Photocatalytic degradation of lignin and lignin models, using titanium dioxide: The role of the hydroxyl radical

Article abstract of DOI:10.1016/S0045-6535(99)00269-6

The role of hydroxyl radicals on the degradation of lignins during a cellulosic pulp bleaching process including a photocatalytic stage, was assessed using peroxyformic acid lignins EL1 and REL1 and two phenolic biphenyl lignin models 1 and 2. The irradiations were performed in the absence of photocatalyst TiO₂ and H₂O₂ (condition a), in the presence of TiO₂ (condition b) and in the presence of H₂O₂ (condition c). The experiments were conducted in alkaline (pH? 11) aqueous ethanol solutions with oxygen bubbling. The relative phenolic content of the irradiated solutions, which is indicative of the involvement of hydroxyl radicals, was determined by ionization absorption spectroscopy. The results obtained show that the catalyzed reaction involves both degradation of the phenolate groups by electron transfer and hydroxylation of the lignin aromatic structure. Benzyl alcohol structural elements in sodium borohydride reduced lignin REL1 and compound 2 were also found as good trapping agents for the hydroxyl radicals. The degradation of EL1 was studied by measuring its fluorescence emission by comparison to the fluorescence of compound 2. The emission spectra indicate that some biphenyl phenolate anions in EL1 are reacting under UV/visible irradiation and some others, probably polyphenolic chromophores emitting less fluorescence, are formed.

Full text of DOI:10.1016/S0045-6535(99)00269-6

Chemosphere 40 (2000) 115±124
Photocatalytic degradation of lignin and lignin models, using
titanium dioxide: the role of the hydroxyl radical
a
Antonio E.H. Machado , Aline M. Furuyama , Sandra Z. Falone ,
a
a
a
Reinaldo Ruggiero , Denilson da Silva Perez , Alain Castellan
b
b,
*
a
Departamento de Qu ꢀõ mica, Universidade Federal de Uberl a^ ndia, LFQM, P.O. Box 593, CEP 38400-089, Uberl a^ ndia, MG, Brazil
b
Laboratoire de Chimie des Substances V eꢀ getales, Institut du Pin, Universit eꢀ Bordeaux I, F-33405, Talence Cedex, France
Received 8 March 1999; accepted 7 June 1999
Abstract
The role of hydroxyl radicals on the degradation of lignins during a cellulosic pulp bleaching process including a
photocatalytic stage, was assessed using peroxyformic acid lignins EL1 and REL1 and two phenolic biphenyl lignin
models 1 and 2. The irradiations were performed in the absence of photocatalyst TiO and H (condition a), in the
presence of TiO (condition b) and in the presence of H (condition c). The experiments were conducted in alkaline
2
2 2
O
2
2 2
O
(
pH ꢀ 11) aqueous ethanol solutions with oxygen bubbling. The relative phenolic content of the irradiated solutions,
which is indicative of the involvement of hydroxyl radicals, was determined by ionization absorption spectroscopy. The
results obtained show that the catalyzed reaction involves both degradation of the phenolate groups by electron transfer
and hydroxylation of the lignin aromatic structure. Benzyl alcohol structural elements in sodium borohydride reduced
lignin REL1 and compound 2 were also found as good trapping agents for the hydroxyl radicals. The degradation of
EL1 was studied by measuring its ¯uorescence emission by comparison to the ¯uorescence of compound 2. The
emission spectra indicate that some biphenyl phenolate anions in EL1 are reacting under UV/visible irradiation and
some others, probably polyphenolic chromophores emitting less ¯uorescence, are formed. Ó 1999 Elsevier Science Ltd.
All rights reserved.
Keywords: Lignin degradation; Hydroxyl radicals; Photocatalysis; Titanium dioxide
1
. Introduction
nolic aromatic structural elements formed during its
biosynthesis by dehydrogenative polymerization of co-
niferyl and synapyl alcohols. So, it is not surprising to
note that photolysis of lignin conduces to similar de-
gradation as observed for phenols, even though non-
phenolic elements are absorbing the excitation light. The
wavelength range required for lignin degradation is
generally of lower energy (between 300 and 400 nm)
than for phenols (<300 nm). This is due to the multi-
functional character of lignin with some groups sensi-
tizing the degradation of the phenolic elements
Upon excitation with ultraviolet light, phenolic
compounds in aerated aqueous solutions, undergo de-
gradation yielding quinones, dimers and sometimes
oligomers as major products (Mazelier et al, 1998).
Lignin which is the second most abundant macromole-
cule in wood contains various phenolic and non-phe-
*
Corresponding author. Tel.: +33-05-56-84-62-80; fax: +33-
5-56-84-64-22.
(
Neumann et al., 1986a, 1986b; Kringstad and Lin,
0
E-mail addresses: aeduardo@ufu.br (A.E.H. Machado),
a.castellan@ipin.u-bordeaux.fr (A. Castellan)
1970; Castellan et al., 1990). Excited a-carbonyl groups
are able to abstract hydrogen from some lignin units,
0045-6535/00/$ - see front matter Ó 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 9 9 ) 0 0 2 6 9 - 6

1
16
A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
preferentially phenolic hydrogen, initiating an oxidative
chain reaction with the participation of ground-state
oxygen, leading to the fragmentation and the formation
of new chromophores such as quinones. It is known that
the action of hydroxyl radicals can induce the formation
of radical sites on a substrate, allowing further oxidation
or disproportionation, the process being accelerated by
oxygen (Gierer, 1997).
2. Experimental
The degradation of the lignin samples EL1 and
REL1 and of the lignin model compounds 1 and 2 was
monitored at di€erent reaction times under similar
conditions to those used for the photobleaching of cel-
lulosic pulps (photocatalysis at alkaline medium,
pH ˆ 11), using UV-visible radiation (k > 300 nm) sup-
plied by a 400 W medium pressure mercury lamp. The
photochemical reactions were performed at 25°C in a
cylindrical Pyrex glass reactor (5 cm diameter) equipped
with a cooling jacket; it was positioned at 8.0 cm from
the lamps. During the reaction, the medium was main-
tained under constant oxygen bubbling. The reactions
were carried out on a solution of lignin or model (25 ml)
in a mixture of ethanol and water (1 V/4 V) at a con-
In a recent study (Perez et al., 1998a), we have used
the semiconductor TiO
of pulps with UV/visible light. We proposed that lignin
degradation in the presence of TiO and light was due to
2
to enhance the photobleaching
2
the combined action of singlet oxygen, hydroxyl and
superoxide radicals which are usually the principal re-
active species generated in such a photocatalytic process
in the presence of oxygen. The following equations
�
1
summarize their formation under photocatalytic condi-
‡
centration of 0.25 mg ml . Three conditions were used
as: (a) with absence of TiO and H ; (b) with presence
of TiO and absence of H ; (c) with presence of H
and absence of TiO . For the experiments performed in
the presence of titanium dioxide, a TiO ± substrate
tions (Serpone, 1997; Ho€man et al., 1995). TiO
�
2
(h )
(e ) represent respectively, the electron-de®-
2
2 2
O
and TiO
2
2
2
O
2
2 2
O
cient and electron-rich parts in the structure of TiO
and S the lignin substrate.
2
,
2
2
weight ratio equal to 1/50 was used, whereas for those in
the presence of hydrogen peroxide, equimolar concen-
tration of H O and substrate was considered.
hm
�
‡
TiO
2
! TiO
2
ꢁe =h †
‡
á
‡
TiO
2
ꢁh † ‡ H
2
O ! TiO
2
2
2
2
‡ HO ‡ H
2
2
‡
‡á
The lignin EL1, from Eucalyptus grandis wood, was
obtained from the peroxyformic acid pulping spent li-
quor, by means of dissolution in sodium hydroxide so-
lution (0.1 N) and precipitation in a solution of
hydrochloric acid, after solvent evaporation. The pre-
cipitate, after centrifugation, was dried before use. The
spectroscopic characterization of this lignin, detailed in
the previous publication (Perez et al., 1998b), will be
referred for the discussion. The lignin model dehyd-
rodivanillin 1 was prepared by oxidative coupling of
vanillin (Elbs and Lerch, 1916). Both the lignin REL1
and the compound 2 (1.5 g) were obtained from lignin
EL1 and compound 1 respectively by action of sodium
borohydride (1 g) in ethanol (100 ml) for four days at
room temperature. Then the reaction mixture was
acidi®ed with dilute hydrochloric acid, the precipitate
was washed with distilled water until neutrality, and
then the material was dried in an oven at 40°C. UV
absorption of the lignin and model compounds showed
complete removal of the carbonyl chromophores.
TiO
2
ꢁh † ‡ S ! TiO
‡ S
‡
�
á
TiO
2
ꢁh † ‡ OH ! TiO
‡ HO
�
� á
TiO
2
ꢁe † ‡ O
2
! TiO
! HO ‡ O
‡ O
2
�
á
� á
OH ‡ O
2
2
á
� á
�
1
HO ‡ O ! OH ‡ O
2
2
In this communication, the role of hydroxyl radicals
in the photocatalytic deligni®cation process was ap-
proached using two soluble lignins. One, EL1, was ob-
tained by peroxyformic acid pulping of Eucalyptus
grandis wood and was previously well characterized by
spectrometric methods (Perez et al., 1998b), the second
one, REL1, was obtained from EL1 by the action of
sodium borohydride which is known to reduce carbonyl
group. The study was also extended to the phenol
biphenyl model compounds 1 and 2 which mimic con-
densed residual lignin. Ionization UV/visible absorption
and ¯uorescence emission spectroscopies were the two
main techniques used (Scheme 1).
Lignin degradation was monitored by size exclusion
chromatography (SEC), using the samples obtained af-
2
ter photolysis in the presence of TiO , followed by sol-
vent evaporation under vacuum at 45°C and dissolution
in tetrahydrofuran (1 ml). The resulting solution (20 ll)
was injected in a Merck±Hitachi HPLC equipped with
four Shodex KF800 columns in series giving an exclu-
3
5
sion limit between 1.5 ´ 10 and 4.0 ´ 10 Da. A Hach
DR-4000U UV/visible spectrometer was also used to
follow the reactions, measuring the variation of the ab-
sorbance at 280 nm for EL1, REL1 and compound 2
and that at 350 nm for compound 1. The phenolic
Scheme 1. Formulae of the studied biphenyl lignin model
compounds.

A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
117
content was evaluated by a di€erential absorption
spectrometry, using a methodology proposed by Wexler
1964). Prior to analysis, two samples of the substrate
(
were prepared at the same concentration by diluting 2 ml
of the photolyzed solutions to 10 ml with HCl (0.1 N)
and with NaOH (0.1 N), respectively. The phenolic hy-
droxyl group content in the sample was directly ex-
pressed as relative absorbance which were measured
with the acidic solution placed in the reference beam.
The ¯uorescence spectra were recorded on a Hitachi
F4500 ¯uorimeter using the non-corrected mode with
the excitation and emission slits adjusted to 2.5 nm. The
spectra were recorded for basic (pH ˆ 11) and almost
neutral (pH ˆ 6) solutions with an absorbance adjusted
around 0.1 at 280 nm.
Fig. 1. Relative changes in the content of phenolic OH during
the photochemical degradation of lignin under three di€erent
conditions: (a) direct photolysis; (b) photocatalysis (TiO
2
/lig-
3
. Results and discussion
nin ˆ 1/50 w/w); (c) photolysis in the presence of H
2
O
2
2 2
(H O /
lignin ˆ 1/1 mol/mol). The phenolic content is expressed in
relative absorbance units given by ionization absorption mea-
surements at 280 nm (Wexler, 1964).
3
.1. Irradiation of EL1 and REL1 lignins
The involvement of hydroxyl radicals in the lignin
photodegradation process was approached by measur-
ing the variation of the phenolic content of the sample
determined by ionization absorption spectroscopy set up
by Wexler (1964). The lignin degradation was induced
by three di€erent photochemical processes with: (a)
photolysis of lignin by UV/visible light in alkaline so-
lution (pH ˆ 11) in the presence of oxygen, the light
promoting the electron transfer from the phenolate an-
ion to the oxygen molecule (Marcoccia et al., 1991); (b)
an oxidative chain with the participation of molecular
oxygen (Scheme 2).
This process might contribute to the observed re-
duction of the free phenolic groups content in lignin EL1
using the process (a). In other respect, the scission of b-
O-4 structures could induce the formation of more free
phenolic groups (Palm et al., 1992) but their content in
9
EL1 was estimated approximately at 0.1/C units (Perez
photocatalyzed assistance by TiO
2
(Perez et al., 1998a);
et al., 1998b) and so they cannot justify a signi®cant
increase of the phenolic content.
and (c) the bene®t from species generated, by the pho-
todecomposition of hydrogen peroxide, which is known
to induce deligni®cation (Tatsumi and Terashima, 1981).
The results of the irradiation of lignin EL1 are pre-
sented in Fig. 1.
The behavior of lignin under the conditions (b) and
(c) was found to be very di€erent from that observed
under direct photolysis (a). In such cases, a rapid initial
decrease in the content of phenolic hydroxyl groups was
observed, followed by a clear tendency of increasing, due
to the incorporation of hydroxyl groups on aromatic
rings, occurring by hydroxylation and/or demethoxyla-
tion (Gierer, 1997).
The curve (a) indicates a sharp decrease of the phe-
nolic content. It was not unexpected considering that the
phenolic content of the lignin was high (ꢀ0.5/C
9
unit
(
Perez et al., 1998b)) and that under direct photolysis the
generation of hydroxyl radicals is not favored. There-
fore, the oxidation of the phenol groups giving phenoxyl
radicals appears to be the main channel of disappear-
ance of these groups. On the other hand, one of the
processes expected during the direct photolysis of lignins
is the excitation of carbonyl groups, which can promote
hydrogen abstraction, preferentially from free phenolic
groups (Neumann et al., 1986a, 1986b), generating
radical sites in the substrate, allowing the propagation of
2
In contrast to the direct irradiation, the TiO pho-
á
tocatalyzed reaction generates hydroxyl radicals ( OH)
�
á
and anion superoxyde (O ) as previously discussed in
2
Section 1. The high reactivity of the hydroxyl radicals is
well known (Klopfer, 1987; Steenken, 1987). This is not
the case for the superoxide radicals, although they have
a long lifetime when compared to the hydroxyl radicals,
their reactivity is comparatively low (Gierer, 1997;
Sawyer, 1988; Mattes and Farid, 1986). However, it has
Scheme 2. Formation of phenoxyl radicals by intermolecular abstraction of phenolic hydrogen by carbonyl groups.

1
18
A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
been shown that the selectivity of the superoxide during
the bleaching process is higher than hydroxyl radicals
the new benzyl alcohol structural elements formed by
reduction of the conjugated a-carbonyl groups of EL1
(Scheme 4). When the ketyl radical incorporates b±O±4
structural elements, cleavage of the b±O±4 bond can
occur generating phenoxyl and phenacyl radicals (Palm
et al., 1992). For other structural elements, the ketyl
radical reverts to a-carbonyl derivative which does not
contribute to the degradative pathway.
(
Gierer, 1997). The initial decrease noted for curve (b)
has to be attributed to the direct photolysis mechanism,
due to the minor content of hydroxyl radicals photo-
catalytically generated in the beginning of the reaction.
This is a consequence of the low quantity of photocat-
alyst, relative to the lignin mass (1/50), selected to re-
produce the nearest as possible pulp photobleaching
conditions (Perez et al., 1998a; Castellan et al., 1999).
This low quantity of photocatalyst allowed the preser-
vation of the cellulosic ®bers after a photocatalyzed pulp
bleaching stage (Perez et al., 1998a). In the initial stage
of the reaction, the rate of consumption of phenolic
groups is higher than the rate of incorporation of hy-
droxyl radicals into the aromatic rings. When a steady-
state concentration of hydroxyl radicals is reached, the
reactions induced by these radicals become more im-
portant and their involvement in the degradation pro-
cess increases. Some of them are indicated in Scheme 3.
When the irradiation was performed in the presence of
hydrogen peroxide (Fig. 1, curve (c)), which is an im-
portant source of hydroxyl radicals, a remarkable in-
crease of the phenolic content was observed. Their
On the basis of knowledge of semiconductor photo-
catalyzed process (Villasenor and Mansilla, 1996), the
expected trend is an increase of the phenolic content in
the beginning of the reaction followed by breakdown of
the aromatic structure of the lignin and its complete
2
oxidation. This situation is reached when the ratio TiO /
lignin equal to 1/5 is used as shown in Fig. 3. Both the
similarity of the behavior observed for curves (b) and (c)
in Fig. 1 and the results presented in Fig. 3 suggest that
hydroxyl radicals play an important role during the
2
TiO -photoassisted process, therefore we must take care
about the amount of photocatalyst used for the
bleaching.
An important fragmentation of the lignin structure
during the photocatalytic process was monitored by size
exclusion chromatography (SEC) (Fig. 4). A reduction
of almost 50% in the average molecular weight of EL1
which was equal to 1.4 kD (Perez et al., 1998b), takes
place after 90 min of irradiation.
formation by homolytic scission of H
2 2
O is probably not
the main process due to the low absorption coecient of
hydrogen peroxide in the wavelength range employed in
this study. Decomposition of hydrogen peroxide by
The photocatalytic degradation noted by SEC is in
contrast with the evolution of the absorbance measured
at 280 nm for the lignin solution where an increase was
observed (Fig. 5). This tendency suggests that the
conjugation of the p system is extended, probably by
oxidation of some lateral chains, giving stilbene or other
conjugated chromophores.
2
‡
2‡
transition metal Fe , Cu (Landucci and Sanyer, 1975)
3
‡
or Mn (Abbot, 1991) by Fenton reaction appears to be a
more productive and realistic source of hydroxyl radicals.
The behavior of the sodium borohydride reduced
lignin REL1 under irradiation conditions (a) and (c) is
very similar to the non-reduced lignin EL1 (Fig. 2). In
contrast, the irradiation of REL1 in the presence of TiO
condition (b)) indicates that the incorporation of hy-
2
(
3.2. Irradiation of lignin models 1 and 2
droxyl group in the lignin polymer is less ecient. This
might be attributed to the existence of a competitive
reaction involving the hydroxyl radicals reacting with
The involvement of hydroxyl radicals in the photo-
catalyzed degradation of lignin was also approached
Scheme 3. Reactions of hydroxyl radicals with lignin structural elements.

A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
119
Fig. 2. Relative changes in the content of phenolic OH during
the photochemical degradation of REL1: (a) direct photolysis;
Fig. 4. Relative changes of average molecular weight number
(Mw) of lignin EL1 solution versus irradiation time when ir-
(
presence of H (see caption of Fig. 1).
b) photocatalysis (TiO
2
/lignin ˆ 1/50 w/w); (c) photolysis in the
2
radiated with TiO . (See condition (b) of Fig. 1).
2
O
2
Scheme 4. Reaction of hydroxyl radicals with benzyl alcohol
structural lignin elements.
Fig. 5. Relative changes of the absorbance at 280 nm of lignin
EL1 solution versus irradiation time when irradiated using the
photocatalytic condition (b) of Fig. 1.
biphenyl dimers 1 and 2. The decrease of the hydroxyl-
ation of the benzene structures noted in the photocata-
lyzed (b) process for REL1 is also observed for
compound 2. This con®rms, once again, the reactivity
between the benzyl alcohol structural elements and hy-
droxyl radicals.
Fig. 3 . Relative changes in the content of phenolic OH during
the photocatalytic degradation of lignin EL1 using larger
quantity of TiO (TiO
is expressed in relative absorbance units given by ionization
absorption measurements at 280 nm (Wexler, 1964).
2
2
/lignin ˆ 1/5 w/w). The phenolic content
The comparative behavior of the lignin models and
the lignin samples shows that the presence of conjugated
carbonyl chromophores does not modify to a great ex-
tent the action of the hydroxyl radicals against the lignin
polymer. Also, the electron transfer between the
phenolate anion and the oxygen molecule seems to be
the main mechanism operating in condition (a). This
remains the observation made on high-yield pulps on the
ineciency of sodium borohydride treatment to stabilize
the pulp against photodegradation (Heitner, 1993). In
all these degradative processes, the conjugated carbonyl
using the biphenyl lignin models 1 and 2 which might be
considered as representative of phenolic condensed
structures.
The relative changes in the content of phenolic hy-
droxyl groups during the photodegradation of the lignin
models 1 and 2 are respectively presented in Figs. 6 and
7. Comparatively, the same trends as it was previously
observed for the lignin EL1 and EL2 are seen for the

1
20
A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
Fig. 6. Relative changes in the content of phenolic OH during
the photochemical degradation of lignin model 1: (a) direct
Fig. 8. Pseudo-®rst order kinetics for the degradation of de-
hydrodivanillin 1 monitored at 350 nm: (a) direct photolysis,
photolysis; (b) photocatalysis (TiO
photolysis in the presence of H (H
2
/lignin ˆ 1/50 w/w); (c)
I
� 1
without TiO
(
2
, k ˆ 0.0458 min ; (b) with the addition of TiO
2
2
O
2
2
O
2
/lignin ˆ 1/1 mol/mol).
I
� 1
.
TiO
2
/1 ˆ 1/50 w/w), k ˆ 0.1227 min
The phenolic content is expressed in relative absorbance units
given by ionization absorption measurements at 350 nm
(
Wexler, 1964).
pulps containing lignin and wood (Davidson et al., 1991;
Castellan et al., 1992; Choudhury et al., 1992; Castellan
and Davidson, 1994; Castellan et al., 1994a, 1994b). The
luminescence of these materials has been mainly attrib-
uted to the lignin macromolecule. The emission spectra
of the lignocellulosic materials in the solid state were
found to be highly dependent on the wavelength of ex-
citation, suggesting that the emission arises from di€er-
ent chromophores present in the lignin macromolecule
(
Davidson et al., 1991). In contrast, Lundquist et al.
1978, 1981) have shown that lignin samples in solution
(
when excited in the wavelength range 250±350 nm gave
very similar emission spectra. This is an indication that
intramolecular energy transfer is occurring in solution
and that the emitted light arises primarily from certain
type of chromophores. It was found that reduction of
carbonyl groups of lignin with sodium borohydride in-
duced a strong increase of the ¯uorescence emission
Fig. 7. Relative changes in the content of phenolic OH during
the photochemical degradation of compound 2: (a) direct
photolysis; (b) photocatalysis (TiO
2
/lignin ˆ 1/50 w/w); (c)
(
Castellan et al., 1992). Quenching by carbonyl groups
photolysis in the presence of H O (see caption of Fig. 1).
2
2
of the ¯uorescence emitted by the aromatic skeleton of
lignin was assumed to be the main reason for this be-
havior (Castellan et al., 1992). Also, the relationships
between the spatial arrangement of residual condensed
lignin in the pulp and its luminescence properties were
approached on non-phenolic biphenyl lignin models by
comparing their luminescence emission and their struc-
tures in the crystalline state determined by X-rays
groups do not appear as the main actors of the break-
down of the lignin polymer.
In Fig. 8 are reported the pseudo-®rst order kinetics
of the disappearance of dehydrodivanillin 1 monitored
by the decrease of the absorbance at 350 nm using
conditions (a) and (b). The photocatalytic process is at
least three times more ecient to promote the degra-
dation than the direct photolysis.
(
Guilardi-Ruggiero et al., 1997, 1999). The torsion angle
in the biphenyl ¯uorophore appeared to be the main
parameter governing the ¯uorescence emission in the
solid and liquid states (Guilardi-Ruggiero et al., 1999).
In this way, we decide to investigate by using ¯uores-
cence spectroscopy the behavior of the EL1 and REL1
lignins under the photobleaching conditions used in this
study.
3.3. Study of the photodegradation of EL1 by ¯uorescence
spectroscopy
Fluorescence spectroscopy has been proven to be a
useful tool in the study of the photochemical behavior of

A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
121
3
.3.1. Fluorescence of EL1 and of compound 2 without
irradiation
The ¯uorescence spectra obtained for neutral and
the absorption spectrum. A bathochromic shift (25 nm)
is observed for the maximum of the ®rst excitation band
when the solution was brought to alkaline pH. Com-
pound 2 displays two emission bands, the ®rst one has a
maximum situated around 350 nm and is not shifted
when the medium was brought to alkaline pH. The
maximum of the second band is shifted from 412 nm in
neutral solution to 420 nm in alkaline pH (ꢂ11), this
band originates from the phenolate anion. One can ob-
serve that the variation of the intensity of the phenolate
emission band in alkaline medium follows the variation
of the intensity of the absorption band with very similar
intensity for excitation wavelengths equal to 290 and 330
nm.
alkaline aqueous acetonitrile solutions of EL1 and its
sodium borohydride reduced derivative, REL1 using
three excitation wavelengths, 290, 310 and 330 nm and
two emission wavelengths, 390 and 430 nm. The exci-
tation and emission ¯uorescence spectra of REL1 and
EL1 were found very similar and for clarity only those
of EL1 are presented in Fig. 9.
For a better understanding of the ¯uorescence
properties of the EL1 and REL1 lignin, the ¯uorescence
emission of compounds 1 and 2 were also studied, these
compounds are considered representative of phenolic
condensed conjugated lignin structure. As previously
observed for other biphenyl derivatives, compound 1
was found non-¯uorescent in our experimental condi-
tions. This observation indicates that the ¯uorescence
observed for EL1 was due to the non-carbonyl part of
the lignin polymer. The ¯uorescence emission spectra of
the lignin model 2 are reported in Fig. 10.
Comparison of the ¯uorescence spectra of the lignin
EL1 and the lignin model 2 indicates that the band at
343 nm observed for EL1 might be due to phenolic
biphenyl ¯uorophore in their non-basic form as well as
that emission in the range 400±500 nm is indicative of
the existence of biphenyl phenolate ¯uorophores. Nev-
ertheless, the non-similarity of both excitation and
emission ¯uorescence spectra recorded for compound 2
and lignin EL1 shows that other chromophores in lignin
EL1 are emitting in our experimental conditions. This
was already observed for wood (Davidson et al., 1991).
The excitation spectra monitored at 390 and 430 nm
for the compound 2 are very similar and correspond to
Fig. 9. Fluorescence spectra of lignin EL1 in aqueous aceto-
nitrile solution (acetonitrile/water 1 V:1 V, temperature ꢀ300
K, fexc ˆ fem 2.5 nm, absorbance at 280 nm ꢀ0.8, neutral so-
lution: pH ꢀ 6, alkaline solution pH ꢀ 11).
Fig. 10. Fluorescence spectra at room of the lignin model 2 in
aqueous acetonitrile solution (acetonitrile 1 V/water 1 V, tem-
perature ꢀ300 K, fexc ˆ fem 2.5 nm, absorbance at 280 nm
ꢀ0.8, neutral solution: pH ꢀ 6, alkaline solution pH ꢀ 11).

1
22
A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
3
.3.2. Fluorescence of EL1 after irradiation
The solutions of lignin EL1 were irradiated during 30
non-irradiated sample EL1 (alkaline medium), the in-
tensity of the emission (range 400±500 nm) is very sim-
ilar when excited at 290 and 330 nm. The same
phenomenon is observed for the biphenyl 2 solution
(alkaline medium). The lower intensity emission ob-
served when the excitation was set at 330 nm (alkaline
medium) seems to be characteristic of the reactivity of
the biphenyl phenolate anion. The same observation was
done when biphenyl compound 2 solutions were irradi-
ated using conditions (a)±(c) (spectra not shown). The
¯uorescence spectroscopy study performed in this work
shows that chromophores, such as the biphenyl phenols,
are destroyed and new ¯uorophores, less emitting, are
formed during the photolysis, may be some polyphenolic
compounds.
min using the conditions of Fig. 1(a)±(c). Then the so-
lutions were diluted in a mixture of acetonitrile/water
(
1 V/1 V), the pH was adjusted to 11 and the ¯uores-
cence recorded using the same experimental conditions
as for Fig. 9. The spectra are shown in Fig. 11.
A clear di€erence observed between the non-irradi-
ated and the irradiated solutions spectra is the relative
decrease of the intensity in the range 300±350 nm of the
excitation spectra recorded for kem equal to 390 nm and
4
30 nm, especially for the latter. This region is charac-
teristic of the biphenyl phenolate anion and this obser-
vation shows the reactivity of these anions under UV
light exposure. Another observation, more or less re-
lated to the ®rst one, is the di€erence in intensity for the
range 400±500 nm between the emission recorded for
kexc equal to 290 nm and 330 nm for the irradiated lignin
solution compared to the non-irradiated one. In the
4
. Conclusions
Although the overall mechanism of the photochem-
2
ical bleaching process using TiO as photocatalyst is not
completely depicted using Eucalyptus peroxyformic acid
soluble lignin in alkaline aqueous ethanol solution, some
of the evidences are presented in this study. The hy-
droxylation of aromatic structures in lignin was found to
occur during photochemical bleaching in the presence of
2 2 2
TiO or H O which are known to generate hydroxyl
radicals under UV/visible irradiation. The incorporation
of hydroxyl groups to the benzene rings leads to the
creation of radical sites allowing the action of both
molecular oxygen and active species such the superoxide
radicals. This results in the structural fragmentation of
lignin. In contrast, the irradiation of the lignin without
2 2 2
TiO or H O is dominated by the electron transfer
between the phenolate anion and ground state oxygen
leading to the breakdown of the phenolic units. This
mechanism is operating in the beginning of the irradia-
2
tion when the concentration of TiO is low compared to
lignin. The role of phenolic biphenyl units in the de-
gradation of the studied lignin was revealed by com-
parison of the photochemistry of compound 1 and 2 in
similar conditions as the lignin samples. Finally, the
degradation of condensed structures, mimicked here by
biphenyl units was assessed using ¯uorescence spec-
troscopy. This observation is highly connected with the
e€ectiveness of the bleaching process once the presence
of those structures in the residual lignin arising from
acid process has been largely reported.
Acknowledgements
The authors are grateful to FAPEMIG for ®nancial
support, and to CAPES/ COFECUB (Project76/95) for
travelling mission between Brazil and France and CNPq
Fig. 11. Fluorescence spectra at room of photoirradiated lignin
EL1 (same experimental conditions as for Figs. 1 and 9).

A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
123
for grants. AC is very grateful to Conseil Regional
d'Aquitaine and UE (FEDER Support) for help in the
photobleaching project.
properties and spatial arrangement of lignin model mole-
0
cules in the solid part. Part 2. Crystal structures 2,2 -di-(1-
0
0
hydroxyethyl)-3,3 -dimethoxybiphenyl
0
and
2,2 -dihy-
droxymethyl-3,3 -dimethoxybiphenyl and their ¯uorescence
emission. J. Mol. Struct., 484, 235±248.
Heitner, C., 1993. Photochemistry of lignocellulosic materials.
In: Heitner, C., Scaiano, J.C. (Eds.), Light-induced yellow-
ing of wood-containing papers, An evolution of the mech-
anism, ACS Symposium Series 531, pp. 1±25.
References
Abbot, J., 1991. Catalytic decomposition of alkaline hydrogen
peroxide in the presence of metal ions: binuclear complex
formation. J. Pulp Pap. Sci. 17, 10±17.
Ho€man, M., Martin, S., Choi, W., Bahnemann, D., 1995.
Environmental applications of semiconductor photocataly-
sis. Chem. Rev. 95, 69±96.
Castellan, A., Davidson, R.S., 1994. Steady state and dynamic
¯
uorescence emission from Abies wood. J. Photochem.
Photobiol. A 78, 275±279.
2 2
Klopfer, W., 1987. How to improve the H O photolysis as an
OH-source in simulated atmospheric degradation testing
chemicals. EPA Newsletter 31, 4±16.
Castellan, A., Colombo, N., Vanucci, C., Fornier de Violet, P.,
Bouas-Laurent, H., 1990. Photodegradation of lignin. A
photochemical study of an O-methylated a-carbonyl b-1
Kringstad, K.P., Lin, S.Y., 1970. Mechanism in the yellowing
of high yield pulps by light. Structure and reactivity of free
radical intermediates in the photodegradation of lignin.
Tappi J. 53, 2296±2301.
0
0
lignin model dimer: 1,2 di(3 4 -dimethoxyphenyl) ethanone
deoxyveratroin). J. Photochem. Photobiol A 51, 451±467.
(
Castellan, A., Nourmamode, A., Noutary, C., Belin, C.,
Fornier de Violet, P., 1992. Photoyellowing of milled-wood
lignin in solid 2-hydroxypropylcellulose ®lms after sodium
borohydride reduction and catalytic hydrogenation in
solution: A ¯uorescence spectroscopic study. J. Wood
Chem. Technol. 12, 19±33.
Landucci, L., Sanyer, N., 1975. In¯uence of transition metals in
oxygen pulping. Tappi J. 58 (2), 60±63.
Lundquist, K., Josefsson, B., Nyquist, G., 1978. Analysis of
lignin products by ¯uorescence spectroscopy. Holzforsc-
hung 32, 27.
Lundquist, K., Egyed, I., Josefsson, B., Nyquist, G., 1981.
Lignin products in pulping liquors and their ¯uorescence
properties. Cell. Chem. Technol. 32, 669±679.
Castellan, A., Choudhury, H., Davidson, R.S., Grelier, S.,
1
994a. Comparative study of stone ground wood pulp and
native wood. 2. Comparison of the ¯uorescence of stone
ground wood pulp and native wood. J. Photochem. Pho-
tobiol. A 81, 117±122.
Marcoccia, B., Reeve, D.W., Goring, D.A., 1991. Photoen-
hanced oxygen deligni®cation of softwood kraft pulp. 1.
The e€ects of some process variable. J. Pulp Pap. Sci. 17,
34±39.
Castellan, A., Choudhury, H., Davidson, R.S., Grelier, S.,
1
994b. Comparative study of stone ground wood pulp and
Mattes, S.L., Farid, S., 1986. In: Padwa, A. (Ed.), Organic
Photochemistry. vol. 6, Marcel Dekker, New York (Chap-
ter. 4).
native wood. 3. Application of ¯uorescence spectroscopy to
a study of the weathering of stone ground wood pulp and
native wood. J. Photochem. Photobiol. A 81, 123±130.
Castellan, A., Perez, D.S., Nourmamode, A., Grelier, S.,
Terrones, M.G.H., Machado, A.E.H., Ruggiero, R., 1999.
The improvement of the bleaching of peroxyformic sugar
cane bagasse pulp by photocatalysis and photosensitization.
J. Braz. Chem. Soc., 10, 197±202.
Mazelier, P., Sarakha, M., Rossi, A., Bolte, M., 1998. The
aqueous photochemistry of 2,6-dimethylphenol. Evidence
for the fragmentation of the a-C±C bond. J. Photochem.
Photobiol. A 115, 117±121.
Neumann, M.G., De Groote, R.A.M.C., Machado, A.E.H.,
1986a. Flash photolysis of lignin. Part 1. Deaerated solu-
tions of dioxane-lignin. Polym. Photochem. 7, 401±406.
Neumann, M.G., De Groote, R.A.M.C., Machado, A.E.H.,
1986b. Flash photolysis of lignin. II. Oxidative photode-
gradation of dioxane-lignin. Polym. Photochem. 7, 461±468.
Palm, W.U., Dresskamp, H., Bouas-Laurent, H., Castellan, A.,
1992. The photochemistry of a-phenoxyacetophenones in-
vestigated by ¯ash CIDNP spectroscopy. Ber. Bunges. 96,
50±61.
Choudhury, H., Collins, S., Davidson, R.S., 1992. The colour
reversion of papers made from high yield pulp. A photo-
chromic process?. J. Photochem. Photobiol. A 69, 109±119.
Davidson, R.S., Dunn, L., Castellan, A., Nourmamode, A.,
1
991. A study of the photobleaching and photoyellowing of
paper containing lignin using ¯uorescence. J. Photochem.
Photobiol. A 58, 349±359.
Elbs, K., Lerch, H., 1916. Uber dehydrodivanillin. J. Pratk.
Chem. 93, 1±9.
Gierer, J., 1997. Formation and involvement of superoxide
Perez, D.S., Castellan, A., Grelier, S., Terrones, M.G.H.,
Machado, A.E.H., Ruggiero, R., Vilarinho, A.L., 1998a.
Photochemical bleaching of chemical pulp catalyzed by
titanium dioxide. J. Photochem. Photobiol. A 115, 73±78.
Perez, D.S., Terrones, M.G.H., Castellan, A., Grelier, S.,
Nourmamode, A., Ruggiero, R., Machado, A.E.H., 1998b.
Peroxyformic acid pulping of Eucalyptus grandis wood chips
and sugar cane bagasse in one stage and characterization of
the isolated lignins. J. Wood Chem. Technol. 18, 333±365.
Sawyer, D.T., 1988. The redox thermodynamics for dioxygen
2
� :
(
processes: A review. Holzforschung 51, 34±46.
O
/H
2 2
O ) and hydroxyl (OH.) radicals in TCF bleaching
Guilardi-Ruggiero, S., Castellan, A., Cotrait, M., Grelier, S.,
Drumond, M.G., Pilo Veloso, D., 1997. Luminescence
properties and spatial arrangement of lignin model mole-
0
cules in the solid part. Part 1. Crystal structures of 5,5 -
0
0
0
diacetyl-2,2 -dibenzyloxy-3,3 -dimethoxybiphenyl and 2,2 -
0
0
dibenzyloxy-5,5 -diformyl-3,3 -dimethoxybiphenyl and their
phosphorescence emission. J. Mol. Struct. 435, 77±87.
�
species (O
2
, O , HOO, HOOH and HOO ) and monoox-
2
á
á
�
á
�
Guilardi-Ruggiero, S., Castellan, A., Grelier, S., Drumond,
M.G., Nourmamode, A., Cotrait, M., 1999. Luminescence
ygen species (O, O , HO., and HO ) in water and aprotic
solvents. Basic Life Sci. 49, 11±20.

1
24
A.E.H. Machado et al. / Chemosphere 40 (2000) 115±124
Serpone, N., 1997. Relative photonic eciencies and quantum
yields in heterogeneous photocatalysis. J. Photochem.
Photobiol A 104, 1±12.
Villasenor, J., Mansilla, H., 1996. E€ect of temperature on kraft
black liquor degradation by ZnO photoassisted catalyst. J.
Photochem. Photobiol. A 93, 205±209.
Steenken, S., 1987. Addition-elimination paths in electron-
transfer reactions between radicals and molecules. Oxida-
tion of organic molecules by the OH radical. J. Chem. Soc.
Faraday Trans. 83, 113±124.
Wexler, A.S., 1964. Characterization of lignosulfonates by UV
spectrometry. Direct and di€erence spectrograms. Anal.
Chem. 36, 213±221.
Tatsumi, K., Terashima, N., 1981. Oxidative degradation
of lignin. 4. Reactivities of monomeric lignin model com-
pounds with hydrogen peroxide. Mok. Gakk. 27, 873±878.