Thermodynamic and kinetic relationships between Lewis acid-Lewis base complexes relevant to the P2Pt(OTf)2-catalyzed Diels-Alder reaction

Article abstract of DOI:10.1021/om0109861

The Pt(II) Lewis acids P₂Pt(OTf)₂ (2; P₂ = dppe (a), R-BINAP (b)), which catalyze the Diels-Alder reaction of acryloyl-N-oxazolidinone (4) with cyclopentadiene (HCp), were generated in situ by activation of P₂Pt(S-BINOL) precursors 1 with 2 equiv of triflic acid (HOTf). Catalysts 2 and catalytically relevant Lewis acid - Lewis base complexes [P₂Pt(L₂)]²⁺-[OTf]₂^- (L₂ = 2 H₂O (3), dienophile 4 (6), Diels-Alder adduct 5 (8)) were characterized by ¹H and ³¹P NMR spectroscopy at 195 K. 2a was also characterized by X-ray crystallography. The thermodynamic relationships between 2, 3, 6, 8, and a BE₄^- analogue of 2 were determined through competitive binding experiments monitored by ³¹P NMR at 195 K, to assess the effects of competitive inhibition of counterions, additives, and product on substrate coordination to the catalyst. These experiments demonstrated that the Lewis bases bind to [P₂Pt]²⁺ with relative strengths BF₄^- ? OTf^- < 4 < 5 ? H₂O. The rates of two ligand exchange processes (dienophile/dienophile (k₁) and dienophile/OTf^- (k₂)) were measured at 270 K by simulation of exchange-broadened ³¹P NMR spectra of 2/6 equilibrium mixtures, revealing that ligand substitution occurs more rapidly in the dppe system than the R-BINAP system (180 times faster for k₂). Finally, the reaction of HCp with 6 to give 8 was monitored by ³¹P and ¹H NMR at 195 K; attack of HCp on Pt-coordinated 4 occurred much faster than ligand exchange, indicating that ligand exchange is the turnover-limiting step of the catalytic cycle.

Full text of DOI:10.1021/om0109861

Organometallics 2002, 21, 1565-1575
1565
Th er m od yn a m ic a n d Kin etic Rela tion sh ip s betw een
Lew is Acid -Lew is Ba se Com p lexes Releva n t to th e
P 2P t(OTf)2-Ca ta lyzed Diels-Ald er Rea ction
Nicole M. Brunkan, Peter S. White, and Michel R. Gagne´*
Department of Chemistry, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-3290
Received November 14, 2001
The Pt(II) Lewis acids P2Pt(OTf)2 (2; P2 ) dppe (a ), R-BINAP (b)), which catalyze the
Diels-Alder reaction of acryloyl-N-oxazolidinone (4) with cyclopentadiene (HCp), were
generated in situ by activation of P2Pt(S-BINOL) precursors 1 with 2 equiv of triflic acid
(HOTf). Catalysts 2 and catalytically relevant Lewis acid-Lewis base complexes [P2Pt(L2)]2+-
[OTf]- (L2 ) 2 H2O (3), dienophile 4 (6), Diels-Alder adduct 5 (8)) were characterized by
2
1H and 31P NMR spectroscopy at 195 K. 2a was also characterized by X-ray crystallography.
-
The thermodynamic relationships between 2, 3, 6, 8, and a BF4 analogue of 2 were
determined through competitive binding experiments monitored by 31P NMR at 195 K, to
assess the effects of competitive inhibition of counterions, additives, and product on substrate
coordination to the catalyst. These experiments demonstrated that the Lewis bases bind to
-
[P2Pt]2+ with relative strengths BF4 , OTf- < 4 < 5 , H2O. The rates of two ligand
exchange processes (dienophile/dienophile (k1) and dienophile/OTf- (k2)) were measured at
270 K by simulation of exchange-broadened 31P NMR spectra of 2/6 equilibrium mixtures,
revealing that ligand substitution occurs more rapidly in the dppe system than the R-BINAP
system (180 times faster for k2). Finally, the reaction of HCp with 6 to give 8 was monitored
1
by 31P and H NMR at 195 K; attack of HCp on Pt-coordinated 4 occurred much faster than
ligand exchange, indicating that ligand exchange is the turnover-limiting step of the catalytic
cycle.
In tr od u ction
and Pt(II) Lewis acids that have emerged over the last
several years.6 Because their stereoelectronic properties
are inherently different from those of early transition
metals and the first-row late metals Cu(II) and Zn(II),
Stereoselective carbon-carbon bond-forming reac-
tions catalyzed by chiral, electrophilic transition-metal
Lewis acid complexes have become essential tools for
the synthesis of many enantiopure organic molecules.1
Although Lewis acids containing transition metals from
nearly all regions of the periodic table, from Sc(III) to
Zn(II),1b have been utilized as catalysts for these
transformations, more attention has historically been
devoted to the study of early-metal and Cu(II)/Zn(II)2
catalysts than to examination of Lewis acids based on
group 8-10 metals. For example, the chemistry of well-
defined, chiral Ru(II)- and Fe(II)-based Lewis acids
(group 8) has only recently been developed,3 while group
9 Lewis acid catalysts are even more rare.4 From group
10, chiral octahedral Ni(II) coordination compounds
have been used as enantioselective Lewis acid catalysts
by Kanemasa.5 However, the most numerous catalysts
from this “center-right” area of the periodic table are
the dicationic, square-planar, diphosphine-based Pd(II)
these [P2M]2+[A]- (P2 ) diphosphine; M ) Pd or Pt;
2
A- ) OTf-, BF4-, PF6-, SbF6-, ClO4-) Lewis acids have
(3) Fe: (a) Ku¨ndig, E. P.; Bourdin, B.; Bernardinelli, G. Angew.
Chem., Int. Ed. Engl. 1994, 33, 2856-1857. (b) Bruin, M. E.; Ku¨ndig,
E. P. Chem. Commun. 1998, 2635-2636. Ru: (c) Ku¨ndig, E. P.;
Saudan, C. M.; Bernardinelli, G. Angew. Chem., Int. Ed. 1999, 38,
1219-1223. (d) Carmona, D.; Cativiela, C.; Elipe, S.; Lahoz, F. J .;
Lamata, M. P.; Lo´pez-Ram de V´ıu, M. P.; Oro, L. A.; Vaga, C.; Viguri,
F. Chem. Commun. 1997, 2351-2352. (e) Davenport, A. J .; Davies, D.
L.; Fawcett, J .; Garratt, S. A.; Russell, D. R. J . Chem. Soc., Dalton
Trans. 2000, 4432-4441. (f) Faller, J . W.; Grimmond, B. J . Organo-
metallics 2001, 20, 2454-2458.
(4) (a) Davenport, A. J .; Davies, D. L.; Fawcett, J .; Garratt, S. A.;
Lad, L.; Russell, D. R. Chem. Commun. 1997, 2347-2348. (b) Carmona,
D.; Lahoz, F. J .; Elipe, S.; Oro, L. A.; Lamata, M. P.; Viguri, F.; Mir,
C.; Cativiela, C.; Viu, M. P. L.-R. d. Organometallics 1998, 17, 2986-
2995.
(5) (a) Kanemasa, S.; Kanai, T. J . Am. Chem. Soc. 2000, 122, 10710-
10711. (b) Kanemasa, S.; Oderaotoshi, Y.; Sakaguchi, S.; Yamamoto,
H.; Tanaka, J .; Wada, E.; Curran, D. P. J . Am. Chem. Soc. 1998, 120,
3074-3088.
(6) (a) Ghosh, A. K.; Matsuda, H. Org. Lett. 1999, 1, 2157-2159.
(b) Pignat, K.; Vallotto, J .; Pinna, F.; Strukul, G. Organometallics 2000,
19, 5160-5167. (c) Hao, J .; Hatano, M.; Mikami, K. Org. Lett. 2000, 2,
4059-4062. (d) Oi, S.; Terada, E.; Ohuchi, K.; Kato, T.; Tachibana,
Y.; Inoue, Y. J . Org. Chem. 1999, 64, 8660-8667, and ref 12 and 13
therein. (e) Hattori, T.; Suzuki, Y.; Uesugi, O.; Oi, S.; Miyana, S. Chem.
