Nitrogen dynamics and removal in a horizontal flow biofilm reactor for wastewater treatment

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Nitrogen dynamics and removal in a horizontal flow biofilm reactor for wastewater treatment E. Clifford a,*, M. Nielsen a, K. Sørensen b, M. Rodgers a a b

Department of Civil Engineering, National University of Ireland, Galway, Ireland Danish Technological Institute, Teknologiparken, Kongsvang Alle´ 29, DK-8000 Aarhus C, Denmark

article info

abstract

Article history:

A horizontal flow biofilm reactor (HFBR) designed for the treatment of synthetic waste-

Received 2 October 2009

water (SWW) was studied to examine the spatial distribution and dynamics of nitrogen

Received in revised form

transformation processes. Detailed analyses of bulk water and biomass samples, giving

26 April 2010

substrate and proportions of ammonia oxidising bacteria (AOB) and nitrite oxidising

Accepted 27 April 2010

bacteria (NOB) gradients in the HFBR, were carried out using chemical analyses, sensor rate

Available online 6 May 2010

measurements and molecular techniques. Based on these results, proposals for the design of HFBR systems are presented.

Keywords:

The HFBR comprised a stack of 60 polystyrene sheets with 10-mm deep frustums. SWW

Biofilm reactor

was intermittently dosed at two points, Sheets 1 and 38, in a 2 to 1 volume ratio respec-

FISH

tively. Removals of 85.7% COD, 97.4% 5-day biochemical oxygen demand (BOD5) and 61.7%

Potential nitrification rates

TN were recorded during the study.

Horizontal flow biofilm reactor

In the nitrification zones of the HFBR, which were separated by a step-feed zone, little

Wastewater treatment

variation in nitrification activity was found, despite decreasing in situ ammonia concentra-

Reactor optimisation

tions. The results further indicate significant simultaneous nitrification and denitrification (SND) activity in the nitrifying zones of the HFBR. Sensor measurements showed a linear increase in potential nitrification rates at temperatures between 7 and 16
C, and similar rates of nitrification were measured at concentrations between 1 and 20 mg NH4-N/l. These results can be used to optimise HFBR reactor design. The HFBR technology could provide an alternative, low maintenance, economically efficient system for carbon and nitrogen removal for low flow wastewater discharges. ª 2010 Elsevier Ltd. All rights reserved.

1.

Introduction

The on-site remediation of low flow wastewater sources requires robust, effective systems that are easy and economic to operate and maintain, and employ simple technologies (Crites and Tchobanoglous, 1998). Technologies vary from septic tanks with percolation areas to package aeration systems with polishing filters such as sand, soil and peat filtration systems (Johnson and Atwater, 1988; Corley et al., 2006). Treatment systems can vary from activated sludge

(e.g. continuous flow or sequencing batch reactor) or attached growth systems (e.g. trickling filters or rotating biofilm contactors) or a combination thereof (USEPA, 1980). To meet the stringent standards for wastewater treatment in the EU, small-scale systems are required to incorporate steps that ensure efficient nitrogen (N) removal, primarily involving the elimination of ammonium-nitrogen (NH4-N) and in sensitive areas a reduction in nitrate-nitrogen (NO3-N) (Water Framework Directive, 2000/60/EC; Groundwater Directive, 2006/118/EC; Surface Water Directive, 75/440/EEC). If a simultaneous

* Corresponding author. Tel.: þ353 (0)91 492762; fax: þ353 (0)91 494507. E-mail address: [email protected] (E. Clifford). 0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2010.04.042

3820

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

Nomenclature AOB BOD5 COD CODf DO FISH HFBR HRT MLSS N NH4-N NH4-Nf NOB NO2-N NO2-Nf

