A contribution to the direct observation of transient phosphanylidene complexes [RP=W(CO)5] (R: Me, Ph): A revisited approach to their electronic structure by UV-photoelectron spectroscopy

Article abstract of DOI:10.1002/ejic.201301033

An original approach involving a coupling in the gas phase between flash vacuum thermolysis (FVT) and UV-photoelectron spectroscopy (UV-PES) allowed the transient terminal electrophilic phosphanylidene complex [MeP=W(CO)₅] to be characterized; this is the first direct observation of this P-Me derivative. This approach also permitted the electronic structure of [PhP=W(CO)₅] to be revisited and confirmed the results obtained with the methylated analogues. In contrast, [p-NC-C₆H₄P=W(CO) ₅] proved to be too reactive to be detected under our experimental conditions. These [RP=W(CO)₅] phosphanylidene complexes (R: Me, Ph) were identified by their ionization potentials, which are real fingerprints . These experimental data, supported by density functional calculations, give an overall electronic cartography of these transient species. For generation in the gas phase of these phosphanylidene complexes, the thermal degradation of two kinds of precursors were investigated. The joint experimental/theoretical approach allowed us to conclude that phosphanorbornadiene complexes are more suitable precursors than phosphirane complexes. Copyright

Full text of DOI:10.1002/ejic.201301033

FULL PAPER
DOI:10.1002/ejic.201301033
A Contribution to the Direct Observation of Transient
Phosphanylidene Complexes [RP=W(CO)5] (R: Me, Ph):
A Revisited Approach to Their Electronic Structure by
UV-Photoelectron Spectroscopy
CLUSTER
ISSUE
Stéphane Labat,[a] Karinne Miqueu,*[a] Jean-Marc Sotiropoulos,[a]
Patrick Baylère,[a] Geneviève Pfister-Guillouzo,[a]
Ngoc Hoa Tran Huy,[b] and François Mathey[b]
Keywords: Electronic structure / Photoelectron spectroscopy / Density functional calculations / Carbene homologues /
Phosphorus
An original approach involving a coupling in the gas phase
between flash vacuum thermolysis (FVT) and UV-photoelec-
tron spectroscopy (UV-PES) allowed the transient terminal
electrophilic phosphanylidene complex [MeP=W(CO)5] to be
characterized; this is the first direct observation of this P–Me
derivative. This approach also permitted the electronic struc-
phosphanylidene complexes (R: Me, Ph) were identified by
their ionization potentials, which are real “fingerprints”.
These experimental data, supported by density functional
calculations, give an overall electronic cartography of these
transient species. For generation in the gas phase of these
phosphanylidene complexes, the thermal degradation of two
ture of [PhP=W(CO)5] to be revisited and confirmed the re- kinds of precursors were investigated. The joint experimen-
sults obtained with the methylated analogues. In contrast, [p-
NC–C6H4P=W(CO)5] proved to be too reactive to be detected
under our experimental conditions. These [RP=W(CO)5]
tal/theoretical approach allowed us to conclude that phos-
phanorbornadiene complexes are more suitable precursors
than phosphirane complexes.
philic complexes have been reviewed.[1j] Many stable nucleo-
philic phosphanylidene complexes have been isolated and
characterized on the basis of their X-ray structures. For in-
stance, in 1987, Lappert et al. synthesized the first nucleo-
philic complex [Cp2W=PMes*] (Mes*: 2,4,6-tBuC6H2).[3]
In contrast, when π-acceptor ligands – especially carbonyl
groups – are used, the charge on the phosphorus is reduced
and the complex behaves as an electrophilic complex. For
a long time, these terminal electrophilic phosphanylidene
complexes were known as short-lived intermediates and
their existence was proved indirectly in solution by trapping
with unsaturated substrates such as alkenes, alkynes and ni-
triles through [1+2] cycloaddition reactions, resulting in the
formation of P-heterocycles.[4] Insertion reactions into O–
H,[4a,5] N–H,[5] C–N,[4g] and C–O[6] bonds were first estab-
lished. More recently, insertion of these electrophilic phos-
phanylidene complexes into C–C,[7] Si–H,[8] C–N,[9] C–X,[10]
Introduction
The chemistry
of
phosphanylidene complexes
[LnM=PR], commonly called phosphinidene complexes, has
been extensively studied over the last thirty years.[1] These
analogues of carbene complexes have attracted significant
attention because they can be used as starting materials in
organophosphorus chemistry. Their electronic structures
parallel those of the nucleophilic Schrock type and the elec-
trophilic Fischer type carbene complexes. They are known
to form complexes with a variety of transition metal and
ligands and the electro/nucleophilicity is strongly influenced
by the latter.[1c,1e,2] With σ-donating ligands on the metal
center – especially pentadienyl ligands – the density on the
phosphorus atom is increased and the complexes become
nucleophilic. Syntheses, properties and reactivity of nucleo-
[a] Université de Pau et des Pays de l’Adour, Institut des Sciences B–H,[11] and C–H[12] bonds have been reported. A number
of precursors, for instance, phosphirane,[13] norbornadi-
Analytiques et de Physico-Chimie pour l’Environnement et les
Matériaux (IPREM) – UMR-CNRS 5254, Equipe Chimie-
ene,[4a] and phosphepine[14] metal complexes are known to
Physique,
Hélioparc 2, Avenue P. Angot, 64053 Pau cedex 09, France
E-mail: karinne.miqueu@univ-pau.fr
generate terminal phosphanylidene complexes in situ
through thermal or photochemical decomposition. The first
example of an electrophilic terminal transition-metal com-
plex phosphanylidene [PhP=W(CO)5] was reported in 1982
by Mathey and Marinetti (Scheme 1);[4a,4b] this was gener-
ated in situ from 7-phosphanorbornadiene complex and
http://iprem-ecp.univ-pau.fr
[b] Division of Chemistry and Biological Chemistry, Nanyang
Technological University,
21 Nanyang Link, Singapore 637371
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201301033.
Eur. J. Inorg. Chem. 2014, 1694–1705
1694
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
trapped with an excess of methanol, 2,3-dimethylbutadiene calculations were performed on [RP=Cr(CO)5] and
or tolane.[4a] The transient [PhP=W(CO)5] complex and its [RP=W(CO)5] complexes[24–27] to gain further insight into
molybdenum analogue [PhP=Mo(CO)5] have also been the effect of substituents on the singlet–triplet gap, their
identified by mass spectrometry after thermolysis at 433 electronic properties, and the bonding situation.[26,27]
and 453 K, respectively, from their corresponding 7-phos- [LnM=PH] complexes (M: groups 4, 5, 6, 8, and 9; Ligand:
phanorbornadiene complex precursors,[4b] suggesting that CO, phosphanyl and Cp) have also been studied, and theo-
in the gas phase these complexes might be convenient gen- retical results have shown that the reactivity and the electro/
erators of terminal phosphanylidene complexes. Thereafter, nucleophilicity of these complexes are influenced by the li-
these precursors have been used to generate new transient gands.[2]
substituted electrophilic phosphanylidene complexes involv-
Electronic structures of phosphanylidene and phosphir-
ing aryl,[4a–4c,5] chloro,[15] fluoro,[16] and amino[13] groups on ene complexes were described in 1985 by using extended-
the phosphorus atom. The reaction occurs in toluene at Huckel theory and ab initio calculations by Pfister-Guil-
383 K or at lower temperature (328 K) with a catalytic louzo et al.,[28] and the effect of the metal center (M: W,
amount of CuCl, and the postulated intermediates have Cr) and ligand on phosphorus has been assessed. The im-
been trapped in situ by ethylenic or acetylenic reagents, af- portant role played by the metal atom in the aromaticity of
fording phosphirane or phosphirene complexes, respec- the system and in the σ delocalization of the phosphirenes
tively.
has been highlighted by an original UV-photoelectron spec-
troscopy (UV-PES)/theoretical approach and this allowed
the observed stabilization of the complexed phosphirene
system to be understood.
