Competitive benzene c-h bond activation versus olefin insertion in a (monomethyl)palladium(ii) β-diketiminate complex

Article abstract of DOI:10.1021/om900171d

The monomethylpalladium(II) complex (COD)Pd(CH₃)Cl (COD = 1,5-cyclooctadiene) is found to undergo both benzene C-H activation and migratory insertion of olefin, with the former faster than the latter, at room temperature in the presence of an a

Full text of DOI:10.1021/om900171d

4400 Organometallics 2009, 28, 4400–4405
DOI: 10.1021/om900171d
Competitive Benzene C-H Bond Activation versus Olefin Insertion in a
(Monomethyl)palladium(II) β-Diketiminate Complex
Bo-Lin Lin, Koyel X. Bhattacharyya, Jay A. Labinger,* and John E. Bercaw*
Arnold and Mabel Beckman Laboratories of Chemical Synthesis, California Institute of Technology,
Pasadena, California 91125
Received March 3, 2009
The monomethylpalladium(II) complex (COD)Pd(CH3)Cl (COD=1,5-cyclooctadiene) is found to
undergo both benzene C-H activation and migratoryinsertion of olefin, withthe former faster than the
latter, at room temperature in the presence of an anionic β-diketiminate ligand, to yield η3-(6-R-
cyclooctenyl)palladium(II) β-diketiminate (R=methyl or phenyl). The reaction is proposed to take
place via the formation of a (monomethyl)palladium(II) β-diketiminate with COD as the fourth ligand,
followed by competitive benzene C-H activation and migratory insertion of olefin. The proposed
mechanism is supported by density functional theory calculations upon a simplified model system.
Coupling of arenes and olefins via aromatic C-H activa-
tion by transition metal complexes provides an efficient and
atom-economic way to construct C-C bonds.1 Most arene-
olefin coupling reactions are limited to activated olefins and
arenes with chelation assistance;2 however, there have been
several recent reports of hydroarylation involving simple
arenes (e.g., benzene and toluene) and nonactivated olefins
at elevated temperatures, using catalysts based on Ir, Ru, and
Pt.3 In general, these reactions proceed by aromatic C-H
activation followed by migratory insertion of olefin into the
resulting metal-aryl bond, to generate a β-aryl-alkyl com-
plex; this key intermediate can undergo aromatic C-H
activation to liberate the C-C coupled product and regen-
erate the aryl complex.4 However, it can also undergo
additional olefin insertions, so that olefin oligmerization/
polymerization can be a competing side reaction (Scheme 1).
Understanding the factors controlling the relative rates of
these two reactions, which determine the selectivity for olefin
Scheme 1
hydroarylation,3,4 is essential for rational design of improved
catalysts. (Side reactions presumably involving β-hydride
elimination of the β-aryl-alkyl complex or olefinic C-H
activation have also been observed under harsh conditions,
e.g., 180 °C reaction temperature).3d,3e
The results for platinum suggest that palladium might also
be a promising metal for catalyzing olefin hydroarylation,
and indeed C-H activation of simple arenes by mono-
methylpalladium(II) has been demonstrated.5 However, re-
lated monomethylpalladium(II) complexes often are highly
active for olefin oligomerization/polymerization,6 so the
competition of Scheme 1 may well be unfavorable in many
cases. Herein, we report a monomethylpalladium(II) system
for which benzene C-H activation and migratory insertion
of olefins are competitive, with the former faster at room
temperature, along with experimental and computational
findings relevant to the mechanism.
*Corresponding author. E-mail: jal@caltech.edu; bercaw@caltech.
edu.
(1) Dyker, G., Ed. Handbook of C-H Transformations; Wiley-VCH:
Weinheim, Germany, 2005.
(2) See the following reviews and references therein: (a) Jia, C. G.;
Kitamura, T.; Fujiwara, Y. Acc. Chem. Res. 2001, 34, 633–39.
(b) Kakiuchi, F.; Murai, S. Acc. Chem. Res. 2002, 35, 826–834. (c) Beccalli,
E. M.; Broggini, G.; Martinelli, M.; Sottocornola, S. Chem. Rev. 2007, 107,
5318–5365.
Results and Discussion
(3) Ir: (a) Matsumoto, T.; Taube, D. J.; Periana, R. A.; Taube, H.;
Yoshida, H. J. Am. Chem. Soc. 2000, 122, 7414–7415. (b) Matsumoto, T.;
Periana, R. A.; Taube, D. J.; Yoshida, H. J. Mol. Catal. A: Chem. 2002, 180,
1-18. Ru: (c) Lail, M.; Arrowood, B. N.; Gunnoe, T. B. J. Am. Chem. Soc.
2003, 125, 7506–7507. (d) Foley, N. A.; Lail, M.; Lee, J. P.; Gunnoe, T. B.;
Cundari, T. R.; Petersen, J. L. J. Am. Chem. Soc. 2007, 129, 6765–6781.
(e) Foley, N. A.; Ke, Z. F.; Gunnoe, T. B.; Cundari, T. R.; Petersen, J. L.
Organometallics 2008, 27, 3007-3017. Pt: (f) Karshtedt, D.; McBee, J. L.;
Bell, A. T.; Tilley, T. D. Organometallics 2006, 25, 1801–1811.
(g) McKeown, B. A.; Foley, N. A.; Lee, J. P.; Gunnoe, T. B. Organometallics
2008, 27, 4031–4033. (h) Luedtke, A. T.; Goldberg, K. I. Angew. Chem., Int.
Ed. 2008, 120, 7808–7810.
Addition of lithium β-diketiminate 2 to a CD2Cl2 solu-
tion containing 1 equiv of (COD)Pd(CH3)Cl (1, COD=1, 5-
(5) Ackerman, L. J.; Sadighi, J. P.; Kurtz, D. M.; Labinger, J. A.;
Bercaw, J. E. Organometallics 2003, 22, 3884–3890.
(6) For a reviewing article, see: (a) Ittel, S. D.; Johnson, L. K.;
Brookhart, M. Chem. Rev. 2000, 100, 1169–1203. For several recent
examples, see: (b) Kochi, T.; Nakamura, A.; Ida, H.; Nozaki, K. J. Am.
Chem. Soc. 2007, 129, 7770–7771. (c) Luo, S.; Vela, J.; Lief, G. R.; Jordan,
R. F. J. Am. Chem. Soc. 2007, 129, 8946–8947. (d) Kochi, T.; Noda, S.;
Yoshimura, K.; Nozaki, K. J. Am. Chem. Soc. 2007, 129, 8948–8949.
(e) Weng, W.; Shen, Z.; Jordan, R. F. J. Am. Chem. Soc. 2007, 129, 15450–
15451.
(4) (a) Oxgaard, J.; Muller, R. P.; Goddard, W. A.; Periana, R. A.
J. Am. Chem. Soc. 2004, 126, 352–363. (b) Oxgaard, J.; Periana, R. A.;
Goddard, W. A. J. Am. Chem. Soc. 2004, 126, 11658–11665.
r
pubs.acs.org/Organometallics
Published on Web 07/10/2009
2009 American Chemical Society

Article
Organometallics, Vol. 28, No. 15, 2009 4401
˚
Figure 1. X-ray crystal structure of compound 3 (CCDC 295981). Selected bond distances (A): Pd(1)-N(1), 2.0708(10); Pd(1)-C(12),
2.0932(16); Pd(1)-C(13), 2.1696(11).
Scheme 2
Scheme 3
cyclooctadiene), at room temperature under nitrogen, results
in a yellow solution with some black precipitate. Crystals ob-
tained from solution were identified by 1H NMR and X-ray
crystallography as η3-(6-methylcyclooctenyl)palladium(II)
β-diketiminate (3, Scheme 2). The 1H NMR spectrum, par-
ticularly the characteristic triplet, singlet, and doublet at
δH =5.03 (the proton at the central carbon of the η3-allyl
fragment), 4.53 (the proton at the central carbon of the
β-diketiminate backbone), and 0.43 (the proton of the
methyl group on the cyclooctenyl moiety) respectively,7
indicates that the complex is symmetrical about the plane
bisecting the diketiminate ligand, while the crystal structure
(Figure 1) shows that the methyl group on the cyclooctenyl
moiety is cis to Pd.
(7) Synthesis of (η3-propenyl)palladium(II) β-diketiminates have re-
cently been reported: Tian, X.; Goddard, R.; Porschke, K. R. Organo-
metallics 2006, 25, 5854–5862.
