Abstract

Three types of odour abatement systems in sewer networks in Australia were studied for 18 months to determine the removals of different compounds. Six volatile sulfurous compounds and seven volatile organic compounds (VOCs) were further investigated. All types of odour abatement systems exhibited good removal of hydrogen sulfide with the biotrickling filters (BTFs) showing the highest consistent removal. Biofilters outperformed BTFs and activated carbon (AC) filters in the removal of dimethyl mono-, di- and tri-sulfide species at the low inlet concentrations typically found. AC filters exhibited little VOC removal with no compound consistently identified as having a removal greater than 0%. Biofilters outperformed BTFs in VOC removal, yet both had high removal variability.

INTRODUCTION

Community tolerance for odours from sewer infrastructure is lessening, providing pressure to water utilities to target and treat odour emissions from sewer networks (Mudliar et al. 2010). Gas phase odour abatement systems within sewer networks typically target hydrogen sulfide (H2S) as an odorant and overall odour indicator. However, other compounds can occur naturally or through trade waste in sewer headspaces, which are either odorous themselves or may impact on the performance of odorant removal in odour abatement systems (Sivret et al. 2013; Wang et al. 2014; Sivret et al. 2017).

Odour abatement systems in sewer networks are predominantly biological or physical adsorption based (Sivret & Stuetz 2010; Shammay et al. 2016a), although a range of other approaches are also provided (Koe & Tan 1990; Davidson et al. 2004; Apgar et al. 2007). Performance studies of odour abatement systems removing odorous compounds at different operating conditions are typically fairly limited in duration and in scope, in that non-H2S compounds are measured either once only or infrequently (Shammay et al. 2016a). This is a result of non-H2S compounds being difficult and costly to measure (Muñoz et al. 2010; Wang et al. 2014). What studies there are show that H2S is generally removed well under most conditions; however, the removal of other odorous compounds is variable (Gostelow et al. 2001; Iranpour et al. 2005; Wang et al. 2014; Shammay et al. 2016a; Shammay et al. 2016b). A summary of removal efficiencies of different compounds relevant to this study found in the literature, adapted from Shammay et al. (2016a), is provided in Table 1.

Table 1

Summary of performance of gas phase odour abatement systems

  Compound Biofilter
 
Biotrickling filter
 
Activated carbon
 
Removal (%) Consistency Removal (%) Consistency Removal (%) Consistency 
Volatile sulfurous compounds Hydrogen sulfide 83 to >99 Good 80 to >99 Good 79 to 100 (22 outlier) Good 
Carbon disulfide 13 to 100 Poor 0 to 40 Poor −∞ to 41 Poor 
Methyl mercaptan 62 to 100 Medium 56 to 78 (−638 outlier) Good 77 to 100 (52 outlier) Good 
Dimethyl sulfide 0 to 100 Poor Poor (3 samples from 1 study) −75 to 93 Poor 
Dimethyl disulfide 97 to >99 (17 outlier) Medium nd – −2,254 to 100 Poor 
Dimethyl trisulfide 51.5 to 100 Poor (6 samples in 1 study) nd – −545 to 100 Poor (19 samples in 1 study with big spread) 
Volatile organic compounds Benzene 25 to 100 Medium 0 to 59 (Generally 20 to 30) Medium 86 to 99 Good 
Xylenes −175 to 100 Poor 0 to 57 (92 is outlier) Medium 88 to 100 (outlier study to as low as −7,713) Medium 
Toluene 43 to 100 Medium 0 to 85 Medium 90 to 100 (2 outliers as low as −852) Medium 
1,3,5-Trimethylbenzene nd – nd – −∞ to 100 Poor 
Ethylbenzene 84 to 88 Poor 21 Poor −∞ to 100 Medium 
Dichlorobenzene −18 to 100 Medium 0 to 48 Medium −47 to 100 Medium 
  Compound Biofilter
 
Biotrickling filter
 
Activated carbon
 
Removal (%) Consistency Removal (%) Consistency Removal (%) Consistency 
Volatile sulfurous compounds Hydrogen sulfide 83 to >99 Good 80 to >99 Good 79 to 100 (22 outlier) Good 
Carbon disulfide 13 to 100 Poor 0 to 40 Poor −∞ to 41 Poor 
Methyl mercaptan 62 to 100 Medium 56 to 78 (−638 outlier) Good 77 to 100 (52 outlier) Good 
Dimethyl sulfide 0 to 100 Poor Poor (3 samples from 1 study) −75 to 93 Poor 
Dimethyl disulfide 97 to >99 (17 outlier) Medium nd – −2,254 to 100 Poor 
Dimethyl trisulfide 51.5 to 100 Poor (6 samples in 1 study) nd – −545 to 100 Poor (19 samples in 1 study with big spread) 
Volatile organic compounds Benzene 25 to 100 Medium 0 to 59 (Generally 20 to 30) Medium 86 to 99 Good 
Xylenes −175 to 100 Poor 0 to 57 (92 is outlier) Medium 88 to 100 (outlier study to as low as −7,713) Medium 
Toluene 43 to 100 Medium 0 to 85 Medium 90 to 100 (2 outliers as low as −852) Medium 
1,3,5-Trimethylbenzene nd – nd – −∞ to 100 Poor 
Ethylbenzene 84 to 88 Poor 21 Poor −∞ to 100 Medium 
Dichlorobenzene −18 to 100 Medium 0 to 48 Medium −47 to 100 Medium 

nd – Not detected.

Good – Multiple studies, greater than 5, with 75% or more of studies showing the same outcome.

Medium – Multiple studies, greater than 3, showing different outcomes or few studies, less than 3, showing the same outcome.

Poor – Few studies, generally less than 3, showing different outcomes.

Negative values given where outlet concentration was greater than inlet concentration.

−∞ – Outlet concentration was detected with none detected on the inlet.

Refer to Supplementary Material for references and inlet concentrations (available with the online version of this paper).

Gaseous treatment methods generally remove H2S well with removals consistently above 80% for biological systems and consistently above 79% for activated carbon (AC) systems. AC systems exhibit consistent methyl mercaptan removals above 72%. All other volatile sulfur compound (VSC) and volatile organic compound (VOC) removals, particularly with halogenated VOC removal, are inconsistent and generally low depending on the technology type. With H2S being identified as less of a universal marker for odour (Sivret et al. 2016), the removal of other odorants is becoming more important.

