Probing the surface polarity of various silicas and other moderately strong solid acids by means of different genuine solvatochromic dyes

Article abstract of DOI:10.1021/jp000231s

Reichardt's E_T(30) as well as Kamlet-Taft's α (hydrogen-bond-donor acidity) and π^* (dipolarity/polarizability) values of various solid acids, e.g., silicas, aluminas, alumosilicates, titanium dioxides, and alumina-functionalized silica particles, are presented. The E_T(30) values of the solid acids were directly measured by UV/vis spectroscopy using 2,6-diphenyl-4-(2,4,6-triphenyl-1-pyridinio)phenolate (1a) and an eicosafluorine-substituted derivative of 1a (1b) as surface polarity indicators. Kamlet-Taft's α and π^* parameters for the various solid acids are analyzed by means of Fe(phen)₂(CN)₂ [cis-dicyano-bis(1,10-phenanthroline)iron(II)] (2) and Michler's ketone [4,4′-bis(N,N-dimethylamino)benzophenone] (3) as solvatochromic surface polarity indicators. The chemical interpretation of the α and π^* parameters and the nature of the surface sites which they reflect are discussed. The correspondence of the UV/vis spectroscopic results to those of the IR-sensitive probe benzophenone and the fluorescence probe pyrene (literature data) is also discussed. It is evident that the solid surface environments observed by the various indicators are moderately strong dipolar/polarizable (π^* = 0.38 to 1.04) and are fairly strong hydrogen-bond donors (α = 1.00 to 1.99). Theoretical E_T(30) values of solids acids are calculated by applying linear solvation energy (LSE) relationships using the independently measured α and π^* values of the solid acids according to E_T(30) = [E_T(30)]₀ + aα + sπ^*. The respective dependence of the α and π^* terms, expressed by the coefficients a and s, upon the measured E_T(30) value for solid acids is discussed in comparison to related multiple LSE correlations known for well-behaved regular solvents and functionalized silicas. The results show in general that values of polarity parameters of strong HBD and dipolar surfaces are strongly influenced by the experimental conditions applied.

Full text of DOI:10.1021/jp000231s

J. Phys. Chem. B 2000, 104, 6417-6428
6417
Probing the Surface Polarity of Various Silicas and Other Moderately Strong Solid Acids
by Means of Different Genuine Solvatochromic Dyes
Stefan Spange,* Elmar Vilsmeier, and Yvonne Zimmermann
Department of Polymer Chemistry, Institute of Chemistry, Chemnitz UniVersity of Technology, Strasse der
Nationen 62, D-09111 Chemnitz, Germany
ReceiVed: January 18, 2000; In Final Form: May 9, 2000
Reichardt’s ET(30) as well as Kamlet-Taft’s R (hydrogen-bond-donor acidity) and π* (dipolarity/polarizability)
values of various solid acids, e.g., silicas, aluminas, alumosilicates, titanium dioxides, and alumina-
functionalized silica particles, are presented. The ET(30) values of the solid acids were directly measured by
UV/vis spectroscopy using 2,6-diphenyl-4-(2,4,6-triphenyl-1-pyridinio)phenolate (1a) and an eicosafluorine-
substituted derivative of 1a (1b) as surface polarity indicators. Kamlet-Taft’s R and π* parameters for the
various solid acids are analyzed by means of Fe(phen)2(CN)2 [cis-dicyano-bis(1,10-phenanthroline)iron(II)]
(2) and Michler’s ketone [4,4′-bis(N,N-dimethylamino)benzophenone] (3) as solvatochromic surface polarity
indicators. The chemical interpretation of the R and π* parameters and the nature of the surface sites which
they reflect are discussed. The correspondence of the UV/vis spectroscopic results to those of the IR-sensitive
probe benzophenone and the fluorescence probe pyrene (literature data) is also discussed. It is evident that
the solid surface environments observed by the various indicators are moderately strong dipolar/polarizable
(π* ) 0.38 to 1.04) and are fairly strong hydrogen-bond donors (R ) 1.00 to 1.99). Theoretical ET(30)
values of solid acids are calculated by applying linear solvation energy (LSE) relationships using the
independently measured R and π* values of the solid acids according to ET(30) ) [ET(30)]o + aR + sπ*.
The respective dependence of the R and π* terms, expressed by the coefficients a and s, upon the measured
ET(30) value for solid acids is discussed in comparison to related multiple LSE correlations known for well-
behaved regular solvents and functionalized silicas. The results show in general that values of polarity
parameters of strong HBD and dipolar surfaces are strongly influenced by the experimental conditions applied.
Introduction
on kinetic, thermodynamic, or spectroscopic data relating to
certain reference reactions.17-22 Significantly, different empirical
solvent polarity scales have been shown to correlate well with
each other, pointing to the existence of an underlying common
feature.18,23,24 (See also the discussion of the meaning of eqs 2,
3, and 4.)
Empirical solvent polarity scales based on spectroscopic
measurements usually employ changes in the UV/vis absorption
maximum of an indicator in different solvents (solvato-
chromism).17 Visible probing with suitable indicator dyes was
also recommended by several authors for investigating the
surface polarities of solids.25-32
Reichardt’s dye and substituted derivatives of it are very
sensitive solvatochromic polarity indicators for detecting gradual
differences in the polarity of various environments.17,22,33-35
We,24,30b,36 and others,31,32 recently reported that the polarity
of silica, organically functionalized silica, or alumina surfaces
can also be investigated by Reichardt’s dye 1a (see eqs 1a and
1b, and Chart 1).
The quantification of the surface polarity and catalytic activity
of solid acid catalysts is an important field of materials
research.1-4 This area of research has been stimulated by the
application of improved adsorption-calorimetric and solid-state
NMR spectroscopic methods and techniques to well-established
solid catalyst materials in examining their surface properties.4-8
These methods require special equipment and are particularly
time-consuming. Recently, a number of original contributions
have been published where the development and use of novel
solid acids are described.9,10 Therefore appropriate methods are
required which allow a rapid check of the relevant surface
properties of a solid acid catalyst. For instance, the catalytic
activity of solid acids can be simply estimated by Hammett
indicators, which can also be applied to nonaqueous media.11-14
However, the original Hammett indicators allow only a rough
estimation of the surface acidity of solid acids. They are not
suitable for observing the overall surface polarity.1,13 To date,
no generally agreed upon definition of the term surface polarity
has emerged.15 In the broadest and most general sense the
surface polarity can be viewed as the sum of all interactive forces
between an adsorbed molecule and the occupied surface site or
sites. This definition is based on the related interpretation of
the term solvent polarity.16 Although solvent polarity is among
the most widely used concepts in chemistry (see refs 16-19
for reviews), there has been little consideration of the changes
in surface polarity associated with structural variations in solid
acids.1-8 Most solvent polarity scales are empirical and are based
ET(30) [kcal mol-1] ) 2.891 × 10-3 νmax (1a) [cm-1]
(1a)
ETN ) [ET(30) - 30.7]/32.4
(1b)
For instance, the ET(30) value is about 58 to 59 kcal mol-1 for
bare silica particles36 and 62 to 65 kcal mol-1 for γ-alumina
powders.32
Unfortunately, Reichardt’s standard dye 1a can be used only
with weak acids because its phenolate oxygen atom becomes
protonated in acidic environments with pKa values lower than
* Author to whom correspondence should be addressed.
10.1021/jp000231s CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/21/2000

6418 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Spange et al.
CHART 1: Formulas of the Probe Dyes Used
dipolarity/polarizability (40%) of the solvents. It was also shown
that the â term of the solvents scarcely contributes to the
ET(30) value of well-behaved regular solvents.17,38
Coordination of protic solvents to the phenolate oxygen of
the betaine dye causes a strong hypsochromic shift because the
electron density of the phenoxide part of the dye is drastically
reduced. Consequently, a similar influence of HBD groups on
the ET(30) value is observed when 1a is adsorbed on surface
silanol groups.36
Several authors have shown that the Kamlet-Taft solvent
parameters R, â, and π* are better suited for the quantitative
description of results of solvent-solid interactions, e.g. chro-
matographic and adsorption processes,39-42 than is the single
ET(30) solvent parameter. Furthermore, the advantage of the
Kamlet-Taft parameters is that they can be converted by means
of multiple LSE relationships into other useful polarity scales,
e.g., the acceptor number (AN) or donor number (DN) scales
of Gutman and the ET(30) scale of Reichardt.18
LSE equations and results from pyrene and 4-(N,N-dimethyl-
amino)benzonitrile (DMAB) fluorescence measurements have
been also used for determining ET(30) values for cation-
exchanged Y-zeolites.43 The surface polarity of these solid acids
was about ET(30) ) 45 to 60 kcal mol-1. However, these data
are not representative, because the two indicators pyrene and
DMAB, respectively, mainly measure the dipolarity/polariz-
ability π* of an environment. Reasonable R values of zeolites
were reported by Dutta and Turbeville, using salicylideneaniline
as a UV/vis probe.27 The authors found that the R value of
zeolites significantly depends on their Al/Si ratios. The R value
was between 1.06 and 1.14 (Rav ) 1.1) for the pure siliceous
form, approximately 1.5 to 1.6 for the alumina phase, and the
maximum R value (R ) 1.85) was found at a Si/Al ratio of
about 8.
π* values for silicas and alumosilicates were also reported
using 4-nitroanisole and related indicators.25 The reported values
changed between π* ) 0.65 and 2.00. We are afraid that the
reported π* values for the silica surface from ref 25 also contain
contributions of acid-base interactions, because the calculated
π* values are dependent on the basicity strength of the indicator
used.30b
An independent method for confirming the accuracy of the
observed R and π* values of solid acids is difficult to find,
because there are only a few probe molecules known which
are suitable for this purpose. There are several known solvent-
dependent processes, the results of which correlate with both,
the R and π* or the â and π* term.44 For instance, the only
π*-sensitive dye 5-(N,N-dimethylamino)-5′-nitro-2,2′-bisthio-
phene45 is not well adsorbed on silica and shows also a nonlinear
UV/vis spectroscopic bathochromic shift on acidic surfaces.30b
There are only two indicators which are known to be selectively
responsive to the HBD term. These are pyridinium N-oxide46
and 4-nitropyridine-N-oxide.47 Unfortunately, the application of
pyridinium-N-oxide to solid acids would require high-resolution
solid state 13C NMR spectroscopy. Thus, a widespread use of
this indicator is not possible. In strong HBD environments, the
UV/vis absorption of 4-nitropyridine-N-oxide is expected at νmax
> 30 000 cm-1 and can therefore not easily be utilized when
the indicator is adsorbed on solid acids.
6.35 Therefore widespread use of the dye 1a for measuring the
surface polarity of a variety of solid acids, e.g., aluminas,
titanium oxides, alumosilicates, is not possible.
However, the interaction of a surface environment with a
solvatochromic dye is a composite of many effects. Acid-base,
dipole-dipole, induced dipole-dipole, and dispersion forces
contribute to the overall adsorption energy of a probe with an
inorganic surface. This means that for each UV/vis spectrum
of an adsorbed solvatochromic dye measured, surface sites of
different polarity as well as different contributions of specific
(acid-base) and non specific (dipolar-polarizable) interactions
must be taken into account.
Multiple intermolecular solute/solvent interactions can be
described by the LSE (linear solvation energy) relationship of
Kamlet and Taft.37 The simplified Kamlet-Taft equation applied
to single solvatochromic shifts, XYZ ) νmax(probe),17 is given
in eq 2:
XYZ ) (XYZ)o + aR + bâ + s(π* + dδ)
(2)
(XYZ)o is the solute property of a reference system, e.g., a
nonpolar medium, R describes the HBD (hydrogen-bond-
donating) acidity, â the HBA (hydrogen-bond-accepting) ability,
and π* the dipolarity/polarizability of the solvents. δ is a
polarizability correction term which is 1.0 for aromatic, 0.5 for
polyhalogenated, and zero for aliphatic solvents; a, b, s, and d
are solvent-independent correlation coefficients.37
First, Marcus has shown that the empirical ET(30) solvent
polarity parameter can be expressed by an LSE relationship
using the Kamlet-Taft solvent parameters R and π* [eq 3]; n
is the number of solvents considered, r is the correlation
coefficient, and sd is the standard deviation.38
In this paper, we employ a UV/vis spectroscopic method
suitable for simultaneously examining the surface acidity R and
the dipolarity/polarizability π* of moderately strong solid acids
by means of the two surface polarity indicators Fe(phen)2(CN)2
(2) and Michler’s ketone (3).30a The transition metal complex
Fe(phen)2(CN)2 [cis-dicyano-bis(1,10-phenanthroline)iron (II)]
ET(30) ) 31.2 + 15.2R + 11.5π* (a/s ) 1.32)
n ) 155, r ) 0.98, sd ) 1.1
(3)
Equation 3 demonstrates that the ET(30) value for solvents
contains contributions of the hydrogen bond acidity (60%) and

Surface Polarity of Various Silicas and Other Solid Acids
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6419
(Chart 1) has been especially recommended as a surface polarity
indicator for acidic surfaces because its solvatochromic band
shift in the UV/vis spectrum is significantly dependent on the
acidity of the environment.48-55
CHART 2: Two Different Structures of 3 Adsorbed on
Brønsted and Lewis Acid Sites of Surfaces
The solvatochromic UV/vis absorption band of 3 is attributed
to a π-π* transition which overlaps with the less intense n-π*
transition.50,56 The solvatochromism of 3 has been investigated
by several authors.50,56,57 On increasing both the HBD capacity
and dipolarity/polarizability of a solvent, the UV/vis absorption
band of 3 undergoes a significant bathochromic shift. The
bathochromic effect is even significantly enhanced when a
strong Lewis acid or a mobile proton interacts additionally with
the carbonyl oxygen atom of 3 because a mesomerically
stabilized oxenium ion is generated.58-60 In contrast, protonation
of the nitrogen atoms of the dimethylamino groups by a
Brønsted acid has a hypsochromic influence on the solvato-
chromic UV/vis absorption band of 3.50 For instance, the UV/
vis absorption band of [4-(CH3)2NC6H4]2CO-Si(CH3)3+ appears
at λ ) 505 nm59 and of {[4-(CH3)2NHC6H4]2CO}2+ at λ ) 310
nm.50 Therefore, the indicator 3 seems suitable as a versatile
UV/vis spectroscopic probe in order to observe different types
of acidic sites and their different acidic strengths.
Because the genuine solvatochromisms of the two indicators
2 and 3 can also be expressed very well in terms of the Kamlet-
Taft solvent parameters R and π*,37,50 their simultaneous use
should allow the corresponding R and π* values of an unknown
solvent or of a heterogeneous system to be determined. Then R
and π* values of surface environments can be determined by
utilizing the UV/vis absorption maxima of the respective surface
polarity indicators [νmax(2) and νmax(3), respectively] when
adsorbed on the solid.30,55 The mathematical equations are given
in the Experimental Section.54
We have already demonstrated that the R value of organically
functionalized silicas can be analyzed precisely by applying the
two indicators 2 and 3.30a The R-value for alkyl-modified silicas
decreases linearly with increasing the surface coverage since
monolayer functionalization is accomplished.30a This result
shows that the average polarity of all HBD sites was measured.
In this paper, we have also presumed that the ET(30) parameter
for functionalized silicas is more strongly affected by the HBD
term than by the dipolarity/polarizability [eq 4 was taken from
ref 37],30a compared to the LSE correlation derived for well-
behaved regular solvents [eq 3 from ref 18].
correlation analyses with directly determined ET(30) values
according to ET(30) ) aR + bπ*.17,30a For this purpose, an
eicosafluorine-substituted phenolate betaine dye 1b61 (see Chart
1) is used. Because of the reduced basicity of the phenolate
moiety, protonation of it by Brønsted acidic sites is suppressed.
It is also well-known that alumina and several alumosilicates
possess Brønsted as well as Lewis acid sites.62 According to
well-established procedures for differentiating between these two
sites, surface titration with pyridine and the sterically hindered
base 2,6-di-tert-butyl pyridine (DTBP) is applied to the solva-
tochromic dyes when adsorbed on the solid acid in an inert
liquid.63,64 It is well-established that DTBP interacts weakly with
common Lewis acids (TiCl4, AlCl3, or the tert-butyl cation),
but it reacts significantly with mobile protons.65 Thus, DTBP
was successfully applied to trap protons in cationic isobutylene
polymerization.66,67 We were able to show that DTBP trapped
mobile protons on the internal surface of MCM-41.64 But we
presume that DTBP is not adsorbed on Lewis acid sites (see
also ref 63).
ET(30) ) 36.1 + 14.84R + 5.33π* (a/s ) 2.78) (4)
n ) 30, r ) 0.96, sd ) 1.12
For these measurements a special UV/vis technique has been
successfully employed.64
For checking the order of magnitude of the R and π* values
of the solid acids determined, the IR probe benzophenone (BP)
is used in addition. Furthermore, this IR probe is of size similar
to that of the solvatochromic indicator 3 and shows a significant
shift of its carbonyl vibration when adsorbed on solids. The
carbonyl stretching vibration (νmax,CO) of BP is also remarkably
solvent-dependent and correlates excellently with the Kamlet-
Taft polarity parameters R and π* [ eq 5]. Equation 5 was taken
from the literature.68
Chemists know from experience that various silica samples,
obtained from several suppliers or synthesized by different
experimental techniques, are often quite different in their surface
properties, e.g., as supports for metallocene catalysts or in
chromatographic applications. Despite the different catalytic
activity of different silica samples, i.e., as supports for polar
reactions, the ET(30) values of various silicas studied are rather
similar and are in the range ET(30) ) 58 ( 1.5 kcal mol-1 30b
.
This result has been reported independently by several
authors.29-31,41 Furthermore, the thermal pretreatment of solid
acids is also important for their surface activity.
The objective of this paper is the determination of the
empirical surface polarity parameters ET(30), R, and π* by
means of the indicators 1a, 2, and 3 for various solid acids
including different silicas, aluminas, titanium oxides, and
alumosilicates. The correspondence of the R and π* values of
solid acids to their ET(30) values are tested by means of
νmax,CO ) 1668.80 - 10.10π* - 11.00R
r ) 0.99, n ) 34
(5)
Pyrene fluorescence data from the literature should be also
suitable for supporting the π* values of the solid acids.69 The
intensity ratio (I1/I3) of the two UV/vis emission bands at λmax
) 354 nm (I1) and 374 nm (I3) of pyrene increases on increasing
the dipolarity/polarizability π* of the solvent [eq 6]. Therefore,

6420 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Spange et al.
source
TABLE 1: Solid Acids Used in This Paper, Their Chemical Constitutions, and Physical Properties
sample
number
BET-surface
specific pore
av pore
diameter [nm]
solid acid sample
SiO2, Aerosil 200
SiO2, Aerosil 300
area [m2g-1
]
volume [cm3g-1
]
1
2
n.d.a
240
n.d.a
423
315
30
390
254
190
340
303
776
268
131
336
52
Degussa
Degussa
Degussa
Merck
Grace
Grace
Grace
Grace
Grace
Grace
Grace
Grace
Grace
Grace
Grace
Degussa
Condea
Condea
Condea
Condea
Condea
Condea
Condea
Condea
Condea
Condea
Condea
b
3
4
SiO2, Aerosil 380
SiO2, KG 60
0.63
1.42
n.d.a
0.90
0.57
0.61
1.02
1.22
0.57
1.68
0.72
0.80
0.27
0.14
0.16
0.46
0.44
0.35
0.36
0.81
0.50
0.34
0.42
0.54
0.79
0.78
0.74
0.77
0.56
0.17
0.50
0.57
9.0
n.d.a
5.5
4.5
6.4
6.0
8.1
1.5
4.5
11.0
5
6
7
8
SiO2, KG SG 432
SiO2, KG LC 1500
SiO2, 1.5% Al2O3 Cogel
AlPO4, SP 8293
9
AlPO4, SP2-8111.09
SiO2/Al2O3, SP 18-8495
SiO2/TiO2, SP 18-8381.01
SiO2, SP 18-8470
SiO2/ZrO2, SP 18-8402.03
Al2O3, SP 18-8510
SiO2/Al2O3, SP2-8515.01
TiO2 P25
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
4.8
10.0
9.3
7.1
3.1
2.5
2.2
2.1
3.5
2.6
1.9
3.2
2.7
3.8
4.4
4.3
4.3
3.3
5.9
10.3
11.4
TiO2 (rutile)
TiO2 (anatas)
30
45
Alumosilicate, Siral 5
Alumosilicate, Siral 10
Alumosilicate, Siral 20
Alumosilicate, Siral 30
Alumosilicate, Siral 40
Alumosilicate, Siral 50
Alumosilicate, Siral 60
Alumosilicate, Siral 80
Al2O3, Pural
296
347
315
351
460
381
356
263
292
413
354
344
361
339
58
c
KG 60 10% Al2O3
KG 60 30% Al2O3
KG 60 50% Al2O3
KG 60 70% Al2O3
KG 60 100% Al2O3
Al2O3 (AlEt2Cl)
Al2O3
b
b
b
b
b
97
100
Condea
Degussa
Al2O3-C
a Not determined. b See experimental part. c Grafted with diethylaluminum cloride (AlEt2Cl) and hydrolyzed with water; the 10% means that
10% Al2O3 is produced per amount of silica.
Dong and Winnik recommended the LSE eq 6 as a tool for
analyzing the value of the π* parameter of the surface
environment.69
Correlations of the examined polarity parameters R and π*
with the catalytic activity of a surface-mediated polar reaction
will be reported in the following paper.74
I1/I3 ) Py ) 0.64 + 1.33(π* - 0.24δ) - 0.25R (6)
n ) 32, r ) 0.9
Experimental Section
Chemicals. 2,6-Diphenyl-4-(2,4,6-triphenyl-1-pyridinio)phe-
nolate (1a) and bis-2,6-(pentafluorophenyl)-4-[bis-2,6-(pen-
tafluorophenyl)-4-phenyl-1-pyridinio]phenolate (1b) were kindly
given by C. Reichardt, University of Marburg. The synthesis
of 1b is still not reported in the literature.61 For supplementary
information see ref 61 or contact Prof. C. Reichardt, University
of Marburg, Gemany.
Michler’s ketone was purchased from Merck and recrystal-
lized twice from ethanol before use. Fe(phen)2(CN)2 was
prepared and purified by the method of Schilt.75 The solvent
1,2-dichloroethane (spectroscopic grade) was purchased from
Merck. It was dried over CaH2, freshly distilled before use, and
stored over carefully dried anhydrous alumina.
The chemical constitutions and physical properties of the solid
acids used in this paper are listed in Table 1.
The silicas, aluminas, alumosilicates, and titanium dioxides
(samples No. 1-27 and Nos. 34, 35) were commercially
available products. The producers of the solid acids are also
indicated in Table 1.
As seen from eq 6, for protic solvents and, accordingly, also
for acidic surfaces, the knowledge of the R value is necessary
in order to separate the π* value from the unit of measurement
Py.
Despite the wide application of pyrene for polarity measure-
ments of solids,70,71 the separation of the π* value from the
unit of measurement Py seems problematic because R and π*
have an opposite influence upon Py. For silicas, a Py value of
adsorbed pyrene of about 1.78 has often been reported.70
However, theoretically possible Py values can be estimated by
means of eq 6 by knowledge of independently calculated R and
π* values for comparison.
Of course, the â term, the hydrogen-bond accepting property,
also contributes to the overall surface polarity. However, the
transition metal complex Cu(tmen)(acac)+, which is a suitable
â indicator for functionalized silicas,72 is not adsorbed well by
acidic surfaces and thus a direct examination of â values is
unlikely to be possible. Suitable â indicators for acidic surfaces
are still not established.72,73 We are afraid this is an intrinsic
problem. However, it is expected that the â term of dried acidic
surfaces should not contribute to the ET(30) value.72 We will
report on this topic in a later paper.
The alumina sample No. 33 was synthesized by a sol-gel
process from diethylaluminum chloride and water. The partially
alumina-functionalized silica particles No. 28-32 were syn-
thesized by reaction of diethylaluminum chloride with silica in
toluene and subsequent hydrolysis with water.

Surface Polarity of Various Silicas and Other Solid Acids
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6421
Procedure for Sample Nos. 28-32. A typical procedure for
synthesizing samples Nos. 28-32 is as follows: A 10-g sample
of silica (KG 60®) was heated for 12 h at 400 °C in a glass
flask. After the sample was cooled to room temperature under
dry argon (CaH2), 100 mL of dry toluene were added. The
desired amount of diethylaluminum chloride was added from a
glass syringe through a septum. The reaction mixture was cooled
in an ice bath during this step. After the addition was complete,
the mixture was refluxed for 5 h. After cooling to room
temperature the modified silica was collected in a closed glass
reactor, washed with three portions of dry toluene, and dried
under argon. The hydrolysis of the surface-bonded diethylalu-
minum chloride was carried out with a water-saturated argon
gas flow over a period of 5 h. After a washing with a small
portion of water, the alumina-modified silica was dried at 110
°C for 4 h. A final thermal treatment at 500 °C over 4 h yielded
a white powder of alumina-functionalized silica, which was then
used for the measurements.
Figure 1. UV/vis absorption spectra series of Fe(phen)2(CN)2 adsorbed
on silica (No. 4) as a function of the probe concentration in 1,2-
dichloroethane, measured by the transmission technique. The corre-
sponding spectra of the dye measured in the reflectance mode for the
suspension and the dried colored silica are marked as strong solid lines.
The concentration of the dye, c, for measurement with the transmission
technique is given in [10-6 moldye/gsilica] (from the bottom to the top):
c ) 2.01, λmax(2) ) 539 nm; c ) 4.01, λmax(2) ) 539 nm; c ) 6.02,
λmax(2) ) 539 nm; c ) 8.02, λmax(2) ) 539 nm; c ) 10.00, λmax(2) )
Procedure for Sample No. 33. The same apparatus was used
as described before for the sample preparation of Nos. 28-32.
In the glass flask, 10 mL of dry toluene were mixed with 10
mL (80.5 mmol) of diethylaluminum chloride. Then a water-
saturated argon flow was passed through the solution over 4 h.
To complete the hydrolysis, 20 mL of water were slowly added.
After phase separation, the toluene phase was washed with water
to remove residual aluminum hydroxide. The aqueous phases
collected were combined and dried in a rotary evaporator under
vacuum. The dried alumina powder was heated at 450 °C for
12 h and finally at 550 °C for 2 h.
540 nm; c ) 14.0, λmax(2) ) 539 nm; c ) 18.1, λmax(2) ) 539 nm; c
) 20.1, λmax(2) ) 538 nm reflectance technique in suspension: 2.65
× 10-6 mol/gsilica, λmax(2) ) 542 nm (uncertain), reflectance technique
without solvent: λmax(2) ) 530 nm. (The UV/vis absorption maximum
is different from that in Table 3 because the silica used for this
experiment was from another batch.)
The equipment employed was a UV/vis spectrometer MCS
400 connected to an immersion cuvette TSM 5A (Zeiss). The
measurement of the UV/vis spectrum of the supernatant solution
is possible in the same cell in which the solid is deposited.
The measurements of silica particle suspensions are advanta-
geous because they can be performed under inert conditions
and reflect better the conditions suitable for liquid chromatog-
raphy and catalysis in organic solvents. The reproducibility of
the UV/vis spectra of the adsorbed dyes is very good. For silicas,
a transparent suspension is obtained in liquid 1,2-dichloroethane.
This allows one to take very good quality transmission spectra
and with excellent reproducibility of less than λmax ( 1 nm.
The position of the UV/vis absorption maximum of the adsorbed
dye on the silica particles remains constant during 1-2 h (see
later).
Figure 1 shows a typical UV/vis spectra series of the dye 2
adsorbed on a silica sample (No. 4). It is clearly seen that the
UV/vis absorption maximum of the adsorbed dye 2 is indepen-
dent of its concentration. Because the silica slurry in 1,2-
dichloroethane is transparent, direct recording of reflectance
spectra is not possible due to light scattering; see the terrible
spectrum shown in Figure 1. However, for silica particle
powders, UV/vis reflectance spectra can be readily monitored
(see the second reflectance spectrum in Figure 1). Unfortunately,
the reflectance technique is not well suited for excluding traces
of water from the air. The hypsochromic effect observed in the
UV/vis reflectance spectrum compared to the transmission
spectrum (∆λ ) -10 nm) is due to moisture traces which
significantly increase the surface polarity of silica. Furthermore,
the influence of the suspension liquid upon the surface polarity
is also of importance. Therefore we have measured the slurries
of the carefully dried solid acid powders in the reflectance mode.
For this purpose the UV/vis spectra of nontransparent particle
suspensions of alumina, titanium dioxide, and of other solids
were recorded by a special reflectance technique. A quartz plate
is used as the bottom of the closed cell containing the solid
sample in a suitable liquid. The sensor head for measuring the
reflectance spectra is located at this quartz plate and the UV/
vis spectrum of the adsorbed dye can be monitored after the
particles are deposited.
Sample Preparation for Determining the Aluminum
Content of the Alumosilicates (Nos. 19-26) by AAS (Atomic
Absorption Spectroscopy). The solid (0.1 g) was placed in a
platinum crucible standing in lukewarm water and mixed with
1 g of ammonium fluoride. Under stirring, sulfuric acid was
added dropwise to the mixture until a clear solution was
obtained. The solution was heated for 15 min and then added
to 10 mL of hydrochloric acid (2 mol/L). The resulting mixture
was transferred to a glass flask which was filled with water to
100 mL. This solution (excluding sample No. 26) was diluted
with water 1:1 because of the high aluminum content and then
measured by AAS.
The calibration for the AAS measurements was carried out
with standard stock solutions containing known aluminum
concentrations.
UV/vis Measurements. The UV/vis absorption maxima of
the dyes 1a, 1b, 2, and 3, respectively, adsorbed on the solid
acid catalyst, have been recorded using a diode array spectrom-
eter with glass fiber optics. The solid acids were heated for 12
h at 400 °C. After cooling the solid sample to room temperature
under dry argon, a solution of the probe dye in 1,2-dichloro-
ethane (DCE) is simply added to the solid material. Care must
be taken to avoid overloading the surface with the indicators
used, as multilayer adsorption is expected in solution at higher
concentrations or interfering absorptions from the nonadsorbed
dye from the solution. Therefore, the amount of dye added was
restricted to 2-3 mg of 1a or 1b per 1 g of solid acid, 0.5 mg
of 2 per 1 g of solid acid, and 0.1 mg of 3 per 1 g of solid acid.
For the UV/vis measurement of transparent slurries, the
amount of the solid acid silica must be limited in order to
achieve a sufficient particle concentration in the continuously
stirred slurry. An amount of 0.1 to 0.2 g silica in 15 mL of the
solvent is very suitable, but a lower amount is also possible.
The amount of the silica portion has no influence on the position
of the UV/vis absorption maximum of the adsorbed probe dye.

6422 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Spange et al.
This technique is very suitable for recording UV/vis spectra
of particles in suspension under inert conditions. The reproduc-
ibility of the recorded UV/vis spectra is very good. The error
on measuring the UV/vis absorption maxima of the dyes
adsorbed on aluminas, alumosilicates, or titanium dioxides, is
λmax ( 1.5 nm when applying the reflectance technique. The
evidence from the measured UV/vis absorption maxima of the
adsorbed dyes was tested by using an independent chemisorbed
species: the triphenylmethylium ion. The positions of the UV/
vis absorption maxima of the triphenylmethylium ion at λmax
) 410 ( 2 nm and 430 ( 2 nm are independent of the solvent
as well as of the counterion.76-78 In other words, they show no
dependence on the environment. For each solid acid powder,
the triphenylmethylium ion UV/vis absorption of chemisorbed
chlorotriphenylmethane is independent of the solid acid used.48
Thus, physical influences upon the shift of the UV/vis absorption
maxima of adsorbed indicator dyes are not really observed
employing the reflectance UV/vis technique for the slurry.
Surface titrations with the sterically hindered base 2,6-di-
tert-butylpyridine (DTBP) and pyridine were carried out directly
in the slurry. The base was added by a microsyringe through a
silicone septum. The spectra were recorded immediately after
the particles were deposited. Three runs were carried out for
each solid acid. The reproducibility is of the same order of
magnitude as measured for the solid particles without the base.
Amounts of nonadsorbed dyes, which possibly compete with
the added base, can be detected in the supernatant solution by
the immersion cell placed directly in the slurry. For all νmax
data reported in this paper, a quantitative adsorption of the dye
from the supernatant solution on the solid acid has been
accomplished.
The quality of eq 8 is sufficient for examining accurate values
of the R parameter. Despite the poor correlation coefficient of
eq 9 in determining π* values, the significance is sufficient for
an estimation. This must be considered by interpreting the
results.
IR Measurements. The infrared spectra were recorded with
a BioRad spectrometer using a special device for recording the
diffusion reflectance infrared Fourier technique (DRIFT) spectra.
For the adsorption of benzophenone (BP) on the solid acids
a cyclohexane stock solution (0.2 mol/L) was prepared.
The dried solid acid (1 g) was poured into a flask under argon
and was immediately covered with 10 mL cyclohexane. The
desired amount of BP (10-1600 µmol) was added. After
shaking for 24 h at room temperature the suspension was
filtered, and the filter residue was washed with 5 mL of
cyclohexane. The solid was dried under high vacuum at room
temperature and stored under argon. The dried solids were
measured by a DRIFT cell that allows one to work with vacuum
or under argon. The IR spectra of the adsorbed BP were obtained
as the difference between the IR spectrum of the sample and
that of the pure solid acid.
Results
The UV/vis spectra of the polarity indicators 1a, 1b, 2, and
3, respectively, adsorbed on the solid acid particles were
immediately recorded after the components were mixed, usually
after a standard period of 5 min. The shape of the UV/vis
spectrum and the position of the absorption maximum remain
constant for 1 h after adsorption using the indicators 1a and 1b
as well as 2. Employing these three indicators, 1,2-dichloroet-
hane was used as solvent, because 1a and 2 are insoluble in
nonpolar solvents such as cyclohexane. As already mentioned,
the dye 1a becomes immediately protonated when adsorbed on
several Brønsted acidic alumosilicates (Nos. 19-26) or aluminas
(Nos. 28-33) which were produced by a sol-gel process from
diethylaluminum chloride as well as on similarly produced
alumina-functionalized silica particles (see Experimental Sec-
tion). Corresponding to Drago’s results,7 the concentration of
the adsorbed dye 1a or 2 did not appear to influence the position
of the UV/vis maximum in the range of concentration used.
See also Figure 1 and explanations in the Experimental Section.
Additionally, Figure 2 parts a and b, show some typical UV/
vis spectra of the dyes 1a and 1b adsorbed on various solid
acids.
The typical intramolecular CT bands of the adsorbed dyes
1a and 1b are very broad in some cases, i.e., for 1a adsorbed
on alumina (No. 35) and for 1b on titanium dioxide (Anatas)
(No. 18). Similar broad UV/vis spectra of 1a on alumina were
reported by Michels and Dorsey.32 This shows that multiple
configurations of the solid surface of alumina are contributing
to the ET(30) value.
The measured UV/vis absorption maxima of the dyes 1a
and 1b, the corresponding ET(30) values [eq 1a], and the
normalized ETN values [eq 1b] for the solid acids used are given
in Table 2.
Calculation of the Surface Polarity Parameters. ET(30)
values can be directly calculated by using the νmax of dye 1a
from the UV/vis spectrum and eq 1a. By using the νmax of the
adsorbed dye 1b, the corresponding ET(1b) values are also
calculated from eq 1a. The standard ET(30) values are then
calculated from eq 7 according to ref 61.
ET(30) ) 1.05ET(1b) - 6.13
r ) 0.98, n ) 38, sd ) 1.33
(7)
The R and π* values were calculated by a correlation analysis
of the measured absorption maxima νmax (in cm-1) of the dyes
2 and 3, using the corresponding multiple correlations of R or
π* ) f[νmax(2) and νmax(3)] with the Kamlet-Taft solvent
parameters as the reference system.50,54 This procedure is
described in detail in ref 54. In this paper we used an extended
set of absorption maxima of 2 and 3 in various solvents. The
following multiple correlations, [eqs 8 and 9], were used to
separate the respective property R or π* from the unit of
measurement of the νmax(indicator) for 2 and 3.
R ) -7.90 + (0.45νmax(2) × 10-3) + (0.02νmax(3) × 10-3)
(8)
Because of the larger absorption coefficient of dye 2, the
concentration of this dye was about one tenth of that of the
dyes 1a and 1b. As discussed, the specific advantage of dye 2
is due to its sensitivity to gradual changes in the acidity of the
surface environment. For example, two different bare silica
samples can be clearly distinguished in their acidity when 2 is
adsorbed: on SG 432 (No. 5) λmax(2) ) 532 nm and on LC
1500 (No. 6) λmax(2) ) 526 nm. This difference is evident,
because the transmission technique allows a very precise
determination of the UV/vis absorption maxima of adsorbed
r ) 0.95, sd ) 0.17, n ) 34, F (significance) ) 0.00
π* ) 13.89 - (0.251νmax(2) × 10-3) - (0.32νmax(3) ×
10-3) (9)
r ) 0.57, sd ) 0.15, n ) 36, F ) 0.00
The multiple regression analyses were done with the SPSS 61
statistics program.

Surface Polarity of Various Silicas and Other Solid Acids
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6423
like environment. The other one at about λmax(3) ) 500 nm is
not observed with either solid acid. The absorption intensity
ratio of these two bands changes with increasing adsorption time,
whereas the position of the maximum wavelength remains
constant during a period of 30 min. As mentioned, the two
absorption maxima of 3 are attributed to two different chemical
constitutions of 3. The absorption maximum at about λmax(3)
) 500 nm is caused by the oxenium ion of 3, which is produced
by selectively complexing the carbonyl oxygen of 3 with either
a proton or a Lewis acid.58,59 Consequently, Lewis and/or
Brønsted acid sites on the surface can be responsible for the
strong complexation of the carbonyl oxygen of 3. To differenti-
ate between the two options, surface titration of adsorbed 3 with
pyridine and DTBP was carried out.
Typical UV/vis spectra series of those surface titration
measurements are shown for an alumina (No. 34, Figure 3a and
an alumosilicate (No. 23, Figure 3b) sample, respectively.
For the alumina sample it is clearly seen that DTBP does
not affect the position of the UV/vis absorption maximum of 3
at λmax(3) ) 390 nm and the band at about λmax(3) ) 450 nm.
The overall absorption intensity of adsorbed 3 increases with
the quantity of DTBP added. We think that mobile protons are
trapped by DTBP which has bonded to 3 at both dimethylamino
groups. It is well-known that 3 dissolved in concentrated sulfuric
acid does not adsorb in the visible region of the spectrum
because it is triply protonated.50
A surface titration with DTBP of 2 adsorbed on the alumina
particles does not change the UV/vis spectrum. This effect
is contrary to the UV/vis spectroscopic result observed for
silicas.18a However, this result unambiguously shows that 2 is
adsorbed on both the Brønsted and Lewis acid sites on alumina
fairly strongly because an external base does not compete
with 2.
The UV/vis absorption band of 3 adsorbed on alumina at
about λmax(3) ) 500 nm disappears when an excess of pure
pyridine is added (see Figure 3c). Consequently, for alumina,
the UV/vis absorption maximum of adsorbed 3 at λmax(3) )
400 nm and the double band with UV/vis absorption maxima
at λmax(3) ) 450 nm and λmax(3) ) 500 nm are caused by the
Lewis acid sites on the surface. Therefore the generation of the
oxenium ion of 3 is caused by the single coordination of the
carbonyl oxygen atom to the acid site of the surface and the
shoulder at about λmax(3) ) 450 nm by a double complexation
at the carbonyl oxygen and the dimethylamino group.58,59
The situation is different when 3 is adsorbed on an alumo-
silicate which has Brønsted as well as Lewis acid sites. The
UV/vis absorption spectrum of 3 adsorbed on alumosilicate (No.
34) is shown in Figure 3b. Two UV/vis absorption bands are
clearly detectable. An excess of DTBP affects some of the
Brønsted acidic sites as indicated by the apparent hypsochromic
shift of the absorption band at λmax(3) ) 500 nm. But a clean
differentiation between Lewis and Brønsted acidic sites on the
surface of alumosilicates by DTBP titration of adsorbed 3 seems
difficult.
In principle, two values of the R and π* parameters of
aluminas, alumosilicates, or titanium oxides can be determined
using the two different λmax(3) values of adsorbed 3. However,
it must be considered that the correlation equations (eqs 8 and
9) for determining the R and π* values are derived from the
Figure 2. (a) UV/vis reflectance absorption spectra of 1a adsorbed
on alumosilicate (No. 26) 1, alumina (No. 35) 2, Aerosil 300 (No. 2)
3, and silica (No. 6) 4, from a solution in 1,2-dichloroethane. (b) UV/
vis reflectance absorption spectra of 1b adsorbed on anatas (No. 18) 1,
alumina (No. 34) 2, and alumosilicate (No. 19) 3, from a solution in
1,2-dichloroethane.
indicators on silica. The probe 2 can also be used as a surface
polarity indicator for alumosilicates and other quite strong solid
acids where Reichardt’s dyes do not work because they become
protonated. The following important observation should be noted
at this point. During recent years we have investigated more
than 50 different silica samples. The surface polarities change
with storage time (weeks, years) depending on the kind of silica
used. This effect is reproducible and can be measured for each
silica sample like a fingerprint.
Employing the indicator 3, different UV/vis spectroscopic
behavior is observed when it is adsorbed on silica or alumina
and alumosilicates, respectively. On the silicas used only the
main UV/vis absorption band at about λmax(3) ) (395 ( 4) nm
is observed. This corresponds well with the expected value of
νmax(3) for commercial silica samples as reported in ref 37. But
it must be emphasized that in pores of mesoporous siliceous
MCM-41 materials, 3 shows two UV/vis absorption maxima,
one at λmax(3) ) 398 nm and another at λmax(3) ) 510 nm. The
latter disappears after addition of DTBP indicating Brønsted
acidic sites (mobile protons) within the channel of MCM-41.64
UV/vis absorption spectra with two or three significant UV/
vis absorption maxima of the indicator 3 are also observed when
it is adsorbed on aluminas, alumosilicates, or titanium dioxides.
Characteristic UV/vis spectra of 3 adsorbed on an alumina and
an alumosilicate sample are shown in Figure 3, parts a and b.
The UV/vis absorption band of 3 at about λmax(3) ) 370-
400 nm relates well to weak Brønsted acidic sites, i.e., a silanol-
ν
max values of 3, measured in a set of well-behaved protic and
nonprotic solvents in the range of νmax(3) ) 30 000 cm-1 in
n-hexane (nonpolar) to νmax(3) ) 24 960 cm-1 in 1,1,1,3,3,3-
hexafluoro-2-propanol (strongly polar).50 This background has
to be considered using the correlation eqs 8 and 9. Consequently,

6424 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Spange et al.
TABLE 2: UV/Vis Absorption Maxima of 1a and 1b, Respectively, Adsorbed on the Solid Acids from a 1,2-Dichloroethane
N
Solution, the Corresponding ET(30) Values Calculated from Eq 1a, and the Normalized ET Values Calculated from Eq 1b
sample
number
ν
max(1a or 1b)
ET(30) measured
ETN calculated
from eq 1b
solid acid sample
SiO2, Aerosil 200
SiO2, Aerosil 300
SiO2, Aerosil 380
SiO2, KG 60
SiO2, KG SG 432
SiO2, KG LC 1500
SiO2, 1.5% Al2O3 Cogel
AlPO4, SP 8293
AlPO4, SP2-8111.09
SiO2/Al2O3, SP 18-8495
SiO2/TiO2, SP 18-8381.01
SiO2, SP 18-8470
SiO2/ZrO2, SP 18-8402.03
Al2O3. SP 18-8510
SiO2/Al2O3, SP2-8515.01
TiO2 P25
TiO2 (Rutile)
TiO2 (Anatas)
Siral 5
Siral 10
Siral 20
Siral 30
Siral 40
Siral 50
[10-3 cm-1
]
[kcal mol-1] [eq 1a]
1
2
3
4
5
6
7
8
9
20.21
20.20
20.49
20.33
20.24
20.45
22.94
n.d.b
57.8
58.1
58.6
58.1
57.9
58.5
62.7
0.84
0.85
0.86
0.85
0.84
0.86
0.99
n.d.b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
n.d.b
20.00
20.83
n.d.b
57.2
59.6
0.82
0.89
21.01
60.1
63.5
0.91
1.01
n.d.b
23.20a
n.d.b
21.93a
22.97a
n.d.b
59.7
62.8
0.90
0.99
22.30a
n.d.b
60.8
63.5
0.93
1.01
23.20a
n.d.b
Siral 60
Siral 80
Al2O3. Pural
KG 60 10% Al2O3
KG 60 30% Al2O3
KG 60 50% Al2O3
KG 60 70% Al2O3
KG 60 100% Al2O3
Al2O3 (AlEt2Cl)
Al2O3
21.87a
18.94
21.88
22.57a
n.d.b
59.5
54.2
62.6
61.6
0.89
0.73
0.98
0.95
22.43a
n.d.b
61.2
0.94
20.47a
21.73a
21.14a
20.24
55.3
59.1
57.3
57.9
0.76
0.88
0.82
0.84
Al2O3-C
a Measured with 1b because 1a is not suitable. b n.d. Not determinable because the dyes become protonated.
unusually large π* values are obtained for the Lewis acidic sites
of alumosilicates due to the strong bathochromically shifted
π-π* transition of chemisorbed 3. However, the solvatochromic
concept is based on the assumption that chemical alterations of
the probe should not occur when solvatochromic shifts are
interpreted in terms of reversible solvation. Therefore, the
calculated R and π* parameters should be handled with caution
when UV/vis absorption maxima values of chemisorbed indica-
tors are used. Probably this bathochromicity can be similarly
interpreted in terms of halochromic behavior because an ionic
form of 3 is generated.60 This interpretation would even
correspond to large values of the π* term for Lewis acid sites.
However, both possible values of the R and π* parameters have
been determined. They may be valid for the Brønsted and Lewis
acid sites and will be discussed later compared with independent
measurements. Utilizing the second UV/vis absorption band at
Testing the Surface Polarity Parameters by an IR and a
Fluorescence Probe. The π* and R values of the solid acids
were used to calculate theoretical values of the carbonyl
stretching vibration of BP when adsorbed on the solid acids
using the LSE eq 6 from ref 68 (see Introduction). BP adsorption
on solid acids was investigated in particular for a silica (No. 4)
[(νmax,CO,(calculated) ) 1645.5 cm-1 (BS)], for an alumina (No.
34) [(νmax,CO,(calculated) ) 1652 cm-1 (BS) and 1630 cm-1 (LS)],
and for an alumosilicate (No. 23) [νmax,CO(calculated) ) 1644 cm-1
(BS) and 1628 cm-1 (LS)]. The calculated νmax,CO values are
given in parentheses after the sample number; LS denotes the
presumed Lewis and BS the presumed Brønsted sites. However,
BP adsorbed on a solid acid shows two clearly separated IR
bands for its carbonyl stretching vibration in each case. The
following results are obtained: for silica [νmax,CO ) 1656 to
1659 cm-1 and νmax,CO ) 1644 to 1648 cm-1; both positions of
λ
max(3) ) 500 nm of adsorbed 3, the Kamlet-Taft parameters
νmax,CO are a function of the concentration of adsorbed BP (see
Figure 4)], for alumina (νmax,CO ) 1656 cm-1 and νmax,CO
)
are assigned as R2 and π*2.
1636 to 1640 cm-1), and for the alumosilicate (νmax,CO ) 1653
The measured UV/vis absorption maxima of the indicators 2
and 3, R and π* values calculated according to eqs 6 and 7
(the first UV/vis absorption maximum of 3 was used for
calculation), and the calculated ET(30) values for 35 various
solid acids are listed in Table 3.
cm-1 and νmax,CO ) 1643 cm-1).
For silica the two IR bands indicate that BP is adsorbed on
two different surface sites. However, the IR band observed at
1656 to 1659 cm-1 indicates a significantly lower polarity,
similar to that of 1,2-dichloroethane or n-butanol. The inter-
pretation of the IR data in terms of a two-state model in which
adsorbed BP molecules are in a rapid equilibrium between a
non-hydrogen-bonded state in which BP is physisorbed on the
silica surface and a state in which BP molecules are hydrogen
Table 4 contains the νmax,2 values of 3, and the corresponding
theoretically possible R2 and π*2 values. A slightly larger
concentration of 3 was used for these measurements, because
the shoulder at λmax(3) ) 500 nm then appears clearer as it does
in Figure 3.

Surface Polarity of Various Silicas and Other Solid Acids
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6425
TABLE 3: UV/vis Absorption Maxima of the Indicator Dyes
2 and 3 Adsorbed on the Solid Acids from a
1,2-Dichloroethane Solution, Calculated r and π* Values
According to Eqs 8 and 9, and ET(30) Values Calculated
from Eqs 3 and 4, Respectively
ET(30)
[kcal mol-1
calculated
from eq 3
ET(30)
[kcal mol-1
calculated
from eq 4
ν
max(2)
sample [10-3
number cm-1
ν
max(3)
[10-3
cm-1
]
]
]
]
R
π*
1
2
3
4
5
6
7
8
9
18.66 25.45 1.07 1.05
18.80 25.51 1.14 1.00
18.83 26.46 1.15 0.97
18.78 25.25 1.13 1.08
18.80 25.58 1.14 0.98
19.05 25.32 1.25 1.00
20.07 25.19 1.99 0.63
20.08 26.11 1.73 0.49
20.28 26.11 1.82 0.43
19.76 26.46 1.59 0.46
59.5
59.9
59.9
60.8
59.7
61.6
68.7
63.1
63.9
60.7
57.7
58.5
58.5
58.8
58.4
60.2
69.3
64.6
65.8
62.3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
18.52
a
18.55 26.46 1.05 0.76
19.23 25.84 1.34 0.79
19.61 2695
55.8
60.8
58.4
65.8
64.6
63.6
61.5
63.3
64.6
63.5
65.4
64.1
62.3
58.5
59.5
57.1
62.3
63.5
61.4
59.8
54.2
60.6
58.3
54.7
55.9
60.4
60.7
66.3
63.3
63.8
61.5
63.3
64.6
63.5
65.4
65.9
63.0
59.4
59.9
58.8
63.0
63.8
62.2
60.3
57.9
62.9
59.8
55.8
1.53 0.33
20.20 25.45 1.77 0.70
19.57 25.04 1.48 0.95
19.80 25.64 1.60 0.71
19.46 25.81 1.44 0.73
19.96 26.60 1.69 0.36
20.16 26.39 1.77 0.38
19.92 26.25 1.66 0.48
20.24 26.11 1.80 0.45
20.33 26.11 1.84 0.43
19.76 25.97 1.58 0.61
19.23 26.46 1.35 0.59
19.23 26.18 1.35 0.68
19.23 26.88 1.36 0.45
19.76 25.97 1.58 0.61
19.84 25.71 1.61 0.68
19.65 26.11 1.53 0.60
19.29 26.18 1.38 0.66
19.27 27.78 1.40 0.16
19.92 26.67 1.67 0.35
19.35 26.67 1.41 0.49
18.69 26.95 1.12 0.56
Figure 3. (a) UV/vis absorption spectra of 3 adsorbed on alumina
(No. 34); titration with 2,6-di-tert-butylpyridine. (1: without DTBP,
2: 1.41 × 10-4, 3: 2.35 × 10-4, 4: 4.70 × 10-4 mol DTBP, 0.5042
g alumina, 6 × 10-7 mol 3, 10 mL dichloromethane); (b) UV/vis
absorption spectra of 3 adsorbed on alumosilicate (No. 23); titration
with 2,6-di-tert-butylpyridine. (1: without DTBP, 2: 2.35 × 10-4 mol
DTBP; 0.5087 g alumosilicate, 1 × 10-6 mol 3, 10 mL dichlo-
romethane); (c) UV/vis absorption spectra of 3 adsorbed on alumina
(No. 34); titration with pyridine. (1: 0, 2: 1.24 × 10-4, 3: 2.48 ×
10-4, 4: 6.19 × 10-4, 5: 1.12 × 10-3 mol pyridine; 0.5039 g alumina,
8 × 10-7 mol 3, 10 mL dichloromethane).
a The UV/vis absorption band of 3 relating to Brønsted acid sites is
not observed for this sample.
TABLE 4: Second, Long-Wavelength UV/vis Absorption
Maxima of 3 Adsorbed from a 1,2-Dichloroethane Solution
Probably on the Lewis Acidic Sites of the Solid Acid,
Apparent r and π* Values, As Well As Theoretically
Calculated ET(30) Values for These Sites from Eq 12
bonded to silanol groups of the silica surface is in accordance
with 13C and 29Si NMR chemical shift and relaxation data of
the acetone/silica system.79 It is presumed that BP adsorption
occurs similar to acetone adsorption on silica. The agreement
between calculated and measured values for the carbonyl
stretching vibration is quite different. For the silica sample (No.
4), the shoulder at about νmax,CO ) 1644 to 1648 cm-1 agrees
very well with the calculated νmax,CO using the measured polarity
parameters from Table 3 and eq 5.
ET(30)
sample
number
ν
max, 2(3)
[kcal mol-1
]
[10-3 cm-1
]
R2
π*2
calculated from eq 12
10
11
13
14
16
21
22
23
24
25
26
32
34
20.04
20.00
19.80
20.28
20.12
20.49
19.80
20.79
20.35
20.28
20.41
20.41
20.45
1.46
0.90
1.21
1.40
1.37
1.54
1.67
1.73
1.47
1.22
1.23
1.24
1.28
2.51
2.84
2.72
2.47
2.53
2.33
2.47
2.13
2.41
2.57
2.53
2.52
2.49
70.5
65.9
68.8
69.6
69.6
70.5
72.8
71.6
70.1
68.1
68.0
68.1
68.4
For the alumina sample the agreement between calculated
and measured data is poor for one of the two Kamlet-Taft
parameter sets [R(2) and π*(2)] and satisfactory for the other one
[R1 and π*1]. But the difference between the calculated
(∆νmax,CO ) 22 cm-1) and measured (∆νmax,CO ) 20 cm-1
)
νmax,CO values of both parameter sets is rather similar. For the
alumosilicate sample the agreement for one of the two IR bands
is rather good. A precise assignment of this IR band of adsorbed
BP on a specific surface site is not possible.
parameters. But the calculated νmax,CO values of adsorbed BP
are too low (high polarity) compared to the values of the
measured carbonyl stretching vibration (lower polarity). It should
be noted that the concentration of BP was about 10 to 50 times
larger than those of the solvatochromic polarity indicators 2
The IR results for BP adsorption support the order of
magnitude of the values of the determined Kamlet-Taft

6426 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Spange et al.
Figure 5. Measured versus calculated [via eq 12] ET(30) values for
nineteen solid acids including silicas, aluminas, titanium dioxides, and
alumosilicates used in this paper.
Figure 4. The two positions of νmax,CO of benzophenone adsorbed on
silica as a function of the concentration of adsorbed benzophenone.
and π* values derived from the short wavelength UV/vis
absorption band of adsorbed 3 (Table 3) have been used for
the following correlation analyses [eqs 10-12].
For the ET(30) values measured by the two dyes 1a and 1b,
the following multiple correlation equations were obtained.
and 3. Therefore we assume that the solvatochromic dyes
preferentially indicate the most acidic silanol groups and Lewis
acid sites. The use of both values of π* and R together give, in
particular, a reasonable agreement between calculated and
measured data for BP adsorption.
Pyrene has been recommended by several authors for the
detection of the dipolarity/polarizability of an environment.65
For Aerosil 300 and Aerosil 380, Py values of 1.48 and 1.65,
respectively, were estimated via eq 7 using the R and π* values
from Table 3. This calculation shows that the measured Py value
from the literature (Py ) 1.78) is larger than the value to be
expected from eq 7. It may be that pyrene and 3 are adsorbed
on quite different surface sites, because 3 is quite a strongly
dipolar molecule compared to the nonpolar probe pyrene.
ET(30) (via 1a) ) 52.19 + 5.12R + 0.99π* (a/s ) 5.13)
(10)
n ) 8, r ) 0.964, sd ) 0.235, F ) 0.0013
ET(30) (via 1b) ) 35.08 + 13.39R + 8.70π* (a/s ) 1.54)
(11)
n ) 11, r ) 0.922, sd ) 1.110, F ) 0.0005
The respective influence of the R and π* terms upon ET(30),
expressed by the coefficient ratio a/s, is different for the two
dyes. It can be clearly shown that dye 1a serves mainly as an
R indicator for acidic surfaces, whereas dye 1b reflects the R
and π* value of acidic surface environments similarly to the
way that 1a does for solvents. The LSE equation derived for
well-behaved regular solvents reflects the a/s ratio in a way
similar to that in eq 11. Consequently, the basicity of a phenolate
betaine dye influences the susceptibility coefficients of the
Kamlet-Taft relation and accordingly the (XYZ)0 value when
applied to HBD surfaces. This interpretation is in accordance
to the result of the Kamlet-Taft correlation analyses for
structural different solvatochromic dyes.23 In ref 23 has been
shown that the coefficient a of eq 2 increases in decreasing the
pKa-value of the solvatochromic indicator.
Discussion
By comparison with the Kamlet-Taft parameters of repre-
sentative pure solvents, such as 2,2,2-trifluoroethanol (R ) 1.51,
π* ) 0.73) and 1,1,1,3,3,3-hexafluoro-2-propanol (R ) 1.96,
π* ) 0.65), it is evident that the solid surface environments
observed by the various indicators are moderately strongly
dipolar/polarizable (π* ) 0.38 to 1.04) and rather strong
hydrogen-bond donors (R ) 1.00 to 1.99). This shows that the
polarity of the solid acids studied covers only a rather small
section of the Reichardt ET(30) scale. Within the framework of
the Kamlet-Taft R scale, the observed differences are more
significant, e.g., from the silicas with R values of about 1 to
alumosilicates with R values of about 2. Hence, for solid acids
the Kamlet-Taft parameters set allows a better differentiating
than do the ET(30) values.
Correlation Analyses of the Polarity Parameters with
Each Other. Using the independently calculated R and π*
values from Table 3, theoretical ET(30)values were calculated
from eqs 3 and 4 and then compared with the experimentally
obtained data.
These calculated ET(30) values are also listed in Tables 2
and 3. A comparison of the calculated and measured ET(30)
values shows a rather good agreement for 80% of the samples
considered. It seems that calculated ET(30) values obtained from
eq 3 agree better with those data which are measured by means
of dye 1b. Dye 1b is more hydrophobic and less basic than dye
1a. Therefore it is expected that this dye is more sensitive toward
the π* term than to the R term. Furthermore, the measured ET-
(30) values have been correlated with the independently
determined R and π* values. It must be noted that only those R
This reflection shows that fluorine-substituted betaine dyes
are suitable surface polarity indicators for moderately strong
solid acids. Taking into account the sum of all measured
ET(30), R, and π* values, eq 12 is obtained:
ET(30) ) 39.56 + 11.60R + 5.60π* (a/s ) 2.07) (12)
n ) 19, r ) 0.87, sd ) 1.17, F ) 0.0000
Altogether the agreement is rather good between calculated
and measured ET(30) values for the nineteen different silicas,
alumosilicates, aluminas, and titanium dioxide samples consid-
ered. Figure 5 shows the plot of the measured versus calculated
ET(30) values according to eq 12.
One outlying point from the theoretical plot is clearly
detectable: for alumina (No. 35) ET(30)(measured) - ET(30)-
(calculated) ) ∆ET(30) ) 2.2 kcal mol-1. However, this

Surface Polarity of Various Silicas and Other Solid Acids
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6427
Figure 6. Influence of the aluminum content of the alumosilicates
used on the value of the Kamlet-Taft R parameter. The aluminum
content was determined by atomic absorption spectroscopy. The
arbitrarily drawn curve is a guide for the reader.
Figure 7. Influence of the aluminum content of alumosilicates on the
value of the Kamlet-Taft π* parameter. The aluminum content was
determined by atomic absorption spectroscopy.
The surface acidity (R value) of alumosilicates (H form) as
a function of the aluminum or silicon content goes through a
maximum. A theoretical curve is difficult to calculate because
the samples, which are available from the commercial suppliers,
show significant differences concerning their Kamlet-Taft
parameters for each different sample produced. The R value of
each point in Figure 6 could be reproduced readily. We
independently determined the corresponding aluminum content
of each sample by AAS (see experimental part). When using
these data as indicated by the producer, a poorly defined curve
is obtained. Our results are in very good qualitative agreement
with the results reported in ref 27. This is a good indication of
the correctness of the determined values.
outlying point is included in the individual correlation in eq
10. Hence, these correlation analyses show that the set of surface
polarity indicators used is suitable for measuring the surface
polarity parameters ET(30), R, and π* of solid acids. Notice
that for each specific silica sample individual R and π* values
are obtained. The measured ET(30) parameter observed by dye
1a and data calculated by eq 4, using the independently
examined R and π* values, agree very well for silicas. This
result suggests that individual surface properties of specific silica
samples are caused by variations of the R and π* terms.
However, a critical reflection concerning the basic concept is
necessary at this point. If each of the two dyes 1a and 1b reflects
the polarity of a surface in another ratio of the R and π* values,
measured by means of the indicators 2 and 3, then it is not
possible in principle to examine independent R and π* values
of solid acids. In other words, a clean separation of the HBD
term from the π* term and from the unit of measurement is not
really justified. This problem exists for all measurements
employing LSE relations.
Figure 7 shows the plot of the value of the π* term as function
of the aluminum content of the alumosilicates used.
The π* value decreases with increasing the aluminum content.
For pure alumina, an increase in the π* value is observed again.
Assuming that the π* values in Figure 7 are derived for the
Brønsted sites, the following interpretation is possible. The
decrease in the π* value is caused by a decrease in the
polarizability of the alumosilicatic framework. However, the
dependence shown in Figure 7 is evident and shows that the
π* value is also significant and a function of the aluminum
content of the alumosilicate.
Using the R2 and π*2 values, a rather poor agreement between
measured and calculated ET(30) values is found. This is indirect
confirmation of the hypothesis that the dyes 1a and 1b have no
access to Lewis acidic surface sites. Perhaps ET(30) values for
Lewis acid sites should really be about 68-72 kcal mol-1. Funke
and Mayr have shown by a correlation of ET(30) values of
solvents with kinetic data that the ET(30) value for liquid SO2,
a moderately strong Lewis acid, approaches a value of ET(30)
Conclusions
) 68 kcal mol-1 80
. It should be mentioned that the ET(30) value
Reichardt’s ET(30) as well as the Kamlet-Taft’s solvent
parameters R and π* can be used as a reference system to
characterize empirically the surface polarity of moderately strong
solid acids, e.g., silicas, aluminas, alumosilicates, and titanium
dioxides. For this purpose three structurally different types of
solvatochromic UV/vis probes, N-phenolate pyridinium betaine
dyes, a mixed ligand transition iron(II) complex, and an aromatic
aminoketone derivative have been found to be suitable. The
results generally show that the surface of solid acids exhibit
solvent-like behavior and that an average polarity is seemingly
measured by the indicator dyes 1a, 1b, and 2. Experimentally
available ET(30) values of solid acids correlate well with
independently calculated R and π* values compared to the
Marcus equation for well-behaved solvents and a similar LSE
relationship derived for functionalized silicas. Because the
coefficients of the Kamlet-Taft LSE equation for the ET(30)
value of solid acids depend on the structure of the indicator
used, care must be taken in interpreting the individual parameters
R and π* of solid acid surfaces. Nevertheless, the Kamlet-
Taft parameters must be handled with caution in order to be
mathematically orthogonal, and they neglect any other higher-
order term.
of liquid SO2 cannot be measured directly using the dye 1a.81
Therefore, we believe that the reported polarity data in Tables
2 and 3 are probably representative only for the Brønsted acidic
sites of solid acids. It is conceivable that the phenolate betaine
dyes have no direct access to Lewis acid sites because the phenyl
rings in the 2- and 6-positions of the phenoxide ring prevent
this interaction.
Correlation of the Kamlet-Taft Parameters of Surfaces
with Structural Data. The variations of the value of the R or
π* term of the silicas are obviously caused by different
molecular surface structures (geminal silanols, isolated silanols,
associated silanols, siloxane bridges, etc.). A classification of
changes in the R or π* as a result of structure modification is
still not possible, because the DRIFT spectra as well as the solid
state 1H NMR and 29Si NMR spectra of the silica samples (Nos.
1-6) are virtually identical.53 This result shows that the surface
polarity indicators employed are more sensitive for the detection
of gradual modifications of the surface polarity than are DRIFT
or solid-state NMR spectroscopy. This topic is still under study
by our group.
Figure 6 shows the plot of the influence of the aluminum
content on the R term of the alumosilicates used in this paper.
The arbitrarily drawn curve is a guide for the reader.
Further research is necessary to evaluate exactly the individual
parameters R and π* for the Lewis as well as Brønsted acidic

6428 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Spange et al.
(40) Rutan, S. C.; Carr, P. W.; Taft, R. W. J. Phys. Chem. 1989, 93,
4292-4297.
(41) Helburn, R. S.; Rutan, S. C.; Pompano, J.; Mitchern, D.; Patterson,
W. T. Anal. Chem. 1994, 66, 610-618.
sites. For this purpose genuine solvatochromic probe dyes of
low basicity and smaller size are required.
Acknowledgment. Financial support, in particular by the
Deutsche Forschungsgemeinschaft and the Fonds der Chemis-
chen Industrie, is gratefully acknowledged. We thank especially
Prof. C. Reichardt, University of Marburg, for generously
providing the eicosafluorine-substituted betaine dye 1b, for
presentation of unpublished results, and for helpful discussions.
(42) Park, J. H.; Carr P. W. J. Chromatogr. 1989, 465, 123-136.
(43) (a) Ramamurthy, V. In Surface Photochem.; Anpo, M., Ed.;
Wiley: New York, 1996; pp 65-115. (b) Ramamurthy, V.; Eaton, D. F.
In Proceedings of the 9th International Zeolite Conference, von Ballmoos,
R., Higgins, J. B., Treacy, M. M. J., Eds.; Butterworths-Heinemann: Boston,
1992, 587.
(44) Kamlet, M. J.; Taft, R. W. Acta Chem. Scand. 1985, B39, 611-
628.
(45) Effenberger, F.; Wu¨rthner, F. Angew. Chem. 1993, 106, 742-743;
Angew. Chem. Int. Ed. Engl. 1993, 32, 719-720.
(46) Schneider, H.; Migron, Y.; Marcus, Y. Z. Phys. Chem. (Frankfurt)
NF 1992, 175, 145-152.
(47) Lagalante, A. F.; Jacobson, R. J.; Bruno, T. J. J. Org. Chem. 1996,
61, 6404-6406.
(48) Soukup, R. W.; Schmid, W. J. Chem. Educ. 1985, 62, 459-462.
(49) Migron, Y.; Marcus, Y. J. Phys. Org. Chem. 1991, 4, 310-314.
(50) Spange, S.; Keutel, D. Justus Liebigs Ann. Chem. 1992, 423-428.
(51) Burgess, J. Spectrochim. Acta 1970, 26A, 1957-1962.
(52) Spange, S.; Heinze, Th.; Klemm, D. Polymer Bull. 1992, 28, 697-
702.
(53) Spange, S.; Graeser, A.; Eschner, M.; Steiner, M.; Jaeger, C.
Unpublished results.
(54) Spange, S.; Keutel, D.; Simon, F. J. Chim. Phys. 1992, 89, 1615-
1622.
(55) Spange, S.; Reuter, A.; Vilsmeier, E.; Keutel, D.; Heinze, Th.;
Linert, W. J. Polym. Sci. 1998, 36, 1945-1955.
(56) Groenen, E. J. J.; Koelman, W. N. J. Chem. Soc., Faraday Trans.
2 1979, 75, 85.
References and Notes
(1) Corma, A. Chem. ReV. 1995, 95, 559-614.
(2) Olah, G. A. In Acidity and Basicity of Solids, Theory, Assessment
and Utility, Fraissard, J., Petrakis, L., Eds.; 1994, Nato ASI Ser., Kluwer
Academic: Dortrecht, 1994; 444, p 305.
(3) Drago, R. S.; Dias, S. C.; Torrealbe, M.; de Lima, I. J. Am. Chem.
Soc. 1997, 119, 4444-4452.
(4) Haw, J. B.; Nicholas, J. B.; Xu, T.; Beck, L. W.; Ferguson, D. B.
Acc. Chem. Res. 1996, 29, 259-267.
(5) Song, X.; Sayari, A. Catal. ReV. Sci. Eng. 1996, 38, 329-412.
(6) Heeribout, L.; Semmer, V.; Batamack, P.; Dore`mieux-Morin, C.;
Vincent, R.; Fraissard, J. Stud. Surf. Sci. Catal. 1996, 101, 831-840.
(7) Chronister, C.; Drago, R. S. J. Am. Chem. Soc. 1993, 115, 4793-
4798.
(8) Maciel, G. E.; Bronnimann, C. E.; Zeigler, R. S.; Ssuer Chuang,
J.; Kinney, D. R.; Keiter, E. A. In The Colloid Chemistry of Silica, Bergna,
H. E., Ed.; American Chemical Society, Washington, DC, 1994, pp 260-
278.
(9) Harmer, M. A.; Sun, Q.; Michalczyk, M. J.; Yang, Z. J. Chem.
Soc., Chem. Commun. 1997, 1803-1804.
(57) Suppan, P. J. Photochem. 1982, 18, 289.
(58) Sekuur, T. J.; Kranenburg, P. Spectrochim Acta 1973, 29A, 807-
811.
(59) Spange, S.; Vilsmeier, E.; Adolph, S.; Fa¨hrmann, A. J. Phys. Org.
Chem. 1999, 12, 547-556.
(10) Bergna, H. E. In The Colloid Chemistry of Silica; American
Chemical Society, Washington, DC, 1994. Bergna, H. E., Ed.; p 1, and
refs therein. AdV. Chem. Ser. 1-47.
(11) Benesi, H. A. J. Am. Chem. Soc. 1956, 78, 5490-5494.
(12) Forni, L. Catal. ReV. 1973, 8, 65-115.
(60) Werts, M. H. V.; Duin, M. H.; Hofstraat, J. W.; Verhoeven, J. W.
J. Chem. Soc., Chem. Commun. 1999, 799-800.
(61) (a) Reichardt, C.; Eschner, M. Unpublished results communicated
to the editors, see Eschner, M. Ph.D. Thesis, University Marburg, 1992.
(b) Reichardt, C.; Asharin-Fard, S.; Blum, A.; Eschner, M.; Mehranpour,
A.-M.; Milart, P.; Niem, T.; Scha¨fer, G.; Wilk, M. Pure Appl. Chem. 1993,
65, 2593-2601.
(62) (a) Drago, R. S.; Dias, J. A.; Maier, T. O. J. Am. Chem. Soc. 1997,
119, 7702-7710. (b) Drago, R. S.; Petrosius, S. C.; Chronister, C. W. Inorg.
Chem. 1994, 33, 367-372.
(63) Lee, C.; Parrillo, D. J.; Gorte, R. J.; Farneth, W. E. J. Am. Chem.
Soc. 1993, 118, 3262-3268.
(64) Spange, S.; Graeser, A.; Zimmermann, Y. Chem. Mater. 1999, 11,
3245-3251.
(65) Gandini, A.; Martinez, A. Macromol. Chem. Symp. 1988, 13/14,
211-234.
(66) Storey, R. F.; Choate, K. R. Macromol. Symp. 1995, 95, 71-78.
(67) Barson, F.; Karam, A. R.; Parent, M. A.; Baird, M. C. Macromol-
ecules 1998, 31, 8439-8447.
(68) Garcia, M. V.; Redondo, M. I. Spectrochim. Acta, Part A 1987,
43, 879-885.
(69) Dong, D. C.; Winnik, M. A. Can. J. Chem. 1984, 62, 2560-2565.
(70) Krasnansky, R.; Thomas, J. K. In The Colloid Chemistry of Silica;
Bergna, H. E., Ed.; American Chemical Society, Washington, DC, 1994; p
223.
(71) Baker, G. A.; Jordan, J. D.; Bright, F. V. J. Sol-Gel Sci. Technol.
1998, 11, 43-54.
(72) Spange, S.; Reuter, A.; Linert, W. Langmuir 1998, 14, 3479-3483.
(73) Spange, S.; Reuter, A.; Prause, S.; Bellmann, C. J. Adhesion Sci.
Technol. 2000, 14, 399-414.
(74) Adolph, S.; Spange, S.; Zimmermann, Y. J. Phys. Chem. B., in
press.
(75) Schilt, A. A. J. Am. Chem. Soc. 1960, 82, 3000-3004.
(76) Baaz, M.; Gutmann, V.; Kunze, O. Monatsh. Chem. 1962, 93,
1162-1175; 1142-1161.
(77) Bentley, A.; Evans, A. G.; Halprem, J. Trans. Faraday Soc. 1951,
47, 711-716.
(78) Leftin, H. P. Carbonium Ions, Olah, P. v. R. Schleyer, Eds.;
Wiley: New York, 1988, 353.
(79) Pan, V. H.; Tao, T.; Zhou, J.-W.; Maciel, G. E. J. Phys. Chem. B
1999, 103, 6930-6943.
(13) Umanski, P.; Engelhardt, J.; Hall, W. K. J. Catal. 1991, 127, 128-
140.
(14) Fargasiu, D.; Ghenciu, A.; Li, J. Q. J. Catal. 1996, 158, 116-127.
(15) Jensen, W. B. In Acid-Base-Interactions; Mittal, K. L., Anderson,
H. R., Eds.; VSP: Utrecht, 1991, pp 3-23.
(16) Mu¨ller, P. Pure Appl. Chem. 1994, 66, 1077-1079.
(17) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358.
(18) Marcus, Y. Chem. Soc. ReV. 1993, 409-416.
(19) Gutmann, V. Coord. Chem. ReV. 1976, 18, 225-240.
(20) Liptay, W. Naturforsch. 1965, 20a, 1441-1471.
(21) Suppan, P. J. Photochem. Photobiol. A 1990, 50, 293-330.
(22) Reichardt, C. SolVents and SolVent effects in Organic Chemistry,
2nd ed.; VCH: Weinheim, 1988, and refs therein.
(23) Novaki, L. P.; El Seoud, O. A. Ber. Bunsen-Ges. Phys. Chem. 1996,
100, 648-655.
(24) Palm, N.; Palm, V. Org. React. (Tartu) 1997, 104, 141-158.
(25) Lindley, S. M.; Flowers, G. C.; Leffler, J. E. J. Org. Chem. 1985,
50, 607-610.
(26) Handreck, G. P.; Smith, T. D. J. Chem. Soc., Faraday Trans. 1
1988, 84, 1847-1852.
(27) Dutta, P. K.; Turbeville W. J. Phys. Chem. 1991, 95, 4087-4092.
(28) Rutan, S. C.; Harris, J. M. J. Chromatogr. A 1993, 656, 197-215.
(29) Spange, S.; Reuter, A.; Schramm, A.; Reichardt, C. Organic
ReactiVity (Tartu) 1995, 29, 951-952.
(30) (a) Spange, S.; Simon, F.; Heublein, G.; Jacobasch, H.-J.; Bo¨rner,
M. Colloid Polym. Sci. 1991, 269, 173-178. (b) Spange, S.; Reuter A.;
Vilsmeier, E. Colloid Polym. Sci. 1996, 274, 59-69.
(31) Taverner, S. J.; Clark, J. H.; Gray, G. W.; Heath, P. A.; Macquarrir,
D. J. Chem. Soc., Chem. Commun. 1997, 1147-1148.
(32) Michels, J. J.; Dorsey, J. G. Langmuir 1990, 6, 414-419.
(33) Dimroth, K.; Reichardt, C.; Siepmann, T.; Bohlmann, F. Justus
Liebigs Ann. Chem. 1963, 661, 1-37.
(34) Reichardt, C.; Harbusch-Go¨rnert, E. Justus Liebigs Ann. Chem.
1983, 721-743.
(35) Spange. S.; Lauterbach, M.; Gyra, A. K.; Reichardt, C. Justus
Liebigs Ann. Chem. 1991, 323-329.
(36) (a) Spange, S.; Reuter, A. Langmuir 1999, 15, 141-150. (b) Spange,
S.; Reuter, A.; Lubda, D. Langmuir 1999, 15, 2103-2111.
(37) (a) Kamlet, M. J. D.; Abboud, J.-L. M.; Abraham, M. H.; Taft, R.
W. J. Org. Chem. 1983, 48, 2877-2887. (b) Taft, R. W.; Kamlet, M. J. J.
Chem. Soc., Perkin Trans. 2 1979, 1723-1727.
(38) Marcus, Y. J. Solution Chem. 1991, 20, 929-944.
(39) Brune, B. J.; Payne, G. F.; Chaubal, M. V. Langmuir 1997, 13,
5766-5769.
(80) Funke, M.; Mayr, H. Eur. J. Chem. 1997, 3, 1214-1222.
(81) Reichardt, C. Unpublished results, communicated to the author.