The intermolecular asymmetric Heck reaction: Mechanistic and computational studies

Article abstract of DOI:10.1002/1522-2675(20011017)84:10<3043::AID-HLCA3043>3.0.CO;2-V

Reactive intermediates in the asymmetric Heck reaction between aryl electrophiles and 2,3-dihydrofuran have been identified by NMR and mass spectrometry, with BINAP or the achiral diphosphanes dppp and dppf as ligands. The major cationic species observed

Full text of DOI:10.1002/1522-2675(20011017)84:10<3043::AID-HLCA3043>3.0.CO;2-V

Helvetica Chimica Acta ± Vol. 84 (2001)
3043
The Intermolecular Asymmetric Heck Reaction: Mechanistic and
Computational Studies1)
by K. K. (Mimi) Hii2), Timothy D. W. Claridge, and John M. Brown*
Dyson Perrins Laboratory, South Parks Rd., Oxford, OX13QY, UK (phone: (‡44)-1865-275642;
fax: (‡44)-1865-275674; e-mail: john.brown@chem.ox.ac.uk)
and Andrew Smith and Robert J. Deeth*
Department of Chemistry, University of Warwick, Coventry, CV47AL, UK (phone: (‡44)-24-76-523187;
fax: (‡44)-24-76-524112; e-mail: mssbq@dredd.csv.warwick.ac.uk)
Dedicated to the memory of Professor Luigi M. Venanzi
Reactive intermediates in the asymmetric Heck reaction between aryl electrophiles and 2,3-dihydrofuran
have been identified by NMR and mass spectrometry, with BINAP or the achiral diphosphanes dppp and dppf as
ligands. The major cationic species observed is an alkylpalladium produced by addition of PdAr to the alkene
followed by two further PdH-mediated isomerisation steps. This last species has been characterised at 608 by
1H-, 13C-, and 31P-NMR, including HMQC techniques. The regiochemistry of PdAr and PdH addition to the
reactant are opposite as defined by the reaction with (2-2H1)-2,3-dihydrofuran. DFT Calculations on the
‡
reaction pathway between [CH2(PH2)]PdPh and 2,3-dihydrofuran reveal several structurally interesting
intermediates involving agostic b-H-atom, ipso-Ph or reactant O-atom bonded to the Pd-atom, and elucidate the
isomerisation pathway.
Introduction. ± The Heck reaction between an unsaturated electrophile and an
alkene may be catalysed by several different classes of Pd complexes and has a
correspondingly diverse mechanistic basis [2]. The reaction normally proceeds through
an addition-elimination pathway such that the reactant and product have sp2-
hybridisation at the site of substitution. For asymmetric synthesis to function, there
needs to be double-bond isomerisation in the course of reaction. Essentially all the
examples of enantioselective Heck reactions in the literature involve cationic
diphosphinepalladium complexes3), and the mechanistic studies reported herein are
based on the intermolecular examples discovered by Hayashi and co-workers [5], and
depicted in Scheme 1. In the addition of an aryl- or vinyl-Pd species to a cyclic alkene,
the initial addition step enforces PdH elimination away from the new C C bond, and,
in doing so, a new stereogenic centre is generated. Much of the published work has
involved aryl additions to 2,3-dihydrofuran (1) with BINAP as ligand, and our studies
have largely followed that example.
1
)
)
)
For preliminary communication of part of this work, see [1].
2
3
Present address: Department of Chemistry, Kingꢀs College, Strand, London, WC2R2LS, UK.
For reviews, see [3a][3b]; but for an exception involving neutral intermediates, see [3c]. For more recent
examples, see [4].

3044
Helvetica Chimica Acta ± Vol. 84 (2001)
Scheme 1. The Intermolecular Asymmetric Heck Reaction of 2,3-Dihydrofuran with Phenyl Triflate. With PP
ligands, the 2,3-dihydrofuran isomer predominates, but with PN ligands the 2,5-dihydrofuran isomer
predominates.
An interesting mechanistic sidelight emerged from the earliest reports of the
specified Heck reaction (see [6] and earlier papers cited therein). Depending on the
reaction conditions, the product was predominantly the 2,5- or 2,3-dihydro isomer 2 or
3, respectively, with excess PPh3 providing the latter. This implies an intermediate that
can partition between elimination and isomerisation. In the detailed studies carried out
by Hayashi and co-workers, both were formed with the latter favored; the enantiomer
excess (e.e.) of the major product was enhanced by a favorable partition of the
diastereoisomers of the first-formed intermediate, minor towards elimination and
major towards isomerisation [5]. Subsequently, Pfaltz and Loiseleur and co-workers
demonstrated that catalysis involving a PN ligand in the phosphinoaryldihydrooxazole
series provided entirely the 2,5-dihydroisomer in high e.e. [7]; other results obtained
were broadly in line with a single double-bond-shift pathway. Subsequent workers have
largely confirmed this regiochemical pattern, with diphosphine ligands giving rise
mainly to double-isomerisation products and with phosphinamine ligands to single-
isomerisation products.
When this work was started, there were no detailed mechanistic studies on the
asymmetric Heck reaction and, indeed, few on the reaction in general [8]. In the
intervening period, there has been an effusion of speculations about the detailed
mechanism of simple Heck reactions, especially on the role of Pd0 ± PdII vs. PdII ± PdIV
cycles [9]4), and the possible involvement of deligated colloidal Pd or Pd nanoparticles
[10]. Direct evidence on the structure of true late intermediates in the catalytic cycle
remains limited to the cationic-diphosphine case from our own work [1][7][11] and a
related contribution from Akermark and co-workers [12]. There are related studies on
alkene-Pd polymerisation or alkene/CO copolymerisation [13] in which comparable
intermediates have been identified and characterised.
Results and Discussion. ± Observation of Reactive Intermediates. In our original
communication on this topic, the observation of a single transient intermediate on the
catalytic pathway of the asymmetric Heck arylation reaction from compound 1 was
For a theoretical comparison of PdIV and PdII pathways, see [9c].
4
)

Helvetica Chimica Acta ± Vol. 84 (2001)
3045
described. This had the surprising structure 4 shown in Fig. 1, being formed by a double
migration of the original Pd arylation adduct. No evidence for a second diastereoisomer
on the alternative enantiomeric pathway was observed, and, at that stage, the distinc-
tive NMR spectrum of 2,5-dihydro-2-phenylfuran (2) (or its 2,3-dihydro-isomer 3) was
not present. Its structure was established by a multifaceted NMR approach, in which
the key experiments were based on the 1H-NMR spectra. The most distinctive feature
in the spectrum of complex 4 was the existence of two separate tertiary low-field H-
atom signals at 5.3 and 5.5 ppm, assigned by a combination of HMQC, 31P-decoupling,
and COSY experiments; the last of these located adjacent CH2 groups and completed
the H-atom connectivity of the ring. The NOE ꢁmapꢀ derived from ROESY
experiments is shown in Fig. 1 and suggestes that the Pd and Ph substituents are cis
related on opposite flanks of the ether moiety. A definitive choice between these two
was made by employing the 2-deuterated analogue of 1; the 1H-resonance at 5.5 ppm
was not observed in the spectrum of the intermediate. On the basis of additional
information from the ROESY experiment, where a cross-peak between the signal at
Fig. 1. NMR characterisation of the intermediate 4 observed by NMR in the range of
708 to 508;
a) connectivities from NOE data obtained by ROESY experiments; b) selected chemical shift and coupling data
derived from (1H,1H)- and (1H,31P)-COSY, -TOCSY, and -HMQC experiments that define the solution structure

3046
Helvetica Chimica Acta ± Vol. 84 (2001)
5.5 ppm of the Hendo C(5) and the signal at 1.7 ppm associated with Hendo C(4) was
observed, a likely geometry for the complex was suggested. In this, the five-membered
ring is in a twist conformation where the Ph group is equatorial and the Pd-atom with its
ligands is axial. This unusual geometry raised the possibility that the ether O-atom
participates in bonding to the Pd-atom. Whilst not proved by the data, the existence of a
unique five-bond coupling between the tertiary benzylic H-atom and one of the two P-
atoms, provided further support for the postulate. DFT Calculations (vide infra) also
support this possibility.
The intermediate 4 observed by NMR is stable at 608 for short periods and
decomposes on warming. GC Analysis of the product shows that it is exclusively 2
(< 5% 3) and is formed in 95% e.e. of the (S)-enantiomer, consistent with the catalytic
results from the work of Hayashi and co-workers [5]. When decomposition of complex
4 was monitored by 31P-NMR, two new species were observed concurrently. Cross-
1
1
reference to the H-NMR spectra, and especially the species formed in the H-NMR
spectrum of (2-2H1)-1 indicates that these are the diastereoisomers of the Pd complex 5
formed by the unselective addition of PdH to reactant 1 but with the Pd-atom now
adjacent to the heteroatom. Not all low-field 1H-resonances of the two diastereoisom-
ers of 5 have been uniquely identified, but there is sufficient information to be clear
about the regiochemistry of addition. Interestingly, the new intermediate 5 does not
react readily with a further molecule of reactant 1 to produce a di- or oligomeric
product by multiple Heck condensation.
A single experiment was carried out with the corresponding complex 6 derived
from the PN-ligand QUINAP. The reaction was considerably slower, and, at 108,
intermediates were not observed. Analysis of the reaction product by GC indicated the
formation of (S)-3 in 95% e.e., demonstrating the alternative regiochemistry of reaction
to the BINAP case.
Similar reaction conditions were established for dppp- or dppf-ligated reactant
complexes. In the first case, the initial intermediate 7 was unstable at 508, and hence
only its 31P-NMR spectrum was recorded. This species provided the second inter-
1
mediate 8 cleanly together with product 2. The H-NMR spectrum augments the
evidence for Pd C(2) bonding since there are three characteristic low-field 1H-
resonances, with the H
C Pd signal as a broad doublet; the low value of vicinal
coupling constants indicates an axial conformation for the C Pd bond, as in the initial
adducts. In the dppf case both 9 and 10 were spectroscopically characterised by 31P-
NMR, the latter exhibiting interesting dynamic behaviour, which precluded interpre-
tation of the 1H-NMR spectrum. The origin of this process equilibrating P(1) and P(2)

Helvetica Chimica Acta ± Vol. 84 (2001)
3047
is unknown but may indicate rapid reversible b-hydride transfer to the Pd-atom with an
intermediate that equivalences the two P-atoms.
DFT Studies of Initial Complex and Insertion Step. The starting point for
computation was the cationic complex 11 carrying both the Ph group and the alkene.
As in previous work [14], the ꢁcomputationalꢀ ligand H2PCH2PH2 was employed, and
the TfO counterion involved in the mechanistic studies described above was replaced
by I ion for ease of computation. With an achiral ligand there are two possible
conformations (exo or endo) linked by rotation about the alkene metal bond, which is
orthogonal to the square plane. The preferred plane of the Ph ring is also orthogonal.
Their respective binding energies, and the computed structure of the lower-energy
isomer are shown in Fig. 2. As expected [15], the b-C-atom of the enol-ether double-
bond is much closer to the square plane than is the a-C-atom, and C(b)-Pd at 2.21 Š is
much shorter than C(a)-Pd at 2.38 Š. In contrast to a free vinyl ether [16], the b-C-
atom is essentially neutral (Mulliken charge 0.02 e) whereas the a-C-atom is strongly
positive (0.58 e). The Ph group is h1-bound to the Pd-atom, and bond lengths within the
ring are typically aromatic. Negative charge is essentially localised on the ipso C-atom
(
0.35 e) providing a charge-based rationalisation for the regiochemistry of the Heck
migration step, since it is transferred exclusively to the positive a-C-atom of the vinyl
ether. The Pd Ph bond is very short. This resulting high trans influence gives rise to a
long trans Pd P bond.
In order for the migration step to occur, the alkene must first rotate into the square
plane, and the productive transition state involves charge-matched bonding of the Ph
group to C(a), indicated in Fig. 2. A confirmatory frequency calculation was
1
undertaken and generated one negative frequency at 234 cm . This frequency was
animated and corresponds to the coupling of C(2) and C(8). The C C distance is
Fig. 2. The reaction pathway of the first step of the Heck phenylation of 2,3-dihydrofuran from complex 11 to
1
complex 12, with DFT energies of the different states quoted in kJ mol

