Role of axial base coordination in isonitrile binding and chalcogen atom transfer to vanadium(III) complexes

Article abstract of DOI:10.1021/ic5017005

The enthalpy of oxygen atom transfer (OAT) to V[(Me₃SiNCH₂CH₂)₃N], 1, forming OV[(Me₃SiNCH₂CH₂)₃N], 1-O, and the enthalpies of sulfur atom transfer (SAT) to 1 and V(N[t-Bu]Ar)₃, 2 (Ar = 3,5-C₆H₃Me₂), forming the corresponding sulfides SV[(Me₃SiNCH₂CH₂)₃N], 1-S, and SV(N[t-Bu]Ar)₃, 2-S, have been measured by solution calorimetry in toluene solution using dbabhNO (dbabhNO = 7-nitroso-2,3:5,6-dibenzo-7-azabicyclo[2.2.1]hepta-2,5- diene) and Ph₃SbS as chalcogen atom transfer reagents. The V-O BDE in 1-O is 6.3 ± 3.2 kcal·mol^-1 lower than the previously reported value for 2-O and the V-S BDE in 1-S is 3.3 ± 3.1 kcal·mol^-1 lower than that in 2-S. These differences are attributed primarily to a weakening of the V-N_axial bond present in complexes of 1 upon oxidation. The rate of reaction of 1 with dbabhNO has been studied by low temperature stopped-flow kinetics. Rate constants for OAT are over 20 times greater than those reported for 2. Adamantyl isonitrile (AdNC) binds rapidly and quantitatively to both 1 and 2 forming high spin adducts of V(III). The enthalpies of ligand addition to 1 and 2 in toluene solution are -19.9 ± 0.6 and -17.1 ± 0.7 kcal·mol^-1, respectively. The more exothermic ligand addition to 1 as compared to 2 is opposite to what was observed for OAT and SAT. This is attributed to less weakening of the V-N_axial bond in ligand binding as opposed to chalcogen atom transfer and is in keeping with structural data and computations. The structures of 1, 1-O, 1-S, 1-CNAd, and 2-CNAd have been determined by X-ray crystallography and are reported. (Graph Presented)

Full text of DOI:10.1021/ic5017005

Article
pubs.acs.org/IC
Role of Axial Base Coordination in Isonitrile Binding and Chalcogen
Atom Transfer to Vanadium(III) Complexes
Subhojit Majumdar,† Julia M. Stauber,‡ Taryn D. Palluccio,§ Xiaochen Cai,† Alexandra Velian,‡
Elena V. Rybak-Akimova,*,§ Manuel Temprado,*,∥ Burjor Captain,*,† Christopher C. Cummins,*,‡
and Carl D. Hoff*,†
†Department of Chemistry, University of Miami, 1301 Memorial Drive, Coral Gables Florida 33021, United States
‡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
∥Department of Analytical Chemistry, Physical Chemistry and Chemical Engineering, Universidad de Alcala,
́
Ctra. Madrid-Barcelona
Km. 33.600, Madrid, 28871, Spain
S
* Supporting Information
ABSTRACT: The enthalpy of oxygen atom transfer (OAT) to
V[(Me3SiNCH2CH2)3N], 1, forming OV[(Me3SiNCH2CH2)3N], 1−
O, and the enthalpies of sulfur atom transfer (SAT) to 1 and V(N[t-
Bu]Ar)3, 2 (Ar = 3,5-C6H3Me2), forming the corresponding sulfides
SV[(Me3SiNCH2CH2)3N], 1−S, and SV(N[t-Bu]Ar)3, 2−S, have been
measured by solution calorimetry in toluene solution using dbabhNO
(dbabhNO = 7-nitroso-2,3:5,6-dibenzo-7-azabicyclo[2.2.1]hepta-2,5-
diene) and Ph3SbS as chalcogen atom transfer reagents. The V−O
BDE in 1−O is 6.3 3.2 kcal·mol−1 lower than the previously reported
value for 2−O and the V−S BDE in 1−S is 3.3 3.1 kcal·mol−1 lower
than that in 2−S. These differences are attributed primarily to a
weakening of the V−Naxial bond present in complexes of 1 upon
oxidation. The rate of reaction of 1 with dbabhNO has been studied by
low temperature stopped-flow kinetics. Rate constants for OAT are over 20 times greater than those reported for 2. Adamantyl
isonitrile (AdNC) binds rapidly and quantitatively to both 1 and 2 forming high spin adducts of V(III). The enthalpies of ligand
addition to 1 and 2 in toluene solution are −19.9 0.6 and −17.1 0.7 kcal·mol−1, respectively. The more exothermic ligand
addition to 1 as compared to 2 is opposite to what was observed for OAT and SAT. This is attributed to less weakening of the
V−Naxial bond in ligand binding as opposed to chalcogen atom transfer and is in keeping with structural data and computations.
The structures of 1, 1−O, 1−S, 1−CNAd, and 2−CNAd have been determined by X-ray crystallography and are reported.
addition, rather than being based on an equatorial set of four
ligands as is the case with porphyrin systems, the molecular
architecture supporting a reactive metal center can be
developed based on a pure X3 ligand set,5 and also by tripodal
ligands that are formed by linking three X groups to a threefold
symmetric anchor. For example, Meyer and co-workers6 have
uncovered novel CO2 chemistry of uranium by tethering three
aryloxide arms to a triazacyclononane anchor. Stack and Holm7
reported subsite differentiation in [4Fe-4S]2+ clusters enabled
by a ligand in which three thiolate moieties were linked to the
1, 3, and 5 positions of a hexa-substituted benzene core.
Deriving inspiration from the phosphatrane chemistry of
Verkade,8 the relatively bulky silyl-substituted triamidoamine
ligand based upon tris(2-aminoethyl)amine (tren) was
introduced to transition-metal chemistry with a report on
INTRODUCTION
■
Reversible coordination of an axial base in an octahedral metal
complex has been widely studied since it plays an important
role in the chemistry of cobalamin1 and hemoglobin.2 Recent
work in the cobalamin area has resulted in isolation and
structural characterization of the first “base off” adduct of
vitamin B-12.3 It is believed that the “base off” form may be the
active intermediate during cobalt-alkyl bond formation. The
binding of oxygen to hemoglobin and its analogs and models is
also known to depend strongly on the nature of the axially
coordinated base. Detailed theoretical studies show a complex
range of behavior in spin state, structure, and reaction
energetics, all dependent upon the nature of the specific base
coordinated axially to iron.4 In spite of considerable work, the
role of axial base coordination in these complex bioinorganic
systems is still not fully resolved.
Coordination of an axial base can play an important role in
Received: July 24, 2014
simpler and easier to study inorganic systems as well. In
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
vanadium and titanium complexes9 and another report on
trigonal monopyramidal complexes of titanium through iron.10
This set the stage for a rich early transition-metal chemistry of
such ligands as summarized by Schrock,11 including applica-
tions to N2 reduction processes.12
rapid reassociation.16 The final case C involves a multilinked
axial base, which is present in the tren type complexes discussed
above. In this case it would appear at first sight that the base
could not dissociate since it is apparently locked in position by
the chelate system. However, lengthening of this bond can
occur with subsequent bond weakening which may approx-
imate a “base off” position.
Differing reactivity patterns can occur in the presence or
absence of an axial base ligand, and this occurs most notably for
13
This work began as an extension of our recent study17 of
oxygen atom transfer (OAT) for a series of nitrogen oxides to
the complex V(N[t-Bu]Ar)3, 2, and comparison of its reaction
energetics to that of the corresponding silylated tren complex
V[(Me3SiNCH2CH2)3N], 1.18 Complex 2 does not contain a
vacant coordination site but rather one that is blocked by close
aryl contact of one of the arene ligands through the single
unoccupied d orbital of V. This interaction is computed to be
on the order of 4 kcal·mol−1.17 The computed structures for
both species with and without this interaction have been
reported previously.17 While a crystal structure of 2 is not
available, the V−arene interaction has been demonstrated
crystallographically for a related complex.19 A related Ti
complex has been shown structurally to contain two such
interactions at the metal center.20 In the nomenclature of
Scheme 1, complex 1 corresponds to a “trapped base” and
complex 2 to a “tethered base” complex which can be viewed as
a hemilabile ligand.
the Mo(III) triamido complexes Mo(N[t-Bu]Ar)3 and the
fragment Mo(tren*) (tren* = triamidoamine ligand substituted
with a range of bulky groups).11,12 Complex Mo(N[t-Bu]Ar)3
and the Mo(tren*) moiety both contain three trigonal anionic
amido donor atoms, but the tren* complexes are augmented by
a trilinked amine donor atom coordinated as an axial base.