Concerning the bonding situation, donor–acceptor inter-
actions of Lewis bases such amine, phosphane, and phos-
phaalkyne with various phosphanylidene complexes carry-
ing a W(CO)5 fragment have been studied by Schoeller et
al.[29] More recently, a joint experimental/theoretical ap-
proach to the study of cationic [M]P=NR2 electrophilic
phosphanylidene complexes of iron, ruthenium and os-
mium has allowed their structures and their bonding situa-
tion to be described.[30] In the same way, bonding situations
in linear and bent cationic electrophilic phosphanylidene
complexes [Cp(CO)2MϵPMe]+ and [Cp(CO)3M=PMe]+
(M: Cr, W, Mo) have been compared to each other by DFT
calculations.[31]
Energy profiles of species involving phosphanylidene
complexes have also been explored by theoretical ap-
proaches. The mechanisms of the copper(I) chloride cata-
lyzed decomposition of 7-phosphanorbornadiene W(CO)5
complex and olefin trapping of the terminal phosphinidene
complex was described by Lammertsma et al.[32] The car-
bon dioxide deoxygenation process using transient model
terminal phosphanylidene (P–Me) tungsten complexes have
Scheme 1. Transient and stable terminal electrophilic phosphanyl-
idene complexes.
Recently, the direct observation of a transient electro- been analyzed by using DFT calculations. The overall reac-
philic phosphanylidene complex has also been reported by tion was found to be exothermic, in agreement with calcu-
the Lammertsma group (Scheme 1). The authors detected lated thermodynamic oxygen transfer.[33]
by gas-phase mass spectrometry (ESI-MS/MS) the reactive
Phosphorus magnetic shielding tensor calculations for
intermediate cationic complex [ArP=W(CO)5]+(OTf)– car- transition-metal compounds such as phosphanylidene were
rying a P-trimethyl ammoniumphenyl substituent. The exis- also investigated by coupled ab initio and DFT approaches
tence of this transient species was demonstrated thanks to and the advantages of each computational method have
its reactivity with propene and 1,4-dioxane, affording the been described.[34]
[2+1] cycloaddition adduct.[17]
With these different results in mind and taking into ac-
Nowadays, few stable cationic electrophilic aminophos- count that only two direct characterizations – by mass spec-
phanylidene complexes have been isolated involving Fe,[18] troscopy – have allowed transient phosphanylidene com-
Co,[19] Mo,[20–22] W, [20,21] Ru,[17] or Os[17] metal centers and plexes, [PhP=W(CO)5] and [ArP=W(CO)5]+, to be gener-
only one example concerns a neutral electrophilic complex ated, we envisaged associating flash vacuum thermolysis
of vanadium prepared by the reaction between (FVT) in the gas phase with UV-photoelectron spec-
Na2CpV(CO)3 with Cl2PNR2 (Scheme 1).[23]
troscopy (UV-PES) to characterize, by their ionization po-
These terminal phosphanylidene complexes have also tentials, even more elusive [RP=W(CO)5] complexes (R:
been the subject of theoretical studies. Ab initio and DFT Me, Ph, p-NC–C6H4). Indeed, UV-PES is the only experi-
Eur. J. Inorg. Chem. 2014, 1694–1705
1695
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
mental technique that provides insight on the energetic po-
Results and Discussion
sitions of all the occupied molecular orbitals of a molecule,
which are real “fingerprints”. Thus, an experimental elec-
tronic cartography of molecules can be obtained and, cou-
pled with appropriate synthetic methods in the gas phase
(FVT or Vacuum Gas Supported Reaction), it is well-suited
to the characterization of transient species such as radicals
or low-coordinated compounds.[35]
Thermal Decomposition of Phosphirane Complexes
We first investigated the thermolysis of phosphirane
complexes 1 and 2 in the gas phase (Scheme 2) and re-
corded the corresponding photoelectronic (PE) spectra in
an attempt to characterize the elusive phosphanylidene
complex [PhP=W(CO)5] (4b) and gain some experimental
insight into its electronic structure.
In our case, two kinds of precursors have been thermo-
lyzed in the gas phase, phosphirane complexes and the well-
known and commonly used 7-phosphanorbornadiene com-
plexes. The reaction was followed in situ by UV-PES. Phos-
phirane complexes are generally too stable to be useful as
very efficient precursors of terminal phosphanylidene com-
plexes except when an amino group is linked to the phos-
phorus atom.[13] However, Mathey and Marinetti have sug-
gested that a three-membered phosphirane complex involv-
ing an aryl group on the phosphorus atom can decompose
between 383 and 423 K.[4d] They confirmed this result by
performing trapping reactions in situ at 423 K with diphen-
ylacetylene in toluene in a sealed tube. The corresponding
phosphirene complex was formed, suggesting that phos-
phirane complexes can be precursors of terminal phos-
phanylidene complexes. For this reason, we selected both
phosphirane complexes 1 and 2 and the 7-phosphanor-
bornadiene complexes 3a–c [R: p-NC–C6H4 (3a), Ph (3b),
Me (3c)] as precursors (Scheme 2) to study their thermal
decomposition and characterize the corresponding phos-
phanylidene complexes 4a–c in the gas phase by UV-PES.
First, PE spectra of the two precursors 1 and 2 were re-
corded at 423 and 383 K (beam temperature: see experi-
mental section), respectively, under a pressure of 10–5 mbar
(Figure 1). The PE spectrum of compound 1 presents a
broad band centered at 7.6 eV, followed by two sharp bands
at 8.6 and 9.2 eV and two other broad and more intense
bands than the first centered at 10.6 and 12.0 eV. The PE
spectrum of compound 2 presents similarities with a first
band centered around 8.0 eV (with two shoulders at 7.6 and
8.4 eV) and two others centered at 9.5 (between 9.2 and
10 eV), followed by a massif around 11.2 eV.
Scheme 2. Thermal decomposition of compounds 1, 2, and 3a–c.
Figure 1. Black spectra correspond to the photoelectronic spectra
of compounds 1 (a) and 2 (b) recorded respectively at 423 and
383 K. Grey spectra correspond to the thermal decomposition of
precursors 1 and 2 recorded at 443 K (partial decomposition).
Main ionizations of the ethylenic compounds, respectively styrene
(a) and dichloroethylene (b) are noted by a star (*).
The assignment and assessment of the nature of the dif-
ferent ionizations in the photoelectronic spectra recorded
after thermal degradation of the precursors was made on
the basis of DFT calculations. A comparison of these latter
results with the experimental data allowed us to gain a
deeper understanding of the thermal decomposition reac-
tion according to the nature of the precursors.
