Carbon 42 (2004) 1285–1290

www.elsevier.com/locate/carbon

Kinetics of pyrolytic carbon infiltration for the preparation

of ceramic/carbon and carbon/carbon composites
J.M. Rosas, J. Bedia-Matamoros, J. Rodr
ıguez-Mirasol *, T. Cordero
Chemical Engineering Department, School of Industrial Engineering, University of M
alaga, Campus de El Ejido, s/n, 29013 M
alaga, Spain
Received 10 July 2003; accepted 28 December 2003

Available online 26 February 2004

Abstract
Carbon/carbon and zeolite/carbon composites have been prepared by pyrolytic carbon infiltration of organic and inorganic
substrates with different porous structures. The chemical vapour infiltration kinetics of these substrates has been studied in a
thermogravimetric system at atmospheric pressure, using benzene as pyrolytic carbon precursor. The rate of pyrolytic carbon
infiltration seems to depend on the porosity of the substrate available to the pyrolytic carbon precursor, irrespective of the nature of
the substrate studied. Activation energy values of about 180 kJ/mol were found for the different substrates used in the temperature
range of 700–800 °C, where the cracking reaction of benzene takes place, predominantly, in a heterogeneous form. At higher
temperatures homogeneous reactions compete with heterogeneous ones and higher values of activation energies (280–380 kJ/mol)
were obtained. The oxidation of the pyrolytic carbon deposited on the different substrates studied takes place in the same range of
temperature, which suggests the presence of a similar pyrolytic carbon structure on substrates of different nature or a similar
accessibility to the deposited layer.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: A. Carbon composites, Pyrolytic carbon; B. Chemical vapor infiltration; D. Surface area; Porosity

1. Introduction petition between homogeneous reaction in the gas phase

and heterogeneous ones on the substrate surface.
Chemical vapour deposition/infiltration (CVD/I) of In the present work, the influence of the nature and
carbon from the pyrolysis of hydrocarbons is a tech- porous structure of the substrate on the CVI of pyrolytic
nique of great interest for different applications, such as carbon has been studied.
the densification of porous materials in the preparation
of C/C composites, for the aeronautic and space indus-
2. Experimental
tries [1] and the preparation of carbon molecular sieves
by partial blocking of micropore entrance [2].
Carbon/carbon and zeolite/carbon composites have
The process of formation of pyrolytic carbon by
been prepared by pyrolytic carbon (Pc) infiltration of
CVD/I is complex and not fully understood yet. Some
carbonaceous and inorganic substrates with different
authors propound that the pyrolytic carbon formation
porous structures in a conventional laboratory hori-
reactions take place in the gas phase [3–5]. Other
zontal tubular furnace (ID, 5 cm). The chemical vapour
researchers, however, state that the structure of the
infiltration kinetics of these substrates has been studied
substrate play an important role in the formation of
in a thermogravimetric system (CI Electronics) at
solid carbon through heterogeneous reactions [6–9].
atmospheric pressure.
Benzinger and coworkers [10,11] and Feron et al. [12]
A microporous Y zeolite Z14US (Davison Chemical
consider that pyrolytic carbon is produced by the com-
Co.) widely used as hydrocarbon cracking catalyst, a
mesoporous MCM-41 zeolite with a Si to Al molar ratio
of 25 and a non-porous zeolitic material (MCM-41-
*
Corresponding author. 1000), prepared by heat treatment of MCM-41 at 1000
E-mail address: mirasol@uma.es (J. Rodr
ıguez-Mirasol). °C, were used as inorganic substrates. Non-porous SP1
0008-6223/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carbon.2004.01.028
1286 J.M. Rosas et al. / Carbon 42 (2004) 1285–1290

graphite, a very narrow microporous char (C800), apparent surface areas (ABET and ADR , respectively) for
prepared by carbonisation of lignin and an activated AC, Z14US and C800 char, with higher values for those
carbon (AC), prepared by partial gasification with water derived from the CO2 adsorption data, can be attributed
vapour of C800 char, were used as carbon substrates. to the lack of achievement of true N2 adsorption equi-
Benzene (2 vol%) in a stream of He was used as the librium, due to restricted diffusion through the pores at
pyrolytic carbon precursor. The benzene cracking reac- the low temperature of adsorption ()196 °C). These
tion temperature ranged from 700 to 900 °C. The CVD results indicate the existence of very narrow micropores
experiments were carried out with about 5 mg of sub- in the structure of these substrates. An opposite
strate when the TGA system was used and 1 g of sub- behaviour is observed for MCM-41 zeolite, which
strate when the reaction took place in the tubular exhibits negligible adsorption of CO2 and adsorption of
reactor. A total flow rate of 150 cm3 (STP)/min was used N2 similar to that found in Z14US zeolite. This zeolite
in both cases. presents a much higher external surface area than
The oxidation behaviour of the carbonaceous sub- Z14US, due basically to a well-developed narrow mes-
strates and the composites materials was studied over oporosity, with total absence of narrow micropores.
the entire burn-off range using non-isothermal thermo- C800, a char with a narrow microporous structure,
gravimetric analysis (air, 150 cm3 (STP)/min; 1 atm) at a shows a relatively high value of the CO2 surface area
heating rate of 5 °C/min. and a low value of the BET one. SP1 graphite is a non-
The porous structure of the different substrate and porous substrate with a low external surface area. Heat
the resulting composites were analysed by adsorption of treatment of MCM-41 at 800 °C slightly reduced the
N2 at )196 °C and CO2 at 0 °C (Omnisorb, Coulter) and micropore volume, without causing a substantial change
their surface texture and chemistry by scanning electron of the mesopore volume (see MCM-41-800 in Table 1).
microscopy (JSM 840 JEOL) and by X-ray photoelec- A more drastic heat treatment of this zeolite to 1000 °C
tron spectroscopy (5700C model Physical Electronic), produced a non-porous zeolitic substrate (MCM-41-
respectively. From the N2 isotherm, the apparent sur- 1000), with an external surface area very similar to that
face area (ABET ) was determined applying the BET shown by SP1 graphite.
equation, the micropore volume (Vt ) and the external Fig. 1 shows the pyrolytic carbon deposited on the
surface area (At ) was calculated using the t-method and activated carbon substrate as a function of reaction
the narrow mesopore volume (Vmespo ) was obtained by
subtracting the micropore volume from the volume of 0.25
Pyrolytic carbon deposited (g Pc/g AC)

N2 adsorbed at a relative pressure of 0.79 [13]. From the

725ºC
CO2 adsorption data the narrow micropore volume 0.20
775ºC
(VDR ) and the apparent surface area (ADR ) was calcu- 750ºC
lated using the Dubinin–Radushkevich equation. 0.15
800ºC

0.10
3. Results and discussion
700ºC
0.05
Table 1 shows the surface properties of the different
substrates used in this study. AC and Z-14US zeolite
0.00
present a microporous structure, with high values of 0 50 100 150 200
t (min)
surface area derived from both N2 and CO2 adsorption
data. The differences observed between the N2 and CO2 Fig. 1. Pyrolytic carbon deposition on AC as a function of reaction
micropore volumes (Vt and VDR , respectively) and time at different temperatures.

Table 1
Surface properties of the different substrate used in this study
Substrates N2 isotherms CO2 isotherms
ABET (m2 /g) Vt (cm3 /g) At (m2 /g) Vmesopor (cm3 /g) ADR (m2 /g) VDR (cm3 /g)
AC 746 0.310 25 0.035 1025 0.386
C800 30 0.013 443 0.168
SP1 <1
Z14US 556 0.248 21 0.015 577 0.217
MCM-41 645 0.420 141 0.121 N N
MCM-41-800 539 0.308 138 0.118 N N
MCM-41-1000 1 0.008 N N
N: negligible.
 J.M. Rosas et al. / Carbon 42 (2004) 1285–1290 1287

time, at different temperatures. Blank experiments were 0.25

carried out without substrate and no mass gain was

Pyrolytic carbon deposited

observed below 825 °C, indicating that no homogeneous 0.2

(g Pc/g Subs.)
chemical vapour deposition takes place below this Z-14US

0.15
temperature under the operation conditions used in this AC
study. MCM-41
0.1
The deposition proceeds initially at a high rate and
after certain mass gain it decreases considerably. The 0.05
MCM-41 1000

initial rate of deposition substantially increases with Graphite

reaction temperature. However, the total amount of 0

pyrolytic carbon deposited decreases with chemical va- 0 20 40 60 80 100 120 140 160 180 200
t (min)
pour deposition temperature. These results suggest that
at low temperature the cracking reaction of benzene take Fig. 3. Pyrolytic carbon deposited on different substrates at 800 °C as
place on the micropore surface of the activated carbon, a function of reaction time.
where pyrolytic carbon is deposited uniformly reducing
the pore dimensions. The reaction proceeds on the
external surface area of the substrate once the pyrolytic substrate at temperatures lower than 800 °C. Pyrolytic
carbon precursor cannot penetrate into the micropores. carbon is deposited very slowly on SP1 graphite as a
The reaction rate increases considerably with tem- consequence of the low porosity of this substrate.
perature and diffusion of benzene through the narrower Fig. 3 compares the pyrolytic carbon deposition as a
micropores limits partially the cracking process inside function of reaction time at 800 °C, for the different
these pores [14]. substrates used in this study. The results obtained for
At higher temperatures the pyrolytic carbon deposi- the pyrolytic carbon deposition on the microporous
tion takes places preferentially on the surface of the Z14US zeolite were very similar to those observed for
wider micropores and at the micropore entrance, AC. AC shows a slightly higher rate of deposition at the
blocking it shortly after the beginning of the deposition initial state of the process than Z14US, probably due to
that continues at lower rate on the external surface. As the higher micropore volume of the former. However,
mentioned above, infiltration of AC with pyrolytic car- Z14US presents a higher total amount of pyrolytic
bon at 800 °C during 120 min produced a composite carbon deposited than that observed for AC. It is likely
material, AC/carbon, with a low N2 surface area and that the wider micropores of this zeolite allow higher
micropore volume. However, high CO2 micropore vol- amount of pyrolytic carbon to be deposited before the
ume and surface area (1090 m2 /g and 0.417 cm3 /g, micropores become partially block, remaining inacces-
respectively) were found for this sample, even higher sible to the pyrolytic carbon precursor. At this stage the
than for the corresponding substrate (AC; Table 1). This deposition continues on the external surface. In this
result suggests that the pore entrance were not com- case, low CO2 surface area and micropore volume were
pletely blocked during this reaction time and that some found after pyrolytic carbon infiltration at 800 °C dur-
narrow micropores have been formed by the pyrolytic ing 2 h. The nature of the surfaces of the AC activated
carbon deposition on the wider micropores. carbon and that of the Z14US zeolite are clearly differ-
Fig. 2 represents the pyrolytic carbon deposited on ent. However, both substrates present a similar CVI
SP1 graphite as a function of reaction time, at different reaction rate at 800 °C (and at lower temperatures, not
temperatures. No deposition was observed on this shown in the text), in spite of Z14US is a catalyst for the
cracking reactions of hydrocarbons.
The initial CVI reaction rate for MCM-41 zeolite is
0.3 much lower than that for AC and Z14US, although
Pyrolyticcarbon deposited

0.25
higher than those for SP1 graphite, C800 char and
MCM-41-1000. The nature of MCM-41 zeolite surface
0.2
(g Pc/g SP1)

850ºC should be more comparable to that of Z14US zeolite

0.15 than to that of the activated carbon. The lower CVI
825ºC
reaction rate shown by MCM-41 seems to be related to
0.1
the absence of narrow microporosity (that measured by
0.05
800ºC
CO2 ) in this zeolite, given that MCM-41-800 presents a
BET apparent surface area similar to that observed for
0
0 50 100 150 200 250 Z14US.
t (min)
The extremely low apparent surface areas of MCM-
Fig. 2. Pyrolytic carbon deposition on SP1 graphite as a function of 41-1000 and SP1 graphite appear to be responsible for
reaction time at different temperatures. the much lower CVI reaction rates of these samples, two
1288 J.M. Rosas et al. / Carbon 42 (2004) 1285–1290

substrates that clearly present surface of different nature.

Finally, C800 char exhibits a CVI reaction rate (not
shown in Fig. 3) very similar to that found for MCM-
41-1000 and SP1 graphite, in spite of this char presents a
CO2 micropore volume and surface area comparable to
those found for Z14US zeolite. However, the small N2
apparent surface area and micropore volume observed
for this char indicate the presence of very small mi-
cropores in this substrate, which may impede the pre-
cursor of the pyrolytic carbon to infiltrate the pore
structure before cracking. In this case the deposition
appears to take place on the external surface of the
substrate, blocking off the pores very fast, at a similar
rate to those observed on SP1 graphite and MCM-41-
1000. This result has also been recently observed for
carbon deposition from methane [8].
It has to be noted the tendency found for CVI on
MCM-41 zeolite. The deposition rate for this zeolite
increases slightly with reaction time. The decrease of the
size of the mesoporosity to microporosity as reaction
proceeds may cause this increase in the reaction rate. It
is also noteworthy that the final deposition rates ob-
served for all the substrate, except for that of MCM-41,
are very similar. In fact, the N2 surface areas and
micropore volumes found for Z14US/carbon and AC/
carbon composites after a deposition process carried out
at 800 °C and 120 min, become very low (<1 m2 /g),
similar to those measured for SP1/carbon, MCM-41-
1000/carbon and C800/carbon composites, obtained at
the same conditions. Nevertheless, MCM-41/carbon
composite obtained at the conditions mentioned above
showed a BET surface area of about 280 m2 /g and an
external surface area of about 80 m2 /g, which would
justify the higher deposition rate observed for this sub-
strate.
Activation energy values of about 180 kJ/mol were
found for AC and Z14US substrates in the range of
temperature studied. Similar values have been obtained
in the literature for heterogeneous decomposition of
different hydrocarbons [9,12,14,15]. Higher values, 280
and 380 kJ/mol were found for SP1 and MCM-41 and
MCM-41-1000, respectively. The experiments for these
substrates were carried out at higher temperatures (800–
875 °C), where homogeneous reactions in the gas phase Fig. 4. Scanning electron micrographs of AC (a), AC/carbon com-
posite (b) and pyrolytic carbon obtained by cracking of benzene at 900
are in competition with heterogeneous one on the sur- °C without any substrate (c).
face [10–12].
Fig. 4 shows the scanning electron micrographs of
pyrolytic carbon obtained by cracking of benzene at 900 of the substrates through heterogeneous reactions. The
°C without any substrate and those corresponding to fact that the surface of the AC/carbon composite ap-
AC and AC/carbon composite. The spherical form of pears almost identical to that of the AC substrate seems
the pyrolytic carbon obtained at 900 °C, where homo- to indicate that the pyrolytic carbon has been deposited
geneous reactions prevail, is not observed on the sur- on the internal surface of the substrate.
face of the AC/carbon composite, very similar to that of Fig. 5 shows the oxidation behaviour for the carbo-
the activated carbon substrate. The results obtained naceous substrates and the composites materials pre-
for the others composites are identical, which indicate pared by CVI of pyrolytic carbon at 800 °C and a
that the decomposition of benzene occurs at the surface reaction time of 2 h. AC presents the lowest oxidation
 J.M. Rosas et al. / Carbon 42 (2004) 1285–1290 1289

Z-14US

N (E)/E
Z-14US/carbon

112 107 102 97

Binding Energy (eV)
Fig. 5. Oxidation resistance profiles for different substrates and com-
posites materials. CVI temperature: 800 °C; CVI time: 120 min. Fig. 6. XPS Si 2p spectra for Z-14US substrate and Z-14US/carbon
composite.

resistance. Its oxidation starts at 400 °C with a low accessibility to the deposited layer. The oxidation of the
reaction rate that increases considerably at temperatures pyrolytic carbon of SP1/carbon composite seems to in-
above 580 °C. The oxidation resistance for this activated crease the reactivity of SP1 graphite of SP1/carbon
carbon is enhanced by deposition of pyrolytic carbon. composite. This may be due to the formation of active
The oxidation of AC/carbon starts at about 550 °C, site on the surface of SP1 during the pyrolytic carbon
which indicates that the pyrolytic carbon deposited is oxidation of the composite.
less reactive than the substrate (AC). Once the pyrolytic
carbon is removed, the oxidation of the composite takes
place at the same rate as that of the AC, suggesting that 4. Conclusions
surface of the activated carbon is now been oxidised
[8,16–18]. The rate of pyrolytic carbon infiltration seems to
The oxidation reaction for the pyrolytic carbon of the depend on the porous structure of the substrate avail-
Z-14US/carbon and MCM-41/carbon start at a tem- able to the pyrolytic carbon precursor, irrespective of
perature similar as that for the AC/carbon composite. the nature of the substrate studied.
The pyrolytic carbon deposited on MCM-41 shows a Activation energy values of about 180 kJ/mol were
lower oxidation resistance than that deposited on Z- found for the substrates used in the temperature range
14US, likely due to the high surface area of MCM-41/ of 973–1073 K, where the cracking reaction of benzene
carbon composite (280 m2 /g). The fact that the Z- takes place, predominantly, in a heterogeneous form. At
14US/carbon exhibit a higher oxidation resistance than higher temperatures homogeneous reactions compete
MCM-41/carbon, both with the same amount of pyro- with heterogeneous ones and higher values of activation
lytic carbon deposited, suggests that only a small frac- energies (280–380 kJ/mol) were obtained.
tion of the total surface area of the carbon deposited on The oxidation of the pyrolytic carbon deposited on the
Z-14US is available to the oxidant gas. Most of the different substrates studied takes place in the same range
surface area of the carbon in this composite must be of temperature, which suggests the presence of a similar
blocked and becomes available to the gas as oxidation pyrolytic carbon structure on substrates of different nat-
proceeds [19]. ure or a similar accessibility to the deposited layer.
Nevertheless, it can be observed a slight increase of
the oxidation rate for Z-14US/carbon with respect to References
that of the MCM-41/carbon at low temperatures, which
can be attributed to pyrolytic carbon bonding to Si on [1] Murdi N. Carbon fiber/carbon composites: production, properties
the external surface of the zeolite. The deconvolution of and applications. In: Marsh H, Heintz EA, Rodr
ıguez-Reinoso F,
the XPS band C 1s for the Z-14US/carbon composite editors. Introduction to carbon technologies. Alicante: Universi-
exhibits a small peak at a binding energy close to 283 eV, dad; 1997. p. 597–633.
[2] Sircar S, Golden TC, Rao MB. Carbon 1996;34:1–12.
which can be ascribed to C–Si bond. It can also be ob- [3] Grisdale RO. J Appl Phys 1953;24:1082–91.
served in the XPS Si 2p spectrum of Fig. 6, the forma- [4] Hirt TJ, Palmer HB. Carbon 1963;1:65–70.
tion of a small peak at about 102 eV that can be ascribed [5] McAllister P, Hendricks JF, Wolf EE. Carbon 1990;28:579–88.
to C–Si bond. [6] Fitzer E. J Phys IV 1991;1:509.
The pyrolytic carbon of the SP1/carbon composite [7] Verma SK, Walker Jr PL. Carbon 1992;30(6):829–36.
[8] Quli FA, Thrower PA, Radovic LR. Carbon 1998;36(11):1623–32.
is oxidised in the same range of temperature than those [9] Freitas MMA, Figueiredo JL. Fuel 2001;80:1–6.
of Z-14US/carbon and MCM-41/carbon ones, suggest- [10] Bezinger W, Becker A, H€ uttinger KJ. Carbon 1996;34:957–66.
ing the presence of a similar structure or a similar [11] Bezinger W, H€ uttinger KJ. Carbon 1996;34:1465–71.
1290 J.M. Rosas et al. / Carbon 42 (2004) 1285–1290

[12] Feron O, Langlais F, Naslain R, Tr
ebault J. Carbon 1999;37: [16] Cordero T, Thrower PA, Radovic LR. Carbon 1992;30:365–74.
1343–53. [17] Rodr
ıguez-Mirasol J, Thrower PA, Radovic LR. Carbon 1993;
[13] Gregg SI, Sing KSW. Adsorption, surface area and porosity. 31:789–99.
London: Academic Press; 1982. [18] Rodr
ıguez-Mirasol J, Thrower PA, Radovic LR. Carbon 1995;33:
[14] Prasetyo I, Do DD. Carbon 1999;37:1909–18. 545–54.
[15] Kawabuchi Y, Kishino M, Kawano S, Whitehurst DD, Mochida [19] Rodr
ıguez-Mirasol J, Cordero T, Radovic LR, Rodr
ıguez JJ.
I. Langmuir 1996;12:4281–5. Chem Mater 1998;10:550–8.