Introduction

The agricultural and industrial sectors as well as household release each day a huge amount of pollutants into surface water, making them nonusable (Telegang 2017). Among them, the most well known are pharmaceutical products, phenolic compounds, and dyes (Bouyarmane 2014). Synthetic dyes are widely used in numerous industries such as textile, printing, food, cosmetic, clinical, paper, leather, pharmaceutical, and food industries (Martínez-Huitle and Brillas 2009). In fact, world production of dyes is estimated at 800,000 T/year and about 140,000 T/year is released in the environment during the fabrication and coloring steps of textiles (Mansour et al. 2011).

Dye effluent may contain chemicals that are toxic, carcinogenic, or mutagenic, to various fish species (Verma et al. 2012). It may also prevent light penetration in water and photosynthesis, which can drive to a lack of dissolved oxygen and upset the biological metabolism processes (Joshi et al. 2004; Assémian et al. 2018).

However, they can be removed by different processes: electrocoagulation (Assémian et al. 2018), Fenton (Briton et al. 2018), coagulation/flocculation (Bouyakoub et al. 2010), adsorption using activated carbon, membrane filtration (Han et al. 2016; Sattar et al. 2017), and combined coagulation/carbon adsorption (Papić et al. 2004). Though these processes give good results, they have some drawbacks. Indeed, adsorption is selective and is just a transfer of pollutants from the aqueous phase to the solid one, and at bottom leads to another contamination. Fenton requires a lot of reagents, and electrocoagulation generates sludge at the end of the process. One way to destroy pollutants without using reagent and generating secondary toxic materials is photocatalysis (Asiri et al. 2011). Photocatalytic degradation, as a cheap and environmentally eco-friendly technology, has drawn extensive attention (Spasiano et al. 2015).

The type of photocatalyst used plays a key role in photocatalytic efficiency. TiO2 has been investigated for the removal of hazardous compounds in water. Owing to its nontoxicity, low-cost, and good stability property compared to other materials (Ansari et al. 2016), TiO2 is currently used in photocatalytic process. The effectiveness of this process on dyes removal has been widely demonstrated in many studies (Sacco et al. 2018; Yang and Yang 2018).

Besides, effectiveness of photocatalytic degradation is strongly related to the operating conditions such as TiO2 concentration (powder), light intensity, pH, concentration of pollutants, and the presence of certain ions. Sometimes, the effectiveness of one parameter can be related to another one. Many publications have dealt with the degradation of dye using TiO2 films (Wu and Zhang 2004; Crişan et al. 2018), TiO2 nanocomposite (Hariprasad et al. 2013; Viet et al. 2018; Sanzone et al. 2018), and in suspension (Hafizah and Sopyan 2015). Among all of them, the best results were obtained with TiO2 in suspension because it provides a better contact between TiO2 and the pollutants. However, the use of TiO2 powder causes the clogging of lamps and requires an expensive filtering process after the treatment. Hence, studies have been oriented on the immobilization of TiO2 nanoparticles but the photocatalytic efficiencies are less than in suspension mode due to the reduction in the contacts between photocatalyst and pollutants. For this reason, β-SiC alveolar foam is a good support for the immobilization of the catalyst. Its three-dimensional structure improves the contact between the organic pollutant and the supported catalyst.

TiO2/β-SiC foam material has proven its effectiveness for the removal of organic pollutants (Marien 2017; M’Bra et al. 2019). Parameters which are commonly studied in photocatalytic process are pH, dye concentration, mass of catalyst, and effect of electron acceptors such as H2O2, KBrO3, K2SO2, etc. (Aliabadi and Sagharigar 2011; Seck et al. 2012; Hamidi and Kacem 2017). To our knowledge, few studies have evaluated the effect of one parameter on another and the interaction between parameters on the removal process. Furthermore, the classical method which consists of varying one factor at a time and keeping others constant is not sufficient to explain the phenomenon entirely, and it is time-consuming. In this case, full factorial design is carried out to achieve the best optimization of the process (Elhalil et al. 2016). This design determines the effect of each factor on the response as well as the manner in which each factor varies with the change in the level of the other factors (Arenas et al. 2007). Interaction effects of different factors could be attained using experimental design (Montgomery 1997; Brasil et al. 2005). This technique is used to reduce the number of experiments, time, overall process cost. The objective of this study was to assess the interaction effects of each parameter and their effects on the photocatalytic performance using TiO2/β-SiC foam.

