Evidence for ligand-centered reactivity of a 17e radical cationic 2H-azaphosphirene complex

Article abstract of DOI:10.1002/ejic.200700658

Reactions of 2H-azaphosphirene complex 1 with electron-rich nitrile derivatives 2a-e in the presence of substoichiometric amounts of ferrocenium hexafluorophosphate yielded regioselectively 2H-1,4,2-diazaphosphole complexes 3a-e under mild conditions. The dependence of the reaction on the amount of ferrocenium hexafluorophosphate is discussed. In the reactions of 1 with electron-poor nitrile derivatives HCN (2f), EtOC(O)CN (2g), and C ₆F₅CN (2h), byproducts possessing a P-F bond are formed, which indicates that the hexafluorophosphate anion is involved in the reaction course. Apart from NMR spectroscopic (3a-f) and X-ray data (3b, 3d-e), a detailed DFT study is included that provides first insight into the reaction pathway. Wiley-VCH Verlag GmbH & Co. KGaA, 2007.

Full text of DOI:10.1002/ejic.200700658

FULL PAPER
DOI: 10.1002/ejic.200700658
Evidence for Ligand-Centered Reactivity of a 17e Radical Cationic
2H-Azaphosphirene Complex
Holger Helten,[a] Christoph Neumann,[b] Arturo Espinosa,[c] Peter G. Jones,[b]
Martin Nieger,[d] and Rainer Streubel*[a]
Keywords: Ligand-centered reactivity / Phosphorus heterocycles / Electron transfer
Reactions of 2H-azaphosphirene complex 1 with electron-
rich nitrile derivatives 2a–e in the presence of substoichio-
metric amounts of ferrocenium hexafluorophosphate yielded
regioselectively 2H-1,4,2-diazaphosphole complexes 3a–e
under mild conditions. The dependence of the reaction on
the amount of ferrocenium hexafluorophosphate is dis-
cussed. In the reactions of 1 with electron-poor nitrile deriva-
ucts possessing a P–F bond are formed, which indicates that
the hexafluorophosphate anion is involved in the reaction
course. Apart from NMR spectroscopic (3a–f) and X-ray data
(3b, 3d–e), a detailed DFT study is included that provides
first insight into the reaction pathway.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
tives HCN (2f), EtOC(O)CN (2g), and C6F5CN (2h), byprod- Germany, 2007)
oxidation.[1,2] In this work, ferrocenium salts are used as
Introduction
mild single-electron oxidants, which have the further advan-
tage that they are more easily recycled than most other
heavy metal oxidants. Ferrocenium salts have been success-
fully applied in organic[8] and organometallic syntheses.[5]
Recently, they were employed in the oxidation of N-hetero-
cyclic carbenes.[9]
In the course of our studies concerning the applicability
of 2H-azaphosphirene metal complexes I[10] (Scheme 1) in
heterocyclic ligand syntheses,[11] we discovered that I un-
dergoes P–N bond-selective ring-expansion reactions in the
presence of substoichiometric amounts of typical single-
electron oxidants such as tetracyanoethylene[12] and/or
ferrocenium hexafluorophosphate.[13]
The generation of radical cationic intermediates has
emerged as a useful synthetic tool in organic[1–4] and orga-
nometallic chemistry.[5] Oxidation of closed-shell organic
molecules by single-electron transfer (SET) results in
strongly destabilized odd-electron species that very often
tend to fragment. Thus, upon pulse radiolysis or 60Co γ
radiolysis, phenylaziridines either form transient radical cat-
ions with retention of the closed ring structure or undergo
ring opening with formation of radical cationic azomethine
ylides depending on their substitution patterns.[6] Although
the spin density of most organometallic radicals is localized
mainly at the metal center, organometallic radicals can
show both metal-centered and/or ligand-centered reactivity,
whereas 17e cationic complexes often show the latter behav-
ior.[7] For the generation of radical cations in solution four
major routes are currently in use: (1) anodic oxidation, (2)
photoinduced electron transfer (PET), (3) oxidation by rad-
icals generated by pulse or γ radiolysis, and (4) chemical
Scheme 1. 2H-Azaphosphirene complexes I and radical cationic
2H-azaphosphirene complexes I·+ (M = Cr, Mo, W; R1, R2 = ubi-
quitous organic substituents).
[a] Institut für Anorganische Chemie, Rheinische Friedrich-
Wilhelms-Universität Bonn,
Gerhard-Domagk-Strasse 1, 53121 Bonn, Germany
Fax: +49-228-73-9616
Preliminary studies showed that these reactions require
polar solvents such as dichloromethane or carbonitriles, re-
actants with a π system having a lone pair of electrons at
one terminus, and substituents that are not (too) bulky. In
all ferrocenium hexafluorophosphate induced ring expan-
sions, the formation of small amounts of ferrocene is ob-
served,[13b,13c] which indicates the transient formation of
radical cationic species such as I·+ (Scheme 1). Further evi-
dence for the intermediate formation of I·+ was obtained by
the finding that reaction rates are reduced if ortho-phenyl-
E-mail: r.streubel@uni-bonn.de
[b] Institut für Anorganische und Analytische Chemie der Tech-
nischen Universität Braunschweig,
Postfach 3329, 38023 Braunschweig, Germany
[c] Departamento de Química Orgánica, Facultad de Química,
Universidad de Murcia,
Campus de Espinardo, 30100 Murcia, Spain
[d] Laboratory of Inorganic Chemistry, Department of Chemistry,
University of Helsinki,
P. O. Box 55 (A.I. Virtasen aukio 1), 00014 Helsinki, Finland
Supporting information for this article is available on the
WWW under http://www.eurjic.org or from the author.
Eur. J. Inorg. Chem. 2007, 4669–4678
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4669

H. Helten, C. Neumann, A. Espinosa, P. G. Jones, M. Nieger, R. Streubel
FULL PAPER
substituted 2H-azaphosphirene complexes with electron-
donating substituents were employed.[13b] Despite all these
indications, fundamental questions remained open regard-
ing reaction pathways, for example, at which stage does the
three-membered ring open and/or do transition-metal coor-
dination spheres increase during the reactions.
Here, reactions of 2H-azaphosphirene complex 1 with ni-
triles of varying donor abilities under SET conditions are
reported. A detailed DFT study on the reaction pathway is
included.
Figure 1. Ratio of 3d over reaction time for the reaction of 1 with
2d in the presence of varying amounts of [FcH]PF6: ᭿ 0.01, ᭹
0.02, ᭡ 0.03, ½ 0.04, o 0.05, · 0.10, + 0.15, - 0.20 equiv.
Results and Discussion
Reactions of 2H-azaphosphirene complex 1[10] with di-
methylcyanamide (2a) and electron-rich five-membered het-
erocyclic nitrile derivatives RCN (2b–e; Scheme 2) in the
presence of substoichiometric amounts (0.05 equiv.) of
ferrocenium hexafluorophosphate ([FcH]PF6) at ambient
temperature regioselectively yielded 2H-1,4,2-diazaphos-
phole complexes 3a–e (Scheme 2). The molecular structures
of 3a–e were unambiguously established by multinuclear
NMR spectroscopic experiments and mass spectrometry.
Scheme 2. Synthesis of 2H-1,4,2-diazaphosphole complexes 3a–h
(a: R = NMe2; b: R = 1,5-dimethyl-2-pyrryl; c: R = 2-furyl; d: R
= 2-thienyl; e: R = 3-thienyl; f: R = H; g: R = CO2Et; h: R =
C6F5).[14]
Figure 2. Ratio of 3d over reaction time for the reaction of 1 with
2d in the presence of 0.025 equiv. of [FcH]PF6. The reaction is com-
plete after 1 d.
We monitored the ring expansion reaction of 1 + 2d Ǟ
The flat trend at the start is quite obvious. The fastest
3d (Scheme 2) at different concentrations of [FcH]PF6 by increase in the amount of 3d is recorded between 29 and
using 31P{1H} NMR spectroscopy (Figure 1). These series 42% conversion. It should be noted that a flat trend at the
revealed that the reaction progression depends strongly on start of the reaction is characteristic for autocatalytic reac-
the amount of [FcH]PF6, but it can be reduced to tions, and in reactions with a very low starting concentra-
0.03 equiv. (0.01 molL–1) without a significant decrease in tion of the autocatalytic species the peak of the reaction
the conversion. If only 0.02 equiv. of [FcH]PF6 was used, rate is recorded at around 50% conversion.[15]
the conversion leveled off at around 80%; in the presence
On the basis of these results and the finding that ferro-
of 0.01 equiv. of [FcH]PF6 even after 3 d the ratio of 3d cene is formed during the reaction, we suppose a radical
did not exceed 18%, and in the absence of [FcH]PF6 no cation chain reaction mechanism initiated by oxidation of
conversion of 1 was detected within 10 d (not shown). For 1 by the ferrocenium cation with formation of radical cation
ratios Ն0.02 equiv. of [FcH]PF6, the curves show a steep 1·+ and ferrocene (Scheme 3).[16]
ascending slope after a comparably flat trend at the start,
The observed acceleration in the reaction rate at the start
and again a flattening at the end. As no resonances other (Figures 1 and 2) can be explained by assuming slow oxi-
than those of 1 and 3d were detected by 31P{1H} NMR dation of 1 followed by a fast ring-expansion reaction of
spectroscopic monitoring during the reaction course, the highly reactive radical cation 1·+ with nitrile 2a–e. The cycle
time dependent decrease of 1 follows the same trend as the is closed by reduction of radical cationic 2H-1,4,2-diaza-
increase of 3d. Because we were interested in understanding phosphole complexes 3·+a–e by 1 to yield 3a–e and 1·+,
the start of the reaction, we decided to monitor the reaction which then can restart the chain reaction. In consequence,
of 1 with 2d with 0.025 equiv. of [FcH]PF6 at more frequent the generation of 1·+ causes its multiplication, and it may
time intervals (Figure 2).
be regarded as the autocatalytic species of this cycle. If
4670
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 4669–4678

Reactivity of a 17e Radical Cationic 2H-Azaphosphirene Complex
an excess of 6.4 equiv. was employed. The 19F{1H} NMR
spectroscopic resonance of the hexafluorophosphate anion
decreased in the course of the reaction, and after 6 d its
31P{1H} NMR spectroscopic resonance was no longer de-
tected.
