Chemosphere 136 (2015) 198–203
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Technical Note
Removing organic and nitrogen content from a highly saline municipal wastewater reverse osmosis concentrate by UV/H2O2–BAC treatment Shovana Pradhan, Linhua Fan ⇑, Felicity A. Roddick School of Civil, Environmental and Chemical Engineering, RMIT University, Melbourne, Victoria 3001, Australia
h i g h l i g h t s UV/H2O2–BAC treatment of a highly saline wastewater ROC was investigated. Approximately 60% DOC and TN was removed from the ROC by the process. Residual H2O2 in BAC feed enhanced DOC removal but limited TN removal.
a r t i c l e
i n f o
Article history: Received 14 March 2015 Received in revised form 7 May 2015 Accepted 12 May 2015 Available online 22 May 2015 Keywords: Reverse osmosis concentrate Biological activated carbon Advanced oxidation processes Organic matter removal Nitrogen removal Wastewater reclamation
a b s t r a c t Reverse osmosis (RO) concentrate (ROC) streams generated from RO-based municipal wastewater reclamation processes pose potential health and environmental risks on their disposal to confined water bodies such as bays. A UV/H2O2 advanced oxidation process followed by a biological activated carbon (BAC) treatment was evaluated at lab-scale for the removal of organic and nutrient content from a highly saline ROC (TDS 16 g L 1, EC 23.5 mS cm 1) for its safe disposal to the receiving environment. Over the 230-day operation of the UV/H2O2–BAC process, the colour and UV absorbance (254 nm) of the ROC were reduced to well below those of the influent to the reclamation process. The concentrations of DOC and total nitrogen (TN) were reduced by approximately 60% at an empty bed contact time (EBCT) of 60 min. The reduction in ammonia nitrogen by the BAC remained high under all conditions tested (>90%). Further investigation confirmed that the presence of residual peroxide in the UV/H2O2 treated ROC was beneficial for DOC removal, but markedly inhibited the activities of the nitrifying bacteria (i.e., nitrite oxidising bacteria) in the BAC system and hence compromised total nitrogen removal. This work demonstrated that the BAC treatment could be acclimated to the very high salinity environment, and could be used as a robust method for the removal of organic matter and nitrogen from the pre-oxidised ROC under optimised conditions. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Over the past decade, reverse osmosis (RO) technology has been used increasingly for full-scale municipal wastewater reclamation plants due to its ability to produce high quality effluent and increasing affordability (Shannon et al., 2008). However, management of the concentrate (also referred to as the brine) streams generated from the RO-based wastewater reclamation processes remains a major challenge for the water industry in applying the technology. The RO concentrate (ROC) contains almost all of the organic contaminants and nutrients from the secondary effluent at elevated levels (i.e., typically 4–6 times higher in concentration). The organic content of the ROC commonly includes ⇑ Corresponding author. E-mail address: [email protected] (L. Fan). http://dx.doi.org/10.1016/j.chemosphere.2015.05.028 0045-6535/Ó 2015 Elsevier Ltd. All rights reserved.
pharmaceuticals, personal care products, pesticides, endocrine disruptors, disinfection by products and other organic species; many of these are toxic and bio-accumulative. Moreover, introduction of excess nutrients to the receiving environment can cause adverse effects such as blooms of harmful algae. The disposal of the untreated ROC to a confined water body (e.g., bays) consequently poses significant risks to the health of humans and ecosystems. Due to the recalcitrant nature of the organic matter in the ROC, several oxidative treatment methods including advanced oxidation processes (AOPs) have been studied as a means of removing and/or enhancing the biodegradability of the organic content. In some recent studies, the homogeneous AOP utilising UVC/H2O2 has been demonstrated as an effective method for degrading the organic matter in the ROC and improving its biodegradability for subsequent biological treatment. One study showed the process (initial H2O2 concentration 11.5 mM, 120 min irradiation) could achieve
199
S. Pradhan et al. / Chemosphere 136 (2015) 198–203
complete decolourisation and 55% reduction in COD, 38% in DOC and 32% in dissolved organic nitrogen (DON) (Bagastyo et al., 2011). Liu et al. (2012) reported that more than 80% DOC removal could be achieved for a municipal wastewater ROC with UVC/H2O2 treatment (initial H2O2 concentration 4 mM, 30 min irradiation), followed by a biodegradable DOC (BDOC) determination as a surrogate biological treatment. The bench-scale study revealed that such treatment was effective for degrading the organic compounds over a wide range of salinity (EC 4.5–11.2 mS cm 1). Based on these studies, low-energy requirement and small-footprint biological processes should be considered as the subsequent treatments to improve the total organic removal and hence the cost-effectiveness of the oxidative treatment processes. Biological activated carbon (BAC) processes can provide simultaneous adsorption of non-biodegradable matter and oxidation of biodegradable matter in a single reactor with microbial activity in a granular activated carbon system (Walker and Weatherley, 1999). The use of BAC to treat pre-oxidised ROC has been reported in several studies. Lee et al. (2009) found that the coupling of a BAC column with ozone pre-oxidation enhanced the organic removal efficiency by 3 times that of BAC alone at an empty bed contact time (EBCT) of 60 min, with 70% TOC removed from the ROC (TDS 1.2 g L 1, EC 2 mS cm 1). In a more recent study, Lu et al. (2013) observed that the integrated treatment of UV/H2O2 followed by BAC removed 59% DOC, 64% colour and 78% UV absorbance at 254 nm (UVA254) from a ROC (TDS 10 g L 1, EC 13.5 mS cm 1). However, the removal efficiencies for nutrients such as total nitrogen (TN) and total phosphate (TP) were fairly low, with only 24% of TN and 17% of TP removed. In an earlier study, Ng et al. (2008) found that BAC treatment following capacitive deionisation could achieve a higher TN removal (91%) from a ROC with relatively low salinity (EC 2 mS cm 1). The low TN removal observed by Lu et al. (2013) may be related to the high salinity of the ROC, as it is well recognised that high salinity of wastewater can greatly affect biological activity, resulting in decreased nitrification and denitrification, thereby affecting overall TN removal (Kargi and Dinçer, 1997). The biological removal of nitrogen at elevated salt concentration can be challenging due to the sensitivity of nitrogen-removing microorganisms to high salt conditions, along with other environmental conditions including temperature, DO concentration, pH, ammonia concentration, heavy metals and C/N ratio (Okabe et al., 1996). However, it is possible for the microorganisms to be acclimated to high salinity environments. The ROC salinity level may be classified as low (TDS < 5 g L 1), medium (5–10 g L 1), high (10–15 g L 1) and extremely high (>15 g L 1). To date, most of the published work has been on ROC with low or medium salinity. Investigation of the removal of organic content and nutrients (particularly nitrogen) from ROC with extremely high level of salinity using UV/H2O2 followed by BAC has not been reported. The aim of this study was to investigate the effectiveness and robustness of a UV/H2O2–BAC process for removing organic matter and nitrogen from a municipal wastewater ROC with extremely high salinity over an extended period of operation (230 days). The impact of the residual H2O2 in the BAC influent on the bio-treatment performance was also studied. The treatment effectiveness was characterised in terms of reductions in DOC, UVA254, colour, COD, total nitrogen (TN), ammonium nitrogen (NH4–N), nitrate nitrogen (NO3–N) and nitrite nitrogen (NO2–N). 2. Materials and methods 2.1. Source of ROC The ROC samples were collected from a local wastewater reclamation facility in which the biologically treated secondary effluent was treated by a UF (0.04 lm)–RO system to remove salts and
other contaminants to produce recycled wastewater. The water recovery of the RO system was approximately 75%. The general characteristics of the ROC samples collected during January to October, 2013 are given in Table 1. The secondary effluent had a high TDS level due to the infiltration of salty groundwater into the sewer system, leading to the extremely high salinity of the ROC. The collected samples were stored at 4 °C and brought to room temperature before use. 2.2. UV/H2O2 treatment of ROC Irradiation was conducted in batch mode using an annular reactor with a centrally mounted lamp. The UVC lamp (254 nm) had an energy input of 39 W (Australian Ultra Violet Services, G36T15NU), and the average fluence was determined as 8.9 mJ s 1 cm 2. During irradiation, samples were mixed and aerated by humidified air. The average irradiation area was 464 cm2 with a path length of 1.94 cm and other UV reactor conditions are reported elsewhere (Liu et al., 2012). The ROC sample (900 mL) was dosed with 3 mM H2O2 and then exposed to UV irradiation for 30 min (UV dose 16 J cm 2) as suggested by Liu et al. (2012) as the optimum conditions for treating the ROC. The treated ROC was then fed to BAC columns under the defined experimental conditions. The residual H2O2 in the UV/H2O2 treated ROC was measured using a photometric method at 551 nm (Bader et al., 1988). The enzyme catalase derived from bovine liver (Sigma, 3691 units per mg solids) was used to quench the residual H2O2 for examining the effect of residual peroxide on the efficiency of BAC treatment. For every 25 mL of the water sample, 10 lL (4000 U mL 1) of catalase was added to remove the residual H2O2. Catalase was also used to decompose residual H2O2 to remove its interference in water quality analyses. The samples were shaken at 100 rpm at room temperature until the concentration of H2O2 was less than 0.5 mg L 1. 2.3. BAC start-up and operation Granular activated carbon (GAC) columns with an inner diameter of 1.5 cm and an effective packing height of 12 cm were set up. Coal-based activated carbon (Activated Carbon Technologies Pty Ltd, Australia) with effective size of 1.2–1.4 mm, density 0.20– 0.30 g cm 3 and surface area >1200 m2 g 1 was used. The carbon was sieved to remove the very fine particles, washed repeatedly with deionised water until the fine particles were removed, and then dried in an oven at 110 °C for 2 days. The carbon media (200 g) was then inoculated by mixing with activated sludge (1 L, MLSS 4000 mg L 1) obtained from the wastewater treatment plant which supplied the secondary effluent to the reclamation facility for 4 days. Nutrients including N, P and C sources were added to the system to promote the growth of microorganisms during the inoculation period (glucose 0.78 g L 1, ammonium chloride
Table 1 Characteristics of the ROC samples (January to October, 2013). Parameter
Value
pH DO (mg L 1) DOC (mg L 1) UVA254 (cm 1) Colour (mg Pt–Co L 1) COD (mg L 1) TN (mg L 1) TP (mg L 1) TDS (g L 1) Chloride (mg L 1) Conductivity (mS cm 1)
7.7 ± 0.4 10.2 ± 1.3 36.0 ± 4.0 0.62 ± 0.02 148 ± 10.0 120 ± 19 21.4 ± 4.5 28.5 ± 1.1 16.6 ± 0.8 7,700 ± 1,800 23.5 ± 1.3
200
S. Pradhan et al. / Chemosphere 136 (2015) 198–203
0.11 g L 1, potassium dihydrogen phosphate 0.033 g L 1) (Lu et al., 2013). The media was then transferred into glass columns which were then fed with UV/H2O2-treated ROC in downflow mode. The TDS of UV/H2O2 treated ROC was increased gradually from 4 to 16 g L 1 over four weeks; DOC, UVA254 and colour of the BAC effluent were continuously monitored until steady values were achieved. The BAC columns were then used for the study on the ROC with original salinity. All BAC tests were run at room temperature (22–26 °C). Two identical BAC columns were operated with EBCT of 60 min on raw and UV/H2O2 treated ROC, respectively, over a period of 230 days. The flow rate of the effluent from the BAC columns was monitored daily and regulated using the outlet valve to maintain the fixed exposure time. Backwashing was carried out every 2 weeks to control the biomass growth and prevent media clogging. The effluent samples were collected periodically (every 3–5 days) and analysed for DOC, COD, UVA254, colour, TN, NH+4–N, NO2 –N, NO3 –N and TP. 2.4. Analytical methods DOC was measured using a Shimadzu TOC-L analyser. Prior to DOC analysis, the samples were diluted (1 in 5) with MilliQ water in order to reduce the possible impact of high TDS of the ROC on the measurement. A spectrophotometer (DR5000, Hach, USA) was used to determine the COD, colour, TN and TP of the water samples according to the Hach Methods Manual. For COD measurement, 0.5 g of mercury sulphate (Sigma) was added to overcome the interference of high chloride concentration as suggested in the standard method (APHA, 2005). UV absorbance was measured at 254 nm (UVA254) using a Shimadzu UV–vis spectrophotometer (UV2-2700). TDS was determined in accordance with Standard Method 2540C (APHA, 2005). The DO and conductivity was measured with a Hach DO meter (LDO 101) and Hach conductivity meter (CDC 401), respectively. The pH was measured with a Hach Sension 156 pH meter. NH+4–N was measured colorimetrically at 660 nm via indophenol formation with sodium salicylate as described by Verdouw et al. (1978), NO3 –N was measured colorimetrically at 420 nm with the sodium salicylate method described by Scheiner (1974) and NO2 –N was measured using Standard Method 4500B (APHA, 2005). All analyses were duplicated and results recorded in terms of mean values and errors/standard deviations. 3. Results and discussion 3.1. Removal efficiencies for organic matter and nutrients by the UV/ H2O2–BAC treatment The UV/H2O2–BAC treatment system was operated for 230 days (February to October, 2013) with EBCT of 60 min on multiple
batches of ROC sample to evaluate its efficiency for the removal of organics and nutrients (Table 2). As a reference, the raw ROC was treated by BAC alone with an identical column under the same operating conditions. After UV/H2O2 pre-treatment of the ROC, there were marked reductions in UVA254 and colour (60% and 86%, respectively), whereas only 15% reduction was achieved for DOC and COD, respectively. The UVA254 reflects the presence of conjugated bonds and aromatic content of organic matter (Chen et al., 2011). The result was consistent with previous studies which showed great cleavage but limited mineralisation of the organic matter in the ROC using such treatment (Umar et al., 2013). For the BAC alone treatment, the removal efficiency for DOC and COD was considerably higher (38% and 32%, respectively) than for the UV/H2O2 treatment alone, although the reductions in UVA254 and colour (64% and 80%, respectively) were comparable for the two treatments. The reduction in DOC by the BAC alone was associated with biodegradation and adsorption. Although the organic molecules in the ROC are generally regarded as low in biodegradability as they were primarily derived from the biologically treated secondary effluent, it was possible that some of the molecules were removed by the microbes with sufficient contact time (Yapsakli and Çeçen, 2010). Using the combined treatment, the removal efficiency for the organic matter was increased significantly, with 57% DOC, 46% COD, 81% UVA254 and 95% colour removed from the ROC. The great improvement in organic matter removal was mainly attributed to the partial degradation of the organic molecules by the oxidative treatment, leading to the production of simpler molecules which could be readily removed by the microbes embedded in the BAC column (Lu et al., 2013). A comparison of the organic matter removal efficiency obtained in the present study and the work done by Lu et al. (2013) using the same treatment train showed that the treatment performance was fairly comparable, although the ROC salinity was markedly different. This might indicate that the change in the TDS of ROC in the range of 10–16 g L 1 did not have any significant impact on the organic matter removal. It is worth noting that the UV/H2O2–BAC treated ROC had similar DOC and COD levels to the secondary effluent used as the influent of the RO-based reclamation process (i.e., 15 cf. 14.5 mg DOC L 1 and 65 cf. 64 mg COD L 1), but significantly lower UVA254 and colour (i.e., 0.12 cf. 0.19 cm 1 and 7 cf. 120 mg Pt–Co L 1). In terms of nutrient removal, minimal reduction was obtained by the UV/H2O2 treatment. The small reduction in TN during the oxidation process was likely due to the oxidation of some nitrogen species to nitrogen gas (Dwyer et al., 2008). With BAC treatment alone, the removal efficiency was fairly high for TN (71%) but still low for TP (8%). TN reduction for the combined treatment system was lower (60%) compared with the BAC only treatment, although TP reduction was improved slightly. This study showed a higher TN removal (60%) compared with the study by Lu et al. (2013), in which the reduction in TN was only 24% for the combined
Table 2 Overall water quality characteristics of the ROC after the various treatments. Treatment Parameter
Raw ROC
UV/H2O2
BAC
UV/H2O2 + BAC
pH DO (mg L 1) DOC (mg L 1) COD (mg L 1) UVA254 (cm 1) Colour (mg Pt–Co L TN (mg L 1) TP (mg L 1)
7.7 ± 0.4 10.2 ± 1.3 35.0 ± 5.0 120.0 ± 9.0 0.623 ± 0.025 138.0 ± 10.0 21.0 ± 3.0 29.0 ± 2.0
8.0 ± 0.2 12.1 ± 0.9 30.0 ± 4.0 (15%) 102.0 ± 8.0 (15%) 0.246 ± 0.033 (60%) 20.0 ± 8.0 (86%) 18.0 ± 3.0 (13%) 28.0 ± 2.0 (2%)
8.4 ± 0.2 7.0 ± 0.9 21.0 ± 4.0 (38%) 82.0 ± 11.0 (32%) 0.23 ± 0.06 (64%) 28.0 ± 8.0 (80%) 6 ± 1.0 (71%) 26.0 ± 1.0 (8%)
8.4 ± 0.2 7.2 ± 1.2 15.0 ± 2.0 (57%) 65.0 ± 7.0 (46%) 0.12 ± 0.03 (81%) 7.0 ± 5.0 (95%) 9.0 ± 2.0 (60%) 24.0 ± 1.0 (15%)
1
)
Note: Results were based on the 57 samples collected during February to October, 2013; values are presented as mean ± 1 standard deviation, average removal efficiency (%) is shown in the bracket.
201
S. Pradhan et al. / Chemosphere 136 (2015) 198–203
treatment system with the same EBCT for the BAC treatment. The higher removal of TN in the present study may be related to the use of a lower H2O2 concentration (3 mM cf. 4 mM), and thus less residual H2O2 after the oxidative process. The impact of residual H2O2 on the removal of organic content and nitrogen was therefore investigated further in this work and reported in Section 3.3. Nevertheless, the TP removal was comparable in both studies (15% in this study) for the combined treatment system. It is known that phosphorus removal is greatly inhibited by increasing salinity (>5 g L 1) (Uygur and Kargı, 2004). As the salinity of ROC used in both studies was very high (TDS 16.6 g L 1 and 10 g L 1), lower removal of TP would then be anticipated. Since chemical coagulation is generally an effective means for TP removal, it could be employed prior to the oxidative treatment for achieving the required removal efficiency for TP, and additional reduction in organic content. The DO of the effluent was 6 mg L 1 for the combined and the BAC alone treatment, which was lower than for the influent (10 mg L 1), suggesting that the activated carbon adsorbed DO and the adsorbed DO was utilised by the microorganisms to biodegrade the contaminants (Jin et al., 2013). During the 230-day operation of the combined treatment, six batches of ROC sample were used for the study. The TDS of the ROC samples varied from 15.6 to 18.2 g L 1 and the DOC varied from 32 to 38 mg L 1. It was observed that the average DOC removal efficiency of the combined treatment over the ROC sample batches varied only slightly (54–59%), whereas the TN removal efficiency varied a little more (55–65%). This indicates that the UV/H2O2–BAC process was a robust treatment system for the ROC for organic matter and nitrogen removal.
3.2. Characterisation of nitrogen removal To obtain a better understanding of the nitrogen removal, the treatments were characterised in terms of the removal efficiency of TN, NH+4–N, NO2 –N and NO3 –N over a 50 days period (August to October, 2013) (Fig. 1). The UV/H2O2 treatment alone was not very effective for removing total nitrogen, with only 15% of TN removed. NO2 –N was removed almost completely, whereas NO3 –N concentration increased by 1–2 mg L 1 after the oxidation treatment. The increase in NO3 –N concentration was attributed to the conversion of NO2 –N to NO3 –N in the presence of oxidant as suggested by Munter (2001). Since NH+4–N has no unsaturated double bonds, it would be difficult for the_OH radical to break down the N–H bonds through electrophilic attack (Haag et al., 1984), hence the slight reduction in the NH+4–N during the treatment was probably due to its volatile nature.
With the combined UV/H2O2–BAC treatment, the reduction in TN, NH+4–N and NO3 –N improved markedly. The removal of nitrogen species in the BAC system was possible through the nitrifica tion–denitrification process (Kalkan et al., 2011). Nitrification is carried out by two different groups of bacteria in the presence of sufficient DO (3–4 mg L 1), including ammonia oxidising bacteria (AOB) which oxidise ammonium to nitrite and nitrite oxidising bacteria (NOB) which further oxidise nitrite to nitrate. During the denitrification process, the nitrate is consumed by the denitrifiers, thereby converting nitrate to nitrous oxide (N2O) or nitrogen gas (N2) (Lee and Welander, 1996). In the present study, 90% NH+4–N removal was obtained with the combined treatment and BAC alone, however only 36% and 70% NO2 –N removal was achieved, respectively. Yapsakli et al. (2010) observed that most ammonium removal occurred as soon as the influent entered the BAC column, indicating that most of the nitrification took place in the uppermost part of the filter media. It is likely that this occurred in the present study. The results inferred that nitrifying bacteria can adapt to the extremely high salinity as there was high removal of ammonium by the combined system, which was in accordance with the fact that nitrifiers could survive at TDS up to 30 g L 1 by acclimation (Glass and Silverstein, 1999). The lower removal of nitrite might be due to inhibition of NOB by hydroxylamine produced during conversion of ammonia to nitrite by the AOB and/or due to the H2O2 present in the BAC influent. The changes in nitrite and nitrate after BAC treatment (Fig. 1) demonstrated that there was effective nitrification taking place. There was 62% and 65% of NO3–N removed by the combined treatment system and BAC alone, respectively, suggesting partial denitrification had taken place. The denitrifiers would be harboured deep inside the biofilm and towards the bottom of the BAC column, where the concentration of DO would be minimal. The COD:N ratio for complete denitrification varies between 7 and 10 (Carrera et al., 2004). In the present study, the ratio was 10, which was sufficient for denitrification. However, complete denitrification was not achieved despite sufficient carbon being available in the influent. This was most likely due to the higher DO level (6–7 mg L 1) of the ROC in the BAC system. Complete denitrification would normally take place when DO is
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Technical Note
Removing organic and nitrogen content from a highly saline municipal wastewater reverse osmosis concentrate by UV/H2O2–BAC treatment Shovana Pradhan, Linhua Fan ⇑, Felicity A. Roddick School of Civil, Environmental and Chemical Engineering, RMIT University, Melbourne, Victoria 3001, Australia
h i g h l i g h t s UV/H2O2–BAC treatment of a highly saline wastewater ROC was investigated. Approximately 60% DOC and TN was removed from the ROC by the process. Residual H2O2 in BAC feed enhanced DOC removal but limited TN removal.
a r t i c l e
i n f o
Article history: Received 14 March 2015 Received in revised form 7 May 2015 Accepted 12 May 2015 Available online 22 May 2015 Keywords: Reverse osmosis concentrate Biological activated carbon Advanced oxidation processes Organic matter removal Nitrogen removal Wastewater reclamation
a b s t r a c t Reverse osmosis (RO) concentrate (ROC) streams generated from RO-based municipal wastewater reclamation processes pose potential health and environmental risks on their disposal to confined water bodies such as bays. A UV/H2O2 advanced oxidation process followed by a biological activated carbon (BAC) treatment was evaluated at lab-scale for the removal of organic and nutrient content from a highly saline ROC (TDS 16 g L 1, EC 23.5 mS cm 1) for its safe disposal to the receiving environment. Over the 230-day operation of the UV/H2O2–BAC process, the colour and UV absorbance (254 nm) of the ROC were reduced to well below those of the influent to the reclamation process. The concentrations of DOC and total nitrogen (TN) were reduced by approximately 60% at an empty bed contact time (EBCT) of 60 min. The reduction in ammonia nitrogen by the BAC remained high under all conditions tested (>90%). Further investigation confirmed that the presence of residual peroxide in the UV/H2O2 treated ROC was beneficial for DOC removal, but markedly inhibited the activities of the nitrifying bacteria (i.e., nitrite oxidising bacteria) in the BAC system and hence compromised total nitrogen removal. This work demonstrated that the BAC treatment could be acclimated to the very high salinity environment, and could be used as a robust method for the removal of organic matter and nitrogen from the pre-oxidised ROC under optimised conditions. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Over the past decade, reverse osmosis (RO) technology has been used increasingly for full-scale municipal wastewater reclamation plants due to its ability to produce high quality effluent and increasing affordability (Shannon et al., 2008). However, management of the concentrate (also referred to as the brine) streams generated from the RO-based wastewater reclamation processes remains a major challenge for the water industry in applying the technology. The RO concentrate (ROC) contains almost all of the organic contaminants and nutrients from the secondary effluent at elevated levels (i.e., typically 4–6 times higher in concentration). The organic content of the ROC commonly includes ⇑ Corresponding author. E-mail address: [email protected] (L. Fan). http://dx.doi.org/10.1016/j.chemosphere.2015.05.028 0045-6535/Ó 2015 Elsevier Ltd. All rights reserved.
pharmaceuticals, personal care products, pesticides, endocrine disruptors, disinfection by products and other organic species; many of these are toxic and bio-accumulative. Moreover, introduction of excess nutrients to the receiving environment can cause adverse effects such as blooms of harmful algae. The disposal of the untreated ROC to a confined water body (e.g., bays) consequently poses significant risks to the health of humans and ecosystems. Due to the recalcitrant nature of the organic matter in the ROC, several oxidative treatment methods including advanced oxidation processes (AOPs) have been studied as a means of removing and/or enhancing the biodegradability of the organic content. In some recent studies, the homogeneous AOP utilising UVC/H2O2 has been demonstrated as an effective method for degrading the organic matter in the ROC and improving its biodegradability for subsequent biological treatment. One study showed the process (initial H2O2 concentration 11.5 mM, 120 min irradiation) could achieve
199
S. Pradhan et al. / Chemosphere 136 (2015) 198–203
complete decolourisation and 55% reduction in COD, 38% in DOC and 32% in dissolved organic nitrogen (DON) (Bagastyo et al., 2011). Liu et al. (2012) reported that more than 80% DOC removal could be achieved for a municipal wastewater ROC with UVC/H2O2 treatment (initial H2O2 concentration 4 mM, 30 min irradiation), followed by a biodegradable DOC (BDOC) determination as a surrogate biological treatment. The bench-scale study revealed that such treatment was effective for degrading the organic compounds over a wide range of salinity (EC 4.5–11.2 mS cm 1). Based on these studies, low-energy requirement and small-footprint biological processes should be considered as the subsequent treatments to improve the total organic removal and hence the cost-effectiveness of the oxidative treatment processes. Biological activated carbon (BAC) processes can provide simultaneous adsorption of non-biodegradable matter and oxidation of biodegradable matter in a single reactor with microbial activity in a granular activated carbon system (Walker and Weatherley, 1999). The use of BAC to treat pre-oxidised ROC has been reported in several studies. Lee et al. (2009) found that the coupling of a BAC column with ozone pre-oxidation enhanced the organic removal efficiency by 3 times that of BAC alone at an empty bed contact time (EBCT) of 60 min, with 70% TOC removed from the ROC (TDS 1.2 g L 1, EC 2 mS cm 1). In a more recent study, Lu et al. (2013) observed that the integrated treatment of UV/H2O2 followed by BAC removed 59% DOC, 64% colour and 78% UV absorbance at 254 nm (UVA254) from a ROC (TDS 10 g L 1, EC 13.5 mS cm 1). However, the removal efficiencies for nutrients such as total nitrogen (TN) and total phosphate (TP) were fairly low, with only 24% of TN and 17% of TP removed. In an earlier study, Ng et al. (2008) found that BAC treatment following capacitive deionisation could achieve a higher TN removal (91%) from a ROC with relatively low salinity (EC 2 mS cm 1). The low TN removal observed by Lu et al. (2013) may be related to the high salinity of the ROC, as it is well recognised that high salinity of wastewater can greatly affect biological activity, resulting in decreased nitrification and denitrification, thereby affecting overall TN removal (Kargi and Dinçer, 1997). The biological removal of nitrogen at elevated salt concentration can be challenging due to the sensitivity of nitrogen-removing microorganisms to high salt conditions, along with other environmental conditions including temperature, DO concentration, pH, ammonia concentration, heavy metals and C/N ratio (Okabe et al., 1996). However, it is possible for the microorganisms to be acclimated to high salinity environments. The ROC salinity level may be classified as low (TDS < 5 g L 1), medium (5–10 g L 1), high (10–15 g L 1) and extremely high (>15 g L 1). To date, most of the published work has been on ROC with low or medium salinity. Investigation of the removal of organic content and nutrients (particularly nitrogen) from ROC with extremely high level of salinity using UV/H2O2 followed by BAC has not been reported. The aim of this study was to investigate the effectiveness and robustness of a UV/H2O2–BAC process for removing organic matter and nitrogen from a municipal wastewater ROC with extremely high salinity over an extended period of operation (230 days). The impact of the residual H2O2 in the BAC influent on the bio-treatment performance was also studied. The treatment effectiveness was characterised in terms of reductions in DOC, UVA254, colour, COD, total nitrogen (TN), ammonium nitrogen (NH4–N), nitrate nitrogen (NO3–N) and nitrite nitrogen (NO2–N). 2. Materials and methods 2.1. Source of ROC The ROC samples were collected from a local wastewater reclamation facility in which the biologically treated secondary effluent was treated by a UF (0.04 lm)–RO system to remove salts and
other contaminants to produce recycled wastewater. The water recovery of the RO system was approximately 75%. The general characteristics of the ROC samples collected during January to October, 2013 are given in Table 1. The secondary effluent had a high TDS level due to the infiltration of salty groundwater into the sewer system, leading to the extremely high salinity of the ROC. The collected samples were stored at 4 °C and brought to room temperature before use. 2.2. UV/H2O2 treatment of ROC Irradiation was conducted in batch mode using an annular reactor with a centrally mounted lamp. The UVC lamp (254 nm) had an energy input of 39 W (Australian Ultra Violet Services, G36T15NU), and the average fluence was determined as 8.9 mJ s 1 cm 2. During irradiation, samples were mixed and aerated by humidified air. The average irradiation area was 464 cm2 with a path length of 1.94 cm and other UV reactor conditions are reported elsewhere (Liu et al., 2012). The ROC sample (900 mL) was dosed with 3 mM H2O2 and then exposed to UV irradiation for 30 min (UV dose 16 J cm 2) as suggested by Liu et al. (2012) as the optimum conditions for treating the ROC. The treated ROC was then fed to BAC columns under the defined experimental conditions. The residual H2O2 in the UV/H2O2 treated ROC was measured using a photometric method at 551 nm (Bader et al., 1988). The enzyme catalase derived from bovine liver (Sigma, 3691 units per mg solids) was used to quench the residual H2O2 for examining the effect of residual peroxide on the efficiency of BAC treatment. For every 25 mL of the water sample, 10 lL (4000 U mL 1) of catalase was added to remove the residual H2O2. Catalase was also used to decompose residual H2O2 to remove its interference in water quality analyses. The samples were shaken at 100 rpm at room temperature until the concentration of H2O2 was less than 0.5 mg L 1. 2.3. BAC start-up and operation Granular activated carbon (GAC) columns with an inner diameter of 1.5 cm and an effective packing height of 12 cm were set up. Coal-based activated carbon (Activated Carbon Technologies Pty Ltd, Australia) with effective size of 1.2–1.4 mm, density 0.20– 0.30 g cm 3 and surface area >1200 m2 g 1 was used. The carbon was sieved to remove the very fine particles, washed repeatedly with deionised water until the fine particles were removed, and then dried in an oven at 110 °C for 2 days. The carbon media (200 g) was then inoculated by mixing with activated sludge (1 L, MLSS 4000 mg L 1) obtained from the wastewater treatment plant which supplied the secondary effluent to the reclamation facility for 4 days. Nutrients including N, P and C sources were added to the system to promote the growth of microorganisms during the inoculation period (glucose 0.78 g L 1, ammonium chloride
Table 1 Characteristics of the ROC samples (January to October, 2013). Parameter
Value
pH DO (mg L 1) DOC (mg L 1) UVA254 (cm 1) Colour (mg Pt–Co L 1) COD (mg L 1) TN (mg L 1) TP (mg L 1) TDS (g L 1) Chloride (mg L 1) Conductivity (mS cm 1)
7.7 ± 0.4 10.2 ± 1.3 36.0 ± 4.0 0.62 ± 0.02 148 ± 10.0 120 ± 19 21.4 ± 4.5 28.5 ± 1.1 16.6 ± 0.8 7,700 ± 1,800 23.5 ± 1.3
200
S. Pradhan et al. / Chemosphere 136 (2015) 198–203
0.11 g L 1, potassium dihydrogen phosphate 0.033 g L 1) (Lu et al., 2013). The media was then transferred into glass columns which were then fed with UV/H2O2-treated ROC in downflow mode. The TDS of UV/H2O2 treated ROC was increased gradually from 4 to 16 g L 1 over four weeks; DOC, UVA254 and colour of the BAC effluent were continuously monitored until steady values were achieved. The BAC columns were then used for the study on the ROC with original salinity. All BAC tests were run at room temperature (22–26 °C). Two identical BAC columns were operated with EBCT of 60 min on raw and UV/H2O2 treated ROC, respectively, over a period of 230 days. The flow rate of the effluent from the BAC columns was monitored daily and regulated using the outlet valve to maintain the fixed exposure time. Backwashing was carried out every 2 weeks to control the biomass growth and prevent media clogging. The effluent samples were collected periodically (every 3–5 days) and analysed for DOC, COD, UVA254, colour, TN, NH+4–N, NO2 –N, NO3 –N and TP. 2.4. Analytical methods DOC was measured using a Shimadzu TOC-L analyser. Prior to DOC analysis, the samples were diluted (1 in 5) with MilliQ water in order to reduce the possible impact of high TDS of the ROC on the measurement. A spectrophotometer (DR5000, Hach, USA) was used to determine the COD, colour, TN and TP of the water samples according to the Hach Methods Manual. For COD measurement, 0.5 g of mercury sulphate (Sigma) was added to overcome the interference of high chloride concentration as suggested in the standard method (APHA, 2005). UV absorbance was measured at 254 nm (UVA254) using a Shimadzu UV–vis spectrophotometer (UV2-2700). TDS was determined in accordance with Standard Method 2540C (APHA, 2005). The DO and conductivity was measured with a Hach DO meter (LDO 101) and Hach conductivity meter (CDC 401), respectively. The pH was measured with a Hach Sension 156 pH meter. NH+4–N was measured colorimetrically at 660 nm via indophenol formation with sodium salicylate as described by Verdouw et al. (1978), NO3 –N was measured colorimetrically at 420 nm with the sodium salicylate method described by Scheiner (1974) and NO2 –N was measured using Standard Method 4500B (APHA, 2005). All analyses were duplicated and results recorded in terms of mean values and errors/standard deviations. 3. Results and discussion 3.1. Removal efficiencies for organic matter and nutrients by the UV/ H2O2–BAC treatment The UV/H2O2–BAC treatment system was operated for 230 days (February to October, 2013) with EBCT of 60 min on multiple
batches of ROC sample to evaluate its efficiency for the removal of organics and nutrients (Table 2). As a reference, the raw ROC was treated by BAC alone with an identical column under the same operating conditions. After UV/H2O2 pre-treatment of the ROC, there were marked reductions in UVA254 and colour (60% and 86%, respectively), whereas only 15% reduction was achieved for DOC and COD, respectively. The UVA254 reflects the presence of conjugated bonds and aromatic content of organic matter (Chen et al., 2011). The result was consistent with previous studies which showed great cleavage but limited mineralisation of the organic matter in the ROC using such treatment (Umar et al., 2013). For the BAC alone treatment, the removal efficiency for DOC and COD was considerably higher (38% and 32%, respectively) than for the UV/H2O2 treatment alone, although the reductions in UVA254 and colour (64% and 80%, respectively) were comparable for the two treatments. The reduction in DOC by the BAC alone was associated with biodegradation and adsorption. Although the organic molecules in the ROC are generally regarded as low in biodegradability as they were primarily derived from the biologically treated secondary effluent, it was possible that some of the molecules were removed by the microbes with sufficient contact time (Yapsakli and Çeçen, 2010). Using the combined treatment, the removal efficiency for the organic matter was increased significantly, with 57% DOC, 46% COD, 81% UVA254 and 95% colour removed from the ROC. The great improvement in organic matter removal was mainly attributed to the partial degradation of the organic molecules by the oxidative treatment, leading to the production of simpler molecules which could be readily removed by the microbes embedded in the BAC column (Lu et al., 2013). A comparison of the organic matter removal efficiency obtained in the present study and the work done by Lu et al. (2013) using the same treatment train showed that the treatment performance was fairly comparable, although the ROC salinity was markedly different. This might indicate that the change in the TDS of ROC in the range of 10–16 g L 1 did not have any significant impact on the organic matter removal. It is worth noting that the UV/H2O2–BAC treated ROC had similar DOC and COD levels to the secondary effluent used as the influent of the RO-based reclamation process (i.e., 15 cf. 14.5 mg DOC L 1 and 65 cf. 64 mg COD L 1), but significantly lower UVA254 and colour (i.e., 0.12 cf. 0.19 cm 1 and 7 cf. 120 mg Pt–Co L 1). In terms of nutrient removal, minimal reduction was obtained by the UV/H2O2 treatment. The small reduction in TN during the oxidation process was likely due to the oxidation of some nitrogen species to nitrogen gas (Dwyer et al., 2008). With BAC treatment alone, the removal efficiency was fairly high for TN (71%) but still low for TP (8%). TN reduction for the combined treatment system was lower (60%) compared with the BAC only treatment, although TP reduction was improved slightly. This study showed a higher TN removal (60%) compared with the study by Lu et al. (2013), in which the reduction in TN was only 24% for the combined
Table 2 Overall water quality characteristics of the ROC after the various treatments. Treatment Parameter
Raw ROC
UV/H2O2
BAC
UV/H2O2 + BAC
pH DO (mg L 1) DOC (mg L 1) COD (mg L 1) UVA254 (cm 1) Colour (mg Pt–Co L TN (mg L 1) TP (mg L 1)
7.7 ± 0.4 10.2 ± 1.3 35.0 ± 5.0 120.0 ± 9.0 0.623 ± 0.025 138.0 ± 10.0 21.0 ± 3.0 29.0 ± 2.0
8.0 ± 0.2 12.1 ± 0.9 30.0 ± 4.0 (15%) 102.0 ± 8.0 (15%) 0.246 ± 0.033 (60%) 20.0 ± 8.0 (86%) 18.0 ± 3.0 (13%) 28.0 ± 2.0 (2%)
8.4 ± 0.2 7.0 ± 0.9 21.0 ± 4.0 (38%) 82.0 ± 11.0 (32%) 0.23 ± 0.06 (64%) 28.0 ± 8.0 (80%) 6 ± 1.0 (71%) 26.0 ± 1.0 (8%)
8.4 ± 0.2 7.2 ± 1.2 15.0 ± 2.0 (57%) 65.0 ± 7.0 (46%) 0.12 ± 0.03 (81%) 7.0 ± 5.0 (95%) 9.0 ± 2.0 (60%) 24.0 ± 1.0 (15%)
1
)
Note: Results were based on the 57 samples collected during February to October, 2013; values are presented as mean ± 1 standard deviation, average removal efficiency (%) is shown in the bracket.
201
S. Pradhan et al. / Chemosphere 136 (2015) 198–203
treatment system with the same EBCT for the BAC treatment. The higher removal of TN in the present study may be related to the use of a lower H2O2 concentration (3 mM cf. 4 mM), and thus less residual H2O2 after the oxidative process. The impact of residual H2O2 on the removal of organic content and nitrogen was therefore investigated further in this work and reported in Section 3.3. Nevertheless, the TP removal was comparable in both studies (15% in this study) for the combined treatment system. It is known that phosphorus removal is greatly inhibited by increasing salinity (>5 g L 1) (Uygur and Kargı, 2004). As the salinity of ROC used in both studies was very high (TDS 16.6 g L 1 and 10 g L 1), lower removal of TP would then be anticipated. Since chemical coagulation is generally an effective means for TP removal, it could be employed prior to the oxidative treatment for achieving the required removal efficiency for TP, and additional reduction in organic content. The DO of the effluent was 6 mg L 1 for the combined and the BAC alone treatment, which was lower than for the influent (10 mg L 1), suggesting that the activated carbon adsorbed DO and the adsorbed DO was utilised by the microorganisms to biodegrade the contaminants (Jin et al., 2013). During the 230-day operation of the combined treatment, six batches of ROC sample were used for the study. The TDS of the ROC samples varied from 15.6 to 18.2 g L 1 and the DOC varied from 32 to 38 mg L 1. It was observed that the average DOC removal efficiency of the combined treatment over the ROC sample batches varied only slightly (54–59%), whereas the TN removal efficiency varied a little more (55–65%). This indicates that the UV/H2O2–BAC process was a robust treatment system for the ROC for organic matter and nitrogen removal.
3.2. Characterisation of nitrogen removal To obtain a better understanding of the nitrogen removal, the treatments were characterised in terms of the removal efficiency of TN, NH+4–N, NO2 –N and NO3 –N over a 50 days period (August to October, 2013) (Fig. 1). The UV/H2O2 treatment alone was not very effective for removing total nitrogen, with only 15% of TN removed. NO2 –N was removed almost completely, whereas NO3 –N concentration increased by 1–2 mg L 1 after the oxidation treatment. The increase in NO3 –N concentration was attributed to the conversion of NO2 –N to NO3 –N in the presence of oxidant as suggested by Munter (2001). Since NH+4–N has no unsaturated double bonds, it would be difficult for the_OH radical to break down the N–H bonds through electrophilic attack (Haag et al., 1984), hence the slight reduction in the NH+4–N during the treatment was probably due to its volatile nature.
With the combined UV/H2O2–BAC treatment, the reduction in TN, NH+4–N and NO3 –N improved markedly. The removal of nitrogen species in the BAC system was possible through the nitrifica tion–denitrification process (Kalkan et al., 2011). Nitrification is carried out by two different groups of bacteria in the presence of sufficient DO (3–4 mg L 1), including ammonia oxidising bacteria (AOB) which oxidise ammonium to nitrite and nitrite oxidising bacteria (NOB) which further oxidise nitrite to nitrate. During the denitrification process, the nitrate is consumed by the denitrifiers, thereby converting nitrate to nitrous oxide (N2O) or nitrogen gas (N2) (Lee and Welander, 1996). In the present study, 90% NH+4–N removal was obtained with the combined treatment and BAC alone, however only 36% and 70% NO2 –N removal was achieved, respectively. Yapsakli et al. (2010) observed that most ammonium removal occurred as soon as the influent entered the BAC column, indicating that most of the nitrification took place in the uppermost part of the filter media. It is likely that this occurred in the present study. The results inferred that nitrifying bacteria can adapt to the extremely high salinity as there was high removal of ammonium by the combined system, which was in accordance with the fact that nitrifiers could survive at TDS up to 30 g L 1 by acclimation (Glass and Silverstein, 1999). The lower removal of nitrite might be due to inhibition of NOB by hydroxylamine produced during conversion of ammonia to nitrite by the AOB and/or due to the H2O2 present in the BAC influent. The changes in nitrite and nitrate after BAC treatment (Fig. 1) demonstrated that there was effective nitrification taking place. There was 62% and 65% of NO3–N removed by the combined treatment system and BAC alone, respectively, suggesting partial denitrification had taken place. The denitrifiers would be harboured deep inside the biofilm and towards the bottom of the BAC column, where the concentration of DO would be minimal. The COD:N ratio for complete denitrification varies between 7 and 10 (Carrera et al., 2004). In the present study, the ratio was 10, which was sufficient for denitrification. However, complete denitrification was not achieved despite sufficient carbon being available in the influent. This was most likely due to the higher DO level (6–7 mg L 1) of the ROC in the BAC system. Complete denitrification would normally take place when DO is