Materials and methods

Materials

RhB dye was procured from Fluka, Germany (purity ≥ 99%), and its characteristics are mentioned in Table 1. Hydrogen peroxide (purity ≥ 95%), nitric acid (purity ≥ 99.5%), and sodium hydroxide (purity ≥ 95%) were provided by Panreac (Spain). All solutions were prepared with distilled water. Cylindrical-shape foam (35 mm diameter; 50 mm length) was provided by SICAT (German). The β-SiC foams were calcinated at 1000 °C for 2 h in order to remove the residual organic carbon. The deposition of TiO2-P25 onto β-SiC foams was made by dip coating. Each foam was completely immersed in a suspension (10 g of TiO2-P25 and 4 mL of TTIP in 200 mL of dry ethanol) for 3 min at 5 rpm. This process was repeated five times. Then, the photocatalytic materials were dried at room temperature for 20 min, avoiding clogging of the alveoli. Subsequently, TiO2/β-SiC foams were placed in an oven at 110 °C overnight to evaporate residual organic compounds. After that, TiO2/β-SiC foams were brought to the furnace at 450 °C for 2 h at a rise rate of 5 °C min−1. The average weight of TiO2 fixed on each foam is 1.5 g.

Table 1 Physicochemical properties of Rhodamine B
Full size table

Photo-reactor and UV source

The photoreactor (Fig. 1) consisted of one coaxial quartz tube (400 mm length, ID 40 mm) placed inside a Pyrex glass cylinder. The quartz tube was filled with the TiO2/β-SiC foam. TiO2 (1.5 g) was coated on each foam giving a total mass of 10 g (foam + TiO2). Four UV-A lamp tubes (Philips F8T5/BL) of 8 Watts with a maximum emission at approximately 365 nm were placed around the quartz tube and within the Pyrex glass cylinder. The total light intensity in the reactor was 14 W/m2.

Fig. 1
figure 1

Schematic representation of experimental setup

Full size image

Experimental procedure

The working solutions were obtained by dilution of a stock solution of 100 ppm prepared by dissolving required mass of RhB dye in distilled water. Degradation tests were performed by circulating the RhB solution within 120 min through the photoreactor containing four TiO2/β-SiC foams, using a peristaltic pump (MasterFlex, model 7520-47) at variable flow rate between 4 and 12 mL s−1. pH was adjusted at a desired value before each test. One milliliter of the solution was extracted at regular time intervals to follow the kinetics of degradation. The residual concentration was determined with a JASCO UV–Vis spectrophotometer (λ = 554 nm).

Experimental design

Two levels were chosen for each parameter, a higher level denoted (+) and a lower level denoted (−). In this study, a full factorial design (FFD) was used (24). It resulted in a combination of all levels of each factor (Assidjo et al. 2005). Thus, 16 experiments were carried out and the chosen parameters were pH (X1), concentration of RhB (X2), volume of H2O2 (X3), and the number of TiO2/β-SiC foam (X4) (Table 2). The experimental domain is shown in Table 2. The response obtained was RhB removal rate (Y).

Table 2 Experimental domain for full factorial design
Full size table

The RhB removal rate was calculated to according Eq. 1:

$$ y_{i} = \frac{{C_{\text{o}} - C_{\text{r}} }}{{C_{\text{o}} }} \times 100 $$
(1)

C0 is the initial RhB concentration and Cr is final RhB concentration.

In a full factorial design, the response studied is considered to be linear for all factors and can be represented by a linear polynomial function where b0 represents the average response. bi is the estimation of the principal effect of the factor i for the response Y and bij the interaction effect between factors i and j for the response Y investigated of RhB removal.

$$ Y \, = \, b_{0} + \, b_{1} X_{1} + \, b_{2} X_{2} + \, b_{3} X_{3} + \, b_{4} X_{4} + \, b_{12} X_{1} X_{2} + \, b_{13} X_{1} X_{3} + \, b_{14} X_{1} X_{4} + \, b_{23} X_{2} X_{3} + \, b_{24} X_{2} X_{4} + \, b_{34} X_{3} X_{4} + \, b_{123} X_{1} X_{2} X_{3} + \, b_{124} X_{1} X_{2} X_{4} + \, b_{134} X_{1} X_{3} X_{4} + \, b_{234} X_{2} X_{3} X_{4} + \, b_{1234} X_{1} X_{2} X_{3} X_{4} $$
(2)

Results and discussion

Comparison of photolysis and photocatalysis photocatalytic

Figure 2 shows both photolysis and photocatalysis results. The photolysis discoloration under UV radiation (365 nm) was weak, with only 20% of removal after 120 min. These results are different from the 90% removal obtained by Hamidi and Kacem (2017) using direct photolysis. This difference may be explained by the fact that they used in their study UVC (254 nm) which produces higher energy. This high photon energy (hν) facilitates the H2O photolysis and the production of hydroxyl radicals according to Eq. 3:

Fig. 2
figure 2

Photocatalytic tests with TiO2/β-SiC foam (CRhB = 10 ppm, flow rate = 4 mL s−1, T = 25 °C)

Full size image
$$ {\text{H}}_{2} {\text{O}} + h\nu \to {\cdot}{\text{OH}} + {\text{H}}{\cdot} $$
(3)

However, in the photocatalytic reaction, RhB was almost fully removed. The discoloration was about 90%. Applying the Langmuir–Hinshelwood equation with solutions highly diluted, the reaction follows an apparent first-order kinetic. ln(C0/Cr) versus time was linear line (Fig. 3) with R2 of 0.979 and an apparent constant rate Kap of 194.10−4 min−1.

Fig. 3
figure 3

Linearization from Langmuir–Hinshelwood model

Full size image

Modeling of the photocatalytic discoloration of RhB

Photocatalytic discoloration efficiency (Y) was measured for each factor, and the results are in Table 3. The predicted value (Ycalc) and the measured values were very close, which indicated a good correlation between the model and experimental results. The coefficients of the polynomial model were calculated using the NemrodW Software (Ano et al. 2019) as follows:

Table 3 Experimental design matrix and response based on the experimental runs and predicted values on discoloration of RhB
Full size table
$$ Y = 74.17 + 1.9X_{1} {-}6.66X_{2} {-}1.98X_{3} + 8.78X_{4} + 0.31X_{1} X_{2} {-}0.34X_{1} X_{3} + 3.92X_{2} X_{3} {-}0.29X_{2} X_{4} + 0.252X_{3} X_{4} $$
(4)

Main and interaction effects were determined through experimental results (Table 4).

Table 4 Value of model coefficient
Full size table

Using the Pareto diagram enables the calculation of the effect (%) of each factor on the response, according to Eq. 5 (García-Gómez et al. 2014):

$$ p = \frac{{b_{i}^{2} }}{{\sum b_{i}^{2} }} \times 100 \left( {i \ne 0} \right) $$
(5)

Figure 4 shows that foam number was the most important factor (53.01%), next RhB initial concentration (30.49%). The most relevant interaction obtained was between initial concentration of RhB and H2O2 volume (10.56%).

Fig. 4
figure 4

Graphical Pareto analysis of effect of pH, concentration of RhB, volume of H2O2 and number of foam samples on the discoloration efficiency

Full size image

Main interactions

The coefficient b0 = 74.171 represents the average value of the response of the 16 assays. Table 4 shows that RhB discoloration was highly positively influenced by foam quantity (b4 = 8.78). RhB removal percentage increases on average 17.56 (2 × 8.78) when foam number was increased from 2 to 4. These results are similar to those of Al-Kahtani (2016) and Antoniou and Dionysiou (2007). They showed that the increase in TiO2 quantity resulted in an increase in active sites on which OH– was oxidized to produce hydroxyl radicals which contribute to RhB removal. The second most important factor according to the Pareto diagram was pollutant’s concentration which has a negative effect on the removal (b2 = − 6.67). The increase in RhB concentration induced a decrease in removal rate on average 13.34% (2 × 6.67), when concentration was increased from 10 to 30 ppm. This result can be explained by a decrease in the transmittance of the solution with the increase in the concentration of RhB, leading to fewer photons reaching the photocatalyst surface, capable of activating it and generating ·OH and O·−2 radicals. Furthermore, a large number of adsorbed RhB molecules would inhibit the reaction between RhB molecules and ·OH radicals as a result of a lower chance of any direct interaction between them (Ammari et al. 2015; Goyal and Kishore 2017).

Among the interaction between two factors, only X2X3 showed a high coefficient (b23 = 3.92) and also a positive effect on the discoloration rate. The interaction is shown in Fig. 5, and the corner at the top represents the combination of the lower and the higher levels of both factors RhB concentration (X2) and peroxide volume (X3).

Fig. 5
figure 5

Interaction between initial concentration of RhB (X2) and hydrogen peroxide volume (X3)

Full size image

When hydrogen peroxide volume was set at the lowest value (1 mL), RhB concentration produced (induced) a negative influence on the removal which rose from 86.73 to 65.57%. On the one hand, when the dose of peroxide was set at the highest level (3 mL), the concentration varied from 79.93 to 69.45%. These results showed that in order to obtain good discoloration, it is necessary to work with a low concentration of dye at a low dose of H2O2. Similar trends were observed by Hamidi and Kacem (2017). They showed that the effectiveness for RhB decolorization is influenced by the ratio [H2O2]/[RhB]. In this study, the optimum ratio was 6.05 when the H2O2 concentration was higher. Two reactions were observed in the production of HO·2 (E0 = 1.7 V) which is less oxidizing than ·OH (E0 = 2.8 V) and not favorable for good removal of organic compounds Eq. (6), and a consumption of hydroxyl radicals (Qourzal et al. 2007; Briton et al. 2019) according to the following reactions in Eq. (7):

$$ {\text{H}}_{2} {\text{O}}_{2} + {\cdot}{\text{OH}} \to {\text{H}}_{2} {\text{O}} + {\text{HO}}_{2}^{{\mathbf{ \cdot }}} $$
(6)
$$ {\text{HO}}_{2}^{{\mathbf{ \cdot }}} + {\cdot}{\text{OH}} \to {\text{H}}_{2} {\text{O}} + {\text{O}}_{2} $$
(7)

Validity of the model

The model is validated if the value of the correlation coefficient R2 is close to 1 (Fu et al. 2007). R2 obtained for the model was 0.991, indicating a good agreement between the predicted values and the measured values. The determined F value which was 55.4654 and the low probability value (Pr > F = 0.000173) indicated that the full factorial design used in this study was validated by the results (Table 5).

Table 5 ANOVA results for full factorial design
Full size table

Study aging of foam

For an application, it is important to evaluate how long the foam can be used without any regeneration; for that purpose, ten different aqueous solutions of RhB (5 L) were consecutively treated with the same foams for 20 h (Fig. 6), without any washing and no regeneration. At the end of each test, the reactor was emptied and new solution was put for the following treatment. The apparent kinetic value constants were very close (Table 6), which means, firstly, that there is no loss of TiO2 particles by washing out and, secondly, that there is no poisoning of the photocatalyst surface by adsorption of the RhB intermediate by-products. At the end of the ten tests, a yield decrease of only 6.7% was observed. TiO2/β-SiC foam was weighed before and after the 30 tests; only slight loss of weight of 3% per sample was observed. These results indicated that the material presented a good stability.

Fig. 6
figure 6

 Kinetics of RhB discoloration for ten solutions treated with TiO2/β-SiC foam

Full size image
Table 6 Apparent kinetics (Kap), coefficient of correlation (R2) and yield (Y%) of ten tests
Full size table

Effect of the recirculation flow rate on the photocatalytic removal

It is necessary to evaluate the effect of the recirculation flow rate. Figure 7 shows that the increase in flow rate induced a slight increase in discoloration rate from 81.4 to 85.6%. The same observation was done by Kouamé et al. (2012) and Sacco et al. (2018). This increase was due to diffusion between RhB and TiO2 catalyst which can be explained by the transfer of O2 an electron scavenger into the liquid phase (Merabet et al. 2009). Thus, it may inhibit the recombination of hole–electron pairs and lead to the oxidation of RhB dye.

Fig. 7
figure 7

Influence of flow rate on RhB removal (CRhB = 10 ppm)

Full size image

Conclusion

A full factorial design (24 test) was used to evaluate the effect of pH, dye concentration, volume of H2O2 and number of TiO2 β-sic foam samples on the photocatalytic removal of RhB. The Pareto analysis of the model terms showed that the photocatalytic process was highly influenced by the number of foam samples (53.01%) and the concentration of Rhodamine (30.49%). only the interaction between RhB concentration and volume of H2O2 has an influence on the removal rate (10.86%).

Analysis of variance showed the high coefficient of determination values (R2 = 0.991), thus ensuring a satisfactory adjustment of the first-order regression model with the experimental data. Analyses of the coefficients show that initial concentration of RhB and volume of H2O2 had a negative effect on the discoloration. However, quantity of TiO2/β-SiC foam and pH had a positive effect.

The study of aging of the foam had shown a good stability of the material, with a loss of only 3% of TiO2 after 30 successive tests. Furthermore, it is important to notice that the increase in the recirculation flow rate results in a slight increase in RhB discoloration rate which goes from 81.4 to 85.6%.