Experimental and theoretical treatment of hydrogen splitting and storage in boron-nitrogen systems

Article abstract of DOI:10.1016/j.jorganchem.2009.03.023

Hydrogen gas serves as a reducing agent and hydrogen atom source in numerous industrially important chemical processes and also has a great potential as a clean power source for fuel cells. In this respect, the reversible storage of hydrogen and the development of new metal-free hydrogenation catalysts are important tasks. Here, we review the recent literature, primarily on cases where the split H₂ forms an N-H?H-B dihydrogen bond. In these systems dihydrogen interaction was found to be the key actor in the hydrogen liberating process. Accordingly, the intramolecular ansa-aminoboranes (where B and N atoms are situated within each other's range) can reversibly activate hydrogen. Moreover, the theoretical studies of the hydrogen splitting by bulky Lewis acid-Lewis base systems are discussed.

Full text of DOI:10.1016/j.jorganchem.2009.03.023

Journal of Organometallic Chemistry 694 (2009) 2654–2660
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Review
Experimental and theoretical treatment of hydrogen splitting and storage
in boron–nitrogen systems
Victor Sumerin a, Felix Schulz b, Martin Nieger a, Michiko Atsumi a, Cong Wang a, Markku Leskelä a,
b
Pekka Pyykkö a, Timo Repo a, , Bernhard Rieger
*
a Department of Chemistry, University of Helsinki, P.O. Box 55, FIN-00014 Helsinki, Finland
b WACKER-Lehrstuhl für Makromolekulare Chemie, Technische Universität München, Lichtenbergstrasse 4, D-85747 Garching bei München, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Hydrogen gas serves as a reducing agent and hydrogen atom source in numerous industrially important
chemical processes and also has a great potential as a clean power source for fuel cells. In this respect, the
reversible storage of hydrogen and the development of new metal-free hydrogenation catalysts are
important tasks. Here, we review the recent literature, primarily on cases where the split H2 forms an
N–Hꢀ ꢀ ꢀH–B dihydrogen bond. In these systems dihydrogen interaction was found to be the key actor
in the hydrogen liberating process. Accordingly, the intramolecular ansa-aminoboranes (where B and N
atoms are situated within each other’s range) can reversibly activate hydrogen. Moreover, the theoretical
studies of the hydrogen splitting by bulky Lewis acid–Lewis base systems are discussed.
Ó 2009 Elsevier B.V. All rights reserved.
Received 30 January 2009
Received in revised form 12 March 2009
Accepted 13 March 2009
Available online 21 March 2009
Keywords:
Hydrogen
Boron
Nitrogen
Activation
Liberation
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2654
2. Characterizing B–Hꢀ ꢀ ꢀH–N dihydrogen bond systems with the Cambridge Structural Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2655
3. Hydrogen activation by amines in combination with B(C6F5)3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2655
4. Reversible hydrogen activation by the ansa-aminoborane 1-N-TMPN-CH2-2-[B(C6F5)2]C6H4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2657
5. Catalytic reduction of imines by the ansa-ammonium-borate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2657
6. Modification of the ansa-aminoborane system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2658
7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2658
8. Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2658
8.1. Physical methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2658
8.2. Synthetic methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2659
8.2.1. Trans-2,6-Dimethyl-2,6-diphenylpiperidinium hydrido[tris(pentafluorophenyl)]borate (5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2659
8.2.2. 1-(2-Bromo-3-methylbenzyl)-2,2,6,6-tetramethylpiperidine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2659
8.2.3. 1-{2-[Bis(pentafluorophenyl)boryl]-3-methylbenzyl}-2,2,6,6-tetramethylpiperidine (8). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2659
8.2.4. Hydrido{2-methyl-6-[(2,2,6,6-tetramethylpiperidinium-1-yl)methyl]phenyl} bis(pentafluorophenyl)borate (9) . . . . . . . . 2659
8.2.5. Dehydrogenation of 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2659
8.2.6. Dehydrogenation of 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2660
8.3. Crystallographic studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2660
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2660
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2660
1. Introduction
Hydrogen bonds are well-known interactions and considered
to be very important in modern physics, chemistry and biology
[1]. In the middle of the 90 s, a new kind of hydrogen bonding
* Corresponding author. Fax: +35 8919150198.
E-mail addresses: timo.repo@helsinki.fi (T. Repo), rieger@tum.de (B. Rieger).
0022-328X/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2009.03.023

V. Sumerin et al. / Journal of Organometallic Chemistry 694 (2009) 2654–2660
2655
(M–Hdꢁꢀ ꢀ ꢀd+H–A) between metal or boron hydrides and classical
proton donor groups, such as NH and OH, were discussed for the
first time [2]. Later, the term ‘‘dihydrogen bond” (DHB) for these
protic–hydridic interactions in the range 1.7–2.2 Å was suggested
by Crabtree et al. [3]. Although a wide array of different organome-
tallic systems contains dihydrogen bonds, non-metals with B–
Hꢀ ꢀ ꢀH–N interactions occupy a prominent role in this study. Fur-
thermore, recent findings by our group highlight partially covalent
DHBs in the ansa-ammonium-borate (1-N-TMPN-CH2-2-
[B(C6F5)2]C6H4, where TMPNH is 2,2,6,6-tetramethylpiperidinium)
with an extremely short H–H distance of the order of 1.6 Å [4].
The phenomenon occurs in the partially covalent DHB which can
be treated as the intermediate stage of the hydrogen splitting, or
inverse liberating process. Accordingly, understanding the mecha-
nism of activation or formation of gaseous H2 by non-metal com-
pounds is essential for learning how to design metal-free
systems for H2 storage and new, metal-free hydrogenation cata-
lysts. This implies understanding the role and the nature of the
DHB interactions in these compounds [5].
