Experimental and computational studies of binding of dinitrogen, nitriles, azides, diazoalkanes, pyridine, and pyrazines to M(PR3) 2(CO)3 (M = Mo, W.; R = Me, Pr)

Article abstract of DOI:10.1021/ic900764e

The enthalpies of binding of a number of N-donor ligands to the complex Mo(PⁱPr₃)₂(CO)₃ in toluene have been determined by solution calorimetry and equilibrium measurements. The measured binding enthalpies span

Full text of DOI:10.1021/ic900764e

Inorg. Chem. 2009, 48, 7891–7904 7891
DOI: 10.1021/ic900764e
Experimental and Computational Studies of Binding of Dinitrogen, Nitriles, Azides,
Diazoalkanes, Pyridine, and Pyrazines to M(PR3)2(CO)3 (M = Mo, W; R = Me, iPr).†
Patrick Achord,‡ Etsuko Fujita,‡ James T. Muckerman,*,‡ Brian Scott,*,§ George C. Fortman, Manuel Temprado,
Xiaochen Cai, Burjor Captain,*, Derek Isrow, John J. Weir, James Eric McDonough, and Carl D. Hoff*,
‡Chemistry Department, Brookhaven National Laboratory, Upton, New York 11973, §Chemistry Division,
MS-J514, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, and Department of Chemistry,
University of Miami, Coral Gables, Florida 33146
Received April 21, 2009
The enthalpies of binding of a number of N-donor ligands to the complex Mo(PiPr3)2(CO)3 in toluene have been determined
by solution calorimetry and equilibrium measurements. The measured binding enthalpies span a range of ∼10 kcal mol-1
:
ΔHbinding = -8.8 ( 1.2 (N2-Mo(PiPr3)2(CO)3); -10.3 ( 0.8 (N2); -11.2 ( 0.4 (AdN3 (Ad = 1-adamantyl)); -13.8 (
0.5 (N2CHSiMe3); -14.9 ( 0.9 (pyrazine=pz); -14.8 ( 0.6 (2,6-Me2pz); -15.5 ( 1.8 (Me2NCN); -16.6 ( 0.4 (CH3CN);
-17.0 ( 0.4 (pyridine); -17.5 ( 0.8 ([4-CH3pz][PF6] (in tetrahydrofuran)); -17.6 ( 0.4 (C6H5CN); -18.6 ( 1.8 (N2CHC
(dO)OEt); and -19.3 ( 2.5 kcal mol-1 (pz)Mo(PiPr3)2(CO)3). The value for the isonitrile AdNC (-29.0 ( 0.3) is 12.3 kcal
mol-1 more exothermic than that of the nitrile AdCN (-16.7 ( 0.6 kcal mol-1). The enthalpies of binding of a range of arene
nitrile ligands were also studied, and remarkably, most nitrile complexes were clustered within a 1 kcal mol-1 range despite
dramatic color changes and variation of νCN. Computed structural and spectroscopic parameters for the complexes Mo
(PiPr3)2(CO)3L are in good agreement with experimental data. Computed binding enthalpies for Mo(PiPr3)2(CO)3L exhibit
considerable scatter and are generally smaller compared to the experimental values, but relative agreement is reasonable.
Computed enthalpies of binding using a larger basis set for Mo(PMe3)2(CO)3L show a better fit to experimental data than that
for Mo(PiPr3)2(CO)3L using a smaller basis set. Crystal structures of Mo(PiPr3)2(CO)3(AdCN), W(PiPr3)2(CO)3(Me2NCN), W
(PiPr3)2(CO)3(2,6-F2C6H3CN), W(PiPr3)2(CO)3(2,4,6-Me3C6H2CN), W(PiPr3)2(CO)3(2,6-Me2pz), W(PiPr3)2(CO)3(AdCN),
Mo(PiPr3)2(CO)3(AdNC), and W(PiPr3)2(CO)3(AdNC) are reported.
Introduction
Previously we have reported4 computational analysis of the
thermochemistry of ligand binding to W(PCy3)2(CO)3 for a
wide range of ligands. This report focuses on the binding of N
donor ligands to Mo(PiPr3)2(CO)3.
The first compounds characterized to bind H2 as a mole-
cule,1 M(PR3)2(CO)3 (M=Cr, Mo, W; R=isopropyl (iPr),
cyclohexyl (Cy)), were discovered nearly 30 years ago.2 They
are ideal for studying physical aspects of binding and oxida-
tive addition at a sterically constrained, active metal center.3
In comparison to N-donor ligands, the organometallic
thermochemistry of P donor ligands has been more exten-
sively investigated.5 Phosphines play an important role in
modifying metal catalyst reactivity, and an understanding of
the factors controlling this important bond has emerged in
terms of the steric and electronic parameters originally
delineated by Tolman6 in 1977. For organometallic com-
plexes, the bond to N donor ligands is generally more labile,
and there have been relatively few studies of N donor ligand
binding in organometallic thermochemistry. Taube7 and co-
workers have investigated enthalpies of binding to [Ru-
(NH3)5(H2O)]2þ in aqueous solution for a range of N-donor
ligands. Dinitrogen was found to be a stronger ligand than
† The authors dedicate this paper to Dr. Gregory J. Kubas, Los Alamos National
Laboratory.
*To whom correspondence should be addressed. E-mail: c.hoff@
miami.edu.
(1) Kubas. G. J. Metal Dihydrogen and σ-Bond Complexes; Kluwer
Academic/Plenum Publishers: New York, 2001.
(2) Kubas, G. J. J. Chem. Soc., Chem. Commun. 1980, 61.
(3) (a) McDonough, J. E.; Weir, J. J.; Sukcharoenphon, K.; Hoff, C. D.;
Kryatova, O. P.; Rybak-Akimova, E. V.; Scott, B. L.; Kubas, G. J.;
Mendiratta, A.; Cummins, C. C. J. Am. Chem. Soc. 2006, 128, 10295. (b)
Lang, R. F.; Ju, T. D.; Kiss, G.; Hoff, C. D.; Bryan, J. C.; Kubas, G. J. J. Am.
Chem. Soc. 1994, 116, 7917. (c) Torres, L.; Moreno, M.; Lluch, J. M. J. Phys.
Chem. A 2001, 105, 4676. (d) Rattan, G; Khalsa, K.; Kubas, G. J.; Unkefer, C. J.;
Van der Sluys, L. S.; Kubat-Martin, K. A. J. Am. Chem. Soc. 1990, 112, 3855. (e)
Gonzalez, A. A.; Hoff, C. D. Inorg. Chem. 1989, 28, 4295. (f) Gonzalez, A. A.;
Zhang, K.; Nolan, S. P.; de la Vega, R. L.; Mukerjee, S. L.; Hoff, C. D.; Kubas, G.
J. Organometallics 1988, 7, 2429.
(5) Hoff, C. D. Prog. Inorg. Chem. 1992, 40, 503.
(6) Tolman, C. A. Chem. Rev. 1977, 77, 313.
(7) (a) Armor, J. N.; Taube, H. J. Am. Chem. Soc. 1970, 92, 6170. (b)
Wishart, J. F.; Taube, H.; Breslauer, K. J.; Isied, S. S. Inorg. Chem. 1984, 23,
2997.
(4) Muckerman, J. T.; Fujita, E.; Hoff, C. D.; Kubas, G. J. J. Phys. Chem.
B 2007, 111, 6815.
r
2009 American Chemical Society
Published on Web 07/21/2009
pubs.acs.org/IC

