close

Вход

Забыли?

вход по аккаунту

?

Noncoordinating AnionsЧFact or Fiction A Survey of Likely Candidates.

код для вставкиСкачать
Reviews
I. Krossing and I. Raabe
Weakly Coordinating Anions
Noncoordinating Anions—Fact or Fiction?
A Survey of Likely Candidates
Ingo Krossing* and Ines Raabe
Keywords:
alkoxymetallates · aluminates · anions ·
carboranes · Lewis acids
Angewandte
Chemie
2066
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200300620
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
Is there anything resembling a truly noncoordinating anion?
Would it not be great to be able to prepare any crazy, beautiful,
or simply useful cationic species that one has in mind, or has
detected by mass spectroscopy? In condensed phases the target
cation has to be partnered with a suitable counteranion. This is
the moment when difficulties arise and many wonderful ideas
end in the sink owing to coordination or decomposition of the
anion. However, maybe these counteranion problems can be
overcome by one of the new weakly coordinating anions
(WCAs). Herein is an overview on the available candidates in
the quest for the least coordinating anion and a summary of
new applications, available starting materials, and general
strategies to introduce a WCA into a system. Some of the
unusual properties of WCA salts such as high solubility in low
dielectric media, pseudo gas-phase conditions in condensed
phases, and the stabilization of weakly bound and low-charged
complexes are rationalized on thermodynamic grounds. Limits
of the WCAs, that is, anion coordination and decomposition,
are shown and a quantum chemical analysis of all types of
WCAs is presented which allows the choice of a particular
WCA to be based on quantative data from a wide range of
different anions.
From the Contents
1. Introduction
2067
2. The Candidates
2067
3. Applications of Weakly Coordination
Anions
2072
4. Available Starting Materials and
Strategies to Introduce Weakly
Coordinating Anions
2073
5. A Rationalization of the Special
Properties of Salts of Weakly
Coordinating Ions Based on
Thermodynamic Considerations
2078
6. Limits of Weakly Coordinating Anions
2081
7. Comparison of the Properties of different
Classes of Weakly Coordinating Anions
Based on Quantum Chemical
Calculations
2084
8. Conclusion and Outlook
2086
1. Introduction
2. The Candidates
Only 25 years ago the term “noncoordinating anion” was
commonly used when a coordinating anion, such as a halide
X , was replaced by a complex anion, such as [CF3SO3] ,
[BF4] , [ClO4] , [AlX4] , or [MF6] (X = Cl–I; M = P, As, Sb,
etc.). However, with the advent of routine X-ray crystallography it became evident that in many cases also a “noncoordinating” complex anion can easily be coordinated.[1] To
account for the coordination of these complex anions, for
example in a [M]-F-BF3 fragment, the term “weakly coordinating anion” (WCA) was coined. This expression allows for
weak coordination but already includes the potential of such
complexes to serve as a precursor of the “noncoordinated”
cation, for example, in catalytic processes. Owing to the
importance of such WCAs both in fundamental[2] and
applied[3] chemistry many efforts were undertaken to finally
reach the ultimate goal of a truly noncoordinating anion.
However, noncoordination is impossible[4] but since S. H.
Strauss' widely cited article in Chemical Reviews on WCAs[2b]
a plethora of new species that closely approximate a noncoordinating anion appeared; many of the latter were initially
published in the patent literature demonstrating their impact
on various applications. Some authors tend to call this new
generation of WCAs “superweak anions”.[5] This article
summarizes the development over the last 10 years and
gives a survey on the most likely candidates for the best
approximation of a noncoordinating anion.
The general prescription to produce a more weakly
coordinating anion appears to be the delocalization of the
negative charge over a large area of non-nucleophilic and
chemically robust moieties. Several approaches to achieve
this goal have recently been used and will be detailed in the
following. However, one should still bear in mind that the
coordinating ability of each anion is limited by its most basic
site, that is, just like a chain breaks with its weakest link, a
WCA will always (albeit weakly) coordinate with its most
nucleophilic, sterically accessible moiety, which may be the
starting point for anion decomposition. The art of constructing the ultimate noncoordinating anion is therefore to realize
a structure without an accessible basic site. This may be
achieved by a combination of steric as well as electronic
effects.
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
2.1. Borate-Based and Related Anions
Exchange of the fluorine atoms in [BF4] ions for phenyl
groups leads to the long known larger [BPh4] ion (Figure 1).
[*] Priv.-Doz. Dr. I. Krossing, Dipl.-Chem. I. Raabe
University of Karlsruhe
Engesserstrasse Geb. 30.45
76128 Karlsruhe (Germany)
Fax: (+ 49) 721-608-4854
E-mail: krossing@chemie.uni-karlsruhe.de
DOI: 10.1002/anie.200300620
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2067
Reviews
I. Krossing and I. Raabe
effect of the CF3 group compared to the 2-C3F7
group.[9] Also the -C6F5 ligand was modified in the 4position by a -CF3,[10] -Si(iPr)3,[11] -SiMe2tBu,[11] or
-C6F4{C(F)(C6F5)2}[12] group. Alternatively the -C6F5
group was exchanged for fluorinated biphenyl or
naphthalene moieties.[13] However, although these
modified ions were in some cases less coordinating
and gave more active catalysts, the extra synthetic
effort to produce these special ions appears to exceed
their potential for wide-spread use and only few
subsequent publications report on the application of
these modified ions.[14]
A further modification of the fluorinated tetraarylborates was to exchange the Lewis acidic boron atom
for aluminum or gallium, however, both compounds
hydrolyze and the aluminum compounds tend to be
explosive.[15] Nevertheless, both aluminates and gallates [M(C6F5)4] were shown to stabilize reactive
cationic polymerization catalysts.[16, 17]
Another approach was more effective: Reaction of
Figure 1. Structures of selected borate based WCAs as superpositions of ball-and-stick
two equivalents of B(C6F5)3 or a related Lewis acid
and space-filling models.
with a strong and hard nucleophile X , such as CN ,[18]
C3N2H3 (imidazolyl)[19, 20] or NH2 .[21] The resulting
dimeric [(F5C6)3B(m-X)B(C6F5)3] borates are simple to
However, this ion is prone to hydrolysis and the phenyl
groups tend to be easily cleaved and are relatively strongly
prepare and surprisingly stable. This stability may be seen
coordinating.[4] To overcome these problems, the phenyl
by the reaction of Na+[(F5C6)3B(m-NH2)B(C6F5)3] with HCl
groups were fluorinated and -C6F5 or -C6H3-3,5-(CF3)2 groups
in diethyl ether which produces the [H(OEt2)2]+[(F5C6)3B(mwere attached to the Lewis acidic central boron atom giving
NH2)B(C6F5)3] salt (Scheme 1). No trace of decomposition
the nowadays seemingly ubiquitous class of [B(C6F5)4] [6] and
to NaCl, H3N·B(C6F5)3, and Et2O·B(C6F5)3 was detected.[21]
[B(ArF)4] (ArF = C6H3-3,5(CF3)2)[7] ions mainly used in
Crystalline [CPh3]+[(F5C6)3B(m-NH2)B(C6F5)3] may even be
[3]
homogenous catalysis. Salts of both ions are commercially
stored in a closed container in air for prolonged periods
without decomposition.[18a]
available which promotes their use in many applications. The
ligands in these anions were then modified with the aim of
making new ions which were less coordinating and more
soluble in hydrocarbon solvents. Hence the CF3 groups in the
[B(ArF)4] ion were replaced by larger perfluoroalkyl groups
giving the modified [B(ArF’)4] ions (ArF’ = C6H3-3,5(RF)2 ;
RF = n-C6F13,[8] n-C4F9,[9] 2-C3F7[9]). The borate with RF = nC6F13 can be used to make cationic, transition-metal catalysis
compatible with fluorous, biphasic recycling techniques.[8]
However, in terms of anion stability, it was shown that the
parent [B(ArF)4] ion with RF = CF3 is more stable against
methanolic sulfuric acid than the modified borate with RF = 2Scheme 1. The reaction of Na+[(F5C6)3B(m-NH2)B(C6F5)3] with HCl
[9]
proceeds without decomposition of the anion.
C3F7 in agreement with the greater electron-withdrawing
Ingo Krossing was born 1968 in Berlin and
studied chemistry at the Ludwig-Maximillians-Universit%t in Munich. His PhD work
was supervised by Prof. H. N.th (completed
1997). After a post doctoral stay with Prof.
Jack Passmore (Feodor Lynen Fellow) at the
University of New Brunswick he returned to
Germany in 1999 and started his Habilitation work in the group of Prof. H. Schn.ckel
in Karlsruhe as Liebig Fellow. Since the completion of the Habilitation in 2002 he is Heisenberg Fellow and group leader at the University of Karlsruhe (TH). For the research
done during his Habilitation he was awarded the ADUC prize of the
German Chemical Society (GDCh).
2068
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Ines Raabe was born in 1977 in Karlsruhe
and studied chemistry at the University of
Karlsruhe (TH). Since the completion of her
diploma thesis in January 2003 she has
worked as a PhD student in the group of
Ingo Krossing. Her topics are the chemistry
of weakly coordinating anions and the stabilization of highly electrophilic species, such
as CX3+ ions (X = halogen), in condensed
phases.
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
In these dimeric [(F5C6)3B(m-X)B(C6F5)3] ions the
boron atoms were again exchanged for aluminum[20] with X
being
imidazolyl.
Interestingly
this
aluminate,
[(F5C6)3Al(m-C3N2H3)Al(C6F5)3] ,
although
sensitive
towards hydrolysis, generated a more active catalyst than
the homologous borate.[20] Sophisticated perfluoroarylbased diborane Lewis acids were designed, of which the
ion [C6F4-1,2-{B(C6F5)2}(m-OCH3)] , derived from C6F41,2-{B(C6F5)2}[22] appears to be one of the most promising.[23] However, the this anion is not stable in the presence
of the [SiEt3]+ ion.[23] Another very stable but relatively
small borate is [B(CF3)4] , which can now be made in good
Figure 2. Structures of selected carborane-based WCAs.
quantities from M+[B(CN)4] and ClF3 (M+ = univalent
cation).[24, 25] This anion is stable towards elemental fluorine
in anhydrous HF solution as well as to sodium in liquid
ammonia! Its silver salt Ag+[B(CF3)4] readily coordinates
nucleophilic carborane-based ion, Xie et al. produced almost
completely halogenated carboranate ions [1-H-CB11X5Y6] [38]
the weak Lewis base, CO and even generates the hitherto
+
unknown [Ag(CO)4] complex at higher CO pressures, as
and [1-Me-CB11X11] (X, Y = Cl, Br, I),[38b] and the partially
[25]
confirmed by vibrational spectroscopy. By comparison of
alkylated [1-Me-CB11H5X6] (X = Cl, Br, I)[39] ions. Michl
+
the equilibrium CO pressure over solid Ag [A] ([A] =
et al. prepared a permethylated [CB11Me12] ion[40] as
WCA) it was concluded that the [B(CF3)4] ion is less
well as an—albeit explosive—pertrifluoromethylated
[CB11(CF3)12] [41] derivative and Strauss et al. synthesized
coordinating than [Nb(OTeF5)6] , [B(OTeF5)4] , and
[25]
[Sb2F11] .
highly fluorinated carboranates including the [1-R-CB11F11]
However, [B(CF3)4] is incompatible with the
+
+
extremely reactive [SiEt3] or [AliPr2] cations as well as with
ions (R = Me, Et, etc.).[42, 43] From the 29Si NMR spectra of the
the strong Lewis acid AsF5. A simpler derivative of the latter
iPr3Sid+CBd “silylium” species (where CBd is a carborane
ion, known since 1960, the [(F3C)BF3] ion was initially
ion) of all the given carborane ions CBd, the [1-R-CB11F11]
+
[26]
+
prepared as [Me3Sn] [(F3C)BF3]
salt, but now its Li salt
(R = Me, Et) ions are the least coordinating.[34d] The halo
as well as of the related [(F3C)2BF2] ion are known and both
genated 10-vertex carboranate ions appear to be more
coordinating than the 12-vertex carboranate ions.[44] A very
appear to be favorable electrolytes for lithium-ion batteries.[27] Also the [FB(CF3)3] ion[28–31] that is almost as stable as
recent addition was the development of a convenient highyield synthesis of [B12F12]2 salts.[45] Despite the dianionic
the homoleptic [B(CF3)4] ion[25] is now readily available. It
seems to be easier to obtain single crystals suitable for X-ray
character, [B12F12]2 stabilizes cationlike species, such as
crystallography from salts of the [FB(CF3)3] ion, presumably,
[AlMe2]d+ or [SiR3]d+ and thus it may be an useful and more
because it is less regular and its salts are less prone to twinning
easily accessible extension of WCA (car)borane ion chemisthan those of the [B(CF3)4] ion.[25] The related [F4xB(RF)x]
try.
Despite their proven capability as weakly coordinating as
ions are described in the literature.[32] .
well as chemically very robust anions, the halogenated
carborane ions are not widely used owing to the expensive
and time consuming multistep procedure of their preparation.
2.2. Carborane-Based Anions
However, very recently a simple and straight forward highyield two-step process giving the [CB11H12] ion from NaBH4,
An alternative to the anions based on a Lewis acidic
central atom strongly bound to ligands that are normally
CHCl3, base, and BF3·OEt2 was published.[46] This method
weakly coordinating (Section 2.1) is the use of a stable,
may open the door for wide application of the halogenated
univalent, polyhedral central moiety, such as the closocarborane ions.
carbaboranate ions [CB11H12] or [CB9H10] . All the exohedral BH bonds in these ions are very stable and only weakly
2.3. Alkoxy- and Aryloxymetallates
coordinating, however, both ions are prone to oxidation. It
was shown in the mid 1980s that (partial) halogenation makes
these ions less coordinating and more stable towards oxidaA recent alternative to the fluorinated tetraarylboratetion. In the last decade the [CB11H6X6] ion (X = Cl, Br)
based anions described in Section 2.1. is the use of poly- or
perfluorinated alkoxy- (ORF) and aryloxy- (OArF) metallates
developed by StGbr et al.[33] and Reed et al.[2a] emerged as one
of the most chemically robust WCAs known to date. Thus
(Figure 3). Oxophilic and strongly Lewis acidic atoms M, such
using the [CB11H6Cl6] ion (Figure 2) and a recent modificaas BIII, AlIII, NbV, TaV, YIII, and LaIII were used as center of the
tion, the [1-H-CB11Me5Cl6] ion,[34d] it was possible to prepare
“ate” complexes [M(ORF)n] [47–51] and [M(OArF)n] .[52, 53]
the free Brønsted acid,[34] a free silylium ion,[35] stable
Compared to the [B(C6F5)4] ion and related borates these
+
+ [36]
fullerene ions C60 and [HC60] , an approximation of the
metallates offer the advantage of being easily and safely
accessible on a lagre scale. Their preparation circumvents the
[AlEt2]+ alumenium ion,[37] and protonated benzene and
need to make LiC6F5, which is required in the generation of
toluene salts that are thermally stable to 150 8C.[34] Other
groups embarked on the journey to synthesize the least
[B(C6F5)4] , related borates, as well as B(C6F5)3, and which is
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2069
Reviews
I. Krossing and I. Raabe
2.4. Teflate-Based Anions
Replacing the small fluorine atoms in [BF4] and [MF6]
ions by the larger univalent OTeF5 moiety leads to the large,
and against electrophiles robust WCAs [B(OTeF5)4] [62] and
[M(OTeF5)6] (Figure 4; M = As,[63] Sb,[64, 65] Bi,[63] Nb[64, 66]) in
Figure 3. Structures of selected fluorinated alkoxy- and aryloxy-metallate based WCAs.
known to be unstable towards LiF elimination; its use has led
to several explosions. The [M(OC6F5)n] ions generate
cationic polymerization catalysts which have activities that
are similar to, or even higher than those of catalysts partnered
with the [B(C6F5)4] ion.[52, 53] However, the fluorinated
aryloxide-based [M(OC6F5)6] ions with M = Nb or Ta are
susceptible to OC6F5 abstraction by sterically open cationic
zirconocene catalysts, such as [Cp2ZrMe]+ (Cp = C5H5).[52] In
addition, the oxygen atoms and the CF bonds of the OArF
ligands in [M(OArF)n] [52] have a tendency to coordinate.
Substitution of OArF for sterically demanding ORF alkoxy
ligands, such as OC(CF3)3 generated the very stable and very
weakly coordinating [Al{OC(CF3)3}4] ion.[47, 48c] The donorfree naked Li+[Al{OC(CF3)3}4] salt is easily prepared in
100 g scale within two days in 94 % yield starting from
commercially available reagents.[54] In contrast to the easily
hydrolyzed alkoxyaluminates, the [Al{OC(CF3)3}4] ion is
stable in nitric acid[47] and also its Brønsted acid
[H(OEt)2]+[Al{OC(CF3)3}4] may be generated in high
yield.[55a] This stability towards hydrolysis was attributed to
the steric shielding of the oxygen atom by the bulky C(CF3)3
ligand and to the electronic stabilization resulting from the
perfluorination. This stabilization is demonstrated by the
increasing acidity of the tertiary alcohols, from the nonfluorinated HO-C(CH3)3 (pKa = 19.3), to the partially fluorinated HO-C(H)(CF3)2 (pKa = 9.5), and finally the perfluorinated HO-C(CF3)3 (pKa = 5.5).[48c] A systematic analysis of
the solid-state contacts of several silver salts of WCAs
including [B(OTeF5)4] and [CB11H6Cl6] showed the
[Al{OC(CF3)3}4] ion[47] to be at least as weakly coordinating
as the [CB11H6Cl6] ion which currently claims the title “least
coordinating ion”.[2a] This conclusion was substantiated by a
number of subsequent publications that showed the weakly
coordinating nature of the [Al{OC(CF3)3}4] ion[47] towards
silver adducts of very weak Lewis bases, such as P4,[56] P4S3,[57]
S8,[58] and C2H4.[55] In addition the [Al{OC(CF3)3}4] ion is
chemically very robust as was shown by the synthesis of its
salts with highly electrophilic cations, such as [PX4]+, [P2X5]+,
[P5X2]+,[59] and [CI3]+.[60] Also the [Al{OC(CF3)3}4] ion was
shown to generate cationic polymerization catalysts of similar
activity to those with the [B(C6F5)4] ion.[61]
2070
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. Structures of selected teflate-based WCAs.
which the negative charge is dispersed over 20 or 30 fluorine
atoms, respectively. The borate [B(OTeF5)4] is less stable
than its group 15 counterparts [M(OTeF5)6] (M = As, Sb, Bi)
and is prone to the loss of one OTeF5 group in the presence of
strong electrophiles, such as [SiR3]+ or even Ag+.[62] There is
no report on the decomposition of the respective arsenate or
antimonate, but the niobate [Nb(OTeF5)6] looses O(TeF5)2 in
the strongly coordinating solvent CH3CN and forms trans[NbO(OTeF5)4(NC-CH3)] .[64] From what has been reported
it appears that the antimonate [Sb(OTeF5)6] is the most
stable WCA of this series. All teflate-based WCAs require the
strict exclusion of moisture, they decompose rapidly in the
presence of only trace amounts of water. Nevertheless, the
borate, arsenate, and antimonate were used to stabilize
complexes of weak Lewis bases, such as CO in
[Ag(CO)2]+ [67] as well as electrophilic [AsX4]+,[68] [SbX4]+ [69]
and [CX3]+ [70] (X = Cl, Br) or [Cl3Te-F-TeCl3]+ [71] ions. To our
knowledge no account on the use of the teflate-based WCAs
in catalysis has appeared.
2.5. Anions Formed by Reaction with Lewis Acids
Instead of using the starting materials that already include
the entire WCA, the ion may also be formed in the course of
the reaction. Normally a strong Lewis acid achieves this by
abstracting an ionic fragment from the substrate. With the
exclusion of some combined Lewis acid/oxidation reactions of
MF5 (M = As, Sb) this approach gives only the desired salt
and thus circumvents the otherwise necessary separation of
product and byproduct. Especially when the generated
molecules, such as Ph3CR (from [Ph3C]+) or OEt2 (from
[H(OEt2)2]+), may react with the desired “naked” cation, the
Lewis acid approach has proven to be very valuable.
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
2.5.1. Anions From AsF5, SbF5, and Related Lewis Acids
The [MF6] ions are a common class of ions that may also
be introduced by metathesis with MI[MVF6] salts. However, by
using very weakly basic conditions throughout (that is, the use
of anhydrous HF (aHF), SO2, SO2ClF, liquid SbF5 or similar
solvents), di-, tri-, and tetranuclear ions that are very weakly
coordinating and increasingly robust towards electrophiles
can be prepared from the reaction of [MF6] with excess MF5
(Figure 5). The [As2F11] ion has been known for a long
(FIA 591 kJ mol1)[82] followed by SbF5 (FIA 503 kJ mol1)[81]
and AlF3 (FIA 481 kJ mol1).[81] Typical FIA values[81] of
classical Lewis acids are (in kJ mol1): AsF5 (443), PF5 (397),
and BF3 (347). However, AuF5 is so reactive that it forms
AuF3 and F2 when dissolved in aHF,[82] therefore SbF5 is the
strongest Lewis acid that can be handled conveniently. We
will use this FIA approach for a general classification of all
types of WCAs described in this article (see below in
section 7.).
2.5.2. Anions From Organometallic Lewis Acids
Several organometallic perfluoroaryl-based
Lewis acids were developed[83] after the observation that the Lewis acidity of B(C6F5)3[6, 84] is high
enough to abstract a methyl ion from a group 14
metallocene, thus forming the [MeB(C6F5)3] ion
and generating active polymerization catalysts
(Figure 6).[85, 86] A review on L-B(C6F5)3 complexes (L = neutral Lewis base) has been published.[87] Following on from the original
reports,[85, 86] the C6F5 group in B(C6F5)3 was
partially or fully replaced by 2-perfluorobiphenyl[88] and 2-perfluoronaphtyl.[89] Later also
diborane Lewis acids, such as ortho-phenylenediborane C6F4-1,2-{B(C6F5)2},[90a] para-phenylenediborane C6F4-1,4-{B(C6F5)2},[91] and octafluoro-
Figure 5. Structures of selected multinuclear fluorometallate-based WCAs.
time,[72] but was only recently structurally verified.[73] The
[Sb2F11] ion is relatively abundant,[79] however, salts of the
[Sb3F16] [74] and [Sb4F21] [75] ions, which are more stable
towards electrophiles, are rare. The [Sb3F16] and [Sb4F21]
ions in particular have helped to solve some very fundamental
questions. With salts of them is was possible to prepare and
fully characterize simple but very reactive cations, such as
[Br2]+ [76] and [Xe2]+.[75] The larger the fluorometallate
[MnF5n+1]is, the less coordinating it is. This tendency as may
be seen in the series of [Au(Xe)n]2+ salts in which the cation
with n = 4[77] is only accessible with the weaker Lewis base
[Sb2F11] but salts with n = 1, 2[78] are also accessible with the
more basic [SbF6] ion. Similar effects were seen in the
chemistry of homoleptic metal carbonyl cations.[79] A problem
associated with using any of the fluorometallates [MnF5n+1] is
that mixtures with varying values of n = 1–4 exist in solution.
This situation makes crystallization more difficult and provides the free Lewis acid MF5, which can act as an oxidizing
agent and thus cause unwanted side reactions.
As a measure for the strengths of the parent Lewis acids
MFn (= A) and also of the stability of the [AF] ion the
affinity of several Lewis acids A towards the fluoride ion
[fluoride ion affinity FIA, Eq. (1)] was examined based on
DH¼FIA
AðgÞ þ FðgÞ ƒƒƒƒ
ƒ!AFðgÞ
ð1Þ
lattice enthalpy considerations[80] as well as quantum chemical
calculations.[81, 82]
From the available computational data[80–82] it follows that
gaseous, monomeric AuF5 is the strongest known Lewis acid
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Figure 6. Structure of the [MeB(C6F5)3] ion formed by methyl abstraction with
B(C6F5)3 and the new organometallic diborane Lewis acid C6F4-1,2-{B(C6F5)2}.
9,10-bis(pentafluorophenyl)-9,10-diboraanthracene[92] were
prepared. From the polymerization results of mixtures of
these organometallic boron Lewis acids with group 14 dimethylmetallocenes it followed that the resulting [MeB(ArF)3]
(ArF = C6F5, perfluorobiphenyl, or perfluoronaphtyl) ions are
more strongly coordinating than the homoleptic [B(C6F5)4]
ion.[3] This result is also apparent in several solid-state
structures with [MeB(ArF)3] ions which include coordinated
[M]MeB moieties, that is, in [(1,2-Me2Cp)2(Me)Zr]+[(mCH3)B(C6F5)3] in which the bridging Zr-Me separation is
about 30 pm longer than the terminal Zr-Me bond.[85] Hence,
no full ionization but tight ion pairing is achieved by the
[MeB(ArF)3] ions (see Section 6.2). Recently the partially
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2071
Reviews
I. Krossing and I. Raabe
fluorinated Lewis acid (N-pyrrolyl)B(C6F5)2 was prepared
and successfully used in the generation of active polymerization catalysts.[93]
3. Applications of Weakly Coordination Anions
It exceeds the scope of this Review to give full reference
to all the possible applications of WCAs. We only scratch the
surface of selected new areas of growing importance in which
the special properties of WCAs are currently exploited and
give reference to recent developments. For important mature
applications, such as group 14 metallocene-based and related
olefin-polymerization reactions[3, 83, 94] and non-metallocene
catalysts,[95] comprehensive and up to date reviews are
available.
3.1. Li+- and Ag+-Ion-Catalyzed Organic Reactions
Organic transformations, such as the Diels–Alder reaction, 1,4-conjugate addition reactions, pericyclic rearrangements and others are catalyzed by the Li+ ion of a WCA.
Traditionally these reactions were performed in concentrated
Li+[ClO4] solution[96, 97] (approx. 3–5 m in Et2O) which is
potentially very dangerous owing to the explosive nature of
the ClO4 ion. The new generation of WCAs provides access
to almost “naked” Li+ ions, which are soluble and catalytically
active even in highly nonpolar solvents, such as toluene[98] and
hexane.[48a] It was shown that very low concentrations of the
Li+[X] catalyst (X = WCA; c = 0.01–0.1m) are sufficient to
promote the reactions. Lithium salts of [Al(ORF)4] (RF =
C(Ph)(CF3)2),[48a]
[B(ArF)4]
(ArF = C6H3(CF3)2),[99]
F
F
[100]
[Nb(OR )6]
(R = C(H)(CF3)2),
[CB12Me12] ,[98] or
[101]
[Co(B9C2H11)2]
were used for such transformations.
Silver ions were used as catalysts[97b] and a recent
investigation showed [Ag(PPh3)]+[CB11H6Br6] to be air and
moisture stable as well as to be the best catalyst for a series of
hetero Diels–Alder reactions. In addition this silver species
can used with the low catalyst loadings of only 0.1 mol %
(compared to 5–10 mol % with classical ions)[102]
3.2. Electrochemistry
The most commonly used supporting electrolytes used in
electrochemistry are of type NR4+[X] (R = Me, Et, Pr, Bu;
X = ClO4, BF4, PF6, CF3SO3). However, regularly the generated oxidized species are very reactive and decompose the
X ion.[103] Although CH2Cl2 is a preferred solvent for
oxidation reactions, in this solvent the precipitation of
insoluble polycationic products at the electrode may complicate voltammetric scans and often a more polar but also more
reactive solvent, such as acetonitrile, must be used.[104]
However, not all the desired oxidized species are compatible
with the more polar solvents. Both types of problems are
avoided when supporting electrolytes with a robust and
large WCA are used. [N(Bu)4]+[B(C6F5)4] [105] and
2072
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[N(Bu)4]+[B{(C6H3(CF3)2}4] [106] were reported to be very
stable towards the formed oxidized species and effectively
solubilize positively charged species formed in anodic processes. Also Li+[CB12Me12] was used and allowed cyclic
voltammetric measurements to be performed even in benzene.[107] The [F3P(C2F5)3] ion[108] is now commercially
available and its salts are more soluble and more stable
than the coresponding PF6 salts.
3.3. Ionic Liquids
Ionic liquids (ILs) are of increasing importance for
industrial and laboratory processes and many are commercially available.[108] More specifically, the nonvolatile ILs are
expected to replace organic solvents in industrial processes
and thus are viewed as “green solvents”. Traditional ILs use
[AlCl4] , [PF6] , and related ions partnered with unsymmetrical large cations, such as EMIM+ (1-ethyl-3-methyl
imidazolium) or quaternary pyridinium salts,[109] but recently
the effect of the ion in the IL on a hydrovinylation reaction
was systematically studied and showed ILs with WCA
counterions,
such
as
[Al{OC(Ph)(CF3)2}4]
or
[B{(C6H3(CF3)2}4] to be the most effective in enhancing the
conversion.[110] A comprehensive examination of ILs based on
imidazolium melts with carborane ions appeared recently.[111]
Also some [Al(ORF)4] –imidazolium melts are known.[48c]
3.4. Lithium-Ion Batteries
Another area of industrial importance is the development
of new stable electrolytes for lithium-ion batteries. Currently
Li+[PF6] in ethylene carbonate or propylene carbonate is
used almost exclusively in practice. However, to further
increase the possible battery current, more stable ions [X]
that still generate highly conducting Li+[X] solutions are
needed. To increase the conductance of the lithium electrolyte both, size and coordinating ability appear to be important. Therefore the ultimate Li+[X] electrolyte should
contain a very small but very weakly coordinating WCA
[X] . Additionally it should be nontoxic, hydrolytically stable,
and be easily prepared from cheap starting materials. Likely
candidates are mainly reported in the patent literature; some
of the recent small but robust WCAs candidates are
[F4xB(CF3)x] (x = 2, 3, 4),[27, 112] [F6xP(CF3)x] (x = 3, 4,
5),[113]
[F3P(C2F5)3] ,[108]
and
[B(R’)2]
(R’ =
[114]
OC(CF3)2C(CF3)2O).
However, contradicting the expectation, larger WCAs, such as [B(ArF)4] ,[115] [M(ORF)4] (M =
B, Al; RF = C(H)(CF3)2 etc.)[48c] were shown to give very
high[115] or even the highest conductivities.[49]
3.5. Extraction of LnIII Ions
LnIII lanthanide ions can be extracted from HNO3
solutions (c > 1.0 m) into the organic phase if partnered with
a suitable extractant. Now it has been found that extraction
into the organic phase and distribution between aqueous and
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
organic phase can be enhanced by several orders of magnitude when the extractant is added as its [B(ArF)4] salt (ArF =
C6H3(CF3)2).[116]
3.6. Photoacid Generators
In recent years diaryl iodonium salts have been widely
used as photoacid generators (PAGs) for cationic polymerization of many different monomers and oligomers. Upon
irradiation with UV light the iodonium salts undergo
irreversible fragmentation and generate a strong Brønsted
acid which initiates cationic polymerization. As a result of
these properties PAGs are used in many applications, such as
photolithography, printing inks, release coatings, optical
fibers, and holography.[117] In most commercial salts [SbF6]
is the counterion, however, because of the toxicity as well as
hygroscopic nature of the iodonium salts of [SbF6] , the
counterion was recently exchanged for a series of WCAs
which included [M(C6F5)4] (M = B,[118] Ga[119]), [CB11H6Br6]
and [(F5C6)3B(m-X)B(C6F5)3]n (X = C3N2H3, n = 1; x = C6F4,
n = 2).[120] The photoactivities of these WCA PAGs were
tested in the polymerization of a nonpolar, epoxy silicon
oligomer as well as a polar, cycloaliphatic epoxide monomer
and indicated that the [M(C6F5)4] (M = B, Ga) and
[(F5C6)3B(m-X)B(C6F5)3] (X = C3N2H3) salts were most
effective.[120] The unwanted side effects of the [SbF6] ion
were eliminated by using these WCA photoacid generators.
4. Available Starting Materials and Strategies to
Introduce Weakly Coordinating Anions
Important factors to be considered when deciding which
of the multitude of known WCAs should be used, are the
availability, accessibility, and price, as well as knowledge of
the desired degree of “noncoordination” necessary for a given
problem. Hence, the right choice normally is a WCA that is
stable and weakly coordinating enough to allow the desired
transformation but not necessarily the ultimate most weakly
coordinating ion. Sometimes the commercial availability of
the WCA will play a deciding role, that is, salts of the
[B{C6H3(CF3)2}4] and [B(C6F5)4] ions, as well as the Lewis
acid B(C6F5)3, can be bought and certainly account for their
seemingly ubiquitous use in homogenous catalysis. However,
attractive alternatives to these established reagents are available and can be made in high yield and in a short time. It is
worth spending a little time to consider using one of the new
WCAs, in particular the fluorinated alkoxy- and aryloxymetallates or the bridged dimeric [(F5C6)3B(m-X)B(C6F5)3] ions.
4.1 Available Starting Materials
Table 1 gives an overview to the preparation of Li+, Ag+,
Tl , M+, CPh3+, H+, H(OEt2)2+, H(L)+, ammonium, and
imidazolium salts of WCAs. A “–” in Table 1 indicates that no
simple preparation of this specific salt is known to us; this is
not necessarily due to (but may be due to) the instability of
the resulting salt.
+
Table 1: Available starting materials for the preparation of WCA salts (RF = fluorinated alkyl, ArF = fluorinated arene).
WCA
Li+
Ag+
Tl+
M+
[CPh3]+
H+
[H(OEt2)2]+ H(L)+
N(R1)(R2)(R3)(R4)+[b]
[Al(ORF)4]
[Al(OArF)4]
[M(ORF)n]
[M(OArF)n]
[47–49]
[49, 52]
Nb[49, 100]
Nb, Ta, Y,
La[49, 52]
–
[47]
–
–
–
[48b, 50]
–
–
–
Cs+ [51]
–
–
–
–
–
–
–
[48c]
–
–
B[53]
B[62]
–
–
–
Y,
La[52]
–
[55a]
–
–
–
Sb, Nb,[64]
B[62]
–
–
–
[51]
[52]
[100]
Nb, Ta, Y,
La[52]
Sb, Nb,[64]
B[62]
–
[123]
CF3,[10] SiiPr3,
SiMe2tBu[11]
–
–
–
–
–
[122b, 128]
[130]
–
–
L = C6Me5H[34c]
–
As, Sb, Bi,[63, 64] Nb,
Ta,[66] B[62]
[129]
[122b, 131]
B,[12][h] 4-CF3[10]
NH2,[21]
CN[18]
A[90][g]
–
–
NH2,[21]
–
C3H3N2[19][f ]
A[23][g]
–
[124b,c]
[34]
[132]
[M(OTeF5)n]
[B{C6H3(CF3)2}4] [121]
[B(C6F5)4]
[6]
4-CF3,[10]
[B(C6F4R)4]
SiiPr3,
SiMe2tBu[11b]
[(F5C6)3B(m-X)C6F4,[17]
–
n [e]
B(C6F5)3]
C3H3N2[19][f ]
other borates
[B(CF3)4] [25] [B(CF3)4] [25]
–
–
–
Sb:Cs+,[65]
Nb,Ta:Cs+ [66]
Na+ [121, 122]
–
–
–
Na+, NH2[21]
–
[1-H-CB11R5X6] [a] –
[124]
–
K+, Cs+:
B(CF3)4 ;[25]
Na+: C[8][i]
Cs+ [33, 124b]
[CB11RxX12x] [c]
[98, 125]
[38, 39]
[125]
[CB9RxX10x] [c]
[CB11RxF12x] [d]
–
–
[44]
[43]
–
–
Cs+,[39–41] Na+,
[39]
K+, Rb+, Cs+ [125]
Cs+ [44]
[44]
Cs+ [43]
[43, 126, 127]
C3H3N2,[17, 20] CN[18]
–
[111]
[133] [133]
C6H6,[34]
Ph2CO,
PhNO2[132]
[133]
[133] [133]
–
–
[133]
–
[44]
[43]
[38]
[a] R = H, Me, X = Cl, Br, I. [b] R1–R4 = H, Me, Et, Ph etc.; also imidazolium salts for ionic liquids. [c] R = H, Me, CF3. [d] R = H, Me, Et. [e] m-X = bridging
ionic ligand, such as CN , imidazolyl, NH2, -C6F4- etc.; n = 1, 2. [f] Also X = 4,5-dimethyl imidazole and benzimidazole. [g] A = [C6F4-1,2-{B(C6F5)2}(mX)] (X = F, OCH3, OC6F5). [h] B = B[{C6F4-4-C(C6F5)2F}4] . [i] C = B(ArF’)4] (ArF’) = C6H3-3,5(C6F13)2).
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2073
Reviews
I. Krossing and I. Raabe
In this section we present common strategies to introduce
a WCA into a system. To illustrate the procedure only
selected examples for each type of reaction are given.
The use of Lewis acids as anion generators may conflict
with the formation of Lewis acid–base adducts. Thus when
one equivalent of SbF5 or AsF5 were reacted with FC(O)NCO
only the oxygen-adduct F5M OC(F)NCO (M = As, Sb) was
isolated [Eq. (7 a)].[140] By using an excess of the Lewis acid
!
4.2. Strategies to Introduce WCAs
4.2.1. Direct Oxidation of the Substrate
Silver salts of WCAs may oxidize transition-metal complexes with an intermediate oxidation potential. Owing to the
high solubility of the WCA salts in weakly coordinating
solvents these oxidation reactions may be performed in
solvents such as CH2Cl2, in which the Ag+/Ag0 couple has the
highest oxidation potential.[134] Thus, ferrocene is oxidized by
Ag+[Al(ORF)4]
(RF = C(CF3)3)
in
CH2Cl2
giving
+
F
[135]
[Cp2Fe] [Al(OR )4] .
Another class of oxidizers are the [NAr3]C+ radical cation
salts of WCAs (Ar = halogenated or nonhalogenated aryl).
The [NAr3C]+ salts are accessible by the reaction of Ag+[Y]
(Y = WCA) with X2 (X = Cl, Br) and NAr3 [Eq. (2)]:
Ar3 N þ 1=2 X2 þ Agþ ½Y ! AgX þ ½Ar3 NCþ ½Y
ð2Þ
The oxidation potential of the [NAr3]C+ salts may be tuned
by varying the degree of halogenation of the aryl ligand.
Oxidizing [NAr3]C+ salts of the carborane ions were recently
used to generate stable fullerene cations [Eq. (3); X =
CB11H6Cl6].[36, 136]
½Ar3 NCþ ½X þ C60 ! C60 Cþ ½X þ NAr3
ð3Þ
For electronic and steric reasons the free amines NAr3
that are formed as byproducts are weak Lewis bases. A
powerful oxidizing agent is the stable [CB11Me12]C radical.[137]
This reagent was used to generate a stable [Bu3Sn]+ ion
[Eq. (4)].[138]
Sn2 Bu6 þ 2 ½CB11 Me12 C ! 2 ½Bu3 Snþ ½CB11 Me12 ð4Þ
The formation of [SbX4]+ (X = Cl, Br) from 2 Sb(OTeF5)3
and 2 X2 provides an example for a disproportionation with
WCA formation. This reaction occurs upon oxidation of
Sb(OTeF5)3 to Sb(OTeF5)3X2 [Eq. (5); X = Cl, Br].[69]
2 SbðOTeF5 Þ3 þ 2 X2 ! ½SbX4 þ ½SbðOTeF5 Þ6 ð5Þ
4.2.2. Reactions with Strong Lewis Acids
A very versatile reaction is the methyl-group abstraction
from [Cp2M(CH3)x] (M = Ti, Zr, Hf, Ta; x = 2, 3) by strong
organometallic Lewis acids.[3] In the classical reactions of
B(C6F5)3 (and related Lewis acids) with [Cp2ZrMe2] a tight
ion pair [Cp2ZrMe]+[MeB(C6F5)3] is formed (see Section 2.5.2). A less common example is the reaction of
[Cp2TaMe3] with two equivalents of Al(C6F5)3 that delivers
the salt [Cp2TaMe2]+[(F5C6)3Al-Me-Al(C6F5)3] that contains
a bridging methyl group in the anion [Eq. (6)].[139]
½Cp2 TaMe3 þ 2 AlðC6 F5 Þ3 !
½Cp2 TaMe2 þ ½ðF5 C6 Þ3 Al-Me-AlðC6 F5 Þ3 2074
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ð6Þ
SbF5 (which enhances its Lewis acidity) even the very stable
CF bond in FC(O)NCO can be heterolytically cleaved
giving the bent [OCNCO]+[Sb3F16] salt [Eq. (7 b)].[140]
4.2.3. Hydride and Alkyl Abstraction
A prototype of hydride- and alkyl-abstraction reactions is
the reaction of the [Ph3C]+ salts of WCAs with [Cp2M(CH3)x]
(M = Ti, Zr, Hf, Ta; x = 2, 3). This reaction gives active olefinpolymerization catalysts, for example [Cp2ZrMe]+[B(C6F5)4]
(and MeCPh3) from [Cp2ZrMe2] and [Ph3C]+[B(C6F5)4] .[3]
Jordan et al. showed[141] that the alkyl-abstracting properties
of [Ph3C]+[B(C6F5)4] can be employed to generate lowcoordinate, cationic aluminum and indium complexes, provided sufficiently bulky ligands are bound to the metal atom
to prevent decomposition of the counterion. Bochmann et al.
showed that the [B(C6F5)4] ion decomposes in the presence
of the sterically open [AlMe2]+ ion. In fact, the reaction of
AlMe3 and [Ph3C]+[B(C6F5)4] in the appropriate stoichiometry provides the best access to Al(C6F5)3.[15b]
It was shown that the tight ion pair [Cp2ZrMe]+[MeB(C6F5)3] can be used to abstract ionic ligands from
transition-metal compounds, for example, Cp from
[Cp3Ta(L)] [L = butadiene; Eq. (8)].[142] An alternative to
methyl abstraction is b-hydride abstraction, provided that the
alkyl ligand bears at least one b-hydride atom. A current
example of this type of reaction is the reaction of AlEt3 and
[Ph3C]+[CB11X6H6] giving an alumenium-like molecular
Et2Al(CB11X6H6) species [Eq. (9); X = Cl, Br].[37]
In the structurally characterized Et2Al(CB11Br6H6) the
aluminum atom[37] forms two comparatively strong AlBr
bonds (at about 256 pm) which reduces the electrophilicity of
the aluminum atom compared to that predicted for the still
unknown free, two-coordinate [AlEt2]+ ion (see Figure 12 in
Section 6.1). A similar but dimeric compound was prepared
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
be used if the substrate is oxidized by Ag+ (see Section 4.2.1).
Two recent examples for the use of silver salts of WCAs in
metathesis reactions involve the preparation of the homoleptic metal carbonyl cations [Rh(CO)4]+[1-Et-CB11F11] [145]
and [Cu(CO)4]+[1-Et-CB11F11] [Eqs. (12) and (13)].[42]
2 Agþ ½1-Et-CB11 F11 þ ½fRhðClÞðCOÞ2 g2 þ 4 CO
! AgCl þ 2 ½RhðCOÞ4 þ ½1-Et-CB11 F11 Agþ ½1-Et-CB11 F11 þ CuCl þ 4 CO
! AgCl þ ½CuðCOÞ4 þ ½1-Et-CB11 F11 by reaction of AlR3 and [Ph3C]+[1-Me-CB11F11] (R = Me,
Et).[126]
AlEt3 þ ½Ph3 Cþ ½CB11 X6 H6 !
Et2 AlðCB11 X6 H6 Þ þ Ph3 CH þ C2 H4
ð9Þ
Straight hydride abstraction is the key to the synthesis of
the “silylium-like” molecules R3Si-CB [Eq. (10)].[143]
½CPh3 þ ½CB þ HSiEt3 ! Et3 Si-CB þ HCPh3
4.2.4. Metathesis Reactions
WCAs can be introduced into a system through metathesis reactions of M+[X] ([X] = WCA, M = univalent
metal, such as Li, Na, Ag, Tl) with labile or sometimes even
covalently bound halides. A recent example from catalysis is
shown in Equation (11) (ArF = C6H3(CF3)2, L = 2-(phosphanylaryl)oxazoline, COD = cyclooctadiene).[144]
! ½ðLÞIrðcodÞþ ½BðArF Þ4 þ NaCl
ð11Þ
In many cases Ag+ is the best cation to abstract the halide
from the substrate, because the lattice potential enthalpies of
the silver halides are exceptionally high and thus provide a
large driving force for the ionization. However, silver may not
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
ð13Þ
To prepare the homoleptic Rh carbonyl cation the pure
donor free silver salt Ag+[1-Et-CB11F11] has to be used. If the
benzene adduct [Ag(C6H6) ]+[1-Et-CB11F11] is used the final
product is [(C6H6)Rh(CO)2]+[1-Et-CB11F11] .[145] This example shows the necessity of using strictly nonbasic conditions
throughout the synthesis of such electrophilic species. Even
solvents commonly viewed as nonbasic, such as benzene, may
interact with the generated active cationic species!
If silver salts of good WCAs, such as [Al(ORF)4] (RF =
C(CF3)3), are used in weakly basic but polar solvents, such as
CH2Cl2, even covalently bound halide ions can be abstracted
from simple molecules, such as CI4 or PX3 (X = Br, I). Thus
the reaction of Ag[Al(ORF)4] with CI4 led to the roomtemperature-stable salt [CI3]+[Al(ORF)4] which was the first
compound of a binary CX cation (X = halogen) characterized by X-ray crystallographic analysis [Eq. (14)].[60]
ð10Þ
The driving force for this reaction is the formation of the
CH bond, which is stronger than the parent SiH bond.
Until very recently all structurally characterized silylium-like
species were either coordinated to the anion or to a solvent
molecule, such as benzene or toluene. However, the crystal
structure of the salt [Mes3Si]+[CB11Cl6Me6] (Mes = 2,4,6trimethylphenyl) was published in 2002.[35] The [Mes3Si]+ ion
is trigonal planar and no contacts to the anion are observed
indicating the presence of a truly three-coordinate, isolated
[Mes3Si]+ ion in this salt.[35] Thus, the long standing “silylium
ion problem”[143] was finally solved by using an advanced
WCA![35]
Naþ ½BðArF Þ4 þ 0:5 ½fIrðcodÞðm-ClÞg2 þ L
ð12Þ
CH2 Cl2
Cl4 þ Ag½AlðORÞ4 ƒƒƒƒƒƒƒƒƒƒ!½Cl3 þ ½AlðORÞ4 þ AgI
RT; exclusion of light
ð14Þ
Prior to this experiment the [CX3]+ ions (X = Cl, Br, I)
were only detected as intermediates at low temperature
(78 8C).[146] This result shows that as long as a reactive cation
is compatible with the (weakly basic and weakly coordinating) solvent, species previously only known as intermediates
may even be stabilized at ambient temperatures provided that
anion–cation interactions are minimized by using a chemically robust and very weakly coordinating anion. This is
underlined by the recent report of the crystal structures and
solution characterization of pure [CX3]+[Sb(OTeF5)6] (X =
Cl, Br) salts.[70]
Solvent choice in the silver-salt metathesis reactions is
important for the success or failure of a planned reaction.
Thus, the reaction of a CH2Cl2 solution of Ag+[Al(ORF)4]
with PX3 (X = Br, I) at 78 8C generated the carbene
analogue [PX2]+ ions intermediates.[59] These intermediates
then inserted as electrophilic carbene analogues into the
XX, PX, or PP bond of X2, PX3, P2I4, or P4 to give the very
reactive, binary PX cations [PX4]+, [P2X5]+, [P3I6]+, and
[P5X2]+ [Eq. (15)].[59]
However, PBr bond activation in PBr3 fails in even
slightly more basic solvents, such as 1,2-Cl2C2H4. PBr3 and
Ag+[Al(ORF)4] mixtures in 1,2-Cl2C2H4 solution did not
react at 25 8C while the same mixture in CH2Cl2 reacts
immediately at 78 8C![59] This result again shows the
importance of the choice of the solvent if the maximum
halide-abstraction power of the Ag+ ion is desired for a
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2075
Reviews
I. Krossing and I. Raabe
reaction. CH2Cl2 appears to be one of the best choices in this
respect.
4.2.5. Reaction with Brønsted Acids
A new class of superacids, that is, Brønsted acids exceeding the acidity of pure H2SO4, was introduced by Reed et al.:
H+[CB] ([CB] = halogenated carborane-based WCA) is
prepared from the reaction of the molecular silyliumcarborane Et3Si-CB with excess liquid HCl [Eq. (16); [CB] =
[CB11R6X6] ; R = H, Me; X = Cl, Br].[34]
HClðlÞ
þ
Et3 Si-CB þ HClexx: ƒƒƒƒƒƒƒƒ!H
½CB þ ClSiEt3
o
196 C!78 C
ð16Þ
The H+[CB] superacids can be made in small scale and
are weighable, temperature stable, and non-oxidizing solids.
In contrast to the classical (oxidizing) superacids HF/SbF5 and
HSO3F/SbF5, the superacidity of H+[CB] is not accompanied
by Lewis acidity. This advantageous feature led to the first
protonation of fullerene (![HC60]+); with conventional
superacids and strong mineral acids only decomposition
occurred [Eq. (17)].[36] The superacid H+[CB] also protonates weakly basic solvents, such as benzene, toluene, and
nitrobenzene.[34]
Hþ ½CB þ C60 ! ½HC60 þ ½CB
ð17Þ
The simpler to prepare, but compared to H+[CB]
weaker, Brønsted acids [H(OEt2)2]+[X] have been widely
used in catalysis since the their preparation was first reported
in 1992 [X] = [B(ArF)4] [122b, 128a] and 2000 [X] =
[B(C6F5)4] .[130] It has become common practise to call
[H(OEt2)2]+[B(ArF)4]
“Brookhart's
acid”
and
[H(OEt2)2]+[B(C6F5)4] “Jutzi's acid”. However, it should be
noted that Taube and Wache[122b] published the preparation of
[H(OEt2)2]+[B(ArF)4] several months prior to Brookhart
et al. and thus it should correctly be called Taube's acid.[128a]
Both acids are often used for protonolysis of transition-metal
alkyls [M]R giving a cation [M]+ and the alkane RH. An
instructive example of the effects that have to be taken into
account for such reactions is the protonolysis of a diiminepalladiumdiethyl complex by two very similar Brønsted
acids [H(OEt2)2]+[B(ArF)4] and [H(OiPr2)2]+[B(ArF)4]
[Eq. (18 a)].[147] Protonolysis with [H(OEt2)2]+[B(ArF)4]
which contains the sterically less-encumbered donor led to
the OEt2 complex of the ethylpalladium cation [Eq. (18 a)]
2076
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
while the reaction with [H(OiPr2)2]+[B(ArF)4] which contains the less-accessible donor OiPr2 gave the desired agostic
ethylpalladium cation [Eq. (18 b)].[147]
Reactions of tertiary ammonium salts [HNR3]+[X]
([X] = WCA) with metal alkyl compounds are related to
the reactions of the protonated ethers [Eq. (19)]. Dimethylanilinium salts are most frequently employed. This approach
½HNR3 þ ½X þ ½Cp2 ZrðCH3 Þ2 ! ½Cp2 ZrðCH3 Þþ ½X þ H-CH3 þ NR3
ð19Þ
is limited to substrates where the inferior Brønsted acidity of
the [HNR3]+ salts is still sufficient to protonate the metal alkyl
compounds, in addition the generated free amine must not
coordinate to the resulting cation. An example of this side
reaction is the coordination of dimethylaniline (Me2NPh),
which is produced by protonolysis in the reaction of
[HNMe2Ph]+[B(C6F5)4] with [(iPr2-ATI)InMe2] (iPr2-ATI =
N,N’-diisopropyl aminotroponiminate) [Eq. (20)]. An elegant
variation of this reaction in which the cation was generated
and the ligand incorporated at the same time was recently
published by Bochmann et al. [Eq. (21)].[148]
The iminium salt of the WCA [B(C6F5)4] reacts with
dialkylzinc complexes to give diimine alkylzinc cations that
serve as catalysts for the ring-opening polymerization of
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
epoxides and e-caprolactone.[148] A related neutral Brønsted
acid is the dipolar pyrroliumborate (H5C4N)B(C6F5)3 that
protonates [Cp2ZrMe2] with liberation of methane and
formation of [Cp2ZrMe]+[(H4C4N)B(C6F5)3] [Eq. (22)].[149]
Equation (24) shows that atomic Xe is a stronger base
towards Au2+ than the ion [Sb2F11] and the solvent hydrogen
fluoride. In all cases where weak ligands are to be coordinated
to produce a given cation, the choice of solvent and WCA is
crucial. To stabilize the resulting Lewis acid–base adducts, the
anion–cation as well as solvent–cation interactions have to be
minimized and together they must be weaker than the
interaction with the weakly basic ligand. This principle may
also be illustrated by a series of Ag-P4 and Ag-S8 complexes:
1) The role of the counterion: In the series of silver salts
Ag[Al(ORF)4] with RF = C(H)(CF3)2, C(Me)(CF3)2, and
C(CF3)3 the perfluorinated ion [Al{OC(CF3)3}4] is the
least nucleophilic. The addition of an excess of ligand L =
P4[56] or S8[58] to Ag[Al(ORF)4] led, for the slightly more
basic ions RF = C(H)(CF3)2, C(Me)(CF3)2, to molecular
adducts (L)Ag[Al(ORF)4] (L = S8,P4) in which the anion is
coordinated directly to the silver center [Eq. (25a)]. In
contrast, with the least basic ion [Al{OC(CF3)3}4] , P4 and
S8 are stronger Lewis bases towards the cation than the
anion and consequently the [Ag(L)]+ ions coordinate a
second equivalent of L and the salts [Ag(L)2]+[Al{OC(CF3)3}4] are formed [Eq. (25 b)].[56, 58]
2) The role of the solvent: When Ag+[Al{OC(CF3)3}4] and
two equivalents of P4 are mixed in CH2Cl2 the equilibrium
shown in Equation (26) ([X] = Ag+[Al{OC(CF3)3}4])
4.2.6. Nucleophilic Substitution at and Addition to the Cation of
a WCA Salt
½AgðsolventÞn þ ½X þ 2 P4 G
in C6 H6
A pure nucleophilic substitution reaction at the cation of a
WCA is rare. An illustrative example of such a reaction is the
substitution of the fluorine atom in [N2F]+ salts by an N3
ligand to give the pentanitrogen cation [N5]+ and HF
[Eq. (23); [X] = [AsF6] , [SbF6] , [Sb2F11]].[150]
½N2 Fþ ½X þ HN3 ! ½N5 þ ½X þ HF
½AuðFHÞ4 2þ ð½X Þ2 þ 4 Xe ! ½AuðXeÞ4 2þ ð½X Þ2 þ 4 HF
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
ð24Þ
þ
H½AgðP4 Þ2 ½X þ n solvent
ð26Þ
lies completely to the right. However, replacing the
weakly basic solvent CH2Cl2 for benzene, which is more
strongly coordinating towards the Ag+ ion, led to decomposition of the [Ag(P4)2]+ ion, replacement of P4 by C6H6,
and formation of [Ag(C6H6)2]+. Thus in benzene the
equilibrium lies completely to the left.[56]
ð23Þ
Other reactions can be described as addition reactions; a
spectacular example would be the addition of four Xe atoms
into the coordination sphere of the Au2+ ion giving a
[Au(Xe)4]2+ salt.[77] However, there is no such thing as a
noncoordinated cation in solution and if the WCA is weakly
basic enough not to interact with the cation then the solvent
will do so. Therefore in this example the addition should be
reformulated as a substitution of the solvent HF by atomic
xenon [Eq. (24); [X] = [Sb2F11]].
in CH2 Cl2
A rare example of a pure addition reaction is the reaction
of O2+ salts with Cl2 giving the dark violet p*–p* bonded
trapezoidal [Cl2O2]+ salts [Eq. (27); [X] = [AsF6] , [SbF6] ,
Oþ2 ½X þ Cl2 ! ½Cl2 O2 þ ½X
ð27Þ
[Sb2F11]].[151] The thermal stability of the [Cl2O2]+ salts
formed increases with the decreasing basicity of the WCAs
and is highest for [Sb2F11] .[151]
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2077
Reviews
I. Krossing and I. Raabe
5. A Rationalization of the Special Properties of
Salts of Weakly Coordinating Ions Based on
Thermodynamic Considerations
Large and weakly coordinating anions offer some special
properties that are best rationalized on the basis of simple
thermodynamic Born–Fajans–Haber cycle considerations.
Overall WCAs enhance the stability of weakly bound and
low-charged species, such compounds often present unusual
bonding situations. Therefore the stabilization provided by
WCAs can be seen in analogy to the chemistry done with
large and bulky organic substituents, such as Wiberg's
Si(tBu)3 ligand[152] or Power's bulky terphenyl ligands[153]
that allow the preparation of otherwise unstable compounds.
Similarly, there is also a close relationship to the chemistry of
large and weakly coordinating cations, such as [NMe4]+,
[C5H6NMe6]+ (1,3,5-hexamethylpiperidinium), [S(NMe2)3]+,
and [P(NMe2)4]+.[154–156]
5.1. Gas-Phase Cations in Condensed Phases?
Many unusual and fundamentally important ions have
been detected in the gas phase by using one of the advanced
mass-spectrometric methods. Often these cationic species are
highly electrophilic or they are only weakly bound Lewis
acid–base complexes that in condensed phases react with the
counterions and/or solvents. Hence, it has often been
impossible to generate stable salts of these species in
condensed phases and to analyze their properties by classical
physical measurements, such as vibrational and NMR
spectroscopy or X-ray crystallography. WCAs can be used
to overcome this deficiency and enable such compounds to be
prepared in solution or the solid state. It is then possible to
verify experimentally the predictions of the sensitive quantum
chemical calculations of these gas-phase cations which often
reside in shallow minima on an extended potential energy
surface. An illustrative example of this approach is the
geometry of the [AgP8]+ ion which in 1995 was produced in
the gas phase and analyzed by MS.[157] Initial calculations
suggested a [Ag(h1-P4)2]+ to be the global minimum.[157]
However, recently it was shown by an X-ray crystal structure
that the global minimum is best described as edge-bound
[Ag(h2-P4)2]+ (see Figure 7).[56] This example clearly shows
that conclusions drawn from gas-phase studies are delicate
and, if possible, should be validated in the condensed phase.
Figure 7. Postulated gas-phase structure (left)[157] and observed solidstate structure (right) of the [Ag(P4)2]+ ion in [Ag(P4)2]+[Al(OR)4]
(R = C(CF3)3).[56]
5.1.1. “Pseudo Gas-Phase Conditions” in Condensed Phases
To stabilize a given gas-phase cation in condensed phases
a counterion and, usually, a solvent have to be introduced. In
2078
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
the solid state the ion–counterion pairs form an ionic lattice in
which the gas-phase cation is subjected to an electrostatic
field absent in the gas phase (= lattice energy). Since lattice
energies are inversely proportional to the sum of the ionic
radii (or volumes) of the constituting ions,[158] the lattice
energies of the WCA salts are very low (Table 2). In fact they
are so low that the lattice energy of the WCA salt
[Ag(S8)2]+[Al(OR)4] (1 in Table 2) of 326 kJ mol1
approaches the values of sublimation enthalpies of molecular
solids of comparable atomic weight, that is, that of C60 or C70
with 175 and 200 kJ mol1, respectively (cf. Mr = 1588 (1), 721
(C60), 841 g mol1 (C70)).[58, 159]
Table 2: Thermochemical volume (Vtherm) and lattice enthalpies of
several M+X salts.
Salt
Li+F
Cs+F
Cs+[AsF6]
Cs+[Al(OR)4] [c]
[Ag(S8)2]+[Al(OR)4] (1)[c]
Vtherm [I3]
27
43
128
776
1169
Upot. [kJ mol1]
1036[a]
740[a]
568[b]
362[b]
326[b]
[a] Experimental value.[159] [b] calculated from the thermochemical volumes.[158] [c] R = C(CF3)3.
Comparing the lattice energy of 1 (326 kJ mol1) to the
lattice energies of typical salts, such as Li+F (1036 kJ mol1)
and Cs+F (740 kJ mol1) clearly shows that the environment
of the ions in [Ag(S8)2]+[Al(OR)4] closer resembles the
situation in the gas phase (or a molecular solid) than to the
strong electrostatic field within a classical salt. Large WCAs
have diameters in the nanometer scale (1.25 nm for
[Al(OR)4] [47] or 1.20 nm for [Sb(OTeF5)6] [64]) and thus the
anions and cations are considerably separated which effectively diminishes coulombic interactions. Owing to the highly
fluorinated surface of most WCAs, dispersive interactions are
weak and not structure determining. Therefore, the environment of the cations within the framework of an ensemble of
large and very weakly coordinating anions, such as
[Al(OR)4] or [Sb(OTeF5)6] , can be described as providing
“pseudo gas-phase conditions” in the solid state.[58]
In solution pseudo gas-phase conditions are also present:
Salts with smaller anions are usually only soluble in polar
media with high dielectric constants, such as ethanol (erel. =
24), CH3CN (erel. = 35), or even strong acids, such as
anhydrous HF (erel. = 83). In contrast WCA salts are generally
very soluble which allows the use of very nonpolar solvents
with low dielectric constants erel. = 2 to 9 (toluene to CH2Cl2).
In these less polar media, the solvation energies that stabilize
the dissolved ions with respect to the gas phase are greatly
reduced. From a plot of the Gibbs solvation energy versus the
dielectric constant (erel.) of the solvent (Figure 8),[160] it can be
seen that the effect of decreasing the dielectric constant of the
solvent from HF (erel. = 83) to ethanol (erel. = 24) is much
smaller than for changing from ethanol to CH2Cl2 (erel. = 9) or
even toluene (erel. = 2). Thus, in low polarity or nonpolar
solvents the ions receive only a minimal stabilization through
the solvation energy. Therefore by using WCAs in combina-
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
enthalpies of A+X were approximated by using the Passmore–Jenkins equation.[158] In Figure 10 the lattice enthalpy
of A+X and the sum of the Gibbs solvation energies of A+
and X are plotted against the thermochemical volume of the
salt A+X .[161] From Figure 10 it can be seen that the solvation
and lattice enthalpies decrease rapidly with the increasing size
Figure 8. Plot of the Gibbs solvation energy DGsolv. calculated with the
Born equation for an univalent ion of radius 200 pm versus the dielectric constant erel. of the solvent.[160]
tion with low-dielectric media, gas-phase conditions (with no
stabilization by solvation) are also approximated in solution.
By the combination of WCAs and low dielectric media gasphase conditions are approximated both in solution and the
solid state. To describe this property the term psuedo gasphase conditions was coined.[58]
5.2. The Effect of Ion Size on Lattice and Solvation Enthalpies
The high solubility of WCA salts (Section 5.1) clearly
results from the decreased lattice potential enthalpies of these
salts which incorporate large counterions with diameters
exceeding one nanometer. Generally a salt is soluble if the
sum of the Gibbs free energies of solvation for the cation A+
and the anion X is larger than the lattice enthalpy of the salt
A+X (see Figure 9; entropic contributions play practically no
role in the solid state, which is why the lattice enthalpy and the
Gibbs lattice energy are almost identical).
Figure 9. Born–Fajans–Harber cycle for a soluble A+X salt.
If the WCA X is very large, then the lattice enthalpy is
low, the sum of the Gibbs solvation energies is then larger
than the lattice enthalpy, and hence the WCA salt is highly
soluble. For example: to determine the effect of the ion size
upon dissolution of a salt A+X in a nonpolar solvent a
simulation was performed in which the ionic radius of A+ was
kept constant at roughly that of a K+ ion, and the ionic radius
of X was changed from very small (halide) to very large
(large WCA). The solvation enthalpies for A+X in three
nonpolar solvents, with dielectric constants of 2.2, 4.8, and 8.9
(similar to toluene, CHCl3, and CH2Cl2), were approximated
with the Born equation.[160] In the next step the lattice
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Figure 10. Plot of the lattice enthalpy versus the solvation Gibbs free
energies for a A+X salt. For the calculataion the ionic radius of A+
was kept constant at roughly that of a K+ ion, and the ionic radius of
X was changed from very small (halide) to very large (large WCA).
The solvation enthalpies for A+X in three nonpolar solvents, with dielectric constants of 2.2, 4.8, and 8.9 (similar to toluene, CHCl3, and
CH2Cl2), were approximated.[160, 161]
of the anions. The shapes of the Gibbs solvation energy and
lattice potential enthalpy curves differ: With increasing size of
the anion the solvation energies relatively soon reach a
plateau region and remain almost constant while the lattice
energy decreases more steadily with increasing anion size.
The less polar the solvent the later the intersection of the
Gibbs solvation energies with the lattice enthalpy curve. This
intersection point indicates the change from insoluble to
soluble. Thus, only very large ions X with a volume of larger
than approximately 890 U3 induce a sufficiently low lattice
enthalpy to dissolve salts A+X in toluene (dashed vertical
line in Figure 10). In more polar solvents, such as CHCl3 or
CH2Cl2, the minimum size of the anion can be smaller, that is
120 U3 or 60 U3, respectively. In agreement with this result
K[Al(OR)4] (thermochemical volume Vtherm. = 762 U3) is
slightly soluble in toluene whereas K[BF4] (Vtherm. = 77 U3) is
only sparingly soluble in CH2Cl2. Of course these values
should not be taken as absolute because an exact evaluation
of solvation effects, dispersion energies, temperature effects
etc. has not been carried out. However, the general trend
holds: large WCAs lead to salts that are soluble in weakly
polar solvents, often even at very low temperatures.
5.3. Stabilization of Weakly Bound Complexes
Apart from being weakly basic and stable to oxidation,
large counterions provide another contribution to the stabilization of weak Lewis acid–base adducts [M(L)n]+[X] (M+ =
univalent cation, L = uncharged weakly bound ligand, X =
WCA): the reduced gain in M+[X] lattice energy upon
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2079
Reviews
I. Krossing and I. Raabe
dissociation of [M(L)n]+[X] into M+[X] and n L (see
Scheme 2 for a Born–Fajans–Haber cycle of the dissociation
of an [Ag(L)2]+[X] species). Considering the cases where
L = S8 and X = [AsF6] or [Al(OR)4] (R = C(CF3)3),[58, 162] the
Scheme 2. Born–Fajans–Haber cycle for the dissociation of a solid
Lewis acid–base adduct [Ag(L)2]+[X] into Ag+[X] and 2 L.
calculated lattice enthalpies for [Ag(S8)2]+[X] (X =
[AsF6] , [Al{OC(CF3)3}4]) are 393 and 327 kJ mol1, those
of the ligand-free salt Ag+[X] are 586 and 361 kJ mol1
(thermochemical volumes).[58, 158] The gas-phase enthalpy
DrH(gas) and the sublimation enthalpy of S8 are the same in
both cases, and therefore the only differences are due to the
lattice enthalpies. For [AsF6] the resulting gain in lattice
enthalpy upon dissociation is 193 kJ mol1 while that for
[Al{OC(CF3)3}4] is only 34 kJ mol1. Therefore [Ag(S8)2]+[Al{OC(CF3)3}4] is more stable by 19334 = 159 kJ mol1
against a dissociation than the [AsF6] salt. In other words,
with the [Al(OR)4] ion, [Ag(L)2]+ Lewis acid–base adducts
can be stabilized for very weak bases L, for which DrH(gas) of
the gas-phase term [Ag(L)2]+(g)![Ag+(g)] + 2 L(g) is by 100 to
160 kJ mol1 less endothermic than that with the smaller
[AsF6] ion.
If the standard state of the ligand L in Scheme 2 is a gas,
the contribution of the sublimation enthalpy DHsubl.(L), which
enhances decomposition, is eliminated. For the above example DHsubl.(S8) = 101 kJ mol1, (for two molecules of S8
202 kJ mol1). If a gas, for example, CO is used as a ligand,
this destabilizing contribution is eliminated. Therefore, it is
possible to prepare and characterize [Ag(CO)2]+ salts,
although the gas-phase dissociation enthalpy DrH(gas) of
[Ag(L)2]+(g)!Ag+(g) + 2 L(g) is only 197 kJ mol1 for L = CO
(experimental value),[163] but 363 kJ mol1 for L = S8 (ab inito
calculation, MP2/TZVPP).[58] DDrH(gas) amounts to
363197 = 166 kJ mol1 which is less than twice the sublimation enthalpy of L = S8 (202 kJ mol1). This leads to the
unusual conclusion that (neglecting entropy) solid
[Ag(S8)2]+[Al{OC(CF3)3}4] is approximately 30 to
40 kJ mol1 less stable towards a dissociation into solid
Ag[Al{OC(CF3)3}4] and solid S8 than (hypothetical)
[Ag(CO)2]+[Al{OC(CF3)3}4] towards decomposition into
solid Ag[Al{OC(CF3)3}4] and gaseous CO. In agreement
with this notion, in 1994 a [Ag(CO)2]+ salt with a, compared
to [Al{OC(CF3)3}4] , smaller [B(OTeF5)4] ion was prepared.[62, 164]
Another example underlines this conclusion: We recently
prepared the [Ag(h2-C2H4)3]+ ion as the [Al{OC(CF3)3}4] salt
(Figure 11).[55] The Ag+ complexes of C2H4 are of fundamental interest as they are seen as the prototypes for transitionmetal–olefin complexes without backbonding. Since they are
2080
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 11. Section of the solid-state structure of [Ag(h2-C2H4)3]+[Al{OC(CF3)3}4] .[55]
difficult to stabilize in condensed phases, they were intensively investigated by theoretical and mass spectrometry (MS)
studies.[165, 166] However, despite many earlier MS experiments,
only [Ag(h2-C2H4)x]+ complexes with x = 1, 2 are known in the
gas phase.[167] The formation and structural characterization
of a [Ag(h2-C2H4)3]+[Al{OC(CF3)3}4] salt (that is, x = 3) was
therefore surprising and shows that the environment provided
by the solid-state arrangement of the [Al{OC(CF3)3}4] ions is
very close to that of the gas phase in thermal equilibrium at
low temperature. The gas-phase addition of a third molecule
of gaseous C2H4 to gaseous [Ag(h2-C2H4)2]+ was impossible in
the mass spectrometer, presumably because of the low Gibbs
binding energy of the third ligand (DG8 = 55 kJ mol1).[55, 166]
This result points to the failure to relax the energy stored in
translational, vibrational, and rotational levels, which in sum
can be larger than the binding energy of the third ligand, thus
preventing coordination. In WCA salts this internal energy
can be removed through intermolecular vibrational coupling
thus allowing equilibrium conditions to be reached at a given
temperature. In this respect the pseudo gas-phase conditions
provided by the best WCAs in the solid state are even better
than the gas phase inside a mass spectrometer where thermal
equilibrium conditions are difficult to reach. In this respect
good WCAs, such as the [Al{OC(CF3)3}4] ion may justly be
called noncoordinating.
5.4. Stabilization of Lower Charged Species: Coulomb Explosion
of Multicharged Cations
Polyatomic, multicharged cations (and anions) are unstable in the gas phase and dissociate into smaller, less charged
species. Thus the simple O22 ion known from many salts in
the solid state is unknown in (equilibrium) gas-phase
chemistry where it is unstable with respect to two O ions.
The driving force for this dissociation is the strong electrostatic repulsion of the two adjacent negative charges in the
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
O22 ion which is relieved upon dissociation (the “Coulomb
explosion”). In the solid state the multicharged polyatomic
ions are stabilized against dissociation into singly charged ion
salts by having much higher lattice energies than singly
charged ion salts.[168–170] This effect is shown by the Kapustiniskii equation for the lattice energy Upot [Eq. (28)] where za
Upot ¼
121:4 za zb n
0:0345
1
kJ mol1
ðra þ rb Þ
ðra þ rb Þ
ð28Þ
and zb are the charge of the ions, ra and rb the ionic radii, and n
is the number of ions per forumla unit (i.e. two for A+B ,
three for A2+[(B)2] etc.).
Thus, with similar ionic radii, the lattice energy of a
A2+[(B)2] or a A3+[(B)3] salt is about three or six times,
respectively that of a A+B salt. Therefore, salts of higher
charged ions, which are unstable in the gas phase, are lattice
stabilized in the solid state.[168–170] Through the use of large
WCAs the singly charged cations can be favored over the
multiply charged cations. Two borderline cases will be
presented in the following subsections.
5.4.2. The Reaction E42+!2 E2+ (E = O, S)
The highly p-bonded gaseous O2+ and S2+ ions have short
E–E separations of 112 and 183 pm, respectively. In the solid
state, singly charged O2+ salts with an average oxidation state
of oxygen of + 0.5 have been known for a long time with a
variety of fluorometallate anions. In contrast, all sulfur
compounds with an average oxidation state of + 0.5 (several
with fluorometallate anions are known) contain the doubly
charged, formal dimer S42+ = (S2+)2 [Eq. (29)].[168]
Why is it that no O42+ salt forms in the solid state? This
can be explained with the Born–Fajans–Haber cycle shown in
Scheme 4. The gas phase contribution is dominant for E = O,
5.4.1. The Reaction [S3N2]2+![SN]+ + [S2N]+
Reaction of S4N4 with AsF5 yields the 6 p HVckel aromatic
[S3N2]2+ as the [AsF6] salt.[171] In the gas phase [S3N2]2+ is
unstable towards disproportionation into [SN]+ and [S2N]+ by
400 kJ mol1 (Scheme 3).[171] An analysis of a suitable Born–
Scheme 3. Coulomb explosion of gaseous [S3N2]2+ giving [SN]+ and
[SNS]+. The solid [S3N2]2+[AsF6]2 is lattice stabilized, but the dissociation products [SN]+[Sb2F11] and [S2N]+[Sb2F11] were obtained with the
larger [Sb2F11] ion.[171]
Fajans–Haber cycle showed[171] [S3N2]2+[(AsF6)2] to be stable
against a disproportionation into [SN]+[AsF6] and
[S2N]+[AsF6] by 28 kJ mol1. By reducing the overall lattice
energies by replacing [AsF6] (Vtherm. = 105 U3) with a larger
ion, such as [Sb2F11] (Vtherm. = 200 U3) only the singly charged
dissociation products [SN]+[Sb2F11] and [S2N]+[Sb2F11] are
formed (Scheme 3; DrH(s) is 10 kJ mol1 in favor of the
A+B salts).[171]
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Scheme 4. Born–Fajans–Haber Cycle rationalizing the formation of
solid O2+A but S42+(A)2 (A = AsF6).
and owing to the short (calculated) OO bond in O42+ of
134 pm[168] the Coulomb repulsion is large and DHdiss(g) is
exothermic by 1035 kJ mol1.[168] Although the lattice
enthalpy of the A2+[(B)2] salt O42+[A]2 is roughly three
times larger than that of the A+B salt, the formation of two
units of solid O2+[AsF6] is favored by 615 kJ mol1 over
O42+[AsF6]2 (Scheme 4).[168] Owing to the larger S–S separation in S42+ (201 pm) and the stronger SS single bonds[172]
DHdiss(g) is exothermic[173] by only 258 kJ mol1. Therefore in
the solid state the higher lattice enthalpy contribution of the
A2+[(B)2] salt dominates and the formation of S42+[(AsF6)2]
is favored by 238 kJ mol1. All attempts to lower the overall
lattice enthalpies by replacing [AsF6] (Vtherm. = 105 U3) for
[Sb(OTeF5)6] (Vtherm. = 725 U3)[65] and thus stabilizing an S2+
salt analogous to the O2+ salt in the solid state failed. A
thermochemical analysis showed that a WCA that would
stabilize a S2+ salt would have to have a thermochemical
volume of at least 5000 U3.[174] However, such extremely large
WCAs are currently unknown.
6. Limits of Weakly Coordinating Anions
After showing the amazing properties of WCAs in
stabilizing salts of unusual and reactive cations it is important
to also give examples of the limits of WCAs in various
applications. Typical limits are anion coordination and anion
decomposition. Selected examples of these limits are detailed
below, others were mentioned in Section 2.
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2081
Reviews
I. Krossing and I. Raabe
6.1. Coordination
Virtually all of the WCAs presented in Section 2 can
coordinate given the right cation. Most prominent examples
are the highly reactive cations [SiR3]+ and [AlR2]+ (R = Me,
Et, iPr) that decompose almost all WCAs. The exceptions are
the halogenated carboranes with which they form molecular
coordination compounds of the type R3Sid+-CBd and R2Ald+CBd (Figure 12). Based on their Si–anion or Al–anion
Figure 12. Molecular “ionlike” coordination compounds of type
R3Sid+ CBd and R2Ald+ CBd (CB = halogenated carborane-based ion).
separations these compounds are clearly molecules;
however some of the properties of the R3Sid+-CBd
and R2Ald+-CBd species suggested that in solution
the character of the parent cations [SiR3]+ and
[AlR2]+ is partially retained and, therefore, species,
such as R3Sid+-CBd and R2Ald+-CBd are also called
“ionlike”.
The more widely used ion [B(ArF)4] (ArF =
C6H3(CF3)2) is susceptible to h3-, h4-, and even h6coordination of the ArF ligand (Figure 13).[175] Studies
show that the weakly coordinating nature of the
[B(ArF)4] ion is primarily due to steric effects. If the
countercations, such as RhI or AgI, are small and
electrophilic enough to fit between the aromatic rings,
coordination is possible and led with RhI to the rapid
irreversible decomposition of the [B(ArF)4] ion
within hours.[175]
A related anion known to be somewhat more coordinating is [MeB(C6F5)3] which often coordinates to the cation
center through the methyl group but additionally may also
form cation–F contacts, such as in [{Cp’2Y}d+{(mCH3)B(C6F5)3}]d (Cp’ = C5H4SiMe3) (Figure 14).[176] Also
the [Sb2F11] ion can be coordinated by a suitable electrophilic
cation, such as Au2+. Thus, upon evacuation of [AuXe4]2+[(Sb2F11)2] the xenon poorer compound [Au(Xe)2(Sb2F11)2] containing coordinated [Sb2F11] ions forms
(Figure 15).
Figure 14. Anion coordination in the [{Cp’2Y}d+{(m-CH3)B(C6F5)3}d]
complex.
Figure 15. Comparison of salt formation in [AuXe4]2+[Sb2F11]2 and ion coordination in [Au(Xe)2(Sb2F11)2].
6.2. Decomposition
Anion coordination as described in Section 6.1 can be the
starting point for the degradation of the WCA provided that
the (possibly only transient) countercation is reactive enough
to induce the decomposition. Often decomposition can be
slowed down or even hindered by working at low temperatures. In general, solid compounds are less sensitive towards
decomposition than WCA salts in solution. Selected typical
examples for WCA degradation are presented below.
6.2.1. Ligand Abstraction without Degradation of the Ligand
Figure 13. h6-Coordination of the ArF ligand in [(cod)Rh(h6-ArFB(ArF)3)]
(COD = cylcooctadiene).
2082
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The dissolved [Me(Ph3P)2Pt(OEt2)]+[B(ArF)4] salt—generated from [(Ph3P)2PtMe2] and [H(OEt2)2]+[B(ArF)4]—is
only stable at temperatures below 30 8C and when allowed
to reach room temperature the ion slowly (days) degrades and
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
serves as a source for the ArF ligand [Eq. (30); ArF =
C6H3(CF3)2].[177]
The decomposition of Ag+[B(OTeF5)4] is slow but that of
the [SiPh3]+ species is so fast that the silylium ion was
postulated as only a short lived intermediate.[62] Electrophilic
zirconocene cations degrade the [Nb(OC6F5)6] WCA. The
pathways in Scheme 7 were described for [Cp*2ZrMe]+[Nb(OC6F5)6] .[52b]
Similarly the [B(C6F5)4] ion rapidly degrades in the
presence of [AlR2]+,[15] [RZn]+ [178] (R = Me, Et) and H+
cations if no donor solvent, such as diethyl ether is present
(Scheme 5) Also sterically open group 14 metallocene cations
decompose the [B(C6F5)4] ion.[3]
Scheme 7. Pathway for the decomposition of [Cp*2ZrMe]+[Nb(OC6F5)6] .
6.2.2. Ligand Abstraction with Degradation of the Ligand
Scheme 5. Decomposition pathway of the [B(C6F5)4] ions in the presence of diethyl zinc.
An
interesting
approach
to
generate
the
[(F5C6)Xe]+[F2B(C6F5)2] salt is anion decomposition starting
from the B(C6F5)3 Lewis acid and XeF2. The intermediate
[FXe]+[FB(C6F5)3] is generated which degrades with formation of (unstable) (F5C6)XeF and FB(C6F5)2 that finally react
to give [(F5C6)Xe]+[F2B(C6F5)2] as a clean product stable
below 14 8C (Scheme 6).[179]
Scheme 6. Synthesis of [(F5C6)Xe]+[F2B(C6F5)2] through the planned
initiation of an anion decomposition.
Also the teflate based WCAs decompose in the presence
of strong electrophiles. Thus the [B(OTeF5)4] ion looses a
OTeF5 group in the presence of electrophiles, such as Ag+ [62]
or [SiPh3]+ [62] [Eq. (31), E = Ag+, [SiPh3]+].
Eþ ½BðOTeF5 Þ4 ! EðOTeF5 Þ þ BðOTeF5 Þ3
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
ð31Þ
A second rarer type of WCA degradation is ligand
abstraction followed by decomposition of the ligand or direct
decomposition without ligand abstraction. The [B(C6F5)4]
ion decomposes in the presence of [Cp2ZrMe]+ to give the
dimeric, fluoride bridged [Cp2(F5C6)Zr-F-Zr(C6F5)Cp2]+ ion
that crystallizes from the mixture with a remaining intact
[B(C6F5)4] ion. In this case, C6F5 abstraction is combined with
a CF bond activation [Eq. (32)].
The formation of fluorine bridges was also observed
during the decomposition of the dissolved [Al(OR)4] ion
(R = C(CF3)3) in the presence of very reactive phosphorus or
phosphorus–halogen cations E+ at temperatures between 10
and 30 8C.[59, 180] In contrast to the solutions, solid samples of
phosphorus–halogen cation salts are stable for days or even
months at ambient temperatures.[59] The decomposition
reaction of the [Al(OR)4] ion presumably starts with ligand
abstraction and the formation of the very strong Lewis acid
Al(OR)3. According to the information in Table 3 (Section 7),
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2083
Reviews
I. Krossing and I. Raabe
Al(OR)3 is a stronger Lewis acid than SbF5 and, consequently,
attacks another [Al(OR)4] ion by abstracting of a fluoride
ion of one of the 12 equivalent CF3 groups giving [FAl(OR)3]
and (C4F8O)Al(OR)3. The epoxide OC4F8 is then replaced by
[FAl(OR)3] in a nucleophilic substitution reaction leading to
the formation of the [(RO)3Al-F-Al(OR)3] ion (Scheme 8).
regenerated AsF5 then reenters this cycle, produces new CF2
which inserts into the CF bond of the CF2 group so that in
the end borates bearing the C(F)(CF3)2 moiety are isolated.
The WCA decomposition example given in Equation (33)
shows that the ligand may not only be extracted as such, but, if
½NbðOTeF5 Þ6 þ CH3 CN !
½O¼NbðOTeF5 Þ4 ðNCCH3 Þ þ OðTeF5 Þ2
ð33Þ
the metal is able to form multiple bonds, may leave as a
volatile molecule. Thus, [Nb(OTeF5)6] decomposes in acetonitrile with formation of [O=Nb(OTeF5)4(NCCH3)] and
gaseous O(TeF5)2.[64]
Scheme 8. Formation of the fluoride-bridged [(RO)3Al-F-Al(OR)3] ion
via an epoxide intermediate.
According to an analysis of the structural parameters[59, 180] as
well as to DFT-calculations[184] (see Table 3), the fluoride
bridged [(RO)3Al-F-Al(OR)3] ion (R = C(CF3)3) is chemically more robust then the already very stable [Al(OR)4] ion.
With a surface made up of 54 CF bonds and its O atoms well
protected by the six R ligands, the fluoride bridged [(RO)3AlF-Al(OR)3] ion is certainly the best candidate for the most
weakly coordinating ion known to date. All other known
WCAs contain at least 18 peripheral CF bonds less.
In a related pathway, the [B(CF3)4] ion also decomposes
in the presence of strong Lewis acids, such as AsF5
(Scheme 9). AsF5 abstracts a fluoride ion from a CF3 group,
this results in the formation of a difluorocarbene CF2
coordinated to a B(CF3)3 Lewis acid. The carbene then
inserts into the CF bond of another CF3 group giving
(F5C2)B(CF3)2, which in turn abstracts a fluoride ion from
[AsF6] leading to [(F5C2)(F)B(CF3)2] and AsF5 (the FIA of
B(CF3)3 is much higher than that of AsF5, see Table 3). The
7. Comparison of the Properties of different Classes
of Weakly Coordinating Anions Based on
Quantum Chemical Calculations
To compare the properties of the very different types of
WCAs, such as the fluoroantimonates with the perfluoroarylborates, a computational approach was chosen. The structures of all the WCAs of the type [M(L)n] (L = monoanionic
ligand), their parent Lewis acids A = M(L)n1, as well as their
fluoride adducts AF = [FM(L)n1] were optimized with
DFT methods at the BP86/SV(P) level. With these calculated
data, the thermodynamic stability and coordinating ability of
the WCAs was established based on the following considerations:
1) All [M(L)n] ions that are based on a Lewis acidic central
atom M are prone to decomposition by ligand abstraction.
A measure for the intrinsic stability of a given WCA is the
Lewis acidity of the parent Lewis acid A (e.g. the B(C6F5)3
acid for the [B(C6F5)4] ion). An established measure for
Lewis acidity is the fluoride ion affinity FIA (see
Section 2.5.1) which is calculated from an isodesmic
reaction.[81, 181] FIA values of the parent Lewis acids A
are included in Table 3. The higher the FIA value, the
more stable the WCA is against ligand abstraction.
2) Additionally we also directly assessed the ligand affinity
LA of all types of WCAs through an isodesmic reaction.[182] The LA is the enthalpy of reaction necessary to
remove the ionic ligand L from the ion [M(L)n]
[Eq. (34)]. The LA is always endothermic and the more
DH¼LA
½MðLÞn ƒƒƒ
ƒ!½MðLÞn1 þ L
Scheme 9. Generation of the F2C carbene ligand and insertion into the
CF bond of fluorated borates.
2084
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ð34Þ
positive the LA value (see Table 3) the more stable the
WCA is against ligand abstraction. However, a word of
caution is needed here: The LA also reflects the stability
of the generated L ion. Thus, if L is stable (e.g. L =
[OC(CF3)3] or [OTeF5]) the LA is relatively low
compared to less stable ionic ligands (such as L =
[C6H5] or [C6H3(CF3)2]).
3) To assess the stability of a WCA towards attack of a hard
or soft electrophile and to eliminate the contribution of
the intrinsic stability of L in (2) the isodesmic decomposition reactions of [M(L)n] with H+ [hard, Eq. (35);
PD = enthalpy of proton-induced decomposition] and
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
Cu+ [soft, Eq. (36); CuD = enthalpy of copper-induced
decomposition] were calculated.
DH¼PD
½MðLÞn þ Hþ ƒƒƒ
ƒ!½MðLÞn1 þ HL
ð35Þ
DH¼CuD
½MðLÞn þ Cuþ ƒƒƒƒ
ƒ!½MðLÞn1 þ CuL
ð36Þ
PD and CuD allow conclusions to be drawn on the
stability of a given WCA of type [M(L)n] upon reaction
with a hard (H+, PD) or soft (Cu+, CuD) electrophile.
Since in Equations (35) and (36) a gaseous ion and a
gaseous cation react giving two neutral species, PD and
CuD are both exothermic. The less negative the PD and
CuD values are (see Table 3), the more stable the WCA is
against electrophilic attack. For gas-phase acidities of
neutral Brønsted acids including H[CB11F12] see ref. [183].
4) The energy of the HOMO of a WCA relates to its
resistance towards oxidation. The lower the HOMO
energy, the more difficult it is to remove an electron and
thus oxidize the WCA.
5) The HOMO–LUMO gap (see Table 3) can be associated
with the resistance of a WCA towards reduction and the
larger the HOMO–LUMO gap, the more stable the anion
is with respect to reduction. Very small gaps, such as those
for [Sb4F21] or [As(OTeF5)6] , are an indication of the
potentially oxidizing character of these anions, which may
interfer with countercations sensitive towards oxidation.
6) A measure of the coordination ability of an anion is the
partial charge of the most negatively charged atom (qneg.),
or, the most negatively charged atom on the surface of the
anion (qsurf. ; see Table 3). It is clear that low partial
charges are an indication for low coordination ability.
However, steric effects may also be of importance, and the
most basic atoms may be hidden in the center of a large
WCA and are, therefore, not available for coordination.
In these cases the charge of the most basic, accessible
surface atom qsurf. is a better measure.
Ligand abstraction and hydrolysis are frequently observed
decomposition pathways for WCAs (see Section 6.2) and,
therefore, the computational approach to calculate LA, PD,
and CuD mimics experimental observations. However, by
calculations, only the underlying thermodynamics can be
assessed, kinetic barriers against decomposition may additionally stabilize a given WCA. The data included in Table 3
may not be taken as absolute, however, since the same
methods were used for all computations, relative trends will
definitely be correct. For the carborane-based anions the FIA,
LA, PD, and CuD cannot directly be assessed and, therefore,
these WCAs were excluded from this approach. Full details of
all calculations will be disclosed in an upcoming full paper.[184]
From Table 3 emerges the outstanding capability of the
[Sb3F16] , [Sb4F21] , and [Sb(OTeF5)6] WCAs to stabilize
highly oxidizing cations even in anhydrous HF solution (see
FIA, PD, HOMO). This stability against oxidation has to be
traded in for sensitiveness against reduction (see HOMO–
LUMO gap) and moisture which greatly reduces the use of
these anions. In terms of coordinating ability, these WCAs are
more coordinating than others (compare qneg. and qsurf.). For
the borate-based ions it can be seen that fluorination greatly
Table 3: Calculated properties of WCAs.
Anion
Symmetry FIA of the Lewis
(anion)
Acid [kJ mol1]
LA [kJ mol1] PD [kJ mol1] CuD [kJ mol1] HOMO [eV] HOMO–
LUMO
gap [eV]
qneg.[a]
qsurf.[a]
[BF4]
[PF6]
[AsF6]
[SbF6]
[Sb2F11] vs. Sb2F10
[Sb3F16] vs. Sb3F15
[Sb4F21] vs. Sb4F20
[B(OTeF5)4]
[As(OTeF5)6]
[Sb(OTeF5)6]
[Al{OC(CF3)3}4]
[(RO)3Al-FAl(OR)3] [c]
[B(C6H5)4]
[B(C6H3(CF3)2)4]
[B(C6F5)4]
[B(CF3)4]
[F4C6{1,2B(C6F5)2}2F] [d]
[F4C6{1,2B(C6F5)2}2OMe] [e]
[F4C6{1,2B(C6F5)2}2OMe] [d]
Td
Oh
Oh
Oh
C1
Ci
C2v
C1
C3
C3
S4
Ci
338
394
426
489
549
582
584
550
593
633
537
685[c]
[b]
274
290
341
342
441
S4
S4
S4
T
C1
342
471
444
552
510
324
382
296
490
328[d]
1402
1251
1256
1136
1224[d]
C1
510
C1
–
1212
1156
1124
1061
1026
994
991
1040
1023
973
1081
983
521
465
433
371
336
303
301
420
403
353
395
297
1.799
2.672
3.149
3.911
5.540
6.342
6.579
5.811
6.335
6.610
4.100
4.987
10.820
8.802
6.284
5.135
4.336
3.886
3.256
2.593
2.204
2.326
6.747
6.500
0.25 (F)
0.44 (F)
0.44 (F)
0.44 (F)
0.40 (F)
0.38 (F)
0.39 (F)
0.40 (F)
0.62 (O)
0.61 (O)
0.24 (O)
0.23 (O)
0.25 (F)
0.44 (F)
0.44 (F)
0.44 (F)
0.40 (F)
0.38 (F)
0.39 (F)
0.40 (F)
0.40 (F)
0.39 (F)
0.20 (F)
0.20 (F)
649
506
538
379
506[d]
2.150
3.789
3.130
3.530
3.274
4.087
3.816
4.196
9.158
3.861
0.45 (B)
0.44 (B)
0.21 (F)
0.58 (B)
0.54 (B)
0.05 (H)
0.22 (F)
0.21 (F)
0.21 (F)
0.22 (F)
586[e]
1061[e] 332[e]
3.101
3.754
0.68 (B) 0.22 (C)
305[d]
1247[d] 529[d]
[b]
[b]
[b]
[b]
[b]
[b]
–
–
–
–
[a] qneg. = The partial charge of the most negatively charged atom; qsurf. the partial charge of the most negatively charged surface atom. [b] LA and FIA
are identical. [c] FIA versus 2 Al(OR)3 ; R = C(CF3)3. [d] Against [C6F5] abstraction (LA) or EC6F5 abstraction (E = H: PD, E = Cu: CuD). [e] Against
OMe abstraction (LA) or EOMe abstraction (E = H: PD, E = Cu: CuD).
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2085
Reviews
I. Krossing and I. Raabe
increases the thermodynamic stability of the anion (see FIA,
LA, PD, CuD, HOMO, HOMO–LUMO gap). The differences between the commercially available WCAs, [B(C6F5)4]
and [B(C6H3(CF3)2)4] , are small, however, diborane-based
ions, such as [F4C6{1,2-B(C6F5)2}2X] (X = F, OMe), are far
more stable. Of all borates the novel [B(CF3)4] ion is the
best. However, the experimentally found decomposition
shown in Scheme 3 was not investigated by this approach.
The stability of the perfluoroalkoxyaluminate [Al(OR)4]
(R = C(CF3)3) with respect to FIA, PD, CuD, HOMO, and
HOMO–LUMO gap is remarkable and higher than that of all
the borates except the [B(CF3)4] ion which has comparable
values. The [Al(OR)4] ion comes even close to the oxidation
resistance and low PDs of the fluoroantimonates and is, in
part, even better than the teflate-based ions. In contrast to the
latter two types of anions, the synthesis of the perfluoroalkoxyaluminate is straight forward, which shows the great
potential of this special type of ion for chemistry. An even
better choice would be the fluoride bridged [(RO)3Al-FAl(OR)3] ion which is among the best WCAs according to
the data in Table 3, but this WCA is only known as a
decomposition product. However, we recently developed a
direct synthesis of the silver salt of this WCA.[184]
simple answer can be given to the question: “Which ion is the
best?”. The full answer will always be found in a balance of
steric and electronic effects of the cation and the anion as well
as anion-stability considerations. However, with the multitude of WCAs known and the guidelines given in the
preceding sections a suitable noncoordinating counterion
that lies below the coordination threshold of the cation in
question should be found in most of the cases.
WCA chemistry is an area of increasing scientific investigation and it can be anticipated that over time suitable ions
will be prepared that allow even the most difficult tasks to be
overcome, such as the preparation of a free and true
alumenium cation [AlR2]+. By using bulky terphenyl ligands
Ar and a dimeric [Li{Al(OR)4}2] ion (R = C(H)(CF3)2) the
heavier homologue, a linear [Ar-Ga+-Ar] cation, was prepared in 2003.[185] However, it remains unlikely that true salts
of the ultimate electrophile, the proton H+, can be stabilized.
The broad interest of applied chemistry (such as homogenous
catalysis, electrolytes for lithium-ion batteries, electrochemistry, ionic liquids) in the development of new, easily
accessible, and cheap WCAs will further fertilize this area
of research and new exciting developments can be expected.
List of Abbreviations
8. Conclusion and Outlook
It is time to answer the question raised in the title:
“Noncoordinating Anions—Fact or Fiction?” From the
preceeding paragraphs it is clear that one has to distinguish
between two fundamental properties of WCAs:
1) the chemical robustness of a WCA towards electrophiles
and oxidizing agents.
2) The coordinating ability of the WCA.
The most chemically robust anions known to date, the
halogenated carborane anions, are more strongly coordinating than other WCAs and even less electrophilic cations, such
as Ag+, are always coordinated by these ions. But owing to
their exceptional stability, they allow the preparation of
ionlike compounds of very electrophilic cations, such as
[SiR3]+ and [AlR2]+ (R = small alkyl group), cations that
decompose all other currently known WCAs. Thus, when a
maximum of WCA stability towards electrophiles is desired,
the halogenated carborane-based WCAs are certainly the
best choice. Of these, the fluorinated derivatives [1RCB11F11] (R = Me, Et) are the least coordinating. However,
if the synthetic cation target is of a more normal electrophilicity, compatible with other WCAs, and consists of a very
weakly bound complex which is easily prone to dissociation,
larger, less coordinating but also less robust ions, such as
[Al(OR)4] (R = C(CF3)3) are probably a better choice; for
example, see the discussion of the [Ag(h2-C2H4)3]+[Al(OR)4]
(R = C(CF3)3) salt in Section 5.3.[55]
In conclusion, and answering the title question, it can be
stated that noncoordinating ions are available as long as the
electrophilicity of the cations is below the coordination
threshold that enforces coordination. Which of the variety
of known WCAs is the best and thus noncoordinating for a
special system has to be explored experimentally and no
2086
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
WCA
[CB]
ArF
RF
teflate
FIA
LA
PD
CuD
Weakly coordinating ion
Halogenated carborane-based anion
Fluorinated aryl ligand
Fluorinated alkyl ligand
[OTeF5]
Fluoride ion affinity
Ligand affinity
Enthalpy of proton-induced decomposition
Enthalpy of copper-induced decomposition
We thank Dr. H. J. Himmel, Prof. H. Schn6ckel, Prof. H.
Willner, Prof. C. A. Reed, Prof. J. Passmore, and Prof. S. H.
Strauss for stimulating discussions. Prof. H. Willner is
acknowledged for making results available prior to publication. I.K. and I.R. like to thank the DFG, the FCI and the
Alexander von Humboldt-Stiftung for financial support.
Received: July 24, 2003 [A620]
[1] W. Beck, K. H. SVnkel, Chem. Rev. 1988, 88, 1405.
[2] Reviews: a) C. Reed, Acc. Chem. Res. 1998, 31, 133; b) S. H.
Strauss, Chem. Rev. 1993, 93, 927.
[3] E. Y.-X. Chen, T. J. Marks, Chem. Rev. 2000, 100, 1391, and
references therein.
[4] a) K. Seppelt, Angew. Chem. 1993, 105, 1074; Angew. Chem.
Int. Ed. Engl. 1993, 32, 1025; b) M. Bochmann, Angew. Chem.
1992, 104, 1206; Angew. Chem. Int. Ed. Engl. 1992, 104, 1181.
[5] A. J. Lupinetti, S. H. Strauss, Chemtracts: Inorg. Chem. 1998,
11, 565.
[6] A. G. Massey, A. J. Park, J. Organomet. Chem. 1964, 2, 245.
[7] a) J. H. Golden, P. F. Mutolo, E. B. Lobrovski, F. J. DiSalvo,
Inorg. Chem. 1994, 33, 5374; b) K. Fujiki, S. Ikeda, H.
Kobayashi, A. Mori, A. Nagira, J. Nie, T. Sonoda, Y. Yagupolskii, Chem. Lett. 2000, 66.
[8] J. van den Broeke, B.-J. Deelman, G. van Koten, Tetrahedron
Lett. 2001, 42, 8085.
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
[9] K. Fujiki, J. Ichikawa, H. Kobayashi, A. Sonoda, T. Sonoda, J.
Fluorine Chem. 2000, 102, 293.
[10] [B{C6F4(4-CF3)}4] : F. A. R. Kaul, G. T. Puchta, H. Schneider,
M. Grosche, D. Mihalios, W. A. Herrmann, J. Organomet.
Chem. 2001, 621, 177.
[11] a) L. Jia, X. Jang, C. L. Stern, T. J. Marks, Organometallics
1997, 16, 842; b) L. Jia, X. Yang, A. Ishihara, T. J. Marks,
Organometallics 1995, 14, 3135.
[12] G. Rodriguez, P. Brant, Organometallics 2001, 20, 2417.
[13] Most of those were unpublished results for the T. J. Marks
group: See page 1398–1399 and references [78a,91,92] in
ref. [3].
[14] One exception: The crystallization of the cyclotrigermenium
salt [(GeSitBu3)3]+: A. Sekiguchi, N. Fukaya, M. Ichinobe, Y.
Ishida, Eur. J. Inorg. Chem. 2000, 1155.
[15] a) P. Biagini, G. Luigli, L. Abis, P. Andeussi (Enichem), US
Patent 5 602 269, 1997; b) M. Bochmann, M. J. Sarsfield,
Organometallics 1998, 17, 5908.
[16] K. Ren, A. Mejiritski, J. H. Malpert, D. C. Neckers, Tetrahedron
Lett. 2000, 41, 8669.
[17] K. Ren, J. H. Malpert, H. Li, H. Gu, D. C. Neckers, Macromolecules 2002, 35, 1632.
[18] a) J. Zhou, S. J. Lancaster, D. A. Walker, S. Beck, M. ThorntonPett, M. Bochmann, J. Am. Chem. Soc. 2001, 123, 223; b) S. J.
Lancaster, D. A. Walker, S. Beck, M. Thornton-Pett, M.
Bochmann, Chem. Commun. 1999, 1533; c) R. E. LaPointe,
WO 99/42467, 1999.
[19] D. Vagedes, G. Erker, R. FrXhlich, J. Organomet. Chem. 2002,
641, 148.
[20] R. E. LaPointe, G. R. Roof, K. A. Abboud, J. Klosin, J. Am.
Chem. Soc. 2000, 122, 9560.
[21] S. J. Lancaster, A. Rodriguez, A. Lara-Sanchez, M. D. Hannant, D. A. Walker, D. H. Hughes, M. Bochmann, Organometallics 2002, 21, 451.
[22] V. C. Williams, W. E. Piers, W. Clegg, S. Collins, T. B. Marder, J.
Am. Chem. Soc. 1999, 121, 3244.
[23] L. D. Henderson, W. E. Piers, G. J. Irvine, R. McDonald,
Organometallics 2002, 21, 340.
[24] M. Schmidt, A. KVhner, E. Bernhardt, H. Willner, EP
1205480A2, 2002.
[25] E. Bernhardt, G. Henkel, H. Willner, G. Pawelke, H. BVrger,
Chem. Eur. J. 2001, 7, 4696.
[26] R. D. Chambers, H. C. Clark, C. J. Willis, J. Am. Chem. Soc.
1960, 82, 5298.
[27] M. Schmidt, A. KVhner, K.-D. Franz, G.-V. RXschenthaler, DE
101 03 189A1, 2002.
[28] E. Bernhardt, M. Finze, C. W. Lehmann, F. Aubke, H. Willner
Angew. Chem. 2003, 115, 2123; Angew. Chem. Int. Ed. 2003, 42,
2071.
[29] D. J. Brauer, H. BVrger, Y. Chebude, G. Pawelke, Inorg. Chem.
1999, 38, 3972.
[30] For a recent review on [(CF3)3BY] compounds see: G.
Pawelke, H. BVrger, Coord. Chem. Rev. 2001, 215, 243.
[31] M. Finze, E. Bernhardt, A. Terheiden, M. Berkei, H. Willner, D.
Christen, H. Oberhammer, F. Aubke, J. Am. Chem. Soc. 2002,
124, 15 385.
[32] a) V. V. Bardin, H. J. Frohn, Main Group Met. Chem. 2002, 25,
589; b) H. J. Frohn, N. Y. Adonin, V. V. Bardin, V. F. Starichenko, J. Fluorine Chem. 2002, 117, 115; c) V. V. Bardin, S. G.
Idemskaya, H. J. Frohn, Z. Anorg. Allg. Chem. 2002, 628, 883.
[33] J. Plesek, T. Jelinek, S. Hermanek, B. StGbr, Collect. Czech.
Chem. Commun. 1986, 51, 819.
[34] a) C. A. Reed, K.-C. Kim, E. S. Stoyanov, D. Stasko, F. S. Tham,
L. J. Mueller, P. D. W. Boyd, J. Am. Chem. Soc. 2003, 125, 1796;
b) C. A. Reed, K.-C. Chan, R. D. Bolskar, L. J. Mueller, Science
2000, 289, 101; c) C. A. Reed, N. L. P. Fackler, K.-C. Kim, D.
Stasko, D. R. Evans, L. J. Mueller, P. D. W. Boyd, C. E. F.
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
www.angewandte.org
Rickard, J. Am. Chem. Soc. 1999, 121, 6314; d) D. Stasko, C. A.
Reed, J. Am. Chem. Soc. 2002, 124, 1148.
K.-C. Kim, C. A. Reed, D. W. Elliott, L. J. Mueller, F. Tham, L.
Lin, J. B. Lambert, Science 2002, 297, 825.
C. A. Reed, K.-C. Kim, R. D. Bolskar, L. J. Mueller, Science
2000, 289, 101.
K.-C. Kim, C. A. Reed, G. S. Long, A. Sen, J. Am. Chem. Soc.
2002, 124, 7662.
a) C.-W. Tsang, Q. Yang, E. Tung-Po Sze, T. C. W. Mak,
D. T. W. Chan, Z. Xie, Inorg. Chem. 2000, 39, 5851; b) Z. Xie,
C.-W. Tsang, E. Tung-Po Sze, Q. Yang, D. T. W. Chan, T. C. W.
Mak, Inorg. Chem. 1998, 37, 6444.
Z. Xie, C.-W. Tsang, F. Xue, T. C. W. Mak, J. Organomet. Chem.
1999, 577, 197.
B. T. King, Z. Janousek, B. GrVner, M. Tramell, B. C. Noll, J.
Michl, J. Am. Chem. Soc. 1996, 118, 3313.
B. T. King, J. Michl, J. Am. Chem. Soc. 2000, 122, 10 255.
S. M. Ivanova, S. V. Ivanov, S. M. Miller, O. P. Anderson, K. A.
Solntsev, S. H. Strauss, Inorg. Chem. 1999, 38, 3756.
a) S. V. Ivanov, J. J. Rockwell, O. G. Polyakov, C. M. Gaudinski,
O. P. Anderson, K. A. Solntsev, S. H. Strauss, J. Am. Chem. Soc.
1998, 120, 4224; b) S. H. Strauss in Contemporary Boron
Chemistry (Ed.: M. Davidson), Royal Society of Chemistry,
Cambridge, 2000, P. 44.
C.-W. Tsang, Q. Yang, E. Tung-Po Sze, T. C. W. Mak, D. T. W.
Chan, Z. Xie, Inorg. Chem. 2000, 39, 3582.
S. V. Ivanov, S. M. Miller, O. P. Anderson, K. A. Solntsev, S. H.
Strauss, J. Am. Chem. Soc. 2003, 125, 4694.
a) A. Franken, B. T. King, J. Rudolph, P. Rao, B. C. Noll, J.
Michl, Collect. Czech. Chem. Commun. 2001, 66, 1238; b) A.
Franken, B. T. King, J. Michl, WO 2002079210, 2002.
I. Krossing, Chem. Eur. J. 2001, 7, 490.
[Al(ORF)4] : a) T. J. Barbarich, S. T. Handy, S. M. Miller, O. P.
Anderson, P. A. Grieco, S. H. Strauss, Organometallics 1996, 15,
3776; b) T. J. Barbarich, S. M. Miller, O. P. Anderson, S. H.
Strauss, J. Mol. Catal. A 1998, 128, 289; c) S. M. Ivanova, B. G.
Nolan, Y. Kobayashi, S. M. Miller, O. P. Anderson, S. H.
Strauss, Chem. Eur. J. 2001, 7, 503.
S. H. Strauss, B. G. Nolan, B. P. Fauber, WO 00/53611, 2000.
M. Gonsior, I. Krossing, N. Mitzel, Z. Anorg. Allg. Chem. 2002,
628, 1821.
I. Krossing, H. Brands, R. Feuerhake, S. Koenig, J. Fluorine
Chem. 2001, 112, 83.
a) Y. Sun, M. V. Metz, C. L. Stern, T. J. Marks, Organometallics
2000, 19, 1625; b) M. V. Metz, Y. Sun, C. L. Stern, T. J. Marks,
Organometallics 2002, 21, 3691.
F. A. R. Kaul, G. T. Puchta, H. Schneider, M. Grosche, D.
Mihalios, W. A. Herrmann, J. Organomet. Chem. 2001, 621, 184.
By scaling up the procedure given in ref. [47]. Note added in
proof (March 30, 2004): This Li salt is now available from
Strem.
a) A. Reisinger, I. Krossing, Diplomarbeit, UniversitZt Karlsruhe 2003; b) I. Krossing, A. Reisinger, Angew. Chem. 2003,
115, 5903; Angew. Chem. Int. Ed. 2003, 42, 5725.
a) I. Krossing, J. Am. Chem. Soc. 2001, 123, 4603; b) I. Krossing,
L. van WVllen, Chem. Eur. J. 2002, 8, 700.
A. Adolf, M. Gonsior, I. Krossing, J. Am. Chem. Soc. 2002, 124,
7111.
T. S. Cameron, A. Decken, I. Dionne, Min Fang, I. Krossing, J.
Passmore, Chem. Eur. J. 2002, 8, 3386.
a) I. Krossing, I. Raabe, Angew. Chem. 2001, 113, 4544; Angew.
Chem. Int. Ed. 2001, 40, 4406; b) M. Gonsior, I. Krossing, L.
MVller, I. Raabe, M. Jansen, L. van WVllen, Chem. Eur. J. 2002,
8, 4475.
I. Krossing, A. Bihlmeier, I. Raabe, N. Trapp, Angew. Chem.
2003, 115, 1569; Angew. Chem. Int. Ed. 2003, 42, 1531.
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2087
Reviews
I. Krossing and I. Raabe
[61] M. Bochmann, I. Krossing, unpublished results. [Ph3C]+[Al{OC(CF3)3}4] was used as an activator for ethene and
propene polymerization with [(SBI)ZrMe2] in the presence of
AliBu3. It was found that the aluminate was an equally efficient
initiator as [Ph3C]+[B(C6F5)4] .
[62] D. M. Van Seggan, P. K. Hurtburt, M. D. Noirot, O. P. Anderson, S. H. Strauss, Inorg. Chem. 1992, 31, 1423.
[63] [E(OTeF5)6] (E = As, Sb, Bi): H. P. A. Mercier, J. C. P. Saunders, G. T. Schrobilgen, J. Am. Chem. Soc. 1994, 116, 2921.
[64] [E(OTeF5)6] (E = Sb, Nb): D. M. Van Seggen, P. K. Hurlburt,
O. P. Anderson, S. H. Strauss, Inorg. Chem. 1995, 34, 3453.
[65] [Sb(OTeF5)6] : T. S. Cameron, I. Krossing, J. Passmore, Inorg.
Chem. 2001, 40, 2001.
[66] K. Moock, K. Seppelt, Z. Anorg. Allg. Chem. 1988, 561, 132.
[67] P. K. Hurlburt, J. J. Rack, J. S. Luck, S. F. Dec, J. D. Webb, O. P.
Anderson, S. H. Strauss, J. Am. Chem. Soc. 1994, 116, 10 003.
[68] M. Gerken, P. Kolb, A. Wegner, H. P. A. Mercier, H. Borrmann, D. A. Dixon, G. J. Schrobilgen, Inorg. Chem. 2000, 39,
2813.
[69] W. J. Casteel, Jr., P. Kolb, N. LeBlond, H. P. A. Mercier, G. J.
Schrobilgen, Inorg. Chem. 1996, 35, 929.
[70] G. J. Schrobilgen, presented at the ACS meeting in New
Orleans, March 2003, INOR 770.
[71] T. S. Cameron, I. Dionne, I. Krossing, J. Passmore, Solid State
Sci. 2002, 4, 1435.
[72] a) S. Brownstein, Can. J. Chem. 1969, 47, 605; b) K. O. Christe,
W. Maya, Inorg. Chem. 1969, 8, 1253; c) P. A. W. Dean, R. J.
Gillespie, R. Hume, J. Chem. Soc. D 1969, 990.
[73] a) R. Minkwitz, F. Neikes, Inorg. Chem. 1999, 38, 5960; b) R.
Minkwitz, C. Hirsch, Eur. J. Inorg. Chem. 1999, 2249; c) K. O.
Christe, X. Zhang, R. Bau, J. Hegge, G. A. Olah, G. K. S.
Prakash, J. A. Sheehy, J. Am. Chem. Soc. 2000, 122, 481.
[74] [Sb3F16] : a) I. Bernhardi, T. Drews, K. Seppelt, Angew. Chem.
1999, 111, 2370; Angew. Chem. Int. Ed. 1999, 38, 2232; b) R.
Faggiani, D. K. Kennepohl, C. J. L. Lock, G. J. Schrobilgen,
Inorg. Chem. 1986, 25, 563; c) T. Drews, W. Koch, K. Seppelt, J.
Am. Chem. Soc. 1999, 121, 4379.
[75] [Sb4F21] : T. Drews, K. Seppelt, Angew. Chem. 1997, 109, 264;
Angew. Chem. Int. Ed. Engl. 1997, 36, 273.
[76] A. J. Edwards, G. R. Jones, R. J. Sills, J. Chem. Soc. Chem.
Commun. 1968, 1527.
[77] S. Seidel, K. Seppelt, Science 2000, 290, 117.
[78] T. Drews, S. Seidel, K. Seppelt, Angew. Chem. 2002, 114, 470;
Angew. Chem. Int. Ed. 2002, 41, 454.
[79] a) H. Willner, M. Bodenbinder, R. BrXchler, G. Hwang, S. J.
Rettig, J. Trotter, B. von Ahsen, U. Westphal, V. Jonas, W. Thiel,
F. Aubke, J. Am. Chem. Soc. 2001, 123, 588; b) H. Willner, F.
Aubke, Angew. Chem. 1997, 109, 2506; Angew. Chem. Int. Ed.
Engl. 1997, 36, 2402; c) H. Willner, F. Aubke, Inorg. Chem.
Highlights 2002, 195.
[80] a) H. D. B. Jenkins, H. K. Roobottom, J. Passmore, L. Glasser,
Inorg. Chem. 1999, 38, 3609; b) T. S. Cameron, R. J. Deeth, I.
Dionne, H. Du, H. D. B. Jenkins, I. Krossing, J. Passmore, H. K.
Roobottom, Inorg. Chem. 2000, 39, 5614; c) S. Brownridge,
H. D. B. Jenkins, I. Krossing, J. Passmore, H. K. Roobottom,
Coord. Chem. Rev. 2000, 197, 397; d) T. E. Mallouk, G. L.
Rosenthal, G. Muller, R. Brusasco, N. Bartlett, Inorg. Chem.
1984, 23, 3167.
[81] K. O. Christe, D. A. Dixon, D. McLemore, W. W. Wilson, J.
Sheehy, J. A. Bootz, J. Fluorine Chem. 2000, 101, 151.
[82] I.-C. Hwang, K. Seppelt, Angew. Chem. 2001, 113, 3803; Angew.
Chem. Int. Ed. 2001, 40, 3690.
[83] Reviews: ref. [3] and a) W. E. Piers, T. Chivers, Chem. Soc. Rev.
1997, 26, 345; b) W. E. Piers, G. J. Irvine, V. C. Williams, Eur. J.
Inorg. Chem. 2000, 2131; c) G. Erker, Acc. Chem. Res. 2001, 34,
309; d) F. P. Gabbaï, Angew. Chem. 2003, 115, 2318; Angew.
Chem. Int. Ed. 2003, 42, 2218; For related Lewis acids based on
2088
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[84]
[85]
[86]
[87]
[88]
[89]
[90]
[91]
[92]
[93]
[94]
[95]
[96]
[97]
[98]
[99]
[100]
[101]
[102]
[103]
[104]
polyfluoroaromatic metalloids see: e) G. M. Brooke, J. Fluorine
Chem. 1997, 86, 1; f) S. C. Cohen, A. G. Massey, Adv. Fluorine
Chem. 1970, 6, 83.
A. G. Massey, J. Park, J. Organomet. Chem. 1966, 5, 218.
a) X. Yang, C. L. Stern, T. J. Marks, J. Am. Chem. Soc. 1991,
113, 3623; b) X. Yang, C. L. Stern, T. J. Marks, J. Am. Chem.
Soc. 1994, 116, 10 015.
J. A. Ewen, M. J. Elder, Eur. Patent Appl. 0 427 697, 1991; J. A.
Ewen, M. J. Elder, US Patent 5 561 092, 1996.
H. Jacobsen, H. Berke, S. DXring, G. Kehr, G. Erker, R.
FrXhlich, O. Meyer, Organometallics 1999, 18, 1724.
a) Y. X. Chen, C. L. Stern, S. Yang, T. J. Marks, J. Am. Chem.
Soc. 1996, 118, 12 451; b) L. Li, C. L. Stern, T. J. Marks,
Organometallics 2000, 19, 3332.
a) L. Li, T. J. Marks, Organometallics 1998, 17, 3996; b) T. J.
Marks, L. Li, Y. X. Chen, M. H. McAdon, P. N. Nickias, WO
WO99/06412, 1999.
a) V. C. Williams, W. E. Piers, W. Clegg, S. Collins, T. B. Marder,
J. Am. Chem. Soc. 1999, 121, 3244; b) V. C. Williams, G. J.
Irvine, W. E. Piers, Z. Li, S. Collins, W. Clegg, M. R. J.
Elsegood, T. B. Marder, Organometallics 2000, 19, 1619.
C. Fritze, F. KVber, H. BXhmen (Hoechst AG), Eur. Patent
Appl. 811 627, 1997.
a) V. C. Williams, C. Dai, Z. Li, S. Collins, W. E. Piers, W. Clegg,
M. R. J. Elsegood, T. B. Marder, Angew. Chem. 1999, 111, 3922;
Angew. Chem. Int. Ed. 1999, 38, 3695; b) M. V. Metz, D. J.
Schwartz, C. L. Stern, P. N. Nickias, T. J. Marks, Angew. Chem.
2000, 112, 1368; Angew. Chem. Int. Ed. 1999, 38, 1312; c) M. H.
McAdon, P. N. Nickias, T. J. Marks, WO WO99/06413, 1999.
G. Kehr, R. FrXhlich, B. Wibbeling, G. Erker, Chem. Eur. J.
2000, 6, 258.
a) H.-H. Brintzinger, D. Fischer, R. MVlhaupt, B. Rieger, R.
Waymouth, Angew. Chem. 1995, 107, 1255; Angew. Chem. Int.
Ed. Engl. 1995, 34, 1143; b) W. E. Piers, Chem. Eur. J. 1998, 4,
13.
a) S. D. Ittel, L. K. Johnson, M. Brookhart, Chem. Rev. 2000,
100, 1169; b) G. J. P. Britovsek, V. C. Gibson, D. F. Wass,
Angew. Chem. 1999, 111, 448; Angew. Chem. Int. Ed. 1999,
38, 429; c) S. Mecking, Coord. Chem. Rev. 2000, 203, 305.
a) R. Braun, J. Sauer, Chem. Ber. 1986, 119, 1269; b) P. A.
Grieco, J. J. Nunes, M. D. Gaul, J. Am. Chem. Soc. 1990, 112,
4595; c) A. Flohr, H. Waldmann, J. Prakt. Chem. 1995, 337, 609.
Reviews: a) A. Kumar, Chem. Rev. 2001, 101, 1; b) S. Saito in
Lewis Acids in Organic Synthesis, Vol. 1 (Ed.: H. Yamamoto),
Wiley-VCH, Weinheim, 2000, p. 9.
S. Moss, B. T. King, A. de Meijere, S. I. Kozushkov, P. E. Eaton,
J. Michl, Org. Lett. 2001, 3, 2001.
K. Fujiki, S. Ikeda, H. Kobayashi, A. Mori, A. Nagira, J. Nie, T.
Sonoda, Y. Yagupolskii, Chem. Lett. 2000, 66.
[Nb(ORF)6] : G. M. Kloster, W. J. Dubaz, P. A. Grieco, D. F.
Shriver, S. H. Strauss, Inorg. Chim. Acta 1997, 263, 195.
a) P. A. Grieco, W. J. DuBay, L. J. Todd, Tetrahedron Lett. 1996,
37, 8707; b) W. J. DuBay, P. A. Grieco, L. J. Todd, J. Org. Chem.
1994, 59, 6898.
N. J. Patmore, C. Hague, J. H. Cotgreave, M. F. Mahon, C. G.
Frost, A. S. Weller, Chem. Eur. J. 2002, 8, 2088.
See, for example: a) N. J. Stone, D. A. Sweigart, A. M. Bond,
Organometallics 1986, 5, 2553; b) C. G. Zoski, D. A. Sweigart,
N. J. Stone, P. H. Rieger, E. Mocellin, T. F. Mann, D. R. Mann,
D. K. Gosser, M. M. Doeff, A. M. Bond, J. Am. Chem. Soc.
1988, 110, 2109.
See, for example:a) R. Rulkens, A. J. Lough, I. Manners, S. R.
Lovelace, C. Grant, W. E. Geiger, J. Am. Chem. Soc. 1996, 118,
12 683; b) B. Grossmann, J. Heinze, E. Herdtweck, F. H.
KXhler, H. NXth, H. Schwenk, M. Spiegler, W. Wachter, B.
Weber, Angew. Chem. 1997, 109, 384; Angew. Chem. Int. Ed.
Engl. 1997, 36, 387.
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Angewandte
Chemie
Weakly Coordinating Anions
[105] a) R. J. LeSuer, W. E. Geiger, Angew. Chem. 2000, 112, 254;
Angew. Chem. Int. Ed. 2000, 39, 248; b) F. Barriere, N. Camire,
W. E. Geiger, U. T. Mueller-Westerhoff, R. Sanders, J. Am.
Chem. Soc. 2002, 124, 7262; c) N. Camire, A. Nafady, W. E.
Geiger, J. Am. Chem. Soc. 2002, 124, 7260; d) N. Camire, U. T.
Mueller-Westerhoff, W. E. Geiger, J. Organomet. Chem. 2001,
637–639, 823.
[106] a) P. G. Gassman, P. A. Deck, Organometallics 1994, 13, 1934;
b) M. G. Hill, W. M. Lamanna, K. R. Mann, Inorg. Chem. 1991,
30, 4687; c) P. G. Gassman, J. R. Sowa, Jr., M. G. Hill, K. R.
Mann, Organometallics 1995, 14, 4879.
[107] L. Pospisil, B. T. King, J. Michl, Electrochim. Acta 1998, 44, 103.
[108] http://www.ionicliquids-merck.de.
[109] Reviews: a) P. Wasserscheid, W. Keim, Angew. Chem. 2000,
112, 3927; Angew. Chem. Int. Ed. 2000, 39, 3772; b) J. D.
Holbrey, Clean Prod. Processes 1999, 1, 2071; c) T. Welton,
Chem. Rev. 1999, 99, 2071; d) P. Wasserscheid, Chem. Unserer
Zeit 2003, 37, 52.
[110] A. BXsmann, G. Francio, E. Janssen, W. Leitner, P. Wasserscheid, Angew. Chem. 2001, 113, 2769; Angew. Chem. Int. Ed.
2001, 40, 2697.
[111] A. S. Larsen, J. D. Holbrey, F. S. Tham, C. A. Reed, J. Am.
Chem. Soc. 2000, 122, 7264.
[112] See also the related straight forward synthesis of K+[F7C3BF3]
and K+[F13C6BF3] : H.-J. Frohn, V. V. Bardin, Z. Anorg. Allg.
Chem. 2001, 627, 15.
[113] a) F. Kita, H. Sakata, A. Kawakami, H. Kamizori, T. Sonoda, H.
Nagashima, N. V. Pavlenko, Y. Yagupolskii, J. Power Sources
2001, 97–98, 581; b) L. M. Yagupolski, L. Yu, J. Yagupolski, J.
Fluorine Chem. 1995, 72, 225; c) N. Ignat'tev, P. Sartori, J.
Fluorine Chem. 2001, 101, 203.
[114] W. Xu, C. A. Austen, Electrochem. Solid-State Lett. 2000, 3, 366.
[115] F. Kita, H. Sakata, S. Sinomoto, A. Kawakami, H. Kamizori, T.
Sonoda, H. Nagashima, J. Nie, N. V. Pavlenko, Y. Yagupolskii,
J. Power Sources 2000, 90, 27.
[116] H. Suzuki, H. Naganawa, S. Tachimori, Phys. Chem. Chem.
Phys. 2003, 5, 726.
[117] J. V. Crivello, Radiat. Curing Polym. Sci. Technol. 1993, 2, 435.
[118] F. Castellanos, J. P. Fouassier, C. Priou, J. Cavezzan, J. Appl.
Polym. Sci. 1996, 60, 705.
[119] a) K. Ren, A. Mejiritski, J. H. Malpert, O. Grinevich, H. Gu,
D. C. Neckers, Tetrahedron Lett. 2000, 41, 8669; b) H. Li, K.
Ren, W. Zhang, J. H. Malpert, D. C. Neckers, Macromolecules
2001, 34, 4161; c) H. Gu, K. Ren, A. Mejiritski, O. Grinevich,
J. H. Malpert, D. C. Neckers, J. Org. Chem. 2001, 66, 4161.
[120] K. Ren, J. H. Malpert, H. Li, H. Gu, D. C. Neckers, Macromolecules 2002, 35, 1632.
[121] H. Nishida, N. Takada, M. Yoshimura, T. Sonoda, H. Kobayashi, Bull. Chem. Soc. Jpn. 1984, 57, 2600.
[122] Further purification of Na+[B(C6H3(CF3)2)4] : a) A. D. Hennis,
J. D. Polley, G. S. Long, A. Sen, D. Yandulov, J. Lipian, G. M.
Benedikt, L. F. Rhodes, J. C. Huffman, Organometallics 2001,
20, 2802; b) R. Taube, S. Wache, J. Organomet. Chem. 1992, 428,
431.
[123] a) J. C. W. Chien, W.-M. Tsai, M. D: Rausch, J. Am. Chem. Soc.
1991, 113, 8570; b) J. A. Ewen, M. J. Elder, Eur. Patent Appl.
0 426 637, 1991; M. Bochmann, S. J. Lancaster, J. Organomet.
Chem. 1992, 434, C1.
[124] a) Z. Xie, B.-M. Wu, T. C. W. Mak, J. Manning, C. A. Reed, J.
Chem. Soc. Dalton Trans. 1997, 1213; b) Z. Xie, J. Manning,
R. W. Reed, R. Mathur, P. D. W. Boyd, A. Benesi, C. A. Reed,
J. Am. Chem. Soc. 1996, 118, 2922; c) Z. Xie, T. Jelinek, R. Bau,
C. A. Reed, J. Am. Chem. Soc. 1994, 116, 1907; d) K. Shelly,
C. A. Reed, Y. J. Lee, W. R. Scheidt, J. Am. Chem. Soc. 1986,
108, 3117; e) D. J. Liston, Y. J. Lee, W. R. Scheidt, C. A. Reed, J.
Am. Chem. Soc. 1989, 111, 6643.
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
[125] a) B. T. King, B. C. Noll, J. Michl, Collect. Czech. Chem.
Commun. 1999, 64, 1001; b) B. T. King, Z. Janusek, J. Michl,
US 5731470 A, 1998; c) B. T. King, Z. Janusek, J. Michl, WO
9631519A1, 1996.
[126] S. H. Strauss, S. V. Ivanov, WO 2002036557, 2002.
[127] S. H. Strauss, S. V. Ivanov, WO 98/43983, 1998.
[128] a) M. Brookhart, B. Grant, A. F. Volpe, Jr., Organometallics
1992, 11, 3920; b) D. J. Tempel, L. K. Johnson, R. L. Huff, P. S.
White, M. Brookhart, J. Am. Chem. Soc. 2000, 122, 6686.
[129] K. Fujiki, M. Kashiwagi, H. Miyamoto, A. Sonoda, J. Ichikawa,
H. Kobayashi, T. Sonoda, J. Fluorine Chem. 1992, 57, 307.
[130] P. Jutzi, C. MVller, A. Stammler, H.-G. Stammler, Organometallics 2000, 19, 1442.
[131] a) G. G. Hladky, D. J. Upton, H. W. Turner, WO 91/09882,
1992; b) X. Yang, C. L. Stern, T. J. Marks, Organometallics
1991, 10, 840; c) H. W. Turner, EP 0 277 004 A1, 1988.
[132] D. Stasko, S. P. Hoffmann, K.-C. Kim, N. L. P. Fackler, A. S.
Larsen, T. Drovetskaya, F. S. Tham, C. A. Reed, C. E. F.
Rickard, P. D. W. Boyd, E. S. Stoyanov, J. Am. Chem. Soc.
2002, 124, 13 869.
[133] C. W. Tsang, Q.-C. Yang, T. C. W. Mak, Z.-W. Xie, Chin. J.
Chem. 2002, 20, 1241.
[134] Review: N. G. Conolly, W. E. Geiger, Chem. Rev. 1996, 96, 877.
[135] M. Gonsior, F. Breher, I. Krossing, unpublished results.
[136] R. D. Bolskar, R. S. Mathur, C. A. Reed, J. Am. Chem. Soc.
1998, 118, 13 093.
[137] B. T. King, B. C. Noll, A. J. McKinley, J. Am. Chem. Soc. 1996,
118, 10 902.
[138] I. Zharov, B. T. King, Z. Havlas, A. Pardi, J. Michl, J. Am.
Chem. Soc. 2000, 122, 10 253.
[139] E. Y.-X. Chen, K. A. Abboud, Organometallics 2000, 19, 5541.
[140] Ref. [74a].
[141] a) A. V. Korolev, E. Ihara, I. A. Guzei, V. G. Joung, Jr., R. F.
Jordan, J. Am. Chem. Soc. 2001, 123, 8291; b) A. V. Korolev, F.
Delpech, S. Dagorne, I. A. Guzei, R. F. Jordan, Organometallics 2001, 20, 3367; c) F. Delpech, I. A. Guzei, R. F. Jordan,
Organometallics 2002, 21, 1167.
[142] a) H. C. Strauch, B. Wibbeling, R. FrXhlich, G. Erker, H.
Jacobsen, H. Berke, Organometallics 1999, 18, 3802; b) H. C.
Strauch, G. Erker, R. FrXhlich, Organometallics 1998, 17, 5746.
[143] [bersichten: a) J. B. Lambert, L. Kania, S. Zhang, Chem. Rev.
1995, 95, 1191; b) C. A. Reed, Acc. Chem. Res. 1998, 31, 325.
[144] S. Kainz, A. Brinkmann, W. Leitner, A. Pfaltz, J. Am. Chem.
Soc. 1999, 121, 6421.
[145] A. J. Lupinetti, M. D. Havighurst, S. M. Miller, O. P. Anderson,
S. H. Strauss, J. Am. Chem. Soc. 1999, 121, 11 920.
[146] a) G. A. Olah, G. Rasul, L. Heiliger, G. K. S. Prakash, J. Am.
Chem. Soc. 1996, 118, 3580; b) G. A. Olah, L. Heiliger, G. K. S.
Prakash, J. Am. Chem. Soc. 1989, 111, 8020; c) H. Vancik, K.
Percac, D. E. Sunko, J. Am. Chem. Soc. 1990, 112, 7418; d) J. W.
Hudgens, R. D. Johnson III, B. P. Tsai, S. A. Kafafi, J. Am.
Chem. Soc. 1990, 112, 5763; e) G. A. Olah, Y. K. Mo, E. G.
Melby, H. C. Lin, J. Org. Chem. 1973, 38, 367; f) G. A. Olah, G.
Rasul, A. K. Yudin, A. Burrichter, G. K. S. Prakash, A. L.
Chistyakov, A. V. Stankevich, I. S. Akhrem, N. P. Gambaryan,
M. E. Vol'pin, J. Am. Chem. Soc. 1996, 118, 1446.
[147] L. H. Shultz, M. Brookhart, Organometallics 2001, 20, 3975.
[148] M. D. Hannant, M. Schormann, M. Bochmann, Dalton Trans.
2002, 4071.
[149] G. Kehr, R. Roesmann, R. FrXhlich, C. Holst, G. Erker, Eur. J.
Inorg. Chem. 2001, 535.
[150] A. Vij, W. W. Wilson, V. Vij, F. S. Tham, J. A. Sheehy, K. O.
Christe, J. Am. Chem. Soc. 2001, 123, 6308.
[151] Ref. [74c].
[152] N. Wiberg, Coord. Chem. Rev. 1997, 163, 217, and references
therein.
www.angewandte.org
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2089
Reviews
I. Krossing and I. Raabe
[153] B. Twamley, S. T. Haubrich, P. P. Power, Adv. Organomet.
Chem. 1999, 44, 1.
[154] E. Lork, D. Viets, R. Mews, H. Oberhammer, Inorg. Chem.
2000, 39, 4838.
[155] A. R. Mahjoub, X. Zhang, K. Seppelt, Chem. Eur. J. 1995, 1,
261.
[156] A. Kolomeitsev, G. Bisky, E. Lork, V. Movchun, E. Rusanov, P.
Kirsch, G.-V. RXschenthaler, Chem. Commun. 1999, 1017.
[157] M. Lin, C. Liu, L. Zheng, Wuli Huaxue Xuebao 1995, 11, 266.
[158] Ref. [80a].
[159] a) D. D. Wagman, W. H. Evans, V. B. Parker, R. H. Schumm, I.
Halow, S. M. Bailey, K. L. Churney, R. L. Nuttal, J. Phys. Chem.
Ref. Data 1982, 11 (Suppl. 2); b) S. G. Lias, J. E. Bartmess, J. F.
Liebman, J. L. Holmes, R. D. Levin, W. G. Mallard, J. Phys.
Chem. Ref. Data 1988, 17 (Suppl. 1); c) CRC Handbook of
Chemistry und Physics, CRC Press, ; d) http://www.nist.gov/
chemistry.
[160] The simple Born equation was used to approximate the Gibbs
solvation energy: DGsolv. = 0.5{z2 e Na/4pe0 r}{11/erel.}. z =
charge of the ion, e = charge of the electron, Na = Avogadro's
constant, e0 = dielectric constant of the vacuum, r = ionic radius,
erel. relative dielectric constant of the medium. For the plot r was
arbitrarily set to 2.0 U.
[161] For the estimation of lattice versus Gibbs solvation energy
effects the size of the cation A+ was arbitrarily set to a
thermochemical volume of 10 U3 equivalent to an ionic radius
of about 1.3 U (similar to that of K+). Then the volume (radius)
of the ion X was changed from 10 U3 (1.3 U) to 1250 U3
(6.7 U) and solvation as well as lattice enthalpies were
calculated for A+X . Gibbs energies of solvation were approximated by the Born equation,[160] lattice enthalpies by the
volume based equation of Jenkins and Passmore.[158]
[162] H. W. Roesky, M. Thomas, J. Schimkowiak, P. G. Jones, W.
Pinkert, G. M. Sheldrick, J. Chem. Soc. Chem. Commun. 1982,
895.
[163] A. J. Lupinetti, V. Jonas, W. Thiel, S. H. Strauss, G. Frenking,
Chem. Eur. J. 1999, 5, 2573.
[164] The thermochemical volume of [Al(OR)4] is 758 U3,[51] that of
[Sb(OTeF5)6] is 725 U3.[65] Therefore the [B(OTeF5)4] ion
should constitute approximately 2/3 of the volume of the
[Sb(OTeF5)6] ion or about 483 U3.
[165] See ref. [166] for an up to date overview.
[166] J. Kanetti, L. C. P. M. de Smet, R. Boom, H. Zuilhof, E. J. R.
SudhXlter, J. Phys. Chem. A 2002, 106, 11 197 – 11 204.
[167] a) B. C. Guo, A. W. Castleman, Jr., Chem. Phys. Lett. 1991, 181,
16; b) D. SchrXder, R. Wesendrup, R. H. Hertwig, T. K. Dargel,
H. Grauel, W. Koch, B. R. Bender, H. Schwarz, Organometallics 2000, 19, 2608.
[168] Ref. [80c]
[169] Ref. [80b].
[170] T. S. Cameron, I. Dionne, H. D. B. Jenkins, S. Parsons, J.
Passmore, H. K. Roobottom, Inorg. Chem. 2000, 39, 2042.
[171] W. V. Brooks, T. S. Cameron, S. Parsons, J. Passmore, M.
Shriver, Inorg. Chem. 1994, 33, 6230.
[172] BE(O-O) = 142 kJ mol1 and BE(S-S) = 267 kJ mol1.
[173] However, the computation turned out to be very delicate and
finally converged at the highly correlated CCSD(T)-cc-pV5Z
level: H. D. B. Jenkins, L. C. Jitariu, I. Krossing, J. Passmore, R.
Suontamo, J. Comput. Chem. 2000, 21, 218.
[174] Ref. [170].
[175] J. Powell, A. Lough, T. Saeed, J. Chem. Soc. Dalton Trans. 1997,
4137.
[176] X. Song, M. Thornton-Pett, M. Bochmann, Organometallics
1998, 17, 1004.
[177] W. V. Konze, B. L. Scott, G. J. Kubas, Chem. Commun. 1999,
1807.
2090
2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[178] D. A. Walker, T. J. Woodman, D. L. Hughes, M. L. Bochmann,
Organometallics 2001, 20, 3772.
[179] H. J. Frohn, S. Jakobs, G. Henkel, Angew. Chem. 1989, 101,
1534; Angew. Chem. Int. Ed. Engl. 1989, 28, 1506.
[180] I. Krossing, Dalton Trans. 2002, 500.
[181] To evaluate the FIA of SbnF2n (n = 2, 3, 4) giving the fluoridebridged ions [SbnF5n+1] the FIA was calculated by two separate
calculations: 1) The FIA of the doubly fluoride-bridged Al2F6
(D2h) for the formation of the singly fluoride-bridged [Al2F7]
ion was calculated based on the average of G2 and CBS-Q
calculations as 501 kJ mol1. This step is non-isodesmic but the
G2 and CBS-Q levels are reported to reproduce experimental
values with a uncertainty of less than 8 kJ mol1 lending
credibility to these values. 2) The FIA of SbnF2n (n = 2, 3, 4)
was then calculated by adding the reaction enthalpy of the
isodesmic reaction [Al2F7] + SbnF2n ![SbnF2n+1] + Al2F6 to
the FIA of Al2F6 (501 kJ mol1; see Equations 37 a–d) for an
example for [Sb2F11]). Similarly the FIA of [(RO)3Al-FAl(OR)3] and [F4C6{1,2-B(C6F5)2}2F] was assessed in isodesmic reactions of [Al2F7] and the Lewis acid 2 Al(OR)3 and
F4C6{1,2-(B(C6F5)2}2 giving 2 AlF3 and the fluoride-bridged ion.
From this reaction enthalpy and the FIA of 2 AlF3—determined as the average of G2 and CBS-Q calculations
(706 kJ mol1)—the FIA of 2 Al(OR)3 and F4C6{1,2(B(C6F5)2}2 was calculated (see Equation 38 a–d) for an example for [(RO)3Al-F-Al(OR)3]).
[182] The LA was partitioned in two separate reactions: the first was
an isodesmic reaction with which even very large systems could
be calculated reliably at the BP86/SVP level [Eq. (39 a)]. The
second reaction contains much smaller species, however, is nonisodesmic. Therefore the computationally much more time
consuming but also more reliable MP2/TZVPP level was
selected to assess the second part (It is currently impossible to
run also the first reaction at the MP2 level). The LA was than
obtained by a simple addition of both equations [Eq. (39 c)].
[183] I. A. Koppel, P. Burk, I. Koppel, I. Leito, T. Sonoda, M.
Mishima, J. Am. Chem. Soc. 2000, 122, 5114.
[184] I. Krossing, I. Raabe, Chem. Eur. J., submitted.
[185] R. J. Wehmschulte, J. M. Steele, J. D. Young, M. A. Khan, J.
Am. Chem. Soc. 2003, 125, 1470.
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2066 – 2090
Документ
Категория
Без категории
Просмотров
2
Размер файла
701 Кб
Теги
anionsчfact, survey, likely, noncoordinate, fictionф, candidatus
1/--страниц
Пожаловаться на содержимое документа