Introduction

Nowadays, water pollution remains one of the major problems worldwide. Discharges of industrial, domestic, and agricultural effluents, among others, contribute to increase the contamination level of surface and groundwater. In many cases, pollutants have harmful effects on different forms of life and ecosystems, causing an environmental and health risk.

Water pollution can be reduced by conventional effluent treatments that involve physical, chemical, and biological methods. However, the removal of bio-recalcitrant or highly toxic pollutants (pesticides, herbicides, heavy metals, chlorinated hydrocarbons, dyes, disinfectants, and pharmaceutical contaminants, among others) by these methods is often highly inefficient or expensive since they preclude usual biological treatment (Oller et al. 2011).

In the last years, advanced oxidation processes (AOPs) have proved to be an efficient alternative to remove bio-recalcitrant and toxic pollutants in water (Comninellis et al. 2008). These processes are based on the generation of hydroxyl radicals (HO·) which rapidly attack organic matter in a nonselective form due to its highly oxidizing power (Ameta et al. 2012; Pignatello et al. 2006).

Among the different AOPs is the well-known Fenton reaction based in the HO· generation from the reaction between H2O2 and Fe(II)/Fe(III) ions in aqueous solution (Neyens and Baeyens 2003). In addition, irradiation with UV and/or visible light contributes to increase the rate of HO· generation by photoreduction of hydroxylated Fe(III) ions (Fe(OH)2+), in the so-called photo-Fenton process (Pignatello et al. 2006). The low cost of solar light as a suitable radiation source for this process has significantly increased the interest in this process as an efficient and cheap method for wastewater treatment (Ammar et al. 2016; Miralles-Cuevas et al. 2017; Ruiz-Aguirre et al. 2017).

In photo-Fenton processes, iron in solution acts as a homogeneous catalyst and the iron concentrations needed to achieve high-performance levels usually exceed the allowed limits for effluent discharges into watercourses. Therefore, iron must be removed by precipitation at the end of the process, leading to the generation of sludge and its disposal. This affects the cost of the process and constitutes a limitation for its large-scale implementation. Furthermore, the pH of the medium affects the photocatalytic performance of the homogeneous process which is efficient in a narrow pH range close to 3 (Huston and Pignatello 1999). This appears as an inconvenience, particularly for natural effluents where pH adjustment becomes a requirement.

To overcome these drawbacks, iron immobilized on solid materials is seen as an interesting alternative which facilitates the separation of the catalyst by sedimentation and allows its recovery and reuse. Moreover, immobilized iron species could not be easily transformed into less photoactive materials, thus allowing the possibility of using a wider pH range than that associated with the homogeneous system. In these heterogeneous photo-Fenton processes, iron species have been supported on different solid materials such as zeolites (González-Olmos et al. 2012), silica (Jusoh et al. 2015), SBA-15 (Zhong et al. 2011), and activated carbon fiber (Lan et al. 2015). The use of clays as support have received special attention due to the high catalytic activity that these materials display after iron incorporation by pillaring processes to produce iron-pillared clays (Fe-PILCs) (De León et al. 2008, 2013, 2017; Bel Hadjltaief et al. 2015; Herney-Ramirez et al. 2010; Martin del Campo et al. 2014). Fe-PILCs are obtained by ion exchange of the interlayer cations of clays by large iron polycations which are then thermally decomposed leading to the formation of pillars between clay platelets and thereby to an increase in interlayer spacing, specific surface area, and pore volume. These materials display significant catalytic activity at initial pH values ranging from 3 to 7 (De León et al. 2008; Najjar et al. 2007). Additionally, catalytic stability tests of these materials also show that their activity remains stable in several consecutive reuses (Najjar et al. 2007; Sum et al. 2005; Xu et al. 2014).

Similar to other photo-activated processes, the photo-Fenton involves photon absorption, in this case by iron species playing a catalytic role, and the consequent formation of activated species that participate in the reaction mechanisms. Because of that, the efficiency of these processes is the result of the intrinsic activity of the catalyst and its efficiency in the use of photons. Chemical and morphological properties of the catalyst, the reaction nature and the experimental conditions such as reactor type and design, range of wavelength and radiation intensity, catalyst loading, reagent initial concentration, and pH are among the variables which determine the efficiency (Satuf et al. 2007). The parameters used to evaluate the process efficiency in the use of radiation are quantum (or absolute) efficiency and photonic (or apparent) efficiency. These allow knowledge about the extent to which the radiation leads to the chemical transformation of the present species and to compare the efficiency of different photo-activated systems. In the case of a heterogeneous photo-Fenton process, the only known antecedent is the recent determination of the photonic efficiency by using iron-modified MCM-41 as catalyst (Benzaquén et al. 2017), but, as far as we know, there are no studies for iron-modified clays. Considering the catalytic properties of Fe-PILCs in photo-Fenton processes and the fact that these materials are cheap and reusable, the study of the efficiency in the use of light could contribute to their application in wastewater treatment.

The present work presents the evaluation of photonic and quantum efficiencies of photo-Fenton processes with regional raw clays pillared with iron (Fe-PILCS) as catalysts, applied to the degradation of 2-chlorophenol chosen as the model pollutant. The influence of the iron content of the catalyst and the catalyst loading is also studied.

Materials and methods

Catalyst preparation

A natural clay extracted from Bañado de Medina, Uruguay, was used as starting raw material. About 80% of the clay is a calcium-rich montmorillonite with low sodium and potassium content (Diano et al. 1994). The clay was dried, ground, and sieved, selecting the fraction of aggregate size less than 250 μm for catalyst preparation.

The catalysts (Fe-PILCs) were prepared as previously reported by De León et al. (2015). The clay was exchanged with an aqueous solution of [Fe3(OCOCH3)7OH·2H2O]NO3 as follows: a solution of the iron complex was poured on an aqueous suspension (10% in weight) of the selected aggregate fraction of the clay until a final ratio of 0.5 mmol of complex per gram of clay was attained. This procedure was repeated using 3.5 mmol of complex per gram of clay. The exchanged clays were recovered by filtration, washed several times with deionized water, dried at 60 °C, and calcined at 400 °C in air atmosphere. The temperature was increased at 1 °C min−1 from room temperature up to 400 °C which was maintained for 2 h. The Fe-PILCs thus obtained were ground with a mortar and sieved. The particles that passed through a sieve of 200 mesh (nominal sieve opening 74 μm) were selected and denoted as C-0.5 and C-3.5, according to the complex/clay ratio used in the preparation.

Catalyst properties

The properties of the starting clay and the catalysts—iron content, specific surface area, and specific micropore volume—were determined in a previous study (De León et al. 2015). These properties are summarized in Table 1.

Table 1 Properties of the clay and the catalysts (De León et al. 2015)
Full size table

Reactor

Catalytic tests were performed in a cylindrical, well-stirred, batch reactor, irradiated from the bottom by a tubular lamp Philips 18 W TL’D/08 (wavelength range 350–400 nm) placed in the focal axis of a cylindrical parabolic reflector (Ortiz de la Plata et al. 2010a). The bottom of the reactor (reactor window) consists of an acrylic plate transparent to UVA radiation and a borosilicate ground glass plate which ensures the entry of diffuse radiation and azimuthal symmetry to the radiation beam. The reactor window area (Aw) is 227 cm2, the internal diameter of the reactor is 17 cm, the reaction volume (VR) is 2000 cm3, and the operating liquid height is 9 cm. The agitation in the reactor is carried out with a blade stirrer with an impeller made of acrylic, which produces an upwardly directed axial flow ensuring the homogenization of the suspension. In addition, the reactor is equipped with four internal transparent glass heat exchangers connected to a thermostatic bath to keep the temperature constant during the reaction. The incident radiation flux in the reactor window (qw) is 5.65 × 10−9 Einstein cm−2 s−1 and was calculated by actinometry measurements using potassium ferrioxalate (Murov et al. 1993).

Catalytic tests

An aqueous solution containing 2-chorophenol (2-CP) and hydrogen peroxide (H2O2) was used in all tests. Initial concentration of 2-CP (C2-CP,0) and H2O2 were 50 mg L−1 and 265 mg L−1, respectively. These values are in the concentration ranges used in previous degradation studies of 2-CP and other chlorophenols by the photo-Fenton process (Karci et al. 2012; Ortiz de la Plata et al. 2010b; Soltani and Lee 2017). The solution pH was adjusted with perchloric acid. Different catalyst loads (Lcat) were employed, and the temperature was maintained at 25 °C. Two thousand cubic centimeter of 2-CP and H2O2 solution was placed in the reactor and thermostated with constant stirring. Then, the catalyst was added and the irradiation was started (reaction time equal to zero). The UVA lamp was previously stabilized for 15 min blocking the passage of light to the reactor. These tests were named as photo-Fenton. Other tests, named as Fenton, were performed in the same conditions but without irradiating the reaction medium.

In addition, complementary photo-Fenton tests at pH 3.0, 5.0, and 7.0 were performed with C-3.5 using a catalyst load of 1.0 g L−1. The initial concentration of 2-CP and H2O2, as well as the temperature, was the same as those used in the tests described above. For these tests, an auxiliary reactor whose schematic representation can be found in De León et al. (2017) was used. The reactor consists in a borosilicate glass tube illuminated by four Philips TLD 18 W/08 lamps place around it. The lamps emit UVA radiation with a maximum emission at 360 nm. The reaction medium (0.5 L), placed in a reservoir, is recirculated through the borosilicate tube using a peristaltic pump. The reservoir is a jacketed vessel whose temperature is kept constant by recirculating water from a thermostatic bath. Additionally, the reservoir is magnetically agitated to keep the catalyst in suspension.

Moreover, in order to characterize the adsorption of 2-CP on the catalysts, dark adsorption tests were carried out at different pH values (3.0, 5.0, and 7.0), using an initial 2-CP concentration of 50 mg L−1.

Analytical methods

Samples of the reaction medium were collected and the catalyst was separated by filtration through a 0.2-μm pore-size cellulose acetate membrane. 2-CP was determined with a HPLC Waters, equipped with a LC-18 Supelcosil reversed-phase column (Supelco) and a dual UV detector. The eluent was a mixture of an aqueous solution of acetic acid (1% v/v)/methanol/acetonitrile 60:30:10 in volume with a flow rate of 1.0 mL min−1. The UV detector wavelength was set at 280 nm. Total organic carbon (TOC) was determined with a Shimadzu TOC-5000 Analyzer. H2O2 was quantified spectrophotometrically at 350 nm using an iodometric technique (Allen et al. 1952). Total iron in solution was quantified colorimetrically at 510 nm with the 1,10-phenantroline technique; the samples were previously treated with ascorbic acid. A Cary100 Bio spectrophotometer was used for absorbance measures.

Efficiency parameters

Photonic efficiency

Photonic (or apparent) efficiency (ηp) can be defined as the ratio between the number of reactant molecules degraded during a given time interval and the total number of incident photons on the irradiated window of the reactor during the same period (Benzaquén et al. 2012; Braslavsky et al. 2011):

$$ {\upeta}_p=\frac{\ \mathrm{number}\ \mathrm{of}\ \mathrm{reactant}\ \mathrm{molecules}\ \mathrm{degraded}}{\mathrm{number}\ \mathrm{of}\ \mathrm{incident}\ \mathrm{photons}} $$
(1)

According to Eq. (1), ηp for 2-CP degradation in the catalytic system under study can be expressed in terms of the experimental data as

$$ {\upeta}_p\kern0.5em =\kern0.5em \frac{\left({C}_{2-\mathrm{CP},\kern0.5em 0\kern0.5em }-{C}_{2-\mathrm{CP},f}\right)\ {V}_R}{q_w\ {A}_w\ \left({t}_f-{t}_0\right)} $$
(2)

where C2-CP,0 and C2-CP,f are the initial and final concentrations of 2-CP, respectively; VR is the reaction volume; qw is the incident radiation flux at the reactor window average over the window area; Aw is the reactor window area and (tf − t0) is the time interval elapsed during the catalytic test, being t0 and tf the initial and final time, respectively.

The extent to which all the organic compounds (phenol and reaction intermediates) attain their mineralization, i.e., its conversion to water, mineral acids, and carbon dioxide, is also important. The photonic efficiency of the mineralization process (ηTOC,p) can be defined as the ratio between the amount of total organic carbon converted during a given time interval and the total number of incident photons on the irradiated window of the reactor during the same period (Benzaquén et al. 2012; Satuf et al. 2007):

$$ {\upeta}_{\mathrm{TOC},p}=\frac{\mathrm{amount}\ \mathrm{of}\ \mathrm{TOC}\ \mathrm{converted}}{\mathrm{number}\ \mathrm{of}\ \mathrm{incident}\ \mathrm{photons}} $$
(3)

According to Eq. (3) and the experimental data, ηTOC,p for 2-CP mineralization process can be determined as

$$ {\upeta}_{\mathrm{TOC},p}\kern0.5em =\kern0.5em \frac{\left({\mathrm{TOC}}_0-{\mathrm{TOC}}_f\right)\ {V}_R}{q_w\ {A}_w\ \left({t}_f-{t}_0\right)} $$
(4)

where TOC0 and TOCf are the initial and final total organic carbon, respectively.

Quantum efficiency

Quantum (or absolute) efficiency (ηq) can be defined as the ratio between the number of reactant molecules degraded during a given time interval and the number of photons absorbed by the species to be activated, over the employed spectral range, during the same period (Benzaquén et al. 2012; Braslavsky et al. 2011):

$$ {\upeta}_q=\frac{\mathrm{number}\ \mathrm{of}\ \mathrm{reactant}\ \mathrm{molecules}\ \mathrm{degraded}}{\mathrm{number}\ \mathrm{of}\ \mathrm{photons}\ \mathrm{absorbed}\ \mathrm{by}\ \mathrm{the}\ \mathrm{species}\ \mathrm{to}\ \mathrm{be}\ \mathrm{activated}} $$
(5)

This parameter takes into account that in heterogenous system photons might be scattered by the catalyst particles in suspension; therefore, not the whole incident radiation is absorbed by the catalyst to promote the reaction.

Besides, the quantum efficiency of the mineralization process (ηTOC,q) can be defined as the ratio between the amount of total organic carbon degraded during a given time interval and the number of photons absorbed during the same period (Benzaquén et al. 2012; Satuf et al. 2007):

$$ {\upeta}_{\mathrm{TOC},q}=\frac{\mathrm{amount}\ \mathrm{of}\ \mathrm{TOC}\ \mathrm{removed}}{\mathrm{number}\ \mathrm{of}\ \mathrm{photons}\ \mathrm{absorbed}\ \mathrm{by}\ \mathrm{the}\ \mathrm{species}\ \mathrm{to}\ \mathrm{be}\ \mathrm{activated}} $$
(6)

In the catalytic system under study, ηq for 2-CP degradation and ηTOC,q for the mineralization process can be calculated as follows:

$$ {\upeta}_q=\kern0.5em \frac{\left({C}_{2-\mathrm{CP},\kern0.5em 0\kern0.5em }-{C}_{2-\mathrm{CP},\kern0.5em f}\right)\ }{{\left\langle {e}^a(x)\right\rangle}_{V_R}\left({t}_f-{t}_0\right)} $$
(7)
$$ {\upeta}_{\mathrm{TOC},q}=\kern0.5em \frac{\left({\mathrm{TOC}}_{\kern0.5em 0\kern0.5em }-{\mathrm{TOC}}_f\right)\ }{{\left\langle {e}^a(x)\right\rangle}_{V_R}\left({t}_f-{t}_0\right)} $$
(8)

where the expressions (C2-CP,0 − C2-CP,f)/(tf − t0) in Eq. (7) and (TOC0 − TOCf)/(tf − t0) in Eq. (8) represent 2-CP degradation and 2-CP mineralization rates, respectively, and \( {\left\langle {e}^a(x)\right\rangle}_{V_R} \) is the local volumetric rate of photon absorption (LVRPA) averaged over the reactor volume, VR. Thus, quantum efficiency determination requires knowledge of the rate of photon absorption in the reactant system.

Radiation model in the reactor and LVRPA determination

The local volumetric rate of photon absorption was calculated by solving the radiative transfer equation (RTE) in the reactor. For this purpose, the one-dimensional, one-directional radiation transport model was applied (Fig. 1):

$$ \mu\ \frac{\partial {I}_{\lambda}\left(x,\kern0.5em \mu \right)}{\partial x}+{\upbeta}_{\lambda }{I}_{\lambda}\left(x,\mu \right)=\kern0.5em \frac{\upsigma_{\lambda }}{2}\ \underset{\mu^{,}=-1}{\overset{1}{\int }}{I}_{\lambda}\left(x,{\mu}^{\prime \kern0.5em }\right)\ p\left(\mu, {\mu}^{\prime}\right)d{\mu}^{\prime \kern0.5em } $$
(9)

where Iλ is the spectral radiation intensity, λ represents the radiation wavelength, x is the axial coordinate, βλ is the spectral extinction coefficient, σλ is the spectral scattering coefficient, μ is the direction cosine of the ray with respect to the propagating direction for which the RTE is written (μ = cos θ), μ′ is the cosine of an arbitrary ray before scattering, and p represents the phase function for scattering. βλ is defined as the sum of the absorption and the scattering coefficients (βλ = σλ + κλ).

Fig. 1
figure 1

Coordinate system for the one-dimensional, one-directional radiation model

Full size image

To model the scattering effects of catalyst particles, the Henyey and Greenstein (HG) phase function was adopted (Satuf et al. 2005):

$$ {p}_{\mathrm{HG},\lambda}\left({\mu}_0\right)=\frac{\left(1-{g}_{\lambda}^2\right)}{{\left(1+{g}_{\lambda}^2-2{g}_{\lambda }{\mu}_0\right)}^{3/2}} $$
(10)

where gλ is the dimensionless asymmetry factor and μ0 is the cosine of the angle between the incident and the scattered rays.

The boundary conditions were assumed as follows:

$$ {\displaystyle \begin{array}{l}x=0\ \left(\mu >0\right),I\left(x,\mu \right)={I}^0\\ {}x={L}_R\ \left(\mu \kern0.50em 0\right),I\left(x,\mu \right)=0\end{array}} $$

To solve the RTE, it is necessary to know the optical properties of the catalyst suspension: the specific absorption coefficient (\( {\upkappa}_{\uplambda}^{\ast } \)), the specific scattering coefficient (\( {\upsigma}_{\uplambda}^{\ast } \)), and the asymmetry factor (gλ). These properties for aqueous suspensions of C-0.5 and C-3.5 were determined in a previous study (De León et al. 2015). Table 2 summarized only the values in the wavelength range of interest for the present study.

Table 2 Optical properties of the catalysts: specific absorption and scattering coefficients, and the asymmetry factor of the Henyey-Greenstein phase function (De León et al. 2015)
Full size table

The RTE was solved applying the discrete ordinate method (DOM) (Duderstadt and Martin 1979), which transforms the integro-differential expression of the RTE into a system of finite-difference algebraic equations that can be solved numerically. This method was applied discretizing the wavelength λ, the spatial coordinate in the x dimension, and the angle in the θ direction, represented by μ = cos θ (Ortiz de la Plata et al. 2010a). The solution of the RTE provides the values of Iλ(x, μ) at each point and each direction in the heterogeneous reacting medium.

Once the radiation intensities were obtained, the LVRPA was calculated as

$$ \mathrm{LVRPA}={e}^a(x)=2\uppi\ \underset{\uplambda}{\int }{\upkappa}_{\uplambda}\ {\int}_{\mu -1}^1{I}_{\uplambda}\left(x,\mu \right) d\mu\ d\uplambda $$
(11)

and finally, the LVRPA average over the reactor volume (VRPA) was calculated according to

$$ {\left\langle {e}^a(x)\right\rangle}_{V_R}=\frac{1}{V_R}\ \underset{V_R}{\int }{e}^a(x)\ dV $$
(12)

Result and discussion

Catalytic tests

Table 3 shows the results of the photo-Fenton tests performed with the catalyst C-3.5 at different pH values. As can be seen, the catalyst showed to be active in the catalytic process at each pH values. However, at pH 3, a significant improvement in its catalytic performance is observed. Considering these results, the catalytic evaluation of C-0.5 and C-3.5 using different catalyst loads, as well as the photon and quantum efficiency determinations, was studied at pH 3.

Table 3 Catalytic performance in photo-Fenton tests at different pH values for C-3.5 and a catalyst load of 1.0 g L−1
Full size table

In addition, dark adsorption effects in the photo-Fenton process was studied to confirm the catalytic nature of the 2-CP removal. The results of dark adsorption tests performed at different pH values with 1 g L−1 of C-3.5 indicated that, at the end of the tests (120 min), 2-CP adsorption on the catalyst surface was very low: 4.1%, 2.4%, and 1.5% at pH 3.0, 5.0, and 7.0, respectively. If these values are compared with the 2-CP conversions reached in the photo-Fenton tests at each pH (Table 3), it can be concluded that the contribution of the adsorption to 2-CP removal in the photo-Fenton tests can be considered negligible.

Figure 2 shows the results of the catalytic test under photo-Fenton conditions using 1.0 g L−1 of C-0.5. A slow decrease of 2-CP (10%) and H2O2 (5%) concentrations is observed during the first 90 min. After this induction period, 2-CP degradation is accelerated and attains 24% at the end of the test (120 min). The observed induction period has been reported in the literature for iron-modified clays used as catalyst in Fenton processes and has been attributed to the time required for the activation of surface iron species and the adsorption of reactants onto the catalyst surface (Carriazo et al. 2003; Luo et al. 2009; Silva et al. 2012; Timofeeva et al. 2005). TOC relative concentration is higher than that of 2-CP during the test and only 14% of the initial TOC is removed at the end. These results reveal the formation of organic reaction intermediates whose conversion into final inorganic products occurs at a lower reaction rate. Although their identification is not an objective of this study, it is worth mentioning that several aromatic compounds such as chlorobenzoquinone, chlorohydroquinone, benzoquinone, and hydroquinone, as well as carboxylic acids such as maleic, fumaric, and oxalic, have been identified as reaction intermediates in the 2-CP photo-Fenton oxidation (Bel Hadjltaief et al. 2018; Ortiz de la Plata et al. 2010b). Iron concentration in the reaction medium remained at very low values and reaches 0.6 mg L−1 at the end of the test. This value corresponds to the leaching of only 1.0% of the iron contained in the fresh catalyst used in the assay.

Fig. 2
figure 2

2-CP degradation by photo-Fenton process with C-0.5 and Lcat = 1.0 g L−1. Dimensionless concentration of 2-CP, H2O2, and TOC, and total iron concentration in solution vs time

Full size image

The results of the catalytic test for the same catalyst (C-0.5) and a catalyst load (0.2 g L−1) are shown in Fig. 3. The same induction period is observed although the decrease of 2-CP concentration was lower than 2% during the first 90 min. Then, 2-CP degradation rate increased but lower values of 2-CP and TOC removal (11% and 3%, respectively) were reached at the end of the test. Iron in solution reaches a final value of 0.2 mg L−1, which corresponds to 1.6% of the iron present in the fresh catalyst.

Fig. 3
figure 3

2-CP degradation by photo-Fenton process with C-0.5 and Lcat = 0.2 g L−1. Dimensionless concentration of 2-CP, H2O2, and TOC, and total iron concentration in solution vs time

Full size image

Figures 4 and 5 show the results of the photo-Fenton test for C-3.5. For a catalyst load of 1.0 g L−1 (Fig. 4), total 2-CP degradation is reached at 120 min of reaction, thus showing the increased reaction rate with respect to C-0.5 as a consequence of the higher iron content in C-3.5. Nevertheless, TOC removal was only 16%, indicating a scarce influence of the iron content in the catalyst on the degradation rate of intermediates. When a lower catalyst load (0.2 g L−1) was used (Fig. 5), the degradation rate decreases and the removal of 2-CP and TOC achieved at the end of the assay was only 53% and 4%, respectively. The amount of iron in solution at the end of this test is 0.6 mg L−1. Thus, for C-3.5, the maximum loss of iron by leaching during the tests does not exceed 1.7% of the iron present in the fresh catalyst.

Fig. 4
figure 4

2-CP degradation by photo-Fenton process with C-3.5 and Lcat = 1.0 g L−1. Dimensionless concentration of 2-CP, H2O2, and TOC, and total iron concentration in solution vs time

Full size image
Fig. 5
figure 5

2-CP degradation by photo-Fenton process with C-3.5 and Lcat = 0.2 g L−1. Dimensionless concentration of 2-CP, H2O2, and TOC, and total iron concentration in solution vs time

Full size image

Table 4 summarizes the main results of the photo-Fenton and Fenton catalytic tests. From these results, it can be inferred that 2-CP degradation is significantly increased along with the catalyst load and its iron content. These results agree with those expected, since both an increase in the iron content of the catalyst and in the catalyst load will produce an increase in the amount of iron available for the reaction. According to the reaction mechanisms generally accepted, the higher amount of iron sites would increase the rate of HO· radical production and the subsequent reactions with 2-chlorophenol. The TOC removal rate also showed a significant increase along with the catalyst load and to a lesser extent also with respect to the iron content. Furthermore, in the Fenton catalytic tests, performed in the absence of light, the conversions of 2-CP and TOC were much lower than those achieved when the reaction medium was irradiated (photo-Fenton catalytic test). These results show the important role of light in the reaction mechanism involved in the degradation and mineralization of 2-CP, as well as the scarce contribution of the Fenton reaction.

Table 4 C-0.5 and C-3.5 catalytic performance in photo-Fenton and Fenton tests
Full size table

Photonic efficiency evaluation

Table 5 presents the values of the photonic efficiency for 2-CP degradation calculated from Eq. (2) at 120 min of reaction, using results of the catalytic experiments. For both catalysts, a 5-fold increase in the catalyst load (from 0.2 to 1.0 g L−1) almost doubles the photonic efficiency ηp. In heterogeneous catalytic systems, the increase of the catalyst load causes the intensification of the radiation scattering and the decrease of light penetration into the reaction medium. Therefore, it is reasonably expected that the increase in ηp will not be proportional to that of the catalyst load.

Table 5 Photonic efficiency of 2-CP degradation under photo-Fenton conditions for different loads of C-0.5 and C-3.5
Full size table

If the photonic efficiencies for C-0.5 and C-3.5 are compared when using the same catalyst load, either 0.2 or 1.0 g L−1 (Table 5), it is observed that ηp is significantly higher for C-3.5 than for C-0.5. This is explained by the higher reaction rate observed with C-3.5 as a consequence of its higher iron content.

Photonic efficiency of 2-CP mineralization process (ηTOC,p) at 120 min of reaction, calculated from Eq. (4) and TOC experimental data obtained in the catalytic evaluation assays, is presented in Table 6. For C-0.5, the increase of the catalyst load from 0.2 to 1.0 g L−1 causes a significant increase of ηTOC,p (4 times). A similar trend is observed when increasing the catalyst load of C-3.5. On the other hand, for equal loads of C-0.5 and C-3.5, ηTOC,p is slightly higher for C-3.5, the catalyst with higher iron content. Thus, it can be inferred that there is some influence of the catalyst iron content on the photonic efficiency of the mineralization process.

Table 6 Photonic efficiency of 2-CP mineralization under photo-Fenton conditions for different loads of C-0.5 and C-3.5
Full size table

Volumetric rate of photon absorption

Figure 6 shows the LVRPA in the reactor as a function of the spatial coordinate x for both catalysts, C-0.5 and C-3.5, and each catalyst load, 0.2 g L−1 and 1.0 g L−1. In all cases, the LVRPA decreases when the distance from the irradiated window located at the reactor bottom is increased. For a catalyst load of 0.2 g L−1, the LVRPA profiles along the x-coordinate are fairly uniform. However, an increase in the catalyst load from 0.2 to 1.0 g L−1 results in a highly non-uniform photon absorption rate close to the irradiated reactor bottom; then, an important decrease of the LVRPA occurs for the catalyst load of 1.0 g L−1 as the depth of the radiation propagation increases. In particular, it is observed how the photon absorption rate becomes insignificant for radiation propagation depths greater than 4 cm. This observation is interesting from a practical point of view, since operating the reactor with catalyst loads of 1 g L−1 and liquid heights higher than 4 cm would not be useful to increase the photon absorption and improve the photo-Fenton process.

Fig. 6
figure 6

LVRPA profiles in the reactor for different loads of C-0.5 and C-3.5

Full size image

From the spatial distribution of LVRPA in the reactor for each catalyst load, the averaged value of LVRPA over the reactor volume (VRPA or \( {\left\langle {e}^a(x)\right\rangle}_{V_R} \)) can be calculated from Eq. (12). Figure 7 shows the results of \( {\left\langle {e}^a(x)\right\rangle}_{V_R} \) for different catalyst loads from 0.2 to 1.0 g L−1 and for C-0.5 and C-3.5. As can be observed for C-0.5, \( {\left\langle {e}^a(x)\right\rangle}_{V_R} \) increases with the catalyst load up to about 0.6 g L−1 and it remains practically constant for higher loads. The same tendency is observed for C-3.5, with \( {\left\langle {e}^a(x)\right\rangle}_{V_R} \) values lower than C-0.5 up to 0.6 g L−1 and constant values reached at about 1.0 g L−1. The maximum radiation absorption is already achieved at approximately 0.6 g L−1 for C-0.5 and 1.0 g L−1 for C-3.5. These values can be ascribed to the maximum load of catalyst in which all the catalyst particles—i.e., all the surface exposed—are totally illuminated. For higher catalyst loads, a screening effect of excess particles occurs, which masks part of the photosensitive surface (Herrmann 2005). Satuf et al. (2007) found that higher amounts of catalyst than those required to achieve maximum radiation absorption do not significantly increase the degradation and mineralization efficiencies. Thus, for practical applications, this optimum catalyst loading should be chosen in order to avoid a useless excess of catalyst and to ensure the total absorption of photons. In the case of the catalytic systems here studied, those optimal catalyst loading values are approximately 0.6 g L−1 for C-0.5 and 1.0 g L−1 for C-3.5.

Fig. 7
figure 7

Local volumetric rate of photon absorption averaged over the reactor volume for different catalyst loads for C-0.5 and C-3.5

Full size image

Quantum efficiency evaluation

From the averaged value of LVRPA over the reactor volume and the 2-CP concentrations as a function of time, the quantum efficiency of 2-CP degradation was calculated from Eq. (7) at 120 min of reaction. The results are presented in Table 7. For both catalysts, the increase of the catalyst load from 0.2 to 1.0 g L−1 increases the ηq process with factors of 2.1 for C-0.5 and 1.5 for C-3.5. For each catalyst load, ηq is significantly higher for C-3.5 than for C-0.5. For Lcat = 1.0 g L−1, the values of ηq for C-3.5 are approximately four times those of C-0.5 and for Lcat = 0.2 g L−1, 5.4 times, while the iron content of C-3.5 does not reach to three times that of C-0.5. This behavior is similar to that observed for the photonic efficiencies (Table 5) and can be explained by similar causes.

Table 7 Quantum efficiency of 2-CP degradation under photo-Fenton conditions for different loads of C-0.5 and C-3.5
Full size table

As expected, for each of the studied cases, the quantum efficiency of 2-CP degradation (Table 7) is greater than the photonic efficiency (Table 5). Because the photons that reach the reactor window are not fully absorbed, the denominator of Eq. (7) is smaller than the denominator of Eq. (2). Therefore, since the respective numerators are equal, the values of ηq are always greater than the corresponding values of ηq. In fact, for each catalyst and catalyst load, the values of ηq practically double those of ηp.

Table 8 shows the ηTOC,q values calculated from Eq. (8) at 120 min of reaction. For both catalysts, the increase of the catalyst load from 0.2 to 1.0 g L−1 causes the quantum mineralization efficiency increase by a factor close to 4 for C-0.5 and 2.9 for C-3.5. For the same catalyst load of each of the catalysts, the values of ηTOC,q are similar, showing a similar behavior to that observed for ηTOC,p values.

Table 8 Quantum efficiency of 2-CP mineralization under photo-Fenton conditions for different loads of C-0.5 and C-3.5
Full size table

As it was verified for degradation efficiencies, for the same reason explained above, the quantum efficiency of mineralization (Table 8) practically doubles the photonic efficiency of mineralization (Table 6) for both catalysts and catalyst loads.

Finally, to get an idea of the relative activity of the Fe-PILCs, it is interesting to compare their catalytic performance with other commonly used materials. Thus, Table 9 includes results obtained with C-3.5 using a catalyst load of 0.2 g L−1 and zero valent iron nanoparticles (ZVI). The values for ZVI were calculated from the information reported by Ortiz de la Plata et al. (2012) for 2-CP degradation, in the same reactor with similar experimental conditions. It can be observed that, even with a lower iron amount, the catalyst C-3.5 achieves a higher degradation level of 2-CP, as well as higher photonic and quantum efficiencies of degradation than ZVI. However, both materials have similar performances in the 2-CP mineralization process.

Table 9 Comparison between the catalysts C-3.5 and nZVI in the degradation of 2-CP in photo-Fenton conditions
Full size table

Conclusions

The iron-pillared clays here studied, C-0.5 and C-3.5, exhibit catalytic activity in the photo-Fenton process applied to 2-CP degradation.

The 2-CP degradation rate is significantly increased along with the catalyst load and the catalyst iron content. The rate of TOC removal also showed a significant increase with the catalyst load and, to a lesser extent, with respect to the iron content. The best catalytic performance was obtained using 1.0 g L−1 of C-3.5, the catalyst with the higher iron content (17.6%).

The photonic and quantum efficiencies of 2-CP degradation depend on both the catalyst load and the iron content of the catalyst. For both catalysts, C-0.5 and C-3.5, the increase of the catalyst load from 0.2 to 1.0 g L−1 significantly increases ηp and ηq. In addition, for equal loads of C-0.5 and C-3.5, ηp and ηq are significantly higher for C-3.5, the catalyst with the higher iron content.

For the mineralization process, photonic and quantum efficiencies depend mainly on the catalyst load, since the increase in the iron content of the catalyst does not have a significant effect on ηTOC,p and ηTOC,q.

These results on photonic and quantum efficiencies demonstrate that it was possible to take advantage of a natural and cheap resource, a regional raw clay, to obtain pillared clay-based catalysts capable of degrading organic pollutants in water.