The mechanism of borane-amine dehydrocoupling with bifunctional ruthenium catalysts

Article abstract of DOI:10.1021/ja311092c

Borane-amine adducts have received considerable attention, both as vectors for chemical hydrogen storage and as precursors for the synthesis of inorganic materials. Transition metal-catalyzed ammonia-borane (H₃N-BH ₃, AB) dehydrocoupling offers, in principle, the possibility of large gravimetric hydrogen release at high rates and the formation of B-N polymers with well-defined microstructure. Several different homogeneous catalysts were reported in the literature. The current mechanistic picture implies that the release of aminoborane (e.g., Ni carbenes and Shvo's catalyst) results in formation of borazine and 2 equiv of H₂, while 1 equiv of H ₂ and polyaminoborane are obtained with catalysts that also couple the dehydroproducts (e.g., Ir and Rh diphosphine and pincer catalysts). However, in comparison with the rapidly growing number of catalysts, the amount of experimental studies that deal with mechanistic details is still limited. Here, we present a comprehensive experimental and theoretical study about the mechanism of AB dehydrocoupling to polyaminoborane with ruthenium amine/amido catalysts, which exhibit particularly high activity. On the basis of kinetics, trapping experiments, polymer characterization by ¹¹B MQMAS solid-state NMR, spectroscopic experiments with model substrates, and density functional theory (DFT) calculations, we propose for the amine catalyst [Ru(H)₂PMe₃{HN(CH₂CH₂PtBu ₂)₂}] two mechanistically connected catalytic cycles that account for both metal-mediated substrate dehydrogenation to aminoborane and catalyzed polymer enchainment by formal aminoborane insertion into a H-NH ₂BH₃ bond. Kinetic results and polymer characterization also indicate that amido catalyst [Ru(H)PMe₃{N(CH₂CH ₂PtBu₂)₂}] does not undergo the same mechanism as was previously proposed in a theoretical study.

Full text of DOI:10.1021/ja311092c

Subscriber access provided by A.A. Lemieux Library | Seattle University
Article
The Mechanism of Borane-Amine Dehydrocoupling
with Bifunctional Ruthenium Catalysts
Alexander N Marziale, Anja Friedrich, Isabel Klopsch, Markus Drees,
Vinicius R Celinski, Jörn Schmedt auf der Günne, and Sven Schneider
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja311092c • Publication Date (Web): 09 Aug 2013
Downloaded from http://pubs.acs.org on August 10, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 39
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The Mechanism of BoraneꢀAmine Dehydrocoupling
with Bifunctional Ruthenium Catalysts
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
‡
†
¶
†
§
Alexander N. Marziale, Anja Friedrich, Isabel Klopsch, Markus Drees, Vinicius R. Celinski, Jörn
§
,†,‡,¶
Schmedt auf der Günne, and Sven Schneider*
Technische Universität München, Department Chemie, Lichtenbergstr. 4, 85747 Garching b. München,
Germany, LudwigꢀMaximiliansꢀUniversität München, Department Chemie und Biochemie,
Butenandtstr. 5ꢀ13, 81377 München, Germany, FriedrichꢀAlexanderꢀUniversität ErlangenꢀNürnberg,
Department Chemie und Pharmazie, Egerlandstr. 1, 91058 Erlangen, Germany, and GeorgꢀAugustꢀ
Universität Göttingen, Tammannstr. 4, 37077 Göttingen, Germany.
sven.schneider@chemie.uniꢀgoettingen.de
RECEIVED DATE
†
Technische Universität München
§
LudwigꢀMaximiliansꢀUniversität
‡
FriedrichꢀAlexanderꢀUniversität ErlangenꢀNürnberg
¶
GeorgꢀAugustꢀUniversität Göttingen
ACS Paragon Plus Environment
1

Journal of the American Chemical Society
Page 2 of 39
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT: Boraneꢀamine adducts have received considerable attention, both as vectors for chemical
hydrogen storage and as precursors for the synthesis of inorganic materials. Transition metal catalyzed
ammoniaꢀborane (H N–BH , AB) dehydrocoupling offers, in principle the possibility of large
3
3
gravimetric hydrogen release at high rates and the formation of B–N polymers with wellꢀdefined
microstructure. Several different homogeneous catalysts were reported in the literature. The current
mechanistic picture implies that the release of aminoborane (e.g. Ni carbenes and Shvo’s catalyst)
results in formation of borazine and 2 equiv. H while one equiv. H and polyaminoborane are obtained
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
with catalysts that also couple the dehydroproducts (e.g. Ir and Rh diphosphine and pincer catalysts).
However, in comparison with their rapidly growing number, the amount of experimental studies that
deal with mechanistic details is still limited. Here, we present a comprehensive experimental and
theoretical study about the mechanism of AB dehydrocoupling to polyaminoborane with ruthenium
amine/amido catalysts, which exhibit particularly high activity. Based on kinetics, trapping experiments,
1
1
polymer characterization by B MQMAS solidꢀstate NMR, spectroscopic experiments with model
substrates, and density functional theory (DFT) calculations, we propose for the amine catalyst
[Ru(H) PMe {HN(CH CH PtBu ) }] two mechanistically connected catalytic cycles that account both
2
3
2
2
2 2
for metal mediated substrate dehydrogenation to aminoborane and catalyzed polymer enchainment by
formal aminoborane insertion into a H–NH BH bond. Kinetic results and polymer characterization also
2
3
indicate that amido catalyst [Ru(H)PMe {N(CH CH PtBu ) }] does not undergo the same mechanism
3
2
2
2 2
as was previously proposed in a theoretical study.
ACS Paragon Plus Environment
2

Page 3 of 39
Journal of the American Chemical Society
INTRODUCTION
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In recent years, borane amine adducts, particularly parent H NꢀBH (AB), attracted great interest as
3
3
1
H ꢀvectors for chemical hydrogen storage, most recently revived by reports on the regeneration of
2
2
spent fuel. Aside from energy storage applications, borane amine dehydrocoupling is also highly
3
attractive for the synthesis of polymeric B–Nꢀmaterials. The exothermic nature of AB dehydrogenation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
renders thermal degradation the simplest approach. However, the onset of H release is difficult to
2
control, and the polymeric dehydrocoupling products are generally structurally illꢀdefined solids with
3
,4
varying degrees of dehydrogenation and crossꢀlinking. Alternatively, catalysis offers in principle the
possibility of enhanced control over H ꢀrelease kinetics and polymer microstructure. Hence,
2
considerable efforts have been made to develop protocols for the dehydrocoupling of parent AB and of
primary and secondary boraneꢀamine adducts, using Brønsted acids and bases, Lewis acidic main group
5,6,7,8
metal compounds, or frustrated Lewis acidꢀbase pairs as catalysts.
However, most work was
arguably dedicated to the application of a wide range of homogeneous transition metal catalysts, such as
9
10
group 4 metallocenes, group 5ꢀ7 carbonyl and nitrosyl complexes, and particularly late transition
11,12,13,14
metal phosphine, carbene, or olefin complexes.
At this point, detailed mechanistic understanding will be crucial for rational catalyst design and the
optimization of selectivities with respect to the formed B–N materials. However, despite all efforts in
catalyst development mechanistic understanding remains limited. Most information is available for Rh,
n+
Ir, and Ru [M(L) ] ꢀbased precatalysts bearing nonꢀchelating, bulky phosphine or Nꢀheterocyclic
2
carbene ligands (n = 1: M = Rh, Ir; L = PiPr , PiBu , PCy , IMes / n = 0: M = Ru; L = PCy ). The
3
3
3
3
coordination chemistry of these metal fragments with boraneꢀamines or dehydrogenation and
dehydrocoupling products, such as aminoboranes or diborazanes (H B–NR , H(H B–NR ) H; R = H,
2
2
2
2 n
15
Me; n = 1, 2), was examined by several groups. Such model compounds feature B–H interactions with
1
6
17
the metal center and B–H oxidative addition has also been directly observed. In contrast,
experimental information about the N–H activation step is less wellꢀestablished. Overall, experimental
and theoretical mechanistic studies for these catalysts are in favor of pathways with formal oxidation
ACS Paragon Plus Environment
3

Journal of the American Chemical Society
Page 4 of 39
state change upon hydrogen transfer to the metal and aminoborane release, followed by H reductive
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8
elimination. To date, this mechanistic work is mostly limited to the dehydrogenation of Me HN–BH
2
3
(Me AB) which forms linear diborazane Me HN–BH –NMe –BH as an intermediate and soluble
2
2
2
2
3
cycloꢀdiaminoborane (Me N–BH ) as the final product. In comparison, dehydrocoupling of AB or
2
2 2
MeH N–BH (MeAB) and AB/MeAB coꢀdehydropolymerization with several catalysts was shown to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
give linear and/or cyclic polyaminoboranes (RHN–BH ) (R = H, Me) of medium to high molecular
2
n
13c,13d,19
weight,
often with low solubility, thereby hampering examination. However, these results and
2
0
the recent synthesis of parent H N–BH –NH –BH fuel speculation about the relevance of linear
3
2
2
3
oligoborazane intermediates for parent AB dehydrocoupling, as well.
Scheme 1. Proposed pathways for the formation of AB dehydrocoupling products with Goldberg and
Heinekey’s Irꢀcatalyst and Baker’s Niꢀcatalyst, respectively.
An elegant experimental study for this catalyst type suggested, not only for dehydrogenation but also
18c
the B–N coupling step to be a metal mediated process.
Also the kinetics of AB and MeAB
dehydropolymerization with a related Ir pincer catalyst support a chain growth mechanism, hence
ACS Paragon Plus Environment
4

Page 5 of 39
Journal of the American Chemical Society
1
3d
indicating catalyzed B–N bond formation.
In contrast, metal centered B–N bond formation was
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
1
considered to be unlikely in a theoretical study. Baker, Dixon, and coꢀworkers proposed that the
1
4b
selectivity of AB dehydropolymerization depends on the fate of intermediate aminoborane: On one
14a,22
hand, the release of more than two equivalents of H , e.g. by a Ni catalyst,
was attributed to the
2
formation of free aminoborane H N–BH , a highly reactive molecule which readily oligomerizes via
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
cyclic dehydrotrimerization product H B–H N–B N H (A), ultimately giving polyborazylene after
3
2
2
2
7
23,24,25
further dehydrocoupling steps (Scheme 1).
On the other hand, catalyst [IrH (POCOP)] (POCOP =
2
C H ꢀ1,3ꢀ(OPtBu ) ), produces polyaminoborane and one equiv. H . The absence of A and aminoborane
6
3
2 2
2
trapping products was rationalized with the assumption that H N–BH is bound during catalysis
2
2
(
Scheme 1), also explaining polymer chain growth observed with this system. This model is of
fundamental importance for both hydrogen storage and B–N polymer formation:
1
. It provides a rationale for the strongly differing yields in H2.
2
. A metal centered B–N coupling process enables control over the polymer microstructure by catalyst
design.
Scheme 2. Fagnou’s mechanism proposed for bifunctional AB dehydrogenation.
Particularly rapid dehydrogenation of AB and MeAB was reported for ruthenium and most recently
1
1a,13d,26,27,28,29,30
also iron precatalysts bearing potentially cooperative ligands.
For example, Williams
29
and coꢀworkers used Shvo’s catalyst for rapid release of 2 equiv. H and formation of borazine. The
2
zero order kinetics in AB for hydrogen evolution were rationalized by turnꢀover limiting H elimination
2
ACS Paragon Plus Environment
5

Journal of the American Chemical Society
Page 6 of 39
from the catalyst after rapid outerꢀsphere transfer of hydrogen from the substrate and release of
aminoborane. However, the product of aminoborane dehydrooligomerization, borazine, serves as a
reversible catalyst poison slowing down catalysis, which could be circumvented with a related second
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
30
generation catalyst. In comparison, Fagnou and coꢀworkers reported extraordinary activities in AB
dehydrogenation using catalyst [Ru(Cl) (iPr PCH CH NH ) ] and 30 equiv. of base (KOtBu) as
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
2
2
2 2
2
6a
activator.
assumption
[Ru(H)(iPr PCH CH NH)(iPr PCH CH NH )] is formed upon activation (Scheme 2). The proposed
Supported by DFT calculations, the authors proposed a catalytic cycle based on the
that the catalytically active hydrido amido complex
2
2
2
2
2
2
2
cycle can be broken down into two parts: 1. Hydrogen transfer from the substrate to the amido catalyst
in two steps with release of free H N–BH and formation of dihydrido amine complex
2
2
[Ru(H) (iPr PCH CH NH ) ]. 2. Subsequent turnꢀover limiting H elimination from the resulting
2
2
2
2
2 2
2
hydrido amine species.
Independently, our group used amido catalyst [RuH(PMe )(PNP)] (1, PNP = N(CH CH PiPr ) ,
3
2
2
2 2
Scheme 4) for AB and Me AB dehydrogenation (1 equiv. H ) with comparable rates as Fagnou’s
2
2
26b,31
catalyst.
A bifunctional mechanism in analogy to Fagnou’s proposal was considered and indicated
by the kinetic isotope effects found for deuterated AB. However, the first order rate dependence on
substrate excludes intramolecular H2 elimination from the corresponding amine complex
[
Ru(H) (PMe )(HPNP)] (2, Scheme 4) to be turnꢀover limiting, as suggested by the computations for
2 3
Fagnou’s system. Furthermore, experimental examination of the H elimination/addition equilibrium of
2
2
6b,32
1
and 2 indicated that this reaction is too slow to be relevant for AB dehydrogenation.
However,
3
3
Brønsted acid catalysis for H elimination was observed and attributed to a protonꢀshuttle mechanism.
2
Accordingly, Nguyen and coꢀworkers proposed a mechanism for AB dehydrogenation with our system
34
in a computational study with strongly truncated model catalysts which picks up this idea (Scheme 3):
H
Proton transfer to the hydride ligand of 2 exhibits low barriers by proton shuttling with the substrate.
Further, they proposed attack of hydridic B–H at the PNP amine proton with release of H and
2
aminoborane. Another equivalent of dihydrogen can then be eliminated from the resulting dihydrogen
ACS Paragon Plus Environment
6

Page 7 of 39
Journal of the American Chemical Society
H
H
H
complex without a barrier to give amido complex 1 . Although the regeneration of 2 from 1 was not
explicitly calculated to complete the catalytic cycle, transfer of hydrogen from the substrate, as in
Fagnou’s mechanism, seems to be a viable path at low energetic cost.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Nguyen’s mechanism proposed for bifunctional AB dehydrogenation.
In this paper, we report an experimental and computational study to clarify the mechanism of AB
dehydrogenation with bifunctional ruthenium amine and amido catalysts. We will show that some of our
experimental results are incongruent with either Fagnou’s or Nguyen’s proposals, leading to a modified
mechanism which accounts for both dehydrogenation and B–N coupling.
RESULTS
In the first section, we describe the H release kinetics for the potentially bifunctional amido and
2
amine precatalysts [RuH(PMe )(PNP)] (1, section 1.1) and [Ru(H) (PMe )(HPNP)] (2, section 1.2)
3
2
3
including H/D kinetic isotope effects for the four substrates H N–BH , H N–BD , D N–BH , and D N–
3
3
3
3
3
3
3
BD . Qualitative comparison with the structurally related (yet inherently not bifunctional) precatalyst
3
[
Ru(H) (PMe )(MePNP)] (3), is given in section 1.3. In section 2 the kinetic studies are complemented
2 3
by stoichiometric experiments to examine substrate/catalyst interaction (section 2.1) and possible
ACS Paragon Plus Environment
7

Journal of the American Chemical Society
Page 8 of 39
pathways of B–N coupling, i.e. aminoborane headꢀtoꢀtail coupling vs. intermediate release of
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
aminoborane (sections 2.1 and 2.2). These mechanistic examinations in combination with the structural
characterization of the aminoborane polymer (section 3) afford a mechanistic proposal, which differs
26a,34
from the previously proposed mechanisms,
and is evaluated computationally in section 4.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 4. Ruthenium PNP complexes discussed throughout the manuscript.
1
. Kinetic studies. 1.1 Dehydrocoupling of AB with [RuH(PMe )(PNP)] (1). Addition of 1 to a
3
solution of AB results in rapid and vigorous evolution of H and precipitation of a colorless
2
dehydrocoupling product (Scheme 5). Initial results on AB dehydrogenation using precatalyst 1 with
ꢀ1
turnover numbers (TONs) up to 8300 and turnover frequencies (TOF) around 20 s were reported in a
2
6b
preceding communication. In the present study, the substrate ([AB] = 0.25 – 2.5 M) and catalyst
0
loadings (0.01 – 0.1 molꢀ%) were varied over a much wider range, with TONs ≥ 10000 at max. [AB]0 /
[
1] ratios (Supporting Information Figure S1). H release up to slightly more than one equiv. [AB]
2 0
suggests the formation of mostly polyaminoborane, i.e. in agreement with the polymer characterization
3
5
11
(
section 3). After release of around 1 equiv. H , the substrate is fully consumed (solution B NMR)
2
and minor amounts of A and borazine are observed, which correlate well with the excess H yield
2
ACS Paragon Plus Environment
8

Page 9 of 39
Journal of the American Chemical Society
(
internal standard NaBPh ), indicating decent error margins for this setup. Higher catalyst loadings
4
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
result in the formation of more borazine at identical initial substrate concentrations [AB] . At high
0
1
31
substrate conversions, catalyst deactivation product 4 (Scheme 4) is observed by H and P NMR
31
spectroscopy.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 5. Dehydrocoupling of AB with catalyst 1.
11
B NMR monitoring of the reaction in a sealed (JꢀYoung) NMR tube reveals considerably slower
rates than volumetric (H ) derivation. Hence, reaction rates were measured eudiometrically at room
2
temperature and the rate law was derived from initial rates to avoid inhibition due to H buildꢀup and
2
diffusion problems owing to product precipitation and to minimize error attributable to borazine
formation at a later stage of the reaction. The data (Figure S1, Supporting Information) are in agreement
with a simple rate law that is first order both in [AB] and [1]:
r = k [1] [AB]
(Eq. 1)
H2
obs1
0
Accordingly, first order plots (ln[AB] vs. time) were also linear over two halfꢀlifes for a wide range of
catalyst and initial substrate concentrations (Figure S1, Supporting Information). Low catalyst loadings
ꢀ1 ꢀ1
(0.05 molꢀ%) gave rate constants around kobs1 = 15 ± 2 M s (Table S1), in excellent agreement with
ꢀ1 ꢀ1
ꢀ1 ꢀ1
the initial rate estimate (kobs1 = 13 M s ). Higher rate constants up to 32 M s were obtained for higher
catalyst loadings (0.25 molꢀ% and 0.1 molꢀ%) (Table S1), coinciding with larger amounts of borazine
being observed.
ACS Paragon Plus Environment
9

Journal of the American Chemical Society
Page 10 of 39
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Kinetic isotope effects for catalysts 1 and 2. Top: First order plot of H N–BH and H N–BD
3
3
3
3
(
[AB] = 0.54 M; [1] = 0.0005 M) dehydrogenation. Center: Zero order plot of D N–BH and D N–
0 0 3 3 3
BD ([AB] = 0.54 M; [1] = 0.0005 M) dehydrogenation. Bottom: First order plot of H N–BH , D N–
3
0
0
3
3
3
BH , D N–BD (all: [AB] = 0.5 M; [2] = 0.0005 M), and H N–BD ([AB] = 1.0 M; [2] = 0.001 M)
3
3
3
0
0
3
3
0
0
dehydrogenation.
ACS Paragon Plus Environment
10

Page 11 of 39
Journal of the American Chemical Society
In our preceding communication, kinetic isotope effects (KIEs) for this reaction were estimated from
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
firstꢀorder plots using Bꢀ (H N–BD ) and Nꢀterminally (D N–BH ) deuterated substrates and fully
3
3
3
3
2
6b
deuterated D N–BD , respectively. While the experimental results were reproduced in the present
3
3
study, reevaluation of the data indicated a more complex picture. As for parent H N–BH (0.54 M; 0.1
3
3
molꢀ% [1]), the rate law for isotopomer H N–BD exhibits firstꢀorder rate dependence on substrate over
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
more than 2 halfꢀlifes after a short initiation period with a KIE k(H N–BH ) / k(H N–BD ) = 1.8 (Figure
3
3
3
3
36
1
, top). However, for the Nꢀdeuterated D N–BH and D N–BD , [AB] vs. time plots are linear over
3 3 3 3
two halfꢀlifes indicating zero order dependence of the rate law on substrate concentration (Figure 1,
center), and hence a different turnover limiting step for these isotopomers. Therefore, KIE’s for these
isotopomers referenced to unlabeled H N–BH could not be derived.
3
3
1.2 Dehydrocoupling of AB with [Ru(H) (PMe )(HPNP)] (2). Both Fagnou’s (Scheme 2) and
2
3
Nguyen’s (Scheme 3) mechanistic proposals involve both dihydrido amine and the hydrido amido
complexes as active catalyst within the catalytic cycles. Hence, complex 2 was employed as catalyst
(Scheme 6) over a wide range of initial substrate concentrations [AB] (0.25 – 1.5 M) and catalyst
0
32b
loadings [Ru] (0.02 ꢀ 0.1 molꢀ%) (Figure S2). Since catalyst 2 is not isolable owing to facile H loss,
0
2
a stock solution was freshly prepared prior to every kinetic run by in situ hydrogenation of 1. As for 1,
initial rates of H evolution (Figure S2) are in agreement with first order dependence of the rate law in
2
catalyst and substrate (Eq. 2):
r = k [2] [AB]
(Eq. 2)
H2
obs2
0
Scheme 6. Dehydrocoupling of AB with precatalyst 2.
ꢀ1 ꢀ1
ꢀ1 ꢀ1
A rate constant kobs2 = 24 ± 2 M s was estimated, i.e. higher than for precatalyst 1 (k
= 13 M s ).
obs1
First order plots (ln[AB] vs. time) were also linear over approximately two halfꢀlifes for an even larger
ACS Paragon Plus Environment
11

Journal of the American Chemical Society
Page 12 of 39
range of catalyst and substrate initial concentrations, as compared with precatalyst 1 (Supporting
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀ1 ꢀ1
Information). The rate constant extracted from these plots (k = 22 M s , Figure S2) is in excellent
agreement with the initial rate estimate indicating that the simple rate law (Eq. 2) is applicable over
large parts of the catalytic reaction.
As for precatalyst 1, KIEs were derived with isotopically labeled substrates. However, with 2 all
isomers (H N–BH , H N–BD , D N–BH , D N–BD ) exhibit first order rate dependence on [AB]
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
3
3
3
3
3
3
(
Figure 1 bottom). Also in contrast with 1, Bꢀterminal deuteration resulted in no KIE (k(H N–BH ) /
3 3
k(H N–BD ) = 1.0; k(D N–BH ) / k(D N–BD ) = 1.1). Large KIEs were obtained for the Nꢀterminally
3
3
3
3
3
3
deuterated isotopomers (k(H N–BH ) / k(D N–BH ) = 5.3; k(H N–BH ) / k(D N–BD ) = 5.9). These
3
3
3
3
3
3
3
3
results indicate N–H bond cleavage, yet no B–H activation in the turnover limiting step and will be
analyzed in combination with the computational results in the discussion section.
1
.3 Dehydrocoupling of AB with [Ru(H) (PMe )(MePNP)] (3). To further probe for the potential
2 3
[
33]
relevance of ligand cooperativity, tertiary amine complex [Ru(H) (PMe )(MePNP)] (3)
was
2
3
employed as a reference catalyst (Scheme 7 and Figure S3). Methylation of the pincer ligand nitrogen
atom prohibits N–H bonding. Thus, a bifunctional mechanism cannot be operative for this catalyst.
Addition of 3 to a solution of AB in THF at room temperature results in evolution of H and
2
precipitation of the polymeric dehydrocoupling product D. As for catalysts 1 and 2, high turnover
ꢀ4
numbers (TON ≥ 10000 at [AB] = 2.5 M and [3] = 1 ꢁ 10 M) were observed. However, the rate of H
0
0
2
evolution is considerably lower by almost two orders of magnitude (Figure 2), as compared with
compounds 1 and 2. Initial rates for a wide range of substrate concentrations ([AB] = 0.25 M – 2.5 M)
0
and catalyst loadings (0.05 molꢀ%, 0.10 molꢀ%, 0.25 molꢀ%) also indicate first order dependence of the
rate law in both catalyst and substrate (Figure S3):
r = k [3] [AB]
(Eq. 3)
H2
obs3
0
ACS Paragon Plus Environment
12

Page 13 of 39
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 7. Dehydrocoupling of AB with precatalyst 3.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
1 ꢀ1
The rate constant derived by the initial rate method (kobs3 = 0.6 M s ) is slightly smaller as compared
ꢀ1 ꢀ1
with the rate constant obtained from firstꢀorder plots (0.9 ± 0.3 M s , Figure S3), yet within statistical
3
7,38
error.
dehydrocoupling mechanism does not substantially change over the course of the reaction.
. Stoichiometric examinations. 2.1 Reactions with model substrates. The reactions of catalysts 2
and 3 with model substrates H N–BEt and Me N–BH were investigated by NMR spectroscopy.
The firstꢀorder plots generally showed good linearity over two halfꢀlives, suggesting that the
2
3
3
3
3
Blocking of the Bꢀ or Nꢀtermini, respectively, by full substitution enabled the examination of initial
substrate interactions with the catalyst without dehydrocoupling. An equimolar mixture of 2 or 3 with
1
31
11
1
Me N–BH did not reveal any reaction on the NMR timescale, as judged by H, P, B, and H 2D
3
3
EXSY NMR spectroscopy. In contrast, a mixture of 2 and H N–BEt (1.5 equivalents in THF) exhibits a
3
3
small equilibrium concentration (< 10 %) of a new compound. This compound could not be isolated, but
was assigned to borane amido complex 5 (Scheme 8). Its spectroscopic data are in agreement with the
31
meridional arrangement of three phosphine ligands ( P NMR), a single hydride ligand (–18.4 ppm),
which is in transꢀposition to a boraneamido and cisꢀ with respect to the phosphine ligands, and with
–
1
further signals assignable to a H N–BEt ligand. Most importantly, the H 2D EXSY NMR spectrum of
2
3
this mixture reveals stereoselective exchange of the H N–BEt N–H protons with the hydride ligand of 2
3
3
that is in synꢀposition to the pincer N–H proton (Figure S5). Quantification of the exchange rates was
not possible, due to peak overlaps. However, this result is in agreement with an exchange mechanism
via transient dihydrogen complex 6a (Scheme 8), which can form the minor equilibrium concentration
of 5a upon loss of H . The stereoselectivity of the N–H/Ru–H exchange is indicative of the stabilization
2
33
of 6a by hydrogen bonding, as previously proposed for proton exchange of 2 with water. Similarly, for
ACS Paragon Plus Environment
13

Journal of the American Chemical Society
Page 14 of 39
31
1
a mixture of H N–BEt and 3 new NMR signals in the P and H (–19.7 ppm) NMR spectra were
3
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
assigned to the formation of around 10% of an amido complex [Ru(NH BEt )(H)(PMe )(MePNP)].
2
3
3
However, no N–H/Ru–H exchange was detected on the NMR timescale of EXSY NMR experiments
max
(
τm
= 1000 ms). Hence, hydride exchange with the substrate protons must proceed at considerably
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
slower rates as compared with complex 2.
Et B
3
H
NH2
P Pr
2
i
i
P Pr
2
H
H
Et B NH3
3
N
Ru
PMe
3
N
Ru
PMe
3
H
2
THF
P
Pr
P
Pr
i
H
i
H
2
2
2
5
BEt3
H N
2
H
H
i
P Pr
H
2
N
Ru
PMe3
P
Pr
i
H
2
6a
Scheme 8. Reaction of 2 with H N–BEt and proposed mechanism via dihydrogen complex 6a. Colors
3
3
1
indicate selected chemical exchange cross peaks observed by H 2D EXSY NMR.
A possible pathway for the formation of linear products, such as polyaminoboranes H N–(H N–
3
2
BH ) –BH , could be the direct headꢀtoꢀtail dehydrocoupling of two boraneꢀamine substrate molecules.
2
n
3
Hence, Me N–BH and H N–BEt were used as model substrates within a crossꢀcoupling experiment in
3
3
3
3
the presence of catalyst 2 (1.0 molꢀ%). Over several hours at room temperature no coupling reaction
1
1
was observed by B NMR (Scheme 9), indicating, that proton and hydride transfer from the same
substrate molecule to the catalyst is required. Furthermore, substrate metathesis by B–N bond cleavage
can be excluded.
H
i
P Pr
2
H
N Ru
P
PMe3
i
Pr H
2
2
Et B-NH3 + H B-NMe3
Et B-NH -BH NMe
3 2 2 3
3
3
Scheme 9. Crossꢀcoupling experiment of Me N–BH and H N–BEt in the presence of 2.
3
3
3
3
ACS Paragon Plus Environment
14

Page 15 of 39
Journal of the American Chemical Society
2
.2 Aminoborane trapping experiments. As a result from the crossꢀcoupling experiment, H seems
2
1
2
to be transferred from the same substrate molecule to the catalyst. Hence, the release of free
aminoborane H N–BH within the mechanism was probed by trapping experiments with cyclohexene,
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
14b
as introduced by Baker and coꢀworkers. AB dehydrocoupling with precatalyst 1 in the presence of
1
1
cyclohexene (21 equivalents with respect to AB) showed no olefin hydroboration products ( B NMR)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
within the timescale of the dehydrocoupling reaction (full substrate consumption by B NMR).
Importantly, only small amounts of polyborazylene and borazine were detected. However, after 20 h at
11
room temperature, rising amounts of borazine and trapping product H N–B(C H ) ( B NMR: δ = 48.4
2
6
11 2
39
ppm) were observed, indicating rearrangement processes, such as slow degradation of the initially
formed, insoluble polyaminoborane. Similarly, trapping experiments with catalyst 2 gave besides
1
1
insoluble polyaminoborane some polyborazylene and borazine by B NMR. No trapping products from
cyclohexene hydroboration were initially observed, but small amounts were detected 24 h after
completion of the dehydrogenation, as similarly found for 1. These results for catalysts 1 and 2 contrast
with trapping experiments for catalyst 3. In this case, not only borazine and small amounts of A are
observed in solution, but also considerable amounts of H N–B(C H ) within the timescale of the
2
6
11 2
dehydrogenation reaction (Figure S4), indicative of the release of free aminoborane.
1
1
Figure 2. B MAS NMR sheared tripleꢀquantum filtered MQMAS spectrum of polymer C.
ACS Paragon Plus Environment
15

Journal of the American Chemical Society
Page 16 of 39
3
. Polymer Characterization. The white, insoluble products resulting from AB dehydrocoupling
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
were further examined. Characterization of B by IR and solidꢀstate B MAS spectroscopy was reported
2
6b
in a preceding communication. Combustion analysis and IR spectra of C and D are in agreement with
13c,13d,26b
the formation of the same product as obtained with catalyst 1 or with an Ir(POCOP) catalyst,
assignable to polyaminoborane H N–(BH –NH ) –BH , rather than the originally proposed cyclic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
2
2 n
3
1
3a
pentaaminoborane (H N–BH ) . As was previously found, some solvent adheres to the polymer even
2
2 5
13c,d
after extended drying in vacuo, which was also observed with other catalysts.
For example,
elemental analysis of D supports the formation of H N–(BH –NH ) –BH ꢁ 0.6 THF, as similarly
3
2
2 19
3
1
3d
reported for AB dehydrocoupling with an Ir(POCOP) catalyst by Staubitz et al. (n ≈ 20). In this study,
1
1
B MultipleꢀQuantum MagicꢀAngleꢀSpinning (MQMAS) NMR spectroscopy was emphasized to
determine the isotropic chemical shift and differentiate between true resonances and secondꢀorder
11
quadrupole line shapes. The B MQMAS NMR spectrum of B (Figure S7) features two signals, one of
higher intensity at δiso = –10.7 ppm (secondꢀorder quadrupolar effect (SOQE) = 1.5 MHz) and one of
lower intensity at δiso = –21.0 ppm (SOQE = 1.4 MHz). These peaks indicate two different boron
environments present in the polymer, assignable to linear main chain BH groups and terminal BH end
2
3
1
1
40
groups, respectively, by comparison with soluble boranes, such as AB (
δ
( B) = –23 ppm) or H N–
3
1
1
20
11
BH –NH –BH (
δ
( B) = –11.6 (BH ) and –22.8 ppm (BH )). While the B spectrum of B
2
2
3
2
3
4
1
considerably differs from early reports of polyaminoborane, it is almost identical with the
13d
11
polyaminoborane prepared by Ir(POCOP)ꢀcatalyzed AB dehydrocoupling.
MQMAS NMR spectrum of polymer C (Figure 2) exhibits a low intensity signal at δiso = –21.0 ppm
BH , SOQE = 1.3 MHz) and the main signal at δiso = –11.1 ppm (BH , SOQE = 1.5 MHz). In addition,
Similarly, the
B
(
3
2
a small signal at δiso = 1.0 ppm (SOQE < 0.4 MHz) is observed. This signal, yet even weaker, is also
visible in polyaminoborane obtained with the Ir(POCOP) catalyst, but absent in B. Owing to its
1
1
42
chemical shift (e.g. Li[B(NHMe)4]:
δ
( B)
= 0.2 ppm) and small SOQE, which indicates a highly
THF
ACS Paragon Plus Environment
16

Page 17 of 39
Journal of the American Chemical Society
symmetric (tetrahedral) environment around the boron atom, this peak is assigned to B(NH ) moieties
2
4
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
within the polymer indicating some crosslinking of the chains.
4. DFT computations. 4.1 Dehydrogenation with model catalyst 2 . Our mechanistic proposal for
Me
Me
Me
AB dehydrocoupling with PMe ꢀtruncated model catalysts 2 and 3 was evaluated using density
2
functional theory (DFT, B3LYP/6ꢀ31+G**). The overall dehydrogenation reaction of AB to H and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0
aminoborane was calculated to be exergonic by ꢀG = –16.5 kcal/mol (Scheme 10), only slightly larger
4
3
than previously reported results at a much higher level of theory (–12.4 kcal/mol; MP4/++3df).
Me
Scheme 10. DFT results for the proposed mechanism of AB dehydrogenation, catalyzed by 2 (black
Me
line) and 3 (red line), respectively.
Me
The calculated mechanism for catalyst 2 was guided by the experimental results presented above.
Me
Me
Initial interaction of 2 with the Nꢀterminal protons of AB leads to dihydrogen complex 6 and is
0
endergonic by ꢀG = +6.5 kcal/mol. Formation of the H ligand (D(H–H) = 0.80 Å) is accompanied by
2
PNP
a hydrogen bridge within the ion pair (D(H BH N–H) = 1.70 Å; D(N –H) = 1.09 Å). This
3
2
ACS Paragon Plus Environment
17

Journal of the American Chemical Society
Page 18 of 39
stabilization results in an energetically considerably more favorable pathway as compared with catalyst
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Me
H
3
(see below). Formation of the heavily truncated model complex 6 was calculated by Nguyen et al.
0
to be even exergonic by ꢀG = –6.2 kcal/mol. However, such strong stabilization is not in agreement
3
4
with our EXSY NMR studies for the real system (section 2.1). Transition state TS is late on the
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
reaction coordinate according to bond lengths (D(H BH N–H) = 1.82 Å; D(H–H) = 0.84 Å).
3
2
Me
Me
Concerted hydrogen loss and borane attack within 6 via TS gives cyclic aminoborane adduct 7 .
2
According to bond distances, the bridging hydride ligand is firmly bound to both the aminoborane
(D(B–HRu) = 1.31 Å) and the metal center (D(Ru–HB) = 1.83 Å), while the bridging amine proton is
PNP
PNP
only weakly hydrogen bound to the aminoborane (D(N –H) = 1.05 Å; D(H BH NꢁꢁꢁꢁHN ) = 1.83
2
2
Å). However, both N and B atoms of the aminoborane are strongly pyramidalized. H release is
2
0
‡
irreversible (ꢀG = –22.2 kcal/mol) and also accompanied by a low kinetic barrier (ꢀG (TS ) = 6.0
2
kcal/mol). In the transition state (TS ), the H ligand is almost completely cleaved from the metal center
2
2
(D(Ru–H) = 2.80 and 2.81 Å; D(H–H) = 0.75 Å) while the approach of the borane moiety is only
completed to a small extend (D(Ru–HB) = 3.16 Å). H ꢀloss represents the highest overall point on the
2
potential Gibbs energy surface within the catalytic cycle. Hence, this step and the preꢀequilibrium of H2ꢀ
complex formation should be turnꢀover limiting. Isotope effects for substrate H/D isotopomers were also
computed and will be compared with experimental values in the discussion section.
Scheme 11. DFT results for the proposed mechanism of catalyzed B–N coupling.
ACS Paragon Plus Environment
18

Page 19 of 39
Journal of the American Chemical Society
An alternative pathway to concerted H ꢀloss and borane attack via TS would be a stepwise
2
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Me
33
mechanism. In fact, a pathway through dihydrogen amido complex 8 (Scheme 4), which loses H
2
Me
and is subsequently attacked by AB was located and the minimum barriers for the two sequences 2
+
Me
Me
Me
Me
Me
Me
AB → 8 + AB → 1 + H + AB → 7 + H (stepwise) and 2 + AB → 6 → 7 + H2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(concerted) are within a few kcal/mol and therefore indistinguishable by DFT. However, with the large
Me
Me
Me
excess of substrate the equilibrium 6
8 + AB will be shifted towards 6 , in favor of the concerted
mechanism (Scheme 10). Furthermore, our kinetic studies suggest, that 1 and 2 do not go through the
H
H
same catalytic cycle. In comparison, Nguyen and coꢀworkers located the transition state for 6 → 8 +
‡
H + H B–NH in their alternative pathway at ꢀG = 16.9 kcal/mol (Scheme 3).
2
2
2
Me
Me
‡
Finally, aminoborane elimination from 7 restores catalyst 2 with a small barrier (ꢀG (TS ) = 8.7
3
kcal/mol). The long bridging B–H (2.57 Å) and N–H (2.45 Å) distances and planarity of the H NBH
2
2
Me
moiety are indicative of a late transition state TS . In contrast to catalyst 3 (see below), this reaction is
3
0
almost thermoneutral (ꢀG = 0.8 kcal/mol), suggesting an equilibrium for aminoborane release.
However, a low energy pathway for aminoborane coupling with boraneꢀamines to linear products was
also found (see below). Hence, a low steady state concentration should arise for H NBH .
2
2
Me
4.2 B–N coupling with model catalyst 2
Previous studies suggested that uncatalyzed
1
4b
oligomerization of aminoborane results in cyclic products. Analysis of the dehydrocoupling products
Me
Me
with catalysts 1 and 2 supports the formation of mainly linear polyaminoborane with minor crossꢀ
linking. The results of the crossꢀcoupling experiments (Section 2.1) rule out the direct dehydrogenative
headꢀtoꢀtailꢀcoupling of two substrate molecules. An alternative pathway is offered by coupling of
aminoborane, the initial product of AB dehydrogenation, with AB (or higher, linear boraneꢀamine
adducts). A mechanism with low activation barriers was found on this route (Scheme 11). In fact, the
Me
same intermediate of the dehydrogenation cycle that results from substrate Nꢀterminal activation (6 ) is
attacked by aminoborane leading to the net insertion into an N–H bond. Formation of linear H B–H N–
3
2
‡
H B–NH exhibits an almost identical barrier as H ꢀloss (ꢀꢀG = 0.3 kcal/mol), providing a model for
2
3
2
ACS Paragon Plus Environment
19

Journal of the American Chemical Society
Page 20 of 39
energetically competitive polyaminoborane enchainment. The observation of aminoborane trapping
product Cy B–NH several hours after consumption of the substrate can also be rationalized: According
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
2
4
to our calculations (and others’ at higher level of theory), aminoborane insertion is almost
thermoneutral, rendering this reaction reversible and therefore mainly driven by precipitation of the
polyaminoborane from solution.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The large excess of AB at catalytic conditions raises the question whether direct reaction of AB with
Me
7
is possible without intermediate release of aminoborane. Several pathways were computed, such as
Me
Nꢀterminal attack of AB at the bridging hydride or at the bridging proton of 6 . However, all efforts
resulted in considerably higher barriers as compared with the mechanism outlined in Scheme 12. A
strong energetic penalty seems to arise from the cleavage of the hydrogen bond without concerted B–H
cleavage and planarization to aminoborane, and vice versa. Although we do not want to fully rule out
the possibility of B–N coupling without intermediate aminoborane release, because of the lack of kinetic
data, the current mechanistic model (Scheme 11) represents an upper rate limit and is in agreement with
turnꢀover limiting hydrogen release.
Me
4
.3 Dehydrogenation with model catalyst 3 . The cooperative effect of the PNP ligand was further
Me
examined by computation of the same mechanistic pathway for catalyst 3 (Scheme 10). Both the
transition states and the dihydrogen complex intermediate that lead to hydrogen elimination are about 10
Me
kcal/mol at higher energy compared with 2 . This result is attributed to the stabilization of the ion pair
in case of 2 by hydrogen bonding with the PNP amine functional group. Hence, this effect is likely to be
somewhat overemphasized within the computations, as hydrogen bonding with other substrate
molecules under catalytic conditions possibly stabilizes this pathway. However, the result is in
qualitative agreement with the experimental observation that 3 catalyzes AB dehydrogenation with rates
Me
almost two orders of magnitude smaller than 2. As for 6 , H release is irreversible and accompanied
2
Me
by formation of a hydride bridged adduct with aminoborane (10 : D(B–HRu) = 1.35 Å; D(Ru–HB) =
1
.83 Å). Owing to the terminal bonding, dissociation of aminoborane is associated with a minute barrier
‡
(
ꢀG = 0.8 kcal/mol) to complete the catalytic cycle.
ACS Paragon Plus Environment
20

Page 21 of 39
Journal of the American Chemical Society
DISCUSSION
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The kinetic, spectroscopic, and computational results for the rapid AB dehydropolymerization to
mainly linear polyaminoborane with our bifunctional ruthenium amine catalyst allow for the
formulation of a mechanistic model. Solely based on DFT calculations, two different mechanisms for
2
6a
34
this type of catalysts had been previously proposed by Fagnou and by Nguyen, respectively.
However, both mechanistic proposals exhibit features which are not in agreement with some of our
experimental results:
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1. In Fagnou’s mechanism (Scheme 2) hydrogen transfer from the substrate to the amido catalyst
species was calculated to be thermodynamically downhill, followed by rate determining, intramolecular
H elimination. Hence, when assuming irreversible loss of gaseous H in an open system, a rate law
2
2
2
9
which is zero order in AB should result, as for Shvo’s catalyst, but first order dependence was found
with our catalysts 1 and 2.
2
. Nguyen’s mechanism (Scheme 3) supports an equilibrium of 2 and AB with an energetically low
lying dihydrogen complex. While such an equilibrium was in fact found with model substrate Et B–NH
3
3
(
Section 2.1) the putative dihydrogen complex 6a is not detected by NMR, suggesting that it should be
more than 10 kJ/mol higher in energy.
. Both mechanisms propose the involvement of the respective amido and amine complexes as active
3
species within the same catalytic cycle. However, despite very similar turnꢀover frequencies starting
from catalysts 1 or 2, respectively, we found considerably different kinetics for the deuterated
isotopomers and different degrees of crossꢀlinking in the resulting polyaminoboranes. These
observations suggest that catalysts 1 and 2 do not catalyze AB dehydrogenation through the same
mechanism.
On the other hand, some of the features within Fagnou’s and Nguyen’s mechanisms could be
confirmed:
ACS Paragon Plus Environment
21

Journal of the American Chemical Society
Page 22 of 39
1
. Both mechanisms include the involvement of the amine functional groups in the sense of
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
bifunctional catalysis. Accordingly, the use of a tertiary amine ligand (catalyst 3) results in rates that are
almost two order of magnitude slower as compared with catalysts 1 and 2.
2
. Nguyen suggested, in contrast to Fagnou’s mechansim, that transfer of a proton from the substrate
Nꢀterminus to a hydride ligand of 2 is the initial step of the catalytic cycle. We had previously
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
demonstrated hydride protonation of 2 by other Brønsted acids with EXSY NMR spectroscopy. This
method similarly indicates stereoselective N–H proton exchange of model substrate Et B–NH with 2,
3
3
as proposed in Nguyen’s mechanism.
Scheme 12. Proposed mechanistic cycles for AB dehydrocoupling with catalyst 2.
The catalytic cycle that we propose for catalyst 2 represents a modification of these two previously
published mechanisms and, in addition, a pathway for polymer formation. The complete mechanism can
be broken down into two connected catalytic cycles (Scheme 12): The left cycle accounts for AB
dehydrogenation with release of H and aminoborane H N–BH , respectively. The right cycle offers a
2
2
2
route for metalꢀmediated oligomerization by catalytic insertion of aminoborane into a N–H bond of the
substrate. The latter pathway via dihydrogen complex 6 was computed to exhibit very low barriers
explaining the formation of mostly linear H N–(BH –NH ) –BH . As for the Ir(POCOP) catalyst, the
3
2
2 n
3
1
1
polyaminoborane then precipitates from solution (THF) at around n ≈ 20 ( B MQꢀMAS NMR
ACS Paragon Plus Environment
22

Page 23 of 39
Journal of the American Chemical Society
spectroscopy), thereby driving the reaction to completion. However, line shape analysis of broad,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
4
unresolved peaks exhibits relatively large uncertainty. Crossꢀlinking of the polymer to a smaller
extend is also indicated by solid state NMR.
Rapid, metalꢀcatalyzed B–N coupling with catalysts 1 and 2 also explains the trapping experiments
(Scheme 13): The initial absence of cyclohexylboranes suggests that hydroboration is kinetically not
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
competitive with B–N coupling. The calculated barriers for dehydrogenation and B–N coupling are
almost identical suggesting a low steadyꢀstate concentration for aminoborane. However, the observation
of the trapping product several hours after completion of the reaction can be explained by slow
degradation of polyaminoborane via the reverse reaction, i.e. aminoborane extrusion from the polymer.
Note that the linear diborazane H N–BH –NH –BH had been reported as a stable molecule, suggesting
3
2
2
3
2
0
that polymer decomposition should be a catalyzed process. Furthermore, B–N coupling with catalyst 2
was not only computed to proceed with low barriers (DFT), but is also almost thermoneutral, hence
reversible. In case of catalyst 3, the relative reaction rates of B–N coupling vs. hydroboration are more
in favor of aminoborane trapping, again indicating that B–N coupling is a metalꢀcatalyzed process.
Scheme 13. Qualitative kinetic model that explains the trapping experiments.
Reversibility of the metal centered B–N coupling process was previously demonstrated by Manners
4
5
and coꢀworkers for MeAB: When starting either from MeAB or from MeH N–BH –MeHN–BH in
2
2
3
the presence of catalyst [IrH (POCOP)] the polymeric dehydropolymerization product (MeHN–BH )
2
2 x
was obtained. Addition of cyclohexene, did not give any hydroboration trapping product of the free
ACS Paragon Plus Environment
23

Journal of the American Chemical Society
Page 24 of 39
aminoborane. Interestingly, the trapping product MeHN–BCy was formed upon metal free, thermal
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
degradation of MeH N–BH –MeHN–BH . These results indicate, as for catalysts 1 and 2, that trapping
2
2
3
by hydroboration cannot compete with catalytic B–N coupling. While the B–N coupling mechanism for
the Ir(POCOP) system is not known, Weller and coꢀworkers proposed a pathway for metal centered
+
dehydrocoupling of two MeAB with [Ir(H) (H ) (PCy ) ] via initial B–H activation, H ꢀloss, and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2 2
3 2
2
1
3e
coupling of two metalꢀbound fragments. Our mechanism via initial N–H activation and outer sphere
coupling represents an alternative pathway.
The proposed dehydrogenation cycle (Scheme 11, left side) starts from dihydrido amine 2 to form
dihydrogen complex 6 upon reaction with AB. The proton shuttle preꢀequilibrium prior to irreversible
loss of H was calculated to go through a considerably lower kinetic barrier compared with the
2
inherently nonꢀbifunctional catalyst 3. It is qualitatively supported by the rapid, stereoselective proton
exchange of 2 with model substrate H N–BEt (EXSY NMR). This mechanism also exhibits lower
3
3
32,33
computed barriers than direct, intramolecular H loss from 2,
as proposed by Fagnou for his system.
2
The dehydrogenation cycle is closed by aminoborane loss from complex 7. This reaction step is
assumed to be irreversible due to rapid aminoborane oligomerization as suggested by the initial absence
of olefin trapping products and the computational identification of a low energy pathway for catalyzed
B–N coupling. Hence, the rate law for H release can be expressed as (Supporting Information):
2
[46]
r = K k [Ru][AB]
(Eq. 5)
(Eq. 6)
H2
1 2
kobs2 = K1k2
This simple kinetic model can be evaluated by comparison of the experimental isotope effects for kobs2
with the computed isotope effects for K k for deuterated substrates D B–NH , H B–ND , and D B–
1
2
3
3
3
3
3
4
7
ND . Such calculations were successfully used in the past to evaluate hydrogenation mechanisms, with
3
the advantage that relative rates are compared, rather than absolute rates. Computation of the isotope
effects for fully deuterated D B–ND was straightforward using [Ru(D) PMe {DN(CH CH PMe }] (d -
3
3
2
3
2
2
2
3
Me
2
) as starting structure owing to the exchange of all hydrides and the amine proton within the catalytic
H
1
H
2
D
1
D
2
cycle. The high computed isotope effect (K k /K k = 5.6, Table 2 and Supporting Information) is
ACS Paragon Plus Environment
24

Page 25 of 39
Journal of the American Chemical Society
H
D
obs2
in good agreement with the experimental (k obs2/k
= 5.9) value. Likewise, Bꢀterminal deuteration
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
gives, both experimentally and computationally, no rate change (Table 1 ‘no scramble’). For the
Me
calculations, the reaction sequence is initiated with [Ru(D) PMe {HN(CH CH PMe }] (d -2 ),
2
3
2
2
2
2
assuming regioselective transfer of B–D to Ru–D and N –H to NPNP–H, as proposed within our
AB
mechanism (Scheme 10 and Supporting Information Scheme S1). Interestingly, the computed value
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H
H
D
D
H
D
(
K k /K k = 1.0) is a combination of an inverse equilibrium isotope effect (K /K = 0.5), which
1 2 1 2 1 1
H
2
D
2
is offset by the normal KIE for HD elimination (k /k = 2.1). However, the computed isotope effects
are complex and cannot be easily rationalized, as several E–H/D (E = N, B, Ru) bonds are
simultaneously formed and broken in both transition states.
Table 1. Comparison of computed and experimental isotope effects.
H
H
D
D
K k / K k
2
H
1
2
1
k
/
obs2
D
no scramble scramble
k
obs2
D B–NH
1.0
8.2
1.7
5.3
1.0
3
3
3
3
H B–ND
3
5.3
5.9
D B–ND
3
5.6
The situation is more complex for Nꢀdeuterated substrate D N–BH . The prediction of the isotope
3
3
effect by DFT with the same approach as for the isotopomer H N–BD , i.e. assuming regioselective Hꢀ
3
3
Me
transfer by using [Ru(H) PMe {DN(CH CH PMe }] (d -2 ) as starting structure (Scheme S1 ‘no
2
3
2
2
2
1
H
1
H
2
D
1
D
2
scramble’) gives a very high value (K k /K k = 8.2), hence failing to reproduce the experimental
H
D
obs2
result (k obs2/k
= 5.3). However, the smaller rate retardation can be rationalized by the consideration
of H/Dꢀscrambling. Our results from 2D EXSY NMR spectroscopy demonstrated rapid substrate Nꢀ
terminal proton exchange with the catalyst hydride ligands and amine proton. This observation suggests
possible H/D scrambling prior to H elimination. In fact, H/D scrambling likely proceeds through the
2
same intermediate as dehydrogenation (6). It is therefore reasonable to assume that scrambling is at least
on the same timescale or faster as dehydrogenation. In that case, the system will then follow the lowest
energy pathway out of the different possible isotopomeric distributions. Hence, the alternative pathways
ACS Paragon Plus Environment
25

Journal of the American Chemical Society
Page 26 of 39
Me
Me
with H/D scrambling prior to H elimination, i.e. starting from d -2 and d -2 for the isotopomers
2
0
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
D N–BH and D N–BH (Supporting Information, Scheme S1 ‘scramble’), respectively, were also
3
3
3
3
accounted for. In fact, for D N–BH the scrambling pathway (Tables 1 and S4) provides a smaller
3
3
H
1
H
2
D
1
D
2
computed isotope effect (K k /K k = 5.3), which will therefore be kinetically preferred, and is in
excellent agreement with experiment. In comparison, the ‘scrambling’ route exhibits a higher normal
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H
1
H
2
D
1
D
2
isotope effect for isotopomer H N–BD (K k /K k = 1.7) than the ‘nonꢀscrambling’ route, also in
3
3
agreement with experiment.
Catalyst 1 exhibits considerably different KIEs with deuterated substrates than catalyst 2. While a KIE
H
D
k
obs1/k
= 1.8 was found for D B–NH , with Nꢀdeuterated substrates H B–ND and D B–ND a shift
obs1 3 3 3 3 3 3
to zero order dependence of the rate law on substrate was observed, indicating a change in the turnꢀover
limiting step. Also, the spectroscopic characterization of the resulting polyaminoborane indicated some
crossꢀlinking for catalyst 2, which was not observed for 1. Based on these results, complexes 1 and 2
seem to catalyze AB dehydrocoupling via different mechanisms, and not as active intermediates within
the same catalytic cycle, as was previously proposed. However, at the present stage we cannot offer a
detailed mechanism for AB dehydrocoupling with catalyst 1 which will be subject to further studies.
CONCLUSIONS
In conclusion, the present paper offers a new mechanistic proposal for the rapid ammoniaꢀborane
dehydrocoupling with bifunctional Ru amine catalyst 2, substantiated by kinetics, aminoborane trapping
experiments, stoichiometric reactions with model substrates, and DFT computations. Strong support for
the deydrogenation cycle comes from a good agreement of the experimental and computed KIEs. The
rapid dehydrogenation rates are attributed to the outerꢀsphere nature of the mechanism which is
associated with low energetic contributions for rearrangement of the first coordination sphere around the
metal and to the cooperative behaviour of the PNP ligand. The proposed mechanism with catalytic
release of aminoborane upon dehydrogenation of AB and subsequent metal catalyzed B–N coupling to
(mainly) polyaminoboranes relies on initial N–H activation as entry to both subꢀcycles. This pathway is
different to the proposed mechanisms with other catalysts:
ACS Paragon Plus Environment
26

Page 27 of 39
Journal of the American Chemical Society
1
. polyaminoborane formation by initial B–H activation, dehydrogenation, and metal centered B–N
1
2
1
3,14b
coupling without aminoborane release (iridium catalysts),
. release of catalytically dehydrogenated aminoborane and subsequent uncatalyzed aminoborane
oligomerization resulting in formation of borazine and polyborazylene (nickel and Shvo’s
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
1
4b,29
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
catalysts).
Low steadyꢀstate concentrations of aminoborane are indicated by the initial absence of trapping
products and typical, cyclic products from aminoborane oligomerization, which is explained by a
catalytic aminoborane enchainment reaction also relying on metalꢀligand cooperativity. Therefore,
catalyst 2 is considered bifunctional in the literal and a figurative sense: (a) two catalyst functional
groups are involved in the turnover limiting steps and (b) both dehydrogenation and BꢀNꢀcoupling are
catalyzed. As an interesting consequence of this two stage mechanism, the employment of two separate
catalysts, one for dehydrogenation and one for B–N coupling or even for polyaminoborane
isomerization in a subsequent treatment, could be a new approach for the synthesis of B–N polymers
with wellꢀdefined microstructure.
The proposed mechanistic model presented in this paper is highly simplified and does not account
quantitatively for some observations, such as polyaminoborane crossꢀlinking, polyborazylene sideꢀ
product formation, or the mechanistic pathway for our original catalyst 1. However, it provides an
excellent basis for further catalyst design and an important prerequisite to control the microstructure of
BNꢀpolymers, which is currently a highly unexplored field.
EXPERIMENTAL SECTION
Materials and Methods. All experiments were carried out under an atmosphere of argon using
Schlenk and gloveꢀbox techniques. Benzene and THF were dried over Na/benzophenone, distilled under
argon and deoxygenated prior to use. Pentane was dried and deoxygenized by passing through columns
packed with activated alumina and Q5, respectively. Deuterated solvents were dried by distillation from
t
Na/K alloy (C D and d ꢀTHF), and deoxygenated by three freezeꢀpumpꢀthaw cycles. KO Bu was
6
6
8
ACS Paragon Plus Environment
27

Journal of the American Chemical Society
Page 28 of 39
purchased from VWR and sublimed prior to use. 1 was synthesized as reported earlier and 2 was
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
prepared in situ from the reaction of 1 with H and used as catalyst as a freshly prepared stock
2
26b
solution. For NMR exchange measurements 2 was prepared in a JꢀYoung NMR tube and the latter
was subsequently evacuated and backfilled with Argon. AB was purchased from SigmaꢀAldrich and
sublimed prior to use. Partially and fully deuterated AB were prepared from deuterated starting
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
j
materials as reported in the literature and were sublimed prior to use.
Analytical Methods. IR spectra were recorded on a Varian FT/IRꢀ670 spectrometer. NMR spectra
were recorded on a Bruker Avance III 400 NMR spectrometer and a JEOL Lambda 400 NMR
1
3
spectrometer and were calibrated to the residual proton resonance and the natural abundance
C
=
resonance of the solvent (C D , δH = 7.16 and δC = 128.06 ppm; d ꢀTHF, δH = 1.72 and 3.57 ppm, δC
6
6
8
31
2
5.3 and 67.4 ppm). P NMR chemical shifts are reported relative to external phosphoric acid (δ 0.0
11
ppm). B NMR chemical shifts are reported relative to external BF etherate. Signal multiplicities are
3
1
abbreviated as: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), br (broad). H 2D EXSY
NMR spectra were obtained at 300 K in d ꢀTHF using the Bruker pulse sequence noesyph with D1 set
8
to 2s.
Solid State NMR experiments were conducted at a magnetic field of 11.7 T on a Bruker Avance IIIꢀ
1
5
00 spectrometer equipped with a 2.5 mm double resonance MAS probe at H frequency of
1
500.13 MHz. Saturation combs were used before all repetition delays. The H resonance of 1%
11
tetramethylsilane in CDCl served as an external secondary reference for the B resonance of BF ꢁEt O
3
3
2
1
1
48
in CDCl using the Ξ values for B as reported by IUPAC.
3
1
1
The B MQMAS was performed using a tripleꢀquantum filtered threeꢀpulse sequence with a zꢀfilter
1
49
and H continuousꢀwave decoupling implemented with a 24ꢀstep nested phase cycle. The repetition
delay was set to 0.8 s and acquired 360 transients/FID at a sample spinning frequency ν of 25 kHz. The
r
secondꢀorder quadrupolar effect (SOQE) parameters and the isotropic chemical shift values were
50
11
determined by moment analysis and fitting from projections taken from the sheared B MQMAS
spectrum.
ACS Paragon Plus Environment
28

Page 29 of 39
Journal of the American Chemical Society
Stoichiometric Experiments. Synthesis of H N–BEt : NH (g) (15 mL) is dried by condensation into
3
3
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a flask with sodium at –78 °C and subsequently condensed trapꢀtoꢀtrap to a solution of BEt in THF (20
3
mL, 20 mmol BEt ). After warming to room temperature and evaporation of NH the solvent is removed
3
3
i. vac. until a colorless oil remains. Further removal of the solvent at elevated temperatures i. vac. leads
to product decomposition. Therefore, the crude product, H NꢀBEt • 0.5 THF, was used as obtained:
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
1
THF
NMR (C D , r.t., [ppm]) H NMR (399.8 MHz): δ 3.44 (2H, m, CH
), 1.96 (3H, br, NH ), 1.50 (2H,
6
6
2
3
THF
3
3
11
m, CH2 ), 0.80 (9H, t, J = 7.9 Hz, CH ), 0.19 (6H, q, J = 7.9 Hz, CH ). B NMR (128.3
HH
3
HH
2
1
THF
MHz): δ –5.3. NMR. (d ꢀTHF, r.t., [ppm]) H NMR (399.8 MHz): δ 3.61 (2H, m, CH
), 3.52 (3H, br,
8
2
THF
2
3
3
11
NH ), 1.77 (2H, m, CH
), 0.74 (9H, t, J = 7.8 Hz, CH ), 0.15 (6H, q, J = 7.8 Hz, CH ). B
3
HH
3
HH
2
1
3
1
THF
THF
NMR (128.3 MHz): δ ꢀ5.9. C { H} NMR (100.6 MHz): δ 68.2 (s, CH
), 26.4 (s, CH2 ), 15.4 (br,
2
13
1
CH ), 10.2 (s, CH ). Assignments were confirmed by C { H} DEPT NMR.
2
3
Reaction of 2 with Et B–NH . 2 (14.7 ꢁmol) is mixed with BEt NH ꢁ 0.5 THF (22.0 ꢁmol, 1.5 equiv)
3
3
3
3
1
31
11
in dry d ꢀTHF in a JꢀYoung NMR tube. Small amounts of 4 are observed in the H, P and B NMR
8
1
spectra. Selected NMR data of 4 (d ꢀTHF, r.t., [ppm]): H NMR (399.8 MHz): δ 6.50 (1H, s br, NH),
8
2
3
BEt3NH2
3
0
.19 (2H, m, RuꢀNH ꢀB), 1.40 (d, J = 7.4 Hz, 9H, P(CH ) ), 0.66 (9H, t, J = 7.6 Hz, CH3
),
2
HP
3 3
HH
3
BEt3NH2
2
31
1
.08 (6H, q, J = 7.6 Hz, CH2
), –18.44 (1 H, q, J = 24.1 Hz, RuꢀH). P{ H} NMR (161.83
HH
HP
2
i
2
11
MHz): δ 62.1 (d, J =29.5 Hz, P Pr ), 10.3 (t, J =29.5 Hz, P(CH ) ). B NMR (128.3 MHz): δ –3.1
PP
2
PP
3 3
(
s). With mixing times of τ = 500 ms and 1000 ms cross peaks between BEt NH and the hydride
m
3 3
1
adjacent to the N–H proton of 2 were observed in the H 2D EXSY NMR spectra (Supporting
Information Figure S5).
Reaction of 3 with Et B–NH . 3 (10.0 mg, 20.1 ꢁmol) is dissolved in dry d ꢀTHF in a JꢀYoung NMR
3
3
8
tube with BEt NH ꢁ 0.5 THF (4.5 mg, 29.8 ꢁmol, 1.5 equiv). With mixing times of τ
m
= 500 ms and
3
3
1
1
000 ms no exchange peaks between NH BEt and the hydrides of 3 were observed by H 2D EXSY
3
3
NMR spectra.
ACS Paragon Plus Environment
29