Functionalized bis(pentafluoroethyl)phosphanes: Improved syntheses and molecular structures in the gas phase

Article abstract of DOI:10.1002/ejic.201300364

(C₂F₅)₂PNEt₂ represents an excellent starting material for the selective synthesis of bis(pentafluoroethyl) phosphane derivatives. The moderately air-sensitive aminophosphane is accessible on a multi-gram scale by treating Cl₂PNEt₂ with C ₂F₅Li. Treatment with gaseous HCl or HBr yielded the corresponding phosphane halides (C₂F₅)₂PCl and the so far unknown (C₂F₅)₂PBr in good yields. The hitherto unknown (C₂F₅)₂PF was obtained by treating (C₂F₅)₂PBr with excess antimony trifluoride. Treatment of (C₂F₅)₂PCl with Bu₃SnH led to the quantitative formation of (C₂F ₅)₂PH. Deprotonation formally yielded the (C ₂F₅)₂P^- anion in a form that was stabilized by coordination to mercury ions to form the complex [Hg{P(C ₂F₅)₂}₂(dppe)]. An improved high-yielding synthesis of (C₂F₅)₂POH was achieved by treating (C₂F₅)₂PNEt₂ with p-toluenesulfonic acid. The gas-phase structures of (C₂F ₅)₂PH and (C₂F₅)₂POH were determined by electron diffraction. The vibrational corrections employed in the data analysis of the diffraction data were derived from molecular dynamics calculations. Both compounds exist in the gas phase mostly as C ₁-symmetric cis,cis conformers with regard the orientation of the C₂F₅ groups relative to the functional groups H and OH. The presence of a second conformer at ambient temperature is likely in both cases. The refined amounts of dominant conformers are 94(6) and 85(6) % for (C₂F₅)₂PH and (C₂F₅) ₂POH, respectively. The conformational behaviour was further explored by potential energy surface scans based on DFT calculations. Important experimental structural parameters for the most stable conformers are r _e(P-C)_average = 1.884(3) A for (C₂F ₅)₂PH and r_e(P-C)_average = 1.894(4) A and r_e(P-O) = 1.582(3) A for (C₂F ₅)₂POH. The different coordination properties of (C ₂F₅)₃P, (C₂F₅) ₂POH, (CF₃)₃P and (CF₃) ₂POH were evaluated by complex formation with [Ni(CO)₄]: the maximum achievable number of CO ligands substituted by (C₂F ₅)₃P is 1, by (C₂F₅)₂POH is 2, by (CF₃)₃P is 3 and by the smallest ligand (CF ₃)₂POH is 4. New synthetic routes to a series of bis(pentafluoroethyl)phosphanes have been developed and the structures of two of them, (C₂F₅)₂PH and the phosphinous acid (C₂F₅)₂POH, were studied by gas electron diffraction, making use of molecular dynamics calculations to support data analysis in an improved approach. Copyright

Full text of DOI:10.1002/ejic.201300364

FULL PAPER
DOI:10.1002/ejic.201300364
Functionalized Bis(pentafluoroethyl)phosphanes: Improved
Syntheses and Molecular Structures in the Gas Phase
Alexander V. Zakharov,[a,b] Yury V. Vishnevskiy,[b] Nadine Allefeld,[c]
Julia Bader,[c] Boris Kurscheid,[c] Simon Steinhauer,[c]
Berthold Hoge,*[c] Beate Neumann,[b] Hans-Georg Stammler,[b]
Raphael J. F. Berger,[b][‡] and Norbert W. Mitzel*[b]
Dedicated to Professor Heinrich Nöth on the occasion of his 85th birthday
Keywords: Phosphorus / Electron diffraction / Structure elucidation / Molecular dynamics / Density functional
calculations / Fluorine
(C2F5)2PNEt2 represents an excellent starting material for the
selective synthesis of bis(pentafluoroethyl)phosphane deriv-
atives. The moderately air-sensitive aminophosphane is ac-
cessible on a multi-gram scale by treating Cl2PNEt2 with
C2F5Li. Treatment with gaseous HCl or HBr yielded the cor-
responding phosphane halides (C2F5)2PCl and the so far un-
known (C2F5)2PBr in good yields. The hitherto unknown
(C2F5)2PF was obtained by treating (C2F5)2PBr with excess
antimony trifluoride. Treatment of (C2F5)2PCl with Bu3SnH
led to the quantitative formation of (C2F5)2PH. Deprotonation
formally yielded the (C2F5)2P– anion in a form that was stabi-
lized by coordination to mercury ions to form the complex
[Hg{P(C2F5)2}2(dppe)]. An improved high-yielding synthesis
of (C2F5)2POH was achieved by treating (C2F5)2PNEt2 with
p-toluenesulfonic acid. The gas-phase structures of (C2F5)2-
PH and (C2F5)2POH were determined by electron diffraction.
The vibrational corrections employed in the data analysis of
the diffraction data were derived from molecular dynamics
calculations. Both compounds exist in the gas phase mostly
as C1-symmetric cis,cis conformers with regard the orienta-
tion of the C2F5 groups relative to the functional groups H
and OH. The presence of a second conformer at ambient
temperature is likely in both cases. The refined amounts of
dominant conformers are 94(6) and 85(6)% for (C2F5)2PH and
(C2F5)2POH, respectively. The conformational behaviour was
further explored by potential energy surface scans based on
DFT calculations. Important experimental structural param-
eters for the most stable conformers are re(P–C)average
=
1.884(3) Å for (C2F5)2PH and re(P–C)average = 1.894(4) Å and
re(P–O) = 1.582(3) Å for (C2F5)2POH. The different coordina-
tion properties of (C2F5)3P, (C2F5)2POH, (CF3)3P and (CF3)2-
POH were evaluated by complex formation with [Ni(CO)4]:
the maximum achievable number of CO ligands substituted
by (C2F5)3P is 1, by (C2F5)2POH is 2, by (CF3)3P is 3 and by
the smallest ligand (CF3)2POH is 4.
pared with the vast number of π-donating phosphanes, π-
accepting ligands are relatively uncommon. One of the most
electron-deficient and π-acidic phosphane derivatives is
P(CF3)3.[2] It is highly volatile and air-sensitive and hence
unsuitable for technical applications. By comparison, the
next highest homologue, the less volatile P(C2F5)3, can be
handled by using common Schlenk techniques. However,
the high steric demand of the pentafluoroethyl groups pre-
vents P(C2F5)3 from being employed as a ligand in coordi-
nation chemistry and not a single complex is known.
P(C2F5)3 in the solid state and gas phase has a Tolman cone
angle of 193 and 191°, respectively;[3] the conformational
behaviour of the C2F5 substituents in P(C2F5)3 depends on
repulsive F···F contacts.[3]
Introduction
Phosphane ligands are essential for the synthesis of metal
complexes used in a variety of catalytic reactions.[1] Com-
[a] Ivanovo State University of Chemistry and Technology,
Department of Physics,
Engels Av. 7, 153000 Ivanovo, Russia
[b] Fakultät für Chemie, Lehrstuhl für Anorganische Chemie und
Strukturchemie, Universität Bielefeld,
Universitätsstraße 25, 33615 Bielefeld, Germany
E-mail: mitzel@uni-bielefeld.de
Homepage: www.uni-bielefeld.de/chemie/arbeitsbereiche/
ac3-mitzel
[c] Fakultät für Chemie, Anorganische Chemie II, Universität
Bielefeld,
Universitätsstraße 25, 33615 Bielefeld, Germany
E-mail: b.hoge@uni-bielefeld.de
Homepage: www.uni-bielefeld.de/chemie/acii/hoge
[‡] Present address: FB Materialwissenschaften und Physik, Paris-
Lodron Universität Salzburg, Hellbrunner Str. 34, 5020
Salzburg, Austria
Substitution of one pentafluoroethyl group by a sterically
less demanding substituent leads to more complicated con-
formational behaviour, as hasbeen demonstrated for (C2F5)2-
PCl, which has also been studied by gas electron diffraction
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300364.
Eur. J. Inorg. Chem. 2013, 3392–3404
3392
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
(GED).[4] Such compounds with sterically more accessible
Exposure of neat (C2F5)2PNEt2 to gaseous hydrogen
phosphorus atoms allow manifold coordination chemistry. chloride led to P–N bond cleavage and quantitative forma-
A prominent example of this is the chelating ligand (C2F5)2- tion of (C2F5)2PCl (Scheme 1). Completion of the reaction
PC2H4P(C2F5)2 with two P(C2F5)2 groups. It forms the ba- was monitored by evaporating all the volatile components
sis for a diverse range of complexes, many of which exhibit followed by further treatment with gaseous HCl until no
promising catalytic activities.[5]
further precipitation of ammonium salts was observed. Af-
(C2F5)2PC2H4P(C2F5)2 was first synthesized by Ernst ter removal of excess HCl, the resulting colourless liquid
and Roddick in 1989 by the reaction of LiC2F5 with was identified as (C2F5)2PCl by multi-nuclear NMR[18] and
Cl2PC2H4PCl2.[6] In contrast to LiCF3, which eliminates IR spectroscopy.[12]
lithium fluoride even at temperatures below –100 °C,[7]
LiC2F5 is stable up to –50 °C.[8] It can be generated in situ
by treating C2F5Cl or C2F5H with nBuLi in diethyl
ether.[6,9] The C2F5 group can also be introduced by the
fluoride-mediated reaction of C2F5SiMe3 with (PhO)2-
PC2H4P(OPh)2.[10]
The reaction of white phosphorus with C2F5I under high
pressure and temperatures is not a favourable alternative to
the synthesis of bis(pentafluoroethyl)phosphane derivatives.
Mono- and bis(pentafluoroethyl)iodophosphane were ob-
tained in this way only in very low yields.[11] Treating the
iodo compounds with silver or mercury chloride furnishes
the corresponding chloro derivatives.[12] Recently published,
significantly improved syntheses of (C2F5)2PCl[4] and
Scheme 1. Reaction pathways for the syntheses of bis(pentafluoro-
(C2F5)2POH[13] start from the technical product (C2F5)3PF2
ethyl)phosphane derivatives (C2F5)2PX (X = H, F, Cl, Br, OH).
and proceed via P(C2F5)3 and (C2F5)2PO– salts as interme-
diates, respectively. The improved protocol for the synthesis
of (C2F5)2POH has allowed the preparation of several
The bromo derivative (C2F5)2PBr was synthesized analo-
stable and catalytically active transition-metal complexes.[13] gously by treating (C2F5)2PNEt2 with gaseous hydrogen
(C2F5)2PH is formed as a side-product in the synthesis of bromide. The yield of the colourless liquid (C2F5)2PBr was
P(C2F5)3 by reduction of (C2F5)3PF2 with NaBH4.[14]
70% (Scheme 1). The 19F NMR spectrum exhibits a signal
(CF3)2PNEt2 has been shown to represent an ideal start- for the CF3 unit at δ = –80.6 ppm and a multiplet of higher
ing material for the synthesis of (trifluoromethyl)phosphane order for the CF2 unit at δ = –112.0/–114.1 ppm. The
derivatives.[15] It is accessible on a multi-gram scale by the higher-order pattern is due to the diastereotopic nature of
Ruppert procedure[16] and is only slightly air-sensitive. The the fluorine atoms of the CF2 unit.
aminophosphane (CF3)2PNEt2 can be transformed by gas-
In contrast, gaseous hydrogen fluoride cannot be used to
eous HBr into (CF3)2PBr and the bis(trifluoromethyl)phos- generate the corresponding fluorophosphane from (C2F5)2-
phinous acid, (CF3)2POH, can be obtained by treating PNEt2. The results of this reaction will be published else-
(CF3)2PNEt2 with excess p-toluenesulfonic acid.
where. We employed antimony trifluoride as an alternative,
In analogy, the known aminophosphane (C2F5)2PNEt2 a reagent with known potential for transformation of halo-
has shown promise as a starting material for the synthesis phosphane derivatives into the corresponding fluorophos-
of new functional bis(pentafluoroethyl)phosphane deriva- phanes.[19] Treatment of (C2F5)2PBr with SbF3 at ambient
tives (C2F5)2PX (X = F, Br) and selective access to known temperature led to the selective formation of the so far un-
compounds (X = H, Cl, OH). In this contribution we dem- known (C2F5)2PF as a colourless liquid within two days
onstrate such improved preparative protocols as well as a (Scheme 1). In contrast, the reaction of (C2F5)2PCl with
detailed analysis of the gas-phase structures of (C2F5)2POH SbF3 remained incomplete even after seven days at ambient
and (C2F5)2PH.
temperature. The 19F NMR spectrum of (C2F5)2PF reflects
the chemically and magnetically non-equivalent fluorine
atoms. A resonance arising from the CF3 group is observed
3
at δ = –82.4 ppm with J(F,F) coupling between the CF3
Results and Discussion
group and the fluorine atom directly attached to the phos-
phorus atom. The resonance of the diastereotopic CF2
group is a high-order pattern. The resonance of the fluorine
Preparation of Bis(pentafluoroethyl)phosphane Derivatives
Previous work has shown that LiC2F5 is well suited to atom directly bound to the phosphorus atom is split by the
the efficient pentafluoroethylation of chlorophosphane de- characteristic 1J(P,F) coupling of 1015 Hz (δ = –216.7 ppm)
rivatives by nucleophilic substitution.[17] It was generated in into a pattern that can be completely described as an [AB]2-
situ according to a literature method[9] and subsequently M6GX (A, B, G, M = F; X = P) spin system. For compari-
treated with Cl2PNEt2. The product, (C2F5)2PNEt2, was son, Table 1 compiles a set of NMR parameters for com-
isolated by vacuum distillation as a colourless liquid.
pounds containing (C2F5)2P units.
Eur. J. Inorg. Chem. 2013, 3392–3404
3393
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 1. Selected 31P and 19F NMR spectroscopic data for bis(pen-
tafluoroethyl)phosphane derivatives.[a]
δ(31P) 3J(P,F) δ(19CF3) δ(19CFaFb) δ(19CFaFb)
[ppm]
[Hz]
[ppm]
[ppm]
[ppm]
[b]
(C2F5)2PNEt2
(C2F5)2PBr
46.8
44.0
62.2
137.6
–50.8
93.0
–
–82.1
–80.6
–80.2
–82.4
–84.8
–82.1
–116.1
–112.0
117.0
–114.1
17
16
13
13
18
(C2F5)2PCl
–115.9[c]
(C2F5)2PF[d,e]
(C2F5)2PH[f]
(C2F5)2POH
–123.7
–103.0
–123.4
–125.0
–109.0
–125.4
[a] Data measured in [D6]acetone. [b] Determined in CDCl3. [c]
Fluorine atoms in the CF2 group are equivalent. [d] δ(31P) =
137.6 ppm [m, d, 1J(P,F) = 1015 Hz]. [e] δ(19F) = –216.7 ppm [d,
1J(P,F) = 1015 Hz, (C2F5)2PF]. [f] δ(31P) = –50.8 ppm [m, d, 1J(P,H)
3
= 231 Hz], J(P,F) = 18 Hz.
Figure 1. Molecular structure of [Hg{P(C2F5)2}(CN)(dppe)] in the
solid state. Hydrogen atoms and toluene as solvent moleucle have
been omitted for clarity. Relevant atoms are shown as displacement
ellipsoids at the 50% probability level. Selected bond lengths [Å]
The P–N bond could also be cleaved by treatment with
sulfonic acids.[15] The reaction of (C2F5)2PNEt2 with at least
5 equivalents of dried p-toluenesulfonic acid in the less vol-
atile solvent 1,6-dibromohexane led to a quantitative con- and angles [°]: Hg1–P1 2.615(1), Hg1–P2 2.565(1), Hg1–P3
2.490(1), Hg1–C27 2.178(3), C27–N1 1.132(3), P3–C28 1.893(3),
version to (C2F5)2POH (Scheme 1). The phosphinous acid
P3–C30 1.884(3), P1–Hg1–P2 82.0(1), P3–Hg1–C27 116.6(1), P1–
Hg1–P3 111.9(1), P2–Hg1–C27 113.7(1).
is the only volatile compound and was isolated by fractional
condensation in yields better than 90%. The experimental
NMR spectroscopic data are in good agreement with litera-
ture data.[13]
sphere with a (C2F5)2P–Hg–P(C2F5)2 angle of 123.7(1)° and
Bis(pentafluoroethyl)phosphane, (C2F5)2PH, was easily
a Ph2P–Hg–PPh2 angle of 80.4(1)°. Possible effects of nega-
accessed from bis(pentafluoroethyl)halophosphanes by
tive hyperconjugation are evident in the elongated C–F dis-
treatment with Bu3SnH (Scheme 1). After the reaction of
tances, the mean for the CF2 units being 1.36 Å, which
neat (C2F5)2PCl with a slight excess of Bu3SnH at room
compares with a similarly long average of 1.345 Å in the
temperature, pure (C2F5)2PH was isolated by fractional
gaseous state of the neutral compound (C2F5)2PH.
condensation in quantitative yield. The NMR spectroscopic
data are in agreement with the literature.[13] IR spectro-
scopic investigations revealed the P–H stretching mode at
2349 cm–1.
Deprotonation of (C2F5)2PH with [Hg(CN)2dppe] led to
the formation of the bis(pentafluoroethyl)phosphanide
anion, (C2F5)2P–, stabilized by coordination to the mercury
cation. The reaction proceeded via the monosubstituted
mercury complex [Hg(CN){P(C2F5)}(dppe)] to give the fi-
nal product [Hg{P(C2F5)2}2(dppe)]. The intermediate was
characterized by NMR spectroscopy [1J(Hg,P) = 1289 Hz]
and crystallography (see Figure 1, which also lists structural
parameters in the caption). The NMR properties of
[Hg{P(C2F5)2}2(dppe)] are comparable to those of its CF3
analogue.[20] The 31P{19F} NMR spectrum exhibits, in ad-
dition to a signal for the dppe ligand [δ(31P) = 11.2 ppm], a
2
triplet at δ = –9.2 ppm with a J(P,P) coupling constant of
58 Hz and 199Hg satellites [1J(Hg,P) = 824 Hz]. Conse-
quently, the resonance in the 199Hg NMR spectrum at δ =
296.2 ppm is a triplet of triplets due to its coupling to dif-
ferent phosphorus atoms.
Figure 2. Molecular structure of [Hg{P(C2F5)2}2(dppe)]·C7H8 in
Recrystallization from toluene yielded single crystals of
[Hg{P(C2F5)2}2(dppe)]·C7H8 suitable for X-ray diffraction.
The complex crystallizes in the monoclinic space group Pc.
The molecular structure of the complex, shown in Figure 2,
is similar to that of the CF3 analogue.[20] The mercury atom
the solid state. Hydrogen atoms and toluene as the solvent molecule
have been omitted for clarity. Relevant atoms are shown as dis-
placement ellipsoids at the 50% probability level. Selected bond
lengths [Å] and angles [°]: Hg1–P1 2.686(1), Hg1–P2 2.580(1),
Hg1–P3 2.507(1), Hg1–P4 2.505(1), P4–C31 1.882(3), P4–C33
1.885(3), P1–Hg1–P2 80.4(1), P3–Hg1–P4 123.7(1), P1–Hg1–P3
exhibits a strongly distorted tetrahedral coordination 108.6(1), P2–Hg1–P4 114.8(1).
Eur. J. Inorg. Chem. 2013, 3392–3404
3394
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Gas-Phase Structures
The thermally averaged molecular structures determined
by experimental gas-phase electron diffraction (GED) rep-
resent the weighted (Boltzmann) averages over all the vi-
brational states. As a consequence, they are geometrically
inconsistent (i.e., internuclear distances cannot be used to
construct a consistent model of Cartesian coordinates); this
is also called the “shrinkage effect”,[21,22] which can be ac-
counted for by applying vibrational corrections to the equi-
librium internuclear distances, which improves the quality
of refined experimental parameters. This also allows a more
accurate comparison between quantum chemically calcu-
lated and experimentally determined structures. There are a
number of approaches for the computation of these correc-
tions. These include the “rectilinear harmonic approxi-
mation”[23,24] (implemented, for example, in the programs
ASYM40[23,24] and SHRINK;[25–28] first approximation:
kh0), the improved “curvilinear harmonic approxi-
mation”[26,28] (implemented in SHRINK; second approxi-
mation: kh1) and the more realistic “anharmonic approxi-
mation” (also implemented in SHRINK; k3, calculated at
the first-order perturbation theory level by using third de-
rivatives of the potential energy[28]). A novel approach in-
volves calculating vibrational corrections by using molecu-
lar dynamics (MD).[29,30] The last two methods were ap-
plied in the two structure determinations in this contri-
bution, and the MD method has been further developed in
this context.
Figure 3. Potential energy surface diagram of (C2F5)2PH for rota-
tion around two P–C bonds [τ(C3–P1–C2–C4) and τ(C2–P1–C3–
C5) dihedral angles] calculated at the B3LYP/6-31++G(d,p) level
of theory.
Table 2. Conformers of (C2F5)2PH, their torsion angles, relative en-
ergies and abundances at 295.15 K calculated at the O3LYP/aug-
cc-pVTZ level of theory.
Conf. τ(C2–P1–C3–C5) τ(C3–P1–C2–C4)
ΔE
ΔG°
x
[°]
[°]
[kcalmol–1] [kcalmol–1] [mol-%]
I
II
III
IV
V
173.4
76.5
170.5
81.5
77.9
–108.8
163.7
155.5
51.0
–113.0
46.2
0.00
0.52
0.62
1.28
1.26
2.07
0.00
0.61
1.10
1.85
2.00
2.00
61.6
21.9
9.6
2.7
2.1
(C2F5)2PH
The (C2F5)2PH molecule has two enantiomers separated
by a very small (0.25 kJmol–1, B3LYP/cc-pVTZ calculation)
barrier to inversion corresponding to a saddle-point struc-
ture of Cs symmetry with an imaginary frequency of
14 cm–1 (B3LYP/cc-pVTZ). However, enantiomers cannot
be distinguished by GED. Consequently, data analysis was
carried out for only one of them.
VI
66.8
2.1
According to the potential energy surface (PES) analysis
(Figure 3), there are six conformers of (C2F5)2PH for each
enantiomer. Table 2 shows the O3LYP/aug-cc-pVTZ ener-
gies and abundances of the conformers; B3LYP/cc-pVTZ
data are provided in the Supporting Information. Most of
the conformers are structurally similar to those of (C2F5)2-
PCl,[4] the difference being that there is only one minimum
for conformer II of (C2F5)2PH instead of two very close
lying minima on the PES for (C2F5)2PCl (conformers 2 and
3 in the notation in ref.[4]). Hence, theory predicts two
major conformers for (C2F5)2PH whereas all the others
have higher energies and abundances that are too low at
the temperature of the GED study to be reliably detectable.
Consequently, the data analysis was restricted to the two
lowest-energy conformers I and II shown in Figure 4.
We undertook molecular dynamics (MD) simulations
and it was interesting to note that conformer II was present
in 18% of the steps in these simulations starting from the
equilibrium structure of the first conformer. Therefore MD
simulations may also be used to estimate the conforma-
tional composition.
Figure 4. Calculated molecular structures of the two major con-
formers of (C2F5)2PH, I (upper) and II (lower).
Eur. J. Inorg. Chem. 2013, 3392–3404
3395
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
The refinement of GED data by using MD corrections
The bond lengths calculated by using the B3LYP func-
for the vibrational parameters yielded the lowest disagree- tional paired with the cc-pVTZ basis set for P, C and H
ment factor (Rf = 4.6%), but differences between the struc- and aug-cc-pVTZ for F are generally overestimated. The
tural parameters obtained from refinements performed by calculated C–C and P–C distances are of poor quality, espe-
using the kh1 and k3 corrections were negligibly small. In- cially when comparing them with experimental re (not rh1)
clusion of conformer II in the amount calculated (22 mol- values. In better agreement with the experiment are the
%) did not lead to significant changes in the Rf factor. bond lengths computed at the O3LYP/aug-cc-pVTZ level
However, inclusion of conformer II in amounts larger than of theory. The C–F distances are especially remarkable, as
25% led to a significant (by the Fisher criterion) increase their differences are within experimental error, but the cal-
in the Rf factor and a model consisting solely of conformer culated C–C and P–C distances still show rather poor agree-
II produced a very high Rf factor (9.4%). Refinement by ment with experimental values. Calculations at the MP2/
varying the conformer ratio (starting with 80% of con- aug-cc-pVTZ level produced values for re(P–C) in agree-
former I) resulted in 94(6)% abundance of the first con- ment with experiment; all F–C distances in this computa-
former with its structural parameters being very similar to tion are very slightly elongated (by ca. 0.005 Å); the C–C
those found by refinement of the single-conformer model. bond length is still too large (by 0.011–0.012 Å), but the
Therefore we can conclude that conformer I is dominant, agreement is much better than in the case of DFT calcula-
with an abundance larger than 75%, although the ratio de- tions. This halogen-rich compound appears to be a difficult
termined from the GED data has rather limited accuracy. case for quantum chemistry because none of the applied
The difference in the energies yielded by DFT computations levels of theory was able to reproduce the experimental ge-
is probably underestimated (as it appears to rise with in- ometry with an accuracy better than 0.01 Å for all bond
creasing level of theory) and the actual abundance of con- lengths.
former II is lower than predicted by calculations.
We can compare the bond lengths with those of the
The major structural parameters of (C2F5)2PH obtained (C2F5)2PCl molecule studied previously by GED.[4] In that
from the GED experiment (single conformer model) by study a two-conformer model was applied and the second
using k3 and MD sets of vibrational corrections, as well as most abundant conformer was detectable; the abundance of
from computations, are presented in Table 3 (the param- the major conformer was 61(5)%. These first and second
eters obtained by using kh1 corrections are available in the conformers of (C2F5)2PCl are similar to those of (C2F5)2-
Supporting Information for comparison). The radial distri- PH, but, because the abundance of the second conformer
bution f(r) curve is presented in Figure 5 and the molecular in our experiment was small and could not be detected with
intensity sM(s) curves are provided in the Supporting Infor- sufficient accuracy, we compare the parameters of the first
mation.
(major) conformer only. Corrections of kh1 type were used
Table 3. Major geometrical parameters for conformer I in the (C2F5)2PH molecule.[a]
GED
B3LYP/cc-pVTZ O3LYP/aug-cc-pVTZ MP2/aug-cc-pVTZ
[c]
[c]
re (k3)[b]
rg (kMD
)
re (kMD
)
re
re
re
r(P1–C2)
r(P1–C3)
r(C2–C4)
r(C3–C5)
r(C2–F6)
r(C2–F8)
r(C3–F7)
r(C3–F9)
r(C4–F10)
r(C4–F12)
r(C4–F14)
r(C5–F11)
r(C5–F13)
r(C5–F15)
α(C2–P1–C3)
α(C4–C2–P1)
α(C5–C3–P1)
τ(C4–C2–C1–C3)
τ(C5–C3–P1–C2)
Rf [%]
1.876(3)
1.880(3)
1.516(3)
1.515(3)
1.344(1)
1.341(1)
1.343(1)
1.341(1)
1.325(1)
1.327(1)
1.318(1)
1.323(1)
1.330(1)
1.318(1)
98.6(6)
111.3(2)
110.4(2)
133(2)
1.897(3)
1.896(3)
1.533(3)
1.535(3)
1.349(1)
1.348(1)
1.349(1)
1.348(1)
1.333(1)
1.337(1)
1.327(1)
1.332(1)
1.336(1)
1.327(1)
1.882(3)
1.886(3)
1.519(3)
1.518(3)
1.346(1)
1.343(1)
1.344(1)
1.343(1)
1.328(1)
1.330(1)
1.322(1)
1.326(1)
1.333(1)
1.323(1)
98.4(8)
1.916
1.920
1.555
1.554
1.359
1.355
1.357
1.355
1.340
1.342
1.332
1.337
1.344
1.332
97.4
1.900
1.905
1.549
1.547
1.347
1.343
1.345
1.343
1.329
1.331
1.323
1.327
1.334
1.323
97.4
1.881
1.886
1.531
1.529
1.352
1.349
1.350
1.348
1.332
1.333
1.324
1.328
1.336
1.324
95.1
112.6(2)
111.7(2)
166(3)
111.7
110.8
162.8
173.2
111.2
110.2
162.4
172.8
111.2
110.2
159.9
168.4
143(2)
4.7
176(3)
4.6
[a] The uncertainties quoted in parentheses are 3σLS. Bond lengths of the same kind (P–C, C–C and C–F) were refined in groups g1–g3
with fixed differences (see the Exp. Sect. for details) and therefore the 3σLS uncertainties represent least-squares analysis errors of these
groups, not individual bond lengths. Distances are given in Å and angles in degrees. [b] Fixed differences from B3LYP/cc-pVTZ calcula-
tions. [c] Fixed differences from O3LYP/aug-cc-pVTZ calculations.
Eur. J. Inorg. Chem. 2013, 3392–3404
3396
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 5. Experimental (dots) and theoretical (line) radial distribution f(r) curves and the difference curve (below) for (C2F5)2PH. Vertical
bars indicate the interatomic distances.
in the study of (C2F5)2PCl,[4] so rg distances from our re- C–F) are in very good agreement with SHRINK. On the
finement, which also used kh1 corrections, are compared. other hand, in our experience, SHRINK calculations with a
The rg(P–C) distances in (C2F5)2PH [1.897(3) and cubic force field can sometimes produce unreasonably large
1.896(3) Å] are comparable to those in (C2F5)2PCl [1.904(3) corrections for bonded distances in molecules with low-fre-
and 1.894(3) Å]. The average rg(C–F) = 1.351(1) Å for CF2 quency vibrational modes.
and rg(C–F) = 1.332(1) Å for CF3, which are the same as
those in (C2F5)2PCl within experimental error.
Nevertheless, anharmonic vibrational corrections (either
from MD or from SHRINK) allow equilibrium bond
Note that the 3σLS uncertainties given in Table 3 repre- lengths to be determined from GED experiments that can
sent only least-squares errors. Because the geometrical pa- be directly compared with theoretically calculated values.
rameters were refined in groups with fixed differences, these The calculations of third derivatives for larger molecules
errors represent uncertainties in the groups, not for individ- requires a prohibitively large amount of computational re-
ual bond lengths. Moreover, there are uncertainties in the sources and it also scales very poorly with the system size,
differences, because their values vary between levels of so, until recently, the analysis of the GED data for such
theory. Although the differences in the r(P–C) and r(C–C) molecules was only possible by using kh1 vibrational correc-
bond lengths yielded by the B3LYP/cc-pVTZ, O3LYP/aug- tions (see the discussion in ref.[32] for an example). It is also
cc-pVTZ and MP2/aug-cc-pVTZ calculations agree to not better than harmonic approaches for the description
within 0.001 Å, the differences in r(C–F) often vary by of modes having even potentials (because cubic force field
0.002–0.004 Å. Therefore the total errors for r(C–F) may be constants for such modes are zero by symmetry). At the
several times larger than the 3σLS uncertainty.
same time, molecular dynamics calculations are possible
It is also interesting to compare the anharmonic vi- even for large (more than 100 atoms) molecules and vi-
brational corrections computed by two approaches, namely brational corrections by this approach are computed di-
by molecular dynamics[29,30] and the SHRINK pro- rectly, without any assumptions about the shape of the gov-
cedure,[31] by using potential energy third derivatives. The erning potential. Therefore the MD approach[29] is a valu-
corrections (k = ra – re) for all the bonded distances in able method for determining equilibrium structures directly
(C2F5)2PH calculated by these two methods are available in from GED experiments and we also applied it to analyse
the Supporting Information. Corrections to r(P–C) and the GED data for (C2F5)2POH.
r(C–C) are remarkably similar, whereas corrections to r(C–
F) computed from the MD trajectories are always smaller
(C2F5)2POH
than those from the SHRINK procedure. The reason for
The (C2F5)2POH molecule has an additional conforma-
this is unclear. On the one hand, it is well known that classi- tional degree of freedom due to the rotation of the O–H
cal MD underestimates vibrational amplitudes and correc- bond around the P–O axis. Therefore each enantiomer can
tions for light atom pairs (see ref.[29,30]), but corrections for theoretically have 12 conformers: six with hydrogen ori-
C–C pairs (which have reduced masses even smaller than ented in the same direction as the two P–C bonds (conform-
Eur. J. Inorg. Chem. 2013, 3392–3404
3397
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
ers A) and six with hydrogen oriented in the opposite direc-
Therefore, to keep the model from being overcom-
tion (conformers B). Figure 6 shows the positions of con- plicated, we limited the analysis to the two lowest-energy
formers A on the PES and Table 4 contains information on conformers with the same orientation of O–H, IA and IIA
both conformers A and B (data determined at the B3LYP/ (see Figure 7), as in the case of (C2F5)2PH.
cc-pVTZ level are available in the Supporting Information).
According to the gas-phase study,[15] (CF3)2POH has two
conformers with different orientations of the O–H bond in
a relative abundance of 82 and 18% at room temperature
[IR and computational data for (CF3)2POH[15]]. However,
because the scattering ability of hydrogen atoms is very low
compared with that of heavier nuclei (C, O, F and P), these
rotamers cannot be reliably distinguished by GED. We
made an attempt to include conformers with different ori-
entations of the O–H unit in (C2F5)2POH in the analysis,
but the disagreement factor Rf remained unchanged.
Figure 7. Calculated molecular structures of conformers IA and
IIA of (C2F5)2POH.
The refinement with kMD corrections yielded the dis-
agreement factor Rf = 5.2%. The refined amount of con-
former IIA was 15(6) mol-% (the 3σLS uncertainty is shown
in parentheses), which is in agreement with the amount of
IIA conformer (and also with the combined amount of IIA
and IIB conformers) predicted by computation. The major
refined structural parameters of (C2F5)2POH obtained from
Figure 6. Potential energy surface of (C2F5)2POH for the rotation
the GED experiment as well as the parameters calculated
at the O3LYP and MP2 levels of theory are listed in Table 5,
radial distribution f(r) curves are presented in Figure 8 and
molecular intensity sM(s) curves are available in the Sup-
porting Information.
around two P–C bonds [τ(C2–P1–C3–C5) and τ(C3–P1–C2–C4)
dihedral angles] calculated at the B3LYP/6-31++G(d,p) level of
theory.
The calculated P–C and C–C bond lengths are overesti-
mated in the O3LYP/aug-cc-pVTZ computations with dif-
ferences larger than the experimental errors, but MP2/aug-
cc-pVTZ yielded values close to the experimental ones. The
calculated values of the P–O bond length are too large at
all levels of theory. The O3LYP/aug-cc-pVTZ approxi-
mation gives C–F distances that are very slightly elongated
(by 0.003 Å). It seems that this compound is also a difficult
case for quantum chemistry like (C2F5)2PH.
Again, we can compare the bond lengths of the major
conformer to those of (C2F5)2PCl studied previously by
GED.[4] The rg(P–C) distances in (C2F5)2POH [1.909(3) and
1.911(3) Å] are virtually the same (within experimental un-
certainties) as in (C2F5)2PCl, whereas the rg(C–C) distances
[1.549(3) and 1.551(3) Å] are longer on average (by
0.022 Å). The average rg(C–F) = 1.347(1) Å for CF2 and
rg(C–F) = 1.328(1) Å for CF3 are slightly shorter (by
0.004 Å) than those in (C2F5)2PCl.[4]
Table 4. Conformers of (C2F5)2POH, their torsion angles, relative
energies and abundances at 295.15 K calculated at the O3LYP/aug-
cc-pVTZ level of theory.[a]
Conf. τ(C2–P1–C3–C5) τ(C3–P1–C2–C4)
ΔE
ΔG°
x
[°]
[°]
[kcalmol–1] [kcalmol–1] [mol-%]
IA
IB
IIA
IIB
IIIA
IIIB
IVA
IVB
VA
VB
VIA
VIB
164.6
167.8
62.2
168.4
166.1
170.4
167.0
50.0
47.7
49.5
45.4
0.00
1.22
0.97
2.10
0.61
1.77
2.08
2.96
2.54
3.27
3.88
4.30
0.00
0.85
1.04
1.63
1.31
1.94
2.68
3.22
2.91
3.45
4.83
4.82
60.7
14.4
10.5
3.9
6.6
2.3
0.7
0.3
0.4
0.2
0.0
0.0
69.5
167.0
168.7
70.8
75.7
70.2
75.0
–99.2
–94.5
–107.6
–109.8
58.7
54.8
[a] Conformers A have H atoms oriented in the same direction as
the two P–C bonds, conformers B have H atoms oriented in the
opposite direction.
Eur. J. Inorg. Chem. 2013, 3392–3404
3398
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 5. Major structural parameters of the (C2F5)2POH molecule.[a]
Conformer IA
O3LYP/aug-cc-pVTZ MP2/aug-cc-pVTZ
Conformer IIA
O3LYP/aug-cc-pVTZ MP2/aug-cc-pVTZ
GED
GED
re (kMD
)
re
re
re (kMD
)
re
re
r(P1–C2)
r(P1–C3)
r(P1–O16)
r(C2–C4)
r(C3–C5)
r(C2–F6)
r(C2–F8)
r(C3–F7)
r(C3–F9)
r(C4–F10)
r(C4–F12)
r(C4–F14)
r(C5–F11)
r(C5–F13)
r(C5–F15)
α(C2–P1–C3)
α(C4–C2–P1)
α(C5–C3–P1)
τ(C4–C2–C1–C3)
τ(C5–C3–P1–C2)
1.888(3)g1
1.900(3)g1
1.582(3)g2
1.530(3)g2
1.526(3)g2
1.342(1)g3
1.347(1)g3
1.337(1)g3
1.349(1)g3
1.326(1)g3
1.325(1)g3
1.323(1)g3
1.324(1)g3
1.327(1)g3
1.322(1)g3
92.7(5)
1.917
1.928
1.623
1.548
1.544
1.345
1.350
1.340
1.352
1.329
1.328
1.326
1.328
1.330
1.325
95.0
1.892
1.904
1.627
1.531
1.527
1.351
1.355
1.346
1.356
1.332
1.329
1.326
1.330
1.333
1.325
93.4
1.886(3)
1.896(3)
1.583(3)
1.529(3)
1.526(3)
1.344(1)
1.346(1)
1.346(1)
1.348(1)
1.326(1)
1.325(1)
1.323(1)
1.321(1)
1.328(1)
1.323(1)
97.0(5)
1.915
1.925
1.624
1.547
1.544
1.347
1.349
1.349
1.351
1.329
1.328
1.326
1.324
1.331
1.327
99.3
1.891
1.901
1.628
1.531
1.530
1.353
1.355
1.354
1.356
1.331
1.329
1.326
1.325
1.334
1.327
97.4
112.9(3)
111.5(3)
177(3)
173(3)
112.7
111.3
168.3
164.5
112.7
111.1
164.9
164.5
112.9(3)
119.6(3)
179(3)
112.7
119.4
170.4
62.2
112.8
118.7
168.4
59.8
71(3)
[a] The uncertainties quoted in parentheses are 3σLS. Bond lengths of the same kind were varied in groups g1–g3 with fixed differences
(see the Exp. Sect. for details) and therefore the 3σLS uncertainties represent least-squares analysis errors of these groups, not individual
bond lengths. Distances are given in Å and angles in degrees.
Figure 8. Experimental (dots) and theoretical (line) radial distribution f(r) curves and difference curve (below) for (C2F5)2POH. Vertical
bars indicate the interatomic distances.
Evaluation of the Steric Demand by Complex Formation
(C2F5)3P ligand, namely [(CO)3Ni{P(C2F5)3}]. The huge
steric demand of (C2F5)3P prevented substitution of a sec-
To demonstrate the different coordination properties of ond CO ligand by this phosphane. This led to the exclusive
(C2F5)3P, (C2F5)2POH, (CF3)3P and (CF3)2POH, in par- formation of [(CO)3Ni{P(C2F5)3}], even when [Ni(CO)4]
ticular, with respect to their spatial requirements, we investi- was treated with a large excess of (C2F5)3P over a long
gated their kinetically controlled reactions with [Ni(CO)4] period of time.
in CH2Cl2 solution. The reaction of an excess of (C2F5)3P
In contrast, the less sterically demanding phosphinous
with [Ni(CO)4] led to the formation and isolation of the acid (C2F5)2POH as well as (CF3)3P allowed the substitu-
first example of a transition-metal complex bearing the tion of two CO ligands in [Ni(CO)4]. This led to the forma-
Eur. J. Inorg. Chem. 2013, 3392–3404
3399
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tion of the corresponding dicarbonyl bis-phosphane com- achieved by first establishing a solid/liquid equilibrium
plexes (cf. Scheme 2). The even smaller steric demand of close to the melting point at 281.2 K, then melting all the
(CF3)2POH was demonstrated by the substitution of all solid except for a tiny crystal seed (by using a thin copper
four CO ligands by phosphane units to afford [Ni{P(CF3)2- wire as external heat source) followed by very slowly lower-
OH}4] under the same reaction conditions.
ing the temperature until the whole capillary was filled with
a single crystalline specimen. Figure 9 shows a part of the
crystal structure with hydrogen-bridged chains. In agree-
ment with the VSEPR model, the C–P–C and O–P–C
angles are reduced compared with in the free phosphinous
acid[13c,15] by around 2° due to the coordination to the
Ni(CO)3 fragment. On the other hand, such a moderate
change in angles can be attributed to packing effects in the
crystal. The P–C and P–O distances are reduced as well by
about 0.02 Å upon coordination.
Conclusions
Bis(pentafluoroethyl)phosphane derivatives are well-
suited starting materials for the synthesis of a variety of
electron-deficient and π-acidic ligands. Starting from non-
toxic, commercially available C2F5H with zero ozone de-
pletion potential, the minor air- and moisture-sensitive ami-
nophosphane (C2F5)2PNEt2 is accessible on a multi-gram
scale. It serves as a convenient starting material for the se-
lective synthesis of bis(pentafluoroethyl)phosphane deriva-
tives (C2F5)2PX (X = H, Cl, OH) and the efficient synthesis
of the hitherto unknown compounds (C2F5)2PF and (C2F5)2-
PBr. Bis(pentafluoroethyl)phosphane is a synthon for nu-
cleophilic bis(pentafluoroethyl)phosphanides, as shown by
the preparation of the complex [Hg{P(C2F5)2}2(dppe)],
whereas the halide derivatives (C2F5)2PX (X = F, Cl, Br)
can be synthetically used as electrophilic bis(pentafluoro-
ethyl)phosphenium synthons.
Scheme 2. Observed complex formation in the reaction of [Ni-
(CO)4] with an excess of different perfluoroalkylphosphane deriva-
tives.
The complex [Ni(CO)3{P(CF3)2OH}] is a colourless li-
quid. A single crystal with one water molecule per formula
unit was obtained by in situ crystallization. This was
The gas-phase structures of (C2F5)2PH and (C2F5)2POH
were determined by electron diffraction. A recently devel-
oped approach to the treatment of vibrational corrections
by molecular dynamics calculations was used in these struc-
tural analyses. For comparison, the average structural pa-
rameters for these compounds as well as for (C2F5)2PCl and
(C2F5)3P are listed in Table 6. For the first three entries, the
functionalized (C2F5)2PX compounds, no significant varia-
tion in the P–C and C–F bond lengths can be observed,
however, some slight variations can be seen in the C–C dis-
tances.
Figure 9. Molecular structure of [Ni(CO)3{P(CF3)2OH}]·H2O in
the solid state. Non-hydrogen atoms are shown as displacement
ellipsoids at the 50% probability level. Selected bond lengths [Å]
and angles [°]: P1–Ni1 2.123(1), P1–C4 1.883(2), P1–C5 1.881(2),
Ni1–C1 1.823(2), P1–O4 1.591(2), O4–O5 2.514(2), C5–F4
1.330(3); C4–P1–C5 97.1(1), Ni1–P1–O4 123.9(1).
A comparison between (C2F5)2POH in the gaseous and
solid phases also shows that in this case typical parameters
are rather invariant, for example, the P–C bonds [solid
1.902(2) Å, average] or C–P–C angles [solid 95.3(1)°].[13c] In
Table 6. Average structural parameters for (C2F5)2PX (X = H, OH, Cl, C2F5).
(C2F5)2PH, re (kMD
)
(C2F5)2POH, re (kMD
)
(C2F5)2PCl,[a] rh1
P(C2F5)3,[b] rh1
P–C [Å]
1.884(3)
1.518(3)
1.344(1)
98.4(8)
1.891(3)
1.528(3)
1.333(1)
97.0(5)
1.902(3)
1.544(3)
1.333(1)
98.2(4)
1.904(3)
1.533(3)
1.339(1)
99.9(3)
C–C [Å]
C–F [Å]
C–P–C [°]
P–C–C [°]
112.2(2)
112.2(3)
112.8(3)
113.2(3)
[a] See ref.[4] [b] See ref.[3]
Eur. J. Inorg. Chem. 2013, 3392–3404
3400
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Synthesis of (C2F5)2PBr: (C2F5)2PNEt2 (2.5 g, 7.2 mmol) was co-
oled to –30 °C and degassed for 5 min. Under vigorous stirring,
gaseous HBr was slowly added. All volatile compounds were re-
moved by condensation and the condensate was repeatedly treated
with gaseous HBr until no further precipitation of ammonium salts
was observed. Excess HBr was removed in vacuo at –78 °C yielding
1.8 g (5.1 mmol, 70%) of a colourless liquid. 13C{19F} NMR (Et2O,
20 °C): δ = 119.1 [d, 2J(C,P) = 22 Hz, CF3], 116.0 [d, 1J(C,P) =
69 Hz, CF2] ppm. 19F NMR (Et2O, 20 °C): δ = –80.6 [m, d, 3J(P,F)
= 17 Hz], –112.0/–114.1 (m, CF2) ppm. 31P NMR (Et2O, 20 °C): δ
contrast, the P–O distances are significantly different, that
is, 1.583(3) Å in the gas and 1.629(1) Å in the solid state.
However, in the latter case, the P–OH function acts as a
hydrogen bridge in the aggregation of (C2F5)2POH into tet-
ramers.
Together with earlier studies on (C2F5)3P, the structure
determination in this contribution provides information on
the steric requirements of (C2F5)3P and (C2F5)2POH as li-
gands in complex chemistry. These data are augmented by
complex formation studies of these ligands and their tri- = 44.0 (m) ppm. IR (gas phase): ν = 612 (w), 628 (w), 749 (m), 949
˜
(m), 1111 (w), 1137 (w), 1167 (m), 1230 (s), 1302 (m) cm–1.
fluoromethyl analogues (CF3)3P and (CF3)2POH in substi-
tution reactions with [Ni(CO)4], which revealed the order
of effective size: (C2F5)3PϾ (C2F5)2POH≈(CF3)3PϾ (CF3)2-
POH.
Synthesis of (C2F5)2PF: (C2F5)2PBr (1.4 g, 4.0 mmol) was con-
densed onto an excess of antimony trifluoride, dried in vacuo for
1 h. The reaction mixture was stirred at room temperature for 48 h
and 1.0 g (3.6 mmol, 90%) of the volatile product was removed by
2
condensation. 13C{19F} NMR (Et2O, 20 °C): δ = 119.1 [d, J(C,P)
1
= 20 Hz, CF3], 116.6 [dd, J(C,P) = 57, 2J(C,F) = 8 Hz, CF2] ppm.
Experimental Section
19F NMR (Et2O, 20 °C): δ = –82.4 [dd, 3J(P,F) = 13, 4J(F,F) =
6 Hz, CF3], –123.7/–125.0 (m, CF2), –216.7 [m, d, 1J(P,F) =
1015 Hz, F] ppm. 31P{1H} NMR (Et2O, 20 °C): δ = 137.6 [m, d,
General: All chemicals were obtained from commercial sources and
used without further purification. Standard high-vacuum tech-
niques were employed throughout all preparative procedures. Non-
volatile compounds were handled under dry N2 by using Schlenk
techniques. The NMR spectra were recorded with a Bruker Model
Avance III 300 spectrometer (31P: 111.92 MHz; 19F: 282.40 MHz;
13C: 75.47 MHz, 1H: 300.13 MHz) with positive shifts (in ppm)
downfield from the external standards [85% orthophosphoric acid
(31P), CCl3F (19F) and TMS (1H)]. IR spectra were recorded with
an ALPHA-FT-IR spectrometer (Bruker) by using a gas cell with
KBr windows.
1J(P,F) = 1015 Hz] ppm. IR (gas phase): ν = 610 (vw), 628 (vw),
˜
751 (vw), 864 (w), 959 (w), 1029 (m), 1171 (m), 1231 (s), 1309
(m) cm–1.
Synthesis of (C2F5)2PH: (C2F5)2PCl (5.8 g, 19 mmol) was con-
densed onto degassed Bu3SnH (6.0 g, 21 mmol). The two-phase
system was stirred at room temperature for 1 h and the product
was removed under static vacuum yielding 5.2 g (19 mmol, 100%)
of a colourless liquid. 1H NMR (Et2O, 20 °C): δ = 5.2 [m, d,
1J(P,H) = 231 Hz] ppm. 13C NMR (Et2O, 20 °C): δ = 119.1 [m, t,
1
1J(C,F) = 289 Hz, CF2], 118.9 [m, q, J(C,F) = 285 Hz, CF3] ppm.
Synthesis of (C2F5)2PNEt2: Caution! LiC2F5 is highly reactive and
tends to decompose violently at temperatures above –50 °C.
13C{19F} NMR (Et2O, 20 °C): δ = 119.1 [m, d, 1J(C,P) = 39 Hz,
CF2], 118.9 [m, d, 2J(C,P) = 19 Hz, CF3] ppm. 19F NMR (Et2O,
A 1.6 m n-butyllithium solution in n-hexane (141 mL, 226 mmol)
dissolved in diethyl ether (400 mL) was degassed at –78 °C and the
solution was stirred for 30 min in an atmosphere of pentafluoroeth-
ane (230 mmol). After the addition of dichloro(diethylamino)phos-
phane (18.1 g, 104 mmol) at –78 °C, the mixture was warmed to
room temperature and stirred overnight. The precipitate was re-
moved by filtration. After evaporation of the solvent, 24.6 g
(72.1 mmol, 69%) of (C2F5)2PNEt2 was obtained by vacuum distil-
3
20 °C): δ = –84.8 [m, d, J(P,F) = 13 Hz, CF3], –103.0/–109.0 (m,
CF2) ppm. 31P NMR (Et2O, 20 °C): δ = –50.8 [m, d, 1J(P,H) =
231 Hz] ppm. IR (gas phase): ν = 404 (vw), 608 (w), 748 (m), 855
˜
(vw), 961 (m), 990 (w), 1004 (w), 1155 (m), 1230 (s), 1312 (m), 2349
(vw) cm–1.
Synthesis of (C2F5)2POH: (C2F5)2PNEt2 (4.1 g, 11.9 mmol) was
added to a suspension of p-toluenesulfonic acid monohydrate
(15.5 g, 81.5 mmol) in 1,6-dibromohexane (75 mL), dried and de-
gassed in vacuo for 1 h. The reaction mixture was stirred at room
temperature for 48 h resulting in a rose-coloured suspension. The
volatile compounds were removed under dynamic vacuum condi-
tions using three traps at –30, –78 and –196 °C to separate minor
amounts of the solvent, (C2F5)2POH and C2F5H. The –78 °C trap
contained 3.2 g (11.1 mmol, 93%) of a colourless liquid identified
as (C2F5)2POH. 1H NMR (CDCl3, 20 °C): δ = 3.3 (s) ppm.
13C{19F} NMR (CDCl3, 20 °C): δ = 119.2 [d, 2J(C,P) = 20 Hz,
CF3], 117.3 (m, CF2) ppm. 19F NMR (CDCl3, 20 °C): δ = –82.1 [m,
1
lation at 33–35 °C as a colourless liquid. H NMR (C6D6, 20 °C):
δ = 1.3 [t, 2J(H,H) = 6.9 Hz, CH3], 3.3 [q, 2J(H,H) = 6.9 Hz,
CH2] ppm. 13C{1H} NMR (C6D6, 20 °C): δ = 46.9 [d, 2J(C,P) =
43 Hz, CH2], 12.6 (m, CH3) ppm. 13C{19F} NMR (C6D6, 20 °C): δ
= 119.4 [d, 2J(C,P) = 25 Hz, CF3], 118.5 [d, 1J(C,P) = 58 Hz,
3
CF2] ppm. 19F NMR (C6D6, 20 °C): δ = –82.1 [d, J(P,F) = 18 Hz,
CF3], –116.1/–117 (m, CF2) ppm. 31P{1H} NMR (C6D6, 20 °C): δ
= 46.8 (m) ppm. IR (gas phase): ν = 405 (vw), 445 (vw), 470 (vw),
˜
618 (vw), 746 (w), 797 (vw), 947 (m), 1035 (w), 1059 (vw), 1126
(m), 1145 (m), 1171 (w), 1224 (s), 1301 (m), 1389 (w), 1472 (vw),
2889 (vw), 2951 (vw), 2980 (vw) cm–1.
3
d, J(P,F) = 18 Hz, CF3], –123.4/–125.4 (m, CF2) ppm. 31P NMR
(CDCl3, 20 °C): δ = 93.0 (m) ppm.
Synthesis of (C2F5)2PCl: (C2F5)2PNEt2 (9.1 g, 27 mmol) was cooled
to –30 °C and degassed for 5 min. Under vigorous stirring, gaseous
HCl was slowly added. All volatile compounds were removed by
condensation and the condensate was repeatedly treated with gas-
eous HCl until no further precipitation of ammonium salts was
observed. Excess HCl was removed in vacuo at –78 °C yielding
6.9 g (23 mmol, 85%) of a colourless liquid. 19F NMR (CDCl3,
General Procedure for the Investigation of the Coordination Proper-
ties of P(C2F5)3, P(CF3)3, (C2F5)2POH and (CF3)2POH by Kinet-
ically Controlled Reactions with [Ni(CO)4]: Four equivalents of the
phosphane derivative were condensed onto a solution of nickel tet-
racarbonyl in CH2Cl2. After stirring for 24 h, the progress of the
reaction was monitored by IR and NMR spectroscopy (some of
the NMR spectra are of higher order, in these cases best estimates
of the coupling constants are provided.).
3
20 °C): δ = –80.2 [d, J(P,F) = 16 Hz, CF3], –115.9 (m, CF2) ppm.
31P NMR (CDCl , 20 °C): δ = 62.2 (m) ppm. IR (gas phase): ν =
˜
3
532 (w), 555 (w), 749 (w), 951 (m), 1117 (w), 1141 (w), 1169 (m),
[Ni(CO)3{P(C2F5)3}]: 31P NMR (CH2Cl2, [D6]acetone ext. lock,
1230 (s), 1304 (m) cm–1.
20 °C): δ = 56.7 [sept, decet, 2J(P,F) = 60, 3J(P,F) = 6 Hz] ppm. 19
F
Eur. J. Inorg. Chem. 2013, 3392–3404
3401
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
3
NMR (CH2Cl2, [D6]acetone ext. lock, 20 °C): δ = –78.5 [d, J(P,F)
matrix least-squares approach.[33] All non-hydrogen atoms were re-
fined with anisotropic displacement factors.
2
= 6 Hz, CF3], –106.9 [d, J(P,F) = 60 Hz, CF2] ppm. IR (CH2Cl2):
ν = 2110 [m, ν (CO)], 2051 [vs, ν (CO)] cm–1. IR (gas phase): ν =
˜
˜
s
as
Data for [Hg(CN){P(C2F5)2}(dppe)]: Colourless crystal, Mr
=
2114 [m, νs(CO)], 2060 [s, νas(12CO)], 2023 [w, νas(13CO)], 1303 [s,
ν(CC)], 1237 [vs, ν(CF), CF3], 1172 [s, ν(CF), CF2], 1107 [w, ν(CF),
CF2], 950 [s, ν(PC)], 751 [w, δ(CF3)], 630 (w) cm–1. MS (EI-TOF,
pos., 30 eV): m/z (%) = 529.9 (33) [M]+, 501.9 (88) [M – CO]+,
473.9 (74) [M – 2CO]+, 454.9 (31) [M – 3CO]+; 387.9 (16)
[P(C2F5)3]+, 141.9 (100) [Ni(CO)3]+, 118.9 (32) [C2F5]+, 113.9 (33)
[Ni(CO)2]+, 100 (31) [C2F4]+, 58 (18) [Ni]+.
¯
986.15, triclinic, space group P1, a = 9.102(1), b = 11.776(2), c =
18.358(2) Å, α = 89.34(1), β = 89.07(1), γ = 76.50(1)°, V =
1913.0(4) Å3, Z = 2, ρcalcd. = 1.712 gcm–3, F(000) = 964, 41781
reflections collected up to θ = 27.5°, 8734 independent reflections
of which 7704 with IϾ2σ(I), 479 parameters. Hydrogen atoms were
taken into account at idealized positions by using a riding model.
R values: R1 = 0.024 for reflections with IϾ2σ(I), wR2 = 0.045 for
all data.
[Ni(CO)3{P(CF3)3}]: 31P NMR (CDCl3, 20 °C): δ = 50.3 [decet,
2J(P,F) = 87 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –58.4 [d,
Data for [Hg{P(C2F5)2}2(dppe)]·C7H8: Colourless crystal, Mr
1229.14, monoclinic, space group Pc, a = 9.664(1), b = 11.690(1),
c = 20.679(2) Å, β = 96.49(1)°, V = 2321.2(3) Å3, Z = 2, ρcalcd.
=
2J(P,F) = 87 Hz] ppm. IR (CH Cl ): ν = 2110 [m, ν (CO)], 2037
˜
2
2
s
[vs, νas(CO)] cm–1.
=
[Ni(CO)2{P(CF3)3}2]: 31P NMR (CDCl3, 20 °C): δ = 52.0 [m,
1.759 gcm–3, F(000) = 1196, 136717 reflections collected up to θ =
30°, 13386 independent reflections of which 12286 with IϾ2σ(I),
596 parameters. Hydrogen atoms were taken into account at ideal-
ized positions by using a riding model. Hydrogen atoms were re-
fined isotropically. R values: R1 = 0.022 for reflections with
IϾ2σ(I), wR2 = 0.039 for all data.
2J(P,F) ≈ 93 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –56.8 [m,
2J(P,F) ≈ 93 Hz] ppm. IR (CH Cl ): ν = 2096 [ν (CO)] cm–1.
˜
2
2
as
[Ni(CO)3{P(C2F5)2OH}]: 31P NMR (CDCl3, 20 °C): δ = 119.2 [m,
2J(P,F) ≈ 73 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –78.7 (m,
CF ), –121.3 [m, 2J(P,F) ≈ 73 Hz, CF ] ppm. IR (CH Cl ): ν = 2100
˜
3
2
2
2
[m, νs(CO)], 2036 [vs, νas(CO)] cm–1.
Data for [Ni(CO)3{P(CF3)2OH}]·H2O: Colourless crystal, Mr
346.75, orthorhombic, space group Pca21, a = 7.988(1), b =
11.855(1), c = 12.752(1) Å, V = 1207.6(1) Å3, Z = 4, ρcalcd.
=
[Ni(CO)2{P(C2F5)2OH}2]: 31P NMR (CDCl3, 20 °C): δ = 124.6 [m,
2J(P,F) ≈ 73 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –78.4 (m,
=
1.907 gcm–3, F(000) = 680, 21971 reflections collected up to θ =
27.0°, 2594 independent reflections of which 2464 with IϾ2σ(I),
177 parameters. R values: R1 = 0.018 for reflections with IϾ2σ(I),
wR2 = 0.0434 for all data.
2
CF3), –124.6 [m, J(P,F) = 83 Hz, CF2] ppm.
[Ni(CO)3{P(CF3)2OH}]: 31P NMR (CDCl3, 20 °C): δ = 112.6 [sept,
2J(P,F) = 91 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –70.3 [d,
2J(P,F) = 91 Hz] ppm. IR (CH Cl ): ν = 2102 [m, ν (CO)], 2038
˜
2
2
s
[vs, νas(CO)] cm–1.
CCDC-920714 (for [Hg(CN){P(C2F5)2}(dppe)]·C7H8), CCDC-
920715 (for [Hg{P(C2F5)2}2(dppe)]·C7H8) and CCDC-920716 (for
[Ni(CO)3{P(CF3)2OH}]·H2O) contain 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.
[Ni(CO)2{P(CF3)2OH}2]: 31P NMR (CDCl3, 20 °C): δ = 115.1 [m,
2J(P,F) ≈ 93 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –69.7 [d,
2J(P,F) = 93 Hz] ppm. IR (CH Cl ): ν = 2075 [m, ν (CO)], 2033
˜
2
2
s
[vs, νas(CO)] cm–1.
[Ni(CO){P(CF3)2OH}3]: 31P NMR (CDCl3, 20 °C): δ = 115.3 [m,
Computational Analysis: Theoretical studies of (C2F5)2PH and
(C2F5)2POH were carried out by DFT and Møller–Plesset second-
order perturbation theory (MP2) calculations. Preliminary calcula-
tions, intended for obtaining starting geometries and Hessian ma-
trices for further higher level computations, as well as for the evalu-
ation of vibrational corrections, were carried out by using the 6-
31++G**[34] basis sets. In the geometry optimizations and Hessian
computations performed by using the B3LYP hybrid functional (re-
ferred to as B3LYP/cc-pVTZ), C, H and P atoms were described by
the triple-ꢀ valence correlation-consistent cc-pVTZ basis sets and F
atoms by the aug-cc-pVTZ basis set.[35] In calculations performed
with the O3LYP hybrid functional and MP2, the aug-cc-pVTZ ba-
sis sets[35] were used for describing all atoms (these levels of theory
are referred to as O3LYP/aug-cc-pVTZ and MP2/aug-cc-pVTZ,
respectively). Preliminary calculations and all B3LYP/cc-pVTZ,
O3LYP/aug-cc-pVTZ and MP2/aug-cc-pVTZ calculation were per-
formed by using the Firefly 7.1.G QC package,[36] which is partially
based on the GAMESS (US)[37] source code. Third derivatives of
the potential energy and two-dimensional potential energy scans
along the C–C–P–C torsional angles were calculated by using the
Gaussian 03 software[38] at the B3LYP/6-31++G** level of theory.
2J(P,F) ≈ 91 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –69.6 [m,
2J(P,F) = 91 Hz] ppm. IR (CH Cl ): ν = 2056 [m, ν(CO)] cm–1.
˜
2
2
[Ni{P(CF3)2OH}4]: 31P NMR (CDCl3, 20 °C): δ = 115.4 [m, 2J(P,F)
≈ 91 Hz] ppm. 19F NMR (CDCl3, 20 °C): δ = –69.2 [m, J(P,F) ≈
2
91 Hz] ppm.
Synthesis of [Hg{P(C2F5)2}2(dppe)]: (C2F5)2PH (8.0 mmol) was
condensed onto a solution of [Hg(CN)2(dppe)] (1.76 g, 2.7 mmol)
in CH2Cl2 and warmed to room temperature. After stirring for
10 min, all the volatile compounds were removed in vacuo yielding
1
2.98 g (2.6 mmol, 96%) of a light-yellow solid. H NMR (CDCl3,
20 °C): δ = 2.7 (s, C2H4), 7.5 (m, C6H5) ppm. 13C{1H} NMR
(CDCl3, 20 °C): δ = 24.4 (m, C2H4), 129.4 (s, C-m), 130.4 (m, C-
i), 131.3 (s, C-p), 132.5 (m, C-o) ppm. 13C{19F} NMR (CDCl3,
20 °C): δ = 120.0 [d, 2J(C,P) = 23 Hz, CF3], 124.1 [d, 1J(C,P) =
76 Hz, CF2] ppm. 19F NMR (CD2Cl2, 247 K): δ = –83.5 [d, 3J(P,F)
2
2
3
= 14 Hz, CF3], –89.3 [dd, J(F,F) = 291, J(P,F) = 65, J(Hg,F) =
250 Hz, CFaF], –98.5 [d, 2J(F,F) = 296, 3J(Hg,F) = 140 Hz,
2
CFFb] ppm. 31P{19F} NMR (CD2Cl2, 247 K): δ = 11.2 [t, J(P,P)
1
2
1
= 57, J(Hg,P) = 212 Hz, PPh2], –9.2 [t, J(P,P) = 58, J(Hg,P) =
824 Hz, P(C2F5)2] ppm. 199Hg{19F} NMR (CD2Cl2, 247 K): δ =
1
1
–296.2 [tt, J(Hg,Pdppe) = 209, J(Hg,PC2F5) = 824] ppm.
Molecular dynamics (MD) calculations were performed by using
the CP2K code.[39] The Quickstep method for density functional
X-ray Crystallography: Data collections for the X-ray diffraction
analyses of [Hg{P(C2F5)2}2(dppe)]·C7H8, [Hg(CN){P(C2F5)2}- calculations using a mixed Gaussian and plane waves approach was
(dppe)] and [Ni(CO)3{P(CF3)2OH}]·H2O were performed with a
Bruker–Nonius KappaCCD diffractometer at 100(2) K by using
graphite-monochromated Mo-Kα radiation (λ = 0.71073 Å). The
structures were solved by direct methods and refined by the full-
applied.[40] To model an isolated molecule by using this periodic
code, a single molecule was simulated in a cubic supercell with a
size of 13 Å. The BLYP (Becke–Lee–Yang–Parr) exchange-corre-
lation functional, Goedecker, Teter and Hutter (GTH) pseudopo-
Eur. J. Inorg. Chem. 2013, 3392–3404
3402
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tentials[41] and the corresponding DZVP basis sets optimized for
the GTH pseudopotentials and the BLYP functional were used.
The equilibrium geometries of (C2F5)2PH (I and II) and (C2F5)2-
POH (conformers IA and IIA) were optimized starting from the
geometries calculated at the 6-31G++(d,p) level, and these struc-
tures were subsequently used to determine vibrational corrections
by using MD simulations. The simulations were performed in the
canonical ensemble(NVT) by using a chain of five Nose–Hoover
thermostats with a time constant of 4 fs to regulate the simulation
temperature at 300 K, approximately the temperature of the GED
experiment. The simulations were performed with a time step of
0.5 fs and lasted for 21 ps [(C2F5)2PH] and 20 ps [(C2F5)2POH].
The technique for calculating vibrational corrections by using mo-
refinement with kMD corrections [and also in the study of (C2F5)2-
POH].
The independent valence angles used were α(C2–P1–C3), α(C4–C2–
P1), α(F6–C2–P1) and α(F10–C4–C2); the differences between
α(C4–C2–P1) and α(C5–C3–P1) were fixed. The position of the F6
atom was described by the α(F6–C2–C4) angle, and its difference
with α(F6–C2–P1) was fixed. The F7, F8 and F9 atoms were de-
scribed in a similar manner with all differences of angles also fixed.
The γ(C4–C2–P1–C3) and γ(F10–C4–C2–P1) dihedral angles were
independent parameters; the differences between γ(C5–C3–P1–C2)
and γ(C4–C2–P1–C3) and γ(F11–C5–C3–P1) and γ(F10–C4–C2–
P1) were fixed. The position of F12 was described by the α(F12–
C4–F10) angle, and its difference with α(F10–C4–C2) was fixed.
The F13, F14 and F15 atoms were described similarly. The position
of the H16 atom was fixed relative to P1. The rms vibrational am-
plitudes were refined in groups corresponding to distinct peaks on
the radial distribution curve.
lecular dynamics data is described in refs.[29,30]
.
In this contribution the GED/MD technique[29] was further devel-
oped to allow computation of vibrational corrections for several
conformers, which are present in the same MD trajectory (or sev-
eral trajectories). Geometries from the trajectories’ time steps are
classified according to the value of a dihedral or torsional angle
such that the conformation and vibrational corrections are com-
puted for each conformer separately by using only geometries that
were determined to belong to this particular conformation.
The model of (C2F5)2POH was very similar to that of (C2F5)2PH,
described above, with an additional three independent parameters
[r(O16–P1), α(O16–P1–C2), γ(O16–P1–C2–C3)] describing the oxy-
gen atom. The position of the H17 atom was fixed relative to that
of O16. All fixed differences were taken from the O3LYP/aug-cc-
pVTZ calculations.
Electron Diffraction: Electron scattering intensities for (C2F5)2PH
and (C2F5)2POH were recorded at room temperature by using a
combination of reusable Fuji and Kodak imaging plates and a
strongly modified Balzers KD-G2 Gas Eldigraph.[42] For both mo-
lecules, two sets of electron diffraction patterns were obtained from
long (L = 500 mm) and short (L = 250 mm) camera distances at
the accelerating voltage of 50 kV [four and three patterns for (C2F5)2-
PH and four and six patterns for (C2F5)2POH from long and short
camera distances, respectively]. The data reduction, which yields
molecular intensity curves (see the Supporting Information), the
electron wavelength determination and the molecular structure re-
finement were performed by using the UNEX program.[43]
Supporting Information (see footnote on the first page of this arti-
cle): Calculated relative energies and abundances for conformers of
(C2F5)2PH and (C2F5)2POH; potential energy surface for B con-
formers of (C2F5)2POH; experimental and theoretical geometrical
parameters of (C2F5)2PH; total and molecular electron diffraction
intensities of (C2F5)2PH and (C2F5)2POH; anharmonic vibrational
corrections for (C2F5)2PH; experimental and theoretical Cartesian
coordinates of (C2F5)2PH and (C2F5)2POH.
Acknowledgments
Several sets of vibrational corrections were used for the analysis of
(C2F5)2PH data. First, kh1 corrections from the SHRINK pro-
gram[31] (so-called “second approximation”), calculated by using
the B3LYP/cc-pVTZ force field, were used. It is a typical approach
for large and flexible molecules successfully used in studies.[32,44,45]
Secondly, anharmonic corrections computed from the molecular
dynamics trajectories[29,30] (kMD) were applied. We also calculated
third derivatives for the major conformer at the B3LYP/6-31++G**
level and computed the anharmonic vibrational corrections by
using SHRINK (k3). The Q2SHRINK and SHREx programs[32]
were used for making input files and extracting data from output
files of SHRINK. The (C2F5)2POH data was analysed by using
corrections computed from the molecular dynamics trajectories
(kMD).
The authors are grateful to the German Academic Exchange Ser-
vice (DAAD) for funding a visit of A. V. Z. to Bielefeld University
and to the Humboldt Foundation for a stipend for Yu. V. V. This
work is part of a Russian–German Cooperation Project (supported
by the Russian Foundation for Basic Research, grant number 12-
03-91333-NNIO_a, and the Deutsche Forschungsgemeinschaft
(DFG), OB28/22-1). Support by the Deutsche Forschungsgemein-
schaft through the program Core Facilities for GED@BI is grate-
fully acknowledged. Merck KGaA, Darmstadt, Germany is ac-
knowledged for financial support. The authors thank Dr. N. Ignat-
iev (Merck KGaA) for helpful discussions.
[1] a) S. Lou, G. C. Fu, Adv. Synth. Catal. 2010, 352, 2081; b)
A. F. Littke, G. C. Fu, Angew. Chem. 1998, 110, 3586; Angew.
Chem. Int. Ed. 1998, 37, 3387; c) A. F. Littke, G. C. Fu, Angew.
Chem. 1999, 111, 2568; Angew. Chem. Int. Ed. 1999, 38, 2411;
d) A. F. Littke, C. Dai, G. C. Fu, J. Am. Chem. Soc. 2000, 122,
4020; e) A. F. Littke, G. C. Fu, J. Org. Chem. 1999, 64, 10; f)
J. Kingston, J. Verkade, J. Org. Chem. 2007, 72, 2816.
[2] W. Mahler, Inorg. Chem. 1979, 18, 352.
[3] S. A. Hayes, R. J. F. Berger, B. Neumann, N. W. Mitzel, J.
Bader, B. Hoge, Dalton Trans. 2010, 39, 5630.
[4] S. A. Hayes, R. J. F. Berger, N. W. Mitzel, J. Bader, B. Hoge,
Chem. Eur. J. 2011, 17, 3968.
[5] S. Basu, N. Arulsamy, D. M. Roddick, Organometallics 2008,
27, 3659.
[6] M. Ernst, D. M. Roddick, Inorg. Chem. 1989, 28, 1624.
[7] R. D. Chambers, Fluorine in Organic Chemistry, Wiley, New
York, 1973, p. 345.
Nine independent parameters were used in the model of (C2F5)2-
PH and were refined by least-squares analysis. The independent
distances were r(C2–P1), r(C4–C2) and r(C2–F6), and the differ-
ences between r(C2–P1) and r(C3–P1), r(C4–C2) and r(C5–C3).
r(C2–F6) and other r(C–F) parameters were fixed at theoretically
calculated values. This was justified by the fact that the differences
are small and not determinable from the GED data and that the
values yielded by calculations at different levels of theory are very
similar. Therefore the approach with flexible restraints based on
different computations does not have an advantage. Initially, all
fixed differences mentioned here were taken from the B3LYP/cc-
pVTZ calculations. Then, when we found that the O3LYP/aug-cc-
pVTZ values of the bond lengths are closer to the experimental
values, we used the differences from these calculations for the final
Eur. J. Inorg. Chem. 2013, 3392–3404
3403
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[8] a) M. Henrich, A. Marhold, A. Kolomeitsev, A. Kadyrov, G.-V.
Röschenthaler, J. Barten (Bayer AG, Leverkusen), EP1266902,
2002 (SciFinder Scholar 2002:963725); b) M. H. Königsmann,
Ph. D. Thesis, University of Bremen, Germany, 2004.
[9] N. E. Shevchenko, V. G. Nenajdenko, G.-V. Röschenthaler, J.
Fluorine Chem. 2008, 129, 390.
[10] M. B. Murphy-Jolly, L. C. Lewis, A. J. M. Caffyn, Chem. Com-
mun. 2005, 4479.
[11] a) F. W. Bennett, H. J. Emeléus, R. N. Haszeldine, J. Chem.
Soc. 1953, 1565; b) H. J. Emeléus, J. D. Smith, J. Chem. Soc.
1959, 375; c) A. H. Cowley, T. A. Furtsch, D. S. Dierdorf, J.
Chem. Soc. C 1970, 523.
[33]
[34]
SHELX-97: G. M. Sheldrick, Acta Crystallogr., Sect. A 2008,
64, 112.
a) W. J. Hehre, R. Ditchfield, J. A. Pople, J. Chem. Phys. 1972,
56, 2257; b) M. M. Francl, W. J. Petro, W. J. Hehre, J. S. Bin-
kley, M. S. Gordon, D. J. DeFrees, J. A. Pople, J. Chem. Phys.
1982, 77, 3654.
a) T. H. Dunning Jr., J. Chem. Phys. 1989, 90, 1007; b) D. E.
Woon, T. H. Dunning Jr., J. Chem. Phys. 1993, 98, 1358.
A. A. Granovsky, Firefly, v. 7.1.G; http://classic.chem.msu.su/
[35]
[36]
[37]
gran/firefly/index.html.
M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert,
M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A.
Nguyen, S. Su, T. L. Windus, M. Dupuis, J. A. Montgomery,
J. Comput. Chem. 1993, 14, 1347.
[12] H. G. Ang, M. E. Redwood, B. O. West, Aust. J. Chem. 1972,
25, 493.
[13] a) B. Hoge, J. Bader, B. Kurscheid, N. Ignatiev, E. Aust (Merck,
Darmstadt, Germany), WO 2010/009818, 2010 (SciFinder
Scholar 2010:111024); b) B. Hoge, J. Bader, H. Beckers, Y. S.
Kim, R. Eujen, H. Willner, N. Ignatiev, Chem. Eur. J. 2009, 15,
3567; c) J. Bader, R. J. F. Berger, H.-G. Stammler, N. W. Mitzel,
B. Hoge, Chem. Eur. J. 2011, 17, 13420; d) B. Kurscheid, L.
Belkoura, B. Hoge, Organometallics 2012, 31, 1329.
[14] U. Welz-Biermann, N. Ignatyev, M. Weiden, M. Schmidt, U.
Heider, A. Miller, H. Willner, P. Sartori (Merck, Darmstadt,
Germany), WO 03/087113, 2003 (SciFinder Scholar
2003:837099).
[15] B. Hoge, P. Garcia, H. Willner, H. Oberhammer, Chem. Eur. J.
2006, 12, 3567.
[16] A. Kolomeitsev, M. Görg, U. Dieckbreder, E. Lork, G.-V.
Röschenthaler, Phosphorus Sulfur Silicon Relat. Elem. 1996,
109–110, 597.
[17] M. F. Ernst, D. M. Roddick, Inorg. Chem. 1989, 28, 1624.
[18] T. Mahmood, J. M. Shreeve, Inorg. Chem. 1986, 25, 3128.
[19] A. B. Burg, G. Brendel, J. Am. Chem. Soc. 1958, 80, 3198.
[20] B. Hoge, C. Thösen, I. Pantenburg, Inorg. Chem. 2001, 40,
3084.
[21] Y. Morino, Acta Crystallogr. 1960, 13, 1107.
[22] Y. Morino, J. Nakamura, P. W. Moore, J. Chem. Phys. 1962,
36, 1050.
[23] L. Hedberg, I. M. Mills, J. Mol. Spectrosc. 1993, 160, 117.
[24] L. Hedberg, I. M. Mills, J. Mol. Spectrosc. 2000, 203, 82.
[25] V. A. Sipachev, THEOCHEM 1985, 22, 143.
[26] V. A. Sipachev, in: Advances in Molecular Structure Research
(Ed.: I. Hargittai), JAI, Greenwich, 1999, vol. 5, p. 323.
[27] V. A. Sipachev, Struct. Chem. 2000, 11, 167.
[28] V. A. Sipachev, J. Mol. Struct. 2001, 67, 567.
[29] D. A. Wann, A. V. Zakharov, A. M. Reilly, P. D. McCaffrey,
D. W. H. Rankin, J. Phys. Chem. A 2009, 113, 9511.
[30] D. A. Wann, R. J. Less, F. Rataboul, P. D. McCaffrey, A. M.
Reilly, H. E. Robertson, P. D. Lickiss, D. W. H. Rankin, Orga-
nometallics 2008, 27, 4183.
[38]
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T.
Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R.
Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morok-
uma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzew-
ski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian 03,
rev. E.01, Gaussian, Inc., Wallingford, CT, 2004.
[39]
CP2K, v. 2.2.334 (development version), CP2K developers
group, 2011, http://cp2k.org; CP2K is available free of charge
on the www: http://cp2k.berlios.de/.
[40] J. Van de Vondele, M. Krack, F. Mohamed, M. Parrinello, T.
Chassaing, J. Hutter, Comput. Phys. Commun. 2005, 167, 103.
[41] a) S. Goedecker, M. Teter, J. Hutter, Phys. Rev. B 1996, 54,
1703; b) C. Hartwigsen, S. Goedecker, J. Hutter, Phys. Rev. B
1998, 58, 3641.
[42] a) R. J. F. Berger, M. Hoffmann, S. A. Hayes, N. W. Mitzel, Z.
Naturforsch. B 2009, 64, 1259; b) W. Zeil, J. Haase, L. Weg-
mann, Z. Instr. 1966, 74, 84.
[43] a) Yu. V. Vishnevskiy, UNEX, v. 1.5, 2011, http://unexprog.org;
b) Yu. V. Vishnevskiy, J. Mol. Struct. 2007, 871, 24.
[44] A. V. Zakharov, S. A. Shlykov, Yu. A. Zhabanov, G. V. Gir-
ichev, Phys. Chem. Chem. Phys. 2009, 11, 3472.
[45] A. V. Zakharov, S. L. Masters, D. A. Wann, S. A. Shlykov,
G. V. Girichev, S. Arrowsmith, D. B. Cordes, P. D. Lickiss,
A. J. P. White, Dalton Trans. 2010, 39, 6960.
[31] a) V. A. Sipachev, J. Mol. Struct. 1985, 121, 143; b) V. A. Sipa-
chev, Struct. Chem. 2000, 11, 167; c) V. A. Sipachev, J. Mol.
Struct. 2001, 67, 567.
[32] A. V. Zakharov, S. A. Shlykov, E. A. Danilova, A. V. Krasnov,
M. K. Islyaikin, G. V. Girichev, Phys. Chem. Chem. Phys. 2009,
11, 8570.
Received: March 17, 2013
Published Online: May 21, 2013
Eur. J. Inorg. Chem. 2013, 3392–3404
3404
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim