Thermodynamic, kinetic, and mechanistic study of oxygen atom transfer from mesityl nitrile oxide to phosphines and to a terminal metal phosphido complex

Article abstract of DOI:10.1021/ic2013599

The enthalpies of oxygen atom transfer (OAT) from mesityl nitrile oxide (MesCNO) to Me₃P, Cy₃P, Ph₃P, and the complex (Ar[^tBu]N)₃MoP (Ar = 3,5-C₆H₃Me ₂) have been measured by solution calorimetry yielding the following P-O bond dissociation enthalpy estimates in toluene solution (±3 kcal mol^-1): Me₃PO [138.5], Cy₃PO [137.6], Ph ₃PO [132.2], (Ar[^tBu]N)₃MoPO [108.9]. The data for (Ar[^tBu]N)₃MoPO yield an estimate of 60.2 kcal mol^-1 for dissociation of PO from (Ar[^tBu]N) ₃MoPO. The mechanism of OAT from MesCNO to R₃P and (Ar[^tBu]N)₃MoP has been investigated by UV-vis and FTIR kinetic studies as well as computationally. Reactivity of R₃P and (Ar[^tBu]N)₃MoP with MesCNO is proposed to occur by nucleophilic attack by the lone pair of electrons on the phosphine or phosphide to the electrophilic C atom of MesCNO forming an adduct rather than direct attack at the terminal O. This mechanism is supported by computational studies. In addition, reaction of the N-heterocyclic carbene SIPr (SIPr = 1,3-bis(diisopropyl)phenylimidazolin-2-ylidene) with MesCNO results in formation of a stable adduct in which the lone pair of the carbene attacks the C atom of MesCNO. The crystal structure of the blue SIPr·MesCNO adduct is reported, and resembles one of the computed structures for attack of the lone pair of electrons of Me₃P on the C atom of MesCNO. Furthermore, this adduct in which the electrophilic C atom of MesCNO is blocked by coordination to the NHC does not undergo OAT with R₃P. However, it does undergo rapid OAT with coordinatively unsaturated metal complexes such as (Ar[ ^tBu]N)₃V since these proceed by attack of the unblocked terminal O site of the SIPr·MesCNO adduct rather than at the blocked C site. OAT from MesCNO to pyridine, tetrahydrothiophene, and (Ar[ ^tBu]N)₃MoN was found not to proceed in spite of thermochemical favorability.

Full text of DOI:10.1021/ic2013599

ARTICLE
pubs.acs.org/IC
Thermodynamic, Kinetic, and Mechanistic Study of Oxygen Atom
Transfer from Mesityl Nitrile Oxide to Phosphines and to a Terminal
Metal Phosphido Complex
Xiaochen Cai,† Subhojit Majumdar,† George C. Fortman,† Luis Manuel Frutos,‡ Manuel Temprado,*,‡
Christopher R. Clough,§ Christopher C. Cummins,*,§ Meaghan E. Germain,|| Taryn Palluccio,||
Elena V. Rybak-Akimova,*,|| Burjor Captain,*,† and Carl D. Hoff*,†
†Department of Chemistry, University of Miami, 1301 Memorial Drive, Coral Gables Florida 33021, United States
‡Department of Physical Chemistry, Universidad de Alcalꢀa, Ctra. Madrid-Barcelona Km. 33,600, Madrid, 28871, Spain
§Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139,
United States
Department of Chemistry, Tufts University, 62 Talbot Avenue, Medford, Massachusetts 02155, United States
S Supporting Information
b
ABSTRACT: The enthalpies of oxygen atom transfer (OAT) from mesityl nitrile
oxide (MesCNO) to Me3P, Cy3P, Ph3P, and the complex (Ar[tBu]N)3MoP (Ar =
3,5-C6H3Me2) have been measured by solution calorimetry yielding the following
PꢀO bond dissociation enthalpy estimates in toluene solution ((3 kcal molꢀ1):
Me3PO [138.5], Cy3PO [137.6], Ph3PO [132.2], (Ar[tBu]N)3MoPO [108.9]. The
data for (Ar[tBu]N)3MoPO yield an estimate of 60.2 kcal molꢀ1 for dissociation of
PO from (Ar[tBu]N)3MoPO. The mechanism of OAT from MesCNO to R3P and
(Ar[tBu]N)3MoP has been investigated by UVꢀvis and FTIR kinetic studies as well
as computationally. Reactivity of R3P and (Ar[tBu]N)3MoP with MesCNO is
proposed to occur by nucleophilic attack by the lone pair of electrons on the
phosphine or phosphide to the electrophilic C atom of MesCNO forming an adduct rather than direct attack at the terminal O. This
mechanism is supported by computational studies. In addition, reaction of the N-heterocyclic carbene SIPr (SIPr = 1,3-
bis(diisopropyl)phenylimidazolin-2-ylidene) with MesCNO results in formation of a stable adduct in which the lone pair of the
carbene attacks the C atom of MesCNO. The crystal structure of the blue SIPr MesCNO adduct is reported, and resembles one of
3
the computed structures for attack of the lone pair of electrons of Me3P on the C atom of MesCNO. Furthermore, this adduct in
which the electrophilic C atom of MesCNO is blocked by coordination to the NHC does not undergo OAT with R3P. However, it
does undergo rapid OAT with coordinatively unsaturated metal complexes such as (Ar[tBu]N)3V since these proceed by attack of
the unblocked terminal O site of the SIPr MesCNO adduct rather than at the blocked C site. OAT from MesCNO to pyridine,
3
tetrahydrothiophene, and (Ar[tBu]N)3MoN was found not to proceed in spite of thermochemical favorability.
’ INTRODUCTION
A new approach to N2O chemistry has also emerged in the form
of its capture by frustrated Lewis pair (FLP) systems.7
Oxygen atom transfer (OAT) is a fundamental step in bio-
chemical and industrial oxidative reactions.1 Selective catalytic
oxidation of substrates is at the heart of development of many
environmentally benign processes,2 and understanding the detailed
mechanisms of OAT reactions will help in design of such catalysts.
Nitrous oxide is of interest as a green oxidant both because the
byproduct of its oxidations is the innocuous N2 molecule, and
because it is a thermodynamically powerful oxidant (De(NꢀO)
ca. 40 kcal molꢀ1).3 A reality to be dealt with in harnessing N2O
for oxidation chemistry is that it exhibits low reactivity and often
can be regarded as essentially an inert gas.4 Nitrous oxide
oxidations may in principle be catalyzed, in furtherance of which
coordination complexes of this small molecule have been
studied.5 Recently, an example of such a complex has been
characterized in the solid state by X-ray crystallography.6
Another common class of compounds frequently used in OAT
processes is the amine-N-oxides. Here as well, transition metal
based catalysis of nonmetal oxidations is often required since a
number of thermodynamically favorable oxygen atom transfer
reactions occur at prohibitively slow rates at room temperature.
An example is transfer of an O atom from pyridine-N-oxide
(PyO) to a phosphine as shown in reaction 1.
Received: June 24, 2011
Published: August 29, 2011
r
2011 American Chemical Society
9620
dx.doi.org/10.1021/ic2013599 Inorg. Chem. 2011, 50, 9620–9630
|

Inorganic Chemistry
ARTICLE
In spite of being exothermic by over 70 kcal molꢀ1 because
of the strong PdO bond formed, reaction 1 does not occur
readily since there is no low energy path available. A simple
explanation is that lone pair/lone pair repulsion yields a high
energy transition state for this very favorable reaction. This
highlights the role kinetic barriers play in oxidative chemistry.8
Transition metal based catalytic systems have been
developed9 to lower this barrier by the two general steps
shown in reactions 2 and 3 which may be representative of a
wide range of enzymatic biochemical and catalytic industrial
oxidations.
does not have not been delineated.
Furthermore, for conversion of the terminal phosphide com-
plex (Ar[tBu]N)3MoP to its oxide, (Ar[tBu]N)3MoPO, the
harsh oxidant dimethyl dioxirane (DMDO) was required, as
both N2O and amine-N-oxides failed to react.21 As described in
this report, MesCNO smoothly effects OAT to (Ar[tBu]N)3MoP,
suggesting that this reagent class ought to be considered more
generally for OAT processes. In this regard, Beck and co-
workers have reported the oxidation of coordinated phosphines
by stable nitrile oxides as shown in reaction 6.22 Fe(CO)5 is
capable of deoxygenating nitrile oxides to nitriles in moderate
yields.23
The step in reaction 2 involves OAT from the oxidant to the
metal complex. Espenson and co-workers have recently shown
that this requires a vacant coordination site at the metal center
to be oxidized.10 The step in reaction 3 involves OAT from the
metal oxide to the substrate. Hall has postulated that this
involves attack of the lone pair of the phosphine on an M=O
π* orbital.11 An alternative mechanism involving coupled
electron transfer/atom transfer has also been proposed for
some OAT steps.12 The factors controlling the kinetics of
reactions 2 and 3 remain an area of great activity. These two
steps serve to bracket the MꢀO bond strength available for a
potential catalyst. Ideally this should be stronger than NꢀO in
PyO but weaker than the PdO bond in R3PO. Holm and co-
workers have highlighted the importance of a thermodynamic
scale for OAT.13
Less well appreciated than N2O and amine-N-oxides for
OAT reactivity is the class of molecules known as nitrile
oxides, RCNO.14 Like nitrous oxide, the 1,3-dipolar nitrile
oxide functional group contains a linear triatomic sequence
with a terminal oxygen atom bonded to an sp hybridized
nitrogen atom. The dative15 NfO bond distances in sub-
stituted benzonitrile oxides are, for example, 1.249(7) and
1.237(10) Å, these being shorter than corresponding distances
for trimethylamine-N-oxide 1.388 Å or pyridine-N-oxide deri-
vatives (1.28ꢀ1.30 Å),16 while longer than that (1.184 Å)
in nitrous oxide.17 The NꢀO bond dissociation enthalpy
(BDE) in MesCNO is only 52.3 kcal molꢀ1, compared to a
value of 63.3 kcal molꢀ1 for the NꢀO BDE in PyO.18 This
makes it a potent OAT reagent thermodynamically, even more
so than molecular oxygen itself as implied by the energetics of
reaction 4.19
In spite of these attractive results we have found only limited
use of MesCNO in literature reports of OAT reactions. A char-
acteristic organic reaction of nitrile oxides is 1,3-cycloaddition.24
Likewise, the use of nitrile oxides in cycloaddition reactions of
organometallic complexes is well documented.25
The work reported here began with an interest in simply
utilizing MesCNO as an oxidant in calorimetric studies, first for
R3P to calibrate it for work in solution calorimetry, and then for
(Ar[tBu]N)3MoP (Ar = 3,5-C6H3Me2) to generate data on the
MoꢀP and PꢀO bond strengths in the known unique terminal
phosphorus monoxide complex (Ar[tBu]N)3MoPO.21 This
work extends our earlier thermodynamic and kinetic studies of
chalcogen atom transfer for E = S, Se, and Te (but not O) for
R3P, NHC (NHC = N-heterocyclic carbene), (Ar[tBu]N)3MoP
and (Ar[tBu]N)3Mo.26 In our studies of sulfur atom transfer
(SAT) to NHCs, Ph3SbS was utilized as a useful single S atom
transfer reagent to NHCs as shown in eq 7.
Ph3SbS þ SIPr f Ph3Sb þ SIPrdS
ð7Þ
It was hoped in the current work to extend this to the OAT by
using MesCNO as a single O atom transfer reagent for NHCs as
shown in eq 8.
To our surprise, full oxidation as shown in reaction 8 did not
occur. The potential oxidation of the NHC to a cyclic urea
stopped at simple adduct formation, and we were able to isolate
and structurally characterize the blue SIPr MesCNO complex in
3
which the lone pair of electrons of SIPr binds to the electropos-
itive C atom of MesCNO. NHC ligands such as SIPr strongly
resemble R3P ligands in their bonding to both metal and
nonmetal acids.27 Isolation of the SIPr MesCNO adduct led
2MesCNO f 2MesCN þ O2
3
ΔH ¼ ꢀ 14:5 kcalmolꢀ1
ð4Þ
us to postulate that the observed kinetic OAT facility of
MesCNO to phosphines might be conferred by initial attack of
R3P at the MesCNO nitrile carbon atom to form an adduct
The endothermic nature of O2 addition to MesCN suggests
that this reagent may be of value for “difficult” oxidations. Indeed
unlike both N2O and amine-N-oxides, nitrile oxides are known to
effect the oxidation of tertiary phosphines within minutes upon
mixing in solution at room temperature according to reaction 5.20
The reasons why reaction 1 requires a catalyst and reaction 5
related to that observed in SIPr MesCNO. That hypothesis was
3
confirmed by detailed computational studies of the reaction
profile which are also reported here.
The primary focus of the current work is the energetics and
mechanism of oxidation of nonmetals by MesCNO. Limited
9621
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
preliminary results on reactions of transition metal complexes
with MesCNO are reported here to confirm mechanistic ideas
regarding these nonmetal reactions. Detailed kinetic and ther-
modynamic studies of metal complex oxidations by MesCNO
will be reported separately.28
reflections was applied using the program SADABS. All structures were
solved by a combination of direct methods and difference Fourier
syntheses, and refined by full-matrix least-squares on F2, by using the
SHELXTL software package.32 Crystal data, data collection parameters,
and results of the analyses are listed in the Supporting Information.
The systematic absences in the intensity data were consistent with the
unique space group P21/c. Half a molecule of toluene from the solvent of
crystallization cocrystallized with the complex and is present in the
asymmetric crystal unit. The toluene molecule is disordered about an
inversion center, and was modeled in part using geometric restraints and
refined with isotropic thermal parameters. However, this disorder could
not be completely resolved, which resulted in the relatively high R value
(7.46%). There is not much disorder present in the nonsolvent part of
the structure. An ORTEP drawing of the structure is shown in the
Results section, see Figure 3. Tables of crystallographic data and the .cif
file are available in the Supporting Information.
’ EXPERIMENTAL SECTION
Unless stated otherwise, all operations were performed in a Vacuum
Atmospheres or MBraun drybox under an atmosphere of purified
nitrogen or argon. Diethyl ether and toluene were dried and deoxyge-
nated by the method of Grubbs.29 C6D6 was purchased from Cambridge
Isotopes. The deuterated solvents were degassed and dried over 4 Å
molecular sieves. The 4 Å sieves and Celite were dried in vacuo overnight
at a temperature just above 200 °C. Mesityl nitrile oxidewasprepared and
recrystallized according to the literature.30 All other compounds were
used asreceived. 1H NMRspectrawere recordedon a BrukerAvance-400
spectrometer at room temperature. 1H NMR chemical shifts are reported
in parts per million (ppm) with respect to the protio impurities
referenced at 7.16 ppm for C6D6 and 2.09 ppm for toluene-d8. FTIR
spectra were obtained using a Perkin-Elmer Spectrum 400 FTIR Spectro-
meter. Calorimetric measurements were made using a Setaram C-80
Calvet microcalorimeter.
Calorimetric Measurement of Reaction of (Ar[tBu]N)3MoP
and R3P with MesCNO. In the glovebox a solution of 0.5093 g
(Ar[tBu]N)3MoP (0.776 mmol) was dissolved in 6 mL of C6D6, and
1 mL of this stock solution was loaded into an NMR tube. The remaining
5 mL of solution were loaded into the Calvet calorimeter cell with
MesCNO (0.0100 g, 0.062 mmol) as limiting reagent. The calorimeter
cell was sealed, taken from the glovebox, and loaded into the Setaram
C-80 calorimeter. Following temperature equilibration, the reaction was
initiated, and the calorimeter rotated to achieve mixing. Following return
to baseline the calorimeter cell was taken into the glovebox, opened, and
1 mL of the solution loaded into an NMR tube. NMR spectra of both the
stock solution and the calorimetry solution were then acquired, and the
reaction was confirmed as quantitative. The enthalpy of three measure-
ments done in this way led to ΔH = ꢀ52.3 ( 0.7 kcal molꢀ1 based on the
Synthesis of (Ar[tBu]N)3MoPO using MesCNO. In the glove-
box, 299 mg (0.456 mmol) of bright yellow (Ar[tBu]N)3MoP21 was
dissolved in 5 mL of CH2Cl2, and the solution was cooled to ꢀ35 °C. A
solution of MesCNO, 84 mg (1.1 equiv) in 2 mL of CH2Cl2 (also cooled
to ꢀ35 °C) was added rapidly to the yellow solution. An additional 1 mL
of chilled CH2Cl2 was used to ensure that all MesCNO had been
transferred to the reaction mixture. Upon warming to room tempera-
ture, the reaction mixture changed from yellow to deep purple in color.
The mixture was stirred for 1 h after mixing to ensure complete reaction.
The mixture was then filtered through a bed of Celite, which was washed
with 5 mL of chilled CH2Cl2. The filtrate was then reduced in volume to
about 50% of its original volume. The purple reaction mixture was
cooled, and 30 mL of thawing acetonitrile was added to precipitate the
product. The cold mixture was filtered on a medium-porosity glass frit,
and the purple solid was washed with 10 mL of cold acetonitrile. The
solid was dried under reduced pressure and yielded (Ar[tBu]N)3MoPO
as a deep purple powder (175 mg, 0.261 mmol, 57%). 1H and 31P NMR
data agreed with previously published data.12
reaction (Ar[tBu]N)3MoP (tol. sol.)
+ MesCNO (solid) f
(Ar[tBu]N)3MoPO (tol. sol.) + MesCN (tol sol.). To these data the
enthalpy of solution of MesCNO in toluene (+4.3 ( 0.1 kcal molꢀ1) was
subtracted to give ΔHrxn = ꢀ56.6 ( 0.8 kcal molꢀ1 with all species in
toluene solution. Reactions of R3P were performed in a similar fashion,
but typically with five independent measurements on each phosphine.
UVꢀvis Kinetic Study of Reaction of (Ar[tBu]N)3MoP and
MesCNO. Solid (Ar[tBu]N)3MoP and MesCNO were dissolved in dry
dichloromethane or dry toluene under Ar in an MBraun glovebox. The
solutions of (Ar[tBu]N)3MoP were loaded into 1 cm path length airtight
quartz cuvettes, and solutions of MesCNO were loaded into Hamilton
gastight syringes. Initial spectra of (Ar[tBu]N)3MoP (0.6 mM) were
collected on a JASCO V-570 UVꢀvis/NIR spectrophotometer, the
MesCNO solution was added to the cuvette via syringe, and the reaction
was monitored in 1ꢀ10 min intervals over a total time of 20ꢀ60 min.
Variable temperature (15ꢀ45 °C) measurements were achieved using the
JASCO PSC-498T temperature controller. The kinetic experiments were
run under pseudo first order conditions with excess MesCNO
(25 mMꢀ100 mM) by monitoring the absorbance increase at λ =
550 nm. Data analysis was performed with IGORpro 5.0 by WaveMetrics.
FTIR Kinetic Study of the Reaction of MesCNO with R3P,
and (Ar[tBu]N)3MoP. In the glovebox, a stock solution of MesCNO
was prepared (0.125 g in 20 mL of freshly distilled toluene). In a separate
vial, 0.100 g of (Ar[tBu]N)3MoP was dissolved in 2 mL of toluene. A
5 mL syringe was loaded with 4 mL of the MesCNO stock solution, and a
2.5 mL syringe was used to load the 2.0 mL of (Ar[tBu]N)3MoP in
toluene. The 4.0 mL solution was loaded into a thermostatted reaction
vessel and allowed to equilibrate with respect to temperature. The
2.0 mL (Ar[tBu]N)3MoP solution was then injected into the reactor and
the timer started. The thermostatted reaction vessel was fitted via thick
wall Teflon tubing lines to a thermostatted FTIR cell kept in a
temperature and environment controlled chamber. A total of 4.0 mL
of solution was flushed under Ar pressure through the tubing and cell
through a valve with tubing leading to a vent. The valve was closed so
Synthesis of the SIPR MesCNO Adduct. In the glovebox,
3
0.400 g of SIPr in 10 mL of toluene was added to 0.1657 g of MesCNO,
and the mixture was shaken to dissolve the MesCNO. The solution
turned blue-violet, and some solid began to precipitate out. It was left
undisturbed in the glovebox overnight. Evaporation of toluene and
addition of heptane led to isolation of the pure adduct in 95% yield.
Analysis for C37H49N3O: theory (found): C 80.5 (79.2), H 8.9 (8.9), N
7.6 (7.0). The mass spectrum showed a strong peak at 552.39 corre-
sponding to P + 1 for C37H49N3O = 551.38. 1H NMR (400M, C6D6):
δ = 6.89 (t, 2H, p-Ph of SIPr), 6.76 (d, 4H, m-Ph of SIPr), 6.45 (s, m-Ph of
MesCNO), 3.64 (t, 4H, CH2 of SIPr), 3.32 (septet, 4H, CH of iPr), 2.04
(s, 3H, p-CH3 of MesCNO), 1.84 (s, 6H, o-CH3 of MesCNO), 1.26 (d,
12H, CH3 of iPr), 0.95 (d, 12H, CH3 of iPr) ppm.
Crystallographic Analyses. Violet single crystals of SIPr MesC-
3
NO suitable for X-ray diffraction analyses obtained by evaporation of
solutions of toluene at 25 °C, crystallized in the monoclinic crystal
system. The data crystal was glued onto the end of a thin glass fiber.
X-ray intensity data were measured by using a Bruker SMART APEX2
CCD-based diffractometer using Mo Kα radiation (λ = 0.71073 Å). The
raw data frames were integrated with the SAINT+ program by using a
narrow-frame integration algorithm.31 Corrections for Lorentz and
polarization effects were also applied with SAINT+. An empirical
absorption correction based on the multiple measurement of equivalent
9622
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
that under Ar pressure, the thermostatted cell was filled with a solution
0.0258 M in MesCNO and 0.0254 M in (Ar[tBu]N)3MoP. A series of
FTIR spectra were collected of the reaction mixture approximately every
20 s through the first two half-lives, and at a slightly slower rate as the
reaction progressed. Data for the nearly equimolar reaction were
analyzed by standard techniques and found to obey second order
kinetics. Averaged rate data are collected in Table 3 for reactions studied
by FTIR spectroscopy.
Table 1. Enthalpies of Reaction of MesCNO and AnP in
Toluene Solution and Derived PdO BDE Dataa
AnP
ΔH
PdO BDE
PdS BDE
(Ar[tBu]N)3MoP
Ph3P
ꢀ56.6 ( 0.8
ꢀ79.9 ( 1.7
ꢀ85.3 ( 1.8
ꢀ86.2 ( 1.7
108.9
132.2
137.6
138.5
78
88
98
94
Cy3P
Me3P
Reaction of the SIPr MesCNO adduct with (Ar[tBu]N)3V
3
a In kcal molꢀ1. For comparison purposes previously reported PdS
and (Ar[tBu]N)3Mo. In the glovebox, 0.010 g of (Ar[tBu]N)3V was
data26 are also included.
dissolved into 1 mL of C6D6, and this solution was syringed onto 0.0095
g of the blue SIPr MesCNO adduct in a vial. The mixture was shaken,
3
polarization functions are added. Single-point energy calculations are
carried out on MP2(full)/6-31G(d) optimized geometries, incorporat-
ing scaled HF/6-31G(d) zero-point vibrational energies, a so-called
higher-level correction to accommodate remaining deficiencies, and
spinꢀorbit correction for atomic species only.
and there was an immediate change in color to red-orange. The sample
was transferred to an NMR tube. Analysis of the NMR spectrum showed
complete conversion to MesCN, (Ar[tBu]N)3VO, and free SIPr. Similar
results were obtained using (Ar[tBu]N)3Mo.
Attempted Reaction of SIPr MesCNO and (Ar[tBu]N)3MoP
3
and Cy3P. In the glovebox, 0.0295 g (Ar[tBu]N)3MoP was dissolved
’ RESULTS
into 2 mL of C6D6, and this solution was syringed onto 0.025 g of the
blue SIPr MesCNO adduct in a vial. The mixture was shaken yielding a
green solution. The sample was transferred to an NMR tube. Analysis of
the NMR spectrum showed no reaction had occurred after a 3 h period.
Reaction of MesCNO and (Ar[tBu]N)3MoP. The initial target
of this work was improved synthesis and determination of the
thermochemistry of the previously characterized purple-blue
complex (Ar[tBu]N)3MoPO.21 Attempted preparation from
PyO as shown in reaction 9 did not occur at room temperature
in toluene solution, which is in keeping with the known inability
of PyO to oxidize phosphines.
3
Similar results were obtained in the attempted reaction of SIPr MesC-
3
NO and Cy3P. In this case the solution remained blue in color, and no
reaction could be detected by NMR spectroscopy.
Attempted Reaction of MesCNO and Pyridine, Tetrahy-
drothiophene and (Ar[tBu]N)3MoN. Several reagents were
checked qualitatively at room temperature for reactivity with MesCNO.
A total of 27.9 mg of MesCNO was dissolved in 1 mL of toluene. To this
was added 1 mL of pyridine, and the solution was monitored by FTIR
spectroscopy. No decrease in the band due to MesCNO was observed
over approximately 4 h. In similar experiments, tetrahydrothiophene and
also (Ar[tBu]N)3MoN were found to not react at room temperature as
followed by NMR and FTIR spectroscopy.
In contrast, reaction of MesCNO with the bright yellow
complex (Ar[tBu]N)3MoP was found to occur rapidly in toluene
solution to yield purple (Ar[tBu]N)3MoPO and MesCN in
quantitative yield as determined by NMR spectroscopy and
shown in reaction 10.
Computational Details. Electronic structure calculations were
carried out using the B3LYP33,34 and M05-2X35 density functionals with
the 6-311G(d,p) and 6-311G(3df,2p) basis sets as implemented in the
Gaussian 09 suite of programs.36 Minimum energy and transition state
structures were optimized by computing analytical energy gradients.
The obtained stationary points were characterized by performing energy
second derivatives, confirming them as minima or first order saddle
points by the number of negative eigenvalues of the Hessian matrix of
the energy (zero and one negative eigenvalues respectively). Computed
electronic energies were corrected for zero-point energy, thermal
energy, and entropic effects to obtain the corresponding thermodynamic
properties H0 and G0. To derive binding energies, the basis set super-
position error (BSSE) was computed using counterpoise calculations.37
For the metal-containing species (Ar[tBu]N)3MoP and (Ar[tBu]N)3MoPO,
optimizations were performed using the Stuttgart-Dresden MWB2838
quasi-relativistic effective core potential and basis including a set
of additional f functions for Mo and the triple-ζ quality basis set
(6-311G(d,p)) for all other elements.
ðAr½tBuꢁNÞ3MoP
þ MesCNO f ðAr½tBuꢁNÞ3MoPO þ MesCN
ð10Þ
Previous preparation of (Ar[tBu]N)3MoPO involved reaction
with dimethyldioxirane at ꢀ78 °C. The pathway in reaction 10
provides a more convenient method for in situ generation of
(Ar[tBu]N)3MoPO and basis for further exploration of its
reactivity.21b
Thermochemistry of OAT from MesCNO to (Ar[tBu]N)3-
MoP or R3P. Solution calorimetric measurements of reactions 5
and 10 were performed in toluene solution at 30 °C using Calvet
calorimetry with solid mesityl nitrile oxide as limiting reagent.
Values for the enthalpies of reaction with all species in toluene
solution are collected in Table 1 together with bond strength
estimates and earlier reported data for sulfur atom transfer
(SAT).26 Data on the BDE in toluene solution are considered
accurate to (3 kcal molꢀ1
.
Intrinsic reaction coordinate (IRC) calculations39 were done to
describe the reaction mechanism for PhCNO and Me3P, providing
the connection between the minimum energy points through the
different transitions states. Further optimization of the final points of
the IRCs with steepest descent algorithm was done to obtain the real
minimum energy structures.
Moreover, the energy of the compounds studied was calculated using
Gaussian-n theory at the G3 level when applicable.40 G3 corresponds
effectively to calculations at the QCISD(T)/G3large level, G3large
being a modification of the 6-311+G(3df,2p) basis set, including more
polarization functions for the second row (3d2f), less on the first row
(2df), and other changes to improve uniformity. In addition, some core
Computational Data on XꢀO Bond Dissociation Enthal-
pies. The XꢀO bond dissociation enthalpy of several species was
calculated according to the procedure described by Lee and
Holm13c computing the enthalpy of reaction with molecular
oxygen for several X/XO couples as shown in Scheme 1 (reaction
(ii)). The XꢀO bond dissociation enthalpy corresponding to
reaction (i) in Scheme 1 was then derived as BDE = ꢀΔHX/XO
+
ΔH0m(O(g)), where ΔH0m(O(g)) = 59.55 kcal molꢀ1 41
.
Computed XꢀO bond dissociation enthalpies performed at
the M05-2X/6-311G(3df,2p) and G3 level, together with experi-
mental values, are collected in Table 2. G3 theory is known for
9623
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
Scheme 1. Thermodynamic Cycle Used to Obtain Computa-
tionally the BDE
Table 2. BDEs (in kcal molꢀ1) for the Reaction XO(g) f
X(g) + O(g) Computed at the M05-2X/6-311G(3df,2p) and
G3 Levels, and Compared to Experimental Dataa
X
M05-2X/6-311G(3df,2p)
G3
experimentalb
NN
38.0 [ꢀ2.0]
48.5
40.8 [0.8]
50.0
40.0
PhCN
MesCN
Pyd
48.1 [ꢀ4.2]
61.7 [ꢀ1.6]
74.6
48.8 [ꢀ3.5]
63.6 [0.3]
78.3
52.3c
63.3
delivering chemical accuracy; however, its high computational
cost most often give preference to the less computationally
demanding DFT methods. Cundari and co-workers42 have
shown that the G3 method performs within the experimental
error in calculating the thermochemistry of OAT reactions, while
DFT methods such as the popular B3LYP functional provided
inadequate results. As can be seen in Table 2, a good agreement
between experimental data and calculated values at the M05-2X/
6-311G(3df,2p) level is obtained for all the species studied. For
the smallest molecules where the G3 methodology is applicable,
similar results are obtained at the M05-2X/6-311G(3df,2p) level.
Calculated values using the B3LYP functional give worse results
as previously shown by Cundari and co-workers42 giving the
worst agreement with the experimental (or G3-calculated) data
with the phosphines (Supporting Information, Table S1).
NP
THTe
Me2S
PP
82.1
84.8
88.3 [1.7]
86.0 [ꢀ0.6]
90.1
86.6
90.1
(Ar[tBu]N)3MoP
105.9 [ꢀ3.0]f
128.2 [1.0]
130.2 [ꢀ2.0]
135.3 [ꢀ2.3]
133.0 [ꢀ5.5]
151.5
108.9g
127.2
132.2g
137.6g
138.5g
CO
128.7 [1.5]
Ph3P
Cy3P
Me3P
SIPr
134.9 [ꢀ3.6]
a BDEcalc ꢀ BDEexp is shown in brackets beside the computed values.
b Unless stated otherwise, values taken from ref 19. c Value taken from ref
18. d Pyridine. e Tetrahydrothiophene. f Using the Stuttgart-Dresden
MWB2838 effective core potential and basis including a set of additional
f functions for Mo and the 6-311G(d,p) basis set for all other elements.
g This work.
Synthesis and Structure of SIPr MesCNO. As discussed
3
earlier, we initially studied the reaction of SIPr and MesCNO
in hopes of using this as an entry to measure enthalpies of OAT to
NHC ligands forming cyclic ureas as shown in eq 8. OAT did not
occur in spite of being thermochemically highly favorable (see
Table 2), but simple binding of the SIPr to MesCNO as shown in
reaction 11.
The crystal structure of the adduct shows binding between the
lone pair of electrons of the NHC at the electrophilic C atom of
MesCNO as shown in Figure 3.
Figure 1. Spectral data as a function of time taken at 15 °C in CH2Cl2
with 25 mM MesCNO ([(Ar[tBu]N)3MoP] = 0.6 mM), over a time
interval of 60 min showing the rise in the band at 550 nm because of
increase in [(Ar[tBu]N)3MoPO].
This mode of binding resembles that to CS2 which is displayed
for both NHC and R3P ligands.27a Louie and co-workers27b have
even reported adducts between CO2 and NHCs. On the basis of
the analogous Lewis basicity of NHCs and R3P compounds, the
isolation and characterization of the adduct shown in Figure 3
suggested that attack of R3P at MesCNO might also proceed by
initial attack at the C atom rather than the O atom and that this
was the explanation of why it and not PyO is an efficient OAT
reagent for phosphines. This prompted kinetic and computa-
tional study of the reaction mechanism of phosphines and
MesCNO. Detailed synthetic and computational study of the
adducts between a range of NHCs and MesCNO are in
progress.43
UVꢀvis Kinetic Study of Reaction of MesCNO with
(Ar[tBu]N)3MoP. Kinetic analysis of the reaction of MesCNO
with (Ar[tBu]N)3MoP was performed on a UVꢀvis spectro-
photometer under pseudo first order conditions with large excess
of MesCNO in both CH2Cl2 and toluene solution. Time-
resolved spectra show the appearance and growth of the peak
at 550 nm that corresponds to the absorbance of independently
prepared (Ar[tBu]N)3MoPO. The exponential growth of this
peak was used for kinetic analysis. Kinetic traces were fit to a
single exponential function, and rate constants were obtained,
kobs = k1[MesCNO] (Supporting Information, Figure S1).
Under high concentrations of MesCNO (100 mM), complete
formation of (Ar[tBu]N)3MoPO was followed by its slow partial
decomposition, which could be seen as a minor decrease in the
intensity of the peak at 550 nm and a slow appearance of a
shoulder at 440 nm over longer periods of time (3ꢀ4 h,
Figure 1). The minor decomposition was judged to not com-
promise kinetic analysis of the (Ar[tBu]N)3MoPO formation
step. The identity of the decomposition products was not
established. Plots of kobs versus [MesCNO] gave straight lines
as shown in Supporting Information, Figure S2, and were used to
derive second-order rate constants in CH2Cl2:k15 °C = 0.11Mꢀ1 sꢀ1;
k22 °C = 0.15 Mꢀ1 sꢀ1; k30 °C = 0.20 Mꢀ1 sꢀ1
.
Identical kinetic experiments were repeated in toluene solu-
tion. Reaction rate as a function of temperature was measured
with [MesCNO] = 50 mM and [(Ar[tBu]N)3MoP] = 0.6 mM at
9624
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
Figure 2. Eyring plot for reaction between MesCNO and (Ar[tBu]N)3-
MoP in toluene.
Figure 4. Eyring plotsfor reactionofMesCNOandSIPr(topline, brown
square); Me3P (second line, blue squares), Cy3P (third line, green
triangles), (p-tolyl)3P (fourth line, pink square), and (Ar[tBu]N)3MoP
(bottom line, orange triangles).
Table 3. Rate Constants at ≈20 °C and Derived Activation
Parameters As Well As Reaction Enthalpies for Interaction
between X and MesCNO
ΔHq
ΔSq
ΔH0
X
k (Mꢀ1 sꢀ1) (kcal molꢀ1) (cal molꢀ1 Kꢀ1) (kcal molꢀ1
)
Me3P
Cy3P
1.0
11.5
9.3
ꢀ19.5
ꢀ28.4
ꢀ30.9
ꢀ26.6
ꢀ27.0
ꢀ30.7
ꢀ86.2
ꢀ85.3
ꢀ79.9a
ꢀ56.6
ꢀ56.6
0.45
0.20
0.07
0.04
3.8
(p-tolyl)3P
9.1
(Ar[tBu]N)3MoPb
(Ar[tBu]N)3MoPc
SIPR
11.0
11.0
7.4
Figure 3. Thermal ellipsoid plot (40% probability) of the SIPr MesC-
3
NO adduct. Selected distances (Å) and angles (deg): C1ꢀC8, 1.521(4);
C8ꢀN2, 1.326(2); C10ꢀN2, 1.451(2); N1ꢀC1ꢀC8ꢀN2, 84.6. Com-
plete structural data is available in Supporting Information.
a Refers to reaction of Ph3P, expected to be similar to (p-tolyl)3P.
b Values obtained by FTIR spectroscopy. c Values obtained by UVꢀvis
spectrophotometry.
five different temperatures: 15, 22, 30, 38, and 45 °C. In the
replicate experiments for the same temperature in each solvent,
kobs in CH2Cl2 and toluene were found to be similar as shown in
the Supporting Information, Table S2. An Eyring plot was
generated for the temperature dependence study in toluene
solution, and activation parameters were calculated (Figure 2).
The graphical analysis provided ΔHq = 11 kcal molꢀ1 and ΔSq =
Cy3P and MesCNO displayed isosbestic points, and the rate of
decay of MesCNO and the rate of buildup of Cy3PO were equal
and opposite in sign. Eyring plots are shown in Figure 4; rate
constants and activation parameters are shown in Table 3.
The entropies of activation are near ꢀ25 cal molꢀ1 Kꢀ1 in
keeping with an associative transition state. The least unfavorable
entropy of activation occurs, as might be expected, for Me3P
since it is sterically the least hindered. The enthalpies of activa-
tion are in the range ΔHq = 10 ( 2 kcal molꢀ1, except for SIPr
which involves only binding and not oxidative addition. The data
determined by FTIR kinetics for (Ar[tBu]N)3MoP are in good
agreement with those determined independently by UVꢀvis
studies. The rate and activation parameters in Table 3 appear in
keeping with a common mechanism to these reactions.
ꢀ27 cal molꢀ1 Kꢀ1
.
FTIR Kinetic Studies of Reaction of MesCNO with R3P,
(Ar[tBu]N)3MoP and SIPr. To compare the rates of oxidation of
tertiary phosphines to that of (Ar[tBu]N)3MoP, it was necessary
to use vibrational spectroscopy because there is no observable
color change when the phosphines are oxidized to the phosphine
oxides because both are colorless. In addition, discovery of binding
of SIPr to MesCNO was investigated kinetically for purposes of
comparison of simple binding of the NHC to binding plus OAT
which occurs for the phosphines and phosphide substrates.
Reactions were run in a thermostatted FTIR cell under argon
atmosphere at a balanced 1:1 ratio of [MesCNO]:[R3P]. As a
check, the kinetics of (Ar[tBu]N)3MoP oxidation was also in-
vestigated by FTIR in toluene. A typical difference plot for
oxidation of (Ar[tBu]N)3MoP by MesCNO is shown in Support-
ing Information, Figure S3, and the linear second order plot
derived from it in Supporting Information, Figure S4.
Theoretical Mechanism for Reaction of PhCNO and Me3P.
Experimental work described above prompted theoretical inves-
tigation of OAT between Me3P and PhCNO as a model system.
The reason OAT does not occur in the SIPr MesCNO adduct is
3
under computational study as part of a more detailed investiga-
tion of these adducts.43 The frontier orbitals for Me3P and
PhCNO calculated at the M05-2X/6-311G(3df,2p) level are
shown in Figure 5. The shape of the orbitals suggests that a
probable reaction pathway would involve attack of the highest
Attempts to detect intermediates in these reactions were not
successful. Even at ꢀ40 °C, FTIR spectral data for the reaction of
9625
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
Me3P was found to be repulsive and did not lead to reaction.
A transition state higher in energy with a syn disposition
of the CipsoꢀCꢀNꢀO moiety was also found (TS1b) which is
shown in Supporting Information, Figure S5. This struc-
ture resembles that of the SIPr MesCNO adduct shown in
3
Figure 3. The lower energy of TS1 compared to TS1b can be
attributed to electrostatic attraction between the positively
charged P atom and the negatively charged O atom. Presum-
ably steric forces do not allow this to be established for the
SIPr MesCNO adduct.43
3
Reaction of the SIPr MesCNO Adducts with R3P, (Ar[tBu]N)3-
3
MoP, or (Ar[tBu]N)3V. In view of the proposed mechanism of
reaction of MesCNO with phosphines, it was of interest to see
how blocking the electrophilic C atom of MesCNO in the
SIPr MesCNO adduct would affect its OAT reactivity. In that
regard, it was important to first establish that the OAT ability of
3
Figure 5. Computed frontier orbitals for Me3P and PhCNO. (A)
HOMO of PhCNO; (B) LUMO of PhCNO; (C) HOMO of Me3P;
(D) LUMO of Me3P.
the SIPr MesCNO adduct was not reduced significantly with
3
respect to free MesCNO. Therefore, reactions 12 and 13 were
investigated qualitatively.
ðAr½tBuꢁNÞ3V þ MesCNO f ðAr½tBuꢁNÞ3VO þ MesCN
ð12Þ
On the basis of visual observation, both reactions 12 and 13
occurred immediately upon mixing at room temperature.
Stopped flow kinetic studies have shown that at room tempera-
ture, reaction of the adduct (reaction 13) is actually more rapid
than OAT of MesCNO itself (reaction 12).28 It is clear that no
significant kinetic reduction in OAT ability through the O atom
site occurred for the SIPr MesCNO adduct compared to free
3
MesCNO in its reaction with (Ar[tBu]N)3V. Thus it can be
concluded that reaction of coordinatively unsaturated metal
Figure 6. IRCs calculated at the M05-2X/6-311G(3df,2p) level.
Relative enthalpies (in kcal molꢀ1) (BSSE corrected), entropies between
parentheses (in cal molꢀ1 Kꢀ1), and Gibbs energies between brackets
(kcal molꢀ1) at T = 298 K.
complexes with the SIPr MesCNO adduct is not impaired by
3
coordination of the NHC; in fact, it is slightly enhanced.28
In contrast, reactivity with R3P is shut down when an NHC
coordinates and blocks the C atom binding site of MesCNO.
Thus, reaction 14 did not proceed at room temperature over a
four hour period.
occupied molecular orbital (HOMO) of Me3P, which is essen-
tially a lone pair orbital (C in Figure 5), to the lowest unoccupied
molecular orbital (LUMO) of PhCNO (B in Figure 5), which is a
delocalized π* orbital with a significant lobe on the C atom
of PhCNO.
This is confirmed in the reaction mechanism obtained by
computing the intrinsic reaction coordinate (IRC) at the same
level of theory which is shown in Figure 6. The mechanism is
composed of three consecutive stages, where three transition
states have to be overcome, the first one assigned as the rate-
limiting step. The structures of the computed transition states
and intermediates are shown in Figure 7. Distances calculated at
the M05-2X/6-311G(3df,2p) level for the reactants, products,
transition states, and intermediates in Figure 6 are available in the
Supporting Information, Table S3.
Several alternative mechanisms were also considered.
Neither minima nor transition states were located corresponding
to a 1,3-cycloaddition of PhNCO to Me3P according to a
concerted mechanism. Likewise, direct OAT from PhNCO to
In a similar way (Ar[tBu]N)3MoP, which is efficiently oxidized
by free MesCNO, undergoes no reaction with the SIPr MesC-
3
NO adduct.
Cases Where MesCNO Did Not Perform OAT. In addition
to the inability of MesCNO to oxidize the N-hetero-
cyclic carbenes to form the corresponding cyclic ureas, several
reactions that are thermodynamically favorable for OAT
from MesCNO were found not to occur at room temperature
in toluene solution. In spite of the fact that reaction 10
occurs, corresponding reaction of (Ar[tBu]N)3MoN as shown
9626
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
Figure 7. Computed structures of intermediates and transition states for reaction of Me3P and PhNCO at the M05-2X/6-311G(3df,2p) level. See
Figure 5 for relative energies.
in reaction 15 was not observed.44
found to differ by ≈5 kcal molꢀ1 from work done by Mortimer
using static bomb calorimetry.47 Classic work in reaction calo-
rimetry by Skinner and co-workers on enthalpies of reaction of
trialkyl phosphines with H2O2 includes in its final paragraph the
possibility that the oxidation reactions may not have been
quantitative because of possible formation of trialkyl phosphine
peroxides and concludes with the statement: “In our opinion, the
thermochemical results given here need independent verification,
possibly by oxidation studies based on some other agent than hydrogen
peroxide”.48 The value for the PdO BDE in Ph3PO determined in
this work in toluene solution of 132.2 ( 3.0 kcal molꢀ1 is in good
agreement with the gas phase value of Domalski9 of 135.4 ( 2.8
kcal molꢀ1. The data for Me3P and Cy3P are also in reasonable
agreement with values quoted for other trialkyl phosphines
reported in the literature.49 In addition, computed BDE values
at the M05-2X/6-311G(3df,2p) level are in excellent agreement
with experimental data as shown in Table 2.
It is of interest to compare the data in Table 1 for OAT to
earlier data for SAT. Surprisingly, there is a smaller gap between
Ph3P and Cy3P bonded to O than to S. This may be due to a
greater significance of π bonding in the PdO compared to
PdS bond in R3PE (E = O, S). Thus, the weaker basicity of
Ph3P compared to Cy3P may be compensated in its bonding to O
by an increased π bonding capacity in Ph3P which would be
expected to be a poorer σ donor but better π acceptor than Cy3P.
For S, since π-bonding would be expected to be of less impor-
tance than for O, the difference between Ph3P and Cy3P would be
greater since it is now dominated by the σ-donor ability of the
phosphine.
Oxidation of tetrahydrothiophene and of pyridine (reactions
16 and 17) did not occur over 4 h at room temperature.
Attempts to oxidize dimethyl sulfoxide with MesCNO also
showed no reaction at room temperature over several hours.
’ DISCUSSION
MesCNO has proven to be a useful reagent for OAT to
phosphines and phosphides for the systems shown in Table 1,
and this work is now being extended to metal complexes.28 OAT
reactions of MesCNO can follow two different reaction channels:
reaction with a Lewis acid may occur at the terminal O, but OAT
reactions with a Lewis base may occur by attack at the electro-
philic C atom. This second mechanistic possibility was envi-
sioned only after stable adducts between NHC and MesCNO
were discovered and shown to involve coordination of the
nucleophilic carbene to the electrophilic C atom of MesCNO
as shown in Figure 3. Since there is a strong similarity between
NHCs and R3P in their coordination chemistry to metals,45 this
suggested that nonmetal oxidation by MesCNO may proceed
through a similar adduct for the phosphine as that found to be
stable for the NHC. The discussion focuses on two areas: (i) the
thermochemistry of OAT and its implication for the MoꢀP bond
in (Ar[tBu]N)3MoPO, and (ii) the mechanism of OAT to
phosphines and phosphides.
The PdO BDE in Me3PO (138.5 kcal molꢀ1) is similar to the
fundamental PdO BDE (140.8 kcal molꢀ1) for the gas phase
diatomic molecule phosphorus monoxide. The enthalpy of OAT
from MesCNO to (Ar[tBu]N)3MoP is nearly 30 kcal molꢀ1
lower than that for Me3P and yields a very low PꢀO BDE.
Detailed discussion of why the PꢀO BDE in (Ar[tBu]N)3MoPO
is so low is not warranted beyond noting that upon loss of
an O atom, the Mo complex has a stronger MoꢀP bond (MoꢀP
BDE of ꢀ92.2 kcal molꢀ1 in (Ar[tBu]N)3MoP vs ꢀ60.2 kcal
molꢀ1 in (Ar[tBu]N)3MoPO, see below), whereas this does not
happen in the case of phosphines. Furthermore, it was considered
of interest to compare the data for (Ar[tBu]N)3MoPO to
PtPdO and NtPdO. Computational studies (see Table 2)
Thermochemistry of OAT Reactions. Much thermochemical
data on organophosphines date back to work done in the 1950s
to 1960s, and modern compilations of experimental data usually
can be traced back to work done a half century ago. An exception
to that is Domalski’s determination by rotating bomb calorimetry
of accurate data for Ph3P and Ph3PO.46 The data for Ph3P were
9627
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
Scheme 2. Measured and Derived Enthalpies for the
(Ar[tBu]N)3MoEO (E = P, N)a
product states.
ΔH ¼9f2:2
ðAr½tBuꢁNÞ3MoP
ðAr½tBuꢁNÞ3Mo þ P
ð18Þ
kcalmolꢀ1
ΔH ¼6f0:2
ðAr½tBuꢁNÞ3MoPO
ðAr½tBuꢁNÞ3Mo þ PO
kcalmolꢀ1
ð19Þ
The difference between the two BDEs is related to the fact that
when (Ar[tBu]N)3MoPO dissociates it produces the stable
molecule PO. There is a large increase in the PꢀO BDE which
follows that dissociation. Subtracting reaction 18 from reaction
19 yields reaction 20.
a All data in kcal molꢀ1. For data relevant to (Ar[tBu]N)3MoNO
binding and the (Ar[tBu]N)3MoN BDE see ref 52.
yielded the following values for the PꢀO bond dissociation
enthalpy (kcal molꢀ1): (Ar[tBu]N)3MoPO [105.9], PtPdO
[90.1], and NtPdO [78.3]. This trend appears to follow
the expected decreasing basicity of the lone pair on P for
(Ar[tBu]N)3MoP versus PtP versus NdP. Moreover, dissocia-
tion of an O atom from (Ar[tBu]N)3MoPO forms (Ar[tBu]N)3MoP;
dissociation of an O from PPO forms PtP, and from NPO forms
NtP. Since it might reasonably be expected that the strongest
triple bond in that series would be NtP, then it seems reasonable
the NPO would be more susceptible to O loss.
ΔH ¼ ꢀ 32
kcalmolꢀ1
ðAr½tBuꢁNÞ3MoPO þ P
ðAr½tBuꢁNÞ3MoP þ PO
ð20Þ
f
The difference in BDEs relates directly to the OAT value in
reaction 20. The fact that PO has a stronger PꢀO Bond (BDE =
140.8 kcal molꢀ1) than does (Ar[tBu]N)3MoPO (BDE = 108.9
kcal molꢀ1) accounts for the apparent discrepancy in bond
length/BDE.
Since the enthalpy of formation of PO is known50 the
thermochemical data determined previously by us for the MoꢀP
bond in (Ar[tBu]N)3MoP51 can be combined with the oxidation
of (Ar[tBu]N)3MoP to (Ar[tBu]N)3MoPO reported here to cal-
culate the enthalpy of coordination of PO to the (Ar[tBu]N)3Mo
fragment as shown in Scheme 2. Data are also included in
Scheme 2 for N for comparison purposes.52 The 60.2 kcal molꢀ1
MoꢀP BDE in (Ar[tBu]N)3MoPO represents a strong metalꢀ
ligand bond but is still lower than the MoꢀN BDE in
Mechanism of OAT from MesCNO to Phosphines and
Phosphides. Computational studies suggest that attack by
Me3P at the O atom of MesCNO leads to a repulsive surface.
The computed mechanism entails a gradual rise in energy as the
reactants approach each other at 4ꢀ5 Å to make an adduct
between the Lewis basic lone pair of electrons on Me3P, and the
Lewis acid site at the electrophilic C atom of MesCNO. The first
transition state (TS1) has the highest energy along the pathway
in keeping with experimental failure to observe buildup of any
intermediate adduct in the reaction by FTIR or UVꢀvis spectro-
scopic studies. The calculated values of ΔHq = 11.6 kcal molꢀ1
and ΔSq = ꢀ38.1 cal molꢀ1 Kꢀ1 are in reasonable agreement
with the experimental values determined by FTIR spectroscopy
in Table 3. In the computed structure of Int1, corresponding to
an adduct between Me3P and the C atom of MesCNO, there is an
electrostatic interaction between the positive P atom (formal
charge = +1) in the adduct and the O of the NO group (formal
charge = ꢀ1). For Me3P steric hindrance to this configuration
may be smaller than for bulkier phosphines. A shallow inter-
mediate state (TS2) is computed along the second part of the
reaction trajectory which corresponds to formation of what is
approximately a cyclic structure in Int2. The enthalpy of activa-
tion for the last transition state (TS3) is ΔHq = 9.9 kcal molꢀ1
comparable to the energetic barrier in the first step; however, the
entropy of activation is ΔSq = 1.2 cal molꢀ1 Kꢀ1 making this step
more favorable in terms of Gibbs energy. Full establishment of
the PdO bond in Me3PO is a powerful driving force for this
highly exothermic reaction which proceeds smoothly to comple-
tion. This pathway avoids the lone pair-lone pair repulsion that
would be present if R3P underwent direct attack at the O atom of
MesCNO. In this type of mechanism the C atom functions
somewhat like a metal;55 R3P coordinates and as a consequence
of its electron donation develops a partial positive charge. The C
atom gains electron density in this interaction, and this is
delocalized to the O atom of MesCNO. It is noteworthy that
blocking the C atom site by prior coordination of an NHC
disrupts OAT activity for nonmetal but not for metal compounds
(Ar[tBu]N)3MoNO of 82.5 kcal molꢀ1
.
It is noteworthy that based on step (I) in Scheme 2, and a value
of 59.5 kcal molꢀ1 for step (III) of Scheme 1 that oxidation of
(Ar[tBu]N)3MoE by 1/2 O2 to (Ar[tBu]N)3MoEO is exother-
mic by 18.6 kcal molꢀ1 for E = N, and 49.3 kcal molꢀ1 for E = P.
Thus, air oxidation of both the nitride and phosphide to the NO
and PO complexes is thermodynamically favored. Oxidation of
(Ar[tBu]N)3MoN is not as favorable as for (Ar[tBu]N)3MoP
and this may be due to in part to the greater strength of the
MotN bond in (Ar[tBu]N)3MoN. The MoꢀN distance in the
nitride is shorter than in the nitrosyl complex. That is in keeping
with the stronger MoꢀN bond in (Ar[tBu]N)3MoN (BDE=
155.3 kcal molꢀ1) than that for (Ar[tBu]N)3MoNO (BDE =
82.5 kcal molꢀ1). However, as first pointed out by Cummins,21
the MoꢀP distance of 2.079 Å in (Ar[tBu]N)3MoPO is 0.04 Å
shorter than the MoꢀP distance in (Ar[tBu]N)3MoP. This has
led to speculation that the longer MoꢀP bond in the phosphide
may in fact be weaker than that in the (Ar[tBu]N)3MoPO
complex. Computations by Frenking and co-workers53 are in
agreement with the experimental bond distances, and they have
also concluded that the MoꢀP BDE in (Ar[tBu]N)3MoPO is
higher than that of (Ar[tBu]N)3MoP. That conclusion is at
odds with the data in Scheme 2 where the MoꢀP BDE is
nevertheless 30 kcal molꢀ1 higher in the phosphide than in
(Ar[tBu]N)3MoPO in spite of its longer distance. Bond
length and bond strength do not always correlate with each
other.54 The BDE is the enthalpy of a homolysis reaction
(eqs 18 and 19 for the MoꢀP bonds in both species) and thus
depends exclusively on the relative stability of reactant and
9628
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
Scheme 3. Oxidation of Nitrile Oxides by Amine Oxidesa
appears that two criteria exist for the OAT via this mechanism:
(i) the reagent must be nucleophilic enough for adduct formation
at the C atom, (ii) and it must be capable of expanded valence for
a cyclic OAT transition state. Coordination of R3P appears to
satisfy both these requirements. The novel remote site attack
displayed here may be of utility in design of other reagents to
achieve difficult OAT reactions to other nonmetal substrates.
Additional synthetic, mechanistic, and thermochemical studies
are in progress to expand understanding and utilization of OAT
reactivity of nitrile oxides.
a See ref 56.
as shown in reactions 13 and 14. This provides support for two
different mechanisms being operative in OAT with MesCNO.
There is literature precedent for the second proposed mecha-
nism. Oxidation of nitrile oxides by amine oxides has been
proposed to yield nitrosocarbonyls which may be trapped by
dienes or other receptors.56 The proposed first step in this
reaction involves nucleophilic attack of the O of the amine oxide
at the C of MesCNO to produce a zwitterionic intermediate
which then dissociates as shown in Scheme 3. The central
intermediate zwitterion proposed resembles the computed inter-
mediate Int1 in Figure 6.
’ ASSOCIATED CONTENT
S
Supporting Information. Spectroscopic, kinetic, structural,
b
and computational data. This material is available free of charge
via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: c.hoff@miami.edu.
Also of interest are the unsuccessful OAT reactions involving
nonmetals that were surveyed here including tetrahydrothio-
phene, pyridine, and (Ar[tBu]N)3MoN. In spite of being com-
puted to be exothermic in each case, for none of these
combinations was OAT observed. An explanation is that for
OAT to occur by the mechanism outlined here for nonmetals,
the nucleophile must be strong enough to form a zwitterionic
intermediate by nucleophilic attack at the C atom of MesCNO;
however, it must also be capable of forming a cyclic inter-
mediate by utilizing expanded valence. It seems plausible that
(Ar[tBu]N)3MoN and Py, even if they were to form an adduct at
C with MesCNO similar to that observed in the reaction of
MesCNO with NHCs, could not readily undergo valence expan-
sionatthe N atombeyondlonepair coordination. Tetrahydrothio-
phene may be capable of undergoing subsequent OAT chemistry
via 1,3-cycloaddition but may not be basic enough to form an
adduct with MesCNO. At this point, we have found only R3P and,
somewhat surprisingly, (Ar[tBu]N)3MoP capable of doing this.
’ ACKNOWLEDGMENT
Support of this work by the National Science Foundation
Grants CHE 0615743 (CDH), 0750140 (ERA), 0724158
(CCC) and by the Spanish Ministry of Science and Innovation
(MICINN) CTQ2009-07120 (M.T.) is gratefully acknowl-
edged. M.T. and L.M.F. acknowledge MICINN for a “Juan de
la Cierva” and a “Ramꢀon y Cajal” contract. The authors wish to
thank Jared S. Silvia (MIT) for a gift of (Ar[tBu]N)3V and Alex
T. Vai (MIT) for help in preparing (Ar[tBu]N)3MoP.
’ REFERENCES
(1) Nam., W. Acc. Chem. Res. 2007, 40, 465. This is the first article in
an issue dedicated to oxygen activation.
(2) Horvꢀath, H. T.; Anastas, P. T. Chem. Rev. 2007, 107, 2167. This is
the first article in an issue dedicated to green chemistry and containing
several articles on green oxidation.
(3) (a) Parmon, V.; Panov, G.; Uriarte, A.; Noskov, A. Catal. Today
2005, 100, 115. (b) Hintz, P. A.; Sowa, M. B.; Ruatta, S. A.; Anderson,
S. L. J. Chem. Phys. 1991, 94, 6446.
’ SUMMARY
(4) (a) Vaughan, G. A.; Rupert, P. B.; Hillhouse, G. L. J. Am. Chem.
Soc. 1987, 109, 5538. (b) Vaughan, G. A.; Hillhouse, G. L.; Lum, R. T.;
Buchwald, S. L.; Rheingold, A. L. J. Am. Chem. Soc. 1988, 110, 7215.
(c) Vaughan, G. A.; Sofield, C. D.; Hillhouse, G. L.; Rheingold, A. L.
J. Am. Chem. Soc. 1989, 111, 5491. (d) Matsunaga, P. T.; Hillhouse,
G. L.; Rheingold, A. L. J. Am. Chem. Soc. 1993, 115, 2075. (e) Koo, K.;
Hillhouse, G. L.; Rheingold, A. L. Organometallics 1995, 14, 456.
(f) Harrold, N. D.; Waterman, R.; Hillhouse, G. L.; Cundari, T. R.
J. Am. Chem. Soc. 2009, 131, 12872.
(5) (a) Tuan, D. F.-T.; Hoffmann, R. Inorg. Chem. 1985, 24, 871.
(b) Tuan, D. F.-T.; Reed, J. W.; Hoffmann, R. THEOCHEM(J. Mol. Struct.)
1991, 232, 111. (c) Armor, J. N.; Taube, H. J. Am. Chem. Soc. 1969,
91, 6874. (d) Armor, J. N.; Taube, H. J. Am. Chem. Soc. 1970, 92, 2560.
(e) Armor, J. N.; Taube, H. J. Chem. Soc. D 1971, 287. (f) Diamantis,
A. A.; Sparrow, G. J. J. Chem. Soc. D 1969, 469. (g) Diamantis, A. A.;
Sparrow, G. J. J. Chem. Soc. D 1970, 819. (h) Diamantis, A. A.; Sparrow,
G. J. J. Colloid Interface Sci. 1974, 47, 455. (i) Diamantis, A.; Sparrow, G.;
Snow, M.; Norman, T. Aust. J. Chem. 1975, 28, 1231. (j) Bottomley, F.;
Crawford, J. R. J. Chem. Soc. D 1971, 200. (k) Bottomley, F.; Crawford,
J. R. J. Am. Chem. Soc. 1972, 94, 9092. (l) Bottomley, F.; Armor, J. N.
Inorg. Synth. 1976, 75. (m) Bottomley, F.; Brooks, W. V. F. Inorg. Chem.
1977, 16, 501.
Mesityl nitrile oxide is a versatile OAT reagent. Reactivity with
coordinatively unsaturated metal complexes occurs rapidly
through initial coordination of the terminal O atom. Oxidation
of R3P and (Ar[tBu]N)3MoP on the other hand occurs in
minutes at room temperature but follows a different mechanism
in which nucleophilic attack at the electrophilic C atom of
MesCNO is followed by a cyclic transition state leading to
OAT and elimination of MesCN. Reaction of MesCNO with
(Ar[tBu]N)3MoP provides ready entry to (Ar[tBu]N)3MoPO
which is a unique terminal phosphorus monoxide complex.
However, the PꢀO BDE in this complex is some 30 kcal molꢀ1
lower than in R3PO and corresponds more closely to computed
values for PPO or NPO. In spite of the fact that the MoꢀP bond
in (Ar[tBu]N)3MoPO is shorter than in (Ar[tBu]N)3MoP, the
MoꢀP BDE in (Ar[tBu]N)3MoPO (60.2 kcal molꢀ1) is lower
than in (Ar[tBu]N)3MoP (92.2 kcal molꢀ1). This may be
attributed to the fact that dissociation of PO results in an increase
in the PꢀO bond strength for the free molecule compared to its
adduct. Neither (Ar[tBu]N)3MoN, THT, nor Py are oxidized
under the same conditions in which R3P oxidation is facile. It
9629
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630

Inorganic Chemistry
ARTICLE
(6) Piro, N. A.; Lichterman, M. F.; Harman, W. H.; Chang, C. J.
J. Am. Chem. Soc. 2011, 133, 2108.
(7) (a) Otten, E.; Neu, R. C.; Stephan, D. W. J. Am. Chem. Soc. 2009,
131, 9918. (b) Neu, R. C.; Otten, E.; Stephan, D. W. Angew. Chem., Int.
Ed. 2009, 48, 9709.
(33) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(34) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(35) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. J. Chem. Theory Comput.
2006, 2, 364.
(36) Frisch, M. J.; et al. Gaussian 09, revision B.01; Gaussian, Inc.:
(8) Viege, A. S.; Slaughter, L. M.; Wolczanski, P. T.; Matsunaga, N.;
Decker, S. A.; Cundari, T. R. J. Am. Chem. Soc. 2001, 123, 6419.
(9) Abu-Omar, M. M. In Physical Inorganic Chemistry. Reactions,
Processes, and Applications; Bakac, A., Ed.; Wiley: Hoboken, NJ, 2010; p
75 and references cited therein.
Wallingford, CT, 2010.
(37) (a) Simon, S.; Duran, M.; Dannenberg, J. J. J. Chem. Phys. 1996,
105, 11024. (b) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553.
(38) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123.
(10) Cai, Y.; Ellern, A.; Espenson, J. H. Inorg. Chem. 2005, 44, 2560.
(11) Keith, J. M.; Tomil, Z. D.; Zaric, S. D.; Hall, M. B. J. Mol. Catal.
A: Chem. 2010, 324, 15.
(12) Park, J.; Morimoto, Y.; Lee, Y. M.; Nam, W.; Fukuzumi, S. J. Am.
Chem. Soc. 2011, 133, 5236.
(13) (a) Holm, R. H. Chem. Rev. 1987, 87, 1401. (b) Holm, R. H.;
Donahue, J. P. Polyhedron 1993, 12, 571. (c) Lee, S. C.; Holm, R. H.
Inorg. Chim. Acta 2008, 361, 1166.
(39) (a) Fukui, K. Acc. Chem. Res. 1981, 14, 363. (b) Hratchian,
H. P.; Schlegel, H. B. J. Chem. Theory Comput. 2005, 1, 61.
(40) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.;
Pople, J. A. J. Chem. Phys. 1998, 109, 7764.
(41) Cox, J. D.; Wagman, D. D.; Medvedev, V. A. CODATA Key
Values for Thermodynamics; Hemisphere: New York, 1989.
(42) Dinescu, A.; Whiteley, C.; Combs, R. R.; Cundari, T. R. J. Phys.
Chem. A 2006, 110, 4053.
(14) Grundmann, C.; Gr€unanger, P. The nitrile oxides: versatile tools of
theoretical and preparative chemistry; Springer-Verlag: Berlin, Germany,
1971.
(15) Haaland, A. Angew. Chem., Int. Ed. Engl. 1989, 28, 992–1007.
(16) Shiro, M.; Yamakawa, M.; Kubota, T. Acta Crystallogr., Sect. B:
Struct. Sci. 1979, 35, 712–716.
(17) Herzberg, G. Electronic spectra and electronic structure of polya-
tomic molecules; Van Nostrand: New York, 1966; p 1.
(18) Luo, Y. R. Handbook of Bond Dissociation Energies in Organic
Compounds; CRC Press: Boca Raton, FL, 2003.
(19) The authors are not aware of this thermodynamically favored
loss of O2 occurring in either catalyzed or uncatalyzed reactions.
(20) (a) Grundmann, C.; Frommeld, H.-D. J. Org. Chem. 1965,
30, 2077. (b) Sicard, G.; Baceiredo, A.; Crocco, G.; Bertrand, G. Angew.
Chem., Int. Ed. Engl. 1988, 27, 301.
(43) Temprado, M.; Cai, X.; Majumdar, S.; Captain, B.; Hoff, C. D.;
work in progress.
(44) Examples of nitride to NO conversion via OAT have been
reported in the literature: Williams, D. S.; Meyer, T. J.; White, P. S. J. Am.
Chem. Soc. 1995, 117, 823.
(45) Díez-Gonzꢀalez, S.; Marion, N.; Nolan, S. P. Chem. Rev. 2009,
109, 3612.
(46) Kirklin, D. R.; Domalski, E. S. J. Chem. Thermodyn. 1988, 20,
754.
(47) Bedford, A. F.; Mortimer, C. T. J. Chem. Soc. 1960, 1622.
(48) Chernick, C. L.; Skinner, H. A. J. Chem. Soc. 1956, 1401.
(49) The PꢀO BDEs for Me3PO from ref 19 of 139 kcal molꢀ1 and
for Bu3PO from ref 10 of 137.2 kcal molꢀ1 are in agreement with the
values of 138.5 and 137.6 kcal molꢀ1 respectively for Me3PO and Cy3PO
in this work.
(21) (a) Johnson, M. J. A.; Odom, A. L.; Cummins, C. C. Chem.
Commun. 1997, 1523–1524. (b) Piro, N. A.; Cummins, C. C. J. Am.
Chem. Soc. 2009, 131, 8764–8765.
(22) Beck, W.; Keubler, M.; Leidl, E.; Nagel, U.; Schaal, M.; Cenini,
S.; Buttero, P. D.; Licandro, E.; Maiorana, S.; Villa, A. C. Chem. Commun.
1981, 446.
(23) Genco, N. A.; Partis, R. A.; Alper, H. J. Org. Chem. 1973, 38, 4365.
(24) Feuer, H., Ed.; Nitrile Oxides, Nitrones, and Nitronates in Organic
Synthesis; John Wiley & Sons: Hoboken, NJ, 2008.
(25) (a) Bokach, N. A.; Khripoun, A. V.; Kukushkin, V. Y.; Haukka,
M.; Pombeiro, A. J. L. Inorg. Chem. 2003, 42, 896. (b) Kuznetsov, M. L.;
Kukushkin, V. Y.; Dement’ev, A. I.; Pombeiro, A. J. L. J. Phys. Chem. A
2003, 107, 6108.
(26) (a) McDonough, J. E.; Mendiratta, A.; Curley, J. J.; Fortman,
G. C.; Fantasia, S.; Cummins, C. C.; Rybak-Akimova, E. V.; Nolan, S. P.;
Hoff, C. D. Inorg. Chem. 2008, 47, 2133. (b) Huang, J.; Schanz, H. J.;
Stevens, E. D.; Nolan, S. P.; Capps, K. B.; Bauer, A.; Hoff, C. D. Inorg.
Chem. 2000, 39, 1042. (c) Capps, K. B.; Wixmerten, B.; Bauer, A.; Hoff,
C. D. Inorg. Chem. 1998, 37, 2861.
(50) Data taken from the NIST Webbook. http://webbook.nist.
gov/chemistry
(51) Stephens, F. H.; Johnson, M. J. A.; Diaconescu, P. L.; Cummins,
C. C.; Kryatova, O. P.; Rybak-Akimova, E. V.; McDonough, J. E.; Hoff,
C. D. J. Am. Chem. Soc. 2005, 127, 15191.
(52) Johnson, A. R.; Baraldo, L. M.; Cherry, J. P. F.; Tsai, Y. C.;
Cummins, C. C.; Kryatov, S. V.; Rybak Akimova, V.; Capps, K. B.; Hoff,
C. D.; Haar, C. M.; Nolan, S. P. J. Am. Chem. Soc. 2001, 123, 7271.
(53) Caramori, G. F.; Frenking, G. Theor. Chem. Acc. 2008, 120, 351.
(54) (a) Frenking, G.; Wichmann, K.; Fr€ohlich, N.; Grobe, J.; Golla,
W.; Le Van, D.; Krebs, B.; L€age, M. Organometallics 2002, 21, 2921.
(b) Fischer, R. A.; Schulte, M. M.; Weiß, J.; Zsolnai, L.; Jacobi, A.; Huttner,
G.; Frenking, G.; Boehme, C.; Vyboishckikov, S. F. J. Am. Chem. Soc.
1998, 120, 1237.
(55) Several recent articles have compared carbenes to metals, see:
Martin, D.; Soleihavoup, M.; Bertrand, G. Chem. Sci. 2011, 2, 389.
(56) Quadrelli, P.; Mella, M.; Invernizzi, A. G.; Caramella, P.
Tetrahedron 1999, 55, 10497.
(27) (a) Delaude, L.; Demonceau, A.; Wouters, J. Eur. J. Inorg. Chem.
2009, 1882–1891. (b) Duong, H. A.; Tekavec, T. N.; Arif, A. M.; Louie,
J. Chem. Commun. 2004, 112–113.
(28) Rybak-Akimova, E. V.; Palluccio, T.; Silvia, J. S.; Cummins,
C. C.; Temprado, M.; Majumdar, S.; Cai, X.; Hoff, C. D., work in progress.
(29) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
’ NOTE ADDED AFTER ASAP PUBLICATION
This paper was published on the Web on August 29, 2011,
with a minor error in equation 4. The corrected version was
reposted on September 1, 2011.
(30) For a description see supporting information files in: Barybin,
M. V.; Diaconescu, P. L.; Cummins, C. C. Inorg. Chem. 2001, 40, 2892
which presents a modified preparation from the original. Grundmann,
C.; Dean, J. M. J. Org. Chem. 1965, 30, 2809.
(31) Apex2 Version 2.2ꢀ0 and SAINT+ Version 7.46A; Bruker
Analytical X-ray System, Inc.: Madison, WI, 2007.
(32) (a) Sheldrick, G. M. SHELXTL, Version 6.1; Bruker Analytical
X-ray Systems, Inc.: Madison, WI, 2000. (b) Sheldrick, G. M. Acta
Crystallogr. 2008, A64, 112–122.
9630
dx.doi.org/10.1021/ic2013599 |Inorg. Chem. 2011, 50, 9620–9630