You are on page 1of 10

Carbon 43 (2005) 786–795

www.elsevier.com/locate/carbon

KOH and NaOH activation mechanisms of multiwalled


carbon nanotubes with different structural organisation
a,b,*
E. Raymundo-Piñero , P. Azaı̈s a, T. Cacciaguerra a, D. Cazorla-Amorós b,
A. Linares-Solano b, F. Béguin a,*
a
CRMD, CNRS-University, 1b Rue de la Férollerie, 45071 Orléans Cedex 02, France
b
Departamento Quimica Inorganica, Universidad de Alicante, Apartado 99, E-03080 Alicante, Spain

Received 23 June 2004; accepted 2 November 2004


Available online 9 December 2004

Abstract

New information was obtained on the mechanism of porosity development during chemical activation by KOH and NaOH using
various multiwalled nanotubes (MWNTs) of different structural organization. The high purity MWNTs were prepared by acetylene
decomposition on a cobalt-based catalyst at different temperatures. The obtained samples ranged from MWNTs with well organised
graphitic walls to nanotubes with disorganised layers mixed with some pyrolytic carbon when decreasing synthesis temperature. The
results of transmission electron microscopy (TEM) observations were linked with gas adsorption measurements and X-ray diffrac-
tion data. They show that NaOH is only effective with disordered materials whereas KOH is effective whatever the structural order.
After reaction of the poorly ordered precursor with KOH, the nanotubular morphology is completely destroyed, whereas it is pre-
served when NaOH is used. However for the more ordered materials, the morphology remains unchanged with both reactants.
Effects of activation are only seen with KOH, which generated a large concentration of defects in the nanotubes walls. The differ-
ences found between KOH and NaOH during activation are related with an additional intercalation step of metallic K or Na pro-
duced during the redox reactions. It is shown that metallic K has the ability to be intercalated in all materials in contrast with Na
which can only intercalate in the very disorganised ones. The conclusions obtained from the study on ordered nanotubes were con-
firmed with an ordered carbon black, demonstrating that the structural organization of the carbon precursor is an important param-
eter which must be taken into account when alkali reactants are used for the activation.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: A. Carbon nanotubes; B. Activation; C. Adsorption, Transmission electron microscopy, X-ray diffraction

1. Introduction from the performance point of view, allowing activated


carbons to be prepared from many kinds of carbona-
Chemical activation has been shown as a very effi- ceous precursors (coals, chars, fibres, etc) [1–18].
cient method to obtain carbons with high surface area Although it is a frequently used process to prepare acti-
and narrow micropore distribution. Among all the vated carbons, the general mechanism of chemical acti-
chemical activation agents, alkaline hydroxides as vation is not well understood, and the various
KOH or NaOH are reported to be highly interesting interpretations found in literature underline the process
complexity [2,5–17]. In general terms, chemical activa-
*
tion by alkalis consists in solid–solid or solid–liquid
Corresponding authors. Tel.: +34 965 903 946; fax: +34 965 903
454 (E. Raymundo-Piñero).
reactions involving the hydroxide reduction and carbon
E-mail addresses: raymundo@ua.es (E. Raymundo-Piñero), oxidation to generate porosity. During the reactions,
beguin@cnrs-orleans.fr (F. Béguin). CO, CO2 and H2 evolution is observed, and additional

0008-6223/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carbon.2004.11.005
E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795 787

reactions between active intermediates produced on the agent/carbon ratio of 4/1 and a maximum heat treat-
surface and the constituents of the gas phase are possi- ment temperature of 800 °C in a N2 flow (250 ml/min).
ble. A recent study suggests that the carbon/MOH After activation the materials were extensively washed
(M = Na or K) reaction mechanism is independent of in diluted HCl and water.
the hydroxide used and consists in the overlapping of re- Transmission electron microscopy (TEM, Philips
dox processes [15,16]. The hydroxide reduction leads to CM20) was used for the observation of the materials be-
H2 and Na or K metals, and carbon is oxidised to fore and after activation. For the analysis, few milli-
carbonates (Na or K) according to the global reaction grams of the MWNTs were dispersed in anhydrous
(1): ethanol and a drop of this suspension was deposited
6MOH þ C $ 2M þ 3H2 þ 2M2 CO3 ð1Þ on a TEM grid covered by a lacey amorphous carbon
film.
In a previous work [18] we have used chemical activa- The porous texture of pristine and activated MWNTs
tion with KOH to effectively develop the microporosity was analysed by N2 and CO2 adsorption at 77 and
of different kinds of multiwalled nanotubes (MWNTs), 273 K, respectively. Surfaces areas were calculated by
obtaining BET specific surface areas around 1000 m2/g applying the BET equation to the N2 isotherm data.
and micropore volumes close to 0.5 cm3/g. The TEM The total and narrow micropore volumes were calcu-
observations showed that the nanotubular morphology lated with the Dubinin–Raduskevich equation from
is preserved after activation although a large concentra- the N2 and CO2 adsorption data, respectively. Mesopore
tion of defects appears on the nanotubes surface. Sur- volume (for pore sizes between about 2–7.5 nm) was cal-
prisingly, NaOH, which is a very effective activating culated from the partial pressure range of the N2 iso-
agent [4], could not activate most kinds of MWNTs, therm between 0.2 < P/Po < 0.7.
in agreement with results found with other types of car- The products of the KOH or NaOH/MWNTs reac-
bonaceous precursors [14]. The common feature of all tion at different temperatures using a 2/1 ratio under
the materials for which NaOH was not an efficient acti- a N2 flow of 250 ml/min were analysed by X-ray diffrac-
vating agent, was the presence of some organisation of tion using CuKa or MoKa radiations (INEL CPS120).
the graphene layers. Therefore, the previous statements Since the activation process produces alkali metals
that carbon activation by NaOH or KOH takes place which are extremely sensitive to air, the products were
through the same reactions [15,16] are not completely transferred from the reaction furnace to a glove box
exact. It is clear that further work is needed to explain under Ar atmosphere and sealed in Lindeman capi-
the different activation behaviour of NaOH and KOH llaries.
as a function of the structural/nanotextural organisation In order to demonstrate that an intercalation step can
of the carbon precursor used. occur in some cases, carbon nanotubes prepared at
In this work, we use the advantage that MWNTs can 600 °C were mixed with KOH or NaOH in a 1/1 ratio
be prepared with different structural organizations, in and heat treated in a close vessel up to 800 °C inside a
order to study the influence of structure/nanotexture glove box under Ar atmosphere. The same experiment
on activation by KOH and NaOH. The pristine and was done using a carbon soot for comparison purposes.
activated MWNTs are analysed by TEM in a way which The products of the reactions were transferred and
is not possible with regular carbons, and linking these sealed into Lindeman capillaries for their further XRD
data with gas adsorption and XRD measurements al- analysis.
lows to demonstrate that the activation process is partly For having some reference of alkali metal intercala-
controlled by the structural organization of the precur- tion, reactions between MWNTs and Na or K metal
sor. These results will help to better understand the dif- were carried out by the two-bulb technique in vapour
ferent activation behaviours of NaOH and KOH phase according to the conditions described in reference
observed with coals of different ranks [16]. [20]. Previously outgassed MWNTs are placed in one
extremity of the tube and in the other extremity an ex-
cess of the distilled metal is introduced. The tube is evac-
2. Experimental uated under high vacuum, then sealed and placed inside
a furnace able to create a temperature gradient between
Carbon nanotubes were prepared by catalytic decom- the two extremities containing the MWNTs and the me-
position of acetylene at 450, 500 or 600 °C on a pow- tal. The temperature of the MWNTs was fixed at 300 °C
dered CoxMg1xO solid solution catalyst according to while metal was kept at lower temperature (250 °C) in
the process described in reference [19]. After synthesis, order to avoid its condensation on the nanotubes. As
the catalyst precursor is eliminated by dissolution in the intercalation compounds are very sensitive to air,
concentrated HCl. the samples were transferred in a glove box under Ar
Chemical activation is performed by physically mix- atmosphere and sealed into Lindeman capillaries for
ing KOH or NaOH with the MWNTs using a activating XRD analysis.
788 E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795

3. Results and discussion pyrolytic carbon. At 500 °C the MWNTs walls are not
as well ordered as in the previous case (Fig. 1b). Finally,
3.1. Chemical activation. Samples characterisation when the synthesis is performed at 450 °C, nanotubular
morphologies are still found, but they are not very well
Nanotubes were prepared by catalytic decomposition organised (Fig. 1c). In addition, a high amount of pyro-
of acetylene at 450 °C (NT450), 500 °C (NT500) and lytic carbon is also noticed either mixed with the nano-
600 °C (NT600). The TEM observations show that high tubes or on their surface. Hence, it is possible to produce
purity nanotubes are obtained at 600 °C and only few a tubular material with different degree of crystallinity
tips are closed. Fig. 1 presents high resolution transmis- by modifying the synthesis temperature.
sion electron microscopy images obtained in the 0 0 2 lat- Fig. 2 contains the N2 adsorption isotherms for the
tice fringe mode for the different samples. It can be MWNTs prepared at 450 °C (Fig. 2a), 500 °C (Fig. 2b)
observed that at 600 °C MWNTs with well organised and 600 °C (Fig. 2c) before and after activation with
carbon walls are obtained (Fig. 1a). The outer surface KOH and NaOH. All pristine nanotubes present similar
is less organised probably due to some deposition of isotherms between type II and IV, with an hysteresis
loop typical of a mesoporous material. Mesoporosity
comes from the tubes entanglement and from the central
canal of the open tubes. Table 1 collects the BET specific
surface areas and the micropore and mesopore volumes
calculated from the N2 and CO2 adsorption isotherms.
When the synthesis temperature of MWNTs decreases,
the specific surface area slightly increases, as a conse-
quence of the increase in micropore volume. This corre-
sponds to the increase of the number of defects observed
in Fig. 1. In Fig. 2, it can be also observed that KOH is a
much more efficient activating agent than NaOH to gen-
erate porosity in MWNTs, and efficiency seems also to
depend on the structural order of the MWNTs used.
For the different materials after KOH activation, there
is an increase of N2 adsorption at low partial pressures
indicating the generation of micropores and the hyster-
esis is also more pronounced than in the pristine tubes,
confirming a wider distribution of mesopores size. Table
1 shows that the specific surface area of KOH activated
MWNTs increases with decreasing the structural/micro-
textural order of the graphene layers of the pristine
materials, i.e. increasing the materials reactivity, reach-
ing 1670 m2/g for the most disordered MWNTs. This in-
crease in specific surface area comes from the generation
of both micro- and mesopores.
On the contrary, NaOH has only some efficiency when
the pristine sample is highly disordered, as NT450. For
this material, the N2 adsorption isotherm (Fig. 2a) re-
veals some increase of N2 adsorption in the low partial
pressure region and changes in the mesoporosity region.
Table 1 shows a noticeable increase of specific surface
area when compared to the pristine material. By con-
trast, a very low increase of specific surface area was
found after activation with NaOH of the nanotubes pre-
pared at 500 and 600 °C (Fig. 2b and c, respectively). In
Table 1 it can be noticed that this small increase corre-
sponds to the generation of some mesopores, as it is also
reflected by the differences found in the hysteresis loop of
the N2 isotherms presented in Fig. 2.
These results are easily understandable after observ-
Fig. 1. TEM 0 0 2 lattice fringe images of MWNTs prepared at (a) ing the TEM images from the materials after reaction
600 °C; (b) 500 °C; (c) 450 °C. with KOH and NaOH. Fig. 3 presents, as an example,
E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795 789

1400 1400

1200 NT450 1200 NT500


NT450KOH NT500KOH
NT450NaOH NT500NaOH
1000
V(cm3/g,STP)

1000

V(cm3/g,STP)
800 800

600 600

400 400

200 200

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
(a) P/Po (b) P/Po

1400

1200 NT600
NT600KOH
NT600NaOH
1000
V(cm3/g,STP)

800

600

400

200

0
0 0.2 0.4 0.6 0.8 1

(c) P/Po

Fig. 2. N2 adsorption isotherms at 77 K for the pristine and activated MWNTs.

Table 1 a mesoporous material, the isotherm of KOH activated


Porous texture characteristics of the pristine and activated MWNTs NT450 is almost type I, typical of a microporous mate-
SBET V microðN2 Þ V microðCO2 Þ Vmeso rial (Fig. 2). By contrast, Fig. 3b shows that when the
(m2/g) (cm3/g) (cm3/g) (cm3/g) same low temperature MWNTs are activated with
NT 450 °C 345 0.14 0.06 0.13 NaOH, the material keeps now the nanotubular mor-
NT 450 °C KOH 1670 0.62 0.39 0.43 phology after the reaction. However, the high resolution
NT 450 °C NaOH 562 0.22 0.14 0.19 images show that many defects have been generated in
NT 500 °C 330 0.13 0.05 0.13 the tubes walls and now all tips are open. These are
NT 500 °C KOH 1220 0.46 0.29 0.41 the reasons for the increase of the micropore and meso-
NT 500 °C NaOH 386 0.14 0.09 0.17
pore volumes of this material after NaOH activation
NT 600 °C 326 0.13 0.03 0.13 (Table 1).
NT 600 °C KOH 868 0.32 0.22 0.29 The transmission electron microscopy characterisa-
NT 600 °C NaOH 377 0.14 0.07 0.16
tions on the more ordered material, NT600, after reac-
tion with KOH and NaOH are presented in Fig. 3c
and d, respectively. In both cases, the nanotubular mor-
the 0 0 2 lattice fringe images for the MWNTs NT450 phology is preserved. The high magnification picture
and NT600 activated with KOH or NaOH. Fig. 3a after KOH activation reveals that the process generates
shows that after reaction of the poorly ordered precur- a large concentration of defects in the nanotubes walls,
sor (NT450) with KOH, the nanotubular morphology giving rise to micropores. In addition, the few nanotube
is completely destroyed. The high magnification image tips which were closed are now open and contribute to
resembles to that of an activated carbon with graphene the mesopore volume accessible for N2 (Table 1). After
layers fully misoriented. These results are in agreement NaOH reaction, the wall structure is preserved and de-
with the differences found between the isotherms of fects are not observed even when a weight loss of
the pristine and activated materials. Whereas the pris- 40 wt% was measured. The tips of all the tubes are now
tine tubes present a type II and IV isotherm typical of open (Fig. 3d), fitting well with the slight mesoporosity
790 E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795

Fig. 3. TEM 0 0 2 lattice fringe images of MWNTs prepared at 450 °C and activated with (a) KOH and (b) NaOH; and of MWNTs prepared at
600 °C and activated with (c) KOH and (d) NaOH.

development observed by gas adsorption when com- avoiding the contact with air. Fig. 4a presents the
paring the pristine and the activated materials (Table 1, XRD patterns for the pristine MWNTs (NT600), the
Fig. 2). physical mixture MWNTs/KOH and the mixture after
The results presented above demonstrate a different reaction at different temperatures. From these XRD re-
behaviour between KOH and NaOH during activation, sults, it can be observed that the redox reaction (carbon
which is related with the structural organisation of the oxidation) starts at around 400 °C, forming K2CO3. At
pristine material. In summary, there are clear evidences 600 °C, the KOH peaks are no longer visible, indicating
to state that KOH can be used as activating agent for that all KOH has been transformed into K2CO3 and if
samples with different degree of crystallinity, being more the reaction temperature is increased up to 700 °C some
efficient with decreasing crystallinity. By contrast, K2O can be detected. Finally, at 800 °C the only potas-
NaOH is only acting as activating agent when the mate- sium species which remain after the reaction is K2O. It
rial structural order is not so high, reacting only where it was also found that K2O is converted into KHCO3 after
exists a high concentration of defects as in the low tem- contact with air, which is given as the main product of
perature MWNTs or at the tips of highly ordered the reaction for all the previous studies where the
MWNTs. The reasons of such a behaviour can be found XRD analysis was performed in air atmosphere [17].
by deep analysis of the mechanism of chemical activa- However, the above observations agree with previous
tion. results [15,16], proposing that the first step in the chem-
ical activation with KOH is a redox reaction which
3.2. About reaction mechanism could be summarized by the global reaction (1). When
the reaction temperature is higher than 700 °C, the fol-
X-ray diffraction (XRD) analyses were performed lowing plausible reactions could explain the generation
with the objective of determining the reaction products of K2O:
formed by the redox reactions between KOH or NaOH K2 CO3 þ C $ K2 O þ 2CO ð2Þ
with carbon during the pyrolysis of the mixture. For this
purpose, mixtures of KOH or NaOH/NT600 in a 2/1 K2 CO3 $ K2 O þ CO2 ð3Þ
mass ratio were treated up to 400, 600, 700 or 800 °C
in N2 atmosphere and further analysed by XRD, while 2K þ CO2 $ K2 O þ CO ð4Þ
E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795 791

KOH K2 O
+ K2CO3 * Co NaOH Na
* + Na2CO3 * Co
* MWNT:KOH 800°C
+ + +
+ + * +
* MWNT:KOH 700°C ++ + *
+ + + ++ ++ * MWNT:NaOH 800°C
+ +
++ + + ++ + * +
+ +* +* MWNT:KOH 600°C + * MWNT:NaOH 700°C
+
+ + + ++ +
+ + * ++ * MWNT:NaOH 600°C
+ +
MWNT:KOH 400°C
MWNT:NaOH 400°C
MWNT:KOH
MWNT:NaOH
MWNT MWNT
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
(a) 2 Theta (b) 2 Theta

Fig. 4. XRD patterns (CuKa) of MWNTs prepared at 600 °C (a) during the reaction with KOH at different temperatures; (b) during the reaction
with NaOH at different temperatures.

These reactions are very favourable for potassium, hav- reduction is intercalated between the graphitic type lay-
ing large negative free energies at the reaction tempera- ers. In the literature, the intercalation of metallic potas-
ture [8] and it is known that alkali metal carbonates can sium has been suggested as one of the steps in the
decompose at temperatures below their melting point in chemical activation of carbon by using KOH
the presence of carbon [15,16,21]. In addition, such reac- [2,9,11,14]. However, only recently, Xue et al. were
tions are supported by the fact that evolution of CO and showing a direct evidence of the formation of graph-
CO2 was observed from TPD experiments over KOH/C ite–potassium intercalation compounds during the acti-
mixtures when temperature is higher than 700 °C vation of MCMB with KOH [17].
[15,16]. Finally, the K2O formed could be also expected In order to have some references to demonstrate any
to react with carbon by the reaction: metal intercalation during the activation by KOH or
NaOH, in a first step, we have checked if the pure metals
K2 O þ C $ 2K þ CO ð5Þ
(K or Na) are able to be intercalated into the different
which has a negative free energy change at around MWNTs precursors (NT450, NT500 and NT600). For
800 °C [8]. this purpose, vapour phase reactions were carried out
The case of NaOH activation can be seen in Fig. 4b between the distilled metal and the MWNTs by the
which shows the XRD patterns for the same pristine two-bulb technique according to the conditions de-
MWNTs, the MWNTs mixed with NaOH and the prod- scribed in the experimental part.
ucts of the reaction at different temperatures. In this Fig. 5a shows the X-ray diffractograms of the pristine
case, the redox reaction (Na2CO3 formation) starts at nanotubes prepared at different temperatures and of
higher temperature (600 °C) than for KOH (400 °C). their reaction products with K. After reaction of potas-
This observation is in agreement with the higher reactiv- sium with the nanotubes prepared at low temperatures
ity of KOH when compared with NaOH. However, con- (NT450 and NT500), the broad d002 peak corresponding
trarily to the case of KOH, we observe that Na2CO3 is to the interplanar distance of the pristine materials
the only product at 700 °C, and at higher temperature (d  0.35 nm) shifts to a higher value at d  0.4 nm. This
(i.e. 800 °C) important peaks of metallic Na are seen result indicates that K insertion/intercalation took place
in addition to the Na2CO3 peaks. This observation is resulting in the separation of the graphene layers. In the
in agreement with the global reaction summarized by case of the MWNTs prepared at higher temperature
Eq. (1). (NT600), the (0 0 2) line of pristine MWNTs completely
After comparison of Fig. 4a and b, it can be noticed disappears. A new peak at ca. 0.55 nm corresponds to
that, as expected, the first step of the activation reaction the 1st stage intercalation compound, whereas two other
by hydroxides is a redox process. Another important re- peak at ca. 0.4 nm and 0.28 nm are attributed to higher
mark is that for the reaction with NaOH (Fig. 4b) metal- stage derivatives [22,23].
lic Na was found on the surface of the material, whereas Fig. 5b shows the results after Na reaction with the
metallic K was not detected after reaction with KOH. same materials. The 0 0 2 line of the most disordered
The reason could be that at the reaction temperature MWNTs (NT450) is shifted to higher interlayer distance
the vapour pressure of potassium is higher than that from d002  0.35 nm to d002  0.4 nm, indicating Na
of sodium. However, these results could be also ex- insertion and separation of the graphene layers. By con-
plained considering that potassium produced by KOH trast, the XRD pattern of MWNTs prepared at higher
792 E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795

KOH. By comparing this XRD pattern after KOH reac-


tion with the one presented in Fig. 5a from the same
K Na material after reaction with K vapour, it is possible to
conclude that these lines correspond to K intercalation
NT450 NT450
compounds. Hence, these data confirm that K is interca-
lated between the graphene layers during the activation
process by KOH. In addition, Fig. 6a shows that if the
KOH reacted sample is exposed to air, the new peaks
K
Na
disappear and only the carbon (0 0 2) line remains. This
is another confirmation that K intercalation compounds
NT500
NT500 were formed during the activation reaction, as it is well
known that these derivatives are unstable in air, spe-
cially due to moisture [27]. The comparative XRD pat-
terns in the region of (0 0 2) line of carbon for the
K pristine and NaOH reacted material are presented in
Na
Fig. 6b. After reaction with NaOH, we could not ob-
NT600 NT600 serve any shift of the (0 0 2) line of carbon. This result
agrees with the data presented in Fig. 5b, indicating that
0.65 0.55 0.45 0.35 0.25 0.15 0.65 0.55 0.45 0.35 0.25 0.15 Na cannot be inserted/intercalated between the graph-
(a) d (nm) (b) d (nm) ene layers of the well organised MWNTs prepared at
600 °C.
Fig. 5. XRD patterns (CuKa) in the region of the (0 0 2) line of carbon The differences found between KOH and NaOH,
for (a) the various MWNTs before and after reaction with K in the gas
phase; (b) the various MWNTs before and after reaction with Na in
regarding their activating behaviour, can be explained
the gas phase. by different intercalation ability of K and Na depending
on the structural order of the pristine material [24–26].
Metallic potassium produced during the redox reaction
temperatures (NT500 and NT600) is unchanged after between MWNTs and KOH can be intercalated into
contact with Na vapour, indicating the absence of metal the nanotubes walls, independently of the structural
insertion/intercalation. order of the material, being responsible for the separation
Hence, it is clearly demonstrated that K can be in-
serted/intercalated in all the MWNTs studied and that
its intercalation process is favoured with the increase
in crystallinity, whereas Na can be inserted/intercalated
to a much lower extent only in the MWNTs of low
structural order. These results are in agreement with
the well known fact that intercalation of alkali metals
as Na or K strongly depends on the crystallinity of the
host carbon [24–26]. In this sense, K can be intercalated Air
into graphite-like structures [24], whereas Na can be
only intercalated in low temperature carbons with
non-graphitic structure [24–26].
After documenting the intercalation behaviour of al-
kali metals (Na and K) in the different MWNTs, we
decided in a second step to check if the same phenome- KOH
non could be observed in the conditions of chemical
NaOH
activation. For this purpose, 1/1 KOH/MWNTs or
NaOH/MWNTs mixtures were heat treated in a close
NT600
vessel up to 800 °C under Ar atmosphere. The experi- NT600
ments were done with the highly ordered MWNTs
(NT600) which were showing the highest differences 0.65 0.55 0.45 0.35 0.25 0.15 0.65 0.55 0.45 0.35 0.25 0.15
during KOH and NaOH activation. Fig. 6a presents (a) d (nm) (b) d (nm)
the XRD patterns in the region of the (0 0 2) line of car-
Fig. 6. XRD patterns (CuKa) in the region of the (0 0 2) line of carbon
bon for the pristine and KOH activated materials. Addi-
for (a) MWNTs prepared at 600 °C, after reaction with KOH in static
tionally to the typical (0 0 2) peak of the pristine material Ar atmosphere and the previous sample after contact with air; (b)
at d  0.35 nm, two new lines can be observed ca. MWNTs prepared at 600 °C and after reaction with NaOH in static Ar
0.28 nm and 0.4 nm after reaction of the nanotubes with atmosphere.
E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795 793

Air

KOH

Fig. 7. TEM 0 0 2 lattice fringe images of the carbon black CS. CS

and degradation of the graphitic layers and thus for the


development of microporosity. On the contrary, sodium
is only able to be intercalated in highly defective materi- 0.65 0.55 0.45 0.35 0.25 0.15
als. Therefore, NaOH only activates the MWNTs pre-
d (nm)
pared at low temperature.
In order to determine if the same results could be ob- Fig. 9. XRD patterns (CuKa) in the region of the (0 0 2) line of carbon
tained from other materials with an ordered nanotex- for the pristine CS, after reaction with KOH in static Ar atmosphere
ture, a carbon black (CS) prepared by CVD (CENIM, and the previous sample after contact with air.
CSIC, Madrid, Spain) has been studied. The 0 0 2 lattice
fringe mode image presented in Fig. 7 confirms that this
material presents a highly ordered nanotexture with gra- tion with NaOH, the increase in N2 adsorption is very
phitic layers concentrically oriented in the spherical par- low, indicating that NaOH was not able to develop
ticles. This material has been reacted with KOH or porosity. Hence, the carbon black and the MWNTs pre-
NaOH in the same conditions previously used for the pared at high temperatures (NT500 and NT600) have a
MWNTs (i.e. temperature: 800 °C; N2 flow of 250 ml/ similar behaviour because of their ordered nanotexture.
min). Fig. 8 shows the N2 adsorption isotherms of the XRD patterns of the (0 0 2) region for carbon after
carbon black before (CS) and after activation with either performing the KOH/carbon black (CS) reaction under
KOH or NaOH (CSKOH and CSNaOH, respectively). a static Ar atmosphere are shown in Fig. 9. By analogy
The pristine carbon black used has a negligible specific with Fig. 6a for the KOH/MWNTs reaction, lines at
surface area. A type I isotherm is observed in Fig. 8 after higher and lower interlayer distances than the (0 0 2) line
the reaction with KOH, that demonstrates the genera- for carbon are appearing, corresponding to the same
tion of micropores, due to KOH activation. After reac- intercalation potassium compound which was detected
in Figs. 5a and 6a for the case of the MWNTs. In Fig.
9 it can be also observed that if the KOH reacted sample
is exposed to air, these new lines found in the reacted
250
material disappear and only the (0 0 2) line for carbon
CS
CSKOH
remains.
200
CSNaOH Therefore, we can confirm that all the conclusions ob-
tained before with MWNTs can be extended to other
V(cm /g,STP)

150
carbonaceous materials and potassium intercalation is
a common step in the activation process of materials
3

100
with high microtextural order.
50

0 4. Conclusion
0 0.2 0.4 0.6 0.8 1
P/Po
The mechanism of chemical activation by KOH and
Fig. 8. N2 adsorption isotherms at 77 K for the pristine carbon black NaOH has been studied in nanotube materials ranging
(CS) and for CS activated with KOH and NaOH. from high purity MWNTs with well organised graphitic
794 E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795

walls to nanotubes with disorganised layers mixed with [3] Lozano-Castelló D, Lillo-Ródenas MA, Cazorla-Amorós D,
some pyrolytic carbon. We have demonstrated that Linares-Solano A. Preparation of activated carbons from Spanish
anthracite. I. Activation with KOH. Carbon 2001;39:741–9.
there is a different behaviour of NaOH and KOH as [4] Lillo-Ródenas MA, Lozano-Castelló D, Cazorla-Amorós D,
activating agents related with the structural organisation Linares-Solano A. Preparation of activated carbons from Spanish
of the pristine material. NaOH is only effective with dis- anthracite. II. Activation with NaOH. Carbon 2001;39:751–9.
ordered materials whereas KOH was found effective [5] Yamashita Y, Ouchi K. Influence of alkali on the carbonization
with all of them. process. I. Carbonization of 3,5-dimethylphenolformaldehyde
resin with NaOH. Carbon 1982;20:41–5.
The chemical activation process starts with similar re- [6] Yamashita Y, Ouchi K. Influence of alkali on the carbonization
dox reactions between carbon and KOH or NaOH, pro- process. II. Carbonization of various coals and asphalt with
ducing metallic K or Na and the metal carbonate NaOH. Carbon 1982;20:47–53.
(K2CO3 or Na2CO3). However, the reaction proceeds [7] Yamashita Y, Ouchi K. Influence of alkali on the carbonization
further with KOH giving K2O. In addition, K metal process. III. Dependence on type of alkali and of alkali earth
compounds. Carbon 1982;20:55–8.
produced from the redox reactions is intercalated be- [8] McKee DW. Gasification of graphite in carbon dioxide and water
tween the graphene layers of the carbon materials; the vapour. The catalytic effects of alkali metal salts. Carbon
higher the structural order, the better intercalation. 1982;20:59–66.
Hence, during the chemical activation with KOH, K [9] Hu Z, Vansant EF. Synthesis and characterization of a controlled-
intercalation can be considered as a relevant step which micropore-size carbonaceous adsorbent produced from walnut
shell. Micropor Mater 1995;3:603–12.
will have an important influence on the total reaction [10] Oh GH, Park CR. Preparation and characteristics of rice-straw-
mechanism. The metal intercalation favours the process based porous carbons with high adsorption capacity. Fuel
by separating the layers, that generates micropores or 2002;81:327–36.
even new high energy sites for the redox reactions. The [11] Guo J, Lua AC. Textural and chemical characterizations of
differences of efficiency found between KOH and NaOH adsorbent prepared from palm shell by potassium hydroxide
impregnation at different stages. J Colloid Interf Sci
for the activation of MWNTs of different crystallinity 2002;254:227–33.
are explained by the fact that K intercalates better than [12] Diaz-Teran J, Nevskaia DM, Fierro JLG, Lopez-Peinado AJ,
Na. Jerez A. Study of chemical activation process of a lignocellulosic
The possibility of obtaining a series of nanotubes of material with KOH by XPS and XRD. Micropor Mesopor Mater
various structural organisation simply by changing the 2003;60:173–81.
[13] Maciá-Agulló JA, Moore BC, Cazorla-Amorós D, Linares-
preparation temperature, was extremely useful for this Solano A. Activation of coal tar pitch carbon fibres: Physical
study. The uncertainty which could appear in the results activation vs. chemical activation. Carbon 2004;42:1361–4.
interpretation using materials from different sources is [14] Maciá-Agulló JA, Moore BC, Cazorla-Amorós D, Linares-
eliminated. Moreover, HRTEM could be efficiently ap- Solano A. Chemical activation by KOH and NaOH of carbon
plied for monitoring the effects of activation, that would materials with different crystallinity. Extended Abstracts, Carbon
2003, Oviedo (Spain): 2003.
not be possible with regular carbons. Hence, it is now [15] Lillo-Ródenas MA, Cazorla-Amorós D, Linares-Solano A.
well demonstrated that the structural organization of Understanding chemical reactions between carbons and NaOH
the carbon precursor is an important parameter which and KOH: an insight into the chemical activation mechanism.
must be taken into account when alkali reactants are Carbon 2003;41:267–75.
used for the activation. [16] Lillo-Ródenas MA, Juan-Juan J, Cazorla-Amorós D, Linares-
Solano A. About reactions occurring during chemical activation
with hydroxides. Carbon 2004;42:1371–5.
[17] Xue R, Shen Z. Formation of graphite–potassium intercalation
compounds during activation of MCMB with KOH. Carbon
Acknowledgments
2003;41:1862–4.
[18] Raymundo-Piñero E, Cazorla-Amorós D, Linares-Solano A,
This research was supported by a Marie Curie fellow- Delpeux S, Frackowiak E, Szostak K, et al. High surface area
ship of the European Community programme ‘‘Improv- nanotubes prepared by chemical activation. Carbon
ing Human Research Potential and the Socio-Economic 2002;40:1614–7.
[19] Delpeux S, Szostak K, Frackowiak E, Bonnamy S, Béguin F.
Knowledge Base’’ under contract number HPMF-CT-
High yield carbon nanotubes from the catalytic decomposition of
2001-01453. acetylene on in-situ formed Co nanoparticles. J Nanosci Nanotec
2002;2:481–4.
[20] Hérold A. Recherches sur les composes dÕinsertion du graphite.
Bull Soc Chim Fr 1995:999–1012.
References [21] Kapteijn F, Abbel G, Moulijn JA. Carbon dioxide gasification of
carbon catalyzed by alkali metals. Reactivity and mechanism.
[1] Marsh H, Yan DS, OÕGrady TM, Wennerberg AN. Formation of Fuel 1984;63:1036–42.
active carbons from cokes using potassium hydroxide. Carbon [22] Béguin F. Crystallochemistry of intercalation in the crystalline
1984;22:603–11. forms of carbon. In: Setton R, Bernier P, Lefrant S, editors.
[2] Otowa T, Tanibata R, Itoh M. Production and adsorption Carbon molecules and materials. London: Taylor and Francis;
characteristics of maxorb: high surface area carbon. Gas Sep Purif 2002.
1993;7:241–5. [23] Méténier K. PhD thesis, Orléans (France), 1999. p. 88–121.
E. Raymundo-Piñero et al. / Carbon 43 (2005) 786–795 795

[24] Tanaike O, Inagaki M. Ternary intercalation compounds of [26] Joncourt L, Mermoux M, Touzain P, Bonnetain L, Dumas D,
carbon materials having a low graphitization degree with alkali Allard B. Sodium reactivity with carbons. J Phys Chem Solids
metals. Carbon 1997;35:831–6. 1996;57:877–82.
[25] Hérold C, Hérold A, Lagrange P. New synthesis routes for donor- [27] Schlögl R, Boehm HP. The reaction of potassium–graphite
type graphite intercalation compounds. J Phys Chem Solids intercalation compounds with water. Carbon 1984;22:
1996;57:655–62. 351–8.

You might also like