Catalytic combustion of methane on substituted strontium ferrites

Fuel 90 (2011) 1245–1256

Contents lists available at ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel

Catalytic combustion of methane on substituted strontium ferrites

Marina V. Bukhtiyarova ⇑, Aleksandra S. Ivanova, Elena M. Slavinskaya, Lyudmila M. Plyasova,Vladimir A. Rogov, Vasily V. Kaichev, Aleksander S. NoskovBoreskov Institute of Catalysis SB RAS, Pr. Akad. Lavrentieva, 5, 630090 Novosibirsk, Russia

a r t i c l e i n f o

Article history:Received 18 May 2010Received in revised form 3 November 2010Accepted 3 November 2010Available online 20 November 2010

Keywords:HexaferriteXRDH2-TPRXPSMethane oxidation

0016-2361/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.fuel.2010.11.012

⇑ Corresponding author. Tel.: +7 383 3269772; fax:E-mail address: [email protected] (M.V. Bukhtiyaro

a b s t r a c t

Sr–hexaferrites prepared by co-precipitation method and calcined at 700–1000 �C have been character-ized by thermogravimetric and differential thermal analysis (TG–DTA), Fourier transformed infraredspectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), hydrogentemperature-programmed reduction (H2-TPR), and Ar adsorption techniques. It has been shown thathexaferrite phase formed after calcination at 700 �C is amorphous and its crystallization occurs at800 �C. Specific surface area (SBET) of the samples calcined at 700 �C is 30–60 m2/g. Reduction in hydrogenproceeds in several steps, Fe(III) in the hexaferrite structure being practically reduced to Fe0. Amount ofhydrogen necessary for the reduction of the samples decrease in the order: SrMn2Fe10O19 > SrFe12O19 >SrMn6Fe6O19 > SrMn2Al10O19. Surface composition of the ferrites differs from bulk. According to XPS data,the surface is enriched with strontium. Sr segregation is most probably explained by the formation ofsurface carbonates and hydroxocarbonates. The main components on the surface are in oxidized states:Mn3+ and Fe3+. Maximum activity in the methane oxidation is achieved for the SrMnxFe12�xO19 (0 6 x6 2)catalysts. These samples are characterized by highest amount of the hexaferrite phase, which promoteschange of oxidation state Mn(Fe)3+

M Mn(Fe)2+.� 2010 Elsevier Ltd. All rights reserved.

1. Introduction of the catalysts applied for afterburning gas emissions in spite of

At the present, unburned hydrocarbons, CO and NOx are mainair pollutants in the cities which are produced by a vast amountof gasoline and diesel internal combustion engine vehicles. Oneof the ways to resolve this ecological problem is to use alternativefuels such as methane or natural gas. However, a major drawbackassociated with its use is the emission of unburned methane, aneffective greenhouse gas which contributes to global atmospherewarming even more than CO2 due to its longer lifetime [1]. There-fore, to comply with modern legislation, highly efficient catalystsfor the complete abatement of unburned methane are needed.

Noble metal-based catalysts like Pt and/or Pd supported onmixed metal-oxides are typically employed for the oxidation ofhydrocarbons [2]. Reducing the high cost of noble metal catalystsusing inexpensive metals for the complete oxidation of hydrocar-bons remains the major research target. Recently, it has been dem-onstrated that doped metal-oxides, perovskites, spinels, and othermixed metal-oxides are active for the methane oxidation [3]. Ex-tremely high stability of methane compared to other hydrocarbonsdemands relatively high temperatures for the methane oxidation.Moreover, high exothermicity of this reaction: CH4 + 2O2 = CO2 +2H2O (DH298 = �802.7 kJ/mol) [4] results in additional overheating

ll rights reserved.

+7 383 3308056.va).

low concentration of methane in exhaust feed. Therefore, the cat-alysts developed for the methane oxidation should be not onlyhighly active, but also should be thermally stable in the tempera-ture range of 800–1000 �C.

Recently it was shown [5,6] that substituted hexaaluminatescan be applied for the methane oxidation. For Mn(Fe)-substitutedhexaaluminates, which differ by nature of components and theirratio, the temperature (T50) corresponding to 50% CH4 conversionare in the range of 560–776 �C. The introduction of iron ions till to-tal substitution of aluminum cations with iron ones and obtainingBaFe12O19 hexaferrite are accompanied by decrease of T50 to 533 �C[7]. It can be supposed that further T50 decrease can be achieved byusing Mn-substituted hexaferrites.

The aim of this study was to synthesize substituted Sr–hexafer-rites and to investigate the influence of nature and content of addi-tive (Mn, Al) on their structure, chemical state of components,texture and catalytic performance in the methane oxidation.

2. Experimental

2.1. Catalyst preparation

Hexaferrites SrMnxFe12�xO19 (x = 0, 1, 2, 6), SrMn6Fe4Al2O19 andSrMn2Al10O19 marked as SF, SM1F, SM2F, SM6F, SM6FA2 andSM2A10, respectively, were prepared by co-precipitation of soluble

1246 M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256

nitrates of Sr, Fe and M (M = Mn(II), Al (III)) at pH = 7.2–7.5 andtemperature of 70 �C with NH4HCO3 as a precipitating agent. Theslurry was aged at 70 �C for 2 h and then filtered [8]. SrMnx-Fe12�xO19 (x = 1, 2) samples marked as SM1F* and SM2F* wereprepared in the same way. Difference is that hydrogen peroxidesolution has been previously added to solution of Mn(II) nitratefor transformation of Mn(II) to Mn(III). The obtained precipitatewas washed and dried in air. The solid was dried at 110 �C for12–14 h and then calcined at 700 �C for 4 h in air flow. The subse-quent calcination of the samples was carried out at 800 and1000 �C for 4 h in a muffler. Except for the SM2A10 sample, thesamples calcined at 700 �C were aged at 800 �C for 14 h in the10% H2O/air gas mixture.

2.2. Catalyst characterization

The catalysts were characterized using elemental analysis, Aradsorption, thermogravimetric and differential thermal analysis,X-ray diffraction, Fourier transformed infrared spectroscopy, X-ray photoelectron spectroscopy, and hydrogen temperature-programmed reduction techniques.

Elemental analysis was performed using the ICP atomic absorp-tion spectroscopy with an accuracy of 0.01–0.03% [9]. The specificsurface area was determined with an accuracy of ±10% by the ther-mal desorption of argon [10].

The thermogravimetric and differential thermal analysis wascarried out on a NETZSCH STA 449C apparatus. Catalysts weretested over the temperature range from room temperature up to1200 �C at the heating rate of 10 �C/min in air. The accuracy ofdetermination of weight losses was ±0.5%.

XRD studies were performed on a ARL X‘TRA diffractometerusing Cu Ka monochromatic radiation (k = 1.5418 Å). X-ray diffrac-tion patterns were recorded in a step scan mode in the 2h rangefrom 10� to 75o with the step of 0.02–0.05o and 3–5 s per stepdepending on the sample crystallinity. The phase identificationwas performed by comparison of the measured set of the interpla-nar distances di and the corresponding intensities of the diffractionmaximums Ii with that found in the ICDD, PDF-2.

Cell parameters of a-Fe2O3 (tetragonal structure) are calculatedfrom equation:

1

d2 ¼h2 þ k2

a2 þ l2

c2 ;

where, h, k, l are Miller Indices, d is interplanar distances, which isdetermined from the experiment. The corresponding number of themost accurate, not overlapping and quite intensive lines is selectedfor calculation of the cell parameters. It is desirable to select thelines in the region of large angles since at the same accuracy ofdetermination of angles the accuracy of determination of interpla-nar distances increases. The least-squares method is employed forspecification of the cell parameters that increases reliability ofdetermination since all possible lines are used. The accuracy ofdetermination of the cell parameters a and c is ±0.005 and±0.01 Å, respectively.

FTIR spectra were recorded in the range of 250–2000 cm�1 on aBomem MB-102 spectrometer. To take the spectra, the sampleswere prepared by the pelletizing with CsI.

