Oxidative coupling of methane over unsupported and alumina-supported samaria catalysts

Applied Catalysis A: General 454 (2013) 100–114

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General

journal homepage: w ww.elsevier.com/locate/apcata

Oxidative coupling of methane over unsupported and

alumina-supported samaria catalysts

Trenton W. Elkins, Helena E. Hagelin-Weaver^∗

Department of Chemical Engineering, University of Florida, Gainesville, FL, United States

a r t i c l e i n f o

a b s t r a c t

Article history:

Being the shortest unsaturated hydrocarbon, ethylene is a valuable feedstock gas for synthesizing longer

chain hydrocarbon products. Using the oxidative coupling of methane to produce ethylene and ethane

(C₂products) has been studied extensively over the past few decades. In this work, samaria nanoparticles

(NPs) and alumina-supported samaria catalysts were prepared using different methods (water/toluene

reverse microemulsion, metal-oleate high temperature decomposition, and incipient-wetness impreg-

nation followed by calcination) and various types of alumina supports (high and low surface area

nanoparticle alumina [n-Al₂O₃(+), n-Al₂O₃(−)] and a porous gamma alumina support [p-Al₂O₃]). The

highest product yields were obtained over Sm₂O₃NPs synthesized using the metal-oleate high temper-

ature decomposition and a nitrate precursor. Using a chloride precursor in the preparation of the Sm₂O₃

NPs resulted in less active and selective catalysts, and should be avoided. While the C₂selectivities were

lower over the Sm₂O₃/Al₂O₃catalysts, the yield per gram of Sm₂O₃were higher compared with the

Sm₂O₃NPs. The best supported catalyst was the Sm₂O₃/n-Al₂O₃(−) prepared using incipient wetness

impregnation of a nitrate precursor, since this leads to the smallest Sm₂O₃particles as well as the largest

coverage of the acidic Al₂O₃support. XRD analysis revealed that high-temperature calcinations form

SmAlO₃on the Al₂O₃-supported catalysts. While this reduced the near surface Sm₂O₃concentration, it

had beneﬁcial effects as it also lowered the Al₂O₃content. Therefore, Al₂O₃-supported Sm₂O₃warrants

further investigation, as surface modiﬁcations of the Al₂O₃can reduce its acidity and lead to higher C₂

selectivities in the oxidative coupling of methane over these catalysts.

Received 24 October 2012

Received in revised form 2 January 2013

Accepted 5 January 2013

Available online xxx

Keywords:

Samarium oxide nanoparticles

Alumina-supported samaria

Methane coupling activity and selectivity

XRD

XPS

Published by Elsevier B.V.

1. Introduction

conversion of ∼20–30% with the maximum observed C₂yield being

in the 20% range [6–10]. The reason for the low C₂yields can be

The conversion of ethylene to higher hydrocarbons is

a

attributed to the thermodynamics and kinetics of the OCM reac-

tion. Both the complete oxidation and the partial oxidation of

methane are exothermic reactions (ꢀH^◦= −890 KJ/mol for com-

plete oxidation vs. ꢀH^◦= −175 KJ/mol for C₂H₆formation, using

standard heats of formation). As the complete oxidation yields

products with a lower energy (CO_x) this is the thermodynamically

favored reaction pathway. The activation energies for the forma-

tion of ethane (E_a= 135 KJ/mol) and ethylene (E_a= 173 KJ/mol) are

considerably higher than the activation energy for the total com-

bustion of methane and ethane (E_a= 66 KJ/mol and E_a= 94 KJ/mol,

respectively) under the same conditions (700 ^◦C and CH₄:O₂= 20:1

using samaria as a catalyst [6]). Research has been focused on try-

ing to increase the activity in the oxidative coupling of methane

and decrease the oxidation rate to CO and CO₂products.

Several catalyst systems have been investigated in detail for

their use in OCM. These include for example, alkali earth metal

oxides catalysts [5,7,11–17], reducible non-transition metal oxides

[8,18–22], and rare-earth oxides (REOs), especially samarium oxide

(samaria) [2,9,10,23–33]. In particular, the lanthanides that form

sesquioxides (Me₂O₃, where Me is a metal ion) and do not form

frequently used industrial process, mainly for polymerization, oxi-

dation, and halogenation reactions [1–3]. Currently, ethylene is

synthesized using steam cracking of gaseous or light liquid hydro-

carbons [4]. Given the abundance of natural gas, however, there

is signiﬁcant interest in using natural gas as the feedstock for the

production of short hydrocarbons. Methane, the main component

(>90%) of natural gas, is one possible raw material. One of the ﬁrst

oxidative coupling of methane (OCM) investigations occurred in

the early 1980s [5] and it was discovered that methane (CH₄) and

oxygen (O₂) fed over a catalyst at elevated temperatures produce

the usual combustion products (CO₂, CO, H₂O,) along with ethane

(C₂H₆), ethylene (C₂H₄), and hydrogen (H₂). The impediment for

using OCM as an industrial process is the low reaction yield for

ethane and ethylene (C₂products). Generally, highly active cat-

alysts for OCM have a C₂selectivity less than 70% at a methane

∗

Corresponding author. Tel.: +1 352 392 6585; fax: +1 352 392 9513.

E-mail address: hweaver@che.uﬂ.edu (H.E. Hagelin-Weaver).

0926-860X/$ – see front matter. Published by Elsevier B.V.

http://dx.doi.org/10.1016/j.apcata.2013.01.010

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

101

higher oxides (such as MeO₂) are successful in converting CH₄to C₂

products. The high thermionic properties along with their natural

basicity contribute to their ability to promote the partial oxidation

pathway and therefore increase the C₂product yield [17]. Con-

ventional transition metal oxide and alkali metal oxide catalysts

typically favor the combustion of both the methane feedstock and

any formed C₂products, and thus are not viable catalyst systems to

study. In particular the irreducible rare earth oxides are active OCM

catalysts, and the major factors playing a role in determining the

reaction conditions. The catalysts were analyzed using BET surface

area measurements, X-ray diffraction analysis and X-ray photoelec-

tron spectroscopy to obtain their surface areas, crystal structures

and surface compositions, and determine if and how these catalysts

properties inﬂuence the catalytic activities and selectivities.

2. Experimental

2.1. Catalyst preparation

C₂selectivity for these REOs are the oxidation/reduction properties

and the oxygen mobility [34].

2.1.1. Samaria nanoparticle synthesis via metal-oleate

Of all the lanthanides, samaria has been the most studied OCM

catalyst, with the initial investigation performed by Otsuka [6]. This

study showed that the C₂rate of production for samaria was almost

3 times greater than the next best REO catalyst (lanthanum oxide)

and these results were later conﬁrmed in subsequent work [9].

Additionally, the reaction scheme for the catalytic surface interac-

tions that occur during C₂H₆production was formulated as shown

in Eqs. (1)–(3). The active site(s) is(are) regenerated through water

desorption from two neighboring Sm³⁺OH⁻(Eq. (4)).

decomposition

REO NPs were synthesized via decomposition of an REO oleate

complex using a slightly modiﬁed literature procedure [37]. Brieﬂy,

16 mmol of the REO precursor (samarium (III) nitrate or chloride,

99.9% REO, REacton, Alfa Aesar) and 48 mmol of sodium oleate

(received from TCI) were mixed together in a two-phase solution

consisting of 28 mL hexane (ACS grade, Fisher), 16 mL ethanol (his-

tological grade, Fisher) and 12 mL deionized (DI) water. The mixture

was stirred vigorously and was allowed to age at reﬂux (60 ^◦C) for

4 h. After reaction, the resulting samarium-oleate complex is in

the organic phase while the salts from the precursors are in the

aqueous phase. The samarium-oleate complex was isolated from

the aqueous phase by means of a separatory funnel and washed

with DI water three times in order to remove any residual salts.

Afterwards, the hexane was removed by a rotary evaporator. The

complex was then mixed with 50 mL of tri-n-octlyamine (TCI, 90%)

and 8 mmol of oleic acid (Alfa Aesar, tech. grade, 90%) for the high

temperature decomposition step. The heating program used for the

decomposition is as follows:

1

2

Sm³⁺V⁻(s) + O₂→ Sm³⁺O⁻(s)

(1)

Sm³⁺O⁻(s) + CH₄→ Sm³⁺OH⁻(s) + ^•CH₃

2^•CH₃→ C₂H₆

2Sm³⁺OH⁻(s) → Sm³⁺O⁻(s) + Sm³⁺V⁻(s) + H₂O

(2)

(3)

(4)

where V⁻indicates a surface oxygen vacancy (or basic site).

These sites, once occupied with a surface oxygen, are responsi-

ble for the hydrogen abstraction from the methane and form the

surface hydroxide plus the ^•CH₃radical. Ethane is formed by the

collision of the ^•CH₃radicals in the gas phase and the secondary

product, ethylene, is produced by the dehydrogenation of ethane

1. Ramp to 100 ^◦C at 8 ^◦C/min.

2. Hold at 100 ^◦C for 20 min (outgas step).

3. Ramp to 350 ^◦C at 4 ^◦C/min.

4. Hold at 350 ^◦C for 3 h.

−

[24]. Zhang et al. discovered the role of O₂as the most proba-

ble active oxygen species for non-reducible REO based catalysts

under OCM reaction conditions using Raman spectroscopy [10].

The supplementary in situ laser Raman spectroscopic study of the

active-oxygen species for a different catalyst supported this claim

[35,36].

After the decomposition step, the NPs were recovered by desta-

bilizing the suspension using ethanol. Sonication of the resulting

mixture for 20 min was done to remove most of the capping agent

remaining on the surface. The NPs were then collected by cen-

trifugation followed by redispersing them in hexane. This process

was repeated several times to maximize capping agent removal.

The samaria NPs were calcined at 800 ^◦C for 4 h unless otherwise

stated and are labeled Sm₂O₃(Cl)-OM and Sm₂O₃(N)-OM according

to Table 1.

In this study, we have investigated the effects of catalyst prepa-

ration method and the inﬂuence of various Al₂O₃supports, on

Sm₂O₃nanoparticles and Al₂O₃-supported Sm₂O₃catalysts in the

oxidative coupling of methane to obtain a better understanding

of the factors inﬂuencing the catalytic activity and selectivity of

Sm₂O₃-based OCM catalysts. While more active multi-component

catalysts have been reported in the literature, no additives or pro-

moters were added in this study to investigate in detail the effects

of the preparation method, the catalyst precursor and the Al₂O₃

support on the oxidative coupling of methane. Two Sm₂O₃precur-

sors (chloride and nitrate salts) were included in the study, and

the Sm₂O₃NP preparation methods investigated include thermal

decomposition of a samarium-oleate complex and a water/toluene

reverse microemulsion technique. The supported catalysts were

prepared using incipient wetness impregnation of the oleate NPs or

a samarium nitrate precursor, and samaria deposition using a mod-

iﬁed reverse microemulsion technique onto three Al₂O₃supports

with varying properties, a high surface area nanoparticle alumina

(n-Al₂O₃(+)) a low surface area nanoparticle alumina (n-Al₂O₃(−)),

and a conventional ␥-Al₂O₃porous catalyst support (p-Al₂O₃). The

effects of calcination temperature on the activity and selectiv-

ity of selected Sm₂O₃NPs and Sm₂O₃/Al₂O₃catalysts were also

investigated. Each catalyst was tested at three different CH₄:O₂

ratios (9:1, 7:1, and 4:1) as well as two reaction temperatures

(740 ^◦C and 800 ^◦C) using one ﬁxed volumetric feed ﬂow rate to

observe trends in the activity and selectivity between catalysts and

2.1.2. Samaria nanoparticle synthesis via microemulsion method

Samaria NPs were also synthesis via a modiﬁed previously

reported [38] reverse microemulsion method for comparison.

Using this method, 22.5 mL of DI water with the dissolved samaria

nitrate salt precursor (4.0 g) was mixed vigorously with 186 mL of

toluene. The amount of precursor was the amount that would yield

a 1.5-g batch of samarium oxide. The microemulsion surfactant,

Tween 80, was then added dropwise under vigorous stirring until

a microemulsion was formed (approximately 2–4 mL of Tween),

afterwards an extra 1–2 mL was added in order to prevent the aque-

ous sodium hydroxide solution from breaking the microemulsion.

15 mL of an aqueous sodium hydroxide solution was then added

dropwise to precipitate samarium hydroxide. 50% excess of the

sodium hydroxide was used in order to maximize the yield of the

samarium hydroxide. The microemulsion solution was allowed to

stir vigorously overnight and was then centrifuged to separate the

solid product from the liquid phase. The hydroxide NPs were then

dispersed in 200 mL of an ethanol/water mixture (50%, v/v). The

solution was sonicated for an hour and centrifuged again to collect

102

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

Table 1

BET surface areas obtained from the different Sm₂O₃nanoparticles and Sm₂O₃/Al₂O₃catalysts.

Catalyst description^a

Sm₂O₃Prep.^b

Catalyst Prep.^c

SSA^d[m²/g]

Particle size from BET [nm]

Particle size from XRD [nm]

Nanoparticles

Sm₂O₃(Cl)-OM

Sm₂O₃(N)-OM

OM

ME

OM

–

8.5

14.5

4.5

3.0

1.1

85

50

160

240

650

25

14

43

29

20

Sm₂O₃(N)-ME

Sm₂O₃(Cl)-OM^h

Sm₂O₃(N)-OM^h

Al₂O₃-Supported Sm₂O₃

Sm₂O₃/p-Al₂O₃-OM

Sm₂O₃/p-Al₂O₃-IM

Sm₂O₃/p-Al₂O₃-ME

Sm₂O₃/n-Al₂O₃(+)-OM

Sm₂O₃/n-Al₂O₃(−)-OM

Sm₂O₃/n-Al₂O₃(+)-IM

Sm₂O₃/n-Al₂O₃(−)-IM

Sm₂O₃/p-Al₂O₃-IM^h

Sm₂O₃/p-Al₂O₃-ME^h

Al₂O₃Supports^e

OM

–

OM

–

IM

ME

IM

ME

145

125

130

85

80

50

43

60

NH₃cm³/g ^f

9.0

8.3

1.6

CO₂cm³/g ^g

0.75

0.6

n-Al₂O₃(+)

–

695

275

260

n-Al₂O₃(−)

p-Al₂O₃

1.75

a

100% Sm₂O₃nanoparticles (NPs), or 20% Sm₂O₃by weight on an Al₂O₃support. (Cl): samarium chloride precursor used for nanoparticle (NP) preparation, and (N):

samarium nitrate precursor used for nanoparticle or catalyst preparation.

b

Sm₂O₃Prep.: OM = Sm₂O₃NPs prepared using the oleate method. ME = Sm₂O₃NPs prepared using reverse microemulsion method.

Catalysts prepared via the incipient wetness impregnation method (IM) using an aqueous solution of samarium nitrate (IM) or a hexane dispersion of OM nanoparticles

c

(OM), or via a modiﬁed reverse microemulsion method (ME).

d

Speciﬁc Surface Area (SSA) as determined by BET analysis. After calcination treatment at 800 ^◦C (or 1000 ^◦C) for the Sm₂O₃NPs and Sm₂O₃/Al₂O₃catalysts, or after drying

at 100 ^◦C for the Al₂O₃supports.

e

Al₂O₃supports; Al₂O₃(+) and Al₂O₃(−) from NanoScale Inc. and ␥-Al₂O₃from Alfa Aesar.

f

NH₃adsorption (standard cm³/g) to determine acidic sites.

g

CO₂adsorption (standard cm³/g) to determine basic sites.

h

Catalysts calcined at 1000 ^◦C before reaction instead of 800 ^◦

C

the ﬁnal product. The NPs were dried overnight at 105 ^◦C before

calcination.

The third catalyst was prepared using a heterogeneous reverse

microemulsion methodology which has been previously reported

in the lab [39]. Modiﬁcations to the procedure were made in

order to be consistent with the procedure for the pure samaria

NP microemulsion experiments. Brieﬂy, 3.4 g of p-Al₂O₃sup-

port was suspended in 22.5 mL DI water prior to the addition

of 2.2 g samarium nitrate salt. The resulting mixture was then

added to 186 mL toluene under vigorous stirring. The addition of

the surfactant and NaOH solution was the same as for the pure

Sm₂O₃microemulsion method. After collection, half of the cata-

lyst was calcined at 800 ^◦C while the other half was calcined at

1000 ^◦C.

2.1.3. Preparation of Al₂O₃-supported Sm₂O₃

Three different types of supported Sm₂O₃catalysts were pre-

pared using two main methods of Sm₂O₃deposition onto the

support, namely, incipient wetness impregnation and a reverse

microemulsion technique. Furthermore, three different alumina

supports were used in the synthesis of the supported catalysts, the

conventional p-Al₂O₃catalyst support from Alfa Aesar, the high

surface area n-Al₂O₃(+) from NanoActive, and the lower surface

area n-Al₂O₃(−) also from NanoActive. In all cases, the samar-

ium nitrate precursor was used to facilitate comparisons. Two

sets of catalysts were prepared using incipient wetness impreg-

nation (IWI) of either an aqueous solution of samarium nitrate

or a hexane dispersion of the samaria NPs formed via the metal

oleate methodology. The third type of catalyst was prepared using

a reverse microemulsion technique to deposit the samaria onto the

support.

For loading the samarium nitrate salt using the IWI method, the

precursor was dissolved in a volume of DI water equal to that of

the catalysts support pore volume, or the minimum amount of sol-

vent which would completely dissolve the precursor, and was then

mixed with the catalyst support until incipient wetness. The sec-

ond IWI method used the samaria NPs prepared using the oleate

method and the samarium nitrate precursor. As post-calcination

NPs, with the capping agent removed, did not disperse in the sol-

vent used, the recovered samaria NPs were dispersed in hexane

before the calcination step. After each loading step the catalyst was

dried to remove the hexane/water from the catalyst pores, and the

loading step was repeated until the desired loading concentration

was obtained (20% Sm₂O₃by weight on the support). After the ﬁnal

impregnation step (usually the 2nd or 3rd impregnation), the cat-

alysts were dried over night at 105 ^◦C and calcined at 800 ^◦C for 4 h

unless stated otherwise.

2.2. Characterization methods

Brunauer–Emmett–Teller (BET) surface area measurements

for all catalysts were performed using a 6-point isotherm on

a Quantrachrome NOVA 1200 instrument operating at liquid

nitrogen temperatures. Particle size calculation based on the

BET surface areas are calculated using Eq. (5). The calculation

is based on the assumption of having monodisperse, spherical

particles.

6

d_p

=

(5)

SA · ꢁ_p

where SA is the speciﬁc surface area [m²/g] and ꢁ_p[g/m³] is the

particle density.

X-ray diffraction (XRD) patterns for select catalysts were col-

lected on a Phillips APD 3720 Xray diffractometer using Cu K␣

˚

radiation (ꢂ = 1.54 A). Catalyst powders were pressed onto double

sided sticky tape and secured on a glass slide. An average crystallite

size was calculated using the Scherrer equation (Eq. (6)).

Kꢂ

d_p

=

(6)

ˇ cos(ꢃ)

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

103

In the Scherrer equation, d_pis the crystallite size (nm), K is the

shape factor (taken as unity), ꢂ is the Cu K␣ radiation wavelength

(nm), ˇ is the peak width at half the maximum intensity (FWHM)

in radians, and ꢃ is the Bragg angle, which is half of the diffraction

angle (2ꢃ).

TEM images were collected using a JEOL 2010F instrument using

an accelerating voltage of 200 kV. Samples were dispersed in hex-

ane and sonicated for 15 min before being loaded onto a 400 mesh

copper grid with a carbon ﬁlm. Several TEM images were recorded

at various locations on the sample and at different magniﬁcations

to conﬁrm that the structures observed were present throughout

the catalyst. For each catalyst representative images are presented

in the ﬁgures.

XPS measurements were performed on a PerkinElmer 5100 XPS

system using an aluminum x-ray source. Samples were dried in an

oven overnight at 105 ^◦C to minimize the water adsorption on the

catalyst surface. For survey scans the time/step was 30 ms with a

0.5 eV step size, and a pass energy of 89.45 eV. For high resolution

(narrow) scans, the time/step was 50 ms, the step size 0.1 eV, and

the pass energy 35.75 eV. Both survey and high-resolution spectra

were collected using a total of 10 scans. Samples were prepared by

placing a thin layer of catalyst on double-sided carbon conductive

tape. Scan shifts were taken into account by referencing to the C

1s peak at 284.6 eV. The Si 2p peak at 99 eV, observed on all three

catalysts, is due to the sample holder or the instrument rather than

the sample. This was conﬁrmed in a different XPS instrument which

uses an aluminum cup as the sample holder (rather than the car-

bon conductive tape) and no Si peak was detected in the spectra

obtained in this system.

ice-water bath. In addition to the CH₄conversions (X_CH₄) deﬁned

as the methane reacted over the total amount of methane fed t o_∗the

reactor (Eq. (7)), a CH₄conversion to C₂and CO_xproducts (X_CH

)

4

was deﬁned and calculated according to Eq. (8). This was done to

obtain an estimate of the coking in the reactions. The C₂selectivity

was deﬁned as the methane reacted to form C₂products over the

methane reacted to C₂plus CO_xproducts (Eq. (9)) and is reported

as a percentage. C₂product yield (Eq. (10)) was calculated as the

methane conversion (X_C^∗_H) (according to Eq. (8)) multiplied by the

4

selectivity (Eq. (9)). Repeat reaction experiments involving reload-

ing the catalyst resulted in standard deviations of 0.5% for both the

CH₄conversion and the C₂selectivity.

CH_4,in− CH_4,out

X_CH

=

(7)

(8)

4

CH_4,in

[2 · C H₆+ C₂H₄+ CO + CO ]

(

)

2

out

X_C^∗_H

4

CH_4,in

sccm C₂products

sccm CH₄reacted to C₂and CO_x

S_C

=

(9)

2

C₂Yield [%] = X_C^∗_H· S_C· 100

(10)

2

4

3. Results and discussion

The prepared Sm₂O₃catalysts were subjected to methane cou-

pling experiments and catalyst characterizations.

3.1. OCM activity measurements

The catalytic activity and selectivity of all catalysts in the oxida-

tive coupling of methane to C₂hydrocarbons were evaluated at

the two different temperatures, 740 ^◦C and 800 ^◦C. The results are

presented in Tables 2 and 3 and Figs. 1–5.

2.3. OCM reaction experiments

Catalytic activity measurements were carried out in a quartz

tube reactor with an inner diameter of 10 mm. The quartz tube

reactor was used in a previously reported reactor system, which

was slightly modiﬁed for the current application [40]. Each cat-

alyst was pressed into a pellet, using a Carver pellet press, then

crushed and sieved to a size range of 180–250 microns. For load-

ing the reactor, 0.4 g of the sieved catalysts (either Sm₂O₃NPs

or Al₂O₃-supported Sm₂O₃) was supported in the quartz reactor

tube between two pieces of quartz wool. CH₄, O₂, and an inter-

nal standard (N₂) were fed through the system at a rate of 120

standard cm³per minute (sccm) (with N₂constant at 23.2 sccm)

using three mass ﬂow controllers (MFCs). This results in a GHSV

of 1760 h⁻¹for the supported catalysts and 880 h⁻¹for the pure

samaria NPs. Three different CH₄:O₂ratios (9:1, 7:1, and 4:1) at

two temperature settings (740 and 800 ^◦C) were used to observe

the effects of the reaction parameters on the activity, selectivity and

product distribution in the oxidative coupling of methane (OCM).

Product analysis was performed on a custom-conﬁgured on-line

Agilent 6890 N gas chromatograph (GC) equipped with both a ther-

mal conductivity detector (TCD) and a ﬂame ionization detector

(FID) in series along with two packed columns, a polar Porapak Q

capillary column and a molecular sieve, also in series with a col-

umn isolation valve in between the columns. The former was used

to separate the CO₂, C₂H₆, and C₂H₄and the latter for CH₄, O₂, N₂,

CO, H₂separation. Calibration curves for product formation and

methane conversion were taken by feeding a known ﬂow rate of

the calibration gas and 23.2 sccm of internal standard (N₂) to the

GC and measuring the calibration gas to N₂TCD peak area ratio. All

calibration curves were linear and regression analysis resulted in ﬁt

values of >99.3%. Measurements were taken at each reaction condi-

tion after 20 min to allow for the reaction to reach steady state and

a repeat was taken to ensure reproducibility. The water produced

during the reaction was separated from the gas phase products

prior to entering the GC using a condenser trap emerged in an

3.1.1. Pure samaria nanoparticles

The samaria NPs were synthesized via the oleate-method using

chloride and nitrate salt precursors and via the reverse microemul-

sion using the nitrate salt to determine the effects of samaria

precursor and preparation method on the Sm₂O₃particle size

and how these parameters affect the OCM activity and selectivity.

Fig. 1A and B display the OCM reaction data for the pure samaria

NPs from the oleate and microemulsion methods at reaction tem-

peratures of 740 ^◦C and 800 ^◦C, respectively. The results are also

summarized in Tables 2 and 3. A few general trends, which are the

same for most catalysts, can be observed from the reaction results

(Fig. 1A and B plus Tables 2 and 3). In most cases, the methane con-

version increases and the selectivity decreases as the CH₄:O₂ratio

decreases from 9:1 to 4:1, i.e. more CH₄reacts but a larger fraction is

converted to CO and CO₂at the higher O₂concentrations. This is to

be expected and is the reason why CH₄concentrations signiﬁcantly

higher than the stoichiometric CH₄:O₂ratio for ethane or ethylene

formation is often used in this reaction [17]. However, since the

decrease in selectivity is smaller than the increase in conversion

with decreasing CH₄:O₂ratio, the highest C₂yields are obtained

at the 4:1 CH₄:O₂ratio for most Sm₂O₃NP catalysts (Table 2).

As can be seen in Table 2, most of the methane reacted over the

Sm₂O₃nanoparticles forms C₂or CO_xproducts. The largest differ-

ences between X_CHand X_C^∗_Hare observed at a CH₄:O₂ratio of 4:1,

4

which at ﬁrst appears unus⁴ual as coking would not be expected

at the highest O₂concentrations. However, the C₂yield is higher

at the lower CH₄:O₂ratios, which can increase coke formation as

the C₂products decompose easier than CH₄. Furthermore, coking

is more likely at higher temperatures, and the increased conver-

sion, in particular to CO_x, will result in more heat released from the

exothermic reactions. This is evident as a larger difference between

104

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

Table 2

Oxidative coupling of methane over all catalysts at 740 ^◦C and 800 ^◦C.

Catalyst^a

CH₄:O₂ratio

T = 740 ^◦

C

T = 800 ^◦

C

X_CH[%]

X_C^∗_H[%]

S_C[%]

Y_C[%]

X_CH[%]

X_C^∗_H[%]

S_C[%]

Y_C[%]

4

2

4

2

4

Sm₂O₃nanoparticles

Sm₂O₃(Cl)-OM

9

7

4

9

7

4

9

7

4

7.5

10.5

18.5

10.0

12.5

19.5

10.5

13.0

17.5

7.5

10.5

17.5

10.0

12.0

18.0

10.0

12.0

16.0

44.0

43.5

58.5

55.5

43.5

50.5

52.5

30.0

3.4

4.6

7.6

6.0

6.7

7.8

5.1

6.3

4.8

8.5

12.5

19.0

10.0

12.0

19.0

8.5

12.5

17.5

10.0

11.5

17.0

8.5

51.0

47.0

41.0

57.0

53.5

40.0

37.5

47.5

28.0

4.4

6.0

7.2

5.8

6.2

6.8

3.2

5.2

4.3

Sm₂O₃(N)-OM

Sm₂O₃(N)-ME

13.0

18.0

11.0

15.5

Sm₂O₃/Al₂O₃catalysts

Sm₂O₃/p-Al₂O₃-OM

9

7

4

9

7

4

9

7

4

6.0

7.5

15.0

6.0

8.5

16.5

8.5

10.5

20.5

16.0

14.0

14.5

18.5

16.5

19.5

18.5

19.0

5.0

7.0

12.5

6.0

8.5

14.0

7.0

6.0

7.5

0.3

0.5

1.3

1.7

2.9

1.4

1.6

3.8

0.9

1.1

1.8

2.5

3.4

5.1

5.0

3.6

6.5

8.5

19.0

7.5

10.0

18.5

9.0

12.0

21.5

–

6.0

7.5

14.5

7.5

9.0

15.0

7.5

9.5

16.5

–

18.0

20.0

24.5

34.0

30.0

28.0

28.5

27.5

–

1.1

1.5

3.6

2.6

2.7

4.2

2.1

2.7

4.5

–

10.5

22.0

20.5

21.0

19.5

19.0

24.5

8.5

10.0

15.0

17.5

23.5

32.0

31.0

24.0

Sm₂O₃/p-Al₂O₃-ME

Sm₂O₃/p-Al₂O₃-IM

8.5

15.5

10.5

11.5

12.0

14.0

14.5

16.0

15.0

p-Al₂O₃

n-Al₂O₃(−)

–

Sm₂O₃/n-Al₂O₃(+)-OM

Sm₂O₃/n-Al₂O₃(+)-IM

Sm₂O₃/n-Al₂O₃(−)-OM

Sm₂O₃/n-Al₂O₃(−)-IM

Sm₂O₃/p-Al₂O₃-ME^b

Sm₂O₃/p-Al₂O₃-IM^b

18.5

20.5

17.0

20.0

19.5

21.5

13.5

15.5

14.5

16.0

15.5

21.0

24.5

23.5

35.0

30.5

27.5

2.8

3.8

3.4

5.6

4.9

4.3

Catalyst descriptions are given in Table 1. X_CH₄: methane conversion [%], X^∗: methane conversion to C₂and CO_xproducts [%] S_C, C₂: selectivity, i.e. methane reacted

a

2

CH

4

to C₂products over methane reacted to C₂and CO_xproducts [%], Y_C: C₂yield (X^∗· S_C₂) [%]

2

CH

4

Catalysts calcined at 1000 ^◦C before reaction instead of 800 ^◦C.

b

Table 3

Product distribution data for the oxidative coupling of methane over all catalysts at 740 ^◦C and 800 ^◦C.

Catalyst^a

CH₄:O₂ratio

Product distribution [%] at 740 ^◦

C

Product distribution [%] at 800 ^◦

C

C₂H₄

C₂H₆

CO₂

CO

C₂H₄

C₂H₆

CO₂

CO

Sm₂O₃nanoparticles

Sm₂O₃(Cl)-OM

9

7

4

9

7

4

9

7

4

17.4

19.3

22.7

25.5

25.4

22.9

23.0

23.9

15.3

26.6

24.7

20.6

33.2

30.0

20.7

27.6

28.7

14.5

44.7

46.4

50.3

35.8

38.6

49.2

41.8

39.4

60.3

11.3

9.7

6.5

5.5

6.0

7.2

7.6

8.0

10.0

29.7

29.8

28.3

36.0

34.7

27.6

21.5

29.4

19.2

21.4

17.3

12.9

21.0

18.9

12.3

16.0

18.3

8.7

40.3

44.9

51.1

34.8

38.0

50.4

49.5

42.4

59.5

8.6

8.0

7.7

8.2

8.4

Sm₂O₃(N)-OM

Sm₂O₃(N)-ME

9.7

13.0

10.0

12.6

Sm₂O₃/Al₂O₃catalysts

Sm₂O₃/p-Al₂O₃-OM

9

7

4

9

7

4

9

7

4

0.8

1.5

4.5

5.8

9.2

7.1

6.5

13.7

2.4

5.4

5.8

6.1

15.9

14.5

11.8

12.4

12.5

10.8

6.1

62.1

61.4

59.7

55.4

57.5

60.7

51.5

52.5

53.5

52.8

78.6

62.2

59.5

54.7

49.9

55.4

53.6

31.7

31.3

29.7

22.9

22.2

18.3

29.0

28.5

22.0

38.6

11.2

23.0

16.9

28.0

18.0

13.6

22.2

8.7

11.0

18.3

19.8

16.4

19.3

15.9

17.8

20.1

–

9.4

8.7

6.3

14.3

13.5

8.8

12.4

10.9

7.6

–

47.5

43.7

46.5

50.1

53.4

43.8

45.2

50.0

–

34.4

32.7

31.7

19.4

19.9

18.6

27.9

26.1

22.4

–

Sm₂O₃/p-Al₂O₃-ME

Sm₂O₃/p-Al₂O₃-IM

p-Al₂O₃

n-Al₂O₃(−)

1.7

7.8

8.5

7.0

–

Sm₂O₃/n-Al₂O₃(+)-OM

Sm₂O₃/n-Al₂O₃(−)-OM

Sm₂O₃/n-Al₂O₃(+)-IM

Sm₂O₃/n-Al₂O₃(−)-IM

Sm₂O₃/p-Al₂O₃-ME^b

Sm₂O₃/p-Al₂O₃-IM^b

14.8

15.2

17.2

24.1

21.5

19.8

6.4

8.2

7.3

10.9

9.0

7.7

49.4

59.6

49.1

48.5

52.3

44.9

29.5

17.1

26.4

16.5

17.2

27.6

13.2

8.4

17.4

17.0

12.0

10.4

9.0

14.7

14.0

12.2

a

Catalyst descriptions are given in Table 1.

Catalysts calcined at 1000 ^◦C before reaction instead of 800 ^◦C.

b

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

105

70

60

50

40

30

20

10

0

70

A

T = 700°C

T = 800°C

60

50

9:1

7:1

40

30

20

10

0

4:1

Sm₂O₃(Cl)-OM T800

Sm₂O₃(Cl)-OM T1000

Sm₂O₃(N)-OM T800

Sm₂O₃(N)-OM T1000

Sm₂O₃(Cl)-OM

Sm₂O₃(N)-OM

Sm₂O₃(N)-ME

0

5

10

15

20

25

0

5

10

15

20

25

CH₄Conversion [%]

70

60

50

40

30

20

10

0

70

60

50

40

30

20

10

0

B

T = 800°C

Sm₂O₃/p-Al₂O₃-IM T800

Sm₂O₃/p-Al₂O₃-IM T1000

Sm₂O₃/p-Al₂O₃-ME T800

Sm₂O₃/p-Al₂O₃-ME T1000

9:1

7:1

4:1

9:1

7:1

4:1

Sm₂O₃(Cl)-OM

Sm₂O₃(N)-OM

Sm₂O₃(N)-ME

0

5

10

15

20

25

0

5

10

15

20

25

CH₄Conversion [%]

Fig. 2. Effects of calcination temperature between 800 ^◦C and 1000 ^◦C in the OCM

reaction data for (A) Sm₂O₃nanoparticles prepared using the oleate decomposition

method with nitrate and chloride precursors and (B) Sm₂O₃/p-Al₂O₃catalysts pre-

pared using incipient wetness impregnation with an aqueous solution of samarium

nitrate and the reverse microemulsion method. The data was collected at a reaction

temperature of 800 ^◦C and at different CH₄:O₂feed ratios, i.e. 4:1, 7:1 and 9:1 (as

indicated by the numbers inside the boxes).

Fig. 1. OCM reaction data for Sm₂O₃nanoparticles prepared using the oleate

decomposition with nitrate and chloride precursors and the reverse microemul-

sion method using a samarium nitrate precursor. The numbers in the boxes indicate

the CH₄:O₂feed ratios, i.e. 4:1, 7:1 and 9:1. (A) reaction temperature of 740 ^◦C and

(B) reaction temperature of 800 ^◦C.

the X_CHand X_C^∗_Hat 800 ^◦C versus 740 ^◦C (Table 2). Coking for all

4

Sm₂O₃NP catalysts was observed on the reactor walls downstream

of the catalyst bed and not on the Sm₂O₃NPs. This is in an oxygen-

deﬁcient region of the reactor, as the O₂conversion in all cases is

at least close to 100%, and indicates that coking over the Sm₂O₃is

negligible.

and (12)). It is interesting to note that at each CH₄:O₂ratio the

C₂H₄:C₂H₆product ratio is the same for the Sm₂O₃NPs regardless

of preparation method and precursor, and this ratio increases from

0.8 to 1.1 as the CH₄:O₂ratio is decreased from 9:1 to 4:1. This

suggests that the formation of the secondary C₂H₄product is not

dependent on the Sm₂O₃surface over these NPs. This observation

is supported by an earlier OCM study over 1% Sr/La₂O₃[24].

The CH₄:O₂ratio also affects the product distribution as can

be seen in Table 3. As expected, the CO₂content in the product

stream increases with increasing O₂concentration. The change in

CO concentration with CH₄:O₂ratio is not as signiﬁcant, and the

trend depends on the speciﬁc Sm₂O₃NPs. The CO concentration

increases slightly over the Sm₂O₃(N)-OM and Sm₂O₃(N)-ME cata-

lysts, while it decreases with increasing O₂concentration over the

Sm₂O₃(Cl)-OM catalyst. Therefore, the CO₂-to-CO product ratio

nearly doubles over the Sm₂O₃(Cl)-OM catalyst, while there is only

a slight increase (if any) in this ratio with decreasing CH₄:O₂ratio

for the Sm₂O₃(N)-OM and Sm₂O₃(N)-ME catalysts. At 740 ^◦C and

the higher CH₄:O₂ratios (9:1 and 7:1) the C₂H₄:C₂H₆product ratio

is below one, i.e. more C₂H₆than C₂H₄is formed, while at the high-

est O₂concentration (CH₄:O₂ratio = 4:1) more C₂H₄is formed than

C₂H₆over all Sm₂O₃NP catalysts. This is consistent with the trend

expected from the stoichiometry, i.e. more oxygen (lower CH₄:O₂

ratio) is required for production of C₂H₄versus C₂H₆(Eqs. (11)

2CH₄+ 1/2O₂→ C₂H₆+H₂O

2CH₄+ O₂→ C₂H₄+ 2H₂O

(11)

(12)

The effects of reaction temperature are different dependent on

whether a chloride or nitrate precursor was used in the prepara-

tion of the Sm₂O₃NPs. While the CH₄conversion is not altered

signiﬁcantly, the C₂selectivity decrease with increasing reaction

temperature (between 740 ^◦C and 800 ^◦C) for the Sm₂O₃NPs pre-

pared using nitrate precursors, irrespective of preparation method

(oleate method or reverse microemulsion). The C₂yields are there-

fore lower at 800 ^◦C versus 740 ^◦C (Table 2). In contrast, the CH₄

conversion and the C₂selectivity both increase with increasing

reaction temperature for the Sm₂O₃(Cl)-OM nanoparticles, for all

except the 4:1 CH₄:O₂ratio (which exhibit a slight decrease in C₂

106

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

selectivity). However, the C₂yields for the Sm₂O₃(Cl)-OM NPs at

800 ^◦C are still lower than for the Sm₂O₃(N)-OM NPs at 740 ^◦C.

The C₂H₄:C₂H₆product ratio is signiﬁcantly higher at 800 ^◦C

than at 740 ^◦C. In fact, at a reaction temperature of 800 ^◦C more C₂H₄

than C₂H₆product is formed at all CH₄:O₂ratios over all Sm₂O₃

NPs. Over the Sm₂O₃(N)-OM NP catalyst, the CO₂content in the

product stream is similar at both 740 ^◦C and 800 ^◦C, but the CO

content is higher at 800 ^◦C, and a higher CO content at the higher

reaction temperature is observed also over the Sm₂O₃(N)-ME NP

catalyst. This is expected since more CO would have enough energy

to desorb from the catalyst surface before undergoing oxidation at

the higher temperature. At 740 ^◦C the CO is more strongly bound to

the surface, and has a higher probability of being oxidized to CO₂

before desorbing. The lower C₂selectivities at 800 ^◦C thus appear

to be mainly due to a higher CO formation rate for these catalysts.

From the reaction data, it is evident that the samaria NPs

prepared using the oleate method with the nitrate precur-

sor outperform both the Sm₂O₃(Cl)-OM and the Sm₂O₃(N)-ME

nanoparticles, since the Sm₂O₃(N)-OM catalyst at most conditions

exhibits both the highest CH₄conversion and the highest C₂selec-

tivity (Fig. 1A and B). This is true for all CH₄:O₂ratios at 740 ^◦C and

most CH₄:O₂ratios at 800 ^◦C (Table 2). Only at 800 ^◦C and a CH₄:O₂

ratio of 4:1 does the Sm₂O₃(Cl)-OM give a higher C₂yield than the

Sm₂O₃(N)-OM catalyst (although the Sm₂O₃(Cl)-OM yield at 800 ^◦C

(7.2%) is not higher than the C₂yield obtained over the Sm₂O₃(N)-

OM catalyst at 740 ^◦C (7.8%)). While a yield of 8% is not the highest

reported in the literature, it is in line with the yields obtained over

pure Sm₂O₃[9]. Higher yields can be obtained at even lower CH₄:O₂

ratios, such as a 12% yield at a CH₄:O₂ratio of 2.5 [28], but ratios

below CH₄:O₂= 4:1 were avoided in the current investigation for

safety reasons, i.e. to make sure that the feed stream had methane

concentrations above the upper ﬂammability limit.

these catalysts are presented in Fig. 3A and B and have been

included in Tables 2 and 3.

Compared to the unsupported Sm₂O₃NP results, the perfor-

mance of the alumina-supported Sm₂O₃is inferior. Not only are

the C₂selectivities signiﬁcantly lower at all temperatures and

CH₄:O₂ratios, the CH₄conversions are also slightly lower for most

Sm₂O₃/p-Al₂O₃catalysts. A lower conversion is expected, since

the amount of Sm₂O₃in the reactor for the alumina-supported

catalysts is only 20% of the amount used for the pure NP Sm₂O₃

reactions. The acidic alumina support is also expected to yield a

much lower C₂selectivity than that observed for pure Sm₂O₃[41].

Moreover, compared to the Sm₂O₃NP catalysts, the X_C^∗_H/X_CHis

4

lower on the Al₂O₃-supported catalysts, particularly at the 4:1

CH₄:O₂ratio. This means that sligthly more coking takes place

over the Al₂O₃-containing catalysts and this was also observed as

a change in color of the catalysts from white to black, as well as a

small amount of black deposits on the reactor walls downstream of

the catalyst bed. Despite the presence of coking, no decrease in CH₄

conversion as a function of time was observed during the experi-

ments, which indicates that the amount of coking was not sufﬁcient

to block the active surface sites. This is consistent with coking

mainly occurring on the Al₂O₃support, and not on the Sm₂O₃(as

inferred from the Sm₂O₃NP experiments, where no coking was

observed on the Sm₂O₃). While coking did not result in deactivation

with time on stream, coking should be minimized as it reduces the

yield to desired products and at longer time-on-stream a decrease

in activity is likely due to reactor blockage. Even though the Al₂O₃-

supported Sm₂O₃catalysts exhibited lower CH₄conversions and

C₂selectivities, the C₂yield per gram of samarium is signiﬁcantly

higher over the Sm₂O₃/p-Al₂O₃-ME and Sm₂O₃/p-Al₂O₃-IM cata-

lysts compared with the Sm₂O₃NP catalysts. This is particularly

true at the 4:1 CH₄:O₂ratio.

The effects of calcining the nanoparticles at 1000 ^◦C instead of

800 ^◦C were also investigated for the Sm₂O₃NPs prepared using

the oleate method. The higher calcination temperature results in a

signiﬁcant decrease in C₂selectivity, and in most cases also a

decrease in CH₄conversion as illustrated in Fig. 2A. After cal-

cination at 1000 ^◦C the performance of the Sm₂O₃(Cl)-OM and

Sm₂O₃(N)-OM NPs are similar, but in most cases this is inferior

to the performance of the catalysts calcined at 800 ^◦C. Only at a

reaction temperature of 740 ^◦C and a CH₄:O₂ratio of 9:1 is there

an advantage to calcining the Sm₂O₃(Cl) NPs at a higher tempera-

ture (not shown), as the CH₄conversion increased from the 7.5% of

the 800 C calcination value to 9.5% (1000 ^◦C calcination) and the C₂

selectivity is similar (44% vs. 45.5%).

Therefore, calcination at a higher temperature is undesirable as

it decreases the C₂selectivity and thus also the C₂yield in most

cases. Only in cases where chlorines may be present on the surface

could a higher calcination temperature improve the C₂yield, but

then only in very special circumstances. Avoiding chloride precur-

sors in catalyst preparation is likely a more effective way to obtain

a higher yield.

Some of the trends in activity and selectivity with CH₄:O₂

ratio and reaction temperature are different for the alumina-

supported Sm₂O₃catalysts compared with the unsupported Sm₂O₃

NPs. While the CH₄conversion increases with decreasing CH₄:O₂

ratio, just as for the unsupported Sm₂O₃NPs, the C₂selectivity

either increases or is not changed signiﬁcantly with decreasing

CH₄:O₂ratio over the alumina-supported Sm₂O₃. This trend in

selectivity is in contrast to what is normally observed over OCM

catalysts, and may be due to different contributions from Sm₂O₃

and Al₂O₃at the various CH₄:O₂ratios. The effect of reaction tem-

perature is also different between the unsupported and supported

Sm₂O₃catalysts. Increasing the reaction temperature from 740 ^◦C

to 800 ^◦C increases both the CH₄conversion and C₂selectivity

over the Sm₂O₃/p-Al₂O₃catalysts. Therefore, while the unsup-

ported Sm₂O₃NP catalysts are superior at a reaction temperature

of 740 ^◦C and higher CH₄:O₂ratios (9:1 and 7:1), at 800 ^◦C and a

CH₄:O₂ratio of 4:1 the Sm₂O₃/p-Al₂O₃catalysts can outperform

the Sm₂O₃(N)-ME NP catalyst and compete with the other Sm₂O₃

NP catalysts.

Evidently, the performance of the Sm₂O₃/p-Al₂O₃catalysts is

dependent on the preparation method. The worst performing cat-

alyst is the Sm₂O₃/p-Al₂O₃-OM catalyst, where the preformed

Sm₂O₃nanoparticles, prepared using the oleate method, are

deposited onto the p-Al₂O₃support using incipient wetness

impregnation. Both the activity and the selectivity are lower on

the Sm₂O₃/p-Al₂O₃-OM catalyst, although the main reason for

the inferior performance is the signiﬁcantly lower C₂selectivity

compared to the other Sm₂O₃/p-Al₂O₃catalysts. Of the three cat-

alysts, the Sm₂O₃/p-Al₂O₃prepared using the incipient wetness

impregnation of an aqueous samarium nitrate solution (Sm₂O₃/p-

Al₂O₃-IM) has the best performance, since the CH₄conversion is

higher and the C₂selectivity is either similar to or higher than those

obtained over the Sm₂O₃/p-Al₂O₃-OM and Sm₂O₃/p-Al₂O₃-ME

catalysts.

3.1.2. Supported Sm₂O₃catalysts – effects of preparation methods

Three types of alumina-supported samaria catalysts were pre-

pared to investigate the effects of samaria deposition method on

the catalytic activity and selectivity of Al₂O₃-supported Sm₂O₃in

the methane coupling reaction. A porous alumina was selected

for this study as this is a typical alumina-based catalyst support.

As described in the catalyst preparation section, the samaria was

deposited using either an incipient wetness impregnation method

or a modiﬁed reverse microemulsion method. For the IWI method

samarium nitrate dissolved in deionized water, or Sm₂O₃(N)-OM

nanoparticles dispersed in hexane, was used during the impreg-

nation of the p-Al₂O₃support. The reaction results obtained from

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

107

are very acidic (NH₃adsorption = 8–9 standard cm³/g) and have

very few basic sites (CO₂adsorption = 0.6–075 standard cm³/g). In

contrast, the p-Al₂O₃support has signiﬁcantly less acidic sites and

an almost equal number of acidic and basic sites: 1.6 sccm/g NH₃

adsorption and 1.75 sccm/g CO₂adsorption, respectively. Further-

more, the p-Al₂O₃and n-Al₂O₃(−) supports have similar surface

areas (260–275 m²/g), while the n-Al₂O₃(+) has a signiﬁcantly

higher surface area (695 m²/g).

70

60

50

40

30

20

10

0

A

T = 740°C

Sm₂O₃/p-Al₂O₃OM

Sm₂O₃/p-Al₂O₃IM

Sm₂O₃/p-Al₂O₃ME

To faciliate interpretation of the reaction data obtained from

the Sm₂O₃/Al₂O₃catalysts, two alumina supports, the p-Al₂O₃

and n-Al₂O₃(−) calcined at 800 ^◦C, were subjected to the reaction

conditions to determine any activity originating from the sup-

port. As can be seen in Table 2, the Al₂O₃supports do exhibit

some activity. At 740 ^◦C and a CH₄:O₂ratio of 4:1 the CH₄con-

version over the p-Al₂O₃support is similar to that obtained over

the Sm₂O₃/p-Al₂O₃-OM and Sm₂O₃/p-Al₂O₃-ME catalysts under

the same conditions. However, the conversion to C₂and CO_xprod-

4:1

9:1

7:1

0

5

10

15

20

25

ucts (i.e. the X_C^∗_H) is lower, indicating that coking occurs over

CH₄Conversion [%]

4

the bare supports. This is evident in the color change from white to

black after reaction. Also, most of the CH₄reacted over the support

is converted to CO₂or CO. The selectivity to C₂products is only 8.5%

over the p-Al₂O₃support.

The n-Al₂O₃(−) support exhibits a slightly different behavior

compared with the p-Al₂O₃support. The CH₄conversion is some-

what lower, but the CH₄conversion to C₂and CO_xproducts is

higher, i.e. there is less coking. The latter is clearly evident as this

support is gray rather than black after reaction. The C₂selectivity is

also higher and so is the CO₂/CO ratio while the C₂H₄/C₂H₆product

ratio is lower.

The reaction data obtained from Sm₂O₃deposited onto the

different Al₂O₃supports using the incipient wetness impreg-

nation method of both the aqueous samarium nitrate solution

and the hexane dispersion of Sm₂O₃-OM NPs are presented in

Figs. 4 and 5. As the trends observed on the Sm₂O₃/p-Al₂O₃cat-

alysts, i.e. the CH₄conversion and C₂selectivity increases with

increasing temperature and decreasing CH₄:O₂ratio, are the same

for all alumina-supported catalysts, irrespective of catalyst prepa-

ration method and speciﬁc alumina support, only results obtained

at the 4:1 CH₄:O₂ratio are presented in Tables 2 and 3 (i.e. the

feed ratio which gives the highest CH₄conversion and C₂selec-

tivity). Despite the differences in speciﬁc surface areas as well

as acid and basic properties, the CH₄conversion does not vary

signiﬁcantly between the various Al₂O₃supports. For most condi-

tions, the variations in CH₄conversion between catalysts prepared

using the different Al₂O₃supports is close to + - 1%. Thus, the

method of Sm₂O₃deposition and the reaction conditions have

larger effects on the CH₄conversion, than the Al₂O₃support prop-

erties. The CH₄conversions (14–16%) for the Sm₂O₃/Al₂O₃-OM

catalysts are consistent with the conversions obtained from the

bare supports, while the CH₄conversions over the Sm₂O₃/Al₂O₃-

IM catalysts are higher (18.5–20.5%). These results suggest that

more support is accessible to reaction on the Sm₂O₃/p-Al₂O₃(OM)

catalyst compared with the Sm₂O₃/p-Al₂O₃(IM) and Sm₂O₃/p-

Al₂O₃(ME) catalysts. The Sm₂O₃dispersion is thus likely highest

on the Sm₂O₃/Al₂O₃(IM) catalysts and lowest on the Sm₂O₃/p-

Al₂O₃(OM) catalysts.

70

B

T = 800°C

Sm₂O₃/p-Al₂O₃-OM

Sm₂O₃/p-Al₂O₃-IM

Sm₂O₃/p-Al₂O₃-ME

60

50

40

30

20

10

0

4:1

9:1

7:1

0

5

10

15

20

25

CH₄Conversion [%]

Fig. 3. OCM reaction data for Sm₂O₃supported on p-Al₂O₃prepared using incipient

wetness impregnation of samarium nitrate (aqueous solution, IM) and a hexane dis-

persion of Sm₂O₃(N)-OM nanoparticles (OM) as well as a modiﬁed microemulsion

method (ME). The numbers in the boxes indicate the CH₄:O₂feed ratios, i.e. 4:1, 7:1

and 9:1. (A) reaction temperature of 740 ^◦C and (B) reaction temperature of 800 ^◦C.

Compared to the Sm₂O₃NPs, the CO₂:CO product ratios are

lower over the Sm₂O₃/p-Al₂O₃catalysts, as signiﬁcantly more CO is

formed over the supported catalysts. In most cases, the C₂H₄:C₂H₆

ratio is also lower for the supported catalysts. Only at a CH₄:O₂

ratio of 4:1 is the C₂H₄:C₂H₆ratio higher for the Sm₂O₃/p-Al₂O₃-IM

catalyst compared with the unsupported catalysts. This is also true

for the Sm₂O₃/p-Al₂O₃-OM catalyst but only at a reaction temper-

ature of 800 ^◦C. Clearly, the presence of an Al₂O₃support alters the

product distribution and results in lower C₂yields, but the differ-

ence in C₂yields is not as large as may be expected from an acidic

support, and, moreover, the yield per gram of Sm₂O₃is higher for

the Sm₂O₃/p-Al₂O₃catalysts compared with the Sm₂O₃NPs. This

indicates that Al₂O₃may be considered as a support in OCM reac-

tions, and it is possible that if the Al₂O₃is doped to reduce the

acidity of the surface, Al₂O₃-supported Sm₂O₃could be a compet-

itive catalyst.

In contrast, the C₂selectivity is strongly inﬂuenced by the Al₂O₃

support. In most cases, catalysts supported on the more acidic n-

Al₂O₃(+) exhibit a lower C₂selectivity than catalysts supported on

the p-Al₂O₃(Table 2 and Fig. 4). The highest C₂selectivity of the

Al₂O₃-supported catalysts is obtained over the Sm₂O₃/n-Al₂O₃(−)-

IM catalyst and this is true for all CH₄:O₂ratios at both reaction

temperatures (Fig. 4). The Sm₂O₃/n-Al₂O₃(−)-OM also has a higher

C₂selectivity than the other Sm₂O₃/Al₂O₃-OM catalysts at most

conditions (Fig. 5). Considering the higher acidity of the n-Al₂O₃(−)

support versus the p-Al₂O₃support, this result is unexpected, but

3.1.3. Supported Sm₂O₃catalysts – effects of alumina support

Three different alumina supports, p-Al₂O₃, n-Al₂O₃(+), and n-

Al₂O₃(−), were selected and loaded with 20% samaria by weight to

determine the effects of alumina properties on Sm₂O₃/Al₂O₃cata-

lysts. These supports were chosen to probe the effects of support

surface areas and varying acidic and basic support properties. As

can be seen in Table 1, both the nanoparticle alumina supports

108

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

70

60

50

40

30

20

10

0

A

T = 740°C

Sm₂O₃/n-Al₂O₃(+)-OM

Sm₂O₃/p-Al₂O₃-OM

Sm₂O₃/n-Al₂O₃(-)-OM

Sm₂O₃/n-Al₂O₃(+)-IM

60

50

40

30

20

10

0

Sm₂O₃/p-Al₂O₃-IM

Sm₂O₃/n-Al₂O₃(-)-IM

4:1

9:1

7:1

4:1

15

7:1

9:1

0

5

10

20

25

0

5

10

15

20

25

CH₄Conversion [%]

70

60

50

40

30

20

10

0

70

60

50

40

30

20

10

0

B

T = 800°C

Sm₂O₃/n-Al₂O₃(+)-IM

Sm₂O₃/p-Al₂O₃-IM

Sm₂O₃/n-Al₂O₃(-)-IM

T = 800°C

Sm₂O₃/n-Al₂O₃(+)-OM

Sm₂O₃/p-Al₂O₃-OM

Sm₂O₃/n-Al₂O₃(-)-OM

9:1

7:1

4:1

7:1

9:1

0

5

10

15

20

25

0

5

10

15

20

25

CH₄Conversion [%]

Fig. 5. OCM reaction data obtained from catalysts prepared using incipient wetness

impregnation of an aqueous samarium nitrate solution on different Al₂O₃supports.

The numbers in the boxes indicate the CH₄:O₂feed ratios, i.e. 4:1, 7:1 and 9:1. (A)

reaction temperature of 740 ^◦C and (B) reaction temperature of 800 ^◦C.

Fig. 4. OCM reaction data for Sm₂O₃supported on different Al₂O₃supports prepared

using incipient wetness impregnation of a hexane dispersion of Sm₂O₃(N)-OM

nanoparticles. The numbers in the boxes indicate the CH₄:O₂feed ratios, i.e. 4:1,

7:1 and 9:1. (A) reaction temperature of 740 ^◦C and (B) reaction temperature of

800 ^◦C.

trend is the same, only the C₂H₄:C₂H₆product ratios are higher

(for the Sm₂O₃/p-Al₂O₃-IM catalyst they increase from 1.3 to 2.7),

which is expected over samaria catalysts [6]. The C₂H₄:C₂H₆and

CO₂:CO product ratios vary more with Al₂O₃supports over the

Sm₂O₃/Al₂O₃-OM catalysts compared with the Sm₂O₃/n-Al₂O₃-

IM catalysts. For example, the higher CO₂:CO product ratio of

the Sm₂O₃/n-Al₂O₃(−)-OM catalyst compared with the Sm₂O₃/p-

Al₂O₃-OM catalyst, reﬂects the higher CO₂:CO product ratio of the

n-Al₂O₃(−) versus the p-Al₂O₃supports. This again is consistent

with more Al₂O₃being exposed on the OM versus the IM series

of catalysts. Thus, the higher CO₂and CO product concentrations

(lower C₂selectivities), observed over the Sm₂O₃/Al₂O₃catalysts

are due to the Al₂O₃support. The best Al₂O₃-supported catalyst

is the Sm₂O₃/n-Al₂O₃(−)-IM, and the main reason is likely that

more of the alumina support is covered by Sm₂O₃on catalysts pre-

pared using impregnation with an aqueous solution of samarium

nitrate compared to the other preparation methods. Furthermore,

the C₂selectivity of the n-Al₂O₃(−) is slightly higher than that of the

p-Al₂O₃. The Sm₂O₃/n-Al₂O₃(−)-IM catalyst can almost compete

with the best pure NP samaria catalyst as the best C₂yield over this

catalyst is 5.6% at 800 ^◦C, while it is 7.8% at 740 ^◦C (or 6.8% at 800 ^◦C)

over the Sm₂O₃(N)-OM NPs. Since the Sm₂O₃content is only 20%

for the Al₂O₃-supported catalysts, the amount of C₂products pro-

duced per gram of Sm₂O₃is more than three times higher over the

indicates that other support properties also inﬂuence the selectiv-

ity. The original surface area can be excluded, as it is similar on

the n-Al₂O₃(−) and p-Al₂O₃supports (Table 1). As for the p-Al₂O₃-

supported catalysts, the impregnation method using the nitrate

precursor (the Sm₂O₃/n-Al₂O₃(−)-IM and Sm₂O₃/n-Al₂O₃(+)-IM

catalysts) results in catalysts with considerably higher C₂yields

than the catalysts prepared using the Sm₂O₃(N)-OM nanoparticles

(the Sm₂O₃/n-Al₂O₃(−)-OM and Sm₂O₃/n-Al₂O₃(+)-OM catalysts),

due to higher CH₄conversions and C₂selectivities. This result sup-

ports the conclusion that the Sm₂O₃dispersion is higher on the

Sm₂O₃/Al₂O₃(IM) catalysts compared with the Sm₂O₃/Al₂O₃(OM)

catalysts.

Using the incipient wetness impregnation method of a nitrate

precursor onto the Al₂O₃supports (the Sm₂O₃/Al₂O₃-IM catalyst

series), the trends in product distribution appear to be similar

over the different Al₂O₃supports. For example, over the Sm₂O₃/p-

Al₂O₃-IM catalyst at 740 ^◦C the C₂H₄:C₂H₆product ratio increases

from 0.6 to 1.3 and the CO₂:CO product ratio increases from 1.8

to 2.4 with decreasing CH₄:O₂ratio (Table 3). These trends are

expected considering the CH₄:O₂stoichiometry of the reactions

(CH₄to C₂H₄versus C₂H₆and CH₄to CO₂versus CO) and are the

same (with similar product ratios) over the other Sm₂O₃/Al₂O₃-IM

catalysts (not shown). Even at a reaction temperature of 800 ^◦C the

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

109

is a signiﬁcant decrease in surface area from 14.5 m²/g to 1.1

m²/g when increasing the calcination temperature from 800 ^◦C and

1000 ^◦C, this only results in a small decrease in activity (1% 0.5%)

and this is true for all reaction temperatures and reactant feed

ratios. However, there is a signiﬁcant decrease in the C₂selectivity

at the higher calcination temperature, which indicates that smaller

Sm₂O₃particles (higher Sm₂O₃surface areas) are beneﬁcial for the

reaction.

Al₂O₃-supported catalysts compared with the pure Sm₂O₃NPs.

Therefore, considering the limited supply of “rare earth oxides”,

using an Al₂O₃support can reduce the amount of REO needed and

by modifying the support to, for example, make it more basic, as

well as adding alkaline earth promoters to the Sm₂O₃it should be

possible to prepare supported samaria catalysts that can signiﬁ-

cantly outperform pure Sm₂O₃catalysts.

3.1.4. Supported Sm₂O₃catalysts – effects of calcination

As expected, due to the high surface areas of the alumina sup-

ports, the BET surface areas of the supported Sm₂O₃catalysts are

signiﬁcantly larger than those of the Sm₂O₃NPs (Table 1). However,

despite the higher original surface area of the n-Al₂O₃(+) support,

the catalysts supported on the p-Al₂O₃exhibit the highest surface

areas. The lowest surface area is obtained from catalysts supported

on the n-Al₂O₃(−). The p-Al₂O₃support is therefore more stable

than the nanoparticle Al₂O₃supports during catalyst preparation

and the high calcination temperatures (800 ^◦C or higher) used in

the investigation. The higher stability of the p-Al₂O₃support is also

evident when comparing catalysts prepared using different meth-

ods. The BET surface areas of the Sm₂O₃/p-Al₂O₃catalysts prepared

using the impregnation or microemulsion techniques vary between

125 and 145 m²/g. In contrast, the BET surface areas of Sm₂O₃

catalysts supported on Al₂O₃(+) vary between 80 and 130 m²/g

dependent on which impregnation techniques was used (Table 1).

The same numbers for the Sm₂O₃/Al₂O₃(−) catalysts are 50 and 85

m²/g.

Comparing the different preparation methods on the same

support, it is evident that the incipient wetness impregnation

(IWI) method using the nitrate precursor results in catalysts with

smaller surface areas compared with those prepared using the same

method and the OM NPs. This suggests that the IWI with the nitrate

precursor blocks more of the pores in the Al₂O₃support during the

impregnation and drying steps, so that the reactants are exposed

to less alumina over this catalyst compared with the IWI using the

OM NPs. This is consistent with the conclusions from the reaction

experiments over the Sm₂O₃/Al₂O₃-OM catalysts. Despite the large

differences in surface areas between catalysts prepared using the

same method but different Al₂O₃-supports, and the similar surface

areas for catalysts prepared using different methods on the same

support, the CH₄conversion vary more with catalyst preparation

method than with the speciﬁc Al₂O₃support. As no strong corre-

lation was observed between the Sm₂O₃surface area and activity,

it is not surprising that the support surface area does not appear to

signiﬁcantly inﬂuence the CH₄conversion.

In contrast, the C₂selectivity varies with both preparation

method and Al₂O₃support. Although the reason for this is not due

to the variation in the overall catalyst surface area, the BET data

reveals some interesting properties that are likely important in

these reactions. As the alumina can provide acidic sites that pro-

mote oxidation to undesired CO_x, optimizing the catalyst synthesis

procedure and use of supports that can provide a high surface area

for the Sm₂O₃while at the same time minimize the number of

exposed acidic sites are important. The properties of the p-Al₂O₃,

n-Al₂O₃(+), and n-Al₂O₃(−) are very different and this is evident

in the product distribution from the catalysts on these supports.

However, the preparation method can also have a signiﬁcant effect

on the selectivity. Due to the high surface area and the acidity

of the n-Al₂O₃(+), the catalysts on this support in general exhibit

the lowest C₂selectivities. However, the lowest C₂selectivity of

all the Al₂O₃-supported catalysts is observed over the Sm₂O₃/p-

Al₂O₃-OM catalyst at a reaction temperature of 740 ^◦C. Comparing

the catalysts prepared using IWI of OM NPs versus the IWI of the

nitrate precursor, it is evident that the OM method results in cata-

lysts with signiﬁcantly lower C₂selectivities. The BET surface area

measurements indicate that this is due to the fact that more of the

Al₂O₃support is exposed on these catalysts. The incipient wetness

temperature

The effects of calcination temperature on selected supported

catalysts were also investigated by calcining Sm₂O₃/p-Al₂O₃-IM

and Sm₂O₃/p-Al₂O₃-ME catalysts at 1000 ^◦C instead of at 800 ^◦C.

Tables 2 and 3, reveal that the effects of calcination temperature on

the supported Sm₂O₃catalysts are dependent on the preparation

method. For the Sm₂O₃/p-Al₂O₃-ME catalyst there is a signiﬁcant

increase in C₂selectivity and a slight increase in CH₄conver-

sion after calcination at the higher temperature (Fig. 2B). This is

true at both reaction temperatures, 740 ^◦C and 800 ^◦C, although

the percent increase in C₂yield after the calcination at 1000 ^◦C

versus 800 ^◦C is higher at 740 ^◦C. While this trend is similar to that

observed over the Sm₂O₃(Cl)-OM NPs, the increase in selectivity is

more signiﬁcant over the Sm₂O₃/p-Al₂O₃-ME catalyst. In contrast,

over the Sm₂O₃/p-Al₂O₃-IM catalyst, there is a slight decrease in

both CH₄conversion and C₂selectivity after calcination at 1000 ^◦C.

While calcination at a higher temperature can be beneﬁcial

for Al₂O₃-supported Sm₂O₃catalysts, it is only under certain cir-

cumstances. Therefore, as was the case for the pure Sm₂O₃NPs,

methods other than high temperature treatments are likely to be

more effective in improving the catalytic activities and selectivities

for Sm₂O₃/Al₂O₃catalysts.

3.2. Samaria catalyst characterizations

The unsupported samaria NPs and the Sm₂O₃/Al₂O₃catalysts

were subjected to catalyst characterizations to determine proper-

ties of importance for the catalytic activities and selectivities and

determine differences in these properties as a function of prepa-

ration methods and alumina supports used. The results from each

technique are presented below.

3.2.1. BET surface area measurements

The BET surface areas of the prepared Sm₂O₃nanoparticles

after calcination at 800 ^◦C are presented in Table 1. The particle

sizes calculated using the BET surface areas are included in the

table. It is evident from these results that the oletate method gives

smaller nanoparticles (larger BET surface area) than the₋microemul-

sion method, and, using the oleate method, the NO₃precursor

results in slightly smaller particles compared to the Cl⁻precur-

sor. The lower BET surface areas do explain the lower activities

of the Sm₂O₃(N)-ME and Sm₂O₃(Cl)-OM NPs compared with the

Sm₂O₃(N)-OM NPs, but no simple correlation between the BET

surface area and the activity or selectivity data obtained over the

Sm₂O₃nanoparticles is evident. For example, the Sm₂O₃(N)-ME

NPs are more active than would be expected from their surface

area, and the Sm₂O₃(N)-OM NPs have the lowest activity per m²

of the Sm₂O₃NPs even though they are the most active per unit

mass. Since the Sm₂O₃nanoparticles were prepared using dif-

ferent methods, and there are likely residues from the catalyst

preparation present on the surface of the NPs (which could cover

some of Sm₂O₃), it is not expected that there will be a straight

correlation between the catalytic activity and the surface area of

these nanoparticles. However, it is also possible that factors other

than the Sm₂O₃surface area are important for a high catalytic

activity. The results from the Sm₂O₃(N)-OM NPs calcined at differ-

ent temperatures indicate that this may be the case. While there

110

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

Sm₂O₃/p-Al₂O₃Catalysts

Sm₂O₃Nanoparticles

A

Calc. 800°C

OM

Sm₂O₃(Cl)-OM

Sm₂O₃(N)-OM

IM

ME

Sm₂O₃(N)-ME

OM/Al₂O₃(-)

20

30

40

50

60

70

2θ

20

30

40

50

60

70

80

2

θ

Fig. 6. XRD data obtained from Sm₂O₃nanoparticles prepared using the oleate

decomposition with nitrate and chloride precursors and the reverse microemul-

sion method using a samarium nitrate precursor. ꢀ: indicates peak positions due to

cubic Sm₂O₃.

Sm₂O₃/p-Al₂O₃Catalysts

B

Calc. 1000°C

IM

impregnation using the nitrate precursor appears to block more

of the pores on the alumina support compared to the IWI of the

OM NPs. The best C₂selectivities for the supported catalysts are

observed over the Sm₂O₃/n-Al₂O₃(−)-IM catalyst, and the BET data

reveal that this is likely due to the fact that a signiﬁcant portion

of the acidic alumina support has been covered during catalyst

preparation.

3.2.2. XRD measurements

The XRD patterns obtained from the NPs synthesized using the

microemulsion method (Sm₂O₃(N)-ME) after calcination at 800 ^◦C

is shown in Fig. 6. The main phase identiﬁed from the XRD plot is the

cubic form of Sm₂O₃with no major impurities. This is important as

the cubic phase has been shown to be more active in the OCM reac-

tion compared with for example the monoclinic phase of Sm₂O₃

[42]. The Sm₂O₃peaks are well-deﬁned revealing a highly crys-

talline sample. The average particle size according to the Scherrer

equation (Eq. (6)) is 48.2 nm. This is signiﬁcantly smaller than the

particle size estimated from the BET surface area measurement and

is likely due to crystal defects in the Sm₂O₃nanoparticles, which

would result in broader XRD peaks. Using the Scherrer equation it

is assumed that the line-broadening is due to particle size effects

only, and thus resulting in XRD particle sizes that are smaller than

the real average particle size. Furthermore, the presence of XRD

invisible amorphous Sm₂O₃cannot be excluded.

The Sm₂O₃(N)-OM NPs yield signiﬁcantly broader XRD peaks

compared to the Sm₂O₃(N)-ME NPs (Fig. 6). This is an indication of

smaller particles, as would be expected from the larger BET surface

area of the Sm₂O₃(N)-OM compared with the Sm₂O₃(N)-ME NPs.

The crystallite size for the Sm₂O₃(N)-OM NPs calculated using the

Scherrer equation is 14.0 nm, i.e. signiﬁcantly smaller than for the

Sm₂O₃(N)-ME NPs.

The slightly broadened peaks in the XRD pattern obtained from

the Sm₂O₃(Cl)-OM NPs are characteristic of the small grain sizes

in the crystalline phase. According to the Scherrer equation, the

average crystallite size for the Sm₂O₃(Cl)-OM NPs is approximately

25.4 nm. While the particle sizes determined from the XRD mea-

surements in all cases are smaller than those calculated using the

BET surface areas, they do follow the same trend, i.e. the Sm₂O₃(N)-

OM are the smallest and the Sm₂O₃(N)-ME are the largest NPs.

The XRD patterns obtained from the Sm₂O₃/p-Al₂O₃catalysts

are presented in Fig. 7A. While the Sm₂O₃(N)-OM NPs are cubic

ME

20

30

40

50

60

70

80

2

θ

Fig. 7. A) XRD data obtained from the Sm₂O₃/p-Al₂O₃-OM, Sm₂O₃/p-Al₂O₃-IM,

Sm₂O₃/p-Al₂O₃-ME and Sm₂O₃/n-Al₂O₃(−)-OM catalysts after calcination at 800 ^◦C.

B) XRD data obtained from the Sm₂O₃/p-Al₂O₃-IM and Sm₂O₃/p-Al₂O₃-ME catalysts

after calcination at 1000 ^◦C. ꢁ: monoclinic Sm₂O₃, ꢀ: Sm₂O, ^ଝ: SmO_0.656, ᭹: ␥-Al₂O₃,

ꢂ: SmAlO₃, and ꢀ: cubic Sm₂O₃.

after calcination at 800 ^◦C, in the presence of the Al₂O₃sup-

port the cubic Sm₂O₃appears to undergo a transformation, as no

peak due to cubic Sm₂O₃is visible in the XRD pattern obtained

from the Sm₂O₃/p-Al₂O₃catalysts and the Sm₂O₃is evidently

less crystalline on the Sm₂O₃/p-Al₂O₃catalysts. In addition to the

broad peaks due to the poorly crystalline ␥-Al₂O₃support, the

XRD patterns obtained from the Sm₂O₃/p-Al₂O₃-ME and Sm₂O₃/p-

Al₂O₃-OM catalysts exhibit two peaks near 2␪ = 30^◦. In contrast, the

Sm₂O₃/n-Al₂O₃(−)-OM catalyst exhibits four peaks in this region.

The differences in the XRD patterns obtained from the Sm₂O₃/p-

Al₂O₃-OM and Sm₂O₃/n-Al₂O₃(−)-OM catalysts, suggest that the

interactions between the Sm₂O₃and the Al₂O₃are different on

the p-Al₂O₃and n-Al₂O₃supports. While the presence of mono-

clinic Sm₂O₃cannot be ruled out on these catalysts, particularly not

on the Sm₂O₃/n-Al₂O₃(−)-OM catalyst, the reference monoclinic

Sm₂O₃peaks (JCPDS [42–1464]) are not a good match (Fig. 7A).

The peaks are not due to fully developed mixed metal oxide phases

either, such as SmAlO₃or Sm₄Al₂O₉[39,43]. However, these peaks

have been observed on a 40% Sm₂O₃on Al₂O₃catalyst [39], and

may thus be a precursor to SmAlO₃formation, or they could be due

to Sm-rich SmO_x(Fig. 7A). The fact that the XRD pattern obtained

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

111

from a spent catalyst (not shown), after several hours on stream,

only shows particle size growth and does not contain any peaks due

to SmAlO₃suggests that the peaks most likely are due to Sm-rich

SmO_x.

Sm₂O₃/p-Al₂O₃-IM catalyst after calcination at 1000 ^◦C is there-

fore likely due to a lower Sm₂O₃surface area, due to both SmAlO₃

formation and Sm₂O₃particle growth. In contrast, the Sm₂O₃/p-

Al₂O₃-ME catalyst after calcination appears to consist of other

SmO_xphases in addition to the cubic Sm₂O₃. The effects of these

SmO_xphases on the catalytic activity and selectivity are not known,

but they do not appear to be detrimental. It is also possible that the

calcination treatment removes (or redistributes) residues that are

left from the catalyst preparation procedure, residues which could

have resulted in the low initial activity and selectivity observed

over this catalyst after calcination at 800 ^◦C.

No peaks due to Sm₂O₃are observed in the XRD pattern obtained

from the Sm₂O₃/p-Al₂O₃-IM catalyst. Thus, the incipient wetness

impregnation using the nitrate precursor, results in Sm₂O₃on the

surface which is either amorphous or has particle sizes below the

detection limit of XRD. While it is possible that the inferior perfor-

mance of the Al₂O₃-supported catalysts compared with the Sm₂O₃

NPs is due to the change in Sm₂O₃crystal phase, it is more likely

that the lower activity and selectivity are due to the lower Sm₂O₃

content and the acidic alumina, respectively, on the Sm₂O₃/p-Al₂O₃

catalysts.

3.2.3. TEM analysis

TEM images obtained from the Sm₂O₃(N)-OM NPs before the

calcination step is presented in Fig. 8A. Small nanoparticles with

diameters between 3 and 5 nm and a narrow size distribution

are observed. After calcination at 800 ^◦C particle agglomeration

is evident, but there are still particles present with diameters

around 5 nm (Fig. 8B). However, signiﬁcantly larger particles are

also present and, thus, the particle size distribution is signﬁcantly

broader. The particle sizes appear to be reasonably consistent

with the 14 nm average crystallite size determined from the XRD

data, but there is signiﬁcant agglomeration which explains the

larger particle size obtained from the BET surface area measure-

ments.

After calcination at 1000 ^◦C, the Sm₂O₃/p-Al₂O₃-IM and

Sm₂O₃/p-Al₂O₃-ME catalysts are much more crystalline (Fig. 7B).

While some peaks due to ␥-Al₂O₃and cubic Sm₂O₃can be observed

in the XRD patterns from these catalysts, the main crystalline phase

is a samarium aluminum oxide (SmAlO₃) [41]. Even though the

transformation from poorly crystalline Sm₂O₃/␥-Al₂O₃to SmAlO₃

appears to be similar on both catalysts, the behavior of the

Sm₂O₃/p-Al₂O₃-IM and Sm₂O₃/p-Al₂O₃-ME catalysts with calci-

nation temperature is very different. This could be due to two

opposing effects, i.e. SmAlO₃formation decreases the surface Al₂O₃

concentration (a positive effect), but it also decreases the Sm₂O₃

at the surface (a negative effect). On the Sm₂O₃/p-Al₂O₃-IM the

high-temperature treatment causes particle growth and/or crys-

tallization of the cubic Sm₂O₃phase. The lower activity on the

The TEM image for the Sm₂O₃(N)-ME NPs (post calcination)

reveal highly crystalline NPs with a particle size range between

30 and 50 nm (Fig. 8C). Thus, these particles are larger than

Fig. 8. TEM images obtained from (A) Sm₂O₃(N)-OM NPs before calcination (scale bar = 10 nm), (B) Sm₂O₃(N)-OM NPs after calcination at 800 ^◦C (scale bar = 20 nm), (C)

Sm₂O₃(N)-ME NPs after calcination at 800 ^◦C (scale bar = 50 nm), and (D) Sm₂O₃/n-Al₂O₃(−)-OM catalyst after calcination at 800 ^◦C (scale bar = 50 nm).

112

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

the Sm₂O₃(N)-OM NPs, but appears to be less agglomerated

after calcination. The particle sizes from TEM is fairly consistent

with the average particle size calculated using the XRD data, but

there are signiﬁcantly larger particles present which again explains

the particle size determined from the BET surface area.

The TEM image of the Sm₂O₃(N)-OM NPs deposited onto the

alumina support reveal reasonably small Sm₂O₃particles dispersed

on the surface of the n-Al₂O₃(−) support. Due to the difference

in the material (samaria is a high Z material, while alumina is a

low Z material), it is possible distinguish between the two oxides

(Fig. 8D).

preparation method on the supported catalysts. The survey spec-

tra are presented in Fig. 9A. To facilitate analysis, the high binding

energy region between 1200 and 900 eV, and the low binding

energy region between 500 and 0 eV are displayed in Fig. 9B and

C, respectively. The main differences between the catalysts are

variations in the relative peak intensities. The Sm₂O₃/p-Al₂O₃-

IM has the lowest carbon content. The higher carbon content on

the other catalysts is likely due to residual carbon on the sur-

face from the catalyst preparation, i.e. from the oleate complex

for the Sm₂O₃/p-Al₂O₃-OM and from the Tween 80 surfactant for

the Sm₂O₃/p-Al₂O₃-ME catalyst. On a carbon-free basis, or look-

ing only at the Al, Sm and O peaks, the Sm/Al and O/(Sm + Al)

peak area ratios are similar for the Sm₂O₃/p-Al₂O₃-ME and

Sm₂O₃/p-Al₂O₃-IM catalysts (Table 4). Perhaps the most impor-

tant difference between the three catalysts is the higher Sm/Al

3.2.4. XPS measurements

XPS spectra were collected from the three Sm₂O₃catalysts

supported on the p-Al₂O₃to determine the effects of catalyst

Sm₂O₃/p-Al₂O₃Catalysts

A

ME

OM

IM

1200

1000

800

600

400

200

0

Binding Energy

Sm₂O₃/p-Al₂O₃Catalysts

B

C

OM

ME

IM

OM

ME

IM

500

400

300

200

100

0

1200

1100

1000

900

Binding Energy

Fig. 9. XPS survey spectra obtained from the Sm₂O₃/p-Al₂O₃-IM, Sm₂O₃/p-Al₂O₃-OM and Sm₂O₃/p-Al₂O₃-ME catalysts. (A) Full scale XPS survey spectra, (B) high binding

energy region, and (C) low binding energy region.

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

113

Table 4

XPS Compositional analysis obtained from Sm₂O₃/p-Al₂O₃catalyst series.

Catalyst^a

O 1s [%]

Sm 3d_5/2[%]

Al 2p_3/2[%]

C 1s [%]

Na 1s [%]

Sm/Al Ratio

O/(Sm + Al) Ratio

Sm₂O₃(N)/p-Al₂O₃-OM

Sm₂O₃(N)/p-Al₂O₃-ME

Sm₂O₃(N)/p-Al₂O₃-IM

57

58

63

11

7

20

23

24

12

11

6

–

1.5

–

0.55

0.30

0.29

1.84

1.96

2.03

a

The following atomic sensitivity factors were used in the calculation of the composition: O 1s = 0.711, Sm 3d_5/2= 2.907, Al 2p = 0.193, C 1s = 0.296, Na 1s 1.685 [44].

peak area ratio observed on the Sm₂O₃/p-Al₂O₃-OM catalyst.

Thus, there is more Sm₂O₃in the near surface region on this

catalyst compared with the other Sm₂O₃/p-Al₂O₃catalysts. This is

not necessarily due to a higher Sm₂O₃surface area on the Sm₂O₃/p-

Al₂O₃-OM catalyst, as it may simply mean that more Sm₂O₃is

deposited onto the outermost surface of the p-Al₂O₃support. The

lower intensity of the Al peaks obtained from this catalyst, com-

pared to the Sm₂O₃/p-Al₂O₃-IM and Sm₂O₃/p-Al₂O₃-ME catalysts,

supports this conclusion (Fig. 9 C). This is also consistent with the

BET surface area measurements on the Sm₂O₃/p-Al₂O₃-OM and

Sm₂O₃/p-Al₂O₃-IM catalysts. It is possible that the Sm₂O₃(N)-OM

NPs do not enter the pores of the p-Al₂O₃as efﬁciently as the

aqueous solutions of samarium nitrate in the Sm₂O₃/p-Al₂O₃-IM

and Sm₂O₃/p-Al₂O₃-ME catalysts. So, while the outermost surface

has more Sm₂O₃exposed, the small pores of the alumina support

may have more alumina exposed, which would be accessible to

gas phase methane and oxygen. This in turn would explain the

poor selectivity of this and the other Sm₂O₃/Al₂O₃-OM catalysts

compared with the Sm₂O₃/p-Al₂O₃-IM and Sm₂O₃/p-Al₂O₃-ME

catalysts. The lowest Sm peak intensities are observed over the

Sm₂O₃/p-Al₂O₃-ME catalyst and this is mainly due to the pres-

ence of sodium on the surface of this catalyst. More residual

sodium (from the catalyst preparation) is present on the surface

of this catalyst compared to the other Sm₂O₃/p-Al₂O₃catalysts.

While the Sm/Al ratio in the near surface region is similar on the

Sm₂O₃/p-Al₂O₃-IM and Sm₂O₃/p-Al₂O₃-ME catalysts, it is possible

that the Na covers both Sm₂O₃and Al₂O₃on the Sm₂O₃/p-Al₂O₃-

ME catalyst. The lower Sm peak intensities in the survey spectrum

obtained from the Sm₂O₃/p-Al₂O₃-ME catalyst indicate that this

is the case. As alkaline earth metal ions often are used to pro-

mote Sm₂O₃-based OCM catalysts, this is likely the reason why

the Sm₂O₃/p-Al₂O₃-ME catalyst has the highest C₂selectivity of

the Sm₂O₃catalysts supported on p-Al₂O₃. However, the large

amount of Na on the surface of this catalyst probably blocks some

of the Sm₂O₃, which would explain the lower CH₄conversion

compared with the Sm₂O₃/p-Al₂O₃-IM catalyst. It is likely that

the heat treatment at 1000 ^◦C immobilizes the Na and causes a

redistribution of the Na on the surface of the Sm₂O₃/p-Al₂O₃-ME

catalyst, so that more Sm₂O₃is present at the outermost surface

compared with before the heat treatment. This together with a

potential promoting effect of the Na on the Sm₂O₃can explain

the improved performance of this catalyst after calcination at

1000 ^◦C.

The high-resolution XPS scans of the Sm 3d_5/2and O 1s regions

are shown in Fig. 10A and B, respectively. The Sm 3d_5/2bind-

ing energy of 1083.0 eV is within the reported range [44]. The

Sm 3d_5/2peak obtained from the Sm₂O₃/p-Al₂O₃-IM catalyst is

slightly broader than the Sm 3d_5/2peaks obtained from the other

Sm₂O₃/p-Al₂O₃catalysts. This is due to an increased intensity at

higher binding energy and could indicate stronger interactions with

the alumina support on this catalyst. The O 1s binding energy of

531.3 eV is expected for O 1s in Al₂O₃. All O 1s peaks are slightly

broadened on the high binding energy side, revealing presence

of surface hydroxyl groups and possibly carbonate species. The

presence of carbonate species is expected over rare earth oxide

catalysts.

Sm₂O₃/p-Al₂O₃Catalysts

A

B

Sm 3d_5/2

O 1s

OM

ME

OM

ME

IM

1090

1085

1080

1075

535

530

525

Binding Energy [eV]

Fig. 10. High-resolution XPS spectra of the (A) Sm 3d_5/2and (B) O 1s peaks obtained from the Sm₂O₃/p-Al₂O₃-IM, Sm₂O₃/p-Al₂O₃-OM and Sm₂O₃/p-Al₂O₃-ME catalysts.

114

T.W. Elkins, H.E. Hagelin-Weaver / Applied Catalysis A: General 454 (2013) 100–114

4. Conclusion

acknowledged for their support during the XRD and XPS mea-

surements and Mr Gill Brubaker for support during the BET

surface area measurements. These analyses were performed at

the Major Analytical Instrumentation Center (MAIC) and the Par-

ticle Engineering Research Center (PERC) at the University of

Florida.

The reaction data obtained from the Sm₂O₃NPs reveal that the

smaller particle sizes of the Sm₂O₃(N)-OM NPs result in higher CH₄

conversions and C₂selectivities at nearly all conditions used in the

investigation. The highest C₂yield obtained over this catalyst was

7.8% at a CH₄:O₂ratio of 4:1 and a reaction temperature of 740 ^◦C

with a GHSV of 880 h⁻¹. This is not an impressive yield, but it can

be signiﬁcantly improved using typical dopants, such as alkaline

earth metals. While several catalysts can produce a reaction yield

greater than that obtained over samaria, it remains one of the best

performing single compound OCM catalysts [9]. Calcination of pure

Sm₂O₃NPs at 1000 ^◦C instead of 800 ^◦C results in particle sintering

and thus a lower surface area with a concomitant reduction in both

activity and selectivity. Only under certain conditions, and if the

catalyst is prepared using a chloride precursor, is there a positive

effect from a higher calcination temperature, but the best perform-

ing catalyst is still the Sm₂O₃(N)-OM NPs prepared using a nitrate

precursor.

References

[1] S. Liu, F. Chen, S. Xie, P. Zeng, X. Du, L. Xu, J. Nat. Gas Chem. 18 (2009)

21–24.

[2] J. Orejas, Chem. Eng. Sci. 56 (2001) 513–522.

[3] W. Liu, A. Eliyas, L. Petrov, Appl. Catal. 61 (1990) 265–274.

[4] H. Cai, A. Krzywicki, M. Oballa, Chem. Eng. Proc. Process Intensiﬁcation 41

(2002) 199–214.

[5] G.E. Keller, M.M. Bhasin, J. Catal. 73 (1982) 9–19.

[6] K. Otsuka, K. Jinno, A. Morikawa, J. Catal. 100 (1986) 353–359.

[7] T. Ito, J.H. Lunsford, Nature 314 (1985) 721–722.

[8] K. Fujimoto, K. Omata, J. Yoshihara, Appl. Catal. 67 (1991) L21–L24.

[9] O. Forlani, S. Rossini, Mater. Chem. Phys. 31 (1992) 155–158.

[10] H.B. Zhang, G.D. Lin, H.L. Wan, Y.D. Liu, W.Z. Weng, J.X. Cai, Y.F. Shen, K.R. Tsai,

Catal. Lett. 73 (2001) 141–147.

[11] D.J. Driscoll, W. Martir, J.X. Wang, J.H. Lunsford, J. Am. Chem. Soc. 107 (1985)

58.

As rare earth oxides are in limited supply, ﬁnding ways to min-

imize the samaria content in a catalyst while still maintaining

the beneﬁts of the pure compound are of great interest. While

a typical Al₂O₃support is not expected to yield highly selective

catalysts, they are readily available with high surface areas. There-

fore, three different Al₂O₃supports with varying properties were

investigated to determine if Al₂O₃is a viable support. The Al₂O₃-

supported catalysts do exhibit a lower C₂selectivity, likely due to

participation from acidic sites on the support, which increases CO_x

formation. In particular, there is signiﬁcantly more CO produced

over the supported catalysts compared with the Sm₂O₃NPs. The

catalyst preparation method appears to inﬂuence the CH₄conver-

sion more than the speciﬁc Al₂O₃support used, despite signiﬁcant

differences between the Al₂O₃support surface areas and acidic

and basic properties. In contrast, the C₂selectivity depends on

both the catalyst preparation method and the Al₂O₃support used.

The best supported catalyst is obtained by impregnating the n-

Al₂O₃(−) support with an aqueous solution of samarium nitrate.

As the C₂selectivity increases with reaction temperature between

740 ^◦C and 800 ^◦C over the supported catalysts, the highest C₂yield

is obtained at a CH₄:O₂ratio of 4:1 and a temperature of 800 ^◦C.

The impregnation method using the nitrate precursor likely yields

the smallest Sm₂O₃particles on the surface of the Al₂O₃supports,

while at the same time blocking a signiﬁcant portion of the highly

acidic, CO_xproducing, alumina sites. The CH₄conversion over the

Sm₂O₃/Al₂O₃(−)-IM catalyst is the same as over the Sm₂O₃(N)-OM

(the best performing pure Sm₂O₃NP catalyst), but the C₂selec-

tivity is slightly lower. While the C₂yield per gram of catalyst is

lower for the supported catalysts, the C₂yield per gram of Sm₂O₃

is higher over several of the Al₂O₃-supported catalysts. Therefore,

by selectively blocking some of the acidic support sites, it will be

possible to increase the C₂selectivity and thus make the catalysts

[12] D. Wang, M.P. Rosynek, J.H. Lunsford, J. Catal. 151 (1995) 155–167.

[13] J.B. Kimble, J.H. Kolts, Energy Prog. 6 (1986) 226–229.

[14] K. Aika, T. Moriyama, N. Takasaki, E. Iwamatsu, J. Chem. Soc. Chem. Commun.

(1986) 1210–1211.

[15] E. Iwamatsu, T. Moriyama, N. Takasaki, K. Aika, J. Chem. Soc. Chem. Commun.

(1987) 19–20.

[16] T. Moriyama, N. Takasaki, E. Iwamatsu, K. Aika, Chem. Lett. (1986) 1165–

1168.

[17] N. Yamagata, K. Tanaka, S. Sakaki, Z. Okazaki, Chem. Lett. (1987) 81–82.

[18] W. Hinsen, W. Bytyn, M. Baerns, Proc. Eight Int. Congress Catal. 3 (1984)

581–592.

[19] J.P. Bartek, J.M. Hupp, J.F. Brazdil, R.K. Grasselli, Catal. Today 3 (1988) 117–126.

[20] K. Asami, S. Hashimoto, T. Shikada, K. Fujimoto, H. Tominaga, Ind. Eng. Chem.

Res. 26 (1987) 1485–1488.

[21] M.Y. Lo, S.K. Agarwal, G. Marcelin, J. Catal. 112 (1988) 168–175.

[22] S.K. Agarwal, R.A. Migone, G. Marcelin, Appl. Catal. 53 (1989) 71–80.

[23] K.D. Campbell, H. Zhang, J.H. Lunsford, J. Phys. Chem. 92 (1988) 750–753.

[24] J.M. Deboy, R.F. Hicks, J. Catal. 113 (1988) 517–524.

[25] A.M. Gaffney, C.A. Jones, J.J. Leonard, J.A. Sofranko, J. Catal. 114 (1988) 422–432.

[26] H.L. Wan, X.P. Zhou, W.Z. Weng, R.Q. Long, Z.S. Chao, W.D. Zhang, M.S. Chen,

J.Z. Luo, S.Q. Zhou, Catal. Today 51 (1999) 161–175.

[27] T. Baidya, N. van Vegten, Y. Jiang, F. Krumeich, A. Baiker, Appl. Catal. A 391

(2011) 205–214.

[28] K. Otsuka, K. Jinno, A. Morikawa, Chem. Lett. (1985) 499–500.

[29] K. Otsuka, K. Jinno, Inorg. Chem. Acta 121 (1986) 237–241.

[30] K. Otsuka, T. Komatsu, Chem. Lett. (1987) 483–484.

[31] J.A. Labinger, K.C. Ott, J. Phys. Chem. 91 (1987) 2682–2684.

[32] H. Imai, T. Tagawa, J. Chem. Soc. Chem. Commun. (1986) 52–53.

[33] K.P. Peil, James G. Goodwin Jr., G. Marcelin, J. Am. Chem. Soc. 112 (1990)

6129–6130.

[34] M.P. Rosynek, Catal. Rev. - Sci. Eng. 16 (1977) 111–154.

[35] K.R. Tsai, D.A. Chen, H.L. Wan, H.B. Zhang, G.D. Lin, P.X. Zhang, Catal. Today 51

(1999) 3–23.

[36] K.R. Tsai, H.B. Zhang, G.D. Lin, H.L. Wan, Y.D. Liu, W.Z. Weng, J.X. Cai, Y.F. Shen,

Abstracts of Papers CATL 028, ACS Nat. Mtg., Boston, 1998.

[37] S.V. Mahajan, J.H. Dickerson, Nanotechnology 18 (2007) 32605–32610.

[38] W. Zhu, L. Xu, J. Ma, R. Yang, Y. Chen, J. Colloid Interface Sci. 340 (2009) 119–

125.

[39] S.D. Jones, L.M. Neal, H.E. Hagelin-Weaver, Appl. Catal. B 84 (2008) 631–

642.

[40] S.D Jones, Activity and Characterization Studies in Methanol Reforming Catal-

ysis: Cu and Cu-ZnO Catalysts and the Role of Nanomaterials, Dissertation,

2008.

even more effective per unit weight of Sm₂O₃, and render Al₂O₃

viable catalyst support for the oxidative coupling of methane.

a

[41] M.J. Capitan, P. Malet, M.A. Centeno, A. Munoz-Paez, I. Carrizosa, J.A. Odriozola,

J. Phys. Chem. 97 (1993) 9233–9249.

[42] S.N. Vereshchagin, J.R.H. Ross, Catal. Today 24 (1995) 285–287.

[43] C.L. Huang, Y.C. Chen, Mater. Res. Bull. 37 (2002) 563–574.

[44] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-Ray Photo-

electron Spectroscopy, Eden Prairie, Minnesota, 1995.

Acknowledgements

The authors gratefully acknowledge NSF support (Chemistry

1026712). Dr. Valentin Craciun and Dr, Eric Lambers are also

Article Doi

Oxidative coupling of methane over unsupported and alumina-supported samaria catalysts

DOI: 10.1016/j.apcata.2013.01.010

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.apcata.2013.01.010

Full text of DOI:10.1016/j.apcata.2013.01.010

Products guided by the article

R&D Labs maybe for 124-38-9

Relevant to this article

Hot Product