3048
Helvetica Chimica Acta ± Vol. 84 (2001)
2.16 Š, significantly longer than in related calculations, and having its likely origin in
the electrostatic attraction between the partners, leading to an earlier transition-state
(TS). In fact, C(2) is more negative at the TS and compensated by increased positive
charge at the ortho positions of the aryl ring. The transition-state for the disfavoured
transfer to C(b) was also found, and is 6.7 kJ mol 1 higher than that for the preferred
pathway. If realised in practice, this would correspond to ꢀ 95% regioselectivity in the
observed direction.
The terminus of the migration step was also subjected to computational analysis,
with the new ground-state 12 shown in Fig. 3. The most striking feature is the direct
Cipso Pd bond. This observation is in contrast to one related computational study of
Ph Pd insertion into ethene by carbene ligands; for the model (bis)diaminocarbene
ligand in 13 weak h2-Pd ipso-ortho bonding was observed at the same stage, whilst the
monocarbene complex 14 exhibited only Pd Cortho bonding [17]. There are, in
addition, well-established and X-ray-characterised examples of ipso bonding in Pd
complexes [18], raising the possibility that this substitution type could be important in
Pd catalysis when an aryl ring is in proximity to the catalytic centre. The relative
stability of the Cipso Pd bond in the present case can be established by introducing a
single H2O molecule into the coordination sphere. This leads to the displacement of the
Ph group and replacement by H2O at that coordination site. The C Pd s-bond rotates
to remove the Ph group from the vicinity of the metal, giving rise to the structure shown
boxed in Fig. 3. Summing the total energies, however, indicates that the ipso-bonded
intermediate is lower in energy than the aqua complex 18 (corrected by the energy of a
1
1
fully dissociated H2O molecule, 1357.4 kJ mol ) by 9.6 kJ mol .
DFT Studies of Pd H Migrations. To generate the alternative Pd-hydrides required
to explain the multiplicity of products in the experimental Heck reactions, possible
pathways for this step were evaluated empirically. Through a linear transit calculation
on structure 12 with constraints on C H distances designed to enforce hydride
migration, the ground-state structure of the next intermediate 15 was obtained in which
the integrity of the Cipso Pd bond was maintained. With these constraints an
improbably high-energy path was revealed, and structure 15 was indicated to be 7 kJ
mol 1 above structure 12 in binding energy. To gain more insight into the pathway the
transformation was run in reverse, emphasising PdH extraction without explicit
migration. The most favorable route discovered revealed a complex multi-minima
1
surface with a calculated energy barrier of 49 kJ mol . An important feature of this
pathway is that the Ph group rotates away from the Pd-atom, weakening the Cipso Pd
bond, and that b-hydride agostic bonding occurs concomitantly, giving structure 16.
This process continues with local minima to the eventual formation of a true h2-alkene-
1
Pd hydride 17 (Fig. 3), 42 kJ mol higher in energy than the original structure 12.
Interestingly, the double bond is in the coordination square plane in this intermediate,
such that the Ph group may interact weakly (Pd Cipso ˆ 2.63 Š) with the dz2 orbital
orthogonal to the same plane. The reverse process provides a model for the path of Pd
migration that is shown in Fig. 3.
If the isomerisation process is repeated once more from complex 15, the resulting
Pd alkyl is adjacent to the ring O-atom. Computation of the ground state of this final
intermediate 19 reveals that the Pd O bond is established in preference to
participation of b-hydride or aryl C-atoms. The calculations support experimental

Helvetica Chimica Acta ± Vol. 84 (2001)
3049
Fig. 3. The reaction pathway for the first isomerisation step from the alkyl-Pd intermediate. For ease of
computation, the ground states prior to and subsequent to isomerisation (compounds 12 and 16, resp.) were first
established, and then the two further intermediate minimum-energy states 17 and 18 were located by working
backwards from 16. The energy of the aqua complex 15 needs to be corrected for the binding energy of H2O
(
1457.4 kJ mol 1) in order to facilitate comparisons.

3050
Helvetica Chimica Acta ± Vol. 84 (2001)
observations in that the ultimate product of the Heck reaction of 2,5-dihydrofuran
depends on a competition between elimination and migration pathways, the latter
governed by a subtle interplay between differing weak ligands and the unsaturated Pd
intermediates (for full details of the computational chemistry, see [19]).
Fig. 4. The structure and energy of the final alkyl Pd intermediate 19 formed by double Pd H-mediated
rearrangement

