International Journal of Mass Spectrometry 245 (2005) 61–67

Photocatalytic degradation of methylene blue on nanocrystalline TiO2:

Surface mass spectrometry of reaction intermediates
Hubert Gnaser a,∗ , Michael R. Savina b , Wallis F. Calaway b , C. Emil Tripa b ,
Igor V. Veryovkin b , Michael J. Pellin b
a Department of Physics, University of Kaiserslautern, D-67663 Kaiserslautern, Germany
b Materials Science Division, Argonne National Laboratory, Argonne, IL 60439, USA

Received 25 May 2005; received in revised form 13 July 2005; accepted 13 July 2005
Available online 11 August 2005

Abstract

Photocatalytic degradation reactions of methylene blue on nanocrystalline TiO2 (nc-TiO2 ) films were studied in situ by surface mass spec-
trometric techniques including secondary ion mass spectrometry, laser desorption direct ion mass spectrometry, and laser desorption/laser
photoion mass spectrometry. The parent ion of methylene blue and/or its reduced form leucomethylene blue was observed with little fragmen-
tation by all methods prior to ultraviolet exposure. The surface composition changed upon ultraviolet (UV) irradiation in air, an observation
ascribed to photocatalytic reactions induced by UV photons: the parent molecule signal diminished and intermediate reaction products such
as sulfoxides and sulfones were detected from the TiO2 surface. After prolonged UV irradiation these species also vanished and the methylene
blue appeared to be almost completely mineralized.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Photocatalysis; Nanocrystalline TiO2 ; Surface mass spectrometry; Methylene blue

1. Introduction before recombining, they may reduce or oxidize adsorbed

molecular species. The ultimate goal would be the complete
Nanocrystalline TiO2 (nc-TiO2 ) has attracted great inter- mineralization of any pollutants present. While many studies
est in recent years due to its many potential applications have demonstrated the feasibility of that approach, detailed
[1–5]. Among these, the photocatalytic properties of nc-TiO2 reaction pathways could be established only in a very lim-
clearly stand out: to degrade organic or biological pollutants ited number of cases. To some extent, this is due to the fact
in waste water or in the atmosphere constitutes a very promis- that most of those investigations were carried out in solution.
ing application [6–8]. Numerous studies [9–13] have inves- Under such conditions, the assessment of the reaction pro-
tigated ultraviolet (UV) photocatalytic degradation reactions cesses occurring on the surface of the nc-TiO2 particles is
employing nc-TiO2 particles in aqueous suspensions. The rather restricted.
underlying physico-chemical mechanisms appear to be clear It is generally assumed that the absorption of photons with
cut [14–16]: TiO2 in the anatase phase (which exhibits the an energy hν ≥ Eg = 3.2 eV induces the following reactions at
higher photocatalytic efficiency) is a semiconductor with a the surface of nc-TiO2 :
band gap of Eg ∼ 3.2 eV. A UV photon with an energy greater
Electron–hole pair formation:
than Eg creates an electron–hole pair in a TiO2 nanocrystal-
lite. If these charge carriers diffuse to the crystallite surface TiO2 + hν → TiO2 (e− + h+ ) (1)
Oxidation of adsorbed water by holes:
∗ Corresponding author. Tel.: +49 631 2054038; fax: +49 631 2052854.
E-mail address: [email protected] (H. Gnaser). (H2 O)ads + h+ → H+ + OH• (2)

1387-3806/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijms.2005.07.003
62 H. Gnaser et al. / International Journal of Mass Spectrometry 245 (2005) 61–67

Ionosorption of adsorbed oxygen by electrons: (120 mg/L) and a single drop was applied onto the surface
spreading over the ∼30 mm2 nc-TiO2 films.
(O2 )ads + e− → O2 •− (3) Both pristine and UV-exposed films were investigated by
mass spectrometry. The UV exposures were done in air by
Oxidation of adsorbed hydroxide ions by holes:
means of a xenon–mercury arc lamp (λ = ∼250–380 nm).
2(OH− )ads + h+ → OH− + OH• (4)
2.2. Mass spectrometry
Transient formation of hydroperoxide radicals via Eq. (3)
and the further reactions: Most of the mass spectra were obtained using the
CHARISMA time-of-flight mass spectrometer, which has
O2 •− + H+ → HOO• (5) been described in detail elsewhere [38,39]. Two means of
2HOO• → H2 O2 + O2 (6) desorbing and ionizing material from the nc-TiO2 surface
were used. A pulsed Nd:YAG laser operating in the third
H2 O2 + e− → OH− + OH• (7) harmonic mode (λ = 355 nm, pulse width = 8 ns) focused to a
∼90 ␮m × 150 ␮m elliptical spot (60◦ incidence) was used
Oxidation of organics by OH• radicals, or directly by holes: to desorb and ionize material in what are referred to as direct
ion spectra. Secondary ion mass spectra (SIMS) were col-
R H + OH• → R• + H2 O (8)
lected by bombarding the surface with either a 5 kV Ar+ beam
R + h+ → R+• → degradation products (9) (900 ns pulse, 250 nA, ∼1 mm spot size, 60◦ incidence) or a
15 kV Ga+ beam (200 ns pulse, 20 nA, 10 ␮m spot, 60◦ inci-
The goal of this work was to use mass spectrometry of dence). Repetition rates were 500 Hz for the laser desorption
desorbed surface species to monitor photocatalytic reactions and 1 kHz for the ion beams.
such as those initiated by Eqs. (8) and (9) and followed In addition, mass spectra were obtained using the SPIRIT
by insertion of oxygen occurring on the surface of nc-TiO2 instrument, which has been described in detail elsewhere
particles. Such reactions have been studied in suspensions [40]. This instrument is similar to CHARISMA except that
containing nc-TiO2 particles, but in many applications, nc- the ion extraction optics are more efficient and the desorp-
TiO2 is used in the form of films supported on substrates [17]. tion laser is a nitrogen laser (λ = 337 nm, pulse width = 10 ns,
Hence, a direct probe of photocatalytic processes occurring 6 ␮m spot size, normal incidence). Owing to a lower pulse
on the nc-TiO2 via mass spectrometry of desorbed surface energy, the N2 laser did not directly ionize the desorbed
species is desirable. In such techniques, atomic and molecu- species as the Nd:YAG laser did, rather the desorbed neutrals
lar species are desorbed by means of ion or laser beams and were post-ionized with a F2 laser (λ = 157 nm, 10 ns pulse
detected either directly as ions or as neutrals upon a suitable width).
post-ionization step [18]. Methylene blue (MB), a common
organic dye, was studied because of its ubiquitous use and
because the removal of dyes from waste waters constitutes an 3. Results and discussion
acute problem [19]. Furthermore, extensive research into the
degradation of methylene blue [20–24] and of many other The objectives of the present work are: (i) to explore the
dyes [25–35] in aqueous nc-TiO2 suspensions has already optimal mass spectrometric approach to study photocatalytic
been performed. In the present work, films treated with MB reactions on the surface of nanocrystalline TiO2 films, (ii)
were analyzed before and after UV exposure, and interme- to follow changes in the molecular composition on the TiO2
diate species in the photocatalytic reaction pathway were surface induced by UV irradiation and thereby derive infor-
detected. mation about photocatalytic reactions, and (iii) to determine
whether the laser photons responsible for desorption in the
mass spectrometer might affect the analysis by themselves
2. Experimental inducing photodegradation reactions.

2.1. Preparation of TiO2 films 3.1. Mass spectrometry of methylene blue on TiO2 films

Nanocrystalline TiO2 particles (anatase phase, of either 6, The 5 kV Ar+ SIMS spectrum of MB on nc-TiO2 is shown
12, or 20 nm particles) were suspended in an ethanol/water in Fig. 1. The 15 kV Ga+ SIMS spectrum was nearly identical.
mixture and deposited on glass substrates by spin coat- In addition to a large number of low-mass peaks due either to
ing. Calcination at 500 ◦ C for 1 h resulted in stoichiometric fragmentation or organic contamination, we observe strong
TiO2 films composed of nanocrystallites with an average peaks corresponding to the MB cation at m/z = 284 (structure
size identical to the original powder material, as determined a in Fig. 2). However, the manifold of peaks in the parent
by X-ray diffraction and transmission electron microscopy ion region shows excesses at m/z = 285 and 286 that are not
[36,37]. Methylene blue (chloride) was dissolved in methanol attributable to the isotopomers of MB. Whereas the expected
 H. Gnaser et al. / International Journal of Mass Spectrometry 245 (2005) 61–67 63

was due merely to a lack of oxygen at the surface (typical

chamber pressure = 2 × 10−9 mbar), the chamber was back-
filled with oxygen to a partial pressure of 2.5 × 10−7 mbar
and the laser was left to irradiate the film for 60 min at
500 Hz at a power just below the desorption threshold
(∼5 × 106 W/cm2 ). This corresponds to an oxygen arrival
rate of 0.25 monolayer/s, or 900 monolayers/h, and a UV dose
of ∼7 × 104 J/cm2 over the course of the irradiation. The test
area of the film was analyzed via 15 kV Ga+ SIMS before and
after irradiation. No changes were observed even though the
integrated UV dose was several orders of magnitude higher
than that required to cause pronounced changes in the spec-
trum when the exposure was done with an arc lamp in air (see
below).
While direct ion mass spectra of MB on nc-TiO2 can be
obtained quite easily as shown in Fig. 3, the spectra exhibit
Fig. 1. Secondary ion mass spectrum (5 keV Ar+ ) of methylene blue on a a pronounced dependence on the power of the desorption
nc-TiO2 film. laser. Generally, increasing the laser power increases the sig-
nal levels, but above a certain level fragmentation dominates
ratio of intensities of the peaks at 284, 285, and 286 due to the extent that the MB parent ions all but vanish at high
to MB alone is 1/0.2/0.06, the observed ratio is 1/0.83/0.22, laser powers. The intensities of Ti and of some typical surface
indicating that a significant amount of signal in the m/z = 284 species (like Na and K) strongly increase and approach satu-
manifold (∼40%) is due to [MB + H]+ . The structure of the ration as the laser power is increased. Fig. 4 shows the laser
m/z = 285 ionic moiety is not known, however its neutral pre- power dependence for some selected atomic and molecular
cursor is likely to be leucomethylene blue (LMB, structure b ions desorbed from a MB-covered nc-TiO2 surface. To ensure
in Fig. 2), the reduced form of methylene blue. Whether the high and stable signals for the MB parent ion with as little
LMB was initially present on the surface or was formed as a fragmentation as possible, a laser power in an intermediate
result of the Ar+ bombardment is not known. range (10–20 ␮J/pulse, ∼1–2 × 107 W/cm2 ) was employed
The SIMS spectrum also shows manifolds of peaks due for the laser desorption experiments.
to [MB (CH2 )n ]+ , corresponding to replacement of methyl The mass spectrum of MB laser desorbed from nc-TiO2
groups with protons. These are likely due to fragmentation of and photoionized with a F2 laser (Fig. 5) was similar to the
the parent ion to produce the well-known homologues of MB direct ion spectrum shown in Fig. 3. The base peak is again
such as Azure B (270 amu), Azure A (256 amu), and Azure [LMB]+ , however the main fragment peak corresponding to
C (240 amu). Unlike the parent, these species show strong Azure B is found at [M − 1]+ rather than [M + 1]+ . This is
peaks at [M − 1]+ and [M − 2]+ . likely due to loss of a methyl group from MB, which results
In contrast to the SIMS spectrum, the laser desorption in a neutral molecule which is then ionized by the F2 laser.
direct ion mass spectrum shows [LMB]+ as the dominant The mass spectrum of Ar+ -sputtered neutrals photoionized
peak, with only a very minor amount of [MB]+ present with the F2 laser is identical to the SIMS spectrum shown
(Fig. 3). The m/z = 285 peak manifold has the expected iso- in Fig. 1. Examining the neutral channel via post-ionization
topomer distribution for LMB, and the ratio of LMB to MB for both ion and laser desorption demonstrated that the direct
is ∼9:1. The homologue peaks (Azure A and Azure B) also ion channel was an accurate reflection of the composition
have the [M + H]+ ions as the dominant species. As in the of the surface constituents for the MB/nc-TiO2 system (at
SIMS spectrum, the [M]+ , [M − 1]+ , and [M − 2]+ peaks of least qualitatively). This is an important observation, since the
Azure A and B are also present, but with much less intensity. ionic and neutral channels of desorption are not often similar.
The dominance of the LMB peak in the direct ion spectrum Since neutral channel analysis is significantly more compli-
is striking given that the organic portion of MB is cationic cated than measuring the direct ion channel and because no
and all that is required to obtain it in the mass spectrum is to additional information was revealed from post-ionized neu-
dissociate it from the Cl− ion. LMB also dominates the direct tral analysis, the majority of the data was obtained with direct
ion spectra of MB on gold and titanium metal surfaces, indi- ion measurements once the fact that the neutral and ion chan-
cating that it is not the product of a surface reaction specific nels track each other was proven.
to nc-TiO2 . Whether LMB is present initially or is produced
during the desorption event is unclear. 3.2. TiO2 films with methylene blue upon UV irradiation
Comparison of the SIMS and laser desorption direct ion
spectra shows that the laser does not induce oxidation reac- UV exposure of the MB-covered nc-TiO2 films in air pro-
tions on the surface even though the photon energy (3.49 eV) duced distinct changes in the visual appearance and mass
is above the band gap of TiO2 (3.2 eV). To test whether this spectra of the samples. The MB-treated films were initially
64 H. Gnaser et al. / International Journal of Mass Spectrometry 245 (2005) 61–67

Fig. 2. Some of the molecules detected in the mass spectra: (a) methylene blue cation, m/z = 284 amu, (b) leucomethylene blue, m/z = 285 amu, (c) the
leucomethylene blue sulfoxide, m/z = 301 amu, (d) the leucomethylene blue sulfone, m/z = 317 amu, and proposed structures for ions at m/z = 157 amu (e) and
m/z = 141 amu (f).

dark blue or violet; upon brief UV exposure they turned to a of 30 nm TiO2 particles. There, it was proposed that OH rad-
light blue and were almost white after long exposures. These icals attack the C S+ = C functional group of MB, which at
color changes correlate with changes in the mass spectra. low pH is in direct Coulombic interaction with the anionic
Fig. 6 shows direct ion spectrum of a specimen exposed to TiO2 surface. It was further argued that the first step in MB
UV light for 5 min at a power density of 6.4 mW/cm2 . The degradation is oxygen addition to the C S+ = C functional
parent ion (LMB) and the Azure B fragment have nearly van- group. The resulting sulfoxide, which had also undergone
ished, and the dominant high-mass peaks are at m/z = 301 and a ring-opening reaction at the N heteroatom, was detected
317. These are likely due to oxidation of the parent molecule. at m/z = 303 amu by GC/MS. This ring-opening can occur
Similar species have been reported in recent work [23] on when the attack takes place on the MB cation with the charge
photocatalytic degradation of MB in an aqueous suspension concentrated on the sulfur atom, however if neutral LMB
 H. Gnaser et al. / International Journal of Mass Spectrometry 245 (2005) 61–67 65

Fig. 5. Mass spectrum obtained by laser desorption (λ = 337 nm) and photo-
Fig. 3. Laser desorption direct ion mass spectrum (λ = 355 nm) of methylene ionization (λ = 157 nm) of neutral species from a nc-TiO2 film covered by
blue on a nc-TiO2 film. methylene blue.

is attacked or if the cationic sulfur is coordinated by TiO2

tion mass spectrometry from pure TiO2 films showed [37]
surface oxygen, no such ring-opening is required and the sul-
the abundant desorption of O2 and OH molecules at low
foxide has structure c in Fig. 2 and is detected at m/z = 301.
temperature; that is, of weakly bound species. In the present
The sulfoxide group can react with a second OH radical
experiment, two fragments representing sulfones are detected
to produce a sulfone. This may proceed on LMB sulfoxide
at m/z = 157 and 141, corresponding to structures e and f in
without disrupting the ring structure as shown in structure d in
Fig. 2. These species derive from dissociation of the cen-
Fig. 2, which we detect at m/z = 317. Note that here the sulfur
tral ring after sulfone formation, accompanied by loss of a
atom is in the maximally stable +6 oxidation state. In addi-
dimethylamino group. They may be fragments produced by
tion, oxidative dissociation of the central ring as described in
laser irradiation of the LMB sulfone, or may be the prod-
ref. [23] can result in various sulfone-containing fragments.
ucts of reactions on the nc-TiO2 surface. We favor the later
The sulfones were not detected in [23], but it is well known
interpretation since we do not see laser-generated fragments
[41] that the sulfoxide is easily oxidized into the sulfone in
corresponding to LMB sulfone that has lost one or both of
the presence of hydroperoxides. The latter are thought to be
its dimethylamino groups. These would have appeared at
formed, at least transiently, via the neutralization of O2 − by
m/z = 274 and 231 in Fig. 3.
protons, cf. Eqs. (5)–(7). Although these species cannot be
The UV dose required to bring about these changes was
monitored in the present experiment, recent thermal desorp-
∼2 J/cm2 , or roughly a factor of 104 lower than the dose deliv-
ered by the laser in the earlier experiment in which the mass

Fig. 4. Intensity of various atomic and molecular ions laser desorbed from
a nc-TiO2 film covered by methylene blue as a function of the laser pulse
energy. Apart from K+ () and Ti+ (
), the methylene blue parent ion,
[MB + 1]+ () and two fragment species at m/z = 242 (♦) and 271 () are Fig. 6. Laser desorption direct ion mass spectrum (λ = 355 nm) of methylene
shown. blue on a nc-TiO2 film after UV irradiation for 5 min at 6.4 mW/cm2 .
66 H. Gnaser et al. / International Journal of Mass Spectrometry 245 (2005) 61–67

observed in both spectra. The relative peak heights between

Figs. 7 and 8 are not indicative of the relative amounts of
material present because the laser intensities were not the
same (12 ␮J for Fig. 7 versus 26 ␮J for Fig. 8). Nevertheless
it is clear that there is much less organic material present
on the sample after the high-intensity UV exposure. This
is corroborated by a visual inspection of the sample, which
shows it to be essentially white. We may conclude, there-
fore, that an appreciable amount of the methylene blue has
been mineralized at the nc-TiO2 surface upon UV irradiation.
Because of their volatile nature, the final mineralization prod-
ucts (e.g., CO2 , SO4 , NH4 , and NO3 ) could not be detected
in this experiment; however, the identification of some reac-
tion intermediates by surface mass spectrometry appears to
open an interesting avenue for further studies of photocat-
alytic reactions on nc-TiO2 films.

Fig. 7. Laser desorption direct ion mass spectrum (λ = 355 nm) of methylene
blue on a nc-TiO2 film after UV irradiation for 1 min at 1 W/cm2 . 4. Conclusions

spectrometer chamber was backfilled with oxygen. Given that Photodegradation reactions of methylene blue on
no reaction was observed under high vacuum conditions or nanocrystalline TiO2 films can be studied in situ using mass
with only oxygen present, this implicates the water-derived spectrometric techniques that analyze material desorbed from
OH radical in the reaction mechanism (e.g., reactions (2), (4), the sample surface. Among the methods studied here, laser
and (8)) rather than molecular oxygen or direct oxidation by desorption direct ion mass spectrometry proved to be a facile
holes (reactions (3) and (9)). technique that reveals several reaction intermediates with lit-
For more intense UV exposures the MB parent ion and tle fragmentation, and without the laser probe itself affecting
the Azure B fragment vanish completely. The spectrum of the reaction. It was shown for the MB/nc-TiO2 system that
Fig. 7 was taken from a specimen irradiated for 1 min at a the neutral and ion species observed are qualitatively sim-
power density of 1 W/cm2 (integrated UV dose ∼60 J/cm2 ), ilar. The first products in the oxidation process induced by
and shows the LMB sulfoxide and sulfone peaks, though UV exposure in air revealed by this method are the sulfox-
their relative intensities are reversed. Most of the other peaks ide and sulfone of leucomethylene blue, the reduced form of
originate from the underlying TiO2 film, as can be inferred methylene blue.
from a comparison with Fig. 8, which is a spectrum of an
untreated nc-TiO2 film taken under the same conditions as
Fig. 7 but without the UV exposure. In particular, the peaks Acknowledgements
at the m/z = 81, 105, 125, 139, 153, 165, 181, and 198 amu are
We are grateful to A. Orendorz for providing the nanocrys-
talline TiO2 films. One of the authors (H.G.) thanks Argonne
National Laboratory for financial support. This work was
supported by the U.S. Department of Energy, BES-Materials
Sciences, under Contract W-31-109-ENG-38.

References

[1] M.A. Fox, M.T. Dulay, Chem. Rev. 93 (1993) 341.

[2] M. Anpo, M. Takeuchi, J. Catal. 216 (2003) 503.
[3] A. Hagfeldt, M. Grätzel, Chem. Rev. 95 (1995) 49.
[4] M. Grätzel, Nature 414 (2001) 338.
[5] L.R. Skubal, N.K. Meshkov, M.C. Vogt, J. Photochem. Photobiol.
A: Chem. 148 (2002) 103.
[6] A. Mills, S. LeHunte, J. Photochem. Photobiol. A: Chem. 108 (1997)
1.
[7] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C:
Photochem. Rev. 1 (2000) 1.
[8] H. Gnaser, B. Huber, C. Ziegler, in: H.S. Nalwa (Ed.), Encyclope-
Fig. 8. Laser desorption direct ion mass spectrum (λ = 355 nm) of a pristine dia of Nanoscience and Nanotechnology, vol. 6, American Scientific
nc-TiO2 film. Publishers, Stevenson Ranch, 2004, p. 505.
 H. Gnaser et al. / International Journal of Mass Spectrometry 245 (2005) 61–67 67

[9] M. Schiavello (Ed.), Photocatalysis and Environment, Kluwer Aca- [26] C. Galindo, P. Jacques, A. Kalt, J. Photochem. Photobiol. A: Chem.
demic Publishers, Dordrecht, 1988. 130 (2000) 35.
[10] D.F. Ollis, H. Al-Ekabi (Eds.), Photocatalytic Purification and Treat- [27] M. Vautier, C. Guillard, J.-M. Herrmann, J. Catal. 201 (2001) 46.
ment of Water and Air, Elsevier, Amsterdam, 1993. [28] J. Grzechulska, A.W. Morawski, Appl. Catal. B: Environ. 36 (2002)
[11] P.V. Kamat, Chem. Rev. 93 (1993) 267. 45.
[12] A. Heller, Acc. Chem. Res. 28 (1995) 503. [29] J. Li, C. Chen, J. Zhao, H. Zhu, J. Orthman, Appl. Catal. B: Environ.
[13] J. Peral, X. Domènech, D.F. Ollis, J. Chem. Technol. Biotechnol. 70 37 (2002) 331.
(1997) 117. [30] Z. Sun, Y. Chen, Q. Ke, Y. Yang, J. Yuan, J. Photochem. Photobiol.
[14] A.L. Linsebigler, G. Lu, J.T. Yates Jr., Chem. Rev. 95 (1995) 735. A: Chem. 149 (2002) 169.
[15] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. [31] M. Sökmen, A. Özkan, J. Photochem. Photobiol. A: Chem. 147
Rev. 95 (1995) 69. (2002) 77.
[16] P.V. Kamat, in: J.H. Fendler (Ed.), Nanoparticles and Nanostructured [32] S. Sakthivel, M.V. Shankar, M. Palanichamy, B. Arabindoo, V.
Films, Wiley-VCH, Weinheim, 1998, p. 207. Murugesan, J. Photochem. Photobiol. A: Chem. 148 (2002) 153.
[17] A. Mills, G. Hill, S. Bhopal, I.P. Parkin, S.A. O’Neill, J. Photochem. [33] S. Al-Qaradawi, S.R. Salman, J. Photochem. Photobiol. A: Chem.
Photobiol. A: Chem. 160 (2003) 185. 148 (2002) 161.
[18] J.C. Vickerman, D. Briggs (Eds.), TOF-SIMS: Surface Ananysis by [34] T. Sauer, G. Cesconeto Neto, H.J. José, R.F.P.M. Moreira, J. Pho-
Mass Spectrometry, IM Publications, Chichester, 2001. tochem. Photobiol. A: Chem. 149 (2002) 147.
[19] H. Zollinger (Ed.), Color Chemistry. Synthesis, Properties and Appli- [35] J.-M. Wu, T.-W. Zhang, J. Photochem. Photobiol. A: Chem. 162
cations of Organic Dyes and Pigments, second ed., VCH, Weinheim, (2004) 171.
1991. [36] B. Huber, H. Gnaser, C. Ziegler, Anal. Bioanal. Chem. 375 (2003)
[20] R.W. Matthews, J. Chem. Soc., Faraday Trans. 85 (1989) 1291. 917.
[21] S. Lakshmi, R. Renganathan, S. Fujita, J. Photochem. Photobiol. A: [37] B. Huber, A. Brodyanski, M. Scheib, A. Orendorz, C. Ziegler, H.
Chem. 188 (1995) 163. Gnaser, Thin Solid Films 472 (2005) 114.
[22] T. Zhang, T. Oyama, A. Aoshima, H. Hidaka, J. Zhao, N. Serpone, [38] Z. Ma, R.N. Thompson, K.R. Lykke, M.J. Pellin, A.M. Davis, Rev.
J. Photochem. Photobiol. A: Chem. 140 (2001) 163. Sci. Instrum. 66 (1995) 3168.
[23] A. Houas, H. Lachheb, M. Ksibi, E. Elaloui, C. Guillard, J.-M. [39] M.R. Savina, M.J. Pellin, C.E. Tripa, I.V. Veryovkin, W.F. Calaway,
Herrmann, Appl. Catal. B: Environ. 31 (2001) 145. A.M. Davis, Geochim. Cosmochim. Acta 67 (2003) 3215.
[24] H. Lachheb, E. Puzenat, A. Houas, M. Ksibi, E. Elaloui, C. Guillard, [40] I.V. Veryovkin, W.F. Calaway, J.F. Moore, M.J. Pellin, D.S. Burnett,
J.-M. Herrmann, Appl. Catal. B: Environ. 39 (2002) 75. Nucl. Instrum. Methods B219-B220 (2004) 473.
[25] G. Liu, T. Wu, J. Zhao, H. Hidaka, N. Serpone, Environ. Sci. Tech- [41] H. Beyer, W. Walter, Lehrbuch der Organischen Chemie, Hirzel Ver-
nol. 33 (1999) 2081. lag, Stuttgart, 1998.