Whereas Mo(N[t-Bu]Ar)3 does not bind N2 to form
monometallic isolable complexes, mononuclear (N2)Mo(tren*)
complexes have been isolated and characterized, and for
specially constructed systems, stepwise reduction to ammonia
via coupled proton transfer/electron transfer has been
proven.12 However, Mo(N[t-Bu]Ar)3 forms a bridging μ-N2
dinuclear complex that splits N2 to form 2 equiv of NMo(N[t-
Bu]Ar)3.14 Interestingly, addition of an excess of selected N-
donor bases to Mo(N[t-Bu]Ar)3 has been shown to lead to
reversible binding of the N-donor and accelerated formation of
[μ-N2][Mo(N[t-Bu]Ar)3]2.15 This may occur through for-
mation of an intermediate base adduct (Nbase)Mo(N[t-Bu]Ar)3
that has a greater affinity for N2 than does Mo(N[t-Bu]Ar)3
itself.
The presence of a similar vanadium/vanadium oxo pair (with
reference to V(N[t-Bu]Ar)3/OV(N[t-Bu]Ar)3 (2/2−O)),
namely, V[(Me3SiNCH2CH2)3N]/OV[(Me3SiNCH2CH2)3N]
(1/1−O), allows direct comparison of the role of a pendant
axial base for two comparable systems. In addition to OAT,
sulfur atom transfer (SAT) and ligand binding of AdNC (Ad =
1-adamantyl) to 1 and 2 have also been investigated. A
combination of structural, thermodynamic, kinetic, and
computational studies is reported aimed at determining
whether chalcogen atom transfer and ligand binding reactions
are more favorable for 1 versus 2.
Three possible modes of axial base binding are shown in
Scheme 1. The first case A involves reversible binding of a free
base ligand in which dissociation occurs with complete loss of
the coordinating base. The second case B involves a base linked
by a single tethered arm that is capable of dissociation but
remains in near proximity toi the metal center and is poised for
Scheme 1. Three Model Systems for Base (Shown as B)
Activation of Trisamido Early Transition Metal Complexes
a
2. EXPERIMENTAL SECTION
All manipulations were performed in a Vacuum Atmospheres glovebox
under an inert atmosphere of purified N2 or Ar. Solvents were
obtained anhydrous and oxygen-free by sparging with N2 and purifying
using a Glass Contours Solvent Purification System (SG Water) or by
distillation from Na/benzophenone under Ar. C6D6 was purchased
from Cambridge Isotope Laboratories, degassed, and stored over
activated 4 Å molecular sieves prior to use. All glassware was oven-
dried at temperatures greater than 170 °C prior to use. NMR spectra
were obtained on a Varian Mercury 300 MHz or Bruker Avance 400
MHz instrument equipped with Oxford Instruments superconducting
magnets. 1,18a SV[(Me3SiNCH2CH2)3N] (1−S),18b 2,21 and dbabh-
NO (dbabhNO = 7-nitroso-2,3:5,6-dibenzo-7-azabicyclo[2.2.1]hepta-
2,5-diene)22 were prepared and isolated according to literature
procedures, while all other reagents were acquired from commercial
sources and used as received or purified by standard procedures.
2.1. Synthesis of V[(Me3SiNCH2CH2)3N] (1). In the glovebox, a
100 mL round-bottom flask containing n-pentane (40 mL) was
charged with Li3[(Me3SiNCH2CH2)3N] (1.675 g, 4.404 mmol, 1
equiv) followed by VCl3(THF)3 (1.639 g, 4.404 mmol, 1 equiv). The
aqua-blue reaction mixture was allowed to stir for a total of 12 h, at
which point it was filtered through a pad of Celite to remove liberated
LiCl. The aqua-blue filtrate was then evaporated to dryness, affording a
waxy blue solid. The blue solids were dissolved in n-pentane (3 mL),
and the solution stored at −35 °C overnight to afford a crop of aqua
colored X-ray quality crystals. A second crop was obtained by
concentration of the remaining solution and storage at −35 °C (1.355
a
Pendant groups on the amido ligands are omitted for simplicity.
B
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
g in 2 crops, 3.303 mmol, 75%). NMR spectral data for 1 are available
in the Supporting Information.
2.6. Chalcogen Exchange Experiments. In the glovebox, 20 mg
(0.045 mmol, 1 equiv) of 1−S and 40 mg (0.066 mmol, 1.5 equiv) of 2
were loaded into a 10 mL screw cap vial. One milliliter of toluene-d8
was added to the solids. A 1H NMR spectrum was taken after about 25
min. The NMR data showed formation of 1 and also the formation of
2−S indicating that ‘S’ atom transfer had occurred. At the end of 2 h
the reaction had gone nearly to completion. Investigation of the
transfer in the reverse direction by reaction 1 and 2−S showed no
reaction over a similar time period. Additionally, O atom transfer
between 1−O and 2 was examined at room temperature and found to
not occur in either direction using procedures identical to those
described above.
2.2. Synthesis of OV[(Me3SiNCH2CH2)3N] (1−O). In the
glovebox, solid Me3NO (36 mg, 0.483 mmol, 1 equiv) was added to
a solution of 1 (198 mg, 0.483 mmol, 1 equiv) in n-pentane (3 mL).
The color of the reaction mixture gradually changed from aqua blue to
amber over the course of 20 min. After 1 h of stirring, all volatiles were
removed from the reaction mixture under reduced pressure. X-ray
quality crystals of 1−O were obtained by dissolution of the waxy
amber solid in n-pentane (1 mL) and storage at −35 °C overnight.
The remaining solution was concentrated and stored at −35 °C to
obtain a second crop of crystals (138 mg in 2 crops, 0.323 mmol,
67%). NMR spectral data for 1−O as well as 1−S are available in the
Supporting Information.
2.7. Calorimetric Measurements. The enthalpy of reaction of 2
and dbabhNO was measured in a Setaram C-80 calorimeter. In the
glovebox, a solution containing a large molar excess of 2 (∼0.50 g in 6
mL C6D6) was prepared and filtered into a clean vial. One milliliter of
this stock solution was loaded into an NMR tube. The remaining 5 mL
of solution was loaded into the Calvet calorimeter cell with dbabhNO
(∼0.0100 g) as limiting reagent. The calorimeter cell was sealed, taken
from the glovebox, and loaded into the calorimeter. Following
temperature equilibration, approximately 1.5 h, the reaction was
initiated and the calorimeter rotated to achieve mixing. Following the
return to the baseline the calorimeter cell was taken into the glovebox,
2.3. Synthesis of AdNCV[(Me3SiNCH2CH2)3N] (1−CNAd). In
the glovebox, a 10 mL screw cap vial was charged with solid 1 (100
mg, 0.244 mmol, 1 equiv) followed by AdNC (39.3 mg, 0.244 mmol, 1
equiv). Then, 1 mL distilled toluene was added to the solid and the
mixture was shaken for 5 min until all the solids were dissolved. The
solution turned dark green in color. Then, the solution was filtered and
the filtrate was poured into one side of a two-legged glass tube. The
tube was sealed under vacuum. The initially empty side of the tube was
kept in dry ice at −78 °C and the whole system was kept in a freezer at
−20 °C. Over a period of several days there was slow transfer of
toluene to the initially empty leg of the apparatus which was held at
dry ice temperature. During this time dark green X-ray-quality crystals
deposited on the wall of the tube. The tube containing the crystals was
inverted and the toluene transferred to the other leg. The dry ice trap
was replaced by a liquid nitrogen trap to freeze all the toluene and the
tube containing the crystals was sealed off under vacuum. Earlier
experiments had shown that even the isolated crystalline product 1−
CNAd decomposed rapidly when exposed to air. The sealed tube
containing the crystals was stored in the freezer until it was taken into
the glovebox for mounting. The mounted crystal in the diffractometer
holder was sealed in an argon filled box and taken to the diffractometer
where the apparatus was mounted under a flow of nitrogen at 100 K.
The reaction was also studied by NMR spectroscopy. In the
glovebox a 1/1 mixture of AdNC and 1 was made from 50 mg 1 and
20.0 mg AdNC which were weighed in a 10 mL screw cap vial and
dissolved in 1.0 mL C6D6. The solution was transferred into a NMR
tube in the glovebox fitted with a screw cap and Teflon-lined silicone
rubber septum. A NMR spectrum was taken after 10 min of the
reaction. NMR spectral data of the air sensitive paramagnetic complex
are available in the Supporting Information.
2.4. Synthesis of AdNCV(N[t-Bu]Ar)3 (2−AdNC). In the
glovebox, a 10 mL screw cap vial was charged with solid 2 (200 mg,
0.345 mmol, 1 equiv) followed by AdNC (56.0 mg, 0.347 mmol, 1.008
equiv). Then, 2 mL distilled toluene was added to the solid and the
mixture was shaken for 5 min until all the solids were dissolved. The
solution turned dark greenish brown in color. A procedure analogous
to that described above for 1−CNAd was used to isolate high quality
crystals for diffraction studies. The reaction was also studied by NMR
spectroscopy. In the glovebox, 50 mg 2 and 14.0 mg AdNC were
weighed in a 10 mL screw cap vial and dissolved in 1.0 mL C6D6. The
solution was transferred into an NMR tube in the glovebox fitted with
a screw cap and Teflon-lined silicone rubber septum. An NMR
spectrum was taken after 12 min of the reaction. NMR spectral data
for the air sensitive paramagnetic complex are available in the
Supporting Information.
