Direct Access to Inversely Polarized Phosphaalkenes from Elemental Phosphorus or Polyphosphides

Article abstract of DOI:10.1002/ejic.201501017

Phosphorus-containing multiple-bond systems have received great interest in various applications but often require elaborate syntheses and special precursors. In this paper, we describe simple methods for the synthesis of imidazoyl phosphinidenes and bis(

Full text of DOI:10.1002/ejic.201501017

DOI: 10.1002/ejic.201501017
Full Paper
CLUSTER
ISSUE
Phosphaalkenes
Direct Access to Inversely Polarized Phosphaalkenes from
Elemental Phosphorus or Polyphosphides
Mario Cicač-Hudi,[a] Johannes Bender,[a] Simon Hugo Schlindwein,[a] Mark Bispinghoff,[b]
Martin Nieger,[c] Hansjörg Grützmacher,*[b] and Dietrich Gudat*[a]
Abstract: Phosphorus-containing multiple-bond systems have ides. NMR spectroscopy studies revealed the formation of a
received great interest in various applications but often require bis(imidazoyl)-substituted tetraphosphatriene and the bicyclic
elaborate syntheses and special precursors. In this paper, we anion HP4– as transient intermediates and indicated the partici-
describe simple methods for the synthesis of imidazoyl phos- pation of tBuOH in the formation of imidazoyl phosphinidenes.
phinidenes and bis(imidazolyl)–P(I) halides from elemental Na3P7 and (Me3Si)3P7 can be obtained from P4 and Pred in excel-
phosphorus or the heptaphosphides Na3P7 and (Me3Si)3P7. The lent yields by a safe and simple method and are versatile pre-
reactions of imidazolium salts with KOtBu and P4 afford mix- cursors for the synthesis of imidazoyl phosphinidenes with
tures of imidazoyl phosphinidenes and Pn compounds and, for varying steric demand.
N-methylated imidazolium salts, also bis(imidazolyl)–P(I) hal-
Introduction
Phosphorus-containing multiple-bond systems have experi-
enced an amazing evolution during recent decades: included
in the elusive “nonexistent compounds” up to the 1960s,[1] they
can now be considered one of the most advanced classes of
compounds with π-bonds between heavier main-group ele-
ments,[2] and their application is receiving interest in diverse
fields such as catalysis[3] and π-conjugated materials.[4] A pre-
requisite for this development was that chemists learned to
Scheme 1. Prototypes of phosphorus-containing double-bond systems: phos-
phamethine cyanines (1, R = Me, Et) and phosphaalkenes 2 (R1 = Me, Cy, tBu,
Ph) and 3.
overcome the low π-bond stability that renders multiple-bond
systems unstable with respect to dimerization or addition reac-
tions.[2] This is often achieved by introducing sterically protect-
ing substituents (kinetic stabilization),[5] electronic stabilization
by π delocalization or a combination of both. Prototypes illus-
trating the stabilization of phosphorus-containing double-bond
systems include phosphamethine cyanines 1[6] (the first isolable
phosphorus–carbon multiple-bond systems ever prepared) and
the first isolable acyclic phosphaalkene 2 (Scheme 1).[8] The sta-
bility of 1 is generally related to the π delocalization associated
with the resonance between canonical structures 1′ and 1′′,[6]
but the presence and delocalization of charge may also play a
role.[7]
The stability of 2 may be related to a combination of steric
protection and the attenuation of the phosphorus electrophilic-
ity by the π-donating siloxy group. This π-donor stabilization
can be increased by introducing two donating amino groups
to afford 3,[9] which is stable even in the absence of steric
shielding.
The need to provide sufficient protection for the π-bonds
implies that phosphorus-containing multiple-bond systems
must often be accessed by elaborate, multistep syntheses or
through the utilization of specially designed precursors, which
may be laborious to prepare. In this work, we report the synthe-
sis of sterically unprotected low-coordinate phosphorus com-
pounds with similar structures to those of 1 and 3 in a one-
step process involving the direct activation of white phos-
phorus.[10] As an alternative entry to these compounds, we also
report a simple and safe synthesis of the heptaphosphides
Na3P7 and (Me3Si)3P7.
[a] Institut für Anorganische Chemie, University of Stuttgart,
Pfaffenwaldring 55, 70550 Stuttgart, Germany
E-mail: gudat@iac.uni-stuttgart.de
http://www.iac.uni-stuttgart.de/akgudat
[b] Department of Chemistry and Applied Biosciences, ETH Zürich,
Vladimir-Prelog-Weg 1, 8093 Zürich, Switzerland
[c] Laboratory of Inorganic Chemistry, Department of Chemistry, University of
Helsinki,
Our studies were inspired by the ground-breaking reports by
the group of Bertrand on the activation of P4 with various types
of stable amino-substituted carbenes.[11–15] Depending on the
P. O. Box 55, 00014 University of Helsinki, Finland
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejic.201501017.
Eur. J. Inorg. Chem. 2016, 649–658
649
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
steric demand and “philicity” of the carbenes, these reactions addressed as bis(carbene) complexes of a univalent P(I) cat-
yield transient cyclic diphosphenes A[11] or polycycles B[14] as ion.[17,18] In addition, solid-state 31P and 13C NMR spectroscopy
initial products (Scheme 2). The former are not directly observ- studies disclosed the presence of species tentatively assigned
able but can be trapped by cycloaddition with a diene or an as organic polyphosphorus compounds.
excess of the carbene.[11,13] The trapping products C and D may
in turn undergo further aggregation or reaction with P4 to yield
species with P8 or P12 scaffolds.[12,14] With less bulky carbenes,
the fragmentation of the P4 skeleton to afford products with P1
to P3 cores (e.g., structures E and F) has also been observed.[13]
We have now modified this approach by replacing the pre-
formed carbene with a mixture of an imidazolium salt and a
strong anion base. Although the combinations of these rea-
gents produces N-heterocyclic carbenes (NHCs),[16] their reactiv-
ity towards P4 does not simply mirror the previously reported
behaviour of NHCs[11–14] but offers access to a new and unprec-
edented reaction channel under these conditions.
Scheme 3. Products observed in the reactions of imidazolium salts 4a–4d[X]
with P4 and KOtBu (4a[X]–6a[X]: X = I, R = Me, R′ = H; 4b[X]–6b[X]: X = I, R =
Me, R′ = Me; 4c[X]–5c: X = Cl, R = mesityl, R′ = H; 4d[X]–5d: X = Cl, R = 2,6-
diisopropylphenyl, R′ = H).
In the reactions of the N-arylated imidazolium salts 4c[Cl]
and 4d[Cl], the phosphaalkenes 5c and 5d as well as soluble
and insoluble organophosphorus materials were observed but
bis(carbene)-stabilized P(I) cations were not. Presumably, their
formation is disfavoured by the higher steric demand of the N-
aryl substituents. Phosphaalkene 5d has been prepared previ-
ously by phosphinidene transfer from a phosphasilene to a
carbene,[19] the reaction of imidazolium salts with Na(OCP) or
(Me3Si)3P7,[20] or the methanolysis of a P-silylated precursor.[24]
The phosphaalkenes (“imidazoyl phosphinidenes”) 5a–5c
were separated by evaporation of the supernatant solutions
and extraction of the residue with benzene and purified by
sublimation. Phosphaalkene 5d was identified only by spectro-
scopic data but could not be isolated. The salts 6a[I] and 6b[I]
were isolated by extraction of the precipitate with CH2Cl2 and
crystallization. The salt 6a[I] was further converted to 6a[OTf]
(OTf = triflate) by anion exchange with Me3SiOTf. The yields
with respect to 4a[I] obtained under optimized reaction condi-
tions of 5a (23 %) and 6a[I] (50 %) indicate that both species
are formed in roughly equimolar amounts and that 73 % of
the imidazolium salts are converted into these products. Even
though no formation of 6c[Cl] was observed in the reaction of
4c[Cl], the phosphaalkene 5c was isolated in a similar yield of
25 %.
Scheme 2. Pathways for the breakdown of P4 by reaction with various types
of stable carbenes (R2C = cyclic or acyclic diaminocarbenes, cyclic amino alkyl
carbenes or diaminocyclopropylidene).[11–14]
Results and Discussion
Three-component reactions between white phosphorus and
equimolar amounts of an imidazolium salt and KOtBu were con-
ducted either by dissolving all of the components together or
by reacting the imidazolium salt and the base first to give an
NHC and tBuOH and then adding P4. Tetrahydrofuran (THF) was
the most suitable solvent, but the reactions also occurred in
diethyl ether or toluene. The resulting mixtures were stirred for
12 to 24 h to give suspensions, which were separated by filtra-
tion. The analysis of the supernatant solutions and the reddish
As white phosphorus is dangerous to handle and difficult to
precipitates by solution or solid-state 31P NMR spectroscopy in- obtain commercially, we investigated different routes to these
dicated that the product distribution depends on the N substit- phosphaalkenes. Na(OCP) and (Me3Si)3P7 have been used previ-
uents of the imidazolium salt but not on the order in which the ously as phosphorus sources in their synthesis.[20,21] However,
components are added.
this approach is limited to N-arylated derivatives. The N-methyl-
and N-isopropylimidazolium iodides or chlorides do not react
with Na(OCP) or (Me3Si)3P7 even at temperatures above 80 °C.
Furthermore, the literature-known synthesis of (Me3Si)3P7 relies
on the use of Na/K alloy, a dangerous reagent unsuitable for
large-scale synthesis.[22] The first step in this procedure is the
formation of the trimetallated P7 cage, which led us to explore
the direct use of trimetal heptaphosphides for the synthesis of
imidazoyl phosphinidenes.
For the N-methylated derivatives 4a[I] and 4b[I], the superna-
tant solutions contained phosphaalkenes 5a and 5b (Scheme 3)
as the dominant products. In addition, 31P NMR spectra re-
corded with long acquisition times revealed low-intensity sig-
nals with complex splitting patterns, which indicated the pres-
ence of soluble, unidentified polyphosphorus compounds. In
some reactions, minor amounts of the cations 6a and 6b (see
below) were also detected. In addition to inorganic salts, the
reddish precipitates contained ionic 6a[I] and 6b[I] as the major
To circumvent the use of Na/K alloy in the heptaphosphide
phosphorus-containing components. Species of this type are synthesis but still achieve a high conversion rate in the reaction
analoguesofthephosphamethinecyanine1[6] buthavealsobeen between sodium and elemental phosphorus, we employed cat-
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
650
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
alytic amounts of naphthalene. Thus, sodium can be used as
the only reductant, and Na3P7 is obtained in quantitative yield
by simply stirring a suspension of white phosphorus, elemental
sodium, and naphthalene (1 mol-% per P) in a 1:1 mixture of
dimethoxyethane (DME) and THF for 3 d at 20 °C (Scheme 4).
It can also be obtained in equally high yields from red phos-
phorus. In that case, the suspension has to be stirred for 2 d at
20 °C and then heated under reflux for 24 h. In both cases, the
orange Na3P7 was collected by filtration, washed with n-hexane
and dried thoroughly. Note that even after 3 d under reduced
pressure and elevated temperature (70 °C), the product still
contains 12 mol-% DME.
Scheme 5. Synthesis of N-alkylated and N-arylated imidazoyl phosphinidenes
5a and 5c–5e from the corresponding imidazolium chlorides and Na3P7. The
imidazolidinyl phosphinidene 5f can be obtained in the same way from the
dihydroimidazolium chloride 4f[Cl]; R = Me (4a, 5a), mesityl (4c, 5c), 2,6-
diisopropylphenyl (4d,f, 5d,f), iPr (4e, 5e).
ably more shielded than those of acyclic C-amino phosphaalk-
enes (cf. δ31P = –62.6 ppm for 3),[9] the presence of a P–C dou-
ble bond is corroborated by the deshielded 13C NMR resonance
1
of the α-carbon atom (δ13C = 178 to 181 ppm). The 13C and H
NMR spectra show only one set of resonances for the nuclei of
the CH/CMe and NR units on both sides of the P–C double
bond; therefore, the barrier for the E/Z isomerization is very low.
The 31P NMR signals of the cations 6a and 6b (δ31P = –113
to –117 ppm) are less shielded than those of the neutral 5a–
5d and match those of known acyclic bis(NHC)–P(I) cations.[17]
The 31P magic-angle spinning (MAS) NMR spectra of the insolu-
ble residues contain a broad, featureless resonance between
Scheme 4. High-yielding, safe and simple synthesis of Na3P7 and (Me3Si)3P7
from elemental phosphorus by employing catalytic amounts of naphthalene.
The insoluble Na3P7 can be converted to soluble
(Me3Si)3P7 in 94 % yield by treating it with trimethylsilyl chlor-
ide in toluene at –80 °C, followed by filtration to remove the
NaCl byproduct and recrystallization from toluene.
We found Na3P7 to be a versatile phosphorus source for the δ = –200 to 200 ppm and a maximum at δ ≈ –2 ppm. The
synthesis of imidazoyl and imidazolidinyl phosphinidenes. The extension of this signal reflects the wide range of chemical
treatment of suspensions of the imidazolium salts 4a[Cl] and shifts known for polyphosphorus frameworks,[25] and the inabil-
4c[Cl]–4f[Cl] in THF with Na3P7 followed by extraction with tolu- ity to resolve individual signals points to a highly disordered
ene and recrystallization or sublimation gives the phosphaalk- structure with a low degree of crystallinity. The 13C MAS spectra
enes 5a and 5c–5f in 60 to 80 % yield based on the imid~azol- of the insoluble solids contain similarly broadened resonances.
ium salts (Scheme 5). The procedure works equally well for N- The observed chemical-shift patterns indicate the presence of
alkylated and N-arylated imidazolium salts, and the only differ- NHC units together with minor amounts of other unidentified
ence is that the former need to be heated under reflux for 16 h, structural motifs. Further assignments, in particular of a possible
whereas the latter react at 20 °C. Imidazoyl phosphinidene 5a connection between the carbon and phosphorus building
can also be obtained from the imidazolium iodide 4a[I], albeit blocks, remained elusive.
in considerably lower yield (25 vs. 71 %).
The structural assignments derived from the spectroscopic
The reaction products 5a–5f and 6a[I] and 6b[I] were identi- studies were confirmed by single-crystal XRD studies of 5a–5c
fied unambiguously from analytical and spectroscopic data. The as well as 6a[OTf][26] and 6b[I]. For 5c, a qualified discussion of
phosphaalkenes 5a–5c and 5e exhibit more shielded 31P NMR the metrical parameters is impeded by an orientational disor-
resonances (δ = –147 to –149 ppm) compared with those of der, which becomes evident in an elongation of the thermal
the previously reported 5d (δ = –136.7 ppm, Table 1).[20] The ellipsoid of the P1 atom along the direction of the P–H bond
same trend in the 31P NMR shifts was observed by Bertrand and and the location of the P-bound hydrogen atom at a split posi-
co-workers for imidazoyl phenylphosphinidenes and attributed tion on either side of the P1–C1 bond (see Figure S9 in the
to the different π-accepting properties of the carbenes.[23] Al- Supporting Information). We interpret these peculiarities as re-
though the 31P NMR resonances of all five species are consider- sulting from the superposition of two molecules with different
1
Table 1. Comparison of 31P NMR shifts [ppm] and JP, H coupling constants [Hz] as well as selected bond lengths [Å] and angles [°] of the imidazoyl phosphin-
idenes 5a–5e, 7 and 8 (in C6D6); Dipp = 2,6-diisopropylphenyl; Mes = mesityl.
5a
5b
5c
5d[19]
5e
7[27]
8[27]
R
Me
Me
Mes
Dipp
iPr
Dipp
Dipp
R′
H
Me
–148.8
165
H
H
H
δ(31P)
1JP, H
C=P
N–C–N
N–C–P
–149.3
165
1.763(3)
104.9(2)
124.0(2)
131.2(2)
–147.3
165
1.747(2)
104.3(2)[b]
127.7(2)
128.0(2)[b]
–136.7
164
1.752(1)
104.3(1)
123.9(1)
131.8(1)
–149.9
166
–143.0
171
1.763(2)
103.7(2)
127.2(2)
129.1(2)
–129.5
–
1.774(1)
103.8(1)
119.9(1)
135.8(1)
1.772(1)[a]
105.1(1)[a]
124.6(1)
130.2(1)[a]
[a] The crystal structure contains a second (disordered) molecule with the following parameters: C=P 1.774(2) Å, N–C–N 105.4(2)°, N–C–P(avg.) 127.3(1)°.
[b] Average values for two disordered molecules.
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
651
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
orientations of the PH unit. This hypothesis is corroborated by CH···X (X = F, I) and CH···O hydrogen-bond interactions. The
the fact that the P–C–N bond angles lack a pronounced dissym- structural characteristics of the cations resemble those of 9[17]
metry (vide infra) but are indistinguishable within experimental (Scheme 6) and feature bent C–P–C units and helical conforma-
error. A similar disorder phenomenon is also observed for one tions, which result from a twist of the heterocyclic rings relative
of two crystallographically independent molecules in crystalline to the central PC2 plane. The P–C bond lengths [1.796(1) to
5b (Figure S10).
1.804(2) Å] are slightly shorter than those in 9 [1.824(2)/1.823(2)
Å[17]], whereas the C–P–C angles [96.3(1) to 98.2(1)°] are similar
[97.4(1)°].[17] The cations differ from each other and from 9
mainly in the twist angles between the PC2 and NHC planes
[interplanar angles 36.2(1) to 47.5(1)° for 6a and 42.7(1) and
43.5(1) for 6b vs. ca. 50 to 60° for 9].[17,29] The perceptibly
smaller twists in 6a and 6b compared with that in 9 may be
attributed to the smaller N-alkyl substituents.
