Molecularly Tailored Nickel Precursor and Support Yield a Stable Methane Dry Reforming Catalyst with Superior Metal Utilization

Article abstract of DOI:10.1021/jacs.7b01625

Syngas production via the dry reforming of methane (DRM) is a highly endothermic process conducted under harsh conditions; hence, the main difficulty resides in generating stable catalysts. This can, in principle, be achieved by reducing coke formation, s

Full text of DOI:10.1021/jacs.7b01625

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Molecularly-Tailored Nickel Precursor and Support Yield a Stable
Methane Dry Reforming Catalyst with Superior Metal Utilization
Tigran Margossian, Kim Larmier, Sung Min Kim, Frank Krumeich,
Alexey Fedorov, Peter Chen, Christoph Mueller, and Christophe Copéret
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 26 Apr 2017
Downloaded from http://pubs.acs.org on April 26, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 12
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Molecularly-Tailored Nickel Precursor and Support Yield a Stable
Methane Dry Reforming Catalyst with Superior Metal Utilization
Tigran Margossian†, Kim Larmier†, Sung Min Kim‡, Frank Krumeich†, Alexey Fedorov†,‡, Peter
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chen†, Christoph R. Müller‡, Christophe Copéret†*
†
Department of Chemistry and Applied Biosciences, Vladimir-Prelog-Weg 1-5 ETH Zürich, CH-8093 Zürich, Swit-
zerland
‡
Department of Mechanical and Process Engineering, Leonhardstrasse 21, ETH Zürich, CH-8092 Zürich, Switzer-
land
KEYWORDS: Ni(II) formate tmeda, nickel(0) supported nanoparticles, deactivation mechanism, stability, methane
dry reforming
ABSTRACT: Syngas production via the Dry Reforming of Methane (DRM) is a highly endothermic process conducted
under harsh conditions, hence the main difficulty resides in generating stable catalysts. This can in principle be achieved
by reducing coke formation, sintering and loss of metal through diffusion in the support. [{Ni(μ2-
OCHO)(OCHO)(tmeda)}2(μ2-OH2)], readily synthesized and soluble in a broad range of solvents, was developed as a
molecular precursor to form pure 2 nm Ni(0) nanoparticles on alumina, the commonly used support in DRM. While such
small nanoparticles prevent coke deposition and increase the initial activity, operando XANES confirms that deactivation
largely occurs through the migration of Ni into the support. However, we show that Ni loss into the support can be miti-
gated through the Mg-doping of alumina, thereby increasing significantly the stability for the dry reforming of methane
via minimization of Ni-migration into the support. The superior performance of our catalytic system is a direct conse-
quence of the molecular design of the metal precursor and the support yielding in a maximization of the amount of acces-
sible metallic nickel in the form of small nanoparticles while preventing coke deposition.
rapid deactivation. One major issue is carbon deposition
ꢀ
INTRODUCTION
(whisker formation) via methane cracking (Eq 3.) or re-
verse Boudouard reaction (Eq 4.), which is particularly
severe for large nanoparticles.5,6 Deactivation by coke
deposition can be reduced in part by metal doping of the
active phase.7-10 An alternative approach is to reduce the
size of the nickel nanoparticles down to 2 nm, which also
partly alleviates whisker formation or the deposition of
carbon while increasing the activity because of a better
metal dispersion.5,11
Syngas is a key feedstock for the production of hydro-
gen and hydrocarbons, making reforming an essential
process of the petrochemical industry.1,2 Steam reforming
of methane (SRM) leads to a syngas with a high content
of hydrogen (H2/CO ratio of 3, Eq. 1). The dry reforming
of methane (DRM), in which steam is replaced by carbon
dioxide, leads to a H2/CO ratio of 1 (Eq. 2). This latter
process converts CO2 and CH4 into two valuable mole-
cules and provides a way to tune the H2/CO ratio of syn-
gas depending on the final application (Fischer-Tropsch
or methanol synthesis for instance).3,4
Methane Cracking:
CH4 = C + 2 H2, ΔH0298K = + 75 kJ.mol–1
(3)
(4)
Steam Reforming of Methane:
CH4 + H2O = CO + 3 H2, ΔH0298K = + 207 kJ.mol–1 (1)
Reverse Boudouard Reaction:
2 CO = CO2 + C, ΔH0298K = - 172 kJ.mol–1
Dry Reforming of Methane:
CH4 + CO2 = 2 CO + 2 H2, ΔH0298K = + 247 kJ.mol–1 (2)
However, obtaining small nickel nanoparticles on a
suitable support for dry reforming is still challenging.12,13
In particular, while the strong interaction of Ni(0) and
alumina support increases the stability of smaller particles
towards sintering, it also leads to the formation of inac-
tive nickel aluminate phase during synthesis or under dry
reforming conditions via migration of Ni from the particle
Since the dry reforming of methane is highly endo-
thermic, catalyst stability is important, in addition to low
cost and high activity. Alumina-supported nickel nano-
particles are suitable candidates because of their low cost
and high initial activity, but they commonly suffer from
ACS Paragon Plus Environment

Journal of the American Chemical Society
to the support.14 Tuning the support, for instance by in-
Page 2 of 12
Materials. Magnesium doped alumina (Mg@Al2O3) was
1
2
3
4
5
6
7
8
serting magnesium into the alumina lattice, can prevent
the formation of mixed nickel and alumina oxide.8,15-18
Thus, one challenge in designing better DRM catalysts is
the generation of small Ni(0) nanoparticles without losing
nickel into the support. The most common preparation
methods to Ni DRM catalysts are based on the impregna-
tion of alumina with nickel salts in aqueous media: condi-
tions under which Ni2+ ions react with a hydrated alumina
surface to form, after calcination, a poorly reducible hy-
drotalcite phase.19,20 The detrimental formation of mixed
nickel-aluminum oxide can be partially mitigated by uti-
lizing specific Ni precursors, such as ethylenediamine Ni
complex on γ-alumina.19-21 However, this method is not
general. Surface organometallic chemistry (SOMC) has
emerged as a powerful approach to generate single site
catalysts. SOMC can also be used to prepare supported
nanoparticles through grafting tailored molecular precur-
sors in a dry organic solvent followed by a subsequent
treatment under H2, yielding small supported particles for
a broad range of metals and supports.22-28 SOMC could
thus be ideal to generate small Ni nanoparticles from
tailored molecular precursors and supports for the prepa-
ration of dry reforming catalysts with minimal loss of Ni
and improved catalytic performances.
prepared via an incipient wetness impregnation (IWI)
method using Mg(NO3)2.6H2O ([Mg2+] = 5.5 M). After
calcination at 650 °C (5°C.min–1) for 5 h, Mg@Al2O3 had a
BET surface area of 199 m2.g-1, corresponding to a cover-
age 20 Mg.nm–2.29,30 After exposure to air, all supports
were dehydroxylated at 500 °C (5 °C.min–1 ramp) for 12 h
in a flow of synthetic air (100 mL.min–1), degassed under
high vacuum (10–5 mbar) for 1 hour and then stored in a
solvent-free glovebox. These supports referred to as
Mg@Al2O3-500, MgO500 and Al2O3-500 were also character-
ized by powder XRD (Figure S1). In case of Al2O3-500, the
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
surface OH density was reported previously to be 1.1
31
.
mmol OH/gsupport
[{Ni(μ2-OCHO)(OCHO)(tmeda)}2(μ2-OH2)] (1). NiCO3
(4.0 g, 34 mmol) was vigorously stirred with 100 mL of
deionized water. An excess of formic acid (10 mL, 277
mmol) was added to the mixture. After 1 h, a pale-green
solution was formed. The solvent was removed under
reduced pressure and the reaction mixture was dried (40
mbar), treated with 100 mL of absolute ethanol and dried
again under vacuum (10–2 mbar) for 16 h, yielding 6.1 g
(95%) of Ni(OCHO)2.2H2O. Ni(OCHO)2.2H2O (4.0 g, 22
mmol) was stirred in 150 mL of deionized water for 40
min at 40 °C. Tetramethylethylenediamine (3.8 mL, 25
mmol, 1.14 equiv) was quickly added to the resulting pale
green homogeneous solution. The solution became dark
blue and a precipitate formed. After 15 min, the reaction
mixture was dried under reduced pressure (40 mbar).
Ethanol (100 mL) was added to the pale-green solid, the
resulting reaction mixture stirred for 10 min and the sol-
vent evaporated to dryness under reduced pressure. The
residue was then extracted in 200 mL of toluene and fil-
Herein we describe the development of an air stable
molecular precursor, easily prepared on a gram scale and
soluble in a broad range of solvents (water, THF, toluene
and pentane), [{Ni(μ2-OCHO)(OCHO)(tmeda)}2(μ2-OH2)]
(1, tmeda = tetramethylethylenediamine), which can be
used to easily generate small, narrowly-dispersed (2.0 ±
1.0 nm) Ni(0) nanoparticles on various supports such as γ-
Al2O3 or Mg-doped alumina, providing highly active and
stable catalysts. Operando X-ray absorption spectroscopy
(XAS) demonstrates that this increase of catalytic perfor-
mance of Ni supported on Mg-doped alumina arises from
the formation of small Ni nanoparticles without loss of Ni
in the support, neither during the preparation nor under
dry-reforming conditions, thanks to the tailored supports
and the use of a practical molecular complex.
tered
through
Celite
to
remove
unreacted
Ni(OCHO)2.2H2O. Concentrating the filtrate and storing
o
at –38 C gave large turquoise crystals of 1 suitable for X-
ray crystallography. CCDC 1522535 contains the supple-
mentary crystallographic data. Two successive recrystalli-
zations from toluene yielded 5.1 g of 1 (82 %). Anal. Calcd
(%) for C16H38Ni2N4O9: C = 35.08 %, H = 6.99 %, N = 10.23
%. Found: C = 35.18 %, H = 7.03 %, N = 10.12 %. IR (KBr):
3020, 2986, 2889, 2843, 2817, 2791, 2780, 2733, 2696,
ꢀ
EXPERIMENTAL SECTION
General. Nickel nitrate (99.5%), aluminum nitrate
(>98.5%), magnesium nitrate (99%), nickel carbonate
(99.995%) and formic acid (≥ 96%) were purchased from
Sigma-Aldrich. Deionized water was purified using a
Purilab instrument (> 10 MΩ.cm). THF and benzene were
distilled from sodium under Ar (benzophenone used as
an indicator of dryness). Impregnation of aqueous solu-
tions was carried out in air. Impregnation or specific ad-
sorption using dry organic solvents were carried out using
a Schlenk line with Ar (grade 4.5) and 10–2 mbar vacuum.
H2 was purified over activated R3-11 BASF catalyst and
activated molecular sieves MS 4 Å prior use. γ-Alumina
was obtained from Sasol (Puralox SBA 200, SBET = 243
m2.g–1,Vpore = 0.6 mL.g-1). For water-based impregnation
routes, the supports were exposed to ambient air. All
other catalysts were prepared under air-free conditions
using standard Schlenk line techniques.
1
2087,1647, 1561, 1464, 1366 cm–1. H NMR (250 MHz, THF-
d8): δ = 27.0, 74.3, 79.9, 88.8, 104.0 ppm.
Catalyst synthesis. Impregnation (I) of Ni(NO3)2 on Al2O3
in water: NiNO3/Al2O3 (IH2O) γ–Al2O3 (2 g) was treated by
IWI with 1.2 mL of an aqueous solution of Ni(NO3)2 (0.426
M). The material was dried at 120 °C (1 °C.min–1) for 12 h in
a flow of synthetic air (80 mL min-1). The nominal nickel
loading is 1.0 wt%.
Impregnation of 1 on Al2O3 in water: 1/Al2O3 (IH2O) γ–Al2O3
(2 g) was treated by IWI with 1.2 mL of an aqueous solu-
tion of 1 (0.426 M) that was preheated to 50 °C for 40 min
to increase the solubility. The material was dried at 120 °C
(1 °C.min–1) for 12 h in a flow of synthetic air (80 mL.min–
1). The nominal nickel loading is 1.0 wt%.
Impregnation of 1 on Al2O3-500 in THF: 1/Al2O3 (ITHF) γ–
Al2O3-500 (2 g) was treated by IWI with 1.2 mL of solution
of 1 (0.426 M) in THF. The material was dried under high
ACS Paragon Plus Environment

Page 3 of 12
Journal of the American Chemical Society
vacuum (10–5 mbar) at room temperature for 16 h. The
were recorded on a Philips CM120 microscope operated at
1
2
3
4
5
6
7
8
nominal nickel loading is 1.0 wt%.
120 kV or on a Jeol 2100F microscope operated at 200 kV.
Particle size distribution (PSD) was determined by ana-
lyzing 100 nanoparticles and fitting a normal distribution
(particle size = mean value ± standard deviation). Scan-
ning transmission electron microscopy images were rec-
orded on an aberration-corrected Hitachi HD2700CS
microscope with a high-angle annular dark field detector
(HAADF STEM). X-ray spectra were measured with an
energy-dispersive detector (EDX) attached to the
HD2700CS. Particle size distributions and EDX spectra
are shown in the supplementary information (Figure S3-
8). The Ni nanoparticles appear bright in the HAADF-
STEM images as Ni has the highest atomic number (Z
contrast) among the elements present, which was also
confirmed by EDX spectroscopy.
Specific adsorption of 1 on Al2O3-500 from THF: 1/Al2O3
(ATHF) Complex 1 (65 mg, 0.12 mmol, 0.11 equiv) was dis-
solved in THF (20 mL) and contacted with 1.0 g of
Al2O3-500 (1.1 mmol AlOH, 1 equiv). The reaction mixture
was gently stirred for 2 h at room temperature. The solid
was washed 3 times with 20 mL of THF and dried under
vacuum (10–5 mbar) for 16 h to yield a light green solid. IR:
3669, 3548, 2980, 2881, 2809, 2791,1662, 1610, 1470, 1394,
1368 cm–1. Elemental analysis: Ni = 1.40 %, C = 2.01 %, H =
0.71 %, N = 0.51 %, which corresponds to the molar ratio
of N/Ni = 1.4 and C/Ni = 7.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Specific adsorption of 1 on Mg@Al2O3-500 from THF: 1/
Mg@Al2O3 (ATHF) The material was prepared as described
above but using Mg@Al2O3-500 as a support instead of
Al2O3-500 and 74 mg of 1 (0.14 mmol). A light green solid
was obtained. Nominal loading: Ni = 1.6 wt%. IR: 3539,
3018, 2979, 2878, 2850, 2806, 2730, 1609, 1469, 1371 cm–1.
o
N2 adsorption. Nitrogen adsorption isotherms at –196 C
were recorded on a Bel-Mini apparatus (Bel-Japan). Prior
to the measurement, the samples were outgassed under
vacuum (ca. 10–3 mbar) at 350 oC for 2 h. The BET method
was applied to calculate the total surface area.34
Reduction of the impregnated materials to give supported
nanoparticles. The dried samples were reduced under a
flow of pure hydrogen at 650 °C (1 °C.min–1) for 8 h. After
outgassing under high vacuum (10–5 mbar), the catalysts
were stored in an Ar atmosphere in a solvent-free glove-
box. The code 2 in the numbering below, refers to the
reduced supported nanoparticles prepared from complex
1. Reference catalyst, Niref/Al2O3. The material was pre-
pared by IWI using 1.2 ml of aqueous solution of
Ni(NO3)2.6H2O (1.58 M). The as-synthesized material with
a nominal 10 wt.% Ni loading was dried at 110 °C overnight
(ca. 12 h under static air) followed by calcination at 800 °C
(5 °C.min–1) for 2 h. The catalyst was reduced at 850 °C in
the fixed bed reactor for 1 h prior to the DRM test.32,33
H2 chemisorption. Chemisorption experiments were car-
ried out on a BELSORB-max apparatus (Bel-Japan). Ap-
proximately 200 mg of reduced catalyst was loaded in an
airtight cell inside a glovebox which was then mounted
on the apparatus. The sample was outgassed at 350 °C for
3 h at 10–6 mbar. The chemisorption measurements were
performed at 25 °C, the pressures at equilibrium were
recorded when the pressure variation was below 0.01 %
per minute. The uptake of hydrogen was calculated by
fitting the isotherm with the Langmuir isotherm model.
Temperature programmed reduction (TPR) studies. TPR
experiments were carried out using a BELCAT-B appa-
ratus (Bel Japan) equipped with a homemade cell allowing
the transfer of the sample from a glovebox to the instru-
ment without exposure to air. The dried sample (250 mg)
was introduced into the cell inside a glovebox. Prior to
opening the cell mounted in the apparatus, the inlet and
outlet were purged with argon for 30 min using a by-pass
system. The measurement was performed from 25 °C to
900 °C in the gas flow consisting of 6% H2/94% He using
a ramp of 1 °C.min–1. The gas released was analyzed using
a calibrated TCD detector and a mass-spectrometer (Bel-
Mass).
Catalytic dry reforming of methane. The DRM reaction
was carried out in a fixed-bed quartz reactor (400 mm
length, 12.6 mm internal diameter). A typical experiment
was performed with 50 mg of catalyst. The catalysts were
o
first reduced in situ in 10 vol% H2/N2 at 650 C for 1 h.
Subsequently, the bed was purged with N2 (100 ml min–1)
for 10 min and then the feed gas was introduced. The total
flow rate of the feed gas was 100 sccm (GHSV of 108000
ml.gcat-1.h-1; composition: 45 % CH4, 45 % CO2 and 10 %
N2). The composition of the off-gas at the outlet was ana-
lyzed with a micro-GC (Thermo Scientific) identifying the
following products: CH4, CO2, H2, CO and N2 (thermal
conductivity detector (TCD); molecular sieve 5A and U-
plot column cartridges).
CO adsorption monitored by FTIR spectroscopy. The ex-
periment was conducted in a 200 mL glass reactor
equipped with IR-transparent CaF2 windows. A thin pellet
of the sample was pressed in a glovebox and loaded to the
reactor using a sample holder. A spectrum was recorded
in an argon atmosphere. The reactor was outgassed (10–5
mbar) for 5 min at room temperature. Subsequently, 100
mbar of CO were introduced to the reactor and a second
spectrum was recorded. The reactor was outgassed and a
third spectrum was recorded under vacuum (10–5 mbar).
Spectra were recorded in a transmission mode on a Ni-
colet 6700 FTIR spectrophotometer-using 64 scans at a
resolution of 2 cm–1.
Physicochemical characterization. Elemental analysis.
The metal content in each catalyst was determined by
inductively coupled plasma optical emission spectroscopy
(ICP-OES). The ICP-OES measurement were carried out
by Pascher laboratories.
Scanning Transmission Electron Microscopy (TEM/STEM).
Prior to TEM analysis, all samples were oxidized by the
slow diffusion of air to the catalysts under argon. After
oxidation, materials were dispersed in ethanol and a
droplet of the suspension was deposited on a lacey carbon
foil supported on a copper grid. Dark field TEM images
1
In-situ deposition of complex 1 followed by H-NMR analy-
sis. The spectra were recorded on a 250 MHz Bruker spec-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 12
trometer. Al2O3-500 (85 mg) was introduced to a J Young Powder X-ray diffraction. The powder XRD measurements
1
2
3
4
5
6
7
8
tube. A solution containing complex 1 (8 mg, 15 μmol) and
FeCp2 (6 mg, 33 μmol, internal standard) in 0.4mL of
THF-d8 was prepared in a J Young tube inside a glovebox.
Spectra were recorded using the following settings: 256
scans, d1 = 0.05 sec, band width = 300 ppm for paramag-
netic samples and 16 scans, d1 = 20 sec, band width = 25
ppm for diamagnetic samples.
were performed using a Bruker-AXS “D8 Advance” dif-
fractometer using Cu Kα monochromatic radiation (λ =
1.5418 Å). The XRD patterns were recorded between 10-
80° (2θ with a step size of 0.033°).
Single crystal XRD. The data were collected on a Bruker
D8 Venture diffractometer equipped with a CCD area
detector using Mo Kα radiation. The crystal was placed in
Paratone and mounted in the beam under a flow of nitro-
gen at 100 K. An empirical absorption correction was
performed with SADABS-2008/1 (Bruker). The structure
was solved with SHELXL35 using intrinsic phasing fol-
lowed by a least-squares refinement (SHELXL-97) using
the OLEX 2−1.2 suite of programs.36 The non-hydrogen
atoms were refined anisotropically. The hydrogen atoms
were placed at the calculated positions (see the Support-
ing Information file for details).
Ultraviolet-visible (UV-Vis) spectroscopy. UV-Vis spectra
were recorded with a resolution of 1 nm in transmission
mode on an Agilent Technologies, Cary Series spectrome-
ter using the corresponding solvent as a reference. Diffuse
reflectance UV-Vis spectra of the solids were recorded
with a resolution of 1 nm on the same spectrometer, using
KBr as a reference. A sample was loaded to an air-tight
cell inside a glovebox. To convert the reflectance data to a
signal proportional to the absorption, the diffuse reflec-
tance spectra were processed using Kubelka-Munk theo-
ry.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ
RESULTS AND DISCUSSIONS
Synthesis of [{Ni(μ2-OCHO)(OCHO)(tmeda)}2(μ2-
Operando dry reforming quick X-ray absorption spectros-
copy. qXAS at the Ni K-edge was measured at the
SuperXAS beamline at the Swiss Light Source (SLS; Paul
Scherrer Institute, Villigen, Switzerland). The SLS is a
third-generation synchrotron operating at a 2.4-GeV elec-
tron energy and a current of 400 mA. The incident beam
was collimated by a Si-coated mirror at 2.8 mrad, mono-
chromatized using a double crystal Si(111) monochroma-
tor, and focused with an Rh-coated toroidal mirror (at 2.8
mrad) down to 100 × 100 μm with a beam intensity of 4-5
x 1011 ph/s. XAS spectra were collected in the transmission
mode using ion chambers filled with He-N2 gas mixtures.
Multiple X-ray adsorption scans (8000 - 9200 eV) were
averaged with a scan time of 5 min. The DRM reaction
was carried out in a fixed-bed capillary quartz reactor (40
mm length, 3 mm internal diameter). A typical experi-
ment was performed with 10 mg of catalyst. The catalysts
were first reduced in situ in 5 vol% H2/N2 at 650 oC for 1 h.
Subsequently, the feed gas was introduced with a total
flow rate of 28 sccm (GHSV of 151200 ml.gcat–1.h–1; compo-
sition: 45 % CH4, 45 % CO2 and 10 % N2). Water was re-
moved from the outlet flow using a membrane separator
(Genie 170 technologies). The composition of the off-gas
at the outlet was analyzed with a micro-GC (Agilent
Technologies) identifying the following products: CH4,
CO2, H2, CO and N2 (thermal conductivity detector
(TCD); molecular sieve 5A and U-plot column cartridges).
Scans were recorded every half hour during the reduction
and the catalytic reaction. Reference NiAl2O4 and NiO
copper oxides were pressed into wafers and sealed in
Kapton tape. All spectra were calibrated with respect to
the spectrum of a Ni reference foil recorded simultane-
ously by setting the first inflexion point to 8333 eV (Ni K-
edge). XANES spectra (8320 – 8420 eV) of the supported
nanoparticles were fitted using a linear combination fit-
ting of Ni foil and NiAl2O4 references. Extended X-ray
absorption fine structure (EXAFS) data were fitted in the
R-space (1-4 Å) after a Fourier transform (k = 1 - 11 Å–1)
using a k weight of 1, 2 and 3 for complex 1 and using a k
weight of 1, 2 and 3 for the specifically adsorbed species.
OH2)], 1.
Complex 1 was synthesized in 82% yield by reacting ca.
1.1 equiv. of tetramethylethylenediamine (tmeda) in water
with nickel formate, prepared from NiCO3 (see experi-
mental section for details). The synthesis can be carried
out on a multigram scale to give pure crystalline complex
1 that is very soluble in a broad range of solvents: 15 g.L–1
o
in water, 115 g.L–1 in toluene, and 170 g.L–1 in THF at 25 C.
Complex 1 was characterized by X-ray crystallography
(Figure 1). It crystalizes as a dimer with octahedral Ni
centers connected by three μ2 bridges (two formate ani-
ons and a water molecule). Each nickel atom is inde-
pendently connected to another κ1 formate ligand and
one κ2 tmeda.
Figure 1. Crystal structure of complex 1. Thermal ellipsoids
were set at 50% probability level and except for μ-H2O, hy-
drogen atoms and the co-crystallized water molecule are
omitted for clarity. Selected bond lengths (Å) and angles
(deg): Ni1–O1 = 2.049(2), Ni1–O2 = 2.099(1), Ni1–O3 =
2.097(2), Ni1–O4 = 2.022(2), Ni1–N1 = 2.168(2), Ni1–N2 =
2.148(2), Ni2–O7 = 2.030(2), Ni2–N4 = 2.178(2), Ni1–O2–Ni2 =
117.15(8), O1–Ni1–O2 = 89.29(8).
Preparation and characterization of supported nick-
el nanoparticles. Using a conventional and prototypical
preparation route, i.e. incipient wetness impregnation (IWI) of
an aqueous solution of nickel nitrate followed by reduction at
650 °C in a hydrogen flow, yields nanoparticles with an averꢀ
age diameter of 4 nm and a comparatively broad size distribuꢀ
tion of ± 2.5 nm (Figure 2a). In contrast, the size of the nanoꢀ
particles is smaller (3.4 ± 1.5 nm) when using precursor 1
(IWI of aqueous solution, Figure 2b). Changing from water to
THF during the deposition step via IWI allows for the forꢀ
ACS Paragon Plus Environment

Page 5 of 12
Journal of the American Chemical Society
mation of even smaller nanoparticles with a narrower distribuꢀ
tion (2.9 ± 1.2 nm, Figure 2c). Hence, avoiding water clearly
1
improves the particle size distribution of the nickel particles
obtained after reduction. Finally, we performed the specific
adsorption of complex 1 in a THF solution on γꢀalumina. The
nanoparticle size distribution obtained with this method was
further improved to 2.0 ± 1.0 nm according to the darkꢀfield
STEM images (Figure 2d). Thus, tuning the precursor, the
conditions and the deposition methods allows for controlling
the particle size and distribution. The XRD patterns of the
materials do not show any specific signature due to the
small Ni nanoparticle size, consistent with the STEM
results (Figure S1). Transmission IR spectra of the reduced
nanoparticles (Figure S2) contain no C–H frequencies
indicating that the Ni surface is free from remaining or-
ganic ligands. The metal dispersion was further deter-
mined by hydrogen chemisorption at 25 °C (Table 1 and
isotherm in Figure S3).
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
20
15
10
5
20
15
10
5
(a)
(b)
0
0
0
2
4
Particlesizeꢀ nm
6
8
10
0
2
4
6
Particlesizeꢀ nm
8
10
(a) dp = 4.0 ± 2.5 nm
(b) dp = 3.4 ± 1.5 nm
5 nm
5 nm
(c)
25
20
15
10
5
(d)
30
20
10
0
0
0
2
4
6
Particlesizeꢀ nm
8
10
0
2
4
Particlesizeꢀ nm
6
8
10
(c) dp = 2.9 ± 1.2 nm
(d) dp = 2.0 ± 1.0 nm
5 nm
5 nm
Figure 2. Particle size distribution as determined by
HAADF-STEM for alumina-supported catalysts. (a)
NiNO3/Al2O3 (IH2O), (b) 2/Al2O3 (IH2O), (c) 2/Al2O3 (ITHF), (d)
2/Al2O3 (ATHF).
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 12
Table 1. Characterization of the as prepared supported nanoparticle catalysts. a: Dehydroxylated at 500 oC; b: Nickel load-
ing determined by elemental analysis; c: measured by HAADF/STEM; d: H2 chemisorption uptake at 25oC using a Langmuir
model; e: dispersion = surface Ni/total Ni assuming that one hydrogen atom adsorbs on each surface Ni (details in ESI). The
amount of total Ni was determined by elemental analysis. “S. Ads.” stands for specific adsorption
1
2
3
4
5
6
Nickelb
(%wt)
Particle sizec
(nm)
H2 uptaked
(mmolH2.gNi-1)
De
Entry
Impregnated material
Reduced material
Support
Method
(%)
7
8
9
1
Ni(NO3)2/Al2O3 (ref)
Niref/Al2O3
γ-Al2O3
γ-Al2O3
IWI
IWI
10
15.0 ± 7.0
4.0 ± 2.5
0.8±0.1
1.3±0.1
9
2
Ni(NO3)2/Al2O3 (IH2O
1/Al2O3 (IH2O
1/Al2O3 (ITHF
1/Al2O3 (ATHF
1/Mg@Al2O3 (ATHF
)
NiNO3/Al2O3 (IH2O
2/Al2O3 (IH2O
2/Al2O3 (ITHF
2/Al2O3 (ATHF
)
0.92
15
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
)
)
γ-Al2O3
IWI
IWI
1.10
3.5 ± 1.5
2.9 ± 1.2
2.0±0.1
2.2±0.2
24
26
4
)
)
γ-Al2O3a
1.05
5
)
)
γ-Al2O3a
S.Ads.
S.Ads.
1.40
1.60
2.0 ± 1.0
1.3 ± 0.7
2.8±0.2
3.0±0.2
33
36
2/Mg@Al2O3
(ATHF
6
)
Mg@Al2O3a
)
We observe that a decrease in the particle size of Ni, as
determined by microscopy, correlates with an increase in
the hydrogen uptake from 1.3 ± 0.1 mmolH2.gNi–1 for
NiNO3/Al2O3 (IH2O) to 2.8 ± 0.2 mmolH2.gNi–1 for 2/Al2O3
(ATHF). Accordingly, the metal dispersion increases as the
metal particle size decrease from D = 15% for NiNO3/Al2O3
(IH2O) to D=33% for 2/Al2O3 (ATHF) (Table 1).
Assuming that the contribution of the bands that are due
to the binding of CO on alumina is constant for different
catalysts, the fraction of Ni(0) bands increases with de-
creasing particle size (Table S2). No characteristic blue
shifted band due to the Ni(II)-CO stretching vibration
around 2173 cm–1 is observed for NiNO3/Al2O3 (IH2O) as
might be anticipated from the incomplete reduction of
this material, suggesting migration of Ni into the bulk of
the support and in line with its low reducibility.
The fraction of Ni(0) in the reduced samples 2/Al2O3
(ATHF) and NiNO3/Al2O3 (IH2O) was further assessed by in
situ XANES spectroscopy at 650 C under H2 flow (Figure
o
3). The XANES spectrum of NiNO3/Al2O3 (IH2O) contains a
white line with a normalized adsorption value that is 20%
higher than that of the Ni foil reference, indicative of the
presence of partially oxidized nickel species. This is con-
sistent with the high temperature peak (960 °C) in the
TPR spectrum of NiNO3/Al2O3 (IH2O) due to the formation
of the mixed metal aluminate, which are difficult to re-
duce even under harsh conditions (Figure S8).
Linear combination fitting of the XANES data for
NiNO3/Al2O3 (IH2O) using the Ni foil and NiAl2O4 refer-
ences (Figure S9) shows that only 60% of total nickel was
reduced at 650 °C under an H2 flow. In contrast, the
XANES spectrum of 2/Al2O3 (ATHF) (Figure 3, b) is very
similar to that of the Ni foil reference, displaying identical
pre-edge (8333 eV) and edge (8345 eV) positions.37 Linear
combination fitting of the XANES spectrum of 2/Al2O3
(ATHF) (Figure S9) indicates that > 97% of nickel is re-
duced to Ni(0) in this sample. Next, the reduced Ni sur-
face was probed by CO adsorption.38 The IR spectra plot-
ted in Figure S13 were obtained by subtracting the spec-
trum of the as-reduced catalyst (a, b) or a respective ref-
erence materials (c, d) from the spectrum recorded in a
CO atmosphere. The alumina support and NiAl2O4 refer-
ence material bind CO (frequency at ν = 2209 cm–1 for
Figure 3. In situ XANES spectrum at the Ni K-edge under
hydrogen flow (650 oC, 1 bar); (a) NiNO3/Al2O3 (IH2O); (b)
2/Al2O3 (ATHF); (c) Ni foil reference; (d) NiAl2O4 reference.
Thus, the analyses presented above demonstrate that
deposition of the precursor using organic solvents leads
to nanoparticles with a decreased mean size and narrow
particle size distribution down to 2.0 ± 1.0 nm (2/Al2O3
(ATHF)). The specific adsorption route also mitigates the
migration of nickel into the alumina matrix and the for-
mation of non-reducible NiAl2O4 species, affording > 95 %
of reduced Ni(0) (Figure 4). In general, the method dis-
cussed here gives smaller and more narrowly dispersed
nanoparticles than those reported previously from a
strongly adsorbed nickel ethylenediamine complex (2-7
nm).21,41
Al2O3
and at ν = 2173 cm–1 for NiAl2O4). The former
700
band is also present in NiNO3/Al2O3 (IH2O) and 2/Al2O3
(ATHF). Multiple bands in the range 1740 cm–1 to 2120 cm–1
are assigned to different coordination sites and vibration
modes of CO on metallic Ni nanoparticles (see ESI).39,40
ACS Paragon Plus Environment

Page 7 of 12
Journal of the American Chemical Society
(a)
(b)
1
2
3
4
5
6
7
Ni
Ni
Ni
"
Ni"
NiAl2O4
γꢀAl2O3
γꢀAl2O3"
Dispersion= 15 %
Dispersion= 33 %
Figure 4. Schematic drawing of samples A) NiNO3/Al2O3
(IH2O) and B) 2/Al2O3 (ATHF
8
9
)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Investigation of the specific adsorption of Ni com-
plex 1. Since the preparation route has a large influence
on the size of the particles, we investigated the interac-
tion of 1 with the surface of γ-alumina. UV-Vis spectros-
copy was used to assess the coordination of the Ni2+ cen-
ter of 1 on the alumina surface (Figure 5). The UV-Vis
spectrum of 1/Al2O3 (IH2O) is very similar to that of the
molecular complex 1 showing bands at 388 nm and 665
nm, that can be assigned to the d-d transition of the octa-
hedral Ni2+(d8) (Figure 5).42 In contrast, the specific ad-
sorption of complex 1 onto alumina from dry THF, (b) on
Figure 5, produces material with bands at 404 and 703
nm. While Ni2+ sites remain octahedral, the bands are red
shifted compared to the bands in 1. This can be explained
by a change in the nickel environment for a weaker field
ligand such as Al-OH.21 Fits of the EXAFS spectra of
1/Al2O3 (ATHF) presented in Figure S19 confirm the pres-
ence of the octahedral Ni2+ center. A nickel-nickel path
allows fitting the feature centered at 3.10 Å thereby
demonstrating that 1 retains its dimeric structure at the
surface of alumina (Figure 6). Complete structure investi-
gation of 1/Al2O3 (ATHF) is described in the supplementary
information section (Figure S14-20).
Figure 6. Proposed structure of a surface coordinated com-
plex of 1 on Al2O3-500
.
Overall, the spectroscopic data is consistent with a ligand
exchange reaction that replaces tmeda in 1 by two Al-OH
and/or Al-O-Al weak field ligands and retaining the di-
meric structure and the octahedral geometry of Ni cen-
ters. On the other hand, impregnation procedures lead
mostly to the precipitation of 1 on the support.
Dry reforming of methane. The performances of the
catalysts prepared under DRM conditions were examined
in a fixed-bed flow reactor. The tests were performed at
650 °C under kinetic regime, far from the equilibrium
conversion (thermodynamic XCH4 = 57%) with methane
conversion below 40% and that is far from a diffusion
regime. Figure 7A shows the rate of methane consump-
tion (in molCH4.moleNi-1.s–1 (rCH4)) after 0.5 h and 20 h on
stream. With a dispersion of 10 % and 15 %, NiRef/Al2O3
and NiNO3/Al2O3 (IH2O) have an initial activity of 0.68 and
0.94 molCH4.moleNi-1.s–1. Further increasing the dispersion
enhances the initial activity to 1.45 molCH4.moleNi-1.s–1 for
2/Al2O3 (IH2O) (D = 24 %) and 1.40 molCH4.moleNi-1.s–1 for
2/Al2O3 (ITHF) (D = 25 %). The smallest supported nickel
nanoparticles have the greatest initial activity of 1.79
molCH4.moleNi-1.s–1 for 2/Al2O3 (ATHF) (D = 33 %). Thus, we
observe that the initial rate of consumption of methane
increases linearly with the nickel dispersion (Figure 8).
