close

Вход

Забыли?

вход по аккаунту

?

On the Mechanism of Fullerene Formation.

код для вставкиСкачать
COMMUNICATIONS
On the Mechanism of Fullerene Formation**
Thilo Belz, H a r a l d Werner, Friedrich Zemlin, U r s u l a
Klengler, Michael Wesemann, Bernd Tesche, E l m a r
Zeitler, Armin Reller, and Robert Schlogl*
Dedicated to Prqfissor Hans Bock
on the occasion of' his cirfli hirthdaj.
Fullerenes['] differ in their molecular character from the currently known forms of elemental carbon such as graphite (including derived turbostratic variations) and diamond. They do
not form extended two- or three-dimensional structures, but
closed molecular systems based on s p ' a r b o n atoms as the
common building block. Fullerenes can currently be prepared
by thermal treatment of graphite['] and pyrolysis of arenes in
flames.[31Separation and purification techniques are also available.[41Graphitic tubules['' and onionlike spheres,"] which are
related to the fullerenes, have been structurally characterized
and produced in microscopic quantities, but are not yet available in pure form and preparative amounts.['] Given both the
diversity of products and methods of production, and the necessity to develop generally applicable syntheses for novel carbon
forms, the question arises as to the mechanism of formation of
these compounds. Methods involving the thermal treatment of
carbon in an electric arc, with fluxes of highly energetic electrons, or with a laser beam have the disadvantage that only
limited control over the selectivity of formation of particular
products has been possible up to now; the exact production
parameters do, however. have a crucial influence on the product
distribution.f8]We chose to investigate the formation of fullerenes in an electric arc despite the low degree of control over the
course of the reaction, since this is now the most important
method for the production of significant amounts of fullerenes.
When more is known about the formation processes, this technique should also have considerable potential in the preparation
of molecular forms of carbon other than soluble fullerenes.
The initial formation of C , fragments containing no heteroatoms is the common step in all synthetic routes to fullerenes.
According to numerous growth theories,['] these fragments
form small rings that combine with C, and C, units to give
additional five- or six-membered rings and eventually fullerenes
(Scheme 1). A second growth theory postulates that the first
step is the formation of large monocyclic compounds. Shock
annealing under certain conditions then results in the formation
of numerous linkages between the carbon atoms in these monocycles to give carbon clusters. These then isomerize to fullerenes.
which are thermodynamically stable under the reaction conditions. Evidence for this sequence was provided by mass-spectrometric experiments in the vapor phase under model conditions
used for the synthesis of fullerenes.['O1The suggested mechanisms'"] differ on the question whether many small rings condense together and oligomerize via polycyclic arenes to form
fullerenes, o r whether a large carbon ring isoinerizes to produce
a fullerene molecule.
All reaction pathways leading to fullerene molecules require
an atmosphere that is as inert as possible. This ensures that the
lower chemical stability of five- compared to that of six-membered rings does not result in a decrease in concentration of the
five-membered ring structures below a critical level. Furthermore, reactants which form stable carbon - heteroatoin bonds
under the reaction conditions should not be present, since they
would prevent the formation of an all-carbon compound.
Previous investigations into the mechanism of formation relied on experiments carried out in the vapor phase with massselected sections of the primary product spectrum obtained
from carbon vaporization. Due to a microstructure that is X-ray
amorphous, fullerene black, which is the main product formed
in the synthesis of fullerenes, has hardly been investigated structurally.[' 21 Studies in this area could provide significant clues on
the mechanism of formation. Assuming that fullerene black
arises from all the failed self-organization processes which lead
to fullerenes, we would expect a microstructure consisting of
platelets if the reaction proceeds by polycondensation of small
rings and a three-dimensional, random microstructure if the
isomerization process dominates.
The experiments were performed with an arc-discharge carbon vaporization apparatus," 31 which produces soluble fullerenes with a 10 % selectivity. Using emission spectroscopy,['41we
were only able to observe the bands corresponding to C, and C ,
units. Figure 1 shows the low-resolution spectra with band assignments.["] The substantial change in the partial pressure of
helium, which causes an increase in yield from 0.025 to l o % , is
clearly not accompanied by a change in the nature of the species
in the plasma. The reactions in the graphite vaporization experiments leading to the formation of the C, clusters and to fuller469
Scheme 1 . Postulated mechanisma for the formation of fullerenes.
[*] Prof. Dr. R. Schlogl, DipLChem T. Belz, DipLChem. H. Werner.
Dr. F. Zemhn, U . Klengler. M. Wesemann. Dr. B. Tesche.
Prof. Dr. E. Zeitler
Fritz-Haber-Insritut der Max-Planck-Gesellschaft
Faradayweg 4 6. D-14195 Berlin ( F R G )
Telefax: Int code + (30)8305-487
[**I
Prof. Dr. A. Reller
Institut fur Anorganische Chemie der Universltit Hamburg (FRG)
This work was supported by the Fonds der Chemischen Industrie. the
Hermann-Willkom-Stiftung and the Bundesministerium fur Forschung und
Technologie (Fullerene Pilot Project)
I
200
#
300
400
-
500
k[nm]
600
700
800
Fig. 1. Emission spectrum of ;in electric arc during fullerene synthesis. Discharge
conditions: graphite rods for spectroscopy (Ringsdorf RW 4). diameter 6. I5 mm,
current 150 A, alternating current. distance between electrodes (regulated) 3.0 mm.
The helium pressure during the measurement of spectra a and b ( x 4) was 10 and
2 x lo4 Pa, respectively. The band at 249 nm IS assigned to the C , fragment; all other
bands belong to the transitions of singlet and triplet C, molecules (232: Mulliken
system: 360, 385: Deslandres d'Aaarnbuja system: 437. 469. 512. 550: Swan system). Refer to the text for a discussion of those signals overlapping with bands of
the CN molecule.
COMMUNICATIONS
enes obviously proceed in the cool part of the carbon cloud
outside the range accessible to emission spectroscopy (diffuse
background in the visible region). The spectra thus give an
insight into the initial precursors of fullerenes, but not into the
formation of larger carbon frameworks. Even the small C, fragment with an expected principal band at 405 nm cannot be detected. This is consistent with earlier observations in laser vaporization experiments that C, fragments were detected by mass
spectrometry in the cold region of the apparatus, but not in the
plasma emission spectrum.[’61This evidence emphasizes the significance of the colder zones within the reaction chamber for the
construction of larger frameworks. The hot zone is, in contrast,
the source of building blocks.” The rate at which the building
blocks are supplied is, however, fundamental to the whole process and is determined by the parameters in the hot zone. as can
be seen from the spectra in Figure 1.
The different intensity ratios of the bands corresponding to
the C 2 fragments in the spectra of Figure 1 arise because of
overlap in the 360 and 385nm regions with bands assigned to
CN molecules, which dominate when traces of nitrogen are
added to the quench gas. Their appearance at low partial pressures of helium (spectrum a) is explained by the Fact that the
carbon fragments react with residual air components in the absence of the inert quench gas. This has consequences for the
growth mechanism.
A systcmatic investigation into the structural and chemical
properties of conventional carbon black and fullerene black
indicates that the two nanocrystalline carbon forms differ
markedly. This can be deduced from the differing microstructure, which has been investigated by means of high-resolution
transmission electron microscopy (TEM) .[’*I In the mesoscopic
region both materials consist of spherical aggregates of many
primary building blocks.””1
Figure 2 clearly reveals that the spherical aggregates in conventional carbon black are formed from many planar basic
structural units (BSU),[”] which are stacked in a parallel
ously curved multilayers of carbon can be recognized inside the
particle. The spacing of parallel layers is similar to that observed
in conventional carbon black or graphite. The carbon layers are
far more extended[221
than those in BSU of conventional carbon
black. Large parts of the particles are composed of amorphous
carbon that is not arranged in two-dimensional layers. but
rather as a tangle of threadlike polymers.
Fig. 3. High-resolution TEM image offullerene black. The experimental conditions
were adjusted to optimize the C, yield. One can see the empty cavities inside the
continuous carbon loops. The enhanced contrast ofthe loops suggests a two-dimensional structure of layered aggregates which are not quite parallel to one another.
The length of the bar represents 10 nm.
If energy input and quench-gas pressure in the plasma chamber are reduced, the threadlike molecules can be observed directly on the periphery of the fullerene black particles. Figure 4
shows a single tangled cluster of these carbon threads. The fact
that continuous curves of different radii are present both in and
perpendicular to the plane of projection of the microscope explains the “interrupted contrast” (cf. “regular contrast” of the
BSU in Fig. 2). A platelike primary structure can thus be ruled
out. Figure 4 provides clear evidence for the formation of large
(possibly irregularly) crosslinked rings of carbon atoms. which
are required for fullerene formation. It also becomes clear why
fullerene black is considerably more susceptible to oxidation
and chlorination reactions than conventional carbon black.“ *I
Fig. 2. High-resolution TEM image of carbon black (industrial sample FW-1 from
Degussa) The spacing between the planar carbon platelets agrees with that between
the I:iyer\ in graphite (approximately 340 pm). The length of the bar represents
10 nni.
arrangement.[”] Stacking errors caused by platelets being tilted
with respect to the stacking axis and by nonplanar objects (BSU
of differing ring size) result in formation of the bent secondary
structure in the carbon black particles. No in-plane correlation
exists between the BSU (turbostratic arrangement) .[’‘I
The microstructure of fullerene black is far more complex and
depends to a large extent On the reaction conditions’ Under
conditions that give rise to high yields of C,,, the Structures
shown in Figure 3 have been observed. A network of continu-
Fig. 4 Section of a high-resolution TEM image of fullerene black. The reaction
conditions were chosen to maximize the ratio of hlgher fullerenes relauve tc) C,,,:
thls reduced the overall yield markedly The length of the bar represents 5 nm.
COMMUNICATIONS
The growth of the carbon threads is largely suppressed if the
plasma apparatus is operated without a quench gas at a residual
pressure of about
bar [231 In this case. the few remaining
collision partners are no longer inert, and chain termination
reactions by incorporation of d heteroatom from the residual
gas occur very frequently (see also text to Fig. 1 ) This is also
reflected i n the extremely low rate of fullerene formation A
highly resolved TEM image of this product is shown in Figure 5
Color coding is used to depict a section through the projection
plane of the microscope The objects with a red border are short
fragments of carbon threads. Their orientation is irregular, and
a very hard but not very dense form of carbon that is completely
amorphous to X-rays results fZ4l The mechanical hardness of
this material suggests that the units in Figure 5 are covalently
bonded
process, but rather from the increased frequency of side reactions.
Our findings complement those obtained in vapor phase experiments[261and show that in the “caldron” of the plasma
reactor both condensation and crosslinking reactions of macrocycles proceed side by side At the same time. cyclic compounds
are formed and fall apart. depending on the energy content of
the molecule and its immediate environment. The “recipe” for
the synthesis of fullerenes, fullerene black. and planar, layered
structures is determined by the local properties of the vapor
phase outside the plasma These observations and the fundamentally different structures of the condensation products point
to kinetic control of the overall reaction The selectivity of the
condensation of the small primary fragments produced in the
plasma is determined by their concentration, by the efficiency of
energy e x ~ h a n g e , ~and
” ~ by heteroatoms disrupting growth and
rearrangement of the carbon fragments
Received April 2, 1994 [268221E]
German version 4 1 i g e ~ Chrrn 1994, 106. 1919
Fig 5 ArtifiCidl color-coded repreqentdlion o l d highly resolved rection ot d T E M
image of h x d tullerene bldck Under the conditions chosen. the weld ot aoluble
fullerenes w.is less than 0 4 % The length ot the bar represents 2’ nm
In summary, the microstructural analysis of different
fullerene black samples clearly indicates that a significant proportion of the vaporized carbon atoms condenses to form
threadlike molecules These combine to ring systems, which
either form intramolecular links, o r fall apart if their energy
content is too high Intermolecular bonds are then formed when
the density is sufficiently high The proportion of ring growth
and shock annealing determines the yield of soluble fullerenes
The majority of rings are apparently not able to react to form
stable fullerenes, since either the ring size cannot be suitably
adjusted. o r bond formation releases too much energy and dissociation and polymerization results
A portion of the vaporized carbon apparently reacts to form
more o r less planar layers This can be explained by early ring
closure and formation of C, rings (x = 5. 6, 7). The conclusion
that a range of ring sizes are formed follows from the observed
continuous curvature of the resulting layers. This reaction pathway. which proceeds parallel to the linking up of larger monocycles. may be the mechanism for the formation of graphitic
tubules and ontonlike spheres It is unlikely that such structures
are formed by linking of monocycles, a condensation mechanism is much more probable
The detrimental effect of heteroatoins acting as chain-terminnting reagents was also clearly shown and is consistent with
evidence from the synthesis of fullerenes [ 2 5 1 It is known that
linear or cyclic intermediates in the formation of complex carbon species can be trapped and thus detected by addition of
tieleroatoms to the quench gns Our observations indicate that
these products do not arise as a result of disruption of the main
1868
(
CCH V r r l a g ~ ~ r ~ c N s i hmhH
u f t D-6Y451
Weiriliivtn
1994
[I] R . E Smalley, Acc. Chem Rex 1992, 29. 98.
[2] H. W. Kroto. J. R. Heath, S. C . O’Brien. R. F. Curl. R. E. Smalley. N u t u t ~
1985.318, 162; W. Krdtschmer. L. Lamb. K. Fostiropoulos. D. Huffinan. ibid.
1990. 347, 354. G . Peters. M. Jansen. Angcir. (‘hem. 1992. 104, 240; ,4ngw.
Chem. l n t . E d Enpl. 1992. 31. 223.
131 T. Baum. S. Loftler. P. Loffler. P. Weilmiinster. K . H. Homann, Ber. Bunsriigrx P l t w Client 1992. 96. 841.
(41 A. Giigel. M . Becker. D . Hammel. L Mindach, J. Rider. T. Simon. M. Wagner.
K. Miillen. Angrw. Clirm. 1992. 104,666:Angel<. Client. h i t . Ed Engl. 1992. 31.
644: H. Werner. D. Bublak. U. Gobel. 8 . Henschke. W. Bensch. R. Schlogl.
ibid. 1992. 104. 909 and 1992, 31. 868.
[5] S . hJma. Nuturr 1991. 354, 56: N Hamada, S. Sawada. A. Oshiyama. P l 7 y
Rev. Lett. 1992. 68. 2579.
(61 D. Ugarte, Nurure 1992. 359. 707.
[7] This remains m e despite the most recent reports on metal-catalyzed production processes, since a selective synthesis is still not possible. R Dagani. Chem
Ens. N e w 1993. 71, 6.
[El The yield ofsoluble fullerenes is a sensitive and complex function of the helium
vapor pressure (used for quenching), the electrode diameter. and the distance
between the electrodes. which also influences the selectivity of formation of the
various fullerenes: hot plasmas with a small electrode separation give increased
quantities of higher fullerenes at lower overall yields. Graphitic tubules can be
found within the carbon slag formed on the negative electrode used in the arc
discharge uithdirect current. Graphitic onionlike spheres apparently arise only
when fullerene black is intensively bombarded with electrons: however. the
desired structure is not obtained in every batch. even when conditions are
identical. It has not been proved beyond doubt whether graphitic onionlike
spheres can be produced by other thermal processes. since transmission electron microscopy has. up to now. been the only possible means of identification.
[9] R. F. Curl, Philos. Trans R . Soc. London A 1993. 343, 19.
[lo] G . van Helden. N . G . Gotts. M T. Bowers. J: Am Chein. Soc. 1993, 115, 4363.
Nirrirre 1993,363.60; G. van Helden. M.-T. Hsu. P. R. Kemper. M. T. Bowers.
J Chrm. P h w . 1991. 95. 3835.
[ I 11 H. Schwarr. Angeir. Chrm. 1993, 105. 1493; Angen. Chem. Inr. Ed. Engl. 1993.
32. i 4 x
1121 H. Werner. M Wohlers. D. Herein. D. Bublak. J. Blocker. A. Reller. R .
Schlogl. Fulleroir Sci. Tethnol. 1993. 1 . 199: H. Werner, D Herein. J. Rldcker.
B. Henschke. U . Tegtmeyer. T. Schedel-Niedrig. M. Keil. A M. Bradshaw. R.
Schlogl. Chrtn Phj.s Lert 1992. 34. 62.
[I31 Modificd ULVlC BUCKY I 1 apparatus equippcd with graphite rods for spectroscopy (6 mm.3 mm 3p3rt): alternating current of 150 A , distance between
electrodes adjusted to give a constant current.
[l4] Diode array spectrophotometer BECKMANK D U 7000; spectra were recorded with the help o f a quartz optical fiher and a quartz window directly in front
of the center of the electric arc.
1151 E. A Rohlfing. J Chiw. Plim. 1988, 89. 6103. and references therein.
[I61 M. Jeunehomme. R. P. Schwenksr. J Chem. Phrs. 1965. 42. 2406
[17] I n plasmas. a large number ofcharged particles are also formed. The electrical
charge can intluence the formation of fullerenes. and this source of building
blocks could give rise to additional selectivity. However. the fact that fullerenes
are also produced under conditions in which charged species are absent suggests that electrically charged building blocks are helpful hut not essential.
[I81 All samples were investigated supported on amorphous carbon films after
deagplomeration in an ultrasonic bath of acetone. Siemens (Elmiskop 101.
0570-0X33 94/18/X-lX6X X 10 00+ 25 (J
Angi’M C h m In/ Ed En?/ 1994. 33 No 18
COMMUNICATIONS
120 kV. double tilt, side entry), Jeol (CX 200. 200 kV, double tilt, top entry),
and Phillips (CM 20, modified with a cryoprojection lens, ultrahigh vacuum in
the sample chamber, single tilt, side entry) instruments were used.
[19] F. Atamny, H. Kollmann, W. Bartl. R. Schlogl, Ultramicroscop) 1993.48, 281.
[20] A. Oberlin. Chem. Phys. Carbon 1989, 22, 1.
[21] F. Atamny. R. Schlogl, A. Reller, Curhon 1992, 30, 1123.
[22] The layered character of the objects, shown with continuous contrast, also
follow from powder X-ray diffraction data, which show a clear peak corresponding to the interlayer separation of 341 pm and a complex. modulated
backgound [13]. This material can be distinguished from the electrode
graphite. which may contaminate the product, by its much lower combustion
temperature [l?].
[23) Since the apparatus was designed for the manufacture of large amounts of
fullerenes. we operated with a regulated electrode separation and continuous
external adjustment of the graphite rods, which led to a leak rate of the same
order as the residual pressure. This factor could only be eliminated if the whole
apparatus was &xed in a glove box with an inert atmosphere.
[24] Presumably. this material is always formed when the apparatus is in operation
and acts as a “getter” to the vapor-phase contaminants (helium 5.0). Moreover, such particles may well act as nuclei for the condensation of large fullerenes. A. Goeres. E. Sedelmayer, Cheni. P/iy.s. Lett. 1991,154,310; H. W. Kroto,
K. McKay, Nuture 1988,331, 328.
[25] T. Grosser, A. Hirsch, ,4ngew. Chem. 1993, 105, 1390; Angew. Chem. I n r . Ed.
En$ 1993,32, 1340.
126) J. M Hunter. J. L. Fye, E. J. Rosecamp. M. E Jarrold, J. Phys. Chem. 1994,98,
1810.
[27] This 13 determined by the thermal conductivity and partial pressure of the
quench gas. as well as by a steep temperature gradient in the gas chamber,
which was achieved by placing a metal plate around the plasma chamber to
provide efficient cooling of the quench gas.
mains of only a few micrometers in size can be seen with a polarizing microscope. Extension results in necking and hardening of
the material; at the same time, the size of the doubly refracting
domains increases to the millimeter range. The observed effects
are not temperature dependent; that is, the structure remains
unchanged up to decomposition of the material. Figure 1 shows
the photograph under a polarizing microscope of a slightly
stretched sample; Figures 2 and 3 show the wide-angle and
small-angle X-ray diffractograms of the unextended sample.
Fig. 1. Photograph under a polarizing microscope of PAA-C,, at an extension of
i.
= 1.5 nm.
“I
7
Synthesis of Very Highly Ordered Liquid
Crystalline Phases by Complex Formation of
Polyacrylic Acid with Cationic Surfactants
Markus Antonjetti* and J. Conrad
6
5l /r”?%
4
Ionic surfactants and oppositely charged polyelectrolytes
spontaneously form addition complexes that precipitate from
aqueous solution. Neighboring group effects, which begin to
operate at very low concentrations, are important for the formation of such complexes. A highly cooperative zipper mechanism
is observed, and the stoichiometry of the complexes (with respect to charge) is usually I :1 - 3 1
The properties of such complexes in nonaqueous solvents and
in the solid state have, however, only recently been investigated.
The poly(n-alkyltrimethylammonium styrene sulfonate) system
shows highly ordered lamellar structures, which clearly adopts
different phase modifications when the chain length of the alkyl
groups is varied.[41The driving force for the formation of these
phases is the incompatibility between the ionic backbone and
alkyl side chains, otherwise known as the amphotropy of the
system.
We now propose a system in which complex formation occurs
at the polyacrylic acid backbone and in which steric effects
prevent a parallel orientation of the n-alkyl chains and formation of the resulting layered structures. For the first time, yet
more diverse phase structures can be achieved in this polymer
system.
The polyelectrolyte-surfactant complex PAA-C,,, obtained
by simple mixing of polyacrylic acid and dodecyltrimethylammonium chloride, is a highly elastic, deformable material.
Even in an unextended cast film, strongly doubly refracting do-
3 t 1
0
.[’
[*] Prof. Dr. M. Antonietti, J. Conrad
Max-Planck-lnstitut fur Kolloid- und Grenzflichenforschung
Kantstrasse 55. D-14513 Teltow (FRG)
Telefax: Int. code + (3328)46-204
A n x e w Cliwi. Int. Ed. Enxl. 1994. 33, No. 18
tI
0 VCH
10
30
20
50
40
201“
70
60
80
90
__t
Fig. 2. Wide-angle diffractogram of PAA-C,,, 2.
counts per second).
= 0.154 nm.
I = intenslty (in
350
t
t
0
0
0.1
0.2
01
03
0 4
-
0.3 0.4 0.5
s / nm-’
0.6 0.7
05
0.7
06
0.8 0.9
1
Fig. 3. Small-angle diffrdctogram of PAA-C,,, the scattering vector is defined as
s = 2sinB/E.. For clarity at higher orders of diffraction, the logarithmic plot of the
filtered spectrum is included. I = intensity (in counts per second).
Verlugsgrsellschafl mbH. D-69451 Weinheim. 1994
0570-0833/94jlXi8-i869 $10.00+ .ZSIO
1869
Документ
Категория
Без категории
Просмотров
4
Размер файла
2 052 Кб
Теги
mechanism, formation, fullerenes
1/--страниц
Пожаловаться на содержимое документа