Synthesis of nanosized spherical cobalt powder by ultrasonic spray pyrolysis

Article abstract of DOI:10.1016/j.materresbull.2006.03.006

The ultrasonic spray pyrolysis (USP) method has been used to prepare nanosized powders of metallic, intermetallic compounds and ceramic materials. Spherical nanosized cobalt powders were obtained by USP of aqueous solutions of cobalt nitrate followed by thermal decomposition of generated aerosols in hydrogen atmosphere. Particle sizes of the produced cobalt powder can be controlled by the change of the concentration of an initial solution. Non-agglomerated spherical nanosized cobalt particles in the range of 158-1001 nm were obtained at 800 °C. A decrease of the concentration of cobalt nitrate decreases the mean particle diameter from 596 to 480 nm. The discrepancy between the experimentally and theoretically obtained values indicates that the partial coalescence of the droplets occur during the formation of aerosol.

Full text of DOI:10.1016/j.materresbull.2006.03.006

Materials Research Bulletin 41 (2006) 1882–1890
www.elsevier.com/locate/matresbu
Synthesis of nanosized spherical cobalt powder by
ultrasonic spray pyrolysis
a,
Sebahattin G u¨ rmen , Sre c˙ ko Stopi c´ , Bernd Friedrich
b
b
*
a
Istanbul Technical University, Metallurgical and Materials Engineering Department, Istanbul, Turkey
b
IME Process Metallurgy and Metal Recycling, RWTH Aachen University, Germany
Received 15 December 2005; received in revised form 8 March 2006; accepted 12 March 2006
Available online 31 March 2006
Abstract
The ultrasonic spray pyrolysis (USP) method has been used to prepare nanosized powders of metallic, intermetallic compounds
and ceramic materials. Spherical nanosized cobalt powders were obtained by USP of aqueous solutions of cobalt nitrate followed by
thermal decomposition of generated aerosols in hydrogen atmosphere. Particle sizes of the produced cobalt powder can be
controlled by the change of the concentration of an initial solution. Non-agglomerated spherical nanosized cobalt particles in the
range of 158–1001 nm were obtained at 800 8C. A decrease of the concentration of cobalt nitrate decreases the mean particle
diameter from 596 to 480 nm. The discrepancy between the experimentally and theoretically obtained values indicates that the
partial coalescence of the droplets occur during the formation of aerosol.
#
2006 Elsevier Ltd. All rights reserved.
Keywords: A. Nanostructures; A. Metals
1
. Introduction
Metal particles smaller than 100 nm in primary particle diameter are generally considered as nanoparticles.
Nanoscale particle research has recently become a very important field in materials science. Such metal nanoparticles
often exhibit very interesting electronic, magnetic, optical, and chemical properties. The reactivity of nanoparticles
depends on its size, shape, surface composition, and surface atomic arrangement. Their high surface-to-volume ratios
have large fractions of metal atoms at surface available for catalysis. In the case of cobalt nanoparticles, they are
expected to possess exceptionally improved catalytic properties [1–4]. Spherical non-agglomerated submicron
particles of complex composition and controlled phase content, suitable for direct application or fabrication of high
technology sintered materials, can be prepared by the spray pyrolysis method [5]. The ultrasonic spray pyrolysis (USP)
is a versatile technique for producing various materials in a wide range of composition, size and morphology. Suh and
Suslick [6] applied the USP technique to synthesize silica-encapsulated metal nanoparticles. USP of solutions is
performed by applying a powerful source of ultrasound to the corresponding solution forming an aerosol with constant
droplet size, which depends on the characteristic of the liquid and the frequency of the ultrasound. At a resonance
frequency of the order of 2 MHz, the droplet size is approximately 2 mm [7–9]. Stopi c´ et al. [7,8] prepared the nickel
*
Corresponding author. Tel.: +90 212 2856862; fax: +90 212 2853427.
E-mail address: gurmen@itu.edu.tr (S. G u¨ rmen).
0025-5408/$ – see front matter # 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2006.03.006