X-ray photoelectron spectroscopy was applied for characteriza-tion of the surface contents of the elements. XPS measurementswere performed on a SPECS’s machine equipped with an X-raysource XR-50M with a twin Al/Ag anode, an ellipsoidal crystalmonochromator FOCUS-500, and a hemispherical electron energyanalyzer PHOIBOS-150. The core-level spectra were typically ob-tained using monochromatic Al Ka radiation (h = 1486.74 eV) andfixed analyzer pass energy of 20 eV under ultrahigh vacuum condi-

tions. During XPS measurements, the static charge was minimizedby a flood gun of electrons. For further calibration of the chargeshift, C1s peak at 284.8 eV from adventitious hydrocarbon wasused. Spectra were background-subtracted using a Shirley fit algo-rithm [11] and then fitted onto separate components. Doniach–Sanjic symmetric function was applied for peak approximation[12]. To quantify the atomic concentration of the present elements,the cross-sections according to Scofield [13] were used.

H2-TPR measurements were carried out in a flow reactorequipped with a thermal conductivity detector. The samples of0.2 g with an average granule size of 0.25–0.5 mm were previouslypretreated in oxygen at 500 �C for 30 min, and then they werecooled to room temperature. A 10 vol.% H2/Ar stream (40 ml/min)was passed over the sample while it was heated from 40 to900 �C at the heating rate of 10 �C/min.

2.3. Catalyst activity test

The catalyst activity in the methane oxidation was tested inaccordance with temperature-programmed reaction (light-off test)in the temperature range from 100 to 625 �C at the heating rate of10 �C/min using a mixture of 0.1 vol.% CH4, 20 vol.% O2, 0.5 vol.% Nebalanced with He at a total flow rate of 500 ml/min. Catalystvolume was 0.6 cm3, space velocity was 50,000 h�1. The gas com-position was analyzed by a quadruple mass-spectrometer SRSQMS-200.

3. Results and discussion

3.1. Phase composition

Non-isothermal temperature-programmed treatments of thefresh air-dried samples resulted in the appearance of exo- andendoeffects on the TG–DTA curves (Fig. 1). When the precursorswere heated from room temperature up to 1200 �C, they lost from22% (SF) to 28% (SM6F) or even 36% (SM6FA2) of their weight. Theendoeffects observed at 128–136 �C were caused by the dehydra-tion of hexaferrites. The effects at 219–266 �C are probably relatedto the decomposition of ammonia salts of the corresponding com-ponents. According to [14], the endoeffects observed at 375–405 �Cfor the SM6F and SM6FA2 samples can be caused by the decompo-sition of iron carbonate. The endoeffects at 462–464 and 511–543 �C are related to the decomposition of manganese carbonateto MnO2 followed by the transformation of MnO2 to Mn2O3

[15,16]. The endoeffects at 840–856 �C are originated due to thedecomposition of strontium carbonate [17]. The endoeffects at980–995 �C can be related both to transition Mn2O3 ? Mn3O4

[16] and formation of spinel (MnxFe1�x)3O4 [18].The exoeffects appears at 650–697 �C and 798 �C for the substi-

tuted and unsubstituted ferrites, respectively (Fig. 1). Exoeffect at798 �C can be caused by crystallization of the SrFe12O19 phase,since BaFe12O19 crystallization occurs at 760 �C [19]. Probably, exo-effect at 650–697 �C is related to crystallization of Sr manganite.Thus, the thermal genesis of the ferrites exhibits the complexbehavior pointing out the presence of the phase transformation.

XRD data (Figs. 2 and 3; Table 1) show that the samples calcinedat 700 �C contain mainly SrCO3 and a-Fe2O3 crystalline phases,whereas SM2A10 hexaaluminate system calcined at the same tem-perature is amorphous [20]. At the same time the SM6F andSM6FA2 samples contains also the Mn2O3 phase (Fig. 2c; Table 1)and the phases which were not identified. It should be stressedthat the phase composition of the SM1F and SM1F* samples cal-cined at 700 �C depends on valence of manganese in the feedstock(Fig. 3a). The introduction of Mn(III) in the sample leads to

Fig. 1. TG–DTA curves of the samples (a) SF, (b) SM6F, (c) SM6FA2.

20 30 40 50 60 70

(a)

1000oC, 4h800oC, 14h (H2O + air)800oC, 4h

700oC, 4h

vvvvvv

vvv oooo

o

o

ooooo

ooooo

o

o

o

o

oo

oo

o

ooo

++

++++

++

+

+

Inte

nsity

2 theta

20 30 40 50 60 70

o

++

oooo

o

ooo

oo

oooo

oo

oo

o

oo

o

o

oo

++++

(b)

Inte

nsity

vvv vv

v

vo

vv

vv

v v

1000oC, 4h

800oC, 14h (H2O+air)

700oC, 4h

2 theta

20 30 40 50 60 70

ooooo

o

oo

oo

o

o

oo

o

oo

oo

o

o

o

o

o

o **

***

*

+++

(c)

Inte

nsity

om

vm

vvmv

m

vmvm

mmv

m

v

m

*

v

v

*

*

1000oC, 4h

800oC, 14h (H2O+air)

700oC, 4h

2 theta

Fig. 2. XRD patterns of the samples (a) SF, (b) SM1F, (c) SM6F calcined at differenttemperature. (+) SrCO3; (s) SrFeO3; (m) Mn2O3; (*) SrMn3O6; (v) Fe2O3; (o)SrFe12O19.

M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256 1247

decrease in the amount of the SrCO3 phase. This result can be con-sidered as increase in the interaction between components.

In good agreement with TG–DTA data, increasing the calcina-tion temperature up to 800 �C promotes crystallization andappearance of the SrFe12O19 phase for the SF sample (Fig. 2a).The SF sample contains the a-Fe2O3 and SrCO3 phases as well.The treatment of the sample at 800 �C for 14 h in the H2O/air gasmixture results in the formation of the same phases, exceptingSrCO3 (Fig. 2a, Table 1). Besides, amount of the hexaferrite phaseincreases and amount of the a-Fe2O3 phase decreases. Thus, ‘‘age-ing’’ the sample in the H2O/air gas mixture accelerates the crystal-

lization of Sr–hexaferrite. The phase composition of the ‘‘aged’’SM1F, SM2F and SM6F samples differs (Table 1). Except for theSrFe12O19 and a-Fe2O3 phases, as the amount of manganese in-creases, the SM2F and SM6F samples contain the slight amountof the Mn2O3 and Mn2O3 + SrMn3O6 phases, respectively. The ob-tained results show that limited amount of manganese is able tobe dissolved in the hexaferrite structure. One can see (Fig. 2) thatan increase in the amount of manganese in the SrMnFe11O19 com-position leads to an increase in the amount of the hexaferrite phaseand to a decrease in the amount of the additional phases. Thefurther increase in the amount of Mn in the SrMn6Fe6O19 samplepromotes an increase in the amount of the additional manga-nese-containing phases (Fig. 2c).

Calcination of the samples at 1000 oC leads to an increase in theamount of the hexaferrite phase and to a decrease in the amount ofthe additional phases: a-Fe2O3, Mn2O3 and SrMn3O6 (Table 1). Cal-cination of the SF sample at 1000 �C promotes the formation of

20 30 40 50 60 70

oo

(a)

oooo

oo

oo

oooo

o

o

o

oo

o

o

oooo vvv

vvv

vv

v

+++

+++

++

SM1F* (1000oC)

SM1F (1000oC)

SM1F* (700oC)

SM1F (700oC)

Inte

nsity

2 theta

28 29 30 31 32 33 34 35 36 37 38 39 40

o o

(b)

mm vv oooooooooo

SM2F*, 1000oC, 4h

SM2F, 1000oC, 4h

SM2F, 800oC, 14h (H2O + air)

SM2F*, 800oC, 14h (H2O + air)

Inte

nsity

2 theta

Fig. 3. XRD patterns of the samples (a) SM1F and SM1F* calcined at different temperature, (b) SM2F and SM2F* calcined at different temperature. (+) SrCO3; (m) Mn2O3;(v) Fe2O3; (o) SrFe12O19.

1248 M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256

practically single-phase hexaferrite. The sample contains a smallamount of the a-Fe2O3 phase as well.

Comparison of the XRD patterns of the SM1F and SM1F*, SM2Fand SM2F* samples calcined at 1000 �C shows (Fig. 3a and b) thatthe introduction of Mn(III) promotes an increase in the amount ofthe a-Fe2O3 phase and a decrease in the amount of the hexaferritephase. For the SM2F* sample, the introduction of Mn(III) leads tothe disappearance of the Mn2O3 phase (Fig. 3b). Thus, manganeseions completely enter in the hexaferrite structure when manga-nese is introduced as Mn(III). This leads to an increase in theamount of the a-Fe2O3 phase.

Phase composition of the SM6F sample calcined at 1000 �C re-mains practically constant (Fig. 2c): ratio of the hexaferrite andadditional phases doesn’t change in comparison with phase com-position of the sample aged at 800 �C. A distinctive feature of theSM6FA2 and SM2A10 samples is the fact that the hexaferrite(hexaaluminate) phase is not detected by XRD even after calcina-tion at 1000 �C. Moreover, the SrAl2O4 phase is formed in theSM2A10 sample (Table 1).

Unit cell parameters of the a-Fe2O3 phase which is observed inall samples calcined at 700 �C change as follows (Fig. 4): a param-eter decreases from 5.04 to 4.97 Å with the introduction of Mn andMn-Al, whereas c parameter remains practically unchanged (theaccuracy of determination of a and c is ±0.005 and ±0.01 Å, respec-tively). Decrease in the cell parameter a with respect to the refer-ence parameter (Table 1) is evidenced of the promoting iron oxideby manganese and aluminum cations and the formation of solidsolution on basis of a-Fe2O3. The change of parameter c lies inthe range of the accuracy (13.74–13.75 Å). Thus, the introduction

of Mn and Al in the a-Fe2O3 structure does not affect thisparameter.

Correct determination of the unit cell parameters of the corre-sponding phases was difficult due to complicated phase composi-tion. For predominant hexaferrite phase a = 5.88 Å, c = 22.99 Åirrespective of the treatment conditions and the introduced addi-tives (Mn, Al). The invariance of the a and c parameters of the hexa-ferrite phase can be probably caused by accuracy of itsdetermination (a: ±0.005; c: ±0.01 Å) or a similarity of ionic radiiof Fe3+ and Mn3+. It is well-known that iron enters in the hexafer-rite structure as a trivalent cation [21]. Difference in ionic radii ofFe3+ and Mn3+ is insignificant (0.67 vs. 0.70 Å), whereas ionic radiusof Mn2+ is considerably higher than that of Fe3+ (0.91 vs. 0.67 Å).Since a and c cell parameters of the hexaferrite remain practicallyconstant, it can be supposed that Mn2+ ions don’t enter into thestructure.

Thus, investigation of the phase composition of the Sr–ferritesshowed that they contain the SrCO3 and a-Fe2O3 phases after cal-cination at 700 �C. The samples with increased amount of manga-nese (SM6F and SM6FA2) contain additional phases. ‘‘Ageing’’ thesamples at 800 �C or its calcination at 1000 �C is accompanied bycrystallization of the hexaferrite phase; its amount is significantlyhigher than amount of the additional phases.

3.2. FTIR spectroscopy

FTIR spectra of the SF sample calcined at 700 and 1000 �C areshown in Fig. 5a. FTIR spectrum (Fig. 5a, curve 2) indicate thatthe absorbance bands at 288, 304, 359, 396, 438, 548, 591 and

4.96

4.97

4.98

4.99

5.00

5.01

5.02

5.03

5.04

5.05

SM6FA2SM6FSM2FSM1FSF

c, A

(α-F

e 2O3)

a, A

( α-F

e 2O3)

13.72

13.73

13.74

13.75

13.76

13.77

13.78

Fig. 4. Changes of unit cell parameters of a-Fe2O3 in the samples calcined at 700 �C.

Table 1Phase composition of hexaferrites.

Sample Phase composition

700 �C, 4 h 800 �C (10%H2O/air),14 h

1000 �C, 4 h

SF Fe2O3

SrCO3

Fe2O3

SrFe12O19

SrFeO3�x

Fe2O3

SrFe12O19

SM1F Fe2O3

SrCO3

Fe2O3

SrFe12O19

Mn2O3

Fe2O3

SrFe12O19

SM1F* Fe2O3

SrCO3

– Fe2O3

SrFe12O19

SM2F Fe2O3

SrCO3

– Fe2O3

SrFe12O19

Mn2O3 (traces)SM2F* – – Fe2O3

SrFe12O19

SM6F Fe2O3

SrCO3

Mn2O3

Fe2O3

SrFe12O19

Mn2O3

SrMn3O6

Fe2O3

SrFe12O19

Mn2O3

SrMn3O6

SM6FA2 Fe2O3

Mn2O3

uncertain phases

– Uncertainphases

SM2A10 Amorphous – SrAl2O4

SrFe12O19 a = 5.886, c = 23.03 ICDD, PDF-2, [00-33-1340]a-Fe2O3 a = 5.035, c = 13.74 ICDD, PDF-2. [00-33-0664]SrFeO3�x a = 3.852 ICDD, PDF-2. [00-40-0906]Mn2O3 a = 9.408 ICDD, PDF-2. [00-41-1442]SrMn3O6�x a = 9.131, b = 2.82,

c = 12.09ICDD, PDF-2. [00-28-1233]

2000 1750 1500 750 500 250

2

1

(a)

30433

0;32

437

739

7

360

Abs

orba

nce

(a.e

.)

Wavenumber, cm-1

2000 1750 1500 750 500 250

Wavenumber, cm-1

458;

438

54859

2

232

323

376

470

552

698

705

859

1456

(b)

3

2

1

857

1457

Abs

orba

nce

(a.e

.)

321

330

492

536

574

666

482

531

572

85714

57

323

470

546

859

1456

Fig. 5. FT-IR spectra of the samples (a) SF calcined at 700 �C (1) and 1000 �C (2),(b) SM1F (1), SM6F (2) and SM6FA2 (3) calcined at 700 �C,

M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256 1249

779 cm�1, relating to SrFe12O19 [22], and low-intensive absorbancebands at 323, 377, 458 and 799 cm�1, relating to a-Fe2O3 [23,24],are observed for the SF sample calcined at 1000 oC. Thus, obtainedresults agree with XRD data, which show that this sample containsthe SrFe12O19 and a-Fe2O3 phases. FTIR-spectrum of the SF samplecalcined at 700 �C (Fig. 5a, curve 1) shows the 232, 322–329, 376(shoulder), 470, 552 and 593 cm�1 absorbance bands which aretypical for the a-Fe2O3 phase [23,24]. The 705–698, 858, 1071–1055 and 1456 cm�1 bands correspond to vibrations of CO2�

3 inSrCO3 [25]. Furthermore, the 329, 553 and 593 cm�1 absorbancebands which correspond to Sr–hexaferrite phase are also observed.Hence, except for SrCO3 and a-Fe2O3, the SF sample calcined at700 �C also contains SrFe12O19. The reason of non-observation ofthe hexaferrite phase by XRD (Table 1) is probably its presencein the highly dispersed state.

For the SM1F, SM6F and SM6FA2 samples calcined at 700 oCthe positions of the absorbance bands and their intensities differ.The presence of manganese in the SM1F sample promotes slightchange of FTIR-spectrum (Fig. 5b, curve 1). It contains practicallythe same absorbance bands as the SF sample; however, the absor-bance bands corresponding to the hexaferrite phase are absent.Moreover, according to [26], an additional absorbance band at1609 cm�1 can be related to molecules of adsorbed water. The fur-ther increase of manganese content in the SM6F sample leads tothe appearance of the 531, 571 and 667 cm�1 absorbance bands,which are typical for Mn2O3 [27,28]. At the same time, the 323,481, 571, 705, 733, 857, 1071, 1457 and 1616 cm�1 absorbancebands, which point out the presence of SrCO3, a-Fe2O3 and mole-cules of adsorbed water, are also observed (Fig. 5b, curve 2). How-ever, amount of these phases in the SM6F sample is smaller thanthat in the SM1F sample. The introduction of aluminum ions inthe SM6FA2 sample practically doesn’t influence on the absor-bance bands positions in the FTIR spectra (Fig. 5b, curve 3). Theabsorbance bands at 320–329, 385, 455–491, 535–537, 567–574,722, 799, 856–857, 1410–1420, 1457–1458 and 1610–1615 cm�1 are typical for a-Fe2O3, Mn2O3, SrCO3 and adsorbedwater.

Thus, TG–DTA, XRD and FTIR data show that the samples cal-cined at 700oC are multiphase. The SrFe12O19 hexaferrite, whichis amorphous, SrCO3 and a-Fe2O3 are observed in the SF sample.The introduction of manganese leads to the transformation of thephase composition: the SrCO3, a-Fe2O3 and Mn2O3 phases areobserved, whereas the SrFe12O19 phase is not detected. An

1250 M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256

additional promoting Mn-containing sample by aluminum doesn’tlead to essential changes in the phase composition, since the SrCO3,a-Fe2O3 and Mn2O3 phases are observed. Ionic radius of Al3+ ions issmaller than that of Mn3+ and Fe3+ ions (0.57 vs. 0.70 and 0.67 Å),thus, probably, Al3+ ions enter in the a-Fe2O3 and Mn2O3 structure.This fact is confirmed by decrease in the a and c cell parameters ofthe a-Fe2O3.

3.3. Textural properties

Together with formation of the phase composition, changes ofspecific surface area occur. The specific surface area depends onthe nature of the components, their ratio and the calcination tem-perature. SBET of the samples calcined at 700 oC is in the range of30–59 m2/g (Table 1). It should be noted that the introduction ofaluminum promotes the formation of the samples with higher spe-cific surface area (Table 1). For the SM2A10 sample this value is145 m2/g, probably, as a result of formation of highly dispersedaluminates which is not detected by XRD. Treatment of the sam-ples at 800 oC for 14 h in the H2O/air gas mixture leads to decreaseof SBET to 7–16 m2/g. SBET of the samples calcined at 1000 oC is lessthan 1 m2/g.

3.4. Temperature-programmed reduction by hydrogen

Since the SF, SM2F and SM6FA2 samples calcined at 700 �C weremultiphase, H2-TPR experiments was carried out for the estimationof their redox properties. One can see (Fig. 6) that the reductionproceeds in several steps depending on the nature of the compo-nents and their ratio. The H2-TPR profile of the SF sample is shownin Fig. 6, curve 1. The peaks with maximum at 408, 520, 708 �C areobserved. According to [29], the first peak corresponds to thereduction of Fe2O3 to Fe3O4. The second peak is associated withthe reduction of Fe3O4 to FeO + Fe. The third low-intensive peakcan be related to the reduction of traces of the hexaferrite phase.

The H2-TPR profile of the SM2F sample is similar (Fig. 6, curve2). The reduction of the SM2F sample proceeds in three steps withmaximum at 424, 544 and 700 �C. The first peak is associated withthe reduction of Fe2O3 and Mn2O3 to Fe3O4 and Mn3O4, respec-tively. The second peak is related to the reduction of Fe3O4

and Mn3O4 to FeO + Fe and MnO, respectively. According to [30],the reduction of Fe3�xMnxO4 oxide proceeds in two steps which

0 200 400 600

5

1

1

2

2

3

mol

H2/g

*s

70

520

408

70

544

424

533

455

382

Temperature, oC

mol

H2/g

*s

Fig. 6. TPR-profiles of the samples (1) SF, (2) SM2F,

are indicated by peaks at 540 and 700 �C. Thus, the second andthird peaks in TPR-profile of the SM2F sample also can be relatedto the reduction of traces of the Fe3�xMnxO4 spinel and Sr–hexaferrite.

The TPR-profile of the SM6FA2 sample is different (Fig. 6, curve3). All peaks are shifted to the higher temperature region and anadditional peak at 382 �C is observed. It is known [31] that thereduction of manganese oxide supported on alumina proceeds byfollowing scheme:

3Mn2O3 þH2 !330 �C

2Mn3O4 þH2O

Mn3O4 þH2 !407 �C

3MnOþH2O

Probably, reduction behavior of the SM6FA2 sample depends onaluminum introduced in the sample. At the same time, reductiontemperatures of Mn2O3 and Fe2O3 are shifted to the lower temper-ature region in comparison with ones related to the reduction ofthe SM2F sample. Thus, peak at 382 �C is associated with thereduction of Fe2O3 and Mn2O3 to Fe3O4 and Mn3O4, respectively.The peak at 455 �C is related to the reduction of Mn3O4 to MnOand Fe3O4 to FeO + Fe. The peaks with maximum at 533 and816 �C can be caused by the reduction of most strong-bonded ionsof Mn and Fe that agrees well with the reduction behavior ofhexaaluminate.

The reduction of the SrMn2Al10O19 hexaaluminate proceeds inthe same way as the reduction of the SM6FA2 sample. The evi-dence of this fact is the presence of two intense peaks at 382 and466 �C (Fig. 6, curve 4). These peaks are related to the reductionMn2O3 to Mn3O4 and Mn3O4 to MnO, respectively. TPR-profileshows one incomplete peak at 895 �C. It means that a part of theMn3+ ions is in a more strongly bonded state (probably, they are lo-cated in the positions of the spinel block of the hexaaluminatestructure [3]) and is not reduced in the studied temperature range(600–900 �C).

According to equation Fe(Mn)2O3 + H2 ? 2Fe(Mn)O + H2O, thereduction of Mn2O3 or Fe2O3 requires 6.3 lmol H2/gcatalyst. Basedon this value, the amount of hydrogen, which is necessary for thereduction of the involved Mn2O3 and/or Fe2O3 phases, was calcu-lated. One can see (Table 2) that the amount of hydrogen con-sumed in the reduction of the SF, SM2F and SM6FA2 samples is12.5, 13.7 and 9.1 lmol H2/gcatalyst, respectively, which is signifi-cantly higher than the calculated values (5.7, 5.7 and 5.0 lmol

800 1000

4

0 200 400 600 800 1000

0.0

.0x10-7

.0x10-6

.5x10-6

.0x10-6

.5x10-6

.0x10-6

Temperature, oC

43

2

18

0

816

(3) SM6FA2 and (4) SM2A10 calcined at 700 �C.

Table 2Influence of composition of samples on their reduction.

Sample Tcalc (�C) Content (wt.%) Reduction temperature (�C) H2 consumption (lmol H2/g)

Mn2O3 Fe2O3 Experiment Calculation

FeO Fe

SF 700 – 90.2 408 520 708 12.5 5.7 17.0SM2F 700 15.0 75.2 424 544 700 13.7 5.7a 15.1a

SM6FA2 700 47.4 32.0 382 455 533 816 9.1 5.0a 9.0a

SM2A10 700 23.6 – 382 466 >600 1.98 – –

a Mn2O3 is reduced to MnO.

Table 3Atomic concentration of elements on hexaferrite surface determined by XPS as well as bulk concentrations from chemical analysis.

Sample [Sr]/[Fe] [Sr]/[Mn] [Mn]/[Fe] [Fe]/[Al] [O]/[RMe]

SF XPS 0.18 – – – 1.8Chemical analysis 0.08 – – 1.5

SM1F XPS 0.27 2 0.14 – 1.9Chemical analysis 0.09 1 0.09 – 1.5

SM6F XPS 0.27 0.45 0.59 – 1.9Chemical analysis 0.17 0.17 1 – 1.5

SM6FA2 XPS 0.53 0.50 1.0 0.48 1.8Chemical analysis 0.25 0.17 0.67 2 1.5

SrCO3 XPS – – – – 2.4

M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256 1251

H2/gcatalyst). Taking into account the fact that the reduction of Fe2O3

to Fe0 requires 18.8 lmol H2/g Fe2O3, the obtained experimentalvalues demonstrate that the reduction of the samples proceedswith the formation of Fe0. At the same time, the amount of hydro-gen of 12.5 lmol H2/gcatalyst, which is necessary for the reduction ofthe SF sample, indicates that 40% Fe2O3 is reduced to FeO(2.26 lmol H2/gcatalyst) and 60% Fe2O3 is reduced to Fe0

(10.20 lmol H2/gcatalyst).For Mn-containing samples, an increase in the amount of hydro-

gen necessary for the reduction of the SM2F sample is observed.This amount is 13.7 lmol H2/gcatalyst. Mn2O3 can be reduced onlyto MnO in the investigated temperature range. Therefore, it canbe supposed that the presence of manganese in the hexaferritestructure promotes the highest reduction of Fe2O3 to Fe0 in com-parison with that for the SF sample. This is confirmed by the calcu-lation: 13.7 lmol H2/gcatalyst corresponds to the reduction of Mn2O3

to MnO (0.94 lmol H2/gcatalyst), 15% Fe2O3 to FeO (0.71 lmol H2/gcatalyst) and 85% Fe2O3 to Fe0 (12.00 lmol H2/gcatalyst). Similar ef-fect is observed when aluminum ions are introduced in theSM6FA2 sample (Table 2). The calculation shows that the amountof hydrogen of 9.1 lmol H2/gcatalyst corresponds to the reductionof Mn2O3 to MnO (2.96 lmol H2/gcatalyst) and Fe2O3 to Fe(6.14 lmol H2/gcatalyst) without the reduction of Fe2O3 to FeO.

The amount of hydrogen consumed on the reduction of theSM2A10 sample is 1.98 lmol H2/gcatalyst (Table 2) which is signifi-cantly higher than the calculated value (1.29 lmol H2/gcatalyst). It iscan be caused by a poor crystallinity of the samples and differentstates of the components on the surface. Namely, the ‘‘surface’’contribution to the reduction leads to the differences in the reduc-tion degree.

Thus, the key feature of the hexaferrite reduction is the forma-tion of Fe0. The amount of metallic iron increases with the intro-duction of manganese and aluminum ions in the samples.

3.5. Surface distribution and states of various elements in hexaferrites

The XPS data indicate that the surface concentrations ofelements in the samples differ from the bulk chemical contents

(Table 3). One can see, the surface of all samples is enriched bySr. Atomic ratios [Sr]/[Fe] and [Sr]/[Mn] are significantly higherthan stoichiometric values (Table 3). Sr segregation is mostprobably explained by the formation of surface carbonates, oxides,and hydroxides. Indeed, oxygen content with respect to the totalamount of cations exceeds stoichiometric value. Moreover, XPSindicate the presence of surface carbonates in the samples. Twofeatures at 284.8 and 289.2 ± 0.3 eV are observed in the C1s spectra(not shown here). The first one corresponds to carbon in hydrocar-bon impurities, whereas the second one corresponds to carbon incarbonate groups [32,33]. It should be noted that the segregationof strontium is very often observed for Sr doped mixed oxides[34,35].

The formation of carbonate, hydroxocarbonate, oxides orhydroxides of strontium on the surface makes difficulties for anal-ysis of the Sr3d spectrum. The Sr3d spectrum is known to representas the unresolved spin–orbit Sr3d5/2–Sr3d3/2 doublet (intensity ra-tio 3:2 and splitting of ca. 1.79 eV). Fig. 7 shows the Sr3d spectrumfor the samples and SrCO3. The Sr3d spectra for the catalysts, ex-cept for the SM6F sample, are described by two doublets withthe Sr3d5/2 binding energies of 132.1–132.6 and 133.6–134.1 eV.According to literature data [36,37], the binding energy for thestrontium ions in SrO, Sr(OH)2 and SrCO3 is in the range of131.7–132.4, 132.8 and 133.4–133.8 eV, respectively. The Sr3d5/2

binding energy of La1�xSrxCrO3 mixed oxide is 131.5–131.7 eV[35]. Correspondingly, the first doublet can be attributed to Sr2+

in the ferrite structure, and the second one can be attributed tothe hydroxocarbonate phase [34,35]. Indeed, the doublet withthe Sr3d5/2 binding energy of 133.8 eV predominates in the spec-trum of SrCO3. Moreover, the relative intensity of the C1s peak at289.2 ± 0.3 eV corresponds to the molar ratio [CO3]/[Sr] of 0.8–1.3. The Sr3d spectrum for the SM6F sample is described by onedoublet with the Sr3d5/2 binding energy of 133.3 eV.

The Fe2p spectrum for the samples contains the spin–orbitFe2p3/2–Fe2p1/2 doublet with maximum intensities at 710.4 and723.9 eV and with the intensity ratio of 2:1 (Fig. 7a). Two addi-tional weak satellites at 718.6 and 733.4 eV are due to chargetransfer multielectron transitions. Usually, the determination of

128 130 132 134 136 138

133.8

132.2

5

2

3

4

1

XPS

Sr3

d In

tens

ity [

arb.

un.

]

Binding Energy [eV]

a

700 705 710 715 720 725 730 735 740

Sat.Sat.

Fe2p3/2

Fe2p3/2

2

3

4

1

XPS

Fe2

p In

tens

ity [

arb.

un.

]

Binding Energy [eV]

b

636 638 640 642 644 646 648

3

21

644.3

641.9

XPS

Mn2

p 3/2

Inte

nsity

[ar

b. u

n.]

Binding Energy [eV]

c

Fig. 7. XPS spectra for: (a) Sr3d of (1) SF, (2) SM1F, (3) SM6F, (4) SM6FA2 and (5)SrCO3; (b) Fe2p of (1) SF, (2) SM1F, (3) SM6F, (4) SM6FA2; (c) Mn2p3/2 of (1) SM1F,(2) SM6F, (3) SM6FA2. The Fe2p lines were used as references.

Table 4Binding energy of surface elements of hexaferrites (eV).

Sample Chemical composition Fe2p3/2 Mn2p3/2 Al2p

SF SrFe12O19 710.4 – –SM1F SrMnFe11O19 710.6 641.9 –SM6F SrMn6Fe6O19 710.3 641.7 –SM6FA2 SrMn6Fe4Al2O19 710.9 641.9 73.9

1252 M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256

the chemical state of iron was carried out using both the Fe2p3/2

binding energy and the shape of the Fe2p spectrum, i.e. presenceor absence of «shake-up satellites». Iron in FeO, Fe3O4, Fe2O3 andFeOOH are characterized by the Fe2p3/2 binding energies in therange of 709.3–709.7, 709.7–710.6, 710.4–711.2 and 710.2–711.05 eV, respectively [38,39]. The feature which differ the Fe2pspectrum of FeO from the Fe2p spectrum of Fe2O3 is the presenceof weak «shake-up satellites» localized on 6 and 8 eV above thebasic photoelectron lines of Fe2p3/2 and Fe2p1/2 [38,39]. The satel-lite structure is absent in the case of Fe3O4. For the ferrites theFe2p3/2 binding energy is in the range of 710.3–710.9 eV (Table 5).The similar value of the Fe2p3/2 binding energy (710.6 eV) is ob-served for NiFe2O4 and CoFe2O4 mixed oxides [38]. Hence, theweak satellite at 718.6 eV which is observed in the Fe2p spectraof the samples indicates the formation of the Fe3+ ions. The intro-duction of Mn leads to decrease in the Fe2p3/2 binding energy,while the introduction of Al leads to its increase (Table 4). Recentlyit was shown [5,40], the Fe2p3/2 binding energy of the LaFeAl11O19

hexaaluminate is 711.4–711.8 eV that is typical for Fe3+. A decreaseof the Fe2p3/2 binding energies from 711.4 to 709.4 eV after theintroduction of Mn was observed for LaFeMn5Al6O19.

Fig. 7b shows the Mn2p3/2 spectra of the samples under studies.The spectra exhibit asymmetric shape. Therefore, two peaks wereapplied for the approximation of experimental curves. The mainpeak corresponds to the Mn2p3/2 line. The second less-intensivepeak shifted to the higher binding energies by 2.4 eV is originated

due to multiplet splitting [41]. The Mn2p3/2 binding energy of641.8–641.9 eV is typical for Mn3+ ions. The similar values of641.0–641.5 eV have been observed for Mn3+ in LaMnO3,La1�xSrxMnO3 oxides [42]. Manganese in MnO, Mn2O3 and MnO2

are characterized by the Mn2p3/2 binding energies in the range of640.6–641.7, 641.5–641.9 and 642.2–642.6 eV, respectively [43].

The Al2p binding energy for the SM6FA2 sample is 73.9 eV(Fig. 7c, Table 5). This value is typical for Al3+ in the hexaaluminatestructure. For comparison, aluminum in a-Al2O3 and AlOOH arecharacterized by the Al2p binding energies of 74.2 eV and73.9 eV, respectively [44]. The Al2p binding energy for the LaFe-Al11O19�d, LaMnAl11O19�d hexaaluminates is in the range of 74.1–74.8 eV [5,40].

Hence, the XPS data showed that iron and manganese aremainly in oxidized state Fe3+ and Mn3+. The samples surface is en-riched with strontium. Sr segregation is most probably explainedby the formation of surface carbonates. This hypothesis is con-firmed by excess of the ratio of oxygen content to the total amountof cations above the stoichiometric value and the appearance of thecorresponding features from the strontium carbonates in Sr3d andC1s spectra. Nevertheless, part of strontium enters in the hexafer-rite structure.

3.6. Catalytic properties of the catalysts

Fig. 8 shows the ‘‘light-off’’ tests in the methane oxidation forthe SF, SM1F, SM2F, SM6F, SM6FA2 and SM2A10 catalysts calcinedat 700 �C. One can see that reaction starts at 300 �C. Temperature of10% CH4 conversion (T10) increases in the order (Fig. 8a):

SrMn2Fe10O19T10 340

< SrFe12O19343

< SrMnFe11O19347

< SrMn6Fe6O19365

< SrMn6Fe4Al2O19380

< SrMn2Al10O19443

The H2 amount consumed during the reduction of the samplesdecreases in the same order:

SrMn2Fe10O19H2 13:7

> SrFe12O1912:5

> SrMnFe11O19 > SrMn6Fe6O19

> SrMn6Fe4Al2O199:1

> SrMn2Al10O192:0

It is well-known that the total reduction degree determines theamount of accessible oxygen. Hence, the initial catalytic activity ofthe samples for the methane oxidation correlates with the amountof accessible oxygen.

Temperatures corresponding to 50% and 90% CH4 conversion(T50, T90) are given in Table 5. The lowest T50 value of 410 �C is ob-served for the SF sample. It is seen that the elevation of the amountof manganese entered in the hexaferrite structure correlates withan increase in T50. The T50 value is 418 and 447 �C for the SM1Fand SM6F samples, respectively. Similarly, the introduction ofAl3+ cations (SM6FA2) till total substitution of the iron cations withaluminum ones (SM2A10) is accompanied by the drastic decreasein the catalytic activity. T50 characteristic for the SrMn2Al10O19

sample (538 �C) is higher by 120 �C than that for SrMn2Fe10O19

(418 �C). Whereas the T50 value for Fe2O3 is 451 �C (Table 5).The specific surface area does not influence on the catalytic

activity of the samples obtained. SBET of a-Fe2O3 is 7.5 m2/g that

10.5

11

11.5

12

12.5

13

1.3 1.35 1.4 1.45 1.5 1.55 1.6 1.65 1.7 1.75 1.81000/T, 1/K

lgW

SF

SM1F

SM2F

SM6F

SM6FA2

SM2A10

SM1F*

SM2F*

(a)

11.8

12

12.2

12.4

12.6

12.8

13

13.2

13.4

lgW

SFSM1FSM2FSM6FSM6FA2

(b)

Table 5Catalytic activity expressed as T10%, T50% and T90% for both fresh and aged hexaferrites.

# Catalyst Tcalc

(oC)SBET

(m2/g)Catalytic activity in CH4

oxidationEa

(kJ/mol)

T10% T50% T90%

FreshSF SrFe12O19 700 30 343 410 462 91SM1F SrMnFe11O19 700 34 347 418 471 97SM1F* 36 343 413 468 92SM2F SrMn2Fe10O19 700 38 340 418 473 98SM2F* 33 334 411 466 94SM6F SrMn6Fe6O19 700 27 365 447 510 80SM6FA2 SrMn6Fe4Al2O19 700 59 380 463 532 92SM2A SrMn2Al10O19 700 145 443 538 606 91F Fe2O3 700 7.5 373 451 514 –

AgedSF SrFe12O19 800 7.0 391 479 549 100SM1F SrMnFe11O19 800 10 371 461 528 97SM1F* 11 371 455 528 –SM2F SrMn2Fe10O19 800 9 377 463 530 95SM2F* – 386 469 549 –SM6F SrMn6Fe6O19 800 6.5 417 525 >610 82SM6FA2 SrMn6Fe4Al2O19 800 16 421 522 598 82

M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256 1253

is 4–5-fold lower than SBET of the concerned samples (Table 1). Theactivity of the samples studied practically does not depend on SBET.In spite of higher value of SBET for the SM6FA2 and SM2A10 sam-

0

10

20

30

40

50

60

70

80

90

100

(a)

CH

4 con

vers

ion,

%

SF SM1F SM2F SM6F SM6FA2 SM2A10

100 200 300 400 500 600

0

10

20

30

40

50

60

70

80

90

100

CH

4 con

vers

ion,

%

Temperature, оС

100 200 300 400 500 600

Temperature, оС

(b)

(b)

SF SM1F SM2F SM6F SM6FA2

Fig. 8. Temperature dependences of the CH4 conversion for hexaferrites catalystscalcined at (a) 700 �C in air and (b) 800 �C in water vapor (H2O + air).

11.61.3 1.35 1.4 1.45 1.5 1.55 1.6 1.65

1000/T, 1/K

Fig. 9. Arrenius plots for hexaferrites catalysts calcined at (a) 700 �C in air and (b)800 �C in water vapor (H2O + air).

ples compared to SBET of the remaining samples (59 and 145 vs.30–39 m2/g), they are less active (Table 5).

Estimation of values of activation energy (Ea) of methane oxida-tion was carried out on initial data of light-off curves (Fig. 8) withchanging methane conversion (X) from 2% to 20%. To estimate Ea athigher values of conversion is not desirable since methane concen-tration over catalyst should be constant. The values of the apparentactivation energy were estimated from the Arrhenius plots on thebase of a CH4 first-order rate equation. Reaction rate is propor-tional to expression:

W � C0=100 � X � NA � V22;400 �m � SBET � 1000

;

where C0 is initial concentration of methane, X is conversion ofmethane, NA is Avogadro constant, V (cm3/sec) is space velocity ofmethane, m (kg) is weight of catalyst, SBET is specific surface area.

According to results obtained, the values of apparent Ea for thesamples calcined at 700 �C lies in the range 80–98 kJ/mol, the high-est activation energy being observed over the SM2F sample(Fig. 9a). The similar values of activation energy are obtained forthe rest of samples besides SM6F. Activation energy for this sampleis 80 kJ/mol (Table 5). It should be noted that the accuracy of deter-mination of the values of Ea is ±8 kJ/mol. The treatment of the sam-ples at 800 �C in the water vapor leads to slightly change of the Ea

values (Fig. 9b). According to [45], activation energy of methaneoxidation in the presence of BaMgAl10O17 and BaMgMnAl10O17 is157 and 86 kJ/mol, respectively, that agrees with our data.

4. Discussion

According to opinion of Jang et al. [46], the catalytic activity inthe methane oxidation mainly is determined by the presence of re-dox sites in the catalyst. The higher being amount of cations with

1254 M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256

highest oxidation state the higher catalytic activity. As it was men-tioned above, the samples calcined at 700 �C contain the SrCO3 anda-Fe2O3 phases (Table 1). At the same time, for the SF sample, theSrFe12O19 hexaferrite phase is observed in highly dispersed state.For the SM6F, SM6FA2 and SM2A10 samples the small amount ofthe Mn2O3 or Mn2O3 + SrMn3O6 phases is present. Probably, theSM6FA2 and SM2A10 samples can also contain aluminum-contain-ing phases. Hence, the activity of the samples is mainly determinedby presence of the a-Fe2O3 phase since strontium carbonate is notactive in the methane oxidation. However, the XPS data show thatthe surface of the samples contains excess of surface oxygen abovethe stoichiometric value as a result of surface segregation of Srleading to formation of surface carbonates or hydroxocarbonates(Table 3). Thus, superstoichiometric oxygen is present on the sam-ple surface; its presence generally promotes an increase in the cat-alytic activity. Furthermore, manganese and iron in the ferritesunder studies is mainly in oxidized state: Fe3+ and Mn3+.

The activity of the catalysts aged at 800 �C for 14 h in the H2O/air gas mixture decreases in comparison with the activity of thefresh catalysts (Fig. 8b). Comparison of T50 values of the freshand aged samples shows (Table 5) that the difference in the tem-peratures is in the range of 42–78 �C. At the same time, the mostactive catalysts are SM1F and SM1F*, the T50 value for these sam-ples is 461 and 455 �C, respectively. The different change of activityis observed for the SM2F and SM2F* samples aged at 800 �C for14 h in the presence of water vapor. T10, T50, T90 increases fromSM2F to SM2F* (Table 5), i.e. activity of the sample obtained usingMn(III) is lower than that of the sample obtained using Mn(II).According to the XRD data, these samples differ by ratio of presentphases (Table 1): amount of the a-Fe2O3 phase is higher andamount of the Mn2O3 phase is lower in the SM2F* sample than thatin the SM2F sample. The differences are revealed in greater extentafter treatment of the samples at 1000 �C (Fig. 3b). The introduc-tion of increased amount of manganese and aluminum is accompa-nied by increase in the T50 value, for the SM6F and SM6FA2samples, this value being practically the same (Table 5).

As it was mentioned above, ‘‘ageing’’ the samples is accompa-nied by crystallization of the hexaferrite phase. Amount of the a-Fe2O3 phase changes and SrCO3 is absent (Table 1). Thus, activityof the ‘‘aged’’ samples can be caused by different ratio of theSrFe12O19 and a-Fe2O3 phases and higher crystallinity. At the same

Fig. 10. Crystal structure of (a) a-Fe2O3 [4

time, it can be noted that an increase in the amount of the hexafer-rite phase leads to activity increase. This fact is confirmed by theT50 value for the SM2F and SM2F* samples differing by ratio ofthe indicated phases, namely: T50 increases (Table 5) with increas-ing the amount of a-Fe2O3 in the SM2F* sample (Fig. 3b).

According to [47], a-Fe2O3 is characterized by a hexagonalstructure in which octahedral sites are occupied by Fe3+ ions(Fig. 10a). Sr–hexaferrite has megnetoplumbite structure(Fig. 10b). The unit cell contains five different crystallographic sitesfor Fe ions. The part of Fe ions is octahedrally coordinated{Fe(4)2O9, Fe(5)3O8, Fe(1)O6}. The Fe ions also form Fe(2)O5 trigonalbyparimds and Fe(3)O4 tetrahedron [21]. Probably, presence ofboth octahedral and tetrahedral sites in the hexaferrite structuredetermines its higher activity as a result of possibility of changeof oxidation state Fe3+

M Fe2+ in comparison with a-Fe2O3.For the substituted Sr–ferrites presence of both manganese and

iron ions able to change the oxidation state Mn(Fe)3+M Mn(Fe)2+

will facilitate the methane oxidation since manganese introductionin the hexaferrite structure results in increase in oxygen mobility(H2 consumption is 13.7 and 12.5 lmol H2/g for the SM2F and SF sam-ples, respectively). However, the results obtained show that there isa limit of manganese concentration for maximum activity ofsubstituted Sr–hexaferrites (Table 5). Analogous conclusion wasmade by authors [5]: among substituted LaFeMnxAl11�xO19 hexa-aluminates the most active is LaFeMnAl10O19; the LaF-eMnxAl11�xO19 catalysts with [ > 2 are characterized by loweractivity. Limited amount of manganese is able to be dissolved inthe hexaferrite structure, probably, as a result of presence ofMn3+ ions with larger ionic radius than that of Fe3+ ions. Thus, in-crease in manganese amount to composition SM6F leads to forma-tion of additional phases, its activities are lower in comparisonwith activity of Sr–hexaferrite.

Thus, observed differences in the catalytic activity of the sam-ples aged at 800 �C in the presence of water vapor can be causedby the different ratio of the hexaferrite and a-Fe2O3 phases. In-crease in the amount of hexaferrite in the samples promotes an in-crease in its activity due to structure features of the hexaferritephase providing presence of redox cycle Fe3+

M Fe2+. The introduc-tion of manganese in the hexaferrite structure to compositionSM1F also results in activity increase. Probably, it is caused by in-crease of mobility of lattice oxygen associated with defectiveness

7] and (b) SrFe12O19 hexaferrite [21].

M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256 1255

of the hexaferrite, which is formed after replacement of Fe3+ byMn3+ with larger ionic radius facilitating methane oxidationMn(Fe)3+ + O2�

lattice M 2Mn(Fe)2+ + ½O2.

5. Conclusions

Series of the SrMnxFe12�xO19 (x = 0, 1, 2, 6) hexaferrites was syn-thesized by co-precipitation of solution of corresponding nitrateswith NH4HCO3 as a precipitating agent at fixed temperature andpH. The samples were characterized by different physico-chemicalmethods and were tested in the methane oxidation. The phasecomposition, specific surface area, redox behaviors and catalyticperformance strongly depend on the manganese and/or aluminumloading. Some specific conclusions can be drawn as follows:

� The unsubstituted SrFe12O19 sample calcined at 700 �C containsSrCO3 and a-Fe2O3. The hexaferrite phase is in highly dispersedstate, and this phase is not detected by XRD. Increasing the cal-cination temperature up to 800–1000 �C promotes the crystalli-zation of the hexaferrite phase. The introduction of Mn(II) orAl(III) ions leads to the appearance of the Mn2O3 and/orMn2O3 + SrMn3O6 phases and some phases which were notidentified at this stage. The introduction of Mn(III) in the sampleat the precipitation stage promotes its more complete enteringin the hexaferrite structure and decrease in the amount of addi-tional Mn-containing phases. Ageing of the samples at 800 �Cfor 14 h in the H2O/air gas mixture promotes the hexaferritecrystallization, whereas the amount of the additional phasesdecreases.� Specific surface area of the SrMnxFe12�xO19 (x = 0, 1, 2, 6) sam-

ples calcined at 700 �C is in the range of 30–59 m2/g. The intro-duction of aluminum in the samples promotes an increase ofSBET. Ageing of the samples at 800 �C in the presence of watervapor leads to decrease in the specific surface area to 7–16 m2/g.� The samples calcined at 700 �C are multiphase. This fact causes

complex behavior of its reduction. The reduction of the samplesproceeds in several steps: Fe2O3(Mn2O3) ? Fe3O4(Mn3O4),Fe3O4(Mn3O4) ? FeO(MnO) + Fe0. The reduction temperatureof the first peak is shifted to the lower temperature region withthe introduction of manganese and aluminum ions in the sam-ples. The key feature of the hexaferrite reduction is formation ofmetallic iron, the amount of which increases with the introduc-tion to SrFe12O19 of such promoting additives as Mn and Al. Theamount of hydrogen necessary for the reduction of the samplesdecreases from 13.7 to 2.0 lmol H2/gcatalyst in the order:SrMn2Fe10O19 > SrFe12O19 > SrMn6Fe6O19 > SrMn2Al10O19.� The samples surface is enriched with strontium. Sr segregation

is most probably explained by the formation of surface carbon-ates or surface oxides/hydroxides. This hypothesis is confirmedby XPS. Nevertheless, part of strontium enters in the hexaferritestructure. Iron and manganese are mainly in oxidized state Fe3+

and Mn3+. It is remarkable that the introduction of manganesein the Sr–hexaferrite leads to decrease in the Fe2p3/2 bindingenergy, and the introduction of aluminum results in its increase.� Hexaferrite compositions are active catalysts in the methane

oxidation. Its initial catalytic activity increases with an increasein the amount of accessible oxygen and decreases with anincrease in amount of manganese and/or aluminum in the cat-alyst. The higher being amount of surface oxygen whichincreases with manganese introduction the higher activity ofthe hexaferrites. The catalysts are characterized by high thermalstability since its treatment at 800 �C for 14 h in the H2O/air gasmixture doesn’t lead to significant decrease in the catalyticactivity.

Acknowledgements

The authors wish to thank kindly collaborators from BoreskovInstitute of Catalysis: G.S. Litvak for TG–DTA measurements, I. Yu.Molina for taking X-ray diffraction patterns, G.N. Kustova for theexperimental support in the FTIR measurements, V.V. Mokrinskiifor ageing catalysts in the H2O/air gas mixture and I. A. Polukhinafor activity tests.

References

[1] Arosio F, Colussi S, Groppi G, Trovarelli A. Regeneration of S-poisoned Pd/Al2O3

catalysts for the combustion of methane. Catal Today 2006;117:569–76.[2] Gelin P, Primet M. Complete oxidation of methane at low temperature over

noble metal based catalysts: a review. Appl Catal B 2002;39:1–37.[3] Choudhary TV, Banerjee S, Choudhary VR. Catalysts for combustion of methane

and lower alkanes. Appl Catal A 2002;234:1–23.[4] Alegre VV, Da Silva MAP, Schmal M. Catalytic combustion of methane over

palladium alumina modified by niobia. Catal Commun 2006;7:314–22.[5] Ren X, Zheng J, Song Y, Liu P. Catalytic properties of Fe and Mn modified

lanthanum hexaaluminates for catalytic combustion of methane. CatalCommun 2008;9:807–10.

[6] Artizzu-Duart P, Millet JM, Guilhaume N, Garbowski E, Primet M. Catalyticcombustion of methane on substituted barium hexaaluminates. Catal Today2000;59:163–77.

[7] Groppi G, Cristiani C, Forzatti P. BaFexAl(12�x)O19 system for high-temperaturecatalytic combustion: physico-chemical characterization and catalytic activity.J Catal 1997;168:95–103.

[8] Groppi G, Cristiani C, Forzatti P. Preparation, characterisation and catalyticactivity of pure and substituted La-hexaaluminate systems for hightemperature catalytic combustion. Appl Catal B 2001;35:137–48.

[9] Price WJ. Analytical atomic-absorption spectroscopy. New York; 1976.[10] Lowell S, Shields JE, Thomas MA, Thommes M. Characterization of porous

solids and powders: surface area, pore size and density. The Netherlands:Springer; 2006.

[11] Shirley DA. High-resolution X-ray photoemission spectrum of the valencebands of gold. Phys Rev B 1972;5:4709–14.

[12] Doniach S, Sanjic M. Many-electron singularity in X-ray photoemission and X-ray line spectra from metals. J Phys C 1970;3:285–91.

[13] Scofield JH. Hartree–Slater subshell photoionization cross-sections at 1254and 1487 eV. J Electron Spectrosc Relat Phenom 1976;8:129–37.

[14] Nekrasov BV. Essentials of general chemistry. 2nd ed. Moscow: Himiya; 1973(in Russian).

[15] Deraz NAM, El-Shobaky GA. Solid–solid interaction between ferric oxide andmanganese carbonate as influenced by lithium oxide doping. ThermochimActa 2001;375:137–45.

[16] Shaheen WM, Selim MM. Thermal decompositions of pure and mixedmanganese carbonate and ammonium molybdate tetrahydrate. J Therm AnalCalorim 2000;59:961–70.

[17] Xiujing Z, Yan F, Qing H, Gang C. Microwave technology synthesized thecathode materials of solid oxide fuel cell La1�xSrxMnO3. Chin J Mater Res2008;22:539–44.

[18] Guillemet-Fritsch S, Navrotsky A, Tailhades P, Coradin H, Wang M.Thermochemistry of iron manganese oxide spinels. J Solid State Chem2005;178:106–13.

[19] Roos W. Formation of chemically coprecipitated barium ferrite. J Am CeramSoc 1980;63:601–3.

[20] Lietti L, Cristiani C, Groppi G, Forzatti P. Preparation, characterization andreactivity of Me–hexaaluminate (Me = Mn, Co, Fe, Ni, Cr) catalysts in thecatalytic combustion of NH3-containing gasified biomasses. Catal Today2000;59:191–204.

[21] Kimura K, Ohgaki M, Tanaka K, Morikawa H, Marumo F. Study of thebipyramidal site in magnetoplumbite-like compounds, SrM12O19 (M = Al, Fe,Ga). J Solid State Chem 1990;87:186–94.

[22] Zhao WY, Wie P, Cheng HB, Tang XF, Zhang QJ. FTIR spectra, lattice shrinkage,and magnetic properties of CoTi-substituted M-type barium hexaferritenanoparticles. J Am Ceram Soc 2007;90:2095–103.

[23] Yurchenko EN, Kustova GN, Batsanov SS. Vibrational spectra of inorganiccompounds. Novosibirsk: Science; 1981 (in Russian).

[24] Onari S, Ari T, Kudo K. Infrared lattice vibrations and dielectric dispersion in a-Fe2O3. Phys Rev B 1977;16:1717–21.

[25] Nakamoto K. FTIR-spectra and Raman spectra of inorganic and coordinationcompounds. Moscow: Mir; 1991 (in Russian).

[26] Valange S, Beauchaud A, Barrault J, Gabelica Z, Daturi M, Can F. Lanthanumoxides for the selective synthesis of phytosterol esters: correlation betweencatalytic and acid–base properties. J Catal 2007;251:113–22.

[27] Das DP, Parida KM. Mn(III) oxide pillared titanium phosphate (TiP) for catalyticdeep oxidation of VOCs. Appl Catal A 2007;324:1–8.

[28] Morales MR, Barbero BP, Cadús LE. Evaluation and characterization of Mn–Cumixed oxide catalysts for ethanol total oxidation: influence of copper content.Fuel 2008;87:1177–86.

1256 M.V. Bukhtiyarova et al. / Fuel 90 (2011) 1245–1256

[29] Liang M, Kang W, Xie K. Comparison of reduction behavior of Fe2O3, ZnO andZnFe2O4 by TPR technique. J Nat Gas Chem 2009;18:110–30.

[30] Oliveira LCA, Fabris JD, Rios RRVA, Mussel WN, Lago RM. Fe3�xMnxO4 catalysts:phase transformations and carbon monoxide oxidation. Appl Catal A: Gen2004;259:253–9.

[31] Álvarez-Galván MC, Pawelec B, De la Peña O’Shea VA, Fierro JLG, Arias PL.Formaldehyde/methanol combustion on alumina-supported manganese-palladium oxide catalyst. Appl Catal B: Environ 2004;51:83–91.

[32] Khassin AA, Yurieva TM, Kaichev VV, Bukhtiyarov VI, Budneva AA, PaukshtisEA, et al. Metal-support interactions in cobalt–aluminum co-precipitatedcatalysts: XPS and CO adsorption studies. J Mol Catal A 2001;175:189–204.

[33] Rida K, Benabbas A, Bouremmad F, Pena MA, Merthinez-Arias A. Surfaceproperties and catalytic performance of La1�xSrxCrO3 perovskite-type oxidesfor CO and C3H6 combustion. Catal Commun 2006;7:963–8.

[34] Bukhtiyarova MV, Ivanova AS, Plyasova LM, Litvak GS, Rogov VA, Kaichev VV,et al. Selective catalytic reduction of nitrogen oxide by ammonia on Mn(Fe)-substituted Sr(La)-aluminates. Appl Catal A 2009;357:193–205.

[35] Rousseau S, Loridant S, Delichere P, Boreave A, Deloume JP, Vernoux P.La(1�x)SrxCo1�yFeyO3 perovskites prepared by sol–gel method: characterizationand relationships with catalytic properties for total oxidation of toluene. ApplCatal B 2009;88:438–47.

[36] Sosulnikov MI, Teterin YA. X-ray photoelectron studies of Ca, Sr and Ba andtheir oxides and carbonates. J Electron Spectrosc Relat Phenom1992;59:111–26.

[37] Dupin JC, Gonbeau D, Vinatier P, Levasseur A. Systematic XPS studies of metaloxides, hydroxides and peroxides. Phys Chem Chem Phys 2000;2:1319–24.

[38] McIntyre NC, Zetaruk DG. X-ray photoelectron spectroscopic studies of ironoxides. Anal Chem 1977;49:1521–9.

[39] Graat PCJ, Somers MAJ. Simultaneous determination of composition andthickness of thin iron-oxide films from XPS Fe 2p spectra. Appl Surf Sci1996;100–101:36–40.

[40] Xu Z, Zhao L, Pang F, Wang L, Niu C. Partial oxidation of methane to synthesisgas over hexaaluminates LaMAl11O19-catalysts. J Natur Gas Chem2007;16:60–3.

[41] Nesbitt HW, Banerjee D. Interpretation of XPS Mn(2p) spectra of Mnoxyhydroxides and constraints on themechanism of MnO2 precipitation. AmMineral 1998;83:305–15.

[42] Ponce S, Pena MA, Fierro JLG. Surface properties and catalytic performance inmethane combustion of Sr-substituted lanthanum manganites. Appl Catal B2000;24:193–205.

[43] Di Castro V, Polzonetti G. XPS study of MnO oxidation. J Electron SpectroscRelat Phenom 1989;48:117–23.

[44] Kosova N, Devyatkina E, Slobodyuk A, Kaichev V. Surface chemistry study ofLiCoO2 coated with alumina. Solid State Ionics 2008;179:1745–9.

[45] Astier M, Garbowski E, Primet M. BaMgAl10O17 as host matrix for Mn in thecatalytic combustion of methane. Catal Lett 2004;95:31–7.

[46] Jang BWL, Nelson RM, Spiveya JJ, Ocal M, Oukaci R, Marcelin G. Catalyticoxidation of methane over hexaaluminates and hexaaluminate-supported Pdcatalysts. Catal Today 1999;47:103–13.

[47] Pailhé N, Wattiaux A, Gaudon M, Demourgues A. Impact of structural featureson pigment properties of a-Fe2O3 haematite. J Solid State Chem2008;181:2697–704.