474 Langmuir 2001, 17, 474-480

Characterization of Microporosity and Mesoporosity in

Carbonaceous Materials by Scanning Tunneling
Microscopy
J. I. Paredes, A. Martı́nez-Alonso,* and J. M. D. Tascón
Instituto Nacional del Carbón, CSIC, La Corredoria s/n, Apartado 73, 33080 Oviedo, Spain

Received June 19, 2000. In Final Form: November 7, 2000

Scanning tunneling microscopy has been employed to characterize the microporous and mesoporous
structure of different carbon materials. First, model micropores artificially created on a highly oriented
pyrolytic graphite substrate by plasma treatment were studied at the atomic scale: the observed increase
in electronic density near the Fermi level around the defect implies an increased adsorptivity of the model
micropore region. Second, activated carbon fibers were studied. A spongy mesoporous texture along with
slit-shaped microporosity (∼1 nm) was observed, accounting for the high adsorption properties of this
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

material. For comparison, nonporous thermally treated carbon black and nonactivated carbon fibers with
a ultramicroporous texture were also investigated. In the former case, as expected, no sign of extensive
microporosity or mesoporosity such as that of the activated carbon fibers was encountered, in agreement
with its poor sorptive capability. In the latter case, the interpretation of the results was rather troublesome,
Downloaded via UNIV PADERBORN on July 29, 2025 at 08:50:54 (UTC).

since the minute pore size rendered a reasonably accurate STM imaging difficult.

Introduction number of different techniques have been employed to

characterize not only the porous structure of these
Porosity in carbon materials plays a fundamental role materials but also their surface chemistry. These tech-
for their application in the adsorption and separation niques include gas adsorption (mainly N2 and CO2),9-13
technology of gases and liquids or in heterogeneous immersion calorimetry,9 small-angle X-ray scattering,13,14
catalysis, where the development of an extensive network inverse gas chromatography and differential thermal
of narrow pores lending high surface areas is required.1-3 analysis,15,16 Fourier transform infrared spectroscopy,15-17
As a particular example, activated carbon fibers have scanning and transmission electron microscopy (SEM/
recently been given considerable attention for their TEM)12,18-22 and electron spin resonance and Raman
potential use as adsorbents in environmental applications, spectroscopy,23 among others. It should be noted that, with
including the removal of a variety of organic contaminants the single exception of SEM/TEM, all these techniques
from water4,5 as well as the extraction from air of SOx and only provide indirect information about the porous struc-
NOx, which are byproducts of fossil fuel combustion for ture of the materials.
energy generation purposes.6 In other cases, such as those
of nuclear graphites and some carbon fiber composites In recent years, the development of scanning probe
employed as first-wall materials in thermonuclear fusion microscopy (SPM), first with the scanning tunneling
devices,7,8 porosity is not desirable since it promotes, on microscope (STM)24 and shortly thereafter its offspring
one hand, the diffusion of oxidizing gases into the inner the atomic force microscope (AFM),25 opened up the
material, leading to its corrosion and subsequent dete-
rioration and, on the other hand, the outgassing of (9) Rodrı́guez-Reinoso, F.; Molina-Sabio, M. Adv. Colloid Interface
Sci. 1998, 76-77, 271.
impurities that contaminate the fusion plasma. (10) Valladares, D. L.; Reinoso, F. R.; Zgrablich, G. Carbon 1998, 36,
Regardless of its specific applications, a thorough 1491.
knowledge of the physicochemical properties of porous (11) Nakashima, M.; Shimada, S.; Inagaki, M.; Centeno, T. A. Carbon
1995, 33, 1301.
carbon materials is essential for their general use in (12) Ryu, S. K.; Kim, S. Y.; Li, Z. J.; Jaroniec, M. J. Colloid Interface
science and technology. Over the years, a considerable Sci. 1999, 220, 157.
(13) Bóta, A.; László, K.; Nagy, L. G.; Copitzky, T. Langmuir 1997,
13, 6502.
* To whom correspondence may be addressed. Telephone num- (14) Cazorla-Amorós, D.; Salinas-Martı́nez de Lecea, C.; Alcañiz-
ber: (+34) 985 28 08 00. Fax number: (+34) 985 29 76 62. E-mail Monge, J.; Gardner, M.; North, A.; Dore, J. Carbon 1998, 36, 309.
address: [email protected]. (15) Putyera, K.; Jagiello, J.; Bandosz, T. J.; Schwarz, J. A. Carbon
(1) McEnaney, B.; Mays, T. J. In Introduction to Carbon Science; 1995, 33, 1047.
Marsh, H., Ed.; Butterworth: London, 1989; Chapter 5. (16) Adib, F.; Bagreev, A.; Bandosz, T. J. Langmuir 2000, 16, 1980.
(2) Rodrı́guez-Reinoso, F.; Linares-Solano, A. In Chemistry and (17) Jia, Y. F.; Thomas, K. M. Langmuir 2000, 16, 1114.
Physics of Carbon; Thrower, P. A., Ed.; Marcel Dekker: New York, (18) Huttepain, M.; Oberlin, A. Carbon 1990, 28, 103.
1989; Vol. 21, Chapter 1. (19) Oshida, K.; Kogiso, K.; Matsubayashi, K.; Takeuchi, K.; Koba-
(3) Jaroniec, M.; Choma, J. In Chemistry and Physics of Carbon; yashi, S.; Endo, M.; Dresselhaus, M. S.; Dresselhaus, G. J. Mater. Res.
Thrower, P. A., Ed.; Marcel Dekker: New York, 1989; Vol. 22, Chapter 1995, 10, 2507.
3. (20) Oya, A.; Yoshida, S.; Alcañiz-Monge, J.; Linares-Solano, A.
(4) Brasquet, C.; Le Cloirec, P. Carbon 1997, 35, 1307. Carbon 1995, 33, 1085.
(5) Brasquet, C.; Le Cloirec, P. Langmuir 1999, 15, 5906. (21) Le Cloirec, P.; Brasquet, C.; Subrenat, E. Energy Fuels 1997, 11,
(6) Mochida, I.; Korai, Y.; Shirahama, M.; Kawano, S.; Hada, T.; Seo, 331.
Y.; Yoshikawa, M.; Yasutake, A. Carbon 2000, 38, 227. (22) Brasquet, C.; Rousseau, B.; Estrade-Szwarckopf, H.; Le Cloirec,
(7) Moormann, R.; Hinssen, H. K.; Krüssenberg, A.-K.; Stauch, B.; P. Carbon 2000, 38, 407.
Wu, C. H. J. Nucl. Mater. 1994, 212-215, 1178. (23) Manivannan, A.; Chirila, M.; Giles, N. C.; Seehra, M. S. Carbon
(8) Barabash, V.; Akiba, M.; Bonal, J. P.; Federici, G.; Matera, R.; 1999, 37, 1741.
Nakamura, K.; Pacher, H. D.; Rödig, M.; Vieider, G.; Wu, C. H. J. Nucl. (24) Binnig, G.; Rohrer, H.; Gerber, Ch.; Weibel, E. Phys. Rev. Lett.
Mater. 1998, 258-263, 149. 1982, 49, 57.

10.1021/la0008535 CCC: $20.00 © 2001 American Chemical Society

Published on Web 12/28/2000
Porosity in Carbonaceous Materials Langmuir, Vol. 17, No. 2, 2001 475

possibility to directly image the surface structures of an

extensive variety of materials26-28 in real space with
unprecedented resolution, down to the atomic scale in the
most favorable cases. Therefore, SPM appears, a priori,
as a promising technique for providing important visual
information of the porous structure of carbon materials
at scales which are not readily accessible by means of
other techniques. However, very few articles have ap-
peared in the literature dealing directly with its applica-
tion to the study of porosity (especially microporosity) in
this type of material,13,22,29-32 the main reason being
limitations inherent to the technique that renders the
imaging of disordered and rough topographies (such as
those typical of porous carbons) difficult to perform.
This work is aimed at studying the microporosity and
mesoporosity of some carbon materials by means of Figure 1. Atomic resolution STM image of a highly oriented
pyrolytic graphite sample showing the effect of atomic vacancies
scanning tunneling microscopy. With that object, STM created by a very short (6 s) oxygen plasma treatment.
results for different carbon materials with different
degrees and types of porosity are presented and discussed, from Digital Instruments. Mechanically prepared Pt/Ir tips were
from highly oriented pyrolytic graphite with artificially employed. All the images were acquired in the constant current
created defects (to investigate porosity from a fundamental mode (variable height), whereby a topographical picture of the
point of view) to more application-oriented activated sample surface is obtained. Typical tunneling conditions were
carbon fibers with high surface areas. between 30 mV and 1 V for the bias voltage and between 0.35
and 1 nA for the tunneling current. It was observed that these
parameters could be varied along a wide range of values with no
Experimental Section observable changes in the corresponding images. The fibrous
Highly oriented pyrolytic graphite (HOPG) samples, from and particulate form of the fibers and the carbon black,
Union Carbide (grade ZYH), were cleaved in air to expose fresh respectively, demanded a careful sample preparation for their
new surfaces. To develop porosity on their surface, oxygen plasma observation by this technique. Small quantities of these materials
treatments were carried out. These were accomplished in a were mounted onto the sample holder by means of a carbon
Technics Plasma 200-G treatment chamber where the oxygen adhesive tape, paying special attention to detect and remove
plasma was generated by means of 2.45 GHz microwave (MW) aggregates protruding a great length from the sample surface
radiation. Two different etching conditions were employed with that could prevent a stable tunneling current signal.
the purpose of creating pores of different sizes: in the first, the To be assured of the reproducibility of the images, all the
HOPG was treated at a MW power of 40 W for 6 s, and in the samples were studied with several different previously unused
other, the MW power was raised to 100 W and the etching time tips and in several different areas. The images were flattened
to 9 min. and then smoothed by a low-pass filter to improve their visual
Activated carbon fibers (ACFs) with a BET specific surface presentation, taking special care to ensure that no distortions
area of 874 m2 g-1, as determined from N2 adsorption at 77 K, or artifacts were introduced by this procedure.
were prepared by physical activation of pyrolyzed Kevlar pulp
with carbon dioxide at a temperature of 1023 K and with a burnoff Results and Discussion
degree of 73%. Details of their preparation procedure and textural Several types of carbon materials with varied degrees
characterization by gas adsorption are given elsewhere.33
of porosity were examined and compared in this work:
Thermally treated (3000 K) carbon black (Sterling FT N880)
highly oriented pyrolytic graphite with shallow pores
with a mean particle diameter of 192 nm34 was obtained from
Cabot. N2 adsorption experiments at 77 K yielded a BET surface created by oxygen plasma etching of the fresh material,
area of 15.3 m2 g-1. Nonactivated carbon fibers were obtained with the intention of characterizing porosity at the atomic
from the pyrolysis of aramid fibers (Nomex 2.2) in a quartz reactor and nanometer scale in a model carbon material; activated
by heating under Ar at 10 K min-1 to 1173 K, followed by cooling carbon fibers, to study the porous structure of a high
to room temperature under Ar. The carbon fibers thus produced surface area and porosity carbonaceous material of interest
yielded a BET surface area of 0.3 m2 g-1 by N2 adsorption at 77 for its potential application as an adsorbent in diverse
K. However, their CO2 (273 K) equivalent microporous surface fields; carbon black was also studied in order to contrast
area, calculated by the DRK method, yielded a value of 750 m2 the characteristic features of a highly porous carbon with
g-1, implying that these fibers, in contrast to the carbon black,
those of other disordered carbon materials of low surface
present a ultramicroporous texture (pore size <0.7 nm).
area and porosity; finally, a ultramicroporous nonactivated
STM measurements on the mentioned samples were performed
in air at room temperature with a Nanoscope Multimode IIIa, carbon fiber with a very small N2 BET surface area but
with a high CO2 equivalent surface area was also
investigated.
(25) Binnig, G.; Quate, C. F.; Gerber, Ch. Phys. Rev. Lett. 1986, 56,
930. HOPG is a highly ordered carbon material whose surface
(26) Magonov, S. N.; Whangbo, M.-H. Surface Analysis with STM is composed of large atomically flat terraces where pits of
and AFM; VCH: Weinheim, 1996. different sizes can be artificially produced by means of
(27) Bottomley, L. A. Anal. Chem. 1998, 70, 425R.
(28) Lillehei, P. T.; Bottomley, L. A. Anal. Chem. 2000, 72, 189R. oxygen plasma etching.35 In this work, these pits are
(29) Daley, M. A.; Tandon, D.; Economy, J.; Hippo, E. J. Carbon 1996, studied as model pores in carbon materials. Figure 1 shows
34, 1191. an atomic resolution image of a highly oriented pyrolytic
(30) Vignal, V.; Morawski, A. W.; Konno, H.; Inagaki, M. J. Mater. graphite sample following a very mild treatment in oxygen
Res. 1999, 14, 1102.
(31) Stoeckli, F.; Centeno, T. A.; Donnet, J.-B.; Pusset, N.; Papirer, plasma for 6 s. Two ring-shaped protuberances about 1
E. Fuel 1995, 74, 1582. nm wide each and bulging 0.1-0.2 nm from the surface,
(32) Hoffman, W. P.; Fernandez, M. B.; Rao, M. B. Carbon 1994, 32, which are never found in the fresh material, can be
1383.
(33) Martı́nez-Alonso, A.; Jamond, M.; Montes-Morán, M. A.; Tascón, appreciated surrounded by the usual STM unperturbed
J. M. D. Microporous Mater. 1997, 11, 303.
(34) Cascarini de Torre, L. E.; Bottani, E. J.; Martı́nez-Alonso, A.; (35) Paredes, J. I.; Martı́nez-Alonso, A.; Tascón, J. M. D. J. Mater.
Cuesta, A.; Garcı́a, A. B.; Tascón, J. M. D. Carbon 1998, 36, 277. Chem. 2000, 10, 1585.
476 Langmuir, Vol. 17, No. 2, 2001 Paredes et al.

atomic arrangement of graphite. Several points have to

be considered to interpret this and other similar images.
First, it is highly unlikely that the protrusions arise from
some impurity molecules adsorbed on the HOPG surface,
since they would become considerably altered or even
wiped out by the STM tip after many consecutive scans
with the tunneling resistance used (∼108 Ω),36 and this
was not the case. Second, atomic oxygen (the main reactive
species in an oxygen plasma) is known to react with carbon
atoms both from defects (that is, carbon atoms with
unsaturated bonds) as well as from basal planes,35 in this
latter case creating atomic vacancies in the very early
stages, which consist of one or at most a few missing carbon
atoms in the lattice, and subsequently evolving into
nanometer-sized pits with extended etching times. This
is different from the case of molecular oxygen, which has
been shown to react, at temperatures below 700 °C,
exclusively with atoms at pre-existing defects on the HOPG Figure 2. STM image showing mesoporosity development on
surface, such as step edges and atomic vacancies, but not the surface of highly oriented pyrolytic graphite following
extended (9 min) oxygen plasma etching.
with carbon atoms from perfect sites of basal planes.37 In
the present case of attack by an oxygen plasma, the It has to be stressed that the enhancement in the
indication that the observed protuberances (Figure 1) are electronic density referred to previously can only be
formed on the originally defect-free and atomically flat distinctly and unmistakably perceived in a model carbon
basal planes, and not (at least exclusively) along step material such as graphite, due to its well-ordered atomi-
edges, arises from the fact that they appear distributed cally flat surface, which allows for purely electronic effects
all over the HOPG surface with a high uniformity and (responsible for the aforementioned enhanced adsorptiv-
density. Densities of several thousand protuberances per ity) not to be eclipsed by topographical effects (variations
square micrometer were typically found, which facilitated in the vertical positions of the atoms constituting the region
the location and imaging of individual protuberances at under study). As will be shown subsequently, in a general
the atomic scale. If the attack had proceeded only along disorganized carbon material, which tends to be very rough
step edges, the protuberances would have only been found even at atomic or near atomic scales, topographical effects
aligned along the steps, but that was not the case. Instead, predominate over electronic ones, and thus micropores
they were found on the terraces between steps (i.e., on the are detected essentially as physical voids in the surface
basal planes) with a highly uniform distribution. Thus, of the material in question, rather than as electronic
the protrusions observed after a very short oxygen plasma density protuberances. Also, in the case of mesopores and
treatment in Figure 1 are ascribed to atomic vacancies larger pores, the scales required to perform their imaging
created by the plasma on the initially defect-free basal are large enough to make electronic effects pass unnoticed
planes. These atomic vacancies, in addition, can be because, if any, they are completely outweighed by
regarded as model micropores in a carbon material. Third, topographical ones.
atomic vacancies in graphite induce an enhancement of Electronic effects are already obscured by topography
the electronic density near the Fermi level (EF) in the in graphite when it is subjected to a long enough plasma
carbon atoms near the vacancy, as has been shown by oxidation. Figure 2 shows a 100 × 100 nm2 image of an
several recent semiempirical and ab initio theoretical HOPG sample after 9 min of plasma oxidation at a MW
calculations,38-40 and changes in that partial electronic power of 100 W. The atomically flat surface characteristic
density near EF are, in a first approximation, what the of the pristine material has been substituted by a rough
STM actually detects41 together with changes in topog- and disorganized one where a great deal of mesoporosity
raphy, i.e., changes in the vertical position of the atoms. has developed. Some of the mesopores have approximately
So, in the case of Figure 1 of atomic scale imaging of an rounded shapes, whereas others are elongated. Their
atomically flat HOPG surface with atomic vacancies, the typical diameters or widths are between 2.0 and 7.7 nm
contrast is giving variations in the electronic density near with an average value of 3.7 nm. A few of them are linked
EF.36 Consequently, the protrusions represent increased to other pores, but the whole porous system does not make
Fermi level electronic density of the atoms comprising up an interconnected network. These mesopores should
the model micropore wall. This enhancement in the be formed by the expansion of micropores (Figure 1)
electronic density of the atoms constituting the micropore created on the HOPG surface during the plasma treatment
wall implies an enhancement of the chemical reactivity/ at random locations.35
adsorptivity of the micropore region,38 especially toward Next, STM images obtained from a representative
polar or polarizable molecules.2 Therefore, the STM visual microporous carbon material are considered. Figure 3a
depiction of the model micropore reveals the atomic level shows a typical picture of activated carbon fibers prepared
foundations of the adsorptive properties of micropores in from Kevlar fiber pulp. The right and bottom parts of the
carbon materials. image strongly suggest a highly spongy texture of the
sample since it displays a great amount of mesoporosity
(36) Hahn, J. R.; Kang, H.; Song, S.; Jeon, I. C. Phys. Rev. B 1996, that seems to penetrate very deeply into the inner
53, R1725. material. The pores of that region have diameters ap-
(37) Tandon, D.; Hippo, E. J.; Marsh, H.; Sebok, E. Carbon 1997, 35,
35. proximately between 4 and 16 nm, with a mean value of
(38) Lee, K. H.; Lee, H. M.; Eun, H. M.; Lee, W. R.; Kim, S.; Kim, D. 8.3 nm; consequently, they can be classified as mesopores.
Surf. Sci. 1994, 321, 267. Though not completely evident in Figure 3a, mesoporosity
(39) Hjort, M.; Stafström, S. Phys. Rev. B 2000, 61, 14089. also spreads over the other parts of that image: Figure
(40) Lee, K. H.; Causá, M.; Park, S. S.; Lee, C.; Suh, Y.; Eun, H. M.;
Kim, D. THEOCHEM-J. Mol. Struct. 2000, 506, 297. 3b presents a three-dimensional picture of the central left
(41) Tersoff, J.; Hamann, D. R. Phys. Rev. B 1985, 31, 805. part of Figure 3a in which the mesopores are patent as
Porosity in Carbonaceous Materials Langmuir, Vol. 17, No. 2, 2001 477

g-1.33 The mesopore size distribution calculated in the

present work by the BJH method indicates that most
mesopores have sizes ranging from 3 to 5 nm. Likewise,
the regularization method according to the density
functional theory44 yielded typical mesopore sizes ranging
from 2.5 to 9 nm, this result being quite consistent with
that attained by STM on the areas of the ACFs that were
interpreted to be originally more crystalline in the starting
material (Figure 3b).
To resolve micropores, very high-resolution imaging has
to be performed and that will depend both on the size of
the probing tip and on the topography of the sample
surface. In general, an SPM image is the result from the
convolution between the tip and the sample surface
features45,46 and therefore it has some degree of distortion
that should be borne in mind if misinterpretations, rather
quantitative than qualitative, are to be avoided. As a rule,
the smaller the tip radius of curvature, the higher the
resolution that can be attained, because the interaction
area between tip and sample is smaller, thus allowing
minute variations in the surface topography to be more
readily detected (even atomic resolution in the most
favorable cases, i.e., in atomically flat surfaces) and also
because a sharper tip can probe features such as narrow
trenches or pores which would otherwise remain unde-
tected. For this reason, deep and narrow features are more
hardly reached than shallow and broad ones; i.e., highly
corrugated topographies are much more difficult to track
than flat or smooth surfaces. The difficulty in attaining
a very high resolution in STM imaging of very rough active
carbons, prepared from poly(ethylene terephthalate) and
cellulose, has been pointed out by Bóta et al.13 recently.
It is for this reason that microporosity, and also meso-
Figure 3. Mesoporous structure of activated carbon fibers as
imaged by STM: (a) general view, the right and bottom parts porosity, was easier to trace in relatively flat areas
present a highly spongy texture; (b) detail of the central left compared with rougher ones.
part of (a), where mesopores are also present. Some images of the microporous structure of the ACFs
are shown in Figure 4. First, it has to be noted that the
well but in this case they do not present a spongy texture height difference between high and low-lying areas in the
such as that of the former case. The mesopores in this images is about 1-2 nm, this being 1 order of magnitude
case have diameters roughly between 3 and 10 nm, with greater than the differences induced by electronic effects
a mean value of 6.5 nm. The reason behind the different (0.1-0.2 nm, as observed previously), which confirms the
development of mesoporosity, both in size and in expansion screening of the latter by purely topographical features.
into the inner material, in two neighboring areas of the Also, the attaining of atomic resolution is precluded by
fibers can be possibly put down to differences in the this disordered and rough topography since, as commented
crystallinity of the feedstock material, which is Kevlar earlier, an ordered (or at least reasonably ordered) and
pulp in the present case. It is known that Kevlar presents atomically flat surface is an essential requirement for that
regions of different crystallinity within the same fiber.42,43 purpose. In all images of Figure 4 a great number of
Thus, small local differences in the degree of crystallization micropores can be observed along with a few small
of the Kevlar pulp could result in different texture mesopores (∼3 nm). The micropores have sizes ranging
developments upon pyrolysis and activation in such a way from about 0.6 to 1.7 nm with an average value of 0.94
that less crystalline areas of the precursor are more prone nm. Considering that only a small fraction of the total
to develop porosity (due to higher reactivity in gasification) sample surface is actually studied by STM at magnifica-
than higher crystallinity areas. In addition, no sign of tions high enough to resolve the micropores, this figure
mesopores (up to 50 nm) larger than those shown above is in reasonable agreement with that deduced from CO2
or macropores (greater than 50 nm) was found on this adsorption at 273 K, which is 1.08 nm,33 and also with the
sample. This observation is consistent with that made by fact that the micropore volume from N2 (77 K) adsorption
Brasquet et al.,22 who noted that CO2 activation, and is greater than that obtained from CO2 (273 K) adsorption
contrary to the case of steam activation, developed no (Table 1), implying the presence of supermicropores.
macropores or very large mesopores in carbon fibers. The Likewise, it has to be noted that a great number of
N2 (77 K) isotherms for the present ACFs are essentially micropores have slit shape, and this is especially noticeable
of type I, typical of microporous solids, but with a capillary in parts a and b of Figure 4, where they appear to be
condensation hysteresis loop at high relative pressures, interconnected to a great extent forming an irregular and
indicating the presence of mesoporosity. The mesoporous winding network. In Figure 4c, they are also quite
volume is only 0.06 cm3 g-1 (Sext ) 36 m2 g-1), to be connected to each other and many of them are slit-shaped
compared with the total porous volume, which is 0.44 cm3 or elongated, though some approximately rounded pores
(42) Rebouillat, S.; Peng, J. C. M.; Donnet, J.-B. Polymer 1999, 40,
7341. (44) Olivier, J. P. J. Porous Mater. 1995, 2, 9.
(43) Graham, J. F.; McCague, C.; Warren, O. L.; Norton, P. R. Polymer (45) Keller, D. J.; Franke, F. S. Surf. Sci. 1993, 294, 409.
2000, 41, 4761. (46) Villarrubia, J. S. Surf. Sci. 1994, 321, 287.
478 Langmuir, Vol. 17, No. 2, 2001 Paredes et al.

Table 1. Geometric and BET (N2, 77 K) Specific Surface Areas and Micropore Volume, Vµp, by the DRK Method from
CO2 (273 K) and N2 (77 K) Adsorption Data, for the Activated Carbon Fibers, Nonactivated Carbon Fibers, and Carbon
Black Studied in This Work
BET area Vµp(DRK) Vµp(DRK)
r F geometric area (N2, 77 K) (CO2, 273 K) (N2, 77 K)
(µm) (g/cm3) (m2/g) (m2/g) (cm3/g) (cm3/g)
ACF ∼2.5 ∼1.5 0.5 874 0.28 0.38
NACF ∼4 ∼1.5 0.3 0.3 0.27 ∼0
carbon black 0.096 1.85 16.9 15.3

Figure 5. STM image showing long narrow microporous

ramifications branching off a mesopore situated in the upper
part of the image.

play an important role in the selective adsorption of

molecules, e.g. planar molecules.
One of the most important functions of mesopores and
larger pores is to provide access for the adsorptive to the
inner part of the material and reach the micropores, where
the majority of adsorption takes place.2 This is illustrated
in Figure 5, where a mesopore about 3-4 nm wide in the
top part of the image branches into a long narrow
microporous crack with a variable width between 0.9 and
1.8 nm. It has to be noted that, although these STM images
reflect exclusively the surface morphology of the sample
and not its internal features, it may be reasonable to
assume that microporous ramifications can also branch
off the mesopore into the inner material and not just into
the surface. However, the possibility that the inner
porosity of the material could not be exactly the same as
that of the surface should not be ruled out, i.e., interiors
might be somewhat different. Likewise, the origin of some
of the mesopores formed in a carbon material like this can
be traced to the expansion of micropores by the oxidation
with the reagent during activation.1
To compare the surface structure of the ACFs with that
of other nonporous carbon materials, some images of a
carbon black surface are shown in Figure 6. This sample
has a BET specific surface area (N2, 77 K) of 15.3 m2 g-1,
which is about 2 orders of magnitude lower than that of
the ACFs of Figures 3-5 (874 m2 g-1). For comparison,
the geometric surface areas per unit mass of these samples
are also given in Table 1. These are calculated, considering
a cilindrical shape for the fibers and a spherical one for
Figure 4. Microporous structure of activated carbon fibers, the carbon black, by the formulas 2/Fr and 3/Fr, respec-
evidencing a great number of slit-shaped as well as ap- tively, where F is their density and r their radius. It can
proximately rounded pores, many of them interconnected. be noticed that the carbon black has a BET surface area
which is very close to the geometric area, implying that
are always found. Elongated porosity has also been the material has very low porosity. By contrast, the BET
observed by STM on oxidized carbon spheres30 and
activated carbon fibers,29 both prepared from a phenolic (47) Stoeckli, H. F. Carbon 1990, 28, 1.
precursor. Slit-shaped or elongated microporosity is known (48) Radhakrishnan, R.; Gubbins, K. E.; Watanabe, A.; Kaneko, K.
to be present in activated carbon materials1,47,48 and may J. Chem. Phys. 1999, 111, 9058.
Porosity in Carbonaceous Materials Langmuir, Vol. 17, No. 2, 2001 479

Figure 7. Nonactivated carbon fiber STM image. A similar

general structure is observed compared with that of the carbon
black (Figure 6a). However, it presents a fine structure which
could be reflecting the ultramicroporous texture of this sample.

g-1, which coincides with their geometric area, but they

present a high adsorptivity toward CO2 at 273 K: the
micropore volume Vµp calculated from CO2 (273 K)
adsorption isotherms by the DRK method yielded a value
of 0.27 cm3 g-1 (corresponding to an equivalent surface
area of 750 m2 g-1), which is very close to that of the ACFs
studied in this work (Table 1). The corresponding mi-
cropore volume from N2 (77 K) adsorption is nearly equal
to zero. This implies that these NACFs have a ultrami-
croporous structure with pore sizes below 0.7 nm, as
typically found with carbonized materials.49 From Figure
7 one can see that the general structure of the fiber is
Figure 6. STM images of thermally treated carbon black similar to that of the carbon black (Figure 6a). However,
showing the virtual absence of microporosity (a) and mesopo- when inspecting it in more detail, the former exhibits a
rosity (b) when compared with the activated carbon fibers great number of minute voids, about 0.4 nm wide, which
(Figures 3-5). are practically absent from the latter. From this, it is
surface area is about 3 orders of magnitude greater than tempting to conclude that those tiny voids are the
the geometrical one for the ACFs (Table 1). These results ultramicropores detected by CO2 adsorption in such a way
are consistent with the interpretation of the STM obser- that, when performing N2 adsorption experiments at 77
vations: the carbon black presents a very poorly developed K, the nitrogen molecules cannot enter the pores due to
porosity, if any (Figure 6a), in comparison with that of the diffusional constraints at this low temperature, so they
ACFs (Figures 4 and 5). Specifically, the former displays “see” a structure quite similar to that of the carbon black
a smaller number of voids than the activated carbon fibers, (as evidenced comparing the general structure of Figures
being considerably shallower. Furthermore, they are not 6a and 7 without taking into account the tiny voids of this
so distinctly interconnected forming networks as in the latter figure), thus yielding a BET surface area very close
case of the ACFs. Likewise, no sign of spongy textures to its geometric area (Table 1), as is the case with the
similar to those of the ACFs (created by the presence of carbon black. On the other hand, when performing CO2
a great number of mesopores penetrating deeply into the adsorption experiments at 273 K, the relatively high
material) was found in the carbon black. In fact, no temperature now allows the molecules to enter the
extensive mesoporosity such as that of the ACFs (Figure ultramicropores presumably seen in Figure 7, resulting
3) was encountered in the mentioned material. This can in the high micropore volume Vµp measured for these
be seen for instance in Figure 6b, which displays an image nonactivated carbon fibers (Table 1). Nevertheless, and
of an entire carbon black particle surrounded by two other despite the fact that this interpretation of the STM results
particles in the upper part. It is rather apparent from this for the nonactivated carbon fibers is consistent with their
image that the material has no mesoporosity which would adsorption experimental results, extreme caution must
promote the access of molecules to its inner part, sup- be taken when interpreting STM features of such a small
porting the fact that the external surface of this sample size (about 0.4 nm in this case) in a material with a rough
is essentially purely geometric as calculated in Table 1. and disordered topography such as the one under con-
All these aspects account for the reduced adsorptive sideration, due to the reasons referred to previously. It
properties of this material. Therefore, as has been shown, should be borne in mind that, possibly, features of that
differences between carbon materials of high and low size in this rough surface might not be probed by the STM
porosity can be evidenced by STM imaging, both at the tip with enough accuracy so as to be completely sure of
micropore and the mesopore scale. their existence. Thus, this question remains open to
Finally, Figure 7 shows an image of a nonactivated discussion.
carbon fiber (NACF) obtained from the pyrolysis of Nomex
aramid fibers. These fibers have a BET specific surface (49) Garrido, J.; Linares-Solano, A.; Martı́n-Martı́nez, J. M.; Molina-
area, obtained from N2 adsorption at 77 K, of only 0.3 m2 Sabio, M.; Rodrı́guez-Reinoso, F.; Torregrosa, R. Langmuir 1987, 3, 76.
480 Langmuir, Vol. 17, No. 2, 2001 Paredes et al.

Conclusions of extensive mesoporosity or microporosity was found in

Scanning tunneling microscopy has proved to be in thermally treated carbon black, in agreement with its poor
general a powerful technique for the visualization of the sorptive capability. The most difficult case to study and
microporous and mesoporous structure of carbon materi- interpret by this technique was that of nonactivated carbon
als, being able to explain the different sorptive properties fibers, since these present a ultramicroporous texture (pore
of different carbon materials, a question never addressed size < 0.7 nm) which renders their imaging troublesome,
before by the mentioned technique. The study of mi- evidencing a limitation of the scanning tunneling mi-
cropores in a model carbon material such as graphite at croscopy technique for accurately tracking extremely small
the atomic scale revealed an enhancement of the electronic features in rough and disordered materials. This means
density near the Fermi level in the atoms constituting the that, although the tentative interpretation of their STM
micropore wall, thus implying the enhanced reactivity features was consistent with their adsorption experimen-
and adsorptivity of the micropores. Activated carbon fibers tal results, further progress in the field of scanning probe
were shown to possess a spongy texture derived from the microscopy techniques should be needed in order to
presence of mesoporosity open to the surface which overcome this limitation and achieve results that were as
promotes the flowing of the adsorptive into the inner reliable as possible.
material to reach the micropores. Likewise, slit-shaped
Acknowledgment. Financial support from CICYT
or elongated micropores with an average width of about
(Project MAT96-0430) and DGICYT (Project PB98-0492)
1 nm were observed to form irregular and sinuous
is gratefully acknowledged. J. I. Paredes is indebted to
networks in the mentioned fibers. These characteristics
the FICYT for fellowship support.
account for the high sorptive capacity of the activated
carbon fibers. On the contrary, and as expected, no sign LA0008535