Fumed Oxides Modified Due To Pyrolysis Of Cyclohexene

Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124
www.elsevier.com/locate/colsurfa

Fumed oxides modified due to pyrolysis of cyclohexene
V.M. Gun’ko a,*, R. Leboda b, V.I. Zarko a, J. Skubiszewska-Zieba b,
W. Grzegorczyk b, E.M. Pakhlov a, E.F. Voronin a, O. Seledets b,
E. Chibowski b
a

b

Institute of Surface Chemistry, 31 Prospect Nauki, Kiev 03680, Ukraine
Faculty of Chemistry, Maria Curie-Sklodowska University, Lublin 20031, Poland
Received 25 May 2002; accepted 5 December 2002

Abstract
Mixed oxides such as fumed X/silica (X 0/Al2O3 (AS), TiO2 (TS)) and CVD-TiO2/fumed silica, initial and covered by
carbon deposit formed on cyclohexene pyrolysis, were characterized by means of carbonization kinetics, nitrogen
adsorption Á/desorption, water adsorption, and electrophoresis methods. Catalytic capability of AS (per gram of the
adsorbent) in cyclohexene pyrolysis (973 K for 4 h) is greater at a low concentration of alumina, but for fumed TS, it is
maximal at large CTiO2. Concentration of pyrocarbon formed on AS or TS is 2 Á/4 times larger than that on silica under
the same pyrolysis conditions, and fumed TS can stronger catalyze pyrolysis than CVD-titania/silica does. A marked
impact of the nature of X/SiO2 on cyclohexene pyrolysis results in variations of the structural, adsorptive, and
electrokinetic properties of C/X/SiO2. Alterations in the pore size distributions (main maxima at Rp between 0.5 and 7
nm) can be due to the availability of various channels between primary particles (5 Á/50 nm) packed in relatively dense
aggregates and between aggregates in more loose agglomerates of aggregates differently filled by grafted pyrocarbon.
The dependencies of z potential of C/X/SiO2 particles on pH are closer to that of C/SiO2 than X/SiO2, i.e. carbon
deposit covers mainly the X/SiO2 interface and X phase possessing catalytic activity in cyclohexene pyrolysis.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: Fumed mixed oxides; Pyrolysis kinetics; Carbon/mixed oxides; Nitrogen adsorption; Pore size distribution; z Potential

1. Introduction

* Corresponding author. Present address: Laboratory of
Surface Electro-Physics, Institute of Surface Chemistry,
National Academy of Sciences of Ukraine, 17 General
Naumov Street, Kiev 03164, Ukraine. Tel.: '/380-44-4449627; fax: '/380-44-424-3567.
E-mail address: (V.M. Gun’ko).

Individual and mixed X/SiO2 oxides (e.g. silica,
alumina/silica, titania/silica, etc.) of different origin covered by carbon deposit of the pre-graphite
structure prepared by means of pyrolysis of
organic precursors can be used as adsorbents,
fillers, etc. [1 Á/10]. Features of C/X/SiO2 materials
depend not only on the carbon concentration (CC),

0927-7757/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0927-7757(02)00606-4


104

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

pyrolysis temperature, organic precursor structure,
and silica substrate type, but also on the concentration (CX) and distribution of the second oxide
in X/SiO2 (e.g. X 0/Al2O3, TiO2) possessing catalytic activity in pyrolysis and related reactions.
These properties can also alter due to variations in
the substrate topology (porous, nonporous but
highly dispersed with large or small polydispersity,
different types of particle packing in aggregates

and agglomerates, broad or narrow particle size
distributions, etc.) and technique of X/SiO2 synthesis [1 Á/15]. Many of the important characteristics
of oxide surfaces covered by carbon deposit
(porosity, adsorptive ability, acidity of surface
sites, etc.) can be studied upon adsorption of polar
and nonpolar probe molecules [2 Á/11]. Various
pyrocarbon/silicas (CS) synthesized by pyrolysis of
organic and organometallic precursors on fumed
silica, silica gel, X/silica (X 0/TiO2, ZnO, etc.) as
substrates and interaction of CS with different
adsorbates, as well as coking of mixed oxide
catalysts due to side processes were investigated
previously [2 Á/6,11 Á/15]. It should be noted that the
size of basal planes in pyrocarbon prepared using
organic or organometallic precursors with a relatively small molecular size (e.g. methylene chloride, cyclohexene, acetylacetonates of Ti, Co, Ni,
Cr, Zn, Zr, etc.) at 750 Á/950 K can reach only
several nanometers [16,17], and pyrocarbon possesses a relatively low specific surface area, e.g.
SBET of carbon deposit per se is approximately
100 Á/120 m2 g(1 at CC !/20 wt.% in C/fumed silica
[3]. Additionally, pyrocarbon can form mainly on
the outer surface of porous silica particles at CC !/
10 wt.%, but at low CC, carbon deposit can cover
mainly the pore walls [7 Á/10,18]; nevertheless, a
significant portion of the oxide surface can remain
practically pure and accessible for adsorbate
molecules over a broad CC range. But in the case
of mixed oxide substrates (e.g. TiO2/SiO2) or
pyrolysis of organometallic precursors on silica
gel (resulting in the formation of new oxide,
silicate or even pure metal clusters), the pyrocarbon distribution depends on the allocation of

active sites (Brønsted and Lewis acid sites) catalyzing carbonization [2 Á/6].
Previously, we studied such properties or characteristics of X/SiO2 (X 0/Al2O3, TiO2, Al2O3/

TiO2 (AST)) as the X phase structure, particle
morphology, concentration and distribution of
Brønsted and Lewis acid sites, as well as molecular
and dissociative adsorption of water, surface
modification by organic and organosilicon compounds, etc. [2 Á/6,11,19Á/26]. Clearly, the appearance of an X phase (which can catalyze pyrolysis
and related reactions of organics) in X/SiO2 can
be responsible for significant changes in the
characteristics not only of oxide surfaces itself
but also of carbon deposit in C/X/SiO2. Elucidation of this impact is useful in order to control the
physicochemical properties of hybrid C/X/SiO2
materials, and controlled variations in CC and
CX, and synthetic conditions allow one to prepare
C/X/SiO2 adsorbents with desirable porosity, polarity of surface, active surface site distribution,
etc.
Despite numerous investigations of highly dispersed or porous oxides covered by carbon
deposit, many questions related to their structures
(especially for mixed fumed oxides) after carbonization of organics and their interactions with such
adsorbates as water (e.g. its associative desorption
and dissociative adsorption from air or alterations
in the z potential and particle aggregation in
aqueous suspensions, as adsorbents are frequently
utilized in aqueous media) still remain unstudied
or unclear. Therefore, the aim of this work was to
study carbonization kinetics, structural and adsorptive characteristics, and z potential of carbon/
X/fumed silica (X 0/Al2O3, TiO2) materials
synthesized by using cyclohexene pyrolysis depending on the concentration of X oxides in fumed
X/SiO2 or silica covered by titania using a

chemical vapor deposition (CVD) technique. A
choice of C6H10 as a precursor was conditioned by
a relatively small size of its molecules (to
provide their effective interaction with active sites
of the X/SiO2 surfaces in channels of aggregates of
primary particles) and the absence of Cl (e.g.
CH2Cl2 was used previously to prepare carbon
deposit on fumed silica [3]), which can effectively
interact with Ti or Al atoms in mixed oxides at
high temperatures that can result in uncontrolled
and undesirable changes in the oxide matrix
structure.


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

2. Experimental
2.1. Materials
Highly dispersed fumed silica (Aerosil A-175
and A-300 having different specific surface areas)
and fumed X/SiO2 (X 0/Al2O3 (AS), TiO2 (TS)) at
different concentrations (CX) of X oxide phase
(Tables 1 and 2) (pilot plant at the Institute of
Surface Chemistry, Kalush, Ukraine) were used as
the initial materials to prepare C/X/SiO2 adsorbents. Additionally, TS and AST samples at
CTiO2 ]/65 wt.% (SBET between 30 and 74 m2
g(1) were used for the comparison of kinetics of
carbonization reaction. CVD-titania/fumed silica
(Institute of Surface Chemistry, Kiev) synthesized
by means of CVD technique using chemisorption

of TiCl4 on fumed silica A-300 as a substrate, and
then hydrolyzed at 423 K was described in detail
elsewhere [21]. The titania phase in fumed TS is
crystalline and consists of a blend of anatase
(major portion) and rutile, but in the case of
CVD-TS, titania particles represent only anatase,
whose phase transition to rutile does not occur
even at 1100 K for 2 h due to the silica impact [21].
Alumina in AS is totally amorphous as well as
pure fumed silica or silica in fumed AS and TS
[19 Á/25].
Carbosil (CS) samples (Maria Curie-Sklodowska University, Lublin) were synthesized using
cyclohexene pyrolysis on oxide substrates at 973 K
for 4 h. The technique of CS sample preparation
was described in detail elsewhere [1,26 Á/28].
2.2. Kinetics of carbonization
Pyrocarbon deposition on oxide samples was
carried out using controlled cyclohexene pyrolysis
under isothermal conditions (973 K for 4 h) in a
flow reactor (coupled with the scales) in the He'/
5% C6H10 stream. To reduce the impact of external
diffusion phenomena on dynamics of carbon
deposit formation, the low-weighted amounts of
samples ( :/100 mg) and flow velocity of the
reaction mixture of 63 dm3 h (1 were used.
Kinetics of carbon grafting was controlled by
means of the gravimetric method.

105

2.3. Nitrogen adsorption
Nitrogen (spectra analyzed grade) adsorption Á/
desorption isotherms were recorded at 77.35 K
using a Micromeritics ASAP 2010 (V-2.00) adsorption analyzer. The structural characteristics of
studied materials calculated on the basis of nitrogen adsorptionÁ/desorption are summarized in
Tables 1 Á/3.
2.4. Electrophoresis
Electrophoretic investigations were performed
using a ZetaPlus (Brookhaven Instruments) z
potential apparatus. Deionized distilled water
(pH 6.95), and 0.2, 1, and 2.5 g of solids per liter
of water were utilized to prepare the suspensions,
which were then treated with an ultrasonic bath
for 3 Á/3.5 h. The pH values measured by an OP208/1 precision digital pH-meter were adjusted by
the addition of 0.1 M HCl or NaOH solutions.
The suspension salinity was changed using NaCl
solution.
2.5. Computation
The specific surface area (SBET) was calculated
according to the standard BET method [29,30]
using adsorption data at relative pressures (p/p0)
between 0.06 and 0.2, where p and p0 denote the
equilibrium and saturation pressures of nitrogen,
respectively. The pore volume VBJH was evaluated
from adsorption data using the BarrettÁ/Joyner Á/
Halenda (BJH) method [31]; then, VBJH and SBET
were used to estimate the average pore diameter
(DBJH). The specific surface area of mesopores
(SK) was calculated using the Kiselev equation
[32]. Also, the mesopore volume distributions

dVmes/dRp were calculated with desorption data
according to the improved theory of capillary
condensation Á/evaporation (with Broekhoff Á/
deBoer Á/Dollimore Á/Heal Á/Dubinin Á/Ulin corrections to the Kelvin equation) [33 Á/35], which links
the adsorbed layer thickness (t) and the radius
(Rp !/1 nm) of filled pores (emptied on desorption), using local isotherm approximation (LIA)
with the program package described in detail
elsewhere [36 Á/38].


106

Sample
A-300
C/A300
A-175
C/A175
AS1
C/AS1
AS3
C/AS3
AS8
C/AS8
AS23
C/AS23

CAl2O3 (wt.%) CC (wt.%) SBET (m2 g (1) SK (m2 g (1) SDS (m2 g (1) VBJH (cm3 g (1) DBJH (nm) DAJ1, p /p0 B/0.1 DAJ2, p /p0 B/0.85 DFRDA, p /p0 B/0.1

1.3
3

8
23

Á/
8.5

308
378

244
320

43
44

0.65
0.96

7.5
8.6

2.367
2.300

2.598
2.574

2.196
2.171

Á/
8.3

185
203

168
177

25
25

0.50
0.56

9.2
9.4

2.328
2.313

2.569
2.572

2.186
2.174

Á/
31.8
Á/

27.5
Á/
31.8
Á/
25.6

203
169
185
199
303
195
347
321

158
140
146
156
241
160
283
266

22
25
29
29
42
30

45
43

0.42
0.38
0.39
0.43
0.66
0.45
0.80
0.77

7.3
7.9
7.5
7.5
7.6
8.1
8.0
8.2

2.406
2.235
2.332
2.311
2.355
2.323
2.348
2.319

2.611
2.564
2.588
2.586
2.592
2.581
2.588
2.579

2.183
2.144
2.186
2.173
2.195
2.180
2.194
2.176

SK is the specific surface area of mesopores determined by Kiselev method, SDS is the specific surface area of micropores determined by Dubinin Á/Stoeckli method, VBJH
and DBJH are the total pore volume and the pore diameter, respectively, estimated according to the BJH method, DAJ is the fractal dimension estimated using Avnir Á/
Jaroniec method, DFRDA is the fractal dimension estimated using the fractal analogue of the Dubinin Á/Astakhov equation.

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Table 1
Structural parameters of initial fumed silica, AS, C/SiO2, and C/AS


Sample CTiO2 (wt.%) CC (wt.%) SBET (m2 g (1) SK (m2 g (1) SDS (m2 g (1) VBJH (cm3 g (1) DBJH (nm) DAJ1, p /p0 B/0.1 DAJ2, p /p0 B/0.85 DFRDA, p /p0 B/0.1
TS2

C/TS2
TS5
C/TS5
TS17
C/TS17
TS22
C/TS22
TS33
C/TS33
TS9
C/TS9
TS20
C/TS20
TS36
C/TS36

1.7a
5a
17a
22a
33a
9b
20b
36b

Á/
20.1
Á/
20.0
Á/

17.4
Á/
18.7
Á/
23.1
Á/
26.3
Á/
17.3
Á/
29.7

313
281
306
268
286
274
247
215
215
166
235
184
84
76
114
83

264

247
252
235
242
233
214
189
188
149
198
152
65
60
87
71

42
34
42
33
41
35
32
30
29
24
32
26
17
15

20
15

0.71
0.77
0.71
0.74
0.69
0.72
0.63
0.60
0.55
0.47
0.57
0.43
0.17
0.16
0.25
0.22

8.1
9.3
8.2
9.4
8.4
9.0
8.7
9.6
9.2
9.8

8.6
8.2
7.3
7.4
8.1
9.2

2.354
2.317
2.361
2.340
2.350
2.320
2.375
2.329
2.358
2.336
2.332
2.304
2.353
2.313
2.413
2.324

2.584
2.570
2.591
2.577
2.582
2.577

2.585
2.573
2.581
2.574
2.583
2.578
2.598
2.589
2.615
2.580

2.191
2.178
2.195
2.187
2.188
2.180
2.197
2.181
2.191
2.185
2.194
2.174
2.192
2.177
2.212
2.178

SK is the specific surface area of mesopores determined by Kiselev method; SDS is the specific surface area of micropores determined by Dubinin Á/Stoeckli method; VBJH
and DBJH are the total pore volume and the pore diameter, respectively, estimated according to the BJH method; DAJ is the fractal dimension estimated using Avnir Á/

Jaroniec method, DFRDA is the fractal dimension estimated using the fractal analogue of the Dubinin Á/Astakhov equation.
a
CVD-TS.
b
Fumed TS.

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Table 2
Structural parameters of initial fumed and CVD-titania/silica and C/TiO2/SiO2

107


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

108

Table 3
Mesopore specific surface areas and fractal dimensions of oxide
and carbon/oxide adsorbents
Adsorbent

SdB
(m2 g (1)

Svt
(m2 g (1)

DFRDA, xmax 0/

3.5 nm

A-300
C/A-300
AS23
C/AS23
TS33
C/TS33
TS36
C/TS36

255
328
294
271
195
152
89
71

249
314
284
265
166
123
89
66

2.476

2.512
2.358
2.424
2.494
2.460
2.493
2.477

SdB and Svt are the specific surface area of mesopores
calculated using the t or v ,t methods, respectively. DFRDA is
the fractal dimension calculated using the fractal analogue of
the Dubinin Á/Astakhov equation at xmax 0/3.5 nm.

Calculation of fractal dimension (D ) [39] was
performed using adsorption data with the linear
relationship between adsorption (ln U ) and nitrogen pressure (ln ln(p0/p )) [40] at p/p0 5/0.1 with no
capillary condensation in mesopores, and at p/
p0 B/0.85 with a marked capillary condensation. It
is pertinent to note that deviations of the t-curves
from the standard t-curve for nitrogen [29] are
observed for all the studied samples only at p/p0 !/
0.75, for which ln U (U is the reduced adsorption)
is a nearly linear function of ln ln(p0/p ), as
determination coefficient:/0.98 Á/0.99 at 0.001 B/
p /p0 B/0.1 and :/0.80 Á/0.84 at p /p0,max 0/0.85, i.e.
the larger the maximal p/p0, the worse is the fit on
linearization, i.e. DAJ value is less uniquely determined. The fractal analogue of the DubininÁ/
Astakhov equation (FRDA) [41] was utilized to
estimate fractal dimension DFRDA at p/p0 5/0.1
and the pore half-width range between xmin 0/0.2

nm and xmax 0/2.5 nm corresponding to main
maxima in the pore size distributions of fumed
oxides linked to micropores (Rp B/1 nm) and the
narrowest mesopores at Rp between 1 and 2.5 nm.
Contribution of micropores was estimated using
the Dubinin Á/Stoeckli (DS) equation [42] with
consideration for adsorption in mesopores using
SK for corresponding corrections. It should be
recognized that the SK values can be overestimated, and the greater the capillary effect (or
larger hysteresis loop in adsorption Á/desorption

isotherms), the larger is the SK overestimation [4 Á/
6,43]. Consequently, microporosity can be underestimated when SK is used for corrections of
isotherms to calculate the micropore parameters.
On the other hand, calculations of the micropore
parameters without such a correction result in an
enhancement of SDS by a factor 2Á/3; however,
VDS increases to a lesser extent. An increase in the
upper limit of Rp on integration for SDS up to 2
nm results in an enhancement of SDS by half (with
consideration of adsorption in mesopores using
SK). However, calculations without correction on
adsorption in mesopores give the SDS values of
190Á/200 m2 g(1 (Rp between 0.2 and 2 nm) and
100Á/130 m2 g(1 (Rp between 0.2 and 1 nm) for A300 or C/A-300, respectively. Consequently, fumed
oxide microporosity expanded to Rp B/2 nm can
correspond to 60Á/65% of a total surface area, i.e.
the pore character of studied adsorbents can be
close to that of microporous adsorbents. This
circumstance allows us to utilize the FRDA

equation [41] at xmax 0/2.5 (Tables 1 and 2) or
3.5 nm (Table 3). To compare the results summarized in Tables 1 and 2, the structural parameters of
some adsorbents were calculated using t or v,t
methods [29,44,45], and obtained values of the
mesopore specific surface area (SdB and Svt, Table
3) are close to SK (Tables 1 and 2). One can
conclude that calculations of the micropore and
mesopore parameters (Tables 1 Á/3) using series of
adsorption equations give results, which differ
relatively slightly and typically are within the
accuracy of the methods used. It should be pointed
out that some equation parameters depend on the
nature of adsorbents, but some others are independent [29]; nevertheless, the calculated structural
characteristics are closely related (Tables 1 Á/3) that
confirms the reliability of the obtained results and
the possibility to use the mentioned equations to
calculate the structural parameters of hybrid
adsorbents.
To compute the pore size distribution f(Rp), the
modified DubininÁ/Astakhov equation [46,47]
ln a0

  ni 
L  
X
W
Ax
ln f i (
kb
W

i
0

(1)

(where L 0/3 or 4 [44]; Wi is the current and W0


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

the limiting adsorption of vapor; A 0/RgT ln(p0/p)
is the differential molar work equal (with inverse
sign) to the variation in the Gibbs free energy,
where Rg denotes the gas constant; ni is the
equation parameter (typically varied between 0.5
and 4); bN2 0/0.33; x is the micropore half-width
or pore radius of mesopores) was used as a local
isotherm Ul in the overall adsorption isotherm in
the form of Fredholm integral equation of the first
kind:
U(T; p)0

g

xmax

Ul (T; p; x)f (x) dx;

(2)

xmin

where f(x) is the unknown distribution function of
a given parameter x . To calculate the f (x ) function, the regularization method can be used, as
solution of Eq. (2) is well-known ill-posed problem
due to a strong influence of noise components on
the experimental data, which do not allow one to
utilize exact inversion formulas or iterative algorithms [48,49]. For this purpose, the CONTIN
program package [49] was modified to utilize
adsorption equations to estimate the pore size
and adsorption energy distributions [44]. The
weights of the Wi /W0 coefficients in Eq. (1) were
estimated within the scope of LIA using adsorption data. One of the reasons of an increase in L
up to 4 in Eq. (1) with the varied ni parameters was
a large non-uniformity of hybrid C/X/SiO2 materials, as it is known [16,29,46,47] that ni in Eq. (1)
depends not only on the pore size distribution but
also on the nature of adsorbent surface, and f(Rp)
is stable at L ]/3 [44]. Additionally, a subsequent
increase in L up to 8 results in only small changes
in the f(Rp) distributions which become smoother
[38]. The k value (:/10 nm kJ mol (1 for hybrid
adsorbents) was estimated comparing parameter
magnitudes obtained using adsorption equations
dependent and independent on the nature of
adsorbents with LIA.
Additionally, the pore size distribution was
calculated using the overall isotherm equation [50]
a0

g

rk (p)

f (Rp ) dr'
rmin

g

rmax
rk (p)

w
Rp

t(p; Rp )f (Rp ) dr;

(3)

where rmin and rmax are the minimal and maximal
half-width or pore radius, respectively; w 0/1 for

109

slit-like pores and 2 for cylindrical pores;
2gnm cos u
;
(4)
Rg T ln(p0 =p)
a
cz

t(p; Rp )0 m
SBET 1 ( z
1 ' 1(nb ( n)zn(1 ( (nb ' 1)zn ' 1(nb ' n)zn'1 
2
2
;
1 ' (c ( 1)z ' 12(cb ( c)zn ( 12(cb ' c)zn'1
rk (p)0t(p; Rp )'

(5)

where b 0/exp(Do /RgT ), Do is the excess of the
evaporation heat due to the interference of the
layering on the opposite wall of pores (determined
as a varied parameter using LIA), t(p ,Rp) the
statistical thickness of adsorbed layer, am the
monolayer capacity, c 0/cs exp((Qp(/Qs)/RgT ), cs
the BET coefficient for adsorption on flat surface
(calculated using LIA), Qs and Qp are the adsorption heat on flat surface and in pores, respectively,
z0/p/p0, and n 0/d /tm is the number (non-integer)
of statistical monolayers. Qs and Qp values were
estimated on the basis of ab initio (B3LYP/631G(d,p) or larger basis sets) calculations of
nitrogen molecule interaction with different adsorbents, and the nitrogen adsorption energy
distributions f(E ) determined using adsorption
data for similar oxides and pyrocarbon/oxides
(Qp :/2Qs for narrow micropores at Rp B/0.4
nm). Desorption data were utilized to compute
the f(Rp) distribution with Eq. (3) and regularization procedure. According to Nguen and Do [50],
Eqs. (3) Á/(5) are valid not only for mesopores but
also for micropores (in Ref. [50], xmin 0/0, but

micropores at the half-width xmin B/0.2 nm are
inaccessible for nitrogen molecules, and therefore,
one can assume that xmin 0/0.2 nm in Eq. (3) has
no influence on the positions of f(Rp) peaks even
for micropores).

3. Results and discussion
3.1. Kinetics of cyclohexene pyrolysis
Differences in the characteristics of Brønsted (Å/
SiO(H)M Å/, M 0/Al, Ti), Lewis (incompletely Ocoordinated Ti or Al atoms), and other sites on TS


110

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

and AS surfaces depending on the synthetic
technique (pyrogenic, CVD, etc.) and X concentration (as changes in CX lead to alterations in the
size and shape of X particles and X/SiO2 interface
structure) reflect in pyrolysis of organics on these
oxides [3 Á/6]. Alumina/silica catalyses pyrolysis at
CX 5/8 wt.% stronger (per gram of the adsorbent)
than AS23 and CVD-TiO2/SiO2 (Fig. 1, CX 0/0
corresponds to A-300). This result is due to several
reasons. Firstly, the lower the CAl2O3, the greater is
the Brønsted acidity of Å/SiO(H)Al Å/ [51,52];
secondly, formation of individual alumina phase
in AS at great CAl2O3 causes the appearance of
large number of sixfold O-coordinated Al atoms
[20], which are not Lewis acid sites and can also be

responsible for a reduction of the Brønsted acidity
of neighboring bridging hydroxyls, i.e. the number
of strong Brønsted and Lewis sites can be reduced
in AS with increasing CAl2O3, which is responsible
for diminution of the catalytic activity of AS23 in
acid Á/base reactions. Thirdly, SBET of AS (Table 1)
is typically larger than that of fumed TS (Table 2);
therefore, normalized CC (divided by SBET, Fig.
1(c) and (d)) is larger for TS36 than those for AS1
or AS8, and similar result is observed for the
normalized carbonization rate constant (Fig. 2),
i.e. catalytic activity of TS36 per m2 is the largest
among studied X/SiO2 oxides.
Fumed TS and AST at CTiO2 ]/65 wt.% (Fig.
1(e)) are more active in cyclohexene pyrolysis than
fumed TS at low CTiO2 or CVD-TS (Fig. 1(b)) due
to the differences in the titania/silica interface
structure, and distribution and morphology of
titania phase per se [19,21,22,25]. In the case of
CVD-TS, titania is distributed in the form of
relatively large particles, whose size strongly grows
with CTiO2. Therefore, the specific catalytic activity
(per m2) of CVD-TS depends slightly on CTiO2 and
maximal for TS33 at CTiO2 0/33 wt.% (Figs. 1(b)
and 2). The reaction rate (Fig. 2) depends slightly
on the carbonization time (maximal for TS36 at
t :/0), i.e. entire poisoning of active surface sites of
X/SiO2 is not observed on cyclohexene pyrolysis at
973 K for 4 h, since pyrocarbon does not cover
totally the oxide surfaces (additionally, pyrocarbon can be a catalyst of this reaction).

Comparing kinetic data for AS and TS (Figs. 1
and 2), one can conclude that in the case of AS, the

Al2O3/SiO2 interfaces with Brønsted and Lewis
sites catalyze pyrolysis of organics, but for TS,
both titania phase and TiO2/SiO2 interfaces possess the catalytic capability, which impacts the
pyrocarbon distribution on the oxide matrix.
However, titania per se in TS and AST, may be,
plays the main role, as the normalized reaction
rate is larger for fumed TS at CTiO2 0/36 wt.% (Fig.
2); however, the largest number of Å/Si Ã/O Ã/TiÅ/ (or
Å/SiÃ/O(H) Ã/TiÅ/) bridges is observed in fumed TS
at CTiO2 :/20 wt.% [19,21]. For CVD-TS, the
reaction rate can be lower due to distribution
features of CVD-TiO2 in the form of relatively
large particles, which cannot provide a great
catalytic activity of CVD-TS due to a small
contribution of the surface area of titania particles
to the overall specific surface area of CVD-TS. In
the case of AS, the reaction rate is maximal at low
CAl2O3 (Figs. 1 and 2), as Brønsted acidity
decreases with CAl2O3. These catalytic features of
X/SiO2 can reflect in the structural adsorptive
characteristics of pyrocarbon and C/X/SiO2 as a
whole.
3.2. Morphology of X/SiO2 and C/X/SiO2
Isotherms of nitrogen adsorption Á/desorption
on fumed silica and pyrocarbon/fumed silica [3]
demonstrate the absence of the saturation effect,
as the adsorption plateau does not appear even at

p/p0 0/1. Such a plateau is typically observed for
silica gels when pores are filled in a significant
portion and nitrogen adsorption does not practically rise with increasing p /p0 !/0.9, as a contribution of the outer surface of silica gel particles to the
specific surface area is very small and capillary
condensation occurs mainly in pores. In the case of
fumed X/SiO2 and C/X/SiO2, the isotherm shape is
close to that for fumed silica and C/SiO2, but their
position drops down with decreasing VBJH and
SBET (Tables 1 and 2).
Fumed oxides possess a complicated structural
hierarchy with nonporous spherical primary particles (diameter: 5 Á/50 nm) packed in aggregates
(100 Á/500 nm) and agglomerates of aggregates (!/
1 mm) possessing strongly different apparent
densities due to changes in the type of packing of
particles in aggregates (apparent density:/0.7 g


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

111

Fig. 1. (a, b) Relative and (c, d) normalized (divided by SBET) carbon contents as a function of the carbonization time for silica and
mixed oxides: (a, c) AS; (b, d) CVD-TS (open symbols) and fumed TS (solid symbols); and (e) fumed TS and AST at large CTiO2 ]/65
wt.%.


112

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Fig. 2. Carbonization reaction rate as a function of carbonization time for (a) fumed silica and AS; and (b) CVD-TS (open
symbols) and fumed TS (solid symbols).

cm (3) and agglomerates (:/0.04 Á/0.10 g cm (3)
[53 Á/57]. Firstly, adsorption of nitrogen can occur
near contacts between adjacent primary particles
in aggregates corresponding to ‘micropores’ and
narrow ‘mesopores’, then in channels (mesopores)
formed by clusters of neighboring particles in
aggregates and also on the external surfaces of
aggregates, which are not closely packed in
agglomerates, and between aggregates, i.e. in large
mesopores or even macropores. Clearly, complete
filling of all these free inter-particle volumes
(especially in agglomerates) cannot be reached
upon nitrogen adsorption on fumed oxides, possessing the apparent density of about 1 Á/3% of the

specific density, even at p/p0 0/1. This circumstance causes the corresponding shape of isotherms at p /p0 0/1 with the plateau independent
on the X phase type and its concentrations in X/
SiO2. However, the appearance of X phase is
responsible for a reduction of the adsorption due
to a decline in the pore volume and specific surface
area with increasing CX especially for fumed TS
(Tables 1 and 2). Additionally, the hysteresis loop
in isotherms for fumed silicas is relatively small
(i.e. capillary effect is weak) in contrast to
mesoporous silica gels [2 Á/5], but it is slightly
larger for mixed X/SiO2 and C/X/SiO2 due to
known changes in the particle (primary and
swarm) morphology in comparison with the initial

fumed silica [53 Á/57]. The absence of the strong
capillary effects for fumed oxides causes more
exact estimation of the specific surface area of
mesopores using the Kiselev equation in comparison with silica gels (whose SK can be nearly twice
as large as SBET [2 Á/6]); as a result, the sum SK'/
SDS is close to SBET (Tables 1 and 2) and the SK
values are akin to SdB and Svt (Table 3).
For all TS samples, carbonization leads to
reduction of the specific surface area (Table 2),
and the isotherms for C/TiO2/SiO2 lie below those
of TiO2/SiO2. The pore volume decreases for all C/
fumed TS samples (large CC per m2 of the
adsorbent), but for C/CVD-TS (smaller CC per
m2), it is seen only for TS22 and TS33 (Table 2).
For C/SiO2 and C/AS3 (AS3 has a minimal specific
surface area and porosity among AS samples),
SBET and VBJH are enlarged, but for other C/AS
samples, they tend to diminish (especially for C/
AS8) in comparison with those for initial AS
(Table 1). This effect for silica (A-300, A-175)
can be linked with relatively low CC values in C/
SiO2 (the lower the CC value, the larger is the
specific surface area of carbon deposit per se) [2 Á/
6,26], as silica does not catalyze pyrolysis in
contrast to AS and TS (Figs. 1 and 2). These
structural effects can be elucidated in detail on the
basis of analysis of the pore volume (Fig. 3) and
size (Figs. 4Á/6) distributions and changes in
adsorbent fractality (Tables 1Á/3).
The dVmes/dRp functions (Fig. 3) obtained using

LIA have two main maxima corresponding to the
mesopore radius of Â/1.5 and Â/2.5 nm, i.e. these


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

113

Fig. 3. Mesopore volume distribution for (a) initial fumed silica, C/SiO2, AS1, and C/AS1; (b) initial CVD-TS and C/CVD-TS; (c) AS
and C/AS; and (d) fumed TS and C/TS.

Rp values are smaller than the size of primary
particles of fumed oxides, and the main dVmes/dRp
peak shifts towards smaller Rp in comparison with
that of the particle size distribution of fumed silica
(or X/SiO2) with increasing average particle size
and decreasing SBET (A-300 0/A-175 and lower)
[55]. These small mesopores can be formed by
primary particles having relatively tight contacts in
aggregates (apparent density of aggregates is
about 30% of the specific density [53,54]), i.e.
these pores represent an accessible surface of
primary particles in aggregates giving a significant

contribution to the specific surface area increasing
with the aggregate size. The spatial structure of
these mesopores can be complicated due to random packing of primary particles of different sizes
in aggregates. Clearly, formation of carbon deposit (especially at CC !/10 wt.% [18]) can occur
mainly on the outer surface of aggregates or in
broad channels in aggregates (if oxide is relatively

inert, such as silica, in pyrolysis of organics) and,
with consideration of a possible pore structure
(particle packing is random and not too dense),
formation of the pore plugs is less probable for


114

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Fig. 4. Pore size distributions calculated with the regularization procedure using (a, b) Eq. (1) at L0/3 for adsorption data
(regularization parameter a0/0.01), and (c, d) Eq. (3) for desorption data (a0/0.001); (a, c) silica, alumina (SBET :/160 m2), and AS;
and (b, d) silica, titania (SBET :/60 m2), and TS.

fumed oxides than that for silica gels having
mainly cylindrical pores between more closely
packed primary particles with smoother pore walls
(i.e. a micropore contribution in SBET for meso-

porous silica gels can be smaller than that for
fumed oxides) [2 Á/6].
For C/AS, the intensity of the first and second
peaks in dVmes/dRp (Fig. 3(a) and (c)) is minimal


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

115

Fig. 5. Pore size distributions for pyrocarbon/oxides calculated with the regularization procedure using (a, b) Eq. (1) and L0/3 for

adsorption data (regularization parameter a 0/0.01), and (c, d) Eq. (3) for desorption data (a0/0.001); (a, c) C/silica and C/AS; and (b,
d) C/silica and C/TS.

for C/AS1 having maximal CC and minimal SBET
and VBJH values among C/AS samples (Table 1).
However, a maximal reduction of the last parameters due to pyrocarbon grafting is observed for
C/AS8, but AS8 possesses larger VBJH and SBET
than those for AS1. Minimal values of CC,
DSBET 0/SBET(C /AS)(/SBET(AS), and DVBJH 0/
VBJH(C/AS)(/VBJH(AS) among C/AS samples are
characterized for C/AS23 due to the availability of
individual amorphous alumina phase, which possesses a lower catalytic capability in pyrolysis in
comparison with the alumina/silica interface
(Brønsted acidic sites). Therefore, the intensity of
dVmes/dRp (Fig. 3) at Rp between 1 and 3 nm for
C/AS23 is maximal among C/AS samples.

The first and second peaks of dVmes/dRp (Fig. 3)
are larger for CVD-TS (and C/CVD-TS) than
those for fumed TS due to the smaller sizes of a
major portion of CVD-TS particles (namely silica
particles in TS, as formation of CVD-titania was
performed on the A-300 substrate, and CVDtitania rather represents individual relatively large
particles than a continuous layer covering the
silica substrate [21]), and SBET of CVD-TS is
significantly larger than that of fumed TS (Table
2). Additionally, the DBJH and VBJH values (Table
2) and dVmes/dRp functions (Fig. 3) show that
channels in aggregates of fumed TS can be
narrower than those of CVD-TS that corresponds

to a marked diminution of the porosity (SBET,


116

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Fig. 6. Pore size distribution of oxides and carbon/oxides calculated with the regularization procedure using Eq. (1) and L 0/4, and
automatically chosen regularization parameter a (determined between 0.005 and 0.0005).


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

VBJH) of fumed TS at CTiO2 ]/20 wt.% in comparison with CVD-TS. It should be noted that in the
case of TS (and C/TS), the third peak (shoulder) of
dV /dRp is observed at Rp :/20Á/25 nm and it
slightly shifts towards larger Rp with increasing
CTiO2 (Fig. 3). These mesopores can be linked with
the inter-aggregate volumes in agglomerates and
broad channels in aggregates between relatively
large TiO2 particles (!/20 nm), which can be filled
by nitrogen only in part at p /p0 !/0.9. A similar
shoulder of dV /dRp is also observed for silica and
AS (Fig. 3(a) and (c)), but it is less marked than
that of TS; however, its shift is seen, e.g. for A-300
and A-175.
The pore size distribution calculated using Eq.
(1) as a kernel of Eq. (2) at L 0/3 (ni and Wi /W0
optimized with LIA) and fixed regularization
parameter a 0/0.01 (parts (a) and (b) of Figs. 4

and 5), or L 0/4 and unfixed a (determined on the
basis of F -test and confidence regions for nonnegative solutions of Eq. (2) [56]) (Fig. 6) for
adsorption data or utilizing Eq. (3) for desorption
data (parts (c) and (d) of Figs. 4 and 5) with the
regularization procedure (a 0/0.001) demonstrate
marked changes in f(Rp) depending on the nature
of X, and concentrations CX in X/SiO2 and CC in
C/X/SiO2. It is pertinent to note that the f(Rp)
distributions at Rp between 1 and 10 nm (Figs. 4Á/
6) are akin to dVmes/dRp (Fig. 3) but differ in halfwidth of peaks; however, f(Rp) gives more informative picture of micro- and mesopore changes
over a large Rp range up to 100 nm. Thus,
utilization of different isotherm equations for
adsorption and desorption data with LIA or
overall adsorption (Eq. (2)) results in relatively
close pore size distributions (Figs. 3 Á/5) that lend
support to the validity of the used methods.
Contact zones between adjacent spherical primary particles can be considered as slit-like
micropores (narrow channels in aggregates are
‘cylindrical’ micropores) with a different accessibility dependent on the size of adsorbate molecules, which impact, e.g. SBET for fumed silica
determined with different adsorbates [53]. Micropores give large one or even two f(Rp) peaks at Rp
between 0.4 and 0.9 nm (Figs. 4Á/6), whose halfwidth depends on the regularization parameter
value (a 0/0.01 Á/0.0005) and used local isotherm

117

with Eq. (1) at L 0/3 or 4 and Eq. (3). Thus, all the
studied samples possess marked microporosity,
whose contribution to f (Rp) seems large due to
the log-scale for Rp (in the linear Rp scale,
micropores give a very narrow peak as half-width

equals to 0.3 Á/0.7 nm).
Fumed TS at CTiO2 ]/20 wt.%, titania and
alumina (SBET B/160 m2 g(1), which consist of
larger primary particles (forming larger channels
in aggregates) than AS or silica, have more marked
peaks in f(Rp) at Rp !/2 nm (Fig. 4). However,
after carbonization, new f(Rp) peaks appear at
Rp !/5 nm and their intensity is larger for C/AS1
(Fig. 5(a)) and C/fumed TS (Fig. 5(b)) containing
greater amounts of pyrocarbon, i.e. these f (Rp)
peaks can be linked to the porosity of grafted
carbon.
Some details of the pore size distributions can be
elucidated using the regularization procedure with
unfixed a (Fig. 6). Relative intensity of the second
f(Rp) peak grows with the size of primary particles
(Fig. 6(a) for A-300 and Fig. 6(b) for A-175), but
micropore contribution (first f(Rp) peak) of C/A300 is higher than that of C/A-175; therefore,
changes in SBET of C/A-300 (in comparison with
A-300) is larger than that of C/A-175 (Table 1) as
well as changes in f(Rp) for C/SiO2 relative silica
(notice that CC for C/A-300 and C/A-175 is very
close). However, the first f(Rp) peak for A-175
shifts toward smaller Rp in comparison with that
for A-300 (due to changes in slit-like micropores
between neighboring primary particles with increasing their diameters) and relative contribution
of the second peak is greater. Changes in f (Rp) for
AS1 and AS23 due to grafted carbon are different,
and in the case of C/AS23, pyrocarbon is more
uniform (Fig. 6(c) and (d)), as the alumina phase

distribution (i.e. distributions of active sites, which
catalyze pyrolysis) depends strongly on CAl2O3 in
fumed alumina/silica [20], and CC in C/AS1 is
larger (Table 1). The titania phase distributions in
CVD-TS and fumed TS are different, as CVDtitania consists of relatively large particles having
weak contacts with the silica substrate [21]; therefore, pore size distribution in CVD-TS depends
mainly on the silica phase and f (Rp) for A-300 and
CVD-TS33 or C/A-300 and C/CVD-TS33 are
similar (Fig. 6(a) and (e)) and relatively large


118

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

amounts of grafted carbon (formed mainly on the
titania surface) have a weak impact on the pore
size distribution of the silica phase. At the same
time, the distribution function f (Rp) of carbon/
fumed TS36 is similar to that of C/AS23 (Fig. 6(d)
and (f)); however, the catalytic capability of TS36
per m2 is the greatest (Figs. 1 and 2).
Thus, changes in the pore (channels in aggregates) structure of fumed oxides due to carbon
grafting (Tables 1 and 2 and Figs. 3Á/6) strongly
alter from those observed for C/silica gels or C/
CVD-titania/silica gel [2 Á/6] because of the differences in (a) structural hierarchies, (b) packing of
primary globules of silicas, and (c) distribution of
the second oxide phase [20 Á/25]. Typically, the
specific surface area of C/fumed X/SiO2 decreases
smaller due to carbon grafting (Tables 1 and 2)

than that for C/silica gels or C/CVD-titania/silica
gel at the same concentration of pyrocarbon
deposit [2 Á/6], as in the case of fumed oxides, the
porosity is linked with more open pores (channels)
formed by primary particles in aggregates, which
are less dense than large silica gel particles having
mesopores with a relatively narrow size distribution, which influences the carbon phase structure.
Also, contribution of the external surfaces of
aggregates of primary particles of fumed oxides
is significantly larger than that for silica gel
particles. This conclusion is also supported by
the shape of f (Rp) (Figs. 4 Á/6) and dV /dRp(Rp)
(Fig. 3) showing marked inter-particle volumes at
Rp up to 50 nm in C/X/SiO2 in contrast to silica
gels [2 Á/6].
Pyrocarbon grafting on the fumed silica surfaces
leads even to a gain of the specific surface area
(Table 1); however, in the case of X/SiO2 (X 0/
fumed titania, alumina, CVD-titania), the SBET
values typically go down (Tables 1 and 2). It is
worth noting that apparent density of C/X/SiO2
can rise in comparison with that of initial X/SiO2.
However, a diminution of the amount of oxide per
cm3 of C/X/SiO2 can be found, and as a result,
SBET is lowered, as the specific surface area of the
carbon deposit (:/100 m2 g(1 at large CC) [3] can
be lower than that of fumed silica. Fractal dimension of C/X/SiO2 materials is lower than that of X/
SiO2 (Tables 1 and 2). This can be caused by an

enhancement of the density of adsorbents due to

reduction of the free volume between primary
particles in aggregates (fumed oxides are mass
fractal [56]), whose walls become smoother, that
gives a diminution of the specific surface area (in
m2 g(1), i.e. surface fractality changes due to the
alterations in the surface topology (roughness
alters per cm3) [2 Á/6]. Additionally, the specific
surface area of the carbon deposit at large CC
values [3,7 Á/10,18] is typically lower than that of
fumed silica. Notice that at CTiO2 5/17 wt.%, the
VBJH value is magnified due to carbon grafting
onto both fumed and CVD-TS; however, the SBET
value decreases (Table 2). In the case of AS, VBJH
typically reduces (Table 1). This effect can be
linked with the differences in CC, as the pyrocarbon concentration is typically lower for C/TiO2/
SiO2 than that for C/Al2O3/SiO2. However, fractal
dimension of C/fumed TS is smaller than that of
C/Al2O3/SiO2 (except C/AS1) or C/CVD-TS (Tables 1 and 2). Comparison of DAJ at different p /p0
(or U ) and fractal dimensions (Df) estimated on
the basis of the density of aggregates (as fumed
oxides are mass fractal) (Df,1 :/2.5 from equation
rcluster 0/rparticle(dcluster/dparticle)Df(3) and agglomerates of aggregates (Df,2 :/2.1 Á/2.2) [53,54] shows
that DAJ2 (p/p0 B/0.85) is close to Df,1. Therefore,
one can assume that filling of inter-particle volume
by nitrogen occurs only in aggregates at large p /p0
values, but filling of the inter-aggregates volumes
in agglomerates (adsorption on the outer surface
of aggregates of primary particles) takes place in
an insignificant portion. The bulk silica (specific
density :/2.2 g cm (3) volume corresponds to :/

0.32 cm3 g(1 of aggregates (as the apparent
density of aggregates is :/0.7 g cm (3) and the
free volume in aggregates is about 0.68 cm3 g(1,
which is in agreement with VBJH of fumed silica
(Table 1). Consequently, this pore volume VBJH
filled at p/p0 :/0.98 (as VBJH was estimated at such
a pressure) corresponds the free volume in aggregates practically without contribution of a large
free volume in agglomerates. This conclusion is in
agreement with the shape of isotherms, which do
not have the plateau at p/p0 0/1. Changes in
DFRDA and DAJ with increasing CC are close
(Tables 1 and 2) and their minima are linked
/


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

with C/A-300, C/TS2, and C/AS1. It should be
mentioned that the observed differences between
DFRDA and DAJ values are in agreement with the
results obtained previously [41]. However, an
increase in xmax to 3.5 nm gives the DFRDA values
(Table 3) between DAJ1 and DAJ2 (Tables 1 and 2).
The observed dependencies of the D values on the
utilized p/p0 range (DAJ1, DAJ2) or xmax value for
DFRDA can be caused by nonuniformity of X/SiO2
and C/X/SiO2 particles (primary particles, aggregates, and agglomerates), i.e. these materials can
be characterized by combination of mass (primary
particles of initial fumed oxides), surface (portion
of tiny CVD-titania and pyrocarbon particles),

and pore (oxide aggregates and large pyrocarbon
particles) fractal [56] components.
The nature of the mixed oxide surfaces influences the water adsorption energy distributions
f(E ) (calculated with modified BET equation akin
to Eq. (5) as a kernel in Eq. (2) assuming c 0/
exp((E(/QL)/RT ), where QL is the liquefaction
heat of water) (Fig. 7) due to the differences in
amounts of Brønsted and Lewis acid sites (adsorption of water on studied AS and TS samples was
described in detail elsewhere [20,21]). The first f (E )
peak corresponds to water molecules weakly
bound to the surface (in large adsorbed water
clusters or molecules adsorbed on weak surface
sites) and having one or two hydrogen bonds per
molecule. The second f (E ) peak is linked to
strongly bound water molecules interacting directly with Brønsted and Lewis acid sites (first
adsorbed monolayer). The last peak energy is
greater for AS than that for silica and TS (AS
possesses stronger Brønsted acidity than TS has
[20 Á/23]); however, it shifts toward larger energy
for titania, but its intensity is very low. Notice that
the activation energy (E") of desorption of intact
water molecules from these oxides (e.g. E" 0/67 kJ
mol (1 at peak temperature Tmax 0/458 K for
AS23, 63 kJ mol(1 and Tmax 0/438 K for TS22,
and 71 kJ mol (1 and Tmax 0/480 K for TiO2 [24])
is in agreement with f(E ) shown in Fig. 8.
Additionally, the second f(E ) peak for AS1 lies
at larger energy than that for AS3 or AS23, which
can be connected with greater acidity of AS at low
CAl2O3 [51,52].

119

Fig. 7. Water adsorption energy distributions calculated using
modified BET equation and regularization procedure at a 0/
0.01 for adsorption data.

3.3. X/SiO2 and C/X/SiO2 particles in aqueous
suspension
Properties of C/X/SiO2 particles in the aqueous
suspensions can be linked to the carbon deposit
morphology, which are of importance for applications of these materials in liquid media. For C/
SiO2, an increase in CC makes the surface more
basic, and a minimum of electrokinetic potential
z (pH) observed for pure silica suspensions at
pH:/2.5 Á/3 disappears even at low CC (a similar
effect has been seen for pure silica with increasing
salinity of the suspension) (Fig. 8) [57]. An
enhancement of the basic properties of C/SiO2
surface is linked with basic sites of carbon phase


120

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Fig. 8. Electrokinetic potential as a function of pH at (a)
different salinity of pure silica suspension (z (pH) is also shown
for C/A-300 at CC 0/9.3 wt.%); and (b) different CC (pyrolysis
of CH2Cl2 at 673 Á/823 K for 40 Á/120 min) in C/A-300 at 0.01 M

NaCl.

(possessing pre-graphite structure and partially
oxidized graphene clusters), for which a pH value
of the isoelectric point (IEP) can be about 9 [3,57Á/
59]. In case of AS, the carbon deposit forms first of
all on the strongest acidic or basic sites (possessing
a maximal catalytic capability in pyrolysis), as the
shape of the z (pH) curves for C/AS changes not
only at pH B/5 (negative charge of particles is
provided by acidic sites) but also at pH !/6 (Fig.
9). For C/AS, pH(IEP) is close to 4, but for pure

Fig. 9. z Potential as a function of pH for (a) AS and C/AS; (b)
fumed TS and C/TS; and (c) CVD-TS and C/CVD-TS at
different concentrations of CX in C/X/SiO2.


V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

AS, it is below 2. Therefore, one can assume that
pyrocarbon in C/AS is distributed nonuniformly,
mainly on the alumina/silica interfaces and alumina phase, as they possess the catalytic ability in
pyrolysis of organics. Additionally, the z (pH)
curves for C/AS are relatively close to that for C/
SiO2; consequently, silica surface and carbon
deposit in C/X/SiO2 are mainly accessible for
liquid water. A similar effect is observed for C/
TS (Fig. 9), but it is not too great as that for C/AS
may be due to closer electrokinetic properties of

titania and silica than those of alumina and silica
[57]. There is tendency of an enhancement of
pH(IEP) with increasing CC practically independent on the nature of X/SiO2 substrate or salinity
of the suspension (Fig. 10); additionally, the range
of pH(IEP) for C/X/SiO2 materials is more narrow
(2.5 Á/4.5) than that of X/SiO2 [57] (Fig. 10). These
effects can be connected with the formation of
pyrocarbon on more active (in pyrolysis) patches
of the X/SiO2 surfaces such as the X/silica interface
and X phase possessing Brønsted and Lewis acid
sites or sites active in redox reactions (anatase), i.e.
silica phase in C/X/SiO2 is covered by pyrocarbon
to a lesser extent and pH(IEP) of C/X/SiO2
corresponds to average value of silica
(pH(IEP) :/2.2) and carbon deposit.
Thus, in the case of such mixed oxides as AS
and TS, formation of carbon deposit due to
pyrolysis of cyclohexene occurs mainly on surface
patches of the X/SiO2 interface and X phase
possessing the catalytic capability in pyrolysis,
which results in the similarity of the electrokinetic
behavior of C/X/SiO2 and C/SiO2 particles in the
aqueous suspensions, as X phase is shielded by
pyrocarbon to a greater extent than silica.

4. Conclusions
For fumed alumina/silica, mainly the interfaces
with Brønsted and Lewis sites catalyze cyclohexene pyrolysis, but in the case of TS, both the
titania phase and TiO2/SiO2 interfaces possess the
catalytic capability, which impacts the carbon

structure and distribution. However, titania phase
per se in TS plays an important role, as the
normalized reaction rate is larger for fumed

121

TS36, but for AS, the reaction rate is maximal at
low CAl2O3. For CVD-TS, the pyrolysis rate can be
lower also due to distribution features of CVDTiO2 in the form of relatively large particles, which
cannot provide a great specific catalytic activity of
CVD-TS due to a small contribution of the surface
area of titania particles to the specific surface area
of CVD-TS as a whole. These effects result in the
differences in the pyrocarbon distribution on oxide
support particles.
The SBET value of C/fumed silica (low CC) is
larger than that for the initial silica in contrast to
pyrocarbon/silica gels. The absence of the plateau
and small hysteresis loops are observed in isotherms of nitrogen adsorptionÁ/desorption on
fumed silica, alumina/silica, titania/silica, CVDTS, and C/X/SiO2, as capillary effects are weak
due to features of channels (free volume) in
aggregates formed with near-spherical primary
particles. The difference in dVmes/dRp functions
for C/X/SiO2 is due to the formation of various
channels in aggregates of primary particles, whose
accessible surface area determines the SBET and
Smes values to a greater extent than the pore
volume, which also depends on the packing types
of primary particles not only in aggregates but also
in agglomerates. The SBET values typically go

down for C/X/SiO2 in comparison with those of
X/SiO2, and fractal dimension drops due to
grafting of tiny carbon particles in the free volume
of aggregates and on their outer surfaces (surfaces
become smoother); however, in the case of pure
silica, carbon particles can be mainly formed on
the outer surface of aggregates, which reflects in
smaller changes in fractal dimension of C/A-175
(relative A-175) in comparison with other C/X/
SiO2 adsorbents. Thus, changes in the distribution
of X oxide in X/SiO2 result in the alterations in
pyrocarbon distribution and its concentration in
C/X/SiO2, which are maximal for AS1 (per g) and
TS36 (per m2).
Changes in electrokinetic behavior of hybrid
particles in the aqueous suspensions due to pyrocarbon grafting show that C deposit can be
nonuniform, as it covers mainly surface patches
of the X/silica interface and X phase possessing
great catalytic activity in pyrolysis of organics than
silica, and in the case of pure silica, pyrocarbon


122

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

Fig. 10. pH of the isoelectric point for aqueous suspensions of C/X/SiO2 as a function of (a) CC in C/A-300 at different salinity; (b)
CAl2O3 in C/AS; (c) CC in C/A; (d) CTiO2 and CC in C/fumed TS; and (e) CTiO2 in CVD-TS and C/CVD-TS, and CC in C/CVD-TS.

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

represents particles formed mainly on the outer
surface of aggregates of primary silica particles,
and SBET of C/SiO2 is larger than that of SiO2.
Consequently, changes in the structure of silica
and X/SiO2 (X distribution, X particle morphology, X concentration, etc.) allow one to control
the distribution and amount of pyrocarbon
(grafted first of all on more active sites), nonuniformity of hybrid surfaces, and as a result the
overall properties of C/X/SiO2 adsorbents.

[14]
[15]
[16]
[17]
[18]
[19]
[20]

Acknowledgements
Financial support from the State Committee for
Scientific Research (KBN, Warsaw, Project No. 3
T09A 03611) is gratefully acknowledged. This
research was partially supported by NATO (Grant
EST.CLG.976890).

[21]

[22]

[23]

References
[24]
[1] (a) R. Leboda, Mater. Chem. Phys. 31 (1992) 243;
(b) R. Leboda, Mater. Chem. Phys. 34 (1993) 123.
[2] R. Leboda, V.M. Gun’ko, M. Marciniak, A.A. Malygin,
A.A. Malkin, W. Grzegorczyk, B.J. Trznadel, E.M.
Pakhlov, E.F. Voronin, J. Colloid. Interf. Sci. 218 (1999)
23.
[3] V.M. Gun’ko, J. Skubiszewska-Zieba, R. Leboda, V.I.
Zarko, Langmuir 16 (2000) 374.
[4] V.M. Gun’ko, R. Leboda, M. Marciniak, W. Grzegorczyk,
J. Skubiszewska-Zieba, A.A. Malygin, A.A. Malkov,
Langmuir 16 (2000) 3227.
[5] V.M. Gun’ko, R. Leboda, J. Skubiszewska-Zieba, J.
Rynkowski, J. Colloid. Interf. Sci. 231 (2000) 13.
[6] V.M. Gun’ko, F. Villie´ras, R. Leboda, M. Marciniak, B.
Charmas, J. Skubiszewska-Zieba, J. Colloid. Interf. Sci.
230 (2000) 320.
[7] K. Kamegawa, H. Yoshida, J. Colloid. Interf. Sci. 159
(1993) 324.
[8] T.J. Bandosz, J. Jagiello, K. Putyera, J.A. Schwarz,
Langmuir 11 (1995) 3964.
[9] T.J. Bandosz, J. Jagiello, K. Putyera, J.A. Schwarz,
Carbon 32 (1994) 659.
[10] T.J. Bandosz, J. Jagiello, B. Andersen, J.A. Schwarz, Clays
Clay Miner. 40 (1992) 306.
[11] V.V. Turov, R. Leboda, Adv. Colloid. Interf. Sci. 79 (1999)
173.

[12] K.A. Cumming, B.W. Wojciechowski, Catal. Rev. Sci.
Eng. 38 (1996) 101.
[13] A. Corma, V. Fornes, B. Imelik, C. Naccache, G.
Coudurier, Y. Ben Tarrit, J.C. Vedrine (Eds.), Catalysis

[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]

123

by Acids and Bases: Studies in Surface Science and
Catalysis, Vol. 20, Elsevier, Amsterdam, 1985.
B.W. Wojciechowski, A. Corma, Catalytic Cracking,
Dekker, New York, 1986.
J.B. Butt, E.E. Petersen, Activation, Deactivation and
Poisoning of Catalysts, Academic Press, New York, 1988.

V.B. Fenelonov, Porous Carbon, Institute of Catalysis,
Novosibirsk, 1995.
X. Bourrat, Carbon 31 (1993) 287.
F. Villieras, R. Leboda, B. Charmas, F. Bardot, G.
Gerard, W. Rudzinski, Carbon 36 (1998) 1501.
V.M. Gun’ko, V.I. Zarko, V.V. Turov, E.F. Voronin, V.A.
Tischenko, A.A. Chuiko, Langmuir 11 (1995) 2115.
V.M. Gun’ko, V.V. Turov, V.I. Zarko, E.F. Voronin, V.A.
Tischenko, V.V. Dudnik, E.M. Pakhlov, A.A. Chuiko,
Langmuir 13 (1997) 1529.
V.M. Gun’ko, V.I. Zarko, V.V. Turov, R. Leboda, E.
Chibowski, L. Holysz, E.M. Pakhlov, E.F. Voronin, V.V.
Dudnik, Yu.I. Gornikov, J. Colloid. Interf. Sci. 198 (1998)
141.
V.M. Gun’ko, V.I. Zarko, V.V. Turov, R. Leboda, E.
Chibowski, E.M. Pakhlov, E.V. Goncharuk, M. Marciniak, E.F. Voronin, A.A. Chuiko, J. Colloid. Interf. Sci.
220 (1999) 302.
V.M. Gun’ko, V.I. Zarko, V.V. Turov, R. Leboda, E.
Chibowski, Langmuir 15 (1999) 5694.
V.M. Gun’ko, V.I. Zarko, B.A. Chuikov, V.V. Dudnik,
Yu.G. Ptushinskii, E.F. Voronin, E.M. Pakhlov, A.A.
Chuiko, Int. J. Mass Spectrom. Ion Processes 172 (1998)
161.
V.M. Gun’ko, V.I. Zarko, E. Chibowski, V.V. Dudnik, R.
Leboda, V.A. Zaets, J. Colloid. Interf. Sci. 188 (1997) 39.
R. Leboda, V.V. Turov, B. Charmas, J. SkubiszewskaZieba, V.M. Gun’ko, J. Colloid. Interf. Sci. 223 (2000) 112.
R. Leboda, Chromatographia 9 (1981) 524.
A. Gierak, R. Leboda, E. Tracz, J. Anal. Appl. Pyrolysis
13 (1988) 89.
A.W. Adamson, A.P. Gast, Physical Chemistry of Surface,

6th ed., Wiley, New York, 1997.
S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and
Porosity, 2nd ed., Academic Press, London, 1982.
E.B. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem.
Soc. 73 (1951) 373.
A.V. Kiselev, V.P. Dreving, A.D. Runov, Dokl. Acad.
Nauk. USSR 46 (1945) 310.
D. Dollimore, C.R. Heal, J. Appl. Chem. 14 (1964) 109.
(a) J.C. Broekhoff, J.H. deBoer, J. Catal. 9 (1967) 8;
(b) J.C. Broekhoff, J.H. deBoer, J. Catal. 10 (1968) 153.
M.M. Dubinin, L.I. Kataeva, B.I. Ulin, Izv. Acad. Nauk.
USSR Ser. Khim. 3 (1977) 510.
V.M. Gun’ko, Zh. Fiz. Khim. 63 (1989) 506.
V.M. Gun’ko, V.M. Lukyanchuk, A.A. Chuiko, Catal.
Catalysts 28 (1992) 82.
V.M. Gun’ko, Teoret. Eksperim. Khim. 36 (2000) 1.
D. Avnir (Ed.), The Fractal Approach to Heterogeneous
Chemistry, Wiley, Chichester, 1989.
D. Avnir, M. Jaroniec, Langmuir 5 (1989) 1431.


124

V.M. Gun’ko et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 103 Á/124

[41] A.P. Terzyk, P.A. Gauden, G. Rychlicki, R. Wojsz,
Colloid. Surf. A 152 (1999) 293.
[42] M.M. Dubinin, F. Stoeckli, J. Colloid. Interf. Sci. 75
(1980) 34.
[43] A.V. Neimark, P.I. Ravikovitch, M. Grun, F. Schuth,

K.K. Unger, J. Colloid. Interf. Sci. 207 (1998) 159.
[44] R.Sh. Mikhail, S. Brunauer, E.E. Bodor, J. Colloid. Interf.
Sci. 26 (1968) 45.
[45] B.C. Lippens, B.G. Linsen, J.H. deBoer, J. Catal. 3 (1964)
32.
[46] M.M. Dubinin, Carbon 23 (1985) 373.
[47] M.M. Dubinin, in: D.A. Cadenhead (Ed.), Progress in
Surface and Membrane Science, vol. 9, Academic Press,
New York, 1975.
[48] M.V. Szombathely, P. Brauer, M. Jaroniec, J. Comput.
Chem. 13 (1992) 17.
[49] S.W. Provencher, Comput. Phys. Commun. 27 (1982) 213,
229.

[50] C. Nguen, D.D. Do, Langmuir 15 (1999) 3608.
[51] K. Tanabe, Solid Acids and Bases, Kodansha, Tokyo,
1970.
[52] V.B. Kazansky, Usp. Khim. 57 (1988) 1937.
[53] H. Barthel, Colloid. Surf. A 101 (1995) 217.
[54] H. Barthel, L. Rosch, J. Weis, in: N. Auner, J. Weis (Eds.),
Organosilicon Chemistry. II. From Molecules to Materials, VCH, Weinheim, 1996, pp. 761 Á/778.
[55] A.G. Degussa, Basic Characteristics of Aerosil, Technical
Bulletin Pigments, Vol. 11, Hanau, 1997.
[56] A.P. Legrand (Ed.), The Surface Properties of Silicas,
Wiley, New York, 1998.
[57] V.M. Gun’ko, V.I. Zarko, R. Leboda, E. Chibowski, Adv.
Colloid. Interf. Sci. 91 (2001) 1.
[58] H.P. Boehm, Carbon 32 (1994) 759.
[59] C.A. Leon y Leon, L.R. Radovic, in: P.L. Walker, Jr., P.A.
Thrower (Eds.), Chemistry and Physics of Carbon, vol. 17,

Marcel Dekker, New York, 1994, pp. 213 Á/310.