Aza-nido-dodecaboranes and -borates from aza-closo-dodecaboranes by the addition of neutral or anionic bases: Mechanism

Article abstract of DOI:10.1002/1521-3749(200203)628:3<632::AID-ZAAC632>3.0.CO;2-8

The addition of neutral (L = py, NEt₃, NHEt₂, NH₂tBu) and anionic Lewis bases (X^- = OH^-, Br^-, N₃^-, Me^-, NHBu^-, NHtBu^-, [FeCp(CO)₂

Full text of DOI:10.1002/1521-3749(200203)628:3<632::AID-ZAAC632>3.0.CO;2-8

Aza-nido-dodecaboranes and -borates from Aza-closo-dodecaboranes by the
Addition of Neutral or Anionic Bases: Mechanism
Peter Paetzold*, Petra Lomme, and Ulli Englert
Aachen, Institut für Anorganische Chemie der Technischen Hochschule
Received October 4th, 2001.
Dedicated to Professor Hans Bürger on the Occasion of his 65th Birthday
Abstract. The addition of neutral (L ϭ py, NEt3, NHEt2, NH2tBu)
and anionic Lewis bases (XϪ ϭ OHϪ, BrϪ, N3Ϫ, MeϪ, NHBuϪ,
NHtBuϪ, [FeCp(CO)2]Ϫ) to aza-closo-dodecaboranes RNB11H11
(1) or to derivatives thereof with boron bound non-hydrogen li-
gands yields nido-clusters RNB11H11L or [RNB11H11X]Ϫ or deriva-
tives thereof, respectively, the non-planar pentagonal aperture
NϪB4ϪB9ϪB8ϪB5 of which hosts a B8ϪB9 hydrogen bridge. The
base is either bound to B8 (3) or B4 (5) or B2 (7). The structures
identified. The position of boron-bound ligands during the addi-
tion of bases to 1 shows, that only vertices of the ortho-belt of 1
are involved in the opening process. A mechanism is made plausible
that starts by the attack of the base at B2 of 1 and opening of the
N-B2 bond, denoted as a [3c,1c]-collocation, to give 2 with an endo-
H atom, whose migration into an adjacent bridge position and ope-
ning of a second BϪN bond, denoted as a [3c,2c]-translocation,
gives 3; this isomer can be transformed into 7 by a sequence of
[3c,2c]-translocations via the isomers 4, 5, and 6. The chiral type 3
species MeNB11H11L with L ϭ NHEt2, NH2tBu undergo a rapid
enantiomerization, for whose mechanism the exchange of the brid-
ging and the amine-H atom has been made plausible.
1
of these adducts are concluded from H and 11B NMR including
2D-NMR spectra, and in the case of MeNB11H11(NHEt2) (type
3) also by a crystal structure analysis. With two of the adducts
MeNB11H11L (L ϭ py, NHEt2), isomers of the type 3, 5, and 7,
and with two of the adducts, MeNB11H11(NH2tBu) and
{MeNB11H11[FeCp(CO)2]}Ϫ, isomers of the type 3 and 7 could be
Keywords: Boron; Azadodecaboranes; Crystal Structure
Aza-nido-dodecaborane und -borate aus Aza-closo-dodecaboranen durch die Addition
neutraler und anionischer Basen: Mechanismus
Inhaltsübersicht. Die Addition neutraler (L ϭ py, NEt3, NHEt2,
NH2tBu) und anionischer Lewis-Basen (XϪ ϭ BrϪ, N3Ϫ, MeϪ,
NHBuϪ, NHtBuϪ, [FeCp(CO)2]Ϫ) an Aza-closo-dodecaborane
RNB11H11 (1) oder an deren Derivate mit Bor-gebundenen Nicht-
wasserstoff-Liganden ergibt nido-Cluster RNB11H11L bzw.
[RNB11H11X]Ϫ bzw. Derivate davon, in deren nicht-planar penta-
gonaler Öffnung NϪB4ϪB9ϪB8ϪB5 die Atome B8 und B9 von
einem H-Atom überbrückt sind. Die Base ist entweder an B8 (3)
oder B4 (5) oder B2 (7) gebunden. Die Struktur dieser Addukte
ergibt sich aus 1H- und 11B-NMR- einschließlich entsprechender
2D-NMR-Spektren, im Falle von MeNB11H11(NHEt2) (Typ 3)
auch durch eine Kristallstrukturanalyse. Bei zwei Addukten
MeNB11H11L (L ϭ py, NHEt2) konnten alle drei Isomere vom Typ
3, 5 und 7, bei zwei Addukten, nämlich MeNB11H11(NH2tBu) und
{MeNB11H11[FeCp(CO)2]}Ϫ, zwei Isomere vom Typ 3 und 7 nach-
gewiesen werden. Die Stellung Bor-gebundener Liganden bei der
Addition von Basen an 1 zeigt, daß nur die ortho-B-Atome von 1
an der Clusteröffnung beteiligt sind. Ein Mechanismus stimmt mit
allen Beobachtungen überein, bei dem zunächst die Base am Atom
B2 von 1 unter Öffnung der Bindung N-B2 angreift, d.i. eine sog.
[3c,1c]-Collocation, bei der ein Cluster 2 mit einem endo-H-Atom
entsteht, dessen Wanderung in eine benachbarte Brückenposition
unter Öffnung einer weiteren B-N-Bindung im Rahmen einer sog.
[3c,2c]-Translocation zu 3 führt; dieses Isomer kann im Zuge weite-
rer [3c,2c]-Translocationen über die Isomeren 4, 5 und 6 in das
Isomer 7 übergehen. Die chiralen Verbindungen MeNB11H11L mit
L ϭ NHEt2, NH2tBu vom Typ 3 unterliegen einer schnellen Enan-
tiomerisierung, für deren Mechanismus der Austausch eines Brük-
ken- und eines Amin-H-Atoms plausibel gemacht wurde.
Introduction
belt) and the six non-adjacent vertices (B7ϪB11: meta-belt;
B12: para-B vertex). The acid-catalyzed electrophilic substi-
tution of H atoms is exclusively restricted to H7ϪH12 [1,2],
whereas a nucleophilic H exchange is possible for H2ϪH6;
The chemistry of the aza-closo-dodecaboranes RNB11H11 is
characterized by a pronounced electronic difference bet-
ween the five vertices adjacent to nitrogen (B2-B6: ortho-
a
typical example is presented by the reaction of
MeNB11H11 with excess iodine chloride, that gives
HNB11Cl5I6 by nucleophilic attack of Cl in position 2Ϫ6
and electrophilic attack of I in position 7Ϫ12 [3]. Neutral
bases L or anionic bases XϪ can add their lone pair to one
of the ortho-BH vertices under opening of two neighbou-
ring BϪN bonds, accompanied by the migration of the cor-
responding exo-H atom into a bridging position. The resul-
* Prof. Dr. P. Paetzold
Institut für Anorganische Chemie
Technische Hochschule Aachen
D-52056 Aachen, Germany
E-mail: peter.paetzold@ac.rwth-aachen.de
632
WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002 0044Ϫ2313/02/628/632Ϫ640 $ 17.50ϩ.50/0 Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640

Aza-nido-dodecaboranes and -borates from Aza-closo-dodecaboranes
1a MeNB11H11 [5]
1b PhNB11H11 [4]
1c MeNB11H10Br (Br in pos. 12) [1]
1d MeNB11H10I (I in pos. 12) [1]
1e MeNB11H5Me5(OTrf) (Me in pos. 7Ϫ11; triflate in pos. 12) [1]
1f PhNB11H5Me5(OTrf) (like 1e) [see Exp. Sect.]
1g NB11H10(BH2NEt2CH2CH2) (the BϪNϪCϪC chain connects
N and B2) [6]
Anionic Bases
According to a principally known route [4], we synthesized
the colourless solids [NH3Bu][MeNB11H11(NHBu)],
Fig. 1 Structure and systematic numbering of the hypothetical
twelve-vertex nido-anion [RNB11H11]2Ϫ (pentagonal aperture in
bold lines) and schematic representation of its characteristic upper
fragment in the case of RNB11H11L and [RNB11H11X]Ϫ
[NH3tBu][PhNB11H11(NHtBu)],
[NEt4][PhNB11H11Br],
and [N(PPh3)2][PhNB11H11(N3)] from the aza-closo-bora-
nes 1a and 1b and excess amines NH2Bu and NH2tBu or
stoichiometric amounts of [NEt4]Br or [N(PPh3)2]N3. All
of the four product anions belong to type 7 (Fig. 2) with
Ϫ
NHBuϪ, NHtBuϪ, BrϪ, and N3 as the anionic bases.
The structure of these nido-anions can be concluded
from the 11B NMR spectra (Table 1). Seven shift values,
four of them with twofold intensity, are in accord with Cs
symmetry, with B2, B12, and B13 on the mirror plane.
Apart from the singlet peak of atom B2, which carries the
ligands X, all the other peaks appear as doublets because
of BϪHexo coupling. The shift values are found in a typical
range (Table 1) and this range corresponds to the values of
ten already known type 7 anions [4,5], including the one
characterised by X-ray structural analysis (R ϭ Me, X ϭ
OMe [5]). Moreover, the 2D-11B/11B NMR spectrum of the
anion [PhNB11H11Br]Ϫ supports the proposed structure. As
in several of such spectra, discussed later, the cross-peaks
for BϪB coupling at the pentagonal aperture are not found.
Moreover, the B2ϪB4 (B2ϪB5) correlation cannot be seen
in any of the products of Table 1, a normal situation for
the BϪB bond of NBB faces in azaborane clusters.
Fig. 2 Proposed mechanism of the opening of closo-RNB11H11 (1)
by bases (see Fig. 1 for type 3 and 7)
The 1:1 reaction of 1a and Na[FeCp(CO)2] in thf gives a
ting nido-NB11 skeleton (Fig. 1) hosts neutral bases in posi-
tion 8 and anionic bases in position 2; the corresponding
examples PhNB11H11(NEt3) [4] and [MeNB11H11(OMe)]Ϫ
[5] had been structurally characterised.
2:3
mixture
of
two
isomeric
products
Na{MeNB11H11[FeCp(CO)2]} after 12 h. On standing, the
amount of the second isomer increases; after 3 d the ratio
is 1:3, and finally the peaks of the first isomer disappear.
The second isomer is of type 7, proven by the same NMR
arguments as just given (Table 1). Again, the cross-peak
B4ϪB9 (and symmetrically equivalent: B5ϪB8) is not
found. Ϫ The first isomer belongs to type 3. It is the first
example of this structure type in the case of anionic nido-
NB11 clusters. Eleven 11B NMR signals, one singlet (B8)
and ten doublets, are found. The chemical shifts are compa-
rable to those of the neutral type 3 cluster 1a·NEt3 (Table
1), which is structurally well characterized [4]; the shift va-
lue of B8, however, deviates considerably from the B8 shift
series in Table 1, because of the particular nature of the
ligand in position 8, certainly.
A mechanism for the reaction of closo-RNB11H11 with
bases had been proposed [2], which leaves the base in a
fixed position at the primarily attacked vertex, whereas the
extra-proton jumps from endo into bridging positions and
vice versa, and the RN apex builds and breaks BϪN bonds
around the pentagon B2ϪB5ϪB8ϪB9ϪB4 (i.e. the ortho-
belt B2ϪB3ϪB4ϪB5ϪB6 in the former closo-structure),
until the final isomer 7 is formed via the endo-H containing
isomers 2, 4, 6 and the bridge-H containing isomers 3 and
5 (Fig. 2).
It looks in the moment, as if neutral bases generally re-
sult in type 3 and anionic base in type 7 products. In order
to get a better founded insight into the opening process of
closo-clusters 1, we extended the opening process to further
neutral and anionic bases, and we started from the simple
type 1 azaboranes RNB11H11 (1a, b) as well as from type 1
azaboranes with substituents at the boron skeleton (1c-g):
The haloborane MeNB11H10I is opened by excess LiMe
to give an oily product Li[MeNB11H10IMe]. According to
the 11B NMR shift values (Table 1) and according to the
2D-11B/11B NMR spectrum, a type 7 product is present;
again, the cross-peak B4ϪB9 is missing. In addition to B2,
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640
633

P. Paetzold, P. Lomme, U. Englert
Table 1 11B NMR chemical shifts of the nido-adducts of 1a؊g with neutral and anionic bases of type 3, 5, and 7
type
B2
B4
B5
B6
B7
B8
B9
B10
B11
B12
B13
1a·py
1a·NHEt2
1a·NH2tBu
1b·NEt3 [4]
Na{1a[FeCp(CO)2]}
1a·py
1a·NHEt2
1g·NH2tBu
1a·py
1a·NHEt2
1a·NH2tBu
1c·NEt3
3
"
"
"
"
5
"
"
7
"
"
"
"
"
"
"
"
"
"
"
"
7.2
Ϫ2.7
Ϫ3.1
6.0
Ϫ1.1
4.5
Ϫ10.3
Ϫ2.7
Ϫ3.1
Ϫ10.6
Ϫ10.7
Ϫ0.6a)
2.4a)
Ϫ10.3
1.0
1.9
Ϫ4.3
Ϫ7.6
Ϫ17.5
Ϫ17.3
Ϫ17.3
Ϫ17.0
Ϫ21.3
Ϫ20.2
Ϫ22.8
Ϫ19.3
Ϫ19.2
Ϫ17.2
Ϫ16.5
Ϫ16.0
Ϫ14.2
Ϫ11.7a)
Ϫ18.6
Ϫ12.5
Ϫ17.3
Ϫ17.1
Ϫ19.5
Ϫ14.5
Ϫ12.5a)
Ϫ18.6
Ϫ19.7
Ϫ18.9
Ϫ19.9
Ϫ16.9
Ϫ16.6
Ϫ19.7
Ϫ16.4
Ϫ19.2
Ϫ17.2
Ϫ16.5
Ϫ16.0
Ϫ14.2
Ϫ11.7a)
Ϫ18.6
Ϫ12.5
Ϫ17.3
Ϫ17.1
Ϫ19.5
Ϫ14.5
Ϫ12.5a)
17.1a)
14.7a)
11.7a)
19.5a)
28.4a)
9.8
11.8
12.5
12.5
3.9
0.3
4.2
2.6
Ϫ0.5
4.9
9.0
1.0
1.9
13.0
14.3
13.3
11.8
13.3
12.5
3.9
0.3
4.2
2.6
Ϫ0.5
4.9
Ϫ18.6
Ϫ19.7
Ϫ18.9
Ϫ19.9
Ϫ16.4
Ϫ14.5
Ϫ19.7
Ϫ15.9
Ϫ14.6
Ϫ6.7
Ϫ10.8
Ϫ12.1
Ϫ11.0
Ϫ3.4a)
Ϫ12.9
Ϫ13.8
Ϫ11.1
Ϫ13.3
Ϫ12.5
Ϫ12.5
Ϫ3.0a)
Ϫ14.2
Ϫ22.7
Ϫ22.5
Ϫ18.4
Ϫ15.9
Ϫ14.5
Ϫ14.0
Ϫ19.3
Ϫ14.6
Ϫ6.7
Ϫ10.8
Ϫ12.1
Ϫ11.0
Ϫ3.4a)
Ϫ12.9
Ϫ13.8
Ϫ11.1
Ϫ13.3
Ϫ12.5
Ϫ12.5
Ϫ3.0a)
6.1
5.4
6.2
9.3
6.8
7.1
6.1
6.5
2.7
Ϫ11.1
Ϫ11.2
0.8a)
Ϫ12.1a)
Ϫ4.2a)
Ϫ5.6
3.0
Ϫ17.3
5.3
Ϫ17.5
Ϫ7.3a)
7.2a)
Ϫ28.3
Ϫ22.7
Ϫ22.5
Ϫ28.7
Ϫ26.5
Ϫ30.6
Ϫ29.8
Ϫ30.5
Ϫ29.6
Ϫ34.1
Ϫ34.4
Ϫ31.9
Ϫ31.6
Ϫ25.7a)
Ϫ33.3
Ϫ30.3
Ϫ35.5
Ϫ30.7
Ϫ32.7
Ϫ28.5
Ϫ25.3a)
Ϫ13.7
Ϫ12.4
Ϫ7.5a)
Ϫ15.0
Ϫ11.1
Ϫ12.5
Ϫ12.1
Ϫ15.8
Ϫ13.2
Ϫ12.9
Ϫ12.5
Ϫ11.1
Ϫ10.0
Ϫ11.1
Ϫ10.0
Ϫ12.1
Ϫ0.4
4.5
Ϫ7.5a)
Ϫ15.0
Ϫ11.1
Ϫ12.5
Ϫ12.1
Ϫ15.8
Ϫ13.2
Ϫ12.9
Ϫ12.5
Ϫ11.1
Ϫ10.0
Ϫ11.1
Ϫ10.0
Ϫ12.1
Ϫ9.8a)
7.5a)
8.8a)
5.3a)
8.8a)
7.4a)
12.8a)
27.9a)
10.9a)
1.5a)
6.0a)
4.4a)
14.3a)
1d·NEt3
1e·NEt3
[NH3Bu][1a(NHBu)]
Na{1a[FeCp(CO)2]}
[NH3tBu][1b(NHtBu)]
[NEt4][1bBr]
[N(PPh3)2][1b(N3)]
Li[1dMe]
6.8
0.4
10.1
0.6
8.3
6.8
0.4
10.1
0.6
8.3
H[1f(OH)]
1.9
1.9
a) Singlet. b) Peak assignments for B6, B8, B10 and for B7, B9, B11 might be interchanged; exo-BH2: δ(11B) ϭ Ϫ4.6 (t, 1J ϭ 97 Hz).
a second singlet is found for B12. Assuming that the BI
vertex in the product is the same as in the starting material,
the position of the BI vertex in the product shows, that
only the ortho-belt of the closo-cluster 1a is involved in the
opening process to give 7 (Fig. 2).
The same conclusion must be drawn from the NMR
spectra of the type 7 product H[PhNB11H5Me5(OTrf)-
(OH)], which is formed from closo-PhNB11H5Me5(OTrf)
(1f; see Exp. Sect.) and aqueous methanol. We had observed
the formation of a strong Brønsted acid earlier, when we
had reacted MeNB11H11 with methanol in CH2Cl2 to give
Hϩ and [MeNB11H11(OMe)]Ϫ [5]. In the present reaction,
water molecules attack the closo-cluster in spite of excess
methanol, giving a hydroxy instead of a methoxy product.
The structure of the type
7
hydroxy anion
[PhNB11H5Me5(OTrf)(OH)]Ϫ is again in accord with the
NMR spectra (Table 1). Five singlets and two doublets in
the expected intensity ratio indicate Cs symmetry. All of the
expected cross-peaks are found in the 2D-11B/11B NMR
spectrum. In the 1H NMR spectrum, a long-range coupling
(3J ϭ 8.0 Hz) between the two exo-H atoms H4 and H5
(δ ϭ 2.88; d) and the bridging H atom (δ ϭ Ϫ3.26; t) is
found. The 1H NMR signal of the OH group disappears in
the water peak (δ ϭ 5.17) because of rapid proton exchange.
The 1H NMR signals of all of the type 7 anions are found
at more or less comparable values, with the highest field for
the bridging protons (Table 2). Certainly, no exo-1H NMR
shift is present, when the base or other exo-ligands are
bound to the corresponding B atoms. The equivalencies be-
cause of Cs symmetry are certainly the same as for the 11B
NMR signals.
lene. We now followed the 1:1 reaction of MeNB11H11 (1a)
with pyridine in thf by NMR. At Ϫ50 °C, the type 3 isomer
is formed, identified by similar 11B NMR shifts as found
for the corresponding NEt3 adduct (Table 1). Again the
shift values for B7 and B10 are accidentally degenerate, and,
moreover, the values for B4 and B5 are degenerate in the
pyridine adduct. The singlet for B8 marks the position of
the base. The 2D-11B/11B NMR cross-peaks are in accord
with a type 3 isomer, though the correlations 2Ϫ6, 2Ϫ7,
4Ϫ6, 5Ϫ8 and 5Ϫ11 cannot be seen.
At about 0 °C a second isomer is formed and is present
together with the type 3 isomer in a 1:1 ratio. The new iso-
mer gives ten 11B NMR signals, one with double intensity
because of accidental degeneracy. A 2D-11B/11B NMR
spectrum of the mixture reveals all the cross-peaks expected
for the type 5 isomer, with exception of the correlations
4Ϫ9, 5Ϫ8, and 2Ϫ6. The NMR shifts of corresponding B
atoms are found at similar values in both isomers; the sin-
glets at B8 (type 3) and B4 (type 5), however, are distinctly
low-field shifted with respect to the corresponding doublets
Pyridine as Neutral Base
The addition of NMe3 and NEt3 to 1b stops at the type 3
adducts [4]. These adducts remain stable even in boiling xy-
634
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640

Aza-nido-dodecaboranes and -borates from Aza-closo-dodecaboranes
1
Table 2 H NMR chemical shifts of the nido-adducts of 1a؊g with neutral and anionic bases of type 3, 5, and 7
type µ-H
H2
H4
H5
H6
H7
H8
H9
H10 H11 H12
H13
R
base, cation
1a·py
1a·NHEt2
3
"
Ϫ0.17 2.63
Ϫ1.01 2.10
2.02 2.15 1.85 1.39
2.10 2.76 1.84 1.16
/
/
3.80
2.76
0.65
1.16
1.39
0.87
3.21
3.34
0.68 2.60
0.87 2.55
7.20, 8.80, 8.45 (2:1:2)
1.45, 3.33, 4.85 (t, q, s;
6:4:1)
1a·NH2tBu
1b·NEt3 [4]
Na{1a[FeCp(CO)2]}
1a·py
"
"
"
5
"
Ϫ1.31 2.50
Ϫ0.90 3.07
Ϫ0.80 2.20
Ϫ1.71 3.41
Ϫ2.61 2.80
2.50 1.28 1.81 1.28
2.28 2.47 2.13 1.60
1.48 1.48 2.20 2.20
/
/
/
1.28
3.66
1.71
3.83
3.75
1.28
1.60
0.84
1.16
1.55
0.90
1.00
0.84
1.16
1.25
3.40
3.29
1.71
3.43
3.28
0.90 2.50
1.00 7.08Ϫ7.32 1.15, 3.30 (t, q; 3:2)
0.13 2.21
0.47 2.18
0.42 2.51
1.52, 6.88, (9:2)
4.67 (Cp)
/
/
1.29 1.08 1.91 3.89
1.33 0.87 1.55 3.75
7.94, 8.40, 8.94 (2:1:2)
1.33, 3.33, 3,72 (t, q, s;
6:4:1)
1.42, 4.89 (9:2)
like type 5
1a·NHEt2
a)
1g·NH2tBu
1a·py
1a·NHEt2
"
7
"
Ϫ1.58 3.22
/
/
1.77 1.77 3.77
3.67
3.83
2.69
1.19
1.29
1.42
1.19
1.29
1.42
3.30
3.43
2.55
0.46
Ϫ2.48
Ϫ3.66
/
/
1.72 1.72 1.72 1.72 3.83
1.75 1.75 1.63 1.63 2.69
0.57 2.17
1.38 2.12
1.45, 3.33, 4.85 (t, q, s;
6:4:1)
1a·NH2tBu
1c·NEt3
1d·NEt3
1e·NEt3
[NH3Bu][1a(NHBu)]
Na{1a[FeCp(CO)2]}
[NH3tBu][1b(NHtBu)]
[NEt4][1bBr]
[N(PPh3)2] [1b(N3)]
Li[1dMe]
"
"
"
"
"
"
"
"
"
"
"
Ϫ4.01
Ϫ3.60
Ϫ2.62
Ϫ3.56
Ϫ3.32
Ϫ2.42
Ϫ3.71
Ϫ2.24
Ϫ3.47
Ϫ2.81
Ϫ3.26
/
/
/
/
/
/
/
/
/
/
/
1.17 1.17 1.72 1.72 2.75
1.64 1.69 2.04 2.04 3.20
1.44 1.44 1.77 1.77 3.48
2.75
3.20
3.48
1.46
3.22
3.22
2.96
3.67
3.43
3.37
1.66
1.64
1.56
1.44
/
0.89
0.84
2.44
1.13
1.15
1.25
/
1.64
1.56
1.44
/
0.89
0.84
2.44
1.13
1.15
1.25
/
1.17
/
/
/
0.15 2.00
0.34 2.15
0.04 2.10
1.31, 6.88 (9:2)
1.43, 3.28 (t, q; 3:2)
1.35, 3.03 (t, q; 3:2)
2.69 2.69
/
/
1.46
/
2.16
1.35, 3.18 (t, q; 3:2)b)
c)
1.30 1.30 1.17 1.17 3.22
1.71 1.71 1.71 1.71 3.22
1.32 1.32 1.63 1.63 2.96
1.92 1.92 1.85 1.85 3.67
2.91 2.91 1.75 1.75 3.43
1.69 1.69 1.41 1.41 3.37
2.38
0.18 2.31
2.98 Ϫ0.02 2.21
4.67 (Cp)d)
1.32
3.24
1.39
/
0.09 7.08Ϫ7.32 1.45, 4.83 (18:4) (base)
0.35 7.22Ϫ7.32 1.27, 3.13 (t, q; 3:2)
0.03 6.87Ϫ7.60e)
0.67 2.11
7.06Ϫ7.20
0.05
f)
H[1f(OH)]
2.88 2.88
/
/
1.66
/
/
a) 1.03, 2.82, 2.88 (3 m; 6:2:2, two non-equivalend ethyl with CHaHb methylene groups), 2.65, 2.96 (2 t; 1:1, C2H4), 2.05 (q; BH2). b) Additional: Ϫ0.03, 0.08,
0.19 (3 s; 2:2:1; 5 BMe). c) 0.96 (t; 2 Me), 1.40, 1.62, 1.70 (3 m; 2:1:1, 4 CH2), 3.10 (t; 2 NCH2), 5.32 (br., NH, NH3) (anionic base and cation). d) Additional:
13C NMR; δ ϭ 43.1 (Me), 84.9 (Cp), 218.0 (CO). e) 7 Ph (R, cation). f) Additional: Ϫ0.29, 0.01, 0.29 (3 s; 1:2:2; 5 BMe); 13C NMR: δ ϭ Ϫ3.8, Ϫ1.1, 1.3 (3 q;
1:2:2; BMe), 120.2 (q, 1J ϭ 317 Hz; CF3), 125.4, 125.7, 129.0, 145.9 (3 d, 1 s; Ph); 19F NMR: δ ϭ Ϫ79.2.
at B8 (type 5) and B4 (type 3), thus indicating the different
fixation of pyridine to the cluster skeleton.
an anionic base. A cluster was formed, whose 11B NMR
spectrum revealed seven signals in the intensity ratio
2:2:1:2:2:1:1 (from high to low field), indicating Cs symme-
try. Most of the type 7 products of Table 1 give the ratio
1:2:2:1:2:2:1; the high-field shift maximum corresponds to
only one B atom, B13, for all of the type 7 isomers. Our
low-temperature adduct 1a·NHEt2 cannot exhibit a type 7
structure. The second symmetric isomer in the hypothetical
reaction sequence of Fig. 2 is the type 2 isomer. Actually,
we found a broad 11B NMR singlet at δ ϭ 14.7, that corre-
Standing at room temperature for a few hours makes a
third isomer observable. The final ratio of all three of the
isomers is 1:1:1, and this ratio remains constant on addition
of excess pyridine or on heating the mixture up to 100 °C.
As a typical type 7 cluster, the third isomer shows only se-
ven 11B NMR peaks in the expected intensity ratio (Table
1), whose chemical shifts correspond well to those of known
type 7 clusters, but are also not far away from those of the
type 3 and type 5 isomer; this is not true for the B2 singlet
of the type 7 isomer, which is high-field shifted with respect
to the B2 doublets of the two other isomers. The correla-
tions 4/5Ϫ8/9, 8/9Ϫ13, and 12Ϫ13 are not found in the 2D-
11B/11B NMR spectrum for the isomer with Cs symmetry.
1
sponds to a H NMR signal at δ ϭ 1.01 in the 2D-1H/11B-
HMQC NMR spectrum; this proton signal can be resolved
as a quartet with a coupling constant, 1J ϭ 80 Hz, too small
for a BϪHexo coupling, but possibly representing BϪHendo
of a type 2 cluster. We could interpret all of the 2D-11B/11B
NMR cross peaks in terms of a type 2 structure (Fig. 3)
with the correlations 2/4Ϫ5/9 (nitrogen bridging), 6Ϫ7/10
(11B NMR signals close together) and 5/9Ϫ8 as the only
missing ones.
1
All these findings are supported by three sets of H NMR
signals for the cluster-H atoms (Table 2).
The constant isomer ratio of 1:1:1 indicates, that the
three isomers with the base pyridine in position 3, 5, or 7
are almost equal in free energy.
All of these findings could also be interpreted in terms
of a rapid equilibration of two type 3 enantiomers (Fig. 3).
This interpretation is supported by a comparison of the 11
B
Diethylamine as Neutral Base
NMR shifts of the type 3 adducts 1a·py and 1b·NEt3. Since
the proposed enantiomerization process creates a pseudo-
equivalency of the boron atoms 2/4, 5/9, 7/10, and 11/13
with the boron atoms 6,8, 12 on the pseudo-mirror plane
(Fig. 3), one would expect the signals of double intensity at
shift values, which are roughly the average values of the
The situation is different, when pyridine is replaced by the
secondary amine NHEt2. We followed the 1:1 reaction of
1a and NHEt2 in CD2Cl2 at Ϫ50 °C by NMR, having in
mind that excess amine would give [NH2Et2][MeNB11H11-
(NEt2)] [4], which is formally the product of a reaction with
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640
635

P. Paetzold, P. Lomme, U. Englert
(B2ϪB6ϪB4 65.2(1)°, B2ϪB7ϪB5 65.58(9)°). The three
skeletal NϪB bonds (150Ϫ154 pm) are considerably shorter
than the five NϪB bonds in a closo-BN11 skeleton (e.g.
170Ϫ173 pm in (PhCH2)NB11H11) and are even shorter
than in cubic boron nitride (157 pm). All of the other clu-
ster edges are found in the expected range, 182Ϫ186 pm,
and the angles are found close to 60°. The bond length bet-
ween the amine-N atom and B8, 156.2(2) pm, is close to the
BϪN length of cubic boron nitride. The crystal structure
supports the assumption of an equilibrium of type 3 enan-
tiomers in solution. When the crystals are dissolved again
in CD2Cl2 at Ϫ40 °C, the solution exhibits the same NMR
spectrum as before.
Fig. 3 MeNB11H11L: type 2 isomer and equilibrium of type 3 en-
antiomers (systematic numbering only for the left type 3 enanti-
omer; identical B atoms retain the same number in the three iso-
mers)
After standing at Ϫ40 °C for only 10 min, the NMR pe-
aks of a second isomer are growing, and this isomer repre-
sents the main component after 3.5 h. Ten 11B NMR peaks,
one of double intensity (accidental degeneracy) reveal an
unsymmetric structure (C1). The chemical shifts are found
at similar values as for the type 5 isomer 1a·py. The
2D-11B/11B NMR spectrum confirms the presence of a type
5 adduct; all of the expected cross-peaks could be observed,
except for the BϪB correlations 4Ϫ9, 5Ϫ8, 2Ϫ6, and 4Ϫ6
(and certainly 2Ϫ4 and 2Ϫ5 because of N bridging). The
signals of B8 and B9 are degenerate. The base is bound to
B4 (singlet).
corresponding B atoms of the non-enantiomerizing type 3
clusters, and just this expectation is nicely met (Table 1).
The same pseudo-equivalency as for the boron atoms holds
1
for the corresponding exo-H atoms, whose H NMR shift
values are also almost average values of the corresponding
peaks of type 3 adducts with tertiary bases (Table 2). We
assume, therefore, that an equilibrium of type 3 enantio-
mers is present after the addition of NHEt2 to 1a at
Ϫ50 °C.
The question arises, why such an equilibrium is observed
with NHEt2, but not with py and NMe3, particularly when
a type 2 isomer would represent the transition state. We
suggest that the amine-H atom is involved in the enantio-
merization process. During this process, the NϪB5 bond
is opened and the NϪB9 bond closed. Simultaneously, the
hydrogen atom Ha migrates from nitrogen to the B5ϪB8
bridging position, while Hb migrates from the B8ϪB9 brid-
ging position to the nitrogen. This process is rapid enough
to bring the corresponding 11B NMR signals to coales-
After 4 h standing at Ϫ40 °C, the 11B NMR peaks of a
third isomer become visible. This isomer is the only one
after 2 w at Ϫ40 °C. It is certainly the stable final type 7
product, proven by the typical intensity ratio (Cs symme-
try), by an ideal 2D-11B/11B NMR spectrum (missing only
the correlations 2Ϫ4/5) and by a singlet for B2 (see Table 1).
The 1:1 addition of NHEt2 to 1a represents the reaction
1
1
sequence of Fig. 2. H NMR spectra (Table 2) confirm the
cence, but too slow to let the H NMR signals of Ha and
structural conclusions by the same type of arguments given
earlier. Different from 1a·py, the type 7 isomer is more sta-
ble than the other isomers. Different from 1b·NMe3 and
1b·NEt3, the activation energies for the isomerizations are
small enough to allow a quantitative reaction at Ϫ40 °C.
Hb coalesce, which differ strongly in frequency (δ ϭ Ϫ1.01,
µ-H; δ ϭ 4.85, NH).
Crystals grow from the CD2Cl2 solution at Ϫ40 °C and
can be isolated. An X-ray structural analysis shows, that
both enantiomers of the type 3 isomer are present in the
crystal. One of the two enantiomers is depicted in Fig. 4.
The structure of the nido-NB11 skeleton is similar to that of
PhNB11H11(NEt3) (type 3) [4]. Long distances B4ϪB9 and
B5ϪB8 (208.4(2), 211.8(2) pm) make plausible that the cor-
responding 2D-11B/11B cross-peaks are not observed. The
bonds B2ϪB4 and B2ϪB5 are also rather long (196.4(2),
196.9(2) pm), allowing bond angles at nitrogen B2ϪN1ϪB4
and B2ϪN1ϪB5 (80.3(2), 81.1(2)°), which are closer to the
tetrahedral angle, preferred by an ammonium-type nitro-
gen, than the angles of a regular triangle. The bond angles
opposite to the edges B2ϪB4 and B2ϪB5 are less enlarged
tert-Butylamine as Neutral Base
The 1:1 reaction of 1a with NH2tBu in [D8]thf at Ϫ50 °C
yields the type 3 adduct. On standing at Ϫ40 °C, the NMR
peaks of the type 7 adduct become observable. This adduct
is the only product after 1 w. No indications of a type 5
isomer were recognized. The structural assignment is possi-
ble because of the similarity of the 11B NMR peaks of the
adducts of 1a with NHEt2 and NH2tBu (Table 1). The type
3 adduct 1a·NH2tBu also undergoes an enantiomerization,
636
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640

Aza-nido-dodecaboranes and -borates from Aza-closo-dodecaboranes
which establishes pseudo-Cs symmetry. We assume the same
enantiomerization mechanism as described for 1a·NHEt2.
Addition of Amines to Substituted Aza-closo-
dodecaboranes
The adducts of 1c؊e with NEt3 are obtained as oily pro-
ducts on a preparative scale at room temperature in
CH2Cl2. Though we have NMR indications from starting
the reaction at Ϫ78 °C, that other isomers might have been
formed as intermediate products, the only products at room
temperature were the type 7 clusters in all three cases. The
structure of the products can clearly be concluded from the
11B NMR spectra: chemical shifts are found in the expected
range and intensities (Table 1), supported by 2D-11B/11B
NMR spectra of 1c·NEt3 (cross-peaks 6/7Ϫ12 missing) and
1d·NEt3 (cross-peaks 4/5Ϫ8/9 and 4/5Ϫ6/7 missing). As
described for [1dMe]Ϫ, the 11B NMR peak of B12 appears
as a singlet (BHal), besides B2 certainly (BϪL), in the ad-
ducts of MeNB11H10Hal with NEt3, whereas we find five
singlets and only two doublets in the spectra of the adduct
1e·NEt3, comparable to [1f(OH)]Ϫ. All these findings show
again that only the ortho-belts of the closo-clusters 1 are
involved in the opening process, according to Fig. 2.
A striking result is, that halogen (Br, I) in the para-posi-
tion of a closo-cluster 1 makes type 7 nido-clusters the final
cluster opening products, whereas the clusters 1, which are
unsubstituted at the boron skeleton give type 3 products,
both with NEt3 as the base. A remarkable electronic effect
on a vertex in an icosahedral cluster is exerted by substitu-
ents in the para-position, well known as antipodal effect [8].
The polyhedral frameworks presented in the series of iso-
mers 2؊7 (Fig. 2) are not far away from an icosahedral
structure, from which they differ by the opening of one or
two BϪN bonds. The B12 vertex is situated in a kind of
para-position. It is plausible that the B12 substituents influ-
ence the BϪN openings and closures during the isomeriza-
tion process (Fig. 2).
[N(PPh3)2]Cl in CH2Cl2 up to 60 °C for 4 d), but can be
replaced by OHϪ (1:2 in thf, 3 h at room temperature):
1b·NMe3 ϩ KOH Ǟ K[1b(OH)] ϩ NMe3
The product is an isomer of type 7, identified by its NMR
spectra [4].
The base BrϪ cannot be replaced by MeϪ ([NEt4][1bBr]
and LiMe, 2:3 in thf, 48 h at room temperature), but can be
replaced by OHϪ (1:5 in thf, 3 d at room temperature) and
also by NEt3 (1:18 in CH2Cl2, 12 h at room temperature).
[NEt4][1bBr] ϩ KOH Ǟ K[1b(OH)] ϩ [NEt4]Br
[NEt4][1bBr] ϩ NEt3 Ǟ 1b·NEt3 ϩ [NEt4]Br
By these two reactions, the type 7 starting cluster either
remains a type 7 cluster (X ϭ OHϪ) or is transformed into
a type 3 cluster (L ϭ NEt3), both identified by the NMR
spectra ([4] or Table 1, respectively).
We assume that the corresponding arachno-clusters
[RNB11H11XL]Ϫ or [RNB11H11XXЈ]2Ϫ are formed as inter-
mediates. It is not yet possible to make conclusive assump-
The aza-closo-dodecaborane 1g offers an example of a
substitution in the ortho-belt. We added the primary amine
NH2tBu in a 1:1 ratio and isolated a colourless solid of the
nido-adduct 1g·NH2tBu. The 2D-11B/11B NMR spectrum
can only be interpreted by assuming a type 5 structure with
an accidental degeneracy of B4/B5 and B6/B11; no cross-
peaks are found for the connections 4/5Ϫ8/9 (as in most),
2Ϫ4/5 (as in all of our nido-NB11 clusters), and 6Ϫ10. The
peak of B4/B5 is a singlet (Table 1).
Obviously, the base NH2tBu attacks 1g in the ortho-belt
in a position non-adjacent to the vertex with the alkyl li-
gand. The type 5 structure of the isolated isomer should be
governed by the optimal steric arrangement of the three
non-hydrogen ligands at the non-planar pentagonal aper-
ture (bold lines in the above scheme).
Fig. 4 Structure of 1a·(NHEt2), type 3 (PLATON [7], 30 % dis-
placement ellipsoids); selected distances:
N1ϪC1 147.0(2), N1ϪB2 150.1(2) N1ϪB4 154.4(2), N1ϪB5 152.8(2),
N2ϪB8 156.2(2), B2ϪB4 196.4(2), B2ϪB5 197.1(2), B2ϪB6 178.3(2),
B2ϪB7 178.4(2), B4ϪB6 186.2(2), B4ϪB9 208.3(2), B4ϪB10 178.5(2),
B5ϪB7 185.1(2), B5ϪB8 211.8(2), B5ϪB11 179.4(2), B6ϪB7 178.5(2),
B6ϪB10 177.4(2), B6ϪB12 174.5(2), B7ϪB11 177.1(2), B7ϪB12 175.6(2),
B8ϪB9 179.0(2), B8ϪB11 171.7(2), B8ϪB13 173.9(2), B9ϪB10 173.2(2),
B9ϪB13 173.9(2), B10ϪB12 179.3(2), B10ϪB13 179.1(2), B11ϪB12
180.1(2), B11ϪB13 181.0(2), B12ϪB13 174.4(2), B8ϪH89 126(1), B9ϪH89
135(1).
Substitution of Bases
The base NMe3 in 1b·NMe3, type 3 [4], cannot be replaced
by ClϪ (heating
a 1:1 solution of 1b·NMe3 and
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640
637

P. Paetzold, P. Lomme, U. Englert
tions about the structure of these intermediates and to make
plausible suggestions for the rearrangements, that are steps
during those substitutions.
Discussion of Results and Conclusions
The three isomers 3, 5, and 7 with a BHB bridge on the
one hand and the three isomers 2, 4, and 6 with an endo-H
atom on the other hand become indistinguishable, when the
nido-anion [NB11H13]Ϫ is formed by the hypothetical addi-
tion of HϪ to closo-NB11H12. In order to describe the bon-
ding situation in this parent nido-anion, we have to take
into account 52 valence electrons and 61 valence orbitals
of the 25 atoms. If we restrict ourselves to localized (3c2e)
(number: t) and (2c2e) bonds (number: y), the octet rule for
boron and nitrogen and the doublet rule for hydrogen will
be obeyed by the balances 3t ϩ 2y ϭ 61 and 2t ϩ 2y ϭ 52,
giving t ϭ 9 and y ϭ 17. Subtracting twelve (2c2e) bonds
for exo-BH and -NH bonds, the skeleton will be built up
by nine (3c2e) and five (2c2e) bonds. We will have three
BN- and two BB-(2c2e) skeletal bonds in the type 7 isomer,
having in mind that BNB-(3c2e) bonds are unfavourable; in
the type 6 isomer, we have to expect two BN- two BB-,
and one BHendo-(2c2e) bonds. The situation is depicted in
Fig. 5.
Actually, the transformation of 6 into 7 and vice versa
includes only two bonds: a BNB-(3c2e) becomes a BN-
(2c2e) bond and a BH-(2c2e) becomes a BHB-(3c2e) bond.
We had called such a type of reaction a [3c,2c]-transloca-
tion [9], which is a common reaction step in borane-cluster
chemistry. The isomer 7 is certainly more stable than 6,
since a BNB-(3c2c) and a BH-(2c2e) bond are worse in
energy than a BN-(2c2e) and BHB-(3c2e) bond.
Fig. 5 (a) One of the possible distributions of nine (3c2e) bonds
(grey triangles) and five (2c2e) bonds (bold lines) in the NB11 skel-
eton of nido-[NB11H13]Ϫ, planarized and divided into a type 6 or
type 7 open NB5 pyramid (upper side of the cluster) and a capped
pentagonal antiprism (with pentagons 2Ϫ5Ϫ8Ϫ9Ϫ4 and
6Ϫ7Ϫ11Ϫ13Ϫ10 and capping apex 12; lower side of the cluster,
which fits with 6 as well as with 7); (b) three canonical structures
of the type 6 open NB5 pyramid, the first one taken from (a); (c)
five canonical structures of the B6 pentagonal pyramid with basis
6Ϫ7Ϫ11Ϫ13Ϫ10 and apex 12, the first one taken from (a)
Going from the parent anion [NB12H13]Ϫ to neutral or
anionic base adducts, RNB11H11L or [RNB11H11X]Ϫ, the
differences in stability between the type 3, 5, and 7 isomers
depend on steric and electronic influences of the base, that
we do not understand in detail. In the case of pyridine as
the base, all three of the isomers seem to be comparable in
stability. In the case of the amines NHEt2 and NH2tBu, on
the other hand, the type 7 isomer is the most stable, which
is formed via the type 3 and the type 5 isomer. In the case
of the tertiary amines NMe3 and NEt3, however, the type
3 isomer is the final product, thermally so stable that its
predominance (e.g. in boiling xylene) seems to be caused
thermodynamically. In the case of kinetic control, the terti-
ary amines should not hinder a simple [3c,2c]-translocation
so effectively, assuming that the formation of 5 from 3 goes
via a transition state or intermediate of type 4 by the same
step as just pointed out for the parent anion. Anionic bases
lead to type 7 final products. A type 3 intermediate was
observable in the case of [FeCp(CO)2]Ϫ as the base.
ortho-belt, thus supporting the proposed mechanism (Fig.
2). In spite of the striking stability of the type 3 product
PhNB11H11(NEt3), the substituted aza-closo-boranes gave
type 7 products with NEt3.
We have no experimental indication, whether the isomers
of type 2, 4, or 6 are intermediates or transition states du-
ring the reaction sequence presented in Fig. 2. Concerning
the formation of the type 2 isomer, we cannot imagine that
the lone pair of a Lewis base would attack anything else
but a vertex of 1. Using again the localized orbital picture,
we find that closo-NB11H12 is built up by 10 (3c2e) and 15
(2c2e) bonds, three of which in the skeleton. One of these
three bonds is certainly a BN bond; more than one (2c2e)
bond is not possible for a cluster atom with the connectivity
five. The attack of a base in the ortho-belt makes a second
BN-(2c2e) bond possible, which is illustrated by the follo-
wing structures, showing only the upper half of the cluster.
Thereby, a BNB-(3c2e) bond and a (1c2e) lone pair are
transformed into two (2c2e) bonds, one of which is a BN
bond, the other one is the bond between the base and bo-
ron. We had called such an elementary reaction a [3c,1c]-
When we started from aza-closo-boranes with substitu-
ents in para- or in para- and meta-position, the substitution
pattern of the products, all of type 7 by the way, showed,
that the ortho-belt is attacked by the base and that isomeri-
zations during the formation of the products occur in the
638
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640

Aza-nido-dodecaboranes and -borates from Aza-closo-dodecaboranes
Table 3 Amounts of starting materials 1 and base and yields of
adducts 1·base of the neutral type 1·L or the anionic type [cat-
ion][1X]
1
mmol base
mmola) 1·base
yield
b)
1a
"
"
"
"
1b
"
"
1c
1d
"
1e
1f
1g
0.69
1.00
1.00
1.38
1.70
0.68
0.32
0.59
0.50
0.56
0.70
0.27
0.23
0.20
NHEt2
0.68
10.1
1.05
1.37
1a·NHEt2
[NH3Bu][1a(NHBu)]
1a·NH2tBu
NH2Bu
NH2tBu
py
b)
b)
collocation [9]. We assume that the type 2 isomers are gene-
rally formed from the aza-closo-dodecaboranes by a
[3c,1c]-collocation.
Summarizing, the results show that Fig. 2 represents the
most plausible mechanism for the opening of aza-closo-do-
decaboranes by Lewis bases.
1a·py
Na[FeCp(CO)2] 1.70
Na{[1a[FeCp(CO)2]}
[NEt4][1bBr]
[N(PPh3)2][1b(N3)]
[NH3tBu][1b(NHtBu)]
1c·NEt3
1d·NEt3
Li[1dMe]
1e·NEt3
H[1f(OH)]
1g·NH2tBu
74
57
78
83
46
67
88
[NEt4]Br
[N(PPh3)2]N3
NH2tBu
NEt3
0.70
0.32
9.5
0.72
0.72
0.64c)
0.72
ca.1d)
0.20
"
LiMe
NEt3
H2O
NH2tBu
65
b)
Experimental
60
a) Amount of amines measured by a calibrated micro-syringe. b) Yield quanti-
tative, according to the NMR spectrum of the reaction mixture. c) 0.40 ml
1.6 M solution in Et2O. d) 1 ml aqueous MeOD.
NMR spectra: Varian UNITY 500; 499.83, 160.36, 125.70,
470.27 MHz for 1H, 11B, 13C, 19F, respectively. We used external
standards for 11B (BF3(OEt2)) and 19F (CFCl3); residual protons
and 13C atoms of the deuterated solvents acted as standards for
1H and 13C, respectively, converted finally into the common tms
standard. CD2Cl2 was the solvent, but [D8]thf was used for 1a,
1a·py, 1a·NH2tBu, Li[1dMe], and Na{1a [FeCp(CO)2]} and a di-
luted solution of H2O in MeOD for H[1f(OH)]. COSY 45 was ap-
plied for pulse recurrence in 2D-11B/11B NMR spectra; cluster-exo-
H atoms were assigned with the aid of 2D-1H/11B HMQC. 11B
and then stirred for 12 h. The adducts were isolated by removing
Ϫ
all volatile materials in vacuo; in the case of N3 as the base, the
product precipitated by adding hexane to the reaction mixture. No
purification took place with NHBuϪ and the ferrate as the bases;
the solid red ferrate product was pure with respect to the NMR
spectra and was present in a 2:3 ratio of the type 3 and 7 isomers;
this ratio changed in favour of the type 7 isomer on standing in thf
solution. The adducts of 1c؊e with NEt3 and Li[1dMe] were iso-
lated as viscous colourless oils; impurities could be extracted by
digesting with hexane or (in the case of Li[1dMe]) Et2O. Extraction
with hexane was also applied to the colourless solids with BrϪ and
1
NMR peaks are listed in Table 1, H NMR peaks in Table 2.
7,8,9,10,11-Pentamethyl-1-phenyl-12-(trifluoromethylsulfonyl)aza-
closo-dodecaborane: A mixture of 1-phenylaza-closo-dodecaborane
(1b) [4] (380 mg, 1.72 mol), methyl trifluoromethanesulfonate (me-
thyl triflate) (2.1 ml), and trifluoromethanesulfonic acid (0.6 ml)
was heated to 140 °C for 20 h. Excess triflate and triflic acid were
removed in vacuo. The black residue was extracted with three por-
tions of hexane (5 ml). Colourless product (420 mg, 56 %) was ob-
tained from the hexane solution, after removal of hexane, by subli-
mation (80 °C/0.001 mbar). Recrystallization from hexane was pos-
sible at Ϫ40 °C. C12H25B11F3NO3S (439.3), found (calcd.): C 32.58
(32.81), H 6.45 (5.74), N 3.03 (3.19).
Ϫ
N3 as the bases. A recrystallisation from CH2Cl2 could be achie-
ved with the salts containing the anions [1bBr]Ϫ and [1b(NHtBu)]Ϫ
and with 1g·NH2tBu. Amounts of the starting materials and yields
of the purified products are listed in Table 3.
1H NMR: δ ϭ 0.55 (s, 15 H, Me), 2.37 (H2ϪH6), 6.55Ϫ6.84 (5H, Ph). Ϫ
11B NMR: δ ϭ Ϫ8.8 (d, J ϭ 165 Hz, B2ϪB6), Ϫ4.2 (s, B7ϪB11), 16.8 (s,
B12). Ϫ 13C NMR: δ ϭ Ϫ4.1 (Me), 119.2 (q, 1J ϭ 314 Hz, CF3) 123.1,
127.9, 129.1, 145.1 (Ph). Ϫ 19F NMR: δ ϭ Ϫ77.4.
Crystal structure of 8-(diethylammonio)-1-methyl-1-aza-nido-dode-
caborane: Data were collected on an ENRAF-Nonius CAD4 dif-
fractometer, using MoKα radiation (λ ϭ 154.184 pm, graphite mo-
nochromator) at 228 K. Space group P21/c (Nr. 14); a ϭ 755.9(2),
b ϭ 1371.6(4), c ϭ 1449.9(3) pm; β ϭ 94.31(2)°; V ϭ 1.4991(6) nm3;
General procedure of the adduct formation: The closo-clusters of
type 1 were dissolved in CH2Cl2 (5 ml); other solvents were used,
however, before the salt Na[FeCp(CO)2] (thf, 2 ml) and the bases
NH2tBu and py (both 1:1; [D8]thf, 3 ml), LiMe (Et2O, 5 ml) were
added, and aqueous MeOD was added to undissolved 1f. The solu-
tions were cooled down to Ϫ50 °C, but were cooled down to
Ϫ78 °C, when LiMe was added to 1d and NEt3 to 1c؊e; MeOD
was added to 1f at room temperature. The amines or salts (in the
case of anionic bases) were added undissolved, except the sodium
ferrate (thf, 5 ml) and LiMe (1.6 molar solution in Et2O). In the
case of the 1:1 reaction of 1a with amines, the adducts were not
isolated from the reaction mixture; NMR samples were taken from
the reaction flask at Ϫ50 °C and measured at different temperatu-
res, as indicated above; crystals of 1a·NHEt2 grew in the flask at
Ϫ40 °C and were isolated for the X-ray structural investigation as
usual. The reaction mixture from 1f and aqueous MeOD was im-
mediately treated by NMR methods. In all of the other cases, the
reaction mixture was brought to room temperature within 30 min
Z
ϭ
4; calculated density 1.029 g cmϪ3
0.60·0.60·0.60 mm3; 3.18 cmϪ1
sec. extinction coeff.
;
crystal size
µ
ϭ
;
0.696·10Ϫ6; range of Θ: 5Ϫ70°. The structure was solved by direct
methods and refined on reflection intensities [10]. Measured reflec-
tions: 4337; observed independent reflns. (I Ͼ σ(I)): 2417; refined
parameters: 264; R ϭ 0.048; Rw ϭ 0.057 (wϪ1 ϭ σ2(Fo)); GOF ϭ
1.657; max. residual electron density: 0.2 e AϪ3. The crystallo-
˚
graphic data have been deposited at the Cambridge Crystallo-
graphic Data Centre as supplementary publication No. CCDC 169
169. Copies of the data can be obtained free of charge on applica-
tion to CCDC, 12 Union Road, Cambridge CB2 1 EZ UK (e-mail:
deposit@ccdc.cam.ac.uk).
Acknowledgement: We are grateful to Deutsche Forschungsgemein-
schaft and to Fonds der Chemischen Industrie for the support of
this work.
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640
639

P. Paetzold, P. Lomme, U. Englert
[5] F. Meyer, J. Müller, P. Paetzold, R. Boese, Angew. Chem. 1992,
104, 1221; Angew. Chem. Int. Ed. Engl. 1992, 31, 1227.
[6] L. Schneider, U. Englert, P. Paetzold, Chem. Ber. 1994, 127, 87.
[7] A. L. Spek, PLATON 94, University of Utrecht, The Nether-
lands, 1994.
References
[1] P. Lomme, M. Roth, U. Englert, P. Paetzold, R. Boese, Chem.
Ber. 1996, 129, 1227.
[2] P. Paetzold, Eur. J. Inorg. Chem. 1998, 143.
˘
´
[8] S. Hermanek, Chem. Rev. 1992, 92, 325.
[9] P. Paetzold in The Borane, Carborane, Carbocation Continuum,
J. Casanova (ed.), Wiley, New York 1998, p. 381.
[10] G. M. Sheldrick, SHELXS 97, Program for Structure Solution,
and SHELXL 97, Program for Structure Refinement, Univer-
sity of Göttingen, Germany, 1997.
[3] H. Yao, P. Lomme, C. Hu, P. Paetzold, Z. Anorg. Allg. Chem.,
2001, 627, 2477.
[4] P. Lomme, F. Meyer, U. Englert, P. Paetzold, Chem. Ber. 1995,
128, 1225.
640
Z. Anorg. Allg. Chem. 2002, 628, 632Ϫ640