A mechanism for the formation of 3 is proposed in
Scheme 3. Salt metathesis between (COD)Pd(CH3)Cl and

4402 Organometallics, Vol. 28, No. 15, 2009
Lin et al.
Scheme 4
Scheme 5
Scheme 6
lithium β-diketiminate affords COD-coordinated mono-
methylpalladium(II) β-diketiminate (4). Migratory insertion
of COD into the Pd-CH3 bond leads to η1-methyl-
cyclooctenylpalladium(II), which further undergoes a series
of β-hydride eliminations and migratory insertions to form
the η3-allyl species 3, presumably thermodynamically pre-
ferred among all the isomers in the ring-walking process.
The cis conformation of 3 suggests that dissociation of
coordinated olefins from the Pd(olefin)(H) intermediates
does not occur.
the molecular cation and the cation of the molecule minus
the phenylcyclooctenyl moiety.
1
Because sharp H NMR signals are observed for N-aryl
methyl groups of both 3 and 5, rotation of the allyl moiety
appears to be slow on the NMR time scale, which is in sharp
contrast to many other (η3-allyl)palladium(II) systems.8
Slow apparent rotation has also been recently noted for
(η3-propenyl)palladium(II) β-diketiminates.7
The ratio of products 3 and 5, by NMR, is approximately
1:6 for the reaction in C6H6 and 1:2 in C6D6. Since these
ratios reflect the relative rates of benzene C-H/C-D activa-
tion by monomethylpalladium(II) and insertion of COD
into the Pd-CH3 bond, and because the latter is isotope
insensitive, the ratio (∼3) reflects the kinetic isotope effect
(KIE) for benzene C-H activation. This suggests that C-H
bond breaking is the rate-determining step in the formation
of 5. Scheme 5 shows the proposed mechanism: after forma-
tion of intermediate 4, displacement of COD by ben-
zene takes place, followed by C-H activation to form
phenylpalladium(II) and release methane. The resultant
phenylpalladium(II) then undergoes migratory insertion
with COD and forms an η1-(phenylcyclooctenyl)palladium-
(II); the ring-walking process discussed above leads to the
η3-allyl species 5.
When the reaction is carried out in benzene (C6H6)
1
instead of CD2Cl2, some 3 still forms, but the H NMR
indicates the presence of a new species 5 as well. The new 1H
NMR signals assigned to 5 (in CD2Cl2) are similar to those
of 1, with a triplet at δH 5.42 and a singlet at δH 4.76; there
are also several new signals in the aromatic region. If the
same reaction is carried out in a closed NMR tube
using C6D6 as the solvent, formation of CH3D was ob-
served, indicating benzene C-D bond activation by
monomethylpalladium(II), and the new aromatic signals
are absent. (The formation of a small amount of CH4 is
observed, presumably due to competing C-H bond activa-
tion of COD or β-diketiminate.) These findings strongly
suggest that 5 is the phenyl analogue of 3, η3-(6-phenyl-
cyclooctenyl)palladium(II) β-diketiminate (Scheme 4). The
assignment is further supported by high-resolution FABþ
mass spectroscopy, which gives signals corresponding to
(8) For a review involving discussions of apparent rotation in (η3-
allyl)palladium(II) systems and its mechanisms, see: Trost, B. M.;
Van Vranken, D. L. Chem. Rev. 1996, 96, 395–422.

Article
Organometallics, Vol. 28, No. 15, 2009 4403
˚
Figure 2. X-ray crystal structure of compound 6 (CCDC 296075). Selected bond distances (A): Pd(1)-N(3), 2.0135(9); Pd(1)-N(1),
2.0262(8); Pd(1)-C(24), 2.0526(9); Pd(1)-N(2), 2.1113(7).
reductive elimination. PdIV-H is generally considered to be a
Scheme 7
high-energy intermediate, which might argue against the
latter, but it cannot be firmly excluded. To compare these
two mechanisms, density functional theory (DFT) calcula-
tions at the level of rB3LYP/LACVP**11 were performed
on a model system with a simplified β-diketiminate ligand.
The computational results suggest that σ-bond metathesis
is indeed preferred over oxidative addition for the conver-
sion of 7 to 8, with the activation free energy of the former
(21.7 kcal/mol) lower than that of the latter by 9.6 kcal/mol
and the product, methane adduct 8, slightly less stable (by
1.3 kcal/mol) than benzene adduct 7.
The competition between benzene C-H activation via
σ-bond metathesis (TSσ) and migratory insertion (TSin) was
subsequently investigated by DFT for the same system, using
cis-2-butene as a simplified model for COD. Coordination
of benzene is calculated to be 7.1 kcal/mol weaker than that
of cis-2-butene (9 þ C6H6 to 7 þ cis-2-butene, Scheme 7).
If 2 equiv of acetonitrile is added to the reaction mixture
Although a transition state for associative displacement of
containing 1 and 2 in benzene, neither 3 nor 5 is formed;
instead monomethylpalladium(II)(acetonitrile) complex 6 is
obtained nearly quantitatively (Scheme 6). The crystal struc-
ture of 6 (Figure 2) shows that (in contrast to 3 and 5) the
cis-2-butene by benzene (a mechanism tentatively suggested
by experimental observations)12 was not located, the activa-
tion barrier for the dissociative mechanism, which should
be approximately equal to the dissociation free energy of
coordinated cis-2-butene from 7, 8.0 kcal/mol, can be taken
coordination plane is a plane of symmetry, consistent with
1
the H NMR spectrum. Related monomethylpalladium(II)
as an upper limit. Hence displacement of olefin by benzene
β-diketiminates have been previously reported.7,9 Although
6 is fairly stable in C6D6 at room temperature, it decomposes
should be fast compared to breaking the benzene C-H bond,
which is consistent with the experimentally observed KIE of
about 3.
to a mixture of uncharacterized products at 80 °C with the
formation of a significant amount of CH3D, indicative of the
activation of a benzene C-D bond. This is in sharp contrast
to a related acetonitrile-ligated monomethylplatinum(II)
complex supported by another monoanionic N,N-based
ligand, which is stable in refluxing benzene.10
(11) Jaguar, version 6.5 ; Schrodinger, LLC: New York, 2005.
(12) Use of the sterically bulkier N,N0-2,6-diisopropylphenyl-substi-
tuted β-diketiminate instead of 2 for the reaction in C6D6 gives only a
trace amount of CH3D, indicating that the activation of the benzene
C-D bond is significantly suppressed. This suggests that the displace-
ment of COD by benzene in Scheme 5 is associative; however, in the
absence of any well-characterized products from this reaction, no firm
conclusion can be drawn. The formation of a significant amount of CH4
is also observed, presumably due to the competing C-H bond activation
of COD or β-diketiminate. Further characterization of the palladium
products is needed to identify the source of the CH4.
The mechanism for benzene C-H activation could involve
either σ-bond metathesis or oxidative addition followed by
(9) Kang, M.; Sen, A. Organometallics 2005, 24, 3508–3515.
(10) Iverson, C. N.; Carter, C. A. G.; Baker, R. T.; Scollard, J. D.;
Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2003, 125, 12674–12675.

4404 Organometallics, Vol. 28, No. 15, 2009
Lin et al.
Further calculations indicate that migratory insertion of
cis-2-butene into Pd-CH3 (9 to 11, Scheme 7) is nearly
thermoneutral and has a 26.5 kcal/mol activation free en-
ergy, lower than that of TSσ by 2.3 kcal/mol. This small
difference appears consistent with the experimental observa-
tion of competitive C-H activation vs migratory insertion,
given the imprecision of DFT calculations as well as the
simplified model system.
In summary, we have demonstrated that our monomethyl-
palladium(II) β-diketiminate can undergo both benzene
C-H activation and migratory insertion of olefin at room
temperature. The simultaneous observation of these two
reactions for metal alkyls has previously only been demon-
strated in the catalytic hydroarylation of olefins at elevated
temperatures.3 In addition, unlike other monomethyl-
palladium(II) systems, where olefin oligomerizations and
polymerizations dominate, benzene C-H activation is some-
what favored in our system. DFT calculations on a simplified
model system suggest that the C-H activation occurs
through a σ-bond metathesis mechanism. Further work is
underway to investigate effects of various anionic ligands on
the competition between benzene C-H activation and mi-
gratory insertion at monomethylpalladium(II), with the goal
of gaining insights leading to the design of Pd(II)-based
catalysts for hydroarylation of olefins.
(75 MHz, CH2Cl2): δ 158.30, 155.20, 131.26, 129.67, 127.75,
127.72, 123.23, 109.13, 93.99, 73.65, 34.64, 29.19, 27.78, 22.13,
21.96, 18.66, 18.41. HRMS (FABþ): m/z calcd for C30H40N2Pd
534.2227, found 534.2207.
Preparation of Compound 5. Lithium N,N0-2,6-dimethyl-
phenyl-substituted β-diketiminate 2 (11.8 mg, 0.0374 mmol)
and (COD)Pd(CH3)Cl (10 mg, 0.0377 mmol) were dissolved in
0.5 mL of C6H6 and stirred overnight at room temperature
under nitrogen. The solution turned yellow immediately upon
mixing with slow precipitation of a black solid. After the
removal of the precipitate and the solvent, a yellow solid was
obtained. The 1H NMR spectrum of the solid in CD2Cl2 shows
that compound 5 is the major product (NMR yield: 81%), which
was further characterized by high-resolution mass spectroscopy.
1H NMR (300 MHz, CD2Cl2): δ 7.20-6.81 (m, 11H), 5.42 (t, J=
7.9, 1H), 4.76 (s, 1H), 2.79 (tt, J=2.8, 12, 1H), 2.66 (q, J=8.0,
2H), 2.27 (s, 6H), 2.19 (s, 6H), 1.57 (s, 6H), 1.39-0.98 (m, 8H).
HRMS (FABþ): m/z calcd for C35H42N2Pd 596.2383, found
596.2424; [M - phenylcyclooctenyl]þ calcd for C21H28N2Pd
414.13, found 414.1061.
Preparation of Compound 6. Lithium N,N0-2,6-dimethylphe-
nyl-substituted β-diketiminate 2 (11.8 mg, 0.0374 mmol),
(COD)Pd(CH3)Cl (10 mg, 0.0377 mmol), and ∼2 equiv of
acetonitrile were dissolved in 0.7 mL of C6H6 and stirred over-
night at room temperature under nitrogen. The solution turned
yellow immediately upon mixing with slow precipitation of a
black solid. After 24 h, the solution was filtered to remove the
precipitate and volatiles were removed under vacuum. A yellow
solid was obtained, consisting entirely (1H NMR, CDCl3) of 6.
Crystals suitable for X-ray structural analysis were obtained
by thermal diffusion of petroleum ether into the filtrate in a
refrigerator (isolated yield>95%). 1H NMR (300 MHz, CDCl3):δ
7.04 (d, J = 7.4, 4H), 6.97-6.78 (m, 2H), 4.72 (s, 1H), 2.32
(s, 6H), 2.22 (s, 6H), 1.57 (s, 3H), 1.53 (s, 3H), 1.53 (s, 3H), -0.57
(s, 3H). 13C NMR (75 MHz, CDCl3): δ 160.53, 159.91, 154.56,
147.69, 134.99, 134.77, 131.57, 130.14, 129.90, 126.21, 125.70,
96.08, 27.27, 25.47, 21.45, 21.28, 4.69, 0.63. HRMS (FABþ):
m/z calcd for C24H31N3Pd 467.1553, found 467.1571.
Computational Methodology. All the structures were fully
optimized by using the restricted hybrid density functional
theory B3LYP method as implemented in the Jaguar 6.5 pro-
gram package.11 The Pd was described by LACVP**, a basis set
consisting of the Wadt and Hay15 relativistic effective core
potentials (RECPs) and valence double-ζ contraction functions.
A modified variant of Pople’s16 all-electron basis set, 6-31G**,
where the six d functions have been reduced to five, was used for
all other atoms. It should be noted that the computation method
used in this study is commonly more reliable in studying trends
than providing absolute numbers for the reactions, although
model calculations with similar methods have been demon-
strated to afford remarkably accurate figures in absolute terms
for migratory insertions.17
Experimental Section
General Information. All air- and/or moisture-sensitive
compounds were manipulated by using standard Schlenk tech-
niques or in a glovebox under a nitrogen atmosphere. All
starting materials are commercially available and used as re-
ceived without further purification. Solvents were dried with
appropriate methods before use and stored under nitrogen.
(COD)Pd(CH3)Cl was synthesized according to the literature
procedure.13 A modified literature procedure was used for the
synthesis of THF-freelithiumβ-diketiminates.14 All NMR spectra
were recorded at room temperature using a Varian Mercury 300
spectrometer. NMR spectra were referenced to TMS using the
residual impurities of the given solvent. Chemical shifts are
reported using the standard δ notation in parts per million;
positive chemical shifts are to a higher frequency from TMS, and
coupling constants are reported in Hz. Multiplicities are re-
ported as follows: singlet (s), doublet (d), doublet of doublets
(dd), doublet of triplets (dt), triplet (t), quartet (q), multiplet (m),
broad resonance (br). The Caltech X-ray Crystallography
Laboratory provided the X-ray analysis. High-resolution
mass spectra were obtained at Caltech Mass Spectrometry
Laboratory.
Preparation of Compound 3. Lithium N,N0-2,6-dimethylphe-
nyl-substituted β-diketiminate 2 (11.8 mg, 0.0374 mmol) and
(COD)Pd(CH3)Cl (10 mg, 0.0377 mmol) was mixed with 0.7 mL
of CD2Cl2 at room temperature under nitrogen in a J-Young
tube. The solution turned yellow immediately upon mixing with
slow precipitation of a black solid. The reaction was complete
after 24 h, with 3 as the major product (NMR yield: 90%). After
removal of the precipitate, crystals suitable for X-ray structural
analysis were obtained by thermal diffusion of petroleum ether
into the filtrate in the refrigerator. 1H NMR (300 MHz, CD2Cl2): δ
6.85-6.80 (m, 4H), 6.69-6.63 (m, 2H), 5.03 (t, J=7.9, 1H), 4.53
(s, 1H), 2.40 (q, J=8.0, 2H), 2.02 (s, 6H), 1.95 (s, 6H), 1.34
(s, 6H), 0.97-0.49 (m, 9H), 0.43 (d, J = 6.6, 3H). 13C NMR
All species were treated as singlets. Harmonic vibrational
frequencies were calculated with the B3LYP method for the
optimized geometries. Zero-point vibrational energy correc-
tions and thermodynamic corrections were obtained by using
unscaled frequencies. Free energies were calculated for each
species at 298 K and 1 atm in the gas phase. All transition
structures possess one and only one imaginary frequency and
(15) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299–310.
(16) (a) Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 16,
217–219. (b) Franel, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon,
M. S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654–3665.
(17) (a) Musaev, D. G.; Svensson, M.; Morokuma, S.; Stromberg, S.;
Zetterberg, K.; Siegbahn, P. E. M. Organometallics 1997, 16, 1933–1945.
(b) Deng, L.; Woo, T. K.; Cavallo, L.; Margl, P. M.; Ziegler, T. J. Am. Chem.
€
(13) Salo, E. V.; Guan, Z. B. Organometallics 2003, 22, 5033–5046.
(14) The preparation of the THF-free lithium salt is performed with
1.6 M n-BuLi/hexane in hexanes: Budzelaar, P. H. M.; de Gelder, R.;
Gal, A. W. Organometallics 1998, 17, 4121–4123.
Soc. 1997, 119, 6177–6186. (c) Albert, K.; Gisdakis, P.; Rosch, N. Organo-
metallics 1998, 17, 1608–1616. (d) Michalak, A.; Ziegler, T. J. Am. Chem.
Soc. 2001, 123, 12266–12278. (e) Lin, B.-L.; Liu, L.; Fu, Y.; Luo, S.-W.;
Chen, Q.; Guo, Q.-X. Organometallics 2004, 23, 2114–2123.

Article
Organometallics, Vol. 28, No. 15, 2009 4405
were further confirmed by following the corresponding normal
mode toward each product and reactant.
suggestions in preparing the manuscript. This work has been
generously supported by BP through the MC2 program.
Acknowledgment. We thank Dr. Michael W. Day and
Lawrence M. Henling for performing the X-ray crystal-
lographic analysis. B.-L.L. also thanks George S. Chen for
Supporting Information Available: Detailed experimental and
computational data. This material is available free of charge via
the Internet at http://pubs.acs.org.