Dianionic amidinates at silicon and germanium centers: Four-, six- and eight-membered rings

Article abstract of DOI:10.1016/j.jorganchem.2007.02.032

At variance to an earlier finding, the reaction of Me₂SiCl₂ with Li[(Me)N-C(Ph)-NH] (1a), in the presence of a base, gives a six-membered ring molecule μ-[(Ph)(MeN)C-N][-SiMe₂-N-C(Ph)-N(Ph)-SiMe₂-] (s3a), wherea

Full text of DOI:10.1016/j.jorganchem.2007.02.032

Journal of Organometallic Chemistry 692 (2007) 2789–2799
www.elsevier.com/locate/jorganchem
Dianionic amidinates at silicon and germanium centers: Four-,
six- and eight-membered rings
a
a
b
a
a
Thomas Segmuller , Peter A. Schluter , Markus Drees , Annette Schier , Stefan Nogai ,
¨
¨
c
b,
a,
*
Norbert W. Mitzel , Thomas Straßner *, Hans H. Karsch
a
Department Chemie, TU Mu¨nchen, Lichtenbergstr. 4, 85747 Garching, Germany
Professur fu¨r Physikalische Organische Chemie, TU Dresden, Mommsenstr. 13, 01062 Dresden, Germany
Institut fu¨r Anorganische und Analytische Chemie, Universita¨t Mu¨nster, Corrensstr. 30, 48149 Mu¨nster, Germany
b
c
Received 2 November 2006; received in revised form 9 February 2007; accepted 9 February 2007
Available online 25 February 2007
Abstract
At variance to an earlier finding, the reaction of Me2SiCl2 with Li[(Me)N–C(Ph)–NH] (1a), in the presence of a base, gives a six-mem-
bered ring molecule l-[(Ph)(MeN)C–N][–SiMe2–N–C(Ph)–N(Ph)–SiMe2–] (s3a), whereas with Li[iPrN–C(Ph)–NH] (1b), a four-mem-
bered ring molecule l-[iPrN(Ph)C–N]2(SiMe2)2 (s4b) was formed. In contrast, with Li[tBuN–C(Ph)–NH] (1c), no such reaction
occurred. Obviously, a delicate influence of steric effects has to be taken into account. In fact, the latter amidinate reacts with GeCl2
to form an eight-membered ring molecule [tBuN–C(Ph)–N–Ge]4 (5c) without adding an additional base.
The compounds are fully characterized and their structures determined by X-ray diffraction. DFT calculations confirm the depen-
dence on steric influences. The relative energies of ground and transition states give a rationalization the ease of transformations of
the various rings via pathways with penta- and hexacoordinate silicon centers, which in turn relates to the experimental results on penta-
and hexacoordinate silicon amidinates and their fluctional behavior in solution.
Ó 2007 Elsevier B.V. All rights reserved.
Keywords: Amidinates; Silicon(IV); Germanium(II); DFT calculations
1. Introduction
mers and, again in principle, give rise to ceramic
materials with designed properties by thermolysis. In prin-
ciple, this concept, though then not intended for the indi-
cated purpose, was realized in an early study of the
reaction of Me2SiCl2 with Li[MeN–C(Ph)–NH], obtained
in situ from LiNHMe and PhCN, in the presence of a base
[1]. This reaction, performed at room temperature, was
accompanied by ‘‘fume evolution’’ and was believed to
yield an eight-membered ring molecule, 2a, in 43% (Eq.
Dianionic amidinates at (organo-)silicon centers are
potentially useful to form polymers with [(R)N–C(Ph)–
N]2ꢀ linkages, mimicking O2ꢀ bridges in silicones or sili-
cates in one-, two- or three-dimensional frameworks. This
dianionic amidinate should be obtained by the reaction
of the monoanionic, unsymmetrical amidinates [(R)N–
C(R0)–NH]ꢀ, in turn easily prepared (R0 = Ph) by reacting
LiNHR with PhCN (Eq. (1)), followed by a reaction with
(organo)silicon chlorides and subsequent deprotonation
by means of a suitable base. Other (organo)element chlo-
rides (e.g. BCl3, AlCl3, TiCl4, etc.) could modify the poly-
1
(2)), by a questionable analysis of the H NMR spectrum
(60 MHz, CW, r.t.). Nevertheless, besides of the exact con-
stitution of this compound, the report seemed reliable
enough to assume the presence of a dianionic amidinate
as bridging ligand between two silicon centers and thus
gave a hint that the above-mentioned concept seemed
feasible.
*
Corresponding authors.
E-mail address: Hans.H.Karsch@lrz.tum.de (H.H. Karsch).
0022-328X/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2007.02.032

2790
T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
N
N
H
R
N
N
Li
an experiment at ꢀ78 °C led to the formation of colorless
crystals of compound s3a isolated in 89.6% yield (Eq. (5)).
RNHLi
+
PhCN
Ph
C
Ph
C
H
R
Me
Me
Me
Li
Si
N
Me
N
R = Me
R = 1a: Me; 1b: iPr; 1c: tBu; 1d: Cy
C
N
C
Ph
ð1Þ
Ph
Si
N
Me
Me
Me
Si
Me
Me
+ 2 Me2SiCl2
+ 2 NEt3
N
s3a
2 1a, 1b
Ph
2 NEt3
N
C
C
N
2 1a
+
2 Me2SiCl2
Me Me
Si
- 2 [HNEt3]Cl
- 2 LiCl
Pri
N
Ph
N
Ph
N
Si
Me
C
C
N
N
R = iPr
iPr
Me
Ph
Si
Me
Me Me
2a
a4b
ð2Þ
ð5Þ
Ph
C
H
The compound was found to have a dimeric structure
[Me2Si–N–C(Ph)–NMe]2 (MS), in line with Scherer’s for-
mula 2a, and appears to be stable in solution at room tem-
perature according to the NMR spectra, but with a
somewhat different constitution. Unfortunately, we could
not draw relevant information from the 13C and 29Si spec-
tra, because room temperature spectra showed only broad
signals, presumably due to their fluctional behavior. How-
ever, in the VT-1H NMR spectra (200 MHz, FT), fluctional
and constitutional effects could be separated. In the tem-
perature range from ꢀ80 to 85 °C two SiCH3 signals with
decreasing shift difference were observed, but with equal
intensity throughout, which is in line with Scherer’s report.
