Zn-Co double metal cyanides as heterogeneous catalysts for hydroamination: A structure-activity relationship

Article abstract of DOI:10.1021/cs300805z

Zn-Co double metal cyanide (DMC) materials are effective heterogeneous catalysts for intermolecular hydroaminations. Using the reaction of 4-isopropylaniline with phenylacetylene as a test, the effect of different catalyst synthesis procedures on the cata

Full text of DOI:10.1021/cs300805z

Research Article
pubs.acs.org/acscatalysis
Zn−Co Double Metal Cyanides as Heterogeneous Catalysts for
Hydroamination: A Structure−Activity Relationship
Annelies Peeters,† Pieterjan Valvekens,† Rob Ameloot,† Gopinathan Sankar,‡ Christine E. A. Kirschhock,†
and Dirk E. De Vos*,†
†Centre for Surface Chemistry and Catalysis, KU Leuven, Kasteelpark Arenberg 23, 3001 Leuven, Belgium
‡Department of Chemistry, University College London, 20 Gordon Street, London WC1H 0AJ, U.K.
S
* Supporting Information
ABSTRACT: Zn−Co double metal cyanide (DMC) materials are effective
heterogeneous catalysts for intermolecular hydroaminations. Using the
reaction of 4-isopropylaniline with phenylacetylene as a test, the effect of
different catalyst synthesis procedures on the catalytic performance is
examined. The best activities are observed for double metal cyanides with a
cubic structure and prepared with a Zn2+ excess, and for nanosized particles
prepared via a reverse emulsion synthesis. Detailed study of the active Zn2+
sites in the cubic material by EXAFS gives evidence for coordinative
vacancies around the Zn, with four cyanide ligands in close proximity of the
Zn. The substrate scope of the hydroaminations was successfully expanded
to both aromatic and aliphatic alkynes and other aromatic and aliphatic
amines. Even with styrenes the reaction proceeded with aromatic amines. The DMC catalysts are truly heterogeneous, possess a
high thermal stability and are perfectly reusable.
KEYWORDS: amination, zinc, intermolecular, EXAFS, heterogeneous catalysis, alkenes
looked at solid catalysts.1 Reports exist on metal-cation
INTRODUCTION
■
exchanged zeolites,4 clays5,6 and heteropolyacids,7−9 and on
The one-step amination of alkenes or alkynes, called hydro-
the Brønsted acid forms of zeolites,10−12 clays,13 ion exchange
amination, is a sustainable route to fine chemicals.1 It is a 100%
atom-efficient process without stoichiometric formation of
resins,14,15 and silica gel.16 Furthermore there are reports on
supported (complexes of) Pd,17,18 Au,19 Cu,20 Zn,21 V,22 and
byproducts (Scheme 1). Generally, the reaction can provide
organolanthanides23,24 on solid carrier materials. In most cases,
two regioisomers, of which the Markovnikov product is
thermodynamically favored.1
hydroamination is limited to intramolecular reactions or more
reactive alkyne substrates or activated alkenes instead of
nonactivated alkenes. For materials that contain Brønsted
Scheme 1. Hydroamination of Alkenes
acid sites the activity is limited by protonation of the amine.
High temperatures and pressures are needed to counteract this,
like in the tert-butylamine production with pentasil and β-type
zeolites, one of the few commercial applications of hydro-
amination.25
Recently, our group has shown that double metal cyanides
(DMCs) and more in particular Zn−Co DMCs are active
Furthermore, hydroamination utilizes some of the most
heterogeneous catalysts for the hydroamination of phenyl-
readily available carbon sources, while synthesis pathways
acetylene with 4-isopropylaniline.26 DMCs are Prussian blue
toward high value amines in industry from alternative starting
analogues, which are synthesized by the addition of a water-
compounds, such as alcohols, alkyl halides, or carbonyl
soluble metal salt (e.g., ZnCl2) to a water-soluble metal cyanide
compounds, overall employ a larger number of steps.2,3 The
salt (e.g., K3[Co(CN)6]). DMCs are mainly used as catalysts in
obtained amines, imines, or enamines have widespread
epoxide polymerization, which is a large scale industrial
applications as bulk chemicals, specialty chemicals or
process.27−29 Reactions like the ring-opening polymerization
(ROP) of propylene oxide and copolymerization of CO2 and
pharmaceuticals. The high activation barriers of hydroamina-
tions however demand the use of efficient catalyst systems. A
generic catalyst for both intra- and intermolecular hydro-
amination of alkyne and alkene feedstock is still lacking. The
desire for a heterogeneous system constitutes an additional
challenge, and hydroamination research has only scarcely
Received: December 12, 2012
Revised: February 11, 2013
© XXXX American Chemical Society
597
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
epoxides have received considerable literature attention.30−34
While DMCs have been reported to be useful for trans-
esterifications,35 hydrolysis of oils,36 other epoxide ring-opening
reactions,37 Prins condensation reactions,38 and synthesis of
hyperbranched polymers,39 they remain a generally underex-
plored class of catalytic materials. Here, we report on a detailed
exploration of the substrate scope of DMCs in hydroamination
reactions, and we elucidate the relation between structure and
activity of the materials. An EXAFS analysis was performed to
acquire a detailed insight into the environment of the active Zn
sites of the catalyst. This technique gives detailed local
information, which is crucial for understanding the structure−
activity relationship.
polyoxyethylene isooctylphenyl ether type surfactant, is used
both as a surfactant for stabilizing the aqueous droplets and as
co-CA (NDMC-Igepal, with N = nano).
The structure of the DMCs is, despite extensive research for
epoxide polymerization, still a matter of debate. The [Co-
(CN)6]3− anion and the Zn2+ cations build a structure which is
terminated by the complexing agents, or by water molecules.
