Curve fits for the tested experimental conditions (Eqs. 1–2) are depicted in Figs. 1–5.
PLEASE, INSERT FIGURE 1 HERE
PLEASE, INSERT FIGURE 2 HERE
PLEASE, INSERT FIGURE 3 HERE
PLEASE, INSERT FIGURE 4 HERE
PLEASE, INSERT FIGURE 5 HERE
Mean R2 values were calculated for the curve fits (Figs. 1–5), with R2 = 0.984 for HA and R2 = 0.964 for FAH. The rate constants fitted in this study are listed in Table 2.
Table 2
First-order rate constants
Constant
|
Experimental conditions (runs)
|
I (1 and 5)
|
II (2 and 6)
|
III (3 and 7)
|
IV (4 and 8)
|
V (9 and 10)
|
kHA (min− 1)
|
4.4 × 10− 3
|
3.0 × 10− 3
|
2.3 × 10− 3
|
3.1 × 10− 3
|
2.0 × 10− 3
|
kFAH (min− 1)
|
7.0 × 10− 4
|
9.0 × 10− 4
|
1.5 × 10− 3
|
1.3 × 10− 3
|
7.5 × 10− 4
|
kHA, kinetic constant for humic acids; kFAH, kinetic constant for fulvic acids + humins. |
It can be seen that kinetic parameters did not vary greatly as a function of pH range or catalyst system. Therefore, we applied analysis of variance (ANOVA) to assess the significance of these variables in organic fraction degradation. However, differences as a function of contaminant type (HA or FAH) are evident.
A study on HA decomposition by FeNi3-SiO-TiO2 photodegradation found a kHA of 15.5 × 10− 3 to 220 × 10− 3 min− 1 for HA concentrations in the range 15 to 2 mg L− 1, respectively (Khodadadi et al., 2020). These results indicate that higher HA concentrations lead to lower first-order rate constants. In the current study, we obtained even lower kHA values, which is consistent with the HA concentrations used here (about 600 mg L− 1). Oskoei et al. (2016) observed a similar behavior for UV/ZnO nanophotocatalysis under pH 4: kHA ranged from 0.038 to 0.112 min− 1 for HA concentrations of 10 to 5 mg L− 1, respectively. He (2013) estimated a kHA value of 9.6 × 10− 3 min− 1 for the photocatalytic degradation of HA by TiO2 under natural sunlight. The referred process is similar to that applied in the current study, and the findings corroborate those described in Table 1. For solar photocatalysis of HA (68 mg L− 1) catalyzed by TiO2 in pH 7.8, a slightly higher kHA value (0.092 min− 1) was estimated (Al-Rasheed and Cardin, 2003).
Rajca and Bodzek (2013) studied FA and HA photodegradation by TiO2 and found a kHA value of 0.0025 min− 1 and a kFA value of 0.0024 min− 1 for initial HA and FA concentrations of 9.6 and 8.8 mg L− 1, respectively. It can be observed that FA rate constants are slightly lower than those for HA. The kinetic data obtained in the current study are in agreement with literature data.
From the ANOVA results presented in Table S1, a regression model for HA conversion after 240 min was developed (Eqs. 9 and 10 for replicates 1 and 2, respectively). The model had an R2 of 0.96, with normal distribution of residuals (p > 0.15, Kolmogorov–Smirnov test) (Fig. S5).
$${HA}_{1}=0.4130-0.03125\left[ZnO\text{–}Ti{O}_{2}\right]+0.1088\left[ZnO\text{–}Ti{O}_{2}\right]\times \left[ZnO\text{–}Ti{O}_{2}\right]+0.07375\left[ZnO\text{–}Ti{O}_{2}\right]\times \left[pH\right]$$
9
$${HA}_{2}=0.3570-0.03125\left[ZnO\text{–}Ti{O}_{2}\right]+0.1088\left[ZnO\text{–}Ti{O}_{2}\right]\times \left[ZnO\text{–}Ti{O}_{2}\right]+0.07375\left[ZnO\text{–}Ti{O}_{2}\right]\times \left[pH\right]$$
10
In the model, run replication was considered a categorical variable, given that higher conversions were obtained in the first run (replicate 1) of each condition than in the second run. The model (Eqs. 9 and 10) showed that the replicate variable was significant. According to the model, the mean difference in HA conversion between replicates 1 and 2 was 5.6% (0.413 − 0.357 = 0.056). Such a difference might be due to the possibility that HA was adsorbed on the plate painted with the catalyst mixture. We highlight that plates were washed with detergent at the end of each run. Thus, it is possible that there was an average decrease in HA adsorption after the first run. Liu et al. (2008) found that HA have affinity for adsorption onto TiO2 surface, particularly under acidic pH conditions (close to 4.0), because of its point of zero charge and zeta potential. Tran et al. (2006) found that carboxylic acids have greater tendency for adsorption, whereas alcohols and long-chain saturated aliphatic compounds do not. Ren et al. (2018) explained that adsorption is undesirable and represents an inhibitory factor in photocatalysis. To decrease the occurrence of adsorption onto TiO2, it is necessary to operate under more alkaline pH conditions. However, in this study, we used a pH greater than 4.
ZnO has a positive zeta potential at pH 6.7 to 9.3 (Mohd Omar et al., 2014). HA molecules, in turn, are negatively charged at pH 2.0 to 10.7. With the increase in solution pH (> 7.0), HA molecules become more easily ionized because of the action of electrostatic attraction forces (Nguyen et al., 2020). However, the same adsorption potential was observed when pH was lower than the point of zero charge, resulting from a complex interaction between attractive and repulsive electrostatic forces, van der Waals bonds, and other interactions with humic substances (Mohd Omar et al., 2014). A similar effect was observed in dairy wastewater, as described by Samanamud et al. (2012). The authors observed enhanced ZnO photocatalytic activity at acidic pH; however, the possible adsorption of contaminants onto the semiconductor surface was not discussed.
Another interesting result of the ANOVA table (Table S1) is the significant interaction between pH and catalyst concentration (p < 0.001). The highest HA conversion was estimated to be achieved using acidic pH (− 1) and high TiO2 concentration (− 1). This result suggests that, when working with binary mixtures, synergy between compounds favors degradation kinetics. According to the model, the action of TiO2 is favored by acidic medium and that of ZnO by alkaline medium, greatly enhancing absorption of the solar spectrum.
The ANOVA table for FAH (Table S2) allowed the construction of a regression model for FAH conversion after 240 min (Eqs. 11 and 12 for replicates 1 and 2, respectively). The model had an R2 of 0.992, with normal distribution of residuals (p > 0.15, Kolmogorov–Smirnov test) (Fig. S6).
$${FAH}_{1}=0.18800+0.06000\left[\text{ZnO–Ti}{\text{O}}_{\text{2}}\right]+0.04500\left[ZnO\text{–}Ti{O}_{2}\right]\times \left[ZnO\text{–}Ti{O}_{2}\right]-0.02000\left[ZnO\text{–}Ti{O}_{2}\right]\times \left[pH\right]$$
11
$${FAH}_{2}=0.13200+0.06000\left[\text{ZnO–Ti}{\text{O}}_{\text{2}}\right]+0.04500\left[\text{ZnO–Ti}{\text{O}}_{\text{2}}\right]\times \left[\text{ZnO–Ti}{\text{O}}_{\text{2}}\right]-0.02000\left[\text{ZnO–Ti}{\text{O}}_{\text{2}}\right]\times \left[pH\right]$$
12
Similar to that found for HA degradation, the categorical replicate variable was significant in the study of FAH degradation (Eqs. 11 and 12) (p < 0.001). The mean difference between FAH conversion replicate runs was of 5.6% (0.188–0.132 = 0.056), suggesting that both fractions (FAH and HA) diffused in a similar manner inside catalyst pores. The model suggests an antagonistic behavior in the preference of the catalyst for FAH over HA. This is because, according to the model (Eqs. 11 and 12), FAH conversion is highest at acidic pH (− 1) and high ZnO concentration (+ 1). The finding is corroborated by the kinetic model for the degradation of humic substances: a higher rate constant for FAH (kFAH = 1.5 × 10− 3 min− 1) was obtained when using higher ZnO concentrations and acidic pH (4.5–5.0) (condition III), in agreement with the observations of Mohd Omar et al. (2014). Thus, the results show a greater affinity of the binary mixture for photocatalytic FAH degradation when high ZnO concentrations are used.