In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle catalysts for hydrogenation reactions

Article abstract of DOI:10.1016/j.cattod.2017.02.049

Core@shell Fe@Fe_xO_y nanoparticles (NPs) have attracted a great deal of interest as potential magnetic supports for catalytic metals via galvanic exchange reactions. In this study Fe@Fe_xO_y/Pd bimetallic NPs were synthesized through galvanic exchange reactions using 50:1, 20:1 and 5:1 molar ratios of Fe@Fe_xO_y NPs to Pd(NO₃)₂. The resulting Fe@Fe_xO_y/Pd NPs have Pd NPs on the Fe oxide surfaces, and still retain their response to external magnetic fields. The materials could be recovered after the reaction by an external magnetic field, and agitation of the solution via a magnetic field led to improvements of mass transfer of the substrates to the catalyst surface for hydrogenation reactions. The Fe@Fe_xO_y/Pd NPs derived from the 5:1 molar ratio of their respective salts (Fe:Pd) exhibited a higher catalytic activity than particles synthesized from 20:1 and 50:1 molar ratios for the hydrogenation of 2-methyl-3-buten-2-ol. The highest turnover frequency reached 3600?h^?1 using ethanol as a solvent. In situ XANES spectra show that the Fe@Fe_xO_y NPs in the Fe@Fe_xO_y/Pd system are easily oxidized when dispersed in water, while they are very stable if ethanol is used as a solvent. This oxidative stability has important implications for the sustainable use of such particles in real world applications.

Full text of DOI:10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
In situ X-ray absorption spectroscopic studies of magnetic
Fe@FexOy/Pd nanoparticle catalysts for hydrogenation reactions
Yali Yaoa, Stefano Rubinob, Byron D. Gatesb, Robert W.J. Scotta,∗, Yongfeng Huc,∗
a Department of Chemistry, 110 Science Place, University of Saskatchewan, Saskatoon, SK S7N 5C9, Canada
b Department of Chemistry and 4D LABS, Simon Fraser University, Burnaby, BC V5A 1S6, Canada
c Canadian Light Source, University of Saskatchewan, Saskatoon, SK S7N 2V3, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 July 2016
Received in revised form 28 January 2017
Accepted 27 February 2017
Available online xxx
Core@shell Fe@FexOy nanoparticles (NPs) have attracted a great deal of interest as potential magnetic
supports for catalytic metals via galvanic exchange reactions. In this study Fe@FexOy/Pd bimetallic NPs
were synthesized through galvanic exchange reactions using 50:1, 20:1 and 5:1 molar ratios of Fe@FexOy
NPs to Pd(NO3)2. The resulting Fe@FexOy/Pd NPs have Pd NPs on the Fe oxide surfaces, and still retain their
response to external magnetic fields. The materials could be recovered after the reaction by an external
magnetic field, and agitation of the solution via a magnetic field led to improvements of mass transfer
of the substrates to the catalyst surface for hydrogenation reactions. The Fe@FexOy/Pd NPs derived from
the 5:1 molar ratio of their respective salts (Fe:Pd) exhibited a higher catalytic activity than particles
synthesized from 20:1 and 50:1 molar ratios for the hydrogenation of 2-methyl-3-buten-2-ol. The highest
turnover frequency reached 3600 h−1 using ethanol as a solvent. In situ XANES spectra show that the
Fe@FexOy NPs in the Fe@FexOy/Pd system are easily oxidized when dispersed in water, while they are
very stable if ethanol is used as a solvent. This oxidative stability has important implications for the
sustainable use of such particles in real world applications.
Keywords:
Fe@FexOy/Pd nanoparticles
in situ XANES study
Hydrogenation reactions
Solvent effect
© 2017 Elsevier B.V. All rights reserved.
1. Introduction
surfaces of Fe3O4 nanochains using poly(cryclotriphosphazene-
co-4,4ꢀ-sulfonyldiphenol) (PZS), and applied these nanochains as
Nanoparticles (NPs) can have dramatically different properties
compared to their bulk counterparts due to their large surface-
to-volume ratios and size-dependent electronic properties [1,2].
They have broad applications in catalysis, drug delivery, and envi-
ronmental remediation [3–6]. Because of their small sizes, the
separation of NPs from solutions after applications can be par-
ticularly problematic. Magnetic Fe or Fe oxide NPs have attracted
tremendous attention, because they can be isolated from the liq-
uid medium by simply applying an external magnetic field [7–9].
Moreover, Fe or Fe oxide NPs can respond to common mag-
netic stirrers, which opens up the possibilities of using them as
nanoscale magnetic stir bars for mass transfer acceleration in
microfluidic systems. For example, Chen’s group assembled Fe3O4
NPs to form 1D chains preserved in a shell of silica [10]. The
resulting Fe3O4 chains remain suspended and stir independently
within a small liquid droplet. Song’s group grafted Pd NPs on the
catalysts and nanometer-sized magnetic stir bars for microscopic
hydrogenation reactions [11]. Considering that metallic Fe NPs also
have interesting magnetic behavior, further studies into their mag-
netic performance and stability are required for applications in
catalysis and microfluidic research.