7892 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
acetonitrile by ≈ 1 kcal mol-1. Computation of the gas-phase
enthalpies of binding of nitriles to Agþ were recently shown8
to span a range of over 30 kcal mol-1 depending on the
substituent X in NtC;X. Experimental data for binding to
Alþ in the gas phase showed a much more modest depen-
dence on R of only 4-5 kcal mol-1.9 Mass spectroscopic
studies of binding of alkyl nitriles and benzonitrile to Niþ and
Coþ showed a linear correlation with the proton affinity of
the nitrile; however, the range was small with benzonitrile
widespread belief that it is a dominant factor controlling N2
binding to metals.
Experimental Section
General Considerations. Unless stated otherwise, all opera-
tions were performed in a drybox under an atmosphere of
purified argon. Toluene, tetrahydrofuran (THF), benzene, hep-
tane, and octane were distilled over sodium metal before use.
Distilled solvents were transferred under out-flow of argon into
vacuum-tight vessels and were degassed before being transferred
into the drybox. C6D6 was purchased from Sigma-Aldrich.
Compounds M(PiPr3)2(CO)3 (M = W, Mo) were prepared;16
AdNC was crystallized from a CH2Cl2/heptane mixture and
sublimed at ≈ 85 °C before use.17 All other compounds were
being favored over acetonitrile by ≈2 kcal mol-1 10
.
Kovacs
and co-workers have shown an opposite trend in binding
enthalpies where acetonitrile is favored over benzonitrile by
≈2 kcal mol-1 in FeIII complexes which are models for nitrile
hydratase.11
1
used as received. H NMR spectra were recorded on a Bruker
This work reports on an experimental and computational
investigation of bonding in the complexes Mo(CO)3-
(PiPr3)2(NX) where NX can be NtN, NtCR, NtN-NR,
NdNdCHR, and other N-atom donors as a function of R to
try to assess the role of π bonding in these systems. There are
reasons to anticipate that π bonding may play an important,
possibly dominant role. It is generally held that in terms of
bonding ability to a metal complex, N2 is intrinsically inferior
to CO both as a σ donor and a π acceptor. Computations by
Frenking12 imply that N2 binds mainly through metalfN2
back-bonding. Similar conclusions have been made regard-
ing binding to [Mn(PCy3)2(CO)3]þ1: “N2 is an exceedingly
poor electron donor, even toward strong electrophiles, where
it is much feebler than the weakest known ligands, e.g.,
CH2Cl2. The complete lack of binding to [Mn(PCy3)2-
(CO)3][B(C6H3(3,5-CF3)2)4] and other electron-poor cationic
complexes indicates that N2 apparently can only be stabilized
on a metal center by a high degree of π-back donation, even
in actinide complexes.”13
The importance of π-backbonding and electron delocali-
zation in Cr(CO)5(CHX) (X=H, OH, OCH3, NH2, NHCH3)
complexes was highlighted by work of Frenking which
showed that differences in the CrdC bond in Cr(CO)5(CHX)
depend mainly on the π-bonding ability of X and that the
strong bond present when X=H (97.9 kcal mol-1) is nearly
30 kcal mol-1 weaker when X=NH2 (68.7 kcal mol-1).14
Gusev has recently15 computed donor properties for a range
of two-electron donors and concluded that nitrogen donors
follow the order NMe3 > NH3 > oxazoline > pyridine >
MeN=CH2 > acetonitrile . N2 with respect to σ donating
properties in metal complexes. The situation with respect
to π-bonding for NX ligands is not clear except for the
AVANCE-400 spectrometer at T=20 °C. Chemical shifts are
reported with respect to internal solvent: 7.16 ppm (C6D6).
FTIR spectra were obtained using a Perkin-Elmer 2000 FTIR/
Microscope system that has been described elsewhere.18 UV-
vis spectra were recorded on a Perkin-Elmer Lambda 900
spectrophotometer.
Synthesis of Mo(PiPr3)2(CO)3(L) (L =2,6-Me2pz (pz = pyra-
zine), Me2NCN, 4-Me2NC6H4CN, C6F5CN, CH3CN, AdCN (Ad=
1-adamantyl), 2,4,6-Me3C6H2CN, 2,6-F2C6H3CN, N2CHSiMe3,
N2CHC(dO)OEt, NC5H5 (pyridine), 4-CF3C6H4CN, [4-Mepz]-
[PF6], C6H5CN, AdNC.). All syntheses were performed in a similar
manner. In a representative preparation a 20 mL scintillation vial
was charged with 0.20 mmol of Mo(PiPr3)2(CO)3 and an equimo-
lar amount of AdNC. 5.0 mL of toluene were added. The reaction
solution instantaneously turned from brown to yellow. The solvent
was then removed under reduced pressure. The solid was then
washed twice with a minimal amount of heptane to remove excess
ligand and M(PiPr3)2(CO)4. A yield of 0.114 g (86% yield) of
yellow crystalline M(PiPr3)2(CO)3(AdNC) was recovered. All
products except the N2CHSiMe3 and N2CHC(dO)OEt com-
plexes were recovered with spectroscopic yields of 100% and
isolated yields of >75%. The complexes of N2CHSiMe3 and
N2CHC(dO)OEt were formed quantitatively based on FTIR
and NMR analysis; however, the initially formed complexes
decomposed to unknown products upon removal of solvent and
isolated yields could not be determined. FTIR spectral data for the
N2CHSiMe3 and N2CHC(dO)OEt complexes are listed in the
Supporting Information, Tables ST-2 and ST-3. 1H NMR shifts
(C6D6) in ppm were: 2,6-Me2pz, 1.23 36H(q), 2.12 6H(m), 8.76 2H
(s), 2.00 6H(s); Me2NCN, 1.39 36H(q), 2.36 6H(m), 1.79 6H(s);
4-Me2NC6H4CN, 1.45 36H(q), 2.44 6H(m), 7.13 2H(d), 6.04 2H(d),
2.21 6H(s); C6F5CN, 1.39 36H(q), 2.41 6H(m); CH3CN, 1.35 36H
(q), 2.32 6H(m), 0.67 3H(s); AdCN, 1.40 36H(q), 2.40 6H(m), 1.59
6H(d), 1.50 3H(m), 1.24 6H(m); 2,4,6-MeC6H2CN, 1.38 36H(q),
2.39 6H(m), 6.36 2H(s), 2.28 6H(s), 1.83 3H(s); 2,6-F2C6H3CN,
1.40 36H(q), 2.42 6H(m), 6.28 1H(m), 6.04 2H(t); NC5H5, 1.25
36H(q), 2.11 6H(m), 8.90 2H, 6.15 2H, 6.61 1H; 4-CF3C6H4, 1.37
36H(q), 2.38 6H(m), 6.92 2H, 6.89 2H; N2CHSiMe3, 1.27 36H(q),
2.22 2.38 6H(m), 2.78 1H(s) 0.03 9H(s); C6H5CN, 1.38 36H(q),
2.38 6H(m), 7.04 2H(d), 6.84 2H(t), 6.74 1H(t); N2CHC(dO)
OC2H5, 1.23 36H(q), 2.24 6H(m); AdNC, 1.38 36H(q), 2.39 6H
(m), 1.75 6H(d), 1.68 3H(m), 1.28 6H(m).
(8) Shoeib, T.; El Aribi, H.; Siu, K. W. M.; Hopkinson, A. C. J. Phys.
Chem. A 2001, 105, 710.
(9) Uppal, J. S.; Staley, R. H. J. Am. Chem. Soc. 1982, 104, 1235. See also:
El. Aribi, H. E.; Rodriquez, C. F.; Shoeib, T.; Ling, Y.; Hopkinson, A. C.; Siu, K.
W. M. J. Phys. Chem. 2002, 106, 8798.
(10) Chen, L. Z.; Miller, J. M. Org. Mass Spectrom. 1992, 27, 883.
(11) Shearer, J.; Jackson, H. L.; Schweitzer, D.; Rittenberg, D. K.; Leavy,
T. M.; Kaminsky, W.; Scarrow, R. C.; Kovacs, J. A. J. Am. Chem. Soc. 2002,
124, 11417.
(12) Ehlers, A. W.; Dapprich, S.; Vyboishchickov, S. F.; Frenking, G.
Organometallics 1996, 15, 105.
Synthesis of W(PiPr3)2(CO)3(L) (L=2,6-Me2pz, Me2NCN,
AdCN, 2,4,6-MeC6H2CN, 2,6-F2C6H3CN, AdNC.). Syntheses
were the same as the Mo analogues. All yields were >75%. 1H
NMR shifts (C6D6) ppm: 2,6-Me2pz, 1.22 36H(q), 2.23 6H(m),
8.89 2H(s), 1.97 6H(s); Me2NCN, 1.37 36H(q), 2.43 6H(m), 1.81
(13) Toupadakis, A.; Kubas, G. J.; King, W. A.; Scott, B. L.; Huhmann-
Vincent, J. Organometallics 1998, 17, 5315.
ꢀ
ꢁ
(14) Cases, M.; Frenking, G.; Duran, M.; Sola, M. Organometallics 2002,
(16) (a) Kubas, G. J. Organomet. Synth. 1986, 3, 264. (b) Khalsa, G. R. K.;
Kubas, G. J.; Unkefer, C. J.; Van Der Sluys, L. S.; Kubat-Martin, K. A. J. Am.
Chem. Soc. 1990, 112, 3866.
(17) Sasaki, T.; Nakanishi, A.; Ohno, M. J. Org. Chem. 1981, 46, 5445.
(18) Ju, T. D.; Capps, K. B.; Roper, G. C.; Hoff, C. D. Inorg. Chim. Acta
1998, 270, 488.
21, 4182.
(15) A range of over 10 kcal mol-1 was computed for the isodesmic
exchange reaction IrCp(CO)(L) þ IrCp(Xe)(PF3)=IrCp(CO)(Xe) þ IrCp(L)
(PF3). A key factor stabilizing IrCp(CO)(L) as a function of L appears to be
the σ donor ability of L. See: Gusev, D. G. Organometallics 2009, 28, 763.

