Langmuir 1996,11, 4386-4392

4386

Activated Carbon Surface Modifications by Nitric Acid,

Hydrogen Peroxide, and Ammonium Peroxydisulfate
Treatments
C. Moreno-Castilla," M. A. Ferro-Garcia, J. P. Joly,? I. Bautista-Toledo,
F. Carrasco-Marin, and J. Rivera-Utrilla
Departamento de Quimica Inorganica, Facultad de Ciencias, Universidad de Granada,
18071 Granada, Spain
Received March 29, 1995. I n Final Form: July 31, 1995@
A series of activated carbons with different degrees of activation was treated with HN03, HzOz, and
(NH4)2S208in order to introduce oxygen surface complexes. The effects of the oxidizingtreatments on the
surface area, the pore texture, and the surface chemical nature were analyzed by means of N2 and COz
adsorption, mercury porosimetry, FTIR, TPD, electrophoretic, and mass titration measurements. Results
obtained show that the HN03 treatment affects the surface area and the porosity of the samples to a
greater extent than the other treatments. Carboxyl groups were essentially fixed after the three treatments,
although ketone and ether groups, as detected by FTIR, were also fixed after the treatments with peroxides.
The most important conclusionwas that the stronger acid groups were fixed after the (NH4hSz08treatment
rather than after the HN03 treatment, in spite of the fact that this latter treatment fixed the largest
number of oxygen complexes that evolved as COz.

Thus, it has been showng-13 that in the adsorption of

inorganic compounds on activated carbons from aqueous
Oxygen surface complexes are formed on activated
solutions the surface chemistry of the adsorbent is, in
carbons when they are treated with oxidizing agents either
general, more important than the surface area and pore
in the gas phase or in solution. The treatments produce
texture ofthe adsorbent. Moreover,the presence ofoxygen
three types of surface oxides: acidic, basic, a n d n e ~ t r a l . l - ~ surface groups on carbons can also affect the behavior of
One of the methods to introduce predominantly acidic
supported c a t a l y ~ t s , ~ and
~ J ~ itJ ~is known that acidic
surface oxides, i.e., carboxylic, phenolic, and lactonic
surface oxides can be impotant in reactions catalyzed by
groups, is to treat the activated carbons with different
activated carbons, such as the catalytic conversion of
oxidizing solutions. In addition to these acidic groups,
alcohols and olefin polymerization and racemization.16-19
basic and neutral suface groups that evolve CO after
The aim of the present paper is to ascertain the changes
heating at high temperature are also introduced. Fixation
in pore texture and surface chemitry of a series of activated
of the acidic groups on the surface of the activated carbons
carbons, prepared with different degrees of activation,
makes it more hydrophilic, decreases its pH of the point
after their treatments with three oxidizing agents in
of zero charge, and increases the negative surface charge
aqueous solution: nitric acid, hydrogen peroxide, and
d e n ~ i t y . ~ >At~the
- l ~same time, the above treatments can
ammonium peroxydisulfate.
affect the surface area and pore texture of the activated
Experimental Section
carbons.
The above changes in the surface chemistry of the
The activatedcarbons were prepared from almond shells. The
sample caused by the formation of acidic oxygen surface
raw material was pyrolyzed at 1173 K in a Nz flow for 1h and
activated in a steam flow at 1123 Kfor different periods of time,
complexes will affect the behavior of the activated carbons
as described elsewhere.20The final particle size range was
when they are used either as adsorbents or as catalysts.
between 0.15 and 0.25mm. The sample numberofthe activated
carbons, the percentage burn-off (BO) obtained during the
* To whom correspondenceshould be addressed.
activation step, and the ash content are given in Table 1.
Universit6 Lyon 1, Institut des Sciences de la MatiBre,
The samples so obtained were oxidized with HN03 and H2Oz
Laboratoire d'Application de la Chimie a YEnvironnement,ESCIL,
and some of them with (NH&SzOs, all reagent grade supplied
69622 Villeurbanne Cedex, France.
by Merck. These oxidized samples will be referred to in the text
Abstract published inAdvance ACSAbstracts, October 1,1995.
as N, 0, or S, respectively. The procedures were as follows: 1
(1)Boehm, H. P. In Advances in Catalysis; Eley, D. D., Pines, H.,
g of carbon was treated with 10 mL of concentrated HN03 (13.9
Weisz P. B., Eds.; Academic Press: New York, 1966; Vol. 16, p 179.

Introduction

(2) Puri, B. R. In Chemistry and Physics of Carbon; Walker, P. L.,

Jr., Ed.; Marcel Dekker: New York, 1970; Vol. 6, p 191.
(3)Mattson, J. S.;Mark, H. B., Jr. Activated Carbon: Surface
Chemistry and Adsorption from Solution; Marcel Dekker; New York,

1971.
(4) Kinoshita, K. Carbon: Electrochemical and Physicochemical
Properties; Wiley: New York, 1988.
(5) Bansal, R. C.; Donnet, J. B.; Stoeckli, H. F.Actiue Carbon;Marcel
Dekker: New York, 1988.
(6) Wen, W. W.;
Sun,S. C. Sep. Sci. Technol. 1981, 16, 1492.
(7) Lau, A. C.; Furlong, D. N.; Healy, T. W.; Grieser, F. Colloids Surf:
1986, 18, 93.
(8) Fuerstenau, D. W.; Rosenbaum, J. M.; You, Y. S. Enegy Fuels
1988,2, 241.
(9) Solar, J. M.; Le6n y Le6n, C. A,; Osseo-Asare, K.; Radovic, L. R.
Carbon 1990,28,369.
(10) Le6n y Le6n, C. A.; Radovic, L. R. In Chemistry and Physics of
Carbon;Thrower, P. A., Ed.;Marcel Dekker, Inc.: New York, 1994;Vol.
24, p 213.

(11)Foger, K. In Catalysis. Science and Technology; Springer

Verlag: Heidelberg, 1984; Vol. 6, p 227.
(12) Rivera-Utrilla, J.; Ferro-Garcia, M. A. Adsorpt. Sci. Technol.
1986, 3, 293.
(13)Bautista-Toledo, I.; Rivera-Utrilla, J.; Ferro-Garcia, M. A,;
Moreno-Castilla, C. Carbon 1994, 32, 93.
(14) Moreno-Castilla. C.; Ferro-Garcia. M. A,: Rivera-Utrilla, J.:Jolv,
J. P. Energy Fuels 1994, 8 , 1233.
(15) Romdn-Martinez, M. C.; Cazorla-Amor6s, D.; Linares-Solano,
A,; Salinas-Martinez de Lecea, C.; Yamashita, H.; Anpo, M. Carbon
1996,33, 3.
(16) Szymanski, G.S.;Rychlicki, G. Pzrem. Chem. 1988, 67,315.
(17) Szymanski, G. S.; Rychlicki, G. Carbon 1991,29, 489.
(18) Szymanski, G. S.;Rychlicki, G. Carbon 1993,31,247.
(19) F'reiss,H.;Lischke,G.;Eckelt,R.;Miessner,
H.;Meyer, K. Carbon
1994,32, 587.
(20) Ldpez-Ramh, M. V.; Moreno-Castilla, C.; Rivera-Utrilla, J.;
Hidalgo-Alvarez, R. Carbon 1993, 31, 815.

0743-74631951241l-4386$09.00/0 0 1995 American Chemical Society

Langmuir, Vol. 11, No. 12, 1995 4387

Activated Carbon Surface Modifications

Table 1. Activation Time, Burn-Off,and Ash Content of
Activated Carbons Used
sample

activation time (h)

%BO

% ash

A1

1.0
2.5
5.0
8.0

18.6
25.0
42.0
61.4

0.04
0.05

A2
A3
A4

solid in the aqueous slurry increases, its pH takes an asymptotic

value which is equivalent to the point of zero net charge of the
solid. 10,22,23

Results

0.08

0.10

M) a t 353 K until dryness, and the residue was washed with

distilled water until all the nitrates were removed (asdetermined
with brucine). The treatment with HzOz was carried out with
1 g of carbon per 10 mL of concentrated H202 (9.8 M) 298 K in
a flask placed in a shaking bath for 48 h, and finally, the oxidation
with (NH&SzO* was carried out with a saturated solution of
this salt in HzS04 1M (1g of carbod10 mL of solution) a t 298
K for 48 h also in a flask placed in a shaking bath. After the
treatment, the samples were washed with distilled water until
absence of sulfates (as determined with BaC12) was reached.
All samples were characterized by Nz and C02 adsorption at
77 and 273 K, respectively. The BET equation was applied to
the N2 adsorption isotherms and the Dubinin-Astakhov equation
to the COz adsorption isotherms. Mercury porosimetry data were
obtained up to a final pressure of 2400 kg.cm-2 using an Autoscan
60. From this technique the surface area in pores greater than
7.5 nm in diameter (S,,t) was obtained, as well as the following
pore volumes: VZof pores with a diameter between 7.5 and 50
nm and V3 of pores with a diameter greater than 50 nm.
The transmission FTIR spectra of some selected samples were
obtained with a Nicolet 20 SXB spectrophotometer using pellets
of KBr containing about 0.5% carbon. These pellets were dried
overnight a t 393 K before the spectra were recorded.
Temperature-programmed desorption (TPD) was carried out
by heating the samples up to 1173 K in a He flow at a heating
rate of 20 K-min-l and recording the amounts of CO, C02, and
HzO with a quadrupole mass spectrometer as a function of
temperature as described elsewhere.21
Electrophoretic mobilities were determined with a Zeta-Sizer
IIe (from Malvern Instruments, England) as described elsewhere.20All experiments were carried out at 298 Kin acylindrical
cell. In each case, 150 mg of fresh sample with a particle size
below 0.063 mm was added to plastic bottles containing 65 mL
offreshly outgassed distilled water. Each bottle was hand-shaken
periodically for 1 day before the electrophoretic mobility measurements were carried out. The pH ofthe slurries was adjusted
by adding either HC1 or NaOH. Each pH data point was
measured with different portions (150 mg) of the same sample.
The electrophoretic mobility values, p,, were converted into
[-potential values according to Smoluchowski's equation (1)

The Dubinin-Astakhov equation (DA)has been applied

to the COz adsorption isotherms obtained on the activated
carbons. The DA equation reads as follow^:^^^^-^^

W=W,exp-

(Pi3"I
-

In this equation (31,W is the amount adsorbed a t a relative

pressure PIPo, WO is the micropore volume, A is the
differential molar work given by eq 4, p is the affinity

A = RT I
n
(
;
)

5 (mV) = 12.8,~~
(108m2/s*V)

(2)

(4)

coefficient, taken28as 0.46, and Eo is the characteristic

energy of adsorption in micropores. The CO2 liquid density
a t 273 K was taken28,29as 1.03 g/cm3.
The DA equation has three unknown parameters: WO,
Eo, and n. These were calculated by applying to the
experimental results a computer program which uses a n
iterative method, as explained e l s e ~ h e r e Values
. ~ ~ ~ ~of ~
WO,Eo, and n are compiled in Table 2.
The DA equation has been shown30 to fit the COZ
adsorption data on activated carbons better than the DR
equation ( n = 2 in eq 3) because of the well-accepted fact
that5~25-27,32-36
the value of n in eq 3 is not constant for all
activated carbons but depends on the percentage of burnoff (BO), Le., on the degree of activation of the sample.
Thus, the value of n decreases when the BO increases,
indicating that the heterogeneity of the microporosity
increases in the same way. Equation 3 has been
~ h o ~ n to~be, based
~ ~ on
, a~theoretical
~ , ~ ~ model involving
adsorption energies and their distribution. In this approach, n reflects the width of the energy distribution.
Dubinin and S t o e ~ k l ifound
~ , ~ ~that there is a linear
relationship between n and the characteristic variable A
of the Dubinin-Radushkevich-Stoeckli equation when
n takes a value between 1 and 2. Thus

n = 2.00 - 1.78 x 106A

where D is the dielectric constant and r] is the viscosity of the
liquid. For water, a t 298 K, eq 1becomes

(3)

(5)

Ais, according to Dubinin and S t o e ~ k l i the

, ~ ,half-width
~~
of the normalized Gaussian obtained for the distribution
of WOwith the structural parameter B of the activated
carbon. Once A is known, one can obtain the differential
curve for the micropore size distribution by applying eq
6, which was deduced by Dubinin and S t ~ e c k l i . ~ ~ ~ ~

(25)Dubinin, M. M.; Astakhov, V. A.Adv. Chem. Ser. 1970,102,69.

Mass titration of aqueous slurries of activated carbons was
carried out following a method described e l ~ e w h e r e , which
~ ~ . ~ ~ . ~ ~ (26)Dubinin, M. M.; Stoeckli, H. F. J. Colloid Interface Sci. 1980,
75,34.
is a modified version of a method proposed by Schwarz et al.24
(27)Dubinin, M.M.Carbon 1989,27,457.
For this purpose, 1 g of carbon was added to 10 mL of CO2-free
(28)Ismail, I. M. K. Carbon 1991,29, 119.
distilled water kept in a plastic bottle at 298 K and hand-shaken
(29)Rodriguez-Reinoso,F.; Linares-Solano, A. In Chemistry and
periodically for 1or 2 days until the pH of the slurry was stabilized.
Physics ofcarbon;Thrower,P. A,, Ed.; Marcel Dekker, Inc.: New York,
1989;VOl. 21,p 1.
After that, a small amount of COz-free distilled water was added
(30)Carrasco-Marin, F.: L6~ez-Ram6n.M. V.: Moreno-Castilla, C.
and again the pH of the slurry was measured after stabilization.
Langmuir 1993,9 , 2758.
This procedure was repeated until the pH ofthe slurry approched
(31)Salas-Peregrin, M. A.; Carrasco-Marin,F.; L6pez-GarzOn, F. J.;
that of the COz-free distilled water. As the concentration of the
Moreno-Castilla, C. Energy Fuels 1994,8, 239.
(32)Marsh, H.; Rand, B. J. Colloid Interface Sci. 1970,33,101.
(21)Ferro-Garcia, M. A.; Utrera-Hidalgo, E.; Rivera-Utrilla, J.;
(33)Freeman, E.M.; Siemieniewska,T.; Marsh, H.;Rand, B. Curbon
Moreno-Castilla, C.; Joly, J. P. Carbon 1993,31, 857.
1970,8,7.
(22)Le6n y Le6n, C. A,;Lizzio, A. A.; Radovic, L. R. In Proceedings
(34)Rand, B.J. Colloid Interface Sci. 1976,56,337.
of the International Carbon Conference; Paris, France, 1990,p 24.
(35)Huber, V.;Stoeckli,H. F.; Hourier, J. P. J . Colloid Interface Sci.
(23)Le6n y L e h , C. A.; Solar, J. M.; Calemma, V.; Radovic, L. R.
1978,67,195.
Carbon 1992,30,797.
(36)Finger, G.; Biilow, M. Carbon 1979,17, 87.
(24)Noh, J. S.;Schwarz, J. A. J. Colloid Interface Sci. 1989,130,
(37)Stoeckli, H.F. Carbon 1981,19, 325.
157.
(38)Stoeckli, H. F. Carbon 1990,28, 1.

4388 Langmuir, Vol. 11, No. 11, 1995

Moreno-Castilla et al.

Table 2. Surface Area and Porosity of the Activated Carbons

EO,

wo

WO(COrr),

sample

kJ/mol

cm3.g-1

cm3.g-1

A1
A1N
A10
A2
A2N
A20
A2S
A3
A3N
A30
A3S
A4
A4N
A40

22.8
22.6
22.6
20.9
20.9
20.9
20.8
16.3
14.2
16.3
15.8
13.4
12.4
13.5

1.96
1.86
1.95
1.74
1.68
1.78
1.83
1.63
1.12
1.39
1.42
1.41
1.09
1.29

0.270
0.259
0.278
0.330
0.275
0.324
0.368
0.520
0.414
0.580
0.576
0.723
0.561
0.737

0.270
0.259
0.278
0.330
0.275
0.324
0.368
0.509
0.335
0.567
0.562
0.497
0.204
0.503

sco*,

Sext,

SN2r
,2Lg-1

m2.g-l

712
683
733
870
725
854
970
1342
800
1495
1482
1310
538
1326

785
650
750
825
660
800
800
1290
740
1200
1060
1600
820
1400

v3,
cm3.g-1

v2 3

m2.g-1

cm3.g-1

18.3
30.7
21.5
22.5
34.6
29.8
21.1
38.2
41.4
42.4
40.5
65.9
34.8
59.1

0.063
0.097
0.071
0.078
0.105
0.089
0.068
0.119
0.128
0.132
0.127
0.225
0.114
0.200

0.068
0.089
0.061
0.078
0.116
0.115
0.074
0.155
0.164
0.166
0.131
0.309
0.194
0.243

7.00

6.02 .

5.m .

with LO as the pore width a t the maximum micropore

distribution curve.
The results obtained from application of eq 6 to C02
adsorption on the activated carbons are depicted in Figures
1-5. In the cases where the micropore distributions
extended beyond a micropore width of 2 nm, the limit of
the micropores, the distribution curves were integrated
up to this limit (2 nm) and the WOvalues so obtained will
be referred to in the text as WO(~,,~).
This method was used
in a previous paper.31 Wo(corr)therefore represents the
micropore volume of the activated carbon, which has been
convertedto an apparent surface area, Sco2given in Table2,
using a valuez9of 0.195 nm2 for the cross-sectional area
of the COZmolecule. Finally, the nitrogen surface area,
S N obtained
~,
from the BET equation (the cross-sectional
area of N2 a t 77 K was taken29as 0.162 nm2),the values
of Sext,Vz, and V3 are also shown in Table 2.
The FTIR spectra of samples A3, A3N, A30, and A3S
are shown in Figure 6, as a n example. Similar results
were obtained with the other oxidized samples. The band
assignments made in this work were based on previous
assignments in the l i t e r a t ~ r e . ~ > ~ ~ - ~ ~
It is well-known that when a carbonaceous material is
subjected to a program of increasing temperature, the
oxygen complexes desorb primarily as C02 and C0.3-5,43,44
Thus, COZresults from the decomposition of carboxyl,
anhydride (acidic), and lactonic groups, whereas CO
results from the decomposition of phenolic, carbonyl,
quinone, pyrone, and anhydride (acidic) groups. TPD
experiments were carried out with all samples, and the
amounts ofCO and C02 evolvedup to 1173Kare compiled
in Table 3. The C02, CO, and H2O desorption profiles are
depicted,in Figure 7 for samples A3N, A30, and A3S, as
an example. Desorption profiles with similar shapes were
obtained for the other oxidized samples. The nature of
the oxidizing agent can be seen to influence the shape of
these profiles.

6403.

3.m

A3

1.m .

0.m
0.75

Infrared Spectroscopy; Academic Press: New York, 1985;Vol. 4,p 169.

(40)Colthup, N. B.; Daly, L. H.; Wiberlay, S. E. Introduction to
Infrared Raman Spectroscopy, 3rd ed.; Academic Press: New York,
1989.
(41) Zawadzki, J.In Chemistry and Physics of Carbon; Thrower, P.
A,, Ed.; Marcel Dekker, Inc.: New York, 1989; Vol. 21, p 147.
(42) Fanning, P. E.; Vannice, M. A. Carbon 1993,31, 721.
(43) Tremblay, G.;Vastola, F.J.;Walker, P. L., Jr. Carbon 1978,I6,
35.
(44) Calo, J. M.; Hall, P. J. In Fundamental Issues in Control of
Carbon GasiFcationReactiuity;Lahaye, J.,Ehrburger, P., Eds.; NATO/
AS1 Series E192; Kluwer Academic: New York, 1991; p 329.

095

115

135

155

1.75

195

215

235

L (nm)

Figure 1. Micropore size distribution for activated carbons

Al-A4.
703

6.m

5.m

6 4.m
$3.03

2.m

1.m .

0.m
0 75

0 85

0 95

I 05

I15

I 25

I 35

t (nm)

Figure 2. Micropore size distribution for activated carbons

from the A1 series.
250

200

6 150

100
050

om
075

(39) Painter, P.; Starsinic, M.; Coleman, M. In Fourier Transform

A4

2.m

085

095

105

115

125

135

145

155

165

175

L (nm)

Figure 3. Micropore size distribution for activated carbons

from the A2 series.

The net surface charge of the activated carbons was

determined by both electrophoresis and mass titration
measurements. According to Corapcioglu and H ~ a n g , ~ ~
the electrophoretic mobility measurements detect the
(45) Corapcioglu, M. 0.;Huang, C. P. Carbon 1987,25,569.

Langmuir, Vol. 21, No. 11, 1995 4389

Activated Carbon Surface Modifications

Table 3. Amount of CO and C02 Evolved after Heating

up to 1173 K in He Flow

2m

A3

075

115

095

155

135

175

215

I95

235

t (nm)

Figure 4. Micropore size distribution for activated carbons

from the A3 series.
2.m 1

A4

150

co,

coz,

sample

mmol-g-l

mmol-g-l

CO/COZ

%Oa

A1
A1N
A10

0.41
5.43
1.47

0.04
3.18
0.49

10.3
1.7
3.0

0.8
18.9
3.9

A2
A2N
A20
A2S

0.32
5.74
1.38
3.38

0.02
3.56
0.40
1.27

16.0
1.6
3.5
2.7

0.6
20.6
3.5
9.5

A3
A3N
A30
A3S

0.24
5.93
1.62
3.40

0.01
3.34
0.45
1.14

24.0
1.8
3.6
3.0

0.4
20.2
4.0
9.1

A4
A4N
A40

0.26
6.22
2.02

0.00
3.27
0.38

1.9
5.3

0.4
20.4
4.4

"From the amount of CO and COZ.

5 100

51

050

000
1

12

I6

16

I4

22

24

28

26

L (nm)

Figure 5. Micropore size distribution for activated carbons

from the A4 series.
157q 1430

291:

34po

1720, I 'I

I ,

I1

,1260

300

400

500

600

700

800

900

lo00

1100

1200

900

lo00

1100

1200

T (K)
I'

51

...._.,.-...-...
0
300

400

500

600

700

800

T (K)
3600

3200

2SOo

2400

2000

1600

izoo

aoo

cm-l
Figure 6. FTIR spectra of activated carbonsfrom the A3 series.
potential a t the shear plane which is adjacent to the
external surface. Therefore these measurements yield
false information on the overall surface characteristics of
granular activated carbons, whereas mass titration measurements involve the transfer of Hf and OH- ions
between the bulk phase and the surface. These two types
of measurements have been used in carbon surface

Figure 7. TPD spectra for oxidized samples of the A3 series.

(-)A3N, (- . ) A 3 0 , and (-- -)A3S. (A)C02 desorptionprofiles,
(B)CO desorption profiles, and ( C ) H20 desorption profiles.

chemistry research to assess the surface charge distribution.9J0,22,23

Thus, the pH of the wetted activated carbon
particles, determined by mass titration, can be taken as
the point of zero charge (pzc), i.e., the pH value below
which all the surface of the carbon accessible to water will
be, on average, positively charged. The isoelectric point
(iep)is determined by electrophoresismeasurements, and
it is the pH value below which the external surface of the

Moreno-Castilla et al.

4390 Langmuir, Vol. 11, No. 11, 1995

activated carbon will be, on average, positively charged.

Therefore, the difference between the pzc and the iep has
been taken as a measurement of the surface charge
d i s t r i b ~ t i o n . Thus,
~ ~ ~ ~when
~ * ~the
~ pzc value is higher
than the iep value, the internal surface of the carbon
particles will be more positively charged than the external
surface, and vice versa.10,2z,23
The 5-potential-pH curves ofthe activated carbons A l A4 and of the oxidized samples from activated carbons A2
and A3 are shown in Figure 8. From these curves, the
pHiepvalues were obtained, which are compiled in Table
4, together with those of the pH,,,.

Discussion
Surface areas and pore textures of all samples used are
given in Table 2 and Figures 1-5, which show the
calculated micropore size distributions. From the results
obtained on the surface characterization the following
points can be deduced. First, for the original activated
carbons (from A1 to A4), the increase in the degree of
activation (%BO in Table 1) brings about an increase in
the S Nand
~ Sextvalues. Sco2also increases but only up to
about 40%BO (sample A3). The mean micropore width,
Lo,is shifted to higher values (Figure 11, and there is a n
increase in Vz and V3 values. All these results indicate
that an increase in %BO produces an increase in micropore
volume up to 40% BO, a widening of the microporosity,
and also a n increase in the meso- and macroporosity of
the samples. At higher BO, sample A4, the microporosity
is centered in the supermicropores (higher limit of
micropores) and in the lower limit ofthe mesopores which
makes this the only sample in which S N>~Sco2,due to
capillary condensation of Nz into the supermicropores. In
the other cases, Sco2 is always similar or quite close to

."I

>ti

sc02

SN2.

Values of Sex,,V Z and

,
V3 increase with respect to the
original activated carbon up to sample A3N (40% BO).
Therefore, up to this degree of activation the destruction
and widening of the micropores by oxidation causes an
and the volume
increase in the external surface area (SeXt)
ofthe meso- and macropores. However, for a higher degree
of activation (sample A4 with 61.4%BO) the nitric acid
treatment (sample A4N) causes a decrease in the Sext,
VZ,
and V, values; so in this case, it appears that the mesoand macroporous network is also partially destroyed, since
as explained before a n increase in the degree of activation
makes the pore walls thinner.
Finally, the treatments with hydrogen peroxide and
ammonium peroxydisulfate yielded fairly similar results
but their effects on the pore texture and surface area of
the original samples were quite different from those
observed with nitric acid. The samples obtained after
both treatments have practically the same surface area

A2

-30

4r -40

S N Z*

Second, when the original activated carbons were

oxidized with HN03 there was a decrease of both S Nand
~
Sco2 with regard to the respective values of the original
samples, and the decrease in these values became more
pronounced with increasing degree of activation of the
original activated carbon, i.e., from A1 to A4. Since a n
increase in the degree of activation makes the pore walls
thinner and, thus, more easily destroyed by the HN03
treatment, this results in a widening of the microporosity
and consequently a diminishing ofthe Sco2and S Nvalues.
~
This is especially significant with sample A4N, which
presents the largest decrease in the Sco2 value, with all
the micropore volume corresponding to the upper limit of
the micropores, Le., the supermicropores (see Figure 51,
showing again capillary condensation of N2 at 77 Kmaking

- 10
o

-801
-

A2S

2 7

60

PH
2o
10

- 10
-20

-30

A3

-40
-4Oi
I

-50

-- 560
01

A3N

-70

A3S
A30

-701

Figure 8. <-potential-pH curves for different activated

carbons.

and pore texture as the original activated carbons, and in

some cases (samples A2S, A30, and A3S) ScoZ =- S N ~
indicating that there are some constrictions due to the
fixation of oxygen surface complexes a t the entry of the
micropores so that there is restricted diffusion of NPa t 77
K.29

Langmuir, Vol. 11, No. 11, 1995 4391

Activated Carbon Surface Modifications

Table 4. pHs of the Point of Zero Charge and Isoelectric
Point of the Activated Carbons
A1
A1N
A10

9.9
2.4
4.8

1.8

A2S

10.4
2.3
4.8
1.4

2.4
nd
2.3
nd

A3
A3N
A30
A3S

10.6
2.4
4.9
1.3

3.4
nd
2.1
nd

A4
A4N
A40

11.1
2.3
4.7

4.1

A2
A2N

A20

nd = not detected.

The surface chemistry of the samples was studied by

FTIR, TPD, electrokinetic and mass titration measurements. Infrared spectroscopy plays a n important part in
the investigations of carbon materials, although this
technique presents some difficulties that stem from their
unique physicochemical proper tie^.^^,^^!^^
The FTIR spectra (Figure 6) of the original sample A3
presents bands a t 2912 and 2848 cm-' due to asymmetric
and symmetric C-H stretching vibrations in aliphatic
CH, CH2, and CH3. The band at 1720cm-l can be assigned
to the C-0 stretching vibration from ketones, aldehydes,
or carboxyl groups, although probably the latter should
be excluded due to the basic character of this carbon (see
Table 4) and the very small amount of CO2 evolved during
the heat treatment up to 1173 K (see Table 3). The band
in the region of 1600cm-' (at 1570cm-l) has been observed
by many authors and has not been interpreted unequivocally. The band is assigned to aromatic ring stretching
or highly conjugated hydrogen-bonded C=0.39,41,42The
bands in the 1000-1200 cm-I region are difficult to assign
because there is superposition of a number of broad
overlapping bands. They cannot therefore be described
in terms of simple motion of specific functional groups or
chemical bonds.39
New bands and modifications of others appear in the
FTIR spectrum of sample A3N. Thus, there is a large
increase in the intensity of the band a t 1720 cm-l, the
C=O stretchingvibration, that together with that a t about
3400 cm-', due to the 0 - H stretching vibration, can be
assigned to carboxyl acid groups. The band a t 1460 cm-I
can be attributed to 0 - H deformation vibrations of
carboxyl groups and/or to the existence of carboxylcarbonate structure^.^^,^^ A new band appears a t 1538
cm-l, and although there is uncertainty in this region as
mentioned above, this might indicate the presence of
nitrate and nitro surface group^.^^,^^ Actually, nitrogen
contents determined by elemental analysis for samples
A3 and A3N were 0.51 and 0.76%, respectively.
The FTIR spectra of samples A30 and A3S are quite
similar to each other but somewhat different from that of
sample A3N. Thus, instead of a single band a t 1720 cm-'
a doublet now appears a t 1720 and 1690 cm-l. The small
band a t 1460 cm-l now appears with a higher intensity
a t 1430 cm-', and a new small band appears a t 1260 cm-'
as a shoulder on the large band centered in the region of
1150-1110 cm-l. These bands can be interpreted as follows: the 1720 cm-l together with those a t 1430 and 3400
cm-l are assigned to carboxyl groups, the band a t 1430
cm-l can also be due to carboxyl-carbonates structures,
(46) Zawadzki, J. Carbon 1980, 18, 281.

and the band a t 1690 cm-' near to that of 1720 cm-l is

probably due to ketone groups.47 Finally the band at 1260
cm-' can be assigned to C-0-C vibrations in ether

structure^.^^
In conclusion, the FTIR technique detected, from a
qualitative point of view, that the treatment with HN03
fixed carboxyl groups and probably nitro and nitrate
groups, whereas the treatments with H202 and (NH4)&08
fixed carboxyl, ketone, and ether groups. These results
are most likely due to the drastic nature of the treatment
with HN03, which partially broke the porous texture of
the activated carbons compared with the other two
oxidizing treaments, which probably does not favor the
formation of ketone and ether groups.
Desorption profiles of C02, CO, and H2O are depicted
in Figure 7 for the oxidized samples from activated carbon
A3. The C02 desorption profile (Figure 7a) of sample A3N
is similar to that of other activated carbons oxidized with
HN03.14,48-51The maximum rate of COZ evolution is a t
575 K, with a shoulder a t around 700 K that tails up to
around 1200 K. This profile is indicative ofthe occurrence
of chemically different complexes and/or the same oxygen
complex existing on energetically different ~ i t e s . ~ ~
Thus,
-~O
anhydride groups which can be formed from two adjacent
carboxylgroups during the TPD runs are more stable than
carboxylic acid groups and, therefore, will evolve CO2 and
CO above 573 K. Possible evidence for the formation of
anhydride groups during the experiments is that the H 2 0
profiles (Figure 7c) present a peak a t around 500 K
inmediately after evolution of physisorbed water. The
COz desorption profile ofA3O sample presents a maximum
desorption rate at around 575 K, as for the A3N sample,
and a tail up to around 1200 Kwith no significant shoulder
a t 700 K. In the case of sample A3S, the maximum
desorption rate of the C02 profile is shifted to a lower
temperature, 525 K, than in the above cases, which
indicates that the thermal stability of the carboxyl groups
present on this sample is lower than that on samples A3N
and A30. After this maximum there is a slight shoulder
in the region of 700 K, possibly due to the presence of
carboxyl anhydride groups. These groups will be present
in a higher proportion in sample A3N due to the fact that
the HN03 treatment fixed the highest amount of oxygen
that evolved as C02. The carboxyl anhydride groups will
also yield CO, probably in the same temperature range
as C 0 2 .
With regard to the CO profiles (see Figure 7b), sample
A3N has a small maximum a t 530 K, which is shifted to
500 Kin sample A3S and does not appear in sample A30.
Sample A3N has a shoulder a t around 775 K (which can
correspond to the shoulder a t around 700 K in the COz
desorption profile due to anhydride carboxyl groups), and
the maximum desorption rate is a t 1000 K, whereas
sample A3S has a peak a t this temperature and the
maximum desorption rate is shifted to 1130 K. The CO
profile of sample A30 does not present any peak and
increases continuously from 600 up to 1170 K.
The amounts of C02 and CO evolved from all samples
are given in Table 3 and the results obtained indicate
that (i) The amount of oxygen fixed after the different
(47) Starsinic, M.;Taylor, R. L.; Walker, P. L., Jr. Carbon 1983,21,
69.
(48)Linares-Solano, A.; Salinas-Martinez de Lecea, C.; CazorlaAmor&, D.;Joly, J. P. In Fundamental Issues in Control of Carbon
Gasification Reactiuity; Lahaye, J., Ehrburger, P., Eds.; NATO/ASI
Series E192; Kluwer Academic: New York, 1991; p 409.
(49) Cazorla-Amor6s, D.;Linares-Solano, A,; Joly, J. P.; SalinasMartinez de Lecea, C. Catal. Today 1991,9, 219.
(50) Otake, Y.; Jenkins, R. G. Carbon 1993, 31, 109.
(51)Vinke, P.; Van der Eijk, M.; Verbree, M.; Voskamp, F.; Van
Bekkum, H. Carbon 1994,32, 675.

Moreno-Castilla et al.

4392 Langmuir, Vol. 11, No. 11, 1995

oxidizing treatments follows the sequence HN03 x(NHJzSz08 > HzOz, and for a given oxidizing agent, the
amount of oxygen fixed does not seem to depend on the
degree of activation of the sample. (ii) As deduced from
the CO/COz ratio values, the three oxidizing treatments
do not fix the same proportion of oxygen groups evolving
as CO and COZ;followingthis ratio the contrary sequence
given above. This could be in agreement with the
appearance ofketone and ether groups on the FTIRspectra
of the oxidized samples with hydrogen peroxide and
ammonium peroxydisulfate.
Results obtained from electrophoretic and mass titration
measurements are given in Figure 8 and Table 4 and
indicate that. (i)For activated carbons Al-A4 both pH,,,
and pH,,, increase with the degree of activation or
percentage of burn-off of the activated carbon, which
means, as reported elsewhere,20 that the increase in
activation also increases the surface basicity of the samples
making it less negatively charged. In these cases, as pH,,,
> pHi,,, the internal surface is always more basic than
the external surface. However, the difference between
these two values decreases from sample AI to A4 due to
the opening of the porosity of the samples. After the three
oxidizing treatments of samples A2 and A3, the pH,,,
values decreased to a greater extent than the pHiepvalues,
although pH,,, was still higher than pHi,,. This means
that the external surface was somewhat more acidic than
the internal surface, but the fixation of the oxygen groups
with an acid character essentially took place on the
internal surface. (ii) The three oxidizing treatments
decreased the pH,,, of the activated carbons to the acid
range, and for the same treatment pH,,, seems to be
independent ofthe degree of activation of the sample. The
pH,,, values after the three oxidizing treatments follow
the order HZOZ> HN03 > (NH4)zSzOs. This sequence
indicates that, after treatment with ammonium peroxydisulfate, the surface was more acidic than after the
treatment with HN03, in spite of the fact that the latter
treatment (samples A2N and A3N) fured a larger number
of oxygen complexes that evolved COz than the former
(samples A2S and A3S). Moreover, the <-potential of
sample A2S is more negative than that of sample A2N in
the pH range of approximately 2-9 and within this range,

samples A3S and A3N show fairly close <-potentialvalues.

These results are quite important and indicate that the
treatment with (NH4)zSzOefavors the formation of stronger acid groups than the treatment with HN03.

Conclusions
In this study, the surface modifications of activated
carbons after their treatments with HN03, HzOz, and
(NH4)&08 have been analyzed. Treatment with HN03
causes a decrease in surface area and in the microporosity
ofthe activated carbons, and these effects are more drastic
with increasing degrees of activation of the original
activated carbon. However, the treatments with H2Oz
and (NH4)2S208were not as drastic, and in some samples
the oxygen surface complexes were fixed a t the entrance
of the micropores so that there was some restricted
diffusion of Nz a t 77 K.
From FTIR, carboxyl groups and probably carboxylcarbonate structures as well as nitro and nitrate groups
were detected after the HN03 treatment, whereas after
the HzOz and (NH4)&08 treatments carboxyl, ketone,
and ether groups and probably carboxyl-carbonate structures were detected. The amount of oxygen fixed after
these treatments followed the sequence HN03 (NH4)ZSz08 > HzOz, and the CO/COZ ratio followed a contrary
sequence, probably due to the higher proportion of ketone
and ether groups with respect to COz-liberatingcomplexes
on samples treated with peroxides.
Finally, the most important is that treatment with
(NH4)zSzOefixed stronger acid groups than treatment with
HN03 in spite of the fact that this latter treatment fixed
the largest amount of oxygen surface complexes evolving
as COZ. There is the added advantage that, unlike the
HN03 treatment, treatment with (NH4)~S208does not
modify the surface area and pore texture of the original
activated carbon to a great extent. These findings can be
important when using activated carbons in reactions
catalyzed by acidic solids.

Acknowledgment. We thank DGICYT (Projects no.

AMB 92-1032 and PB94-0754) for the financial support.
LA950248F