Fig. 2. The B–Hꢀ ꢀ ꢀH angle versus the N–Hꢀ ꢀ ꢀH angle of aminoboranes with H–H
distances smaller than 2.2 Å.
2. Characterizing B–Hꢀ ꢀ ꢀH–N dihydrogen bond systems with the
Cambridge Structural Database
Significant interest in DHB interactions during the past 15 years
has resulted in a couple of excellent reviews [3,6]. However, to the
best of our knowledge, there are no recent systematic Cambridge
Structural Database (CSD) searches incorporating all X-ray
structures that have been published. We performed a search for
B–Hꢀ ꢀ ꢀH–N contacts in the CSD and 290 examples of intra- and
intermolecular DHBs with short d(Hꢀ ꢀ ꢀH) interactions, much smal-
ler than twice the van der Waals radius of hydrogen (2.4 Å), were
identified in 176 X-ray crystal structures (Fig. 1) [7].
The Hꢀ ꢀ ꢀH distances are usually in the range of 1.7–2.2 Å, and
the N–Hꢀ ꢀ ꢀH angles are typically ranging from 130° to 170°, while
the B–Hꢀ ꢀ ꢀH angles tend to be strongly bent (in most cases 95–
125°) (Fig. 2) [8]. The nature of these side-on bonds was explained
by Crabtree and co-workers as an interaction between the electron
Scheme 1. ((Dibenzylamino)ethyl)dicarbollyl, [nido-7-NHBnþ2 ðCH2Þ -8-Me-7; 8-C2
2
B9Hꢁ10] [11] (where Bn is benzyl).
Scheme 1) [11]. Thus, this finding allowed us to analyze the geom-
etry of the key ‘‘intermediate” of the reversible hydrogen activation
process based on the ‘‘frozen” X-ray data. It should be noted that
usually such compounds are not stable enough because of their ten-
dency to the spontaneous liberation of hydrogen. Accordingly, the
precursor to ammonia–borane – ammonium borohydride (NH4BH4)
decomposes at temperatures above ꢁ40 °C to give H2 and has a half-
life of about 6 h at room temperature [12,13]. However, the given
compound 1 is not able to form dihydrogen, because the negative
charged hydride ion is not presented in the structure of DHB.
donating
r B–H bond and the protonic N–H bond [3]. Thus,
B–Hꢀ ꢀ ꢀH–N systems are maximizing the attractive Coulomb inter-
action and their energies of interaction become substantial (12–
29 kJ/mol). The simplest and best-studied compound, which con-
tains DHBs and 19.6 wt% of hydrogen, the polar (5.2 D) ammo-
nia–borane complex ðNH3 ꢂ BH3Þ has the strikingly high melting
point of +104 °C relative to the isoelectronic substances ethane
(ꢁ181 °C) or to the also polar fluoromethane (ꢁ141 °C, 1.8 D)
mainly because of the DHB stabilization [9].
Although one of the most interesting DHB interactions with a
length less than 1.65 Å was discussed only from a theoretical point
of view [10], the shortest reliable intramolecular contact that we
found in the CSD is close to 1.6 Å (hNHH = 154°, hHHB = 125°,
3. Hydrogen activation by amines in combination with B(C6F5)3
In 2003 Roesler and Piers predicted that unusual non-metal sys-
tems based on ‘separated Lewis acid–Lewis base’ pairs would be
suitable for reversibly activating H2 (Scheme 2) [14]. Specifically,
they not only speculated that compound 3 might be a ‘‘dihydrogen
storage device, able to release H2 upon heating or during a chem-
ical reaction, regenerating 2”, but also suggested that strong DHB
interactions would play a key role in this process. Unfortunately
compound 2 was not able to cleave H2 and the system 3 was never
characterized. However, the in situ generated 1-Ph2NH-2-
[HB(C6F5)2]C6H4 (3) spontaneously liberates hydrogen.
Despite the fact that the hydrogen–hydrogen bond is remark-
ably strong (432 kJ/mol) and non-polar, 3 years later, this approach
was successfully applied for the reversible hydrogen activation by
non-metal systems based on ‘separated’ phosphine–borane pairs
Fig. 1. Dependence of the numbers of the Hꢀ ꢀ ꢀH contacts on the Hꢀ ꢀ ꢀH distance in
Scheme 2. Unsuccessful reversible hydrogen activation by Piers’s aminoborane 3
[14].
aminoboranes found in the Cambridge Structural Database.

2656
V. Sumerin et al. / Journal of Organometallic Chemistry 694 (2009) 2654–2660
Scheme 3. Proposed mechanism for the heterolytic cleavage of H2 by TMPNH and B(C6F5)3.
coworkers [21], and clear theoretical support for this idea was
found by Rokob et al. [20a]. Thereby, for hydrogen activation by
t-Bu3P and B(C6F5)3 the transition state (P–Hꢀ ꢀ ꢀH–B) was almost
linear, with H–H only slightly lengthened from 0.74 to 0.79 Å,
and the activation energy was 43.5 kJ/mol. The H2 was polarized
and chemically activated by the totally frustrated system.
Later we examined the heterolytic splitting of hydrogen by
trans-2,6-dimethyl-2,6-diphenylpiperidine and B(C6F5)3. This reac-
tion led to the formation of product 5 in quantitative yield after 3 h
at room temperature. Interestingly, when a toluene solution of 5
(0.1 M) was refluxed at 110 °C in a closed system under reduced
pressure for 36 h, a 50% conversion of 5 to the starting materials
was observed (Scheme 4). This finding is in contrast to previous
compounds [TMPNH2]+[BH(C6F5)3]ꢁ (4) and [t-BuNH2Bn]+-
[BH(C6F5)3]ꢁ (where Bn is benzyl) which are not able to liberate
hydrogen upon heating [19,22].
Scheme 4. Reversible hydrogen activation by trans-2,6-dimethyl-2,6-diphenylpi-
peridine and B(C6F5)3.
[15,16]. Moreover, a number of theoretical papers and reviews on
hydrogen storage in Nꢀ ꢀ ꢀB systems have been published [17,18].
We reported the facile splitting of hydrogen by 2,2,6,6-tetra-
methylpiperidine (TMPNH) in combination with B(C6F5)3 in 2008
[19]. Whereas this reaction gave the stable hydrogenated ionic
product 4, a remarkable intermediate 4a, with a strong N–Hꢀ ꢀ ꢀH–
B interaction between ammonium and borate was observed by
NMR spectroscopy (Scheme 3). Moreover, a recently reported con-
certed Lewis acid–Lewis base mechanism of heterolytic hydrogen
splitting is quite consistent with the proposed mechanism [20],
in which the bulky amine and borane are needed to prevent the
formation of a classical Lewis donor–acceptor adduct. Instead of
forming a favorable N–B bond, the N with its lone electron pair
To gain further insight into the mechanism of this reaction, we
performed X-ray diffraction experiments (Fig. 3). While the X-ray
crystallographic study of 4 showed isolated ions with the shortest
NHꢀ ꢀ ꢀHB distance of 2.97 Å, connected only by a framework of
NHꢀ ꢀ ꢀF (2.07 Å) hydrogen bonds [19], in the X-ray diffraction struc-
and the B with its 2pp acceptor orbital remain at a certain distance.
This makes it possible to use them as proton and hydride accep-
tors, respectively, upon the cleavage of dihydrogen. This situation
was previously characterized as ‘frustration’ by Stephan and
ture of 5 strong phenyl-perfluorophenyl
p
–
p
stacking and NHꢀ ꢀ ꢀF
(2.34 Å), CHꢀ ꢀ ꢀF (2.52 Å) hydrogen bonding interactions were found.
These non-covalent bonds between cation and anion resulted in the
Fig. 3. X-ray structure of 4 and 5 [19,32].

V. Sumerin et al. / Journal of Organometallic Chemistry 694 (2009) 2654–2660
2657
Scheme 5. Synthesis of the ansa-aminoborane 6.
short DHB interaction close to 1.89 Å (hNHH = 161°, hHHB = 157°)
[23]. According to these data binding together the ‘separated Lewis
acid–Lewis base’ pairs has a greater effect on the properties of the
whole system than reducing the Lewis basicity.
ting with system 6 occurs via a quasilinear (N–HꢀꢀꢀH–B) transition
state at H–H of 0.78 Å, with the activation energy of 14 kJ/mol and
the activation free energy of 62 kJ/mol [4,13,25]. The extra Coulomb
attraction between the two ions at 3.32 Å (the calculated B–N dis-
tance for system 7) would produce an attraction energy of 413 kJ/
mol, and this would be comparable with the amount of energy re-
quired for the heterolytic cleavage of H2, 432 kJ/mol [15]. According
to this calculation, ‘Coulomb pays for Heitler-London’ hypothesis was
enounced [4], although this is only an order-of-magnitude estimate.
The idea somewhat parallels the suggestion by Simons’ group, that
the Coulomb stabilization of Rydberg states by a nearby ion would
play a role in the splitting of S–S bonds in proteins [26].
4. Reversible hydrogen activation by the ansa-aminoborane 1-
N-TMPN-CH2-2-[B(C6F5)2]C6H4
The observation described above, the discovery of compound 1
with a partially covalent DHB and the previous work of Piers in-
spired us to attempt the synthesis of an intramolecular system
with a pseudo five-membered NCCCB-ring [14]. In this respect
we designed the ansa-aminoborane 6 and developed an effective
and common procedure for the preparation of dual ‘Lewis acid–Le-
wis base’ systems (Scheme 5) [4].
5. Catalytic reduction of imines by the ansa-ammonium-borate
1-(2-Bromobenzyl)-2,2,6,6-tetramethylpiperidine was synthe-
sized through alkylation of TMPNH by 2-bromobenzylbromide in
the presence of a base. It was then smoothly lithiated with t-BuLi
at ꢁ70 °C and the lithiated derivative was transmetallated with
(C6F5)2BCl to give the final product 6 in a total yield of 55%. Thus,
the ansa-aminoborane 6 has been prepared on a gram scale (up
to 5 g) by an efficient two-step synthesis from readily available
precursors.
To continue our investigations, we examined the ansa-ammo-
nium-borate 7 in hydrogenations of non-sterically demanding imi-
nes and enamines under mild conditions (110 °C, 2 atm of H2,
Table 1).
While previous non-metal systems have shown only stoichiom-
etric hydrogenation of the imine from benzaldehyde and benzyl-
amine, employing ansa-ammonium-borate 7 as catalyst in this
reaction resulted in 51% conversion of imine to amine (Table 1, en-
try 1) [22,27]. Continuing the reduction for up to the 48 h showed
negligible influence on yield (60%, Table 1, entry 2). However,
increasing the amount of catalyst from 4 to 8 mol% led to signifi-
cantly enhanced conversion after 12 h (99%, Table 1, entry 3).
The reduction of C6H5CH2N@CPh(CH3) or CH3N@CPh(CH3) gave
The compound 6 reacted rapidly with H2 at 20 °C forming the air-
and moisture stable 1-N-TMPNH-CH2-2-[HB(C6F5)2]C6H4 in quanti-
tativeyield(Scheme6). Heatingatoluenesolutionof7 undervacuum
at 110 °C (in analogy to 5) up to 20 h resulted in the quantitative
recovery of product 6 (Scheme 8). Accordingly, ansa-ammonium-bo-
rate 7 loses hydrogen much faster than the non-bridged system 5.
In order to understand such a significant effect of the covalent
bridge on the reactivity, the structure of 7 was determined by
X-ray crystallography and studied theoretically (Fig. 4). While the
X-ray data showed that the NHꢀꢀꢀHB distance is not as short as those
found forcompound1 (1.78 Å, compared with1.6 Å) [4], thetheoret-
ical investigations, including solvation, led to a HꢀꢀꢀH distance of
1.51 Å. Furthermore, preliminary neutron diffraction experiments
also indicate an H–H distance of 1.58 Å [24]. The fact that they are
shorter than the apparent X-ray value of 1.78 Å, may largely be be-
cause the X-ray technique detects electron-density distribution
andlocatestheelectron-densitymaximaoftheatoms, whileneutron
scattering and calculations experiments locate the nuclei. Moreover,
not only the short HꢀꢀꢀH contact indicates the presence of a strong
DHB interaction, but also the experimental topological parameters
such as the NHH (154°) and BHH (125°) angles are almost the same
in compounds 1 and 7. Accordingly, these results confirm the view
that the formation of the stable, partially covalent DHB occurs only
under a favorable geometry of the system and is one important pre-
requisite for the hydrogen liberation reaction. Further theoretical
study of reaction path and energetic showed that the hydrogen split-
N-benzyl-a-methylbenzylamine or N-methyl-a-methylbenzyl-
amine, respectively, in almost quantitative conversions (Table 1,
entries 4 and 5), while hydrogenation of less steric imines such
as CH3N@CPh(H) and CH3N@CCH2Ph(CH3) provided only traces
of the corresponding amines (4%, Table 1, entries 7 and 8).
These data support the standpoint that the rate-determining
step of the hydrogenation mechanism is the suppression of the cat-
alytic activity by amine–borane adduct formation 7a and therefore
the sum of the Lewis acidic and the steric factors of the catalyst and
imine, respectively, play a decisive role (Scheme 7).
Scheme 6. Reversible hydrogen activation by the ansa-aminoborane 6.
Fig. 4. X-ray structure of 7 [4].

2658
V. Sumerin et al. / Journal of Organometallic Chemistry 694 (2009) 2654–2660
Table 1
tion reactions [15c]. In this respect, the new ansa-aminoborane 8
with an electron donating group in ortho-position in relation to
the boron was synthesized in the same manner as 6 (Scheme 5).
Interestingly, the electron donating group dramatically increases
the time of the hydrogen splitting (1 week instead of a few minutes)
as expected. It does not affect the molecular hydrogen formation
reaction (Scheme 8). This phenomenon can only be explained by
the steric or electronic influence of the CH3 group on the structure
of the ansa-ammonium-borate 9. Although, the X-ray crystallo-
graphic study of 9 showed only insignificant changes in the
rigid geometry (in comparison with 7), the typical DHB interaction
of 1.90 Å (hNHH = 133° and hBHH = 126°) was detected instead of
the partially covalent bond in system 7 (Fig. 5) [23].
Hence not only must the geometry of aminoboranes be appropri-
ate to form strong DHB interactions, but the dual Lewis acidity/basi-
city must be correctly tuned in terms of combined efforts in order to
obtain a partially covalent DHB interaction and a successful revers-
ible hydrogen activation. Accordingly, the ansa-ammonium-borate
7 is able to split and form hydrogen in a facile way because: (1) the
acidity and basicity of the active centers are sufficiently matched
and(2)thefavorablegeometryfortheformationofthepartiallycova-
lent DHB interaction, whichplaysa keyrole inthis process, is present.
Catalytic reduction of imines by the ansa-ammonium-borate 7.
Entry
Substrate
Time (h)
24
Amine
Yield (%)a
51b
1
N
N
H
N
H
2
48
60b
N
N
N
N
N
N
H
3
4
12
6
99c
99b
N
H
5
6
12
‘
99b
4b
N
H
N
H
H
N
N
7
24
4b
Yield was determined by 1H NMR spectroscopy.
Catalyst 7 (4 mol%).
Catalyst 7 (8 mol%).
a
b
c
6. Modification of the ansa-aminoborane system
On the basis of the above results and the known chemistry of
boranes, we proposed that the further reduction of Lewis acidity
at the active boron center should not only lower the temperature
needed for the hydrogen liberation process significantly, but should
also lead to an increasing the activity of our catalyst in hydrogena-
7. Conclusion
In summary, the short (less than 1.9 Å) and very short (less than
1.7 Å) dihydrogen bonds in boron–nitrogen systems are the key ac-
tors in the hydrogen liberating process. However, much of the en-
ergy, needed for splitting H2, may come from the Coulomb
attraction between the two counterions. Further, careful design
of ‘Lewis acid–Lewis base’ systems, with appropriate geometry
and matching of acidity and basicity, is essential for the synthesis
of metal-free compounds for H2 storage and hydrogenation cata-
lysts. In fact this evolution is needed before any such applications
can become practical and be effectively used.
8. Experimental
Scheme 7. Adduct of catalyst 7 with N-methyl-1-phenylethanamine 7a.
8.1. Physical methods
All experiments were performed on double-manifold H2(Ar)/
vacuum lines or in an argon glove box (MBraun Labmaster 130).
Solvents were dried by an MBraun solvent purification system
(MB SPS-800). Hydrogen gas was purchased from AGA Ab and
passed through a drying unit prior to use. All organic reagents were
purchased from Sigma–Aldrich and purified by conventional meth-
ods. HRMS (ESI+-TOF) mass spectra were recorded on a Bruker
micrOTOF mass spectrometer. NMR experiments were performed
Scheme 8. Reversible hydrogen activation by the ansa-aminoborane 8.
Fig. 5. X-ray structures of the ansa-ammonium-borates 9 (left) and 7 (right) [4,33].

V. Sumerin et al. / Journal of Organometallic Chemistry 694 (2009) 2654–2660
2659
on a Bruker ARX-300 spectrometer (1H, 13C, 19F) or Bruker DPX-400
M solution of t-BuLi in pentane (3.6 mL, 6.1 mmol). The solution was
allowed to warm up to room temperature and stirred over night. The
solvent was removed in vacuo and the solid residue suspended in
hexane (5 mL). Filtration yielded the lithium derivative as a fine pale
yellow powder. Without further purification the crude lithium salt
was dissolved in 17.5 mL of a 40:60 hexane/toluene mixture and
cooled to ꢁ20 °C. A solution of (C6F5)2BCl (1.141 g, 3 mmol) in tolu-
ene (5 mL) was added dropwise over 5 min. Immediately, intense
bright yellow coloration indicated formation of the product. The
reaction mixture was stirred over night at room temperature and
the solvent was removed in vacuo. The solid residue was suspended
inhexane(15 mL), thecontentsfiltered. Thesolventpresentinthefil-
trate was removed to give 0.797 g of 1-{2-[bis(pentafluorophe-
(
11B). 1H and 13C NMR spectra are referenced to Me4Si by
referencing the residual solvent peak. 11B, 19F NMR spectra were
referenced externally to BFꢂ Et2O at 0 ppm and CF3CO2H at
3
ꢁ78.5 ppm relative to CFCl3 at 0 ppm, respectively. trans-2,6-Di-
methyl-2,6-diphenylpiperidine, 2-bromo-3-methylbenzyl bromide
and (C6F5)2BCl were prepared by the literature methods [28–30].
8.2. Synthetic methods
8.2.1. Trans-2,6-Dimethyl-2,6-diphenylpiperidinium
hydrido[tris(pentafluorophenyl)]borate (5)
In a glove box, a 25-mL flame-dried Schlenk tube equipped with a
stir bar, a Teflon stopcock and a cap (GlindemannÒ-sealing rings were
used for conical joints instead of grease) was charged with B(C6F5)3
(0.2 mmol, 102.4 mg), 1 mL dry toluene and trans-2,6-dimethyl-2,6-
diphenylpiperidine (0.2 mmol, 53.1 mg). The reaction was degassed
once with a freeze–pump–thaw cycle and refilled with H2 (1 atm).
The reaction wasstirred at RTfor 3 h. All volatiles wereremoved in va-
cuo to give the product 5 (155.9 mg, 100% yield) as a white solid. Crys-
tals suitable for X-ray diffraction were grown from a C6D6 solution at
20 °C. Anal. Calc. for C37H25BF15N: C, 57.02; H, 3.23; N, 1.80. Found: C,
57.30; H, 3.20; N, 1.85%. 1H NMR (CD2Cl2, 300 MHz): d 7.50–7.42 (m,
nyl)boryl]-3-methylbenzyl}-2,2,6,6-tetramethylpiperidine
(yield
45%) as an orange oil. 1H NMR (C6D6, 300 MHz): d 7.90 (d, 1H,
3JHH = 8.0 Hz, ortho-(C6F5)2B-C6H3), d 7.26 (t, 1H, 3JHH = 7.3 Hz, para-
3
(C6F5)2B-C6H3), d 6.91 (d, 1H, JHH = 7.3 Hz, ortho-TMPCH2-C6H3), d
3.70 (s, 2H, TMPCH2), d 2.05 (s, 3H, C6H3CH3), d 1.37 (br. s, 6H,
CH2CH2CH2), d 0.81 (br. s, 12H, CH3). 11B NMR (C6D5CD3, 128 MHz):
d 44.12 (br. s). 13C NMR (C6D6, 75 MHz): d 150.18 (s, quaternary car-
bon of CH3C6H3), d 148.64 (s, quaternary carbon of TMPCH2C6H3), d
1
1
147.09 (dm, JCF = 250 Hz, ortho-C6F5), d 144.50 (dm, JCF = 260 Hz,
1
para-C6F5), d 138.61 (dm, JCF = 255 Hz, meta-C6F5), d 130.64 (s,
para-(C6F5)2B-C6H3), d 127.40 (s, ortho-CH3-C6H3), d 127.40 (s,
ortho-TMPCH2-C6H3), d 64.90 (s, NC(CH3)2CH2), d 50.66 (s, TMPCH2),
d 44.85 (s, CH2CH2CH2), d 30.54 (br. s, CH3), d 22.95 (s, C6H3CH3), d
22.70 (br. s, CH3), d 18.55 (s, CH2CH2CH2). Quaternary carbons of
C6F5 ring and (C6F5)2BC6H3 were not observed. 19F NMR (C6D6,
1
10H, C6H5), d 5.99 (br. t, 2H, JNH = 49 Hz, NH2), d 3.58 (br. q, 1H,
1JBH = 86 Hz), d 2.69 (m, 2H, CH2), d 2.11 (m, 4H, CH2), d 1.44 (s, 6H,
1
CH3). 11B NMR (CD2Cl2, 128 MHz): d ꢁ24.48 (d, JBH = 86 Hz). 13C
1
NMR (C6D6, 75 MHz): d 148.87 (dm, JCF = 241 Hz, ortho-C6F5), d
138.81 (dm, 1JCF = 247 Hz, para-C6F5), d 138.15 (s, Quaternary carbon
3
282 MHz): d ꢁ129.40 (d, 6F, JFF = 21 Hz, ortho-C6F5), d ꢁ145.25 (t,
1
of C6H5), d 137.35 (dm, JCF = 255 Hz, meta-C6F5), d 130.05 (s, para-
3
3F, JFF = 20 Hz, para-C6F5), d -161.11 (m, 6F, meta-C6F5). HRMS
C6H5), d 129.83 (s, meta-C6H5), d 124.35 (s, ortho-C6H5), d 65.48 (s,
NC(CH3)(Ph)CH2), d 32.62 (s, NC(CH3)(Ph)CH2), d 28.17 (s, CH3), d
ESI+-TOF: C29H26BNFꢂ10Hþ; Calc. 590.2077. Found: 590.2076.
16.18 (s, CH2). Quaternary carbon of C6F5 ring was not observed. 19
F
8.2.4. Hydrido{2-methyl-6-[(2,2,6,6-tetramethylpiperidinium-1-
yl)methyl]phenyl} bis(pentafluorophenyl)borate (9)
NMR (CD2Cl2, 282 MHz): d ꢁ134.83 (d, 6F, 3JFF = 21 Hz, ortho-C6F5), d
ꢁ164.48 (t, 3F, 3JFF = 21 Hz, para-C6F5), d ꢁ167.69 (m, 6F, meta-C6F5).
In a glove box, a 100-mL flame-dried Schlenk tube equipped
with a stir bar, a Teflon stopcock and a glass stopper (Glinde-
mannÒ-sealing rings were used for conical joints instead of grease)
was charged with 1-{2-[bis(pentafluorophenyl)boryl]-3-methyl-
benzyl}-2,2,6,6-tetramethylpiperidine (8, 1.0 mmol, 0.589 g) and
10 mL dry toluene. The reaction was degassed once with a freeze–
pump–thaw cycle and refilled with H2 (1 atm). The reaction was
stirred at 1000 rpm at room temperature for 7 days. All volatiles
were removed in vacuo. The solid residue suspended in hexane
(15 mL), the contents filtered to give 0.591 g of the product as a
white solid (yield 100%). 1H NMR (CD2Cl2, 300 MHz): d 7.07–6.95
8.2.2. 1-(2-Bromo-3-methylbenzyl)-2,2,6,6-tetramethylpiperidine
A dry 25 mL Schlenk tube was charged with 1.413 g (10 mmol)
2,2,6,6-tetramethylpiperidine, 2.640 g (10 mmol) 2-bromo-3-
methylbenzyl bromide, 1.6 g (11.6 mmol) K2CO3 and 0.166 g KI
(1 mmol) in 18 mL of dry acetone. The mixture was heated at
95 °C for 48 h. The Schlenk tube was cooled and the contents fil-
tered. The solvent present in the filtrate was removed under re-
duced pressure. The crude product was dissolved in 50 mL of Et2O
and extracted twice with 100 mL of 0.1 M HCl. The aqueous solu-
tion was basified to pH 12 with KOH and extracted twice with
50 mL of CH2Cl2. The resulting organic layer was dried over
K2CO3, passed through a short column with silica gel and rotovaped
to give 2.85 g of 1-(2-bromo-3-methylbenzyl)-2,2,6,6-tetramethyl-
piperidine (yield 88%) as white crystals. 1H NMR (CDCl3, 300 MHz):
3
(m, 3H, C6H3), d 6.11 (br. s, 1H, NH), d 4.57 (d, 2H, JHH = 5.5 Hz
1
TMPNHCH2), d 3.67 (br. q, 1H, JBH = 76 Hz, BH), d 2.05 (s, 3H,
C6H3CH3), d 1.73 (m, 6H, CH2CH2CH2), d 1.53 (s, 6H, CH3), d 1.29
(s, 6H, CH3). 11B NMR (CD2Cl2, 128 MHz): d ꢁ21.69 (d, 1JBH = 76 Hz).
13C NMR (CDCl3, 75 MHz): d 148.29 (dm, 1JCF = 230 Hz, ortho-C6F5), d
3
3
1
d 7.73 (d, 1H, JHH = 7.7 Hz, ortho-CH3-C6H3), d 7.17 (t, 1H, JHH
=
146.55 (s, quaternary carbon of CH3C6H3), d 138.14 (dm, JCF
=
3
1
7.5 Hz, para-Br-C6H3), d 7.06 (d, 1H, JHH = 7.8 Hz, ortho-TMPCH2-
C6H3), d 3.73 (s, 2H, TMPCH2), d 2.41 (s, 3H, –C6H3CH3), d 1.73 (br.
s, 2H, CH2CH2CH2), d 1.54 (br. s, 4H, CH2CH2CH2), d 1.10 (br. s, 6H,
CH3), d 0.86 (br. s, 6H, CH3). 13C NMR (CDCl3, 75 MHz): d 144.01
(s, quaternary carbon of TMPCH2C6H3), d 137.18 (s, quaternary car-
bon of CH3C6H3), d 128.15 (s, ortho-CH3C6H3), d 127.97 (s ortho-
TMPCH2-C6H3), d 125.97 (m, quaternary carbon of Br-C6H3), d
124.89 (s, para-Br-C6H3), d 54.83 (s, NC(CH3)2CH2), d 49.50 (s,
TMPCH2), d 41.30 (s, CH2CH2CH2), d 33.12 (br. s, CH3), d 23.30 (s,
C6H3CH3), d 21.73 (br. s, CH3), d 17.89 (s, CH2CH2CH2).
245 Hz, para-C6F5), d 136.82 (dm, JCF = 250 Hz, meta-C6F5), d
134.95 (s, quaternary carbon of TMPCH2C6H3), d 131.14 (s, para-
(C6F5)2BH-C6H3), d 125.28 (s, ortho-CH3C6H3), d 125.22 (s, ortho-
TMPCH2-C6H3), d 67.60 (s, NC(CH3)2CH2), d 54.85 (s, TMPNHCH2),
d 41.35 (s, CH2CH2CH2), d 31.62 (br. s, CH3), d 23.34 (s, C6H3CH3), d
21.57 (br. s, CH3), d 15.74 (s, CH2CH2CH2). Quaternary carbons of
C6F5 ring and (C6F5)2BC6H3 were not observed. 19F NMR (CD2Cl2,
3
282 MHz): d ꢁ134.06 (d, 6F, JFF = 21 Hz, ortho-C6F5), d ꢁ164.33 (t,
3
3F, JFF = 20 Hz, para-C6F5), d ꢁ167.45 (m, 6F, meta-C6F5). HRMS
ESI+-TOF: C29H28BNFꢂ10Naþ; Calc. 614.2052. Found: 614.2059.
8.2.3. 1-{2-[Bis(pentafluorophenyl)boryl]-3-methylbenzyl}-2,2,6,6-
tetramethylpiperidine (8)
1-(2-Bromo-3-methylbenzyl)-2,2,6,6-tetramethylpiperidine
(0.973 g, 3 mmol) was lithiated at ꢁ70 °C in Et2O (10 mL) using a 1.7
8.2.5. Dehydrogenation of 5
In a glove box, a 25-mL flame-dried Schlenk tube equipped with
a stir bar, a Teflon stopcock and a glass stopper (GlindemannÒ-
sealing rings were used for conical joints instead of grease) was

2660
V. Sumerin et al. / Journal of Organometallic Chemistry 694 (2009) 2654–2660
[2] T. Richardson, S. deGala, R.H. Crabtree, P.E.M. Siegbahn, J. Am. Chem. Soc. 117
Table 2
(1995) 12875–12876.
Crystal data of compounds 5 and 9.
[3] R.H. Crabtree, P.E.M. Siegbahn, O. Eisenstein, A.L. Rheingold, T.F. Koetzle, Acc.
Chem. Res. 29 (1996) 348–354.
5
9
[4] (a) V. Sumerin, F. Schulz, M. Atsumi, C. Wang, M. Nieger, M. Leskelä, T. Repo, P.
Pyykkö, B. Rieger, J. Am. Chem. Soc. 130 (2008) 14117–14119;
(b) CCDC 694515 (7).
[5] (a) G.J. Kubas, Science 314 (2006) 1096–1097;
(b) R.H. Crabtree, Science 282 (1998) 2000–2001;
(c) A.L. Kenward, W.E. Piers, Angew. Chem., Int. Ed. 47 (2008) 38–41.
[6] (a) R. Custelcean, J.E. Jackson, Chem. Rev. 101 (2001) 1963–1980;
(b) V.I. Bakhmutov, Dihydrogen Bond: Principles, Experiments, and
Applications, Wiley-VCH, New York, 2008.
[7] CSD version 5.29, November 2007 + 2 updates (January 2008, August 2008).
[8] The N–H (1.03 Å) distances were normalized.
[9] W.T. Klooster, T.F. Koetzle, P.E.M. Siegbahn, T.B. Richardson, R.H. Crabtree, J.
Am. Chem. Soc. 121 (1999) 6337–6343.
[10] S.J. Grabowski, W.A. Sokalski, J. Leszczynski, Chem. Phys. 337 (2007) 68–76.
[11] (a) Y.-J. Lee, J.-D. Lee, H.-J. Jeong, K.-C. Son, J. Ko, M. Cheong, S.O. Kang,
Organometallics 24 (2005) 3008–3019;
Formula
C37H25BF15N –
C6D6
863.53
C29H28BF10N
Formula weight
T (K)
Wavelength (Å)
Crystal system
Space group
a (Å)
591.33
123(2)
0.71073
Monoclinic
P21/c
15.335(2)
11.451(1)
14.863(1)
90.59(1)
2609.8(4)
4
123(2)
0.71073
Monoclinic
P21/c
16.612(1)
9.962(1)
22.626(2)
94.55(1)
3732.5(5)
4
b (Å)
c (Å)
b (°)
V (Å3)
Z
Dcalc (Mg mꢁ3
)
1.537
0.141
1744
1.505
0.137
1216
l
(mmꢁ1
F(000)
)
(b) CCDC 215487 (1).
[12] R.W. Parry, D.R. Schultz, P.R. Girardot, J. Am. Chem. Soc. 80 (1958) 1–3.
[13] (a) Theoretical calculations suggest very complicated proton dynamics,
already for the molecular NH4BH4, see: L.Ya. Baranov, O.P. Charkin, J. Struct.
Chem. 30 (1989) 721;
Crystal size (mm)
Crystal color
hmax (°)
Reflections collected
Independent reflections
Rint
Observed reflections
Goodness-of-fit on F2
Parameters
0.35 ꢃ 0.25 ꢃ 0.15 0.50 ꢃ 0.30 ꢃ 0.15 mm
Colorless crystals
27.5
Colorless crystals
27.5
40288
5979
0.0029
4712
1.056
377
0.321/ꢁ0.237
34901
8523
0.0344
5963
1.038
550
0.293/ꢁ0.245
(b) I. Alkorta, J. Elguero, C. Foces-Foces, J. Chem. Soc., Chem. Commun. (1996)
1633–1634.
[14] R. Roesler, W.E. Piers, M. Parvez, J. Organomet. Chem. 680 (2003) 218–222.
[15] W.B. Tolman, in: J.W. Tye, Y. Darensbourg, M.B. Hall (Eds.), Activation of Small
Molecules, Wiley-VCH, Weinheim, Germany, 2006, pp. 121ꢁ158.
[16] (a) G.C. Welch, R.R.S. Juan, J.D. Masuda, D.W. Stephan, Science 314 (2006)
1124–1126;
Largest difference in peak and hole
(b) H. Wang, R. Fröhlich, G. Kehr, G. Erker, Chem. Commun. (2008) 5966–
5968;
(c) M. Ullrich, A.J. Lough, D.W. Stephan, J. Am. Chem. Soc. 131 (2009) 52–
53;
(e Åꢁ3
)
(I))
R1 (I > 2
r
0.0448
0.0919
0.0368
0.0976
wR2 (all data)
(d) G.C. Welch, D.W. Stephan, J. Am. Chem. Soc. 129 (2007) 1880–1881;
(e) P. Spies, G. Erker, G. Kehr, K. Bergander, R. Froehlich, S. Grimme, D.W.
Stephan, Chem. Commun. (2007) 5072–5074;
(f) S.J. Geier, T.M. Gilbert, D.W. Stephan, J. Am. Chem. Soc. 130 (2008) 12632–
12633;
charged with 5 (0.1 mmol, 77.9 mg) and 1.0 mL of dry deuterated
toluene (d8). The reaction was degassed once with a freeze–
pump–thaw cycle, stirred at 1000 rpm and at 110 °C for 36 h.
NMR experiments (1H, 11B, 13C and 19F) of the solution showed a
50% conversion to trans-2,6-dimethyl-2,6-diphenylpiperidine and
B(C6F5)3. No side reactions were detected.
(g) D.P. Huber, G. Kehr, K. Bergander, R. Froehlich, G. Erker, S. Tanino, Y. Ohki,
K. Tatsumi, Organometallics 27 (2008) 5279–5284.
[17] (a) C.W. Hamilton, R.T. Baker, A. Staubitzc, I. Manners, Chem. Soc. Rev. 38
(2009) 279–293;
(b) T.B. Marder, Angew. Chem., Int. Ed. 46 (2007) 8116–8118.
[18] (a) A. Staubitz, M. Besora, J.N. Harvey, I. Manners, Inorg. Chem. 47 (2008)
5910–5918;
(b) J.T. Tanskanen, M. Linnolahti, A.J. Karttunen, T.A. Pakkanen, J. Phys. Chem.
C 112 (2008) 2418–2422.
[19] (a) V. Sumerin, F. Schulz, M. Nieger, M. Leskelä, T. Repo, B. Rieger, Angew.
Chem., Int. Ed. 47 (2008) 6001–6003;
(b) CCDC 678875 (4).
[20] (a) T.A. Rokob, A. Hamza, A. Stirling, T. Soos, I. Papai, Angew. Chem., Int. Ed. 47
(2008) 2435–2438;
(b) T.A. Rokob, A. Hamza, A. Stirling, I. Papai, J. Am. Chem. Soc. 131 (2009)
2029–2036;
(c) Y. Guo, S.-H. Li, Inorg. Chem. 47 (2008) 6212–6219.
[21] (a) J.S.J. McCahill, G.C. Welch, D.W. Stephan, Angew. Chem., Int. Ed. 46 (2007)
4968–4971;
8.2.6. Dehydrogenation of 9
In a glove box, a 25-mL flame-dried Schlenk tube equipped with
a stir bar, a Teflon stopcock and a glass stopper (GlindemannÒ-
sealing rings were used for conical joints instead of grease) was
charged with 9 (0.1 mmol, 59.1 mg), 1.0 mL dry toluene. The reac-
tion was degassed once with a freeze–pump–thaw cycle, stirred at
1000 rpm and at 110 °C for 20 h. All volatiles were removed in va-
cuo to give 58.9 mg of 8 (yield 100%) as an orange oil.
8.3. Crystallographic studies
(b) G.C. Welch, L. Cabrera, P.A. Chase, E. Hollink, J.D. Masuda, P.-R. Wei, D.W.
Stephan, Dalton Trans. (2007) 3407–3414;
(c) D.W. Stephan, Org. Biomol. Chem. 6 (2008) 1535–1539.
[22] P.A. Chase, T. Jurca, D.W. Stephan, Chem. Commun. (2008) 1701–1703.
[23] N–H distances are non-normalized.
The single-crystal X-ray diffraction studies of 5 and 9 were car-
ried out on a Bruker-Nonius Kappa-CCD diffractometer at 123(2) K
using Mo Ka radiation (k = 0.71073 Å). Direct methods (SHELXS-97)
[24] CCDC 713967 (5); see experimental part.
were used for structure solution, and full-matrix least-squares
refinement on F2 (SHELXL-97) [31]. H atoms were localized by differ-
ence Fourier synthesis and refined using a riding model (H(N) and
H(B) free) (see Table 2).
[25] Neutron diffraction experiments are underway.
[26] Inclusion or omission of solvent effects using a Polarized Continuum Model
(PCM) (for benzene) had no major effect.
[27] (a) A. Sawicka, P. Skurski, R.R. Hudgins, J. Simons, J. Phys. Chem. B 107 (2003)
13505–13511;
(b) I. Anusiewicz, J. Berdys-Kochanska, P. Skurski, J. Simons, J. Phys. Chem. A
110 (2006) 1261–1266.
[28] (a) P.A. Chase, G.C. Welch, T. Jurca, D.W. Stephan, Angew. Chem., Int. Ed. 46
(2007) 8050–8053;
Acknowledgements
(b) D. Chen, J. Klankermayer, Chem. Commun. (2008) 2130–2131;
(c) P. Spies, S. Schwendemann, S. Lange, G. Kehr, R. Fróhlich, G. Erker, Angew.
Chem., Int. Ed. 47 (2008) 7543–7546.
Support was provided by the DAAD (D/05/51658), Academy of
Finland (123248), the Finnish CoE of Computational Molecular Sci-
ence, Magnus Ehrnrooth Foundation (C.W.), Academy of Finland,
and JSPS (M.A.). CSC, Espoo provided computing resources.
[29] CCDC 713968 (9); see experimental part.
[30] (a) J. Einhorn, C. Einhorn, F. Ratajczak, A. Durif, M.-T. Averbuch, J.-L. Pierre,
Tetrahedron Lett. 39 (1998) 2565–2568;
(b) S. Cicchi, M. Bonanni, F. Cardona, J. Revuelta, A. Goti, Org. Lett. 5 (2003)
1773–1776.
References
[31] L.D. Lange, J.Y. Corey, N.P. Rath, Organometallics 10 (1991) 3189–3196.
[32] D.J. Parks, W.E. Piers, G.P.A. Yap, Organometallics 17 (1998) 5492–5503.
[33] G.M. Sheldrick, Acta Crystallogr. A64 (2008) 112–122.
[1] G.R. Desiraju, T. Steiner, The Weak Hydrogen Bond in Structural Chemistry and
Biology;, Oxford University Press, Chichester, UK, 1999.