Both Fe(0) and Fe(II) have relatively low standard electrode
potentials (Fe2+/Fe0, Eo = −0.447 V, Fe3+/Fe2+, Eo = 0.771 V), so
½
½
they can be used to reduce another metal onto their surfaces to form
bimetallic NPs through galvanic exchange reactions [12]. Many
bimetallic NPs, such as FePd, FeCu, FeRu, have been synthesized
successfully by this method [13–16]. Recently, we showed that
Fe@FexOy NPs are very effective reductants for galvanic exchange
reactions which can fully reduce Pd or Cu salts onto their sur-
faces to form Pd or Cu NPs [17]. We also demonstrated that these
galvanic reactions can be monitored by in situ X-ray absorption
spectroscopy (XAS). Depending on the metal deposited on the
Fe@FexOy NPs, the resulting bimetallic NPs can be used to fur-
ther catalyze a variety of organic reactions. For example, Moores’
group has shown that Fe@FexOy@Cu NPs are active for the het-
erogeneous azide-alkyne click reactions and the cyclopropanation
of diazoesters with styrene derivatives [14,18], and Zhou and
∗
Corresponding authors.
E-mail addresses: robert.scott@usask.ca (R.W.J. Scott),
yongfeng.hu@lightsource.ca (Y. Hu).
http://dx.doi.org/10.1016/j.cattod.2017.02.049
0920-5861/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
2
Y. Yao et al. / Catalysis Today xxx (2017) xxx–xxx
Fig. 1. High-angle annular dark-field STEM images and EDX elemental maps of the Fe@FexOy/Pd NPs prepared from 20:1 (A and B) and 5:1 (C and D) molar ratios of Fe@FexOy
NPs to Pd (II).
−
coworkers have reported that Fe@FexOy@Pd NPs are active cat-
alysts for Suzuki-Miyaura cross-coupling reactions [13]. Pd can
catalyze not only C C coupling reactions, but also a myriad of
hydrogenation reactions, and Pd/C is widely used as a catalyst
for hydrogenation reactions [19–21]. A variety of supports have
also been investigated to further improve the catalytic activity,
stability and selectivity of Pd sites in different hydrogenation reac-
tions. He and coworkers reported that covalent trizaine (CTF)
framework-supported Pd NPs exhibit ca. 3.6 times faster reaction
rates than Pd/C for the catalytic hydrogenation of N-methylpyrrole;
the accelerated rate was attributed to intensified electronic inter-
actions between the Pd NPs and the CTF [22]. The Fe3O4-PZS-Pd
nanochains show much better stability during cycling tests for the
hydrogenation of styrene in comparison with commercial Pd/C,
because the interaction between PZS and Pd can efficiently pre-
vent the aggregation of Pd NPs and the magnetic Fe3O4 component
can assist with recovering these nanomaterials with almost no
loss [11]. Studies of Fe-supported Pd NPs for catalytic hydrogena-
tions are rare [23]. For many applications of the bimetallic NPs
based on Fe as mentioned above, it is of great interest to inves-
tigate the performance of Fe@FexOy@Pd NPs for hydrogenation
reactions.
The study of the nature of a catalyst and the identity of its
active site is important to improve its performance in catalytic
reactions. In situ and operando characterizations of catalysts have
been widely practiced by researchers, because they can directly
probe catalysts under working conditions. Among various in situ
techniques, X-ray absorption near edge structure (XANES) spectra
are sensitive to the oxidation state and coordination environment
of an element and can be used to probe short-range order within
materials, and thus it is a useful technique for the characterization
of metal NPs in solution [24–27]. We have previously shown that
both Fe metal oxidation and galvanic redox reactions of Fe@FexOy
particles with Pd(II) and Cu(II) can be followed by in situ XANES
[17,28]. Coordination environment changes can also be followed;
for example, recently we monitored the reaction of Pd(II) acetate
with Au25(SC8H9)18 clusters, and found that Pd(II) acetate was
converted to Pd(II) thiolate species through the reaction with S
−
atoms in the staple motifs (-S-Au-S-Au-S-) of the Au25(SC8H9)18
clusters as determined from the changes in the XANES spectra
[29]. Based on the results of in situ Au L3-edge XANES spectra on
supported-partially oxidized Au NPs, Haider et al. proposed that
metallic Au is the active species in the aerobic liquid-phase oxida-
tion of alcohols. The conversion of 1-phenylethanol in the oxidation
reaction increased with the concomitant reduction of Au species
and did not decrease even after the disappearance of oxidized Au
species [30].
Herein, we have synthesized Fe@FexOy/Pd NPs with different
molar ratios of Fe to Pd (50:1, 20:1 and 5:1) through galvanic
exchange reactions. We also applied the resulting Fe@FexOy/Pd
NPs for the hydrogenation of 2-methyl-3-buten-2-ol in a solution
of either water or ethanol, and showed they could be magneti-
cally recovered and/or agitated during the reaction. Additionally,
we carried out in situ XANES experiments to study catalytic
speciation in different solvents during the hydrogenation reac-
tion, to better understand differences in catalytic behavior in
the different solvents. The magnetic response of 50:1 and 20:1
Fe@FexOy/Pd NPs shows that they hold promise for the design
of real nanometer-sized magnetic stir bars for microscopic reac-
tions. The results of the studies indicated that Fe@FexOy/Pd NPs
have a higher catalytic activity in ethanol compared to water for
hydrogenation reactions, and in situ XANES experiments reveal
that these NPs are more stable in ethanol solution, whereas
further oxidation of the Fe cores occurs in the presence of
water.
2. Experimental section
2.1. Materials
All chemicals were used as received without further purifi-
cation. Iron (II) sulfate heptahydrate, methylene blue and
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
Y. Yao et al. / Catalysis Today xxx (2017) xxx–xxx
3
2-methyl-2-butanol were purchased from Sigma-Aldrich.
Poly(vinylpyrrolidone) (M. W. 58,000 g/mol), 2-methyl-3-buten-
2-ol and palladium (II) nitrate hydrate were purchased from Alfa
Aesar. Sodium borohydride, ethanol and methanol (HPLC grade)
were purchased from Fisher Scientific. Sulfuric acid was purchased
from EMD Millipore. Eighteen Mꢀ·cm Milli-Q water (Millipore,
Bedford, MA) was used for all syntheses.
in the sample can be displayed as an elemental map. By select-
ing several windows, different elemental maps can be obtained
simultaneously from a single scan, provided that the ionization
energies do not significantly overlap. False color maps can be also
superimposed to determine where two or more elements coexist
and in what relative amounts. Scanning time and beam intensity
were chosen as a compromise between signal-to-noise ratio and
radiation damage to the sample.
Fe K-edge and Pd L-edge XANES spectra were collected at the
Soft X-ray Microcharacterization Beamline (SXRMB) at the Cana-
dian Light Source (CLS). The measurements were conducted under
an ambient atmosphere for observing the Fe K-edge and under
helium for obtaining the Pd L3-edge spectra. Decreasing the beam
flux by defocusing and/or filtering the beam with Kapton filters and
stirring the sample via magnetic stirring were used to avoid sample
damage due to photoreduction. Liquid cells (SPEX CertiPrep Dispos-
able XRF X-Cell sample cups) were covered with a 4 ␮m ultralene
film (purchased from Fisher Scientific, Ottawa, ON) and used for
XANES analysis. The data were analyzed using the IFEFFIT software
package [31,32].
2.2. Synthesis of Fe@FexOy/Pd NPs
Fe@FexOy NPs were synthesized by reducing FeSO4·7H2O
(5.0 mmol) by NaBH4 (25 mmol) in a 1:1 water/methanol (v/v) mix-
ture (20 mL) in the presence of PVP (10 mmol based on monomer
unit) under nitrogen gas as previously reported [28]. First a 100 mL
round bottom flask with FeSO4·7H2O solution (8 mL in water) was
purged by nitrogen gas, then PVP solution (10 mL in methanol) was
added to the flask and the mixture was stirred for 10 min. Finally, a
solution of NaBH4 (2 mL in water) was added dropwise. After 30 min
stirring, 5.0 mL of 1 M H2SO4 was added to remove any excess
NaBH4 before injecting a Pd(II) nitrate solution (20 mM) to react
with the Fe@FexOy NPs by galvanic exchange reactions, following
our previous method [17].
The 1H NMR spectra were recorded in CDCl3 on a Bruker
500 MHz Advance NMR spectrometer. Chemical shifts were
recorded in parts per million (ppm), using the residual solvent peak
for calibration. The composition of the reaction mixture was ana-
lyzed by a gas chromatography (GC, Agilent Technologies 7890A)
with a flame ionization detector and a HP-5 capillary column
(30 m × 0.32 mm × 0.25 ␮m, J&W Scientific).
2.3. Hydrogenation of methylene blue
A solution of NaBH4 with Fe@FexOy/Pd NPs was prepared by
dispersing 0.50 mL of the above solution of Fe@FexOy/Pd NPs and
0.038 g sodium borohydride in 10 mL water. A 25 ␮L aliquot of
this mixture was injected into a 25 ␮L methylene blue solution
(0.050 mM), which was dropped onto a hydrophobic Teflon plate
located on top of a magnetic stirrer.
3. Results and discussions
3.1. Synthesis and STEM analysis of Fe@FexOy/Pd NPs
The Fe@FexOy NPs (16.0 6.0 nm) were synthesized by reduc-
ing FeSO4·7H2O with NaBH4 in a mixture of ethanol and water (1:1
volume ratio) containing PVP as a stabilizer [28]. Different molar
ratios of Fe@FexOy/Pd (prepared as 50:1, 20:1, and 5:1 based on
Fe:Pd molar ratios) NPs were synthesized by mixing the solution
of Fe@FexOy NPs with a Pd(II) nitrate solution under nitrogen gas
purging as described previously [17]. Fig. 1A and C shows dark field
STEM images of the particles synthesized at the 20:1 and 5:1 Fe:Pd
molar ratios, respectively. The Fe@FexOy particles self-assemble
into larger chain-like nanostructures, presumably due to magnetic
interactions between particles in solution (Fig. S1) [28]. As shown
in the STEM elemental maps (Fig. 1B), the 20:1 Fe@FexOy/Pd NPs
have isolated small Pd NPs (3 2 nm) deposited on the surfaces of
the Fe@FexOy NPs. In the system of 5:1 Fe@FexOy/Pd NPs most of
the Fe@FexOy NPs have been consumed in the galvanic exchange
formation, leaving a hybrid material that contains a fairly well-
dispersed Pd on the surfaces of these Fe oxidation supports (Fig. 1D).
It was very difficult to resolve individual Pd NPs in the samples
containing a higher Pd loading. These results agree with previous
EXAFS results on Fe@FexOy/Pd NPs from our group; the lowest first
shell Pd coordination numbers, and thus highest dispersions, were
seen for 5:1 Fe:Pd ratios [17].
2.4. Hydrogenation of 2-methyl-3-buten-2-ol
A solution of 0.010 mmol Fe@FexOy/Pd NPs (molar concentra-
tion determined based on Pd content) in 5.0 mL water was added
to a round bottom flask filled with 1.1 atm hydrogen gas. To this
solution, a 0.20 mL 2-methyl-3-buten-2-ol (1.9 mmol) solution was
added to obtain a 190:1 substrate to catalyst molar ratio. The pres-
sure was monitored by a differential pressure manometer (407910,
Extech Instrument), and the reaction progress was also followed by
1H NMR by extracting products from the reaction solutions with
2 mL aliquots of CDCl3. Turnover frequencies (TOFs) were mea-
sured by plotting either product conversion or H2 consumption (as
a proxy for product conversion, a direct correspondence between
NMR results and H2 pressure decay was always observed) over
reaction time. These results are reported in either (moles product
formed/moles Pd)/time or (moles H2 consumed/moles Pd)/time.
2.5. Characterization
Transmission electron microscopy (TEM) analyses of the NPs
were initially conducted using a HT7700 microscope (Hitachi
High-Technologies) operating at 100 kV. The samples were pre-
pared by drop-casting one drop of dilute, aqueous sample onto
a carbon-coated 200 mesh copper grid (Electron Microscopy Sci-
ences, Hatfield, PA).
3.2. Magnetic recovery/agitation with Fe@FexOy/Pd NPs
Elemental maps were obtained using an FEI Osiris STEM
equipped with ChemiSTEM Technology integrating the signal from
four Energy Dispersive X-ray (EDX) spectrometers. A 200 kV elec-
tron beam was focused to a nanometer-sized spot and scanned
across the sample to excite electrons from the core shells. As the
electron beam is raster scanned across the sample, a full X-ray
spectrum is collected for each pixel. By selecting the appropriate
energy window, a specific element is selected and its distribution
Fe supported bimetallic NPs have been applied in a variety
of catalytic reactions [13–16,33,34]. The presence of Fe in these
bimetallic NPs not only allows them to be easily separated from
reaction solutions by an external magnet, but can also act as a redox
scavenger to redeposit the leached metal and avoid the contami-
nation of the organic product. Fe atoms have a strong magnetic
moment due to their high number of unpaired 3d electrons, thus
Fe and Fe oxide materials exhibit strong magnetic properties [35].
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
4
Y. Yao et al. / Catalysis Today xxx (2017) xxx–xxx
Scheme 1. The hydrogenation of 2-methyl-3-buten-2-ol.
Fig. 3. The rate of hydrogen consumption in the hydrogenation of 2-methyl-3-
buten-2-ol in water using the 50:1 (red circle), 20:1 (blue triangle) and 5:1 (black
square) Fe@FexOy/Pd NPs as catalysts. (For interpretation of the references to color
in this figure legend, the reader is referred to the web version of this article.)
(MB) within small liquid droplets. Solutions containing either 50:1
or 20:1 Fe@FexOy/Pd NPs and NaBH4 were injected into the MB
droplet. MB could be reduced by the NaBH4 and converted to a col-
orless leucomethylene blue solution in 30 s when agitated by the
50:1 or 20:1 Fe@FexOy/Pd NPs (Fig. S4). In comparison, it was dif-
ficult to reduce MB by the addition of NaBH4 without the stirring
of NPs. This study reveals that the agitation of the 50:1 and 20:1
Fe@FexOy/Pd NPs using an external magnetic field can effectively
mix the reactants and improve the reaction rate.
Fig. 2. The conversion of 2-methyl-3-buten-2-ol to 2-methyl-2-butanol in the
hydrogenation reaction (black square) using the 50:1 Fe@FexOy/Pd NPs as catalysts
without stirring, and (orange dot) using 50:1 Fe@FexOy/Pd NPs as both catalysts and
magnetic stirrers in a 5 mL solution. (For interpretation of the references to color in
this figure legend, the reader is referred to the web version of this article.)
The large surface area to volume ratio of Fe and Fe oxide NPs can
result in increased remanence and coercivity and can give these NPs
improved magnetic properties in comparison to their bulk counter-
parts [36,37]. The fact that the original Fe@FexOy NPs self-assemble
into larger domains may allow them to be used for stirring at the
microscopic scale. Indeed this is the case for the Fe@FexOy/Pd NPs
synthesized with 50:1 and 20:1 molar ratios of Fe@FexOy NPs to
Pd(II). The resulting NPs have strong responses to the magnetic
field of a common stir plate, and allow for a solution to be stirred
in the absence of a macroscopic stir bar (Fig. S2). However, the
Fe@FexOy/Pd NPs synthesized with a 5:1 molar ratio of Fe@FexOy
NPs to Pd(II) do not agitate a solution upon magnetic stirring.
In addition, we examined the relative ease of magnetically sepa-
rating each of the NP systems from solution. The 50:1 and 20:1
Fe@FexOy/Pd NPs could be separated by an external magnet from
the reaction solutions after 9 s (Fig. S3), while the separation of
the 5:1 Fe@FexOy/Pd NPs from solution takes a much longer time
(more than 30 s). These results indicate the possibility of using
these bimetallic catalysts as magnetically recoverable catalysts and
magnetic stirrers, provided that the Pd loading is not too high.
To further test whether Fe@FexOy/Pd NPs can improve the reac-
tion rate for catalytic reactions by alleviating mass transfer issues,
we applied the Fe@FexOy/Pd NPs as catalysts and microscopic
stirrers for the hydrogenation of 2-methyl-3-buten-2-ol in water
under a 1.1 atm H2 gas at 25 ◦C (Scheme 1). The addition of the
50:1 Fe@FexOy/Pd NPs as catalysts yielded a TOF of 21 h−1 for the
hydrogenation of 2-methyl-3-buten-2-ol to 2-methyl-2-butanol as
determined by NMR in the absence of stirring, whereas the TOF
was improved to 34 h−1 by stirring the solution using the 50:1
Fe@FexOy/Pd NPs (Fig. 2). This result confirms that the NPs can
improve mass-transfer of the substrates to the catalyst surface. As
the temperature of the reaction flask was kept constant at 25 ◦C, we
do not believe that magnetic heating was a factor. As this test was
performed on a 5 mL sample and not a droplet with a small volume,
the difference between the stirred and non-stirred states is not sig-
nificant. A microscopic model test system was also carried out using
Fe@FexOy/Pd NPs to catalyze the hydrogenation of methylene blue
3.3. Catalytic behavior in the hydrogenation of
2-methyl-3-buten-2-ol
Having established that the Fe@FexOy/Pd NPs can be effective
catalysts, we sought to optimize the catalyst system in terms of the
Pd loading, solvent composition, and long-term stability of the cat-
alyst. For these studies larger volumes of solutions were used along
with a macroscopic magnetic stir bar (egg shaped, 7/8 in. × 3/8 in.)
rotated at 1600 rpm. Fig. S5 shows that the rotation with a macro-
scopic magnetic stir bar can speed up the reaction rate much more
at higher temperatures compared to the reaction rates at the same
temperatures without any stirring (Fig. S6), which shows that stir-
ring can minimize mass transfer effects for the hydrogenation
reactions and thus allow a better comparison of actual catalytic
activities. The progress of the reaction was monitored by measur-
ing the decrease in pressure of H2 filled in a round bottom flask as
a function of reaction time. When Fe@FexOy NPs (in the absence of
Pd) were used as the catalyst for the hydrogenation of 2-methyl-
3-buten-2-ol in water at 1.1 atm H2 (g) and 25 ◦C, there was no
decrease in the pressure of H2. However, Fe@FexOy/Pd NP catalysts
(50:1, 20:1 and 5:1 Fe:Pd molar ratios) exhibit high TOFs. As shown
in Fig. 3, the reaction follows a zero order rate law, and reaches
completion once all the substrate is consumed. The decrease of the
moles of the H2 gas in the reaction from 0.0482 mol to 0.0463 mol
reveals that the consumption of H2 is 1:1 with respect to the sub-
strate 2-methyl-3-buten-2-ol (1.9 mmol used for the reaction). The
5:1 Fe@FexOy/Pd NPs showed the highest TOF (ca. 1300 h−1) in
comparison to the 50:1 and 20:1 Fe@FexOy/Pd NPs (TOF of 385 h−1
and 785 h−1, respectively). As noted earlier, Fe@FexOy NP controls
showed no activity for hydrogenations, confirming that Pd is the
active catalyst for this reaction. The higher TOF of 5:1 Fe@FexOy/Pd
NPs compared to 50:1 and 20:1 Fe@FexOy/Pd NPs suggests that
the Pd is more dispersed and has a higher surface area for the
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
Y. Yao et al. / Catalysis Today xxx (2017) xxx–xxx
5
Fig. 4. The rates of hydrogen consumption in the hydrogenation of 2-methyl-
3-buten-2-ol in water using the 20:1 Fe@FexOy/Pd NPs recycled for the use in
sequential catalytic cycles.
5:1 Fe@FexOy/Pd NPs, which is in agreement with the results of
elemental EDX maps shown earlier.
The reusability of the Fe@FexOy/Pd NPs was also examined by
using 20:1 Fe@FexOy/Pd NPs as an example. These NPs were recy-
cled by isolating the NPs from solution using centrifugation. The
results of this evaluation are summarized in Fig. 4. Surprisingly, the
Fe@FexOy/Pd NPs exhibited a higher catalytic activity upon recy-
cling. However, it also became successively more difficult to fully
recover the material with an external magnet. The TEM images
show that there are fewer and/or significantly smaller Fe@FexOy
NPs remaining in the bimetallic NPs system as the number of cycles
increased (Fig. 5). The result suggests that the Fe@FexOy NPs are
degrading over time. This conclusion is also supported by ex situ
XANES measurements, as shown in Fig. 6. There is almost no change
in the speciation of the zerovalent Pd NPs as judged by compar-
ing the Pd L3-edge XANES spectra after different recycling tests
(Fig. 6A), while the Fe K-edge XANES spectra show a significant shift
to higher energies upon using the catalysts over 6 cycles (Fig. 6B).
These results are consistent with the progressive oxidation of the
Fe@FexOy NPs to Fe(II) and Fe(III) oxides [28].
To determine whether the oxidation of Fe@FexOy NPs takes
place during the hydrogenation reaction or during the recycling
process, in situ XANES studies were carried out. The in situ setup is
shown in Fig. 7, and consisted of a liquid cell loaded with Fe@FexOy
NPs and 2-methyl-3-buten-2-ol in water, and placed inside an alu-
minum box which was located on top of a magnetic stirrer (not
shown) to ensure the constant mixing of the reaction mixture.
Hydrogen gas was bubbled into the liquid cell to start the reac-
Fig. 6. The Pd L-edge (A) and Fe K-edge (B) XANES spectra of the 20:1 molar ratio
Fe@FexOy/Pd NPs.
tion and the reaction was monitored at the Pd L3 and Fe K edges,
as shown in Fig. 8. No significant change of the in situ Pd L3-
edge XANES spectra was observed (Fig. 8A), which is not overly
surprising as Pd typically remains in the zerovalent state dur-
ing hydrogenation reactions [38]. However, slight oxidation of the
Fe@FexOy NPs could be observed from in situ Fe K-edge XANES
spectra (Fig. 8B). Through fitting the Fe K-edge XANES spectra
using a linear combination fitting with standards (Fe foil, FeSO4
and Fe(NO3)3) as we did in our previous work [28], it is revealed
that the Fe@FexOy NPs in the Fe@FexOy/Pd NP systems contain 56%
Fe(0) (Table S1). These standards were chosen as they led to the
best fits of the experimental data; we were not able to fit the data
Fig. 5. TEM images of the 20:1 Fe@FexOy/Pd NPs: (A) before hydrogenation reaction; (B) after 4 cycles; and (C) after 6 cycles.
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
6
Y. Yao et al. / Catalysis Today xxx (2017) xxx–xxx
Fig. 7. The hydrogenation setup for in situ fluorescence XANES studies.
Fig. 9. The Fe K-edge XANES spectra of the 20:1 Fe@FexOy/Pd NPs in the hydrogena-
tion reaction using ethanol as a solvent.
Fig. 10. The rates of hydrogen consumption for the hydrogenation of 2-methyl-3-
buten-2-ol using the 5:1 Fe@FexOy/Pd NPs as catalysts in either water or ethanol.
oxidized to Fe(III) during the recycling process when the samples
are exposed to air.
In order to address this oxidation damage to the NPs, we
switched the solvent from water to ethanol for the hydrogena-
tion reactions, as ethanol should not react with Fe(0). The Pd
L3-edge XANES spectra indicate that the Pd NPs have no change
before and after the hydrogenation reaction in ethanol (Fig. S7).
There is also no change observed in the in situ Fe K-edge XANES
spectra (Fig. 9). These results confirm that the Fe@FexOy NPs are
quite stable in ethanol. The Fe@FexOy/Pd NPs also showed a much
higher catalytic activity in ethanol than in water, as shown in
Fig. 10. The 5:1 Fe@FexOy/Pd NPs showed improvements in TOF
from ca. 1300 h−1 in water to 3600 h−1 in ethanol, an increase
of 2.8 times in magnitude. The NMR and GC results confirmed
that the conversion of 2-methyl-3-buten-2-ol to 2-methylbutan-
2-ol is completed in 4 min when catalyzed by 5:1 Fe@FexOy/Pd
NPs with a 190:1 substrate to catalyst ratio in ethanol. No change
was seen in the TOF for the hydrogenation reaction upon recy-
cling the catalysts for three cycles in ethanol, and no significant
changes were seen by TEM (Fig. S8). This catalytic result is a very
high TOF for the hydrogenation of 2-methyl-3-buten-2-ol under
ambient conditions, and compares favourably with many excel-
lent hydrogenation catalysts in the literature [40–43]. The factors
determining solvent influence in the hydrogenation reactions have
Fig. 8. The Pd L-edge (A) and Fe K-edge (B) XANES spectra of the 20:1 molar ratio
Fe@FexOy/Pd NPs in the hydrogenation reaction using water as a solvent.
using standards involving different oxide and hydroxide phases of
Fe. This Fe(0) content decreases to 49% after 30 min of the hydro-
genation reaction in water, and all the Fe(0) was oxidized after 6
cycles (to 60% Fe(II) and 40% Fe(III), the fitted data from Fig. 6B,
Table S2). This result indicates that the oxidation of Fe@FexOy NPs
happens during the reaction. We note that water is the solvent for
this reaction; water can react with exposed Fe in Fe@FexOy NPs
and oxidize Fe(0) to Fe(II), which is likely the reason for oxidation
of the Fe@FexOy NPs in the reaction [39]. The Fe(II) is then further
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049

G Model
CATTOD-10638; No. of Pages7
ARTICLE IN PRESS
Y. Yao et al. / Catalysis Today xxx (2017) xxx–xxx
7
been studied by many groups [38,44–48]. The solubility of H2 (g)
seems to be the most significant factor influencing the rate of this
reaction along with the high stability of the support in ethanol.
The solubility of H2 (2.98 × 10−3 mol/L) in ethanol is significantly
higher than that in water (0.81 × 10−3 mol/L) [49]. These results
are also consistent with Nikoshvili group’s studies on the selective
and solvent dependent hydrogenation of 2-methyl-3-butyn-2-ol to
2-methyl-3-buten-2-ol over Pd NPs stabilized in hypercrosslinked
polystyrene [38].
[6] R. Sachan, A. Malasi, J. Ge, S. Yadavali, H. Krishna, A. Gangopadhyay, H. Garcia,
G. Duscher, R. Kalyanaraman, ACS Nano 8 (2014) 9790–9798.
[7] A. Schejn, T. Mazet, V. Falk, L. Balan, L. Aranda, G. Medjahdi, R. Schneider,
Dalton Trans. 44 (2015) 10136–10140.
[8] R. Hudson, Y. Feng, R.S. Varma, A. Moores, Green Chem. 16 (2014) 4493–4505.
[9] A.Y. Li, M. Kaushik, C.-J. Li, A. Moores, ACS Sustainable Chem. Eng. 4 (2016)
965–973.
[10] W.H. Chong, L.K. Chin, R.L.S. Tan, A.Q. H.Wang, H. Liu, Angew. Chem. Int. Ed.
52 (2013) 8570–8573.
[11] S. Yang, C. Cao, Y. Sun, P. Huang, F. Wei, W. Song, Angew. Chem. Int. Ed. 54
(2015) 2661–2664.
[12] CRC Handbook of Chemistry and Physics, 96th ed., CRC Press, Cleveland, OH,
2010.
[13] S. Zhou, M. Johnson, J.G.C. Veinot, Chem. Commun. 46 (2010) 2411–2413.
[14] R. Hudson, C.J. Li, A. Moores, Green Chem. 14 (2012) 622–624.
[15] R. Hudson, V. Chazelle, M. Bateman, R. Roy, C.J. Li, A. Moores, A.C.S. Sustain,
Chem. Eng. 3 (2015) 814–820.
[16] Y. Zhang, Y. Su, X. Zhou, C. Dai, A.A. Keller, J. Hazard Mater. 263 (2013)
685–693.
[17] Y. Yao, C. Patzig, Y. Hu, R.W.J. Scott, J. Phys. Chem. C 119 (2015) 21209–21218.
[18] S. Ishikawa, R. Hudson, M. Masnadi, M. Bateman, A. Castonguay, A. Moores,
C.J. Li, Tetrahedron 70 (2014) 6162–6168.
[19] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for
Organic Synthesis, John Wiley & Sons, New York, 2001.
[20] P.N. Rylander, Hydrogenation Methods, Academic Press, New York, 1985.
[21] H. Sajiki, T. Ikaea, K. Hirota, Tetrahedron Lett. 42 (2001) 7699–7701.
[22] T. He, L. Liu, G. Wu, P. Chen, J. Mater. Chem. A 3 (2015) 16235–16241.
[23] J. Yuan, Pesticides 46 (2007) 827–829.
[24] R.M. Anderson, L. Zhang, J.A. Loussaert, A.I. Frenkel, G. Henkelman, R.M.
Crooks, ACS Nano 10 (2013) 9345–9353.
[25] L. Wang, S. Zhang, Y. Zhu, A. Patlolla, J. Shan, H. Yoshida, S. Takeda, A.I.
Frenkel, F. Tao, ACS Catal. 3 (2013) 1011–1019.
[26] J. Ohyama, K. Teramura, T. Shishido, Y. Hitomi, K. Kato, H. Tanida, T. Uruga, T.
Tanaka, Chem. Phys. Lett. 507 (2011) 105–110.
4. Conclusion
A suspension of either 50:1 or 20:1 Fe@FexOy/Pd NPs can act
as both magnetically recoverable catalysts and magnetic stirrers to
improve the mixing of reactants in catalytic hydrogenation reac-
tions. For studies on the catalytic activity of the Fe@FexOy/Pd NPs
for the hydrogenation of 2-methyl-3-buten-2-ol, the 5:1 molar
ratio Fe@FexOy/Pd NPs exhibit a high TOF of 3600 h−1 using ethanol
as the solvent. In situ XANES spectra show that water as a sol-
vent in the hydrogenation reaction can oxidize the Fe(0) core in
Fe@FexOy/Pd NPs, while the same Fe@FexOy/Pd NPs are very stable
when dispersed in ethanol. This result shows that while these com-
posite materials have viability as low cost, magnetically recoverable
catalysts, they do not have long-term stability in water which may
limit their possible applications.
[27] A. Maclennan, A. Banerjee, Y. Hu, J.T. Miller, R.W.J. Scott, ACS Catal. 3 (2013)
1411–1419.
Acknowledgements
[28] Y. Yao, Y. Hu, R.W.J. Scott, J. Phys. Chem. C 118 (2014) 22317–22324.
[29] A. Shivhare, K.E. Lee, Y. Hu, R.W.J. Scott, J. Phys. Chem. C 119 (2015)
23279–23284.
The authors thank William Barrett for assistance with hydro-
genation reactions. The authors acknowledge financial assistance
from the National Sciences and Engineering Research Council of
Canada “(NSERC Discovery Grant No. 1077758) and the Canada
Research Chairs Program (B.D. Gates, Grant No. 950-215846).
XANES experiments described in this paper were performed at
the Canadian Light Source, which is supported by NSERC, the
National Research Council Canada, the Canadian Institutes of Health
Research, the Province of Saskatchewan, Western Economic Diver-
sification Canada, and the University of Saskatchewan. This work
also made use of 4D LABS (www.4dlabs.ca) shared facilities sup-
ported by the Canada Foundation for Innovation (CFI), British
Columbia Knowledge Development Fund (BCKDF), Western Eco-
nomic Diversification Canada, and Simon Fraser University.
[30] P. Haider, J.D. Grunwaldt, R. Seidel, A. Baiker, J. Catal. 250 (2007) 313–323.
[31] M. Newville, J. Synchrotron Radiat. 8 (2001) 322–324.
[32] B. Ravel, M. Newville, J. Synchrotron Radiat. 12 (2005) 537–541.
[33] S. Kovács, K. Zih-Perényi, Á. Révész, Z. Novák, Synthesis 44 (2012) 3722–3730.
[34] S. Kovács, Z. Novák, Org. Biomol. Chem. 9 (2011) 711–716.
[35] P. Wasilewski, G. Kletetschka, Geophys. Res. Lett. 26 (1999) 2275–2278.
[36] B. Issa, I.M. Obaidat, B.A. Albiss, Y. Haik, Int. J. Mol. Sci. 14 (2013)
21266–21305.
[37] D.K. Kim, Y. Zhang, W. Voit, K.V. Rao, M. Muhammed, J. Magn. Magn. Mater.
225 (2001) 30–36.
[38] L. Nikoshvili, E. Shimanskaya, A. Bykov, I. Yuranov, L. Kiwi-Minsker, E. Sulman,
Catal. Today 241 (2015) 179–188.
[39] X. Li, D.W. Elliott, W. Zhang, Rev. Crit. Env. Sci. Technol. 31 (2006) 111–122.
[40] R. Ma, N. Semagina, J. Phys. Chem. C 114 (2010) 15417–15423.
[41] J. Shen, N. Semagina, ACS Catal. 4 (2014) 268–279.
[42] A. Banerjee, R.W.J. Scott, Green Chem. 17 (2015) 1597–1604.
[43] Y. Niu, L.K. Yeung, R.M. Crooks, J. Am. Chem. Soc. 123 (2001) 6840–6846.
[44] E. Sadeghmoghaddam, H. Gu, Y.S. Shon, ACS Catal. 2 (2012) 1838–1845.
[45] N.M. Bertero, A.F. Trasarti, C.R. Apesteguia, A. Marchi, J. Appl. Catal. A: Gen.
394 (2011) 228–238.
References
[1] M. Haruta, T. Kobayashi, H. Sano, N. Yamada, Chem. Lett. 2 (1987) 405–408.
[2] R.V. Jagadeesh, A. Surkus, H. Junge, M. Pohl, J. Radnik, H. Huan, V.
Schünemann, A. Brückner, M. Beller, Science 342 (2013) 1073–1076.
[3] H. Zhang, M. Jin, Y. Xia, Chem. S oc. Rev. 41 (2012) 8035–8049.
[4] T. Sun, Y. Zhang, B. Pang, D.C. Hyun, M. Yang, Y. Xia, Angew. Chem. Int. Ed. 53
(2014) 12320–12364.
[46] R.A. Rajadhyskasha, S.I. Karwa, Chem. Eng. Sci. 41 (1986) 1765–1770.
[47] R.L. Augustine, R.W. Warner, M.J. Melnick, J. Org. Chem. 49 (1984) 4853–4856.
[48] E. Toukoniitty, P. Mäki-Arvela, J. Kuusisto, V. Nieminen, J. Päivärinta, M.
Hotokka, T. Salmi, D. Yu Murzin, J. Mol. Catal. 192 (2002) 135–151.
[49] M. Ghavre, S. Morrissey, N. Gathergood, Hydrogenation in ionic liquids, Ionic
Liq. Appl. Perspect. (2011) 331–392.
[5] A. Kaur, U.J. Gupta, Mater. Chem. 19 (2009) 8279–8289.
Please cite this article in press as: Y. Yao, et al., In situ X-ray absorption spectroscopic studies of magnetic Fe@FexOy/Pd nanoparticle
catalysts for hydrogenation reactions, Catal. Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.02.049