Full Papers
doi.org/10.1002/ejic.202100153
further radical reactions (Scheme 5, f) or another PR3 addition
donor structure, either the ferrocenylphosphonium salts
[CpFeC5H4PR3](PF6) as the products of ring CÀ H functionalization
or the half-sandwich complexes [CpFe(PR3)3](PF6) as the prod-
ucts of Cp-ring replacement.
depending on the reaction conditions (see above). The
described substitution process (Scheme 5, a–e) has the first-
order kinetics with respect to PR3, so the first introduction of a P
donor (Scheme 5, either a, b, or c) has the highest activation
barrier.[23] The same trend was observed in our case: the first
nucleophilic attack (Scheme 4, ii) has the highest activation
energy. Finally, the reactions shown in Scheme 5 proceed
Both these reactions are rather complex from the mecha-
nistic point of view; all their key stages are associated with
interconversions of odd-electron intermediates. The initial and
common for both reactions stage is the nucleophilic addition of
the P donor to the cyclopentadienyl ring, which leads to the
17e η4-phosphoniocyclopentadiene complex IM1. The subse-
quent addition of the PÀ OR nucleophile to IM1 proceeds at the
metal atom and leads to the 19e adduct IM3 stabilized by a
partial delocalization of the “19th” electron on the phosphorous
ligand. Then, the ring replacement proceeds as alternating
processes of the partial decoordination of the η4-phosphoniocy-
clopentadiene ligand and nucleophilic addition of the P donor
to the metal; the oxidation of 19e IM7 by the initial ferrocenium
salt completes the formation of half-sandwich product 3. For
bad π-acids, the addition to the metal in IM1 is energetically
unfavorable and the reaction proceeds as its oxidation followed
by the deprotonation of the resulting intermediate IM2.
°
smoothly at À 20 C, while similar 18e isostructural complexes
react only at higher temperatures (e.g. [CpFe(C6H5Cl)]+ reacts
with P(OEt)3 at 150 C for 18 h[47]). Thus, it can be estimated that
°
the corresponding substitution in odd-electron complexes
proceeds about 109 times faster than in 18e complexes.
Following this observation, the reactions of ferrocenium with
PÀ OR nucleophiles studied in the present work proceed in mild
conditions too.
As was shown above, the proposed mechanism of the Cp-
ring replacement by a PÀ OR nucleophile is based on some
features reported earlier: the substitution itself proceed via
interconversion of 17e and 19e intermediates, phosphorus
reagent attacks the singly occupied metal-centered orbital, and
the resulting 19e complexes are stabilized by low-energy
acceptor orbitals of a phosphorus ligand. Both ring CÀ H
functionalization and Cp-ring replacement reactions have the Experimental and Computational Section
common exo-addition step (Scheme 1, a and Scheme 4, i,
respectively). It can be assumed that IM1 and IM2 are always in
General Methods: All reactions were performed in an argon
atmosphere using Schlenk technique. All workup procedures were
equilibrium regardless of the nature of a P donor (Scheme 1, b).
Therefore, the further reactions are governed by the possibilities
of the nucleophilic attack on the SOMO of IM1 (Scheme 4, ii) to
give IM3 and the deprotonation of IM2 (Scheme 1, c). For good
π-acceptors (phosphites and phosphonites), the former process
is feasible due to sufficient stabilization of the corresponding
19e adduct IM3. The rate of the 17e–19e substitution is high
and, therefore, this process will dominate. On the contrary, for
bad π-acceptors (which are usually good bases, e.g. tertiary
phosphines and aminophosphines), the intermediate IM3 is
destabilized, so the deprotonation of IM2 occurs. In some rare
cases (d–f), the rates of both processes are similar, so the
reaction leads to the formation of both products 2 and 3. The
considerations presented in this Section are in agreement with
the classification of the phosphorus reagents by their TEPs
presented earlier: thus, the higher value of TEP implies the
higher π-acidity of the corresponding PR3. However, such a
simple classification cannot cover all subtle mechanistic features
that govern the outcome in the reactions of ferrocenium with
PÀ OR nucleophiles and cannot provide its absolutely accurate
prediction.
performed in air. Organic solvents were dried using standard
procedures and distilled prior to use. Commercially available
acetylferrocene and phosphites P(OMe)3, P(OEt)3, and P(OPh)3 were
used as received. PhP(OMe)2,[49] Ph2P(OMe),[50] Ph2P(OEt),[51] iPr2P-
[53]
(OMe), iPr2P(OEt),[52] and PhP(OiPr)2 were prepared according to
the published protocols from the corresponding chlorophosphines,
dry alcohols, dry triethylamine as a base, and dry petroleum ether
as a solvent. Ferrocenium hexafluorophosphate and tetrafluorobo-
rate were prepared as described.[54] Silica gel (Merck 60, 70–
230 mesh) was used for column chromatography. Petroleum ether
°
refers to the 40–70 C boiling fraction.
NMR spectra were recorded on a Bruker AMX 400 spectrometer in
(CD3)2CO at r.t. using the standard pulse sequences (residual
internal C3HD5O (δ(1H) 2.05), internal C3D6O (δ(13C) 29.84), and
external H3PO4 85% aq. (δ(31P) 0). Chemical shifts are given in parts
per million (ppm). Elemental analysis was performed on a Carlo
Erba 1106 automated CHN analyzer at the Laboratory of Micro-
analysis of INEOS RAS.
X-ray Diffraction Study: The crystals were grown from dichloro-
methane/ether by slow diffusion using enriched chromatographic
fraction for a given compound. The complex 3e·BF4 ·2(CH2Cl2) was
obtained by the described experimental procedure but using
ferrocenium tetrafluoroborate instead of ferrocenium hexafluoro-
phosphate. Single-crystal X-ray diffraction experiments were carried
out with a Bruker SMART APEX II diffractometer[55] (graphite-
monochromated Mo Kα radiation, λ=0.71073 Å, ω-scan technique).
Semiempirical absorption correction based on equivalent reflec-
tions was applied using the SADABS program.[56] The structures
were solved by direct methods and refined by the full-matrix least-
squares technique against F2 with anisotropic thermal parameters
for non-hydrogen atoms with the SHELXL program.[57] The hydro-
gen atoms were placed geometrically and included in the structure
factors calculations in the riding motion approximation. Crystallo-
Conclusions
Redox activation of organometallic compounds increases their
reactivity, which makes possible to carry out those reactions
that do not proceed for the initial stable, as a rule, 18e
complexes.[36,37,40,46,48] In the present study, we have found that
ferrocenium hexafluorophosphate reacts with PÀ OR nucleo-
philes under mild conditions and gives, depending on the P
Eur. J. Inorg. Chem. 2021, 1601–1610
1608
© 2021 Wiley-VCH GmbH