Thermally induced dehydrogenation of amine-borane adducts and ammonia-borane by group 6 cyclopentadienyl complexes having single and triple metal-metal bonds

Article abstract of DOI:10.1002/ejic.201300629

Treatment of solutions of ammonia-borane (NH₃·BH ₃, AB) with catalytic amounts (5 mol-%) of the singly bonded dimers [M₂Cp₂(CO)₆] [M = Cr (1a), Mo (1b), W (1c); Cp = cyclopentadienyl] under mild thermal activation (333 K) led to the progressive dehydrogenation of the adduct and quantitative conversions were achieved after 12, 24, and >34 h, respectively. At the initial stages of these reactions (low conversions), the major products were cyclic and branched oligomers of aminoborane (NH₂=BH₂). However, at longer reaction times (high conversions), the major products were, in all cases, borazine, [HNBH] ₃, and polyborazylene, [NBH_x] (x < 1), whereas other minor products were derived from B-N bond-cleavage processes. Over the course of these reactions, complexes 1a-c were transformed into the corresponding mononuclear hydrides [MCpH(CO)₃] [M = Cr (2a), Mo (2b), W (2c)], which are supposed to be the catalytically active species in these processes, as also supported by similar catalytic activity exhibited by pure samples of the dihydride [Mo₂Cp₂(H)₂(μ-Ph ₂PCH₂PPh₂)(CO)₂] (2b′). Under similar conditions, 1a-c were also active catalysts for the dehydrogenation of adducts derived from substituted amines (tBuH₂N·BH₃ and Me₂HN·BH₃), although the rate of dehydrogenation was significantly lower than that of AB. This lower activity follows from deprotonation of hydrides 2 by the free amines, which are in turn generated through B-N bond-cleavage processes. The dehydrogenation products of tBuH₂N·BH₃ are also derived from oligomerization processes of the corresponding aminoborane (tBuHN=BH₂), which in this case was identified in the reaction mixtures, but even at long reaction times, the formation of the borazine-like product was not complete, and the reaction mixture contained significant amounts of (poorly defined) soluble polymeric materials. For Me₂HN·BH₃, the major product obtained in all of the reactions was cyclic dimer [Me₂N=BH ₂]₂. Similar studies were performed with triply bonded complexes [Mo₂Cp₂(CO)₄] (3b) and [Mo ₂Cp₂(μ-Ph₂PCH₂PPh ₂)(CO)₂] (3b′), which displayed similar catalytic activity while remaining essentially unperturbed along the reactions, and these complexes yielded product distributions that were similar to those observed for singly bonded dimers 1a-c. Readily accessible group 6 binuclear cyclopentadienyl complexes having single and triple M-M bonds are efficient catalysts for the dehydrogenation of a range of amine-borane adducts, including ammonia-borane, under mild thermal activation (333 K). Copyright

Full text of DOI:10.1002/ejic.201300629

FULL PAPER
DOI:10.1002/ejic.201300629
Thermally Induced Dehydrogenation of Amine–Borane
Adducts and Ammonia–Borane by Group 6
Cyclopentadienyl Complexes Having Single and Triple
Metal–Metal Bonds
Daniel García-Vivó,*[a] Estefanía Huergo,[a] Miguel A. Ruiz,[a] and
Raquel Travieso-Puente[a]
Keywords: Boranes / Amines / Dehydrogenation / Metal–metal bonds / Group 6 elements
Treatment of solutions of ammonia–borane (NH3·BH3, AB)
nation of adducts derived from substituted amines
with catalytic amounts (5 mol-%) of the singly bonded dimers
(tBuH2N·BH3 and Me2HN·BH3), although the rate of dehy-
[M2Cp2(CO)6] [M = Cr (1a), Mo (1b), W (1c); Cp = cyclopen- drogenation was significantly lower than that of AB. This
tadienyl] under mild thermal activation (333 K) led to the pro-
gressive dehydrogenation of the adduct and quantitative
conversions were achieved after 12, 24, and Ͼ34 h, respec-
tively. At the initial stages of these reactions (low conver-
sions), the major products were cyclic and branched oligo-
mers of aminoborane (NH2=BH2). However, at longer reac-
tion times (high conversions), the major products were, in all
cases, borazine, [HNBH]3, and polyborazylene, [NBHx] (xϽ
1), whereas other minor products were derived from B–N
bond-cleavage processes. Over the course of these reactions,
complexes 1a–c were transformed into the corresponding
mononuclear hydrides [MCpH(CO)3] [M = Cr (2a), Mo (2b),
W (2c)], which are supposed to be the catalytically active spe-
cies in these processes, as also supported by similar catalytic
activity exhibited by pure samples of the dihydride
[Mo2Cp2(H)2(μ-Ph2PCH2PPh2)(CO)2] (2bЈ). Under similar
lower activity follows from deprotonation of hydrides 2 by
the free amines, which are in turn generated through B–N
bond-cleavage processes. The dehydrogenation products of
tBuH2N·BH3 are also derived from oligomerization processes
of the corresponding aminoborane (tBuHN=BH2), which in
this case was identified in the reaction mixtures, but even at
long reaction times, the formation of the borazine-like prod-
uct was not complete, and the reaction mixture contained
significant amounts of (poorly defined) soluble polymeric ma-
terials. For Me2HN·BH3, the major product obtained in all of
the reactions was cyclic dimer [Me2N=BH2]2. Similar studies
were performed with triply bonded complexes [Mo2Cp2-
(CO)4] (3b) and [Mo2Cp2(μ-Ph2PCH2PPh2)(CO)2] (3bЈ), which
displayed similar catalytic activity while remaining essen-
tially unperturbed along the reactions, and these complexes
yielded product distributions that were similar to those ob-
conditions, 1a–c were also active catalysts for the dehydroge- served for singly bonded dimers 1a–c.
gen because of their high gravimetric hydrogen capacity,
Introduction
stability, and controlled hydrogen release. Therefore, it is
not surprising that these “simple” adducts have attracted
much research recently,[3] including numerous studies re-
lated to the development of low-energy regeneration routes
of the B–N spent fuels.[4] The interest in these substances
also stems from their potential as precursors of new inor-
ganic materials (i.e., polymers,[5] ceramics,[6] etc.) and as re-
ducing and hydroborating reagents in organic chemistry.[7]
Prompted by all of these promising applications, the devel-
opment of novel transition-metal-based catalysts able to
promote and control the dehydrocoupling/dehydrogenation
reactions of amine–borane adducts has bloomed in recent
years.[3] Most of the systems used so far are complexes
based on second- and third-row transition metals including
Rh,[8] Ir,[9] Ru,[10] Re,[11] and Pd,[12] which are expensive and
in some cases act in a heterogeneous manner rather than a
homogeneous manner. More scarce are the catalytic sys-
One of the biggest technological challenges of modern
societies is the development of cleaner energy sources that
can mitigate the current dependence on fossil fuels and alle-
viate the greenhouse effect associated with carbon-based
technologies. Over the last decades, hydrogen has appeared
as a long-term sustainable solution able to meet our future
energy demands;[1] however, the safe and economical stor-
age of hydrogen remains one of the major barriers for its
broad implementation.[2] In this context, amine–borane ad-
ducts, and in particular ammonia–borane (NH3·BH3, AB),
are promising candidates for the chemical storage of hydro-
[a] Departamento de Química Orgánica e Inorgánica / IUQOEM,
Universidad de Oviedo,
33071 Oviedo, Spain
E-mail: garciavdaniel@uniovi.es
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300629.
Eur. J. Inorg. Chem. 2013, 4998–5008
4998
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tems based on first-row transition metals (i.e., Ti,[13] Cr,[14]
Results
Ni,[15] and Fe[16] catalysts). Unfortunately, many of these
systems suffer from instability under the reaction condi-
tions, which further complicates the mechanistic landscape
of these reactions. Quite relevant to this work, Manners
et al. recently reported the dehydrocoupling reactions of
Dehydrogenation Reactions of AB Catalyzed by Complexes
1a–c (M–M) and 3b (MoϵMo)
Following a procedure similar to that used by Manners
primary and secondary amine–borane adducts and of AB and co-workers for iron dimer [Fe2Cp2(CO)4],[16a] we
itself by using the [Fe2Cp2(CO)4] (Cp = cyclopentadienyl) started our study by adding catalytic amounts (5 mol-%)
iron dimer under photochemical activation.[16a] Elegant of singly bonded complexes 1a–c to THF solutions of the
parallel experiments excluded the presence of hetero- unsubstituted adduct, NH3·BH3 (AB, 0.23 m). The reaction
geneous process, which suggests the formation of the rather progress was evaluated in situ by 11B and 11B{1H} NMR
unstable mononuclear hydride [FeCpH(CO)2] over the spectroscopy, and all of the dehydrogenation products were
course of these reactions; however, such a species was only identified by comparison of their NMR resonances with
detected in stoichiometric experiments and was not iden- those reported previously in the literature. Although no sig-
tified in the catalytic reactions.
nificant reaction was observed at room temperature, pro-
Encouraged by these results, we reasoned that the related gressive dehydrogenation was observed if these mixtures
binuclear [M2Cp2(CO)6] complexes of group 6 metals [M = were heated at 333 K to give mixtures of products
Cr (1a), Mo (1b), and W (1c)] might display significant (Scheme 1 and Table 1). All of the complexes under study
catalytic activity in these sorts of dehydrogenative processes. displayed significant catalytic activity (reactions faster than
In particular, these dimers are known to undergo M–M the blank run) and yielded almost-quantitative conversions
bond homolysis to generate the 17-electron [MCp(CO)3]· after 12, 24, and Ͼ34 h for 1a, 1b, and 1c, respectively. It
radicals in a reaction that takes place spontaneously in was also evident that the metal center has a strong influence
solution at room temperature for the Cr substrate[17] and on the activity, as the activity decreased for heavier metals,
under photochemical conditions (or strong thermal acti- which is an observation that might be rationalized on the
vation) for the Mo and W complexes.[18] Moreover, the cor- basis of the different strengths of the M–M and M–H
responding hydrides, [MCpH(CO)3] (2a–c), are relatively bonds in the species involved in these reactions (vide infra).
stable molecules that have been extensively studied,[19] and In spite of the different rates, the products obtained in all
they are known to undergo dehydrogenation processes un- of these reactions were quite similar, and their relative ratio
der mild conditions to regenerate the corresponding binu- evolved with the reaction time, as expected. Thus, in the
clear complexes.[20] Herein, we report that group 6 binuclear initial stages of the reaction, the major dehydrogenation
dimers 1a–c are effective catalysts, via mentioned mononu- products observed in the presence of heavier congeners 1b
clear hydrides 2, for the dehydrocoupling of a range of and 1c were those following from the formal loss of one
amine–borane adducts including AB under mild thermal equivalent of H2, in particular the cyclolinear trimer, B-
conditions (333 K). To extend the scope of this study, we (cyclodiborazanyl)aminoborohydride (BCDB),[15b] and the
also explored the catalytic activity of the readily accessible cyclic trimer (CTB).[23] These two species are intermediates
30-electron complex [Mo2Cp2(CO)4] (3b)[21] and that of the in the dehydrogenation reaction, as prolonged reaction
dppm-bridged derivatives [Mo2Cp2(μ-dppm)(CO)n] [dppm times led to quantitative conversions of the parent adduct
= Ph2PCH2PPh2; n = 4 (1bЈ), 2 (3bЈ)].[22] Although the tri- and, most importantly, to the formation of borazine (BH–
ply bonded complexes have stronger intermetallic bonds, NH)3 (formal loss of 2 equiv. of H2)[8a] and polyborazylene
the presence of the diphosphane ligand is expected to in- (BNHx, xϽ 1, PB,[15b] loss of Ͼ2 equiv. of H2) as major
crease the electron density and increase the stability of these dehydrogenation products. However, chromium complex 1a
complexes towards degradation of the dimetal center. As displayed a significantly higher reaction rate, and even at
discussed below, we found that the triply bonded complexes short reaction times the major dehydrogenation product
exhibit higher catalytic activity than their corresponding was borazine. Notably, besides these major dehydrogenation
singly bonded analogues in some of the reactions examined, products, we could also identify the presence of μ-aminodi-
which proves, for the first time, that complexes with mul- borane, B2H5(μ-NH2) (μ-ADB), in the reaction mixtures[24]
tiple M–M bonds can be efficient catalysts for the dehydro- and small amounts of the BH3·THF adduct, both of which
genation of amine–borane adducts.
should be formed through B–N bond-cleavage processes of
Scheme 1.
Eur. J. Inorg. Chem. 2013, 4998–5008
4999
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 1. Thermal dehydrogenation of AB catalyzed by complexes 1a–c and 3b (5 mol-%).[a,b]
Complex
t [h]
Product distribution [%]
Conv. [%]
TOF [h–1][c]
μ-ADB
BH3·THF BCDB
CTB[d]
Borazine
PB[e]
1a
6
34
26
19
23
19
24
4
2
2
3
0
1
2
0
0
0
0
0
0
1
3
0
10
7
1
41
25
9
56
63
41
33
54
32
28
7
20
11
4
19
10
9
36
19
19
17
23
22
16
13
24
30
49
53
13
34
49
4
4
5
20
4
11
7
0
2
6
8
3
8
9
18
0
81
94
100
40
78
97
14
40
66
80
42
65
92
99
20
12
24
6
12
24
6
12
24
34
6
12
24
34
24
0.8
1b
1c
0.4
6
9
10
13
7
11
11
12
0
24
29
13
27
35
47
4
0.3
0.8
3b
Blank
72
[a] Reactions were run with [cat] = 0.012 m and [AB]0 = 0.23 m in THF (0.5 mL). The conversion and product distribution were determined
by integration of the corresponding signals in the 11B{1H} NMR spectra. [b] Reaction times were not necessarily optimized. [c] Turnover
frequency (TOF) = (%conversion/%loading)/time, with %loading = 10 (for 1a–c) and 5 (for 3b), see text. [d] The corresponding signal
in the 11B{1H} NMR spectrum overlaps one of the resonances of BCDB, and therefore, the percentage was calculated by difference.
[e] The product distribution was calculated on the basis of the corresponding monomer (-BNHx-) (xϽ 1).
the parent adduct.[15b] Actually, the former was recently of singly bonded 1b, as illustrated by the effective TOF val-
prepared in high yield from the reaction of AB with ues at 24 h (0.8 vs. 0.4 h–1).
BH3·THF.[24c]
Triply bonded tetracarbonyl 3b also catalyzed the dehy-
drogenation of AB under the same reaction conditions.
Quite remarkably, this complex achieved quantitative con-
versions after approximately 24 h, which led to a reaction
Dehydrogenation Reactions of AB Catalyzed by
Diphosphane-Bridged Complexes
rate and product distribution similar to those of singly
Treatment of singly bonded dppm-bridged complex 1bЈ
bonded 1b (vide infra), although the formation of products with 20 equiv. (5 mol-% cat.) of AB under the same condi-
derived from B–N bond-cleavage processes were partially tions as those used for complexes 1a–c did not lead to sig-
suppressed.
