The Al(ORF)3/H2O/phosphane [RF = C(CF3)3] System - Protonation of phosphanes and absolute bronsted acidity

Article abstract of DOI:10.1002/ejic.201300031

The synthesis of the classical, neutral donor-acceptor adducts Ph ₂MeP-/Ph₃P-/Ph₃As-Al(OR^F) ₃ and H₂O-Al(OR^F)₃ [1, 2, 3, 4, OR^F = OC(CF₃)₃] is reported. The intermediate H₂O-Al(OR^F)₃ (4) was generated by substitution of PhF in PhF-Al(OR^F)₃ with H₂O and was analyzed in a long-term NMR study over 22 days. This Bronsted acidic system was used in orienting experiments to protonate phosphanes such as PMePh₂, PPh₃, PCy₃, P(tBu)₃, and PCy₂[2,4,6-(iPr)₃C₆H₂]. Depending on the use of one or two equivalents of PhF-Al(OR^F)₃, the new weakly coordinating anions [(R^FO)₃Al(μ-OH) Al(OR^F)₃]^- or [HOAl(OR^F) ₃]^- were obtained. However, in dependence of the steric bulk of the phosphanes, stable and unreactive R₃P-Al(OR ^F)₃ adducts were also observed in the NMR experiments. The absolute acidity of the key H₂O-Al(OR^F)₃ adduct was evaluated by the relaxed COSMO cluster-continuum (rCCC, COSMO = conductor-like screening model) model in fluorobenzene solution. For a 0.001 M solution of H₂O-Al(OR^F)₃, the medium acidity resulted as -986 kJ mol^-1 or a pH_abs value of 173. Long-term hydrolysis of H₂O-Al(OR^F)₃ (4), probably to give HOR^F and HOAl(OR^F)₂ followed by trimerization, gave [HOAl(OR^F)₂]₃ (10), which was identified by X-ray diffraction. Small donor ligands such as Ph ₂MeP, Ph₃P, Ph₃As, or even H₂O form classical donor-acceptor adducts with the Lewis superacid Al(OR ^F)₃ [R^F = C(CF₃)₃]. FLP-like (FLP = frustrated Lewis pair) combinations of this Lewis acid, phosphanes, and water then lead to [HPR₃]⁺ and two new weakly coordinating anions [HOAl(OR^F)₃]^- and [(^FRO)₃Al(μ-OH)Al(OR^F)₃] ^-. The absolute acidity of H₂O-Al(OR^F) ₃ is evaluated.

Full text of DOI:10.1002/ejic.201300031

FULL PAPER
DOI:10.1002/ejic.201300031
The Al(ORF)3/H2O/Phosphane [RF = C(CF3)3] System –
Protonation of Phosphanes and Absolute Brønsted Acidity
Anne Kraft,[a] Josephine Possart,[a] Harald Scherer,[a]
Jennifer Beck,[a] Daniel Himmel,[a] and Ingo Krossing*[a]
Keywords: Acidity / Lewis acids / Frustrated Lewis pairs / Phosphanes / Arsanes / Weakly coordinating anions
The synthesis of the classical, neutral donor–acceptor ad-
ducts Ph2MeP–/Ph3P–/Ph3As–Al(ORF)3 and H2O–Al(ORF)3 [1,
2, 3, 4, ORF = OC(CF3)3] is reported. The intermediate H2O–
Al(ORF)3 (4) was generated by substitution of PhF in PhF–
Al(ORF)3 with H2O and was analyzed in a long-term NMR
study over 22 days. This Brønsted acidic system was used in
orienting experiments to protonate phosphanes such as
dependence of the steric bulk of the phosphanes, stable and
unreactive R3P–Al(ORF)3 adducts were also observed in the
NMR experiments. The absolute acidity of the key H2O–
Al(ORF)3 adduct was evaluated by the relaxed COSMO clus-
ter-continuum (rCCC, COSMO = conductor-like screening
model) model in fluorobenzene solution. For a 0.001
M solu-
tion of H2O–Al(ORF)3, the medium acidity resulted as
PMePh2, PPh3, PCy3, P(tBu)3, and PCy2[2,4,6-(iPr)3C6H2]. De- –986 kJmol–1 or a pHabs value of 173. Long-term hydrolysis
pending on the use of one or two equivalents of PhF–Al-
(ORF)3, the new weakly coordinating anions [(RFO)3Al(μ-
OH)Al(ORF)3]– or [HOAl(ORF)3]– were obtained. However, in
of H2O–Al(ORF)3 (4), probably to give HORF and HOAl-
(ORF)2 followed by trimerization, gave [HOAl(ORF)2]3 (10),
which was identified by X-ray diffraction.
based cationic complexes stabilized by [(C6F5)3B(μ-OH)B-
(C6F5)3]–[13] or [(C6F5)3BOH···H2O–B(C6F5)3]–.[14,15] This
system and other donor–acceptor complexes were also
studied theoretically by Timoshkin and Frenking.[5] In these
complexes, the Brønsted acidity of H2O is significantly in-
creased owing to the complexed Lewis acid. Thus, a
stronger Lewis acid should further increase the proton acid-
ity. The use of one of the strongest monomeric Lewis su-
peracids, Al(ORF)3,[16,17] should result in one of the highest
available Brønsted acidities.
Our route into this topic was indirect. We were inspired
by the recent work of Stephan and Erker dealing with the
activation of small molecules such as H2, CO2, and ole-
fins,[18–21] and we were interested to test the monomeric
Lewis superacid Al(ORF)3 [ORF = OC(CF3)3] for FLP-like
(FLP = frustrated Lewis pair) chemistry. Closely related
FLP systems containing phosphanes and boranes had al-
ready been investigated theoretically and experimen-
tally.[22–24] The great advantage of phosphane bases is the
huge variety of available compounds. Also, the mixtures are
easy to investigate by NMR spectroscopy. Thus, 1:1 mix-
tures of various phosphanes and PhF–Al(ORF)3 under
1 atm H2 were analyzed. In all tested reactions, no H2 acti-
vation could be realized. Instead, mainly adduct formation
was observed with the sterically less demanding phosphanes
such as PPh3.[25] Nevertheless, in many reactions protonated
phosphonium ions were detected as side-products by NMR
spectroscopy. After solving the X-ray structure of a crystal-
line material that included a protonated [HPR3]+ phos-
phonium cation partnered with a new hydroxo aluminate
Introduction
The importance of a controlled synthesis of alumoxanes
and alumoxane hydroxides, which are used, for example,
as catalysts for polymerization reactions, has inspired the
interest of many research groups.[1–4] As a completely water-
free environment is almost impossible to realize and side-
reactions are often thermodynamically favored, the hydroly-
sis products cannot be neglected and change the Lewis acid
reactivity in many cases.[5] Thus, Roesky and others system-
atically analyzed the influence of water on group 13 organo-
metallic compounds, for example.[2] A special challenge still
remains with the synthesis of neutral Lewis acid–base ad-
ducts of water and alanes. To the best of our knowledge,
only a few neutral adducts with a 1:1 constitution can be
cited here: H2O–AlMes3,[6] H2O–Al(C6F5)3,[7] H2O–Al-
(OSiPh3)3(THF)2,[8] H2O–AlMe(OAr)2(THF)2 (–40 °C,
THF = tetrahydrofuran),[9] H2O–AlX3 (X = Br, Cl).[10]
Closely related to this type of adducts and intensively
studied is the H2O–B(C6F5)3 system, for which a number
of water complexes were also structurally characterized (see
refs.[11,12] and literature cited therein). For example, H2O–
B(C6F5)3 was used to generate [M(η-C5H5)2]+ (M = Cr, Fe,
Co),[11] an In-based triple-decker sandwich complex, or Ir-
[a] Institut für Anorganische und Analytische Chemie, Freiburger
Materialforschungszentrum (FMF) and Freiburg Institute for
Advanced Studies (FRIAS), Universität Freiburg,
Albertstr. 19, 79104 Freiburg, Germany.
Homepage: http://portal.uni-freiburg.de/molchem/lehrstuhl
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300031
Eur. J. Inorg. Chem. 2013, 3054–3062
3054
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[HO(Al(ORF)3)2]–, it was clear to us that traces of water shifted by 4.9 ppm in comparison to that of free water in
were present in the reaction mixtures. After determination PhF (δ = 1.13 ppm). The low-intensity signal at δ =
of the H2O content of the solvent as 9 ppm by Karl Fischer 3.57 ppm (Figure 1) resulted from the free alcohol
titration, we systematically studied the influence of water HOC(CF3)3. The intensity of the alcohol signal increased
on the phosphane/Al(ORF)3 system. Therefore, in initial ex- over time, as expected for the slow hydrolysis of Al(ORF)3;
periments we protonated phosphanes with intentionally the signal at δ = 6.03 ppm decreased accordingly. The same
formed H2O–Al(ORF)3 (4). We observed the formation of tendencies were observed in the 19F NMR spectra. Further
the two new weakly coordinating anions [HOAl(ORF)3]– support for our assignment gives the well fitting, DFT-cal-
and [(FRO)3Al(μ-OH)Al(ORF)3]–, which are isostructural culated 1H NMR chemical shift of 4 of δ = 6.32 ppm
to the B(C6F5)3 analogs.[11,13,15] The formation and hydroly- [(RI-)BP86/def2-TZVPP,[32,33] reference system H2O in PhF
sis of 4 was monitored by NMR experiments. In one of at δ = 1.13 ppm].
these reactions, the trimerization product [HOAl(ORF)2]3
(10) was characterized by single-crystal X-ray crystallogra-
phy.
Results and Discussion
The general route to the donor–acceptor systems follows
Equation (1). With PMePh2, PPh3, and AsPh3, stable ad-
duct formation was directly observed; the three adducts 1–
3 were characterized by NMR spectroscopy and single-crys-
tal X-ray crystallography. Furthermore, their calculated
dissociation enthalpies of +92, +77, and +60 kJmol–1
Figure 1. 1H NMR spectra (400.17 MHz, 298 K, PhF) of 4 mea-
underline the strong Lewis acidity of Al(ORF)3 [ΔrH°(g)
,
1
sured directly and after 22 days. Spectra were calibrated to the H
(RI-)BP86/def-TZVP,[26–31] 298 K, Equation (1)]. As ex-
pected for donor ligands, the phosphanes and the arsane
aremorestronglycoordinatedthantheweakbasesPhFandSO2
(–23 and –49 kJmol–1).
signal of the o-hydrogen atoms of PhF at δ = 7.14 ppm. The signal
at δ = 5.39 ppm is so far unidentified.
Protonation of Phosphanes
Sterically demanding phosphanes such as PCy2(2,4,6-
(iPr)3C6H2), PCy3, and P(tBu)3 [Tolman cone[34] angles:
P(tBu)3 182°,[35] PCy3 170°[35]] as well as less crowded ones
such as PPh3 and PMePh2 (PPh3 145°,[34,35] PMePh2
136°[34,35]) were (partially) protonated with the H2O/PhF/
Al(ORF)3 mixture (Table 1) in NMR tube reactions. De-
pending on the use of one or two equivalents of PhF–
Al(ORF)3, the terminal hydroxide [HOAl(ORF)3]– or the
bridged [(FRO)3Al(μ-OH)Al(ORF)3]– anion formed. The
calculated reaction enthalpies agree with the experimental
observations. The formation of the hydroxo-bridged anion
is more favored at the (RI-)BP86/def-TZVP level (Table 1);
values in italics were calculated by using Grimme’s D3[36,37]
dispersion correction, and Gibbs solvation energies were
(1)
To learn more about the stability of Al(ORF)3 in the
presence of water, we performed an NMR study of 4 in
monofluorobenzene, which revealed the unexpected sta-
bility of adduct 4. Thus, we dissolved PhF–Al(ORF)3 in
PhF that contained 173 ppm of H2O as measured by a Karl
Fischer titration. In agreement with the hard σ-donor prop-
erties of H2O and the oxophilicity of aluminum, H2O sub-
stituted the weak ligand PhF in the Lewis superacid PhF–
Al(ORF)3. In NMR spectra of directly prepared 4, as well
as in spectra recorded previously of other systems, we ob-
served a sharp singlet at δ ≈ 6 ppm, which showed no cross-
Table 1. COSMO-solvated computed Gibbs energies [kJmol–1] for
the protonation of phosphanes [(RI-)BP86/def-TZVP level; εr(PhF)
= 5.42]. Values in italics were calculated by including Grimme’s D3
dispersion correction.
peaks in an H–13C HMBC or H–13C heteronuclear single
1
1
n Al(ORF)3
quantum coherence (HSQC) spectra. Thus, we suggest that
this signal belongs to a proton bound to a heteroatom. H
–
1
PR3 + H2O–Al(ORF)3
PR3
[R3PH]+ + [HO(Al(ORF)3)1+n
]
n = 0, 1
NMR spectra of the moist PhF solution were recorded di-
rectly, and after two, five, 30, and 45 h and 22 days (Fig-
ure 1 and Supporting Information). Here, also no cross-
peaks in the HMBC and HSQC spectra were detected.
ΔrG°(solv)
Compound n = 0
1
PMePh2
PPh3
PCy3
P(tBu)3
5
6
7
8
9
11
18
14 –54
21 –47
–118
–111
–40 –108 –172
–42 –106 –174
1
–44
–42
–20
Thus, we assigned the main signal in the H NMR spectra
at δ = 6.03 ppm to the hydrogen atoms of complexed H2O.
PCy2[2,4,6-(iPr)3C6H2]
–22 –85
–154
This 1H NMR signal of H2O is significantly downfield
Eur. J. Inorg. Chem. 2013, 3054–3062
3055
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
calculated with the conductor-like screening model ppm).[44] The 31P NMR chemical shifts are in the same
(COSMO)[38,39] by using εr(PhF) = 5.42.[40]
range as those observed for similar Al/B–P complexes
However, the reaction mixtures always showed several [MePh2P–BBr3 δ(31P) = –9.2,[43] MePh2P–AlMe3 δ(31P) =
byproducts in the NMR spectra. In particular, the forma- –24.2,[45] Ph3P–AlMe3 δ(31P) = –7.3 ppm].[45] For 2, the dis-
tion of competing adducts was observed for phosphanes tance between the two maxima in the 31P NMR spectra
such as PPh3 or PMePh2 with smaller Tolman cone angle.
did not change at different magnetic field strengths. This
All experimental observations are closely related to the characteristic shape of the 31P NMR signal results from P–
known equilibrium behavior of H2O–B(C6F5)3.[15] They are Al coupling. This proves that the phosphane is covalently
in good agreement with the theoretically predicted adduct bound to the Lewis acid. The complex formation and sta-
stabilities of the weak donor–acceptor adducts of B(C6F5)3 bility in solution could finally be proven with a 1H–19F het-
and Al(C6F5)3: thus, the gas phase energy for the formation eronuclear Overhauser effect spectroscopy (HOESY) ex-
of H2O–Al(ORF)3 of –94 kJmol–1 is in the same order of periment (see Supporting Information). The NMR spectro-
magnitude as that of H2O–Al(C6F5)3 (–112 kJmol–1) and, scopic data of 3 in 1,2-F2C6H4 are listed in the Exp. Sect.,
as expected, is higher than that of H2O–B(C6F5)3 but the proton chemical shifts can only be assigned to a
[–37 kJmol–1, all calculated at the (RI-)BP86/def2-TZVPP range of signals. However, NMR measurements of 3 in
level].[5,41] In analogy to the exothermic formation of CD2Cl2 solution showed the formation of another com-
[(C6F5)3X(μ-OH)X(C6F5)3]– {X = Al –199, B –63 kJmol–1, pound that is probably [Ph3As–CD2Cl]+[ClAl(ORF)3]–.
(RI-)BP86/def2-TZVPP},[5] the formation of [(FRO)3Al(μ- Dissociation of 3 to the weak base AsPh3 and Al(ORF)3
OH)Al(ORF)3]– is also thermodynamically favored by a might form an FLP system that attacks CD2Cl2 with chlor-
similar amount [–144 kJmol–1, Equation (2)].
ide abstraction and alkylation of the arsane. The signals at
δ = 51.2 ppm in the 27Al NMR spectrum and at δ =
–75.5 ppm in the 19F NMR spectrum are indicative of the
known [ClAl(ORF)3]– anion.[16] Furthermore, the signals
(2)
1
with H chemical shifts of 7.62, 7.78, and 7.92 ppm (ortho,
meta, para) are too sharp to result from an adduct (free
AsPh3: one broad signal at 7.34 pm). Consistent with the
increased central atom size (P vs. As) and the decreased
calculated adduct stability, dissociation of 3 and a further
reaction with the solvent CD2Cl2 seems to be reasonable.
Quantum chemical calculations support this assumption
[Equation (3), Gibbs solvation energies were calculated at
the (RI-)BP86/def-TZVP level with the COSMO model,
εr(CH2Cl2) = 8.93,[40] value in italics includes Grimme’s D3
correction, values are given in kJmol–1].
In a further attempt to crystallize compound 4, we iso-
lated crystals of the hydrolyzed product [HOAl(ORF)2]3
(10), which probably resulted from trimerization of the pri-
mary hydrolysis product HOAl(ORF)2. This is supported by
quantum chemical calculations and the similar behavior of
HOB(C6F5)2 (Figure 2).[42]
(3)
Figure 2. Hypothetic hydrolysis pathway leading to 10, and the cal-
culated energies (ΔrG°(solv)) at the (RI-)BP86/def-TZVP level at
298 K, COSMO model εr(PhF) = 5.42;[40] values are given in
A similar chloride abstraction from CD2Cl2 was also ob-
served when dissociated (CF3)3COH–Al(ORF)3 was reacted
kJmol–1, values in italics include Grimme’s D3 dispersion correc- with Me6C6 to yield [Me6C6–CD2Cl]+[ClAl(ORF)3]– and
tion.
free HOC(CF3)3 (NMR discussion, see Supporting Infor-
mation).
The protonation of the phosphanes with 4 was analyzed
with NMR spectroscopy. In all cases, the typical doublet of
the JP,H coupling of 400–500 Hz was observed in the H
NMR Spectroscopy
1
1
The NMR spectra of adducts 1 and 2 are as expected and 31P NMR spectra (Table 2).
and will only be discussed briefly. The 19F NMR signals at
Nevertheless, several side-products, such as unreacted
δ = –75.4 and –74.3 ppm as well as the broad 27Al NMR starting materials, neutral competing adducts, or other un-
signals at δ = 43.9 and 41.4 ppm are typical for strongly characterized hydrolysis products, were detected in low-to-
bound Al(ORF)3 complexes.[16] The line-width of the 27Al medium yields. Although a straight forward characteriza-
signal is too broad to show splitting by the P–Al coupling. tion of the cations was possible, the anion part was not as
1
The 31P NMR spectra showed a broad signal at δ = easy to interpret. In the H NMR spectra, singlets at δ ≈
[43]
–26.6 ppm for 1 (cf. free PMePh2 in THF: –26.1 ppm, in 3.5 ppm were observed, which showed no cross-peaks, in-
1
1
CD2Cl2: –27.2 ppm) and a broad signal with two maxima dicative of hydroxo groups, in H–13C HMBC or H–13C
at the borders at δ = –11.1 ppm for 2 (cf. free PPh3: –6.0 HSQC spectra. Rather, in a H–19F HOESY spectrum a
1
Eur. J. Inorg. Chem. 2013, 3054–3062
3056
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 2. Experimental 31P and 1H NMR shifts of [HPR3]+ mea-
sured in PhF at room temperature.
The Al–P distances range from 244.1(1) in 1 to 244.7(2)
pm in 2 and are significantly shorter than those in other
R3Al–phosphane adducts owing to the strong electron-
withdrawing effect of the fluorinated alkoxy ligands at the
aluminum atom [R = Me 253.29(6) pm, R = Et 254.13(4)
pm,[46] R = PhSe, 2*Me 251.7(1) pm].[47] The shortening of
the Al–P bond length by 0.6 pm in 1 compared to that in 2
can be explained by the additional inductive effect of the
methyl group and the reduced Tolman cone angle (136 vs.
145°).[34,35] The Al–As distance in 3 of 252.83(4) pm is in
the same range as that in (I/Cl)3Al–AsPh3 (251.4,
251.9 pm).[48] Consistent with the increasing ligand size and
the weaker σ-donor ability of the ligand, the average Al–O
distances decrease from 1 to 2 to 3 (171.8, 171.4, 170.7 pm).
However, according to the higher donor strength of the
phosphane and arsane ligands, the average Al–O distance
increases by about 1.2 to 3.6 pm, if compared to the weaker
donor adducts Do–Al(ORF)3 with Do = PhF, 1,2-F2C6H4,
or SO2.[16]
Cation
31P δ [ppm],
1H δ [ppm],
1JH,P [Hz]
1JP,H [Hz]
{HPCy2[2,4,6-(iPr)3C6H2]}+ 14.4 (m), 443
5.95 (m), 443
7.41 (m), 497
4.31 (m), 426
4.38 (m), 440
6.62 (m), 495
[HPPh3]+
7.95 (m), 497
61.2 (m), 426
34.9 (m), 440
2.3 (m), 495
[HP(tBu)3]+
[HPCy3]+
[HPMePh2]+
1
cross-peak of the H signal to a 19F resonance in the range
typical for perfluoro tert-butoxy groups was observed. With
1
this H–19F HOESY spectrum, we can prove the existence
of a [HO(Al(ORF)3)n]– anion, but the exact number of n
can be 1 or 2. The calculated 1H chemical shifts of
[HOAl(ORF)3]– and [HO(Al(ORF)3)2]– are 0.89 (0.84) and
3.45 (3.65) ppm [(RI-)BP86/def-TZVP ((RI-)BP86/def2-
TZVPP)]; the reference system is H2O in PhF with δ(1H) =
1.13 ppm. In agreement with the thermodynamic calcula-
tions, which favor the formation of [HO(Al(ORF)3)2]–, the
signal is better assigned to the hydroxo-bridged anion. Nev-
ertheless, further investigations dealing with this type of an
anion need to be performed. However, according to our
experience, almost quantitative conversion was observed
with the more sterically hindered phosphanes [e.g., PPh3
and (tBu)3P were almost quantitatively protonated] and
when the stoichiometry was strictly controlled.
Protonated Phosphonium Salts
The molecular structures of 7 and 9 consist of proton-
ated phosphonium ions stabilized by aluminates with ter-
minal (7) or μ-bridged (9) hydroxo groups (Figures 4 and
5). The terminal Al–O4 bond in 7 of 170.9(2) pm is shorter
than that in [{LAl(OH)}2(μ-O)]·0.5(C7H8)·C6H14 reported
by Roesky [173.8(3) pm[3]] and the average Al–O distances
to the alkoxy ligands. The central Al atom is tetrahedrally
coordinated with an average O–Al–O angle of 109.4°. The
shortest anion–cation contact is a C–H···F interaction of
248.8 pm. Owing to the presence of a hydrogen bridge, the
P···O4 distance of 313.8 pm is considerably smaller than the
sum of the van der Waals radii (152 + 180 pm[49]). However,
Crystal Structure Determinations
The molecular structures of 1–3, 7, 9, and 10 were char-
acterized by single-crystal structure analysis.
Donor–Acceptor Adducts
The classical phosphane/arsane–alane adducts 1, 2, and the hydrogen atom attached to O4 was included in the re-
3 are isostructural; the P or As atoms of the ligands are finement only at a calculated position. As the proton bound
bound to the Al atom (Figure 3), and the ORF and phos- to the phosphorus atom was found in the difference Fourier
phane/arsane substituents are arranged in a staggered fash- map and charge equalization is required, this procedure is
ion.
justified.
Figure 3. Molecular structures of 1 (left) and 2 (right). Thermal ellipsoids are drawn at the 50% probability level. Selected distances [pm]:
1: P–Al 244.1(1), av. Al–O 171.8, av. P–CMe/Ph 182.1. ΣO–Al–O 342.7°; 2: P–Al 244.7(2), av. Al–O 171.4, ΣO–Al–O 343.4°. Adduct 3 is
isostructural with 2 and has an As–Al distance of 252.83(4), av. Al–O 170.7, av. As–CPh 192.8 pm, ΣO–Al–O 342.3°.
Eur. J. Inorg. Chem. 2013, 3054–3062
3057
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
ing Al(ORF)2 and O–H groups and is shown in Figure 6.
Compound 10 is isostructural to the fluorine analog, which
is known from the decomposition of Al(ORF)3.[16] At two
of the three hydroxo groups, H bridging to the fluorine
atoms of two PhF molecules can be observed and further
supports our assignment. These interactions with O···F sep-
arations of 274.0 and 284.2 pm can be classified as rather
strong hydrogen bonds (see Figure 6) and even coordinate
the weak donor PhF. As expected, these distances are elon-
gated in comparison with the intramolecular hydrogen
bonds in [(B6F5)2BOH]3 (O–H to ortho fluorine of C6F5
ligand, 267 pm[42]), but they are in the same range as those
in MnF3·3H2O (262–276 pm) or the O···N distance in
[(tBu)2Al(μ-OH)]3·2MeCN.[52,53] Several [R2Al(μ-OH)]3
compounds have previously been described. Most of them
exhibit a “twist-boat conformation”.[53] Also the closely re-
lated [(C6F5)2BOH]3 shows a “C2 twist-boat conforma-
tion”.[42] The average Al–O–Al angle in 10 of 138.6° is sig-
nificantly reduced in comparison to those in [(tBu)2Al(μ-
OH)]3 (without solvent 142.0°, with coordinated MeCN
141.7°, with THF 140.5°)[53] or [(FRO)2Al(μ-F)]3
(146.4°).[16] The average cyclic Al–O bond length in 10
(179.4 pm) lies between those in [(tBu)2Al(μ-OH)]3
(184.9 pm)[53] and [(MCIMP)Al(μ-OH···THF)]3 (177.7 pm,
“almost coplanar”)[4] and is in the same range as those of
the tetrameric aluminopolysiloxanes reported by Veith
[R2Al(μ-OH)]4 (with pyridine av. Al–O 176.9 pm,[54] with
Et2O av. 180.0 pm[55]). The difference of the cyclic Al–O
bond lengths between coordinated and non-coordinated hy-
droxo groups is interesting. In all three cases, only two sol-
vent molecules are coordinated and the difference between
the non-coordinated and coordinated Al–O bond lengths
varies significantly {[(tBu)2Al(μ-OH)]3·2L: L = THF, Δ =
+1.1 pm; L = MeCN, Δ = +1.8 pm; 10: L = PhF, Δ =
–1.2 pm}. A possible explanation is that the coordinated
solvent molecule in 10 is almost parallel to the [Al(μ-
Figure 4. Molecular structure of 7. Thermal ellipsoids are drawn
at the 50% probability level. Selected distances [pm]: P–H1 128(3),
Al–O4 170.9(2), av. Al–O 173.4, av. P–CCy 182.1, P···O4 313.8,
(P)H1···O4 265.9; av. O–Al–O 109.4°.
Figure 5. Molecular structure of 9. Thermal ellipsoids are drawn
at the 50% probability level. Selected distances [pm]: Al1–O7
181.9(4), Al2–O7 182.0(3), P–H1 152.7, av. Al–O 171.2. The Al1–
O7–Al2 angle is 148.4°.
As expected, the bridged Al–O distances in 9 are elon-
gated by around 10 pm compared to the Al–O bond lengths
in 7 or those to the terminal OC(CF3)3 groups, as the O
atom is shared between two Al(ORF)3 moieties. The posi-
tion of the hydrogen atom on the hydroxy group was pri-
marily found on the difference Fourier map and was finally
restrained by using a DFIX instruction (0.9 Å).[50] Similar
to the Cl analog,[51] but in contrast to the F analog, the
anion exhibits a bend with an Al1–O7–Al2 angle of 148.4°,
which is almost 9° larger than that in [(C6F5)3B(μ-OH)-
B(C6F5)3]– [139.6(5)°].[15] The shortest anion–cation dis-
tance is a C–H···F distance of 249.3 pm. Three intramolecu-
lar O–H···F contacts within the anion between 235.0 and
260.6 pm were also observed.
Figure 6. Molecular structure of 10 with two hydrogen-bonded PhF
molecules. Thermal ellipsoids are drawn at the 50% probability
level. Disorder in the C(CF3)3 groups is removed for clarity. Se-
lected distances [pm]: Al1–O10 179.7(2), Al1–O11 180.0(2), Al2–
O11 178.9(2), Al2–O12 178.4(3), Al3–O10 180.4(2), Al3–O12
178.7(3), O11···F56 274.0, O10···F55 284.2, O12···F41 292.4. Se-
lected angles [°]: Al1–O10–Al3 137.0(1), Al2–O11–Al1 137.7(1),
Hydrolysis Product (10)
The trimeric hydrolysis product [HOAl(ORF)2]3 (10) con-
sists of an almost planar six-membered ring with alternat- Al2–O12–Al3 141.2(2).
Eur. J. Inorg. Chem. 2013, 3054–3062
3058
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
OH)]3 plane (29.0 and 22.6° between the [Al–O]3 plane and nor computationally investigated. Using our recently estab-
the PhF planes), whereas in [(tBu)2Al(μ-OH)]3·2L, L is lished rCCC model,[58] we calculated ΔsolvG°(H+,PhF), as
more vertically positioned.[53]
–887 kJmol–1 (details in the Supporting Information). This
is in line with expectations. As a bulk solvent, according
to our calculations, fluorobenzene is more basic than 1,2-
dichloroethane (–860 kJmol–1) and is slightly less basic
than sulfur dioxide (–898 kJmol–1).[58] The comparison
with the aqueous system is more spectacular. According to
the experimental ΔsolvG°(H+,H2O) value of –1105 kJmol–1
(rCCC model: –1107 kJmol–1), as a bulk solvent water is
more basic by 218 kJmol–1 or 38 pH units, or in other
words: a fluorobenzene solution of pH 38 is about as acidic
as one molar aqueous hydrochloric acid of aqueous pH 0!
By using this anchor point, we calculated the pKa value
of 4 in fluorobenzene solution by constructing the Born–
Fajans–Haber cycle (BFHC) shown in Scheme 1. The stan-
dard Gibbs solvation energies (COSMO, εr = 5.42) of the
water adduct (4) and its conjugate base were calculated at
the BP86/def-TZVP level of theory.
Brønsted Acidity Considerations
Gas Phase Acidity of H2O–Al(ORF)3
Thermodynamically, the Brønsted acidity of molecules is
best expressed by the gas phase acidity (GA = standard
Gibbs energy of deprotonation in the gas phase). The lower
the GA, the more acidic the molecule. According to our
calculated GA value of 1148 kJmol–1 at the BP86/def-
TZVP level of theory, H2O–Al(ORF)3 (4) is more acidic
than, for example, HSO3F (GA = 1233 kJmol–1), but less
acidic than our recently published perfluorinated alcohol
adduct H[Al(ORF)4] (GA = 1041 kJmol–1), both calculated
at the same level.[56]
Absolute Acidity of the H2O–Al(ORF)3/PhF System
In agreement with the increased acidity of the coordi-
nated H2O, experimentally observed by the downfield shift
of the protons, we were interested to evaluate the absolute
acidity of H2O–Al(ORF)3 (4) in regard to the absolute pH
scale introduced in 2010.[57,58] In this medium-independent
acidity scale, the absolute chemical potential of the proton
μabs(H+), with the gaseous proton as a reference, is used
as a universal measure of Brønsted acidity. In a solvent S,
μabs(H+) depends on the standard Gibbs solvation energy
of the proton ΔsolvG°(H+,S) and the pH of the solution
through Equation (4).
Scheme 1. Born–Fajans–Haber cycle to calculate the pKa of 4 in
PhF.
According to our calculated pKa of 31.7, a solution of 4
in pure PhF is only very scarcely dissociated. Using basic
chemistry knowledge, the pH of a 1 mm solution of 4 in
PhF can then be calculated according to Equation (6).
pH(PhF) = ½[31.7 – log(0.001)] = 17.4
(6)
The absolute acidity is then obtained by inserting this value
into Equation (4) to give Equation (7):
(4)
μabs(H+) = –887 kJmol–1 – 5.71 kJmol–1 ϫ17.4 = –986 kJmol–1 or
pHabs = 172.6
(7)
This can be made more comprehendible by the absolute
pHabs, defined as in Equation (5).
Thus, according to Equation (7), a millimolar solution of
4 in fluorobenzene has a pHabs of 172.6, which makes it
only slightly less acidic than pure H2SO4 with a pHabs of
171 (see Supporting Information). An excess of the Lewis
acid Al(ORF)3, which removes [HO–Al(ORF)3]– from the
protolysis equilibrium by complexation, should clearly lead
to a superacidic solution [cf. to Equation (2) above: ΔrH for
the complexation is –144 kJmol–1].
(5)
In other words, with the knowledge of ΔsolvG°(H+,S) as
anchor point, one can convert pH(S) in an individual me-
dium into absolute acidity values pHabs, which can be com-
pared over medium boundaries. Single ion Gibbs solvation
energies are very difficult to obtain experimentally, but can
Conclusions
We have described our preliminary attempts to use the
be computed quantum chemically with an error of 15– superacid Al(ORF)3 for FLP-like chemistry. Dihydrogen ac-
20 kJmol–1 (≈ 3 pH units).
tivations with various phosphane/Al(ORF)3 systems were
For fluorobenzene, the standard solvent for this chemis- not successful. Instead, the activation of the omnipresent,
try, the anchor point was hitherto neither experimentally small molecule water was observed even at a low concentra-
Eur. J. Inorg. Chem. 2013, 3054–3062
3059
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tion of 9 ppm. Therefore, H2O–Al(ORF)3 (4) was synthe-
sized by substitution of PhF in PhF–Al(ORF)3 with H2O
and was analyzed in a long-term NMR study. Over 22 days,
crystallographic 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.
Crystal Structure Data for MePh2P–Al(ORF)3 (1): a = 10.895(3), b
= 28.144(7), c = 13.905(3) Å; α = 90, β = 128.759(14), γ = 90°; P21/
c, Z = 4, wR2 = 0.1028, R1 = 0.0382. Ph3P–Al(ORF)3 (2): a =
1
the H and 19F NMR signals of alcohol HOC(CF3)3 in-
creased as was expected for hydrolysis to give HOC(CF3)3
and HOAl(ORF)2. In analogy to the B(C6F5)3 system, a
crystalline hydrolysis product 10 was formed, which proba-
bly resulted from trimerization of the initially formed
HOAl(ORF)2. Trimer 10 contains a six-membered ring with
bridging hydroxo groups.
¯
17.691, b = 17.691, c = 19.840 Å; α = 90, β = 90, γ = 120°; P1, Z
= 6, wR2 = 0.1365, R1 = 0.0476. Ph3As–Al(ORF)3 (3): a =
10.7515(2), b = 17.2947(4), c = 21.1997(5) Å; α = 90, β =
112.4620(10), γ = 90°; P21/c, Z = 4, wR2 = 0.0690, R1 = 0.0290.
[HPCy3][HOAl(ORF)3] (7): a = 12.4547(2), b = 17.3208(3), c =
20.0236(3) Å; α = 90, β = 114.5130(10), γ = 90°; P21/c, Z = 4, wR2
The absolute acidity of 4 was evaluated, according to the
recently introduced absolute pH scale.[57,58] The absolute
chemical potential of a 0.001 m solution of 4 in PhF was
established as –986 kJmol–1 and it has a pHabs value of 173.
This is nearly as acidic as pure H2SO4 (pHabs = 171). The
resulting increase of the Brønsted acidity of H2O by com-
plexation with the Lewis superacid Al(ORF)3 was used to
protonate phosphanes. Herein, PMePh2, PPh3, PCy3,
P(tBu)3, and PCy2(2,4,6-(iPr)3C6H2) were protonated and
crystallized with the two new weakly coordinating anions
[(FRO)3Al(μ-OH)Al(ORF)3]– and [HOAl(ORF)3]–. How-
ever, formation of classical, neutral donor–acceptor adducts
between the less encumbered Ph2MeP, Ph3P, and Ph3As and
the superacid Al(ORF)3 was also observed. The dissolution
of 3 in CD2Cl2 resulted in the formation of a FLP system,
which afforded the activation of CD2Cl2 to give the halide
abstraction product [ClAl(ORF)3]– and alkylated arsane
[CD2ClAsPh3]+. Even if dihydrogen activation has not been
achieved, the interesting behavior between water, phos-
phanes, and Al(ORF)3 allows for a better understanding of
the undesirable side-reactions of the superacid Al(ORF)3.
=
0.1374, R1
=
0.0517. [HO(Al(ORF)3)2][HPCy2(2,4,6-
(iPr)3C6H2)] (9): a = 12.9204(6), b = 18.1265(9), c = 29.660(2) Å; α
= 90, β = 93.603(7), γ = 90°; P21/c, Z = 4, wR2 = 0.1887, R1
=
0.0662. [HOAl(ORF)2]3·2PhF (10·2PhF) a = 12.7572(8), b =
13.7652(10), c = 16.7687(12) Å; α = 83.638(6), β = 79.132(6), γ =
76.968(5)°; Pν, Z = 2, wR = 0.1702, R = 0.0588.
˜
2
1
Attenuated total reflectance (ATR) IR spectra were recorded with
a Nicolet Magna-IR 760 FTIR spectrometer with the software
OMNIC. An advanced ATR correction was perfomed using the
standard values. Raman measurements were performed with a
Bruker RAM II Fourier transform Raman module for a VERTEX
70 spectrometer with a Nd-YAG laser and the OPUS software. The
excitation wavelength was 1064 nm.
General Procedure for Adduct Synthesis: In a special J. Young flask
PhF–Al(OC(CF3)3)3 dissolved in PhF or perfluorohexane and
phosphane or arsane (1 equiv.) dissolved in PhF or perfluorohex-
ane were mixed at –20 °C. The reaction mixture was stirred for up
to two hours and crystallized at –20 or –40 °C. After decantation
of the liquid and removal of the residual solvent from the crystal-
line mass in vacuo, the compounds were isolated as colorless pow-
ders in good to almost quantitative yields.
Ph2MeP–Al(OC(CF3)3)3 (1): Ph2MeP (0.033 g, 0.16 mmol) in per-
fluorohexane, PhF–Al(OC(CF3)3)3 (0.136 g, 0.16 mmol, 1 equiv.).
A colorless precipitate appeared immediately, yield 0.07 g, 47%.
1H NMR (400.17 MHz, CD2Cl2, 298 K): δ = 7.59 (m, 2 H, p-
C6H5), 7.50 (m, 8 H, o/m-C6H5), 1.99 (d, J = 9.3 Hz, 3 H, CH3)
ppm. 19F NMR (376.54 MHz, CD2Cl2, 298 K): δ = –75.4 {s, 27 F,
Ph2MeP–Al(OC(CF3)3)3} ppm. 27Al NMR (104.27 MHz, CD2Cl2,
298 K): δ = 43.9 {br. s, 1 Al, Ph2MeP–Al(OC(CF3)3)3} ppm. 31P
NMR (161.99 MHz, CD2Cl2, 298 K): δ = –26.6 {m, 1 P, Ph2MeP–
Experimental Section
General: All reactions were performed by using standard Schlenk
and vacuum techniques. Chemicals were handled under an argon
atmosphere or in a glove box. All phosphanes and AsPh3 were used
as provided. PhF, CD2Cl2, and perfluorohexane were dried with
CaH2 and distilled afterwards or were directly condensed into the
reaction mixtures. PhF–Al(ORF)3 was synthesized as previously de-
scribed.[17] NMR spectra were measured with special J. Young or
flame-sealed NMR tubes with a Bruker Avance II+ 400 MHz WB
or Bruker DPX 200 spectrometer. Spectra of CD2Cl2 solutions
were calibrated to the 1H signal of CHDCl2 at δ = 5.32 ppm. Spec-
tra of PhF solutions (without deuterium lock) were calibrated to
Al(OC(CF ) ) } ppm. IR (Diamond ATR): ν = 445 (w), 538 (w),
˜
3 3 3
573 (w), 692 (w), 727 (m), 739 (w), 854 (w), 894 (w), 955 (w), 974
(s), 1108 (w), 1182 (m), 1218 (s), 1249 (vs), 1266 (s), 1300 (m), 1353
(m), 1442 (w), 3070 (vw) cm–1. FT Raman: ν = 228 (w), 270 (w),
˜
327 (w), 350 (w), 539 (w), 564 (w), 618 (w), 676 (w), 691 (w), 751
(w), 811 (w), 1002 (s), 1029 (w), 1108 (w), 1164 (vw), 1192 (vw),
1275 (vw), 1425 (vw), 1593 (m), 2756 (w), 2907 (w), 2938 (w), 2971
(vw), 3055 (m), 3072 (m) cm–1.
the 19F signal at δ = –113.1 ppm with respect to CFCl3 and the H
1
signal of the o-H of PhF at δ = 7.14 ppm. Spectra in 1,2-F2C6H4
solutions were calibrated to the 1H signal of the o-H of the solvent
at δ = 6.96 ppm and the 19F signal at –139.43 ppm (tetramethylsil-
ane, TMS). The field corrections of the 31P and 27Al spectra were
adjusted to the proton calibration. The Topspin 2.1 software was
used.
Ph3P–Al(OC(CF3)3)3 (2): PhF–Al(OC(CF3)3)3 (0.40 g, 0.48 mmol)
in PhF (3 mL), PPh3 (0.127 g, 0.48 mmol, 1 equiv.) in PhF (2 mL),
yield 0.451 g, 94%. 1H NMR (400.17 MHz, CD2Cl2, 298 K): δ =
7.62 (m, 3 H, p-Hs), 7.50 (m, 6 H, m-Hs), 7.39 (m, 6 H, o-Hs) ppm.
19F NMR (376.53 MHz, CD2Cl2, 298 K): δ = –74.3 {s, 27 F, Ph3P–
Al(OC(CF3)3)3}. 27Al NMR (104.27 MHz, CD2Cl2, 298 K): δ =
41.4 {br. s, Ph3P–Al(OC(CF3)3)3} ppm. 31P NMR (161.99 MHz,
CD2Cl2, 298 K): δ = –11.1 {m, distance between the two maxima
Crystal structure measurements were performed with a Bruker
Quazar APEX2 CCD area detector or a Rigaku R-Axis SPIDER
diffractometer with Mo-Kα X-ray sources. SHELX,[50] SIR2004,[59]
and OLEX2[60] were used to solve and refine the structures. PLA-
TON[61] was used for twin analysis or to change the space group.
Graphics of the crystal structures were drawn with ORTEP.[62]
1060 Hz, Ph P–Al(OC(CF ) ) } ppm. IR (Diamond ATR): ν = 434
˜
3
3 3 3
(w), 497 (w), 537 (w), 565 (vw), 693 (m), 713 (w), 727 (m), 747 (s),
808 (w), 849 (w), 897 (w), 973 (s), 1001 (w), 1027 (w), 1101 (m),
1179 (s), 1215 (vs), 1246 (vs), 1264 (s), 1300 (m), 1350 (w), 1438
CCDC-893508 (for 1), -893671 (for 2), -893442 (for 3), -893723 (for
7), -893628 (for 9), and -914824 (for 10) contain the supplementary
Eur. J. Inorg. Chem. 2013, 3054–3062
3060
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
(w), 1485 (vw), 1507 (vw), 1540 (vw), 1559 (vw), 3066 (vw) cm–1.
[4] H.-L. Chen, B.-T. Ko, B.-H. Huang, C.-C. Lin, Organometallics
2001, 20, 5076–5083.
FT Raman: ν = 224 (w), 251 (w), 269 (w), 326 (w), 366 (w), 539
˜
[5] A. Y. Timoshkin, G. Frenking, Organometallics 2008, 27, 371–
(w), 564 (w), 618 (vw), 691 (vw), 714 (vw), 752 (vw), 809 (w), 1003
(vs), 1029 (m), 1102 (m), 1162 (w), 1187 (w), 1247 (w), 1578 (m),
1590 (vs), 3070 (vs) cm–1.
380.
[6] J. Storre, A. Klemp, H. W. Roesky, H.-G. Schmidt, M. Nolte-
meyer, R. Fleischer, D. Stalke, J. Am. Chem. Soc. 1996, 118,
1380–1386.
Ph3As–Al(OC(CF3)3)3 (3): AsPh3 (0.037 g, 0.12 mmol), PhF–Al-
[7] D. Chakraborty, E. Y.-X. Chen, Organometallics 2003, 22, 207–
(OC(CF3)3)3 (0.10 g, 0.12 mmol, 1 equiv.) in PhF (3 mL). Almost
1
210.
quantitative conversion presumed according to NMR analysis. H
[8] A. W. Apblett, A. C. Warren, A. R. Barron, Can. J. Chem.
1992, 70, 771–778.
[9] R. A. Stapleton, B. R. Galan, S. Collins, R. S. Simons, J. C.
Garrison, W. J. Youngs, J. Am. Chem. Soc. 2003, 125, 9246–
9247.
[10] A. A. Malkov, I. P. Romm, E. N. Guryanova, Y. G. Noskov,
E. S. Petrov, Polyhedron 1997, 16, 4081–4086.
[11] L. H. Doerrer, M. L. H. Green, J. Chem. Soc., Dalton Trans.
1999, 4325–4329.
[12] C. Bergquist, B. M. Bridgewater, C. J. Harlan, J. R. Norton,
R. A. Friesner, G. Parkin, J. Am. Chem. Soc. 2000, 122, 10581–
10590.
[13] A. H. Cowley, C. L. B. Macdonald, J. S. Silverman, J. D.
Gorden, A. Voigt, Chem. Commun. 2001, 175–176.
[14] A. Di Saverio, F. Focante, I. Camurati, L. Resconi, T. Be-
ringhelli, G. D’Alfonso, D. Donghi, D. Maggioni, P. Mercand-
elli, A. Sironi, Inorg. Chem. 2005, 44, 5030–5041.
[15] A. A. Danopoulos, J. R. Galsworthy, M. L. H. Green, L. H.
Doerrer, S. Cafferkey, M. B. Hursthouse, Chem. Commun.
1998, 2529–2560.
[16] A. Kraft, N. Trapp, D. Himmel, H. Böhrer, P. Schlüter, H.
Scherer, I. Krossing, Chem. Eur. J. 2012, 18, 9371–9380.
[17] L. O. Müller, D. Himmel, J. Stauffer, G. Steinfeld, J. Slattery,
G. Santiso-Quiñones, V. Brecht, I. Krossing, Angew. Chem.
2008, 120, 7772; Angew. Chem. Int. Ed. 2008, 47, 7659–7663.
[18] D. W. Stephan, G. Erker, Angew. Chem. 2010, 122, 50; Angew.
Chem. Int. Ed. 2010, 49, 46–76.
NMR (400.17 MHz, 1,2-F2C6H4, 298 K): δ = 7.35–7.50 (m, C6H5)
ppm. 19F NMR (376.54 MHz, 1,2-F2C6H4, 298 K): δ = –75.4 {s,
side-product, probably PhF– or 1,2-F2C6H4–Al(OC(CF3)3)3},
–75.2 {s, 27 F, Ph3As–Al(OC(CF3)3)3} ppm. 27Al NMR
(104.27 MHz, 1,2-F2C6H4, 298 K): δ = 41.9 {br. s, 1 Al, Ph3As–
Al(OC(CF3)3)3}.
NMR Tube Reactions: In special J. Young NMR tubes, 1:1 or 1:2
mixtures of phosphane/PhF–Al(OC(CF3)3)3 were dissolved in mo-
ist PhF (173 ppm water). NMR spectra were measured directly, and
1
the data are included in Table 2 (31P and H) and the Supporting
Information.
PCy2(2,4,6-(iPr)3C6H2) 1:2: PCy2(2,4,6-(iPr)3C6H2) (0.024 g,
0.06 mmol) and PhF–Al(OC(CF3)3)3 (0.099 g, 0.12 mmol, 2 equiv.)
in PhF (1 mL, 173 ppm H2O).
P(tBu)3 1:2: P(tBu)3 (0.012 g, 0.06) and PhF–Al(OC(CF3)3)3
(0.099 g, 0.12 mmol, 2 equiv.) in PhF (1 mL, 173 ppm H2O).
PCy3 1:2: PCy3 (0.017 g, 0.06 mmol) and PhF–Al(OC(CF3)3)3
(0.099 g, 0.12 mmol, 2 equiv.) in PhF (1 mL, 173 ppm H2O).
PCy3 1:1 (1bar H2): PCy3 (0.021 g, 0.07 mmol) and PhF–Al-
(OC(CF3)3)3 (0.058 g, 0.07 mmol, 1 equiv.) in PhF (1 mL).
PPh3 1:2: PPh3 (0.01 g, 0.038 mmol) and PhF–Al(OC(CF3)3)3
(0.063 g, 0.076 mmol, 2 equiv.) in PhF (0.8 mL, 173 ppm H2O).
[19] G. Ménard, D. W. Stephan, Angew. Chem. 2011, 123, 8546; An-
gew. Chem. Int. Ed. 2011, 50, 8396–8399.
[20] G. Ménard, D. W. Stephan, Angew. Chem. 2012, 124, 8397; An-
gew. Chem. Int. Ed. 2012, 51, 8272–8275.
PMePh2 1:2: PMePh2 (0.02 g, 0.10 mmol) and PhF–Al(OC-
(CF3)3)3 (0.16 g, 0.20 mmol, 2 equiv.) in PhF (0.8 mL, 173 ppm
H2O).
[21] G. Ménard, D. W. Stephan, Angew. Chem. 2012, 124, 4485; An-
gew. Chem. Int. Ed. 2012, 51, 4409–4412.
H2O–Al(OC(CF3)3)3 (4): In a special J. Young NMR tube with a
Teflon® valve, PhF–Al(OC(CF3)3)3 (0.020 g, 0.024 mmol) was dis-
solved in PhF (0.6 mL, 173 ppm H2O). A series of NMR spectra
were recorded at various times (Figure 1 and Supporting Infor-
mation). 1H NMR (400.17 MHz, PhF, 298 K): δ = 6.03 {s, 2 H,
H2O–Al(OC(CF3)3)3}, 3.57 [s, 1 H, HOC(CF3)3] ppm. 19F NMR
(376.54 MHz, PhF, 298 K): δ = –74.53 [s, 9F, HOC(CF3)3], –75.53
{s, 27F, PhF–Al(OC(CF3)3)3}, –75.57 {27 F, H2O–Al(OC(CF3)3)3}
ppm. 27Al NMR (104.27 MHz, PhF, 298 K): no signal.
[22] D. Wu, D. Jia, A. Liu, L. Liu, J. Guo, Chem. Phys. Lett. 2012,
541, 1–6.
[23] G. C. Welch, D. W. Stephan, J. Am. Chem. Soc. 2007, 129,
1880–1881.
[24] S. Grimme, H. Kruse, L. Goerigk, G. Erker, Angew. Chem.
2010, 122, 1444; Angew. Chem. Int. Ed. 2010, 49, 1402–1405.
[25] Quantum chemical calculations presume a possible activation
of hydrogen, which has so far not been realized experimentally,
thus further systematic investigations need to be conducted in
our laboratories.
[26] K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs,
Chem. Phys. Lett. 1995, 242, 652–660.
[27] K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs,
Chem. Phys. Lett. 1995, 240, 283.
Supporting Information (see footnote on the first page of this arti-
cle): Details of the quantum chemical calculations, NMR dis-
cussions, and crystallographic data.
[28] K. Eichkorn, F. Weigend, O. Treutler, R. Ahlrichs, Theor.
Chem. Acc. 1997, 97, 119–124.
Acknowledgments
[29]
[30]
[31]
A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100.
J. P. Perdew, Phys. Rev. B 1986, 33, 8822–8824.
A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100,
5829–5835.
F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7,
3297–3305.
F. Weigend, M. Häser, H. Patzelt, R. Ahlrichs, Chem. Phys.
Lett. 1998, 294, 143–152.
C. A. Tolman, Chem. Rev. 1977, 77, 313–348.
C. E. Housecroft, A. G. Sharpe, Inorganic chemistry, 2nd ed.,
Pearson Education Limited, 2005.
The authors are grateful for financial support by the Fonds der
Chemischen Industrie, the Albert-Ludwigs-Universität Freiburg,
the Deutsche Forschungsgemeinschaft (DFG) and the European
Research Council through the UniChem project.
[32]
[33]
[1] E. Y.-X. Chen, T. J. Marks, Chem. Rev. 2000, 100, 1391–1434.
[2] H. W. Roesky, M. G. Walawalkar, R. Murugavel, Acc. Chem.
Res. 2001, 34, 201–211.
[3] G. Bai, H. W. Roesky, J. Li, M. Noltemeyer, H.-G. Schmidt,
Angew. Chem. 2003, 115, 5660; Angew. Chem. Int. Ed. 2003,
42, 5502–5506.
[34]
[35]
[36]
S. Grimme, J. Comput. Chem. 2006, 27, 1787–1799.
Eur. J. Inorg. Chem. 2013, 3054–3062
3061
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[37]
[38]
[39]
[40]
[41]
[48] E. Conrad, J. Pickup, N. Burford, R. McDonald, M. J. Fergu-
son, Can. J. Chem. 2010, 88, 797–803.
[49] A. Bondi, J. Phys. Chem. 1964, 68, 441–451.
[50] G. M. Sheldrick, Acta Crystallogr., Sect. A 2008, 64, 112–122.
[51] A. Kraft, J. Beck, I. Krossing, Chem. Eur. J. 2011, 17, 12975–
12980.
[52] M. Molinier, W. Massa, J. Fluorine Chem. 1992, 57, 139–146.
[53] M. R. Mason, J. M. Smith, S. G. Bott, A. R. Barron, J. Am.
Chem. Soc. 1993, 115, 4971–4984.
[54] M. Veith, M. Jarczyk, V. Huch, Angew. Chem. 1998, 110, 109;
Angew. Chem. Int. Ed. 1998, 37, 105–108.
[55] M. Veith, M. Jarczyk, V. Huch, Angew. Chem. 1997, 109, 140;
Angew. Chem. Int. Ed. Engl. 1997, 36, 117–119.
[56] A. Kraft, J. Beck, G. Steinfeld, H. Scherer, D. Himmel, I.
Krossing, Organometallics 2012, 31, 7485–7491.
[57] D. Himmel, S. K. Goll, I. Leito, I. Krossing, Angew. Chem.
2010, 122, 7037; Angew. Chem. Int. Ed. 2010, 49, 6885–6888.
[58] D. Himmel, S. K. Goll, I. Leito, I. Krossing, Chem. Eur. J.
2011, 17, 5808–5826.
[59] M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L.
Cascarano, L. D. Caro, C. Giacovazzo, G. Polidori, R. Spagna,
J. Appl. Crystallogr. 2005, 38, 381–388.
[60] O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard,
H. Puschmann, J. Appl. Crystallogr. 2009, 42, 339–341.
[61] A. L. Spek, Acta Crystallogr., Sect. D, Bio. Cryst. 2009, 65,
148–155.
S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys.
2010, 132, 154104–154119.
A. Klamt, Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2011, 1,
699–709.
A. Klamt, G. Schüürmann, J. Chem. Soc. Perkin Trans. 2 1993,
799–805.
D. R. Lide, CRC Handbook of Chemistry and Physics, 87th ed.,
Taylor and Francis, Boca Raton, 2007.
To the best of our understanding, the calculated values taken
from the literature are based on the self-consistent field (scf)
energies without zero-point energy (ZPE). To be in agreement
with these values and to prove the postulated tendency, we cal-
culated the complexation energy of 4 to be –101 kJmol–1 and
that of the bridged anion to –152 kJmol–1 with the same
method. However, the tendencies remain unchanged with or
without ZPE included.
[42] T. Beringhelli, G. D’Alfonso, D. Donghi, D. Maggioni, P. Mer-
candelli, A. Sironi, Organometallics 2003, 22, 1588–1590.
[43] F. Dornhaus, M. Bolte, H.-W. Lerner, M. Wagner, Eur. J. Inorg.
Chem. 2006, 1777–1785.
[44] S. O. Grim, A. W. Yankowsky, J. Org. Chem. 1977, 42, 1236–
1239.
[45] A. R. Barron, J. Chem. Soc., Dalton Trans. 1988, 3047–3050.
[46] V. V. Shatunov, A. A. Korlyukov, A. V. Lebedev, V. D. Shelud-
yakov, B. I. Kozyrkin, V. Y. Orlov, J. Organomet. Chem. 2011,
696, 2238–2251.
[47] R. Kumar, D. G. Dick, S. U. Ghazi, M. Taghiof, M. J. Heeg,
J. P. Oliver, Organometallics 1995, 14, 1601–1607.
[62] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565.
Received: January 10, 2013
Published Online: May 10, 2013
Eur. J. Inorg. Chem. 2013, 3054–3062
3062
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim