DOI: 10.2478/s11532-007-0010-3
Research article
CEJC 5(2) 2007 420–454
Behaviour of pure water and water mixture with
benzene or chloroform adsorbed onto ordered
mesoporous silicas
Vladimir M. Gun’ko1∗ , Vladimir V. Turov1, Alexander V. Turov2,
Vladimir I. Zarko1, Vasiliy I. Gerda2† Victor V. Yanishpolskii1,
Inna S. Berezovska1, Valentin A. Tertykh1
1
Institute of Surface Chemistry,
03164 Kiev, Ukraine
2
Taras Shevchenko University,
Kiev 01030, Ukraine
Received 13 September 2006; accepted 15 November 2006
Abstract: Structural characteristics of synthesized ordered mesoporous silicas MCM-41, MCM-48 and
SBA-15 were studied using XRD, nitrogen adsorption and FTIR methods. Pure water and mixtures with
water/benzene and water/chloroform-d adsorbed onto silicas were studied by 1 H NMR spectroscopy with
layer-by-layer freezing-out of bulk and interfacial liquids. Concentrated aqueous suspensions of MCM-48
and SBA-15 were studied by thermally stimulated depolarization current (TSDC) method. Benzene and
chloroform−d can displace a portion of water to broad pores from the pore walls and from narrower pores,
especially in the case of a large excess of an organic solvent. This process is accompanied by diminution
of both interaction energy of water with an adsorbent surface and freezing temperature depression of
adsorbed water. The effect of nonpolar benzene on pore water is much stronger than that of weakly polar
chloroform-d. Modifications of the Gibbs-Thomson relation to describe the freezing point depression of
mixtures of immiscible liquids confined in pores allow us to determine distribution functions of sizes of
structures with unfrozen pore water and benzene.
c Versita Warsaw and Springer-Verlag Berlin Heidelberg. All rights reserved.
Keywords: MCM-41, MCM-48, SBA-15, XRD, FTIR, nitrogen adsorption, structural characteristics,
pore liquid, 1 H NMR, TSDC, cryoporometry, relaxometry
∗
†
E-mail: [email protected]
Former address: Pisarzhevskii Institute of Physical Chemistry, 31 Prospect Nauki, Kiev, Ukraine
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
1
421
Introduction
Mesoporous silicas with ordered structure of pores (M41S, SBA, FSM, etc.) are interesting as adsorbents, catalyst carriers, etc. [1–7]. Morphological, textural and adsorptive
characteristics of these materials have been studied and well documented [1–11]. Typically ordered mesoporous silicas have a narrow size distribution of pores (PSD) of a
cylindrical shape. The structural characteristics of these silicas depend on the molecular
size and the structure of a template, composition of a reaction solution (e.g. the presence
of alcohol) and reaction conditions (temperature, time, pH, etc.). Their adsorption properties depend on the PSD and the structure of pore walls. Surface silanols (2-4 μmol/m2
for MCM-41 [11], 4-5OH/nm2 for different silicas) serve as the main adsorption sites
for polar adsorbates [2–6]. The oxygen atoms in the siloxane bonds ≡Si-O-Si≡ possess
weaker electron-donor properties than that in silanols; therefore, these bonds demonstrate
much weaker hydrophilic properties than silanols. Dispersive forces are predominantly
responsible for the interaction of nonpolar and weakly polar adsorbates with the siloxane
bonds, as well as with a whole silica surface. These structural features of silicas allow
polar and nonpolar adsorbates to interact with different surface sites due to the hydrogen
and/or dispersion binding. Therefore, the structure of the interfacial layers of mixtures
of different immiscible or poorly miscible polar and nonpolar liquids or fluids in pores of
silicas can have heterogeneous structure. Additionally, competitive adsorption of different adsorbates can lead to the displacement of one adsorbate by another from pores and
pore walls. The behaviour of water/organic mixtures in the form of solutions, emulsions
or separated phases, which is of importance from practical point of view, can be strongly
different at a plane surface, in narrow pores and in the bulk liquids. This difference is due
to lower solubility of organics in the interfacial water [12] and lower molecular mobility in
the interfacial layers [13] than in the bulk. Numerous papers with detailed descriptions of
the synthesis and the characterization of ordered mesoporous silicas have been published.
Therefore, these aspects are shown in the present paper only with respect to the effects
of confined space and the surface chemistry on pore water and its mixtures with other
liquids (benzene and chloroform).
Previously the behaviour of pore water in mesoporous silicas was studied by 1 H NMR
spectroscopy (giving temperature dependence of transverse relaxation time and chemical shift of the proton resonance), XRD (freezing/melting behaviour and ice crystalline
structure), thermally stimulated depolarization current, TSDC, (relaxation phenomena
for pore water depending on temperature and pore structure), FTIR and Raman spectroscopies (characteristics of adsorption sites and water binding), and other methods
[1–11, 13–16]. For instance, Morishige and Iwasaki [16] studied the freezing/melting behaviour of water in partially filled pores of SBA-15 (average pore radius R = 3.9 nm)
by XRD measurements depending on temperature and pore filling. They showed that
freezing temperature increased with increasing pore filling. The freezing of the pore water
results in formation of ice microcrystals with almost the same structure and size, irrespective of the different states of the pore water [16]. There are several methods used
Unauthenticated
Download Date | 6/18/17 9:10 AM
422
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
for estimation of the pore size distributions based on a specific behaviour of pore liquids.
For instance, the cryoporometry [17–19] is based on the NMR measurements and the
Gibbs-Thomson relation for melting point depression of liquids confined in pores. The
relaxometry uses the enhanced relaxation of molecules at a pore surface to calculate the
pore diameter [18] and normally assumes rapid exchange between molecules at the surface
and in the pores. The thermoporometry is used for characterizing pore structure from the
melting or freezing point depression of a liquid confined in pores, by reason of the added
contribution of surface curvature to the phase-transition free energy [20]. The effects of
organic solvents on the behaviour of pore water adsorbed onto carbon adsorbents [21–23]
and mesoporous silica gels [24] were studied by the 1 H NMR method. It was shown that
this behaviour depends on molecular size, polarity, polarizability, electron-donor properties, concentration of organic solvents, as well as on the solubility of water in organics and
vice versa. Nonpolar solvents (immiscible with water) displace pore water from narrow
pores of carbon adsorbents more effectively than weakly polar or polar ones (soluble in
bulk water). This is due to the different nature of forces giving the main contribution to
interaction energy of polar and nonpolar molecules with polar, weakly polar and nonpolar
surface sites and adsorbed water molecules forming clusters and droplets near polar sites
in pores or at the entrances into the pores [21–23]. In the case of silica gel Si-40 with
narrow mesopores (close in size to that of MCM-48 studied here) the effects of organics
on pore water differ from that observed for activated carbons because of the difference
in the structural and adsorption properties of the adsorbents [13, 24]. In general, the
structurization of adsorbed mixtures with water/organics occurs in the confined space
of pores to minimize Gibbs free energy of the system. Clearly, the structural characteristics of these mixtures depend on the structure and the chemistry of both adsorbates
and adsorbents [21–24]. Additionally, the order of the loading of liquids (as well as other
treatments of pore liquid mixtures) can affect this structurization because the molecular
mobility of liquids in narrow pores is much lower than that in the bulk. This restricts or
slows down the rearrangement of the structure of the pore liquid mixtures.
The interfacial behaviour of water/organics, which can be very important for the practical application of porous materials as adsorbents, catalysts, etc., can be investigated
by using both NMR cryoporometry and relaxometry techniques and TSDC for a deeper
understanding of the structurization of liquids confined in pores. Therefore the aim of
the present work is to study (i) the structural characteristics of synthesized ordered mesoporous silicas MCM-41, MCM-48 and SBA-15 by XRD, adsorption and FTIR methods
to explain the difference in their adsorption characteristics with respect to polar (water),
weakly polar (chloroform) and nonpolar (benzene) adsorbates; (ii) the relationships between these characteristics and the behaviour of adsorbed pure water and its mixtures
with benzene or chloroform-d over a wide temperature range and on variation of the
mixture composition by using the 1 H NMR spectroscopy (as both cryoporometry and
relaxometry) with layer-by-layer freezing-out of bulk and pore liquids at 170-283 K; and
(iii) concentrated aqueous suspensions of MCM-48 and SBA-15 by the TSDC method at
90-265 K.
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
2
423
Experimental
2.1 Materials
Ordered mesoporous silica MCM-41 with spherical particles was synthesized according to
a technique described in detail elsewhere [25] using cetyltrimethyl ammonium bromide
(CTAB) as a template and tetraethoxysilane (TEOS) as a precursor of silica at a molar
ratio of components TEOS : CTAB : NH3 : H2 O : C2 H5 OH = 1 : 0.3 : 11 : 144 : 58.
CTAB (2.5 g) was solved in distilled water (50 ml) and then 20 ml of 25% ammonium
solution and 80 ml of ethanol (96% ethanol + 4% water) were added. The solution was
stirred by a magnetic stirrer (250 rpm) for 15 min. Then 5 ml of TEOS was added.
The system was stirred at room temperature for 2 h. Sample 1 was filtered just after
this stirring but sample 2 was filtered in 5 h after the stirring. Sediments were washed
in 100 ml of water and 100 ml of ethanol, then dried at 363 K for 5 h and treated at
823 K for 5 h to remove the template. This synthesis technique differs slightly from the
standard one used previously [2–6]; however, addition of alcohol to the reaction mixture
and the corresponding effects on the MCM-41 structure were studied previously [28].
The structural characteristics of the synthesized two samples of MCM-41 are shown in
Table 1. The pore size of MCM-41, as well as SBA-15 and MCM-48, was calculated on
the basis of the XRD and adsorption data as described elsewhere [26, 27].
Synthesis of MCM-48 [29] was carried out using a reaction mixture with CTAB
(Aldrich), TEOS (Alfa Aesar, 99.9%), NaOH (‘Khimreactiv’) at a molar ratio TEOS :
CTAB : NaOH : H2 O = 1 : 0.5 : 0.54 : 110. TEOS (34 ml) was slowly added to aqueous
solution (300 ml) of NaOH (3.3 g) and CTAB (27.8 g) stirred by a magnetic stirrer for
2 h at room temperature. Then the homogeneous mixture was hydrothermally treated
at 383 K for 72 h. The product was filtrated, washed using distilled water and dried at
room temperature. The organic compounds were removed by calcination of the samples
in air flow (9 L/h) initially heated at a rate 24 K/h and then at 823 K for 6 h.
SBA-15 [30] was synthesized by using a nonionic oligomeric alkyl-ethylene oxide surfactant (Pluronic P123 (Aldrich)) as a structure-directing agent dissolved in water (120
ml) and HCl (20 ml, 37 %) at 313 K. TEOS (10 ml) was added to this solution and stirred
at 313 K for 20 h. Then the reaction mixture was hydrothermally treated at 373 K for
24 h without stirring. The product was filtered, washed with distilled water and dried at
373 K. The organic compounds were removed on calcination as in the case of MCM-48
preparation.
Benzene (melting point Tm,∞,b = 278 K) and chloroform-d (Tm,∞,c = 209 K) used
on the NMR measurements are poorly soluble in water and visa versa because benzene
is nonpolar (dielectric constant ε = 2.27) and chloroform is weakly polar (ε = 4.71).
Chloroform-d was used to avoid contribution of protons of an organic solvent to the 1 H
NMR signal. Benzene was used to measure the relationship between the amounts of
unfrozen adsorbates dependent on temperature and PSD of silica.
Unauthenticated
Download Date | 6/18/17 9:10 AM
1104
809
1324
974
MCM-48
SBA-15
1.333
1.174
0.702
0.604
Vp
(cm3 /g)
165
340
26
13
Smic
(m2 /g)
807
975
1170
956
Smes
(m2 /g)
2
7
1
1
Smac
(m2 /g)
λ
is
2 sin Θm
k2
l2 )0.5
2d100
√
3
0.045
0.130
0.010
0.018
Vmac
(cm3 /g)
0.18
0.14
0.43
0.38
a
Δwcyl
15.3
12.3
29.9
27.5
χa
(%)
0.02
0.07
0.19
0.17
b
Δwcyl
2.0
6.5
16.0
14.5
χb
(%)
ρ0 = 2.2 g/cm3 is the true density of amorphous silicas;
1.231
0.956
0.683
0.583
Vmes
(cm3 /g)
is the distance between pore centers for MCM-41 and SBA-15 [26];
+
+
for MCM-48.
a0 = d211
a s
2 b s
2
N2 = 0.162 nm ;
N2 = 0.135 nm ; *Parameters for MCM-48 were calculated from the d211 spacing as described elsewhere [27].
(h2
ρ0 Vp
;
1+ρ0 Vp
0.058
0.094
0.009
0.003
Vmic
(cm3 /g)
the spacing value and Θm is the angle corresponding to (hkl) reflection peak;
twall = a0 − dXRD is the thickness of pore walls and a0 =
dhkl =
√
dXRD = 1.213d100 εis the pore diameter for MCM-41 and SBA-15 [26], ε =
788
997
969
1196
b
SBET
(m2 /g)
MCM-41(2)
a
SBET
(m2 /g)
MCM-41(1)
Sample
6.10
3.92
2.34
2.45
dN2
(nm)
11.43
3.25*
3.09
3.11
dXRD
(nm)
10.91
3.53
3.27
3.40
dhkl
(nm)
9.55
3.76
3.14
3.36
dP SD
(nm)
1.17
0.69*
0.69
0.82
twall
(nm)
Table 1 Structural characteristics of MCM-41 (two samples), MCM-48 and SBA-15 calculated from the nitrogen adsorption isotherms.
Diameter mesopores dN 2 = 4Vmes /Smes ; dP SD is the diameter of pores corresponding to the main maximum of the IPSD.
424
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
425
Small angle X-ray diffraction (XRD) patterns (Fig. 1a) were recorded over 2θ = 0.5–
10 range using DRON4-07 and DRON-3M diffractometers (LOMO and Bourevestnik,
Inc., St. Petersburg) with Cu Kα (λ = 0.15418 nm) radiation and a Ni filter.
An initial MCM-41 sample and a sample treated at 473 K for 2 h were stirred with dry
KBr (1:20) to record the FTIR spectra using a FTIR ThermoNicolet spectrophotometer
over the 400-4000 cm−1 range. The FTIR spectra of MCM-48 and SBA stirred with
dry KBr powder (1:100) were recorded using a FTIR Nicolet Nexus spectrophotometer
without thermal pre-treatment of silicas.
o
2.2
1
H NMR
The 1 H NMR spectroscopy with layer-by-layer freezing-out of bulk and interfacial liquids
was used to determine the characteristics of interaction of water with silica surface affected
by benzene or chloroform-d. This technique was described in detail elsewhere [13, 31, 32].
The 1 H NMR spectra were recorded using a Varian 400 Mercury spectrometer and a
Bruker WP-100 SY spectrometer of high resolution with the probing 90◦ or 40o pulses
at duration of 4-5 μs. On the measurements of the 1 H NMR spectra temperature was
controlled using a Bruker VT-1000 device. Relative mean errors were ±10% for 1 H NMR
signal intensity and ±1 K for temperature. Measurements of amounts of unfrozen water
or water/benzene were carried out upon heating of samples preliminarily cooled to 170190 K for the prevention of supercooling. This procedure provides more equilibrium state
of pore water than that on cooling of the system (some portion of pore water can be
in the supercooling state on cooling). The 1 H NMR spectra recorded here include the
signals only of nonfreezable mobile water and benzene molecules. The signals of water
molecules from ice and frozen benzene (as well as from silanols) do not contribute the 1 H
NMR spectra because of features of the measurement technique and the short duration
(∼10−6 s) of transverse relaxation of protons in immobile structures.
The transverse relaxation times were measured with Carr-Purcell-Meiboom-Gill (CPMG)
pulse sequence: 90x-(τ -180y − τ -echo)n with 90o pulse of 3 μs and total echo time of 500
μs. A total of eight scans were used for each measurement. The relaxation time between
scans was 0.5 s. Equilibration time of each temperature was 5 min. To calculate a distribution function of transverse relaxation time f (T 2), CPMG echo decay envelopes (I(t))
were used as the left term of integral equation
T
2max
I(t) = A
exp(−
T 2min
t
)f (T 2)dT 2,
T2
(1)
where t is the time, T 2min and T 2max are the minimal (10−6 s) and maximal (1 s) T 2
values on integration respectively, and A is a constant. Solution of Eq. (1) is well known
ill-posed problem due to the impact of noise on measured data, which does not allow
one to utilize exact inversion formulas or iterative algorithms. Therefore, Eq. (1) can
be solved using a regularization procedure based on the CONTIN algorithm [33] under
nonnegativity condition (f (T 2) ≥ 0 at any T 2) and an unfixed value of the regularization
Unauthenticated
Download Date | 6/18/17 9:10 AM
426
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 1 (a) XRD patterns of ordered mesoporous silicas; (b) nitrogen adsorption isotherms;
(c) αS plots (silica gel Si-1000 as a reference) corresponding to the nitrogen adsorption
isotherms; and incremental pore size distributions for (d) SBA-15 with CONTIN and
CONTIN/MEM-j and the model of cylindrical pores (d); and (e, f) MCM-41 ((e) sample
2 and (f) samples 1 and 2), MCM-48 and SBA-15 with respect of the pore volume with the
model of (e) cylindrical pores (CONTIN/MEM-0 (solid line) and CONTIN (dashed line))
and (f) cylindrical pores plus voids between spherical particles (CONTIN) calculated with
self-consisting regularization procedure.
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
427
parameter (α) determined on the basis of the F-test and confidence regions using the
parsimony principle. It should be noted that the solution of Eq. (1) is insensitive to the
A value which can affect only relative intensity of f (T 2). A similar procedure applied
the NMR data to calculate f (T 2) was described previously [34, 35].
The condition of freezing of pore water (or benzene) corresponds to the equality of the
Gibbs free energies of molecules of water and ice. Lowering the temperature of freezing of
structured pore water (Tf < 273.15 K) is defined by diminution of its Gibbs free energy
caused by strong inter-molecular interactions (ΔG = G − G0 < 0, where G0 is the free
energy of ice at 273.15 K) that disrupt the hydrogen bond network characteristic of the
bulk water. The temperature dependence of the Gibbs free energy of ice [36] can be
approximated as follows:
ΔGice = 0.0295 − 0.0413ΔT + 6.64369 × 10−5(ΔT )2 + 2.27708 × 10−8(ΔT )3 (kJ/mol) (2)
where ΔT = 273.16 - T at T ≤ 273.15 K. The fact that pore water may remain unfrozen
at T < 273 K suggests that its Gibbs free energy remains lower than that of bulk water
or bulk ice Giw < Gice because of interaction with a solid surface. Further lowering of
temperature reduces this inequality until the point Tc , at which a certain amount of bound
water becomes frozen. At Tc the relationship ΔGw = ΔGice pertains, where ΔGiw =
Giw (T )−G0w (G0w denotes the Gibbs free energy of unperturbed bulk water at 273.15 K and
superscript i stands for interfaces).It is assumed that neither Gice , nor ΔGice are influenced
by the solid surface. The area under the ΔG(Cuw ) curve (temperature dependences
ΔG(T ) and amounts of unfrozen pore water Cuw (T ) can be simply transformed into the
relationship ΔG(Cuw )) determines overall changes in the Gibbs free energy of interfacial
water
max
Cuw
γS = A
ΔGdCuw ,
(3)
0
max
is the total amount of unfrozen water at T → 273 K, and A is a constant
where Cuw
dependent on the type of units used in this equation. Weakly bound water corresponding
to a portion of unfrozen water with the free energy slightly lowered by inter-molecular
interactions with a solid surface freezes at temperature close to 273 K. Strongly bound
water can be unfrozen even on significant cooling of the system and corresponds to the
maximum disturbed water layer at the interfaces. The thickness of the layers of each
type of water or water + benzene (Cus and Cuw denoting amounts of strongly and weakly
bound liquids) and the maximum values for lowering of the Gibbs free energy of these
liquids (ΔGs and ΔGw ) can be estimated from a linear extrapolation of appropriate linear
portions of the ΔG(Cu ) graphs to the corresponding axis. Notice that there are two types
of water (parameter subscript ‘w’) and benzene (subscript ‘b’) or chloroform-d amounts:
(i) unfrozen fraction (Cu = Cuw + Cub ) and (ii) total amounts of water (h), benzene
(b) and chloroform-d (c) in ml per gram of dry silica. Derivative dC u /dT (calculated
using spline approximation of Cu (T )) as a function of temperature can be used to show
characteristic changes in the amounts of structured unfrozen water and water/benzene
Unauthenticated
Download Date | 6/18/17 9:10 AM
428
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
dependent on temperature. A portion of unfrozen water or water/benzene corresponding
to ΔG > -0.5 kJ/mol may be considered as a liquid weakly bound in pores.
Water or benzene can be frozen in narrower pores at lower temperatures that can
be described by the Gibbs-Thomson (GT) relation for freezing point depression for pore
liquids [17, 19, 37]
ΔTm = Tm,∞ − Tm (R) = −
k
2σsl Tm,∞
= ,
ΔHf ρR
R
(4)
where Tm (R) is the melting temperature of ice or solid benzene in pores of radius R, Tm,∞
is the bulk melting temperature, ρ is the density of the solid, σ sl is the energy of solidliquid interaction, ΔHf is the bulk enthalpy of fusion, and k is a constant. According to
Aksnes and Kimtys [19], kb = 44 K nm for benzene freezing in porous silicas. For pure
water adsorbed in pores of silica kw = 67 K nm. Eq. (4) can be utilized to calculate the
incremental pore size distribution (IPSD) on the basis of the temperature dependence
Cu (T ) estimated from the 1 H NMR spectra recorded at T < Tm,∞ for individually adsorbed liquids. This approach was employed for different materials and gave the IPSDs
close to those calculated on the basis of nitrogen adsorption/desorption isotherms [13].
In the case of a mixture of immiscible liquids (e.g., water and benzene or chloroform-d),
Eq. (4) can be used for both liquids with consideration for their amounts in a mixture
and their temperature-dependent signal intensity (obtained on deconvolution of the total
signal)
Ai
dVu,i
dCu,i
= (T − Tm,∞,i )2
,
(5)
dR
ki
dT
where i denotes an adsorbate number, Cu,i(T ) is the integral intensity of a δH band for
the i-th adsorbate as a function of temperature, and Ai is a weight constant dependent on
the molecular volume (vi ), the number of protons (ni ) in a molecule of the i-th adsorbate
and the used units. Notice that nvww ≈ nvbb because of vb /vw ≈ 2.83 and nb /nw = 3. In
the case of the water/benzene mixtures the σ sl,w value determined for the water/silica
system was multiplied by ΔGs,h,b/ΔGs,h (where ΔGs,h and ΔGs,h,b are changes in the free
energy of strongly bound water without and with the presence of benzene respectively)
for consideration for the effect of benzene. Clearly this effect leads to a certain diminution
of the kw value varied between 67 and 40 K nm for different water/benzene mixtures. For
simplicity, the kb value was assumed as a constant (44 K nm) for all the studied systems
because benzene is being in contact with the pore walls since it can displace water from
pores and the pore walls. Eqs. (4) and (5) allow us to calculate the distribution functions
of the sizes of pores filled by an individual adsorbate (e.g. water) and an incremental size
distribution (ISD) of structures in a liquid mixture (e.g. water/benzene) filling pores.
In the last case the pore size can correspond to a sum of the sizes (thickness) of the
pore liquid layers if liquids simultaneously locate in the same pores. In the case of a
mixture of water with chloroform-d, the latter does not contribute the 1 H NMR spectra.
However, chloroform-d can fill a portion of pores and affect the occupation of them by
water, i.e. change the Cuw (T ) shape. Eqs. (4) and (5) can be transformed into integral
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
429
equation (IGT), replacing dV /dR by f (R), converting dC /dT to dC /dR with Eq. (5)
and integrating by R,
Rmax Cu,i(T ) = A
Rmin
ki
(Tm,∞,i − Tm,i (R))R
2
fi (R)dR
(6)
where Rmax and Rmin are the maximal and minimal pore radii (or sizes of unfrozen
liquid structures) respectively, i is the index corresponding to i-th pore liquid, and A is
a normalization factor.
In the case of the use of the relaxometry method for estimation of the sizes of the water
and benzene structures (layers, clusters, etc.) filling pores, Eq. (1) can be transformed
for consideration for the dependence of the CPMG echo decay envelopes (i.e. transverse
relaxation time) on the pore size
Rmax
Ii (t) = B
exp(−
Rmin
t
(Tm,∞,i − Tm,i (R))
)
fi (R)dR,
T 2i,m
ki
(7)
where B is a normalization factor, and ki is a constant analogous to that in Eqs. (4) and
(5). Eqs. (6) and (7) can be solved using the regularization procedure [33] similar to that
applied to Eq. (1).
Pores of MCM-41 (sample 2 was chosen for the NMR measurements because it has
larger Vp and SBET values than sample 1) were partially filled by water at hydration h
= 0.186 or 0.317 ml of water per gram of dry MCM-41, then a portion of benzene (b =
0.196 ml per gram of dry MCM-41) or chloroform-d (c = 5.0 ml/g) was added, then a
portion of water was added that gives h = 0.37 ml/g, and then a portion of benzene was
added that gives b= 0.4 ml/g. Sample at h = 0.37 ml/g and b = 5.8 ml/g was prepared
on the first addition of water (h = 0.37 ml/g) and then benzene (b = 5.8 ml/g). In the
case of MCM-48 (h = 0.2 ml/g and b = 0.2 ml/g) and SBA-15 (h = 0.16 ml/g and b =
0.13 ml/g) water was adsorbed before benzene.
2.3 Nitrogen adsorption
Low-temperature (77.4 K) adsorption/desorption isotherms of nitrogen (Fig. 1b), which
is type IV in the IUPAC classification [38] (however, for MCM-41 it tends to type I), were
measured using a Micromeritics ASAP 2405N adsorption analyzer. The specific surface
area SBET (Table 1) was calculated according to the standard BET method [38] but
at different pressure ranges p/p0 = 0.05-0.17 (MCM-41), 0.05-0.2 (MCM-48), and 0.050.23 (SBA-15) to avoid overestimation of the SBET values because of the beginning of
the formation of the second layer with adsorbed nitrogen in narrow mesopores at lower
pressures [39]. The dependence of this effect on the used pressure range results in different
dependences of the SBET values on the high boundary value p/p0,max for the studied silicas
(Fig. 2). The pore volume Vp was estimated at p/p0 ≈ 0.98-0.99 (where p and p0 denote
the equilibrium and saturation pressures of nitrogen respectively) converting the adsorbed
Unauthenticated
Download Date | 6/18/17 9:10 AM
430
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
amount a0.98 (in cm3 STP of gaseous nitrogen per gram of the adsorbent) to the liquid
adsorbate Vp ≈ 0.0015468a0.98.
Fig. 2 SBET as a function of the maximal p/p0 value of the pressure range (0.05 –
p/p0,max ) used on calculation of the SBET values for MCM-41 (sample 2), MCM-48 and
SBA-15.
Pore size distributions (PSD, fV (R) and fS (R) with respect to the pore volume and
the specific surface area respectively) were calculated using an overall isotherm equation
based on the equation proposed by Nguyen and Do for carbon adsorbents with slitlike
pores [40] and modified for cylindrical pores and voids between spherical particles [41, 42].
This equation was solved by means of a regularization procedure based on the CONTIN
algorithm [33] as described elsewhere [41]. The differential PSDs fV (R) and fS (R) were
converted to incremental PSDs (IPSD) (Figs. 1e and 1f). The fV (R) and fS (R) functions
∗
were used to calculate contributions of micropores (Smic
, Vmic ) at R < 1 nm, mesopores
∗
∗
(Smes , Vmes ) at 1 < R < 25 nm, and macropores (Smac , Vmac ) at 25 < R < 100 nm to
the specific surface area and the total porosity. Additionally, fS (R) was used to estimate
deviation of the pore shape from the model of cylindrical pores or a complex model
with cylindrical pores and voids between spherical particles (used for estimation of the
textural porosity of the studied samples) using a self-consisting (to better fit the nitrogen
adsorption isotherms) regularization [41, 42].
Δw =
SBET
R
max
Rmin
fS (R)dR
−1 =
SBET
R
max
Rmin
w
(f (R)
R V
−
−1
(8)
V (R)
)dR
R
where Rmax and Rmin are the maximal and minimal pore radii respectively, V (R) is the
∗
∗
, Smes
and
volume of pores at the radius R, and w = 2 for cylindrical pores. The Smic
∗
∗
Smac values were corrected by multiplication by (Δwcyl + 1) that gives S (Δwcyl + 1) =
Ssum = Smic + Smes + Smac = SBET (Table 1).
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
431
To better describe the porosity of the studied samples an additional regularizer was
derived using maximum entropy principle [43] applied to the distribution function f (R)
which can be written as N-dimension vector (N is the number of the grid points for f )
V AR + α2 (1 −
S(p(f))
) → min,
Smax
(9)
where VAR is the regularizer, α is the regularization parameter,
p0 (f) = f,
p1i (f) = fi+1 − fi + (fmax − fmin ); i = 1, . . . N − 1,
p2i (f¯) = fi+1 − 2fi + fi−1 + 2(fmax − fmin ), i = 2, . . . N − 1,
S(f) = −
N
sk ln(sk ),
k=1
sk =
fk
N
k=1
,
fk
Smax = − ln(
1
).
N
vector corresponds to the maximum entropy principle of the j-order [43].
The pj (f)
This procedure was used to modify the CONTIN algorithm [33] (CONTIN/MEM-j where
j denotes the order of pj (f)). A self-consisting regularization procedure with an unfixed
regularization parameter (to better fit the isotherms) was used on CONTIN/MEM-j
calculations. The COTIN/MEM-0 procedure was also applied to IGT equation (6).
To compute the αS plots [38] (Fig. 1c), silica gel Si-1000 was used as a standard
adsorbent [44].
2.4 Thermally stimulated depolarisation current (TSDC)
The tablets (diameter 30 mm, thickness ∼1 mm) with the frozen aqueous suspension
of silicas (16.7 wt%, i.e. h = 5 ml/g) were polarised by the electrostatic field at the
intensity Fp = 200-300 kV/m at 260 K then cooled to 90 K with the field still applied
and heated without the field at a heating rate β = 0.05 K/s. The current evolving due
to sample depolarisation [37, 45] was recorded by an electrometer over the 10−15 –10−7 A
range. Relative mean errors for measured TSD current were δ I = ±5%, δ T = ±2 K for
temperature, δ β = ±5% for the temperature change rate. Modified Eqs. (4) and (6) with
k as a linear function of temperature (k(T ) = 40 + 56 (T − 90) K nm at T between 90 and
270 K) obtained on the basis of the calibration curves for silica gels Si-40 and Si-60 was
used for estimations of the IPSDs on the basis of the TSDC data [13, 37].
Unauthenticated
Download Date | 6/18/17 9:10 AM
432
3
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Results and discussion
3.1 Structural characteristics of silicas
Figure 1a shows the XRD patterns for MCM-41 (observed reflections suggest that it can
be assigned to 2D hexagonal structure), SBA-15 (hexagonal structure) and MCM-48 (3D
cubic structure) [26, 27]. For hexagonal structure the lattice constant a0 = 2d√100
corre3
sponding to the distance between pore centers, the pore size dXRD = 1.213d100
ρ0 Vp
1+ρ0 Vp
(ρ0 = 2.2 g/cm3 is the true density of amorphous silica) and the pore wall thickness
twall = a0 – dXRD can be estimated from the d100 spacing value [26]. The dXRD and twall
values for MCM-48 (Table 1) were calculated using the d211 spacing value as described
elsewhere [27].
The average pore wall thickness is between two (0.62 nm) and four (1.24 nm) silica
layers for the studied samples (Table 1, twall ). The pore sizes estimated from the XRD
(Table 1, dXRD ) and adsorption data (dN 2 and dP SD ) are close (with one exception of SBA15). A significant difference between the dN 2 and dXRD values for SBA-15 is due to the
contribution of narrow mesopores at 1 < R < 2 nm (Figs. 1e and 1f) (these pores can be
attributed to inner-wall pores), which are out of the main mesopore peak, on calculation
of dN 2 . The dP SD value for the main mesopore peak for SBA-15 is closer to dXRD than
dN 2 . Calculation of the dN 2 value only for this peak gives 9.92 nm which is closer to the
dXRD value. Consequently, on the calculation of the size of the main mesopores from the
adsorption data, not only external surface and the corresponding pore volume but also
the surface and the volume of narrow pores should be taken into account. Notice that a
certain difference between the pore sizes determined from the XRD and adsorption data
can be caused by neglecting or accounting of the size of surface atoms.
The shapes of the nitrogen adsorption/desorption isotherms and hysteresis loops
(Fig. 1b) reveal that MCM-48 and SBA-15 have the porosity of two types [46]: (i)
framework-confined porosity caused by relatively uniform channels of the template framework (tested by the presence of a significant adsorption step at p/p0 between 0.1 and 0.5),
and (ii) textural porosity caused by intraaggregates voids and spaces formed by interparticle contacts (tested by the presence of the hysteresis loop at p/p0 > 0.5). The presence
of the textural porosity is also confirmed by the IPSD for MCM-48 and SBA-15 characterized by certain contribution of large mesopores and macropores (Figs. 1d, 1e and 1f).
The porosity of the second type for MCM-41 is very small because the isotherm does not
include the hysteresis loop and the adsorption step at p/p0 > 0.5 (Fig. 1b).
The Δwcyl values (Table 1) suggest that pore wall surfaces are not smooth because
the specific surface area (S ∗ ) calculated using the model of cylindrical pores is smaller
than the SBET value, especially in the case of MCM-41 samples. Notice that the use of
∗
the Smes
value which was not multiplied by (Δwcyl + 1) on calculation of the diameter
of the mesopores gives 3.33 nm for MCM-41 (sample 2). This value is close to dXRD and
dP SD . There are, at least, two reasons resulting in S ∗ < SBET : (i) certain roughness of the
pore walls which cannot be described in the framework of the model of ideal cylindrical
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
433
pores; and (ii) overestimation of the SBET value because of the nitrogen adsorption in
the second and next layers (i.e. pore filling) at p/p0 < 0.3 and d < 4 nm and because of
overestimation of the cross-sectional area for a nitrogen molecule (sN 2 = 0.162 nm2 ) for
hydroxylated silica surfaces. For instance, this area of 0.135 nm2 gives better agreement
with the SBET values obtained from the Ar adsorption data [39]. Diminution of the
p/p0,max values on the calculations of the SBET values allows us to reduce the second
effect. The use of reduced sN 2 = 0.135 nm2 [39] gives much smaller SBET and Δwcyl
values (Table 1); however, inequality S ∗ < SBET remains. Therefore, one can assume
that this residual inequality S ∗ < SBET is due to a nonideal shape of the pores, e.g.
the roughness of the pore walls. Consequently, consideration for the roughness of the
pore walls, which can be estimated as χ = 100(SBET – S ∗ )/SBET = 100Δwcyl /(Δwcyl +
1) (%) (Table 1), is important on estimation of the pore size from the adsorption data.
For instance, the χ value is maximal for MCM-41 samples; therefore underestimation
of the dN 2 values is maximal for these silicas. In other words, the pore walls of MCM41 samples are rougher than that of MCM-48 or SBA-15. However, according to the
IPSDs obtained with the model of cylindrical pores, contribution of micropores (which
can be also linked to the roughness of the pore walls) for MCM-41 is lower than that
for other silicas. This roughness can be modelled by tiny spherical-like hills and the
corresponding shallow hollows on the surface of cylindrical mesopores; i.e. the pore walls
are composed of spherical globules (∼ 1 nm in diameter). Additionally, the textural
porosity can be modelled as voids between spherical globules of larger sizes (10-100 nm
or more). This allows us to use a complex model of pores of a cylindrical shape and voids
between spherical particles with self-consisting regularization procedure used to obtain
independent f (R) functions for both pore shapes. The complex model of pores adds
intensity of the IPSD of micropores for all samples (Fig. 1f); however, the main peaks
of mesopores remain nearly the same. This model gives Δw values that are much lower
(e.g. 0.01 for MCM-48 and 0.12 for sample 2 of MCM-41) than those obtained using the
model of cylindrical pores with different sN 2 values.
The studied samples of MCM-41 (especially sample 2) do not have the ideal conventional structure. For instance, the nitrogen adsorption/desorption isotherm does not
have a large adsorption/desorption step (Fig. 1b) characteristic of MCM-41 synthesized
using C16 TAB but without alcohol [2–6]. However, this step is more clearly observed
for sample 1 than for sample 2 of MCM-41 (Fig. 1b). Alcohol in the reaction solution
interacting with the template molecules can affect the micellar/tubular, supramolecular
structure and, therefore, the finished mesostructure of silica; and the roughness of the
pore walls increases (especially on additional aging of sample 2). Notice that similar
XRD patterns and nitrogen adsorption/desorption isotherms were previously observed
for MCM-41 synthesized with alcohol in the reaction solution [28]. The synthesis of
MCM-41 without alcohol gives materials characterized by a typical shape of the nitrogen
adsorption isotherms with a typical step shape [7–10, 39, 47]. The reaction time effect
is well observed on comparison of the structural characteristics of two MCM-41 samples
(Table 1, Fig. 1) synthesized under nearly the same conditions only at different time of
Unauthenticated
Download Date | 6/18/17 9:10 AM
434
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
the filtration stage after slurry stirring.
The IPSDs (Figs. 1d, 1e and 1f) demonstrate the presence of large and narrow pores in
addition to the main narrow mesopores for all the studied silicas. The IPSDs of MCM-48
and SBA-15 shift towards larger pore sizes in comparison with MCM-41. The αS plots
(Fig. 1c) and the IPSDs (Figs. 1d, 1e and 1f) reveal a certain contribution (however,
relatively small) by micropores at R < 1 nm. Notice that the model of cylindrical pores
(Fig. 1e) (with both CONTIN and CONTIN/MEM-0 regularizations) does not give micropores for MCM-41 in contrast to the complex model of pores (Fig. 1f). Calculations of
the IPSDs for SBA-15 with CONTIN/MEM-j at j = 0, 1, and 2 (Fig. 1d) show that the
main differences are observed at R < 0.7 nm and R > 30 nm; consequently calculations
for other silicas were carried out with CONTIN/MEM-0 (Fig. 1e). The studied silicas
are characterized by different IPSDs; therefore the behaviour of pore liquids in these adsorbents can depend differently on temperature at T < 273 K according to Eqs. (4)-(7)
[13]. To study the influence of benzene and chloroform-d on the structure of pore water
the measurements were started from the amounts of adsorbed liquids much smaller than
the pore volume of silicas.
3.2 Behaviour of pore liquids
3.2.1 MCM-48 and SBA-15
In the case of partial filling of pores by water and benzene the latter tends to displace a
portion of water from narrow pores to larger ones or from the pore walls to the center of
pores. This is well seen on comparison of the 1 H NMR spectra of adsorbed pure water
and water/benzene mixture (Fig. 3). The 1 H NMR spectra show a stronger decrease
in the signal intensity of the proton resonance of adsorbed water at δH ≈ 5 ppm at T
< 230-240 K with the presence of benzene. Therefore the relationship between the ΔG
and Cuw values (Fig. 4) demonstrates diminution of the Cuw value for the water/benzene
mixture in comparison with pure water at the same ΔG and temperature values. These
effects can be explained by the displacement of water by benzene from the pore walls
or narrow pores into larger ones. These effects can seem unusual since water can form
strong hydrogen bonds with a silica surface in contrast to benzene. However, dispersive
interaction of benzene with siloxane bonds can be characterized much higher energy than
that for water. Additionally, there is a phase boundary between water and benzene with
high Gibbs free energy. To reduce this energy and the area of contact between water and
benzene the letter can displace water into large pores where water forms larger clusters
and droplets than in narrow pores. Typically water tends to form clustered coverage of a
silica surface and a continuous layer is not formed at relative high hydration h < 0.3 g/g
[13, 37]. Therefore such clustered coverage of a silica surface by water leads to a great
area of contact between water clusters and benzene. Therefore, formation of much larger
water structures (in larger pores) with smaller contact area with the benzene phase is
energetically favorable.
The f (T 2) distribution functions (Fig. 5) demonstrate that the proton transverse reUnauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
435
Fig. 3 Chemical shit of proton resonance of (a) water (h = 0.2 ml/g) and (b) water/benzene mixture (h = 0.2 ml/g, b = 0.2 ml/g) adsorbed onto MCM-48; (c) water (h
= 0.16 ml/g) and (d) water/benzene mixture (h = 0.16 ml/g, b = 0.13 ml/g) adsorbed
onto SBA-15.
laxation of pore benzene is more complex (as characterized by bi- or trimodal f (T 2)
distributions) in comparison with pore water (mainly monomodal f (T 2) distributions).
Additionally, there are both faster and slower relaxation processes of benzene in comparison with water. These results can be caused by the difference in the spatial distribution
of benzene and water in pores and the effects of the pore size and the structure of the pore
walls on the T 2 values because of the corresponding changes in the molecular mobility of
adsorbed liquids. This conclusion is confirmed by calculations of the pore water and benzene structures (Figs. 6-8) using Eqs. (4)-(7). The analysis of the f (T 2) and ISD graphs
for pore liquids reveals that these liquids can (i) fill narrow pores (e.g. micropores at R <
1 nm), (ii) adsorb onto the pore walls in the form of adsorbed films in larger pores (R > 1
nm) in which certain free space remains far from the pore walls because of partial filling
of pores, and (iii) form clusters, nanodomains, droplets, films or continuous condensate
[13] in large mesopores (R > 10 nm) or macropores (R > 25 nm). The thickness (size) of
Unauthenticated
Download Date | 6/18/17 9:10 AM
436
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 4 Relationships between amounts of unfrozen water and changes in its Gibbs free
energy for water and water/benzene adsorbed onto MCM-48 and SBA-15.
the layers (clusters) with unfrozen liquids decreases with lowering temperature (Figs. 6
and 7) because of partial (layer-by-layer) freezing-out of adsorbed liquids. Pore filling by
the liquids is larger for MCM-48 than for SBA-15 because of the difference in the pore
volume of silicas (Table 1, Vp ) and in the amounts of the pore liquids. The structures
formed in MCM-48 and SBA-15 have the size close to that of the main mesopores (Fig. 8);
however, the ISD curves are below the IPSDs because of partial filling of pores.
Similar results with respect to various structurization of adsorbed water in pores of
different sizes were obtained at much higher hydration of these samples (h ≈ 5 ml/g) by
using the TSDC method (Fig. 9). The relaxed water structures correspond to mesopore
sizes of MCM-48 but they are slightly smaller than the pore size of SBA-15. The dipolar
and direct current (dc) relaxations of water adsorbed on different mesoporous silicas are
characterized by two typical low-temperature (LT) and high-temperature (HT) bands in
the TSDC thermograms (Fig. 9a). The effect of pore walls is minimal for a fraction of
water located in large pores, e.g., of SBA-15 (Fig. 9c) or silica gel Si-60 [37]. MCM-48
and SBA-15 (with a more ordered pore structure than that of silica gels) have a larger
effect on the adsorbed water than silica gel [37] and the LT peak shifts towards higher
temperatures (Fig. 9a). The dipolar relaxation of water frozen in pores of different sizes
occurs step-by-step if the pore size is larger than the size of the relaxed water structures.
For instance, in the case of SBA-15 with relatively large pores, the size of water clusters
relaxing at certain temperatures is smaller than the size of pores (Fig. 9c) because water
film freezes at the pore walls at lower temperatures (it is in the liquid state under the
applied field for a longer time) than layers far from the pore walls. A similar effect was
observed on the 1 H NMR investigations of water in silica gels [24]. In the case of MCM48 (Fig. 9b) with narrower pores than SBA-15 water can freeze in pores over a narrow
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
437
Fig. 5 Distribution functions of proton transverse relaxation time (T 2) of (a) benzene
and (b) water adsorbed onto MCM-48 as a mixture with C6 H6 and H2 O (h = 0.2 ml/g, b
= 0.2 ml/g); (c) benzene and (d) water adsorbed onto SBA-15 as a mixture at h = 0.16
ml/g and b = 0.13 ml/g.
temperature range (a similar effect was observed for silica gel Si-40 [13, 24]) but at lower
temperatures. A small amount of water (small clusters) can be frozen at the pore walls at
lower temperatures. According to the TSDC data, mobile water molecules (participating
in the dc relaxation on proton transferring) appear at T ≈ 210 K (Fig. 9a) that is
in agreement with the 1 H NMR spectra (Fig. 3) showing that mobile water molecules
appear at T ≥ 213 K.
The freezing/thawing of pore liquids in relative large mesopores of SBA-15, especially
on total filling of pores (on the TSDC measurements), can occur step-by-step. This
can result in very nonuniform frozen/unfrozen structures of liquids mixed in pores. To
diminish these effects the pore liquids can be studied on the adsorption onto MCM-41
possessing narrower pores than MCM-48 and SBA-15 (Fig. 1). For the NMR measurements sample 2 of MCM-41 was used because it has a larger adsorption capacity, narrower
pores and a larger roughness of the pore walls in comparison with other samples (Table 1)
Unauthenticated
Download Date | 6/18/17 9:10 AM
438
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 6 Pore size distributions calculated on the basis of the nitrogen adsorption isotherm
and ISDs based on CPMG echo decay envelopes for (a) benzene and (b) water adsorbed
onto MCM-48 as a mixture with 200 mg C6 H6 and 200 mg H2 O per gram of silica.
that can provide the appearance of unusual weakly associated water [13]. Additionally,
water adsorbed alone in narrower pores at partial filling of them can form smaller clusters
that cause a larger area of boundary contacts between water and subsequently adsorbed
benzene and stronger effects of rearrangement of water clusters to reduce the boundary
contact area with benzene.
3.2.2 MCM-41
The 1 H NMR spectra of the water/benzene mixtures at different h and b values include
signals of water (∼5 ppm) and benzene (∼7 ppm) differently dependent on temperature
(Fig. 10) similarly to those for MCM-48 and SBA-15 (Fig. 3). In the case of the sample
at h = 0.37 ml/g and b = 5.8 ml/g, a low-intensity signal is also observed at δ H =
1.2-1.4 ppm (Fig. 10d) which corresponds to weakly associated water [13]. Notice that
the roughness of the pore walls (providing formation of very small water clusters) can
be one of the necessary conditions for observation of weakly associated water. The main
peak of adsorbed water at 5 ppm is close to that of liquid water characterized by the
average number of the hydrogen bonds per a molecule close to four. Consequently, the
major portion of water in the pores of MCM-41 is strongly associated because it keeps
the hydrogen bond network causing the mentioned magnitude of δ H . If the total volume
of the adsorbates (at h = 0.186 ml/g and b = 0.196 ml/g) is substantially lower than
the pore volume Vp that the relationship of the signal intensities of water and benzene
slightly decrease with lowering temperature to 223 K. This is due to filling of the narrowest
pores that provides the maximal freezing point depression. The intensity of the water
and benzene signals weakly changes at 220-280 K (Fig. 11c, h = 0.186 ml/g and b =
0.196 ml/g). In this case the temperature dependence of the signal is nearly the same
as for pure water adsorbed on MCM-41 in air at h = 0.186 ml/g. Notice that Cu (T ) in
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
439
Fig. 7 Incremental size distributions of liquid structures calculated on the basis of CPMG
echo decay envelopes for (a and b) benzene and (c) water adsorbed onto SBA-15 as a
mixture at h = 0.16 ml/g and b = 0.13 ml/g; IPSDV based on the nitrogen adsorption
isotherm is shown.
Unauthenticated
Download Date | 6/18/17 9:10 AM
440
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 8 IPSDs for (a) MCM-48 and (b) SBA-15 calculated on the basis of the nitrogen adsorption isotherms and the ISDs based on the 1 H NMR spectra of water and
water/benzene calculated using IGT (Eq. (5)) and MEM-0 with Eq. (9).
Figs. 11c and 11d correspond to the sum Cu = Cuw + Cub for unfrozen water (subscript
‘uw’) and benzene (subscript ‘ub’). Fig. 11e shows only the Cuw (T ) graphs calculated
by deconvolution of the 1 H NMR spectra.
If the volume of the adsorbates is greater than the total pore volume (e.g. h = 0.37
ml/g and b = 5.8 ml/g) that a portion of benzene locating out of pores can freeze at T
between Tm,∞,b ≈ 278 K and Tm,∞,w ≈ 273 K because of small freezing point depression
for benzene mixed with water even out of pores since Tm,∞,b > Tm,∞,w . Composition of the
frozen portions of liquids at T < Tm,∞,i depends on the affinity of the adsorbates to the
pore walls, a clustered structure of pre-adsorbed water, order of the loading of liquids into
pores and a period of time before measurements (if the equilibrium state is not reached
before freezing). According to the data shown in Fig. 10e, the relationship of the signal
intensities of water and benzene is practically identical at 243 K independent of the initial
relationships of the amounts of components (h = 0.186 ml/g and b = 0.196 or 0.4 ml/g).
In the case of a small excess of the liquids (h = 0.37 ml/g and b = 0.4 ml/g) and the same
amount of water but a great excess of benzene (h = 0.37 ml/g and b = 5.8 ml/g) the
main signals at ∼5 ppm (shoulder) and ∼7 ppm (peak) have a similar shape; however,
an additional signal of weakly associated water appears at 1.2-1.5 ppm. This suggests
that a portion of water is strongly clustered, i.e. it is distributed in the form of very
small clusters (locating in shallow micropores in rough walls of mesopores) or individual
molecules dissolved in benzene in silica pores. It is possible that both the roughness of the
pore walls of MCM-41 and residual organic functionalities promote appearance of weakly
associated water [13]. However, the chemical shift of the last signal does not practically
change with temperature (Fig. 10d) in contrast to that of water dissolved in benzene and
characterized by decreased signal intensity with lowering temperature. Consequently,
there are some effects of the silica surface (micropores and residual organic groups), i.e.
its structure and nature, on weakly associated water. Thus, residual organic groups at
the silica surface, the roughness of the pore walls of MCM-41 (sample 2) (Table 1, χ), and
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
441
Fig. 9 TSDC thermograms for aqueous suspensions (h = 5 ml/g) of silicas (Fp = 200300 kV/m); and incremental pore size distributions calculated on the basis of nitrogen
desorption isotherms for dry powders and the size distributions of water clusters relaxing
on the TSDC measurements of aqueous suspensions (h = 5 ml/g) for (b) MCM-48 (using
modified GT and IGT equations), and (c) SBA-15 (modified IGT).
Unauthenticated
Download Date | 6/18/17 9:10 AM
442
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 10 1 H NMR spectra of water/benzene mixture adsorbed on MCM-41 at different
temperatures and different amounts of water and benzene (b): h = (a, b, e) 0.186, (c, d,
e) 0.37 ml/g and b = (a, e) 0.196, (b, c, e) 0.4 and (d, e) 5.8 ml/g (sample 2).
significant excess of the organic solvent can facilitate the appearance of weakly associated
water at δH = 1.2-1.5 ppm.
In all the cases of the water/benzene mixtures, a portion of water is displaced from
the pore walls and narrow pores into larger ones (Figs. 11- 13) to reduce the contact
area between these liquids. Adsorbed pure water forms several types of structures (e.g.
clusters, thin surface films and structures more completely filling certain narrow pores)
because the ISDs have two overlapping peaks (Fig. 13a), despite the IPSD having only
one main peak of mesopores (Fig. 1). The formation of more distinct water (as well as
benzene) structures is observed on the adsorption of water/benzene mixtures because of
the effects of benzene (Fig. 13). In the case of the water/chloroform-d mixture (Fig. 13e),
water is mainly in narrow mesopores of MCM-41 (only a small portion of adsorbed water
is in pores at R > 2 nm) similar to that for adsorbed pure water (Fig. 13a). Consequently,
the effect of nonpolar benzene on the pore water structure is stronger than that of weakly
polar chloroform-d. This is due to a smaller excess of the Gibbs free energy on interaction
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
443
of chloroform with clustered water than that on interaction of benzene with this water. It
should be noted that at T ≤ 213 K only benzene (signal at 7.2 ppm) remains unfrozen in
the water/benzene mixture filling pores because adsorbed water completely freezes at T
< 233 K. However, in the case of adsorbed pure water, its portion freezes at much lower
temperatures. In other words, benzene weakens the interaction of water with the pore
walls because water displaced from narrow pores forms larger structures.
The amounts of water adsorbed onto silica in air (h = 0.186 ml/g) or on addition of a
small amount of benzene (h = 0.186 ml/g and b = 0.196 ml/g) provide filling of approximately quarter of the total pore volume by water and high clusterization of adsorbed
water, and approximately half volume (h = 0.317 and 0.37 ml/g respectively) in the case
of addition of chloroform-d (c = 5.0 ml/g) and benzene (b = 0.4 and 5.8 ml/g). A vertical
section observed on the ΔG(Cuw ) graphs (Fig. 11b) corresponds to non-freezing of pore
water over a wide temperature range (Fig. 11a). This pore water is strongly bound to the
silica surface (ΔG < -1 kJ/mol, Fig. 11b), and the smaller the adsorbed water amount,
the lower the freezing temperature of the water. This effect is caused by predominant
occupation of the narrowest pores (characterized by the maximal adsorption potential)
on a partial filling of them (Fig. 13a). At a low hydration the adsorbed water can form
clusters or thin films smaller in size (thickness) than pores. An increase in the hydration
up to h = 5.2 ml/g leads to filling of larger pores because the incremental size distribution,
ISD, of the structures with unfrozen pore water shifts toward larger R values (Fig. 13a).
However, the amount of unfrozen water located in pores at h = 5.2 ml/g remains smaller
than the total pore volume. The cause for this effect can be explained by : (i) air bubbles
remaining in pores inhibit filling of them by water; (ii) a small portion of pyrocarbon and
organics remains (Fig. 14) in pores of silica after its calcination on the template removal;
and (iii) a portion of water (far from the pore walls) filling broader mesopores at R > 4
nm (Fig. 1) can be frozen at temperature close to 273 K. However, it is the first two factors that take a predominant role because contribution of broad mesopores in MCM-41 is
very low (Fig. 1) and weakly bound water (which can be located in broad pores and thus
satisfy the condition ΔG > −0.5 kJ/mol) is not observed (Fig. 11b). A slightly decreased
hydrophilicity of the pore walls of MCM-41 due to small amounts of pyrocarbon and
trace organic functionalities can strengthen bubble plugs in narrow mesopores at R < 2
nm in which the capillary condensation of water is absent. The FTIR spectra (Fig. 14,
MCM-41 (sample 1), MCM-48 and SBA-15) show that small amounts of the template
remain in all the samples as indicated by the low three peaks at 2925 cm−1 (νCH,as ), 2850
cm−1 (νCH,s ) and 1465 cm−1 (δCH2,as ) corresponding to vibrations of the CH2 groups.
Additional heating of MCM-41 in air at 473 K for 2 h affects only adsorbed water, the
intensity of broad band of the O-H stretching vibrations of water at 3200 cm−1 decreases
after heating and the intensity of free silanols at 3740 cm−1 increases and a shoulder at
3650 cm−1 appears. The presence of low amounts of residual organics is typical for silicas
synthesized using organic templates because of the difficulty of complete removal of them
from pores. Residual hydrophobic CH2 groups can affect the amounts of pore water (i.e.
pore filling by water) and its structure. However, the water/benzene mixture can fill the
Unauthenticated
Download Date | 6/18/17 9:10 AM
444
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 11 Amounts of unfrozen (a, e) water and (c, d) water and benzene, and (b, f) the
relationship between changes in the Gibbs free energy of adsorbed water and unfrozen
water amounts on MCM-41 (sample 2); (d) shows an initial portion of the graphs shown
in (c).
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
445
Fig. 12 Amounts of unfrozen benzene for different water/benzene mixtures adsorbed
onto MCM-41 (sample 2).
total pore volume that may confirm the above-mentioned assumption.
The enthalpy of immersion of MCM-41 in water Δim H = −102 mJ/m2 and of the
enthalpy of its wetting in water Δw H = −78 mJ/m2 [48]. Re-calculation of the γS
values (Table 2) with consideration for a portion of filled pores (multiplying γS by z =
103 ρuw Vp /Cuw |T =273K because Cuw is given in mg/g and assuming ρuw = 1 g/cm3 ) gives
the values (Table 2, zγ S ) close to Δim H at different h values for the air and CDCl3
media. In the case of the water/benzene mixtures, the zγ S values related to the total
changes in the Gibbs free energy of the pore water is much smaller (by a factor of 3-4)
because of the displacement of adsorbed water by benzene from narrow pores and the
pore walls. This effect is caused by the fact that direct interaction of benzene with the
silica surface instead of interaction with water can cause a reduction of the Gibbs free
energy of the system on simultaneous diminution of the interfacial contact area between
poorly immiscible benzene and water. The zγ S (as analogous to Δim H) values for MCM48 and SBA-15 are close to that of MCM-41 (Table 2). An increase in these values with
decreasing h value can be caused by higher energy of the adsorption of water in narrower
pores that can give a slight overestimation of the zγ S values.
Chloroform can fill a portion of the pores and displace a fraction of water from the
pore walls and the narrowest pores into larger pores (comp. Figs. 1, 13a, and 13e) that
leads to a starting freezing temperature depression (corresponding to dC w /dT = 0) by
approximately 20 K (Figs. 11a and 15, curves 2 and 6). The corresponding maximum
temperature below which pore water begins to freeze determines minimal reduction of the
Gibbs free energy caused by adsorption interactions with the pore walls. These changes
are close for the air and aqueous media (bending of the ΔG(Cuw ) curves at -2 kJ/mol
characteristic for strongly bound water). However, in the case of the CDCl3 medium it is
Unauthenticated
Download Date | 6/18/17 9:10 AM
446
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 13 Distribution function of sizes of structures unfrozen liquids in MCM-41 (sample
2) pores: (a) pure water, (b, c, d) water and benzene, and (e) water (with the presence
of chloroform-d) calculated using Eqs. (4) and (5).
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
447
Table 2 Energetic characteristics of interfacial water adsorbed individually or in mixture
with benzene or chloroform-d onto MCM-41 (sample 2), MCM-48 and SBA-15.
Parameter
MCM-41
γ S , mJ/m2
γ S , mJ/m2
γ S,w , mJ/m2
γ S,w , mJ/m2
γ S,w , mJ/m2
γ S,w , mJ/m2
zγ S , mJ/m2
zγ S , mJ/m2
zγ S,w , mJ/m2
zγ S,w , mJ/m2
zγ S,w , mJ/m2
zγ S,w , mJ/m2
zγ S,w , mJ/m2 (MCM-48)
zγ S,w , mJ/m2 (SBA-15)
68.0 (h = 5.2 ml/g)
26.0 (h = 0.186 ml/g)
25.0 (h = 0.186 ml/g, c = 5.0 ml/g)
10.4 (h = 0.186 ml/g, b = 0.196 ml/g)
15.6 (h = 0.37 ml/g, b = 0.4 ml/g)
12.0 (h = 0.37 ml/g, b = 5.8 ml/g)
100.1 (h = 5.2 ml/g)
98.1 (h = 0.186 ml/g)
94.3 (h = 0.186 ml/g, c = 5.0 ml/g)
39.3 (h = 0.186 ml/g, b = 0.196 ml/g)
29.6 (h = 0.37 ml/g, b = 0.4 ml/g)
22.8 (h = 0.37 ml/g, b = 5.8 ml/g)
101 (h = 0.2 ml/g)
116 (h = 0.16 ml/g)
equal to approximately -1 kJ/mol (Fig. 11). The extrapolation of the ΔG(Cuw ) curves to
the Y-axis determines the maximum reduction in the Gibbs free energy of the first surface
layer of water at ΔGs ≈ −4 kJ/mol [13]. This ΔGs value is lower (by -1.5 kJ/mol) for the
air and aqueous media in comparison with CDCl3 (ΔGs ≈ -2.5 kJ/mol). Consequently,
pore chloroform-d causes certain diminution of pore water interaction with the MCM-41
surface; however, weakly bound water at ΔG > -0.5 kJ/mol is absent with the presence of
CDCl3 as in the case of adsorbed pure water (Fig. 11b). Freezing of pore fluid is sensitive
to its states dependent on the degree of pore filling, and the adsorbed film (closest to the
pore walls) and fluid at the middle of pores freeze at different temperatures [16]. Therefore
the effect of chloroform-d on the pore water can be explained by the displacement of a
portion of water from the narrowest pores into larger ones or from the pore walls towards
the middle of pores or out of pores. For instance, water fills pores at R > 1 nm (Fig. 13e),
and there are two maxima of the water structures (ISD), as well as in the case of the
adsorption of pure water (Fig. 13a); however, the left portion of the ISD is “cut” by the
pore chloroform-d. This suggests nonuniform freezing of pore water or water/chloroformd because thin liquid layers or small clusters locating close to the pore walls freeze at
lower temperatures than liquid domains locating far from the pore surface that causes
appearance of two ISD peaks of water.
Benzene affects the characteristics of the pore water more strongly than chloroform-d
(Figs. 11-13 and 14) because it can more strongly displace water from the pore walls
and pores (Fig. 13), especially at its excess at b = 5.8 ml/g (Figs. 13c and 13d). An
ISD peak at R < 1 nm of unfrozen benzene (Figs. 13b, 13c, and 13d) can correspond
to a thin layer locating at the pore walls when water is displaced from the pore walls
Unauthenticated
Download Date | 6/18/17 9:10 AM
448
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
Fig. 14 FTIR spectra of (a) MCM-41 (sample 1) initial and treated in air at 473 K for
2 h, and (b) MCM-48 and SBA-15.
or out of pores. This effect causes diminution of the freezing point depression of water
(comp. Figs. 11a and 11c), and changes in the Gibbs free energy of adsorbed water are
smaller (Fig. 11f) than in the case of the adsorption of pure water or water/chloroformd (Fig. 11b). The benzene/water mixture can fill total pore volume because the Cu
value corresponds to the Vp value (Fig. 11c). It is probable that clustered structure
forms (especially in the case of sequential addition of water and benzene to MCM-41)
in which contacts of benzene with the pore walls can be predominant because the ISDs
related to benzene shift toward smaller R values in comparison with the ISDs linked
to water (Figs. 13b, 13c, and 13d). There is certain broadening of these ISDs because
of incomplete compensation of the interference of water and benzene on calculations.
This rearrangement of pore liquids diminishes the interaction energy of water with the
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
449
Fig. 15 Derivative dC u /dT as a function of temperature at different amounts of water,
benzene and chloroform-d (sample 2) for (a) total amounts of water or water/benzene;
(b) water and (c) benzene adsorbed onto MCM-41; and (d) IPSD calculated on the basis
of the nitrogen adsorption isotherm and ISD for water/benzene adsorbed onto MCM-41
(using GT and IGT equations).
silica surface (Fig. 13f). A major fraction of water forms relatively large structures,
e.g. domains and superdomains [13], responsible for the 1 H NMR signal at δH ≈ 5 ppm
(Fig. 10). On the other hand, a minor portion of water at large excess of benzene (b
= 5.8 ml/g) can be assigned to weakly associated (δ H = 1.2-1.4 ppm, Fig. 10d) but
strongly bound (as it freezes at very low temperatures) water. Another portion of the
pore water is weakly bound (ΔG > -0.5 kJ/mol, Fig. 11f) because its interaction with
benzene occurs without formation of strong hydrogen bonds characteristic for the bulk
water (δ H ≈ 5 ppm) or water interacting with the silica surface (at the same δ H values).
The effects of the surface forces and intermolecular interaction between benzene and
water molecules, resulting in the structurization of the mixture in pores, ensure the
minimum value of Gibbs free energy for this complex system. However, interaction energy
Unauthenticated
Download Date | 6/18/17 9:10 AM
450
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
of the water-water, water-silica, benzene-benzene, and benzene-silica types is greater
than that of water-benzene. Therefore the ΔG(Cuw ) curves for the pore water in the
benzene/water mixtures (Fig. 11f) locate over the curves of the pure water (Fig. 11b),
and the benzene/water/MCM-41 system should tend to reach a state with a minimal
area of the water/benzene interfaces in pores. For instance, maxima of the ISDs at R
= 1.0-1.5 nm related to adsorbed unfrozen water (Figs. 13b, 13c and 13d) correspond to
minima of the ISD related to adsorbed benzene. Additionally, benzene fills a fraction of
the narrowest pores (Fig. 13) and changes in the Cub (T) values on benzene interaction
with silica (Fig. 12) are larger than that of water (Fig. 11). According to Aksnes and
Kimtys [19], all benzene freezes in silica gel pores (average R ≈ 2 nm) at T > 227 K that
is in agreement with the dC u /dT peak at 230-250 K of benzene adsorbed on MCM-41
(Fig. 15). Notice that MCM-41 has narrower pores than the mentioned silica gel, i.e.
the freezing temperatures of pore benzene adsorbed onto MCM-41 can be lower. The
dC u /dT peak at 270 K corresponds to thawing of benzene in larger pores and far from
the pore walls because its melting point in the bulk is only slightly higher (∼278 K). A
minor portion of adsorbed water, which corresponds to weakly bound one, can thaw at
temperatures close to 270 K (Figs. 11e and 11f).
A simple additive model of water/benzene mixture completely filling pores (h= 0.37
ml/g and b=5.8 ml/g) based on Eq. (5)
dVuw (Ruw ) dVub (Rub )
dVu (Ruw + Rub )
=
+
(10)
dR
dR
dR
gives the distribution function close in R values to IPSDN 2 but with a more complex shape
(Fig. 15d). This is due to the presence of several structures with benzene and water as
reflected in the f (T 2) distributions (Fig. 5) and the ISDs (Figs. 6-8, 13 and 15d). This
complexity causes certain broadening of the ISD. Calculation of the ISD with the IGT
equation (Eq. (6)) gives a complex curve (because of the pore benzene effects), which,
however, has a maximum close to that of IPSD.
4
Conclusion
The study of the structural characteristics of ordered mesoporous silicas using an improved procedure (CONTIN/MEM-j) or a complex model of cylindrical pores and voids
between spherical particles suggests that the use of alcohol in the reaction solution and
additional aging of the reaction product causes increased roughness of the pore walls.
The textural porosity of MCM-41 samples synthesized with the presence of ethanol is
lower than that of MCM-48 and SBA-15 synthesized without alcohol. The roughness of
the pore walls and residual organics provide the appearance of weakly associated water
adsorbed in a mixture with an organic solvent.
The investigations of pure water and water/chloroform-d or water/benzene mixtures
adsorbed onto MCM-41 by using the 1 H NMR spectroscopy with layer-by-layer freezingout of bulk and pore liquids show that both chloroform-d and benzene can displace a
portion of water from narrow pores and/or from the pore walls to reduce the contact
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
451
area between immiscible liquids. The effects of nonpolar benzene are stronger than that
of weakly polar chloroform-d. This effect of benzene leads to diminution of freezing temperature depression of adsorbed water and its interaction energy with the pore walls.
Water fills only a portion of MCM-41 pores at high hydration (h = 5.2 ml/g); however,
the water/benzene mixture fills total pore volume. The modified Gibbs-Thomson relation for freezing point depression and equations related to transverse relation processes
in adsorbed water/benzene mixtures elucidate the difference in structurization and rearrangement of pure pore water and water/benzene or water/chloroform-d mixtures. This
structurization depends on the pore size distribution, amounts of pore liquids, their clusterization, the order of their adsorption, the pore wall structure and residual amounts
of organic templates. One could expect a smaller effect of benzene or other nonpolar
solvents on the structure of water adsorbed onto more hydrophilic surfaces than that of
silica, e.g. mixed mesoporous oxides, on which water can form a continuous layer.
Acknowledgment
V.I.G. is grateful to Prof. V.G. Il’in for the help in the characterization of some silica
samples.
References
[1] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt,
C.T.W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins and J.L.
Schlenker: “A new family of mesoporous molecular sieves prepared with liquid crystal
templates”, J. Am. Chem. Soc., Vol. 114, (1992), pp. 10834–10843.
[2] L. Bonneviot, F. Belard, C. Danumah, S. Giasson and S. Kaliaguine (Eds.): Mesoporous Molecular Sieves, Studies in Surface Science and Catalysis, Vol. 117, Elsevier,
Amsterdam, 1998.
[3] A. Davidson: “Modifying the walls of mesoporous silicas prepared by supramoleculartemplating”, Curr. Opin. Colloid In. Sci., Vol. 7, (2002), pp. 92–106.
[4] U. Ciesla and F. Schüth: “Ordered mesoporous materials”, Micropor. Mesopor.
Mater., Vol. 27, (1999), pp. 131–149.
[5] C.J. Brinker and G.W. Scherer: Sol-Gel Science, The Physics and Chemistry of
Sol-Gel Processing, Academic Press, New York, 1990.
[6] L.P. Legrand (Ed.): The Surface Properties of Silicas, Wiley, New York, 1998.
[7] M.M.L. Ribeiro Carrott, A.J.E. Candeias, P.J.M. Carrott, P.I. Ravikovitch, A.V.
Neimark and A.D. Sequeira: “Adsorption of nitrogen, neopentane, n-hexane, benzene
and methanol for evaluation of pore size in silica grades of MCM-41”, Micropor.
Mesopor. Mater. Vol. 47, (2001), pp. 323–337.
[8] W. Zhou and J. Klinowski: “The mechanism of channel formation in the mesoporous
molecular sieve MCM-41”, Chem. Phys. Lett., Vol. 292, (1998), pp. 207–212.
Unauthenticated
Download Date | 6/18/17 9:10 AM
452
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
[9] V.V. Guliants, M.A. Carreon and Y.S. Lin: “Ordered mesoporous and macroporous
inorganic films and membranes”, J. Membrane Sci., Vol. 235, (2004), 53–72.
[10] M. Jaroniec, M. Kruk, H.J. Shin, R. Ryoo, Y. Sakamoto and O. Terasaki: “Comprehensive characterization of highly ordered MCM-41 silicas using nitrogen adsorption,
thermogravimetry, X-ray difraction and transmission electron microscopy”, Micropor. Mesopor. Mater., Vol. 48, (2001), pp. 127–134.
[11] D. Kumar, K. Shumacher, C. du Fresne von Hohenesche, M. Grün and K.K. Unger:
“MCM-41, MCM-48 and related mesoporous adsorbents: their synthesis and characterisation”, Colloid Surf. A, Vol. 187-188, (2001), pp. 109–116.
[12] M. Chaplin: “Water Structure and Behaviour”, http://www.lsbu.ac.uk/water/.
[13] V.M. Gun’ko, V.V. Turov, V.M. Bogatyrev, V.I. Zarko, R. Leboda, E.V. Goncharuk,
A.A. Novza, A.V. Turov and A.A. Chuiko: “Unusual properties of water at hydrophilic/hydrophobic interfaces”, Adv. Colloid Interfac. Sci., Vol. 118, (2005), pp.
125–172.
[14] S. Sklari, H. Rahiala, V. Stathopoulos, J. Rosenhjlm and P. Pomonis: “The influence
of surface acid density on the freezing behavior of water confined in mesoporous
MCM-41 solids”, Micropor. Mesopor. Mater., Vol. 49, (2001), pp. 1–13.
[15] M. Sliwinska-Bartkowiak, G. Dudzoak, R. Gras, R. Sikorski, R. Radhakrishnan and
K.E. Gubbins: “Freezing behavior in porous glasses and MCM-41”, Colloid Surf. A,
Vol. 187-188, (2001), pp. 523–529.
[16] K. Morishige and H. Iwasaki: “X-ray study of freezing and melting of water confined
within SBA-15”, Langmuir, Vol. 19, (2003), pp. 2808–2811.
[17] J.H. Strange, J. Mitchell and J.B.W. Webber: “Pore surface exploration by NMR”,
Magn. Res. Imaging, Vol. 21, (2003), pp. 221–226.
[18] D.P. Gallegos, K. Munn, D.M. Smith and D.L. Stermer: “A NMR technique for the
analysis of pore structure: application to materials with well-defined pore structure”,
J. Coll. In. Sci., Vol. 119, (1986), pp. 127–140.
[19] D.W. Aksnes and L. Kimtys: “1H and 2 H NMR studies of benzene confined in porous
solids: melting point depression and pore size distribution”, Solid State Nucl. Mag.,
Vol. 25, (2004), pp. 146–152.
[20] M.R. Landry: “Thermoporometry by differential scanning calorimetry: experimental
considerations and applications”, Thermochim. Acta, Vol. 433, (2005), pp. 27–50.
[21] V.M. Gun’ko, V.V. Turov, J. Skubiszewska-Zieba,
R. Leboda, M.D. Tsapko and D.
Palijczuk: “Structural characteristics of a carbon adsorbent and influence of organic
solvents on interfacial water”, Appl. Surf. Sci., Vol. 214, (2003), pp. 178–189.
[22] V.M. Gun’ko, V.V. Turov, R. Leboda, J. Skubiszewska-Zieba,
M.D. Tsapko and
D. Palijczuk: “Influence of organics on structure of water adsorbed on activated
carbons”, Adsorption, Vol. 11, (2005), pp.163–168.
[23] V.V. Turov, G.P. Prikhodko, S.Y. Brychka and M.D. Tsapko: “Joint Adsorption of
Benzene and Water on Carbon Nanotubes “, Russ. J. Phys. Chem., Vol. 80, (2006),
pp. 591–596.
Unauthenticated
Download Date | 6/18/17 9:10 AM
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
453
[24] V.V. Turov, V.M. Gun’ko, M.D. Tsapko, V.M. Bogatyrev, J. Skubiszewska-Zieba,
R. Leboda and J. Riczkowski: “Influence of organic solvents on interfacial water at
surfaces of silica gel and partially silylated fumed silica”, Appl. Surf. Sci., Vol. 229,
(2004), pp. 197–213.
[25] I.S. Berezovska, V.V. Yanishpolsky and V.A. Tertykh: “Nanoporous silicas with
uniform spherical morphology of particles”, Nanostruct. Mater. Sci., Vol. 2-6, (2005),
pp. 24–28.
[26] A. Wang and T. Kabe: “Fine-tuning of pore size of MCM-41 by adjusting the initial
pH of the synthesis mixture”, Chem. Commun., Vol. 20, (1999), pp. 2067–2068.
[27] K. Schumacher, P.I. Ravikovitch, A. Du Chesne, A.V. Neimark and K.K. Unger:
“Characterization of MCM-48 Materials”, Langmuir, Vol. 16, (2000), pp. 4648–4654.
[28] A. Lind, C. du Fresne von Hohenesche, J.-H. Smått, M. Lindén and K.K. Unger:
“Spherical silica agglomerates possessing hierarchical porosity prepared by spray
drying of MCM-41 and MCM-48 nanospheres”, Micropor. Mesopor. Mater., Vol.
66, (2003), pp. 219–227.
[29] R. Ryoo, S.H. Joo and S. Jun: “Synthesis of highly ordered carbon molecular
sieves via template-mediated structural transformation”, J. Phys. Chem. B, Vol.
103, (1999), pp. 7743–7746.
[30] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka and D.G. Stucky: “Nonionic triblock and
star diblock copolymer and oligomeric surfactant syntheses of highly ordered, hydrothermally stable, mesoporous silica structures”, J. Am. Chem. Soc., Vol. 120,
(1998), pp. 6024–6036.
[31] V.M. Gun’ko, V.I. Zarko, V.V. Turov, E.F. Voronin, I.F. Mironyuk and A.A. Chuiko:
“Structural and adsorptive characteristics of fumed silicas in different media”, In:
H.E. Bergna (Ed.): Coll. Silica: Fundamentals and Applications, Taylor & Francis
LLC, Salisbury, 2005, pp. 499–530.
[32] V.M. Gun’ko and V.V. Turov: “Structure of hydrogen bonds and 1 H NMR spectra
of water at the interface of oxides”, Langmuir, Vol. 15, (1999), pp. 6405–6415.
[33] S.W. Provencher: “A constrained regularization method for inverting data represented by linear algebraic or integral equations”, Comp. Phys. Comm., Vol. 27,
(1982), pp. 213–227.
[34] H.-R. Tang, J. Godward and B. Hills: “The distribution of water in native starch
granules – a multinuclear NMR study” Carbohydrate Polym., Vol. 43, (2000), pp.
375–387.
[35] S.-G. Choi and W.L. Kerr: “Effects of chemical modification of wheat strach on
molecular mobility as studied by pulsed 1 H NMR”, Lebensm.-Wiss. U.-Technol.,
Vol. 36, (2003), pp. 105–112.
[36] V.P. Glushko (Ed.): Handbook of Thermodynamic Properties of Individual Substances, Nauka, Moskow, 1978.
[37] V.M. Gun’ko, V.I. Zarko, E.V. Goncharuk, L.S. Andriyko, V.V. Turov, Y.M. Nychiporuk, R. Leboda, J. Skubiszewska-Zieba,
A.L. Gabchak, V.D. Osovskii, Y.G.
Ptushinskii, G.R. Yurchenko, O.A. Mishchuk, P.P. Gorbik, P. Pissis, and J.P. Blitz:
Unauthenticated
Download Date | 6/18/17 9:10 AM
454
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
V.M. Gun’ko et al. / Central European Journal of Chemistry 5(2) 2007 420–454
“TSDC spectroscopy of relaxational and interfacial phenomena”, Adv. Colloid Interface Sci., in press.
S.J. Gregg and K.S.W. Sing: Adsorption, Surface Area and Porosity, 2nd ed., Academic Press, London, 1982.
M. Thommes, R. Köhn and M. Fröba: “Sorption and pore condensation behavior of pure fluids in mesoporous MCM-48 silica, MCM-41 silica, SBA-15 silica and
controlled-pore glass at temperatures above and below the bulk triple point”, Appl.
Surf. Sci., Vol. 196, (2002), pp. 239–249.
C. Nguyen and D.D. Do: “A new method for the characterization of porous materials”, Langmuir, Vol. 15, (1999), pp. 3608–3615.
V.M. Gun’ko and S.V. Mikhalovsky: “Evaluation of slitlike porosity of carbon adsorbents”, Carbon, Vol. 42, (2004), pp. 843–849.
V.M. Gun’ko, O. Seledets, J. Skubiszewska-Zieba,
V.I. Zarko, R. Leboda, W. Janusz
and S. Chibowski: “Phosphorus-containing pyrocarbon deposits on silica gel Si-100”,
Micropor. Mesopor. Mater., Vol. 87, (2005), pp. 133–145.
W.B. Muniz, F.M. Ramos and H.F. de Campos Velho: “Entropy- and Tikhonovbased regularization techniques applied to the backwards heat equation”, Comput.
Mathem. Applic., Vol. 40, (2000), pp. 1071–1084.
M. Jaroniec, M. Kruk and J.P. Olivier: “Standard nitrogen adsorption data for
characterization of nanoporous silicas”, Langmuir, Vol. 15, (1999), pp. 5410–5413.
G.M. Sesler (Ed.): Electrets, Topics in Applied Physics, Vol. 33, Springer, Berlin,
1980.
P.T. Tanev and T.J. Pinnavaia: “Mesoporous silica molecular sieves prepared by
ionic and neutral surfactant templating: a comparison of physical properties”, Chem.
Mater., Vol. 8, (1996), pp. 2068–2079.
J. Wawryszczuk, J. Goworek, R. Zaleski and T. Goworek:”Positron lifetime in mesoporous silica of MCM-41 type”, Langmuir, Vol. 19, (2003), pp. 2599–2605.
M.J. Meziani, J. Zajac, J.-M. Douillard, D.J. Jones, S. Partyka and J. Rozière:
“Evaluation of surface enthalpy of porous aluminosilicates of the MCM-41 type using
immersional calorimetry: effect of the pore size and framework Si:Al ratio”, J. Colloid
Interf. Sci., Vol. 233, (2001), pp. 219–226.
Unauthenticated
Download Date | 6/18/17 9:10 AM