Preparation of stable AsBr4+ and I2As-PI 3+ salts. Why didn't we succeed to prepare AsI 4+ and As2X5+? A combined experimental and theoretical study

Article abstract of DOI:10.1039/b417629d

In analogy to our successful PX₂⁺ insertion reactions, an AsX₂⁺ insertion route was explored to obtain new arsenic halogen cations. Two new salts were prepared: AsBr₄⁺[Al(OR)₄]^-, starting from AsBr₃, Br₂ and Ag[Al(OR)₄], and I ₂As-PI₃⁺[Al(OR)]₄ from AsI ₃, PI₃ and Ag[Al(OR)₄] (R = C(CF ₃)₃). The first cation is formally a product of an AsBr₂⁺ insertion into the Br₂ molecule and the latter clearly a PI₂⁺ insertion into the As-I bond of the AsI₃ molecule. Both compounds were characterized by IR and NMR spectroscopy, the first also by its X-ray structure. Reactions of Ag[Al(OR)₄] with AsI₃ do not lead to ionization and Agi formation but rather lead to a marginally stable Ag(AsI₃)₂⁺[Al(OR)]₄ salt. Despite many attempts we failed to prepare other PX-cation analogues such as AsI ₄⁺, As₂X₅⁺ and P ₄AsX₂⁺ (X = Br, I). To explain these negative results the thermodynamics of the formation of EX₂⁺, EX₄⁺ and E₂X₅⁺ (E = As, P; X = Br, I) was carefully analyzed with MP2/TZVPP calculations and inclusion of entropy and solvation effects. We show that As₂Br₅ ⁺ is in very rapid equilibrium with AsBr₂⁺ and AsBr₃ (ΔG°(CH₂Cl₂) = +30 kJ mol ^-1). The extremely reactive AsBr₂⁺ cation available in the equilibrium accounts for the observed decomposition of the [Al(OR)₄]^- anion. By contrast, the stability of AsI ₃ against Ag[Al(OR)₄] appears to be kinetic and, if prepared by a suitable route, As₂I₅⁺ would be expected to have a stability intermediate between the known P₂I ₅⁺ and P₂Br₅⁺. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b417629d

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
+
+
Preparation of stable AsBr4 and I2As–PI3 salts. Why didn’t we
+
+
succeed to prepare AsI4 and As2X5 ? A combined experimental and
theoretical study†
Marcin Gonsiora and Ingo Krossing*b
a Universita¨t Karlsruhe, Institut fu¨r Anorganische Chemie, Engesserstr. Geb. 30.45, 76128,
Karlsruhe, Germany. E-mail: marcin.gonsior@chemie.uni-karlsruhe.de
b Ecole Polytechnique Fe´de´rale de Lausanne (EPFL), Laboratory of Inorganic and Coordination
Chemistry (LCIC), ISIC-BCH, 1015, Lausanne, Switzerland. E-mail: ingo.krossing@epfl.ch
Received 22nd November 2004, Accepted 16th February 2005
First published as an Advance Article on the web 2nd March 2005
In analogy to our successful “PX2+” insertion reactions, an “AsX2+” insertion route was explored to obtain new
arsenic halogen cations. Two new salts were prepared: AsBr4+[Al(OR)4]−, starting from AsBr3, Br2 and Ag[Al(OR)4],
and I2As–PI3+[Al(OR)]4 from AsI3, PI3 and Ag[Al(OR)4] (R = C(CF3)3). The first cation is formally a product of an
“AsBr2+” insertion into the Br2 molecule and the latter clearly a “PI2+” insertion into the As–I bond of the AsI3
molecule. Both compounds were characterized by IR and NMR spectroscopy, the first also by its X-ray structure.
Reactions of Ag[Al(OR)4] with AsI3 do not lead to ionization and AgI formation but rather lead to a marginally
stable Ag(AsI3)2+[Al(OR)]4+salt. Despite many attempts we failed to prepare other PX-cation analogues such as
+
AsI4+, As2X5 and P4AsX2 (X = Br, I). To explain these negative results the thermodynamics of the formation of
EX2+, EX4+ and E2X5+ (E = As, P; X = Br, I) was carefully analyzed with MP2/TZVPP calculations and inclusion of
entropy and solvation effects. We show that As2Br5+ is in very rapid equilibrium with AsBr2+ and AsBr3 (DG◦
=
(CH Cl
2
)
2
+30 kJ mol−1). The extremely reactive AsBr2 cation available in the equilibrium accounts for the observed
+
decomposition of the [Al(OR)4]− anion. By contrast, the stability of AsI3 against Ag[Al(OR)4] appears to be kinetic
+
and, if prepared by a suitable route, As2I5 would be expected to have a stability intermediate between the known
+
+
P2I5 and P2Br5
.
intermediate generated in situ from phosphorus trihalide and
the silver salt Ag[Al(OR)4] (R = C(CF3)3) (see Scheme 1).37–40
Introduction
Few reports on binary As–X species (X = halogen) have
appeared and binary As–X cations with more than one arsenic
atom are unknown. The limited spectrum of well characterized
binary As–X neutrals ranges from AsF5,1,2 very unstable AsCl5
(decomposition above −50 ◦C),3–7 and AsCl F5− (n = 1–4)8–12
n
n
to AsX3 (X = F, Cl, Br, I) and easily decomposing As2I4.13
These rare examples may be contrasted by the wide range of
phosphorus halides known: PX5 (X = F, Cl, Br),13,14 PX3 (X =
F, Cl, Br, I),13 P2X4 (X = F, Cl, Br, I),13,15 P6X6,16 P3I5,17 P4X2
(X = Br, Cl)15 and P7X3 (X = Br, I).15 Turning to the cations, the
Scheme 1 Formation of the “PX2+” intermediate and its insertion into
the Z–Y bond of a simple molecule (e.g. I2P–I, Br–Br, PI2–PI2 etc.).
situation is similar: while several salts of PCl4+, PBr4 and PI4
+
+
+
are well characterized by X-ray, 31P NMR and vibrational spec-
troscopy as well as quantum-chemical calculations18–25 (PF4+ was
Insertion of the PX2 cation (X = Cl, Br, I) into the Z–Y
bond of simple molecules with Z = Br, I, P or S and Y =
Br, I, P led to several new binary and ternary P–X cations in
good yields.24,25,35,41 This carbenoid PX2+ cation is highly reactive
and the above insertion reactions require a robust counterion
to avoid decomposition. The only Weakly Coordinating Anion
(WCA) that was compatible with the cations, was the fluorinated
alkoxyaluminate [Al(OR)4]− with R = C(CF3)3.36,37,39 For a
comparison of the stabilities of several WCAs see our earlier
work.42,43 The intermediately generated “PX2+” cation does not
decompose this anion at low temperature (−78 to −30 ◦C) and
the insertion products are stable with those anions in solution
between −30 and 0 ◦C (some even for 15 min at RT). Therefore,
we were interested to transfer this pathway to arsenic chemistry.
Thus, using an “AsX2+” intermediate generated in situ from
AsX3 (X = Br, I) and Ag[Al(OR)4], it should be possible to
synthesize this class of compounds by insertion reactions (see
Fig. 1 for target cations).
observed by Raman spectroscopy),26,27 the AsX4 cations char-
+
−
28
acterized by an X-ray structure are limited to AsCl4+AsF6
,
AsCl4+[As(OTeF5)6]− and AsBr4+[As(F)(OTeF5)5]−.29 The na-
ture of AsI4+ and AsF4+ is somewhat questionable; these cations
are reported to be highly unstable salts decomposing above
−78 ◦C and were described by Raman and IR spectroscopy.30,31
+
The only other binary As–X cations known are of type AsX2
(AsIII) and were observed in a mass spectrometer.32,33 Other As–
X cations such as the As analogues of the low valent phosphorus
cations P2X5+ (X = Br, I),24,34 P3I6
,
25,35 or phosphorus rich P5X2
+
+
(X = Cl, Br, I)25,35,36 still remain unknown in arsenic chemistry.
A new method to prepare hitherto unknown binary phospho-
rus halogen cations is the insertion of a “PX2+” intermediate—
a positively charged carbene analogue—into the P–P or P–X
(X = Br, I) bonds of neutral P X molecules forming the
n
y
species mentioned above. The “PX2+” cation is a short lived
However, the transfer of this insertion methodology from P to
As turned out to be rather difficult and the only compounds that
could be synthesized by this route were AsBr4+[Al(OR)4]− and
I2AsPI3+[Al(OR)4]−. In this contribution we describe successful
and unsuccessful synthetic efforts directed to the preparation of
† Electronic supplementary information (ESI) available: Detailed exper-
imental and theoretical discussion. See http://www.rsc.org/suppdata/
dt/b4/b417629d/
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3
1 2 0 3
©

View Article Online
+
+
energies for the formation of PX4 and P2X5+ are by 35–
65 kJ mol−1 more favorable than those for AsX4 and As2X5
,
+
but, when including the formation of solid AgX, also the arsenic
cations should form.
Syntheses
All experiments were started at −78 ◦C to avoid anion decompo-
sition by in situ formed “AsX2+”. Than the mixtures were allowed
to warm to about −30 ◦C, or, if no AgX precipitation was
observed, also warmed to ambient temperature. In the following
R denotes the C(CF3)3 group.
Reactions of AsX3 and Ag[Al(OR)4]. We attempted to gen-
+
erate the AsX2 intermediate according to eqn. (7).
Fig. 1 Arsenic and arsenic–phosphorus halogen cations that were
targeted by insertion of “AsX2+” into the X–X, E–X and P–P bonds of
X2, EX3 and P4 (E = P, As; X = Br, I).
AsX3 + Ag[Al(OR)4]
→ AsX2+[Al(OR)4]− + AgX (X = Br, I)
(7)
For AsBr3, precipitation of AgBr was observed and after
removing all volatiles of the mixture the weight of the reaction
residue was reduced to about a half, typical for anion decompo-
sition (also NMR). Solid AsBr3 and solid Ag[Al(OR)4] ground
in the glove box for 15 min react already in the solid state at
room temperature and decompose the anion (mass balance,
IR, NMR). From this solid-state reaction (after dissolving the
residue in THF–CH2Cl2 mixture) the decomposition product
[(RO)2AlF(THF)]2 was isolated and completely characterized
(X-ray, IR, NMR). By contrast, AsI3 did not react under the
same conditions.
new binary As–X cations and analyze the reasons for the failure
with the additional help of detailed ab initio calculations.
Results
Due to the at first sight rather confusing experimental results, we
start with the general thermochemistry underlying the proposed
reactions leading to the products as in Fig. 1.
Initial thermodynamic calculations
A word of caution before starting: The thermodynamics of a
reaction may be predicted by quantum chemical calculations.
However, the underlying kinetics is not assessed by this ap-
proach. Thus experiments are needed to verify the predictions
made based on the computed thermodynamics. Since earlier
work24,35 showed the necessity to employ ab initio and not
DFT methods, we exclusively used the balanced MP2/TZVPP
level44–46 throughout the work. Energies and geometries of the
binary P–X cations obtained with this level are in very good
agreement with the experiment and thus lend confidence to
conclusions drawn from these calculations. All enthalpies and
free energies given in this section include the ZPEs, entropic and
thermal contributions to the free energy at 298 K as well as
the solvation energies (in CH2Cl2 using the COSMO solvation
model47).
+
The synthesis of As2X5 salts was attempted according to
eqn. (8):
2AsX3 + Ag[Al(OR)4]
→ As2X5+[Al(OR)4]− + AgX (X = Br, I)
(8)
No visible reaction occurred below −30 ◦C. Rising the
temperature to −20 ◦C led to precipitation of AgBr but only the
decomposition of the anion was observed (IR, NMR, mass bal-
ance). In all those experiments with AsBr3, the only crystalline
material that we isolated was AsBr3 (unit cell determination, IR).
When 1 or 2 equiv. of AsI3 were reacted with Ag[Al(OR)4]
no reaction was observed even at room temperature. No
precipitation of AgI occurred even when we changed the solvent
to a CS2–CH2Cl2 mixture to improve the solubility of AsI3.
The reaction mixture left at room temperature for weeks in a
sealed NMR tube showed no decomposition of the anion and
still no precipitation of AgI. From cooled AsI3/Ag[Al(OR)4]
mixtures only AsI3 precipitated. We never had any indication
+
+
Thermodynamics of the formation of E2X5 and EX4 (E =
P, As). We assessed the enthalpies for the formation of the
“EX2+” intermediate and EX4 or E2X5 starting from the
+
+
elemental trihalide, dihalogen and the in CH2Cl2 solution
for the precipitation of Ag(AsI3) +[Al(OR)4]− salts, although
x
+
prevailing Ag(CH2Cl2)3 cation (see Table 1).
they may be present in solution. This may be concluded from
the essentially better solubility of AsI3 in Ag[Al(OR)4]–CH2Cl2
solution than in pure CH2Cl2 and the intense red color of
the solution. Pure AsI3 is nearly insoluble in CH2Cl2 and the
solution is only little colored. Interestingly, the IR-spectra of the
solid-state reaction of 2 equiv. AsI3 and the “naked” silver salt
ground in the glove-box for 15 min showed that AsI3 molecules
Excluding the formation of solid AgX, the energetics of
+
the formation of all arsenic cations, the PX2 intermediate
+
and P2Br5 are endergonic in solution. However, as shown
◦
experimentally, PX2 formation proceeds even at −78 C.24,25,35
In agreement with this, all reactions are exergonic, if the
formation of solid AgX is assumed. Overall, the solution Gibbs
+
Table 1 DH◦(g), DG◦(g), DG◦
for the formation of EX4 and E2X5+. The values in parenthesis in italics include the formation of solid AgX.48
+
(CH2Cl2
)
Eqn.
Reaction (MP2/TZVPP)
DH◦
DG◦
DG◦
(g)
(g)
(CH2Cl2)
+
+
(1a)
(1b)
(2a)
(2b)
(3a)
(3b)
(4a)
(4b)
(5a)
(5b)
(6a)
(6b)
AsBr3 + Ag(CH2Cl2)3 → AsBr2 + AgBr + 3CH2Cl2
+378
+329
+158
+143
+235
+151
+378
+329
+97
+244
+196
+79
+130 (−87)
+101 (−96)
+14 (−203)
+12 (−185)
+101 (−116)
+30 (−167)
+130 (−87)
+104 (−93)
−49 (−266)
−18 (−215)
+48 (−169)
−6 (−191)
+
+
AsI3 + Ag(CH2Cl2)3 → AsI2 + AgI + 3CH2Cl2
+
+
AsBr3 + Br2 + Ag(CH2Cl2)3 →AsBr4 + AgBr + 3CH2Cl2
+
+
AsI3 + I2 + Ag(CH2Cl2)3 → AsI4 + AgI + 3CH2Cl2
+67
+
+
2AsBr3 + Ag(CH2Cl2)3 → As2Br5 + AgBr + 3CH2Cl2
+155
+72
+
+
2AsI3 + Ag(CH2Cl2)3 → As2I5 + AgI + 3CH2Cl2
+
+
PBr3 + Ag(CH2Cl2)3 → PBr2 + AgBr + 3CH2Cl2
+243
+198
+21
+
+
PI3 + Ag(CH2Cl2)3 → PI2 + AgI + 3CH2Cl2
+
+
PBr3 + Br2 + Ag(CH2Cl2)3 → PBr4 + AgBr + 3CH2Cl2
+
+
PI3 + I2 + Ag(CH2Cl2)3 → PI4 + AgI + 3CH2Cl2
+107
+181
+107
+36
+
+
2PBr3 + Ag(CH2Cl2)3 → P2Br5 + AgBr + 3CH2Cl2
+106
+37
+
+
2PI3 + Ag(CH2Cl2)3 → P2I5 + AgI + 3CH2Cl2
1 2 0 4
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3

View Article Online
coordinate to the Ag+ cation. In this reaction the anion was not
decomposed unlike in the solid-state reaction between AsI3, I2
and Ag[Al(OR)4] (see below) and Ag(AsI3)x+[Al(OR)4]− with x
likely to be 2 formed.
AgI but also to a complete decomposition of the anion (13C,
19F NMR, mass balance)—presumably due to the reaction of
iodine and the silver salt (see below, eqns. (13) and (14)). No
single crystals could be isolated, except those of AsI3 which were
always isolated from the non-volatile residue of this reaction
(Raman, IR, X-ray).
To exclude solvent effects, the solid-state reaction between
AsI3, I2 and Ag[Al(OR)4] was carried out in the glove box.
However, only decomposition of the anion occurred (IR).
To confirm that the reaction between I2 and Ag+ leads to
anion decomposition (see eqns. (13) and (14)) we reacted I2
and Ag[Al(OR)4] at room temperature and observed complete
decomposition of the anion after precipitation of AgI.
The reaction of AsBr3, P4 and Ag[Al(OR)4]. In a reaction
+
analogously to the synthesis of P5X2 (X = Br, I) we reacted
AsBr3, P4 and Ag[Al(OR)4] to obtain P4AsBr2+[Al(OR)4]−
(eqn. (9)).
P4 + AsBr3 + Ag[Al(OR)4] → P4AsBr2+[Al(OR)4]− + AgBr (9)
The “AsBr2+” intermediate was anticipated to react with P4 to
+
the P4AsBr2 salt. However, after precipitation of AgBr above
−30 ◦C only decomposition of the anion without noticeable
I2 + Ag+ → “I+” + AgI
(13)
degradation of white phosphorus was observed. In the 31P NMR
only the signal of dissolved Ag(P4) + (x = 1, 2) was visible. This
x
ꢀꢀI+ꢀꢀ + [Al(OR)4]− → decomposition
(14)
is similar to our earlier reaction of PCl3, P4 and Ag[Al(OR)4]
that led to Ag(P4)2+[(RO)3Al–F–Al(OR)3]−.36
“I+” in eqns. (13) and (14) denote the hypothetical species
formed from I2 and Ag+ upon precipitation of the silver iodide.
The reaction of AsI3, PI3 and Ag[Al(OR)4]. After the un-
successful experiments with AsI3 we carried out the reaction
between AsI3, PI3 and Ag[Al(OR)4] to obtain the mixed AsPI5
+
Crystal structures
cation (eqn. (10)).
Crystal structure of AsBr4+[Al(OR)4]−. The air- and
moisture-sensitive yellowish block-shaped crystals of 1 were
grown from a concentrated CH2Cl2 solution at −80 ◦C.
AsI3 + PI3 + Ag[Al(OR)4] → AsPI5+[Al(OR)4]− + AgI (10)
In the 31P NMR spectra of the reaction mixture (243 K, in
CD2Cl2), two groups of signals were found: one that was assigned
The cation in 1 (Fig. 2) forms an almost ideal tetrahe-
dron with Br–As–Br angles between 108.93(2) and 110.54(2)◦,
on average 109.47(2)◦. The As–Br bond lengths range from
+
+
to the P2I5 cation—the known product of a PI2 insertion into
+
the P–I bond of PI3 and another which we assign to the I2As–PI3
˚
˚
2.2144(5) to 2.2318(5) A and are 2.2257(5) A on average. The
cation. Considering that AsI3 does not react with Ag[Al(OR)4],
[Al(OR)4]− anion has the usual geometry.50 This is the second
+
+
the AsI2PI3 cation is the only likely product of a PI2 insertion
+
structural report of the AsBr4 cation. The As–Br distances
(except P2I5+, eqn. (11)).
in 1 are in good agreement with the previously characterized
+
−
˚
AsBr4 [As(F)(OTeF5)5] (2.225(2)–2.236(2) A, average 2.232(2)
29
˚
A) and our quantum chemically calculated value of the As–
Br distance of 2.231 A (MP2/TZVPP). There is only one weak,
˚
˚
As1 · · · F1 contact at 3.552 A longer than the van der Waals limit
˚
(sum of van der Waals radii 3.5 A), but each of the bromine
atoms exhibits at least one shorter fluorine contact, i.e. Br1
(11)
˚
˚
˚
3.223 A, Br2 3.254 A, Br3 3.098 A, while Br4 has two close
The intensities of the signals of both cations in the 31P NMR
are 1 : 1. Neither addition of CS2 (to dissolve more AsI3) nor
additional AsI3 (1 equiv.) changed the product ratio. Another
reaction with 6 equiv. of AsI3, 1 equiv. of PI3 and the silver
salt was performed in CH2Cl2–CS2 (3 : 1) mixture for improved
solubility of AsI3. Again the product ratio was the same and we
could not isolate a pure I2As–PI3+ salt from the reaction mixture.
However, we succeeded to characterize I2As–PI3+[Al(OR)4]− as
˚
contacts at 3.237 and 3.266 A (sum of van der Waals radii of Br
˚
and F = 3.4 A). In total 17 weak Br–F contacts are formed (see
Fig. 3 and Table A, ESI†).
+
a mixture with P2I5 also by IR spectroscopy.
Reactions of AsX3, X2 and Ag[Al(OR)4]. Successful preparation
of AsBr4+[Al(OR)4]−. The reactions according to eqn. (12)
were carried out at −78 ◦C and kept below −30 ◦C.
AsX3 + X2 + Ag[Al(OR)4]
→ AsX4+[Al(OR)4]− + AgX (X = Br, I)
(12)
However, only the preparation of AsBr4+[Al(OR)4]− pro-
ceeded as indicated in eqn. (12). After precipitation of AgBr and
low-temperature work-up we isolated slightly yellowish crystals
in 54% yield. A X-ray single crystal structure determination
showed the crystals to be AsBr4+[Al(OR)4]− (1). The salt 1 was
also characterized by IR spectroscopy and by 75As NMR.
Fig. 2 Section of the solid-state structure of AsBr4+[Al(OR)4]−.
+
Thermal ellipsoids were drawn at the 25% probability level. Selected
In the attempted AsI4 synthesis, precipitation of yellow AgI
◦
˚
was not observed below −30 ◦C but only voluminous dark solids
bond lengths (A) and angles ( ): As1–Br1 2.2318(5), As1–Br2 2.2284(4),
As1–Br3 2.2144(5), As1–Br4 2.2283(4), Al1–O1 1.664(2), Al1–O2
1.679(2), Al1–O3 1.7030(9), Al1–O4 1.7039(9); Br3–As1–Br2 109.02(2),
Br3–As1–Br4 108.93(2), Br2–As1–Br4 110.54(2), Br3–As1–Br1
109.43(2), Br2–As1–Br1 109.64(2), Br4–As1–Br1 109.27(2).
precipitated at −78 ◦C. These solids could be reversibly dissolved
between −50 and −30 ◦C. We suspect that the precipitated solids
+
consist of a complex Ag(I2) salt, since very recently it was
x
found that the silver cation forms stable adducts with iodine
given the right counterion (see Ag(I2)+[AsF6]−).49 AsI3 is almost
insoluble in CH2Cl2 at low temperature and due to the volume
and shape of the precipitate, solid I2 could also be excluded.
The multitude of Br · · · F contacts suggests that the charge
is distributed equally over the four bromine atoms. Sim-
ilar to the AsBr4+[As(F)(OTeF5)5]− crystal structure there
are no substantial Br4As+ · · · F interactions. In contrast, in
◦
Warming the reaction mixture above 0 C led to formation of
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3
1 2 0 5

View Article Online
Fig. 5 Solid [(RO)2Al(l-F)(THF)]2 with 25% probability ellipsoids.
+
Fig. 3 Weak fluorine contacts of the AsBr4 cation in 1. The F–Br
˚
contacts are shown within the sum of van der Waals radii of 3.4 A with a
tolerance of +5%. Thermal ellipsoids were drawn at the 25% probability
level.
AsCl4+[As(OTeF5)6]− several closer Cl4As+ · · · F contacts were
found (from 3.377 to 3.432 A).29 In the solid state, the salt 1 forms
˚
a distorted CsCl structure with the almost spherical [Al(OR)4]−
+
anion forming a distorted primitive cubic array and the AsBr4
cations residing in the center of the cubic interstices (Fig. 4).
Fig. 6 The core of the solid-state structure of [(RO)2Al(l-F)(THF)]2.
All the –C(CF3)3 groups were omitted for clarity. Thermal ellipsoids
(except for isotropic H) were drawn at the 25% probability level.
Crystal structure of [(RO)2Al(l-F)(THF)]2. Colorless plates
of [(RO)2Al(l-F)(THF)]2 were grown from a CH2Cl2–THF solu-
tion at −28 ◦C. Neutral [(RO)2Al(l-F)(THF)]2 is a dimer, which
forms a centrosymmetric planar four-membered Al2F2 ring as
the central structural building block. The aluminium atom in ev-
ery (RO)2Al(l-F)(THF) unit has a five-fold coordination in be-
tween a square pyramid and a trigonal bipyramid (Figs. 5 and 6).
˚
than Al–OTHF (1.826(2) A) in the Lewis acid–base adduct
(THF)·Al(OR)3.36
NMR and vibrational spectroscopy
Similar building blocks were earlier found in [(RꢀOH)Tl]
75As NMR of AsBr4+[Al(OR)4]−. Since 75As is a quadrupolar
nucleus (I = 3/2, 100% abundance), only species with a high
local symmetry can be expected to give signals in the 75As NMR
spectrum. We recorded a spectrum of freshly prepared 1 in a
NMR tube reaction in CD2Cl2 solvent (according to eqn. (12),
[Al(ORꢀ)3F]2
(Rꢀ
=
C(H)(CF3)2) or [PhC(CH(SiMe3)2)
51
N(SiMe3)AlF2]2 and [(RꢀO)(Me)Al(ORꢀ)2AlF(Me)]2 (Rꢀ =
52
53
2,6-i-Pr2C6H3). There are two sets of Al–F distances in
˚
[(RO)2AlF(THF)]2: 1.791(1) and 1.887(1) A which may be
compared to the Al–F distances in ionic [(RꢀOH)Tl][Al(ORꢀ)3F]2
X = Br). At −65 ◦C a singlet appeared at d −148.2 (s, m1/2
=
51
(1.839(6) and 1.919(6) A), in neutral [PhC(C(SiMe3)2)
983.2 Hz) which is assigned to the AsBr4+ cation (Fig. C. ESI†).
This chemical shift is in agreement with the experimental value of
AsBr4+[AsF(OTeF5)5]− (−134.3 ppm in SO2ClF at r.t.) recorded
by Schrobilgen et al.29
˚
ꢀ
˚
N(SiMe3)AlF2]2 (1.829(3) and 1.850(3) A) and [(R O)(Me)
Al(ORꢀ)2AlF(Me)]2 (1.812(2) and 1.879(2) A). The Al–OTHF
˚
˚
dative bond (1.895(1) A) in [(RO)2AlF(THF)]2 is longer
Fig. 4 Distorted cubic primitive packing of the ions in AsBr4+[Al(OR)4]−. The AsBr4 cation is surrounded by eight [Al(OR)4]− anions forming a
+
distorted cubic coordination sphere. All R groups were omitted for clarity. The AlO4 units are shown as tetrahedra.
1 2 0 6
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3

View Article Online
IR spectrum of AsBr4+[Al(OR)4]−. The IR spectrum of
crystalline 1 recorded in Nujol (Fig. 7) shows, apart from the
anion bands, an intense absorption at 354 cm−1: the T2 mode
of AsBr4+. All other bands are due to the anion (no unassigned
vibration, Table F, ESI†).
As was observed). The 31P chemical shift of 2 at −183.6 ppm is
+
close to the phosphonium resonance of the –PI3+ moiety in P2I5
at d(31P) −157.2. The similarity of the chemical shifts suggests an
+
“exchange” of the I2+P– substituent in I2P–PI3 against I2As–X
giving the I2As–PI3 cation. The integration of the signals
+
+
showed the ratio P2I5 : AsI2PI3 to be 1 : 1.
+
IR spectrum of I2As–PI3
. The IR spectra of the mixture
+
+
of I2As–PI3 and P2I5 cations were recorded in Nujol between
CsI plates. The isolation of pure 2 from the reaction mixture
+
failed. However, we could assign all bands of I2As–PI3 in
a microcrystalline fraction (I) and fraction (II), that resulted
from removing the solvent of the reaction residue (see Fig. B
and section, ESI†). It should be noted that vibrations at 635 and
864 cm−1 indicate partial decomposition of the anion ([(RO)3Al–
F–Al(OR)3]− formation36). All attempts to obtain a Raman
spectrum of fraction I/II failed due to immediate decomposition
in the laser beam. The frequency calculation performed at
the MP2/TZVPP level for 2 is in good agreement with our
assignments in Table 2. For more details, the interested reader is
referred to a deposited section.
Computational results
All the species present in this paper were calculated at the
MP2/TZVPP level with the exception of molecules participating
in the ligand abstraction reaction according to eqn. (15). Due to
Fig. 7 IR spectrum of a Nujol mull of crystalline 1 showing the T2
vibration of AsBr4+ cation at 354 cm−1. The anion vibrations are marked
by an asterisk.
−
their size, EX2OR (E = As, P; X = Br, I), Al(OR)3 and Al(OR)4
(R = C(CF3)3) were only calculated at the BP86/SVP level.
Binary and ternary P/As/X cations. The BP86/SVP level
This result is in very good agreement with the calculated
378 cm−1 value at the MP2/TZVPP and 337 cm−1 value
at the BP86/SV(P) level. The m(T2) in the Raman spectrum
of AsBr4+[AsF(OTeF5)5]− was reported at 353 cm−1.29 In
+
gave unsatisfying results for the As2X5 cations. Neither an
+
˚
appropriate As–As distance, i.e. in As2Br5 2.701 A and in
+
˚
As2I5 2.675 A at the BP86/SVP level (cf. MP2 geometries
of the As2X5 cations in Fig. 9), nor a true minimum could
be reached (the optimized As2Br5 structure always had one
AsBr4+AsF6 this vibration was found at 349 cm−1 (Raman
−
+
spectrum).54,55 It may be noted that m(T2) increases with decreas-
ing coordination power of the anion, i.e.: m(T2, [AsF6]−) < m(T2,
[FAs(OTeF5)5]−) < m(T2, [Al(OR)4]−) or 349 < 353 < 354 cm−1.
+
imaginary frequency). Similar results and the failure of DFT has
been reported for the lighter homologous binary P–X cations.
In contrast, the MP2/TZVPP calculated As–As distances in
31P NMR spectrum of I2As–PI3
.
The 31P NMR spectrum of
+
+
˚
the AsI2PI3+/P2I5 salt mixture was recorded at −80 ◦C over a
+
As2X5 (X = Br, I), 2.493 and 2.472 A, are in very good agree-
ment with experimental As–As bond lengths in neutral cyclo-
period of one month (Fig. 8).
56
2−
57
˚
˚
(AsPh)6 (2.459 A ) or anionic As6Br8 (2.430(3)–2.465(3) A. ).
Therefore, and due to our experience with the homologous P–X
systems, we exclusively rely on the MP2/TZVPP calculations
for the following discussion. All calculated species are shown
in Fig. 9 which include important distances and bond angles,
partial charges and shared electron numbers for selected atoms
and bonds.
PX3, PX2+, P2X5+ and PX4+ (X = Br, I) used in the thermody-
namic calculations were calculated earlier;24 in this paper neither
their geometries nor structural parameters are shown (their
total energies, ZPEs, COSMO solvation energies, thermal and
entropic contributions are given in the supplemental material).
The arsenic cations and molecules have similar geometries to
the homologous phosphorus species. Fig. 9 also includes the
+
+
geometry of AsI2PI3 and PI2AsI3 cations.
+
Ag(CH2Cl2)3
. To get more reliable results for our thermo-
dynamic calculations we replaced the naked Ag+ ion by its
tris(methylene chloride) complex (see Fig. 10) that was found
in the crystal structure of (CH2Cl2)3Ag+[Al(OR)4]−,58 as well as
(CH2Cl2)3Ag+[(RO)3Al–F–Al(OR)3]−.36 The structure of such a
solvated silver cation is also present in CH2Cl2 solution and, as
will be shown, is important for the entropic balance. The starting
geometry was taken from the X-ray data and calculated in C2
symmetry.36,58
Fig. 8 31P NMR spectrum of the products of the reaction between PI3,
AsI3 and Ag[Al(OR)4] (eqn. (10)). In the range from +600 to −600 ppm
only signals of P2I5+ (two doublets) and I2As–PI3+ (s) are observed. The
spectrum was recorded at −80 ◦C.
No anion d◦ecomposition was visible when the sample was
+
stored at −28 C. Two groups of signals were found: the P2I5
+
+
cation was observed at d(31P) +129.7 (doublet of the I2P–
Anion decomposition by EX2 and E2X5
.
To estimate the
1
+
moiety, JPP = 325.1 Hz) and at d(31P) −157.2 (doublet of
reactivity of the EX2 salts in a decomposition reaction (eqn.
(15)), we calculated the four products of a RO− ligand abstrac-
tion from the [Al(OR)4]− anion (Fig. A, ESI†): EX2OR (C1)
(X = Br, I; E = P, As). Those species were only calculated at
–PI3+, 1JPP = 325.1 Hz) which is in good agreement with earlier
+
data.24 Another broader signal (see Fig. 8) we assign to the –PI3
+
moiety in I2As–PI3 appeared at d(31P) −183.6 (no coupling to
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3
1 2 0 7

View Article Online
+
+
+
Table 2 The IR frequencies of the mixture of I2As–PI3 and P2I5 cations. The vibrations of the pure P2I5 salt are also included as well as the
MP2/TZVPP calculated frequencies of the I2As–PI3 and P2I5 cationsc
+
+
+
+
I2As–PI3 MP2/TZVPP
P2I5 MP2/TZVPP
P2I5+ in (I)/(II)24
Pure P2I5
Int. (sym.) (IR:%)
Int. (sym.) (IR:%)
+
+
Fractions (I)/(II)
I2As–PI3 in (I)/(II)
440a
440a
—
408 (vs)
—
375 (s)
—
440a
412 (s)
430 (sh)
413 (s)
—
401 (s)
—
364 (s)
330 (s)
—
443 (aꢀ) (98)
—
462 (aꢀ) (59)
437 (aꢀꢀ) (100)
—
412 (s)
408 (vs)
402 (s)
375 (s)
364 (s)
334 (m)
247 (m)
237 (w)
223 (m)
206 (w)b
433 (aꢀꢀ) (100)
—
402 (s)
—
364 (s)
334 (m)
—
—
223 (m)
—
420 (aꢀ) (66)
—
394 (aꢀ) (34)
—
389 (aꢀꢀ) (63)
354 (aꢀ) (23)
—
—
247 (m)
237 (w)
—
268 (aꢀꢀ) (38)
254 (aꢀ) (11)
—
—
223 (m)
—
235 (aꢀ) (23)
—
—
—
185 (aꢀ) (14)
127 (aꢀ) (4)
100 (aꢀꢀ) (0)
92 (aꢀ) (2)
84 (aꢀ) (0)
69 (aꢀꢀ) (0)
63 aꢀ (0)
50 (aꢀ) (0)
42 (aꢀꢀ) (0)
25 (aꢀꢀ) (0)
142 (aꢀ) (2)
107 (aꢀꢀ) (0)
105 (aꢀ) (1)
88 (aꢀ) (0)
+
+
a One vibration of I2As–PI3 (calc. 443 cm−1) overlapped with P2I5 and anion vibration (at 445 cm−1) and is therefore not seen. b AsI3 (cf. Raman
spectrum of pure AsI3 at 207 cm−1). c “Fraction (I)” denotes the first fraction of more crystalline product, the “Fraction (II)” fraction was obtained
from the filtrate after crystallization by removing the solvent in vacuo.
+
+
+
+
Fig. 9 The MP2/TZVPP minimum geometries of the arsenic-halogen species As2X5 (Cs), AsX4 (Td), AsX2 (C2v), AsX3 (C3v), AsI2PI3 (Cs)
+
and PI2AsI3 (Cs) (X = Br, I). Calculated partial charges (“Paboon”) are given in italics, shared electron numbers (SEN) are given in italics and are
underlined. All species are true minima with no imaginary frequencies.
the BP86/SVP level, since the geometries of the other reaction
participators in eqn. (15), Al(OR)3 and Al(OR)4−, could only be
calculated at this lower level.40 However, the used reactions are
all isodesmic. Therefore, we expect the results to be reliable. The
results are included in the Discussion below (Table 3).
is formed only above −30 ◦C, while “PX2+” (X = Br, I)
intermediates are already formed at −78 ◦C. Moreover, “AsBr2
”
+
is much more reactive than “PX2+” and completely decomposed
+
the anion in all reactions. Neither P4AsBr2 (eqn. (9)) nor
+
As2Br5 (eqn. (8), X = Br) could be obtained. The only As
containing product was the AsBr4+ salt obtained by the reaction
of AsBr3, Br2 and Ag[Al(OR)4]. In contrast to the smooth
Discussion
+
reactions of PI3 with the silver salt leading to the P2I5 salt,
no reaction occurred between AsI3 and Ag[Al(OR)4] and an
“AsI2+” intermediate was never formed. Instead of AsI3, diiodine
reacted with the silver cation in the reaction between AsI3, I2 and
According to the information in the Initial calculations section,
+
+
the formation of AsX2 is almost as likely as that of PX2
.
However, the experiments proved that the “AsBr2+” intermediate
1 2 0 8
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3

View Article Online
Table 4 DH◦(g), DG◦ and DG◦
for equilibrium (16)
(CH2Cl2)
(g)
Reaction
DH◦
DG◦
DG◦
(g)
(g)
(CH2Cl2)
+
+
As2Br5 → AsBr3 + AsBr2
+144
+178
+197
+222
+90
+30
+
+
As2I5 → AsI3 + AsI2
+124
+138
+161
+70
+82
+
+
P2Br5 → PBr3 + PBr2
+
+
P2I5 → PI3 + PI2
+109
PX2+, if not involved in rapid exchange, was below the detection
limit at −90 ◦C. However, supporting the equilibrium in
eqn. (16), the P2Br5+[Al(OR)4]− salt can be used as a source
+
of “PBr2+” in an insertion reaction with P4 giving cleanly P5Br2
and PBr3. What is the situation for AsX2+? Thermodynamic
calculations show clearly that in solution the participation of
AsBr2 in such equilibrium is substantially greater than in the
+
Fig. 10 The silver tris(methylene chloride) complex Ag(CH2Cl2)3+ (C2).
Calculated partial charges (“Paboon”) are given in italics, shared electron
numbers (SEN) are given in italics and are underlined. This species is a
true minimum with no imaginary frequencies.
case of all other cations (Table 4).
+
+
E2X5 ꢀ EX3 + EX2 (E = As, P; X = Br, I)
(16)
Equilibrium (16) is only endergonic by +30 kJ mol−1 for
As2Br5+. In contrast, all other cations form considerably less
likely (endergonic by +70 to +109 kJ mol−1). More specifically,
Ag[Al(OR)4], forming an unknown but very reactive species that
completely decomposed the anion. In the reaction between AsI3,
PI3 and the silver salt only the products of the insertion of a
“PI2+” intermediate in the P–I bond of PI3 (→ P2I5+) and As–I
bond of AsI3 (→ I2As–PI3+) were obtained. No evidence for
the presence of an “AsI2+” intermediate was observed. In the
following we want to shed light on these puzzling facts.
PBr2 formation from P2Br5 is by +52 kJ mol−1 less favorable
+
+
+
than AsBr2 from As2Br5+. This means that the latter is far
more reactive and the equilibrium concentration of AsBr2+ is so
high that it immediately decomposes the anion. Eqn. (16) may
further be shifted to the right hand side by complex formation of
+
+
EX2 with CH2Cl2. It was shown24 that complexation of PBr2
with CH2Cl2 giving [PBr2 · · · ClCH2Cl]+ is favorable by about
+
On the AsBr2 intermediate
50 kJ mol−1. In summary it is therefore likely to assume that by
We showed that the formation of “AsBr2+” only led to the
decomposition of the anion. Why does the decomposition of the
+
additional CH2Cl2 coordination, the As2Br5 equilibrium (16)
may completely be shifted to the right hand side. This is in
good agreement with our experiments. The reaction of AsBr3
and Ag[Al(OR)4] did not occur below −30 ◦C, probably due to
kinetic reasons, and above this temperature the intermediately
+
+
anion dominate over the formation of As2Br5 and P4AsBr2
?
A possible answer is that the “AsBr2+” cation is very reactive
and is too short lived to insert in the As–Br or P–P bond of
AsBr3 or P4.59 To support this assumption, we compared the
reactivity of “EX2+” in a simple isodesmic ligand abstraction
reaction (eqn. (15)) from which the relative (thermodynamic)
+
formed AsBr2 completely decomposed the anion.
+
From the preceding it is clear that E2X5 (in equilibrium
+
with EX2 and EX3) is the true reactive cation in solution.
stability of the EX2 cations with the [Al(OR)4]− anion may be
+
To understand the experimental reactivity trends, one has
extracted (E = As, P; X = Br, I; Table 3).60
EX2+[Al(OR)4]− → EX2OR + Al(OR)3
+
to assess the anion stability in the presence of the E2X5
cation. Therefore, we combined the ligand abstraction reactions
+
in eqn. (15) (BP86/SVP level) with the E2X5 equilibrium
(E = P, As; X = Br, I; R = C(CF3)3)
(15)
(16) (MP2/TZVPP level) to obtain the thermodynamics of
+
According to the calculated thermodynamic data
AsI2+[Al(OR)4]− would be more stable than PI2+[Al(OR)4]− and
AsBr2+[Al(OR)4]− should be more stable than PBr2+[Al(OR)4]−.
This contrasts with the experiment. The observed reactivity
of the “EX2+” intermediate increases according to: “PI2+” <
“PBr2+” ꢁ “AsBr2+”. The calculations indicate that gaseous
decomposition of the anion by the E2X5 cations (by addition
of the enthalpies of both reactions). The values of DH(g),
DG◦ and DG◦
for this decomposition are collected in
(g)
(CH Cl )
2 2
Table 5.
The DG◦
values in Table 5 are in very good
(CH Cl
2
)
2
agreement with the experiments. The decomposition of the
+
PX2 is indeed more reactive, however, at low temperature in
P2Br5+[Al(OR)4]− salt occurs at 0 ◦C and proceeds quickly at
+
solution it has to be almost absent, since we have never found
room temperature◦(few hours), the P2I5 salt can be handled
+
24
+
+
direct evidence for the presence of PX2
.
In solution the PX2
with less care at 0 C and the As2Br5 ◦ salt (better described as
cations are incorporated in equilibrium (16) (E = P).
AsBr2+) does not survive even at −30 C!
In the above equilibrium practically no free PX2+ is present but
According to Table 5 and in comparison to known P2Br5+, the
+
only P2X5 at low temperature. It was shown experimentally24
relative stability of the As2I5+ cation looks promising. Therefore,
that even if a 1 : 1 Ag+ : PX3 stoichiometry was present, only the
As2I5 would be a good candidate to be prepared by a suitable
+
+
P2X5 cation is formed in the reaction and the NMR-signal of
(new) route.
Table 3 DH◦(g), DG◦(g), DG◦
of the isodesmic decomposition according to eqn. (15) calculated at the BP86/SVP level
)
(CH2Cl2
Decomposition of EX2+[Al(OR)4]−
DH◦
DG◦
DG◦
a
a
(g)
(g)
(CH2Cl2)
+
−
AsBr2 + Al(OR)4 → AsBr2OR + Al(OR)3
−397
−357
−426
−387
−393
−354
−418
−380
−88
+
−
AsI2 + Al(OR)4 → AsI2OR + Al(OR)3
−68
−111
−91
+
−
PBr2 + Al(OR)4 → PBr2OR + Al(OR)3
+
−
PI2 + Al(OR)4 → PI2OR + Al(OR)3
a The enthalpies in the gas phase are high since two charged species react to give two neutrals.
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3
1 2 0 9

View Article Online
Table 5 DH◦(g), DG◦ and DG◦
for the decomposition of E2X5+[Al(OR)4]−
(CH2Cl2)
(g)
Reaction of the decomposition of E2X5+[Al(OR)]4
DH◦
DG◦
DG◦
−
(g)
(g)
(CH2Cl2)
+
−
As2Br5 + Al(OR)4 → AsBr2OR + AsBr3 + Al(OR)3
−254
−179
−230
−166
−303
−230
−281
−219
−59
+3
+
−
As2I5 + Al(OR)4 → AsI2OR + AsI3 + Al(OR)3
+
−
P2Br5 + Al(OR)4 → PBr2OR + PBr3 + Al(OR)3
−29
+19
+
−
P2I5 + Al(OR)4 → PI2OR + PI3 + Al(OR)3
+
+
The formation of AsI2–PI3
The failure to prepare As2X5
+
The reaction between PI3, AsI3 and Ag[Al(OR)4] proceeds
It was shown earlier that P2X5 is in equilibrium with the
“PX2+” cation, likely as its CH2Cl2 solvate. For X = Br
this equilibrium may be used as an experimental source of
“PBr2+”.24 According to the current investigations, the same
equilibrium takes places for As2Br5+, but lies in this case almost
completely on the (solvated) “AsBr2+” side. This carbenoid
cation is to reactive to be handled in the presence of the
[Al(OR)4]− anion even at very low temperature. Therefore,
all attempts to prepare As2Br5+[Al(OR)4]−, AsBr2+[Al(OR)4]−
as well as P4AsBr2+[Al(OR)4]− salts failed and instead led to
+
according to a similar mechanism to that for the E2X5 cations.
I2As–PI3+ would be the product of a “PI2+” insertion in the As–I
bond and I2P–AsI3+ the product of an “AsI2+” insertion into the
P–I bond. The latter is by +27 kJ mol−1 less favorable than the
first (MP2/TZVPP), which is in agreement with the experiment
+
+
(I2As–PI3 formation). Therefore, AsI2–PI3 is the kinetically
+
and thermodynamically favored product. However, P2I5 was
also formed in the reaction mixture. Thus AsI3 and PI3 compete
in the insertion reaction with “PI2+” (see eqn. (17)).
+
anion decomposition. The As2I5 cation appears to be a likely
candidate for a successful synthesis with a stability that would be
+
expected to lie between the known P2Br5 and P2I5+. However,
+
the only in low concentration formed Ag(AsI3)2 complex is
kinetically stable against heterolytic As–I bond cleavage and no
+
evidence for the formation of As2I5 could be presented.
Experimental
All manipulations were performed using standard Schlenk or
dry box techniques and a dinitrogen or argon atmosphere (H2O
and O2 < 1 ppm). Apparatus were closed by J. Young valves.
All solvents were rigorously dried over P2O5 and degassed prior
to use and stored under N2. AsBr3 was prepared from arsenic
trioxide, sulfur and bromine according to the method described
in the literature61 and AsI3 was prepared from As2O3, KI and
HCl according to the literature.62,63 PI3 was prepared by reacting
white phosphorus and iodine in CS2. PBr3 (Fluka), I2 and Br2
(Merck) were purified prior to use by distillation, sublimation
or recrystallization. Purity of those halides was checked by
Raman spectroscopy. The silver aluminate Ag[Al(OR)4] was
prepared according to the literature.39 Raman and IR spectra
were recorded using a 1064 nm laser on a Bruker IFS 66v
spectrometer equipped with the Raman modul FRA 106. IR
spectra were recorded in Nujol mull between CsI plates. NMR
spectra (Bruker AC250) of sealed samples were run in CD2Cl2
and were referenced towards the solvent (1H, 13C) and external
aqueous AlCl3 (27Al), 85% H3PO4 (31P), CF3COOH (19F) or
NaAsF6 in d3-acetonitrile.
(17)
DG◦
for reaction (17a) is −203 kJ mol−1 and for
(CH Cl
2
)
2
(17b) −199 kJ mol−1 (including AgI(s)). The almost equal
thermodynamics of the reactions explains the 1 : 1 ratio of the
products of the reaction. Interestingly, excess AsI3 added to the
reaction mixture did not change the product ratio. In agreement
with this, the reaction of P2I5+ with AsI3 (eqn. (18)) is endergonic
by +4 kJ mol−1 (MP2/TZVPP).
◦
= +4 kJ mol−1 (18)
Additionally, the poor solubility of AsI3 at low temperature
influences eqn. (18): The equilibrium concentration of PI3 is
higher than that of AsI3 and the maximum concentration of
+
+
P2I5 + AsI3 → I2AsPI3 + PI3; DG
(CH Cl
)
2
2
+
AsI3 thus limits the amount of I2As–PI3 formed.
Conclusions
+
+
Preparation of AsBr4 and I2As–PI3
+
AsBr4 is the only As–X cation that could be prepared using
the “AsX2+” insertion route. AsBr4+[Al(OR)4]− is the second
structurally characterized representative of a AsBr4 cation. It
Reactions between AsI3, PI3 and Ag[Al(OR)4]
+
was also characterized by IR and 75As NMR spectroscopy. The
successfully prepared I2As–PI3+[Al(OR)4]− salt is, to the best
of our knowledge, the first mixed As/P-halogen cation. It is
the product of a “PI2+” insertion into the As–I bond of the AsI3
molecule. Its structure could be assigned based on 31P NMR and
IR spectroscopy (as a mixture with P2I5+) as well as quantum-
chemical calculations.
Reaction (1) (in CH2Cl2). Ag[Al(OR)4] (0.953 g, 0.887 mmol)
and AsI3 (0.410 g, 0.887 mmol) and were loaded into a two-bulb
vessel with an internal dropping funnel (two pieces connected
by a flange with viton O-ring and Teflon inlay). PI3 (0.368 g,
0.887 mmol) was placed in the dropping funnel and dissolved
in CH2Cl2 (30 ml). A few ml of CH2Cl2 were placed in stirred
mixture of AsI3/Ag[Al(OR)4]. The flask was cooled to about
−78 ◦C and the dissolved PI3 was added slowly in the course
of 60 min. After addition of the entire PI3, the reaction mixture
was stirred for 3 days at temperatures below −30 ◦C. A 31P
NMR spectrum of about 3 ml of the clear orange solution was
prepared by removing all volatiles in vacuo and addition CD2Cl2
(1 ml) (sample 1). An additional equivalent of AsI3 (0.402 g,
0.884 mmol) was added to the mixture via the funnel supported
by CS2 (5 ml). After 2 days stirring at −30 ◦C from 2 ml of the so-
lution another NMR sample was prepared as before (sample 2).
+
The failure to prepare AsI4
In contrast to all other investigated Ag+/X2/EX3 mixtures,
AsI4+[Al(OR)4] salts can not be prepared by this route, although
its formation would thermodynamically be possible. Instead it
+
is likely that a Ag(I2) cation forms as an intermediate which
x
reacts above 0 ◦C by separation of solid AgI giving products that
completely decompose the anion.
1 2 1 0
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3

View Article Online
Sample 1: 13C NMR (63 MHz, CD2Cl2, −30 ◦C): d 121.0
(q, JCF = 292.5 Hz, CF3 of the [Al(OR)4]− anion); 27Al NMR
(78 MHz, CD2Cl2, −20 ◦C): d 37.1 (s, m1/2 = 25 Hz); 31P NMR
(101 MHz, CD2Cl2, −30 ◦C): d 129.7 (d, JPP = 325.1 Hz, P2I5+),
−157.3 (d, JPP = 325.1 Hz, P2I5+), −183.7 (s, m1/2 = 40 Hz, AsI2–
PI3+). Sample 2: 13C NMR (63 MHz, CD2Cl2, −30 ◦C): d 121.0
(q, JCF = 292.4 Hz, CF3 of the [Al(OR)4]− anion); 27Al NMR
(78 MHz, CD2Cl2, −30 ◦C): d 37.4 (s, m1/2 = 15 Hz); 31P NMR
(101 MHz, CD2Cl2, −30 ◦C): d 129.7 (d, JPP = 325.3 Hz, P2I5+),
−157.3 (d, JPP = 325.3 Hz, P2I5+), −183.6 (s, m1/2 = 40 Hz, AsI2–
3 days to finish the reaction. The pale yellow-brownish solution
was filtered from the solid AgBr, concentrated to about a half
and crystallized. AsBr4+[Al(OR)4]− (R = C(CF3)3) (0.233 g, 54%
yield) was obtained from CH2Cl2 as slightly yellow block-shaped
crystals, sensitive to air and moisture. The crystals are stable
under inert atmosphere at r.t. for at least 2 months. The solution
is stable for a few hours at 0 ◦C.
IR (CsI, Nujol mull): m 354 (m, T2 AsBr4+); anion: 281 (w),
315 (w), 330 (vw), 367 (w), 380 (w), 445 (m), 536 (mw), 561 (mw),
572 (w), 727 (s), 756 (w), 833 (m), 973 (vs), 1074 (w, sh), 1131
(m, sh), 1164 (ms), 1222 (vs), 1244 (vs), 1274 (vs), 1298 (s), 1352
(ms) cm−1.
PI3+); 31P NMR (101 MHz, CD2Cl2, −50 ◦C): d 128.0 (d, JPP
=
323.4 Hz, P2I5+), −156.9 (d, JPP = 323.4 Hz, P2I5+), −182.8
(s, m1/2 = 40 Hz, AsI2–PI3+).
Preparation in the NMR tube. Ag[Al(OC(CF3)3)4] (0.503 g,
0.434 mmol) and AsBr3 (0.137 g, 0.434 mmol) were transferred
into a NMR tube glass-blown onto a J. Young valve followed by
condensing 1 ml of CD2Cl2 onto the mixture at −78 ◦C. At this
temperature Br2 (0.069 g, 0.022 ml, 0.434 mmol) was added
to the mixture with a calibrated 100 ll syringe. The NMR
tube was flame-sealed under vacuum and stored overnight at
−78 ◦C. Before the spectra were recorded, the tube was warmed
Reaction (2) (with
6 equiv. of AsI3 in CH2Cl2–CS2).
Ag[Al(OR)4] (0.902 g, 0.773 mmol) was loaded in one bulb of
the two bulbed vessel followed by CH2Cl2 (5 ml). AsI3 (1.758 g,
3.865 mmol) and PI3 (0.320 g, 0.773 mmol) were loaded into the
other bulb followed by CH2Cl2–CS2 mixture in 3 : 1 ratio (15 ml).
However, even at room temperature the AsI3 remained mostly
undissolved. Both bulbs of the vessel were cooled at −78 ◦C and
solution of the silver salt was poured through the frit plate onto
the red susp◦ension of AsI3 and PI3. Precipitation of AgI started
below −30 C and the solution became more orange in color.
The reaction was stirred at about −30 ◦C for 6 h and stored for
4 days in the refrigerator at −30 ◦C. After this, the solvent was
completely removed in vacuo. About 5 ml CH2Cl2 was added to
the reaction mixture. The resulting solution was cooled to about
−20 ◦C and filtered at this temperature to get rid off the excess
AsI3. The solution was concentrated to about a half and stored
at −80 ◦C. At this temperature after 2 days microcrystalline
orange–reddish solids precipitated (fraction I, together with
characteristic red crystals of AsI3). The residual solution was
evaporated in such a way that the at −80 ◦C precipitated solids
were not contaminated, giving a yellow–orange residue (fraction
II). The IR and NMR spectra of both fractions were recorded.
The separation of the I2As–PI3+ from P2I5+ salt was unsuccessful
and both fractions contain both products but in a different ratio.
It should be noted that due to the use of CS2 (lowering of the
polarity of the solution) a partial anion decomposition occurred
(see 13C NMR).
◦
to −20 C, shaken a few times until the precipitation of AgBr
appeared to be complete.
13C NMR (63 MHz, CD2Cl2, −30 ◦C): d 120.9 (q, JCF
=
292.7 Hz, CF3); 27Al NMR (78 MHz, CD2Cl2, +25 ◦C): d 36.0
(s, m1/2 = 15 Hz); 75As NMR (51 MHz, CD2Cl2, −70 ◦C): d
−148.2 (s, m1/2 = 983.2 Hz).
Formation of [(RO)2AlF(THF)]2. Ag[Al(OR)4] (1.077 g,
0.862 mmol) and AsBr3 (0.588 g, 1.724 mmol) were transferred
into a two bulbed vessel under argon atmosphere. The mixture
was kept at room temperature. The reaction started immedi-
ately after mixing the starting materials (occasionally shaken)
resulting in a softening of the mixture simultaneously with
changing of the color. The slightly brown mixture became liquid
for a short time and than solidified again with precipitation of
AgBr. The mixture was extracted three times with pentane and
concentrated. Thereby crystals of AsBr3 formed. After removal
of all volatiles from the extract a white residue remained, that
got dark within a few minutes. After removing all volatiles the
residue weighted 0.836 g (loss of weight 49%). The residue was
dissolved in 3 : 1 CH2Cl2–THF and stored at −30 ◦C. Colorless
crystals of [(RO)2AlF(THF)]2 were grown from this solution and
isolated (0.473 g, 88% yield based on Al in Ag[Al(OR)4]).
1H NMR (250 MHz, CD2Cl2, +25 ◦C): d 4.28 (m, O–CH2,
THF), 2.13 (m, CH2, THF); 13C NMR (63 MHz, CD2Cl2,
+25 ◦C): d 121.1 (q, JCF = 290.8 Hz, CF3), 74.5 (s, –OCH2–,
THF), 73.7 (s, OCH2, THF), 25.5 (s, CH2, THF), 25.3 (s, CH2,
THF).
Fraction I: 13C NMR (63 MHz, CD2Cl2, −30 ◦C): d 120.4 (q,
JCF = 290.9 Hz, CF3 of decomposed anion), 120.9 (q, JCF
=
291.7 Hz, CF3 of [Al(OR)4]−); 31P NMR (101 MHz, CD2Cl2,
◦
−30 C): d 129.5 (d, JPP = 324.5 Hz, P2I5+), −157.0 (d, JPP
=
324.6 Hz, P2I5+), −183.3 (s, m1/2 = 77 Hz, AsI2–PI3+). Fraction II:
13C NMR (63 MHz, CD2Cl2, −30 ◦C): d 120.4 (q, JCF = 290.7 Hz,
CF3 of decomposed anion), 120.9 (q, JCF = 291.9 Hz, CF3 of
not decomposed anion); 31P NMR (101 MHz, CD2Cl2, −30 ◦C):
d 129.5 (d, JPP = 324.4 Hz, P2I5+), −157.0 (d, JPP = 324.5 Hz,
P2I5+), −183.3 (s, m1/2 = 80 Hz, AsI2–PI3+). IR: 223 (m, P2I5+),
237 (w, I2As–PI3+), 247 (m, I2As–PI3+), 290 (m, [Al(OR)4]−),
315 (m, [Al(OR)4]−), 334 (m, P2I5+), 364 (s, P2I5+), 375
(s, I2As–PI3+), 402 (s, P2I5+), 408 (vs, I2As–PI3+), 412 (s, P2I5+),
X-Ray crystal structure determinations
Data collection for X-ray structure determinations (Table 6)
were performed on a STOE IPDS II diffractometer using
440 (m, overlapping bands of I2As–PI3+, P2I5 and [Al(OR)4]−
+
˚
graphite-monochromated Mo–Ka (0.71073 A) radiation. Single
vibration at 446 cm−1), 446 (s, [Al(OR)4]−), 537 (s, [Al(OR)4]−),
560 (m, [Al(OR)4]−), 579 (m, [Al(OR)4]−), 636 (m, [(RO)3Al–
F–Al(OR)3]−), 728 (s, [Al(OR)4]−), 755 (w, [Al(OR)4]−), 831
(m, [Al(OR)4]−), 864 (m, [(RO)3Al–F–Al(OR)3]−), 975 (vs,
[Al(OR)4]−), 1136 (s, [Al(OR)4]−), 1161 (s, [Al(OR)4]−), 1216 (vs,
[Al(OR)4]−), 1248 (vs, [Al(OR)4]−), 1258 (vs, [Al(OR)4]−), 1275
(vs, [Al(OR)4]−), 1305 (s, [Al(OR)4]−), 1350 (s, [Al(OR)4]−).
crystals were mounted in perfluoroether oil on top of a
glass fiber and then brought into the cold stream of a low-
temperature device so that the oil solidified. All calculations
were performed on PC’s using the SHELX97 software package.
The structures were solved by direct methods and successive
interpretation of the difference Fourier maps, followed by least-
squares refinement. The crystals of 1 were racemically twinned.
Solving structure in the Pna21 space group and applying the
appropriate twin law gave in the final refinement R1 = 0.0394
(wR2 = 0.0861) see table below. Six (–CF3) groups of the anion
in 1 had to be split over two positions and the disordered atoms
were included in the refinement giving an occupation number of
30% for the disordered atoms. The positions of all CF3 groups
were fixed with SADI restraints. The hydrogen atoms of the
coordinated THF molecule in [(RO)2AlF(THF)]2 were found in
Synthesis of AsBr4+[Al(OR)4]−:
Typical large-scale synthesis. Ag[Al(OC(CF3)3)4] (0.368 g,
0.317 mmol) was weighed into a two bulbed vessel. AsBr3
(0.100 g, 0.317 mmol) and Br2 (0.051 g, 0.317 mmol) followed
◦
by CH2Cl2 (10 ml) were added at −78 C. The flask was kept
◦
at −78 C and was occasionally shaken until the precipitation
◦
of AgBr started. The flask was stored in a −80 C freezer for
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3
1 2 1 1

View Article Online
Table 6 X-Ray structure data for AsBr4+[Al(OR)4]− and [(RO)2AlF(THF)]2
AsBr4+[Al(OR)4]−
[(RO)2AlF(THF)]2
Crystal size/mm
Crystal system
Space group
0.3 × 0.1 × 0.1
0.2 × 0.6 × 0.8
Orthorhombic
Monoclinic
Pna21
19.062(4)
13.874(3)
13.588(3)
90
90
90
3593.6(1)
P21/n
9.652(2)
12.727(3)
15.596(3)
90
96.34(3)
90
1904.1(7)
˚
a/A
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
3
˚
V/A
Z
4
2
Dc/Mg m−3
2.517
5.634
2.052
0.304
l/mm−1
Max., min. trans.
Index ranges, hkl
2h/◦
0.3910, 0.5391
—
−21 to 21, −17 to 16, −16 to 16
−9 to 11, −13 to 15, −18 to 19
51.9
51.7
T/K
200(2)
22082
6572
3485
100(2)
7799
3426
2749
Refl. collected
Refl. unique
Refl. observed (2r)
Rint
0.0707
0.0411
GOOF/GOOF restrained
Final R/wR2 (2r)
0.855/0.852
0.0394/0.0861
0.0925/0.1022
0.529
1.009/1.009
0.0301/0.0710
0.0435/0.0750
0.266
Final R/wR2 (all data)
−3
˚
Larg. res. peak/e A
the difference Fourier map and included isotropically into the
refinement. All other atoms were refined anisotropically.
CCDC reference numbers 256551 and 256552.
See http://www.rsc.org/suppdata/dt/b4/b417629d/ for cry-
stallographic data in CIF or other electronic format.
9 F. Claus, M. Glaser, V. Woelfel and R. Minkwitz, Z. Anorg. Allg.
Chem., 1984, 517, 207–214.
10 V. F. Claus and R. Minkwitz, Z. Anorg. Allg. Chem., 1983, 501, 19–26.
11 R. Minkwitz and H. Prenzel, Z. Anorg. Allg. Chem., 1986, 534, 150–
152.
12 R. Minkwitz, H. Prenzel, A. Schardey and H. Oberhammer, Inorg.
Chem., 1987, 26, 2730–2732.
Computational details
13 A. F. Holleman, E. Wiberg and N. Wiberg,Lehrbuch der Anorganis-
chen Chemie, Walter de Gruyter, Berlin and New York, 101st edn.,
1995.
All computations were done with the program
TURBOMOLE.64,65 The geometries of all species were
optimized at the (RI-)MP2 level46 with the triple f valence
polarization (two d and one f functions) TZVPP basis set.44,45
The 28 and 46 electron cores of Ag and I were replaced
by a quasi relativistic effective core potential.47 All species
were also fully optimized at the BP86/SV(P) (DFT-)level66–70
albeit these geometries are not shown. Approximate solvation
energies (CH2Cl2 solution with er = 8.93) were calculated with
the COSMO model71 at the BP86/SV(P) (DFT-)level using
the MP2/TZVPP geometries. Frequency calculations were
performed for all species and structures represent true minima
without imaginary frequencies on the respective hypersurface.
For thermodynamic calculations the zero point energy and
thermal contributions to the enthalpy and the free energy at
298 K have been included.72 The calculation of the thermal
contributions to the enthalpy and entropic contributions to
the free energy were done with TURBOMOLE. For all species
a modified Roby–Davidson population analysis based on
occupation numbers (Paboon) has been performed using the
(RI-)MP2/TZVPP electron density.
14 PI5 is present in many text books as “compound in bottle”, however,
we found that its characterization is limited only to the melting point!
According to our unpublished experimental results, PI5 could not be
obtained by the route by which it was claimed to be synthesized
(starting from POCl3 and LiI, see ref. 13). This will be the subject of
a paper, currently in preparation.
15 B. W. Tattershall and N. L. Kendall, Polyhedron, 1994, 13, 1517–
1521.
16 M. Baudler, D. Grenz, U. Arndt, H. Budzikiewicz and M. Feher,
Chem. Ber., 1988, 121, 1707–1709.
17 K. B. Dillon and B. Y. Xue, Inorg. Chim. Acta, 2001, 320, 172–173.
18 C. Aubauer, M. Kaupp, T. M. Klapo¨tke, H. No¨th, H. Piotrowski, W.
Schnick, J. Senker and M. Suter, J. Chem. Soc., Dalton Trans., 2001,
1880–1889.
19 C. Aubauer, M. Klapo¨tke and A. Schulz, Int. J. Vib. Spec., 1999, 3.
20 C. Aubauer and T. M. Klapo¨tke, Int. J. Vib. Spec., 2001, 5.
21 M. Kaupp, C. Aubauer, G. Engelhardt, T. M. Klapo¨tke and O. L.
Malkina, J. Chem. Phys., 1999, 110, 3897–3902.
22 S. Pohl, Z. Anorg. Allg. Chem., 1983, 498, 15–19.
23 I. Tornieporth-Oetting and T. Klapo¨tke, J. Chem. Soc., Chem.
Commun., 1990, 132–133.
24 M. Gonsior, I. Krossing, L. Mu¨ller, I. Raabe, M. Jansen and L. Van
Wullen, Chem. Eur. J., 2002, 8, 4475–4492.
25 I. Krossing and I. Raabe, Angew. Chem., Int. Ed., 2001, 40, 4406–
4409.
References
26 G. S. H. Chen and J. Passmore, J. Chem. Soc., Dalton Trans., 1979,
1251–1256.
1 J. Petersen, E. Lork and R. Mews, Chem. Commun., 1996, 1897–1898.
2 M. Binnewies, M. Ja¨ckel and H. Willner, Anorganische und All-
gemeine Chemie, Spektrum Akademischer Verlag, Heidelberg and
Berlin, 1st edn., 2004.
3 I. R. Beattie and K. M. S. Livingston, J. Chem. Soc. A, 1969, 5,
859–860.
4 F. J. J. Brinkmann, H. Gerding and K. Olie, Recl. Trav. Chim. Pays-
Bas, 1969, 88, 1358–1360.
5 S. Haupt and K. Seppelt, Z. Anorg. Allg. Chem., 2002, 628, 729–734.
6 K. Seppelt, Z. Anorg. Allg. Chem., 1977, 434, 5–15.
7 K. Seppelt, Angew. Chem., 1976, 88, 410–411.
8 F. Claus, M. Glaser and R. Minkwitz, Z. Anorg. Allg. Chem., 1983,
506, 178–184.
27 G. S. H. Chen and J. Passmore, J. Chem. Soc., Chem. Commun., 1973,
559.
28 R. Minkwitz, J. Nowicki and H. Borrmann, Z. Anorg. Allg. Chem.,
1991, 596, 93–98.
29 M. Gerken, P. Kolb, A. Wegner, H. P. A. Mercier, H. Borrmann, D. A.
Dixon and G. J. Schrobilgen, Inorg. Chem., 2000, 39, 2813–2824.
30 I. Tornieporth-Oetting and T. Klapo¨tke, Angew. Chem., 1989, 101,
1742–1744.
31 M. Broschag, T. M. Klapo¨tke and I. C. Tornieporth-Oetting, J. Chem.
Soc., Chem. Commun., 1992, 446–448.
32 R. E. Pabst, M. C. Sharpe, J. L. Margrave and J. L. Franklin,
Int. J. Mass Spectrom. Ion Phys., 1980, 33, 187–199.
1 2 1 2
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3

View Article Online
33 H. L. Sievers, H.-F. Gru¨tzmacher, H. Gru¨tzmacher and S. Pitter,
J. Am. Chem. Soc., 1995, 117, 2313–2320.
34 C. Aubauer, G. Engelhardt, T. M. Klapo¨tke and A. Schulz, J. Chem.
Soc., Dalton Trans., 1999, 1729–1734.
35 I. Krossing, J. Chem. Soc., Dalton Trans., 2002, 500–512.
36 A. Bihlmeier, M. Gonsior, I. Raabe, N. Trapp and I. Krossing, Chem.
Eur. J., 2004, 10, 5041–5051.
37 I. Krossing, H. Brands, R. Feuerhake and S. Koenig, J. Fluorine
Chem., 2001, 112, 83–90.
38 I. Krossing, J. Am. Chem. Soc., 2001, 123, 4603–4604.
39 I. Krossing, Chem. Eur. J., 2001, 7, 490–502.
40 I. Krossing and I. Raabe, Angew. Chem., Int. Ed., 2004, 43, 2066–
2090.
41 M. Gonsior, I. Krossing and E. Matem, Eur. J. Chem., submitted.
42 H. D. B. Jenkins, I. Krossing, J. Passmore and I. Raabe, J. Fluorine
Chem., 2004, 125, 1585–1592.
43 I. Krossing and I. Raabe, Chem. Eur. J., 2004, 10, 5017–5030.
44 A. Scha¨fer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97, 2571.
45 A. Scha¨fer, C. Huber and R. Ahlrichs, J. Chem. Phys., 1994, 100,
5829.
55 T. Klapo¨tke, J. Passmore and E. G. Awere, J. Chem. Soc., Chem.
Commun., 1988, 1426–1427.
56 A. L. Rheingold and P. J. Sullivan, Organometallics, 1983, 2, 327–331.
57 U. Mu¨ller and H. Sinnig, Angew. Chem., 1989, 101, 187–188.
58 I. Krossing and A. Reisinger, unpublished results.
59 For the reaction between AsBr3, P4 and Ag[Al(OR)4] in solution
a complex [(P4)Ag(Br2AsBr)]+ should be formed as a intermediate
leading to [P4AsBr2]+. This complex, however could undergo a dis-
mutation reaction into [Ag(P4)2]+ and [Ag(Br2AsBr)2]+ according to
the equation: 2[(P4)Ag(Br2AsBr)]+ → [Ag(P4)2]+ + [Ag(Br2AsBr)2]+.
DG◦(CH2Cl2) = +9 kJ mol−1 of this reaction suggests that all the
species are present in the solution. Since the reaction of AsBr3 with
silver salt (formation of [Ag(Br2AsBr)2]+) decomposes the anion at
−30 ◦C, it seems to be very reasonable that the [(P4)Ag(Br2AsBr)]+
complex formed at low temperature does not react below −30 ◦C
and the consumed [Ag(Br2AsBr)2]+ shifted the equilibrium of the
reaction completely to the side of the dismutation products.
+
60 The energies of the formation of EX2 cations were assessed
considering all the species to be in solution (solvated). Additionally,
+
we performed an estimation of the EX2 formation in a suitable
46 F. Weigend and M. Ha¨ser, Theor. Chim. Acta, 1997, 97, 331.
47 W. Kuechle, M. Dolg, H. Stoll and H. Preuss, Mol. Phys., 1991, 74,
1245.
Born–Haber cycle including the sublimation enthalpies of EX3. The
obtained “new” values of the DG◦(CH2Cl2) of formation of EX2
,
+
however, differ only very slightly and do not account for the differing
+
48 The sublimation energies for the formation of solid silver halide were
behavior of EX2 cations (see ESI†).
included. The sublimation energy for the process AgI(g) →AgI(s)
=
61 G. Oddo and U. Giachery, Gazz. Chim. Ital., 1924, 53, 56.
62 P. G. Patfrnosto, Rev. Facultad Cienc. Quim. (Univ. La Plata), 1930,
7, 43–46.
63 J. C. Bailar, Jr., Inorg. Synth., 1939, 35, 103–104.
64 R. Ahlrichs, M. Ba¨r, M. Ha¨ser, H. Horn and C. Ko¨lmel, Chem. Phys.
Lett., 1989, 162, 165.
−196.9 kJ mol−1 and for AgBr(g) →AgBr(s) = −217.4 kJ mol−1
.
49 T. S. Cameron, J. Passmore and X. Wang, Angew. Chem., Int. Ed.,
2004, 43, 1995–1998.
50 The [Al(OR)4]− anion has the usual geometry with Al–O bond
˚
lengths at 1.687(2) A on average and Al–O–C angles that range
from 153.7(1) to 162.9(1)◦. Therefore, the structure of the anion is
affected by librational motion, but correcting all bond distances for
librations by a riding model gives the normal [Al(OR)4]− values of
65 M. v. Arnim and R. Ahlrichs, J. Chem. Phys., 1999, 111, 9183.
66 A. D. Becke, Phys. Rev. A: At. Mol. Opt. Phys., 1988, 38, 3098–3100.
67 K. Eichkorn, O. Treutler, H. Oehm, M. Haeser and R. Ahlrichs,
Chem. Phys. Lett., 1995, 242, 652–660.
+
−
˚
1.731–1.740 A (in [Ag(C2H4Cl2)3] [Al(OR)4] the non corrected Al–
˚
O bond lengths are 1.714(3)–1.736(3) A).
68 J. C. Slater, Phys. Rev., 1951, 81, 385–390.
51 M. Gonsior, I. Krossing and N. Mitzel, Z. Anorg. Allg. Chem., 2002,
69 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200–
628, 1821–1830.
1211.
52 C. Cui, H. W. Roesky, M. Noltemeyer, M. F. Lappert, G. Schmidt
and H. Hao, Organometallics, 1999, 18, 2256–2261.
53 C. Rennekamp, H. Wessel, H.W. Roesky, P. Mu¨ller, H.-G. Schmidt,
M. Noltemeyer, I. Uso´n and A. R. Barron, Inorg. Chem., 1999, 38,
5235–5240.
70 J. P. Perdew, Phys. Rev. B, 1986, 33, 8822.
71 A. Klamt and G. Schu¨rmann, J. Chem. Soc., Perkin. Trans., 1993, 2,
799.
72 Thermal and entropic contributions to the enthalpy and free energy
were obtained with TURBOMOLE software package with an
exception of Ag+ for which the above contribution were calculated
with Gaussian98W at the semiempirical PM3 level.
54 T. Klapo¨tke and J. Passmore, J. Chem. Soc., Dalton Trans., 1990,
3815–3822.
D a l t o n T r a n s . , 2 0 0 5 , 1 2 0 3 – 1 2 1 3
1 2 1 3