Some preliminary characterization of some materials were
already reported before,26 and relevant data were selected to
illustrate the properties of these materials. X-ray powder
diffraction analysis was performed on the DMC samples
(Figure 1). This shows that all DMC samples are crystalline.
Most show reflections at 14.9°, 17.1°, 24.3°, 34.8°, and 39.1°
2θ, typical of a cubic lattice structure with space group Fm3m,
as was described by Mullica et al.41 DMC-PTMEG, DMC-P123
and NDMC-Igepal display the most pronounced phase purity
of the cubic phase. However, additional lines were observed for
several materials, for instance when PEG was used as the co-
CA, or when Zn compounds other than ZnCl2 were used.
Reflections at 14.5°, 17.0°, 20.6°, 23.7°, and 24.7° 2θ are readily
assigned to the monoclinic phase with space group P11m, as
described by Wojdel et al.42 Despite the close overlap of some
of the lines arising from the two different phases, the
observation of reflections at e.g. 20.6° and 23.7° 2θ is clearly
diagnostic for the presence of the monoclinic phase. The
relation between the 3D structure of the DMC materials and
their catalytic activity will be addressed in detail below (vide
infra).
Selected SEM images of DMC-PTMEG, DMC-pure and
NDMC-Igepal show that the synthesis procedure affects
particle size, morphology and particle size distribution of the
catalyst to a great extent (Figure 2). Comparison of DMC-
PTMEG and DMC-pure, both synthesized via the standard
procedure, shows the importance of using complexing agents.
DMC-pure shows large crystalline particles, many of which
present a morphology of truncated octahedra, consistent with
the cubic symmetry. In contrast, DMC-PTMEG displays
smaller but highly coagulated and nonuniform particles. The
SEM image of NDMC-Igepal however shows very small,
uniformly sized particles. This indicates that the reverse
emulsion technique indeed allows for a more controlled and
fine particle size of the DMC material.
Pyridine adsorption followed by infrared (IR) spectroscopy
was performed on NDMC-Igepal. Figure 3 shows the difference
spectrum of NMDC-Igepal before and after adsorption of
pyridine. The IR bands at 1610 and 1451 cm−1 prove the
presence of Lewis acid sites, while Brønsted acid sites are
absent, as indicated by the absence of peaks at 1639 and 1545
cm−1.43 This shows that Zn−Co DMCs are strictly Lewis acid
solid catalysts. The catalytically active Lewis acid sites are
presumed to be associated with Zn sites in the DMC
structure.37 Further evidence for this will be given below.
As some of the hexacyanocobaltate anions are missing in the
cubic structure due to charge considerations (vide infra), the
structure has substantial porosity, giving rise to high internal
surface areas. The porous framework has voids of about 8.5 Å,
but they are connected through small windows of only about
∼4.2 Å diameter.44 In view of these properties, double metal
cyanides have been studied for hydrogen storage and also for
light alkane separations, etc.45,46 An indication of the pore size
can be obtained by performing nitrogen sorption on the DMC
samples. Horvath−Kawazoe (HK) pore size estimations yielded
average values of about 4.8 Å. However, as most reactants for
RESULTS AND DISCUSSION
■
Synthesis and Characterization of Zn−Co DMCs. Zn−
Co DMC catalysts were prepared by various synthesis
techniques, as described in the Experimental Section. An
overview of the relevant synthesis parameters is given in Table
1. The Zn(II)−Co(III) catalysts in this paper are typically
Table 1. Overview of Synthesis Parameters of Prepared Zn−
Co−DMCs
catalyst
Co-CA
PTMEG
CA
Zn source
ZnCl2
DMC-PTMEG
DMC-PEG
DMC-P123
DMC-X
tBuOH
tBuOH
tBuOH
tBuOH
−
PEG
ZnCl2
P123
ZnCl2
−
−
−
ZnCl2
DMC-pure
DMC-1:1
ZnCl2
−
ZnCl2
DMC-sulfate
DMC-triflate
NDMC-Igepal
PTMEG
PTMEG
Igepal CA-520
tBuOH
tBuOH
tBuOH
ZnSO4.7H2O
Zn(CF3SO3)2
ZnCl2
synthesized by the addition of an excess of ZnCl2 to
K3[Co(CN)6], but other possible metal combinations are
Cu(II)−Co(III), Zn(II)−Fe(II), Zn(II)−Fe(III), Cu(II)−Fe-
(II), Ni(II)−Co(III), etc. Additives have been proposed to
make the DMCs more active. A complexing agent (CA), often
an alcohol like tert-butanol (tBuOH), is used alongside
cocomplexing agents (co-CA) which are polyethers like
poly(tetramethylene ether) glycol (PTMEG), polyethylene
glycol (PEG) or Pluronic P123 (containing polyethylene glycol
and polypropylene glycol segments, P123). The role of these
complexing agents is multifold, but they are believed to affect
the active site accessibility and particle size of the DMCs by
coordinating on peripheral metal cations and simultaneously
acting as protecting agents.33 DMC-PTMEG, DMC-PEG, and
DMC-P123 were prepared analogously, but with different co-
t
CAs, while DMC-X was prepared with only BuOH, in the
absence of a co-CA. For DMC-pure, no alcohol additives were
used at all. This is also the case for DMC-1:1, but here the ratio
of zinc to cobalt in the synthesis mixture was 1:1, while mostly
a Zn/Co ratio of 10:1 was applied. To investigate the influence
of the anion of the used zinc salt, ZnCl2 was replaced by
ZnSO4·7H2O and Zn(CF3SO3)2, resulting in DMC-sulfate and
DMC-triflate, respectively.
A higher hydroamination activity is expected for small
particles.26 Smaller particle size can be obtained by using a
reverse emulsion technique, in which an aqueous phase is
dispersed in a cyclohexane phase.40 Igepal CA-520, a
598
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
Figure 2. SEM images of (a) DMC-PTMEG, (b) DMC-pure, and (c)
NDMC-Igepal. Reproduced with permission from ref 26. Copyright
2011 The Royal Society of Chemistry.
Figure 3. IR difference spectrum: Subtraction of absorbance before
and after adsorption of pyridine to NDMC-Igepal. Pyridine was
adsorbed at 323 K (25 mbar); spectrum was recorded after evacuation
at 448 K.
the hydroaminations contain aromatic rings, they are expected
to react mainly on the outer particle surface. The nitrogen
sorption isotherms obtained are shown in Figure 4, for a
standard material (DMC-PTMEG) and for NDMC-Igepal
prepared in a reverse emulsion. BET surface areas of
respectively 464 and 552 m2/g were obtained. For NDMC-
Igepal, the steeper rise of the isotherm at intermediate relative
pressures is indicative of a larger external surface area, in
accordance with the smaller overall particle size of this
material.26
Thermogravimetric analysis of the DMCs was performed
under oxygen flow. The weight loss and its derivative as a
function of the temperature are shown in Figure 5 for DMC-
PTMEG, DMC-pure, DMC-X, and NDMC-Igepal as repre-
sentative materials. All of them show a very sharp mass decay
when the cyanide ligands are burned off. TGA analysis also
Figure 1. X-ray diffractograms of synthesized DMC catalysts: (a)
DMC-PTMEG, (b) DMC-PEG, (c) DMC-P123, (d) DMC-X, (e)
DMC-pure, (f) DMC-1:1, (g) DMC-sulfate, (h) DMC-triflate, and (i)
NDMC-Igepal.
599
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
Figure 4. Nitrogen sorption isotherms (77 K) of (a) DMC-PTMEG and (b) NDMC-Igepal.
Figure 5. TGA and DTGA profiles for (a) DMC-PTMEG, (b) DMC-pure, (c) DMC-X, and (d) NDMC-Igepal.
shows that physisorbed water is removed fast in a first stage
(<100 °C), followed by removal of BuOH, the co-CA and
indicates there are hardly traces of ZnCl2 left. In view of the
excess of Zn2+ in the synthesis mixture and in the structure, it is
expected that zinc terminates the crystals and is abundant on
the outer surface.
Catalytic Activity and Substrate Scope. To probe the
catalytic activity of the different DMC catalysts, a test reaction
was used. The hydroamination of 4-isopropylaniline 1 and
phenylacetylene 2 results in the formation of the enamine
hydroamination product (Scheme 2), which further tautomer-
izes to the thermodynamically more stable imine product (4-
isopropylphenyl)(1-phenylethylidene)amine 3. The reaction
t
finally the cyanide ligands, which under oxidizing conditions are
removed as (CN)2.47 At this stage the Zn−Co cyanide salt is
converted to the corresponding oxides Co2O3 and ZnO, which
is known as Rinmann’s green, and is easily recognized by the
dark green color.48
Elemental analysis was performed with EDX (Figure 6). The
Zn/Co/Cl molar ratio calculated is 1.85:1:0.03, for DMC-
PTMEG, meaning that the Zn:Co ratio is slightly higher than
the stoichiometrically expected 1.5. The low chlorine content
600
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
Figure 6. EDX analysis of Zn−Co−DMC−PTMEG. Note that the Au signal originated from Au sputtering on the sample for conductivity.
features as evident from Figure 1 a-c: the best activities are
found for nearly phase-pure cubic materials. While the reason
for the poor results with PEG is not immediately clear, it should
be noted that the molecular weight (MW) of the PEG used
(2000 g/mol) is significantly higher than that of the PTMEG
(1000 g/mol). The presence of the PTMEG co-CA also has a
beneficial effect on the hydroamination activity in comparison
with DMC-X, prepared without any co-CA (Table 2, entry d vs
Scheme 2. Test Reaction: Hydroamination of
Phenylacetylene with 4-Isopropylaniline
was conducted at 383 K for 24 h in the presence of the DMC
catalyst in toluene solvent. All reactions resulted in the
Markovnikov-type addition product, as expected for Lewis
acidic catalysts.1 For detailed experimental procedures and
characterization of products, see Supporting Information.
Earlier, we reported that Zn(II)−Co(III) catalysts were the
most active DMCs, while Zn(II)−Fe(II), Zn(II)−Fe(III),
Cu(II)−Fe(II), and Cu(II)−Co(III) showed negligible cata-
lytic activity in the hydroamination test reaction.26 Table 2 lists
t
a). Finally, omission of the CA BuOH resulted again in a low
hydroamination activity, which in this case can be related to the
low outer surface area available for catalysis (entry e). NDMC-
Igepal, prepared by the reverse emulsion technique, was the
most active DMC material. Again, the high activity seems to be
connected to the cubic structure, combined in this case with a
large outer surface (entry h).
Assuming that the zinc sites of the Zn−Co DMC-PTMEG
material are the active sites for the hydroamination reactions,
the choice of the Zn salt for catalyst preparation may affect the
catalytic activity. The nature of the zinc precursor was indeed
found to have a large influence: sulfate- or triflate-based Zn salts
resulted in DMCs with much lower hydroamination activities
(Table 2, entries f and g vs a). In first approximation, this anion
effect can again be traced back to the phases formed: the cubic
phase is the predominant one formed in the presence of ZnCl2,
while significant fractions of side phases are present when other
precursors are employed (Figure 1a, g, and h). This is in
agreement with literature; Huang and co-workers also found an
effect of the metal salt precursor on the structure of the
DMC.30 Only with ZnCl2 a cubic structure was obtained, while
with ZnSO4 and Zn(CH3COO)2 a different phase was formed.
For polymerization of propylene oxide only the cubic form was
highly effective, while the others had a low activity.30 On the
other hand, it cannot be excluded that some chloride ions
would remain coordinated to the Zn2+ in the final catalyst.
However, from EDX measurements (Figure 6) it can be
concluded that this amount of Cl− on the surface is only a trace.
As ZnCl2 clearly plays a crucial role in the catalyst preparation,
the catalytic activity of ZnCl2 was also tested as such, even if
ZnCl2 is only partially soluble in the reaction conditions.
However, no hydroamination products were found (entry i),
indicating that the dispersed, coordinated zinc sites in Zn−Co
DMCs are essential for activity. K3[Co(CN)6] was equally
inactive (entry j). This shows that the catalytic activity of the
DMCs is not merely a summation of the catalytic effect of the
respective metals, but results from the unique structure of the
Table 2. Hydroamination Yields (%) of 3 Obtained with
Different Zn−Co-Based DMC Catalysts
a
b
entry
a
catalyst
yield (%)
92
DMC-PTMEG
c
92
b
c
d
e
f
DMC-PEG
DMC-P123
DMC-X
18
67
79
40
8
DMC-pure
DMC-sulfate
DMC-triflate
NDMC-Igepal
g
h
i
28
96
0
d
ZnCl2
d
j
K3[Co(CN)6]
0
a
Conditions: 1 mmol of 4-isopropylaniline and 0.5 mmol of
phenylacetylene with 50 mg of catalyst in 1 mL of toluene. Reactions
b
were carried out at 383 K for 24 h. Calculation of yield based on
c
phenylacetylene, GC yield. Yield of hydroamination product after
d
third consecutive run. 5 mol % of the metal salt (with respect to
phenylacetylene) was added.
the hydroamination yields for the DMC catalysts as well as for
their metal salt precursors. Poly(tetramethylene ether) glycol
(PTMEG) was found to be one of the most suitable co-CAs,
with the Pluronic coblock polymer (P123) based material being
somewhat less active (Table 2, entries a and c). A low activity
was found for the polyethylene glycol (PEG) containing
material (entry b). These trends coincide with the structural
601
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
a
Table 3. Hydroamination Yields (%) Obtained with Different Amines and Alkynes
a
Conditions: 1 mmol of amine and 1 mmol of alkyne with 50 mg of NDMC-Igepal catalyst in 1 mL of toluene. Reactions were carried out at 423 K
b
c
for 24 h. Calculation of yield based on amine, GC yield. 0.5 mmol of amine instead of 1 mmol and reaction at 473 K.
DMC and the influence of this structure on the properties of
the Zn2+ ion.
appears that an electron donating substituent in the para
position of the aniline results in higher hydroamination yields.
These substituents make the amine more nucleophilic for
attack on the alkyne moiety. In entries c, d, and e, an aliphatic
amine or alkyne is used as substrate. The hydroamination of 1-
octyne with 4-isopropylaniline (Table 3, entry c) proceeds
smoothly with the NDMC-Igepal catalyst. Aliphatic amines
appear to be more difficult reactants. Reacting 1-hexylamine
with phenylacetylene at 423 K, a yield of only 5% was obtained
(entry d). However, when working with an excess of the
phenylacetylene at 473 K this yield was raised to 48%. At this
temperature there was however a significant byproduct
formation because of phenylacetylene oligomerization. Also
with 1-octyne a 46% yield of the hydroamination product could
be obtained (Table 3, entry e). Heterogeneous catalysts that
can catalyze the intermolecular reaction of aliphatic alkynes
with aromatic amines have been reported, but aliphatic amines
have till now only rarely been used successfully. This indicates
that DMCs are general and versatile catalysts for alkyne
hydroaminations.
For DMCs in epoxide polymerization, it has been found that
a ZnCl2 excess with respect to K3[Co(CN)6] in the synthesis
beneficially affects the activity. In DMC literature this is often
attributed to the higher activity of a less crystalline material, and
the use of a co-CA is supposed to have a similar effect.31,32
However, our data do not indicate that the most active
materials have a decreased crystallinity. Rather, both the ZnCl2
excess and the CAs result in materials with small particle size,
with Zn preferably located on the external surface. XRD
patterns show that both the 1:1 material and the DMC
prepared with an excess of ZnCl2 display sharp peaks (Figure 1,
entry e and f). While most catalysts presented in Table 1 were
prepared with a 10:1 Zn:Co ratio, the material prepared using a
1:1 Zn:Co ratio in the synthesis was also tested. This resulted
in an activity decrease by more than 20%, suggesting that Zn
cations terminating the lattice at the external surface play an
important role in the catalysis.
NDMC-Igepal, prepared by the reverse emulsion technique,
yielded very small, uniform and phase pure particles which
showed the highest activity obtained so far. These were
employed in further experiments to expand the substrate scope
of the catalyst. Table 3 shows the hydroamination yields of
aromatic and aliphatic amines with both aromatic and aliphatic
alkynes. The reactions were conducted with equimolar amounts
of amine and alkyne at 423 K for 24 h. From entries a and b it
While the results with alkynes are encouraging, the
intermolecular hydroamination of olefins in mild reaction
conditions remains challenging. Therefore, a range of styrenes
were tested in the hydroamination with anilines (Table 4).
Again reactions were conducted at 423 K for 24 h. Promising
results were obtained with 4-methoxy and 4-ethoxystyrene.
When working with 0.5 mmol of 4-isopropylaniline instead of 1
602
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
a
Table 4. Hydroamination Yields (%) Obtained with Different Anilines and Styrenes
a
Conditions: 1 mmol of aniline and 1 mmol of styrene with 50 mg of NDMC-Igepal catalyst in 1 mL of toluene. Reactions were carried out at 423 K
b
c
d
for 24 h. Calculation of yield based on amine, GC yield. 0.5 mmol of aniline instead of 1 mmol. Combined yield.
mmol, a 27% yield could be obtained with 4-methoxystyrene.
The lower concentration of the amine probably facilitates
coordination of the styrene to the active Lewis acid sites, giving
a higher hydroamination yield. This is in line with literature
findings on hydroamination by late transition metals. For Zn2+
ion exchanged Beta-zeolite, Penzien et al.4 suggested that the
mechanism comprises a nucleophilic attack of the amine on the
metal-coordinated unsaturated bond. This coordination makes
the π-system susceptible for attack by a nucleophile.
In comparison with the nonsubstituted styrene, the hydro-
amination yield is higher with electron donating substituents in
the para position (Table 4, entries f, g, and h). When a styrene
with an electron withdrawing substituent in this position was
tested, such as 4-bromostyrene (not shown), no hydro-
amination product was formed at all, indicating that an
increased electron density on the styrene double bond
enhances the hydroamination rate. This may indicate that a
sufficient electron density, as provided by alkoxy substituents, is
required to allow coordinative bonding of the double bond on
the Zn sites. With the sterically demanding 2,4,6-trimethylstyr-
ene only very little hydroamination product was obtained
(<1%), just like for 4-tert-butyl styrene (<1%). When 4-
isopropylaniline was replaced with aniline, good yields were
obtained, but next to the hydroamination product also the
hydroarylation product was retrieved (Table 4, entries i and j).
Further tests show that Zn−Co DMCs are stable recyclable
heterogeneous catalysts. This was demonstrated by a hot
filtration test, a recycling test and structure analysis. In the hot
filtration test a reaction mixture was prepared analogously to
the conditions used in Table 2 with DMC-PMEG catalyst. A
sample was taken after 2 h of reaction and filtered without
cooling. The resulting filtrate was left stirring for an additional 4
h and no further catalytic conversion was observed, indicating
that no active species leach from the heterogeneous catalyst.
DMC-PTMEG also kept its high activity after three consecutive
runs (see Table 2, entry a), indicating that it is a stable and
recyclable catalyst.
Figure 7 shows the X-ray diffractograms of NDMC-Igepal
before reaction and after reaction at 383, 423, and 473 K. This
clearly demonstrates that the structure of the catalyst is
preserved at the first two reaction temperatures. However, at
the high reaction temperatures of 473 K, changes in the catalyst
structure occur after 24 h, which is most likely due to
prolonged contact with solvent or reactants at this reaction
temperature. We can therefore conclude that NDMC-Igepal is
a stable catalyst up to the reaction temperature of 423 K.
603
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
Figure 7. X-ray diffractograms of NDMC-Igepal before reaction (a),
after reaction at 383 K (b), after reaction at 423 K (c) and after
reaction at 473 K (d).
Figure 8. X-ray diffractograms of Zn−Co DMC samples: (a) highly
cubic material, (b) simulated model of a cubic Zn−Co DMC
according to Mullica et al.,41 (c) highly monoclinic synthesized
material, and (d) simulated model of a monoclinic Zn−Co DMC
according to Wojdel et al.42
Structure−Activity Relationship. So far, the most active
catalytic materials all display a cubic crystal structure with
substantial phase purity. The cubic Fm3m model as proposed
by Mullica et al. demands vacancies in the octahedral lattice due
to the unequal charge of the hexacyanocobaltate anion
[Co(CN)6]3− and the Zn2+ cation, which creates open
coordination sites associated with Zn.41 In the monoclinic
phase however, Zn is tetrahedrally coordinated and no free
coordination sites are encountered.42
Materials with high phase purity were prepared for further
investigation. For the cubic structure a procedure described by
Kuyper and Boxhoorn was followed without addition of
complexing agents.49 The X-ray diffractogram of this material
is shown in Figure 8a and is compared with the cubic structure
model based on Mullica et al.41 (Figure 8, b), which confirms
that the material is purely cubic. Subsequently, an attempt was
made to prepare a material with a high content of the
monoclinic side phase. This was accomplished with a lower Zn/
Co ratio and the use of zinc sulfate instead of zinc chloride as
Zn source.45,50 The X-ray diffractogram is shown in Figure 8c.
Reflections at 20.6° and 23.7° 2θ together with high intensity of
the reflection at 14.5° 2θ are diagnostic for the monoclinic
phase.
the monoclinic side phase (82% vs 42%). This is another
indication of the higher activity of a cubic phase in comparison
with the monoclinic side phase.
The cubic DMC of Figure 8a was characterized by EXAFS.
From a detailed study of both the Co-edge and Zn-edge data,
the environments of the Co ions and the active Zn sites can be
described more precisely. The results of the Co-edge EXAFS
analysis are presented in Figure 9. The Fourier transformed
data fit very well with the proposed cubic model for the first
two coordination shells around Co: six carbon atoms can be
fitted at a distance of 1.90 Å and further 6 nitrogen atoms at
3.00 Å. The high intensity of the second shell is due to the
linear arrangement of the cyanide ligand; a Co−C−N bond
angle of 180° enhances forward scattering and results in high
intensities of the radial distribution curve. This confirms that
the hexacyanocobaltate anions are incorporated in intact
octahedral coordination geometry in the structure of the
catalyst; cobalt is fully coordinated, leaving no sites for catalysis.
The intact 6-fold coordination of cobalt in the DMC complex is
supported by the strong bonding nature of the CN− ligand;
strong σ-donation and π-backbonding result in a stable
complex. In this context the hexacyanocobaltate anion plays a
role of structure builder by placing the cyano-ligands in exactly
When comparing the activity of these materials with high
phase purity for the test reaction of 1 and 2, it was again found
that the cubic form was the best catalyst. The yield of the
hydroamination product 3 could be almost doubled with the
cubic form rather than using the material with a high fraction of
604
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
Figure 10. Zn−K-edge data: (a) k3-weighted raw EXAFS data, (b)
Fourier transform fit, and (c) local environment around Zn in the Zn−
Co DMC structure. Four ligands are located in close proximity, while a
fifth neighbor is at a nonbonding distance (not shown).
Figure 9. Co−K-edge data: (a) k3-weighted raw EXAFS data, (b)
Fourier transform fit, and (c) local environment around Co in the
Zn−Co DMC structure.
the right position in the double metal cyanide complex to allow
catalysis at the Zn-sites.
This means that the number of nitrogen neighbors around Zn
can theoretically vary between 2 and 6. EXAFS data indeed
demonstrate that the number of neighbors around Zn is well
below 6. A few feasible structure models were tested and a
satisfactory fit was obtained with 4 nitrogen neighbors in close
proximity to the Zn, even if models with 5 or 3 nitrogen
neighbors also appeared possible. The best fit obtained is
shown in Figure 10c in which 4 nitrogen neighbors are present
at a close distance of 1.98 Å and a fifth neighbor is further away
The Zn-edge data are shown in Figure 10. While Zn should
be in an octahedral site, according to the cubic model, the
structure should contain a large number of hexacyanocobaltate
vacancies due to the charge neutrality of the system. This would
lower the actual coordination number around Zn from 6 to an
average of 4. The observed average Fm3m structure implies a
random distribution of these vacancies throughout the lattice.
605
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
at a nonbonding distance of 2.48 Å, resulting in the small
shoulder visible in the Fourier transform radial distribution.
This model is based on the unsymmetrical environment of
the ligands around zinc in an incompletely coordinated
octahedral position which lacks an inversion center. As the
material was prepared without any CAs or co-CAs, and was
vacuum dehydrated prior to EXAFS analysis, the Zn in the
dehydrated cubic material probably adopts a pseudotetrahedral
coordination, in which it is moving slightly out of its position in
the octahedral lattice. This implies that the angle of the Zn−N
bonds is not 180° as proposed from the diffraction based
model, but is locally distorted. This results in smaller Zn−N
distances compared to the cubic crystallographic model of
Mullica et al. (1.96 vs 2.09 Å) while a fifth neighbor is at a
larger distance (2.48 vs 2.09 Å). This EXAFS study confirms
that free coordination sites exist, which allow the Zn cations to
function as Lewis acid catalytic sites.
dissolving 10 mmol of ZnCl2 in 2 mL of deionized water, while
solution C contained 2 mmol of K3[Co(CN)6] in 2 mL of
deionized water. Solution B was first added dropwise to
solution A under vigorous stirring. Next, solution C was added
dropwise to the mixture, and left stirring for another hour at
room temperature. The resulting particles were centrifuged and
washed three times with tBuOH (NDMC-Igepal). The powder
was then dried first at 333 K followed by drying at 353 K under
vacuum overnight.
Instrumentation. XRD measurements were performed on
a STOE Stadi MP operating in Bragg−Brentano mode.
Measurements were carried out with CuKα radiation with a
wavelength of 1.54060 Å using a linear position sensitive
detector. SEM images were recorded on a Philips XL30 FEG
after coating with a thin film of Au, followed by EDX
recordings at 15 kV for chemical analysis. The EDX data were
processed using EDAX software. IR measurements were
performed on a Nicolet 6700 spectrometer using thin self-
supporting wafers ( 10 mg cm−2), prepared by application of
pressure on the powder sample. The wafer was placed in a cell
under vacuum and was heated to 448 K before an IR spectrum
was recorded. The cell was cooled down to 323 K and pyridine
was allowed to diffuse into the cell. Physisorbed pyridine was
removed by heating up to 448 K under vacuum, before
recording the IR spectrum. Nitrogen physisorption measure-
ments were conducted at 77 K on a Quantachrome AS1Win
instrument, after evacuation of the powder sample overnight at
373 K in vacuum. TGA measurements were carried out on a
TGA Q500 of TA Instruments with a heating rate of 10 °C/
min under O2-atmosphere, in a high throughput mode. Zn and
Co K-edge X-ray absorption spectra (XAS) were collected at
B18 at the Diamond Light Source which operates at 3GeV. The
beamline is equipped with Si(111) double crystal mono-
chromator, ion chambers for measuring incident and trans-
mitted beam intensity and a 9 element Ge fluorescence detector
for measurement in fluorescence mode. All the measurements
were carried out in fluorescence mode and typically 6 scans
were averaged to produce the spectra. X-ray absorption spectra
were processed using ATHENA51 software and subsequent
analyses of the EXAFS data were performed using EX-
CURVE.52
CONCLUSION
■
Zn−Co double metal cyanides are excellent catalysts for a wide
range of intermolecular hydroamination reactions, particularly
using alkynes. They are easily prepared, stable, reusable and
truly heterogeneous. The high activity of the material is
ascribed to Lewis acid zinc sites; typically the Zn is in a pseudo-
octahedral environment, but instead of 6, only 4 or 5 cyanide
ligands surround the Zn, as shown by an EXAFS study.
Although double metal cyanides are underexplored in
heterogeneous organic catalysis, these catalysts offer a high
potential for many applications because of the high level of
possible variation in e.g. active metals, particle size, morphology
and hence activity.
EXPERIMENTAL SECTION
■
General Procedure for the Synthesis of Zn−Co DMC
Catalysts. Standard DMCs.33 Solution A is prepared by
dissolving 10 mmol of a metal salt (ZnCl2, ZnSO4·7H2O
(DMC-sulfate) or Zn(CF3SO3)2 (DMC-triflate)) in 100 mL of
deionized water. To this solution, 1 mmol of cocomplexing
agent (co-CA) is added. This co-CA can be poly-
(tetramethylene ether) glycol (1000 g mol−1; DMC-
PTMEG), polyethylene glycol (2000 g mol−1; DMC-PEG) or
Pluronic P123 triblock polymer (5750 g mol−1; DMC-P123).
Solution A is then stirred magnetically at 353 K. Solution B is
prepared by dissolving 1 mmol of a metal cyanide salt
K3[Co(CN)6] in 10 mL of deionized water. Solution B is
added dropwise to solution A under vigorous stirring. After the
addition, the mixture is left to stir for another 30 min at 353 K.
ASSOCIATED CONTENT
■
S
* Supporting Information
General procedure for the catalysis experiments, 1H NMR
spectroscopic data, and GC-MS fragmentation data of all
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
t
Hereafter, 25 mL of the complexing agent (CA) BuOH is
added and the final mixture is stirred for another 3 h at 353 K.
The synthesis can be repeated without addition of co-CA
(DMC-X) or without addition of both CA and co-CA (DMC-
pure). DMC-1:1 was also prepared without addition of CAs,
but with equimolar amounts of Zn and Co. The resulting solids
were centrifuged and washed three times with a 1:1 mixture of
tBuOH and water. They are subsequently dried at 333 K
overnight, followed by drying at 353 K under vacuum
overnight.
AUTHOR INFORMATION
Corresponding Author
*E-mail: dirk.devos@biw.kuleuven.be. Fax: (+32) 16 321 998.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
A.P. acknowledges the Fund for Scientific Research Flanders
(FWO-Vlaanderen) and L’Orea
financial support (Aspirant of FWO). D.E.D.V. thanks the
Belgian Federal Government (IAP 7/05) and KU Leuven for
the Methusalem CASAS grant. The authors thank the Diamond
Light Source (U.K.) for beam time at the B18 beamline.
Reverse Emulsion Synthesis.40 Solution A was prepared by
the addition of 10 g of Igepal CA-520 emulsifier and 7.7 mL of
́
l/UNESCO Belgium for
t
CA BuOH to cyclohexane (109 mL) constituting the apolar
phase. This solution was stirred at room temperature with a
mechanical stirrer (300 rpm). Solution B was prepared by
606
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607

ACS Catalysis
Research Article
(30) Huang, Y. J.; Qi, G. R.; Chen, L. S. Appl. Catal., A 2003, 240,
263−271.
ABBREVIATIONS
■
DMC, double metal cyanide; CA, complexing agent; co-CA,
cocomplexing agent; tBuOH, tert-butanol; PTMEG, poly-
(tetramethylene ether) glycol; PEG, polyethylene glycol;
P123, Pluronic P123; NDMC, nano double metal cyanide;
Igepal, Igepal CA520; EXAFS, extended X-ray absorption fine
structure
(31) Zhang, X. H.; Hua, Z. J.; Chen, S.; Liu, F.; Sun, X. K.; Qi, G. R.
Appl. Catal., A 2007, 325, 91−98.
(32) Lee, S. H.; Lee, I. K.; Ha, J. Y.; Jo, J. K.; Park, I.; Ha, C. S.; Suh,
H.; Kim, I. Ind. Eng. Chem. Res. 2010, 49, 4107−4116.
(33) Lee, I. K.; Ha, J. Y.; Cao, C.; Park, D. W.; Ha, C. S.; Kim, I.
Catal. Today 2009, 148, 389−397.
(34) Dienes, Y.; Leitner, W.; Muller, M. G. J.; Offermans, W. K.;
Reier, T.; Reinholdt, A.; Weirich, T. E.; Muller, T. E. Green Chem.
2012, 14, 1168−1177.
REFERENCES
■
(1) Muller, T. E.; Hultzsch, K. C.; Yus, M.; Foubelo, F.; Tada, M.
Chem. Rev. 2008, 108, 3795−3892.
(2) Brunet, J. J.; Neibecker, D. In Catalytic Heterofunctionalization;
(35) Sreeprasanth, P. S.; Srivastava, R.; Srinivas, D.; Ratnasamy, P.
Appl. Catal., A 2006, 314, 148−159.
(36) Satyarthi, J. K.; Srinivas, D.; Ratnasamy, P. Appl. Catal., A 2011,
391, 427−435.
Togni, A., Grutzmacher, H., Eds.; Wiley-VCH: Weinheim, Germany,
̈
2001, p 91−141.
(37) Saikia, L.; Satyarthi, J. K.; Gonnade, R.; Srinivas, D.; Ratnasamy,
P. Catal. Lett. 2008, 123, 24−31.
(3) Hayes, K. S. Appl. Catal., A 2001, 221, 187−195.
(4) Penzien, J. C.; Haessner, C.; Jentys, A.; Kohler, K.; Muller, T. E.;
Lercher, J. A. J. Catal. 2004, 221, 302−312.
(38) Patil, M. V.; Yadav, M. K.; Jasra, R. V. J. Mol. Catal. A: Chem.
2007, 273, 39−47.
(5) Shanbhag, G. V.; Halligudi, S. B. J. Mol. Catal. A: Chem. 2004,
222, 223−228.
(39) Sebastian, J.; Srinivas, D. Chem. Commun. 2011, 47.
(40) Yi, M. J.; Byun, S. H.; Ha, C. S.; Park, D. W.; Kim, I. Solid State
Ionics 2004, 172, 139−144.
(6) Joseph, T.; Shanbhag, G. V.; Halligudi, S. B. J. Mol. Catal. A:
Chem. 2005, 236, 139−144.
(41) Mullica, D. F.; Milligan, W. O.; Beall, G. W.; Reeves, W. L. Acta
Crystallogr., Sect. B 1978, 34, 3558−3561.
(7) Shanbhag, G. V.; Halligudi, S. B. Open Org. Chem. J. 2008, 2, 52−
57.
(42) Wojdel, J. C.; Bromley, S. T.; Illas, F.; Jansen, J. C. J. Mol. Model.
2007, 13, 751−756.
(8) Lingaiah, N.; Babu, N. S.; Reddy, K. M.; Prasad, P. S. S.;
Suryanarayana, I. Chem. Commun. 2007, 278−279.
(9) Yang, L.; Xu, L. W.; Xia, C. G. Tetrahedron Lett. 2008, 49, 2882−
2885.
(43) Emeis, C. A. J. Catal. 1993, 141, 347−354.
(44) Krap, C. P.; Zamora, B.; Reguera, L.; Reguera, E. Microporous
Mesoporous Mater. 2009, 120, 414−420.
(10) Mizuno, N.; Tabata, M.; Uematsu, T.; Iwamoto, M. J. Catal.
1994, 146, 249−256.
(45) Autie-Castro, G.; Autie, M.; Reguera, E.; Santamaria-Gonzalez,
J.; Moreno-Tost, R.; Rodriguez-Castellon, E.; Jimenez-Lopez, A. Surf.
Interface Anal. 2009, 41, 730−734.
(11) Lequitte, M.; Figueras, F.; Moreau, C.; Hub, S. J. Catal. 1996,
163, 255−261.
(46) Krap, C. P.; Balmaseda, J.; del Castillo, L. F.; Zamora, B.;
Reguera, E. Energy Fuels 2010, 24, 581−589.
(12) Jimenez, O.; Muller, T. E.; Schwieger, W.; Lercher, J. A. J. Catal.
2006, 239, 42−50.
(47) Zboril, R.; Machala, L.; Mashlan, M.; Sharma, V. Cryst. Growth
Des. 2004, 4, 1317−1325.
(13) Motokura, K.; Nakagiri, N.; Mori, K.; Mizugaki, T.; Ebitani, K.;
Jitsukawa, K.; Kaneda, K. Org. Lett. 2006, 8, 4617−4620.
(14) Bhanushali, M. J.; Nandurkar, N. S.; Bhor, M. D.; Bhanage, B.
M. Catal. Commun. 2008, 9, 425−430.
(48) Natta, G.; Passerini, L. Gazz. Chim. Ital. 1929, 59, 620−642.
(49) Kuyper, J.; Boxhoorn, G. J. Catal. 1987, 105, 163−174.
(50) Rodriguez-Hernandez, J.; Reguera, E.; Lima, E.; Balmaseda, J.;
Martinez-Garcia, R.; Yee-Madeira, H. J. Phys. Chem. Solids 2007, 68,
1630−1642.
(15) Qureshi, Z. S.; Deshmukh, K. M.; Tambade, P. J.; Dhake, K. P.;
Bhanage, B. M. Eur. J. Org. Chem. 2010, 6233−6238.
(16) Liu, P. N.; Xia, F.; Zhao, Z. L.; Wang, Q. W.; Ren, Y. J.
Tetrahedron Lett. 2011, 52, 6113−6117.
(17) Richmond, M. K.; Scott, S. L.; Yap, G. P. A.; Alper, H.
Organometallics 2002, 21, 3395−3400.
(18) Tada, M.; Shimamoto, M.; Sasaki, T.; Iwasawa, Y. Chem.
Commun. 2004, 2562−2563.
(51) Ravel, B.; Newville, M. J. Synchrotron Radiat. 2005, 12, 537−
541.
(52) Gurman, S. J.; Binsted, N.; Ross, I. J. Phys. C: Solid State Phys.
1984, 17, 143−151.
(19) Corma, A.; Gonzalez-Arellano, C.; Iglesias, M.; Navarro, M. T.;
Sanchez, F. Chem. Commun. 2008, 6218−6220.
(20) Shanbhag, G. V.; Joseph, T.; Halligudi, S. B. J. Catal. 2007, 250,
274−282.
(21) Duncan, C.; Biradar, A. V.; Asefa, T. ACS Catal. 2011, 1, 736−
750.
(22) Maurya, M. R.; Arya, A.; Kumar, U.; Kumar, A.; Avecilla, F.;
Pessoa, J. C. Dalton Trans. 2009, 9555−9566.
(23) Zhao, J. Q.; Marks, T. J. Organometallics 2006, 25, 4763−4772.
(24) Le Roux, E.; Liang, Y. C.; Storz, M. P.; Anwander, R. J. Am.
Chem. Soc. 2010, 132, 16368−16371.
(25) Holderich, W.; Taglieber, V.; Pohl, H. H.; Kummer, R.; Baur, K.
̈
G. Process for amine manufacture. German Patent DE3634247C1,
1987.
(26) Peeters, A.; Valvekens, P.; Vermoortele, F.; Ameloot, R.;
Kirschhock, C.; De Vos, D. Chem. Commun. 2011, 47, 4114−4116.
(27) Le-Khac, B.; Chester, W. Highly active double metal cyanide
catalysts. U.S. Patent 5,714,428, 1998.
(28) Hoffmann, J.; Ehlers, S.; Klinksiek, B.; Klesczewski, B.; Steinlein,
C.; Obendorf, L.; Pielartzik, H.; Pazos, J. F. Process for producing
polyether polyols. U.S. Patent 2002/0198278 A1, 2002.
(29) Ostrowski, T.; Ruppel, R.; Baum, S. Process for preparing
polyether polyols. U.S. Patent 2006/223979 A1, 2006.
607
dx.doi.org/10.1021/cs300805z | ACS Catal. 2013, 3, 597−607