Helvetica Chimica Acta ± Vol. 84 (2001)
3051
Conclusions. ± Overall, the work reported herein reinforces and refines previous
ideas about the asymmetric Heck reaction. The diverse product distributions observed
by Hayashi and co-workers, in which the proportions of products 2 and 3 are highly
sensitive to added AcOH and the nature of the base [5], may be economically
explained. Similarly, the preponderance of product 3 when PN ligated catalysts are
employed can be explained. The initial observation of Larock and co-workers that
excess PPh3 diverted the reaction pathway from predominantly 3 to predominantly 2,
can also be explained by the same mechanistic model. The detail is set out in Scheme 2,
an essential new factor being the easy accessibility and stability of the doubly
isomerised Pd alkyl. With an achiral ligand, the product of the first addition step
provides a branching point. The options are direct elimination of Pd H and release of
alkene, or dyotropic rearrangement to give the isomeric Pd alkyl. If the intermediate
Pd alkene hydride has the capacity to lose a ligand by dissociation (monophosphine
not in excess, or PN chelate) then alkene release is the preferred route. If the
intermediate is long-lived enough to rearrange (PP chelate), then a second dyotropic
step occurs rapidly so that the resting state is the doubly rearranged alkyl. The ultimate
product can only be the 2-arylated 2,3-dihydrofuran. For the asymmetric cases, two
possible diastereoisomers are formed in the initial step, and both proceed along the
pathways described. If the first intermediate is unstable (PN chelate), the product
enantiomer ratio is determined only by the stereoselectivity of the initial addition of the
Pd Ar group to the alkene. In cases where further rearrangement occurs (PP chelate),
the enantiomer ratio is determined by two factors ± the stereoselectivity of the initial
addition step, and also the branching of the first intermediate between alkene release
and isomerisation. In the stoichiometric NMR experiments, the sole intermediate
observed is one diastereoisomer from the double-rearrangement process, without any
evidence of product release at that stage. In the catalytic experiments of Hayashi and
co-workers, the more favored enantiomer of the product is formed via more stable
intermediates. A consequence of this is that the less-favored diastereoisomeric pathway
is more likely to lead to 3 rather than to 2. Together with our previous work, these
observations demonstrate that the Pd H formed in the Heck reaction is efficiently
captured by an excess of the alkene in the absence of base. When base is present,
however, this pathway is averted [20]. These previous calculations indicate that a Pd-
agostic b-hydride of significant lifetime would also be responsive to base-promoted
alkene formation.
The DFT calculations endorse the veracity of this interpretation, and offer insights
into the structure and energetics of the intermediates involved. The first-formed alkyl
complex is computed to possess strong Pd Cipso bonding, maintained in the first
rearrangement product. This isomerisation is endothermic. The second rearrangement
gives rise to the most stable of the three positional isomers of the adduct, whose
additional stability is conferred by significant Pd O bonding. At intermediate stages b-
agostic Pd H bonding can produce additional shallow minima. The energetics of the
various intermediates described in the computational part are summarised in Scheme 2.
The broad indications from these studies are borne out in practice ± the first-formed h2-
alkene adduct remains unobserved in realistic models for the Heck reaction because it
has a low-energy migration pathway in which the first alkyl is formed. The initially
formed alkyl Pd cation is higher in energy than the first intermediate by 32 kJ mol 1 in

3052
Helvetica Chimica Acta ± Vol. 84 (2001)
the model calculations so that the migration step is exothermic and probably
irreversible. Formation of a Pd hydride by b-elimination is endothermic by > 35 kJ
1
mol . The order of stability of the three alkyl Pd intermediates in the DFT
calculations is final > initial > intermediate, indicating the value of the additional
Pd O stabilisation in the final species.
We thank EPSRC for support (to K. K. H.), the University of Warwick for provision of computer
equipment and financial support (for A. S.), and Johnson-Matthey for the loan of Pd salts. We thank Prof. A.
Pfaltz for a useful discussion.
Scheme 2. Pathways for the Asymmetric Phenylation of 2,3-Dihydrofuran. The partitioning of the first-formed
Pd H complex B determines the course of reaction. With (S)-BINAP as ligand, hydride return is the main
pathway, giving the transient isomer C, which further rearranges to the most stable Pd alkyl D, the precursor of
Å
the major product. On the minor enantiomer pathway, Pd H complex B preferentially partitions to the 2,5-
Å
dihydrofuran rather than further rearrangement through C. Collapse to product through B or B is the exclusive
pathway when PN ligands are employed [7].

Helvetica Chimica Acta ± Vol. 84 (2001)
3053
Experimental Part
General. Experimental protocols and resources are taken from previous papers on this topic [21].
Computational Details. The DFT calculations were obtained with the Amsterdam Density Functional
(ADF) program suite Version 2.3 [22]. Geometries were optimised at the local density approximation level and
with binding energies computed subsequently with the Becke88/Perdew86 gradient-corrected functional
[23][24]. The frozen-core approximation [25] was applied with orbitals 1s-3d frozen on Pd-atoms, 1s-2p on P-
atoms and the 1 s orbital frozen for O- and C-atoms. Basis sets comprised triply-z ‡ polarisation STO expansions
(basis IV) throughout.
Preparations. General Procedures for the Preparation of [Pd(I)(Ph)P2]. Method A: A soln. of Li2(COT)
[26] (0.26m in Et2O, 1.5 equiv.) was added slowly to a stirred soln. of the corresponding [PdCl2P2] complex in
THF at 788. The resulting dark red soln. was stirred for 30 min followed by the addition of PhI (2 equiv.). The
mixture was then allowed to warm to r.t. and stirred overnight. The mixture was evaporated, and petroleum
ether (40 ± 608) was added to precipitate the crude product, which was filtered off and recrystallized from THF/
pentane to give the required compound as a creamy solid. Yields were between 65 ± 75%.
Method B.
A Schlenk tube charged with [Pd(I)(Ph)(tmeda)] [27] (150 mg, 0.35 mmol) and the
corresponding diphosphine (0.35 mmol) was purged-filled with Ar twice. Degassed toluene (3 ml) was added
and the soln. was stirred for 5 h, during which time the required compound precipitated from the soln. as a solid.
This was filtered off, washed (Et2O), and dried. A second crop of compound may be recovered by evaporating
the mother liquor to dryness, followed by the addition of Et2O. Typical yields are 65 ± 85%.
[Pd(BINAP)(I)(Ph)]. The complex decomposes rapidly at r.t. in chlorinated solvents, therefore the NMR
spectra were recorded at 243 K. 1H-NMR (500 MHz, CD2Cl2, 243 K): 6.47 (dd, J(H,H) ˆ 2.2, 9.8, H C(3));
6.66 (br. t, PPh); 6.71 (ddd, J(H,H) ˆ 1.5, 3.3, 7.3, 1 H); 6.74 (ddd, J(H,H) ˆ 1.1, 1.2, 7.2, H
Cortho); 6.86 ± 6.97
(m, 7 H); 6.95 (td, J(H,H) ˆ 1.2, 6.8, H C(4)); 7.04 (ddd, J(H,H) ˆ 1.0, 1.3, 8.8, H C(9)); 7.12 (m, 2 H); 7.18
(ddd, J(H,H) ˆ 1.3, 6.8, 8.8, H C(10)); 7.29 (ddd, J(H,H) ˆ 1.2, 6.8, 8.1, H C(5)); 7.34 (dt, J(H,H) ˆ 8.8,
J(P,H) ˆ 1.0, H C(8)); 7.37 (ddd, J(H,H) ˆ 1.3, 6.8, 8.1, H C(11)); 7.44 (br. m, 4 H, PPh); 7.51 ± 7.55 (m, 4 H);
7.58 (dd, J(H,H) ˆ 1.2, 8.3, H C(6)); 7.69 (dd, J(H,H) ˆ 1.3, 8.3, H C'ortho); 7.74 (br. m, 4 H, PPh); 7.78
(dd, J(H,H) ˆ 8.6, J(P,H) ˆ 2.2,
H
C(2)); 7.93 (dd, J(H,H) ˆ 8.6, J(P,H) ˆ 11.0,
H
C(1)). 31P-NMR
‡
(202 MHz, CD2Cl2 , 243 K): 19.8 (d, J ˆ 39); 24.3 (d, J ˆ 39). ES-MS (MeOH, 25 eV): 805 ([M I]
‡
[Pd(BINAP)(Ph)] ).
[Pd(dppp)(I)(Ph)]. 1H-NMR (500 MHz, CD2Cl2): 7.1 ± 7.9 (m, 20 H, PPh); 6.89 (td, J(H,H) ˆ 7.8,
J(P,H) ˆ 2.2, 7.8, 2 H, H C(3)); 6.53 (ddd, J(H,H) ˆ 7.3, 7.6, J(P,H) ˆ 5.5, 2 H, H C(2)); 6.45 (t, J(H,H) ˆ
7.4, H C(4)); 2.55 (m,CH2P); 2.44 (m, CH2P); 1.80 (m, CH2). 13C-NMR (125.7 MHz, CD2Cl2): 19.5
(d, J(P,C) ˆ 4, CH2); 27.6 (dd, J(P,C) ˆ 3, 19, CH2P); 29.1 (dd, J(P,C) ˆ 7, 25, CH2P); 122.0 (s, C(4)); 127.3
(d, J(P,C) ˆ 9, C(2)); 138.0 (d, J(P,C) ˆ 3, C(3)); 156.7 (dd, J(P,C) ˆ 3, 128, C(1)). 31P-NMR (202 MHz,
CD2Cl2): 8.8 (d, J ˆ 52), 12.2 (d, J ˆ 52).
[Pd(I)(Ph)(R)-(QUINAP)]. Super Hydride ꢂ (1m soln. in THF, 0.5 ml, 0.5 mmol) was added to a stirred
suspension of [PdCl2(R)-(QUINAP)] (117 mg, 0.2 mmol) and 1,5-cyclooctadiene (0.1 cm3) in THF (5 ml) at
788. When effervescence ceased after 15 min, PhI (0.1 ml) was added and the soln. was warmed to r.t. very
slowly (over 5 h). The solvents were then removed completely, and the addition of pentane precipitated the
crude product as a brown solid, which was recrystallized from toluene/pentane to give the required complex as a
buff-coloured solid (50 mg, 35%). 1H-NMR (500 MHz, (D6)acetone): 6.52 (tt, J(H,H) ˆ 1.2, 7.2, 1 H); 6.62
(br. t, J(H,H) ˆ 7.5); 6.91 (dddd, J(H,H) ˆ 1.0, 1.0, 1.9, 8.7, H C(3)); 6.96 (ddd, J(H,H) ˆ 1.3, 8.5, J(P,H) ˆ
11.4, 1 H); 7.07 (ddd, J(H,H) ˆ 1.0, 1.5, 8.6, H C(9)); 7.17 (dd, J(H,H) ˆ 7.4, J(P,H) ˆ 2.6, 1 H); 7.16 ± 7.26 (m);
7.18 (ddd, J(H,H) ˆ 1.5, 6.8, 8.3, H C(11)); 7.30 ± 7.50 (m); 7.33 (ddd, J(H,H) ˆ 1.0, 6.8, 8.7, H C(4)); 7.40
(ddd, J(H,H) ˆ 1.1, 6.8, 8.6, H C(10)); 7.42 (m, H C(7)); 7.68 (ddd, J(H,H) ˆ 1.0, 6.8, 8.1, 1 H); 7.71
(ddd, J(H,H) ˆ 1.9, 6.8, 8.3, H C(5)); 7.76 ± 7.80 (m); 7.89 (ddd, J(H,H) ˆ 1.0, 1.0, 8.3, H C(6)); 7.92
(dd, J(H,H) ˆ 1.0, 6.3, H C(2)); 8.04 (ddd, J(H,H) ˆ 0.8, 8.5, 8.5, 1 H); 8.17 (dd, J(H,H) ˆ 1.0, 8.3, H C(12));
8.27 (ddd, J(H,H) ˆ 1.0, 8.6, J(P,H) ˆ 1.8, H C(8)); 9.65 (d, J(H,H) ˆ 6.3, H C(1)). 31P-NMR (202 MHz,
(D6)acetone): 27.5.
Preparation of (2-2H1)-2,3-Dihydrofuran by a Modified Procedure [28]. In a dry Schlenk tube, BuLi (2.5m in
hexane, 13 ml) was added rapidly to a mixture of freshly distilled 2,3-dihydrofuran (2 g, 28.5 mmol) and tmeda
(0.6 g), whereupon vigorous effervescence and precipitation occurred. The mixture was then cooled (water
bath) and the addition was completed with stirring. The solid was allowed to settle before the liquid was
decanted. The residue was washed with dry pentane (3 Â 6 ml), and dried in vacuo (Caution: highly pyrophoric
solid). The solid was subsequently suspended in dry toluene (2 ml) and cooled to 308. D2O was added

3054
Helvetica Chimica Acta ± Vol. 84 (2001)
dropwise very slowly, and the soln. was allowed to warm to r.t. gradually. The deuterated 2,3-dihydrofuran was
obtained in 60% yield by distillation of the mixture over a 4'' Vigreux column. 1H-NMR (CDCl3): 2.33 (dt, J ˆ
2.5, 9.5, CH2); 4.03 (t, J ˆ 9.5, CH2O); 4.68 (t, J ˆ 2.5, ˆCH).
Reactions with 2,3-Dihydrofuran. Typical Procedure for the Reaction of [Pd(P2)(Ph)(S)]TfO with 2,3-
Dihydrofuran. AgOTf (1.1 equiv.) was added to a soln. of the corresponding [Pd(Ar)(diphosphane)(I)]
(0.02 mmol) in (D8)THF (0.5 ml) at 788. After 10 min of vigorous stirring, the mixture was centrifuged at
608. The pale yellow soln. was decanted quickly into a NMR tube containing the dihydrofuran (6 ml). The
mixture was transferred into the NMR spectrometer cooled at 608. The first intermediate was observed as a
single diastereoisomer 4, which was converted to the diastereoisomeric second intermediates 5a and 5b by
raising the probe temp. gradually to 208, where it decomposed to give the turnover product and a mixture of
two diastereoisomers. When the mixture was analysed, the turnover product (S)-2,3-dihydro-2-phenylfuran was
found in 95% e.e. (chiral GC).
P2 ˆ (S)-BINAP: 1H-NMR (500 MHz, 238 K): 5.50 (ddd, J(P,H) ˆ 8.5, 12.0, J(H,H) ˆ 4.8, H C(5)); 5.30
(dd, J(P,H) ˆ 15.5, J(H,H) ˆ 3.6, H C(2)); 2.70 (dddd, J(H,H) ˆ 5, 7, 12, 18, H C(4)); 2.20 (ddd, J(H,H) ˆ
4.8, 11, 18, H C(4)); 1.70 (ddddd, J(P,H) ˆ 1.0, J(H,H) ˆ 3.6, 5, 11, 14.7, H C(3)); 1.0 (ddd, J(P,H) ˆ 7.3,
J(H,H) ˆ 7.0, 14.7, H C(3)). 13C-NMR (125.7 MHz, 238 K): 28.6 (s, C(3)); 32.5 (s, C(4)); 92.7 (s, C(2)); 97.2
(d, J(P,C) ˆ 17, 60, C(5)). 31P-NMR (202 MHz, 238 K): 22.5 (d, J(P,P) ˆ 35); 37.0 (d, J(P,P) ˆ 35). ES-MS,
MeOH, 20 eV): 878 ([M ‡ 1]), 800 (M Ph).
1
P2 ˆ dppf: H-NMR (500 MHz, (D8)THF, 238 K): 1.02 (ddd, J(H,H) ˆ 6.6, 15, J(P,H) ˆ 7, H C(4)); 1.51
(m, H C(4')); 2.17 (br. m, H C(3)); 2.46 (br. m, H C(3')); 4.12 (m, 1 H of Cp); 4.17 (m, 2 H of Cp); 4.41
(m, 1 H of Cp); 4.45 (m, 1 H of Cp); 4.50 (m, 1 H of Cp); 4.55 (m, 1 H of Cp); 4.72 (m, 2 H of Cp); 4.76 (m, 1 H
of Cp); 4.97 (dd, J(P,H) ˆ 20, J(H,H) ˆ 3.7, H C(5)); 5.57 (ddd, J(P,H) ˆ 12, J(H,H) ˆ 4, 12, H C(2)).
13C-NMR (125.7 MHz, 238 K): 32.5 (s, C(4)); 92.7 (s, C(2)); 98.8 (s, C(5)). 31P-NMR (202 MHz, (D8)-THF,
228 K): 13.1 (d, J(P,P) ˆ 18); 36.0 (d, J(P,P) ˆ 18.
P2 ˆ dppp: unstable with respect to product elimination at 233 K. 31P-NMR (202 MHz, 213 K): 0.7
(J(P,P) ˆ 60); 31.7 (J(P,P) ˆ 60). ES-MS (MeOH, 40 eV): 807 ([PdP2(Ph) ‡ C4H6O]), 731 ([PdP2 ‡ C4H6O]).
Product Characterisation: P2 ˆ (S)-BINAP: Formation of two diastereoisomers was observed at 243 K.
Diastereoisomer A. 1H-NMR (500 MHz, (D8)THF, 243 K): 5.14 (dd, J ˆ 3.5, J(P,H) ˆ 17.3, 1 H); 4.75
(br. t, J(H,H) ˆ 7.9 ) .
Diastereoisomer B. 1H-NMR (500 MHz, (D8)THF, 243 K): 4.68 (t, J(H,H) ˆ 8.5, 1 H); 4.60 (dd, J(H,H) ˆ
3.9, J(P,H) ˆ 15.8). 31P-NMR (202 MHz): 19.6 (J(P,P) ˆ 27); 22.4 (J(P,P) ˆ 34); 37.7 (J(P,P) ˆ 27); 38.8
(J(P,P) ˆ 34).
P2 ˆ dppf: 31P-NMR (202 MHz, 233 K): 14.4 (d, J(P,P) ˆ 22), 36.0 (d, J(P,P) ˆ 22).
P2 ˆ dppp: 1H-NMR (500 MHz, (D8)THF, 233 K): 1.20 (m, H C(4)); 1.31 (m, H C(4)); 1.81
(m, H C(3)); 2.05 ± 2.30 (br. m, CH2); 2.48 (m, H C(3)); 2.73 (m, 1 H, CH2P); 2.82 (ddd, J(H,H) ˆ 8.0,
15.3, J(P,H) ˆ 15.3, 1 H, CH2P); 2.94 (br. t, 1 H, CH2P); 3.30 (br. m, 1 H, CH2P); 3.92 (br. m, H C(2)); 4.11
(br. t, J(H,H) ˆ 8.0, H C(2)); 4.62 (dd, J(H,H) ˆ 3.8, J(P,H) ˆ 16.1, H C(5)). 31P-NMR (202 MHz, 233 K):
2.2 (J(P,P) ˆ 47), 31.4 (J(P,P) ˆ 47).
Reaction of [Pd(Ph)(R)-(QUINAP)(S)][TfO] with 2,3-Dihydrofuran. AgOTf (3.5 mg) was added to a
soln. of the aryl-Pd halide complex (10 mg, 0.01 mmol) in (D8)THF (0.75 ml) at 788. 31P-NMR revealed the
formation of a new species (25.0 ppm) which was stable up to 108, whereupon slow decomposition took place. In
the presence of 2,3-dihydrofuran, this was also found to be the reaction temp. Analysis of the org. product
formed in situ was found to be the (S)-2,5-dihydro product in 95% e.e., as analysed by chiral GC.
Catalytic Studies. 1a (P2 ˆ (S)-BINAP) (5 mol-%) was generated by the addition of AgOTf (6 mg) to a
soln. of [Pd(I)(P2)(Ph)] (20 mg, 0.02 mmol) in THF at 788. 2,3-Dihydrofuran (155 mg, 2.2 mmol), Et3N
(135 mg, 1.3 mmol) and PhOTf (100 mg, 0.44 mmol) were added sequentially. The mixture was gradually
warmed to r.t., then warmed to 408. The reaction progress was monitored by GC-MS, which indicated the
formation of one single regioisomer. After 22 h (99% consumption of PhOTf), the mixture was poured into
pentane (20 ml) and filtered through a short column of silica gel. The filtrate was evaporated to yield a yellow
oily residue, which was found to be the 2,3-regioisomer only (1H-NMR) and the (S)-enantiomer in 74% e.e.
(chiral GC; lit.: 74% e.e. (S)).
[Pd(S)-(BINAP)(Ph)(S)]TfO was generated as before from the addition of AgOTf (7 mg) to a stirred
soln. [Pd(S)-(BINAP)(I)(Ph)] in THF (1.5 ml) at 788. After the removal of the precipitated Ag salt by
centrifugation, the pale yellow soln. was transferred cold to a Schlenk tube containing PhOTf (113 mg,
0.5 mmol) 2,3-dihyrofuran (0.21 ml) and Bu4NOAc (450 mg, 1.5 mmol). The mixture was warmed slowly to r.t.,
then warmed at 408 for 72 h. The reaction only achieved 31% conversion as indicated by GC, with the exclusive

Helvetica Chimica Acta ± Vol. 84 (2001)
3055
formation of one regioisomer. The mixture was poured into Et2O, and washed with H2O. The org. layer was
separated, dried (MgSO4), and brought to dryness. Analysis of the residue revealed the presence of only 2,5-
dihydro-2-phenylfuran (3), racemic by Cydex-B GC analysis.
REFERENCES
[1] K. K. Hii, T. D. W. Claridge, J. M. Brown, Angew. Chem., Int. Ed. 1997, 36, 984.
[2] I. P. Beletskaya, A. V. Cheprakov, Chem. Rev. 2000, 100, 3009; G. T. Crisp, Chem. Soc. Rev. 1998, 27, 427; A.
Demeijere, F. E. Meyer, Angew. Chem., Int. Ed. 1995, 33, 2379.
[3] a) P. J. Guiry, A. J. Hennessy, J. P. Cahill, Top. Catal. 1997, 4, 311; b) M. Shibasaki, E. M. Vogl, J. Organomet.
Chem. 1999, 576, 1; c) L. E. Overman, D. J. Poon, Angew. Chem., Int. Ed. 1997, 36, 518.
[4] N. G. Andersen, R. McDonald, B. A. Keay, Tetrahedron: Asymmetry 2001, 12, 263; A. V. Malkov, M. Bella,
I. G. Stara, P. Kocovsky, Tetrahedron Lett. 2001, 42, 3045; S. R. Gilbertson, Z. Fu, Org. Lett. 2001, 3, 161;
S. R. Gilbertson, Z. C. Fu, D. J. Xie, Tetrahedron Lett. 2001, 42, 365; K. Yonehara, K. Mori, T. Hashizume,
K. G. Chung, K. Ohe, S. Uemura, J. Organomet. Chem. 2000, 603, 40; N. G. Andersen, M. Parvez, B. A. Keay,
Org. Lett. 2000, 2, 2817; T. Benincori, E. Cesarotti, O. Piccolo, F. Sannicolo, J. Org. Chem. 2000, 65, 2043;
W. P. Deng, X. L. Hou, L. X. Dai, X. W. Dong, Chem. Commun. 2000, 1483; S. R. Gilbertson, D. G. Genov,
A. L. Rheingold, Org. Lett. 2000, 2, 2885; Y. Hashimoto, Y. Horie, M. Hayashi, K. Saigo, Tetrahedron:
Asymmetry 2000, 11, 2205; A. J. Hennessy, D. J. Connolly, Y. M. Malone, P. J. Guiry, Tetrahedron Lett. 2000,
41, 7757; A. J. Hennessy, Y. M. Malone, P. J. Guiry, Tetrahedron Lett. 2000, 41, 2261; M. Ogasawara, K.
Yoshida, T. Hayashi, Heterocycles 2000, 52, 195; K. Selvakumar, M. Valentini, P. S. Pregosin, A. Albinati, F.
Eisentrager, Organometallics 2000, 19, 1299; L. F. Tietze, K. Thede, Synlett 2000, 1470; A. J. Hennessy,
Y. M. Malone, P. J. Guiry, Tetrahedron Lett. 1999, 40, 9163; L. F. Tietze, K. Thede, F. Sannicolo, Chem.
Commun. 1999, 1811; M. Tschoerner, P. S. Pregosin, A. Albinati, Organometallics 1999, 18, 670.
[5] T. Hayashi, A. Kubo, F. Ozawa, Pure Appl. Chem. 1992, 64, 421; F. Ozawa, T. Hayashi, J. Organomet. Chem.
1992, 428, 267; F. Ozawa, A. Kubo, T. Hayashi, Tetrahedron Lett. 1992, 33, 1485; F. Ozawa, A. Kubo, Y.
Matsumoto, T. Hayashi, E. Nishioka, K. Yanagi, K. Moriguchi, Organometallics 1993, 12, 4188.
[6] R. C. Larock, W. H. Gong, J. Org. Chem. 1990, 55, 407.
[7] O. Loiseleur, M. Hayashi, M. Keenan, N. Schmees, A. Pfaltz, J. Organomet. Chem. 1999, 576, 16; O.
Loiseleur, M. Hayashi, N. Schmees, A. Pfaltz, Synthesis 1997, 1338; O. Loiseleur, P. Meier, A. Pfaltz, Angew.
Chem., Int. Ed. 1996, 35, 200.
[8] J. M. Brown, K. K. Hii, Angew. Chem., Int. Ed. 1996, 35, 657; C. Amatore, E. Carre, A. Jutand, M. A.
Mbarki, G. Meyer, Organometallics 1995, 14, 5605; W. A. Herrmann, C. Brossmer, K. Ofele, M. Beller, H.
Fischer, J. Organomet. Chem. 1995, 491, C 1.
[9] a) B. L. Shaw, New J. Chem. 1998, 22, 77; b) B. L. Shaw, S. D. Perera, E. A. Staley, Chem. Commun. 1998,
1361; Milstein; c) A. Sundermann, O. Uzan, J. M. L. Martin, Chem. ± Eur. J. 2001, 7, 1703.
[10] A. Biffis, M. Zecca, M. Basato, Eur. J. Inorg. Chem. 2001, 1131; F. Y. Zhao, M. Shirai, M. Arai, J. Mol.
Catal., A 2000, 154, 39; J. H. Ding, D. L. Gin, Chem. Mater. 2000, 12, 22; M. T. Reetz, E. Westermann,
Angew. Chem., Int. Ed. 2000, 39, 165.
[11] J. M. Brown, J. J. Perez-Torrente, N. W. Alcock, H. J. Clase, Organometallics 1995, 14, 207.
[12] M. Ludwig, S. Stromberg, M. Svensson, B. Akermark, Organometallics 1999, 18, 970.
[13] C. S. Shultz, J. Ledford, J. M. DeSimone, M. Brookhart, J. Am. Chem. Soc. 2000, 122, 6351; D. J. Tempel,
L. K. Johnson, R. L. Huff, P. S. White, M. Brookhart, J. Am. Chem. Soc. 2000, 122, 6686.
[14] R. J. Deeth, A. Smith, K. K. Hii, J. M. Brown, Tetrahedron Lett. 1998, 39, 3229.
[15] D. W. Price, M. G. B. Drew, K. K. Hii, J. M. Brown, Chem.±Eur. J. 2000, 6, 4587.
[16] L. R. Domingo, Eur. J. Org. Chem. 2000, 2265; L. R. Domingo, M. Arno, J. Andres, J. Org. Chem. 1999, 64, 5867.
[17] K. Albert, P. Gisdakis, N. Rosch, Organometallics 1998, 17, 1608.
[18] W. Lesueur, E. Solari, C. Floriani, A. Chiesi-Villa, C. Rizzoli, Inorg. Chem. 1997, 36, 3354.
[19] A. Smith, Ph.D. Thesis, Warwick University, 2001.
[20] K. K. Hii, unpublished work.
[21] A. Jutand, K. K. M. Hii, M. Thornton-Pett, J. M. Brown, Organometallics 1999, 18, 5367.
[22] ADF version 2.3.0, Theoretical Chemistry, Vrije Universiteit, Amsterdam, Netherlands, 1997.
[23] A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098.
[24] J. P. Perdew, Phys. Rev. B: Condens. Matter 1986, 33, 8822.
[25] E. J. Baerends, D. E. Ellis, P. Ros, Theor. Chim. Acta 1972, 27, 339.

3056
Helvetica Chimica Acta ± Vol. 84 (2001)
[26] J. L. Spencer, Inorg. Synth. 1976, 19, 214.
[27] P. L. Alsters, J. Boersma, G. van Koten, Organometallics 1993, 12, 1629.
[28] F. T. Oakes, J. F. Sebastian, J. Org. Chem. 1980, 45, 4959.
Received June 14, 2001