DFT and experimental studies on the PtX2/X--catalyzed olefin hydroamination: Effect of halogen, amine basicity, and olefin on activity, regioselectivity, and catalyst deactivation

Article abstract of DOI:10.1021/om2009149

A DFT/B3LYP study with inclusion of solvent and temperature effects has probed the olefin activation mechanism for the intermolecular hydroamination of ethylene and 1-hexene by aniline derivatives catalyzed by the PtX ₂/X^- system on the basis of a variety of experimental results, including new experiments on catalyst deactivation. For ethylene and aniline, the calculated ΔG?_cycle between the resting state [PtX₃(C₂H₄)]^-, 1^X, and the TOF-determining transition state of the C-H reductive elimination from [PtX₃(H)-(CH₂CH₂NHPh)]^-, TS2 ^X, is slightly smaller for X = Br than for Cl or I. The ΔG?_cycle decreases as the aniline basicity decreases. For the slightly less efficient hydroamination of 1-hexene, ΔG? _cycle is greater than that for the hydroamination of ethylene, with a preference for the Markovnikov addition, in agreement with experiment and with essentially equivalent ΔG?_cycle values for the Br and I systems. In general, the results of the calculations are in agreement with the experimental observation. A clear-cut comparison of trends is hampered by the small energy differences and by the possibility, proven in certain cases, that the reaction parameters under investigation affect the catalyst degradation rate in addition to its intrinsic activity. Extrapolation of the computational study to the fluoride system suggests that this should be even more active. However, experimental studies show that this is not the case. The reason for this anomaly has been traced to the basicity of the fluoride ion, which triggers more rapid catalyst decomposition. A bonding analysis of 1^X indicates a significant push-pull π interaction between the C₂H₄ and the trans-F ligand.

Full text of DOI:10.1021/om2009149

Article
pubs.acs.org/Organometallics
DFT and Experimental Studies on the PtX2/X−-Catalyzed Olefin
Hydroamination: Effect of Halogen, Amine Basicity, and Olefin on
Activity, Regioselectivity, and Catalyst Deactivation
Pavel A. Dub,† Aurelien Bethegnies,† and Rinaldo Poli*,†,‡
́
́
†CNRS, LCC (Laboratoire de Chimie de Coordination), Universite
Toulouse, France
́
de Toulouse, UPS, INP, 205, route de Narbonne, F-31077
‡Institut Universitaire de France, 103, bd Saint-Michel, 75005 Paris, France
S
* Supporting Information
ABSTRACT: A DFT/B3LYP study with inclusion of solvent and temperature
effects has probed the olefin activation mechanism for the intermolecular
hydroamination of ethylene and 1-hexene by aniline derivatives catalyzed by the
PtX2/X− system on the basis of a variety of experimental results, including new
experiments on catalyst deactivation. For ethylene and aniline, the calculated
ΔG‡cycle between the resting state [PtX3(C2H4)]−, 1X, and the TOF-determining
transition state of the C−H reductive elimination from [PtX3(H)-
(CH2CH2NHPh)]−, TS2X, is slightly smaller for X = Br than for Cl or I. The
ΔG‡
decreases as the aniline basicity decreases. For the slightly less efficient hydroamination of 1-hexene, ΔG‡
is greater
cycle
cycle
than that for the hydroamination of ethylene, with a preference for the Markovnikov addition, in agreement with experiment and
with essentially equivalent ΔG‡cycle values for the Br and I systems. In general, the results of the calculations are in agreement with
the experimental observation. A clear-cut comparison of trends is hampered by the small energy differences and by the possibility,
proven in certain cases, that the reaction parameters under investigation affect the catalyst degradation rate in addition to its
intrinsic activity. Extrapolation of the computational study to the fluoride system suggests that this should be even more active.
However, experimental studies show that this is not the case. The reason for this anomaly has been traced to the basicity of the
fluoride ion, which triggers more rapid catalyst decomposition. A bonding analysis of 1X indicates a significant push−pull π
interaction between the C2H4 and the trans-F ligand.
in the hydroamination of ethylene was obtained with the
catalyst PtBr2/nBu4PBr, whereas hydroamination of 1-hexene is
most productive with PtBr2/nBu4PI,29 an observation that has
not yet been rationalized. Halide ion effects have also been
observed for other types of catalyzed transformations, but
poorly understood in most cases.30−33 Another interesting
observation, made for the hydroamination of ethylene, is that
1. INTRODUCTION
Hydroamination, the direct addition of an amine N−H bond
across an unsaturated C−C bond, is a challenging trans-
formation for which catalytic approaches are attracting much
interest in both academia and industry worldwide.1 Alkenes are
the most difficult substrates to hydroaminate,2 especially for the
intermolecular version and when using nonactivated substrates,
namely, the simple α-olefins.3 An additional challenge concerns
the N−H addition regiochemistry in favor of the anti-
Markovnikov product for monosubstituted olefins,4 listed as
one of the “Ten Challenges for Catalysis”.2 For the
homogeneously catalyzed hydroamination of ethylene and
other nonactivated olefins, limited success was achieved by
using Li,5 lanthanides,6,7 Zr,8 and late transition metals, such as
Fe,9 Ru,10−12 Rh,13−15 Ir,14,15 Pd,16,17 Ag,18 Au,19−22 and
notably Pt.23−25
Investigations initiated in our team by J.-J. Brunet have
shown that the hydroamination of ethylene or 1-hexene by
anilines is catalyzed by PtX2 in the presence of nBu4PX (X = Cl,
Br, I) as an activator, this being one of the most efficient
catalytic systems reported for this challenging transforma-
tion.26−29 The hydroamination catalytic activity is slightly lower
for 1-hexene than for ethylene under the same conditions. The
highest productivity (expressed by the turnover number, TON)
the TON is inversely proportional to the ArNH2 basicity.26
A
greater productivity for less basic amines (anilines, carbox-
amides, and sulfonamides) has also been reported for the
intermolecular hydroamination of ethylene catalyzed by other
23
Pt systems, notably PtCl2(C2H4)2 and Pt(COD)(OTf)2.24
Finally, a high Markovnikov selectivity (95%) is observed for
the 1-hexene hydroamination.27
The hydroamination reaction with late-transition-metal
complexes as catalysts can, in principle, follow an olefin
activation or an amine activation mechanism, and the former
appears as the most likely option on the basis of experimental
evidence for group 10 metals.29 The mechanism has been
analyzed by a few computational investigations. Restricting the
citations to the intermolecular version and to late-transition-
Received: September 29, 2011
Published: December 6, 2011
© 2011 American Chemical Society
294
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
metal catalysts, a first report by Senn et al. addressed the model
NH3 addition to ethylene catalyzed by the [MCl(PH3)-
(C2H4)]z+ complexes of group 9 (z = 0) and 10 (z = 1)
metals.34 The study addressed only the olefin activation
pathway for the group 10 metals and showed that the TOF-
determining step is the H−CH2CH2NH2 bond formation by
reductive elimination from a PtIV aminoalkyl hydride
intermediate, whereas the NH3 nucleophilic addition to the
activated ethylene is thermodynamically and kinetically
favorable. More recently, Tsipis and Kefalidis, using the “Pt0”
model complex Pt(C2H4)(PH3), explored only the amine
activation pathway, finding the reaction to be limited by the
product reductive elimination step from the PtII amido alkyl
intermediate.35 Other computational studies (e.g., on gold-
catalyzed diene36 or ethylene37 hydroamination or palladium-
catalyzed vinylarenes hydroamination38) have also explored
solely the olefin activation mechanism.
In a recent article, we have reported a detailed DFT analysis
of the catalytic cycle for the C2H4/PhNH2 reaction catalyzed by
the PtBr2/Br− system.39 This study indicated that the preferred
pathway is the olefin activation, rather than the amine
activation. The best pathway, shown schematically in Scheme
Figure 1. Free energy profile of the aniline addition to ethylene
catalyzed by PtBr2/Br−.39 Values are ΔGCPCM (aniline) relative to
[PtBr3(C2H4)]− in kcal mol−1 at 298.15 K (and at 423.15 K in
parentheses).
Scheme 1. Mechanism of the PtBr2/Br−-Catalyzed
Hydroamination of C2H2 by Aniline39
determined by the free energy difference (ΔG‡cycle) between
the resting state and the TOF-determining transition state,41,42
which are, respectively, 1Br and TS2Br.
The aim of this article is to test the proposed catalytic cycle,
by means of additional DFT investigations, to check whether it
fits with the other observations highlighted above, and to
possibly provide their rationalization. Additional catalytic
experiments have also been carried out, showing how the
observed TON is a reflection not only of the catalytic activity
(TOF, related to the ΔG‡cycle) but also of catalyst
decomposition, which has been traced to the decomposition
of the zwitterionic intermediate (4), by reaction with an
external base.
2. RESULTS AND DISCUSSION
As mentioned in the Introduction, the prototypical hydro-
amination process investigated in our laboratory has been the
addition of aniline to ethylene, for which the highest TONs
were found for the PtX2/X− catalytic system with X = Br.
Therefore, the previously reported comprehensive computa-
tional study of the catalytic cycle was restricted to this halide
and to ethylene as the olefin and aniline as the amine.39 The
1, involves bromide and olefin coordination to give the resting-
state complex [PtBr3(C2H4)]− (1Br). Other off-loop species,
namely, trans-PtBr2(C2H4)(PhNH2) (2Br), cis-PtBr2(C2H4)-
(PhNH2), trans- and cis-PtBr2(C2H4)2, and trans-
PtBr2(PhNH2)2 (3Br) were also found at relatively low energy
(in agreement with experimental equilibrium studies),40 the
lowest ones being 2Br and 3Br.
The essential features of the catalytic cycle are also shown as
an energetic profile in Figure 1. The aniline nucleophilic
addition to 1Br yields the zwitterionic complex
[PtBr3(CH2CH2NH2Ph)]− (4Br) through the transition state
TS1Br. Although the nucleophilic addition to other olefin
complexes, notably 2Br and trans-PtBr2(C2H4)2, has a lower
barrier, only 4Br is the productive intermediate. The latter easily
ΔG‡
value resulting from the study (35.6 kcal mol−1) is in
cycle
relatively good agreement, given the computational approx-
imations, with the experimentally measured catalytic turnover
frequency (38 h−1 at 150 °C,26 which yields the upper estimate
of 28.9 kcal/mol). The discrepancy between the calculation and
the experiment (ca. 7 kcal mol−1) is not unreasonable given the
usual precision of the DFT method, considering the additional
approximation involved in modeling solvation effects for ionic
species and the estimation of thermodynamic parameters in a
condensed phase.43 For comparison purposes, the additional
calculations presented in this contribution were carried out at
the same level of theory, with inclusion of the gas-phase free
energy and solvent effects at the CPCM level in aniline, but
without consideration of ion-pairing effects. The reasons for
these choices have been discussed previously.39,40,43
transfers
a proton to the Pt atom to yield
[PtHBr3(CH2CH2NHPh)]− (5Br), which then proceeds to
C−H reductive elimination through the TOF-determining
transition state TS2Br with formation of the transient σ-
complex [PtBr3(k2:C,H-HCH2CH2NHPh)]− (6Br). Liberation
of the product from 6Br (or from a rearranged isomeric form
w it h a n N- bo nd ed h ydro a min a tio n pro du ct ,
[PtBr3(NHEtPh)]−, 7Br) via ligand substitution by a new
C2H4 molecule regenerates 1Br. The overall activation barrier is
a. Halide Effect in the Hydroamination of Ethylene.
All energies are given with respect to the reference complex
295
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
[PtX3(C2H4)]−, 1X, taking into account the values of all the
species that are added to or subtracted from the system. The
results are shown in Figure 2. As mentioned above, the
1X with the molecules present under catalytic conditions (X−,
aniline, and ethylene), were at even greater energy for X = Br,43
and their optimization for the other halogen systems was not
pursued.
The energy gap between 1X and complexes 2X and 3X is
much greater for X = F (12.4 kcal/mol for 2F and 15.9 kcal/mol
for 3F), lower and relatively halogen-independent for the three
heavier halogens (ca. 4 kcal/mol for both series). Although this
has no relevance to the catalyst activity, since the resting state
remains, in all cases, the 1X system, it is interesting to
understand this “fluoride effect”. With this objective, we have
carried out more detailed analyses of the electronic structure
and charge distribution for the 1X, 2X, and 3X complexes, and
the most relevant results are shown in Table 1.
As the halogen becomes less electronegative, the halogen
atom natural charge predictably changes to less negative values
for all series of complexes. The metal charge changes even
more dramatically (for instance, from 1.022 for 1F to −0.251
for 1I, a change of more than 1 unit). Counterintuitively, the
ethylene C atoms in the 1X and 2X series are more negatively
charged in the systems with the more electronegative halogen
atoms, where the metal is more positively charged. The change
is particularly pronounced on going from F to Cl and greater
for the 1X system (+0.041) than for the 2X system (+0.021).
This suggests that there is greater metal−ethylene back-
bonding, yielding greater contribution of the platinacyclopro-
pane limiting form, for the fluoride system. This result, contrary
to the expectation based on the halide electronegativity and the
computed charge on the metal atom, is, nevertheless, in
agreement with the computed longer CC distance in 1F and
2F, decreasing on going down the halogen series. Hence, the
fluoride coligands, even though pulling more electron density
away from the metal atom via the σ-bonding framework,
provide more electron density in return via the π mechanism.
In other words, the F− ligands are worse σ-donors, but better π-
donors, than the other halides and allow the metal to provide
more back-bonding to the ethylene ligand. Given the
orientation of the ethylene ligand perpendicular to the
Figure 2. Relative Gibbs free energies of [PtX3(C2H4)]− (1X), trans-
PtX2(C2H4)(PhNH2) (2X), trans-PtX2(PhNH2)2 (3X), and the TOF-
determining transition state (TS2X) for the ethylene/aniline process.
Values are ΔGCPCM (aniline) relative to 1X in kcal mol−1 at 298.15 K
(and at 423.15 K in parentheses).
previously reported investigation had shown that complex 1Br is
the catalyst resting state for the X = Br system.39 The next-
lowest ethylene-containing system was found to be trans-
PtBr2(C2H4)(PhNH2), 2Br, only 4.0 kcal mol−1 higher than the
resting state, whereas trans-PtBr2(PhNH2)2 (3Br) is even closer
at +3.6 kcal mol−1. Because the effect of X on the relative
energy of compounds 1X, 2X, and 3X could not be predicted a
priori, we have also run calculations of all systems 2X and 3X to
ensure that we always consider the correct form of the resting
state. As shown in Figure 2, the lowest-energy species is, in each
case, complex 1X, and this remains true at 150 °C (catalytic
conditions) according to the calculations. Other potential off-
loop intermediates, obtained by ligand-exchange processes from
Table 1. Selected Natural Charge and Bond Distances in Systems 1X, 2X, 3X, and TS2X
a
qPt (NBO)
qX trans (NBO)
qX cis (NBO)
qC (NBO)
Pt−C (d/Å)
CC (d/Å)
1F
1.022
0.269
−0.650
−0.445
−0.382
−0.274
−0.640
−0.430
−0.367
−0.266
−0.523
−0.482
−0.480
−0.478
2.089
2.131
2.145
2.166
1.415
1.410
1Cl
1Br
1I
0.070
1.408
−0.251
qPt (NBO)
1.405
a
qN (NBO)
qX cis (NBO)
qC (NBO)
Pt−C (d/Å)
CC (d/Å)
Pt−N (d/Å)
2F
0.875
0.350
0.216
0.002
−0.842
−0.821
−0.821
−0.822
qPt (NBO)
−0.624
−0.406
−0.338
−0.230
−0.467
−0.446
−0.446
−0.444
2.136
2.158
2.164
2.173
1.400
1.398
1.398
1.398
2.123
2.133
2.137
2.151
2Cl
2Br
2I
b
a
qN (NBO)
qX cis (NBO)
Pt−N (d/Å)
3F
0.754
0.312
−0.815
−0.794
−0.795
−0.796
−0.658
−0.468
−0.406
−0.305
2.063
2.075
2.081
2.092
3Cl
3Br
3I
0.191
−0.006
qPt (NBO)
qX trans (NBO)
qH (NBO)
qC (NBO)
Pt−C (d/Å)
CC (d/Å)
Pt−H (d/Å)
C−H (d/Å)
TS2F
TS2Cl
TS2Br
TS2I
0.873
0.132
−0.629
−0.424
−0.364
−0.261
0.191
0.212
0.209
0.204
−0.661
−0.587
−0.577
−0.569
2.195
2.203
2.212
2.226
1.584
1.571
1.571
1.573
1.435
1.548
1.567
1.580
−0.065
−0.378
a
b
Average value for the cis-X atoms in systems 1X, 2X, and 3X. Average value for the two N atoms in system 3X.
296
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
coordination plane, this interaction occurs via the metal dxz
orbital, where x is the axis along the C2H4(center)-Pt-Xtrans
vector. A qualitative orbital interaction diagram of the π
interaction for the 1X system is shown in Figure 3.
back-bonding as the halogen gets lighter, with the biggest gap
between Cl and F.
For systems 2X, the ligand trans to ethylene (aniline) is
common to all complexes and the cis halogen atoms give rise to
a weaker π effect. The resulting qC and CC distance variations
are smaller, although still in the direction expected for a more
important contribution of the π-bonding mechanism. Counter-
intuitively, the calculated natural charge on the aniline N atoms
in systems 2X and 3X is also more negative for the complexes
containing more electronegative halogen atoms, the biggest gap
being once again between the F and Cl systems, although this
variation is less pronounced relative to that on the ethylene C
atoms in system 1X, which is paralleled by shorter Pt−N
distances. In this case, it does not seem possible to attribute the
observed change to a π-bonding interaction, since aniline is not
a π acid. A σ effect through the hardness/softness concept45−47
could be held responsible for this trend, the harder F atom
rendering the Pt−N bond more ionic/less covalent. The Pt−N
natural charge difference gets smaller on going to the heavier
halide systems, which is accompanied by a lengthening of the
Pt−N bond.
Now the trend observed for the energy change on going
from 1X to 2X can be rationalized. Indeed, the replacement of a
π-donating X− ligand with the π-neutral PhNH2 decreases the
effect of the special π-stabilizing effect of the F− ion and flattens
the relative stability between the species, probably because of a
more balanced compensation between σ and π effects. On
going further from 2X to 3X, the ethylene ligand is replaced by a
second aniline ligand and the residual small π synergism that
was present in 2X is no longer possible. Hence, the fluoride
system is destabilized by a greater amount than the other three
systems, for which the exchange is essentially thermoneutral.
The transition states TS2X were also optimized; see Figure 2.
The free energy relative to the resting state, G(TS2X) − G(1X),
is a direct measure of ΔG‡cycle, related to the turnover frequency
under standard conditions.41,42 This value varies in a non-
monotonous way, increasing from F to Cl, then decreasing
from Cl to Br, and finally increasing again from Br to I. Note
that, for the three systems that were investigated experimentally
(Cl, Br, I), the computational result is in qualitative agreement
with the experimental results shown in Table 2.28 However, we
Figure 3. Orbital interaction responsible for the Pt−C2H4 back-
bonding in complexes 1X (see text). The orbital energies are in
electronvolts. Representative contour diagrams are only shown for the
fluoride system (for the complete set of MOs, see the Supporting
Information).
To a first approximation, only the trans-X atom participates
in the interaction with its pz orbital. The [PtX3]− fragment on
the left of the interaction diagram participates with two filled
orbitals, resulting from the in-phase and out-of-phase
combination of the trans-X pz orbitals and the Pt dxz orbital,
which results in no net X−Pt π-bonding. The greater
electronegativity in the order F > Cl > Br > I is reflected in
an energy decrease along the same sequence, with the exception
of the F system, for which the out-of-phase combination is
raised above those of the other halogen systems, whereas the
in-phase combination is unusually stabilized relative to those of
the other three systems. This is already a clear indication of the
stronger π interaction occurring for the F system. Now, both
orbitals have the correct symmetry to interact and be stabilized
by the ethylene CC π* combination (on the right of the
interaction diagram), a phenomenon widely known as push−
pull interaction.44 The net result is transmission of electron
density from the halogen pz orbital to the ethylene π* orbital
with the consequent increase of C−C distance and natural
charge of the ethylene C atoms. As shown in Figure 3, the
establishment of these interactions clearly leads to a stronger
stabilization for the out-of-phase combination in the case of the
F system. The donor orbital (the out-of-phase Ftrans−Pt xz
combination at −0.093 eV) is much more Pt(dxz) in character
and higher in energy relative to the other halogen systems, thus
resulting in a stronger transfer of electron density from the F
atom pz orbital to the CC π* combination. This effect is partly
counterbalanced by that of the in-phase combination, which
grows in importance on going from F to I because the in-phase
combination becomes more and more Pt(dxz)-like as X
becomes heavier (hence, stronger back-bonding to ethylene),
whereas the out-of-phase combination becomes more and more
X(pz)-like (hence, weaker back-bonding to ethylene). The
overall balance is much in favor of stronger (XtransPt)−(C2H4)
Table 2. Hydroamination of Ethylene by Aniline Catalyzed
a
by [PtBr2]/nBu4PX
total TON 10 equiv total TON 65 equiv
total TON 150
equiv nBu4PX
nBu4PX
nBu4PX
nBu4PX
nBu4PF
nBu4PCl
nBu4PBr
nBu4PI
not studied
not studied
105
not studied
50
80
88
5
150
135
130
100
a
Data from ref 28. Conditions: PtBr2, 0.13 mmol; PhNH2, 45 mmol;
p(C2H4) = 25 bar at 298 K (ca. 100 mmol); 150 °C; 10 h.
must be cautious to make a direct comparison, because only
TON values were reported for the different halide systems. The
ΔG‡
parameter can be related by the Eyring expression to
cycle
the turnover frequency (TOF), which was not measured, and
not to the TON. Given that the TON values were obtained
under the same conditions and after the same period (10 h),
they should be proportional to the TOF, assuming no catalyst
degradation. However, we know that the catalyst is irreversibly
degraded during the catalytic runs to metallic Pt, which is
297
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
inactive for the hydroamination process.26,48 Hence, the greater
TON observed for the Br system might be simply the result of
slower catalyst degradation. Note also that the maximum
activity for each halide system (in terms of TON, Table 2) was
observed for a different excess amount of the activating salt (65
equiv for the chloride, 10 equiv for the bromide and iodide).
The computed values relate, on the other hand, to the standard
conditions (1 M concentrations for all species).
catalytic activities and an identical selectivity could be achieved
by activation with aqueous NaBr.48
Hence, we could proceed to catalytic tests using n-
Bu4NF·3H2O. The experiment, carried out under the classical
conditions, gave disappointing results, which did not
significantly change after drying the salt (Table 3, runs 4 and
5). This result could remain consistent with the high activity
suggested by the calculated ΔG‡
only if the fluoride system
cycle
One relevant and interesting result of the calculations is that
the lowest ΔG‡cycle is, in fact, calculated for the F system, which
has not yet been experimentally investigated (nBu4PF is less
readily available), suggesting that the PtX2/X− catalyst may
have a greater catalytic efficiency when X = F. Note that a
positive effect of fluoride was experimentally demonstrated for
the iridium-catalyzed hydroamination of norbornene by
aniline.49 For this reason, as well as for removing the doubt
about catalyst degradation, new experimental measurements
were carried out for this catalytic process with monitoring of
the kinetic profile, in the presence of an equivalent amount of
all halides, I−, Br−, Cl−, and F− (see next section).
is affected by a more severe catalyst decomposition process or if
fluoride exerts another function in addition to coordinating the
Pt center. The next series of experiments aimed at obtaining the
relative initial activities through a determination of the kinetic
profile. These experiments were carried out under nearly
identical conditions as those of the previous runs (e.g., those
shown in Tables 2 and 3), with double amounts of all reagents
in order to have enough material in the autoclave for multiple
sampling. The only exception was C2H4, which was maintained
at the same pressure. However, its available amount relative to
aniline is still sufficient to satisfy the reaction stoichiometry.
The results are reported in Figure 4.
