A TPD-MS study of the interaction of coumarins and their heterocyclic derivatives with a surface of fumed silica and nanosized oxides CeO 2/SiO2, TiO2/SiO2, Al 2O3/SiO2

Article abstract of DOI:10.1002/jms.1765

The interactions between coumarins and the surface of fumed SiO ₂, CeO₂/SiO₂, TiO₂/SiO₂ and Al₂O₃/SiO₂ were assessed by means of temperature-programed desorption mass spectrometry. The different stages of the thermolysis of coumarin were identified and an analysis of the underlying reactions was performed. The kinetic parameters of the involved reactions were thus obtained. The decomposition of thiazolyl-substituted coumarins was found to proceed through a 'thiazole-thiazine' ring expansion in the adsorbed state. A linear correlation between the sigma constants (Σσ) of the coumarin substituents and the activation energy of CO₂ formation was obtained. Copyright

Full text of DOI:10.1002/jms.1765

Research Article
Received: 09 December 2009
Accepted: 09 May 2010
Published online in Wiley Interscience: 8 June 2010
(www.interscience.com) DOI 10.1002/jms.1765
A TPD-MS study of the interaction of coumarins
and their heterocyclic derivatives with a
surface of fumed silica and nanosized oxides
CeO2/SiO2, TiO2/SiO2, Al2O3/SiO2
Kostiantyn Kulyk,a Valentyna Ishchenko,b Borys Palyanytsya,c
Volodymyr Khylya,b Mykola Borysenkoa and Tetiana Kulykc∗
The interactions between coumarins and the surface of fumed SiO2, CeO2/SiO2, TiO2/SiO2 and Al2O3/SiO2 were assessed by
means of temperature-programed desorption mass spectrometry. The different stages of the thermolysis of coumarin were
identified and an analysis of the underlying reactions was performed. The kinetic parameters of the involved reactions were
thus obtained. The decomposition of thiazolyl-substituted coumarins was found to proceed through a ‘thiazole–thiazine’ ring
expansion in the adsorbed state. A linear correlation between the sigma constants (ꢀσ) of the coumarin substituents and the
c
activation energy of CO2 formation was obtained. Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Keywords: TPD-MS; coumarins; fumed silica; nanosized oxides; kinetic parameters
Introduction
related sets of other reactions which differ from the first one
by one-type changes in the reagent structures and experimental
conditions. As it has been shown in previous works,[16–18] these
correlations allow to describe quantitatively the changes in the
reaction medium, substrate structure (Hammett equation[19,20]
and Taft equation[21,22]) or reagent structure (Brønsted catalysis
equation[23]). Such information could be obtained from the
analysis of TPD-MS data. A review by Kraus[24] is specially devoted
to the description of linear correlations of substrate reactivity in
heterogeneous catalytic reactions.
With respect to the chosen adsorbate, their importance cannot
be overemphasized. Coumarins are compounds derived from
α-pyrone and are important biologically active phytogenous
substances. A great value is specially attributed to synthetic,
modified coumarins in which the nucleus is modified with
heterocyclic groups. These groups are chosen for their biological
activity and the final products find wide use as medicines[15]
because of their good balance of high biological activity and
low toxicity. This activity comprises many practical medicine
The study of the thermal transformations of coumarins is of
practical value due to the widespread application of these
compounds as matrixes in matrix-assisted laser desorption and
ionization mass spectrometry (MALDI MS). The derivatization
properties of coumarins are intimately related to their acid–base
properties, particularly to their ability to form hydrogen-bonded
and donor–acceptor complexes with analyte molecules that have
proton-donating or electron-donating groups. Another important
though less reported aspect of the effectiveness of coumarins
as MALDI MS matrixes is their pattern of thermal transformation.
Some studies have been published[1–4] and models of ionization
have been suggested such as those of disproportionational,
polyphotonic and thermal ionization. It seems however evident
that the mechanisms of ionization of molecules during MALDI MS
tests have not been fully elucidated and need a further deep study
in order to develop fundamental models. Therefore investigation
of mechanisms of thermal reactions of coumarins by means of
temperature-programed desorption mass spectrometry (TPD-MS)
could be useful for the identification of the ionization mechanisms
inMALDIMSandinsurface-assistedlaserdesorptionandionization
massspectrometry(SALDIMS).[5] TPD-MStechniqueiswidelyused
in surface chemistry for the study of adsorption and desorption
processes, surface reactions and surface properties of various
objects.[6–14]
∗
Correspondenceto:Tetiana Kulyk,DepartmentofAmourphousandStructurally
Ordered Oxides, Chuiko Institute of Surface Chemistry of National Academy
of Sciences of Ukraine, 17 Generala Naumova Street, Kyiv 03164, Ukraine.
E-mail: tanyakulyk@gala.net
a
Department of Oxide Nanocomposites, Chuiko Institute of Surface Chemistry
of National Academy of Sciences of Ukraine, 17 Generala Naumova Street, Kyiv
03164, Ukraine
Particularly a methodology should be developed to link
empirical kinetic data such as global reaction rates, energies
and entropies of activation to fundamental molecular structure
features. An appropriate one seems to be the building of
quantitative structure–activity relationships based on the linear
free energy relationship (LFER) principle.[15] The LFER principle
is based on the linear correlation of the logarithm of the rate
constant (or equilibrium constant) of one reaction with the
b
c
Department of Organic Chemistry, Taras Shevchenko Kyiv National University,
64 Volodymyrska Street, Kyiv 01033, Ukraine
Department of Amourphous and Structurally Ordered Oxides, Chuiko Institute
of Surface Chemistry of National Academy of Sciences of Ukraine, 17 Generala
Naumova Street, Kyiv 03164, Ukraine
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.