The PE spectra of the two precursors were assigned on
the basis of DFT calculations carried out at the B3LYP/
Eur. J. Inorg. Chem. 2014, 1694–1705
1696
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
LanL2DZ(W), 6-31G*(H,C,N,O,P,Cl) level of theory (see complexes 3a–c (Scheme 2), which are described as more
computational details). In the privileged staggered form of suitable precursors for the generation of phosphanylidene
these phosphirane compounds (see the Supporting Infor- complexes (lower temperature of thermal decomposition
mation, Table S1), the phenyl ring linked to the phosphorus than phosphirane complexes). Monitoring on line by UV-
atom is almost perpendicular to the bisecting plane of the PES of the thermolysis of 3a–c showed that decomposition
P–C bonds (dihedral angle ca. 70°). For compound 1, two was complete. Indeed, for compound 3a, from 373 to 473 K
minima have been found on the potential energy surface at and under a pressure of 10–5 mbar, only the dimethyl 4,5-
almost the same energy (0.1 kcalmol–1), with the two dimethylphthalate (main ionizations at 9 and 10.2 eV) was
phenyl rings in either the cis or trans position. Comparison identified by UV-PES, indicating that the precursor is to-
of the theoretical and experimental results allowed the dif- tally decomposed upon thermolysis. In this case, the phos-
ferent ionizations on the PE spectra to be assigned. Fig- phanylidene complex [p-NC–C6H4P=W(CO)5] (4a) seems
ure S1 (see the Supporting Information), which represents to be too reactive and was not observed under our experi-
the interaction diagram between the parent phenylphos- mental conditions, even after long accumulation times. The
phirane and the W(CO)5 fragments, allows the assignment presence of a black powder in the capillary tube used for
described below to be easily understood. The first band, the experiments is consistent with this result. In the same
centered at 7.6 eV for 1 and 8 eV for 2, corresponds to the way, for compounds 3b and 3c, the first PE spectra recorded
three d ionizations of the metal center [dxy(W), dxz(W), and at 443 and 433 K, respectively, under a pressure of
dyz(W)] without any interaction. The bands centered at 3.5ϫ10–5 mbar correspond to dimethyl 4,5-dimethylphthal-
9.2 eV (1) and 9.5 eV (2) come from the ejection of an elec- ate, which is also in agreement with total decomposition of
tron of the π orbitals of the phenyl ring linked to the phos- the precursors. After several accumulations, new ionization
phorus atom. For compound 2, the π orbitals of the phenyl signals appear in the PE spectra, with, in particular, a first
group linked to the three PCC rings appear at lower energy band at low energies (7.6–7.7 eV) that is consistent with the
(band between 8 and 9 eV) because of the difference of elec- formation of a new species involving a transition-metal cen-
tronegativity between the phosphorus (χ: 2.19) and carbon ter. Figure 2 shows the PE spectra recorded around 430–
(χ: 2.55) atoms.[36] The ionizations between 10.5 and 440 K from complexes 3b and 3c. The grey spectra corre-
11.5 eV are attributed to the σPC and σPW orbitals and to spond to the superposition of the PE spectrum of the new
the chlorine lone pair orbitals.
species generated upon thermolysis of 4b and 4c
At higher temperature (443 K), only partial decomposi- (Scheme 2), and that of dimethyl 4,5-dimethylphthalate
tion of the two precursors occurs. Indeed, the PE spectra (light-grey spectrum). The PE spectrum arising from ther-
of compounds 1 and 2 changed, with the appearance of mal decomposition of 3b (Figure 2, a), presents a first ion-
characteristic bands of styrene and 1,2-dichloroethylene, ization at 7.6 eV, followed by a broad band centered at
respectively. The four first ionizations of styrene appear at 9.2 eV with a shoulder at 8.4 eV and a more intense band
8.47, 8.65, 9.25 and 10.53 eV as fine bands with almost the centered at 10.8 eV.
same intensity.[37] For isodichloroethylene, the two first ion-
These ionizations are similar to those previously ob-
izations appear at 10 and 11.67 eV, the first ionization is served upon partial decomposition of phosphirane com-
a broad band and the second is fine.[37] In Figure 1, new plexes 1 and 2. The PE spectrum recorded at 443 K from
ionizations appear at the same position, both in grey spec- 3c (Figure 2, b) is similar to the spectrum previously de-
trum of Figure 1a and b (7.6, 8.4, 9.2 and 10.0 eV), and scribed, with a first band at 7.7 eV, a second, sharper band
could be attributed to the short-lived phosphanylidene at 9.0 eV, and a third at 10.8 eV with a shoulder at 10.2 eV.
complex [PhP=W(CO)5]. Unfortunately, these new ioniza- The main difference between these two spectra is the band
tions are mixed with those of the precursors. Increasing the centered at 9 eV. This latter is broad and presents a shoulder
temperature from 443 to 523 K, in an attempt to obtain at 8.4 eV when the phosphorus atom is substituted by a
total decomposition of compounds 1 and 2, led, more par- phenyl ring, in agreement with the presence of the π2Ph and
Ph
Ph
ticularly, to the disappearance of the first ionization cen- the π1 ionizations of this phenyl group (π1 can interact
tered at 8 eV. Only the ionizations of the 1,2-dicholoroethy- with the phosphanylidene skeleton and not π2Ph).
lene and the styrene were retained, showing that the species
Theoretical calculations and trapping experiments al-
formed during the thermolysis is thermally unstable. Taking lowed us to confirm that the new compounds, generated in
into account these results, it seems that these phosphiranes the ionization chamber of the UV photoelectron spectrome-
are not good precursors with which to characterize unam- ter and characterized for the first time by their ionization
biguously phosphanylidene complexes in gas phase because potentials, correspond to the two short-lived phosphanylid-
their total thermal degradation needs excessively high tem- ene complexes [PhP=W(CO)5] (4b) and [MeP=W(CO)5]
peratures, leading to decomposition of the transient phos- (4c). Indeed, to probe the formation of 4a–c in the gas
phanylidene complex [PhP=W(CO)5].
phase, thermolysis of the precursors 3a–c was also realized
in a vacuum line outside the ionization chamber of the
spectrometer in the presence of trans stilbene. The corre-
sponding phosphirane complexes 5a–c were identified by
Thermal Decomposition of 7-Phosphanorbornadiene
Complexes
To overcome the difficulties described above, we studied 31P NMR spectroscopic analysis. The characteristic chemi-
the thermal decomposition of 7-phosphanorbornadiene cal shifts in CD2Cl2 resonate at –134.9 ppm (5a, R: p-NC–
Eur. J. Inorg. Chem. 2014, 1694–1705
1697
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Information, HOMO–2 in Figure S2-b). Nevertheless, as
previously observed[28] by Pfister-Guillouzo et al. for
[HP=Cr(CO)5], the phosphorus M–P–R bond angle is
greater than 90°. Following the opening of the bond angle
from 90 to ca. 115°, the stabilization of this bent structure
comes from the presence of a secondary stabilizing interac-
tion between the σPC orbital and the empty dz2(W) orbital
and a decrease of the destabilizing interaction between the
two occupied σPC orbitals and the dxz(W) orbital.[38]
Thus, in this bent structure the first ionic state energy,
calculated at respectively 7.48 and 7.76 eV for [PhP=
W(CO)5] (4b) and [MeP=W(CO)5] (4c) (ΔSCF calculations)
corresponds to σ-donation from the occupied p phosphan-
ylidene lone pair [np(P), in σ plane] into the dz2(W) orbital
[weak participation] with also an antibonding interaction
with the dxz(W) orbital (MO2 in Figures 3 and 4). This re-
sult is in agreement with the experimental ionization poten-
tial observed at low energies (band 1 in the PE spectra) in
Figure 2 (7.6 eV for phenyl compound 4b, and 7.7 eV for
methylated compound 4c).
Figures 3 and 4, which correspond to the interaction dia-
gram between the methyl or phenyl-phosphanylidene,
respectively, and the W(CO)5 fragments allow all the other
ionizations of 4b and 4c in the PE spectra of Figure 2 to be
assigned. Thus, for [MeP=W(CO)5] (4c), the second broad
band center at 9 eV can be assigned to the ejection of an
electron from the dxy(W) orbital without interaction with
the phosphanylidene fragment (MO3), the bonding combi-
nation between the occupied p phosphorus lone pair np(P)
and the dxz and dz2 orbitals (MO4) as well as the orbital
corresponding to back-donation from the dyz(W) to the
empty p phosphorus orbital [pπ(P)] of the phosphanylidene
(MO5). Finally, the σPC ionization corresponds to the band
at 10.8 eV (MO6).
Figure 2. Photoelectronic spectra of the thermal decomposition of
compounds 3b (a) and 3c (b) at 443 and 433 K, respectively, under
a pressure of 3.5ϫ10–5 mbar. Dark-grey spectra correspond to
phosphanylidene complexes 4b (a) and 4c (b) spectra superimposed
with that of dimethyl 4,5-dimethylphthalate (light-grey spectrum).
