Application of aerosol techniques to study the catalytic formation of methane on gasborne nickel nanoparticles

Article abstract of DOI:10.1021/jp0115594

A well known reaction, the so-called methanation reaction over a Ni catalyst, i.e., the formation of methane from CO and hydrogen, was studied to demonstrate the possibilities of the aerosol technique. Reaction order and activation energy conformed to generally accepted values from supported Ni catalysts. The turnover rate (TOR) decreased strongly during the first 10 sec as the reaction proceeded toward a steady value. The decrease correlated with a buildup of about 0.3 monolayer equivalents of carbon on the particle surface measured by TGA and a decline in particle photoelectric activity found via measurement by aerosol photoemission spectroscopy. Order-of-magnitude changes were induced in TOR via defined changes in particle morphology induced by aerosol restructuring techniques preceeding exposure to the catalytic reaction. Aerosol catalysis has potential to develop new catalysts and could be an avenue for studying the putative relationship between combustion aerosols and the formation of dioxin.

Full text of DOI:10.1021/jp0115594

8958
J. Phys. Chem. A 2001, 105, 8958-8963
Application of Aerosol Techniques to Study the Catalytic Formation of Methane on
Gasborne Nickel Nanoparticles
Alfred P. Weber,* Martin Seipenbusch, and Gerhard Kasper
Institut fu¨r Mechanische Verfahrenstechnik und Mechanik, UniVersita¨t Karlsruhe (TH),
D-76128 Karlsruhe, Germany
ReceiVed: April 25, 2001; In Final Form: July 13, 2001
“Aerosol catalysis” is shown to be a powerful tool for investigating the catalytic properties of freshly formed
nanoparticles in situ and without substrate interference. The method is first outlined conceptually, followed
by an illustrative application to the catalytic formation of methane on a nickel nanoaerosol. Reaction order
and activation energy were found conform with generally accepted values from supported Ni catalysts. The
TOR decreases strongly during the first 10 s as the reaction proceeeds toward a steady value. The decrease
correlates with a buildup of about 0.3 monolayer equivalents of carbon on the particle surface measured by
TGA and a decline in particle photoelectric activity observed via measurement by aerosol photoemission
spectroscopy (APES). APES is shown to be capable of detecting the progressive degradation of the freshly
formed particle surface due to a heterogeneous surface reaction on a millisecond time scale. Furthermore, it
was possible to induce order-of-magnitude changes in TOR via defined changes in particle morphology,
induced by aerosol restructuring techniques preceding exposure to the catalytic reaction.
1. Introduction
high purity, and over a wide range of pressures. Aerosol
conferences, workshops, and the pertinent archival literature give
ample testimony to that effect. Second, one can proceed without
removing these freshly generated particles from the gas stream
to initiate a specific chemical reaction by mixing them in a
second flow reactor with the required educt molecules and then
continue to study its progress downstream. Since the reactor
axis is in effect the reaction coordinate, one has a high degree
of control over the size, morphology, and surface properties of
the particulate phase. This represents some compensation for
the fact that aerosols cannot be stabilized or manipulated as
elegantly as colloids by means of electrical bilayers. More
importantly, however, aerosol science has at its disposal a range
of on-line tools for characterizing the particles, which offer
unique ways to correlate particle properties with chemical
kinetics and under actual process conditions, i.e., also at
atmospheric or higher pressures. “Aerosol catalysis” may
therefore eventually contribute to closing the so-called pressure
gap between UHV methods of investigation and the real process
and become another tool for catalyst development.
In contemporary parlance, the “active phase” of a catalyst
consists of nanoparticles. To preserve their high degree of
dispersion and also to provide a suitable form of packaging and
exposure in a chemical reactor, these nanoparticles are usually
supported on much larger granules of an “inert” material such
as Al2O3 or MgO. Experiments under real reaction conditions
are usually done with such materials. However, supported
catalysts are not ideal for investigating the fundamental proper-
ties of nanoparticles or materials because the various multistep
methods for their preparation usually result in a loss of control
over the size distribution and morphology of the nanophase and
it also becomes difficult to separate the influence of the support
material and other extraneous factors. Research has thus resorted
to a variety of sophisticated tools for probing and modifying
pure catalyst surfaces, many of which require ultrahigh vacuum
to understand how molecules interact with catalysts.
Rapid advances during the past decade for preparing and
handling “clusters” or “nanoparticles” have also opened up new
avenues to study catalysts. Today, there we have an almost
unlimited array of possibilities for making very well-defined
particles from a few nanometers in diameter upward, especially
by liquid-phase processes. This has also spurred fundamental
investigations of unsupported colloidal catalysts. Schmid et al.,1
for example, demonstrated how variations in size of ligand-
stabilized metal clusters correspond to systematic changes in
catalytic properties.
Despite its potential, the aerosol route with all its possibilities
has so far not been used broadly, and pertinent literature is
scarce. Glikin2,3 has made extraordinary claims with regard to
the effects of the aerosol state on the catalytic activity for the
total oxidation of acetic acid over an iron oxide catalyst.
However, it is doubtful whether the aerosols they used were
sufficiently characterized and quantitatively understood. Nev-
ertheless, these claims have contributed to sparking our own
interest.4
The aerosol routesoften called gas-phase synthesis or chemi-
cal vapor synthesis in order to stress analogies with surface
coating by CVDsprovides another conceptually interesting
approach to study unsupported as well as supported catalysts.
For one, various types of techniques are available for on-demand
production of well-defined nanoparticles in flow reactors, with
The aim of this article is to demonstrate the possibilities of
the aerosol technique by investigating a well know reaction,
namely, the so-called methanation reaction over a Ni catalyst,
i.e., the formation of methane from carbon monoxide and
hydrogen.
* Corresponding author. E-mail: alfred.weber@ciw.uni-karlsruhe.de.
FAX: 0049 721 608 6563.
10.1021/jp0115594 CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/06/2001

Methane on Gasborne Nickel Nanoparticles
J. Phys. Chem. A, Vol. 105, No. 39, 2001 8959
3.1. Aerosol Particle Generation. Nickel nanoparticles are
generated with a spark discharge generator (SDG).6 In this
device, two Ni electrodes are placed in an inert gas stream
(99.99% N2). A capacitance of 20 nF parallel to the electrodes
is periodically charged by a constant current until the break-
through voltage is reached. The ensuing spark vaporizes small
but fairly constant amounts of metal, which then form Ni
nanoparticles of about 4 nm diameter by homogeneous nucle-
ation. These so-called primary particles undergo collisions
resulting in chainlike agglomerates in the size range of tens to
a few hundred nanometers (Figure 2a). The agglomerate number
concentration is controlled by the discharge frequency and was
in the range of 5 × 106 cm-3. On-line measurements with a
scanning mobility particle sizer (SMPS) give continuous data
on agglomerate size distribution and number concentration.
3.2. Structure Modification of the Agglomerates. As
illustrated in Figure 2a, the Ni agglomerates have a high
porosity, i.e., a large percentage of exposed surface area. One
can rearrange the primary particles to form more compact
agglomerates with only minor loss of overall BET-equivalent
surface area (Figure 3). This “restructuring” is induced thermally
by passing the aerosol through a heated tube furnace. Once
mobile enough to move, the primary particles rearrange into a
lower energy state, i.e., in a more compact agglomerate
structure,7 as illustrated in Figure 2b. Above about 450 °C, one
observes the onset of sintering accompanied by a rapid decrease
of BET surface area, followed by melting of the nickel, which
ultimately leaves solid spherical particles (see Figure 2c,d). At
the exit of the furnace, the aerosol is at room temperature.
The exposed surface area per average particle shown in Figure
3 is determined from its mobility equivalent size measured by
electrical mobility analysis (see below), which is known to be
a good representation of projected area.8,9
Figure 1. Schematic diagram of the experimental setup
2. Model Reaction CO + 3H2 f CH4 + H2O
The formation of methane from carbon monoxide and
hydrogen over a Ni catalyst, the so-called methanation reaction
was chosen as a simple and very well characterized system for
which the literature offers ample data on classical supported
catalysts. In addition, the rate of CH4 formation can be translated
directly into an overall turnover rate (TOR), defined as
number of product molecules
number of active sites‚time
TOR )
(1)
because methane is formed almost exclusively (>98%). The
TOR is related to the reaction rate r, i.e., the volume of methane
produced per mass of catalyst and time, in mL mg-1 s-1
rNA
TOR )
(2)
VMSBET
δ
3.3. Catalysis. After conditioning, the educt gases CO and
H2 are added to the aerosol immediately before entering the
flow reactor, which is heated, thereby initiating the catalytic
reaction. The point at which a constant maximum temperature
is reached is defined as t ) 0 for the reaction. Changing gas
composition and particle properties are monitored by a succes-
sion of sampling points which correspond to residence times
between a few seconds and a few minutes. At that point, the
reaction and any further growth processes within the aerosol
are quenched by dilution with a known flow of nitrogen, and
the aerosol sample is split for particle characterization and for
gas analysis. The aerosol characterization will be discussed
below. For the gas analysis, the particles are first removed by
filtration, and the gas composition is then determined on-line
with an FTIR spectrometer (Bruker Model Vector 22) equipped
with an 8.8 m cell. Stable concentrations in the optical cell were
obtained about 5 min after a change in input conditions,
corresponding to three exchanges of the cell volume.
where NA is the Avogadro number, VM the molar volume, SBET
the specific surface area of the catalyst, and δ the number of
active sites per unit surface area.
The reaction order for hydrogen is known to be close to unity,
while the reaction order for carbon monoxide is reported to vary
between -1 and 0.5, depending on CO partial pressure.5
However, by keeping the amount of produced methane low with
low educt concentrations and short reaction times, the initial
concentrations of CO and H2 remain virtually constant through-
out the reactor. Under these conditions, the reaction orders of
the individual educt gases need not be considered for the
determination of the TOR. They were determined nevertheless
in separate experiments by varying the educt initial concentra-
tions to establish a comparison with the classical route.
Moreover, the overall activation energy for the formation of
methane was determined from an Arrhenius-type plot.
3. Experimental Setup
4. Particle Characterization Techniques
As outlined above, the basic idea is to start with gasborne
nanoparticles produced freshly in a continuous process, thereby
guaranteeing constant and reproducible initial conditions. Im-
mediately after generation, the aerosol may undergo modifica-
tion steps to adjust particle structure and certain surface
conditions (for which an operational definition shall be given
later on) to have a broader range of variability than afforded
by the generator alone. These particle properties are all measured
and monitored as far as possible with on-line techniques, as
described in Section 4. In a third reactor, the actual catalytic
reaction is initiated and studied. A schematic diagram of the
experimental setup is shown in Figure 1.
The Ni particles were characterized by on-line and off-line
techniques. The latter include total surface area determination
by standard nitrogen adsorption (BET, Quantachrom Nova
2000), transmission electron microscopy (TEM, Zeiss Leo EM
109), and thermogravimetric analysis (TGA, Netsch TG 209)
coupled with an FTIR spectrometer to analyze adsorbates on
the particle surface qualitatively and quantitatively.
Aerosol particle size distributions and concentrations were
principally determined on-line by a scanning mobility particle
sizer (SMPS, combining TSI Models 3071 and 3022). This
technique is based on fractionating particles according to their

8960 J. Phys. Chem. A, Vol. 105, No. 39, 2001
Weber et al.
Figure 2. (a) Agglomerate of Ni nanoparticles at room temperature. (b) Collapsing of the agglomerate after heating to 450 °C. (c) Partial sintering
of the agglomerate after heating to 600 °C. (d) High degree of sintering after heating to 900 °C, but complete sintering of the agglomerate to a
spherical particle is not achieved yet
mobility in an electric field and is amply described in the
literature.10,11
the generation and conditioning process or, subsequently, during
participation in the catalytic reaction.
The photoelectric yield Y is the probability, per incident
photon of energy hν above the photothreshold Φ, of emitting
an electron from a particle. In the vicinity of the photothreshold,
Y is given by the Fowler-Nordheim equation function14
4.1. Aerosol Photoemission Spectroscopy (APES). Since
its first introduction,12 APES has become a valuable tool for
probing the surface of particles in their gas-borne state. APES
is based on photon-induced electron emission from the particle
surface, which leads to a stable charge level on the aerosol that
is determined on-line via simultaneous measurements of total
electrical charge and total particle number concentration. For
small particles (<300 nm), the escape probability of the
photoelectron is close to unity,13 even at atmospheric pressure.
Aside from its dependence on particle surface area (and hence
size), the electron emission yield, expressed as average number
of unit charges per particle, is a highly sensitive function of
both surface composition and photon energy. APES is therefore
an ideal tool to track changes in particle surface, either during
Y ) c(hν - Φ)m
(3)
where m ) 2 holds for metallic particles or graphite. In practice,
c and Φ are determined by plotting xY over a range of energies
hν. The emission constant c is proportional to the local density
of states at the Fermi level; Φ is an effective work function.
Changes in surface chemical composition by reaction or
adsorption may affect the work function Φ and/or the local
density of states and hence c. Because of the nonideal nature

Methane on Gasborne Nickel Nanoparticles
J. Phys. Chem. A, Vol. 105, No. 39, 2001 8961
Figure 4. Temporal evolution of the photoelectric and catalytic activity
of unconditioned Ni nanoparticles (reaction temperature of 450 °C).
Figure 3. Exposed surface area (based on mobility equivalent diameter)
and total surface area (based on BET) of agglomerates of Ni
nanoparticles as a function of temperature. The numbers along the curve
give the primary particle size.
reaction depend directly on the LDOS (EF).20 This correlation
is shown in Figure 4, which will be discussed below.
5. Results and Discussion
of an aerosol system, there is no established theoretical
correlation, but a great deal of empirical work has been done,
especially related to combustion15 and adsorption,16 which has
proven APES to be a very sensitive technique to track and in
some cases quantify changes occurring on the surface. However,
APES has, to our knowledge, not been correlated with catalytic
activity.
In practice, Φ and c are determined from the electron
emission rate per particle for a given irradiation spectrum (also
called photoelectric activity):
5.1. Reaction Order and Activation Energy. The order of
a reaction Ri with regard to constituent i is obtained by plotting
the reaction rate r versus partial pressure of this constituent.
For RCO, the results obtained in our reactor are shown below in
Table 1 for Ni aerosol coming directly from the spark discharge
generator without any restructuring. The values range from
-0.52 to -0.15 and evidently fall within the range published
for supported Ni catalysts.
Similarly, the activation energy agreed well with literature
values for temperatures of up to 350 °C, as shown in Table 2
below.4
I∆t
N
) YFDSp
(4)
TABLE 1: Reaction Orders for CO on Ni Aerosol Catalysts
partial pressure of CO [Pa]
reaction order of CO [-]
where FD is the photon flux density and Sp the particle surface
relevant for photoemission. Initially, uncharged gas-borne
nanoparticles are irradiated for a fixed period of time ∆t (well
below any saturation effects) with monochromatic photons from
a broad-band UV source coupled with a monochromator. The
photon flux incident on the particles is measured with a UV-
sensitive photomultiplier. The irradiation time ∆t is given by
the residence time in the photo chamber. The resulting net
average charge per particle, a direct measure for the yield, is
determined from simultaneous measurements of the aerosol
current I with a Faraday cup electrometer (FCE) and the total
number concentration N with a condensation particle counter
(CPC).17
5250
7210
11 010
-0.15
-0.48
-0.52
TABLE 2: Activation Energies Measured with Ni Aerosol
Compared to Literature for Supported Catalysts
catalyst type activation energy [kJ/mol]
this work4 (T e 350 °C) aerosol
96
103
107
109
Vannice21
Vannice21
Goodman22
supported
supported
supported
Above 350 °C, the activation energy and the reaction rate
becomes progressively smaller than those extrapolated from the
lower temperature behavior due to the accelerated formation
rate of adsorbed carbon atoms on the catalyst surface. If the
hydrogenation rate cannot keep up with the dissociation of CO
at higher temperatures, carbon builds up on the particle surface,
which may transform into graphite. Goodman22 found that the
formation of relatively inactive graphite on a single-crystal
nickel catalyst requires both a critical temperature of 375 °C
and a prerequisite surface carbide level. As will be shown below,
a buildup of carbon on the surface of our Ni aerosol particles
was indeed observed. In summary, these findings indicate that
the key features of the methanation reaction (i.e., overall
activation energy, reaction order and poisoning effects) are
similar for supported and gas-borne catalyst particles.
Equation 4 further requires a value for the surface area per
18
particle, Sp. It has been shown that the mobility equivalent
diameter of an agglomerate particle as determined by SMPS is
the appropriate measure for the exposed surface area and hence
also for the surface available to photoemission. This was
confirmed in earlier experiments19 with agglomerates of Pt
primary particles. When applying the thermal restructuring
technique described above to these agglomerates, it was indeed
found that the photon induced average charge per particle
remained proportional to the square of the mobility equivalent
diameter, as long as c and Φ were constant.
4.2. Relationship between APES and Catalytic Activity.
In the following discussion of the particle properties, extensive
use will be made of the correlation between photoelectric and
catalytic activity. The reason these two activities are, at any
time, proportional to each other is that both are directly related
to the local density of states (LDOS) at the Fermi energy level.
On one hand, the possibility for the emission of a photoelectron
(given by the emission constant) is determined by the LDOS
(EF). On the other hand, main steps in the course of a catalytic
5.2. Temporal Evolution of Catalytic and Photoelectric
Activity. As mentioned above, the catalytic activity of Ni aerosol
particles drops substantially above 350 °C due to the “poisoning”
by adsorbed carbon. The kinetics of such a poisoning is shown
in Figure 4. Beside the temporal evolution of the catalytic
activity of Ni nanoparticles during the methanation reaction at

8962 J. Phys. Chem. A, Vol. 105, No. 39, 2001
Weber et al.
Figure 6. Catalytic and photoelectric activity as a function of particle
preconditioning temperature (residence time of 38 s).
Figure 5. Photoactivity vs the amount of carbon deposited on the
particles (reaction temperature of 450 °C).
primary particle size. Figure 6 shows the catalytic and photo-
electric activity as a function of temperature, to which the
particles were exposed before reaction. As known from former
work by Schmidt-Ott et al.,23 the photoelectric activity, which
is correlated to the catalytic one, is reduced for decreasing
mobility equivalent diameter. However, in our case, the catalytic
activity even increases when the mobility diameter decreases
in the temperature range from 450 to 600 °C. Therefore, this
increase cannot be related to structural changes of the agglomer-
ates. The explanation may rather be related to the impurities in
the nickel (<1%). It is known that at elevated temperatures
impurities can move to the particle surface. This cleaning of
the particle interior is even more pronounced for nanoparticles.24
We hypothesize that the impurities may desorb or react and
desorb from the Ni surface resulting in cleaner catalytic particles.
This hypothesis is supported by the APES measurements, which
show for the particles heated to 600 °C emission constants close
to the value reported for ultrapure Ni nanoparticles.25
450 °C, the photoelectric activity is shown also. The catalytic
activity decreases substantially over a few seconds. After about
5 s, the reaction rate approaches a constant value, however, at
a low level compared to the initial activity. The initial decay of
the catalytic activity during the first seconds is enormous.
However, the amount of methane produced in the first mil-
liseconds is so small that a direct detection with FTIR is beyond
the measuring capabilities. With the current setup, the shortest
measured residence time was 250 ms. Here, as an alternative,
the correlation between photoelectric and catalytic activity17 can
be used. This means that instead of measuring the amount of
catalytically produced gas-phase products, the activity of the
particles is determined over the photoelectric activity. This
approach should thus be capable, providing a measure for the
activity of freshly formed surfaces on a time scale of as low as
a few milliseconds.
For temperatures above 625 °C, the catalytic activity de-
creases substantially with temperature, although no significant
change of mobility diameter or of the specific surface area has
been observed. Therefore, this decrease can be related to internal
changes only. In fact, Baumann26 reports that metal particles
approaching their sintering temperature can undergo changes
of the crystalline structure resulting in a decreased photoactivity.
For nanometer particles, the sintering temperature may be
reduced by several hundred degrees compared to the bulk
melting point. Since photoactivity and catalytic activity are
correlated via the local density of states at the Fermi energy
level, such changes of the crystal structure would affect the
catalytic activity of nanoparticles in the same way.
To investigate the influence of the deposited carbon on the
photoelectric behavior, samples were taken after different
residence times in the reactor and analyzed by APES and for
carbon content. The amount of carbon deposited on the
nanoparticle agglomerates was measured by TGA and FTIR.
The results are shown in Figure 5. The amount of carbon is
translated into a monolayer equivalent, i.e., the number of
monolayers if the carbon atoms would be distributed homoge-
neously over the total particle surface. It is obvious from Figure
5 that a drastic decrease of the photoactivity occurs when carbon
deposition starts. At several monolayers, the photoelectric
behavior of the Ni particles approaches the one of graphite
particles. The interesting point is that less than one monolayer
equivalent is needed to reduce the photoactivity and, therefore,
the catalytic activity of the nanoparticles successfully. This
indicates that even at this small scale the surface of 4 nm
particles contains sites of higher and lower activity. The carbon
seems to deposit on the active sites first leading to a very
efficient poisoning of the catalyst.
5.3. Influence of Agglomerate Structure Modification on
Catalytic Activity. So far, the behavior of untempered particles
has been studied. Next, the influence of the reduction of the
BET surface, i.e., the increase of primary particle size, at
constant aerosol mass concentration on the catalytic activity will
be discussed. Since the temperature in the reactor was 450 °C,
restructuring effects below this temperature were not investi-
gated. As shown in Figure 3 for temperatures above 450 °C,
only moderate changes of the agglomerate mobility were
observed, indicating that the apparent projection area of the
agglomerate was only reduced by a little. However, for
temperatures above 450 °C, a significant decrease of the specific
surface area was found, which is related to an increase of the
In any case, it is obvious that the particle history can have
an important influence on their catalytic performance.
Summary and Conclusions
A new approach to on-line investigations of heterogeneous
catalytic reactions on gasborne nanoparticles is presented. It
combines particle formation, restructuring and measurement
techniques via the aerosol route to obtain well-defined nano-
particle properties with measurements of gas phase composition
by FTIR spectroscopy. Time-resolved measurements in a flow
reactor permitted observation of the kinetics of catalyst deac-
tivation due to the formation of submonolayers of carbon. It
was possible to correlate the resulting decline in TOR with a
change in particle photoelectric activity, measured in real time
by APES. In addition, the influence of deliberately imposed
changes in particle morphology on the catalytic performance
was observed.
“Aerosol catalysis” is first of all shown to be a new tool for
understanding the relationship between particle properties and

Methane on Gasborne Nickel Nanoparticles
J. Phys. Chem. A, Vol. 105, No. 39, 2001 8963
(9) Katzer, M.; Weber, A. P.; Kasper, G. J. Aerosol Sci., in press.
(10) Wang, S. C.; Flagan, R. C. Aerosol Sci. Technol. 1990, 13, 230.
(11) Fissan, H.; Hummes, D.; Stratmann, F.; Bu¨scher, P.; Neumann,
S.; Pui, D. Y. H. Chen, D. Aerosol Sci. Technol. 1996, 24, 1.
(12) Schmidt-Ott, A.; Schurtenberger, P.; Siegmann, H.-C. Phys. ReV.
Lett. 1980, 45, 1284.
catalytic activity. It has potential for developing new catalysts
and may also be an avenue for investigating the putative
relationship between combustion aerosols and the formation of
dioxin.27
(13) Filipov, A.; Schmidt-Ott, A.; Fendel, W. J. Aerosol Sci. 1993, 24,
S501.
Acknowledgment. The financial support of this project by
the Ministerium fu¨r Wissenschaft und Kunst des Landes Baden-
Wu¨rttemberg, Germany, within the interfacultative research
group “Nanotechnology-controlled functional elements on the
nanometer scale” is highly appreciated.
(14) Schleicher, B.; Burtscher, H.; Siegmann, H.-C. Appl. Phys. Lett.
1993, 63, 1191.
(15) Burtscher, H. J. Aerosol Sci. 1992, 23, 549.
(16) Burtscher, H.; Schmidt-Ott, A. J. Aerosol Sci. 1986, 17, 699.
(17) Weber, A. P.; Seipenbusch, M.; Kasper, G. Chem. Ing. Technol.
2000, 72 (8), 879.
References and Notes
(18) Schmidt-Ott, A.; Baltensperger, U.; Jost, D. T.; Ga¨ggeler, H. W.
J. Aerosol Sci. 1990, 21, 711.
(1) Schmid, G.; Maihack, V.; Lantermann, F.; Peschel, S. J. Chem.
Soc., Dalton Trans. 1996, 5, 589.
(2) Glikin, M. A. Theor. Found. Chem. Eng. 1996, 30, 390.
(3) Glikin, M. A.; Kutakova, D.; Prin, E. Chem. Eng. Sci. 1999, 54,
4337.
(4) Weber, A. P.; Seipenbusch, M.; Thanner, Ch.; Kasper, G. J.
Nanoparticle Res. 1999, 1, 253.
(5) Klose, J.; Baerns, M. J. Catal. 1984, 85, 105.
(6) Schwyn, S.; Garwin, E.; Schmidt-Ott, A. J. Aerosol Sci. 1987, 19,
639.
(19) Weber, A. P.; Burkert, K.; Kasper, G. AIChE Proc. 1998, 1, 285.
(20) Thomas, J. M.; Thomas, W. J. Principles and Practice of
Heterogeneous Catalysis; VCH: Weinheim, Germany, 1997.
(21) Vannice, M. A. Catal. ReV. Sci. Eng. 1976, 14, 153.
(22) Goodman, D. W.; Kelley, R. D.; Madey, T. E.; Yates, J. T., Jr. J.
Catal. 1980, 63, 226.
(23) Schmidt-Ott, A. J. Aerosol Sci. 1988, 19, 553.
(24) Weller, H. Curr. Opin. Colloid Interface Sci. 1998, 3, 194.
(25) Mu¨ller, U.; Ammann, M.; Burtscher, H.; Schmidt-Ott, A. Phys.
ReV. B 1991, 44(15) , 8284.
(26) Baumann, F. Z. Physik 1960, 158, 607.
(27) Huang, H.; Buekens, A. Chemosphere 1995, 31, 4099.
(7) Weber, A. P.; Friedlander, S. K. J. Aerosol Sci. 1997, 28, 179.
(8) Rogak, S. N.; Flagan, R. C.; Nguyen, H. V. Aerosol Sci. Technol.
1993, 18, 25.