S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
1883
powder by ultrasonic spray pyrolysis method. The authors state that powders had a spherical morphology and were
substantially crystalline. Ultrasonic spray pyrolysis enables the subsequent reactions to occur in a very small volume
leading to an ultra fine powder. The different form of produced particles is present due to the process conditions
leading to either volume or surface precipitation within the droplet and the subsequent conversion to a dense, solid
particle or one hollow, shell-like particle. From the viewpoint of the application of the as-generated particles for
advanced materials synthesis, particle morphology is of great interest. It is presumed that certain particle morphology
is formed during the evaporation/drying stage that encountered processes of evaporation and diffusion of both the
solvent and solute, changing in droplet temperature and crust formation [8]. Messing et al. [10] mentioned that
developments in process control, atomisation, and system design are required for wider commercialisation of spray
pyrolysis process. USP is a useful tool for large-scale or small-scale production of particles with controlled particle
size. The fabrication of cobalt nanosized powder in this way has not yet been investigated. And only a few studies for
the synthesis of cobalt particles by gas phase chemical reaction have been reported in the literature [11]. Cobalt
nanoparticles were produced by the hydrogen reduction of cobalt chloride in a multistage tubular aerosol flow reactor
by Jang et al. [11]. Reaction zone temperature, preheating temperature, mole fraction of CoCl and H , and residence
2
2
time were considered as key process variables for the control of particle size and size distribution. Ranging from 50 to
8 nm in average diameter, cobalt nanoparticles with narrow size distribution were synthesized throughout the above-
mentioned experiments.
7
A novel route to prepare spherical nanopowders of cobalt using ultrasonic spray pyrolysis (USP) will be
investigated. A controlled particle size will be realized through the choice of precursor obtained from cemented
tungsten carbide scrap and solution concentration as well as by changing the aerosol decomposition parameters.
Because of a higher surface to volume ratio and an enhanced reactivity the prepared nanopowder should be ideally
effective in catalysis in order to accelerate chemical reactions. This hydrometallurgical method will utilize hydrogen
reduction of a produced aerosol under dynamic conditions.
2
. Experimental
2
.1. Materials
A purified leach solution from co-extraction experiments from cemented tungsten carbide scrap using nitric acid
was used as starting material for this research (for details, see Refs. [12,13]). The chemical composition of the
cemented tungsten carbide scrap used in our experiments is mainly as follows (wt.%): 75.86 W, 8.14 Co, and 6.07 C.
The concentration of the obtained final Co solution is between 0.08 and 0.04 mol/l.
2
.2. Experimental procedure
The nanostructured materials were synthesized using the ultrasonic spray pyrolysis method. Fig. 1 shows the
schematics of the apparatus. The experiments were done at 700–900 8C starting from cobalt nitrate solutions with
concentrations of 0.08 mol Co/l and 0.04 mol/l. Very fine droplets of the aerosol were obtained in an ultrasonic
atomizer Pyrosol 7901 (Ramine Baghai Instrumentation, with a frequency of 800 kHz). The liquid feed rate amounted
between 12 and 14 ml per hour. The aerosol was transported by H -carrier/reduction gas via a quartz tube (0.7 m
2
length and 0.02 m diameter) to an electrical heated furnace (Str o¨ hlein, Germany) with a temperature control of ꢀ1 8C.
Because of the safety reasons and to create an inert atmosphere, nitrogen with a flow rate of 1 l/min was used prior to
the reduction process. Under spray pyrolysis conditions in hydrogen atmosphere and at a flow rate of 1 l/min, the
dynamic (continuous) reduction took place in the quartz tube reactor (heated zone 280 mm). The residence time,
calculated from the ratio of the volume of the reaction zone and the carried gas flow, was about 1 s with the assumption
that the rate of droplets and the carried gas are equal. According to model one particle from one droplet [9] we
supposed that this short reaction time is sufficient for the transformation of droplets to metal particle in the hydrogen
atmosphere. An X-ray diffractometer (Siemens D 5000) and a scanning electron microscope (Zeiss DSM 982 Gemini)
were used for the characterization of the obtained cobalt powders. SEM images were used to observe the surface
morphology of particles formed at different reaction temperatures. The powders were first dispersed in ethanol and
inserted in ultrasonic bath for 30 min and then the suspension was dispersed dropwise onto a glassy carbon slide to
make a thick film, which later coated by palladium for SEM characterization. The particle size was examined using the

1884
S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
Fig. 1. Schematic drawing of experimental apparatus for the synthesis of cobalt nanoparticles from Co-nitrate-solutions.
Table 1
Composition of precursor solutions, conditions of the process, and descriptions of the obtained powders (test period = 2 h, volume of used
solution = 260 ml)
No.
Concentration of precursor
solution (mol/l)
Temperature (8C)
Flow rate
Characteristic of the product
2
of H (l/min)
1
2
3
4
Co(NO
Co(NO
Co(NO
Co(NO
3
)
3
)
3
)
3
)
2
2
2
2
, 0.08 M
, 0.08 M
, 0.08 M
, 0.04 M
700
800
900
800
1
1
1
1
Spherical, nanostructured cobalt powder
Spherical, nanostructured cobalt powder
Spherical, nanostructured cobalt powder
Spherical, nanostructured cobalt powder
areal analysis method with a semi-automatic image analyzer (Videoplan, Kontron). ICP (Spectroflame Modula EOP)
analysis was used for quantitative chemical composition.
Table 1 contains data for chemical composition of the applied solutions, conditions of the production process, and a
short description of the obtained powders.
3
. Results and discussion
3
.1. X-ray analysis of cobalt powder
Fig. 2 shows X-ray diffraction (XRD) patterns of the cobalt powders. The X-ray analysis of the powders, produced
at 700, 800 and 900 8C from Co(NO ) solution in H atmosphere by ultrasonic spray pyrolysis, indicated the presence
3
2
2
of pure cobalt powders.
The final product (Co particles) contains no impurities such as tungsten.
.2. Thermodynamic analysis of hydrogen reduction
3
The reaction for the formation of Co metal from cobalt nitrate can be described as in Eq. (1). The thermodynamic
analysis is done using Fact Sage Software and in the temperature range of 25–1000 8C (Fig. 3).
CoðNO3Þ2 þ 2H2 ! Co þ 2NO2 þ 2H2O
The values of Gibb’s free energy (DG ) for the reaction 1 in the temperature range up to 1000 8C confirm the
(1)
0
probability for formation of Co from Co(NO ) by hydrogen reduction. This probability for formation of Co increases
at elevated temperatures.
3
2

S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
1885
Fig. 2. X-ray analysis of the USP cobalt powders (700, 800 and 900 8C).
3
.3. The effect of precursor solution concentration
The effects of the precursor drop size on size distribution of the particles and their morphology after
spray pyrolysis have been reported by Tsai et al. [14]. This study examines the effect of the precursor concentration in
the range of 0.04–0.08 M under the conditions of 2 h run time, 800 8C reduction temperature, and 1 l/min
H2 volumetric flow rate. SEM micrographs of the obtained cobalt powders are shown in Fig. 4a and b,
respectively.
In both cases the cobalt powders are almost spherical with comparable particle morphology. Under the
same conditions the reduced concentration of cobalt in the precursor solution decreases the particle size and the
uniformity of the powder. Some agglomeration of particles is particularly present in both cases the surface
morphology seems to be smoother at low Co concentration. The increase in the number of larger particles with
cobalt concentration appears to indicate that they are produced by the coagulation, surface reaction, and sintering.
However, smaller particles that seemed to be grown by the coagulation and surface reaction are observed at low
cobalt concentrations. In difference to non-agglomerated particles by ultrasonic spray pyrolysis nanosized
primary particles prepared in the gas phase [14] form aggregates because of interactive forces between particles.
Such aggregates change into large particles when the temperature in the gas phase is high enough to sinter the
aggregates.
Fig. 3. Dependence of the Gibb’s energy on the temperature.

1886
S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
Fig. 4. (a) SEM micrograph of the cobalt powder obtained at 800 8C with cobalt concentration of 0.08 M Co/l. (b) SEM micrograph of the cobalt
powder obtained at 800 8C with cobalt concentration 0.04 M Co/l.
3
.4. Comparative analysis of powders obtained under static and dynamic conditions
In a previous investigation cobalt powders were produced from a similar solution by chemical precipitation, and the
calcination and reduction were performed under static conditions [13]. The cobalt hydroxide precipitates were placed
in a quartz tube without any displacement for 2 h. Micron size particles started to agglomerate at increased
temperatures of 800 8C. Also particles with irregular shapes have been produced (Fig. 5).
Compared to the static conditions the droplets in this paper were reduced in the reaction furnace by hydrogen in a
very short time, leaving the quartz tube after approximately 1 s.
The difference between the results in dynamic and static conditions using the same furnace, the same temperature,
and the initial concentration of solution is a consequence of different retention times and fluid dynamics in the reaction
zone. Under static conditions the retention time of 2 h is long enough for an agglomeration and growth of particles.
Due to these reasons an ultrasonic atomizer produces spherical droplets of Co(NO ) . It can be presumed that in the
3
2
first stage water is removed from the surface of the droplets and that the solution is being concentrated. Such the
concentration of the solution at the droplet’s surface is much higher than in the inner part of the droplets. When the
solution reaches its saturation point, compounds crystallize, decompose thermally, and form a product shell. This is the
reason why ideally spherical non-agglomerated powders can be produced using the ultrasonic spray pyrolysis.

S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
1887
Fig. 5. SEM micrograph of a cobalt powder obtained under static conditions at 800 8C [13].
3
.5. Comparative analysis of theoretically expected particle diameter and the experimentally obtained values
The relation between the droplet diameter of the atomized solution and the frequency of the ultrasound source was
studied by Tsai et al. [14]. It was shown that the mean diameter of the aerosol droplets develops according to the
relationship:
2
1=3
D ¼ 0:34ð8pg=r f Þ
where D is the mean droplet diameter, g is the surface tension of the solution, r is the density of the solution and f is the
(2)
frequency.
Using the parameters of this study (g, 72.9 ꢁ 10ꢂ3 N m ; r, 1 g cm ; f, 800 kHz) the calculated droplet diameter
ꢂ1
ꢂ3
amounts to D = 4.26 mm.
The dependence of the mean droplet diameter by the frequency of the used ultrasonic atomizer is shown in Fig. 6. It
can be seen that the mean droplet diameter decreases with increasing frequency of the ultrasonic atomizer.
Using this rule for the droplet size the expected value of the mean particle diameter of the cobalt powder can be
calculated depending on the initial concentration of the Co(NO ) solution. It is assumed that each droplet is
3
2
transformed into a particle and that during the atomisation no coalescence occurs. If D is the particle diameter; D is
p
the droplet diameter; Cprecursor is the concentration of the water of Co(NO ) -solution and r is the density of cobalt,
3 2
Co
the following equation can be developed:
1
=3
D ¼ DðC
M =M
r Þ
precursor Co
(3)
p
precursor Co
Fig. 6. Dependence of the mean droplet diameter by the frequency of ultrasonic atomizer.

1888
S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
Fig. 7. Dependence of particle diameter by the concentration of cobalt in solution.
The calculated particle diameter of Co (D ) varies from 352 to 274 nm, by changing the precursor concentration
p
from 0.08 to 0.04 mol/l, while keeping the following data constant: D, 4.26 mm; M , 58.93 g/mol; Mprecursor,
Co
3
82.93 g/mol; r , 7.84 g/cm . The fact is clearly seen in Fig. 7.
1
Co
Experimental values of the minimal, maximal and mean particle diameters and theoretical calculated values of the
particle diameter of cobalt are given in Table 2.
The experimental results confirmed that the decrease of the concentration of cobalt nitrate decreases the particle
size. In both cases the calculated values of particle diameter of cobalt are between the minimal and mean experimental
obtained values (157.97 nm < 352 nm (calculated value for 0.08 M of cobalt nitrate) < 595.88 nm and
1
43.94 nm < 274 nm (calculated value for 0.04 M of cobalt nitrate) < 479.83 nm, respectively). The maximal
experimental values of particle diameters of cobalt of 766.27 and 1001.00 nm are consequence of the coalescence of
droplets during the transport by hydrogen. We have supposed that different sized droplets produced by an ultrasonic
generator have different rising velocities. Larger droplets can catch smaller ones along their way into the furnace. Also
the droplet collisions could be caused by the turbulent velocity of hydrogen.
This calculation is done by the conventionally accepted one particle drop mechanism [14]. The principle of the
proposed mechanism is based on a single-step high temperature aerosol decomposition (one droplet/one particle),
ultrasonically generated from precursor solutions of cobalt nitrate. A careful comparison of precursor drop sizes with the
product particle sizes reveals that in addition to the conventionally accepted one particle per drop mechanism, spray
pyrolysis may also involve the gas to particle conversion mechanism, which creates nanoparticles much smaller than
those predicted by the one particle per drop mechanism alone according to Tsai et al. [14]. Thus, uniform nanoparticles
can be produced by spray pyrolysis at proper conditions using uniform drops of precursor salts with low vapour pressure.
Fig. 8 presents scaled SEM picture of particles received from 700 8C, 0.08 mol/l cobalt nitrate solution USP tests. The
particles between 50.2 and 73 nm seen in Fig. 8 constitute one big spherical particle of cobalt powder.
In contrast to ideally spherical particles by ultrasonic spray pyrolysis method Jang et al. [11] prepared the spherical
cobalt nanoparticles in a range between 55 and 78 nm by hydrogen reduction of cobalt chloride in the gas phase at
9
particles was observed. The production rate of Co nanoparticles was 1 g/h at 0.0862% of CoCl , that was the same in
00 8C. Primary particles looked like single spherical crystals of nearly uniform size and the directional linkage of
2
ultrasonic pyrolysis method.
Table 2
Experimental and theoretical values of diameter of cobalt particles
Concentration (mol/l)
0
.08
0.04
Maximum
Minimum
157.97
Mean
Maximum
Minimum
143.94
Mean
Experimental values (nm)
Theoretical values (nm)
1001.00
352.0
595.88
766.27
274.0
479.83

S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
1889
Fig. 8. SEM micrograph of USP cobalt powder obtained at 700 8C with a precursor cobalt concentration of 0.08 mol/l (magnification 50,000ꢁ).
The results, presented here, are very promising and further detailed investigations including the physical properties
have been started.
4
. Conclusion
Ultrasonic spray pyrolysis (USP) is used for the synthesis of fine, spherical, nanosized cobalt particles from a
dispersed cobalt nitrate solution formed during the recycling of waste cemented carbide. In the temperature range
between 700 and 900 8C the aerosol droplets of 4 mm were decomposed by hydrogen reduction in order to produce
cobalt powder. The experimental results confirmed that the decrease of the concentration of cobalt nitrate decreases the
particle size. In both cases the calculated values of particle diameter of cobalt ranged between the minimal and mean
experimental obtained values (157.97 nm < 352 nm (calculated value for 0.08 M of cobalt nitrate) < 595.88 nm and
1
43.94 nm < 274 nm (calculated value for 0.04 M of cobalt nitrate) < 479.83 nm, respectively). The maximal
experimental values of particle diameters of cobalt of 766.27 and 1001.00 nm are consequence of the coalescence of
droplets during the transport by hydrogen into the furnace. One comparison between the powder morphology of cobalt
obtained in static and dynamic conditions emphasizes that the ultrasonic spray pyrolysis method is a good choice for
the production of the particles with controlled morphology.
Acknowledgment
We thank the Alexander von Humboldt Foundation for providing research fellowships to Dr. Sebahattin G u¨ rmen
allowing an 18-month stay at the IME Process Metallurgy and Metal Recycling, RWTH Aachen University, where this
work was performed in context of a process development to recycle hard metal scrap.
References
[
[
[
1] B. Zhou, S. Hermans, S. Somorjai, Nanotechnology in Catalysis, Kluwer Academic/Plenum Publishers, New York, 2004, pp. 361–399.
2] R.W. Siegel, Nanophase Materials: Synthesis, Structure, and Properties, Springer Verlag, 1994.
3] V. Jokanovic, in: A.M. Spasic, J.P. Hsu (Eds.), Nanodesigning Surfactant Science Series 130, CRC Press, Taylor & Francis Group, Boca Raton,
2006, pp. 513–533.
[
[
[
[
[
4] S. Stopi c´ , P. Dvorak, B. Friedrich, World of Metallurgy—Erzmetall 58 (2005) 195–201.
5] T. Kodas, United States Patent 6,316,100 (2001).
6] W.H. Suh, K.S. Suslick, J. Am. Chem. Soc. 127 (34) (2005) 12007–12010.
7] S. Stopi c´ , I. Ili c´ , D. Uskokovi c´ , Mater. Lett. 24 (1995) 369–376.
8] S. Stopi c´ , I. Ili c´ , D. Uskokovi c´ , Int. J. Powder Metall. 32 (1) (1996) 59–65.

1890
S. G u¨ rmen et al. / Materials Research Bulletin 41 (2006) 1882–1890
[9] J.H. Kim, V.I. Babushok, T.A. Germer, G.W. Mulholland, S.H. Ehrman, J. Mater. Res. 18 (2003) 1614–1622.
[
[
[
[
[
10] G.L. Messing, S.C. Zhang, G.V. Jayanthy, J. Am. Ceram. Soc. 76 (1993) 2707–2726.
11] H.D. Jang, D.W. Hwang, D.P. Kim, H.C. Kim, B.Y. Lee, I.B. Jeong, Mater. Res. Bull. 39 (2004) 63–70.
12] S. G u¨ rmen, B. Friedrich, World of Metallurgy—Erzmetall 57 (3) (2004) 143–147.
13] S. G u¨ rmen, S. Stopi c´ , B. Friedrich, World of Metallurgy—Erzmetall 57 (6) (2004) 321–326.
14] S.C. Tsai, Y.L. Song, C.S. Tsai, C.C. Yang, W.Y. Chiu, H.M. Lin, J. Mater. Sci. 39 (2004) 3647–3657.