For the sake of completion, we have also carried out a natural
charge analysis on the TOF-determining transition state TS2X;
see Table 1. The same trends observed above for the halide and
metal charge in 1X are also observed in TS2X. The negative
charge on the Pt-bonded C atom decreases systematically on
moving from F to I, indicating that the alkyl group is still
sufficiently bonded to the metal to feel the competitive
inductive effect of the halide ligands. The H atom moves from
Pt to C as a proton (positive charge, approximately halogen-
independent) and the transition state has an “earlier” character
(the C−H distance becomes longer, whereas the Pt−C and
Pt−H remain approximately unchanged) as the X atoms
become heavier.
b. New Catalytic Studies. A comparison of the four halide
salts is legitimate if all conditions are perfectly identical except
for the nature of the halide ion. Namely, the cation should be
the same. The effect of chloride, bromide, and iodide was
studied using the n-butylphosphonium salts, all being
commercially available (Cl, Br) or easily synthesized (I). In
the case of the fluoride salt, we did not find a convenient access
to nBu4PF, and so we wondered whether the nature of the
cation is really important. A readily available fluoride salt is
nBu4NF·3H2O. To see whether using this salt for comparison
could be justified, we have run a few additional catalytic
experiments with the following bromide salts as catalyst
additives: nBu4PBr (classical conditions),26 n-Bu4NBr, and n-
Bu4NBr·3H2O. These three salts give essentially equivalent
results (see Table 3, runs 1−3), showing that the process is not
Figure 4. Kinetic profiles of the hydroamination of C2H4 by aniline,
catalyzed by PtBr2/nBu4PX (X = Cl, Br, I) or PtBr2/nBu4NF·3H2O.
Conditions: PtBr2, 0.26 mmol; salt additive, 2.6 mmol; PhNH2, 90
mmol; p(C2H4) = 25 bar at 298 K (ca. 100 mmol); 150 °C.
The analysis of these data leads to several considerations.
The first one is the limited catalyst lifetime in each case. At the
end of the catalytic run, massive amounts of both aniline and
ethylene are still present in the autoclave, but the catalytic
activity is reduced essentially to zero after just 6 h, at least for
the most-active Br and I systems. Metallic platinum, as
previously reported,26,48 was found in the product mixture.
The second consideration is that, limited to the heavier
halogens, Cl, Br and I, the initial activity is in qualitative
agreement with the original report by Brunet et al., which was
based on the TON after 10 h (activity in the order Br > I ≫
Cl). Thus, even though the system is heavily affected by catalyst
decomposition, a qualitative direct relationship between initial
TOF and TON exists.26 The third observation is that the
fluoride system, contrary to the computational prediction, is the
least-active system. Its initial slope is not any greater than that
of the chloride system, and furthermore, it seems to deactivate
faster, the percent yield no longer increasing significantly after
45 min, whereas the other halogen systems continue to produce
the N-ethylaniline product. The data, therefore, suggest that
fluoride not only leads to a faster catalyst decomposition but
also exerts another function in addition to coordinating the Pt
center.
Table 3. Hydroamination of Ethylene by Aniline Catalyzed
a
by [PtBr2]/nBu4PX
run
additive
nBu4PBr
TON PhNHEt TON PhNEt2 TON quinaldine
1
2
3
4
5
97
119
104
6
1
2
2
0
0
10
8
nBu4NBr
nBu4NBr·3H2O
nBu4NF
7
2
nBu4NF·3H2O
2
2
a
Conditions: PtBr2, 0.13 mmol; salt additive, 1.3 mmol; PhNH2, 45
mmol; p(C2H4) = 25 bar at 298 K (ca. 100 mmol); 150°C; 10 h.
significantly affected by the cation nature and by the presence
of water. Incidentally, we had already established that similar
To rationalize this result, we consider that one important
difference between F− and the heavier halides is its basicity. We
298
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
have recently shown that the use of a basic amine, notably
NHEt2, leads to a more favorable nucleophilic addition to the
coordinated ethylene, with generation of equilibrium amounts
of the zwitterionic intermediate that could be fully charac-
terized (contrary to aniline for which this intermediate is
unobservable), but the excess amine leads to a deprotonation
equilibrium with the aminoalkyl species, incapable of
proceeding along the hydroamination cycle, and leads to the
system decomposition with generation of metallic platinum, as
also pointed out by other previous related studies,50−52 by a
pathway that is not yet understood.53 Metallic platinum was
proven catalytically inactive for the hydroamination process.
Hence, it may be proposed that the F− ion leads to a faster
catalyst deactivation through its stronger basic character. The
NHEt2-induced zwitterion deprotonation was also analyzed by
DFT calculations.53 We did not consider it useful to carry out
additional calculations in the presence of an extra halide ion.
On the other hand, we checked the hypothesis of a base-related
deactivation through additional experimental studies.
run could be carried out in regular glassware since aniline boils
at 184 °C. The result is dissolution of PtBr2 to yield a
Bordeaux-red solution, from which a powdery brown
precipitate deposited upon cooling. This product may
correspond to (nBu4P)[PtBr3(PhNH2)], previously obtained
by anion exchange from K2PtCl4 in the presence of aniline.40
For the purpose of this test, however, the exact nature of this
product is irrelevant and characterization was not pursued. The
important point is that the solution was filtered and the solid
washed out completely with a variety of solvents (dichloro-
methane, acetone, ethanol...), leaving on the filter only traces of
insoluble residue that could possibly correspond to metallic
platinum. A second experiment was carried out in the presence
of ethylene and in the absence of aniline. Because aniline is also
the catalysis solvent in the classical procedure,26 the test was
carried out by replacing aniline with toluene. The result of this
experiment is the complete dissolution of PtBr2, yielding a clear
red solution and notably containing no visible amounts of black
Pt0. Precipitation by addition of diethyl ether yielded an orange
crystalline product, the 1H NMR of which confirmed that this is
the known nBu4P[PtBr3(C2H4)] by comparison with an
authentic sample.40
As we have recently reported and already mentioned above,
the [PtBr3(C2H4)]− system is also catalytically active when
generated through the use of aqueous NaBr (the K2PtCl4
precatalyst was more conveniently used in this case).48 The
same conditions have now been used in the presence of added
bases, namely, NHEt2, NEt3, and 1,8-diazabicyclo[5.4.0]undec-
7-ene (DBU). As shown by the results in Table 4, the addition
These two tests clearly indicate the thermal stability of PtII
and resistance to reduction in the coordination environment
provided by bromide anions in combination with either
ethylene or aniline. The reduction must, therefore, be a
consequence of the combined action of both reagents. A
working hypothesis is that the base acts on intermediate 4Br
(Table 1), obtained after nucleophilic addition of aniline to the
coordinated ethylene. A logical possibility is deprotonation of
the ammonium function, but other possibilities may also be
considered, such as reversible β-H elimination, followed by
deprotonation of the resulting Pt(II) hydride species, whereas
the deprotonation of the Pt(IV) hydride complex subsequently
formed along the hydroamination cycle (5Br) would not lead to
the generation of metallic platinum. These hypotheses for the
catalyst decomposition will be further probed from the
experimental and computational viewpoints in forthcoming
investigations and published in due course.
Table 4. [K2PtCl4]/NaBr-Catalyzed Hydroamination of
a
Ethylene by Aniline in the Presence of Bases
run
base
TON PhNHEt
TON PhNEt2
TON quinaldine
1
2
3
4
85
1
1
0
0
0
8
∼0.5
0
NHEt2 his
NEt3
0.5
0.4
DBU
0
a
Conditions: K2PtCl4, 0.13 mmol; NaBr, 19.5 mmol; PhNH2, 45
mmol; p(C2H4) = 25 bar at 298 K (ca. 100 mmol); H2O, 15 mL; base,
1.3 mmol; 150 °C; 10 h.
of the base completely suppresses the catalytic activity. This is
also and especially true for bases, such as NEt3 and DBU, that
combine high basicity with poor nucleophilicity. In the absence
of any external base (fluoride or a stronger amine), the catalyst
is still deactivated, though more slowly, by the action of aniline
itself. Fluoride is a weaker base (pKa = 3.18) than aniline (pKa
= 4.62) in water, but its basicity is enhanced by up to 3 orders
of magnitude in lower polarity solvents.54 On the basis of these
results, it is also possible to rationalize the previously
highlighted positive effect of a catalytic amount of a strong
acid as an additive to the PtBr2/Br− system.26 The strong acid
(CF3COOH in the reported study) in the PhNH2-rich medium
d. Effect of the ArNH2 Basicity. As mentioned in the
Introduction, the catalytic activity for the Brunet system, as
measured by the TON obtained under the same conditions for
a variety of substituted anilines ArNH2, is inversely propor-
tional to the amine basicity: 22 for 4-MeOC6H4 (pKa = 5.34),
55 for C6H5 (pKa = 5.08), 90 for 4-ClC6H4 (pKa = 4.15), and
110 for 2-ClC6H4 (pKa = 2.65).26 In light of the discussion
offered in the preceding sections, this trend would simply result
from a different rate of catalyst deactivation, faster for more
basic anilines. We were interested, however, to see whether the
electronic structure of the aniline would also affect the intrinsic
activity in terms of the ΔG‡cycle. Therefore, we have run
additional calculations with two modified anilines, 4-
MeOC6H4NH2 and 4-ClC6H4NH2, restricted to the bromide
system. These calculations included the transition states
TS2BrOMe and TS2BrCl, and the corresponding complexes
trans-[PtBr2(C2H4)(ArNH2)] (2BrOMe with Ar = 4-MeOC6H4
and 2BrCl with Ar = 4-ClC6H4). We did not calculate the
corresponding trans-[PtBr2(ArNH2)2] derivatives (3). The
results are compared with those of the previously calculated
2Br and TS2Br systems in Figure 5. Once again, species 1Br is, in
every case, the most stable one, both at room temperature and
at the temperature used for catalysts (resting state). The
+
generates an equivalent amount of PhNH3 , which buffers the
action of free aniline in the deprotonation of the zwitterionic
intermediate and, therefore, retards the deactivation process.
c. Additional Experimental Mechanistic Studies. To
learn more about how the basicity affects the catalyst stability, a
few additional experiments have been carried out. First, two
control experiments probed the catalyst stability in a mixture
where one of the two reaction components (aniline and
ethylene) is left out. The experiments were run under the same
conditions of time (10 h) and temperature (150 °C) as the
catalytic runs. A first experiment involved warming up PtBr2 in
pure aniline in the presence of nBu4PBr without ethylene. This
calculations yield ΔG‡
values growing with the aniline
cycle
299
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
amination fail when applied to higher olefins. Previous work on
the PtX2/X− catalytic system, however, has shown a relatively
high activity also for the hydroamination of 1-hexene, even
though slightly lower than for the ethylene hydroamination
(under the same conditions, namely, 10 h at 150 °C with 0.3%
catalyst and 65 equiv of nBu4PBr per Pt; 1-hexene yields 57
turnovers27 vs 131 for ethylene26). This activity difference is
not likely due to a difference in catalyst deactivation rate, since
the deactivation process is now known to be related to the
action of an external base (which ethylene and 1-hexene are
not) on the zwitterionic intermediate and the above
comparative productivities were obtained under strictly
identical conditions except for the olefin. Therefore, the
different activity is probably related to a greater activation
barrier (ΔG‡cycle) when going from ethylene to the higher
olefin. It was also shown that the Markovnikov product is
highly favored (95:5),27 and the best results (highest TON)
were obtained, in this case, in the presence of the iodide salt.28
Our aim is to use computational chemistry to compare the
aniline addition to 1-hexene and ethylene with the PtX2/X−
catalytic system. The calculation of the thermodynamic gain of
the reaction (see Table 6) shows that the hydroamination of 1-
hexene is slightly less favorable than that of ethylene and that
the Markovnikov product is slightly favored on the H and G
scales in the gas phase, although the preference reverses in favor
of the anti-Markovnikov product when considering the solvent
effect of aniline by the C-PCM.
Figure 5. Relative Gibbs free energy of the trans-PtBr2(C2H4)-
(ArNH2) (Ar = 4-C6H4OMe, 2BrOMe; C6H5, 2Br; 4-C6H4Cl, 2BrCl) and
the corresponding TOF-determining transition states (TS2BrOMe
,
TS2Br, TS2BrCl). Values are ΔGCPCM (aniline) relative to
[PtBr3(C2H4)]− (1Br) in kcal mol−1 at 298.15 K (and at 423.15 K
in parentheses).
basicity, which agrees with the experimental observed greater
activity for the addition of less basic anilines. Hence, the
observed activity trend may result from an intrinsic aniline
basicity effect on ΔG‡cycle, or from a lower propensity of the less
basic aniline to deactivate the catalyst, or from a combination of
these two effects.
The natural charge analysis of transition states TS2BrCl and
TS2BrOMe (Table 5) do not reveal dramatic changes relative to
1. Markovnikov vs anti-Markovnikov Regioselectivity.
Calculations to address this issue were carried out only for the
bromide system, since the 95:5 regioselectivity is essentially
independent of the nature of the halide.28 Following our
previous work on the catalytic cycle for ethylene (see the
Introduction), at least five 1-hexene complexes are potentially
susceptible to aniline nucleophilic addition, namely, [PtBr3(1-
hexene)]−, cis- and trans-[PtBr2(1-hexene)(PhNH2)], and cis-
and trans-[PtBr2(1-hexene)2].
Table 5. Selected Natural Charge and Bond Distances in
Systems TS2BrCl and TSBrOMe
TS2BrCl
TS2BrOMe
qPt (NBO)
−0.067
−0.359
0.209
−0.066
−0.365
0.208
qtrans‑Br (NBO)
qH (NBO)
The calculations started with complexes [PtBr3(1-hexene)]−
(1Brhex) and trans-[PtBr2(1-hexene)(PhNH2)] (2Brhex), which
represent the 1-hexene analogues of the resting state and the
next most stable olefin-containing complex for the ethylene
system. Relative to [PtBr4]2−, 1-hexene coordination provides a
qC (NBO)
−0.579
1.566
−0.574
1.570
C−H, d (Å)
Pt−H, d (Å)
Pt−C, d (Å)
1.572
1.570
2.211
2.213
smaller stabilization than ethylene by 4.7 kcal mol−1 for 1Br
the corresponding parameters of TS2Br analyzed above in Table
1. In particular, the charge on the Pt center remains
approximately unaffected. The most notable change is the
less negatively charged Br− for the p-chloroaniline derivative
relative to the other two derivatives. As the aniline becomes
more basic, the Pt-bonded C atom engaged in the reductive
elimination becomes less negatively charged, while the
geometry of the PtHC moiety experiences very minor
variations.
hex
(−9.9 kcal mol−1 relative to −14.6 kcal mol−1 for 1Br) and by
3.3 kcal mol−1 for 2Brhex (−7.4 kcal mol−1 relative to −10.7 kcal
mol−1 for 2Br). Calculations on the 1-hexene analogues of the
other, higher-energy systems were not carried out. We recall,
however, that compound trans-[PtBr2(PhNH2)2], 3Br, is placed
at −11.0 kcal mol−1 relative to [PtBr4]2− on the GCPCM
(aniline) scale at 298 K.39 Hence, while 1Br is the resting
state for the ethylene system, the analogous complex 1Br is
hex
slightly less stabilized than 3Br. Calculations, including the
e. Hydroamination of 1-Hexene. As highlighted in a
thermochemical corrections at 423.15 K, however, place
recent review,29 many active catalysts for ethylene hydro-
compound 1Br
0.25 kcal/mol lower in relative
hex
Table 6. Thermodynamic Data for the Hydroamination of Ethylene and 1-Hexene by Aniline in the Gas Phase and in Aniline as
a
Solvent
reaction
ΔS° (e.u.)
ΔH° (kcal/mol)
ΔG° (kcal/mol)
ΔGCPCM (kcal/mol)
b
C2H4 + PhNH2 → PhNHEt
−37.4 (−37.4)
−41.9 (−41.6)
−17.2 (−17.2)
−13.4 (−13.3)
−12.1 (−12.1)
−6.0 (−1.4)
−0.9 (4.4)
0.4 (5.6)
−8.0 (−6.3)
−1.1 (1.0)
−1.5 (0.4)
C4H9CHCH2 + PhNH2 → C4H9CH(NHPh)CH3
C4H9CHCH2 + PhNH2 → C4H9CH2CH2NHPh
−41.8 (−41.9)
a
b
The data refer to 298.15 K, with those at 423.15 K in parentheses. Data from ref 39.
300
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
GCPCM(aniline) than 3Br. Therefore, the resting state of the
syn attacks (Msyn and Asyn). For each type of attack, there are
actually two different pathways, depending on the direction in
which the attached C atom will twist. Only the lower-energy
transition state for each type of attack is shown in Figure 8. The
catalytic cycle for the 1-hexene hydroamination is either 1Br
hex
or the off-loop species 3Br depending on the temperature. It
should also be mentioned that, under conditions similar, but
not totally identical, to those of the hydroamination catalysis,
[PtBr4]2− was recently shown to afford [PtBr5]3− in the
presence of a large excess of a bromide salt and that the latter
species does not react with 1-hexene.55
Because it was shown that the only productive zwitterionic
intermediate for the ethylene system is that derived from the
anionic tribromo complex (Figure 1), we have only examined
the aniline addition to 1Brhex, considering, however, both the
CH2 and the CHBu groups to yield zwitterionic complexes of
Markovnikov (4BrhexM) and anti-Markovnikov (4BrhexA) types.
Several local minima were found for each addition product, the
most stable ones being outlined in Figure 6. Both types of
Figure 8. Views of the lowest-energy transition states for the four
types of nucleophilic addition of PhNH2 to coordinated 1-hexene in
1Brhex. Values of ΔGCPCM (aniline) in kcal mol−1 relative to [PtBr3(1-
hexene)]− (1Brhex) at 298.15 K (and at 423.15 K in parentheses) and
of the calculated imaginary frequency in cm−1 are also shown.
Figure 6. Geometries and relative Gibbs free energies of the most
stable products of PhNH2 addition to the CHBr (4BrhexM) and CH2
(4BrhexA) groups of the coordinated 1-hexene in [PtBr3(1-hexene)]−
(1Brhex). Values are ΔGCPCM (aniline) in kcal mol−1 relative to 1Brhex at
298.15 K (and at 423.15 K in parentheses).
calculations indicate that for each C atom, the anti attack is
favored (the same result was previously found39 for the PhNH2
to ethylene in 1Br) and that the Markovnikov attack is
kinetically preferred by ca. 1 kcal mol−1. Hence, the lowest
transition state is TS1BrhexManti, leading to the Markovnikov
zwitterionic complex 4BrhexM. However, the difference between
the two barriers is irrelevant for the regioselectivity of the
catalysis reaction, because the turnover rate for both
Markovnikov and anti-Markovnikov products is determined
by the final reductive elimination step (vide infra).
The next step in the mechanism is proton transfer from the
ammonium function of the zwitterionic intermediate to the Pt
atom, yielding the 5-coordinate 5BrhexM and 5BrhexA complexes
shown in Figure 9. Like for the zwitterionic precursors 4BrhexM
and 4BrhexA, complexes 5Brhex exhibit a stabilizing intramolecular
N−H···Br interaction and the Markovnikov product is slightly
more stable.
PhNH2 additions are slightly less favorable thermodynamically
relative to the addition to ethylene (18.2 kcal mol−1; see Figure
1), with 4BrhexM being less destabilized than 4BrhexA. The
geometries of the two products are closely related to that of the
ethylene analogue 4Br, with a short N−H···Pt hydrogen bond,
predisposed for proton transfer to the platinum center, and a
less tight hydrogen bond connecting a second N−H bond with
one of the bromine atoms.
We have also analyzed the activation barriers for the aniline
nucleophilic addition (TS1Brhex). The shape of the LUMO of
complex 1Br (Figure 7) clearly suggests that the C2 atom
hex
Starting from these complexes, the product is generated by
C−H reductive elimination through the TOF-determining
transition state TS2Brhex, leading to a σ-complex 6Brhex, where
the newly formed C−H bond remains placed in correspond-
ence to one of the coordination sites of the square-planar PtII
center. All geometries are depicted in Figure 10. For the anti-
Markovnikov product 6BrhexA, no significant interaction with
the C atom is observed but the H atom remains in close contact
with the Pt center (the Pt···H−C angle is more open). Another
peculiarity, revealed by IRC calculations from the optimized
transition states, is that another local minimum exists for the
Markovnikov pathway between the hydride complex 5BrhexM
Figure 7. Three mutually orthogonal views of the LUMO (E = 0.0424
hartree) of complex 1Brhex, with indication of all possible pathways for
nucleophilic aniline addition.
should be more easily attacked by the aniline nucleophile. The
NBO analysis yields the charges of −0.48 for C1 and −0.24 for
C2. Therefore, charge control also favors attack at the C2 atom.
The figure also indicates the four possible pathways for
nucleophilic addition: two anti attacks (Manti and Aanti) and two
301
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
the N-bonded complex switches the relative energy order, the
anti-Markovnikov product becoming now the preferred
regioisomer.
The two competitive pathways are compared in the energy
diagram of Figure 11. Like for the hydroamination of ethylene,
Figure 9. Geometries and relative Gibbs free energies of the PtIV
hydride intermediates 5BrhexM and 5BrhexA. Values are ΔGCPCM
(aniline) in kcal mol−1 relative to [PtBr3(1-hexene)]− (1Brhex) at
298.15 K (and at 423.15 K in parentheses).
Figure 11. Relative Gibbs free energy pathways for the PtBr2/Br−-
catalyzed hydroamination of 1-hexene. Values are ΔGCPCM (aniline) in
kcal mol−1 relative to [PtBr3(1-hexene)]− (1Brhex) at 298.15 K (and at
423.15 K in parentheses).
the calculation shows that the TOF-determining transition state
is that of the reductive elimination process forming the C−H
bond between the aminoalkyl and hydride ligands. The
nucleophilic addition step occurs with a transition state
TS1Br
at lower energy. Therefore, the different energy
hex
barriers at the nucleophilic addition step level have no kinetic
consequence. The only interesting difference is the relative
thermodynamic stability of the zwitterionic products (4BrhexM
is present at greater relative concentration than its isomer
4BrhexA under catalytic conditions). The overall rate is
determined by the free energy difference between TS2Br
hex
and the resting state, which is calculated as trans-
PtBr2(PhNH2)2 (although complex 1Br
is calculated as
hex
slightly more stable at 150 °C). Because the interesting
CPCM
parameter here is the ΔΔG‡
cycle
for the two regioisomers
and the resting state is shared by the two cycles, whether the
resting state is trans-PtBr2(PhNH2)2 or [PtBr3(1-hexene)]− (or
even [PtBr5]3−),55 the result does not change. From this
Figure 10. Geometries and relative Gibbs free energies of the
transition states and products, TS2Br and 6Brhex, of the C−H
hex
ΔΔG‡ CPCM (2.6 and 3.7 kcal mol−1 at 298.15 and 423.15 K,
reductive elimination from the hydride complexes 5Brhex, as well as of
cycle
the products of rearrangement of the N-1- or N-2-hexylaniline, 7Br
.
respectively) an expected M/A ratio of 98.8:1.2 can be
predicted at both temperatures (the greater energy difference
at the higher temperature is balanced off exactly by the different
RT term), which can be considered in qualitatively good
agreement with the observed 95:5 ratio.28
hex
Values are ΔGCPCM (aniline) in kcal mol−1 relative to [PtBr3(1-
hexene)]− (1Brhex) at 298.15 K (and at 423.15 K in parentheses).
and the transition state, characterized by a trigonal-bipyramidal
geometry with equatorial Br, H, and CH2CH(Bu)NHPh
ligands, at higher energy (31.2 kcal mol−1) than the square-
pyramidal isomer 5BrhexM. This type of isomer could not be
located for the anti-Markovnikov regioisomer. For both
regiochemistries, the N−H···Br interaction persists from the
In conclusion for this section, the results of the calculations
based on the previously proposed mechanism39 are in
agreement with a number of experimental observations. (i)
The catalytic activity for the hydroamination of 1-hexene
(TON = 37)27 is lower than that for ethylene (TON = 92)26
under the same conditions, consistent with a greater ΔG‡cycle in
the former case. The difference between the two barriers at 150
°C (37.7 kcal mol−1 for the Markovnikov addition to 1-hexene
and 35.6 kcal mol−1 for addition to ethylene) is actually too
large to justify such a small activity difference, but it is in the
right direction. Besides computational errors, a possible reason
leading to an artificially smaller activity difference is a faster
catalyst degradation in the presence of ethylene relative to 1-
starting hydride complex 5Br through the transition state
hex
TS2Br
and is still found in the product 6Brhex. The
hex
hydroamination product, N-1- or N-2-hexylaniline, is weakly
bonded to the Pt center through the C−H bond and can
rearrange to yield a complex with an N-bonded hydroamination
product, 7Brhex, shown again in Figure 10 for both
regiochemistries. The Markovnikov pathway has the lower-
energy transition state and σ-complex, but the rearrangement to
302
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
hexene, but this does not seem likely on the basis of the
considerations already made above. (ii) The catalytic activity
for addition to 1-hexene is greater for the Markovnikov
addition mode, with the calculated ΔΔGCPCM qualitatively
agreeing with the observed 95:5 ratio.
3. CONCLUSIONS
Application of the catalytic mechanism recently established by a
DFT study for the PtBt2/Br−-catalyzed N−H addition of
aniline to ethylene39 has provided the following results: (i) the
activation barrier (ΔG‡cycle) for the C2H4/PhNH2 reaction
catalyzed by PtX2/X− is similar for Br, Cl, and I, though
marginally smaller for the Br system, which affords the highest
activity; (ii) the ΔG‡cycle slightly increases as the aniline basicity
increases (p-chloroaniline < aniline < p-methoxyaniline); (iii)
the ΔG‡cycle for the 1-hexene/PhNH2 reaction is slightly greater
2. Halide Effect. Like for the aniline addition to ethylene,
we applied the computational approach to analyze the ΔG‡
cycle
for the addition of aniline to 1-hexene catalyzed by different
PtX2/X− systems. Calculation of the various [PtX3(1-hexene)]−
(1Xhex) and comparison with the trans-PtX2(PhNH2)2 (3X)
systems already discussed above reveal that the lowest-energy
system (resting state) is the former for X = F and the latter for
the other three halogens (see Figure 12). The much greater
than that for the C2H4/PhNH2 reaction; (iv) ΔG‡
for the
cycle
Markovnikov addition of aniline to 1-hexene is lower than that
for the anti-Markovnikov addition; and (v) the ΔΔG‡
(M/
cycle
A) for the 1-hexene/PhNH2 reaction is relatively halide-
independent. Some of these results are in reasonable agree-
ment, at least qualitatively, with the experimental evidence
(notably, the marked preference for the Markovnikov addition),
whereas others are less reliable or in fortuitous agreement
because of the small computed differences where the
computational errors may not perfectly cancel out,56,57 and
because the observed trends in catalytic activity for the
previously published studies (expressed by the TON) are
affected by catalyst decomposition, which may not systemati-
cally be in direct relation with the turnover frequency.
Overall, it seems that all the available experimental results for
the Pt-catalyzed olefin hydroamination are reasonably well
rationalized by the proposed mechanism (Figure 1),39 after
allowance is made for the base-induced catalyst decomposition.
For instance, decomposition seems to negatively affect the
catalytic performance in the presence of fluoride ions, contrary
to the computational prediction. The mechanism of Figure 1
may be considered a good basis for further mechanistic
elaborations and catalyst design, which also requires, however, a
fuller understanding of the catalyst degradation mechanism.
Additional studies into the catalyst degradation, learning how
to prevent it or retard it, are warranted. It is also interesting to
Figure 12. Relative Gibbs free energies for systems [PtX3(1-hexene)]−
(1Xhex), trans-PtX2(PhNH2)2 (3X), and the transition states of the rate-
determining C−H reductive elimination steps (TS2Xhex) for both
Markovnikov and anti-Markovnikov products related to the PtX2/X−-
catalyzed hydroamination of 1-hexene. Values are ΔGCPCM (aniline) in
kcal mol−1 relative to the resting state in each case, at 298.15 K (and at
423.15 K in parentheses).
stabilization of 1Xhex relative to 3X for the fluoride system can be
rationalized in the same way as that of 1X, since the Pt−C2H4
and Pt−hexene bonds have similar constitutions.
conceive other coordination geometries allowing a ΔG‡
cycle
reduction, namely, a reduction of the free energy difference
between the resting state and the transition state of the TOF-
determining C−H reductive elimination step. Work in these
directions is currently ongoing.
The ΔGCPCM values of the TOF-determining transition state
for the two regiochemistries, TS2XhexM and TS2XhexA, are also
reported in Figure 12 relative to the resting state. The
ΔGCPCM(TS2XhexM) − ΔGCPCM(TS2XhexA) value is approx-
imately constant (between 2.6 kcal mol−1 for the Br system and
3.8 kcal mol−1 for the I system), in agreement with a small
effect of the halide nature on the M/A ratio. When considering
4. EXPERIMENTAL SECTION
a. Computational Details. All geometry optimizations were
carried out by the DFT approach with the Gaussian 03 suite of
programs58 using the B3LYP functional.59−61 All atoms except Pt and I
were described by the standard 6-31+G* basis set, which includes both
polarization and diffuse functions that are necessary to allow angular
and radial flexibility to the highly anionic systems. The I atom was
described by the LANL08d basis and the Pt atom by the LANL2TZ(f)
basis, which are uncontracted versions of LANL2DZ and include an
ECP.62 The I basis set contains a d polarization and a diffuse function
and the Pt basis set contains an f polarization function. The geometry
optimizations were carried out on isolated molecules in the gas phase.
Transition states were optimized using the synchronous transit-guided
quasi-Newton (STQN) method in the gas phase. Frequency
calculations were carried out for all optimized geometries in order
to verify their nature as local minima, for the calculation of
thermodynamic parameters at 298.15 K and at 423.15 K under the
ideal gas and harmonic approximations, and for the identification of all
transition states (one imaginary frequency). Solvent effects were
introduced by means of C-PCM63,64 single-point calculations on the
gas-phase optimized geometries in aniline (ε = 6.89), which was
considered to best approximate the reaction solvent. The solvent
cavity was created by a series of overlapping spheres by the default
only the heavier halides, the ΔG‡
for the main
cycle
(Markovnikov) pathway is greater for the Cl system, and
smaller and essentially identical for the Br and I systems,
whereas the iodide system was experimentally found to yield a
greater TON. As already concluded above in sections 2a−2c,
the experimental result is largely affected by the rate at which
the catalyst decomposes, whereas the computational limitations
and approximations make the ΔG‡
differences between the
cycle
different halide systems little reliable. Hence, a perfect match
between the trends in ΔG‡cycle and the observed TON is not to
be expected. The results of the calculations qualitatively fit with
the observation of a relatively halide-independent M/A
regioselectivity, but a finer analysis of such small effects is
probably meaningless.
303
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
UA0 model (or SPHEREONH for the delocalized proton atoms), and
all standard settings as implemented in Gaussian 03 were used for the
C-PCM calculations. The reaction free energy changes in solution
were calculated as ΔGsolut = ΔGgas + Δ(ΔGsolv) and corrected for the
change of standard state from the gas phase (1 atm) to solution (1
M).65 The approach of optimizing the geometries in the gas phase and
then keeping the geometry frozen for the C-PCM calculation was
preferred because of convergence problems of the geometry
optimization in the presence of the C-PCM, especially for molecules
containing weak interactions (i.e., hydrogen bonds).
b. Catalytic Runs. 1. Yield vs Time Monitoring. Catalytic
experiments were conducted in a 100 mL stainless steel autoclave
without a glass liner under magnetic stirring. In a typical procedure,
the autoclave was charged with PtBr2 (0.26 mmol) and the appropriate
salt additive (2.6 mmol), closed, and submitted to vacuum/argon
cycles. Aniline (8.3 mL, 90 mmol) was added to the autoclave by a
syringe through a valve equipped with a septum and the ethylene
pressure adjusted to 25 bar (ca. 100 mmol). The temperature was then
raised to 150 °C. Samples were withdrawn at regular time intervals,
treated with the external standard (N,N-dibutylaniline), and extracted
with diethyl ether. The organic layer was analyzed by GC.
2. In the Presence of Added Base. In a typical procedure, K2PtCl4
(0.13 mmol) and NaBr (19.5 mmol) were introduced in the autoclave,
which was then closed and submitted to vacuum/argon cycles.
Subsequently, water (15 mL), aniline (4.15 mL, 45 mmol), and
additional base (1.3 mmol) were added to the autoclave by a syringe
through a valve equipped with a septum, and the ethylene pressure was
adjusted to 25 bar (ca. 100 mmol). The temperature was then raised to
150 °C. After 10 h, the autoclave was allowed to cool to room
temperature and slowly vented. The entire reaction mixture was
treated with the external standard (N,N-dibutylaniline) and extracted
with dichloromethane (ca. 60 mL). The organic layer was analyzed by
GC.
(5) Khedkar, V.; Tillack, A.; Benisch, C.; Melder, J. P.; Beller, M. J.
Mol. Catal. A 2005, 241, 175−183.
(6) Ryu, J. S.; Li, G. Y.; Marks, T. J. J. Am. Chem. Soc. 2003, 125,
12584−12605.
(7) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004, 37, 673−686.
(8) Yang, L.; Xu, L. W.; Zhou, W.; Gao, Y. H.; Sun, W.; Xia, C. G.
Synlett 2009, 1167−1171.
(9) Gardner, D. M.; Clark, R. T. U.S. Patent 4,454,321, 1984.
(10) Gardner, D. M.; Clark, R. T. European Patent 39,061, 1981.
(11) Schaffrath, H.; Keim, W. J. Mol. Catal. A 2001, 168, 9−14.
(12) Yi, C. S.; Yun, S. Y. Org. Lett. 2005, 7, 2181−2183.
(13) Krukowka, E.; Taube, R.; Steinborn, D. DD Patent 296,909,
1991.
(14) Coulson, D. R. Tetrahedron Lett. 1971, 429−430.
(15) Coulson, D. R. U.S. Patent 3,758,586, 1973.
(16) Diamond, S. E.; Mares, F.; Szalkiewicz, A. Fundam. Res.
Homogeneous Catal. 1979, 3, 345−358.
(17) Diamond, S. E.; Mares, F. U.S. Patent 4,215,218, 1980.
(18) Giner, X.; Najera, C. Synlett 2009, 3211−3213.
(19) Liu, X. Y.; Li, C. H.; Che, C. M. Org. Lett. 2006, 8, 2707−2710.
(20) Zhang, J. L.; Yang, C. G.; He, C. J. Am. Chem. Soc. 2006, 128,
1798−1799.
(21) Zhang, Z. B.; Lee, S. D.; Widenhoefer, R. A. J. Am. Chem. Soc.
2009, 131, 5372−5373.
(22) Giner, X.; Najera, C. Org. Lett. 2008, 10, 2919−2922.
(23) Wang, X.; Widenhoefer, R. A. Organometallics 2004, 23, 1649−
1651.
(24) Karshtedt, D.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2005,
127, 12640−12646.
(25) McBee, J. L.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2008,
130, 16562−16571.
(26) Brunet, J. J.; Cadena, M.; Chu, N. C.; Diallo, O.; Jacob, K.;
Mothes, E. Organometallics 2004, 23, 1264−1268.
(27) Brunet, J. J.; Chu, N. C.; Diallo, O. Organometallics 2005, 24,
3104−3110.
ASSOCIATED CONTENT
* Supporting Information
■
S
(28) Rodriguez-Zubiri, M.; Anguille, S.; Brunet, J.-J. J. Mol. Catal. A
2007, 271, 145−150.
MO contour plots of all orbitals shown in Figure 3 and
Cartesian coordinates for all new optimized structures. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(29) Brunet, J.-J.; Chu, N.-C.; Rodriguez-Zubiri, M. Eur. J. Inorg.
Chem. 2007, 4711−4722.
(30) Jeffery, T. Tetrahedron 1996, 52, 10113−10130.
(31) Amatore, C.; Jutand, A. J. Organomet. Chem. 1999, 576, 254−
278.
AUTHOR INFORMATION
Corresponding Author
*Fax: (+) 33-561553003. E-mail: rinaldo.poli@lcc-toulouse.fr.
■
(32) Handy, S. T.; Okello, M. Tetrahedron Lett. 2003, 44, 8395−
8397.
(33) Maitlis, P. M.; Haynes, A.; James, B. R.; Catellani, M.; Chiusoli,
G. P. Dalton Trans. 2004, 3409−3419.
(34) Senn, H. M.; Blochl, P. E.; Togni, A. J. Am. Chem. Soc. 2000,
122, 4098−4107.
ACKNOWLEDGMENTS
■
We are grateful to the ANR (Agence Nationale de la
Recherche, Grant No. NT09_442499), to the CNRS (Centre
National de la Recherche Scientifique), and to the IUF (Institut
Universitaire de France) for support of this work. This work
was granted access to the HPC resources of CINES under the
allocation 2010-086343 made by GENCI (Grand Equipement
National de Calcul Intensif) and to the resources of the CICT
(Centre Interuniversitaire de Calcul de Toulouse, project
(35) Tsipis, C. A.; Kefalidis, C. E. J. Organomet. Chem. 2007, 692,
5245−5255.
(36) Kovacs, G.; Ujaque, G.; Lledos
853−864.
(37) Kovacs, G.; Lledos, A.; Ujaque, G. Organometallics 2010, 29,
5919−5926.
́
, A. J. Am. Chem. Soc. 2008, 130,
(38) Vo, L. K.; Singleton, D. A. Org. Lett. 2004, 6, 2469−2472.
(39) Dub, P. A.; Poli, R. J. Am. Chem. Soc. 2010, 132, 13799−13812.
(40) Dub, P. A.; Rodriguez-Zubiri, M.; Daran, J.-C.; Brunet, J.-J.; Poli,
R. Organometallics 2009, 28, 4764−4777.
CALMIP). P.A.D. thanks the MENESR (Minister
̀
e de
́
l′Education Nationale, de l′Enseignement Super
́
ieur et de la
(41) Poli, R. Comments Inorg. Chem. 2009, 30, 177−228.
(42) Kozuch, S.; Shaik, S. Acc. Chem. Res. 2011, 44, 101−110.
(43) Dub, P. A.; Poli, R. J. Mol. Catal. A 2010, 324, 89−96.
(44) Caulton, K. G. New J. Chem. 1994, 18, 25−41.
(45) Drago, R. S.; Wong, N. M. Inorg. Chem. 1995, 34, 4004−4007.
(46) Liu, G.-H.; Parr, R. G. J. Am. Chem. Soc. 1995, 117, 3179−3188.
(47) Pearson, R. G. Coord. Chem. Rev. 1990, 100, 403−425.
(48) Dub, P. A.; Rodriguez-Zubiri, M.; Baudequin, C.; Poli, R. Green
Chem. 2010, 1392−1396.
Recherche, France) for a Ph.D. fellowship.
REFERENCES
■
(1) Muller, T. E.; Hultzsch, K. C.; Yus, M.; Foubelo, F.; Tada, M.
̈
Chem. Rev. 2008, 108, 3795−3892.
(2) Haggins, J. Chem. Eng. News 1993, 71, 23−27.
(3) Brunet, J.-J.; Neibecker, D. In Catalytic Heterofunctionalization;
Togni, A., Grutzmacher, H., Eds.; VCH: Weinheim, Germany, 2001;
̈
(49) Dorta, R.; Egli, P.; Zurcher, F.; Togni, A. J. Am. Chem. Soc. 1997,
119, 10857−10858.
pp 91−141.
(4) Borman, S. Chem. Eng. News 2004, 82, 42−43.
304
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305

Organometallics
Article
(50) Pryadun, R.; Sukumaran, D.; Bogadi, R.; Atwood, J. D. J. Am.
Chem. Soc. 2004, 126, 12414−12420.
(51) Balacco, G.; Natile, G. J. Chem. Soc., Dalton Trans. 1990, 3021−
3024.
(52) Al-Najjar, I. M.; Green, M. J. Chem. Soc., Dalton Trans. 1979,
1651−1656.
(53) (a) Dub, P. A.; Daran, J.-C.; Levina, V. A.; Belkova, N. V.;
Shubina, E. S.; Poli, R. J. Organomet. Chem. 2011, 696, 1174−1183.
(b) Dub, P. A.; Bet
5167−5172.
́
hegnies, A.; Poli, R. Eur. J. Inorg. Chem. 2011,
(54) Landini, D.; Maia, A.; Rampoldi, A. J. Org. Chem. 1989, 54,
328−332.
́ ́ ́
(55) Aullon, G.; Gomez, K.; Gonzalez, G.; Jansat, S.; Martínez, M.;
Poli, R.; Rodríguez-Zubiri, M. Inorg. Chem. 2011, 50, 5628−5636.
(56) Wheeler, S. E.; Moran, A.; Pieniazek, S. N.; Houk, K. N. J. Phys.
Chem. A 2009, 113, 10376−10384.
(57) Pieniazek, S. N.; Clemente, F. R.; Houk, K. N. Angew. Chem., Int.
Ed. 2008, 47, 7746−7749.
(58) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J.
B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev,
O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.;
Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,
J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision
C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(59) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.
(60) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785−
789.
(61) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett.
1989, 157, 200−206.
(62) Roy, L. E.; Hay, P. J.; Martin, R. L. J. Chem. Theory Comput.
2008, 4, 1029−1031.
(63) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995−2001.
(64) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669−681.
(65) Winget, P.; Cramer, C. J.; Truhlar, D. G. Theor. Chem. Acc.
2004, 112, 217−227.
305
dx.doi.org/10.1021/om2009149 | Organometallics 2012, 31, 294−305