Fine-tuning the energy barrier for metal-mediated dinitrogen N≡N bond cleavage

Article abstract of DOI:10.1021/ja505309j

Experimental data support a mechanism for N≡N bond cleavage within a series of group 5 bimetallic dinitrogen complexes of general formula, {Cp*M[N(ⁱPr)C(R)N(ⁱPr)]}₂(μ-N ₂) (Cp* = η⁵-C₅Me₅) (M = Nb, Ta), that proceeds in solution through an intramolecular end-on-bridged (μ- η¹:η¹-N ₂) to side-on-bridged (μ- η²: η²-N₂) isomerization process to quantitatively provide the corresponding bimetallic bis(μ-nitrido) complexes, {Cp*M[N( ⁱPr)C(R)N(ⁱPr)](μ-N)}₂. It is further demonstrated that subtle changes in the steric and electronic features of the distal R-substituent, where R = Me, Ph and NMe₂, can serve to modulate the magnitude of the free energy barrier height for N≡N bond cleavage as assessed by kinetic studies and experimentally derived activation parameters. The origin of the contrasting kinetic stability of the first-row congener, {Cp*V[N(ⁱPr)C(Me)N(ⁱPr)]} ₂(μ- η¹:η¹-N₂) toward N≡N bond cleavage is rationalized in terms of a ground-state electronic structure that favors a significantly less-reduced μ-N₂ fragment.

Full text of DOI:10.1021/ja505309j

Communication
pubs.acs.org/JACS
Fine-Tuning the Energy Barrier for Metal-Mediated Dinitrogen NN
Bond Cleavage
Andrew J. Keane, Brendan L. Yonke, Masakazu Hirotsu, Peter Y. Zavalij, and Lawrence R. Sita*
Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742, United States
S
* Supporting Information
Scheme 1
ABSTRACT: Experimental data support a mechanism for
NN bond cleavage within a series of group 5 bimetallic
dinitrogen complexes of general formula, {Cp*M[N(iPr)-
C(R)N(iPr)]}2(μ-N2) (Cp* = η5-C5Me5) (M = Nb, Ta),
that proceeds in solution through an intramolecular “end-
on-bridged” (μ-η1:η1-N2) to “side-on-bridged” (μ-η2:η2-
N2) isomerization process to quantitatively provide the
corresponding bimetallic bis(μ-nitrido) complexes,
{Cp*M[N(iPr)C(R)N(iPr)](μ-N)}2. It is further demon-
strated that subtle changes in the steric and electronic
features of the distal R-substituent, where R = Me, Ph and
NMe2, can serve to modulate the magnitude of the free
energy barrier height for NN bond cleavage as assessed
by kinetic studies and experimentally derived activation
parameters. The origin of the contrasting kinetic stability
of the first-row congener, {Cp*V[N(iPr)C(Me)N-
(iPr)]}2(μ-η1:η1-N2) toward NN bond cleavage is
ization process as the most likely pathway involved, rather than
rationalized in terms of a ground-state electronic structure
the alternative of initial homolytic NN bond fragmentation to
that favors a significantly less-reduced μ-N2 fragment.
a transient terminal nitride species followed by bimolecular
recombination (cf. upper and lower paths of Scheme 1,
respectively). Finally, we establish that the sharply contrasting
kinetic stability of the first-row congener, {Cp*V[N(iPr)C(Me)-
N(iPr)]}2(μ-N2) (5), is most likely the consequence of a change
in the ground-state electronic structure of this bimetallic
dinitrogen complex that favors more reduced metal centers
and a less-reduced central μ-N2 fragment.4
central requirement of nitrogen fixation, which encom-
Apasses all chemical and biochemical processes that convert
molecular dinitrogen (N2) into more reduced nitrogen-
containing products, and most essentially, ammonia (NH3), is
a reaction pathway that can ultimately surmount the high free
energy barrier associated with breaking the exceptionally strong
NN triple bond.1−4 Many seminal advances have been made
over the past 100 years with delineating the mechanisms of
different metal-mediated nitrogen fixation schemes. However, to
date, no experimental system has yet been discovered or
devisedwhether of chemical or biochemical inspirationthat
permits one to systematically probe the impact of changes in
molecular and electronic structure on the free energy barrier
height of a well-established NN bond cleaving process through
the programmed synthesis of a family of substitutionally
differentiated derivatives. Herein, we now present such a system
that is based on the thermal conversion of group 5 bimetallic
dinitrogen complexes of general formula, {Cp*M[N(iPr)C(R)-
N(iPr)]}2(μ-N2) (Cp* = η5-C5Me5), where M = Ta (1) and Nb
(2), to the corresponding bimetallic bis(μ-nitrido) complexes,
{Cp*M[N(iPr)C(R)N(iPr)](μ-N)}2, where M = Ta (3) and Nb
(4), respectively, according to Scheme 1.5 Importantly, through
synthetic manipulation of the steric and electronic nature of the
distal R-substituent of 1 and 2, and specifically, where R = Me
(a), Ph (b) and NMe2 (c), we provide experimental data in
support of an intramolecular μ-η1:η1-N2 → μ-η2:η2-N2 isomer-
In 2007, we reported that the ditantalum “end-on-bridged”
dinitrogen complex 1a could be isolated as a paramagnetic
crystalline material through chemical reduction of the tantalum
trichloride, Cp*M[N(iPr)C(R)N(iPr)]Cl3, where M = Ta and R
= Me (6a), in tetrahydrofuran (THF) solution at low
temperature using 4 equiv of potassium graphite (KC8) as the
reductant according to Scheme 2.5−7 In hydrocarbon (e.g.,
Scheme 2
Received: May 27, 2014
Published: June 24, 2014
© 2014 American Chemical Society
9906
dx.doi.org/10.1021/ja505309j | J. Am. Chem. Soc. 2014, 136, 9906−9909

Journal of the American Chemical Society
Communication
Figure 1. Molecular structures (30% thermal ellipsoids) of (a) 1b, (b) 2b and (c) 5. Hydrogen atoms have been removed for the sake of clarity.8
toluene or pentane) solution, pure samples of 1a were observed
to quantitatively convert to the ditantalum bis(μ-nitrido)
complex 3a at temperatures above 0 °C according to Scheme 1.
In the present work, similar KC8 reductions of the trichlorides
6b and 6c, where R = Ph and NMe2, respectively, provided the
corresponding complexes 1b and 1c in good yields.8
Interestingly, both of these derivatives proved to be more
kinetically stable in solution at room temperature than 1a (vide
infra). Upon moving to the second row, chemical reduction of
the niobium trichloride 7a (see Scheme 2) in standard fashion
now provided a green-colored solution that only persisted at
temperatures below −30 °C. Upon warming to room temper-
ature, this same solution rapidly turned pale-orange in color, and
Figure 2. (a) Scanning UV−vis spectroscopy for (15N2, 98%)-1a →
after the typical workup, no trace of the desired dinitrogen
(15N2, 98%)-3a at T = 318.15 K.8 (b) Kinetic analysis of [1a]t vs time as
complex 2a could be obtained, but rather, only the bis(μ-nitrido)
monitored at λ = 625 nm.
product 4a was isolated (see Scheme 1).8 However, similar to the
̀
increased stability of 1b vis-a-vis 1a, chemical reduction of the
converting to a cis, trans isomeric mixture of 3a, bulk thermolysis
in hydrocarbon solutions of 1b, 1c, and 2b provided excellent
isolated yields of the corresponding diamagnetic bimetallic bis(μ-
nitrido) complexes 3b, 3c, and 4b, respectively, as crystalline
mixtures. Fractional crystallization and single-crystal X-ray
analysis then confirmed the structural identity of either a trans
or cis isomer in each case, along with solid-state geometric
parameters that confirmed the absence of N−N bonding [cf.
calculated d(NN) values >2.55 Å].8
niobium trichloride 7b (R = Ph) now provided the paramagnetic
diniobium dinitrogen complex 4b (58% yield) that likewise
displayed good kinetic stability in solution at 0 °C. Finally, to
complete an entire group 5 isostructural series, chemical
reduction of Cp*V[N(iPr)C(Me)N(iPr)]Cl (8), but now
using 0.5% sodium amalgam (NaHg) as the reductant in THF
at −30 °C, produced an excellent 91% yield of the paramagnetic
divanadium dinitrogen complex 5 that surprisingly proved to be
thermally robust in solution up to temperatures of 90 °C.8 It can
be further noted that, except for M = Cr, synthetic derivatives of
{Cp*M[N(iPr)C(R)N(iPr)]}2(μ-N2) have now been prepared
for the entire set of groups 4−6 metals (M = Ti,7 Zr,6 Hf,6 V, Nb,
Ta,5 Mo,7 and W7).
Detailed kinetic investigations of the thermal conversions
presented in Scheme 1 were conducted for methylcyclohexane
solutions using variable-temperature scanning UV−vis spectros-
copy.8 Figure 2a presents typical data obtained at a given
temperature by this method that further confirmed the
quantitative nature of each thermal reaction through the
appearance of sharply focused isosbestic points in an overlay
plot of spectra collected at fixed time intervals. By monitoring the
rate of disappearance of the bimetallic dinitrogen complex as a
function of time at a single fixed wavelength (see Figure 2b),
kinetic rate data at five different temperatures (in triplicate) over
a temperature range (cf. 313−333 K for 1a and 2b, 328−353 K
for 1c, and 343−368 K for 1b) were collected and fit to the
Eyring equation as shown in Figure 3, and from which, the set of
activation parameters provided in Table 1 were derived.8 Further
studies revealed that kinetic parameters remained invariant over a
range of concentrations and that initial rates were consistent with
being first-order in dinitrogen complex. Finally, kinetic analysis
of the thermal conversion of isotopically labeled (15N2, 98%)-1a
to (15N2, 98%)-3a was performed in similar fashion (see Table 1).
Based on the experimentally derived enthalpic and entropic
activation parameters, ΔH⧧ and ΔS⧧, respectively, of Table 1,
ΔG⧧ values were computed for the fixed temperature of 338.15 K
and these were found to be consistent with the observed trend in
thermal stability of 1b (R = Ph) > 1c (NMe2) > 1a (Me). A
comparison of the kinetic data and activation parameters for the
Analytical characterization data for the new group 5 bimetallic
dinitrogen complexes 1b, 1c, 2b, and 5 were consistent with the
expected empirical formulas, and single crystal X-ray analyses
served to establish the μ-η1:η1-N2 coordination motif in the solid
state according to the structural data presented in Figure 1. Of
particular interest are the relative magnitudes of the nitrogen−
nitrogen bond distances d(NN) as a function of the metal row
number. In this respect, the values displayed by the second- and
third-row derivatives of 1 and 2 are consistent with a strong
degree of NN bond activation and a formal [M(IV), d1,
M(IV), d1] (M = Nb, Ta) open-shell electronic ground state,
whereas those for the first row analog 5 are more in keeping with
a much smaller degree of N2 activation and more reduced metal
centers, such as a formal [V(II), d3, V(II), d3] electronic
structure, [cf. 1a: d(NN) = 1.313(4) Å; 1b: 1.308(5) Å; 1c:
1.306(5) Å; 2b: 1.300(3) Å; 5: 1.225(2) Å].4,5,7,9
As already previously noted, derivatives of 1 and 2 exhibited
profound differences in kinetic stability in solution depending
upon the nature of the distal R-substituent, and qualitatively, in
order of decreasing stability for R = Ph > NMe2 > Me. Further, in
keeping with the original observation of 1a quantitatively
9907
dx.doi.org/10.1021/ja505309j | J. Am. Chem. Soc. 2014, 136, 9906−9909

Journal of the American Chemical Society
Communication
Chart 1
within a structurally characterized reduced product has been
reported by Floriani et al.10 Fryzuk et al.11 have also previously
proposed that thermal conversion of a diniobium μ-η1:η1-N2
complex to a product arising from NN bond cleavage and
formal intramolecular insertion of a metal nitrido group into a
niobium−phosphorus bond of the supporting ligand framework
proceeds via a diniobium side-on-bridged μ-η2:η2-N2 intermedi-
ate. Although additional experimental support for such an
intermediate or the proposed bimetallic μ-η1:η1-N2 → μ-η2:η2-N2
thermal rearrangement step was not provided, a computational
investigation of the relative thermodynamic stabilities of
simplified structural models for μ-η1:η1-N2 vs μ-η2:η2-N2
coordination appeared to put this hypothesis on firmer
theoretical ground.12 In our own studies, we have also previously
stated that the most likely mechanism for thermal conversion of
1a to 3a involves an intramolecular μ-η1:η1-N2 to μ-η2:η2-N2
structural isomerization that occurs prior to NN bond
cleavage according to the upper pathway of Scheme 1.5,7 The
basis for this conjecture rested on structural data obtained for the
set of related group 4 bimetallic dinitrogen derivatives, {(η5-
C5Me4R′)M[N(R′′)C(R)N(R′′)]}2(μ-η2:η2-N2) (M = Zr and
Hf) (I), which display exceedingly large d(NN) values that
increase as the magnitude of nonbonded steric interactions
within the supporting ligand environment decrease as shown in
Chart 1.6 Due to the inability of the formal group 4 M(IV, d0)
metal centers in these complexes to deliver any additional
reducing electron equivalents to the central NN bond, these
structures can be viewed as representing “arrested” transition
states for the proposed isomerization mechanism of Scheme 1.10
More recently, Musaev et al.13 have published the results of a
detailed computational study of a simplified structural model for
1a → 3a which provides additional theoretical support for our
hypothesis.
As presented in the lower half of Scheme 1, a possible
alternative mechanism for NN bond cleavage within bimetallic
end-on-bridged dinitrogen complexes is via homolytic NN
bond fragmentation to produce two equivalents of a metal
terminal nitride product.4 Experimental precedent for this
pathway rests almost entirely with the seminal structural and
kinetic studies performed by Cummins et al.14 for the
dimolybdenum dinitrogen complex, {(Ar[tBu]N)3Mo}2(μ-
η1:η1-N2) (Ar = 3,5-C6H3Me2) (II) that undergoes thermally
induced NN bond fragmentation to produce 2 equiv of the
corresponding terminal metal nitride, (Ar[tBu]N)3Mo(N) (III).
Detailed kinetic and computational studies conducted by this
group and others support a reaction profile that proceeds
through a ‘trans-zigzag’ transition state leading to homolytic N
N bond cleavage and terminal metal nitride products.9,15
Since the magnitudes (i.e., negative ΔS⧧) and trends in
activation parameters presented in Table 1 do not unequivocally
establish that an intramolecular isomerization mechanism is
favored over homolytic NN bond fragmentation for
thermolysis of the present set of group 5 bimetallic dinitrogen
complexes, several additional experimental results and lines of
inquiry were pursued. Thus, to begin, comparing the
Figure 3. Temperature-dependent first-order rate constants with least-
squares fit to the Eyring equation, k = (kBT/h) exp(ΔS⧧/R) exp(−ΔH⧧/
RT), for the conversion of 1a → 3a (circle), 1c → 3c (square) and 1b →
3b (triangle). For a similar analysis of data for the conversion of 2b →
4b, see Supporting Information.8
Table 1. Experimentally-Derived Activation Parameters for
Thermal Conversion of 1 → 3 and 2b → 4b
ab
,
b
b
ΔG⧧ (kcal/mol)
ΔH⧧ (kcal/mol)
ΔS⧧ (cal/(mol K))
1a
24.3(3)
24.4(2)
25.5(3)
26.6(3)
24.4(1)
21.5(3)
21.8(2)
20.6(3)
21.7(3)
20.2(1)
−8.4(9)
−7.4(6)
−14.6(8)
−14.4(8)
−12.4(5)
1a (15N2)
1c
1b
2b
a
b
Calculated at 338.15 K. Error reported at the 95% confidence
interval.
second-row derivative 2b (R = Ph) reveals that it is similar in
stability to 1a [cf. ΔG⧧ (338.15 K) = 24.3(3) kcal mol−1 for 1a vs
24.4(1) kcal mol−1 for 2b] (see Table 1).8 Intriguingly, all ΔS⧧
values are negative in magnitude, which implies a more highly
ordered transition state that must be reached for the rate-
determining step along the reaction pathway. Indeed, as Table 1
reveals, the greatest contributing factor for the increase in ΔG⧧
values in the order of 1a < 1c < 1b is the large 6 cal mol−1 K−1
negative change in ΔS⧧ associated with the more sterically
encumbered R = Ph and NMe2 substituents of 1b and 1c,
respectively, relative to R = Me for 1a. In analyzing data for the
former pair more closely, the notable difference in the enthalpic
term [cf. ΔH⧧ = 20.6(3) kcal mol−1 for 1c vs 21.7(3) kcal mol−1
for 1b] might suggest more stabilization of the formally oxidized
metal centers of the transition state by the electron-rich NMe2
substituent relative to the Ph group. Finally, in comparing the
second- and third-row derivatives bearing the same R = Ph
substituent, it is notable from the data of Table 1 that the much
smaller calculated ΔG⧧ (338.15 K) value of 24.4(1) kcal mol−1
obtained for 2b relative to that of 1b at 26.6(3) kcal mol−1 is
largely due to a significant reduction in the change in enthalpy of
activation term, ΔH⧧, which is 20.2(1) kcal mol−1 for 2b vs
21.7(3) kcal mol−1 for 1b. In summary, these kinetic data
strongly suggest that, at least for the present supporting ligand
environment, diniobium μ-η1:η1-N2 complexes will possess an
intrinsically lower potential energy barrier for NN bond
cleavage relative to a ditantalum μ-η1:η1-N2 congener and that a
sterically demanding distal R-substituent is critical for kinetically
stabilizing an isolable diniobium dinitrogen derivative.
Upon sodium metal reduction in dimethoxyethane (DME)
solution, rearrangement of a diniobium end-on-bridged μ-η1:η1-
N2 moiety to a diniobium side-on-bridged μ-η2:η2-N2 group that
is stabilized by secondary sodium−nitrogen bonding interactions
9908
dx.doi.org/10.1021/ja505309j | J. Am. Chem. Soc. 2014, 136, 9906−9909

Journal of the American Chemical Society
Communication
experimentally derived activation parameters obtained for 15N2
(98%)-labeled 1a with those for unlabeled 1a provided a very low
to virtually nonexistent 14N2/15N2 kinetic isotope effect of
1.03(3) at 318.15 K, which would be consistent with an
isomerization mechanism in which no significant NN bond
cleavage had yet occurred in the transition state.14 Second, a
crossover experiment was conducted in which a concentrated
solution consisting of a 1:1 mixture of 1a and 1c in THF, which
was chosen as the solvent to maintain a homogeneous solution of
starting materials and products, was heated to 70 °C for 3.5 h
within a sealed tube. Analysis of the crude product mixture by
electrospray ionization mass spectrometry subsequently revealed
only the presence of parent molecular ions [M + H]+ at 943.21
and 1001.26 m/z that correspond to 3a and 3c, respectively, and
with no evidence being obtained for the theoretical crossover
product, [{Cp*Ta[N(iPr)C(Me)N(iPr)]}{Cp*Ta[N(iPr)C-
(NMe2)N(iPr)]}(μ-N)2], which would be expected to exhibit a
molecular ion [M + H]+ at 972.45 m/z. Finally, we undertook the
synthesis of a comparable set of isolable ditantalum “side-on”
dinitrogen complexes in which the magnitude of steric
interactions within the supporting ligand environment remains
approximately the same as 1. Gratifyingly, introduction of nitrous
oxide (N2O) into toluene solutions of 1a−c at −30 °C readily
generated a new product in each case and in high yield that could
be isolated as a diamagnetic, purple crystalline material.8 As
depicted in Scheme 3, analytic, spectroscopic, and single-crystal
X-ray data served to establish the identity of these products as
being, {Cp*Ta[N(iPr)C(R)N(iPr)]}2(μ-η2:η2-N2)(μ-O) where
R = Me (9a), Ph (9b), and NMe2 (9c).8 Importantly, the latter
structural data unequivocally confirm the ability of a [Ta(V),
Ta(V)] dinitrogen complex to adopt side-on μ-N2 coordina-
tion.16 For all three structures, the unique Ta2N2 fragment is
characterized as being highly asymmetric with respect to d(MN)
values, and the d(NN) parameter is large thereby suggesting a
high degree of formal NN bond activation [cf. 1.493(12),
1.499(4), and 1.504(4) Å for 9a−c, respectively].8
indeed, addition of 4 equiv of 2,6-dimethylphenylisocyanide to a
pentane solution of 5 provided a 89% yield of the structurally
characterized V(II) bis(isocyanide) complex, Cp*V[N(iPr)C-
(Me)N(iPr)]}[CN(2,6-Me2C6H3)]2 (10), according to Scheme
4.7,8
In summary, the present work serves to experimentally
establish a thermal intramolecular pathway by which the
exceptionally strong bond of N2 can be cleaved via bimetallic
coordination and with external control over the free energy
barrier.
ASSOCIATED CONTENT
* Supporting Information
Experimental details and crystallographic analyses. This material
is available free of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
lsita@umd.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Funding for this work was provided by the Department of
Energy, Basic Energy Sciences (Grant DE-SC0002217) for
which we are grateful.
REFERENCES
■
(1) Smil, V. Enriching the Earth: Fritz Haber, Carl Bosch, and the
Transformation of World Food Production; MIT Press: Cambridge, MA,
2001.
(2) (a) Howard, J. B.; Rees, D. C. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
17088. (b) Hu, Y.; Ribbe, M. W. J. Biol. Inorg. Chem. DOI: 10.1007/
s00775-014-1137-2.
(3) Ertl, G. Catal. Rev. Sci. Eng. 1980, 21, 201.
(4) (a) Chirik, P. J. Dalton Trans. 2007, 16. (b) Fryzuk, M. D. Acc.
Chem. Res. 2009, 42, 127. (c) Tanabe, Y.; Nishibayashi, Y. Coord. Chem.
Rev. 2013, 257, 2551.
(5) Hirotsu, M.; Fontaine, P. P.; Epshteyn, A.; Zavalij, P. Y.; Sita, L. R. J.
Am. Chem. Soc. 2007, 129, 9284.
(6) Hirotsu, M.; Fontaine, P. P.; Zavalij, P. Y.; Sita, L. R. J. Am. Chem.
Soc. 2007, 129, 12690.
(7) Fontaine, P. P.; Yonke, B. L.; Zavalij, P. Y.; Sita, L. R. J. Am. Chem.
Soc. 2010, 132, 12273.
(8) Details are provided in the SI.
As a final consideration, we have begun to probe the possible
origin(s) of the unique thermal solution stability displayed by the
̀
first-row divanadium dinitrogen complex 5 vis-a-vis that of 1 and
2. Mindiola et al.9 have previously explored the prohibitive
kinetic barrier that exists for achieving NN bond cleavage
within a [V(II), V(II)] (μ-η1:η1-N2) complex. In the present
study, the structural parameters obtained for 5 are highly
suggestive of a similar ground-state electronic configuration, and
(9) Tran, B. L.; Pinter, B.; Nichols, A. J.; Konopka, F. T.; Thompson,
R.; Chen, C.-H.; Krzystek, J.; Ozarowski, A.; Telser, J.; Baik, M.-H.;
Meyer, K.; Mindiola, D. J. J. Am. Chem. Soc. 2012, 134, 13035.
(10) Caselli, A.; Solari, E.; Scopelliti, R.; Floriani, C.; Re, N.; Rizzoli, C.;
Chiesi-Villa, A. J. Am. Chem. Soc. 2000, 122, 3652.
Scheme 3
(11) Fryzuk, M. D.; Kozak, C. M.; Bowdridge, M. R.; Patrick, B. O.;
Rettig, S. J. J. Am. Chem. Soc. 2002, 124, 8389.
(12) Christian, G. J.; Terrett, R. N. L.; Stranger, R.; Cavigliasso, G.;
Yates, B. F. Chem.Eur. J. 2009, 15, 11373.
(13) Zhang, W.; Tang, Y.; Lei, M.; Morokuma, K.; Musaev, D. G. Inorg.
Chem. 2011, 50, 9481.
(14) (a) Laplaza, C. E.; Cummins, C. C. Science 1995, 268, 861.
(b) Laplaza, C. E.; Johnson, M. J. A.; Peters, J.; Odom, A. L.; Kim, E.;
Cummins, C. C.; George, G. N.; Pickering, I. J. J. Am. Chem. Soc. 1996,
118, 8623. (c) Curley, J. J.; Cook, T. R.; Reece, S. Y.; Muller, P.;
Cummins, C. C. J. Am. Chem. Soc. 2008, 130, 9394.
(15) (a) Cui, Q.; Musaev, D. G.; Svensson, M.; Sieber, S.; Morokuma,
K. J. Am. Chem. Soc. 1995, 117, 12366. (b) Brookes, N. J.; Graham, D. C.;
Christian, G.; Stranger, R.; Yates, B. F. J. Comput. Chem. 2009, 30, 2146.
(16) Fryzuk, M. D.; Johnson, S. A.; Rettig, S. J. J. Am. Chem. Soc. 1998,
120, 11024.
Scheme 4
9909
dx.doi.org/10.1021/ja505309j | J. Am. Chem. Soc. 2014, 136, 9906−9909