Simple Catalytic Approach for Removal of Analytical Interferences Caused by Hydrogen Peroxide in a Standard Chemical Oxygen Demand Test

The difference between measured COD (${\mathrm{COD}}_{\mathrm{meas}}$ after ${\mathrm{H}}_2{\mathrm{O}}_2$ addition) and original COD (${\mathrm{COD}}_{\mathrm{org}}$ before ${\mathrm{H}}_2{\mathrm{O}}_2$ addition) is the excess in COD caused by ${\mathrm{H}}_2{\mathrm{O}}_2$, as a ratio in ${\mathrm{mg}}_{\mathrm{COD}}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$. Table S1 lists the ratios determined in this work and in the literature for different wastewaters, whereby interference ratios are in the range $10.2-19.72{\mathrm{mg}}_{\mathrm{COD}}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$. These results demonstrate the need for a uniform experimental approach to remove ${\mathrm{H}}_2{\mathrm{O}}_2$ from wastewater samples without changing COD.

To overcome the difficulties in determining COD values in effluents containing ${\mathrm{H}}_2{\mathrm{O}}_2$, different methods were investigated. Eq. (2) was first used to calculate a ${\mathrm{COD}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ of $620\mathrm{mg}/\mathrm{L}$ for all tested wastewaters based on the ${\mathrm{H}}_2{\mathrm{O}}_2$ added to each sample ($44.5{\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}/\mathrm{L}$). ${\mathrm{COD}}_{\mathrm{meas}}$, measured in the presence of ${\mathrm{H}}_2{\mathrm{O}}_2$, was corrected by ${\mathrm{COD}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ to ${\mathrm{COD}}_{\mathrm{cal}}$ and compared with ${\mathrm{COD}}_{\mathrm{org}}$ of the original wastewater (Table 1). Results of the comparison are summarized in Table 2. The difference (ΔCOD) between ${\mathrm{COD}}_{\mathrm{cal}}$ and ${\mathrm{COD}}_{\mathrm{org}}$ indicated the predictive quality of the COD calculation based on measured or known ${\mathrm{H}}_2{\mathrm{O}}_2$ concentrations for different wastewater types.

ΔCOD ranged from $-126$ to $+147\mathrm{mg}/\mathrm{L}$, where negative ΔCOD indicated that predicted COD was lower than original (measured) COD. The relative error was between $-10.5\%$ (for leachate) and 76.6% (for CID) and increased with decreasing COD values. Eq. (2) may therefore be useful to estimate possible interference with an error of $\pm 130–150\mathrm{mg}/\mathrm{L}$ in COD at an ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration of $44.5\mathrm{mmol}/\mathrm{L}$. However, this error is often not acceptable for evaluations of COD removal efficiency. These results and the interferences shown in Table S1 confirm that the interference ratio is influenced by general wastewater characteristics and not only by ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration [as stated by Kang et al. (1999)] in Eq. (2).

The carbonate approach was then tested for SWW containing $90{\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ under the conditions proposed by Wu and Englehardt (2012) [90°C and high ${\mathrm{Na}}_2{\mathrm{CO}}_3$ ($38{\mathrm{mmol}}_{{\mathrm{Na}}_2{\mathrm{CO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$)], which resulted in a pH of 11. Samples were taken during the 60-min reaction of SWW at 90°C to determine COD and ${\mathrm{H}}_2{\mathrm{O}}_2$ concentrations at different reaction times. The standardized values (${\mathrm{COD}}_{\mathrm{t}}/{\mathrm{COD}}_{\mathrm{org}}$) are shown in Fig. S1. During the first 5 min, $90\mathrm{mmol}/\mathrm{L}$ of ${\mathrm{H}}_2{\mathrm{O}}_2$ was removed but COD also decreased by $\sim 10\%$. COD was reduced by $\sim 16\%$ in a reaction time of more than 5 min, indicating a competing reaction of COD oxidation caused by oxidative species, as well as ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition. Therefore, the role of carbonate/bicarbonate species and their pH effect were studied in further detail.

To study the effect of ${\mathrm{NaHCO}}_3$ on ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition, the decrease in ${\mathrm{H}}_2{\mathrm{O}}_2$ in two samples of deionized water and 1 M ${\mathrm{NaHCO}}_3$ was observed at ambient temperature. After shaking both samples for 24 h, ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration decreased from $44.5{\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}/\mathrm{L}$ to 43.2 and $7.2{\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}/\mathrm{L}$ in deionized water and 1 M ${\mathrm{NaHCO}}_3$ solution, respectively. Iodometric measurement confirmed this result, which was directly visible by the different color densities for the deionized water and the ${\mathrm{NaHCO}}_3$ solution, as shown in Fig. 3(a). The calculated ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition rate was approximately $1.55\mathrm{mmol}/\mathrm{L}/\mathrm{h}$ in 1 M ${\mathrm{NaHCO}}_3$ solution, while in deionized water ${\mathrm{H}}_2{\mathrm{O}}_2$ was almost stable for one day.

During shaking, the generation and decomposition of ${\mathrm{HCO}}_4^{–}$ was expected according to Reaction 3 followed by Pathways B and C, generating the final species: ${\mathrm{HCO}}_3^{–}$, ${\mathrm{H}}_2\mathrm{O}$, and ${\mathrm{O}}_2$. The generation of oxygen was confirmed by the formation of oxygen bubbles observed in samples containing ${\mathrm{NaHCO}}_3$. The bubbles started to grow on the wall of sample tubes 20 min after ${\mathrm{NaHCO}}_3$ was added [Fig. 3(b)]. Dissolved oxygen was measured at ambient temperature (∼ 20°C) in both samples, at $20.52\mathrm{mg}/\mathrm{L}$ for the ${\mathrm{NaHCO}}_3$ solution and $9.63\mathrm{mg}/\mathrm{L}$ for the deionized water, indicating higher saturation with pure oxygen generated by ${\mathrm{NaHCO}}_3$ in solution.

Table 3 summarizes the progress of ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition and pH to determine the appropriate dose of ${\mathrm{NaHCO}}_3$ to completely remove ${\mathrm{H}}_2{\mathrm{O}}_2$. The addition of ${\mathrm{NaHCO}}_3$ shifted the pH value of SWW from almost neutral (pH 7.1) to a basic range ($\mathrm{pH}\sim 9$). ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration decreased at ambient temperature nonlinearly with ${\mathrm{NaHCO}}_3$ concentration over 24 h. In the first 5 h, only $2.1\mathrm{mmol}/\mathrm{L}{\mathrm{H}}_2{\mathrm{O}}_2$ decomposed at an ${\mathrm{NaHCO}}_3$ concentration of $0.5\mathrm{mol}/\mathrm{L}$ (Sample 1), whereas $\sim 8$ and $20\mathrm{mmol}/\mathrm{L}{\mathrm{H}}_2{\mathrm{O}}_2$ decomposed when the ${\mathrm{NaHCO}}_3$ concentration was 1 and $2\mathrm{mol}/\mathrm{L}$, respectively. After 24 h, $6\mathrm{mmol}/\mathrm{L}$ of ${\mathrm{H}}_2{\mathrm{O}}_2$ was still found in Sample 1, whereas ${\mathrm{H}}_2{\mathrm{O}}_2$ was completely removed in Samples 2 and 3. The minimum dose of ${\mathrm{NaHCO}}_3$ for complete ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition of the samples with residual ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration of up to $44.5\mathrm{mmol}/\mathrm{L}$ was calculated as $22.5{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ (with 1 M ${\mathrm{NaHCO}}_3$ solution).

To clarify the applicability of the proposed method at higher ${\mathrm{H}}_2{\mathrm{O}}_2$ concentrations, the last run was repeated with an ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration of $90\mathrm{mmol}/\mathrm{L}$. As shown in Table 3, after 24 h shaking of SWW samples with 1 M ${\mathrm{NaHCO}}_3$ solution, an ${\mathrm{H}}_2{\mathrm{O}}_2$ residual of $1.2\mathrm{mmol}/\mathrm{L}$ was still found (Sample 4). A 2-mol/L increase in ${\mathrm{NaHCO}}_3$ ensured the complete decomposition of ${\mathrm{H}}_2{\mathrm{O}}_2$ after 24 h (Sample 5), which met the ratio of $22.2{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$. Calculation of the concentration ratio (${\mathrm{NaHCO}}_3{:\mathrm{H}}_2{\mathrm{O}}_2$) for samples with complete ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition gave a minimum value of $22.5{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ (Sample 2) for $44.5\mathrm{mmol}/\mathrm{L}{\mathrm{H}}_2{\mathrm{O}}_2$, and $22.2{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ for $90\mathrm{mmol}/\mathrm{L}{\mathrm{H}}_2{\mathrm{O}}_2$. Therefore, we propose a value of $22.5{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ as a minimum ratio to ensure complete ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition regardless of ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration.

The proposed method was then applied to the four wastewaters (Table 1) containing ${\mathrm{H}}_2{\mathrm{O}}_2$ at two concentrations (44.5 and $90\mathrm{mmol}/\mathrm{L}$). For each wastewater, four samples were prepared by mixing wastewater with ${\mathrm{NaHCO}}_3$ solutions of 1 (Samples A and B) and $2\mathrm{mol}/\mathrm{L}$ (Samples C and D) ($1:1$ volume per volume). The $22.5{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ ratio was almost met in Samples B and D ($44.5{\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}/\mathrm{L}$ and $90{\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}/\mathrm{L}$, respectively). Samples A and C were run as blanks without ${\mathrm{H}}_2{\mathrm{O}}_2$. After 24-h shaking, no ${\mathrm{H}}_2{\mathrm{O}}_2$ was detected in Samples B and D for $\mathrm{all}$ wastewater types, as shown in Table 4, demonstrating that ${\mathrm{H}}_2{\mathrm{O}}_2$ was effectively decomposed by ${\mathrm{NaHCO}}_3·\text{Additionally}$, the comparison of blank and treated samples indicated minimal changes considering the standard deviation between original COD and COD after ${\mathrm{NaHCO}}_3$ pretreatment.

Under the experimental conditions, there appeared to be more than one factor influencing the decomposition of ${\mathrm{H}}_2{\mathrm{O}}_2$ into its final products (${\mathrm{O}}_2{,\mathrm{H}}_2\mathrm{O}$, and ${\mathrm{HCO}}_3^{–}$). The first factor was the high ${\mathrm{NaHCO}}_3$ concentration versus the low ${\mathrm{H}}_2{\mathrm{O}}_2$ concentration ($22.5{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$). The high concentration of ${\mathrm{NaHCO}}_3$ activated the decomposition reaction (Reaction 3) forming ${\mathrm{HCO}}_4^{–}$ (Bakhmutova-Albert et al. 2010), which was subsequently decomposed through the preferred pathways (B and C). The second factor was the basic pH value of all samples (Table 4). A basic pH of 8.62–8.82 increased the abundance of ${\mathrm{HCO}}_3^{–}$ species according to lime-carbonic acid equilibrium (Zeebe and Wolf-Gladrow 2005); however, it also supported the self-decomposition of ${\mathrm{H}}_2{\mathrm{O}}_2$ according to Reaction 4 (Kolyagin and Kornienko 2003).

The extrapolation of COD results in Table 4 was supported by the formation of oxygen bubbles observed in Samples B and D (active ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition by ${\mathrm{HCO}}_3^{–}$), as shown in Fig. 3(b). This observation confirmed the dominance of ${\mathrm{O}}_2\text{-}\mathrm{generating}$ reactions in Pathways B and C until ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition was complete.

A comparison of Sample A with Sample B and Sample C with Sample D for each wastewater showed very close COD values, with a deviation of $4\mathrm{mg}/\mathrm{L}$ (9.5% relative to $42\mathrm{mg}/\mathrm{L}$) to $64\mathrm{mg}/\mathrm{L}$ (4.8% relative to $1.328\mathrm{mg}/\mathrm{L}$). Additionally, Samples A and C (Table 4) confirmed the original COD values for all wastewater samples (Table 1) with absolute deviations between 0 and $4\mathrm{mg}/\mathrm{L}$. These results indicated that the addition of ${\mathrm{NaHCO}}_3\le 2\mathrm{mol}/\mathrm{L}$ did not interfere with the COD measurements (see Runs A and C) and that the decomposition of ${\mathrm{H}}_2{\mathrm{O}}_2$ was complete (see Runs B and D), while the original sample values were verified after ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition (Table 4).

Additionally, the calculated standard deviation (SD) of the COD analysis (Table 4) relating to the accuracy of the spectrophotometer (1%) and sample preparation was approximately 1%–11% relative to the original COD. Therefore, the deviations derived by the proposed approach were in the same range as those derived by the standard method. Indeed, Samples B and D, with ${\mathrm{NaHCO}}_3$ solution, showed a small decrease in COD (Table 4) compared with the COD of the original sample (${\mathrm{COD}}_{\mathrm{org}}$ in Table 1). This small decrease may indicate that the competing reaction pathway (A in Fig. 1) made only a small contribution to COD oxidation, as proposed by Yang et al. (2019), because ${\mathrm{HCO}}_4^{–}$ reacts with electron-rich compounds, such as organic sulfides (Richardson et al. 2000), amines (Balagam and Richardson 2008), and alkenes (Yao and Richardson 2000), which may exist in untreated wastewater. In untreated leachate, this fraction seemed to be higher than in the other wastewaters examined, resulting in the highest SD, which could be neglected, however, due to the much higher COD value of the leachate.

In treated wastewaters, especially those treated by AOPs including the addition or generation of ${\mathrm{H}}_2{\mathrm{O}}_2$, organic electron-rich compounds is not expected due to their oxidation by AOPs. Therefore, the competition reaction (A in Fig. 1) is expected to be less relevant for treated wastewater and proposed pretreatments with ${\mathrm{NaHCO}}_3$. This approach would be more useful for correct COD measurements, which are essential to determine the required limit of COD effluent or to evaluate the COD removal effeciency by AOPs. In summary, the proposed approach at ambient temperature removed ${\mathrm{H}}_2{\mathrm{O}}_2$ without any significant change in the COD of original samples for the four wastewaters.

The 24-h pretreatment described is sometimes unacceptable due to delays such as for process evaluation. A typical method to accelerate chemical reactions is increasing the temperature. A simple approach for the described method of ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition is to use the same COD standard test equipment for heating and decomposition. The thermal heat block system commonly includes preset temperatures (e.g., 70°C, 90°C, 100°C, 148°C) for the reaction of organic compounds in the standard COD test. Here, for the mixture of SWW and ${\mathrm{NaHCO}}_3$ as well as for subsequent heating and reactions at a fixed standard temperature, empty cuvettes appropriate for the thermal heat block system were used.

The ${\mathrm{SWW}/\mathrm{NaHCO}}_3$ mixture, with the proposed ratio of $22.5{\mathrm{mmol}}_{{\mathrm{NaHCO}}_3}{/\mathrm{mmol}}_{{\mathrm{H}}_2{\mathrm{O}}_2}$ ($\mathrm{pH}\sim 9$) was heated at 70°C, 100°C, and 148°C to accelerate the decomposition of ${\mathrm{H}}_2{\mathrm{O}}_2$. Each experiment was stopped when the ${\mathrm{H}}_2{\mathrm{O}}_2$ was no longer present. Fig. 4 shows that decomposition was complete after 5 min at 148°C. In contrast to the results at ambient temperature ($\sim 20°\mathrm{C}$), $\sim 22\%$ COD removal occurred, indicating oxidation of the original COD.

At 100°C and 5 min reaction time, 6 mmol ${\mathrm{H}}_2{\mathrm{O}}_2$ remained in the samples: a 12% increase in COD compared with the original tested wastewater. Further heating at 100°C for 30 min caused an additional decrease in COD, to the same concentration as at 148°C. The results for 100°C and 148°C indicated oxidation of a constant fraction of COD in the SWW by ${\mathrm{NaHCO}}_3\text{-}\text{activated}{\mathrm{H}}_2{\mathrm{O}}_2$ following the oxidation pathway (A in Fig. 1). Parallel to these experiments, blank samples of ${\mathrm{SWW}/\mathrm{NaHCO}}_3$ without ${\mathrm{H}}_2{\mathrm{O}}_2$ were heated at 70°C, 100°C, and 148°C for 60, 30, and 20 min, respectively. No COD removal was detected. A decrease in COD at $\ge 100°\mathrm{C}$ is expected to be catalytic oxidation by the ${\mathrm{NaHCO}}_3/{\mathrm{H}}_2{\mathrm{O}}_2$ mixture.

At 70°C, ${\mathrm{H}}_2{\mathrm{O}}_2$ decreased from 90 to 6 and $2\mathrm{mmol}/\mathrm{L}$ after 20 and 30 min, respectively. The measured COD after this pretreatment was approximately equal to ${\mathrm{COD}}_{\mathrm{org}}$ before treatment. Therefore, 30-min pretreatment at 70°C seems to accelerate ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition without COD removal. The optimal temperature for a wide range of water and wastewater types is recommended for further study.

To further elucidate the effect of pH on ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition, experiments were repeated four times at 70°C and side by side for ${\mathrm{NaHCO}}_3$ and ${\mathrm{Na}}_2{\mathrm{CO}}_3$ as ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition catalysts. This resulted in pHs of $\sim 9$ and $\sim 11$, respectively. Fig. 5 summarizes the results with mean values and SD (details of each experiment are provided in Table S2). The carbonate catalyst decomposed ${\mathrm{H}}_2{\mathrm{O}}_2$ approximately 4–5 times faster than the bicarbonate catalyst. Additionally, the COD of the original SWW sample decreased by $211\mathrm{mg}/\mathrm{L}$ with the carbonate catalyst (from 1,195 to $984\mathrm{mg}/\mathrm{L}$) after 1 h heating, whereas COD removal was four times lower ($55\mathrm{mg}/\mathrm{L}$) during ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition by the bicarbonate catalyst. The main difference between the carbonate and the bicarbonate was pH.

To confirm the importance of pH, additional experiments were carried out using sodium bicarbonate, in which pH was increased from $\sim 9$ to $\sim 10.7$ by NaOH addition, followed by ${\mathrm{H}}_2{\mathrm{O}}_2$ addition and a 30-min reaction at 70°C. According to lime-carbonic acid equilibrium (Zeebe and Wolf-Gladrow 2005), this action is supposed to force most ${\mathrm{NaHCO}}_3$ to change into ${\mathrm{Na}}_2{\mathrm{CO}}_3$ prior to the activation reaction of ${\mathrm{H}}_2{\mathrm{O}}_2$. Fig. S2 shows that ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition increased in the first 20 min while COD degraded to below the original concentration after 30 min, similar to COD degradation with ${\mathrm{Na}}_2{\mathrm{CO}}_3$ (Fig. 5). The difference between the effect of carbonate and that of bicarbonate on COD removal is most probably related to the quantity and quality of oxidative species formed at different pH values. Other than ${\mathrm{HCO}}_4^{–}$ and ${\mathrm{HO}}_2^{–}$ as oxidative species, singlet oxygen (${}^1{\mathrm{O}}_2$) was probably generated through the recombination of ${\mathrm{O}}_2^{–}$ formed during ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition, as observed by Bokare and Choi (2015) for alkaline solutions. Moreover, the copresence of ${\mathrm{O}}_2^{–}$ and ${\mathrm{H}}_2{\mathrm{O}}_2$ could lead to the slow generation of hydroxyl radicals via the Haber-Weis mechanism (Bokare and Choi 2015). Using NBT (1.5 mM) to detect ${\mathrm{O}}_2^{–}$ caused the occurrence of deep-blue formazan pigment precipitates after just 1 min in the carbonate sample, while the bicarbonate was still light yellow (Fig. S3). The dark color of the carbonate sample indicated the generation of ${\mathrm{O}}_2^{–}$ radicals produced via the catalytical decomposition of ${\mathrm{H}}_2{\mathrm{O}}_2$ (by carbonate). Continuous shaking for 24 h caused the reaction between ${\mathrm{O}}_2^{–}$ and NBT in both samples to become much darker, with more precipitates in the carbonate sample than in the bicarbonate sample, which was light blue with few precipitates in [Figs. S3(b and c)]. This confirmed that ${\mathrm{NaHCO}}_3$ enabled much less ${\mathrm{O}}_2^{–}$ species generation than ${\mathrm{Na}}_2{\mathrm{CO}}_3$, which was responsible for the original COD that remained unchanged. This suggests that ${\mathrm{NaHCO}}_3$ is beneficial for ${\mathrm{H}}_2{\mathrm{O}}_2$ decomposition to avoid interference in spectrophotometric COD tests. A quantitative determination of those oxidative species will be a topic for future research.