Photocatalytic interconversion of nitrogen-containing benzene derivatives

Article abstract of DOI:10.1039/a607883d

The role of the electrons and holes at the surface of semiconductor oxides (TiO₂ and WO₃) in heterogeneous photocatalysis has been investigated in aqueous media for the reactions involving the series: nitrobenzene, nitrosobenzene, phenylhydroxylamine, aniline, and the related compound, 4-nitrosophenol. Qualitative and quantitative evaluation of most intermediates and their time evolution suggest that the reductive pathways are important and even predominant under a variety of experimental conditions. This aspect is not only true at the beginning of the process or for the readily reducible structures, but also during the entire degradation process. Each compound of the series is converted to all the others, even though in widely different amounts. In the early part of the photocatalytic process, with nitrosobenzene and with phenylhydroxylamine, even in the presence of oxygen, the nitrogen substituent undergoes simultaneous oxidation and reduction at comparable rates, so that very little change in the mean oxidation state of the system is observed. This suggests that photogenerated electrons have a controlling role, particularly for some compounds in the early steps of the photocatalytic transformation. For 4-nitrosophenol and p-benzoquinone, in the early steps of degradation the reductive pathways represent the main route, even in the presence of oxygen. As a consequence, for some compounds the presence of an excess of oxygen in the reacting atmosphere decreases the degradation rate, instead of promoting it, as is commonly observed in photocatalysis. It is also remarkable that, for some compounds examined, the redox reactions at the nitrogen-containing substituent have a comparable or even more important role than the hydroxylation of the aromatic ring, which was the predominant degradation pathway for most of the other aromatic compounds.

Full text of DOI:10.1039/a607883d

View Article Online / Journal Homepage / Table of Contents for this issue
Photocatalytic interconversion of nitrogen-containing benzene
derivatives
Paola Piccinini, Claudio Minero, Marco Vincenti and Ezio Pelizzetti*
`
Dipartimento di Chimica Analitica, Universita di T orino, via Pietro Giuria 5, 10125 T orino, Italy
The role of the electrons and holes at the surface of semiconductor oxides (TiO and WO ) in heterogeneous photocatalysis has
2
3
been investigated in aqueous media for the reactions involving the series: nitrobenzene, nitrosobenzene, phenylhydroxylamine,
aniline, and the related compound, 4-nitrosophenol. Qualitative and quantitative evaluation of most intermediates and their time
evolution suggest that the reductive pathways are important and even predominant under a variety of experimental conditions.
This aspect is not only true at the beginning of the process or for the readily reducible structures, but also during the entire
degradation process.
Each compound of the series is converted to all the others, even though in widely di†erent amounts. In the early part of the
photocatalytic process, with nitrosobenzene and with phenylhydroxylamine, even in the presence of oxygen, the nitrogen substit-
uent undergoes simultaneous oxidation and reduction at comparable rates, so that very little change in the mean oxidation state
of the system is observed. This suggests that photogenerated electrons have a controlling role, particularly for some compounds
in the early steps of the photocatalytic transformation. For 4-nitrosophenol and p-benzoquinone, in the early steps of degradation
the reductive pathways represent the main route, even in the presence of oxygen. As a consequence, for some compounds the
presence of an excess of oxygen in the reacting atmosphere decreases the degradation rate, instead of promoting it, as is com-
monly observed in photocatalysis. It is also remarkable that, for some compounds examined, the redox reactions at the nitrogen-
containing substituent have a comparable or even more important role than the hydroxylation of the aromatic ring, which was
the predominant degradation pathway for most of the other aromatic compounds.
attention is addressed to the transformation of the nitrogen
functional groups and to the interconversion of the com-
Introduction
Light-induced electron-transfer reactions on semiconductor
oxides represent a very active research area for the implica-
tions that these reactions have on environmental processes
and on the construction of devices for energy production. Het-
erogeneous photocatalytic processes have proved capable of
converting solar energy into chemical or electrical energy, pro-
vided new synthetic routes to organic chemistry and have
o†ered valuable alternatives to conventional methods for the
removal of organic pollutants from waste- and ground water.1
In the last domain, titanium dioxide-mediated photo-
catalytic processes in aerated aqueous slurries were shown to
attack even highly refractory compounds, which could be
totally converted into carbon dioxide and inorganic ions. The
possible mechanisms for the overall oxidation process have
been considered in detail, although some questions remain
unanswered.1h6
Organic compounds containing nitrogen atoms are
extremely common in nature (amino acids and proteins) and
several classes of man-made substances of environmental
concern have nitrogen atoms in their structure (herbicides and
pesticides, drugs, explosives, dyes etc.). As a consequence,
increasing attention has been addressed to the photocatalytic
degradation mechanism of nitrogen-containing functional
groups and the ultimate fate of the nitrogen moiety.7h16 The
reduction of substituted nitrobenzenes has been carefully
investigated, both for its ecotoxicological importance in the
subsurface environment and for the treatment of hazardous
wastes and contaminated soils.17h19
pounds in the series as a function of experimental conditions.
Investigation of further degradation steps leading to mineral-
ization into carbon dioxide and inorganic nitrogen com-
pounds will be presented in
a
forthcoming study.
Nitrobenzene has been studied in detail recently,12,13 so only
a few further experiments involving it are presented here as a
basis for comparison with the other substrates.
