Mild and Selective Catalytic Hydrogenation of the C=C Bond in α,β-Unsaturated Carbonyl Compounds Using Supported Palladium Nanoparticles

Article abstract of DOI:10.1002/chem.201600878

Chemoselective reduction of the C=C bond in a variety of α,β-unsaturated carbonyl compounds using supported palladium nanoparticles is reported. Three different heterogeneous catalysts were compared using 1 atm of H₂: 1) nano-Pd on a metal-organic framework (MOF: Pd⁰-MIL-101-NH₂(Cr)), 2) nano-Pd on a siliceous mesocellular foam (MCF: Pd⁰-AmP-MCF), and 3) commercially available palladium on carbon (Pd/C). Initial studies showed that the Pd@MOF and Pd@MCF nanocatalysts were superior in activity and selectivity compared to commercial Pd/C. Both Pd⁰-MIL-101-NH₂(Cr) and Pd⁰-AmP-MCF were capable of delivering the desired products in very short reaction times (10-90 min) with low loadings of Pd (0.5-1 mol %). Additionally, the two catalytic systems exhibited high recyclability and very low levels of metal leaching.

Full text of DOI:10.1002/chem.201600878

DOI: 10.1002/chem.201600878
Full Paper
&
Heterogeneous Catalysis
Mild and Selective Catalytic Hydrogenation of the C=C Bond in
a,b-Unsaturated Carbonyl Compounds Using Supported
Palladium Nanoparticles
Anuja Nagendiran+,[a] Vlad Pascanu+,[a] Antonio Bermejo Gómez,[a, c] Greco Gonzµlez Miera,[a]
Cheuk-Wai Tai,[b] Oscar Verho,*[a] BelØn Martín-Matute,*[a] and Jan-E. Bäckvall*[a]
Abstract: Chemoselective reduction of the C=C bond in a va-
riety of a,b-unsaturated carbonyl compounds using support-
ed palladium nanoparticles is reported. Three different heter-
ogeneous catalysts were compared using 1 atm of H2:
1) nano-Pd on a metal–organic framework (MOF: Pd0-MIL-
101-NH2(Cr)), 2) nano-Pd on a siliceous mesocellular foam
(MCF: Pd0-AmP-MCF), and 3) commercially available palladi-
um on carbon (Pd/C). Initial studies showed that the
Pd@MOF and Pd@MCF nanocatalysts were superior in activi-
ty and selectivity compared to commercial Pd/C. Both Pd0-
MIL-101-NH2(Cr) and Pd0-AmP-MCF were capable of deliver-
ing the desired products in very short reaction times (10–
90 min) with low loadings of Pd (0.5–1 mol%). Additionally,
the two catalytic systems exhibited high recyclability and
very low levels of metal leaching.
as supports in heterogeneous catalysis.[4] On the other hand,
porous siliceous materials such as mesocellular foams (MCFs),
although amorphous, consist of an adjustable three-dimen-
sional network of pores and display a significantly broader
range of pore sizes of up to 50 nm. Therefore they can pro-
mote a more efficient mass transfer. The siliceous MCF that has
been the subject of work in our group[5] delivers a high surface
concentration of silanol groups that can be further functional-
ized. In a similar way, the inner surface of MOFs’ pores can be
tailored by anchoring functional groups onto the organic link-
ers, either directly in the synthesis process or through post-
synthetic modification techniques. Both MOFs[6] and MCFs[7]
have been successfully used to support metallic nanocatalysts.
The confined environment created around the catalyst can
lead to a different reactivity than that observed in solution.
Often, this reactivity is also dependent on the size and shape
of the nanoparticles.[8] Furthermore, the porous nature of the
supports prevents aggregation of the metal. Finely dispersed
nanoparticles with narrow size distributions can easily be ob-
tained and protected in the absence of additional surfactants
and their surface-dependent catalytic properties can be pre-
served.
Introduction
Catalytic hydrogenation reactions have proven to be highly im-
portant transformations in organic synthesis. In 2013, one
fourth of all drugs, marketed or in clinical trials, involved at
least one hydrogenation step in their synthetic route.[1] The
chemoselective hydrogenation of olefinic bonds in the pres-
ence of other reducible functional groups such as aldehydes or
ketones is an important but particularly challenging transfor-
mation. Most catalysts, especially heterogeneous ones, are se-
verely limited in terms of selectivity and their applicability is re-
stricted to non-challenging substrates.[2]
Metal–organic frameworks (MOFs)[3] are crystalline porous
materials with very high surface areas and tunable pore sizes
from micro- to mesopores, that have been successfully applied
[a] Dr. A. Nagendiran,+ V. Pascanu,+ Dr. A. Bermejo Gómez, G. Gonzµlez Miera,
Dr. O. Verho, Prof. Dr. B. Martín-Matute, Prof. Dr. J.-E. Bäckvall
Department of Organic Chemistry and
Berzelii Centre EXSELENT on Porous Materials
Arrhenius Laboratory, Stockholm University
106 91 Stockholm (Sweden)
E-mail: Oscar@organ.su.se
belen.martin.matute@su.se
In this study we have evaluated and compared the efficiency
of palladium nanoparticles supported on siliceous mesocellular
foam (Pd0-AmP-MCF, 8 wt% Pd)[9] and on a metal–organic
framework (Pd0-MIL-101-NH2(Cr), 8 wt% Pd)[4b,10] with that of
Pd/C. Both porous matrices have been pre-functionalized with
pendant ÀNH2 groups, in order to stabilize the nanoparticles
and to minimize Pd leaching upon repeated uses. Herein, we
demonstrate the efficiency of the two porous nanocatalysts in
the chemoselective reduction of the olefinic double bond of
different a,b-unsaturated carbonyl compounds.
jan.e.backvall@su.se
[b] Dr. C.-W. Tai
Department of Materials and Environmental Chemistry
Stockholm University, 106 91 Stockholm (Sweden)
[c] Dr. A. Bermejo Gómez
Current address:
AstraZeneca Translational Science Center at Karolinska Institute
171 65 Stockholm (Sweden)
[+] These authors contributed equally to this work.
Supporting information for this article can be found under http://
dx.doi.org/10.1002/chem.201600878.
Chem. Eur. J. 2016, 22, 7184 – 7189
7184
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
alyst, the reaction had in general very poor selectivities (50–
73%) and low conversions (36–68%) in all the tested solvents
(Table 1, entries 10–13). The reactions catalyzed by Pd/C were
generally slower than those catalyzed by the other two nano-
palladium catalysts, and among them Pd0-AmP-MCF afforded
the highest rates. This may be explained by a facilitated mass
transfer of the reagents into the MCF material through the
wider pore windows (13 nm for AmP-MCF vs. 1.6 nm for MIL-
101-NH2(Cr)). It is noteworthy that allylic alcohol 4a could not
be observed at any moment in any of the experiments de-
scribed until this point. To rule out the intermediacy of 4a in
the formation of 2a (Pathway D) we attempted the isomeriza-
tion of the allylic alcohol (4a). However, none of the three cat-
alysts afforded aldehyde 2a. Furthermore, it was demonstrated
that 4a is rapidly reduced to 3a (Pathway C’) under the reac-
tion conditions (i.e., 1 mol% of Pd0-MIL-101-NH2(Cr) under
1 atm H2 reduces cinnamyl alcohol (4a) into 3-phenylpropan-
1-ol (3a) with a TOF three times higher than in the case of cin-
namaldehyde). In further control experiments, the attempted
hydrogenation of 2a (stirred overnight under 1 atm H2) in the
presence either Pd0-AmP-MCF or Pd0-MIL-101-NH2(Cr) did not
produce compound 3a. The absence of 3a in these experi-
ments indicates that over-reduction via Pathway C does not
occur under the present conditions, and that 3a is only
formed from intermediate allylic alcohol 4a, which quickly re-
duces to form 3a in the reaction mixture (Pathways B+C’).
We also considered that the readily oxidizable cinnamalde-
hyde (1a) could be contaminated with trace amounts of cin-
namic acid. This species could then activate the carbonyl
moiety towards hydrogenation (via hydrogen bonding), which
would favor formation of the over-reduced product 3a via
Pathways B+C’. However, 3a also was observed in similar
amounts when freshly distilled cinnamaldehyde was used.
Results and Discussion
The hydrogenation of cinnamaldehyde (1a) was chosen as the
model reaction. The possible pathways the reaction can take
are illustrated in Scheme 1. Under ideal conditions, the only
operating pathway would be the selective reduction of the
C=C double bond (Pathway A), leading to exclusive formation
of 3-phenylpropanal (2a). Over-reduction to 3-phenylpropan-1-
ol (3a; pathways C and C’) is favored by longer reaction times
and/or harsher reaction conditions, such as higher temperature
and increased H2 pressure. Additionally, the selectivity can be
affected by other factors such as the type and morphology of
the support, metal dispersion, steric constraints and electronic
effects. Literature reports based on Ru and Pt nanoparticles
demonstrated that the size of the nanoparticles does play an
important role on the selectivity and that larger nanoparticles
and aggregates are able to promote selective formation of the
allylic alcohol 4a.[11] On these grounds we hypothesized that
small, finely dispersed nanoparticles supported on porous
MOFs and MCFs would have the right shape, size and environ-
ment to favor Pathway A over B and C.
Scheme 1. Possible reaction pathways in the hydrogenation of cinnamalde-
hyde.
We set out to find the ideal conditions for a selective trans-
formation by evaluating the behavior of the two catalysts (Pd0-
AmP-MCF and Pd0-MIL-101-NH2(Cr)) together with that of Pd/C
in different solvents. As shown in Table 1, the efficiency of Pd0-
MIL-101-NH2(Cr) was mostly unaffected by the choice of sol-
vent. Overall, excellent conversions (>99%) and relatively high
selectivities (89–90%) were obtained within 90 min in all of the
solvents examined (Table 1, entries 3–5). Shorter reaction times
afforded lower conversions (Table 1, entries 1 and 2). On the
contrary, the catalytic performances of Pd0-AmP-MCF (Table 1,
entries 6–9) and of Pd/C (Table 1, entries 10–13) were depen-
dent on the solvent to a large extent. With Pd0-AmP-MCF, high
conversions were obtained in both acetone and EtOAc (88 and
96%, respectively) in only 30 min; however, the selectivity to-
wards the desired 3-phenylpropanal (2a) was higher in the
latter solvent (86 vs. 94%; Table 1, entries 6 and 7).[12] Interest-
ingly, although full conversion was achieved when the reaction
time was prolonged to 60 min in EtOAc, a minor increase in
over-reduction of 2a to 3a could be observed (6% vs 8%;
Table 1, entry 7 vs. 8). Running the reaction in toluene resulted
in lower conversion (66%) and the selectivity decreased to
87% (Table 1, entry 9). For the commercially available Pd/C cat-
Table 1. Initial screening of solvents and reaction time for the hydroge-
nation of cinnamaldehyde (1a).[a]
Entry Catalyst
Solvent
t
Conv. 2a/3a[b] 2a
[min] [%][b]
[%][b]
1
2
3
4
5
6
7
8
Pd0-MIL-101-NH2(Cr) acetone 30
acetone 60
55 85:15
86 87:13
>99 89:11
>99 90:10
>99 89:11
88 86:14
96 94:6
>99 92:8
66 87:13
40 55:45
68 55:45
50 73:27
36 50:50
47
75
89
90
89
76
90
92
57
22
37
37
18
acetone 90
EtOAc
90
toluene 90
acetone 30
Pd0-AmP-MCF
Pd/C[c]
EtOAc
EtOAc
30
60
9
toluene 30
acetone 30
acetone 60
10
11
12
13
EtOAc
30
toluene 30
[a] Unless otherwise noted, reactions were performed on a 0.8 mmol
1
scale under 1 atm of H2 at 208C. [b] Determined by H NMR spectroscopy
from the crude reaction mixture. [c] 10 wt% Pd.
Chem. Eur. J. 2016, 22, 7184 – 7189
www.chemeurj.org
7185 ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Therefore, the effect of a base additive such as K2CO3, which
could neutralize acidic species in-situ, was investigated for the
three different catalysts (Table 2). When Pd0-MIL-101-NH2(Cr)
was used, the addition of K2CO3 (0.1 equiv) was detrimental for
the selectivity (Table 2, entries 1–3 vs. Table 1, entries 3–5).
However, these results are consistent with our[13] and others’[14]
recent investigations on the stability of the MIL-101(Cr) frame-
work in basic media. The presence of K2CO3 leads to a degrada-
tion of the porous framework, which can no longer prevent
the agglomeration of Pd nanoparticles. Large Pd aggregates
can then be responsible for the decreased selectivity as dis-
cussed above.[11] Conversely, for the Pd0-AmP-MCF catalyst,
which is not base-sensitive, the presence of K2CO3 had a negli-
gible effect on the selectivity in EtOAc (Table 2, entry 4 vs.
Table 1, entry 7). However, in acetone the selectivity increased
to >99% in the presence of K2CO3, and quantitative formation
of 2a was observed within 60 min (Table 2, entries 5 and 6).
Compared to the two nanocatalysts, Pd/C still exhibited lower
conversions (29–52%) and selectivities (57–82%), even in the
presence of base (Table 2, entries 7–10). Other solvents and ad-
ditives were investigated under similar conditions but this did
not lead to any further improvements of the reaction (see the
Supporting Information for further details).
the effect of the substrate concentration. Gratifyingly, we
found that lowering the substrate concentration from 0.4 to
0.2m without the addition of base, improved the selectivity
significantly (Table 2, entries 11–15). This surprising effect could
be attributed to the presence of Cr Lewis acidic sites in the
structure of the MOF, which are prone to be exposed on the
surface of the crystals and coordinate to the carbonylic oxygen
atoms, activating them towards the undesired reduction of the
C=O. Interestingly, in all experiments catalyzed by Pd@MOF,
the by-product (3a) is formed in the earliest stages of the reac-
tion. By diluting the reaction mixture from 0.4 to 0.2m, the
concentration of exposed acidic Cr sites is reduced, decreasing
their chances to activate C=O bonds towards undesired reduc-
tion. On the other hand, the dilution of the reaction medium
could simply render the whole catalyst less active, thus im-
proving the overall selectivity. Changing the solvent still did
not drastically affect the outcome of the reactions; quantitative
conversions and similarly high selectivities (94–95%) were re-
corded in each of the three solvents that were examined after
90 min of reaction time. Further diluting the reaction had no
beneficial effect and running the reaction under neat condi-
tions afforded a very low conversion with a selectivity of 92%
(Table 2, entry 16). For comparison, the use of Pd/C under simi-
lar conditions (0.2m, no base) resulted in lower conversion
(69%) and a very poor selectivity (60%) after 90 min (Table 2,
entries 17–19). For the reactions with Pd0-AmP-MCF, only negli-
gible improvements in selectivity were observed when the re-
action was performed without base and with a substrate con-
centration of 0.2m compared to the reaction with 0.4m.
Since the addition of base did not improve the selectivity in
the reactions catalyzed by Pd0-MIL-101-NH2(Cr), further optimi-
zation of this catalytic system was carried out by investigating
Table 2. Influence of base addition and dilution on the Pd-catalyzed hy-
drogenation reactions.[a]
The screening experiments described clearly demonstrate
that Pd0-MIL-101-NH2(Cr) and Pd0-AmP-MCF display a signifi-
cantly higher selectivity for the reduction of the olefinic bond
in a,b-unsaturated carbonyl compounds compared to commer-
cially available Pd/C. Even more, the remarkable reactivity of
the two nanocatalysts enables them to operate under very
mild conditions, avoiding over-reduction of the reaction prod-
uct. Under the optimized conditions, the well-controlled size
and shape of the encapsulated nanoparticles (ca. 2–3 nm; see
Supporting Information, Figure S1) in MIL-101-NH2(Cr) and in
AmP-MCF inhibits the undesired reduction of the C=O bond
(Pathway B, Figure 1), previously observed with larger nanopar-
ticles.[11b] To the best of our knowledge, the results reported
herein for the Pd@MCF and Pd@MOF catalysts are unsurpassed
in terms of TOF and selectivity (see the Supporting Information
for details) by any heterogeneous Pd catalysts in the selective
hydrogenation of cinnamaldehyde under ambient conditions
(room temperature and 1 atm of H2). A review of recent litera-
ture on the topic revealed that only catalysts operating under
much harsher conditions (up to 708C and up to 20 atm H2)
manage to return higher turnover frequencies and comparable
selectivities.
Entry Catalyst
Solvent
t
Conv. 2a/3a[b] 2a
[min] [%][b]
[%][b]
1
Pd0-MIL-101-NH2(Cr) toluene
90
90
90
30
30
60
30
30
60
30
90
30
60
90
90
960
30
60
90
>99
>99
>99
62
88:12
69:31
74:26
95:5
89
69
74
2
acetone
EtOAc
EtOAc
acetone
acetone
toluene
acetone
acetone
EtOAc
3
4
5
6
Pd0-AmP-MCF
Pd/C (10wt%)
59
67 >99:1
67
>99 >99:1
>99
28
18
30
32
95
37
67
7[c]
8[c]
9[c]
10[c]
11[d]
12[d]
13[d]
14[d]
15[d]
16[d]
17[d]
18[d]
19[d]
38
29
52
39
>99
42
73:27
62:38
57:43
82:18
95:5
Pd0-MIL-101-NH2(Cr) toluene
acetone
acetone
acetone
EtOAc
88:12
93:7
72
>99
>99
25
95:5
94:6
92:8
95
94
23
neat
Pd/C[e]
acetone
acetone
acetone
37
57
69
60:40
60:40
60:40
22
34
41
Next, the catalytic efficiency of Pd0-MIL-101-NH2(Cr) and Pd0-
AmP-MCF was compared for the chemoselective hydrogena-
tion of crotonaldehyde (1b) and methacrolein (1c), which are
challenging substrates but highly interesting, since butanal
(2b) and isobutanal (2c) are industrially important chemical in-
termediates.[15] Although both catalysts had no difficulties in
[a] Unless otherwise noted, reactions were performed on a 0.8 mmol
scale under 1 atm of H2 at room temperature. [b] Determined by H NMR
spectroscopy of the crude reaction mixture. [c] Traces of 4a were ob-
served when the reaction was allowed to stir, overnight. [d] Reaction di-
luted from 0.4 to 0.2m and no base additive used. [e] 10 wt% Pd.
1
Chem. Eur. J. 2016, 22, 7184 – 7189
www.chemeurj.org
7186
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. Unless otherwise noted, the reactions were performed on a 0.8 mmol scale, in 4 mL of acetone (0.2m for Pd0-MIL-101-NH2(Cr)) or 2 mL of EtOAc
1
(0.4m for Pd0-AmP-MCF) at 208C and under H2 atmosphere (balloon, 1 atm), for the indicated time. The progress of the reaction was monitored by H NMR
analysis of the crude reaction mixture. [a] The reaction was run in acetone in the presence of 0.1 equiv of K2CO3.
providing full conversions, Pd0-AmP-MCF was found to be
slightly faster while also affording excellent selectivity, even
when using a loading of Pd as low as 0.5 mol%. On the other
hand, the slightly lower selectivity of the Pd@MOF catalyst
(95% and 90% for 2b and 2c, respectively) can be again at-
tributed to the presence of Lewis acidic Cr sites on the frame-
work, which are even more easily accessed by the small mole-
cules of substrates 1b and 1c.
The selectivity was also excellent for linear a,b-unsaturated
ketones. Both aromatic (1d) and aliphatic (1e) substrates were
efficiently hydrogenated with complete selectivity by both cat-
alysts, the reaction being faster for Pd0-AmP-MCF. Cyclic sub-
strates (1 f–1h) were also well tolerated even when the double
bond was substituted (1 f and 1h), and full conversions were
achieved in a matter of minutes, displaying the versatility of
these Pd nanocatalysts. Only trace amounts of alcohols 3 f–3h
were observed for Pd0-MIL-101-NH2(Cr) while Pd0-AmP-MCF
provided the desired products 2 f–2h with complete selectivi-
ty. However, in the case of the cyclic enone 1g the formation
of the over-reduced byproduct 3g could be observed on pro-
longed reaction times. Gratifyingly, all products 2b–2h were
obtained in excellent yields and selectivities without the addi-
tion of base.
Stability and leaching
The initial recycling studies were conducted with the model
substrate cinnamaldehyde (1a), in the presence K2CO3 and
using Pd0-AmP-MCF as the catalyst. The reaction time was
shortened to 30 min in order to clearly asses the catalyst effi-
ciency for each cycle. During three consecutive cycles, a gradual
decrease (67, 53 and 43%) in the conversion towards the de-
sired product 2a was observed, although the selectivity was
unchanged. Since the recycling of cinnamaldehyde was not
optimal for Pd0-AmP-MCF, we did not evaluate the recyclability
of Pd0-MIL-101-NH2(Cr) for this substrate. However, when
ketone 1d was subjected to recycling studies, both Pd0-MIL-
101-NH2(Cr) and Pd0-AmP-MCF showed a very stable behavior
with no differences in reactivity over five runs (Table 3). In each
case, full conversion and excellent selectivity towards 2d were
obtained after 75 min for Pd0-MIL-101-NH2(Cr) and after 30 min
for Pd0-AmP-MCF. Over the five runs, the amount of leached
Chem. Eur. J. 2016, 22, 7184 – 7189
www.chemeurj.org
7187
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Pd was kept at barely detectable levels by ICP-OES (ꢀ0.2 ppm)
for both catalysts (Table 3). The analysis of HAADF-STEM and
TEM images of Pd0-MIL-101-NH2(Cr) revealed that in general,
most nanoparticles retain their original size and shape (see
Supporting Information, Figures S1 and S2). In the fresh cata-
lyst about 85% of the nanoparticles fit in a very narrow size
distribution range between 1.5 and 2.5 nm and altogether
more than 95% of the nanoparticles are smaller than the pore
size limit of the MOF (34 Š). After five recycling runs, some re-
distribution of Pd is observed but still about 87% of the nano-
particles remain under the size limit imposed by the MOFs
pores. The other 13% of Pd nanoparticles, which are larger in
size, are located on the surface of the MOF crystallites. This
can happen due to Pd migration or degradation of the MOF
by mechanical factors. Still, maintaining most of the Pd species
in a very narrow size range shows that the porous support is
actively protecting the catalyst against the natural Ostwald rip-
ening and thus prolongs its lifetime. The XRPD patterns of
fresh and recycled Pd0-MIL-101-NH2(Cr) after five runs were
identical with no distinguishable loss in crystallinity (Support-
ing Information, Figure S3). However, small amounts of Cr
were detected in the third and fifth run (Table 3), most likely
caused by the mechanical stress to which the MOF crystals are
subjected during reaction cycles under vigorous stirring.
of them remaining under 3.5 nm. This is in accordance with
the previous results where Pd0-AmP-MCF had been used for
hydrogenation of nitroarenes and no agglomeration was de-
tected.[16]
Conclusions
The two composite materials, Pd0-MIL-101-NH2(Cr) and Pd0-
AmP-MCF have shown excellent activity in the selective hydro-
genation of the C=C bond in various a,b-unsaturated carbonyl
compounds, with all substrates reaching a full conversion
within 90 min under only 1 atm of H2. Under the optimized re-
action conditions, both Pd0-AmP-MCF and Pd0-MIL-101-NH2(Cr)
are able to prevent agglomeration of Pd nanoparticles, pro-
longing their lifetime and maintaining their size and facet de-
pendent properties. This property provides both catalysts with
a remarkable control of the chemoselectivity of the reaction,
which is not achieved with commercially available Pd/C. Over
multiple recycling runs the contamination of leached metallic
species was kept at a minimum level. However, in this compa-
rative study, Pd0-AmP-MCF proved to be slightly more effective
than Pd0-MIL-101-NH2(Cr), and its excellent performance at low
catalyst loadings makes it highly attractive for hydrogenations
in industrial settings.
Table 3. Metal leaching during recycling experiments.[a]
Experimental Section
General information
Pd0-MIL-101-NH2(Cr)
Run Conv. of Selectivity Pd
Pd0-AmP-MCF
Conv. Selectivity Pd
Reagents and solvents were used as obtained from commercial
suppliers without further purification, except cinnamaldehyde,
which was distilled under reduced pressure before use. Pd0-MIL-
101-NH2(Cr) (8.40 wt%) and Pd0-AmP-MCF (7.90 wt%) were synthe-
sized according to previously reported procedures, described in
the Supporting Information.
Cr
1d [%] 2d/3d
leaching leaching [%]
2d/3d
leaching
[ppm]
[ppm]
[ppm]
1
2
3
4
5
>99
>99
>99
>99
>99
>99:1
>99:1
>99:1
>99:1
>99:1
<0.1
n.d.
<0.1
n.d.
0.1
n.d.
0.6
n.d.
1.8
>99 >99:1
>99 >99:1
>99 >99:1
>99 >99:1
>99 >99:1
<0.1
n.d.
<0.1
n.d.
1H NMR spectra were recorded at 400 MHz on a Bruker Advance
1
spectrometer. H chemical shifts (d) are reported in ppm from tet-
0.1
0.2
ramethylsilane with the solvent resonance as the internal standard
(CDCl3: dH =7.26 ppm). Elemental analysis: determinations of Pd
and Cr were performed by inductively coupled plasma optical
emission spectrometry (ICP-OES) with a Varian Vista MPX instru-
ment. Medac Ltd. in the UK carried out the elemental analysis. X-
ray powder siffraction (XRPD) data were collected on a PANalytical
X’Pert PRO diffractometer in reflectance Bragg–Brentano geometry
equipped with a pixel detector and using CuKa1 radiation. MIL-101
samples were suspended in EtOH, dispersed on zero-background
Si plates and dried at room temperature. The morphology of “sam-
ples” was examined in a JEOL-2100F transmission electron micro-
scope operated at 200 kV and equipped with a Schottky-type field
emission gun. The crushed samples were dispersed in ethanol and
then onto a TEM Cu grid with holey carbon supporting films.
Bright-field and high-resolution TEM (HRTEM) images were record-
ed by a CCD camera (Gatan Ultrascan 1000). High-angle annular
dark-field scanning transmission electron microscopy (HAADF-
STEM), also known as Z-contrast, images for determining the size
and distribution of Pd nanoparticles were acquired using a JEOL
ADF detector. The probe size and camera length used are 0.20 nm
and 8 cm, respectively.
[a] Conditions for reactions with Pd0-MIL-101-NH2(Cr): The reactions were
performed on a 0.8 mmol scale, in 4 mL of acetone at 208C and under H2
atmosphere (balloon, 1 atm), for 75 min. Conditions for reactions with
Pd0-AmP-MCF: The reactions were performed on a 0.8 mmol scale, in
2 mL of EtOAc at 208C and under H2 atmosphere (balloon, 1 atm), for
30 min. It was shown in the first run that the reaction of Pd0-AmP-MCF re-
quired at least 10 min for full conversion.
When the Pd0-AmP-MCF was subjected to a similar HAADF-
STEM and TEM analysis, a slightly wider size distribution was
observed in the fresh catalyst with only 40% of the nanoparti-
cles in the 1.5–2.5 nm range and 50% in the 2.5–3.5 nm range,
while the remaining 10% of the nanoparticles were larger than
4 nm (see Supporting Information, Figures S4 and S5). This is
not surprising considering the wider pore apertures of the
MCF materials. However, we were delighted to find that after
five recycling runs, the level of Pd agglomeration is barely de-
tectable and most nanoparticles retain their size with ca. 90%
Chem. Eur. J. 2016, 22, 7184 – 7189
www.chemeurj.org
7188
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
General procedure for the Pd0-catalyzed hydrogenation of
enals and enones
17483–17493; c) V. Pascanu, P. R. Hansen, A. Bermejo Gómez, C. Ayats,
A. E. Platero-Prats, M. J. Johansson, M. À. Pericàs, B. Martín-Matute,
ChemSusChem 2015, 8, 123–130; d) V. Pascanu, F. Carson, M. V. Solano,
J. Su, X. Zou, M. J. Johansson, B. Martín-Matute, Chem. Eur. J. 2016, 22,
3729–3737.
The corresponding enal or enone (0.8 mmol) and Pd0-MCF/MOF
(4–8.0 mmol Pd, 0.50–1 mol%) were suspended in the chosen sol-
vent (MCF: EtOAc, 2 mL or MOF: acetone, 4 mL) in a screw-capped
Radleyꢁ tube. The reaction vessel was then evacuated and filled
with hydrogen gas from a balloon in three repeating cycles. The
reaction was allowed to stir at room temperature for the indicated
time with the H2 balloon attached, after which the reaction was
stopped. Aliquots were taken from the reaction mixture, filtered
[5] For recent developments of catalytic processes with Pd0-AmP-MCF, see:
a) E. Bratt, O. Verho, M. J. Johansson, J.-E. Bäckvall, J. Org. Chem. 2014,
79, 3946; b) K. Engstrçm, E. V. Johnston, O. Verho, K. P. J. Gustafson, M.
Shakeri, C.-W. Tai, J.-E. Bäckvall, Angew. Chem. Int. Ed. 2013, 52, 14006–
14010; Angew. Chem. 2013, 125, 14256–14260; c) J. Malmgren, A. Na-
gendiran, C.-W. Tai, J.-E. Bäckvall, B. Olofsson, Chem. Eur. J. 2014, 20,
13531–13535.d) O. Verho, A. Nagendiran, E. V. Johnston, C.-W. Tai, J.-E.
Bäckvall, ChemCatChem 2013, 5, 612–618.
1
through Celite, eluted with CDCl3 (1 mL), and analyzed by H NMR
[6] For a comprehensive review on MOF supported nanoparticles, see: A.
Dhakshinamoorthy, H. Garcia, Chem. Soc. Rev. 2012, 41, 5262–5284.
[7] For a comprehensive review on MCF, see: A. Taguchi, F. Schüth, Micro-
porous Mesoporous Mater. 2005, 77, 1–45.
spectroscopy to determine the conversion and selectivity. See the
Supporting Information for more details.
[8] a) G. Collins, M. Schmidt, C. O’Dwyer, J. D. Holmes, G. P. McGlacken,
Angew. Chem. Int. Ed. 2014, 53, 4142–4145; Angew. Chem. 2014, 126,
4226–4229; b) J. Zhang, C. Feng, Y. Deng, L. Liu, Y. Wu, B. Shen, C.
Zhong, W. Hu, Chem. Mater. 2014, 26, 1213–1218.
[9] For MCF synthesis, see: a) Y. Han, S. S. Lee, J. Y. Ying, Chem. Mater. 2006,
18, 643–649; b) M. Shakeri, K. Kawakami, Microporous Mesoporous
Mater. 2009, 118, 115–120; for loading of Pd nanoparticles and applica-
tions see: c) E. V. Johnston, O. Verho, M. D. Kärkäs, M. Shakeri, C.-W. Tai,
P. Palmgren, K. Eriksson, S. Oscarsson, J.-E. Bäckvall, Chem. Eur. J. 2012,
18, 12202–12206.
Acknowledgements
The Swedish Research Council (VR), the Swedish Governmental
Agency for Innovation Systems (VINNOVA), the Berzelii Center
EXSELENT, and the Knut and Alice Wallenberg Foundation are
gratefully acknowledged for project grants. The Knut and Alice
Wallenberg Foundation is also acknowledged for an equip-
ment grant for the electron microscopy facilities at Stockholm
University. The help of Dr. Fabian Carson for measuring XRPD
patterns at Stockholm University is gratefully acknowledged.
[10] For MOF synthesis see: a) G. FØrey, C. Mellot-Draznieks, C. Serre, F. Mill-
ange, J. Dutour, S. SurblØ, I. Margiolaki, Science 2005, 309, 2040–2042;
b) For NH2 post-synthetic functionalization see: S. Bernt, V. Guillerm, C.
Serre, N. Stock, Chem. Commun. 2011, 47, 2838–2840.
[11] a) C. Vriamont, T. Haynes, E. McCague-Murphy, F. Pennetreau, O. Riant, S.
Hermans, J. Catal. 2015, 329, 389–400; b) L. J. Durndell, C. M. A. Parlett,
N. S. Hondow, M. A. Isaacs, K. Wilson, A. F. Lee, Sci. Rep. 2015, 5, 9425.
[12] It should be noted that the high selectivity was observed with Pd0-
AmP-MCF with a Pd loading of 7.90 wt%.
Keywords: heterogeneous catalysis
metal–organic framework · palladium nanoparticles · selective
hydrogenation
· mesocellular foam ·
[13] F. Carson, V. Pascanu, A. Bermejo Gómez, Y. Zhang, A. E. Platero-Prats, X.
Zou, B. Martín-Matute, Chem. Eur. J. 2015, 21, 10896–10902.
[14] J. Gascon, A. Corma, F. Kapteijn, F. X. LlabrØs i Xamena, ACS Catal. 2014,
4, 361–378.
[15] a) V. Babu, A. Thapliyal, G. K. Patel, Biofuel Production, Wiley, New York,
2013; b) E. R. Riegel, J. A. Kent, Kent and Riegel’s Handbook of Industrial
Chemistry and Biotechnology, Springer, New York, 2007; c) M. A. Benve-
nuto, Industrial Chemistry: for Advanced Students, de Gruyter, Berlin,
2015; d) S. Lee, Methane and its Derivatives, Taylor & Francis, 1996.
[16] O. Verho, K. P. J. Gustafson, A. Nagendiran, C.-w. Tai, J.-E. Bäckvall, Chem-
CatChem 2014, 6, 3153–3159.
[1] G. Dormµn, L. Kocsis, R. Jones, F. Darvas, J. Chem. Health Saf. 2013, 20,
3–8.
[2] a) S. Fujiwara, N. Takanashi, R. Nishiyabu, Y. Kubo, Green Chem. 2014, 16,
3230–3236; b) S. Ganji, S. Mutyala, C. K. P. Neeli, K. S. R. Rao, D. R. Burri,
RSC Adv. 2013, 3, 11533–11538; c) Z. Wei, Y. Gong, T. Xiong, P. Zhang, H.
Li, Y. Wang, Catal. Sci. Technol. 2015, 5, 397–404; d) M. S. Ide, B. Hao, M.
Neurock, R. J. Davis, ACS Catal. 2012, 2, 671–683; e) J.-H. Yang, D. Ma,
RSC Adv. 2013, 3, 10131–10134; f) L. Chen, H. Chen, Y. Li, Chem.
Commun. 2014, 50, 14752.
[3] H. Furukawa, K. E. Cordova, M. O’Keeffe, O. M. Yaghi, Science 2013, 341,
1230444.
[4] For recent developments of catalytic processes with MOFs, see: a) A. H.
Chughtai, N. Ahmad, H. A. Younus, A. Laypkov, F. Verpoort, Chem. Soc.
Rev. 2015, 44, 6804–6849; for loading of Pd nanoparticles on function-
alized MOF: b) V. Pascanu, Q. Yao, A. Bermejo Gómez, M. Gustafsson, Y.
Yun, W. Wan, L. Samain, X. Zou, B. Martín-Matute, Chem. Eur. J. 2013, 19,
Received: February 24, 2016
Published online on April 25, 2016
Chem. Eur. J. 2016, 22, 7184 – 7189
www.chemeurj.org
7189
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim