Chemosphere 111 (2014) 478–484
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Concentrations of trace substances in sewage sludge from 28 wastewater treatment works in the UK Vera Jones a,⇑, Mike Gardner a, Brian Ellor b a b
Atkins Limited, The Hub, 500 Park Avenue, Aztec West, Bristol BS32 4RZ, UK UKWIR, Queen Anne Gate, London SW1H 9BT, UK
h i g h l i g h t s Sludge samples from 28 UK WwTWs were analysed for >40 trace substances. Concentrations were broadly similar across all sludge samples. Concentrations were generally below regulatory standards for sludge. Predicted concentrations in soil indicated negligible environmental risk.
a r t i c l e
i n f o
Article history: Received 23 December 2013 Received in revised form 8 April 2014 Accepted 9 April 2014
Handling Editor: I. Cousins Keywords: Sewage Sludge Directive Sludge Priority substances Trace substances Chemical Investigations Programme
a b s t r a c t Concentrations of trace substances in sewage sludge have been measured in a survey of 28 wastewater treatment works (WwTWs) in the UK carried out over a period of 12 months. Approximately 250 samples were analysed for more than 40 trace contaminants, including trace metals, pharmaceuticals, polycyclic aromatic hydrocarbons (PAHs), ‘emerging’ and regulated organic pollutants. All substances investigated were found to be present in at least some of the sludges sampled. Concentrations were relatively homogenous across all the WwTWs, irrespective of the treatment process, influent and effluent concentrations, and the location of the sludge sampling point within each works. Analysis of the results against existing regulatory and proposed thresholds suggested that levels are mostly below the limits set in the Sewage Sludge Directive, and proposed new limits for sludge used in agriculture. Predicted soil concentrations after application of sewage sludge to land were below the predicted no effect concentrations (PNEC) for all determinands. Predicted concentrations of pharmaceuticals in soil were also below thresholds deemed to indicate negligible environmental risk. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Sewage sludge is increasingly applied to agricultural land, a practice that is deemed both economically and environmentally advantageous (e.g. Eriksson et al., 2011). Production of sludge is expected to increase from 11.5 M tonnes of dried sludge (2010) to 13 M tonnes of dried sludge by 2020 (Palfrey, 2010). In parallel, raised public awareness of environmental pollution and new environmental regulation (e.g. the ‘Water Framework Directive’ (Official Journal of the European Commission, 2000); the ‘Priority Substances Directive’ (Official Journal of the European Commission, 2008)) have led to an increased interest in the presence of trace substances in wastewaters and effluents (WwTWs; e.g. Rule et al., 2006; Gasperi et al., 2008; Clara et al., 2012). Greater awareness of the presence of ⇑ Corresponding author. Tel.: +44 1925 238558. E-mail address: [email protected] (V. Jones). http://dx.doi.org/10.1016/j.chemosphere.2014.04.025 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
contaminants in wastewater raises the issue of the extent to which various chemicals are present in sewage sludge and consequently might be transferred to land, impacting soil organisms or being transferred up the food chain. As a result, the presence of trace chemicals in sewage sludge has been investigated in several studies over the last decades (see for example, reviews by Eriksson et al., 2008; Fytili and Zabaniotou, 2008; Eriksen et al., 2009; Clarke and Smith, 2011). The current paper is an account of the concentrations of trace substances in sludge, determined as part of the Chemical Investigations Programme (CIP) in the UK. The CIP was a £25 M nationally-coordinated investigation of the risks posed by trace contaminants in the wastewater treatment process in the UK, examining effluent quality, the effectiveness of different treatment processes and sources of substances in the sewer catchment (Gardner et al., 2012; 2013). Determinations of sludge quality were made, as part of the wider CIP process investigations, at 28 WwTW sites.
479
V. Jones et al. / Chemosphere 111 (2014) 478–484
2. Materials and methods WwTWs for process investigation were selected to be representative of different process types applied across UK sites. 14 biological-filter (BF) WwTWs, 12 activated sludge process (ASP) WwTWs, one biological nutrient removal (BNR) WwTW and one membrane bio-reactor (MBR) WwTW were included in the study. Sludge samples were collected at the participating works over a period of approximately 12 months, with 7–15 sampling occasions at each site. Sludge samples were collected at one selected point per WwTW. Samples consisted of primary sludge (collected from the primary settlement tank), secondary/biological sludge (e.g. humus sludge) or mixed sludge (mixture of primary and secondary/ biological sludge) and were analysed for suites of substances including nutrients, metals, emerging and regulated organics, polybrominated diphenyl ethers (BDEs) and pharmaceuticals. The frequency and extent of the sampling meant that results presented here for each parameter are based on approximately 250 samples, sometimes with additional replicates taken on the same date. A smaller number of data points (derived from 6 or 7 WwTWs) were available for a limited number of determinands; these were diclofenac, nonylphenol and its mono, di- and tri- ethoxylates. Where
WwTWs collected replicate sludge samples on the same sampling occasion, concentrations across replicates have been averaged and the average value has been used in all subsequent statistical calculations. All laboratories taking part in this study were accredited to ISO17025 standard for their quality systems. An Aqua Regia digest was applied for the analysis of metals, and organic substances were analysed by LC/MS or GS/MS. All concentrations are reported as mg kg1 dry weight, with the exception of dry solids concentrations which are reported as a percentage. Where results were reported as less than the limit of detection (LOD), the approach was taken of substituting half the reported value (as specified at EU level; Official Journal of the European Commission, 2009).
3. Results 3.1. Trace substances concentrators in sludge Key statistics for all parameters considered are presented in Table 1. All substances included in the analysis suite for this study were detected in at least some of the sludge samples. In the metals group, zinc exhibited the highest concentrations, with a median
Table 1 Summary of sludge data. sd: standard deviations, 25%ile: 25th percentile, 75%ile: 75th percentile CoV: coefficient of variation; WwTW: wastewater treatment works. Mean
Median
sd
25% ile
75% ile
CoV
Number of WwTWs sampled
Dry solids % Nitrogen Phosphorus Potassium
2.6 41733 19898 3313
1.3 38409 17742 2171
2.3 12418 10836 2795
0.9 32941 10491 1574
4.3 53281 27397 4195
0.88 0.30 0.55 0.84
25 28 28 27
Metals Nickel Lead Copper Zinc Cadmium Mercury Silver
29.9 68.9 344 607 0.8 0.7 2.7
25.1 48.3 269 505 0.7 0.7 1.7
19.9 52.4 228 309 0.5 0.2 3.8
19.3 38.1 172 454 0.6 0.6 0.5
32.8 82.6 414 642 0.9 0.8 2.7
0.67 0.76 0.66 0.51 0.61 0.34 1.44
28 28 28 28 28 28 28
BDEs 2,20 ,4,40 -tetrabromodiphenyl ether (PBDE47) 2,20 ,4,40 ,5-pentabromodiphenyl ether (PBDE99) 2,20 ,4,40 ,6-pentabromodiphenyl ether (PBDE100) 2,20 ,4,40 ,5,50 -hexabromodiphenyl ether (PBDE153) 2,20 ,4,40 ,5,60 -hexabromodiphenyl ether (PBDE154)
0.023 0.032 0.007 0.006 0.005
0.021 0.033 0.006 0.005 0.005
0.014 0.019 0.004 0.003 0.002
0.015 0.022 0.005 0.004 0.003
0.031 0.043 0.009 0.008 0.005
0.59 0.60 0.56 0.58 0.53
28 28 28 28 28
‘Emerging’ and regulated organic substances Diethylhexylphthalate (DEHP) Nonylphenol 4-nonylphenol Tributyltin compounds (Tributyltin-cation; TBT) Triclosan Bentazone Bisphenol-A Nonylphenol Monoethoxylate Nonylphenol Diethoxylate Nonylphenol Triethoxylate
19.0 4.4 0.02 4.9 0.07 0.34 5.0 1.1 176.0
11.0 3.8 0.02 4.7 0.02 0.21 6.1 1.1 0.6
20.6 2.9 0.01 3.1 0.06 0.35 2.9 0.6 1.7
3.0 2.3 0.01 2.1 0.02 0.12 2.7 1.0 0.3
30.8 5.8 0.02 7.0 0.11 0.56 7.1 1.3 0.9
1.08 0.67 0.70 0.64 0.95 1.05 0.59 0.51 1.41
28 28 28 28 28 28 6 6 6
PAHs Anthracene Fluoranthene Naphthalene Benzo(a)pyrene Benzo(b)fluoranthene Benzo(k)fluoranthene Benzo(g,h,i)perylene Indeno(1,2,3-cd)pyrene
0.11 0.70 0.43 0.38 0.38 0.29 0.33 0.32
0.10 0.52 0.21 0.32 0.31 0.25 0.27 0.28
0.07 0.51 0.73 0.20 0.21 0.21 0.22 0.18
0.07 0.41 0.10 0.19 0.25 0.13 0.16 0.18
0.13 0.77 0.40 0.50 0.46 0.38 0.44 0.42
0.66 0.73 1.71 0.52 0.54 0.74 0.66 0.58
28 28 28 28 28 28 28 28
Pharmaceuticals Diclofenac Ibuprofen Propranolol Erythromycin Ofloxacin Oxytetracycline Fluoxetine
0.06 0.27 0.14 0.06 0.22 7.63 0.13
0.07 0.22 0.12 0.05 0.20 4.00 0.12
0.03 0.19 0.08 0.04 0.12 9.25 0.05
0.05 0.12 0.10 0.03 0.14 2.65 0.09
0.07 0.39 0.18 0.06 0.27 8.66 0.18
0.41 0.69 0.54 0.63 0.56 1.21 0.42
7 28 28 28 28 28 28
480
V. Jones et al. / Chemosphere 111 (2014) 478–484
value of 505 mg kg1, followed by copper (median: 269 mg kg1). Concentrations of nickel and lead were an order of magnitude lower (median: 25.1 and 48.3 mg kg1, respectively). The median silver concentration was 1.7 mg kg1, while cadmium and mercury concentrations were a further order of magnitude lower (both with a median of: 0.7 mg kg1). BDE concentrations were in many cases below the LOD, which was either 0.01 or 0.002 mg kg1, depending on the analytical laboratory. Owing to the large discrepancies in LODs and prevalence of less than results, the results for BDE 28 have been excluded from further analysis. Within the remaining BDEs, the highest concentrations were reported for BDE 99 (median: 0.033 mg kg1), followed by BDE 47 (median: 0.021 mg kg1). Concentrations of BDE 100, 153 and 154 were an order of magnitude lower, ranging from 0.005 mg kg1 to 0.006 mg kg1 (median values). Amongst the PAHs, the highest concentrations were reported for fluoranthene (median: 0.52 mg kg1). Concentrations were in the same order of magnitude, but slightly lower, for all other PAHs measured, ranging from a median value of 0.32 mg kg1 for benzo(a)pyrene to 0.10 mg kg1 for anthracene. The highest concentrations within the pharmaceuticals group were noted for oxytetracycline, with a median value of 4.00 mg kg1 across all works. Concentrations for the majority of other pharmaceuticals were an order of magnitude lower, with a median ibuprofen concentration of 0.22 mg kg1 and ofloxacin concentration of 0.20 mg kg1. Erythromycin and diclofenac concentrations were a further order of magnitude lower, with median concentrations of 0.05 mg kg1 and 0.07 mg kg1, respectively. A number of other emerging and regulated organic substances were considered in this study. Concentrations of the plasticiser diethylhexylphthalate (DEHP) were highest amongst this group, with a median concentration of 11.0 mg kg1. Levels of nonylphenol, a surfactant derivative with endocrine disruptor properties, exhibited a median concentration of 3.8 mg kg1. The median concentration of the surfactants nonylphenol ethoxylates (only measured at 6 WwTWs) was highest for nonylphenol monoethoxylate (6.1 mg kg1), followed by di-ethoxylate (1.1 mg kg1) and tri-ethoxylate (0.6 mg kg1). The anti-microbial agent triclosan also exhibited high concentrations in the CIP sludge samples, relative to other parameters in the emerging and regulated substances group, with a median concentration of 4.7 mg kg1. The biocide tributyl tin (TBT) was found at variable concentrations across works, with a median concentration of 0.02 mg kg1 and a coefficient of variation (CoV) of 0.70. 3.2. Do trace substances concentrations vary significantly between different types of sludge? Sludge samples were divided into three groups: those collected at a primary sludge, secondary/biological sludge or mixed sludge sampling point (Fig. 1). Visual analysis of data graphs was then undertaken, and, where this suggested noticeable differences between sludge sample groups, it was followed by single-factor analysis of variance (ANOVA). For the majority of trace substances, there were no major differences in concentrations between the three different sludge types sampled (see for example, copper plot in Fig. 1). There were some exceptions to this observation. Triclosan, propanolol, ibuprofen, erythromycin and nonylphenol concentrations differed to a marked extent between the three sludge groups (p < 0.05), with levels being generally higher in the primary and mixed sludge samples than in the secondary/biological samples (Fig. 1). We also examined of the relative hydrophobicity of these substances (using log kow as an indicator of hydrophobicity). This analysis suggested that log kow is not a good indicator for the presence of trace substances in different types of sludge, which is likely to be affected more significantly by mechanisms such as biodegradation.
3.3. Do trace substance concentrations vary significantly between different types of wastewater treatment plants? Sludge data were then divided into two broad groups for this analysis: those collected at WwTWs operating BF and those employing ASP. Data for a BNR and a MBR WwTW were not included in this analysis, as they represented single data points. Results for the BF and ASP groups were checked for apparent differences visually on data plots and then by single-factor ANOVA. The results only suggested significant differences (p < 0.05) in concentrations between ASP and BF plants for BDE 47, with BDE 47 concentrations being higher in sludge collected at BF plants compared to those collected at ASP plants (Fig. 2). Results for all other parameters did not show any significant difference between ASP and BF plants, suggesting that the type of treatment process employed at different WwTWs does not have an overall significant effect on the concentrations of trace chemicals in sludge. 3.4. Concentrations in sludge versus influent to effluent fractional removal within WwTWs The relationship between median fractional removal (influent to effluent) at each WwTWs and each substance (Gardner et al., 2013) and median concentration in sludge was examined by regression analysis. The aim was to assess whether, for any of the chemicals considered, works with higher influent to effluent removal also exhibited higher concentrations in sludge. The results of this analysis did not indicate any statistically significant relationships between fractional removals and concentration in sludge for any of the substances considered. 3.5. Does influent and effluent quality affect sludge concentrations? In a similar way to above, the relationship between influent/ effluent concentration and concentration in sludge was investigated by means of regression. Mean sludge concentrations were plotted against mean influent and then effluent concentrations for each WwTW (Gardner et al., 2013). This indicated the presence of a potential link between these two variables for three metals only (copper, lead and nickel), although the relationship was not statistically significant. For all other determinands, levels in the influent or effluent did not appear to be related in any statistically discernible way to concentrations in sludge. Metals are the most conservative of the trace substances considered in this study. It is, therefore, not surprising that there is a link between influent/ effluent and sludge concentrations for these determinands, which are not significantly affected by degradation processes during the transfer of a substance from wastewater to sludge. 4. Discussion 4.1. Homogeneity in trace substances concentrations across WwTW The analysis described above indicates a broad homogeneity in the trace substance concentrations across different sludge samples considered. To confirm this further, the between-WwTWs coefficient of variation (CoV) for influent and then effluent concentrations was compared to the CoV for sludge concentrations for each substance. This comparison indicated that the CoV (and hence between-works variability) was greater in influent concentrations than in sludge concentrations for 25 out of the 37 determinands considered in both effluent and in sludge. The same comparison for effluent indicated that the between-WwTWs CoV was higher for effluent concentrations than for sludge for 26 out of the 37 determinands considered in both effluent and in sludge. This suggests
V. Jones et al. / Chemosphere 111 (2014) 478–484
481
Fig. 1. Concentration of selected determinands across all WwTW sampled, grouped by type of sludge sampled.
Fig. 2. Concentration of BDE47 across all WwTW sampled, grouped by treatment process employed at the WwTW. Data for the membrane bioreactor (MBR) and biological nutrient removal (BNR) works sampled are also shown on the plot for information, although these data were not included in the ANOVA described in the main text.
482
V. Jones et al. / Chemosphere 111 (2014) 478–484
then that, although the quality of influent (and effluent discharged) is relatively variable between WwTWs, this variability is greatly decreased when it comes to sludge. It is possible that this is the result of the long residence time that sludge exhibits, which means it is an overall well-mixed and highly-degraded entity. It may also indicate that a number of factors affect sludge concentrations without one factor being the predominant control. The overall effect is broadly consistent concentrations across the considerable number of samples analysed for this study. It is also worth noting that sludge levels appear broadly homogenous, despite the fact that samples were analysed by several different laboratories for this study. 4.2. Comparison with trace substances concentrations in sludge reported in literature In order to help set the results of this study into context, a comparison with selected previous studies was undertaken. We refer below to two reviews of trace substances in sludge, which covered
some of the substances considered here: a European Commission study (EC, 2001 and references therein) reviewing data from several European countries and the US, and the US EPA Targeted National Sewage Sludge Survey (EPA, 2009). For some of the contaminants reference is made to other smaller-scale studies. Nickel, lead, copper, zinc, cadmium and mercury concentrations reported here are in the same order of magnitude as those reported in EC (2001; e.g. copper concentrations ranging from 103 to 641 mg kg1). In contrast, CIP lead concentrations are generally an order of magnitude lower than those reported in the EC (2001) study (22–283 mg kg1). Similarly, the CIP silver concentrations are an order of magnitude lower than the silver concentration quoted in the EC report (2001; 48.2 mg kg1, derived from a US study) and the EPA survey (2009; median of 13.50 mg kg1). BDE levels reported for the CIP samples are generally lower than BDE levels reported by the EPA (2009; e.g. median BDE47 concentration: 0.575 mg kg1 and median BDE99 concentration: 0.610 mg kg1), but broadly in agreement with concentrations
Fig. 3. Estimated concentrations of trace metals, organic substances and PAHs in soil after application of sludge, based on the highest average sludge concentration recorded amongst the wastewater treatment works in this study and a conversion as in UKWIR (2013), and relevant soil predicted no-effect concentrations (PNEC) based on review by Eriksen et al. (2009) for BDE99 and nonyl-phenol (NOP) and as in Amorim et al. (2010) for triclosan (TRICL). Note the log scale on the x-axis. NIT: total nickel; PBT: total lead, CUT: total copper; ZNT: total zinc; CDT: total cadmium; NOP: nonylphenol; TRICL: triclosan; ANT: anthracene; FLU: fluoranthene; NAP: naphthalene; BAP: benzo(a)pyrene; BBF: benzo(b)fluoranthene; BBK: benzo(k)fluoranthene; BGHIP: benzo(g,h,i)perylene.
483
V. Jones et al. / Chemosphere 111 (2014) 478–484
reported in Eriksson et al. (2011) based on analysis of treated sludge from 4 European cities (PBDE concentrations:
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Concentrations of trace substances in sewage sludge from 28 wastewater treatment works in the UK Vera Jones a,⇑, Mike Gardner a, Brian Ellor b a b
Atkins Limited, The Hub, 500 Park Avenue, Aztec West, Bristol BS32 4RZ, UK UKWIR, Queen Anne Gate, London SW1H 9BT, UK
h i g h l i g h t s Sludge samples from 28 UK WwTWs were analysed for >40 trace substances. Concentrations were broadly similar across all sludge samples. Concentrations were generally below regulatory standards for sludge. Predicted concentrations in soil indicated negligible environmental risk.
a r t i c l e
i n f o
Article history: Received 23 December 2013 Received in revised form 8 April 2014 Accepted 9 April 2014
Handling Editor: I. Cousins Keywords: Sewage Sludge Directive Sludge Priority substances Trace substances Chemical Investigations Programme
a b s t r a c t Concentrations of trace substances in sewage sludge have been measured in a survey of 28 wastewater treatment works (WwTWs) in the UK carried out over a period of 12 months. Approximately 250 samples were analysed for more than 40 trace contaminants, including trace metals, pharmaceuticals, polycyclic aromatic hydrocarbons (PAHs), ‘emerging’ and regulated organic pollutants. All substances investigated were found to be present in at least some of the sludges sampled. Concentrations were relatively homogenous across all the WwTWs, irrespective of the treatment process, influent and effluent concentrations, and the location of the sludge sampling point within each works. Analysis of the results against existing regulatory and proposed thresholds suggested that levels are mostly below the limits set in the Sewage Sludge Directive, and proposed new limits for sludge used in agriculture. Predicted soil concentrations after application of sewage sludge to land were below the predicted no effect concentrations (PNEC) for all determinands. Predicted concentrations of pharmaceuticals in soil were also below thresholds deemed to indicate negligible environmental risk. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Sewage sludge is increasingly applied to agricultural land, a practice that is deemed both economically and environmentally advantageous (e.g. Eriksson et al., 2011). Production of sludge is expected to increase from 11.5 M tonnes of dried sludge (2010) to 13 M tonnes of dried sludge by 2020 (Palfrey, 2010). In parallel, raised public awareness of environmental pollution and new environmental regulation (e.g. the ‘Water Framework Directive’ (Official Journal of the European Commission, 2000); the ‘Priority Substances Directive’ (Official Journal of the European Commission, 2008)) have led to an increased interest in the presence of trace substances in wastewaters and effluents (WwTWs; e.g. Rule et al., 2006; Gasperi et al., 2008; Clara et al., 2012). Greater awareness of the presence of ⇑ Corresponding author. Tel.: +44 1925 238558. E-mail address: [email protected] (V. Jones). http://dx.doi.org/10.1016/j.chemosphere.2014.04.025 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
contaminants in wastewater raises the issue of the extent to which various chemicals are present in sewage sludge and consequently might be transferred to land, impacting soil organisms or being transferred up the food chain. As a result, the presence of trace chemicals in sewage sludge has been investigated in several studies over the last decades (see for example, reviews by Eriksson et al., 2008; Fytili and Zabaniotou, 2008; Eriksen et al., 2009; Clarke and Smith, 2011). The current paper is an account of the concentrations of trace substances in sludge, determined as part of the Chemical Investigations Programme (CIP) in the UK. The CIP was a £25 M nationally-coordinated investigation of the risks posed by trace contaminants in the wastewater treatment process in the UK, examining effluent quality, the effectiveness of different treatment processes and sources of substances in the sewer catchment (Gardner et al., 2012; 2013). Determinations of sludge quality were made, as part of the wider CIP process investigations, at 28 WwTW sites.
479
V. Jones et al. / Chemosphere 111 (2014) 478–484
2. Materials and methods WwTWs for process investigation were selected to be representative of different process types applied across UK sites. 14 biological-filter (BF) WwTWs, 12 activated sludge process (ASP) WwTWs, one biological nutrient removal (BNR) WwTW and one membrane bio-reactor (MBR) WwTW were included in the study. Sludge samples were collected at the participating works over a period of approximately 12 months, with 7–15 sampling occasions at each site. Sludge samples were collected at one selected point per WwTW. Samples consisted of primary sludge (collected from the primary settlement tank), secondary/biological sludge (e.g. humus sludge) or mixed sludge (mixture of primary and secondary/ biological sludge) and were analysed for suites of substances including nutrients, metals, emerging and regulated organics, polybrominated diphenyl ethers (BDEs) and pharmaceuticals. The frequency and extent of the sampling meant that results presented here for each parameter are based on approximately 250 samples, sometimes with additional replicates taken on the same date. A smaller number of data points (derived from 6 or 7 WwTWs) were available for a limited number of determinands; these were diclofenac, nonylphenol and its mono, di- and tri- ethoxylates. Where
WwTWs collected replicate sludge samples on the same sampling occasion, concentrations across replicates have been averaged and the average value has been used in all subsequent statistical calculations. All laboratories taking part in this study were accredited to ISO17025 standard for their quality systems. An Aqua Regia digest was applied for the analysis of metals, and organic substances were analysed by LC/MS or GS/MS. All concentrations are reported as mg kg1 dry weight, with the exception of dry solids concentrations which are reported as a percentage. Where results were reported as less than the limit of detection (LOD), the approach was taken of substituting half the reported value (as specified at EU level; Official Journal of the European Commission, 2009).
3. Results 3.1. Trace substances concentrators in sludge Key statistics for all parameters considered are presented in Table 1. All substances included in the analysis suite for this study were detected in at least some of the sludge samples. In the metals group, zinc exhibited the highest concentrations, with a median
Table 1 Summary of sludge data. sd: standard deviations, 25%ile: 25th percentile, 75%ile: 75th percentile CoV: coefficient of variation; WwTW: wastewater treatment works. Mean
Median
sd
25% ile
75% ile
CoV
Number of WwTWs sampled
Dry solids % Nitrogen Phosphorus Potassium
2.6 41733 19898 3313
1.3 38409 17742 2171
2.3 12418 10836 2795
0.9 32941 10491 1574
4.3 53281 27397 4195
0.88 0.30 0.55 0.84
25 28 28 27
Metals Nickel Lead Copper Zinc Cadmium Mercury Silver
29.9 68.9 344 607 0.8 0.7 2.7
25.1 48.3 269 505 0.7 0.7 1.7
19.9 52.4 228 309 0.5 0.2 3.8
19.3 38.1 172 454 0.6 0.6 0.5
32.8 82.6 414 642 0.9 0.8 2.7
0.67 0.76 0.66 0.51 0.61 0.34 1.44
28 28 28 28 28 28 28
BDEs 2,20 ,4,40 -tetrabromodiphenyl ether (PBDE47) 2,20 ,4,40 ,5-pentabromodiphenyl ether (PBDE99) 2,20 ,4,40 ,6-pentabromodiphenyl ether (PBDE100) 2,20 ,4,40 ,5,50 -hexabromodiphenyl ether (PBDE153) 2,20 ,4,40 ,5,60 -hexabromodiphenyl ether (PBDE154)
0.023 0.032 0.007 0.006 0.005
0.021 0.033 0.006 0.005 0.005
0.014 0.019 0.004 0.003 0.002
0.015 0.022 0.005 0.004 0.003
0.031 0.043 0.009 0.008 0.005
0.59 0.60 0.56 0.58 0.53
28 28 28 28 28
‘Emerging’ and regulated organic substances Diethylhexylphthalate (DEHP) Nonylphenol 4-nonylphenol Tributyltin compounds (Tributyltin-cation; TBT) Triclosan Bentazone Bisphenol-A Nonylphenol Monoethoxylate Nonylphenol Diethoxylate Nonylphenol Triethoxylate
19.0 4.4 0.02 4.9 0.07 0.34 5.0 1.1 176.0
11.0 3.8 0.02 4.7 0.02 0.21 6.1 1.1 0.6
20.6 2.9 0.01 3.1 0.06 0.35 2.9 0.6 1.7
3.0 2.3 0.01 2.1 0.02 0.12 2.7 1.0 0.3
30.8 5.8 0.02 7.0 0.11 0.56 7.1 1.3 0.9
1.08 0.67 0.70 0.64 0.95 1.05 0.59 0.51 1.41
28 28 28 28 28 28 6 6 6
PAHs Anthracene Fluoranthene Naphthalene Benzo(a)pyrene Benzo(b)fluoranthene Benzo(k)fluoranthene Benzo(g,h,i)perylene Indeno(1,2,3-cd)pyrene
0.11 0.70 0.43 0.38 0.38 0.29 0.33 0.32
0.10 0.52 0.21 0.32 0.31 0.25 0.27 0.28
0.07 0.51 0.73 0.20 0.21 0.21 0.22 0.18
0.07 0.41 0.10 0.19 0.25 0.13 0.16 0.18
0.13 0.77 0.40 0.50 0.46 0.38 0.44 0.42
0.66 0.73 1.71 0.52 0.54 0.74 0.66 0.58
28 28 28 28 28 28 28 28
Pharmaceuticals Diclofenac Ibuprofen Propranolol Erythromycin Ofloxacin Oxytetracycline Fluoxetine
0.06 0.27 0.14 0.06 0.22 7.63 0.13
0.07 0.22 0.12 0.05 0.20 4.00 0.12
0.03 0.19 0.08 0.04 0.12 9.25 0.05
0.05 0.12 0.10 0.03 0.14 2.65 0.09
0.07 0.39 0.18 0.06 0.27 8.66 0.18
0.41 0.69 0.54 0.63 0.56 1.21 0.42
7 28 28 28 28 28 28
480
V. Jones et al. / Chemosphere 111 (2014) 478–484
value of 505 mg kg1, followed by copper (median: 269 mg kg1). Concentrations of nickel and lead were an order of magnitude lower (median: 25.1 and 48.3 mg kg1, respectively). The median silver concentration was 1.7 mg kg1, while cadmium and mercury concentrations were a further order of magnitude lower (both with a median of: 0.7 mg kg1). BDE concentrations were in many cases below the LOD, which was either 0.01 or 0.002 mg kg1, depending on the analytical laboratory. Owing to the large discrepancies in LODs and prevalence of less than results, the results for BDE 28 have been excluded from further analysis. Within the remaining BDEs, the highest concentrations were reported for BDE 99 (median: 0.033 mg kg1), followed by BDE 47 (median: 0.021 mg kg1). Concentrations of BDE 100, 153 and 154 were an order of magnitude lower, ranging from 0.005 mg kg1 to 0.006 mg kg1 (median values). Amongst the PAHs, the highest concentrations were reported for fluoranthene (median: 0.52 mg kg1). Concentrations were in the same order of magnitude, but slightly lower, for all other PAHs measured, ranging from a median value of 0.32 mg kg1 for benzo(a)pyrene to 0.10 mg kg1 for anthracene. The highest concentrations within the pharmaceuticals group were noted for oxytetracycline, with a median value of 4.00 mg kg1 across all works. Concentrations for the majority of other pharmaceuticals were an order of magnitude lower, with a median ibuprofen concentration of 0.22 mg kg1 and ofloxacin concentration of 0.20 mg kg1. Erythromycin and diclofenac concentrations were a further order of magnitude lower, with median concentrations of 0.05 mg kg1 and 0.07 mg kg1, respectively. A number of other emerging and regulated organic substances were considered in this study. Concentrations of the plasticiser diethylhexylphthalate (DEHP) were highest amongst this group, with a median concentration of 11.0 mg kg1. Levels of nonylphenol, a surfactant derivative with endocrine disruptor properties, exhibited a median concentration of 3.8 mg kg1. The median concentration of the surfactants nonylphenol ethoxylates (only measured at 6 WwTWs) was highest for nonylphenol monoethoxylate (6.1 mg kg1), followed by di-ethoxylate (1.1 mg kg1) and tri-ethoxylate (0.6 mg kg1). The anti-microbial agent triclosan also exhibited high concentrations in the CIP sludge samples, relative to other parameters in the emerging and regulated substances group, with a median concentration of 4.7 mg kg1. The biocide tributyl tin (TBT) was found at variable concentrations across works, with a median concentration of 0.02 mg kg1 and a coefficient of variation (CoV) of 0.70. 3.2. Do trace substances concentrations vary significantly between different types of sludge? Sludge samples were divided into three groups: those collected at a primary sludge, secondary/biological sludge or mixed sludge sampling point (Fig. 1). Visual analysis of data graphs was then undertaken, and, where this suggested noticeable differences between sludge sample groups, it was followed by single-factor analysis of variance (ANOVA). For the majority of trace substances, there were no major differences in concentrations between the three different sludge types sampled (see for example, copper plot in Fig. 1). There were some exceptions to this observation. Triclosan, propanolol, ibuprofen, erythromycin and nonylphenol concentrations differed to a marked extent between the three sludge groups (p < 0.05), with levels being generally higher in the primary and mixed sludge samples than in the secondary/biological samples (Fig. 1). We also examined of the relative hydrophobicity of these substances (using log kow as an indicator of hydrophobicity). This analysis suggested that log kow is not a good indicator for the presence of trace substances in different types of sludge, which is likely to be affected more significantly by mechanisms such as biodegradation.
3.3. Do trace substance concentrations vary significantly between different types of wastewater treatment plants? Sludge data were then divided into two broad groups for this analysis: those collected at WwTWs operating BF and those employing ASP. Data for a BNR and a MBR WwTW were not included in this analysis, as they represented single data points. Results for the BF and ASP groups were checked for apparent differences visually on data plots and then by single-factor ANOVA. The results only suggested significant differences (p < 0.05) in concentrations between ASP and BF plants for BDE 47, with BDE 47 concentrations being higher in sludge collected at BF plants compared to those collected at ASP plants (Fig. 2). Results for all other parameters did not show any significant difference between ASP and BF plants, suggesting that the type of treatment process employed at different WwTWs does not have an overall significant effect on the concentrations of trace chemicals in sludge. 3.4. Concentrations in sludge versus influent to effluent fractional removal within WwTWs The relationship between median fractional removal (influent to effluent) at each WwTWs and each substance (Gardner et al., 2013) and median concentration in sludge was examined by regression analysis. The aim was to assess whether, for any of the chemicals considered, works with higher influent to effluent removal also exhibited higher concentrations in sludge. The results of this analysis did not indicate any statistically significant relationships between fractional removals and concentration in sludge for any of the substances considered. 3.5. Does influent and effluent quality affect sludge concentrations? In a similar way to above, the relationship between influent/ effluent concentration and concentration in sludge was investigated by means of regression. Mean sludge concentrations were plotted against mean influent and then effluent concentrations for each WwTW (Gardner et al., 2013). This indicated the presence of a potential link between these two variables for three metals only (copper, lead and nickel), although the relationship was not statistically significant. For all other determinands, levels in the influent or effluent did not appear to be related in any statistically discernible way to concentrations in sludge. Metals are the most conservative of the trace substances considered in this study. It is, therefore, not surprising that there is a link between influent/ effluent and sludge concentrations for these determinands, which are not significantly affected by degradation processes during the transfer of a substance from wastewater to sludge. 4. Discussion 4.1. Homogeneity in trace substances concentrations across WwTW The analysis described above indicates a broad homogeneity in the trace substance concentrations across different sludge samples considered. To confirm this further, the between-WwTWs coefficient of variation (CoV) for influent and then effluent concentrations was compared to the CoV for sludge concentrations for each substance. This comparison indicated that the CoV (and hence between-works variability) was greater in influent concentrations than in sludge concentrations for 25 out of the 37 determinands considered in both effluent and in sludge. The same comparison for effluent indicated that the between-WwTWs CoV was higher for effluent concentrations than for sludge for 26 out of the 37 determinands considered in both effluent and in sludge. This suggests
V. Jones et al. / Chemosphere 111 (2014) 478–484
481
Fig. 1. Concentration of selected determinands across all WwTW sampled, grouped by type of sludge sampled.
Fig. 2. Concentration of BDE47 across all WwTW sampled, grouped by treatment process employed at the WwTW. Data for the membrane bioreactor (MBR) and biological nutrient removal (BNR) works sampled are also shown on the plot for information, although these data were not included in the ANOVA described in the main text.
482
V. Jones et al. / Chemosphere 111 (2014) 478–484
then that, although the quality of influent (and effluent discharged) is relatively variable between WwTWs, this variability is greatly decreased when it comes to sludge. It is possible that this is the result of the long residence time that sludge exhibits, which means it is an overall well-mixed and highly-degraded entity. It may also indicate that a number of factors affect sludge concentrations without one factor being the predominant control. The overall effect is broadly consistent concentrations across the considerable number of samples analysed for this study. It is also worth noting that sludge levels appear broadly homogenous, despite the fact that samples were analysed by several different laboratories for this study. 4.2. Comparison with trace substances concentrations in sludge reported in literature In order to help set the results of this study into context, a comparison with selected previous studies was undertaken. We refer below to two reviews of trace substances in sludge, which covered
some of the substances considered here: a European Commission study (EC, 2001 and references therein) reviewing data from several European countries and the US, and the US EPA Targeted National Sewage Sludge Survey (EPA, 2009). For some of the contaminants reference is made to other smaller-scale studies. Nickel, lead, copper, zinc, cadmium and mercury concentrations reported here are in the same order of magnitude as those reported in EC (2001; e.g. copper concentrations ranging from 103 to 641 mg kg1). In contrast, CIP lead concentrations are generally an order of magnitude lower than those reported in the EC (2001) study (22–283 mg kg1). Similarly, the CIP silver concentrations are an order of magnitude lower than the silver concentration quoted in the EC report (2001; 48.2 mg kg1, derived from a US study) and the EPA survey (2009; median of 13.50 mg kg1). BDE levels reported for the CIP samples are generally lower than BDE levels reported by the EPA (2009; e.g. median BDE47 concentration: 0.575 mg kg1 and median BDE99 concentration: 0.610 mg kg1), but broadly in agreement with concentrations
Fig. 3. Estimated concentrations of trace metals, organic substances and PAHs in soil after application of sludge, based on the highest average sludge concentration recorded amongst the wastewater treatment works in this study and a conversion as in UKWIR (2013), and relevant soil predicted no-effect concentrations (PNEC) based on review by Eriksen et al. (2009) for BDE99 and nonyl-phenol (NOP) and as in Amorim et al. (2010) for triclosan (TRICL). Note the log scale on the x-axis. NIT: total nickel; PBT: total lead, CUT: total copper; ZNT: total zinc; CDT: total cadmium; NOP: nonylphenol; TRICL: triclosan; ANT: anthracene; FLU: fluoranthene; NAP: naphthalene; BAP: benzo(a)pyrene; BBF: benzo(b)fluoranthene; BBK: benzo(k)fluoranthene; BGHIP: benzo(g,h,i)perylene.
483
V. Jones et al. / Chemosphere 111 (2014) 478–484
reported in Eriksson et al. (2011) based on analysis of treated sludge from 4 European cities (PBDE concentrations:
