Environmental Science and Pollution Research

https://doi.org/10.1007/s11356-018-2431-2

ADVANCED OXIDATION TECHNOLOGIES: STATE-OF-THE-ART IN IBERO-AMERICAN

COUNTRIES

CFD modeling of a UV-A LED baffled flat-plate photoreactor

for environment applications: a mining wastewater case
John Steven Devia-Orjuela 1 & Luis Andrés Betancourt-Buitrago 1 & Fiderman Machuca-Martinez 1

Received: 31 January 2018 / Accepted: 28 May 2018

# Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract
The use of ultraviolet light in photoreactors for wastewater treatment has become popular as an alternative of known chemical
oxidative substances. UV LED light represents cheaper, robust, and versatile alternative to traditional UV lamps. In this study, it
was designed and evaluated a photoreactor with an approach of chemical fluid dynamics (CFD) and experimental validation. The
evaluation consisted of (1) CFD velocity profile analysis, (2) characterization of the average light distribution with potassium
ferrioxalate actinometry, (3) degradation of a typical recalcitrant metallic cyanocomplex Fe(CN)63−, and (4) scavenger effect
analysis in the photodegradation using potassium persulfate. Actinometrical essay concluded that the system was able to receive
1.93 μE/s. The reactor operated under turbulent regime and best result for Fe(CN)63− degradation was obtained at 4 h of
operation, using 5-W UV-A LEDs, with pH ~ 7 and 10 mM de S2O82−. Baffled photoreactor demonstrated to be useful for this
type of illumination and wastewater treatment.

Keywords Photoreactor . UV LED . Cyanocomplex . CFD . Actinometry . Photodegradation

Introduction several processes have been well described to remove

Gold extraction represents an important part of all economies in
the countries. Cyanidation is the hydrometallurgical technique
most used to dissolve the gold within the mineral from low-
grade ores (de Andrade Lima and Hodouin 2005). The process
has an efficiency of about 99% in the extraction and generates
xenobiotic and recalcitrant wastewater. However, cyanide not
only dissolves gold but other metals present in the ores such as
Fe, Ni, Cu, and Co, among others (Johnson 2015).
Degradation of cyanide-metal complexes byproducts from
gold mining wastewater requires more than just solar illumina-
tion to decompose themselves into less-hazardous compounds.
Also, pH alteration is not enough to hydrolyze Fe, Co, and Cu
cyanocomplexes to its free cyanide form (CN−) and HCN(g). The
iron and cobalt cyanocomplexes are the most difficult to decom-
pose (Adams 2013). For the gold mining wastewater treatment,
Responsible editor: Suresh Pillai
* Fiderman Machuca-Martinez
fiderman.machuca@correounivalle.edu.co
1
Escuela de Ingenieria Química, Universidad del Valle, Calle 13 No.
100-00, Cali, Colombia
cyanometallic compounds in wastewater. The most known pro-
cesses are alkali-chlorination, H2O2 oxidation, and biological
aerobic-anaerobic treatments (Kuyucak and Akcil 2013).
Photocatalysis (PC) is one of the advanced oxidation process
(AOPs) that uses artificial or natural light to promote charges
from a valence band to a conduction band in a semiconductor,
leading to redox reaction that degrades the pollutant (Nakata
and Fujishima 2012). The electron excitation is done using light
at a wavelength with enough energy to overcome the band gap
in the semiconductor. Photocatalytic processes are a good op-
tion where solar irradiation is abundant or otherwise, where
there is enough electric energy to use artificial illumination such
as Xe and Hg lamps. In general, the photocatalytic removal of
metals occurs by using the redox potential of the conduction
and valence band of the semiconductor against the metal in
solution, then oxidation or reduction can be done under certain
conditions of pH by direct or indirect reaction. The complex
CuCN–Cu(CN)3 have been degraded in a traditional batch
photoreactor using supported and suspended TiO2 (Bozzi et
al. 2004; Barakat 2004; Vilhunen and Sillanpää 2010). Iron
cyanocomplex Fe(CN)63− have been degraded using suspended
TiO2 and supported TiO2/SiO2, TiO2/GrSiO2 with degradations
of + 90% (van Grieken et al. 2005; López-Muñoz et al. 2005).

Light-emitting diodes (LED) are a less-consuming type of
lamps that can address different photoreactor configuration to
solve the problem of high electric energy consumption by arti-
ficial illumination (Izadifard et al. 2013). Also, this has been
used as an alternative to traditional Hg and Xe lamps for waste-
water treatment with satisfactory results (Yeh et al. 2014). This
type of light source has several advantages such as selective
wavelength application, more robust and long-lasting usability
of about 100 times more than traditional lamps (10.000 h of
life-time), low release of heat, and stable linearity for irradia-
tion-current, and it can be operated with pulsed illumination (Jo
and Tayade 2014). These characteristics make possible to de-
sign several types of photoreactors different to traditional CPC,
flat-plate or annular photoreactor, due to the manipulation of the
light source position, and then wide geometries can be applied.
Novel photoreactors using UV LED had been developed
for degradation of phenol (Jamali et al. 2013; Levchuk et al.
2015), for drinking water purification (Izadifard et al. 2013),
dye degradation such as methylene blue, red 23, and methyl
orange (Dai et al. 2014; Tokode et al. 2014a, b; Jo and Tayade
2014; Rasoulifard et al. 2015); for agriculture antibiotics deg-
radation (Malkhasian et al. 2014); and for pesticides degrada-
tion (Hossaini et al. 2014). Other applications had been on the
enhancement of the metal removal such as chromium (Liu et
al. 2014), and for bacteria disinfection like Escherichia coli
(Jenny et al. 2014; Marugán et al. 2015).
The chemical fluid dynamics (CFD) is a tool that can be
used for designing reactors and has a high flexibility and
breadth of applications. The advantages of CFD include the
ability to visualize several species, velocities profiles, vortices,
and other time-dependent variables at any point of the volume
(Bennett and Wiggins 2003). The CFD modeling had been
applied for tubular photoreactor with different light source
positions inside the reactive volume (Mohajerani et al.
2012), to study the photocatalytic degradation of acetone
(Vincent et al. 2011), for acetaldehyde degradation in
microphotoreactors (Verbruggen et al. 2015; van Walsem et
al. 2016), and water disinfection (Jenny et al. 2014).
In our previous work, we demonstrated UV LED are feasible
to be implemented in laboratory photoreactors for methylene blue
degradation (Betancourt Buitrago et al. 2017), but to the best of
our knowledge, no evaluation has been done for a baffled + flat-
plate + UV LED photoreactor to the degradation of Fe(CN)63− as
a typical mining wastewater pollutant. In this paper, we studied
the performance of a baffled flat-plate quartz photoreactor, for the
degradation of hexacyanoferrate as a gold mining cyanocomplex
in presence of oxygen and peroxydisulfate.
Material and methods
Hydrodynamic simulation was done using Comsol
Multiphysics 5.3a on a computer with Intel core i5 4590S
Environ Sci Pollut Res
CPU 3 GHz and 8 GB RAM. First, it was analyzed the veloc-
ity distribution into the photoreactor with and without baffles.
Then, it was used the Helmolthz light profile to estimate the
radiant field into the reactor and so the spacemen of the baffles
(Vaiano et al. 2015b). After that, a chemical actinometry using
potassium ferrioxalate was done to estimate the fluency of
photons to the fluid (Kuhn et al. 2004; Union et al. 2005).
Therefore, a typical evaluation of a photoreactor was done:
(1) catalyst load, (2) initial concentration, (3) addition of
S2O82− as electron scavenger, (4) apparent kinetics for the
photoreactor, and (5) comparison of the photocatalytic process
capacity of treatment for the iron-cyanide complex with other
studies.
Chemicals and equipment
It was used potassium hexacyanoferrate K3Fe(CN)6 (CAS
13746-66-2, Panreac AppliChem) as the major cyanocomplex
present in mining wastewater. As a catalyst, it was used
suspended TiO2 P25 Aeroxide CAS 13463-67-7. A UV-Vis
Spectrophotometer Vasco V-730 was used to follow the
Fe(CN)63− at 302.3 nm (Arellano and Martínez 2010). The
photoreactor was adapted from a quartz cuvette with
102.5 × 45 × 12.5 mm (typical 100-mm UV-vis path length
cuvette) and with all the sides transparent. Baffles were done
in acrylic (35 mm × 10 mm), and three 5-W UV LED were
used for one side of the cuvette powered in parallel at 4.5 V-
DC, and 1 A. No variation in the LED power was done in the
experiments.
Operation conditions
All solutions were prepared under normal temperature and
pressure with deionized water II. Air was bubbled to the
photoreactor at a rate of 1.5 L/min in an external bubbler
vessel. Total volume treated was 0.35 L with an illuminated
volume of 40 mL. A 50-W centrifugal pump was used to
pump the fluid in the photoreactor. No more than 10% of the
total volume were taken in five aliquots of 6 mL (30-mL
samples/350-mL treated volume) to measure degradation of
the complex.
Results and discussion
CFD: velocity profile
A mesh of 107,035 elements divided into 61,563 ele-
ments tetrahedral, 1712 elements pyramid, 43,760 ele-
ments prism, 10,598 elements triangular, 243 elements
quadrilateral, 824 elements edge, and 48 elements vertex
was used. Figure 1 describes the flat-plate-baffled geom-
etry and the UV LED size used.
Environ Sci Pollut Res

a b c

+ -
+ -
+ -

Fig. 1 Photoreactor configuration and 5-W UV-A LED source of light. a UV-diode. b View 3D reactor. c Meshing system

  
It was necessary to evaluate the Reynolds number from the μ ϵ ϵ2
following definition of hydraulic radius for the geometry of ρðu∙∇ Þϵ ¼ ∇ ∙ μ þ T ∇ ϵ þ C ϵ1 pk −C ϵ2 ρ ð5Þ
σϵ k k
our system (Bird 2012):
ϵ ¼ ep ð6Þ
cross section avaliable for flow volume available for flow
Rh ¼ ¼ k2
wet perimeter wet total surface μT ¼ ρC μ ð7Þ
ϵ
By definition, the Reynolds number is as follows: h  i 2
pk ¼ μT ∇ u : ∇ u þ ð∇ uÞT − ρk∇ ∙u ð8Þ
Deq vρ 4Rh vρ 3
Re ¼ ¼ ð1Þ
μ μ where:
where u fluid velocity
k turbulent kinetic energy
μ viscosity of the fluid
p pressure
v fluid velocity
ep reason for turbulent dissipation
p density of the fluid
σϵ, Cϵ1, Cϵ2, Cμ, σT parameters
The simulation solved the following equations: Navier μT turbulent viscosity
Stokes for steady state turbulent regime
h  i In Table 1, it is shown the boundary conditions.
ρðu∙∇ Þu¼∇ ∙ −pIþμ ∇ u þ ð∇ uÞT þ F ð2Þ A higher velocity also leads to a stronger collision in the
recombination zone and hence better mixing of the reactants
(Bawornruttanaboonya et al. 2017). To improve the hydrody-
namic turbulence conditions inside the photoreactor, the inlet
u fluid velocity
speed was varied resulting in a variation in the flow regime
F forces on the system
and Reynolds number. In addition, three baffles were added at
p pressure
different lengths and location in order to obtain the highest
With that, the continuity equation for incompressible flow:
Table 1 Inlet and outlet conditions
ρ∇ ∙ðuÞ¼0 ð3Þ
Inlet Outlet
For the turbulent state, the turbulence model k-ε is defined
u = − U0n ½−pIþðμ þ μT Þ ð∇ uþð∇ uÞ T Þ− 23
by the following expressions (Bird 2012): Uref=U0 2 ρkI n¼−p0* n
3=4
   2  k ¼ 32 U re f LT ; ϵ ¼ C μ 3=4 kLT p0∗ ≤ p0
LT = 0.07 × 0.0125 ∇k ∙ n = 0
ρðu∙∇ Þu ¼ ∇ ∙ −pI þ ðμ þ μT Þ ∇ u þ ð∇ uÞT − ρkI þ F ð4Þ
3 ∇ϵ ∙ n = 0
 Environ Sci Pollut Res

Fig. 2 Velocity profile for 30-mm baffled equidistant with a 0.025 m/s in the entrance and laminar profile, b 0.1 m/s in the entrance and laminar profile,
and c 0.2 m/s in the entrance and turbulent profile

residence time. Figure 2 shows the hydrodynamic behavior This increase in the average residence time indicates the
inside the reactor for the three different input speeds 0.025, photoreactor baffles could enhance the photocatalytic reaction
0.1, and 2 m/s. It can be seen that as the inlet speed increases, since the reactive volume is limited to the illuminated zone.
the dead zones are reduced, which favors mixing and de- However, a modification of the baffle length and position was
creases the mass transfer limitation, so the reactive fluid keeps required to increase the average velocity inside the cuvette. As
in constant motion. it can be seen in Fig. 3, the velocity profiles at 0.2 m/s without
A high flow rate leads to a decrease in residence time. For baffles (Fig. 3a) are enhanced to a better hydrodynamic be-
an input speed of 0.2 m/s without baffles, the average velocity havior for the same inlet velocity of the reactor, but with 35-
oscillates around 0.058 m/s, while with the presence of baffles mm baffle length and an asymmetrical arrange inside the cu-
the average velocity is around 0.091 m/s. This represents an vette (Fig. 3b). The 35-mm asymmetrical baffled photoreactor
increase of 57% in average speed. If we consider the volumet- resulted in 24% of increase in the average volumetric
ric residence time defined as: Reynolds number (Fig. 3c), from 1340 to 1662 against the
30-mm baffle’s length at the same asymmetrical position in-
V
τ¼ ρ ð9Þ side the reactor and a 4% increase of pressure drop (Fig. 3d)
ṁ instead of 30 mm asymmetrical. Although average volumetric
where Reynolds is lower than 2000, it is important to note that esti-
mation includes zones with lower local velocities, even
τ residence time
though this configuration reached a good mixing into the re-
V total reactive
actor resulting convenient for the photochemical degradation
volume
(Ismail et al. 2009). A better experimental mixing is expected
ṁ average mass flow
due to the bubbling in the recirculating dark zone. Since there
is no significant change in the pressure drop into the cuvette
Then, residence time is increased by 21% when passing to by the increase in the baffle length, 35 mm was the selected for
τ = 3.43 s without baffles to 4.16 s with baffles according to the photoreactor to obtain a maximum velocity of 0.32 m/s
Table 2. and a maximum Reynolds number greater than 5000.
Reynolds number characterizes the movement of a fluid
Table 2 Estimation of velocities and average residence times within the
reactor
and is a measure of the mixing inside the reactor. For some
microchannel photoreactor, it was found that the influence on
Configuration τ average residence time (s) Average velocity the number of Reynolds is not significant for mixing (Bennett
and Wiggins 2003); however, for catalytic microreactor, it has
0.2 without baffles 3.43 0.058
been reported that the conversion of the reactor studied in-
0.2 with baffles 4.16 0.091
creased to a high Reynolds due to collisions or recombination
Percentage increase 21% 57%
of the molecules (Bawornruttanaboonya et al. 2017). Other
Environ Sci Pollut Res

Fig. 3 Velocity profile for 0.2 m/s in the entrance and turbulent profile with asymmetrical baffles. a Without baffle and b with baffle 35 mm. c Reynolds
number inside the cuvette. d Pressure drop inside the cuvette (Pa)

CFD studies evaluate this variable and conclude that the
micromixing efficiency was greatly enhanced by increasing
the Reynolds number (Guo et al. 2017), which could be
achieved by high inlet liquid flow velocity which is consistent
with the mixing results obtained in this work.
CFD: radiant field
2003), hyperbolic one of Kubelka-Munk (Palmisano et al.
2013), and the Helmholtz equation (Vaiano et al. 2015a).
The Helmholtz equation is selected because, unlike the
other models, it has the ability to predict the radiant field in
the three spatial coordinates with a simpler approximation.
The Helmholtz equation is expressed in mathematical terms
as follows:

∇ ∙ð−c∇ I Þ þ aI ¼ F ð10Þ
For the description of the radiant field, different light models
have been reported in the literature such as bidirectional scat-
tering model (Motegh et al. 2013), two-flux model (Brucato F source term = 0
and Rizzuti 1997), six-flux model (Brucato et al. 2006), two- a absorption coefficient [cm−1]
flux model with simpler radiation attenuation terms (Li Puma b dispersion coefficient [cm−1]

Fig. 4 Light position of the LEDs. a Radiant field 2D. b Radiant field 3D with three cuts
 Environ Sci Pollut Res

Table 3 Actinometrical
comparison of different for Type of photoreactors/light source Actinometrical Wavelength Photon flux
photoreactors systems method (nm)

Batch 1 L/lamp UV de mercury 150 W (López-Muñoz Potassium > 300 13.7 μE s−1
et al. 2009) ferrioxalate
Batch 3 L/lamp UV mercury 8 W (D. C. Cid et al. 2012) Potassium 253 0.8 μE s−1
ferrioxalate
Microreactor 0.51 mL/UV LED (Aillet et al. 2014) Potassium 365 0.38 μE s−1
ferrioxalate
CSTR 100 mL/UV-A LED (Ferreira et al. 2016) Photo detector UV 375 0.76 μE s−1
Batch 150 mL/UV-A LED (Davididou et al. 2017) Potassium 365 3.32 μE s−1
ferrioxalate
This study Potassium 365 1.93 μE s−1
ferrioxalate

c diffusion coefficient [cm]
I light intensity [W/cm2]
1 UV LEDs (Canterino et al. 2008; Satuf et al. 2011; D. C. Cid
c¼ ½cm
3 ð a þ bÞ
where a = 97.5E−2 cm−1, and b = 1492E−2 cm−1 for a TiO2
load of 0.5 g/L (Otálvaro-Marín et al. 2014). Subject to the
following Dirichlet boundary condition in the geometric
spaces where the LEDs are located:
I ¼ r ¼ 5W=m2
As it can be seen in Fig. 4a, enough illumination is reached
for the LED positions inside the channel, with about 1.242 W/
m2 as an average irradiation. Figure 4b shows the position of
only three of them in each channel where same average illu-
mination was estimated for all the photoreactor.
2011; Aillet et al. 2014) and UV-A LEDs (Davididou et al.
2017). Although the configuration had a medium photon flux,
it was superior to other photocatalytic reactors reported using
ð11Þ
et al. 2012; Aillet et al. 2014; Karunakaran et al. 2015; Valari
et al. 2015; Ferreira et al. 2016). Thus, it is expected a good
harvesting of light in the photocatalytic activation of the TiO2
for this type of UV illumination.
Theoretical calculation of photon fluency to the
photoreactor is shown in Table 4. This approach considers
one wavelength of 400 nm, three UV LEDs at 5 W/m2 each,
and a lightning semiconductor diode area of 4 mm, as seen in
Fig. 1a. Thus, harvesting of light from LEDs in this arrange is
about 77% of the supplied by the UV LEDs.
AOP’s process evaluation

Effect of catalyst loading

Actinometry measurement
Chemical actinometry is a technique that measures the fluency
of photons that are absorbed by a fluid, in this case the reactive
illuminated volume of the photoreactor (Goldstein and Rabani
2008). For this photoreactor, the potassium ferrioxalate acti-
nometry estimated a fluency of photons of 1.93 μE/s. The
values are reported in the Table 3.
In spite of the small reaction volume, there was enough
light utilization from the UV LED, resulting in photon absorp-
tion close to other system using mercury lamps (Satuf et al.
The load of the catalyst represents the amount of the semicon-
ductor in the photoreactor. The range of operation has to be
low enough to save the catalyst from losing, and high enough
to reach good degradation until the light screening appears for
excess of TiO2. Figure 5 shows the behavior for the photocat-
alytic degradation from 0.0 to 1.0 g/L TiO2.
It can be seen an increasing in the slope until 0.5 g/L; after
this value, the decrease in the degradation evidences the
screening effect characteristic in heterogeneous photoreactors
with suspended catalyst. Then, selecting loads superior to

Table 4 Calculation for λ = 400 nm, E = 5 W/m2, n = 3 UV LEDs, 4 mm of lightning area per LED

Energy for Number of photons for 5 W Flux of photons Flow photons dosed to the Experimental Efficiency light
photon ep (J) (number of photons/m2/s) (μmol/s/m2) reactor (μmol/s) actinometry (μmol/s) capture (%)
N
Ep ¼ hc
λ N p ¼ EEp E QF ¼ N Ap I = nALEDQEF Iexp η ¼ I exp
I

4.969E−19 1.006E+19 1.671E−05 2.52 1.93 77%

Environ Sci Pollut Res

1.0
concentrations in time for 20, 50, and 100 ppm. It can be seen
that in 20 and 50 ppm, there is a similar kinetic behavior until
0.8 90 min; however, at a concentration of 100 ppm, the degrada-
tion is slower, decreasing the kinetics behavior after 20 min.
This may be related to the loss of photocatalytic activity
% Degradation

0.6
(poisoning) due to Fe precipitation on the TiO2 (Kuhn and
90 min Young 2005). Also, the change in the kinetics behavior repre-
30 min
0.4 sents a different phenomenon occurring inside the
photoreactor, such as a possible increase in the electron-hole
pair recombination due to excess of oxides, hydroxides, or
0.2
metallic Fe on TiO2 surface. In that way, a superficial area,
porosity, and active sites diminishing could be presented due
0.0 to irreversible chemisorption of Fe affecting the kinetic rate
0.0 0.2 0.4 0.6 0.8 1.0 order of the reaction when this concentration of iron
Catalyst load (g TiO2/L)
cyanocomplex was used. The highest percentage of degrada-
Fig. 5 Increase in the efficiency of the degradation for different catalyst tion was 62%, which was achieved for a concentration of
loads for 100 ppm K3Fe(CN)6 at 11 pH. Values for 0.0 g/L correspond to 20 ppm. It has been reported percentages of degradation about
the photolysis at 30 and 90 min
23, 50, and 42% in 350 min (Aguado et al. 2002) (Table 4).

0.5 g/L may represent waste of semiconductor in case it was
expensive and it was required a recovery after the process.
Although the load depends of the photoreactor geometry and
the capability of harvesting the light, a load of 0.5 g/L is
suggested as a common value and it was fixed for the follow-
ing experiments (Aguado et al. 2002; Osathaphan et al. 2008).
Nevertheless, it has been reported catalyst loads from 0.1 up to
5 g/L; however, loads superior to 2 g TiO2/L are not recom-
mended in suspended heterogeneous photocatalysis.
Initial concentration of K3Fe(CN)6 effect
Initial concentration of the cyanocomplex was evaluated to
reach the maximum capability of degradation in the
photoreactor. Figure 6 shows the variation of initial
pH effect in the degradation
The pH of the solution affects the zero charge point (ZPC) of
the TiO2 (Kosmulski 2011). In typical treatment of cyanide
wastewater, peroxide addition is useful to eliminate recalci-
trant cyanometallic complexes (Lee et al. 2004). In a photo-
catalytic system, the addition of peroxides is known to be an
electron scavenger, modifying the mechanism and enhancing
the formation of hydroxyl radicals (Song et al. 2012). Figure 7
shows the variation of pH and S2O82− addition as electron
acceptor.
For this cyanocomplex, it was evidenced a better behavior
when pH was close to 7, which is close to the TiO2 ZPC of 6.8
(McCullagh et al. 2011). Also, when peroxydisulfate as elec-
tron scavenger was added, a better photocatalytic degradation
was found evidencing the chemical activation of S2O82− and a

Fig. 6 Initial concentration effect 1.0

of K3Fe(CN)6 at 5 W/m2 and 20 ppm
0.9 50 ppm
pH = 11, catalyst load 0.5 g TiO2
/L 0.8
100 ppm
Photolysis
0.7

0.6
C/C0

Light
0.5
Off
0.4

0.3

0.2

0.1

0.0
-20 0 20 40 60 80 100 120 140 160
Time (min)
 Environ Sci Pollut Res

1.0 1.0
a) b)
100 ppm+S2O82-
0.8 100 ppm+S2O82-
0.8
100 ppm
100 pm

0.6 Light
Light Off
0.6 Off

C/C0
C/C0

0.4

0.4
0.2

0.2
0.0

0.0 -0.2
-20 0 20 40 60 80 100 120 140 160 180 200 220 240 -20 0 20 40 60 80 100 120 140 160 180 200 220 240
Time (min) Time (min)

Fig. 7 Effect of pH in the addition of peroxydisulfate as electron acceptor. One hundred parts per million of K3Fe(CN)6. a pH = 11 with 10 mM S2O82−.
b pH = 7 with 10 mM S2O82−

synergistic effect at pH 7, but it implies a risks of toxicity due where θ ¼ V TVþVR

R
is the diluted factor that means the propor-
to the release of free cyanide in the form of HCN(g) which has tion of the illuminated section with the rest of the reactor. For
a pKa of 9 (Kuyucak and Akcil 2013). On the other hand, at an the photocatalytic reactions, the law of speed follows the mod-
alkaline pH Fig. 7a, it can be seen that the degradation of the el of Langmuir-Hinshelwood;
cyanocomplex is inhibited which could be to the metal depo-
sition on the catalyst as previously mentioned. θkKC
r¼ ð14Þ
1 þ KC
where K is the adsorption constant and k is the kinetic con-
Kinetic analysis
stant. The apparent pseudo-first-order constant k′ can be ob-
tained when it is satisfied KC<<1. Then Eq. (13) can be writ-
For a recirculating batch photoreactor, the mass balance for
ten as:
the solute in the reactor and bubbling tank, it can be expressed
as follows: C 0
ln ¼ −θk t ð15Þ
d
 dC T CR
VR
∫0 rdV R ¼ ∫C R dV R þ V T ð12Þ
dt dt Table 5 shows the kinetic parameters (K, k, and k′) for the
where VR is the illuminated volume and VT is the total volume different concentration values.
(illuminated volume + volume of bubbling tank). If we as- Table 5 shows the values of the adsorption (K), kinetic (k),
sume that the illuminated volume does not have a spatial de- and pseudo constant (k′). It can be seen that the apparent con-
pendence and that the differences in the concentration of input stant increases as the concentration decreases, congruent to
and output are very small, the law of speed can be represented previous photocatalytic studies (Shrestha et al. 2012), making
as follows the photocatalytic process better for lower initial concentration.
In the other hand, the criteria or KC<<1 in Table 5 is not
dC T V T þ V R dC T 1 satisfied, then the pseudo-first-order kinetic in Eq. (14) should
r¼ ¼ ð13Þ
dt VR dt θ not be employed to describe the photocatalytic dynamic.

Table 5 Kinetic apparent pseudo-

first-order constant Initial Adsorption constant Kinetic constant Dilution factor θ Pseudo apparent
concentration (ppm) K (L mg−1) k (L−1.min−1) constant of reaction k′
(min−1)

100 1.27 0.32 0.1242 0.01690

50 1.20 0.42 0.1242 0.03784
20 0.89 0.19 0.1242 0.03945
Environ Sci Pollut Res
Table 6 Comparison of treatment capability of the UV LED photoreactor with other systems for K3Fe(CN)6 degradation
System evaluated Catalyst/source of light
Concentric parabolic reactor glass TiO2 0.6 [g/L]/sun
tube (Arellano and Martínez 2010) Photolysis/sun
3.2 g/L K3Fe(CN)6 degraded
Cylindrical batch reactor (Aguado et al. 2002) TiO2 degussa P25 0.5 [g/L] and
SBA-15/20%TiO2. 150-W UV-A
Medium Pressure lamp 150 W
210 mg/L Fe(CN)63−
This study TiO2 degussa P25 0.5 [g/L]/15 W/m2
Treatment capacity
Considering the material balance in our previous UV
LED photoreactor, with photocatalytic bubbling separat-
ed tanks and the dilution factor, the capability of treat-
ment can be compared to other systems (see Table 6).
Thus, the system is as good as other parabolic solar
photoreactors (Arellano and Martínez 2010), but still
lower than other and 600-W lamp photoreactors. The
other works obtained degradations below 50% at high
powers and cylindrical geometries with dilution factors
equal to 1. That may be related to the lack of illumina-
tion for using only three UV LEDs instead double illu-
mination for both sides of the quartz photoreactor,
reaching higher photon fluency, increasing the absorp-
tion efficiency of the photocatalyst and then the photo-
catalytic degradation rate.
Conclusions
This novel photoreactor resulted in a good alternative to
treat synthetic mining cyanide wastewater. It was dem-
onstrated UV LED are good alternative to degrade Fe
cyanocomplex using low-energy-consuming LED lamps.
The pH resulted in an important variable on the photo-
catalytic degradation. Values close to pH 7 (neutral)
favored the degradation of the Fe cyanocomplex; how-
ever, this point of operation is not suggested for iron
cyanometallic complex degradation due to a possible
gas evolution of hydrogen cyanide in the photocatalytic
treatment.
The CFD design tool characterized the hydrodynamics of
the reactor, which allowed to improve the mixing conditions
through the manipulation of the variables analyzed in this
paper (inlet velocity, flow regime and the addition of baffles),
and also the characterization of the radiant field allowed to
give the values of light intensity, needed to evaluate kinetics
of photocatalytic process.
Degradation (%)/time Complex degradation
[g Fe(CN)63− ∙ (gTiO2)−1 h−1]
96/24 h 0.058
75/24 h
23/350 min 0.046
50/350 min
42/350 min
97/4 h 0.050
Acknowledgments Acknowledgements are given to the Project
BRecuperación de oro y tratamiento de aguas residuales de minería
aurífera CI 2832^ funded by Colciencias and Universidad del Valle and
the Young Researcher Program 2015 of Colciencias Francisco Jose
Caldas.
References
Adams MD (2013) Impact of recycling cyanide and its reaction products
on upstream unit operations. Miner Eng 53:241–255. https://doi.org/
10.1016/j.mineng.2013.04.012
Aguado J, Van Grieken R, López-Muñoz M et al (2002) Removal of
cyanides in wastewater by supported TiO2-based photocatalysts.
Catal Today 75:95–102. https://doi.org/10.1016/S0920-5861(02)
00049-4
Aillet T, Loubiere K, Dechy-Cabaret O, Prat L (2014) Accurate measure-
ment of the photon flux received inside two continuous flow
microphotoreactors by actinometry. Int J Chem React Eng 12:1–
13. https://doi.org/10.1515/ijcre-2013-0121
Arellano CAP, Martínez SS (2010) Effects of pH on the degradation of
aqueous ferricyanide by photolysis and photocatalysis under solar
radiation. Sol Energy Mater Sol Cells 94:327–332. https://doi.org/
10.1016/j.solmat.2009.10.008
Barakat M (2004) Removal of toxic cyanide and cu(II) ions from water by
illuminated TiO2 catalyst. Appl Catal B Environ 53:13–20. https://
doi.org/10.1016/j.apcatb.2004.05.003
Bawornruttanaboonya K, Devahastin S, Mujumdar AS, Laosiripojana N
(2017) A computational fluid dynamic evaluation of a new
microreactor design for catalytic partial oxidation of methane. Int J
Heat Mass Transf 115:174–185. https://doi.org/10.1016/j.
ijheatmasstransfer.2017.08.027
Bennett J, Wiggins C (2003) A computational study of mixing
microchannel flows. arXiv Prepr cond-mat/0307482 3–7
Betancourt Buitrago LA, Ossa Echeverry O, Rodriguez Vallejo JC et al
(2017) An approach to the utilization of artificial high power LED
UV-A radiation in photoreactors for methylene blue degradation.
Photochem Photobiol 16:79–85. https://doi.org/10.1039/
C6PP00230G
Bird R (2012) Fenómenos de transporte. Saudi Med J 33:3–8. https://doi.
org/10.1073/pnas.0703993104
Bozzi A, Guasaquillo I, Kiwi J (2004) Accelerated removal of cyanides
from industrial effluents by supported TiO2 photo-catalysts. Appl
Catal B Environ 51:203–211. https://doi.org/10.1016/j.apcatb.2004.
02.014
Brucato A, Rizzuti L (1997) Simplified modeling of radiant fields in
heterogeneous photoreactors. 2. Limiting Btwo-flux^ model for the

case of reflectance greater than zero. Ind Eng Chem Res 36:4748–
4755. https://doi.org/10.1021/ie960260i
Brucato A, Cassano AE, Grisafi F, Montante G, Rizzuti L, Vella G (2006)
Estimating radiant fields in flat heterogeneous photoreactors by the
six-flux model. AICHE J 52:3882–3890. https://doi.org/10.1002/
aic.10984
Canterino M, Di Somma I, Marotta R, Andreozzi R (2008) Kinetic in-
vestigation of Cu(II) ions photoreduction in presence of titanium
dioxide and formic acid. Water Res 42:4498–4506. https://doi.org/
10.1016/j.watres.2008.07.035
Cid LDC, Grande DCM, Acosta EO, Ginzberg B (2012) Removal of
Cr(VI) and humic acid by heterogeneous photocatalysis in a labora-
tory reactor and a pilot reactor. Ind Eng Chem Res 51:9468–9474.
https://doi.org/10.1021/ie2010687
Dai K, Lu L, Liang C, Dai J, Zhu G, Liu Z, Liu Q, Zhang Y (2014)
Graphene oxide modified ZnO nanorods hybrid with high reusable
photocatalytic activity under UV-LED irradiation. Mater Chem Phys
143:1410–1416. https://doi.org/10.1016/j.matchemphys.2013.11.
055
Davididou K, Monteagudo JM, Chatzisymeon E, et al (2017)
Degradation and mineralization of antipyrine by UV-A LED
photo-Fenton reaction intensified by ferrioxalate with addition of
persulfate. https://doi.org/10.1016/j.seppur.2016.08.021
de Andrade Lima LRP, Hodouin D (2005) Optimization of reactor vol-
umes for gold cyanidation. Miner Eng 18:671–679. https://doi.org/
10.1016/j.mineng.2004.12.007
Ferreira LC, Lucas MS, Fernandes JR, Tavares PB (2016) Photocatalytic
oxidation of Reactive Black 5 with UV-A LEDs. J Environ Chem
Eng 4:109–114. https://doi.org/10.1016/j.jece.2015.10.042
Goldstein S, Rabani J (2008) The ferrioxalate and iodide-iodate actinom-
eters in the UV region. J Photochem Photobiol A Chem 193:50–55.
https://doi.org/10.1016/j.jphotochem.2007.06.006
Guo T, Ruan B, Liu Z, Jamal MA, Wen L, Chen J (2017) Numerical and
experimental investigations of liquid mixing in two-stage micro-im-
pinging stream reactors. Chin J Chem Eng 25:391–400. https://doi.
org/10.1016/j.cjche.2016.11.015
Hossaini H, Moussavi G, Farrokhi M (2014) The investigation of the
LED-activated FeFNS-TiO2 nanocatalyst for photocatalytic degra-
dation and mineralization of organophosphate pesticides in water.
Water Res 59:130–144. https://doi.org/10.1016/j.watres.2014.04.
009
Ismail I, Abdel-Monem N, Fateen SE, Abdelazeem W (2009) Treatment
of a synthetic solution of galvanization effluent via the conversion of
sodium cyanide into an insoluble safe complex. J Hazard Mater 166:
978–983. https://doi.org/10.1016/j.jhazmat.2008.12.005
Izadifard M, Achari G, Langford C (2013) Application of photocatalysts
and LED light sources in drinking water treatment. Catalysts 3:726–
743. https://doi.org/10.3390/catal3030726
Jamali A, Vanraes R, Hanselaer P, Van Gerven T (2013) A batch LED
reactor for the photocatalytic degradation of phenol. Chem Eng
Process Process Intensif 71:43–50. https://doi.org/10.1016/j.cep.
2013.03.010
Jenny RM, Simmons OD, Shatalov M, Ducoste JJ (2014) Modeling a
continuous flow ultraviolet light emitting diode reactor using com-
putational fluid dynamics. Chem Eng Sci 116:524–535. https://doi.
org/10.1016/j.ces.2014.05.020
Jo W, Tayade RJ (2014) New generation energy-efficient light source for
photocatalysis: LEDs for environmental applications. Ind Eng Chem
Res 53:2073–2084. https://doi.org/10.1021/ie404176g
Johnson CA (2015) The fate of cyanide in leach wastes at gold mines: an
environmental perspective. Appl Geochem 57:194–205. https://doi.
org/10.1016/j.apgeochem.2014.05.023
Karunakaran C, Vasumathi D, Vinayagamoorthy P (2015) Enhanced pho-
tocatalytic Fe 3 + reduction with H 2 O 2 generation by TiO 2
anatase / rutile blend. Indian J Chem 54:1076–1084
Environ Sci Pollut Res
Kosmulski M (2011) The pH-dependent surface charging and points of
zero charge. J Colloid Interface Sci 353:1–15. https://doi.org/10.
1016/j.jcis.2010.08.023
Kuhn DD, Young TC (2005) Photolytic degradation of hexacyanoferrate
(II) in aqueous media: the determination of the degradation kinetics.
Chemosphere 60:1222–1230. https://doi.org/10.1016/j.
chemosphere.2005.02.011
Kuhn HJ, Braslavsky SE, Schmidt R (2004) International Union of Pure
and Applied Chemistry—chemical Actinometry. IUPAC Tech Rep
1–47
Kuyucak N, Akcil A (2013) Cyanide and removal options from effluents
in gold mining and metallurgical processes. Miner Eng 50–51:13–
29. https://doi.org/10.1016/j.mineng.2013.05.027
Lee T-Y, Kwon Y-S, Kim D-S (2004) Oxidative treatment of cyanide in
wastewater using hydrogen peroxide and homogeneous catalyst. J
Environ Sci Health A 39:787–801. https://doi.org/10.1081/ESE-
120027742
Levchuk I, Rueda-Márquez JJ, Suihkonen S, Manzano MA, Sillanpää M
(2015) Application of UVA-LED based photocatalysis for plywood
mill wastewater treatment. Sep Purif Technol 143:1–5. https://doi.
org/10.1016/j.seppur.2015.01.007
Li Puma G (2003) Modeling of thin-film slurry photocatalytic reactors
affected by radiation scattering. Environ Sci Technol 37:5783–5791.
https://doi.org/10.1021/es0300362
Liu X, Pan L, Lv T, Sun Z (2014) CdS sensitized TiO2 film for photo-
catalytic reduction of Cr(VI) by microwave-assisted chemical bath
deposition method. J Alloys Compd 583:390–395. https://doi.org/
10.1016/j.jallcom.2013.08.193
López-Muñoz M-J, Van Grieken R, Aguado J, Marugán J (2005) Role of
the support on the activity of silica-supported TiO2 photocatalysts:
structure of the TiO2/SBA-15 photocatalysts. Catal Today 101:307–
314. https://doi.org/10.1016/j.cattod.2005.03.017
López-Muñoz MJ, Aguado J, van Grieken R, Marugán J (2009)
Simultaneous photocatalytic reduction of silver and oxidation of
cyanide from dicyanoargentate solutions. Appl Catal B Environ
86:53–62. https://doi.org/10.1016/j.apcatb.2008.07.022
Malkhasian AYS, Izadifard M, Achari G, Langford CH (2014)
Photocatalytic degradation of agricultural antibiotics using a UV-
LED light source. J Environ Sci Health B 49:35–40. https://doi.
org/10.1080/03601234.2013.836871
Marugán J, van Grieken R, Pablos C, Satuf ML, Cassano AE, Alfano OM
(2015) Kinetic modelling of Escherichia coli inactivation in a pho-
tocatalytic wall reactor. Catal Today 240:9–15. https://doi.org/10.
1016/j.cattod.2014.03.005
McCullagh C, Skillen N, Adams M, Robertson PKJ (2011)
Photocatalytic reactors for environmental remediation: a review. J
Chem Technol Biotechnol 86:1002–1017. https://doi.org/10.1002/
jctb.2650
Mohajerani M, Mehrvar M, Ein-mozaffari F (2012) Computational fluid
dynamics (CFD) modeling of photochemical reactors. Appl Comput
Fluid Dyn 155–176
Motegh M, Ruud van Ommen J, Appel PW, Mudde RF, Kreutzer MT
(2013) Bubbles scatter light, yet that does not hurt the performance
of bubbly slurry photocatalytic reactors. Chem Eng Sci 100:506–
514. https://doi.org/10.1016/j.ces.2013.02.022
Nakata K, Fujishima A (2012) TiO2 photocatalysis: design and applica-
tions. J Photochem Photobiol C: Photochem Rev 13:169–189.
https://doi.org/10.1016/j.jphotochemrev.2012.06.001
Osathaphan K, Chucherdwatanasak B, Rachdawong P, Sharma VK (2008)
Photocatalytic oxidation of cyanide in aqueous titanium dioxide sus-
pensions: effect of ethylenediaminetetraacetate. Sol Energy 82:1031–
1036. https://doi.org/10.1016/j.solener.2008.04.007
Otálvaro-Marín HL, Mueses MA, Machuca-Martínez F (2014) Boundary
layer of photon absorption applied to heterogeneous photocatalytic
solar flat plate reactor design. Int J Photoenergy 2014:1–8. https://
doi.org/10.1155/2014/930439
Environ Sci Pollut Res
Palmisano G, Loddo V, Augugliaro V (2013) Two-dimensional modeling
of an externally irradiated slurry photoreactor. Int J Chem React Eng
11:675–685. https://doi.org/10.1515/ijcre-2012-0049
Rasoulifard MH, Fazli M, Eskandarian MR (2015) Performance of the
light-emitting-diodes in a continuous photoreactor for degradation
of Direct Red 23 using UV-LED/S2O82—process. J Ind Eng Chem
24:121–126. https://doi.org/10.1016/j.jiec.2014.09.018
Satuf ML, Pierrestegui MJ, Rossini L, Brandi RJ, Alfano OM (2011)
Kinetic modeling of azo dyes photocatalytic degradation in aqueous
TiO 2 suspensions. Toxicity and biodegradability evaluation. Catal
Today 161:121–126. https://doi.org/10.1016/j.cattod.2010.11.018
Shrestha TB, Seo GM, Basel MT, Kalita M, Wang H, Villanueva D, Pyle
M, Balivada S, Rachakatla RS, Shinogle H, Thapa PS, Moore D,
Troyer DL, Bossmann SH (2012) Stem cell-based photodynamic
therapy. Photochem Photobiol Sci 11:1251–1258. https://doi.org/
10.1039/c2pp05417e
Song S, Hong F, He Z, Cai Q, Chen J (2012) AgIO3-modified AgI/TiO2
composites for photocatalytic degradation of p-chlorophenol under
visible light irradiation. J Colloid Interface Sci 378:159–166
Tokode O, Prabhu R, Lawton LA, Robertson PKJ (2014a) Mathematical
modelling of quantum yield enhancements of methyl orange photo-
oxidation in aqueous TiO2 suspensions under controlled periodic
UV LED illumination. Appl Catal B Environ 156–157:398–403.
https://doi.org/10.1016/j.apcatb.2014.03.046
Tokode O, Prabhu R, Lawton LA, Robertson PKJ (2014b) The effect of
pH on the photonic efficiency of the destruction of methyl orange
under controlled periodic illumination with UV-LED sources. Chem
Eng J 246:337–342. https://doi.org/10.1016/j.cej.2014.03.002
Union I, Pure OF, Chemistry A (2005) International Union of Pure and
Applied Chemistry Chemical Actinometry. 1–47
Vaiano V, Sacco O, Pisano D, Sannino D, Ciambelli P (2015a) From the
design to the development of a continuous fixed bed photoreactor
for photocatalytic degradation of organic pollutants in wastewater.
Chem Eng Sci 137:152–160. https://doi.org/10.1016/j.ces.2015.06.
023
Vaiano V, Sacco O, Pisano D et al (2015b) From the design to the devel-
opment of a continuous fi xed bed photoreactor for photocatalytic
degradation of organic pollutants in wastewater. Chem Eng Sci 137:
152–160
Valari M, Antoniadis A, Mantzavinos D, Poulios I (2015) Photocatalytic
reduction of Cr ( VI ) over titania suspensions. Catal Today 252:
190–194. https://doi.org/10.1016/j.cattod.2014.10.014
van Grieken R, Aguado J, López-Muñoz M-J, Marugán J (2005)
Photocatalytic degradation of iron–cyanocomplexes by TiO2 based
catalysts. Appl Catal B Environ 55:201–211. https://doi.org/10.
1016/j.apcatb.2004.08.008
van Walsem J, Verbruggen SW, Modde B, Lenaerts S, Denys S (2016)
CFD investigation of a multi-tube photocatalytic reactor in non-
steady- state conditions. Chem Eng J 304:808–816. https://doi.org/
10.1016/j.cej.2016.07.028
Verbruggen SW, Lenaerts S, Denys S (2015) Analytic versus CFD ap-
proach for kinetic modeling of gas phase photocatalysis. Chem Eng
J 262:1–8
Vilhunen S, Sillanpää M (2010) Recent developments in photochemical
and chemical AOPs in water treatment: a mini-review. Rev Environ
Sci Biotechnol 9:323–330. https://doi.org/10.1007/s11157-010-
9216-5
Vincent G, Schaer E, Marquaire PM, Zahraa O (2011) CFD modelling of
an annular reactor, application to the photocatalytic degradation of
acetone. Process Saf Environ Prot 89:35–40. https://doi.org/10.
1016/j.psep.2010.08.004
Yeh N, Yeh P, Shih N, Byadgi O, Chih Cheng T (2014) Applications of
light-emitting diodes in researches conducted in aquatic environ-
ment. Renew Sust Energ Rev 32:611–618. https://doi.org/10.1016/
j.rser.2014.01.047