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7893
6H(s); 2,4,6-MeC6H2CN, 1.36 36H(q), 2.47 6H(m), 6.36 2H(s),
2.28 6H(s), 1.82 3H(s); 2,6-F2C6H3CN, 1.38 36H(q), 2.48 6H
(m), 6.28 1H(m), 6.03 2H(t); AdCN, 1.39 36H(q), 2.47 6H(m),
1.60 6H(d), 1.51 3H(m), 1.26 6H(m); AdNC, 1.37 36H(q), 2.45
6H(m), 1.75 6H(d), 1.68 3H(m), 1.29 6H(m).
was considered as uniform, as any changes due to reaction were
small compared to experimental error. Relative concentrations
of the organometallic complexes were determined by band
shape analysis of the spectra (Supporting Information, Figure
SF-2). The absorbance maxima of the major and minor bands
were measured. Assuming a symmetrical band shape, it was
determined that the absorbance of the major band at the peak
maximum position of the minor band was 0.19 times the major
band height. This was done by measuring the absorbance
equidistant on the low wavenumber side. From this the ratio
of the two bands was determined.
Calorimetric Measurement of AdNC-Driven Displacement of
L from Mo(PiPr3)2(CO)3L. In a glovebox, a solution of the
target compound was prepared in 5 mL of freshly distilled
toluene, and transferred via syringe into the calorimeter cell. A
few milligrams of freshly sublimed AdNC were loaded into the
solid sample holder of the calorimeter cell. The cell was sealed,
removed from the glovebox, and loaded into a Setaram C-80
calorimeter. After thermal equilibration, the reaction was ini-
tiated and followed to completion at 30 °C. Following return to
baseline, the cell was returned to the glovebox, and its contents
examined by FTIR to confirm complete conversion. The re-
ported average values are based on four to six independent
determinations for the enthalpy of reaction based on solid
AdNC. A value (all species solvated) was determined from the
measured values and the enthalpy of solution of AdNC in
toluene (ΔHsoln=3.1 ( 0.1 kcal mol-1).
Calorimetric Measurements of Nitriles. The determination of
the enthalpy of binding of a series of aromatic nitriles was
carried out using a crystallized sample of Mo(PiPr3)2-
(CO)3(AdCN) as the limiting reagent following the isonitrile
displacement procedure. In a typical measurement, a few
milligrams of solid Mo(PiPr3)2(CO)3(AdCN) were added to a
solution of the free nitrile in toluene solution in the calorimeter.
Reactions were quantitative under these conditions as
confirmed by FTIR spectroscopy after disassembly of the
calorimeter.
Keq for Binding of N2 to Mo(PiPr3)2(CO)3. A thermostatted
medium-pressure FTIR cell with CaF2 windows (Harrick Scien-
tific) was fitted with a 40 mL stainless steel bomb, valves,
pressure sensor, and a thermistor probe inserted directly into
the solution. This assembly was loaded in the glovebox with
0.3050 g (0.6100 mmol) Mo(PiPr3)2(CO)3 and 20.0 mL of 4.0
mM Mo(CO)6 in toluene. The CO-stretching frequency (νco at
1984 cm-1) of Mo(CO)6 was used in data analysis as a calibra-
tion standard to correct for increased path lengths under the
higher pressures of N2. Initially, the cell was pressured to ∼280
psi N2 at room temperature, to obtain a maximum value for the
absorbance due to Mo(PiPr3)2(CO)3(N2). That maximum value
was then used to calculate the relative concentrations by means
of difference. The pressure was released to ∼15 psi, and the
temperature was incrementally varied from 336 to 284 K.
Corrections were made to the observed pressure of N2 because
of increased toluene vapor pressure at higher temperature.
Values for Keq were interpreted in units of atm-1. At the end
of the experiment the cell was repressurized to ∼280 psi at room
temperature, and the sample was again examined by FTIR
spectroscopy. The resultant spectrum was ultimately compared
to the initial spectrum, indicating no decomposition of the
reactive Mo(PiPr3)2(CO)3 complex. Variable temperature FTIR
spectral data are shown in the Supporting Information, Figure
SF-1. Three independent experiments led to values of ΔH =
Crystallographic Analyses. Yellow single crystals of Mo
(PiPr3)2(CO)3(AdNC), Mo(PiPr3)2(CO)3(AdCN) and W(PiPr3)2-
(CO)3(AdCN); orange single crystals of W(PiPr3)2(CO)3(2,4,6-
Me3C6H2CN) and maroon single crystals of W(PiPr3)2(CO)3-
(2,6-F2C6H2CN) suitable for X-ray diffraction analyses were
obtained by evaporation of a mixture of benzene and octane at
21 °C under an inert atmosphere of argon. Blue single crystals of
W(PiPr3)2(CO)3(2,6-Me2pz), yellow single crystals of W(PiPr3)2-
(CO)3(Me2NCN), and yellow single crystals of W(PiPr3)2-
(CO)3(AdNC) were grown from a layered solution of heptane/
toluene that was placed in a -20 °C freezer for 1 month. Once
separated from the mother liquor, the solid crystals were fairly air
stable. Each crystal was glued onto the end of a thin glass fiber; no
decomposition was observed during the data collection period.
X-ray intensity data were measured using a Bruker SMART
APEX2 CCD-based diffractometer with Mo KR radiation
19
˚
(λ=0.71073 A). Raw data frames were integrated with the
SAINTþ program using a narrow-frame integration algo-
rithm.19 Corrections for Lorentz and polarization effects were
also applied with SAINTþ. An empirical absorption correction
based on the multiple measurement of equivalent reflections was
applied using SADABS. All structures were solved by a combi-
nation of direct methods and difference Fourier syntheses, and
refined by full-matrix least-squares on F2 using the SHELXTL
software package.20 All non-hydrogen atoms were refined with
anisotropic displacement parameters. Hydrogen atoms were
placed in geometrically idealized positions and included as stan-
dard riding atoms during the least-squares refinements. Crystal
data, data collection parameters, and results of the analyses are
listed in Supporting Information, Tables ST-4 and ST-5.
-10.3 ( 0.8 kcal mol-1 and ΔS=-32.2 ( 2.6 cal mol-1 K-1
.
Calorimetric Measurements of the Enthalpy of Binding of W
(PiPr3)2(CO)3(N2) to W(PiPr3)2(CO)3. In a representative
experiment, 19.3 mg of crystalline (μ-N2)[W(PiPr3)2(CO)3]2
was loaded into the solid compartment of a calorimeter cell.
The cell was then loaded with 5 mL of toluene and then sealed
under a nitrogen atmosphere. The sealed cell was taken from the
glovebox and loaded into a Setaram C-80 calorimeter. After
thermal equilibration, the reaction was initiated and followed to
completion at 30 °C. Following return to baseline, the cell was
taken back into the glovebox, and its contents examined by
FTIR to confirm complete conversion to W(PiPr3)2(CO)3(N2).
The reported average values are based on three independent
experiments. From the measured values and the enthalpy of
solution of (μ-N2)[W(PiPr3)2(CO)3]2 in toluene under an Ar
atmosphere (ΔH = þ6.7 ( 0.3 kcal mol-1), a value with all
species in solution for the reaction of (μ-N2)[W(PiPr3)2(CO)3]2
with N2 to form W(PiPr3)2(CO)3(N2) was determined to be
Compounds Mo(PiPr3)2(CO)3(AdCN), Mo(PiPr3)2(CO)3-
(AdNC), W(PiPr3)2(CO)3(2,4,6-Me3C6H2CN), and W(PiPr3)2-
(CO)3(2,6-F2C6H2CN) crystallized in the monoclinic crystal
system. For compounds Mo(PiPr3)2(CO)3(AdCN), Mo(PiPr3)2-
(CO)3(AdNC), and W(PiPr3)2(CO)3(2,6-F2C6H2CN) the sys-
tematic absences in the intensity data were consistent with the
unique space group P21/n. With Z=8, there are two formula
equivalents of the molecule present in the asymmetric crystal
unit for these three compounds. Compounds Mo(PiPr3)2-
(CO)3(AdCN) and Mo(PiPr3)2(CO)3(AdNC) are isomorphous
and isostructural. For compound W(PiPr3)2(CO)3(2,4,6-
Me3C6H2CN) the systematic absences in the intensity data were
consistent with the unique space group P21/c.
ΔH=-1.5 ( 0.6 kcal mol-1
.
Keq for Binding of Pyrazine to Mo(PiPr3)2(CO)3. The mod-
ified FTIR cell assembly was loaded in the glovebox with
0.1939 g (0.4 mmol) of Mo(PiPr3)2(CO)3, 0.5025 g (6.3 mmol)
of pyrazine, and 12 mL of C6D6. The concentration of pyrazine
(19) Apex2 Version 2.2-0 and SAINTþ Version 7.46A; Bruker Analytical
X-ray System, Inc.: Madison, WI, 2007.
(20) Sheldrick, G. M. SHELXTL Version 6.1; Bruker Analytical X-ray
Systems, Inc.: Madison, WI, 2000.

7894 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
Compounds W(PiPr3)2(CO)3(AdCN) and W(PiPr3)2(CO)3-
(2,6-Me2pz) crystallized in the monoclinic crystal system. For
compound W(PiPr3)2(CO)3(AdCN) the systematic absences in
the intensity data were consistent with the unique space group
P212121. There is minor disorder present in the adamantyl group
(atoms C51-C59) which was not modeled because of satisfac-
tory low R factors, R1=3.45%, in the final stages of the ref-
inement. The carbon atoms of the disordered adamantyl group
were refined with isotropic displacement parameters. For com-
pound W(PiPr3)2(CO)3(2,6-Me2pz) the systematic absences in
the intensity data were consistent with the unique space group
Pbca. With Z = 16, there are two formula equivalents of the
molecule present in the asymmetric crystal unit.
the complexes. Since this behavior complicates the direct
comparison of binding enthalpies, we ran a parallel set of
calculations on the corresponding PMe3 compounds. As an
additional test of robustness, we performed single-point
calculations on all optimized PMe3 structures using the
MWB28 ECP and basis27 for Mo (MWB60 ECP and basis27
for W) containing one set of f functions, and the 6-311G(d,p) 5d
basis28 for all other elements. The polarization functions in the
larger basis sets can, in principle, better describe the delicate
balance between σ donation and π backdonation. Finally, TD-
B3LYP calculations were carried out on most of the PiPr3
solution-phase structures.
Crystals of W(PiPr3)2(CO)3(AdNC) and W(PiPr3)2(CO)3-
(Me2NCN), recrystallized from heptane, were mounted in a
nylon cryoloop from Paratone-N oil under argon gas flow,
placed on a Bruker diffractometer, and cooled to 141 K using
a Bruker Kryoflex cryostat located at LANL. The data for the
AdNC crystal were collected on a Bruker P4/1k-CCD, while the
data for the Me2NCN crystal were collected on a Bruker APEX
II CCD. Both instruments were equipped with sealed, graphite
Results
Enthalpies of Ligand Binding. Previous ligand binding
studies have focused mainly on the cyclohexyl phosphine
complexes M(PCy3)2(CO)3 (M = Cr, Mo, W),29 although
some data on the more soluble isopropyl phosphine com-
plexes have been reported.30 For the N-donor ligands studied
here, several techniques were used to assemble experimental
data: calorimetric studies of ligand binding and displace-
ment, variable temperature equilibrium studies, as well as
ligand competition equilibrium studies at fixed temperature.
These data and how they were derived are presented sepa-
rately and then combined later to form a stability series (see
Table 4) for comparison to computational data.
Enthalpies of Reaction with 1-Adamantyl Isocyanide.
Calorimetric measurements were made of the enthalpy of
displacement of weakly bound NX ligands by AdNC as
shown in reaction 1 and summarized in Supporting
Information, Table ST-1.
˚
monochromatized Mo KR X-ray sources (λ= 0.71073 A). For
each of the two structures a hemisphere of data was collected
employing j or ω scans and 0.50° frame widths. All data were
corrected and absorption (SADABS)21 and Lorentz-polariza-
tion effects. Decay of reflection intensity was monitored and
corrected via analysis of redundant frames. The structures were
solved using Direct methods and difference Fourier techniques.
All hydrogen atom positions were idealized, and rode on the
atom to which they were attached. The final refinement included
anisotropic temperature factors on all non-hydrogen atoms.
Structure solution, refinement, graphics, and creation of pub-
lication materials were performed using SHELXTL.20 Addi-
tional details of data collection may be found in the Supporting
Information, Tables ST-6 and ST-7.
Computational Details. All electronic structure calculations
used the B3LYP hybrid density functional method as imple-
mented in the Gaussian 03 suite of programs.22 Structures were
optimized using the LANL2DZ basis, LANL2DZ ECP23 and
basis for the transition-metal centers and P atoms, and the
Dunning-Huzinaga D95 V24 basis for all other atoms. Optimi-
zations were first carried out in the gas phase to obtain the basis
set superposition error (BSSE) using counterpoise calcula-
tions.25 The structures were then reoptimized and vibrational
frequencies calculated in “toluene solution” (or “THF solu-
tion”) using a polarizable continuum model (PCM)26 and
UAHF radii. Computed electronic energies were corrected for
zero-point energy, thermal energy, and entropic effects to obtain
the corresponding thermodynamic properties H° and G° in
solution. Adding gas-phase BSSE corrections to H° provided
the binding enthalpies.
In a typical measurement, a few milligrams of solid
AdNC (recrystallized and sublimed) were added to
a solution of the complex in the calorimeter. Reaction
is rapid and quantitative under these conditions. The
enthalpy of binding of AdNC to Mo(PiPr3)2(CO)3, as
(27) (a) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123. (b) Martin, J. M. L.; Sundermann, A. J. Chem.
Phys. 2001, 114, 3408.
(28) (a) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639. (b)
Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650.
(29) (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) Weir, J. J.; McDonough, J. E.; Fortman, G.;
Isrow, D.; Hoff, C. D.; Scott, B.; Kubas, G. J. Inorg. Chem. 2007, 46, 652. (c)
McDonough, J. E.; Weir, J. J.; Sukcharoenphon, K.; Hoff, C. D.; Kryatova, O. P.;
Rybak-Akimova, E. V.; Scott, B. L.; Kubas, G. J.; Mendiratta, A.; Cummins, C. C. J.
Am. Chem. Soc. 2006, 128, 10295. (d) Lang, R. F.; Ju, T. D.; Kiss, B.; Hoff, C. D.;
Bryan, J. C.; Kubas, G. J. J. Am. Chem. Soc. 1994, 116, 7917. (e) Kubas, G. J.;
Burns, C. J.; Khalsa, G. R. K.; Van Der Sluys, L. S.; Kiss, G.; Hoff, C. D.
Organometallics 1992, 11, 3390. (f) Zhang, K.; Gonzalez, A. A.; Mukerjee, S. L.;
Chou, S. J.; Hoff, C. D.; Kubat-Martin, K. A.; Barnhart, D.; Kubas, G. J. J. Am.
Chem. Soc. 1991, 113, 9170. (g) Gonzalez, A. A.; Hoff, C. D. Inorg. Chem. 1989,
28, 4285. (h) Gonzalez, A. A.; Hoff, C. D. Inorg. Chem. 1989, 28, 4295. (i)
Gonzalez, A. A.; Zhang, K.; Nolan, S. P.; de la Vega, R. L.; Mukerjee, S. L.; Hoff, C.
D.; Kubas, G. J. Organometallics 1988, 7, 2429. (j) Gonzalez, A. A.; Mukerjee, S.
L.; Chou, S. L.; Kai, Z.; Hoff, C. D. J. Am. Chem. Soc. 1988, 110, 4419.
(30) Data on thermochemistry of PiPr3 can be found in ref 29 (e) and:
Bender, B. R.; Kubas, G. J.; Jones, L. H.; Swanson, B. I.; Eckert, J.; Capps,
K. B.; Hoff, C. D J. Am. Chem. Soc. 1997, 119, 9179.
When possible, starting structures were modeled after crystal
structures. We found that optimized structures were dependent
on initial structure guesses, pointing toward multiple local
minima. It became apparent that this level of theory poorly
i
described the conformations of the Pr groups in some of
::
::
(21) Sheldrick, G. SADABS 2.10; University of Gottingen: Gottingen,
Germany, 2001.
(22) Frisch, M. J. et al. Gaussian 03, revision D.01; Gaussian, Inc.:
Wallingford, CT, 2004.
(23) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (b) Wadt, W.
R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. (c) Hay, P. J.; Wadt, W. R. J. Chem.
Phys. 1985, 82, 299.
(24) Dunning, T. H. Jr., Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., III, Ed.; Plenum: New York, 1976; Vol. 3, pp 1-28.
(25) (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.
(26) (a) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117. (b)
Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117. (c) Mennucci, B.;
Tomasi, J. J. Chem. Phys. 1997, 106, 5151.