Table 1. 31P{1H} NMR spectroscopic data of 3a–f.
3a[a]
3b[a]
3c[a]
3d[a]
3e[a]
3f[b]
δ / ppm
102.1
240.3
108.4
233.9
110.6
230.6
110.5
229.5
109.1
229.8
105.5
225.1
1JW,P / Hz
[a] In CDCl3. [b] In CH2Cl2.
Scheme 3. Proposed chain reaction of 2H-azaphosphirene complex
1 with nitrile derivatives 2a–e initiated by the ferrocenium cation
[FcH]+.
The 31P{1H} NMR spectroscopic data of 3b–e, displayed
in Table 1, are similar, and each complex has a resonance
at about 110 ppm. Concerning the magnitudes of the tung-
sten–phosphorus coupling constant, a definite trend can be
1
0.02 equiv. of [FcH]PF6 was employed (Figure 1), the chain seen: JW,P constants increase with increasing donor abil-
reaction was terminated at some point. Obviously, ity[18] of the substituent at C-5 in the order: 3-thienyl ≈ 2-
0.01 equiv. of [FcH]PF6 is not sufficient for the successful thienyl Ͻ 2-furyl Ͻ 1,5-dimethyl-2-pyrryl. The tungsten–
initiation of the chain reaction (Figure 1).
phosphorus coupling constant of complex 3a, which has a
In contrast to 2a–e, the reaction of 1 with HCN (2f) pro- very strong electron-donating substituent (NMe2) attached
ceeded much more slowly and was unselective. Reaction to C-5, is about 6 Hz larger than that of complex 3b,
monitoring showed that even after a short time (1.5 h) a whereas the 31P NMR spectroscopic resonance is shifted by
variety of 31P NMR spectroscopic resonances was detected, around 6 ppm to higher field. Consequently, the smallest
including a resonance at δ = 105.5 ppm (Table 1) with a tungsten–phosphorus coupling constant (1JW,P = 225.1 Hz)
tungsten–phosphorus coupling constant of 225.1 Hz and a was exhibited by complex 3f, which has no substituent in
phosphorus–hydrogen coupling constant of 34.3 Hz, which the C-5 position, which lends further support to the assign-
revealed the formation of 3f (ca. 16%). Even though an ment made beforehand. A further consequence can be ex-
excess of 2f (2.6 equiv.) was employed, after 4 d the reaction pected from this trend: for 3g and 3h even lower values are
solution still contained 16% (ratio estimation by 31P{1H} expected.
NMR spectroscopic resonance integration) of unreacted 1
To examine the aforementioned formation of products
and only 20% of 3f, besides numerous unidentified byprod- with a P–F bond, we investigated the role of the
ucts. Some of the latter showed large phosphorus–fluorine counteranion. In the reaction of 1 with 2d we exchanged
coupling constants (see Experimental Section), thus reveal- hexafluorophosphate for the weaker nucleophilic counter-
ing that the hexafluorophosphate anion, which is usually anion tetraphenylborate, which could not serve as a fluoride
regarded as weakly coordinating, had taken part in the re- source. The outcome was a clean reaction yielding complex
action course. As reported recently, highly electrophilic cat- 4d (by 31P NMR spectroscopy), which also indicated that
ionic species are able to abstract fluoride from the hexafluo- the counteranion did not play a crucial role for the ring
rophosphate anion.[17] Byproducts bearing a P–F bond expansions of 1 as induced by ferrocenium salts. So, the
were also found in the reaction of 1 with electron-poor ni- products bearing a P–F bond in the reaction of 1 with 2f
trile derivative 2g (see below and Experimental Section). and 2g do not constitute intermediates of the ring-expan-
After 6 d, 41% of 1 remained unreacted, and a product with sion reaction pathway. Instead, these should be regarded as
a
31P NMR spectroscopic resonance at δ = 110.7 ppm byproducts emerging from the reaction of transiently
(21%) was observed that featured a tungsten–phosphorus formed highly electrophilic species with hexafluorophos-
coupling constant of 227.6 Hz. These data are too close to phate through fluoride transfer. To elucidate further the ef-
those of 3b–e and, therefore, an assignment to 3g is not fect that [FcH]BPh4 might have on reactions of 1, we
supported; even more vexing is the finding that in the reac- looked again at the reactions with electron-poor nitrile de-
tion of 1 with 2f a byproduct having analogous 31P{1H} rivatives 2f–h by using this SET reagent. Under these condi-
NMR spectroscopic data was detected. Reaction of 1 with tions, 1 was treated with 2f to afford the product (again)
2h showed a similar behavior to that of 1 with 2g, and having a 31P{1H} NMR spectroscopic resonance at δ =
again, a product with a 31P NMR spectroscopic resonance 110.6 ppm (1JW,P = 227.6 Hz; ca. 15%) and at δ =
at δ = 110.7 ppm with a tungsten–phosphorus coupling 105.5 ppm (1JW,P = 225.1 Hz: ca. 5%) that we assign to 3f
constant of 228.9 Hz was observed (12% after 6 d) as well and traces of unidentified byproducts; ca. 66% of 1 re-
as numerous byproducts, some of which showed large phos- mained unreacted, which did not change significantly
phorus–fluorine coupling constants (see Experimental Sec- within 4 d. Reactions of 1 with 2g and 2h in the presence
tion). Even after 6 d, 48% of 1 Ϫ and the entire amount of of [FcH]BPh4 showed almost no conversion of 1 within 1 d.
2h [by 19F{1H} NMR spectroscopy] Ϫ remained unreacted, Nevertheless, the 31P{1H} NMR spectroscopic resonance at
thus indicating that 2h had not reacted at all even though δ = 110.6 ppm (1JW,P = 227.6 Hz) slowly increased after sev-
Eur. J. Inorg. Chem. 2007, 4669–4678
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4671

H. Helten, C. Neumann, A. Espinosa, P. G. Jones, M. Nieger, R. Streubel
FULL PAPER
eral days in both reactions, and again, complex 1 did not
react with 2h.
The molecular structures of 3b and 3d–e were determined
by single-crystal X-ray diffraction analysis[19] (Figures 3, 4,
and 5).
Figure 5. Molecular structure of complex 3e in the crystal (50%
probability level, hydrogen atoms are omitted for clarity). Selected
bond lengths [Å] and angles [°]: W–P 2.5321(5), P–C8 1.8335(16),
P–N1 1.7042(14), P–C6 1.8767(16), C6–N2 1.2990(21), N2–C7
1.4213(22), C7–N1 1.2903(21), C7–C21 1.4592(22), C6–C15
1.4675(23), N1–P–C6 90.02(7), P–C6–N2 109.14(12), C6–N2–C7
110.33(14), N2–C7–N1 120.74(14), C7–N1–P 109.31(12), P–C6–
C15 128.99(12), N1–C7–C21 122.28(16).
Figure 3. Molecular structure of complex 3b in the crystal (30%
probability level, hydrogen atoms are omitted for clarity). Selected
bond lengths [Å] and angles [°]: W–P 2.5324(13), P–C8 1.8345(47),
P–N1 1.7006(39), P–C6 1.8701(45), C6–N2 1.2947(57), N2–C7
1.4344(57), C7–N1 1.2946(59), C7–C21 1.4228(62), C6–C15
1.4876(60), N1–P–C6 89.86(19), P–C6–N2 110.00(32), C6–N2–C7
109.92(37), N2–C7–N1 119.98(40), C7–N1–P 109.80(31), P–C6–
C15 127.42(33), N1–C7–C21 124.78(44).
and the particular cyclic substituent and the torsion angles
between the planes of the cyclic substituents are given in
Table 2. The smallest deviation from planarity is found in
complex 3b. Bond lengths and angles in the diazaphosphole
ring systems of 3b and 3d–e are almost identical.
Table 2. Torsion angles of ring planes[a] for 3b and 3d–e.
