Removal of MIB and geosmin using granular activated carbon with and without MIEX pre-treatment

Журнал(книга):

Раздел науки:

Читать онлайн: 
Текст статьи: 
Mary Drikas*, Mike Dixon, Jim Morran Australian Water Quality Centre, South Australian Water Corporation, GPO Box 1751, Adelaide, South Australia 5001, Australia a r t i c l e i n f o Article history: Received 8 May 2009 Received in revised form 11 August 2009 Accepted 14 August 2009 Available online 21 August 2009 Keywords: Biodegradation Dissolved organic carbon Granular activated carbon Geosmin 2-methylisoborneol MIEX a b s t r a c t This study assessed the impact of MIEX pre-treatment, followed by either coagulation or microfiltration (MF), on the effectiveness of pilot granular activated carbon (GAC) filters for the removal of the taste and odour compounds, 2-methylisoborneol (MIB) and geosmin, from a surface drinking water source over a 2-year period. Complete removal of MIB and geosmin was achieved by all GAC filters for the first 10 months, suggesting that the available adsorption capacity was sufficient to compensate for differences in dissolved organic carbon (DOC) entering the GAC filters. Reduction of empty bed contact time (EBCT), in all but one GAC filter, resulted in breakthrough of spiked MIB and geosmin, with initial results inconclusive regarding the impact of MIEX pre-treatment. MIB and geosmin removal increased over the ensuing 12 months until complete removal of both MIB and geosmin was again achieved in all but one GAC filter, which had been pre-chlorinated. Autoclaving and washing the GAC filters had minimal impact on geosmin removal but reduced MIB removal by 30% in all but the prechlorinated filter, confirming that biodegradation impacted MIB removal. The impact of biodegradation was greater than any impact on GAC adsorption arising from DOC differences due to MIEX pre-treatment. It is not clear whether, at a lower initial EBCT, MIEX pre-treatment may have impacted on the adsorption capacity of the virgin GAC. The GAC filter maintained at the longer EBCT, which was also pre-chlorinated, completely removed MIB and geosmin for the period of the study, suggesting that the greater adsorption capacity was compensating for any decrease in biological degradation. ª 2009 Elsevier Ltd. All rights reserved. 1. Introduction Algal blooms have long been of concern to the drinking water industry due to the tastes and odours that are often associated with their presence. Of particular importance are the cyanobacteria or blue-green algae which may produce earthy– musty odours caused by the compounds 2-methylisoborneol (MIB) and geosmin. Both compounds are volatile saturated tertiary alcohols that can be detected at extremely low concentrations of between 10 and 20 ng/L. The low threshold of detection of these compounds can result in many customer complaints of taste and odour when other water quality characteristics, such as number of algal cells, are acceptable. Therefore much of the water industry concern has focussed on these particular compounds. Conventional treatment comprising coagulation/sedimentation/filtration can be very effective for the removal of algal cells when the process is optimised, but is ineffective for removal of dissolved algal metabolites released from these cells (Ando et al., 1992; Chow et al., 1998). Additional processes incorporating * Corresponding author. Tel.: þ61 8 7424 2110; fax: þ61 8 7003 2110. E-mail address: mary.drikas@sawater.com.au (M. Drikas). Available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/watres 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.08.016 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 oxidation or adsorption are generally required to remove these dissolved metabolites. Activated carbon has been shown to be an effective adsorbent for the removal of MIB and geosmin (Lalezary et al., 1986; Cook et al., 2001). Activated carbon can be applied as a powder in the early stages of the conventional treatment process when algal episodes are infrequent or, if algal blooms are a regular occurrence, it is more economic to include granular activated carbon (GAC) filters following the rapid filtration step in the conventional treatment process. The effectiveness of activated carbon for removal of contaminants can be impacted by the water quality of the contaminated water and particularly the presence of dissolved organic carbon (DOC) (Gillogly et al., 1999; Knappe et al., 1999; Newcombe and Cook, 2002; Li et al., 2003a,b; Quinlivan et al., 2005). DOC consists of a mixture of organic compounds of varying chemical characteristics within a wide range of molecular weight ranges that can also be adsorbed by activated carbon. For example, direct competition of the smaller DOC can occur with MIB/geosmin (molecular weight 169 and 182 g/mol respectively) for similar sized adsorption sites on the activated carbon. In addition, due to the higher concentration of DOC in water (mg/L) compared with the concentration of MIB/geosmin present (ng/L), it also likely that the larger DOC may adsorb onto larger pores and therefore prevent access of MIB/geosmin to the smaller sites that are thermodynamically preferred for adsorption (Newcombe et al., 1997; Cook et al., 2001; Newcombe et al., 2002b; Hepplewhite et al., 2004). Reduction of DOC prior to application of activated carbon has been shown to improve removal of MIB/geosmin by activated carbon (Newcombe and Cook, 2002; Newcombe et al., 2002b). The MIEX DOC process was developed specifically to remove DOC from raw water sources utilising a contact process rather than column filtration (Morran et al., 1996; Bursill et al., 2002). As the MIEX DOC process only removes dissolved organics, it is necessary to link the process with another technique to remove particulates. The application of this process in operating treatment plants is increasing (Drikas et al., 2003b; Hammann et al., 2004; Nestlerode et al., 2006; Warton et al., 2007) following the commissioning of the first MIEX plant at Mt Pleasant in South Australia in August 2001. The Mt Pleasant Water Treatment Plant (WTP) encompasses the MIEX DOC process and also enables comparison of two subsequent turbidity removal processes – conventional treatment (comprising coagulation, flocculation, sedimentation, and filtration) and submerged microfiltration (MF) (Drikas et al., 2003b). This has proven the effectiveness of operating MIEX in two possible scenarios – retrofitting into a conventional treatment plant or in a greenfield operation utilising MF to remove particulates. The successful operation of this plant has shown that considerable DOC removal can be achieved using MIEX combined with a second step for removal of turbidity. It was anticipated that the lower DOC resulting from MIEX treatment may improve the effectiveness of GAC filters for removal of problem contaminants and prolong filter life. To validate this theory and enable direct comparison of water quality with and without MIEX, pilot GAC filters were constructed and operated over a 2-year period at the Mt Pleasant WTP. The direct comparison of the efficiency for removal of MIB/geosmin by GAC following coagulation or MF with and without MIEX pre-treatment is discussed. 2. Materials and methods 2.1. Treatment processes In late July 2005, five pilot filters utilising coal based GAC (Calgon F400, Calgon Carbon Corporation. Pittsburgh, USA) were established with 20 min empty bed contact time (EBCT) to treat five water sources. Details of the treatment trains supplying feed waters to the GAC columns are summarised in Table 1. A conventional pilot plant without MIEX pre-treatment established on site at the Mt Pleasant WTP enabled direct comparison with the full scale operation of the MIEX pretreated water. The conventional pilot plant consisted of coagulation, flocculation, sedimentation and rapid filtration. The alum dose was chosen both using a model and jar tests to achieve the optimum DOC removal (defined as the point of diminishing return, where an additional 10 mg/L alum produces <0.1 mg/L DOC reduction) (van Leeuwen et al., 2005). The pH was not optimised but was between 6.5 and 6.8 throughout the study. The alum dose over the study period averaged 40 mg/L (as Al2(SO4)3$18H2O). Stream 1 at the Mt Pleasant WTP incorporates MIEX followed by conventional treatment (comprising coagulation, flocculation, sedimentation, rapid filtration). Stream 2 incorporates MIEX followed by CMF-S (submerged continuous microfiltration) with polyvinylidene fluoride (PVDF) membranes which have a nominal pore size of 0.04 mm. During the period July2005 to June2007, MIEX was applied at an average resin dose of 8 mL/L for 10 min contact time followed by sedimentation andremovaloftheresinbeforeenteringtheseparateparticulate removal processes. The resin was recirculated in a continuous process with 10% taken off for regeneration. Coagulation in Stream 1 during this period utilised an average alum dose of 6.5 mg/L (as Al2(SO4)3$18H2O) and 0.2 mg/L poly dimethyl diallyl ammonium chloride (DADMAC) as a coagulant aid. The MF pilot plant consisted of a single module CMF-S membrane pilot plant incorporating a suspended membrane module the same as that used in the Mt Pleasant WTP. Two separate membrane modules were used in the MF pilot Table 1 – Description of treatment trains supplying GAC filters. Conv Conventional treatment – pilot plant GAC 1 MIEX Coag WTP Stream 1 – MIEX treatment followed by conventional treatment GAC 2 Stream 2 WTP Stream 2 – MIEX treatment followed by submerged microfiltration GAC 3 Raw MF Raw water followed by submerged MF pilot plant GAC 4 MIEX MF WTP Stream 1 MIEX Treated water prior to coagulation followed by submerged MF pilot plant GAC 5 5152 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 plant unit in a one week on, one week off rotation. The source water for one module was raw water and the other MIEX treated water sourced from Stream 1 (prior to coagulation). This enabled direct comparison of the membrane’s performance with and without MIEX using identical operating and cleaning conditions. In June 2006, after 10 months operation of the GAC filters, the EBCT of the GAC filters was reduced to 5 min, in all but GAC filter 3, to accelerate the reduction in adsorption capacity. This was to accentuate differences in removal of algal metabolites, within the time frame of the 2-year study, as all GAC filters had achieved complete removal of MIB/geosmin in spiking trials undertaken to date. The reduced EBCT was achieved by removing all the GAC from the filters, mixing to ensure homogeneity, and placing a representative reduced volume of filter media back into each respective GAC filter. This was to ensure that the GAC removed was not taken from the top of each GAC filter, which would have reduced adsorption capacity as a result of the poorer water quality entering the filter. GAC 3, sourced from Stream 2 at Mt Pleasant, was retained at 20 min EBCT to enable a comparison of removal and enable an estimation of the life of the GAC filters at the longer contact time. 2.2. Analyses Samples were taken on a regular basis from the raw water and before and after GAC filtration. Samples for UV absorbance and DOC were filtered through 0.45 mm membranes. UV Absorbance (UV254) was measured at 254 nm through a 1 cm quartz cell and DOC was measured using a Sievers 820 Total Organic Carbon Analyser (GE Analytical Instruments, USA). Specific UV absorbance (SUVA) was calculated as (UV254  100)/DOC in/m/mg/L. Samples for MIB and geosmin analysis were pre-concentrated on a solid phase microextraction syringe fibre (Supelco, Australia) and concentrations determined on a gas chromatography–mass spectrometry system (Agilent Technologies, Australia) against quantified labelled internal standards (Ultrafine Chemicals, UK) according to the method by Graham and Hayes (1998). Reporting limit was 4 ng/L for MIB and 2 ng/L for geosmin. All results for geosmin and MIB removal are quoted as percentage removal, calculated from the difference between the concentration prior to and after GAC contact. As the spiking trials were undertaken over periods varying from several hours to several days, the values quoted are the average of the percent removal achieved over the trial period. The number of bacteria in water entering or leaving the GAC filters was enumerated using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA, USA) by the method of Hoefel et al. (2003). 2.3. Spiking trials The efficiency of each GAC filter for removal of MIB and geosmin was tested by spiking the feed water entering the GAC filters at regular intervals. Spiking was undertaken for periods extending between 5 h and 3 days. MIB and geosmin (Ultrafine Chemicals, UK) stock solutions were prepared using 10 mg of each compound dissolved in 200 mL MilliQ water. This stock solution was then diluted using MilliQ water and used to dose the inlet to the GAC filters to achieve concentrations between 50 and 200 ng/L of each compound. Initially, spiking trials were undertaken over 3 days with samples taken twice a day from all locations but, as the initial trials showed similar removal over the entire period, and to minimise analytical costs, subsequent trials were reduced to 5 h with only one sample taken before and after each GAC filter. 3. Results and discussion 3.1. Water supplying GAC filters Raw water quality over the study period from July 2005 to June 2007 consisted of turbidity ranging from 8.7 to 60 Nephelometric turbidity units, colour from 6 to 23 Hazen units and DOC from 2.7 to 5.8 mg/L. The water quality, as measured by DOC, in the raw water and achieved after the various treatment trains is summarised in Figs. 1 and 2. Fig. 1 illustrates the DOC of the water supplying the GAC filters after treatment with the coagulation processes (GAC 1 and GAC 2 respectively), using either conventional treatment with alum (Conv) or MIEX pre-treatment followed by coagulation (MIEX Coag). MIEX Coag consistently achieved lower DOC levels (ranging between 1.0 and 2.8 mg/L) than Conv (range 1.6–4.2 mg/L), over the period of the study. Consequently, the DOC remaining after GAC filtration was also lower for the MIEX Coag water (post GAC 2) than the Conv water (post GAC 1). Other studies have also shown improved DOC removal using MIEX compared with conventional treatment (Singer and Bilyk, 2002; Drikas et al., 2003a; Morran et al., 2004; Boyer and Singer, 2005; Warton et al., 2007; Jarvis et al., 2008). The reduction in EBCT of the GAC filters from 20 min to 5 min in June 2006 (after 10 months operation) is shown by a dashed line. The GAC adsorption capacity had already started to decline at this time and after the reduction in EBCT both GAC filters were only removing minimal DOC. However different DOC concentrations entering the GAC filters were maintained over the entire period of the study. Fig. 2 shows the DOC of water supplying the GAC filters after passage through the MF pilot plant (GAC 4 and GAC 5 respectively), either using raw water followed by MF (Raw MF) 6 5 4 3 2 1 0 17/7/05 25/10/05 2/2/06 13/5/06 21/8/06 29/11/06 9/3/07 17/6/07 DOC mg/L Raw Pre GAC 1 Post GAC 1 Pre GAC 2 Post GAC 2 Fig. 1 – DOC before and after GAC filtration (using coagulation processes) (line denotes change in GAC EBCT from 20 min to 5 min). w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 5153 or MIEX pre-treatment followed by MF (MIEX MF). The difference in DOC concentration entering the GAC filters due to MIEX pre-treatment is strongly accentuated with the MF streams, with a difference of at least 1 mg/L maintained for the period of the study. This difference is not surprising as the MF, with a nominal pore size of 0.04 mm, would be expected to remove only minimal DOC whereas both coagulation and MIEX treatment remove DOC. After the EBCT was reduced to 5 min, shown by a dashed line, GAC filtration again resulted in little further DOC removal in both streams, confirming that the adsorption capacity for DOC was minimal at this EBCT. The higher DOC concentration entering the GAC filters with the Conv and Raw MF train would be expected to result in a greater decrease in GAC adsorption capacity with time (Kilduff et al., 1998; Gillogly et al., 1999; Knappe et al., 1999; Li et al., 2003a,b; Yu et al., 2009) and also result in stronger direct competition for the removal of contaminants (Cook et al., 2001; Newcombe et al., 2002b; Shih et al., 2003). The different treatment trains resulted not only in different DOC concentrations but also different DOC character. This was illustrated by calculating the SUVA achieved for each of the treatment streams. SUVA has been linked with aromaticity and conjugation of DOC with higher SUVA waters showing better removal of DOC by coagulation (Archer and Singer, 2006) and better GAC adsorption (Kitis et al., 2001). The average of the SUVA of the water entering each of the GAC filters over the period of the study is summarised in Table 2. The raw water DOC concentration and character varied over the period of the study and this is apparent from the high standard deviation of the SUVA for the raw water. The MIEX pre-treatment resulted in lower SUVA with the MIEX Coag (GAC 2) producing slightly lower SUVA overall than the MIEX MF (GAC 5). This suggests that the MIEX treatment was able to remove more organics of an aromatic nature than that of conventional treatment (GAC 1). The raw MF (GAC 4) resulted in the least change in SUVA compared with the raw water, consequently having the water with the highest SUVA supplying the GAC filters. The impact of the different DOC concentrations, as a result of including MIEX pre-treatment, on the capacity of the GACs to remove MIB and geosmin was assessed. 3.2. Removal of MIB and geosmin Spiking trials of MIB and geosmin into the GAC filters were first undertaken in August 2005, a month after the GAC filters had been commissioned. MIB and geosmin, at concentrations between 50 and 200 ng/L, were dosed into each GAC filter. Five samples were taken over 3 days. Whilst there was some variation in the inlet concentrations over this period, no MIB and geosmin was detected in the outlet of any of the GAC filters. This was despite significant differences in DOC concentrations in water entering the GAC filters between the streams with and without MIEX treatment. Source water for GAC 1 and 2 had DOC concentrations of 2.7 mg/L and 2 mg/L respectively, whilst source water for GAC 4 and 5 had DOC concentrations of 4 mg/L and 2 mg/L respectively. It appears that when using virgin GAC and 20 min EBCT, the large amount of adsorption capacity available compensates for any difference in competition for adsorption sites caused by different DOC concentrations. Further spiking trials were undertaken after 5 and 9 months operation of the GAC filters and complete removal of all MIB and geosmin was still maintained. This was despite the continuous exposure of the GACs to significantly different DOC concentrations over the 9-month period. It is apparent from Figs. 1 and 2 that the uptake of DOC during this 9-month period had reduced the capacity of the GAC to continue to remove DOC, as evidenced from the increase in DOC post GAC for all GAC filters regardless of source water feeding the GAC. The difference in DOC concentrations pre and post GAC diminished with time over this period indicating reduced GAC adsorption capacity. This reduction can likely be attributed to both filling of pores of sizes suitable for adsorption of DOC and/or blocking of access to these pores by adsorption of larger DOC in larger transport pores (Pelekani and Snoeyink, 1999; Newcombe et al., 2002a; Hepplewhite et al., 2004; Li et al., 2003a,b; Quinlivan et al., 2005; Yu et al., 2009). It appears that this decrease in adsorption sites and, consequently, ability to remove DOC was not indicative of the capacity of the GAC to remove MIB and geosmin. After 9 months all GACs, while removing less DOC, continued to remove between 0.3 and 0.7 mg/L DOC whereas the concentration of MIB and geosmin removed was 200 ng/L. It appears that there was still sufficient adsorption capacity available in all GACs to enable adsorption of these comparatively small concentrations of MIB and geosmin. In June 2006, after approximately 10 months operation of the GAC filters, the EBCT of all the GAC filters, except GAC 3, was reduced to 5 min to accelerate the reduction in adsorption capacity. This was to accentuate differences between the GAC filters in removal of MIB/geosmin. GAC 3, sourced from the MIEX MF stream at Mt Pleasant (Stream 2), was retained at 20 min EBCT to enable a comparison of removal and an 6 5 4 3 2 1 0 17/7/05 25/10/05 2/2/06 13/5/06 21/8/06 29/11/06 9/3/07 17/6/07 DOC (mg/L) Raw Pre GAC 4 Post GAC 4 Pre GAC 5 Post GAC 5 Fig. 2 – DOC before and after GAC filtration (using microfiltration processes) (line denotes change in GAC EBCT from 20 min to 5 min). Table 2 – Average SUVA (/m/mg/L) of the water entering the GAC filters over the period of the study. Raw GAC 1 GAC 2 GAC 4 GAC 5 2.88  1.35 1.57  0.21 0.96  0.18 2.07  0.23 1.15  0.19 5154 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 estimation of the life of the GAC filters with the longer contact time. Fig. 3 illustrates the percentage removal obtained for both geosmin and MIB respectively, together with the standard deviation of the averaged data, in June 2006 following the reduction in EBCT for GAC 1, 2, 4 and 5. The lower EBCT was successful in reducing the removal of MIB and geosmin by the various GACs and all filters recorded between 80 and 90% removal of geosmin and between 50 and 80% removal of MIB; definitively less than 100% removal for both MIB and geosmin. GAC 3, with EBCT of 20 min, continued to remove all MIB and geosmin. The removal of geosmin by all GACs (Fig. 3a) was greater than that of MIB (Fig. 3b). Other studies have also observed the improved removal of geosmin (Lalezary et al., 1986; Cook et al., 2001; Newcombe and Cook, 2002). Although both molecules are of similar size, it has been suggested by Newcombe and Cook (2002) that the flatter shape of the geosmin molecule compared with the more rounded or boxlike MIB molecule enables it to access more thermodynamically favourable adsorption sites. The difference in DOC concentration of the source waters to the GAC filters was significant in June 2006 with GAC 1 receiving water containing 3.1 mg/L DOC and GAC 2 receiving water containing 1.8 mg/L DOC. Similarly GAC 4 was receiving water containing 3.7 mg/L DOC while GAC 5 was receiving only 1.7 mg/L DOC. Based on difference in DOC concentration, this suggests that GAC 2 and 5 should achieve the highest removal with GAC 1 and 4 having the poorest removal based on competition of geosmin and MIB with DOC for adsorption sites. Within experimental error, the most significant difference in removal of either MIB or geosmin was GAC 1 which achieved poorer removal than the other GACs and less than its MIEX treated equivalent, GAC 2; this was most evident for removal of MIB. This was consistent with expectations based on DOC differences in the source waters but inconsistent with GAC 4, which, although also receiving more DOC, did not show similar poor removal. In fact, while the removal of geosmin by GAC 2, 4 and 5 was similar, GAC 4 appeared to show better removal of MIB. This was unexpected based on the DOC concentration of the source water to GAC 4 but may have been impacted by the character of this water which had higher SUVA than the source waters of the other GACs. Cook et al. (2001) have previously shown that adsorption of MIB by activated carbon is very sensitive to the character and/or the concentration of the DOC whereas the adsorption of geosmin is not. Higher SUVA is indicative of larger more aromatic organics which would not be expected to compete as strongly with the smaller geosmin and MIB, hence resulting in increased removal (Newcombe et al., 1997; Newcombe et al., 2002b). However this contradicts the reduced adsorption of MIB observed for GAC 1 where the source water for GAC 1 also had higher SUVA than the MIEX treated waters. A number of simultaneous competing effects may have contributed to these observations; source water for GAC 4 with the highest DOC concentration (strongest competition) and highest SUVA (least competition) resulted in the best MIB adsorption, source waters to GAC 2 and 5 with the lowest DOC concentration (least competition) and lowest SUVA (strongest competition) resulted in medium MIB removal while source water to GAC 1 with the medium DOC concentration and medium SUVA resulted in the poorest MIB adsorption. It was anticipated that these differences would be further accentuated in the spiking trials undertaken over the next few months as the GAC adsorption capacity was further reduced (Fig. 3). However the September 2006 trial showed similar removals of geosmin for all GACs, within experimental error (Fig. 3a). The MIB data (Fig. 3b) showed a reversal in the previously observed trend with GAC 2 showing poorer MIB removal than GAC 1, and than the other GACs. By the next spiking trial, in December 2006, the difference was eliminated with all GACs behaving similarly with 100% MIB and geosmin removal achieved by February 2007, except for GAC 2. GAC 2 consistently performed worse for MIB removal than all the other GACs over this period. This was unexpected based on differences in DOC of the source waters feeding the GACs. GAC 4 received water with the highest DOC concentration followed by GAC 1 while GAC 5 and 2 received water with the least DOC. This would have been expected to lead to greater pore blockage for GAC 1 and 4 which had both received more DOC over their lifetime and also to more direct competition for adsorption sites with those waters. However this reduction in GAC adsorption capacity was only evident with GAC 1 in the June 2006 spiking trial and had been eliminated by the next spiking trial in September 2006. In addition there was an unexpected poor removal of MIB, and 60 80 100 May-06 Jul-06 Aug-06 Oct-06 Nov-06 Jan-07 Mar-07 May-06 Jul-06 Aug-06 Oct-06 Nov-06 Jan-07 Mar-07 % Geosmin removal GAC 1 GAC 2 GAC 4 GAC 5 40 60 80 100 % MIB Removal GAC 1 GAC 2 GAC 4 GAC 5 a b Fig. 3 – Percentage removal of geosmin and MIB by GAC filters from June 06 to February 07 (a) geosmin (b) MIB. w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 5155 to a lesser extent possibly geosmin, with GAC 2 in later spiking trials. The poorer adsorption by GAC 2 could conceivably be attributed to the presence in the source water of a smaller amount of DOC but which consisted of lower SUVA that could be expected to compete more strongly with the MIB for adsorption sites. However, this is inconsistent with the behaviour of GAC 5 which also had similar source water DOC concentration and character as GAC 2 but did not result in similar poor removal of MIB. Therefore this theory has been discarded as the explanation of the poor performance of GAC 2. Interestingly, geosmin removal within each GAC filter increased with time (Fig. 3a), rather than decreasing, with 100% removal again being achieved in all GAC filters after the additional 8 months. The removal of MIB (Fig. 3b) also showed an increased removal with time within each of the individual filters. This could not be attributed to loss of adsorption capacity of the GAC as this would result in decreased removal with time, not increased. Newcombe et al. (1996) reported that with an EBCT of 20 min, 18 months was the maximum time that their GAC could be expected to reduce MIB below the odour threshold. However, it is well established that MIB and geosmin can be biodegraded by a variety of micro-organisms (Izaguirre et al., 1988, Hoefel et al., 2006). In addition, a number of studies have shown that MIB and geosmin can be removed biologically within both sand and GAC filters (Hrudey et al., 1995; Elhadi et al., 2004; Elhadi et al., 2006). It appeared likely that biological degradation was responsible for the increased removal with time of these metabolites within each of the GAC filters. If biological degradation was responsible for the increased removal with time then the cause of the reduced removal in GAC 2 may also be a result of some biological impact. GAC 2 was sourced from Stream 1 at the Mt Pleasant WTP which incorporates MIEX followed by conventional treatment (MIEX Coag). Discussion with the plant operators identified that, without our knowledge, pre-chlorination was introduced into the flash mixing stage of the conventional stage of the treatment process in Stream 1 in February 2006. This was both to prevent algal growth through the treatment process and to decrease pH to improve coagulation as the low alum doses applied did not reduce pH to within the optimal coagulation range. Although no chlorine was detected at the inlet to GAC 2 when measured in February 2007 it was considered that prechlorination may have had an impact on the biological growth in GAC 2. Bacterial counts entering all the GAC filters in February 2007 were measured by flow cytometry and the results are summarised in Table 3. The bacterial counts entering GAC 2 were an order of magnitude lower than those entering the other GAC filters (except for GAC 3) thus establishing a biodegradation removal mechanism would have been much slower. This supported the view that less bacterial growth in GAC 2 may have been responsible for the poorer MIB removal observed through GAC 2. Bacterial counts entering GAC 3 were approximately half that entering the other GAC filters, excluding GAC 2. GAC 3 was sourced from the Mt Pleasant WTP Stream 2 which comprises MIEX followed by MF. Chlorine is dosed at the outlet of the balance tank supplying water to the MF unit to achieve residual chlorine of 0.2–0.3 mg/L on a continuous basis to prevent biofouling on the MF membranes. Again, although no chlorine was detected at the inlet to GAC 3, this may have contributed to the lower bacterial counts. To validate the cause of the reduced removal in GAC 2, sterilisation of the GAC filters was undertaken. The carbon was removed from each of the four GAC filters (with 5 min EBCT), autoclaved at 121 C for 15 min to inactivate the bacteria and then washed with MilliQ water to remove the bacteria. The GAC was then placed back into the respective filters and filtration restarted. The effectiveness of the autoclaving and washing process was assessed by measuring bacterial counts leaving each GAC filter prior to autoclaving and after removing, autoclaving and washing the GAC and allowing filtration to stabilise for 1 week. Fig. 4 shows that bacterial counts leaving the filters were high prior to autoclaving but that after autoclaving bacterial counts were significantly lower in all but GAC 2, sourced from Stream 1 at the Mt Pleasant WTP (MIEX Coag). While the lack of change in bacterial counts for GAC 2 could be considered to support the theory of lower bacterial growth in this filter the data may have been compromised as unfortunately on the date of sampling the GAC filters, Stream 1 at the plant was not flowing. This resulted in the source water to GAC 2 being supplied from stagnant rapid filters in the Stream 1 process and this may have impacted the results for GAC 2. However, as the other GAC filters had reduced bacterial counts, it was considered that the autoclaving process used had been successful. In particular, GAC 4 and 5 had bacterial counts reduced by 2 orders of magnitude. Spiking of both MIB and geosmin was undertaken prior to and after autoclaving. The results are shown in Fig. 5. Prior to autoclaving, all GAC filters, except GAC 2, were removing similar amounts of geosmin (Fig. 5a) (within 10%). After autoclaving, the percentage removal of geosmin for all GAC filters did not change significantly, with little difference Table 3 – Bacterial counts entering GAC filters. Bacterial counts (cells/mL)  105 GAC 1 2.37 GAC 2 0.36 GAC 3 1.34 GAC 4 5.24 GAC 5 8.29 1.00E+03 5.10E+04 1.01E+05 1.51E+05 2.01E+05 2.51E+05 3.01E+05 3.51E+05 4.01E+05 GAC 1 GAC 2 GAC 4 GAC 5 Bacterial Counts after GAC (cells/mL) Before autoclaving After autoclaving Fig. 4 – Bacterial counts leaving GAC filters before and after autoclaving. 5156 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 between GACs and less than 20% difference from that prior to autoclaving. This suggests that either autoclaving was not effective in inactivating the bacteria responsible for biodegradation or that biodegradation was not a significant cause of the removal of geosmin in these GAC filters. The high capacity for adsorption of geosmin by GAC is well established (Lalezary et al., 1986; Cook et al., 2001; Newcombe and Cook, 2002) so it is feasible that the GAC maintained considerable adsorption capacity for geosmin at this time. Persson et al. (2007) showed that even after almost four years, the GAC was still able to remove geosmin by adsorption. However, other studies on biological degradation of MIB and geosmin (Elhadi et al., 2004; Ho et al., 2007) have shown that geosmin is also preferentially removed by biodegradation. It is therefore also possible that there was a high population of bacteria responsible for geosmin degradation, and that the autoclaving and washing process was not effective in removing sufficient numbers of these organisms to reduce their impact on geosmin removal. As the removal of geosmin increased with time, it is clear that biodegradation was occurring but it is not possible to ascertain from this study the extent of removal of geosmin that can be attributed to biodegradation and/or adsorption. However the MIB results in Fig. 5b show a clear difference before and after autoclaving for all GACs except GAC 2. Prior to autoclaving, GAC 2 showed significantly less MIB removal than the other GACs. After autoclaving, the adsorption capacity was reduced by about 30% for GAC 1, 4 and 5 with removal then similar for all four GACs. This confirms that biological degradation was contributing significantly to the removal of MIB in GAC filters 1, 4 and 5 but much less to GAC 2. It also appears that these bacteria were not responsible for geosmin biodegradation and were easier to inactivate and remove by autoclaving and washing the GAC than any bacteria that may have been responsible for geosmin biodegradation. Another interesting aspect is that, assuming autoclaving was effective in inactivating bacteria, it appears that these carbons were continuing to remove significant amounts of MIB (70%) by adsorption. Persson et al. (2007) also showed that GAC was still able to remove MIB by adsorption even after almost four years operation. GAC 3 continued to remove all MIB/geosmin in these spiking trials. This GAC was sourced from Stream 2 of the Mt Pleasant WTP which comprised MIEX followed by submerged MF where chlorine was dosed to achieve residual chlorine of 0.2–0.3 mg/L on a continuous basis prior to the MF contact tank. The chlorine would also have been expected to have some impact on the biological degradation. Due to the large volume of GAC present in this filter, this GAC was not autoclaved so the difference before and after autoclaving could not be assessed. However, as this GAC continued to remove all MIB/geosmin, it was concluded that the larger surface area available at the longer EBCT (20 min) compensated for any decrease in biological degradation that may have been caused by chlorination prior to the GAC filtration. The impact of the reduction in EBCT on the adsorption capacity of the GAC filter is clear when comparing the DOC removed by the two GACs both receiving water which has been MIEX pre-treated followed by MF – that of GAC 3 (Stream 2) and that of GAC 5 (pilot plant MIEX MF). The EBCT of GAC 3 remained at 20 min for the duration of the study when the EBCT of the other GAC filters was reduced to 5 min. Fig. 6 clearly illustrates that, whilst the DOC removal for these two GAC filters was similar initially, after the reduction in EBCT for GAC 5, the DOC removal by GAC 3 remained higher than that by GAC 5. This supports the hypothesis that the increased EBCT in GAC 3 would compensate for reduced biological growth by providing more total surface area for adsorption and/or biodegradation of geosmin and MIB. The impact of increasing EBCT on removal of a range of compounds has been well established (Lee et al., 1981; Crittenden et al., 1993; Gillogly et al., 1999). At the EBCT of 20 min, sufficient adsorption sites were available on the virgin GAC filters to enable complete 0 25 50 75 100 % Geosmin Removal Before Autoclaving After Autoclaving 0 25 50 75 100 GAC 1 GAC 2 GAC 4 GAC 5 GAC 1 GAC 2 GAC 4 GAC 5 % MIB Removal Before Autoclaving After Autoclaving a b Fig. 5 – Percentage removal of geosmin and MIB by GAC filters before and after autoclaving (a) geosmin (b) MIB. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 17/7/05 25/10/05 2/2/06 13/5/06 21/8/06 29/11/06 9/3/07 17/6/07 DOC Removed by GAC (mg/L) GAC 3 GAC 5 Fig. 6 – DOC removed by GAC 3 and 5 (line denotes change in EBCT for GAC 5 from 20 min to 5 min). w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 5157 removal of MIB/geosmin, regardless of any difference in DOC entering each GAC. Once this carbon had been in use for some time and its capacity for DOC removal was virtually exhausted, the impact of MIEX pre-treatment on the adsorption of either MIB or geosmin by this particular carbon was unclear and/or its impact was less than that of biological degradation. It is not clear from this study whether, if the EBCT had been lower at the onset of the study, the resultant reduced adsorption capacity of the virgin GAC may have been impacted by the MIEX pre-treatment and the difference in DOC. 4. Conclusion This study showed that:  MIEX pre-treatment consistently achieved higher DOC removal than the treatment trains without MIEX over the entire period of the study, both before and after GAC filtration.  MIB and geosmin, spiked at concentrations between 50 and 200 ng/L, were totally removed by all GAC filters for the first 10 months of the study indicating that the available adsorption capacity at an EBCT of 20 min was sufficient to compensate for any difference in DOC entering the GAC filters.  At the reduced EBCT of 5 min, MIB/geosmin breakthrough occurred with initial results providing contradictory results regarding the impact of MIEX pre-treatment.  Further spiking trials showed an increase in removal efficiency with time in all GAC filters with similar removal achieved in all GAC filters, except GAC 2. This suggests that biodegradation was responsible for the increased removal. GAC 2, receiving pre-chlorinated water, had bacterial counts an order of magnitude lower than those entering the other GAC filters thus establishing a biodegradation removal mechanism would be much slower.  Spiking trials undertaken before and after autoclaving the filters had minimal impact on geosmin removal, suggesting either the removal mechanism was predominantly adsorption and/or the bacteria responsible for biodegradation were not removed by the autoclaving process used in this study  Autoclaving reduced removal of MIB by about 30% in all filters except GAC 2, confirming that biological degradation was contributing significantly to the removal of MIB in all GAC filters, except GAC 2. It also appears that these carbons were continuing to remove significant amounts of MIB (70%) by adsorption.  The impact of biodegradation was greater than differences in DOC resulting from MIEX pre-treatment. It is not clear whether, at a reduced EBCT, MIEX pre-treatment may have impacted on the removal of MIB and geosmin by the virgin GAC.  GAC 3, also pre-chlorinated, continued to remove all MIB/ geosmin suggesting that the greater adsorption capacity at the EBCT of 20 min was compensating for any decrease in biological degradation that may have been caused by chlorination prior to GAC filtration. r e f e r e n c e s Ando, A., Miwa, M., Kajino, M., Tatsumi, S., 1992. Removal of musty-odorous compounds in water and retained in algal cells through water purification processes. Water Science and Technology 25 (2), 299–306. Archer, A.D., Singer, P.C., 2006. An evaluation of the relationship between SUVA and NOM coagulation using the ICR database. Journal of the American Water Works Association 98 (7), 110–123. Boyer, T.H., Singer, P.C., 2005. Bench-scale testing of a magnetic ion exchange resin for removal of disinfection by-product precursors. Water Research 39 (7), 1265–1276. Bursill, D., van Leeuwen, J., Drikas, M., 2002. Problems related to particulate and dissolved components in water: the importance of organic matter. Water Science and Technology: Water Supply 2 (1), 155–162. Chow, C.W.K., House, J., Velzeboer, R.M.A., Drikas, M., Burch, M.D., Steffensen, D.A., 1998. The effect of ferric chloride flocculation on cyanobacterial cells. Water Research 32 (3), 808–814. Cook, D., Newcombe, G., Sztajnbok, P., 2001. The application of powdered activated carbon for MIB and geosmin removal: predicting PAC doses in four raw waters. Water Research 35 (5), 1325–1333. Crittenden, J.C., Vaitheeswaran, K., Hand, D.W., Howe, E.W., Aieta, E.M., 1993. Removal of dissolved organic carbon using granular activated carbon. Water Research 27 (4), 715–721. Drikas, M., Chow, C., Cook, D., 2003a. The impact of recalcitrant organic character on disinfection stability, trihalomethane formation and bacterial regrowth: an evaluation of magnetic ion exchange resin (MIEX) and alum coagulation. Journal of Water Supply: Research and Technology – Aqua 52 (7), 475–487. Drikas, M., Morran, J.Y., Cook, D., Bursill, D.B., 2003b. Operating the MIEX process with microfiltration or coagulation. In: Proceedings of the American Water Works Association Water Quality Technology Conference, Philadelphia, USA. Elhadi, S.L.N., Huck, P.M., Slawson, R.M., 2004. Removal of geosmin and 2-methylisoborneol by biological filtration. Water Science and Technology 49 (9), 273–280. Elhadi, S.L.N., Huck, P.M., Slawson, R.M., 2006. Factors affecting the removal of geosmin and MIB in drinking water biofilters. Journal of the American Water Works Association 98 (8), 108–119. Gillogly, T.E.T., Snoeyink, V.L., Vogel, J.C., Wilson, C.M., Royal, R.P., 1999. Determining GAC bed life. Journal of the American Water Works Association 91 (8), 98–110. Graham, D., Hayes, K.P., 1998. Application of solid phase microextraction for the analysis of off-flavours in water. In: Proceedings of the Australian Water Association WaterTECH Conference, Brisbane, Australia. Hammann, D., Bourke, M., Topham, C., 2004. Field report – evaluation of a magnetic ion exchange to meet DBP regs at the village of Palm Springs. Journal of the American Water Works Association 96 (2), 46–50. Hepplewhite, C., Newcombe, G., Knappe, D.R.U., 2004. NOM and MIB, who wins in the competition for activated carbon adsorption sites? Water Science and Technology 49 (9), 257–265. Ho, L., Hoefel, H., Bock, F., Saint, C.P., Newcombe, G., 2007. Biodegradation rates of 2-methylisoborneol (MIB) and geosmin through sand filters and in bioreactors. Chemosphere 66 (11), 2210–2218. Hoefel, D., Grooby, W.L., Monis, P.T., Andrews, S., Saint, C.P., 2003. Enumeration of water-borne bacteria using viability assays and flow cytometry: a comparison to culture-based techniques. Journal of Microbiological Methods 55 (3), 585–597. 5158 w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 Hoefel, D., Ho, L., Aunkofer, W., Monis, P.T., Keegan, A., Newcombe, G., Saint, C.P., 2006. Cooperative biodegradation of geosmin by a consortium comprising three gram-negative bacteria isolated from the biofilm of a sand filter column. Letters in Applied Microbiology 43 (4), 417–423. Hrudey, S.E., Huck, P.M., Mitton, M.J., Kenefick, S.L., 1995. Evaluation of odour removal by pilot-scale biological treatment process trains during spring runoff in an icecovered river. Water Science and Technology 31 (11), 195–201. Izaguirre, G., Wolfe, R.L., Means, E.G., 1988. Bacterial degradation of 2-methylisoborneol. Water Science and Technology 20 (8/9), 205–210. Jarvis, P., Mergen, M., Banks, J., McIntosh, B., Parson, S.A., Jefferson, B., 2008. Pilot scale comparison of enhanced coagulation with magnetic resin plus coagulation systems. Environmental Science and Technology 42 (4), 1276–1282. Kilduff, J.E., Karanfil, T., Weber Jr., W.J., 1998. TCE adsorption by GAC preloaded with humic substances. Journal of the American Water Works Association 90 (5), 76–89. Kitis, M., Karanfil, T., Kilduff, J.E., Wigton, A., 2001. The reactivity of natural organic matter to disinfection byproducts formation and its relation to specific ultraviolet absorbance. Water Science and Technology 43 (2), 9–16. Knappe, D.R.U., Snoeyink, V.L., Prados, P.R., Prados, M.J., Bourbigot, M.M., 1999. Atrazine removal by preloaded GAC. Journal of the American Water Works Association 91 (10), 97–109. Lalezary, S., Pirbazari, M., McGuire, M.J., 1986. Evaluating activated carbons for removing low concentrations of taste and odor producing organics. Journal of the American Water Works Association 78 (11), 76–82. Lee, M.C., Snoeyink, V.L., Crittenden, J.C., 1981. Activated carbon adsorption of humic substances. Journal of the American Water Works Association 73 (8), 440–446. Li, Q., Snoeyink, V.L., Marinas, B.J., Campos, C., 2003a. Elucidating competitive adsorption mechanisms of atrazine and NOM using model compounds. Water Research 37 (4), 773–784. Li, Q., Snoeyink, V.L., Marinas, B.J., Campos, C., 2003b. Pore blockage effect of NOM on atrazine adsorption kinetics of PAC: the roles of PAC pore size distribution and NOM molecular weight. Water Research 37 (20), 4863–4872. Morran, J.Y., Bursill, D.B., Drikas, M., Nguyen, H., 1996. A new technique for the removal of natural organic matter. In: Proceedings of the Australian Water Association WaterTECH Conference, Sydney, Australia. Morran, J.Y., Drikas, M., Cook, D., Bursill, D.B., 2004. Comparison of MIEX treatment and coagulation on NOM character. Water Science and Technology: Water Supply 4 (4), 129–137. Nestlerode, F., Teply, M., Bourke, M., 2006. Field report – Green Valley upgrade gets to the root of disinfection by-product problem. Journal of the American Water Works Association 98 (10), 58–60. Newcombe, G., Collett, A., Drikas, M., Roberts, B., 1996. Granular activated carbon pilot plant studies. Water 23 (3), 29–31. Newcombe, G., Cook, D., 2002. Influences on the removal of tastes and odours by PAC. Journal of Water Supply: Research and Technology – Aqua 51 (8), 463–474. Newcombe, G., Drikas, M., Hayes, R., 1997. The Influence of characterised natural organic material on activated carbon adsorption: II. Effect of pore volume distribution and adsorption of MIB. Water Research 31 (5), 1065–1073. Newcombe, G., Morrison, J., Hepplewhite, C., 2002a. Simultaneous adsorption of MIB and NOM onto activated carbon. I. Characterisation of the system and NOM adsorption Carbon. 40 (12), 2135–2146. Newcombe, G., Morrison, J., Hepplewhite, C., Knappe, D.R.U., 2002b. Simultaneous adsorption of MIB and NOM onto activated carbon: II. Competitive effects. Carbon 40 (12), 2147–2156. Pelekani, C., Snoeyink, V.L., 1999. Competitive adsorption in natural water: role of activated carbon pore size. Water Research 33 (5), 1209–1219. Persson, F., Heinicke, G., Hedberg, T., Hermansson, M., Uhl, W., 2007. Removal of geosmin and MIB by biofiltration – an investigation discriminating between adsorption and biodegradation. Environmental Technology 28 (1), 95–104. Quinlivan, P.A., Li, L., Knappe, R.U.D., 2005. Effects of activated carbon characteristics on the simultaneous adsorption of aqueous organic micropollutants and natural organic matter. Water Research 39 (8), 1663–1673. Shih, T.C., Wangpaichitr, M., Suffet, M., 2003. Evaluation of granular activated carbon technology for the removal of methyl tertiary butyl ether (MTBE) from drinking water. Water Research 37 (2), 375–385. Singer, P.C., Bilyk, K., 2002. Enhanced coagulation using a magnetic ion exchange resin. Water Research 36 (16), 4009–4022. van Leeuwen, J., Daly, R., Holmes, M., 2005. Modelling the treatment of drinking water to maximize dissolved organic matter removal and minimize disinfection by-product formation. Desalination 176 (1–3), 81–89. Warton, B., Heitz, A., Zappia, L.R., Franzmann, P.D., Masters, D., Joll, C.A., Alessandrino, M., Allpike, B., O’Leary, B., Kagi, R.I., 2007. Magnetic ion exchange drinking water treatment in a large-scale facility. Journal of the American Water Works Association 99 (1), 89–101. Yu, Z., Peldszus, S., Huck, P.M., 2009. Adsorption of selected pharmaceuticals and an endocrine disrupting compound by granular activated carbon. 1. Adsorption capacity and kinetics. Environmental Science and Technology 43 (5), 1467–1473. w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 1 5 1 – 5 1 5 9 5159

Комментировать