, 41, 4500;
b) A. Biffis, M. Braga, M. Basato, Adv. Synth. Catal. 2004, 346,
451, and references therein.
[7] For related approaches involving fluorous catalysts, see a) O.
Yamazaki, X. Hao, A. Yoshida, J. Nishikido, Tetrahedron Lett.
2003, 44, 8791; b) C. D. Ablan, J. P. Hallett, K. N. West, R. S.
Jones, C. A. Eckert, C. L. Liotta, P. G. Jessop, Chem. Commun.
2003, 2972.
[8] For another recent approach to catalyst recovery by precipita-
tion, see V. K. Dioumaev, R. M. Bullock, Nature 2003, 424, 530.
[9] a) L. V. Dinh, J. A. Gladysz, Tetrahedron Lett. 1999, 40, 8995;
b) L. V. Dinh, J. A. Gladysz, New J. Chem. 2005, 29, 173.
[10] J. J. J. Juliette, D. Rutherford, I. T. Horvꢁth, J. A. Gladysz, J. Am.
Chem. Soc. 1999, 121, 2696.
[11] For further discussion of this point, see Chapter 4 of refer-
ence [3] and J. A. Gladysz, Pure Appl. Chem. 2001, 73, 1319.
[12] When 1-Rf8 is employed in the reaction shown in Figure 1,
identical yields are obtained. However, some catalyst (rest state)
precipitates during the reaction, presumably as a result of the
decreasing polarity of the solution as the product ether is
formed. For a related phenomenon, see ref. [8]. Also, 1-Rf8 does
not completely dissolve in dibutyl ether at 558C at the
concentrations and loadings (1.00–0.15 mol%) used.
In summary, the above data constitute, to our knowledge,
the first catalyst recycling method involving common Teflon
tape. Attractive interactions between the fluorous domains of
the catalyst and the tape are clearly operative. However,
other forms of Teflon—such as used to coat stir bars—are
ineffective, and additional work is needed to identify the
optimum morphologies.[14] There are also tantalizing possi-
bilities with other fluoropolymers, as well as for fabricating
reactors or components thereof that could efficiently release
and recapture thermomorphic fluorous catalysts as a function
of temperature. Furthermore, precoated “catalyst-on-a-tape”
systems offer many advantages for delivery alone, such as the
[13] J. A. Widegren, R. G. Finke, J. Mol. Catal. A 2003, 198, 317.
[14] A reviewer has stressed that absorption isotherms and surface
area measurements can provide much insight regarding the
mode of catalyst attachment. Although these experiments are
beyond the scope of this preliminary communication, such data
are planned for a later full paper.
Angew. Chem. Int. Ed. 2005, 44, 4095 –4097
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4097