A comparison of selected bond lengths and angles for differ-
ent imidazoyl phosphinidenes is given in Table 1. The molecular
structures of 5a (Figure 1) and the ordered molecule of 5b fea-
ture a planar arrangement of all heavy atoms and the (freely
refined) H1 atoms (largest deviations from least-squares plane
0.012 and 0.018 Å for H1 in 5a and 5b, respectively). The P=C
distances [5a: 1.763(3) Å, 5b: 1.772(1) Å] differ insignificantly
from those in 5d [1.752(1) Å][20] or C-lithiated 7 [Scheme 6,
1.763(2) Å].[27] All P=C distances are close to the upper limit of
reported double-bond lengths in C-aminophosphaalkenes (1.70
to 1.76 Å);[28] a marginally longer bond was reported recently
for the P-silylated derivative 8 [Scheme 6, 1.774(1) Å].[24] The
bond lengths in the N-heterocyclic ring as well as the N–C–N
and N–C–P angles of 5a, 5b and 5d are nearly identical. The C–
P–H angles [5a: 96(1)°, 5b: 96(2)°] are also similar to that in 5d
[94.5(7)°][20] and somewhat more acute than that in
7
[102(2)°].[27] The substantial difference between the P–C–N an-
gles at the E and Z positions relative to the P–H bond [5a:
124.0(2) vs. 131.2(2)°; 5b: 124.6(1) vs. 130.2(1)°] is a typical fea-
ture of P=C double-bond systems.[28] Consequently, the signifi-
cantly different steric and electronic properties of the imidazoyl
units are not reflected in the structural data of the phosphaalk-
enes 5a and 5d. We previously observed the same feature when
comparing the structural data of 5d to 5f, which bears a satu-
rated imidazolidinyl unit with a CH2–CH2 bridge in the NHC
backbone.[21]
Figure 2. Molecular structure of one of the cations of crystalline 6a[OTf].
Thermal ellipsoids are drawn at 50 % probability level. Selected bond lengths
[Å] and angles [°] (the values in brackets refer to the second crystallographi-
cally independent unit; the data for 6b[I] are given in the Supporting Infor-
mation): P1–C9 1.797(1) [1.800(1)], P1–C2 1.796(1) [1.796(1)], C2–N3 1.357(2)
[1.356(2)], C2–N6 1.350(2) [1.351(2)], N3–C4 1.384(2) [1.378(2)], N3–C8
1.461(2) [1.461(2)], C4–C5 1.349(2) [1.343(2)], C5–N6 1.381(2) [1.385(2)], N6–
C7 1.464(2) [1.462(2)], C9–N13 1.356(2) [1.353(2)], C9–N10 1.361(2) [1.357(2)],
N10–C11 1.385(2) [1.381(2)], N10–C15 1.462(2) [1.462(2)], C11–C12 1.345(2)
[1.344(3)], C12–N13 1.385(2) [1.381(2)], N13–C14 1.464(2) [1.463(2)], C9–P1–
C2 98.2(1) [96.3(1)].
For species with similar molecular structures to those of 5a
and 6a (e.g., 5d,[20] 7,[27] or P-silylated 8),[20] P–C distances in
the range of single bonds are commonly observed. This obser-
vation has raised a vivid debate on the nature of the P–C bond-
ing, and compounds of type 5 are described either as inversely
polarized phosphaalkenes with a polarized covalent σ-bond
and an inversely polarized π-bond[28] or as phosphinidene–NHC
adducts featuring a superposition of two dative bonds.[30] The
observed structural and spectroscopic features of 5a are in ac-
cordance with the results of a recent computational study,[31]
which depicts the bonding in terms of the leading resonance
structures 5a′–5a′′′ (Scheme 7). This description is an exact
match of the picture invoked for inversely polarized phos-
phaalkenes,[28] but computational predictions of relatively low
dissociation energies for the P–C bond indicate that these com-
pounds may present a borderline case. For 6a, a natural bond
order (NBO) analysis suggests that a canonical structure 6a′ is
the leading resonance structure. However, a second-order per-
turbation analysis indicates that hyperconjugation between the
phosphorus lone pairs and the adjacent π*(CN) orbital exerts a
similar stabilizing effect to the π conjugation in the NCN unit
Figure 1. Molecular structure of crystalline 5a. Thermal ellipsoids are drawn
at 50 % probability level. Selected bond lengths [Å] and angles [°]: P1–C1
1.763(3), P1–H1 1.32(3), N1–C1 1.361(3), N1–C2 1.388(3), C1–N2 1.368(3), N2–
C3 1.389(3), C2–C3 1.336(4), C1–P1–H1 96(1), N1–C1–N2 104.9(2), N1–C1–P1
124.0(2), N2–C1–P1 131.2(2).
Scheme 6. Related imidazoyl phosphinidenes 7[27] and 8[24] and a related
bis(NHC)–P(I) cation 9;[17] R = 2,6-diisopropylphenyl.
Crystalline 6a[OTf] and 6b[I] contain two or one crystallo- of the heterocycle and, thus, suggests that the central CPC unit
graphically independent [P(NHC)2]+ cations (Figures 2 and S11), still exhibits a substantial π-bonding contribution regardless of
which connect to the triflate or iodide anions through weak its torsional distortion.
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
652
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
As previous accounts on reactions of P4 with carbenes do
not mention the observation of analogues of 5a–5d,[11–14] we
attribute their formation to the participation of tBuOH, which
is a byproduct of the deprotonation of imidazolium salts and is
not separated in the reactions reported here. This conjecture is
supported by the finding that the formation of 5a is suppressed
if the deprotonation of the imidazolium salt is performed with
sodium hydride.[34] It is corroborated further by the outcome of
experiments conducted with deuterium-labelled isotopomers
of 4a[I] as starting materials. The most decisive results were
obtained from studies involving [D3]4a[I], in which all three
ring-carbon atoms were uniformly deuterated. In this case, the
reactions produced a blend of isotopomers [Dn]5a (n = 1 to 3;
detected by analysis of the 31P NMR spectra of the reaction
mixtures, Figure 3, a), which feature both P–D and P–H substitu-
tion. This product distribution can be consistently explained if
we assume that the P–D isotopomers arise from proton-transfer
reactions with tBuOD (formed by the initial deprotonation of
the deuterated imidazolium salt) and that the P–H isotopomers
arise from reactions with residual protic impurities that dilute
the isotope labels. This hypothesis was corroborated by the ob-
servation that the fraction of P–H isotopomers increased sub-
stantially if the reaction was conducted in a solvent that had
previously been deliberately treated with a drop of H2O (Fig-
ure 3, b).[35,36]
Scheme 7. Leading resonance structures of 5a[31] and 6a.
The formation of comparable molar quantities of 5a and 5b
as well as 6a and 6b in the reactions between P4, tBuOH, and
the appropriate imidazolium salts reveals a formal parallel to
the well-known disproportionation of P4 in alkaline medium to
give phosphane and hypophosphorous acid: P4 + 4H2O → 2PH3
+ 2H3PO2.
However, as the formation of the phosphaalkenes 5c and 5d
is not accompanied by the generation of detectable amounts
of phosphamethine cyanines (which suggests that both pro-
ducts are formed through independent routes) and the occur-
rence of a substantial amount of insoluble phosphorus-contain-
ing material in all reactions has to be accounted,[32] we must
assume that the reactions reported here follow a different path-
way. To cast light on the formation of 5, we studied the reac-
tions of the imidazolium salts 4a[I] and 4c[Cl] with P4 and
tBuOK at low temperature through 31P NMR spectroscopy.
The treatment of 4c[Cl] with P4 and tBuOK at –20 °C gave a
solution with a 31P NMR spectrum that showed the signals of
5c, unreacted P4, and two transient AA′XX′ patterns in a ratio
of ca. 4:1 (see Figure S7). The parameters derived from spectral
simulation of the signals of the more abundant component
[δ31P = 451.3 (AA′), 43.7 ppm (XX′) JA,A′/X,X′ = –498/ 34 Hz; JA,X/
= –382/288 Hz][33] closely resemble those of the tetraphos-
A,X′
phatrienes of Bertrand (C, Scheme 2),[11,12] and we assign the
observed intermediates as the N-mesityl derivatives E/Z-10 on
this basis (Scheme 8).[33] The formation of these species was
explained[11,12] as resulting from the sequential attack of two
nucleophilic carbenes on P4 during the initial stages of the reac-
tion. Although analogous derivatives were unobservable in the
reaction of 4a[I] (presumably as their lifetimes are much shorter
owing to the lower steric demand of the N substituents), the
activation of white phosphorus by sterically less demanding
carbenes may give rise to cationic P(I)–bis(carbene) adducts (F,
Scheme 2).[13] Consequently, the formation of 6a and 6b in the
reaction of P4 with 4a[I]/4b[I] and tBuOH can support the inter-
pretation that this transformation is likewise initiated by carb-
ene-activation of P4.
Figure 3. Expansion of the 31P NMR spectra of solutions obtained from the
reaction of [D3]4a[I] with KOtBu and P4 in (a) THF and (b) THF with a drop of
H2O. The shown signals are attributable to the H and D isotopomers of 5a
(1: H2-NHC=PH, 2: HD-NHC=PH, 3: D2-NHC=PH, 4: HD-NHC=PD, 5: D2-NHC=
PD). Note that the spectra were recorded at different magnetic field strengths
1
(101.2 vs. 161.9 MHz); thus, the JP, D couplings have different scaling.
Similar experiments on substrates with specific isotope labels
at the 2- ([D1]4a) or 4,5-positions of the ring ([D2]4a) gave less
conclusive results. In this case, the known propensity of NHCs
or NHC/imidazolium mixtures to exchange their ring H atoms
with protic solvents or substrates[37,38] led to scrambling of the
isotope labels during the reaction, which precluded further
mechanistic interpretation. This exchange can also explain the
formation of CH-protonated products from [D3]4a.
As the deuteration experiments allowed us to identify the
imidazolium ions as the initial source of (at least part of) the P-
bound hydrogen atom in 5a, we wondered if the cations could
also be employed directly in proton-transfer reactions. There-
fore, we studied the reaction of P4 with a 3:2 mixture of 4a[I]
and KOtBu.[39] Although 5a still formed under these conditions,
Scheme 8. Transient intermediates E/Z-10 (R = mesityl) and 11[40] observed
during the reactions of imidazolium salt and /KOtBu with P4.
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
653
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the workup of the reaction mixture afforded a reduced yield of Conclusions
this product and a sizeable amount of recovered imidazolium
We have demonstrated the one-step synthesis of the phos-
phaalkenes 5a–5d with varying steric bulk and salts 6a[X] and
6b[X] (X = I, OTf) with bis(NHC)-stabilized phosphorus cations
from P4 in presence of an imidazolium salt and KOtBu. Further-
more, we have established a safe and high-yielding procedure
for the conversion of white or red phosphorus into the hepta-
phosphides Na3P7 and (Me3Si)3P7. Although the latter was used
previously only for the synthesis of N-arylated imidazoyl phos-
phinidenes and did not give access to N-alkyl derivatives, Na3P7
is a versatile precursor for the synthesis of both types of phos-
phaalkenes. This allows tuning of the reactivity by the steric
demand and, thus, might open up new reaction pathways and
applications.
At first glance, the conversion of elemental phosphorus into
a mixture of species with formally lower (5a and 5b) and higher
(6a and 6b) phosphorus oxidation states resembles the alkaline
disproportionation of elemental phosphorus to give PH3 and
hypophosphorous acid. However, the finding that the formal
disproportionation products are clearly generated through in-
dependent reaction channels and the observation of polyphos-
phorus species as coproducts indicate that this reaction pro-
ceeds by a different mechanism. Spectroscopic studies of low-
temperature reactions suggest that phosphaalkenes 5 are
formed through an initial carbene-activation of P4 and subse-
quent trapping of the formed products by tBuOH, which arises
as a byproduct of the carbene generation. Thus, the simultane-
ous reaction of P4 with a carbene and a weak Brønsted acid,
which may be considered a “frustrated Brønsted pair”, might
open new avenues for the formation of low-coordinate phos-
phorus species or P4 activation, if the currently rather low atom
economy is neglected. We are now exploring the use of a wider
range of imidazolium precursors and attempting to optimize
the reaction conditions to improve the atom economy.
salt. Thus, we conclude that the deprotonation of the imid-
azolium ions by a stoichiometric quantity of base is mandatory
and that the carbene formed through deprotonation, not the
imidazolium ion, is the active reagent in the formation of 5a.
In addition to the resonances of P4 and 5a, the 31P NMR
spectra recorded during the reaction of 4a[I] with P4 and KOtBu
at –20 °C showed a characteristic signal pattern [AK2MX spin
1
system with A, K, M = 31P, X = H; δ31P = 68.9 (A), –358.0 (K),
–330.0 ppm (M); δ1H = –4.3 ppm (X); JA,K = –194, JA,M = 12,
JK,M = –244, JA,X = 16, JK,X = 128, JM,X = 3.6 Hz] attributable to
the tetraphosphide anion 11 (Scheme 8).[40] The signals of 11
and P4 persisted as long as a temperature of –20 °C was main-
tained but disappeared when the mixture was warmed to ambi-
ent temperature to give rise to complex, unidentified line pat-
terns. As the anion 11 was reported to decay at room tempera-
ture to P-richer phosphides[40] and polyphosphides in general
are known to equilibrate with each other or with P4 under the
assembly of new polyphosphorus frameworks,[41,42] we inter-
pret the spectral changes as an indication that 11 and P4 had
reacted to yield higher polyphosphide(s) with as yet unknown
molecular structures.[43]
Although 11 was originally obtained by the reduction of P4
with Na/K naphthalenide,[40] tetraphosphabicyclo[1.1.0]butane
anions are also discussed as the initial products of the activa-
tion of P4 by nucleophiles.[42] We consider this pathway as the
most likely explanation for the formation of 11 in the present
case. Nevertheless, we can discuss two alternatives. One is the
formation of 11 and other polyphosphides after the decay of
one of the initial products of the carbene-activation of P4, simi-
lar to the formation of [(carbene)2P]+ cations together with
anionic species observed by Bertrand and co-workers.[14] The
other is the direct nucleophilic activation of P4 by KOtBu. The
formation of products attributed as phosphides in a binary re-
action of both components was established independently (see
Supporting Information). However, as the monitoring of a reac-
tion of an in situ generated carbene/tBuOH mixture with P4
indicated that the formation of observable amounts of soluble
or insoluble phosphides[44] lagged behind that of 5a, we as-
sume that the second alternative is less plausible.
Experimental Section
General Conditions: All manipulations were performed under dry
argon in Schlenk apparatus or a glovebox. Solvents were dried by
standard procedures. White phosphorus was washed with Et2O and
subsequently sublimed at 4 × 10–2 bar and 50 °C. The solution NMR
spectra were recorded with Bruker Avance AV 400 or AV 250 instru-
ments (1H: 400.1/250.0 MHz, 13C: 100.5/62.9 MHz, 31P: 161.9/
101.2 MHz). The chemical shifts are referenced to external tetra-
methylsilane (TMS; 1H, 13C) or 85 % H3PO4 (Ξ = 40.480747 MHz,
31P). Coupling constants are given as absolute values. The solid-
state NMR spectra were recorded with a Bruker Avance 400 spec-
trometer (13C: 100.5 MHz, 31P: 161.9 MHz) equipped with a 4 mm
MAS probe. All experiments were performed under MAS with spin-
It should be noted that the observation of phosphide forma-
tion in reactions of P4 with imidazolium salts and tBuOH estab-
lishes a parallel to the reported synthesis of 5d from 4d[Cl] and
(Me3Si)3P7 via an intermediate salt [NHC–H][P7(SiMe3)2].[20] Even
if the mechanism of this transformation is still unknown, in this
case, it can be envisaged that 5d may be formed by an alterna-
tive route that bypasses a direct carbene-activation of P4. We
consider this option less likely for the reactions studied here,
as it would provide no forthright explanation for the observed
formation of transient tetraphosphatrienes 10 in the synthesis
of 5c. However, as proton transfers or interconversion between
polyphosphides and P4 are reversible processes and, in princi-
ple, may occur in either direction, an analysis of the interplay
of the individual reaction steps observed in the formation of 5
is exceedingly difficult. Thus, we are unable to give a detailed
mechanistic explanation of this reaction.
1
ning speeds between 8 and 12 kHz and high-power H decoupling
during data acquisition. Cross-polarization (CP) was applied with a
ramp-shaped contact pulse and mixing times of 5 (31P) and 2.7 ms
(
13C). The FTIR spectra were recorded with a Thermo Scientific Nico-
let iS5 instrument equipped with an iD5 attenuated total reflec-
tance (ATR) accessory. Elemental analyses were performed with an
Elementar Micro Cube. The imidazolium salts 4a[I], 4b[I],[45]
4a[Cl],[46] 4c[Cl], 4d[Cl],[47] 4e[Cl] [48] and 4f[Cl][49] were synthesized
according to literature procedures. Melting points were determined
with a Büchi B-545 melting point apparatus with samples in sealed
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
654
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
capillaries. 1,3-Dimethyl[D1]imidazolium iodide and 1,3-[D3]dimeth-
ylimidazolium chloride were prepared by analogy to the corre-
and the mixture was stirred overnight. The solvent was removed
under reduced pressure. The remaining bright yellow solid was sus-
sponding 1-butyl-3-methyl imidazolium tetrafluoroborates by H/D pended in CH2Cl2 (5 mL), and the colourless solid was removed by
exchange in D2O or D2O/cat. NaOH.[38] Purification of the crude
products was accomplished through a modified procedure, which
is described in the Supporting Information. 1,3-[D2]Dimethylimid-
azolium chloride was prepared by subjecting the D3 derivative to
H/D exchange in H2O and analogous workup. The degree of deuter-
ation was determined by 1H NMR spectroscopy (see Supporting
Information).
filtration. The filtrate was concentrated to a volume of 1 mL and
stored at –15 °C to yield 6a[OTf] (724 mg, 1.94 mmol, yield 97 %)
as a bright yellow crystalline precipitate, m.p. >165 °C (dec.). 1H
3
NMR (CD2Cl2, 400 MHz): δ = 7.19 (d, JH,H = 0.5 Hz, 4 H, =CH), 3.51
(s, 12 H, NCH3) ppm. 31P NMR (THF, 101.2 MHz): δ = –117.2 (s) ppm.
31P NMR (CD2Cl2, 101.2 MHz): δ = –115.7 (s) ppm. 13C{1H} NMR
(CD2Cl2, 62.9 MHz): δ = 159.9 (d, 1JC,P = 91 Hz, P=C), 123.2 (d, 2JC,P
2.8 Hz, =CH), 121.5 (q, JC,F = 321 Hz, CF3), 37.1 (d, JC,P = 7.1 Hz,
=
1
2
Compounds 5a/6a[X] from the Reaction of P4 with 4a[I] and
KOtBu: A mixture of solid 4a[I] (2.00 g, 8.93 mmol), KOtBu (1.00 g,
8.93 mmol) and white phosphorus (552 mg, 4.46 mmol) was dis-
persed in THF (100 mL), and the resulting slurry was agitated by
magnetic stirring. The liquid phase eventually adopted a yellow col-
our, and a red precipitate formed. Stirring was continued for 24 h.
The solids were then separated by filtration through a sinter plate
and washed with THF (2 × 25 mL). The filtrate was evaporated to
NCH3) ppm. IR: ν = 3121 (w), 3082 (m), 2923 (m), 1630 (w), 1563 (s),
˜
1469 (m), 1379 (m), 1250 (s), 1228 (vs), 1148 (vs), 1029 (vs), 763 (m),
637 (vs) cm–1. C11H16F3N4O3PS (372.30): calcd. C 35.49, H 4.33, N
15.05; found C 35.40, H 4.39, N 14.80.
Compounds 5b/6b[X] from the Reaction of P4 with 4b[I]/KOtBu:
A mixture of solid 4a[I] (401 mg, 1.59 mmol), KOtBu (175 mg,
1.59 mmol) and white phosphorus (99 mg, 0.80 mmol) was dis-
dryness under reduced pressure. The remaining residue was dis- persed in THF (16 mL), and the resulting slurry was agitated by
persed in benzene (50 mL), and the dispersion was filtered. The
new filtrate was again evaporated to dryness. The sublimation of
the solid residue at 50 °C and 4 × 10–2 bar afforded 5a (268 mg,
2.09 mmol, yield 23 % with respect to 4a[I]) as a colourless solid
(m.p. 85 °C). Single crystals suitable for an XRD study were obtained
by recrystallization from benzene. The red solid remaining after the
magnetic stirring. The liquid phase eventually adopted a yellow col-
our, and a red precipitate formed. Stirring was continued for 24 h.
The solids were then separated by filtration through a sinter plate
and washed with THF (2 × 6 mL). The filtrate was evaporated to
dryness under reduced pressure. The remaining residue was dis-
persed in benzene (50 mL), and the dispersion was filtered. The
filtration of the original reaction mixture was extracted with CH2Cl2 new filtrate was again evaporated to dryness. The sublimation of
(2 × 80 mL). The combined extracts were concentrated under re- the solid residue at 95 °C and 4 × 10–2 bar afforded 5b (64 mg,
duced pressure to a total volume of 10 mL, and the resulting solu-
0.41 mmol, yield 26 % with respect to 4b[I]) as a colourless solid
tion was stored at –15 °C. A crystalline, bright yellow precipitate (m.p. 134 °C). The red solid remaining after the filtration of the
formed and was collected by filtration to give 6a[I] (836 mg,
2.39 mmol, yield 54 % with respect to 4a[I]), m.p. >150 °C (dec.).
The residue of the last filtration was dried under reduced pressure
and characterized by CP/MAS 31P and 13C NMR spectroscopy (see
Supporting Information).
original reaction mixture was extracted with CH2Cl2 (2 × 15 mL).
The combined extracts were concentrated under reduced pressure
to a total volume of 3 mL, and the resulting solution was stored at
–15 °C. A bright yellow crystalline precipitate formed and was col-
lected by filtration to give 6b[I] (304 mg, 0.75 mmol, yield 47 %
with respect to 4b[I]); m.p. >140 °C (dec.).
1,3-Dimethyl-2-phosphanylidene-2,3-dihydro-1H-imidazole
(5a): 1H NMR ([D8]THF, 250 MHz): δ = 6.69 (s, 2 H, HC=), 3.26 (d,
1,2,3,4-Tetramethyl-2-phosphanylidene-2,3-dihydro-1H-imid-
1
1
1
4
4JP, H = 0.9 Hz, 6 H, NCH3), 1.68 (d, JP, H = 164 Hz, 1 H, PH) ppm. H azole (5b): H NMR (C6D6, 250 MHz): δ = 2.83 (d, JP, H = 0.5 Hz, 6
4
1
NMR (C6D6, 250 MHz): δ = 5.56 (d, JP, H = 0.5 Hz, =CH), 2.73 (d,
H, NCH3), 2.51 (d, JP, H = 165 Hz, 1 H, PH), 1.25 (s, 6 H, CH3) ppm.
1
4
4JP, H = 1.0 Hz, 6 H, NCH3), 2.38 (d, JP, H = 165 Hz, 1 H, PH) ppm. 31P
1H NMR ([D8]THF, 250 MHz): δ = 3.22 (d, JP, H = 0.6 Hz, 6 H, NCH3),
NMR ([D8]THF, 101.2 MHz): δ = –155 (d, JP, H = 164 Hz) ppm. 31P
2.03 (s, 6 H, CH3), 1.71 (d, JP, H = 164 Hz, 1 H, PH) ppm. 31P NMR
1
1
NMR (C6D6, 101.2 MHz): δ = –149.3 (d, JP, H = 165 Hz) ppm. 13C{1H} (C6D6, 101.2 MHz): δ = –148.8 (d, JP, H = 165 Hz) ppm. 31P NMR
1
1
NMR ([D8]THF, 62.9 MHz): δ = 178.0 (d, 1JC,P = 66 Hz, P=C), 118.0 (d,
([D8]THF, 101.2 MHz): δ = –153.5 (d, JP, H = 164 Hz) ppm. 13C{1H}
1
2JP, C = 3.4 Hz, =CH), 34.6 (d, 2JP, C = 6.9 Hz, NCH3) ppm. 13C{1H} NMR
NMR ([D8]THF, 62.9 MHz): δ = 174.7 (d, JP, C = 89 Hz, P=C), 120.7 (d,
1
1
2
(C6D6, 62.9 MHz): δ = 176.0 (d, JC,P = 66 Hz, P=C), 116.0 (d, JP, C
3.0 Hz, =CH), 33.5 (d, JC,P = 6.7 Hz, NCH3) ppm. IR: ν = 3161 (s),
3127 (s), 3105 (s), 2311 (vs), 1483 (m), 1437 (m), 1391 (m), 1338 (m),
1233 (s), 1156 (s), 885 (m), 699 (vs), 646 (m), 628 (w) cm–1. C5H9N2P
(128.11): calcd. C 46.88, H 7.08, N 21.87; found C 46.80, H 7.09, N
21.80.
=
1JC,C = 3.2 Hz, =C), 30.7 (s, NCH3), 7.5 (s, CH3) ppm. IR: ν = 2928 (w),
˜
2
2295 (vs), 1664 (w), 1459 (w), 1433 (w), 1381 (vs), 1345 (vs), 1221
(m), 1164 (s), 1096 (m), 921 (m) cm–1. C7H13N2P (156.17): C 53.84, H
8.39, N 17.94; found C 53.44, H 8.36, N 17.75.
˜
Bis[1,2,3,4-tetramethyl-1H-imidazol-2(3H)-ylidene]phos-
1
phenium Iodide (6b[I]): H NMR (CD2Cl2, 250 MHz): δ = 3.47 (s, 12
Bis[1,3-dimethyl-1H-imidazol-2(3H)-ylidene]phosphenium Iod- H, CCH3), 2.25 (s, 12 H, NCH3) ppm. 31P NMR (CD2Cl2, 101.2 MHz):
1
ide (6a[I]): H NMR (CD2Cl2, 250 MHz): δ = 7.78 (s, 4 H, =CH), 3.55
δ = –113.0 (s) ppm. 13C{1H} NMR (CD2Cl2, 62.9 MHz): δ = 159.0 (d,
4
2
(d, JP, H = 0.5 Hz, 12 H, NCH3) ppm. 31P NMR (CD2Cl2, 101.2 MHz):
1JC,P = 77 Hz, P=C), 126.8 (s, =C), 34.4 (d, JC,P = 7.9 Hz, NCH3), 9.92
δ = –117.3 (s) ppm. 31P CP/MAS NMR (161.9 MHz): δ = –119 ppm.
(s, CCH3) ppm.
1
13C{1H} NMR (CD2Cl2, 62.9 MHz): δ = 159.0 (d, JC,P = 77 Hz, P=C),
Reactions of P4 with 4c[Cl]/4d[Cl] and KOtBu: A mixture of solid
imidazolium salt (4c[Cl]: 500 mg, 1.47 mmol; 4d[Cl]: 625 mg,
1.47 mmol), KOtBu (165 mg, 1.47 mmol) and white phosphorus
(91 mg, 0.73 mmol) was dispersed in THF (20 mL), and the resulting
slurry was agitated by magnetic stirring. The liquid phase eventually
turned violet, and a dark precipitate formed. Stirring was continued
for 48 h. Compound 5d was characterized by 31P NMR spectro-
scopic analysis of the reaction mixture. For the isolation of 5c, the
solids were separated by filtration through a sinter plate. The filtrate
2
2
123.0 (d, JC,P = 2.8 Hz, =CH), 36.9 (d, JC,P = 7.0 Hz, NCH3) ppm. IR:
ν = 3138 (w), 3070 (m), 1662 [v(s)], 1565 (s), 1466 (w), 1437 (w),
˜
1386 (m), 1238 (s), 1174 (m), 1147 (m), 1089 (vs), 904 (m), 760 (s),
740 (w), 656 (w), 619 (s) cm–1. C10H16IN4P (350.14): calcd. C 34.30,
H 4.61, N 16.00; found C 34.30, H 4.74, N 15.70.
Bis[1,3-dimethyl-1H-imidazol-2(3H)-ylidene]phosphenium Tri-
flate (6a[OTf]): Trimethylsilyl triflate (0.40 mL, 489 mg, 2.20 mmol)
was added to a solution of 6a[I] (700 mg, 2.00 mmol) in THF (5 mL),
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
655
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
was concentrated to ca. 5 mL and stored at r.t. to produce 5c
The combined filtrates were evaporated to dryness, and the residue
(123 mg, 0.37 mmol, yield 25 % with respect to 4c[Cl]) as a crystal- was extracted with toluene (5 × 3 mL). The combined extracts were
line precipitate. Further purification was performed by sublimation
at 50 °C and 4 × 10–2 bar to afford the product (98 mg,0.29 mmol,
yield 20 % with respect to 4c[Cl]).
filtered through Celite, and the solvent was removed under reduced
pressure. The yellow residue was recrystallized from toluene at
–30 °C to yield the phosphaalkenes 5a and 5e as pale yellow crys-
talline solids (5a: 340 mg, 2.61 mmol, 70 %; 5e: 456 mg, 2.48 mmol,
66 %).
1,3-Dimesityl-2-phosphanylidene-2,3-dihydro-1H-imidazole
(5c): 1H NMR (C6D6, 250 MHz): δ = 6.76 (s, 4 H, m-CH), 5.90 (s, 2
1
H, =CH), 2.21 (s, 12 H, o-CH3), 2.08 (s, 6 H, p-CH3), 1.78 (d, JP, H
=
1,3-Diisopropyl-2-phosphanylidene-2,3-dihydro-1H-imidazole
(5e): M.p. 132–133 °C. 1H NMR (300 MHz, C6D6): δ = 6.12 (s, 2 H,
1
165 Hz, PH) ppm. H NMR ([D8]THF, 250 MHz): δ = 6.98 (s, 4 H, m-
CH), 6.82 (s, 2 H, =CH), 2.31 (s, 12 H, p-CH3), 2.14 (s, 6 H, o-CH3),
3
4
HC=CH), 4.49 [sept d, JH, H = 6.7, JP, H = 2.7 Hz, 2 H, CH-
1.26 (d, JP, H = 164 Hz, PH) ppm. 31P NMR (C6D6, 101.2 MHz): δ =
1
1
3
(CH3)2], 2.63 (d, JP, H = 165.9 Hz, 1 H, P–H), 0.95 [d, JH,H = 6.7 Hz,
12 H, CH(CH3)2] ppm. 13C{1H} NMR (75 MHz, C6D6): δ = 173.7 (d,
1JP, C = 92.9 Hz, C=P), 113.2 (s, HC=CH), 48.8 [s, CH(CH3)2], 21.2 [s,
CH(CH3)2] ppm. 31P{1H} NMR (THF, 121 MHz): δ = –149.9 (s) ppm.
–147.3 (d, JP, H = 165 Hz) ppm. 31P NMR ([D8]THF, 101.2 MHz): δ =
1
–149.6 (d, 1JP, H = 164 Hz) ppm. 13C NMR (C6D6, 62.9 MHz): δ = 129.3
(s, =CH), 127.8 (s, m-CH), 20.6 (s, p-CH3), 17.8 (s, o-CH3) ppm. 13C
1
31P NMR (THF, 121 MHz): δ = –149.9 (d, JP, H = 166.1 Hz) ppm.
1
NMR ([D8]THF, 62.9 MHz): δ = 180.0 (d, JP, C = 93.5 Hz, P=C), 129.0
(s, m-CH), 118.0 (s, =CH), 20.3 (s, o-CH3), 17.3 (s, p-CH3) ppm.
C9H17N2P (184.22): calcd. C 58.7, H 9.3, N 15.2; found C 58.4, H 9.5,
N 14.9.
1,3-Bis(2,6-diisopropylphenyl)-2-phosphanylidene-2,3-dihydro-
1H-imidazole (5d): 31P NMR (C6D6, 101.2 MHz): δ = –135.0 (d, 1JP, H
=
N-Arylated Imidazoyl Phosphinidenes from [(Na3P7)(DME)x]:
The corresponding imidazolium salt (4c[Cl]: 1.28 g; 4d[Cl]: 1.59 g;
4f[Cl]: 1.60 g; 3.75 mmol; 3 equiv.) and Na3P7 (500 mg, 1.75 mmol,
1.35 equiv.) were suspended in THF (10 mL), and the suspension
was stirred at 20 °C for 16 h. The brown-black precipitate was re-
moved from the red solution by filtration through a G3 glass frit
and washed with THF (3 × 3 mL). Further workup was performed
as described above for the N-alkylated species (5c: 760 mg,
2.25 mmol, 60 %; 5d: 1.26 g, 3.00 mmol, 80 %; 5f: 1.04 g, 2.46 mmol,
66 %). The elemental analysis data as well as the 1H, 13C and 31P
NMR spectra correspond to the data given above for 5c and the
previously reported data for 5d[20] and 5f.[21]
165 Hz) ppm.
[(Na3P7)(DME)x]: A suspension of white phosphorus (10.0 g,
80.7 mmol, 1.75 equiv.), freshly cut sodium (3.17 g, 138 mmol,
3 equiv.) and naphthalene (400 mg, 3.13 mmol, 0.07 equiv.) in a
mixture of THF (50 mL) and DME (50 mL) was stirred for 3 d at
20 °C until a bright orange suspension formed. The suspension was
concentrated, and the yellow precipitate was collected by filtration
and washed thoroughly with n-hexane (10 × 10 mL) to remove the
naphthalene. Drying at reduced pressure and 20 °C for 6 h yielded
trisodium heptaphosphide as an orange powder, which was soluble
in THF. Further drying at reduced pressure and 60 °C for 3 d yielded
an insoluble orange-red powder. The phosphorus content was de-
termined photometrically to be 72.5 %, which corresponds to a re-
sidual DME content of x ≈ 0.12 (13.5 g, 45.5 mmol, yield 99 %).
Crystal Structure Determinations: X-ray diffraction studies of 5a–
5c, 6a[I], 6b[I] and 6a[OTf] were performed with a Bruker diffrac-
tometer equipped with a Kappa APEX II Duo CCD detector and a
KRYO-FLEX cooling device with Mo-Kα radiation (λ = 0.71073 Å) at
T = 100 K. The structures were solved by direct methods (SHELXS-
97[50]) and refined with a full-matrix least-squares scheme on F2
(SHELXL-2014 and SHELXL-97).[50] Semiempirical absorption correc-
tions were applied for all structures. Non-hydrogen atoms were re-
fined anisotropically, and H atoms were refined isotropically with a
riding model [H(P) free]. Crystalline 5b contains two crystallographi-
cally independent molecules. One of these resides at a special posi-
tion with crystallographic C2 symmetry and features a H(P) atom
that is disordered over two positions. The H1 atom in 5c is disor-
dered over two positions. Owing to the low quality of the crystal,
the data of 6a[I] is included in the Supporting Information but was
not deposited at the CCDC.
Alternative synthesis from red phosphorus: Red phosphorus (40.0 g,
1.29 mol, 7 equiv.), freshly cut sodium (12.7 g, 0.550 mol, 3 equiv.)
and naphthalene (4.95 g, 38.7 mmol, 0.21 equiv.) were suspended
in a mixture of THF (50 mL) and DME (50 mL). The dark green
suspension was stirred at 20 °C for 18 h until all of the sodium was
consumed. The resulting black suspension was heated under reflux
for 24 h until a bright orange suspension formed. The insoluble
yellow precipitate was collected by filtration, thoroughly washed
with n-hexane (6 × 40 mL) to remove the naphthalene and dried at
reduced pressure and 60 °C for 3 d (54.0 g, 0.182 mol, yield 99 %).
31P{1H}NMR (THF, 101 MHz): δ = –119.0 (br s) ppm. [(Na3P7)-
(C4H10O2)0.12]: calcd. C 1.9, H 0.4, N 0; found C 2.0, H 0.6, N 0.1.
(Me3Si)3P7 from [(Na3P7)(DME)x]: To a stirred suspension of
[(Na3P7)(DME)0.12] (1.43 g, 4.81 mmol, 1 equiv.) in toluene (10 mL)
was added trimethylsilyl chloride (2.3 mL, 18 mmol, 1.25 equiv.) at
–80 °C over 15 min. The suspension was stirred and slowly warmed
to 20 °C over 3 h. The pale precipitate was removed from the yellow
solution by filtration through Celite. The removal of the solvent un-
der reduced pressure and recrystallization from toluene at –30 °C
yielded tris(trimethylsilyl) heptaphosphide as a pale yellow powder
(1.96 g, 4.49 mmol, yield 94 %). The elemental analysis data and the
1H, 13C and 31P NMR spectra were identical to the previously re-
ported data.[22]
CCDC-1414812 (for 5a), -1422252 (for 5b), -1414813 (for 5c),
-1414811 (for 6a[OTf]) and –1422307 (for 6b[I]) contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Compound 5a: Colourless crystals, C5H9N2P, M: 128.11 g mol–1
,
crystal size: 0.36 × 0.16 × 0.11 mm, monoclinic, space group P21/n,
a = 6.7979(19) Å, b = 11.661(3) Å, c = 8.641(2) Å, β = 91.021(15)°,
V = 684.8(3) A3, Z = 4, ρ(calcd.) = 1.243 Mg m3, F(000) = 272, θmax
=
28.34°, μ = 0.299 mm–1, 5924 reflections measured, 1685 unique
reflections [Rint = 0.047] for structure solution and refinement with
78 parameters, R1 = 0.065 [for 1156 reflections with I > 2σ(I)], wR2 =
N-Alkylated Imidazoyl Phosphinidenes from [(Na3P7)(DME)x]:
The corresponding imidazolium salt (4a[Cl]: 497 mg; 4e[Cl]: 708 mg;
3.75 mmol, 3 equiv.) and [(Na3P7)(DME)x] (500 mg, 1.68 mmol,
1.35 equiv.) were suspended in THF (10 mL), and the mixture was
heated under reflux for 16 h. The resulting orange suspension was
filtered through a G3 glass frit and washed with THF (3 × 3 mL).
0.189, largest diff. peak and hole: 1.221 and –0.612 e Å–3
.
Compound 5b: Colourless crystals, C7H13N2P, M: 156.16 g mol–1
,
crystal size: 0.52 × 0.40 × 0.34 mm, monoclinic, space group C2/c,
a = 22.265(2) Å, b = 8.6445(6) Å, c = 14.0303(12) Å, β = 107.512(5)°,
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
656
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
V = 2575.3(4) A3, Z = 12, ρ(calcd.) = 1.208 Mg m3, F(000) = 1008,
θmax = 30.56°, μ = 0.251 mm–1, 16120 reflections measured, 3950
unique reflections [Rint = 0.026] for structure solution and refine-
ment with 150 parameters and one restraint, R1 = 0.040 [for 2846
reflections with I > 2σ(I)], wR2 = 0.114, largest diff. peak and hole:
Dalton Trans. 2007, 5505; e) C. Ganter, in: Phosphorus Ligands, in: Asym-
metric Catalysis (Ed.: A. Börner), Wiley-VCH, Weinheim, Germany, 2008;
p. 393–406; f) J. I. Bates, J. Dugal-Tessier, D. P. Gates, Dalton Trans. 2010,
39, 3151; g) M. Nicolas, G. C. Fu, in: Chiral Ferrocenes in Asymmetric Cataly-
sis (Eds.: L.-X. Dai, X.-L. Hou), Wiley-VCH, Weinheim, Germany, 2010, ch.
11, p. 307–335.
0.435 and –0.292 e Å–3
.
[4] For selected reviews, see: a) M. Hissler, P. W. Dyer, R. Réau, in: Topics in
Current Chemistry, vol. 250 (Ed.: J.-P. Majoral), Springer, Berlin, 2005, p.
127–163; b) P. W. Dyer, R. Réau, in: Functional Organic Materials: Synthe-
ses, Strategies, and Applications (Eds.: T. J. J. Müller, U. H. F. Bunz), Wiley-
VCH, Weinheim, Germany, 2007, p. 119–178; c) P. W. Siu, D. P. Gates, in:
Conjugated Polymer Synthesis: Methods and Reactions (Ed.: Y. Chujo),
Wiley-VCH, Weinheim, Germany, 2010, p. 215–227; d) J. D. Protasiewicz,
in: Comprehensive Inorganic Chemistry II, vol. 1 (Eds.: J. Reedijk, K. Poep-
pelmeier), Elsevier, Amsterdam, 2013, p. 325–348.
Compound 5c: Yellow crystals, C21H25N2P, M: 336.40 g mol–1, crystal
size: 0.23 × 0.21 × 0.10 mm, monoclinic, space group P21/n, a =
8.2558(9) Å, b = 14.6092(17) Å, c = 16.1057(14) Å, β = 102.010(4)°,
V = 1900.0(3) A3, Z = 4, ρ(calcd.) = 1.176 Mg m3, F(000) = 720, θmax
=
27.48°, μ = 0.149 mm–1, 17209 reflections measured, 4341 unique
reflections [Rint = 0.074] for structure solution and refinement with
230 parameters, R1 = 0.056 [for 2439 reflections with I > 2σ(I)],
wR2 = 0.134, largest diff. peak and hole: 0.272 and –0.321 e Å–3
.
[5] P. P. Power, J. Organomet. Chem. 2004, 689, 3904.
[6] a) K. Dimroth, P. Hoffmann, Angew. Chem. Int. Ed. Engl. 1964, 3, 384;
Angew. Chem. 1964, 76, 433; b) K. Dimroth, Top. Curr. Chem. 1973, 38, 1.
[7] D. Schmid, S. Loscher, D. Gudat, D. Bubrin, I. Hartenbach, Th. Schleid, Z.
Benkö, L. Nyulászi, Chem. Commun. 2009, 830–832.
Compound 6a[OTf]: Bright yellow crystals, C10H16N4P·CF3O3S, M:
372.31 g mol–1, crystal size: 0.36 × 0.30 × 0.24 mm, triclinic, space
group P1, a = 11.6579(9) Å, b = 12.3283(9) Å, c = 12.9947(10) Å, α =
¯
116.937(3)°, β = 102.470(4)°, γ = 91.388(4)°, V = 1609.7(2) A3, Z = 4,
ρ(calcd.) = 1.536 Mg m3, F(000) = 768, θmax = 27.48°, μ = 0.350 mm–
1, 65509 reflections measured, 7368 unique reflections [Rint = 0.021]
for structure solution and refinement with 415 parameters, R1 =
0.028 [for 6673 reflections with I > 2σ(I)], wR2 = 0.074, largest diff.
[8] G. Becker, Z. Anorg. Allg. Chem. 1976, 423, 242.
[9] K. Issleib, E. Leißring, M. Riemer, H. Oehme, Z. Chem. 1983, 23, 99–100.
[10] In a few special cases, there is precedent for the direct access of multiply
bonded species from elemental phosphorus, for example, P5–, see: a) M.
Baudler, D. Düster, D. Ouzounis, Z. Anorg. Allg. Chem. 1987, 544, 47; b)
M. Baudler, S. Akpapoglou, D. Ouzounis, F. Wasgestian, B. Meinigke, H.
Budzikiewicz, H. Münster, Angew. Chem. Int. Ed. Engl. 1988, 27, 280–281;
Angew. Chem. 1988, 100, 288.
[11] J. D. Masuda, W. W. Schoeller, B. Donnadieu, G. Bertrand, Angew. Chem.
Int. Ed. 2007, 46, 7052–7055; Angew. Chem. 2007, 119, 7182–7185.
[12] J. D. Masuda, W. W. Schoeller, B. Donnadieu, G. Bertrand, J. Am. Chem.
Soc. 2007, 129, 14180–14181.
peak and hole: 0.967 and –0.330 e Å–3
.
Compound 6b[I]: Bright yellow crystals, C14H24IN4P, M:
406.24 g mol–1, crystal size: 0.26 × 0.22 × 0.20 mm, monoclinic,
space group P21/n, a = 11.7734(5) Å, b = 11.4627(5) Å, c =
13.1048(6) Å, β = 90.714(2)°, V = 1768.42(13) A3, Z = 4, ρ(calcd.) =
1.526 Mg m3, F(000) = 816, θmax = 36.35°, μ = 1.899 mm–1, 68773
reflections measured, 8502 unique reflections [Rint = 0.042] for
structure solution and refinement with 189 parameters, R1 = 0.029
[for I > 2σ(I)], wR2 = 0.061, largest diff. peak and hole: 2.067 and
[13]
[14]
[15]
[16]
O. Back, G. Kuchenbeiser, B. Donnadieu, G. Bertrand, Angew. Chem. Int.
Ed. 2009, 48, 5530–5533; Angew. Chem. 2009, 121, 5638.
C. D. Martin, C. M. Weinstein, C. E. Moore, A. L. Rheingold, G. Bertrand,
Chem. Commun. 2013, 49, 4486–4488.
For a review, see: D. Martin, M. Soleilhavoup, G. Bertrand, Chem. Sci.
2011, 2, 389–399.
a) H. J. Schönherr, H. W. Wanzlick, Chem. Ber. 1970, 103, 1037–1046; b)
N. I. Korotkikh, G. F. Rayenko, O. P. Shvaika, T. M. Pekhtereva, A. H. Cowley,
J. N. Jones, C. L. B. MacDonald, J. Org. Chem. 2003, 68, 5762–5765; c) W.
Zuo, P. Braunstein, Dalton Trans. 2012, 41, 636–643; d) P. K. Majhi, S.
Sauerbrey, G. Schnakenburg, A. J. Arduengo, R. Streubel, Inorg. Chem.
2012, 51, 10408–10416.
–1.231 e Å–3
.
Supporting Information (see footnote on the first page of this
article): Synthetic procedures for deuterated 1,3-dimethylimid-
azolium salts; 31P NMR spectra of reaction mixtures of 4a/KOtBu/P4
and 4c/KOtBu/P4 obtained under different reaction conditions; CP/
MAS NMR spectra of the solid residue of a reaction of 4a/KOtBu/
P4 before and after extraction with CH2Cl2; representations of the
molecular structures of 5b[I], 5c[I] and 6b[I], and the preliminary
molecular structure of 6a[I].
[17]
B. D. Ellis, C. A. Dyker, A. Decken, C. L. B. Macdonald, Chem. Commun.
2005, 1965–1967.
[18]
[19]
B. D. Ellis, C. L. B. Macdonald, Coord. Chem. Rev. 2007, 251, 936–973.
a) K. Hansen, T. Szilvási, B. Blom, S. Inoue, J. Epping, M. Driess, J. Am.
Chem. Soc. 2013, 135, 11795–11798.
A. M. Tondreau, Z. Benkö, J. R. Harmer, H. Grützmacher, Chem. Sci. 2014,
5, 1545–1554.
M. Bispinghoff, A. M. Tondreau, H. Grützmacher, C. A. Faradji, P. G. Pringle,
Dalton Trans. 2015, DOI: 10.1039/C5DT01741F.
H. Schmidbaur, A. Bauer, Phosphorus Sulfur Silicon Relat. Elem. 1995, 102,
217–219.
O. Back, M. Henry-Ellinger, C. D. Martin, D. Martin, G. Bertrand, Angew.
Chem. Int. Ed. 2013, 52, 2939–2943; Angew. Chem. 2013, 125, 3011.
A. Doddi, D. Bockfeld, T. Bannenberg, P. G. Jones, M. Tamm, Angew. Chem.
Int. Ed. 2014, 53, 13568–13572; Angew. Chem. 2014, 126, 13786–13790.
a) J. Schmedt auf der Günne, S. Kaczmarek, L. v. Wüllen, H. Eckert, D.
Paschke, A. J. Foecker, W. Jeitschko, J. Solid State Chem. 1999, 147, 341;
b) K. Eichele, R. E. Wasylishen, J. F. Corrigan, N. J. Taylor, A. J. Carty, K. W.
Feindel, G. M. Bernard, J. Am. Chem. Soc. 2002, 124, 1541; c) A. Pfitzner,
M. F. Bräu, J. Zweck, G. Brunklaus, H. Eckert, Angew. Chem. Int. Ed. 2004,
43, 4228; Angew. Chem. 2004, 116, 4324; d) M. Scheer, L. J. Gregoriades,
M. Zabel, M. Sierka, L. Zhang, H. Eckert, Eur. J. Inorg. Chem. 2007, 2775;
e) T. Wiegand, H. Eckert, S. Grimme, D. Hoppe, M. Ruck, Chem. Eur. J.
2011, 17, 8739–8748.
Acknowledgments
[20]
[21]
[22]
[23]
[24]
[25]
The authors thank Dr. W. Frey (Institute of Organic Chemistry,
University of Stuttgart, Germany) for the collection of X-ray data
sets. Furthermore, the University of Stuttgart, Germany, ETH Zü-
rich, Switzerland, and the European Union (Marie Curie ITN Sus-
Phos, grant agreement number 317404) are acknowledged for
financial support.
Keywords: Carbenes · Phosphorus · Phosphaalkenes · P–P
activation · Nitrogen heterocycles
[1] W. F. Dahsent, Nonexistent Compounds, Marcel Dekker, New York, 1965.
[2] For overviews, see: a) P. P. Power, Chem. Rev. 1999, 99, 3463–3503; b)
R. C. Fischer, P. P. Power, Chem. Rev. 2010, 110, 3877–3923.
[3] For selected reviews, see: a) L. Weber, Angew. Chem. Int. Ed. 2002, 41,
563; Angew. Chem. 2002, 114, 583; b) P. Le Floch, Coord. Chem. Rev. 2006,
250, 627; c) G. Fu, Acc. Chem. Res. 2006, 39, 853; d) C. Müller, D. Vogt,
[26]
A single-crystal XRD study was also performed on 6a[I]. Although the
structure could be solved, satisfactory refinement remained unfeasible
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
657
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
owing to the low quality of the crystal. The preliminary results included
in the Supporting Information confirm the constitutional assignment
and reveal similar metrical parameters of the cation to those in 6a[OTf].
Y. Wang, Y. Xie, M. Y. Abraham, R. J. Gilliard Jr., P. Wei, H. F. Schaefer III, P.
v. R. Schleyer, G. H. Robinson, Organometallics 2010, 29, 4778–4780.
[39] By using the notation [NHC–H+] for an imidazolium salt, the formation
of a mixture of 5a and 6a may be described formally by the following
overall equation: 3[NHC–H+]I–
+ 1/2P4 + 2KOtBu → NHC=PH +
[(NHC)2P+]I– + 2tBuOH + 2KI. Thus, the formation of 5a does not formally
require a further stoichiometric amount of base beyond the two equiva-
lents needed for the generation of [(NHC)2P+]I– (6a).
[27]
[28]
[29]
[30]
[31]
[32]
L. Weber, Eur. J. Inorg. Chem. 2000, 2425–2441.
[40] M. Baudler, C. Adamek, S. Opiela, H. Budzikiewicz, D. Ouzounis, Angew.
Chem. Int. Ed. Engl. 1988, 27, 1059–1061; Angew. Chem. 1988, 100, 1110–
1111.
[41] M. Baudler, K. Glinka, Chem. Rev. 1993, 93, 1623–1667.
[42] M. Scheer, G. Balázs, A. Seitz, Chem. Rev. 2010, 110, 4236–4256.
For 6a[I], the angles differ further: 28.8(7) and 52.6(6)°.
G. Frison, A. Sevin, J. Organomet. Chem. 2002, 643–644, 105–111.
Z. Kelemen, R. Streubel, L. Nyulászi, RSC Adv. 2015, 5, 41795–41802.
Attempts to reduce the amount of P4 led to a decrease of the yield
of 5a and gave solutions that still contained unreacted carbene, which
suggests that an initial P4/NHC molar ratio of 1:2 is needed to ensure
the complete consumption of the imidazolium salt.
[43] For
a specific reaction of an anionic tetraphosphabicyclo[1.1.0]-
butane derivative with P4 to give an octaphosphide anion, see: W. T. K.
Chan, F. García, A. D. Hopkins, L. C. Martin, M. McPartlin, D. S. Wright,
Angew. Chem. Int. Ed. 2007, 46, 3084–3086; Angew. Chem. 2007, 119,
3144.
[33] Spectral simulation of the signals of the minor component at δ = 328.4
and 50.8 ppm was prevented by the insufficient signal-to-noise (S/N)
ratio. On the basis of the observed chemical shifts, we assign the major
component as the E-isomer and the minor one as the Z-isomer.
[34] In this case, deprotonation yields H2 as a byproduct, which evolves from
the reaction mixture. Further reaction with P4 gave only intractable insol-
uble phosphorus compounds and was not pursued further.
[35] The presence of water may at first glance seem incompatible with the
proposed reaction, but NHCs may coexist with small amounts of H2O,
in: organic solvents (see ref.[37]), and we have also verified the stability
of 6a under such conditions.
[44] The formation of solid phosphides was recognized through the reddish
colour of the formed precipitate.
[45] D. J. Williams, M. R. Fawcett-Brown, R. R. Raye, D. Van Derveer, Y. T. Pang,
R. L. Jones, K. L. Bergbauer, Heteroat. Chem. 1993, 4, 409–414.
[46] M. Smiglak, J. D. Holbrey, S. T. Griffin, W. M. Reichert, R. P. Swatloski, A. R.
Katritzky, H. Yang, D. Zhang, K. Kirichenko, R. D. Rogers, Green Chem.
2007, 9, 90–98.
[47] L. Hintermann, Beilstein J. Org. Chem. 2007, 3, 1–5.
[48] T. Schaub, U. Radius, in: Inorganic Syntheses, vol. 35 (Ed.: T. B. Rauchfuss),
John Wiley & Sons, Hoboken, 2010, p. 78–79.
[36] Similarly, the formation of products with P–D substituents was promoted
when D2O was added to the reaction mixture; see Supporting Informa-
tion.
[49] K. M. Kuhn, R. H. Grubbs, Org. Lett. 2008, 10, 2075–2077.
[50] G. M. Sheldrick, Acta Crystallogr., Sect. A 2008, 64, 112.
[37] O. Hollóczki, P. Terleczky, D. Szieberth, G. Mourgas, D. Gudat, L. Nyulászi,
J. Am. Chem. Soc. 2011, 133, 780–789.
Received: September 7, 2015
[38] R. Giernoth, D. Bankmann, Eur. J. Org. Chem. 2008, 2881–2886.
Published Online: October 22, 2015
Eur. J. Inorg. Chem. 2016, 649–658
www.eurjic.org
658
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim