Research Article
Preparation, characterisation andmodification of carbon-based monolithicrods for chromatographic applications
A range of porous carbon-based monolithic (PCM) rods with flow-through pore sizes of 1,
2, 5 and 10 mm, were produced using a silica particle template method. The rods were
characterised using SEM and energy-dispersive X-ray spectroscopy, BET surface area and
porous structure analysis, dilatometry and thermal gravimetry. SEM evaluation of the
carbon monolithic structures revealed an interconnected rigid bimodal porous structure
and energy-dispersive X-ray spectroscopy analysis verified the quantitative removal of the
embedded silica beads. The specific surface areas of the 1, 2, 5 and 10 mm rods were 178,
154, 84 and 125 m2/g after pyrolysis and silica removal, respectively. Shrinkage of the
monolithic rods during pyrolysis is proportional to the particle size of the silica used and
ranged from 9 to 12%. Mercury porosimetry showed a narrow distribution of pore sizes,
with an average of �700 nm for the 1 mm carbon monolith. The suitability of bare and
surface oxidised PCM rods for the use as a stationary phase for reversed and normal phase
LC was explored. The additional modification of PCM rods with gold micro-particles
followed by 6-mercaptohexanoic acid was performed and ion-exchange properties were
evaluated.
Keywords: Gold micro-particles / LC porous carbon monoliths / Silica geltemplateDOI 10.1002/jssc.200900845
1 Introduction
Monolithic porous columns have many advantageous
characteristics to be used in various LC techniques as
compared to traditional particle packed columns [1–3]. The
most significant of them is a relatively high flow through
porosity, leading to low-column pressure drops, with the
potential for use with elevated flow rates, leading to faster
separations. Early reports on the production and use of
monolithic columns in GC were made by Ross and
Jefferson [4]. The approach was extended to LC as early as
1974 by Hansen and Sievers [5]. Polymer monoliths quickly
gained popularity since their early demonstration in the late
1980s due to their relative ease of fabrication and have
shown specific advantages for application in bio-separations
[6, 7]. Nakanishi and Soga reported a significant advance in
1991 in the preparation of silica based monolithic rods,
produced via phase separation by polymerisation of tetra-
methoxysilane in the presence of poly(ethylene oxide) [8, 9].
A meso-porous structured silica skeleton was achieved
through hydrothermal treatment of the silica rod with
ammonium hydroxide. Since this pioneering work, silica-
based monolithic columns have arguably become the most
popular phases currently applied to small molecule separa-
tions [10, 11], including organic and inorganic ions [12–14].
Recently, other alternatives to silica materials for the
preparation of monolithic columns have been investigated,
including zirconia [15], hafnia [16], titania [17] and carbon
monoliths. These materials and corresponding columns
exhibit better hydrolytical and thermal stability properties
than their silica counterparts, but the inorganic oxide based
materials are more reactive.
Carbon monoliths exhibit the advantages of both inor-
ganic oxides and organic polymers, being resistant to swel-
ling and hydrolysis. Once formed, they are structurally rigid
and resistant to attack from strongly acidic or basic solu-
tions, and stable in almost all organic solvents. Additionally,
carbon monoliths exhibit a temperature stability well
beyond most normal LC operating conditions. Additionally,
being conductive materials PGC columns are also suitable
for the use in electrochemically modulated LC or, if
suitably functionalised, in electrochemically switched ion
exchange [18–22]. Various authors have recently reported
Ali H. Eltmimi1
Leon Barron1
Aran Rafferty2
John P. Hanrahan3
Olga Fedyanina4
Ekaterina Nesterenko1
Pavel N. Nesterenko4
Brett Paull1
1Irish Separation Science Cluster,National Centre for SensorResearch, Dublin City University,Glasnevin, Dublin, Ireland
2School of Mechanical andManufacturing Engineering,Dublin City University, Ireland
3Glantreo, Rubicon Centre, CITCampus, Bishoptown, Cork,Ireland
4ACROSS-Australian Centre forResearch on Separation Science,School of Chemical Sciences,University of Tasmania, Hobart,TAS, Australia
Received December 16, 2009Revised January 6, 2010Accepted January 8, 2010
Abbreviations: EDX, energy-dispersive X-ray spectroscopy;
HF, hydrofluoric acid; ICP-OES, inductively coupledplasma–optical emission spectroscopy; PCM, porouscarbon-based monolithic
Correspondence: Professor Brett Paull, Irish Separation ScienceCluster, National Centre for Sensor Research, Dublin CityUniversity, Glasnevin, Dublin 9, IrelandE-mail: [email protected]: 1353-1-7005503
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
J. Sep. Sci. 2010, 33, 1231–1243 1231
the preparation in various forms of porous carbon mono-
lithic structures [23–31]. Mareche et al. [32] proposed a new
kind of porous monolithic phase using exfoliated graphite
as a starting material. The exfoliated graphite material was
compressed into adjustable shapes and the porosity and
density of carbon block was shown to depend upon the
compression rate. Alvarez and Fuertes [33] prepared
carbon monoliths by using silica monoliths as sacrificial
templates. They retain the foamy vesicular structure and
exhibit a high surface area of 1800 m2/g, with a large
porosity made up of the framework-confined meso-pores of
around 3.4 nm. Recently, Xu et al. [34] reported a method for
the synthesis of a carbon monolith for chromatographic
applications with a tri-modal porous structure of macro/
macro/meso-pores by nano-casting and phase separation. In
this reported method a silica monolith was used as
the hard template and a mixture of styrene and divi-
nylbenzene, in the presence of initiator and porogenic
reagent (dodecanol), was filled into the void space of the
monolithic template.
The use of template particles, in particular silica based,
for the preparation of ordered nano- and micro-structured
carbon materials has been the topic of much attention in
recent years in the production of so-called ‘‘inverse opals’’.
For example, Zakhidov et al. [35] synthesised carbon inverse
opals by infiltrating silica opal plates with a phenolic resin.
They cured the resin at low temperature, pyrolysed the resin
then removed the silica with aqueous hydrofluoric acid
(HF). They also reported that the X-ray diffraction indicated
an amorphous structure consistent with a glassy carbon.
Further work followed in this area [36–39], and most
recently Guan et al. [40] in 2007 synthesised carbon micro-
particle arrays with different particle morphologies from
phenolic resin using a similar inverse opal templating
process.
The idea of using spherical particles as templates for the
preparation of monolithic rods of dimensions and physical
stability suitable for chromatographic application was first
generated in 1966 by John Knox, who proposed an ‘‘inverse
column’’ in which the stationary phase structure had the
same shape as the mobile phase (gas phase) in the particle
packed column [41]. Knox stated how such a structure could
be made ‘‘by filling a column with beads of some low
melting point substance, fusing the points of contact, filling
the gas space with a thermosetting resin, and, after curing,
melting out the original beads’’. However, it was only
recently that Liang et al. [42] using such an approach
reported on the preparation of a carbon-based monolithic
column for chromatography by pyrolysing a carbon rod
made of a copolymer of a resorcinol/iron(III) complex and
formaldehyde using silica particles as a template. SEM
images revealed macro-pores with diameters between 5 and
10 mm, while BET measurements show reasonable porosity
(0.5229 mL/g) for the monoliths and a developed surface
area of 163 m2/g. Preliminary separations of n-alkylben-
zenes were achieved with dichloromethane and methanol as
the mobile phase, with 1% n-hexane as an additive.
In the work presented here, a detailed examination of
the production, and subsequent physical characterisation, of
carbon monolithic rods of 1–10 mm macro-pore size is
presented, based on an adapted and optimised templating
procedure first proposed by Liang et al. [42]. Synthetic
carbon rods of up to 10 cm in length were produced using a
simple thermally initiated process, based upon pyrolysing a
cast rod made of a copolymer of a 25:75 phenol resin mixed
with formaldehyde in the presence of silica beads (1, 2, 5
and 10 mm in particle size, respectively). These rods were
physically characterised, and then clad using a variety of
methods to ensure elimination of all housing voids and to
enable chromatographic evaluation. Finally, the surface
modification of the carbon rods with gold nano/micro-
particles for further chromatographic application was also
investigated.
2 Materials and methods
2.1 Chemicals
Resorcinol (99%), formaldehyde (37 wt% aqueous solution),
ethyl alcohol, toluene (99.9%), ethylbenzene (99.91%),
butylbenzene (99.9%) and HF acid (38–40%) were obtained
from Sigma-Aldrich (Gillingham, UK). 1-Butanol and ferric
chloride (99%) were obtained from Riedel-De Haen, Seelze
(Hannover, Germany). Nucleosil 5 and 10 mm silica beads
(surface area 5 359 and 124 m2/g, respectively, pore sizes
5 50 and 100 A, respectively) were obtained from Macher-
ey-Nagel (Duren, Germany) and 1 and 2 mm silica particles
(surface area 5 809 and 617 m2/g, respectively, pore sizes
5 47.5 and 50.1 A, respectively) were obtained from
Glantreo (The Rubicon Centre, CIT Campus, Bishopstown,
Ireland). The approximate particle size range measured
using SEM for each silica gel (1, 2, 5 and 10 mm) was
0.8–1.6 mm, 1.4–2.2 mm, 3.8–5.2 mm and 7.9–12.4 mm,
respectively. Araldite epoxy resin was obtained from Bostik
Findley (Stafford, UK). Gold(III) chloride 99.99% and 6-
mercaptohexanoic acid 90% were obtained from Sigma-
Aldrich. Reagent water was obtained from a Millipore Milli-
Q water purification unit (Millipore, Bedford, MA, USA)
and was 18.2 MO or better. All chemicals were used as
received from the manufacturers.
2.2 Instrumentation and characterisation
A GFL water bath, model 1013, from Laborgerateborse
GmbH (Burladingen, Germany) was used to prepare the
precursor carbon rod. An EHRET thermovacuum oven from
Ehret Labor and Pharmatechnik GmbH, KG, Emmendin-
gen, Germany was used to complete the polymerisation of
the precursor carbon rod. For the pyrolysis of the rod, a
desktop alumina tube furnace, model GSL1300X, from MTI
(Richmond, USA) was used. The surface morphology of the
carbon monoliths was examined using Hitachi SEM/
J. Sep. Sci. 2010, 33, 1231–12431232 A. H. Eltmimi et al.
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
energy-dispersive X-ray spectroscopy (SEM/EDX) model
S-3000N VP, Oxford, UK. High-resolution SEM images of
the monolith surface morphology were achieved using a
field emission Hitachi S-5500 SEM (Hitachi High Technol-
ogies America, USA). Inductively coupled plasma–optical
emission spectroscopy (ICP-OES) was used for the analysis
of gold, using a Liberty 220 series instrument (Varian, Palo
Alto, USA). To characterise the 1, 2, 5 and 10 mm templated
carbon monolith structures, a surface area analyzer, model
2375 (Micromeritics Gemini, Georgia, USA) was used to
measure the specific surface area and pore volume using the
nitrogen adsorption/desorption technique. Dilatometric
analysis was carried out using a programmable horizontal
pushrod dilatometer, model Dil 402-E (Netzsch, Emmen-
dingen, Germany) to study the effect of the pyrolysis on the
carbon monoliths. Thermal analysis was carried out using
differential thermal analysis/thermo gravimetric analysis
(DTA/TGA), model STA 1500 (Stanton Redcroft, UK).
Additionally, pore size distribution of the carbon monoliths
was characterised by mercury porosimetry, model Autoscan
33 (Quantachrome, UK) with a nominal measurement
range of 3.2 nm to 5 mm over a range of pressure from
atmospheric to 228 MPa. Infra-red spectra of the monolithic
carbon material surface were recorded using a Spectrum GX
FT-IR instrument (Perkin-Elmer, Milano, Italy). Monolith
cladding images were obtained using a SVM340 Syncro-
nised Video Microscope from LabSmith (Livermore, CA,
USA). The carbon monolithic rods were then sealed into
short chromatographic columns and evaluated for flow-
through porosity and backpressure profiles, retention
selectivity for a range of test solutes, and ruggedness over
extended use.
2.3 Preparation of the carbon monolith rods
A modified procedure [42] was used for the preparation of
the carbon monolithic rods. A 1 g portion of silica particles
(either 1, 2, 5 or 10 mm) was dispersed in 1.5 g of 1-butanol
and sonicated for 1 h. A 0.18 g portion of ferric chloride
followed by the addition of 0.367 g of resorcinol was then
added into the silica suspension and dissolved by gentle
shaking. After the addition of resorcinol, the mixture
solution directly turned dark and the resorcinol/Fe(III)
complex was formed. A 0.3 g portion of an ice cooled, 37%
formaldehyde solution in water was introduced into the
mixture in one step with further gentle agitation. The
mixture was kept in an ice-water bath for 1 h with constant
stirring. After removal from the ice-water bath, the mixture
was slowly transferred into 7 mm id glass tubes which were
capped when filled. These tubes were then placed in a 901C
hot water bath for 15 h. The mixture polymerised into a
solid rod inside the glass tube. The resin rod detached from
the glass tube wall due to shrinking caused by polymerisa-
tion. The polymer rods were aged in the glass tube overnight
in the hot-water bath. Crack-free phenolic resin/silica rods
were then removed from the glass tubes and kept within a
fumehood for 72 h to slowly evaporate the majority of
residual solvent. Finally, the rods were thoroughly dried in a
vacuum oven at 801C overnight and further cured at 1351C
for 4 h to ensure complete polymerisation.
A horizontal tube furnace purged with N2 gas was
employed to pyrolyse the rods. The temperature was first
ramped from room temperature to 8001C at 2.51C/min, and
then held at 8001C for 2 h, to ensure complete carbonisa-
tion. A second ramp took place from 800 to 12501C at 101C/
min. The temperature was kept constant at 12501C for a
further 1 h. Then the furnace was allowed to cool naturally
to room temperature.
The silica particles and the iron catalyst were removed
from the rods by etching in concentrated HF acid (38–40%)
for 3 h and subsequently washed away with copious
amounts of distilled water. The porous carbon rods obtained
were finally thoroughly dried within a fumehood for 24 h.
2.4 Preparation of carbon/gold composite monolith
New carbon/gold composite monoliths were prepared
following a procedure adapted from that reported by
Dekanski et al. [43] detailing the modification of glassy
carbon surfaces with silver particles. The carbon monolithic
rod was immersed in a 100 mM chlorauric acid in pH 1.8
aqueous solution overnight followed by drying for 24 h. The
degree of gold functionalisation of the carbon rods was
evaluated through digestion in HCl and analysed for gold
content using ICP-OES. Carbon/gold composite monoliths
were further modified with 6-mercaptohexanoic acid,
following an adapted procedure reported by Huo and
Worden [44], by immersing the rod in 10 mM of the acid
solution overnight followed by drying for 24 h. A second
modification procedure was also investigated to modify
carbon rods after column cladding (see below), which
simply involved pumping 100 mM chlorauric acid (pH 1.8)
through the cladded carbon monolith overnight, followed by
flushing the column with deionised water.
2.5 Cladding of the carbon rod and chromatographic
characterisation
The initial range of monolithic carbon rods were clad in
heat-shrinkable PTFE tubing by heating at 3401C. Encased
carbon rods were then sealed (making sure all air pockets
were eliminated) into short PEEK HPLC columns (typically
30� 6 mm id) with epoxy resin. This column could then be
connected to a standard HPLC system. Later rods were
completely encased directly within epoxy resin of �5 mm
thickness, into which chromatographic tubing connections
were sealed directly interfacing with the end of the
monolithic rod.
Chromatographic characterisation of monoliths was
carried out on an Agilent 1200 chromatograph comprising a
vacuum degasser, quaternary pump, autosampler, column
J. Sep. Sci. 2010, 33, 1231–1243 Liquid Chromatography 1233
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
oven and UV-Vis detector. Both reversed-phase and normal-
phase chromatographic selectivity was evaluated for each
new column. Bare carbon monolithic columns were tested
for normal-phase selectivity through injections of 15 mL of
model mixtures containing 10 mg/mL each of butylbenzene,
acetophenone and nitrobenzene in hexane, with hexane as
the mobile phase delivered at 1 mL/min, with UV detection
at 254 nm. Reversed-phase chromatographic evaluation of
the monolithic rods was based on the retention of various
substituted phenols in a water–methanol (1:9) mobile
phase, delivered at 0.15 mL/min, with detection at 278 nm.
Methylbenzene exhibited no retention in either normal or
reversed-phase mode under the conditions used and so was
used as a dead-time marker for each new monolithic rod
produced. The gold and 6-mercaptohexanoic acid modified
carbon monoliths were tested for ion exchange capacity by
injection of 10 mL of a 1 mg/mL solution of imidazole in
1 mM sodium acetate, with photometric detection at
280 nm. 1 mM sodium acetate/acetic acid buffer solution
was used as an eluent at 1 mL/min flow-rate.
3 Results and discussion
The synthesis of porous carbon-based monolithic (PCM)
rods includes three main steps: preparation of silica particle
embedded polymer rods, drying and pyrolysis of the
organopolymer phase, and etching of silica phase. The
adapted method developed results in an intermediate glassy
carbon–graphite monolithic material, with a reported
graphite index value of �0.2 [42]. The intermediate structure
and degree of disorder would suggest some degree of
undesirable surface functionality, likely to negatively affect
upon chromatographic selectivity and efficiency. However,
PCM of the macro-porous dimensions produced here have
not previously been evaluated chromatographically, there-
fore potential selectivity and efficiency is unknown.
Physically, the ideal porous monolith for application as a
stationary phase within LC is of course one exhibiting a
completely homogeneous structure and uniform mono-
functional surface. Thus it is necessary to avoid any cracks
or large internal cavities formed by bubbles in the
preparation of the PCM rods. Both cracks and cavities can
appear during the drying and pyrolysis stages of forming
intermediate resorcinol-formaldehyde resin rods with
embedded silica particles, so the most critical part of this
synthetic process is the careful drying of such resin rods.
3.1 Preparation and pyrolysis of silica-embedded
precursor rods
As described in Section 2.3, silica particles of 1, 2, 5 and
10 mm diameters were used in this work. The support
surface area was typically in the range 300–900 m2/g, with
pore diameters in the range 4–40 nm. Using these silica
particles, monoliths were made according to the procedure
detailed in Section 2.3. To achieve crack-free and bubble-free
rods, the drying procedure was found to be critical. Rods
were aged overnight at 901C after the phenolic resin was
formed. Evolution of solvent molecules during aging from
within the resin creates stresses, which can result in cracks
within the wet gel. Slow aging allows the steady formation a
highly cross-linked bulk polymer to resist these stresses.
Additionally, keeping the ageing temperature below the
boiling point of the solvent used, here butanol, limits the
formation of bubbles inside the rod. The aged wet rod was
then dried slowly at room temperature inside a fumehood.
Careful control of the pyrolysis step, where carbonisa-
tion of phenol–formaldehyde resin occurs, is also a critical
aspect of monolith formation. The pyrolysis step carbonises
the inter-particulate phenolic resin to form the monolithic
structure, which is then subject to acid treatment to
remove the silica particles. Shrinkage will occur as a
result of carbonisation, as during this step water vapour,
carbon monoxide, and hydrogen gases are produced. Here,
dilatometry was used to investigate the shrinkage of
samples during the pyrolysis heating cycle. Figure 1 shows
the shrinkages measured for the four samples studied here.
For each of the rods, significant shrinkage was observed
during the pyrolysis process as shown in Fig. 1. Shrinkage
commenced at approximately 2001C and was rapid up to
approximately 8001C. During this step a slow heating rate of
2.51C/min was used, to minimise the potential for thermal
shock, rapid gas evolution and cracking. A plateau in the
shrinkage occurs coinciding with the 2 h isothermal hold at
8001C. The shrinkage slows down as the structure is
consolidated. After the isothermal hold, the heating rate is
increased to 101C/min and heating proceeds to 12501C.
During this new heating phase, a second phase of rapid
shrinkage is observed. The temperature is then held
constant at 12501C for 1 h. The observed behaviour was
different for the 1 and 2 mm samples, compared to the 5 and
10 mm samples. For the former samples, a sharp shrinkage
was observed during the 101C/min heating ramp to 12501C.
During the isothermal hold at 12501C these samples tended
to plateau out.
The 5 and 10 mm samples exhibited a steady rapid
shrinkage from 8001C to the end of the isothermal hold at
12501C. All of the samples continued to contract, almost
identically, during cooling. From Fig. 1 it can be seen that
the overall shrinkage of the rods ranged from approximately
9% to almost 12%. Shrinkage increased proportional to the
particle size of the silica used. The larger the particle size
the larger the gaps between particles. The shrinkage
observed in Fig. 1 can be partly attributed to carbonisation
of the resin located between the silica beads. Further
shrinkage can be attributed to the non-uniformity of silica
particles themselves. During pyrolysis the silica particles
will rearrange themselves and shift slightly until such time
as they become locked in position. Another possible reason
for high shrinkage could be non-uniform infiltration of
phenolic resin into the void spaces between silica particles.
Infiltration can be difficult to achieve and is dependent on
J. Sep. Sci. 2010, 33, 1231–12431234 A. H. Eltmimi et al.
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
the conditions used to mix the phenolic resin and silica
beads (and achieve full wetting of the silica), including the
use of stirring or use of a vacuum to force the resin into the
voids. The viscosity of the resin solution is another variable
which can determine the degree of infiltration.
Ko et al. [45] have carried out an in-depth study of the
behaviour of phenolic resins during pyrolysis. Ko et al.showed how volatiles such as H2O, CO, CO2, H2 and
other gases were removed during pyrolysis, leading to
weight loss and shrinkage. Significantly, Ko et al. showed
that 91% of the total gases evolved during pyrolysis occurred
in the temperature range of 400–7001C. Below 7001C
shrinkage of the phenolic resin increased linearly with
temperature and was attributed to the rearrangement of the
carbon structure. Above 10001C the shrinkage neared
completion due to the packing and cross-linking of the
glassy carbon structure. Total shrinkage of approximately
17% was observed. In the study conducted by Ko et al., a
total weight loss of phenolic resin of 435% was observed up
to a temperature of 24001C. Briefly, 94% of this weight loss
occurred below 9001C.
DTA/TGA analysis was used to investigate thermal
behaviour and weight loss of the 1 mm silica powder and also
the 1 mm silica–carbon rod during pyrolysis. For the powder
sample, a weight loss of 7.4% was observed up to 12501C.
This is expected since porous silica generally acts as a
sponge for organics and contaminants.
For the combined silica–carbon sample, a total weight
loss of 50% was observed (Fig. 2). For the carbon–silica
sample, a sharp endothermic peak was observed at
approximately 551C. A broad endothermic peak occurred
between 240 and 4001C. The disturbances observed at 800
and 12501C are artefacts created by the pyrolysis profile
which contained isothermal dwell times at these tempera-
tures.
Clearly, the weight loss is dramatic and is accompanied
by a vigorous exothermic reaction, even at a heating rate of
2.51C/min. According to Ko et al. [45] the condensation
0
400
800
1200
1600
0 200 400 600 800
Time (min)
Tem
pera
ture
(°C
)
-0.15
-0.11
-0.07
-0.03
0.01
dL/L
o
Temp
dL/L0 1 µm
2 µm
5 µm
10 µm
Figure 1. Dilatometry profilesof the 1, 2, 5 and 10 mm silica-particle templated monolithicrods subjected to a tempera-ture programmed pyrolysiscycle.
40
50
60
70
80
90
100
20 220 420 620 820 1020 1220
Temperature (°C)
% W
eigh
t rem
aini
ng
-2.00E-05
-1.00E-05
0.00E+00
1.00E-05
2.00E-05
3.00E-05
4.00E-05
5.00E-05
6.00E-05
Del
ta T
(mic
rovo
lts)
% weight remaining
Delta T
Figure 2. Thermo-gravimetric beha-viour of the 1 mm silica–carbon rodduring pyrolysis under flowing nitro-gen.
J. Sep. Sci. 2010, 33, 1231–1243 Liquid Chromatography 1235
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
of polymer structures results in gas evolution and pore
creation in the carbonised sample. Above 7001C the effects
of chemical densification is greater than the formation of
pores. It can be seen from Fig. 2 that above approximately
4001C the bulk of the burn-off activity and associated weight
loss has been completed. However, it would be reasonable to
suggest that even lower heating ramp rates than 2.51C/min
in the region up to 4001C could be beneficial, such is the
volatility of the reaction in this region.
3.2 Characterisation of porous structure of carbon
monoliths
The above careful control of monolith preparation and
pyrolysis conditions resulted in the ability to produce carbon
monoliths of up to 100 mm in length without internal and
surface fissures or cracks. Optical microscopy of the
monolith surface showed a smooth homogenous surface
structure along the complete length of the monolithic rod
(Fig. 3A). SEM analysis of the monolithic rods in cross-
section revealed interconnected rigid porous structures for
each version of the monolith and verified the quantitative
removal of the embedded silica particles using HF acid
(Fig. 3B, 1 mm particle templated rod; 3C, 2 mm particle
templated rod; 3D; 5 mm particle templated rod; and 3E,
10 mm particle templated rod). HR-SEM was used to image
the smooth carbon surface morphology, although the
presence of isolated pores in the carbon skeleton of the
order 10–20 nm could also be identified (Fig. 3F).
EDX analysis carried out on cross sections of the
carbon monolith samples (1, 2, 5 and 10 mm) gave further
evidence that the silica was totally removed from the carbon
rod after treatment in HF acid (Fig. 4). Only negligible trace
amounts of silica residue could be detected in the HF
treated rods.
Surface area analysis (BET) was employed to investigate
characteristics of the monoliths before and after pyrolysis.
The silica particles (powder) were also characterised. Table 1
summarises the measured surface areas and pore volumes
of the silica particles and monoliths investigated in this
study. The 1 mm silica particles exhibited a high surface area
of 934 m2/g, indicative of a highly porous powder. The
powder exhibited a classic Type 1 isotherm, indicative of
micro-porous materials. The pore volume distribution
peaked at 2.1 nm (Fig. 5A). A similar isotherm was observed
for the 2 mm particles (powder).
In terms of the monolithic carbon–silica samples, lower
surface areas, compared to the silica particles, were
observed. For the 1 mm carbon–silica monolith before
pyrolysis, a surface area of 315 m2/g was recorded. The
surface area in this case is made up of the non-porous
phenolic resin element and the accessible silica element of
the composite. In this case it may be expected that this
monolithic sample would show very similar characteristics
to the 1 mm particle sample. However, a Type 4 isotherm,
commonly associated with the presence of meso-porosity,
was observed (Fig. 5B). Capillary condensation gave rise to a
hysteresis loop at the mid-range of the pressures investi-
gated. A hysteresis loop was also observed close to unity,
similar to that observed for the 1 mm particle sample. A peak
was observed for pores between 2–9 nm. Presumably, this is
the same porosity as was found for the powder sample. In
this sample, however, there is a distinct uptake in volume
Figure 3. Top: (A) Photograph of a 90�4 mm porous carbonmonolith, with inset showing microscope image of homoge-nous surface. Middle: SEM images of carbon monoliths with (B)1 mm macro-pores, (C) 2 mm macro-pores, (D) 5 mm macro-poresand (E) 10 mm macro-pores after removal of the silica beads.Magnification equal for images (B)–(E). Bottom: HR-SEM imageof the carbon surface morphology of the 5 mm templatedmonolith.
J. Sep. Sci. 2010, 33, 1231–12431236 A. H. Eltmimi et al.
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
absorbed beyond 15 nm and this continues to rise to 45 nm
and beyond. This could be attributed to pores formed by the
contact points between silica particles which were not
completely filled by phenolic resin.
Following pyrolysis and HF treatment, the surface area
of the 1 mm sample dropped to 178 m2/g. This further drop
is due to the removal of the porous silica beads, which were
responsible for the high surface area prior to pyrolysis, thus,
leaving a skeleton of large pores (approximately 1 mm). The
adsorption-desorption isotherm for the sample after pyro-
lysis was also a Type 4 isotherm, but in this case it exhibited
hysteresis from relative pressure 0.4 to unity, indicative of
pores in the meso-range and beyond. Pores were shown to
exist at high volumes below 15 nm but also up to 200 nm
and further, beyond the range of the instrument. This was
expected since pores of approximately 1 mm were present
following removal of the silica.
Mercury porosimetry was carried out to further inves-
tigate the silica particles and monoliths. The 1 mm silica
powder was compacted to form a disc. This was done to
minimise the possibility of powder being sucked into the
mercury under vacuum. It also allowed for more direct
comparison with the monolithic samples. Powders behave
differently under conditions of mercury filling in terms of
filling of free space between particles. In terms of pore
volume, a large distribution peak was observed at approxi-
mately 500 nm. This was attributed to the spaces between
the contact points of the 1 mm silica spheres. No other pores
were observed for the volume distribution. However, a
different picture emerged when the pore number fraction
was taken into consideration. In this case, a very small peak
was observed at 500 nm. However, a sharp rise was observed
at approximately 30 nm and continued down to 6 nm,
something that was entirely absent from the volume
distribution. Clearly, large quantities of small pores exist,
but these possess a low pore volume. It is believed that these
pores correspond to the pores within the silica particles, as
found earlier by BET. In the volume distribution, the
presence of these small pores is being masked by a smaller
number of large volume pores.
Porosimetry was then carried out on the carbon–silica
(1 mm) monolith before pyrolysis. The pore distribution of
Figure 4. EDX images of carbonmonolith samples (left) before and(right) after HF treatment to removethe silica template particles.
Table 1. Surface area and pore volumes measurements of the silica powders and carbon monoliths
Nominal
particle size
Silica particles (mm) Surface area of
monolith (m2/g)
Pore volume of
monolith (mL/g)
Macro-pore size
distributiona) (mm)
Total
porosity
Measureda) average
particle size
Measureda) particle
size %RSD
Surface area
(m2/g)
Pre-
pyrolysis
Post-
pyrolysis
Pre-
pyrolysis
Post-
pyrolysis
Range Mean
1 1.3 19 934 315 178 0.24 0.23 0.5–1.2 0.7 0.25
2 1.8 16 616 235 154 0.48 0.23 1.5–2.0 1.6 0.66
5 4.0 34 359 196 84 0.30 0.095 4.5–6.0 5.3 0.70
10 9.9 16 339 195 125 0.57 0.32 9.2–10.1 9.7 –
a) Determined using SEM.
J. Sep. Sci. 2010, 33, 1231–1243 Liquid Chromatography 1237
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
this sample is shown in Fig. 6A. This sample exhibited
pores of approximately 230 nm. This was somewhat unex-
pected, in that it may well have been reasonable to expect a
result similar to the compacted 1 mm silica powder,
assuming that the phenolic resin was non-porous. Pore
number fraction analysis (data not shown) revealed a minor
peak at approximately 230 nm, but the main activity occur-
red at 6–7 nm and 10–20 nm. These latter two peaks are
believed to correspond to the internal pores in the silica
beads. Less clear was the presence of the volumetric peak at
230 nm. There are a number of possible reasons for this as
previously mentioned, including impartial infiltration of
phenolic resin into the void spaces between silica particles,
or too-high viscosity of the resin solution affecting degree of
infiltration. The raw volume intruded for this sample was
1.23 cm3/g as compared with 2.36 cm3/g for the compacted
powder, adding weight to the assumption that the silica
particles were not totally coated, and the spaces between
silica particles were not completely filled.
Following pyrolysis and silica removal, the pore distri-
bution changes again, as shown in Fig. 6B. As can be seen,
the pore distribution forms a sharp peak at approximately
800 nm. The raw volume intruded has increased to
3.15 cm3/g indicative of the increase in pore size. The sharp
peak found at 250 nm prior to pyrolysis (attributed to inter-
particle voids) is no longer in evidence. However, the large
peak at 800 nm does possess a tail which extends to
approximately 150 nm. The large peak observed in Fig. 6B
ranges from 1200 to 500 nm, and is due to the pores left by
the removal of the 1 mm silica particles. The pores are
expected to be less than 1 mm in diameter due to shrinkage
during pyrolysis, which was shown in Fig. 1 to be approxi-
mately 9% for the sample made from 1 mm silica particles.
Other variances are expected due to the imperfect particle
size distribution of the silica particles, i.e. �0.8–1.6 mm from
SEM averaging measurements.
The pore number fraction data was examined. The pore
number fraction is the number of pores in a narrow range of
pore sizes relative to the total number of pores in the
measurement range. The pore number fraction is found by
dividing the number of pores in a small interval by the total
number of pores. The resulting number represents the
fractional amount of all the pores which are found in that
particular interval. The calculation of pore number fraction
is based on the assumptions that all pores are cylindrical
and are of equal length. Pore number fraction is a dimen-
sionless quantity. When the pore number fraction data was
examined, it is clear that many other fine pores also exist in
the pyrolysed sample (Fig. 7).
The predominant pore size (in volume terms) are the
pores remaining after the removal of the 1 mm silica parti-
cles. However, other pores also exist, ranging from 9 nm to
-10
65
140
215
290
365
0 2 4 6 8 10Pore Diameter (nm)
Incr
emen
tal P
ore
Are
a (m
2 .g-1
)
140
170
200
230
260
0 0.5 1Relative pressure (P/Po)
adsorption
desorption
0
0.009
0.018
0.027
0.036
0.045
0 10 20 30 40 50
Pore diameter (nm)
Incr
emen
tal p
ore
vol (
cm3 .
g-1
)
B
A
0
100
200
300
0 0.5 1Relative pressure (P/Po)
adsorption
desorption
Figure 5. (A) Adsorption-desorption isotherm (inset) and poredistribution of 1 mm silica powder. (B) Isotherm (inset) and poredistribution of the 1 mm carbon–silica monolith before pyrolysis.
-0.1
0.3
0.7
1.1
10 100 1000 10000Diameter (nm)
Max Ø = 753 nmB
-0.1
0.2
0.5
0.8
10 100 1000 10000Diameter (nm)
dV/d
log
(cm
3 .g-1
)
Max Ø = 231 nmA
Figure 6. Pore distribution for1 mm silica–resin monolith (A)before pyrolysis and (B) afterpyrolysis and removal of silicaparticles.
J. Sep. Sci. 2010, 33, 1231–12431238 A. H. Eltmimi et al.
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
approximately 70 nm. This phenomenon has been observed
by other authors. For example, Liang et al. [42] reported four
types of pores in their final samples prepared using 10 mm
template silica. These were macro-pores formed from the
original silica spheres, macro-pores from the contact points
of the spheres, meso-pores formed by silica diffusion into
carbon and micro-pores formed by pyrolysis of phenol–
formaldehyde resin.
3.3 Housing, modification and chromatographic
evaluation
Housing/cladding the carbon monoliths was carried out
using two approaches as detailed in Section 2.5. Both
approaches proved successful with evaluation using optical
microscopy confirming complete elimination of voids along
the monolithic edge (Fig. 8). Both cladding techniques were
subjected to applied column backpressures of up to
2.41 MPa without sign of stress-induced voids. Typical
monolith backpressures for the column housed/clad mono-
liths ranged from 0.77 MPa on a 5 mm templated
(30� 6 mm id) bare carbon monolith, to 2.21 MPa on a
1 mm templated (73� 4 mm id) carbon/gold composite
monolithic rod, with applied flow rates of 1 mL/min.
Column LC can provide important information on the
quality and properties of the prepared carbon monolith.
First, the right peak shape or elution profile of any retained
solute will confirm the absence of cracks or significant
internal cavities within the monolithic rod and reflect the
uniformity of the porous structures. Second, the recording
of backpressure at the top of carbon monolith column
provides information on the permeability of the material
and the interconnectivity of the pores within the monolith.
Finally, the evaluation of selectivity and retention of model
substances can provide information on adsorption proper-
ties of the surface of the bare monoliths.
All of the prepared monoliths had well-developed
surface areas of between 84 and 178 m2/g (Table 1).
However, pure glassy carbon is known to be a relatively poor
LC phase, showing only weak retention for both polar and
non-polar solutes under normal and reversed-phase condi-
tions [46]. Despite this limitation, here due to the inter-
mediate composition of the newly structured PCM rods, it
was important to evaluate the selectivity and efficiency
exhibited. Using a short 30� 6 mm id 5 mm PCM rod rela-
tively weak (ok 5 1) retention of aromatic hydrocarbons
including toluene, ethylbenzene and butylbenzene was
noted in hexane. These results are in a good agreement with
data reporting the weak retention of benzene, toluene and
xylenes on porous graphitic carbon (Hypercarb) in heptanes,
obtained by Kaliszan et al. [47].
The PCM phase exhibited stronger retention for more
polar molecules under normal-phase LC conditions (as
shown later in Fig. 11A). The conditioning of the carbon
surface using mineral acids was investigated to potentially
increase retention of polar solutes and homogenise the
surface functionality. The surface was conditioned sequen-
tially using 0.5 and 2 M solutions of sulphuric acid, followed
by a 2 M solution of 50:50 sulphuric/nitric acid. Following
such conditioning steps the retention factors for the more
polar species such as nitrobenzenene and acetophenone
were increased significantly in comparison with butylben-
zene under normal-phase conditions, indicating increased
presence of polar sites on the carbon surface (Table 2).
(These sites were confirmed through FT-IR evaluation of the
carbon surface showing the presence of various oxygen
containing functional groups. Figure 9 shows a typical FT-
IR spectrum obtained. The broad absorption peak at
Figure 7. Microscopy imagesof (A) epoxy resin clad 5 mmtemplated carbon monolithand (B) a 5 mm templatedcarbon monolithic rod cladwithin a heat-treated Teflonsheath with an epoxy exteriorsealing layer for housing with-in a PEEK column.
Figure 8. Pore number fraction of pores for the 1 mm carbonmonolith after pyrolysis.
J. Sep. Sci. 2010, 33, 1231–1243 Liquid Chromatography 1239
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
3429 cm�1 corresponds to the stretching O–H vibrations.
The characteristic C 5 O stretching vibration band is seen at
1701 cm�1. One more broad band at 1152 cm�1 was prob-
ably caused by a superposition of bands associated with
vibrations of the C-O-C group. A the same time the presence
of very small bands associated with asymmetric methyl
group vibrations at 2960 cm�1 or with asymmetric methy-
lene group vibrations at 2930 cm�1 prove the almost
complete pyrolysis of the original polymer.)
The retention of polar molecules such as phenols on
carbonaceous stationary phases in methanol–water mobile
phases is mainly due to ‘‘the polar retention effect on
graphite’’ [48], when the charge of the polar adsorbate
approaching the surface induces the appearance of a dipole
at the surface. Under these conditions, hydrophobic inter-
actions play a less important role. Because of the presence of
numerous oxygen containing groups at the surface of the
PCM, the retention of phenols can be increased through the
formation of hydrogen bonds. To investigate the retention
mechanism further, the retention of a number of substi-
tuted phenols was determined on a 50� 6 mm id PCM
column. The obtained results (Table 3) show the retention
order of the various phenols on the PCM, which were of a
very similar order to that recently recorded on a porous
graphitic carbon column (Hypercarb) [49]. However,
obviously there was no strong correlation between the
retention of phenols and their hydrophobicity and pKa
values.
The PCM rods were further modified with gold nano-
particles by self-assembly in solution, following sponta-
neous induced reduction upon functional groups at the
surface. Initial modification of each rod was carried out by
simple immersion of the monolithic rods in 0.1 M tetra-
chloroauric acid. Initial experiments showed the dimen-
sions of the surface immobilised gold particles were
proportional to immersion time. Gold particles ranging
from 50 to 150 nm in high degrees of surface coverage could
be obtained using this technique, as shown in Fig. 10,
which compares a typical SEM of an unmodified (Fig. 10A)
and gold modified 5 mm templated PCM (Fig. 10B). A
typical gold-modified carbon monolith was subsequently
immersed in HCl and the digest was analysed for gold
Table 2. Retention factors for test solutes on glassy carbon
monolith following surface modification with mineral
acids
Monolith surface modification Retention factor, k
Butylbenzene Nitrobenzene
No modification 0.29 1.4
0.5 M H2SO4 0.32 2.0
2.0 M H2SO4 0.32 3.9
2.0 M HNO3/H2SO4 0.33 9.1
Figure 10. SEM images of the 5 mm templated carbon monolith (A) before modification and (B) after modification with gold nano/micro-particles (50–150 nm) through immersion of the carbon rod, (C) after modification through dynamic modification via pumpingtetrachloroauric acid solution through clad monolith.
Table 3. Retention volume (VR, mL) of phenols in methanol–-
water (9:1) eluent on porous glassy carbon monolith
and selectivity (k) for same on porous graphitic carbon
[49]
Phenol pKa log P VR (mL) k [38]
2-tert-Butyl-4-methylphenol 11.72 3.97 1.84 0.57
Phenol 9.99 1.46 3.27 0.30
4-Methylphenol 10.26 1.94 5.02 0.76
2-Methoxyphenol 9.98 1.32 5.69 0.83
4-Methoxyphenol 10.10 1.58 9.23 0.91
2-Chlorophenol 8.56 2.15 11.72 0.95
Figure 9. FT-IR spectrum of the carbon monolith surface.
J. Sep. Sci. 2010, 33, 1231–12431240 A. H. Eltmimi et al.
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
content using ICP-OES. The results indicated a 1.62% w/w
Au content using the above modification method. Conduc-
tivity measurements confirmed the gold modified carbon
monolithic exhibited higher conductance (322 S/m) than
recorded for the bare PCM rod (176 S/m).
Overnight immersion of the gold modified monolithic
phase in 10 mM of 6-mercaptohexanoic acid solution,
followed by drying for 24 h, was carried out to convert the
surface of modified monolith to a weak cation exchanger.
The modified rod was then clad as described above and
evaluated chromatographically, using imidazole as a stan-
dard solute, with a 1 mM aqueous acetic acid/acetate mobile
phase (pH 3.96). Figure 11B shows the retention of the
imidazole cation on the short 30 mm gold particle modified
5 mm templated PCM. Although the efficiency was low,
calculated using N ¼ 5:55ðtR=W1=2Þ2 as 1270 N/m, for the
imidazole peak, it was clear that cation exchange based
retention was evident.
A 73� 4 mm id 1 mm templated PCM was clad within
an epoxy resin column and subsequently modified with gold
through pumping at 1 mL/min 100 mM tetrachlorauric acid
(pH 1.8) through the cladded carbon monolith overnight.
Further modification with 6-mercaptohexanoic acid was
achieved again through pumping the solution through the
previously modified monolith. Using this dynamic modifi-
cation technique a considerably lower degree of surface
coverage was achieved (equal to only 0.011% w/w Au for a
5 mm templated monolith, as shown in Fig. 10C). Using the
above moderately modified 1 mm templated column, the
retention of neutral resorcinol and cationic imidazole was
investigated.
The 1 mm templated monolith exhibited both a signifi-
cantly reduced void volume compared to the 5 mm templated
monolith. Given the total volumes of the two monolithic
rods, namely 0.849 cm3 for the short 5 mm templated
monolith compared to 0.917 cm3 for the longer 1 mm
templated rod, the observed void volumes correspond to a
large difference in both column density and overall porosity,
equivalent to �25% total porosity for the 1 mm rod
compared to �70% for the 5 mm templated material. Given
the surface areas of the two materials and the column
dimensions, based upon a glassy carbon density of 1.5 g
cm3, this corresponds to total surface areas of �180 m2 for
the 1 mm column and only �30 m2 for the 5 mm column.
This reflects the low surface area of the 5 mm material noted
in Table 1, indicative of non-ideal homogeneity and packing
density in the preparation of the original silica-polymer mix,
resulting in potential presence of internal voids and larger
than expected overall porosity. Chromatographically, this
difference is reflected in the observed peak shapes shown in
Fig. 11, where the 1 mm templated rod exhibited an
improvement in both peak width (approximately 2.1 min at
-0.5
0.5
1.5
2.5
3.5
4.5
5.5
6.5
0 2 4
Abs
(m
AU
)
A B C D
1.2 min
O
0.7 min
-0.15
0.05
0.25
0.45
0.65
0 2 4
t00.6 min
Imidazole1.6 min
NH
N
-20
30
80
130
180
230
0 1 2
30 oC Rt = 0.27 min
60 oC Rt = 0.30 min
-40
10
60
110
160
210
260
310
0 1 2Time (min)
NH
N
30 oC Rt = 0.23 min
60 oC Rt = 0.62 min
HO OH
Figure 11. (A) Separation of butylbenzene and acetophenone (100 mg/L) in hexane on a 30� 6 mm id 5 mm templated carbon monolithiccolumn. (B) Retention of imidazole cation (1 mg/L) on a gold-mercaptohexanoic acid modified 30� 6 mm id 5 mm templated carbonmonolithic column using a 1 mM sodium acetate mobile phase, with photometric detection at 280 nm. (C) Peak shape for resorcinol onthe gold-mercaptohexanoic acid modified 73� 4 mm id 1 mm templated carbon monolith over 30–601C. (D) Increasing retention ofimidazole on the gold-mercaptohexanoic acid modified 73�4 mm id 1 mm templated carbon monolith over 30–601C.
J. Sep. Sci. 2010, 33, 1231–1243 Liquid Chromatography 1241
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
baseline for imidazole on the 5 mm modified rod compared
to �1.5 min on the 1 mm phase under similar conditions)
and symmetry (Fig. 11C and D), although a reduced overall
ion exchange capacity was apparent from direct retention
comparisons between the two gold modified monoliths
under similar chromatographic conditions.
The effect of temperature upon retention was investi-
gated to confirm the dominance of ion exchange in the
retention mechanism on the modified monoliths. Despite
low capacity the effect upon retention of imidazole
compared with the resorcinol was significant, with no
temperature effect upon retention or peak shape noted
between 30 and 601C for the later neutral solute, but a
significant increase in retention observed for the imidazole
cation. This effect was also seen with 5 and 10 mm templated
gold–mercaptohexanoic acid modified monoliths, with the
reproducibility of peak shapes for both retained and unre-
tained species following multiple temperature variations
confirming thermal stability of both the surface modifica-
tion and the carbon rod cladding itself (Fig. 11C shows six
overlaid chromatograms spanning the temperature range
30–601C).
The results show clearly that the potential for higher
capacity modified monoliths exists through utilising the
immersion based modification approach on the monoliths
of higher surface area and greater density, namely the 1 and
2 mm templated materials. This, together with alternative
applications for these modified glassy carbon monoliths,
including electrochemical studies, is the subject of ongoing
research.
4 Concluding remarks
The research shows the technical feasibility of production of
monolithic carbon monolithic rods of dimensions suited to
future potential chromatographic applications, based upon
template particles ranging from 1 to 10 mm. In this study,
the synthetic process described was reliable and can
reproducibly produce rods of specific physical characteris-
tics. At this stage in the work it is not possible to state how
minor differences in the physical structure of the rods
would ultimately show themselves, for example, in chro-
matographic performance. This would need to be deter-
mined in further work on covalently modified rods.
The physical properties of the intermediate glassy –
graphitic carbon monoliths has been fully characterised,
together with evaluation of simple surface modification
techniques resulting in a stable gold–mercaptohexanoic
acid modified material, which exhibits temperature stable
and reproducible ion exchange capacity. The results
obtained to-date show limited suitability to the efficient
separation of small molecules in the monolithic formats so
far investigated, although the potential application to larger
bio-molecules is under investigation, which may provide
better separation efficiencies. These new composite conduc-
tive materials have potential further novel application within
separation science, together with possibilities in flow
through electrodes, sensors and electrochemical micro-
reactors.
The authors acknowledge Science Foundation Ireland forfunding this project (Grant 05/RFP/CHE0011) and thankBrendan Twamley for his assistance with SEM imaging. BrettPaull would also like to thank the Royal Society of Chemistry forfunding provided to complete this work under the JournalsGrants for International Authors Scheme.
The authors have declared no conflict of interest.
5 References
[1] Minakuchi, H., Nakanishi, K., Soga, N., Ishizuka, N.,Tanaka, N., Anal. Chem. 1996, 68, 3498–3501.
[2] Tennikova, T. B., Belenkii, B. G., Svec, F., J. Liq. Chro-matogr. 1990, 13, 63–70.
[3] Paull, B., Nesterenko, P. N., Trends anal. Chem. 2005,24, 295–303.
[4] Ross, W. D., Jefferson, R. T., J. Chromatogr. Sci. 1970, 8,386–389.
[5] Hansen, L. C., Sievers, R. E., J. Chromatogr. Sci. 1974,99, 123–140.
[6] Liao, J. L., Zhang, R., Hjerten, S., J. Chromatogr. 1991,586, 21–26.
[7] Svec, F., Frechet, J. M. J., Anal. Chem. 1992, 64,820–822.
[8] Nakanishi, K., Soga, N., J. Am. Ceram. Soc. 1991, 74,2518–2530.
[9] Nakanishi, K., Soga, N., J. Non-Cryst. Solids 1992, 139,1–13.
[10] Cabrera, K., J. Sep. Sci. 2004, 27, 843–852.
[11] Motokawa, M., Ohira, M., Minakuchi, H., Nakanishi, K.,Tanaka, N ., J. Sep. Sci. 2006, 29, 2471–2477.
[12] Glenn, K. M., Lucy, C. A., Haddad, P. R., J. Chromatogr.A 2007, 1155, 8–14.
[13] O’Riordain, C., Gillespie, E., Connolly, D., Nesterenko, P.N., Paull, B ., J. Chromatogr. A 2007, 1142, 185–193.
[14] Paull, B., Nesterenko, P. N., Trends Anal. Chem. 2005,24, 295–303.
[15] Randon, J., Huguet, S., Piram, A., Puy, G., Demesmay,C., Rocca, J. L., J. Chromatogr. A 2006, 1109,19–25.
[16] Hoth, D. C., Rivera, J. G., Colon, L. A., J. Chromatogr. A2005, 1079, 392–396.
[17] Zizkovsky, V., Kucera, R., Klimes, J., Dohnal, J.,J. Chromatogr. A 2008, 1189, 83–91.
[18] Antrim, R. F., Scherrer, R. A., Yacynych, A. M., Anal.Chim. Acta 1984, 164, 283–286.
[19] Ge, H. L., Wallace, G. G., J. Liq. Chrom. 1990, 13,3245–3260.
[20] Keller, D. W., Ponton, L. M., Porter, M. D., J. Chroma-togr. A 2005, 1089, 72–81.
J. Sep. Sci. 2010, 33, 1231–12431242 A. H. Eltmimi et al.
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
[21] Keller, D. W., Porter, M. D., Anal. Chem. 2005, 77,7399–7407.
[22] Muna, G. W., Swope, V. M., Swain, G. M., Porter, M. D.,J. Chromatogr. A 2008, 1210, 154–159.
[23] Amadou, J., Begin, D., Nguyen, P., Tessonnier, J. P.,Dintzer, T., Vanhaecke, E., Ledoux, M. J., Pha-Huu, C.,Carbon 2006, 44, 2587–2589.
[24] Job, N., Thery, A., Pirard, R., Marien, J., Kocon, L.,Rouzaud, J. N., Beguin, F., Pirard, J. P., Carbon 2005, 43,2481–2494.
[25] Job, N., Sabatier, F., Pirard, J. P., Crine, M., Leonard, A.,Carbon 2006, 44, 2534–2542.
[26] Laszlo, K., Onyestyak, G., Rochas, C., Geissler, E.,Carbon 2005, 43, 2402–2405.
[27] Shi, Z. G., Feng, Y. Q., Xu, L., Da, S. L., Liu, Y., Mater.Chem. Phys. 2006, 97, 472–475.
[28] Siyasukh, A., Maneeprom, P., Larpkiattaworn, S.,Tonanon, N., Tanthapanichakoon, W., Tamon, H.,Charinpanitkul, T., Carbon 2008, 46, 1309–1315.
[29] Tonanon, N., Siyasukh, A., Wareenin, Y., Char-inpanitkul, T., Tanthapanichakoon, W., Nishihara, H.,Mukai, S. R., Tamon, H., Carbon 2005, 43, 2808–2811.
[30] Wang, D., Smith, N. L., Budd, P. M., Polym. Intern. 2005,54, 297–303.
[31] Lu, X., Caps, R., Fricke, J., Alviso, C. T., Pekala, R. W., J.Non-Cryst Solids 1995, 188, 226–234.
[32] Mareche, J. F., Begin, D., Furdin, G., Puricelli, S., Pajak,J., Albiniak, A., Jasienko-Halat, M., Siemieniewska, T.,Carbon 2001, 39, 771–773.
[33] Alvarez, S., Fuertes, A. B., Mater. Lett. 2007, 61,2378–2381.
[34] Xu, L. Y., Shi, Z. G., Feng, Y. Q., Microporous Meso-porous Mat. 2008, 115, 618–623.
[35] Zakhidov, A. A., Baughman, R. H., Iqbal, Z., Cui, C. X.,Khayrullin, I., Dantas, S. O., Marti, I., Ralchenko, V. G.,Science 1998, 282, 897–901.
[36] Zakhidov, A. A., Khayrullin, I. I., Baughman, R. H.,Iqbal, Z., Yoshino, K., Kawagishi, Y., Tatsuhara, S.,Nanostruct. Mater. 1999, 12, 1089–1095.
[37] Kajii, H., Take, H., Yoshino, K., Synthetic Met. 2001, 121,1315–1316.
[38] Take, H., Matsumoto, T., Hiwatashi, S., Nakayama, T.,Niihara, K., Yoshino, K., Jpn. J. Appl. Phys. 2004, 43,4453–4457.
[39] Aliev, A. E., Lee, S. B., Baughman, R. H., Zakhidov, A. A.,J. Luminesc. 2007, 125, 11.
[40] Guan, G. Q., Kusakabe, K., Ozono, H., Taneda, M.,Uehara, M., Maeda, H., J. Mater. Sci. 2007, 42,10196–10202.
[41] Knox, J. H., Anal. Chem. 1966, 38, 253–261.
[42] Liang, C., Dai, S., Guiochon, G., Anal. Chem. 2003, 75,4904–4912.
[43] Dekanski, A., Stevanovic, J., Stevanovic, R., Jovanovic,V. M., Carbon 2001, 39, 1207–1216.
[44] Huo, Q., Worden, J. G., J. Nanopart. Res. 2007, 9,1013–1025.
[45] Ko, T., Kuo, W., Chang, Y., J. Appl. Polym. Sci. 2000, 81,1084–1089.
[46] Gilbert, M. T., Knox, J. H., Kaur, B., Chromatographia1982, 16, 138–146.
[47] Kaliszan, R., Osmialowski, K., Bassler, B. J., Hartwick,R. A., J. Chromatogr. 1990, 499, 333–344.
[48] Knox, J. H., Ross, P., Advances in Chromatography,vol. 37, Marcel Dekker, New York 1997, p. 73.
[49] Fedyanina O. N., Nesterenko P. N., Russ. J. Phys. Chem.A 2010, 84, 476–480.
J. Sep. Sci. 2010, 33, 1231–1243 Liquid Chromatography 1243
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Top Related