Conversion of dimethyl ether to 2,2,3-trimethylbutane over a Cu/BEA catalyst: Role of Cu sites in hydrogen incorporation

Article abstract of DOI:10.1021/cs501876w

Recently, it has been demonstrated that methanol and/or dimethyl ether can be converted into branched alkanes at low temperatures and pressures over large-pore acidic zeolites such as H-BEA. This process achieves high selectivity to branched C₄ (e.g., isobutane) and C₇ (e.g., 2,2,3-trimethylbutane) hydrocarbons. However, the direct homologation of methanol or dimethyl ether into alkanes and water is hydrogen-deficient, resulting in the formation of unsaturated alkylated aromatic residues, which reduce yield and can contribute to catalyst deactivation. In this paper we describe a Cu-modified H-BEA catalyst that is able to incorporate hydrogen from gas-phase H₂ cofed with dimethyl ether into the desired branched alkane products while maintaining the high C₄ and C₇ carbon selectivity of the parent H-BEA. This hydrogen incorporation is achieved through the combination of metallic Cu nanoparticles present on the external surface of the zeolite, which perform H₂ activation and olefin hydrogenation, and Lewis acidic ion-exchanged cationic Cu present within the H-BEA pores, which promotes hydrogen transfer. With cofed H₂, this multifunctional catalyst achieved a 2-fold increase in hydrocarbon productivity in comparison to H-BEA and shifted selectivity toward products favored by the olefin catalytic cycle over the aromatic catalytic cycle.

Full text of DOI:10.1021/cs501876w

Research Article
pubs.acs.org/acscatalysis
Conversion of Dimethyl Ether to 2,2,3-Trimethylbutane over
a Cu/BEA Catalyst: Role of Cu Sites in Hydrogen Incorporation
Joshua A. Schaidle,*,† Daniel A. Ruddy,‡ Susan E. Habas,† Ming Pan,† Guanghui Zhang,§
Jeffrey T. Miller,§ and Jesse E. Hensley†
‡
†National Bioenergy Center and Chemistry and Nanoscience Center, National Renewable Energy Laboratory, Golden,
Colorado 80401, United States
§Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne, Illinois 60439, United States
S
* Supporting Information
ABSTRACT: Recently, it has been demonstrated that methanol and/or dimethyl ether can be converted into branched alkanes
at low temperatures and pressures over large-pore acidic zeolites such as H-BEA. This process achieves high selectivity to
branched C4 (e.g., isobutane) and C7 (e.g., 2,2,3-trimethylbutane) hydrocarbons. However, the direct homologation of methanol
or dimethyl ether into alkanes and water is hydrogen-deficient, resulting in the formation of unsaturated alkylated aromatic
residues, which reduce yield and can contribute to catalyst deactivation. In this paper we describe a Cu-modified H-BEA catalyst
that is able to incorporate hydrogen from gas-phase H2 cofed with dimethyl ether into the desired branched alkane products
while maintaining the high C4 and C7 carbon selectivity of the parent H-BEA. This hydrogen incorporation is achieved through
the combination of metallic Cu nanoparticles present on the external surface of the zeolite, which perform H2 activation and
olefin hydrogenation, and Lewis acidic ion-exchanged cationic Cu present within the H-BEA pores, which promotes hydrogen
transfer. With cofed H2, this multifunctional catalyst achieved a 2-fold increase in hydrocarbon productivity in comparison to
H-BEA and shifted selectivity toward products favored by the olefin catalytic cycle over the aromatic catalytic cycle.
KEYWORDS: triptane, dimethyl ether, homologation, zeolite, H-BEA, Cu/H-BEA, olefin catalytic cycle, aromatic catalytic cycle,
hydrogen incorporation
C4 and C7 paraffins and olefins is controlled by the relative rates
of methylation-to-hydrogen transfer, carbocation stability, and
facile β scission of C8+ chains.4 Due to the hydrogen deficiency
of DME homologation to alkane products, unsaturated residues
are also formed (e.g., alkylated aromatics such as hexamethyl-
benzene (HMB)), which decrease the yield of the desired
alkanes and can cause catalyst deactivation.4,6 The stoichio-
metric reaction for DME homologation to form triptane and
HMB is provided in eq 1.
1. INTRODUCTION
Zeolite materials, combining both Brønsted acidity and
confinement effects, possess unique catalytic properties directly
relevant to biomass-to-fuels conversion processes. Specifically,
the C1 chemistry of methanol conversion over acidic zeolites
presents a promising route for the conversion of biomass-
derived synthesis gas into a variety of hydrocarbons: light
olefins, aromatics, branched alkanes, and gasoline-range hydro-
carbons.1−3 The types of hydrocarbons produced are controlled
by the zeolite type and the operating conditions. Of particular
interest is the recent work by Iglesia and co-workers
demonstrating the selective synthesis of triptane (2,2,3-
trimethylbutane) and isobutane by reacting dimethyl ether
(DME) over large-pore acidic zeolites (e.g., H-BEA) at
relatively low temperatures (180−220 °C) and pressures
(60−250 kPa of DME).4−7 The high selectivity to branched
33CH3OCH3 → 6C7H16 + 33H2O + 2C6(CH3)6[HMB]
(1)
Received: July 15, 2014
Revised: January 23, 2015
© XXXX American Chemical Society
1794
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
To address this hydrogen deficiency, Simonetti et al. demon-
strated that alkanes cofed with DME could provide the required
hydrogen when a hydride transfer cocatalyst, adamantane, was
present.6 However, the use of low-volatility adamantane in a
gas-phase industrial process could create additional challenges.
Therefore, we hypothesized that the incorporation of a metal
into/onto the H-BEA zeolite could activate H2 and incorporate
hydrogen into the products with only minimal effect on the
high selectivity to C4 and C7 hydrocarbons. To do so, we
postulated that the metal would have to possess the following
properties: (1) H2 activation at temperatures relevant to DME
homologation, (2) formation of stable ion-exchanged species
within the H-BEA zeolite, (3) kinetically slow alkene hydro-
genation to prevent removal of reactive alkene intermediates
from the chain growth pathway, and (4) alkane dehydrogen-
ation, enabling alkanes to be reincorporated into the chain
growth pathway and indicating that H2 could be activated in
the reverse reaction. For these reasons, first-row transition
metals are an attractive target, especially Cu. The key
objectives of this work were to determine the effect of Cu
modification on the activation and incorporation of H2 during
DME homologation and to elucidate the role of the Cu sites
on H atom management.
Herein, we describe the physical, chemical, and catalytic
properties of a Cu-modified H-BEA catalyst for the selective
homologation of DME to hydrocarbons at 200 °C. The
Cu/BEA material was characterized by X-ray diffraction
(XRD), transmission electron microscopy (TEM), and X-ray
absorption spectroscopy (XAS). The acid sites on Cu/BEA
were investigated using NH3 temperature-programmed desorp-
tion (TPD) and pyridine adsorption diffuse-reflectance FT-IR
spectroscopy (py-DRIFTS) and compared to the parent
H-BEA catalyst. The catalytic properties of Cu/BEA were
tested for DME homologation both with and without H2
cofeed, and the discharged catalysts were analyzed for residual
alkylated aromatics using NMR spectroscopy. Further, H−D
exchange and isobutane dehydrogenation experiments were
performed to investigate the role of ionic versus metallic Cu
species on the observed catalytic reactivity.
recommended by the supplier and converted the BEA from the
ammonium form to the proton form. This catalyst will be
denoted as H-BEA. The Cu-NH4-BEA catalyst was activated as
described for NH4-BEA (denoted ox-Cu/BEA). In some cases,
this material was cooled to 300 °C directly after activation and
exposed to flowing H2 for a minimum of 2 h to reduce Cu
(denoted red-Cu/BEA).
2.2. Catalyst Characterization. 2.2.1. Elemental Analysis.
Elemental analysis of Cu was determined by Galbraith
Laboratories, Inc. (Knoxville, TN) using inductively coupled
plasma atomic emission spectroscopy.
2.2.2. X-ray Diffraction. Powder X-ray diffraction data were
collected using a Rigaku Ultima IV diffractometer with a Cu Kα
source. Diffractograms were collected in the 2θ range of
10−60° at a scan rate of 2°/min. Samples (10−20 mg) were
supported on a glass sample holder with a 0.2 mm recessed
sample area and were pressed into the recession with a glass
slide to obtain a uniform z-axis height. Cu and CuO crystallite
sizes were calculated using the Scherrer equation.
2.2.3. Transmission Electron Microscopy. TEM was per-
formed on red-Cu/BEA on an FEI G20 Tecnai TEM operating
at 200 kV. Following reduction, a sample of the red-Cu/BEA
was dispersed in anhydrous toluene under air-free conditions
and dropped onto a carbon-coated Cu mesh grid. Air exposure
was minimized prior to TEM imaging.
2.2.4. H2 Temperature-Programmed Reduction. Temperature-
programmed reduction (TPR) experiments were carried out
in a U-shaped quartz tube at atmospheric pressure. The re-
actor was loaded with a mixture of 50 mg of Cu-NH4-BEA
catalysts and 300 mg of quartz chips (30−40 mesh). The
catalyst sample was activated by heating to 500 °C at 2 °C/min
and holding at 500 °C for 2 h in flowing 1% O2/He (20 mL/min).
The sample was then cooled to room temperature in flow-
ing He before heating to 600 °C at 10 °C/min in 2% H2/He
(20.4 mL/min). The gas composition was monitored by a mass
spectrometer.
2.2.5. Total Acid Site Titration: NH3 TPD. The total number
of acid sites were determined using NH3 TPD on an Altamira
Instruments AMI-390 system with gas flow rates of 25 mL/min.
Catalyst samples (ca. 250 mg) were loaded into a 1/2 in. quartz
U-tube reactor and held as a fixed bed between plugs of
quartz wool. Samples were calcined in flowing 10% O2/Ar at
2 °C/min to 500 °C and then held at this temperature for 5 h.
For the red-Cu/BEA catalyst, the sample was subsequently
reduced in flowing 10% H2/Ar at 300 °C for 5 h. The samples
were cooled to 120 °C in flowing He and then saturated with
flowing 10% NH3/He for 30 min. Excess and/or physisorbed
NH3 was removed by holding the samples at 120 °C in flowing
He for 1 h. TPD of NH3 was performed by heating the sample
from 120 to 500 °C at 30 °C/min. The temperature was then
held at 500 °C for 30 min to allow any remaining NH3 to
desorb without increasing the temperature above the 500 °C
activation temperature, as recommended elsewhere.8 Desorbed
NH3 was measured with a thermal conductivity detector, and
calibration was performed after each experiment by introducing
five pulses of 10% NH3/He from a 5 mL sample loop into a
stream of flowing He.
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation. Beta zeolite (BEA) having a
SiO2/Al2O3 ratio of 27 was obtained in ammonium form from
Tosoh. The NH4-BEA powder was used as received and had a
particle/agglomerate size range of 45−125 μm. For Cu/BEA
catalysts, an aqueous solution of Cu(NO3)2·2.5H2O (0.961 g in
9 mL of deionized water) was added dropwise to powdered
NH4-BEA (4.95 g) to reach the incipient wetness point. The
slurry was briefly mixed and then dried in an oven at 50 °C
overnight to give Cu-NH4-BEA. The targeted Cu loading was
5 wt %, and the actual measured loading (by elemental analysis)
was 4.3 wt %. The measured weight loading corresponds to
∼60% of the theoretical ion-exchange capacity of the parent
NH4-BEA. A 5% Cu/SiO2 control sample was also prepared by the
analogous incipient wetness method by adding Cu(NO3)2·2.5H2O
(0.92 g in 27 mL of water) to silica (4.75 g, Sipernat-50, Evonik:
particle size 50 μm). The measured Cu loading was 5.3 wt %.
The NH4-BEA and Cu-NH4-BEA catalysts were activated
prior to characterization and catalytic testing. The NH4-BEA
catalyst was activated in a flowing oxidizing mixture (air, 1%
O2/He, or 10% O2/Ar depending upon the experiment) with
heating to 500 °C at 2 °C/min. The catalyst was then held
at 500 °C for a minimum of 5 h. This procedure was
2.2.6. Brønsted and Lewis Acid Site Titration: Pyridine
DRIFTS. The relative amounts of Lewis and Brønsted acid sites
were determined by py-DRIFTS using a Thermo Nicolet iS50
FT-IR spectrometer operating at 4 cm−1 resolution with a
Harrick praying mantis attachment and Si windows operated at
ambient pressure. As-prepared materials were activated in situ
1795
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
at 500 °C in flowing air (50 mL/min, 2 °C/min from ambient,
10 h soak) and then purged with flowing N2 at 150 °C for
30 min. For red-Cu/BEA, the catalyst was reduced in 5% H2
at 300 °C for 4 h. After pretreatment, the sample was held at
150 °C, and saturated pyridine vapor was introduced with N2
for 5 min. The sample was then heated to 300 °C under flowing
N2 (or 5% H2 for red-Cu/BEA) to remove excess and/or
physisorbed pyridine. The absorption peaks at 1545 cm−1
(Brønsted) and 1445 cm−1 (Lewis) and their relative absorption
coefficients (εB/εL = 0.76) were used to determine the relative
Brønsted/Lewis acid site ratios, per common practice.9,10
spectrum (edge energy 8979.0 eV) was acquired simultaneously
with each measurement for energy calibration.
Catalysts were treated in a continuous-flow reactor, which
consisted of a quartz tube (1 in. o.d., 10 in. length) sealed with
Kapton windows by two Ultra-Torr fittings. Ball valves were
welded to each Ultra-Torr fitting and served as the gas inlet and
outlet. An internal K type thermocouple (Omega) was placed
against the catalyst sample holder to monitor temperature.
Catalyst samples were pressed into a cylindrical sample holder
consisting of six wells, forming a self-supporting wafer. The
catalyst loading was predetermined to ensure an absorbance
(μx) of approximately 1.0. The catalysts were treated at the
indicated temperature in air or H2 and cooled to room tem-
perature before obtaining spectra to minimize contributions
due to thermal effects. XANES and EXAFS spectra taken at
high temperature (500 and 300 °C) were similar to those at
room temperature except for the thermal disorder in the
EXAFS. Traces of oxygen and moisture in the He were re-
moved by means of a purifier (Matheson PUR-Gas Triple
Purifier Cartridge).
1
2
2.2.7. Quantification of Hexamethylbenzene: H and H
NMR. The amount of HMB remaining on the catalyst after
reaction was determined using 1H NMR spectroscopy.
1
2
Solution-phase H and H NMR spectra were recorded using
a Varian Inova 400 MHz spectrometer. Post-reaction catalyst
samples (ca. 100 mg) were combined with solvent (0.7 mL of
CDCl3 for 1H; 0.7 mL of CHCl3 for 2H) and internal standard
(10 μL CH2Cl2) in a sealed NMR tube. The tube was shaken
briefly and analyzed for solubilized organics.
2.2.8. H−D Exchange. H−D exchange experiments were
carried out in a U-shaped quartz tube at atmospheric pressure.
Samples of as-prepared materials (ca. 50 mg) were converted to
the proton form in situ by heating at 2 °C/min to 500 °C and
holding for 2 h under flowing 1% O2/He (20 mL/min). For the
red-Cu/BEA catalyst, the sample was then reduced at 300 °C
for 2 h under flowing H2 (20 mL/min, undiluted, 99.999%).
The samples were cooled to room temperature in He and then
heated at 10 °C/min to 800 °C in flowing D2 (Matheson,
99.999%) at 20 mL/min. A mass spectrometer was used to
measure the effluent gas composition.
2.2.9. Isobutane Dehydrogenation. The dehydrogenation
of isobutane was performed in a U-shaped quartz tube at
atmospheric pressure. The as-prepared catalyst (ca. 100 mg)
was mixed with 300 mg of quartz chips (30−40 mesh) and then
loosely packed into the reactor with quartz wool. Catalysts were
activated by heating to 500 °C at 2 °C/min in flowing 1%
O2/He (20 mL/min) and holding at 500 °C for 2 h. For red-
Cu/BEA, the catalyst was reduced at 300 °C for 2 h in flowing
H2 (20 mL/min, undiluted, 99.999%). After pretreatment, the
catalyst sample was heated from room temperature to 600 °C
at 10 °C/min in 1% isobutene/He at a flow rate of 5 mL/min.
A mass spectrometer was used to monitor the reaction
products.
2.2.10. X-ray Absorption Spectroscopy. X-ray absorption
measurements were acquired on the bending magnet beamline
(10-BM) of the Materials Research Collaborative Access
Team (MRCAT) at the Advanced Photon Source, Argonne
National Laboratory. Photon energies were selected using a
water-cooled, double-crystal Si(111) monochromator, which
was detuned by approximately 50% to reduce harmonic reflec-
tions. Measurements were made in transmission mode, and
data points were acquired in three separate regions (energies
relative to the elemental Cu K edge): a pre-edge region (−250
to −30 eV, step size 10 eV, dwell time 0.25 s), the X-ray ab-
sorption near-edge structure (XANES) region (−30 to +30 eV,
step size 0.5 eV, dwell time 0.25 s), and the extended X-ray
absorption fine structure (EXAFS) region (to 13 Å−1, step
size 0.07 Å−1, dwell time 1 s). The ionization chambers were
optimized for the maximum current with linear response
(∼1010 photons detected/s) with 10% absorption (93% N2 and
7% He) in the incident ion chamber and 70% absorption
(50% N2 and 50% Ar) in the transmission detector. A Cu foil
The edge energy of the XANES spectrum was determined
from the inflection point in the leading edge: i.e., the maximum
in the first derivative of the XANES spectrum. Experimental
Cu−O phase shift and backscattering amplitude were obtained
from the reference compound copper(II) acetylacetonate
(Cu(acac)2; 4 Cu−O at 1.92 Å) and used to fit the data.11
Background removal and normalization procedures were
carried out using standard methods. Standard procedures for
normalization and background subtraction were performed
using WinXAS 3.2 software.12 The coordination parameters
were obtained by a least-squares fit in R space of the nearest-
neighbor, k2-weighted Fourier transform data.
2.3. Dimethyl Ether Homologation Catalytic Measure-
ments. Catalyst powders (∼0.6 g) were diluted with low-
surface-area, inert silicon carbide in a 1:6 mass ratio of catalyst
to diluent to prevent channeling, avoid problems with axial
dispersion, and minimize temperature gradients in the bed. The
catalysts were loaded into a stainless steel tubular packed bed
reactor and positioned within the isothermal zone using quartz
chips and quartz wool. A four-point thermocouple positioned
within the catalyst bed was used to monitor reaction
temperature. The reaction temperature during an experiment
was maintained within 0.5 °C of the set point. Prior to reac-
tion rate measurements, the catalysts were activated by ramping
from room temperature to 500 °C at 2 °C/min in flowing dry
air at 150 mL min−1 gcat−1 and holding at 500 °C for 10 h. For
the red-Cu/BEA catalyst, the sample was cooled to 300 °C in
inert gas following air activation at 500 °C, and then exposed to
H2 flowing at 150 mL min−1 gcat−1. The catalyst was held at
300 °C in H2 for 5 h before cooling to the reaction tempera-
ture. A physical mixture of 5% Cu/SiO2 and NH4-BEA (0.3 g of
each material) was also tested (denoted H-BEA-Cu/SiO2) and
was activated using the same procedure as the red-Cu/BEA
catalyst.
All reactions were performed at 200 °C and atmospheric
pressure, and DME conversion was maintained below 20%.
Two feed mixtures were tested: (1) 7.1 mL/min of DME and
4.9 mL/min of Ar (60 mol % DME and 40 mol % Ar) and (2)
7.1 mL/min of DME, 7.1 mL/min of H2, and 1 mL/min of Ar
(47 mol % DME, 47 mol % H2, and 6 mol % Ar). The reaction
conditions without H2 cofeed were similar to those tested by
Ahn et al.4 Single-component gases were fed to the reactor via
calibrated Brooks thermal mass flow controllers. Reactor inlet
1796
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
and outlet gases were sampled through heated (170 °C) lines
to an Agilent 7890 GC instrument equipped with a flame
ionization detector for analysis of oxygenates and hydrocarbons
and two thermal conductivity detectors for analysis of
permanent gases and water. GC responses for reactants and
products were calibrated using Scott Master Class gas standards
and gravimetrically prepared liquid standards. The effluent was
sampled every 45 min. Catalyst performance was evaluated
solely from inlet flow and GC measurements using Ar as an
internal standard. Hydrocarbon production rates were
normalized by the total number of acid sites as determined
by NH3-TPD.
for Cu(0), Cu(I)-SSZ-13 (Cu(I)-zeolite) and Cu2O were used
as Cu(I) references, and Cu(II)-SSZ-13 (Cu(II)-zeolite) and
CuO were used as Cu(II) references.17 The reference spectra
for the standards are provided in Figure S3 in the Supporting
Information. For ox-Cu/BEA, the best fit was a mixture of CuO
and Cu(II)-zeolite with relative concentrations of 45% and
55%, respectively. Inclusion of Cu(I) standards gave similar
quality fits but only amounted to less than 5% Cu(I). For red-
Cu/BEA, there is a high fraction of metallic Cu (∼86%) but a
residual fraction that is unreduced. The best fit suggests that
the unreduced portion is a mixture of Cu(I)-zeolite (4%) and
Cu(II)-zeolite (10%). The reduction of Cu(II) ions within
zeolite pores is reported to occur stepwise, Cu(II) → Cu(I) →
Cu(0), and the temperatures at which these reductions
occur are dependent upon the zeolite type, Cu ion location
within the pore, Si/Al ratio, and ion exchange level of Cu.18−23
For Cu-BEA, the reductions of Cu(II) to Cu(I) and of Cu(I) to
Cu(0) have been reported to occur at 200 and 390 °C,
respectively.23 This finding would suggest that all of the ionic
Cu in the red-Cu/BEA sample should be present as Cu(I);
however, ∼10% of the Cu remained as Cu(II). The presence of
Cu(II) in the red-Cu/BEA sample is likely due to the low
concentration of ion-exchanged Cu (the fraction of Cu present
as Cu(II) in the red-Cu/BEA sample corresponds to 8% of the
ion exchange capacity of the parent NH4-BEA), as it has been
reported that the Cu ion reduction temperature increases with
decreased Cu loading.22,23
3. RESULTS AND DISCUSSION
3.1. Structural and Chemical Characterization of the
Cu Species. The H-BEA and Cu/BEA catalysts were char-
acterized using a suite of bulk- and surface-sensitive techniques
to probe their chemical and structural properties. The bulk
crystal structure of H-BEA matched the known reported
structure for BEA zeolite,13 as confirmed by XRD (Figure S1 in
the Supporting Information). The addition of Cu resulted in
CuO domains (crystallite size of 30 nm) for ox-Cu/BEA and
metallic Cu domains (crystallite size of 43 nm) for red-Cu/BEA.
The formation of large CuO domains upon air oxidation at
500 °C is due to the high Cu loading and the deposition of Cu
via incipient wetness (instead of ion exchange methods). How-
ever, it is likely that a small amount of Cu was ion-exchanged
into the BEA pores during impregnation. The H2 TPR profile
of ox-Cu/BEA showed only one strong peak at approximately
255 °C, attributed to the one-step reduction of CuO to metallic
Cu (Figure S2 in the Supporting Information).14−16 TEM
images confirmed the presence of metallic Cu nanoparticles, as
shown in Figure 1. The nanoparticles were polydisperse with
The k2-weighted magnitude of the Fourier transform of the
red-Cu/BEA sample is shown in Figure 2b along with that of
the Cu foil. In the region typical of Cu−O bonds (i.e., ∼1.7 Å),
there is a small feature; however, this feature is too small to fit
directly since the scattering from metallic Cu is much larger
than that of Cu−O. In order to fit the Cu−O contribution, the
Cu−Cu coordination was first determined (9.5 at 2.54 Å). By
scaling ψ from the Cu foil to the Cu−Cu coordination number
in the catalyst, the metallic Cu scattering was subtracted from
the red-Cu/BEA spectrum, leaving only the Cu−O scattering
contribution. The Cu−O coordination number was determined
to be 1.3 with a bond distance of 1.92 Å (Table S1 in the
Supporting Information). The presence of oxidized Cu and the
low Cu−O coordination number suggest the presence of ion-
exchanged Lewis acidic cationic Cu.
3.2. Acid Site Characterization. Acid site densities were
determined by NH3-TPD (Figure 3a) and are given in Table 1.
The H-BEA TPD profile resembled literature reports,24,25
exhibiting two distinct desorption peaks at 334 and 470 °C,
attributed to weak and strong acid sites, respectively. The
addition of Cu resulted in an increase in the total number of
acid sites and additional features in the TPD profile in com-
parison to H-BEA, specifically a low-temperature desorption
peak at ∼275 °C. The ox-Cu/BEA profile also exhibited a sharp
peak at 342 °C, attributed to the decomposition of a Cu
ammonia complex (i.e., NH3 bonded at Cu cationic species).26,27
The lack of this sharp peak for the red-Cu/BEA sample is in
agreement with the XAS results, showing that only a small
amount of cationic Cu was present for the red-Cu/BEA sample.
Py-DRIFTS was performed to investigate the acidic character
(i.e., Lewis vs Brønsted) of the catalysts.9 Figure 3b presents
the IR spectra between 1400 and 1650 cm−1 for H-BEA,
ox-Cu/BEA, and red-Cu/BEA after adsorption of pyridine at
150 °C followed by purging with N2. The peaks at 1545, 1490,
and ∼1450 cm−1 correspond to pyridine bound to Brønsted
sites, Brønsted and Lewis sites, and Lewis sites, respectively.
Figure 1. TEM images showing the distribution of (a) small
(<10 nm) and (b) larger (>20 nm) metallic Cu particles associated
with red-Cu/BEA.
diameters ranging from 2 to 60 nm. It should be noted that no
effort was made to control the size of these metallic nanoparticles.
Figure 2a shows the XANES spectra collected at the Cu K
edge for the as-synthesized Cu-NH4-BEA in air at room tem-
perature, ox-Cu/BEA, red-Cu/BEA, and Cu foil. The ox-Cu/BEA
sample has a small pre-edge feature characteristic of Cu(II),
while the edge energy of the red-Cu/BEA catalyst is identical
with that of Cu foil. The XANES spectrum of the reduced
sample, however, has a shape slightly different from that of Cu
foil, consistent with a small fraction of oxidized Cu.
The XANES spectra were fit using a linear combination of
Cu(0), Cu(I), and Cu(II) standards to estimate the relative
concentrations of each phase. Cu foil was used as the standard
1797
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
Figure 2. X-ray absorption spectra for as-prepared Cu/BEA following various treatments: (a) XANES and (b) EXAFS. Reference spectra for a Cu
foil are provided for comparison.
Figure 3. Surface acidity characterization for the H-BEA, ox-Cu/BEA, and red-Cu/BEA catalysts using (a) NH3-TPD and (b) py-DRIFTS.
Table 1. Acidity Results from NH3-TPD and py-DRIFTS
NH3 desorption peak temp (°C)
a
b
catalyst
H-BEA
total acid site (μmol/g) Brønsted/Lewis ratio
Brønsted acid site (μmol/g) Lewis acid site (μmol/g)
1
2
3
4
1642
2058
1918
16.8
0.6
1550
772
92
1286
581
334
277
275
470
342
371
c
ox-Cu/BEA
red-Cu/BEA
427
500
c
2.3
1337
500
a
Total acid site density determined from NH3-TPD. The catalyst sample weight after the activation procedure was used to normalize the results.
Brønsted/Lewis acid site ratio was determined from py-DRIFTS. Desorption peak occurred during 30 min hold at 500 °C.
b
c
When the integrated areas of the peaks at 1545 and 1450 cm−1
are compared, the relative ratio of Brønsted to Lewis sites can
be determined (assuming the same extinction coefficient for
py-Al and py-Cu species at 1450 cm−1).9,28−30 The Brønsted/
Lewis ratios for H-BEA, ox-Cu/BEA, and red-Cu/BEA were
16.8, 0.6, and 2.3, respectively. The addition of Cu decreased
the density of Brønsted sites in comparison to H-BEA, due to
conversion of Brønsted sites into Lewis sites by ion-exchanged
Cu and pore blocking by large CuO and Cu domains. A small
peak at 1575 cm−1 was also observed for the Cu-containing
samples (more prominent for ox-Cu/BEA) and has been
attributed to pyridine adsorption to a Cu(II)−OH site.30−32 Cu
provided an additional Lewis site (1608 cm−1), weaker than
tetrahedral Al(III) (1635 cm−1) and comparable in strength to
octahedral Al(III) (1612 cm−1).33 The intensity of this Cu-
based Lewis site at 1608 cm−1 decreased after reduction,
consistent with Brønsted/Lewis ratios determined from peak
area integrations at 1545 and 1450 cm−1. These results
demonstrate the significant increase in Lewis acidity with the
presence of Cu, suggesting Cu(I) or Cu(II) sites, which is in
good agreement with the NH3-TPD and XAS fits.
3.3. Catalytic Testing and Mechanistic Studies.
3.3.1. Dimethyl Ether Homologation. The catalytic perform-
ance of ox-Cu/BEA, red-Cu/BEA, and a physical mixture of
Cu/SiO2 with H-BEA (H-BEA-Cu/SiO2) is compared to that
of the parent H-BEA in Figure 4. The initial C−C bond
formation for DME homologation may be facilitated by trace
amounts of hydrocarbon (i.e., olefin) contaminants in the DME
feed. The C1−C7 hydrocarbon production rate for H-BEA with-
out H2 cofeed was similar to that with H2 cofeed (Figure 4a).
At 1.6 h time on stream (TOS), the H-BEA catalyst exhibited a
hydrocarbon production rate of 827 μmol C molsite−1 s−1. This
value is slightly higher than that reported by Ahn et al. under
similar conditions (740 μmol C molAl s−1 at 1.2 h TOS).4
−1
This difference may be due to slightly different SiO2/Al2O3
ratios of the parent BEA materials and acid site determination
methods (theoretical Al density vs experimental NH3-TPD).
The red-Cu/BEA catalyst without H2 in the feed exhibited a
1798
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
Figure 4. DME homologation rates and product selectivities: (a) hydrocarbon production rates; (b) methanol-free carbon selectivities at 6 h TOS;
(c) C7 paraffin/olefin molar ratios; (d) percent difference in C4−C7 paraffin/olefin molar ratios in comparison to H-BEA at 6 h TOS. For H-BEA-
Cu/SiO2, the hydrocarbon production rate in (a) is based on the number of acid sites present on the H-BEA. The reactions were performed using
0.6 g of catalyst at 200 °C and ∼1 atm and a feed flow rate of either 12 mL/min (60 mol % DME, 40 mol % Ar) for experiments without H2 cofeed
or 15.2 mL/min (47 mol % DME, 47 mol % H2, 6 mol % Ar) for experiments with H2 cofeed.
hydrocarbon production rate slightly lower than that of H-BEA,
while ox-Cu/BEA without H2 showed no detectable hydro-
carbon production. We hypothesize that this lack of activity for
ox-Cu/BEA may be due to Cu ions inhibiting the formation of
surface methoxides or alkoxides by (1) the formation of
Cu−OH species (as discussed in section 3.2) that are
responsible for the observed Brønsted acidity and (2) the
elimination of Brønsted −OH groups via ion exchange.
However, when H2 was included in the feed, the production
rate for red-Cu/BEA was ∼2 times greater than that for H-BEA
on both a gravimetric and per site basis (1689 μmol C molsite−1 s−1
at 1.5 h TOS). A similar increase in performance was also ob-
served for the physical mixture of Cu/SiO2 and H-BEA,
suggesting that metallic Cu may be responsible for the
improved production rate. In further support, the presence of
H2 for ox-Cu/BEA resulted in an increase in hydrocarbon
productivity between 1 and 3 h TOS (achieving rates similar to
red-Cu/BEA), likely due to reduction of CuO to metallic Cu.
The significant increase in productivity for the red-Cu/BEA
catalyst in the presence of H2 appeared to have a negligible
effect on the methanol-free carbon selectivity, as shown in
Figure 4b. As observed by Ahn et al.,4 high selectivity was
achieved to both C4 and C7 hydrocarbons. The C7 selectivities
at 6 h TOS for H-BEA, H-BEA + H2, ox-Cu/BEA + H2, red-
Cu/BEA, red-Cu/BEA + H2, and H-BEA-Cu/SiO2 + H2 were
31%, 30%, 32%, 30%, 32%, and 33%, respectively (error on
measurements is 1%). H2 cofeed also had minimal effect on
the product selectivity of H-BEA. However, Cu-containing cata-
lysts modified the degree of product saturation in comparison
to H-BEA, as shown in Figure 4c,d. The C7 paraffin/olefin
molar ratio over red-Cu/BEA + H2 (and ox-Cu/BEA + H2 after
3 h TOS) was more than 2.5 times greater than that of H-BEA
(Figure 4c). Interestingly, the physical mixture of H-BEA and
Cu/SiO2 behaved similarly to H-BEA, suggesting minimal
hydrogenation of triptene and other C7 olefins by metallic Cu
supported on SiO2. In general, the red-Cu/BEA catalyst with
H2 cofeed yielded products with significantly increased
paraffin/olefin ratios for all carbon numbers in comparison to
H-BEA (Figure 4d and Figures S4 and S5 in the Supporting
Information). The physical mixture of H-BEA and Cu/SiO2 did
increase paraffin/olefin ratios over H-BEA, but only for C2 and
C3 hydrocarbons. Metallic Cu nanoparticles are known to
perform olefin hydrogenation, albeit at a rate much lower than
that for platinum-group metals.34 These results suggest that (1)
metallic Cu performs olefin hydrogenation of ethylene and
propylene primarily, (2) the increased selectivity toward
paraffins for C4−C7 hydrocarbons over red-Cu/BEA + H2 was
due to either close proximity of the metallic Cu to the BEA
zeolite active site and/or the presence of cationic Cu within the
BEA pores, and (3) the increased hydrocarbon production rate
for red-Cu/BEA + H2 and H-BEA-Cu/SiO2 + H2 is related to
the increase in paraffin/olefin ratio for C2 and C3 hydrocarbons.
These preliminary conclusions will be discussed in greater
detail in section 3.4.
Without H2 in the feed, red-Cu/BEA demonstrated a higher
selectivity toward C5−C7 olefins compared to H-BEA (Figure 4c,d
and Figure S4 in the Supporting Information). The paraffin/olefin
ratio over red-Cu/BEA for C5, C6, and C7 hydrocarbons at 6 h
1799
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
Figure 5. H2 production rate and H2/(CH4 + CO + CO2) molar
formation rate ratio as a function of TOS for red-Cu/BEA experiment.
The dashed lines correspond to the maximum H2/(CH4 + CO + CO2)
values achievable on the basis of only methanol decomposition-WGS
or only DME decomposition-WGS. The reaction was performed using
0.6 g of catalyst at 200 °C, ∼1 atm, and a feed flow rate of 12 mL/min
(60 mol % DME, 40 mol % Ar).
Figure 6. Hydrogen (m/z 2) evolution during isobutane dehydrogena-
tion experiments for H-BEA, ox-Cu/BEA, and red-Cu/BEA.
Alkane dehydrogenation requires C−H bond activation (i.e.,
hydrogen abstraction) and recombinative desorption of hydro-
gen atoms to form H2. Simonetti et al. demonstrated that
H-BEA can dehydrogenate branched alkanes (i.e., isobutane)
via hydride-transfer reactions with surface alkoxides at 200 °C.6
However, isobutane dehydrogenation over H-BEA was not
observed until approximately 500 °C (Figure 6). The reported
rate-limiting step for alkane dehydrogenation over zeolites is
hydrogen recombination.37 These results suggest that cationic
Cu species catalyze recombinative hydrogen desorption, similar
to ion-exchanged Ga in Ga/H-ZSM5 during propane
aromatization.37−39 Thus, for DME homologation over red-
Cu/BEA without cofed H2, hydrogen abstracted from branched
alkanes can either be transferred to alkoxides on the zeolite
surface via a bimolecular hydride-transfer step6 or recombine
and desorb on cationic Cu species. As this result supports the
formation of cationic Cu−Hx or CuOx−H species under
reaction conditions, cofed H2 likely dissociates on these cationic
Cu species, resulting in hydrogenation of olefins or hydrogen
transfer to surface alkoxides. This proposed phenomenon
agrees with the trend in paraffin/olefin ratios observed for
red-Cu/BEA + H2 in comparison to H-BEA-Cu/SiO2 + H2
(Figure 4d and Figure S5 in the Supporting Information),
which suggests that cationic Cu species are required to achieve
an increase in C4−C7 paraffin/olefin ratios. The effect of
cationic Cu species on C−H activation still needs to be deter-
mined but is outside the scope of this paper.
3.3.3. H−D Exchange and D2 Incorporation. Due to the
significant increases in productivity and paraffin selectivity for
red-Cu/BEA + H2, we probed the activation of H2 over H-BEA,
ox-Cu/BEA, and red-Cu/BEA using H−D exchange experi-
ments. Figure 7 shows the evolution of HD during isotopic
exchange of D2 with hydrogen atoms over the activated
catalysts. The hydrogen comes from −OH groups on the
catalyst surface, including Brønsted protons from zeolitic-
bridged −OH groups, silanols, or −OH groups at extraframe-
work Al and Cu atoms (e.g., Cu(II)−OH). The red-Cu/BEA
catalyst was capable of activating D2 and evolving HD at a
significantly lower temperature than H-BEA. The maximum
rate of HD evolution was observed at ∼180 °C for red-Cu/BEA
and ∼500 °C for H-BEA. The ox-Cu/BEA sample also exhibited
a low-temperature peak for HD evolution (∼200 °C); however,
D2O production was also observed at this temperature, indicating
that CuO particles were reduced to metallic Cu from the D2
in the feed (in agreement with DME homologation results
TOS decreased by 41%, 31%, and 24% in comparison to
H-BEA, respectively. Additionally, H2 production was observed
during DME homologation over red-Cu/BEA (Figure 5).
No H2 production was observed over H-BEA. The production
of H2 could occur via DME/methanol decomposition over
metallic Cu35 or alkane dehydrogenation over Lewis acidic
cationic Cu sites.36 The decomposition of DME would result in
the formation of H2, CH4, and CO at a stoichiometry of 1:1:1,
whereas the decomposition of methanol would produce H2 and
CO at a stoichiometry of 2:1. However, since metallic Cu is an
active water-gas shift (WGS) catalyst and water is produced as a
product of DME homologation, CO could also undergo WGS
to produce CO2 and H2. Thus, assuming that H2 is being
produced only via decomposition and WGS, the maximum
attainable value for a ratio of the molar formation rates of
H2/(CH4 + CO + CO2) would be 3 for only methanol de-
composition combined with WGS and 1 for only DME
decomposition combined with WGS (corresponding to the
dashed lines in Figure 5). The H2/(CH4 + CO + CO2) molar
formation rate ratio is shown in Figure 5 for red-Cu/BEA, and
the values were greater than 3 throughout the experiment. The
high concentration of DME in the feed and the increase in
CH4 selectivity observed for red-Cu/BEA (Figure 4b; 1.7% for
red-Cu/BEA vs 0.7% for H-BEA) suggest that DME decom-
position likely contributed to more H2 production than methanol
decomposition. These results suggest that, while methanol/DME
decomposition was occurring during DME homologation over
red-Cu/BEA, it was not the only source of H2 production; alkane
dehydrogenation was also occurring.
3.3.2. Isobutane Dehydrogenation. Temperature-programmed
reactions with isobutane were performed to test the dehydro-
genation activity of the H-BEA, ox-Cu/BEA, and red-Cu/BEA
catalysts, and the results are provided in Figure 6. H2 evolution
over H-BEA did not occur until ∼500 °C, whereas it was
observed at 275−300 °C for ox-Cu/BEA and red-Cu/BEA.
The only products observed were H2 and isobutylene. The sim-
ilarity of the H2 evolution onset temperatures for ox-Cu/BEA
and red-Cu/BEA suggests that the low-temperature dehydro-
genation activity is most likely attributed to cationic Cu species
(potentially in combination with Brønsted acid sites), not
metallic Cu sites.
1800
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
Figure 7. HD (m/z 3) evolution during H−D exchange experiments
for H-BEA, ox-Cu/BEA, and red-Cu/BEA.
showing an increase in hydrocarbon productivity after 3 h
TOS). The total amounts of HD evolved (i.e., the area under
the curve) were nearly identical for H-BEA and red-Cu/BEA,
suggesting that the D2 activated by the Cu can reach all of
the H atoms present within the zeolite. The total amount of
HD evolved for ox-Cu/BEA was approximately 65% of that for
H-BEA. Proton mobility and H2 spillover on aluminosilicate
surfaces is a debated topic in the literature.40 It has been
reported that acidic protons within zeolites can be mobile at
elevated temperatures, and their mobility may be enhanced by
the presence of H2O.41−43 The spillover of H2 onto a defect-
free zeolite surface is believed to be energetically impossible;
however, H2 spillover onto zeolite surfaces with defects or
carbonaceous deposits may be possible.40 Recently, a study
by Im et al. demonstrated experimentally and theoretically
that H2 spillover can occur on an aluminosilicate surface at
DME homologation temperatures via a defect generation−
annihilation pathway.44 Although H−D exchange is not direct
proof of H2 spillover, these results indicate that activated
hydrogen is available during DME homologation over
red-Cu/BEA for hydrogen transfer or hydrogenation reactions
that may occur at defect sites, cationic Cu sites, and metallic Cu
sites or on carbonaceous deposits.
The activation and incorporation of H2 was further explored
through isotopically labeled experiments. D2 was cofed with
DME over H-BEA and red-Cu/BEA, and the incorporation of
deuterium into the hydrocarbon products was monitored using
mass spectrometry. The mass spectra for triptane collected at
6 h TOS are shown in Figure 8. When H2 is replaced with D2
over H-BEA, there is no change in the resulting mass spectrum
for triptane. In contrast, over red-Cu/BEA, significant in-
corporation of deuterium into the desired product was ob-
served. This deuterium incorporation was observed for all
representative products, including HMB (Figures S6−S8 in the
Supporting Information).
Figure 8. Mass spectra for triptane collected at 6 h TOS from
(a) H-BEA + H2, (b) H-BEA + D2, and (c) red-Cu/BEA + D2
experiments. The reactions were performed using 0.6 g of catalyst at
200 °C, ∼1 atm, and a feed flow rate of 15.2 mL/min (47 mol %
DME, 47 mol % H2 or D2, 6 mol % Ar).
the relative propagation of the olefin and aromatic cycles can
be determined by calculating the ratio of ethylene to
2-methylbutane and 2-methylbutene (2MB) in the products.3,46
Accordingly, we determined the ratios of the molar formation
rates of ethane and ethylene to 2MB (rC /r2MB) and the ratios
2
of carbon present in HMB to carbon present in hydrocarbons
(CHMB/CHC) to probe the effect of H2 introduction on (1) the
relative rates of production of hydrocarbons from the aromatic
and olefin cycles and (2) alkyl-aromatic formation via the olefin
catalytic cycle. Table 2 provides these ratios for the catalysts
with and without cofed H2. Ethane was included with ethylene
Table 2. Effect of Cu Addition on Product Selectivity
a
b
c
d
experiment
H-BEA
H-BEA + H2
rC /r2MB
rC /r2MB
CHMB/CHC
2
2
0.23
0.21
0.08
0.24
0.08
0.08
0.16
0.15
0.07
0.24
0.08
0.09
0.12
0.07
0.03
0.14
0.04
0.03
ox-Cu/BEA + H2
red-Cu/BEA
red-Cu/BEA + H2
H-BEA-Cu/SiO2 + H2
a
The reactions were performed using 0.6 g of catalyst at 200 °C,
3.4. Mechanistic Interpretations. As discussed earlier, the
direct homologation of DME to alkanes is hydrogen deficient.
The necessary hydrogen atoms are provided by the formation
of alkylated aromatic byproducts, such as HMB. For the red-
Cu/BEA catalyst, hydrogen atoms are provided by the
activation of gaseous H2 and are incorporated into the hydro-
carbon products, suggesting that the formation of alkylated
aromatics may be reduced. It has been reported that the
conversion of methanol to hydrocarbons occurs via a dual
catalytic cycle of olefin and aromatic methylation3,45,46 and that
∼1 atm, and a feed flow rate of either 12 mL/min (60 mol % DME,
40 mol % Ar) for experiments without H2 cofeed or 15.2 mL/min
(47 mol % DME, 47 mol % H2, 6 mol % Ar) for experiments with H2
b
cofeed. Ratio of molar formation rate of ethane and ethylene to molar
formation rate of 2-methylbutane and 2-methylbutenes at 6 h TOS.
c
Ratio of molar formation rate of ethane and ethylene to molar
formation rate of 2-methylbutane and 2-methylbutenes at a DME
d
conversion of 11% (different TOS). Ratio of total carbon present in
HMB to total carbon in C1−C7 hydrocarbons produced over an entire
experiment.
1801
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
Research Article
due to probable olefin hydrogenation over metallic Cu. Cal-
culated either at the same TOS or at the same conversion,
red-Cu/BEA + H2 and H-BEA-Cu/SiO2 + H2 exhibited lower
H2, it is proposed that these cationic Cu species activate H2
and either hydrogenate olefins or transfer hydrogen to surface
alkoxide species to produce alkanes, as evidenced by the
increase in C4−C7 paraffin/olefin ratios for red-Cu/BEA + H2.
The relative rates of dehydrogenation and hydrogenation/
hydrogen transfer over the cationic Cu species as a function of
H2 concentration in the feed is an area of future investigation.
The incorporation of gas-phase D2 into hydrocarbon
products over red-Cu/BEA was observed; however, the exact
mechanism for this incorporation is unclear. The incorpora-
tion could be explained by one of, or a combination of, the
following: (1) a D atom on the metallic Cu or cationic Cu
exchanges with the H atom of a hydroxyl on the BEA surface
near the perimeter of the Cu particle/ion, followed by
subsequent OH−OD exchange on the zeolite surface,40 (2)
olefin hydrogenation with D2 activated over metallic Cu and/or
cationic Cu species, and the formed alkane undergoing
hydrogen abstraction and intermolecular hydride transfer to a
surface alkoxide within the zeolite, (3) H−D scrambling for
DME, methanol, water, and/or alkanes on the metallic Cu
surface, followed by desorption and transport into the zeolite,
and (4) deuterium spillover from metallic Cu sites to defect
sites or carbonaceous species on the BEA zeolite, followed
by transport of the D (possibly by a defect generation−
annihilation pathway44) to a cationic Cu site or other defect site
capable of performing hydrogenation/hydrogen transfer.
Research to further elucidate the mechanism(s) of H−D
exchange and hydrogen incorporation over red-Cu/BEA and its
exact role in the catalytic process is ongoing.
rC /r2MB ratios than H-BEA, suggesting that the addition of
2
metallic Cu promoted propagation of the olefin cycle. This shift
was further confirmed by the CHMB/CHC ratios. Red-Cu/
BEA + H2 exhibited a CHMB/CHC ratio of 0.04 in comparison to
that of 0.12 for H-BEA. The decreased HMB production for
red-Cu/BEA + H2 is not likely due to hydrogenation of HMB
on metallic Cu, as supported metallic Cu catalysts have been
reported to be inactive for aromatic ring hydrogenation under
similar reaction conditions.47,48 As discussed previously, the ox-
Cu/BEA + H2 behaved similarly to red-Cu/BEA + H2 as the
cofed H2 resulted in the reduction of CuO to metallic Cu under
reaction conditions. The red-Cu/BEA without cofed H2
exhibited the highest values for rC /r2MB and CHMB/CHC ratios,
2
consistent with previous results showing its increased selectivity
toward olefins via dehydrogenation. Although red-Cu/BEA +
H2 exhibited a decreased selectivity toward HMB, its de-
activation profile was similar to that of H-BEA (Figure S9 in the
Supporting Information), suggesting that carbon deposition
and/or site blocking is still occurring at a similar rate despite
enhanced hydrogenation rates and increased selectivity to
hydrocarbons over HMB. Research is ongoing to identify the
primary cause of deactivation during the early stages of DME
homologation.
Considering the above results, the red-Cu/BEA catalyst
possesses both metallic Cu sites and Lewis acidic cationic
Cu sites, and both types of sites appear necessary to achieve
increases in hydrocarbon productivity, increases in paraffin/
olefin ratios, and decreases in selectivity toward aromatics.
On the basis of TEM, a large fraction of the metallic Cu
nanoparticles were larger than the diameter of the H-BEA pores
and thus were located on the external surface of the zeolite.
Given the comparison to Cu/SiO2, the metallic Cu sites
activate H2 at low temperature, decompose some DME into H2,
CO, and CH4, and hydrogenate ethylene and propylene.
Although C2 and C3 hydrocarbons accounted for less than 10%
of the carbon in the products, the hydrogenation of ethylene
and propylene over metallic Cu sites appeared to be directly
responsible for both increased hydrocarbon productivity and
decreased selectivity toward aromatics, i.e. HMB. We propose
that the hydrogenation of ethylene and propylene prevented
their entry into the aromatic catalytic cycle via aromatization,
and thus caused the observed changes in hydrocarbon
productivity and aromatic selectivity.
The cationic Cu sites had low Cu−O coordination numbers
and their XAS spectra resembled those for other Cu ion-
exchanged zeolites,17 suggesting that the cationic Cu of the red-
Cu/BEA catalyst was ion-exchanged within the pores of the
zeolite. These cationic Cu species clearly introduced significant
Lewis acidity on the basis of py-DRIFTS, and these sites
catalyzed alkane dehydrogenation (i.e., hydrogen recombinative
desorption). These findings suggest the formation of cationic
Cu−Hx or CuOx−H species under reaction conditions that can,
either independently or in conjunction with a Brønsted acid
site, activate C−H bonds, abstract hydrogen, and perform
recombinative hydrogen desorption. The dehydrogenation
activity of these Cu species suggests that they can facilitate
reincorporation of “terminal” alkanes into the chain growth
pool during DME homologation via direct dehydrogenation or
hydride abstraction and intermolecular hydrogen transfer
between an alkane and a surface alkoxide. In the presence of
4. CONCLUSION
A Cu-modified BEA zeolite, possessing multiple catalytic
functionalities (metallic Cu sites, Lewis acidic cationic Cu
sites, Brønsted acidity), was demonstrated to incorporate gas-
phase H2 into the desired products, to increase hydrocarbon
productivity per site, to increase paraffin selectivity, and to
decrease aromatic selectivity while maintaining high selectivity
to C4 and C7 hydrocarbons. The metallic Cu sites performed
H2 activation and ethylene/propylene hydrogenation, while the
cationic Cu sites facilitated hydrogen transfer and hydro-
genation of C4−C7 olefins produced via the olefin catalytic
cycle. These two effects favor the production of hydrocarbons
via the olefin pathway over the aromatic pathway.
These findings suggest that H2 incorporation during DME
homologation over a Cu-modified zeolite can achieve high
productivity to high-value gasoline-range alkanes, with lower
aromatic formation. The combination of metallic and Lewis
acidic functionality with Brønsted acid zeolite chemistry
presents a pathway toward designing optimized catalysts for
homologation reactions in which hydrogen incorporation is
desired. Future work will focus on tuning performance by
balancing the relative rates of each of the reaction steps.
ASSOCIATED CONTENT
■
S
* Supporting Information
The following file is available free of charge on the ACS
Publications website at DOI: 10.1021/cs501876w.
XRD patterns (Figure S1), H2 TPR profile for Cu-NH4-
BEA (Figure S2), XANES reference spectra (Figure S3),
paraffin/olefin molar ratios as a function of TOS (Figure
S4), percent difference in C2−C3 paraffin/olefin molar
ratios in comparison to H-BEA at 6 h TOS (Figure S5),
1802
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803

ACS Catalysis
mass spectra for propylene and isobutane collected
Research Article
(23) Kwak, J. H.; Tran, D.; Burton, S. D.; Szanyi, J.; Lee, J. H.; Peden,
C. H. F. J. Catal. 2012, 287, 203−209.
during deuterium incorporation experiments (Figures S6
and S7), 2H NMR spectra for spent catalysts (Figure S8),
deactivation profiles (Figure S9), and EXAFS fitting
results (Table S1) (PDF)
(24) Wang, H.; Xin, W. Catal. Lett. 2001, 76, 225−229.
(25) Barthel, N.; Finiels, A.; Moreau, C.; Jacquot, R.; Spagnol, M. J.
Mol. Catal. A: Chem. 2001, 169, 163−169.
(26) Salker, A. V.; Weisweiler, W. Appl. Catal., A 2000, 203, 221−
229.
(27) Putluru, S. S. R.; Riisager, A.; Fehrmann, R. Appl. Catal., B 2011,
AUTHOR INFORMATION
■
101, 183−188.
Corresponding Author
*E-mail for J.A.S.: Joshua.Schaidle@nrel.gov.
(28) Hughes, T. R.; White, H. M. J. Phys. Chem. 1967, 71, 2192−
2201.
(29) Amin, N. A. S.; Anggoro, D. D. J. Nat. Gas Chem. 2003, 12,
123−134.
Notes
The authors declare no competing financial interest.
(30) Zaki, M. I.; Hasan, M. A.; Al-Sagheer, F. A.; Pasupulety, L.
Colloids Surf., A 2001, 190, 261−274.
ACKNOWLEDGMENTS
(31) Parry, E. J. Catal. 1963, 2, 371−379.
■
(32) Bezrodna, T.; Puchkovska, G.; Shimanovska, V.;
Chashechnikova, I.; Khalyavka, T.; Baran, J. Appl. Surf. Sci. 2003,
214, 222−231.
This work was supported by the Laboratory Directed Research
and Development Program at the National Renewable Energy
Laboratory and the Department of Energy’s Bioenergy
Technology Office under Contract no. DE-AC36-08-
GO28308. X-ray Absorption Spectroscopy was performed at
the Advanced Photon Source at Argonne National Laboratory
with the assistance of members of the Materials Research
Collaborative Access Team (supported by the Department of
Energy and the MRCAT member institutions).
(33) Zaki, M. I.; Hasan, M. A.; Pasupulety, L. Langmuir 2001, 17,
768−774.
(34) Best, R. J.; Russell, W. W. J. Am. Chem. Soc. 1954, 76, 838−842.
(35) Fisher, I. A.; Bell, A. T. J. Catal. 1999, 184, 357−376.
(36) Sievers, C.; Onda, A.; Guzman, A.; Otillinger, K. S.; Olindo, R.;
Lercher, J. A. J. Phys. Chem. C 2006, 111, 210−218.
(37) Biscardi, J. A.; Iglesia, E. Catal. Today 1996, 31, 207−231.
(38) Meitzner, G. D.; Iglesia, E.; Baumgartner, J. E.; Huang, E. S. J.
Catal. 1993, 140, 209−225.
REFERENCES
■
(39) Krishnamurthy, G.; Bhan, A.; Delgass, W. N. J. Catal. 2010, 271,
370−385.
(1) Chang, C. D.; Silvestri, A. J. J. Catal. 1977, 47, 249−259.
(2) Haw, J. F.; Song, W.; Marcus, D. M.; Nicholas, J. B. Acc. Chem.
Res. 2003, 36, 317−326.
(40) Prins, R. Chem. Rev. 2012, 112, 2714−2738.
(41) Sarv, P.; Tuherm, T.; Lippmaa, E.; Keskinen, K.; Root, A. J. Phys.
Chem. 1995, 99, 13763−13768.
(3) Ilias, S.; Bhan, A. ACS Catal. 2012, 3, 18−31.
(4) Ahn, J. H.; Temel, B.; Iglesia, E. Angew. Chem., Int. Ed. 2009, 48,
3814−3816.
(42) Ryder, J. A.; Chakraborty, A. K.; Bell, A. T. J. Phys. Chem. B
2000, 104, 6998−7011.
(5) Simonetti, D. A.; Ahn, J. H.; Iglesia, E. J. Catal. 2011, 277, 173−
195.
(43) Sierka, M.; Sauer, J. J. Phys. Chem. B 2001, 105, 1603−1613.
(44) Im, J.; Shin, H.; Jang, H.; Kim, H.; Choi, M. Nat. Commun.
2014, 5, 3370.
(6) Simonetti, D. A.; Ahn, J. H.; Iglesia, E. ChemCatChem 2011, 3,
704−718.
(45) Ilias, S.; Bhan, A. J. Catal. 2012, 290, 186−192.
(46) Ilias, S.; Khare, R.; Malek, A.; Bhan, A. J. Catal. 2013, 303, 135−
140.
(47) Emmett, P.; Skau, N. J. Am. Chem. Soc. 1943, 65, 1029−1035.
(48) Saadi, A.; Rassoul, Z.; Bettahar, M. J. Mol. Catal. A: Chem. 2000,
164, 205−216.
(7) Hazari, N.; Iglesia, E.; Labinger, J. A.; Simonetti, D. A. Acc. Chem.
Res. 2012, 45, 653−662.
(8) Kenvin, J. MicroReport: Micromeritics Instrument Corporation
2004, 15, 8−10.
(9) Emeis, C. A. J. Catal. 1993, 141, 347−354.
(10) Maier, S. M.; Jentys, A.; Lercher, J. A. J. Phys. Chem. C 2011,
115, 8005−8013.
(11) Golchoubian, H. Asian J. Chem. 2008, 20, 5834−5838.
(12) Ressler, T. J. Synchrotron Radiat. 1998, 5, 118−122.
(13) Newsam, J.; Treacy, M. M.; Koetsier, W.; De Gruyter, C. Proc.
R. Soc. A 1988, 420, 375−405.
(14) Richter, M.; Fait, M. J. G.; Eckelt, R.; Schreier, E.; Schneider,
M.; Pohl, M. M.; Fricke, R. Appl. Catal., B 2007, 73, 269−281.
́
(15) Szegedi, A.; Kon
Horvath, Z. E.; Biro,
266.
́ ́ ́ ́ ́
ya, Z.; Mehn, D.; Solymar, E.; Pal-Borbely, G.;
́
́
L. P.; Kiricsi, I. Appl. Catal., A 2004, 272, 257−
(16) Kefirov, R.; Penkova, A.; Hadjiivanov, K.; Dzwigaj, S.; Che, M.
Microporous Mesoporous Mater. 2008, 116, 180−187.
(17) Bates, S. A.; Verma, A. A.; Paolucci, C.; Parekh, A. A.; Anggara,
T.; Yezerets, A.; Schneider, W. F.; Miller, J. T.; Delgass, W. N.;
Ribeiro, F. H. J. Catal. 2014, 312, 87−97.
(18) Yan, J.; Lei, G.-D.; Sachtler, W.; Kung, H. J. Catal. 1996, 161,
43−54.
(19) Yan, J. Y.; Sachtler, W. M. H.; Kung, H. H. Catal. Today 1997,
33, 279−290.
(20) Kieger, S.; Delahay, G.; Coq, B.; Neveu, B. J. Catal. 1999, 183,
267−280.
(21) Ded
Mater. 1999, 32, 63−74.
(22) Bulanek, R.; Wichterlova,
2001, 31, 13−25.
̌ ̌
ecek, J.; Wichterlova, B.; Kubat, P. Microporous Mesoporous
́
́
́
B.; Sobalík, Z.; Tichy, J. Appl. Catal., B
1803
DOI: 10.1021/cs501876w
ACS Catal. 2015, 5, 1794−1803