TPD-MS of silica-supported coumarins
therapeutic effects such as blood anticoagulation, capillary tubes
strengthening, hypotension, antispasmodic, anti-inflammatory
and cholagogue effects.[25,26]
Al2O3, amount of aluminum atoms in surface layer 6.2%) and
fumed TiO2/SiO2 (Sa = 156 m2g−1, 14% TiO2, amount of titanium
atoms in surface layer 7.8%) were supplied by the Research
and Experimental plant of the Chuiko Institute of Surface
The specific interest on the coumarin–silica system follows the
venue on modern medical drugs. The development of complex
drugsbasedonbiologicallyactivecompoundsandnanosizedsilica
has drawn much interest because of their demonstrated longer
therapeutic effect.[27] For example, some derivatives of another
natural oxygen heterocycles γ -pyrones and chromones[28–30] and
biologically active substituted cinnamic acids have been lately
studied.[31,32] Despite the importance of the gathered empirical
evidence, the development of complex drugs demands more
detailed information about the interactions in the adsorbed
state, the structure of surface complexes and their mechanism of
formation. Such information could be obtained from the analysis
of TPD-MS data.[31,32] Of special interest would be the application
of this technique to the elucidation of the binding mechanisms of
coumarins on the silica surface.
Chemistry (Kalush, Ukraine). Nanocomposite CeO2/SiO2 (Sa
=
265 m2g−1, 6.6% CeO2) was prepared using powdery fumed silica
and acetylacetonate of cerium [CH3COCH C(CH3)O]3Ce·H2O
(Aldrich). Modification of silica with cerium acetylacetonate was
performed by a liquid-phase method (with carbon tetrachloride
as a solvent).[33]
Coumarins 1–4 and 4-methylthiazole were supplied by Fluka.
Coumarins 5–7 were synthesized at the Department of Organic
Chemistry of Taras Shevchenko Kyiv National University. NMR
spectral data. Coumarine 5. (δ, ppm) 9.15–4-H (1H, s); 7.1–7.6–5-H
(1H, m); 7.1–7.6–6-H (1H, m); 7.1–7.6–7-H (1H, m); 10.35–8H (1H,
broadened s, OH). δ H (Het) 7.52 (2H, m, 4-H, 7-H), 7.20 (2H, m,
5-H, 6-H). Coumarine 6. 8.85–4-H (1H, s); 7.81–5-H (1H, d, J3 = 8);
6.64–6-H (1H, dd, J3 = 8, J4 = 2.5); 10.91–7-H (1H, broadened
s, OH); 6.83–8H (1H, d, J4 = 2.5). δ H (Het) 2.44 (3H, s, Me), 7.36
(1H, s, 5-H thiazole). Coumarine 7. 8.73–4-H (1H, s); 6.80–5-H (1H,
d, J3 = 9); 7.44–6-H (1H, d, J3 = 9); 4.23 – (8-CH2); 1.92 (4H, m,
2CH2); 2.77 (4H, m, 2CH2N). δ H (Het) 2.52 (3H, s, Me), 7.01 (1H,
s, 5-H thiazole). Samples of silica (SiO2, CeO2/SiO2, Al2O3/SiO2
and TiO2/SiO2) supported coumarin with loadings of 0.3, 0.4 and
0.6 mmol g−1 were obtained by impregnation. In a typical sample
preparation session a portion (1 g) of fumed silica was gradually
mixed with ethanol solution (25 ml) of coumarin. The sample was
then aged at room temperature for 24 h and dried on air.
This work focuses on the study of some of the most used
coumarin drugs, namely the hydroxy- and methoxy-substituted
coumarins and the synthetic thiazole and benzothiazole deriva-
tives of α-pyrone.
Experimental
Materials
Powdery fumed silica (Sa = 300 m2g−1) with a specific surface
area of 300 m2g−1, fumed Al2O3/SiO2 (Sa = 207 m2g−1, 1.2%
Temperature-programed desorption mass spectrometry
CH3
TPD-MS experiments were performed in an MKh-7304A monopole
mass spectrometer (Electron, Sumy, Ukraine) with electron impact
ionization,adaptedforthermodesorptionmeasurements.Atypical
test comprised placing a 20 mg sample on the bottom of a
molybdenum-quartz ampoule (5.4 mm in diameter, a length ◦of
20 cm and the volume 12 ml), evacuating to ∼5·10−5 Pa at ∼20 C
andthenheatingat0.15 ◦C s−1 fromroomtemperatureto∼750 ◦C.
For all the samples, the sample vials were filled approximately
1/16 full, what helped (as it was shown in the works[8,10]) to limit
interparticle-diffusion effects. The works[8,10,34–36] demonstrate
that limiting sample volume along with the high vacuum should
further limit readsorption and diffusion resistance. The volatile
pyrolysis products passed through a high-vacuum valve into
the ionization chamber of the mass spectrometer where they
were ionized and fragmented by electron impact. After mass
separation in the mass analyzer, the ion current due to desorption
and thermolysis was amplified with a VEU-6 secondary-electron
multiplier. The mass spectra were recorded and analyzed using
a computer-based data acquisition and processing setup. The
mass spectra were recorded within 1–210 amu. During each
TPD-MS experiment, ∼240 spectra were recorded. During a
thermodesorptionexperimentthesamplewasheatedslowlywhile
keeping ahighrateof evacuationof thevolatilepyrolysisproducts.
Diffusion effects can thus be neglected and the intensity of the ion
current can be considered proportional to the desorption rate.
The possible existence of mass transfer resistances was assessed
by calculating the Thiele modulus at the conditions of maximum
desorption rate. Values of the specific surface area (Sg), the
pore volume (Vg), average pore radius (rp) and the particle
diameter (dp) of the powdery fumed silicas used in this work
have been previously reported in several works by Gun’ko
O
O
HO
HO
O
O
1
2
CH3
O
O
HO
CH3O
O
O
O
CH3
3
4
CH3
N
N
S
S
HO
O
O
O
O
OH
5
6
CH3
N
H
S
O
O
HO
N
7
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