nificant dehydrogenation of the adduct. Instead, the metal
The fate of the metal complexes over the course of the complex was transformed completely into a mixture of
1
above reactions was examined by H NMR and IR spec- products containing three known complexes as major orga-
troscopy, and it turned out to be different for the singly nometallic species: triply bonded dicarbonyl 3bЈ (Figure 1),
and triply bonded complexes. The former compounds were phosphanylmethyl complex [Mo2Cp2(μ-Ph2PCH2)(μ-PPh2)-
found in all cases to transform into corresponding mononu- (CO)2],[22b] and dihydride [Mo2Cp2(H)2(μ-dppm)(CO)4]
clear hydrides 2. In fact, for chromium complex 1a this (2bЈ).[22a] Whereas the first two complexes are genuine prod-
transformation takes place spontaneously upon mixing with ucts of the thermal decarbonylation of 1bЈ in the absence
AB at room temperature. In contrast, for the heavier conge- of AB,[22b] the dihydride complex is not, and its formation
ners no reaction was observed at room temperature, and should follow a pathway similar to that of mononuclear hy-
the corresponding hydrides were only formed after heating drides 2a–c, as discussed below. To check if the low conver-
these mixtures. Thus, molybdenum dimer 1b was converted sions observed in the reaction of 1bЈ were derived from
completely into hydride 2b only after 12 h of reaction, small amounts of 2bЈ generated during the reaction, we de-
whereas the tungsten compound was converted into its hy- cided to test the catalytic activity of pure samples of dihy-
dride after nearly 34 h. In contrast, triply bonded complex dride 2bЈ (Table 2). As it can be seen, this was the case,
3b remained essentially unperturbed over the course of the as the catalytic activity of 2bЈ at short reaction times was
dehydrogenation reaction. Of course, this implies the occur- comparable to that of hexacarbonyl dimer 1b; however, at
rence of different reaction mechanisms for the singly and high conversions the rate of dehydrogenation became slower
triply bonded complexes, as discussed below. Moreover, the and more than 46 h was required to achieve quantitative
formation of mononuclear hydrides 2 as the active species conversion of the adduct. Although the dehydrogenation
in the reactions of complexes 1a–c implies that these reac- products are similar to those observed in the reactions of
tions are actually running at a higher (ideally 10 mol-%) complexes 1a–c and 3b, there are some significant differ-
catalyst loading. Therefore, it can be guessed that the ac- ences if dihydride 2bЈ is used as the catalyst. First, a com-
tivity of triply bonded 3b (which is actually running at paratively lower amount of borazine is generated at high
5 mol-% loading) would be comparatively higher than that conversions (i.e., 28% at 46 h vs. 49% at 24 h for 1b), which
Eur. J. Inorg. Chem. 2013, 4998–5008
5000
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 2. Thermal dehydrogenation of AB catalyzed by complexes 2bЈ and 3bЈ (5 mol-%).[a,b]
Complex t [h]
Product distribution [%]
dppm·BH3 μ-ADB
Conv. [%] TOF [h–1][c]
BH3·THF BCDB
CTB[d]
Borazine
PB[e]
2bЈ
3bЈ
6
5
7
5
8
7
5
5
6
0
0
1
3
5
0
0
3
0
0
0
0
0
0
0
0
76
55
39
30
28
61
48
34
12
22
31
31
31
15
26
30
5
2
4
6
8
7
6
6
7
46
67
12
24
34
46
6
12
24
12
18
20
22
13
15
20
78
84
92
43
56
75
0.3
0.2
0.6
[a] Reactions were run with [cat] = 0.012 m and [AB]0 = 0.23 m in THF (0.5 mL). The conversion and product distribution were determined
by integration of the corresponding signals in the 11B{1H} NMR spectra. [b] Reaction times were not necessarily optimized. [c] TOF =
(%conversion/%loading)/time, with %loading = 10 (for 2bЈ) and 5 (for 3bЈ). [d] The corresponding signal in the 11B{1H} NMR spectrum
overlaps one of the resonances of the BCDB, and therefore, the percentage was calculated by difference. [e] The product distribution was
calculated on the basis of the corresponding monomer (-BNHx-) (xϽ 1).
implies that the dehydrogenation of the intermediate species sions the reaction mixture contained higher amounts of
(i.e., BCDB and CTB) is significantly slower in this case. borazine and PB (65 vs. 27%).
Second, the formation of products derived from B–N bond-
cleavage processes is also comparatively lower, in particular
Dehydrogenation Reactions of tBuNH2·BH3 (tBuAB)
that of μ-ADB [≈3% for 2bЈ vs. 24% for 1b]. Finally, al-
though complex 2bЈ remains essentially unaltered over the
course of the reaction, a minor decomposition pathway
leads to the liberation of a small amount of dppm, which
under the working conditions is rapidly transformed into
the corresponding BH3 adduct (dppm·BH3).[25]
To evaluate the scope of amine–borane adducts that
might be efficiently dehydrogenated with the binuclear com-
plexes under study, we tested the catalytic activity of some
of these complexes (the most active ones) in the dehydroge-
nation of the primary tBuNH2·BH3 (tBuAB) adduct. In
agreement with previous studies, we found that this adduct
is significantly more thermally robust than AB,[3g] as a 1.3 m
solution of tBuAB in THF remained stable under the reac-
tion conditions used in our work (333 K). However, ad-
dition of catalytic amounts (5mol-%) of complexes 1a, 1b,
2bЈ, 3b, and 3bЈ[26] to these solutions induced dehydrogena-
tion of tBuAB to yield a mixture of products (Scheme 2
and Table 3). We note, however, that these reactions were
comparatively slower than those observed for AB.
Figure 1. Complexes used in this work.
The second diphosphane-bridged complex tested as a
catalyst was triply bonded [Mo2Cp2(μ-dppm)(CO)2]
(3bЈ),[22,26] which is a diphosphane-substituted analogue of
Scheme 2.
In terms of conversion, chromium complex 1a was found
3b. This complex also promoted the dehydrogenation of AB to be again the most active catalyst among the electron-
(Table 2). In the initial stages of the reaction, this complex precise complexes, and it achieved nearly quantitative con-
induced conversions that were slightly higher than those versions after roughly 50 h of reaction. Analogous molyb-
found for 3b, but both complexes required approximately denum complex 1b was less active, with only about 62%
24 h to achieve almost quantitative conversions. However, conversion after 44 h. Even lower activity was exhibited in
even if the conversions were similar, the degree of dehydro- this case by diphosphane-bridged dihydride 2bЈ, for which
genation was higher for tetracarbonyl 3b, as at high conver- less than 3% conversion was achieved after 12 h, and there-
Eur. J. Inorg. Chem. 2013, 4998–5008
5001
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 3. Thermal dehydrogenation of tBuAB catalyzed by complexes 1a, 1b, 2bЈ, 3b, and 3bЈ (5 mol-%).[a,b]
Complex
t
Product distribution [%]
Conv. TOF[d]
dppm·
BH3
B2H4(μ-
NHtBu)(μ-H)
(H2B–
NHtBu)n
Poly. Poly. (H2B–
(HB–
NtBu)3
[h]
HB(NHtBu)2 tBuHN=BH2
[%]
[h–1
]
[c]
I[c]
II[c]
NHtBu)2–3
1a
1b
14
44
148
14
44
50
62
12
14
108
12
24
70
–
–
–
–
–
–
–
0
–
–
1
1
1
37
24
4
50
28
23
15
0
52
16
18
13
4
8
4
1
0
4
3
1
1
0
8
2
1
1
2
10
30
29
6
38
42
49
0
0
3
–
–
0
2
9
0
0
0
1
0
0
0
0
0
1
8
33
10
1
31
9
9
6
0
8
32
20
34
43
0
5
24
0
2
9
10
0
0
16
0
42
82
94
20
62
72
90
3
16
11
4
12
22
5
4
2
2
0
0
14
25
21
20
0.06
16
14
16
100
32
17
35
27
17
0.15
0.16
0.26
2bЈ
3b
9
88
46
62
90
3bЈ
3
12
0
[a] Reactions were run with [cat] = 0.07 m and [tBuAB]0 = 1.3 m in THF (0.7 mL). Conversion and product distribution were determined
by integration of the corresponding signals in the 11B{1H} NMR spectra. [b] Reaction times were not necessarily optimized. [c] Product
distribution was calculated on the basis of the corresponding monomer. [d] TOF = (%conversion/%loading)/time, with %loading = 10
(for 1a, 1b, and 2bЈ) and 5 (for 3b and 3bЈ).
fore, it was not further explored in this reaction. As for the tion pathway (by progressively reducing the amount of hy-
triply bonded complexes, tetracarbonyl 3b required longer drides 2 in solution). In contrast, triply bonded complexes
reaction times than singly bonded 1b, with conversions 3b and 3bЈ remained essentially unperturbed over the
close to 90% after approximately 108 h of reaction; in con- course of these reactions in a similar way to that observed
trast, related diphosphane-bridged complex 3bЈ exhibited for the AB dehydrogenations, which thus allows the pres-
remarkably high activity, with quantitative conversions af- ence of roughly constant amounts of active metallic species
ter about 70 h, and it was the most active catalyst among in the solutions.
all of the binuclear compounds here examined.