Ammonia oxidising bacteria 5-Day biochemical oxygen demand Chemical oxygen demand Filtered COD Dissolved oxygen Fluorescent in situ hybridization Horizontal flow biofilm reactor Hydraulic retention time Mixed liquor suspended solids Nitrogen Ammonium-nitrogen Filtered NH4-N Nitrite oxidising bacteria Nitrite-nitrogen Filtered NO2-N

reduction in NH4-N and NO3-N is required, the establishment of suitable conditions to facilitate both nitrification and denitrification is necessary. This represents a challenge to engineers when designing on-site wastewater treatment systems, because the two processes are catalyzed by physiologically distinct groups of micro-organisms i.e., autotrophic nitrifiers and heterotrophic denitrifiers, which have fundamentally different metabolic requirements (Lowe et al., 2008). Nitrogen removal is normally realized by sequentially alternating between oxic and anoxic conditions or by the creation of separated zones with suitable conditions for nitrification and denitrification, respectively. Alternatively, high rates of simultaneous nitrification and denitrification (SND) can be achieved, in activated sludge and biofilm type systems alike, at operational conditions where both oxic and anoxic micro-environments are present (Randall, 1992; Henze et al., 1997). Nitrification can occur at the liquid/biomass interface, while denitrification of nitrate (or nitrite) may be found in deeper sub-surface biomass zones (Von Mu¨nch et al., 1996; Helmer and Kunst, 1998). Combined carbonaceous oxidation, nitrification and denitrification systems, based on complex and high flow systems, can require frequent and costly maintenance (Ferguson et al., 2003; EPA, 2007). The horizontal flow biofilm reactor (HFBR) is a simple technology with a flexible design that provides a new effective technology for carbon and nitrogen removal from domestic wastewaters (Rodgers et al., 2004; Rodgers and Clifford, 2009). In contrast to conventional trickling filters, sand filters and other media filters, the HFBR technology enables easy access to the different reactor regions, hence allowing for the examination of contaminant and biomass distributions, microbial community structures and process transformation rates in various regions of the reactor. Such information can help scientists and engineers to understand the factors affecting treatment performance and lead to more efficient system design (Lydmark et al., 2006). In this study an HFBR was monitored for its efficacy in removing carbon and nitrogen from a domestic strength synthetic wastewater (SWW). The spatial arrangement of microbial populations in an HFBR, with a particular focus on

NO3-N NO3-Nf NOX ORP PE SND SS SRT STE SWW TN TNf TS TPA TSPA VSS WWTP

Nitrate-nitrogen Filtered NO3-N Oxidised nitrogen Oxidationereduction potential Persons equivalent Simultaneous nitrification and denitrification Suspended solids Solids retention time (days) Septic tank effluent Synthetic wastewater Total nitrogen Filtered TN Total solids Total plan area (m2) Top surface plan area (m2) Volatile suspended solids Wastewater treatment plant

nitrogen dynamics, was investigated. To obtain detailed information on system performance, traditional methodologies for characterisation of wastewater parameters were combined with advanced bio-sensing technology and molecular tools. This approach provided data that can be used by engineers and scientists to design HFBR systems and optimise nitrogen removal.

2.

Materials and methods

2.1.

Design and operation

The HFBR comprised a vertical stack of 60 horizontal polystyrene sheets (Terram Ltd.) that were arranged in sets of 10 sheets, on polypropylene shelves positioned one above another and supported on a simple frame (Fig. 1). Each sheet measured 340 mm  310 mm in plan and had 10-mm deep frustums, giving a potential volume of 1.24 l/sheet. Three sides of each sheet were turned up to a height of 10 mm in order to direct the flow along the sheets to its exit end. The top surface plan area (TSPA) of each sheet was 0.1054 m2 (0.34 m  0.31 m). The total plan surface area (TPA) of the media was 6.324 m2 (TSPA  60). SWW was pumped from a feed tank using a Masterflex L/S economy peristaltic pump. The SWW was applied intermittently for 10 min every hour onto the top sheet at 29.4 l/day (Sheet 1) and at a step-feed point at 14.7 l/day (Sheet 38). This equated to a TSPA hydraulic load of 418.4 l/m2 day on the unit. Wastewater flowed along one sheet, dropped to a sheet underneath, and then flowed in the opposite direction on that sheet, and so on down through the reactor. The pump was calibrated, the tubing cleaned, and the feed tank cleaned and refilled every second day. The SWW (Table 1) was based on a mix by Odegaard and Rusten (1980). The unit, seeded on Day 0, with activated sludge taken from a municipal activated sludge treatment plant (WWTP), was located in a temperature controlled room at 11
C and the system was operated for a total of 450 days. The STU was seeded with 30 l of SWW mixed with 15 l of nitrifying activated sludge taken from the

3821

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

Fig. 1 e Schematic of step-feed HFBR.

WWTP. The mixed liquor suspended solid (MLSS) concentration of the activated sludge was 5478 mg SS/l. This was dosed onto the unit as per the normal dosing regime, which is detailed above. When the seed mixture was used up, the influent feed tank was cleaned and the SWW alone was used as a substrate for the study.

2.2.

Sampling and analyses

During the study, samples of the influent were taken from the feed tank and effluent samples were taken as the treated wastewater dropped off the final sheet (Sheet 60). Samples were taken from various sheets down through the reactor using a 10 ml pipette with a 3-mm tip opening. Filtered

chemical oxygen demand (CODf), suspended solid (SS) samples and volatile suspended solids (VSS), after filtration using 1.2 mm Whatman GF/C microfibre filters, were tested in accordance with standard methods (APHA, 1995). Filtered and unfiltered TN was measured using a DR/2010 spectrophotometer (HACH Company). Filtered NH4-N, NO2-N and NO3-N were measured using a Konelab 20 Nutrient Analyser. Dissolved oxygen (DO) concentrations were obtained with a WTW model CellOx 325 electrode and a WTW 330 meter, pH with a WTW SenTix 21 pH electrode and a WTW pH 320 meter and oxidationereduction potential (ORP) with a Dolmen 23 Redox electrode and a WTW pH 340/ION meter. Unfiltered 5-day biochemical oxygen demand (BOD5) was analysed using WTW OxiTop meters. Particle sizes were

Table 1 e Composition and characteristics of SWW. Standard deviation shown in parenthesis. Constituent Glucose Yeast extract Dried milk Urea NH4Cl Na2NPO4$12H2Oa KHCO3a

Concentration (mg/l)

Constituent

Concentration (mg/l)

Contaminant

200.0 30.0 120.0 30.0 60.0 100.0 50.0

NaHCO3a MgSO4$7H2Oa FeSO4$7H2Oa MnSO4$H2Oa CaCl2$6H2Oa

130.0 50.0 2.0 2.0 3.0

Total COD Soluble COD Total BOD5 TN Total NH4-N Total NO3-N SS

a Used in experimental solution for rate measurement experiments.

Concentration (mg/l) 387.7 356.0 283.2 37.6 23.4 0.7 56.1

(55.9) (39.8) (31.2) (3.8) (2.2) (0.2) (17.7)

3822

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

Table 2 e Formamide concentrations and washing buffer stringencies for FISH studies. Probe

Sequence

EUB338 NSO1225

GCT GCC TCC CGT AGG AGT CGC CAT TGT ATT ACG TGT GA

Ntspa685 Nit3 CompNit3

CAC CGG GAA TTC CGC GCT CCT C CCT GTG CTC CAT GCT CCG CCT GTG CTC CAG GCT CCG

Target organisms

% Formamide

Reference

Most bacteria Ammonium-oxidising b-proteobacteria Genus Nitrospira (388 of 759) Genus Nitrobacter (77 of 104)

0e50 35

Amann et al., 1990 Mobarry et al., 1996

20 40

Hovanec et al., 1998 Wagner et al., 1996 Wagner et al., 1996

analysed using a Micro Plus (Mastersizer TM) laser analyser with a size range of 0e500 mm. All instruments were calibrated in accordance with the manufacturers’ instructions before testing.

2.3.

Process rates

Nitrification and denitrification rate analyses of HFBR biomass were performed in a 300-ml batch reactor (CAL 300, Unisense A/S). The reactor had ports for sensor insertion and was equipped with an air stone for oxygen control and biomass mixing. For temperature control, the reactor was placed in a refrigerated water bath. Nitrification (ammonia oxidation) and denitrification (nitrite reduction) rates were measured on-line using   a sensitive macroscale NO x (NO2 þ NO3 ) biosensor (Unisense A/S) with a detection limit of about 1 mM NO x and a 90% response time of about 3 min (Nielsen et al., 2002). The biosensor gave precise information on process transformation rates of the small biomass samples that were retrieved from the HFBR system. The sensor was calibrated (2-point) in situ in the batch reactor through 25 and 100 ml additions from a 100-mM NO 3 stock solution. A Clark type mini-scale O2 sensor with a 300-mm tip and a response time (t90) of about 5 s was used to monitor the oxygen water level in the batch reactor. Suspended biomass samples (15e30 ml) were collected from HFBR sheets with a 20-ml syringe. To remove back ground concentrations of NHþ 4 , NOx and organic constituents, the samples were washed three times in 200 ml of experimental solution, which comprised the SWW constituents, but without the nitrogen and organic components (Table 1). In each washing step, a sample was allowed to settle for about 20 min before decanting occurred. The remaining sample was then diluted, using the experimental solution, to a working volume of 150 ml and was transferred to the batch reactor. After sensor calibration and establishment of the different test conditions (see below), the production or consumption of oxidised N was monitored with the NO x biosensor, typically for a period of about one hour, to obtain a stable process rate measurement. For the kinetic studies on nitrification, rate estimates were obtained at different ammonia levels in sequence by stepwise addition of substrate to the reactor. All experiments were replicated for three biomass samples with the exception of the kinetic studies that were conducted for one sample only. The sample biomass density (VSS/ml) was determined at the end of the experiment. The potential nitrification activity was measured for biomass samples from ten sheet locations, in the presence of 20 mg/l NH4-N and at air-saturated oxygen conditions (air bubbling).

The effect of temperature on nitrification activity was examined in a range between 7 and 16
C for biomass from Sheet 29, and the effect of ammonia substrate was examined at levels between 0.5 and 20 mg/l NH4-N for Sheets 17, 29, and 38. Potential denitrification rates were determined for biomass from four sheets below the step-feed point, in the presence of 200 mg/l COD, 2.5 mg/l NO3-N and anoxic conditions (N2 bubbling). Substrate additions to the batch reactor were made  from stock solutions: 100 mM NHþ 4 , 100 mM NO3 , and 10,000 mg/l COD.

2.4.

FISH studies

HFBR biomass samples were removed using a pipette, and 0.5 ml of material was fixed overnight at 4
C in 2 volumes of PBS buffer (1.09 g l1 Na2HPO4, 0.32 g l1 NaH2PO4, 9.0 g l1 NaCl, pH 7.2) with 4% paraformaldehyde. The fixed sample was centrifuged (10,000 g, 1 min), and the pellet was washed twice by re-suspension and centrifugation in 1 PBS buffer. The sample was stored at 20
C in 1 ml of a 1:1 ethanol/PBS solution. Prior to hybridization, 0.1 ml of fixed sample was spotted onto gelatinized (0.1% gelatine þ 0.01% CrKSO4) glass microscope slides, air dried, and dehydrated by immersion in 50, 70 and 100% ethanol solutions (5 min each). FISH was carried out using CY3-labelled probes for general as well as ammonia- or nitrite-oxidising bacteria, probespecific formamide concentrations and washing buffer stringencies from the literature (Table 2). Hybridization and washing buffers were prepared as previously described (Amann et al., 1990; Wagner et al., 1996). DAPI (final concentration 50 ng/ml) was added to the hybridization buffer in order to obtain total cell numbers in the samples. Each spot on the microscope slide was covered with 100 ml of hybridization buffer and hybridization was performed for 2 h at 46
C. Slides were subsequently immersed in washing buffer at 48
C for 15 min and finally air dried. Microscopy was performed on a Nikon epifluorescence microscope equipped with suitable filters and digital imaging system. Image analysis was performed using UVIgel software. The percentage of AOB and NOB to total bacteria was evaluated by comparing the area of cells hybridized with each probe to the total area covered by bacteria stained with DAPI. At least 50 randomly distributed microscope images taken at 1000 magnification were used for each quantification.

3.

Results

3.1.

HFBR system performance

The HFBR system was operated for a total of 450 days. After an initial start-up period with gradually improving performance

3823

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

Table 3 e Overall performance of the HFBR. Steady state (average of 66 observations between Days 31e450) Contaminant

Average influent (mg/l)

St. deviation influent

Average effluent (mg/l)

St. deviation effluent

% Reduction

387.7 356.0 283.1 59.1 22.0 0.7 37.6 31.3 11.5

55.9 39.0 26.1 17.1 4.6 0.2 3.8 4.9 0.5

55.6 49.5 7.5 16.2 1.0 11.7 14.4 11.7 9.9

13.7 13.7 3.9 9.5 0.4 2.9 1.1 1.1 2.5

85.7 86.1 97.4 72.6 95.7

Total COD CODf BOD5 SS Total NH4-N Total NO3-N TN TNf Total PO4-P

(Days 0e30), the system stabilised and remained in pseudo steady state until the end of the experimental period. Between Days 31 and 450 average removals of 97.4% BOD5, 85.7% COD, 95.7% NH4-N and 61.7% TN were observed (Table 3). The stability of the system during this period was confirmed by consistent results over time (Fig. 2). The hydraulic retention time measured using a bromide tracer, between Days 252 and 256, was 11.1 h.

3.2.

e 61.7 62.6 13.9

the formation of a thick black/grey biofilm. A slight sulphurous smell and a distinct grey/white surface layer revealed the presence of anaerobic conditions inside biofilm surface layers and activity of sulphide-oxidising organisms. Thinner, brownish biofilms were characteristic of the sheets from the nitrification zones. Biofilms from the step-feed region, between Sheets 38e40, were similar in colour to that observed at the top of the unit, but without the presence of a white layer.

Biomass analysis

Biomass concentration and characteristics varied significantly within the HFBR. VSS concentrations in samples ranged from 3800 to 8700 mg/l in the mineralisation/carbon removal zone (Sheets 1e10), decreasing to below 2000 mg/l in the nitrification zones (Sheets 10e37 and Sheets 40e60, respectively). VSS concentrations in the step-feed region (Sheets 38e40) averaged about 3400 mg/l. During steady state solids concentrations on each sheet remained reasonably steady indicating low excess solid production. The largest average particle sizes were observed on Sheets 1 and 40, in both cases of about 215 mm. Average particle sizes were significantly smaller in the nitrification zones with minimum averages of about 40 mm measured on Sheets 29, 31 and 58 (Fig. 3). Visual observations of the biomass indicated a pronounced zonation of the microbial community. The carbon rich conditions at the top between Sheets 1 and 14 of the unit led to

3.3.

Reactor processes

3.3.1.

Carbon removal, mineralisation and dissolved oxygen

Sheets 1e14 and Sheets 38e40 (below the step-feed) were characterised by high rates of carbon removal and mineralisation (Fig. 4). The COD removal rate in the top 14 sheets can be related to the sheet number or surface area, in this region, using a linear relationship;  y  17:907  ; CODfy ¼ 0:0577

R2 ¼ 0:9962

(1)

where y ¼ sheet number counted vertically from the top and [CODfy] ¼ concentration of CODf on sheet y.

VSS and SS (mg/l) 0

2000

4000

6000

8000

10000

12000

0

SS VSS Average Size 10

Effluent COD Influent COD Influent TN Effluent TN

20 50

40 300

30

200

TN ( m g / l)

COD ( m g / l )

400

Sh eet Nu m b er

500

60

30

40

20

50 100

10

0

0

50

100

150

20 0

250

300

35 0

400

0 450

Days

Fig. 2 e Influent and effluent COD and TN concentrations during steady state.

60 50

70

90

110

130

150

170

190

210

230

250

Average size (mm)

Fig. 3 e SS, VSS and average particle sizes during steady state. Standard error bars shown for VSS and average size.

3824

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

3.3.2.

Nitrification

Between sheets 10e37 and 40e60 almost full nitrification occurred (Fig. 5), with reductions from about 28 to 2 mg NH4Nf/l and 10 to 1 mg NH4-Nf/l, respectively. For each zone the NH4-Nf removal rate was related to sheet number as follows;

Nitroge n (mg/l) 0

10

20

30

0

10

20

Sheet Number

On Sheet 14 most biodegradable organic matter had been removed. At the step-feed the CODf concentration increased to about 95 mg/l; about 70% of CODf removal occurred between Sheets 38 and 40. On Sheet 1, the calculated surface removal rate was about 16.3 g CODf/m2.day. In comparison, the surface removal rate between Sheets 38 and 40 was 11.2 g CODf/m2.day. The measured DO concentrations in the top reactor section and in the step-feed region were below about 2 mg/l. The DO concentrations increased steadily in the nitrification zones from Sheets 10 to 37 and Sheets 40 to 60.

30

40

NH4-N 50

NO 3-N TN

 y 36:852  ; NH4  Nfy ¼ 0:997  y60:71  ; NH4  Nfy ¼ 1:837

R2 ¼ 0:982; ð10  y  37Þ

(2a)

R2 ¼ 0:968; ð41  y  60Þ

(2b)

where, y ¼ sheet number and, [NH4-Nfy] ¼ concentration of soluble NH4-N on Sheet y. The area specific nitrification rate between Sheets 10 and 37 was about 30% higher than between Sheets 41 and 60. As a result of nitrification, NO3-N concentrations increased in these regions, from 3.4 to 17.4 mg/l (between Sheets 10 and 37) and from 6 to 11 mg/l (between Sheets 41 and 60), respectively (Fig. 5). NO2-N concentrations remained below 1.5 mg/l.

3.3.3.

Denitrification

Average net reductions from 12 mg NO3-N/l to 6 mg NO3-N/l were measured between Sheets 38 (step-feed) and 40. The additional carbon introduced by the step-feed was nearly completely removed by Sheet 40. Removals of 2.5 g CODf/g

CO D (mg/l) 0

50

100

150

200

250

300

350

400

0

10

Sheet Number

20

30

COD DO Linear (COD)

40

50

60

0

1

2

3

4

5

6

7

8

9

DO (mg/l)

Fig. 4 e Profile CODf and DO during steady state. Standard error bars shown.

60

Fig. 5 e Nitrogen profiles during steady state. Standard error bars shown.

NO3-N, and 4 g CODf/g TN were observed between Sheets 38 and 40. TN concentrations decreased throughout the upper and lower nitrification zones (Fig. 5).

3.4.

Sensor rate measurements

In a broad zone between Sheets 10e37 and again between Sheets 43e60 the biomass demonstrated high ammonia oxidising capacity (Fig. 6(a)). The measured rates in the upper nitrification zone before the step-feed point were generally higher than in the lower zone, with maximum productions of about 0.02 mg NO x -N/mg VSS day and 0.006 mg NO x -N/mg VSS day, respectively. Markedly lower specific activities were observed for biomass samples from areas with high COD availability (Sheets 8 and 40), showing average production rates between 0.001 and 0.002 mg NO xN/mg VSS day. Denitrification rate measurements revealed potential rates  of 0.17 mg NO x -N/mg VSS day on Sheet 40, 0.09 NOx -N/mg  VSS day on Sheet 43, 0.07 NOx -N/mg VSS day on Sheet 28 and 0.05 NO x -N/mg VSS day at Sheet 54. The potential rate dropped by about 50% between Sheets 40 and 43, with a more gradual reduction, thereafter. This profile correlated well with the CODf profiles. Rate measurements on Sheets 17, 29 and 37 of the reactor showed that nitrification activity was affected by ammonia availability at concentrations between 0.5 and 5 mg NH4-N/l (Fig. 6(b)). Compared to the measured maximum activities, the process rate at 0.5 mg NH4-N/l was reduced in the range 20e30%. Using the experimental results from Sheets 17, 29 and 37, the half saturation constant (KS-NH), at 11
C, was calculated at 0.15 mg NH4-N/l (standard deviation 0.05 mg NH4-N/l). Ranges of between 0.15 and 5.4 mg NH4-N/l have been reported by Copp and Murphy (1995). It should be noted that a somewhat higher apparent value for KS-NH is expected

3825

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

a

3.5.

-

NOx production (mg/mg VSS.day) 0

0.01

0.02

0.03

0

Sheet number

10 Production

20

Consumption 30 40 50 60 0

0.05

0.1

0.15

0.2

-

NOx consumption (mg/mg VSS.day)

b 0.015

Sheet17 Sheet 29 Sheet 37

0.005

0

0

5

10

15

0.015 Sheet 29 y = 0.0009x - 0.0018 2 R = 0.9882

0.01

-

0.005

0 0

5

10 o Temperature ( C)

15

4.

Discussion

4.1.

Nitrogen transformation processes in the HFBR

20

+ NH4 -N (mg/l)

c NOx pro ductio n (mg /mg VSS.da y)

Probes for general bacteria, ammonium-oxidising betaproteobacteria (AOB), and the nitrite-oxidising genera Nitrospira and Nitrobacter (NOB) were applied to describe the spatial distribution of nitrifiers in the HFBR. The FISH probe EUB338 targets most but not all bacterial groups. In particular members of the Planctomycetales and the Verrumicrobia may not be targeted though. Their presence has been found to be low in other systems (Witzig et al., 2002; Mattila et al., 2007). If these groups were present in the reactor, the ratio of ammonium and nitrite oxidizers is likely to have only been slightly overestimated (Daims et al., 1999). AOB as well as Nitrospira type NOB were observed throughout the unit (Fig. 7). No hybridization was observed with the probe Nit3, which is specific for members of the genus Nitrobacter. The AOB and NOB proportions of the total cell counts were low on sheets at the top of the reactor, increased between Sheets 12 and 35, decreased dramatically at the step-feed, and increased again below this. This is consistent with a high activity of AOB and NOB mainly on sheets with low COD levels.

0.01

-

NOx pro ductio n (mg / mg VSS.da y)

0.02

FISH studies

20

The boustrophedonic nature of the water flow regime leads to the formation of distinct COD, oxygen, NH4-N and NO3-N gradients along the vertical axis of the HFBR. The studied unit could broadly be divided into: (1) a top zone (Sheets 1e15) with reduced DO concentrations and high organic carbon availability; (2) an upper oxidised zone (Sheets 16e37) with low organic carbon availability; (3) a narrow anoxic zone (Sheets 38e40) at the step-feed with high organic carbon availability; and (4) a lower oxidised zone (Sheets 41e60) with low organic carbon availability. Nitrifying populations were observed in most regions of the HFBR, and nitrifying zones accounted for more than 80% of the total reactor surface area. The nitrifier presence was

Fig. 6 e (a) Potential nitrification and denitrification rates. Standard error bars shown. (b) Effect of NH4-N concentration on potential nitrification rates for 3 sheets. (c) Effect of temperature on potential nitrification rates for Sheet 29. Standard error bars shown.

