Cryptocatalytic 1,2-alkene migration in a ?-alkyl palladium diene complex

Full text of DOI:10.1002/anie.200901468

Communications
DOI: 10.1002/anie.200901468
Reaction Mechanism
Cryptocatalytic 1,2-Alkene Migration in a s-Alkyl Palladium Diene
Complex**
Louise A. Evans, Natalie Fey, Guy C. Lloyd-Jones,* M. Paz Muꢀoz, and Paul A. Slatford
Alkene migration is frequently encountered in transition-
metal-catalyzed reactions as part of the productive cycle[1] or
as a competing process.[2] The observation of alkene migration
in isolated n-metallo-1-alkene complexes[3] is rare,[3a,4] and
provides a valuable opportunity for mechanistic study, which
has the potential to yield information for the control of
selectivity in related catalytic reactions.
Scheme 2. Generation and b-hydride elimination of complexes 4 and 7,
and their equilibration mediated by TfOH. Empirical first-order rate
constants (sꢀ1) for 3!4: (1.0ꢁ0.4)ꢀ10ꢀ2; 4!5: (3.4ꢁ0.4)ꢀ10ꢀ3
;
6!7: (1.5ꢁ0.1)ꢀ10ꢀ5; 7!8: (6.0ꢁ0.2)ꢀ10ꢀ6. Tf=trifluoromethane-
sulfonyl.
Scheme 1. Thermal isomerization of 1 to 2 (toluene, 1108C), with a
proposed (k1-dppp)PtIVH(h3-allyl) intermediate.[4a] dppp=propane-l,3-
diylbis(diphenylphosphane).
nation, and 4 is no exception, thus generating triene 5 in the
presence of a proton sponge (1,8-bis(dimethylamino)naph-
thalene) at ꢀ608C.[6a] The isomeric h5-alkyldiene complex 7
can be generated analogously from the 1,5-diene complex 6[7]
at 508C[8] and undergoes elimination of a b hydrogen to give
triene 8 at 218C.
A recent example involves alkene migration in cis-Pt(k2-
dppp)(alkenyl)2 complexes (1!2; Scheme 1),[4a] for which a
unimolecular mechanism involving a PtIV(H) complex was
proposed on the basis of 1) reactions being faster in dilute
solution,[5] 2) there being no inhibition by Hg metal, and 3) a
rate retardation upon addition of excess dppp.[4a] Herein, we
report our investigation of an analogous migration in a
palladium complex (4!7; Scheme 2) by using NMR spec-
troscopy, mass spectrometry, and isotopic labeling (2H, 13C,
108Pd). The outcome eliminates the appealingly simple intra-
ꢀ
molecular allylic C H insertion process analogous to
Scheme 1, and reinforces the caveat that what appears on
first inspection to be a relatively simple intramolecular
process, rarely is.
The palladium complexes 4 and 7 were readily prepared.
Cationic diallyl malonate complex 3 (Scheme 2) undergoes
intramolecular allyl palladation at ꢀ208C to give the neutral
h5-alkyldiene complex 4.[6] s-Alkyl palladium complexes
bearing b-hydrogen atoms are usually susceptible to elimi-
Figure 1. X-ray structures of 4 and 7·H2O.[9] All protons apart from
those on C8–C10 (highlighted in purple) are omitted for clarity.
Complex 7 crystallized with a water molecule bound to Pd in place of
the triflate counterion (not shown), which has a weak intermolecular
contact (Pd–O=248 pm). In 4, only the palladium-bound oxygen atom
of the triflate is shown.
In the absence of base, 4 and 7 are remarkably stable, thus
allowing detailed structural analysis (Figure 1).[9] However,
over a period of months in CDCl3 solution, 4 underwent 1,2-
alkene migration at C9 and C10 as well as diastereofacial
inversion at C1 and C2 to generate the thermodynamically
more stable complex 7.
[*] L. A. Evans, Dr. N. Fey, Prof. Dr. G. C. Lloyd-Jones, Dr. M. P. Muꢁoz,
Dr. P. A. Slatford
School of Chemistry, University of Bristol
Cantock’s Close, Bristol, BS81TS (UK)
Fax: (+44)117-929-8611
E-mail: guy.lloyd-jones@bris.ac.uk
Acid-catalyzed alkene migration in iridium complexes has
been reported by Shaw and co-workers.[4b] Accordingly, the
addition of one equivalent of TfOH to 4 in CDCl3 (28.6 mm)
resulted in reversible and diastereoselective protonation of
the malonate carbonyl group and induced clean 1,2-alkene
migration to generate 7.[10] Although protonation of the
[**] Prof. Dr. J. N. Harvey (Bristol) is thanked for helpful discussions
regarding DFT calculations. We thank the Spanish Ministry of
Science and AstraZeneca for support. G.C.L.J. holds a Royal Society
Wolfson Research Merit Award.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200901468.
6262
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6262 –6265

Angewandte
Chemie
carbonyl group was immediate, the migration occurred after a
variable and sometimes extreme induction period (up to 8 h)
during which the time-average H NMR signal of (TfOH)n
labeled C6. This outcome also rules out mechanism E, as a
1,3-suprafacial shift in 4b can only involve 1H. Meanwhile, the
13C-labeled C10 in 7 was partially deuterated, suggestive of
intermolecularity between 4b and 4b’ in the alkene migration
step.
1
migrated from d = (12.75 ꢁ 0.15) to (13.5 ꢁ 0.6) ppm. Upon
switching to C6D6 as the solvent, so that dimeric acid (TfOH)2
more rapidly monomerizes,[11] the kinetics of alkene migra-
tion (4!7) became more reproducible but remained sigmoi-
dal in character (Figure 2a). The analysis of the maximal rate
(d[7]/dt)max as a function of TfOH concentration indicates an
approximately first-order dependency (see filled circles in
Figure 2b).
Experiment 4 confirmed that there is intermolecular
transfer of a hydrogen atom, but not palladium, as the
coreaction of 4c (ꢂ 95% 108Pd) with 4d gave two distinct
isotope clusters [2Hꢃ3,108Pd]-7 and [2Hꢂ3]-7 upon ESI-MS
analysis. Simulation of the spectrum indicated that multiple
1H/2H transfers had occurred, for example, for [2Hn,108Pd]-7,
n = 1 (10%), n = 2 (5%), and n = 3 (2.5%). Mechanism D
would allow hydride exchange by dimerization of the Pd(H)
intermediate 9. However, for this process to facilitate multiple
transfers, generation of 9 would need to be reversible and 1H/
2H exchange would only occur at C8.
Experiment 5 confirmed 1H/2H exchange in the substrate,
but ruled out mechanism D because after 50% conversion
into 7 the remaining 4d had become partially deuterated at
C10 but had not lost any deuterium atoms at C8 or C6. The
two complexes are isobaric by ESI-MS, and key to the
analysis of 4 without interference by 7 was their difference in
reactivity towards the proton sponge (Scheme 2), thus allow-
ing the generation of triene [2Hn]-5, as a proxy for 4d, without
the generation of 8.
Experiments 1 to 5 excluded all reasonable mechanisms
for hydrogen migration which directly involved the palladium
centre in 4 or the proton from the TfOH (mechanisms A to
E).[15] In further experiments, it was found that analogues of 4
that lacked the conformationally unrestricted malonate ester
moieties did not undergo migration at C9 and C10[16] (even
Figure 2. a) Temporal evolution of 7 during isomerization of 4
(29 mm) with TfOH (1 equiv) in C6D6, see the Supporting Information
for full details. b) Relationship between maximal rate (y axis) and initial
concentration (x axis) of TfOH (filled circles) and 4 (open circles).
We considered a number of mechanisms (Scheme 3) to
explain the alkene migration at C9 and C10 (4!7), including
[4a,c,d,f]
acid-mediated[4b,12] (A) and allylic C H insertion
(D)
ꢀ
mechanisms. Other possibilities include mechanisms involv-
ing the generation of a hydride[4g,h] through b-hydride
elimination[6a]/re-addition sequences emanating from C4 (B)
or C6 (C), as well as a palladium-assisted 1,3-suprafacial shift
of a hydrogen atom (E).[13]
By using a library of 2H, 13C, and 108Pd labeled[14] forms of
4, we conducted reactions with TfOH(D) in C6D6. We
employed a combination of 13C{1H} NMR spectroscopy
(1JCD and DdCH,D) and ESI-MS to quantify and locate H-
2
populations. The key experiments are summarized in Experi-
ments (1)–(5) in Scheme 4.
Experiment 1 ruled out any direct involvement of the
acid, (e.g. A) as migration induced by TfOD resulted in no
deuterium incorporation in either 4 or 7, even when eight
equivalents of TfOD was used.
Experiment 2 ruled out mechanism B (and any involve-
2
13
=
ment of alkene C1 C2), as the reaction of 4a gave [ Hn, C1]-7
Scheme 3. Six mechanisms (A to F) for alkene migration 4!7, with the key
hydrogen-migration source and destinations indicated (as H and H), see
text for full details. Dissociative[6a] diastereofacial inversion at C1 and C2 can
occur before, during, or after alkene migration at C9 and C10. E=CO2Me.
2
without any detectable incorporation of H at C1 or C4.
Experiment 3 ruled out mechanism C, as the reaction of
2
4b/4b’ gave [2Hn,13C1]-7 in which no H was detected at 13C-
Angew. Chem. Int. Ed. 2009, 48, 6262 –6265
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6263

Communications
In summary, the simplicity of mechanisms A to E make
them appealing as explanations for alkene migration in n-
metallo-1-alkene complexes.[4] However, for the case of 4!
7, isotopic labeling experiments rule out all of these
processes. In analogous migrations, the possibility of a
“cryptocatalytic” intermolecular mechanism (F, Scheme 3)
should thus be considered, whether it be a stoichiometric
process[4] or an integral part of a catalytic cycle.[1]
Received: March 17, 2009
Revised: June 3, 2009
Published online: July 17, 2009
Keywords: alkene migration · homogeneous catalysis ·
.
hydrides · isotopic labeling · palladium
[1] See for example: a) F. Ozawa, A. Kubo, T. Hayashi, J. Am.
Chem. Soc. 1991, 113, 1417 – 1419; b) O. Loiseleur, P. Meier,
A. Pfaltz, Angew. Chem. 1996, 108, 218 – 220; Angew. Chem.
Int. Ed. Engl. 1996, 35, 200 – 202; c) T. G. Kilroy, A. J.
Hennessy, D. J. Connolly, Y. M. Malone, A. Farrell, P. J.
Guiry, J. Mol. Catal. A 2003, 196, 65 – 81; for elegant labeling
studies of this system, see: d) B. M. M. Wheatley, B. A. Keay,
J. Org. Chem. 2007, 72, 7253 – 7259.
[2] See for example: a) S. J. Miller, H. E. Blackwell, R. H.
Grubbs, J. Am. Chem. Soc. 1996, 118, 9606 – 9614; b) A.
Fꢀrstner, O. R. Thiel, L. Akermann, J.-J. Schanz, S. P. Nolan, J.
Org. Chem. 2000, 65, 2204 – 2207; c) B. Schmidt, Eur. J. Org.
Chem. 2004, 1865 – 1880.
Scheme 4. 2H, 13C, and 108Pd labeling to probe the mechanism of 4!7.
Conditions and reagents: a) TfOD (1–8 equiv), C6D6; b) TfOH (1 equiv),
C6D6.
[3] a) A. Sivaramakrishna, H. S. Clayton, C. Kaschula, J. R. Moss,
Coord. Chem. Rev. 2007, 251, 1294 – 1308; b) E. Hager, A.
Sivaramakrishna, H. S. Clayton, M. M. Mogorosi, J. R. Moss,
Chem. Rev. 2008, 108, 1668 – 1688.
when coreacted with 4) and thus the carbonyl group is
essential for reactivity.
In earlier studies we showed that the “elimination” of
triene 5 from s-alkyl palladium complex 4 (Scheme 2)
proceeded indirectly: a Pd(H) species was generated in very
low equilibrium concentration with 4 by reversible dissocia-
tion at C1 and C2 as well as syn-b-hydride elimination at C4,
and it is this Pd(H) complex that is deprotonated.[6a] In the
absence of base, trace quantities of any Pd(H) species
irreversibly liberated from this equilibrium will be capable
of catalyzing 1,2-alkene migration in the bulk complex.[17]
[4] a) A. Sivaramakrishna, H. Su, J. R. Moss, Organometallics 2007,
26, 5786 – 5790; b) A. J. Deeming, B. L. Shaw, R. E. Stainbank, J.
Chem. Soc. A 1971, 374 – 376; c) B. T. Heaton, D. J. A. McCaf-
frey, J. Chem. Soc. Dalton 1979, 1078 – 1083; d) I.-H. Wang, G. R.
Dobson, J. Organomet. Chem. 1988, 356, 77 – 84; e) H. Matstj-
zaka, Y. Hiroe, M. Iwasaki, Y. Ishii, Y. Koyast, M. Hidai, Chem.
Lett. 1988, 377 – 380; f) D. V. Krupenya, S. I. Selivanov, S. P.
Tunik, M. Haukka, T. A. Pakkanen, Dalton Trans. 2004, 2541 –
2549; g) E. Royo, S. Acebrꢁn, M. E. G. Mosquera, P. Royo,
Organometallics 2007, 26, 3831 – 3839; h) G. Chahboun, C. E.
Petrisor, E. Gꢁmez-Bengoa, E. Royo, T. Cuenca, Eur. J. Inorg.
Chem. 2009, 1514 – 1520; i) A. Sivaramakrishna, B. C. E. Makhu-
bela, F. Zheng, H. Su, G. S. Smith, J. R. Moss, Polyhedron 2008,
27, 44 – 52; j) A. Sivaramakrishna, P. Mushonga, I. L. Rogers, F.
Zheng, R. J. Haines, E. Norlander, J. R. Moss, Polyhedron 2008,
27, 1911 – 1916; k) T. Mahamo, F. Zheng, A. Sivaramakrishna,
J. R. Moss, G. Smith, J. Organomet. Chem. 2008, 693, 103 – 108;
l) F. Zheng, A. Sivaramakrishna, J. R. Moss, Inorg. Chim. Acta
2008, 361, 2871 – 2878.
[5] The temporal fractional conversion should in fact be concen-
tration independent. There is also an induction period evident
with 1 (R = butenyl), see Figure 4 in reference [4a].
[6] a) G. C. Lloyd-Jones, P. A. Slatford, J. Am. Chem. Soc. 2004, 126,
2690 – 2691; b) The OTf group is uniquely effective for the
generation of 4, analogous reactions using BF4, PF6, SbF6, OTs,
or OAc failed. Ts = 4-toluenesulfonyl.
[7] L. A. Evans, N. Fey, J. N. Harvey, D. Hose, G. C. Lloyd-Jones, P.
Murray, A. G. Orpen, R. Osborne, G. J. Owen-Smith, M. Purdie,
J. Am. Chem. Soc. 2008, 130, 14471 – 14473.
[8] Complex 6 acts as an efficient reagent for the generation of 3
from the corresponding the 1,6-diene. DFT calculations on
Mechanism
F (Scheme 3) in which reaction of the
[(L)nPd(H)] complex 10 (L = alkene[6a,17e]) with 4 to generate
11 accounts for all of the isotopic labeling results: reversible
hydropalladation at C9 and C10 of 4 allows intermolecular
1H/2H exchange at C10 and 1,2-alkene migration, without any
cross-over of 108Pd in 5,9-dipalladium complex 11.[18] The
malonate carbonyl serves to precoordinate the [(L)nPd(H)]
catalyst 10, thus delivering it to C9 and C10 and stabilizing the
resulting s-alkyl/palladium intermediate; similar oxo-chelate
derivatives have been identified in cycloisomerization.[17g]
In the absence of added TfOH, complex 4 undergoes slow
alkene migration, and thus the acid[10] acts to accelerate the
process rather than facilitate it.[19] The approximately first-
order dependency on TfOH concentration suggests that it
assists in product liberation from the resting state (11) of the
catalytic cycle. Complex 11 cannot undergo syn-b-hydride
elimination at C8 without ring opening of the oxo-chelate
species.[20] Protonation of the carbonyl group in 11 by TfOH[10]
would assist this process.
6264 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6262 –6265

Angewandte
Chemie
1
model complexes suggest the difference in insertion rates (21.5
versus 30.1 kcalmolꢀ1) arises from interactions between the
terminal methyl group and the approaching allyl group in 6, see
the Supporting Information.
[17] The intermolecular H/2H transfers are reminiscent of a palla-
dium-catalyzed diene cycloisomerization mediated by a Pd(H)
species that has been generated in situ. Such reactions com-
monly display induction periods and pseudo zero-order sub-
strate dependency: a) K. L. Bray, I. J. S. Fairlamb, J. P. H.
Charmant, G. C. Lloyd-Jones, Chem. Eur. J. 2001, 7, 4205 –
4215; b) K. L. Bray, I. J. S. Fairlamb, J.-P. Kaiser, G. C. Lloyd-
Jones, P. A. Slatford, Top. Catal. 2002, 19, 49 – 59; c) G. C. Lloyd-
Jones, Org. Biomol. Chem. 2003, 1, 215 – 236; d) K. L. Bray, G. C.
Lloyd-Jones, M. P. Muꢂoz, P. A. Slatford, E. H. P. Tan, A. R.
Tyler-Mahon, P. A. Worthington, Chem. Eur. J. 2006, 12, 8650 –
8663; e) R. S. Widenhoefer, Acc. Chem. Res. 2002, 35, 905 – 913;
f) L. A. Goj, G. A. Cisneros, W. Yang, R. A. Widenhoefer, J.
Organomet. Chem. 2003, 687, 498 – 507.
[9] CCDC 720419 (7.H2O) and 720420 (4) contain the supplemen-
tary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[10] HBF4·Et2O, and CF3CO2H were also effective, however, AcOH,
TsOH, and HCl were not. Reactions induced by HBF4·Et2O
were the most efficient but varied substantially from batch to
batch, even though the addition of NaF or TBAF to reactions
conducted in the presence of TfOH had no effect. TBAF = tetra-
n-butylammonium fluoride.
[11] E. S. Stoyanov, K.-C. Kim, C. A. Reed, J. Phys. Chem. A 2004,
108, 9310 – 9315.
[12] M. P. Muꢂoz, G. C. Lloyd-Jones, Eur. J. Org. Chem. 2009, 516 –
524.
[18] Addition of a stoichiometric quantity of [(PCy3)2Pd(H)Cl] (Cy =
cyclohexyl; prepared according to I. D. Hills, G. C. Fu, J. Am.
Chem. Soc. 2004, 126, 13178 – 13179) to 4 in CDCl3 led to very
slow decomposition to give a mixture of products, including 5
(ca. 50% over 12 h), but with no trace of 7 evident by 1H NMR
analysis. Addition of 10 mol% [PdCl2(MeCN)2] to 4 in CDCl3
(2 h) and subsequent addition of 10 mol% Ph3SiH resulted in no
significant change over 12 h.
[19] TfOH may also generate the active catalyst 10 from 4, for
example by demethylative lactonization (see reference [12])
then b-hydride elimination at C6. Despite careful GCMS/NMR
analysis, we were unable to identify any coproducts from this
process.
[13] Preliminary DFT studies of intramolecular mechanisms identi-
fied feasible intermediates for B and C (12 and 13 kcalmolꢀ1
above 4), but the calculated barriers in excess of 40 kcalmolꢀ1
for C, D, and E would not be energetically accessible under the
general reaction conditions (218C), and no suitable transition
state was located for pathway B. See the Supporting Information
for a discussion.
[14] a) P. Drozdzewski, M. Musiala, M. Kubiak, Aust. J. Chem. 2006,
59, 329 – 335; b) P. Drozdzewski, A. Brozyna, M. Kubiak, T. Lis,
Vib. Spectrosc. 2006, 40, 118 – 126, for preparation of simple
104Pd/110Pd mixed-label coordination complexes.
[20] A five-ring oxo-chelate with palladium s-bonded to C8 is also
feasible, provided that the requisite Pd,H dyotropy (see
reference [1d]) is stereospecific and reversible, so that there is
no 2H/1H scrambling between C8 and C9.
[15] See the Supporting Information for additional reactions of 4, and
analogues, that reinforce this conclusion.
[16] Complexes tested: Z = diethyl, ethyl methyl, and di(tert-butyl)
malonate, N-tosylamide, 4’,4’-dimethyl-3’,5’-dioxan-2’,6’-dionyl,
and 3’,5’-dioxan-4-onyl. See the Supporting Information.
Angew. Chem. Int. Ed. 2009, 48, 6262 –6265
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6265