C. M. Frech et al.
and 12–20, Table 1). Whereas almost identical activities
were noticed by increasing the water content from 2 to
20%,[22] only a slight decrease in activity was obtained when
the amount of water was increased further. Thus, reactions
performed in pure H2O exhibit the lowest rates. Moreover,
catalysis stopped after about 10 min and the release of ap-
proximately 2.2 equivalents of H2. On the other hand, com-
not have an effect on neither the conversion rates nor on
the amount of hydrogen generated (in any of the reactions
performed):[29] 1) In all the reactions examined, sigmoidal-
shaped kinetics were observed, which is characteristic of
metal-particle formation and autocatalytic surface growth
that can lead to soluble, monodisperse nanoclusters or in-
soluble bulk-metal formation.[31] 2) The UV/Vis spectrum of
the reaction mixtures exhibit a continuous absorption with a
steep rise in absorbance at shorter wavelengths—a clear in-
dication of the presence of rhodium nanoparticles.[32] 3) A
black solution was instantly formed after the addition of sol-
plex 1 also shows (in contrast to [{RhClACTHNUTRGEN(UNG cod)}2]) an excellent
performance under anhydrous conditions.[23] For example,
when the dehydrogenation reactions were carried out in
pure THF, approximately 1.5 equivalents of H2 were re-
À
À
leased from H3N BH3 in about 8 min (Figure 3). Moreover,
utions of H3N BH3 in THF/H2O to the solutions of 1–4 in
THF, of which a solid with metallic appearance (insoluble in
common organic solvents and water) precipitated after the
reaction had run to completion. The filtrate was inactive in
À
the catalytic dehydrogenation of H3N BH3, whereas cataly-
sis was resumed upon addition of more substrate to the pre-
cipitate. 4) The addition of H2O to solutions of 1 in THF
lead to instant complex degradation, accompanied by the
formation of a black solid with metallic appearance, which
À
is active in the catalytic dehydrogenation of H3N BH3. The
filtrate on the other hand, showed no activity. 5) The steric
bulk and the s-donor strength of 1,1’,1’’-(phosphanetriyl)tri-
piperidine, 1,1’-(cyclohexylphosphanediyl)dipiperidine, 1-(di-
cyclohexylphosphanyl)piperidine, and tricyclohexylphos-
phane are almost identical,[33] implying similar catalytic ac-
tivities to be a molecular mechanism. However, whereas 2
operates at a similar level of activity compared to 1, a drop
in activity was noticed for 3. A further (significant) retarda-
tion was found for the phosphine analogue—the most stable
compound of this series[16]—as expected when nanoparticles
are involved in the catalytic cycle.[17]
À
Figure 3. Volumetric monitoring of the H2 release from H3N BH3 cata-
lyzed by 1 with different THF/H2O ratios.[16]
whereas NH3 and B(OH)3 have been generated in the pres-
ence of H2O, B-(cyclodiborazanyl)aminoborohydride (d=
À24.2, À11.2, and À5.2 ppm), borazine (d=30.8 ppm), and
polyborazylene (d=28.3 ppm) were reaction products under
anhydrous reaction conditions.[24] The signals due to bora-
zine and polyborazylene became dominant after some time.
However, when H2O was added to the reaction mixtures in-
stant release of additional H2, accompanied by the forma-
tion of NH3 and B(OH)3 was obtained, implying that the
borane species formed are hydrolyzed (in a first step) to
The following experimental observations indicate a cata-
lytic mechanisms based on the sequential dehydrogenation
À
of H3N BH3Àn(OH)n (n=0–2) and hydrolysis of H2N=
BH2Àn(OH)n (n=0–2): 1) The reaction products of the cata-
À
lytic dehydrogenation of H3N BH3 in the presence of H2O
are NH3 and B(OH)3, of which the latter is in equilibrium
with BO2 and other borate species.[18] 2) The catalytic dehy-
À
À
drogenation of H3N BH3 in the absence of H2O leads to the
release of ꢀ1.5 equivalents of H2 and is accompanied by the
formation of B-(cyclodiborazanyl)aminoborohydride, bora-
zine, and polyborazylene,[24] which implies that rhodium
nanoparticles are susceptible to dehydrogenation of both
À
À
H3N BH2(OH) and H3N BH(OH)2, respectively, which are
susceptible to dehydrogenation in the presence of rhodium
nanoparticles.[25]
À
H3N BH3 and H2B=NH2. However, the addition of H2O to
The same trends, but successively lower conversion rates
were obtained when 1,1’,1’’-(phosphanetriyl)tripiperidine
was substituted by 1,1’-(phosphanetriyl)dipiperidine, 1-
(phosphanetriyl)piperidine, and tricyclohexylphosphane,
which allows the efficiency of the H2 release to be con-
trolled by simple modifications to the aminophosphine
ligand.[16] The phosphine-based system, however, shows by
far the lowest catalytic activity, under all the reaction condi-
tions applied.
these reaction mixtures led to the instant release of another
1.5 equivalents of H2 and the formation of NH3 and
B(OH)3.[25] 3) The dehydrogenation of H3N BH3 performed
À
in D2O generates approximately 82% H2 and only about
1
18% HD (as indicated by H NMR spectroscopy) and, thus,
excludes mechanisms involving rhodium-catalyzed dissocia-
[25,34]
À
À
tion and hydrolysis of H3N BH3 and H3N BH2(OH).
Expectedly, no reaction occurred (neither in the absence
À
nor in the presence of H2O) when Me3N BH3 was applied
The following experimental observations indicate the in-
volvement of rhodium nanoparticles in the catalytic
cycles,[28] even though the presence of metallic mercury did
in dehydrogenation reactions. On the other hand,
ꢀ3.0 equivalents of H2 was released within 3.6 min in the
À
presence of 1.0 mol% of 1 when iPr2HN BH3 (0.5m; THF/
4734
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4732 – 4736