Experimental
Reagents
Nitrobenzene (NB), o-, m- and p-nitrophenols (NP), nitroso-
benzene (NOB), 4-nitrosophenol (4-NOP), aniline (A), o-, m-
and p-aminophenols (AP), hydroquinone (HQ), p-benzo-
quinone (BQ) and hydroxylamine were purchased from
Aldrich (Steinheim, Germany) at the highest purity available
and were used without further puriÐcation. Phenylhydroxyl-
amine (PHA) was synthesized by reduction of nitrobenzene
with zinc dust20 and recrystallized twice from cold water.
Titanium dioxide TiO P25 (BET area 55 m2 g~1, Degussa
2
A. G., Frankfurt, Germany) was irradiated in aerated aqueous
suspension for at least 12 h and washed with deionized, bidis-
tilled water in order to eliminate both organic impurities and
inorganic ions possibly adsorbed on the photocatalyst. Titania
loaded with 1% of platinum (Pt0) was prepared by photo-
catalytic reduction of hexachloroplatinate ions. The aqueous
TiO slurry was added to Na PtCl (Aldrich) and illuminated
The reaction pathways by which nitrobenzene, nitroso-
benzene, phenylhydroxylamine, aniline and a few related
derivatives are transformed in oxygenated and deoxygenated
aqueous solution by the action of irradiated semiconductors
have been investigated and, in the present report, particular
2
2
6
until H was detected. The powder was recovered and washed
2
with deionized, bidistilled water. WO of BET surface area 1
m2 g~1 21 was purchased from Ventron.
3
J. Chem. Soc., Faraday T rans., 1997, 93(10), 1993È2000
1993

View Article Online
Irradiation experiments
selection was based on maximum analytical sensitivity. Under
these conditions, retention times were (in parentheses, min):
NB (7.91), 2-NP (6.34), 3-NP (3.98), 4-NP (3.64), NOB (9.04),
Suspensions of the catalyst were prepared at di†erent concen-
trations and then mixed with aqueous solutions of the organic
substrates directly in the cells used for the irradiation experi-
ments (Ðnal volume, 5 ml). Typical concentrations were 10 mg
4
4
-NOP (4.33), PHA (4.55), A (3.28), 2-AP (7.30), 3-AP (5.80),
-AP (4.40), BQ (4.35), HQ (3.48), respectively.
l~1 (10 ppm) of organics and 200 mg l~1 (ppm) of TiO . The
2
cells containing the reaction slurry were kept in a water bath
in the dark until they reached the working temperature of the
lamp housing (ca. 50 ¡C). Irradiation was carried out in cylin-
drical Pyrex glass cells (4.0 cm diameter, 2.5 cm height), con-
taining 5 ml of aqueous suspension and the desired amount of
photocatalyst and substrate. Irradiation conditions have been
described elsewhere.22,23 Total photon Ñux (340È400 nm) in
the cell was 1.35 ] 10~5 ein min~1 (1 ein \ 1 mole of
photons). At Ðxed time intervals the irradiation was stopped,
the whole sample was Ðltered through 0.45 lm cellulose
acetate membranes (Millipore HA), and the Ðltrates were
analysed without further clean-up. In the experiments carried
out in the absence of oxygen, the cells were purged with
helium for 0.5 h before the organic substance was added and
subsequently irradiated.
Results and Discussion
Primary events and reactive species at irradiated TiO
2
The reactions and species, occurring at the surface of irradi-
ated semiconductors, have been discussed in detail else-
where.1h6 In the forthcoming paragraphs reactions of
hydroxyl radical and aquated electrons in aqueous homoge-
neous solutions are referred to for comparison. However, het-
erogeneous photocatalytic processes involve reactions at the
surface/solution interfaces where (i) the oxidizing species can
be either holes (more oxidizing than aqueous hydroxyl
radical)25 or trapped holes (i.e. ~OH , less oxidizing than
aqueous hydroxyl radical); (ii) the reducing species can be
ads
conduction band electrons or trapped electrons, both being
much less reducing than aqueous hydrated electrons. For the
sake of simplicity, the photocatalytic process is usually
depicted by considering only the main processes, where the
organic substrates (initial and intermediates) can react
through direct interaction with the carriers and with the
trapped holes (hereafter called ~OH radicals). Photogenerated
electrons and hydroxyl radicals concurrently promote the
reduction and the oxidation of the nitrogen-containing group
and induce the hydroxylation of the aromatic moiety. Oxida-
tive steps are presented both as hydrogen abstraction and OH
addition, but these are only symbolic representations of the
actual two-step mechanisms involving surface-bound species.
However, besides the solvent, other species present at the
interface or in solution (e.g. HO ~/O ~~, O , H O ) may par-
Analytical procedures
The aqueous solutions obtained upon Ðltration of the reaction
slurry were extracted three times with CH Cl . The organic
2
2
extracts were concentrated to small volume under vacuum
and then to ca. 300 ll in conical vials under a stream of argon.
The concentrated extracts were analysed by gas
chromatographyÈmass spectrometry (GCÈMS) in electron
impact (EI) mode, from which the semivolatile products were
identiÐed. Qualitative and quantitative determinations were
carried out using authentic standards. Benzoquinonemono-
imine, hydroxybenzoquinonemonoimine and hydroxy-
benzoquinone were identiÐed only on the basis of their mass
spectra. Aniline and hydroxylamine were determined using
GCÈMS, after derivatization with hexylchloroformate.24
GCÈMS experiments were performed on a Finnigan-MAT 95
Q hybrid instrument, with magnetic, electrostatic and quadru-
pole analysers. Ions were collected at the Ðrst detector, located
after the electrostatic sector. Samples were injected splitless
into a Varian 3400 gas chromatograph equipped with a J&W
DB-5MS 30 m ] 0.25 mm capillary column, 0.25 lm Ðlm
thickness. The oven temperature was programmed as follows:
isothermal at 50 ¡C for 3 min, from 50 ¡C to 300 ¡C at 12 ¡C
min~1, isothermal at 300 ¡C for 10 min.
2
2
2
2 2
ticipate in the degradation scheme leading Ðnally to mineral-
ization of the organics.
All the starting structures involved, and the initial interme-
diates, detected by GCÈMS and HPLC are considered in
Scheme 1 that will be discussed in detail later. Although not
quantiÐed precisely, benzoquinone, benzoquinonemonoimine,
hydroxybenzoquinonemonoimine and hydroxybenzoquinone
were clearly identiÐed by GCÈMS.
Qualitative and quantitative determinations of initial com-
pounds and reaction intermediates were carried out by high-
performance liquid chromatography (HPLC) using authentic
standards. Determinations were performed by UV detection
Interconversion of the functional groups
Nitrosobenzene. In the investigations reported pre-
viously,12,13 only small amounts of reduced species (NOB,
PHA, A) were found to be produced from the photocatalytic
degradation of nitrobenzene and 4-nitrophenol (4-NP).
During the degradation of 4-NP, 4-aminophenol was produc-
ed at a maximum concentration of 2% of the initial substrate
concentration (after 2 min irradiation). The high degradation
rate constants found for NOB and PHA in the present study
under identical conditions justify such low concentrations of
reduced compounds. High rate constants for the decomposi-
tion of the reduced species can account for the attainment of
low steady-state concentrations in the degradation of NB and
(
model L6200 and model 6000 Hitachi-Merck pumps with
model L4200 UV) and using an RP-18 column (Lichrocart 5
lm, 150 mm long). Di†erent eluents were used, depending on
the analytes under study: acetonitrileÈphosphate bu†er
5
] 10~2 M, pH 2.8 (eluent A); acetonitrileÈ1-hexanesulfonic
acid 1 ] 10~2 M in phosphate bu†er 5 ] 10~2 M, pH 2.8
eluent B). The bu†er was used for eluting acid compounds in
(
their undissociated form and to avoid peak tailing. Solutions
of pure authentic standards were used for quantitative assess-
ment, whenever commercially available.
Nitrobenzene and nitrophenols were detected at 210 nm
using either eluent A or B at 35/65 v/v ratio; nitrosobenzene
was determined at 307 nm using the same eluents; 4-
nitrosophenol was also detected at 307 nm using both eluent
A and B at 13/87 v/v ratio; phenylhydroxylamine and aniline
at 197 nm using eluent B at 14/86 v/v ratio; 2- and 3-
aminophenol were determined at 208 nm using eluent B at
4-nitrophenol, even in the case when reduction is the major
degradation path.
The photolysis of NOB has been reported in non-aqueous
solvents (freon,26 ethanol27), assuming, for the initial step:
hl
C H NO ÈÈÈ
Õ
C H ~ ] NO
(1)
6
5
6 5
7/93 v/v ratio; 4-aminophenol at 208 nm with eluent B at 5/95
v/v ratio; BQ at 246 nm using eluent A at 10/90 v/v ratio; HQ
at 195 nm using eluent A at 5/95 v/v ratio. The wavelength
Owing to the small amount of phenol (\0.01%) found in
the present experiments and the slow degradation of NOB in
1994
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Scheme 1 Early steps of oxidative and reductive photocatalytic pathways proposed for degradation of NB, NOB, PHA, A, NOP, NP, AP, BQ
and HQ. Only the para-isomer is shown for clarity (when applicable). The ring hydroxylation is illustrated by two consecutive OH attacks,
initially leading to H abstraction and subsequently to OH addition onto the radical formed.
the absence of the photocatalyst (15% of NOB was decom-
posed after 5 h irradiation), the contribution of direct photoly-
sis [eqn. (1)] appears to be modest in the overall degradation
process.
diation, for ca. 80% of the organic matter present in the
system.
The analysis of product evolution at short irradiation times
under photocatalytic conditions reveals that a large part of
the photocatalytic reactions takes place at the nitrogen-
In the presence of a small amount of TiO (200 mg l~1) and
2
UV light, nitrosobenzene disappears very rapidly (see Fig. 1).
Table 1 reports the NOB decay rate constants as a function of
both the catalyst and initial substrate concentrations. The
behaviour is saturative, as usually reported for photocatalytic
transformations. As proposed and veriÐed recently,28,29 the
saturative behaviour as a function of catalyst and substrate
concentration can be ascribed to primary photochemical pro-
cesses, independent of the adsorption of substrates. The e†ects
of temperature, pH, catalyst loading and composition of the
gas phase on the decay rate constant are reported in Table 2.