Commun. 2000, 73-74. (f) Ferraris, D.; Young, B.; Dudding, T.; Lectka,
T. J . Am. Chem. Soc. 1998, 120, 4548-4549. (g) Hori, K.; Kodama, H.;
Ohta, T.; Furukawa, I. J . Org. Chem. 1999, 64, 5017-5023. (h) Hori,
K.; Ito, J .; Ohta, T.; Furukawa, I. Tetrahedron 1998, 54, 12737-12744.
(i) Fu¨rstner, A.; Voigtla¨nder, D.; Schrader, W.; Giebel, D.; Reetz, M.
T. Org. Lett. 2001, 3, 417-420.
(1) (a) J acobsen, E. N.; Pfaltz, A.; Yamamoto, H., Eds. Comprehen-
sive Asymmetric Catalysis; Springer-Verlag: Berlin, 1999; pp 912-
1306. (b) Yamamoto, H., Ed. Lewis Acids in Organic Synthesis; Wiley-
VCH: New York, 2000. (c) Noyori, R. Asymmetric Catalysis in Organic
Synthesis; Wiley: New York, 1994. (d) Carmona, D.; Lamata, M. P.;
Oro, L. A. Coord. Chem. Rev. 2000, 200-202, 717-772. (e) J ørgensen,
K. A. Angew. Chem., Int. Ed. 2000, 39, 3558-3588. (f) Dias, L. C. J .
Braz. Chem. Soc. 1997, 8, 289-332.
(2) Cu: (a) J ohnson, J . S.; Evans, D. A. Acc. Chem. Res. 2000, 33,
325-335. (b) J ørgensen, K. A.; J ohanssen, M.; Yao, S.; Audrain, H.;
Thorhauge, J . Acc. Chem. Res. 1999, 32, 605-613.
10.1021/om0109861 CCC: $22.00 © 2002 American Chemical Society
Publication on Web 03/19/2002

1566 Organometallics, Vol. 21, No. 8, 2002
Brunkan et al.
Sch em e 1
the potential to uniquely affect the reactivity and
selectivity of the reactions they catalyze.
One significant difference between [P2M]2+ Lewis
acids and their early metal or first-row counterparts is
the softness of the Pd and Pt metal centers, which
increases their carbophilicity relative to their oxophi-
licity.7 The effects of this enhanced carbophilicity on the
coordination chemistry of P2MII complexes include
preferential binding of enolate ligands via metal-carbon
rather than metal-oxygen bonds.8 The observation of
unorthodox C,O binding modes for binaphthol or bi-
phenanthrol-type ligands in sterically congested Pt and
Pd complexes (in contrast to the O,O′-binding modes
seen in early-transition-metal compounds) has also been
attributed to the greater softness of the group 10
metals.9 With respect to catalysis, the impact of the
carbophilicity of a [P2Pd]2+[A]- Lewis acid catalyst on
2
complexes may be important in defining the mecha-
nisms of [P2M]2+-catalyzed reactions.
the mechanism of imine and aldehyde alkylation by
Mukaiyama-type nucleophiles was demonstrated re-
cently in work by Sodeoka and by Fujimura.10 In polar
solvents at relatively high reaction temperatures, these
reactions were found to proceed through C-bound Pd
enolate intermediates formed by reaction of the catalyst
with the nucleophile, rather than by traditional Lewis
acid activation of the aldehyde or imine substrates via
Pd-O or Pd-N coordination. This behavior is unprec-
edented with conventional early metal Lewis acid
catalysts. However, the same [P2Pd]2+[A]-2 catalyst may
also act as a traditional Lewis acid in analogous
reactions involving more activated imine substrates,
lower reaction temperatures, and less polar solvents, as
demonstrated by Lectka.6f
Clearly, group 10 [P2M]2+ transition-metal complexes
have the potential to exhibit new and beneficial reactiv-
ity, compared to traditional p-block, early-metal, and
Cu(II)/Zn(II) Lewis acids, when employed as catalysts
for organic transformations. Thus far, however, the
mechanisms of [P2M]2+-catalyzed reactions have not
been investigated in sufficient detail to understand or
predict the diverse reactivity offered by these Pd and
Pt Lewis acids. Therefore, we have undertaken a
comprehensive investigation of the Pt(II) Lewis acids
P2Pt(OTf)2 (2; P2 ) dppe (a ), R-BINAP (b)) and their
The lability of ligands coordinated to the metal center
may also differ greatly in complexes involving early- and
late-transition-metal Lewis acids. Square-planar, 16-
electron Pd(II) and especially Pt(II) complexes undergo
ligand exchange reactions much more slowly than
complexes containing substitutionally labile, octahedral
early-metal centers or first-row metals such as Cu(II)
or Zn(II).1b Moreover, ligand exchange almost always
proceeds via associative mechanisms in P2MII com-
plexes,11 even when extremely weakly coordinated
ligands such as σ-bound methane are involved.12 Since
substrate/product ligand exchange is essential to achieve
turnover in a Lewis acid catalyzed reaction, the unique
ligand exchange characteristics peculiar to group 10
activity as catalysts for the Diels-Alder reaction of
acryloyl-N-oxazolidinone (4) with cyclopentadiene (HCp;
eq 1), with the ultimate aim of understanding the
(7) Pearson, R. G. Hard and Soft Acids and Bases; Dowden,
Hutchinson, and Ross: Stroudsburg, PA, 1973.
fundamental principles underlying the mechanism(s)
available to these catalysts. Specific questions addressed
during the research include the following. (1) Do the
reactions proceed via a true Lewis acid mechanism: i.e.,
does the catalyst activate the substrate by coordination
of the dienophile to Pt through the carbonyl groups
(Scheme 1)? (2) Does competitive coordination of coun-
terions, product, or additives inhibit substrate binding
to the catalyst? (3) What is the turnover-limiting step
of the catalytic cycle? (4) How do the stereoelectronic
properties of the diphosphine ligand P2 affect the
behavior of the Lewis acid catalyst?
(8) (a) Veya, P.; Floriani, C.; Chiesi-Villa, A.; Risoli, C. Organome-
tallics 1993, 12, 4899-4907. Examples of late-metal C-bound eno-
lates: (b) Bergman, R. G.; Kaplan, A. W. Organometallics 1997, 16,
1106-1108. (c) Hsu, R. H.; Chen, J . T.; Lee, G. H.; Wang, Y.
Organometallics 1997, 16, 1159-1166. (d) Hamann, B. C.; Hartwig,
J . F. J . Am. Chem. Soc. 1997, 119, 12382-12383. A late-metal O-bound
enolate: (e) Slough, G. A.; Bergman, R. G.; Heathcock, C. H. J . Am.
Chem. Soc. 1989, 111, 938-949.
(9) (a) Brunkan, N. M.; White, P. S.; Gagne´, M. R. J . Am. Chem.
Soc. 1998, 120, 11002-11003. (b) Bergens, S. H.; Leung, P.-H.; Bosnich,
B. Organometallics 1990, 9, 2406-2408. (c) See also: Kocovsky, P.;
Vyskocil, S.; Cisarova, I.; Sejbal, J .; Tislerova, I.; Smrcina, M.; Lloyd-
J ones, G. C.; Stephen, S. C.; Butts, C. P.; Murray, M.; Langer, V. J .
Am. Chem. Soc. 1999, 121, 7714-7715.
(10) (a) Fujii, A.; Hagiwara, E.; Sodeoka, M. J . Am. Chem. Soc. 1999,
121, 5450-5458. (b) Fujii, A.; Sodeoka, M. Tetrahedron Lett. 1999, 40,
8011-8014. (c) Fujimura, O. J . Am. Chem. Soc. 1998, 120, 10032-
10039.
(11) Cross, R. J . Adv. Inorg. Chem. 1989, 34, 219-292.
(12) J ohansson, L.; Tilset, M. J . Am. Chem. Soc. 2001, 123, 739-
740.
Note that questions such as the first three above are
of general interest for many Lewis acid catalyzed
reactions; however, the data needed to answer them are
difficult to obtain for most p-block, early-transition-

The P2Pt(OTf)2-Catalyzed Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1567
metal, or even Cu(II)/Zn(II) Lewis acids.13 For example,
solid-state14 or in situ characterization15 of only a few
catalyst-substrate or catalyst-product complexes has
been reported, while issues such as competitive coordi-
nation of counterions, reaction products, or additives to
the catalyst have rarely been addressed.15a The P2Pt-
(OTf)2 Lewis acids examined here, on the other hand,
are uniquely suited to mechanistic studies because of
their good stability, convenient catalytic rates, and slow
rates of ligand substitution reactions and because of the
fact that they contain spin-active 31P and 195Pt nuclei.
These characteristics enabled direct, in situ observation
of active catalysts, catalyst-substrate and catalyst-
Sch em e 2
2b was examined initially because Ghosh reported that
20 mol % of it, prepared by protonation of [R-BINAP]-
Pt(salicylate) with 2 equiv of triflic acid (HOTf), cata-
lyzed the Diels-Alder reaction (eq 1) in 96-98% ee.6a
Ghosh also found that addition of 2 equiv of water to
the reaction mixture dramatically increased the turn-
over rate (complete conversion in 1 h rather than 20 h
at -40 °C).6a The dppe Lewis acid 2a was of interest
because it is a soluble model for a Lewis acid in a
molecularly imprinted polymer,17 which we hoped to
generate by protonation of a polymer-imprinted ana-
logue of 1a with 2 equiv of HOTf (Scheme 2, vide infra),
as part of our ongoing research into the effect of
associated chiral cavities on molecularly imprinted
catalysts.18
1
product complexes using 31P and H NMR spectroscopy,
made investigation of both the thermodynamic relation-
ships between the complexes and the kinetics of ligand
exchange reactions feasible via 31P NMR experiments,
and even facilitated direct observation of Pt-catalyzed
Diels-Alder reactions at low temperature by 31P and
1H NMR.
The results of our examination of P2Pt(OTf)2 Lewis
acid Diels-Alder catalysts are presented in two con-
secutive papers. This paper describes a detailed inves-
tigation of the binding of catalytically relevant Lewis
bases (counterions, substrate, product, and water) to the
[P2Pt]2+ Lewis acid catalysts. The Lewis acids 2 were
generated in situ, characterized spectroscopically and
crystallographically (P2 ) dppe), and allowed to react
with Lewis bases, forming complexes that were char-
acterized by 31P and 1H NMR at 195 K. The relative
binding strengths of the Lewis bases to [P2Pt]2+ were
determined through competition experiments, the ki-
netics of ligand exchange were explored via simulation
of dynamic 31P NMR spectra, and the relative rates of
ligand exchange vs cycloaddition were revealed when
2-catalyzed Diels-Alder reactions (eq 1) were monitored
by NMR at 195 K. The second paper (immediately
following) describes the last experiments more thor-
oughly, since they revealed that the Pt Lewis acids
undergo an unusual decomposition reaction under ca-
talysis conditions and that phosphine-dependent de-
composition rates actually cause the Diels-Alder reac-
Resu lts
(1) F or m a tion a n d Ch a r a cter iza tion of P 2P t-
(OTf)2 (2). The Pt(II) Lewis acids P2Pt(OTf)2 (2; P2
)
dppe (a ); R-BINAP (b)) were generated in situ from the
yellow air- and water-stable precursors P2Pt(S-BINOL)
(1), which were prepared as previously described.19
Protonation of 1 with 2 equiv of freshly distilled triflic
acid (HOTf) at room temperature in CD2Cl2 under dry
conditions cleanly generated new, colorless, symmetric
P2Pt species that were assigned to 2 on the basis of 1H
and 31P NMR data (Scheme 2; Table 1).20 The 530-540
Hz increase in the P-Pt coupling constant on conversion
of 1 to 2 was diagnostic, as it indicated that BINOL was
replaced by a ligand (triflate) that exerts a poorer trans
influence on phosphorus and is therefore more weakly
coordinated to Pt than BINOL.21 One equivalent of free
1
S-BINOL was observed by H NMR.
tion to proceed via different mechanisms when P2
)
When the 1a protonation reaction was carried out in
dry chlorobenzene rather than CD2Cl2, colorless single
crystals of 2a were obtained from the solution upon
standing overnight. X-ray crystallographic analysis
yielded the structure shown in Figure 1. Rigorously dry
conditions were necessary to isolate 2a , since the bis-
dppe vs R-BINAP.16 Note that the R-BINAP catalyst
(13) Notable exceptions: (a) J acquith, J . B.; Levy, C. J .; Bondar, G.
V.; Wang, S.; Collins, S. Organometallics 1998, 17, 914-925. (b) Lin,
S.; Bondar, G. V.; Levy, C. J .; Collins, S. J . Org. Chem. 1998, 63, 1885-
1892. (c) Bonnesen, P. V.; Puckett, C. L.; Honeychuck, R. V.; Hersh,
W. H. J . Am. Chem. Soc. 1989, 111, 6070-6081.
(aqua) complex [(dppe)Pt(OH2)2]2+[OTf]- (3a ) is also
(14) Crystallographic characterization of catalyst-substrate com-
plexes: (a) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Angew.
Chem., Int. Ed. Engl. 1990, 29, 256-272. (b) Gothelf, K.; Hazell, R.
G.; J ørgensen, K. A. J . Am. Chem. Soc. 1995, 117, 4435-4436. (c)
Evans, D. A.; Rovis, T.; Kozlowski, M. C.; Downey, W.; Tedrow, J . S.
J . Am. Chem. Soc. 2000, 122, 9134-9142. See also ref 3d,e.
(15) In situ investigation (by 1H NMR) of the interactions of
substrates with p-block catalysts such as BF3, SnCl4, and MeAlCl2:
(a) Hunt, I. R.; Rogers, C.; Woo, S.; Rauk, A.; Keay, B. R. J . Am. Chem.
Soc. 1995, 117, 1049-1056. (b) Denmark, S. E.; Almstead, N. G. J .
Am. Chem. Soc. 1993, 115, 3133-3139. (c) Denmark, S. E.; Henke, B.
R.; Weber, E. J . Am. Chem. Soc. 1987, 109, 2512-2514. (d) Castellino,
S. J . Org. Chem. 1990, 55, 5197-5200. (e) Castellino, S.; Dwight, W.
J . J . Am. Chem. Soc. 1993, 115, 2986-2987. (f) Childs, R. F.;
Mulholland, D. L.; Nixon, A. Can. J . Chem. 1982, 60, 801-808. (g)
Hawkins, J . M.; Loren, S.; Nambu, M. J . Am. Chem. Soc. 1994, 116,
1657-1660. (h) Gajewski, J . J .; Ngernmeesri, P. Org. Lett. 2000, 2,
2813-2815. (i) Corey, E. J .; Loh, T.-P.; Sarshar, S.; Azimioara, M.
Tetrahedron Lett. 1992, 33, 6945-6948. (j) Cardillo, G.; Gentilucci, L.;
Gianotti, M.; Tolomelli, A. Org. Lett. 2001, 3, 1165-1167.
2
crystalline and very stable (Scheme 3).22 Only one other
P2Pt(OTf)2 species, [1,2-C6H4(PMePh)2]Pt(OTf)2, has
(17) For reviews of molecular imprinting see: (a) Molecularly
Imprinted Polymers: Man-made Mimics of Anitbodies and Their
Applications in Analytical Chemistry; Sellergren, B., Ed.; Elsevier:
Amsterdam, 2001. (b) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995,
34, 1812-1832.
(18) Brunkan, N. M.; Gagne´, M. R. J . Am. Chem. Soc. 2000, 122,
6217-6225.
(19) Brunkan, N. M.; White, P. S.; Gagne´, M. R. Angew. Chem., Int.
Ed. 1998, 37, 1579-1582.
(20) See the Supporting Information for the 1H NMR spectrum.
(21) (a) Pidcock, A.; Richards, R. E.; Venanzi, L. M. J . Chem. Soc. A
1966, 1701-1710. (b) Appleton, T. G.; Bennett, M. A. Inorg. Chem.
1978, 17, 738-747.
(22) (a) Fallis, S.; Anderson, G. K.; Rath, N. P. Organometallics 1991,
10, 3180-3184. (b) Gorla, F.; Venanzi, L. M. Helv. Chim. Acta 1990,
73, 690-697.
(16) See: Brunkan, N. M.; Gagne´, M. R. Organometallics 2002, 21,
1576-1582.

1568 Organometallics, Vol. 21, No. 8, 2002
Ta ble 1. 31P NMR Da ta for P 2P tII Com p lexes in CD2Cl2
Brunkan et al.
a
Pt complex
temp (K)
δ (ppm)
1J P-Pt (Hz)
2J P-P (Hz)
(dppe)Pt(S-BINOL) (1a )
(R-BINAP)Pt(S-BINOL) (1b)
(dppe)Pt(OTf)2 (2a )
room temp
room temp
195
195
195
27.2
5.2
40.4
1.6
37.7
3.5
39.9
35.9
3.8
3640
3670
4170
4210
3950
4050
4080
4000
4200
3990
4080
3990
4270
4050
4110
4020
4210
4020
4210
4020
(R-BINAP)Pt(OTf)2 (2b)
[(dppe)Pt(OH2)2]2+[OTf-]2 (3a )
[(R-BINAP)Pt(OH2)2]2+[OTf-]2 (3b)
[(dppe)Pt(4)]2+[OTf-]2 (6a )
195
195
9.7
27.4
9.7
[(R-BINAP)Pt(4)]2+[OTf-]2 (6b)
[(dppe)Pt(4)]2+[BF4-]2 (7a )
195
195
195
195
195
195
2.9
39.9
36.0
4.0
[(R-BINAP)Pt(4)]2+[BF4-]2 (7b)
[(dppe)Pt(2S-5)]2+[OTf-]2 (8a )
27.4
10.6
27.5
27.7
2.6
39.5
36.4
3.6
2.8
3.3
[(R-BINAP)Pt(2S-5)]2+[OTf-]2 (R,2S-8b)
[(R-BINAP)Pt(2R-5)]2+[OTf-]2 (R,2R-8b)
3.0
a
1J P-Pt values determined from spectra acquired at 195 K have (30 Hz uncertainty, as Pt satellites were very broad at this temperature.
Sch em e 3
(2) F or m a t ion a n d Ch a r a ct er iza t ion of [P 2P t -
(L2)]2+[OTf]-2. Addition of neutral, Lewis basic ligands
L2 (L2 ) 2 H2O, acryloyl-N-oxazolidinone (4), Diels-
Alder adduct (5)) to P2Pt(OTf)2 in CD2Cl2 solution at
room temperature generated the Lewis acid-Lewis base
adducts [P2Pt(L2)]2+[OTf]- (Scheme 3), which were
2
1
characterized by 31P and H NMR at 195 K (Table 1).20
In the case of water, adding 2 equiv quantitatively
converted 2 to 3; however, addition of up to 10 equiv of
4 or 5 to 2 produced an equilibrium mixture of 2 and
the Lewis acid-Lewis base complex 6 or 8. At 195 K,
OTf-/L2 exchange was slow on the NMR time scale, and
F igu r e 1. ORTEP drawing of (dppe)Pt(OTf)2 (2a ). Selected
bond lengths (Å): Pt-P1 ) 2.2158(14), Pt-P2 ) 2.2151-
(14), Pt-O1 ) 2.120(4), Pt-O2 ) 2.138(4). Selected bond
angles (deg): P1-Pt-P2 ) 85.74(5), O1-Pt-O4 ) 87.51-
(15), P1-Pt-O1 ) 98.69(11), P2-Pt-O4 ) 87.98(11).
separate sharp resonances for the [P2Pt(L2)]2+[OTf]-
2
and residual P2Pt(OTf)2 species were observed; at room
temperature, rapid OTf-/L2 exchange produced a single
broad resonance in the 31P NMR spectra of some of the
reaction mixtures (vide infra).
been crystallographically characterized;23 bond lengths
and angles for 2a (Figure 1) are similar (within 0.01 Å
and 7°, respectively) to those of the 1,2-C6H4(PMePh)2
complex. Although a few crystals of 2a were isolated
for X-ray analysis, attempts to prepare macroscopic
quantities of 2 were unsuccessful, as the isolated
products always contained the diaqua complex 3. Thus,
2 was prepared in situ from 1 and used immediately.
The P-Pt coupling constant of the symmetric diaqua
complex 3 was found to be ∼200 Hz smaller than that
of 2, indicating that water coordinates more strongly
to Pt than does triflate (Table 1).21 In the case of 3a ,
the Pt-coordinated water resonates at 7.28 ppm in the
1H NMR spectrum.20 The white air- and water-stable
dppe complex 3a was isolated as described in the
literature;22 however, analogous attempts to isolate 3b
2+
cleanly yielded the µ-OH dimer [(R-BINAP)Pt(µ-OH)]2
-
[OTf]- instead of the desired diaqua species.24,25
(23) Appelt, A.; Ariaratnam, V.; Willis, A. C.; Wild, S. B. Tetrahe-
2
1
dron: Asymmetry 1990, 1, 9-12. Also note that the J P-Pt value for
Pt-oxazolidinone complexes 6 were generated by
adding 2 equiv of HOTf to 1 (forming 2 in situ) in the
[1,2-C6H4(PMePh)2]Pt(OTf)2 is 4034 Hz, 140 Hz smaller than that of
2a .

The P2Pt(OTf)2-Catalyzed Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1569
Sch em e 4
presence of 2-10 equiv of dienophile 4. 31P NMR spectra
of 6 consisted of two doublets with Pt satellites, with
P-Pt coupling constants slightly smaller (10-220 Hz)
than those of the corresponding P2Pt(OTf)2 species
(Table 1). The loss of symmetry in 6 and the magnitudes
of the P-Pt coupling constants are consistent with the
expected (and productive) bidentate coordination of 4
to Pt through its carbonyl oxygens (Scheme 3). Reso-
nances for coordinated 4 were also well-separated from
F igu r e 2. 31P NMR spectrum (400 MHz) of an experiment
in which 4 and 2S-5 competed for coordination to 2a ,
yielding an equilibrium mixture of 6a and 8a (entry 5,
Table 2). 3a is an impurity, and • denotes unidentified
asymmetric minor products of the reaction of 2a with 2S-
5. Platinum satellites are not shown.
1
free 4 in the H NMR spectrum.20 Similar Pt-oxazoli-
(3) Mea su r em en t of Equ ilibr iu m Con sta n ts for
dinone complexes with BF4- rather than OTf- counter-
ions were obtained by protonation of 1 with 2 equiv of
HBF4‚OMe2 in the presence of 5 equiv of 4 (Scheme 4).
Although protonation of 1 with HBF4‚OMe2 in the
absence of 4 generates a mixture of symmetric and
asymmetric species by 31P NMR,26 [P2Pt(4)]2+[BF4]-2 (7)
formed quantitatively when 4 was present, producing
NMR spectra almost identical with those of 6 (Table 1).
Generating 2 in the presence of 1.5-2 equiv of
enantioenriched Diels-Alder adduct 2S-5 (95% endo,
th e Rela tive Bin d in g of Lew is Ba ses to [P 2P t]2+
.
The relative abilities of the OTf- and BF4- counterions
and neutral ligands L2 to coordinate to the [P2Pt]2+
Lewis acids were quantified by measuring equilibrium
constants (Keq) for the reactions in Table 2. The data in
entries 1-3 were derived from the experiments de-
scribed in section 2 as follows: Keq values for the 2/6
(OTf-/4) equilibria in entry 1 were calculated via
integration of the 31P NMR resonances for Pt-oxazoli-
dinone complex 6 and residual 2 in spectra obtained
when 2 was generated in the presence of 2-10 equiv of
4. The data in entry 2 represent a lower limit of Keq for
the 2/3 (OTf-/OH2) equilibrium, since addition of only
2 equiv of water to 2 completely converted it to diaqua
complex 3. Similarly, entry 3 represents a lower limit
of Keq for the “P2Pt(BF4)2”/7 (BF4-/4) equilibrium, as
addition of 5 equiv of 4 to the Lewis acid generated by
protonation of 1 with HBF4‚OMe2 quantitatively pro-
duced 7.
92% ee) yielded Pt-endo-5 complexes [P2Pt(2S-5)]2+
-
[OTf]-2 (8a and R,2S-8b), with 31P NMR spectra similar
to those of the Pt-oxazolidinone complexes 6, as the
major reaction products (∼90%; Scheme 3). Additionally,
four to five minor asymmetric species that could not be
identified were observed by 31P NMR in the dppe
reaction (see Figure 2), while a single minor species was
present in the R-BINAP case. Using racemic rather than
enantioenriched Diels-Alder adduct 5 (90% endo) for
the R-BINAP reaction produced the same major species,
which was now accompanied by a different minor
species (35%). This new species was assigned to the
diastereomer [(R-BINAP)Pt(2R-5)]2+[OTf]-2 (R,2R-8b),
formed when the 2R rather than the 2S enantiomer of
5 coordinates to the chiral [(R-BINAP)Pt]2+ fragment.27
Resonances for free 5, coordinated 2S-5, and coordinated
Thermodynamic relationships between the cationic
complexes [P2Pt(L2)]2+[OTf]- were explored through
2
experiments in which two different ligands L2 competed
for binding to 2. For example, the abilities of 4 and 5 to
compete with water for coordination to Pt were tested
by exposing isolated dppe diaqua complex 3a to 10 equiv
of either 4 or rac-5 in CD2Cl2 solution. 195 K 31P NMR
spectra of these reactions showed no trace of 6a or 8a ,
allowing calculation of the upper limit of Keq for the 3/6
(H2O/4) and 3/8 (H2O/5) equilibria in entry 4 (Table 2).
In the R-BINAP case, nonisolable diaqua complex 3b
was generated in situ by titration of 2b with a stock
solution of water in CH2Cl2 until 98-99% pure 3b was
attained (by 31P NMR), and then 10 equiv of 4 or 5 was
added to the reaction mixture. Initially, 10 equiv of 4
appeared to convert 10% of 3b to 6b; however, adding
10 equiv more of 4 to the reaction did not increase the
1
2R-5 were distinguishable in the H NMR spectrum.20
In catalytic reactions, (R-BINAP)Pt(OTf)2 produces 2S-5
as the major product (96-98% ee).6a
(24) The complex {[R-BINAP]Pt(OH2)2]2+ (3b) has not been reported,
although the Pd analogue and the monoaqua complex species {[R-
BINAP]Pt(OH2)(OTf)}+[OTf]- are known: (a) Fujii, A.; Hagiwara, E.;
Sodeoka, M. J . Am. Chem. Soc. 1999, 121, 5450-5458. (b) Fujii, A.;
Sodeoka, M. Tetrahedron Lett. 1999, 40, 8011-8014. (c) Stang, P. J .;
Olenyuk, B.; Arif, A. M. Organometallics 1995, 14, 5281-5289.
(25) Complexes of the type [P2Pt(µ-OH)]22+ are well-known; see for
example: (a) Bandini, A. L.; Banditelli, G.; Demartin, F.; Manassero,
M.; Minghetti, G. Gazz. Chim. Ital. 1993, 123, 417-423. (b) Li, J . J .;
Li, W.; Sharp, P. R. Inorg. Chem. 1996, 35, 604-613.
(26) The symmetric (1J P-Pt ≈ 4100 Hz) and one of the asymmetric
products appear to be di- and mono-Me2O-coordinated [P2Pt]2+ com-
plexes (two OMe2 peaks by 1H NMR). A second asymmetric species,
the Pt2(µ-BINOL) dimer A, was also observed; it forms whenever less
than 2 equiv of acid (HOTf or HBF4‚O(CH3)2) is added to 1a . See ref
16 and a manuscript in progress for details.
(27) Note that when rac-5 (90% endo) was used for the dppe
experiment, an additional species (1:1 ratio compared to the original
major species) whose 31P NMR spectrum was almost coincident with
that of 8a was observed, even though the dppe ligand is achiral. Two
sets of resonances due to Pt-coordinated 5 were seen by 1H NMR. At
present, these data cannot be satisfactorily explained.

1570 Organometallics, Vol. 21, No. 8, 2002
Ta ble 2. Equ ilibr iu m Con sta n ts for Rela tive Liga n d Coor d in a tion to [P 2P t]2+
Brunkan et al.
a
b
Average over 5-7 experiments with 1.5-10 equiv of 4. Calculated assuming a detection limit of 1% for minor species by 31P NMR.
c All major and minor resonances of 8 were included in the calculation of Keq, while residual 2 (<3%) was not. Experiments with rac- and
2S-5 gave the same Keq value. Two experiments gave the same Keq value.
d
amount of 6b present. Ambiguous results were also
observed when 3b was treated with rac-5; no 8b was
formed, but unidentified resonances appeared upfield
of 3b and grew disproportionately large when more
rac-5 was added to the solution.
To directly compare the abilities of dienophile 4
(reaction substrate) and Diels-Alder adduct 5 (reaction
product) to bind to the Pt Lewis acids, 2 was generated
in the presence of both 10 equiv of 4 and 1.5-2 equiv of
2S-5. Figure 2 shows the 31P NMR spectrum afforded
by the dppe version of this experiment; Keq for the 6/8
(4/5) equilibrium in entry 5 of Table 2 was calculated
from the spectral integration data. Keq for the R-BINAP
case was measured similarly.
Finally, the dppe and R-BINAP Lewis acids P2Pt-
(OTf)2 (2a and 2b, respectively) were allowed to compete
F igu r e 3. 31P NMR spectrum (400 MHz) of an experiment
directly for binding to dienophile or Diels-Alder adduct.
in which 2a and 2b competed for coordination to 4, yielding
When 2 equiv each of Lewis acids 2a and 2b was
an equilibrium mixture of 2a , 2b, 6a , and 6b (entry 1,
Table 3). Platinum satellites are not shown.
generated in the presence of 1 equiv of 4, no free 4 was
observed by H NMR, while 2a , 2b, 6a , and 6b were
1
visible in the 31P NMR spectrum of the equilibrium
reaction mixture (Figure 3), affording the Keq value in
entry 1 of Table 3. An analogous competition experiment
conducted with 2S-5 instead of 4 gave Keq for entry 2.
To better contrast the relative 4/5 binding preferences
of the two [P2Pt]2+ Lewis acids, the equilibrium in entry
3 (Table 3) was calculated in two ways: (1) by addition
of the reverse of entry 1 to entry 2 or (2) by addition of
the reverse of the R-BINAP 6/8 equilibrium in entry 5
of Table 2 to the dppe 6/8 equilibrium in entry 5 of Table
2. Two Keq values for the new equilibrium were calcu-
lated from the two sets of data, providing a convenient
check of the internal consistency of the thermodynamic
data in Tables 2 and 3.

The P2Pt(OTf)2-Catalyzed Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1571
Ta ble 3. Equ ilibr iu m Con sta n ts for 2a /2b Com p etition Exp er im en ts
a
b
Single experiment. Calculated from entries 1 and 2, Table 3: 2.8/1.6 ) 1.8. c Calculated from entry 5, Table 2: 20/2.5 ) 9.2.
Ta ble 4. Liga n d Exch a n ge Ra tesa for P 2P tII
Sch em e 5
Com p lexes in CD2Cl2 a t 270 K
dppe
R-BINAP
k1
k2
8400(300)
26(1)
47(1)
2(0.2)
a
Uncertainties in k1 and k2, given in parentheses, were
estimated by noting the range of values that produced visually
indistinguishable simulated 31P NMR spectra.
complexes. Therefore, exchange rates for the P2PtII
complexes are reported and compared at this temper-
ature (Table 4).29
Attempts to measure the rates of other ligand ex-
change processes using line shape analysis were not
successful. For example, broadened spectra of a 2/6/8
equilibrium mixture (Table 2, entry 5) proved impossible
to simulate because too many types of exchange (dieno-
phile/dienophile, cycloadduct/cycloadduct, dienophile/
cycloadduct, OTf-/dienophile, and OTf-/cycloadduct)
were occurring at once, while BF4-/dienophile and H2O/
dienophile exchange rates were inaccessible because the
equilibrium strongly favored one of the complexes (Table
2, entries 3 and 4).
(4) Mea su r em en t of Liga n d Exch a n ge Ra tes via
Dyn a m ic 31P NMR. The rates of dienophile/dienophile
and OTf-/dienophile exchange were examined as a
function of temperature via line shape analysis of
broadened 31P NMR spectra of a 2/6 equilibrium mixture
(Table 2, entry 1). The spectra were acquired at tem-
peratures between 190 and 320 K and were simulated
using the program DNMR3 in conjunction with Spin-
Works, yielding the rates (k, s-1) of ligand exchange at
each temperature.28 Since two different exchange pro-
cesses were apparent from the 31P spectra, two different
rates (k1 and k2) were varied to fit the experimental
data. In the first type of exchange (k1), which occurred
at lower temperatures, the two doublets of Pt-oxazo-
lidinone complex 6 broadened and eventually coalesced
into a single resonance, indicating mutual exchange of
the nonequivalent phosphorus nuclei PA and PB in 6
(Scheme 5). The second type of exchange (k2) was
marked by broadening and coalescence of the reso-
nances for 2 and 6, indicating that OTf-/4 ligand
substitution occurs (Scheme 5). Although k1 and k2 were
determined over a range of temperatures in both the
dppe and R-BINAP systems, 270 K was the only
temperature at which well-defined values for both k1
and k2 could be obtained for both the R-BINAP and dppe
Discu ssion
(1) In Situ Ch a r a cter iza tion of Lew is Acid -
Lew is Ba se Com p lexes. A prerequisite to examination
of the mechanism of any transition-metal-catalyzed
organic transformation is the ability to observe, in situ,
the organometallic species present during the reaction.
As pointed out in the Introduction, a combination of
fortuitous properties make the P2PtII Lewis acid-Lewis
base complexes formed during the P2Pt(OTf)2-catalyzed
Diels-Alder reaction (eq 1) amenable to characteriza-
1
tion by 31P and H NMR spectroscopy at 195 K. Using
NMR methods, the active catalysts 2, as well as
catalyst-substrate complexes 6, catalyst-product com-
plexes 8, and even complexes (3) of the Pt Lewis acids
(28) SpinWorks Version 1.2 is copyright 1999-2001 by Kirk Marat,
The University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2.
The original authors of DNMR3 were G. Binsch and D. Kleier of the
University of Notre Dame, South Bend, IN 46556.
(29) See the Supporting Information for plots of ln k1 and ln k2 vs
1/T (K) for the dppe and R-BINAP complexes.

1572 Organometallics, Vol. 21, No. 8, 2002
Brunkan et al.
with water (a common impurity or additive in these
reactions) could be observed as stable, nonfluxional
species in CD2Cl2 solution at low temperature. 31P NMR
spectra were particularly informative, as they indicated
that both dienophile 4 and Diels-Alder adduct 5
coordinate to [P2Pt]2+ in a bidentate fashion via their
carbonyl group oxygens, producing nonsymmetric che-
late complexes (6 and 8) with large P-Pt coupling
constants, consistent with formation of weak Pt-OdC
bonds (Scheme 3, Table 1). Observation of this “tradi-
tional” coordination mode for 4 and 5 strongly suggests
that the P2Pt(OTf)2 Lewis acids do catalyze the Diels-
Alder reaction by a typical Lewis acid mechanism
(Scheme 1). Thus assured that complexes 2, 6, and 8
(and sometimes 3) are stable and present in solution
during 2-catalyzed Diels-Alder reactions, the thermo-
dynamic (section 2 below) and kinetic relationships
(section 3 below) between these complexes were inves-
tigated to further illuminate the details of the catalytic
reaction mechanism.
the substrate to the catalyst may also affect the ef-
ficiency of Lewis acid catalyzed reactions. For example,
additives such as water, alcohols, and amines have been
shown to affect the turnover rates of some Lewis acid
catalyzed reactions.31 With the R-BINAP Lewis acid 2b,
Ghosh observed significant acceleration of Pt-catalyzed
Diels-Alder reactions when 2 equiv of water was added
to the catalyst (from 20 to 1 h for 100% conversion, 20
mol % catalyst).6a However, the experiments described
here clearly show that water binds more strongly to [P2-
Pt]2+ than any of the other Lewis bases examined. This
conclusion is supported by the following facts. (1) The
P-Pt coupling constants of diaqua complexes 3 are
smaller than those of any other P2PtII species, indicating
that water exerts the strongest trans influence on
phosphorus of any of the oxygen-based ligands studied
herein (Table 1).21 (2) Two equivalents of water dis-
placed OTf- from P2Pt(OTf)2 quantitatively (Table 2,
entry 2), demonstrating that water binds much more
strongly to Pt than triflate. (3) Neither dienophile 4 nor
cycloadduct 5 was able to substantially displace Pt-
coordinated water from 3 (Table 2, entry 4), although
the observation of species other than 3b in the R-BINAP
experiments may indicate that displacement of water
by 4 or 5 is more thermodynamically favorable in this
system than in the dppe case. Thus, we predicted (and
observed),32 on the basis of thermodynamic consider-
ations, that water should act as a competitive inhibitor
during Diels-Alder catalysis, rather than accelerating
catalysis as observed by Ghosh.
(2) Th er m od yn a m ic Rela tion sh ip s: Com p etitive
Bin d in g of Lew is Ba ses to [P 2P t]2+. Dien op h ile (4)
vs Cou n ter ion (OTf- or BF 4-). The ability of a Lewis
acid to coordinate and activate Lewis basic substrates
is fundamentally important to successful Lewis acid
catalysis. The favorable triflate/dienophile equilibrium
constant (Table 2, entry 1) observed when 4 was added
to P2Pt(OTf)2 (2) demonstrates a thermodynamic bias
for formation of the complex (6) necessary for dienophile
activation. Although triflate is generally considered to
be the “most coordinating” of the weakly coordinating
counterions, coordination of OTf- to Pt evidently does
not significantly inhibit dienophile binding to the Lewis
acid. Nevertheless, experiments performed by Ghosh
and in our own laboratory have shown that “P2Pt(BF4)2”
Lewis acids are more active than OTf- catalysts for the
Diels-Alder reaction in eq 1;6a,30 therefore, the effect
of the counterion on dienophile binding to [P2Pt]2+ was
also examined (Table 2, entry 3). As expected, displace-
ment of BF4- by 4 is more thermodynamically favorable
than displacement of OTf- (Keq >1100 for BF4- vs 60-
90 for OTf-), consistent with the conventional wisdom
Dien op h ile (4) vs Diels-Ald er Ad d u ct (5). During
Lewis acid catalyzed reactions, competitive coordination
of reaction product to the catalyst (product inhibition)
may significantly inhibit catalytic turnover, making
high catalyst loading necessary to achieve satisfactory
conversion. To assess the potential for product inhibition
in P2Pt(OTf)2-catalyzed Diels-Alder reactions, competi-
tion experiments were conducted to see whether sub-
strate 4 or product 5 binds more strongly to [P2Pt]2+
.
These reactions showed that both the dppe and R-
BINAP Lewis acids exhibit small thermodynamic pref-
erences for coordination to 2S-5 rather than 4 (Table 2,
entry 5). The fact that [(dppe)Pt]2+ prefers 2S-5 over 4
more strongly than does [(R-BINAP)Pt]2+ (Keq ) 20 vs
2.5) suggests that steric factors may be more important
in determining the relative binding of Lewis bases to
2b than they are in the case of 2a . For example,
catalyst-product complex R,2S-8b may be destabilized
relative to catalyst-dienophile complex 6b in the R-
BINAP system due to steric interactions of the bulky
R-BINAP ligand with 2S-5, which are not as severe with
that BF4 is “less coordinating” than OTf-.
-
Comparison of Keq values for the dppe and R-BINAP
equilibria in entry 1 of Table 2 reveals that Keq is only
slightly larger in the R-BINAP case than in the dppe
case (90 vs 60), despite significant steric and electronic
differences between the two diphosphines. The larger,
less basic, triaryl-substituted phosphine R-BINAP should
generate an [(R-BINAP)Pt]2+ fragment that is bulkier
but significantly more electrophilic than the [(dppe)Pt]2+
Lewis acid. These properties could cause steric inhibi-
tion of ligand binding and associative ligand exchange
processes in the R-BINAP Lewis acid or they might
cause it to bind Lewis bases more strongly (potentially
activating coordinated dienophiles more strongly to
attack by dienes). Nevertheless, these differences are
not manifested in the similar binding preferences of the
dppe and R-BINAP [P2Pt]2+ catalysts for 4 over OTf-.
(31) (a) Koh, J . H.; Larsen, A. O.; Gagne´, M. R. Org. Lett. 2001, 3,
1233-1236. (b) Vogl, E. M.; Gro¨ger, H.; Shibasaki, M. Angew. Chem.,
Int. Ed. Engl. 1999, 38, 1570-1577.
(32) In fact, our own Diels-Alder catalysis studies, conducted as
reported in the following paper using catalysts prepared in situ from
1, are consistent with water acting as a poison in the present system,
in contrast to Ghosh’s results. For example, at -40 °C the addition of
1, 2, and 4 equiv of H2O to (R-BINAP)Pt(OTf)2 led to decreases in
product conversion (eq 1, 140 min) from 99% to 96, 73, and 47%,
respectively (however, the ee was 96-98% in all cases). Although we
have no satisfactory explanation for the discrepancy between these
observations and Ghosh’s, we note that reactivity differences do occur,
Wa ter vs Dien op h ile (4) or Diels-Ald er Ad d u ct
(5). Competitive coordination of Lewis bases other than
depending on how the [P2Pt]2+ catalyst was generated (e.g. P2PtCl2
+
AgOTf vs P2Pt(BINOL) + HOTf). Apparently the Ghosh methodology
of protonating the salicylate accesses a highly selective, albeit unchar-
acterized, catalyst that is subject to strong rate accelerations in the
presence of water.
(30) (a) Becker, J . J .; Gagne´, M. R. Unpublished results. (b) Evans,
D. A.; Murray, J . A.; Matt, P. V.; Norcross, R. D.; Miller, S. J . Angew.
Chem., Int. Ed. Engl. 1995, 34, 798-800.

The P2Pt(OTf)2-Catalyzed Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1573
Sch em e 6
the smaller, flatter ligand 4. These steric differences
likely matter less in the less bulky dppe system, making
6a and 8a closer in energy than 6b and R,2S-8b.
Superimposed on the steric differences between dieno-
phile and product are their electronic differences. Treat-
ing the imide as an ester leads to the prediction that
2S-5 should be more basic than dienophile 4, since the
carbonyl group of an R,â-unsaturated ester is less basic
than that of the saturated analogue (although the
opposite is true for aldehydes and ketones).33 Thus, we
infer that the binding constant for P2 ) dppe more
closely reflects the inherent basicity of the competing
ligands, while the binding constant for P2 ) R-BINAP
is attenuated by the opposing effects of basicity and the
increased congestion of the product/BINAP complex.
Moreover, the difference in the dppe and R-BINAP Keq
values also indicates that product inhibition should be
more severe in 2a -catalyzed than in 2b-catalyzed Diels-
Alder reactions.
A more direct way to compare the 4/2S-5 binding
preferences of the dppe and R-BINAP Lewis acids,
particularly with respect to the potential for product
inhibition, was provided by experiments in which 2a
and 2b competed for binding to 4 or to 2S-5 (Table 3,
entries 1-2). The small Keq value of 2.8 for the equi-
librium in entry 1 reflects a slight thermodynamic
preference of R-BINAP Lewis acid 2b, relative to dppe
Lewis acid 2a , for binding 4 over OTf-. The smaller Keq
value (1.6) for entry 2 shows that both catalysts bind 5
over OTf-, with similarly favorable thermodynamics.
Taken together, these results again indicate that the
ground-state energy difference between dppe complexes
6a and 8a is greater than the corresponding difference
between R-BINAP complexes 6b and R,2S-8b (Keq > 1
in entry 3, Table 3); therefore, product inhibition should
be greater in 2a -catalyzed Diels-Alder reactions. As a
check of the internal consistency of the Keq data, Keq
for the equilibrium in entry 3 was also calculated from
the equilibrium constants for the dppe and R-BINAP
4/5 equilibria in entry 5 of Table 2. Within experimental
error, this Keq value matched that calculated from
entries 1 and 2 of Table 3 (1.8 ( 1.2 vs 9.2 ( 6.1).
Su m m a r y of th e Rela tive Bin d in g Str en gth s of
Lew is Ba ses to [P 2P t]2+. A scale of relative binding
strengths for the catalytically important Lewis bases
4, 5, H2O, OTf-, and BF4- to Diels-Alder catalysts [P2-
the competing diastereomeric Diels-Alder transition
states must be significantly more different in energy
(k2S/k2R ) 98/2, ∆∆Gq
) 1.7 kcal mol-1 at -50 °C)
2S,2R
to produce 2S-5 in the observed 96% ee (Scheme 6). This
difference between the diastereomeric ground and tran-
sition states is likely due to stereoelectronically imposed
orbital overlap between the incoming diene and the
activating carbonyl group of the Pt-coordinated dieno-
phile in the transition state, an interaction which
rigidifies this structure and creates a conformational
restriction not present in 8b. The additional rotational
degree of freedom in the coordinated product relaxes the
diastereomer-discriminating steric interactions between
R-BINAP and 2S- or 2R-5 and levels the ground-state
energies of R,2S- and R,2R-8b (Scheme 6).
(3) Kin etic Rela tion sh ip s: Liga n d Exch a n ge in
Lew is Acid -Lew is Ba se Com p lexes. Since ligand
substitution reactions are essential to achieve both
substrate activation and product release (turnover) in
a catalytic cycle, the rates of ligand exchange for the
[P2Pt]2+ Lewis acid-Lewis base complexes were inves-
tigated by dynamic 31P NMR spectroscopy. Simulation
of exchange-broadened spectra of a P2Pt(OTf)2/[P2Pt-
(4)]2+ (2/6) equilibrium mixture (entry 1, Table 2)
afforded the rates of two different exchange processes
(k1 and k2) at 270 K for both the dppe and R-BINAP
complexes (Table 4). As described in Results, the first
exchange process (rate ) k1) equilibrates the non-
equivalent phosphorus nuclei PA and PB of Pt-oxazo-
lidinone complex 6, while the second type of exchange
(rate ) k2) corresponds to OTf-/dienophile ligand sub-
stitution (Scheme 5). Possible mechanisms for the
former exchange process include attack of either OTf-
or a second molecule of 4 on the Pt center of 6, followed
by pseudorotation of the five-coordinate intermediate
and release of the nucleophile.34 Such a dynamic process
does not result in net displacement of the dienophile
from the catalyst (i.e., does not affect turnover during
catalysis). This exchange occurs more quickly than the
OTf-/4 exchange process measured by k2 in both the
Pt]2+ at 195 K was assembled from the thermodynamic
-
data described above. For dppe, the scale is BF4
,
OTf- < 4 < 5 , H2O, which also correlates with the
P-Pt coupling constants observed for complexes of the
Lewis bases with [P2Pt]2+ 1J P-Pt ) 4170 (OTf (2a )) >
:
4080/4000 (4 (6a )) g 4200/3950 (5 (8a )) > 3950 (H2O
(3a )) (Table 1; see also Scheme 7, vide infra). As
expected, the weakest ligands exert the weakest trans
influence on the diphosphine ligand.21 A similar binding
scale constructed for the R-BINAP Lewis acid includes
the differential binding of cycloadduct enantiomers 2S-
-
and 2R-5: BF4 , OTf- < 4 < 2R-5 < 2S-5 , H2O.
Although the major cycloadduct enantiomer produced
by [(R-BINAP)Pt]2+ Lewis acid catalysis, 2S-5, does
coordinate more strongly to the Lewis acid than the
minor enantiomer 2R-5 (65:35 thermodynamic prefer-
ence, ∆G2S,2R ) 0.24 kcal mol-1 at -78 °C), note that
(34) The putative 5-coordinate intermediate obtained from OTf-
addition to 6 is also a reasonable intermediate in the OTf-/4 exchange
process.
(33) For instructive references, see ref 15a,b.

1574 Organometallics, Vol. 21, No. 8, 2002
Brunkan et al.
Sch em e 7
investigate which scenario applies to the 2-catalyzed
Diels-Alder reactions in question, 5 equiv of HCp was
added to Pt-dienophile complexes 6 at 195 K, and the
1
reaction to give 8 was monitored by 31P and H NMR
at this temperature. These NMR spectra showed that
all of the Pt-coordinated 4 present was immediately
converted to Pt-coordinated 5 on addition of diene; no
trace of 6 was visible by NMR at any time after HCp
addition in both the dppe and R-BINAP systems.35 Most
important, no turnover was observed in the R-BINAP
reaction at this temperature, confirming that cycload-
dition (kDA) is faster than ligand exchange (k1 and k2),
making the latter the turnover-limiting step of the
catalytic cycle (path A, Scheme 6). Consistent with this
interpretation, the turnover frequency of 2b-catalyzed
Diels-Alder reactions (reported in the following paper)
are smaller than and within an order of magnitude of
the OTf-/4 exchange rates k2 determined here (Table
4).16
Because ligand exchange rather than cycloaddition is
the turnover-limiting step of the P2Pt(OTf)2-catalyzed
Diels-Alder reaction, the organometallic species present
in the reaction solution (2, 6, and 8) never reach
thermodynamic equilibrium under catalysis conditions.
Instead, 6 is quickly converted to 8 by the cycloaddition
step as soon as it forms via the ligand substitution step;
thus, only small, steady-state concentrations of 6 are
expected to be present in solution; the catalyst resting
state will be 8. Under these conditions, the thermody-
namic relationships between 2, 6, and 8 probably have
little effect on the turnover rates of P2Pt(OTf)2-catalyzed
Diels-Alder reactions.
dppe and R-BINAP cases (k1/k2 ) 320 for dppe; 24 for
R-BINAP). More importantly, however, a significant
difference in the ligand exchange rates of dppe and
R-BINAP Pt complexes was observed: k1 was 180 times
faster for dppe than for R-BINAP complexes, while k2
was 13 times faster for dppe. These diphosphine-
dependent rates may differ due to disparities in the
steric bulk of the two phosphines (the bulky R-BINAP
ligand may inhibit associative ligand attack on the Pt
center), disparities in ligand electronics (the more
electrophilic (R-BINAP)Pt center may bind more strongly
to Lewis bases, making ligand displacement less facile),
or a combination of both these factors.
Con clu sion s
Through in situ observation of P2Pt(OTf)2 catalysts
2, catalyst-substrate complexes 6, and catalyst-
1
product complexes 8 using 31P and H NMR spectros-
(4) Kin etics, Th er m od yn a m ics, a n d P t Diels-
Ald er Ca t a lysis. The thermodynamic and kinetic
relationships between complexes of the Lewis bases
OTf-, BF4-, 4, 5, and H2O with the [P2Pt]2+ Lewis acids
are summarized in the left half of the simplified free
energy diagram in Scheme 7. (Free energy values ∆G
and ∆Gq were calculated from the Keq and rate data, at
195 and 270 K, respectively.) This picture contrasts the
small ground-state energy differences between the P2-
PtII complexes (∆G ) 0.5-4 kcal mol-1) with the much
larger energy barriers observed for interconversion of
the complexes via ligand substitution reactions (∆Gq )
11-15 kcal mol-1); the latter values are also more
variable than the former, depending on both the mech-
anism for ligand substitution (k1 vs k2) and the identity
of the diphosphine (dppe vs R-BINAP).
The right half of the diagram depicts two possible
free-energy barriers to the interconversion of 6 and 8
via a Pt-catalyzed cycloaddition reaction (Scheme 7). If
the barrier to cycloaddition is lower than the barriers
observed for ligand exchange (path A), then HCp attack
on Pt-bound 4 will occur at a faster rate than ligand
exchange (kDA > k1 and k2), making ligand exchange the
turnover-limiting step of the catalytic cycle (Scheme 1).
On the other hand, a cycloaddition barrier higher than
the barriers to ligand substitution (path B) indicates
that the cycloaddition step itself is the slow, turnover-
limiting step of the catalytic cycle (Scheme 1). To
copy at 195 K, we have confirmed that 2-catalyzed
Diels-Alder reactions (eq 1) proceed by the expected
Lewis acid activation of dienophile 4, via coordination
of its carbonyl groups to the Pt catalyst (Scheme 1).
Investigation of the thermodynamic relationships be-
tween 2, a BF4- analogue of 2, and the Pt Lewis acid-
Lewis base complexes 3, 6, and 8 led to the following
conclusions. (1) Coordination of OTf- to Pt does not
thermodynamically impede binding of substrate 4 to the
-
Lewis acid catalyst, although BF4 inhibits dienophile
binding even less, consistent with its perception as a
“less coordinating” counterion than OTf-. (2) Water does
competitively inhibit binding of 4 to the catalyst. (3)
Product inhibition should be more significant in (dppe)-
Pt(OTf)2-catalyzed than in (R-BINAP)Pt(OTf)2-catalyzed
Diels-Alder reactions. The thermodynamic relation-
ships between the complexes studied were also sum-
marized by a scale of the relative binding strengths of
-
the Lewis bases to [P2Pt]2+ (BF4 , OTf- < 4 < 5 ,
H2O).
Investigation of the kinetic relationships between P2-
Pt(OTf)2 and [P2Pt(4)]2+[OTf]- via simulation of dy-
2
namic 31P NMR spectra of a 2/6 equilibrium mixture
revealed that two different exchange processes (4/4′ and
(35) Additional mechanistic information was obtained by monitoring
the Pt-catalyzed Diels-Alder reactions in situ by 31P and 1H NMR at
195 K. See ref 16.

The P2Pt(OTf)2-Catalyzed Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1575
OTf-/4) occur and that the rates (s-1) of both are 1-2
orders of magnitude faster for dppe than for R-BINAP
complexes. Also, experiments in which the reaction of
exchange more rapid. Such an enhanced ligand substi-
tution rate may explain the observation that P2Pt(SbF6)2
Lewis acids derived from A and AgSbF6 (P2 ) S-2,2′-
1
HCp with 6 was monitored by 31P and H NMR at 195
K enabled comparison of the rates of 6/8 interconversion
via cycloaddition reactions with the rates of 6/8 inter-
conversion via ligand exchange reactions. Observation
of immediate and quantitative conversion of 6 to 8 upon
addition of HCp, but no catalytic turnover, clearly
demonstrated that cycloaddition is much faster than
ligand substitution, making ligand exchange the turn-
over-limiting step of the catalytic cycle (Scheme 1).
From all the kinetic and thermodynamic data re-
ported here, the rates of ligand substitution processes
emerge as the most important factors controlling the
activity of P2Pt(OTf)2 Lewis acids 2 as catalysts for the
Diels-Alder reaction (eq 1). The fact that ligand ex-
change rather than cycloaddition kinetics determine the
overall efficiency of [P2M]2+-catalyzed reactions leads
to the obvious conclusion that the turnover rates of these
processes may be increased by accelerating ligand
substitution reactions, and not by further enhancing the
metal center’s electrophilicity. Indeed, our data indicate
that despite being considered “soft”, late-metal Lewis
acids are extremely electrophilic, as judged by the fact
that Pt-coordinated dienophiles undergo cycloaddition
very rapidly at 195 K (clearly, 4 is strongly activated
by coordination to Pt). This scenario, in which ligand
substitution rates rather than electrophilicity dominate
catalytic activity, suggests that [P2Pd]2+ Lewis acid
catalysts are more active than their Pt analogues for
transformations such as the Diels-Alder reaction be-
cause ligand substitution dynamics are faster, and not
because the Pd centers are more electrophilic than Pt.
While perhaps counterintuitive in the context of
typical substitution-labile Lewis acids, it is possible that
when ligand exchange is turnover-limiting, decreasing
the electrophilicity of the metal center might increase
its catalytic activity by diminishing the strength of the
exchanging Pt-ligand bonds enough to make ligand
bis[(4-R-C6H4)phosphino]-6,6′-(OMe)2-1,1′-biphenyl; R )
tBu, OMe, H, CF3)31a are more active catalysts for
glyoxylate-ene reactions when they contain electron-rich
rather than electron-poor diphosphines, even though
complexes derived from electron-poor ligands are pre-
sumably more electrophilic.
In conclusion, the combination of slow ligand substi-
tution rates with excellent electrophilicity inherent in
[P2M]2+ complexes suggests that these late-metal Lewis
acid catalysts will continue to display reactivity and
selectivity patterns that are distinct from those of early-
metal, Cu(II)/Zn(II), and p-block Lewis acids.
Ack n ow led gm en t. We graciously thank the NIH
(Grant No. GM60578), the Petroleum Research Fund,
administered by the American Chemical Society, and
Union Carbide for support of this research. We also
thank the UNC Board of Governors and the ACS
Division of Organic Chemistry for graduate Fellowships
(N.M.B.). M.R.G. is a Camille-Dreyfus Teacher-Scholar
(2000-2004).
Su p p or tin g In for m a tion Ava ila ble: Text giving experi-
mental details, plots of ln k1 and ln k2 vs 1/T (K) for 2a /6a
and 2b/6b ligand exchange reactions, tables of atomic param-
eters, isotropic and anisotropic thermal parameters, bond
distances, bond angles, and bond torsion angles for 2a , and
1H NMR spectra of 2a , 2b, 3a , 3b, 6a , 6b, 7a , 7b, 8a , and
R,2S-8b. This material is available free of charge via the
Internet at http://pubs.acs.org.
OM0109861