Research Article

pubs.acs.org/acscatalysis

Differences in the Nature of Active Sites for Methane Dry Reforming

and Methane Steam Reforming over Nickel Aluminate Catalysts
Jessica L. Rogers,†,‡,§ Michael C. Mangarella,† Andrew D. D’Amico,∥ James R. Gallagher,⊥
Michael R. Dutzer,†,‡ Eli Stavitski,# Jeffrey T. Miller,⊥,$ and Carsten Sievers*,†,‡
†
School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332, United States
‡
Renewable Bioproducts Institute, Georgia Institute of Technology, Atlanta, Georgia 30332, United States
§
The Dow Chemical Company, Freeport, Texas 77541, United States
∥
Micromeritics Instrument Corporation, Norcross, Georgia 30093, United States
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

⊥
Argonne National Laboratory, Chemical Technology Division, Argonne, Illinois 60430, United States
#
National Synchrotron Light Source II, Brookhaven National Laboratory, Upton, New York 11973, United States
$
School of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907, United States
Downloaded via SILPAKORN UNIV on August 14, 2024 at 14:31:35 (UTC).

*
S Supporting Information

ABSTRACT: The Pechini synthesis was used to prepare nickel aluminate

catalysts with the compositions NiAl4O7, NiAl2O4, and Ni2Al2O5. The
samples have been characterized by N2 physisorption, temperature-
programmed reduction (TPR), temperature-programmed oxidation
(TPO), X-ray diffraction (XRD), X-ray photoelectron spectroscopy
(XPS), transmission electron microscopy (TEM), and X-ray absorption
spectroscopy (XAS). Characterization results indicate unique structural
properties and excellent regeneration potential of nickel aluminates.
Prepared samples were tested when unreduced and reduced prior to
reaction for methane dry reforming and methane steam reforming reactivity.
NiAl2O4 in the reduced and unreduced state as well as NiAl4O7 in the reduced state are active and stable for methane dry
reforming due to the presence of 4-fold coordinated oxidized nickel. The limited amount of metallic nickel in these samples
minimizes carbon deposition. On the other hand, the presence of metallic nickel is required for methane steam reforming.
Ni2Al2O5 in the reduced and unreduced states and NiAl2O4 in the reduced state are found to be active for methane steam
reforming due to the presence of sufficiently small nickel nanoparticles that catalyze the reaction without accumulating
carbonaceous deposits.
KEYWORDS: spinel, synthesis gas, X-ray absorption spectroscopy, coordination number, regeneration

■ INTRODUCTION
Nickel catalysts supported on γ-Al2O3 have been extensively
plumbites,29−31 while materials of the formula ABO3 are known
as perovskites.32 While it is generally proposed that nickel
used for various high-temperature applications, such as aluminates have an inverse spinel structure with nickel primarily
hydrogenation, hydrocarbon reforming, water-gas shift promo- occupying octahedral sites,33−36 some studies have reported
tion, and ammonia decomposition.1−5 One of the major that nickel primarily occupies tetrahedral sites.19,37 Further, it
challenges is the potential deactivation of these catalysts during has been reported that at high temperatures the distribution of
high-temperature reactions by sintering,6,7 carbon deposi- ions in the octahedral and tetrahedral sites can change,19,35,38
tion,6−8 or loss of active sites due to formation of indicating mobility of nickel within the host framework. These
NiAl2O4.9,10 The formation of supposedly inactive NiAl2O4 materials are of great interest for methane steam reforming and
has been cited as a major mode of deactivation for supported methane dry reforming reactions due to their resistance to
Ni/Al2O3 catalysts.9−11 sintering and the decreased formation of carbonaceous
Recently, bulk materials of the formula AB2O4/ABO3 have deposits.16,28,30 The reactivity of AB2O4/ABO3 materials has
regained attention for reactions such as methane steam been ascribed to many mechanisms: the ability of individual
reforming,12−14 partial oxidation of methane,15−17 methane atoms within the bulk structure to move reversibly in and out of
dry reforming,7,16,18−21 oxidative coupling of methane,22 the lattice,16,23 nanospinodal decomposition,24,39 the reactivity
automotive emissions control,23,24 diesel steam reforming,25,26
oxidative ethane dehydrogenation,27 tetradecane reforming,28 Received: April 21, 2016
and tar reforming.8 Spinels of the formula AB2O4 are widely Revised: July 17, 2016
accepted to have a structure similar to that of magneto- Published: July 20, 2016

© 2016 American Chemical Society 5873 DOI: 10.1021/acscatal.6b01133

ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

of oxygen contained within the material,15,18,22,27,28,31,34 the Pechini Synthesis of Nickel Aluminate Catalysts.
formation of metal nanoparticles within the bulk structure Nickel aluminates were synthesized using the Pechini syn-
framework,8,9,19,30,40,41 and the formation of nickel oxide thesis.46 The metal ions were chelated with citric acid and
particles supported by the nickel aluminate structure.20 polyesterified with ethylene glycol, forming a polymeric resin.
Several methods are commonly used to prepare mixed oxide The amount of nickel nitrate and aluminum nitrate used during
catalysts: solid-state reaction,35 coprecipitation,6,14,31,42,43 sol− synthesis was calculated on the basis of the desired content of
gel method,13,20,44 combustion synthesis,19 alkoxide meth- nickel and aluminum in the spinel structure. Metal nitrate and
od,17,18,24 and Pechini synthesis.16,40,45 In the Pechini synthesis, citric acid (citric acid:([Ni2+] + [Al3+])) in a ratio of 3:1 (molar
metal precursors are incorporated into a polymeric resin, which basis) were added to 500 mL of water, and the mixture was
is calcined in a subsequent step.45 The method is very effective stirred until all components were visibly dissolved. The mixture
for providing uniform multicomponent oxides because the resin was then heated to 70 °C and stirred for an additional 30 min
reduces differences in chemical reactivity of the different metal to ensure complete mixing and dissolution of the metal citrate
ions during the formation of the oxide. Therefore, the uniform complexes. Ethylene glycol was then added at 60 wt % relative
composition is maintained when the resin gradually decom- to the amount of citric acid used, and the mixture was slowly
poses during calcination, and segregation of the mixed oxide is heated to 120 °C. All of the water was allowed to evaporate
minimized.46,47 Another advantage of the Pechini synthesis is until the polymeric resin was formed. The produced resin was
that a much lower calcination temperature is required for aged for 2 days at room temperature prior to calcination to
synthesis in comparison to that required for a solid-state ensure homogeneity.
reaction to form the same multicomponent oxide.36 The resin was calcined using a gas hourly space velocity
Numerous studies using nickel catalysts reported that the (GHSV) of 33 mL of air min−1 gresin−1. The resin was first
active species in reforming reactions are the reduced nickel heated to 400 °C at a ramp rate of 5 °C/min and held at 400
metal particles.8,9,19,21,41 Rivas et al. found that LaNiO3 °C for 1 h until all ethylene glycol was removed. The mixed-
perovskites prepared by the Pechini synthesis exhibited the metal oxide was then further heated to 1000 °C at 5 °C/min
highest stability for methane steam reforming.9 The stability and held for 1 h. The resulting mixed-metal oxide was ground
was said to result from the formation of highly dispersed using a mortar and pestle and sieved to 75−90 μm for all
reduced nickel crystallites after reduction that remained in close subsequent experiments.
contact with the La2O3 substrate. Enger et al. studied nickel Methane Dry Reforming and Methane Steam Reform-
aluminates prepared by wet impregnation for methane steam ing Reactions. Reactivity tests were carried out under
reforming.41 They found that, after reduction in hydrogen, atmospheric pressure in a laboratory scale vertical plug flow
nickel particles were formed and were able to activate methane reactor, which was operated isothermally (Lindberg Minimite
at temperatures greater than 627 °C. Ribeiro et al. studied furnace). Brooks mass flow controllers (Model 4850) were
NiAl2O4 prepared by combustion synthesis for methane dry used to supply gas feeds. Steam was generated in the setup
reforming.19 They concluded that the reactivity of the nickel using an ISCO syringe pump (Model 500 D), which fed into a
aluminate was due to the migration of nickel to the surface, heated T-piece that utilized nitrogen as the sweep gas. Quartz
forming reduced nickel nanoparticles that coexisted with bulk beads were used in the T-piece to provide additional surface
NiAl2O4. area to allow for complete volatilization of the water feed. To
The present work investigates the activity and stability of further ensure steady flow of steam, 4 feet of heated line was
well-defined nickel aluminates of three compositions prepared installed after the T-piece to provide additional residence time
by Pechini synthesis. The materials have been thoroughly before the reactants entered the reactor. An online GC (Bruker
characterized using a suite of physicochemical techniques. The 450 RGA) was used to analyze reactor effluents. The GC was
effect of the composition and structure of NiAl4O7, NiAl2O4, equipped with a dual TCD and an FID. One TCD was used for
and Ni2Al2O5 on the performance in methane steam reforming quantification of hydrogen, while the other quantified carbon
and methane dry reforming is examined. Most importantly, monoxide, methane, carbon dioxide, and nitrogen. The FID
active sites for reforming reactions and the formation of was used to confirm methane quantification from the TCD.
carbonaceous deposits are identified. Linear calibration curves were created for the FID and TCDs

■
using various compositions (within the range of experimental
concentrations) of CO2, CH4, H2, and N2, flowing through an
EXPERIMENTAL SECTION
empty reactor at standard reaction temperature and pressure.
Materials. Chemicals used in the synthesis, Ni(NO3)2· Reactor effluent samples were taken every 15 min. Thermal
6H2O (99.999% trace metals basis), ethylene glycol (spec- conversion was investigated by running reactions in an empty
trophotometric grade, >99%), and citric acid (ACS reagent, reactor tube and was found to be negligible in all experiments.
99.5% purity), were used as received from Sigma-Aldrich. Conversion levels obtained at different flow rates (at constant
Al(NO3)3·9H2O (ACS grade, 98+%) was used as received from space time) and particle sizes showed no effect of these
Strem Chemicals. Nickel aluminates were benchmarked against variables, confirming that mass transport effects were not
a commercial 50 wt % Ni/α-Al2O3 (High Fuel) catalyst present. A GHSV of 52400 h−1 (CO2/CH4) was employed in
purchased from Alfa Aesar. Deionized water was used methane dry reforming experiments and 65500 h−1 (CO2/
throughout this study. Methane (UHP), nitrogen (UHP), CH4/N2) in methane steam reforming experiments (dry basis).
and CO2 (research grade) from Airgas were used during When reduction was employed prior to reactivity tests, the
reactivity experiments without further purification. Argon samples were heated to 600 °C at 5 °C/min in a 20% H2/N2
(UHP), 10% H2/Ar (UHP), helium (UHP), and 10% O2/He mixture and held for 2 h before the temperature was increased
(UHP) from Airgas were used for the temperature- to 700 °C. All reactivity tests were conducted at 700 °C for 12
programmed reduction (TPR) and temperature-programmed h. In the case of methane dry reforming studies, the feed gas
oxidation (TPO) experiments. was stoichiometric carbon dioxide and methane. For methane
5874 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

steam reforming studies, the feed was 20% methane with super twin field emission gun TEM. The TEM was operated at
balance nitrogen, and the steam to carbon ratio (S/C) was 2.4. 300 keV and was equipped with a single tilt tomography holder
Physicochemical Characterization. Elemental Analysis. from Fischione Instrument.50 The samples for TEM observa-
The metal content in each of the nickel aluminates was tions were prepared by creating a powder/ethanol mixture that
determined using inductively coupled plasma optical emission was added dropwise to 200 mesh copper grids with carbon film
spectroscopy (ICP-OES). The ICP-OES measurements were TEM grids.
carried out by Galbraith Laboratories. Carbon Content. Spent catalysts were analyzed for carbon
N2 Physisorption. N2 physisorption measurements were content by Galbraith Laboratories using GLI procedure ME-14.
carried out using a Micromeritics ASAP 2020 instrument. X-ray Photoelectron Spectroscopy (XPS). XPS measure-
Approximately 1.0 g of sample was used in each measurement. ments were performed on a Thermo K α XPS equipped with a
Samples were outgassed at 300 °C for 1 h prior to analysis. The monochromatic small-spot X-ray source using an aluminum
surface area was calculated using the BET method, and the pore anode (Al Kα, hν = 1486.6 eV). The background pressure was
size and volume were determined using the BJH method 4.9 × 10−8 bar, and 4 × 10−7 bar of argon was used during
applied to the desorption branch of the isotherm.48 measurement to prevent sample charging. Binding energies
TPR and TPO. TPR and TPO experiments were carried out were referenced to the sample stage, which contains built-in
on a Micromeritics AutoChem II 2920 instrument equipped calibration standards of copper, silver, and gold.
with a cold trap and a TCD detector. Approximately 0.15 g of X-ray Absorption Spectroscopy (XAS). EXAFS and XANES
sample was used in each set of experiments. Three sets of Ni K-edge spectra were collected at beamline 9-BM-C at the
experiments were performed on each sample: TPR, TPR-TPO, Advanced Photon Source at Argonne National Laboratories
and TPR-TPO-TPR2. During the TPR segment of each of the (Proposal GUP-38563). A Si(111) double monochromator was
experiments, the sample was heated from 25 to 1000 °C (10 used to select the beam energy for all measurements. The beam
°C/min) and held at 1000 °C for 30 min in a 10% H2/Ar size was 800 × 1000 μm. Samples were pressed into wafers 4
mixture (UHP), and hydrogen consumption was recorded. In mm in diameter using a cylindrical wafer press that contains six
experiments where TPO was performed after TPR, the sample slots which can accommodate samples. The beam was
was flushed under Ar (UHP) for 30 min at 900 °C and then internally calibrated with a metallic Ni reference foil, and
cooled to 35 °C under argon flow. Subsequently, the carrier gas spectra were collected in transmission mode. A spectrum was
was changed to 10% O2/He (UHP), and the sample first taken of each of the fresh samples. The samples were then
temperature was ramped from 35 to 1000 °C (10 °C/min) reduced at 900 °C in 5% H2/Ar for 2 h in an external furnace.
and held for 30 min. In experiments where TPR2 was The samples were then flushed with nitrogen and cooled to
performed following TPO measurements, the sample was room temperature, and the Ni K-edge spectra were obtained
cooled in Ar from the final TPO temperature of 1000 to 50 °C without exposing the sample to air. Reduced samples were then
and held for 30 min. The gas was then changed to 10% H2/Ar, reoxidized at 900 °C for 2 h in an external furnace, flushed with
and the TPR measurement, as described before, was repeated. nitrogen, and cooled to room temperature, and the Ni K-edge
In the control experiment, TPR followed directly by TPO was spectra were obtained without exposing the sample to air. Next,
performed as described above. fresh samples were heated to 700 °C in inert gas. The samples
X-ray Diffraction (XRD). Powder XRD patterns were were then exposed to a stoichiometric CH4/CO2 mixture for 1
measured on a Philips X’pert diffractometer equipped with an h. The samples were then flushed with nitrogen and cooled to
X’celerator module. Diffractograms were obtained with Cu Kα room temperature, and the Ni K-edge spectra were obtained
radiation for 2θ = 10−90° with a step size of 0.0167°. without exposing the sample to air.
In Situ XRD. In situ XRD reduction was performed at the Data were analyzed using WinXAS. Two polynomial fits were
X18a beamline at the National Synchrotron Light Source in used for background correction. For the pre-edge region, a
Brookhaven, NY. XRD patterns were acquired with a linear fit was used and, for the post edge, a cubic spline. The
PerkinElmer (PE) amorphous silicon area detector with 2048 edge energy, E0, was identified from the inflection point in the
× 2048 pixels and a 200 × 200 μm2 pixel size. The X-ray edge: i.e., the first maximum in the first derivative of the
photon energy was set to 20 keV. Before the diffractograms XANES spectrum. EXAFS data were then converted to k space
were collected, the detector was calibrated using a LaB6 using a mue-fit function with six spline nodes and k2 weighting.
standard. For each spectrum, 10 exposures of 4 s duration The Fourier transform of the k2 weighted EXAFS oscillation
were recorded for both the diffraction pattern and the dark (Δk = 2.7−10.9) converted the data from k space to R space to
current, and the dark current was subtracted. A Clausen cell,49 obtain a radial distribution function. The peak of interest was
which allows for the flow of hydrogen over the sample, was isolated using inverse Fourier transform. The data were curve-
used for acquisition of XRD data. The sample powder was fitted by standard methods using experimental phase shift and
loosely packed into a 1.0 mm o.d./0.9 mm i.d. quartz capillary. backscattering amplitudes derived from standard compounds of
The capillary was connected to 1/16 in. Swagelok style fittings known structure: Ni foil (12 Ni−Ni at 2.49 Å) and NiO (6 Ni−
with Vespel ferrules. An Omega thermocouple was inserted O at 2.09 Å). Difference spectra were created by subtracting
into the capillary and placed adjacent to and contacting the reaction spectra from the as-is spectra.

■
catalyst bed. Plots of the 2θ data were created from the
recorded XRD patterns using the Datasqueeze software
developed by Paul Heiney. Collected data were converted to RESULTS
Cu Kα energy. Physiochemical Properties of Prepared Catalysts. The
Transmission Electron Microscopy (TEM). TEM, high angle nickel and aluminum concentrations in the nickel aluminate
annular dark field (HAADF) scanning transmission electron samples were determined by ICP-OES (Table 1). Stoichio-
microscopy (STEM) images, and energy dispersive X-ray metric nickel aluminate (NiAl2O4) has 33.2 wt % Ni. To
(EDX) spectrometry were collected using an FEI Tecnai F30 maintain charge neutrality, the sample with 20 wt % nickel was
5875 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Table 1. Characterization of Nickel Aluminates by Elemental

Analysis, Ammonia TPD, N2 Physisorption, Temperature-
Programmed Reduction, and Temperature-Programmed
Oxidation
sample
NiAl4O7 NiAl2O4 Ni2Al2O5
Ni contenta (wt %) 20 34 45
Al contenta (wt %) 36 31 20
surface areab (m2/g) 41 37 25
pore volumec (cm3/g) 0.08 0.06 0.05
pore diameterd (nm) 5 7 6
TPR H2 uptakee (mmol/g) 3.3 5.5 7.6
TPO O2 uptakef (mmol/g) 1.4 2.1 3.1
TPR2 H2 uptakee (mmol/g) 3.3 5.5 7.6
a
Elemental analysis determined from ICP-OES at Galbraith
Laboratories. bCalculated from N2 physisorption data using the BET
equation. cCalculated from N2 physisorption data using the BJH
equation applied to the desorption branch of the adsorption isotherm.
d
Calculated from N2 physisorption data using 4V/A from BJH. eTotal
H2 consumption from TPR data. fTotal O2 consumption from TPO
data from 35 to 1000 °C.

designated NiAl4O7, the sample with 34 wt % nickel as

NiAl2O4, and the sample with 45 wt % nickel as Ni2Al2O5.
Nitrogen Physisorption. The surface areas, pore volumes,
and pore sizes of all catalysts were determined by N2
physisorption (Figure S1 in the Supporting Information and
Table 1). From the IUPAC classification all isotherms are type
IV, indicating a mesoporous solid.48 The isotherm of NiAl4O7
showed an H2 hysteresis, indicating complex pores with ill-
defined shape and size. The isotherm for NiAl2O4 contained a
H1 hysteresis, which is representative of a narrow distribution
of uniform (cylindrical) pores.48 The isotherm for Ni2Al2O5
contained a H4 hysteresis, which is attributed to the presence
of narrow slitlike pores. The presence of a less than Figure 1. Temperature-programmed reduction of NiAl4O7, NiAl2O4,
stoichiometric amount of nickel (NiAl4O7) caused irregularities and Ni2Al2O5 for fresh samples (solid lines) and for samples after TPO
in the spinel structure. A stoichiometric amount of nickel (dashed lines).
(NiAl2O4) created an ordered, uniform spinel structure. When
the nickel content was increased further (Ni2Al2O5), the pores
of the spinel structure became narrower. This might be caused the theoretical value was 5.7 mmol of H2/g (Figure S3 in the
by the accumulation of nickel or nickel oxide particles in void Supporting Information).
spaces of the slitlike pores. For all samples, the majority of the Temperature-programmed oxidation was performed on the
pore volume is distributed among pores between 4 and 30 nm. samples after temperature-programmed reduction (Figure 2).
The total pore volume decreased with increasing nickel content The total quantities of oxygen consumption (Table 1)
(Figure S2 in the Supporting Information). indicated that NiAl4O7 was 84% reoxidized, NiAl2O4 was 76%
Temperature-Programmed Reduction (TPR) and Temper- reoxidized, and Ni2Al2O5 was 80% reoxidized assuming an H2/
ature-Programmed Oxidation (TPO). NiAl4O7, NiAl2O4, and O2 stoichiometry of 2/1. However, the calculated extent of
Ni2Al2O5 exhibited high-temperature reduction peaks corre- reoxidation is lower than the actual extent of reduction because
sponding to reduction of nickel aluminate species at 870, 780, oxygen uptake at room temperature was excluded in the
and 800 °C, respectively (Figure 1). NiAl2O4 had a small peak integration of the TPO profile, as the starting temperature of
at 360 °C corresponding to nickel oxide that exists on the the TPO was approximately 35 °C. The TPO profiles showed
surface of the bulk structure. Ni2Al2O5 exhibited a large that oxidation occurred as soon as the temperature was
bimodal peak centered at 480 °C corresponding to the increased, indicating that some oxidation had likely already
reduction of two different types of NiO that existed on the taken place. For NiAl2O4, approximately 10% additional oxygen
surface of bulk aluminate.51,52 The area of the low-temperature consumption was measured, in a separate isothermal
peak was approximately half of the total hydrogen consumed, chemisorption experiment, at 35 °C. The presence of low-
indicating a NiO:NiAl2O4 ratio of about 1. All samples also and high-temperature O2 consumption peaks indicated that
exhibited a small shoulder at 960 °C. The values of hydrogen multiple types of nickel were present in the samples. NiAl4O7
consumption (Table 1) corresponded to almost complete showed a broad low-temperature oxidation peak at 440 °C and
reduction of all nickel present in the samples. The assignment a small, high-temperature oxidation peak at 970 °C. NiAl2O4
of NiAl2O4 was further validated by the TPR, where the had a primary oxidation peak at 320 °C and two smaller peaks
measured value of H2 consumption was 5.5 mmol of H2/g and at 900 and 1000 °C. Ni2Al2O5 showed a low-temperature peak
5876 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Figure 2. Temperature-programmed oxidation (TPO) of nickel

aluminates.

at 380 °C and two higher temperature peaks at 850 and 1000

°C.
Finally, the samples were exposed to a second TPR
treatment (TPR2) after TPO (Figure 1, dashed lines). The
values of H2 uptake (Table 1) were identical with those in the
initial TPR, indicating that the materials were oxidized to their
original level of oxygen content. The reduction temperature for
bulk aluminate species increased by 40 °C for NiAl4O7, 70 °C
for NiAl2O4, and 20 °C for Ni2Al2O5. For Ni2Al2O5 the bimodal
peak exhibited during TPR narrowed during TPR2 and was
attributed to agglomeration of nickel species in this sample
during the preceding treatment steps.
X-ray Diffraction (XRD). XRD measurements were per-
formed on fresh samples and after TPR, TPO, and TPR2
experiments. The dominating peaks in the diffractograms of
fresh NiAl4O7, NiAl2O4, and Ni2Al2O5 (Figure 3) were matched
to a NiAl2O4 phase using HighScore Plus and the International
Centre for Diffraction Data (ICDD) database. The diffracto-
gram of Ni2Al2O5 also exhibited additional peaks corresponding Figure 3. XRD of (a) NiAl4O7, (b) NiA2O4, and (c) NiAl2O5 after
various treatments.
to NiO. After the TPR experiment, only metallic nickel was
observed in the diffractograms. This phase may also be present,
but undetectable, in fresh and oxidized samples due to the (Figure 5b), and NiAl2O4 (Figure 5c,d). Pores and voids can
overlap of the XRD peaks of bulk NiAl2O4. XRD of samples be easily identified in the TEM images of NiAl2O4 and NiAl4O7.
after TPR, TPO, and TPR2 demonstrated the mobility of Note that the black spots result from strong diffraction of
nickel in the nickel aluminate structure by the ability of the particles with a specific orientation to the electron beam.53
material to return to a state similar to that of the fresh sample They are not indicative of a specific composition of the
after high-temperature treatments. particles. Grain sizes were calculated using dark field TEM
Samples were exposed to a hydrogen atmosphere, and the images (Figure S6 in the Supporting Information), which
temperature of the XRD reactor was incrementally increased to allowed ease of identification of individual grains due to the
the maximum temperature of the apparatus (T = 700 °C) to increased contrast. High angle annular dark field (HAADF)
collect in situ XRD patterns. No change in crystallite structure scanning transmission electron micrographs (STEM) were
was observed for NiAl4O7 or NiAl2O4 (Figures S2 and S3 in the acquired, and energy dispersive X-ray spectrometry (EDX) was
Supporting Information, respectively), as expected, on the basis performed (Table 2 and Figures S7−S9 in the Supporting
of high-temperature reduction peaks indicated by TPR. For Information). The atomic composition was found to be similar
Ni2Al2O5, formation of metallic nickel was observed in the to that obtained using ICP-OES (Table 1), with the differences
presence of hydrogen at temperatures as low as 650 °C (Figure in oxidation attributed to the sample size used during the
4). STEM-EDX measurement.
Transmission Electron Microscopy. Bright field TEM X-ray Photoelectron Spectroscopy. X-ray photoelectron
images were collected for Ni2Al2O5 (Figure 5a), NiAl4O7 (XPS) spectra were obtained for fresh aluminate samples
5877 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

a decrease in methane conversion of 8% (X(CH4)final = 24%).

The H2/CO ratio obtained during methane dry reforming over
NiAl2O4 in the reduced and unreduced states was found to 0.8.
For reduced NiAl4O7 the ratio was 0.9. Ni2Al2O5 and
commercial 50 wt % Ni/α-Al2O3 were very active for methane
conversion under dry reforming conditions, but rapid coking
required a reactor shutdown.
The catalysts were also evaluated for steam reforming at 700
°C starting in unreduced and reduced states. Ni2Al2O5 was
active for methane steam reforming when it was unreduced or
reduced prior to reaction (Figure 7). The conversion of
methane decreased relative to the initial conversion by only 7%
(X(CH4)final = 75%) for reduced Ni2Al2O5 and 2% (X(CH4)final
= 76%) for unreduced Ni2Al2O5 over 12 h on stream. NiAl2O4
was active for methane steam reforming when it was reduced
prior to reaction (Figure 7), and methane conversion decreased
by 2% over 12 h (X(CH4)final = 76%). The commercial 50 wt %
Ni/α-Al2O3 also showed sustained activity for methane steam
reforming when it was unreduced prior to reaction, and the
conversion of methane decreased by 1% (X(CH4)final = 84%)
over 12 h on stream. The conversion over commercial 50 wt %
Ni/αAl2O3 after 6 h on stream (X(CH4)initial = 85%) was
slightly greater than that of the nickel aluminates (X(CH4)initial
Figure 4. Ni2Al2O5 exposed to hydrogen as a function of increasing
temperature.
= 78−83%). However, the hydrogen concentration was higher
than the detection limit of the online GC, indicating that the
H2/CO ratio was extremely high. The H2/CO ratio obtained
(Figure S10 in the Supporting Information). The Ni 2p scan for methane steam reforming over nickel aluminates was
revealed two main peaks, Ni 2p3/2 and Ni 2p1/2, and their approximately 5 for unreduced Ni2Al2O5 and reduced NiAl2O4.
respective satellite peaks. The splitting of the main peak is due For Ni2Al2O5 reduced prior to reaction, the H2/CO ratio was
to “spin−orbit coupling” between the electron spin and the approximately 4.
angular momentum vector.54 The satellite peaks next to the Carbonaceous Deposits. The carbon content on spent
main peaks occur because there are two final states with catalysts that were active and sufficiently stable for methane
different occupation of 3d orbitals due to 4s−3d interactions.3 steam reforming and methane dry reforming was determined
Gardner et al. studied Ni-substituted aluminates by XPS and (Table 5). After methane dry reforming, nickel aluminates had
found that the separation between the satellite peak and the carbon contents between 2.8 and 12.6 wt %. The commercial
main peak of 6.3 ± 0.3 eV is due to Ni bound in the aluminate 50 wt % Ni/α-Al2O3 had a carbon content of 79 wt % when
structure and that a separation of 7.1 ± 0.1 eV is due to Ni in reduced prior to reaction after as short as 3 h on stream. For
NiO. The main peaks in our samples had a separation from the methane steam reforming, nickel aluminates had carbon
satellite peak of 6.3 ± 0.1 eV, which indicated that surface contents of less than 1.0 wt %, and the commercial 50 wt %
nickel was primarily in the form of the aluminate structure. A Ni/α-Al2O3 had a carbon content of 10.9 wt %.
binding energy (BE) of 531.3 eV was found for oxygen in X-ray Absorption Spectroscopy (XAS). XANES spectra
NiAl2O4, and the O 1s BE increased with increasing oxygen of NiAl4O7, NiAl2O4, and Ni2Al2O5 after reduction at 900 °C
content. In all samples, the Ni/Al ratio was found to be about 5 are presented in Figure 8a. The white line intensity of the Ni K-
times lower than the theoretical ratio (Table 3), which may edge, which is indicative of the coordinative environment of Ni
indicate that a certain fraction exists in a nickel-deficient species, increased in the following order: Ni2Al2O5 < NiAl2O4 <
aluminum−oxygen-rich phase deposited on the surface of the NiAl4O7. Further, the Ni2+ pre-edge is not easily observed in
NiAl2O4 spinel.9 any of the samples, indicating reduction to metallic nickel.
Methane Dry Reforming and Methane Steam Reform- Linear combination fitting of the XANES region showed that
ing Reactivity. NiAl4O7, NiAl2O4, Ni2Al2O5, and a commercial Ni2Al2O5 was completely reduced, NiAl2O4 was 85% reduced,
50 wt % Ni/α-Al2O3 catalyst from Alfa Aesar were tested for and NiAl4O7 was 65% reduced. The EXAFS curve fitting results
methane dry reforming activity at 700 °C. The activity of the for NiAl4O7 (Table 6), NiAl2O4 (Table 7), and Ni2Al2O5
samples was tested by starting in unreduced and reduced states (Table 8) confirmed that after exposure to H2 at 900 °C
prior to reaction. Methane conversion over active samples after metallic nickel species were present. The true Ni−Ni
6 h on stream is reported in Table 4. NiAl4O7 was active for coordination numbers of NiAl4O7, NiAl2O4, and Ni2Al2O5
methane dry reforming when it was reduced prior to reaction were found by normalizing the coordination number of Ni−
(Figure 6a) and had stable conversion of both methane and Ni found in the EXAFS fitting by the fraction of metallic Ni
carbon dioxide over the time period examined. Unreduced found from the linear combination fitting of the XANES. After
NiAl4O7 had low but increasing conversion. NiAl2O4 was active determination of the true Ni−Ni coordination number, the
for methane dry reforming in the reduced and unreduced states nickel nanoparticle size present in reduced samples can be
(Figure 6b). The conversion over reduced NiAl2O4 decreased estimated.55 Ni2Al2O5 was estimated to have nickel nano-
relative to the initial conversion by 6% (X(CO2)final = 61%) for particles of approximately 3−4 nm, NiAl2O4 of 6−7 nm, and
carbon dioxide and 8% (X(CH4)final = 24%) of methane. NiAl4O7 of 8−10 nm. After reduction, the samples were
Unreduced NiAl2O4 had a stable carbon dioxide conversion and oxidized at 900 °C. After oxidation, the Ni−O coordination
5878 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Figure 5. Transmission electron micrographs of fresh Ni2Al2O5 (a), NiAl4O7 (b), and NiAl2O4 (c, d).

Table 2. Average Grain Size Calculated from Dark Field Table 4. Methane Conversion over NiAl4O7, NiAl2O4,
TEM Image and Sample Composition from STEM-EDX Ni2Al2O5, and Commercial 50 wt % Ni/α-Al2O3 at 700 °C
Mapping after 6 h on Stream for Methane Steam Reforming and
Methane Dry Reforming
avg
grain avg Ni avg Al avg O empirical X(CH4)a (%) initial/final
size content content content formula from
(nm) (mol %) (mol %) (mol %) STEM-EDX reduced unreduced reduced CH4 unreduced
CH4 dry CH4 dry steam CH4 steam
NiAl4O7 14 9.3 37.8 52.9 NiAl4O6 reforming reforming reforming reforming
NiAl2O4 23 12.4 26.7 60.9 NiAl2O5
NiAl4O7 43 12 not active not active
Ni2Al2O5 14 22.5 24.6 52.9 Ni2Al2O4
NiAl2O4 32 32 78 not active
Ni2Al2O5 rapidly rapidly 82 78
Table 3. Surface Compositions of Fresh Catalysts cokedb cokedb
Determined by XPS 50 wt % cokedc rapidly rapidly 85
Ni/α-Al2O3 cokedb cokedb
BE (eV) Ni/Al a
Methane conversions were assessed at t = 6 h. bCoked requiring
Ni 2p3/2 O 1s Al 2p exptl theor reactor shutdown at t < 1 h. cCoked requiring reactor shutdown at t =
NiAl4O7 856.3 531.5 74.8 0.04 0.25 3 h.
NiAl2O4 856.3 531.3 74.6 0.11 0.5
Ni2Al2O5 856.2 531.1 74.6 0.21 1.0
The oxidation state of nickel aluminates after exposure to
methane dry reforming reaction conditions was also inves-
number was found to return to the value found in the fresh tigated (Figure 8b−d). The designation “Sample-NR” is used
samples. The coordination numbers for fresh and reoxidized when the sample was not subjected to a reductive pretreatment
samples were approximately 5, 4.5, and 4 for NiAl4O7, NiAl2O4, prior to exposure to reaction conditions and the designation
and Ni2Al2O5, respectively. “Sample-R” is used for samples that were reduced in H2 at 500
5879 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Figure 6. CO2/ CH4 conversion and H2/CO yield for methane dry reforming reaction at 700 °C, 1/1 CH4/CO2, reduced at 600 °C (2 h) prior to
reaction (R) and unreduced prior to reaction (NR) for (a) NiAl4O7 and (b) NiAl2O4.

°C prior to exposure to reaction conditions (Tables 6−8). were created. This was achieved by subtracting the EXAFS
Linear combination fitting of the XANES region using NiO and region of the spectrum of the fresh sample from the spectrum
Ni foil as references indicated that after dry reforming after methane dry reforming (Figure S11 in the Supporting
Ni2Al2O5-NR contained 90% Ni2+and 10% Ni0. After reduction
Information). EXAFS fitting of the difference spectrum using a
at 500 °C, Ni2Al2O5-R was composed of 10% Ni2+ and 90%
Ni0, but this catalyst was partially oxidized to 30% Ni2+ and 70% Ni foil standard indicated a Ni−Ni coordination number of 0.6
Ni0 after exposure to dry reforming reaction conditions (Table in Ni2Al2O5-NR, a trace of metallic Ni in NiAl2O4-NR, and no
8). After dry reforming, NiAl2O4-NR contained only Ni2+, as metallic Ni in NiAl4O7-NR (Figure 9). The true coordination
evidenced from the pre-edge of the XANES; however, this number is determined by dividing the experimental coordina-
could not be fit with a NiO XANES reference. After 500 °C tion number by the fraction of metallic Ni determined from the
reduction the catalyst contained 40% Ni2+ and 60% Ni0, which XANES. Thus, the true coordination number of metallic Ni in
was partially oxidized to 90% Ni2+ and 10% Ni0 after
Ni2Al2O5-NR is approximately 6, since only 10% of the nickel
subsequent exposure to methane dry reforming reaction
conditions. In contrast, it was not possible to fit the spectra was present as nickel nanoparticles. From the true coordination
of NiAl4O7-NR, NiAl4O7-R, and NiAl2O4-NR with a linear number, the nickel nanoparticles sizes are estimated to be 1−2
combination of the spectra of Ni0 and NiO despite a decrease nm.55 NiAl2O4-NR contained little or no Ni−Ni coordination
in white line intensity. The pre-edge features in the XANES after exposure to reaction conditions; therefore, the difference
region of these samples indicated the presence of Ni2+ species spectra were fit using a Ni−O scattering path (Figures S11−
with a XANES shape different from that of NiO. Therefore, the S14 in the Supporting Information). For NiAl2O4-R, the sample
reduction of the white line intensities during methane dry
was found to contain a higher amount of metallic nickel than
reforming is attributed to a change in the coordination sphere
of Ni2+ species which remain part of the oxide phase. did NiAl2O4-NR, which explains the increase in the amount of
To gain further insight into the changes in structure of nickel carbonaceous deposits observed for this sample when it is
aluminates during methane dry reforming, difference spectra reduced prior to exposure to the reaction conditions.
5880 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Figure 7. (a) Methane conversion and (b) H2/CO ratio methane steam reforming at 700 °C, S/C = 2.4, for Ni2Al2O5 and NiAl2O4 reduced at 600
°C (2 h) prior to reaction (R) and unreduced prior to reaction (NR).

Table 5. Carbon Analysis of Spent Samples Utilized in 12 h diffusion of metal ions during the synthesis is reduced by
Reactor Experiments gradual decomposition of the resin.47,56,57
The nickel aluminates prepared in the present study had
amt of carbonaceous deposits on spent samples (wt % C)
moderate surface areas (25−40 m2/g) and were mesoporous,
reduced unreduced reduced CH4 unreduced allowing reactants to access a significant part of the material.
CH4 dry CH4 dry steam CH4 steam
reforming reforming reforming reforming The grain size of fresh samples found from TEM ranges from
NiAl4O7 5.0 2.8 NDa NDa 14 to 23 nm. As expected, STEM-EDX analysis indicated
NiAl2O4 12.6 7.5 1.1 NDa nanoscale stoichiometry similar to the bulk stoichiometry
Ni2Al2O5 NDb NDb <0.5 <0.5 determined by ICP. However, XPS, which only analyzes the
50 wt % 79.0c NDb NDb 10.9 sample up to a depth of about 5 nm,58 showed that these
Ni/α-Al2O3
a
materials have an aluminum- and oxygen-rich surface at room
Not determined (ND) because samples were not active. bNot temperature. In agreement with previous studies,47,56,57 our
determined (ND) because samples coked, causing reactor shutdown at results indicate that the Pechini synthesis provides materials
t < 1 h. cReactor was shut down at t = 3 h due to coking. The carbon
content was determined for comparative purposes. that have uniform compositions, a high degree of crystallinity,
and small grain sizes. In contrast, synthesis of stoichiometric
NiAl2O4 by a solid-state reaction was found to produce a

■ DISCUSSION
Formation of Nickel Aluminates. The formation of a
material with several types of NiO in addition to bulk NiAl2O4.9
The nonuniform catalysts severely sintered and formed large
quantities of carbonaceous deposits during catalytic methane
rigid polymeric resin during Pechini synthesis allows for the decomposition. In contrast to the current study, synthesis of
formation of highly uniform nickel aluminates because the NiAl2O4 by a solid-state reaction produced a catalyst with
5881 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Figure 8. Ni K-edge XANES from 8.32 to 8.37 keV of NiAl4O7, NiAl2O4, and Ni2Al2O5 (a) after reduction at 900 °C and (b−d) fresh samples and
samples after exposure to methane dry reforming (CH4/CO2) reaction conditions at 700 °C. NR denotes samples unreduced prior to exposure to
reaction gas.

Table 6. Curve-Fitting Results for the Ni K-Edge EXAFS of the NiAl4O7

XANES fraction
treatment Ni2+ Ni0 scatter N R, Å Δσ2 (×103) E0, eV comment
as is 1.0 Ni−O 5.1 2.06 0.0 −5.0
900 H2 0.4 0.6 Ni−O 2.0 2.05 0.0 4.4 8−10 nm
Ni−Ni 7.5a 2.50 0.0 0.9
900 O2 1.0 Ni−O 5.4 2.08 5.0 0.4
CH4/CO2 reforming 700 °C (NR) 1.0 Ni−O 4.6 2.06 0.0 −1.0 change in Ni−O no Ni0
500 °C H2 pretreatment 0.4 0.6 Ni−O 1.8 2.06 0.0 −5.0 >10 nm Ni0
Ni−Ni 7.3a 2.49 0.0 −13.7
CH4/CO2 reforming 700 °C (R) 1.0 Ni−O 5.0 2.06 0.0 −5.3 500 °C H2 pretreatment
a
Calculated using empirical correlation of H2 chemisorption and EXAFS CN for platinum.55

several NiO species that sintered and coked during methane high catalytic activity and a long lifetime. In a separate study by
reforming. Kizaki et al., La(Fe1−xPdx)O3 perovskites prepared by the
Mobility of Nickel Species in Nickel Aluminates. alkoxide method were found to have excellent stability and
Materials that are similar to nickel aluminates have shown activity for automotive emissions control due to nanospinodal
unique regenerative capabilities. Nishihata et al. studied Pd decomposition.24 In nanospinodal decomposition, the perov-
perovskites prepared by the alkoxide method for automotive skite decomposes into LaFeO3 and PdFeO3 at high temper-
emissions control.23 It was found that Pd was able to move out atures, but the crystal structure itself is maintained. Li et al.
of the perovskite lattice under reducing conditions and back in examined Ni−Fe−Mg−Al-O catalysts prepared by coprecipita-
its original position under an oxidizing atmosphere, giving it tion for tar reforming of biomass-derived syngas.8 Under
5882 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Table 7. Curve-Fitting Results for the Ni K-Edge EXAFS of the NiAl2O4

XANES
fraction
treatment Ni2+ Ni0 scatter N R, Å Δσ2 (×103) E0, eV comment
as is 1.0 Ni−O 4.4 2.07 0.0 −5.0
900 H2 0.15 0.85 Ni−Ni 8.9a 2.50 1.0 0.3 6−7 nm Ni0
900 O2 1.0 Ni−O 4.5 2.07 0.4 −1.8
CH4/CO2 reforming 700 °C (NR) 1.0 Ni−O 3.6 2.07 0.0 −0.4 change in Ni−O trace Ni0
Ni−O difference 1.0b 2.06 0.0 −1.1
500 °C H2 pretreatment 0.4 0.6 Ni−O 1.8 2.07 0.0 −5.2 >10 nm Ni0
Ni−Ni 7.1a 2.49 0.0 −13.8
CH4/CO2 reforming 700 °C (R) 0.9 0.1 Ni−O 3.9 2.08 0.0 −5.0 500 °C H2 pretreatment trace metallic Ni
Ni−Ni difference 2.1a,b 2.49 4.0 −13.2
a
Calculated using empirical correlation of H2 chemisorption and EXAFS CN for platinum.55 bDifference spectrum created by subtracting the
spectrum of the fresh sample from the spectrum of the sample after exposure to described experimental conditions.

Table 8. Curve-Fitting Results for the Ni K-Edge EXAFS of the Ni2Al2O5

XANES fraction
treatment Ni2+ Ni0 scatter N R, Å Δσ2 (×103) E0, eV comment
as is 1.0 Ni−O 4.2 2.09 1.0 0.2
900 H2 1.0 Ni−Ni 8.5a 2.49 1.0 0.7 3−4 nm Ni0
900 O2 1.0 Ni−O 3.7 2.10 0.0 0.6
CH4/CO2 reforming 700 °C (NR) 0.9 0.1 Ni−O 4.0 2.09 1.0 0.4 1−2 nm Ni0
Ni−Ni 6.0a,b 2.45 5.0 −12.7
500 °C H2 pretreatment 0.1 0.9 Ni−Ni 9.1a 2.49 0.0 −13.8 7−8 nm Ni0
CH4/CO2 reforming 700 °C (R) 0.3 0.7 Ni−O 3.9 2.08 0.0 −5.0 3−4 nm Ni0
Ni−Ni 5.8a,b 2.49 1.0 −13.5
a
Calculated using empirical correlation of H2 chemisorption and EXAFS CN for platinum.55 bDifference spectrum created by subtracting the
spectrum of the fresh sample from the spectrum of the sample after exposure to the described experimental conditions.

reducing conditions, uniform Ni−Fe nanoparticles were formed reduced, and Ni2Al2O5 can be 100% reduced. After oxidation at
from the bulk mixed oxide, and under oxidizing conditions, the 900 °C, the Ni−O coordination number of nickel aluminates
Ni−Fe nanoparticles were reincorporated into the bulk was the same as that in the fresh samples. We conclude that the
structure. ability of these materials to return to their initial state after
In the present study, TPR/TPO, XRD, and XAS exposure to oxidizing and reducing conditions at high
demonstrated the regeneration potential of nickel aluminates temperature indicates their potential application for many
at high temperatures in terms of reversible changes of their high-temperature reactions such as methane steam reforming
oxidation state. All samples were almost fully reduced (TPR), and methane dry reforming.
reoxidized (TPO), and then re-reduced (TPR2). The TPR Active Site for Methane Steam and Dry Reforming
profiles showed that NiAl4O7 and NiAl2O4 had predominantly and Their Role in Carbon Formation. Methane dry
high-temperature reduction peaks associated with reduction of reforming occurs according to the stoichiometric equation59,60
bulk nickel aluminate. NiO and bulk nickel aluminates were CH4 + CO2 ↔ 2CO + 2H 2 ΔH °298K = 247 kJ/mol
reduced in Ni2Al2O5. The presence of NiO was indicated by the
(1)
low-temperature reduction peak. The broad TPO patterns
indicated that there were several types of oxygen acceptor sites while methane steam reforming (eq 2) follows the equation
present within the samples. TPR2 showed that the reduction
CH4 + H 2O ↔ CO + 3H 2 ΔH °298K = 206 kJ/mol
peak temperatures were only slightly increased, which indicated
that the mobility of nickel was only slightly reduced after (2)
repeated exposure to high temperatures. The low-temperature The water-gas shift reaction (eq 3) and formation of
region of the TPR2 profile of Ni2Al2O5 was dominated by a carbonaceous species (eqs 4−6) can also occur under the
peak around 340 °C instead of the bimodal peak observed for experimental conditions and can have a significant effect on the
the fresh sample, which is likely due to agglomeration of nickel target reactions (eqs 1 and 2).
oxide nanoparticles on the aluminate surface creating somewhat
larger nickel nanoparticles. XRD was performed on samples CO + H 2O ↔ CO2 + H 2 ΔH °298K = 41 kJ/mol (3)
after each TPR/TPO treatment and demonstrated the mobility
of nickel particles out of the host framework by the ability of CH4 ↔ C + 2H 2 ΔH °298K = 75 kJ/mol (4)
the materials to return to a structure similar to their initial
structure under conditions more severe than those used during 2CO ↔ C + CO2 ΔH °298K = − 171 kJ/mol (5)
reactivity studies. XAS analysis complemented TPR/TPO and
XRD results. Specifically, XANES indicated that, after reduction CO + H 2 ↔ C + H 2O ΔH °298K = − 131 kJ/mol
at 900 °C, NiAl4O7 can be 65% reduced, NiAl2O4 can be 85% (6)

5883 DOI: 10.1021/acscatal.6b01133

ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

Figure 9. Ni K-edge FT of the difference EXAFS of nickel aluminates as synthesized minus catalyst after exposure to methane dry reforming (CH4/
CO2) reaction conditions (NR).

H2/CO ratios of less than unity during methane dry dry reforming activity with limited coke formation due to the
reforming are commonly attributed to the inverse water-gas formation of small metal particles that form from the aluminate
shift reaction.61 H2/CO ratios from 3 to 16 have been observed structure at the same temperature of the catalytic reaction.21
during methane steam reforming and can be attributed to the Many studies using nickel catalysts reported that the active
forward water-gas shift reaction.13 The different directions of species in reforming reactions are the reduced nickel metal
the water-gas shift (eq 3) during methane steam reforming and particles.8,9,19,21,41 Our study found that NiAl2O4 was active for
methane dry reforming, respectively, result from the different methane dry reforming in the unreduced and reduced states.
compositions of the feed for these reactions. Reduced NiAl4O7 was immediately active for methane dry
Methane steam reforming has been proposed to proceed via reforming, and the conversion of methane and carbon dioxide
dehydrogenation of methane to form surface carbon, which is was higher than that over NiAl2O4 in either the reduced or
oxidized by adsorbed oxygen from water dissociation.14 unreduced state. The absence of metallic Ni in nonreduced or
Methane dry reforming is thought to proceed via decom- reduced NiAl4O7 and the low amount of metallic nickel present
position of methane to form surface carbon and hydrogen and in nonreduced or reduced NiAl2O4 during methane dry
reaction of surface carbon with carbon dioxide.20 Fundamen- reforming provides strong evidence that the active site
tally, the mechanisms of methane steam reforming and responsible for stable activity is an oxide. It is speculated that,
methane dry reforming may not be significantly different.61 in the case of reduced NiAl4O7, the reduction facilitates the
However, the thermodynamics of methane dry reforming favor formation of accessible 4-fold coordinated nickel oxide species
formation of nickel carbide or carbon, while methane steam that are active and stable for methane dry reforming. Further,
reforming favors the formation of nickel or nickel oxide.61 Deng the activity of unreduced NiAl4O7 was low and slowly increased
et al. showed that LaPrFeNiO3 catalysts prepared by a modified during time on stream because it takes longer for the Ni atoms
Pechini synthesis were deactivated by formation of stable, to migrate to an environment with a Ni−O coordination
filamentous carbon during dry reforming, while during number around 4, where they are accessible and catalytically
oxidative methane reforming (CH4:CO2:O2 = 1:0.7:0.3) no active. In the case of NiAl2O4, the coordination of the as-
carbonaceous deposits were observed.16 Under reducing synthesized oxide was close to 4, which explains the observation
conditions, Ni-deficient nickel aluminates prepared by the of the same activity for methane dry reforming in reduced and
sol−gel method formed dispersed nickel nanoparticles that unreduced samples. Fourfold-coordinated oxidic nickel species
were active for methane steam reforming.13 Due to the small can be present in a square-planar geometry, which would allow
size of these particles, carbon formation was very low. Similarly, for nickel in such coordination to be accessible to reactants.62
it was reported that NiAl2O4 prepared by a sol−gel had high In a related study, tetrahedrally coordinated nickel was
5884 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

proposed as an active site for low weight loading Ni/Al2O3 if the syngas product is to be used for anything other than
catalysts in the highly selective oxidative dehydrogenation of hydrogen production. The conversion of methane decreased
ethane to ethylene.63 during the 12 h study by less than 7% for all samples. The
Importantly, the active oxide, presumably 4-fold-coordinated commercially available 50 wt % Ni/α-Al2O3 catalyst formed 10
Ni2+species, appears to play little to no part in the formation of times the amount of carbonaceous deposits in comparison to
carbonaceous deposits. The formation of such deposits is the active nickel aluminates. These results indicate that reduced
enhanced when reduced nickel particles are present. Specifi- Ni2Al2O5 provides optimal performance for methane steam
cally, NiAl4O7-R and NiAl2O4-R/NR had only trace amounts of reforming while limiting the formation of carbon.
metallic Ni at the beginning of the reaction (Tables 6 and 7)
and formed limited amounts of carbonaceous deposits during
dry reforming (Table 5). Deactivation of stable nickel
■ CONCLUSIONS
Nickel aluminates prepared by Pechini synthesis are excellent
aluminates was less than 6% of carbon dioxide and 8% of candidates for high-temperature reforming reaction catalysts.
methane during 12 h on stream, and the H2/CO ratio was very Physicochemical characterization revealed unique structural
close to unity. properties, indicated a high degree of mobility of nickel in the
During the dry reforming reaction over Ni2Al2O5, 10% of the aluminate structure, and demonstrated the regeneration
Ni was reduced to nickel nanoparticles of 1−2 nm when the properties of nickel aluminates under harsh reaction conditions.
sample was not reduced prior to reaction (Table 8) and 70% of The minimal change to the materials upon redox cycling, as
the Ni was reduced to nickel nanoparticles of 1−2 nm when the seen by TPRs and TPO, indicates that it may be possible to
sample was reduced prior to reaction. The rapid accumulation regenerate these materials after some level of coking has
of carbonaceous deposits on this catalyst indicated that the occurred. Fourfold-coordinated nickel species on NiAl4O7 and
presence of metallic nickel favors the formation of substantial NiAl2O4 are identified as potential active sites for methane dry
amounts of carbon deposits. Similar to the case for Ni2Al2O5, reforming, providing high and stable activity. These sites form
the commercial 50 wt % Ni/α-Al2O3 catalyst accumulated few carbonaceous deposits under dry reforming conditions,
substantial amounts of deposits (79 wt % carbon in only 3 h on whereas significant amounts of such deposits are formed on the
stream), which is attributed to the presence of metallic Ni metallic nickel particles present in reduced NiAl2O4 and
particles. Note that the formation of carbonaceous deposits on unreduced/reduced Ni2Al2O5. The metallic nickel nanoparticles
metallic nickel is thermodynamically favored under methane that cause the formation of deposits in methane dry reforming
dry reforming conditions and that this reaction appears to are believed to be the active sites for methane steam reforming.
involve nickel carbide/carbon species as intermediates. In Reduced NiAl2O4 and unreduced/reduced Ni2Al2O5 had high
summary, these results indicate reduced NiAl4O7 contained the and stable activity during methane steam reforming studies. In
right proportion of 4-fold-coordinated nickel to provide contrast to dry reforming, reduced nickel particles are the active
optimal performance for methane dry reforming while limiting sites for this reaction.
carbon formation.
In the case of methane steam reforming, reduced nickel
nanoparticles are the active phase.14 Reduced and unreduced
■ ASSOCIATED CONTENT
* Supporting Information
S
Ni2Al2O5 and reduced NiAl2O4 were active for methane steam The Supporting Information is available free of charge on the
reforming and had H2:CO ratios of between 4 and 5. The ACS Publications website at DOI: 10.1021/acscatal.6b01133.
slightly lower methane conversion of the unreduced Ni2Al2O5 Equations for calculations of conversions and yields, TPR
and reduced NiAl2O4 is attributed to lower densities of nickel data, in situ X-ray diffractograms, TEM diffraction
nanoparticles relative to that of reduced Ni2Al2O5. This is in patterns and images, HAADF-STEM-EDX results, X-
agreement with in situ XRD measurements, which indicated the ray photoelectron spectra, and EXAFS data (PDF)

■
formation of metallic nickel in Ni2Al2O5 in hydrogen at
temperatures as low as 650 °C. Reduced NiAl2O4 is believed to
AUTHOR INFORMATION
exhibit activity for methane steam reforming because of an
increase in metallic nickel, as metallic nickel particles >10 nm Corresponding Author
were observed for NiAl2O4 reduced at 500 °C from XAS. It is *C.S.: tel, +1-404-385-7685; fax, +1-404-894-2866; e-mail,
speculated that unreduced NiAl2O4 was not active for methane [email protected].
steam reforming because it only had few or no nickel Notes
nanoparticles, as indicated by the XAS results of this unreduced The authors declare no competing financial interest.
sample during methane dry reforming. XAS measurements
showed that 10% of the Ni in Ni2Al2O5 was reduced during
methane dry reforming. We speculate that a similar amount of
■ ACKNOWLEDGMENTS
Derrick Flick, Brien Stears, Mark Siddoway, and Billy Bardin
reduced nickel nanoparticles exists when Ni2Al2O5 is used for from the Dow Chemical Company are thanked for fruitful
methane steam reforming without prior reduction. When discussions. The assistance of Tianpin Wu at beamline 9-BM-C
Ni2Al2O5 was reduced at 500 °C, it was found to contain 90% at Argonne National Laboratory in collecting XAS measure-
Ni0 nickel and nanoparticles of 7−8 nm exist, which can explain ments is greatly appreciated (Proposal GUP-38563). The
the higher activity observed for Ni2Al2O5-R. Carbon formation authors also wish to thank Yong Ding for acquiring TEM
was very low on active aluminate samples, which could indicate images. We thank Justin Notestein and Andrew Korinda for
that the nickel nanoparticles remained small throughout the their help in orientation and data collection techniques at
reaction, as carbon nucleation requires nickel particles of a Argonne National Laboratory. Rich discussions regarding XRD
certain size.21 Unreduced commercial 50 wt % Ni/α-Al2O3 was with Dr. Z. John Zhang are acknowledged. We also thank
slightly more active than the nickel aluminates. However, the Steven Ehrlich for his help with the synchrotron experiment.
H2/CO ratio was extremely high, which is an undesirable result The work was funded by The Dow Chemical Company and a
5885 DOI: 10.1021/acscatal.6b01133
ACS Catal. 2016, 6, 5873−5886
ACS Catalysis Research Article

fellowship from the Renewable Bioproducts Institute at the (29) Collongues, R.; Gourier, D.; Kahnharari, A.; Lejus, A. M.; Thery,
Georgia Institute of Technology. J.T.M. and J.R.G. were funded J.; Vivien, D. Annu. Rev. Mater. Sci. 1990, 20, 51−82.
by the U.S. Department of Energy, Office of Basic Energy (30) Chu, W. L.; Yang, W. S.; Lin, L. W. Appl. Catal., A 2002, 235,
Sciences, Chemical Sciences, under Contract DE-AC-02- 39−45.
(31) Gardner, T. H.; Spivey, J. J.; Kugler, E. L.; Campos, A.; Hissam,
06CH11357. Use of the Advanced Photon Source was J. C.; Roy, A. D. J. Phys. Chem. C 2010, 114, 7888−7894.
supported by the U.S. Department of Energy, Office of (32) Tejuca, L. G.; Fierro, J. L. G. Properties and applications of
Science, Office of Basic Energy Sciences, under Contract No. perovskite-type oxides; Marcel Dekker: New York, 1993.
DE-AC02-06CH11357. This research used the X18A beamline (33) Gaudon, M.; Robertson, L. C.; Lataste, E.; Duttine, M.;
of the National Synchrotron Light Source, a U.S. Department Menetrier, M.; Demourgues, A. Ceram. Int. 2014, 40, 5201−5207.
of Energy (DOE) Office of Science User Facility operated for (34) Murthy, I.; Swamy, C. S. J. Mater. Sci. 1993, 28, 1194−1198.
the DOE Office of Science by Brookhaven National (35) Han, Y. S.; Li, J. B.; Ning, X. S.; Chi, B. J. Am. Ceram. Soc. 2004,
Laboratory, under Contract No. DE-AC02-98CH10886. 87, 1347−1349.

■
(36) Han, Y. S.; Li, J. B.; Ning, X. S.; Yang, X. Z.; Chi, B. Mater. Sci.
Eng., A 2004, 369, 241−244.
REFERENCES (37) San Jose-Alonso, D.; Illan-Gomez, M. J.; Roman-Martinez, M.
(1) Yung, M. A.; Jablonski, W. S.; Magrini-Bair, K. A. Energy Fuels C. Int. J. Hydrogen Energy 2013, 38, 2230−2239.
2009, 23, 1874−1887. (38) Zhang, Z. J.; Wang, Z. L.; Chakoumakos, B. C.; Yin, J. S. J. Am.
(2) Sutton, D.; Kelleher, B.; Ross, J. R. H. Fuel Process. Technol. 2001, Chem. Soc. 1998, 120, 1800−1804.
73, 155−173. (39) Hamada, I.; Uozumi, A.; Morikawa, Y.; Yanase, A.; Katayama-
(3) Winkler, A.; Borchert, H.; Al-Shamery, K. Surf. Sci. 2006, 600, Yoshida, H. J. Am. Chem. Soc. 2011, 133, 18506−18509.
3036−3044. (40) Katz, M. B.; Zhang, S.; Duan, Y.; Wang, H.; Fang, M.; Zhang,
(4) Muirhead-Gould, J. S.; Laidler, K. J. Trans. Faraday Soc. 1967, 63, K.; Li, B.; Graham, G. W.; Pan, X. J. Catal. 2012, 293, 145−148.
944−952. (41) Enger, B. C.; Lodeng, R.; Walmsley, J.; Holmen, A. Appl. Catal.,
(5) Ewbank, J. L.; Kovarik, L.; Diallo, F. Z.; Sievers, C. Appl. Catal., A A 2010, 383, 119−127.
2015, 494, 57−67. (42) Jeevanandam, P.; Koltypin, Y.; Gedanken, A. Nano Lett. 2001, 1,
(6) Bhattacharyya, A.; Chang, V. W. Stud. Surf. Sci. Catal. 1994, 88, 263−266.
207−213. (43) Santiago, M.; Perez-Ramirez, J. Environ. Sci. Technol. 2007, 41,
(7) Xu, Z. L.; Zhen, M.; Bi, Y. L.; Zhen, K. J. Appl. Catal., A 2000, 1704−1709.
198, 267−273. (44) Cui, H. T.; Zayat, M.; Levy, D. J. Sol-Gel Sci. Technol. 2005, 35,
(8) Li, D.; Koike, M.; Wang, L.; Nakagawa, Y.; Xu, Y.; Tomishige, K. 175−181.
ChemSusChem 2014, 7, 510−522. (45) Pechini, M. U.S. Patent 3,438,723, 1969.
(9) Rivas, M. E.; Fierro, J. L. G.; Guil-Lopez, R.; Pena, M. A.; La (46) Gama, L.; Ribeiro, M. A.; Barros, B. S.; Kiminami, R. H. A.;
Parola, V.; Goldwasser, M. R. Catal. Today 2008, 133-135, 367−373. Weber, I. T.; Costa, A. J. Alloys Compd. 2009, 483, 453−455.
(10) Cimino, A.; Lo Jacono, M.; Schiavello, M. J. Phys. Chem. 1971, (47) Kakihana, M.; Yoshimura, M. Bull. Chem. Soc. Jpn. 1999, 72,
75, 1044−1050. 1427−1443.
(11) Bartholomew, C. H.; Farrauto, R. J. J. Catal. 1976, 45, 41−53. (48) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.;
(12) Ross, J. R. H.; Steel, M. C. F.; Zeiniisfahani, A. J. Catal. 1978, 52, Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem.
280−290. 1985, 57, 603−619.
(13) Salhi, N.; Petit, C.; Kiennemann, A. Stud. Surf. Sci. Catal. 2008, (49) Stampfl, C. Catal. Today 2005, 105, 17−35.
174, 1335−1338. (50) Ding, Y.; Liu, Y.; Niu, S. M.; Wu, W. Z.; Wang, Z. L. J. Appl.
(14) Alubaid, A.; Wolf, E. E. Appl. Catal. 1988, 40, 73−85. Phys. 2014, 116, 154304.
(15) Chu, W. L.; Yang, W. S.; Lin, L. W. Catal. Lett. 2001, 74, 139− (51) Zielinski, J. J. Catal. 1982, 76, 157−163.
144. (52) Lopez-Fonseca, R.; Jimenez-Gonzalez, C.; de Rivas, B.;
(16) Deng, J.; Cai, M.; Sun, W.; Liao, X.; Chu, W.; Zhao, X. S. Gutierrez-Ortiz, J. I. Appl. Catal., A 2012, 437-438, 53−62.
(53) Williams, D.; Carter, B. Transmission Electron Microscopy: A
ChemSusChem 2013, 6, 2061−2065.
Textbook for Materials Science, 2nd ed.; Springer: New York, 2009.
(17) Machida, M.; Eguchi, K.; Arai, H. J. Catal. 1989, 120, 377−386.
(54) Briggs, D.; Heyden, E.; London, S. Handbook of X-ray and
(18) Zhang, K.; Zhou, G. D.; Li, J.; Cheng, T. X. Catal. Commun.
Ultraviolet Photoelctron Spectroscopy; Heyden & Son: London, 1977.
2009, 10, 1816−1820.
(55) Miller, J. T.; Kropf, A. J.; Zha, Y.; Regalbuto, J. R.; Delannoy, L.;
(19) Ribeiro, N. F. P.; Neto, R. C. R.; Moya, S. F.; Souza, M. M. V.
Louis, C.; Bus, E.; van Bokhoven, J. A. J. Catal. 2006, 240, 222−234.
M; Schmal, M. Int. J. Hydrogen Energy 2010, 35, 11725−11732.
(56) Liu, X. M.; Gao, W. L. Mater. Manuf. Processes 2012, 27, 905−
(20) Gonzalez, A. R.; Asencios, Y. J. O.; Assaf, E. M.; Assaf, J. M.
909.
Appl. Surf. Sci. 2013, 280, 876−887.
(57) Ferri, E. A. V.; Santos, I.; Radovanovic, E.; Bonzanini, R.;
(21) Sahli, N.; Petit, C.; Roger, A. C.; Kiennemann, A.; Libs, S.;
Girotto, E. M. J. Braz. Chem. Soc. 2008, 19, 1153−1157.
Bettahar, M. M. Catal. Today 2006, 113, 187−193.
(58) Appapillai, A. T.; Mansour, A. N.; Cho, J.; Shao-Horn, Y. Chem.
(22) Golpasha, R. D.; Karami, D.; Ahmadi, R.; Bagherzadeh, E. React.
Mater. 2007, 19, 5748−5757.
Kinet. Catal. Lett. 1993, 51, 393−400.
(59) Zhang, J. G.; Wang, H.; Dalai, A. K. J. Catal. 2007, 249, 300−
(23) Nishihata, Y.; Mizuki, J.; Akao, T.; Tanaka, H.; Uenishi, M.;
310.
Kimura, M.; Okamoto, T.; Hamada, N. Nature 2002, 418, 164−167.
(60) Xu, J.; Froment, G. F. AIChE J. 1989, 35, 88−96.
(24) Kizaki, H.; Kusakabe, K.; Nogami, S.; Katayama-Yoshida, H.
(61) Wang, S. B.; Lu, G. Q. M.; Millar, G. J. Energy Fuels 1996, 10,
Appl. Phys. Express 2008, 1, 104001. 896−904.
(25) Achouri, I. E.; Abatzoglou, N.; Fauteux-Lefebvre, C.; Braidy, N.
(62) Wiberg, N.; Holleman, A. F.; Wiberg, E. Inorganic Chemistry;
Catal. Today 2013, 207, 13−20. Academic Press: San Diego, CA, 2001.
(26) Fauteux-Lefebvre, C.; Abatzoglou, N.; Braidy, N.; Achouri, I. E. (63) Heracleous, E.; Lee, A. F.; Wilson, K.; Lemonidou, A. A. J. Catal.
J. Power Sources 2011, 196, 7673−7680. 2005, 231, 159−171.
(27) Skoufa, Z.; Xantri, G.; Heracleous, E.; Lemonidou, A. A. Appl.
Catal., A 2014, 471, 107−117.
(28) Gardner, T. H.; Shekhawat, D.; Berry, D. A.; Smith, M. W.;
Salazar, M.; Kugler, E. L. Appl. Catal., A 2007, 323, 1−8.

5886 DOI: 10.1021/acscatal.6b01133

ACS Catal. 2016, 6, 5873−5886