The dehydrogenation products formed in the above reac-
Dehydrogenation Reactions of Me2HN·BH3 (MeAB)
tions (Scheme 2) were slightly different from those observed
for the AB reactions, although they are similar to those
reported for other catalytic systems.[14,27] The most striking
differences are the identification of small amounts of the
simplest dehydrogenation product, the aminoborane
tBuHN=BH2,[14] and the presence of poorly defined, pre-
sumably polymeric species [11B NMR: δ = 41 (polymer I)
and 21 ppm (polymer II)].[27a] Notably, aminoboranes are
key intermediates in the formation of other reaction prod-
ucts by oligomerization[16a] or redistribution reactions.[28] In
any case, these products were also accompanied by the cor-
responding borazine-like and cycloborazane products, and
the former was favored at long reaction times but was never
obtained quantitatively. As found in the AB reactions, we
also identified some products derived from B–N bond-
cleavage processes, such as μ-aminodiborane [B2H4(μ-
NHtBu)(μ-H)] and diaminoborane [HB(NHtBu)2],[14b]
which were formed in significant amounts.
A second critical difference in the reactions of unsubsti-
tuted AB was found for the metallic species present in the
reaction media. In the tBuAB reactions, singly bonded
complexes 1a and 1b were again completely transformed
after 14 h, but the major organometallic species present
were mononuclear anions [MCp(CO)3]– (4a, 4b) along with
small amounts of expected hydrides 2a and 2b. As discussed
below, these anions are inactive in the dehydrogenation of
amine–borane adducts, and therefore, their formation un-
der the catalytic conditions is a drawback to the use of com-
The next step was to test our complexes in the dehydro-
genation of borane adducts of secondary amines, for which
we used Me2NH·BH3 (MeAB). This adduct has been
widely used in studies of this type, because in most cases its
dehydrogenation is quite simple and yields almost exclu-
[13c]
sively the cyclic aminoborane [H2BNMe2]2
and possible
reaction intermediates usually remain soluble.[3]
As observed for tBuAB, THF solutions of MeAB (1.3 m)
remained stable under our particular reaction conditions,
but dehydrogenation could be induced by the dimetal com-
plexes (Table 4 and Scheme 3). Remarkably, the most active
Table 4. Thermal dehydrogenation of MeAB catalyzed by com-
plexes 1a, 1b, 2bЈ, 3b, and 3bЈ (5 mol-% catalyst loading).[a,b]
Complex
t
Product distribution [%]
Conv. TOF
[h–1 [c]
[h] Me2N=BH2 HB(NMe2)2 (H2B-NMe2)2 [%]
]
1a
1b
14
14
40
24
14
68
12
2
0
0
3
0
0
0
5
0
0
5
0
0
1
93
90
11
99
18
15
88
100
0.64
0.25
100
100
92
100
100
99
2bЈ
3b
0.26
1.67
3bЈ
[a] Reactions were run with [cat] = 0.07 m and [MeAB]0 = 1.3 m in
THF (0.7 mL). Conversion and product distribution were deter-
mined by integration of the corresponding signals in the 11B{1H}
NMR spectra. [b] Reaction times were not necessarily optimized.
[c] TOF = (%conversion/%loading)/time, with %loading = 10 (for
plexes 1a and 1b, because it actually represents a deactiva- 1a, 1b, and 2bЈ) and 5 (3b and 3bЈ).
Eur. J. Inorg. Chem. 2013, 4998–5008
5002
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
catalyst for the dehydrogenation of this adduct was again bonded complexes remained essentially stable over the
diphosphane-bridged triply bonded complex 3bЈ, which course of their reactions, and therefore, these species should
achieved quantitative conversions at comparatively short re- be the active catalysts in the corresponding dehydrogena-
action times (≈12 h).[26] As found in the other reactions dis- tion reactions. Thus, two alternative pathways should be
cussed above, chromium complex 1a was also remarkably formulated for these reactions, depending on the dimetal
active, and it provided almost quantitative conversions after complex being used.
only 14 h of reaction.
Proposed Reaction Mechanism for Singly Bonded
Complexes 1a–c and 2bЈ
For tricarbonyl dimers 1a–c, the catalytic reactions
should start with the formation of mononuclear hydrides 2
(Scheme 4). This can be fully rationalized by recalling the
well-established trend of the parent dimers to undergo
homolytic M–M bond cleavage to yield metal-centered rad-
Scheme 3.
icals of formula [MCp(CO)3] (A in Scheme 4).[17,18] These
In contrast, slower reactions were observed in the pres-
ence of molybdenum complexes 1b, 2bЈ, and 3b. In any case,
irrespective of the catalyst used, the major dehydrogenation
product was [H2BNMe2]2.[13c] For some of the complexes,
this was the only product observed in the reactions. Albeit
in very small amounts, two other dehydrogenation species
were observed for some of the systems: the monomeric ami-
noborane H2B=NMe2 (a precursor of the corresponding
cyclic dimer just mentioned)[13c,29] and the diaminoborane
HB(NMe2)2.[13c,30]
As for the metallic species present over the course of the
catalytic reactions, an analogous situation to that discussed
for the tBuAB reactions was found. Thus, the triply bonded
complexes remained essentially unperturbed over the course
of the reactions, as did dihydride complex 2bЈ. However,
singly bonded complexes 1a and 1b were completely trans-
formed into a mixture of mononuclear hydrides 2 and cor-
responding anions 4. In this case, however, the formation
of anions 4a and 4b was not as favorable as that found in
the tBuAB reactions, and in the reactions of molybdenum
complex 1b, hydride 2b remained as the major species,
whereas for the chromium complex an approximately equi-
molar mixture of 2a/4a was observed.
17-electron radicals are highly reactive, and they can evolve
through a number of reactions such as simple recombina-
tion (to give the parent dimers)[31] or hydrogen/halogen ab-
straction,[31a,32] among others. In our case, hydrogen ab-
straction from the amine–borane adducts would take place
rapidly owing to the large excess amount of these adducts
present in the initial stages of the reaction, which would
generate hydrides 2a–c (Scheme 4). By assuming that hydro-
gen abstraction is a fast process, the extent of the formation
of the corresponding hydrides would then be related to the
strength of the respective intermetallic bonds in the metal
dimers (Cr ϽϽ MoϽ W). This is actually supported by our
experimental observations, as fast conversion into hydride
Discussion
As noted in the preceding section, two different species
are involved in the dehydrogenation reactions studied. For
the singly bonded complexes, mononuclear hydrides 2 are
invariably formed in all of the reactions studied, in a way
similar to that observed by Manners and co-workers for
[Fe2Cp2(CO)4] in stoichiometric experiments.[16a] Moreover,
pure samples of dihydride 2bЈ exhibits activity similar to
that observed for molybdenum dimer 1b, but it remains es-
sentially unaltered. Also, in the reactions of adducts derived
from substituted amines, we observed the formation of sig-
nificant amounts of mononuclear anions 4, which were ac-
companied by a significant decrease in the rates of the dehy-
drogenation reaction. Therefore, it seems reasonable to as-
sume that the hydrides are either the true catalysts or the
catalyst precursors in these reactions. In contrast, the triply Scheme 4.
Eur. J. Inorg. Chem. 2013, 4998–5008
5003
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
2 takes place only for chromium dimer 1a (weakest M–M bond energies involved (N–H Ͼ B–H). Even if we have not
bond), whereas for the molybdenum and tungsten com- been able to identify this type of species in our reactions,
plexes this transformation only goes to completion at we note that the coordination mode of the boryl ligand in
longer reaction times. In addition, separate experiments C has been previously observed in related species,[3b] such
proved that molybdenum hydride 2b was formed rapidly as the mononuclear complexes of formula [MX(CO)3-
and quantitatively upon photochemical irradiation (10 min, (BH2PMe3)] [X = Cp, C5Me5 (Cp*); M = Mo, W] pre-
298 K) of a typical reaction mixture of AB and 1b (5 mol- viously reported by Shimoi and co-workers.[34] Interestingly,
% loading) without causing significant dehydrogenation of the latter complexes were prepared through photolysis of
the adduct; even more importantly, heating these mixtures the methyl derivatives [MX(CO)3Me] in the presence of the
led to a dehydrogenation reaction that was somewhat faster phosphane–borane adducts with concomitant evolution of
than that of the dark reactions.
