ISSN 00201685, Inorganic Materials, 2011, Vol. 47, No. 1, pp. 36–40. © Pleiades Publishing, Ltd., 2011. Original Russian Text © Yu.G. Morozov, O.V. Belousova, M.V. Kuznetsov, 2011, published in Neorganicheskie Materialy, 2011, Vol. 47, No. 1, pp. 41–46.
Preparation of Nickel Nanoparticles for Catalytic Applications Yu. G. Morozov, O. V. Belousova, and M. V. Kuznetsov Institute of Structural Macrokinetics and Materials Science, Russian Academy of Sciences, ul. Akademika Osip’yana 8, Chernogolovka, Noginskii raion, Moscow oblast, 142432 Russia email:
[email protected] Received January 14, 2010
Abstract—Spherical oxidized nickel particles 15 to 200 nm in average size have been produced by a crucible less aerosol method involving metal vapor condensation in an inert gas flow and oxidation processes. The par ticles have been characterized by scanning electron microscopy, Xray microanalysis, Xray diffraction, BET surface area measurements, and vibratingsample magnetometry. The process parameters have been opti mized for the preparation of particles with tailored size, specific surface area, and saturation magnetization. A dc electric field applied to the condensation zone during the oxidation process reduces the size and increases the extent of oxidation of the particles. We have studied lowtemperature oxidation of carbon mon oxide and propane on nickel nanopowders differing in particle size and extent of oxidation. The nanoparticles with optimized characteristics have been shown to have a marked catalytic effect on these processes. DOI: 10.1134/S0020168510121027
INTRODUCTION
was applied in some cases to the particle condensa tion/cooling zone. The particles thus produced were entrained by the gas flow through a cooler and then col lected on filter cloth. The composition and crystal structure of the resultant nanomaterials were determined by Xray diffraction on a DRON3M powder diffractometer (CuKα radiation). Powder morphology was examined by scanning electron microscopy (SEM) using Ultra Plus and LEO 1450 instruments (Carl Zeiss, Germany). The latter was equipped with an Oxford Instruments INCA Energy 300 energy dispersive Xray analysis system, which was used to perform elemental analysis of the materials. Electron micrographs were analyzed using SIAMS600 image processing software. The specific surface area of loose nanopowder samples was determined by fourpoint nitrogen BET measurements using a META SORBIM instrument. Magnetic characteristics of slightly com pacted nanopowder samples (for vibrofixation in a nylon ampule) were measured at room temperature in mag netic fields of up to 0.8 MA/m using an M4500 vibrating sample magnetometer (EG&G PARC, USA).
Nanomaterials possess many valuable physical prop erties [1]. The key issues in nanomaterials research can be briefly formulated as follows: targeted synthesis of nano particles of controlled size and shape, investigation of their fundamental physical properties, and the use of such nanometersized standard building blocks to assemble ensembles as key components of multifunc tional devices. The objectives of this work were to produce ensem bles of nickel nanoparticles of controlled average size and oxidation state, investigate their functional characteris tics, and determine their potential catalytic applications. EXPERIMENTAL Nickel nanopowders were prepared by a modified crucibleless levitation jet method [2, 3]. In this method, a droplet of metallic nickel is suspended in a quartz tube of 14mm inner diameter and heated to melting and vapor ization onset by the electromagnetic field (0.44 MHz) of a countercurrent inductor. The droplet levitates in the tube and is exposed to a downward inert gas (Ar or He) stream, responsible for the formation of an evaporant condensation zone, which determines the size of the resulting particles. The source material used was NP1 nickel wire 0.2 mm in diameter, which continuously fed the droplet with nickel at a constant feed rate using a pur posedesigned feeder. Owing to the selfpurification effect [2], the metal vapor was free from refractory impu rities, which remained in the droplet. To obtain oxide nanoparticles, O2 gas was added to the flow upstream or downstream (in the nanoparticle condensation zone) of the droplet [4]. In the latter instance, a dc electric field
RESULTS AND DISCUSSION Morphological and structural characterization. Fig ure 1 shows SEM micrographs of various nanoparticles of two average sizes. The larger particles (sample 005), consisting almost entirely of pure Ni (Fig. 1a), are seen to be spherical in shape (typical of the particles obtained), with a slight scatter in size and partial agglomeration due to magnetic attraction forces. The small particles (sample 015), consisting almost entirely of pure NiO (Fig. 1b), are cubic in shape [5]. Using a comparative morphological analysis of projections of particles in SEM micrographs 36
PREPARATION OF NICKEL NANOPARTICLES FOR CATALYTIC APPLICATIONS
1 µm (b)
(а)
37
100 nm
Fig. 1. SEM micrographs of nanoparticles differing in average size: (a) high nickel content (sample 005), (b) high nickel oxide content (sample 015).
with SIAMS600 (equivalentcircle diameter), we deter mined the average nanoparticle diameter d. The results for a number of nanopowders are presented in the table, together with the main adjustable parameters of the prep aration process: the inertgas flow rate, metallic wire feed rate, and oxygen flow rate. It is well seen that identical parameters of highly oxidized particles can be obtained under slightly different conditions, as distinct from the almost complete uniqueness of conditions for the prepa ration of weakly oxidized particles. The data in the table can, in principle, be used to con struct a simple phenomenological model and derive a relationship between the average nanoparticle size and process parameters, an analogue of that reported by Kondrat’eva et al. [3]. We have, however, an additional, essential factor related to the behavior of oxygen gas, which was delivered to the particle condensation zone in different ways and, accordingly, participated only in the oxidation of the particles or, in addition, influenced the dimensions of the condensation zone. In the latter case, it is well seen that the average particle size as a function of oxygen flow rate has a marked minimum. It is worth pointing out that, in the range of nickel feed rates studied here, the oxygen flow rate needed to oxidize the entire nickel is about one order of magnitude lower than the minimum oxygen flow rate in our experiments. There fore, the increase in particle size with oxygen flow rate is caused by an increase in total gas pressure in the conden sation zone, which should have such an effect according to Gen and Miller [2]. Xray diffraction patterns showed only reflections from metallic nickel and NiO in different ratios, depend ing on the extent of oxidation of the particles. Energy dis persive Xray analysis data also indicated that all of the nanoparticles contained only nickel and oxygen. The best agreement with the Xray diffraction data was observed for almost pure nickel and an almost pure oxide. In the intermediate composition range, the oxy gen content typically exceeded that evaluated from sto ichiometry (by up to several percent). INORGANIC MATERIALS
Vol. 47
No. 1
2011
Magnetic measurements.Magnetic measurements on slightly compacted samples indicated that, when no spe cial measures were taken to isolate nanoparticles from each other, almost all of the nanoparticles showed hyster esis loops, in contrast to what was reported by Petrov et al. [6]. The table gives the main characteristics of the loops: saturation mass magnetization σs, remanent mass mag netization σr, and coercive force Hc. The saturation mag netization of the nanoparticle ensembles is seen to sys tematically decrease only with an increase in the extent of oxidation of the nanoparticles [7], as would be expected because fcc NiO is known to be a paramagnet with a 293K mass susceptibility of 9.6 × 10–9 m3/kg. The rela tive remanent magnetization of the nanoparticle ensem bles has a maximum value just above 0.5, in agreement with earlier data [7]. At the same time, the behavior of their coercive force depends on the particle size and dif fers significantly from that in highly dilute nanoparticle ensembles [6], which suggests that the samples contained chains of interacting ferromagnetic particles. With densities of bulk nickel and nickel oxide ρNi = 8.9 g/cm3 and ρNiO = 6.7 g/cm3, the average core diame ter of a nanoparticle composed of a nickel core and nickel oxide shell can be found as
d Ni = d 3 V , V =
1.27 −1 , ⎛ σσs ⎞ 0.27 + ⎜ ⎟ ⎝ σσ sb ⎠
(1)
where V is the volume fraction of metallic nickel, and σsb is the saturation mass magnetization of pure bulk nickel (55 A m2/kg). The data in the table show that V varies from 98.3 to 0.34%, depending on nanoparticle preparation conditions. In the chain model of Jacobs and Bean [8], the coer cive force of a chain of singledomain ferromagnetic par ticles with coherent rotation of all their magnetization vectors is given by
38
MOROZOV et al.
Main parameters of the nanoparticles σs, σr, Hc, Sample d, nm V, % 2 A m /kg A m2/kg kA/m
S × 10–3, m2/kg
Preparation conditions dNi, nm Ni × 106, Ar × 106, He × 106, O × 106, m3/s E, 2 kg/s kV/m m3/s m3/s
004
180 ± 20 53.8
98.3
6.06
11.5
3.68 ± 0.40 183
2.64
11
–
–
–
010
126 ± 12 40.6
78.2
5.68
15.5
5.66 ± 0.01 116
1.22
39
–
97, downstream
–
005
120 ± 12 51.0
94.2
9.60
16.2
5.73 ± 0.14 117
1.22
39
–
–
–
006
111 ± 10 36.7
71.8
7.02
18.9
6.50 ± 0.02 100
1.22
39
–
3.1, downstream
–
008
108 ± 11 32.0
63.9
4.69
16.6
6.86 ± 0.06
93
1.22
39
–
49, downstream
–
007
105 ± 10 36.1
70.8
5.54
17.2
6.87 ± 0.17
94
1.22
39
–
11, downstream
–
42
–
–
–
–
001
70 ± 7
49.8
92.4
15.9
21.2 10.02 ± 0.40
67
1.67
–
002
51 ± 5
47.2
88.5
19.6
29.3 13.75 ± 0.12
021
33 ± 3
0.49
014
32 ± 3
5.7
023
32 ± 3
0.21
020
31 ± 3
45.9
1.13 11.2 0.48 86.5
49
1.54
–
24
0.17
26.1 26.81 ± 0.10 6.8
0.86
–
167
21, upstream
–
2.68
29.5 26.85 ± 0.12
15
1.22
–
322
33, downstream
–
0.04
15.6 28.31 ± 0.00
5
0.53
–
167
–
–
18.2
23.8 22.55 ± 0.16
29
0.75
–
322
–
–
016
30 ± 3
46.4
87.3
19.7
25.7 22.93 ± 0.13
29
0.86
–
167
–
–
003
29 ± 3
44.3
84.0
20.3
31.5 24.60 ± 0.27
27
1.22
–
278
–
–
013
28 ± 3
46.0
86.6
20.9
29.4 25.10 ± 0.27
26
1.22
–
322
–
–
019
24 ± 2
167
11, upstream
–
028
21 ± 2
0.15
015
15 ± 1
5.0
11.3
011
50 ± 5
18.7
012
46 ± 4
017 018
0.30
0.70
0.06
20.1 36.95 ± 0.11 4.6
0.86
–
0.34
0.02
14.5 43.59 ± 0.78 2.9
0.53
–
167
5.6, upstream
–
1.94
21.8 58.88 ± 0.26 7.1
1.22
–
322
11, downstream
–
39.6
5.14
22.9 15.87 ± 0.01
37
1.54
–
86
33, downstream
–
15.7
33.7
5.99
26.6 17.66 ± 0.07
32
1.54
–
86
33, downstream
250
28 ± 3
10.6
23.3
5.51
30.7 29.54 ± 0.16
17
0.86
–
167
11, downstream
–
22 ± 2
4.4
10.0
2.07
30.1 39.78 ± 0.12
10
0.86
–
167
11, downstream
250
n
H c,n
(n − j) = πρ Niσ s , nj 3 j =1
∑
(2)
where n is the number of particles in the chain. Figure 2 plots the measured coercive force against average nickel core diameter, dNi, for a number of nanopowder samples. The dotted lines indicate the coercive force calculated for different values of n in (2). Analysis of the data in Fig. 2 leads us to conclude that, in all of the slightly compacted nanopowder samples studied, magnetic cohesion forces produce chains of at most four particles. The table presents the BET surface area S of loose nanopowder samples with different particle sizes. Figure 3 plots the average particle diameter (cube edge length) determined in SEM micrographs against the spe cific surface area of the powder. Taking into account the true density of the material under investigation, we find that the data are well represented by the relation
d=
6 × 10 ≈ 0.674 × 10 . ⎛ ⎛ ρ NiO ⎞⎞ S (1 + 0.274V ) S ρ Ni ⎜1 + V ⎜1 − ⎟⎟ ρ Ni ⎠⎠ ⎝ ⎝ 9
5
(3)
Relation (3) leads us to propose a simple and conve nient combined method for determining the average nanoparticle size in ensembles where the interparticle contact area is insignificant. Clearly, it is the large free surface area in nanoparticle ensembles which is of key importance in catalytic applications for heterogeneous chemical reactions involving the gas phase. Thus, know ing the specific surface area of a sample and evaluating V from magnetic measurements, one does not need to per form more complex electronmicroscopic studies for accurately determining the average nanoparticle size in the ensembles. In the bottom part of the table, we present experi mental data that illustrate the influence of a dc electric field E = 250 kV/m applied downstream of the nickel droplet evaporated in flowing helium gas on the parame ters of nanoparticles produced under two distinct sets of INORGANIC MATERIALS
Vol. 47
No. 1
2011
PREPARATION OF NICKEL NANOPARTICLES FOR CATALYTIC APPLICATIONS Нс, kА/m
d, nm 4 3
30
250 200
25 20
150
2
100
15
50
10 20
40 60
80 100 120 140 160 180 200 dNi, nm
Fig. 2. Coercive force of nanoparticles as a function of average nickel core diameter (the values at the dotted lines indicate the number of particles per chain).
conditions. It is well seen that in both cases the electric field markedly reduces the average size and nickel metal content of the particles. The likely reason for this is that the oxidation of metallic particles modifies the diffusion flow in the oxide layer [9]. Since the process involves a liquid phase, the electric field may also influence melt solidification processes [10]. It is reasonable to assume that the electric field reduces the nucleation and growth rates, eventually reducing the size of the resultant nano particles. Catalytic performance. We carried out experiments intended to assess the catalytic activity of the nanoparti cles for the lowtemperature oxidation of a model mix ture of carbon monoxide and hydrocarbons. The cata lytic activity of the synthesized catalysts was studied using an apparatus described elsewhere [11]. The apparatus comprised a cylindrical flow reactor with a bend of the gas flow and a catalyst layer supported on fibrous material, a thermocouple for monitoring the temperature in the reactor, a chromatograph, a gas flow meter, and a cylinder filled with a model gas mixture. The reactor design reproduced the operation of real exhaust catalytic converters, where the catalyst is heated by the hot exhaust gas. The reactor had the form of two coaxial quartz tubes and was heated predominantly by the model gas mixture flow, heated at the reactor wall. The catalyst charge was 1 cm3 of nanopowder, and the relative volu metric gas flow rate was up to 33.3 s–1. The composition of the model gas mixture was similar to the automobile exhaust composition: 0.2 vol % C3H8, 0.5 vol % CO, and 1.8 vol % O2, with nitrogen making up the balance to 100%. The reaction products were identified using a gas analyzer. The experiments were performed in the tem perature range 370 to 870 K at 50K intervals. INORGANIC MATERIALS
Vol. 47
No. 1
2011
0
10
20
30 40 –3 2 S × 10 , m /kg
50
Fig. 3. Measured average nanoparticle size against the specific surface area of the powder.
The results on the catalytic activity of different nano particles for the oxidation of CO and propane are pre sented in Figs. 4 and 5. The data lead us to conclude that the activity of the nanocatalysts increases with increasing specific surface area. CO conversion reaches 80% near 470 K for the best catalyst, consisting of highly oxidized 15nmdiameter particles (Fig. 1b), and near 510 K for the largest parti cles. Propane conversion reaches 80% only at 570 K for the largest particles. The catalytic activity for the latter reaction depends more strongly on the nickel oxide con tent of the particles than on the nanoparticle size (spe cific surface area) in such samples [12]. A complete understanding of the role the size and phase factors play
100 CO conversion, %
0
39
80 1 2 3 4 5
60 40 20 0 350
400
450
500
550 600 Temperature, K
Fig. 4. Temperature effect on the oxidation of carbon monox ide in the presence of catalysts from different nanoparticles (same numbers as in the table): (1) 008, (2) 019, (3) 005, (4) 016, (5) 015.
40
MOROZOV et al.
This research was supported by the Russian Founda tion for Basic Research, grant no. 090800149a.
100
C3H8 conversion, %
80
REFERENCES
60
1 2 3 4 5
40 20 0 400
450
500
550
600
650 700 750 Temperature, K
Fig. 5. Temperature effect on the oxidation of propane in the presence of catalysts from different nanoparticles (same numbers as in the table): (1) 019, (2) 016, (3) 008, (4) 015, (5) 005.
in determining the performance of the nanocatalysts requires additional studies, in particular, in the field of catalytic synthesis. CONCLUSIONS Nickel nanoparticles of controlled size in the range 15 to 200 nm can be produced by a crucibleless aerosol method involving metal vapor condensation in an inert gas flow at different oxygen contents. In this process, the extent of oxidation of the nanoparticles can be varied from 1.5 to 99.7%. A dc electric field applied to the condensation/cool ing zone can be used to control the extent of oxidation and the average size of the nanoparticles. The best catalytic performance for the lowtempera ture catalytic oxidation of carbon monoxide is offered by the smallest and oxidized nanoparticles. At the same time, propane oxidation is better catalyzed by the largest, weakly oxidized nanoparticles. ACKNOWLEDGMENTS We are grateful to N.A. Vakin for performing the cat alytic work.
1. Gusev, A.I., Nanomaterialy, nanostruktury, nanotekh nologii (Nanomaterials, Nanostructures, and Nanotech nologies), Moscow: Fizmatlit, 2007. 2. Gen, M.Ya. and Miller, A.V., A Levitation Method for Producing Ultrafine Metal Powders, Poverkhnost, 1983, no. 2, pp. 150–154. 3. Kondrat’eva, T.A., Morozov, Yu.G., and Chernov, E.A., Effect of Preparation Conditions on the Properties of Ultrafine Nickel Powder, Poroshk. Metall. (Kiev), 1987, no. 10, pp. 19–22. 4. Morozov, Yu.G., Belousova, O.V., and Kuznetsov, M.V., Synthesis of Magnetic Nanooxides by the CrucibleFree Aerosol Method, Yugoslav Materials Science Society Conf. YUCOMAT 2008, Herceg Novi, 2008, p. 11. 5. Grigorevskii, A.V., Mazo, D.I., and Chizhov, P.E., Mag netic Properties and Structure of Ultrafine Nickel Pow ders, IV Vsesoyuznyi simpozium “Svoistva malykh chastits i ostrovkovykh metallicheskikh plenok” (IV AllUnion Symp. Properties of Small Particles and Metallic Island Films, Sumy, 1985), Kiev: Naukova Dumka, 1985, pp. 104–105. 6. Petrov, A.E., Petinov, V.I., and Shevchenko, V.V., Mag netic Properties of Small AerosolDerived Nickel Parti cles between 4.2 and 300 K, Fiz. Tverd. Tela (Leningrad), 1972, vol. 14, no. 10, pp. 3031–3036. 7. Petrov, Yu.I., Fizika malykh chastits (Physics of Small Particles), Moscow: Nauka, 1982. 8. Jacobs, I.S. and Bean, C.P., An Approach to Elongated FineParticle Magnets, Phys. Rev., 1955, vol. 100, no. 4, pp. 1060–1067. 9. Tret’yakov, Yu.D., Khimiya nestekhiometrichnykh okislov (Chemistry of Nonstoichiometric Oxides), Moscow: Mosk. Gos. Univ., 1974. 10. Kozlovskii, M.I., Burchakova, V.I., and Melent’ev, I.I., Elektricheskoe pole i kristallizatsiya (Electric Field and Crystallization), Chisinau: Shtiintsa, 1976. 11. Borshch, V.N., Zhuk, S.Ya., Vakin, N.A., et al., SHS Produced βSialons As Supports for Oxidation Catalysts, Int. J. SelfPropag. HighTemp. Synth., 2009, vol. 18, no. 1, pp. 38–41. 12. Morozov, Yu.G., Belousova, O.V., and Kuznetsov, M.V., Nickel Nanoparticles for Catalytic Applications, Yugoslav Materials Science Society Conf. YUCOMAT 2009, Herceg Novi, 2009, p. 56.
INORGANIC MATERIALS
Vol. 47
No. 1
2011
