Science of the Total Environment 648 (2019) 601–608

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Comparison of different detoxification pilot plants for the treatment of

industrial wastewater by solar photo-Fenton: Are raceway pond reactors
a feasible option?
A. Cabrera-Reina a,⁎, S. Miralles-Cuevas a, G. Rivas b,c, J.A. Sánchez Pérez b,c
a
University School of Mechanical Engineering (EUDIM), University of Tarapacá, Avda. General Velásquez 1775, Arica, Chile
b
Chemical Engineering Department, University of Almería, Ctra. Sacramento s/n, Almería, Spain
c
Centro de Investigaciones de Energía Solar (CIESOL), Joint centre University of Almería – CIEMAT, Ctra. Sacramento s/n, Almería, Spain

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Treatment capacities of different pilot

plants for solar photo-Fenton are com-
pared.
• Raceway pond is used to treat industrial
wastewater at different liquid depths.
• Flat collector and CPC photoreactor are
used to treat industrial wastewater.
• Economic treatment capacity is defined
for photoreactor comparison.

a r t i c l e i n f o a b s t r a c t

Article history: This paper represents a first approach in the study of photoreactor selection to treat industrial wastewater
Received 25 April 2018 using solar photo-Fenton. In this context, simulated textile industry effluent containing a mixture of four
Received in revised form 24 July 2018 dyes at different initial dissolved organic carbon (DOC) concentrations (45, 90, 180 and 270 mg/L) was
Accepted 10 August 2018
treated by using three different solar reactor geometries: (i) tubular (5 cm diameter) provided with com-
Available online 15 August 2018
pound parabolic collector (ii) tubular (5 cm diameter) provided with flat collector and (iii) open channels
Editor: Paola Verlicchi forming raceway ponds with two liquid depths (5 and 15 cm). For comparison purposes, mineralisation
percentages over 75% and chronic toxicity reduction were set as treatment goals. Regardless of the initial
Keywords: DOC concentration, negligible differences in terms of treatment time and hydrogen peroxide consumptions
Solar photoreactor were found between the flat collector and compound parabolic collector photoreactors. Conversely, the
Photo-Fenton treatment in the raceway pond reactors always resulted in higher values. In spite of this, when the
Dye wastewater photoreactors were compared in terms of treatment capacity (mg of DOC removed/m2 min) the raceway
CPC pond reactor at 15 cm of liquid depth presented the best results, with values as much as two or three
Raceway
times higher than those of the tubular reactors, except for the wastewater with 270 mg of DOC/L for
which the raceway pond reactor at 5 cm liquid depth became the best option. When the treatment capacity
is modified to include the photoreactor investment (mg of DOC removed/€ m2) the differences between the

⁎ Corresponding author.
E-mail address: acabrera@limza.cl (A. Cabrera-Reina).