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7895
Scheme 1. Enthalpies of Binding of Pyrazine as Mononuclear and
Bridging Forms
good agreement with the independently measured calori-
metric value (see Supporting Information, Table S-T1)
for binding of 2,6-dimethylpyrazine (ΔH = -14.8 (
0.6 kcal mol-1), as shown in reaction 4.
The ligand 2,6 dimethylpyrazine does not form a
dinuclear complex; presumably steric factors prohibit
binding to the N atom bearing ortho methyl groups.
The close agreement between the derived enthalpy of
binding of monodentate pyrazine and 2,6-dimethylpyr-
azine appears reasonable and gives support to the mea-
sured data.
Enthalpy of Binding of N2 for the Mononuclear and
Bridging Forms. Binding of N2 to Mo(PiPr3)2(CO)3 is not
quantitative, but an equilibrium is rapidly established as
shown in reaction 5.
shown in reaction 2, was measured as a reference point
and gave ΔH=-29.0 ( 0.3 kcal mol-1 with all species in
toluene solution.
Since reaction 2 involves displacement of a C-H agostic
bond with an estimated bond strength of 7 ( 3 kcal mol-1
(see later discussion), this value must be added on to
obtain estimates of the absolute Mo-ligand bond strength.
Enthalpy of Binding of Pyrazine for Mononuclear and
Bridging Forms. The enthalpy of reaction of pyrazine with
2 equiv of Mo(PiPr3)2(CO)3 to form μ-(C4H4N2)[Mo-
(PiPr3)2(CO)3]2 was measured by reaction with AdNC
as -34.2 ( 1.6 kcal mol-1 as shown in Supporting
Information, Table ST-1. This corresponds to the sum
of the enthalpies of binding to form the mononuclear
complex (C4H4N2)Mo(PiPr3)2(CO)3 followed by its reac-
tion to form the dinuclear complex μ-(C4H4N2)[Mo-
(PiPr3)2(CO)3]2. The difference between these two enthal-
pies of binding can be determined by the equilibrium
shown in reaction 3.
Using FTIR techniques, the value of Keq for this
reaction was measured as a function of temperature,
and the data are shown in Supporting Information,
Figure SF-3.
In spite of experimental scatter in the individually mea-
sured values, due in part to the highly air-sensitive nature
of the solutions, the derived thermodynamic parameters,
ΔH=-10.3 ( 0.8 kcal mol-1 and ΔS=-32.2 ( 2.6 cal
mol-1 K-1, are in keeping with data for binding of N2
to Mo(PCy3)2(CO)3 for which we have previously found27
ΔH=-9.0 kcal mol-1 and ΔS=-32.1 cal mol-1 K-1. The
slightly more exothermic binding of N2 to the PiPr3 com-
plex compared to the PCy3 analogue has been observed for
other ligands as well, and is attributed to differences in the
agostic bond strength, as well as to steric factors.31
It was of interest to estimate the enthalpy of binding of
N2 in forming the bridging complex μ-(N2)[Mo(PiPr3)2-
(CO)3]2, particularly for comparison to the formation of
μ-(C4H4N2)[Mo(PiPr3)2(CO)3]2 described above. In spite
of the fact that this complex is stable as a solid, its stability
in solution prevented quantitative measurement. How-
ever, the corresponding W complex is more stable, and we
were able to determine the difference between the binding
energies of N2 in the mononuclear and bridging com-
plexes by simple measurement of the enthalpies of solu-
tion. Dissolution of the bridging dinuclear complex under
argon does not result in any reaction, and its enthalpy
corresponds to reaction 6.
The bridging complex is favored over the monometallic
species, and large excesses of pyrazine are required to
produce detectable amounts of the mononuclear adduct.
Variable temperature FTIR studies of reaction 3 yielded
Keq as a function of temperature as shown in Supporting
Information, Figure SF-2 from which it was determined
that ΔH=þ4.4 ( 0.2 kcal mol-1 and ΔS=þ5.8 ( 0.3 cal
mol-1 K-1. The enthalpy of reaction 3, when combined
with the enthalpy of binding of (-34.2 ( 1.6 kcal mol-1
)
for μ-(C4H4N2)[Mo(PiPr3)2(CO)3]2 can be used to derive
the thermochemical data shown in Scheme 1.
These data indicate that the second enthalpy of bind-
ing, -19.3 kcal mol-1, is significantly stronger than the
first, -14.9 kcal mol-1. Since steric factors would be
expected to oppose it, electronic factors must favor for-
mation of the μ-bridging dimer.
The derived data for binding of pyrazine to form the
(31) Kubas, G. J.; Burns, C. J.; Khalsa, G. R. J.; Van Der Sluys, L. S.;
Kiss, G.; Hoff, C. D. Organometallics 1992, 11, 3390.
mononuclear complex, ΔH=-14.9 ( 0.9 kcal mol-1, is in

7896 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
In contrast, dissolution of the bridging complex under 1
atm N2 results in quantitative conversion to the mono-
nuclear complex as shown in reaction 7.
Enthalpy of Binding of Dimethylcyanamide and Ethyl-
diazoacetate. Because of low solubility of the complexes
formed, it was difficult to make reliable calorimetric
measurements for the enthalpies of binding of these two
ligands. Instead, estimates of Keq for binding of ligands
near to them in bond strength were carried out. Quanti-
tative FTIR studies at room temperature showed that the
values for Keq for reactions 10 and 11 are 7 ( 2 and 5 ( 2,
respectively.
Subtracting reaction 7 from reaction 6 yields a ΔH=
þ1.5 ( 0.6 kcal mol-1 for reaction 8.
Assuming a similar behavior for Mo and W,32 we
estimate a value of -8.8 ( 1.2 kcal mol-1 for the enthalpy
of binding of Mo(PiPr3)2(CO)3(N2) to form the bridging
species.
Binding of Nitriles: Calorimetric and Spectroscopic
Data. As shown in the data in Supporting Information,
Table ST-1, the difference in enthalpy of binding of
benzonitrile and the alkyl nitriles acetonitrile and ada-
mantyl nitrile is only 1.0 ( 0.5 kcal mol-1. Because of the
exothermic nature of the reactions with AdNC, it was
decided that more accurate data for a range of nitriles
might be obtained by measuring the smaller enthalpy of
reaction when an arene nitrile displaces adamantyl nitrile
as shown in reaction 9.
This corresponds to values of ΔG at 298 K of -1.2 ( 0.3
and -1.0 ( 0.3 kcal mol-1 for reactions 10 and 11, res-
pectively. This implies that Me2NCN has a free energy of
binding 1.2 kcal mol-1 weaker than AdCN, and that N2C
(H)C(dO)OEt has a free energy of binding 1.0 kcal mol-1
stronger than benzonitrile. The free energy changes are
attributed to be primarily due to enthalpic rather than
entropic factors, an approximation that is likely valid
within experimental errors for these reactions.33
Quinuclidine. It is known that primary and secondary
amines bind to the Kubas complexes, but tertiary amines
do not.34 This is typically ascribed to steric factors at the
sterically crowded metal center which prohibit binding of
bulky amines. Even in the presence of a large excess of
quinuclidine, no detectable complex was observed at
room temperature.
In these reactions, the readily prepared and purified
AdCN complex, the structure of which is shown in
Supporting Information Figure SF-4, was used as the
limiting reagent and the experimental enthalpy of reac-
tion measured in its reactions with excess ArCN. The
measured enthalpies of these reactions for a set of sub-
stituted benzonitriles (Ar = 4-NMe2C6H4; 2,4,6-
Me3C6H2; 2,6-F2C6H3; 4-CF3C6H4; F5C6) were all small,
and the procedural dilution of the polar nitriles was found
to play a role in the calorimetric measurements nearly as
significant as the reactions themselves. The enthalpies
measured experimentally for reaction 9 for a series of
benzonitriles, as well as the derived enthalpies of binding,
are collected in Table 1.
X-ray Structure Determinations. Crystal structures of
W(PiPr3)2(CO)3(2,4,6-Me3C6H2CN), W(PiPr3)2(CO)3(2,6-
Me2pz), and W(PiPr3)2(CO)3(Me2NCN) are shown in
Figures 2-4. Additional structures are shown in the Support-
ing Information Figures SF-3 (M=Mo; L=AdNC), SF-4 (W
= Mo; L = AdNC), SF-5 (M = Mo; L = AdCN), SF-6
(M=W; L=AdCN), and SF-7 (M=W; L=2,6-F2C6H2CN).
Tables 2 and 3 recapitulate the highlighted bond dis-
tances and angles of the X-ray crystal structures, and
compare them to computational results. Universally, the
overall mononuclear geometries were slightly distorted
i
octahedra. The Pr groups show two “axial” and one
“equatorial” geometries per phosphine. The rotation of
such groups, as well as the absolute stereochemistry of
each “axial” Pr, is not uniform across the ligand spec-
Because the entire enthalpy range was <1 kcal mol-1
,
i
the bond strengths in solution indicate only a small
dependence of the Mo-N bond on R in the complexes.
As shown in Figure 1, this limited range of bond strengths
is accompanied by a dramatic change in color for these
adducts. Spectroscopic data are summarized in Support-
ing Information, Table ST-2.
trum. This is consistent with the distortions, as complexes
must accommodate the ligands and this involves a non-
negligible steric component. This is also consistent with
the non-binding of quinuclidine; such a ligand is so bulky
(33) Estimation of the entropy of ligand substitution of 4-CF3C6H4CtN
and Mo(PiPr3)2(CO)3(4-Me2NC6H4CN) yielded an estimate of ΔS°=-5 ( 2
cal mol-1 K-1. At room temperature this corresponds to about 1.5 kcal mol-1
in terms of free energy. The error limits on ΔH° derived in this way were
increased by this amount.
(34) Wasserman, H. J.; Kubas, G. J.; Ryan, R. R. J. Am. Chem. Soc. 1986,
108, 2294.
(32) Work on related ligand binding shows that enthalpies of binding of
Mo and W closely resemble each other, with binding to Mo being ≈ 0.9 times
that of W. Because of the low value of this difference, and the high
experimental error, no change was made. Under 1 atm of N2 pressure we
see no evidence for any of the dinuclear complex in solution, only terminal
binding and unsaturated complex were detected by FTIR.

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7897
Table 1. Enthalpies of Binding (kcal mol-1) Determined by Reaction of Mo(PiPr3)2(CO)3(AdCN) with Several Benzonitriles (Reaction 9)
a Values with all species in toluene solution.
Figure 1. Colors of the Mo(PiPr3)2(CO)3(RCN) complexes studied.
From left to right: R = Ad; 4-NMe2C6H4; 2,4,6-Me3C6H2; C6H5; 2,6-
F2C6H3; 4-CF3C6H4; F5C6.
that its mild σ-donor properties are overwhelmed. As is
the case with most low-valent Mo and W complexes,
derivatives with the same ligand produced nearly identi-
cal crystal structures.
In agreement with the solution calorimetric data, the R
group of RCN did not introduce large differences in the
length of the CtN bond. The largest differences were
Figure 2. ORTEP diagram of W(PiPr3)2(CO)3(2,4,6-Me3C6H2CN)
showing 35% probability thermal ellipsoids. Hydrogenatoms are omitted
for clarity. Selected bond lengths and angles are given in Tables 2 and 3.
i
˚
observed between W(P Pr3)2(CO)3(AdCN) (1.153(6) A)
i
and Mo(P Pr3)2(CO)3(AdCN) (1.140(4) A). Both values
˚
˚
are similar to those of 1.153(6) and 1.158(6) A for trans-
(2,4,6-Me3C6H2CN)2Mo(N[iPr]Ar)3,36 and 1.140(9) A
˚
˚
˚
1.140(4) A [Mo] and 1.153(6) [W]) and 1.141 A for 2,6-
fac-W(dppm)(CO)3(CH3CN).37
F2C6H3CN39 (bound CtN distance: 1.145(9) A [W]). In
˚
Comparison of the structures of bound nitriles to the
free nitriles in the condensed phase reveals only a minimal
lengthening of the CtN bond. X-ray crystal structures
the case of 2,4,6-Me3C6H2CN, the CtN bond length is
1.160 A.40 This is significantly longer than the CtN bond
˚
˚
length to 1.145(6) A [W] in the bound complex. The
38
˚
report 1.139(4) A for AdCN (bound CtN distances:
reason for this apparent strengthening of the CtN bond
is unknown and is contrary to the FTIR spectral data. It
has been noted that for the μ-(N2)[W(CO)3(PiPr3)2]2
(36) Tsai, Y.-C.; Stephens, F. H.; Meyer, K.; Mendiratta, A.; Gheorghiu,
M. D.; Cummins, C. C. Organometallics 2003, 22, 2902.
(37) Darensbourg, D. J.; Zalewski, D. J.; Plepys, C.; Campana, C. Inorg.
Chem. 1987, 26, 3727.
(39) Britton, D. Acta Crystallogr., Sect. E 2004, 60, 2189.
(40) Britton, D. Cryst. Struct. Commun. 1979, 8, 667.
(38) Gibbons, C. S.; Trotter, J. Can. J. Chem. 1973, 51, 87.

7898 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
with nitrile adducts are consistent with literature values:
37
2.190(5) A in fac-W(dppm)(CO)3(CH3CN), 2.227(12)
˚
˚
and 2.204(12) A for two separate forms of W(CO)5-
(AdCN),42 and 2.21 A in W(CO)3(CH3CN)3.
43
˚
The most remarkable feature of W(PiPr3)2(CO)3-
(Me2NCN) is that the amide N exhibits trigonal planar
geometry. This behavior has been previously reported for
other complexes such as Cr(CO)5(NCNEt2),44 trans-[Fe-
(Et2PCH2CH2PEt2)2(NCNEt2)2][BF4]2,45 and trans-Mo-
(Ph2PCH2CH2PPh2)2(NCNEt2)(N2).46 This geometric
preference for trigonal planar geometry of the amide is
attributed to delocalization of the amide lone pair of
electrons and significant contribution of the sp2 hybri-
dized N resonance form as shown in reaction 12.
Figure 3. ORTEP diagram of W(PiPr3)2(CO)3(2,6-Me2pz) showing
35% probability thermal ellipsoids. Hydrogen atoms are omitted for
clarity. Selected bond lengths and angles are given in Tables 2 and 3.
Structures of the AdNC complexes are shown in the
Supporting Information, Figures SF-3 and SF-4. As
˚
expected, the M-C bonds (2.139(3) A in the Mo complex;
2.113(6) A in the W complex) are much shorter than the
˚
˚
˚
M-N (2.220(3) A in the Mo complex; 2.195(4) A in the W
complex) bond of the nitrile complexes. This reflects
increased σ-basicity and π-acidity of the isonitrile. Car-
bon is a poorer electrophile than N, and thus a stronger
electron σ-donor. For a similar reason the π* lobes
centered on the C of the isonitrile are larger than those
of the analogous π* lobes centered on N of the nitrile.
This results in a more favorable overlap with the metal d
orbitals in the isonitrile compared to the nitrile and thus
stronger backbonding. Both the W-C bond and the NtC
bond are consistent with similar literature values: CpW
(CO)3(μ-PPh2)W(CO)4(iPrNC)47 (R(W-C)=2.144(11) A,
˚
Figure 4. ORTEPdiagramofW(PiPr3)2(CO)3(Me2NCN)showing35%
probability thermal ellipsoids. Hydrogen atoms are omitted for clarity.
Selected bond lengths and angles are given in Tables 2 and 3. In spite of
repeated attempts, we were not able to grow crystals of μ-(C4H4N2)[M
(PiPr3)2(CO)3]2 (M = Mo, W). The structure of the bridging complex μ-
(N2)[W(PiPr3)2(CO)3]2 has been reported.35 Selected experimental and
computed (see later section) bond distances and angles of X-ray crystal
structures that were determined are listed in Tables 2 and 3.
1
48
˚
R(NtC)=1.139(13) A), cis-W(η -dppm)(CO)4(C6H5NC)
˚
˚
(R(W-C)=2.133(8) A, R(NtC)=1.117(10) A), Mo(CO)3-
49
(tBuNC)3 (R(Mo-C)=2.140(4), 2.148(4), and 2.156(4) A,
˚
˚
R(NtC)=1.159(5), 1.162(5), and 1.146(6) A).
It is also worth noting that Tutt and Zink have
reported50 a W-N bond length of 2.26(1) A for the
˚
structure of W(CO)5(C5H5N) which is nearly identical
to the W-N distance in the 2,6-Me2pz structure. A single
crystal X-ray structure has also been reported for the free
2,6-Me2pz ligand.51 Binding of the 2,6-Me3N2C4H2 does
not significantly alter the geometry of the ligand.
˚
complex the NtN distance is 1.136(6) A, suggesting that
electron donation into π*ΝΝ is minimal.41 This was
explained in terms of resonance structures in which the
W-NtN-W (not WtN-NtW) dominated. For W-
NtC-R, no stable resonance structure can be drawn.
Thus, it may be that backbonding is hampered by a lack
of resonance within the ligand itself.
(42) Jefford, V. J.; Schriver, M. J.; Zaworotko, M. J. Can. J. Chem. 1996,
74, 107.
(43) Hamilton, E. J. M.; Smith, D. E.; Welch, A. J. Acta. Crystallogr.,
Sect.C: Cryst. Struct. Commun. 1987, 43, 1214.
Steric interactions with the phosphines seem not to play
a large role in nitrile binding. The CN may be thought of
as a “spacer” between two bulky groups. In fact, the
shortest W-N bond lengths were observed for benzoni-
triles with electron-withdrawing fluorine substituents in
the ortho position. The W-N distance was 2.130(6) A for
2,6-F2C6H2CN and 2.166(4) A for 2,4,6-Me3C6H2CN,
˚
but 2.192(5) A for Me2NCN. This is consistent with the
(44) Ohnet, M. -N.; Spasojevic-de Brie, A.; Dao, N. Q.; Schweiss, P.;
Braden, M.; Fischer, H.; Reindl, D. J. Chem. Soc., Dalton Trans. 1995, 665.
ꢀ
(45) Martins, L. M. R. D. S.; Frausto da Silva, J. J. R.; Pombeiro, A. J. L.;
Henderson, R. A.; Evans, D. J.; Benetollo, F.; Bombieri, G.; Michelin, R. A.
Inorg. Chim. Acta 1999, 291, 39.
(46) Cunha, S. M. P. R. M.; Guedes da Silva, M. F. C.; Pombeiro, A. J. L.
Inorg. Chem. 2003, 42, 2157.
(47) Shyu, S.-G.; Singh, R.; Su, C.-J; Lin, K.-J. Eur. J. Inorg. Chem. 2002,
1343.
˚
˚
backbonding being enhanced by the electron-withdraw-
ing fluorines. The M-N bond distances for the complexes
(48) Knorr, M.; Jourdain, I.; Lentz, D.; Willemsen, S.; Strohmann, C. J.
Organomet. Chem. 2003, 684, 216.
(49) Imhof, W.; Halbauer, K.; Donnecke, D.; Gorls, H. Acta Crystallogr.,
Sect. E 2006, 62, 462.
(50) Tutt, L.; Zink, Z. I. J. Am. Chem. Soc. 1986, 108, 5830.
(51) Kaiser-Morris, E.; Cousson, A.; Paulus, W.; Fillaux, F. Acta.
Crystallogr., Sect. E 2001, 57, 1116.
(41) Butts, M. D.; Bryan, J. C.; Lou, X.-L.; Kubas, G. J. Inorg. Chem.
1997, 36, 3341.

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7899
i
˚
Table 2. Selected M(P Pr3)2(CO)3L Experimental and Computed (in italics) Bond Distances (A)
M1-N1
N1-C49
Mo(PiPr3)2CO3
C11-O11
C12-O12
C13-O13
AdNC
AdCN
2.139(3),a 2.115a
2.220(3), 2.246
1.158(3), 1.193
1.140(4), 1.176
1.147(4), 1.193
1.147(4), 1.204
1.139(3), 1.196
1.144(4), 1.197
1.162(4), 1.204
1.170(4), 1.210
W(PiPr3)2CO3
AdNC
AdCN
2,4,6-Me3C6H2CN
Me2NCN
2,6-F2C6H3CN
2,6-Me2pz
2.113(6),a 2.117a
2.195(4), 2.177
2.166(4), 2.133
2.192(5), 2.185
2.130(6)
1.171(7), 1.195
1.153(6), 1.178
1.145(6), 1.185
1.151(7),b 1.187
1.145(9)
1.159(6), 1.200
1.130(7), 1.202
1.150(6), 1.200
1.178(6),c 1.203
1.169(10)
1.155(6), 1.199
1.177(8), 1.200
1.168(6), 1.201
1.178(6),c 1.201
1.146(11)
1.167(7), 1.207
1.166(8), 1.213
1.172(6), 1.210
1.165(5),d 1.215
1.170(10)
2.257(4)
n/a
1.155(6)
1.158(6)
1.169(6)
a The ligands in these cases are isonitriles and the bond distance corresponds to M1-C49. The atoms referred to in these cases correspond to the
following. b N1-C12. c C11-O2. d C10-O1.
Table 3. Selected Experimental and Computed (in italics) Bond Angles (deg) for M(PiPr3)2(CO)3L Complexes.a
experiment and computation values
M1-N1-C49
P1-M1-P2
N1-M1-P1 N1-M1-P2
C11-M1-C12
N1-M1-C11 N1-M1-C12
Mo(PiPr3)2CO3
AdNC
AdCN
178.0(2),b 176.9
178.4(3), 177.8
174.30(2), 176.5
173.60(3), 174.8
89.05(7),c 91.6
85.33(7),c 85.1
88.43(7), 90.3
85.24(7), 84.8
170.3(1), 172.8
166.8(2), 171.9
94.7(1), 93.1
94.9(1), 92.4
96.4(1), 94.0
96.8(1), 93.2
W(PiPr3)2CO3
AdNC
178.1(5),b 177.8
176.9(5), 178.0
175.1(4), 176.7
180.0(2),d 178.1
177.6(7)
173.71(5), 175.9
173.06(5), 174.2
182.92(4), 178.4
172.25(4),e174.9
174.64(9)
88.2(2),c 91.1
86.1(2),c 85.1
86.0(1), 84.4
87.2(1), 90.2
87.5(1), 86.1
94.9(1), 95.4
86.13(2),f 84.6
86.13(2),f 90.4
89.2(2)
170.9(2), 172.8
167.8(3), 171.3
172.6(2), 172.1
166.3(2)g170.7
170.7(4)
96.4(2), 93.0
90.6(2), 92.6
98.0(3), 94.6
94.2(2), 93.2
91.5(2), 89.7
86.7(2), 90.0
96.8(1),h 94.1
96.8(1),h 94.2
96.8(3)
AdCN
2,4,6-Me3C6H2CN
Me2NCN
2,6-F2C6H3CN
2,6-Me2pz
86.7(2)
86.4(1)
93.5(1)
90.9(4)
94.4(2)
95.6(2)
n/a
179.52(4)
167.1(2)
a Bond angles shown by italic fonts are from values obtained from B3LYP calculations. b The ligands in these cases are isonitriles and the bond angle
corresponds to M1-C49-N1. c The ligands in these cases are isonitriles and the bond angle corresponds to C49-M1-PX (X = 1,2). The bond angles in
these cases refer to the following. d W1-N1-C12. e P1-W1-P1i. f N1-W1-P1. g C11-W1-C11. h N1-W1-C10.
Kaim et al. have reported52 the structure of [W-
Mo-NX bond is discussed later. The focus on this section
is simply on assessing the correlation of these data.
Selected B3LYP optimized bond distances and angles
are shown together with X-ray single-crystal diffraction
results in Tables 2 and 3. As seen in the data in Tables 2
and 3, there is generally good agreement between com-
puted and measured structures. There appears to be a
tendency for the computed M-N bond distances to be
slightly shorter and all other distances slightly longer than
experimental. However, none of these present any sig-
nificant differences, thus providing evidence that the
computed structures are reliable and that the computed
minimum-energy configurations are essentially correct.
The complexes that exhibited the largest structural
difference between the gas phase and PCM-toluene solu-
tion were those of AdN3, so we selected those complexes
to investigate whether the computed gas-phase BSSEs
were adequate for the solvated structures. This was
done by recomputing the electronic energy of the
(PCy3)2(CO)3(N-MeC4H4N2)][PF6]. The W-N bond
was reported as 2.101(10) A. This bond length is greatly
˚
reduced from that which was determined in W(PiPr3)2-
˚
(CO)3(2,6-Me2pz) (2.257(4) A). This difference in bond
length can be attributed to the higher degree of π-back-
donation from the metal fragment to the pyrazinium
ligand compared to neutral 2,6-Me2pz because of the
stronger π-acid character of the cationic acceptor.
Computational Results. Computational data generated
in this work can be compared directly to independent
experimental determinations in the three major areas of
structure, energetics, and spectroscopy. The interrela-
tionship of these areas in assessing the nature of the
(35) Butts, M. D.; Bryan, J. C.; Luo, X.-L.; Kubas, G. J. Inorg. Chem.
1997, 36, 3341.
(52) Bruns, W.; Hausen, H.-D.; Kaim, W.; Schulz, A. J. Organomet.
Chem. 1993, 444, 12.

7900 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
Table 4. Experimental and Computational Binding Enthalpies (kcal mol-1) for M(PR3)2(CO)3L
ΔH(PiPr3)b
ΔH(PMe3)c
ΔHlarge(PMe3)d
a
M
L
ΔHexpt
Mo
H2
N2
-2.4
-12.4
6.0
-7.3
-11.2
-16.5
-12.5
-17.9
-20.8
-14.6
-30.3
-17.8
-17.6
-19.6
-19.2
-19.2
-20.7
-20.1
-20.6
-17.9
-5.9
-10.3 ( 0.8
-8.8 ( 1.2
-13.8 ( 0.5
-18.6 ( 1.8f
-11.2 ( 0.4
-29.0 ( 0.3
-16.6 ( 0.4
-16.7 ( 0.6
-17.6 ( 0.4
-16.7 ( 0.6
-16.3 ( 0.6
-17.1 ( 0.6
-16.7 ( 0.6
-16.4 ( 0.6
-15.5 ( 1.8i
n.a.
-18.1
-10.3
-18.6
-22.5
-16.1
-28.9
-17.3
-17.2
-19.1
-18.6
-18.8
-20.5
-19.7
-21.3
-17.6
-7.9
Mo(PR3)2(CO)3N2
N2CHSiMe3
N2CHC(dO)OEt
N3Ad
AdNC
CH3CN
-11.9
-17.1
-10.9
-21.1
-10.3
-9.9
-12.5
-11.5
-12.2
-13.5
-13.5
-15.3
-10.6
11.2
AdCN
C6H5CN
2,4,6-Me3C6H2CN
4-NMe2C6H4CN
4-CF3C6H4CN
2,6-F2C6H3CN
C6F5CN
Me2NCN
Quinuclidine
C5H5N (pyridine)
C4H4N2 (pz) bridged
C4H4N2 (pz)
Mo(PR3)2(CO)3pz
2,6-Me2pz
-17.0 ( 0.4
-34.2 ( 1.6 j
-14.9 ( 0.9
-19.3 ( 2.5
-14.8 ( 0.6
-17.5 ( 0.8k
-8.5
-17.9
-16.7
-8.6
0.3
-8.7
-16.7k
-0.7k
-17.7
-13.5
-17.9
-23.6k
-11.6k
-16.6
-13.0
-16.8
-23.2k
-9.8k
[4-Mepz][PF6]
THF
W
N2
-18.0
1.4
-28.3
-14.7
-17.0
-15.6
-24.3
-16.8
-37.5
-23.1
-25.1
-23.5
-18.7
-13.4
-35.3
-21.5
-23.0
-21.1
W(PR3)2(CO)3N2
AdNC
AdCN
2,4,6-Me3C6H2CN
Me2NCN
a Experimental enthalpy of binding of L to Mo(PiPr3)2(CO)3 with all species in toluene solution unless stated otherwise. b Computational enthalpy of
binding of L to Mo(PiPr3)2(CO)3. c Computational enthalpy of binding of L to M(PMe3)2(CO)3. d Computational enthalpy of binding of L to M
(PMe3)2(CO)3 using the larger basis set. e Based on ligand displacement by AdNC. f Based on Keq with PhCN. g Determined from the reaction of Mo
(PiPr3)2(CO)3 with AdNC. h Determined by reaction of Mo(PiPr3)2(CO)3(AdCN) with benzonitriles. i Based on Keq with AdCN. j Enthalpy
corresponding to the binding of both nitrogens of pyrazine to form the bridging species. k Measured (computed) in THF.
optimized PCM-toluene structure with the bulky adaman-
tyl azide ligand in the gas phase as a single-point
counterpoise calculation. The BSSEs for both the Mo
(PiPr3)2(CO)3(AdN3) and Mo(PMe3)2(CO)3(AdN3) com-
plexes were found to be close to those of their gas-phase
structure counterparts (5.4 vs 5.3 and 4.9 vs 4.5 kcal mol-1
,
respectively).
Experimentally derived (see earlier section) values of
ΔH° for the binding of the various ligands studied to Mo-
(PiPr3)2(CO)3 are compared to BSSE-corrected computed
values, ΔH°BSSE, in Table 4. Also presented in Table 4 are
the computed values of ΔH°BSSE for binding of the same
ligands to Mo(PMe3)2(CO)3 and of selected ligands to
W(PiPr3)2(CO)3 and W(PMe3)2(CO)3. The calculated
binding enthalpies for the complexes with PMe3 ligands
are listed for both the small (LANL2DZ) and large
(MWB28 or MWB60 ECP and basis for the metal center;
6-311G(d,p) 5d for all other elements) basis sets.
Figure 5. Plot of calculated N-binding bond dissociation energies
(= -ΔH°bind) for Mo(PiPr3)2(CO)3 vs experimental values for Mo-
(PiPr3)2(CO)3. The line drawn is not the best fit, but a line with slope =
1 to show that most of the calculated enthalpies of binding are lower than
the experimental values. (Inset) Plot of calculated BDE values for Mo-
(PiPr3)2(CO)3 vs. calculated BDE values for Mo(PMe3)2(CO)3. The line
drawn indicates BDE(iPr) = BDE(Me) - 7 kcal mol-1 (see text).
A plot of the computational versus experimental en-
thalpies of binding to Mo(PiPr3)2(CO)3 (Figure 5) ex-
cludes two points: the N2- and the pyrazine-bridged
complexes, [Mo(PiPr3)2(CO)3]2(N2) and [Mo(PiPr3)2-
(CO)3]2(C4H4N2). These are predicted by calculations to
have local minima that are unstable with respect to
dissociation, in contrast to the experimental fact of their
existence. This can be explained by poorly described van
der Waals interactions controlling the conformations
of the bulky PiPr3 groups, weak bonding with the
N2 and pyrazine bridging ligands, and/or steric effects.
Neglecting these two points, the average deviation of the
calculated from the experimental binding enthalpies is
þ4.0 kcal mol-1, but there is considerable scatter in the
data. As shown in Figure 5, and in Table 4, only for N2 is

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7901
the computed bond dissociation energy (= -ΔH°bind
)
larger than the experimental one, a possible consequence
of overestimation of steric repulsive factors for other
ligands.
It is noteworthy that the six ligands for which the
calculated and experimental values are in best agreement
are N2, N3Ad, N2CHSiMe3, C6F5CN, [4-Me-pz]þ, and
N2CHC(dO)OEt, and the next-best group is 2,6-
F2C6H3CN and 4-CF3C6H4CN. These ligands, except
for N2 which is a very weak σ donor, are all of moderate
σ donor strength and moderate to good π acceptor
strength. Binding in the complexes containing these li-
gands could most likely be described adequately by a
basis set without great flexibility for treating subtleties in
the balance between σ donation and π backbonding.
Computation at the small and large basis set level were
performed for binding of ligands to Mo(PMe3)2(CO)3
and data is collected in Table 4. It is evident from the inset
in Figure 5 that the correlation between calculated BDE
values for Mo(PiPr3)2(CO)3 and calculated values for Mo
(PMe3)2(CO)3 is much better than the calculated and
experimental values. In fact, the correlation between the
calculated values for the two complexes is well repre-
sented by the relation BDE(iPr) = BDE(Me) - 7 kcal
mol-1. This indicates that the intrinsic BDEs are roughly
the same and that the net BDEs differ by the dissociation
Figure 6. Plot of calculated N-binding enthalpies for Mo(PMe3)2(CO)3
using a large basis set vs experimental values for Mo(PiPr3)2(CO)3. The
line drawn is the best fit through the points. Data for the dinuclear
complexes [Mo(PiPr3)2(CO)3]2(C4H4N2) and [Mo(PiPr3)2(CO)3]2(N2) are
not included.
i
energy of the agostic bond in the Pr complex which
is ≈7 kcal mol-1. As shown in Figure 6, there is a better
correlation between experimental data obtained for
Mo(PiPr3)2(CO)3 and computed data for Mo(PMe3)2-
(CO)3 using the expanded basis set. These results indicate
that polarization functions and the triple-ζ basis do,
indeed, better describe the delicate balance between σ
donation and π backbonding in these complexes. The
smaller basis set appears to lack this flexibility for both
i
the Pr and Me complexes. Since Mo(PMe3)2(CO)3 does
)
not have an agostic bond (estimated as ≈7 ( 3 kcal mol-1
this is in keeping with the computed intercept of 5.5 kcal
mol-1 in Figure 6, and the generally higher computed
enthalpies of binding for Mo(PMe3)2(CO)3 as seen in
Table 4. In this case, the experimental values should be
adjusted upward by the agostic bond which we estimate as
being ≈7 kcal mol-1 as shown in the Supporting Infor-
mation, Table ST-1 where absolute BDE values are
estimated.
It should be pointed out that computation of fac-Mo
(PMe3)2(CO)3(L), in which the PMe3 groups are trans to
each other, defies experimental measurement and can only
be determined computationally. The mer-Mo(PMe3)2-
(CO)3(L) isomer, in which the PMe3 groups are cis to each
other, is the thermodynamically more stable isomer for
the smaller phosphine ligand. The strong correlation in
Figure 6 implies that exchanging the relatively small PMe3
ligands for the sterically more demanding PiPr3 ligands
does not cause significant steric repulsion to the bound
ligand set studied here. The main result is to enforce the
fac over the mer geometry in these systems. For bulky
ligands, such as quinuclidine (see Table 4) that observation
is obviated, but for the majority of ligands studied here
steric repulsion appears minor.
Figure 7. Comparison of computed and observed νCO stretching vibra-
tions. Red points, CO stretching vibrations in normal modes that involve
primarly a single CO ligand, “1 CO”, and tend to have the lowest-energy
CO frequencies; blue points, those that primarily involve two CO ligands,
“2 CO”, and tend to have intermediate-energy CO frequencies; and
black points, those that involve all three CO ligands, “3 CO”, and tend to
have the highest-energy CO frequencies. The line indicates νCO(Calc) =
ν
CO(Expt) - 50 cm-1
.
40 kcal mol-1 as determined by Smith.53 It is reasonable
to expect similar enthalpies of binding of the isoelectronic
RNtC and CtO ligands.
Computed data for UV-vis and IR spectra of M;
NtCR and M;CtO are collected with experimental
values in Supporting Information, Table ST-3. As shown
in Figure 7 below, computed data are in good agreement
with the trends in CO stretching frequencies, νCO, but bands
were offset by ≈ 50 cm-1. Computed UV-vis spectra are
shown in Supporting Information, Figure SF-9. The agree-
ment between computed and experimental UV-vis data is
shown in Figure 8. As was the case with vibrational
The estimated value of the absolute binding enthalpy
of AdNC to Mo(PiPr3)2(CO)3 of 36 kcal mol-1 is in the
range of the CtO enthalpy of binding to Mo(CO)5 of
(53) Lewis, K. E.; Golden, D. M.; Smith, G. P. J. Am. Chem. Soc. 1985,
106, 3905.

7902 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
Figure 8. Comparison of computed and observed UV-vis absorption
maxima.
Figure 9. Plot of ΔνNC vibration vs UV-vis absorbance maximum for
complexes of Mo(PiPr3)2(CO)3(NtCAr) (Ar = C6F5, 2,6-F2C6H3, 4-
CF3C6H4, 2,4,6-Me3C6H2, C6H5, 4-NMe2C6H4).
computations, computed UV-vis data are in excellent
relative agreement with experimental observations.
6-Me3C6H2CN < C6H5CN < 2,6-F2C6H3CN ≈ 4-CF3C6H4-
CN < C6F5CN. This order is opposite to what would be
expected for σ-donor ability, as judged by gas-phase proton
affinity (PA). We could not find experimental literature data
for C6F5CN (this is what is in Table 4), so we have computed
the gas-phase PAto be 179.6 kcalmol-1. The computedvalue
for C6H5CN was 200.2 kcal mol-1, which is in reasonable
Discussion
The main goal of this work was to evaluate the range of
bond energies and the nature of the bonding in Mo(PiPr3)
(CO)3(NX). As shown in Table 4, data for W parallels that
for Mo but with a more exothermic enthalpy of binding, as
was also found in previous work.4,29 In spite of the bulky
nature of the phosphine ligands, structural studies show that
for NtC;R, and presumably for NdNdCHR and
NdNdN;R complexes as well, steric repulsion does not
play a determining role because of the presence of a “spacer”
atom. It is likely that only for pyridine, pyrazine, or amine
ligands that steric repulsion becomes important as can be seen
by comparing the structures of W complexes of 2,4,6-
Me3C6H2CN in Figure 3 and dimethyl pyrazine in Figure 4.
There is little steric repulsion evident for the ortho-methyl
groups of the nitrile. It is clear, however, that for the dimethyl
pyrazine complex shown in Figure 3 binding of the N atom
with ortho-methyl groups would result in severe crowding. In
spite of the thermodynamic stability of the μ-bridging form of
pyrazine, no evidence for this was found for dimethyl pyra-
zine. Computation and experiment confirm that quinuclidine
does not bind to Mo(PiPr3)2(CO)3, presumably for steric
reasons since computation predicts exothermic binding to
Mo(PMe3)2(CO)3. In spite of these comments, for the ligands
studied with “spacer” atoms in place, for example, NtC;R,
NdNdCHR, and NtN;NR, it seems likely that repulsion
is no more severe in their binding than it is for NtN itself.
The most surprising result, and one which gives some insight
into the role of π-bonding, is the nearly constant enthalpy of
binding of arene nitriles as shown in Table 1. This is in spite of
a dramatic change in color as shown in Figure 1. Accompany-
ing the color changes and λmax are IR changes in both νCN and
νCO in the complexes, all the while maintaining a nearly
constant enthalpy of binding. There is a reasonable correlation
between changes in the UV-vis and IR peak positions as
shown in Figure 9. The larger the decrease in the nitrile stretch-
ing frequency upon coordination to the metal resulted in a
longer wavelength and a darker color of the metal complexes.
The spectroscopic changes depicted in Figure 9 are
indicative of π-backbonding effects becoming inc-
reasingly important in the order: 4-Me2N-C6H4CN < 2,4,
agreement with the experimental value of 193.9 kcal mol-1 54
.
The difference in PA of benzonitrile and perfluoronitrile is
thus on the order of 20 kcal mol-1 favoring benzonitrile.
Differences in binding enthalpies on that order of magnitude
would not be expected in the mildly exothermic overall
binding to Mo; however, the near constant values shown in
Table 1 indicate that gains in π-backbonding are compen-
sated for by losses in σ donation, indicating a roughly equal
importance for nitrile binding to Mo(PiPR3)2(CO)3.
Since steric factors can also play a role, a further compar-
ison is to examine the binding of 4-NMe2C6H4CN and 4-
CF3C6H4CN as ligands since the group in the para position
should not alter the steric interactions significantly. It was
found that 4-CF3C6H4CN bound more exothermically,
yielding a computed value of ΔH°= -0.8 ( 0.2 kcal mol-1
for the reaction shown in reaction 13.
FTIR spectroscopic studies showed that at 20 °C in toluene
solution, Keq=0.3 ( 0.2 for the reaction shown in reaction 13,
which corresponds to a value of ΔG°=þ0.7 ( 0.3 kcal mol-1
.
This generates an estimated value of ΔS°=-5(2calmol-1 K-1
for the nitrile substitution reaction. A low value for the entropy
of exchange would be expected and is observed. There is an
apparent increase in backbonding to the nitrile ligand since, as
(54) Afeefy, H. Y.; Liebman, J. F.; Stein, S. E. Neutral Thermochemical
Data. In NIST Chemistry WebBook, NIST Standard Reference Database
Number 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards
and Technology: Gaithersburg, MD, June 2005; U.S. Patent 20899 (http://web-
book.nist.gov).

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7903
Scheme 2. Enthalpies (kcal mol-1) of Ligand Substitution Viewed As
Carbene Transfer
Additional evidence for the role of π-backbonding is found
in enthalpies of binding of NtN;X ligands which contain
(in at least one hybrid resonance form) a formal NtN bond.
Experimental binding enthalpies spanned 10 kcal mol-1
:
Mo(PiPr3)2(CO)3(N2) < N2 < N3Ad < N2CHSiMe3 < N2-
CHC(dO)OEt (Table 4). Experimental displacement of N2 is
exothermic by 3.5 kcal mol-1 for N2CHSiMe3 and by 8.3 kcal
mol-1 for N2CHC(dO)OEt. These reactions may be viewed
(from a thermodynamic perspective) as carbene transfers
(Scheme 2).
In this view, the increased electron density on the uncoor-
dinated lone pair of complexed N2 would make it more basic
and form a better bond to an electrophillic carbene relative to
free N2. Interestingly, the CO stretch of N2CHC(dO)OEt
exhibits a bathochromic shift upon coordination to the metal
fragment. This is in keeping with increased single-bond
character of the CO bond as a result of the resonance form
on the right shown in reaction 14. The somewhat surprising
delocalization from Mo to the carbonyl oxygen of N2CHC
(dO)OEt is also apparent in the HOMO shown in the
Supporting Information, Figure SF-8. This type of delocali-
zation may beresponsible for thesurprisingly high stability of
this bound carbene.
Figure 10. Selected HOMO (left) and LUMO (right) orbitals for Mo
(PiPr3)2(CO)3(NtC-R) Complexes (R = CH3 (a and b), C6H5 (c and d),
and C6F5 (e and f). HOMO and LUMO orbitals for all complexes can be
viewed in the Supporting Information, Figure SF-8.
shown in the Supporting Information, Table ST-2, the differ-
ence between coordinated nitrile and free nitrile changes from
10 cm-1 for the 4-NMe2C6H4CN ligand to 45 cm-1 for the
4-CF3C6H4CN ligand. Concomitant with that change, how-
ever, there is a rise in average CO frequency of about 10 cm-1. A
simple interpretation would be that the gain in backbonding to
the nitrile is offset by a loss in backbonding to the CO ligand.
The nature of the π-bonding in nitriles is more com-
plex than the spectroscopic data shown in Figure 9
imply. Computed highest occupied molecular orbitals
(HOMOs) and lowest unoccupied molecular orbitals
(LUMOs) for the Mo complexes of CH3CN, C6H5CN,
and C6F5CN are shown in Figure 10.
It is more difficult to assess the role of π-bonding in nitriles
since the frontier π* orbital no longer has its dominant lobe
adjacent to the metal. The HOMO to LUMO transition
shown in Figure 10 for NtC-CH3 appears to be largely
localized on the metal. In going from NtC-C6H5 to NtC-
C6F5, the HOMO shows increasing contribution of π*
orbitals of the arene and the LUMO shows increasing
contribution of metal based orbitals. The result is that the
optical transition in Mo(PiPr3)2(CO)3(NtC-C6F5) shows
less charge-transfer character in the sense that the difference
between HOMO and LUMO net electron densities has been
decreased relative to Mo(PiPr3)2(CO)3(NtCCH3). Examina-
tion of nitrile stretching frequencies alone could lead to
underestimation of the extent of π-backbonding nitrile ligand
since part of the backbonding is into the arene antibonding
orbital.
N-heterocyclic Ligands. Quinuclidine does not bind to
the Mo(PiPr3)2(CO)3 complex, but computation says that
it should bind to the less bulky Mo(PMe3)(CO)3 complex
(Table 1). As discussed later, the bonding site at Mo-
(PiPr3)2(CO)3 is amphoteric, and binding of σ-donors
such as quinuclidine should be favorable except for steric
repulsions. Binding of ammonia and primary ammines is
known for Mo(PiPr3)2(CO)3,1 however, these ligands
were not investigated because of the complexity of
H-bonding in these systems. The primary σ-donor ligands
PPh2Me and P(OMe)3 actually form stronger bonds to
W(PCy3)(CO)3 than do any of the N donors investigated
here.29f,29g Pyridine is a stronger σ donor and weaker π
acceptor than are nitriles, but, as shown in the data in
Table 4, binds with equal enthalpy as nitriles. The fact
that the methyl pyrazine cation binds ≈ 2.5 kcal/mol
more strongly than pyrazine is indicative of the greater
π-backbonding ability of the cation. This also shows up
clearly in the computed HOMOs for these ligands as
shown in Supporting Information, Figure SF-8.

7904 Inorganic Chemistry, Vol. 48, No. 16, 2009
Achord et al.
The preferred binding of bridging versus terminal
pyrazine (4.4 kcal mol-1) as opposed to the preferred
binding of N2 as a terminal rather than a bridging ligand
(≈ 1.5 kcal mol-1) warrants additional comment. The
lower stability of the μ-N2 binding architecture may be
due to steric factors which should be alleviated in the
μ-pyrazine complex. Computation does not estimate the
stability of either complex well, but does a somewhat
better job on the Mo(PMe3)2(CO)3 complexes. The rea-
sons for the disparity between experiment and theory in
calculation of the bridging pyrazine stability are not
known but may be due to overestimation of steric repul-
sive terms as discussed earlier. Additional work in this
area with other metal complexes is planned and may add
additional insight.
Comparisons to Gas-Phase Proton Affinities. The gas-
phase proton affinity of N2 is 118 kcal mol-1 compared to
186 kcal mol-1 for CH3CN and 194 kcal mol-1 for
C6H5CN. Taking an average of 190 kcal mol-1 for the
nitriles, the ratio of PAs is 118/190=0.62. In comparing
the estimated absolute bond strengths to Mo(PiPr3)2-
(CO)3 of ≈ 17 kcal mol-1 for N2 and 24 kcal mol-1 for
nitriles (see Supporting Information, Table ST-1), it can
be seen that the bond strength ratio to the Mo complex is
≈ 17/24=0.7. The close agreement of these ratios does
not allow discounting the role of σ-donation for N2, and
implies that the extent of σ donation may be roughly
proportional to the relative proton affinity of N2. In
addition, it should be noted that binding of N2 results
in an increase in the average wavenumber of vibrations of
the CO ligands in Mo(PiPr3)2(CO)3. Thus, in a crude
sense, N2 could be said to be more electron donating than
the agostic bond which it displaces. The largest N2 vibra-
tional shift, ΔνNN=-178 cm-1, occurs for N2 binding to
Mo(PiPr3)2(CO)3. This compares to the largest shift,
ΔνNCR of -68 cm-1, upon coordination by a nitrile for
C6F5CN. As discussed earlier, in backbonding to nitriles
the frequency shift of the nitrile may not reflect the
total situation with respect to backbonding since a sig-
nificant contribution is into the π* orbitals of the arene.
It seems likely that as a fraction of the net bonding, back-
bonding plays a larger role in N2 binding than it does in
C6F5CN binding. Discounting the role of σ-donation in
binding of N2 does not, however, seem warranted for Mo-
(PiPr3)2(CO)3.
strongly than PMe2Ph,4 the W-PMe2Ph bond (which relies
primarily on the σ-donor ability of the phosphine) is never-
theless quite strong in spite of the steric repulsion engendered
by forming W(PCy3)2(CO)3(PMe2Ph). The ligands investi-
gated here, dinitrogen, nitriles, azides, diazoalkanes, and so
on, are also amphoteric in the sense that both σ-donor ability
and π-acceptor character play important roles in the binding.
A range of ≈10 kcal mol-1 was found in the enthalpies of
ligand binding, and good correlation was found between
theory and experiment in terms of structure and spectral
parameters for the complexes involved. A reasonable
correlation between experimental and computational
bond strengths for the NX ligands studied here has been
achieved with the exception of bridging dinuclear struc-
tures, which require further work beyond the scope of this
paper. No simple picture emerges; a blend of factors is
involved in the bonding. The planar nature of the dimethyl
amino group in bound Me2NCN as shown in Figure 4 is an
indication of π-bonding effects as is the significantly
lowered frequency of the carbonyl ligand in N2CHC
(dO)OEt upon coordination. As discussed for the HOMO
and LUMO pictures of nitrile binding shown in Figure 10,
it can be difficult to assess spectroscopically how signifi-
cant π-back bonding is. The near constant enthalpies of
binding of nitrile ligands in spite of a range of spectro-
scopic and color changes show that even as the nature of
the metal-ligand bond is systematically changing, its
interaction energy may remain constant. That result was
surprising, but seems in accord with what has been pre-
viously learned from studies1 of H2 interactions at the
binding site in the complexes M(CO)3(PR3)2, M=Cr, Mo,
W; R=iPr, Cy. Investigation of binding of NX ligands to a
wider range of metal complexes is in progress to see if the
binding trends found here apply to other metal complexes
where the σ acceptor/π donor properties of the metal
complex may not be as closely balanced.
Acknowledgment. The authors dedicate this paper to
Dr. Gregory J. Kubas. Support of this work by the
National Science Foundation (Grant CHE 0615743)
is gratefully acknowledged (C.D.H.). The work at
Brookhaven National Laboratory is funded under
contract DE-AC02-98CH10886 with the U.S. Depart-
ment of Energy and supported by its Division
of Chemical Sciences, Geosciences, & Biosciences, Of-
fice of Basic Energy Sciences. M.T. thanks the Spanish
Ministry of Science and Innovation and the Fulbright
commission under Fellowship FU-2006-1738.
Conclusions
The interaction between ligands and metals depends on the
properties of both. This work investigates that matchup
between NX donor/acceptor ligands and acceptor/donor
properties of M(PiPr3)2(CO)3, M=Mo,W. The amphoteric
nature of the active site at W(PiPr3)2(CO)3 is illustrated by its
most famous reaction: it can accommodate dihydrogen as
either a molecular hydrogen complex or a dihydride. Earlier
work has shown that while W(PCy3)2(CO)3 binds CO more
Supporting Information Available: Supplemental FTIR spec-
tra, computational UV-vis spectra, computational drawings of
HOMO and LUMO orbitals of M(NR)(PiPr3)2(CO)3, XRD
structures, XRD data tables, and CIF files for all single crystal
structure determinations. This material is available free of
charge via the Internet at http://pubs.acs.org.