Diazaphosphole / Diazaphosphole /
3-Hetaryl /
5-Ph
5-Ph
3-hetaryl
3b
3d
3e
6.68(17)
4.29
4.24(8)
1.88(22)
3.47 / 9.96[b]
2.79(8)
5.25(20)
3.52 / 7.63[b]
2.26(9)
[a] Torsion angles with respect to least square planes; all atoms of
particular ring system involved. [b] Molecule is disordered in the
crystal. Values are given for both conformations.
Theoretical Approach
The stationary points of a plausible ring-expansion path-
way were calculated by DFT methods [B3LYP/TZVP//ECP-
60-MWB (W)] on a model system for complex 1, i.e. 4. For
the nitrile component, three model compounds were used,
which might be classified according to their nucleophilicity:
(1) medium (2f: R = H), (2) strong (2i: R = NH2), and (3)
weak (2j: R = CN). The influence of the solvent was taken
into account by employing the COSMO approach with ε
= 8.93 to mimic dichloromethane. It was shown that this
approach is appropriate.[21] In consequence, computed rela-
tive free energies (Tables 3 and 4) for reactions in solution
Figure 4. Molecular structure of complex 3d in the crystal (50%
probability level, hydrogen atoms and minor disordered part of the
thiophene substituent are omitted for clarity). Selected bond
lengths [Å] and angles [°]: W1–P1 2.531(10), P1–C3 1.835(3), P1–
N2 1.707(3), P1–C1 1.877(4), C1–N1 1.304(5), N1–C2 1.424(5),
C2–N2 1.293(5), C2–C16 1.441(5), C1–C10 1.457(6), N2–P1–C1
90.1(2), P1–C1–N1 109.2(3), C1–N1–C2 110.1(3), N1–C2–N2
121.0(3), C2–N2–P1 109.1(3), P1–C1–C10 129.3(3), N2–C2–C16
121.5(4).
Complexes 3b and 3d,e are isotypic. They feature almost are discussed hereafter; computed stationary points are
planar phosphorus heterocycles and, to a large extent, a shown in Schemes 4 and 5. All relative energies, enthalpies,
coplanar arrangement of the three ring systems. The torsion and free energies are available in the Supporting Infor-
angles between the planes of the phosphorus heterocycle mation.
4672
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 4669–4678

Reactivity of a 17e Radical Cationic 2H-Azaphosphirene Complex
Table 3. Calculated thermochemical data for reactions shown in
Scheme 4 (all values in kJmol–1).
϶
Reaction
∆G298
∆RG298
[a]
i
ii
4 + FcH+ Ǟ4·+ + FcH
4·+ + HCN (2f)Ǟ5·+f
4·+ + H2NCN (2i)Ǟ5·+i
+139.2
+24.0
–13.4
–39.5
–102.6
–96.2
–87.5
+59.8
[b]
iii
iv
v
4·+ + NCCN (2j)Ǟ6·+j + HCN (2f) +74.7
5·+iǞ7·+i
+22.3
–1.7
+5.0
5·+fǞ6·+f + HCN (2f)
5·+iǞ6·+i + HCN (2f)
[a] The activation barrier was not calculated. We assume an outer
sphere mechanism. [b] Transition state could not be located.
Scheme 5. Calculated pathways for the reactions of radical cationic
nitrilium phosphane ylide complexes 6·+f,i,j with HCN (2f) (f: R =
H, i: R = NH2, j: R = CN).
Scheme 4. Computed pathways for the reactions of 2H-azaphos-
phirene complex 4·+ with nitrile derivatives 2f (R = H), 2i (R =
NH2), and 2j (R = CN).
Figure 6. Highest singly occupied molecular orbitals (α-SOMO) of
4·+ (left) and 5·+f (right).
Table 4. Calculated thermochemical data for reactions shown in
Scheme 5 (all values in kJmol–1).
϶
Reaction
vi
∆G298
∆RG298
phenyl in 1 by a hydrogen atom, and thus radical cation 4·+
is destabilized with respect to 4. We also expect the free
energy for the real system (compound 1) to be somewhat
lower but still endergonic, so the equilibrium between 1 and
1·+ may support a low concentration of radical cation 1·+.
Most synthetically applicable radical cation chain reactions
require an endergonic SET preequilibrium to initiate selec-
tive follow up reactions of highly electrophilic radical cat-
ions.[1] To save computing time we investigated the hyper-
surface of radical cationic model system 4·+ by assuming
that all stationary points of the reaction are affected in ap-
proximately the same way.
The highest singly occupied molecular orbital (α-SOMO)
of 17e complex 4·+, depicted in Figure 6, is mainly localized
at the metal center. Only a minor participation of the nitro-
gen atom is apparent. An intermediate heptacoordinated
tungsten complex, however, bearing one of the nitriles 2f,i,j
6·+f + HCN (2f)Ǟ7·+f +99.1
–20.5
–15.0
–52.6
–41.4
–42.6
–78.9
+75.1
+108.7
+50.3
–95.6
–123.7
–102.9
–12.7
–8.3
6·+i + HCN (2f)Ǟ7·+i
6·+j + HCN (2f)Ǟ7·+j
+100.6
+94.9
vii
viii
ix
6·+f + HCN (2f)Ǟ8·+f +116.9
6·+i + HCN (2f)Ǟ8·+i
6·+j + HCN (2f)Ǟ8·+j
+114.3
+92.9
6·+f + HCN (2f)Ǟ9·+f +90.3
[a]
6·+i + HCN (2f)Ǟ9·+i
6·+j + HCN (2f)Ǟ9·+j
9·+fǞ7·+f
+77.2
+28.8
[a]
9·+iǞ7·+i
9·+jǞ7·+j
+32.3
[b]
7·+f + 4Ǟ7f + 4·+
7·+i + 4Ǟ7i + 4·+
7·+j + 4Ǟ7j + 4·+
[b]
[b]
–22.2
[a] Transition state could not be located. [b] The activation barrier
was not calculated. We assume an outer sphere mechanism.
The overall free energy balance for net ring-expansion as ligand could not be located as an energy minimum. This
reactions is exergonic (cf. Schemes 4 and 5; 4 + 2fǞ7f: ruled out a metal-centered reactivity of 4·+ towards nitriles.
∆RG298
=
–104 kJmol–1,
4
+
2iǞ7i: ∆RG298
=
Instead, a plausible pathway is the ring opening of 4·+ by
–124 kJmol–1, and 4 + 2jǞ7j: ∆RG298 = –114 kJmol–1). nucleophilic attack of nitriles 2f,i with formation of inter-
Upon oxidation of 4, 17e cationic complex 4·+ is predicted mediates 5·+f,i (Scheme 4, reaction ii). The highest singly
to retain a closed-ring structure (Scheme 4); an isomer with occupied molecular orbital of 5·+f (Figure 6, right) also has
a broken P–N bond could not be localized as an energy a large contribution of the metal center, whereas a certain
minimum. The oxidation by the ferrocenium cation contribution of the nitrogen atom is apparent. This can be
(Scheme 4, reaction i) is predicted to be highly endergonic rationalized by the resonance structure given in Scheme 4.
(∆RG298 = +139 kJmol–1), which may be due to simplifi- The barrier TSiif for the formation of 5·+f (R = H) is
϶
cation resulting from the replacement of CH(SiMe3)2 and ∆G298 = +60 kJmol–1. For R = NH2 no transition state
Eur. J. Inorg. Chem. 2007, 4669–4678
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4673

H. Helten, C. Neumann, A. Espinosa, P. G. Jones, M. Nieger, R. Streubel
FULL PAPER
Figure 7. Transition states of the nucleophilic attack of 2f on 4·+ (TSiif, left), on 4 (TSiiif0, right) and of 2j on 4·+ (TSiiij, middle).
could be located even in a constrained search, which indi- and on the other hand, upon the attack of HCN or H2NCN
cates a very flat barrier. The formation of 5·+f is slightly on 4·+ the HCN substitution can proceed in two steps, that
endergonic (∆RG298 = +24 kJmol–1), whereas the formation is, under intermediate formation of 5·+f,i. Alternatively, for
of 5·+i is slightly exergonic (∆RG298 = –13 kJmol–1). With 5·+i (R = NH2), cyclization to 7·+i is feasible.
2j (R = CN), no intermediate 5·+j could be located. The
In the highest singly occupied molecular orbital (α-
binding of NCCN (through N) to the phosphorus center of SOMO) of 6·+f (Figure 8, left) a significant contribution of
4·+ is associated with dissociation of the HCN moiety of the phosphorus atom is clear (as represented by the reso-
4·+ and exergonic formation of radical cationic nitrilium nance structure in Scheme 4), which is in accord with the
phosphane ylide complex 6·+j (Scheme 4, reaction iii; interpretation that 6·+f lacks one electron of the lone pair
∆RG298 = –40 kJmol–1). The dissociation of HCN is indi- that is assigned to phosphorus in neutral nitrilium phos-
cated in transition state TSiiij (Figure 7, middle) by the phane ylide complexes.[22] In 17e cationic 2H-1,4,2-diaza-
imaginary frequency, a 0.15 Å longer P–C distance and a phosphole complex 7·+f, the α-SOMO is almost exclusively
4.7° smaller P–C–N angle relative to ring-opening transi- located at the metal center (Figure 8, right).
tion state TSiif (Figure 7, left). For the one-step substitution
of HCN from 4·+ by nucleophilic attack of 2j, the free acti-
϶
vation energy TSiiij was determined as ∆G298
=
+75 kJmol–1. A further interesting question was to clarify
the outcome of the nucleophilic attack of nitriles at the
phosphorus atom of neutral 2H-azaphosphirene complex 4.
