Influence of nitrogen on carbon arc plasma and formation of fullerenes

Article abstract of DOI:10.1016/S0009-2614(01)00333-5

The influence of nitrogen on the process of fullerene formation in the carbon arc discharge in a helium/nitrogen atmosphere has been studied. Applying optical emission spectroscopy and using the self-absorption effect, the spatial distribution of the temp

Full text of DOI:10.1016/S0009-2614(01)00333-5

25 May 2001
Chemical Physics Letters 340 12001) 1±6
www.elsevier.nl/locate/cplett
In¯uence of nitrogen on carbon arc plasma and
formation of fullerenes
H. Lange *, A. Huczko
Physical Chemistry Division, Laboratory of Plasma Chemistry, Department of Chemistry, Warsaw University, ul. Pasteur 1, 02-093
Warsaw, Poland
Received 23 January 2001; in ®nal form 6 March 2001
Abstract
The in¯uence of nitrogen on the process of fullerene formation in the carbon arc discharge in a helium/nitrogen
atmosphere has been studied. Applying optical emission spectroscopy and using the self-absorption e€ect, the spatial
‡
distribution of the temperatures and chemical species, i.e. C2ꢀa3Pu; v00 ˆ 0† and CNꢀX2R ; v00 ˆ 0† radicals produced
by the d.c. arc discharge between pure graphite electrodes, were determined across the arc gap. The soot product was
analyzed spectrophotometrically to determine C60 content. C60 was detected even when pure nitrogen was used 1at
about 0.5 wt%). The relative content of higher fullerenes was measured by mass spectrometry and their formation was
found to be signi®cantly enhanced in the presence of nitrogen. Ó 2001 Elsevier Science B.V. All rights reserved.
1. Introduction
equilibrium relationship. What happens to these
species in the arc itself is, of course, quite another
matter and various processes like ionization and
fragmentation can either remove them or generate
non-equilibrium species distributions. Thus, the
composition of the vapor resulting from the arc
ablation of carbonaceous bodies in high tempera-
ture environments is of interest, with respect to the
formation mechanism and yield of nanocarbons.
A great deal of work has been reported that
correlates fullerene species production and yields
with the operational parameters involved, pro-
viding an extensive and very useful body of engi-
neering data. These parametric studies have
provided, however, little insight into the detailed
mechanism of fullerene formation, indispensable
for synthesis scale-up. The process involves some
interesting and unusual chemistry [4]. Fullerene
formation represents growth from an initial species
with perhaps 1±10 carbon atoms to nanoparticles,
The properties of carbon vapor are of interest in
several areas of theoretical chemistry and physics
as well as for a variety of technologies and in-
dustrial uses. While much of the chemical interest
in carbon vapor has been recently connected with
the formation of novel carbon nanostructures, i.e.
the fullerenes [1] and nanotubes [2], the physical
interest has been largely connected with the spec-
tra of the stellar atmosphere or of ¯ames, in which
spectral lines assigned to various carbon species
have been observed [3]. From the considerations
regarding the d.c. carbon arcs it seems probable
that the distribution of Cn species leaving the an-
ode crater resembles, but is not equal to, an
* Corresponding author. Fax: +48-22-822-5996.
E-mail address: lanhub@chem.uw.edu.pl 1H. Lange).
0009-2614/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 9 - 2 6 1 4 1 0 1 ) 0 0 3 3 3 - 5

2
H. Lange, A. Huczko / Chemical Physics Letters 340 -2001) 1±6
with hundreds of `well ordered' carbon atoms 1the
higher fullerenes). Such rearrangements take place
at high temperatures, where the driving force from
entropy strongly opposes such atomic aggrega-
tions. With increasing pressure in the arc reaction
chamber the formation of chains from carbon
species becomes more extensive. At the optimal
pressure of 13.3 kPa [5] the development of the
three-dimensional network that is characteristic of
fullerenes is observed, while at higher pressures,
over-saturation in the system may lead to the
creation of graphite sheets and nanotubes.
Fullerenes have been routinely produced in the
helium carbon arc plasma for more than 10 years.
The involvement of other gases in the process may
result in decreased production costs. In the case of
nitrogen, the heterohedral substitution of carbon
atom1s) by nitrogen in the fullerene structures
could also yield novel compounds with properties
of amazing interest [6]. In the present study, the
formation of fullerenes in He=N2 mixtures was
investigated.
generally accepted, plays an important role in the
formation mechanism for fullerenes and nanotu-
bes. Hence, it is of considerable interest to be able
to quantitatively study its presence in the reaction
zone [10]. It should be mentioned here that there
have been several studies [11±13] on the applica-
tion of emission spectroscopy to the characteriza-
tion of the carbon arcs in the process of fullerene
formation, but these studies all deal with helium
plasma and qualitative data only were reported.
In the present Letter, which we believe is the
®rst account of research on the quantitative diag-
nostics of the carbon arc in a chemically reactive
environment, the in¯uence of nitrogen on the
plasma characteristics and the formation of fulle-
renes are reported. The composition of the arc
generated solid products was also studied by
means of various techniques.
2. Experimental
To obtain knowledge of the plasma character-
istics, the temperature and the plasma composition
were evaluated by optical emission spectroscopy
[7]. The spectral characterization of the arc zone
can provide new insights into the mechanism in-
volved in the transformation of fullerene nano-
structure precursors [8,9]. The C2 and CN radical
distributions were measured across the arc plasma.
Of the numerous possible intermediate carbon/ni-
trogen species these radicals are the only ones that
can be quantitatively measured using conventional
optical emission spectroscopy. The C2 radical, as is
The reaction system and the experimental
procedures employed have been presented else-
where [14]. Graphite electrodes, 6 mm in diame-
ter, were subjected to d.c. arcing in He=N2 gas
mixtures under static conditions and pressures of
about 13.3 kPa. The electrode gap was held
within 1 mm by means of an opto-electronic
system [14]. Table 1 contains the averaged pro-
cessing variables. Compared to the pure helium
plasma [8] higher anode erosion can be related to
the chemical etching of graphite by the plasma
activated nitrogen.
Table 1
Operating parameters and anode sublimation rates
No.
He ‡ %N2
I
1A)
U
1V)
p
1mb)
msub
1mg sÀ1
)
1
2
0
0
76
74
78
81
73
80
70
75
67
75
20.8
20
128
140
132
144
123
150
134
161
152
166
5.3
5.6
6.2
9
3
10
10
25
25
50
50
100
100
21.5
19.9
21.8
20.5
20.4
20.2
21.7
24
4
5
6.3
8
6
7
6.2
8
8
9
10
5
10

H. Lange, A. Huczko / Chemical Physics Letters 340 -2001) 1±6
3
tents are shown in Fig. 1. The self-absorption is
clearly visible in the position of the P10±0) band
head where many rotational lines overlap and
cause a disappearance of the band head in the case
of a high nitrogen content in the gas mixture.
The soot produced by the arc sublimation of the
anode was subsequently re¯uxed in dry toluene
prior to C60 determination by the conventional
spectrophotometric measurements performed at
329 nm [8]. For the identi®cation of the content of
higher fullerenes, the soot was analyzed by the
Laser Desorption TOF MS technique.
3. Results and discussion
3.1. Thermodynamic considerations
To evaluate the composition of reactants the
thermodynamic equilibrium calculations of a car-
bon/nitrogen mixture have been carried out for the
temperature range between 1000 and 6000 K, with
pressure ®xed at 15 kPa and the input C/N ratio
equal to 1 1as an example). From the results pre-
sented in Fig. 2 one can predict that at the tem-
peratures prevailing in the arc, i.e. between 4000
and 6000 K, the plasma consists mainly of C, N2,
‡
‡
Fig. 1. Experimental examples of CNꢀB2R ; v0 ˆ 0 ! X2R ;
v00 ˆ 0; 388, 3 nm) emission band emitted by arc discharge in
He=N2 atmosphere.
The CCD ST8 camera, coupled with a spec-
trograph, was used for the plasma diagnostics. The
C2 contents were determined using the well-pro-
nounced self-absorption e€ect in the d3Pg;
v0 ˆ 0 ! a3Pu; v00 ˆ 0, 1516.5 nm) transition. The
column densities of C2 1the product of average
concentration and column length along a given
chord in the arc cross-section) were calculated on
the basis of the temperatures and integrated in-
tensities of the normalized C2 1d±a, 0±0) band
using an a priori calculated functional dependence
of the integrated intensities on the C2 column
density for various temperatures [7]. Temperatures
were evaluated from the Boltzmann plots, using
the resolved rotational components of this band.
‡
‡
Since
3 nm) band is also strongly self-absorbed, a similar
procedure was elaborated and applied in order to
CNꢀB2R ; v0 ˆ 0 ! X2R ; v00 ˆ 0; 388;
‡
®nd the CNꢀX2R ; v00 ˆ 0) column density across
the arc [15]. Examples of the experimental inten-
sity distributions within this band emitted from the
C=N2=He arc plasma for di€erent nitrogen con-
Fig. 2. Equilibrium composition of carbon/nitrogen plasma:
p ˆ 15 kPa, [C]/[N] ˆ 1.

4
H. Lange, A. Huczko / Chemical Physics Letters 340 -2001) 1±6
N, CN and C2 species. The lower temperatures
favor the formation of higher carbon species, e.g.
C5 and C2N2. The temperature dependence for the
stability of the CN and C2 radicals is much the
same, with a maximum concentration at around
4200 K. Obviously, the starting molar ratio of the
reactant will change the relative content of species
under consideration.
It must be noted, however, that in the actual arc
plasma system di€erent process variables can dis-
tinctly in¯uence the theoretical reaction course.
For example, rapid quenching may lead to the ®-
nal concentrations far exceeding those predicted
from the equilibrium conditions [16]. Thus, equi-
librium thermodynamics are often of limited use in
predicting the ®nal compositions, owing to subli-
mation, condensation, and kinetic barrier e€ects
but they do contribute to a better understanding of
the complex plasma phenomena.
3.2. Plasma temperature and C2 content
The average temperature distributions across
the arc gap are shown in Fig. 3A. Regardless of the
nitrogen content in the arc chamber, the temper-
ature pro®les are similar with a maximum close to
6000 K along the chord crossing the arc axis, and
3500 K on the plasma edge. The observed minima
in the region close to the arc center are a conse-
quence of a systematic error which results from the
self-absorption e€ect. It was shown earlier [15] that
an increase in the C2 density 1and, consequently,
an increase in self-absorption) can lower the tem-
perature values, evaluated from the Boltzmann
plots, by about 10%. This is exactly the case for the
higher contents of N2 150% or 100%) in the present
study.
The radial distributions of the column density
of the C2ꢀa2Pu; v00 ˆ 0† radicals are shown in Fig.
3B. Both for pure carbon plasma and under the
presence of nitrogen, the C2 species are mostly
concentrated near the arc axis, with the distribu-
tions being slightly expanded in the latter case. At
lower nitrogen concentrations 1up to 25% N2) the
C2 column densities are close to the one without
nitrogen with slightly di€erent radial distributions
only. It must be admitted, however, that because
of the smaller gap, the accuracy of the positioning
Fig. 3. Radial distributions of A: temperature, B: column
density of C2ꢀa3Pu; v00 ˆ 0† and C: column density of
CNꢀX2R ; v00 ˆ 0† in the arc plasma.
‡

H. Lange, A. Huczko / Chemical Physics Letters 340 -2001) 1±6
5
of the plasma cross-section image onto the spec-
trograph slit was within Æ0.2 mm. Thus even a
small horizontal displacement can yield results
from a di€erent arc cross-section than was in-
tended.
At higher nitrogen contents a distinct increase
of C2 radicals is observed. Since the dissociation
energy of C2 16.2 eV) is signi®cantly lower com-
pared to that of CN 17.8 eV) such a feature was
rather unexpected. As it follows from thermody-
namics 1Fig. 2), the presence of N2 should rather
lower somehow the C2 equilibrium concentration
because of the CN formation. It is also very likely
that the expansion of carbon species is hampered
by the heavier nitrogen atoms and molecules
1compared to helium) thus resulting in the higher
local C2 concentrations in the arc gap.
was found from the IR measurements of the gas-
eous reactants that the concentration of cyanogene
after 15 min of the arc operation in nitrogen was
relatively small. The amount of C2N2 formed in
the arc discharge in the pure N2 gas was only
about three times higher in comparison to the
amount detected in the case of the discharge in
pure helium, contaminated by nitrogen only
1<0.05%). This e€ect can be related to such factors
as low dissociation degree of N2 at the prevailing
temperatures 1Fig. 2) and the relatively ine€ective
gas mixing due to the limited convection and dif-
fusion processes in the arc gap. Also, the real
carbon arc discharge conditions may not strictly
match the conditions of the local thermodynamic
equilibrium [18] as mentioned earlier.
3.4. Yield of C60 and higher fullerenes
3.3. CN and C2N2 content
The measured contents of C60 fullerene in the
soot collected from the di€erent tests are presented
in Fig. 4. It is evident that the nitrogen drastically
lowers the C60 formation yield. However, it is in-
teresting to note that even in the case of pure ni-
trogen, C60 is still formed 1ca. 0.5 wt%).
The column density distributions of CNꢀX2R ;
‡
v00 ˆ 0† radicals are shown in Fig. 3C. As expected,
with an increase of the nitrogen in the gas mixture
the CN densities are higher. Formation of CN
radicals depends on the penetration of the nitrogen
within the carbon cloud expanding from the arc
gap. The mixing of the oppositely directed ¯uxes
of nitrogen and carbon species, and also inherent
temporal instabilities of the arc discharge [17], are
responsible for the observed non-regular radial
distribution of the CN radicals.
One can see that, in contrast to C2, the CN
species are widely spread beyond the arc region
and are abundant at the plasma edge. A transition
from the column densities to the average local
concentrations can be simply made by dividing
column densities by the radial coordinates of the
radicals, which for CN are much greater 1ꢁ7 mm)
than for C2 1ꢂ5 mm). Since the column densities of
both radicals are close to each other, one can
conclude that the expanding carbon gas e€ectively
constrains the penetration of the nitrogen into the
relatively narrow arc gap. It also shows that the
CN radicals are formed mainly in the cooler zone
just outside the electrodes where the mixing of
carbon gas with nitrogen is not spatially disturbed.
The recombination of CN radicals results in the
formation of the stable C2N2 gas. Surprisingly, it
Fig. 4. Yield of C60 vs. N2 content in He=N2 mixtures.

6
H. Lange, A. Huczko / Chemical Physics Letters 340 -2001) 1±6
Acknowledgements
This work was supported by the Committee for
Scienti®c Research 1KBN) through the Depart-
ment of Chemistry of Warsaw, under grant no. 7
T09A 020 20. Thanks are due to Prof. H. Shino-
hara and Dr. A. Koshio, Department of Chemis-
try, Nagoya University 1Nagoya, Japan) for their
important collaborative MS works.
References
Fig. 5. MS of soot produced under di€erent N2 content in
He=N2 mixtures: 0%, 10% and 100%.
[1] H.W. Kroto, R. Heath, S.C. O'Brien, R.F. Curl, R.E.
Smalley, Nature 318 11985) 162.
[2] S. Iijima, Nature 354 11991) 54.
[3] H. Kroto, Angew. Chem. Int. Ed. Engl. 36 11997) 1578.
[4] S.J. Harris, A.M. Weiner, Ann. Rev. Phys. Chem. 36
11985) 31.
It is known that during the arcing of the
graphite in helium not only C60; but also the
higher, mainly C70, fullerenes are formed. There-
fore the relative mass distributions of these species
in the solid products were measured by LD TOF
MS technique. The resulting mass spectra, nor-
malized to the C60 peak, are presented in Fig. 5.
The content of the higher fullerenes, particularly
the so-called `missing fullerenes' C72; C74 and C80
[19], in the case of the pure He arc is, as expected,
low. However, the positive in¯uence of the nitro-
gen on the formation of these species is evident,
and resembles the one observed earlier when an
iron-doped graphite anode was arced in helium
[20]. It is clearly seen that in the presence of ni-
trogen the formation of higher fullerenes is dis-
tinctly enhanced. Thus, one can conclude that
nitrogen atoms or molecules play a role in gener-
ating these, otherwise `missing', fullerene species.
In considering the fullerene composition in the
soot, it has to be kept in mind that the ®nal yields
also depend on the eciency of the extraction
method. The spectral analyses of the toluene ex-
tracts from the collected soot samples did not re-
veal any new absorption features, which could be
assigned to the nitrogen-doped heterofullerenes
[6]. These relatively unstable species, if present,
cannot, however, be extracted by the `conven-
tional' fullerene solvents such as toluene used in
this study.
[5] Y. Saito, M. Inagaki, H. Shinohara, H. Nagashima, M.
Ohkohchi, Y. Ando, Chem. Phys. Lett. 200 11992) 643.
[6] A. Huczko, Full. Sci. Technol. 5 11997) 1091.
[7] H. Lange, K. Saidane, M. Raza®nimanana, A. Gleizes, J.
Phys. D 32 11999) 1024.
[8] A. Huczko, H. Lange, P. Byszewski, M. Popøawska, A.
Starski, J. Phys. Chem. A 101 11996) 1267.
[9] H. Lange, A. Huczko, P. Byszewski, E. Mizera, H.
Shinohara, Chem. Phys. Lett. 289 11998) 174.
[10] Y. Xia, Y. Mu, S. Li, H. Zhang, C. Tan, L. Mei, Nucl.
Instrum. Meth. B 155 11999) 395.
[11] D.V. Afanas'ev, A.A. Bogdanov, G.A. Dyuzev, A.A.
Kruglikov, Mol. Mat. 7 11996) 93.

[12] T. Belz, R. Schlogl, in: H. Kuzmany et al. 1Ed.), Proceed-
ings of the International Winterschool on Fullerenes and
Fullerene Nanostructures, Kirchberg, Austria, 1996, p. 45.
[13] X. Zhao, T. Okazaki, A. Kasuya, H. Shimoyama, Y.
Ando, Jpn. J. Appl. Phys. 38 11999) 6041.
[14] H. Lange, P. Baranowski, A. Huczko, P. Byszewski, Rev.
Sci. Instrum. 68 11997) 3723.
[15] H. Lange, in: M. Hrabovsky, M. Konrad, V. Kopecky
1Eds.), 14th International Symposium on Plasma Chemis-
try, Symposium Proceedings vol. I, Institute of Plasma
Physics, Prague, 1999, p. 221.
[16] Y. Pang Tsui, H.Y. Cheh, Plasma Chem. Plasma Process 2
11982) 387.
[17] J. Abrahamson, Carbon 12 11974) 111.
[18] S.C. Snyder, G.D. Lassahn, L.D. Reynolds, Phys. Rev. E
48 11993) 4124.
[19] T.S.M. Wan, H.-W. Zhang, T. Nakane, Z. Xu, M.
Inakuma, H. Shinohara, K. Kobayashi, S.J. Nagase, Am.
Chem. Soc. 120 11998) 6806.
[20] A. Huczko, H. Lange, T. Sogabe, J. Phys. Chem. A 104
12000) 10708.