Topp and Starratt showed that about 10% of 14C-NP was extractable at 40 d, which was in general agreement with the current study . The formation of bound residue is considered a decontamination process, as the chemical has become an integral part of the soil matrix. In this study, bound residue was quantified by combustion of extracted soil samples. Figure 3.2 shows the fractions of bound residue in the incubated soils at 112 d. The levels of bound residue were significantly different among the PPCP/EDCs and followed the overall order BPA > NP > DCL ≥ NPX. For example, at 112 d of incubation in Maricopa soil, bound residue accounted for 66.2, 36.3, 29.6, and 14.9% of the spiked 14C-labeled BPA, NP, DCL, and NPX, respectively . However, no significant difference was noted among the different soils, except for the sterilized Irvine soil, which had significantly reduced levels of bound residue for most compounds. For example, at 112, bound residue for NP was 45.4 ± 16.3% in Irvine soil, 43.9 ± 8.2% in the compost amended Irvine soil, 17.1 ± 6.5% in the sterilized Irvine soil, 36.3 ± 1.2% in Maricopa soil,vertical gardening systems and 34.8 ± 4.6% in Ventura soil . Few studies have examined bound residues of PPCP/EDCs, as such analysis requires the use of 14C-labeling and combustion of solvent-extracted samples.
In Fent et al. , 79% was determined to be in the form of bound residue following incubation of 14C-BPA for 120 d, which was slightly higher than that found in this study . Kreuzig et al. measured the bound residue at 44 – 78% of the spiked 14C-DCL after 102 d of incubation in two soils, which was greater than in the current study for 14C-DCL . Such differences may be attributed to the different soil properties, such as organic carbon content, and to the specific 14C-labeling positions among the studies. The tendency of an organic compound to become bound is strongly affected by the hydrophobic partitioning to soil organic matter . Since the soils used in this study had similar low organic content , this may explain why the levels of bound residue for a compound were similar among soils. Due to this partitioning process, compounds with higher hydrophobicity typically become more bound than compounds with lower hydrophobicity . DCL and NPX are both ionizable compounds, with pKa values of 4.0 and 4.19, respectively, suggesting that they were partly ionized under the experimental conditions used . Ionic compounds are much more polar than neutral compounds like BPA and NP, which may explain why NP and BPA had more bound residue and less extractable residue than DCL and NPX. Microbially-mediated mineralization represents the complete breakdown of an organic compound and is therefore regarded as an environmentally beneficial decontamination process. Throughout the soil incubation in this study, the mineralized 14CO2 was continuously sequestered in NaOH solution and periodically measured. Figure 3.3 shows the cumulative 14C mineralization for each compound in the different soil treatments.
The final mineralized fractions were significantly different among the different PPCP/EDCs in the same soil and followed the order NPX > DCL > BPA > NP. For example, in Irvine soil the cumulative fraction mineralized at the end of 112 d of incubation reached 74.8 ± 2.4% of the initially spiked amount for NPX, 65.7 ± 3.6% for DCL, 22.9 ± 0.1% for BPA, and only 9.2 ± 3.7% for NP. These differences were likely related to the extractable fraction of each compound that may be bio-available for microbial metabolism . This relationship was supported by the fact that mineralization followed the same order as the abundance of the extractable fraction. In addition, it must be noted that the location of the 14C-label may have also contributed to the different mineralization rates. As shown in Figure 3.1, 14C was present on a substituent group in DCL and NPX, while the aromatic ring was labeled for BPA and NP. The mineralization rate differed among the soil treatments. For BPA and DCL, mineralization in Maricopa soil was more limited in comparison to Irvine or Ventura soil . For example, at the end of the 112 d incubation, the cumulative mineralized fraction for 14C-DCL in Maricopa soil was 49.8 ± 3.9%, lower than that in Irvine or Ventura soil . On the other hand, mineralization for NPX and NP was generally similar among the three soils. For example, the fractions of 14C-NP mineralized after 112 d were 9.2 ± 3.7%, 10.0 ± 0.6%, and 10.7 ± 0.8% for Maricopa, Irvine, and Ventura soils, respectively.
Since the extractable fraction was not different among the soils for a compound, the differential mineralization rates suggested that the variation was likely due to differences in microbial population and activity in these soils. Both Irvine and Ventura soils were from agricultural fields not previously exposed to treated wastewater, while Maricopa soil was from an uncultivated area used as a groundwater recharge basin for over 10 years. It is likely that agricultural cultivation and exposure to treated wastewater, respectively, may have led to the establishment of specific microbial communities in these soils, resulting in preferential transformations of some compounds in a given soil. Sterilization of Irvine soil significantly decreased mineralization of PPCP/EDCs , confirming the role of soil microorganisms in the transformation of these PPCP/EDCs. In addition, amendment of compost to the Irvine soil generally resulted in decreased mineralization, with the exception of NPX which was quickly mineralized in all non-sterilized soils. For example, mineralization of BPA decreased from 22.9 ± 0.1% in Irvine soil to 17.5 ± 0.3% after compost addition .In Fent et al. , 13.1 – 19.3% of the spiked 14C-BPA was mineralized after 120 d in four soils, which was in good agreement with the 14.2 – 22.9% range observed for unmodified soils in this study. Mineralization of 14C-DCL was monitored for 102 d in Kreuzig et al. and was found to be 13% of the spiked amount, which was substantially smaller than that in the current study . This reduced mineralization may be attributed to the difference in the 14C labeling position between the two studies. Topp et al. reported that about 50% of 14C-NPX was mineralized after 27 d of incubation, and the rapid mineralization was in agreement with the current study . In a separate study, Topp and Starratt observed that 40% of the initially spiked 14C-NP was mineralized after 40 d of incubation at 30 °C in a sandy soil ,vertical tower for strawberries which was greater than that measured in this study . On the other hand, Shan et al. reported that only 5% of spiked 14C-NP was mineralized after 58 d of incubation at room temperature. Since 14C-NP was labeled on the aromatic ring in all these studies, the higher mineralization in Topp and Starratt may be partly attributed to the use of a higher incubation temperature in that study. Mineralization was the major loss pathway for DCL and NPX, which amounted to 49.8 – 68.3% and 69.2 – 78.3% of the initially spiked 14C, respectively. In comparison, formation of bound residue appeared to be the predominant dissipation pathway for BPA and NP in the soils considered in this study, accounting for 53.0 – 66.2% and 34.8 – 45.4% of the initially spiked 14C, respectively. At the end of 112 d of incubation, the extractable fraction for each compound was consistently smaller than the mineralized or bound residue fraction, suggesting that these PPCP/EDCs were mostly removed in three months through mineralization or formation of bound residue.
Concurrently, the potentially bio-available extractable residue greatly diminished. Residues extracted from Irvine, Maricopa, and Ventura soils were further analyzed to characterize the composition of extractable 14C after 14 d and 112 d of incubation. Extracted 14C was identified as the parent compound, transformation products appearing during the HPLC run, and transformation products appearing in the SPE filtrate. The results for Irvine soil are shown in Figure 3.4. Extensive transformation of parent PPCP/EDCs was evident in all soils for most compounds. For example, after 112 d of incubation, parent compounds accounted for only ND – 13.8% of the extractable 14C for BPA and 2.4 – 8.4% for NP. The remaining extractable 14C was in the form of transformation products recovered early in the HPLC run or in the SPE filtrate, suggesting that transformation led to the formation of intermediates more polar than the parent. Differences among soil types were also evident. For example, while no parent compound was detected at the end of the incubation for DCL and NPX in Maricopa soil, the majority of the extractable residue was found as the parent for DCL and NPX in Ventura soil. A first-order decay model was used to fit the dissipation of parent compounds in the different treatments. The calculated half-lives ranged from 1.4 to 5.4 d for all PPCP/EDCs in the unmodified soils . The test compounds were relatively more persistent in Ventura soil, and less persistent in Maricopa soil, likely reflecting differences in the native microbial communities. The half-lives of BPA and NP in this study were generally similar to those previously reported . However, the half-lives calculated for DCL and NPX were somewhat shorter than those reported by Xu et al. or Lin and Gan . This difference may be caused by the different soils and experimental conditions used. Overall, none of the PPCP/EDCs considered in this study exhibited significant persistence in soil as the parent compound. When compared to chromatograms of the parent compound, many transformation products were evident in the soil extracts, and the relative presence of transformation products in the extractable 14C generally increased over time . For example, in Maricopa soil treated with 14C-NPX, 62.1% of the extractable 14C was associated with transformation products at 14 d and the fraction increased to about 100% at 112 d. For NP treatments, 14C was also detected in the SPE filtrate. Since preliminary experiments showed that 14C-NP was quantitatively retained by the SPE cartridge, the 14C in the SPE filtrate may be assumed to be polar compounds not adsorbed by the cartridge sorbent. The extensive transformation of spiked PPCP/EDCs in soil extracts suggests the importance of considering degradation intermediates in addition to the parent compound . Samples from BPA and DCL treatments were further analyzed on UPLC/MS/MS to tentatively identify degradation intermediates. Authentic standards were used to verify the identity by matching retention time and mass transitions of the isolated peaks . In solvent extracts from soil treated with BPA, 4-hydroxyacetophenone , 4- hydroxybenzaldehyde , and 4-hydroxybenzoic acid were detected as transformation products . While information on BPA degradation in soil is very limited, it was suggested by Spivack et al. that BPA may undergo oxidative rearrangement to form 1,2-bis-2-propanol, which is then dehydrated to 4,4′-dihydroxy-a-methylstilbene. Oxidative cleavage may then result in HBA and HA, and further oxidation of HBA forms HBacid. At 14 d, extracts of Maricopa and Ventura soils treated with DCL showed the presence of 5-hydroxydiclofenac , as well as 2,6-dichlorobenzoic acid in Ventura soilonly. A small amount of 2,4-dichlorobenzoic acid was detected in Irvine soil. At the end of 112 d of incubation, 5HD was detected in all soils, while 24DCB and 26DCB were found in Irvine soil, 24DCB and 3,5-dichlorobenzoic acid in Maricopa soil, and 24DCB in Ventura soil. It is likely that oxidation of DCL led to the formation of 5HD, and both DCL and 5HD may serve as precursors to DCB through Ndealkylation of the biphenyl compounds followed by carboxylation . 4′-Hydroxydiclofenac was analyzed for, but not detected in any sample, in contrast to other observations made using microbial culture or human metabolic enzymes . Only a few previous studies examined the transformation products of PPCP/EDCs in soil, sediment, or sewage. In a soil incubated with gram negative bacteria, Spivack et al. identified some of the same intermediates of BPA as in this study. However, even though degradation of DCL was evaluated in sewage , sediment , and in fungal cultures , no effort was made to identify the specific dichlorobenzoic acid isomers. Little information is available about the toxicity of these transformation products as compared to their parent forms.