As indicated in the transition state, denoted as TSiiif0 (Fig-
ure 7, right), dissociation of HCN occurs also from neutral
4 upon attack of 2f, and this reaction is also exergonic
(∆RG298 = –15 kJmol–1) but the barrier is significantly
higher (∆G298϶ = +92 kJmol–1) than for corresponding rad-
ical cation 4·+. This, together with steric factors for real sys-
tem 1, might be the reason why 2H-azaphosphirene com-
plex 1 does not react with nitriles in the absence of single-
electron oxidants at ambient temperature.
Figure 8. Highest singly occupied molecular orbitals (α-SOMO) of
6·+f (left) and 7·+f (right).
In all reactions of 4·+ with the three model nitrile deriva-
tives there are plausible pathways leading to the formation
Compound 5·+i (R = NH2) can undergo cyclization to of 6·+f,i,j with release of HCN, which in turn can (in prin-
7·+i, a process that is predicted to be exergonic (Scheme 4, ciple) react with 6·+f,i,j in two regiochemically different
reaction iv; ∆RG298 = –103 kJmol–1) and proceeding over a [3+2] cycloaddition reactions, that is, with formation of rad-
϶
barrier with TSivi (∆G298 = +22 kJmol–1). For 5·+f (R = ical cationic 2H-1,4,2-diazaphosphole complexes 7·+f,i,j
H), transition state TSivf for the cyclization could not be (Scheme 5, reaction vi) or radical cationic 2H-1,3,2-diaza-
located. A constrained decrease in the CN distance is asso- phosphole complexes 8·+f,i,j (Scheme 5, reaction vii). All re-
ciated with the breaking of the P–C bond. The dissociation actions are exergonic (Table 4) and 2H-1,3,2-diazaphos-
of HCN from 5·+f proceeds over a very flat barrier TSvf phole complexes 8·+f,i,j are ca. 20–30 kJmol–1 more stable
϶
϶
(∆G298 = –2 kJmol–1, ∆E298 = +13 kJmol–1) yielding than corresponding 2H-1,4,2-diazaphosphole complexes
HCN and radical cationic nitrilium phosphane ylide com- 7·+f,i,j. For the reactions of 6·+f,i (R = H and R = NH2)
plex 6·+f (Scheme 4, reaction v; ∆RG298 = –96 kJmol–1). For with HCN the barriers TSvif,i for the formation of 7·+f,i are
6·+i (R = NH2), the HCN dissociation constitutes a compet- 18 and 14 kJmol–1 lower than those for the formation of
ing pathway to the cyclization (Scheme 4, reaction iv). Cal- 8·+f,i (TSviif,i). For the reactions of 6·+j (R = CN) with
϶
culated free activation energy for TSvi is ∆G298
=
=
HCN the barriers TSvij and TSviij are approximately equal.
For the reaction of 6·+f,i,j with HCN a further alternative
+5 kJmol–1, and the free reaction energy is ∆RG298
–88 kJmol–1. In summary, the association of NCCN to 4·+ pathway is plausible, namely a stepwise mechanism starting
is simultaneously associated with a substitution of HCN, with the addition of HCN to 6·+f,i,j, which leads to the
4674
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 4669–4678

Reactivity of a 17e Radical Cationic 2H-Azaphosphirene Complex
formation of a C–N bond to yield intermediates 9·+f,i,j lectively 2H-1,4,2-diazaphosphole complexes 7·+. As experi-
(Scheme 5, reaction viii), with subsequent cyclization to mentally shown, complex 1 reacts with weak nucleophilic
7·+f,i,j (Scheme 5, reaction ix). The barriers for C–N bond nitrile derivatives 2g,h in the presence of ferrocenium salts
϶
formation TSviiif,j (∆G298 = +90 kJmol–1 for the reaction neither to 2H-1,4,2-diazaphosphole complexes 3g,h nor to
϶
of HCN with 6·+f and ∆G298 = +77 kJmol–1 for the reac- 2H-1,3,2-diazaphosphole complexes 8g,h. If hexafluoro-
tion with 6·+j) are slightly lower than those for the corre- phosphate is present, products bearing a P–F bond are
sponding direct [3+2] cycloaddition TSvif,j, whereas the formed. So, in the presence of weak nucleophilic nitrile de-
barriers for the second step TSixf,j are almost negligible rivatives a completely different reaction course is taken,
(∆G298϶ = +29 kJmol–1 for 9·+f and ∆G298϶ = +32 kJmol–1 which proceeds with the participation of the hexafluoro-
for 9·+j). The formation of 9·+f,i,j (Scheme 5, reaction viii) phosphate anion.
is predicted to be endergonic for our three model reactions,
that is, for the reaction of HCN with 6·+f ∆RG298
+75 kJmol–1, for the reaction with 6·+j ∆RG298
=
=
Experimental Section
+50 kJmol–1, and especially for the reaction with 6·+i
∆RG298 = +109 kJmol–1.
General Procedures: All reactions and manipulations were carried
out under an atmosphere of deoxygenated, dried nitrogen or argon
by using standard Schlenk techniques with conventional glassware.
Solvents were dried according to standard procedures. 2H-Azapho-
sphirene complex 1 was synthesized according to the method de-
scribed in the literature.[10d] NMR spectra were recorded with a
Bruker AC-200 spectrometer or a Bruker Avance 300 spectrometer
with [D6]benzene or [D]chloroform as solvent and internal stan-
dard; shifts are given relative to external tetramethylsilane (1H,
13C), neat CFCl3 (19F), and 85% H3PO4 (31P). Mass spectra were
recorded with a Finnigan MAT 8430 (70 eV) or a Kratos Concept
1H spectrometer (FAB+, mNBA); apart from m/z values of the
molecule ions, only m/z values having intensities more than 10%
are given. Infrared spectra were recorded with a Biorad FTIR 165
(selected data given) or a Bruker FTIR IFS113V spectrometer. UV/
Vis spectra were recorded with a Shimadzu UV-1650 PC spectrom-
eter. Melting points were obtained with a Büchi 535 capillary appa-
ratus or a Büchi apparatus Type S. The values are not corrected.
Elemental analyses were performed using a Carlo Erba analytical
gas chromatograph or an Elementar VarioEL instrument.
Once 7·+f,i,j are formed, the cycle is closed by the oxi-
dation of 2H-azaphosphirene complex 4 by 7·+f,i,j (for 7·+f:
∆RG298 = –13 kJmol–1, for 7·+i: ∆RG298 = –8 kJmol–1, for
7·+j: ∆RG298 = –22 kJmol–1) to yield 2H-1,4,2-diazaphos-
phole complexes 7f,i,j and 4·+, which then can restart the
chain reaction (Scheme 3).
Conclusions
A combined experimental and theoretical investigation
on the ring expansion of 2H-azaphosphirene complex 1
with a wide variety of electronically different nitriles re-
vealed that electron-rich nitrile derivatives 2a–e furnish re-
gioselectively 2H-1,4,2-diazaphosphole complexes 3a–e in
the presence of substoichiometric amounts of ferrocenium
hexafluorophosphate. The reaction course shows character-
istics of an autocatalytic reaction and is interpretable in
terms of a ligand-centered reactivity of an organometallic
radical cation (1·+). In contrast, reactions of complex 1 with
2f–h proceeded differently. With HCN (2f) complex 3f was
formed slowly and not selectively; byproducts having a P–
F bond were also formed. This provides the first strong evi-
dence that transiently formed highly electrophilic phospho-
rus-containing species had reacted with hexafluorophos-
phate through fluoride transfer. In reactions of complex 1
with weak nucleophilic nitrile derivatives 2g,h, byproducts
bearing a P–F bond were also formed, but the formation
of 2H-1,4,2-diazaphosphole complexes 3g,h could not be
detected.
General Procedure for the Synthesis of 3a–e: To a solution of 2H-
azaphosphirene complex 1 (617 mg, 1.00 mmol) in CH2Cl2 (3 mL)
was added the appropriate nitrile (2a: 122 µL, 1.50 mmol; 2b:
120 mg, 1.00 mmol; 2c: 180 µL, 2.06 mmol; 2d: 94 µL, 1.01 mmol;
2e: 94 µL, 1.03 mmol) and ferrocenium hexafluorophosphate
(17 mg, 0.05 mmol). The reaction mixture was stirred for 6 h at
ambient temperature [reaction monitoring by 31P{1H} NMR spec-
troscopy]. After the solvent was removed in vacuo, the product was
separated and purified by column chromatography (SiO2, –10 °C,
petroleum ether ether/Et2O, 97.5:2.5).
Complex 3a: Yield: 406 mg (0.59 mmol, 59%). Yellow solid, recrys-
tallized from n-pentane. M.p. 122 °C. 1H NMR (300.13 MHz,
C6D6, 30 °C): δ = –0.22 (s, 9 H, SiMe3), 0.50 (s, 9 H, SiMe3), 1.01
2
[d, JP,H = 5.8 Hz, 1 H, CH(SiMe3)2], 2.78 (s, 3 H, NMe), 2.92 (s,
The computational results can be summarized as follows:
The barrier of the nucleophilic attack of the nitrile at radi-
3 H, NMe), 7.11 (mc, 3 H, meta, para Ph), 8.17 (mc, 2 H, ortho Ph)
3
ppm. 13C{1H} NMR (75.5 MHz, C6D6, 30 °C): δ = 3.1 [d, JP,C
=
=
cal cationic complex 4·+ depends strongly on its nucleophi- 1.6 Hz, Si(CH3)3], 4.1 [d, JP,C = 2.6 Hz, Si(CH3)3], 22.0 [d, JP,C
3
1
5.8 Hz, CH(SiMe3)2], 37.6 (s, NCH3), 37.6 (s, NCH3), 128.8 (s, para
licity. As a result of steric factors in the real system of 1,
this step may be crucial. The further course of the ring ex-
pansion with strong nucleophilic nitrile derivatives[23] can
proceed by direct cyclization of the intermediately formed
complex 5·+ or with moderately nucleophilic nitrile deriva-
tives by formation of complex 6·+ accompanied by the re-
lease of nitrile and subsequent [3+2] cycloaddition. Because
the reactions are performed at ambient temperature, the ki-
netically controlled pathway proceeding via the more favor-
3
2
Ph), 131.6 (d, JP,C = 1.9 Hz, ortho Ph), 132.8 (d, JP,C = 20.7 Hz,
ipso Ph), 133.3 (s, meta Ph), 165.0 (s, PNC), 198.6 (dsat
,
2JP,C
=
1
2
6.5 Hz, JW,C = 126.7 Hz, cis CO), 199.6 (d, JP,C = 22.3 Hz, trans
[1+4]
CO), 200.3 (d,
J
P,C
= 25.5 Hz, PCN) ppm. 31P{1H} NMR
1
(121.5 MHz, C6D6, 30 °C): δ = 101.8 (ssat, JW,P = 241.6 Hz) ppm.
IR (KBr): ν = 2068 (CO), 1980 (CO), 1922 (CO) cm–1. MS (FAB+,
˜
mNBA): m/z (%) = 688.1 (12) [M + H]+, 659.1 (49) [M – CO]+,
631.1 (38) [M – 2 CO]+, 603.1 (10) [M – 3 CO]+, 364.2 (100) [M –
W(CO)5 + H]+. C22H30N3O5PSi2W (687.48): calcd. C 38.44, H
able transition state may be preferred, which yields regiose- 4.40, N 6.11; found C 38.49, H 4.54, N 5.86.
Eur. J. Inorg. Chem. 2007, 4669–4678
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4675

H. Helten, C. Neumann, A. Espinosa, P. G. Jones, M. Nieger, R. Streubel
FULL PAPER
2
Complex 3b: Yield: 398 mg (0.54 mmol, 54%). Orange solid, recrys-
tallized from n-pentane. M.p. 123 °C. 1H NMR (200 MHz, CDCl3,
25 °C): δ = –0.10 (s, 9 H, SiMe3), 0.52 (s, 9 H, SiMe3), 1.19 (d,
2JP,H = 3.9 Hz, 1 H, CHSiMe3), 2.37 (s, 3 H, CMe), 4.08 (s, 3 H,
NMe), 6.11 (d, 3JH,H = 3.9 Hz, 1 H, 3-pyrryl), 7.52 (mc, 5 H, meta,
25 °C): δ = 0.00 (s, 9 H, SiMe3), 0.64 (s, 9 H, SiMe3), 1.30 (d, JP,H
= 4.1 Hz, 1 H, CHSiMe3), 6.87 (mc, 1 H, 4-thienyl), 7.65 (mc, 3 H,
3
meta, para Ph), 7.99 (d, JH,H = 4.9 Hz, 1 H, 5-thienyl), 8.34 (d,
3JH,H = 6.8 Hz, 2 H, ortho Ph) ppm. 13C{1H} NMR (50.3 MHz,
3
3
25 °C, CDCl3): δ = 2.9 [d, JP,C = 1.9 Hz, Si(CH3)3], 3.7 [d, JP,C
=
4
3
para Ph, 4-pyrryl), 8.23 (dd, JH,H = 2.2 Hz, JH,H = 7.6 Hz, 2 H,
2.6 Hz, Si(CH3)3], 18.6 [d, 1JP,C = 4.8 Hz, CH(SiMe3)2], 126.6 (s, 2-
ortho Ph) ppm. 13C{1H} NMR (50.3 MHz, CDCl3, 25 °C): δ = 2.9
thienyl), 128.1 (s, 5-thienyl), 128.6 (s, meta Ph), 131.2 (d, JP,C
=
3
[d, 3JP,C = 2.0 Hz, Si(CH3)3], 3.7 [d, 3JP,C = 2.7 Hz, Si(CH3)3], 13.0 2.1 Hz, ortho Ph), 132.6 (d, 2JP,C = 23.8 Hz, ipso Ph), 133.4 (s, para
1
3
(s, CMe), 19.8 [d, JP,C = 5.4 Hz, CH(SiMe3)2], 109.0 (s, 4-pyrryl),
122.4 (s, 3-pyrryl), 127.6 (d, JP,C = 12.1 Hz, 2-pyrryl), 128.7 (s,
Ph), 133.6 (s, 4-thienyl), 137.2 (d, JP,C = 13.6 Hz, 3-thienyl), 165.0
2
[2+3]
(d,
(d,
J
J
= 4.9 Hz, PNC), 197.1 (d, 2JP,C = 5.9 Hz, cis CO), 198.2
P,C
3
2
[1+4]
2
meta Ph), 131.2 (d, JP,C = 2.0 Hz, ortho Ph), 132.4 (d, JP,C
=
= 22.4 Hz, PCN), 201.8 (d, JP,C = 22.8 Hz, trans CO)
P,C
23.5 Hz, ipso Ph), 132.9 (s, para Ph), 139.7 (s, 5-pyrryl), 162.3 (d, ppm. 31P{1H} NMR (81.0 MHz, CDCl3, 25 °C): δ = 109.1 (ssat
,
[2+3]
2
J
= 5.8 Hz, PNC), 197.5 (d, JP,C = 6.4 Hz, cis CO), 197.0
1JW,P = 229.8 Hz). IR (KBr): ν = 2073 (s, CO), 2000 (s, CO), 1921
˜
P,C
[1+4]
2
(d,
J
= 22.7 Hz, PCN), 198.8 (d, JP,C = 13.5 Hz, trans CO)
(s, sh, CO), 1253 (m, thienyl), 834 (w, thienyl) cm–1. MS (EI, 70 eV,
184W): m/z (%) = 726 (12) [M]+, 698 (26) [M – CO]+, 670 (100)
[M – 2 CO]+, 533 (26) [M – 3 CO – C5H3NS]+, 477 (38) [M – 5
CO – C5H3NS]+, 73 (74) [SiMe3]+. C24H27N2O5PSSi2W (726.53):
calcd. C 39.69, H 3.75, N 3.86, S 4.41; found C 39.34, H 3.98, N
P,C
ppm. 31P{1H} NMR (81.0 MHz, CDCl3, 25 °C): δ = 108.4 (ssat
,
1JW,P = 233.9 Hz) ppm. IR (KBr): ν = 2071 (CO), 1994 (CO), 1926
˜
(CO), 1873 (CO) cm–1. MS (EI, 70 eV, 184W): m/z (%) = 726 (8)
[M]+, 698 (16) [M – CO]+, 670 (32) [M – 2 CO]+, 533 (52) [M – 3
CO – C7H9N2]+, 73 (100) [SiMe3]+. C26H32N3O5PSi2W (737.53): 3.46, S 4.36.
calcd. C 42.22, H 4.51, N 5.48; found C 42.29, H 4.50, N 5.51.
Investigation of the Dependence of the Reaction Progression of 1
Complex 3c: Yield: 404 mg (0.57 mmol, 57%). Orange solid, recrys-
tallized from n-pentane. M.p. 119 °C. 1H NMR (200 MHz, CDCl3,
25 °C): δ = –0.10 (s, 9 H, SiMe3), 0.54 (s, 9 H, SiMe3), 1.16 (d, 1
with 2d on the Amount of Ferrocenium Hexafluorophosphate: To a
solution of 2H-azaphosphirene complex 1 (123 mg, 0.20 mmol) in
CH2Cl2 (0.6 mL) was added 2-thiophene carbonitrile 2d (19 µL,
0.20 mmol) and the appropriate amount of ferrocenium hexafluo-
2
3
H, JP,H = 4.1 Hz, 1 H, CHSiMe3), 6.64 (dd, 3JH,H = 3.5 Hz, JH,H
= 1.8 Hz, 2 H, 4-furanyl), 7.56 (mc, 4 H, meta, para Ph, 4-furanyl), rophosphate (0.7 mg, 0.002 mmol; 1.3 mg, 0.004 mmol; 2.0 mg,
3
4
7.72 (mc, JH,H = 1.8 Hz, 1 H, 5-furanyl), 8.21 (dd, 2 H, JH,H
=
0.006 mmol; 2.6 mg, 0.008 mmol; 3.3 mg, 0.010 mmol; 6.6 mg,
0.020 mmol; 9.9 mg, 0.030 mmol; 13.2 mg, 0.040 mmol). Product
ratios were estimated by 31P{1H} NMR spectroscopic resonance
integration (30 °C, 100 scans each, measurement duration 159 s,
2.7 Hz, JH,H = 5.9 Hz, 2 H, ortho Ph) ppm. 13C{1H} NMR
(50.3 MHz, CDCl3, 25 °C): δ = 2.9 [d, JP,C = 2.0 Hz, Si(CH3)3],
3.7 [d, JP,C = 2.8 Hz, Si(CH3)3], 18.8 [d, JP,C = 5.2 Hz, CH-
3
3
3
1
(SiMe3)2], 112.5 (s, 4-furanyl), 119.6 (s, 3-furanyl), 128.9 (s, meta recorded reaction time corresponds to the end of the respective
3
2
Ph), 131.3 (d, JP,C = 2.1 Hz, ortho Ph), 132.2 (d, JP,C = 22.7 Hz,
measurement). For the investigations corresponding to Figure 2, a
solution of [FcH]PF6 (1.6 mg, 0.005 mmol) and 2-thiophene car-
bonitrile 2d (19 µL, 0.20 mmol) in CH2Cl2 was added to 2H-aza-
phosphirene complex 1 (123 mg, 0.20 mmol) to ensure that the en-
tire amount of [FcH]PF6 was diluted at the start of the reaction.
3
ipso Ph), 133.6 (s, para Ph), 147.4 (s, 5-furanyl), 149.3 (d, JP,C
=
[2+3]
14.5 Hz, 2-furanyl), 160.2 (d,
J
P,C
[1+4]
= 3.8 Hz, PNC), 197.1 (d,
2JP,C = 6.1 Hz, cis CO), 198.2 (d,
J
P,C
= 22.4 Hz, PCN), 202.1
(d, JP,C = 22.7 Hz, trans CO) ppm. 31P{1H} NMR (81.0 MHz,
2
CDCl3, 25 °C): δ = 110.6 (ssat,
1JW,P = 230.6 Hz). MS (EI, 70 eV, Product ratios were estimated by 31P{1H} NMR spectroscopic res-
184W): m/z (%) = 710 (11) [M]+, 682 (11) [M – CO]+, 654 (85) [M –
onance integration (30 °C, 32 scans each, measurement duration
3 CO]+, 533 (52) [M – 3 CO – C5H3NO]+, 477 (52) [M – 5 CO – 56 s).
C5H3NO]+, 73 (100) [SiMe3]+. C24H27N2O6PSi2W (710.47): calcd.
Attempted Synthesis of Complex 3f with the Use of Ferrocenium
C 40.57, H 3.83, N 3.94; found C 40.72, H 3.83, N 3.94.
Hexafluorophosphate: To a solution of 2H-azaphosphirene complex
1 (123 mg, 0.20 mmol) in CH2Cl2 (0.6 mL) was added hydrogen
cyanide 2f (20 µL, 0.51 mmol) and ferrocenium hexafluorophos-
Complex 3d: Yield: 371 mg (0.51 mmol, 51%). Orange solid, recrys-
tallized from n-pentane. M.p. 112 °C. 1H NMR (200 MHz, CDCl3,
25 °C): δ = –0.09 (s, 9 H, SiMe3), 0.53 (s, 9 H, SiMe3), 1.19 [d, 2JP,H phate (12 mg, 0.04 mmol). The reaction mixture was stirred at am-
= 4.0 Hz, 1 H, CH(SiMe3)2], 7.24 (mc, 1 H, 3-thienyl), 7.57 (mc, 4 bient temperature [reaction monitored by 31P{1H} NMR spec-
H, meta, para Ph, 4-thienyl), 8.22 (mc, 1 H, 2-thienyl), 8.24 (dd,
troscopy]. After 4 d the reaction mixture containing unidentified
products A–F and 3f was analyzed by 31P{1H} NMR and 19F{1H}
NMR spectroscopy at 30 °C [ratio estimation by 31P{1H} NMR
resonance integration]. 31P{1H} NMR (121.5 MHz, CH2Cl2): δ =
3
4JH,H = 1.7 Hz, JH,H = 7.5 Hz, 2 H, ortho Ph) ppm. 13C{1H}
3
NMR (50.3 MHz, C6D6, 25 °C): δ = 2.9 [d, JP,C = 2.1 Hz, Si-
3
1
(CH3)3], 3.7 [d, JP,C = 2.6 Hz, Si(CH3)3], 18.9 [d, JP,C = 5.0 Hz,
CH(SiMe3)2], 128.5 (s, 4-thienyl), 128.9 (s, meta Ph), 131.4 (d, 3JP,C
206.9 [d, JP,F = 855.8 Hz, A (1%)], 206.2 [d, JP,F = 841.8 Hz, B
1
1
= 2.1 Hz, ortho Ph), 132.3 (d, 2JP,C = 22.5 Hz, ipso Ph), 133.1 (s, 5-
(3%)], 197.3 [d, JP,F = 824.0 Hz, C (9%)], 191.2 [d, JP,F =
1
1
3
thienyl), 133.6 (s, para Ph), 134.0 (s, 3-thienyl), 138.5 (d, JP,C
=
989.3 Hz, D (4%)], 161.2 [ssat
,
1JW,P = 289.9 Hz, E (6%)], 110.7
14.4 Hz, 2-thienyl), 164.4 (d,
J
P,C
= 4.0 Hz, PNC), 197.1 (d,
[ssat
,
1JW,P = 227.6 Hz, F (5%)], 105.5 [ssat
,
1JW,P = 225.1 Hz, 3f
[2+3]
2JP,C = 6.3 Hz, cis CO), 198.2 (d, 2JP,C = 22.4 Hz, trans CO), 201.3
(20%)], –109.9 [ssat, JW,P = 293.7 Hz, 1 (16%)], –143.5 [sept, JP,F
1
1
[1+4]
–
(d,
J
P,C
= 23.2 Hz, PCN) ppm. 31P{1H} NMR (81.0 MHz,
=
714.0 Hz, PF6 (3%)] ppm. 19F{1H} NMR (282.4 MHz,
1
1
1
C6D6, 25 °C): δ = 110.5 (ssat, JW,P = 229.5 Hz) ppm. IR (KBr): ν
CH2Cl2): δ = –31.0 [d, JP,F = 991.8 Hz, D (4%)], –72.5 [d, JP,F =
˜
–
1
= 2952 (CH), 2899 (CH), 2073 (CO), 2001 (CO), 1909 (CO), 1413 713.6 Hz, PF6 (3%)], –109.1 [d, JP,F = 841.5 Hz, B (3%)], –111.4
1
1
(thienyl), 1253 (thienyl) cm–1. MS (EI, 70 eV, 184W): m/z (%) = 726 [d, JP,F = 854.9 Hz, A (1%)], –117.2 [d, JP,F = 823.5 Hz, C (9%)]
(8) [M]+, 698 (16) [M – CO]+, 670 (32) [M – 2 CO]+, 73 (100) ppm.
[SiMe3]+. C24H27N2O5PSSi2W (726.53): calcd. C 39.68, H 3.75, N
Attempted Synthesis of Complex 3g with the Use of Ferrocenium
Hexafluorophosphate: To a solution of 2H-azaphosphirene complex
3.86, S 4.41; found C 39.82, H 3.70, N 4.64, S 3.79.
Complex 3e: Yield: 378 mg (0.52 mmol, 52%). Orange solid, recrys-
tallized from n-pentane. M.p. 112 °C. 1H NMR (200 MHz, CDCl3,
1 (123 mg, 0.20 mmol) in CH2Cl2 (0.6 mL) was added ethyl cyano-
formate 2g (30 µL, 0.30 mmol) and ferrocenium hexafluorophos-
4676
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 4669–4678

Reactivity of a 17e Radical Cationic 2H-Azaphosphirene Complex
phate (12 mg, 0.04 mmol). The reaction mixture was stirred at am-
bient temperature [reaction monitored by 31P{1H} NMR spec-
troscopy]. After 6 d, the reaction mixture was analyzed by 31P{1H}
NMR and 19F{1H} NMR spectroscopy at 30 °C [ratio estimation
by 31P{1H} NMR resonance integration]. 31P{1H} NMR
163.3 [C (1%)], 110.7 [D (1%)], 97.4 [E (1%)], 54.5 [F (1%)], 34.1
[G (10%)], –110.1 [ssat, JW,P = 293.7 Hz, 1 (81%)] ppm.
1
Attempted Synthesis of Complex 3h with the Use of Ferrocenium
Tetraphenylborate: To a solution of 2H-azaphosphirene complex 1
(123 mg, 0.20 mmol) in CH2Cl2 (0.6 mL) was added 2,3,4,5,6-
pentafluorobenzonitrile 2h (160 µL, 1.27 mmol) and ferrocenium
hexafluorophosphate (18 mg, 0.04 mmol). The reaction mixture
was stirred at ambient temperature [reaction monitored by 31P{1H}
NMR spectroscopy]. After 6 d the reaction mixture containing un-
identified products A–H and 2h was analyzed by 31P{1H} NMR
and 19F{1H} NMR spectroscopy at 30 °C [ratio estimation by
31P{1H} NMR resonance integration]. 31P{1H} NMR (121.5 MHz,
1
(121.5 MHz, CH2Cl2): δ = 197.3 [d, JP,F = 822.7 Hz, A (12%)],
1
177.2 [d, JP,F = 967.7 Hz, B (2%)], 160.8 [ssat
,
1JW,P = 284.8 Hz,
1
C (6%)], 110.7 [ssat
,
1JW,P = 227.6 Hz, D (21%)], 18.5 [d, JP,F
=
1044.0 Hz, E (3%)], –110.1 [ssat, 1JW,P = 293.7 Hz, 1 (41%)], –143.8
1
–
[sept, JP,F
=
714.0 Hz, PF6 (6%)] ppm. 19F{1H} NMR
1
(282.4 MHz, CH2Cl2): δ = –66.7 [d, JP,F = 1046.8 Hz, E (3%)],
1
–
1
–73.8 [d, JP,F = 712.5 Hz, PF6 (6%)], –74.4 [d, JP,F = 968.3 Hz,
1
B (2%)], –117.2 [d, JP,F = 823.5 Hz, A (12%)] ppm.
CH2Cl2): δ = 183.0 [ssat
,
1JW,P = 254.3 Hz, A (9%)], 176.2 [ssat
,
1
1JW,P = 277.2 Hz, B (3%)], 172.5 [ssat, JW,P = 309.0 Hz, C (6%)],
126.1 [D (1%)], 110.7 [E (1%)], 95.5 [ssat, 1JW,P = 277.2 Hz, F (3%)],
Attempted Synthesis of Complex 3h with the Use of Ferrocenium
Hexafluorophosphate: To a solution of 2H-azaphosphirene complex
1 (123 mg, 0.20 mmol) in CH2Cl2 (0.6 mL) was added 2,3,4,5,6-
pentafluorobenzonitrile 2h (160 µL, 1.27 mmol) and ferrocenium
hexafluorophosphate (12 mg, 0.04 mmol). The reaction mixture
was stirred at ambient temperature [reaction monitored by 31P{1H}
NMR spectroscopy]. After 6 d the reaction mixture containing un-
identified products A–F, 1, and 2h was analyzed by 31P{1H} NMR
and 19F{1H} NMR spectroscopy at 30 °C [ratio estimation by
31P{1H} NMR resonance integration]. 31P{1H} NMR (121.5 MHz,
54.4 [ssat, ,
1JW,P = 270.8 Hz, G (1%)], 40.0 [H (1%)], –110.4 [ssat
1JW,P = 295.0 Hz, 1 (69%)] ppm. 19F{1H} NMR: δ = –134.0 (mc,
2h), –144.7 (mc, 2h), –160.6 (mc, 2h) ppm.
Theoretical Methods: DFT calculations were carried out with the
TURBOMOLE V5.8 program package.[24] For optimizations the
gradient corrected exchange functional by Becke[25] (B88) in combi-
nation with the gradient corrected correlation functional by Lee,
Yang, and Parr[26] (LYP) with the RI approximation[27] and the
valence-double-ζ basis set SV(P)[28] was used. For tungsten the ef-
fective core potential ECP-60-MWB[29] derived from the Stuttgart-
Dresden group was used. The influence of the polar solvent was
taken into account by employing the COSMO approach[30] with ε
= 8.93. For cavity construction the atomic radii of Bondi[31] were
used, which are obtained from crystallographic data. For tungsten
the atomic radius was set to 2.2230 Å. The stationary points were
characterized by numerical vibrational frequencies calculations by
using the SNF 3.3.0 program package.[32] Single point calculations
were carried out by using the Three Parameter Hybrid Functional
Becke3[33] (B3) in combination with the LYP correlation func-
tional[26] and the valence-triple-ζ basis set TZVP[34] and ECP-60-
MWB for tungsten.
1
1
CH2Cl2): δ = 197.4 [d, JP,F = 824.0 Hz, JW,P = 286.1 Hz, A
(11%)], 183.1 [s, B (5%)], 176.3 [s, C (5%)], 160.9 [ssat
,
1JW,P
=
1
282.3 Hz, D (5%)], 110.7 [ssat, JW,P = 228.9 Hz, E (12%)], 18.4 [d,
1JP,F = 1041.4 Hz, F (3%)], –110.4 [ssat, 1JW,P = 293.7 Hz, 1 (48%)]
ppm. 19F{1H} NMR (282.4 MHz, CH2Cl2): δ = –67.6 [d, JP,F
=
1
1
–
1041.2 Hz, F (3%)], –75.5 (d, JP,F = 713.6 Hz, PF6 ), –117.6 [d,
1JP,F = 823.5 Hz, A (11%)], –134.0 (mc, 2h), –144.7 (mc, 2h), –160.6
(mc, 2h) ppm.
Synthesis of 3d with the Use of Ferrocenium Tetraphenylborate: To
a solution of 2H-azaphosphirene complex 1 (62 mg, 0.10 mmol) in
CH2Cl2 (0.3 mL) was added 2-thiophene carbonitrile 2d (10 µL,
0.11 mmol) and ferrocenium tetraphenylborate (9 mg, 0.02 mmol).
The reaction mixture was stirred for 8 h at ambient temperature,
and the complete formation of 3d was evidenced by analysis of the
reaction mixture by 31P{1H} NMR spectroscopy.
Supporting Information (see footnote on the first page of this arti-
cle): Relative energies, enthalpies, and free energies for the reactions
depicted in Schemes 4 and 5.
Attempted Synthesis of Complex 3f with the Use of Ferrocenium
Tetraphenylborate: To a solution of 2H-azaphosphirene complex 1
(123 mg, 0.20 mmol) in CH2Cl2 (0.6 mL) was added hydrogen cya-
nide 2f (20 µL, 0.51 mmol) and ferrocenium tetraphenylborate
(18 mg, 0.04 mmol). The reaction mixture was stirred at ambient
temperature [reaction monitored by 31P{1H} NMR spectroscopy].
After 4 d the reaction mixture containing unidentified products A–
D and 1 was analyzed by 31P{1H} NMR spectroscopy at 30 °C
[ratio estimation by 31P{1H} NMR resonance integration]. 31P{1H}
Acknowledgments
We are grateful to the Deutsche Forschungsgemeinschaft, the
Fonds der Chemischen Industrie (Kekulé Stipendium for H. H.),
the SFB 624 “Template”, the John von Neumann Institute for
Computing (HBN12), and Prof. Dr. B. Engels for computing time.
A. E. wishes to acknowledge the grants from MEC-Spain (CTQ
2004–02201) and Fundación Séneca (CARM-Spain) (02970/PI/05).
We thank G. von Frantzius for fruitful discussions.
NMR (121.5 MHz, CH2Cl2): δ = 110.6 [ssat
,
1JW,P = 227.6 Hz, A
1JW,P
(13%)], 105.5 [ssat
,
1JW,P = 223.8 Hz, 3f (3%)], 20.7 [ssat
,
=
236.5 Hz, B (3%)], 13.2 [ssat, 1JW,P = 236.5 Hz, C (8%)], –68.9 [ssat
,
1JW,P = 242.9 Hz, D (17%)], –109.9 [ssat, 1JW,P = 293.7 Hz, 1 (37%)]
ppm.
[1] M. Schmittel, A. Burghart, Angew. Chem. 1997, 109, 2658–
2699; Angew. Chem. Int. Ed. Engl. 1997, 36, 2550–2589.
[2] E. Baciocchi, M. Bietti, O. Lanzalunga, Acc. Chem. Res. 2000,
33, 243–251.
[3] M. Mella, M. Fagnoni, M. Freccero, A. Albini, Chem. Soc.
Rev. 1998, 27, 81–89.
[4] A. A. Fokin, P. R. Schreiner, Chem. Rev. 2002, 102, 1551–1593.
[5] N. G. Connelly, W. E. Geiger, Chem. Rev. 1996, 96, 877–910.
[6] C. Gaebert, J. Mattay, M. Toubartz, S. Steenken, B. Müller, T.
Bally, Chem. Eur. J. 2005, 11, 1294–1304.
[7] K. E. Torraca, L. McElwee-White, Coord. Chem. Rev. 2000,
206–207, 469–491.
Attempted Synthesis of Complex 3g with the Use of Ferrocenium
Tetraphenylborate: To a solution of 2H-azaphosphirene complex 1
(123 mg, 0.20 mmol) in CH2Cl2 (0.6 mL) was added ethyl cyanofor-
mate 2g (30 µL, 0.30 mmol) and ferrocenium tetraphenylborate
(18 mg, 0.04 mmol). The reaction mixture was stirred at ambient
temperature [reaction monitored by 31P{1H} NMR spectroscopy].
After 6 d the reaction mixture containing unidentified products A–
G and 1 was analyzed by 31P{1H} NMR spectroscopy at 30 °C
[ratio estimation by 31P{1H} NMR resonance integration]. 31P{1H}
NMR (121.5 MHz, CH2Cl2): δ = 204.9 [A (1%)], 184.5 [B (1%)],
Eur. J. Inorg. Chem. 2007, 4669–4678
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4677

H. Helten, C. Neumann, A. Espinosa, P. G. Jones, M. Nieger, R. Streubel
FULL PAPER
[8] For example, see: U. Jahn, P. Hartmann, Chem. Commun. 1998,
209–210, and references cited therein.
tions = 14224 (6590), Rint = 0. 0716, µ = 4.124 mm–1, 318 re-
fined parameters, 51 restraints, R1 [for IϾ2σ(I)] = 0. 0361, wR2
(for all data) = 0.0935, max./min. residual electron density =
2.745/–3.017 eÅ–3. X-ray crystallographic analysis of 3b, 3e:
Crystals were obtained from a concentrated n-pentane solu-
tion. Data were collected with a Bruker SMART 1000 CCD
diffractometer equipped with a low-temperature device. Struc-
ture solution and refinement as above. 3b: C26H32N3O5PSi2W;
[9] T. Ramnial, I. McKenzie, B. Gorodetsky, E. M. W. Tsang,
J. A. C. Clyburne, Chem. Commun. 2004, 1054–1055.
[10] a) R. Streubel, A. Kusenberg, J. Jeske, P. G. Jones, Angew.
Chem. 1994, 106, 2564–2566; Angew. Chem. Int. Ed. Engl. 1994,
33, 2427–2429; b) R. Streubel, J. Jeske, P. G. Jones, R. Herbst-
Irmer, Angew. Chem. 1994, 106, 115–117; Angew. Chem. Int.
Ed. Engl. 1994, 33, 80–82; c) R. Streubel, A. Kusenberg, Phos-
phorus Sulfur Relat. Elem. 1994, 93–94, 281–284; d) R. Streu-
bel, A. Ostrowski, S. Priemer, U. Rohde, J. Jeske, P. G. Jones,
Eur. J. Inorg. Chem. 1998, 257–261.
[11] a) F. G. N. Cloke, P. B. Hitchcock, J. F. Nixon, U. Schiemann,
R. Streubel, D. J. Wilson, Chem. Commun. 2000, 1659–1660;
b) H. Wilkens, A. Ostrowski, J. Jeske, F. Ruthe, P. G. Jones, R.
Streubel, Organometallics 1999, 18, 5627–5642; c) R. Streubel,
H. Wilkens, P. G. Jones, Chem. Commun. 1999, 2127–2128; d)
H. Wilkens, F. Ruthe, P. G. Jones, R. Streubel, Chem. Eur. J.
1998, 4, 1542–1553; e) R. Streubel, H. Wilkens, A. Ostrowski,
C. Neumann, F. Ruthe, P. G. Jones, Angew. Chem. 1997, 109,
1549–1550; Angew. Chem. Int. Ed. Engl. 1997, 36, 1492–1493.
[12] R. Streubel, H. Wilkens, P. G. Jones, Chem. Eur. J. 2000, 6,
3997–4000.
[13] a) R. Streubel, C. Neumann, P. G. Jones, J. Chem. Soc. Dalton
Trans. 2000, 2495–2496; b) C. Neumann, E. Ionescu, U. Schie-
mann, M. Schlenker, M. Bode, F. Ruthe, P. G. Jones, R. Streu-
bel, J. Organomet. Chem. 2002, 643–644, 253–264; c) C. Neum-
ann, A. Prehn Junquera, C. Wismach, P. G. Jones, R. Streubel,
Tetrahedron 2003, 59, 6213–6220.
[14] The numbering of atoms in heterocyclic complexes 3a–h ac-
cording to the Hantzsch-Widmann-Patterson nomenclature is
used in this paper.
¯
crystal size 0.19ϫ0.14ϫ0.06 mm, triclinic, P1 (No.2), a =
10.823(2) Å, b = 12.207(2) Å, c = 13.395(3) Å, α = 96.101(6)°,
β = 110.126(6)°, γ = 107.666(6)°, V = 1538.7(5) Å3, Z = 2,
ρcalcd. = 1.592 Mgm–3, 2θmax = 60°, collected (independent) re-
flections = 30876 (8933), Rint = 0.0501, µ = 3.92 mm–1, 351
refined parameters, 63 restraints, R1 [for IϾ2σ(I)] = 0. 0451,
wR2 (for all data) = 0.1125, max./min. residual electron density
=
3.80/–5.13 eÅ–3. 3e: C24H27N2O5PSSi2W; crystal size
¯
0.35ϫ0.23ϫ0.16 mm, triclinic, P1 (No. 2), a = 10.8501(12) Å,
b = 12.2138(12) Å, c = 13.1138(14) Å, α = 96.171(3)°, β =
110.016(3)°, γ = 110.191(3)°, V = 1481.5(3) Å3, Z = 2, ρcalcd.
=
1.629 Mgm–3, 2θmax = 61°, collected (independent) reflections
= 24334 (8967), Rint = 0. 0250, µ = 4.14 mm–1, 331 refined
parameters, 31 restraints, R1 [for IϾ2σ(I)] = 0. 0179, wR2 (for
all data) = 0.0455, max./min. residual electron density =
1.417/–1.308 eÅ–3. CCDC-646819 (for 3b), -639472 (for 3d),
and -646820 (for 3e) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[20] a) SHELXS-97: G. M. Sheldrick, Acta Crystallogr., Sect. A
1990, 46, 467–473; b) G. M. Sheldrick, SHELXL-97, Univer-
sity of Göttingen, 1997.
[21] H. Helten, T. Schirmeister, B. Engels, J. Phys. Chem. A 2004,
[15] P. W. Atkins, Physical Chemistry, 5th ed., Oxford University
Press, Oxford, 1994.
[16] Cyclovoltammetric measurements did not reveal conclusive re-
sults.
[17] M. A. Alvarez, G. García, M. E. García, V. Riera, M. A. Ruiz,
Organometallics 1999, 18, 4509–4517.
108, 7691–7701.
[22] R. Streubel, Top. Curr. Chem. 2002, 223, 91–101.
[23] The model reaction using 4 and H2NCN may be representative
for the reaction of 1 with Me2NCN (2a), but also for the reac-
tions with the electron-rich five-membered heterocyclic nitrile
derivatives 2b–e.
[18] The ease of electrophilic aromatic substitution increases in the
order: thiopheneϽfuraneϽpyrrol. For example, see: D. T.
[24] TURBOMOLE V5.8: see http://www.cosmologic.de/Quantum-
Chemistry/ main_turbomole.html.
Davies, Aromatic Heterocyclic Chemistry, Oxford University [25] A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100.
Press, Oxford, 1992.
[26] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[27] a) K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs,
Chem. Phys. Lett. 1995, 240, 283–290; b) K. Eichkorn, O.
Treutler, H. Öhm, M. Häser, R. Ahlrichs, Chem. Phys. Lett.
1995, 242, 652–660; c) K. Eichkorn, F. Weigend, O. Treutler,
R. Ahlrichs, Theor. Chem. Acc. 1997, 97, 119–124.
[28] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97,
2571–2577.
[29] D. Andrae, U. Häußermann, M. Dolg, H. Stoll, H. Preuß,
Theor. Chim. Acta 1990, 77, 123–141.
[30] A. Klamt, G. Schürmann, J. Chem. Soc. Perkin Trans. 2 1993,
799–805.
[31] A. Bondi, J. Phys. Chem. 1964, 68, 441–451.
[32] J. Neugebauer, M. Reiher, C. Kind, B. A. Hess, J. Comput.
Chem. 2002, 23, 895–910.
[33] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[34] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100,
5829–5835.
[19] X-ray crystallographic analysis of 3d: Suitable yellow single
crystals of 3d were obtained from a concentrated n-pentane
solution upon decreasing the temperature from ambient tem-
perature to +4 °C. Data were collected with a Nonius Kap-
paCCD diffractometer equipped with a low-temperature device
(Cryostream, Oxford Cryosystems) at 123 K by using graphite
monochromated Mo-Kα radiation (λ = 0.71073 Å). The struc-
ture was solved by Patterson methods (SHELXS-97)[20a] and
refined by full-matrix least-squares on F2 (SHELXL-97).[20b]
All non-hydrogen atoms were refined anisotropically. The hy-
drogen atoms were included isotropically by using the riding
model on the bound atoms. The thiophene substituent is disor-
dered [occupancy: 0.81(1): 0.19(1)]. Semiempirical absorption
correction was carried out from equivalents (min./max. trans-
missions = 0.27295/0.61685). C24H27N2O5PSSi2W; crystal size
¯
0.45ϫ0.30ϫ0.12 mm, triclinic, P1 (No.2), a = 10.9114(2) Å, b
= 12.1675(2) Å, c = 13.1114(2) Å, α = 96.695(1)°, β =
110.108(1)°, γ = 109.554(1)°, V = 1486.63(4) Å3, Z = 2, ρcalcd.
= 1.623 Mgm–3, 2θmax = 55°, collected (independent) reflec-
Received: June 27, 2007
Published Online: August 22, 2007
4678
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 4669–4678