Note: Descriptions are shown in the official language in which they were submitted.
Z00~4ZS
A novel isotropically reinforced micro-
composite i9 described. An entirely fluid-phase~method
has been devised for producing a net-shape filamentary
reinforced structure. The process depends for its
success on the ability to generate in situ, within a
shaped mold, a three-dimensional random weave of carbon
filaments by catalytic decomposition of a hydrocarbon
feed. Almost any desired filamentary structure can be
produced by utilizing chemical vapor deposition to
modify the surfaces of the filaments. Infiltration of
filler matrix materials can be achieved by adaptation
of existing materials technologies.
Background of the Invention
Processes for the catalytic production of
carbon from hydrocarbon gases and C~ were proposed and
patented as early as 1920 (U.S. Patents 1,3S2,162;
1,868,921; 1,882,813). These patents identified the
product as "carbon black", but it is clear from the
experimental conditions that filamentary carbon was
formed. The preferred catalysts were iron, cobalt and
nickel or their oxides.
More recent work on filamentous carbon
synthesis has been reported by a number of academic and
industrial organizations throughout the world. Baker
and Harris published a comprehensive review of the
Z~ ;25
field in 1978 ("Chem. and Phys. of Carbon", 14, ~3-165
[19781). Most of the work has centered on the use of
Fe as the hydrocarbon decomposition catalyst, although
many of the ~atent disclosures claim Group VIII metals
in general. The preferred gases are CO, the Cl-C3
alkanes and benzene, but much broader classes of hydro-
carbons are often claimed.
Baker and co-workers have carried out
extensive studies on the catalytic formation of
filamentous carbon, by decomposition of acetylene at
temperatures between approximately 500 to 975C in the
presence of Fe, Co and Cr catalysts supported on single
crystals of graphite and silicon ("J. Catal." 30(1),
86-95 [1973]), or over nickel films ("J. Catal." 26(1),
51-62 [1972~). Each of the filaments was observed to
have a catalyst particle at its growing tip, where the
diameter of the filaments was fixed by that of the
catalyst particle. The filaments' diameter and length
varied respectively between 0.01-0.15 microns and
0.5-8.0 microns. Filament growth followed rando~ paths
forming loops, spirals and other shapes. Growth rate
varied inversely with catalyst particle size. The
filaments stopped growing when the catalyst particle
was completely covered with a carbon layer. Baker also
studied the formation of carbon filaments from other
hydrocarbon gases such as ethylene, benzene,
1,3-butadiene, allene and propyne ("Carbon", 13(3),
245-6 [1975]).
~ .S. Patent 4,5fi5,683 (D.J.C. Yates and R. T.
Baker) discloses FeO as a catalyst for carbon filament
synthesis. The FeO, formed by steam treatment of Fe at
700C, is reacted with acetylene or ethane at 700C.
Z~O`~Z5
U.S. Patent 3,816,609 discloses a process for
the production of a hydrogen-rich stream from a hydro-
carbon feed gas such as propane. The hydrocarbon feed
is first converted to filamentary carbon using a
supported Group VIII non-noble metal catalyst. The
carbon is then gasified using steam to produce the
hydrogen-rich gas stream.
U.S. Patents 4,435,376 and 4,~18,575 are
directed to the synthesis of filamentary carbon from
hydrocarbons and a (Ni,Ti)-based catalyst which has
been promoted with phosphorus. The addition of phos-
phorus is claimed to result in filaments of decreased
diameter and length and increased surface area, such
that the "microfibrous carbon" is a good candidate as a
reinforcing agent.
Department of Energy Report No.
DOE/MC/14400-1551, described a process for making
filamentary carbon by the catalytic reduction of a
carbon-containing gas using iron as the catalyst. In
one preferred embodiment of the process, carbon is
deposited on an iron-based catalyst from a CO/hydrogen
ga~ mixture in the 300-700C temperature range at a
pressure of 1-100 atmospheres. The carbon produced is
called "ferrous carbon" and is described as fibrous,
particulate material in which the metal catalyst
particles are intimately dispersed as nodules through-
out the fibrous carbon growth.
Royama and Endo have developed a process for
growing graphitic fibers at about 1000C in which a
gaseous mixture of benzene and hydrogen is passed
20~
-- 4
through a reaction pipe coated with very fine particles
of Fe (Japan Economic Journal, 17 [December 1981]). The
fibers are reported to grow in a two-stage process (J.
Crystal Growth 32(3), 335-349 [1976]). The growth
process begins with the catalytic formation of very
thin filaments which are then thickened by the
pyrolytic deposition of carbon. The carbon fibers are
typically 10 microns in diameter and several cm long. A
1982 Showa Denko K.K. patent ~Japanese Kohai 57/117622)
discloses that carbon filaments may be prepared by
carbonizing a gaseous mixture containing benzene and
hydrogen at 1000C in the presence of Fe particles with
a particle size less than 0.03 microns or the use of a
suspension of Fe particles sprayed into a reaction
chamber at 1000C with a flowing mixture of henzene and
hydrogen (Japanese Kohai 58/1180615).
G. G. Tibbetts and co-workers at General
Motors developed processes for the growth of carbon
filaments using methane or natural gas as the
hydrocarbon gas at about 1000C. Catalyst particles
are obtained ~rom carburized stainless steel tubes
(U.S. Patent 4,391,787) or by wetting the inside of
steel tubes with an aqueous ferric nitrate solution
("Carbon" 23(4), 423-430 ~1985~), or by growing a thick
layer of oxide on the inside of the tube (U.S. Patent
4,497,788). Also disclosed is a process for growing
graphite fibers on a ceramic substrate pretreated by
evaporating a ferric nitrate solution to deposit an
iron compound (U.S. Patent 4,565,684). In the first of
two carbon-growth stages, a mixture of 5-15 vol~
methane and hydrogen is passed over the ceramic heated
to between 600-1200C (preferentially 1000-1100C).
nuring this stage catalytic growth of thin carbon
` 200~4LZS
filaments occurs. The second growth stage is then
initiated by increasing the methane concentration in
the gas to 25 vol% of higher. This results in the
thickening of the filaments due to pyrolytic deposition
of carbon into fibers with diameters between 5-15
microns and 1-3 cm long.
In 1974 Nishiyama and Tamai ("J. Catal.",
33~1), 98-107 [1974~) reported the formation of fibrous
carbon on Ni/Cu alloy sheets and alloy powders from the
decomposition of benzene in the 580-900C temperature
range. For both the sheet and powder cases, a large
number of metallic particles were present in the carbon
possessing the same composition as the substrate. Por
the catalysts in both the sheet and powder form, the
deposition rate was higher for the alloys containing
40-80% Ni than for pure Ni. In some follow-up work in
1976 the authors reported on the beneficial effect of
addlng hydrogen to the benzene stream on the rate of
formation of the fibroùs carbon under certain
conditlons ~"J. Catal." 45(1), 1-511976]).
In 1985, Bernardo et al ("J. Catal." 96(2),
517-53~ 11985]) studied the deposition of carbon on
sllica supported Ni/Cu catalysts erom a methane-steam
mixture at 500-900C. The carbon deposits from alloys
with 50-100% Ni were filaments with a less dense core
and a metallic particle at the tip.
Neither Nishiyama and Tamal, nor Bernardo et
al. diccovered the surprlsing space-filling capabllity
of filamentary carbon growth from Cu/Ni catalysts when
ethane or ethylene are used as the hydrocarbon feed gas
200~5
in the temperature range 500-700C. Neither did these
workers report the primarily bi-directional, and at
times multi-directional, growth pattern characterizing
the process of this invention.
srief Summary of the Invention
The present invention is a versatile process
for making net-shape microcomposite structures which
overcomes the limitations of today's composite
technology. The process relies on the catalytic growth
of carbon filaments at temperatures typically less than
about 1000C from gas-phase precursors. ~n example is
the catalytic growth of thin filaments of carbon from
gaseous hydrocarbons, e.g., ethane at 700C. The
specific process of this invention involves rapid
catalytic growth of carbon filaments which eventually
expand to fill the available space in a shaped mold.
Furthermore, it is in the nature of the growth process
that the filaments intertwine to form a three-
dimencional random weave (self-woven network), which
has some structural integrity as a free-standing form.
~y appropriate choice of catalyst particles, filaments
as small as 0.01 micron in diameter can be produced.
The carbon networks can be further modified
with one or more of several surface treatment
techniques, e.g., chemical vapor deposition, electro-
depocition, electro-less deposition, to tailor the
structure and properties of the filamentary network. An
example is the chemical vapor deposition of pyrolytic
graphite on the original carbon filaments for improved
strength, and subsequent electrodeposition of nickel to
promote wetting to an aluminum matrix.
200~25
-- 7 --
Thus, it is clear that this invention, as
will be described below, overcomes the limitations of
existing composite technology. The new capabilities
provide the opportunity to produce composite structures
directly to net shape with minimum handling and with an
isotropic reinforcement in thin section.
Detailed Description of the Invention
The reinforcing elements for the synthesis of
the microcomposites of this invention are carbon
filaments grown catalytically at elevated temperatures
using hydrocarbon gases and a metal alloy catalyst. The
carbon filaments range in diameter from 0.01 micron to
about 2 microns, may be several hundred microns long
and are grown in a randomly intertwined network, see
~ig. 1, at a volume density ranging from about 2% to
about 20%.
The carbon filaments are further
characterized in that they predominantly show a
bidirectional, see Fig. 2, and at times a multi-
directional, see Fig. 3, growth mode. Thus, more than
one carbon filament grows from a single metal catalyst
particle. Further, there is generally a one-to-one
correspondence between the diameter of the catalyst
particle and the diameter of the carbon filament~
The filaments may be modified by coating them
with another material ~sing chemical vapor deposition.
200~2S
Brief Description of the Figures
Figure 1 is a photomicrograph showing a
randomly interwoven network of carbon filaments. The
bright dots in the filaments represent catalyst
particles.
Figure 2 is a photomicrograph more
particularly showing the catalyst particles in the
filaments. The diameters of the filaments are clearly
a function of the diameter of the catalyst particles
and bi-directional growth of carbon filaments from
single catalyst particles is evident.
Figure 3 is a photomicrograph showing an
example of multi-directional growth from a single
catalyst particle. The multi-directional growth is
particularly evident from the catalyst particles A, B,
C and D.
Fiqure 4 is a photomicrograph showing a
filamentary carbon network that has been coated with a
ceramic material. Figure 5 shows the same coated
network at a higher magnification.
Figure 6 shows a cross-sectioned view of a
mold for forming microcomposites.
Figure 7 shows in cross-section, a reactor
and mold for forming a cylindrical, hollow net-shape
microcomposite. 1 is a 2.5 centimeter (cm) quartz
tube, 2 is glass frits, 3 is a 1.25 cm graphite rod and
4 is mold cavity defined by the wall of the quartz tube
and the glass frits.
2~0~4L25
The invention is illustrated diagramatically
in Figures 6 and 7. In the first step of the process,
the filamentary carbon network is formed in-situ within
a mold by catalytic decomposition of the hydrocarbon
feed, utilizing metal alloy catalyst particles that are
applied to the walls of the mold. The seeded mold is
placed into a furnace or heated externally to the
desired temperature and a hydrocarbon gas is allowed to
flow through the mold. The filamentary network is
permitted to grow until the mold cavity is filled with
the desired volume fraction of filaments. The natural
tendency of the random weave of filaments uniformly to
fill the available space in the mold ensures faithful
replication of the internal surface features of the
mold, as well as isotropic reinforcement in the final
composite. The filaments may then be modified by
chemical vapor deposition ~CVD) for example.
In the second step of the process, the matrix
material is introduced as a liquid, and allowed to fill
up all available space between the filanents in the
mold, thereby forming the shaped composite structure.
The infiltrant ~matrix) is a polymeric material, but it
could also be a metallic or ceramic material. The
resulting shaped composite part, manufactured without
any traditional handling, is removed from the mold, and
the mold is recycled.
Thin section composites may be synthesized by
this process by choosing catalyst particles that are
significantly smaller than about 1 micron. Since there
is a one-to-one correspondence between the diameter of
200~a~25
-- 10 --
a filament and the dimensions of the catalyst particle,
clearly a thin section, say less than 1 millimeter
thick, may be filled with a random weave structure of
ultra-fine carbon filaments without any handling.
The catalytically grown carbon filaments may
be modified by coating them with another material using
chemical vapor deposition. The CVD coating is applied
by passing an appropriate volatile precursor through
the filamentary carbon network inside the mold at an
appropriate temperature as is well known to those
skilled in the art of chemical vapor deposition and
infiltration. ~ny desired thickness of a shape con-
forming coating (deposit) of a ceramic, metal or carbon
may be applied to the network of carbon filaments.
Such flexibility in filament processing is
particularly advantageous, because it opens up new
possibilities for designing filaments with specific
bulk or ~urface properties. In many metal matrix
composite systems, resistance of the filaments to
dissolution in the melt is an essential requirement.
For example in the fabrication of single crystal
composite turbine blades, utilizing investment cast
ceramic shell molds, only a few CVD-coated filaments,
e.g., A1203, ZrO2, HfO2-coated filaments, would resist
dissolution in the melt. On the other hand, in lower
melting point alloy systems, the most challenging
problem is to achieve good wetting between matrix and
filament to ensure proper melt infiltration and
composite strengthening. Coating of the carbon
filaments with a thin layer of nickel by CVD is one way
to ensure good wetting with aluminum alloys, for
example.
20~
-- 11 --
An intriguing aspect of chemical vapor
deposition is the deliberate construction of artificial
nano-scale multi-layers on the original filaments. When
the scale is sufficiently fine enough, it should be
possible to exploit the well-known super-modulus effect
for achieving exceptional stiffness of the composite
filament. Controlling interfacial bond-strength
between layers in a multi-layer structure is another
method of improving fracture toughness.
The matrix material (infiltrant) may be a
polymer, elastomer, metal, alloy or a ceramic and is
used in a liquid state during infiltration of the
filamentary network. Polymer infiltration may also be
achieved by an in-situ process where the monomer is
allowed to polymerize inside the mold.
Although it is well known that several
transition metals, primarily Co, Ni and Fe, will act as
catalysts to convert hydrocarbon gases and C0 to
filamentary carbon, no catalysts have been identified
that will produce a rapid, voluminous growth that tends
to fill available space. Our discovery that certain
alloy systems will produce such a growth now makes
possible the synthesis of net-shape isotropically re-
inforced microcomposites.
We have discovered that two classes of metal
alloy systems yield such space-filling growths. ~oth
classes are distinguished by metal combinations that
form a series of solid solutions over their whole com-
position range and are a combination of a transition
X0~ 25
- 12 -
metal which is known to be a relatively good catalyst
for filamentous carbon growth and one which shows no
catalytic activity whatever, or one which is a poor
catalyst.
The preferred alloy system of the first type
is based on the system Ni/Cu, an example of a Group
VIII metal, Ni, that is known to be a catalyst for
filamentous carbon formation, and a Group Is metal, Cu,
that is not a catalyst for filamentous carbon
formation. Ni and Cu form a series of solid solutions
over their whole composition range. The preferred
composition range for the purposes of this invention is
from about 20 wt% Ni to about 90 wt~ Ni, with the range
from about 40 wt% Ni to about B0 wt% most preferred.
When Ni is used as a catalyst for the
formation of filamentous carbon from ethane or
ethylene, the filaments tend to be relatively short
with an aspect ratio of about less than 10 and the Ni
catalyst particle i8 found at the tip of the filament.
The Ni tend~ to become deactivated relatively quickly,
most likely by being coated with a layer of carbon
restricting further access of the hydrocarbon
molecules. This results in a relatively inefficient
process and relatively small yields of carbon per gram
of catalyst and very little intergrowth of the
filaments. Unexpectedly when Cu is added to the Ni,
the alloy catalyst forms very long filaments with
aspect ratios generally greater than about 200. The
actual aspect ratio is not known definitively because
it is difficult to find the ends of any one particular
filament in the intertwined network.
200~5
-- 13 --
Filaments grown from Ni/Cu alloys are further
distinguished ~rom those grown from Ni by the location
of the catalyst particle within the filament. Rather
than being at the tip of the filament, the catalyst
particle is predominantly found at the midpoint of two
filaments. Thus a single catalyst particle is found to
grow two carbon filaments in opposite directions. The
two filaments are believed to grow simultaneously. We
have also observed instances of a single Ni/Cu particle
growing as many as six carbon filaments, suggestive of
the six faces of a cube, all active for the growth of a
carbon filament. This extraordinary multi-directional
growth is believed to be responsible for the remarkably
rapid and space-filling tendency of filamentary carbon
growths from the Ni/Cu systems. Although previous
workers have reported that the addition of Cu to Ni
increases the latters activity for filamentary carbon
growth, the predominantly bi-directional growth and the
space-filling ability of this catalyst system when used
with ethane or ethylene as the hydrocarbon gas was not
recognized~ Previous worker~ used benzene and methane
as the hydrocarbon gases.
We have also discovered that other com-
binations of Group VIII-IB solid solution metal alloys
such as Ni/Au, Co/Au, Fe/Au and Co/Cu also show
significantly enhanced catalytic activity over the pure
Group VIII metal and that the alloy systems produce
predominantly bi-directional carbon filament growths.
As in the case of the Ni/Cu system, these alloys are
combinations of known catalysts for filamentous carbon
formation, namely Ni, Fe and Co, and a non-catalyst,
namely Cu and Au.
200~ 5
A preferred example of the second class of
metal alloy catalyst is the Ni/Pd system. Ni and Pd
also form a series of solid solutions over their whole
composition range. Whereas Ni is a relatively active
catalyst for filamentary carbon growth, Pd is a
relatively poor catalyst. When an alloy of Ni/Pd of
about 50/50 wt% is used with ethane or ethylene, an
almost "explosive" type of filamentary carbon growth is
obtained. Like the Group VIII-IB combinations
described above, the carbon growth tends to fill all
available space, but at a much more ra~id rate, in
minutes rather than in tens of minutes. The carbon
filaments are also found to be predominantly
bi-directional, i.e., at least two carbon filaments
grow from a single catalyst particle. Because of the
very rapid space-filling growth with this catalyst
system, the carbon growth tends to be of a very low
volume density, typically about 2 vol ~ rather than the
more common 5 to about 10 vol % within the Ni/Cu
system. Another novel and distinguishing feature of
this type of growth is that it tends to be "sponge-
like" with some resiliency reminding one of
sponge-rubber. Such a filamentary carbon growth has
not been reported heretofore. The reason for this
latter property is not understood.
The unexpected and newly discovered
beneficial nature of the addition of Pd to Ni has also
been observed with another Group VIII base metal/Pd
combination which forms a solid solution, namely Co/Pd.
200~5
In this case also, the alloy system shows significantly
higher catalytic activity for carbon growth than the
non-noble Group VIII metal, the carbon filaments are
oredominantly bi-directional and the carbon growth has
some resiliency and tends to fill available space.
The preferred form of the catalyst is a fine
powder, although bulk forms such as rolled sheet or
thin films may also be employed. There appears to be
an approximate one-to-one correspondence between the
size of the catalyst particle and the diameter of the
carbon filament and generally carbon filaments with
diameters less than about 1 micron are preferred for
the purposes of this invention. Surprisingly even when
bulk material or powder significantly larger than 1
micron, up to 40 microns for example, is used, the
alloy disintegrates during the filamentary carbon
growth process such that particles ranging in size from
about 0.01 micron to about 2 microns are generated.
The process responsible for this advantageous dis-
integration is believed to be similar to one known as
"metal dusting".
For some applications it may be preferred to
manufacture approximately monodisperse carbon fila-
ments, i.e., filaments with substantially the same
diameter. In this case it is clearly advantageous to
start with alloy catalyst particles that are all sub-
stantially of the same diameter and of a size sub-
stantially equal to the desired diameter of the carbon
filaments. Such alloy powders could be produced by
aerosol production from the melt or by thermal
evaporation at relatively high pressures such that a
metal "smoke" is generated, or by pyrolysis of a
volatile organo-metallic precursor.
20~ 25
- 16 -
Alloy particles may also be synthesized
in-situ in a mold if desired. Such well-known
techniques as evaporating an aqueous solution of the
salts of Ni and Cu, e.g. nitrates, chlorides, etc.,
followed by calcining to the metal oxides and then
reducing the oxides to the metal alloy by heating in
hydrogen, may also be employed. The concentration of
the metal salts is adjusted such that the desired alloy
composition is obtained. The walls of a mold could
thus be seeded with the desired metal alloy catalyst by
wetting the walls with the starting aqueous solution
and forming the metal alloy particles in-situ. Another
technique that may be employed is to deposit films of
the constituents metals onto the walls of the mold by
electroplating or electro-less plating, for example,
and then heating the deposited films to form the alloy
by interdiffusion. When the hydrocarbon gas is intro-
duced into the mold, the film will disintegrate during
the filamenta~ry carbon growth process.
Minor impurities in the metal catalyst do not
appear to have significant effects. Thus, Monel powder
of nominal 70 wt% Ni and 30 wt% Cu compositions with
less than about 1 wt% each of Mn and Fe as impurities
has been found to be as useful as nominally pure Ni/Cu
alloys. Ni/Cu sheet of nominal 55 wt% Ni and 45 wt~ Cu
composition with minor amounts of Mn and Fe has also
been found to be an effective catalyst for filamentary
carbon growth.
It has been found that one gram o Ni/Cu
alloy catalyst can generate at least 100 gm of
filamentary carbon before the catalyst particles become
Zoc~ 5
inactive. The actual catalyst loading of a mold can
therefore be adjusted such that the desired volume
density is achieved. The metal alloy particles can at
least in part be leached out of the carbon filamentary
network with acids if so desired.
Hydrocarbon gases may be converted to a form
of carbon by thermal pyrolysis alone. This carbon may
be either in particulate form, commonly known as soot,
formed by gas phase nucleation and practiced in the
manufacture of carbon black, or in thin film, pyrolytic
form when hydrocarbons are decomposed at very high
temperatures, generally above about 900C. The latter
is essentially an example of a process more commonly
known as chemical vapor deposition.
Catalytic filamentary carbon growth requires
contact between a metal catalyst particle at elevated
temperatures and a carbon-bearing gas. Although a
number of gaqes such as CO and various hydrocarbon
gases have been used in the past, the preferred gases
for the purposes of the present invention are ethylene
and ethane. Formation of soot or pyrolytic carbon
reduces the overall efficiency of the process and may
interfere as well with the catalytic activity of the
metal catalyst particles, and is therefore avoided in
the practice of this invention.
It has been proposed by Baker that carbon
source gases that undergo an exothermic decomposition
reaction to elemental carbon are required for
filamentary carbon growth. Thus gases such as CO,
acetylene, ethylene and butadiene readily form
20~
- 18 -
filamentous carbon, whereas gases such as methane
should not. Filamentous carbon growth from methane is
believed to require the thermal conversion of the
methane to less stable molecules prior to catalytic
conversion to carbon.
For the purposes of this invention, the
hydrocarbon gases are chosen such that they form
insignificant amounts of soot or pyrolytic carbon under
the processing conditions, i.e., they will decompose to
carbon only in the presence of the metal alloy
catalyst. Ethylene is preferred in the temperature
range 550 to 650C, while ethane is preferred in the
temperature range 650 to 750C. It is clear, however,
that if one wishes to coat the already formed carbon
filaments with a layer of pyrolytic carbon, the
temperature may be raised or other less thermally
stable hydrocarbons may be used.
Ethane and ethylene, the preferred hydro-
carbon source gases of the present invention, are
readily available and relatively inexpensive. Ethane
requires somewhat higher temperatures for filamentous
carbon growth than does ethylene. The reason for this
is believed to be that the ethane first needs to be
converted to ethylene by pyrolysis before the metal
alloy catalyst can form filamentous carbon.
Benzene is relatively e%pensive and
carcinogenic, while CO is also highly toxic. Methane
is found to require very high reaction temperatures,
above about 900C, for any filamentary carbon growth to
20C)~Dt25
-- 19 --
occur with the alloy catalysts of the present
invention. Under these conditions, pyrolytic carbon
deposition also takes place which tends to coat the
catalyst particles with a layer of carbon and render
them inactive relatively quickly.
Acetylene will form filamentary carbon with
Ni/Cu alloy catalysts at te~peratures as low as 300C.
Pure acetylene, however, has a tendency to readily form
soot due to pyrolysis at temperatures as low as about
450C.
The purity of the feed gases does not appear
to be a critical factor. So-called "chemically pure"
grades, aproximately 98~ purity, of ethane and ethylene
have been used successfully without further
purification. Although it is likely that some
impurities such as hydrogen sulfide may poison the
alloy catalysts, this is not known at the present time.
Although nominally pure ethane and ethylene are
preferred, mixtures of these two gases, as well as
mlxtures with inert gases such as nitrogen or argon may
also be employed. It has also been found that the
deliberate addition of hydrogen in the range from about
10 vol% to about 90 vol% prolongs the time that a Ni/Cu
catalyst remains active. We speculate that the
hydrogen keeps the catalyst particles relatively free
of deposited carbon films.
r~nder the preferred conditions, as much as 80
mole% of the ethylene or ethane being fed to the
reactor has been converted to carbon in the form of
carbon filaments. The by-products of this
2C~0~2~
- 20 -
decomposition have been analyzed by gas chromatography
and found to be primarily hydrogen, with some methane,
as well as smaller amounts of simple saturated
hydrocarbon molecules such as propane and butane.
~ ny unreacted feed gas may be partially
recycled, i.e., mixed with pure feed gas to make the
process even more efficient. The effluent gas stream
from the reactor may also be used as the source of
hydrogen if that is desired during the filamentary
growth.
The temperature range available for the
practice of this invention with the preferred catalyst
systems, i.e., Ni/Cu or Ni/Pd, is from about 300C to
about 800C and is determined primarily by the hydro-
carbon gas used. At lower temperatures the rates of
carbon growth are not sufficiently rapid, while at
higher temperatures, the catalyst particles tend to
become coated with a carbon coating rendering them
~nactive.
It has been shown that, for Ni/Cu and
acetylene, temperatures as low as 300C may be used.
For ethylene the temperature may range from about 500C
to 650C. For ethane, the temperature may range from
about 650C to about 800C. In the latter case it is
believed that the ethane needs to be pyrolyzed or
thermally converted to ethylene before it can be
catalytically converted to filamentary carbon. Thus,
one can envision preheating the ethane to the required
temperature and then letting it contact the catalyst at
a lower temperature.
20C)~2S
Although this invention has been practiced
only in an isothermal mode, there may advantages to
growing the filaments at different temperatures in
order to control how quickly and uniformly the
filamentary network fills up a mold. High temperature
heat treatments as high as 2500C and above are also
contemplated in order to graphitize the carbon
filaments. The temperature used for coating the
filaments by chemical vapor deposition, if practiced,
is chosen to fit the precursor and rate of deposition
desired, as is well known in the art.
Flow rates are chosen to optimize the growth
rate of the carbon filaments and are better defined in
terms of residence or contact time. A typical contact
time is of order of 20 sec, although shorter as well as
longer contact times have been used successfully. The
actual flow rate used will depend on the volume of the
mold (or re~actor) and the processing temperature and is
adjusted to achieve the de~ired residence time. Very
short residence times result in relatively inefficient
use of the hydrocarbon gas.
~ lthough the invention has been practiced
only at ambient atmospheric pressure, it is well known
that catalytic filamentary carbon can be synthesized at
pressures below and above atmospheric pressure as well.
Chemical vapor deposition is generally practiced at
atmospheric pressure, or subatmospheric pressure.
~ lthough the present invention addres~es the
synthesis of composite structures, it is clear that the
as-grown filamentary networks, with or without the
2~)0~4L;;:5
benefit of chemical vapor deposition surface
modification, and, or selective oxidation, have utility
in their own right. Carbon filaments or chemical vapor
deposition coated carbon filaments may be employed as
porous structures and materials, for example as
catalyst sup~orts, filtration media, and thermal
insulators.
The carbon filaments are electrically con-
ductive and thus may also find utility as high surface
area battery electrodes or in electrically conducting
membranes.
The catalysts disclosed in this invention are
highly efficient in converting certain hydrocarbons to
carbon, with the major byproduct being hydrogen gas.
This same technology, therefore, at least as practiced
in the synthesis of carbon filaments, can also be con-
sidered as a means of converting hydrocarbons to
hydrogen, valuable hoth as a fuel as well as chemical.
~xample 1
Two pieces of metal foil, one made of Ni, the
other of Ni/Cu alloy were placed in a 2.5 centimeter
diameter quartz reactor inside a 90 cm long furnace and
heated to 700C under ~rgon flowing at 200 cc/min. The
nominal composition of the alloy was about 45 wt.% Ni,
55 wt.~ Cu and less than 1 wt.~ of Fe and Mn. The
growth of filamentary carbon could be observed through
an optical window at the exit of the quartz reactor. At
temperature, the ~rgon was replaced by flowing ethane
at 100 cc/min. After 2 hours, the ethane was purged
with Argon and the reactor cooled down. The Ni foil
X0C~ 5
- 23 -
had an approximately 2 mm thick black growth of
filamentary carbon on it and the weight increase of the
sample gave a growth rate of - 20 mg/hr per cm2 of
geometric surface area. The Ni/Cu foil in contrast had
grown a 13 mm thick dense layer of carbon which
extended to the walls of the quartz tube. The weight
increase of the sample gave a growth rate of
approximately 200 mg/hr per cm2 of geometric surface
area. The Ni/Cu foil was embedded in the center of the
filamentary carbon growth and mechanically intact.
Figure 2 shows the bidirectional carbon filaments grown
from the Ni/Cu foil in an SEM (Scanning Electron
Microscope) photomicrograph. This example demonstrates
the greater catalytic activity of the Ni/Cu alloy over
pure Ni above and the space-filling tendency of the
carbon filaments from the Ni/Cu alloy. The catalyst
particles in the filaments were examined by x-ray
analysis and found to contain both Cu and Ni.
Example 2
Example 1 was repeated but no metal catalyst
was placed in the quartz reactor. After 2 hours, the
quartz tube was inspected and found to contain no
carbon deposits showing that under these conditions, no
pyrolysis takes place and the filamentary carbon is a
catalytic product.
Example 3
Example 1 was repeated but the quartz reactor
was loaded with a piece of high purity Cu foil. No
filamentary carbon growth occurred and the Cu foil
gained no weight, showing that Cu is not a catalyst for
filamentary carbon growth.
20~ 5
- 24 -
Example 4
Example 1 was repeated except that in ad-
~ition to the Ni foil and Ni/Cu foil, a third sample of
~i/Cu powder (Cerac; 70/30 wt.~) was also included. The
Ni/Cu powder grew a plug of filamentary carbon which
filled the diameter of the quartz reactor and was about
15 cm in length. The initial and final weights were
Ni: 0.2349 g -~ 0.2357 gm; Ni/Cu foil: 0.3996 -~ 1.0749
gm; Ni/Cu powder: 0.409 gm --~ 6.8870 gm.
Example 5
A commercial Ni/Pd based melt-spun alloy
(Allied Chemical) was exposed to ethylene at 600C in a
quartz reactor as in Example 7. After an induction
period of about 2 minutes, filamentary carbon growth
started and filled the full diameter of the quartz
reactor in approximately 5 minutes. No further growth
appeared to occur after approximately 15 minutes. The
resulting filamentary growth had a sponge-like quality,
less dense than growths for Ni/Cu powders or Ni/Cu
foil. Scanning Electron Microscope (SEM) examination
of the carbon filaments showed mainly bidirectional
growth, similar to ~igure 2 of Example 1, from single
catalyst particles. The latter contained Pd and Ni by
x-ray analysis.
Example 6
A 13 mm diameter Cu pipe was coated with Ni
using commercial electroless Ni and placed in a clean
quartz reactor. The conditions of Example 1 were
20~ 5
repeated except that 10% H2 in Argon at 200 cc/min. was
run for 1 hour, followed by ethane at 100 cc/min. for 4
hours. At the end of the experiment, the Cu pipe was
completely filled Wit~l a growth of filamentary carbon.
x-ray analysis of the catalyst particles in the carbon
filaments showed Cu, Ni and some P. The P is known to
be present in the initial electroless Ni deposit. It
is believed that the initial heat treatment, for 1 hour
in this case, formed a Ni/Cu alloy surface layer on the
internal of the Cu tube which provided the Ni/Cu
catalyst particles. The P does not appear to impede
the catalytic process. This example also illustrates
the space-filling ability of the filamentary carbon
grown for Ni/Cu alloys. No carbon growth occurred on
the external surface of the Cu tube since Cu is
non-catalytic (as shown in Example 3).
Example 7
A quartz substrate was coated by partially
overlapping 0.3 micron thick films of Cu and Ni
produced by sputter deposition. The coated quartz
slide was placed in a quartz reactor as in Example 1
and heated to 600C in a 10% H2/Ar mixture and kept at
600C for 1 hour. The gas was then changed to ethylene
at 100 cc/min and the experiment continued for 1 hour.
The part of the slide covered by Cu remained shiny with
no carbon growth; the Ni region was visibly dark with a
carbon growth lesc than 1 mm thick whereas the
overlapping Ni/Cu region had a growth of carbon
approximately 8 mm in thickness. The heat treatment
before exposure to the hydrocarbon gas caused the Cu
and Ni films to interdiffuse and form a Ni/Cu alloy.
The experiment demonstrates, as Examples 1, 2 and 3
that Cu is non-catalytic, and that the addition of Cu
to Ni in alloy form greatly enhances the latter's
activity for filamentary carbon formation.
Example 8
Example 7 was repeated except that the Cu
film was replaced by a sputtered film of Au. The
results were similar showing that Au is non-catalytic
and that a Ni/Au alloy is more active than Ni alone.
~y x-ray analysis, the catalyst particles in the carbon
filaments contained both Ni and Au.
Example 9
Example 7 was repeated except that the Cu was
replaced by a sputtered film of Ag. The Ag area
remained shiny showing that Ag, like Cu and Au, is
non-catalytic for carbon formation. The overlapping
Ni/Ag region showed areas where more carbon filaments
grew than in the Ni region. The interpretation is that
Ni/Ag solid solutions are more difficult to form than
solid solutions of Ni/Cu or Ni/~u, but that Ag can also
enhance the catalytic activity of Ni for filamentary
carbon growth.
Example 10
Example 7 was repeated except that the Ni
film was replaced by Pd and the Cu film by Au. After
exposure to ethylene at 600C, the Pd region showed
only a slight blackening, the Au region remained shiny
whereas the Pd/Au region showed an 0.7 mm thick growth
ZO50~ 5
of filamentary carbon. The experiment demonstrates
that the addition of Au to Pd, which form a solid
solution, greatly enhances Pd's activity for
filamentary carbon formation.
Example 11
Example 10 was repeated except that the Pd
was replaced by a Co film. The Co/Au region showed a
filamentary carbon growth approximately twice as thick
as that of the Co region, showing that Au also enhances
the catalytic activity of Co for filamentary carbon
formation.
Example 12
Example 7 was repeated except that the Cu
film was replaced by a sputtered Pd film. The Ni
region grew a filamentary carbon growth less than 1 mm
thick, whereas the Ni/Pd region had a growth of
filamentary carbon which was more than 10 mm thick in
some areas. The Pd region showed only a spotty growth
of carbon. This is an example of two Group VIII metals
which form a solid solution where the alloys are
catalytically more active than either metal alone.
Example 13
Example 12 was repeated except that the Ni
was replaced by Co. The Co and Pd regions showed only
partial blackening due to filamentary carbon growth,
whereas the Co/Pd region showed a growth approximately
1 mm thick. This is another example of a Group
VIII-VIII combination which is more active than either
metal alone.
20~ 5
- 28 -
Example 14
Example 1 was repeated except that the
reactor was loaded with Ni/Cu powder and the ethane
replaced by methane. At 700C no obvious carbon growth
occurred. The temperature was increased in 50C steps
and filamentary carbon growth was observed to begin at
a temperature of approximately 900C. Growth ceased
after about 10 minutes and did not fill the diameter of
the reactor. The walls of the quartz reactor showed a
black carbonaceous deposit at the end of the run. The
experiment demonstrates that CH4 as a source gas
requires temperatures which also induce pyrolytic
deposition of carbon. This pyrolytic carbon quickly
poisons the catalyst particles and impedes catalytic
growth of carbon filaments.
Example 15
A filamentary carbon growth from Ni/Cu powder
catalyst as in Example l from ethane at 700C was
further treated by coating it with SiCXNy chemical
vapor deposition (CVD) coating derived from the
pyrolysis of Hexamethyldisilazane (HMDS). After the
growth of the filamentary carbon, the ethane was
replaced by Argon saturated with HMDS vapor by bubbling
the Argon through a reservoir of liquid HMDS. The CVD
infiltration was carried out for 3 hours. Scanning
electron microscope (SEM) and X-ray examination of the
carbon filaments after this HMDS exposure showed that
the carbon filaments were coated with a conformal Si
2b~ 5
- 29 -
containing layer. The experiment demonstrates the
in-situ production of Cv~ modified filamentary carbon
growths which could then be used as reinforcing agents
for bulk composites.
Example 16
Example 15 was repeated except that the HMDS
was replaced by Tetraethoxysilane and the temperature
was lowered to 550C during the CVD infiltration. This
precursor is known to deposit SiO2 coatings. SEM
examination verified that the carbon filamentary growth
was conformally coated with a Si containing layer,
believed to be Si2
Example 17
Example 15 was repeated except that liquid
HMDS at a rate of 1 ml/min was injected into the
reactor after raising the temperature to 900C. The
higher temperature was chosen to achieve high
deposition rates. ~fter a 30 minute exposure, the
sample was cooled and removed from the reactor. The
filamentary carbon growth now had a ceramic-like
appearance and after fractioning was examined in a SEM.
Figures 5 and 6 show photomicrographs of such a
fracture surface. The 0.2 micron filamentary carbon is
visible extending from the conformal SiCXNy coating.
The composite growth is not fully dense but bridging of
the filamentary growth is clearly evident.
s
- 30 -
Example 18
A filamentary carbon growth from a Ni/Cu foil
experiment as in Example 1 was placed into a mold press
and impregnated with a low viscosity epoxy resin (E. F.
Fuller, Inc.). The mold was heated to 50C under
pressure and allowed to cool overnight. The resulting
filamentary carbon reinforced epoxy composite was
sectioned and examined in an SEM~ The filamentary
carbon growth was found to be completely and uniformly
infiltrated by the epoxy resin.
Example 19
A mold, shown diagrammatically in Figure 7,
consisting of a 1.2~ mm diameter graphite rod supported
by two porous glass frits at either end was seeded
with powdered Mi/Cu catalyst by sprinkling the powder
onto the graphite rod. The mold was placed into a 2.5
cm quartz reactor as in Example 1 and after heating to
600C exposed to ethylene at a 100 cc/min. flow rate
for 4 hours. At the end of the run the mold was
removed from the quartz reactor. A filamentary carbon
growth had formed in the available space restricted by
the walls of the quartz reactor, the graphite rod and
porous quartz discs. The resulting tubular form of
filamentary carbon was then infiltrated with an epoxy
resin as in Example 18 giving a filamentary carbon
reinforced epoxy composite tube.