CH4. Two other important observations were reported by
Notably, abstraction of hydrogen from amine– and phos- Shimoi: (1) The Cp derivatives were found to be rather ther-
phane–borane adducts by different radical species has been mally unstable. (2) The use of the Me3N·BH3 adduct in-
extensively studied by Roberts and co-workers,[33] who ob- stead of Me3P·BH3 did not yield stable boryl complexes.
served preferential activation of the B–H bonds of the Therefore, it seems reasonable that, under our reaction con-
amine–borane adducts in all cases to yield B-centered radi- ditions, intermediates C would be rather unstable species.
cals of formula [RxH3–xN-BH2]· (B in Scheme 4) as prod- In this case, a simple β-elimination step of the coordinated
ucts of kinetic control. Depending on the substituents and amine–boryl ligand would regenerate hydrides 2 while re-
on the particular experimental variables, these transient leasing aminoborane, in a process which would likely in-
amine–boryl species were observed to evolve through vari- volve reversible loss of CO to generate unsaturated species
ous pathways, which all led to the quenching of the boryl akin to those proposed for the related reactions of the
radical. Among these pathways, they observed the isomer- methyl derivative just mentioned. We should finally note
ization to aminyl–borane radicals (thermodynamically that, although all of the preceding discussion explains the
more stable) and the release of hydrogen to yield aminobor- formation of products in which only one equivalent of hy-
anes [RxH2–xN=BH2].[33] In a similar way, tricarbonyl radi- drogen is lost, it can be assumed that similar processes oc-
cals A would abstract hydrogen from the amine–borane ad- curring for already dehydrogenated species would then ac-
ducts to yield hydrides 2 and the corresponding monomeric count for the formation of products in which two or more
aminoboranes, which is consistent with the detection of sig- equivalents of hydrogen were released, such as borazine or
nificant amounts of these monomers in the catalytic reac- polyborazylene.
tions of the substituted amine–borane tBuAB and MeAB
We should stress that both pathways I and II (Scheme 4)
adducts. Moreover, the liberation of free aminoborane spe- would imply the consumption of the amine–borane adduct
cies would account for the formation of most of the dehy- with production of only hydrogen and the corresponding
drogenation products observed in our catalytic reactions aminoboranes as primary products, so we cannot rule out
through off-metal processes such as oligomerization [i.e., either mechanism. However, there are two experimental ob-
(RxH2–xN-BH2)n] by reaction with unreacted adduct (i.e., servations that would suggest the prevalence of pathway II.
BCDB) or by redistribution processes.[28] Such off-metal First, in the reaction with molybdenum complex 1b, even
processes of aminoboranes have been recently proposed to at long reaction times (when the concentration of unreacted
occur in reactions of photoactivated precatalysts amine–borane is low), the only metal-containing species in
[Fe2Cp2(CO)4][16a] and [M(CO)6] (M = Cr, Mo, W).[14a]
solution is hydride 2b (or its deprotonated anion 4b); how-
Up to this point, only stoichiometric conversion of the ever, if pathway I was to be dominant we would expect the
amine–borane adducts into aminoborane and hydrogen presence of the parent dimer (generated through radical
would be attained. To explain the observed full conversions coupling reactions) at long reaction times. Second, as men-
of the amine–borane adducts, a catalytic cycle should take tioned above, amine–boryl radicals are liberated in path-
place on the basis of hydrides 2. For this, however, two reac- way I rather than aminoboranes, and it is known that these
tion pathways are conceivable. First, hydrides 2 could ex- species can evolve through various reaction paths to yield
perience thermal dehydrogenation to regenerate 17-electron different products,[33] that is, radical homocoupling reac-
radicals A, which then would react again with more amine– tions; however, in all of our reactions, the formation of
borane to yield parent hydrides 2 (pathway I in Scheme 4). these products was not observed.
Indeed, hydrides 2a–c have been reported to undergo ther-
As noted above, for all of the borane adducts studied we
mal dehydrogenation to yield dimers 1a–c via radicals A.[20] observed the formation of products with B/N ratios dif-
As an alternative route (pathway II in Scheme 4), direct re- ferent from unity, which necessarily are formed through B–
action of hydrides 2a–c with the amine–borane adducts N bond-cleavage reactions. These products were observed
might liberate hydrogen to generate unstable amine-stabi- previously in other transition-metal-catalyzed dehydrogena-
lized boryl intermediate C (Scheme 4). Notably, we cannot tion reactions, and they are presumably formed through re-
exclude the alternative activation of an N–H bond of the actions of the free amine or borane entities with the amino-
amine–borane adduct to yield a borane-stabilized amido borane[15b] or amine–borane.[24c] However, in the reactions
complex; however, we reasoned that such a process would of complexes 1a–c, the presence of free amines has an ad-
be of higher energy by taking into account the different ditional negative consequence: the formation of variable
Eur. J. Inorg. Chem. 2013, 4998–5008
5004
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
amounts of anions 4, particularly in the reactions of the a coordination mode has been previously observed for the
tBuAB and MeAB adducts. As it can be seen from the data tetrahydroborate ligand[37] {i.e., [Mn2(μ-η2:η2-BH4)(μ-H)-
in Table 5, hydrides 2a–c are moderately acidic molecules (CO)6(μ-dppm)][37a] and [Mo2Cp2(μ-SMe)2(μ-CCH2Ph)(μ-
that are susceptible to undergo deprotonation if confronted κ1:κ1-BH4)][37b]}, and it has been recently substantiated in
with bases, and then, the presence of free amine in the reac- the borane- and boryl-bridged dirhodium cation [Rh2H2(μ-
tion media could explain the formation of anions 4 by sim- κ1:η2-H2BNMe3)2(μ-η2:η2-H3BNMe3)(PCy3)2]2+ [38]
In any
.
ple deprotonation (Scheme 5). Indeed, an independent ex- case, the coordination of these ligands (formally acting as
periment revealed that the addition of diisopropylamine to four-electron donors) converts intermediate D into a satu-
a freshly prepared solution of molybdenum hydride 2b led rated 34-electron complex with a single intermetallic bond.
immediately to the corresponding anion 4b. However, if we The next step might involve the oxidative addition of one
keep in mind that ammonia is only slightly less basic than of the coordinated B–H bonds to yield complex E with hy-
the substituted tBuNH2 or Me2NH amines, the fact that we dride and amine–boryl ligands akin to that proposed for
did not observe the formation of anions 4 in the reactions the singly bonded complexes. Although the presence of two
of AB must be attributed to its higher volatility, especially metal centers increases the number of possible coordination
in reactions running at 333 K, which would effectively re- modes of this group, a simple β-elimination step would lead
duce its presence in solution. In agreement with this, the again to the liberation of the corresponding aminoborane
reactions of the tBuAB adduct (which would liberate the and the generation of dihydride complex F. Such an inter-
amine with the highest boiling point) are those for which mediate would be related to [Mo2Cp*2(CO)4(μ-H)2], a com-
higher amounts of anions 4 were observed. In any case, the plex reported some time ago by Alt and co-workers as a
formation of these anions is always accompanied by a dras- product of the hydrogenation of [Mo2Cp*2(CO)4] under
tic decrease in rate of dehydrogenation.
photolytic conditions in a process found to be fully revers-
ible.[39] These authors noticed that no dihydride complex
was formed for the analogous Cp complex [Mo2Cp2-
(CO)4] under similar conditions, and therefore, it would not
be surprising that in our reactions putative dihydrides F
would quickly release hydrogen to regenerate catalytically
active species 3.
Table 5. pK values and boiling points for free amines and hydrides
2a–c.
Compound
pKa
pKb
b.p. [K]
[a]
NH3
4.79
3.55
3.36
2.95
240
317
280
357
–
–
–
[a]
tBuNH2
Me2NH[a]
iPr2NH[a]
[CrCpH(CO)3][b]
[MoCpH(CO)3][b]
[WCpH(CO)3][b]
13.3
13.9
16.1
[a] Values in H2O solution taken from ref.[35] [b] Values in NCMe
solution taken from ref.[36]
Scheme 5.
Proposed Reaction Mechanism for Triply Bonded
Complexes 3b and 3bЈ
Scheme 6.
As noted above, the catalytic reactions of the triply
bonded complexes should involve a mechanism that is dif-
Overall, this catalytic cycle would consume the amine–
ferent from the one just discussed for the singly bonded borane adducts to produce hydrogen and free aminobor-
compounds, as complexes 3b and 3bЈ undergo no net trans- anes as primary products. Further interaction of complexes
formation over the course of their catalytic reactions. Pre- 3b and 3bЈ with the oligomerization products of the latter
sumably, the first step in this case would be the simple coor- aminoboranes would be much slower if we take into ac-
dination of the amine–borane adduct to the dimetal center count the significantly slower dehydrogenations observed
through one or two of the B–H bonds (D in Scheme 6). for the tBuAB and MeAB adducts (relative to AB) in the
Clearly, this step would be greatly facilitated by the elec- presence of complexes 3. If we keep in mind that the first
tronic and coordinative unsaturation of the dimetal center step of the mechanism is the coordination of the amine–
in these 30-electron complexes. Although we could not borane adducts to the dimetal center, we can guess that all
identify any intermediate in our reactions, we note that such of this might have a steric origin, that is, a more difficult
Eur. J. Inorg. Chem. 2013, 4998–5008
5005
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
approach to the dimetal center because of unfavorable in- tives. This different behavior correlates with the corre-
teractions with the alkyl chains. sponding M–M bond strength (W Ͼ MoϾϾ Cr) in the
A related observation that might be counterintuitive is parent dimers. In the reactions with adducts derived from
that complex 3bЈ (having a bulky dppm bridging ligand) substituted amines, B–N bond dissociation of the adduct
displays higher dehydrogenation rates than its tetracarbonyl causes a significant decrease in the rate of dehydrogenation,
relative 3b for some adducts. However, we must recall that because the liberated amine deprotonates the hydride com-
although these two complexes are isoelectronic, they are not plexes to yield the inactive mononuclear anions [MCp-
isostructural: in complex 3b four carbonyl ligands bridge (CO)3]–, which then reduces the concentration of hydride
the dimetallic center, whereas in 3bЈ there are only two CO complexes in solution. In contrast, multiple metal–metal
bridging ligands and the diphosphane ligand occupies two bonding confers higher stability to the dinuclear systems in
essentially terminal positions in a forced cis disposition. these reactions, as deduced from the fact that the
Eventually, this renders a more accessible dimetal center in [Mo2Cp2(CO)2(L2)] [L2 = (CO)2 and μ-dppm] complexes re-
dppm-bridged complex 3bЈ.
mained essentially stable during the catalytic reactions. In
spite of this, these complexes displayed similar or even
higher activity than their singly bonded relatives in the de-
hydrogenation of some of the adducts studied, which proves
that the electronic and coordinative unsaturation associated
Conclusions
We have found that readily accessible group 6 binuclear with the triple intermetallic bonds have a beneficial effect
cyclopentadienyl complexes having single and triple M–M in terms of catalytic activity. Moreover, to the best of our
bonds are efficient catalysts for the dehydrogenation of a knowledge, those unsaturated complexes are the first com-
range of amine–borane adducts, including ammonia–bor- plexes with multiple M–M bonds shown to display catalytic
ane, under mild thermal activation (333 K). The activity ob- activity in the dehydrogenation of amine–borane adducts,
served for these complexes is comparable to that reported which then raises the question whether other complexes
by Manners and co-workers for the [Fe2Cp2(CO)4] iron di- with multiple intermetallic bonds might display similar or
mer under photochemical conditions[16a] with the advan- even higher activity in these types of processes, a question
tage that in our case these reactions occur under thermal that we are currently addressing in our laboratory.
conditions rather than photochemical activation. Although
these systems are not close to the fastest catalysts developed
Experimental Section
so far (based on Rh and Ir complexes, with TOF values
close to 6000 h–1),[3h,5a,8e] their activity is comparable to a
broad range of mononuclear transition-metal complexes
used to induce these transformations (with TOF values of
0.1–2.8 h–1).[3h]
General Procedures: All reactions and manipulations were per-
formed under a nitrogen atmosphere by using standard Schlenk or
glove box techniques. Solvents were purified according to literature
procedures[40] and distilled under a nitrogen atmosphere prior to
For all of the adducts studied, the dehydrogenation prod- use. NH3·BH3,[41] [M2Cp2(CO)6] [M
=
Cr (1a),
W
(1c)],[42]
(3b),[21]
[Mo2Cp2(μ-dppm)(CO)4]
(1bЈ),[22]
[Mo2Cp2(CO)4]
ucts are similar to those observed for other transition-metal
catalysts and their formation can be rationalized by the oc-
currence of two alternative reaction pathways: (1) neat de-
hydrogenation processes (dominant) and (2) B–N bond-
cleavage processes. The latter leads to products in which the
ratio B/N is different from unity, whereas the former implies
the metal-mediated generation of the corresponding (RxH2–
xN=BH2) aminoboranes as intermediate species, which
could be identified for the tBuNH2 and Me2NH adducts.
[Mo2Cp2(μ-dppm)(CO)2] (3bЈ),[22] and [Mo2Cp2H2(μ-dppm)(CO)4]
(2bЈ)[22a] were prepared as described previously. All other reagents
were obtained from the usual commercial suppliers and used as
received, except for the Me2NH·BH3 and tBuNH2·BH3 adducts,
which were sublimed twice and stored under a nitrogen atmosphere.
IR stretching frequencies were measured in solution by using CaF2
windows. NMR spectra were routinely recorded at 300.13 (1H),
121.50 (31P{1H}), and 161.47 MHz (11B and 11B{1H}) at 290 K in
THF solutions unless otherwise stated. To eliminate the B–O broad
Once formed, however, the aminoboranes would evolve resonance arising from the borosilicate glass of the NMR tube and
probe, all of the 11B FIDs were processed with back linear predic-
through essentially off-metal processes (oligomerization or
redistribution). In turn, some of the products thus gener-
ated are able to undergo further dehydrogenation mainly
tion (≈24 points) by using the MestReNova program.[43] Under
these processing conditions, total conversions and relative distribu-
tions of the dehydrogenation products in the catalytic reactions
leading to the formation of borazine-like products, which
might be slightly affected by the different relaxation times of the B
then represent a net loss of two equivalents of hydrogen. In
nuclei involved, so the data provided in Tables 1–4 must be taken
the case of the [M2Cp2(CO)6] (M = Cr, Mo, W) singly
as approximate product distributions rather than quantitative de-
bonded complexes, the metal center and, in particular, the
termination of all the species present.
different M–M bond strengths have a strong influence on
Example of Dehydrogenation Reaction for AB: In a typical experi-
the final catalytic activity, which drops significantly when
ment, the required amount of catalyst (0.006 mmol) was added to
going down in the group. Under the reaction conditions
used, these complexes are transformed into the catalytically
an NMR tube fitted with a Young valve. Then, a solution of AB
[0.23 m in THF/C6D6 (10:0.1), 0.5 mL, 0.115 mmol] was added, and
the valve was closed. The sample was then heated at 333 K for a
fixed time (6, 12 h, etc.), and the overall conversion and product
active mononuclear hydrides [MCpH(CO)3], in a reaction
which is instantaneous and complete only for the Cr dimer,
whereas it is significantly slower for the Mo and W deriva- distribution were evaluated by 11B NMR spectroscopy.
Eur. J. Inorg. Chem. 2013, 4998–5008
5006
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
S. Frueh, R. Kellett, C. Mallery, T. Molter, W. S. Willis, C.
King’ondu, S. L. Suib, Inorg. Chem. 2011, 50, 783.
Example of Dehydrogenation Reaction for tBuAB and MeAB: In a
typical experiment, the required amount of catalyst (0.046 mmol)
and the corresponding adduct (0.91 mmol) were dissolved in THF/
C6D6 (10:0.1, 0.7 mL). The resulting solution was then transferred
to an NMR tube fitted with a Young valve, and the valve was
closed. The headspace volume was vacuumed by freeze–pump cy-
cles, and the tube was sealed under static vacuum in the final cycle.
The sample was then heated at 333 K for a fixed time (6, 12 h, etc.),
and the overall conversion and product distribution were evaluated
as above.
[7]
[8]
See, for example: a) G. C. Andrews, T. C. Crawford, Tetrahe-
dron Lett. 1980, 21, 693; b) G. C. Andrews, Tetrahedron Lett.
1980, 21, 697; c) G. Ménard, D. W. Stephan, J. Am. Chem.
Soc. 2010, 132, 1796; d) P. V. Ramachandran, P. D. Gagare, K.
Sakavuyi, P. Clark, Tetrahedron Lett. 2010, 51, 3167.
a) C. A. Jaska, K. Temple, A. J. Lough, I. Manners, Chem.
Commun. 2001, 962; b) C. A. Jaska, K. Temple, A. J. Lough,
I. Manners, J. Am. Chem. Soc. 2003, 125, 9424; c) C. A. Jaska,
I. Manners, J. Am. Chem. Soc. 2004, 126, 9776; d) M. E. Sloan,
T. J. Clark, I. Manners, Inorg. Chem. 2009, 48, 2429; e) R. Dal-
lanegra, A. P. M. Robertson, A. B. Chaplin, I. Manners, A. S.
Weller, Chem. Commun. 2011, 47, 3763; f) L. J. Sewell, G. C.
Lloyd-Jones, A. S. Weller, J. Am. Chem. Soc. 2012, 134, 3598.
M. C. Denney, V. Pons, T. J. Hebden, D. M. Heinekey, K. I.
Goldberg, J. Am. Chem. Soc. 2006, 128, 12048.
Supporting Information (see footnote on the first page of this arti-
cle): 11B{1H} NMR stacking plots for selected dehydrogenation re-
actions.
[9]
[10]
a) N. Blaquiere, S. Diallo-Garcia, S. I. Gorelsky, D. A. Black,
K. Fagnou, J. Am. Chem. Soc. 2008, 130, 14034; b) M. Käß,
A. Friedrich, M. Drees, S. Schneider, Angew. Chem. 2009, 121,
922; Angew. Chem. Int. Ed. 2009, 48, 905; c) B. L. Conley, T. J.
Williams, Chem. Commun. 2010, 46, 4815; d) B. L. Conley, D.
Guess, T. J. Williams, J. Am. Chem. Soc. 2011, 133, 14212; e)
A. Friedrich, M. Drees, S. Schneider, Chem. Eur. J. 2009, 15,
10339.
Acknowledgments
We thank the European Commission (Project PERG08-GA-2010-
276958) for supporting this work and the Consejería de Educación
of Asturias for a “Clarín” postdoctoral reintegration grant (D. G.-
V.).
[11]
[12]
[13]
a) Y. Jiang, H. Berke, Chem. Commun. 2007, 3571; b) Y. Jiang,
O. Blacque, T. Fox, C. M. Frech, H. Berke, Organometallics
2009, 28, 5493.
S.-K. Kim, W.-S. Han, T.-J. Kim, T.-Y. Kim, S. W. Nam, M.
Mitoraj, Ł. Piekos, A. Michalak, S.-J. Hwang, S. O. Kang, J.
Am. Chem. Soc. 2010, 132, 9954.
a) T. Beweries, S. Hansen, M. Kessler, M. Klahn, U. Rosenthal,
Dalton Trans. 2011, 40, 7689; b) M. E. Sloan, A. Staubitz, T. J.
Clark, C. A. Russell, G. C. Lloyd-Jones, I. Manners, J. Am.
Chem. Soc. 2010, 132, 3831; c) T. J. Clark, C. A. Russell, I.
Manners, J. Am. Chem. Soc. 2006, 128, 9582.
a) Y. Kawano, M. Uruichi, M. Shimoi, S. Taki, T. Kawaguchi,
T. Kakizawa, H. Ogino, J. Am. Chem. Soc. 2009, 131, 14946;
b) B. Bera, B. R. Jagirdar, Inorg. Chim. Acta 2011, 372, 200.
a) R. J. Keaton, J. M. Blacquiere, R. T. Baker, J. Am. Chem.
Soc. 2007, 129, 1844; b) V. Pons, R. T. Baker, N. K. Szymczak,
D. J. Heldebrant, J. C. Linehan, M. H. Matus, D. J. Grant,
D. A. Dixon, Chem. Commun. 2008, 6597.
a) J. R. Vance, A. P. M. Robertson, K. Lee, I. Manners, Chem.
Eur. J. 2011, 17, 4099; b) R. T. Baker, J. C. Gordon, C. W.
Hamilton, N. J. Henson, P. H. Lin, S. Maguire, M. Murugesu,
B. L. Scott, N. C. Smythe, J. Am. Chem. Soc. 2012, 134, 5598.
S. J. McLain, J. Am. Chem. Soc. 1988, 110, 643.
See, for example: a) J. Peters, M. W. George, J. J. Turner, Orga-
nometallics 1995, 14, 1503; b) I. G. Virrels, M. W. George,
F. P. A. Johnson, J. J. Turner, J. R. Westwell, Organometallics
1995, 14, 5203; c) C. R. Baiz, R. McCanne, K. J. Kubarych, J.
Am. Chem. Soc. 2009, 131, 13590; d) R. Kania, A. I. Stewart,
I. P. Clark, G. M. Greetham, A. W. Parker, M. Towrie, N. T.
Hunt, Phys. Chem. Chem. Phys. 2010, 12, 1051; e) J. F. Ca-
hoon, M. F. Kling, K. R. Sawyer, L. K. Andersen, C. B. Harris,
J. Mol. Struct. 2008, 890, 328; f) J. F. Cahoon, M. F. Kling,
K. R. Sawyer, H. Frei, C. B. Harris, J. Am. Chem. Soc. 2006,
128, 3152; g) T. S. Chong, P. Li, W. K. Leong, W. Y. Fan, J.
Organomet. Chem. 2005, 690, 4132.
[1] See, for example: a) W. E. Winsche, K. C. Hoffman, F. J. Salz-
ano, Science 1973, 180, 1325; b) L. Schlapbach, A. Züttel, Na-
ture 2001, 414, 353; c) A. Züttel, A. Borgschulte, L. Schlapbach
(Eds.), Hydrogen as a Future Energy Carrier, Wiley-VCH,
Weinheim, 2008; d) A. Cho, Science 2004, 303, 942; e) L.
Schlapbach, Nature 2009, 460, 809.
[2] M. Hirscher (Ed.), Handbook of Hydrogen Storage: New Mate-
rials for Future Energy Storage, Wiley-VCH, Weinheim, 2009.
[3] For some recent reviews, see: a) T. B. Marder, Angew. Chem.
2007, 119, 8262; Angew. Chem. Int. Ed. 2007, 46, 8116; b) F. H.
Stephens, V. Pons, R. T. Baker, Dalton Trans. 2007, 2613; c) A.
Karkamkar, C. Aardahl, T. Autrey, Mater. Matters 2007, 2, 6;
d) B. Peng, J. Chen, Energy Environ. Sci. 2008, 1, 479; e) A.
Staubitz, A. P. M. Robertson, M. E. Sloan, I. Manners, Chem.
Rev. 2010, 110, 4023; f) A. Staubitz, A. P. M. Robertson, I.
Manners, Chem. Rev. 2010, 110, 4079; g) C. W. Hamilton, R. T.
Baker, A. Staubitz, I. Manners, Chem. Soc. Rev. 2009, 38, 279;
h) R. J. Less, R. L. Melen, D. S. Wright, RSC Adv. 2012, 2,
2191.
[4] See, for example: a) A. D. Sutton, A. K. Burrell, D. A. Dixon,
E. B. Garner III, J. C. Gordon, T. Nakagawa, K. C. Ott, J. P.
Robinson, M. Vasiliu, Science 2011, 331, 1426; b) P. V. Rama-
chandran, P. D. Gagare, Inorg. Chem. 2007, 46, 7810; c) S.
Hausdorf, F. Baitalow, G. Wolf, F. O. R. L. Mertens, Int. J. Hy-
drogen Energy 2008, 33, 608; d) B. L. Davis, D. A. Dixon, E. B.
Garner, J. C. Gordon, M. H. Matus, B. Scott, F. H. Stephens,
Angew. Chem. 2009, 121, 6944; Angew. Chem. Int. Ed. 2009,
48, 6812; e) P. G. Campbell, L. N. Zakharov, D. J. Grant, D. A.
Dixon, S. Y. Liu, J. Am. Chem. Soc. 2010, 132, 3289; f) T. J.
Carter, J. W. Kampf, N. K. Szymczak, Angew. Chem. 2012, 124,
13345; Angew. Chem. Int. Ed. 2012, 51, 13168.
[5] See, for example: a) A. Staubitz, M. E. Sloan, A. P. M. Robert-
son, A. Friedrich, S. Schneider, P. J. Gates, J. S. Günne, I. Man-
ners, J. Am. Chem. Soc. 2010, 132, 13332; b) C. W. Ewing, A.
Marchione, D. W. Himmelberger, P. J. Carroll, L. G. Sneddon,
J. Am. Chem. Soc. 2011, 133, 17093; c) T. He, J. H. Wang, G. T.
Wu, H. Kim, T. Proffen, A. A. Wu, W. Li, T. Liu, Z. T. Xiong,
C. Z. Wu, H. L. Chu, J. P. Guo, T. Autrey, T. Zhang, P. Chen,
Chem. Eur. J. 2010, 16, 12814; d) D. W. Himmelberger, C. W.
Yoon, M. E. Bluhm, P. J. Carroll, L. G. Sneddon, J. Am. Chem.
Soc. 2009, 133, 14101.
[14]
[15]
[16]
[17]
[18]
[19]
R. P. L. Burchell, P. Sirsch, A. Decken, G. S. McGrady, Dalton
Trans. 2009, 5851 and references cited therein.
J. Kleinberg, Inorg. Synth. 1963, 7, 136.
M. D. Curtis, M. S. Hay, Inorg. Synth. 1990, 28, 152.
a) V. Riera, M. A. Ruiz, F. Villafañe, Organometallics 1992, 11,
2854; b) G. García, M. E. García, S. Melón, V. Riera, M. A.
Ruiz, F. Villafañe, Organometallics 1997, 16, 624.
D. F. Gaines, R. Schaeffer, J. Am. Chem. Soc. 1963, 85, 3592.
a) H. I. Schlesinger, D. M. Ritter, A. B. Burg, J. Am. Chem.
Soc. 1938, 60, 1296; b) H. I. Schlesinger, D. M. Ritter, A. B.
[20]
[21]
[22]
[6] See, for example: a) E. J. M. Hamilton, S. E. Dolan, C. M.
Mann, H. O. Colijn, C. A. McDonald, S. G. Shore, Science
1993, 260, 659; b) Z. Liu, L. Song, S. Zhao, J. Huang, L. Ma,
J. Zhang, J. Lou, P. M. Ajayan, Nano Lett. 2011, 11, 2032; c)
[23]
[24]
Eur. J. Inorg. Chem. 2013, 4998–5008
5007
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Burg, J. Am. Chem. Soc. 1938, 60, 2297; c) X. Chen, J.-C. Zhao,
S. G. Shore, J. Am. Chem. Soc. 2010, 132, 10658.
Chem. Soc. Perkin Trans. 2 1985, 1723; c) I. G. Green, B. P.
Roberts, J. Chem. Soc. Perkin Trans. 2 1986, 1597.
[25] a) D. R. Martin, C. M. Merkel, J. P. Ruiz, Inorg. Chim. Acta
1986, 115, L29; b) N. Merle, G. Kociok-Köhn, M. F. Mahon,
C. G. Frost, G. D. Ruggerio, A. S. Weller, M. C. Willis, Dalton
Trans. 2004, 3883; c) J. Langer, V. K. Pálfi, H. Görls, M. Reiher,
M. Westerhausen, Chem. Commun. 2013, 49, 1121.
[26] Complex 3bЈ was found to be very air sensitive, and it slowly
decomposed upon prolonged storage under a nitrogen atmo-
sphere even at 253 K. The use of partially decomposed 3bЈ as
a catalyst modified the reaction rate, total conversion, and
product distribution. However, all of the data reported in the
text are derived from reactions in which such decomposition
was negligible, as judged by analysis of the reaction mixtures
by 31P{1H} NMR spectroscopy.
[34]
Y. Kawano, T. Yasue, M. Shimoi, J. Am. Chem. Soc. 1999, 121,
11744.
[35]
[36]
[37]
H. K. Hall, J. Am. Chem. Soc. 1957, 79, 5441.
R. F. Jordan, J. R. Norton, J. Am. Chem. Soc. 1982, 104, 1255.
a) R. Carreño, V. Riera, M. A. Ruiz, Organometallics 1993, 12,
1946; b) A. Le Goff, C. Le Roy, F. Y. Pétillon, P. Schollhammer,
J. Talarmin, Organometallics 2007, 26, 3607; c) P.
Schollhammer, N. Cabon, F. Y. Pétillon, J. Talarmin, K. W.
Muir, Chem. Commun. 2000, 2137; d) N. Cabon, F. Y. Pétillon,
P. Schollhammer, J. Talarmin, K. W. Muir, Dalton Trans. 2004,
2708.
[38]
A. B. Chaplin, A. S. Weller, Angew. Chem. 2010, 122, 591; An-
gew. Chem. Int. Ed. 2010, 49, 581.
[27] a) M. M. Hansmann, R. L. Melen, D. S. Wright, Chem. Sci.
2011, 2, 1554; b) C. Y. Tang, A. L. Thompson, S. Aldridge, J.
Am. Chem. Soc. 2010, 132, 1057.
[39] H. G. Alt, K. A. Mahmoud, A. J. Rest, Angew. Chem. 1983,
95, 569; Angew. Chem. Int. Ed. Engl. 1983, 22, 544.
[40] W. L. F. Armarego, C. L. L. Chai, Purification of Laboratory
Chemicals, 5th ed., Butterworth-Heinemann, Oxford, 2003.
[41] D. J. Heldebrant, A. Karkamkar, J. C. Linehan, T. Autrey, En-
ergy Environ. Sci. 2008, 1, 156.
[42] A. R. Manning, P. Hackett, R. Birdwhistell, P. Soye, Inorg.
Synth. 1990, 28, 148.
[43] MestReNova – NMR spectroscopic data processing software,
Mestrelab Research S.L. 2010, Santiago de Compostela,
www.mestrelab.com.
[28] H. C. Johnson, A. S. Weller, J. Organomet. Chem. 2012, 721–
722, 17.
[29] C. J. Stevens, R. Dallanegra, A. B. Chaplin, A. S. Weller, S. A.
Macgregor, B. Ward, D. McKay, G. Alcaraz, S. Sabo-Etienne,
Chem. Eur. J. 2011, 17, 3011.
[30] E. Wiberg, A. Bolz, Z. Anorg. Chem. 1948, 257, 131.
[31] a) S. L. Scott, J. H. Espenson, Z. Zhu, J. Am. Chem. Soc. 1993,
115, 1789; b) Q. Yao, A. Bakac, J. H. Espenson, Organometal-
lics 1993, 12, 2010.
[32] J. H. Espenson, J. Mol. Liq. 1995, 65/66, 205.
[33] a) J. A. Baban, B. P. Roberts, J. Chem. Soc. Perkin Trans. 2
1984, 1717; b) J. A. Baban, V. P. J. Marti, B. P. Roberts, J.
Received: May 15, 2013
Published Online: August 9, 2013
Eur. J. Inorg. Chem. 2013, 4998–5008
5008
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim