Electrochemical Degradation of Carbamazepine: A Case Study of Quantication of “10, 11 Dihydro-10 Hydroxy Carbamazepine And 10, 11-Epoxycarbamazepine” As The Main By-Products Using LC-TOF/MS.

Carbamazepine (CBZ) is one of the most widely used antiepileptic drugs in Malaysia, so, it was detected in wastewater frequently. The electrochemical treatment process has been applied for the degradation of CBZ using graphite-PVC as an anode. However, two main by-products, namely, 10,11-dihydro10-hydroxy carbamazepine (HDX-CBZ) and 10,11-epoxycarbamazepine (EPX-CBZ) have been analysed and quantied using liquid chromatography-time of ight/mass spectrometry (LC-TOF/MS). HDX-CBZ and EPX-CBZ were analysed in positive ionisation mode and were separated chromatographically using 5 mm, 2 mm´150 mm C18 column at a ow rate of 0.3 mL/min. To improve sensitivity and detectability, SPE was applied as a pre-concentration step for the treated carbamazepine samples to extract and pre-concentrate HDX-CBZ and EPX-CBZ. However, three different solvents, namely, methyl tertiary butyl ether, acetone and methanol, have been optimized to enhance the recovery. The recovery was 85% and 92% for HDX-CBZ and EPX-CBZ, respectively, in the presence of methanol. The limit of quantication (LOQ) was 0.588 and 0.109 µg/L for both by-products, respectively. The concentration of HDX-CBZ and EPX-CBZ was 343 and 144 μg/L, respectively, after 20 min of treatment, then, it was decreased to 17.2 and 9.8 μg/L at 40 min. Finally, both by-products were eliminated after 60 min.


Introduction
The emergence and fate of pharmaceuticals in the aquatic environment of Malaysia have become a major concern. These compounds have been frequently found in water bodies as well as in reused water, Compared to other chemical-oxidation treatment methods or biological treatment processes Li et al. 2021), electrochemical processes are preferred for many reasons: (i) high e ciency in removing contaminants from very complex matrices, as reported in a previous study (Mussa et al. 2015), (ii) it could be applied at normal conditions in terms of temperature and pressure so no more volatilization process occurred and also no sludge formation, (iii) the reaction itself can only be stopped when there is a lack of power and can be quickly recovered once the interoperability issue is resolved, (iv) it could be safe to the environment as the reagent is "electron", and (v) electrochemical process does not require more chemical reagents.
The anodic-electrochemical oxidation process is simply based on realising the electron to the solution followed by the formation of Cl 2 /H 2 O as the main oxidising agent. However, it could be classi ed into the direct and indirect oxidation process on the surface of the anode and/or at the bulk solution (Moreira et al. 2017). The big challenge in the degradation of pollutants is "what are the by-products after degradation and how to identify and quantify these unknown compounds?" The transformation products are varied, depending on the conditions of the experiment, type of electrolyte and so on. It was reported that after electrochemical degradation of carbamazepine in the presence of chlorine, the by-products Carbamazepine has been degraded using 100 mL-cell consists of the cathode, a platinum sheet of 1.5 × 1.5 cm dimensions. The opposite electrode was graphite-PVC anode as a pellet (20 mm diameter). However, both anode and cathode electrodes were connected to the DC power supply (CPX200 DUAL, 35 V 10 A PSU, Thurl, Ltd, St. Ives Cambridgeshire, England).
It was electrochemically treated for 20 min in the presence of a xed amount of 0.5 g NaCl and 5 volt. Then, the solutions were ltered and extracted before being injected into LC-TOF/MS.

Extraction of the by-products
For sensitive and accurate by-product identi cation, a solid-phase extraction procedure was applied for 100 ml of the treated sample. Before using SPE cartridges, the samples were ltered to exclude any particular matter. HLB cartridges were preconditioned by methanol and water in 2 mL. After that, all ltered samples were passed through the HLB cartridges using a 10-sample vacuum manifold at 3 ml/min under vacuum. The drying process was applied for 5 min at 15 mL/min for the cartridges to remove any expected residual water. Then, the by-products were eluted by dispersing 5 ml of methanol, then, evaporated to dry in a gas stream of N 2 . 1.0 ml of MeOH -DIW (10)  Quanti cation and method validation After reconstitution of the by-products, they were extracted accurately using a very narrow window "0.02 Da" in LC-TOF/MS instrument. The identi cation and con rmation of the by-products in the treated samples were based on two factors: (a) the exact value of the mass to charge ratio (m/z) of the main molecular ion and (b) the retained time (Rt) for the target by-product. The instrumental limit of quanti cation (IQL) was assessed by successive injections of the diluted standard solution until it reached a concentration level equal to an s/n ratio of ≥10. Calibration curves were generated for each by-product by matching the respective peak area with the concentrations of each analyte using a linear regression model. Linearity was good, in which the correlation coe cient (R 2 ) was > 0.993). The LOQ quanti cation limit (µg/L) for all methods was calculated using the following equation (Vieno et al. 2006): (1) Where IQL is the instrumental quanti cation limit (µg/L), R (%) is the recovery of the by-product in the sample matrix at the concentration level of IQL, and CF is the concentration factor (100).
To determine the recovery, ve replicates were performed at spiking levels of 50 and 10 µg/L for HDX-CBZ and EPX-CBZ, respectively. The recoveries were evaluated by comparing the solid phase extracted samples to non-extracted standard solutions as presented below: (2) Where A SP is the peak area of the analyte in the solid-phase extracted sample and A S is the peak area in the solvent-based standard solution.

Results And Discussion
HDX-CBZ and EPX-CBZ by-products were selected based on their appearance as the main products after electrochemical degradation of carbamazepine. The separation pro le for both by-products was presented in (Figure 1). The intensity of the peak area is not the same for both products and varied strongly. Therefore, this variation may be due to the diversity of physicochemical properties under electrospray ionization conditions. ToF screening and con rmation Liquid chromatography/time-of-ight/mass spectrometry is a developing technique involving the application of the electrospray ionization method to enhance the detection method. Furthermore, a sharp and nice peak exhibited good chromatographic separation in terms of sensitivity and selectivity for the quanti cation of target compounds in the samples. However, the by-products were analyzed in positive ionization mode (PI) because it exhibited a high signal-to-noise ratio for both by-products. The quanti cation and monitoring in treated samples were also considered in the PI mode. For enhancing the sensitivity of detecting trace levels of the pharmaceuticals using LC-TOF/MS, a very narrow range window of 0.02 Da has been applied to extract a target peak. It was observed that reducing the mass window plays an important role to increase the detection limit and avoiding any interferences from other organic compounds.
TOF/MS compartment has been applied for con rmation and quanti cation of the protonated molecular ion of an organic compound. This analysis was investigated by the elemental formula and mass errors, which has been achieved using the software of "Brucker Daltons Data Analysis program". An example of EPX-CBZ has been presented in Figure 2.
However, the elemental formula of EPX-CBZ (C 15 H 12 N 2 NaO 2 ) as sodium adduct molecular ion has a mass error of −2.9 ppm. In addition, some characteristics are shown in Table 2, including the mass measurement, retention time, mass error and elemental composition. Table 2 Accurate mass measurements obtained by LC-TOF/MS for both by-products (HDX-CBZ and EPX-CBZ) after electrochemical oxidation treatment of carbamazepine

Method validation
An external calibration curve has been built to exhibit the linearity represented by the correlation coe cient (R 2 ). A wide range of concentrations ranged between 50 and 1000 µg/L for HDX-CBZ while it was ranged between 10 and 1000 µg/L for EPX-CBZ (Table 3 and Figure 3). Good linearity (R 2 ≥ 0.993) was observed. For the calculation of instrumental quanti cation limit (IQL), the lowest concentration has been injected directly to the LC-ToF/MS, achieving a signal-to-noise ratio (S/N) of 10. HDX-CBZ and EPX-CBZ have IQL of 50 and 10 µg/L, respectively. The limit of quanti cation (LOQ) for the applied method was calculated using equation (1). They were 0.588 and 0.109 µg/L for HDX-CBZ and EPX-CBZ, respectively. Compared to the reported study by (Teixeira et al. 2013), the limits of quanti cation ranged between 25.9 and 39.0 µg/L, which is higher than the LOQs reported in the present study. The reason is related to the type of instrument used for method validation. However, LC-TOF/MS exhibited high sensitivity compared to the cyclic voltammetry method. Three concentrations (50, 100 and 200 µg/L) have been tested to investigate the inter-day repeatability and intra-day reproducibility in three replications. Accepted results have been obtained, in which the RSD% ≤ 6.7% for intra-day precision and 11.7% for inter-day precision as presented in Table 4. Accuracy was also considered in the present study, in which the recovery was good for both compounds through the SPE method.

Effect of elution solvent
Three solvents have been examined to achieve the best recovery for both by-products. However, methyl tertiary butyl ether (MTBE), acetone and methanol were investigated separately. The lowest recovery was observed with MTBE (5 × 1 mL) while the best recovery was achieved with methanol (5 × 1 mL).
Both by-products were prepared in the concentration of IQL value for each one, then, subjected to SPE method after that eluted with 5 mL of solvent. The by-products HDX-CBZ and EPX-CBZ exhibited the highest recovery of 85% and 92%, respectively with methanol as eluent. However, low recoveries were observed with MTBA and acetone solvents compared to methanol. It could be concluded that methanol, a polar solvent, compared to other solvents, resulted in increased recovery of both by-products. Finally, methanol has been selected as the best eluent for both by-products as presented in Figure 4. It is well known that in the electrochemical degradation process, a low concentration of by-products could be produced. However, the identi cation and quanti cation required an accurate and sensitive instrument, such as LC-TOF/MS.

Identi cation of EDX-CBZ and HDX-CBZ by-products
Two main by-products (HDX-CBZ: 6.092 min and EPX-CBZ: 6.478 min) were generated during the electrochemical degradation of CBZ. It was reported that HDX-CBZ and EPX-CBZ have been formed as main by-products after electrochemical degradation of carbamazepine using novel blue-colored TiO 2 nanotubes anode. The previous study was focusing on the qualitative analysis, not quanti cation since no standard for both by-products was calibrated and both by-products were detected as protonated molecular ions [M+H] + (Xu et al. 2021). Figure 5 shows the mass fragmentation chromatogram of both by-products, HDX-CBZ and EPX-CBZ as sodium adduct molecular ion ([M+Na] + . High S/N ratio was obtained for both by-products due to the pre-concentration sample using solid phase extraction protocol (3 cc HLB cartridges) and also to the optimal choice for the mobile phase and elution program as explained in sections 3.2 and 3.3. A very narrow window of "0.02 Da" has been applied for the extraction of chromatographic peaks using the Bruker Daltonic Analysis software.
Monitoring the by-products after degradation process It is well known that after the degradation of carbamazepine, the compound itself could be reduced with time due to the degradation process. The treated sample of carbamazepine has been retained and separated on a Gemini 5 µm NX 110Å C18 column (2 mm × 150 mm, Phenomenex) using liquid chromatography-time of ight/mass spectrometry. Carbamazepine chromatograms pro le as extracted ion chromatogram (EIC) after 0, 20, 40 and 60 min of the electrochemical treatment process are well presented in Figure 6.
The e ciency of the electrochemical process has been evaluated by monitoring the by-products for 60 min. After 20 min of the electrochemical degradation process, two main by-products HDX-CBZ and EPX-CBZ were detected and analyzed using LC-TOF/MS at retention times of 6.092 and 6.478 min, respectively. The main HDX-CBZ and EPX-CBZ by-products were quanti ed based on the linear regression equation as mentioned previously. Both by-products HDX-CBZ and EPX-CBZ were formed at the maximum concentrations of 343 and 144 µg/L, respectively at 20 min of the electrochemical treatment, then, they were decreased to 17.2 and 9.8 µg/L, respectively at 40 min as shown in Figure 7. This behavior was in agreement with a previous study reported by (Xu et al. 2021). He noticed that at the rst 20 min, the maximum yield of the by-products was achieved. After that, all by-products disappeared within 80 min.

Conclusions
An analytical method using liquid chromatography time of ight mass spectrometry was developed and validated for the identi cation and quanti cation of HDX-CBZ and EPX-CBZ as main by-products after electrochemical degradation of CBZ. Buying standard products helps in the quanti cation and identi cation of these by-products in treated samples as well as applying solid-phase extraction to reconstitute them. Recovery was 85% and 92% for HDX-CBZ and EPX-CBZ, however, the limit of quanti cation for both by-products was 0.588 and 0.109 µg/L, respectively. Furthermore, the fate of the by-products was investigated after monitoring the by-products for 60 min of electrochemical degradation of carbamazepine. It was observed that the maximum amount has been formed at 20 min, then, reduced with time until it disappeared. HDX-CBZ and EPX-CBZ were formed at concentrations of 343.6 and 144 µg/L, respectively at 20 min of electrochemical degradation of carbamazepine.
Ethics approval and consent to participate Not applicable.
Consent for publication Not applicable.

Competing interests The authors declare no competing interests
Funding Not applicable Availability of data and materials Not applicable  Example of the application Bruker Daltonics DataAnalysis software: the analysis of EPX-CBZ by-product.  Effect of elution solvent on recovery of the by-products; HDX-CBZ and EPX-CBZ.

Figure 5
The pro le of elucidating the by-products during the electrochemical oxidation process: (top) HDX-CBZ; (bottom) EPX-CBZ.

Figure 6
Liquid chromatography-time of ight/mass spectrometry chromatograms for carbamazepine under different conditions: 5V, 0.5 g NaCl and 0-60 min.