Analysis of the impacts of nitrogen, phosphorus and potassium supplementation on biofilter performance for organic carbon removal was studied on laboratory-scale biofilter columns. Three dual media biofilter columns were fed with synthetic raw water C:N:P ratios of 546:24:1, 100:10:1, and 25:5:1 (w/w) to simulate nutrient limited and two nutrient supplemented conditions, respectively. Research found that air-scour versus water only backwash improved the nutrient limited dissolved organic carbon (DOC) removal by 8%. In addition, nutrient supplementation and backwash alteration improved DOC removals by 19% for the 25:5:1 column and 14% for the 100:10:1 column. Potassium supplementation with the 25:5:1 C:N:P ratio column had no discernible effect on DOC removal. No correlation with phospholipid (7–474 nmol P/g media) and adenosine triphosphate (ATP) (0.6 × 105–32.74 × 105 pg ATP/g media) values with DOC removal were found. Nutrient availability was found to influence DOC removal, demonstrating its importance when utilizing biofiltration for treatment of source waters.

INTRODUCTION

Natural organic matter (NOM) is a ubiquitous component of surface waters. If left untreated through the water treatment process, it has the potential to cause biological regrowth in water distribution systems. It can also lead to a higher chlorine demand and the formation of disinfection by-products. Biofiltration is a sustainable technology that has the ability to remove NOM through biodegradation mechanisms of bacteria attached to filter media. For processes employing disinfection downstream of filtration, optimizing the filtration process could prove beneficial in terms of reducing the chemicals required and could decrease the formation of disinfection by-products (Badawy et al. 2012; Azzeh et al. 2015).

Currently, a wide range of physical parameters such as backwashing strategy, hydraulic loading rate and empty bed contact time (EBCT) are employed when designing filters for drinking water treatment. Further, many utilities choose to use chlorinated or chloraminated backwash water in an attempt to increase filter run times and thereby maximize water production. These backwash procedures, however, can be detrimental to biological activity within a potential biofilter. While physical parameters such as EBCT play an important role in the functioning of a filtration process, current filtration design practices do not account for optimizing biological growth on the media. However, biological degradation of dissolved organic contaminants is an added potential advantage of biofiltration (Feng et al. 2013; Liao et al. 2013). Furthermore, past research has shown that biodegradation has led to prolonged bed life of granular activated carbon (GAC) biofilters (Scharf et al. 2010; Yapsakli & Çeçen 2010).

Studies found that the biodegradation of organics and performance of the biofilter depend on the type of microorganisms growing on the biofilter media (McDowall et al. 2009; Zhang et al. 2010) and the amount of essential nutrients in the biofilter system (Boon et al. 2011; Lauderdale et al. 2012). Therefore, the microbial activity is often believed to be correlated to their access to nutrients and deficiency of any individual nutrient that limits the growth of microbes. A C:N:P ratio of less than 100:10:1 on a wt basis has been shown to be detrimental to obtaining an optimized biofiltration process (LeChevallier et al. 1991). Coagulants such as alum (used commonly in water treatment) have proven effective in precipitating phosphorus (Tchobanoglous et al. 2013). It is believed that as a result of the effectiveness of the coagulation process and/or source water limitations, phosphorus concentrations to downstream filter influents are usually <0.01 mg/L (Nyfenneger et al. 2013) at typical water treatment facilities. It has been demonstrated that phosphorus concentrations below 5 μg/L impede the typical transport mechanism of phosphorus through membranes within bacterial cells (Rosenberg 1987). Therefore, in order to promote cell development and biofilm growth, it might be prudent to have phosphorus concentrations greater than 5 μg/L in biofilter influents. In a previous study conducted by Sang et al. (2003), biofilters dosed with 25 μg -P/L were compared with columns not receiving phosphorus. It was found that the total organic carbon (TOC) removals increased from 17.1% to 21.3% for the phosphorus limited column to 22%–26.6% for the phosphorus supplemented column.

Further, given the complex nature of bacterial biomass attached to filter media, it is indeed possible that several other factors play a key role in organics removal. In addition to nitrogen and phosphorus, nutrients such as potassium may also play a role in biofilter development. For instance, in wastewater systems, Brdjanovic et al. (1996) investigated the effects of potassium on biological phosphate removal in a sequenced batch reactor operating in anaerobic-aerobic-settling mode. Potassium limited conditions (7.15 × 10−4 mol K/mol P in the influent) showed no biological phosphorus uptake when compared to a ratio of 1 mol K/mol P in the influent, where complete biological phosphate removal was observed. Although there exist fundamental differences between wastewater treatment and biofilters for drinking water treatment, it is possible that potassium addition would result in an increased uptake of phosphorus thereby positively influencing organic carbon removal. Addition of a mixture of different nutrients besides potassium was deemed beyond the scope of this research as the goal was to minimize variability and study the effects of one particular nutrient (in this case, potassium) on system performance.

The research objectives of this work are to investigate the impact of nitrogen and phosphorus enhancement on biofiltration performance from an organics removal perspective. Analysis of the impacts of potassium is also discussed. Finally, a comparison between water only backwash and backwash with water and air scour is discussed.

MATERIALS AND METHODS

Biofilter setup

Research was conducted with three identical laboratory-scale biofilter columns comprised of 520 mm of GAC (effective size, ES = 0.7 mm) overlaying 180 mm of sand (ES = 0.5 mm) followed by a synthetic underdrain material. The columns were 1,250 mm tall with an internal diameter of 50 mm. The columns had an EBCT of 14 minutes with a hydraulic loading rate of 3 m/h (100 mL/min).

Experimental methodology

The three dual media laboratory-scale biofiltration columns were operated in parallel over the duration of the experiments using a simulated surface water (SRW) to mimic the Ottawa River (Ontario, Canada) water. During the filter conditioning phase, nutrient limited conditions (C:N:P = 546:24:1 w/w) were provided to the three columns to simulate the nutrient limited conditions in the Ottawa River water. In Phase I of the study, supplementary nitrogen and phosphorus were added to two of the three biofilters to achieve C:N:P ratios of 25:5:1 and 100:10:1 (w/w), respectively, while one biofilter was kept at the nutrient limited condition (control unit) over the duration of the study. In Phase II, potassium was supplemented to the column with a 25:5:1 C:N:P at a ratio of 1 mol K/mol P as per Brdjanovic et al. (1996). Table 1 provides details of the various phases of this research.

Table 1

Experimental phase description

Phase Comments C:N:P ratio (w/w) Column abbreviationa Duration (days) 
N/A Filter conditioning
Water only backwash 
546:24:1 (All columns) 1-NL
2-NL
3-NL 
104 
Collapse pulsing backwash procedure employed 25:5:1 1-NS A 117 
100:10:1 2-NS B 
546:24:1 3-NL 
II Collapse pulsing backwash procedure employed 25:5:1 1-NS A + K 94 
100:10:1 2-NS B 
546:24:1 3-NL 
Phase Comments C:N:P ratio (w/w) Column abbreviationa Duration (days) 
N/A Filter conditioning
Water only backwash 
546:24:1 (All columns) 1-NL
2-NL
3-NL 
104 
Collapse pulsing backwash procedure employed 25:5:1 1-NS A 117 
100:10:1 2-NS B 
546:24:1 3-NL 
II Collapse pulsing backwash procedure employed 25:5:1 1-NS A + K 94 
100:10:1 2-NS B 
546:24:1 3-NL 

a1, 2, 3, column numbers; NS A, nutrient supplement condition A; NS B, nutrient supplement condition B; NL, nutrient limited condition; K, potassium.

All filters were backwashed once a week. During filter conditioning, a water only backwash procedure with a backwash velocity of 30 m/hr was employed to achieve 30% GAC bed expansion. In Phase I and II, an air scour assisted, collapse pulsing procedure was introduced. The procedure consisted of 6 minutes combined air and water wash (backwash velocity = 12 m/hr and air velocity = 72 m/hr) followed by a high rate water wash (backwash velocity = 30 m/hr with 30% GAC bed expansion) for 2 minutes and finally a low rate water wash (backwash velocity = 10 m/hr) for 2 minutes.

Synthetic raw water

Stock solutions were prepared weekly with the requisite chemicals in 1 L volumetric flasks, which were thoroughly rinsed with phosphate free soap and deionized water. The stock solutions were mixed with dechlorinated tap water in two separate 50 L tanks for carbon and nutrients, respectively. The SRW was fed to the biofilter columns using an in-line static mixer. The carbon dosing solution, prepared with glyoxal (Fisher Scientific, Canada, Ontario), acetic acid (Fisher Scientific, Canada, Ontario) and formic acid (Fisher Scientific, Canada, Ontario). The nutrient dosing solution contained nitrogen, phosphorus and micronutrients, prepared using sodium nitrate (Sigma-Aldrich, Canada, Ontario) and in some cases potassium dihydrogen phosphate (Fisher Scientific, Canada, Ontario) was also supplemented. For nitrogen, phosphorus supplementation, sodium nitrate and potassium dihydrogen phosphate were used. This provided C:N:P ratios of 25:5:1 (Tchobanoglous et al. 2013) and 100:10:1 (Anderson et al. 2008) on a w/w basis. Potassium was supplemented in the form of potassium sulphate (Sigma-Aldrich, Canada, Ontario).

CONSTITUENT ANALYSIS

Water constituents including temperature, pH, total TOC, dissolved organic carbon (DOC), potassium, and total phosphorus were analyzed. TOC and DOC were measured in accordance with Standard Methods (APHA 2012), method 5310C- persulfate ultraviolet oxidation. DOC samples were first filtered using a 0.45 μm filter and then measured according to the same method as TOC. TOC analysis was performed using a SHIMADZU TOC-VCPH/CPN analyzer (SHIMADZU, Canada). Phosphorus and potassium were analyzed according to HACH PhosVer 3 method (#8190) and tetraphenylborate method (#8049), respectively, using HACH DR2800 spectrophotometer. Temperature and pH measurements were acquired using an ORION pH meter with attached pH probe (Thermo Scientific, Canada, Ontario).

Ultraviolet absorbance (UVA) was measured in accordance with Standard Methods 5910B at 254 nm (APHA 2012) using UV visible spectrophotometer (Spectronic Unicam). Specific ultraviolet absorbance (SUVA) was calculated as the UV absorbance values at 254 nm normalized by the DOC concentration. All samples were collected two to three times weekly in duplicate.

Microbial analysis

Phospholipid analysis was carried out in accordance with the procedure described by Wang et al. (1995) to measure the biomass attached to the surface of the GAC media. One gram of GAC media was drawn from two sampling ports located at 5 cm and 18 cm from the top of the filter media, respectively. Phospholipids were extracted from the cells attached to the media and measured calorimetrically at the wavelength of 610 nm. These measurements represent biomass from the top and the middle of each column. Furthermore, to quantify the amount of microbiological activity in the biofilters, adenosine triphosphate (ATP) analysis was carried out using the deposit surface analysis test kit (LuminUltra Technologies Ltd, NB). The ATP samples were introduced to solution containing enzyme Luciferase to produce light and the light was detected in a luminometer (PhotonMaster) as relative light units and converted to ATP concentrations according to manufacturer protocols.

RESULTS AND DISCUSSION

Phase I – Impact of nutrient supplementation

Biofilter columns were initially run under nutrient limited conditions for a period of 104 days. DOC and UVA were monitored for 104 days of operation when steady state DOC removals were achieved. Respective average DOC removals for the columns were 13.5 ± 9.3% for 1-NL, 13.7 ± 9.3% for 2-NL and 15.2 ± 8.7% for 3-NL. A paired t-test analyses of the DOC removals in three biofilter columns showed no statistical difference (p > 0.05); indicating the columns were performing similarly.

Once the columns had reached steady state under the nutrient limited conditions, phase I was initiated. At this phase, supplementary nitrogen and phosphorus were added to two of the three biofilter columns to achieve a C:N:P (w/w) ratio of 25:5:1 and 100:10:1, respectively. Figure 1 shows the percentage of DOC removal in the three biofilter columns operating at the different nutrient conditions over phase I. Although the %DOC removals have large fluctuations, there is an overall increasing trend in removal indicating that both changes had a positive impact on the biofilter operations. After day 171, the system exhibited more stable behaviour.
Figure 1

Effect of temperature on %DOC removals over Phase 1.

Figure 1

Effect of temperature on %DOC removals over Phase 1.

Nutrient supplementation was found to positively impact DOC removals as compared to the conditioning phase. On average, DOC removal for 1-NS A was found to be 10% higher than the nutrient limited control column (32% removal vs 23% removal). This increase was less apparent in 2-NS B, where a 4% increase was observed compared to the control (28% removal versus 23% removal). The average standard deviation of each sample over the course of phase I at each sampling point was found to be 8.6%.

Temperature has been found to be an important parameter for organic carbon removal in drinking water biofilters when comparing 20 °C to 5 °C conditions (Moll et al. 1999; Emelko et al. 2006). Prior to day 171, water temperature entering the columns remained fairly constant in the range of 18–20 °C. However, beyond this point, temperature decreased gradually from 20 °C to 13 °C due to the seasonal temperature changes in the tap water used for influent SRW. An analysis of the DOC results by time and possibly temperature is shown in Table 2. It was concluded that the modest drop in temperature from 20 to 13 °C did not impact the DOC removal. It is more likely that the change in backwash condition at the start of Phase1 resulted in the system taking longer to reach steady state than expected, although this time period is in line with research by others (Liu et al. 2001).

Table 2

Effect of temperature on %DOC removal (avg ± std dev)

Column Overall %DOC removal (n = 34) %DOC removal (before day 171, n = 16)
Temp 20 °C 
%DOC removal (after day 171, n = 18)
Temp decrease to 13 °C 
1-NS A 33.5 ± 9.7 35.3 ± 11.9 31.9 ± 7.3 
2-NS B 28.6 ± 10.5 31.8 ± 9.8 25.9 ± 10.5 
3-NL 23.5 ± 9.4 21.4 ± 12 25.3 ± 6.3 
Column Overall %DOC removal (n = 34) %DOC removal (before day 171, n = 16)
Temp 20 °C 
%DOC removal (after day 171, n = 18)
Temp decrease to 13 °C 
1-NS A 33.5 ± 9.7 35.3 ± 11.9 31.9 ± 7.3 
2-NS B 28.6 ± 10.5 31.8 ± 9.8 25.9 ± 10.5 
3-NL 23.5 ± 9.4 21.4 ± 12 25.3 ± 6.3 

While the overall removal for 1-NS A and 3-NL remained the most consistent over the Phase I study, it is noted that the overall DOC removal for 2-NS B reduced by approximately 6% over this period. In addition, the column with the highest nutrient ratio (1-NS A) has the highest overall DOC removal at 33.5% compared to 2-NS B and 3 NL, 28.6% and 23.5%, respectively. A paired t-test confirms the difference in values between 1-NS A and 3-NL (p < 0.05), while the results for the other two were not significant.

As compared to DOC removals, larger fluctuations were observed in the system when it came to analyzing SUVA removals. Although a reduction in DOC and UVA values were observed for all biofilter columns, analysis of the SUVA values showed a different trend. It was found that in phase I, effluent SUVA values were consistently higher than influent SUVA values for all columns. On average, effluent SUVA values were 29%, 25% and 10% higher than the influent for 1-NS A, 2-NS B and 3-NL, respectively, which is in agreement with a previous study by Basu & Huck (2004).

Phase II – Impact of potassium addition on biofilter performance

Potassium has been found to be an important parameter in the biological uptake of phosphorus and hence carbon uptake in wastewater treatment systems (Brdjanovic et al. 1996). However, there is no information on the effects of potassium addition on the performance of drinking water biofilters. The aim of this phase of the study was to investigate the impacts of potassium supplementation on carbon uptake in a drinking water treatment. The other conditions for all columns remained the same from Phase I. During Phase II, the water temperature was approximately 11 °C throughout the course of this experiment.

Between Phase I and II, the columns were shut down for two weeks and then conditioned for three weeks before regular sampling was initiated. No significant differences (p > 0.05) were found in the overall DOC removals between the two nutrient supplemented columns 1-NS A + K and 2-NS B, suggesting that the potassium supplementation did not improve the biofilter performance with respect to DOC removal (see Figure 2). However, several spikes in DOC removals were observed, with removals as high as 75% being achieved, which could not be explained and require further investigation. However, Figure 2 shows statistically significant differences between the nutrient supplemented columns and the nutrient limited column i.e. 1-NS A + K vs 3-NL (p < 0.05); and 2-NS B vs 3-NL (p < 0.05).
Figure 2

Average %DOC removal for Phase II.

Figure 2

Average %DOC removal for Phase II.

From Figure 3 it is easy to observe that the nutrient supplementation increased the DOC removal by approximately 10–12% compared to the nutrient limited column. However, other studies observed higher improvements in DOC removal after addition of supplementary nutrients. For instance, Lauderdale et al. (2012) achieved approximately 75% greater removal from nutrient enhanced biofilters as opposed to their nutrient limited control. Basu & Huck (2004) have reported TOC removals as high as 50% for columns receiving a C:N:P of 15:5:1 (w/w) on a weight basis. It is possible that more nutrients need to be present in the synthetic raw water to promote a higher level of DOC removal; or that variations in backwash procedure may have influenced results, as was seen here with water only versus water + air scour backwash. Futhermore, we do not observe any particular benefit in the addition of potassium to the NSA column when compared across all the conditions.
Figure 3

Average %DOC removal for all columns over conditioning phase, Phase I and Phase II (in the conditioning phase all columns operated under nutrient limited conditions).

Figure 3

Average %DOC removal for all columns over conditioning phase, Phase I and Phase II (in the conditioning phase all columns operated under nutrient limited conditions).

Impact of backwashing procedures on biofilters performances

In Phase I, the backwash procedure was altered to include air scour which resulted in a measurable improvement in DOC removal in the nutrient limited column (3-NL) from 15.2% to 23.5% (statistically significant with p < 0.05). The effects of the new backwashing regime on 3-NL are outlined in Table 3.

Table 3

%DOC removals (avg ± std dev) over the conditioning phase, Phase I and Phase II

Nutrient condition %DOC removal (conditioning phase) %DOC removal (Phase I) %DOC removal (Phase II) 
1-NS A 13.5 ± 9.3 33.5 ± 9.7 33.5 ± 14.7 
2-NS B 13.7 ± 9.3 28.6 ± 10.5 32.9 ± 14.6 
3-NL 15.2 ± 8.7 23.5 ± 9.4 21.3 ± 12.4 
Nutrient condition %DOC removal (conditioning phase) %DOC removal (Phase I) %DOC removal (Phase II) 
1-NS A 13.5 ± 9.3 33.5 ± 9.7 33.5 ± 14.7 
2-NS B 13.7 ± 9.3 28.6 ± 10.5 32.9 ± 14.6 
3-NL 15.2 ± 8.7 23.5 ± 9.4 21.3 ± 12.4 

This improvement in DOC removals associated with backwashing highlights the importance of implementing optimized backwash procedures for biofiltration to ensure maximum DOC removal. This is in contrast to results reported at similar temperature ranges by Emelko et al. (2006); however, it should be noted that the Emelko et al. study was conducted under non-nutrient limited conditions. This may indicate a greater sensitivity for DOC removal under conditions of stress i.e. nutrient limited conditions. Furthermore, there was no statistical differences between DOC removals in the nutrient limited column (3-NL) between Phase I and II, as was expected when conditions were maintained (p > 0.05). Further, the data seem to highlight the importance of appropriate backwash strategies when dealing with a nutrient limited water source.

Taking into account that the approximate 8% increase in DOC removal for the control column is attributed to the air-scour assisted backwashing, it can be hypothesized that the further increase in DOC removal for 1-NS A and 2-NS B can be attributed to nutrient supplementation. Furthermore, no significant differences were found between 2-NS B biofilter (C:N:P = 100:10:1) compared with 1-NS A biofilter (C:N:P = 25:5:1); implying that a large range of C:N:P ratios or water quality conditions are acceptable for moderate DOC removal. However, more studies should be carried over a wider range of raw water conditions, including micronutrient addition, in order to determine the reproducibility of these results, in particular given the relatively large standard deviation associated with the average removal values.

Moreover, it is believed that nutrient uptake in the active portion of the biofilm is limited to a certain amount that would be required by the bacterial community to perform various metabolic functions. Nutrient availability beyond a certain ratio would thus cease to be beneficial. Indeed, previous research by Romani et al. (2004) has shown a reduction in extracellular enzyme activity (β-glucosidase and phosphatase) within a biofilm as a result of increasing the phosphorus concentration and altering the N:P ratio from 4.34 to 1.46 on a w/w basis under carbon limited conditions.

Analysing the SUVA data showed consistently higher SUVA values in the effluent samples for all observed columns in Phase II. This would be consistent with what was observed in Phase I. Thus, it is concluded that SUVA removal is not indicative of filter performance when carbon sources do not contain aromatic compounds. While this may not have implications for larger treatment plant treating surface/ground water (traditional source waters would be rich in aromatic NOM content), it could have potential implications for future studies conducted using synthetic raw water similar to what was used in this study.

Biomass quantification

Phospholipid analysis of biomass media was conducted by collecting GAC media samples from sampling ports located at 5 cm and 18 cm depths from the top of the GAC media. Results in literature regarding phospholipid analysis and DOC removals have been conflicting (Urfer & Huck 2000; Fonseca et al. 2001; Emelko et al. 2006; Boon et al. 2011; Liao et al. 2013). For instance, Urfer & Huck (2000), observed anthracite phospholipid values of approximately 10–120 nmol P/cm3 media for four dual media anthracite sand filters. It should be noted that despite three of these filters being operated under varying conditions such as periodic ozone dosage and periodic ozone/H2O2 dosage, phospholipid values for each column were similar. For all columns, biodegradable organic matter (BOM) removals (measured as acetate and formate) remained above 95%, thus providing a positive correlation between phospholipid measurements and BOM removal. In an alternate study, Emelko et al. (2006) observed top of the filter biomass concentrations of 18–27 nmol P/cm3 media in dual media anthracite filters and 12–16 nmol P/cm3 media in GAC/sand filters despite both filters showing comparable TOC removal (∼20%).

As noted, there was no clear correlation between biomass concentrations and DOC removal. As can be seen from Figure 4(b) and 4(c), despite low phospholipid counts, %DOC removals remain in the 30% range.
Figure 4

Impact of biomass concentration on DOC removals for all columns (Phase I). (a) 1-NS A; (b) 2-NS B; (c) 3-NL; (d) correlation between phospholipid counts and %DOC removal.

Figure 4

Impact of biomass concentration on DOC removals for all columns (Phase I). (a) 1-NS A; (b) 2-NS B; (c) 3-NL; (d) correlation between phospholipid counts and %DOC removal.

From Figure 4, it is evident that the nutrient supplemented columns displayed higher phospholipid counts (144 nmol P/g media and 126 nmol P/g media on average, respectively, for 1-NS A and 2-NS B) compared to the nutrient limited control column (76 nmol P/g media average). Interestingly, however, top of the filter biomass for both nutrient supplemented columns were comparable (154 nmol P/g media for 1-NS A and 155 nmol P/g media for 2-NS B) and higher than that of the nutrient limited column (57 nmol P/g media for 3-NL). This is also reflected in the higher average %DOC removal for the nutrient supplemented columns as opposed to the nutrient limited columns (Table 3). Statistically, no discernible difference was observed in the biomass concentrations at either 5 cm or 18 cm depth for 1-NS A and 2-NS B (p > 0.05). This was unexpected, as the overall DOC removal over Phase I was slightly greater for 1-NS A than 2-NS B (p > 0.05) and it was expected that the higher removal would be a result of higher biomass concentrations throughout the filter bed. Further, it is apparent that there were large variations in phospholipid values versus DOC removals, and while general trends can be visualized, it was difficult to draw a clear data correlation between DOC removal and phospholipid biomass concentration (Figure 4(d)). This lack of a correlation points to the fact that a better genomic analysis is required for optimizing biofiltration research.

Figure 5(a) shows the phospholipid counts for 1-NS A + K for Phase 2. A steep reduction in biomass concentration can be observed on day 282. Indeed, top of the filter phospholipid values decreased from 474 nmol P/g media on day 274 to 86 nmol P/g media on day 282 (an 82% reduction). Further, Figure 5(b) shows the top of the filter ATP values for 1-NS A + K between days 274 and 285 decreased from 3.273 × 103 ng ATP/g sample to 1.018 × 103 ng ATP/g sample which corresponds to a 69% decrease. This would be in stark contrast to what was observed for 2-NS B and 3-NL (Table 4). Thus, two potential hypotheses are presented here to establish a rationale for the overall reduction in biomass quantity for 1-NS A, and it is believed that either one or both of these theories may be responsible. (a) The phospholipid biomass concentrations reduce as a result of the lower temperature in Phase II (11 °C over the entire phase). Indeed, there is precedence for this in literature, as Emelko et al. (2006) observed markedly reduced phospholipid values for GAC media when studied at temperatures of 21–25 °C (approximately 12–15 nmol P/cm3 of media) versus 1–3 °C (approximately 3–6 nmol P/cm3 of media). (b) It is hypothesized that the reduction in biomass concentration was due to a shedding event that was encountered potentially as a result of potassium addition.
Table 4

ATP values for 2-NS B and 3-NL at 5 cm and 18 cm depths (Phase II)

Days of operation Nutrient condition
 
2-NS B
 
3-NL
 
GAC 1 (5 cm) (ng ATP/g sample) GAC 2 (18 cm) (ng ATP/g sample) GAC 1 (5 cm) (ng ATP/g sample) GAC 2 (18 cm) (ng ATP/g sample) 
259 7.03 × 102 0.66 × 102 2.66 × 102 1.97 × 102 
274 11.99 × 102 16.52 × 102 1.24 × 102 2.49 × 102 
285 9.78 × 102 11.23 × 102 2.12 × 102 2.97 × 102 
293 16.88 × 102 17.6 × 102 2.39 × 102 2.64 × 102 
300 12.48 × 102 10.99 × 102 2.09 × 102 2.49 × 102 
308 4.96 × 102 0.86 × 102 0.6 × 102 0.16 × 102 
315 5.85 × 102 1.99 × 102 1.05 × 102 0.82 × 102 
Days of operation Nutrient condition
 
2-NS B
 
3-NL
 
GAC 1 (5 cm) (ng ATP/g sample) GAC 2 (18 cm) (ng ATP/g sample) GAC 1 (5 cm) (ng ATP/g sample) GAC 2 (18 cm) (ng ATP/g sample) 
259 7.03 × 102 0.66 × 102 2.66 × 102 1.97 × 102 
274 11.99 × 102 16.52 × 102 1.24 × 102 2.49 × 102 
285 9.78 × 102 11.23 × 102 2.12 × 102 2.97 × 102 
293 16.88 × 102 17.6 × 102 2.39 × 102 2.64 × 102 
300 12.48 × 102 10.99 × 102 2.09 × 102 2.49 × 102 
308 4.96 × 102 0.86 × 102 0.6 × 102 0.16 × 102 
315 5.85 × 102 1.99 × 102 1.05 × 102 0.82 × 102 
Figure 5

Biomass concentration and potassium removal for 1-NS A + K. (a) Phospholipid profile; (b) ATP profile; (c) %Potassium removal.

Figure 5

Biomass concentration and potassium removal for 1-NS A + K. (a) Phospholipid profile; (b) ATP profile; (c) %Potassium removal.

As seen in Figure 5(a), the addition of potassium appears to have had a positive impact on phospholipid biomass growth initially. Indeed, the average phospholipid values prior to the shedding event show a statistical difference (p < 0.05) when compared with the phospholipid counts for 1-NS A from Phase I. Thus, potassium supplementation did have a positive impact on biomass growth.

However, as per the potential shedding event indicated in Figure 5(a) and 5(b), perhaps the quantity of potassium being dosed in this study far exceeded the requirements of a bacterial cell in a drinking water treatment scenario. Brdjanovic et al. (1996) provided the potassium to phosphorus mol ratio (1 mol of potassium per mole of phosphorus) on which this study is based. It should be noted however that that particular study was conducted on wastewater systems, which are known to have higher organic loading than drinking water systems. Figure 5(c) indicates the potassium removal trends for phase II. It was believed that studying this trend would give an idea as to the microbial community's potassium uptake ability. From Figure 5, it is evident that around day 281 0% removal of potassium was observed, whereas for the previous 20 days removals were in the 10–80% range.

It has been established that potassium is a key intracellular component responsible for regulating the internal osmotic pressure of bacterial cells, and aids in maintaining cell turgor (rigidity) depending on the osmolarity of the cell environment (Bertrand 2015). Anderson et al. (2008) sought to monitor potassium as a potential surrogate for endotoxin release following filter shut down. They observed an increase in potassium concentrations at the effluent following periods of 18 hour shut downs. They reasoned that within that shut down period, the environment within the biofilter changed to one where the bacteria needed to release potassium to maintain internal osmotic pressure. Cell lysis is also provided as a potential reason for this increase in effluent potassium. Thus, it is possible in our case that potassium accumulation reached a stage where it hindered the physiological and chemical balances within the cell wall (such as its internal osmotic pressure), resulting in cell lysis. Indeed, the data does somewhat support this theory, with higher effluent potassium concentrations observed on days 289 and 296 (Figure 5(c)), which could be caused as a result of cell lysis. As mentioned previously, Figure 5(a) and 5(b) show reductions of 82% and 69% in top of the filter phospholipid and ATP biomass concentrations, respectively, from day 274 to day 282 (phospholipid) and day 274 to day 285 (ATP). This would further support the cell lysis theory. Following this shedding event, Figure 5(a) indicates an increasing trend in phospholipid biomass from day 297 onwards, which corresponds to an increase in potassium removal (uptake) observed in Figure 5(c) around this time.

CONCLUSIONS

The impacts of nitrogen, phosphorus and potassium supplementation on organics removal in drinking water biofilters were studied. In addition, a comparison of water wash only and air-scour assisted backwash were studied with a biofilter under nutrient limited conditions.

  • Nutrient limited conditions (C:N:P = 546:24:1 on a w/w basis) resulted in DOC removals of 15% with a water only backwash condition and improved to 23% with an air-assisted backwash.

  • Nitrogen and phosphorus supplementation improved DOC removal to 32.2% (C:N:P 25:5:1) and 27.5% (C:N:P 100:10:1).

  • No clear correlation could be drawn between phospholipid counts of biomass and ATP concentrations with DOC removal. It should be noted that phospholipid analysis is not necessarily representative of the active biomass, while ATP is only found within living, viable cells. However, neither resulted in a good indicator of DOC removal in this study.

  • Potassium supplementation of 1 mol K per mol P did not impact DOC removals. However, it may have contributed to a bacterial sloughing event in the biofilters, based on an observed reduction in ATP values (i.e. 3.27 × 106 pg ATP/g sample down to 1.02 × 106 pg ATP/g sample). Future research should consider alternative ratios of potassium to phosphorus and other micronutrients to determine their influence in biofiltration DOC removal.

ACKNOWLEDGEMENTS

The authors would like to recognize NSERC (National Science and Engineering Research Canada) for financially supporting this project. In addition, we would like to give our thanks to all the members of the Basu Research Group who supported this project, as well as Robert (Bob) LeCraw from MSFilter Ltd and Shawn Cleary from Humber College, Toronto, ON for their input into practical issues with biofiltration.

REFERENCES

REFERENCES
Anderson
W. B.
Mayfield
C. I.
Huck
P. M.
2008
Endotoxin release from biologically active bench-scale drinking water anthracite/sand filters
.
Journal of Water Supply: Research and Technology - AQUA
57
(
8
),
585
597
.
APHA/AWWA/WEF
2012
Standard Methods for the Examination of Water and Wastewater
,
22nd edn
.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
Azzeh
J.
Taylor-Edmonds
L.
Andrews
R. C.
2015
Engineered biofiltration for ultrafiltration fouling mitigation and disinfection by-product precursor control
.
Water Science and Technology
15
(
1
),
124
133
.
Badawy
M. I.
Gad-Allah
T. A.
Ali
M. E. M.
Yoon
Y.
2012
Minimization of the formation of disinfection byproducts
.
Chemosphere
89
(
3
),
235
240
.
Bertrand
J.
2015
Environmental Microbiology: Fundamentals and Applications
.
Springer
,
Dordrecht
,
The Netherlands
.
Brdjanovic
D.
Hooijmans
C. M.
van Loosdrecht
M. C. M.
Alaerts
G. J.
Heijnen
J. J.
1996
The dynamic effects of potassium limitation on biological phosphorus removal
.
Water Research
30
(
10
),
2323
2328
.
Emelko
M. B.
Huck
P. M.
Coffey
B. M.
Smith
E. F.
2006
Effects of media, backwash, and temperature on full-scale biological filtration
.
Journal American Water Works Association
98
(
12
),
61
73
.
Feng
S.
Chen
C.
Wang
Q. F.
Zhang
X. J.
Yang
Z. Y.
Xie
S. G.
2013
Characterization of microbial communities in granular activated carbon-sand dual media filter for drinking water treatment
.
International Journal of Environmental Science and Technology
10
(
5
),
917
922
.
Fonseca
A. C.
Summers
R. S.
Hernandez
M. T.
2001
Comparative measurements of microbial activity in drinking water biofilters
.
Water Research
35
(
16
),
3817
3824
.
Lauderdale
C.
Chadik
P.
Kirisits
M. J.
Brown
J.
2012
Engineered biofiltration: enhanced biofilter performance through nutrient and peroxide addition
.
Journal American Water Works Association
104
(
5
),
E298
E309
.
LeChevallier
M. W.
Schulz
W.
Lee
R. G.
1991
Bacterial nutrients in drinking water
.
Applied and Environmental Microbiology
57
(
3
),
857
862
.
Liao
X.
Chen
C.
Wang
Z.
Wan
R.
Chang
C. H.
Zhang
X.
Xie
S.
2013
Changes of biomass and bacterial communities in biological activated carbon filters for drinking water treatment
.
Process Biochemistry
48
(
2
),
312
316
.
Liu
X.
Huck
P. M.
Slawson
R. M.
2001
Factors affecting drinking water biofiltration: filtration
.
Journal American Water Works Association
93
(
12
),
90
101
.
Moll
D. M.
Summers
R. S.
Fonseca
A. C.
Matheis
W.
1999
Impact of temperature on drinking water biofilter performance and microbial community structure
.
Environmental Science and Technology
33
(
14
),
2377
2382
.
Nyfenneger
J.
Lauderdale
C.
Brown
J.
Scheitlin
K.
2013
Engineered biofiltration for drinking water treatment: optimization and strategies to enhance performance
.
Technical article.
Florida Water Resources Journal
11
,
12
17
.
Romani
A. M.
Giorgi
A.
Acuna
V.
Sabater
S.
2004
The influence of substratum type and nutrient supply on biofilm organic matter utilization in streams
.
Limnology and Oceanography
49
(
5
),
1713
1721
.
Rosenberg
H.
1987
Phosphate transport in prokaryotes
. In:
Ion Transport in Prokaryotes
,
1st edn
(
Rosen
B. P.
, ed.).
Academic Press
,
Chicago
,
USA
, pp.
85
130
.
Scharf
R. G.
Johnston
R. W.
Semmens
M. J.
Hozalski
R. M.
2010
Comparison of batch sorption tests, pilot studies, and modeling for estimating GAC bed life
.
Water Research
44
(
3
),
769
780
.
Tchobanoglous
G.
Burton
F. L.
Stensel
H. D.
2013
Wastewater Engineering: Treatment and Reuse
.
McGraw Hill
,
New York
,
USA
.
Wang
J. Z.
Summers
R. S.
Miltner
R. J.
1995
Biofiltration performance: part 1, relationship to biomass
.
Journal American Water Works Association
87
(
12
),
55
63
.
Zhang
D.
Li
W.
Wang
K.
Zhang
L.
2010
Bacterial community dynamics and its effects during biological activated carbon filter process for drinking water treatment
. In:
2nd International Conference on Chemical, Biological and Environmental Engineering
.