Fig. 1 shows the evolution of major and minor detected inter-
mediates corresponding to NOB degradation. Although not
comprehensive, the species reported account, after 1 min irra-
Table 1 Monoexponential decay rate constant (min~1) for di†erent
concentrations of NOB and TiO under air-saturated conditions
2
NOB concentration
TiO
2
/
mg l~1 9.3]10~5 M 1.9]10~4 M 4.7]10~4 M 9.3]10~4 M
50
00
00
00
0.66 ^ 0.07
1.21 ^ 0.10
1.39 ^ 0.18
1.39 ^ 0.09
1.49 ^ 0.11
È
È
È
1
2
5
È
0.81 ^ 0.03
È
È
0.39 ^ 0.02
È
È
0.24 ^ 0.01
È
1
000
È
È
È
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
1995

View Article Online
NB and further oxidation of the latter appears to be the most
favourable. However, the direct hydroxylation of the NOB
aromatic ring, yielding three nitrosophenol isomers, is also
likely to represent an important competing process, as sug-
gested by the detection of 4-NOP, despite its fast degradation
kinetics (see later) and its consequent low tendency to accu-
mulate.
The degradation route represented by the oxidation of the
functional group (from nitroso to nitro) is peculiar to this class
of compounds, since previous investigations on benzene deriv-
atives indicated the hydroxylation of the aromatic ring to be
the most important process. It is worth noting that exposure
of dilute NOB solution to 60Co c-rays under N O atmosphere
2
led to nitrobenzene formation.30 It has been reported that,
under pulse radiolytic conditions, ~OH radicals in homoge-
neous solutions give C H NO ~~ from NOB (k \ 1 ] 1010 l
6
5
2
mol~1 s~1).30 This radical (radical R in Scheme 1) is a weak
1
acid (pK \ 3.2). Further reactions of this radical in homoge-
a
neous solution are either: (i) direct e~ abstraction by ~OH
radical to give NB; (ii) disproportionation in the presence of
NOB30 to produce NB and nitrosobenzene radical
Fig. 1 Photocatalytic degradation of NOB (9.3 ] 10~5 M) on TiO
2
(
C H NOH~, a very weak acid with pK \ 11.7, reported in
(
200 mg l~1) at pH 5.8 as a function of irradiation time: (A) NOB and
6
5
a
time evolution of the main intermediates, NB, PHA, 4-AP and A; (B)
time evolution of minor intermediates, NP, 4-NOP
Scheme 1 as R );30 radical R can, in turn, (iii) undergo elec-
2
2
tron exchange with C H NO ~~ (R ) to give NB and PHA
k \ 5 ] 105 l mol~1 s~1 for the dismutation)30 or (iv) directly
6
5
2
1
(
acquire an electron to give PHA.
The reductive reactions of NOB are reported in the lower
part of Scheme 1. NOB reacts with electrons in homogeneous
solution at a rate slightly higher than with ~OH radicals
containing substituents and involves e~ and h` in comparable
extents. For NCB degradation, NB and PHA (peaks after ca.
1
min) account for ca. 25% each of the initial concentration of
(
k \ 4 ] 1010 l mol~1 s~1)30 giving the nitrosobenzene
NOB (75% degraded). The symbolic reaction (2) exempliÐes
this reaction course.
radical (C H NOH~, R ). Further reduction of this radical by
6
5
2
electron capture and charge neutralization (addition of one
proton) produces PHA. As cited above, radical R2
can
`
H2O
2
C H NO ÈÈÈ
Õ
C H NO ] C H NHOH
6 5 6 5
(2)
6
5
2
exchange electrons with other species in solution, whose con-
The average oxidation number of nitrogen, in the organic
compounds present in the reaction mixture, as a function of
irradiation time, is quite signiÐcant. The value of the average
nitrogen oxidation number (by assuming ]3 for NB, ]1 for
NOB, [1 for PHA, [3 for A) is varying in the range ^0.5 in
the initial 30 min of reaction.
These considerations and the nature of the intermediates
detected suggest the occurrence of the reactions depicted in
the upper part of Scheme 1 as viable early steps in the oxida-
tive degradation process. Among the possible oxidative routes
accounting for NOB degradation, the conversion of NOB to
centrations possibly inÑuence the reaction kinetics.
In the electrochemical reduction of NOB pH appears to be
the most important factor controlling the route of reduction,
although the products may be inÑuenced by the nature of the
cathode.31 In weakly acidic media, as in the present experi-
ments, aromatic hydroxylamines and nitroso components are
formed. However, the aromatic hydroxylamines are not
readily reduced (at least at the metallic cathode) and may
rearrange to p-AP (see Scheme 1). PHA and 4-AP give rise to
formation of quinonemonoimine or to further hydroxylation,
oxidation and ring opening.
Table 2 Monoexponential decay rate constants (min~1) for NOB (9.3 ] 10~5 M), PHA (9.2 ] 10~5 M) and A (1.1 ] 10~4 M) in the presence of
TiO (200 mg L~1) as a function of di†erent temperatures, gaseous phases and pH
2
temperature
compound
NOB
20 ¡C
35 ¡C
50 ¡C
0.73 ^ 0.07
0.80 ^ 0.07
gas phase
0.90 ^ 0.17
He
0.1 M tert-
butanol
]
air
O2
He
NOB
PHA
1.39 ^ 0.18
0.88 ^ 0.08
0.86 ^ 0.05
2.45 ^ 0.40
0.12 ^ 0.02
3.04 ^ 0.33
0.43 ^ 0.03
È
pH
3.2
5.8
6.3
11.1
NOB
A
0.99 ^ 0.11
0.12 ^ 0.01
1.39 ^ 0.18
È
1.03 ^ 0.17
0.45 ^ 0.01
È
0.15 ^ 0.01
The compound concentration was followed as a function of the irradiation time. The photocatalytic disappearance of the starting materials can
be reasonably represented by pseudo-Ðrst-order kinetics:25 C \ C exp ([kt). The reported Ðrst-order degradation constants refer to experiments
0
in which the decrease of C with time followed monoexponential decay law with good correlation at least to C /2e.
0
1996
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Apart from the preliminary transformations of the sub-
strates, mainly involving their functional group, the progres-
sive attack of the aromatic ring of NB,13 phenol,32 NOPs,
PHA and 4-AP by means of photocatalytic reactions is indi-
cated by the detection of several hydroxylic and quinonoid
derivatives.
phenylhydroxylamine reached ca. 40% of the initial NOB
concentration after 30 s irradiation and rapidly converted to
aniline, which became the major product reaching 55% after 6
min. As in the case of Fig. 2(A), the rate of NOB disap-
pearance was very close to the initial rate of PHA formation.
These results can be rationalized by considering that the con-
duction band electrons are completely available for reacting
with the organic substrate, while ~OH radicals are trapped by
tert-butanol.
In the present photocatalytic study, the change from air
atmosphere to pure oxygen decreased the NOB disappearance
rate (see Table 2), contrasting with the expectation for an
oxidation-driven degradation process. The slight detrimental
e†ect on the rate of NOB disappearance is probably due to
the competition of oxygen with organics for reduction by con-
duction band electrons. These results are quite unusual in
photocatalysis, where a strong direct dependence of the
photocatalytic degradation rate on the oxygen partial pressure
has been reported for most compounds, as a result of the
prevalence of oxidative processes.33
The deposition of Pt0 onto TiO [Fig. 3(A)] slightly
2
decreases the NOB degradation rate (k \ 0.86 ^ 0.08 min~1)
and modiÐes the product distribution for reduced species,
such as PHA, and the time evolution of nitrophenols. The
platinum islands were reported to act as pools of electrons.35
For example, they lead to hydrogen evolution in the so-called
“sacriÐcial systemsÏ for photocatalytic water cleavage.36 They
also produced a more efficient reduction of oxygen.37,38 This
e†ect decreases the availability of electrons for the reduction
of the organic compounds.
In our experiments, when He was substituted for O (Fig.
2
2
), the initial NOB disappearance increased substantially
(
Table 2), and PHA became by far the major product, with
The temperature at which the photocatalytic degradation
was carried out showed little e†ect on the degradation rate of
NOB (Table 2). While some authors have obtained data
exhibiting an Arrhenius type behaviour and activation ener-
gies ranging from 10 to 16 kJ mol~1,6 other studies report
very low values of activation energy and even a decrease in
the disappearance rate with the increase of the temperature.39
These activation energies are comparable with those measured
almost 70% yield after 2 min irradiation. Interestingly, the
rate of NOB disappearance is very close to the initial PHA
formation rate. NB was formed to a much lower extent (only
3
%) than in air-saturated conditions, while aniline tended to
accumulate slowly. The absence of an antagonist for electrons,
such as O , increased the reduction pathway rate, as indicated
by the high PHA concentration. However, the comparison
2
between Fig. 1(A) and Fig. 2(A) indicates that holes (h`) or
for the reactions of both ~OH radicals and e ~ with several
aq
trapped holes (formally ~OH ), are both kinetically inade-
organic substrates in homogeneous phase. It has been report-
ads
quate oxidizers for NOB, while O molecules may participate
ed that E ranges from 11 to 18 kJ mol~1 for the reaction of
2
act
in the initial oxidation step as photoadsorbed species accord-
e ~ with nitrobenzene and 4-nitrophenol, and from 5 to 10 kJ
aq
ing to the following reaction:
mol~1 for the reaction of ~OH with formate and 2-methyl
propan-2-ol.40
R-NO ] O
] 2H` ] 2e~ ] R-NO ] H O (3)
CB
2(ads)
2
2
A variation in pH is likely to a†ect both the charge of the
This kinetic inhibition (see Table 2) is less pronounced for
PHA (which is more easily oxidized than NOB, see later),
although in this case also the absence of oxygen makes the
reductive pathways more important than the oxidative ones.
The scavenging of holes or trapped holes may proceed
through the formation of surface peroxo-species34 or by
hydroxylation of the PHA aromatic ring (accounting for 20%
of the organic carbon content not detected).
TiO surface and the speciation of the organic substrate. In
2
the case of NOB, no major e†ect is expected because of the
aprotic character of the organic (in the investigated pH range).
Although TiO P25 has a pH of 6.25,34 no dramatic varia-
2
zc
tion of the NOB degradation rate is observed in the pH 3È11
range (Table 2). This is the usual behaviour reported for
neutral molecules.
The photocatalytic activity of another semiconductor oxide,
WO ,41 was compared with TiO . WO presents a ]0.3 V
The relative signiÐcance of reduction pathways (see Scheme
3
2
3
1) is enhanced by the presence of an ~OH scavenger such as
increment in the reduction potential of the conduction band
tert-butanol [Fig. 2(B)]. Under these experimental conditions,
nitrobenzene was not produced during the degradation while
Fig. 2 Disappearance of NOB and formation of intermediates in the
photocatalytic degradation carried out (A) under helium atmosphere
and (B) under helium atmosphere and in the presence of 0.1 M tert-
butanol (other conditions as in Fig. 1)
Fig. 3 Photocatalytic degradation of NOB (9.3 ] 10~5 M at pH 5.8)
and time evolution of intermediates: NB, PHA, 4-AP, A, 4-NOP and
NP: (A) on TiO loaded with 1% of Pt (200 mg l~1) and (B) on WO
2
3
(1 g l~1)
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
1997

View Article Online
with respect to TiO . The degradation rate of NOB in the
minimum value when both electron and hole scavengers are
absent. This observation can be easily rationalized by the
kinetic model proposed recently.28 From these results we can
also deduce that oxygen plays an important role, not only as
an electron scavenger, but also in the oxidative chain
(reactions not reported in Scheme 1).
The degradation pathways of PHA are depicted within the
general Scheme 1. The initial oxidative attack on the molecule
could either involve the nitrogen-containing group, giving
nitrosobenzene (see the preceding discussion), or the aromatic
ring, giving hydroxyphenylhydroxylamine isomers. The detec-
tion of these species is problematic as discussed previously.
When the oxidative path is limited by tert-butanol scavenging,
only the reduction of the nitrogen-containing group remains
accessible, and consequently the overall degradation rate is
decreased.
2
presence of 1 g l~1 WO is slower [see Fig. 3(B)]. As for tita-
3
nium dioxide, NB and PHA are the most important interme-
diates, reaching 20% and 35% of the initial NOB
concentration, respectively. Despite the di†erent reduction
potential of the conduction band, the reductive pathways still
play the predominant role, as proved by large PHA forma-
tion. 4-NOP, 2- and 4-NP are formed in lower abundances,
ranging from 2% to 7% of the initial NOB concentration.
Phenylhydroxylamine. The photodegradation of phenylhy-
droxylamine occurs by both oxidative and reductive pathways
to comparable extents. This deduction is based on the observ-
ation [Fig. 4(A)] that oxidation and reduction intermediates
are formed approximately in the same amount ([4-AP] ] [A]
vs. [NOB] ] [NOP]).
Under He atmosphere [Fig. 4(B)], the PHA degradation
rate decreased considerably in comparison to the standard
conditions (Table 2). Concurrently, aniline formation
increased considerably, reaching 15% instead of 1% as in the
presence of oxygen, while 4-aminophenol became the pre-
dominant intermediate reaching ca. 30% with respect to the
initial PHA concentration after 15 min. Conversely, the abun-
dance of oxidized products declined. Nitrosobenzene concen-
tration rose to 15%, as in air, but after 15 min irradiation
reached a constant 8% value, while it kept decreasing in air
atmosphere until it disappeared. Nitrobenzene showed very
similar behaviour.
The addition of tert-butanol [Fig. 4(C)] consistently
favoured the formation of aniline (ca. 80%). Nitrobenzene and
nitrosobenzene reached only 5% of the initial PHA concentra-
tion. The degradation rate constant for PHA (see Table 2) was
intermediate between those obtained in air and in helium
atmosphere. Note that the degradation rate reached its
Aniline. Aniline proved much more stable than the other
substrates studied so far. In fact, it was virtually unaltered
when irradiated in the absence of the catalyst, while it
degraded slowly under photocatalytic conditions (Table 2).
During photodegradation, 2-aminophenol was the predomi-
nant intermediate [Fig. 5(A)] reaching a maximum of 10%,
with respect to the initial aniline concentration, after 3 min.
Nitrobenzene and 4-aminophenol were formed at trace level.
The e†ect of pH on the degradation rate constant is di†er-
ent from that observed for NOB (Table 2), as the rate
increased in moving to more basic solutions. Since, in aniline
phototransformation, the oxidative pathways are predomi-
nant, an improved transfer of photogenerated holes to OH~,
adsorbed at an increased density onto the surface, can account
for the results. The primary oxidation products, amino-
phenols, can undergo further oxidation at the nitrogen-
containing group or ring hydroxylation (Scheme 1). As
already found in the preceding experiments on phenylhy-
droxylamine, very low amounts of aminophenols are
observed, conÐrming that these species are extremely reactive
under photocatalytic conditions.
E†ect of organic compound structure: the 4-nitrosophenol case
To assess the inÑuence of related structures on the photo-
catalytic transformation pathways, the process involving 4-
nitrosophenol has been investigated. Under the conditions
typically adopted in our experiments, direct photolysis halves
Fig. 5 (A) Disappearance of A (1.1 ] 10~4 M) and formation of inter-
Fig. 4 Photocatalytic degradation of PHA (9.2 ] 10~5 M, 200 mg
mediates in the photocatalytic degradation in the presence of 200 mg
l~1 of TiO at pH 5.8) and formation of intermediates (A) under air,
l~1 TiO at pH 6.3. (B) Disappearance of 4-NOP (4.4 ] 10~5 M) and
2
2
(
B) under He atmosphere and (C) under He and in the presence of 0.1
formation of intermediates in the photocatalytic degradation in the
M tert-butanol
presence of 10 mg l~1 of TiO at pH 5.8.
2
1998 J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Table 3 Monoexponential decay rate constants (min~1) for the degradation of BQ (9.2 ] 10~5 M), HQ (9.2 ] 10~5 M) and BQ in the presence of
HQ (4.6 ] 10~5 M) using TiO at di†erent concentrations as catalyst
2
TiO2 (200 ppm)
TiO (50 ppm)
0.1 M tert-
butanol
0.1 M
NaCl
2
in water
in water
BQ
HQ
BQ (]HQ)
2.07 ^ 0.20
4.43 ^ 0.44
0.18 ^ 0.02
3.17 ^ 0.23
2.71 ^ 0.29
1.28 ^ 0.08
È
È
È
È
È
È
4-NOP concentration after 16 min and under photocatalytic
conditions 4-NOP disappears in less than 1 min. To follow
the photocatalytic degradation, the catalyst concentration had
to be drastically diminished; in the presence of as little as 10
mg l~1 of TiO the half-life of 4-NOP is reduced to ca. 4 min
2
[
k \ 0.19 ^ 0.03 min~1, see Fig. 5(B)]. The detected interme-
diates are 4-aminophenol (4-AP), HQ and 4-nitrophenol (4-
NP). These species, together with 4-NOP, account for 97% of
the initial organic content at 5 min of irradiation. Traces of
phenol, BQ and quinonemonoimine were also detected. Their
involvement in the degradation mechanism will be discussed
later. From the detected intermediates, the reductive and oxi-
dative pathways depicted in Scheme 1 can be proposed for the
degradation process. 4-Hydroxyphenylhydroxylamine was not
available as an authentic standard nor was clear evidence of
its presence found, possibly because of its labile structure. If
this species was formed, as our attribution of HPLC peaks
suggests, its maximum concentration did not exceed some
2
È3% of the initial NOP amount. The same applies also to
Fig. 6 Disappearance of BQ (9.2 ] 10~5 M) and time evolution of
dihydroxynitrosophenol isomers [through path (c)], which are
evident in HPLC proÐles, but difficult to quantify. The
reduction process yielding 4-AP represents the most impor-
tant degradation path for NOP. Oxidative paths lead to the
formation of hydroquinone and, in lower abundance, to 4-NP
HQ during the photocatalytic degradation (A) on TiO (200 mg l~1)
2
at pH 5.8; (B) on TiO (200 mg l~1) at pH 5.8 in the presence of
2
0
.01 M tert-butanol
[
path (a)]. However, paths (b) and (c) are likely to account for
These results support the prominent role of reductive paths in
the degradation of compounds which may exist in a quino-
noid tautomeric form.
small fractions of the overall NOP degradation.
In aqueous solution, 4-NOP is present principally in the
benzoquinonemonoxime tautomeric form.42 From our data, it
appears that this structure can be more easily reduced than
oxidized under photocatalytic conditions. This assumption
explains both the predominant production of ammonium ions
The authors are very grateful for the Ðnancial support of
CIEMAT, MURST, CNR and The Project “Sistema Lagunare
VenezianoÏ.
(
forthcoming paper) and the very low concentration of 4-
nitrophenol formed, in comparison with the much more rele-
vant production of nitrobenzene from the NOB degradation.
To check the importance of the reductive processes further,
the transformation process of BQ was also investigated (see
Fig. 6). HQ was formed with an 80% yield (with respect to the
initial BQ concentration) after 1 min irradiation. BQ half-life
decay in these conditions was as low as 0.25 min (see Table 3).
When the experiment was carried out in the presence of 0.01 M
tert-butanol, the same amount of HQ was formed, but con-
secutive HQ depletion was signiÐcantly reduced. Similar
results were found when 0.1 M sodium chloride was added to
act as a hole scavenger.43 In separate experiments on HQ and
phenol, HQ was mainly degraded by oxidative processes,
yielding trihydroxybenzenes and products of ring opening, but
only traces of BQ. This implies that the scavenging of ~OH
radical decreases the rate of HQ disappearance, as here
reported.
References
1
Photocatalysis. Fundamentals and Applications, ed. N. Serpone
and E. Pelizzetti, Wiley, New York, 1989.
2
3
C. S. Turchi and D. F. Ollis, J. Catal., 1990, 122, 178.
N. Serpone, E. Pelizzetti and H. Hidaka, in Photochemical and
Photoelectrochemical Conversion and Storage of Solar Energy, ed.
Z. W. Tian and Y. Cao, Int. Academic Publishers, Beijing, 1993,
p. 33.
4
M. R. Ho†mann, S. T. Martin, W. Choi and D. W. Bahnemann,
Chem. Rev., 1995, 95, 69.
A. Linsebigler, G. Lu and J. T. Yates, Chem. Rev., 1995, 95, 735.
D. F. Bahnemann, J. Cunningham, M. A. Fox, E. Pelizzetti, P.
Pichat and N. Serpone, in Surface and Aquatic Photochemistry,
ed. G. R. Helz, R. G. Zepp and D. G. Crosby, Lewis Publishing,
Boca Raton, FL, 1993, p. 261.
5
6
7
8
G. K. C. Low, S. R. McEvoy and R. W. Matthews, Environ. Sci.
T echnol., 1991, 25, 460.
E. Pelizzetti, V. Maurino, C. Minero, V. Carlin, E. Pramauro, O.
Zerbinati and M. L. Tosato, Environ. Sci. T echnol., 1990, 24,
When an equimolar mixture of HQ and BQ was irradiated
in the presence of titanium dioxide (see Table 3), the HQ con-
centration increased to 160% of its initial value after 1 min
and decreased to 60% after 6 min, while the p-benzoquinone
disappearance rate was una†ected by the simultaneous pres-
1
559.
9
V. Augugliaro, L. Palmisano, M. Schiavello, A. Sclafani, L. Mar-
chese, G. Martra and F. Miano, Appl. Catal., 1991, 69, 323.
R. W. Matthews, J. Chem. Soc., Faraday T rans. 1, 1989, 85, 1291.
J. C. DÏOliveira, C. Guillard, C. Maillard and P. Pichat, J.
Environ. Sci. Health A, 1993, 28, 941.
1
1
0
1
ence of HQ. Quinonoid structures appear to be efficient e~
CB
12 E. Pelizzetti, C. Minero, P. Piccinini and M. Vincenti, Coord.
Chem. Rev., 1993, 125, 183.
13 E. Pelizzetti, C. Minero, P. Piccinini and M. Vincenti, Chemo-
sphere, 1994, 28, 1229.
scavengers, even in aerated solutions. Interestingly, generation
of ~OH and ~H radicals in sonolysis experiments on aqueous
solutions of BQ also produced HQ in considerable yield.44
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
1999

View Article Online
1
4
C. Maillard, C. Guillard, P. Pichat and M. A. Fox, New J. Chem.,
992, 16, 821.
Z. Wang and C. Kutal, Chemosphere, 1995, 30, 1125.
D. C. Schmelling and K. A. Gray, in Photocatalytic PuriÐcation
and T reatment of W ater and Air, ed. D. F. Ollis and H. Al-Ekabi,
Elsevier, Amsterdam, 1994, p. 625.
30 K. D. Asmus, G. Beck, A. Henglein and A. Wigger, Ber. Bunsen-
Ges. Phys. Chem., 1966, 70, 869.
1
1
1
5
6
31 P. Zuman and B. Shah, Chem. Rev., 1994, 94, 1621.
32 V. Augugliaro, L. Palmisano, A. Sclafani, C. Minero and E. Peliz-
zetti, T oxicol. Environ. Chem., 1988, 16, 89.
33 M. Barbeni, E. Pramauro, E. Pelizzetti, E. Borgarello, M. Graet-
zel and N. Serpone, Nouv. J. Chem., 1984, 8, 547.
17
18
19
20
21
22
23
24
25
26
C. G. Heijman, E. Grieder, C. Holliger and R. P. Schwarzem-
bach, Environ. Sci. T echnol., 1995, 29, 775.
34 J. Augustynski, Struct. Bonding (Berlin), 1988, 69, 1.
35 B. Kraeutler and A. J. Bard, J. Am. Chem. Soc., 1978, 100, 4317.
36 K. Kalyanasundaram, M. Graetzel and E. Pelizzetti, Coord.
Chem. Rev., 1986, 69, 57.
J. Klausen, S. P. Trober, S. B. Haderlein and R. P. Schwarzem-
bach, Environ. Sci. T echnol., 1995, 29, 2396.
A. Agrawal and P. G. Tratnyek, Environ. Sci. T echnol., 1996, 30,
1
53.
37 I. Izumi, W. W. Dunn, K. O. Wilbourn, F. R. Fan and A. J. Bard,
J. Phys. Chem., 1980, 84, 3207.
O. Kamm, Organic Syntheses, Wiley, New York, 1932, Collective
I, p. 435.
38 J. N. Kesselman, G. A. Shreve, M. R. Ho†mann and N. S. Lewis,
J. Phys. Chem., 1994, 98, 13385.
E. Pelizzetti, C. Minero and V. Carlin, New J. Chem., 1993, 17,
3
15.
39 E. Pelizzetti, C. Minero, E. Borgarello, L. Tinucci and N.
Serpone, L angmuir, 1993, 9, 2995.
E. Pelizzetti, C. Minero, V. Maurino, A. Sclafani, H. Hidaka and
N. Serpone, Environ. Sci. T echnol., 1989, 23, 1380.
C. Minero, E. Pelizzetti, S. Malato and J. Blanco, Chemosphere,
40 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross, J.
Phys. Chem. Ref. Data, 1988, 17, 513.
1
993, 26, 2103.
41 S. R. Morrison, Electrochemistry at Semiconductor and Oxidized
Metal Electrodes, Plenum Press, New York, 1980.
42 R. K. Norris and S. Sternhell, Aust. J. Chem., 1966, 19, 841.
43 C. Minero, E. Pelizzetti, M. Ristorto, M. Sega and M. Vincenti,
unpublished work.
C. Minero, M. Vincenti, S. Lago and E. Pelizzetti, Fresenius J.
Anal. Chem., 1994, 350, 403.
D. Lawless, N. Serpone and D. Meisel, J. Phys. Chem., 1991, 95,
5
166.
C. Chatgilialoglu and K. U. Ingold, J. Am. Chem. Soc. 1981, 103,
44 N. Serpone, R. Terzian, P. Colarusso, C. Minero, E. Pelizzetti
and H. Hidaka, Res. Chem. Interm., 1992, 18, 183.
4
833.
2
2
2
7
8
9
R. Tanikaga, Bull. Chem. Soc. Jpn., 1969, 42, 210.
C. Minero, Sol. Energy Materials Sol. Cells, 1995, 38, 421.
C. Minero, E. Pelizzetti, S. Malato and J. Blanco, Solar Energy,
1
996, 56, 421.
Paper 6/07883D; Received 20th November, 1996
2000
J. Chem. Soc., Faraday T rans., 1997, V ol. 93