Alkaline-Earth-Catalysed Cross-Dehydrocoupling of Amines and Hydrosilanes: Reactivity Trends, Scope and Mechanism

Article abstract of DOI:10.1002/chem.201504316

Alkaline-earth (Ae=Ca, Sr, Ba) complexes are shown to catalyse the chemoselective cross-dehydrocoupling (CDC) of amines and hydrosilanes. Key trends were delineated in the benchmark couplings of Ph₃SiH with pyrrolidine or tBuNH₂. Ae{

Full text of DOI:10.1002/chem.201504316

DOI: 10.1002/chem.201504316
Full Paper
&
Synthetic Methods |Hot Paper|
Alkaline-Earth-Catalysed Cross-Dehydrocoupling of Amines and
Hydrosilanes: Reactivity Trends, Scope and Mechanism
ClØment Bellini,[a] Vincent Dorcet,[b] Jean-FranÅois Carpentier,*[a] Sven Tobisch,*[c] and
Yann Sarazin*[a]
Abstract: Alkaline-earth (Ae=Ca, Sr, Ba) complexes are
shown to catalyse the chemoselective cross-dehydrocou-
pling (CDC) of amines and hydrosilanes. Key trends were de-
lineated in the benchmark couplings of Ph3SiH with pyrroli-
dine or tBuNH2. Ae{E(SiMe3)2}2·(THF)x (E=N, CH; x=2–3) are
more efficient than {N^N}Ae{E(SiMe3)2}·(THF)n (E=N, CH; n=
1–2) complexes (where {N^N}À ={ArN(o-C6H4)C(H)=NAr}À
with Ar=2,6-iPr2-C6H3) bearing an iminoanilide ligand, and
alkyl precatalysts are better than amido analogues. Turnover
frequencies (TOFs) increase in the order Ca<Sr<Ba. Ba{CH-
(SiMe3)2}2·(THF)3 displays the best performance (TOF up to
3600 hÀ1). The substrate scope (>30 products) includes dia-
mines and di(hydrosilane)s. Kinetic analysis of the Ba-pro-
moted CDC of pyrrolidine and Ph3SiH shows that 1) the ki-
netic law is rate=k[Ba]1[amine]0[hydrosilane]1, 2) electron-
withdrawing p-substituents on the arylhydrosilane improve
the reaction rate and 3) a maximal kinetic isotopic effect
(kSiH/kSiD =4.7) is seen for Ph3SiX (X=H, D). DFT calculations
identified the prevailing mechanism; instead of an inaccessi-
ble s-bond-breaking metathesis pathway, the CDC appears
to follow a stepwise reaction path with NÀSi bond-forming
nucleophilic attack of the catalytically competent Ba pyrro-
lide onto the incoming silane, followed by rate limiting hy-
drogen-atom transfer to barium. The participation of a Ba
silyl species is prevented energetically. The reactivity trend
Ca<Sr<Ba results from greater accessibility of the metal
centre and decreasing AeÀNamide bond strength upon de-
scending Group 2.
Introduction
formance elastomers, synthetic precursors for ceramic thin
films and monoliths, and in biomedicine.[2]
The catalysis of cross-dehydrocoupling (CDC) reactions be-
tween EÀH and E’ÀH moieties leading to the formation of EÀE’
bonds and release of dihydrogen as a by-product (E and E’ are
electronegative (N, O, P) and more electropositive (B, Si) p-
block elements, respectively) is attracting growing interest in
main-group chemistry.[1] The resulting molecular or macromo-
lecular species are useful as emissive materials, high-per-
The formation of NÀSi bonds to yield silazanes is of interest
to our research program. Silazanes are valuable as ligands in
coordination chemistry,[3] bases,[4] silylating agents[5] or protect-
ing groups for amines, indoles and anilines in organic synthe-
sis.[6] In addition, some oligo- and polysilazanes make for excel-
lent precursors of Si3N4 ceramics.[7] Traditional processes for the
production of silazanes require metalation of amines with
alkali bases or aminolysis of chlorosilanes,[7a,8] but these proce-
dures involve the undesirable formation of stoichiometric
amounts of salts or HCl by-products. The impetus is now shift-
ing towards cost- and atom-efficient methods, such as the ex-
pedient catalytic NÀH/HÀSi CDC of hydrosilanes with HNR1R2
amines (R1, R2 =H, alkyl, aryl). Middle- and late-transition-metal
compounds in which the low-valent metal centre can undergo
oxidative addition, for example, Pd Al2O3-supported species
and graphene-supported nanoparticles,[9] [Ru3(CO)12] and
[a] C. Bellini, Prof. Dr. J.-F. Carpentier, Dr. Y. Sarazin
Organometallics: Materials and Catalysis Department
Institut des Sciences Chimiques de Rennes
UMR 6226 CNRS—UniversitØ de Rennes 1
Campus de Beaulieu, 35042 Rennes (France)
E-mail: jean-francois.carpentier@univ-rennes1.fr
yann.sarazin@univ-rennes1.fr
[b] Dr. V. Dorcet
Centre de DiffractomØtrie des Rayons X
Institut des Sciences Chimiques de Rennes
UMR 6226 CNRS—UniversitØ de Rennes 1, 35042 Rennes (France)
[Rh6(CO)16] clusters,[10] or discrete Ru2+ [6b,d] Rh+[6c,11] or Cr0[12]
,
complexes, first dominated the arena; a copper(I) system was
also described.[13] Notable results were later achieved with
hard, oxophilic metal ions for which s-bond metathesis or Si-
[c] Dr. S. Tobisch
University of St Andrews, School of Chemistry, Purdie Building
North Haugh, St Andrews, KY16 9ST (UK)
E-mail: st40@st-andrews.ac.uk
to-metal hydride transfer are prominent mechanistic features.
[14]
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201504316. It contains structural characteri-
sation of the DFT-optimised key species involved in the various pathways
examined, crystallographic data (including a CIF file) for complex 7-Ba,
NMR spectra and mass spectra of all new silazanes and NMR spectra for all
new complexes.
Some success was met in the 1990s with Cp2TiMe2,
and
more recently Al3+ [15] Y3+ [16] Yb2+ [17] U4+ [18] Zn2+[19] and espe-
,
,
,
,
cially Ae2+ (Ae=alkaline-earth metal)[17c,20] catalysts have
shown great promise. Additionally, B(C6F5)3 is useful for the
coupling of hydrosilanes with aromatic amines.[6e]
Chem. Eur. J. 2016, 22, 4564 – 4583
4564
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Amongst these, alkaline-earth catalysts are particularly effi-
cient, achieving high selectivity in the coupling of hydrosilanes
and amines, hydrazines or ammonia. Harder’s Ca–azametallacy-
clopropane Ca(HMPA)3(h2-Ph2CNPh) couples aliphatic amines
or aniline to Ph3SiH with a metal loading of 3–10 mol% at
208C in 0.5–24 h.[17c] Sadow’s {ToM}Mg{N(SiMe3)2} (ToM =tris(oxa-
zolinyl)borato) couples primary amines, ammonia or hydrazine
with phenyl-substituted hydrosilanes.[20a] The kinetic rate law is
first order in [silane] and [precatalyst] and zero order in
[amine]; mechanistic investigations suggested that the reac-
tion proceeds by a turnover-limiting nucleophilic attack of the
Mg amide onto the silane followed by rapid hydride transfer
from hypervalent Si to the Mg centre. Hill reported that the
homoleptic amido precatalysts Ae{N(SiMe3)2}2 (Ae=Mg, Ca, Sr)
promote the CDC of hydrosilanes and aliphatic amines or
bulky anilines at 20–608C (metal loading 5–10 mol%),[20b] but
no data was provided for their Ba congener. The Ca precatalyst
performed substantially better than its Mg and Sr analogues,
as exemplified in the coupling of Et2NH with Ph2SiH2. However,
several subtleties associated with the bond polarity and charge
density of the Ae ion became apparent upon mechanistic in-
vestigation. The kinetic rate law was first order in [metal] and
[amine] and zero order in [silane] for the Mg and Ca precata-
lysts, but it was second order in [metal] and first order in
[amine] and [silane] with the Sr one. In a recent communica-
tion, we reported that highly active and selective Ba com-
plexes surpass their Ca and Sr analogues for chemoselective
NÀH/HÀSi CDC reactions;[20c] the main mechanistic features
were delineated by a combination of experiments and DFT cal-
culations. The full scope of this Ae-catalysed reaction is pre-
sented here, together with a thorough DFT analysis of the rival
mechanistic pathways. A complete set of data that highlights
the differences between Ca, Sr and Ba (ionic radius rionic =1.00,
1.18 and 1.35 Š, respectively) is provided.
Results and Discussion
Precatalyst synthesis
A broad range of Ae precatalysts, some devoid of bulky ancil-
lary ligands and some stabilised instead by the bidentate imi-
noanilide {N^N}À (ꢀ[ArN(o-C6H4)C(H)=NAr]À, Ar=2,6-iPr2-C6H3),
were initially screened in the CDC of amines and hydrosilanes
(1–8, Figure 1). Most were prepared according to known litera-
ture protocols. The silanido compounds Ae(SiPh3)2·(THF)3 (5-Sr,
5-Ba) and the heteroleptic complex {N^N}Ba{N(SiMe2H)2}·(THF)2
(7-Ba) were synthesised here for the first time. The pyrrolido
complexes {N^N}Sr{N(CH2)4}·(HN(CH2)4) (9-Sr), in which some
degree of stabilisation is ensured by a coordinated molecule of
pyrrolidine (HN(CH2)4),[21] and {N^N}Ba{N(SiMe3)2}·(HN(CH2)4)2
(10-Ba) were also prepared for the mechanistic studies.
The yellow complex 5-Sr (d29Si =À12.14 ppm) and the
orange compound 5-Ba (d29Si =À12.11 ppm) were obtained in
32 and 86% yield (unoptimised), respectively, upon salt meta-
thesis between AeI2 and freshly prepared Ph3SiK[22] (2 equiv) in
THF.[23] The orange complex 7-Ba supported by the bulky bi-
dentate iminoanilide {N^N}À was synthesised by a salt meta-
[24]
thesis reaction of {N^N}H, anhydrous BaI2 and KN(SiMe2H)2
(2 equiv) following a known one-pot protocol.[25] The orange Sr
pyrrolide complex 9-Sr was prepared in pentane upon aminol-
ysis of the alkyl complex 8-Sr with excess pyrrolidine (3 equiv).
On the other hand, treatment of 6-Ba with pyrrolidine
(2 equiv) merely afforded the bis-adduct 10-Ba by displace-
ment of THF, with no evidence for the release of HN(SiMe3)2
and formation of a Ba pyrrolide.[26] Diagnostic resonances in its
1H NMR spectrum include broad overlapping multiplets at d=
2.49 and 1.27 ppm (8H each, belonging to the coordinated
pyrrolidine molecule and pyrrolide) and a broad singlet at d=
0.94 ppm for the two HNpyrrolidine acidic hydrogen atoms.[21] The
presence of N(SiMe2H)À moieties in 7-Ba compared to the
À
more traditional N(SiMe3)2 amido group in 6-Ba is known to
impart greater stability and kinetic inertness by formation of
intramolecular Ba···HÀSi b-agostic interactions, although this
may at times come at the expense of catalytic efficacy.[24,27] The
agostic contact in 7-Ba was detected spectroscopically: the
Figure 1. Ae precatalysts screened for NÀH/HÀSi CDC reactions.
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4565
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
coupling constant 1J(H,Si)=160 Hz in the 1H NMR spectrum
and the stretching frequency n˜SiÀH =2004 cmÀ1 in the FTIR
spectrum of the complex both testified to the existence of
agostic bonding of mild intensity. Agostic distortions are
better visualised in the molecular solid-state structure of the
complex determined by single-crystal X-ray diffraction
(Figure 2). A large discrepancy is noted between the Ba2-N53-
Si3 (129.7(2)8) and Ba2-N53-Si4 (105.31(19)8) angles. Moreover,
the four atoms Ba2, N53, Si4 and H4 (with the Si4ÀH4 bond
pointing towards the metal) are nearly perfectly coplanar,
whereas Ba2, N53, Si3 and H3 are not. Both observations imply
the presence of a single, mild Ba2···H4ÀSi4 agostic interac-
tion,[24–25,27–28] resulting in a formally six-coordinate Ba2 atom.
By comparison, the two Ba-N-Si angles (114.85(15)8 and
119.78(14)8) in 6-Ba show less difference than the angles in 7-
Ba, and the BaÀNamide bond length in 6-Ba (2.623(3) Š) is elon-
gated relative to that in 7-Ba (2.601(4) Š).[29]
1) independently of the identity of the ligands, the catalytic
activity increases systematically in the order Ca<Sr<Ba
(Table 1, entries 9, 10 and 11; entries 15, 16 and 18; en-
tries 20, 21 and 22; entries 24, 25 and 26);[30] the Ba preca-
talysts afforded unmatched overall turnover frequency
(TOF) values calculated at 75–99% conversion up to
3600 hÀ1 (Table 1, entry 13).
2) the simple bis-amido complexes 1 are more effective than
their direct heteroleptic counterparts 6 (Table 1, entry 9
versus 15, entry 10 versus 16, entry 11 versus 17); the same
is true for the bis-alkyl complexes 4 and their heteroleptic
derivatives 8 (Table 1, entry 14 versus 19, entry 22 versus
26).
3) the presence of coordinated THF molecules in bis-amido
precatalysts bears no influence on the overall catalytic effi-
ciency (Table 1, entry 3 versus 4, entry 10 versus 12).
4) the alkyl complexes are significantly more effective than
their amido analogues (Table 1, entry 1 versus 2, entry 11
versus 13, entry 23 versus 26).
5) the complexes incorporating N(SiMe2H)À afford very little, if
any, catalytic activity (Table 1, entries 5 and 7).
6) the bis-silanido complex 5-Sr (Table 1, entry 6) is inactive,
whereas the pyrrolido compound 9-Sr is a competent cata-
lyst (Table 1, entry 8);[31] these observations have heavy im-
plications regarding the nature of the reaction mechanism
(see below).
7) even in the presence of excess Ph3SiH (2 equiv versus the
amine), the very mildly reactive tBuNH2 solely affords the
mono-coupled product tBuNHSiPh3 with 100% chemoselec-
tivity and no detectable formation of the decoupled prod-
uct tBuN(SiPh3)2 (Table 1, entries 20–26), which highlights
the difficulty in making RN(SiPh3)2 disilazanes (R=alkyl etc.)
from primary amines by using this procedure.
From the foregoing, the bis-alkyl barium complex 4-Ba is
evidently the precatalyst of choice for this catalytic reaction. It
is reasonably easy to prepare on a multi-gram scale compared
with its more intricate heteroleptic derivative 8-Ba. The more-
robust THF-coordinated amido analogue 1-Ba, which is admit-
tedly easier to synthesise, also exhibited impressive per-
formance and therefore constitutes an excellent alternative.
These two precatalysts were used to explore the substrate
scope of the reaction.
Figure 2. ORTEP representation (ellipsoids at the 50% probability level) of
the molecular solid-state structure of {N^N}Ba{N(SiMe2H)2}·(THF)2 (7-Ba). Only
one of the two independent but geometrically equivalent molecules found
in the asymmetric unit, with the main component of the disordered THF
molecule corresponding to O96, is depicted. Hydrogen atoms on the carbon
atoms are omitted and the carbon atoms of the coordinated THF molecules
are represented in a shaded tone for clarity. Selected bond lengths [Š] and
angles [8]: Ba2ÀN53 2.601(4), Ba2ÀN51 2.676(4), Ba2ÀN52 2.764(4), Ba2ÀO96
2.774(3), Ba2ÀO91 2.811(3); Si4-N53-Ba2 105.31(19), Si3-N53-Ba2 129.7(2).
It is disappointing, but not entirely unexpected, that com-
plexes 3-Ba and 7-Ba bearing a tetramethyldisilazide moiety
display very poor performance. Similar observations were
made for the related Ca complex {LO}CaN(SiMe2H)2·(THF) com-
plex ({LO}À =aminoether-phenolato ligand) used for the cataly-
sis of intramolecular hydroamination reactions.[27] It was dem-
onstrated that this complex reacted with 2,2-dimethylpent-4-
en-1-amine (=H2NR) upon release of H2 by self-catalysed dehy-
drocoupling to yield the catalytically inactive species {LO}CaN-
(SiMe2H)(SiMe2NHR)x . One can assume that a similar deactiva-
tion process occurs in the present situation with 3-Ba and 7-
Ba, especially because of the nature of the catalytic reaction
under investigation here.
Precatalyst screening
Complexes 1–9 were first assessed as precatalysts in NÀH/HÀSi
CDC reactions in two benchmark reactions: the coupling of
Ph3SiH with the highly reactive pyrrolidine or with the more re-
luctant tert-butylamine (Scheme 1).From the results of this pre-
liminary screening (Table 1), the following conclusions
emerged:
Scheme 1. Benchmark CDC reactions for initial Ae precatalyst screening.
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4566
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
and experimental conditions (i.e. temperature, reac-
tion time and substrate loading).
Table 1. Ae-catalysed CDC of amines and Ph3SiH.[a]
R1R2NÀH
Product
Precat. N/Si/Ae
T
[K]
t
Conv. TOF
Arylhydrosilanes (Ph3SiH, Ph2SiH2 or PhSiH3) react
very quickly with pyrrolidine (Scheme 2). The dihy-
drosilane Ph2SiH2 selectively yields Ph2SiH(N(CH2)4) (I)
or Ph2Si(N(CH2)4)2 (II) depending on the reaction con-
ditions; the corresponding TON (up to 800 equiv of
amine per metal) are otherwise unmatched. With 1-
Ba as the precatalyst, reaction with PhSiH3 leads to
facile formation of PhSiH(N(CH2)4)2 (X) or even to the
tri-coupled PhSi(N(CH2)4)3 (XI), but the high reactivity
of the PhSiH3/pyrrolidine pair of substrates precludes
clean formation of the mono-coupled product
PhSiH2(N(CH2)4) (IX). The presence of a para electron-
withdrawing group increases the reaction rate, as
seen for (p-CF3-C6H4)PhSi(N(CH2)4)2 (III): complete con-
version is achieved within 1 h (compare with the for-
mation of I and II). This observation will be discussed
in more detail in the kinetic and mechanistic analysis
of this catalytic process (see below). Benzylamine is
as good a substrate as pyrrolidine, with high chemo-
selectivity, TON and reaction rates being achieved at
298 K to afford (C6H5CH2NH)SiPh3 (VI) and
(C6H5CH2NH)2SiPh2 (VII). Such selectivity is remark-
able, and formation of the N-di-coupled product
(C6H5CH2)N(SiPh3)2 was never observed. This suggests
limited ability of (SiR3)xNH3Àx silazanes (x=1, 2) to
engage in CDC with hydrosilanes, a reckoning al-
ready made during the preliminary screening (see
above) and further reinforced by the complete ab-
sence of V in the attempted catalysed coupling of
Ph3SiH and HN(SiMe3)2.
[%][b]
[hÀ1
]
[c]
[min]
1
1-Ca
20:20:1
298
2”60 29
3
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
4-Ca
1-Ba
2-Ba
3-Ba
5-Sr
7-Ba
9-Sr
1-Ca
1-Sr
1-Ba
2-Sr
4-Ba
4-Ba
6-Ca
6-Sr
6-Ba
6-Ba
8-Ba
20:20:1
20:20:1
20:20:1
20:20:1
20:20:1
20:20:1
20:20:1
298
298
298
298
298
298
298
2”60 74
2”60 91
2”60 87
7
9
9
<1
0
<1
9
2”60
2”60
3
0
2”60 11
2”60 92
400:400:1 298 15”60 72
400:400:1 298 15”60 99
400:400:1 298 15
400:400:1 298 15”60 99
400:400:1 298
400:400:1 298 15
400:400:1 298 15”60 42
400:400:1 298 15”60 74
400:400:1 298 15
400:400:1 298 15”60 99
19
26
99
1584
26
3600
1584
11
20
512
26
5
75
99
32
400:400:1 298 15
63
0
1008
20
8-Ca
20:20:1
333
2”60
2”60
0
21
22
23
24
25
26
8-Sr
8-Ba
1-Ba
4-Ca
4-Sr
4-Ba
20:20:1
20:20:1
20:20:1
20:20:1
20:20:1
20:20:1
333
333
333
333
333
333
5
<1
3
5
1
8
2”60 25
2”60 50
2”60
9
2”60 82
2”60 95
10
[a] Reactions in C6D6 (0.5 mL), with [Ae]0 =10.0 mm and [amine]=0.2m for [amine]0/
[Ph3SiH]0/[precat.]0 =20:20:1 or 20:40:1, and [amine]=4.0m for [amine]0/[Ph3SiH]0/
[precat.]0 =400:400:1; reactions times are not optimised. In the case of tBuNH2 (en-
tries 20–26), the formation of tBuN(SiPh3)2 was never detected. [b] Conversion of the
starting amine, as determined by 1H NMR spectroscopy. [c] Overall TOF calculated at
maximal conversion.
Aliphatic silanes are moderately reactive. The syn-
thesis of Et3Si(N(CH2)4) (IV) from triethylhydrosilane
and pyrrolidine requires forcing conditions, even with
Reaction scope
4-Ba as the precatalyst. Bulky amines also react with great diffi-
culty. With 2-adamantylamine (adam-NH2), no conversion to
(adam-NH)SiPh3 (XII) was detected spectroscopically, and CDC
of tetramethylpiperidine with Ph3SiH only reached 16% con-
version to VIII after 1 h at 333 K. Aromatic amines exhibit
lower reactivity than aliphatic amines, presumably on account
of their reduced nucleophilicity. Hence, 2,6-diisopropylaniline
can be paired with Ph2SiH2 to give (2,6-iPr2ÀC6H3ÀNH)SiHPh2
(XIV) with good selectivity after a prolonged reaction time at
333 K, but the corresponding coupling to generate XV from
the bulkier Ph3SiH failed entirely. The reactions with less-en-
cumbered anilines proceeded more smoothly. Clean formation
of (C6H5N)SiHPh2 (XIII) was achieved at 298 K from a 40:20 mix-
ture of aniline and Ph2SiH2, whereas a 78:22 mixture of XIII
and its di-coupled derivative (C6H5N)2SiPh2 was obtained at
333 K; clean synthesis of this latter product could not be ach-
ieved under the chosen experimental conditions.
The substrate scope of the catalytic reaction was evaluated
with precatalysts 1-Ba and 4-Ba (the heteroleptic and less-
active precatalyst 6-Ba was only used twice in specific cases
for very reactive substrates) and a variety of silanes and
amines. The results for this screening are summarised in
Scheme 2 for reactions involving monoamines and mono(hy-
drosilane)s, and the coupling of di(hydrosilane)s with monoa-
mines and that of diamines with mono(hydrosilane)s are col-
lected in Schemes 3 and 4, respectively.[32] The reactions were
performed in C6D6 (identical outcomes were obtained for reac-
tions carried out in C6D5Cl and in 5:1 mixtures of C6D6/1,2-
C6H4F2, which have higher dielectric constants than C6D6),[33]
1
and the substrate conversion was monitored by H NMR spec-
troscopy. All new products were characterised by standard an-
alytical methods (1H, 13C and 29Si NMR spectroscopy and mass
spectrometry). Generally speaking, the chemoselectivity and
measured conversions were excellent and, in the case of po-
tentially difunctional substrates, the course of the reaction and
specific formation of a given product could be controlled by
judicious choice of the precatalyst (the more-active precatalyst
4-Ba was used for more reluctant reactions and/or substrates)
The reactivity of di(hydrosilane)s follows that detailed for
mono(hydrosilane)s, and gives access to original di(silazane)s
(Scheme 3). The reagent 1,2-bis(dimethylsilyl)ethane is rather
unreactive. It undergoes CDC with pyrrolidine only after 12–
15 h, and preferably at 333 K, to give mixtures of ((CH2)4N)Me2-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4567
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 2. Ba-catalysed CDC of mono(amine)s and mono(hydrosilane)s. Reactions were performed in C6D6 with precatalysts 1-Ba, 4-Ba or 6-Ba. For each prod-
uct (I–XV) the [amine]0/[silane]0/[precat.]0 ratio, reaction temperature and time, precatalyst, substrate conversion (related to the conversion of the initial hydro-
silane) and chemoselectivity between the mono-, di- and tri-coupled silazane products are indicated. Newly formed bonds are shown in bold.
SiCH2CH2SiHMe2 (XVIII) and preponderantly its di-coupled ana-
logue ((CH2)4N)Me2SiCH2CH2SiMe2(N(CH2)4) (XIX) with limited
selectivity, even with precatalyst 4-Ba. Greater reactivity is
seen for 1,4-bis(diphenylsilyl)benzene, which couples twice
with pyrrolidine or benzylamine under mild conditions to
produce ((CH2)4N) Ph2SiÀC6H4ÀSiPh2(N(CH2)4) (XVI) and
(C6H5CH2NH)SiPh2ÀC6H4ÀSiPh2(NHCH2C6H5) (XVII), respectively,
with good chemoselectivity. The reactivity of 1,4-bis(dimethylsi-
lyl)benzene towards pyrrolidine is also rather satisfactory: 1-Ba
reaction can be oriented towards the formation of
Ph3SiNHCH2CH2NH2 (XXII) or Ph3SiNHCH2CH2NHSiPh3 (XXIII) de-
pending on the initial Ph3SiH concentration; in no case
was further coupling to (Ph3Si)2NCH2CH2NHSiPh3 or even
(Ph3Si)2NCH2CH2N(SiPh3)2 observed, another indication that
ÀNHSiR3 silazanes engage in CDC with hydrosilanes at best
with great difficulty. From an N,N’-dimethylethylenediamine/
Ph3SiH feed ratio of 20:40, precatalysts 1-Ba (at 298 K) and 4-
Ba (at 333 K) gave the mono-coupled Ph3Si(Me)NCH2CH2NHMe
and 4-Ba both afford clean access to ((CH2)4N)Me2SiÀC6H4À (XXV) and the di-coupled Ph3Si(Me)NCH2CH2N(Me)SiPh3 (XXVI),
SiMe2(N(CH2)4) (XXI), whereas 6-Ba allows preponderant forma- respectively. Similarly, on adjusting the initial hydrosilane con-
tion of the mono-coupled derivative, ((CH2)4N)SiMe2ÀC6H4À tent of the reaction mixture, CDC of Ph3SiH with piperazine at
SiMe2H (XX).
298 K catalysed by 1-Ba readily affords the mono- and di-cou-
pled products, XXVIII and XXIX, with high chemoselectivity. By
pairing N,N’-dibenzylethylenediamine with Ph3SiH, Ph3Si(Bn)-
The utilisation of diamines leads to the formation of valuable
di(silazane)s (Scheme 4). Starting from 1,2-ethylenediamine, the
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4568
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 3. Ba-catalysed CDC of amines and di(hydrosilane)s. Reactions were carried out in C6D6, using precatalysts 1-Ba, 4-Ba or 6-Ba. For each product (XVI–
XXI) the [amine]0/[di(hydrosilane)]0/[precat.]0 ratio, reaction temperature and time, precatalyst, substrate conversion (related to the conversion of the initial di-
(hydrosilane)) and chemoselectivity between mono-, di- and tri-coupled silazane products are indicated. Newly-formed bonds are shown in bold.
NCH2CH2NHBn (XXXI) and Ph3Si(Bn)NCH2CH2N(Bn)SiPh3 (XXXII)
can be accessed, although the selectivity towards XXXI was
only moderate (83%). Diamines such as 1,4-phenylenedimeth-
anamine and its N,N’-dimethyl-substituted congeners display
mixed reactivity towards with Ph3SiH. The di-coupled com-
molecular weight polymers (Mn =3–10”103 gmolÀ1) from N,N’-
dibenzylethylenediamine and Ph2SiH2.[32]
Kinetic analysis
pounds
Ph3SiHNCH2ÀC6H4ÀCH2NHSiPh3
(XXIV)
and
Ph3Si(Me)NCH2ÀC6H4ÀCH2N(Me)SiPh3 (XXVII) could be obtained
Kinetic investigations were carried out to improve our under-
smoothly with excellent selectivity. However, BnHNCH2ÀC6H4À standing of these Ae-mediated NÀH/HÀSi CDC reactions. The
CH2NHBn is far less reactive: the mono-coupled Bn(Ph3Si)-
NCH2ÀC6H4ÀCH2NHBn (XXX) is formed exclusively but in poor
yield (16%), even at 333 K with precatalyst 4-Ba.[34]
main lines of the analysis of the coupling of pyrrolidine with
Ph3SiH, our benchmark substrates, catalysed by the bulky
amido precatalyst 6-Ba (6-Ba is less active than 1-Ba and 4-Ba
and lends itself well to monitoring the substrate conversion by
1H NMR spectroscopy) were reported in our earlier communica-
tion.[20c] They are recalled here, together with comparative data
for the Ca, Sr and Ba precatalysts.
Overall, the Ba precatalysts 1-Ba and 4-Ba display a remark-
able combination of productivity, activity and chemoselectivity
in NÀH/HÀSi CDC reactions, and they give access to a wide
range of original mono- and disilazanes. Notably, several of
these products still possess NÀH and/or SiÀH reactive function-
al groups, which renders them useful synthetic intermediates
that could be used for further functionalisation reactions. The
performance displayed by 1-Ba and 4-Ba in the CDC of ditopic
diamines and di(hydrosilane)s suggests that they might be
good precatalysts for dehydrogenative polymerisation to
afford polycarbosilazanes; for instance, we have prepared low
From reactions carried out in C6D6 at 298 K, aimed at investi-
gating the influence of the substrate concentrations ([pyrroli-
dine]0 or [Ph3SiH]0 was varied over a 21-fold concentration
range; Figures 3 and 4) and precatalyst concentration ([6-Ba]0
was studied over a 15-fold concentration range; Figure 5), the
second-order rate law corresponding to Equation (1) was es-
tablished, with k=4.802(13)”10À3 mÀ1 sÀ1.
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4569
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 4. Ba-catalysed CDC of diamines and Ph3SiH. Reactions were carried out in C6D6 with precatalysts 1-Ba or 4-Ba. For each product (XXII–XXXII) the [di-
amine]0/[Ph3SiH]0/[precat.]0 ratio, reaction temperature and time, precatalyst, substrate conversion (related to the conversion of the initial diamine) and che-
moselectivity between the mono-, di- and tri-coupled silazane products are indicated. Newly-formed bonds are shown in bold.
1
0
1
ð1Þ
R ¼ Àd½pyrrolidineŠ=dt ¼ k ½6-BaŠ ½pyrrolidineŠ ½Ph3SiHŠ
An identical kinetic rate law was reported by Sadow and co-
workers for NÀH/HÀSi dehydrocoupling reactions catalysed by
{ToM}Mg{N(SiMe3)2}.[20a] On the other hand, Hill and co-workers
found different rate laws for the bis-amido precatalysts Ae{N-
(SiMe3)2}2, which were found to be of the form R=
k[Ae]1[pyrrolidine]1[triphenylsilane]0 for Ae=Mg and Ca, and
R=k[Ae]2[pyrrolidine]1[triphenylsilane]1 for Ae=Sr.[20b]
For experiments run under otherwise identical conditions,
a strong kinetic isotopic effect (KIE) of kPh3SiH/kPh3SiD =4.7(1) was
measured for the coupling of Ph3SiX (X=H, duplicate experi-
ments; D, triplicate experiments from two batches of deuterat-
ed hydrosilane; the mean values of kPh3SiH and kPh3SiD were used
to obtain the KIE) with HN(CH2)4 (Figure 6), that is, within the
Figure 3. Plot of [HN(CH2)4]=kobst as a function of time for the CDC of pyrro-
lidine and Ph3SiH catalysed by 6-Ba, showing partial zero-order kinetics in
[pyrrolidine]. Conditions: HN(CH2)4 (8.0 mL, 0.1 mmol), Ph3SiH (300 mg,
1.15 mmol), [6-Ba]0 =0.0179m ( , kobs =2.53”10À4 sÀ1) or 0.0097m ( ,
obs =9.83”10À5 sÀ1), T=298 K, C6D6.
&
^
k
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4570
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the SiÀH bond intervenes as a key event of the turnover-limit-
ing step.
The activation parameters DH° =15.6(23) kcalmolÀ1 and
DS° =À13.3(7.5) calmolÀ1 KÀ1 (DG° =19.6(1) kcalmolÀ1 at
298 K) were determined by an Eyring analysis, with six data
points in the temperature range 293–318 K (Figure 7).[36] The
negative value of DS°, diagnostic of an associative mechanism,
is small relative to those given for other Mg–Sr CDC precata-
lysts;[20a–b] this is possibly a consequence of the large radius of
the Ba2+ ion (rionic =1.38 Š), which may induce a less-constrain-
ed arrangement in the transition state.
Figure 4. Plot of ln([Ph3SiH]0/[Ph3SiH]t)=kobst as a function of time for the
CDC of pyrrolidine and Ph3SiH catalysed by 6-Ba, indicating partial first-
order kinetics in [Ph3SiH]. T=298 K, C6D6, HN(CH2)4 (160 mL, 2.0 mmol),
Ph3SiH (10.0 mg, 0.04 mmol), [6-Ba]=4.0 mm (”, kobs =3.33”10À6 sÀ1),
11.0 mm ( , kobs =4.67”10À5 sÀ1), 26.0 mm ( , kobs =8.67”10À5 sÀ1), 30.0 mm
~
^
(
, kobs =1.27”10À4 sÀ1) and 60.0 mm ( , kobs =2.77”10À4 sÀ1).
&
*
Figure 7. Eyring plot for the CDC of pyrrolidine and Ph3SiH catalysed by 6-
Ba; k was determined from the plot of ln([Ph3SiH]0/[Ph3SiH]t)=kobst with
k
obs =k[6-Ba]1[HN(CH2)4]0. HN(CH2)4 (8.0 mL, 0.10 mmol), Ph3SiH (26 mg,
0.10 mmol) and 6-Ba (3.0 mg, 0.0034 mmol), C6D6, T=293–318 K.
A Hammett analysis was performed for the coupling of pyr-
rolidine with Ph2(p-X-C6H4)SiH (X=Me, OMe, F, and CF3). It
showed that in comparison with Ph3SiH (X=H), the reaction
rates increased very substantially with electron-withdrawing
groups (X=F, CF3) and decreased with electron-donating sub-
stituents (X=Me, OMe), in the order X=OMe<Me<H<F<
CF3 (Table 2). The Hammett plot [Eq. (2)] is a straight line with
a positive slope of 1=2.0(2) (Figure 8),[37] which clearly indi-
cates that the activation barrier of the catalytic event is low-
ered by electron-withdrawing p-substituents, presumably by
stabilisation of a developing negative charge on the silicon
atom.
Figure 5. Linear plot of kobs =k[6-Ba] as a function of [Ba]0 for the CDC of
pyrrolidine and Ph3SiH catalysed by 6-Ba, showing partial first-order kinetics
in [6-Ba] with k=4.80”10À3 mÀ1 À1. Conditions: HN(CH2)4 (160 mL,
s
2.0 mmol), Ph3SiH (10 mg, 0.04 mmol), [6-Ba]=4.0, 11.0, 26.0, 30.0 and
60.0 mm, T=298 K, C6D6.
Figure 6. Plot of ln([XSiPh3]0/[XSiPh3]t)=kobst as a function of time for X=H
or D for the CDC of pyrrolidine and Ph3SiX catalysed by 6-Ba. Conditions:
HN(CH2)4 (12.0 mL, 0.15 mmol), Ph3SiH or Ph3SiD (39 mg, 0.15 mmol) and 6-
^
Ba (4.4 mg, 5.0 mmol), T=298 K, C6D6. Ph3SiH: kobs =3.25”10À4
(
) and
*
( ) and
2.50”10À4 sÀ1
(
); Ph3SiD: kobs =1.00”10À4 ( ) 4.67”10
À5
~
&
3.67”10À5 sÀ1 (”).
Figure 8. Hammett plot ln(kX/kH)=sp(X)1 showing the reaction-rate accelera-
tion with electron-withdrawing groups on the silane for the coupling of
HN(CH2)4 and Ph2(p-X-C6H4)SiH (X=OMe, Me, H, F, or CF3) catalysed by 6-Ba.
Experimental conditions: C6D6 (0.5 mL), T=298 K, HN(CH2)4 (3.0m), Ph2(p-XÀ
C6H4)SiH (0.06m).
accuracy of the method the result is commensurate with the
maximal theoretical value.[35] No substantial KIE was detected
for the amine, kHN(CH2)4/kDN(CH2)4 =1.0(1). Both findings are consis-
tent with the kinetic rate law and clearly imply that rupture of
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4571
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ever, this must be mitigated because there is now
a large body of evidence that shows activity varies in
the order Ba<Sr<Ca in the cyclohydroamination of
aminoalkenes.[41] Additionally, no clear trend emerges
in some other cases, for example, in the hydroalkoxy-
lation/cyclisation of alkynyl alcohols.[42] The size of
the Ae ion, its polarizability, electron surface potential
and ability to bind incoming substrates and polarise
their reactive bonds are all unquestionably some of
the features that define the reactivity of the pertain-
ing complexes and their catalytic ability.
Table 2. Hammett analysis for the CDC of pyrrolidine and Ph2(p-XÀC6H4)SiH (X=H,
Me, OMe, F and CF3) catalysed by 6-Ba.[a]
X=OMe
[6-Ba]0 kobs
[mM] [105 sÀ1
X=Me
[6-Ba]0 kobs
[mM] [105 sÀ1
X=H
[6-Ba]0 kobs
[mM] [105 sÀ1
X=F
[6-Ba]0 kobs
[mM] [105 sÀ1
X=CF3
[6-Ba]0 kobs
[mM] [105 sÀ1
]
]
]
]
]
11.0
17.0
28.0
39.0
50.0
2.67
5.17
6.67
9.83
11.2
10.0
17.0
24.0
31.0
42.0
0.33
3.67
4.83
7.33
14.0
4.0
0.33
4.67
8.67
12.7
27.7
5.0
2.00
9.00
13.8
16.3
18.7
5.0
3.83
11.3
27.5
39.2
64.2
11.0
26.0
30.0
60.0
19.0
27.0
32.0
39.0
10.0
17.0
29.0
50.0
[a] Reactions in C6D6 (0.5 mL) at 298 K, [pyrrolidine]0 =3.0m, [Ph2(p-XÀC6H4)SiH]0 =
0.06m.
Computational investigations
lnðkX=kHÞ ¼ spðXÞ1
ð2Þ
To inform ourselves about which of the several con-
ceivable mechanistic scenarios prevails, we embarked on a de-
tailed computational analysis of alternative pathways for the
CDC of pyrrolidine (A) and Ph3SiH (S) catalysed by {N^N}Ba-
{N(SiMe3)2}·(THF)2 (6-Ba, denoted thereafter as C2·(T)2). Various
mechanistic pathways for a catalytically active Ba pyrrolide
(Scheme 5) or alternatively a Ba silyl species (Scheme 6) have
been thoroughly examined, the most accessible of which has
been briefly characterised in our previous communication.[20c]
Taking the route for generation of the silazane product (P) that
commences from the {N^N}Ae pyrrolide by evolving through
a {N^N}Ae hydride intermediate that is converted back into
the {N^N}Ae pyrrolide compound thereafter (Scheme 5) as an
example, the aforementioned kinetic data support both step-
wise and concerted pathways, which further adds to the mech-
anistic diversity. The computational methodology employed
(dispersion-corrected B97-D3 in conjunction with triple-z basis
sets and a sound treatment of bulk solvent effects; see the
Computational methodology in the Experimental section) ade-
quately simulated authentic reaction conditions. The validity of
the computational protocol to reliably map the free-energy
landscape of Ae-mediated hydroelementation reactions has
been substantiated before,[39,41g] and this has allowed mecha-
nistic conclusions with substantial predictive value to be
drawn.
This finding is consistent with the finding that the catalysed
reaction leading to the formation of (p-CF3ÀC6H4)PhSi(N(CH2)4)2
(III) is more rapid than those yielding PhSiH(N(CH2)4)2 (I) and
Ph2Si(N(CH2)4)2 (II) (see Scheme 2).
From the qualitative experiments performed in the prelimi-
nary precatalyst screening (see Table 1) it had emerged that
the catalytic activity increased in the order Ca<Sr<Ba. To
better quantify this observation, the conversion of pyrrolidine
during its CDC with Ph3SiH catalysed by 6-Ca, 6-Sr or 6-Ba at
298 K was monitored by NMR spectroscopy. The observed rate
constants obtained from the semi-logarithmic plot of substrate
conversion versus time (Figure 9) corroborated the earlier find-
Herein we report the full account of examined pathways
that are conceivable for delivery of the silazane product and
also of the disilane by-product in the presence of precatalyst
6-Ba (C2·(T)2). A second investigation scrutinizes the effect of
the alkaline-earth metal (Ae=Ca, Sr) on the energetics of rele-
vant elementary steps, which thereby provides invaluable in-
sights into crucial structure–activity relationships of the CDC
mediated by alkaline-earth iminoanilides.
Figure 9. Comparative semi-logarithmic plot of pyrrolidine conversion versus
^
reaction time for the CDC of pyrrolidine with Ph3SiH catalysed by 6-Ca (
,
obs =3.83”10À5 sÀ1), 6-Sr ( , kobs =15.7”10À5 sÀ1) and 6-Ba ( ,
~
&
k
k
obs =32.5”10À5 sÀ1). Experimental conditions: [pyrrolidine]0 =0.20m, [pyrro-
lidine]0/[Ph3SiH]0/[precatalyst]0 =20:20:1, C6D6 (0.5 mL), T=298 K.
ing that under the chosen experimental conditions, the rate
constants increased by about one order of magnitude from
calcium (kobs =3.83”10À5 sÀ1) to strontium (kobs =15.7”
10À5 sÀ1) and to the most efficacious metal barium (kobs =32.5”
10À5 sÀ1).
Pathways starting from the {N^N}Ba pyrrolide complex
A recent computational study[39] of styrene hydroamination re-
vealed that initial transformation of the {N^N}Ba silylamide
C2·(T)2 precatalyst into the {N^N}Ba pyrrolide compound C3,
which is predominantly present as the bis-amine adduct
C3·(A)2 and exhibits no propensity towards dimer formation, is
sufficiently facile kinetically. For catalytically competent silane
adducts (C3·S·(A)n), a single adducted amine molecule greatly
For a given ligand set, the Ca<Sr<Ba activity trend has
often been reported for a variety of catalytic reactions; for in-
stance in ring-opening polymerisation of cyclic esters,[38] inter-
molecular hydroamination and hydrophosphination of activat-
ed alkenes[29,39] and hydrophosphonylation of ketones.[40] How-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4572
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 5. Plausible mechanistic pathways for iminoanilide alkaline-earth-catalysed CDC of amines and silanes with {N^N}Ae{N(CH2)4} (C3) as the catalytically
competent compound and triphenylsilane (S) and pyrrolidine (A) substrates ({N^N}À ={ArN(o-C6H4)C(H)=NAr}À with Ar=2,6-iPr2-C6H3).
stabilises C3·S, whereas the presence of the rather bulky
Ph3SiH compound makes it impossible for the (N^N)Ba centre
to accommodate another pyrrolidine molecule. Hence, C3·S
and its mono-amine analogue C3·S·A are likely to participate in
accessible pathways for catalyst turnover.
dergo b-hydride elimination. No substantial structural reorgani-
sation is required for transfer of the hydrogen atom from the
five-coordinate silicon atom onto the Ae centre to proceed
through a TS structure with distances of approximately 2.45
and 2.40 Š for vanishing SiÀH and emerging BaÀH bonds (see
Figure S1 in the Supporting Information), respectively, and
generate the silazane product P and {N^N}Ba hydride C5.
The most accessible pathway for stepwise silazane genera-
tion benefits from one associated pyrrolidine spectator mole-
cule participating at all stages of the process (C3·S·A!C4·A!
C5·P·A) on both kinetic and thermodynamic grounds
(Figure 10). As far as the intrinsic reactivity is concerned, both
the NÀSi bond-forming nucleophilic attack of the Ba pyrrolide
(DG°int =4.5 kcalmolÀ1 relative to C3·S·A) and the hydrogen-
atom transfer (DG°int =4.8 kcalmolÀ1 relative to C4·A) are kinet-
ically feasible to an astonishingly comparable amount. In terms
of observable catalytic performance, the nucleophilic attack is
predicted to be kinetically affordable (DG° =17.4 kcalmolÀ1 rel-
ative to C3·(A)2), as is b-hydride elimination (DG° =19.0 kcal
molÀ1 relative to C3·(A)2). The unfavourable thermodynamic
profile of the nucleophilic attack favours C3·S·A over C4·A,
which makes the second hydrogen-atom transfer more kineti-
We start with the examination of a stepwise pathway
(Scheme 5) for silazane formation that comprises nucleophilic
attack of the {N^N}Ae pyrrolide at the silane to furnish the
transient silicate intermediate C4 and subsequent hydrogen-
atom transfer to the Ae centre, which gives rise to the
{N^N}Ae hydride C5 with concomitant release of the silazane
product P. The initial nucleophilic attack of a suitably nucleo-
philic pyrrolide nitrogen centre at S commencing from C3·S or
C3·S·A, featuring only a loosely associated silane molecule,
evolves through a transition-state (TS) structure that describes
NÀSi bond formation occurring outside of the immediate vicin-
ity of the Ae centre at distances of 2.40–2.50 Š (see Figure S1
in the Supporting Information) for the emerging NÀSi bond.
Following the reaction path further, the TS structure decays
into a metastable nucleophilic {N^N}Ba silicate intermediate C4
that features a five-coordinate silicon centre with an elongated
SiÀH linkage, thus indicating the distinct aptitude of C4 to un-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4573
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 6. Plausible mechanistic pathways for iminoanilide alkaline-earth-catalysed CDC of amines and silanes, with {N^N}Ae{Si(C6H5)3} (C6) as the catalytically
competent compound and triphenylsilane (S) and pyrrolidine (A) substrates. A rival pathway for silane homocoupling is also shown. ({N^N}À ={ArN(o-
C6H4)C(H)=NAr}À with Ar=2,6-iPr2-C6H3).
cally demanding. As can be anticipated, the incoming pyrroli-
dine moiety is likely to readily displace the silazane product P
from C5·P·A.[43] It renders the stepwise C3·(A)2!C3·S·A!
C4·A!C5·P·A!C5·(A)2 silazane formation almost thermoneu-
tral; the second hydrogen-atom transfer to Ba dictates the
overall kinetics of this process.
thus this is unlikely to have any relevance for Ba catalyst turn-
over. Whether these findings hold true for more compact alka-
line earths is analysed below.
The subsequent protonolysis of pyrrolidine by the {N^N}Ba
hydride converts C5 back into the catalytically competent
{N^N}Ba pyrrolide for another catalyst turnover. The process
starting from either the mono- or bis-pyrrolidine C5 adduct
evolves through a metathesis-type TS structure (see Figure S2
in the Supporting Information) that decays into amine-free and
amine-adduct forms of the {N^N}Ba pyrrolide compound
through facile liberation of H2 (1 equiv). The participation of
one adducted amine spectator molecule facilitates the process
greatly on thermodynamic and kinetic grounds, thereby paral-
leling the findings for the preceding nucleophilic attack step,
and moreover causes the direct expulsion of H2 after
TS[C5·(A)2–C3·A] is traversed. In light of a kinetically favourable
(DG° =9.6 kcalmolÀ1 relative to C5·(A)2) pyrrolidine protonoly-
sis (C5·(A)2!C3·A+H2(+A)!C3·(A)2) that is almost thermoneu-
On the other hand, despite all our efforts, a low-energy TS
structure describing BaÀN/SiÀH s-bond-breaking metathesis
representing the concerted silazane-generating analogue (see
Scheme 5) could not be located. Careful examination of the
metathesis transformation with the aid of a reaction-path opti-
misation (chain-of-states) method starting from suitable esti-
mated initial TS structures resulted in minimum-energy path-
ways that described either nucleophilic attack or hydrogen-
atom transfer (depending on the initial TS starting point) as
energetically prevalent alternatives. From this, one can con-
clude with some confidence that higher kinetic demands
would be associated with the alternative concerted pathway,
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4574
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 10. Nucleophilic attack of the {N^N}Ba pyrrolide C3 at Ph3SiH (S) together with subsequent hydrogen-atom transfer to barium at the transient silicate
intermediate C4 to afford {N^N}Ba hydride C5 and the silazane product (P).[43,44a] Energies are given in kcalmolÀ1
.
Figure 11. Protonolysis of pyrrolidine by {N^N}Ba hydride C5.[43,44a] Energies are given in kcalmolÀ1
.
tral (Figure 11), hydride and pyrrolide complexes C5·(A)2 and
C3·(A)2 can readily interconvert, thus they are expected to par-
ticipate in a mobile equilibrium.
volves an additionally associated spectator amine molecule,
but favours the {N^N}Ba pyrrolide C6·(A)2 (DG=8.0 kcalmolÀ1).
It characterises the C3·(A)2+S(ÀA)!C6·(A)2 pathway to gener-
ate the {N^N}Ba silyl complex as kinetically viable. Moreover,
the C6·(A)2 (and C6·A) precursors to silazane formation are
somewhat more stable than the encounter species C3·S·A (and
C3·S) that are involved in the alternative mechanistic branch
for silazane formation, hence they can be expected to be
populated in similar appreciable amounts.
Pathways commencing from the {N^N}Ba silyl
We now consider the mechanistic branch that starts from
{N^N}Ba silyl compound C6 (Scheme 6). The BaÀN amido s-
bond silanolysis of {N^N}Ba pyrrolide C3 has been studied as
a plausible route that leads to the generation of C6. The pro-
cess to start from the thermodynamically prevalent C3·(A)2
sees the initial displacement of pyrrolidine by triphenylsilane,
which leads to C3·S·A or, after the release of another adducted
amine molecule, to C3·S encounter species and to proceed
thereafter via a metathesis-type TS structure to furnish mono-
or bis-amine adducts of C6 (see Figure S3 in the Supporting In-
formation). The exchange of the pyrrolide moiety for the silyl
group at C3·(S)2 requires a barrier of 21.3 kcalmolÀ1 to over-
come along the most accessible pathway (Figure 12), which in-
The generation of silazane product P from the {N^N}Ba silyl
C6·(A)n preferably evolves through TS[C6·(A)2–C5·P·A], which
describes concerted BaÀSi/NÀH s-bond-breaking metathesis in
the presence of an adducted pyrrolidine spectator molecule
(see Figure S4 in the Supporting Information). An additional
pyrrolidine unit readily displaces the silazane P from C5·P·A to
yield C5·(A)2,[43] which renders the overall process virtually ther-
moneutral (relative to C3·(A)2). Despite its favourable thermo-
dynamics, the prohibitively high barrier (in excess of 64 kcal
molÀ1; Figure 13) makes the SiÀN/BaÀH s-bond-forming meta-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4575
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 12. BaÀN pyrrolide s-bond protonolysis at C3 by silane substrate S.[43,44a] Energies are given in kcalmolÀ1
.
Figure 13. BaÀSi/NÀH s-bond-breaking metathesis at {N^N}Ba silyl compound C6.[43,44a] Energies are given in kcalmolÀ1
.
thesis kinetically inaccessible. This result parallels the findings
of a previous computational study on silane dehydrocoupling
catalysed by some rare-earth compounds.[45]
of substituted silanes, and also for the substantial primary KIE.
Moreover, the Hammett analysis is inconsistent with those for
other s-bond metathesis processes, which have not shown
substantial charge accumulation in the TS,[46] and provides fur-
ther support for stepwise silazane formation. The DFT-assessed
activation energy for turnover-limiting b-hydride elimination
satisfactorily matches the empirically determined Eyring pa-
rameters.
Moreover, the protonolysis of silane by the {N^N}Ba hydride
(see Figure S5 in the Supporting Information) that would
follow thereafter, thereby regenerating the {N^N}Ba silyl com-
pound, is found to be more demanding energetically
(Figure 14) when compared to the related protonolysis of pyr-
rolidine (Figure 11) via the alternative reaction branch. It is
worth noting that the barrier associated with the prevalent
C5·S·A!C6·A+H2(+A)!C6·(A)2 pathway for silane protonolysis
exceeds the overall barrier for silazane generation starting
from the {N^N}Ba pyrrolide complex C3.
Taking all these findings together, one can confidently dis-
miss the reaction branch that is initiated by the conversion of
{N^N}Ba pyrrolide C3 (or {N^N}Ba silylamide starting material
C2·(T)2) into the {N^N}Ba silyl C6 and involves BaÀSi/NÀH s-
bond-breaking metathesis followed by silane protonolysis in
the presence of an energetically prevalent pathway for step-
wise silazane formation (Figures 10 and 11) with the {N^N}Ba
pyrrolide C3 as the catalytically active compound. The turn-
over-limiting b-hydride elimination accounts for the rate accel-
eration observed for a silane with a para electron-withdrawing
group, the negative charge build-up at silicon (given that
a transient silicate intermediate featuring a hypervalent silicon
centre is involved), which was indicated by a Hammett analysis
These findings rationalise the distinct deviation in reactivity
observed for Sr–pyrrolide 9-Sr and bis-silanide 5-Sr (see
Table 1). On the one hand, 9-Sr does not require further trans-
formation prior to passing through the most accessible path-
way, and thus is competent to catalyse the CDC. In contrast,
an Ae silyl compound is seen to be incapable of effecting sila-
zane formation and would thus make an initial conversion into
an Ae amide indispensable for CDC catalysis to proceed. Com-
plex 5-Sr appears less effective in mediating such a transforma-
tion and/or subsequent nucleophilic attack and b-hydride elim-
ination steps, hence rendering 5-Sr catalytically inactive.
Although disilanes are not part of the observed product
spectrum, we thought it would be informative to examine the
rival homocoupling of Ph3SiH (S) in the presence of {N^N}Ba
silyl C6 (see Figure S6 in the Supporting Information). Both
stepwise and concerted pathways are conceivable, but a step-
wise pathway (dashed arrows in Scheme 6) that bears a formal
similarity with the one shown in Figure 10 is energetically prev-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4576
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 14. Protonolysis of the silane by the {N^N}Ba hydride C5.[43,44a] Energies are given in kcalmolÀ1
.
alent. As discovered for all the steps studied thus far, one asso-
ciated amine spectator molecule A greatly facilitates silane ho-
mocoupling. In light of their formal analogy, it is tempting to
compare the inherent propensity of the {NN}Ba silyl C6 and
{NN}Ba pyrrolide C3 to nucleophilically attack triphenylsilane S
and to effect b-H elimination from a five-coordinate silicon
centre onto the alkaline earth. Considering the intrinsic barri-
ers, SiÀSi bond formation via C6·S·A!C7·A is more kinetically
and thermodynamically demanding (DG°int/DGint =13.8/9.9 kcal
molÀ1 relative to C6·S·A; Figure 15) than C3·S·A!C4·A NÀSi
bond formation (DG°int/DGint =4.5/1.3 kcalmolÀ1 relative to
C3·S·A; Figure 10); this reflects the enhanced nucleophilicity of
the pyrrolide N centre together with some steric constraints as-
sociated with silane homocoupling. By contrast, b-H elimina-
tion from the hypervalent silicon centre at transient silicate in-
termediates C7·A (DG°int =5.4 kcalmolÀ1 relative to C7·A;
Figure 15) and C4·A (DG°int =4.8 kcalmolÀ1 relative to C4·A;
Figure 10) is equally found highly facile. Despite its associated
low intrinsic barrier, the fact that the nucleophilic attack of the
{N^N}Ba silyl is strongly uphill means that the second hydro-
gen transfer step determines the overall kinetics for the gener-
ation of Si2Ph6 and {N^N}Ba hydride C5. Figure 15 reveals a pro-
hibitively large overall barrier (DG° =36.2 kcalmolÀ1 relative to
C3·(A)2) for the rival silane homocoupling, which remains un-
achievable due to its non-competitive kinetic demands versus
the most accessible stepwise pathway for silazane formation
(Figures 10 and 11).
Comparison of {N^N}Ae(NR2)-mediated CDC reactions
(Ae=Ca, Sr, Ba)
We start with the examination of the accessibility of the path-
way for concerted silazane formation (shown in Scheme 5) for
the Ca catalyst analogue. In contrast to the findings for
barium, the TS structures for AeÀN/SiÀH s-bond-breaking
metathesis could be located for the more compact Ca centre;
the key structural and energy features are summarised in Fig-
ures S7 and S8 (see the Supporting Information). Irrespective
of whether a spectator pyrrolidine molecule participates in
metathesis pathways, they are distinctly disfavoured kinetically
Figure 15. Nucleophilic attack of the {N^N}Ba silyl C6 at Ph3SiH (S) together with subsequent hydrogen-atom transfer to barium at the transient intermediate
C7 to afford {N^N}Ba hydride C5 and the disilane product Si2Ph6.[43,44a] Energies are given in kcalmolÀ1
.
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4577
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 16. Nucleophilic attack of the {N^N}Ae pyrrolide C3 at the triphenylsilane S together with subsequent hydrogen transfer to the alkaline earth at the
transient silicate intermediate C4 to afford {N^N}Ae hydride C5 and the silazane (P) product (Ae=Ca/Sr/Ba).[43, 44] Energies are given in kcalmolÀ1
.
(DDG° >14 kcalmolÀ1; Figure S8 in the Supporting Informa-
tion) relative to the corresponding pathway for stepwise sila-
zane production (Figure 16). The pronounced energy gap be-
tween the stepwise and concerted pathways assessed for the
Ca catalyst further corroborates our above conclusion that a s-
bond metathesis pathway remains inaccessible in the presence
of an energetically prevalent stepwise pathway for all the alka-
line earths studied.
C3·(A)2 catalyst resting state and the C3·S·A precursor species
for nucleophilic pyrrolide attack at S for ever more compact
metal centres becomes understandable. The overall barrier as-
sociated with nucleophilic attack is likewise seen to increase
regularly (Ca>Sr>Ba). Of particular note is the modest, but
regular, decrease of the intrinsic activation energy for ever
heavier alkaline earths (DG°int =5.3, 4.7 and 4.5 kcalmolÀ1 for
Ca, Sr and Ba, respectively, relative to C3·S·A) for the most ac-
cessible C3·S·A!C4·A pathway. Hence, limitations in the acces-
sibility of the Ae centre together with variations in the AeÀN
pyrrolide bond strength contribute towards the greater energy
demands for lighter alkaline earths, with the former factor
likely to be dominant. The subsequent b-hydride elimination
appears to be turnover limiting for all the systems studied, and
its total barrier follows a regular Ca>Sr>Ba trend (Figure 16).
It is perhaps instructive to look more closely at the intrinsic
process energetics (DG°int =7.4, 6.5 and 6.1 kcalmolÀ1 and
DGint =3.0, À0.9 and À2.7 kcalmolÀ1 for Ca, Sr and Ba, respec-
tively, relative to C3·S·A). Again a regular trend is displayed,
Focusing exclusively on the prevalent silazane-generating
pathways shown in Figures 10 and 11, we can analyse the
extent to which the nature of the alkaline earth influences the
relevant elementary steps. The generation of silazane P after
consecutive nucleophilic attack and b-hydride elimination
(Figure 16) and also subsequent conversion of the {N^N}Ae hy-
dride C5 back into the catalytically competent {N^N}Ae pyrro-
lide C3 (Figure 17) benefit from involvement of an adducted
pyrrolidine spectator molecule. In light of the increased acces-
sibility of the alkaline earth upon going down the Group 2, the
regular growth of the thermodynamic gap between the
Figure 17. Protonolysis of pyrrolidine by the {N^N}Ae hydride C5 (Ae=Ca/Sr/Ba).[43, 44] Energies are given in kcalmolÀ1
.
Chem. Eur. J. 2016, 22, 4564 – 4583 www.chemeurj.org ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4578

Full Paper
which indicates that AeÀH bond formation is more kinetically
and thermodynamically favourable upon going down Group 2.
Hence, a sterically less restricted barium centre in combination
with an increased aptitude for AeÀH bond formation give rise
to the enhanced performance of the Ba catalyst relative to its
lighter analogues.
tial hydrosilane (e.g. Ph3SiH, Et3SiH), stoichiometric cascade
couplings of diamines (H2NCH2ÀC6H4ÀCH2NH2, piperazine, etc.)
or primary amines (e.g. BnNH2) with hydrodisilanes (e.g.
Ph2SiH2) can yield unusual polysilazanes. Encoded and/or func-
tionalised (macro)molecules can be prepared from judiciously
selected bifunctional substrates, such as p-substituted arylsi-
lanes. The Ae-promoted dehydrocoupling of amines with bor-
anes and the coupling of silanes with phosphines are other
natural extensions of this work. Our continued efforts will be
detailed in forthcoming reports.
The regeneration of the catalytically competent {N^N}Ae
pyrrolide C3 through protonolysis of pyrrolidine A by the
{N^N}Ae hydride C5 with concomitant release of H2 (Figure 17)
is significantly less kinetically demanding than stepwise sila-
zane formation (Figure 16) for the studied family of iminoani-
lide alkaline-earth catalysts. Given that the thermodynamic
force that drives this conversion is rather small, {N^N}Ae pyrro-
lide and {N^N}Ae hydride compounds can be expected to par-
ticipate in a mobile equilibrium.
Experimental Section
General considerations
All manipulations were performed under an inert atmosphere by
using standard Schlenk techniques or in a dry, solvent-free glove-
box (Jacomex; O2 <1 ppm, H2O<5 ppm). AeI2 beads (99.999%, Al-
drich) was used as purchased. THF was distilled under argon from
Na/benzophenone prior to use. Pentane, toluene, dichlorome-
thane, and Et2O were collected from MBraun SPS-800 purification
alumina columns. Deuterated solvents (Eurisotop, Saclay, France)
were stored in sealed ampoules over activated 3 Š molecular
sieves and degassed by several freeze–thaw cycles. Ae{N(Si-
Me3)2}2·(THF)2,[47] Ae{N(SiMe3)2}2,[47a] Ae{N(SiMe2H)2}2,[25,48] Ae{CH(Si-
Me3)2}2·(THF)3,[49] {NN}Ae{N(SiMe3)2}·(THF)x[20c,25] and {NN}Ae{CH(Si-
Me3)2}·(THF)x[20c,25] (Ae=Ca, Sr or Ba) were prepared by following lit-
erature procedures.
Conclusion
Alkaline-earth metal complexes constitute excellent chemose-
lective and versatile precatalysts for the production for a range
of silazanes by CDC of amines and hydrosilanes. The scope of
the reaction is large and includes diamines and di(hydrosi-
lane)s. Characteristically, the catalytic activity increases substan-
tially upon going down Group 2 according to (Mg !)Ca<Sr<
Ba. This observation has been rationalised by DFT calculations,
which show that greater accessibility of the metal centre and,
to a lesser extent, decreasing AeÀNamide bond strength down
the Group 2 series, explain this trend. However, these conclu-
sions solely apply to the CDC catalysis presented herein, and
should not be extended to all Ae-mediated catalytic processes.
The barium precatalysts Ba{CH(SiMe3)2}2·(THF)3 and Ba{N(Si-
Me3)2}2·(THF)2 display impressive performance, and because
their synthesis is relatively facile they are recommended for
the catalytic silazane formation.
NMR spectra were recorded with Bruker AM-400 or AM-500 spec-
trometers. All chemical shifts (d) [ppm] were determined relative to
the residual signal of the deuterated solvent and calibrated against
SiMe4. Assignment of the signals was assisted by 1D (1H, 13C{1H})
and 2D (COSY, HMBC, and HMQC) NMR experiments.
HRMS data were recorded with a Bruker MicrOTOF-Q II mass spec-
trometer equipped with an APCI (Atmospheric Pressure Chemical
Ionisation) source in positive mode by direct introduction (ASAP—
Atmospheric Solids Analysis Probe) at 3708C.
The operative mechanism of the Ae-mediated CDC catalysis
has been established with a high level of confidence following
a kinetic analysis and DFT investigations of several rival path-
ways. In the presence of a catalytically competent iminoanilide
Ba amide or Ba alkyl complex, pathways that involve a Ba silyl
to mediate CDC catalysis or to furnish a disilane by-product
are non-competitive. This is fully consistent with the experi-
mental data: the formation of these species was not detected
spectroscopically during the catalytic CDC reactions. Instead,
complementary experimental and computational data provide
conclusive evidence that CDC of amines and organosilanes fol-
lows a stepwise reaction path that consists of 1) NÀSi bond-
forming nucleophilic attack of the barium pyrrolide onto the
silane, followed by 2) turnover-limiting hydrogen-atom transfer
to the barium centre. A related mechanism was proposed for
the CDC precatalyst {ToM}Mg{N(SiMe3)2}, for which nucleophilic
attack was proposed to be ratedetermining.[20a]
Combustion analyses of 5-Sr (C48H54O3Si2Sr, 822.75 gmolÀ1), 5-Ba
(C48H54O3Si2Ba,
872.45 gmolÀ1),
7-Ba
(C43H69N3O2Si2Ba,
853.54 gmolÀ1) and 9-Sr (C39H56N4Sr, 668.53 gmolÀ1) were per-
formed by with a Carlo–Erba analyzer. Despite repeated (>4) at-
tempts on different batches of each of these compounds (includ-
ing for crystals of 7-Ba, which were also characterised by XRD), we
were unable to obtain satisfactory and reproducible elemental
analyses for these compounds. This can certainly be attributed to
the extreme air- and moisture-sensitivity of the complexes, and
also to the presence of substantial amounts of silicon in the sam-
ples leading to the formation of non-pyrolisable silicon carbides.
Complex 5-Sr
SrI2 (300 mg, 0.8 mmol) was suspended in THF (10 mL) and stirred
at 608C until dissolution was ensured (ca. 60 min). KSiPh3 (500 mg,
1.6 mmol) was dissolved in THF (10 mL) and added dropwise via
cannula to the SrI2 solution at RT. The reaction mixture was stirred
for 2.5 h and a white precipitate gradually appeared. The precipi-
tate was removed by filtration and the solvent was removed under
vacuum. Complex 5-Sr was obtained as a yellow powder (230 mg,
32%) after washing with pentane (3”5 mL). 1H NMR (C6D6,
500.1 MHz, 298 K): d=7.76 (m, 12H; o-C6H5), 7.47 (m, 12H; m-C6H5),
7.12 (m, 6H; p-C6H5), 3.24 (m, 12H; OCH2CH2), 1.24 ppm (m, 12H;
OCH2CH2); 13C{1H} NMR (C6D6, 125.8 MHz, 298 K): d=136.9 (o-C6H5),
The chemoselectivity and compatibility of these Ae precata-
lysts with monoamines, diamines, hydrosilanes and di(hydrosi-
lane)s opens up a range of synthetic possibilities. The synthesis
of original polycarbosilazanes by coupling of diamines and di-
(hydrosilane)s is an area where these Ae-mediated CDC pro-
cesses can prove invaluable.[32] Moreover, starting from an ini-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4579
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
136.4 (m-C6H5), 129.9 (p-C6H5), 68.4 (OCH2CH2), 25.6 ppm (OCH2CH2).
29Si{1H} NMR (C6D6, 79.5 MHz, 298 K): d=À12.14 ppm.
H), 6.21 (m, 2H; arom-H), 3.23 (br, 2H; CH(CH3)2), 3.05 (br, 2H;
CH(CH3)2), 2.49 (br, 8H; HNCH2CH2 (4H) and SrNCH2CH2 (4H)), 1.27
(br, 8H; HNCH2CH2 (4H) and SrNCH2CH2 (4H)), 1.23 (d, 3J(H,H)=
6.9 Hz, 6H; CH(CH3)2), 1.11 (m, 12H; CH(CH3)2), 1.06 (d, 3J(H,H)=
6.7 Hz, 6H; CH(CH3)2), 0.94 ppm (brs, 1H; HN(CH2)4); 13C{1H} NMR
(C6D6, 125.8 MHz, 298 K): d=172.58 (CH=N), 160.0 (N=CHÀi-C6H4),
149.7 (ArNÀi-C6H3), 143.8 (CH=NÀi-C6H3), 140.7 (ArNÀo-C6H3), 139.5
(CH=NÀo-C6H3), 133.5 (NÀi-C6H4), 126.4 (CH=NÀo-C6H3), 124.7 (N=
CHÀm-C6H4), 124.4 (N=CHÀp-C6H4), 119.6 (ArNÀp-C6H3), 116.4 (NÀo-
C6H4), 111.7 (CH=NÀp-C6H3), 49.2 (BaNCH2CH2), 34.5 (HNCH2CH2),
29.3 (BaNCH2CH2), 28.7 (CH(CH3)2), 26.5 (CH(CH3)2), 25.7 (HNCH2CH2),
24.6 (CH(CH3)2), 23.8 (CH(CH3)2), 22.0 (CH(CH3)2), 22.7 ppm
(CH(CH3)2).
Complex 5-Ba
BaI2 (328 mg, 0.8 mmol) was suspended in THF (10 mL) and stirred
at 608C until dissolution was ensured (ca. 60 min). KSiPh3 (500 mg,
1.6 mmol) was dissolved in THF (10 mL) and added dropwise via
cannula to the BaI2 solution at RT. The reaction mixture was stirred
for 2.5 h and a white precipitate gradually appeared. The precipi-
tate was removed by filtration and the solvent was removed under
vacuum. Complex 5-Ba was obtained as an orange powder
(600 mg, 86%) after washing with pentane (3”5 mL). 1H NMR
(C6D6, 500.1 MHz, 298 K): d=7.68 (m, 12H; o-C6H5), 7.46 (m, 12H;
m-C6H5), 6.96 (m, 6H; p-C6H5), 3.54 (m, 12H; OCH2CH2), 1.40 ppm
(m, 12H; OCH2CH2); 13C{1H} NMR (C6D6, 125.8 MHz, 298 K): d=
136.63 (o-C6H5), 130.4 (m-C6H5), 125.9 (p-C6H5), 68.3 (OCH2CH2),
25.7 ppm (OCH2CH2); 29Si{1H} NMR (C6D6, 79.5 MHz, 298 K): d=
À12.11 ppm.
NMR-scale generation of complex 10-Ba
In a glovebox, {N^N}Ba{N(SiMe3)2}·(THF)2 (10 mg, 0.011 mmol) was
added to an NMR tube. The NMR tube was stored in a Schlenk
tube, which was removed from the glovebox and connected to
a Schlenk manifold. Standard Schlenk techniques were used for
subsequent manipulations. C6D6 (0.5 mL) and pyrrolidine (ꢁ2.0 mL,
0.022 mmol) were added to the NMR tube. The NMR tube was
sealed and shaken vigorously, and then put into an oil bath at
258C for 2 h. The reaction mixture was directly analysed by
Complex 7-Ba
Anhydrous BaI2 beads (446 mg, 1.14 mmol) were suspended in THF
(10 mL) and activated at 608C for 60 min. KN(SiMe2H)2 (373 mg,
2.18 mmol) and {N^N}H (480 mg, 1.09 mmol) were dissolved in
THF (10 mL) and the solution was added dropwise via cannula to
the BaI2 solution at RT. The reaction mixture was stirred for 2.5 h
and a white precipitate gradually appeared. The precipitate was re-
moved by filtration and the solvent was pumped off under
vacuum. Complex 7-Ba was extracted with pentane (3”5 mL) and
isolated as orange crystals (521 mg, 56%) by crystallisation at
À278C. Crystals suitable for X-ray diffraction studies were selected
1
1H NMR spectroscopy. H NMR (C6D6, 500.1 MHz, 298 K): d=8.05 (s,
1H; CH=N), 7.19 (d, 3J(H,H)=7.6 Hz, 2H; arom-H), 7.16 (m, 1H;
arom-H), 7.11 (m, 4H; arom-H), 6.91 (td, 3J(H,H)=8.6 Hz, 4J(H,H)=
3
1.8 Hz, 1H; arom-H), 6.29 (t, J(H,H)=8.0 Hz, 1H; arom-H), 6.22 (d,
3J(H,H)=8.8 Hz, 1H; arom-H), 3.50 (br, 2H; CH(CH3)2), 3.27 (hept,
3J(H,H)=6.8 Hz, 2H; CH(CH3)2), 2.41 (br, 8H; HN(CH2)4), 1.33 (d,
3J(H,H)=7.0 Hz, 6H; CH(CH3)2), 1.27 (br, 20H; HNCH2CH2 (8H) and
CH(CH3)2 (12H)), 1.23 (d, 3J(H,H)=6.7 Hz, 6H; CH(CH3)2), 0.60 (br,
2H; HN(CH2)4), 0.21 ppm (s, 18H; Si(CH3)3); 13C{1H} NMR (C6D6,
125.8 MHz, 298 K): d=169.3 (CH=N), 158.0 (N=CHÀi-C6H4), 149.8
(ArNÀi-C6H3), 146.6 (CH=NÀi-C6H3), 141.0 (ArNÀo-C6H3), 139.6 (CH=
NÀo-C6H3), 133.9 (NÀi-C6H4), 126.0 (CH=NÀo-C6H3), 124.9 (N=CHÀm-
C6H4), 124.9 (N=CHCp-C6H4), 124.7 (ArNÀp-C6H3), 118.6 (NÀo-C6H4),
111.1 (CH=NÀp-C6H3), 50.3 (HNCH2CH2), 30.5 (CH(CH3)2), 29.4
(CH(CH3)2), 28.8 (HNCH2CH2), 26.4 (CH(CH3)2), 26.2 (CH(CH3)2), 25.7
(CH(CH3)2), 24.0 (CH(CH3)2), 5.8 ppm (Si(CH3)3).
1
from this batch. H NMR (C6D6, 500.1 MHz, 298 K): d=8.04 (s, 1H;
3
CH=N), 7.28 (d, J(H,H)=7.6 Hz, 2H; arom-H), 7.19 (m, 1H; arom-H),
7.12 (m, 3H; arom-H), 7.00 (m, 1H; arom-H), 6.92 (td, 3J(H,H)=
8.6 Hz, 4J(H,H)=1.8 Hz, 1H; arom-H), 6.28 (t, 3J(H,H)=8.2 Hz, 1H;
arom-H), 6.22 (d, 3J(H,H)=8.7 Hz, 1H; arom-H), 4.71 (br, 1J(H,Si)=
160 Hz, 2H; Si(CH3)2H), 3.36 (m, 10H; OCH2CH2 (8H) and CH(CH3)2
(2H)), 3.18 (q, 3J(H,H)=6.8 Hz, 2H; CH(CH3)2), 1.31 (m, 20H;
OCH2CH2 (8H) and CH(CH3)2 (12H)), 1.23 (d, 3J(H,H)=6.7 Hz, 6H;
CH(CH3)2), 1.17 (d, 3J(H,H)=6.7 Hz, 6H; CH(CH3)2), 0.24 ppm (br,
12H; Si(CH3)2H). 13C{1H} NMR (C6D6, 125.8 MHz, 298 K): d=169.2
(CH=N), 158.0 (N=CHÀi-C6H4), 149.5 (ArNÀi-C6H3), 146.6 (CH=NÀi-
C6H3), 144.6 (NÀi-C6H4), 140.8 (ArNÀo-C6H3), 139.6 (N=CHÀo-C6H4),
133.7 (NÀo-C6H4), 125.7 (CH=NÀo-C6H3), 125.3 (ArNÀm-C6H3), 124.8
(CH=NÀm-C6H3), 124.7 (N=CHÀm-C6H4), 118.6 (CH=NÀi-C6H3), 118.4
(ArNÀp-C6H3), 110.5 (N=CHÀp-C6H4), 68.9 (OCH2CH2), 29.2 (CH(CH3)2),
28.5 (CH(CH3)2), 26.4 (CH(CH3)2), 26.2 (CH(CH3)2), 25.7 (OCH2CH2),
25.4 (CH(CH3)2), 23.8 (CH(CH3)2), 5.1 ppm (Si(CH3)2H); FTIR (Nujol in
KBr plates): n˜ =2964 (s), 2887 (s), 2004 (s), 1984 (sh), 1601 (sh),
1579 (m), 1380 (s), 1377 (s), 1359 (w), 1338 (s), 1234 (s), 1151 (s),
1097 (s), 1037 (w), 1004 (m), 947 (s), 916 (s), 887 (w), 827 cmÀ1 (m).
Typical procedure for catalytic CDC reactions
The following standard protocol was used for the catalytic CDC re-
actions. In a glovebox, precatalyst 1–9 (0.005 mmol, 1 equiv) was
loaded into an NMR tube. The NMR tube was stored in a Schlenk
tube, which was then removed from the glovebox and connected
to a double-manifold Schlenk line. Standard Schlenk techniques
were used for subsequent manipulations. The amine (n”
0.005 mmol, n equiv) and silane (n”0.005 mmol, n equiv) were
added to the NMR tube via syringe. The NMR tube was sealed, re-
moved from the Schlenk tube, shaken vigorously, then placed in
an oil bath at 298 or 333 K. After the required time period, the re-
action was quenched by addition of “wet” C6D6 at RT. Substrate
conversion was determined from the 1H NMR spectrum of the reac-
tion mixture, by comparing the relative intensities of characteristic
resonances of the substrates and products. All silazanes were iso-
lated as solids after quenching the reaction mixture, evaporating
the volatile compounds, washing the residue with pentane and
drying it to constant weight.
Complex 9-Sr
{N^N}Sr{CH(SiMe3)2}2·(THF)2 (250 mg, 0.29 mmol) was dissolved in
pentane (10 mL). Pyrrolidine (69 mL, 0.86 mmol) was diluted in pen-
tane (5 mL) and the solution was added dropwise via cannula to
the precatalyst solution. The reaction mixture was stirred for 2.5 h
at RT. At the end of the reaction the solution was stored at À278C
and 9-Sr was isolated as an orange solid after precipitation from
1
solution (102 mg, 51%). H NMR (C6D6, 500.1 MHz, 298 K): d=8.07
(s, 1H; CH=N), 7.25 (m, 3H; arom-H), 7.08 (m, 3H; arom-H), 7.01 (m,
1H; arom-H), 6.79 (td, 3J(H,H)=6.8 Hz, 4J(H,H)=3.4 Hz, 1H; arom-
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4580
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tional protocol employed for reliably mapping the energy land-
scape of Ae-mediated hydroamination has been substituted be-
fore,[39,41g] and this allowed mechanistic conclusions with substan-
tial predictive value to be drawn.
Procedure for kinetic measurements
The following protocol was used for catalytic CDC reactions moni-
1
tored by H NMR spectroscopy. In a glovebox, precatalyst 6-Ba and
Ph3SiH were loaded into an NMR tube. The NMR tube was stored
in a Schlenk tube, which was then removed from the glovebox
and connected to a double-manifold Schlenk line. Pyrrolidine and
C6D6 (0.5 mL) were added to the NMR tube via syringe. The NMR
tube was sealed, removed from the Schlenk tube, vigorously
shaken, and inserted into the probe of a Bruker AM 500 NMR spec-
trometer preheated to the required temperature. The reaction ki-
netics were monitored starting from this point using the mul-
ti zgvd command of the TopSpin package (D1=0.2 s; DS=0;
NSꢂ8). The conversion was determined on the basis of amine con-
sumption over the course of three or more half-lives, by comparing
the relative intensities of characteristic resonances of the sub-
strates and products.
The reaction pathways were explored by
a chain-of-states
method[60] as implemented in the module woelfling in the TURBO-
MOLE suite of programmes, which makes use of reasonably
chosen reactant and product structures to approximate the mini-
mum-energy path (MEP). The reactant and product states to be
linked to the associated transition state were identified. The ap-
proximate saddle points connected with the MEP were subjected
to an exact localisation of the TS structures. No structural simplifi-
cation of any of the key species involved was imposed, nor have
symmetry constraints have been imposed in any case. The DFT cal-
culations simulated the authentic reaction conditions by treating
the bulk effects of the benzene solvent by a consistent continuum
model in the form of the conductor-like screening model for realis-
tic solvents (COSMO-RS)[61] as implemented in COSMOtherm.[62] This
solvation model includes continuum electrostatic and also solvent-
cavitation and solute-solvent dispersion effects through surface-
proportional terms, and it also refers properly to a 1m standard
state. The free solvation enthalpy was assessed with the aid of
COSMO-RS at the BP86/(SDD+def2-TZVPD)//BP86-D3/(SDD+
TZVP) level of approximation. Geometry optimisation and frequen-
cy calculations were also performed at the BP86-D3/(SDD+
SV(P))[58c] level to confirm the nature of all optimised key structures
and to determine the thermodynamic parameters (298 K, 1 atm)
under the conventional ideal-gas, rigid-rotor and quantum-me-
chanical harmonic-oscillator approximations. This level of basis-set
quality is known to be reliable for the assessment of structural pa-
rameters and vibrational frequencies,[63] thus it allows an affordable
and accurate determination of thermodynamic state functions. As
far as the vibrational partition function is concerned, all frequen-
cies below 60 cmÀ1 were replaced by a value of 60 cmÀ1 to correct
for the quantum-mechanical harmonic-oscillator approximation for
such normal modes (the quasi-harmonic approximation),[64] and
thermochemical contributions to enthalpy and entropy were com-
puted from the resulting partition function. The entropy contribu-
tions for condensed-phase conditions were estimated based on
computed gas-phase entropies (without any scaling applied) by
employing Okuno’s procedure.[65] The mechanistic conclusions
drawn in this study were based on the Gibbs free-energy profile of
the catalytic cycle assessed at the B97-D3(COSMO-RS)/(SDD+def2-
TZVP) level of approximation for experimental condensed-phase
conditions. The calculated structures were visualised by employing
the StrukEd program,[66] which was also used for the preparation of
3D molecular drawings.
X-ray diffraction crystallography
Crystals of 7-Ba suitable for X-ray diffraction analysis were obtained
by recrystallisation from pentane at À278C. Diffraction data were
collected at 150(2) K by using a Bruker APEX CCD diffractometer
with graphite-monochromated MoKa radiation (l=0.71073 Š). A
combination of w and F scans was carried out to obtain a unique
data set. The crystal structures were solved by direct methods, the
remaining atoms were located from difference Fourier synthesis
followed by full-matrix least-squares refinement based on F2 (pro-
grammes SIR97 and SHELXL-97).[50] Carbon- and oxygen-bound hy-
drogen atoms were placed at calculated positions and forced to
ride on the attached atom. All non-hydrogen atoms were refined
with anisotropic displacement parameters. The locations of the
largest peaks in the final difference Fourier map calculation, as well
as the magnitude of the residual electron densities, were of no
chemical significance. Relevant collection and refinement data are
given in the Supporting information. CCDC 1426816 (7-Ba) con-
tains the supplementary crystallographic data for this paper. These
data are provided free of charge by The Cambridge Crystallograph-
ic Data Centre.
Computational methodology
All calculations based on Kohn–Sham DFT[51] were performed by
using the program package TURBOMOLE[52] with flexible triple-z
basis sets. The Becke–Perdew (BP86)[53] generalised gradient ap-
proximation (GGA) functional within the RI-J integral approxima-
tion[54] in conjunction with appropriate auxiliary basis sets was
used for structure optimisation. Empirical dispersion corrections by
Grimme (D3 with Becke–Johnson damping)[55] were used to ac-
count for non-covalent interactions. For Ba and Sr, we used the
Stuttgart–Dresden scalar-relativistic effective core potential (SDD,
46 and 28 core electrons for Ba and Sr, respectively)[56] in combina-
tion with the (7s7p5d1f)/[6s4p3d1f] (def2-TZVPP) valence basis
set;[57] Ca was treated by the (17s12p4d)/[6s5p3d] (def2-TZVPP) all-
electron basis set.[57] All remaining elements were represented by
Ahlrich’s valence triple-z TZVP basis set[58a–b] with polarisation func-
tions on all atoms. Final potential energies were obtained by
single-point calculations on BP86-D3 optimised structures with the
B97-D[59] GGA functional (together with D3(BJ) empirical dispersion
correction)[55] in conjunction with the aforementioned ECP/basis
set for the Ae atoms and a def2-TZVP[54] basis set for all remaining
Acknowledgements
This work was sponsored by the French Ministry for Higher
Education and Research (Ph.D. grant to C.B.). The authors
thank Stephen Boyer (London Metropolitan University) for per-
forming the elemental analyses.
Keywords: alkaline-earth metals · barium · density functional
calculations · cross-dehydrocoupling · reaction mechanisms
[1] a) J. F. Harrod, Coord. Chem. Rev. 2000, 206–207, 493; b) R. Waterman,
Chem. Soc. Rev. 2013, 42, 5629; c) E. M. Leitao, T. Jurca, I. Manners, Nat.
Chem. 2013, 5, 817; d) R. L. Melen, Chem. Soc. Rev. DOI: 10.1039/
c5s00521c.
elements (B97-D3/(SDD+def2-TZVP)//BP86-D3/(SDD+TZVP)).
large integration grid (m4 in TURBOMOLE notation) and tight SCF
convergence criteria have been used. The validity of the computa-
A
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4581
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[2] a) T. Wideman, L. G. Sneddon, Chem. Mater. 1996, 8, 3; b) J. E. Mark,
H. R. Allcock, R. West, Inorganic Polymers, 2nd ed., Oxford University
Press, 2005; c) T. Baumgartner, R. RØau, Chem. Rev. 2006, 106, 4681;
d) A. Nunns, J. Gwyther, I. Manners, Polymer 2013, 54, 1269; e) A. K.
Franz, S. O. Wilson, J. Med. Chem. 2013, 56, 388.
[3] M. Lappert, P. P. Power, A. Protchenko, A. Seeber, Metal Amide Chemistry,
Wiley, Chichester, 2009.
[4] L. F. Fieser, M. Fieser, Reagents in Organic Chemistry, Wiley, New York,
1967.
Maron, J. Okuda, Chem. Commun. 2014, 50, 2311; b) V. Leich, T. P. Span-
iol, J. Okuda, Inorg. Chem. 2015, 54, 4927. The difference in the
29Si NMR chemical shifts between Ca(SiPh3)2·(THF)4 and 5-Sr/5-Ba
(d29Si =À12.14 and À12.11 ppm) reflects the presence of one extra THF
molecule in the Ca complex, and these values altogether hint at highly
ionic Ae···Si bonding. For other examples of Ae silanides, see: c) W.
Teng, U. Englich, K. Ruhlandt-Senge, Angew. Chem. Int. Ed. 2003, 42,
3661; Angew. Chem. 2003, 115, 3789; d) D. J. Liptrot, M. Arrowsmith,
A. L. Colebatch, T. J. Hadlington, M. S. Hill, G. Kociok-Kçhn, M. F. Mahon,
Angew. Chem. Int. Ed. 2015, 54, 15280; Angew. Chem. 2015, 127, 15495.
[24] Y. Sarazin, D. Ros¸ca, V. Poirier, T. Roisnel, A. Silvestru, L. Maron, J.-F. Car-
pentier, Organometallics 2010, 29, 6569.
[5] a) Y. Tanabe, M. Murakami, K. Kitaichi, Y. Yoshida, Tetrahedron Lett. 1994,
35, 8409; b) Y. Tanabe, T. Misaki, M. Kurihara, A. Iida, Y. Nishii, Chem.
Commun. 2002, 1628.
´
[6] a) P. J. Kocienski, Protecting Groups, 3rd ed., Thieme, Stuttgart, 2005,
[25] B. Liu, T. Roisnel, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 2013, 19,
13445.
pp. 595–599; b) D. W. Robbins, T. A. Boebel, J. F. Hartwig, J. Am. Chem.
Soc. 2010, 132, 4068; c) S. Itagaki, K. Kamata, K. Yamaguchi, N. Mizuno,
Chem. Commun. 2012, 48, 9269; d) C. D. F. Kçnigs, M. F. Müller, N. Aigua-
bella, H. F. T. Klare, M. Oestreich, Chem. Commun. 2013, 49, 1506; e) L.
Greb, S. Tamke, J. Paradies, Chem. Commun. 2014, 50, 2318.
[7] a) D. Seyferth, G. H. Wiseman, C. Prud’homme, J. Am. Ceram. Soc. 1983,
66, C-13; b) D. Seyferth, G. H. Wiseman, J. Am. Ceram. Soc. 1984, 67, C-
132; c) Y. D. Blum, K. B. Schwartz, R. M. Laine, J. Mater. Sci. 1989, 24,
1707; d) D. Seyferth, J. M. Schwark, R. M. Stewart, Organometallics 1989,
8, 1980; e) E. Duguet, M. Schappacher, A. Soum, Macromolecules 1992,
25, 4835; f) N. R. Dando, A. J. Perrotta, C. Strohmann, R. M. Stewart, D.
Seyferth, Chem. Mater. 1993, 5, 1624; g) M. Birot, J.-P. Pilot, J. Dunogu›s,
Chem. Rev. 1995, 95, 1443; h) D. Su, Y. Li, F. Hou, X. Yan, J. Am. Ceram.
Soc. 2014, 97, 1311.
[8] R. Fessenden, J. S. Fessenden, Chem. Rev. 1961, 61, 361.
[9] a) L. H. Sommer, J. D. Citron, J. Org. Chem. 1967, 32, 2470; b) J. F. Blan-
dez, I. Esteve-Adell, M. Alvaro, H. García, Catal. Sci. Technol. 2015, 5,
2167.
[10] a) Y. Blum, R. M. Laine, Organometallics 1986, 5, 2081; b) C. Biran, Y. D.
Blum, R. Glaser, D. S. Tse, K. A. Youngdahl, R. M. Laine, J. Mol. Catal.
1988, 48, 183; c) C. K. Toh, H. T. Poh, C. S. Lim, W. Y. Fan, J. Organomet.
Chem. 2012, 717, 9.
[11] W. D. Wang, R. Eisenberg, Organometallics 1991, 10, 2222.
[12] E. J. Matarasso-Tchiroukhine, Chem. Soc. Chem. Commun. 1990, 681.
[13] H. Q. Liu, J. F. Harrod, Can. J. Chem. 1992, 70, 107.
[14] a) H. Q. Liu, J. F. Harrod, Organometallics 1992, 11, 822; b) for the related
coupling of hydrosilanes and hydrazines, see: J. He, H. Q. Liu, J. F.
Harrod, R. Hynes, Organometallics 1994, 13, 336.
[15] L. K. Allen, R. García-Rodríguez, D. S. Wright, Dalton Trans. 2015, 44,
12112.
[16] A. E. Nako, W. Chen, A. J. P. White, M. R. Crimmin, Organometallics 2015,
34, 4369.
[26] In a similar fashion, we checked that the NMR-scale reaction (C6D6,
298 K) of pyrrolidine (2 equiv) with bis-alkyl complex 4-Ba releases
CH2(SiMe3)2 quantitatively, whereas the evolution of free HN(SiMe3)2
during the reaction with bis-amide 1-Ba cannot be detected spectro-
scopically. See the Supporting Information (Figure S9–S11) for details.
[27] B. Liu, T. Roisnel, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 2013, 19, 2784.
[28] a) O. Michel, K. W. Tçrnroos, C. Maichle-Mçssmer, R. Anwander, Chem.
Eur. J. 2011, 17, 4964; b) O. Michel, K. W. Tçrnroos, C. Maichle-Mçssmer,
R. Anwander, Eur. J. Inorg. Chem. 2012, 44.
[29] B. Liu, T. Roisnel, J.-F. Carpentier, Y. Sarazin, Angew. Chem. Int. Ed. 2012,
51, 4943; Angew. Chem. 2012, 124, 5027.
[30] Under identical conditions, the magnesium complexes did not afford
significant catalytic activity. For instance, [Mg{N(SiMe3)2}2]2 only afforded
2% conversion of the substrates after 15 h reaction at 298 K with [pyr-
rolidine]0 =4.0m and [pyrrolidine]0/[Ph3SiH]0/[Mg]0 =400:400:1.
[31] Under identical conditions to those given in Table 1, entry 6, the bis-si-
lazido Ba complex 5-Ba also proved totally inactive. On the other hand,
the Ba analogue of the pyrrolido complex 9-Sr fully converted the sub-
strates under the same conditions (i.e. those corresponding to Table 1,
entry 8). However, due to the uncertainty as to the identity of this Ba
species (see ref. [21]), we have considered these results in the simplest
qualitative manner.
[32] Dehydropolymerisation reactions involving the catalytic coupling of
(di)amines and di(hydrosilane)s are specific applications and will be re-
ported elsewhere: C. Bellini, C. Orione, J.-F. Carpentier, Y. Sarazin, Angew.
Chem. Int. Ed. 2016, 10.1002/anie.201511342 in press.
[33] For instance, under discriminating conditions ([diamine]0/[silane]0/[1-
Ba]0 =20:40:1, [diamine]0 =0.2m, 298 K, 2.5 h), substrate conversion in
the coupling of N,N’-dimethylethylenediamine with Ph3SiH catalysed by
1-Ba reached 43, 45 and 40% in C6D6, C6D5Cl and C6D6/1,2-C6H4F2, re-
spectively.
[17] a) K. Takaki, T. Kamata, Y. Miura, T. Shishido, K. Takehira, J. Org. Chem.
1999, 64, 3891; b) K. Takaki, K. Komeyama, K. Takehira, Tetrahedron
2003, 59, 10381; c) F. Buch, S. Harder, Organometallics 2007, 26, 5132;
d) W. Xie, H. Hu, C. Cui, Angew. Chem. Int. Ed. 2012, 51, 11141; Angew.
Chem. 2012, 124, 11303.
[34] This disappointing result subsequently led us to assess the feasibility of
the dehydrocoupling between Bn2NH and Ph3SiH. Very low conversions
to Bn2NÀSiPh3 were recorded with 1-Ba (3%) and 4-Ba (7%) after 12 h
at 333 K, with [Bn2NH]0/[Ph3SiH]0/[Ba]0 =20:20:1 and [Bn2NH]0 =0.1m in
C6D6. We are unable to explain the very poor reactivity of these diben-
zyl-substituted amines in CDC reactions other than by arguing that
steric factors may be unfavourable: the formation of the Ba–NBn2 spe-
cies required for catalytic turnover (see computational investigations)
may be sterically prohibited for these congested amines.
[18] J. X. Wang, A. K. Dash, J. C. Berthet, M. Ephritikhine, M. S. Eisen, J. Orga-
nomet. Chem. 2000, 610, 49.
[19] T. Tsuchimoto, Y. Iketani, M. Sekine, Chem. Eur. J. 2012, 18, 9500.
[20] a) J. F. Dunne, S. R. Neal, J. Engelkemier, A. Ellern, A. D. Sadow, J. Am.
Chem. Soc. 2011, 133, 16782; b) M. S. Hill, D. J. Liptrot, D. J. MacDougall,
M. F. Mahon, T. P. Robinson, Chem. Sci. 2013, 4, 4212; c) C. Bellini, J.-F.
Carpentier, S. Tobisch, Y. Sarazin, Angew. Chem. Int. Ed. 2015, 54, 7679;
Angew. Chem. 2015, 127, 7789.
[35] The maximal kinetic isotope effect was calculated according to Equa-
tions (3) and (4):
kSi–D/kSi–H =eÀl
l=(hcvSi–H)/(2kB T)”{1À(mSiH/mSiD
(3)
(4)
1/2
) }
[21] The NMR spectroscopic data were unambiguous regarding the formula-
tion and structure of 9-Sr, which clearly contained one coordinated pyr-
rolidine molecule. On the other hand, we could not clearly assess the
in which h is the Planck constant, c is the speed of light, kB is the Boltz-
mann constant, mSiH and mSiD are the relevant effective masses, and vSi–H
is the wavenumber for the stretching vibration of the NÀH bond in the
FTIR spectrum of the amine recorded at the temperature T. See: P.
Atkins, J. de Paula, Physical Chemistry, 7th ed., Oxford University Press,
Chichester 2002, pp. 888–889. We measured vSi–H =2117 cmÀ1 at T=
290.5 K, giving kSi–H/kSi–D =4.35.
1
exact formulation of its Ba congener 9-Ba: the H NMR spectrum of the
complex obtained suggested the presence of one to two pyrrolidine
molecules, and we were not able to remove the doubt as to its actual
composition. Elemental analysis was not informative, and we failed in
our attempts to obtain X-ray quality crystals (see Figure S8 in the Sup-
porting Information for details).
[36] Arrhenius analysis for the same data set gave Ea =16.2(23) kcalmolÀ1
.
[37] For the values of sigma, see: C. G. Swain, E. C. Lupton, J. Am. Chem. Soc.
1968, 90, 4328.
[38] B. Liu, T. Roisnel, J.-P. GuØgan, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J.
2012, 18, 6289.
[22] R. J. P. Corriu, C. GuØrin, J. Chem. Soc. Chem. Commun. 1980, 168.
[23] The pale-yellow Ca derivative Ca(SiPh3)2·(THF)4 (d29Si =À13.99 ppm), in
which the metal is stabilised by four THF molecules instead of three as
in 5-Sr and 5-Ba, was recently reported: a) V. Leich, T. P. Spaniol, L.
[39] S. Tobisch, Chem. Eur. J. 2014, 20, 8988.
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4582
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[40] B. Liu, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 2012, 18, 13259.
[41] a) S. Datta, M. T. Gamer, P. W. Roesky, Organometallics 2008, 27, 1207;
b) M. Arrowsmith, M. S. Hill, G. Kociok-Kçhn, Organometallics 2009, 28,
1730; c) M. Arrowsmith, M. S. Hill, G. Kociok-Kçhn, Organometallics
2011, 30, 1291; d) J. Jenter, R. Kçppe, P. W. Roesky, Organometallics
2011, 30, 1404; e) M. Arrowsmith, M. R. Crimmin, A. G. M. Barrett, M. S.
Hill, G. Kociok-Kçhn, P. A. Procopiou, Organometallics 2011, 30, 1493;
f) M. Arrowsmith, M. S. Hill, G. Kociok-Kçhn, Organometallics 2014, 33,
206; g) S. Tobisch, Chem. Eur. J. 2015, 21, 6765.
[42] C. Brinkmann, A. G. M. Barrett, M. S. Hill, P. A. Procopiou, S. Reid, Organo-
metallics 2012, 31, 7287.
[43] a) Substrate association and dissociation at the various intermediates is
presumably facile and rapid equilibria are expected. b) Examination by
a linear-transit approach gave no indication that such processes are as-
sociated with a significant enthalpy barrier.
c) R. Ahlrichs, F. Furche, C. Hättig, W. Klopper, M. Sierka, F. Weigend,
TURBOMOLE, version 6.6, University of Karlsruhe, Karlsruhe, Germany,
2014, http://www.turbomole.com.
[53] a) P. A. M. Dirac, Proc. R. Soc. London Ser. A 1929, 123, 714; b) J. C. Slater,
Phys. Rev. 1951, 81, 385; c) S. Vosko, L. Wilk, M. Nusiar, Can. J. Phys.
1980, 58, 1200; d) A. D. Becke, Phys. Rev. A 1988, 38, 3098; e) J. P.
Perdew, Phys. Rev. B 1986, 33, 8822.
[54] a) O. Vahtras, J. Almlçf, M. W. Feyereisen, Chem. Phys. Lett. 1993, 213,
514; b) K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs, Chem.
Phys. Lett. 1995, 242, 652.
[55] a) S. Grimme, J. Anthony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104; b) S. Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32,
1456; c) http://toc.uni-muenster.de/DFTD3/getd3.html.
[56] D. Andrae, U. Häußermann, M. Dolg, H. Stoll, H. Preuß, Theor. Chim.
Acta 1990, 77, 123.
[44] a) The prevalent pyrrolidine bis-amine adduct C3·(A)2 of the catalytically
competent {N^N}Ba{N(CH2)4} pyrrolide complex (with the appropriate
number of substrate, silazane product or THF molecules) was chosen as
the reference for the relative free energies [kcalmolÀ1]. b) Free energies
are given for Ca, Sr and Ba analogues of the relevant key species.
[45] a) L. Perrin, O. Eisenstein, L. Maron, New J. Chem. 2007, 31, 549; b) L.
Perrin, L. Maron, O. Eisenstein, T. D. Tilley, Organometallics 2009, 28,
3767.
[57] a) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297; b) F.
Weigend, Phys. Chem. Chem. Phys. 2006, 8, 1057.
[58] a) A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829; b) K.
Eichkorn, F. Weigend, O. Treutler, R. Ahlrichs, Theor. Chem. Acc. 1997, 97,
119; c) A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571.
[59] a) S. Grimme, J. Comput. Chem. 2006, 27, 1787; b) A. D. Becke, J. Chem.
Phys. 1997, 107, 8554.
[60] P. Plessow, J. Chem. Theory Comput. 2013, 9, 1305.
[46] M. E. Thompson, S. M. Baxter, A. R. Bulls, B. J. Burger, M. C. Nolan, B. D.
Santarsiero, W. E. Schaefer, J. E. Bercaw, J. Am. Chem. Soc. 1987, 109,
203.
[47] a) M. Westerhausen, Inorg. Chem. 1991, 30, 96; b) J. M. Boncella, C. J.
Coston, J. K. Cammack, Polyhedron 1991, 10, 769; c) Y. Sarazin, R. H.
Howard, D. L. Hughes, S. M. Humphrey, M. Bochmann, Dalton Trans.
2006, 340.
[48] N. Romero, S.-C. Ros¸ca, Y. Sarazin, J.-F. Carpentier, L. Vendier, S. Mallet-
Ladeira, C. Dinoi, M. Etienne, Chem. Eur. J. 2015, 21, 4115.
[49] M. R. Crimmin, A. G. M. Barrett, M. S. Hill, D. J. MacDougall, M. F. Mahon,
P. A. Procopiou, Chem. Eur. J. 2008, 14, 11292.
[61] a) A. Klamt, G. Schüürmann, J. Chem. Soc. Perkin Trans. 2 1993, 799;
b) A. Klamt, J. Phys. Chem. 1995, 99, 2224; c) F. Eckert, A. Klamt, AIChE J.
2002, 48, 369; d) A. Klamt, F. Eckert, W. Arlt, Annu. Rev. Chem. Biomol.
Eng. 2010, 1, 101; e) A. Klamt, WIREs Comput. Mol. Sci. 2011, 1, 699.
[62] F. Eckert, A. Klamt, COSMOtherm, Version C3.0, Release 13.01, COSMO-
logic GmbH & Co. KG, Leverkusen, Germany, 2012.
[63] W. Koch, M. C. Holthausen, A Chemist’s Guide to Density Functional
Theory, 2nd ed., Wiley-VCH, Weinheim, 2001.
[64] a) J. A. Keith, R. J. Nielsen, J. Oxgaard, W. A. Goddard, J. Am. Chem. Soc.
2007, 129, 12342; b) R. F. Ribeiro, A. V. Marenich, C. J. Cramer, D. G. Truh-
lar, J. Phys. Chem. B 2011, 115, 14556.
[50] a) G. M. Sheldrick, SHELXS-97, Program for the Determination of Crystal
Structures, University of Gçttingen, Germany, 1997; b) G. M. Sheldrick,
SHELXL-97, Program for the Refinement of Crystal Structures, University
of Gçttingen, Germany, 1997.
[65] Y. Okuno, Chem. Eur. J. 1997, 3, 212.
[66] For further details, see http://www.struked.de.
[51] R. G. Parr, W. Yang, Density-Functional Theory of Atoms and Molecules,
Oxford University Press, Oxford, 1989.
Received: October 27, 2015
[52] a) R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kçlmel, Chem. Phys. Lett.
1989, 162, 165; b) O. Treutler, R. Ahlrichs, J. Chem. Phys. 1995, 102, 346;
Published online on February 11, 2016
Chem. Eur. J. 2016, 22, 4564 – 4583
www.chemeurj.org
4583
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim