Atomic contributions from spin-orbit coupling to 29Si NMR chemical shifts in metallasilatrane complexes

Article abstract of DOI:10.1002/chem.201200746

New members of a novel class of metallasilatrane complexes [X-Si-(μ-mt)₄-M-Y], with M=Ni, Pd, Pt, X=F, Cl, Y=Cl, Br, I, and mt=2-mercapto-1-methylimidazolide, have been synthesized and characterized structurally by X-ray diffraction and by ²⁹Si solid-state NMR. Spin-orbit (SO) effects on the ²⁹Si chemical shifts induced by the metal, by the sulfur atoms in the ligand, and by heavy halide ligands Y=Cl, Br, I were investigated with the help of relativistic density functional calculations. Operators used in the calculations were constructed such that SO coupling can selectively be switched off for certain atoms. The unexpectedly large SO effects on the ²⁹Si shielding in the Ni complex with X=Y=Cl reported recently originate directly from the Ni atom, not from other moderately heavy atoms in the complex. With respect to Pd, SO effects are amplified for Ni owing to its smaller ligand-field splitting, despite the smaller nuclear charge. In the X=Cl, Y=Cl, Br, I series of complexes the Y ligand strongly modulates the ²⁹Si shift by amplifying or suppressing the metal SO effects. The pronounced delocalization of the partially covalent M←Y bond plays an important role in modulating the ²⁹Si shielding. We also demonstrate an influence from the X ligand on the ²⁹Si SO shielding contributions originating at Y. The NMR spectra for [X-Si-(μ-mt) ₄-M-Y] must be interpreted mainly based on electronic and relativistic effects, rather than structural differences between the complexes. The results highlight the sometimes unintuitive role of SO coupling in NMR spectra of complexes containing heavy atoms. All in a spin! The class of metallasilatrane complexes [X-Si-(μ-mt)₄-M-Y] with M=Ni, Pd, and Pt, previously reported for X=Y=Cl (shown here), has been extended by new members with X=F and Y=Br and I. New synthetic routes, structural characterizations by X-ray diffraction, and ²⁹Si solid-state NMR data are reported. Spin-orbit effects on the ²⁹Si chemical shifts were investigated with the help of relativistic density functional calculations. Copyright

Full text of DOI:10.1002/chem.201200746

FULL PAPER
DOI: 10.1002/chem.201200746
Atomic Contributions from Spin-Orbit Coupling to 29Si NMR Chemical Shifts
in Metallasilatrane Complexes
Jochen Autschbach,*[a] Kiplangat Sutter,[a] Lionel A. Truflandier,[a] Erica Brendler,[b] and
Jçrg Wagler[c]
Abstract: New members of a novel
class of metallasilatrane complexes [X-
Si-(m-mt)4-M-Y], with M=Ni, Pd, Pt,
X=F, Cl, Y =Cl, Br, I, and mt=2-mer-
capto-1-methylimidazolide, have been
synthesized and characterized structur-
ally by X-ray diffraction and by 29Si
solid-state NMR. Spin-orbit (SO) ef-
fects on the 29Si chemical shifts induced
by the metal, by the sulfur atoms in the
ligand, and by heavy halide ligands Y=
Cl, Br, I were investigated with the
help of relativistic density functional
calculations. Operators used in the cal-
culations were constructed such that
SO coupling can selectively be switch-
ed off for certain atoms. The unexpect-
edly large SO effects on the 29Si shield-
ing in the Ni complex with X=Y=Cl
reported recently originate directly
from the Ni atom, not from other mod-
erately heavy atoms in the complex.
With respect to Pd, SO effects are am-
plified for Ni owing to its smaller
ligand-field splitting, despite the small-
er nuclear charge. In the X=Cl, Y=
Cl, Br, I series of complexes the Y
ligand strongly modulates the 29Si shift
by amplifying or suppressing the metal
SO effects. The pronounced delocaliza-
!
tion of the partially covalent M
Y
bond plays an important role in modu-
lating the 29Si shielding. We also dem-
onstrate an influence from the X ligand
on the 29Si SO shielding contributions
originating at Y. The NMR spectra for
[X-Si-(m-mt)4-M-Y] must be interpret-
ed mainly based on electronic and rela-
tivistic effects, rather than structural
differences between the complexes.
The results highlight the sometimes un-
intuitive role of SO coupling in NMR
spectra of complexes containing heavy
atoms.
Keywords: density functional calcu-
lations · metallasilatranes · NMR
spectroscopy · silicon · structure
elucidation
Introduction
ecules upon substitution of light elements with heavy coun-
terparts that otherwise have similar chemical properties. A
well-known example is the effect on NMR shifts of light
nuclei due to the presence of a heavy element in the vicinity.
Substitution of Cl by Br and I, for instance, causes carbon
and proton nuclei bound to the halide to be more strongly
shielded magnetically, and increasingly so as the halide be-
comes heavier (normal halogen dependence or NHD). Al-
ready several decades ago, this heavy atom effect on NMR
shielding was identified as a consequence of spin-orbit (SO)
coupling at the heavy atom[5] which is a purely relativistic
effect. NHD-like effects are known to be caused by heavy
transition metals as well.[2,6,7] A recent interesting example
is the SO effect on proton chemical shifts in b-agostic inter-
actions in late-transition-metal complexes.[8] A similarity be-
tween NMR J coupling and SO effects on nuclear shielding
has been demonstrated by calculations.[9,10] For instance,
both exhibit similar Karplus-like geometry dependencies
and both are mediated by covalent interactions between the
light and the heavy atom and are strongly dependent on the
s character of the bonding orbitals. Other experimental ob-
servations of the consequences of relativity in NMR spectro-
scopy include a study by Forgeron et al., who measured for
XeF2 a 1000 ppm Xe chemical-shift tensor component paral-
lel to the molecular axis.[11] In the absence of relativistic ef-
fects this component would be zero by symmetry. Kantola
et al. have shown that relativistic effects in methyl halides
Relativistic effects in the chemistry of heavy element com-
pounds have long fascinated scientists.[1] In particular when
NMR parameters of molecules with heavy atoms are of con-
cern, relativistic effects can be spectacularly large.[2–4] Since
the world is fully relativistic, so to say, relativistic effects are
not observable in a strict sense. However, they manifest in
numerous trends or anomalies seen in the properties of mol-
[a] Prof. J. Autschbach, K. Sutter, Dr. L. A. Truflandier
Department of Chemistry, University at Buffalo
State University of New York Buffalo, NY 14260-3000 (USA)
E-mail: jochena@buffalo.edu
[b] Dr. E. Brendler
Institut fꢀr Analytische Chemie
Technische Universitꢁt Bergakademie Freiberg
09596 Freiberg (Germany)
[c] Dr. J. Wagler
Institut fꢀr Anorganische Chemie
Technische Universitꢁt Bergakademie Freiberg
09596 Freiberg (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201200746. It contains details of
the experimental procedures and the X-ray structure analyses, along
with additional computational data (calculated atomic charges, results
for COSMO calculations, calculated results for optimized geometries,
full plot of calculated versus experimental 29Si shifts).
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

not only modifies the isotropic carbon shifts (the aforemen-
tioned NHD), but also have a strong impact on the 13C-
shielding tensor anisotropies.[12] The impact on the tensor
properties of nuclear magnetic shielding evidently has im-
portant implications regarding the interpretation of solid-
state NMR spectra. We note in passing that relativistic ef-
fects on J-coupling constants can be extremely large.[2]
We have recently reported
plexes and correlated the trends in the SO effects with cal-
culated excitation energies and with the magnitude of orbi-
tal energy gaps of metal-centered orbitals.[17]
Herein, we use an all-electron theoretical method in
which the SO coupling for selected atoms in a molecule can
be “switched off”, to study in detail the origin of SO effects
on the 29Si shielding in [X-Si-(m-mt)4-M-Y] metallasilatranes.
On the experimental front, the family of compounds has
been extended significantly to include systematic variations
of the ligand Y=Cl, Br, I, with X=Cl. The structures of the
complexes have been analyzed by single-crystal X-ray dif-
fraction (XRD), and 29Si solid-state NMR data have been
recorded. In addition, several members of a family of analo-
gous compounds with X=F have been synthesized and char-
acterized by solid-state NMR and XRD.
the synthesis and theoretical
characterization of a new class
of metallasilatrane complexes
[X-Si-(m-mt)4-M-Y]
(mt=2-
mercapto-1-methylimidazolide),
as shown in Figure 1, with M=
Pt, Pd, Ni, and X=Y=Cl.[13,14]
The dative metal–Si bond is in-
dicated by an arrow in the
chemical formula in Figure 1.
There are interesting structural
analogies to a recently reported
Pd–Sb complex (also with a
linear arrangement of the type
Figure 1. Metallasilatrane com-
plexes [X-Si-(m-mt)4-M-Y]
studied in this work. M is a
Group 10 metal, X and Y are
halide ligands, and mt is 2-mer-
capto-1-methylimidazolide.
We demonstrate herein that the Ni anomaly for the X=
Y=Cl complexes is not related to SO effects originating at
the Cl and S ligands, but directly sourced at the metal
!
center and transferred through the Si M bond. The large
magnitude of the Ni SO effects must therefore be a result of
the smaller ligand-field splitting for Ni, which renders the
complex more susceptible to external perturbations (the
magnetic field from the NMR spectrometer in this case) and
amplifies the SO effects on the neighboring Si atom directly
at the metal center. The complexes with X=F and Y=Cl
exhibit similar trends experimentally, with 29Si being most
strongly shielded for Pt and least for Pd, not Ni. For X=Cl,
the SO effects induced by the Y ligand are modest. An anal-
ysis shows that in this case only part of the variation of the
29Si shift with respect to changing Y originates from SO cou-
pling at the Y ligand. Rather, variations of the ligand Y
modulate the SO effects of the metal on 29Si. This modula-
X-E-M-Y), which acts as an optical fluoride sensor.[15] See
also reports of the syntheses and characterizations of
!
E
Au bonded systems with E=Si, Sn, which afford signifi-
cantly bent F-E-Au-Cl moieties.[16] Due to the strongly cova-
!
lent character of the Si M bonds in the metallasilatrane
complexes, electron spin-orbit (SO) coupling was shown to
have a profound impact on the 29Si NMR shielding tensor of
these systems. The isotropic shift of silicon is moved “up-
field” (i.e., Si is more shielded) and the shielding tensor
span is drastically reduced by SO effects. Since SO coupling
is a relativistic effect with a leading order of Z2/c2, as a func-
tion of the nuclear charge Z and the speed of light c, the
impact on the 29Si shielding tensor was found to be largest
for the Pt compound. However, a surprising “Ni anomaly”
was observed, namely the SO effects on the 29Si shielding
tensor were larger for Ni than for Pd. On the basis of the
formal scaling of relativistic effects with the nuclear charges
of the atoms in a compound, this finding has no simple ex-
planation. Structural variations among the metal triad do
not cause the observed 29Si shift variations. In reference [13],
we speculated that SO coupling originating at other heavy
atoms present in the system, such as the Cl and S ligands,
might be causing or contributing to an apparent Ni anomaly.
This hypothesis was motivated by analyses of the electronic
structure of the X=Y=Cl complexes, showing a significant
degree of electron delocalization within the Cl-M-Si-Cl
moiety and its first coordination sphere. However, a direct
theoretical proof of the origin of the unexpectedly large SO
effects on the 29Si shielding tensor in the Ni compound has
so far not been possible. Potentially, another reason for the
observed anomaly is the smaller ligand-field splitting for the
3d complex, which would render it more susceptible to per-
turbations, including those from SO coupling. Indeed, a
recent study by Straka, Kaupp, et al. found similar non-mo-
notonous SO trends on proton shifts in Group 9 metal com-
ꢀ
tion is correlated with the delocalization of the M Y bond
toward the Si center upon varying Y. For one of the X=F
complexes the SO effect from Y is shown to be significantly
enhanced compared to X=Cl, demonstrating a complex in-
terplay of ligand substitution effects in the X-Si-M-Y moiety
due to its delocalized electronic structure.
The theoretical method used for the SO analyses, and
other computational details, are described in the Experi-
mental Section along with synthetic routes used to prepare
new members of the [X-Si-(m-mt)4-M-Y] family of com-
pounds. Experimental and calculated NMR data for X-Si-
(m-mt)4-M-Y with X=Cl, F, Y=Cl, Br, I, and M=Ni, Pd,
Pt, are presented and discussed in the Results and Discus-
sion section. Structural features of the new complexes deter-
mined by XRD are also discussed in this section. A brief
summary is then given under the heading Conclusions.
Experimental Section
Our pioneering paddlewheel complexes [Cl-Si-(m-mt)4-M-Cl] (M=Pt, Pd,
Ni)[13,14] are now joined by a variety of related complexes of the general
type [X-Si-(m-mt)4-M-Y] with X=Cl, F; M=Pt, Pd, Ni and Y=I, Br, Cl,
which became accessible through the synthetic routes depicted in
Figure 2 (for details see the Supporting Information). The general ap-
proach B follows the original procedure A, that is, using two equivalents
&
2
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Metallasilatrane Complexes
FULL PAPER
mt)4PdBr]·8CHCl3), CCDC-857801 (3[FSi
857802 ([ClSi(m-mt)4PdI]·2CHCl3; modification in Cc), CCDC-857803
([Si(m-mt)4PtCl]·3CHCl3), CCDC-857804 ([FSi(m-mt)4PtCl]), CCDC-
857805 (3[ClSi(m-mt)4PtBr]·8CHCl3), CCDC-857806 (3[FSi(m-
mt)4NiCl]·8CHCl3), CCDC-857807 ([FSi(m-mt)4PtI]·2CHCl3), CCDC-
857808 ([ClSi(m-mt)4PtI]·2CHCl3) and CCDC-857809 ([ClSi(m-
ACHTUGNTERN(NUNG m-mt)4PdCl]·8CHCl3), CCDC-
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
mt)4PdI]·2CH3CN) contain the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_re-
quest/cif.
Part of this article is concerned with the calculation of atomic contribu-
tions to spin-orbit (SO) relativistic effects on the silicon chemical shifts
discussed in the Introduction. One possibility to split scalar and SO rela-
tivistic effects calculated for a molecule into atomic contribution would
be to determine these effects directly to some order in cꢀ2 of perturbation
theory, with c being the speed of light, and to perform a Mulliken parti-
tioning of the results. Another possibility would be a selective use of
spin-orbit versus spin-free effective relativistic core potentials, or selec-
tive use of all-electron atomic mean-field SO integrals (for an application
of the latter to obtain an atomic g-tensor partitioning see reference [19]).
We chose a different route for the present all-electron relativistic compu-
tations. The zeroth-order regular approximation (ZORA) was used to
treat relativistic effects variationally in the NMR calculations. The
ZORA one-electron operator used in density functional theory (DFT)
with a local effective self-consistent potential V (in atomic units)[20] is
given by Equation (1) with k defined by Equation (2).
1
2
1
2
ð1Þ
ð2Þ
h ¼ V þ pk ꢁ p þ sðpk ꢁ pÞ
2c2
2c2ꢀV0
k ¼
In Equation (1), s is the 3-vector of 2ꢃ2 Pauli spin matrices, with compo-
nents s=(sx, sy, sz), and p=ꢀi5 is the momentum operator. The second
term on the right hand side of Equation (1) contains the nonrelativistic
kinetic energy as well as the relativistic corrections to the kinetic and po-
tential energy operators furnished by the ZORA approximation. The last
term is the ZORA spin-orbit operator (using pꢃp=0 and/or kꢀ1=
Figure 2. Routes applied for the syntheses of the compounds reported
herein.
kV’/ACHTNUGRTNEUNG
(2c2), the SO operator may be written in various alternative
ways[21,22]).
of ClSi(mt)3 as a source of the Si-(m-mt)4 moiety in the final paddlewheel
compound. The presence of additional halides in the reaction mixture af-
fords the final product with the soft halide (Br or I instead of Cl) bound
to the transition metal and the hard halide (route B’, F instead of Cl)
When the ZORA operator is derived, one finds V’=V in Equations (1)
and (2). In practice, for a number of reasons, the potential V’ used for
the construction of k in Equation (2) is usually approximated by a
model potential. For instance, the ZORA implementation in the Amster-
dam density functional (ADF) program used for this work employs a
sum of atomic nuclear, Hartree, and exchange-correlation (XC) poten-
tials,[23] as in Equation (3).
ꢀ
bound to Si. nBu4N+BF4 proved to be a mild source of fluoride for that
purpose. The related approach C furthermore reveals both the suitability
of (Ph3P)2NiACHTUNGTRENNUNG(NCS)2 as a starting material for the synthesis of F-Si-(m-
mt)4-Ni-Cl and thus the preference of Ni to coordinate the rather hard Cl
atom instead of the competing soft thiocyanate S atom. In a similar fash-
ion, the two different halides used in route D adhere to the same hard/
soft preferences. This route proved X=Cl suitable for generating com-
plexes of the type [X-Si-(m-mt)4-Pd-Y], whereas in the absence of Cl
((Ph3P)2PdBr2 and/or (Ph3P)2PdI2 used without chloride source, i.e.,
Si(mt)4 used as silane source[18]) synthesis of such paddlewheel complexes
failed. Finally, route E reveals halide exchange as a convenient route, de-
pending on the availability of [Cl-Si-(m-mt)4-M-Cl] (M=Pt, Pd) as a
starting material. Along one or more of these routes the following set of
complexes was synthesized: [X-Si-(m-mt)4-M-Y] with X,M,Y=F,Pt,Cl;
F,Pt,I; F,Pd,Cl; F,Ni,Cl; Cl,Pt,Cl;[14] Cl,Pt,Br; Cl,Pt,I; Cl,Pd,Cl;[14]
Cl,Pd,Br; Cl,Pd,I; Cl,Ni,Cl.[14] These complexes were characterized by
single-crystal X-ray crystallography and solid-state 29Si CP/MAS NMR
spectroscopy, both kinds of data serving the major subject of this paper.
Furthermore, some of these complexes were obtained as different sol-
vates or different modifications, which thus permit insights into the flexi-
bility of their seemingly rigid molecular architecture, which actually re-
sponds to the crystal packing.
atoms
X
V0 ꢂ
VA
ð3Þ
A
A related method of using a model potential in constructing k was pio-
neered by van Wꢀllen.[24] The ZORA functionality in the NWChem
quantum chemistry package[25,26] employs a sum of atomic nuclear and
Hartree potentials. Such approximations in the ZORA operator have
been demonstrated to be a practical and accurate way to enable ZORA
relativistic computations for large systems. In a similar fashion, the mag-
netic perturbation operators arising from the ZORA approximation,[2,27]
which contain k and derivatives thereof, can be implemented straight-
forwardly by using an approximation such as Equation (3) and numerical
integration.
Equation (3) leads to a straightforward “atoms in molecules” partitioning
of the potential in k. For k!1 one recovers the nonrelativistic limit of
h, that is, hnrel =V+(1/2)p·p, as well as the nonrelativistic limits of the
magnetic perturbation operators used in NMR calculations. If the poten-
tial of a particular atom or a selected set of atoms is excluded from V’ of
Equation (3), the resulting “partial” ZORA operator effectively treats
this (set of) atom(s) as nonrelativistic, while other atoms for which the
CCDC-857798 ([ClSi
ACHTUNGTRENNUNG(m-mt)4PtI]·2CH3CN), CCDC-857799 ([ClSiACHTUNGTRENNUNG
mt)4PdI]·2CHCl3; modification in C2/c), CCDC-857800 (3[ClSiAHCTUNTGRENNUNG
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

J. Autschbach et al.
potential is included in V’ are treated relativistically. The approach in es-
sence amounts to a position-dependent scaling of the speed of light in
order to suppress relativistic effects locally, with the atomic potentials in
k providing a suitable and convenient set of switching functions. Fur-
ther, one may use a partial V’ only in the SO term, along with the full V’
in the scalar ZORA operator. This then creates a relativistic operator in
which scalar relativistic effects are fully present, but SO coupling is sup-
pressed for selected atoms. It needs to be stressed that the suppressed
“atom in molecule” contributions are specific to the molecule under con-
sideration and not predetermined for a given atom type. A similar idea
has previously been applied by van Wꢀllen in ZORA calculations of the
magnetic anisotropy of a molecular magnet.[28]
Results and Discussion
Experimental structures: All molecules [X-Si-(m-mt)4-M-Y]
were found to reflect the shape of a fourfold propeller with
the methimazolyl moieties tilted in the same direction. Both
Si and M are located within a square planar environment of
four N or four S atoms, respectively, almost perfectly in
plane. Flexibility of the molecular architecture is reflected
by the variable bond lengths along the XSiMY axis and by
the variable tilt angles of opposing methimazolyl groups for
one compound. These features are summarized in Table 1.
All computations were carried out with a locally modified 2011 developer
version of ADF.[29] A new keyword, partialzora, reads a list of atoms or
atom types for which the atomic potentials are subsequently excluded in
the construction of k. Since the focus of this work is on SO effects on
NMR shielding, the suppression of relativistic effects has been imple-
mented separately for the scalar and the SO ZORA operator for the
ground-state DFT calculations, as outlined in the previous paragraph. In
the computations discussed below, the partial ZORA functionality has
been applied only to suppress SO effects, while keeping scalar relativistic
effects for all atoms. Modifications have been made accordingly in the
NMR nuclear magnetic shielding module in ADF. Here, the partial
ZORA functionality can be chosen separately for magnetic perturbation
operators originating from the scalar and the SO terms of the ZORA
Hamiltonian, respectively, for consistency with the separation of scalar
and SO effects in the ground state calculations.
ꢀ
With respect to M Si bond lengths, one can find an over-
all trend of decreasing interatomic distance when replacing
Ni by Pd by Pt. This is rationalized by the different orbital
contributions to the metal–silicon bond and the resultant in-
crease in covalent bond character when going to Pt (vide
infra) as well as typical variations in the metal atomic radii.
ꢀ
This M Si bond shortening has direct influence on the
ꢀ
trans-disposed Si X bond, which responds with bond length-
ening, to pronounced extent in the case of M=Pt.
As to the influence of the halide, replacing the Si-bound
ꢀ
Cl by F generally results in M Si bond lengthening. We can
qualitatively attribute this behavior to the thermodynamical-
The ADF package supports scaled-ZORA calculations in which the orbi-
tal energies and the total energy are scaled to improve them towards
fully relativistic results.[21] A similar concept has been suggested for NMR
calculations.[30] The partial ZORA implementation is presently incompat-
ible with scaled ZORA. Data reported in the Results and Discussion for
partial-ZORA shielding contributions are therefore based on unscaled
ZORA calculations. For consistency with our previous work on metallasi-
latrane complexes, all other 29Si shielding data reported herein were ob-
tained with scaled ZORA. For the 29Si shielding, the effects from ZORA
scaling are small (below 2 ppm) and almost completely cancel in the
chemical shifts.
Further details of the computations are as follows: In the 29Si NMR
shielding calculations, a combination of the Vosko–Wilk–Nusair (VWN)
local density functional[31] with revised Perdew–Burke–Ernzerhof[32]
(revPBE) generalized gradient approximation (GGA) terms was utilized
for compatibility with previous work. A hybrid variant of PBE with 25%
Hartree–Fock exchange, PBE0,[33] was used in order to investigate the
sensitivity of the results to approximations in the functional. To better
describe the metal (M) and silicon valence shells in the complexes, a flex-
ible and highly polarized Slater-type orbital (STO) basis set of quadru-
ple-z quality (QZ4P) was used for M and Si. For the ligand atoms in the
complexes, a triple-z doubly polarized STO basis set (TZ2P) designed for
ZORA computations was utilized. NMR chemical shifts were calculated
with the ZORA NMR shielding code of the ADF package.[30,34] All cal-
culations were performed with the X-ray diffraction (XRD) structures, in
selected cases after optimization of the hydrogen positions (HOP).
ꢀ
ꢀ
ly weaker Si Cl bond (compared to Si F), which can re-
spond more easily to trans-disposed lone-pair donation by
ꢀ
bond lengthening, thus strengthening the trans-Si M bond.
ꢀ
Similar features were observed for N!Si X moieties (X=
Cl, F), that is, a shorter N!Si bond for X=Cl.[36] A similar
trend, although to marginal extent, is found for the metal-
bound halide when Cl is replaced by Br by I.
ꢀ
In some crystal structures with several independent M Y
bonds one can see notable differences in M Y bond lengths,
in particular for entries 4, 6, 11, 12, and 16 in Table 1, which
reflect weak M Y bonding and great flexibility of these
ꢀ
ꢀ
bonds resulting therefrom.
Finally, the tilt angle of opposing methimazolyl ring
planes appears to be also very flexible. Although one can
discern the influence of the Si-bound halide (pronounced
tilt angle for X=Cl, resulting from intramolecular H-X-re-
pulsion, Figure 3), for the com-
pounds with the smaller F atom
at silicon crystal packing effects
seem to determine this tilt
angle to a pronounced extent.
This is particularly evident from
In order to investigate the impact of a polarized environment on the cal-
culated NMR parameters, additional computations were performed with
the conductor-like screening model (COSMO) as implemented in
ADF.[35] Whereas the dielectric constants (e) of standard insulating mate-
rials are typically below 10, we chose to investigate the polarization
effect of water (e=78), assuming that this gives an estimation of the
upper limit of the crystal packing or long-range electrostatic interactions
in the solids, as well as the influence of aqueous solvent in the case of
liquid-state NMR. We found that the application of COSMO led to
mean variations of the silicon shielding of about ꢃ5 ppm depending on
halides and the metal centers. Results are reported in Table S7 of the
Supporting Information. The overall effect is not dramatic. The observed
trends for the 29Si NMR parameters discussed herein can therefore be
considered as “intrinsic” to the complexes and not dominantly caused by
crystal environments in the solids.
entry 2 in Table 1, with tilt
angles ranging between 12.7(6)
and 39.0(4)8. The space-filling
Figure 3. Propeller tilt of op-
posing methimazolyl moieties
enforced by repulsion between
the Si-bound halide and methi-
mazolyl-4-H atoms.
ꢀ
models in Figure 4 show the H
X contacts for [F-Si-(m-mt)4-Pt-
Cl] (in I4) and for [Cl-Si-(m-
mt)4-Pd-I] (in Cc).
Overall performance of the calculations: Figure 5 shows the
correlation between ZORA calculations of Si chemical
shifts and experiment for representative members of the [X-
Si-(m-mt)4-M-Y] complex family that are analyzed in detail
&
4
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Metallasilatrane Complexes
FULL PAPER
Table 1. Tilt angle[a] of opposing methimazolyl moieties [8] and bond lengths d of the X-Si-M-Y axes [ꢄ] of
previously published (entries 6, 7, 11, 16) and herein reported crystal structures of compounds of the type [X-
Si-(m-mt)4-M-Y]. For crystal structures with more than one crystallograpically independent X-Si-M-Y axis each
line refers to one molecule.
pounds and selected HOP
structures can be found in the
Supporting Information (Fig-
ure S15). Optimization of the
hydrogen positions did not sys-
tematically improve the agree-
ment with experiment, and
therefore the following discus-
sion is based on calculations
with the XRD structures. In ad-
dition to scalar (SC) ZORA
and SO calculations with the
non-hybrid PBE functional,
which was employed in refer-
ence [13], we report here re-
sults obtained with the PBE0
hybrid functional (25% exact
exchange). The full set of data
is collected in Table 2. The
visual comparison of the differ-
ent calculations in Figure 5
clearly demonstrates the signifi-
cant improvements from includ-
ing SO coupling in the calcula-
tions in particular for the Pt
systems. The resulting chemical
shifts for the [X-Si-(m-mt)4-M-
Y] complexes for the most part
exhibit the correct trends from
the calculations when compar-
ing substitutions of the metal
(as in the [Cl-Si-(m-mt)4-M-Cl]
triad) and upon changes in the
Y halide ligand (example: [Cl-
Si-(m-mt)4-Pt-Y]). The shifts cal-
culated with the PBE0 hybrid
are overall closer to experiment
for the metallasilatrane com-
plexes, whereas the effect from
Tilt angle[a]
d
(X Si)
d
N
dACTHNGUTERNN(UG M Y)
ꢀ
ꢀ
ꢀ
1
2
N
G
19.1(3)
1.682(6)
1.645(6)
1.649(5)
1.640(7)
1.651(7)
1.652(7)
1.650(7)
1.646(7)
1.642(7)
1.648(7)
1.649(6)
1.641(6)
1.614(3)
1.609(5)
1.647(4)
1.631(2)
1.632(3)
2.194(2)
2.211(4)
2.205(4)
2.218(3)
2.204(1)
2.196(2)
2.205(2)
2.203(1)
2.151(1)
2.173(3)
2.172(3)
2.153(2)
2.170(3)
2.167(3)
2.172(1)
2.167(1)
2.166(1)
2.143(1)
2.151(3)
2.164(2)
2.461(3)
2.486(2)
2.480(2)
2.467(3)
2.485(3)
2.504(3)
2.500(3)
2.500(3)
2.499(3)
2.506(3)
2.500(3)
2.503(3)
2.567(1)
2.537(2)
2.528(2)
2.604(1)
2.606(2)
2.469(2)
2.452(3)
2.447(3)
2.457(2)
2.473(1)
2.478(1)
2.485(2)
2.468(1)
2.569(1)
2.546(2)
2.527(2)
2.574(1)
2.543(3)
2.533(2)
2.547(1)
2.563(1)
2.545(1)
2.598(1)
2.588(2)
2.564(2)
2.592(4)
2.619(2)
2.640(2)
2.613(3)
2.621(3)
2.923(1)
2.923(1)
2.925(1)
2.922(1)
2.922(1)
2.924(1)
2.923(1)
2.796(1)
2.651(3)
2.651(2)
2.512(1)
2.532(1)
2.711(1)
2.609(4)
2.598(3)
2.629(2)
2.741(1)
2.762(1)
2.928(1)
2.919(1)
2.791(1)
2.646(2)
2.622(2)
2.900(1)
2.751(1)
2.726(1)
2.970(2)
2.997(1)
2.984(1)
2.612(1)
2.498(2)
2.479(2)
24.5(5), 28.3(5);
38.1(4), 39.0(4);
35.8(4), 36.8(4);
12.7(6), 19.5(6)
27.9(4), 27.4(4);
27.9(3), 29.4(4);
29.5(3), 28.0(4);
29.5(3), 27.9(4);
29.0(3), 27.8(3);
27.6(3), 27.9(4);
27.2(3), 28.1(3)
42.2(3), 41.9(3);
44.5(2);
3
4
ACHTUNGTRENNUNG(m-mt)4PdCl] 8CHCl3
48.7(2)
5
6
A
41.5(2), 41.1(1);
32.8(2), 38.9(2)
50.7(3), 50.4(3);
52.6(3);
[14]
[18]
51.0(2)
7
8
N
44.9(2), 46.5(2)
40.4(2), 47.4(1);
48.8(2), 48.6(1)
40.9(2), 45.1(1)
48.4(1), 49.6(1)
53.0(2), 52.7(2);
52.7(2);
52.8(2)
53.5(6), 53.0(7);
58.0(6);
ACHTUNGTRENNUNG
9
10
11
ACHTUNGTRENNUNG(m-mt)4PtI] 2CHCl3
ACHTUNGTRENNUNG
[14]
ACHTUNGTRENNUNG(m-mt)4PdCl] 8CHCl3
12
(m-mt)4PdBr] 8CHCl3
51.7(7)
13
14
15
16
(m-mt)4PdI] 3CHCl3 (Cc)
50.5(1), 54.0(1)
49.1(1), 45.5(1)
50.7(1), 51.4(1)
51.5(2), 50.0(3);
52.6(3);
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
[18]
ACHTUNGTRENNUNG(m-mt)4NiCl] 8CHCl3
50.8(2)
[a] This tilt angle was determined as the angle between the least-squares-planes of two opposing five-mem-
bered rings.
the hybrid is small for the benchmark set. For the X=F or
Cl, Y=Cl, systems with M=Ni, Pd, Pt, the calculated 29Si
shifts for PBE give the ordering Pt<Pd<Ni, whereas exper-
imentally the ordering is Pt<Ni<Pd. The PBE0 calcula-
tions are improved, giving NiꢂPd for X=F and Ni<Pd for
X=Cl due to an increase of the positive SO contributions in
the Ni complexes. However, already with PBE one finds
that the SO contributions to the 29Si shielding are larger for
Ni than for Pd (the Ni “anomaly” discussed in the introduc-
tion. Because of the need for repeated calculations exclud-
ing one or more sets of atoms from V’ in Equation (2), the
detailed partial-ZORA analysis was carried out with the
PBE functional in order to keep the computational effort
manageable.
Figure 4. Space filling models for [F-Si-(m-mt)4-Pt-Cl] (left, from solvent
free crystal structure in I4) and for [Cl-Si-(m-mt)4-Pd-I] (right, from crys-
tal structure of the chloroform solvate in Cc), view along XSiMY axis to-
wards the Si-bound halide X.
in this work, along with a set of hypervalent Si com-
pounds[37,38] used herein and in our previous work[13] to
benchmark the calculations. An extended plot for all com-
NMR shielding analysis for [Cl-Si-(m-mt)4-M-Cl]: In Table 3,
full spin-orbit relativistic 29Si shielding constants for [Cl-Si-
AHCTUNGTERG(NNUN m-mt)4-MCl] are compared with calculations in which SO
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

J. Autschbach et al.
Figure 6. SO contributions to the 29Si isotropic shielding from selected
atom types for [Cl-Si-(m-mt)4-M-Cl], M=Pt, Pd, and Ni (PBE; for S, the
sum is given). See also Table 3.
atoms in Tables 3 and 4 (column (M,Si,Cl,S)). A certain
degree of non-additivity must be expected, since modifica-
tions of the electron distribution due to SO coupling on one
atom may affect the electron distribution elsewhere in the
molecule because of the self-consistent nature of the calcu-
lations. The high degree of additivity of the SO effects on
the 29Si shielding tensor allows to draw meaningful conclu-
sions from comparing the data for complexes with different
metals.
We consider 29Si shift variations of about 5 ppm between
the different complexes as attributable to slightly differing
structural parameters and crystal packing effects (see end of
the Experimental Section and Tables S6 and S7 in the Sup-
porting Information). Minor shift variations are therefore
excluded from a detailed analysis. One then finds that the
SO effect of Si on its own isotropic shielding (a HAHA
effect) is roughly constant, around 9 ppm, and almost negli-
gible for the tensor span. The sulfur cage around the metal
neither causes significant SO effects on the silicon isotropic
shielding, nor does it influence the tensor span. We remind
of our recent findings in reference [13] about the pro-
nounced electron delocalization in the Si-(m-mt)4-M skele-
ton. The analysis of the partial-ZORA results proves that
despite this delocalization, the Ni anomaly discussed in the
Introduction is not caused by sulfur SO contributions to the
29Si shielding in the three systems. Further, despite the com-
bined chlorine SO contributions being on the order of
5 ppm, they are not large enough to explain the overall
large SO effect on the 29Si shielding in the Ni compound.
The partial-ZORA analysis conclusively identifies the
metal atom itself as a strong source of spin-orbit coupling
induced magnetic shielding of the nearby Si nucleus and
identifies the reason for the Ni anomaly. In the Ni, Pd, Pt
triad, the individual metal contributions are 17.2, 12.5, and
29.4 ppm, for siso, and ꢀ26.7, ꢀ21.0, and ꢀ66.5 ppm for W,
respectively. Thus, the direct metal contributions are in fact
larger in magnitude for Ni than for Pd, unlike what would
be expected simply by considering the nuclear charges of
these atoms. As pointed out in the Introduction, the leading
order of relativistic effects is formally Z2/c2. For Pt versus
Pd the Z2 ratio is (78/46)2 =2.88. The calculated ratio of Pt
versus Pd SO contributions to the 29Si shielding is 29.4/
12.5=2.35. This is smaller than the formal nuclear-charge-
based ratio, but reasonably close. For Pd versus Ni the ratio
of the square of nuclear charges is (46/28)2 =2.70. The calcu-
Figure 5. Correlation between experimental and calculated (scalar (SC)
and spin-orbit (SO) ZORA) 29Si NMR chemical shifts for [Cl-Si-(m-mt)4-
M-Cl] and [Cl-Si-(m-mt)4-Pt-Y] (Y=Cl, Br, I) complexes (see Table 2 for
the full set of data) and for a set of hypervalent Si compounds (see Tacke
et al., references [37,38]) previously used in reference [13] to benchmark
the calculations. Reference: TMS. Compound 3c from reference [36], a
silicon complex with SiO4N2 coordination sphere, was used as a secon-
dary reference in the calculation of 29Si chemical shifts (d
ꢀ183 ppm).
ACHTUNGTRENN(UNG exptl)=
effects have been switched off for selected atom types, using
the partial-ZORA procedure outlined in the Experimental
Section. The two rightmost columns demonstrate that the
spin-orbit NMR code produces exactly the shielding con-
stants from the scalar relativistic branch of the program if
SO effects for all atom types are suppressed simultaneously.
This is an important sanity test of the procedure. A corre-
sponding set of data is collected in Table 4 for the 29Si
shielding tensor span W. As shown in reference [13], the
MAS solid-state NMR spectra are strongly affected by the
tensor span, and only the simulated spectra based on the SO
calculations agree with the experimental spectra obtained
for slow spinning rates. For the isotropic shielding, Figure 6
provides a visual summary of the results along with separate
contributions from the different chlorine ligands.
A number of findings emerge from the data set. First, it is
noted that the SO effects on 29Si from the neighboring metal
M, Cl, and S atoms combined with the “heavy atom–heavy
atom” (HAHA) SO effect of Si on its own shielding are to
a very good degree additive (within one to two ppm). This
is emphasized by comparing the data columns with the sum
of the SO effects from M, Si, Cl, S (noted as ꢀDSO in the
tables) to the SO effect from simultaneously excluding these
&
6
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Metallasilatrane Complexes
FULL PAPER
Table 2. Calculated 29Si NMR shielding [ppm] and chemical shifts [ppm] for the complexes [X-Si-(m-mt)4-M-
Y] using XRD structures. A good agreement to experimental values is achieved when spin orbit effects are
factored.
sponse to an external magnetic
field due to SO coupling. Simi-
lar conclusions have been
drawn recently by Kaupp,
Straka, et al. in a study of non-
monotonous SO trends induced
in proton chemical shifts in
structurally related Co, Rh, and
Ir complexes.[17] The significant-
ly smaller ligand-field splitting
for Co caused large SO induced
chemical shifts for nearby pro-
tons, with qualitatively similar
trends for the Co, Rh, Ir triad
as found here for the M=Ni,
Pd, Pt metallasilatranes. Kaupp,
Straka, et al. were able to cor-
relate the calculated trends
with averaged energy gaps of
orbitals centered on the metal.
Since the partial-ZORA analy-
sis directly and unequivocally
points to the Ni center as the
source of the SO anomaly in
the Ni, Pd, Pt triad of our com-
plexes we forego further analy-
sis of the excitation spectra.
[b]
Contributions[a]
sp sd sso
siso
diso
SC
SO
SC
SO
Exptl
A
G
PBE
PBE0
PBE
E
ꢀ389.6 862.7 46.9
(XRD) ꢀ375.3 873.5 48.8
(XRD)
ꢀ383.3 863.1 36.1
PBE0
(XRD) ꢀ378.8 873.4 45.7
PBE(XRD)
ꢀ371.1 863.0 25.2
PBE0
(XRD) ꢀ368.9 875.0 29.6
PBE(XRD)
ꢀ382.7 860.8 28.5
PBE0
(XRD) ꢀ383.8 875.4 44.0
(m-mt)4-Pt-Cl][c] PBE
(XRD)
ꢀ378.4 867.2 44.1
PBE0
(XRD) ꢀ375.7 875.8 53.9
PBE(XRD)
ꢀ383.6 868.6 37.9
PBE0
(XRD) ꢀ378.1 875.1 51.3
PBE(XRD)
ꢀ384.3 867.0 37.0
PBE0
(XRD) ꢀ379.0 875.4 49.0
PBE(HOP)
ꢀ402.5 867.0 50.9
PBE0
(HOP) ꢀ395.4 874.3 65.4
(m-mt)4-Pd-Cl][c] PBE
(XRD)
ꢀ375.6 866.4 29.0
PBE0
(XRD) ꢀ370.4 876.2 32.4
PBE(XRD)
ꢀ376.8 865.6 32.8
PBE0
(XRD) ꢀ373.6 876.1 36.0
PBE(HOP)
ꢀ378.4 865.9 33.1
PBE0
(HOP) ꢀ373.9 875.7 36.0
PBE(XRD)
ꢀ380.9 864.9 31.9
PBE0
(XRD) ꢀ377.4 875.8 36.1
PBE(HOP)
ꢀ383.9 865.1 32.1
PBE0
(HOP) ꢀ395.4 874.3 65.4
(m-mt)4-Ni-Cl][c] PBE
(XRD)
ꢀ381.0 864.2 31.7
PBE0
(XRD) ꢀ380.8 875.8 47.6
473.5 520.0
498.2 547.0
479.8 516.0
494.6 540.3
491.9 517.1
506.1 535.7
478.1 506.5
491.5 535.5
488.5 532.5
500.1 554.0
483.7 521.6
497.3 548.6
482.6 519.7
496.4 545.5
464.5 515.4
478.9 544.3
490.9 519.9
505.8 538.1
488.9 521.6
502.5 538.5
487.6 520.7
501.8 537.9
484.0 515.9
498.3 534.5
481.2 513.4
478.9 544.3
483.2 514.9
495.0 542.6
ꢀ148.2 ꢀ183.5 ꢀ212.7
ꢀ158.6 ꢀ203.1
E
G
ꢀ154.6 ꢀ179.5 ꢀ197.5
ꢀ155.0 ꢀ189.4
E
G
ꢀ166.6 ꢀ180.5 ꢀ182.3
ꢀ166.5 ꢀ184.8
R
N
ꢀ152.8 ꢀ170.0 ꢀ186.6
ꢀ151.9 ꢀ184.6
G
E
R
ꢀ163.5 ꢀ196.1 ꢀ216.4
ꢀ160.5 ꢀ203.1
T
R
R
ꢀ159.9 ꢀ186.4 ꢀ208.5
ꢀ157.6 ꢀ197.7
R
E
R
ꢀ157.4 ꢀ183.2 ꢀ198.5
ꢀ156.8 ꢀ194.6
N
ꢀ139.2 ꢀ178.9
ꢀ139.3 ꢀ193.4
G
T
G
ꢀ165.6 ꢀ183.4 ꢀ182.4
ꢀ166.2 ꢀ187.2
R
A
R
ꢀ163.6 ꢀ185.2 ꢀ179.8
ꢀ162.9 ꢀ187.6
E
ꢀ162.3 ꢀ184.2
ꢀ162.2 ꢀ187.0
A
A
G
ꢀ158.7 ꢀ179.4 ꢀ175.3
ꢀ158.7 ꢀ183.6
G
ꢀ156.0 ꢀ176.9
ꢀ139.3 ꢀ193.4
A
A
C
ꢀ158.0 ꢀ178.4 ꢀ188.0
ꢀ155.4 ꢀ191.7
[a] Shielding contributions in SO calculations: p = paramagnetic, d = diamagnetic. [b] Chemical shifts with re-
spect to TMS. Compound 3c from reference [36], a silicon complex with SiO4N2 coordination sphere was used
as a secondary reference (experimental chemical shift of ꢀ183 ppm) in the calculations: s=325.3/336.5 and
339.6/350.9 ppm for SR/SO ZORA PBE and ZORA PBE0, respectively; see reference [13] for details.
[c] From reference [13].
SO effects in [X-Si-(m-mt)4-M-
Y] induced by the Y ligand:
Computed and experimental
data for a larger set of [X-Si-(m-
mt)4-M-Y] complexes that are
structurally similar to those
shown in Figure 6, but with dif-
ferent halide ligand combina-
tions X, Y, are collected in
Table 2. A plot of calculated
versus experimental data can
be found in the Supporting In-
formation (Figure S15). As dis-
cussed above, the overall agree-
ment of the SO calculations
with experiment is reasonable
considering that the experimen-
tal data were obtained from
solid-state NMR and the calcu-
lations were performed on iso-
lated [X-Si-(m-mt)4-M-Y] units.
In order to investigate SO ef-
fects on the 29Si shifts from the
Table 3. Calculated isotropic 29Si NMR shielding (siso) for [Cl-Si-(m-mt)4-M-Cl] and spin-orbit contributions
[ppm] using the partial-ZORA decomposition (PBE functional).
siso
SO (SOS)[a]
Atomic potentials not included in V’ and k [Eqs. (2) and (3)]
Si Cl (Cl,S) (Si,Cl,S) (M,Si,Cl,S)
siso
SC[a]
M
M
S
G
N
all
Ni 516.3
Pd 521.2ACHTUNGTRENNUNG
N
512.1
515.1
528.9
502.8
505.5
520.4
485.1
493.0
490.9
484.6
492.2
490.3
484.6
492.2
490.3
Pt
533.8
ꢀ
M,Si,Cl,SDSO
DSO[b]
SOꢀSC[c]
Ni
Pd
Pt
30.6
27.9
42.7
17.2
12.5
29.4
9.1
9.4
8.3
3.5
4.8
4.0
0.7
1.4
0.9
4.2
6.1
4.9
13.5
15.7
13.4
31.1
28.2
42.9
31.7
29.0
43.5
31.7
29.0
43.5
[a] SO=unscaled spin-orbit ZORA. In parentheses: SOS=scaled spin-orbit ZORA, for comparison with
Table 2. SC=unscaled scalar ZORA. [b] DSO=partial-ZORA difference from full SO; ꢀ=summation of in-
dividual DSO contributions from metal (M), Cl, S and Si. If SO effects are additive the sum should be compa-
rable to disabling SO effects for these atoms simultaneously (column “(M, Si, Cl, S)”). [c] SOꢀSC=total SO
effect obtained as difference between full SO and the scalar ZORA calculation.
lated ratio of the SO contributions for Pd versus Ni is 12.5/
17.2=0.73 instead. The only conceivable reason why the SO
terms for Ni are so large compared to Pd is the smaller
ligand-field splitting in the Ni system. This reduces the
energy gaps between occupied and unoccupied orbitals at
the Ni center and locally enhances the magnitude of the re-
halide Y, we have selected the series [Cl-Si-(m-mt)4-Pt-Y]
with Y=Cl, Br, I, along with the Pd counterparts for addi-
tional partial-ZORA analyses. The data are collected in
Table 5. The results for the [F-Si-(m-mt)4-Pt-I] system, also
listed in Table 5, are discussed in the following subsection.
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

J. Autschbach et al.
Table 4. Calculated 29Si NMR shielding tensor span (W) for [Cl-Si-(m-mt)4-M-Cl] and spin-orbit contributions
structures). For the X=Cl plati-
num systems, the SO PBE cal-
culations produce a deshielding
of 13 ppm going from Y=Cl to
I, which is less strong as the
[ppm] using the partial-ZORA decomposition (PBE functional).[a]
W
Atomic potentials not included in V’ and k [Eqs. (2) and (3)]
W
M
SO
M
Si
Cl
S
(Cl,S)
(Si,Cl,S)
(M,Si,Cl,S)
all
scalar
Ni
Pd
Pt
57.2
56.0
31.3
84.0
77.0
97.8
56.2
55.0
27.9
61.8
62.5
36.2
55.8
55.7
31.4
60.4
62.2
36.2
59.6
61.6
33.0
87.7
83.7
102.5
88.6
84.9
103.2 103.2
88.6
84.9
trend
seen
experimentally
(18 ppm). The PBE0 data for
the Pt complexes, while overall
closer to experiment, underesti-
mate the trend more strongly.
For M =Pd with Y=Cl, Br, I,
the calculated decrease of the
29Si shielding from Cl to I is
4 ppm (PBE) and therefore
also underestimates the experi-
mental trend. More important-
ly, the calculations reproduce
the aforementioned finding that
the shielding trends induced by
the Y ligand are modulated by
the nature of the metal and
much smaller in magnitude for
the Pd complexes than for Pt.
Comparison of scalar and SO
relativistic data for the Pt com-
plex series indicates that even
without SO coupling, there is a
Si deshielding of 6 ppm when
going from Y=Cl to I (PBE).
For the Pd series, the corre-
sponding Si shielding decrease
is 7 ppm from Y=Cl to I, that
is, very similar to the Pt series.
For further analysis we turn to
trends in the partial-ZORA Si
shielding constants, collected in
Table 5. In the Pt series with
ꢀ
M,Si,Cl,SDSO
DSO
SOꢀSC
Ni
Pd
Pt
ꢀ28.9
ꢀ26.2
ꢀ68.0
ꢀ26.7
ꢀ21.0
ꢀ66.5
1.0
1.0
3.4
ꢀ4.6
ꢀ6.5
ꢀ4.9
1.4
0.3
ꢀ0.1
ꢀ3.2
ꢀ6.2
ꢀ4.9
ꢀ2.4
ꢀ5.5
ꢀ1.7
ꢀ30.5
ꢀ27.6
ꢀ71.2
ꢀ31.4 ꢀ31.4
ꢀ28.8 ꢀ28.8
ꢀ71.9 ꢀ71.9
[a] See footnotes of Table 3 for further details.
Table 5. Calculated 29Si NMR shielding constants and partial-ZORA analysis for selected [X-Si-(m-mt)4-M-Y]
complexes (PBE functional).[a]
siso
Atomic potentials not included in V’
siso
SO
A
M
Y
Si
scalar
A
T
533.8
522.9
520.9
521.2
522.9
517.2
516.2
517.3
N
504.4
499.2
498.5
508.7
510.9
506.1
499.0
490.7
529.8
514.2
513.5
516.4
514.7
507.7
512.8
498.8
525.5
519.1
512.2
511.8
514.0
508.2
507.1
510.4
490.3
485.2
484.2
492.2
490.2
485.6
484.6
481.1
N
N
R
E
E
U
N
N
G
R
E
N
E
R
ꢀ
M,Y,SiDSO
DSO
SOꢀSC
E
G
41.7
36.1
38.6
26.6
29.2
29.5
29.9
51.8
29.4
23.7
22.4
12.4
12.0
11.1
17.2
26.6
4.0
8.6
7.4
4.8
8.2
9.5
3.5
18.4
8.3
3.8
8.7
9.4
9.0
8.9
9.1
6.8
43.5
37.7
36.7
29.0
32.7
31.5
31.7
48.4
G
ACHTUNGTRENNUNG
N
A
G
G
G
R
G
N
E
N
A
E
[a] See footnotes of Table 3 for further details. Values in parantheses were calculated using scaled-ZORA.
We focus here on [X-Si-(m-mt)4-M-Y] with X=Cl and
Y=Cl, Br, I. Experimentally, for M=Pt the Si chemical
shift becomes more positive along Y=Cl, Br, I by about
18 ppm (8 ppm from Cl to Br, and 10 ppm from Br to I).
The shielding trend obtained by increasing the atomic
number of the Y halide ligand is in the opposite direction as
the effect from replacing Pd with Pt: Si becomes less shield-
ed for fixed M along the Y=Cl, Br, I triad, and more shield-
ed for fixed X and Y when going from M=Pd to M=Pt.
For the Pd systems, the variation of the 29Si shielding for Y
going from Cl to I is only 7 ppm. This shows that these 29Si
shift trends are not only attributable to relativistic effects at
the Y ligand, but must also be modulated by differences in
X=Cl and Y=Cl, Br, I, one finds a decrease of the metal
SO induced shielding from Y=Cl to I, by about 7 ppm (dif-
ferences between the complexes in the DSO column). This
trend dominates the overall SO effects along the series
(keeping in mind that there is an equally important de-
shielding trend present at the scalar relativistic level). The
direct SO contributions from the Y halide ligands to the 29Si
shielding in the Pt series are positive, 4, 9, and 7 ppm for
Y=Cl, Br, and I, respectively.
The atomic SO contributions for the X=Cl series are
therefore rather counter-intuitive. The naive expectation
would be that contributions for Pt are relatively constant
among the set of complexes, and the trend seen experimen-
tally is caused by SO contributions from Y of increasingly
negative magnitude as its nuclear charge increases. Based
on the analysis, however, the SO trends stem from a Y-
halide-modulated quenching of the metal heavy atom effect
as Z(Y) increases, while the DSO(Y) contributions are
always positive. For the Pt complexes, this mechanism is ac-
companied by an increase in the magnitude of the negative
ꢀ
the Si-M Y bonding. This finding needs to be emphasized.
The SO trends from the Y ligands are so different for the Pt
and the Pd sets of complexes, because the metal plays some
role in how the spin density at Y induced by the magnetic
field via SO coupling is transferred to 29Si.
For the comparison of calculated with experimental data
refer to the scalar and SO PBE results of Table 2 (XRD
&
8
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Metallasilatrane Complexes
FULL PAPER
paramagnetic orbital shielding contribution sp (Table 2)
from Y=Cl to I, which reinforces the trend. For the Pd
series, the direct SO effects from Y are similar to the Pt
compounds (i.e., they are to a good degree additive), where-
as the much smaller SO effects originating at the metal are
hardly modulated by the nature of Y. The differences be-
tween the 29Si shifts in the M=Pt and the M=Pd series are
thus predominantly caused by the Y-halide-induced modula-
tion of the metal SO effects on Si (DSO(M)).
and comparable for Br and I. The largest variations for Cl
versus Br versus I are found in DEorb and partially counter-
balanced by DEelstat. The Pauli terms are largest for the
iodine systems. The Si–metal covalency is strongest for Pt,
which contributes to the strong 29Si SO effects for this
metal, and comparable (and weaker) for Pd and Ni. There is
!
a slight decrease in the WBIs for the Si M bonds upon
changing Y from Cl to Br to I for M=Pd. Based on the
NBO[39] analysis of the compounds, the character of the
metal–silicon bonds appears to be very similar for a given
metal but different ligands Y.
Bonding analysis for selected [X-Si-(m-mt)4-M-Y] systems:
For an analysis of the bonding in selected [X-Si-(m-mt)4-M-
Y] complexes, consider trends for the calculated atomic
charges (see Table S5 in the Supporting Information), and
more detailed analyses of the bonding in a selected subset
of the compounds determined by means of a natural bond
orbital (NBO) procedure,[39,40] Tables 6 and S8 (the latter is
in the Supporting Information). All bonding analyses were
carried out based on scalar ZORA calculations.
Additional NBO data for some of the complexes are col-
ꢀ
lected in Table 6, but this time for the M Y bonding local-
ized orbitals (which are predominantly centered on the Y li-
ꢀ
gands). Table 6 also lists calculated WBIs for M Y and
Si···Y interactions. The low bond order between M and Y
classifies the bonding as predominantly ionic, but with some
noticeable covalent character. The latter must be present for
SO effects on the Y halide to manifest directly in the 29Si
shielding tensors (Table 5). In-
terestingly, there are WBIs for
Si···Y that are of comparable
magnitude as (in fact even
!
Table 6. Composition of M Y bonding localized orbitals (natural localized molecular orbitals from NBO
procedure). WBI=Wiberg bond index.
!
WBI
M Y
M
Y orbital composition
M
ꢀ
larger than) those for M Y.
ꢀ
ꢀ
X Si
Si···Y
0.22
Si
Y
The compositions of the rele-
Cl-Si-Pt-Cl
Cl-Si-Pt-I
0.62
0.60
0.69
0.69
0.67
0.69
0.45
0.16
0.19
0.12
0.12
0.14
0.13
0.17
9%
4%
s
84%
sG
ꢀ
vant M Y bonding orbitals in-
s
A
E
N
A
E
U
ACHTUNGTRENNUNG
0.28
0.17
0.19
0.22
0.16
0.28
dicate a significant degree of
delocalization of the partially
s
A
R
s
A
E
U
sN
ACHTUNGTRENNUNG
Cl-Si-Pd-Cl
Cl-Si-Pd-Br
Cl-Si-Pd-I
Cl-Si-Ni-Cl
F-Si-Pt-I
ꢀ
covalent M Y bond over to the
s
A
E
s
A
E
U
s
N
ACHTUNGTRENNUNG
Si center. This delocalization is
increased for the Y=I systems
relative to Y=Cl and Br. This
finding provides a rational for
the larger Pauli terms for the
iodine complexes in the energy
decompositions of Table S8
(Supporting Information). Fur-
ther, the s character of the de-
s
A
E
s
A
E
U
sR
ACHTUNGTRENNUNG
s
A
E
s
A
E
U
sE
ACHTUNGTRENNUNG
s
A
A
s
A
E
U
sN
ACHTUNGTRENNUNG
s
G
R
G
sE
G
G
sN
ACHTUNGTRENNUNG
ꢀ
Trends in calculated atomic charges in the complexes
(Table S5) do not exhibit evident correlations with the 29Si
shielding trends for the Y=Cl to I series. For X=Cl and
M=Pt, Hirshfeld charges of Y become less negative going
from Cl to I, whereas the Mulliken charges become more
negative. Similarly opposing trends are seen for X=F and
M=Pd, Ni. Concerted charge trends for Y are obtained for
X=F and M=Pt. Variations in the charge may occur be-
cause of the different electron donor strengths of the metal
halide ligands Y. In all cases, about 0.5 e of electronic charge
is transferred from Y to the metal. For a subset of the com-
localized M Y bonding orbitals on the Si end can be used
to rationalize some of the trends seen for the partial-ZORA
analysis of Table 5, such as for the M=Pd and Y=Cl, Br, I,
series: For a larger s character on Si one expects a larger
direct SO contribution originating from the Y atom. More
generally, the larger Si···Y WBIs for the iodine compounds
ꢀ
correlate with slightly smaller Si M WBIs for Y=I (versus
Y=Cl for the same metal, Table 6 versus Table S8 in the
Supporting Information). Although subtle variations in the
orbital decompositions complicate the picture, based on the
orbital decompositions this trend to be correlated with re-
duced magnitudes of SO effects on 29Si from the metal that
plexes,
relaxation
energies
DErelax =DEorb +DEPauli +
!
DEelstat +DEprep with respect to promolecular fragments have
been calculated. The data are collected in Table S8 of the
Supporting Information, along with Wiberg bond indices
(WBIs) and atomic hybrid orbital compositions of the
are transmitted by the Si M dative bonds. This is indeed
reflected in the smaller SO contributions on 29Si from the
metal in the iodine complexes, as shown in the analysis of
Table 5.
!
Si M localized bond orbitals as well as a description of the
In addition to data for the Cl-Si-M-Y systems, Table 6
also contains a bonding analysis for F-Si-Pt-I. The partial-
ZORA analysis of the 29Si shielding of this system in Table 6
various energy terms. The data set in Table S8 (Supporting
Information) shows that DE is largest in magnitude for Cl,
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

J. Autschbach et al.
shows, in comparison to its X=Cl counterpart, a somewhat
increased SO contribution from Pt to the 29Si shielding and
a significantly enhanced SO contribution from the iodine Y
ligand. Therefore, not only does the Y ligand modulate the
metal-induced SO shielding of 29Si, but in addition the X-
ligand is able to modulate the SO effects from Y (and the
metal, to some degree). The bonding analysis of Table 6 in-
X=Cl and different metals, the analysis showed that the SO
contributions from Y are approximately transferable be-
tween these complexes. For a given M, there is a minor de-
crease in the Si shielding when going from Y=Cl to the
heavier halides. More important are 29Si shielding trends
caused by the metal SO coupling modulated by the nature
of Y. This effect is much stronger for M=Pt than for M=
Pd or Ni. The pronounced delocalization of the partially co-
!
dicates that the degree of covalency of the M Y bond is
little affected by changing X from Cl to F. However, there is
!
valent M Y bond was shown to play an important role in
a significant drop in the X Si bond order (more ionic X F
these 29Si shielding trends. Moreover, the nature of the
ligand X has been shown to modulate SO contributions to
the 29Si shielding from Y and from the metal, further enrich-
ing the picture.
ꢀ
ꢀ
bonding with X=F), along with a noticeable change in the
!
atomic s/p/d contributions in the delocalized M Y bonding
orbital. Without the NMR analysis of Table 5, it would be
difficult to guess the consequences for the 29Si shielding, but
evidently the electronic delocalization in the X-Si-M-Y moi-
eties complicates the analysis of the spectra by modulations
of the Y-induced SO effects on 29Si by the ligand X.
The calculations reproduce the main trends for 29Si NMR
seen experimentally among the [X-Si-(m-mt)4-M-Y] com-
plexes. There are minor structural differences between the
members of this family of compounds. However, the NMR
spectra can be interpreted based on “intrinsic” electronic
and relativistic effects rather than minor structural differen-
ces or crystal packing effects.
Conclusion
The present study highlights the sometimes unintuitive
role of SO coupling in NMR spectra of complexes contain-
ing heavy atoms, and the useful role that computations in
general,[2,3] and judiciously chosen decompositions of theo-
retical data in particular, can play in uncovering the reasons
for major and minor trends found in the NMR spectra of
structurally and electronically related metal complexes. It is
likely that similar SO effects on NMR spectra, driven by co-
valency and delocalization, are present in other families of
complexes with light and heavy metals from the same peri-
odic group. When discussing NMR trends, one should there-
fore keep in mind that such effects can be more significant
than those from slight structural differences among a set of
related complexes, and end up dominating the observed
trends in NMR spectra.
The class of metallasilatrane complexes [X-Si-(m-mt)4-M-Y]
with M=Ni, Pd, and Pt, previously reported for X=Y=Cl,
has been extended by new members with X=F and Y=Br
and I. Reported herein are new synthetic routes, structural
characterizations by X-ray diffraction, and 29Si solid-state
NMR data. For the X=Y=Cl triad, the calculations report-
ed in reference [13] showed a “Ni anomaly” in the sense
that the SO effects on the 29Si shift were unexpectedly
larger for M=Ni than for M=Pd. Since SO coupling is a
relativistic effect, it should in leading order increase as Z2
where Z is the nuclear charge of the metal. The partial-
ZORA calculations implemented for this study ruled out
that the SO effects in the Ni and Pd complexes are dominat-
ed by heavier main group atoms in the complexes, such as
the S atoms in the mercapto ligand. The larger SO effects in
Ni versus Pd are directly linked to the metal atom. The
most likely mechanism is then the amplification of SO ef-
fects at the metal due to the smaller ligand-field splitting for
Ni as compared to Pd. This conclusion is strongly supported
by recent findings by Straka, Kaupp, et al. who reported
similar “3d metal” anomalies in SO NMR effects on hydro-
gen chemical shifts in Group 9 complexes.[17] In refer-
ence [17], the absence of heavier ligand atoms allowed to
correlate SO effects directly with the metal and the associat-
ed ligand-field splitting, without mandating an equivalent of
the partial-ZORA analysis adopted here. Taking the evi-
dence from reference [17] together with the analysis data re-
ported herein establishes the origin of the non-monotonic
trends of the 29Si shifts in the [X-Si-(m-mt)4-M-Y] family of
complexes for different metals as a combination of chemical
periodic trends (smaller ligand-field splitting for 3d metals)
and relativistic SO effects. It is likely that similar effects can
be found for other classes of metal complexes.
Acknowledgements
The computational study of Si chemical shifts has received financial sup-
port from the National Science Foundation (CHE 0952253). The partial-
ZORA NMR functionality has been developed by J.A. with support by
the US Department of Energy, Office of Basic Energy Sciences, Heavy
Element Chemistry program under Grant No. DE-FG02-09ER16066. We
thank the Center for Computational Research (CCR) at the University
at Buffalo for continuing support of our research projects.
[1] P. Pyykkç, Chem. Rev. 1988, 88, 563–594.
[2] J. Autschbach, S. Zheng, Annu. Rep. NMR Spectrosc. 2009, 67, 1–
95.
[3] A. Bagno, G. Saielli, Theor. Chem. Acc. 2007, 117, 603–619.
[4] M. Kaupp in Relativistic Electronic Structure Theory, Vol. 2 (Ed.: P.
Schwerdtfeger), Elsevier, Amsterdam (The Netherlands), 2004.
[5] Y. Nomura, Y. Takeuchi, N. Nakagawa, Tetrahedron Lett. 1969, 10,
639–642.
[6] M. Bꢀhl, Ann. Rep. NMR Spectrosc. 2008, 64, 77–125.
[7] J. Autschbach in Calculation of NMR and EPR Parameters. Theory
and Applications (Eds.: M. Kaupp, M. Bꢀhl, V. G. Malkin), Wiley-
VCH, Weinheim (Germany), 2004, pp. 227–247.
The availability of systematic data for Y=Cl, Br, I series
of complexes allowed to investigate if SO effects originating
at a heavy Y ligand influences the 29Si chemical shifts. For
&
10
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Metallasilatrane Complexes
FULL PAPER
[8] W. Scherer, V. Herz, A. Brꢀck, C. Hauf, F. Reiner, S. Altmannshof-
er, D. Leusser, D. Stalke, Angew. Chem. Int. Ed. 2011, 50, 2845–
2849.
[9] M. Kaupp, O. L. Malkina, V. G. Malkin, P. Pyykkç, Chem. Eur. J.
1998, 4, 118–126.
[10] J. Autschbach, J. Chem. Phys. 2007, 127, 124106.
[11] M. A. M. Forgeron, R. E. Wasylishen, G. H. Penner, J. Phys. Chem.
A 2004, 108, 4751–4758.
[12] A. M. Kantola, P. Lantto, J. Vaara, J. Jokisaari, Phys. Chem. Chem.
Phys. 2010, 12, 2679–2692.
[13] L. A. Truflandier, E. Brendler, J. Wagler, J. Autschbach, Angew.
Chem. 2011, 123, 269–273; Angew. Chem. Int. Ed. 2011, 50, 255–
259.
[14] J. Wagler, E. Brendler, Angew. Chem. 2010, 122, 634–637; Angew.
Chem. Int. Ed. 2010, 49, 624–627.
[15] C. R. Wade, I. Ke, F. P. Gabbai, Angew. Chem. 2012, 124, 493–496;
Angew. Chem. Int. Ed. 2012, 51, 478–481.
[16] P. Gualco, T. Lin, M. Sircoglou, M. Mercy, S. Ladeira, G. Bouhadir,
L. Perez, A. Amgoune, L. Maron, F. Gabbai, D. Bourissou, Angew.
Chem. 2009, 121, 10076–10079; Angew. Chem. Int. Ed. 2009, 48,
9892–9895.
[17] M. Hyvꢁrinen, J. Vaara, A. Goldammer, B. Kutzky, K. Hegetsch-
weiler, M. Kaupp, M. Straka, J. Am. Chem. Soc. 2009, 131, 11909–
11918.
[18] J. Wagler, E. Brendler, T. Langer, R. Pçttgen, T. Heine, L. Zhech-
kov, Chem. Eur. J. 2010, 16, 13429–13434.
[19] M. Kaupp, C. Remenyi, J. Vaara, O. L. Malkina, V. G. Malkin, J.
Am. Chem. Soc. 2002, 124, 2709–2722.
[20] E. van Lenthe, E. J. Baerends, J. G. Snijders, J. Chem. Phys. 1993, 99,
4597–4610.
[21] E. van Lenthe, Ph.D. thesis, Vrije Universiteit Amsterdam (The
Netherlands), 1996.
[22] J. Autschbach, B. Pritchard, Theor. Chem. Acc. 2011, 129, 453–466.
[23] P. H. T. Philipsen, E. van Lenthe, J. G. Snijders, E. J. Baerends, Phys.
Rev. B 1997, 56, 13556–13562.
[24] C. van Wꢀllen, J. Chem. Phys. 1998, 109, 392–399.
[25] P. Nichols, N. Govind, E. J. Bylaska, W. A. de Jong, J. Chem. Theory
Comput. 2009, 5, 491–499.
[26] F. Aquino, N. Govind, J. Autschbach, J. Chem. Theory Comput.
2010, 6, 2669–2686.
M. Barysz, Y. Ishikawa), Springer, Dordrecht, 2010, Chapter 12,
pp. 521–598.
[28] C. van Wꢀllen, J. Chem. Phys. 2009, 130, 194109–194109–14.
[29] Amsterdam Density Functional: E. J. Baerends, T. Ziegler, J.
Autschbach, D. Bashford, A. Bꢅrces, F. M. Bickelhaupt, C. Bo, P. M.
Boerrigter, L. Cavallo, D. P. Chong, L. Deng, R. M. Dickson, D. E.
Ellis, M. van Faassen, L. Fan, T. H. Fischer, C. Fonseca Guerra, A.
Ghysels, A. Giammona, S. J. A. van Gisbergen, A. W. Gçtz, J. A.
Groeneveld, O. V. Gritsenko, M. Grꢀning, S. Gusarov, F. E. Harris,
P. van den Hoek, C. R. Jacob, H. Jacobsen, L. Jensen, J. W. Kamin-
ski, G. van Kessel, F. Kootstra, A. Kovalenko, M. V. Krykunov, E.
van Lenthe, D. A. McCormack, A. Michalak, M. Mitoraj, J. Neuge-
bauer, V. P. Nicu, L. Noodleman, V. P. Osinga, S. Patchkovskii,
P. H. T. Philipsen, D. Post, C. C. Pye, W. Ravenek, J. I. Rodrꢆguez, P.
Ros, P. R. T. Schipper, G. Schreckenbach, J. S. Seldenthuis, M. Seth,
J. G. Snijders, M. Solꢇ, M. Swart, D. Swerhone, G. te Velde, P. Ver-
nooijs, L. Versluis, L. Visscher, O. Visser, F. Wang, T. A. Wesolow-
ski, E. M. van Wezenbeek, G. Wiesenekker, S. K. Wolff, T. K. Woo,
A. L. Yakovlev, SCM, Theoretical Chemistry, Vrije Universiteit,
Amsterdam (The Netherlands), http://www.scm.com.
[30] S. K. Wolff, T. Ziegler, E. van Lenthe, E. J. Baerends, J. Chem. Phys.
1999, 110, 7689–7698.
[31] S. H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 1980, 58, 1200–1211.
[32] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1998, 80, 891.
[33] C. Adamo, V. Barone, J. Chem. Phys. 1999, 110, 6158–6170.
[34] J. Autschbach, E. Zurek, J. Phys. Chem. A 2003, 107, 4967–4972.
[35] C. C. Pye, T. Ziegler, Theor. Chem. Acc. 1999, 101, 396–408.
[36] A. Kꢁmpfe, E. Kroke, J. Wagler, Eur. J. Inorg. Chem. 2009, 1027–
1035.
[37] S. Metz, C. Burschka, D. Platte, R. Tacke, Angew. Chem. 2007, 119,
7136–7139; Angew. Chem. Int. Ed. 2007, 46, 7006–7009.
[38] S. Metz, C. Burschka, R. Tacke, Organometallics 2009, 28, 2311–
2317.
[39] F. Weinhold in Encyclopedia of Computational Chemistry (P. von R.
Schleyer), Wiley, Chichester (UK), 1998, pp. 1792–1811.
[40] NBO 5: E. D. Glendening, J. K. Badenhoop, A. E. Reed, J. E. Car-
penter, J. A. Bohmann, C. M. Morales, F. Weinhold, Theoretical
Chemistry Institute, University of Wisconsin, Madison. http://
www.chem.wisc.edu/~nbo5.
[27] J. Autschbach in Challenges and Advances in Computational Chem-
istry and Physics, Vol. 10—Relativistic Methods for Chemists (Eds.:
Received: March 5, 2012
Published online: && &&, 0000
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

J. Autschbach et al.
All in a spin! The class of metallasila-
trane complexes [X-Si-(m-mt)4-M-Y]
with M=Ni, Pd, and Pt, previously
reported for X=Y=Cl (shown here),
has been extended by new members
with X=F and Y=Br and I. New syn-
thetic routes, structural characteriza-
tions by X-ray diffraction, and 29Si
solid-state NMR data are reported.
Spin-orbit effects on the 29Si chemical
shifts were investigated with the help
of relativistic density functional calcu-
lations.
Metallasilatrane Complexes
J. Autschbach,* K. Sutter,
L. A. Truflandier, E. Brendler,
J. Wagler ....................... &&&& — &&&&
Atomic Contributions from Spin-Orbit
Coupling to 29Si NMR Chemical Shifts
in Metallasilatrane Complexes
&
12
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!