For compound [PhP=W(CO)5] (4b), a similar assignment
can be made (see Figure 4) with, additionally, the ioniza-
Ph
tions of the phenyl ring (π1 and π2Ph), which can either
be with or without interaction with the phosphanylidene
C6H4), –120.4 ppm (5b, R: Ph), and –139.8 ppm (5c, R: complex. The shoulder of the second band in the PE spec-
Me), consistent with the reported results obtained after trum (8.4 eV) can be attributed to the ejection of an elec-
trapping reactions in solution (R: Ph, –120.8 ppm and R: tron from the sole dxy(W) orbital (MO3), the bonding com-
Me, –142.1 ppm in C6D6).[4c]
bination between the p phosphanylidene lone pair np(P)
DFT calculations allowed all the ionizations in the PE and the dxz and dz2 orbitals (MO5) as well as the orbital
spectra shown in Figure 2 to be assigned and confirmed corresponding to back-donation from the dyz(W) orbital to
the formation of complexes 4b and 4c. In agreement with the empty pπ(P) orbital of the phosphanylidene in anti-
Ph
previous theoretical studies,[26,29] the singlet state was found bonding combination with the π1 orbital of the phenyl
to be the lowest energy state, with only a weak energy dif- ring (MO4). The ionization of the π2Ph orbital of the phenyl
ference with the triplet state (7.6 kcalmol–1 for 4a, (MO6) appears at 9.2 eV and at the bottom of the band
3.5 kcalmol–1 for 4b, and 7.7 kcalmol–1 for 4c). For this sin- centered at 10.8 eV the bonding combination between the
Ph
glet state, the privileged staggered conformation presents a pπ(P) vacant of the phosphanylidene and the π1 orbital
W–P–R plane that bisects the P–W–CO plane (RPCCCO
:
of the phenyl ring (MO7). In 4b, the presence in the band
Ph
41.4° for 4a, 44° for 4b, and 46.5° for 4c) and the W–P–R centered at 9.2 eV of the ionization of the π2 orbital of
bond angle is 117.48° for 4a, 117.73° for 4b, and 111.74° the phenyl ring is consistent with a broader band compared
for 4c (bent structure, see the Supporting Information, to that in the methylated compound 4c. Moreover, it is
Table S2). In a first approach, it is expected that the privi- noteworthy that the ionization corresponding to the back-
Ph
leged arrangement should present an R–P–W bond angle donation dyz(W)Ǟpπ(P), involving the π1 of the phenyl
close to 90°, because it would lead to the maximal p phos- ring, which is almost coplanar with the R–P–W plane (dihe-
phorus lone pair np(P)/dz2(W) overlap (see the Supporting dral angle: 1.4°), appears at lower energy (8.4 eV in 4b vs.
Eur. J. Inorg. Chem. 2014, 1694–1705
1698
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 3. Assignment of the PE spectrum of 4c using the interaction diagram between the methylphosphanylidene and tungsten hexacarb-
onyl fragments. Kohn–Sham values (in eV) coming from DFT calculations. Calculated ionization energies (in eV) in bold. Main interac-
tions (- - -) and secondary interactions (······) involved in the interaction diagram.
8.8 eV in 4c) because of phenyl ring delocalization into the of 0.3 to 0.4 e– is corroborated by the presence of stabilizing
pπ(P) orbital (antibonding contribution). The bonding interactions between this πP=W natural orbital and the
component [pπ(P) + π1Ph] is in the shoulder of the band WCO5 fragment (mainly π*C=O orbital).
centered at 10.8 eV. Finally, the band centered at 10.8 eV
Table 1. NBO analysis for the πP=W orbital in 4a–c. Participation
of each atom in the bond (%P, %W), hybridization of each atom
(%s, p and d) and occupation (occ.) in electron.
can be assigned to the ejection of an electron from the σPC
orbital (MO8).
When we compare the energetic positions of the dxy(W)
orbital without any interaction (8.4 eV) with the d orbital
of tungsten involved in the back-donation [dyz(W)Ǟpπ(P)] –
bands at 9 eV for the methylated compound and 8.4/10.1 eV
for the phenyl compound (interaction with π1Ph) – the back-
donation is relatively important (Ͼ 0.6 eV). This is in agree-
ment with the Charge Decomposition Analysis[39] calcu-
lated with the CDA 2.2 program, for which a back-donation
P
W
%s
%P
%W occ.
%p %d
%p %d
4a 25.12 74.88 1.65
4b 23.34 76.66 1.64
4c 25.28 74.72 1.66
99.61 0.39
99.60 0.40
99.54 0.46
0.0
0.0
0.01 99.99
0.01 99.98
0.12 1.51 98.37
Analysis of the energetic position of the LUMO (Fig-
of 0.28 electrons is found for these phosphanylidene com- ure 5) as well as the Fukui indices (global electrophilicity
plexes and with the Natural Bond Orbital (NBO)[40] analy- ω, condensed Fukui index fk+) clarifies why the phosphan-
sis, which localizes at the first order a πP=W bond between ylidene complex 4a could not be characterized by UV-PES
the phosphorus and tungsten atoms (Table 1). The latter, under our experimental conditions. The global electrophi-
which corresponds to the interaction between the dyz(W) licity ω of a molecule has been defined by Parr et al. in
orbital (98.4 to 100% d) of the tungsten atom and the pπ(P) terms of electronic potential μ and chemical hardness η (see
orbital (99.5 to 99.6% p) of the phosphorus atom, is computational details);[41] a high value of ω means a high
strongly localized on the metal center (from 75 to 77%) and electrophilicity. According to the theoretical results detailed
presents an occupation of 1.6 to 1.7 e–. This delocalization in Table 2, 4a (ω ≈ 2.7) is more electrophilic than 4b and 4c,
Eur. J. Inorg. Chem. 2014, 1694–1705
1699
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 4. Assignment of the PE spectrum of 4b using the interaction diagram between the phenylphosphanylidene and tungsten hexacarb-
onyl fragments. Kohn–Sham values (in eV) coming from DFT calculations. Calculated ionization energies (in eV) in bold. Main interac-
tions (- - -) and secondary interactions (······) involved in the interaction diagram.
which present similar electrophilicities (ω ≈ 2.1). Moreover, quence, this is the more electrophilic reactive site. Looking
these phosphanylidene complexes have the largest value of at the LUMO of 4a–c (Figure 5), which corresponds to the
+
the condensed Fukui[42] index fk (local electrophilicity) at antibonding combination between the 3pπ(P) orbital and
the phosphorus atom (Table 2: 0.25 to 0.27). In conse-
Table 2. Global and local properties for 4a–c.[a]
Global properties
Local property
+
I
A
μ
η
ω
k
fk
4a
4b
4c
7.72 2.65 –5.18 5.07 2.65
7.48 2.02 –4.76 5.48 2.07
7.76 2.02 –4.89 5.74 2.08
P
W
C
0.2526
0.1921
–0.0137
P
W
C
0.2602
0.1867
–0.0021
P
W
C
0.2786
0.2068
0.0055
[a] Vertical ionization potential, I (eV); Electron affinity, A (eV);
Electronic chemical potential, μ (eV); Chemical hardness, η (eV);
+
Global electrophilicity, ω (eV); Local electrophilicity fk of atom k
(P, W and C linked to phosphorus atom) evaluated from NPA
charges (NBO calculation).
Figure 5. Plot (cutoff: 0.05) and energetic positions of the LUMO
in phosphanylidene complexes 4a–c.
Eur. J. Inorg. Chem. 2014, 1694–1705
1700
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
the dyz(W) orbital – with a strong character of 3pπ(P) – this (see Figure 1). In contrast, for complexes 3a–c the reaction
LUMO is more accessible in 4a (–4.1 eV) than in 4b is exergonic (ΔG: –9.3 kcalmol–1 for 4a, –19.6 kcalmol–1 for
(–3.6 eV) and 4c (–3.7 eV).[43] Consequently, this difference 4b, and –13.8 kcalmol–1 for 4c; ΔH: –4.1 kcalmol–1 for 4a,
around 0.5 eV in the LUMO as well as the important global –4.9 kcalmol–1 for 4b, and –1.9 kcalmol–1 for 4c) and the
electrophilicity of 4a can be the reason why, at high tem- activation barriers are around 21–24 kcalmol–1 (ΔG#: from
perature, this phosphanylidene complex can easily recom- 21.4 kcalmol–1 for 4a to 22.9 kcalmol–1 for 4c; ΔH#: from
bine or dimerize, even under gas-phase conditions.
22.8 kcalmol–1 for 4a to 24.3 kcalmol–1 for 4c) lower than
in phosphirane precursor 2. The decomposition reaction
also occurs with a concerted pathway and a nonsymmetric
TS (see Table 3: PC1 ≈ 2.1–2.2 Å and PC2 ≈ 2.8–2.9 Å in
TS vs. PC1: 1.91 Å and PC2: 1.92 Å in 3a–c). These results
are consistent with those previously described by Lam-
mertsma et al.,[32] on a parent complex (H on phosphorus
and benzene instead of dimethylphthalate; ΔG#:
Theoretical Studies of the Thermal Decomposition of
Phosphiranes and 7-Phosphanorbornadiene Complexes
To gain more insight into the thermal decomposition of
phosphirane and 7-phosphanorbornadiene complexes,
DFT calculations were carried out. Figure 6 details the
energetic profiles leading to the phosphanylidene complexes
from phosphirane complex 2 and 7-phosphanorbornadiene
complexes 3a–c. For 2, the reaction is predicted to be ender-
gonic (ΔG +12.0 kcalmol–1; ΔH +25.9 kcalmol–1). The en-
ergy barrier for this concerted mechanism involving a non-
symmetric transition state (see Table 3: PC1: 2.414 Å and
PC2: 3.030 Å in TS1) is calculated to be around
27 kcalmol–1 (ΔG#: 26.9 kcalmol–1; ΔH#: 27.5 kcalmol–1).
Consequently, the reaction can be reversible, which is in
agreement with the experimental results because the photo-
electronic spectrum recorded during the thermic degrada-
tion of the phosphirane complexes corresponds to a mix-
ture of the precursor and the phosphanylidene complex 4b
Table 3. Geometrical parameters (bond lengths in Å, bond and di-
hedral angles in °) for the complexes 2, 3a–c and their correspond-
ing transition states (TS).
P–C1
P–C2
P=W
R–P
RPW
RPWCCO
2
1.882
1.912
1.912
1.914
1.862
1.917
1.918
1.916
3.030
2.789
2.809
2.901
2.553
2.540
2.551
2.556
2.052
2.500
2.513
2.507
1.828
1.844
1.839
1.850
1.830
1.846
1.831
1.861
122.03
112.88
113.29
112.38
116.94
111.92
112.79
111.77
–65.5
–119.1
–59.7
–60.0
126.0
–111.5
–112.2
–111.2
3a
3b
3c
TS2 2.414
TS3a 2.183
TS3b 2.079
TS3c 2.169
Figure 6. Energy profile (ΔG or ΔH values in kcalmol–1) for the thermal decomposition of phosphirane complex 2 and 7-phosphanor-
bornadiene complexes 3a–c [R: p-NC–C6H4 (a), Ph (b), Me (c)]; ΔH values given in parentheses. For reactions involving TS2 and TS3a–c
the number in superscript, respectively 2 and 3a–c, corresponds to the number of the precursor.
,
Eur. J. Inorg. Chem. 2014, 1694–1705
1701
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
19.3 kcalmol–1 and ΔG: –8.3 kcalmol–1). These theoretical the π1Ph orbital of the aryl substituent). This back-donation
results show that the substituent on the phosphorus atom is relatively important because a difference of around 0.6 eV
has no significant influence on the kinetics of the reaction is observed with the pure dxy(W) orbital.
(similar barrier) and presents a more pronounced effect on
the thermodynamic stability of the final products.
The joint experimental/theoretical approach allows fur-
ther insight onto the thermal decomposition of two kinds
It is noteworthy that the decomplexation process from of precursors involved in this study to be gained. As pre-
3a–c cannot compete with the dimethyl 4,5-dimethyl- viously observed in solution in many studies, the phosphir-
phthalate elimination because the reaction is strongly endo- ane complexes are less suitable precursors in the gas phase
thermic [ΔH(decomplexation) ≈ 40 kcalmol–1 vs. ΔH#3a–c than the phosphanorbornadiene complexes. Indeed, the ac-
≈ 23 kcalmol–1].[44]
tivation barriers of their thermal decomposition are slightly
With these different theoretical results in mind, it is easily higher (27 vs. 23 kcalmol–1) and, especially, the generation
understandable why phosphanorbornadiene complexes 3a– of the transient phosphanylidene complexes 4 is strongly
c are more suitable precursors than the phosphirane com- endergonic, involving a reversible reaction (mixture of ini-
plexes for generation in the gas phase of terminal phos- tial reactants 1 or 2 and phosphanylidene complex 4b in
phanylidene complexes 4a–c. Indeed, for the phosphanor- the PE spectra). In contrast, from phosphanorbornadiene
bornadiene complexes, the loss of dimethyl 4,5-dimethyl- complexes 3, dimethyl 4,5-dimethylphthalate elimination is
phatalate can easily occur under our experimental condi- complete and leads exclusively to the transient species 4b–c
tions (thermal degradation at around 453 K in the gas (exergonic reaction).
phase) because the activation barriers are calculated to be
around 21–24 kcalmol–1. Moreover, formation of the short-
lived complexes 4a–c is thermodynamically favored (exer-
Experimental Section
gonic reaction compared with 2). However, this reaction is
General: All reactions and manipulations were carried out under
less exergonic with 4a than with 4b and 4c. Thus, this obser-
an atmosphere of dry argon using standard Schlenk techniques.
vation together with a more pronounced electrophilicity of
the [p-NC–C6H4P=W(CO)5] complex (4a) and a more ac-
cessible LUMO accounts for our inability to identify this
transient species by UV-PES.
Phosphirane complexes 1[4c] and 2,[10d] and 7-phosphanorbornadi-
ene complexes 3a–c[4b] were prepared as previously reported.
For the trapping experiments of phosphanylidene complexes 4a–
c with trans stilbene, the NMR 31P spectra of the corresponding
phosphirane complexes 5a–c were recorded with a Bruker Avance
400 MHz instrument. Chemical shifts are expressed in parts per
Conclusion
million, relative to external 80% H3PO4 solvent signal.
The photoelectron spectra were recorded with a homemade
The transient electrophilic phosphanylidene complex
“PhenixI” instrument equipped with 127° cylindrical analyzer and
[MeP=W(CO)5] has been characterized for the first time in
monitored by a microcomputer supplemented with a digital-analog
the gas phase by its ionization potentials using the coupling
of FVT and UV-PES techniques. At the same time, the elec-
converter. This equipment was specially designed to characterize
transient species. In practice, we increased the ionization chamber
vacuum to 10–6 mbar (all connections are reduced to the necessary
and joints are adapted), which led to better detection and signal
strength. The spectra were calibrated with the known autoioniza-
tion of helium at 4.99 eV [He(II)/He(I)] and xenon ionizations at
12.13 and 13.44 eV. All the products were thermally decomposed
as a result of the beam heat (heating from 298–523 K), which was
embedded into the target chamber. Consequently, the different tem-
peratures presented in this publication correspond to the beam
temperature and not to the sample temperature. To obtain stable
and repeatable spectra, a stable temperature of the beam and a
stable pressure in the system was maintained.
tronic structure of its aryl analogue [PhP=W(CO)5] has
been revisited by using this original experimental approach.
Unfortunately, substitution in the para-position of the
phenyl group by a cyano substituent (p-NC–C6H4) does not
allow the identification of the corresponding terminal phos-
phanylidene complex. This result is in agreement with a
more pronounced electrophilic character of 4a than 4b or
4c and a more accessible LUMO involving dimerization or
polymerization of 4a, even in the gas phase. The characteri-
zation of [MeP=W(CO)5] and [PhP=W(CO)5] complexes
have been corroborated by trapping reactions in the gas
phase with trans-stilbene, and by DFT calculations that al-
Computational Section
low the different ionization potentials in the PE spectra to Calculations were carried out with the Gaussian 03 program[45] on
the real compounds 1–4 at the DFT level of theory by using the
hybrid functional B3LYP,[46] which is a three-parameter functional
developed by Becke that combines the Becke gradient-corrected ex-
change functional and the Lee–Yang–Parr and Vosko–Wilk–Nusair
correlation functionals with part of exact HF exchange energy. The
tungsten atom was treated with the LanL2DZ pseudo-potential in
combination with its adapted basis set.[47] All the other atoms (C,
P, H, P, O, N, Cl) were described with a 6-31G(d) double-ζ basis
set.[48] Geometry optimizations were carried out without any sym-
metry restrictions, the nature of the extrema (minima or TS) was
be assigned. The transient phosphanylidene complexes 4b–
c present especially a first ionization potential at 7.6–7.7 eV
corresponding to σ-donation from the p phosphanylidene
lone pair [np(P)] into the dz2 orbital of the metal center with
also an antibonding interaction with the dxz(W) orbital
[nP(P) – dxz(W) + dz2(W)]. This low ionization energy is
consistent with a high reactivity of these short-lived species.
The ionization corresponding to the back-donation
dyzǞpπ(P) appears at 9 eV for the methylated compound
and 8.4/10.1 eV for the phenyl compound (interaction with verified with analytical frequency calculations. All total energies
Eur. J. Inorg. Chem. 2014, 1694–1705
1702
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
and Gibbs free energies were zero-point energy (ZPE) and tempera-
ture corrected by using unscaled density functional frequencies.
The connection between the transition states and the correspond-
ing minima was confirmed by IRC calculations.[49]
tential. The aim was to directly compare the KS energies to the
experimental energies. Moreover, Werstiuk[55] and Rademacher,[56]
developed and successfully applied a routine for the interpretation
of PE spectra based on B3LYP theory. This routine requires the
calculation of the first vertical ionization potential of the molecule
(I; see above). Calculated orbital energies –ε are then shifted uni-
formly so that the HOMO energy equals that of I, giving the higher
ionization potential. Hoffmann[57] underlined that the KS orbitals
are a good basis for qualitative interpretation of molecular orbitals.
He concluded that if we want to go a step beyond a qualitative
interpretation and look at orbital energies as rough ionization po-
tentials and if the DFT calculations are done with commonly used
potentials, then it appears we must take the absolute constant and
linear orbital energy shift into account by applying a suitable ax+b
scaling. In the case of large molecules, the ordering of levels found
using the DFT calculations also seems to be qualitatively well-
translated by using HF methods. This fact, despite the evident lim-
its of correlation, gives a good base from which to discuss the ion-
izations assignments that cannot be reached by direct calculation.
The charge decomposition analysis (CDA) was carried out with the
CDA 2.2 program of Frenking.[39] The orbital contributions to the
charge distributions were divided into four parts: (i) the mixing of
the occupied orbitals of the ligand and the unoccupied MOs of
the metal fragment (LigandǞW donation d), (ii) the mixing of the
unoccupied orbitals of the ligand and the occupied MOs of the
metal fragment (LigandǟW back-donation b), (iii) the mixing of
the occupied orbitals of the ligand and the occupied orbitals of the
metal fragment (ligand ↔W repulsive polarization r), and (iv) the
mixing of the unoccupied orbitals of the ligand and the unoccupied
orbitals of the metal fragment (residual term Δ).
The nature of the P=W bond in the different phosphanylidene
complexes 4a–c was studied by using Natural Bond Orbital (NBO)
analysis (NBO-5 program).[40] The natural charges (NPA charges)
+
involved in the calculation of the local electrophilicity fk (see be-
Supporting Information (see footnote on the first page of this arti-
cle): Theoretical details and Cartesian coordinates for the opti-
mized structures.
low) were also computed with this program.
Plots of the geometrical structures and molecular orbitals were
made with the Molekel (cutoff: 0.05 for MO)[50] or Gaussview[51]
programs.
Acknowledgments
The global electrophilicity, ω, of atoms and molecules was defined
by Parr et al.[41] as: ω = μ2/(2η), with μ being the electronic chemical
potential (the negative of electronegativity) and η the chemical
hardness. The electronic chemical potential μ and the chemical
hardness η were approached by μ = (I + A)/2 and η = (I – A), in
terms of the vertical ionization potential I and electron affinity A.
The term I was evaluated from I (eV) = Ecation – Eneutral where
Eneutral is the total energy of neutral species at an optimized geome-
try and Ecation is the total energy of the corresponding cation calcu-
lated at the optimized geometry of the neutral species. Term A was
evaluated from A (eV) = Eneutral – Eanion, where Eneutral is the total
energy of neutral species at an optimized geometry and Eanion is the
total energy of the corresponding anion calculated at the optimized
geometry of the neutral species.
Financial support from the Centre National de la Recherche Sci-
entifique (CNRS), the Université de Pau et des Pays de l’Adour
(UPPA) and the Ministère de la Recherche is gratefully acknowl-
edged. The authors also acknowledge the EPCP team of IPREM
at UPPA, and especially Abdel Khouk, for the 31P NMR analysis.
Part of the theoretical work was granted access to HPC resources
of IDRIS under allocation 2013 (i2013080045) made by Grand
Equipement National de Calcul Intensif (GENCI).
[1] a) A. H. Cowley, Acc. Chem. Res. 1997, 30, 445–451; b) A. H.
Cowley, A. R. Barron, Acc. Chem. Res. 1988, 21, 81–87; c) F.
Mathey, N. H. T. Huy, A. Marinetti, Helv. Chim. Acta 2001,
84, 2938–2957; d) K. Lammertsma, M. J. M. Vlaar, Eur. J. Org.
Chem. 2002, 1127–1138; e) A. W. Ehlers, E. J. Baerends, K.
Lammertsma, J. Am. Chem. Soc. 2002, 124, 2831–2838; f) R.
Streubel, A. Ostrowski, H. Wilkens, F. Ruthe, J. Jeske, P. G.
Jones, Angew. Chem. Int. Ed. Engl. 1997, 36, 378–381; g) K.
Lammertsma, in: New Aspects in Phosphorus Chemistry III, in:
Topics in Current Chemistry (Ed.: J.-P. Majoral), Springer,
Berlin, Heidelberg, 2003, p. 95–119; h) F. Mathey, Dalton
Trans. 2007, 1861–1868; i) R. Manju, Synlett 2007, 2078–2079;
j) H. Aktas, J. C. Slootweg, K. Lammertsma, Angew. Chem.
2010, 122, 2148; Angew. Chem. Int. Ed. 2010, 49, 2102–2113.
[2] a) G. Frison, F. Mathey, A. Sevin, J. Organomet. Chem. 1998,
570, 225–234; b) A. W. Ehlers, E. J. Baerends, K. Lammertsma,
J. Am. Chem. Soc. 2002, 124, 2831–2838.
[3] a) P. B. Hitchcock, M. F. Lappert, W.-P. Leung, J. Chem. Soc.,
Chem. Commun. 1987, 1282–1283; b) R. Bohra, P. B. Hitch-
cock, M. F. Lappert, W.-P. Leung, Polyhedron 1989, 8, 1884.
[4] a) A. Marinetti, F. Mathey, J. Fischer, A. Mitschler, J. Am.
Chem. Soc. 1982, 104, 4484–4485; b) A. Marinetti, F. Mathey,
J. Fischer, A. Mitschler, J. Chem. Soc., Chem. Commun. 1982,
667–668; c) A. Marinetti, F. Mathey, Organometallics 1984, 3,
456–461; d) A. Marinetti, C. Charrier, F. Mathey, J. Fischer,
Organometallics 1985, 4, 2134–2138; e) K. Lammertsma, P.
Chand, S. W. Yang, J. T. Hung, Organometallics 1988, 7, 1875–
1876; f) J. B. M. Wit, G. T. van Eijkel, J. J. Franciscus, T. J. J.
Kanter, M. Schakel, A. W. Ehlers, M. Lutz, A. L. Spek, K.
Lammertsma, Angew. Chem. 1999, 111, 2716; Angew. Chem.
The Fukui function f(r) is defined as the first derivative of the elec-
tronic density ρ(r) of a system, with respect to the number of elec-
trons N at a constant external potential v(r):[42] f(r) = [Ѩρ(r)/ѨN]v(r)
= [δμ/δv(r)]N. It has been shown that a high value of the Fukui
function f(r) means a high reactivity of the site r.[52] To describe
site selectivity of an atom in a molecule, it is necessary to condense
the values of f(r) around each atomic site into a single value that
characterizes the atom in a molecule. Thus, for an atom k in a
molecule, depending upon the type of electron transfer, three dif-
ferent types of condensed Fukui function of the atom k exist.
+
fk = [qk(N + 1) – qk(N)], for nucleophilic attack
–
fk = [qk(N) – qk(N – 1)], for electrophilic attack
0
fk = 1/2[qk(N + 1) – qk(N – 1)], for radical attack
where qk is the gross population of atom k in the molecule.[42,53]
+
Here, fk was used to help understand the reactivity of 4a–c.
For the assignment of the bands in the PE spectra, the following
method was used. For large molecules, association between ioniza-
tion potential and Kohn–Sham (KS) energies calculated with com-
mon functionals (Koopmans’ theorem like) was largely discussed.
Arduengo and co-workers[54] first used DFT calculations at the
nonlocal level to assign their photoelectron spectra. They applied
a uniform shift to the orbital energies considering the difference
between the calculated and experimental molecular ionization po-
Eur. J. Inorg. Chem. 2014, 1694–1705
1703
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Int. Ed. 1999, 38, 2596–2599; g) M. J. M. Vlaar, P. Valkier,
F. J. J. De Kanter, M. Schakel, A. W. Ehlers, A. L. Spek, M.
Lutz, K. Lammertsma, Chem. Eur. J. 2001, 7, 3551–3557.
A. Marinetti, F. Mathey, Organometallics 1982, 1, 1488–1492.
S. Shah, J. D. Protasiewicz, Coord. Chem. Rev. 2000, 210, 181–
201.
R. E. Bulo, L. Trion, A. W. Ehlers, F. J. J. De Kanter, M.
Schakel, M. Lutz, A. L. Spek, K. Lammertsma, Chem. Eur. J.
2004, 10, 5332–5337.
[33]
[34]
[35]
C. Schulten, G. Von Frantzius, G. Schnakenburg, A. Espinosa,
R. Streubel, Chem. Sci. 2012, 3, 3526–3533.
K. W. Feindel, R. E. Wasylishen, Can. J. Chem. 2004, 82, 27–
44.
[5]
[6]
a) X. Zeng, M. Ge, Z. Sun, D. Wang, Inorg. Chem. 2005, 44,
9283–9287; b) X. Yang, Z. Sun, M. Ge, S. Zheng, L. Meng, D.
Wang, ChemPhysChem 2002, 3, 963–966; c) W. Jing, S. Zheng,
Z. Xinjiang, Y. Xiaojun, G. Maofa, W. Dianxun, Angew. Chem.
2001, 113, 3145; Angew. Chem. Int. Ed. 2001, 40, 3055–3057;
d) A. Marinetti, J. Fischer, F. Mathey, J. Am. Chem. Soc. 1985,
107, 5001–5002; e) K. Miqueu, J.-M. Sotiropoulos, G. Pfister-
Guillouzo, V. D. Romanenko, New J. Chem. 2001, 25, 930–938;
f) K. Miqueu, J.-M. Sotiropoulos, G. Pfister-Guillouzo, A.-C.
Gaumont, J.-M. Denis, Organometallics 2001, 20, 143–147.
L. Pauling, The Nature of the Chemical Bond, Cornell Univ.,
USA, 3rd ed. 1960.
K. Kimura, Handbook of HeI photoelectron spectra of funda-
mental organic molecules: ionization energies, ab initio assign-
ments and valence electronic structure for 200 molecules, Japan
Scientific Societies Press, New York, Halsted Press, 1981, p. 1–
268.
[7]
[8]
[9]
K. Vaheesar, T. M. Bolton, A. L. L. East, B. T. Sterenberg, Or-
ganometallics 2010, 29, 484–490.
R. Streubel, A. Ostrowski, H. Wilkens, F. Ruthe, J. Jeske, P. G.
Jones, Angew. Chem. 1997, 109, 409; Angew. Chem. Int. Ed.
Engl. 1997, 36, 378–381.
a) A. A. Khan, C. Wismach, P. G. Jones, R. Streubel, Dalton
Trans. 2003, 2483–2487; b) A. A. Khan, C. Wismach, P. G.
Jones, R. Streubel, Chem. Commun. 2003, 2892–2893; c) A.
Özbolat, A. A. Khan, G. Frantzius, M. Nieger, R. Streubel,
Angew. Chem. 2007, 119, 2150; Angew. Chem. Int. Ed. 2007,
46, 2104–2107; d) N. H. T. Huy, F. Mathey, J. Org. Chem. 2000,
65, 652–654.
R. Streubel, A. Ostrowski, H. Wilkens, F. Ruthe, J. Jeske, P. G.
Jones, Angew. Chem. 1997, 109, 409; Angew. Chem. Int. Ed.
Engl. 1997, 36, 378–381.
R. E. Bulo, A. W. Ehlers, F. J. J. De Kanter, M. Schakel, M.
Lutz, A. L. Spek, K. Lammertsma, B. Wang, Chem. Eur. J.
2004, 10, 2732–2738.
[10]
[36]
[37]
[11]
[12]
[38]
For the structure with a C–P–W bond angle of 90°, the HOMO
is the antibonding combination between the σPC and the
dxz(W) orbitals, the HOMO–4 the bonding combination and
the HOMO–1 corresponds to the stabilizing interaction be-
2
tween the np(P) and the d (W) orbitals (see the Supporting
Information, Figure S2-b). zFor the bent structure of 4c, the
[13]
[14]
F. Mercier, B. Deschamps, F. Mathey, J. Am. Chem. Soc. 1989,
111, 9098–9100.
M. L. G. Borst, R. E. Bulo, C. W. Winkel, D. J. Gibney, A. W.
Ehlers, M. Schakel, M. Lutz, A. L. Spek, K. Lammertsma, J.
Am. Chem. Soc. 2005, 127, 5800–5801.
2
HOMO corresponds to a mixture of the np(P), the dz (W)
[weak participation] and the dxz(W) [strong participation] or-
bitals (see HOMO in the Supporting Information, Figure S2-
a). For this bent structure, a secondary stabilizing interaction
2
appears between the σPC orbital and the dz (W) orbitals (see
[15]
[16]
[17]
M. P. Duffy, F. Mathey, J. Am. Chem. Soc. 2009, 131, 7534–
7535.
the Supporting Information HOMO–4 in Figure S2-a and
ZOOM of this orbital). Moreover, the destabilizing interaction
between the σPC and the dxz(W) orbitals strongly decrease com-
pared with the structure with the C–P–W bond angle at 90°
(see the Supporting Information, Figure S2-b HOMO and
HOMO–4 compared to HOMO–4 in Figure S2-a).
C. Compain, B. Donnadieu, F. Mathey, Organometallics 2006,
25, 540–543.
H. Jansen, M. C. Samuels, E. P. A. Couzijn, J. C. Slootweg,
A. W. Ehlers, P. Chen, K. Lammertsma, Chem. Eur. J. 2010,
16, 1454–1458.
B. T. Sterenberg, K. A. Udachin, A. J. Carty, Organometallics
2003, 22, 3927–3932.
J. Sánchez-Nieves, B. T. Sterenberg, K. A. Udachin, A. J.
Carty, J. Am. Chem. Soc. 2003, 125, 2404–2405.
B. T. Sterenberg, A. J. Carty, J. Organomet. Chem. 2001, 617,
696–701.
B. T. Sterenberg, K. A. Udachin, A. J. Carty, Organometallics
2001, 20, 2657–2659.
B. T. Sterenberg, O. S. Senturk, K. A. Udachin, A. J. Carty, Or-
ganometallics 2007, 26, 925–937.
T. W. Graham, K. A. Udachin, M. Z. Zgierski, A. J. Carty, Or-
ganometallics 2011, 30, 1382–1388.
S. Creve, K. Pierloot, M. T. Nguyen, Chem. Phys. Lett. 1998,
285, 429–437.
S. Creve, K. Pierloot, M. T. Nguyen, L. G. Vanquickenborne,
Eur. J. Inorg. Chem. 1999, 107–115.
A. W. Ehlers, K. Lammertsma, E. J. Baerends, Organometallics
1998, 17, 2738–2742.
J.-G. Lee, J. E. Boggs, A. H. Cowley, Polyhedron 1986, 5, 1027–
1029.
D. Gonbeau, G. Pfister-Guillouzo, A. Marinetti, F. Mathey,
Inorg. Chem. 1985, 24, 4133–4140.
S. Grigoleit, A. Alijah, A. B. Rozhenko, R. Streubel, W. W.
Schoeller, J. Organomet. Chem. 2002, 643, 223–230.
K. K. Pandey, P. Tiwari, P. Patidar, J. Phys. Chem. A 2012, 116,
11753–11762.
[39]
[40]
S. Dapprich, G. Frenking, J. Phys. Chem. 1995, 99, 9352–9362.
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
a) NBO 5.0. E. D. Glendening, J. K. Badenhoop, A. E. Reed,
J. E. Carpenter, J. A. Bohmann, C. M. Morales, F. Weinhold,
Theoretical Chemistry Institute, University of Wisconsin,
Madison, 2001; b) A. E. Reed, L. A. Curtiss, F. Weinhold,
Chem. Rev. 1988, 88, 899–926; c) J. P. Foster, F. Weinhold, J.
Am. Chem. Soc. 1980, 102, 7211–7218; d) A. E. Reed, F. Wein-
hold, J. Chem. Phys. 1985, 83, 1736–1740.
R. G. Parr, L. v. Szentpaly, S. Liu, J. Am. Chem. Soc. 1999,
121, 1922–1924.
a) R. G. Parr, W. Yang, J. Am. Chem. Soc. 1984, 106, 4049–
4050; b) W. Yang, W. J. Mortier, J. Am. Chem. Soc. 1986, 108,
5708–5711.
[41]
[42]
[43]
[44]
Kohn–Sham energies of the orbitals involved in the LUMO of
4a–c are: –4.05 eV for 3pπ (P) – dyz(W) in HP=W(CO)5;
Ph
Ph
0.10 eV for π*1 in benzene and –1.23 eV for π*1 + π*CϵN
in p-NC–C6H4.
The decomplexation process [precursor (3a–c)ǞW(CO)5
+
corresponding phosphanorbornadienes] cannot compete with
dimethyl 4,5-dimethylphthalate elimination because the decom-
plexation is strongly endothermic, ΔH(decomplexation)
≈ 40 kcalmol–1 (from 3a: ΔH = 37.9 kcalmol–1; from 3b: ΔH =
39.5 kcalmol–1; from 3c: ΔH = 39.8. kcalmol–1). To exclude
the entropy problem when monomolecular and bimolecular
processes are compared, ΔH values have been used. Indeed,
dimethyl 4,5-dimethylphthalate elimination can be considered
as a monomolecular process because the activation barriers
(ΔH#3a–c ≈ 23 kcalmol–1, see Figure 6) from 3a–c to TS have
been calculated from the difference between compound 3a–c
and TS, whereas the decomplexation reaction is a bimolecular
process because ΔH(decomplexation) values have been calcu-
K. K. Pandey, A. Lledós, J. Organomet. Chem. 2010, 695, 206–
214.
K. Lammertsma, A. W. Ehlers, M. L. McKee, J. Am. Chem.
Soc. 2003, 125, 14750–14759.
Eur. J. Inorg. Chem. 2014, 1694–1705
1704
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
lated from the difference between 3 and the sum of [4a–c +
W(CO)5].
[50]
[51]
[52]
[53]
a) Molekel, v. 5.4.0.8, U. Varetto, Swiss National Supercom-
puting Centre, Lugano, Switzerland.
Gaussview 5.0, Gaussian Inc., 340 Quinnipiac St Bldg 40,
Wallingford, CT 06492 USA.
P. K. Chattaraj, S. Nath, A. B. Sannigrahi, J. Phys. Chem.
1994, 98, 9143–9145.
a) Y. Yang, R. G. Parr, Proc. Natl. Acad. Sci. NSA 1985, 82,
6723–6726; b) M. Berkowitz, S. K. Ghosh, R. G. Parr, J. Am.
Chem. Soc. 1985, 107, 6811–6814; c) M. Berkowitz, R. G. Parr,
J. Chem. Phys. 1988, 88, 2554–2557; d) P. W. Ayers, M. Levy,
Theor. Chem. Acc. 2000, 103, 353–360; e) P. W. Ayers, R. G.
Parr, J. Am. Chem. Soc. 2000, 122, 2010–2018.
[45] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Mont-
gomery, R. E. Stratman, J. C. Burant, S. Dapprich, J. M. Mil-
lam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pom-
elli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.
Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowswi, J. V. Ortiz, A. G.
Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R. Martin, D. J. Fox, T. Keith, M. A.
Al–Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M.
Challacombe, P. M. W. Gill, B. Jonhson, W. Chen, M. W.
Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle, J. A. Po-
ple, Gaussian 03, D.02 ed. Pittsburgh PA, 2003.
[54] A. J. Arduengo III, H. Bock, H. Chen, M. Denk, D. A. Dixon,
J. C. Green, W. A. Hermann, N. L. Jones, M. Wagner, R. West,
J. Am. Chem. Soc. 1994, 116, 6641–6649.
[55] a) H. M. Muchall, N. H. Werstiuk, B. Choudhurry, Can. J.
Chem. 1998, 76, 221–227; b) H. M. Muchall, N. H. Werstiuk,
B. Choudhurry, J. Ma, J. Warkentin, J. P. Pezacki, Can. J.
Chem. 1998, 76, 238–240.
[46] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) C. Lee,
W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789; c) K.
Burke, J. P. Perdew, Y. Wang, Electronic Density Functional
Theory: Recent Progress and New Directions (Eds.: J. F. Dob-
son, G. Vignale, M. P. Das), Plenum Press, New York, 1998.
[47] a) P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283; b)
W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284–98; c) W. R.
Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284–298.
[48] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28, 213–
222.
[49] a) C. Gonzalez, H. B. Schlegel, J. Chem. Phys. 1989, 90, 2154–
2161; b) C. Gonzalez, H. B. Schlegel, J. Phys. Chem. 1990, 94,
5523–5527.
[56] a) H. M. Muchall, P. Rademacher, J. Mol. Struct. 1997, 435,
157–167; b) H. M. Muchall, P. Rademacher, J. Mol. Struct.
1998, 471, 189–194.
[57] R. Stowasser, R. Hoffmann, J. Am. Chem. Soc. 1999, 121,
3414–3420.
Received: August 8, 2013
Published Online: September 24, 2013
Eur. J. Inorg. Chem. 2014, 1694–1705
1705
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim