Effect of ethoxylate number and alkyl chain length on the pathway and kinetics of linear alcohol ethoxylate biodegradation in activated sludge

• PDF
• Split View
• • Article contents
  • Figures & tables
  • Video
  • Audio
  • Supplementary Data

• Cite

  Cite

  Nina R Itrich, Thomas W Federle, Effect of ethoxylate number and alkyl chain length on the pathway and kinetics of linear alcohol ethoxylate biodegradation in activated sludge, Environmental Toxicology and Chemistry, Volume 23, Issue 12, 1 December 2004, Pages 2790–2798, https://doi.org/10.1897/04-053.1

  Close

Abstract

Batch activated‐sludge die‐away studies were conducted with various pure homologs to determine the effect of ethoxylate number and alkyl chain length on the kinetics of primary and ultimate biodegradation of linear alcohol ethoxylates. The ¹⁴C‐(ethoxylate) homologs C₁₄E₁, C₁₄E₃, C₁₄E₆, and C₁₄E₉ were used to investigate the effect of ethoxylate number, and ¹⁴C‐(ethoxylate) homologs C₁₂E₆, C₁₄E₆, and C₁₆E₆ were used to examine the effect of chain length. Activated sludge was dosed with a trace concentration (0.2 μM) of each homolog, and the disappearance of parent, formation of metabolites, production of ¹⁴CO₂, and uptake into solids were monitored with time. Ethoxylate number had little effect on the first‐order decay rates for primary biodegradation, which ranged from 61 to 78 h⁻¹. However, alkyl chain length had a larger effect, with the C₁₆ chain‐length homolog exhibiting a slower rate of parent decay (18 h⁻¹) compared to its corresponding C₁₂ and C₁₄ homologs (61–69 h⁻¹). Ethoxylate number affected the mechanism of biodegradation, with fission of the central ether bond to yield the corresponding fatty alcohol and (poly)ethylene glycol group increasing in dominance with increasing ethoxylate number. Based upon the measured rates of primary biodegradation, removal of parent during activated‐sludge treatment was predicted to range between 99.7 and 99.8% for all homologs except C₁₆E₆, which had a predicted removal of 98.9%. Based upon the measured rates of ultimate biodegradation, removal of ethoxylate‐containing metabolites was predicted to exceed 83% for all homologs. These predictions corresponded closely with previously published removal measurements in laboratory continuous activated‐sludge systems and actual treatment plants.

INTRODUCTION

Alcohol ethoxylates are a family of nonionic surfactants characterized by a hydrophobic alkyl chain and a polar ethoxylate head group. Commercial linear alcohol ethoxylates (LAEs) are complex mixtures of homologs with alkyl chains of varying lengths with a range of ethoxymers. Linear alcohol ethoxylate homologs are identified by using the convention C_xE_y, where x is the total carbon number in the alkyl chain and y is the number of ethoxylate groups (OCH₂CH₂) in the molecule. Mixtures are similarly identified, with x and y designating the range or average chain length and ethoxylate number. Detergent‐range LAEs typically encompass C_12–18 alkyl chain lengths and an ethoxylate range of 1 to 22 units. In 2000, combined production of LAEs in North America and Western Europe exceeded 435,000 metric tons, with approximately 60% being used in household laundry products [1].

The biodegradability of alcohol ethoxylates has been studied extensively in numerous screening tests [2,3], and detergent‐range mixtures usually are ready biodegradable. In synthesizing the results of more than 250 screening tests, Swisher [3] concluded that ultimate biodegradation was favored by linearity of the hydrophobe and shortness of the ethoxylate chain. Biodegradation of LAEs has been observed to proceed by three different pathways [3]. The most common, central fission, occurs when the ether bond connecting the hydrophobe to the ethoxylate chain is cleaved, leading to the formation of fatty alcohol and polyethylene glycol (PEG), which are biodegraded independently. Alternatively, the terminal carbon of the alkyl chain can undergo omega oxidation, which is followed by inward biodegradation involving beta oxidation. Lastly, the terminal end of the hydrophile can undergo hydrolytic or oxidative attack. Although the most frequently reported pathway for biodegradation is central fission followed by rapid degradation of the resulting alcohol and slower degradation of the PEG [4–6], the mode of initial attack appears to be linked to experimental conditions and the complexity of the hydrophobe [5].

Although screening tests are useful to demonstrate the inherent and ultimate biodegradability of chemicals, more realistic simulation tests are needed to determine the actual kinetics and pathways in the environment. An example of such testing was performed by Larson and Games [7], who examined the mineralization kinetics of two different LAE homologs, labeled in both the alkyl and ethoxylate moieties, in river water. In the case of a C₁₂E₉ homolog, the alkyl chain was converted to CO₂ more than twice as fast as the ethoxylate moiety, whereas similar rates of mineralization were observed both for the alkyl and ethoxylate moieties of a C₁₆E₃ homolog. A continuous activated‐sludge (CAS) test is a dynamic simulation of fate during activated‐sludge wastewater treatment. It accurately predicts effluent concentrations of parent and metabolites that can enter a receiving stream. However, the CAS test is limited in that the results are only directly applicable to a single set of operating conditions, defined by the hydraulic retention time (HRT) and solids retention time (SRT) [8,9]. Four CAS studies on LAEs have been reported in the literature. Steber and Wierich [10] conducted a study on a single ¹⁴C‐radiolabeled C₁₈E₇ homolog, and Kravetz et al. [11] evaluated a commercial mixture with alkyl chain lengths C_12–15 and an average ethoxylate number of nine by using a dual‐labeled test material that was ³H labeled in the hydrophobe and ¹⁴C labeled in the ethoxylate chain. More recently, Battersby et al. [12] assessed the removal of both C_12–15E₃ and C_12–15E₇ mixtures in a CAS system, whereas Szymanski et al. [13] examined a C₁₂E₁₀ mixture. To date, no studies have systematically evaluated both the kinetics and pathways of LAE biodegradation as a function of chain length and ethoxylate number under realistic wastewater treatment conditions.

This paper describes a series of radiolabeled batch activated‐sludge die‐away [14] tests that were used to examine the biodegradation kinetics of specific LAE homologs under realistic activated‐sludge sewage‐treatment conditions. A trace level of each test chemical was dosed to activated sludge and the rates of primary and metabolite degradation, as well as mineralization, were followed by using specific radioanalytical methods. The progression of metabolites also was examined to elucidate the mechanism by which the LAE components biodegraded. This testing was conducted in two phases with relatively pure radiolabeled homologs. The ¹⁴C‐(ethoxylate) homologs C₁₄E₁, C₁₄E₃, C₁₄E₆, and C₁₄E₉ were evaluated to investigate the effect of ethoxylate number, and ¹⁴C‐(ethoxylate) homologs C₁₂E₆, C₁₄E₆, and C₁₆E₆ were tested to examine the effect of chain length on biodegradation. The C₁₄E₆ served as a benchmark during both phases of testing.

MATERIALS AND METHODS

Test materials

The ¹⁴C homologs were synthesized at the Procter and Gamble Company (Cincinnati, OH, USA). For each homolog, the ¹⁴C label was located in a single carbon position within the ethoxylate moiety. In the case of C₁₄E₁, the ¹⁴C syntheses involved addition of an ethoxylate group by reacting the starting fatty alcohol with ¹⁴C‐chloroacetic acid. In the case of C₁₄E₃, the synthesis involved reacting a purified C₁₄E₂ LAE with ¹⁴C‐bromoacetic acid. In the case of the other homologs, the synthesis involved addition of a radiolabled ethoxylate to the starting alcohol with ¹⁴C‐bromoacetic acid followed by addition of five or more additional ethoxylate units. Radiochemical purity was determined by radio–thin‐layer chromatography (radio‐TLC) analysis by using a Bioscan Imaging 200 System (Bioscan, Washington, DC). Specific activity was determined with a Beckman LS 6500 scintillation counter (Beckman Instruments, Fullerton, CA, USA). Table 1 shows the structure and ¹⁴C position for each homolog as well as molecular weight, radiochemical purity, and specific activity. In the case of the ¹⁴C‐C₁₄E₃, 76% of the total impurity was ¹⁴C‐C₁₄E₂, whereas approximately 50% of the total impurity in the ¹⁴C‐C₁₂E₆ was ¹⁴C‐C₁₂E₅. All other impurities could not be identified.

Overview of biodegradation testing

The testing was performed in two phases. In phase 1, the effect of ethoxylate number on biodegradation kinetics was examined by using C₁₄E₁, C₁₄E₃, C₁₄E₆, and C₁₄E₉. In phase 2, the effect of chain length was examined by using C₁₂E₆, C₁₄E₆, and C₁₆E₆. The test procedure has been described previously by Federle and Itrich [14]. Briefly, a trace concentration of radiolabeled test material was dosed to freshly collected activated sludge in an open test system. Periodically, sludge samples were collected and lyophilized, and the dried solids were extracted. The disappearance of parent and progression of metabolite formation and decay were monitored over time by TLC of the extracts with radioactivity detection. Production of ¹⁴CO₂ was determined by comparing total radioactivity in a bioactive treatment compared to that in an abiotic control by using liquid scintillation counting. Incorporation into bio‐mass was monitored by combustion and liquid scintillation counting analysis of the extracted sludge solids.

Test system

Activated‐sludge mixed liquor was freshly collected on two occasions from Downingtown Regional Water Pollution Control Center (Downingtown, PA, USA). This plant receives predominantly domestic wastewater. The solids level of the mixed‐liquor suspended solids was adjusted to 2,500 mg/L before use. Two treatments were prepared for each test material that consisted of 1 L each of biologically active sludge and abiotic control sludge in 2‐L flasks. The abiotic control was prepared by amending the sludge with mercuric chloride at 1 g/L followed by autoclaving for 90 min. Each of the ¹⁴C test materials was dosed at a concentration of 0.2 μM in both treatments, which were continually mixed on a shaker table and incubated at 20 ± 2°C. Samples were collected for both mineralization and chemical analysis at each sampling interval. In an effort to accurately assess the early events in the kinetics of biodegradation, rapid sampling occurred over the first 8 h of the test. Subsamples were taken at 1, 5, 15, 30, 45, 60, and 90 min. Subsequent samples were collected every 2 h until the 8‐h time point, and then removed every 24 to 48 h for the remainder of the experiment.

Analysis of mineralization

Mineralization was based on the difference between the total radioactivity in the bioactive and abiotic treatments after acidification. Triplicate 1‐ml aliquots of sludge were taken from each treatment and placed into separate 20‐ml glass scintillation vials containing 1 ml of 0.5% HCl. The acidified samples were incubated overnight, mixed with 15 ml of Ultima Gold Scintillation Cocktail (Packard, Meriden, CT, USA), and analyzed by liquid scintillation counting.

Analysis of parent and metabolite

At each sampling, 10 ml of sludge from each treatment was transferred to a 15‐ml screwtop centrifuge tube and immediately frozen in a dry ice‐acetone bath. The frozen samples were stored at −80°C until lyophilized by using a Virtis bench‐top model 3.3L freeze dryer (Virtis, Gardiner, NY, USA). The lyophilized sludge from phase 1 was extracted twice with 5 ml of methanol:acetone (2:1, v/v) to recover parent and metabolites, whereas the samples from phase 2 were extracted twice with methanol. The sludge solids from both phases were subsequently extracted twice with 5 ml of water. After removing aliquots for liquid scintillation counting, the solvent extracts were concentrated under nitrogen, and the water extracts were frozen and lyophilized. Both were reconstituted in a minimal volume of the original diluent, and subsamples were spotted onto prechanneled 60 A Silica Gel 60 TLC plates (Whatman, Clifton, NJ, USA). The plates used for the C₁₄E₁ homolog were developed in ethyl acetate:hexane (1:1, v/v), whereas the plates used with the other homologs were developed in chloroform:methanol:formic acid (90:10:1, v/v/v). After development, the plates were allowed to dry and were then scanned with a Bioscan Imaging 200 System.

Analysis of sludge solids for biomass incorporation

The extracted solids were quantitatively transferred to microfuge tubes and centrifuged. The supernatant was decanted, and the solids were combusted in a Packard Sample Oxidizer (model 307, Hewlett‐Packard, Avondale, PA, USA) system and analyzed by liquid scintillation counting.

Kinetic and statistical analyses

RESULTS AND DISCUSSION

The average recovery of radioactivity from the abiotic controls exceeded 93% for all the homologs, and the majority (>98%) of this radioactivity was present as parent at the termination of the experiments (Table 2). The bulk of the remaining radioactivity was associated with the solids. The average recovery from the biologically active treatments was consistently lower than that from the abiotic controls (82–88%), which is attributed to the difficulty in accurately assessing mineralization to ¹⁴CO₂. In this study, mineralization was determined by comparing liquid scintillation counts of mixed liquor samples from the biotic and abiotic treatments. These samples contain a high level of suspended solids that reduce counting efficiency. Although alternate approaches for assessing mineralization, such as trapping of evolved ¹⁴CO₂ with alkali, are not limited by counting efficiency, they have other limitations, including trapping efficiency and physical transfer rates that mask the true rate of mineralization. Given the kinetic limitations of other approaches for measuring mineralization, the rapid rates of mineralization expected, and abundant evidence showing that LAEs degrade completely, the method to measure mineralization was selected based upon its ability to provide the most accurate kinetic information.

Table 3 shows the final disposition of radioactivity in the test systems. In the abiotic controls, the bulk of the radioactivity (>97% of the total recovered) was still in the form of parent with a small residual associated with solids (0.4‐1.9% of initial). In the biotic treatments, none of the radioactivity was present as parent alcohol ethoxylate and no more than 4% was in the form of metabolites that could be resolved through chromatography. The majority of the radioactivity that was not mineralized was associated with the solids (7–14% of initial) or was present as extractable but chromatographically indistinct materials (3–7% of initial).

Figure 1 shows chromatograms of the methanol:acetone extracts from the biotic treatments after a 5‐min incubation period during phase 1 with C₁₄ homologs with different ethoxylate numbers. Similar chromatograms (not shown) of extracts from the abiotic controls exhibit a single major peak representing the parent test materials with very small minor peaks representing radiochemical impurities. The chromatograms from biotic treatments in phase 2 are not shown but have a similar pattern to that observed for C₁₄E₆ during phase 1. In addition to showing the presence of residual parent, the 5‐min sample chromatograms for all the homologs revealed the appearance of poorly resolved polar materials that remained at the origin or migrated just above it, which are designated as polar metabolites I. In addition, chromatograms for some homologs showed the appearances of a metabolite somewhat more polar than the parent, which is designated as polar metabolite II, and a nonpolar metabolite, which migrated farther than the parent. The distribution and position of these peaks combined with previous information on likely metabolites provide insight into the mechanisms of initial biodegradation. Patterson et al. [15] isolated acidic and neutral PEGs during the biodegradation of a commercial C_16–18E₉ mixture and concluded that initial primary biodegradation occurs via two simultaneous but independent pathways. These included central fission, involving the cleavage of the ether bond connecting the hydrophobe to the ethoxylate, and omega oxidation of the terminal methyl group in the alkyl chain to yield a carboxyalkyl ethoxylate. The occurrence of these simultaneous and independent modes of attack has since been confirmed in a pure culture with Pseudomonas sp. strain SC25A [16] and in bench‐scale activated‐sludge systems [10]. Central fission of the test materials used in this study would yield ¹⁴C‐PEGs, which migrate just above the origin, appearing to the far left of the parent in the chromatograms. The actual R_f values for specific PEGs vary with chain length. Omega oxidation of the terminal methyl group in the alkyl chain hydrophobe, and subsequent shortening of the chain through beta oxidation, would yield carboxyalkyl ethoxylates, which initially would be slightly more polar than the parent and then become increasingly more polar and impossible to resolve from PEGs as the alkyl chain becomes shortened. Although not previously reported as a primary degradation mechanism for LAEs, oxidative attack of the terminal ethoxylate alcohol to form a carboxylate has been reported for alkyl phenol ethoxylates [11,15]. Such a material would be less polar than the parent under the chromatographic conditions used in this study and migrate farther up the plate than the parent.

Figures 2 and 3 show the disappearance of the parent, formation and disappearance of metabolites, incorporation into solids, and formation of ¹⁴CO₂ as a function of time during the first 48 h. Disappearance of parent was extremely rapid, being nearly complete within the first hour and was associated with the instantaneous appearance of metabolites, which transiently accounted for up to 80% of the total radioactivity, depending upon the homolog. After this rapid appearance, metabolite levels decreased with the concurrent formation of ¹⁴CO₂. Both the disappearance of metabolites and the production of ¹⁴CO₂ were biphasic and occurred initially at a very fast rate and then at a slower rate. The transition from the fast to slow rates occurred after 1 to 4 h, depending upon the homolog. In addition to being converted to metabolites and mineralized to CO₂, some of the radiolabel was incorporated into biomass. Table 3 shows the level of radioactivity associated with the solids at the end of the study (168 h). This level was substantially greater in the biotic compared to the abiotic treatments, consistent with incorporation into biomass. In the abiotic controls, the average radioactivity associated with the solids ranged from 0.4 to 1.9% and did not change with time, thus reflecting materials that could not be recovered through extraction. Figures 2 and 3 show that in the biotic treatments, the level of radioactivity associated with the solids increased with time and reached a maximum within the first 24 h, followed in most cases by a gradual decline. On average, the percentage of radioactivity associated with the solids was highest for the higher ethoxylate materials (E₆ and E₉), which could reflect the different anabolic fates of the different metabolites.

It was not possible to definitively identify all metabolites because of their low concentrations and transient nature as well as a lack of putative standards. Nevertheless, inspection of the metabolite patterns (Fig. 1) and the progression of their formation and disappearance (Figs. 2 and 3) indicate that ethoxylate number significantly affected the relative importance of the various pathways for biodegradation. In the case of C₁₄E₁, the disappearance of the parent was associated with the immediate appearance of materials, which were both more polar and less polar than the parent. Because the C₁₄E₁ homolog has only one ethoxylate group and the radiolabel in the C₁₄E₃ homolog is located in the terminal ethoxylate group, central fission would yield ₁₄C‐ethylene glycol, which would be highly polar and remain close to the origin. Oxidative attack of the terminal methyl group of the alkyl chain would result in the formation of a ¹⁴C‐carboxyalkyl ethoxylate, which would be more polar than the parent and appear to the left of the parent, whereas omega oxidation of the terminal ethoxylate would result in the formation of ¹⁴C‐alkyl ethoxy carboxylates, which would be less polar and migrate to the right of parent. In the chromatograms from the earliest time points, the polar material is likely the central fission product, ethylene glycol, whereas the less‐polar metabolite can only be the alkyl ethoxylate carboxylate. Thus, it appears that initial attack on the C₁₄E₁ homolog involved both central fission and oxidation of the terminal ethoxylate, with the latter process being more predominant. Notably, no evidence was found of carboxyalkyl ethoxylates being formed.

In the case of the C₁₄E₃ homolog, the initial disappearance of parent was associated with the appearance of three materials, a polar material close to the origin, a material just to the left and somewhat more polar than the parent, and a material less polar than the parent at a low level. The first corresponds to triethylene glycol; the second likely corresponds to the omega oxidation product of the alkyl chain, and the third can only represent alkyl ethoxy carboxylate. Hence, three modes of attack are probably operative for this homolog, central fission, omega oxidation of terminal methyl of the hydrophobe, and omega oxidation of the terminal ethoxylate, with the first two being the most prominent. In the case of C₁₄E₆, only two peaks initially arise with the disappearance of parent, a polar peak corresponding to PEG and a smaller peak slightly more polar than the parent, which is likely the carboxyalkyl ethoxylate. Hence, for this homolog, central fission appears to predominate, whereas omega oxidation of the hydrophobe occurs to a smaller extent. This same pattern also is observed with this homolog during phase 2 and with C₁₂E₆ and C₁₆E₆. In the case of C₁₄E₉, only one metabolite is initially formed, consistent with central fission being the sole mechanism of primary biodegradation for this homolog. Therefore, in activated sludge, the mode of attack for most materials involved central fission and the predominance of this route increased as ethoxylate number increased. This observation is consistent with results from CAS studies. Steber and Wierich [17], who used radiolabeled C₁₈E₇ model compounds labeled in both the alkyl and ethoxylate moieties, reported that the level of neutral PEG intermediates in the effluent of a CAS system was higher than that for carboxyalkyl ethoxylates. Neutral PEG‐like intermediates were found in effluents of CAS systems dosed with C_14–15E₁₁, C_14–15E₁₅, and C_14–15E₂₀ [18], and PEG was enriched in the effluent of CAS units fed C_12–15E₃ and C_12–15E₇ [12] as well as C₁₂E₁₀ [13]. Notably, although this study confirms that central fission is a predominant mechanism of biodegradation, it is the first to show that the degree of this dominance can be affected by ethoxylate number.

Table 4 shows the kinetic constants describing the decay of parent, disappearance of total parent and metabolites, and production of ¹⁴CO₂. Data were fit to a two‐compartment first‐order model, which assumes two pools of material, biodegraded at two distinct rates. In the case of parent decay, the first rate (k₁) describes the biodegradation of the largest and most bioavailable pool (A%), which is most likely to escape sewage treatment. The second rate (k₂) describes the biodegradation of a smaller and less bioavailable pool (B%), consisting of sorbed chemical. This sequestered parent can be viewed as an artifact of the batch system and is not relevant under continuous‐flow conditions and would minimally contribute to the level of parent escaping into effluent under actual activated‐sludge conditions because of its association with sludge solids. The k₁ of primary biodegradation, which describes the rate at which the majority of the compound degrades, ranged from 61.6 to 78.4 h⁻¹ and increased with increasing ethoxylate number for the C₁₄ chain‐length materials in phase 1. These rates represent half‐lives ranging from 0.5 to 0.7 min, and the rate correlated with ethoxylate number (r = 0.92; p ≤ 0.1). In the second set of experiments, the effect of hydrophobe length on biodegradation was examined by using materials with six ethoxylate units. The C₁₄E₆ was common to both sets of experiments and behaved similarly, with k₁ values of 70.1 ± 3.2 h⁻¹ and 61.4 ± 3.1 h⁻¹ in phases 1 and 2, respectively, which were performed with sludge samples obtained three weeks apart. In phase 2, k₁ decreased with increasing chain length and ranged from 17.9 to 69.2 h⁻¹, with half‐lives ranging from 0.6 to 2.3 min. Although not significant because of the low number of degrees of freedom, the correlation coefficient between k₁ and chain length equaled −0.93 in phase 2. When combining the k₁ data from both phases, the only substantial correlation existed between k₁ and chain length (r = −0.74; p ≤ 0.1). Both alkyl chain length and ethoxylate number can affect hydrophobicity and sorption. To assess this combined effect, k₁ values from both phases were compared to sorption coefficients (K_d) previously reported for these homologs [19]. In this analysis, k₁ and K_d were negatively correlated (r = −0.87; p < 0.01), further indicating that biodegradation rate and hydrophobicity are inversely related.

Because decay of parent was extremely rapid compared to that for the metabolites, the rates describing the disappearance of combined parent and metabolites primarily describe the biodegradation of the metabolites. The vast majority of this disappearance was described by k₁, which ranged in phase 1 from 0.80 to 2.7 h⁻¹ and increased with increasing ethoxylate number for the C₁₄‐chain‐length materials. These rates represented half‐lives ranging from 15 to 84 min and were significantly correlated with ethoxylate number (r = 0.98; p ≤ 0.05). The k₁ describing combined parent and metabolite biodegradation for C₁₄E₆ was 1.7 h⁻¹ in both phases 1 and 2, showing the reproducibility of the test and the ability to directly compare results from both experimental phases. In phase 2, k₁ did not vary in a consistent pattern with chain length, ranging from 1.0 to 1.7 h⁻¹ with half‐lives ranging from 24 to 42 min. Rates describing the mineralization of the homologs to ¹⁴CO₂ were comparable in magnitude to the rates describing the decay of combined parent and metabolites in both experimental phases. The mineralization rate (k₁) ranged from 1.0 to 4.1 h⁻¹.

The percent of intact LAE present in influent predicted to occur in the dissolved and sorbed form in effluent is shown in Table 5. For all homologs, total removal of parent was predicted to exceed 98%, and with the exception of C₁₆E₆, where removal was predicted to fall between 99.7 and 99.8%. With the exception of C₁₆E₆ parent LAE, the contribution of LAE sorbed to effluent solids was less than 8% of the total amount in effluent. Removal of C₁₆E₆ was predicted to be lower than the other homologs because of its slower primary biodegradation rate. In addition, the contribution of sorbed C₁₆E₆ in effluent (17%) was much greater because of its higher sorption. These predictions for removal efficiency are comparable to those measured in CAS tests in the laboratory and in monitoring of actual wastewater plants in the field. Kravetz et al. [23] reported that parent removal of C_12–15E₉ based upon analysis of cobalt thiocyanate active substances in a CAS system operated with an HRT of 8 h exceeded 98%. Steber and Wierich [10] reported approximately 99% removal for a model C₁₈E₇ homolog in a CAS system operated with a 3‐h HRT. More recently, Szymanski et al. [13] reported an average removal of 96.8% for a C₁₂E₁₀ in CAS system operated with an HRT of 3.5 h, based upon a bismuth active substance and indirect tensammetric methods. Finally, Battersby et al. [12] reported that >99.9% of C_12–15E₃ and C_12–15E₇ were removed in a CAS system operated with an HRT of 6 h, based upon high‐performance liquid chromatography (HPLC) with fluorescence detection. Similar removals have been reported in the field. Gledhill et al. [24] reported removals of 99% for LAE at an Oklahoma, USA, activated‐sludge plant when using HPLC detection. McAvoy et al. [25] used gas chromatography with mass selective detection to measure total LAE concentration in raw influent and effluent from four activated‐sludge and six trickling‐filter treatment plants. Removals averaged 99.1 ± 0.9% and 93.8 ± 3.8% for activated‐sludge and trickling‐filter plants, respectively, excluding overloaded or upset plants. Notably, removal was inversely correlated with the level of solids in the effluents. Matthijs et al. [26] monitored seven municipal treatment plants in The Netherlands, and found average removals between 99.1 and 99.9% for the various plants.

The total load of parent and metabolites in effluent was likewise predicted by using the k₁ describing the biodegradation of combined parent and metabolites. In this analysis, the metabolites were assumed to be highly soluble and not removed by sorption. Thus, the predicted removal is totally a function of the biodegradation rate (k₁) and HRT. Because the radiolabel is localized in the ethoxylate moieties, this analysis is only relevant for metabolites containing ethoxylate moieties, such as ethylene glycol, PEGs, carboxylated PEGs, carboxyalkyl ethoxylates, and alkyl ethoxy carboxylates. Based upon this prediction, the molar percent of LAE transformed to metabolites escaping treatment ranged from 6 to 17%, depending upon homolog, with C₁₄E¹ having the highest predicted level of metabolites in effluent. However, it should be noted that the major metabolites from this homolog were less polar and would likely be removed by sorption, making this estimate high, because the K_d is assumed to be zero. Nevertheless, there is a strong inverse relationship between the predicted level of metabolites in the effluent and ethoxylate number.

Predicted removal based upon first‐order degradation rates and sorption coefficients can be compared to results from published CAS studies. Battersby et al. [12] reported that 88 to 102% and 83 to 97% of the dissolved carbon added in the form of C_12–15E₃ and C_12–15E₇, respectively, was removed in a CAS system operated with a 6‐h HRT. Szymanski et al. [13] measured the level of long‐chain and short‐chain PEG in effluent from a CAS system fed C₁₂E₁₀ operated with an HRT of 3.5 h and concluded that these measured levels equaled 21% of that theoretically formed from central fission. When using the rates from the current study, the level of metabolite predicted to be in effluent when assuming an HRT of 3.5 h is 23% for C₁₂E₆. In general, total removal of parent and metabolites is difficult to assess from a CAS system unless radiolabeled materials are used because of the complexity of the resulting metabolites and the difficulty in establishing a mass balance. The best study for making such a comparison is that of Steber and Wierich [10], who reported that approximately 24% of the radioactivity from a ¹⁴C‐(ethoxylate) C₁₈E₇ homolog dosed to a CAS system operated with a 3‐h HRT was present in the effluent. Sixteen percent of the radioactivity was in the form of neutral metabolites, presumably PEGs, and 8% was in the form of acidic metabolites, carboxyalkyl ethoxylates [17]. For comparison, when using the data from the current study and assuming an HRT of 3 h, the level of metabolites in effluent predicted for C₁₆E₆ would be 22%.

In conclusion, this study shows that a range of homologs of alcohol ethoxylate undergo rapid primary and ultimate biodegradation in activated sludge. However, the mechanism and the rates for this biodegradation vary as a function of ethoxylate number and alkyl chain length. Furthermore, this study demonstrates that the rates measured in an activated‐sludge die‐away test can be used with a reasonably high level of confidence to predict the level of parent and metabolites escaping wastewater treatment.

Acknowledgements

We are grateful to Kay Marks (Roy F. Weston) for her assistance in conducting this research and for her many years of dedicated service to Procter and Gamble in the field of environmental fate research and testing.

REFERENCES

Copyright © 2004 SETAC

×