Page 1
Ionization and protonation of aromatic diamines by sorption
in zeolites
S. Marquisa, A. Moissettea,*, C. Bremarda, H. Vezinb
aLaboratoire de Spectrochimie Infrarouge et Raman UMR-CNRS 8516, Bat. C5 Universite des Sciences et Technologies de Lille,
59655 Villeneuve d’Ascq cedex, FrancebLaboratoire de Chimie Organique et Macromoleculaire ESA-CNRS 8009, Bat. C3 Universite des Sciences et Technologies de Lille,
59655 Villeneuve d’Ascq cedex, France
Received 2 September 2002; accepted 23 September 2002
Abstract
In situ diffuse reflectance UV visible, Raman scattering and EPR spectroscopies were used to monitor spontaneous ionization
of N,N,N0,N0-tetramethyl-p-phenylenediamine (TMPD) by direct exposure to dehydrated MnZSM-5 zeolite (M ¼ Naþ, Hþ;
n ¼ 0,3,6). The TMPD†þ radical cation is found to be generated in low yield in purely siliceous silicalite-1 whereas TMPD†þ is
generated in high yield in aluminated NanZSM-5 and HnZSM-5. The ejected electron is also characterized though electronic
absorption spectra. The charge separation was found to be persistent over several months. The tight fit between the shape of
TMPD and the pore size of straight channels of zeolites is considered to be the main factor responsible for the stabilization of
the TMPD radical ion by preventing rapid electron back transfer. Within NanZSM-5, the amounts of TMPD†þ and trapped
electron were found to decrease over long time and to recombine to molecular TMPD. In contrast, in acidic HnZSM-5 zeolite
the charge recombination generates diprotonated TMPDH22þ occluded species. Furthermore, within HnZSM-5, protonation
appears competitive to the ionization efficiency of zeolite.
q 2003 Elsevier Science B.V. All rights reserved.
Keywords: Zeolite; Aromatic amine; Radical cation; Ionization; Protonation
1. Introduction
A variety of organic radical cations can be
generated spontaneously by inclusion of their
precursors in the void space of porous materials
such as zeolites [1–7]. The tight fit between the
shape of the occluded species and the pore size of
zeolites is considered to be an important factor
responsible for the stabilization of radical cations
[3,6]. The strong electron donor aromatic amines are
known to have low ionization potential in the gas
phase and to exhibit proton affinity. The rod-shaped
N,N,N0,N0-tetramethyl-p-phenylenediamine (TMPD)
molecules can penetrate within the channels of
ZSM-5 type of zeolites [7]. The fully N-methylated
derivative was chosen to avoid further intrazeolite
conversion. The spectroscopy and behavior of
TMPD and protonated species as well as radical
cation in solution are well documented [8–12].
0022-2860/03/$ - see front matter q 2003 Elsevier Science B.V. All rights reserved.
PII: S0 02 2 -2 86 0 (0 2) 00 6 46 -4
Journal of Molecular Structure 651–653 (2003) 305–314
www.elsevier.com/locate/molstruc
* Corresponding author. Fax: þ33-3-20436755.
E-mail address: [email protected] (A. Moissette).
Page 2
After dehydration, ZSM-5 zeolites are microporous
crystalline aluminosilicates which contain two types
of intersecting channels with openings of 5.5 mm in
size. The silicalite-1 (Si/Al , 1000) is the purely
siliceous analogue of the ZSM-5 zeolite. The
incorporation of aluminum (III) in the framework
induces the presence of counterbalancing extrafra-
mework cation in MnZSM-5 zeolites (M ¼ Naþ,
Hþ; n ¼ 3, 6).
In this paper, we report results obtained upon
incorporating TMPD in MnZSM-5 zeolites with the
chemical composition Mn(AlO2)n(SiO2)962n. The
resulting samples were probed by the application of
diffuse reflectance UV-visible absorption (DRUVv),
electron paramagnetic resonance (EPR) and Raman
scattering spectroscopies.
2. Experimental
2.1. Materials
The sodium and NH4þ-exchanged ZSM-5 samples
(Si/Al ¼ 13, 27, particle size ,1 mm) were obtained
from VAW Aluminum (Schwandorf, Germany).
Crystals of silicalite-1 (Si/Al . 1000) were syn-
thesised in high purity according to the fluoride
medium procedure and was a gift of Dr J. Patarin
(Laboratoire des Materiaux Mineraux, ESA 7016,
CNRS- ENSC Mulhouse). All the zeolite samples
were used after a calcination procedure up to 500 8C.
The unit cell compositions of the calcined and
dehydrated MnZSM-5 (M ¼ Naþ, Hþ, n ¼ 0, 3.4,
6.6) were found to be Si96O192, Na3.4(AlO2)3.4
(SiO2)92.6, Na6.6(AlO2)6.6(SiO2)89.4, H3.4(AlO2)3.4
(SiO2)92.6, and H6.6(AlO2)6.6(SiO2)89.4 from elemental
analysis. In the following text, we use the notations
Na3ZSM-5, Na6ZSM-5, H3ZSM-5, H6ZSM-5 for
convenience. The crystallinity of the samples was
checked by XRD and was not reduced by the thermal
treatment. TMPD molecules (C10H16N2) were syn-
thesized and purified as described previously [11].
Pure and dry Ar and O2 gas were used.
2.2. Sorption of TMPD
Weighted amounts (,1.4 g) of freshly calcined
zeolite were held under vacuum and then under dry
argon. Weighted amount of TMPD, corresponding to
one mole per unit cell (UC), was introduced into the
cell under dry argon and the powder mixture was
shaken. The sample was transferred under dry argon
in a Suprasil quartz glass cuvette for DRUVv and FT-
Raman experiments and in a sealed quartz tube for
EPR measurements.
2.3. Instrumentation
The UV-visible absorption spectra of the sample
were recorded between 200 and 800 nm using a Cary
3 spectrometer. The instrument was equipped with an
integrating sphere to study the powdered zeolite
samples through diffuse reflectance; the correspond-
ing bare zeolite was used as the reference. The
DRUVv spectra were plotted as the Kubelka–Munk
function: FðRÞ ¼ ð1 2 RÞ2=2R ¼ K=S; where R rep-
resents the ratio of the diffuse reflectance of the loaded
zeolite to that of the dehydrated neat zeolite, K
designates an absorption coefficient proportional to
the concentration of the chromophore and S the
scattering coefficient of the powder.
Raman scattering spectra were collected on a
LabRAM spectrometer (Instruments S.A) equipped
with a Peltier-cooled Charge Coupled Device. The
excitation wavelength used was 632.8 nm with low
laser power to avoid irreversible laser effects for the
sample. The laser lines were supplied by ionized
argon laser and Helium–Neon laser. The spec-
trometer calibration was verified using the Raman
lines of silicon. This resulted in an accuracy of less
than ^1 cm21.
A BRUKER IFS 88 instrument was used as a FT-
Raman spectrometer with a cw Nd:YAG laser at
1064 nm as excitation source. A laser power of 100–
200 mW was used. The spectra (3500–150 cm21)
were recorded with a resolution of 2 cm21 using 400
scans. The Opus BRUKER software was used for
spectral acquisition, data treatment and plotting.
The (X-band) EPR spectra were recorded at room
temperature on a Bruker ESP-300 spectrometer. The
EPR signals were double integrated using Bruker
software and the spin concentration was determined
relatively to a reference standard.
The structural modeling calculations were per-
formed on a Silicon Graphics workstation using
Cerius2 (version 3.8) package from Molecular
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314306
Page 3
Simulations. The zeolite structural parameters, force
field and calculation details used to carry out the
minimization procedure of the non-bonding energy
were as described previously [13].
The data processing of the UV-visible absorption
spectra recorded during the course of the sorption
was performed using the SIMPLISMA (SIMPle-to-
use Interactive Self-modeling Mixture Analysis)
approach [14]. This method was applied to extract
the characteristic spectra of species generated
through sorption from many spectral data, which
resolves spectrum of mixture into pure component
spectra without any prior information. In addition
the program calculates the relative concentration of
the species. The accuracy of the calculation is
given by residuals, which represent the difference
between the reconstructed, and the original
data. This coefficient can be seen as a standard
deviation and was defined previously [15]. The
values are between 0 and 1. The algorithm for
SIMPLISMA calculation was described in detail
elsewhere [15].
3. Results and discussion
3.1. Sorption of TMPD into silicalite-1 (Si/Al ¼ 1000)
The mere exposure of solid TMPD to dehydrated
silicalite-1 powder generates weak purple coloration.
The EPR experiments provide evidence of ioniz-
ation. The formation of radical cation is detected by
the characteristic absorption features (611, 562,
518 nm) exhibited in the DRUVv spectra recorded
after the mixing of powders [7]. The sorption of
TMPD molecules within the void space of zeolite
was monitored through the increase of the absorp-
tion bands at 276 and 325 nm. The sorption course
of TMPD at loading corresponding to one mole per
unit cell (1 TMPD/UC) was monitored as a function
of time by FT Raman spectrometry. The progressive
disappearance of solid TMPD is observed through
the decrease of the Raman bands at 777 and
1620 cm21 whereas the occupancy of the void
space is observed through the appearance of
Raman bands at 783 and 1630 cm21. The ionization
yield was found to be too weak to be detected by
Raman spectrometry with off resonance conditions.
However, after a long exposure, TMPD†þ was
detected in low yield through the band at
1632 cm21.
Energy minimization procedures of the nonbond-
ing interactions between rigid planar molecule and
fixed zeolite framework provide a reasonable
structural picture of the ability of TMPD to pass
through the pore openings of silicalite-1. The
molecules were found to penetrate into the internal
void space of the zeolite and the preferred
locations lay in straight channels in the
vicinity of the intersection with zigzag channel
with the C–N bonds running along the b direction
of the cell, Fig. 1.
3.2. Sorption of TMPD into NanZSM5 (Si/Al ¼ 27, 13)
Na3ZSM-5 (Si/Al ¼ 27)—The exposure of solid
TMPD to dehydrated Na3ZSM-5 generates rapidly
a deep purple colour. The EPR experiments
indicate an intense signal in agreement with
ionization of TMPD in high yield upon sorption.
DRUVv spectra recorded as a function of time
after the mixing of the powders with 1 TMPD/UC
loading exhibit new absorption bands in the 500–
650 nm region characteristic of TMPD radical
cation, TMPD†þ (Fig. 2). The SIMPLISMA
approach carried out on the spectra set provides
evidence of three extracted spectra (Fig. 3). The
first extracted spectrum is assigned to molecular
TMPD. The second extracted spectrum displays the
spectral features at 611, 562, 518 and 325 nm of
TMPD†þ. The spectrum with broad band observed
in the 350–600 nm region is assigned with respect
to previous works [6] to the electron transferred to
the zeolite framework after the spontaneous TMPD
ionization. The broad band observed is explained
by different environments and so different
local fields for electrons during the sorption course,
Fig. 3(c) [16]. Over one year, DRUVv spectra
provide evidence of progressive decrease of radical
cation and electron signatures. These decreases are
correlated with a significant increase of the 270 nm
band intensity characteristic of molecular TMPD.
This feature shows that recombination of radical
cation occurs through the capture of an electron
and formation of neutral TMPD occluded in the
zeolite channels.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314 307
Page 4
The sorption process of TMPD at 1 TMPD/UC
loading was monitored as a function of time by FT-
Raman spectrometry. Over one month, the Raman
features of solid TMPD disappear nearly totally
whereas the Raman bands of TMPD†þ at 1633, 1508,
1418 cm21 are found to increase significantly. After
more than 3 months, the characteristic bands of neutral
occluded TMPD centered at 777 and 1620 cm21
increase at the expense of TMPD†þ consistent with
the recombination of radical cation as occluded TMPD.
Using the 632 nm excitation line, the resonance
Raman spectra for TMPD occluded within Na3ZSM-5
displays exclusively the spectral features of radical
cation.
Na6ZSM-5 (Si/Al ¼ 13)—The ionization process
is found to be more rapid for TMPD adsorbed into
Na6ZSM-5 compared to TMPD within Na3ZSM-5 but
the process and maximum yield of TMPD†þ are
found to be analogous before the charge
recombination.
3.3. Sorption of TMPD into acidic HnZSM-5
(Si/Al ¼ 13, 27)
H3ZSM-5 (Si/Al ¼ 27)—The mixing of TMPD
and dehydrated zeolite induces nearly instantaneous
deep purple coloration. The EPR experiments
indicate an intense signal in agreement with
ionization of TMPD in high yield upon sorption.
DRUVv spectra recorded as a function of time from
the mixing of the powders exhibit new absorption
bands in the 500–650 nm spectral range character-
istic of TMPD radical cation. The maximum
intensity of TMPDþ† band is reached within 3
days. Concomitantly to the TMPDþ† bands in the
visible region, we can also observe the significant
increase of the 325 nm and 260 nm bands, Fig. 4(a).
These two bands are assigned to contribution of
both TMPD†þ and TMPD. The 325 nm band is
mainly due to TMPD†þ whereas the 260 nm one
comes essentially from TMPD. Within 7 days,
Fig. 1. Modeling of the structure of TMPD occluded in the straight channel of silicalite-1. Black and shaded sticks represent the O and Si atoms
of the (SiO2)96 framework, respectively. The white, dark grey and black cylinders represent the H, C and N atoms of the C10H16N2 molecule,
respectively.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314308
Page 5
the 260 nm band is no more visible in agreement
with nearly total ionization of TMPD. However, it
should be noted that a new weaker band is observed
at 255 nm. This band is assigned to TMPDH22þ, the
diprotonated form of TMPD by comparison with the
spectrum of diprotonated TMPD obtained in con-
centrated chloride acid solution (12 M), (not shown).
Over longer period of time, the TMPD†þ contri-
bution slowly decreases relatively to the weak band
of protonated TMPD. One year after the mixture of
the powders, the DRUVv spectrum displays only
weak contribution of TMPD†þ, Fig. 4(b). It should
be noted that the extinction coefficients of
TMPDH22þ are markedly weaker than the radical
Fig. 2. DRUV spectra recorded at room temperature during the course of adsorption of TMPD into Na3ZSM-5 [Na3.4(AlO2)3.4(SiO2)92.6] zeolite
by mixing of powders. (a) 5 min to 48 h (b) 48 h to 4 months.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314 309
Page 6
cation ones [8]. Thus, assumption is made that
TMPD†þ recombination within acid zeolite chan-
nels occurs exclusively through diprotonation of
TMPD and that a mixture of TMPD†; and
TMPDH22þ is observed one year after the beginning
of the reaction.
Data processing using the SIMPLISMA program
of numerous diffuse reflectance UV-visible spectra
recorded over several months during the course of
TMPD sorption into calcined H3(AlO2)3(SiO2)93
provides evidence of four independent spectra
(Fig. 5). The first extracted spectrum is assigned to
solid TMPD. The second spectrum exhibits the
characteristic features of TMPD†þ. The spectrum
with broad band observed in the 350–600 nm region
is assigned to the ejected electron transferred to the
zeolite framework after the spontaneous TMPD
ionization [6]. The shape of the spectral signature of
the electron differs slightly from that observed for
electron within NaZSM-5. The fourth spectrum
corresponds to TMPDH22þ.
Protonation is only a minor phenomenon compared
to ionization during the first steps of the sorption
process. However, the increase of concentration of
TMPDH22þ corresponds to the TMPD†þ decrease.
Thus, it is obvious that TMPDH22þ essentially
comes from TMPD†þ recombination with electron.
The neutral molecule TMPD contribution is found to
fall to zero in agreement with total ionization and
protonation.
The sorption process of TMPD (1 TMPD/UC) was
monitored as a function of time by FT-Raman
spectrometry. Over several days, the Raman features
evolved from the characteristic spectra of solid TMPD
[11,12] to spectra characteristic of TMPD†þ (Fig. 6).
In these off resonance conditions, the prominent line
of neutral TMPD centered at 777 cm21 disappears
after 5 days. The observation of only radical cation
features with high signal/background ratio provides
evidence of nearly complete ionization of TMPD
molecule. After about 20 days, the Raman spectra
display supplementary weak lines at 797, 900, 995,
1132, 1467 cm21 and shoulder at ca. 1610 –
1620 cm21. These lines are assigned to TMPDH22þ,
by comparison with Raman spectrum recorded in
hydrochloride concentrated solution, Fig. 6 [11]. One
year and a half after the mixing of the powders, the
Raman spectrum is still representative of a mixture of
TMPD†þ and TMPDH22þ. No Raman line of molecu-
lar TMPD was found even for very long time after the
mixture.
Resonance Raman scattering was carried out using
the 632.5 nm exciting laser line within the electronic
absorption of TMPD†þ. The resonance Raman
Fig. 3. Pure DRUV spectra extracted using the SIMPLISMA method of the DRUV data recorded during the sorption of TMPD into
Na3.4(AlO2)3.4(SiO2)92.6: (a) TMPD; (b) radical cation TMPD†þ; (c) ejected electron.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314310
Page 7
spectra obtained several hours to one year after the
mixing of the solids display exclusively the spectral
features of TMPD†þ.
The DRUVv intensities measured at 611 nm as a
function of time for 0.25, 0.5 and 1 TMPD/UC
loadings indicate the TMPD†þ yields. The maximum
of generated TMPD†þ yield is found to be approxi-
mately proportional to the loading value. However,
the ionization yield value deduced from double
integrated EPR signals does not correspond to
Fig. 4. DRUV spectra recorded at room temperature during the course of adsorption of Anth in H3ZSM-5 [H3.4(AlO2)3.4(SiO2)92.6] zeolite by
mixing of powders. (a) 5 min to 48 h (b) 48 h to 6 months.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314 311
Page 8
proportional behaviour (Fig. 7). The increase of
loading generates probably ejected electrons in high
yield. The close proximity of trapped electrons can
lead to antiferromagnetic coupling and induce a
decrease of the spin density.
H6ZSM-5 (Si/Al ¼ 13)—The ionization and pro-
tonation processes which occur upon sorption of
TMPD in H6ZSM-5 were found to be analogous to
those observed for H3ZSM-5. However, the main
difference is the weaker ionization yield as deduced
from UV-visible, EPR and Raman experiments.
3.4. Discussion
EPR, UV-visible and Raman spectra recorded
during the sorption of TMPD in MnZSM-5
(M ¼ Naþ, Hþ; n ¼ 0, 3, 6) zeolites provide evidence
of ionization in all the zeolites under study.
TMPD†þ is produced in low yield in purely
siliceous silicalite-1 whereas TMPD†þ is generated
in high yield in aluminated NanZSM-5. The ejected
electron is also characterized though electronic
absorption spectra. These paramagnetic
species are stabilised in high yield over more
than one month at room temperature. After long
times of exposure, the amounts of TMPD†þ and
trapped electron decrease and indicate charge
recombination to generate occluded molecular
TMPD.
The sorption of TMPD in acidic HnZSM-5
zeolite induces durable charge separation in high
yield for several weeks. After long times of
exposure the amounts of TMPD†þ and trapped
electron decrease and indicate charge recombina-
tion. However, in contrast to non acidic zeolites,
charge recombination generates diprotonated
TMPDH22þ occluded species. The protonation
appears competitive to the ionization efficiency of
HnZSM-5 zeolite. The ionization maximum yield of
TMPD appears somewhat weaker in H6ZSM-5 than
in H3ZSM-5 whereas the protonation was found to
be more rapid.
The ability of acid zeolites to act as single
electron acceptors and to generate spontaneously
significant amounts of radical cation upon sorption
of organic compounds is well documented. So far,
ionization is known to occur through abstraction of
electron by Lewis acid sites in acid zeolites [6].
The electron acceptor site is a prerequisite for the
spontaneous ionization. For instance, it was
reported previously that the yield of ionization of
biphenyl increases gradually with the temperature
Fig. 5. Pure DRUV spectra extracted using the SIMPLISMA method of the DRUV data recorded during the sorption of TMPD into
H3.4(AlO2)3.4(SiO2)92.6: (a) TMPD; (b) radical cation TMPD†þ; (c) ejected electron; (d) diprotonated TMPDH22þ.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314312
Page 9
of thermal treatment under oxygen from 200
to 600 8C and is maximum at high aluminum
content [17].
In the case of aromatic amines with low ionization
potential (6.6 eV) the presence of strong electron
acceptor sites such as Lewis acid sites is not
necessary. It should be noted that Lewis acid sites
were obtained by calcination of acid zeolites under
oxygen. After gentle dehydration of non acidic
NanZSM-5, ionization of TMPD occurs in high
yield in the present work. The strong electrostatic
field in the porous void as well as efficient trapping
sites of radical cation and ejected electron can be
invoked to explain this behavior.
Moreover, TMPD is known to undergo readily
and spontaneously ionization in polar solvents such
as acetonitrile, alcohols and water [8]. Therefore,
there is no doubt that the highly polar environment
constituted by the zeolite framework provides a
suitable environment capable to ionize spon-
taneously in high yield the TMPD molecule.
Analogous behaviour was previously reported [16]
for Na sorption in aluminated zeolite. Upon
sorption, Na atom is ionized producing Naþ ion
and an electron that is trapped in the framework.
This electron absorbs visible light and is respon-
sible for the change of colour. The presence of
Brønsted and Lewis acid sites in HnZSM-5 does
not provide supplementary ionization efficiency but
induces protonation of the amine function. These
protonation properties appear competitive with the
ionization ability.
The tight fit between the rod shape TMPD†þ
radical cation and the pore size of ZSM-5 zeolite is an
important factor responsible for the stabilization
of radical cation. It appears that the presence of
aluminium in the siliceous framework is also a
requirement to prevent rapid electron back transfer.
At this stage of the study, it is possible to give a
reasonable scheme of the sorption of TMPD in ZSM-5
zeolite. The MnZSM-5 induces the uptake of TMPD
molecule at the openings of pores. Simultaneous rapid
ionization occurs through abstraction of electron by
the framework before any protonation by Brønsted
acid sites in acidic zeolites. The sorption goes to
completion by slow diffusion of radical cations and
electron migration within the framework. During the
slow diffusion of radical cation, the uptake of an
electron of the framework induces charge recombina-
tion of molecular sorbate and generates unstable
electron-hole pairs before the rapid electron-hole
annihilation. In the case of acidic HnZSM-5,
the charge recombination is concerted with
the protonation of aromatic amines to provide
quaternary salts of amines.
Acknowledgements
The authors are very grateful to Dr B. Sombret for
assistance and advice while using FT-Raman spec-
trometry. The Centre d’Etudes et de Recherches
Lasers et Applications (CERLA) is supported by the
Ministere charge de la recherche, the region Nord/Pas
Fig. 6. FT-Raman spectra (1064 nm excitation laser line) recorded
during the sorption of TMPD in H3.4ZSM-5 [H3.4(AlO2)3.4
(SiO2)92.6] zeolite by mixing of powders. (a) TMPD at the solid
state (b) 48 h after the mixing of powders; (c) 20 days; (d) 120 days;
(e) 1.5 year; (f) TMPDH22þ recorded in 12 M hydrochloride
solution.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314 313
Page 10
de Calais, and the Fonds Europeen de Developpement
Economique des Regions.
References
[1] V. Ramamurthy, J.V. Caspar, D.R. Corbin, J. Am. Chem. Soc.
113 (1991) 600.
[2] V. Ramamurthy, P. Lakshiminarasimhan, P.C. Grey,
L. Johnston, Chem. Commun. (1998) 2411.
[3] A. Corma, V. Fornes, M.S. Gallatero, H. Garcia, C.J. Gomez-
Garcia, Phys. Chem. Chem. Phys. 3 (2001) 1218.
[4] I. Gener, A. Moissette, C. Bremard, Chem. Commun. (2000)
1563.
[5] J.K. Thomas, Chem. Rev. 93 (1993) 301.
[6] A. Moissette, H. Vezin, I. Gener, J. Patarin, C. Bremard,
Angew. Chem. Int. Ed. 41 (2002) 1241.
[7] A. Moissette, I. Gener, C. Bremard, J. Raman Spectrosc. 33
(2002) 381.
[8] M. Lee, D.J. Jang, D. Kim, S.S. Lee, B.H. Boo, Bull. Korean
Chem. Soc. 13 (1992) 17.
[9] Y. Hirata, A. Nogata, N. Mataga, Chem. Phys. Lett. 189
(1992) 159.
[10] S. Nakamura, N. Kanamaru, S. Nohara, H. Nakamura, Y. Saito,
J. Tanara, M. Sumitani, N. Nakashima, K. Yoshihara, Bull.
Chem. Soc. Jpn 57 (1984) 145.
[11] V. Guichard, A. Bourkba, M.F. Lautie, O. Poizat, Spectro-
chim. Acta 45A (1989) 187.
[12] O. Poizat, A. Bourkba, G. Buntinx, A. Deffontaine,
M. Bridoux, J. Chem. Phys. 87 (1987) 6379.
[13] A. Moissette, I. Gener, C. Bremard, J. Phys. Chem. B 105
(2001) 5647.
[14] W. Windig, J. Guilment, J. Anal. Chem. 63 (1991) 1425.
[15] A. Moissette, Y. Batonneau, C. Bremard, J. Am. Chem. Soc.
123 (2001) 12325.
[16] N.P. Blake, H. Metiu, J. Chem. Phys. 103 (1995) 4455.
[17] I. Gener, A. Moissette, H. Vezin, J. Patarin, C. Bremard, Stud.
Surf. Sc. Cat. 135 (2001) 2272.
Fig. 7. Double integrated EPR signal as a function of time for several loadings: 0.25 (solid line), 0.5 (dashed and dotted line), 1 (dashed line) and
2 (dotted line) TMPD molecules per unit cell.
S. Marquis et al. / Journal of Molecular Structure 651–653 (2003) 305–314314