Formation of Nanoparticles by Photolysis from Metal and Carbon Bearing Molecules

Article abstract of DOI:10.1524/zpch.217.11.1361.20483

The formation of particles following the photolysis of C₃O ₂, Fe(CO)₅ and Mo(CO)₆, diluted with Ar or He was registered at room temperature. Particle growth was followed by taking light extinction profiles at 633 nm and at 220 nm for various mixture compositions and pressures. The particles obtained at different conditions were analyzed using transition electron microscope (TEM) technique. It was found that in the pure undiluted gases at the partial pressures shown in the pictures no light absorption and no particle formation could be observed. Light absorption started for partial pressures of the diluent gas > 10 mbar. A comparison of particle size measured here at room temperature with data obtained at elevated temperature shows that the data obtained here fit well to the elevated temperature data.

Full text of DOI:10.1524/zpch.217.11.1361.20483

Z. Phys. Chem. 217 (2003) 1361–1368
by Oldenbourg Wissenschaftsverlag, München
Formation of Nanoparticles by Photolysis from
Metal and Carbon Bearing Molecules
By A. Emelianov2, A. Eremin2, H. Jander1, and H. Gg. Wagner1,
∗
1
Institut für Physikalische Chemie, Universität Göttingen, Tammannstr. 6,
D-37077 Göttingen, Germany
Institute for High Energy Density RAS, Izhorskaya str. 13/19, 127412 Moscow, Russia
2
(Received June 23, 2003; accepted in revised form August 11, 2003)
Particle Formation / Photolysis of C3O2 / Fe(CO)5 Mo(CO)6 at Room
Temperature
The formation of particles following the photolysis of C3O2, Fe(CO)5 and Mo(CO)6, di-
luted with Ar or He was registered at room temperature. Particle growth was followed by
taking light extinction profiles at 633 nm and at 220 nm for various mixture compositions
and pressures. The particles obtained at different conditions were analyzed using transition
electron microscope (TEM) technique. It was found that in the pure undiluted gases at the
partial pressures shown in the pictures no light absorption and no particle formation could be
observed. Light absorption started for partial pressures of the diluent gas > 10 mbar. A com-
parison of particle size measured here at room temperature with data obtained at elevated
temperature shows that the data obtained here fit well to the elevated temperature data.
1. Introduction
The process of small condensed particles formation from the supersaturated
vapor became of interest in recent years due to the numerous applications of
nanoparticles and clusters. Two different kinds of nanoparticles are metal and
carbon particles. The first ones are usually assumed to have uniform structure,
like frozen drops, while carbon nanoparticles are formed in numerous vari-
ants of different structures. For a long time the commonly used ways to get
carbon nanoparticles like soot, were to produce them via combustion and py-
rolysis of hydrocarbons. During the last decade the interest in fullerens and
carbon nanotubes synthesis initiated the application of the different kinds of
discharge and laser plasma. All these processes generate the vapor however
with poorly known parameters, therefore kinetic mechanisms of cluster growth
could hardly be analyzed and controlled.
*
Corresponding author. E-mail: jkupfer@gwdg.de

1362
A. Emelianov et al.
Here we report about another way for the formation of various nanopar-
ticles and their study. As initial substance for the formation of carbon nanopar-
ticles carbon suboxide (C3O2) can be used. According to [1], under the influ-
ence of UV radiation with λ < 207 nm carbon suboxide decomposes forming
carbon atoms and two CO molecules. Thus, supersaturated carbon vapor with
well known and easily controlled properties could be obtained. Then, de-
pending on conditions (concentration, temperature, kind of diluent gas and
its pressure) the condensation and formation of carbon particles, (with differ-
ent properties) takes place. Even more easily one can get metal nanoparticles,
using metal pentacarbonyls, like Fe(CO)5 or Mo(CO)6. The formation of iron
atoms [2] and small iron clusters (up to 10 atoms) [3] during Fe(CO)5 pho-
tolysis in the range 193–400 nm was already observed. The application of
photodissociation of metal carbonyls is also known as a way to grow thin ul-
trapure metal films on surfaces [4]. However it seems that the formation of
metal nanoparticles in the volume of supersaturated metal vapor was not yet
investigated in detail.
2. Experimental
The experiments described here were performed using a very simple set up,
consisting of a cylindrical brass cell of 5 mm diameter and 70 mm length with
quartz windows at the ends. The cell could be evacuated down to 1 µbar and
filled with the various mixtures through a side tube. As a source of radiation
a usual high pressure Xe lamp of total power 150 W was used. The optical
scheme was arranged to irradiate the whole volume inside the cell. The meas-
ured total power flux of radiation J, entering into the cell, was about 5 W/cm2.
To observe the condensed particle formation the attenuation of the Xe lamp ra-
diation at the wavelengths 220 nm along with the attenuation of a He–Ne laser
beam (633 nm) were measured. After several experiments samples of particles,
deposited on the walls, were analyzed using transition electron microscope
(TEM) technique.
In the experiments performed mixtures of C3O2, Fe(CO)5 and Mo(CO)6
diluted by Ar (or He) were used. The partial pressure of C3O2 and Fe(CO)5
was varied in the range 10–10000 µbar, pressure of Mo(CO)6 was varied in
the range 10–80 µbar (limited by the saturation pressure at room temperature).
The maximum pressure of Ar used was ≤ 1 bar. The process of condensed par-
ticle formation was observed at all pressures used. It proved to be essentially
dependent on the diluent gas pressure. At diluent gas pressures below 10 mbar
no extinction signals were observed. In Fig. 1 typical examples of extinction
profiles, measured at 633 nm and 220 nm are shown. At the beginning of an ex-
periment only the He–Ne laser is switched on. Then the Xe-lamp is fired and
immediately the signal of radiation at 220 nm falls down. The He–Ne laser sig-
nal first rises due to the additional radiation from the lamp and then also steeply

Formation of Nanoparticles by Photolysis . . .
1363
Fig. 1. Examples of extinction profiles at 633 nm and 220 nm in Fe(CO)5 and C3O2. At the
beginning of experiment only He–Ne laser is switched on. At t = t1 Xe-lamp is fired. After
the stationary level of extinction is attained (at t2) the evacuation of cell is open.
decreases. After the stationary level of extinction is attained the cell is evacu-
ated which results in the immediate increase of both signals, proofing that most
particles were formed in the volume and not on the windows. Similar profiles
were recorded in all mixtures. It was observed that the time scale of extinction
rise in C3O2 was noticeably longer then in metal carbonyls. Tests performed in
pure Ar did not show any extinction after the lamp firing.
3. Results and discussion
The main goal of these preliminary experiments was to observe new phe-
nomena of metal and carbon nanoparticle formation directly in the volume of

1364
A. Emelianov et al.
supersaturated vapor yielded from UV photolysis of metal carbonyls and car-
bon suboxide. According to the few reference data the best yield of pure Fe
atoms from Fe(CO)5 should appear at photolysis in the range 200–300 nm [3],
a similar spectral range one could assume for Mo(CO)6. However up to now no
results on metal nanoparticles formation in the volume of metal vapor at room
temperature were reported. The observed multistage process, ending by the
metal nanoparticles growth and formation seems to be an interesting and com-
plicated phenomenon. In carbon suboxide the situation is even more intricate.
Authors of [5] demonstrated that radiation in the range 300 nm > λ > 207 nm,
yields:
C3O2 +hν (207–300 nm) → C2O+CO
(1)
while only radiation shorter than 207 nm could yield carbon atoms:
C3O2 +hν (< 207 nm) → C+2CO .
(2)
Since only a small part of the Xe-lamp radiation, transparent for the air and the
quartz windows could cause the carbon vapor yield, the observed high yield
of carbon particles was rather unexpected. A possible explanation of this phe-
nomenon could be that process (1) proceeds in the mixture in both directions
and all recombining C3O2 molecules get (earlier or later) necessary quanta
λ < 207 nm.
It is easy to estimate the time, necessary for the photolysis of the studied
mixtures. For metal carbonyls the active part am of the whole Xe-lamp spec-
trum could be accepted as am = 0.1–0.2, for C3O2 the active part of spectrum
ac is much less, it is not more than 0.02. The average photon energy for disso-
ciation of metal carbonyls Em is about 5 eV, for C3O2 – Ec is about 6 eV. Thus,
the dissociative photon flux Fi could be expressed as:
Fi = ai J/Ei [phot/cm2 s] .
(3)
Here index i = m stands for metal particles, and i = c for carbon particles.
For metal carbonyls Fm is about 1018 phot/cm2 s, for C3O2 Fc is about
1017 phot/cm2 s. The absorbed part of the flux ∆Fi is
∆Fi = Fi [1−exp(−σnl)]
and for optically thin conditions
∆Fi ≈ Fi σnl
(4)
(5)
where n is the concentration of absorbing molecules, σ is the absorption cross
section, which could be estimated according to [1] as 10−18–10−19 cm2. The
number of molecules on the pass of the flux is N = nl. If for the yield of one

Formation of Nanoparticles by Photolysis . . .
1365
atom mi photons are necessary (i.e. for Fe(CO)5 mm = 5, for C3O2 it was as-
sumed that mc = 1), then one can estimate the total time of photolysis t of the
mixture as:
ti ≈ mi N/∆Fi ≈ mi/Fiσ .
(6)
The time of photolysis ti does not depend on the concentration n. For metal
carbonyls it is
tm ∼ 5–50 s
(7)
and for carbon suboxide
tc ∼ 10–100 s .
(8)
These rough estimations are in good agreement with the observed time profiles
of the extinction rise.
An interesting phenomenon observed in the experiments is the strong de-
pendence of the extinction level on the diluent gas pressure. First, it was
noticed, that in all pure, undiluted gases of metal carbonyls and carbon subox-
ide no light absorption and no particle formation could be observed. Addition
of diluent gas with partial pressure > 10 mbar resulted in the appearance of
light extinction, see Fig. 1. The obtained data of the attenuation of passed ra-
diation were evaluated using Lambert–Beer law and normalized by the total
available concentration of vapor atoms n and the optical length l:
D = − ln(I/Io)/nl .
(9)
Here D – is the normalized optical density, Io – initial intensity of the refer-
ence radiation at the given wavelength, I – current intensity of the radiation
passed through the cell. In Fig. 2 the resulting optical densities observed at the
different partial pressures of various substances are given in dependence of the
pressure of the diluent gas. One can see that the rise of the diluent gas pressure
causes the increase of D, indicating the increase of particle yield. This increase
approaches the maximum at pressures of about 1 bar. On the other hand, a rise
of the partial pressure of reacting molecules leads to the fall of the normalized
optical density (compare data for 10 µbar and 100 µbar of Fe(CO)5 in Fig. 2).
This interesting fact can be probably related to the essential heat release during
particle formation and following overheating of the growing clusters, restrain-
ing their growth, especially at high concentrations of supersaturated vapor and
low pressures of diluter. An additional evidence for this version is the lower op-
tical density observed in He in comparison with Ar at the same partial pressures
of Fe(CO)5.
It is important to underline, that extinction measurements could give only
indirect information of forming particles, while the actual size and the structure

1366
A. Emelianov et al.
Fig. 2. Resulting normalized optical density D, observed in various mixtures depending on
diluent gas pressure.
of the particles could not be extracted from these data. One can mention only
one interesting evidence of the change of particle properties during the obser-
vation – the time profiles of extinction at 633 nm and 220 nm look differently.
This is clearly seen in C3O2 (Fig. 1b) – extinction at 633 nm rises much slower
than extinction at 220 nm. Most likely this phenomenon reflects the rise of the
particles during the coagulation process, when their size becomes comparable
with the wavelength and scattering becomes more and more essential.
The most direct data about the final particle size and structure gave the
electron microscope analysis of the particles, deposited on the cell walls after
experiments. In Fig. 3(a,b) the examples of TEM images of iron and carbon
particles are given. For experiments with iron particles one can see the large
plates (lamellas) with the size up to 2 µm which most probably were formed
partly on the walls, and besides that several smaller spheres with the size
150–250 nm could be clearly observed. In C3O2 photolysis experiments only
such spheres with the sizes from 200 to 400 nm could be observed. These facts
demonstrate the results of direct particle formation in the volume of supersatu-
rated vapor.
The presented data are perhaps the first observation of the metal and carbon
vapor condensation at well known pressures and room temperatures. It is there-
fore interesting to compare these data of the resulting particle size with recent

Formation of Nanoparticles by Photolysis . . .
1367
Fig. 3. Examples of TEM images of iron (a) and carbon (b) particles.
Fig. 4. Comparison of the particle size formed in photolysis experiments described here
(open symbols) with shock wave pyrolysis LII-sizing measurements in equal mixtures
(filled symbols).

1368
A. Emelianov et al.
observations of similar processes in shock tubes, at elevated temperatures. In
Fig. 4 the preliminary results of final particle size measurements in the present
photolysis experiments are given together with the temperature dependencies
of carbon and iron particle size measured by laser induced incandecsence (LII)
in the recent experiments [6] and [7]. One can see that all data roughly obey the
same temperature dependence of the size discussed in [8]. Therefore further ex-
periments on nanoparticles photosynthesis could present valuable information
on the mechanisms of condensed particle growth in the volume of condensating
vapor.
Acknowledgement
The support of these investigations by DFG, The Göttingen Academy of Sci-
ences, and the RFBR is gratefully acknowledged.
References
1. H. Okabe, Photochemistry of small molecules, J. Wiley & Sons Inc., New York
(1978).
2. L. Banares, T. Baumert, M. Bergt, B. Kiefer, and G. Gerber, J. Chem. Phys. 108
(1998) 5799.
3. F. Huisken, B. Koln, R. Alexanderescu, and I. Mrorjan, Eur. Phys. J. D 9 (1999) 144.
4. R. L. Jackson et al., in Laser Microfabrication (1999), pp. 385–451.
5. W. Braun, A. M. Bass, D. D. Davis, and J. D. Simmons, Proc. Roy. Soc. A 312
(1969) 417.
6. R. Starke, B. Kock, P. Roth, A. Eremin, E. Gurentsov et al., Combust. Flame, in
print (2003).
7. A. Eremin, E. Gurentsov, P. Roth, and R. Starke, Int. Symp. Combustion and Atmo-
spheric Pollution, St. Petersburg, Russia (2003).
8. A. Emelianov, A. Eremin, H. Jander, and H. Gg. Wagner, Z. Phys. Chem. 217
(2003) 1.