This clearly points to Ni(0) being the active phase in DRM
in line with our recent study.10 This data contradicts re-
cent proposal describing four coordinated Ni2+ as active
sites in DRM.43 From the slope of the plot rCH4 = f(D), the
initial turnover frequency for the consumption of me-
thane per nickel surface atom is estimated to 5.5 s-1 (at
650 °C).
Thus, increasing the dispersion clearly increases the (ini-
tial) rate of DRM per gNi. In this respect the catalyst with
the higher dispersion prepared by the specific adsorption
route from THF is the most efficient – 2/Al2O3 (ATHF).
However, this sample undergoes a rapid deactivation after
several hours on stream and displays very poor perfor-
mance after 20 h on stream with an activity loss of 95 %
(0.08 molCH4.moleNi-1.s–1). Similarly, NiRef/Al2O3 and
NiNO3/Al2O3 (IH2O) undergo a significant deactivation with
a performance loss of 75%, yielding an activity of 0.25
molCH4.moleNil-1.s–1 after 20 h. Nickel catalysts with an
intermediate dispersion – 2/Al2O3 (IH2O) and 2/Al2O3
(ITHF) – show an improved performance after 20 h with an
activity of approximately 0.70 molCH4.moleNi-1.s–1. We
Figure 5. Background subtracted UV-Vis spectra of materials
1/Al2O3 (IH2O) (a), 1/Al2O3 (ATHF) (b) and 1 in THF solution
(c).
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 12
therefore further investigated the possible deactivation with a white line intensity increasing of by 10% attributed
1
2
3
4
5
6
7
8
mechanisms by examining the formation of coke and the
oxidation state of the metal under reaction condition.
to the formation of NiAl2O4. The loss in activity clearly
parallels the decrease in Ni(0) content with time on
stream as it is converted into NiAl2O4. This data is also
consistent with Ni(0) being the active species in DRM.
Investigation of deactivation mechanisms. Imaging of
the spent catalyst allows a better understanding of the
different deactivation pathways. Carbon whiskers and
sintered metal nanoparticles (>10 nm) are only observed
for the spent NiNO3/Al2O3 (IH2O) and Niref/Al2O3 while
2/Al2O3 (ATHF) still have supported particles with 2 nm
size (Figure S22-24) with no obvious carbon contamina-
tion. Quantification of the carbon content by TGA indi-
cated that 3.10-4 mole of coke formed per mole of carbon
converted after 20 h on stream for the high dispersion
samples (D > 26 %) and a 7.10–4 % for NiNO3/Al2O3 (IH2O).
Thus, larger nickel nanoparticles have a higher tendency
to form coke than smaller nanoparticles, in line with
previous studies conducted on silica.5
The main deactivation mechanism for this sample is thus
the oxidation/migration of Ni(0) into the alumina lattice.
Similar experiment was performed on NiNO3/Al2O3 (IH2O),
which shows that in this case NiAl2O4 is initially present
after reduction and the Ni(0) content does not signifi-
cantly diminish over the course of the reaction (Figure
S26-27) while the activity drops. Thus, in this case, deac-
tivation is mostly linked to the coke formation.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Overall, two distinct deactivation mechanisms come into
play in this set of samples. Larger nickel particles deacti-
vate mostly by coke formation (NiNO3/Al2O3 (IH2O)).
Smaller nickel nanoparticles (2/Al2O3 (ATHF)) while allow
for an increased initial reaction rate and lower coking,
deactivate due to the strong interactions between the
metal and the support, leading to the migration of Ni into
the support to generate nickel aluminate. Indeed, it was
possible to regenerate after 10h on stream the activity of the
samples with a large particle size by an oxidative/reductive
treatment. This is expected as the main deactivation mechaꢀ
nism is coke formation. However, this regeneration treatment
is not successful for the highly dispersed sample (2/Al2O3
(ATHF)); it decreases further the activity, possibly due to the
increased formation of nickelꢀaluminate like species (Figure
S22).
Decreasing the nickel nanoparticle size allows reducing
coke formation while improving activity. Thus, the strong
deactivation of 2/Al2O3 (ATHF) cannot be attributed to the
formation of carbon, since the samples 2/Al2O3 (IH2O) and
2/Al2O3 (ITHF), although forming more coke according to
TGA analysis, deactivate much less over 20 h (by about 50
% each). We then examined the evolution of the oxida-
tion state of nickel by operando XAS under DRM condi-
tions. Ni(0) content was determined by linear combina-
tion fitting of the XANES Ni K-edge spectra (using Ni foil
and NiAl2O4 references, Figure S25) and plotted against
TOS (Figure 7B2). After reduction (t = 0 h), 2/Al2O3
(ATHF) have a Ni(0) content above 95%.
In an early stage of the reaction (t = 0.17 h) the XANES
spectrum of 2/Al2O3 (ATHF) shows already some change
ACS Paragon Plus Environment

Page 9 of 12
Journal of the American Chemical Society
ꢀ3
A)
A)
Rate0.5 h
Rate20 h
1
2
3
Coke
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NNii
Ni
Ni
Ni
Ni
Ni
γꢀAl2O
Ni
Ni
"
Ni"
Mg@Al2O
3"
γꢀAl2O
3"
NiAl2O
γꢀAl2O
4
γꢀAl2O
γꢀAl2O
3
γꢀAl2O
3
3
3
3
NiNO3/Al O (IH2O
)
2/Al O (IH2O
2 3
)
2/Al O (A )
2 3 THF
ref
Ni /Al O
2/Al O (I )
2 3 THF
2/Mg@Al O
2
Sample
3
2
3
2
3
24 % 26 %
15 %
33 %
Dispersion 10 %
36 %
B)
3)
2)
1)
/Al2O (A ) Operando
3 THF
/Mg@Al2O (A ) Operando
3 THF
NiAl2O reference
4
NiAl2O reference
4
time
time
time
time
o
Figure 7. A) Dry reforming performance of the catalysts prepared at 650 C using 100 mL.min-1 of CO2/CH4/N2 (0.45/0.45/0.1), 1
Ni μmol (25-50 mg) of catalyst, GHSV=98’400 mL.h-1.gCat-1. Rate of methane consumption in molCH4.moleNi-1.s–1 for 0.5 h and 20 h
TOS (left axis) and coke formed.(c
arbon converted-1) in % for the different catalysts (right axis, blue column). B) Operando
XANES dry reforming experiment 1)XANES spectra recorded during the operando experiment for 2/Al2O3 (ATHF) 2) Performance
of the catalysts prepared at 650 oC using 23 mL.min-1 of CO2/CH4/N2 (0.45/0.45/0.1), 0.25 Ni μmol (5-10 mg) of catalyst,
GHSV=98400 mL.h-1.gCat-1. Rate of methane consumption in molCH4.moleNi-1.s–1 for the line/markers (left axis) and Ni(0) content
(%) for the line and dotted line (right axis) 3) XANES spectra recorded during the operando experiment for 2/Mg@Al2O3 (ATHF
)
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 12
Tailoring the support and the catalyst performance. cles by migration of Ni into the support matrix, possibly
1
2
3
4
5
6
7
8
In order to prevent the migration of the metal into the
support deactivating the ‘smaller’ nanoparticles, we intro-
duced Mg2+ ions in the lattice of alumina by impregnating
magnesium nitrate solution on γ-alumina followed by
calcination.29,30 After calcination, the diffraction pattern
resembles that of the starting γ-Al2O3, although the peaks
are shifted by 1o towards lower angles with main peaks at
36.5 , 45.4 , 66.2 (Figure S1). This data indicates an
expansion of the lattice, likely due to the insertion of Mg2+
ions into the structure.44
due to a decreased quantity of defects, thereby helping
maintaining high activity over 20 h of time on stream.
2/Mg@Al2O3 (ATHF) still undergoes modest deactivation
that is not related to carbon formation, since no signifi-
cant coke formation is observed (TGA and TEM) but is
attributed to a partial sintering of the particles during
reaction leading to a particle size of 2.7 ± 2.5 nm after 20 h
on stream (Figure S34-35). This is consistent with the
behavior of small Ni particles,5 and is also supported by
the fact that regeneration attempts by oxidative treatment
had no effect on the activity (Figure S22).
o
o
o
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
No segregation of extended MgO domains could be found
neither by TEM mapping (Figure S28) nor by powder
XRD (Figure S1); hence this procedure likely generates a
homogeneous MgAl2O4-like layer on the alumina. The
deposition of 1 on Mg@Al2O3 pretreated at 500 °C under
synthetic airflow was carried out by the specific adsorp-
tion procedure, leading to a similar a surface coordinated
complex as on the alumina surface (Figure S29-31).
Finally, we assessed the activity of the best performing cataꢀ
lyst 2/Mg@Al2O3 (ATHF) at a higher reaction temperature
(900 °C, Figure S23). The conversion was kept below 40 % to
ensure operation in the kinetic regime (space velocity was
increased to GHSV = 1350000 mL h1 gcatꢀ1). The catalyst deacꢀ
tivates by about 50 % within the first five hours on stream, but
then retains a stable activity of 7.6 molCH4 molNiꢀ1 sꢀ1 over more
than 15 h. Hence, the synthesis method proposed here allows
for the preparation of an efficient catalyst that can operate
even under harsh reaction conditions.
ꢀ
CONCLUSION
We have shown how mechanistic insight into catalyst
deactivation allows designing a very active and stable
nickel catalyst on alumina-based supports for dry reform-
ing by increasing metal dispersion and preventing deacti-
vation pathways (coke formation and metal migration to
the support). This improvement was achieved by tailoring
the nickel precursor, the deposition method and the sup-
port. Nickel dispersions as high as 33 % could be obtained
using an easily accessible Ni(II) precursor, namely [{Ni(μ2-
OCHO)(OCHO)(tmeda)}2(μ2-OH2)]
(tmeda
=tetramethylethylenediamine), soluble in a broad range
of solvents, by depositing this precursor on the surface of
the support by a specific adsorption procedure in an or-
ganic solvent. After reduction, small and narrowly dis-
persed supported nanoparticles are obtained (2.0 ± 1.0
nm). We have shown that a higher degree of dispersion
increases the initial DR rate. Using a combination of op-
erando spectroscopy and adsorption techniques we
demonstrate that Ni(0) is the active phase for DRM. Small
Ni particles not only allow for an increased activity, but
they also strongly reduce coke formation. Nevertheless,
such small particles dispersed on γ-alumina deactivate
very quickly by migrating into the alumina support form-
ing NiAl2O4, according to operando XAS experiments.
Incorporating magnesium into the alumina lattice sup-
presses the latter deactivation mechanism occurring on
small size nanoparticles and enables high and stable ac-
tivity over 20 h on stream. As a result, a straightforward
route to prepare highly dispersed supported nickel nano-
particles for DRM by rational tailoring of the metal pre-
cursor and the support was demonstrated. The resulting
sample has an unprecedented metal utilization and main-
tains high specific activity over several hours on stream
thanks to a molecular understanding of the catalyst at
every step of the preparation and during the reaction
conditions by operando XANES spectroscopy.
Figure 8. Rate of consumption of methane per
molCH4.moleNi-1.s–1 after 0.5h TOS vs. the initial metal dis-
persion D (%).
After reduction of the Ni2+ surface species under a flow of
hydrogen, 1.3 ± 0.8 nm nickel nanoparticles were observed
on Mg@Al2O3 surfaces (Figure S32, metal loading of 1.6 %,
respectively) containing 94% of Ni(0) according to
XANES (Figure S33). The thus-obtained 2/Mg@Al2O3
(ATHF) displayed an initial activity of 1.92 molCH4.moleNi-1.s–
1
slightly higher than its pure aluminum counterpart,
2/Al2O3 (ATHF), in line with the slightly higher metal dis-
persion (see Figure 8). More importantly, the magnesium-
doped alumina supported catalysts retained a relatively
high activity after 20 h TOS with a rate of methane con-
sumption of 1.17 molCH4.moleNi-1.s–1 (about 40 % of deacti-
vation), in sharp contrast to what is observed on pure
alumina. Operando XANES analysis shows no significant
change in the oxidation state of this sample, with only a
slight decrease in the Ni(0) content from 94 % to about
90 %, which is substantially smaller than the observed
catalyst deactivation (Figure 7B3). Thus, modifying γ-
alumina by inserting Mg2+ ions in the structure signifi-
cantly reduces the deactivation of small nickel nanoparti-
ACS Paragon Plus Environment

Page 11 of 12
Journal of the American Chemical Society
ASSOCIATED CONTENT
1
2
3
4
AUTHOR INFORMATION
Supporting Information
Corresponding Author
The Supporting Information is available free of charge on the
ACS Publications website.
* C.C.: tel, +41 44 633 93 94; e-mail, ccoperet@ethz.ch
5
6
7
8
Funding Sources
Single XRD details data of complex 1, the powder X-
ray diffractograms, FTIR of the reduced sample,
HAADF-STEM-EDX results, H2 chemisorption iso-
therm, TPR data, XANES data, FTIR spectra for CO
adsorption experiment, H NMR spectra, selectivity
data for DRM tests.
This research was funded by ETH Zürich and the Swiss Na-
tional Funding (SNF) in relation with Swiss Competence
Centers for Energy Research (SCCER Heat and Electricity
Storage).
1
9
Notes
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The authors declare no competing financial interest.
ACKNOWLEDGMENT
The authors thank the SCCER Heat and Energy Storage and
ETH Zürich (ETH-57_12-2) for financial supports. We
acknowledge the Paul Scherrer Institut, Villigen, Switzerland
for provision of synchrotron radiation beam time at beamline
Super XAS (20160169) of the SLS and would like to thank Dr.
Olga V. Safonova and Dr. Maarten Nachtegaal for assistance.
We thank ScopeM (ETH Zürich) for providing STEM meas-
uring time. We also would like to thank P. Trüssel and R.
Mäder (ETH Zürich) for their contribution building the op-
erando reactor set-up.
REFERENCES
(1) Navarro, R. M.; Peña, M. A.; Fierro, J. L. G. Chem. Rev. 2007, 107,
3952.
(2) Maity, S. K. Renew. Sustainaible Energy Rev. 2015, 43, 1427.
(3) Olah, G. A.; Goeppert, A.; Czaun, M.; Mathew, T.; May, R. B.;
Prakash, G. K. S. J. Am. Chem. Soc. 2015, 137, 8720.
(4) Olah, G. A.; Goeppert, A.; Czaun, M.; Prakash, G. K. S. J. Am. Chem.
Soc. 2013, 135, 648.
(5) Baudouin, D.; Rodemerck, U.; Krumeich, F.; Mallmann, A. d.; Szeto,
K. C.; Ménard, H.; Veyre, L.; Candy, J.ꢀP.; Webb, P. B.; Thieuleux, C.;
Copéret, C. J. Catal. 2013, 297, 27.
(6) Kim, J.ꢀH.; Suh, D. J.; Park, T.ꢀJ.; Kim, K.ꢀL. Appl. Catal. A 2000,
197, 191.
(7) Ashok, J.; Kawi, S. ACS Catal. 2014, 4, 289.
(8) Theofanidis, S. A.; Galvita, V. V.; Poelman, H.; Marin, G. B. ACS
Catal. 2015, 5, 3028.
(9) Pakhare, D.; Spivey, J. Chem. Soc. Rev. 2014, 43, 7813.
(10) Kim, S. M.; Abdala, P. M.; Margossian, T.; Hosseini, D.; Foppa, L.;
Armutlulu, A.; van Beek, W.; ComasꢀVives, A.; Copéret, C.; Mueller, C.
J. Am. Chem. Soc. 2017, 139, 1937.
(11) Bengaard, H. S.; Nørskov, J. K.; Sehested, J.; Clausen, B. S.; Nielsen,
L. P.; Molenbroek, A. M.; RostrupꢀNielsen, J. R. J. Catal. 2002, 209, 365.
(12) Baudouin, D.; Szeto, K. C.; Laurent, P.; De Mallmann, A.; Fenet, B.;
Veyre, L.; Rodemerck, U.; Copéret, C.; Thieuleux, C. J. Am. Chem. Soc.
2012, 134, 20624.
(13) Jiao, L.; Regalbuto, J. R. J. Catal. 2008, 260, 329.
(14) Phillips, B.; Hutta, J. J.; Warshaw, I. J. Am. Ceram. Soc. 1963, 46,
579.
(15) Mazzoni, A. D.; Sainz, M. A.; Caballero, A.; Aglietti, E. F. Mater.
Chem. Phys. 2003, 78, 30.
(16) Guo, J.; Lou, H.; Zhao, H.; Chai, D.; Zheng, X. Appl. Catal. A 2004,
273, 75.
(17) Tsyganok, A. I.; Tsunoda, T.; Hamakawa, S.; Suzuki, K.; Takehira,
K.; Hayakawa, T. J. Catal. 2003, 213, 191.
(18) Guo, J.; Lou, H.; Zheng, X. Carbon 2007, 45, 1314.
(19) Carrier, X.; d'Espinose de la Caillerie, J.ꢀB.; Lambert, J.ꢀF.; Che, M.
J. Am. Chem. Soc. 1999, 121, 3377.
(20) d'Espinose de la Caillerie, J.ꢀB.; Kermarec, M.; Clause, O. J. Am.
Chem. Soc. 1995, 117, 11471.
(21) Négrier, F.; Marceau, E.; Che, M.; Giraudon, J.ꢀM.; Gengembre, L.;
Löfberg, A. J. Phys. Chem. B. 2005, 109, 2836.
(22) Low, J. E.; FoelskeꢀSchmitz, A.; Krumeich, F.; Worle, M.; Baudouin,
D.; Rascon, F.; Coperet, C. Dalton Trans. 2013, 42, 12620.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 12
(23) Oakton, E.; Vile, G.; S. Levine, D.; Zocher, E.; Baudouin, D.; Perezꢀ
(35) Sheldrick, G. Acta Crystallogr., Sect. C 2015, 71, 3.
(36) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.;
Puschmann, H. J. Appl. Crystallogr. 2009, 42, 339.
(37) JeanꢀMarie, A.; GribovalꢀConstant, A.; Khodakov, A. Y.; Monflier,
E.; Diehl, F. Chem. Comm. 2011, 47, 10767.
(38) Stavitski, E.; Weckhuysen, B. M. Chem. Soc. Rev. 2010, 39, 4615.
(39) Mette, K.; Kühl, S.; Tarasov, A.; Willinger, M. G.; Kröhnert, J.;
Wrabetz, S.; Trunschke, A.; Scherzer, M.; Girgsdies, F.; Düdder, H.;
Kähler, K.; Ortega, K. F.; Muhler, M.; Schlögl, R.; Behrens, M.;
Lunkenbein, T. ACS Catal. 2016, 6, 7238.
Ramirez, J.; Coperet, C. Dalton Trans. 2014, 43, 15138.
(24) Siddiqi, G.; Mougel, V.; Coperet, C. Dalton Trans. 2015, 44, 14349.
(25) Furman, K.; Baudouin, D.; Margossian, T.; Sabnis, K. D.; Cui, Y.;
Ribeiro, F. H.; Copéret, C. J. Catal. 2015, 324, 9.
(26) Héroguel, F.; Gebert, D.; Detwiler, M. D.; Zemlyanov, D. Y.;
Baudouin, D.; Copéret, C. J. Catal. 2014, 316, 260.
(27) Larmier, K.; Tada, S.; ComasꢀVives, A.; Copéret, C. J. Phys. Chem.
Lett. 2016, 7, 3259.
1
2
3
4
5
6
7
8
(28) Gajan, D.; Guillois, K.; Delichère, P.; Basset, J.ꢀM.; Candy, J.ꢀP.;
Caps, V.; Coperet, C.; Lesage, A.; Emsley, L. J. Am. Chem. Soc. 2009,
131, 14667.
(40) Poncelet, G.; Centeno, M. A.; Molina, R. Appl. Catal. A 2005, 288,
232.
9
(29) Kumar, P. A.; Reddy, M. P.; HyunꢀSook, B.; Phil, H. H. Catal. Lett.
2009, 131, 85.
(41) Bentaleb, F.; Che, M.; Dubreuil, A.ꢀC.; Thomazeau, C.; Marceau, E.
Catal. Today 2014, 235, 250.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(30) Maniecki, T. P.; BawolakꢀOlczak, K.; Mierczyński, P.;
Maniukiewicz, W.; Jóźwiak, W. K. Chem. Eng. J. 2009, 154, 142.
(31) Wischert, R.; Laurent, P.; Copéret, C.; Delbecq, F.; Sautet, P. J. Am.
Chem. Soc. 2012, 134, 14430.
(32) Fouskas, A.; Kollia, M.; Kambolis, A.; Papadopoulou, C.; Matralis,
H. Appl. Catal. A 2014, 474, 125.
(33) Muraza, O.; Galadima, A. Int. J. Energy Res. 2015, 39, 1196.
(34) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60,
309.
(42) Bonneviot, L.; Legendre, O.; Kermarec, M.; Olivier, D.; Che, M. J.
Colloid Interface Sci. 1990, 134, 534.
(43) Rogers, J. L.; Mangarella, M. C.; D’Amico, A. D.; Gallagher, J. R.;
Dutzer, M. R.; Stavitski, E.; Miller, J. T.; Sievers, C. ACS Catal. 2016, 6,
5873.
(44) Zhang, H.; Zhang, M.; Han, J.; Ying, G.; Guo, H.; Xu, C.; Shen, H.;
Song, N. Appl. Surf. Sci. 2011, 257, 5012.
TOC
ACS Paragon Plus Environment