K. Kulyk et al.
et al.:[37,38] Sg = 294–300 m2 g−1, Vg = 0.44–0.57 cm3 g−1
,
silica surface were calculated using the personally designed com-
puter program. This program enables the calculation of various
nonisothermal kinetic parameters of the reactions in the con-
densed phase and on the surface of highly dispersed oxides, such
as the order of the reaction (n), the pre-exponential factor (ν),
activation energy (E=), and activation entropy (dS=) from TPD-MS
data. The nonisothermal parameters were calculated only for well-
resolved peaks, the peaks for which the form and position on the
temperature scale were reproduced in several experiments. These
nonisothermal parameters were calculated from the traces of well-
resolved peaks for which form and position on the temperature
scale were repeated in several experiments.
rp = 3.0–3.2 nm, dp = 15 µm). Due to the low pressure used
in the experiment only wall-molecule collisions were assumed to
be present and the diffusivity was considered to be of the Knudsen
type. The classical formulas for the Thiele modulus (h) and the
Knudsen diffusivity (DK) were taken from Carberry’s textbook.[39]
The effective diffusivity inside the pores was calculated as the
product of DK by the particle porosity (ε = 0.565), disregarding
the effect of tortuosity and restriction factors. The kinetic constant
in Thiele’s formula was taken equal to kd after assuming that
desorption is the rate-limiting step in the mechanism. Then at the
temperatureofmaximumrateofdesorptionofcoumarinsfromthe
silicasurface(≈700 K)thevaluesfoundfortherelevantparameters
are: DK = 5 × 10−4 cm2 s−1; kd = 0.0045 s−1; h = 0.0057. As it
can be seen the reacting system is in the kinetic regime (h ꢁ 0.4),
i.e. mass diffusional constraints are absent.
The procedure of obtaining kinetic parameters from TPD-MS
data has been extensively described in a number of works and
reviews.[6–12,29,40–42] The Arrhenius plot method was used in
this work. This method is conveniently described in full detail
elsewhere[29,42] but a short description is written in the following
paragraphs.
Temperature-programed desorption mass spectrometry
and kinetic parameters
The process of desorption from solid surface can be described
bykineticEqn (1), andthedesorptionrateconstantkd byArrhenius
The nonisothermal parameters of desorption (thermal decompo-
sition) and chemical reactions in the condensed phase and on
Figure 1. Mass spectrum at 470 ◦C (a) and thermal desorption traces of the ions with m/z = 44 and 78 (b), corresponding to the decomposition of
silica-supported umbelliferone (1) (0.3 mmol g−1).
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2010, 45, 750–761

TPD-MS of silica-supported coumarins
Eqn (2). Substitution of Eqn (1) into Eqn (2) yields Eqn (3):
where kb is Boltzmann’s constant, NA is Avogadro’s constant, h is
Planck’s constant and Tmax is the temperature of the maximum
desorption rate.
dθ/dt = −kd θn
(1)
(2)
(3)
kd = ν exp(−E=/RT)
dθ/dt = θn ν exp(−E=/RT)
Results and Discussion
where θ is the surface coverage (θ = 0–1), t is time (s), kd is
the desorption rate constant, n is the order of the reaction, ν is
the pre-exponential factor (1/s), E= is the activation energy for
desorption (J/mol), R is the universal gas constant (J/mol K) and T
is temperature (K).
If the E= and ν rates do not depend on the temperature or
surface concentration, the desorption order can be obtained by
plotting ln(θ−n dθ/dt) as a function of 1/RT. From this plot the
value of the reaction order, n, can be obtained, which is equal
to either 1 or 2. The true value is the one that gives a linear
dependence with a slope −E= (tgα = −E=) and with a high
correlation coefficient (R2 = 0.95–1). The substitution of E= into
Eqn (3) permits obtaining ν, for ln(ν) = E=/RT at ln(kd) = 0.
The value of the entropy of activation dS= can be obtained with
Eqn (4):
The study of coumarins (1–7) in bulk condensed form has shown
that in this state they possess high thermal stability. When heated
in a vacuum they sublimate and form a thin layer over the cold
areas of the mass spectrometer ampoule. For this reason, no
mass spectra of bulk coumarins could be obtained in the TPD-MS
experiments. In order to obtain such spectra, the direct injection
or field desorption methods should be used.
The thermolysis data of the silica-supported coumarins (1–7)
showedthepresenceofamoleculariononlyinthecaseofthemass
spectrum of the methoxy-derivative 4. The other silica-supported
coumarins were seemingly subjected to thermal degradation
because their mass spectra lacked the molecular ion in the whole
scanned temperature range.
The first step of the thermal decomposition of the silica-
supported coumarins (1–3) and their heterocyclic derivatives with
free hydroxyl groups (5–7) is decarboxylation. The TPD traces of
dS= = kb NA ln(h ν/kb Tmax
)
(4)
Figure 2. Mass spectrum at 391 ◦C (a) and thermal desorption traces of the ions with m/z = 44, 34, 78, 135 and 149 (b), corresponding to the thermal
decomposition of silica-supported 3-(2-benzothiazol)-8-hydroxycoumarine (6) (0.4 mmol g−1).
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

K. Kulyk et al.
Figure 3. Mass spectrum at 350 ◦C (a) and thermal desorption traces of the ions with m/z = 176, 148 and 133 (b), corresponding to the desorption of
silica-supported 7-methoxycoumarin (4) (0.3 mmol g−1).
O
MeO
O
176 Da
-
e
.
+
(m/z 176)
M
+
.
[M - CO]
(m/z 148)
(m/z 133)
(m/z 105)
.
+
[M - CO - CH3]
[M - CO2 - CO + H]
.
+
Scheme 1. EI mass fragmentation of 7-methoxy-coumarin (4).
200 ◦C, a molecular ion with m/z = 176 appears along with
its rearrangement counterparts of m/z = 148, 133 and 105.[43]
These are formed in the ion source of the mass spectrometer
under the influence of electrons (Scheme 1).
Figure 4. Correlation between the activation energy of decarboxylation of
silica-supportedcoumarinsandtheconstantsofthecoumarinsubstituents.
theionwithm/z 44(CO2)haveanidenticalshapeandtheirmaxima
occur at the same temperature (Figs 1 and 2).
Unlike1–3and5–7coumarins,theCO2 TPDtraceof7-methoxy-
coumarin (4) has no distinguishable peaks above 100 ◦C (Fig. 3).
7-methoxy-coumarin sublimates easily from the silica surface.
The other spectra indicate that at temperatures higher than
This great difference in surface thermolysis patterns is caused
by the change of of the –OH group to a –OCH3 group. The
presence and number of adsorbate OH groups play a crucial role
in the adsorption over polar silica surfaces. Coumarins could be
chemisorbed over silica in several ways and the a priori two most
feasible are depicted in Scheme 2(a) and (b). The transition from
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2010, 45, 750–761

TPD-MS of silica-supported coumarins
O
O
O
H
O
O
O
H
O
t °C
H
O
H
O
O
Si
Si
Si
Si
Si
Si
O
H
O
OH
Si
O
O
Si
Si
(a)
O
O
O
O
O
O
- H O
2
H
H
H
H
O
O
O
O
O
Si
Si
Si
(b)
Scheme 2. (a) The chemisorption and subsequent protonation of oxygen in pyran-cycle carbonyls of 7-hydroxy-coumarin (1). (b) The chemisorption of
7-hydroxy-coumarin (1) on the surface silanol group.
physical adsorption to chemisorption would occur by hydroxyl-
silanol condensation and water evolution during the heating of
the sample ampoule in the TPD test.
approximate formula:[40–42]
E= = RTmax ln(B ln B), where (B = n ν Tmax Cmaxn−1/b)
(5)
The chemisorption and subsequent protonation of oxygen in
pyran-cycle carbonyls or of nitrogen in thiazole and benzothiazole
substituents would occur by proton donation from a silanol group.
This protonation produces a redistribution of the electron density
of coumarins. As a result, coumarin is degraded during the
temperature increase of the TPD test. The weakest bonds are
broken and carbon dioxide is formed. This is not seen in the case
of the TPD test of bulk coumarin. For this reason, hydroxylated
coumarins cannot desorb from the surface in their molecular form
while methoxylated ones (4) can.
In this formula ν = νaverage = 1.47 · 104 s−1, n is the reaction
n−1
order, Cmax
is the concentration of the adsorbate at Tmax and
b is the value of the sample heating rate. The thus calculated
value E= has only a slight dependence on the B parameter.[40] It
has been shown in the work of Redhead[40] that the logarithmic
dependence of the activation energy on ν helps to decrease the
error sensitivity of the analysis results. Thus reliable estimates of
the activation energy can be obtained. For this reason for the
first-order model and νaverage = 1.47 × 104 s−1, the following
approximate correlation can be used:
The kinetic parameters of the first- and second-order decar-
boxylation reaction were calculated (Table 1). The results indicate
that the reaction is well described by a first-order model because
in this case the correlation coefficient is the highest. The pre-
exponential factors of all the seven samples studied were similar
and had the same magnitude order, νaverage = 1.47 · 104 s−1. This
is an indication that the structure of the transient state of the
decarboxylation reaction for this series of coumarins is the same.
In this case, the isoentropic series is studied. If we define Tmax as
the temperature at which the decarboxylation rate is maximum,
we can see that Tmax spans a range of about 100 ◦C for the set
of coumarins and has the following order: 3: 502 > 1: 481 > 2:
475 > 7:458 > 5:427 > 6:408 ◦C. DecarboxylationratesandTmax
seem therefore to be the features most affected by the presence of
different substituents in the coumarin molecule. The lowest values
of Tmax are displayed by the heterocyclic coumarin derivatives (5
and 6). However a regular change in activation energy is not found
in this studied series. This is because the energy of activation was
calculated for the entire maximum, and the rated value of Eact
was greatly influenced by the form of the front or back edges. For
this reason, the energy of activation was calculated according to
E= = 15 RTmax
(6)
The calculation of the activation energy with this approximate
formula made it possible to obtain a linear correlation (R2 = 0.978)
between the sigma constants (ꢀσ) of the coumarin substituents
and the activation energies of decarboxylation of coumarins 1, 2,
5 and 6 (Fig. 4). The deviation from the correlation displayed by
coumarins 3 and 7 is attributed to the ortho-effect of the sub-
stituents. The sigma constants were obtained from the experimen-
tal data by calculating the effect of the substituents on the value
of pK for the model series of corresponding coumarin acids. Calcu-
lations were made with the Analytical Data Handling and Physic-
ochemical Prediction modules of the ACD Labs Software package.
The observed linear correlation shows that the LFER principle
applies to the free energy of reaction and the activation energy
of the correlated reaction series. The decrease of the activation
energy at increasing values of the electron-seeking force of the
substituents indicates that the electron density in the transition
state grows, i.e. the reaction parameter ρ has a positive value.
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

K. Kulyk et al.
Table 1. Kinetic parameters of the chemical reactions of silica-supported coumarins
Sample, amount on
silica surface
E=,
ν, s−1 (n = 1);
Reaction product
Tmax
,
◦C m/z
n
kJ·mol−1 l·mol−1 s−1 (n = 2) dS=, cal·K−1·mol−1
R2
SiO2
Coumarin 1, 0.3 mmol g−1
CO2
CO2
481
475
44
44
1
2
1
2
1
1
2
1
2
1
1
1
2
1
1
1
1
1
1
99
177
100
197
166
99
1.36 × 104
8.49 × 109
2.10 × 104
3.25 × 1011
1.03 × 107
1.21 × 104
1.37 × 109
1.88 × 104
3.04 × 109
1.33 × 104
5.24 × 1010
1.50 × 104
1.12 × 1011
1.60 × 104
1.55 × 106
2.11 × 104
1.66 × 105
2.5 × 107
−41.6
−14.9
−40.6
−7.7
0.952
0.944
0.959
0.949
0.932
0.952
0.848
0.955
0.888
0.954
0.969
0.950
0.927
0.953
0.941
0.96
Coumarin 2, 0.6 mmol g−1
44
44
Phenylacetylene
CO2
617
502
102
44
−28.6
−41.8
−18.6
−40.6
−16.8
−41.4
−20.4
−38.6
−9.6
Coumarin 3, 0.3 mmol g−1
Coumarin 5, 0.4 mmol g−1
163
92
CO2
427
44
160
96
4-methylthiazole
4H-3-methyl-1,4-thiazine
CO2
457
437
408
99
113
44
153
91
Coumarin 6, 0.4 mmol g−1
Coumarin 7, 0.3 mmol g−1
175
104
126
99
Benzothiazole
Benzo-4H-1,4-thiazine
CO2
512
465
458
487
449
109
135
149
44
−41.1
−32
−40.5
−36.5
−26.4
−29.6
4-methyl-thiazole
99
113
137
63
0.980
0.960
0.954
4H-3-methyl-1,4-thiazine
113
113
4-Methylthiazole, 0.3 mmol g−1 4H-3-methyl-1,4-thiazine
CeO2/SiO2
Coumarin 7, 0.3 mmol g−1
2.57 × 106
CO2
420
433
406
44
99
1
1
1
90
100
109
1.75 × 104
9.18 × 104
1.17 × 106
−40.6
−37.5
−32.5
0.950
0.968
0.977
4-methylthiazole
4H-3-methyl-1,4-thiazine
113
TiO2/SiO2
Coumarin 7, 0.3 mmol g−1
CO2
440
407
44
1
1
87a
113b
∼104
∼106
–
–
–
–
4H-3-methyl-1,4-thiazine
113
Al2O3/SiO2
Coumarin 7, 0.3 mmol g−1
CO2
350
510
370
479
44
44
1
1
1
1
76a
96a
107b
125b
∼104
∼104
∼106
∼106
–
–
–
–
–
–
–
–
4H-3-methyl-1,4-thiazine
113
113
a The activation energy was calculated according to approximation formula E= = 15 · RTmax
b The activation energy was calculated according to approximation formula E= = 20 · RTmax
.
.
It must also be remarked that as the correlations obtained from
the LFER principle are empirical it is not necessary to know the
exact absolute values of the activation energy.[15] For example,
sometimes only some information about the order of magnitude
for the reaction parameter ρ is enough for clarifying the reaction
mechanism. These minimum and approximate pieces of informa-
tion are usually the answers to the questions:[15] Is the constant
ρ positive or negative? Is its numerical value large or small? Are
there any appreciable aberrations from the curve linearity?[15]
After CO2 is detached, the coumarin fragments that remain
adsorbed seemingly polymerize on the silica surface. Upon further
heating above 500 ◦C some coumarin samples (1–3, 5–7) release
ions with m/z = 102 and 78. Coumarin 2 also displays a TPD
trace of m/z = 128 (Fig. 5) indicating[43] that naphthalene (128
Da), phenylacetylene (102 Da) and benzene (78 Da) are released
during the degradation of the polymer. The traces of thermal
desorption of these compounds are those with m/z = 128, 102
and 78. These traces have well-resolved peaks (Fig. 5) that can
be analyzed to obtain the corresponding kinetic parameters of
desorption. Table 1 contains the values of these parameters for
the case of the formation of phenylacetylene.
From the previous analysis of the TPD-MS data, the thermal
transformations of silica-supported coumarins 1–3 can be
described by a common mechanism exemplified in Scheme 3
for the case of coumarin 2.
Coumarins5–7arefoundtoundergofurtherthermolysisstages
(Figs 6 and 7) that correspond to the thermal transformations of
the heterocyclic substituent. Coumarin 5 is found to have TPD
peaks in the traces of m/z = 113, 99 and 71 (Fig. 6). Coumarin 6
has similar peaks at m/z = 149, 135 and 108 (Fig. 7). According
to the AIST database,[43] the lines in the mass spectra shown
in Figs 6 and 7 correspond to 4-methylthiazole (m/z = 99)
and benzothiazole (m/z = 135). Most likely, the second stage
of thermolysis of coumarins 5 and 6 occurs with emission
of molecular 4-methylthiazole (Scheme 4) and benzothiazole
(Scheme 5), respectively.
It is interesting to note that in the case of coumarins with
heterocyclic substituents (5–7) the desorption products have
a molecular mass that differs from that of the heterocyclic
substituent by 14 amu. These are the products with m/z = 113
(5 and 7) and m/z = 149 (6). The maximum in the TPD trace
of the ion with m/z = 113 can be explained by the emission of
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2010, 45, 750–761

TPD-MS of silica-supported coumarins
Figure 5. Mass spectrum at 704 ◦C (a) and thermal desorption traces of the ions with m/z = 128, 102 and 78 (b), corresponding to the thermal
decomposition of silica-supported 4 methyl-7-hydroxycoumarin (2) (0.6 mmol g−1).
CH
T
max = 617 oC
m/z 102
m/z 128
CH3
O
Tmax = 637 oC
Stage 2
Stage 1
CO2
T
max = 475 oC
m/z 44
HO
O
ads.
Tmax = 750 oC
m/z 78
Scheme 3. The thermal decomposition of silica supported 4-methyl-7-hydroxycoumarin (2).
molecular 4H-3-methyl-1,4-thiazine (113 Da). In the case of the
ion with m/z = 149, the peak can be attributed to the release
of benzo-4I-1,4-thiazine (149 Da) (Schemes 4 and 5). Thiazine
formation is thus likely the result of the rearrangements of thiazole
and benzothiazole substituents on the silica surface.
The complex chemical reactions involving the heterocyclic
substituents (4-methylthiazole and benzothiazole) and occurring
on the silica surface at high temperatures are associated to the
peaks in the traces of ions with m/z = 32 and 34 that correspond
to the formation of H2S (Figs 2, 6 and 7).
Additional TPD-MS tests of silica samples impregnated with
4-methylthiazole show that the mass spectrum at 105 ◦C has a
line with m/z = 113. This ion has a TPD trace with a maximum
at Tmax = 110 ◦C (Fig. 8; Table 1). The mass spectrum of bulk
4-methylthiazole, obtained by flooding the sample ampoule with
its vapors, does not contain the line of m/z = 113. These
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

K. Kulyk et al.
Figure 6. Mass spectrum at447 ◦C(a) and thermaldesorption traces of the ions with m/z = 32, 71, 78, 99and 113(b), correspondingto the decomposition
of silica-supported 3-(4-methyl-1,3-thiazol-2-yl)-7-hydroxycoumarin (5) (0.4 mmol g−1).
H
CH3
max = 437 oC
N
T
CH3
S
N
O
m/z 113
Stage 3
max = 750 oC
Stage 1
Stage 2
CO2
T
S
T
max = 427 oC
CH3
m/z 44
N
m/z 78
HO
O
Tmax = 457 oC
ads.
S
m/z 99
Scheme 4. The decomposition of silica supported 3-(4-methyl-2-thiazole)-7-hydroxycoumarin (5).
H
N
Tmax = 465 oC
S
N
m/z 149
Stage 3
Stage 1
Stage 2
CO2
m/z 44
S
Tmax = 408 oC
T
max = 750 oC
m/z 78
O
O
N
ads.
Tmax = 512 oC
OH
S
m/z 135
Scheme 5. The thermal decomposition of silica supported 3-(2-benzothiazole)-8-hydroxycoumarine (6).
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2010, 45, 750–761

TPD-MS of silica-supported coumarins
Figure 7. Mass spectrum at 518 ◦C (a) and thermal desorption traces of the ions with m/z = 149, 135, 108, 78, 34 and 32 (b), corresponding to the thermal
decomposition of silica-supported 3-(1,3-benzothiazol-2-yl)-8-hydroxycoumarin (6) (0.4 mmol g−1).
data confirm the presence of rearrangements of isolated 4-
methylthiazole when grafted on the silica surface though the
exact mechanism has not been yet elucidated.
substituted coumarins. Kinetic parameters (n, E=, ν, dS=) of
the silica-assisted reaction of thiazine formation were also cal-
culated for the CeO2/SiO2, Al2O3/SiO2 and TiO2/SiO2 systems
(Fig. 9; Table 1). The activation energy of the thiazole–thiazine
TheTPDtracesoftheionswithm/z = 99and113corresponding
to the thermolysis of coumarin 5 and those of m/z = 135 and
149 corresponding to coumarin 6 have no overlapping peak
temperatures (Figs 6 and 7). This indicates that two different
surfacechemicalprocessesareoccurring ineachcase. Thepeaksof
releaseofmolecularheterocyclicsubstituentsarewiderandshifted
to higher temperatures as compared to those corresponding to
the release of products of rearrangement.
The maximum rate of 4-methylthiazole formation on silica
occurs at Tmax = 457 ◦C. This is shifted to lower temperatures
by 55 ◦C in the case of benzothiazole (Tmax = 502 ◦C). It
can be seen that the ion with m/z = 113 has a maximum
release rate at a lower temperature (Tmax = 437 ◦C) than
4-methylthiazole. The same occurs with the ion of m/z =
149 in comparison to benzothiazole (Tmax = 465 ◦C). This
means that 4-methylthiazole and benzothiazole desorb from
the surface at higher temperatures than the products of surface
rearrangement.
rearrangement decreased in the order: SiO2 > CeO2/SiO2
≈
TiO2/SiO2 > Al2O3/SiO2. It is interesting to note that thermodes-
orption curve for m/z 113 have two maximums (Tmax = 370 ◦C and
Tmax = 510 ◦C) during thermolysis on the surface of Al2O3/SiO2.
Perhaps this phenomenon could be attributed to contribution of
Lewis acid centers on thermodestruction process.
Formation of CO2 on surface of mixed oxides also occurs at
lower temperatures with respect to pure fumed silica (Fig. 10;
Table 1). Two maximums on the stage of carbon dioxide formation
are observed on the surface of Al2O3/SiO2.
Conclusions
It has been found that free hydroxyls groups in positions 7 and 8
of the coumarin nucleus play a crucial role in the chemisorption
of these molecules to the silica surface. Thermolysis stages were
found to be related to certain structural elements of natural and
The thiazole–thiazine rearrangement occurred during the py-
rolysis of all the thiazole heterocycles, i.e. pure thiazole and
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

K. Kulyk et al.
Figure 8. Mass spectrum at 105 ◦C (a) and TPD-MS traces of a silica sample impregnated with 4-methylthiazole (b).
Figure 9. Thermal desorption traces of the ions with m/z = 113 cor-
responding to the thermal decomposition of pure silica and mixed
oxide-supported 7-hydroxy-3-(4-methyl-1,3-thiazol-2-yl)-8-(pyrrolidin-1-
ylmethyl)coumarine (7).
Figure 10. Thermal desorption traces of the ions with m/z = 44 cor-
responding to the thermal decomposition of pure silica and mixed
oxide-supported 7-hydroxy-3-(4-methyl-1,3-thiazol-2-yl)-8-(pyrrolidin-1-
ylmethyl)coumarine (7).
modified coumarins: the coumarin nucleus, the thiazole and the
benzothiazole substituents.
The thermal degradation of the studied silica-supported
coumarins starts with the detachment of a CO2 molecule and
the further release of molecular benzene, naphthalene and
phenylacetylene at temperatures over 500 ◦C.
A linear correlation between the activation energy for de-
carboxylation of the silica-supported coumarins and the sigma
constants (ꢀσ) of their substituents was obtained.
Heterocyclic coumarine derivatives are found to have addi-
tional thermolysis stages that encompass the release of molec-
ular heterocyclic compounds, particularly 4-methylthiazole and
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2010, 45, 750–761

TPD-MS of silica-supported coumarins
benzothiazole. Rearrangements in the adsorbed state proceeded
through a ‘thiazole–thiazine’ ring expansion.
The activation energies of the thiazole–thiazine rearrangement
[20] L. P. Hammet. PhysicalOrganicChemistry, 2nd ed. McGraw-Hill: New
York, 1970.
[21] R. W. Taft. Polar and Steric Substituent Constants for Aliphatic and
o-Benzoate Groups from Rates of Esterification and Hydrolysis of
Esters. J. Am. Chem. Soc. 1952, 74, 3120.
[22] R. W. Taft, C. Grob. Separation of polar and resonance effects in the
ionization of 4-substituted pyridinium ions. J. Am. Chem. Soc. 1974,
96, 1236.
[23] B. J. Bronsted. Die katalytische Zersetzung des Nitramids und ihre
physikalisch-chemische Bedeutung. Z. Phys. Chem. 1924, 108, 185.
[24] M. Kraus. Advances in Catalysis, vol. 17. Academic Press: New York
and London, 1967, 75.
and reaction of CO2 formation decreased in the order: SiO2
>
CeO2/SiO2 ≈ TiO2/SiO2 > Al2O3/SiO2.
Acknowledgements
The authors thank Prof. V.M. Gun’ko and Dr V.I. Zarko (Chuiko
Institute of Surface Chemistry, Kyiv) for some oxide samples.
[25] V. P. Khylya, O. V. Shablykina, V. V. Ishchenko. Selected Methods for
Synthesis and Modification of Heterocycles, vol. 2,V. G. Kartsev (Ed).
IBS PRESS: Moscow, 2003, 518 (in Russian).
References
[26] M. E. Perelson, Yu.N. Sheinker, A. A. Savina. Spectra and Structures of
Coumarins, Chromones and Xanthones. Meditsina: Moscow, 1975.
[27] Medical Chemistry and Clinical Applications of Silica Dioxide,
A. A. Chuiko (ED). Naukova Dumka: Kyiv, 2003, 416 (in Russian).
[28] T. V. Kulik, N. P. Galagan, V. K. Pogoreliy, V. A. Pokrovskiy, Advances
in Mass Spectrometry, Proceedings of the 15th International Mass
Spectrometry Conference, vol. 15,E. Gelpi (Ed). John Wiley&Sons:
Chichester, 2001, 491.
[29] V. A. Pokrovskiy, N. P. Galagan, T. V. Kulik. Chemistry, Physics and
Technology of Surface, vol 4–6,O. O. Chuiko (Ed). Institute of Surface
chemistry: Kyiv, 2001, 332.
[30] T. V. Kulyk, V. V. Ischenko, B. B. Palyanytsya, V. P. Khylya, K. S. Kulyk,
V. N. Barvinchenko, V. K. Pogorelyi. A study on interactions of
coumarins and their heterocyclic derivatives with a surface of
silicabydesorptionmassspectrometrymethod. Mass-spectrometria
2007, 4, 179 (in Russian).
[31] T. V. Kulik, N. A. Lipkovska, V. N. Barvinchenko, B. B. Palyanytsya,
O. A. Kazakova, O. A. Dovbiy, V. K. Pogorelyi. Interactions between
bioactive ferulic acid and fumed silica by UV-VIS spectroscopy FT-IR,
TPD MS investigation and quantum chemical methods. J. Colloid
Interface Sci. 2009, 339, 60.
[32] T. V. Kulik, V. N. Barvinchenko, B. B. Palyanytsya, O. V. Smirnova,
V. K. Pogoreliy, A. A. Chuiko. ADesorptionMassSpectrometryStudy
of the Interaction of Cinnamic Acid with Silica Surface. Russ. J. Phys.
Chem. 2007, 81, 83.
[33] M. V. Borysenko, K. S. Kulyk, M. V. Ignatovych, E. N. Poddenezhny,
A. A. Boiko, A. O. Dobrodey. Nanomaterials and Supramolecular
Structures, A. P. Shpak, P. P. Gorbyk (Eds). Springer: Dordrecht, The
Netherlands, 2009, 227.
[1] F. Hillenkamp, M. Karas. Matrix-assisted laser desorption/ionisation,
an experience. Int. J. Mass Spectrom. 2000, 200, 71.
[2] K. Breuker, R. Knochenmuss, R. Zenobi. Proton transfer reactions of
matrix-assisted laser desorption/ionization matrix monomers and
dimers. J. Am. Soc. Mass Spectrom. 1999, 10, 1111.
[3] R. Zenobi, R. Knochenmuss. Ion formation in MALDI mass
spectrometry. Mass Spectrom. Rev. 1999, 17, 337.
[4] R. Knochenmuss,A. Stortelder,K. Breuker,R. Zenobi.Secondaryion-
molecule reactions in matrix-assisted laser desorption/ionization. J.
Mass spectrom. 2000, 35, 1237.
[5] S. Alimpiev, A. Grechnikov, J. Sunner, V. Karavanskii, Ya. Simanovsky,
S. Zhabin, S. Nikiforov. On the role of defects and surface chemistry
forsurface-assisted laserdesorption ionization from silicon. J.Chem.
Phys. 2008, 128, 14711.
[6] R. J. Cvetanovic, Y. Amenomiya.
A temperature programmed
desorption technique for investigation of practical catalysts. Catal.
Rev. 1972, 6, 21.
[7] J. A. Delgado, J. M. Gomez. Estimation of Adsorption Parameters
From Temperature-Programed-Desorption Thermograms: Applica-
tiontotheAdsorptionofCarbonDioxideontoNa-andH-Mordenite.
Langmuir 2005, 21, 9555.
[8] S. I. Nicholl, J. W. Talley. Development of thermal programmed
desorption mass spectrometry methods for environmental
applications. Chemosphere 2006, 63, 132.
[9] J. B. Miller, H. R. Siddiqui, S. M. Gates, J. N. Russell, J. T. Yates.
Extraction of kinetic parameters in temperature programmed
desorption: a comparison of methods. J. Chem. Phys. 1987, 87,
6725.
[34] W. Rudzinski, T. Borowiecki, A. Dominko, T. Panczyk.
A New
[10] S. I. Nicholl, J. W. Talley, S. Silliman. Model verification of thermal
programmed desorption-mass spectrometry for estimation of
release energy values for polycyclic aromatic hydrocarbons on
mineral sorbents. Environ. Toxicol. Chem. 2004, 23, 2545.
[11] T. Panczyk, P. Szabelski. The influence of lateral interactions
between adsorbed molecules on adsorption kinetics. A statistical
rate theory approach. J. Phys. Chem. B 2003, 107, 5586.
[12] W. Rudzinski, T. Panczyk. Remarksonthecurrentstateofadsorption
kinetic theories for heterogeneous solid surfaces: a comparison of
the ART and the SRT approaches. Langmuir 2002, 18, 439.
[13] V. V. Brei, V. M. Gun’ko, V. V. Dudnik, A. A. Chuiko. Study of Kinetics
andMechanismsofSomeUnimolecularReactionsonSilicaSurfaces.
Langmuir 1992, 8, 1968.
[14] V. M. Gun’ko, I. F. Mironyuk, V. I. Zarko, E. F. Voronin, V. V. Turov, E.
M. Pakhlov, E. V. Goncharuk, Y. M. Nychiporuk, N. N. Vlasova, P. P.
Gorbik, O. A. Mishchuk, A. A. Chuiko, T. V. Kulik, B. B. Palyanytsya,
S. V. Pakhovchishin, J. Skubiszewska-Zi ba, W. Janusz, R. Leboda.
Morphology and surface properties of fumed silicas. J. Colloid
Interface Sci. 2005, 289, 427.
Quantitative Interpretation of Temperature-Programmed
Desorption Spectra from Heterogeneous Solid Surfaces, Based
on Statistical Rate Theory of Interfacial Transport: The Effects of
Simultaneous Readsorption. Langmuir 1999, 15, 6386.
[35] J. Park, R. T. Yang. Predicting Adsorption Isotherms of Low-Volatile
Compounds by Temperature Programmed Desorption: Iodine on
Carbon. Langmuir 2005, 21, 5055.
[36] J. Joly, A. Perrard. Determination of the Heat of Adsorption of
Ammonia on Zeolites from Temperature-Programmed Desorption
Experiments. Langmuir 2001, 17, 15.
[37] V. M. Gun’ko, J. Skubiszewska-Zi ba, R. Leboda, V. I. Zarko. Fumed
Silica Modified Due to Pyrolysis of Methylene Chloride. Langmuir
2000, 16, 374.
[38] J. P. Blitz, V. M. Gun’ko (Eds). Surface Chemistry in Biomedical and
Environmental Science. Springer: Netherlands, Dordrecht, 2006,
135.
[39] J. J. Carberry. Chemical and Catalytic Reaction Engineering. McGraw-
Hill: New York, 1976.
[40] P. A. Redhead. Thermal Desorption of Gases. Vacuum 1962, 12, 245.
[41] M. U. Kisliyk, V. V. Rozanov. Temperature-programmed desorption
and temperature-programmed reaction - methods of studying of
kinetics and mechanisms of catalytic processes. Kinet. Katal. 1995,
36, 89.
[42] D. P. Woodruff, T. A. Delchar. Modern Techniques of Surface Science.
Cambridge University Press: London, 1986.
[43] SDBSWeb. http://www.aist.go.jp/RIODB/SDBS/ (National Institute
of Advanced Industrial Science and Technology, Last accessed:
March 21, 2009).
[15] P. Sykes. A Guidebook to Mechanism in Organic Chemistry. Longman
Scientific&Technical: London, 1986.
[16] H. H. Jaffe. A Reexamination of the Hammett Equation. Chem. Rev.
1953, 53, 191.
[17] P. R. Wells. Linear Free Energy Relationships. Chem. Rev. 1963, 63,
171.
[18] P. R. Wells, Linear Free Energy Relationships. Academic Press: New
York, 1968.
[19] L. P. Hammet. Some Relations between Reaction Rates and
Equilibrium Constants. Chem. Rev. 1935, 17, 125.
c
J. Mass. Spectrom. 2010, 45, 750–761
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms