Communications
doi.org/10.1002/cctc.202100536
ChemCatChem
even though using excess amounts of alcohol probably dues to
[25,54]
the steric hindrance.
(1)
The time course profiles for the yield over Ir/TIO-10
catalysed alkylation of ammonia with 1a in both air and argon
atmosphere are shown in Figure 1. The catalysis behaviours in
air and argon are slightly different. When the reaction
performed in air, the reaction proceeded slightly faster and
gave the desired product in 19% yield in the initial hour. As the
reaction proceeded, the yield of the product dramatically
increased to 83% within 2 hours and then levelled off with
further increases in the reaction time. In the following time,
there was no substantial increase in production and finally
achieved a 90% yield after 18 hours. As for the reaction in
argon, the catalytic reaction gave the product in a yield of 47%
in the initial hour. Then, the reaction proceeded smoothly and
completed the conversion of all substrates within 6 hours. A
separate experiment reveals that the formation of tribenzyl-
amine from benzylamine and benzyl alcohol was rapidly
finished in 2 hours both in air or argon atmosphere [Eq. (1)],
suggesting the reaction rate hinges on the formation of
benzylamine from benzyl alcohol and ammonia. As one
possible reason for the different time course profiles in air or
argon, the oxidation of benzyl alcohol to benzaldehyde by air
and subsequent conversion to imines might accelerate the
Figure 1. Time-course profiles of the yields (%) of 2a from aqueous ammonia
with 1a over Ir/TIO-10 catalyst at 140°C. The conditions of in air reactions:
3
NH (2 mmol, 28% aqueous solution), 1a (10 mmol), catalyst (3 mol% as Ir,
6 wt%, 200 mg); The conditions of in argon reactions: NH (0.5 mmol, 28%
3
aqueous solution), 1a (2.5 mmol), catalyst (3 mol% as Ir, 6 wt%, 50 mg).
probably due to the strong adsorption of benzaldehyde 1b on
the surface of the catalysts (see below). Note that the
dehydrogenation of benzyl alcohol to benzaldehyde over
titania-supported iridium catalyst was completely suppressed in
the presence of an equimolar amount of benzaldehyde
[Eq. (S4)] as previously reported by us in a similar catalytic
system (see Figure S5 of [40]). This suggests that benzaldehyde
poisons the active sites on the surface of the present catalyst. In
the reactions shown in Eq. (S1) to Eq. (S4), the presence of a
significant amount of benzaldehyde 1b will hamper the
reaction. In the present alkylation of ammonia by benzylic
alcohols, the rapid transformation of aldehydes to imine would
prevent the poisoning of the catalyst.
[55–57]
reaction at the early stage.
The further detailed study on
the different behaviours in the reactions in air and argon is
ongoing. Consequently, the presence of air is proved not to
significantly hamper the reaction.
In addition, we further examined the reactivity of the
possible intermediate, 3a and 3b. The reaction of 3a or 3b
with benzyl alcohol 1a produced a small amount of the desired
product 2a [Eq. (S5) and (S6)], while over 99% of product 2a
was rapidly achieved in 2-hours from 1c and 1a [Eq. (1)]. Again,
the reaction of 3a or 3b with 1b gave 2a in low yields. These
results suggest that the multi-alkylation proceeds via 3a and
3b kept in the coordination sphere of the surface iridium
species rather than “free” 3a or 3b. The catalytic reaction of 1c
and 1d smoothly proceeded to give corresponding tertiary
amines 2b, 2c, and 2d together with imine 3c in a 2-hour
reaction, and the formation of 3a was not observed [Eq. (S7)].
This indicates the predominance of the dialkylation rather than
To understand the reaction mechanisms, a serial of separate
experiments was performed. Since Ir/TiO is well known for its
2
[55–57]
high performance in the aerobic oxidation of 1a,
we first
investigated the conversion of 1a in the presence of the
catalyst for one hour in argon or air [Eq. (S1)]. The result shows
that 36% or 27% of 1a was converted to benzaldehyde and its
dimers in argon or air, respectively. On the other hand, the
alkylation of aqueous ammonia with free benzaldehyde was
slow in the presence of the iridium catalyst [Eq. (S2)], probably
because of the poisoning by benzaldehyde (see below). Without
the catalyst the formation of 1c as well as the corresponding
imines were not detected under the present conditions,
suggesting the present catalyst have a promotional role for this
step. It is generally accepted that the hydrogen transfer
[59]
the self-condensation of 1c. Therefore, the present catalytic
transformation would proceed through hydrogen auto-transfer
pathways as demonstrated in Scheme 2, which is based on that
[29,35]
[19,25,58]
[25]
pathway over heterogeneous
or homogeneous
cata-
previously proposed for homogeneous catalysis. The first N-
lyst proceeds via the oxidation of alcohol and the following
amination of aldehyde. Also, the reaction of benzylamine 1c
and benzaldehyde 1b gave N-benzylidenebenzylamine 3a in
low yield (19%), while the reaction in the absence of the
catalyst smoothly proceeded [Eq. (S3)]. This suggests that the
catalyst retarded the transformation from aldehyde to imine,
alkylation proceeds via the rate-determining dehydrogenation
of benzylic alcohols to aldehydes and hydride species on the
surface of the iridium nanoparticles. The thus formed aldehyde
would immediately react with ammonia, probably adsorbed on
the weak acidic sites of the titania surface, to form an imine.
The hydrogenation of imine by the surface hydride species
ChemCatChem 2021, 13, 1–7
4
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!