Also observed were two independent NCH3 signals at
ꢀ60 °C. This hardly is in line with formula 2a, but rather
is consistent with formula s3a. Furthermore, the postula-
tion of the presence of two different conformers seems ob-
solete (a constitutional eight-membered ring isomer could
not be ruled out rigorously at this stage, however). More-
over, at ꢀ80 °C, one of the two NCH3 signals is split in
turn into two signals (ratio 2:1), which can be interpreted
as hint for the steric congestion within the six-membered
ring of s3a and hence hindered rotation of the N–CH3
group. From the spectra no information was gained about
the rotational behavior (or position) of the external
MeNC(Ph)-group, i.e. about the presence of s3a or a3a
(syn or anti rotamers; Eq. (6)).
base
no reaction or
unidentified
products resp.
1c
+
SiCl4
N
SiCl3
N
But
tBu
N
Ph
C
SiCl3
N
H
ð3Þ
ð4Þ
NEt3
unidentified mixture of products
1c, d
+
Me2SiCl2
Earlier work by us [2,3] has shown, however, that there is
obviously a severe steric problem: the reactions of amidi-
nates 1c and 1d with silicon centers under deprotonation
conditions could not be achieved with various bases and/
or clear results were not obtained (Eq. (4)). Therefore we
decided to reinvestigate reaction (2) and include also the
reaction with 1b. Preliminary results have been reported
[4,5].
To get insight into the steric influence of the substituents
on the preferred ring size we calculated the energies of iso-
mers with different ring size and the transition states for the
ring size changes. We intent to get an explanation why the
change of the substituent (alkyl or aryl) leads to different
crystal structures.
Me
Me
N
Me
Si
Me
Si
Me
N
Ph
Me
Ph
C
N
2. Results and discussion
C
N
Rotation
Ph
N
Me
N
N
C
N
Si
The lithium amidinate 1a [1,3,6], prepared from Me2NLi
and PhCN was allowed to react in situ with Me2SiCl2 in the
presence of NEt3 in Et2O solution according to Scherer’s
procedure (Eq. (2)). Whereas repeated reactions conducted
at room temperature did not lead to reproducible reactions,
C
Si
Me
Me
Me
Me
Ph
Me
s3a
a3a
ð6Þ

T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
2791
To get more insight into the nature of the compound, or of
its possible isomer, a X-ray diffraction study of s3a was
undertaken.
not in plane with the C1 and C2 centered triangles
(interplane angles N1–C1–N2/phenyl ring 41.1° and
N3–C2–N4/phenyl ring 58.2°), respectively. One explana-
tion would be the repulsive interaction of the rings with
the N–CH3 groups (which is less, however with the
N–iPr groups in a4b – see below), thus indicating some
steric constraints. Secondly, the rotational position of
the ‘‘free’’ amidinate nitrogen atom N4 enables this
nitrogen atom to come into contact with the electroni-
Colorless crystals of s3a (orthorhombic, space group
Pna21, with four molecules/cell) were grown from pen-
tane. The molecular structure, together with selected
bond distances and angles, is shown in Fig. 1. The com-
pound consists of two dianionic amidinate ligands and
two SiCl2 fragments, one ligand bridges the two silicon
atoms by its two nitrogen (N1, N2), the other one by
only one nitrogen (N3) atom, whereas the fourth, dicoor-
dinated nitrogen (N4) atom remains uncoordinated.
Thus, the tetracoordinated silicon atoms, the almost pla-
nar, tricoordinated atoms N1, C1, N3 and the dicoordi-
nated N2 atom form a slightly twisted six-membered
ring. Remarkably, the C1–N2 and C2–N4 distances are
almost equal and shorter (double bonds) than the other
C–N distances (single bonds). The Si–N distances
˚
cally least saturated Si1 atom: Si–N4: 2.713 A; cf.:
semi-chelating
(unisobidentate):
(CF3)2CHAN@C
˚
(ClAC6H4)N(2,6-Me2C6H3) ASiCl3: 2.684(4) A [7]; non-
chelating (monodentate): NH(tBu)C(Ph)N–SiCl3 and
[NH(tBu)C(Ph)N]2SiCl2: Si–N > 3.9 A [2,8]. This is
˚
clearly less than the sum of Si–N van-der-Waals radii
˚
(3.7 A) and characterizes the semi-chelating nature of
the respective amidinate ligand and suggests the presence
of a simple, low lying pathway to pentacoordination at
Si(1), a prerequisite for ring expansion (see below). Also,
the angles N4–C2–N3 (116.6(2)°) and C2–N3–Si1
(112.9(1)°) are less than 120°, the four atoms being
approximately coplanar. Though not too strongly con-
vincing, all other data describing the coordination sphere
around Si1 compared to those of Si2 augment this inter-
˚
(1.732–1.778 A, with Si–N2 being the shortest distance)
are longer than the short Si–N bonds in other tetracoor-
dinated silicon compounds with non-chelating amidinate
˚
ligands (1.625–1.651 A; see below), but compare well
with that in an example with a semi-chelating amidinate
˚
ligand (1.735 A; see below). Two additional features
deserve further comment: Firstly, both phenyl rings are
pretation (‘‘real’’ pentacoordination would afford Si–
˚
N
ax = 1.92–1.96 A [2–5]). In fact, most amidinates tend
to form pentacoordinated and even hexacoordinated sili-
con species and fluctional behavior of pentacoordinate
silicon amidinates is a common feature [2–5].
Obviously, since there are steric constraints to be taken
into account as earlier mentioned, amidinates 1c and 1d are
not readily deprotonated at silicon centers with a base. To
further clarify the scope of the reaction, 1b was used (Eq.
(5)). The reaction was performed in ether at r.t. Colorless
crystals of an again dimeric compound were obtained from
toluene solutions of a4b (36.5%).
In contrast to s3a, the NMR spectra of a4b show –
independently from temperature – only one (set of) sig-
nal(s) for the Si–CH3 and N[CH(CH3)2] groups (1H,
{1H}13C), the 29Si resonance lies at 3.88 ppm. Together
with the MS spectrum, the presence of a four-membered
ring cyclo-disilazane (diaza-disila-cyclobutane) structure
was deduced and finally confirmed by the result of a
X-ray diffraction study.
In fact, a4b crystallizes in a monoclinic cell (P21/c) as a
centrosymmetric, essentially planar cyclo-disilazane mole-
cule, i.e. the amidinate ligands are bound only by one nitro-
gen atom each to silicon. The molecular structure is shown
in Fig. 2, together with selected bond distances. From the
structure it is obvious that the two imide units are in anti
conformation, indicating that not the syn compound s4b,
but the anti a4b has been formed. There are no remarkable
deviations from the respective data of known compounds
of this type [9–18], whereby a common feature are the small
N–Si–N and the larger Si–N–Si angles, although in a4b,
these values lie at the low (83.9(1)°) and the high end
Fig. 1. Molecular structure of compound s3a (a: view from above, b: side
view; ORTEP drawing with 50% probability ellipsoids). Selected bond
˚
distances (A) and angles (°): Si1–N1 1.780(2), Si1–N3 1.762(2), Si2–N2
1.732(2), Si2–N3 1.755(2), C1–N1 1.388(3), C2–N3 1.403(3), C2–N4
1.287(3); N3–Si1–N1 102.76(9), Si1–N1–C1 118.9(1), N1–C1–N2 123.1(2),
C1–N2–Si2 123.4(2), N2–Si2–N3 106.5(1), Si1–N3–C2 112.8(2), Si2–N3–
C2 127.6(2), C2–N4–C3 120.4(2).
(96.2(1)°) with
a
non-bonding Si. . .Si distance of

2792
T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
The NMR spectra (1H, {1H}–13C) gave no clear result,
because only one (set of) signal(s) was obtained for the
respective functional groups of the ligand. Nevertheless,
the self-deprotonation was evident from the absence of
an N–H signal and the isolation and NMR spectroscopic
identification of the amidine NH(tBu)AC(Ph)@NH. The
mass spectrum of 5c revealed the presence of a tetra-
meric compound {Ge[(NH)2(tBu)C(Ph)]}4, but no deci-
sion could be made about the true nature: either a
cubane like structure [22] or a twisted eight-membered
ring within a 16-membered macrocycle (there is a unique
structural analogue in tin chemistry [23]) seemed most
likely.
Fig. 2. Molecular structure of compound a4b (ORTEP drawing with 50%
˚
probability ellipsoids). Selected bond distances (A) and angles (°): Si1–N1
1.760(1), Si1–N1A 1.759(1), N1–C3 1.384(2), N2–C3 1.283(2), Si1. . .Si1A
2.6180(7); Si1–N1–Si1A 127.50(9), N1–Si1–N1A 83.85(5), Si1–N1–C3
135.82(10), Si1A–N1–C3 127.50(9), C3–N2–C4 119.82(12).
N22
C20
N21
Ge1A
Ge2
˚
2.618(1) A. As before in s3a, the phenyl rings are not in
plane with the carbon substituents at C3/C3A. In this case,
they are almost perpendicular (interplane angle N1–C3–
N2/phenyl ring 84.4°), obviously for steric reasons, exclud-
ing electronic interaction. And as well, the exocyclic
amidinate nitrogen atoms N2/N2A, forming a double bond
N2
C1
Ge2A
Ge1
N1
˚
to C3/C3A (1.283(2) A), are not completely ‘‘innocent’’:
the four atoms Si1, N1A, C3A, N2A and Si1A, N1, C3,
N2 almost form a plane, thus enabling the nitrogen to
˚
approach the silicon atom (3.095 A). This is more than
the respective value in s3a, but again clearly much less than
the covalent radii would suggest. In variance to s3a, the
four-membered ring of a4b is much less flexible and thus
cannot adapt to the needs for a pre-pentacoordinated con-
formation at silicon. This in turn explains the increase in
ring size for a4b to be a somewhat more energetic pathway
than in s3a (see DFT calculations).
C20
N22
Ge1
N21
N1
N2
It emerges from these findings that (small) steric changes
exert a dominant influence on the nature of products; big-
ger groups prevent double deprotonation of amidines at
all. To reduce this steric hindrance, we tested the same type
of reaction at a somewhat bigger, but comparable coordi-
nation center: we replaced Si(IV) by Ge(II). Germanium
(II) has been demonstrated to be a suitable coordination
center for amidinates [2,4,19–21]. And indeed, by reacting
GeCl2 Æ dioxane with the then sterically unfavorable amidi-
nate 1c, no addition of another base was required for
deprotonation, due to spontaneous deprotonation by a sec-
ond equivalent of 1c and 5c was obtained as colorless crys-
tals from toluene (Eq. (7)) [3,5]
Ge2
C1
Fig. 3. Molecular structure of compound 5c (a: view from above, b: side
view; ORTEP drawing with 50% probability ellipsoids; substituent carbon
atoms are depicted as wire-frame model for clarity). Selected bond
˚
distances (A) and angles (°): Ge1–N1 2.043(1), Ge1–N2 1.990(1), Ge2–
N21 1.994(1), Ge2–N22 2.040(1), Ge2–N2 1.901(1), Ge1A–N21 1.902(1);
N1–C1–N2 110.4(1), N2–Ge1–N1 65.42(5), C1–N1–C2 131.6(1), C1–N1–
Ge1 90.92(9), C2–N1–Ge1 136.61(10), N21A–Ge1–N1 92.87(5), C1–N2–
Ge1 93.03(9), C1–N2–Ge1 92.74(9), Ge1–N2–Ge2 134.16(7), C20–N21–
Ge1A 132.48(19), C20–N21–Ge2 92.88(9), Ge1A–N21–Ge2 134.67(7),
C20–N22–C21 131.9(1), C20–N22–Ge2 90.97(9), C21–N22–Ge 136.0(1),
N2–Ge2–N21 91.02(5), N21–Ge2–N22 64.41(5), N21A–Ge1–N2 91.03(5).
Et2O
.
Li[NH(tBu)-C(Ph)=N] + GeCl2 C4H8O2
1/4 {Ge[N(tBu)C(Ph)N]}4
ð7Þ
- 78˚C
5c
1c

T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
2793
˚
Therefore, a X-ray diffraction study of 5c was under-
taken, which revealed indeed the presence of the latter type
of arrangement in the molecule, leading to four germanium
spirocycles at four edges of an eight-membered ring, con-
sisting of four-membered, almost planar, chelating amidi-
nate rings at Ge, whose ‘‘free’’ (non-tBu substituted) N
atoms bridge the Ge atoms and thus generating the eight-
membered, centrosymmetric ring, in common for the four
Ge-spirocenters. The molecular structure is shown in
Fig. 3, together with selected bond distances and angles.
The quite unusual conformation of the puckered eight-
membered ring at first glance seemed to be due to steric
N21# 1.902(1) A] Ge–N distances, the latter being the
shortest due to the formal negative charge on nitrogen
and some ring strain in the four-membered chelate rings
incorporating the essentially planar N1, C1, N2/N21,
C20, N22 and the pyramidal Ge1/Ge2 atoms. Within
these rings, the C–N distances represent delocalized,
somewhat elongated double bonds (C1–N1 1.323(2)/
C20–N21 1.329(2) and C1–N2 1.331(2)/C20–N22
˚
1.324(2) A), in remarkable contrast to s3a and a4b, and
the N–C–N angles (N1–C1–N2 110.4(1)/N21–C20–N22
110.6(1)°) correspond to the more acute N–Ge–N angles
(N1–Ge1–N2 65.4(1)/N21–Ge2–N22 65.4(1)°), whereas
the endo-chelate angles at N(C1–N1–Ge1 90.9(1)/C1–
N2–Ge1 93.0(1) and C20–N22–Ge2 91.0(1)/C20–N21–
Ge2 92.9(1)°) deviate only slightly from 90°, in line with
the other angles around Ge. As the other angles of the
chelating C1/C20 and N1, N2/N21, N22 ring atoms
add to planarity, the sp2 carbon atoms seem more ‘‘resis-
tant’’ (harder) to the needs of chelate ring formation
than the sp2 nitrogen atoms. And in line with expecta-
tion from s3a and a4b, the phenyl rings adopt an almost
perpendicular position to the planes at C1/C20 (inter-
plane angles N1–C1–N2/phenyl ring 63.3° and N21–
C20–N22/phenyl ring 34.7°).
t
needs: all phenyl groups are in an axial, all butyl groups
in an equatorial position. At closer sight, however, compar-
ison can be made with the related structure of (NSF)4
[24,25]: [NX3(sp2)]2Ge(IV) vs. ½F;NXꢀ2 ðsp2ÞꢁSðIVÞ and simi-
lar comparison holds, including also the 16-membered mac-
rocycle, for the aforementioned tin compound [23]. Thus
electronic needs seem to favor this elusive conformation.
Formal cyclotetramerisation of cyclic germanium(II)
amidindiide {Ge[N(tBu)–C(Ph)–N]} via the Ge–N linkage
defines intramolecular [Ge1–N1(tBu) 2.043(1)/Ge2–
N22(tBu) 2.040(1) and Ge1–N2 1.990(1)/Ge2–N21
1.994(1)] and intermolecular [Ge2–N2 1.901(1)/Ge1–
Me
N
R
Me
Si
R
N
Me
Si
Me
N
Me
Si
N
Me
T
T
R
T
T
R
N
T
N
T
R
N
R
N
N
N
N
Si
Si
N
Si Me
TS
TS
Me Me
Me Me
Me
s4a: R=Me, T=Ph
s4b: R=iPr, T=Ph
s4c: R=tBu, T=Ph
s4e: R=Me, T=Me
s3a: R=Me, T=Ph
s3b: R=iPr, T=Ph
s3c: R=tBu, T=Ph
s3e: R=Me, T=Me
s2a: R=Me, T=Ph
s2b: R=iPr, T=Ph
s2c: R=tBu, T=Ph
s2e: R=Me, T=Me
TS
TS
R
N
R
Me
Me
N
Me
Si
Me
Me
Me
N
T
Si
N
T
N
R
T
Si
N
T
N
N
Si
Me
N
N
Si
T
TS
TS
R
T
NSi
Me
Me
Me Me
N
Me
R
R
a2a: R=Me, T=Ph
a2b: R=iPr, T=Ph
a2c: R=tBu, T=Ph
a2e: R=Me, T=Me
a3a: R=Me, T=Ph
a3b: R=iPr, T=Ph
a3c: R=tBu, T=Ph
a3e: R=Me, T=Me
a4a: R=Me, T=Ph
a4b: R=iPr, T=Ph
a4c: R=tBu, T=Ph
a4d: R=Me, T=Me
Scheme 1. Overview of the calculated pathways.
Table 1
Free energies (DG, kcal/mol) of all ring isomers (relative to the six-membered anti compounds a3a, a3b, a3c, or a3e, respectively)
Ring size !
8 (anti)
6 (anti)
4 (anti)
8 (syn)
6 (syn)
4 (syn)
R = Me, T = Ph
+6.5 (a2a)
+11.7 (a2b)
0.0 (a3a)
0.0 (a3b)
ꢀ0.9 (a4a)
+8.1 (s2a)
+14.4 (s2b)
ꢀ0.4 (s3a)
ꢀ0.1 (s3b)
+0.3 (s3c)
ꢀ2.3 (s3e)
ꢀ2.0 (s4a)
R = iPr, T = Ph
ꢀ11.8 (a4b)
ꢀ10.7 (s4b)
R = tBu, T = Ph
R = Me, T = Me
+17.1 (a2c)
+6.4 (a2e)
0.0 (a3c)
0.0 (a3e)
ꢀ12.6 (a4c)
+19.4 (s2c)
+10.1 (s2e)
ꢀ11.6 (s4c)
ꢀ1.5 (a4e)
ꢀ1.9 (s4e)

2794
T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
TSs3a>s4a
TSs3b>s3b
20
TSa3a>a4a
TSa3b>a4b
15.8
TSa3a>s3a
TSa3b>s3b
13.9
15
10
5
TSs4a>a4a
13.1
10.3
11.8
11.2
8.1
TSs4b>a4b
2.4
a4a
-0.9
-0.1
-0.4
0
1
2
3
4
5
6
7
8
9
0.0
-0.9
a4a
-2.0
s4a
s3b
s3a
a3a / a3b
-5
six-membered ring isomers
anti syn
-10
a4b
-10.7
a2b
-11.7
s4b
four-membered ring isomers
-11.7
-15
four-membered ring isomer
anti
syn anti
Scheme 2. Reaction pathways for R = iPr (solid line) and R = Me (dashed line). The pathways to (or from) the eight-membered rings have been omitted
for clarity.
3. DFT calculations
of the barrier height for the formation of the different
species, we therefore took a look at the kinetics and
the location of the transition states in order to explain
the experimental result.
Scheme 2 gives an overview over the reaction path-
ways for the conversion between the four- and the
six-membered compounds for the two experimentally
investigated systems. Eq. (8) gives an example for the
conversion of the (syn) six-membered ring to the corre-
sponding four-membered isomer. Also the two rotational
transition states that connect the syn and the anti
conformers of the two ring systems have been considered
(see Eq. (9)).
To determine the preferred ring size of the silicon–nitro-
gen heterocyclic compounds 2–4 in dependence of the N-
and C-substituents, DFT calculations (B3LYP/6-31G*)
were carried out to compare the energies of the four-,
six-, and eight-membered isomers and to locate the transi-
tion states for the transformation between different ring
sizes (see Scheme 1).
But not only the ring size itself leads to different isomers,
furthermore syn-/anti-conformers for the four- and the six-
membered rings have to be considered (see Scheme 1). The
relative free energies of the ground states of the two exper-
imentally investigated systems (a: R = Me/T = Ph,
b: R = iPr/T = Ph) and of the two systems studied only
computationally (c: R = tBu/T = Ph, e: R = T = Me) are
given in Table 1.
From Table 1 it is evident, that the originally postulated
eight-membered ring [1] is endergonic compared to the
smaller ring isomers. The larger isopropyl substituent
favors the four-membered ring by 11–12 kcal/mol with
respect to the six-membered ring. Additionally, the four-
membered anti (a4b) ring conformer is 1.1 kcal/mol more
exergonic than the syn (s4b) compound. This is in agree-
ment with the crystal structure that could be determined
for compound a4b.
For R = Me the energy differences between the four-
and the six-membered rings are smaller with only
1–2 kcal/mol. In this case the syn conformer of the
four-membered ring (s4a) is the most exergonic species.
Experimentally a crystal structure of the six-membered
ring (syn) s3a was determined. This might be a result
Atomic distances:
Si1-N1 1.95
Si1-N2 2.11
Si2-N2 1.73
N1-C1 1.34
N2-C1 1.32
Atomic distances:
Si1-N1 1.93
Si1-N2 2.15
Si2-N2 1.73
N1-C1 1.34
N2-C1 1.32
Si1
Si1
N1
N1
Si2
C1
N2
C1
Si2
N2
Fig. 4. Transition states TSs3a > s4a (left) and TSs3b > s4b (right)
together with selected atomic distances.

T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
2795
R
N
Me
N
Me
Me
Si
Me
Si
Me
N
R
T
T
N
T
Me
N
T
R
Si
Si
T
N
N
N
T
N
N
N
Si
R
N
Si
R
Me
Me
Me Me
Me Me
R
a4a: R=Me, T=Ph
a4b: R=iPr, T=Ph
a4c: R=tBu, T=Ph
a4e: R=Me, T=Me
TSa3a>a4a: R=Me, T=Ph
TSa3b>a4b: R=iPr, T=Ph
TSa3c>a4c: R=tBu, T=Ph
TSa3e>a4e: R=Me, T=Me
a3a: R=Me, T=Ph
a3b: R=iPr, T=Ph
a3c: R=tBu, T=Ph
a3e: R=Me, T=Me
ð8Þ
ð9Þ
R
R
R
Me
Me
Me
Me
N
N
T
Me
N
T
Me
N
T
Si
N
Si
N
R
Si
N
N
T
N
R
N
N
Si
Si
Si
T
Me Me
Me Me
Me Me
N
T
R
TSa3a>s3a: R=Me, T=Ph
TSa3b>s3b: R=iPr, T=Ph
TSa3c>s3c: R=tBu, T=Ph
TSa3e>s4e: R=Me, T=Me
s3a: R=Me, T=Ph
s3b: R=iPr, T=Ph
s3c: R=tBu, T=Ph
s3d: R=Me, T=Me
a3a: R=Me, T=Ph
a3b: R=iPr, T=Ph
a3c: R=tBu, T=Ph
a3e: R=Me, T=Me
The barrier heights for the conversion of the six-mem-
bered rings to the four-membered isomers are similar
10 kcal/mol and the thermodynamically most favorable
species s4b has to overcome higher barriers than what
is needed for the formation of s3a, which was observed
in the crystal structure. But the small energy differences
of the isomers also call for a rapid conversion in solution
as there is no species which is really favored thermody-
namically. The structures of the transition states
TSs3a > s4a and TSs3b > s4b are compared in Fig. 4.
They are very similar, only small differences in the Si1–
N2 and Si1–N1 distances can be observed.
for
R = Me
(TSa3a > a4a,
13.9 kcal/mol
and
TSs3a > s4a, 15.8 kcal/mol) and R = iPr (TSa3b > a4b,
11.8 kcal/mol and TSs3b > s4b, 13.1 kcal/mol). While in
the case of the isopropyl substituent the lower barrier
leads to the thermodynamically most favorable species,
which has also been observed in the crystal structure,
the situation is different for the methyl substituent. The
ground states are energetically less favored by about
TSs3e>s3e
TSs3c>s3c
16.7
20
15
10
5
TSa3e>a4e
TSa3c>a4c
14.3
TSa3c>s3c
TSa3e>s3e
13.4
TSs4e>a4e
12.4
8.9
8.6
4.9
TSs4c>a4c
0.5
0.3
s3c
s3e
1
2
3
4
5
6
7
8
9
0
0.0
a3c / a3e
-1.6
a4e
-1.6
a4e
-1.9
-2.3
-5
six-membered ring isomers
anti syn
s4e
-10
-15
a4c
-11.6
a4c
-12.5
-12.5
s4c
four-membered ring isomers
syn anti
four-membered ring isomer
anti
Scheme 3. Reaction pathways for the system R = tBu/T = Ph (solid line) and R = T = Me (dashed line). The pathways to (or from) the eight-membered
rings have been omitted for clarity.

2796
T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
For the sterically more demanding R = tBu an even
ment, relative intensity). NMR spectra were recorded on a
JEOL GX 270 spectrometer (1H 270.17, 13C{1H}
67.94 MHz). Chemical shifts refer to TMS (d = 0) as inter-
nal standard and are reported in ppm, coupling constants J
in Hz. Measurements were carried out in C6D6 (25 °C)
and/or at varied temperatures in [d8] toluene.
Amidinates 1a, 1b, 1d [6] were prepared as described in
the literature, 1c according to this procedure [3]. All
reagents, including Me2SiCl2 and GeCl2 Æ dioxane, were
from commercial sources and purified, when necessary.
more exergonic formation of the four-membered ring is
predicted. In analogy to what was observed for the isopro-
pyl group, the anti isomer a4c is the most feasible species
(Scheme 3). To check whether the substituent T on the car-
bon atom can exert an influence we changed the phenyl
substituent in system a (T = Ph, R = Me) to the smaller
methyl group (e: R = T = Me). For the methyl/methyl
compounds (e) the different isomers show no real preferred
conformation, a mixture of more than one of the isomers
and rapid interconversion is predicted for that substituent
combination.
5.2. {Me2Si[N–C(Ph)–N(Me)]}2 (s3a)
4. Conclusion
To a solution of 21.25 mmol MeNH2 (0.66 g) in 50 ml
Et2O were added 6.06 ml of a 1.65 molar solution of nBuLi
in hexane (10 mmol) at ꢀ78 °C by means of a pipette.
Slowly, the reaction mixture is allowed to warm to r.t.
and after 20 h, the solvent and excess MeNH2 was removed
in vacuo. The residue was again suspended in 50 ml of Et2O
and at ꢀ78°C, 9.80 mmol of PhCN (1.01 g) were added
under stirring. After warming to r.t. and further stirring
for 1 h, 9.80 mmol of NEt3 (0.99 g) were added. After 1 h
and cooling to ꢀ78 °C, 9.80 mmol of Me2SiCl2 (1.26 g)
were added and again allowed to warm to r.t. After 18 h,
volatiles were removed in vacuo and the solid, white residue
was extracted with several portions of pentane (50 ml).
Slow evaporation of the solvent followed by cooling to
10 °C for several days afforded large, colorless crystals
(m.p. 138 °C) of s3a Yield: 1.62 g (86.9%). C20H28N4Si2:
MS(CI, 150 eV): m/z (%): 381 (90) [M], 365 (100) [MꢀMe],
349 (17) [Mꢀ2Me], 191 (29) [Me2SiNC(Ph)@NMe], 174 (6)
[MeSiNC(Ph)@NMe]. 1H NMR ([d8]toluene): Si(CH3):
ꢀ80 °C: d 0.24 [s, 6H], 0.81 [s, 6H]; ꢀ60 °C: d 0.28 [s,
6H], 0.74 [s, 6H]; 40 °C: d 0.12 [s, 6H], 0.57 [s, 6H]; 60
°C: d 0.15 [s (br), 6H], 0.57 [s (br), 6H]; 85 °C: d 0.17 [s,
6H], 0.42 [s, 6H]; –N(CH3): ꢀ 80 °C: d 2.37 [s, 1H, –
N(H–CH2)], 2.40 [s, 2H, –N(HC–H2)], 2.97 [s, 3H,
@NCH3]; ꢀ60 °C: d 2.45 [s, 3H, –N(CH3)], 2.93 [s, 3H,
@N(CH3)]; 40 °C: d 2.65 [s (br), 6H]; 60 °C: d 2.65 [s,
6H]; 85 °C: d 2.60 [s, 6H]; –C6H5: ꢀ80 to 85 °C: d 6.93–
7.18 [m, 10H]. 13C{1H} NMR (C6D6, 25 °C): d 0.97 [s,
Si(CH3)], 25.24 [s, N(CH3)], 127.28 [s, p-C6H5], 128.81 [s,
m-C6H5], 129.00 [s, o-C6H5], 137.21 [s, i-C6H5]; 29Si{1H}
NMR (C6D6, 25 °C): d ꢀ0.52 (s). Anal. Calc. for
C20H28N4Si2: C, 63.11; H, 7.41; N, 14.72. Found: C,
60.02; H, 7.80; N, 12.93%.
The deprotonation of N–H amidinates at silicon centers
by a base is very sensitive to steric influences. With N-
methyl or N-ipropyl substituents, six- or four-membered
ring dimers may be obtained, with N-tbutyl substituents,
deprotonation needs the bigger coordination center Ge(II),
whereby a tetrameric arrangement is formed. Formation of
eight-membered ring dimers, as described in the earlier lit-
erature, is not reproducible. Nevertheless, in line with ear-
lier observations on the fluctional nature of amidinate
ligands at silicon centers, low energy pathways exist for
the mutual rearrangement of those dimeric ring molecules
via higher coordination numbers at silicon and the four-,
six- and eight-membered rings differ energetically, together
with their respective rotational isomers, not grossly. The
existing differences favor the four-membered rings, unless
the C- and N-substituents are very small (C–Me, N–Me:
six-membered ring). Obviously, solid-state effects may
overcome this small difference and lead also to a six-mem-
bered ring for C–Ph, N–Me substitution.
In order to achieve the amidinate bridge formation in
polymeric silicon compounds, the size of the C and N sub-
stituents must be reduced further. Hydrogen substituted or
cyclic amidinates would be candidates for further studies in
this field.
5. Experimental
5.1. General
All operations were performed under an atmosphere of
dry, oxygen free nitrogen and with thoroughly dried glass-
ware by means of standard high-vacuum-line techniques.
Solvents were distilled under nitrogen from K–Na alloy.
Elemental analyses were performed with a Vario EL
CHN analyzer of the Analytisches Laboratorium der TU
Mu¨nchen. In our hands, silicon compounds of this kind
tend to form silicon carbide during combustion and conse-
quently, elemental analysis data are not reliable, particu-
larly so for carbon. Therefore, additional mass spectra
were recorded for analytical purposes. Chemical ionisation
(CI) mass spectral data were obtained employing a Varian
Mat 311A spectrometer; peaks are reported as m/z (assign-
5.3. [Me2SiNC(Ph) = NiPr]2 (a4b)
The analogous procedure as for s3a was used, but
Me2SiCl2 was added at r.t. and for the extraction, toluene
was used. Reagents: 15.23 mmol iPrNH2 (0.9 g),
13.09 mmol (7.7 ml of 1.7 molar solution) of nBuLi,
13.07 mmol PhCN (1.53 g), 13.04 mmol NEt3 (1.32 g),
9.8 mmol Me2SiCl2 (1.26 g). Colorless crystals (m.p.
146 °C) are obtained. Yield:1.04 g, 36.5%. C24H36N4Si2:
i
MS (CI, 150 eV): m/z (%): 436 (63) [M], 393 (33) [Mꢀ Pr],

T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
2797
318 (100) [Mꢀ2Me, ꢀ2iPr], 290 (15) [Mꢀ4Me, ꢀ2iPr], 219
(36) [Me2SiNC(Ph)@NiPr]. 1H NMR (C6D6, 25 °C): d 0.48
(s, 12H, SiCH3), 1.12 (d, 12H, 3J = 6.7, CH3CH), 3.27
(sept, 2H, 3J = 6.7, CH3CH), 7.04 [m, C6H5]. 13C{1H}
NMR (C6D6, 25 °C): d 1.79 (SiCH3), 25.45 (CH3CH),
50.31 (CH3CH), 127.12 (p-C6H5), 128.18 (m-C6H5),
128.47 (o-C6H5), 138.01 (i-C6H5), 158.01 (NCN).
29Si{1H} NMR (C6D6, 25 °C): d 3.88. Anal. Calc. for
C24H36N4Si2: C, 66.00; H, 8.31; N, 12.83. Found: C,
61.61; H, 8.09; N, 11.60%.
tions (70 ml) gave colorless crystals of 5c (m.p. 186 °C)
after slowly evaporating the solvent to dryness. Yield
0.31 g (60%). In an independent experiment, DBU in
THF was used as a base and 5c was obtained in 89.7%
yield. C44H56N8Ge4: MS (74Ge, CI, 150 eV): m/z (%): 989
(25) [M], 249 (10) [M/4], 177 (100) [H2NC(Ph)NHtBu].
1H NMR (C6D6, 25 °C): d 1.15 (s, 36H, CCH3), 7.16–
7.62 (m, 20H, C6H5). 13C{1H} NMR (C6D6, 25 °C): d
32.14 (CCH3), 52.47 (CCH3), 125.94, 127.63, 129.31,
177.97 (C6H5), 164.23 (NCN). Anal. Calc. for C, 53.52;
H, 5.72; N, 11.35. Found: C, 52.68; H, 5.82; N, 10.70%.
5.4. {Ge[N(tBu)C(Ph)N]}4 (5c)
5.5. X-ray crystallography
To a suspension of 5.76 mmol Li[NH(tBu)AC(Ph)@N]
(1.05 g) in 50 ml Et2O was added slowly 2.88 mmol
GeCl2 Æ dioxane (0.67 g) at ꢀ78 °C under stirring and the
mixture allowed to warm to r.t. After 20 h, the solvent
was removed in vacuo and the residue extracted several
times with pentane. The combined pentane (50 ml) solu-
Crystals of s3a, a4b, and 5c were prepared under argon in
a matrix of perfluorinated polyether. Specimens of suitable
quality and size were mounted on the ends of quartz fibers
in F06206R oil and used for measurements of precise cell
constants and intensity data collection on a Nonius
DIP2020 diffractometer, using graphite-monochromated
Mo Ka radiation (k = 0.71073 A). The structures were
solved by a combination of direct methods (SHELXS-97 [26])
and difference-Fourier syntheses and refined by full matrix
tions contained
a
brown, solid residue (0.31 g):
NH(tBu)AC(Ph)@NH and 5c after drying (NMR). The
solid residue of the former extraction was extracted with
several portions of toluene. The combined toluene solu-
˚
Table 2
Crystal data, data collection, and structure refinement for compounds s3a, a4b, and 5c
s3a
a4b
5c
Crystal data
Formula
Mr
C
20H28N4Si2
C24H36N4Si2
436.75
Monoclinic
P21/c
8.6880(2)
8.9290(2)
16.6550(5)
90
97.642(1)
90
1280.5(1)
1.133
C44H56Ge4N8
987.33
Monoclinic
C2/c
23.1928(3)
10.2179(1)
23.2167(2)
90
90.232(1)
90
320.64
Orthorhombic
Pna21
14.3687(3)
9.9001(2)
15.1940(3)
90
90
90
2161.4(1)
1.170
Crystal system
Space group
˚
a (A)
˚
b (A)
˚
c (A)
a (°)
b (°)
c (°)
3
˚
V (A )
5501.9(1)
1.192
qcalc (g cmꢀ3
)
Z
4
2
4
F(000)
l(Mo Ka) (mmꢀ1
816
0.175
472
0.156
2016
2.197
)
Data collection
T (K)
143
143
143
hkl Range
Measured reflections
0 ! 18, 0 ! 12, ꢀ19 ! 0
0 ! 10, 0 ! 11, ꢀ21 ! 20
0 ! 30, 0 ! 13, ꢀ30 ! 29
71646
2703
82580
Unique reflections [Rint
Reflections used for refinement
]
2470 [0.030]
2470
2703 [0.039]
2703
6300 [0.037]
6300
Refinement
Refined parameters
235
140
366
Final R values [I P 2r(I)]
R1
a
0.0341
0.0950
0.096
0.0389
0.0998
–
0.0240
0.0588
–
b
wR2
Absolute structure parameter
(shifꢀt/3error)max
<0.001
0.208/ꢀ0.274
<0.001
0.203/ꢀ0.330
<0.001
0.269/ꢀ0.410
˚
(e A
)
ꢀ
ꢁ
ꢀ
ꢁ
2
w ¼ 1=½r2 F o2 þ ðapÞ þ bpꢁ; p ¼ F o2 þ 2F 2c =3; a = 0.0743 (s3a), 0.0492 (a4b), 0.0368 (5c); b = 0.21 (s3a), 0.57 (a4b), 0.62 (5c).
P
P
a
R ¼ ðjF oj ꢀ jF cjÞ= jF oj.
ꢀ
ꢁ
P
P
2
2
1=2
b
wR2 ¼ f½ w F 2o ꢀ F c2 ꢁ= ½wðF 2oÞ ꢁg
.

2798
T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
least-squares calculations on F2 (SHELXL-97 [26]). In the case
of 5c the scattering contribution of unidentified and highly
disordered solvent in voids of the crystal structure was
accounted for by means of the program SQUEEZE [27]. It
was carefully checked, that this treatment did not affect the
structural parameters: details are given in the supporting
information. The thermal motion was treated anisotropi-
cally for all non-hydrogen atoms. All hydrogen atoms were
calculated and allowed to ride on their parent atoms with
fixed isotropic contributions. Further information on crystal
data, data collection and structure refinement are summa-
rized in Table 2. Important interatomic distances and angles
are shown in the corresponding figure captions.
References
[1] O.J. Scherer, P. Hornig, Angew. Chem. 79 (1967) 60.
[2] H.H. Karsch, P.A. Schlueter, in: N. Auner, J. Weis (Eds.), Organo-
silicon Chemistry IV – From Molecules to Materials, Wiley-VCH,
Weinheim, 2000, p. 287.
[3] T. Segmuller, PhD thesis, Technische Universita¨t, Munchen, 2003.
¨
¨
[4] H.H. Karsch, T. Segmuller, in: N. Auner, J. Weis (Eds.), Organo-
¨
silicon Compounds VI – From Molecules to Materials, Wiley-VCH,
Weinheim, 2005, p. 194.
[5] H.H. Karsch, T. Segmueller, in: N. Auner, J. Weis (Eds.), Organo-
silicon Chemistry V – From Molecules to Materials, Wiley-VCH,
Weinheim, 2003, p. 270.
[6] O.J. Scherer, P. Hornig, Chem. Ber. 101 (1968) 2533.
[7] H.H. Karsch, P.A. Schlueter, in: N. Auner, J. Weis (Eds.), Organo-
silicon Chemistry III – From Molecules to Materials, Wiley-VCH,
Weinheim, 1998, p. 53.
[8] H.H. Karsch, F. Bienlein, in: N. Auner, J. Weis (Eds.), Organosilicon
Chemistry II – From Molecules to Materials, Wiley-VCH, Weinheim,
1995, p. 133.
5.6. DFT calculations
All calculations were performed with the software pack-
age GAUSSIAN03 [28] using the density functional/Hartree–
Fock hybrid model Becke3LYP [29–32] and the split
valence double-f (DZ) basis set 6-31G* [33–35]. No sym-
metry or internal coordinate constraints were applied dur-
ing optimizations. All reported intermediates were verified
as true minima by the absence of negative eigenvalues in
the vibrational frequency analysis. Transition-state struc-
tures were located using the Berny algorithm [36] until
the Hessian matrix had only one imaginary eigenvalue.
The identity of all transition states was confirmed by
IRC calculations and by animating the negative eigenvec-
tor coordinate with MOLDEN [37] and GaussView [38].
Approximate free energies (DG) and enthalpies (DH)
were obtained through thermochemical analysis of fre-
quency calculations, using the thermal correction to Gibbs
free energy as reported by GAUSSIAN-03. This takes into
account zero-point effects, thermal enthalpy corrections
and entropy. All energies reported in this paper, unless
otherwise noted, are free energies or enthalpies at 298 K,
using unscaled frequencies. All transition states are max-
ima on the electronic potential energy surface (PES), which
may not correspond to maxima on the free energy surface.
[9] P.J. Wheatley, J. Chem. Soc. (1962) 1721.
[10] L. Parkanyi, G. Argay, P. Hencsei, J. Nagy, J. Organomet. Chem.
116 (1976) 299.
[11] W. Clegg, U. Klingebiel, G.M. Sheldrick, N. Vater, Z. Anorg. Allg.
Chem. 482 (1981) 88.
[12] L. Bihatsi, P. Hencsei, L. Parkanyi, J. Organomet. Chem. 219 (1981)
145.
[13] W. Clegg, U. Klingebiel, G.M. Sheldrick, Z. Naturforsch. 37B (1982)
423.
[14] A. Szollosy, L. Parkanyi, L. Bihatsi, P. Hencsei, J. Organomet.
Chem. 251 (1983) 159.
[15] W. Clegg, M. Haase, G.M. Sheldrick, N. Vater, Acta Crystallogr.,
Sect. C 40 (1984) 871.
[16] N. Auner, E. Herdtweck, E. Penzenstaudler, Acta Crystallogr., Sect.
C 49 (1993) 359.
[17] L. Parkanyi, L. Bihatsi, P. Hencsei, Z. Kristallogr. 209 (1994) 905.
[18] S.A.A. Shah, H.W. Roesky, P. Lubini, H.-G. Schmidt, Acta
Crystallogr., Sect. C 52 (1996) 2810.
[19] H.H. Karsch, P.A. Schlueter, M. Reisky, Eur. J. Inorg. Chem. (1998)
433.
[20] S.R. Foley, C. Bensimon, D.S. Richeson, J. Am. Chem. Soc. 119
(1997) 10359.
[21] S.R. Foley, Y. Zhou, G.P.A. Yap, D.S. Richeson, Inorg. Chem. 39
(2000) 924.
[22] M. Veith, Angew. Chem. 99 (1987) 1.
[23] M. Veith, M. Grosser, J. Organometal. Chem. 229 (1982) 247.
[24] G.A. Wiegers, A. Vos, Acta Crystallogr. 14 (1961) 562.
[25] G.A. Wiegers, A. Vos, Acta Crystallogr. 16 (1963) 152.
[26] G.M. Sheldrick, SHELXL-97 – Program for Crystal Structure Solution,
Acknowledgments
Institut fur Anorganische Chemie der Universita¨t, Go¨ttingen, Ger-
¨
many, 1997.
This research was financially supported by the Fonds
der Chemischen Industrie and the Lorenz foundation.
[27] A.L. Spek, Acta Crystallogr., Sect. A 46 (1990) C34.
[28] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, J.A. Montgomery Jr., T. Vreven, K.N. Kudin, J.C.
Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B.
Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H.
Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene,
X. Li, J.E. Knox, H.P. Hratchian, J.B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin,
R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma,
G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S.
Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D.
Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G.
Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T.
Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challa-
Appendix A. Supplementary material
CCDC 636083, 626000 and 636082 contain the supple-
mentary crystallographic data for s3a, a4b and 5c. These
data can be obtained free of charge via http://www.ccdc.ca-
m.ac.uk/conts/retrieving.html, or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge
CB2 1EZ, UK; fax: (+44) 1223-336-033; or e-mail: depos-
it@ccdc.cam.ac.uk. Supplementary data associated with
this article can be found, in the online version, at
doi:10.1016/j.jorganchem.2007.02.032.

T. Segmu¨ller et al. / Journal of Organometallic Chemistry 692 (2007) 2789–2799
2799
combe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C.
Gonzalez, J.A. Pople, Gaussian 03, Revision C.02, Gaussian Inc.,
Wallington, CT, 2004.
[34] M.M. Mancl, W.J. Petro, W.J. Hehre, J.S. Binkley, M.S. Gordon,
D.J. DeFree, J.A. Pople, J. Chem. Phys. 77 (1982) 3654.
[35] V. Rassolov, J.A. Pople, M. Ratner, T.L. Windus, J. Chem. Phys.
109 (1998) 1223.
[36] H.B. Schlegel, J. Comp. Chem. 3 (1982) 214.
[37] G. Schaftenaar, J.H. Noordik, J. Comput.-Aided Mol. Des. 14 (2000)
123.
[38] R.I. Dennington, T. Keith, J.M. Millam, K. Eppinnett, W.L.
Hovell, R. Gilliland, GaussView, Gaussian Inc., Wallingford, CT,
2003.
[29] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B: Condens. Matter 37
(1988) 785.
[30] S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200.
[31] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
[32] P.J.D. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, J. Phys.
Chem. 98 (1994) 11623.
[33] P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213.