AC systems tend to be more cost effective than other technologies in treating foul air with low H2S concentrations. They also have a shorter start up time than biological alternatives and can treat sporadic flows where biological systems tend to only partially treat peak concentrations. However, AC systems release contaminants more often than biological systems and suffer from high operating costs due to the need for carbon changeout. Biofilters are appropriate for relatively low stable loads and are cost effective where space is available. There are many long-term studies showing the effectiveness of biofilters; however, removal of VOCs is generally low and inconsistent. Biotrickling filters (BTFs) provide a better level of control over removal rates than biofilters, particularly with intermittent peak concentrations, and are more cost effective for higher H2S concentrations than AC systems (Shammay et al. 2016a).

METHODOLOGY

Site descriptions

Duplicate samples were taken from the inlet and outlet of 18 odour abatement systems in sewer networks (approximately 19 samples per site) or at the headworks of treatment plants in Sydney, Melbourne and Perth, Australia. In general, sites were selected so that they were not downstream of each other, in different catchments (where possible), contained mixed sewage (from both domestic and trade waste), and had a range of upstream chemical dosing.

Sites selected employed AC, BTFs or biofilters (BFs) as the most common forms of odour abatement in sewer networks (Sivret & Stuetz 2010). Details of sites investigated, including their empty bed residence times (EBRTs), are provided in Table 2.

Table 2

Sites investigated for AC, BTFs and biofilters (BFs) performance

Site Year installed Note 1 City Media type Upstream chemical dosing Type Approx. EBRT Volume Species investigated after data filtering 
AC-1 2003 Sydney AC – Sodium hydroxide (NaOH) Impregnated Ferrous chloride Closed-topped 18.0 s 72 m3 VOCs only 
AC-2 2008 Sydney AC – 50% Caustic impregnated/50% Virgin None Closed-topped 9.5 s 9.5 m3 VOCs only 
AC-3 2003 Sydney AC – Sodium iodide (NaI) impregnated Ferrous chloride Closed-topped 3.0 s 12 m3  
AC-4 2004 Sydney AC – Caustic impregnated None Closed-topped 2.7 s 5.3 m3 VOCs only 
AC-5 2010 Sydney AC – Caustic impregnated Ferrous chloride Closed-topped 3.0 s 12 m3 VSCs and VOCs 
AC-6 2003 Sydney AC – Sodium iodide (NaI) impregnated Ferrous chloride Closed-topped 3.5 s 11 m3 VSCs and VOCs 
AC-7 2003 Sydney AC – Sodium iodide (NaI) impregnated Ferrous chloride Closed-topped 4.4 s 13 m3 VSCs only 
AC-8 2009 Melbourne AC – 75% activated alumina/25% proprietary ‘high H2S’ activated carbon None Closed-topped 5.4 s 0.8 m3 VSCs only 
AC-9 2006 Melbourne AC – Caustic impregnated/virgin carbon blend (proportions unknown) None Closed-topped Unknown Note 3 Unknown VSCs only 
BF-1 2006 Sydney BF – 65% bark chip, 25% soil, 5% peat, 5% manure Ferrous chloride/Magnesium hydroxide Open-topped 76 s 21 m3 VSCs only 
BF-2 2009 Sydney BF – 45% shredded wood, 35% composted wood, 10% pine bark nuggets, 2% compost, 3% mushroom compost, 5% limestone Magnesium hydroxide Open-topped 25 s 56 m3 VSCs and VOCs 
BF-3 2009 Melbourne BF – Proprietary blend None Closed-topped 52 s 8.6 m3 VSCs and VOCs 
BF-4 2010 Melbourne BF – Proprietary blend None Closed-topped 61 s 6.3 m3 VSCs only 
BF-5 Note 2 2011 Perth BF – Proprietary blend None Closed-topped 34 s 8.6 m3 VSCs and VOCs 
BTF-1 2010 Sydney BTF – Polyurethane foam None Closed-topped 5.5 s 3.0 m3 VSCs and VOCs 
BTF-3 2007 Perth BTF – HD-QPAC Oxygen Closed-topped 256 s 71 m3 VSCs and VOCs 
Site Year installed Note 1 City Media type Upstream chemical dosing Type Approx. EBRT Volume Species investigated after data filtering 
AC-1 2003 Sydney AC – Sodium hydroxide (NaOH) Impregnated Ferrous chloride Closed-topped 18.0 s 72 m3 VOCs only 
AC-2 2008 Sydney AC – 50% Caustic impregnated/50% Virgin None Closed-topped 9.5 s 9.5 m3 VOCs only 
AC-3 2003 Sydney AC – Sodium iodide (NaI) impregnated Ferrous chloride Closed-topped 3.0 s 12 m3  
AC-4 2004 Sydney AC – Caustic impregnated None Closed-topped 2.7 s 5.3 m3 VOCs only 
AC-5 2010 Sydney AC – Caustic impregnated Ferrous chloride Closed-topped 3.0 s 12 m3 VSCs and VOCs 
AC-6 2003 Sydney AC – Sodium iodide (NaI) impregnated Ferrous chloride Closed-topped 3.5 s 11 m3 VSCs and VOCs 
AC-7 2003 Sydney AC – Sodium iodide (NaI) impregnated Ferrous chloride Closed-topped 4.4 s 13 m3 VSCs only 
AC-8 2009 Melbourne AC – 75% activated alumina/25% proprietary ‘high H2S’ activated carbon None Closed-topped 5.4 s 0.8 m3 VSCs only 
AC-9 2006 Melbourne AC – Caustic impregnated/virgin carbon blend (proportions unknown) None Closed-topped Unknown Note 3 Unknown VSCs only 
BF-1 2006 Sydney BF – 65% bark chip, 25% soil, 5% peat, 5% manure Ferrous chloride/Magnesium hydroxide Open-topped 76 s 21 m3 VSCs only 
BF-2 2009 Sydney BF – 45% shredded wood, 35% composted wood, 10% pine bark nuggets, 2% compost, 3% mushroom compost, 5% limestone Magnesium hydroxide Open-topped 25 s 56 m3 VSCs and VOCs 
BF-3 2009 Melbourne BF – Proprietary blend None Closed-topped 52 s 8.6 m3 VSCs and VOCs 
BF-4 2010 Melbourne BF – Proprietary blend None Closed-topped 61 s 6.3 m3 VSCs only 
BF-5 Note 2 2011 Perth BF – Proprietary blend None Closed-topped 34 s 8.6 m3 VSCs and VOCs 
BTF-1 2010 Sydney BTF – Polyurethane foam None Closed-topped 5.5 s 3.0 m3 VSCs and VOCs 
BTF-3 2007 Perth BTF – HD-QPAC Oxygen Closed-topped 256 s 71 m3 VSCs and VOCs 

Note 1 – All AC sites had AC replaced close to the start of the monitoring period.

Note 2 – First stage only of two-stage process investigated.

Note 3 – 0.39 m3/s flow rate.

Sample analysis

VOC analysis

VOC analysis was performed in duplicate using gas chromatography coupled to mass spectrometry (GC-MS) (Agilent 6890NGC, 5973NMSD, Agilent Technologies). Helium was used as the carrier gas at a flow rate of 1.8 mL/min. The temperature program for the GC was initially held at 50 °C for 2 min and then raised to 220 °C at a rate of 15 °C /min. The mass spectrometer acquired data in scan mode with mass-to-charge (m/z) intervals ranging from 35 to 335. The identity of compounds was verified using reference standards, by initially matching the mass spectra with a mass spectrum library (NIST02 library) available in the GC-MS system.

VSCs analysis

VSC bag samples were connected to an Air Server (CIA 8, Markes International, UK) and collected onto a specialised sulfur cold trap (U-T6SUL, Markes International, UK) prior to injection. Sample analysis was performed using a gas chromatograph equipped with a sulfur chemiluminescence detector (SCD) (7890N GC and 355 SCD, Agilent Technologies, USA) with helium as the carrier gas (flow rate of 1 mL/min). The gas chromatograph column temperature was initially held at 37 °C for 3 min, then raised at a rate of 15 WC/min to 225 °C, and then held for 2 min. DB-VRX (30 m × 0.25 mm × 1.4 μm) (Cheng et al. 2007) was used as the analytical column for both VOC and VSC separation.

The sample and chemical analysis methodologies are described in more detail in Sivret et al. (2016) and Wang et al. (2015).

RESULTS

Data analysis

Data were filtered to remove spikes and to separate rain events and non-warm periods. The remaining data were evaluated for completeness. Sites where there were insufficient data to make a meaningful analysis (generally <5 data points) were removed. Thirteen sites (six AC filters, five biofilters and two BTFs) were deemed to have sufficient information to analyse six VSCs further. Eleven sites (six AC filters, three BFs and two BTFs) were deemed to have sufficient information to analyse seven VOCs further. VSCs and VOCs investigated are listed in Table 3. The removal of these VSCs and VOCs during warm and dry weather were also compared to non-warm periods and to wet periods as discussed further in Shammay et al. (2016b), with the resulting graphs found in Figures S-1 to S-6 in the Supplementary Material (available with the online version of this paper).

Table 3

VSCs and VOCs investigated in AC filters, biofilters and BTFs

Compound Median concentration (μg/m3Appearance frequency (%) 95th percentile odour activity value Odour threshold value (μg/m3) (Nagata 2003Odorant rankinga (Sivret et al. 2016
Volatile sulfurous compounds 
 Hydrogen sulfide 1,880 100 68,400 0.576 
 Methyl mercaptan 293 93.7 16,600 0.14 
 Carbon disulfide 11.0 91.7 0.053 659 
 Dimethyl sulfide 65.4 93.3 3.22 77 
 Dimethyl disulfide 8.72 91.3 5.60 8.5 
 Dimethyl trisulfide 6.40 88.5 – – N/A 
Volatile organic compounds 
 Ethylbenzene 2.37 79.9 0.16 744 
m,p-Xylene 7.35 75.7 1.45 254 
o-Xylene 0.31 33.8 0.074 1,664 
 Toluene 31.7 91.3 0.42 1,254 
 Trimethylbenzene 31.8 74.5 0.77 595 
 Trichloromethane 55.4 82.9 0.021 18,707 
 1,4-Dichlorobenzene 1.05 40.3 – – N/A 
Compound Median concentration (μg/m3Appearance frequency (%) 95th percentile odour activity value Odour threshold value (μg/m3) (Nagata 2003Odorant rankinga (Sivret et al. 2016
Volatile sulfurous compounds 
 Hydrogen sulfide 1,880 100 68,400 0.576 
 Methyl mercaptan 293 93.7 16,600 0.14 
 Carbon disulfide 11.0 91.7 0.053 659 
 Dimethyl sulfide 65.4 93.3 3.22 77 
 Dimethyl disulfide 8.72 91.3 5.60 8.5 
 Dimethyl trisulfide 6.40 88.5 – – N/A 
Volatile organic compounds 
 Ethylbenzene 2.37 79.9 0.16 744 
m,p-Xylene 7.35 75.7 1.45 254 
o-Xylene 0.31 33.8 0.074 1,664 
 Toluene 31.7 91.3 0.42 1,254 
 Trimethylbenzene 31.8 74.5 0.77 595 
 Trichloromethane 55.4 82.9 0.021 18,707 
 1,4-Dichlorobenzene 1.05 40.3 – – N/A 

aK – Key, H – High, L – Low, M – Medium: odorant rankings based on methodology defined in Sivret et al. (2016).

VSC removal

Figure 1 summarises the removal of all VSCs within the study for each odour abatement system type. Data summarised in Figure 1 are only for the warm, dry periods under investigation as this is the period that most odour abatement systems are designed to treat. A comparison of performance between warm and dry, cold and dry, and wet periods for each of the sites is provided in the Supplementary Material. It was found that, in general, warm and dry periods were the times when the units performed the most poorly.

Figure 1

Removal of volatile sulfurous compounds.

Figure 1

Removal of volatile sulfurous compounds.

Activated carbon filters

Four of the six AC filters analysed for VSC removal performed well in terms of H2S and MM removal under warm and dry conditions. The two remaining units still exhibited high removals compared to other VSCs. Overall, H2S removal was greater than 90% in 74% of samples, where MM removal was greater than 90% in 90% of samples. This corroborates with what has been previously reported in the industry for H2S and MM removal in AC systems (Wallis et al. 1996; Adams & Thomason 2004; Ball et al. 2004; Kyi 2014; Mohamed et al. 2014; Wang et al. 2014).

All AC filters exhibited poor removal for CS2, DMS, DMDS and DMTS, which may be due to incompatible AC type used or too short EBRTs (Goyal et al. 2008; SWC 2011). In the case of DMDS, this may be due to the breakdown of MM generating DMDS (Bashkova et al. 2002). In most locations and for most non-H2S and non-MM compounds, removal efficiency during non-warm periods was greater than during warm periods. This may be due to the removal mechanisms for DMS, DMDS, DMTS and CS2 being based on physisorption, which can be enhanced at lower temperatures, as opposed to the removal mechanisms for H2S and MM being based on chemisorption, which is less temperature dependent (Shammay et al. 2016a).

Biofilters

H2S was removed well in BFs with removals over 90% exhibited over 98% of measurements. MM exhibited similar removals to H2S, with removals over 90% exhibited over 97% of measurements. This is consistent with the literature, where MM removal was reported as either similar or less than H2S removal (Wolstenholme & Finger 1994; Deshusses et al. 2001; Converse et al. 2003; Alix & Williams 2006; Wang et al. 2014).

Of all the VSCs, carbon disulfide exhibited the most consistently poor removal efficiency across all BFs, particularly at the open-topped units. No measurements recorded over 99% removal efficiency and only 4.7% of measurements had removals above 90%. Relatively few literature sources discuss CS2 removal, perhaps as it is not commonly identified as an odorant of concern (Sivret et al. 2016). The few found reported mixed removals, with some reporting low removals in the range of 13–36% (Deshusses et al. 2001; Converse et al. 2003; Sorensen & Baadsgard 2004) whilst others reported removals up to 100% (Giggey et al. 1994; Wang et al. 2014). Rojo et al. (2012) found the accumulation of biodegradation byproducts impacts the removal of CS2 in laboratory-scale biofilters. The open-topped units in this study exhibited better performance during non-warm weather compared to warm weather, and the closed-topped units showed little difference between warm and non-warm weather. This implies, as expected, that the temperature within the closed-topped units is more stable than in the open-topped units. It also implies that CS2 removal is inversely related to ambient temperature, which may mean that CS2 is removed more from sorption than bioactivity (McNevin & Barford 2000).

Compared to H2S and MM, the dimethyl mono-, di- and tri-methyl sulfides were removed poorly, with median removals between 82–92% and greater than 90% removal occurring only 41–57% of the time. This low and inconsistent removal agrees with some previous studies (Deshusses et al. 2001; Converse et al. 2003; Wang et al. 2014), but not with others (Giggey et al. 1994; Sorensen & Baadsgard 2004; Alix & Williams 2006). DMS was generally removed well, with DMDS and DMTS exhibiting poorer removals compared to DMS at most BF sites.

Biotrickling filters

BTFs performed well in terms of H2S removal under all conditions, with 100% of measurements showing greater than 90% removal and 65% of measurements showing greater than 99% removal. This is consistent with literature, where typically greater than 99% removal is recorded for BTFs even at low EBRTs (Webster et al. 2000; Cox & Deshusses 2001; Deshusses et al. 2001; Gabriel et al. 2002; Gilani et al. 2004; Morton et al. 2004; Rousseille et al. 2011).

Similar to BFs, MM exhibited fair removals; however, not as high as H2S, with median removal of 82.7% recorded and removals over 90% exhibited over 78% of measurements. This is greater than that recorded in the literature of 56–78% (Deshusses et al. 2001; Converse et al. 2003) with one study as low as 21% (Rousseille et al. 2011).

Greater variability was exhibited for MM removal in BTFs when compared with MM removal in BFs and H2S removal in BTFs. Similar to BFs, this higher variability was not impacted by EBRT with BTF-1 (5.5 s EBRT) exhibiting similar or lower variability than BTF-3 (256 s EBRT), which agrees with other studies (Patria et al. 2001; Lebrero et al. 2012a).

Poor and variable removals were recorded for CS2, with median removals of 51% identified. No measurements recorded over 99% removal efficiency and only 4.7% of measurements had removals above 90%. Relatively few literature sources discuss CS2 removal. The few found reported either no removal (Deshusses et al. 2001; Converse et al. 2003) or low removals of 30–40% (Gabriel et al. 2002).

The dimethyl mono-, di- and tri-methyl sulfides were removed poorly with high variability. DMS exhibited the lowest removals with only 19.6% of all samples showing removals over 50%, compared with 29.8% and 41.5% of samples for DMDS and DMTS respectively. Significant negative removals were identified for all three compounds, the worst being for DMDS where 59% of measurements showed a negative removal compared with 41% and 45% for DMS and DMTS respectively. Only one full-scale, field-based study could be identified which reported DMS removal in BTFs and it showed a 0% removal (Converse et al. 2003). DMDS and DMTS removals in full-scale, field based BTFs could not be found in the literature.

Comparison

H2S removal was good throughout all types of odour abatement systems as can be expected as most odour abatement systems are designed to remove hydrogen sulfide. H2S removal was most consistent in BTFs compared to other technologies, with no negative removals identified.

AC filters exhibited the highest median methyl mercaptan removal; however, there were occasions where removal efficiency reduced and became negative. ACs exhibited poor removal of CS2 and DMDS with median removals below 0%. This supports other studies investigating the removal of carbon disulfide and dimethyl disulfide from ACs, which identified low or negative removals (Adams & Thomason 2004; Ball et al. 2004; Kyi 2014).

BTFs exhibited the poorest methyl mercaptan removal of the three odour abatement systems investigated (Deshusses et al. 2001; Gabriel et al. 2002; Rousseille et al. 2011). DMS and DMTS also exhibited particularly poor removal with median removals below 0%, however no other studies could be identified which investigated the removal of these compounds.

BFs exhibited the highest and most consistent removal of CS2, DMS, DMDS and DMTS; however, removal was still fairly poor with frequent removal efficiencies less than 0%. Other studies have shown a wide range of removals for these substances (Giggey et al. 1994; Deshusses et al. 2001; Sorensen & Baadsgard 2004; Alix & Williams 2006).

Overall, whilst the number of compounds sampled and the length of sampling of this study is greater than most others, the results are somewhat consistent with previous studies identifying removal efficiencies of VSCs from each type of odour abatement systems in sewer networks (Shammay et al. 2016a).

VOC removal

Figure 2 summarises the removal of all VOCs within the study for each odour abatement system type. BFs exhibited the highest removal and least variability for most VOC species, however there was still a large variability identified. m,p-Xylene, the only VOC that has been identified as a ‘High’ risk odorant (Sivret et al. 2016), was removed the highest and most consistently by BFs.

Figure 2

Removal of volatile organic compounds.

Figure 2

Removal of volatile organic compounds.

Activated carbon filters

All compounds at all locations exhibited a wide range of removal efficiencies. No compound was identified where removal rates were consistently above 0%. This may be due to compounds being adsorbed between periods when spot samples were taken and occasionally desorbing. This would imply that these VOCs are only weakly adsorbed and can either be desorbed by atmospheric fluctuations or the presence of other compounds, or can be produced as part of the degradation of other compounds (Shammay et al. 2016a).

Units with higher EBRTs generally exhibited a greater removal, however there were specific locations, such as AC-2, which did not show this trend despite having an EBRT of 9.5 s.

Biofilters

In general, VOC removal was poorer than VSC removal, with negative removals being more often reported. Negative removals could be due to a number of factors, such as the generation of VOCs as biological byproducts (Mrowiec et al. 2005) or the desorption of previously adsorbed VOCs.

1,4-Dichlorobenzene had relatively good removal at BF-5, with poorer removals at BF-2. Median removal of 90.2% was measured across all sites and conditions, with removals greater than 50% being recorded for 68% of measurements. This level of removal is consistent with the findings of some (Sorensen & Baadsgard 2004) but greater than that reported by others (Deshusses et al. 2001; Converse et al. 2003). Removal of halogenated VOCs in general in biological systems has been reported as poor, or requiring long EBRTs to treat, predominantly due to their hydrophobicity (Cox et al. 2002; Delhoménie & Heitz 2005; Iranpour et al. 2005; Kennes et al. 2009; Muñoz et al. 2015).

Ethylbenzene also had relatively good removal at one site (BF-5) with poorer removals at others (BF-2 and BF-3 in particular). Median removal of 62.3% was measured across all sites, with removals greater than 50% being recorded for 52.4% of measurements. Only one previous study was identified that reported on the removal of ethylbenzene (Wang et al. 2014), which reported higher levels of removal than measured in this study.

Removal of all isomers of xylene was relatively poor across all sites, with median removals of 66.4% and 62.9% for m,p- and o-isomers, respectively. This is consistent with the findings from some (Deshusses et al. 2001; Converse et al. 2003; Wang et al. 2014), however higher removals have been reported when granular AC was used as biofilter media (Webster et al. 1996). These removals are greater than those reported by others (Wolstenholme & Finger 1994; Chitwood et al. 1999).

Toluene was removed relatively well from two sites (BF-2 and BF-5) but exhibited poor removal from a third (BF-3). Median removal of 76.9% was recorded with removals above 50% occurring in 62.1% of measurements. This is consistent with some studies (Webster et al. 1996; Chitwood et al. 1999; Deshusses et al. 2001; Converse et al. 2003; Wang et al. 2014) yet others (Ergas et al. 1995; Sorensen & Baadsgard 2004) reported higher removals. Toluene has been known to be formed during biogenic processes, particularly during anaerobic digestion (Mrowiec et al. 2005) however only one site (BF-3) exhibited consistent negative removals of toluene. This may indicate that BF-3 was operating under anaerobic conditions, or had pockets of anaerobic areas if short-circuiting and overwetting occurred.

Total trimethylbenzene was well removed at two sites (BF-2 and BF-5) and poorly removed at a third site (BF-3). A median removal of 95.0% was recorded across all sites and conditions with removals above 50% occurred in 87.1% of measurements. No literature could be found reporting the removal of total trimethylbenzene in field-based biofilters; however, a maximum removal of 70% was identified in a single-contaminant loaded biofilter under laboratory conditions (Delhoménie et al. 2003).

Trichloromethane exhibited relatively poor removals compared to other VOCs in this study with a median removal of 33.0% being recorded, despite having a similarly low Henry's law constant. Removals above 50% occurred in 33.4% of measurements. No literature could be found relating to the removal of trichloromethane in field-based biofilters; however, removal of dichloromethane has had mixed removal efficiencies reported in biofilters ranging from negative values to as high as 98% (Webster et al. 1996; Chitwood et al. 1999; Deshusses et al. 2001; Converse et al. 2003; Sorensen & Baadsgard 2004).

Biotrickling filters

Similar to BFs, VOC removal was generally poorer than VSC removals.

1,4-Dichlorobenzene was removed poorly at both sites with the second lowest median removal (after trichloromethane) of 14.9% measured across all sites, with removals greater than 50% being recorded for 40.6% of measurements. At BTF-3, removal increased with lower temperatures. No literature could be found for comparison on 1,4-dichlorobenzene removal in BTFs in field-based full-scale systems.

Removal of ethylbenzene was similar to that exhibited by 1,4-dichlorobenzene. Median removal of 32.1% was measured across all sites and conditions with removals greater than 50% being recorded for 38.7% of measurements. Only one study was identified which reported on the removal of ethylbenzene (Gabriel et al. 2002) which reported lower levels of removal, 21%, than measured in this study but at shorter EBRTs (1.6–2.3 s).

Removal of all isomers of xylene was mixed across all sites, with median removals of 56.3% and 46.5% for m, p- and o- isomers respectively. This is greater than the findings from most studies (Torres et al. 1996; Webster et al. 2000; Deshusses et al. 2001; Gabriel et al. 2002; Converse et al. 2003; Gilani et al. 2004).

Removal of toluene was mixed between the two BTFs. Overall median removal of 62.1% was recorded, which was lower than from the BFs; however, removals above 50% were found to occur in 63.7% of measurements, which is similar to the BFs. This is similar to those reported by Torres et al. (1996), greater than those reported by others (Deshusses et al. 2001; Gabriel et al. 2002; Converse et al. 2003; Gilani et al. 2004) and lower than those reported by Webster et al. (2000) and in laboratory-scale results (Lebrero et al. 2012b).

Total trimethylbenzene showed the greatest removal of all VOCs under investigation, similar to BFs; however, the overall median removal was lower than BFs. A median removal of 81.5% was recorded across all sites and conditions, with removals above 50% occurring in 76.8% of measurements. No literature could be found reporting the removal of total trimethylbenzene in field-based BTFs.

Trichloromethane exhibited the lowest removals compared to other VOCs in this study, with a median removal of −11.2% being recorded. Removals above 50% occurred in only 11% of measurements, which is lower than those experienced in BFs. No literature could be found relating to the removal of trichloromethane in field-based BTFs; however, removal of dichloromethane has been widely reported as poor in biologically based systems (Webster et al. 1996; Chitwood et al. 1999; Deshusses et al. 2001; Converse et al. 2003; Sorensen & Baadsgard 2004).

Comparison

The biological based filters (both BFs and BTFs) showed a greater removal of total trimethylbenzene than in the AC systems. BFs had the highest toluene removal, yet BTFs had the least variable toluene removal. BFs exhibited the highest and most consistent xylene removal. BFs and AC filters exhibited similar removal of 1,4-dichloro benzene.

In general, AC filters exhibited little VOC removal, with only 1,4-dichloro benzene consistently identified as having a removal greater than 0%.

These findings generally support previous investigations; however, no data have been found in the literature for some of the compounds/odour abatement system combinations under investigation (Wang et al. 2014; Shammay et al. 2016a; Shammay et al. 2016b).

The results imply that, if there are more compounds than H2S and MM, the ability to treat these other compounds should form part of the technology selection process. If DMS, DMDS, DMTS or many VOCs are present, a biological system may be more beneficial than an AC system at treating these compounds. Two-stage systems, whilst outside the scope of this study, would be expected to provide additional compound removal. In two-stage systems, careful consideration should be given as to whether the second stage is biological or physical, depending on the type of residual compounds expected after the first stage.

CONCLUSIONS

Most odour abatement systems are designed to treat H2S due to the ease of measurement and the wealth of literature on how H2S is degraded in odour abatement systems. However, other compounds are present in sewer atmospheres, which, when discharged, can be the cause of odour complaints or could potentially impact on the removal of odorants in odour abatement systems.

Three odour abatement systems commonly found in sewer networks: AC, BTFs and biofilters, were evaluated for long-term removal of common volatile sulfurous and organic compounds. Whilst all odour abatement systems were found to readily remove hydrogen sulfide, methyl mercaptan was best removed by AC and biofilters showed the greatest and most consistent removal of all other VSCs.

Biofilters exhibited the greatest and most consistent removal for most VOCs, with the exception of toluene and ethylbenzene, which was most readily removed by BTFs. AC systems exhibited the poorest removal of all VOCs.

Given the differences in removal efficiencies of different compounds by different odour abatement systems, a knowledge of what is in the foul air to be treated is essential prior to designing odour abatement systems. If an odour abatement system is designed purely to remove hydrogen sulfide, other compounds will not be treated as effectively leading to a greater risk of odour complaints.

ACKNOWLEDGEMENTS

This work is supported by the Australian Research Council Linkage Project (LP0882016) with industry support from Barwon Regional Water Corporation, Gold Coast Water, Hunter Water Corporation, Melbourne Water, South Australian Water Corporation, South East Water, Sydney Water Corporation, Veolia Water Australia, Water Corporation of Western Australia and WQRA, as well as CH2MHill Australia. Ari Shammay is supported by Stantec.

REFERENCES

REFERENCES
Adams
,
D.
&
Thomason
,
K.
2004
Remarkably Low Maintenance and Effective Odor Control From Media-Based Odor Control Systems
.
Water Environment Federation and Air and Waste Management Association
,
Washington
,
USA
.
Alix
,
C.
&
Williams
,
T.
2006
Biofilters make good neighbors
.
Water Environment & Technology
18
(
3
),
57
59
.
Apgar
,
D.
,
Witherspoon
,
J.
,
Easter
,
C.
,
Basrai
,
S. S.
,
Dillon
,
C.
,
Torres
,
E.
,
Corsi
,
R.
,
Forbers
,
B.
,
Quigley
,
C.
,
Ward
,
M.
,
Morton
,
R. L.
,
Weiss
,
J.
,
Bowker
,
R. P. G.
,
Davidson
,
S.
,
Wolstenholme
,
P.
,
Joyce
,
J.
&
Stuetz
,
R. M.
2007
Minimization of odors and corrosion in collection systems - Phase 1
.
Water Environment Research Foundation
,
Alexandria, VA
,
USA
.
Ball
,
B.
,
Apgar
,
D.
&
Maloney
,
J.
2004
Performance Evaluation of the Phoenix Odor Control System
.
Water Environment Foundation and Air and Water Management Association
,
Washington
,
USA
.
Bashkova
,
S.
,
Bagreev
,
A.
&
Bandosz
,
T. J.
2002
Effect of surface characteristics on adsorption of methyl mercaptan on activated carbons
.
Industrial & Engineering Chemistry Research
41
(
17
),
4346
4352
.
Chitwood
,
D. E.
,
Devinny
,
J. S.
&
Edward Reynolds
,
F.
1999
Evaluation of a two-stage biofilter for treatment of POTW waste air
.
Environmental Progress
18
(
3
),
212
221
.
Converse
,
B. M.
,
Schroeder
,
E. D.
,
Iranpour
,
R.
,
Cox
,
H. H. J.
&
Deshusses
,
M. A.
2003
Odor and volatile organic compound removal from wastewater treatment plant headworks ventilation air using a biofilter
.
Water Environment Research
75
(
5
),
444
454
.
Cox
,
H. H. J.
,
Deshusses
,
M. A.
2001
Bioreactors for Waste Gas Treatment, Chapter 4: Biotrickling Filters
(
Kennes
,
C.
&
Veiga
,
M. C.
, eds).
Kluwer Academic Publishers
,
Dordrecht
,
The Netherlands
.
Cox
,
H. H.
,
Deshusses
,
M. A.
,
Converse
,
B. M.
,
Schroeder
,
E. D.
&
Iranpour
,
R.
2002
Odor and volatile organic compound treatment by biotrickling filters: pilot-scale studies at hyperion treatment plant
.
Water Environ. Res.
74
(
6
),
557
563
.
Davidson
,
S.
,
Green
,
J.
,
Mann
,
J.
&
Lamb
,
E.
2004
Design challenges in sewer foul air extraction and treatment
.
Water Environment Federation and Air and Waste Management Association
,
Washington, DC, USA
.
Delhoménie
,
M.-C.
&
Heitz
,
M.
2005
Biofiltration of air: a review
.
Critical Reviews in Biotechnology
25
(
1–2
),
53
72
.
Delhoménie
,
M.-C.
,
Bibeau
,
L.
,
Gendron
,
J.
,
Brzezinski
,
R.
&
Heitz
,
M.
2003
Degradation of toluene, xylene, and trimethylbenzene vapors by biofiltration: a comparison
.
Journal of the Air & Waste Management Association
53
(
2
),
217
226
.
Deshusses
,
M. A.
,
Cox
,
H. H. J.
,
Schroeder
,
E. D.
&
Converse
,
B. M.
2001
Controlling Odor and VOCs with Bioreactors
.
WERF
,
Alexandria, VA
.
Ergas
,
S. J.
,
Schroeder
,
E. D.
,
Chang
,
D. P. Y.
&
Morton
,
R. L.
1995
Control of volatile organic compound emissions using a compost biofilter
.
Water Environment Research
67
(
5
),
816
821
.
Gabriel
,
D.
,
Cox
,
H. H. J.
,
Brown
,
J.
,
Torres
,
E.
&
Deshusses
,
M. A.
2002
Biotrickling Filters for POTWs Air Treatment Full-Scale Experience with a Converted Scrubber
.
Water Environment Federation
,
Alburquerque, NM
.
Giggey
,
M. D.
,
Dwinal
,
C. A.
,
Pinnette
,
J. R.
&
O'Brien
,
M. A.
1994
Performance Testing of Biofilters in a Cold Climate
.
Water Environment Federation
,
Alexandria, VA
, pp.
4
29
/24–39
.
Gilani
,
A. A.
,
Tim Haug
,
R.
,
Wojlaw
,
J. A.
&
Van Durme
,
G. P.
2004
Los Angeles Selects Bio-Trickling/Activated Carbon for Collection System Odor Control
.
Water Environment Federation and Air and Waste Management Association
,
Washington
,
USA
.
Gostelow
,
P.
,
Parsons
,
S. A.
&
Stuetz
,
R. M.
2001
Odour measurements for sewage treatment works
.
Water Res.
35
(
3
),
579
597
.
Goyal
,
M.
,
Dhawan
,
R.
&
Bhagat
,
M.
2008
Adsorption of dimethyl sulfide vapors by activated carbons
.
Colloids and Surfaces A: Physicochemical and Engineering Aspects
322
(
1–3
),
164
169
.
Iranpour
,
R.
,
Cox
,
H. H. J.
,
Deshusses
,
M. A.
&
Schroeder
,
E. D.
2005
Literature review of air pollution control biofilters and biotrickling filters for odor and volatile organic compound removal
.
Environ. Prog.
24
(
3
),
254
267
.
Kennes
,
C.
,
Rene
,
E. R.
&
Veiga
,
M. C.
2009
Bioprocesses for air pollution control
.
J. Chem. Technol. Biotechnol.
84
,
1419
1436
.
Koe
,
L. C. C.
&
Tan
,
N. C.
1990
Field performance of activated carbon adsorption for sewage air
.
Journal of Environmental Engineering
116
(
4
),
721
734
.
Kyi
,
M.
2014
Trickling Filters Foul Air Characteristics and Treatment with Activated Carbon Media
.
Water Environment Foundation
,
Miami, FL
.
Lebrero
,
R.
,
Rodríguez
,
E.
,
Estrada
,
J. M.
,
García-Encina
,
P. A.
&
Muñoz
,
R.
2012a
Odor abatement in biotrickling filters: effect of the EBRT on methyl mercaptan and hydrophobic VOCs removal
.
Bioresource Technology
109
(
0
),
38
45.
Lebrero
,
R.
,
Estrada
,
J. M.
,
Muñoz
,
R.
&
Quijano
,
G.
2012b
Toluene mass transfer characterization in a biotrickling filter
.
Biochemical Engineering Journal
60
,
44
49
.
McNevin
,
D.
&
Barford
,
J.
2000
Biofiltration as an odour abatement strategy
.
Biochemical Engineering Journal
5
(
3
),
231
242
.
Mohamed
,
F.
,
Kim
,
J.
,
Huang
,
R.
,
Nu
,
H. T.
&
Lorenzo
,
V.
2014
Efficient control of odors and VOC emissions via activated carbon technology
.
Water Environment Research
86
(
7
),
594
605
.
Morton
,
R. L.
,
Bao
,
M.
,
Ackman
,
P.
,
Tang
,
C.
,
Horvath
,
R.
&
Stahl
,
J.
2004
Comparison of Different Biotrickling Filter Technologies For Treating Air from Wastewater Treatment Facilities
.
Water Environment Federation and Air and Waste Management Association
,
Washington
,
USA
.
Mrowiec
,
B.
,
Suschka
,
J.
&
Keener
,
T. C.
2005
Formation and biodegradation of toluene in the anaerobic sludge digestion process
.
Water Environment Research
77
(
3
),
274
278
.
Mudliar
,
S.
,
Giri
,
B.
,
Padoley
,
K.
,
Satpute
,
D.
,
Dixit
,
R.
,
Bhatt
,
P.
,
Pandey
,
R.
,
Juwarkar
,
A.
&
Vaidya
,
A.
2010
Bioreactors for treatment of VOCs and odours – A review
.
Journal of Environmental Management
91
(
5
),
1039
1054
.
Muñoz
,
R.
,
Sivret
,
E. C.
,
Parcsi
,
G.
,
Lebrero
,
R.
,
Wang
,
X.
,
Suffet
,
I. H.
&
Stuetz
,
R. M.
2010
Monitoring techniques for odour abatement assessment
.
Water Research
44
(
18
),
5129
5149
.
Muñoz
,
R.
,
Malhautier
,
L.
,
Fanlo
,
J.-L.
&
Quijano
,
G.
2015
Biological technologies for the treatment of atmospheric pollutants
.
International Journal of Environmental Analytical Chemistry
95
(
10
),
950
967
.
Nagata
,
Y.
2003
Measurement of Odor Threshold by Triangle Odor Bag Method
.
Ministry of the Environment (ed.)
,
Government of Japan
,
Japan
, pp.
118
127
.
Patria
,
L.
,
Cathelain
,
M.
,
Laurens
,
P.
&
Barbere
,
J. P.
2001
Odour removal with a trickling filter at a small WWTP strongly influenced by the tourism season
.
Water Sci. Technol.
44
(
2–3
),
243
249
.
Rojo
,
N.
,
Muñoz
,
R.
,
Gallastegui
,
G.
,
Barona
,
A.
,
Gurtubay
,
L.
,
Prenafeta-Boldú
,
F. X.
&
Elías
,
A.
2012
Carbon disulfide biofiltration: influence of the accumulation of biodegradation products on biomass development
.
Journal of Chemical Technology & Biotechnology
87
(
6
),
764
771
.
Rousseille
,
F.
,
Nastasi
,
V.
&
Koers
,
B.
2011
Hydrogen Sulfide Emissions and its Removal by Biotrickling Filters: an Interesting Feedback from as Samra Waste Water Treatment Plant in Jordan
.
Crouna, U.d. (ed.)
,
A Coruna
,
Spain
.
Shammay
,
A.
,
Sivret
,
E. C.
,
Le-Minh
,
N.
,
Lebrero Fernandez
,
R.
,
Evanson
,
I.
&
Stuetz
,
R. M.
2016a
Review of odour abatement in sewer networks
.
Journal of Environmental Chemical Engineering
4
,
3866
3881
.
Shammay
,
A.
,
Sivret
,
E. C.
,
Wang
,
B.
,
Evanson
,
I.
&
Stuetz
,
R. M.
2016b
Performance of field based activated carbon systems for odour control
.
Chemical Engineering Transactions
53
,
277
282
.
Sivret
,
E. C.
&
Stuetz
,
R. M.
2010
Sewer odour abatement practices – an Australian survey
.
Water
37
,
77
81
.
Sivret
,
E. C.
,
Wang
,
B.
,
Parcsi
,
G.
&
Stuetz
,
R. M.
2013
Sewer Odours: More Than Just Rotten Eggs
.
CASANZ
,
Sydney
,
Australia
.
Sivret
,
E. C.
,
Wang
,
B.
,
Parcsi
,
G.
&
Stuetz
,
R. M.
2016
Prioritisation of odorants emitted from sewers using odour activity values
.
Water Research
88
,
308
321
.
Sivret
,
E. C.
,
Le-Minh
,
N.
,
Wang
,
B.
,
Wang
,
X.
&
Stuetz
,
R. M.
2017
Dynamics of volatile sulfur compounds and volatile organic compounds in sewer headspace air
.
Journal of Environmental Engineering
143
(
2
),
04016080
.
Sorensen
,
H. W.
&
Baadsgard
,
M.
2004
Biofilter Inspection and Testing Program Results in Lower Biofilter O&M Costs and Improves Biofilter Performance
.
Water Environment Federation and Air and Waste Management Association
,
Washington
,
USA
.
SWC (Sydney Water Corporation)
2011
Odour Control Unit Standard Specification
.
Sydney Water Corporation
,
Sydney, NSW
.
Torres
,
E.
,
Basrai
,
S. S.
&
Kogan
,
V.
1996
Evaluation of Two Biotechnologies Controlling POTW Air Emissions
.
Reynolds Group, Los Angeles, CA
, pp.
182
197
.
Wallis
,
M.
,
Witherspoon
,
J.
,
Suzuki
,
S.
,
Card
,
T. R.
,
Mayer
,
G.
&
Chang
,
D. P. Y.
1996
A Critical Review of Odor Control Equipment for Toxic Air Emissions Reduction: Final Report
.
Water Environment Research Foundation
,
Alexandria, VA
.
Wang
,
B.
,
Sivret
,
E. C.
,
Parcsi
,
G.
,
Wang
,
X.
,
Le
,
N. M.
,
Kenny
,
S.
,
Bustamante
,
H.
&
Stuetz
,
R. M.
2014
Is H2S a suitable process indicator for odour abatement performance of sewer odours?
Water Sci. Technol.
69
(
1
),
92
98
.
Wang
,
B.
,
Sivret
,
E. C.
,
Parcsi
,
G.
&
Stuetz
,
R. M.
2015
Determination of VOSCs in sewer headspace air using TD–GC–SCD
.
Talanta
137
(
0
),
71
79
.
Webster
,
T. S.
,
Devinny
,
J. S.
,
Torres
,
E. M.
&
Basrai
,
S. S.
1996
Biofiltration of odors, toxics and volatile organic compounds from publicly owned treatment works
.
Environmental Progress
15
(
3
),
141
147
.
Webster
,
T. S.
,
Togna
,
A. P.
,
Guarini
,
W. J.
,
Hooker
,
B.
,
Tran
,
H.
,
Sanfedele
,
J.
&
Olsen
,
J.
2000
Treatment of Vapor Emissions Generated from an Industrial Wastewater Treatment Plant Using a Full-Scale Biotrickling Filter Reactor
.
Air and Waste Management Association
,
Pittsburg, PA
.
Wolstenholme
,
P.
&
Finger
,
R.
1994
Long-Term Odor and VOC Performance Tests on Biofilters
.
Water Environment Federation
,
Alexandria, VA
.

Supplementary data