Chemosphere 50 (2003) 7178

www.elsevier.com/locate/chemosphere

Photo-Fenton treatment of water containing natural

phenolic pollutants
Wolfgang Gernjak a, Thomas Krutzler a, Andreas Glaser a, Sixto Malato
Julia Caceres b, Rupert Bauer a, A.R. Fern

andez-Alba c
a

b,*

Institute of Physical and Theoretical Chemistry, Vienna University of Technology, Veterinarplatz 1, A-1210 Vienna, Austria
b
CIEMATPlataforma Solar de Almer
a, P.O. Box 22, E-04200 Tabernas, Almer
a, Spain
c
Pesticide Residue Research Group, University of Almer
a, 04071 Almer
a, Spain
Received 10 December 2001; received in revised form 15 July 2002; accepted 15 July 2002

Abstract
Phenolic compounds are known to be present in high concentrations in various types of agro-industrial wastes. As
they are highly biorecalcitrant, the possibility of treatment by advanced oxidation processes should be investigated.
In this work, six model phenolic compounds (vanillin, protocatechuic acid, syringic acid, p-coumaric acid, gallic acid
and L -tyrosine) were chosen for a demonstration of degradation by photo-Fenton reaction, under articial light in
laboratory experiments in Vienna and under sunlight in pilot-plant experiments at the Plataforma Solar de Almer
a in
Spain.
All compounds were completely mineralised. No non-degradable intermediates were produced, either in experiments
with single substances or in a more complex matrix of a mixture of phenolic compounds. The expected selectivity of the
photo-Fenton reaction for aromatic compounds was proven by comparison of the decrease in total organic carbon with
the removal of total phenolic content.
2002 Elsevier Science Ltd. All rights reserved.
Keywords: Phenols; Photocatalysis; Photochemical treatment; Photo-Fenton; Water detoxication

1. Introduction
Various natural phenols and their condensation
products, such as tannins or lignins, are present in several types of agro-industrial wastes. Examples of such
wastewater originate in wood-processing industries, as
for instance, paper mills, masonite (fabrication of breboard), debarking of wood (Tello, 2001) or the cork
industry (Moreno, 1994), from tanneries using natural
tannins that still exist in great numbers (Bruneton,
1986), from grape pressing in the wine industry (vinasse)

Corresponding author. Tel.: +34-950387940; fax: +34950365015.

E-mail address: sixto.malato@psa.es (S. Malato).

(Tello, 2001), olive oil milling in the Mediterranean area

(Gonz

alez-Vila et al., 1992), coee production (Tello,
2001) or intensive stock raising (Maureen et al., 1982).
Although the toxicity of phenolic compounds is not as
high as, for example, pesticides or heavy metals, their
high concentration (up to several grams per litre) often
inhibits or even eliminates abundant bacteriological
populations in municipal biological wastewater treatment plants. Thus common biological wastewater treatment is not possible (Borja et al., 1996). In such cases,
advanced oxidation processes (AOPs) should be considered a viable alternative (Legrini et al., 1993). Some
research on the degradation of phenolic compounds with
AOPs has already been done (e.g.: Benitez et al., 1997;
Herrera et al., 1998; Miranda et al., 2000, 2001). AOPs,
while making use of dierent reaction systems, are all

0045-6535/03/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 2 ) 0 0 4 0 3 - 4

72

W. Gernjak et al. / Chemosphere 50 (2003) 7178

Scheme 1. Molecular structures of the selected compounds.

characterised by the same chemical feature: production

of OH radicals ( OH). The versatility of AOPs is also
enhanced by the fact that they oer dierent options for
OH radical production, thus allowing them to conform
to specic treatment requirements. The main disadvantage of AOPs is their high cost. The application of solar
technologies to these processes could help diminish that
problem by reducing the energy consumption required
for generating UV radiation. From the point of view of
their practical application, heterogeneous TiO2 photocatalysis and homogeneous photo-Fenton, are the AOPs
for which the solar technologies have been most extensively studied and developed (Malato et al., 1999; Alfano
et al., 2000; Blanco et al., 2000; Blanco and Malato,
2001). The latter is known to be especially apt for the
treatment of aromatic pollutants because of its great effectiveness in degrading this type of organic compounds
(Huston and Pignatello, 1999). Considering that wastewater is not only polluted by phenolic compounds, but
also to a great extent, by other organic matter, the
achievement of selective treatment would result in eventually lower process costs.
To evaluate the viability of application of the photoFenton reaction to the above-mentioned wastewater, the
six frequently abundant phenolic compounds depicted in
Scheme 1 were chosen for preliminary testing of degradability.

2. Methods
2.1. Analysis
Ferrous sulphate FeSO4
7H2 O and sulphuric acid
were analytical grade (Merck). Hydrogen peroxide (30%
w/v) was analytical grade (Merck) for laboratory purposes and technical grade for the pilot-plant experiments. Total organic carbon (TOC) and inorganic

carbon (IC) were measured by means of a Shimadzu

TOC 5000 analyser with an ASI 5000 auto sampler
(laboratory experiments, Vienna) and a Shimadzu TOC
5050 analyser at Plataforma Solar de Almer
a (PSA,
Spain). H2 O2 in solution was determined by iodometric
titration and pH was measured with a glass electrode.
Iron concentrations were obtained by photometric measurement with 1,10-phenanthroline (following ISO 6332)
using a Shimadzu UV160A spectrophotometer (laboratory experiments, Vienna) and a Unicam-2 spectrophotometer (PSA, Spain). Chemical oxygen demand (COD)
was determined by Merck Spectroquant cuvette tests.
Total phenol content was measured by the Folin-Ciocalteau method (Singleton and Rossi, 1965). Original
substance content was determined by HPLC-UV analysis using a C18 column at acid pH (Spectra Physics
IsoChrom LC Pump, Spectra Physics Spectra 200 programmable wavelength detector with a Shandon 25/5
Hypersil BDS C18 5 lm column in Vienna and Hewlett
Packard, series 1050 equipped with LUNA 5 l-C18,
3  150 mm from Phenomex at PSA, Spain). Elution
conditions and detection wavelength (k) were dierent
for each of the six phenols: Vanillin (30/70, k 310 nm),
protocatechuic acid (15/85, k 256 nm), syringic acid
(30/70, k 272 nm), p-coumaric acid (35/65, k 290
nm), gallic acid (5/95, k 268 nm) and L -tyrosine (0/
100, k 276 nm). Solvent A Methanol, Solvent B
water with 2% acetic acid.
2.2. Experimental set-up
The experiments in Vienna were performed in a
round-bottomed 2-l ask with a magnetic stirrer using a
medium-pressure mercury lamp with water-cooled optics as the radiation source. Radiation entered the ask
through a at Duran glass window. The reaction vessel
was light-insulated with aluminium foil. The experiments at the Plataforma Solar de Almer
a (PSA) were
performed in two dierent non-concentrating pilot-plant
reactors; the newly designed Falling Film Reactor
(FFR, see Fig. 1) and in a conventional CPC (Compound Parabolic Collector) reactor described in detail
elsewhere (Blanco et al., 2000).
The main FFR components are the at plate, top
distributor, bottom receiver, a centrifugal pump and a
batch tank located underneath the at plate. The batch
tank outlet feeds the suction side of the centrifugal pump
(PAN WORLD, Japan. Model: NH-100 PX, 100 W)
that pumps the uid to the top distributor (stainless
steel) of the FFR. From there, the uid ows down the
at plate to the bottom receiver (stainless steel), and
back to the batch tank again. The water thus circulates
in a closed circuit formed by the batch tank, the upper
distributor, the at plate and the bottom receiver. The
at plate consists of a square white sheet of HDPE
(High-density polyethylene, 1 m2 ) tilted 37 (local lati-

W. Gernjak et al. / Chemosphere 50 (2003) 7178

73

y
vw
Cu Hv Nw Ox Ox ! uCO2
H2 O wHNO3
2
2
1

Fig. 1. FFR diagram.

tude, PSA). Piping consists of several HDPE tubes

connected by standard polyethylene quick connectors.
Total reactor volume is 19 l.
The CPC is made up of three modules also inclined
37, the batch tank and a centrifugal pump. Each of
the three modules consists of eight parallel horizontal
borosilicate-glass tubes (1500 mm long  32 mm outer
diameter  1:4 mm wall thickness). Behind each tube,
two parabolic aluminium mirrors redirect the radiation
toward the tubes. The modules are mounted next to
each other and connected in such a way that the water
ows sequentially through all the tubes before owing
to the batch tank (borosilicate glass). From the batch
tank, the water is recirculated by the centrifugal pump
(PAN WORLD, Japan, Model: NH-100 PX, 100 W)
to the CPC modules through HDPE tubing with
standard polyethylene quick connectors. The volume in
the glass CPC module tubes is 24 l and the total volume of the reactor is 35 l. The total illuminated area is
3.08 m2 .
The initial hydrogen peroxide concentration was always according to the calculated stoichiometric demand
for complete mineralisation to CO2 . The amount of
hydrogen peroxide necessary was calculated as follows.
Cu Hv Nw Ox is mineralised as in Eq. (1). Its COD can be
calculated using Eqs. (2) and (3). As H2 O2 sets free one
half of an oxygen molecule, the required stoichiometric
hydrogen peroxide concentration can be calculated by
multiplying the model substance COD by 34/16, which
is the ratio of the molecular weight of H2 O2 and the
atomic weight of oxygen:

1
1
y 2u m w  x
2
5

CODCu Hv Nw Ox 16y cCu Hv Nw Ox mg=l

pH was adjusted to 2.8 with sulphuric acid in all the

tests. In the ask reactor, the experimental procedure
was as follows: First, the pH was adjusted with sulphuric
acid, and then ferrous sulphate and the model substance
were added. Hydrogen peroxide was added after homogenisation. This marked the beginning of the experiment. After 5 min, irradiation was started. While pH
was not kept constant during the experiments, the temperature was maintained at 20 C. The procedure for the
experiments in the CPC was analogous to Vienna, the
only dierence being that it took longer to homogenise
the solution.
In the FFR, the irradiated area could not be covered
during recirculation. Therefore, at the beginning of the
experiment (addition of hydrogen peroxide) no real
dark reaction could be measured. To nd a compromise between low irradiation, yet complete homogenisation of the reaction solution, the rst sample was
taken 5 min after the reaction began. Syringic acid
turned out to be insuciently soluble to perform an
experiment at the desired concentration in the FFR.
Temperature could not be kept constant in the pilot
plants, and varied from 30 to 40 C. Global UV (300
400 nm) irradiance was recorded in the solar plant by a
Kipp and Zonen CUV3 detector with the same 37 inclination as the reactor modules so that incident radiation could be evaluated as a function of time of day,
cloudiness and other environmental variations. The experiments performed during this research project are
listed in Table 1.

3. Results and discussion

3.1. Monitored parametersExample: degradation of pcoumaric acid (experiment E4)
HPLC measurements demonstrated that p-coumaric
acid oxidised rapidly into intermediates. The dark Fenton reaction had already degraded 30% of the original
phenolic compound in the rst 5 min. After 20 min of
irradiation more than 90% of the initial amount was
removed (cf. Fig. 2). Degradation was also comparably
fast or even faster in all the other experiments with single
substances.
The phenol index is commonly measured with the 4aminoantipyrine reagent (Council directive, 1979), but
this has a poor or no response to para-substituted

74

W. Gernjak et al. / Chemosphere 50 (2003) 7178

Table 1
Experiments performed
Experiment

Performed

Reactor

Volume (l)

Substance

H2 O2 (mM)

Iron [mM]

E1
E2
E3
E4
E5
E6
E7
E8
E9
E10
E11
E12
E13
E14

Vienna
Vienna
Vienna
Vienna
Vienna
Vienna
PSA
PSA
PSA
PSA
PSA
PSA
PSA
PSA

Flask
Flask
Flask
Flask
Flask
Flask
FFR
FFR
FFR
FFR
FFR
FFR
FFR
FFR

2
2
2
2
2
2
19
19
19
19
19
19
19
19

17
12
13
19
23.5
18
17
17
17
12
13
19
23.5
85

0.1
0.1
0.1
0.1
0.1
0.1
0.05
0.2
1.0
0.2
0.2
0.2
0.2
0.2

E15
E16
E17

PSA
PSA
PSA

CPC
CPC
CPC

35
35
35

Vanillin
Gallic acid
Protocatechuic acid
Coumaric acid
L -Tyrosine
Syringic acid
Vanillin
Vanillin
Vanillin
Gallic acid
Protocatechuic acid
Coumaric acid
L -Tyrosine
Mixture of vanillin, gallic,
protocatechuic, coumaric acid
and L -tyrosine
Vanillin
Gallic acid
Protocatechuic acid

17
12
13

0.05
0.05
0.05

Initial concentration of phenolic compounds was 1 mM in all tests (1 mM each in the mixture).

Fig. 2. Degradation of p-coumaric acid and carbon mass balance (Experiment E4) in laboratory reactor.

phenolic compounds (Box, 1983). It was therefore

measured with the also widely employed (Box, 1983)
Folin-Ciocalteau phenol reagent (Singleton and Rossi,
1965), which is based on the reduction of phosphomolybdic and phosphotungstic acids in alkali media by
phenolate anions. Partly reduced heteropolyacids form
molybdeniumtungsten blue, which is measured colorimetrically. Caeic acid was used for calibration. However, dierent phenolic compound respond dierently to
this method, depending, among other things, on the
number of hydroxyl moieties on the aromatic ring (Box,
1983), as shown in Fig. 3.
Fig. 3 depicts the phenol index and TOC during the
degradation of gallic acid (initial concentration 1 mM,
Experiment E2) and syringic acid (initial concentration 1
mM, Experiment E6). The phenol index measured at the

Fig. 3. Phenol index: gallic vs. syringic acid in laboratory reactor (Experiments E2 and E6).

beginning was 130 mg/l for gallic acid (instead of

170 mg=l 1 mM) and 80 mg/l for syringic acid (instead
of 198 mg=l 1 mM). In the experiment with gallic
acid, the phenol index decreased rapidly, because the
aromatic ring already has three hydroxyl groups and
further oxidation opens the ring. Syringic acid has only
one hydroxyl group. In the rst phase of its degradation,
more hydroxylated intermediates are formed, which react more strongly with the Folin-Ciocalteau phenol
reagent. Although apparently illogical, the amount of
phenolic compound (measured by phenol index) increases at the beginning of the experiment, but this is
only because of the hydroxylation of the syringic acid
ring. Thus, the phenol index is a useful, but only approximate way of estimating phenolic compounds con-

W. Gernjak et al. / Chemosphere 50 (2003) 7178

centration when no more sophisticated method (such as

HPLC) is available.
The phenol index and COD dropped more rapidly
than the TOC (cf. Fig. 2), however, the correlation between the two parameters in this case, could not be
observed in other experiments. The correlation between
peroxide consumption and COD decrease was reproducible in all experiments. An interesting point is that
during the rst few minutes, in spite of a signicant reaction, there was no, or very little, decrease in peroxide.
Usually, TOC degradation was slower at the beginning and accelerated towards the end, as is typically the
case in most of the degradation experiments. On the
other hand, due to the mineralisation process, IC rose
continuously, reaching 40% of the initial TOC at the
end. An hundred percent has never been reached due to
the acid pH of the process. As TOC analysis is quicker
than COD, it was preferred for monitoring of the experiments.

75

most other experiments. The problem is the resulting

iron-separation step required before further treatment or
discharge after the photocatalytic treatment. Normally
the corresponding iron concentrations used are not
much higher than the limits for direct discharge to a
biological municipal wastewater treatment plant (in the
range of tens of mg/l, dierent amounts are permitted in
the EU, USA, Switzerland, etc.). On the other hand, it
demonstrates that even with very low iron concentrations (0.05 mM) mineralisation is possible. Also, in the
degradation of a mixture of phenolic compounds (Experiment E14) a TOC of more than 540 mg/l was eciently degraded with just 0.2 mM iron (cf. Section 3.6).
In the single substance experiment series, the iron
concentration was chosen in the way that it provided
appropriate reaction rates (not too high) for correct,
reliable and immediate analysis. This diered among
reactors due to dierences in reactor performance.

3.2. Iron concentration

3.3. Laboratory experiments in Vienna: single substance

comparison

The reaction rate is a crucial parameter for evaluation of the application costs, because it determines the
size of the photoreactor required for treatment of a
given wastewater and, therefore, signicantly inuences,
not only the investment, but also the operating costs.
Consequently faster reaction rates are usually desirable
(Goswami et al., 1997) and, depending on the iron
concentration, the reaction has a maximum rate.
Fig. 4 shows the degradation of TOC for vanillin
with three dierent iron concentrations (Experiments
E7, E8 and E9, cf. Table 1). By approximately quadrupling the iron concentration from 0.05 to 0.2 mM, the
incident UV-radiation per litre necessary to achieve almost complete mineralisation was halved. A similar effect was reached by quintupling the iron concentration
to 1 mM. This shows the potential for increasing reaction rates with higher iron concentrations than used in

Complete decomposition of original substances in all

the experiments (Experiments E1-E6) after 1 h irradiation was demonstrated by HPLC analyses. The dark
reaction alone had already removed 4090% of the initial concentration. The decrease in the total phenol index
was not as rapid, most likely due to the formation of
hydroxylated by-products during the degradation process (cf. Section 3.1.). Therefore, in those compounds
having only one hydroxyl group (vanillin, syringic acid)
the total phenol index even increased at the beginning.
The degradation of TOC with regard to the total
removal of organic matter in the solution was considered the main aim of the experiments. TOC degradation
in the experiments with single substances is depicted in
Fig. 5. It has been demonstrated in longer experiments
(not shown in the gure) that given sucient treatment
time, all substances can be completely mineralised. The

Fig. 4. Eect of the iron concentration. Vanillin 1 mM and

H2 O2 17 mM in FFR.

Fig. 5. Degradation of TOC in single substances (Experiments

E1E6) in laboratory reactor.

76

W. Gernjak et al. / Chemosphere 50 (2003) 7178

six model compounds may be ordered by degradation

time as follows: Gallic acid > protocatechuic acid >
coumaric acid
l-tyrosine > syringic acid > vanillin.
The dierence in behaviour of TOC degradation rate
during the course of the experiment is considerable.
TOC for gallic acid and L -tyrosine showed steady degradation, while protocatechuic and coumaric acid TOC
degradation rates slowed down at the end of the experiment and vanillin accelerated. Finally, the behaviour
of syringic acid diered even more, indicating the formation of hard to degrade by-products. Depending on
the substance, such by-products are formed in dierent
phases of the degradation process. Further investigation
concerning the nature of this behaviour would be of
interest, but was not within the scope of this study,
which primarily attempted to demonstrate the degradation of selected model compounds.

et al., 1996), 10 kJ/l incident radiation in the pilot plant

would be approximately equivalent to 40 min in a ask
experiment. Another critical point of these approximate
calculations is that the quantum eciency of the photoFenton reaction is not the same at all wavelengths.
The geometries and materials in the reactors are also
quite dierent, leading to dierent eciencies with regard to radiation. Another dierence is that the mixture
with air is more homogeneous in the FFR and the CPC
due to more violent agitation. In the ask experiment,
IC rose to 50% of initial TOC with substances showing
quick degradation. This is obviously a disadvantage, as
bicarbonate ions are known radical scavengers. This
phenomenon was observed to a lesser extent in the CPC
and not at all in the FFR (cf. Section 3.1.).

3.4. Scale-up from laboratory to pilot plant

As in the ask experiments, the original substance

degraded very quickly in the FFR (E8, E10E13) and in
the CPC (E15E17). The total phenol index also behaved similarly. The shape of the TOC degradation
curves diered from those obtained in the ask experiments (compare Figs. 5 and 6). The TOC degradation
rate did not vary as strongly during the course of the
experiments in the FFR and only in the nal phase did
the reaction slow down considerably. This indicates that
by-products formed in the process, which have an inhibiting eect under ask reaction conditions, do not
inhibit degradation in the FFR. This might be attributed
to the FFR geometry, since because of the short path
length of the incident radiation and the laminar ow of
the solution over the irradiated area, intermediates
generated do not absorb enough light to inhibit disappearance of TOC. This implies that the FFR photoreactor is not appropriate for very dilute solutions. This
did not occur in the CPC experiments.
Even more obvious was that the ve phenolic compounds tested did not require the same amount of inci-

where QUVn , QUVn1 are irradiated UV-energy of samples

n and (n  1) (kJ/l), Dtn is the time interval between two
sampling points (s), UVGN is the average incident radiation on the irradiated area (W/m2 ), A is the irradiated
area (m2 ) and V is the total plant volume (l).
It should be noted that the PSA radiometer measures
only the UV bandwidth of the solar spectrum, while the
photo-Fenton reaction also makes use of part of the
visible spectrum, thereby limiting possibilities for comparison of laboratory and pilot-plant experiments. 10
kJ/l incident UV-radiation in the pilot-plant reactors
would therefore equal ask irradiation time of around 8
min with 40-W lamp output radiation. Assuming that, in
fact, the radiation available in the solar spectrum for the
photo-Fenton reaction (Bauer et al., 1999) is about ve
times higher than the UV-radiation measured (Curc

o

Fig. 6. TOC degradation of single substances (Experiments E8,

E10E13) in the FFR.

There are some fundamental dierences between the

experiments in the ask and in the pilot plants. First of
all, dierent iron concentrations were chosen for the
reactors (cf. Table 1 and Section 3.2.) to produce a
reasonable reaction time for the analytical procedures.
Second, the source of radiation is dierent. In the laboratory, a 400-W medium-pressure mercury lamp, with
40 W main output radiation in the 300400 nm range,
was used for the reactor geometry.
The pilot-plant reactors were irradiated by sunlight.
Obviously, the intensity of the solar radiation depends
on the season, length of day and weather. So it is not
only advisable, but also more correct, to use the incident
radiation per litre solution (kJ/l) as the abscissa instead
of reaction time to allow comparison of experiments.
The incident radiation can be calculated by using Eq. (4)
(Malato et al., 2000):
QUVn QUVn1

Dtn UVGN A
1000 V

3.5. Pilot-plant experiments: single substance comparison

W. Gernjak et al. / Chemosphere 50 (2003) 7178

dent energy to achieve nal TOC degradation. The ve

substances can be ordered by TOC degradation speed
as follows: Gallic acid initial COD: 190 mg/l >
protocatechuic acid initial COD: 256 mg/l > vanillin
initial COD: 286 mg/l > p-coumaric acid initial COD:
271 mg/l > l-tyrosine initial COD: 323 mg/l. This
behaviour, which corresponds quite well with initial the
COD as indicated, was conrmed in the three CPC experiments.
The most outstanding dierence between the FFR
and the laboratory ask was in the experiments with
vanillin, which was the most resistant in the ask. A
likely explanation is, that, as vanillin has signicant
absorbance at higher wavelengths (and, therefore, its
possible reaction intermediates also) than the other
model compounds, a quenching eect is produced by
vanillin and its intermediates in the wavelength range of
the UV-lamp. Fig. 7 shows the absorption spectra of the
compounds tested. As the photo-Fenton reaction makes
use of light up to wavelengths of about 580 nm (Bauer
et al., 1999), a major part of the radiation available in
the solar spectrum and capable of inducing a reaction is
not aected by this quenching eect (Curc

o et al., 1996).
This could also explain the more constant TOC degradation rates mentioned above.
Furthermore, it may be observed that although pcoumaric acid also absorbs at higher wavelengths, there
is no dierence between the lab reactor (illuminated with
UV-lamp) and the solar reactor. This means that parent
compound light absorption is not a crucial parameter
for its mineralisation behaviour. It is therefore thought
that light absorption by vanillin intermediates is related
to its low mineralisation rate in the lab reactor.
Thus two statements may be made. In the FFR and
the CPC all selected model compounds could be degraded in a short time, with just a small concentration of
iron to permit correct analysis of the experimental
samples by slowing down the reaction rates. The use of
sunlight is not only an advantage in so far as it saves

77

Fig. 8. Degradation of mixture (Experiment E14) of soluble

phenols in FFR.

electricity, and thereby money, but also increases the

process quantum eciency.
3.6. Pilot-plant experimentmixture of ve soluble
phenolic compounds (E14)
To simulate a more complex matrix, such as could
appear in real wastewater, another experiment with a
mixture of all ve soluble phenolic compounds was
performed. Phenolic compounds concentration was 1
mM each, that is, ve times the amount of phenolic
compounds in single substance tests, while iron concentration remained the same. Decomposition of the
original substances was again nearly complete after 25
min of reaction time (1.1 kJ/l incident radiation). Surprisingly, degradation of protocatechuic acid was the
slowest of the ve. This was conrmed in a second run
under the same conditions. Therefore, degradation of
the substances in the mixture diered from the singlesubstance experiments in only one case. This conrms
that degradation rates cannot easily be predicted in a
complex matrix. Nevertheless, in this experiment the
main goal (total mineralisation of TOC) was still
achieved in a reasonable length of time. Also, the same
total amount of contaminants, as in the ve singlesubstance experiments, was degraded with less incident
UV-radiation since the solution was more concentrated
(cf. Figs. 6 and 8).

4. Conclusions

Fig. 7. UV spectra of the selected phenols.

All the model phenolic compounds tested were easily

degraded by means of the photo-Fenton reaction in
experiments with single substances, as well as in a more
complex matrix simulated by a mixture of phenols.
The total phenol index dropped much more rapidly
than the TOC. As the phenolic content is mainly responsible for the toxicity of many dierent euents,
photo-Fenton might well be a good choice for treatment

78

W. Gernjak et al. / Chemosphere 50 (2003) 7178

of this type of wastewater, reducing its harmful eect.

This could be achieved in a shorter time than total
mineralisation without requiring large amounts of hydrogen peroxide.
In a comparison of experiments with UV-lamp in the
laboratory and with natural sunlight in the pilot plant,
performance under sunlight was observed to be superior,
another important fact for possible industrial application.

Acknowledgements
The authors wish to thank the EC-DG Research
Improving Human Potential Programme (number of
contract HPR I-CT-1999-00013) and the LAGAR
Project (EC Fourth Framework Programme, number of
contract FAIR-CT98-3807, DG12-SSMI). The authors
also thank Mrs. Deborah Fuldauer for the correction of
the English.
References
Alfano, O.M., Bahnemann, D., Cassano, A.E., Dillert, D.,
Goslich, R., 2000. Photocatalysis in water environments
using articial and solar light. Catalysis Today 58, 199
230.
Bauer, R., Waldner, G., Fallmann, H., Hager, S., Klare, M.,
Krutzler, T., Malato, S., Maletzky, P., 1999. The photoFenton reaction and the TiO2 /UV process for waste water
treatmentNovel developments. Catalysis Today 53, 131
144.
Benitez, F.J., Beltran-Heredia, J., Acero, J., Gonz
alez, T., 1997.
Simultaneous photodegradation and ozonation plus UV
radiation of phenolic acidsMajor pollutants in agroindustrial waste waters. Journal of Chemical Technology
and Biotechnology 70, 253260.
Blanco, J., Malato, S., Fern
andez, P., Vidal, A., Morales, A.,
Trincado, P., Oliveira, J.C., Minero, C., Musci, M., Casalle,
C., Brunote, M., Tratzky, S., Dischinger, N., Funken,
K.-H., Sattler, C., Vincent, M., Collares-Pereira, M.,
Mendes, J.F., Rangel, C.M., 2000. Compound parabolic
concentrator technology development to commercial solar
detoxication applications. Solar Energy 67, 317330.
Blanco, J., Malato, S., 2001. Solar Detoxication. UNESCO,
Natural Sciences, World Solar Programme 19962005.
Available from <http://www.unesco.org/science/wsp>.
Borja, R., Banks, C.J., Maestro-Dur
an, R., Alba, J., 1996. The
eects of the most important constituents of olive mill
wastewater on batch anaerobic methanogenesis. Environmental Technology 17, 167174.
Box, J.D., 1983. Investigation of the Folin-Ciocalteau phenol
reagent for the determination of polyphenolic substances in
natural waters. Water Research 17, 511525.
Bruneton, J., 1986. Elementos de toqu
mica y de farmacognosia. Editorial Acribia, S.A. Zaragoza, Espa~
na, pp. 176
181.

Council Directive 79/869/EEC of 9 October 1979 concerning

the methods of measurement and frequencies of sampling
and analysis of surface water intended for the abstraction of
drinking water in the Member States. Available from <http:
//europa.eu.int/eur-lex/en/consleg/pdf/1979/en_1979L0869_
an_001.pdf >.
Curc

o, D., Malato, S., Blanco, J., Jim
enez, J., Marco, P., 1996.
Photocatalytic degradation of phenol: Comparison between
pilot-plant scale and laboratory results. Solar Energy 67,
387400.
Gonz
alez-Vila, F.J., Verdejo, T., Mart
n, F., 1992. Characterization of wastes from olive and sugarbeet processing
industries and eects of their application upon the organic
fraction of agricultural soils. International Journal of
Environmental Analytical Chemistry 46, 213222.
Goswami, D.Y., Sharma, S.K., Mathur, G.D., Jotshi, C.K.,
1997. Techno-economic analysis of solar detoxication systems. Journal of Solar Energy Engineering 119, 108113.
Herrera, F., Pulgarin, C., Nadtochenko, V., Kiwi, J., 1998.
Accelerated photo-oxidation of concentrated p-coumaric
acid in homogeneous solution. Mechanistic studies, intermediates and precursors formed in the dark. Applied
Catalysis B: Environmental 17, 141156.
Huston, P.L., Pignatello, J.J., 1999. Degradation of selected
pesticide active ingredients and commercial formulations in
water by the photo-assisted fenton reaction. Water Research
33, 12381246.
Legrini, O., Oliveros, E., Braun, A.M., 1993. Photochemical
processes for water treatment. Chemical Reviews 93, 671
698.
Malato, S., Blanco, J., Herrmann, J.M., 1999. Solar catalysis
for water decontamination. Catalysis Today 54, special issue
with 18 articles.
Malato, S., Blanco, J., Richter, C., Fern
andez, P., Maldonado,
M.I., 2000. Solar photocatalytic mineralization of commercial pesticides: Oxamyl. Solar Energy Materials Solar Cells
64, 114.
Maureen, A., Ram
rez, E., Garraway, J.L., 1982. Plant growth
inhibitory activity of extracts of raw and treated pig slurry.
Journal of Science of Food and Agriculture 33, 11891196.
Miranda, M.A., Galindo, F., Amat, A.M., Arques, A., 2000.
Pyrylium salt-photosensitized degradation of phenolic contaminants derived from cinnamic acid with solar light.
Correlation of the observed reactivities with uorescence
quenching. Applied Catalysis B: Environmental 28, 127133.
Miranda, M.A., Galindo, F., Amat, A.M., Arques, A., 2001.
Pyrylium salt-photosensitized degradation of phenolic contaminants present in olive oil wastewaters with solar light.
Part II. Benzoic acid derivatives. Applied Catalysis B:
Environmental 30, 437444.
Moreno, M.T., 1994. Estudio del residuo de la industria del
corcho como sustrato hort
cola. Ph.D. dissertation, University of Sevilla, Spain.
Singleton, V.L., Rossi, J.A., 1965. Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents.
American Journal of Enology and Viticulture 16, 144158.
Tello, E., 2001. Optimizaci

on de tecnolog
as fotocatal
ticas de
oxidaci

on avanzadas al tratamiento de residuos l
quidos del
laboratorio. Ph.D. dissertation, University of Las Palmas
de Gran Canaria, Spain.