Effects of Li doping on MgO-supported Sm2O3 and TbOx catalysts in the oxidative coupling of methane

Research Article

pubs.acs.org/acscatalysis

Eﬀects of Li Doping on MgO-Supported Sm₂O₃and TbO_xCatalysts in

the Oxidative Coupling of Methane

Trenton W. Elkins,^†Bjorn Neumann,^‡Marcus Baumer,^‡and Helena E. Hagelin-Weaver*^,†

̈

^†Department of Chemical Engineering, University of Florida, Gainesville, Florida 32611, United States

^‡Institute of Applied and Physical Chemistry & Center for Environmental Research and Sustainable Technology, University Bremen,

Leobener Strasse UFT, 28359 Bremen, Germany

ABSTRACT: Rare earth oxides (REOs), particularly the sesqui-

oxides, such as Sm₂O₃and La₂O₃, have been investigated as

promising catalysts in the oxidative coupling of methane (OCM).

Much less attention has been paid to the reducible REOs because

they are expected to give oxidation products, such as CO and CO₂

(CO_x), rather than the desirable ethane and ethylene (C₂₊). Because

Li addition can improve the performance of Sm₂O₃in the OCM

reaction and Li/MgO is commonly used as a reference OCM

catalyst, the eﬀects of lithium addition to a reducible oxide, TbO_x,

were investigated in detail in this study and compared with a Sm₂O₃

catalyst, which is the best single component OCM catalyst. Because

of the well-documented volatility of lithium under OCM conditions, particularly for the Li/MgO system, the stability of lithium-

doped samaria and terbia catalysts was examined as a function of preparation methods in this study. As expected, terbia

supported on nanoparticle magnesia (n-MgO) is not a very active or selective OCM catalyst, and most of the observed selectivity

toward C₂₊products is likely due to the n-MgO support. In contrast, Li-doped TbO_x/n-MgO prepared using a coimpregnation

method yields a highly active and selective catalyst. The Li-TbO_x/n-MgO catalyst yields the same methane conversion as pure

Sm₂O₃, and has a higher C₂₊selectivity than the Li-Sm₂O₃/n-MgO catalyst. The stability of the Li-TbO_x/n-MgO catalyst is also

higher than that of the Li-Sm₂O₃/n-MgO catalyst, and the loss of activity for the lithium-doped terbia catalyst appears to be the

same as for the undoped Sm₂O₃/n-MgO catalyst (and undoped TbO_x/n-MgO). The characterization data indicate stronger

interactions between Li and TbO_xthan between Li and Sm₂O₃, which may explain the higher stability of the Li-TbO_x/n-MgO

catalysts. There are also indications that Li enters the TbO_xlattice and reduces TbO_1.81, to Tb₂O₃during reaction, which can

explain the higher C₂₊selectivity compared with undoped TbO_x/n-MgO. Furthermore, the Li-TbO_x/n-MgO catalyst in this

study is active at lower temperatures (600−700 °C) than typically used in the OCM (around 800 °C). Therefore, the Li-TbO_x/

n-MgO catalysts have potential to be very eﬀective OCM catalysts, even though undoped TbO_x/n-MgO catalysts are more

selective toward CO_xthan C₂₊products.

KEYWORDS: oxidative coupling of methane, samarium oxide, terbium oxide, lithium doping, magnesium oxide support, XPS, XRD,

DRIFTS

lithium in the magnesium oxide lattice together with the

volatility of LiOH at the high reaction temperature.⁶

1. INTRODUCTION

The oxidative coupling of methane (OCM) has been

extensively studied over the past few decades since the initial

work of Keller and Bhasin.¹Ethylene, one of the two C₂

hydrocarbon products in OCM, is a highly valuable feedstock

material for synthesizing longer chain hydrocarbons, polymers,

and other products. Early work in the OCM ﬁeld found that

doping a magnesium oxide support with lithium produces an

active catalyst under OCM conditions with a relatively high C₂

product selectivity.^2,3This catalyst system has since been a

reference system in many studies due to its activity, selectivity,

and simplicity. However, lithium is volatile under OCM

reaction conditions, which is detrimental to the catalyst’s

long-term stability, and it has been found that within the ﬁrst

few hours of reaction, the majority of the lithium will leave the

surface of the catalyst, resulting in a sharp decrease in activity.^4,5

The underlying issue has been attributed to the low solubility of

In addition to Li-MgO, a large number of catalysts with

diﬀerent compositions have been investigated in the oxidative

coupling of methane.⁷⁻³¹The best-performing catalyst to date

is the complex, multicomponent Mn-Na₂WO₄/SiO₂catalyst

because of its high activity and long-term stability in multiple

reactor types.²⁸⁻³¹However, this catalyst requires temperatures

in excess of 800 °C for optimum activity and selectivity.³¹

Among the highest performing single-component OCM

catalysts are rare-earth oxides (REOs) in the lanthanide series,

particularly Sm₂O₃.¹⁶⁻²⁴In general, sesquioxides, such as

Sm₂O₃and La₂O₃, which do not form higher oxides, are the

Received: January 31, 2014

Revised: April 21, 2014

Published: April 25, 2014

1972

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

most eﬀective in selectively promoting CH₄coupling to ethane

and ethylene,²²and reducible REOs (e.g. CeO₂, Tb₄O₇,

Pr₆O₁₁) are selective only after doping with, for example, alkali

metals.^27,32Addition of alkali and alkaline-earth metals to

Sm₂O₃, and other nonreducible REOs, such as La₂O₃, has

proven to be beneﬁcial to the system’s C₂yield,²⁶but the

activity (i.e. CH₄conversion) can be negatively inﬂuenced

under certain reaction conditions.^26,33Furthermore, deactiva-

tion issues similar to the Li/MgO have been noted, particularly

for the Li-doped catalysts. Although reducible REOs doped

with alkali and alkaline earth metals have appeared in the

literature, no systematic investigation is available on how the

lithium inﬂuences the REO catalyst and if the eﬀects are

dependent on the speciﬁc REO used. Furthermore, very few

studies using terbium oxide as the active component are

available in the literature,^20,26,27even though there are

indications that terbia-based OCM catalysts can beneﬁt greatly

from alkali doping.²⁷

For these reasons, it was decided to investigate the eﬀects of

Li doping on both Sm₂O₃and TbO_xto determine if the eﬀects

of Li are diﬀerent over the two catalysts and if it is possible to

prepare a highly active TbO_x-based catalyst with activity below

800 °C. Because the activity and selectivity can be dependent

on catalyst preparation methods and catalyst precursors, three

diﬀerent preparation methods were examined: a micro-

emulsion-assisted, precipitation-deposition method and an

incipient wetness impregnation method in which the lithium

was added either with the rare earth oxide precursor

(coimpregnation) or after the impregnation with the rare

earth oxide (sequential impregnation). The eﬀect of the lithium

precursor was also explored using lithium nitrate and lithium

chloride. Although we have shown previously that a 20%

samaria loading on an alumina support can be an active catalyst

in OCM,²⁵the focus in this study was on a more conventional

OCM support, that is, magnesia, to enhance the C₂₊yield

compared with the alumina-supported catalysts. Furthermore,

this choice of support facilitates comparisons with Li/MgO,

which is a commonly used OCM catalyst.

REO/n-MgO for undoped and lithium-doped catalysts,

respectively.

2.1.2. Rare-Earth Oxide Catalysts via Microemulsion

Method. Samaria- and terbia-based catalysts were also

synthesized via a reverse micelle, water in toluene, micro-

emulsion method using n-MgO as the support material. The

procedure has been implemented in previous research using an

aluminum oxide support.²⁵Brieﬂy, the support was suspended

in 22.5 mL of DI H₂O, and the REO precursor salt was added

to the aqueous solution and allowed to stir for 15 min. The

aqueous phase was then added to 187 mL of toluene dropwise,

and afterward, a surfactant, Tween 80 (Acros Organics), was

added to the mixture dropwise (∼4 mL) until emulsion

formation. An aqueous phase of NaOH (50% stoichiometric

excess) was added dropwise to the system to produce the rare-

earth hydroxide in the aqueous phase inside the reverse

micelles. The resulting mixture was allowed to stir overnight

and then was centrifuged to collect the product, which was

washed in a water/ethanol (50/50 vol %) solution to clean the

particles. The catalysts were then dried at 105 °C overnight and

calcined at 800 °C for 4 h.

2.1.3. Li/n-MgO Reference Systems Preparation. Because

Li/MgO catalysts have been studied in the OCM reaction, two

2.5/97.5 Li/n-MgO reference catalysts were included in the

study for comparison. One catalyst, {Li(N)/n-MgO}, was

prepared using the lithium nitrate salt precursor and the same

method (IWI) used to prepare the Li-doped Sm₂O₃/n-MgO

and TbO_x/n-MgO catalysts. The second, {Li(C)/n-MgO}, was

prepared by following a slurry method previously reported by

Lunsford and co-workers.³This method involves slowly adding

Li₂CO₃(99.99%, Alfa Aesar) to an aqueous magnesium oxide

slurry, which was then stirred under gentle heating until a paste

formed. The resulting mixture was dried overnight and then

calcined at 800 °C for 4 h.

2.2. Catalyst Characterization Methods. 2.2.1. Surface

Area Analysis. Brunauer−Emmett−Teller (BET) surface area

measurements were performed on all catalysts using a 6-point

isotherm on a Quantachrome NOVA 1200 instrument

operating at liquid nitrogen temperatures. All correlation values

were >0.999.

2. EXPERIMENTAL SECTION

2.2.2. Electron Microscopy. Transmission electron micros-

copy (TEM), including energy dispersive X-ray spectroscopy

(EDX), was performed using an FEI Tecnai F20 S-TWIN

microscope with an operating voltage of 200 kV. TEM images

were acquired with a slow-scan CCD camera with an integrated

Gatan image ﬁlter, model 2001. For the preparation of the

TEM grids, a small amount of each sample was suspended in

acetone and ultrasonicated. Finally, a droplet (25 μL) was

placed on a carbon-coated copper grid.

2.2.3. X-ray Diﬀraction Analysis. Powder X-ray diﬀraction

(PXRD) data were acquired for the catalysts using a

PANalytical X’Pert MPD Pro l diﬀractometer in Bragg−

Brentano geometry. The setup was equipped with a secondary

Ni ﬁlter, Cu Kα1,2 radiation and a X’Celerator multistrip

detector.

2.2.4. X-ray Photoelectron Spectroscopy. Most XPS

measurements were collected on a PerkinElmer 5100 XPS

system using an aluminum X-ray source (fresh samples). A thin

layer of catalyst was placed on double-sided carbon conductive

tape for sample preparation. In all cases, samples were dried in

an oven overnight at 105 °C to minimize the water adsorption

on the catalyst surface. On the PerkinElmer 5100 XPS system,

survey scans were collected using a time/step of 30 ms with a

2.1. Catalyst Preparation. 2.1.1. Rare-Earth Oxide-Based

Catalysts via Incipient Wetness Impregnation. Samaria- and

terbia-based catalysts were synthesized via incipient wetness

impregnation (IWI) onto a low-surface-area (20.4 m²/g)

nanoparticle magnesium oxide support (n-MgO received

from NanoScale Materials Inc.³⁴) using an aqueous solution

of the REO precursor (samarium(III) nitrate hydrate or

terbium(III) nitrate hydrate, 99.9% REO, Alfa Aesar). Brieﬂy,

an appropriate amount of the REO precursor was dissolved into

a volume of DI water equal to the volume of the pores of the

support. The solution was then added dropwise to the support

under continuous stirring until incipient wetness. The resulting

catalysts were dried at 80 °C for 3 h and overnight at 105 °C

before calcination at 800 °C for 4 h. A lithium nitrate (LiNO₃

hydrate, Puratronic, 99.999% metal basis) or lithium chloride

(LiCl₃hydrate, Puratronic, 99.998%) precursor was also added

to the REO precursor solution for preparation of coimpreg-

nated lithium-doped samples. For the sequential impregnation,

the REO precursor was ﬁrst added to the support, and the

lithium precursor was added after drying. The weight

percentage of the REOs was always 20% (metal oxide basis),

and the lithium loading was 2.5%, resulting in catalyst

compositions of 20/80 REO/n-MgO and 2.5/20/77.5 Li-

1973

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

0.5 eV step size and a pass energy of 89.45 eV, and the time/

step was 50 ms with a step size of 0.1 eV and a pass energy of

35.75 eV for the narrow scans. Survey and narrow scan spectra

were collected using a total of 10 and 50 scans, respectively.

Because the spent catalysts were problematic in the ultrahigh

vacuum environment as a result of excessive outgassing, one

spent and one fresh catalyst were analyzed using a diﬀerent XPS

instrument with a lower resolution. This instrument is

equipped with a double-pass, cylindrical-mirror analyzer (PHI

model 25-270AR) and a Mg Kα X-ray source (PHI 04-151).

Data collection was accomplished using a computer-interfaced,

digital pulse-counting circuit³⁵followed by smoothing with

digital ﬁltering techniques. Spectra were taken in the retarding

mode with a pass energy of 50 eV for survey spectra and 25 eV

for narrow scans on this instrument. For the C 1s, including the

Li 1s region, 25 scans were collected; for the Tb 4d peaks, a

total of 100 scans were collected to get a suﬃcient signal-to-

noise. Samples for this system were pressed into aluminum

cups before insertion into the ultrahigh vacuum chamber.

Charge shift corrections were made on all samples by assuming

a C 1s signal of 284.6 eV.³⁶No signs of diﬀerential charging

were observed in this study.

To obtain a more quantitative estimate of the contribution

from diﬀerent oxygen species on the surface of these catalysts,

the O 1s peak was deconvoluted on selected catalysts by ﬁtting

the data using two Gaussian functions with full width at half

maxima of 2.0 eV. Although there are more than two oxygen

species on the surface of these catalysts (e.g. MgO, REO, and

hydroxyl plus carbonate groups), it is exceedingly diﬃcult to

unambiguously separate the contributions from these oxide

species on the catalysts. Therefore, the deconvolutions were

done to separate contributions from MgO/REO (O 1s peaks

with binding energies of ∼529.3 eV) and hydroxyls/carbonates

(with typical binding energies around 531.5 to 532 eV).

2.2.5. Diﬀuse Reﬂectance Infrared Spectroscopy. In situ

DRIFTS were performed on a Nicolet 6700 FTIR equipped

with a Praying Mantis reaction chamber apparatus. Spectra

were collected using 256 scans and a resolution of 4 cm⁻¹.

Potassium bromide (KBr) was used to dilute the catalyst to

10% weight. Samples were degassed in situ at 450 °C under an

inert ﬂow for 30 min prior to the experiments. Backgrounds

were collected after the outgassing procedure. The catalysts

were then exposed to CO₂at 50 or 250 °C for 10 min and

allowed to outgas for 5 min under a He ﬂow before the data

was collected. The spectra are presented as a percent of

absorbance (i.e., an absorbance of 0% indicates no adsorbed

species).

(TCD) and a ﬂame ionization detector (FID) preceded by a

methanizer to convert all CO and CO₂to CH₄.²⁵Each catalyst

was pelletized using a Carver pellet press, then crushed using a

mortar and pestle and sieved to a size range of 180−250 μm.

For loading the reactor, 0.4 g of the sieved catalyst was

supported in the quartz reactor tube between two pieces of

quartz wool. CH₄, O₂, and an internal standard (N₂) were fed

through the system at a rate of 160 standard cm³(sccm) (with

N₂constant at 23.2 sccm) using three mass ﬂow controllers

(MFCs), which results in a GHSV of 2400 h⁻¹. This is the

highest ﬂow rate that will give the maximum yield under the

conditions in the study. The CH₄/O₂feed ratio was held

constant at 4:1 for all experiments because this is the

stoichiometric amount needed for conversion of methane to

ethane and water, and it was also shown to give the highest

CH₄conversion at a reasonable C₂selectivity in a previous

study.²⁵Although lower CH₄/O₂feed ratios can give a higher

methane conversion, the selectivity decreases signiﬁcantly as

the oxygen concentration is increased. Furthermore, because

the oxidation to CO and CO₂is more exothermic than OCM

reaction pathways, issues with temperature control can arise at

lower CH₄/O₂feed ratios. For these reasons, the CH₄/O₂feed

ratio was held constant at 4:1.

Quantitative analyses of products were done using the TCD

and FID peak area ratios compared with the reference N₂(all

calibration curves were linear, and regression analysis resulted

in ﬁts of ≥99.9%). The reaction temperatures reported are

measured by a thermocouple supporting the catalyst bed on the

eﬄuent side of the reactor. Two methane conversions are

reported in this study. The ﬁrst is a “total” methane conversion

(X_CH), which is calculated by the diﬀerence of the methane

4

entering the reactor system and the unconverted methane in

the reactor eﬄuent as determined by the GC (eq 1). The

second methane conversion (X*_CH, eq 2) is the fraction of

4

methane converted to CO_xand C₂₊products (CO₂, CO, and

C₂plus C₃products), which are the products used to determine

the C₂₊selectivity (S_C, eq 3). The C₂₊yield (y_C), is calculated

2+

by multiplying the methane conversion (X*_CH) and the C₂₊

4

selectivity (S_C) as described in eq 4. Using the X*_CHgives a

2+

4

more accurate C₂₊yield that can be obtained in the reaction,

even though this is not commonly accounted for, and

furthermore, the diﬀerence between the X_CHand X_C*_H

4

methane conversions appears to be mainly due to carbon

deposition on the reactor walls downstream of the catalyst bed

and, thus, most likely due to decomposition of the more

reactive products. The diﬀerence in methane conversion is not

due to signiﬁcant amounts of formed oxygenates, that is,

condensable byproducts because DART-MS analysis on the

condensate detected only trace amounts of formic acid, too low

to quantify accurately.

2.2.6. CO₂Temperature-Programmed Desorption. Tem-

perature-programmed desorption measurements after CO₂

adsorption were performed on a CHEMBET 3000 instrument.

A 250 mg portion of catalyst was loaded into a quartz u-tube

and was outgassed under a He ﬂow at 800 °C for 30 min prior

to CO₂exposure at 200 °C for 30 min. After allowing physically

absorbed CO₂to outgas at 200 °C, the catalyst was heated to

800 °C at a rate of 10 °C/min. Temperature measurements

were recorded by a thermocouple inside the quartz u-tube (on

the outlet side), and a TCD detector measured the amount of

desorbed gas.

2.3. OCM Reaction Experiments. The catalyst testing was

performed on a previously reported system consisting of a

quartz tube reactor with an inner diameter of 10 mm and an

online gas chromatograph (GC) for product monitoring

equipped with two detectors, a thermal conductivity detector

CH_4,in− CH_4,out

X_CH

=

4

CH_4,in

(1)

*

X_CH

=

4

[2·(C₂H₆+ C₂H₄) + 3·(C₃H₈) + CO + CO₂]_out

CH_4,in

(2)

1974

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Figure 1. Reaction data during oxidative coupling of methane as a function of temperature obtained over (a) Sm₂O₃/n-MgO, (b) Li-Sm₂O₃/n-MgO,

□

△

and (c) Li-TbO_x/n-MgO catalysts at a CH₄/O₂ratio of 4:1. ( ) Methane conversion (X_CH); (blue ) methane conversion to C₂₊and CO_x

4

▽

products (X*_CH); (☆) selectivity to C₂₊products (S_C); and (red ) C₂₊product Yield (Y_C).

4

2+

sccm CH₄reacted to C₂and C₃products

temperatures for time-on-stream experiments, the CH₄

conversion, C₂₊selectivity, and C₂₊product yield was

monitored as a function of reaction temperature for Sm₂O₃/

n-MgO, Li-Sm₂O₃/n-MgO, and Li-TbO_x/n-MgO (Figure 1).

The TbO_x/n-MgO catalyst was not included in this experiment

because an undoped TbO_xcatalyst would be expected to be

more selective toward CO and CO₂than to C₂products

compared with the other catalysts in this study. The

temperature of optimum C₂₊product selectivity for the

TbO_x/n-MgO catalyst would therefore not be as important,

as for the other catalysts. As evidenced in Figure 1, the three

catalysts included in the study exhibit very diﬀerent behavior

with reaction temperature.

The Sm₂O₃/n-MgO catalyst is surprisingly active, even at

550 °C with a CH₄conversion of 22% and a selectivity of 44%.

Similar values have been reported previously for Sm₂O₃,

although for reaction temperatures of 700 °C and above.^23,37,38

In fact, the conversion and selectivity for the Sm₂O₃/n-MgO

catalyst under these conditions are slightly better than those

observed at a reaction temperature of 740 °C over the pure

Sm₂O₃nanoparticles in our previous study,²⁵even though the

rare earth oxide content is signiﬁcantly lower in the n-MgO-

supported catalyst. As expected,²³increasing the temperature

increases the C₂₊selectivity, and a maximum selectivity of 53%

is observed at 675 °C over this catalyst. This is also the

temperature of maximum C₂₊yield (11.4%) because the CH₄

conversion remains around 23% 1% in the temperature range

under investigation (500−800 °C).

Doping the Sm₂O₃/n-MgO catalyst with Li improves its

performance, but the catalyst requires higher temperatures to

activate. Below 600 °C, very little methane is converted, and

the only observed products are CO₂with a small fraction of CO

(Figure 1b). A temperature above 600 °C is required for this

catalyst to exhibit a signiﬁcant conversion; however, once the

catalyst is activated, the activity and selectivity increase very

rapidly between 600 and 625 °C. The maximum C₂₊yield

(15.1%) from the Li-Sm₂O₃/n-MgO catalyst is obtained at 650

°C, where the CH₄conversion is 26% and the C₂₊selectivity

65%. If the temperature is increased above 650 °C, there is a

gradual decrease in selectivity while the CH₄conversion

S_C

Y

=

2+

sccm CH₄reacted to CO , C₂, and C₃products

(3)

(4)

x

*

= X ·S

CH₄C₂₊

C₂₊

The catalytic activity and selectivity were initially evaluated as

a function of temperature to determine a suitable temperature

interval for longer time-on-stream experiments. In an initial test

reaction, product monitoring began at 300 °C, after which the

temperature was increased by intervals of 100 °C until evidence

of methane conversion was observed. After this experiment, a

new catalyst was loaded into the reactor and brought to a

temperature 100 °C lower than the activation temperature

determined in the test reaction. The activity and selectivity of

selected catalysts were then measured as a function of

temperature by increasing the temperature in 50 °C intervals

until the temperature reached 800 °C. From these experiments,

two reaction temperatures (650 and 700 °C) were selected, and

the OCM activity was monitored at these temperatures for 8 h.

In each of these experiments (all catalysts and each reaction

temperature), a new, fresh catalyst was loaded into the reactor

before monitoring the reactor eﬄuent composition. The ﬁrst

measurement was taken 15 min after the reactor temperature

reached a steady state. Subsequent measurements were taken in

30 min intervals.

Because Li-doped catalysts are known to deactivate, the best

lithium-doped catalyst was subjected to an extended 30 h

reaction experiment to test the long-term stability. For the ﬁrst

2 h, measurements were taken every 30 min. Afterward, the

collection interval was lengthened to every hour.

3. RESULTS AND DISCUSSION

A number of Li-doped and pure REO catalysts supported on n-

MgO were prepared using diﬀerent methods and lithium

precursors and then subjected to methane coupling experi-

ments and catalyst characterizations to determine the eﬀects of

Li on the REO catalyst properties and if the eﬀect is dependent

on the preparation method, lithium precursor, or speciﬁc REO

used.

3.1. OCM Reaction Results. 3.1.1. OCM Activity Measure-

ments: Eﬀects of Temperature. To determine the best

remains around 25%

0.5%. The decrease in selectivity is

1975

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Table 1. Reaction Data for the Oxidative Coupling of Methane at 700 °C Obtained from MgO-Supported, Undoped and Li-

Doped, Sm₂O₃and TbO_xCatalysts

product selectivity, %

a

b

c

d

e

f

catalyst description

Li precursor

prep. method

X_CH, %

X_C*_H, %

S_C, %

Y_C, %

C₂H₄

C₂H₆

CO₂

CO

4

2

Sm₂O₃/n-MgO

none

IM

24.3

23.8

25.0

22.5

24.4

26.7

25.1

24.8

19.9

24.1

25.0

26.0

24.9

23.7

23.8

21.3

20.8

21.2

20.4

21.1

20.8

22.2

22.0

17.2

18.6

21.8

22.1

23.9

22.1

22.0

21.6

52.3

51.7

63.2

56.2

62.5

61.1

58.2

61.1

38.9

44.3

63.1

58.1

63.8

63.4

57.5

55.4

11.1

10.8

13.4

11.5

13.2

12.7

12.9

12.7

6.7

26.3

23.8

38.1

32.3

38.2

38.8

33.6

31.8

16.3

17.7

37.9

29.7

36.8

37.9

28.8

25.7

25.3

27.4

23.7

23.0

22.8

21.1

23.4

27.2

22.3

26.2

23.8

27.2

25.5

24.1

27.8

29.0

44.3

45.7

34.3

39.5

33.6

32.8

39.1

37.5

59.3

54.6

35.6

41.4

36.2

35.2

41.5

43.4

4.1

3.1

3.9

5.2

5.4

7.3

3.9

3.5

2.1

1.5

2.7

1.6

2.7

1.9

none

ME

CIM

SIM

CIM

SIM

IM

Li-Sm₂O₃/n-MgO-ME

LiC-Sm₂O₃/n-MgO-ME

Li-Sm₂O₃/n-MgO

LiNO₃

LiCl₃

LiNO₃

LiCl₃

none

Li-Sm₂O₃/n-MgO-SIM

LiC-Sm₂O₃/n-MgO

LiC-Sm₂O₃/n-MgO-SIM

TbO_x/n-MgO

TbO_x/n-MgO-ME

none

ME

CIM

SIM

CIM

SIM

8.3

Li-TbO_x/n-MgO-ME

LiC-TbO_x/n-MgO-ME

Li-TbO_x/n-MgO

LiNO₃

LiCl₃

LiNO₃

LiCl₃

13.8

12.8

15.3

14.0

12.6

12.0

Li-TbO_x/n-MgO-SIM

LiC-TbO_x/n-MgO

LiC-TbO_x/n-MgO-SIM

22.7

a

b

REO loading is 20% by weight. Li loading is 2.5% by weight. Preparation: IM = incipient wetness impregnation, ME = microemulsion, CIM =

c

d

coimpregnated, SIM = sequential impregnation. Overall methane conversion. Calculated as (CH_4in− CH_4out)/CH_4in. Methane conversion to CO_x

and C₂products only. Percent of CH₄converted to C₂₊products. Calculated as X_C*_H·S_C

e

f

.

4

2+

likely due to removal of Li from the catalyst surface, from the

extended time-on-stream as well as the increasing temperature.

It is interesting to note that the C₂H₄/C₂H₆ratio is higher over

the Li-doped compared with the undoped Sm₂O₃/n-MgO

catalyst, and the ratio increases with increasing temperature

(not shown). A higher C₂H₄/C₂H₆ratio with increasing

temperature is expected,²³and addition of an alkaline-earth

metal, such as Sr, to a Sm₂O₃catalyst has been shown

previously to increase this ratio.³⁸

and above, more C₂H₄is formed than C₂H₆. Thus, although the

maximum yield is obtained at 650 °C, the concentration of the

more valuable C₂H₄product increases with temperature, and

the decrease in C₂yield is not drastic between 650 and 800 °C.

For example, the C₂₊yield decreases from 14.9% to 12.6%

between 650 and 800 °C, while the C₂H₄/C₂H₆ratio increases

from 1.1 to 2.5 in this temperature range.

3.1.2. Eﬀects of Catalysts Preparation Method and Li

Precursor. The eﬀects of the catalyst preparation method and

lithium precursor on the MgO-supported Sm₂O₃and TbO_x

catalysts were investigated at a reaction temperature of 700 °C.

The activity, selectivity, and product distribution are presented

in Table 1. The Sm₂O₃/n-MgO-ME catalyst (prepared using

the microemulsion method) yields results very similar to those

for the Sm₂O₃/n-MgO catalyst prepared using incipient

wetness impregnation. The main diﬀerences between these

two catalysts are a slightly lower C₂H₄/C₂H₆ratio and

somewhat higher CO₂/CO ratio for the Sm₂O₃/n-MgO-ME

catalyst. The same trend in product ratios is observed over the

TbO_x/n-MgO catalysts, but in this case, the TbO_x/n-MgO-ME

catalyst is both more active and selective than the TbO_x/n-

MgO catalyst prepared via IWI. However, the advantage with

using the microemulsion technique to prepare TbO_xcatalysts is

diminished once lithium is added to the catalyst. There is not a

signiﬁcant diﬀerence in performance of the Li-Sm₂O₃/n-MgO

and Li-Sm₂O₃/n-MgO-ME catalysts, either. Thus, the more

cost-eﬀective incipient wetness impregnation can be used to

prepare these catalysts without any negative eﬀects.

For the samaria catalysts, there is not a signiﬁcant diﬀerence

between catalysts prepared using diﬀerent lithium precursors,

regardless of preparation method used (Table 1). Furthermore,

the preparation methods, microemulsion versus co- or

sequential impregnation, do not have a signiﬁcant eﬀect on

the activity and selectivity over the samaria catalysts. In

contrast, the LiNO₃precursor results in higher C₂₊yields

compared with the LiCl precursor for the Li-TbO_x/n-MgO

catalysts. This result is independent of the preparation methods

used in this study. The higher yield over the TbO_x/n-MgO

The behavior of the Li-TbO_x/n-MgO catalyst is diﬀerent

from both the Sm₂O₃-based catalysts. It has a more gradual

light-oﬀ with increasing temperature than the Li-doped Sm₂O₃/

n-MgO catalyst. The yield is similar to that obtained over the

Sm₂O₃/n-MgO catalyst at 550 °C, despite the lower CH₄

conversion. Thus, the similar yield is due to a higher C₂₊

selectivity, and interestingly, the C₂₊selectivity is higher over

the Li-TbO_x/n-MgO catalyst than either of the Sm₂O₃-based

catalysts at most of the temperatures investigated. Between 550

and 650 °C, the CH₄conversion increases from 16% to 24%,

and the C₂₊selectivity, from 56% to 65%. Therefore, the

maximum C₂₊yield obtained over this catalyst is 14.9%, which

is similar to that obtained over the Sm₂O₃/n-MgO catalyst, and

it is obtained at a temperature as low as 650 °C. Compared

with the Li-Sm₂O₃/n-MgO catalyst, the CO₂/CO ratio is

signiﬁcantly higher on the Li-TbO_x/n-MgO catalyst because of

a lower CO concentration. Above 650 °C, there is a slight

decline in the C₂₊selectivity due to an increase in both CO and

CO₂product formation. A decrease in C₂₊selectivity with

temperature over Sr-doped Sm₂O₃catalysts has been

observed,³⁷even though most OCM catalysts, such as Li/

MgO³⁹and the state-of-the-art Mn-Na₂WO₄/SiO₂catalyst,⁴⁰

appears to yield an increase in C₂₊selectivity with temperature.

The decrease in C₂₊selectivity with temperature is likely due to

the rather low CH₄/O₂ratio because at higher CH₄/O₂ratios,

REO catalysts are more likely to exhibit an increase in C₂₊

selectivity with temperature because the O₂concentration is

low and can limit the CO_xformation.^16,23,41As expected, the

C₂H₄/C₂H₆ratio increases with temperature,¹⁶and at 650 °C

1976

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

catalysts prepared using the LiNO₃precursor is due to a higher

selectivity compared with the catalysts prepared using the LiCl

precursor.

Considering the results from the lithium precursor and

preparation methods study, the LiNO₃precursor and the

simpler coimpregnation method were selected to prepare the

undoped and Li-doped samaria and terbia catalysts for the time-

on-stream study.

Table 3. Product Distribution Obtained in the Oxidative

Coupling of Methane over MgO-Supported, Undoped and

Li-Doped, Sm₂O₃and TbO_xCatalysts at the Beginning of

the Experiment As Well As after 8 h On-Stream

product distribution, %

rxn

temp,

°C

C₂H₄/

C₂H₆

ratio

CO₂/

CO

ratio

catalyst

description

a

C₂H₄C₂H₆CO₂

CO

3.1.3. Time-On-Stream OCM Activity Measurements.

Because Li-doped catalysts are known to deactivate with time

on-stream, the catalytic activity and selectivity of the Sm₂O₃/n-

MgO and TbO_x/n-MgO catalysts with and without lithium

doping were evaluated at 650 and 700 °C for 8 h. These two

temperatures were selected because the maximum C₂₊yield was

observed in this range for all the catalysts under investigation.

Although the undoped TbO_x/n-MgO catalyst was not included

in the temperature study, it was included in these measure-

ments for comparison and also to determine the deactivation

rate of a terbia catalyst without added lithium. The n-MgO

support and two Li/n-MgO catalysts were included in this

study to determine the eﬀects of the support and the Li/MgO

combination in these catalysts as well as to have a reference for

the stability of the lithium containing catalysts. The results are

presented in Tables 2 and 3 and Figures 2 and 3.

b

n-MgO

Li(N)

650

22.4

23.2

7.3

17.4

17.3

37.9

39.9

19.7

20.5

38.6

44.6

18.1

21.4

26.8

26.3

25.0

25.1

25.6

37.5

22.7

19.1

23.6

21.2

22.4

21.7

31.2

27.1

30.0

23.4

18.2

54.6

54.8

45.9

44.0

32.2

28.4

45.8

40.5

31.0

25.4

42.4

47.6

41.5

46.7

33.0

36.8

36.0

33.5

61.0

66.0

57.3

60.0

37.2

36.4

36.7

38.3

44.1

5.6

4.7

8.9

8.6

10.8

16.4

3.7

3.9

10.8

16.2

4.8

4.0

4.8

4.0

5.5

6.8

3.2

10.3

2.0

1.5

2.6

2.2

0.4

2.9

0.5

3.0

5.6

1.3

0.2

1.9

1.7

0.3

0.2

2.2

1.7

0.9

0.8

1.1

1.0

1.4

0.5

1.7

1.9

0.6

0.5

0.8

0.7

1.0

1.2

1.1

1.5

1.8

9.8

11.7

5.2

b

700

b

650

c

650

7.5

5.1

b

700

37.3

34.8

11.9

11.0

40.2

37.0

26.0

22.1

28.7

24.2

35.9

18.9

38.1

37.2

13.4

11.3

17.7

16.1

31.2

33.6

35.3

32.1

3.0

c

700

1.7

b

Li(C)

650

12.4

10.4

2.9

c

650

b

700

c

700

1.6

b

Sm₂O₃

Li-Sm₂O₃

TbO_x

650

8.8

c

650

11.8

8.6

b

700

c

700

11.8

6.0

b

650

c

650

5.4

b

700

11.3

3.2

c

Table 2. Reaction Data for the Oxidative Coupling of

Methane Obtained from MgO-Supported, Undoped and Li-

Doped, Sm₂O₃and TbO_xCatalysts after 8 h on-Stream

700

b

650

30.5

43.0

22.0

26.8

93.0

12.7

73.4

12.6

7.9

c

650

b

700

Y_C

rxn

2

c

700

X_CH

b⁴

,

X_CH

c⁴

,

S_C,

%

Y_C,

%

catalyst

temp,

loss,

SSA,

2

d

2

e

b

a

f

g

m²/g

20.4

Li-TbO_x

650

description

°C

%

c

h

650

n-MgO

650

700

650

700

650

700

650

700

650

700

650

700

650

700

22.1

21.9

3.6

17.8

18.0

2.4

39.8

40.4

47.4

55.7

55.6

59.1

49.0

49.9

56.5

57.3

32.7

38.0

62.1

60.0

50.2

7.1

7.3

b

700

c

700

Li(N)

1.1

5.4

28.5

17.4

41.3

12.1

13.0

74.6

24.4

18.9

9.8

2.1

3.0

d

700

15.9

5.1

12.9

3.6

7.2

a

All catalysts are supported on n-MgO, and REO loading is 20% by

weight. Li loading is 2.5% by weight. Initial reaction data point.

Reaction data after 8 h on-stream. Results after 30 h on-stream.

Li(C)

2.0

b

17.9

22.0

22.4

5.9

13.2

20.3

5.2

7.8

c

d

Sm₂O₃

Li-Sm₂O₃

TbO_x

10.0

10.1

2.9

26.0

2.7

1), it is higher than some reported values for doped MgO.¹¹

This suggests that the MgO nanoparticles contain impurities

that are favorable in the OCM reaction because undoped MgO

typically exhibits lower selectivities.³This may be due to trace

amounts of sodium in the n-MgO because this and a trace

amount of chlorine are the only impurities detectable in the

XPS spectrum obtained from the pure nanoparticles (not

shown).

Adding lithium to the magnesia support actually decreases

the activity; the reference Li/n-MgO catalysts exhibit a lower

methane conversion at 650 °C (Table 2). However, the C₂₊

selectivity is higher over the Li/n-MgO compared with the n-

MgO support. The Li/n-MgO catalyst prepared using the

carbonate precursor {Li(C)/n-MgO} performs better than the

Li(N)/n-MgO catalyst in terms of both C₂₊selectivity and CH₄

conversion. Both Li/n-MgO catalysts display some deactiva-

tion, but the activity is so low that it is diﬃcult to comment on

the severity of the activity loss (Figure 2).

22.7

16.7

18.9

21.9

23.0

22.2

19.6

15.7

17.3

20.0

21.1

18.8

11.2

5.2

21.8

5.9

6.6

Li-TbO_x

12.5

12.6

9.5

16.9

19.0

29.6

i

700

a

All catalysts are supported on n-MgO, and REO loading is 20% by

b

weight. Li loading is 2.5% by weight. Overall methane conversion.

c

Calculated as (CH_4in− CH_4out)/CH_4in. Gas phase CH₄conversion to

d

CO_xand C₂products only. Percent of CH₄converted to C₂products.

e

f

Calculated as X*_CH·S_C. Percent loss of activity after 8 h on-stream.

4

2

g

h

SSA: speciﬁc surface area (m²/g). Initial reaction data point

i

collected for n-MgO support only. Reaction data after 30 h time

on-stream.

3.1.3.1. n-MgO and Li/n-MgO Catalysts. The n-MgO

support is surprisingly active at the reaction temperatures

investigated (Table 2). A methane conversion of 22% is similar

to the value obtained for pure Sm₂O₃and only slightly lower

than that obtained over the Sm₂O₃/n-MgO catalyst (24%), and

although the C₂₊selectivity (40%) is lower than that obtained

over pure Sm₂O₃(44%)²³and Sm₂O₃/n-MgO (52%) (Table

At a reaction temperature of 700 °C, the activity and

selectivity for both Li/n-MgO catalysts are signiﬁcantly higher

compared with the 650 °C results. The Li(C)/n-MgO catalyst

exhibits the better initial performance of the two catalysts with

a CH₄conversion of almost 20% and a C₂₊selectivity of 59.1%.

1977

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

▼

Figure 2. Oxidative coupling of methane reaction data stability experiments at 650 °C for diﬀerent catalysts supported on n-MgO: (red ) Li-TbO_x,

■

◆

▲

(dark blue ) Li-Sm₂O₃, (black ) Sm₂O₃, (green ) TbO_x, (light blue right-pointing triangle) Li(N), (orange left-pointing triangle) Li(C).

Figure 3. Oxidative coupling of methane reaction data for stability experiments at 700 °C for diﬀerent catalyst systems supported on n-MgO: (red

■

◆

▼

▲

) Li-TbO_x, (dark blue ) Li-Sm₂O₃, (black ) Sm₂O₃, (green ) TbO_x, (light blue right-pointing triangle) Li(N), (orange left-pointing triangle)

L(C).

However, compared with the Li(N)/n-MgO catalyst, the

Li(C)/n-MgO catalyst experience a signiﬁcantly higher drop

in activity, and after 8 h on-stream, the CH₄conversion is only

12.9%, which is similar to the 12.8% conversion obtained from

the Li(N)/n-MgO catalyst. Although there is a small drop in

C₂₊selectivity during the ﬁrst 2 h for both Li/n-MgO catalysts,

the C₂₊selectivity is fairly stable during the rest of the run. With

time on-stream, slightly more CO is formed, while the CO₂and

the C₂H₄concentrations are both lower compared with the

initial product distribution (Table 3). Therefore, the activity

loss (lower methane conversion) is the main reason for the

lower C₂₊yield after 8 h on-stream for these catalysts. The

product distributions over the two Li/n-MgO catalysts are

similar (Table 3) but depend strongly on the temperature. At a

reaction temperature of 650 °C, the C₂H₄/C₂H₆ratio is

between 0.2 and 0.3, whereas at 700 °C, this ratio is between

1.7 and 2.2 for the Li/n-MgO catalysts. There is also a

signiﬁcant decrease in the CO₂/CO ratio when the temperature

is increased from 650 to 700 °C. Although the C₂H₄/C₂H₆

ratio is not altered signiﬁcantly, the CO₂/CO ratio decreases

with time on-stream (Table 3).

In summary, the Li/n-MgO catalysts are not very active at

650 °C, but they do exhibit a reasonable initial activity and

selectivity in the OCM reaction at 700 °C. However, as

expected,⁴they are not very stable catalysts. The initially more

active Li(C)/n-MgO catalyst exhibits a more severe deactiva-

tion and, thus, has an activity similar to the Li(N)/n-MgO

catalyst after the 8 h on-stream. The main reason for the loss in

yield over these catalysts with time on-stream is due to a lower

methane conversion.

3.1.3.2. Sm₂O₃/n-MgO Catalysts. The Sm₂O₃/n-MgO

catalyst yields very similar time-on-stream results at the two

reaction temperatures (650 and 700 °C), with the C₂₊yields

after 8 h being 10.0% and 10.1%, respectively (Figures 2 and 3).

This is in the range reported for pure Sm₂O₃catalysts.¹⁷

Compared with the Li/n-MgO catalysts, the CH₄conversion is

higher, but the C₂₊selectivity is lower (or similar) over the

Sm₂O₃/n-MgO catalyst. The initial C₂₊yield obtained over the

1978

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Li(C)/n-MgO catalyst is slightly higher than that obtained over

the Sm₂O₃/n-MgO catalyst; however, the Sm₂O₃/n-MgO

catalyst is signiﬁcantly more stable with time on-stream. At

both temperatures, the Sm₂O₃/n-MgO catalyst exhibits only a

slight decrease in C₂₊yield with time on-stream, which is

mainly due to a loss in the C₂₊selectivity, although there is also

a slight decrease in methane conversion over the 8 h

experiment. Therefore, after 8 h on-stream, the Sm₂O₃/n-

MgO catalyst is signiﬁcantly more active than either of the Li/

MgO catalysts. The C₂H₄/C₂H₆ratio decreases with increasing

time on-stream for the Sm₂O₃/n-MgO catalyst as a result of a

decrease in the amount of C₂H₄product because the C₂H₆

formation is uniform for the 8 h experiment. The CO₂product

formation increases with time on-stream while the CO

formation is fairly constant. Thus, the loss in C₂₊yield appears

to occur from complete oxidation of C₂H₄to CO₂and H₂O.

This trend is observed at both reaction temperatures (Table 3).

Although the CO₂/CO ratio is roughly the same, there is a

slight increase in the C₂H₄/C₂H₆ratio between 650 and 700

°C, but the increase is not suﬃcient to warrant a higher

reaction temperature. Therefore, because there is no beneﬁt to

increasing the temperature, it would be recommended to

operate the Sm₂O₃/n-MgO catalyst at 650 °C instead of 700

°C.

The eﬀects of adding lithium to the Sm₂O₃/n-MgO catalyst

are dependent on the temperature of reaction. At 650 °C,

doping with lithium increases the C₂₊selectivity but decreases

the CH₄conversion compared with undoped Sm₂O₃/n-MgO

catalyst. This eﬀect has been observed previously.²⁶The net

result is that the initial C₂₊yields are the same over the Sm₂O₃/

n-MgO and Li-Sm₂O₃/n-MgO catalysts. However, after only a

few hours on-stream, the CH₄conversion has dropped from

17% to 5.8%, indicating that this is not a stable catalyst. There is

also a drastic change in the C₂H₄/C₂H₆ratio with time on-

stream (Table 3). Initially, more C₂H₄is formed, but after 8 h

on-stream, signiﬁcantly more C₂H₆is formed than C₂H₄. At a

reaction temperature of 700 °C, both the C₂₊selectivity and

CH₄conversion are initially higher over the lithium-doped as

compared with the undoped Sm₂O₃/n-MgO catalyst. However,

compared with the Sm₂O₃/n-MgO catalyst, both the C₂₊

selectivity and CH₄conversion exhibit a faster decline with

time on-stream over the Li-Sm₂O₃/n-MgO catalyst. Never-

theless, the Li-Sm₂O₃/n-MgO catalyst is more stable at 700 °C

than at 650 °C, which is consistent with observations for Li/

MgO catalysts.⁴Although the initial C₂₊yield is signiﬁcantly

higher over the Li-Sm₂O₃/n-MgO catalyst, after 8 h on-stream,

the C₂₊yield approaches that of the Sm₂O₃/n-MgO catalyst.

The loss in C₂₊yield is more signiﬁcant for the Li-Sm₂O₃/n-

MgO compared with the Sm₂O₃/n-MgO catalyst, but it is not

as high as for the Li(C)/n-MgO catalyst. Furthermore, the

major cause of the loss in C₂₊yield is due to decreased C₂₊

selectivity (after 8 h on-stream, more CO₂and less C₂H₄is

formed compared with the initial product distribution) over the

Li-Sm₂O₃/n-MgO catalysts, rather a drop in CH₄conversion as

observed for the Li/n-MgO catalysts. The inﬂuence of Li

doping on the product distribution at 700 °C is an increase in

the C₂H₄/C₂H₆ratio and decrease in the CO₂/CO ratio (less

CO₂and more CO is formed); therefore, the change in product

distribution with time on-stream, a lower C₂H₄/C₂H₆ratio and

a higher CO₂/CO ratio, is consistent with loss of lithium.

3.1.3.3. TbO_x/n-MgO Catalysts. The lower C₂₊yield

obtained over the TbO_x/n-MgO catalyst compared with the

Sm₂O₃/n-MgO is expected because terbia is a better oxidation

catalyst than samaria. Nevertheless, the initial C₂₊selectivity is

reasonably high, between 30% and 40%, which is likely due to

the n-MgO support, but it decreases with time on-stream. Both

the C₂products C₂H₄and C₂H₆decrease, while the

contribution from CO₂in the product stream is higher after

8 h on-stream. In addition to the lower C₂₊selectivity, the CH₄

conversion is also slightly lower than the conversions obtained

from the n-MgO support and the Sm₂O₃/n-MgO catalyst

(Table 2). Because the activity is so low for the Li/MgO

catalysts, the C₂₊yield at 650 °C is higher over the TbO_x/n-

MgO catalyst despite the low C₂₊selectivity. At 700 °C, the C₂₊

yield is higher than at 650 °C, but it is now lower than the C₂₊

yields obtained over the Li/MgO catalysts, at least initially.

Because the TbO_x/n-MgO catalyst is more stable than the Li/

MgO catalysts, the diﬀerence in C₂₊yields between these

catalysts is only 1% after 8 h on-stream.

Adding lithium to the TbO_x/n-MgO catalyst increases both

the CH₄conversion and the C₂₊selectivity. In fact, at 650 °C,

the increase in the C₂₊yield is over 100% when lithium is added

to the MgO-supported TbO_xcatalyst. The majority of the

improvement is from the enhanced C₂₊selectivity, which nearly

doubles with the addition of lithium (32.7% for undoped and

62.1% for Li-doped TbO_x/n-MgO). This is a very high value

for a TbO_xcatalyst, and it is a high selectivity also for other

OCM catalysts at a CH₄/O₂ratio as low as 4.²³Furthermore,

the Li-TbO_x/n-MgO catalyst is signiﬁcantly more stable than

the Li-Sm₂O₃/n-MgO catalyst, particularly at 650 °C. The

deactivation rate for this catalyst appears to be similar to that of

the undoped TbO_x/n-MgO catalyst. As for the Sm₂O₃-based

catalysts, the activity loss is mainly due to a drop in C₂₊

selectivity with time on-stream over the TbO_x/n-MgO catalyst,

as opposed to the drop in CH₄conversion observed over the

Li/MgO catalysts. The eﬀects of Li doping on the product

distribution of TbO_x/n-MgO catalysts is similar to the eﬀects of

adding Li to Sm₂O₃/n-MgO catalyst (at 700 °C), that is, the

C₂H₄/C₂H₆ratio increases, and the CO₂/CO ratio decreases.

In fact, even at 650 °C, the Li-TbO_x/n-MgO catalyst achieves a

C₂H₄/C₂H₆ratio >1.0, which is remarkable because the most

active Sm₂O₃/Al₂O₃catalyst required 800 °C to do so.²⁵If

deactivation was due solely to Li leaving the catalyst, it may be

expected that the C₂H₄/C₂H₆ratio will decrease with time on-

stream. In contrast, the C₂H₄/C₂H₆ratio increases with time

on-stream over the Li-TbO_x/n-MgO catalyst, which means that

more of the formed C₂H₆is converted to C₂H₄. The

concentration of both CO and CO₂is also increased with

time on-stream. This may suggest that the contribution from

surface reaction increases with time on-stream, compared with

the gas phase radical reaction to form C₂H₆; however, there is

not a signiﬁcant change in the surface area of this catalyst with

time on-stream.

Previous studies using a lithium-doped magnesium oxide

catalyst⁴found that the severity of catalyst deactivation was

strongly correlated with the initial loading of lithium on the

catalyst. Loadings of 2% and above underwent severe

deactivation within the ﬁrst few hours of exposure to OCM

reaction conditions and, depending on the preparation method

for the Li containing catalyst, could reach as high as 70% loss of

activity. The optimum lithium doping in the study was found to

be 0.5%, although even at this Li loading, the deactivation was

signiﬁcant. Our Li-REO/n-MgO catalysts deactivate at a much

slower rate than both the Li/MgO catalysts studied by Arndt

and the two reference systems prepared in our lab. Particularly

interesting is the behavior by the Li-TbO_x/n-MgO catalysts at

1979

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

650 °C compared with the undoped TbO_x/n-MgO sample.

After 8 h of time on-stream, the activity loss for the two

catalysts is 18.9% and 16.9%, respectively. However, at 700 °C,

the deactivation for the lithium-doped sample is greater than

the undoped terbia catalyst. This is in contrast to the Li/MgO

systems reported in the literature, for which the higher reaction

temperatures resulted in more stable catalysts.⁴

3.1.4. Long-Term Stability Test of Li-TbO_x/n-MgO. A long-

term stability test was performed at a reaction temperature of

700 °C on the most active and stable catalyst, the Li-TbO_x/n-

MgO, to investigate the behavior at longer times on-stream.

The reaction results (CH₄conversion, C₂₊selectivity, and C₂₊

yield) are presented in Figure 4 and have been added to Tables

than the Li/n-MgO catalyst, ﬁnding an alternative promoter is

highly desirable.

3.2. Catalyst Characterization. The magnesia-supported

REO catalysts with and without lithium doping, as well as the

nanoparticle magnesia support, were subjected to catalyst

characterizations to determine properties of importance for the

catalytic activities and selectivities and determine the eﬀects on

the catalyst during the time on-stream experiments. The results

from each technique are presented below.

3.2.1. BET Surface Area Measurements. The BET surface

area of the undoped and lithium-doped REO supported on

magnesia catalysts after calcinations at 800 °C are presented in

Table 2. The surface area of the nanoparticle MgO support

(20.4 m²/g after calcination at 800 °C) was also measured and

included for reference. The as-prepared TbO_x/n-MgO and

Sm₂O₃/n-MgO catalysts have surface areas of 21.8 and 26.0

m²/g, respectively. Thus, adding the REOs to the n-MgO

support results in a slight increase in the overall surface area

versus the bare support surface area. In contrast, the surface

areas of the as-prepared Li-doped TbO_x/n-MgO and Sm₂O₃/n-

MgO catalysts are almost an order of magnitude smaller than

the undoped catalysts at only 5.9 and 2.7 m²/g, respectively. A

lower surface area after addition of lithium has been observed

previously²and is consistent with the surface areas of the two

Li/MgO reference systems, 2.1 and 3.0 m²/g for the Li(N)/n-

MgO and Li(C)/n-MgO, respectively. The BET data reveal

that there is no simple correlation between the overall catalyst

surface area and the activity or selectivity for these catalysts.

Once there is suﬃcient active surface area to activate methane

toward hydrogen abstraction, increasing the surface area of the

catalyst can decrease the selectivity and may therefore result in

a lower product yield. Thus, even though the surface areas are

rather low for a catalyst, this indicates that the surface area is

not the limiting factor for these catalysts under the conditions

used in the study.

3.2.2. XRD Analysis. The XRD patterns of the fresh Li-

doped and undoped TbO_x/n-MgO and Sm₂O₃/n-MgO are

shown in Figures 5 and 6. The XRD pattern of the pure n-MgO

support after calcinations at 800 °C for 4 h was also collected

(not displayed) to make sure that the particle size of the

support is not signiﬁcantly altered by deposition of the rare

earth oxides. Table 4 reveals that the MgO particle size on the

Figure 4. Oxidative coupling of methane reaction results for long-term

△

stability testing of Li−TbO_x/n-MgO at 700 °C: (blue ) methane

▽

conversion to gas phase products, (☆) C₂₊selectivity, (red ) C₂₊

yield.

2 and 3. Despite the promising results of the 8 h experiment,

this catalyst continues to deactivate at a low but fairly constant

rate during the 30 h on-stream. The 30% loss in C₂₊yield after

30 h on-stream (initial yield of 13.5% and ﬁnal yield of 9.5%) is

due to a small drop in CH₄conversion (from 22% to 18.5%) as

well as a more signiﬁcant loss in C₂₊selectivity (from 60% to

50%). Of course, this 30% loss in C₂₊yield after 30 h is to be

compared with the 40% loss in C₂₊yield over the Li(C)/n-

MgO catalyst after only 8 h on-stream. The C₂H₆product ﬂow

rate decreases over twice as much as the C₂H₄ﬂow rate (20%

for C₂H₄compared to 42% for C₂H₆). These results suggest

that the activity loss is due to a lower methyl hydrogen

abstraction rate with time on-stream, that is, the catalyst is

losing active sites. However, as more of the C₂H₆product is

dehydrogenated to C₂H₄, the dehydrogenation sites do not

seem to be aﬀected to the same extent. In addition, the surface

sites that generate CO and CO₂appear to increase. Because the

C₂₊yield never reaches a steady-state value during the long-

term reaction study, it must be assumed that this catalyst would

continue to lose activity, and that, even though it is more stable

Figure 5. XRD data obtained from fresh undoped, Li-doped, and spent

Li-doped Sm₂O₃/n-MgO. The labels indicate the following crystal

○

▽

phases: , Sm₂O₃(cubic); , Sm2O3 (monoclinic); ★, MgO.

1980

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

to refer to the phase as TbO_x, where 1.75 ≤ x ≤ 2. The XRD

pattern obtained from the fresh Li-TbO_x/n-MgO catalyst has

two peaks in addition to the ones originating from the TbO_x

and MgO phases. These two peaks are reasonably consistent

with reference peaks from hexagonal Li₂CO₃, according to the

database [JCPDS 072−6635], although other Li-containing

phases cannot be excluded. Compared with the reference

hexagonal Li₂CO₃peaks, the peaks in the XRD pattern

obtained from the fresh Li-TbO_x/n-MgO catalyst are shifted to

slightly higher 2θ values, indicating that the hexagonal Li₂CO₃

lattice is distorted in the catalyst, and this could be due to

interactions with the TbO_xon the surface.

The presence of Li-related peaks in the XRD pattern is

diﬀerent from the behavior of the Sm₂O₃/n-MgO catalyst, in

which no peaks due to Li-based compounds could be detected

using XRD. The peaks due to the TbO_1.81phase obtained from

the Li-TbO_x/n-MgO are smaller compared with the TbO_x/n-

MgO catalyst, whereas the particle size according to the

Scherrer equation is slightly larger (Table 4). This may be due

to a lower crystallinity of the Li-TbO_x/n-MgO catalyst

compared with the TbO_x/n-MgO. This behavior is also

diﬀerent from that of the undoped and Li-doped Sm₂O₃/n-

MgO catalysts and suggests that the interactions between Li

and TbO_xare diﬀerent from the interactions between Li and

Sm₂O₃. As for the Sm₂O₃-based catalysts, the MgO particles on

the Li-TbO_x/n-MgO catalyst are larger than on the undoped

TbO_x/n-MgO catalyst (Table 4). This indicates that the

lithium is not incorporated into the bulk phase of MgO,

although this has been observed under certain conditions.^42,43

The monoclinic and cubic Sm₂O₃mixture remains after

exposure to reaction conditions at 700 °C for 8 h, which

suggests that either an amount of lithium suﬃcient to stabilize

the monoclinic Sm₂O₃remains on the surface or the lithium-

induced formation of monoclinic Sm₂O₃is irreversible (Figure

6). Particle growth of the MgO is apparent in the XRD pattern

obtained from the spent Li-Sm₂O₃/n-MgO catalyst: the peaks

are signiﬁcantly narrower compared with the fresh catalyst, and

this is further supported by particle size calculations shown in

Table 4. The crystallite size of cubic Sm₂O₃, however, appears

to decrease slightly (from 54.6 to 48.1 nm) while the crystallite

size of the monoclinic Sm₂O₃is increased, as is evident in

Figure 5. This indicates that some cubic Sm₂O₃has been

converted to the monoclinic phase, which is expected because

prolonged exposure to temperatures as high as 800 °C may

induce phase transformation to the monoclinic phase, which is

the stable phase at higher temperatures.⁴⁴The increase in

monoclinic Sm₂O₃after time on-stream may explain some of

the catalyst deactivation, because the monoclinic Sm₂O₃phase

is less active and selective than the cubic Sm₂O₃phase.²⁶In

addition, some particle growth is expected after prolonged

exposure to high temperatures, which will reduce the Sm₂O₃

surface area available for reaction.

Figure 6. XRD data obtained from fresh undoped, Li-doped, and spent

Li-doped TbO_x/n-MgO catalysts. The labels indicate the following

◊

▼

●

crystal phases: , TbO_x; , Li₂CO₃; , Tb₂O₃; and ★, MgO.

Table 4. Particle Size Calculations Using the Scherrer

Equation and XRD Data Obtained from Selected Catalysts

particle size [nm]

a

catalyst

Sm₂O₃

TbO_x

MgO

b

n-MgO

20.0

21.3

37.1

42.7

23.7

34.8

40.7

b

Sm₂O₃/n-MgO

11.4

54.6

48.2

b

Li-Sm₂O₃/n-MgO

c

Li-Sm₂O₃/n-MgO spent

b

TbO_x/n-MgO

11.9

b

d

Li-TbO_x/n-MgO

16

c

Li-TbO_x/n-MgO spent

39.1

a

Cubic phase of Sm₂O₃. The peaks due to monoclinic Sm₂O₃were too

b

small for particle size determination. After calcination at 800 °C.

c

d

After 8 h on-stream. The signal-to-noise ratio for the TbO_x

crystalline peaks for this sample was very low, so the accuracy of the

calculation is lower compared with the other peaks.

bare support and the undoped TbO_x/n-MgO and Sm₂O₃/n-

MgO are very similar. The main phases present in the fresh

Sm₂O₃/n-MgO catalyst are cubic Sm₂O₃and cubic MgO with

no major impurities. In contrast, the coprecipitation of lithium

and samaria results in a monoclinic form of Sm₂O₃in addition

to the cubic Sm₂O₃phase on the n-MgO support (Figure 5).

No lithium-containing compounds were detected in the XRD

analysis, which means that the added lithium does not form

large crystalline particles of lithium-containing phases under

these conditions. However, it does not rule out formation of

amorphous or small segregated lithium compounds on the

surface because they would be XRD-invisible. Particle size

calculations using the Scherrer equation reveal that both the

cubic Sm₂O₃and MgO crystallites are larger on the Li-doped

versus undoped Sm₂O₃/n-MgO (Table 4). This is consistent

with the lower BET surface area of the Li-doped catalyst.

The XRD pattern obtained from the fresh TbO_x/n-MgO

suggests that the crystalline phase present is a slightly reduced

TbO₂phase, that is, a TbO_2−δphase where 0 ≤ δ ≤ 0.25.

Several reference XRD patterns from the literature for diﬀerent

terbia samples, particularly TbO₂, Tb₄O₇, and a TbO_1.81, were

compared to assist in the assignment of the TbO_xphase. The

reference data that ﬁt most closely the peaks in the XRD

pattern obtained from the TbO_x/n-MgO are assigned to a

TbO_1.81phase. However, the match is not exact, which leads us

The Li-containing phase is no longer present in the XRD

pattern obtained from the spent Li-TbO_x/n-MgO catalyst,

which may be indicative of lithium leaving the surface of the

catalyst during reaction. The TbO_x-related peaks are signiﬁ-

cantly more intense on the spent catalyst compared with the

fresh, which reveals substantial particle growth. The TbO_x

particle sizes increase from an average of 16 to 39 nm (Table

4). The peaks are also shifted from the original TbO_1.81phase

identiﬁed on the fresh Li-TbO_x/n-MgO catalyst. The peaks in

the XRD pattern obtained from the spent Li-TbO_x/n-MgO

catalyst are more consistent with the sesquioxide Tb₂O₃phase

1981

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

[JCPDS 76−156], which would mean that the TbO_1.81was

reduced during reaction. Despite removal of the Li-containing

phase, which was present on the fresh catalyst, the Li-TbO_x/n-

MgO catalyst is more stable than the other Li-doped catalysts in

the study. The deactivation rate of this catalyst is similar to the

undoped Sm₂O₃/n-MgO catalyst. This could indicate that

deactivation due to loss of active surface area from sintering

(particle growth) is an important deactivation pathway, in

addition to Li-removal, over the Li-TbO_x/n-MgO catalysts.

3.2.3. TEM Analysis. Selected representative TEM images for

lithium-doped and undoped Sm₂O₃/n-MgO and TbO_x/n-MgO

catalysts after calcination at 800 °C for 4 h are presented in

Figure 7. The Sm₂O₃/n-MgO and TbO_x/n-MgO catalysts

consist of particles with average particle sizes between 20 and

50 nm. This is reasonably consistent with the particle sizes

calculated using line-broadening of the XRD peaks. In addition,

the particles in the Sm₂O₃/n-MgO catalyst appear to be slightly

smaller than the particles of the TbO_x/n-MgO catalyst, which is

consistent with the higher surface area of the Sm₂O₃/n-MgO

catalyst. After doping with lithium there is a dramatic increase

in the particle size of the catalyst, consistent with the decrease

in BET surface area. The trend is also consistent with the

increase in particle size determined from the XRD data, but the

magnitude of the change is not captured with XRD. TEM

indicates that the particles are larger than what is determined

from XRD, which suggests that each particle in TEM consists

of more than one crystallite. The particles appear to be more

agglomerated in the Li-TbO_x/n-MgO catalyst, which could

explain the signiﬁcant reduction in surface area of this catalyst

compared with the bare support or the undoped TbO_x/n-MgO

catalyst. EDX was performed on the Li-REO/n-MgO catalysts

in an attempt to identify lithium on the surface, but the

attempts were unsuccessful because the atomic similarities

between magnesium and lithium complicates the process.

3.3.4. XPS Measurements. The XPS data obtained from the

fresh Li-doped and undoped Sm₂O₃/n-MgO and TbO_x/n-

MgO catalysts are presented in Figures 8 and 9. The

composition analysis from these narrow scans has been

included in Table 5. The Mg 2p peaks due to the MgO

support reveal that the electronic structure of the support is not

signiﬁcantly aﬀected by the addition of rare earth oxide or

lithium (Figure 8 a). The presence of lithium on the surface of

the Li-Sm₂O₃/n-MgO and Li-TbO_x/n-MgO catalysts is evident

as small peaks at a binding energy of 54.5 eV. The slightly

higher peak intensity of the Li 1s peak in the spectrum obtained

from the Li-Sm₂O₃/n-MgO catalyst indicates that the near-

surface concentration of lithium is higher on this catalyst

compared with the Li-TbO_x/n-MgO catalyst. This is also seen

in the compositional analysis (Table 5) and appears to be due

to terbia covering the lithium (the Tb 3d_5/2signal intensity is

slightly higher on the Li-doped catalysts). For the Sm₂O₃-based

catalysts, the addition of lithium reduces the intensity of the Sm

3d_5/2peak (revealing a lower surface concentration compared

to the undoped catalyst; Table 5), which suggests that the

lithium is covering some of the samaria on this catalyst surface.

The magnesium concentration is also slightly lower, which

could be due to some coverage by lithium species. There is no

apparent shift in the Sm 3d_5/2peak for the lithium-containing

catalyst (Figure 8 b), suggesting the lithium does not have a

charge eﬀect on the samaria on the surface. In contrast, lithium

addition to the TbO_x/n-MgO results in a one-eV shift in the Tb

3d_5/2peak to higher binding energies (Figure 8 c). This is an

indication of strong interactions between lithium and terbia and

Figure 7. TEM images of undoped and lithium-doped REO/n-MgO

catalysts at two diﬀerent magniﬁcations: Sm, Sm₂O₃/n-MgO; Li-Sm,

Li-Sm₂O₃/n-MgO; Tb, TbO_x/n-MgO; and Li-Tb, Li-TbO_x/n-MgO.

Left: low magniﬁcation; scale-bar, 100 nm. Right: high resolution;

scale bar, 50 nm.

could explain the diﬀerences in behavior between the Sm₂O₃/n-

MgO and TbO_x/n-MgO catalysts. It is possible that a small

amount of the added Li enters the TbO_xlattice and forms an

amorphous (and XRD-invisible) mixed Li-TbO_yphase. This

would explain the lower intensities of the TbO_xpeaks in the

XRD pattern and the shift to higher binding energies of the Tb

3d_5/2peaks after Li doping as well as the lower intensity of the

Li 1s peak in the XPS spectra obtained from the Li-TbO_x/n-

MgO compared with the Li-Sm₂O₃/n-MgO.

The O 1s and C 1s peaks obtained from the undoped and

lithium-doped REO-supported catalysts also contain valuable

information regarding these catalysts. Because some of O 1s

1982

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Figure 8. Narrow-scan XPS spectra of the (a) Li 1s and Mg 2p, (b) Sm 3d_5/2, and (c) Tb 3d_5/2binding energy regions obtained from undoped and

lithium-doped REO/n-MgO catalysts.

Figure 9. Narrow-scan XPS spectra of the (a) O 1s where the red trace is the Gaussian peak ﬁt of the collected data and (b) C 1s binding energy

regions obtained from fresh undoped and lithium doped REO/n-MgO catalysts, and (c) Tb 4d peaks obtained from fresh (black) and spent (blue)

Li-TbO_x/n-MgO catalysts.

Table 5. XPS Compositional Analysis of Selected Catalysts

a

b

catalyst

O 1s, %

Tb 3d_5/2, %

Sm 3d_5/2, %

Mg 2p, %

Li 1s, %

REO/Mg ratio

O/(REO+Mg) ratio

Li/REO ratio

O^II/O^I,

Sm₂O₃/n-MgO

Li-Sm₂O₃/n-MgO

TbO_x/n-MgO

56

68

64

65

8

2

36

21

31

23

0.22

0.10

0.16

0.26

1.27

3.09

1.78

2.24

0.49

2.48

1.04

3.86

9

6

4.5

1

5

6

Li-TbO_x/n-MgO

a

The following atomic sensitivity factors were used in the calculation of the compositions: O 1s, 0.711; Sm 3d_5/2= 2.907; Tb 3d_5/2= 3; Mg 2p =

0.12; Li 1s = 0.025.³⁶Peak area ratios of O 1s peaks from the peak ﬁtting (for details, please see the Experimental section). The O^Iis the O 1s peak

between 529.1 and 529.5 eV, which is due to the MgO support and the rare earth oxides. The O^IIpeak located between 530.9 and 531.5 eV is due to

hydroxyl groups and carbonates.

b

contribution is from carbonates (CO₂adsorbed on the surface

of these catalysts), it is useful to look at the O 1s and C 1s

peaks together (Figure 9a,b). The binding energy of the main

O 1s peak obtained from the Sm₂O₃/n-MgO catalyst is 529.3

eV, consistent with the values reported for MgO and

Sm₂O₃;³⁶⁻⁴⁵however, a signiﬁcant shoulder is present at

binding energies between 530.9 and 531.5 eV, revealing the

presence of surface hydroxyl and carbonate groups. The small

C 1s peak at 289.0 eV conﬁrms the presence of surface

carbonates.

Addition of lithium to the Sm₂O₃/n-MgO catalyst increases

the contribution from the state with an O 1s binding energy of

531.0 eV. The peak ﬁtting to deconvolute the contributions

from the metal oxides (MgO and Sm₂O₃oxygen species labeled

O^I) and the carbonates plus potential hydroxyl groups (oxygen

species labeled O^II) is shown in Figure 9 and reveals that the

1983

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

O^II/O^Iratio increases from 0.5 to 2.5 with the addition of

lithium (Table 5). The main contribution to the O^IIspecies

after addition of lithium is due to carbonates, as evident in the

high C 1s peak intensity at a binding energy of 289.2 eV.

Because CO₂is an acidic molecule that adsorbs onto basic sites,

adsorption of CO₂on the surface of these catalysts indicates the

presence of basic sites. This is important because it is believed

that basic oxygen sites are responsible for the hydrogen removal

from methane, the ﬁrst step in the oxidative coupling of

methane.³Because Li₂CO₃was observed in the XRD pattern

obtained from the Li-TbOx/n-MgO catalyst, it is possible that

Li₂CO₃is also present on the surface of the Li-Sm₂O₃/n-MgO

catalyst. However, in the case of the Li-Sm₂O₃/n-MgO catalyst,

the surface amount appears to be too small for detection with

XRD, or the Li₂CO₃on this catalyst is amorphous.

The O 1s peak obtained from the TbO_x/n-MgO catalyst has

a higher contribution from the state at 531.0 eV (O^II/O^Iratio

of 1.0) compared with the undoped Sm₂O₃/n-MgO catalyst

(O^II/O^Iratio of 0.5). This is due to a higher contribution from

adsorbed carbonates on the TbO_x/n-MgO catalyst (versus the

Sm₂O₃/n-MgO) as evident in the C 1s spectra. Similar to the

Sm₂O₃-based catalysts, addition of lithium signiﬁcantly

increases the contribution from the 531.0 eV state (Figure

9a). The O^II/O^Iratio increases by a factor of 4 after lithium

addition, and the majority of this increase is due to carbon

dioxide adsorption, as is evident in the higher intensity of the

289.2 eV C 1s carbonate peak.

3.3.5. Temperature-Programmed Desorption of CO₂.

Because a signiﬁcant amount of CO₂was observed with XPS

on the fresh doped and undoped REO/n-MgO catalysts, the

temperature-programmed desorption (TPD) of CO₂from the

TbO_x/n-MgO catalysts was investigated to determine the

strength of CO₂adsorption on these catalysts. TPD was carried

out after a CO₂dosing at 200 °C because CO₂dosing at 50 °C

did not yield suﬃcient chemisorption which was apparent by

poor signal-to-noise. Thus, it appears that the CO₂adsorption

is activated and requires a minimum temperature. A temper-

ature of 200 °C was chosen to ensure CO₂adsorption on the

surface while limiting CO₂desorption from the most weakly

bound sites before the TPD experiments. The n-MgO has a

small amount of basic sites, but most of them are weakly basic.

At a temperature of 600 °C, most of the CO₂has been removed

from the surface (Figure 10). The CO₂adsorption on the

The lithium peak intensity is very close to the noise level in

the spectrum obtained from the spent Li-TbO_x/n-MgO catalyst

(Figure 8a). This is consistent with the XRD data, which

revealed removal of the lithium-containing compound. Even

though most of the Li appears to have been removed after

reaction, the Li-doped TbO_x/n-MgO catalyst still has a

signiﬁcantly higher C₂₊selectivity compared with the undoped

catalyst, which suggests that either some residual lithium is still

on the surface, or the positive eﬀects of lithium are present even

after the lithium has been removed. The Tb 4d peaks obtained

from the spent Li-TbO_x/n-MgO catalyst reveal that the original

TbO_1.81phase is at least partially reduced (Figure 9 c). The

intensities of the Tb⁴⁺-related peaks are lower after reaction,

which is consistent with the formation of a Tb₂O₃phase as

observed with XRD after reaction. Thus, the XPS data indicate

that more CO₂is adsorbed on the TbO_x/n-MgO catalyst (O^II/

O^Iratio of 1.0) compared with the Sm₂O₃/n-MgO catalyst

(O^II/O^Iratio of 0.5), and the amount of CO₂adsorbed is

increased on both catalysts after doping with lithium, as

revealed by the O^II/O^Iratios for the two lithium-doped samples

(2.5 for Li-Sm₂O₃/n-MgO and 3.9 for Li-TbO_x/n-MgO).

Again, the higher contribution of adsorbed CO₂is due in part

to the presence of Li₂CO₃on the surface of the catalyst. A

larger contribution from the XPS peaks due to adsorbed CO₂

indicates a higher number of basic sites and indicates that the

Li-TbO_x/n-MgO is the most basic catalyst of the ones included

in this study. In addition, the TbO_x/n-MgO catalyst has a

higher number of basic sites compared with the Sm₂O₃/n-MgO

catalyst. This is somewhat unexpected because basic sites have

been shown to be important in the OCM reaction.²⁶If only the

number of basic sites were important, it would be expected that

the Sm₂O₃-based catalyst would have more adsorbed CO₂

because Sm₂O₃catalysts are more active than TbO_x-based

catalysts. Clearly, the character of the basic site, which would

aﬀect the strength of CO₂adsorption, is very important in this

reaction.

Figure 10. CO₂TPD data obtained from the n-MgO support and

selected n-MgO-supported catalysts after CO₂dosing at 200 °C for 30

min. The red line indicates the temperature ramp; green, n-MgO

support; brown, Sm₂O₃/n-MgO; blue, TbO_x/n-MgO; and black, Li-

TbO_x/n-MgO.

Sm₂O₃/n-MgO and TbO_x/n-MgO catalysts is higher than on

the n-MgO support; however, the adsorbed CO₂is completely

desorbed also on these catalysts by 500−600 °C (Figure 10).

Thus, the strength of CO₂adsorption is similar on the n-MgO

support and the REO/n-MgO catalysts, but more CO₂is

adsorbed in the presence of the REOs. In stark contrast, very

little CO₂is desorbed from the Li-TbO_x/n-MgO catalyst before

the temperature reaches 700 °C. The adsorption strength of

CO₂on the Li-TbO_x/n-MgO catalyst is clearly much higher,

and the amount of CO₂adsorbed is also higher than on the

TbO_x/n-MgO catalyst. This is consistent with the XPS data,

and on the basis of these data, the same trend is expected for

the Sm₂O₃/n-MgO catalysts; that is, that the Li doping

increases the amount and the strength of CO₂adsorption.

However, because of the volatility of lithium, which was clearly

observed as deposits on the instrument after the analysis of the

Li-TbO_x/n-MgO catalyst, it was decided not to run the Li-

Sm₂O₃/n-MgO or Li/n-MgO catalysts to protect the equip-

ment, and instead, it was decided to investigate the catalysts

using DRIFTS.

3.3.6. DRIFTS Experiments. To probe the nature of the

surface sites on the catalysts in this study (mainly the surface

oxygen), in situ DRIFTS experiments after CO₂adsorption at

two diﬀerent temperatures were performed on the fresh

1984

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Figure 11. DRIFTS background spectra collected after outgassing at 450 °C for reference compounds and lithium-doped and undoped REO/n-

MgO. Dashed lines mark positions for diﬀerent carbonate species: blue dashed line, monodentate carbonates on MgO; orange dashed line,

monodentate carbonate interactions with Li species;⁴⁸red dotted/dashed line, bidentate carbonates on Sm₂O₃; and pink dashed line, monodentate

carbonates on Sm₂O₃.⁴⁶

catalysts, the n-MgO support, the Li/n-MgO catalyst, and two

pure samaria and terbia nanoparticle (NP) samples (prepared

using the microemulsion method) (Figures 11−13. The nature

of the surface sites are important because surface oxygen is

responsible for the hydrogen abstraction from methane, the key

step of the OCM reaction.²The properties of the surface

oxygen species are also important because they can react with

the methyl radicals or the C₂₊products and form surface

methoxy species, which in turn can undergo oxidation to CO

and CO₂.

The CO₂molecules interact with the surface oxide and

hydroxide sites to form carbonate and bicarbonate species,

respectively. In-depth analysis of the DRIFTS spectra enables

us to distinguish between diﬀerent types of surface oxygens.⁴⁶

1985

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Figure 12. DRIFTS data obtained from n-MgO, Li/n-MgO as well as Sm₂O₃and TbO_xnanoparticles after CO₂adsorption at (a) 50 and (b) 250

°C. Solid gray line in part b is the DRIFTS data obtained after CO₂exposure at 50 °C. The dashed vertical lines indicate the following: (black)

bidentate carbonates on MgO, (blue) monodentate carbonates on MgO, (gray dotted/dashed) bicarbonates on MgO, (red dotted/dashed)

bidentate carbonates on Sm₂O₃, and (pink) monodentate carbonates on Sm₂O₃,⁴⁶and (orange) carbonate interactions with Li species.⁴⁸

At least four types of surface oxygens have been shown to exist

on the surface of magnesia, and these sites result in diﬀerent

carbonate species.⁴⁷The two main types of carbonates,

bidentate and monodentate carbonates, are present in the

1700−1200 cm⁻¹region as a result of CO₂interaction with

basic oxygen sites. The features in this region are due to bond-

stretching carbonyl (CO) vibrations of the diﬀerent

carbonate structures.

3.3.6.1. Background Spectra. Because the maximum

temperature of the heating mantle used in the DRIFTS

experiments is 450 °C, it is expected that most, if not all, of the

adsorbed CO₂will be removed from the TbO_x/n-MgO and

Sm₂O₃/n-MgO catalyst, and most of the adsorbed CO₂will

remain on the Li-TbO_x/n-MgO and Li-Sm₂O₃/n-MgO

catalysts after the 450 °C outgassing procedure. Because the

surface after outgassing at 450 °C for 30 min is used to remove

1986

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

Figure 13. DRIFTS data obtained from Sm₂O₃/n-MgO, Li-Sm₂O₃/n-MgO, TbO_x/n-MgO, and Li-TbO_x/n-MgO catalysts after CO₂adsorption at

(a) 50 and (b) 250 °C. Solid black spectra lines are for the REO/n-MgO samples, and the lighter gray lines shows either the n-MgO support

DRIFTS information or the Li-MgO DRIFTS spectra that the catalyst matches. The dashed vertical lines indicate the following: (black) bidentate

carbonates on MgO, (blue) monodentate carbonates on MgO, (gray dotted/dashed) bicarbonates on MgO, (red dotted/dashed) bidentate

carbonates on Sm₂O₃, (pink) monodentate carbonates on Sm₂O₃,⁴⁶and (orange) carbonate interactions with Li species.⁴⁸

background adsorption in the DRIFTS analysis, these spectra

with a KBr background subtraction are shown in Figure 11. The

amount of CO₂remaining on the surface of the n-MgO support

after outgassing is very low, and the same is true for the TbO_x

nanoparticles, as is evident in Figure 11. In contrast, some CO₂

remains on the surface of Sm₂O₃nanoparticles, and a signiﬁcant

amount of CO₂is present on the Li/n-MgO catalyst after heat

treatment at 450 °C. This suggests that the surfaces of the n-

MgO support and the TbO_xNPs do not have a signiﬁcant

amount of highly basic sites capable of strongly binding CO₂.

The Sm₂O₃NPs are more basic than the TbO_xNPs and the n-

MgO support, and the broad peaks reveal that both

monodentate and bidentate carbonates are present on the

surface. Compared with the Sm₂O₃NPs, more CO₂is bound to

1987

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

the surface of the Li/n-MgO, and the majority of the species on

the surface are monodentate carbonates associated with the

lithium.⁴⁸

concentration of monodentate species is signiﬁcantly increased

on the pure n-MgO support surface (Figures 12a,b). Because

the Li/n-MgO sample already has a large number of

monodentate species on the surface (higher than the amount

observed on the n-MgO after the 250 °C exposure), mainly the

bidentate carbonate species increase after the 250 °C CO₂

exposure. There are also new carbonate peaks on the catalyst’s

surface after doping with lithium. These peaks are most likely

due to carbonates chemisorbed on lithium because the peaks

match well with data obtained on a lithium-doped Al₂O₃-MgO

catalyst system.⁴⁸

As for the n-MgO support, a temperature of 250 °C appears

to be suﬃcient to remove the bicarbonates and yield more

strongly bound mono- and bidentate carbonate species. In both

cases, the amount of CO₂adsorbed on the surface is higher at

250 °C compared with adsorption at 50 °C. Some of the

bicarbonate species are removed after the 250 °C exposure also

on the TbO_xand Sm₂O₃NPs, but only on the TbO_xNPs is

there a slight increase in the mono- and bidentate carbonate

species compared with the 50 °C exposure. It appears that the

amount of CO₂adsorbed on the Sm₂O₃NPs at 250 °C is

smaller than the amount present after dosing at 50 °C. In

contrast, the n-MgO, Li/n-MgO, and TbO_xall have more CO₂

adsorbed on surface after CO₂exposure at the higher

temperature.

The DRIFTS spectra with a KBr background subtraction

obtained from the REO/n-MgO and Li-REO/n-MgO catalysts

after heat treatment at 450 °C are also shown in Figure 11. The

spectra from the Sm₂O₃/n-MgO and TbO_x/n-MgO catalysts

are similar and resemble the spectra obtained from the pure n-

MgO support. Thus, the amount of CO₂remaining on the

surface after outgassing at 450 °C for 30 min is very similar on

these catalysts and appears to be mainly due to the support. In

contrast, both Li-doped REO/n-MgO catalysts have a

signiﬁcant amount of carbonates present on the surface after

the heat treatment. This reveals strongly bound CO₂on the Li-

containing Sm₂O₃/n-MgO and TbO_x/n-MgO catalysts, which

is consistent with the TPD data (and the XPS measurements).

The amount of CO₂on the surface of these catalysts is higher

than on the surface of the Li/n-MgO, but the carbonate species

appear to be the same, that is, mainly monodentate lithium

carbonates. The DRIFTS data not only conﬁrm that CO₂is

strongly bound on Li-containing catalysts, it also corroborate

that a temperature of 450 °C is suﬃcient to remove most of the

CO₂adsorbed on the TbO_x/n-MgO and Sm₂O₃/n-MgO

catalysts.

The spectra in Figure 11 are used to remove the background

in the DRIFTS data obtained from each of the various catalysts

and reference samples after CO₂dosing at 50 and 250 °C.

3.3.6.1. Pure and Reference Components. Results of the

DRIFTS experiments on the n-MgO support, the Li/n-MgO

sample, and the two Sm₂O₃and TbO_xNP catalysts are shown

in Figure 12. The DRIFTS spectrum obtained from the

magnesia support after CO₂adsorption at 50 °C is the most

complex of the pure component spectra (Figure 12a), as would

be expected becaue it is known to have several diﬀerent basic

sites.⁴⁶In fact, all four carbonate species identiﬁed by Davydov

and co-workers⁴⁶bicarbonate, monodentate carbonate, and

two (asymmetric and symmetric) bidentate carbonatesare

present after CO₂adsorption at 50 °C (Figure 12a). Evidently,

all these species are relatively weakly bound, as most of them

would be removed below 500 °C, according to the TPD

measurements. Carbon dioxide adsorption at 50 °C on the Li/

n-MgO sample results in both bidentate carbonates and

bicarbonates, in addition to the monodentate carbonates that

were already present on the surface after outgassing (Figure

11). Comparing the n-MgO support and the Li/n-MgO, the n-

MgO support has a higher contribution from bidentate

carbonates and bicarbonates, and the monodentate carbonates

dominate on the Li/n-MgO.

CO₂adsorption on the Sm₂O₃NPs does not increase the

concentration of monodentate carbonate species; only

bidentate and bicarbonate species are increased after the CO₂

dose at 50 °C (Figure 12). On the TbO_xNPs that had very

little CO₂on the surface after the outgassing procedure, a small

amount of monodentate carbonates together with bidentate

carbonates and bicarbonates are present after the CO₂

exposure. The wavenumbers of the carbonates are close to

those obtained from the reference values for carbonates on

Sm₂O₃. The relative peak heights reveal that the concentration

of bidentate carbonates is higher than the concentration of

monodentate carbonates on the pure terbia nanoparticles.

As the temperature of the CO₂dosing is increased to 250 °C,

a smaller amount of bidentate carbonate species is present on

the surface, the bicarbonate species are removed, and the

3.3.6.2. Sm₂O₃/n-MgO and TbO_x/n-MgO Catalysts. Con-

sistent with the data obtained after outgassing at 450 °C

(Figure 11), the DRIFTS spectra obtained from the Sm₂O₃/n-

MgO and TbO_x/n-MgO catalysts are dominated by CO₂

adsorbed on the MgO support (Figure 13); however, compared

with the IR spectra obtained from the pure MgO support, more

CO₂is adsorbed on the REO/n-MgO catalysts (compare

Figures 12 and 13). More mono- and bidentate carbonate

species are observed on the surfaces of the Sm₂O₃/n-MgO and

TbO_x/n-MgO catalysts, compared with the n-MgO after CO₂

exposure at 50 °C. Because the CO₂adsorption is dominated

by the n-MgO support, the IR spectra obtained from the

Sm₂O₃/n-MgO and TbO_x/n-MgO catalysts are very similar,

despite the diﬀerences in the CO₂adsorption on the Sm₂O₃

and TbO_xNPs. Thus, on the supported catalysts, the

concentration of monodentate carbonates is higher than on

any of the pure components.

At a CO₂adsorption temperature of 250 °C, the spectra

obtained from both the Sm₂O₃/n-MgO and TbO_x/n-MgO

catalysts more closely resemble the spectrum obtained from n-

MgO compared with the data obtained after dosing at 50 °C

(Figure13a,b). This is due to the fact that the main species on

the surfaces of the n-MgO support and the REO/n-MgO

catalysts after this exposure are monodentate carbonates. The

main diﬀerence between the supports and the catalysts is that

the signal intensity is higher on the catalysts, that is, more

monodentate carbonates are present on the REO/n-MgO

catalysts compared with the n-MgO support. It appears that

slightly more CO₂is adsorbed on the TbO_x/n-MgO catalyst

compared with the Sm₂O₃/n-MgO catalyst; however, quanti-

ﬁcation is diﬃcult because the CO₂remaining on the surface

after outgassing is marginally higher on the Sm₂O₃/n-MgO

catalyst compared with the TbO_x/n-MgO catalyst.

3.3.6.3. Li-Sm₂O₃/n-MgO and Li-TbO_x/n-MgO Catalysts.

Adding Li to Sm₂O₃/n-MgO and TbO_x/n-MgO catalysts

results in a signiﬁcant increase in the strongly bound CO₂on

the surface, as evident in Figure 11. Considering the amount of

CO₂present on the surface in the form of monodentates, it is

1988

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

not surprising that the CO₂uptake on the Li-REO/n-MgO

catalysts is low (Figure 13). It appears that the surface is almost

saturated because very little CO₂can be added to the Li-

Sm₂O₃/n-MgO, irrespective of the dosing temperature (Figure

13a,b). In contrast, some bidentate carbonates are present on

the surface of the Li-TbO_x/n-MgO catalyst after dosing CO₂at

50 °C. Increasing the dosing temperature to 250 °C increases

both the monodentate and the bidentate carbonate concen-

trations on this catalyst.

ACKNOWLEDGMENTS

■

The authors gratefully acknowledge NSF support (Chemistry

1026712). Dr. Valentin Craciun, Dr. Eric Lambers, and Mr. Gill

Brubaker are acknowledged for the analytical equipment

training and technical assistance during XRD, XPS, and BET

surface area measurements. These analyses were performed at

the Major Analytical Instrumentation Center (MAIC) and the

Particle Engineering Research Center (PERC) at the University

of Florida (UF). Dr. Volkmar Zielasek (IAPC) and Dr.

Johannes Birkenstock (Geoscience Department) are gratefully

acknowledged for their assistance in collecting the TEM and

XRD data, and Dr. Kari Basso (UF Chemistry mass

spectrometry services), for assistance with MS analysis of the

condensate. We are also thankful for Mr. Xianhe Deng’s

assistance during the 30-h reaction experiment.

4. CONCLUSION

The reaction data obtained in this study reveal that

coimpregnating lithium and a terbia precursor yields a catalyst

that is highly active and selective in the oxidative coupling of

methane, despite the fact that terbia itself is not a selective

OCM catalyst. Addition of Li more than doubled the C₂₊yield

of the TbO_x/n-MgO catalysts, regardless of preparation

method, and the lithium nitrate precursor consistently gave a

higher C₂₊selectivity compared with the lithium carbonate

precursor. This increase in C₂₊yield with Li addition is due to

both an increase in methane conversion and an appreciably

higher C₂₊selectivity (TbO_xis typically highly selective to CO

and CO₂). The Li-TbO_x/n-MgO also exhibit a lower activity

loss with time on-stream compared with both Li/n-MgO and

Li-Sm₂O₃/n-MgO catalysts. In fact, the Li-TbO_x/n-MgO

catalyst exhibited less deactivation after 30 h on-stream than

the Li/MgO after just 8 h on-stream under the same

conditions. Furthermore, the Li-TbO_x/n-MgO catalyst is highly

active and selective at a temperature as low as 650 °C. The

lower OCM reaction temperature is an important factor in the

application toward an industrial process. Addition of Li to the

Sm₂O₃/n-MgO also increased the C₂₊yield, but the activity loss

is higher and temperature-dependent for this catalyst.

Li induces reduction of the TbO_xphase to Tb₂O₃, as

evidenced by both XRD and XPS, which is likely more active

toward C₂₊products than the TbO_x. Addition of Li also shifts

the Tb 3d core-level electrons to higher binding energies, which

is another indication of strong Li-TbO_xinteractions. This, in

turn, could explain the higher stability of the Li-TbO_x/n-MgO

compared with the other Li-containing catalysts. For both the

Sm₂O₃/n-MgO and the TbO_x/n-MgO catalysts, addition of Li

increased the number and strength of the basic sites. More CO₂

adsorption is observed on the Li-doped REO/n-MgO catalysts,

and the binding strength of the CO₂is higher compared with

the undoped catalysts. These highly basic sites may be required

for the hydrogen abstraction from methane to initiate the

coupling reaction. Although the activity and stability of the Li-

TbO_x/n-MgO catalyst system is lower than the Mn-Na₂WO₄/

SiO₂catalyst, the potential beneﬁt of using the Li-TbO_x/n-

MgO catalyst is the lower reaction temperature (600−700 °C

versus 800−875 °C for the Mn-Na₂WO₄/SiO₂system). Future

work on this system is therefore focused on utilizing other alkali

or alkali earth metals to address the issue of lithium

volatilization and ﬁnd a more stable REO-based catalyst.

REFERENCES

■

(1) Keller, G. E.; Bhasin, M. M. J. Catal. 1982, 73, 9−19.

(2) Ito, T.; J. Wang, J.; Lin, C.; Lunsford, J. H. J. Am. Chem. Soc.

1985, 107, 5062−5068.

(3) Driscoll, D. J.; Martir, W.; Wang, W.; Lunsford, J. H. J. Am. Chem.

Soc. 1985, 107, 58−63.

(4) Arndt, S.; Simon, U.; Heitz, S.; Berthold, A.; Beck, B.; Gorke, O.;

̈

Epping, J. D.; Otremba, T.; Aksu, Y.; Irran, E.; Laugel, G.; Driess, M.;

Schubert, H.; Schomacker, R. Top. Catal. 2011, 54, 1266−1285.

̈

(5) Mirodatos, C.; Perrichon, V.; Durupty, M. C.; Moral, P. Stud.

Surf. Sci. Catal. 1987, 34, 183−195.

(6) Korf, S. J.; Roos, J. A.; de Brujin, N. A.; van Ommen, J. G.; Ross,

J. R. H. Catal. Today 1988, 2, 535−545.

(7) Tang, L.; Yamaguchi, D.; Wong, L.; Burke, N.; Chiang, K. Catal.

Today 2011, 178, 172−180.

(8) Simon, U.; Arndt, S.; Otremba, T.; Schlingmann, T.; Gorke, O.;

̈

Dinse, K. P. Catal. Commun. 2012, 18, 132−136.

(9) Langfeld, K.; Frank, B.; Strempel, V. E.; Berger-Karin, C.;

Weinberg, G.; Kondratenko, E. V.; Schomacker, R. Appl. Catal., A

̈

2012, 417−418, 145−152.

(10) Wang, D.; Rosynek, M. P.; Lunsford, J. H. J. Catal. 1995, 151,

155−167.

(11) Campbell, K. D.; Lunsford, J. H. J. Phys. Chem. 1988, 92, 5792−

5796.

(12) Maksimov, N. G.; Selyutin, G. E.; Anshits, A. G.; Kondratenko,

E. V.; Roguleva, V. G. Catal. Today 1998, 42, 279−281.

(13) Leveles, L.; Fuchs, S.; Seshan, K.; Lercher, J. A.; Lefferts, L. Appl.

Catal., A 2002, 227, 287−297.

(14) Spinicci, R.; Marini, P.; De Rossi, S.; Faticanti, M.; Porta, P. J.

Mol. Catal. A: Chem. 2001, 176, 253−265.

(15) Machocki, A.; Jezior, R. Chem. Eng. J. 2008, 137, 643−652.

(16) Otsuka, K.; Jinno, K.; Morkawa, A. J. Catal. 1986, 100, 353−

359.

(17) Forlani, O.; Rossini, S. Mater. Chem. Phys. 1992, 31, 155−158.

(18) Zhang, H. B.; Lin, G. D.; Wan, H. L.; Liu, Y. D.; Weng, W. Z.;

Cai, J. X.; Shen, Y. F.; Tsai, K. R. Catal. Lett. 2001, 73, 141−147.

(19) Vereshchagin, S. N.; Ross, J. R. H. Catal. Today 1995, 24, 285−

287.

(20) Al-Zahrani, S. M.; Elbashir, N. O.; Abasaeed, A. E.; Abdulwahed,

M. J. Mol. Catal. A: Chem. 2004, 218, 179−186.

(21) Neumann, B.; Elkins, T. W.; Dreher, W.; Hagelin-Weaver, H.;

Nino, J. C.; Baumer, M. Catal. Sci. Technol. 2012, 3, 89−93.

̈

(22) Campbell, K. D.; Zhang, H.; Lunsford, J. H. J. Phys. Chem. 1988,

92, 750−753.

(23) Dedov, A. G.; Loktev, A. S.; Moiseev, I. I.; Aboukais, A.;

Lamonier, J. F.; Filimonov, I. N. Appl. Catal., A 2003, 245, 209−220.

(24) Deboy, J. M.; Hicks, R. F. J. Catal. 1988, 113, 517−524.

(25) Elkins, T. W.; Hagelin-Weaver, H. E. Appl. Catal., A 2013, 454,

100−114.

(26) Korf, S. J.; Roos, J. A.; Diphoorn, J. M.; J. Veehof, R. H.; van

Ommen, J. G.; Ross, J. R. H. Catal. Today 1989, 4, 279−292.

AUTHOR INFORMATION

■

Corresponding Author

*Phone: 352-392-6585. E-mail: hweaver@che.u ﬂ .edu.

Notes

The authors declare no competing ﬁnancial interest.

1989

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

ACS Catalysis

Research Article

(27) Gaffney, A. M.; Jones, C. A.; Leonard, J. J.; Sofranko, J. A. J.

Catal. 1988, 114, 422−432.

(28) Simon, U.; Gorke, O.; Berthold, A.; Arndt, S.; Schomacker, R.;

̈

Schubert, H. Chem. Eng. J. 2011, 168, 1352−1359.

(29) Pak, S.; Qiu, P.; Lunsford, J. H. J. Catal. 1998, 179, 222−230.

(30) Kiu, H. T.; Wang, X. L.; Yang, D. X.; Gao, R. X.; Wang, Z. L.;

Yang, J. J. Nat. Gas Chem. 2008, 17, 59−63.

(31) Arndt, S.; Otremba, T.; Simon, U.; Yildiz, M.; Schubert, H.;

Schomacker, R. Appl. Catal., A 2012, 425−426, 53−61.

(32) Yu, Z. Q.; Yang, X. M.; Lunsford, J. H.; Rosynek, M. P. J. Catal.

1995, 154, 163−173.

(33) Hutchings, G. J.; Scurrel, M. S.; Woodhouse, J. R. Catal. Today

1989, 4, 371−381.

(34) NanoScale Materials Inc.: http://nanoactive.interkan.net/

content/nanoactive_materials/nanoactive_home.asp, accessed on 10/

01/2008.

(35) Gilbert, R. E.; Cox, D. F.; Hoflund, G. B. Rev. Sci. Instrum. 1982,

53, 1281−1284.

(36) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. In

Handbook of X-ray Photoelectron Spectroscopy, 2nd ed.; Chastain, J.,

King, R. C. Jr., Eds.; Physical Electronics Inc.: Eden Prairie, MN, 1995.

(37) Wan, H. L.; Zhou, X. P.; Weng, W. Z.; Long, R. Q.; Chao, Z. S.;

Zhang, W. D.; Chen, M. S.; Luo, J. Z.; Zhou, S. Q. Catal. Today 1999,

51, 161−175.

(38) Choudhary, V. R.; Mulla, S. A. R.; Rane, V. H. J. Chem. Technol.

Biotechnol. 1998, 71, 167−172.

(39) Farsi, A.; Moradi, A.; Ghader, S.; Shadravan, V. Chin. J. Chem.

Phys. 2011, 24, 70−76.

(40) Jeon, W.; Lee, J. Y.; Leea, M.; Choi, J.-W.; Haa, J.-M.; Suha, D.

J.; Kim, I. W. Appl. Catal., A 2013, 464−465, 68−77.

(41) Papa, F.; Luminita, P.; Osiceanu, P.; Birjega, R.; Akane, M.;

Balint, I. J. Mol. Catal. A: Chem. 2011, 346, 46−54.

(42) Aritani, H.; Yamada, H.; Nishio, T.; Imamura, S.; Hasegawa, S.;

Tanaka, T.; Yoshida, S. Chem. Lett. 1999, 28, 359−360.

(43) Aritani, H.; Yamada, H.; Nishio, T.; Shiono, T.; Imamura, S.;

Kudo, M.; Hasegawa, S.; Tanaka, T.; Yoshida, S. J. Phys. Chem. B 2000,

104, 10133−10143.

(44) Esquivel, M. R.; Bohe,

2007, 42, 553−562.

(45) NIST Standard X-ray Photoelectron Spectroscopy Reference

Database 20, Version 4.1; http://srdata.nist.gov/xps/, accessed on 12/

16/2013.

́

A. E.; Pasquevich, D. M. Mater. Res. Bull.

(46) Davydov, A. In Molecular Spectroscopy of Oxide Catalyst Surfaces,

1st ed.; Sheppard, N. T., Ed.; John Wiley & Sons Ltd: Chichester,

2003.

(47) Davydov, A. A. Chem. Eng. Technol. 1995, 18, 7−11.

(48) Escobar-Alarcon, L.; Klimova, T.; Escobar-Aguilar, J.; Romero,

S.; Morales-Ramirez, C.; Solis-Casados, D. Fuel 2013, 110, 278−285.

1990

dx.doi.org/10.1021/cs500138j | ACS Catal. 2014, 4, 1972−1990

Article Doi

Effects of Li doping on MgO-supported Sm2O3 and TbOx catalysts in the oxidative coupling of methane

DOI: 10.1021/cs500138j

Source and publish data:

Authors:

Article abstract of DOI:10.1021/cs500138j

Full text of DOI:10.1021/cs500138j

Products guided by the article

R&D Labs maybe for 124-38-9

Relevant to this article

Hot Product