1
opened, and 1 mL of the solution was loaded into an NMR tube. H
NMR spectra of both the stock solution and the calorimetry solution
were then acquired and the reaction was confirmed as quantitative.
Four measurements done in this way yielded ΔH = −132.5 0.9 kcal·
mol−1, which includes adjustment for the endothermic enthalpy of
solution of dbabhNO in toluene (+3.8
0.4 kcal·mol−1)17 for a
combined enthalpy in solution of ΔH = −136.3
Strictly analogous procedures were used to measure other OAT and
SAT reaction enthalpies.
1.3 kcal·mol−1.
2.8. Stopped Flow Kinetic Measurements. Toluene solutions
of complex 1 and dbabhNO were prepared in an MBraun glovebox
filled with ultra-high-purity argon (Airgas) and loaded into Hamilton
gastight syringes. Time-resolved visible spectra were acquired at low
temperatures (−80 to −62 °C) using a Hi-Tech Scientific Kinet Asyst
SF-61DX2 Cryo Stopped-Flow system (TgK Scientific Ltd.) equipped
with a quartz tungsten halogen light source, a J&M TIDAS diode array
detector and a Brandenburg 4479 Series PMT monochromator. The
instrument was equipped with stainless steel plumbing lined with
PEEK tubing and a 1.00 cm3 quartz mixing cell submerged in an
ethanol cooling bath. The temperature in the mixing cell was
maintained to 0.1 °C using a CAL 3200 automatic temperature
controller. Data acquisition was performed using TIDAS-DAQ and/or
Kinetic Studio software programs and mixing times were 2−3 ms. All
flow lines were washed extensively with Ar-saturated dry toluene prior
to loading reactant solutions. Experiments were performed in a single-
mixing mode of the instrument with a 1:1 (v/v) mixing ratio.
Reactions were studied under pseudo-first-order conditions using
excess dbabhNO. Rates of formation of 1−O were measured by fitting
kinetic traces at λ = 600 nm to a single exponential equation, rate =
−A × exp(−kobst) + C (Kinetic Studio). Observed rate constants
represent an average of three measurements which gave an acceptable
standard deviation (within 10%). All remaining parameters derived
from kinetic data are reported with their standard deviations. All
concentrations are reported after mixing in the stopped flow cell.
2.9. Computational Details. Electronic structure calculations
were carried out using the BP8623 density functional with the
Stuttgart-Dresden MDF1024 fully relativistic effective core potential
and basis for V including a set of additional f functions and the triple-ζ
quality basis set (6-311G(d,p)) for all other elements as implemented
in the Gaussian 09 suite of programs.25 Minimum energy structures
were optimized by computing analytical energy gradients. The
obtained stationary points were characterized by performing energy
second derivatives, confirming them as minima by the number of
negative eigenvalues of the Hessian matrix of the energy. Since the
computed Hessians of a couple of optimized structures were found to
possess small negative eigenvalues, the zero-point vibrational effects
were not included in the final energies. The V−O bond dissociation
2.5. FTIR Studies of Reaction of 1 and 2 with AdNC. In the
glovebox solid AdNC (∼15 mg) and excess 1 (∼150 mg) were
weighed into a screw cap vial. Two milliliters of toluene was added to
the vial, which was closed and shaken until all solids had dissolved. An
aliquot of the dark green solution was transferred to a CaF2 FTIR
solution cell, which was sealed and taken from the glovebox to the
spectrometer. A broad band at 2133 cm−1 was assigned to ν-NC for
1−CNAd. No sign of an absorbance due to free AdNC at 2129 cm−1
was detected. In a similar way a broad band at 2151 cm−1 was assigned
to ν-NC for 2−CNAd.
C
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
energies were derived by determining the enthalpy of reaction with
molecular oxygen for the LnV/LnVO couples as previously described.17
Alternatively, the V−S bond dissociation energies were obtained
directly through the homolytic V−S dissociation reaction.
Relaxed potential energy surface scans along the V−Naxial bond
length were performed for the open-shell model complex V-
[(HNCH2CH2)3N] (in which the trimethylsilyl substituent has been
replaced by H), 1m, and for the corresponding closed-shell oxo
derivative 1m−O using the 6-311++G(3df,2p) basis set for all
nonvanadium atoms.
3. RESULTS
3.1. Structural Analysis of 1, 1−O, and 1−S. Although
the synthesis of 1 was reported earlier by Schrock,18 its
characterization did not include structural determination by
single crystal X-ray diffraction. A slightly modified preparation
and crystallization from n-pentane at −35 °C yielded high
quality crystals of 1, allowing for accurate determination of its
structure as shown in Figure 1. The structure of 1 closely
resembles that of V[Me2(tBu)SiNCH2CH2)3N] reported
previously.10
Figure 2. Solid-state molecular structure of 1−O with ellipsoids at the
50% probability level rendered using PLATON.26a Hydrogen atoms
are omitted for clarity. Selected interatomic distances (Å) and angles
(deg) (two complexes were found in the solution denoted here with
distances and angles for V1 and V2): V1−O1 1.610(2), V1−N2
1.908(2), V1−N3 1.903(2), V1−N1 1.902(2), V1−N11 2.288(2),
N2−V1−N3 116.72(9), N2−V1−N1 115.19(9), N1−V1−N3
118.33(9), O1−V1−N11 177.72(8), V2−O2 1.614(2), V2−N5
1.901(2), V2−N4 1.916(2), V2−N6 1.894(2),V2−N22 2.321(2),
N4−V2−N5 117.91(8), N5−V2−N6 115.92(9), N6−V2−N4
115.52(9), O2−V2−N22 177.39(8).
Figure 1. Solid-state molecular structure of 1 with ellipsoids at the
50% probability level rendered using PLATON.26a Hydrogen atoms
are omitted for clarity. Selected interatomic distances (Å) and angles
(deg): V1−N1 1.927(2), V1−N2 1.931(2), V1−N3 1.931(2), V1−N4
2.070(2), N2−V1−N3 119.28(8), N3−V1−N1 119.58(8), N2−V1−
N1 118.85(8), N4−V1−N1 85.31(7).
Figure 3. Solid-state molecular structure of 1−S with ellipsoids at the
50% probability level rendered using PLATON.26a Hydrogen atoms
are omitted for clarity. Selected interatomic distances (Å) and angles
(deg): V1−S1 2.071(1), V1−N3 1.890(2), V1−N1 1.892(2), V1−N2
1.895(2), V1−N11 2.378(2), N1−V1−N3 115.92(8), N2−V1−N1
114.82(8), N2−V1−N3 115.81(8), S1−V1−N11 179.34(5), S1−V1−
N3 102.26(6).
Preparation of 1−O and 1−S proceeded by reaction of 1
with Me3NO and S8, respectively. Crystallization from n-
pentane at −35 °C yielded diffraction quality crystals of 1−O
and 1−S, whose structures are shown in Figures 2 and 3.
The V−O interatomic distance of 1.612(4) Å in 1−O27 is
slightly longer than the 1.59 Å value estimated for the VO
triple bond based on covalent radii.28 The V−S interatomic
distance of 2.071(1) Å in 1−S has a value consistent with that
predicted28 for a VS double bond of 2.06 Å. Both the V−O
and the V−S bond distances are slightly longer in 1−E than in
their 2−E (E = O, 1.590(3) Å and S 2.053(2) Å) analogues.
3.2. Synthesis and Structures of AdNC Adducts of 1
and 2. Reaction of AdNC with both 1 and 2 resulted in rapid
formation of green or green-brown adducts 1−CNAd and 2−
CNAd. NMR data (see Supporting Information) showed both
to be paramagnetic high spin complexes and thus best
formulated as d2 S = 1 vanadium(III) complexes. FTIR data
in toluene solution showed a shift to higher wavenumber than
free AdNC for both adducts indicating no significant back-
donation from metal to ligand: ν(AdNC) = 2129 cm−1; ν(1−
CNAd) = 2133 cm−1; ν(2−CNAd) = 2151 cm−1. The band
due to 2−CNAd is broad and this may be due to different
configurational isomers being present in solution. The crystal
structure of 2−CNAd (see Figure 5) reveals that in the solid
state the anilide ligands adopts an arrangement where one of
the three aryl rings points downward away from the bound
isonitrile and two point upward toward it. This type of
configuration has been found for d2 molybdenum ketimide
complexes.29 It is plausible that in solution different rotational
isomers exist in addition to that shown in the solid state.
The structures of 1−CNAd and 2−CNAd are shown in
Figures 4 and 5 below (full structural data are available in
Supporting Information).
For both complexes, the V−C−N angle is nearly linear and
the CN distances are similar (1.155 Å for 1−CNAd and
1.159 Å for 2−CNAd). The CN distance in free AdNC has
not been reported, but that in MeNC has30 and is 1.167 Å. The
D
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
3.3. Comparison of Structural Data. Selected measured
and calculated structural parameters for 1, 1−O, 1−S, 1−
CNAd, and 2−CNAd are collected in Table 1. For comparative
purposes, reported data for 2−O17 and SV(N[t-Bu]Ar)3 (2−
S)31 are also shown. Upon oxidation of the vanadium center
from V(+3) in 1 and 2 to V(+5) in 1−E and 2−E (E = O,S),
the experimental and calculated interatomic distances between
the vanadium atom and the three equatorial amido nitrogens
shorten as expected. In contrast, a weakening of the V−Naxial
bond occurs in 1 upon oxidation of the vanadium center, as
indicated by the lengthening of the V−Naxial bond distances
which go from 2.07 Å in 1 to 2.30 (1−O) and 2.39 Å (1−S).
During this process, pyramidalization occurs at V as evidenced
by a decrease in the sum of the Neq−V−Neq angles for the three
amides. For a planar structure, Σ(Neq−V−Neq) would be 360°
and this quantity for 1 (357.7°) indicates a nearly planar
structure. The Σ(Neq−V−Neq) value drops to 346.5° for 1−S
and is even lower for 2−S (331.6°). There is a good correlation
between the degree of pyramidalization quantified by Σ(Neq−
V−Neq) and the lengthening of the V−Naxial distance observed
in all derivatives of 1 that have been structurally characterized
by X-ray diffraction as shown in Figure S4 in the Supporting
Information.
Figure 4. Solid-state molecular structure of 1−CNAd with thermal
ellipsoids drawn at the 50% probability level rendered using
ORTEP.26b Selected interatomic distances (Å) and angles (deg):
V1−N3 1.951(4), V1−N1 1.955(4), V1−N2 1.967(4), V1−C16
2.114(5), V1−N4 2.162(4), N5−C16 1.155(7), N3−V1−N1
118.47(19), N3−V1−N2 118.99(19), N1−V1−N2 118.25(19), N3−
V1−C16 96.40(19), N1−V1−C16 97.13(19), N2−V1−C16 97.2(2),
N3−V1−N4 83.36(17), N1−V1−N4 83.20(18), C16−V1−N4
179.7(2), C43−N10−C44 179.3(7).
Alternatively, AdNC binding causes a less pronounced
pyramidalization in the V(III) adducts, and consequently, the
V−Nax bond lengthening in 1-CNAd is less marked than in 1−
O and 1−S.
In addition, the three N atoms of the amido donors define a
plane. Changes in the V−Naxial bond length in 1 comprise an
upward motion of the V atom and a downward motion of the
N
axial atom away from this plane. For 1 itself, the V atom is 0.17
Å above the plane and the Naxial is 1.90 Å below it. The majority
of the bond lengthening that occurs upon oxidation is due to
further elevation of the V atom above the plane (approximately
70%) with a minor amount (about 30%) of motion of the Naxial
bond away from this plane due to straightening of the linking
ethylene bridges. This correlation is fairly uniform for the
complexes as shown in Figure S5 in the Supporting
Information.
3.4. Solution Calorimetric Measurements. In order to
allow direct comparison of the bond to chalcogenides for
complexes 1 and 2, the enthalpy of reaction 1 was measured in
Figure 5. Solid-state molecular structure of 2−CNAd with thermal
ellipsoids drawn at the 50% probability level rendered using
ORTEP.26b Selected interatomic distances (Å) and angles (deg):
V1−N1 1.933(4), V1−N2 1.944(3), V1−N3 1.928(3), V1−C41
2.068(5), N4−C41 1.159(5), N3−V1−N1 113.94(15), N3−V1−N2
120.21(14), N1−V1−N2 121.90(14), N3−V1−C41 98.33(16), N1−
V1−C41 98.82(16), N2−V1−C41 92.96(16), N4−C41−V1 177.0(4),
C41−N4−C42 175.7(4).
toluene solution by reaction calorimetry. Reaction 1 is rapid
and quantitative as determined by NMR spectroscopy. The
value ΔHRxn1 = −136.3 1.3 kcal·mol−1 compares directly to
that previously reported17 for dbabhNO and 2 of ΔH = −142.6
1.9 kcal·mol−1 and gives a direct difference of 6.3 3.2 kcal·
mol−1 for the V−O BDEs. Based on the previous value of 154
kcal·mol−1 for the V−O BDE in 2−O,17 a value of 148 kcal·
mol−1 is derived for the V−O BDE in 1−O.
slightly shorter CN distances in the adducts, coupled with
the FTIR data, suggest that CN bond strengthening occurs
upon complexation and that little donation of electron density
from the dxz and dyz orbitals on V to vacant π* orbitals on
AdNC is present. Furthermore, the V−C bond length is 0.046
Å longer in 1−CNAd than in 2−CNAd analogously to what
was observed for the corresponding oxo and sulfide derivatives.
Likewise, the enthalpies of SAT to complexes 1 and 2 using
Ph3SbS as SAT reagent were measured as shown in eqs 2 and 3.
The experimental data, ΔHRxn2 = −31.0 1.8 and ΔHRxn3
=
−34.3 2.1 kcal·mol−1, give a direct measure of the difference
in V−S BDE of 3.3 3.1 kcal·mol−1 with again a lower V−S
BDE in 1−S than in 2−S. As an additional check, the chalcogen
exchange reactions shown in eq 4 were investigated. In spite of
E
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
a
Table 1. Comparison of Experimental and Calculated Distances (Å) and Angles (deg) for Selected Complexes
V−E
V−Nax
(V−Neq)avg
Σ(Neq−V−Neq)
V−N3 plane
Nax−N3 plane
meas
calc
--
meas
calc
meas
calc
meas
calc
meas
meas
1
--
2.07
2.16
2.30
2.39
--
2.11
2.27
2.45
2.57
--
1.93
1.96
1.90
1.90
--
1.95
1.96
1.91
1.89
1.93
1.95
1.90
1.90
358
356
350
347
--
358
353
346
341
356
353
337
333
0.17
0.24
0.35
0.41
--
1.90
1.92
1.94
1.98
--
1-CNAd
2.11
1.61
2.08
--
2.04
1.61
2.08
--
b
1-O
1-S
2
2-CNAd
2.07
1.59
2.05
2.00
1.60
2.06
--
--
1.94
1.88
1.88
356
333
332
0.23
0.58
0.59
--
c
2-O
--
--
--
d
2-S
--
--
--
a
For simplicity, and to show significant changes only, distances and angles have been rounded off so that the estimated standard deviation < (1) for
b
all data. The full data set, including ESD values, is shown in figure captions and is also available in Supporting Information. For 1−O two molecules
occur in the unit cell. Within experimental error there are no differences in the bond lengths except for the V−Naxial bond lengths which are 2.288(2)
c
d
and 2.321(2) Å. The average value in Table 1 of 2.30 therefore has an uncertainty of 0.02 Å. Data taken from ref 17. Data taken from ref 31.
2 results in rapid and quantitative binding. Solution calorimetric
studies in toluene yielded ΔH = −19.9 0.6 and ΔH = −17.1
0.7 kcal·mol−1 for formation of 1−CNAd and 2−CNAd,
respectively. Thus, in contrast to OAT reactivity, ligand binding
studies showed that the V−C BDE in 1−CNAd is 2.8 1.3
kcal·mol−1 higher than in 2−CNAd. A summary of the
thermochemical BDEs derived in this work is summarized in
Table 2.
Table 2. Experimental V−E (E = O, S) BDE data (kcal·
a
mol−1)
1−X
2−X
V−O BDE (X = O)
V−S BDE (X = S)
148 3 [148.9]
98 3 [96.7]
154 317 [153.1]
101 317 [97.1]
17.1 0.7 [18.3]
V−C BDE (X = AdNC)
19.9 0.6 [18.7]
a
Values from quantum chemical calculations are in brackets.
3.5. Kinetics of OAT from dbabhNO to 1. The kinetics of
the OAT reaction between 1 and dbabhNO was studied using
the stopped-flow method with spectrophotometric registration.
Experiments were performed at low temperatures (−80, −71,
and −62 °C) under pseudo-first-order conditions using 1 (0.5
mM) and excess dbabhNO (1.5 to 9 mM). The time-resolved
spectral changes and kinetic traces at λ = 600 nm at −80 °C are
shown in Figure 6. Similar to the kinetics of OAT from
dbabhNO to 2,17 absorbance bands assignable to an
intermediate adduct were not detected in the time-resolved
spectra, and only clean conversion of 1 to 1−O was observed.
Due to the rapidity of the reaction, all quantitative measure-
ments were made from data collected in single wavelength
mode. At temperatures above −62 °C, the reaction rate became
the predicted thermodynamic favorability, the OAT reaction
was not observed to occur in either the forward or reverse
direction. This is in keeping with the fact that OAT between
OPMe3 and 2 also does not occur in spite of favorable
thermodynamic factors.17 In addition, kinetic factors may also
favor formation of S-bridged rather than O-bridged inter-
mediates for steric reasons. Indeed, SAT from 1−S to 2
producing 2−S and 1 was observed, but only in that direction.
This observation is in keeping with the generally greater
tendency to form bridged species of heavier chalcogenide and
pnictogen complexes.32 In view of the small thermochemical
difference predicting a higher V−S BDE in 2−S, it gave
confidence to the experimental data that reaction 4 did occur as
predicted in the forward direction but not in the reverse.
The lower V−O BDE in 1−O as compared to 2−O
prompted investigation to see if that relation carried over to
simple ligand binding. Addition of AdNC to solutions of 1 and
Figure 6. Left: Time resolved spectra obtained from the reaction
between 1 (0.5 mM) and dbabhNO (1.5 mM) at −80 °C acquired
over 1 s. Right: Kinetic traces at λ = 600 nm with varying dbabhNO
concentrations at −80 °C.
F
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 7. Left: Plots of kobs versus [dbabhNO] for 1−O formation over a temperature range of −80 to −62 °C. Right: Eyring plot with activation
parameters.
too fast to measure. The formation rates of 1−O were
measured at λ = 600 nm and all kinetic traces fit well to a single
exponential equation.
distances 0.15 and 0.18 Å longer than those observed in the
solid-state molecular structures of 1−O and 1−S, respectively.
However, as discussed below, in spite of these significant
discrepancies, the energetic differences between the computed
structures for 1−O and 1−S and the corresponding ones with a
fixed V−Naxial bond length at the observed crystallographic
values are only on the order of 1 kcal·mol−1.
Structural data discussed above showed lengthening and
weakening of the V−Naxial bond (1) or displacement of a weak
V-arene interaction (2) upon oxidation of the corresponding
V(III) complexes. The definition of BDEs often involves bond
rearrangements,33 which play a role in the V−O BDE in 1−O
and 2−O. Thus, it is of utility to analyze the effect that the
weakening of the V−Naxial bond has on reaction energetics in
the oxidation of 1. In this regard, a relaxed potential energy
surface scan along the V−Naxial bond length was performed for
the model complexes V[(HNCH2CH2)3N], 1m, and OV-
[(HNCH2CH2)3N], 1m−O, in which the trimethylsilyl
substituents in 1 and 1−O have been replaced by H. Computed
energies for 1m and 1m−O as a function of the V−Naxial bond
length are shown in Figure 8. The computed minimum energy
Observed rate constants depended linearly on [dbabhNO]
and slopes of the lines yielded second order rate constants for
formation of 1−O as shown in Figure 7. Although some
evidence for reversibility was observed owing to the increase in
final absorbance values as [dbabhNO] increased and in the
values of the y-intercepts as temperature rose, this was not
investigated further since the observed reaction rate is reaching
the limits of the stopped-flow instrument (thus the accuracy of
the rate constants decreases as the reaction rates increase with
increased temperature and concentrations of reagents). The
accuracy of the second order rate constants reported in Table 3
Table 3. Temperature Dependent Second Order (kapp) Rate
Constants and Activation Parameters for 1−O and 2−O
Formation Measured at λ = 600 nm (1−O) or 453 nm (2−
O)
a
kapp (×103 M−1·s−1
)
T (°C)
1−O
2−O17
−80
−71
−62
−53
15.6 0.5
26.0 0.8
39.7 1.5
not measured
1.01 0.03
1.82 0.09
3.84 0.30
6.3 0.3
too fast
3.8 0.1
−19
ΔH‡ (kcal·mol−1
)
ΔS‡ (cal·mol−1·K−1
)
1
−13
1
a
kapp = k1k2/(k−1 + k2) = k1/(1 + k−1/k2) (see ref 17 for further
details).
is sufficient to quantify the rate of OAT. An Eyring plot was
used to derive activation parameters (ΔH‡ = 3.8 kcal·mol−1 and
ΔS‡ = −19 cal·mol−1·K−1) for the OAT reaction between 1 and
dbabhNO and is shown in Figure 7. Second order rate
constants are provided in Table 3. While linear regression
analysis gives rather low values of standard deviations for the
values of ΔH‡ and ΔS‡ (see Figure 7), the relatively narrow
temperature range accessible for kinetic experiments limits
somewhat the accuracy of the derived activation parameters.
Nevertheless, negative sign of activation entropy and relatively
low activation enthalpy are established with confidence.
3.6. Computational Results. DFT calculations were
performed to additionally check the V−E BDE values derived
experimentally and to provide additional insight into the factors
influencing the differing V−O BDE values for 1−O and 2−O.
The calculated structural and energetic values are in good
agreement with experimentally found values as can be seen in
Tables 1 and 2. Nevertheless, calculations predict V−Naxial
Figure 8. Change in energy with V−Naxial bond length for 1m and
1m−O. At each fixed V−Naxial bond length, the minimum energy
structure was computed and the deviations from the minimum energy
of 1m (at 2.09 Å) and 1m−O (at 2.43 Å) are shown. For simplicity,
bridging ethylene groups have been omitted from the core structure.
configuration for 1m−O occurs with an elongated and
weakened V−Naxial bond at a distance of 2.43 Å, which
compares with the experimental distance of 2.30 Å. As the V−
Naxial bond length in 1m−O moves away from its minimum
value at 2.43 Å, there is less than a 1 kcal·mol−1 rise to the
observed crystallographic value. This small energy difference
may be overcome by packing forces.34 A steeper gradient
occurs as the potential energy surface rises to a value of 8.3
G
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
kcal·mol−1 higher at a fixed V−Naxial bond length of 2.09 Å
which is the computed distance for 1m being in good
agreement with the experimental value (2.070(2) Å). Likewise,
the energy dependence of 1m as a function of the V−Naxial
bond length increases steadily to a value of 8.8 kcal·mol−1 at the
distance of 2.43 Å, which is the computed V−Naxial bond length
for 1m−O.
As shown in Table 3, the OAT reaction between 1 and
dbabhNO is faster than the corresponding reaction with 2.
Since the first step in the OAT process is adduct formation, the
LUMOs of the unsaturated vanadium(III) complexes 1 and 2
were calculated to compare the potential sites for attack by an
incoming base and are shown in Figure 9. For complex 2, the
Starting with the open shell (S = 1) model V-
[(HNCH2CH2)3N], NRT analysis gives a total (spin α plus
spin β) natural atomic valence of 3.95 at vanadium, signifying
this atom’s participation in four electron-pair bonds; the three
amide nitrogen atoms are characterized by a valence of 3.01
(three electron-pair bonds), while the amine nitrogen enjoys a
natural valence of 3.98 and should be referred to accordingly as
tetravalent. The bond between vanadium and the amine
nitrogen is polar in nature, being composed 89.9% of a nitrogen
hybrid that is 21% s and 79% p in character (average of α and β
spin). The degree of the vanadium-amide bond polarity is
actually similar, but the nitrogen hybrid orbital involved in that
bond type is much richer in s character (>40%). The NPA
analysis is revealing in that the degree of negative charge on
each amide nitrogen (−1.1) is essentially double that carried by
the amine nitrogen (−0.5). The computed natural charge
densities on the V atom were as follows: 1 (+1.78); 1m−O
(+1.71); 1m−S (+1.25). Finally, we find that a single resonance
structure (two unpaired electrons on V and four V−N single
bonds) dominates the wave function (>50% in both α and β
spin manifolds). An interesting feature of the bonding in
V[(HNCH2CH2)3N], in which the vanadium and three amide
nitrogen atoms are coplanar, is the detection of three
corresponding N:-V-:N three-center, four-electron hyper-
bonds;41 the latter “ω bonds” disappear with the umbrella
flexing of the VN(amide)3 moiety that comes with addition of
an axial VE (E = O, S) multiple bond.
On going to the oxo and sulfido derivatives, we are interested
in how the bonding between vanadium and the axial amine
nitrogen is affected. First, we note that the picture obtained
from NPA analysis in terms of charge distribution has changed
little for the nitrogen atoms (ca. −0.9 for each amide nitrogen;
−0.5 for the amine nitrogen; essentially the same for oxo and
sulfido). At the same time, NRT analysis gives the central metal
atom a natural valence of 5.8 for both the oxo and sulfido
systems, such that vanadium forms six bonds in these systems:
the three metal-amide bonds, and the metal-chalcogenido triple
bond (bVO = 2.8, bVS = 2.7). In this analysis, the tertiary amine
lone pair previously (in the case of V[(HNCH2CH2)3N])
classified as a polar bond pair now falls below the threshold
cutoff to so qualify and is designated instead as an amine lone
pair (see Figure 10). This is in conjunction with an increase in
V−N(amine) interatomic distance of ca. 0.2−0.3 Å on going
from the unsaturated vanadium(III) complex to the vanadium-
(V) complexes bearing the axial metal−ligand triple bonds as
discussed previously. In addition, the oxo and sulfide derivatives
are not so dominated by a single Lewis structure; instead, they
are more delocalized with the major Lewis structure given ca.
25% weight and all others, including ones described with a
bond of the N: → (VE σ*) variety, accruing weight of less
than ca. 5%.
Given that in these derivatives bearing a metal-chalcogen
triple bond, the interaction between the tertiary amine and the
vanadium center is characterized not as a first-order bond but as
a second-order (hyperconjugative) interaction, it can be gauged
by the energy derived for it by second-order perturbation
theory analysis of the Fock matrix in the NBO basis
(ΔE(2)n→σ*):42 72 kcal·mol−1 for E = O, 80 kcal·mol−1 for E
= S; in both cases, this particular second-order delocalization
energy is the greatest of all in the molecule, the next closest in
energy (ca. 40 kcal·mol−1) representing amide lone pair
donation into metal-amide σ* orbitals.
Figure 9. Computed LUMOs for 1 (a) and 2 with (b) and without (c)
the weak interaction with a pendant arene.
V(III) center participates in a weak interaction with the arene
group of one amido ligand as discussed earlier. For comparison
purposes, the LUMO for the corresponding structure of 2
without this arene interaction was also computed. As can be
seen in Figure 9, the LUMO of 1 is more accessible for attack
by an OAT reagent such as dbabhNO.
3.7. NBO and NRT Analysis of a Model System. To gain
a zeroth-order description of the bonding in these vanadium
systems, the model open-shell complex V[(HNCH2CH2)3N]
and the corresponding closed-shell oxo and sulfido derivatives
have also been geometry-optimized at the MP2 level of
theory35 using a def2-TZVPP basis set36 as provided by the
ORCA program system.37 The structures so obtained (see
Supporting Information for the optimized coordinates) are in
good qualitative agreement with those determined experimen-
tally for the trimethylsilyl derivatives, and in the analysis that
follows, the focus is on gaining a qualitative understanding of
the bonding. All three model systems were subjected not only
to standard natural bond orbital38 and natural population
analysis39 (NBO, NPA) to afford natural orbitals and atomic
charges, but also to natural resonance theory (NRT)40 analysis
to afford values for the NRT valence (V) and bond order (b),
as well as the weightings of contributing resonance structures.
H
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 2. Thermochemical Cycle Comparing OAT to a
a
Complex and Its Coordinated Base (B) Adduct
a
ΔHi = ΔHbase binding to M, ΔHii = ΔHOAT to M‑B, ΔHiii = ΔHOAT to M
,
and ΔHiv = ΔHbase binding to M≡O
.
tethered or linked base adduct. Nevertheless, it is clearly
applicable here as well, but rather than direct measurement of a
BDE, the energy change associated with bond lengthening
rather than complete breaking must be calculated.
Figure 10. Natural localized molecular orbital (NLMO) correspond-
ing to the tertiary amine lone pair of SV[H(NCH2CH2)3N]; note the
minor degree of contribution from the VS σ* orbital.
In the absence of other factors, it would be expected that
OAT to a metal complex with an axially coordinated base
would be more exothermic than OAT to the same complex
without base coordination. If the only factor to be considered
was σ-donation from a “hard” N-donor base, it would be
expected that the enthalpy of base binding would be higher for
the higher oxidation state oxo complex than for the lower
4. DISCUSSION
The goal of this work is to assess the role that axial base ligation
plays in the kinetics and thermodynamics of OAT, SAT, and
dative ligand binding to V(III) complexes 1 and 2, and this will
be analyzed separately in the following sections.
oxidation state parent complex since ΔΔHOAT = ΔΔHbase binding
.
In fact, for simple gas phase ions it appears that base binding is
more exothermic to the higher oxidation state metal.44
Experimental data for 1 and 2, however, do not follow that
expected order. When allowances are made for breaking the
weak arene interaction present in 2 then OAT and SAT to 1 are
10 and 7 kcal·mol−1, respectively, more exothermic for 2. In
contrast, the enthalpies of AdNC binding are more exothermic
for 1 than for 2. This is attributed to differences in the extent of
weakening of the V−Naxial bond in 1−CNAd and 1−O or 1−S.
The variations in V−Naxial bond length observed here are in
keeping with a recent redetermination of the structure of
nitrosylcobalamin in which Co−Naxial bond length crystallo-
graphic data are summarized and show that this distance spans
a range from 1.92 to 2.35 Å depending upon the specific
complex.45
Computational results can be combined to assess the role of
the V−Naxial bond as shown in Scheme 3. Step (e) in Scheme 3
corresponds to the thermochemically measured BDE in which
1−O in its most stable configuration (V−Naxial = 2.43 Å)
dissociates yielding an O atom and 1, also in its most stable
configuration (V−Naxial = 2.09 Å). The computed and
measured data are in good agreement as discussed earlier.
Step (a) involves lengthening of the V−Naxial bond in 1−O as
it goes from a distance of 2.09 Å to 2.43 Å. This step is
computed to be exothermic by 8.3 kcal·mol−1 (Figure 8). Step
(c) involves shortening of the V−Naxial bond in 1 as it goes
from a distance of 2.43 Å to 2.09 Å. This step is computed to be
exothermic by 8.8 kcal·mol−1 (Figure 8). These data allow
comparison of three OAT reactions:
4.1. Energetics of OAT and SAT to 1 and 2. Direct use
of calorimetric data indicates that the OAT and SAT from 1 to
2 are exothermic by −6.3 3.2 and −3.3 3.1 kcal·mol−1,
respectively. The combined experimental errors are relatively
large because this conclusion involves taking the difference
between reactions that are individually exothermic by ∼100
kcal/mol. Competition experiments verified the relative order
of thermochemical stability for the SAT reaction but for kinetic
reasons did not occur in either direction for OAT. Computa-
tional studies are in relatively good agreement with the
experimental data and yield the same structural and energetic
trends.
The V−O and V−S BDEs are 6 and 3 kcal·mol−1 higher,
respectively, for 2−O and 2−S as compared to the 1−E
complexes as can be seen in Table 2. These thermochemical
differences would be enhanced by 4 kcal·mol−1 if the arene
interaction was not present in 2, increasing the V−O and V−S
BDEs to 10 and 7 kcal·mol−1, respectively. Although the V−O
and V−S bond are about 0.02 Å shorter in the oxo and sulfido
derivatives of 2 than in the corresponding 1−E analogues (see
Table 1), the thermochemical differences observed are higher
than might be expected for this small variation on the V−E
bonds.
With respect to OAT, thermochemical analysis where
complete base dissociation occurs is straightforward as shown
in Scheme 2.
There is a direct relationship between the relative enthalpies
of OAT and base binding, since ΔΔHOAT = {ΔHOAT
−
M‑B
ΔHOATM} = ΔΔHbind = {ΔHbind
− ΔHbindM}. Thus, the
M=O
difference in enthalpies of OAT is equal to the difference in
enthalpies of base binding. This kind of thermochemical
analysis is common for oxidative addition reactions where the
base is free to fully dissociate,43 but is less common for a
A. OAT to the “base on” complex 1 to produce 1−O also in
the “base on” configuration with V−Naxial = 2.09 Å, as
shown in step (d). This reaction is exothermic by 140.6
kcal·mol−1.
I
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
simple lattice energy considerations will favor a trigonal planar
arrangement since deviations from that geometry bring the
anionic X ligands into closer proximity without increasing the
anion−cation interactions; (2) from a π-bonding perspective,
the planar amides can be viewed as being sp2 hybridized with a
filled p orbital oriented to overlap with vacant orbitals on the V
center. During oxidation, the V(III) complex is converted to
V(V) which goes from a d2 to a d0 configuration. If 1−O and
2−O are viewed as being V(V), then π-donation from px and py
orbitals on the O atom will now displace the amido π-
interaction, with the result that the amido N atoms will now
move down (or the V atom moves up) and rotate so that
electron density that was previously involved in π-bonding to V
now generates a repulsive interaction with the bonding
electrons on the axial N atom. This atom will also move
away, but since it is linked via relatively rigid ethylene tethers,
the major motion to accommodate the V−Naxial bond
lengthening is a motion of the V atom out of the plane of
the amido ligands and toward the O atom π-donor.
In the presence of such complex rearrangements it is not
strictly correct to ascribe the effects in terms of simply
weakening and lengthening of the V−Naxial bondthere is
more going on than just that. Nevertheless, it must be kept in
mind that the “base on/base off” effect is ascribed to the
difference between 1 and 2. The pyramidalization effects that
occur upon oxidation are similar for both 1 and 2; in fact, 2
undergoes a more dramatic pyramidalization since it is less
restricted than 1. The presence of the axial base bond in 1
serves to inhibit pyramidalization and a balance is reached
between strengthening the V−O bond and weakening the axial
base bond. For purposes of simplicity these effects are covered
under the blanket term “weakening of the V−Naxial bond”.
4.3. Binding of AdNC to 1 and 2. Contrary to what was
observed for OAT and SAT, the V−C BDEs in the AdNC
adducts is about 3 kcal·mol−1 higher for 1−CNAd than for 2−
CNAd. In spite of that, the V−C bond length is 0.046 Å shorter
in 2−CNAd, which would seem to indicate a stronger V−C
bond in the latter complex. The reversal in relative
thermodynamic stability of the AdNC adducts of 1 and 2 as
compared to the chalcogenide species is in keeping with the
role of base binding. The paramagnetic adducts 1−CNAd and
2−CNAd show little spectroscopic or structural signs that
electron transfer has occurred. The ν(CN) stretch is shifted to
higher frequency for the adducts as compared to the free
isonitrile, and the AdNC bound ligand is linear. A similar
observation has been made by Nocera and co-workers for a
related complex.47 In addition, the CNAd bond is shorter in
the adducts than in the free isonitrile. This is all in keeping with
formulation as the complexes are high spin V(III) adducts with
little back-donation. Binding without significant electron
transfer in 1 and 2 is in marked contrast to the AdNC and t-
BuNC adducts of Mo(N[t-Bu]Ar)3, which we have reported
earlier.48 In those adducts the ν(CN) is shifted over 400 cm−1
to lower frequency and the isonitrile ligand adopts a bent rather
than linear geometry. In addition, the V−Naxial bond in 1−
CNAd is not significantly lengthened and only changes from
2.09 to 2.16 Å. As shown in Figure 8, this bond lengthening
corresponds to only ∼1 kcal·mol−1. In binding of AdNC to 2,
as shown in the structure in Figure 5, the V−arene interaction
is fully displaced. The fact that 2 loses this interaction (4 kcal·
mol−1) and 1 weakens its V−Naxial bond (1 kcal·mol−1) leads to
the prediction that if the V−C BDEs in 1−CNAd and 2−
CNAd are fundamentally the same, then 1 should be 3 kcal·
Scheme 3. Computed Cycle for 1−O BDE as a Function of
V−Naxial Bond Length
B. OAT to the “base off” complex 1 to produce 1−O also in
the “base off” configuration with V−Naxial = 2.43 Å, the
reverse of step (b). This reaction is exothermic by 157.7
kcal·mol−1.
C. OAT to the “base on” complex 1 (V−Naxial = 2.09 Å) to
produce 1−O in the “base off” configuration (V−Naxial
=
2.43 Å). This corresponds to the normal BDE defined for
the reaction and is shown as step (e), the reverse of
which is exothermic by 148.9 kcal·mol−1.
It is of interest to compare what these computed BDE data
for the V−O bond would be after removing the axial base
contribution to 1−O and the arene interaction in 2−O. As
illustrated by case “B″ above, the 1−O BDE would be 157.7
kcal·mol−1 if the bond to the Naxial base was not strengthened
upon deoxygenation. This is quite close to the value of 158
kcal·mol−1 for 2−O if the 4 kcal·mol−1 arene interaction was
not reformed during deoxygenation. This very close agreement
may be somewhat fortuitous since the 1−S and 2−S BDE
values differ by only 3.3 3.1 kcal·mol−1 in spite of a greater
lengthening of the V−Naxial bond in 1−S than in 1−O. The
same trend is present; however, it must be recognized that
experimental errors in both calorimetry and computational
results are on the order of ∼3 kcal·mol−1 and agreement
outside that range is not to be expected.
Finally, it should be noted that this type of bond lengthening
upon oxidation has been observed by Kol and co-workers for
conversion of V(III) to V(V)-oxo complexes chelated by an
amine tris(phenolate) ligand system.46 Bond lengthening was
found to occur to an even greater extent for the amine
tris(phenolate) compared to the triamidoamine system studied
here. The difference in the V−N bonds in 1 (2.070 Å) and 1−
O (2.288 Å) compares to that found for the amine
tris(phenolate) complexes (2.121 vs 2.4697 Å).
4.2. Qualitative Discussion of Bonding Effects in OAT
to 1. The conclusion that the presence of the base serves to
decrease rather than increase the V−O BDE in 1 disagrees with
the intuitive argument that the V−Naxial bond will be stronger
as the formal oxidation state of the V center increases. While
this would be expected for simple gas phase systems with few
ligands and a number of available orbitals, the more complex
bonding picture in 1−O shows behavior opposite of this
expectation. A simplified view is that for both 1 and 2 the
approximately trigonal planar structure of the amide ligands can
be attributed to two forces: (1) for an ionic MX3 complex,
J
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 4. Proposed Mechanism of Reaction of 1 and dbabhNO
mol−1 more exothermic in its binding of AdNC relative to 1.
The calorimetric result of 2.8 1.3 kcal·mol−1 is in agreement
with this estimate. It should be noted that both enthalpies of
binding are about 10 kcal·mol−1 less exothermic than that
reported earlier for AdNCMo(N[t-Bu]Ar)3.
loosely and a “base on/base off” equilibrium can be readily
established. The current work compares structural, thermody-
namic and kinetic aspects of reactivity of two V(III) trisamido
complexes, both of which have an additional site blocked by a
flexible ligand. Complex 2, while formally three-coordinate, in
fact has an additional site blocked by a weak interaction to a
pendant arene. Complex 1 differs in that the fourth
coordination site is to a tertiary amine that is part of the
[(Me3SiNCH2CH2)3N] ligand itself and is “locked” in position
and not able to fully dissociate. The reactions studied were
OAT, SAT, and binding of AdNC. The transformations
occurring during the reactions of these systems, which are
relatively simple compared to more complex bioinorganic
systems, are nevertheless complicated and the whole system
must be analyzed as “one piece of cloth”. Several conclusions
applicable to complexes 1 and 2 are as follows:
1. The “intrinsic” V−O BDEs in these systems are on the
order of 158 kcal·mol−1, and this is decreased by ∼4 kcal·mol−1
in 2−O by establishment of a weak V−arene interaction and by
∼9 kcal·mol−1 in 1−O by establishment of a stronger V−Naxial
interaction upon deoxygenation.
2. Upon OAT and SAT to 1, there is significant lengthening
and weakening of the V−Naxial bond and computational studies
indicate that this polar bond pair present in 1 is better
described instead as an amine lone pair with a minor degree of
N:→(VE σ*) second-order contribution in complexes 1−O
and 1−S.
3. In contrast to OAT and SAT, ligand addition of AdNC is
more exothermic to 1 than to 2, in spite of 2−CNAd having a
shorter V−C bond length than 1−CNAd. A plausible
explanation for this observation is the fact that binding of
AdNC to 1 only slightly weakens and lengthens the V−Naxial
bond in 1, while AdNC binding to 2 results in complete loss of
the V−arene interaction exhibited by 2.
4. In spite of being less thermodynamically favorable, OAT
occurs more rapidly to 1 than it does to 2 due to the more open
nature of the vacant axial site opposite the donor axial amine.
This is reflected primarily in the greater enthalpy of activation
for addition to 2 due to the necessity of breaking the weak V−
arene interaction.
5. For highly exothermic oxidative addition reactions such as
OAT and SAT, the 6−9 kcal·mol−1 thermochemical differences
associated with weakening of the V−Naxial bond does not
significantly change the reaction energetics. For reactions more
nearly thermoneutral, however, it could play a significant role
and alter the reaction profile.
4.4. Kinetic Differences between 1 and 2. Rapid OAT
from dbabhNO to 1 results in clean formation of 1−O, without
the generation of any observable intermediates (even at
temperatures as low as −80 °C). This behavior is similar to
the previously reported OAT from dbabhNO to 2,17 such as
that shown in Scheme 4 was proposed for a wide range of OAT
reagents (PhNO, PyO, N2O, and carbene adducts of N2O or
MesCNO). This general two step mechanism yielded four
different subcategories of reactivity based on the relative values
of the specific rate constants involved. For OAT from
dbabhNO to both 1 and 2, kinetic data are consistent with a
steady state formation of an adduct between metal and
substrate with no detectable buildup of intermediate in the
relatively rapid reaction. The experimental value for kobs in that
case corresponds to k1k2/(k−1 + k2). Since there are no
spectroscopic signatures detected for formation of the
intermediate adduct, it is not possible to go beyond the
conclusion that kobsd for 1 has a rate constant approximately 20
times larger than does 2 at these low temperatures, and that it
has an enthalpy of activation ∼3 kcal/mol lower. A plausible
explanation for this is the more open nature of the vacant site at
1 compared to 2 as illustrated in their computed LUMOs in
Figure 9.
A similar decrease in ligand binding rates of AdNC was
observed for another seemingly less sterically hindered complex
of molybdenum(III) (molybdaziridine hydride, Mo(H)-
(Me2CNAr)(N[i-Pr]Ar)2 compared to the three-coordinate
Mo(N[t-Bu]Ar)3).43 In this case also, the slower rate of
reaction was attributed to the necessity of generating a vacant
site at the metal center prior to ligand addition. The
preliminary interpretation of these kinetic results is that the
slower rate of reaction of 2 compared to 1 is most likely due to
a slower rate of ligand binding (k1) with a higher enthalpy of
activation due to the necessity of first cleaving the V−arene
interaction in 2 to generate an available vacant site at the metal
center as illustrated in the computed LUMOs in Figure 9.
5. CONCLUSIONS
The role of an axial base in influencing metal center reactivity
has been widely studied for vitamin B12 and analogous
complexes. In those systems, the pendant axial base is tethered
K
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(19) Ruppa, K. B. P.; Desmangles, N.; Gambarotta, S.; Yap, G.;
Rheingold, A. L. Inorg. Chem. 1997, 36, 1194.
(20) Wanandi, P. W.; Davis, W. M.; Cummins, C. C.; Russell, M. A.;
ASSOCIATED CONTENT
* Supporting Information
Additional experimental details including NMR spectroscopic
data, full crystallographic data, and additional computational
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
■
S
Wilcox, D. E. J. Am. Chem. Soc. 1995, 117, 2110.
(21) Brask, J. K.; Fickes, M. G.; Sangtrirutnugul, P.; Dura-
̀ ̀
Vila, V.;
Odom, A. L.; Cummins, C. C. Chem. Commun. 2001, 1676.
(22) Carpino, L. A.; Padykula, R. E.; Barr, D. E.; Hall, F. H.; Krause,
J. G.; Dufresne, R. F.; Thoman, C. J. J. Org. Chem. 1988, 53, 2565.
(23) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Perdew, J. P.
Phys. Rev. B 1986, 33, 8822.
AUTHOR INFORMATION
Corresponding Authors
■
(24) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987,
*E-mail: elena.rybak-akimova@tufts.edu.
*E-mail: manuel.temprado@uah.es.
*E-mail: captain@miami.edu.
*E-mail: ccummins@mit.edu.
*E-mail: c.hoff@miami.edu.
86, 866.
(25) Frisch, M. J. et al. Gaussian 09, revision C.01; Gaussian, Inc.:
Wallingford, CT, 2010.
(26) (a) Spek, A. L. J. Appl. Crystallogr. 2003, 7. (b) Farrugia, L. J. J.
Appl. Crystallogr. 2012, 45, 849.
(27) This reflects the average of 1.610(2) and 1.614(2) for the V−O
distances in the two molecules in the unit cell.
Notes
The authors declare no competing financial interest.
(28) Pyykko, P.; Atsumi, M. Chem.Eur. J. 2009, 15, 12770.
̈
(29) Curley, J. J.; Sceats, E. L.; Cummins, C. C. J. Am. Chem. Soc.
2006, 128, 14036.
ACKNOWLEDGMENTS
■
(30) The CN bond length of MeNC(g) is 1.167 Å as determined by
microwave absorption spectroscopy. Kessler, M.; Ring, H.;
Trambarudo, R.; Gordy, W. Phys. Rev. 1950, 79, 54.
(31) Brask, J. K.; Dura-
Chem.Commun. 2002, 902.
This material is based upon work supported by the National
Science Foundation under Grants CHE-1362118 (C.C.C.),
CHE-1412909, CHE-1229426, CHE-0750140 (E.R.A.), CHE-
0615743 (C.D.H), the Spanish Ministry of Economy and
Competitiveness (MINECO) under CTQ2012-36966 (M.T.),
and award of a GAANN fellowship (T.P.).
̀ ̀
Vila, V.; Diaconescu, P. L.; Cummins, C. C.
(32) Johnson, M. J. A.; Lee, P. M.; Odom, A. L.; Davis, W. M.;
Cummins, C. C. Angew. Chem., Int. Ed. 1997, 36, 87.
(33) For example cleaving the CC bond of ketene (H2CCO)
produces H2C and CO. The formal definition of the CC bond in
ketene is lower than expected since the CO bond is converted to a
CO concomitantly with CC bond cleavage. See: Blanksby, S. J.;
Ellison, G. B. Acc. Chem. Res. 2003, 36, 255.
(34) Thompson, H. P. G.; Day, G. M. Chem. Sci. 2014, 5, 3173.
(35) Møller, C.; Plesset, M. S. Phys. Rev. 1934, 46, 618.
(36) Weigend, F.; Ahlrichs, R. Phys. Chem. 2005, 7, 3297.
(37) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73.
(38) (a) Glendening, E. D.; Landis, C. R.; Weinhold, F. WIREs
Comput. Mol. Sci. 2012, 2, 1. (b) Glendening, E. D.; Badenhoop, J. K.;
Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Landis,
C. R.; Weinhold, F. NBO 6.0; http://nbo6.chem.wisc.edu/; Theoreti-
cal Chemistry Institute; University of Wisconsin, Madison, WI, 2013.
(c) Glendening, E. D.; Landis, C. R.; Weinhold, F. J. Comput. Chem.
2013, 34, 1429.
(39) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,
83, 735.
(40) (a) Glendening, E. D.; Weinhold, F. J. Comput. Chem. 1998, 19,
593. (b) Glendening, E. D.; Weinhold, F. J. Comput. Chem. 1998, 19,
610. (c) Glendening, E. D.; Badenhoop, J. K.; Weinhold, F. J. Comput.
Chem. 1998, 19, 628.
(41) Weinhold, F.; Landis, C. R. Discovering chemistry with natural
bond orbitals; Wiley: Hoboken, NJ, 2012.
(42) Weinhold, F.; Landis, C. R. Valency and bonding: a natural bond
orbital donor-acceptor perspective; Cambridge University Press: Cam-
bridge, U.K., 2005.
REFERENCES
■
(1) (a) Randaccio, L.; Brancatelli, G.; Demitri, N.; Dreos, R.; Hickey,
N.; Siega, P.; Geremia, S. Inorg. Chem. 2013, 52, 13392. (b) Hamza, M.
S. A.; Zou, X.; Banka, R.; Brown, K. L.; van Eldik, R. Dalton Trans.
2005, 782. (c) Rovira, C.; Biarnes, X. Inorg. Chem. 2004, 43, 6628.
(2) Kharitonov, V. G.; Sharma, V. S.; Magde, D.; Koesling, D.
Biochemistry 1997, 36, 6814.
(3) Moore, T. C.; Newmister, S. A.; Rayment, I.; Escalante-Semerena,
J. C. Biochemistry 2012, 51, 9647.
(4) Liao, M.; Huang, M.; Watts, J. D. J. Phys. Chem. A 2010, 114,
9554.
(5) (a) Cummins, C. C. Prog. Inorg. Chem. 1998, 47, 685. (b) Green,
M. L. H. J. Organomet. Chem. 1995, 500, 127.
(6) (a) Castro-Rodriguez, I.; Nakai, H.; Zakharov, L. N.; Rheingold,
A. L.; Meyer, K. Science 2004, 305, 1757. (b) Castro-Rodriguez, I.;
Meyer, K. J. Am. Chem. Soc. 2005, 127, 11242.
(7) Stack, T. D. P.; Holm, R. H. J. Am. Chem. Soc. 1988, 110, 2484.
(8) Verkade, J. G. Acc. Chem. Res. 1993, 26, 483.
(9) Cummins, C. C.; Schrock, R. R.; Davis, W. M. Organometallics
1992, 11, 1452.
(10) Cummins, C. C.; Lee, J.; Schrock, R. R.; Davis, W. M. Angew.
Chem., Int. Ed. 1992, 31, 1501.
(11) Schrock, R. R. Acc. Chem. Res. 1997, 30, 9.
(12) Schrock, R. R. Acc. Chem. Res. 2005, 38, 955.
(13) Laplaza, C. E.; Odom, A. L.; Davis, W. M.; Cummins, C. C.;
Protasiewicz, J. D. J. Am. Chem. Soc. 1995, 117, 4999.
(14) Cherry, J.-P. F.; Johnson, A. R.; Baraldo, L. M.; Tsai, Y.-C.;
Cummins, C. C.; Kryatov, S. V.; Rybak-Akimova, E. V.; Capps, K. B.;
Hoff, C. D.; Haarand, C. M.; Nolan, S. P. J. Am. Chem. Soc. 2001, 123,
7271.
(15) Tsai, Y. C.; Cummins, C. C. Inorg. Chim. Acta 2003, 345, 63.
(16) This can also be viewed as a “hemilabile” ligand: Slone, C. S.;
Weinberger, D. A.; Mirkin, C. Prog. Inorg. Chem. 1999, 48, 233.
(17) Palluccio, T. D.; Rybak-Akimova, E. V.; Cai, X.; Majumdar, S.;
Chui, M.; Temprado, M.; Silvia, J. S.; Cozzolino, A. F.; Tofan, D.;
Cummins, C. C.; Captain, B.; Hoff, C. D. J. Am. Chem. Soc. 2013, 135,
11357.
(43) Hoff, C. D. Prog. Inorg. Chem. 1992, 42, 503.
(44) Nose, H.; Rogers, M. T. ChemPlusChem 2013, 78, 1109.
(45) Hassanin, H. A.; El-Shahat, M. F.; DeBeer, S.; Smith, C. A.;
Brasch, N. E. Dalton Trans. 2010, 39, 10626.
(46) Groysman, S.; Goldberg, I.; Goldschmidt, Z.; Kol, M. Inorg.
Chem. 2005, 44, 5073.
(47) Groysman, S.; Villagran, D.; Freedman, D. E.; Nocera, D. G.
́
Chem. Commun. 2011, 47, 10242.
(48) Stephens, F. H.; Figueroa, J. S.; Cummins, C. C.; Kryatova, O.
P.; Kryatov, S. V.; Rybak-Akimova, E. V.; McDonough, J. E.; Hoff, C.
D. Organometallics 2004, 23, 3126.
(18) (a) Cummins, C. C.; Schrock, R. R. Inorg. Chem. 1994, 33, 395.
(b) Cummins, C. C.; Schrock, R. R.; Davis, W. M. Inorg. Chem. 1994,
33, 1448.
L
dx.doi.org/10.1021/ic5017005 | Inorg. Chem. XXXX, XXX, XXX−XXX