D.V. Jawale et al. / Tetrahedron xxx (2014) 1e6
5
In terms of kinetics, when nanohybrids were compared to each
other with respect to the total content of gold, smaller particles
seemed to be the most active. However, comparison of ‘active’ gold
surface atoms content (those in contact with the medium) of s-
AuCNT and l-AuCNT showed much less differences. The calculation
of the exposed fraction of gold atoms was based on the work of
5
3
Boudart and Dj eꢀ ga-Mariadassou who suggested that the fraction
of surface atoms (or dispersion, D) of a particle could be calculated
according to the following formula where d is the diameter of the
particle in nanometer:
D ¼ 0:9=d
This simple formula provides a good approximation for gold
nanoparticles of spherical shape. More simply, the fraction of metal
exposed is the inverse of the metal particle size in nanometer. Thus,
a metal particle of 1 nm exposes nearly every metal atom, and a 2 nm
metal particle exposes about 45% of the metal atoms, etc. Therefore,
when the turnover frequencies were adjusted using the above cal-
culation, the s-AuCNT and l-AuCNT catalysts showed very similar
activities for the oxidation of dimethylphenyl silane, benzyl alcohol,
and for the reductive amination of benzaldehyde. This observation
suggests that the size effect between the two catalysts was essen-
tially due to the difference in catalytically active surface area exposed
to the medium. In the case of the hydroquinone oxidation, the dif-
ference in surface area cannot entirely account for the observed ef-
fect. In fact, even when the TOF values were corrected by considering
only surface atoms, s-AuCNT was still twice as active as l-AuCNT,
suggesting the occurrence of another factor. It is still unclear
whether the surface roughness of small AuNP (i.e., the larger number
Fig. 6. Early kinetics of the oxidation of 8 into 9 catalyzed by s-AuCNT (A) or l-AuCNT
B). In both cases, values are reported according to either the total amount of gold (a)
or to surface atoms only (b).
(
calculated by only taking into account the surface gold atoms. In-
deed, and as opposed to the transformations discussed above for
alcohol/silane oxidations and reductive amination, the calculated
turnover frequency values are clearly superior in the case of small
ꢀ1
AuNP. The TOFsurf value for s-AuCNT was 5980 h and that of l-
ꢀ
1
AuCNT was 2890 h
.
The recyclability of the catalyst was evaluated as described
above over five runs, using 0.13 mol % of catalyst. With s-AuCNT, for
each cycle, complete conversion was obtained after 4 h, whereas
27 h were required with l-AuCNT (Table 4). No decrease in catalytic
activity was observed over the course of the experiment, neither
with s-AuCNT nor with l-AuCNT. TEM analysis of the AuCNT hybrids
after the fifth run showed no significant morphological alteration.
15,17
of coordinatively unsaturated surface atoms)
can be responsible
for the extra activity of 3 nm over 20 nm particles. The influence of
surface heterogeneity found in smaller metal particles has been put
forward in previous reports from the literature but mainly for
transformation involving very small substrates such as carbon
Table 4
a
Recycling of the catalysts for hydroquinone oxidation
18
s-AuCNT
Run
l-AuCNT
Run
monoxide. In addition, hydroxyquinone diffusion hindrance on the
surface of larger AuNP may also impede the overall efficacy of the
corresponding AuCNT nanohybrid. However, it appears that for the
three first transformations, surface heterogeneity and/or diffusion
hindrance do not interfere with the catalytic activity.
Yieldb (%)
Yieldb (%)
Time (min)
Time (min)
1
2
3
4
5
240
240
240
240
240
98
97
99
96
98
1
2
3
4
5
1620
1620
1620
1620
1620
92
91
91
93
90
4. Conclusion
a
Conditions: tert-butyl-1,4-hydroquinone (0.23 mmol), catalyst (0.13 mol %)
suspended in water, K
under air.
2 3 3
CO (1 equiv), CHCl /water 3:1 (2 mL), room temperature,
We have shown that the AuCNT nanohybrid efficiently promotes
b
Isolated yields.
various organic transformations, whatever the size of the sup-
ported gold nanoparticles (diameter of 3 or 20 nm). However, some
differences were observed as regards turnover frequency values. As
a general trend, the smaller the particle size, the higher the TOF, but
size dependency became less significant when considering gold
surface atoms only as opposed to total gold content. This comment
does not apply to the catalytic oxidation of hydroquinone as the
recorded TOFsurf values were dissimilar for the two catalysts, but
the exact origin of this discrepancy is not fully understood yet.
Nevertheless, one could expect some benefits from the use of larger
CNT-supported gold nanoparticles, such as (i) lower sintering rate
of the catalyst (which classically leads to a loss in selectivity and
3
. Discussion
We investigated herein how nanoparticle size affects the effi-
ciency of CNT-gold nanohybrids. Two different types of nano-
particles were assembled on the surface of the carbon nanotubes
using a layer-by-layer approach. The resulting nanohybrids were
then utilized in the catalytic oxidation of phenyldimethylsilane,
benzyl alcohol, and hydroxyquinone and also in the direct reductive
amination of benzaldehyde. Some previously reported control ex-
periments run either with gold nanoparticles (AuNP in the absence
of CNT) or DANTAePDADMAC-coated carbon nanotube support
5
4
activity) under the reaction conditions and (ii) the commercial
availability of some large citrate-caped AuNP that can readily be
assembled into CNT-based nanohybrids.
(
devoid of any nanoparticle) indicated very low conversion yields
for the former catalyst and no reaction at all for the latter. It thus
appears that the multicomponent catalyst is more active than the
metallic species itself (AuNP). This observation can be rationalized
by the fact that the supported metallic nanoparticles are less prone
to aggregation than their colloidal counterparts. In addition,
nanotubes are electronically active and stabilization of transient
higher oxidation states of gold are anticipated by collaborative in-
teractions between the nanotube and the metal.
5. Experimental section
5.1. Typical procedure for silane oxidation
To a solution of dimethylphenyl silane (1, 0.2 mmol) in THF
(2 mL), under air, the catalyst (0.1 mol %) suspended in water was