(dppe)Pt(OTf)2 and (R-BINAP)Pt(OTf)2 Lewis acids catalyze the Diels-Alder reaction by different mechanisms

Article abstract of DOI:10.1021/om010987t

In situ observation, by ³¹P and ¹H NMR at 195-240 K, of acryloyl-N-oxazolidinone (4)/cyclopentadiene (HCp) Diels-Alder reactions catalyzed by the Pt(II) Lewis acids P₂Pt(OTf)₂ (2; P₂ = dppe (a), R-BINAP (b)) revealed that the ditriflate catalysts are consumed by HCp at different rates to produce the catalytically inactive 18-electron complexes [P₂Pt(η⁵-Cp)]⁺-[OTf]^- (9). Mechanistic studies point to formal loss of HOTf from the proposed intermediate diene complex [P₂Pt(HCp)]²⁺[OTf]₂^-, facilitated by a necessary second equivalent of HCp; ill-defined HCp/HOTf byproducts, whose ¹H NMR resonances include a broad singlet at 12.7 ppm, are also formed. Both catalyst decomposition and Diels-Alder catalysis proceeded faster in reactions involving dppe Lewis acid 2a than in those involving R-BINAP Lewis acid 2b. In the 2b reaction at 195 K, Pt-bound dienophile 4 was immediately converted to the Pt-bound Diels-Alder adduct 2S-5 on addition of HCp; however, no turnover occurred and no η⁵-Cp complex was detected until the temperature was raised to 225 K. At this temperature, no low-field resonances were observed by ¹H NMR, a highly enantiopure Diels-Alder adduct was obtained, and TOFs less than the rates of ligand exchange were observed. These data are consistent with catalysis by a chiral [P₂Pt]²⁺ Lewis acid, in which ligand substitution is the turnover-limiting step of the catalytic cycle. In contrast, 9a and Diels-Alder adduct rac-5 formed rapidly at 195 K in the 2a reaction. The observation of low-field HCp/HOTf resonances (¹H NMR) and TOFs greater than the rates of ligand exchange together suggest that, in this case, the Diels-Alder reaction is catalyzed primarily by a Br?nsted acid mechanism involving the HCp/HOTf byproduct of 2a decomposition, rather than by [(dppe)-Pt]²⁺.

Full text of DOI:10.1021/om010987t

1576
Organometallics 2002, 21, 1576-1582
(d p p e)P t(OTf)2 a n d (R-BINAP )P t(OTf)2 Lew is Acid s
Ca ta lyze th e Diels-Ald er Rea ction by Differ en t
Mech a n ism s
Nicole M. Brunkan and Michel R. Gagne´*
Department of Chemistry, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-3290
Received November 14, 2001
1
In situ observation, by 31P and H NMR at 195-240 K, of acryloyl-N-oxazolidinone (4)/
cyclopentadiene (HCp) Diels-Alder reactions catalyzed by the Pt(II) Lewis acids P2Pt(OTf)2
(2; P2 ) dppe (a ), R-BINAP (b)) revealed that the ditriflate catalysts are consumed by HCp
at different rates to produce the catalytically inactive 18-electron complexes [P2Pt(η5-Cp)]+-
[OTf]- (9). Mechanistic studies point to formal loss of HOTf from the proposed intermediate
diene complex [P2Pt(HCp)]2+[OTf]- , facilitated by a necessary second equivalent of HCp;
2
1
ill-defined HCp/HOTf byproducts, whose H NMR resonances include a broad singlet at 12.7
ppm, are also formed. Both catalyst decomposition and Diels-Alder catalysis proceeded faster
in reactions involving dppe Lewis acid 2a than in those involving R-BINAP Lewis acid 2b.
In the 2b reaction at 195 K, Pt-bound dienophile 4 was immediately converted to the Pt-
bound Diels-Alder adduct 2S-5 on addition of HCp; however, no turnover occurred and no
η5-Cp complex was detected until the temperature was raised to 225 K. At this temperature,
no low-field resonances were observed by 1H NMR, a highly enantiopure Diels-Alder adduct
was obtained, and TOFs less than the rates of ligand exchange were observed. These data
are consistent with catalysis by a chiral [P2Pt]2+ Lewis acid, in which ligand substitution is
the turnover-limiting step of the catalytic cycle. In contrast, 9a and Diels-Alder adduct
rac-5 formed rapidly at 195 K in the 2a reaction. The observation of low-field HCp/HOTf
resonances (1H NMR) and TOFs greater than the rates of ligand exchange together suggest
that, in this case, the Diels-Alder reaction is catalyzed primarily by a Brønsted acid
mechanism involving the HCp/HOTf byproduct of 2a decomposition, rather than by [(dppe)-
Pt]2+.
In tr od u ction
oxophilic than their predecessors.2 This trait has already
been shown to alter the mechanisms of [P2Pd]2+
-
Group 10 transition-metal complexes of the type
[P2M]2+[A]-2 (P2 ) diphosphine; M ) Pd, Pt; A- ) OTf-,
BF4-, PF6-, SbF6-, ClO4-) have recently been shown to
act as excellent Lewis acid catalysts and, when equipped
with chiral diphosphines, have demonstrated the ability
to perform highly enantioselective transformations.1
Because of their unique stereoelectronic properties,
which differ substantially from those of p-block, octa-
hedral early-metal, or even first-row late-metal Cu(II)
or Zn(II) Lewis acids, these square-planar, 16-electron
catalysts could impart unprecedented reactivity and
selectivity to Lewis acid catalyzed reactions. For ex-
ample, the metal centers in this new class of catalysts
are considerably “softer” than those of any other type
of Lewis acid, making them more carbophilic and less
catalyzed imine or aldehyde alkylations by Mukaiyama-
type nucleophiles.3 Additionally, the fact that ligand
exchange is much less facile for group 10 P2MII com-
plexes than for their early-metal or first-row counter-
parts (and always proceeds via associative mecha-
nisms)4 could change the way that catalytic turnover is
achieved in these Lewis acid catalyzed reactions. How-
ever, little information concerning the effects of these
unusual properties on the mechanisms of [P2M]2+
-
catalyzed reactions has so far been reported. Thus, we
elected to study in detail the mechanisms of Diels-Alder
reactions (eq 1) catalyzed by the P2Pt(OTf)2 Lewis acids
2 (P2 ) dppe (a ), R-BINAP (b)), with the ultimate aim
(1) (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.
(2) (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.
(3) (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.
(4) (a) Cross, R. J . Adv. Inorg. Chem. 1989, 34, 219-292. (b)
J ohansson, L.; Tilset, M. J . Am. Chem. Soc. 2001, 123, 739-740.
10.1021/om010987t CCC: $22.00 © 2002 American Chemical Society
Publication on Web 03/19/2002

Lewis Acid Catalysis of the Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1577
Sch em e 1
of understanding these reactions at a sufficiently de-
tailed level to reveal their unique properties.
This paper is the second in a set of two that describe
our investigation of Lewis acids 2 and their roles as
Diels-Alder catalysts. In the first paper,5 we reported
that the active catalysts can be generated in situ by
addition of 2 equiv of triflic acid (HOTf) to the precursors
P2Pt(S-BINOL) (1). Lewis acids 2, as well as complexes
6 and 8, which were obtained by reaction of 2 with the
catalytically relevant Lewis bases 4 (dienophile; eq 1)
and 2S-5 (Diels-Alder adduct; eq 1), respectively,
proved amenable to characterization by 31P and 1H NMR
at 195 K; spectroscopic data were consistent with
bidentate coordination of 4 and 2S-5 to Pt (see Scheme
1). Investigation of the thermodynamic and kinetic
relationships between catalysts 2, catalyst-substrate
complexes 6, and catalyst-product complexes 8 revealed
that ground-state energy differences between the species
are small (<2.4 kcal mol-1 at 195 K) compared to the
barriers for their interconversion by associative ligand
substitution (11-15 kcal mol-1 at 270 K). These barriers
are also diphosphine dependent, as significantly faster
ligand exchange rates were measured (through line
shape analysis of dynamic 31P NMR spectra) for dppe
than for R-BINAP complexes. Finally, monitoring the
reaction of cyclopentadiene (HCp) with catalyst-
substrate complexes 6 by 31P and 1H NMR at 195 K
demonstrated that diene attack on the Pt-coordinated
dienophile 4 occurs more rapidly than ligand substitu-
tion (vide infra), making ligand exchange the turnover-
limiting step of the catalytic cycle.
in conjunction with a temperature-dependent quenching
of the reactive byproduct then cause Diels-Alder reac-
tions catalyzed by 2a and 2b to proceed by different
mechanisms.
Resu lts a n d Discu ssion
(1) In Situ Obser va tion of 2b-Ca ta lyzed Diels-
1
Ald er Rea ction s by 31P a n d H NMR. The Lewis acid
catalyst (R-BINAP)Pt(OTf)2 (2b) was generated at room
temperature by addition of 2 equiv of triflic acid (HOTf)
to binaphtholate 1b in the presence of 10 equiv of
dienophile 4, immediately producing the catalyst-
substrate complex 6b (Scheme 1). Since HOTf itself is
an efficient catalyst for the Diels-Alder reaction, care
was taken to avoid adding excess acid during catalyst
activation.6 31P and 1H NMR spectra of the reaction
mixture, acquired at 195 K, revealed that 91% of the
Lewis acid 2b was converted to the Pt-dienophile
complex 6b under these conditions (Figure 1A,G).7
Resonances for free and coordinated 4 were easily
1
distinguished in the H NMR spectrum (Figure 1G).
Collectively, the data presented in the preceding
paper point to ligand substitution rates, not electrophi-
licity of the metal center, as the key factor controlling
the activity of [P2Pt]2+ Diels-Alder catalysts. Consistent
with this scenario, the data further suggest that faster
turnover should be observed in Diels-Alder reactions
catalyzed by the dppe Lewis acid 2a rather than
R-BINAP Lewis acid 2b (despite the weaker electro-
philicity of 2a relative to 2b), simply because ligand
exchange should be turnover-limiting and it occurs more
rapidly in dppe than R-BINAP complexes. To test this
hypothesis, we investigated the effects of the diphos-
phine P2 on catalytic activity by directly comparing 2a -
and 2b-catalyzed Diels-Alder reactions (eq 1) in situ
by 31P and 1H NMR at 195 K. These experiments,
alluded to in the first paper because they demonstrated
that cycloaddition proceeds faster than ligand substitu-
tion, provided additional mechanistic information that
is presented here in detail. Essentially, the experiments
revealed that the ditriflate Lewis acids 2 react with HCp
under catalysis conditions to produce catalytically inac-
tive 18-electron Pt(II) complexes [P2Pt(η5-Cp)]+[OTf]-
(9) and a poorly characterized HOTf/HCp byproduct.
Diphosphine-dependent rates of catalyst decomposition
Fifty equivalents of freshly cracked HCp was then
added to the reaction mixture at 195 K, and low-
temperature 31P and 1H NMR spectra were recorded
periodically (Scheme 1 and Figure 1). The first spectra8
acquired after HCp addition showed only the Pt-
cycloadduct complex R,2S-8b and no Pt-dienophile
complex 6 (Figure 1B), indicating that the Pt-coordi-
nated dienophile (4*, Figure 1G) was immediately
converted to a Pt-coordinated Diels-Alder adduct (5*,
Figure 1H) by HCp. Despite the rapid cycloaddition step,
however, essentially no free product 5 appeared in the
1H NMR spectrum while it remained at 195 K, even
after 50 min. Clearly, the 2b -catalyzed Diels-Alder
reaction does not appreciably turn over at this temper-
ature. As discussed in the preceding paper, these results
indicate that the stereochemistry-determining cycload-
dition of HCp and 6 occurs faster than ligand exchange,
making ligand substitution the turnover-limiting step
of the catalytic cycle.5
(6) In the R-BINAP system, addition of 1.9 equiv of HOTf to yellow-
orange 1b produces catalyst solutions that remain slightly yellow
instead of losing all color, indicating that protonation of 1b is not
complete. Also, the fact that highly enantiopure product 2S-5 is
obtained from the 2b-catalyzed Diels-Alder reaction (96-98% ee)
indicates that catalysis by excess HOTf does not occur.
(7) Residual 2b (4%) and a small amount of [P2Pt(OH2)2]2+[OTf]-
2
(5) Brunkan, N. M.; White, P. S.; Gagne´, M. R. Organometallics
2002, 21, 1565-1575.
(3b) were also present.
(8) Acquired after 4 min reaction time; not shown in Figure 1.

1578 Organometallics, Vol. 21, No. 8, 2002
Brunkan and Gagne´
Ta ble 1. Tu r n over F r equ en cies for Ca ta lysis of Eq
1 by 2a a n d 2b Com p a r ed w ith Dien op h ile/OTf-
Liga n d Exch a n ge Ra tes (k2)a
Lewis acid
temp (K)
104(TOF) (s-1
)
104k2 (s-1
)
2b
2b
2b
2a
195
210
225
195
0.6
3.3
12
1.5
18
150
16
26
a
Dienophile/OTf- ligand exchange rates (k2) were measured by
line-shape analysis of dynamic 31P NMR spectra at higher tem-
peratures and extrapolated to the indicated temperatures (see the
Supporting Information).
pleted five turnovers in 70 min (Figure 1J ,K). Concur-
rently, however, a singlet at 3.1 ppm with an unusually
large P-Pt coupling of 4587 Hz (1J P-Pt ) 4210 Hz for
2b) grew slowly in the 31P NMR spectrum of the reaction
mixture (Figure 1D). The growth of this new complex
(9b) coincided with both the loss of R,2S-8b (Figure 1D)
1
and the appearance of a singlet at 5.4 ppm in the H
NMR spectrum (Figure 1J ,K). When the NMR temper-
ature was raised to 240 K, rapid Diels-Alder catalysis
occurred until all the dienophile present was converted
to 2S-5 (another four turnovers in 35 min; Figure 2).
Formation of 9b was also accelerated (Figure 1E), and
large, broad resonances characteristic of poly(HCp) grew
1
F igu r e 1. 31P (A-F) and H (G-L) NMR spectra acquired
during the 2b-catalyzed Diels-Alder reaction (eq 1). On
the 31P side, • denotes 3b and 9 denotes R,2R-8b. On the
1H side, 4* denotes Pt-coordinated 4, 5* indicates Pt-
coordinated 5, and * denotes HCp polymer.
1
in the H NMR spectrum at 5.6-5.8 and 1.0-3.0 ppm
(Figure 1L).10 Finally, raising the reaction temperature
to 255 K increased the rates of both 9b production (43%
conversion at the end of the experiment; Figure 1F) and
HCp polymerization. Upon removal from the probe, the
NMR tube contained a bright orange solution that
turned black when warmed to room temperature. Vis-
cous, white polymeric material also appeared in the
reaction vessel at room temperature.
Chiral HPLC analysis of the cycloadduct obtained
from this reaction confirmed that it was 97% ee 2S-5,
consistent with the enantioselectivities reported by
Ghosh.1a Preparative-scale 2b-catalyzed reactions, moni-
tored over time by GC analysis, proceeded at rates
similar to that of the NMR-scale reaction described
above and likewise yielded 96-98% ee 2S-5.11 Reactions
conducted at 195 K reached 30-50% completion and
then died, indicating that turnover, although slow at
this temperature, is not completely absent.
Disregarding (for the moment) the unanticipated
appearance of 9b, the 2b-catalyzed Diels-Alder reaction
proceeded in a manner consistent with the data pre-
sented in the preceding paper, which indicated that
cycloaddition should be rapid and ligand substitution
should be turnover-limiting in the catalytic cycle.5 In
situ monitoring of 2S-5 formation at different temper-
atures allowed calculation of crude turnover frequencies
(TOFs) at these temperatures (Table 1). These TOFs
were expected to roughly correlate with the dienophile/
OTf- ligand exchange rates k2 measured in the previous
paper, if ligand substitution actually is the slowest step
of the catalytic cycle.5 Indeed, comparison of TOFs and
k2 values (Table 1) showed that, at each temperature,
catalysis rates were similar to but slightly slower than
the 4/OTf- exchange rates obtained from 31P NMR line-
F igu r e 2. Conversion vs time for the 2b-catalyzed Diels-
Alder reaction (eq 1): (b) % 4; (2) % 5; (4) % Pt-coordinated
5; (() % 9b. The reaction temperature was increased at the
times indicated by the vertical lines.
When the NMR probe temperature was raised to 210
K, the 2b-catalyzed Diels-Alder reaction began to turn
1
over sluggishly (Figure 1C,I). Although H NMR spectra
of the reaction mixture were complex, all resonances
could be assigned to 8b, 4, 5, free S-BINOL, or HCp,9
and the reaction progress could be assessed by integrat-
1
ing the H NMR resonances for free 4, free 5, and Pt-
coordinated 5 at 5.95, 5.85, and 5.75 ppm, respectively
(Figure 1). The evolution of the 2b-catalyzed reaction
in eq 1, determined from spectral integration data, is
plotted in Figure 2.
After one turnover was completed (50 min at 210 K),
the NMR temperature was elevated to 225 K (Figure
2), at which point the reaction accelerated and com-
(10) (a) Ma, J .; Corcoran, R. C. J . Am. Chem. Soc. 1991, 113, 8973-
8975 and ref 23 therein. (b) Ouchi, M.; Kamigaito, M.; Sawamoto, M.
Macromolecules 2001, 34, 3176-3181.
(11) Preparative-scale reaction conditions were exactly the same as
those for NMR reactions, except 3 times less concentrated.
(9) See the Supporting Information for the complete 1H NMR
spectrum.

Lewis Acid Catalysis of the Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1579
1
1
F igu r e 3. 31P (A-F) and H (G-L) NMR spectra acquired during the 2a -catalyzed Diels-Alder reaction (eq 1). On the H
side, • denotes the HOTf/HCp byproduct, 4* indicates Pt-coordinated 4, 5* denotes Pt-coordinated 5, and * denotes HCp
polymer.
shape analysis. These data further support a mecha-
nistic scenario in which ligand substitution rates control
the activity of the Pt Lewis acid catalyst.
(2) In Situ Obser va tion of 2a -Ca ta lyzed Diels-
Ald er Rea ction s by 31P a n d 1H NMR. Since 4/OTf-
ligand exchange occurs more quickly in dppe than in
R-BINAP complexes, 2a -catalyzed Diels-Alder reac-
tions were predicted to proceed faster than their 2b-
catalyzed counterparts.5 To test this prediction experi-
mentally, a reaction promoted by dppe Lewis acid 2a
1
was similarly monitored by H and 31P NMR (Figure
3). Unexpectedly, significant differences between the
dppe and R-BINAP reactions were observed.
Initially, generating Pt Lewis acid catalyst 2a from
precursor 1a and HOTf in the presence of 10 equiv of
F igu r e 4. Conversion vs time for the 2a -catalyzed Diels-
Alder reaction (eq 1): (b) % 4; (2) % 5; (() % 9a . The
reaction temperature was increased at the times indicated
by the vertical lines.
dienophile 4 afforded the expected equilibrium mixture
of 2a and Pt-dienophile complex 6a (Figure 3A,G).
Addition of 50 equiv of HCp to the reaction mixture at
195 K immediately and quantitatively converted 6a to
8a , as in the R-BINAP reaction (Figure 3B,H). However,
unlike in the 2b case, conversion of the active [P2Pt]2+
catalyst to a symmetric Pt complex with a large P-Pt
no excess HOTf had been added during catalyst activa-
tion.12 Therefore, unlike in the R-BINAP case, it ap-
peared that HOTf was generated by reaction of P2Pt-
(OTf)2 with HCp.
1
coupling constant (42.1 ppm, J P-Pt ) 4512 Hz; 9a )
occurred rapidly at 195 K, yielding 20% 9a after only 8
min reaction time (Figure 3B)! In the 1H NMR spec-
trum, unidentified resonances appeared at 5.75 ppm
(analogous to the resonance at 5.4 ppm in the R-BINAP
reaction) and 12.7 ppm; the 12.7 ppm resonance re-
mained approximately constant in size (0.2 H compared
to S-BINOL) throughout the reaction (Figure 3H-L).
Because of its chemical shift, this downfield signal was
initially assigned to HOTf (vide infra). Note that this
resonance was not present before HCp was added to the
reaction mixture (Figure 3G), and it was certain that
The Diels-Alder reaction proceeded rapidly at 195
K (Figure 3I,J ), reaching complete conversion in 65 min,
while 9a grew in the 31P NMR spectrum (Figure 3C,D;
see also Figure 4). When the Diels-Alder reaction was
complete, the NMR probe temperature was raised to 210
K, which accelerated formation of 9a (Figure 3E).
(12) A small amount (3%) of a complex that we propose to be a Pt2-
(µ-BINOL) dimer was present in the reaction mixture after addition
of HOTf. This species only forms when less than 2 equiv of acid is added
to 1a (manuscript in preparation).

1580 Organometallics, Vol. 21, No. 8, 2002
Brunkan and Gagne´
Sch em e 2
Sch em e 3
Additionally, HCp oligomers and polymers appeared by
1H NMR (Figure 3K). Increasing the probe temperature
to 225 K eventually produced 9a in 78% yield (Figure
3F,L), and removal of the NMR tube from the probe
revealed a bright yellow reaction solution. As in the
R-BINAP case, this solution turned black on warming
to room temperature, and viscous, polymeric material
was deposited in the NMR tube. A 31P NMR spectrum
of this sample at 300 K consisted exclusively of 9a , while
1H NMR showed that all free HCp had polymerized.10,13
Preparative-scale reactions proceeded at rates similar
to those of the NMR-scale reaction.11
HCp, as shown in Scheme 2A, is not. Several experi-
ments were carried out to probe the mechanism by
which 9a forms from 2a and HCp. First, to assess the
feasibility of an intermediate η2- or η4-coordinated diene
on the reaction surface, we investigated the Pt(II)
coordination chemistry of cyclohexadiene (CHD), a less
reactive diene than HCp. Preliminary experiments
showed that a combination of 2a and CHD does not
produce a complex akin to 9a , although it does generate
a black, polymer-containing solution and small amounts
of benzene on warming to room temperature. When 50
equiv of CHD was added to 2a at 195 K, a 31P NMR
spectrum of the reaction mixture consisted of two
singlets 24 Hz apart (53.1 ppm), with a single set of
broad Pt satellites (1J Pt-P ) 3580 Hz), as well as residual
2a (41%). Though the 31P spectrum is not readily
interpretable, the magnitude of the P-Pt coupling
constant is in the range reported for phosphorus ligands
trans to η2-coordinated olefins in Pt(II) complexes,16
suggesting that coordination of a 1,3-diene to the
electrophilic [P2Pt]2+ center is feasible, and could rea-
sonably represent a first step in the conversion of 2 to
9 by HCp.
(3) Id en t ifica t ion of 9 a s [P 2P t (η5-Cp )]+[OTf]-.
Because adding HCp to [P2Pt]2+ precipitated the forma-
tion of complex 9 (at least in the dppe case), we
attempted to prepare 9a independently by adding 2
equiv of HCp to 2a at 195 K (Scheme 2A). Indeed,
quantitative conversion of colorless 2a to yellow 9a was
observed by 31P NMR under these conditions. A 1H NMR
spectrum of the reaction mixture contained a sharp
triplet (3J P-H ) 1.5 Hz) with Pt satellites (2J Pt-H ) 11.8
Hz) at 5.8 ppm (5 H), consistent with coordination of
an η5-Cp ligand to the P2PtII fragment (Scheme 2A). On
the basis of these NMR data, 9a was identified as the
18-electron complex [(dppe)Pt(η5-Cp)]+[OTf]-, previously
prepared by Anderson from [(dppe)Pt(OH2)2]2+[OTf]-
2
(3a ) and TlCp (Scheme 2B).14 Small broad resonances
at 12.2 and 17.5 ppm (0.2 H each) and in the aliphatic
region of the spectrum were also observed and were
assigned to ill-defined HOTf/HCp byproducts. When it
was warmed to room temperature, the reaction mixture
turned brown, and the downfield resonances disap-
Counterintuitively, addition of 1 equiv rather than 2
equiv of HCp to 2a at 195 K produced 9a in only 50%
yield (Scheme 3A). This observation led to the hypoth-
esis that the second equivalent of HCp might function
as the base that deprotonates the [P2Pt(η4-HCp)]2+
complex to give the η5-Cp ligand. However, the logical
byproduct of such a reaction, cyclopentallyl triflate (10),
was never observed by 1H NMR, even though the related
species allyl triflate is stable at low temperature and
1
peared from the H NMR spectrum.
The analogous R-BINAP η5-Cp complex 9b was
prepared independently as a crystalline, red-orange
solid by addition of 1 equiv of TlCp to the µ-chloro dimer
[(R-BINAP)Pt(µ-Cl)]22+[OTf]- , in a modification of Ander-
1
gives sharp, well-defined H NMR resonances between
2
son’s protocol for 9a . The 31P and H (5.4 ppm, J P-H
)
1
3
4.9 and 6.0 ppm.17 The spectrum of the present reaction
contained only broad resonances downfield of 12 and
upfield of 3 ppm, casting doubt on the role of 10. Equally
puzzling was the fact that 2 equiv of HCp was required
to produce 1 equiv of 9a from (dppe)PtCO3 and 1 equiv
of HOTf in the presence of HCp (Scheme 3B). We
predicted that an intermediate “[P2PtOH]+[OTf]-”18 or
“[(dppe)Pt(CO3H)]+[OTf]-” species would be sufficiently
2
1.6 Hz; J Pt-H ≈ 7.2 Hz) NMR spectra of 9b identified
it as the unknown decomposition product from the
R-BINAP Diels-Alder reaction (Figure 1).15 Control
experiments showed that 9a and 9b (10 mol %) are not
catalysts for the Diels-Alder reaction, consistent with
their 18-electron formulations.
(4) Mech a n ism of 9a F or m a tion fr om 2a a n d
HCp . While Anderson’s metathetical synthesis of 9a is
intuitively reasonable, the production of 9a from 2a and
(16) (a) Cross, R. J .; Davidson, M. F. J . Chem. Soc., Dalton Trans.
1987, 139-143. (b) Kurosawa, H.; Asada, N.; Urabe, A.; Emoto, M. J .
Organomet. Chem. 1984, 272, 321-329.
(13) The cycloadduct obtained from 2a -catalyzed Diels-Alder reac-
tions was racemic, as expected for the achiral dppe-based catalyst.
(14) Fallis, S.; Rodriguez, L.; Anderson, G. K.; Rath, N. P. Organo-
metallics 1993, 12, 3851-3855.
(15) The spectroscopic data are similar to those of a recently isolated
[(biphep)Pt(η5-Cp)]+[OTf]- complex: Becker, J . J .; White, P. S.; Gagne´,
M. R. J . Am. Chem. Soc. 2001, 123, 9478-9479.
(17) (a) Vedejs, E.; Engler, D. A.; Mullins, M. J . J . Org. Chem. 1977,
42, 3109-3113. (b) Beard, C. D.; Baum, K.; Grakauskas, V. J . Org.
Chem. 1973, 38, 3673-3677.
(18) [P2Pt(µ-OH)]22+[A]-2 complexes are well-known in the literature.
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.

Lewis Acid Catalysis of the Diels-Alder Reaction
Organometallics, Vol. 21, No. 8, 2002 1581
Sch em e 4
basic to deprotonate the acidic Pt(HCp) intermediate to
afford 9a and 1 equiv of carbonic acid or water/CO2
(Scheme 3B). However, 1 equiv of HCp converted only
50% of the putative monotriflate intermediate to 9a .19
The identity of the 12.7 ppm resonance in 1H NMR
spectra of the 2a -catalyzed Diels-Alder reaction was
probed by combining 4, 5, S-BINOL, and HCp with 1
equiv of HOTf in separate NMR tubes. These species
were all present in the 2a -catalyzed reaction mixture,
and each could conceivably deprotonate Pt-coordinated
HCp and produce an acid-base adduct with a resonance
at 12.7 ppm. However, combining HOTf with the Lewis
bases 4, 5, and S-BINOL at 195 K generated protonated
products with the acidic proton appearing at 9 ppm (S-
BINOL, 3 H),20 17 ppm (cycloadduct 5, 1 H),21 or 18 ppm
(dienophile 4, 1 H) rather than 12.7 ppm in the 1H NMR
spectrum (CD2Cl2; free HOTf at 9.25 ppm). In the case
1
of HCp, a H NMR spectrum acquired immediately after
addition of HOTf at 195 K already consisted primarily
of unidentified broad resonances upfield of 6 ppm, but
a resonance at 12.3 ppm, similar to the one observed in
the catalytic reaction mixture, was also observed, sug-
gesting that the downfield peak observed during 2a
catalysis is derived from a combination of HCp and
HOTf. If this is indeed the case, then the proposed
intermediate [(dppe)Pt(HCp)]2+[OTf]- must be acidic
2
enough to protonate free cyclopentadiene, pointing to a
remarkably electrophilic [P2Pt]2+ fragment.
oligomerization, since evidence of HCp polymerization
does not appear in the 1H NMR spectrum until the
reaction temperature is raised to 225 K. In contrast,
decomposition of [P2Pt]2+ to 9b does not occur at 195 K
in the R-BINAP case but begins slowly when the
reaction temperature is raised to 225 K. At this tem-
perature, however, no resonances for acidic protons are
observed in the downfield region of the 1H NMR
spectrum; instead, HCp oligomers appear immediately.
Thus, the HOTf/HCp product formed at 225 K must
initiate HCp polymerization, rather than getting in-
volved in Brønsted catalysis of the Diels-Alder reaction.
The high enantioselectivity of 2b-catalyzed reactions
(96-98% ee, as reported by Ghosh)1a also indicates that
a conventional Lewis acid pathway mediated by a chiral
metal complex (Scheme 4A), rather than an achiral
proton-catalyzed process (Scheme 4B), is the operative
mechanism in 2b-catalyzed reactions.
The rapid rates observed for Diels-Alder reactions
involving dppe Lewis acid 2a provide additional evi-
dence that these reactions probably proceed primarily
through the Brønsted acid catalyzed route shown in
Scheme 4B, which is expected to turn over more quickly
than the [P2Pt]2+-mediated cycle. (Diels-Alder reactions
catalyzed by 10 mol % of HOTf are complete in <5 min
at 195 K.) Comparison of the catalyst TOF at 195 K with
the dienophile/OTf- ligand exchange rate k2 measured
in the preceding paper (Table 1)5 reveals that in contrast
to the situation observed for the R-BINAP Lewis acid-
catalyzed reaction (catalysis rates slower than dienophile/
OTf- exchange were measured), turnover actually oc-
curs faster than dienophile/OTf- ligand exchange in the
dppe system. This discrepancy between the dppe and
R-BINAP cases strongly suggests that the rate of the
2a -promoted Diels-Alder reactions is augmented by a
(5) d p p e vs R-BINAP : Mech a n ism s of 2a - a n d 2b-
Ca ta lyzed Diels-Ald er Rea ction s. Figures 2 (R-
BINAP) and 4 (dppe) show the time courses of 2b- and
2a -catalyzed Diels-Alder reactions (eq 1), respectively,
as well as decomposition of [P2Pt]2+ to the η5-Cp
complexes 9. Comparison of these plots reveals that both
Diels-Alder catalysis and production of 9 occur more
rapidly for [(dppe)Pt]2+ than for [(R-BINAP)Pt]2+. Al-
though the fact that 2a -catalyzed Diels-Alder reactions
proceed faster than their 2b-catalyzed counterparts is
consistent with our initial, simple hypothesis that faster
ligand exchange rates for dppe complexes should en-
hance turnover rates for 2a -catalyzed reactions,5 the
mechanistic data described here suggest a situation
more complex than previously appreciated.
The key difference between the dppe and R-BINAP
Pt-catalyzed Diels-Alder reactions is the difference in
rates of [P2Pt]2+ decomposition to the η5-Cp complexes
9, which mirrors the difference in ligand exchange rates
observed previously. In 2a -catalyzed reactions (dppe),
addition of HCp to the reaction mixture at 195 K rapidly
produces 9a and an unknown form of HOTf/HCp,
observed by 1H NMR as a broad resonance at 12.7 ppm,
which is evidently acidic enough to catalyze the Diels-
Alder reaction by a Brønsted acid mechanism that is
rapid at 195 K (Scheme 4A).22 The HOTf/HCp product
formed at 195 K apparently does not initiate HCp
(19) Besides 9a (50% yield), the reaction of (dppe)PtCO3 with 1 equiv
of HOTf and 1 equiv of HCp (Scheme 3B) afforded the µ-OH dimer
[(dppe)Pt(µ-OH)]22+[OTf]-2, the same product formed in the absence
of HCp.18
(20) The 9 ppm resonance in the 195 K 1H NMR spectrum of
protonated S-BINOL integrated for 3 H, indicating that exchange of
the acidic proton with the OH protons of S-BINOL is rapid on the NMR
time scale.
(21) Warming to room temperature caused reaction of 5 with HOTf
(complete in 10-15 min), producing new, unidentified species and a
very small resonance at 12.5 ppm in the 1H NMR spectrum. No reaction
was observed in the case of dienophile 4.
(22) A carbonium ion catalyst (in analogy to trityl cation) cannot be
ruled out, though an H+ mechanism seems more plausible.

1582 Organometallics, Vol. 21, No. 8, 2002
Brunkan and Gagne´
non-Pt, Brønsted acid catalysis pathway, making dif-
ferent mechanisms operative in Diels-Alder reactions
involving Pt catalysts 2a and 2b.
acid catalysts can be drawn. (1) Despite the change in
mechanism documented herein, the data still show that
ligand exchange rates dominate the activity of [P2Pt]2+
catalysts, though not in the direct way that was
envisioned at the onset of the work. (2) Rapid (-78 °C)
conversion of Pt-coordinated dienophile to Pt-coordinat-
ed Diels-Alder product demonstrates the significant
Lewis acidity and activating potential of the [P2Pt]2+
fragment. Moreover, the high enantioselectivities ob-
served with the bulky R-BINAP catalyst show that even
when the stereochemistry-determining step of a reaction
is fast, the chiral P2 ligand is extremely efficient at
directing reaction selectivity. (3) Although ultimately
detrimental to the catalyst’s integrity in this case, the
carbophilicity of [P2Pt]2+ enhances the rate of a unique
reactivity pathway (9 formation) that is atypical for
conventional Lewis acids. During other transformations,
this type of reactivity may prove beneficial, especially
in combination with the highly electrophilic character
of these Pt catalysts.
Su m m a r y a n d Con clu sion s
Diels-Alder reactions catalyzed by the P2Pt(OTf)2
Lewis acids 2a (dppe) and 2b (R-BINAP) were moni-
tored in situ at 195 K by 31P and 1H NMR. These
experiments confirmed that ligand substitution is the
turnover-limiting step of the 2b-catalyzed reaction and
that the reaction turns over at a rate similar to
previously measured 4/OTf- ligand exchange rates.
Unexpectedly, decomposition of the [P2Pt]2+ catalysts
by reaction with HCp to give the noncatalysts [P2Pt(η5-
Cp)]+[OTf]- occurred in both the dppe and R-BINAP
systems. Mechanistic studies of the dppe case revealed
that a second equivalent of HCp was required to convert
2a to 9a , presumably to deprotonate a (highly acidic)
[(dppe)Pt(HCp)]2+[OTf]- intermediate and form an
2
acidic HOTf/HCp species with a resonance at 12.7 ppm.
Comparison of the NMR data for 2a - and 2b-promoted
Diels-Alder reactions suggest that the two reactions
proceed by different mechanisms because 2a and 2b
decompose to 9 at different temperatures. In the dppe
case, 9a forms rapidly at 195 K; at this temperature,
the acidic HOTf/HCp byproducts that are formed do not
initiate HCp polymerization but, instead, promote the
Diels-Alder reaction by a Brønsted acid catalyzed route
(Scheme 4B). In the R-BINAP reaction, decomposition
of [P2Pt]2+ to 9b does not occur until 225 K. At this
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, the ACS Division
of Organic Chemistry, and the UNC Department of
Chemistry (Venable Award) for graduate Fellowships
(N.M.B.). M.R.G. is a Camille-Dreyfus Teacher-Scholar
(2000-2004).
1
temperature the acidic species visible in the H NMR
spectrum of the 2a reaction (12.7 ppm) is not observed,
apparently because it more quickly initiates HCp po-
lymerization. Therefore, the 2b-promoted Diels-Alder
reaction proceeds via a true Pt-catalyzed mechanism
(Scheme 4A).
Su p p or tin g In for m a tion Ava ila ble: Text giving ad-
ditional experimental details and figures giving 1H NMR
spectra of 2a - and 2b-catalyzed Diels-Alder reactions. This
material is available free of charge via the Internet at
http://pubs.acs.org.
From the experiments described above, several broad
conclusions regarding the reactivity of [P2Pt]2+ Lewis
OM010987T