https://doi.org/10.1016/j.scitotenv.2018.08.143
0048-9697/© 2018 Elsevier B.V. All rights reserved.
602 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608
raceway pond reactor at both liquid depths and the photoreactors with solar collectors increased by as
much as two orders of magnitude, which demonstrates the potential application of the former for the treatment
of industrial wastewater by solar photo-Fenton.
1. Introduction
Any industrial application of a photocatalytic process requires a
photoreactor, which is distinctive from classic reactors because the
rates of photochemical reactions are functions of the position in the
reaction zone. This is due to the unavoidable non-uniformity of the
radiation field caused by photon absorption (homogeneous media) or
absorption and scattering (heterogeneous media) (Alfano et al., 2009).
Water treatment photoreactors can essentially be divided into two
different types: (i) Concentrating reactors (ii) Non-concentrating
reactors (Malato et al., 2007). Within the first category, the most used
photoreactors were probably parabolic-trough collector (PTCs), which
consist of a structure that supports a reflective concentrating parabolic
surface. Although the PTC has been proved effective for wastewater
treatment (Spasiano et al., 2015), its use has been neglected due to
high cost. On the other hand, Compound Parabolic Collectors (CPC)
considered non-concentrating devices or low concentrating devices
(1–1.5 sun), are the most extensively used photoreactors in this field.
CPCs present two connected parabolic mirrors and an absorber tube in
the axis. Thanks to the design of the reflector surface most of the
incident radiation, including the diffuse component, can be used so
that the light is reflected through the upper part of the tube, favouring
the complete illumination of the system (Goswami et al., 1997).
In spite of the importance of radiation distribution in the tubes, it has
been reported that the photo-Fenton process can reach photosaturation
from a certain UV radiation level (Carra et al., 2014a) and also that the
production of radical excess can be counter-productive for the process
(Cabrera Reina et al., 2017). These statements are extremely important
as photo-Fenton could be also efficiently run with more simple
photoreactors selecting adequate operation conditions. In this sense,
very interesting works have been recently carried out using non-
concentrating low cost raceway pond reactors (RPR) for the treatment
of contaminants of emerging concern (μg/L–ng/L) (De Obra et al.,
2017; Arzate et al., 2017; Carra et al., 2014b). The results have proved
that, for micropollutant removal by solar photo-Fenton, UV light excess
likely occurs when using tubular reactors and, therefore, increasing the
path length could be a good solution to improve process efficiency.
Industrial wastewater treatment requires a much higher amount of
radicals as the pollutants are present in the order of mg/L and, conse-
quently, the above-explained considerations are not valid for this par-
ticular case. So far, the CPC photoreactor is considered the best choice
when treating industrial (toxic and non-biodegradable) wastewater
by solar photo-Fenton although some key points could also support
the use of different detoxification plants. Industrial wastewater is
characterised by its important variability. A wide range of characteris-
tics accounting for inorganic salts, dissolved organic carbon (DOC) con-
centration, volatile compounds, suspended solids, or colour, among
many others can be found depending, mainly, on the industry sector
(Doumic et al., 2015; Chatzisymeon et al., 2006).
Although major efforts are being made to study new photoreactor
configurations, this research field seems to be still in its infancy,
especially considering that most of the photoreactor development
carried out in the 90s was done based on heterogeneous (TiO2)
photocatalysis experiments. While 333 scientific papers can be found
when searching for the terms “photo-Fenton + CPC” in a specific
search engine (www.sciencedirect.com), only 24 were found when
searching for “photo-Fenton + Raceway”. Furthermore, intermediate
configurations between RPRs and CPC photoreactors like the use of
© 2018 Elsevier B.V. All rights reserved.
simpler solar collectors, for instance the flat collector (FC), remain
almost unexplored.
This paper represents a first step in the study of photoreactor
selection to treat industrial wastewater using solar photo-Fenton. In
this context, simulated textile industry effluent containing a mixture
of four dyes at different initial DOC concentrations (45, 90, 180 and
270 mg/L) was treated using three different solar detoxification pilot
plants: (i) CPC photoreactor (ii) FC photoreactor and (iii) RPR. Two liq-
uid depths (5 and 15 cm) were used for RPR while 5 cm tube diameter
was used for tubular reactors. With the aim of achieving mineralisation
percentages over 75% and reducing chronic toxicity, at least, to slight
levels, the photoreactor performances were compared in terms of treat-
ment time, hydrogen peroxide consumption, accumulated energy (QUV)
and treatment capacity (TC) expressed as the mass of mineralised DOC
per unit of time and photoreactor surface unit. Also a new parameter,
the economic treatment capacity (TCE) was defined.
2. Materials and methods
2.1. Reagents and analytical methods
Textile industry effluent was simulated by dissolving a mixture of
four dyes with each compound providing equal DOC concentration to
wastewater. The dyes were purchased from Sigma-Aldrich, in tap
water: (i) Acid Red 1 (CAS Number 3734-67-6, Molecular Weight
(MW) 509.42 g/mol), (ii) Acid Yellow 17 (CAS Number 6359-98-4,
MW 551.29 g/mol), (iii) Reactive Black 5 (CAS Number 17095–24-8,
MW 991.82 g/mol) and (iv) Reactive Orange 16 (CAS Number 12225-
83-1, MW 617.54 g/mol). Their chemical structures are shown in
Fig. 1. The catalyst, Fe(II), was added in the reaction bulk as FeSO4⋅7H2O
(99%), which was obtained from Fluka. The oxidant agent, hydrogen
peroxide (H2O2; 33% w/v), was provided by Sigma-Aldrich. Sulphuric
acid (95–97%), used for pH adjustment, was purchased from J.T. Baker.
The decontamination process was followed measuring the DOC con-
centration using a Shimadzu-VCHP TOC analyser fitted with an auto-
sampler system. The hydrogen peroxide concentration was determined
with the ammonium metavanadate colorimetric method proposed by
Nogueira et al. 2005 (Nogueira et al., 2005) while iron species concen-
tration in water was measured by means of the o-phenanthroline
standardised spectrometric method (ISO 6332). Absorbance measure-
ments were carried out with a Hach Lange DR 5000 spectrophotometer.
2.2. Solar detoxification pilot plants
The experiments were carried out in three types of photoreactors
installed at CIESOL, Centro de Investigaciones de la Energía Solar,
Almería (Spain):
a) CPC photoreactor: this solar detoxification plant consist of two twin
loops each containing two Pyrex glass tubes fitted onto the focus of
two CPC mirrors, tilted 37° (local latitude). The total illuminated sur-
face per loop is 0.42 m2, the illuminated volume is 4.77 L and total
volume is 7 L. The simulated industrial effluent is recirculated
through the system (reservoir tank, piping and glass tubes) by
means of a centrifugal pump.
b) FC photoreactor: this solar detoxification plant follows exactly the
same scheme as the previous one but using a flat mirror, also tilted
37° (local latitude), instead of CPCs.
 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608
Fig. 1. Chemical structures of Acid Red 1, Acid Yellow 17, Reactive Black 5 and Reactive Orange 16. (For interpretation of the references to colour in this figure legend, the reader is referred
to the web version of this article.)
c) RPR: these pilot plants consist of two Polyvinyl Chloride (PVC) RPRs
whose total volumes are 18 L (5 cm liquid depth) and 80 L (15 cm
liquid depth). The illuminated surfaces are 0.36 m2 for 5 cm liquid
depth and 0.53 m2 for 15 cm liquid depth. Each RPR includes a
paddlewheel connected to an engine to obtain good mixing condi-
tions and homogeneous flow during the treatment.
As photochemical processes depend on the accumulated energy
arriving at the reactor surface instead of treatment time, the solar UV ra-
diation (W/m2) was measured using a Delta Ohm radiometer (model LP
UVA 02 AV) with a 327–384 nm spectral response range. This device
was mounted on a horizontal platform when the RPRs reactors were
used while for the CPC and FC photoreactors it was mounted on a plat-
form tilted 37°. The pH and temperature were monitored by means of
CRISON 5335 sensor. These instruments, sensor and UV radiometer,
were connected to a data acquisition card (Labjack U12) to allow the
measurements to be stored on a computer.
2.3. Experimental procedure
In a typical run, the corresponding treatment volume (7 L for the CPC
and FC photoreactors, 18 L for the RPR with 5 cm liquid depth and 80 L
for the RPR with 15 cm liquid depth) containing the four-dye mixture
and the catalyst was loaded in the corresponding pilot plant. Notice
that before iron addition, the pH was adjusted to 2.8 in order to avoid
its precipitation. After 5 min of dark recirculation in the photoreactor,
the initial amount of H2O2 was added to the reaction bulk and, immedi-
ately afterwards, the photoreactor was uncovered initiating the experi-
ment and data acquisition.
The selected initial iron concentration -Fe(II)- was 20 mg/L, as this is
considered the optimal concentration when using the solar photo-
Fenton process in CPC photoreactors with a 5 cm tube diameter
(Malato et al., 2004; Malato et al., 2009). The initial H2O2 concentration
was established at 600 mg/L and, afterwards, the remaining concentra-
tion in solution was determined in each sample adding 400 mg/L every
time this value was below 200 mg/L. In this way, the assays were always
run under H2O2 excess assuring that this factor never limited the
603
decontamination process. On the other hand, the initial DOC was
selected between 45 and 270 mg/L based on two studies found in the
bibliography (Doumic et al., 2015; Chatzisymeon et al., 2006). Both
Doumic et al., 2015 and Chatzisymeon et al., 2006 use simulated textile
effluent according to the information obtained from textile dyeing
companies, Erfoc-Acabamentos Texteis S.A. (Portugal) and EPILEKTOS
S.A. (Greece), respectively. The sampling times were: 0, 5, 10 min and,
from that moment, every 10 min until the end of the assays. The
experiments were carried out during the end of July and the beginning
of August (2017) and each experiment was started between 12:00 noon
and 1:00 pm so that most of them can be directly compared in terms of
treatment time as negligible differences between average UV radiation
values during the experiments were found (see Table 1). The operation
temperatures were also very similar in all the assays, although slightly
lower in the RPRs experiments (2–3 °C). In any case, the accumulated
energy received at the photoreactor surface (QUV) was also calculated
(Eq. (1)) for each sample:
Ai
Q UV;nþ1 ¼ Q UV þ ðtnþ1 −tn Þ∙UV∙ ð1Þ
V
Where the subscript “n” refers to the sample number, UV (W/m2) is
the average UV solar radiation measured between treatment times
tn and tn+1, Ai (m2) is the irradiated surface of the photoreactor and
V (L) corresponds to the total volume of wastewater loaded in the
pilot plant.
2.4. Chronic toxicity bioassay
Simulated textile effluents were tested before and after photo-
Fenton treatment in a concentration series of 100, 50, 25, 12.5 and
6.25% (v/v) to evaluate chronic toxicity. Each test was repeated at
least two times. The pH was adjusted to pH ~ 7, and the remaining
hydrogen peroxide was removed with bovine liver catalase. Toxicity re-
sults were expressed as a percentage of toxic effect (inhibition).
Chronic toxicity was evaluated using the commercial Protoxkit F™
toxicity test. The growth inhibition of the ciliate protozoo Tetrahynema
termophila after 24 h exposure was performed following the Protoxkit
F (1998) standard operational procedure. The test is based on the
604 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608

Table 1
Average UV radiation values during the experiments.

Wastewater initial DOC concentration

45 mg/L 90 mg/L 180 mg/L 270 mg/L

CPC photoreactor 27.9 ± 1.1 W/m2 27.7 ± 1.3 W/m2 26.1 ± 1.7 W/m2 26.1 ± 2.3 W/m2
FC photoreactor 26.4 ± 0.9 W/m2 26.6 ± 1.5 W/m2 25.6 ± 2.5 W/m2 21.1 ± 2.7 W/m2
RPR at 5 cm liquid depth 24.1 ± 1.7 W/m2 24.1 ± 1.7 W/m2 25.8 ± 1.9 W/m2 26.8 ± 2.6 W/m2
RPR at 15 cm liquid depth 24.3 ± 0.9 W/m2 24.7 ± 1.4 W/m2 26.0 ± 1.9 W/m2 26.4 ± 2.3 W/m2

turnover of substrate into ciliate biomass. While normal proliferating
cell cultures clear the substrate suspension in 24 h, inhibited culture
growth is reflected by the remaining turbidity. Optical density measure-
ments of the turbidity quantify the degree of inhibition. Freshwater
was used as negative control and K2Cr2O7 (56 mg/L Cr6+) was used as
positive control (Freitas et al., 2017).
3. Results and discussion
3.1. Evaluation of the Fenton reaction
The Fenton contribution for each industrial wastewater (45, 90, 180
and 270 mg of DOC/L) was evaluated prior to the solar photo-Fenton
decontamination assays. As this process occurs in the dark, the
experiments were only carried out in the CPC photoreactor. In all the
cases, 20 mg of Fe/L and 600 mg of H2O2/L were selected as initial
reagent concentrations while (ambient) temperature was almost
constant (30–32 °C) during each experiment.
Independent of the wastewater type, there was an important DOC
initial decrease and, afterwards, the DOC concentration remained con-
stant for the next 25 min (Fig. 2). At that time, the experiments were
concluded resulting in 24.3%, 22.7%, 21.2% and 17.1% final
mineralisation percentages for the 45, 90, 180 and 270 mg of DOC/L
wastewaters, respectively. The mineralisation profiles responded to
typical Fenton behaviour in which, initially, Fe(II) reacts with H2O2
yielding hydroxyl radicals and Fe(III). Then, the process was stopped
due to the extremely low reaction rate between Fe(III) and H2O2
(Pignatello et al., 2006). A linear increase of the DOC removed in relation
to the wastewater initial DOC concentration was observed (inset Fig. 2),
which indicates that the selected H2O2:iron ratio provided enough
initial radicals not to limit the mineralisation process.
From this point, this work is focused on the comparison of the differ-
ent pilot plants for each level of initial DOC in the simulated industrial
1 parison were established in 75% mineralisation for the 45 mg of DOC/L
0.8
50
DOC removed (mg/L)
wastewater. This comparison is evaluated in terms of: (i) treatment
time, H2O2 consumption and QUV (ii) TC and (iii) TCE.
3.2. Chronic toxicity and treatment goal selection
The objective mineralisation percentages used for the photoreactor
comparison were selected based on the results obtained in the CPC
photoreactor, which included chronic toxicity analysis. An initial and
final screening was performed to elucidate possible risks of long-term
toxicity. The samples were evaluated at least twice, but only for the
CPC photoreactor, as toxicity is independent of photoreactor geometry.
According to the effluent classification proposed by Persoone et al.,
2003 the toxicity can be represented by the percentage of inhibition
and immobilisation (percentage effect, PE). The hazard classification
system for natural waters defines Class I for PE b 20%, Class II for
20 b PE b 50%, Class III for 50 b PE b 100%, Class IV when PE = 100%
in at least one test and Class V when PE = 100% in all tests. All classes
are accompanied by a concise judgment: Class I: no chronic toxicity,
Class II: slight chronic toxicity, Class III: chronic toxicity, Class IV: high
chronic toxicity and Class V: very high chronic toxicity.
In this sense, in Table 2 the percentages of inhibition/immobilisation
for initial and final samples as well as their classification are shown.
Despite most of the toxicity tests in scientific works being performed
with Algae sp., the use of decomposers such as the ciliate protozoo se-
lected for this study (Tetrahynema termophila) represent excellent eco-
toxicological test species due to their important role in the microbial
foodweb, wide distribution and abundance (Gerhardt et al., 2010). The
initial simulated textile effluent with 45 DOC mg/L presented chronic
toxicity (Class III) while, for the rest of initial concentrations, the efflu-
ents presented high chronic toxicity (Class IV). The photo-Fenton treat-
ment decreased the chronic toxicity in each case although it could only
be completely removed (Class I) for the simulated effluent with the
lowest initial concentration (45 mg of DOC/L). When the initial DOC
concentration was equal to or higher than 90 mg of DOC/L, the solar
treatment could reduce the toxicity to Class II, i.e., slight chronic toxicity.
Based on these results, the treatment goals for the photoreactors com-
wastewater, 80% mineralisation for the 90 mg of DOC/L wastewater
and 85% mineralisation for the 180 and 270 mg of DOC/L wastewaters.
Colour was also measured in each assay; nevertheless, as complete
colour removal was achieved at low mineralisation levels this variable
was not used for the comparison of the photoreactors. These results

45
0.6
DOC/DOC0

40 also show that the inner filter effect, typically found on textile wastewa-
35
30 45 mg/L ters, can be considered a minor factor.
0.4 25 90 mg/L
20
15
180 mg/L
Table 2
10 270 mg/L
5 Percentages of inhibition (Chronic; T. thermophila) and toxicity classification.
0.2 0
0 200 Wastewater initial DOC Chronic toxicity Hazard classification
Wastewater initial DOC (mg/L) (mg/L) (%Inhibition of T. termophila) (S0/SF)
0 (S0/SF)a
0 10 20 30 45 86.3/7.3 III-I
time (min) 90 100.0/29.2 IV-II
180 100.0/38.2 IV-II
270 100.0/43.6 IV-II
Fig. 2. Fenton degradation profiles of the simulated industrial wastewater at the different
a
initial DOC concentrations studied. S0: initial sample, SF: final sample.
 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608 605

3.3. Comparison of the photoreactors in terms of treatment time and QUV

Average solar UV radiation (W/m2)

7
The degradation of the industrial wastewater with 45 mg of DOC/L 25
CPC and FC
showed no relevant difference when using the CPC or FC photoreactor. 6
The treatments reached the 75% mineralisation level after 14 min and RW 5cm 20
17 min, respectively, with equivalent degradation profiles and similar 5

H2O2 consumptions, 1096 and 972 mg/L, respectively. QUV was higher
for the FC photoreactor (1.71 kJ/L) than for the CPC photoreactor
(1.44 kJ/L). The treatments in the RPR resulted in higher treatment
times, increasing to 27 and 39 min for the 5 cm and 15 cm liquid depths,
respectively, although this increase cannot be considered significant
taking into account sampling times. Indeed, in Fig. 3 where the compar-
ison of the different degradation profiles for this wastewater is shown, it
can be seen how the degradation curve of the RPR at 5 cm liquid depth is
just slightly delayed with respect to the CPC and FC photoreactors. On
the other hand, at 15 cm liquid depth, this delay becomes more impor-
tant. In a similar way, the H2O2 consumption in the RPR at 5 cm of liquid
depth was almost equivalent to those of the CPC and FC photoreactors;
however, the consumption increased more than 40% when the RPR at
15 cm of liquid depth was used. Notice that these consumptions should
only be considered as reference values because the treatments were
carried out under H2O2 excess. This favours side reactions not related
to mineralisation but with dissolved oxygen generation increasing
H2O2 consumption (Santos-Juanes et al., 2011); consequently, these
values can be optimized. QUV in the RPRs resulted in much lower values,
0.82 kJ/L and 0.31 kJ/L for the 5 cm and 15 cm liquid depths, respectively.
It must be taken into account that QUV represents the energy arriving at
the photoreactor surface per unit of liquid volume. Under the same solar
conditions, photoreactors with different geometries and configurations
receive different amounts of energy per photoreactor surface unit.
This is represented in Fig. 4 where the variation of QUV in each type
of photoreactor during the same summer day at noon in Almería is
shown. This highlights a problem with using QUV as a variable to track
treatment advance because, although it is very useful to compare exper-
iments carried out in different days in the same photoreactor, it does not
appear to be a good option for comparing the performance of different
photoreactors. Thus, this particular case of photoreactor comparison en-
courages the necessity of exploring new reference variables.
The results corresponding to the treatment of the industrial waste-
water with 90 mg of DOC/L (Fig. 5) were very similar to those of the
45 mg of DOC/L industrial wastewater. No difference was found
between the CPC and the FC photoreactors in terms of treatment time
(24 and 22 min, respectively) and just a minor variation in H2O2
consumption (1247 and 1097 mg/L, respectively). QUV was higher in
the CPC photoreactor (2.48 kJ/L) than in the FC photoreactor
(2.16 kJ/L), which means that the process was carried out under light
QUV (kJ/L)
RW 15cm
4 15
Solar UV
3
10
2
5
1
0 0
0 20 40 60
time (min)
Fig. 4. Evolution of QUV in the different pilot plants during a typical summer day at noon.
excess as the extra 0.32 kJ/L arriving at the photoreactor surface did
not represent any improvement of the decontamination process.
Furthermore, as the CPC mirror provides better radiation distribution
than the FC mirror, this difference in the energy available between the
photoreactors was actually higher than 0.32 kJ/L. Conversely, the
treatment time needed to achieve the objective mineralisation level in
the RPRs increased to 37 min when using 5 cm of liquid depth and to
46 min when using 15 cm of liquid depth. Only the H2O2 consumption
of the 15 cm liquid depth experiment increased slightly to 1344 mg/L.
The consumption for the 5 cm of liquid depth assay resulted in
1277 mg/L, which is equivalent to that obtained in the CPC photoreactor.
In the case of the 180 mg of DOC/L industrial wastewater (Fig. 6), the
performances of the CPC and FC photoreactors were also very similar.
The treatment times to reach 85% mineralisation were 67 and 68 min
and the H2O2 consumptions were 1875 and 1892 mg/L, respectively.
The QUV value in the CPC photoreactor (6.82 kJ/L) was again higher
than in the FC photoreactor (5.72 kJ/L), therefore, it can be concluded
that the process was carried out under clear UV radiation excess in the
CPC. The RPR assays presented much higher treatment times. When
the liquid depth was 5 cm, the treatment time needed was 97 min
while, for the 15 cm liquid depth, this time was 250 min. This was the
first time in which the process was actually affected by the lower radia-
tion level and worse radiation distribution in the RPR reactors respect to
the photoreactors with solar collectors as the treatment times increased
45% and 273%, respectively. The low radical production rate for the

CPC CPC
1 1 FC
FC
RW 5 cm
0.8 RW 5 cm 0.8
DOC/DOC0

RW 15 cm
DOC/DOC0

RW 15 cm 0.6
0.6

0.4 0.4

0.2 0.2

0 0
0 20 40 60 0 20 40 60 80
time (min) time (min)

Fig. 3. Degradation profiles of the wastewater at 45 mg of DOC/L in the CPC, FC and RPR Fig. 5. Degradation profiles of the wastewater at 90 mg of DOC/L in the CPC, FC and RPR
(5 cm and 15 cm). (5 cm and 15 cm).
606 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608

148 min at 5 cm of liquid depth, which is 59 min more than the treat-
CPC ment in the CPC photoreactor. In this case it is clear that the lower radi-
1 FC ation level arriving at the photoreactor surface together with the worse
RW 5 cm radiation distribution in the RPR is significantly affecting the photo-
0.8 Fenton cycle, diminishing the amount of radicals produced and limiting
RW 15 cm
the organic matter mineralisation. This important difference in terms of
DOC/DOC0

treatment time also affected the H2O2 consumption, which at the end of
0.6
the process was also higher, 2923 mg/L. Nevertheless, QUV was still
much lower (3.10 kJ/L) than in the CPC and the FC photoreactors. On
0.4 the other hand, the experiment with 15 cm of liquid depth did not
reach the objective mineralisation level after 250 min. At that illumina-
0.2 tion time only 30% of the initial DOC could be degraded and half of this
percentage was degraded within the first 5 min when the Fenton (dark)
0 reaction took place. Thus, the feasibility of the RPR application to indus-
0 50 100 150 200 250 trial wastewater treatment seems to be limited at this point, although
time (min) diminishing the catalyst concentration could be a possible solution in
order to expand this limit of application at high liquid depths.
Fig. 6. Degradation profiles of the wastewater at 180 mg of DOC/L in the CPC, FC and RPR
(5 cm and 15 cm). 3.4. Comparison of the photoreactors in terms of treatment capacity

Treatment capacity (TC), which can be calculated using Eq. 2, is a

15 cm liquid depth allows two stages to be distinguished in the degra-
dation profile, which is in accordance with the results published by
Casas López et al. 2010 (Casas López et al., 2010). In a first step, the rad-
icals attack the dyes molecules but without mineralisation while, in a
second stage, these attacks result in DOC decrease because of the higher
simplicity of the generated organic matter. On the other hand, the dif-
ferences in H2O2 consumptions were very similar to the above-
commented experimental sets. For the 5 cm liquid depth assay the
H2O2 consumption was 2030 mg/L while it increased to 2821 mg/L for
the 15 cm liquid depth assay.
The experimental set with wastewater at 270 mg of DOC/L pre-
sented the most important differences of the whole study (Fig. 7). The
treatment time needed to achieve the objective mineralisation percent-
age in the FC photoreactor was slightly higher than in the CPC
photoreactor, resulting in 15 min of additional treatment time. Never-
theless, taking into account that the experiment in the FC photoreactor
was carried out under an average UV solar radiation of 21.1 W/m2
which is considerably lower than the average UV solar radiation of the
experiment in the CPC photoreactor, 26.1 W/m2, the increase in the
treatment time was probably caused by the different solar light levels
instead of the photoreactor characteristics. The H2O2 consumption in
the FC photoreactor was not affected, resulting in 2304 mg/L, which is
equivalent to the 2260 mg/L needed in the CPC photoreactor. With re-
spect to the treatment in the RPRs, the illumination time increased to
useful variable for the comparison of different photoreactors as it repre-
sents the amount of DOC removed per photoreactor surface unit and
time unit.
Om ∙½DOC0 ∙VR
TC ¼ ð2Þ
Ai ∙t
Where Om is the objective mineralisation level (expressed as a decimal),
[DOC]0 is the initial concentration of the wastewater (mg/L), VR is the
treatment volume (L) and t is the treatment time needed to achieve
the objective mineralisation (min).
The TC variation in the different pilot plants with the wastewater ini-
tial DOC concentration is shown in Fig. 8. The TC was always higher in
the RPRs than in the CPC and the FC photoreactors. In spite of the higher
treatment times needed when the process was carried out without solar
collectors, this increase was not high enough to compensate for the
higher volume to surface ratios of the RPRs (50 and 150 L/m2 for the
5 cm and 15 cm liquid depths, respectively) compared with that of the
CPC and FC photoreactors (16.7 L/m2). The 15 cm liquid depth always
resulted in the best option, except for the treatment of the 270 mg of
DOC/L wastewater where the objective mineralisation could not be
achieved after 250 min. For the 45 mg of DOC/L wastewater, the TC in
the RPR at 15 cm of liquid depth was twice the TC of the RPR at 5 cm
of liquid depth and more than three and four times the TCs of the CPC

CPC
CPC
FC
1 FC
(mg of DOC removed/m2 min)

RW 5 cm 200.0 RW 5cm
RW 15cm
Treatment capacity

0.8 RW 15 cm
DOC/DOC0

150.0
0.6
100.0
0.4

0.2 50.0

0 0.0
0 50 100 150 200 250 45 90 180 270
time (min) Wasterwater initial DOC concentration (mg/L)

Fig. 7. Degradation profiles of the wastewater at 270 mg of DOC/L in the CPC, FC and RPR Fig. 8. Comparison of treatment capability (TC) achieved in each pilot plant for the
(5 cm and 15 cm). different simulated industrial wastewaters.
 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608 607

and FC photoreactors, respectively. These differences were even higher Table 3

when treating the 90 mg of DOC/L wastewater as the TCs in the RPRs in- Economic treatment capacity of each photoreactor type.

creased considerably at both liquid depths while it slightly increased in Economic treatment capacity (TCE)
the CPC and FC photoreactors. The effect of Fenton is so important to (mg of DOC removed/€·min)
this initial wastewater concentration that increasing RPR depth is an ad- Wastewater initial DOC concentration (mg/L) 45 90 180 270
vantage instead of a disadvantage. For the next wastewater concentration, 2
Installation surface: 10,000 m
180 mg of DOC/L, the TC of the RPR at 15 cm liquid depth was 81.0 mg of CPC photoreactor 0.32 0.37 0.26 0.30
DOC removed/m2·min, which is very similar to the TC of the RPR at 5 cm FC photoreactor 0.31 0.50 0.32 0.32
liquid depth, 69.4 mg of DOC removed/m2·min. The process efficiency in RPR at 5 cm liquid depth 8.30 12.26 9.26 9.15
the deepest RPR is being negatively affected by the higher amount of ini- RPR at 15 cm liquid depth 11.48 19.78 7.20 –
tial organic matter, however, the TC values in the RPRs are still much Installation surface: 1000 m2
higher than the TCs of the CPC and FC photoreactors, 33.4 and 33.3 mg CPC photoreactor 0.08 0.09 0.07 0.08
of DOC removed/m2·min, respectively. Consequently, the use of RPR for FC photoreactor 0.08 0.13 0.08 0.08
RPR at 5 cm liquid depth 2.08 3.07 2.31 2.29
the treatment is still justified under these conditions even though an
RPR at 15 cm liquid depth 2.87 4.95 1.80 –
application limit can be expected from this point due to the important ef-
ficiency loss in the RPR at 15 cm liquid depth. Indeed, for the highest
wastewater initial DOC concentration, the use of the RPR at 5 cm of liquid
depth resulted in the best option as the treatment at 15 cm liquid depth wastewater by solar photo-Fenton. It should be taken into account that
could not be completed. There was almost no change of the TC the whole study was carried out in summer around noon so that light
(68.6 mg of DOC removed/m2 min) with respect to the previous waste- availability was always high, which was favourable for the FC
water DOC concentration, therefore, as the TCs of the CPC and the FC photoreactor and especially for the RPRs. Performance in the morning
photoreactors were also very similar to those of the 180 mg of DOC/L and during winter should be also studied in order to determine whether
wastewater, 38.1 and 33.1 mg of DOC removed/m2 min respectively, or not the CPC photoreactor can compensate the important TC differ-
the differences in photoreactor performance remained unchanged. The ences with respect to RPR or if the FC still presents negligible differences
low TC variation in the RPR at 5 cm is probably an indication that its ap- against the CPC under these conditions. Additional studies should also
plication limit is still a long way from being reached. Further studies in optimise reagent consumptions in each pilot plant so that this effect
order to find the initial wastewater DOC concentration for which the TC can be included in the economic treatment capacity. It is well known
of the RW at 5 cm matches the TC of the CPC and FC photoreactors that H2O2 consumption represents a very important percentage of
could be very interesting. The use of intermediate liquid depths within photo-Fenton total cost (Sánchez Pérez et al., 2013) and also that iron
the range used in this work could also provide useful information. In concentration should be optimized depending on RPR liquid depth.
any case, these results suggest the potential application of the RPR to
treat industrial wastewater. 4. Conclusions

3.5. Comparison of the photoreactors in terms of economic treatment
capacity
The reported price of CPC photoreactor is about 500 €/m2 for an in-
stallation of 1000 m2 (Miralles-Cuevas et al., 2016) while this value can
be diminished to 129 €/m2 if the size of the treatment plant is increased
to 10,000 m2 (Tânia et al., 2016). Based on these costs and the informa-
tion given by different suppliers, it has been estimated that
photoreactor cost can be reduced by 20% due to the use of a FC instead
of a CPC, i.e. the cost is 400 €/m2 and 103 €/m2 for the 1000 and
10,000 m2 plants, respectively. The cost of the RPR has been reported
as 10 €/m2 for an installation of 10,000 m2 (Carra et al., 2015) and this
is the price considered for the 5 cm RPR. The same scaling-up factor of
the CPC plant was considered to obtain the RPR cost for the 1000 m2 sys-
tem resulting in 40 €/m2. In order to be conservative and due to the
higher size of the RPR at 15 cm liquid depth, 20 €/m2 and 60 €/m2
have been considered for the 10,000 and 1000 m2 plants. Taking this in-
formation into account a new variable was defined, the economic treat-
ment capacity (TCE), which represents the amount of DOC removed per
time unit and photoreactor investment monetary unit (Eq. 3):
TCE ¼ TC=Cphotoreactor
Where Cphotoreactor is the acquisition cost of the different photoreactors
(€/m2).
The TCE values for each pilot plant and wastewater treated are
presented in Table 3. Due to the RPRs low cost, the important differences
already observed between the RPRs and the CPC and FC photoreactors
in terms of TC are even increased when the photoreactor investment
is included in the comparison varying by as much as two orders of
magnitude.
In spite of these promising results, further studies are needed to de-
termine the most adequate photoreactor for the treatment of industrial
Simulated textile industrial effluent at different DOC initial concen-
trations has been treated by solar photo-Fenton in CPC photoreactor,
FC photoreactor (5 cm diameter tubes both) and RPR at 5 cm and
15 cm liquid depths.
For the simulated wastewaters with 45 and 90 mg of DOC/L, the rel-
ative importance of the Fenton reaction together with the low treat-
ment times needed to achieve the mineralisation goals made the RPR
the best option in terms of TC, in spite of the slightly higher treatment
times obtained with respect to the tubular photoreactors. As the impor-
tance of the solar UV radiation is limited, increasing the liquid depth
from 5 cm to 15 cm allowed, approximately, doubling the TC that
reached 129.1 and 222.5 mg of DOC removed/m2 min in the latter for
the 45 and 90 mg of DOC/L wastewaters, respectively.
The same trend was found when the wastewater initial DOC was in-
creased to 180 mg/L. However, the TC of the RPR at 15 cm liquid depth
was significantly reduced so that the differences with the rest of the
pilot plants were lower. The higher amount of organic matter in the
simulated effluent increased the relative importance of photo-Fenton
in the global decontamination process so that the radiation level and
its distribution in the photoreactor started to take on greater impor-
tance, but not enough to compensate for the higher volume to surface
ð3Þ ratio of the RPR.
In the case of the highest polluted wastewater, the treatment in the
RPR at 15 cm liquid depth could not be finished; indeed, only 30%
mineralisation was achieved after 250 min. The treatment time needed
to reach the mineralisation goal in the RPR at 5 cm was almost 1 h higher
than in the CPC photoreactor. Nevertheless, the TC of the RPR was still
significantly higher than those of the tubular reactors, 68.6 mg of DOC re-
moved/m2 min vs 38.1 and 33.1 mg of DOC removed/m2 min for the CPC
photoreactor and the FC photoreactor, respectively. These results showed
the feasibility of using the RPR to treat industrial wastewater but also that
it presents an application limit, which depends on wastewater character-
istics among other variables.
608 A. Cabrera-Reina et al. / Science of the Total Environment 648 (2019) 601–608
During the whole study, negligible differences were found between
the FC and the CPC photoreactors in terms of treatment time, hydrogen
peroxide consumption and QUV. With respect to this last variable, it
seems not to be a good option to compare pilot plants when these present
different geometries or configurations, i.e., horizontal vs tilted reactors.
Finally, a new variable that includes the photoreactor investment in
the treatment capacity was defined (TCE –mg of DOC removed/€ min-).
In these terms, the differences between the RPR and the photoreactors
with solar collector were even higher, varying by as much as two orders
of magnitude. Thus, the potential application of RPR to treat industrial
wastewater by solar photo-Fenton has been proved. Further studies
where this low cost non-concentrating reactor is operated under differ-
ent UV solar radiation levels could be an interesting continuation for this
work.
Acknowledgments
This work was funded by CONICYT (FONDECYT/INICIACIÓN
11160680). The authors want to thank the Solar Energy Research Center
(CONICYT/FONDAP/15110019), Junta de Andalucía (Andalusian
Regional Government, P12-RNM-1437) and European Regional Devel-
opment Fund (ERDF).
References
Alfano, O.M., Irazoqui, H.A., Cassano, A.E., 2009. The local and observed photochemical re-
action rates revisited. Photochem. Photobiol. Sci. 8, 1047–1058. https://doi.org/
10.1039/b900191c.
Arzate, S., García Sánchez, J.L., Soriano-Molina, P., Casas López, J.L., Campos-Mañas, M.C.,
Agüera, A., Sánchez Pérez, J.A., 2017. Effect of residence time on micropollutant re-
moval in WWTP secondary effluents by continuous solar photo-Fenton process in
raceway pond reactors. Chem. Eng. J. 316, 1114–1121.
Cabrera Reina, A., Miralles Cuevas, S., Casas López, J.L., Sánchez Pérez, J.A., 2017.
Pyrimethanil degradation by photo-Fenton process: influence of iron and irradiance
level on treatment cost. Sci. Total Environ. 605–606, 230–237.
Carra, I., García Sánchez, J.L., Casas López, J.L., Malato, S., Sánchez Pérez, J.A., 2014a. Phe-
nomenological study and application of the combined influence of iron concentration
and irradiance on the photo-Fenton process to remove micropollutants Sci. Total En-
viron. 478, 123–132.
Carra, I., Santos-Juanes, L., Acién Fernández, F.G., Malato, S., Sánchez Pérez, J.A.,
2014b. New approach to solar photo-Fenton operation. Raceway ponds as tertiary
treatment technology. J. Hazard. Mater. 279, 322–329. https://doi.org/10.1016/
jhazmat.2014.07.010.
Carra, I., Sirtori, C., Ponce-Robles, L., Sánchez Pérez, J.A., Malato, S., Agüera, A., 2015. Deg-
radation and monitoring of acetamiprid, thiabendazole and their transformation
products in an agro-food industry effluent during solar photo-Fenton treatment in
a raceway pond reactor. Chemosphere 130, 73–81.
Casas López, J.L., Cabrera Reina, A., Ortega Gómez, E., Ballesteros Martín, M.M., Malato
Rodríguez, S., Sánchez Pérez, J.A., 2010. Integration of solar photocatalysis and mem-
brane bioreactor for pesticides degradation. Sep. Sci. Technol. 45, 1571–1578.
Chatzisymeon, E., Xekoukoulotakis, N.P., Coz, A., Kalogerakis, N., Mantzavinos, D., 2006.
Electrochemical treatment of textile dyes and dyehouse effluents. J. Hazard. Mater.
137, 998–1007.
De Obra, I., Ponce-Robles, L., Miralles-Cuevas, S., Oller, I., Malato, S., Sánchez Pérez, J.A.,
2017. Microcontaminant removal in secondary effluents by solar photo-Fenton at
circumneutral pH in raceway pond reactors. Catal. Today 287, 10–14. https://doi.
org/10.1016/j.cattod.2016.12.028.
Doumic, L.I., Soares, P.A., Ayude, M.A., Cassanello, M., Boaventura, R.A.R., Vilar, V.J.P., 2015.
Enhancement of a solar photo-Fenton reaction by using ferrioxalate complexes for
the treatment of a synthetic cotton-textile dyeing wastewater. Chem. Eng. J. 277,
86–96.
Freitas, A.M., Rivas, G., Campos-Mañas, M.C., Casas López, J.L., Agüera, A., Sánchez Pérez,
J.A., 2017. Ecotoxicity evaluation of a WWTP effluent treated by solar photo-Fenton
at neutral pH in a raceway pond reactor. Environ. Sci. Pollut. Res. 24, 1093–1104.
https://doi.org/10.1007/s11356-016-7101-7.
Gerhardt, A., Ud-Daula, A., Schramm, K.W., 2010. Tetrahymena spp. (Protista, Ciliophora)
as test species in rapid multilevel ecotoxicity tests. Acta Protozool. 49, 271–280.
https://doi.org/10.1111/j.1755-3768.2011.02269.x.
Goswami, D.Y., Sharma, S.K., Mathur, G.D., Jotshi, C.K., 1997. Techno-economic analysis of
solar detoxification systems. J. Sol. Energy Eng. 119, 108–113.
Malato, S., Gálvez, J.B., Rubio, M.I.M., Ibáñez, P.F., Padilla, D.A., Pereira, M.C., 2004. Engi-
neering of solar photocatalytic collectors. Sol. Energy 77, 513–524.
Malato, S., Blanco, J., Alarcon, C.D., Maldonado, M.I., Fernandez-Ibanez, P., 2007. Photo-
catalytic decontamination and disinfection of water with solar collectors. Catal.
Today 49, 122–137.
Malato, S., Fernández-Ibáñez, P., Maldonado, M.I., Blanco, J., Gernjak, W., 2009. Decontam-
ination and disinfection of water by solar photocatalysis: recent overview and trends.
Catal. Today 147, 1–59.
Miralles-Cuevas, S., Oller, I., Agüera, A., Sánchez Pérez, J.A., Sánchez-Moreno, R., Malato, S.,
2016. Is the combination of nanofiltration membranes and AOPs for removing
microcontaminants cost effective in real municipal wastewater effluents? Environ.
Sci.: Water Res. Technol. 2, 511–520.
Nogueira, R.F.P., Oliveira, M.C., Paterlini, W.C., 2005. Simple and fast spectrophotometric
determination of H2O2 in photo-Fenton reactions using metavanadate. Talanta 66,
86–91.
Persoone, G., Marsalek, B., Blinova, I., Torokne, A., Zarina, D., Manusadzianas, L., Nalecz-
Jawecki, G., Tofan, L., Stepanova, N., Tothova, L., Kolar, B., 2003. A practical and
user-friendly toxicity classification system with microbiotests for natural waters
and wastewaters. Environ. Toxicol. 18, 395–402.
Pignatello, J.J., Oliveros, E., McKay, A., 2006. Advanced oxidation processes for organic con-
taminant destruction based on the Fenton reaction and related chemistry. Crit. Rev.
Environ. Sci. Technol. 36, 1–84.
Sánchez Pérez, J.A., Román Sánchez, I.M., Carra, I., Cabrera Reina, A., Casas López, J.L.,
Malato, S., 2013. Economic evaluation of a combined photo-Fenton/MBR process
using pesticides as model pollutant. Factors affecting costs. J. Hazard. Mater.
244–245, 195–203.
Santos-Juanes, L., García, J.L., Casas, J.L., Oller, I., Malato, S., Sánchez Pérez, J.A., 2011. Dis-
solved oxygen concentration: a key parameter in monitoring the photo-Fenton pro-
cess. Appl. Catal. B Environ. 104, 316–323.
Spasiano, D., Marotta, R., Malato, S., Fernandez-Ibañez, P., Di Somma, I., 2015. Solar
photocatalysis: Materials, reactors, some commercial, and pre-industrialized applica-
tions. A comprehensive approach. Appl. Catal. B Environ. 170–171, 90–123.
Tânia, F.C.V., Silva, Amélia Fonseca, Saraiva, Isabel, Boaventura, Rui A.R., Vilar, Vítor J.P.,
2016. Scale-up and cost analysis of a photo-Fenton system for sanitary landfill leach-
ate treatment. Chem. Eng. J. 283, 76–88.