3.2.1. Persistent Compounds
Six compounds including carbamazepine, candesartan, olmesartan, primidone, as well as the transformation products DiOH-CBZ and PEMA did not show attenuation when passing through the column, since concentrations measured in the column outlet were similar to those measured in the column inflow solution (see Figure S1 in the Supplementary Material
). The fact that persistence prevailed in both columns is an indication for the general persistence of these compounds in the aquatic environment, apparently independent of the prevailing hydrochemical conditions and sediment composition.
In case of carbamazepine the environmental persistence has been frequently pointed out (e.g., [43
]). The poor biodegradability is accompanied by a low tendency to sorb onto soils and sediment (e.g., [46
]), pointing towards its high environmental relevance and ability to enter various aquatic compartments including groundwater. DiOH-CBZ, known as the predominant human metabolite evolving from carbamazepine and frequently detected in wastewater treatment plant effluents [48
], likewise persisted along the flow path at concentrations around 0.5 µg L−1
In lab-scaled sewage treatment plants olmesartan and candesartan were shown to be rather poorly degradable with total elimination percentages <20% [29
]. Similar to our results, Hellauer et al. [50
] found candesartan and olmesartan to be biologically persistent in a column system simulating bank filtration conditions. In the same study, the process of ozonation led to an efficient removal of both compounds. Further, Khan and Nicell [51
] expected candesartan to be highly mobile and persistent and suggested its prioritization for further studies.
Primidone and its metabolite PEMA were detected in rather small concentrations of 7 ng L−1
in the surface water used as column influent solution, and—independent of the prevailing redox conditions—no removal was observed along the flow path. The environmental persistence of primidone was already pointed out by other authors in the course of field studies (e.g., [11
]) and lab studies (e.g., [21
]). Although the number of studies on the environmental behavior of the primidone metabolite PEMA is small, its poor biodegradability has been revealed by Hass et al. [52
] and Nham et al. [54
3.2.2. Reactive Compounds
Due to high consumption rates, as well as largely unaffected passage through the human body and persistence during wastewater treatment [44
], the artificial sweetener acesulfame is generally detected at elevated concentrations in treated wastewater [56
] and therefore used as an indicator for wastewater influenced surface waters [58
] and groundwater [60
]. Presuming a low sorption affinity and recalcitrance to microbial degradation, acesulfame has formerly been proposed as an ideal anthropogenic marker [45
]. However, other previous studies revealed that—under certain conditions—acesulfame is actually prone to microbial degradation [14
]. Thereby, acesulfame degradation is largely affected by (i) temperature, (ii) redox conditions, and (iii) biodegradable carbon content [64
]. This is supported by our findings, as acesulfame was attenuated under oxic conditions with degradation rate constants of 14.6 day−1
(core A) and 10.4 day−1
(core B), and under suboxic conditions (λ = 2 day−1
), but persisted under anoxic conditions (Figure 3
a). According to Kahl et al. [64
], first evidences questioning the recalcitrance of acesulfame came up in 2014. Based on their findings the same authors hypothesized, that acesulfame degrading species evolved during the last few years—for example, due to horizontal gene transfer.
The phenazone type metabolite FAA has formerly been stated to behave redox dependent [34
]. This was also proven during this study, as fast degradation was recognized for the oxic zone while FAA persisted under anoxic conditions (Figure 3
b). Degradation rate constants observed under oxic conditions were 24.7 day−1
(core A) and 7.8 day−1
(core B), which are higher than those published elsewhere [23
]. However, rate constants of similar magnitude (5.7 day−1
and 1.4 day−1
) have already been noticed under oxic conditions within sandy columns [14
A similar picture—characterized by an efficient removal within the upper oxic zone paired with persistence within the remaining suboxic to anoxic part of the column—emerges for metoprolol, pregabalin and valsartan acid (Figure 3
e–g). Based on field data, Nödler et al. [31
] suggested valsartan acid to behave persistent during bank filtration. However, evaluating the data for valsartan acid of this study yielded the highest degradation rate constant of all compounds (31.5 day−1
), corresponding to a half-life time of 0.5 h (core A). In accordance, Hellauer et al. [50
] found valsartan acid to be efficiently attenuated after aeration of a test system, while persistence was noticed in an anoxic reference system.
With regard to metoprolol, for which degradation rate constants of 19.7 day−1
(core A) and 4.9 day−1
(core B) have been observed, our results concur with results reported by other authors, who found a strong redox dependency of metoprolol degradation by means of laboratory experiments [18
], and also confirm our previous findings [66
The anticonvulsant gabapentin behaved different along the distinct cores—while the concentration decreased along core A with degradation rate constants of 1.1 day−1
in the upper (oxic to suboxic) part and 0.2 day−1
in the lower (anoxic) part, persistence was observed along core B (Figure 3
c). Hence, the availability of oxygen does not seem to be the controlling factor for gabapentin degradation. These findings differ from those reported by Henning et al. [68
] and Hellauer et al. [50
], who observed by means of batch and column experiments, respectively, a redox dependent degradation of gabapentin with enhanced removal under aerobic conditions. Since one major difference between both cores is the organic carbon content, which is higher in core B, low carbon contents may favor gabapentin removal. However, by investigating the influence of particular organic carbon on oxygen consumption and attenuation of organic trace compounds, Filter et al. [19
] detected neither a distinct impact of the carbon content nor any correlation with the prevailing redox regime. Indications for biodegradation of gabapentin in GAC fixed-bed and tertiary filtration systems have also been found in pilot-scale studies on advanced water and wastewater treatment steps [33
]. Further studies for clarification are needed here.
Gabapentin-lactam, the quantitatively most relevant transformation product evolving from Gabapentin [68
], showed enhanced attenuation under oxic conditions whereas the concentration remained constant under reducing conditions (Figure 3
d). The degradation rate constants obtained were 23.0 day−1
for core A and 3.6 day−1
for core B. These results share similarities with those of Henning et al. [68
], who described also a redox dependent degradation of gabapentin-lactam, even though the reported degradation rate constants of 0.06 day−1
were clearly lower.
The attenuation pattern of oxypurinol appeared to be quite different. While persistence was noticed within core A, decreasing concentrations along core B were detected. The degradation rate constants ranged from 0.7 day−1
in the upper part to 2.2 day−1
in the lower part. Hence, oxypurinol was more efficiently attenuated under strongly reducing (sulfidic) conditions. By investigating its degradation during managed aquifer recharge, Hellauer et al. [71
] found oxypurinol to be persistent during two meters of infiltration under oxic conditions.
By comparing the degradation rate constants observed during this study as shown in Figure 4
, it becomes evident that highest removal rates mostly appeared under oxic conditions (blue bars). Four compounds, namely FAA, metoprolol, pregabalin and valsartan acid, where solely degraded under oxic conditions, while acesulfame and gabapentin-lactam also were prone to degradation under suboxic conditions. Deviating from that, the removal of oxypurinol and gabapentin seems not primarily to be controlled by the redox environment, as oxypurinol concentration decreased only along core B while gabapentin concentrations only decreased along core A.
Further, it is apparent from Figure 4
that degradation rate constants within the upper, oxic zone observed considering core A (blue, solid bars) are systematically higher than those observed in core B (blue, striped bars). Highest discrepancies appeared for gabapentin-lactam, for which the removal within the upper (oxic) part of core A was 6 times larger than in the upper (oxic) part of core B. By looking at the sediment characteristics, the main evident difference is the content of organic carbon determined by the loss of ignition, as it is by factor 10 higher in the upper zone of core B than of core A (Figure 2
). Thus, it seems reasonable that the organic carbon content of the sediment influenced the degradation processes. Consistently, Kahl et al. [64
] suggested the removal of acesulfame to be most efficient when the availability of biodegradable organic carbon is low. However, during investigations the influence of a compost layer on the attenuation of organic micropollutants, Schaffer et al. [72
] observed an enhanced degradation when levels of biodegradable dissolved organic carbon were higher. Since this relation remains unclear, further research needs to be done.