1

2

3

4

0

10

Sheet number

for the HFBR unit, compared with the measured kinetics, due to higher mass transfer limitations in the non turbulent flow conditions of the HFBR. The potential ammonia oxidation rate varied significantly with temperature in the tested range between 7 and 16
C. The activity/temperature correlation can be described by a linear
function of about 0.001 mg NO x -N/mg VSS day C (Fig. 6(c)). In
comparison, the rate at 16 C was about a factor of 3 higher than at 7
C.

% AOB and NOB of total populations 0

AOB

NOB

20

30

40

50

60

Fig. 7 e AOB and NOB distribution in the HFBR.

5

3826

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

consistent with in situ CODf and DO concentrations. The area specific activity in the lower nitrification zone (Sheets 41e60) was approximately 30% lower than in the upper nitrification zone (Sheets 16e37). This variation in dynamics may be related to the lower retention time on each sheet due to the increased flow from the step-feed. The average filtered TN removal of the step-feed HFBR was 62.6% (Table 3). With clarification and removal of effluent SS this could be increased to 70%. The low excess solid production indicates that much of the net nitrogen removal occurred due to denitrification processes with cell synthesis accounting for limited net nitrogen removal. In the step-feed region of the reactor, optimal conditions for denitrification existed due to the convergence of the high NO 3 /low COD HFBR stream and the low NO 3 /high COD step-feed stream. Over 20% of the nitrogen was removed between Sheets 38 and 40. Removal ratios were similar to reported ratios for activated sludge systems, of 4.0 g BOD5/g NO3-N removed, 2.9 g BOD5/g NO3-N removed, and 3.5e4.5 g COD/g TN removed (Barth, 1968; Gray, 1999; Henze et al., 1997), respectively. Interestingly, about two thirds of the total N removal occurred in other reactor zones, which suggests the presence of broad reactor zones with SND activity. The sheet design with 10-mm deep frustums may have helped the development of anoxic regions. Biesterfield et al. (2003) noted denitrification occurring within biofilms even when the bulk aqueous phase had DO concentrations of up to 5 mg O2/l. Suspended biomass, transported along with the water flow, may expose heterotrophic biomass to changing environmental conditions, resulting in cell decay, and hence provide a source of organic carbon. Effluent SS

concentrations were low and remained steady throughout the study further indicating low excess solid production.

4.2.

Design implications for HFBR systems

On three typical domestic residential sites in Ireland, septic tank effluent (STE) concentrations between 60 and 85 mg N/l were measured with flows averaging 102 l/person day (equivalent to between 6.1 and 8.7 g TN/person day) (Gill et al., 2007). Based on the results of this study, a 60-sheet HFBR system with a plan surface area of about 0.5 m2/person, may achieve full nitrification and 62% TN removal, when dosed with a similar STE at a TSPA loading rate of 14 g TN/m2.day. In situations where carbon removal only is required a 20-sheet HFBR system without a step-feed could ensure organic carbon removal. For carbon removal and nitrification at the proposed loading rate it is envisaged that a 40e45 sheet unit without a step-feed would suffice. The kinetic data indicate that with typical nitrogen concentrations encountered in wastewaters from domestic dwellings or small communities, HFBR nitrification rates may not be overly limited by ammonium concentrations. Potential removal rates were similar for ammonium concentrations above 1e5 mg NH4-N/l and thus the measured potential rates in the experimental unit could provide a design basis for nitrification in pilot scale units. The experimental unit was operated at 10
as this is typical of wastewater effluents in Ireland. Because nitrification rates in wastewater systems are significantly affected by temperature, wastewater reactor designers should take this into

Table 4 e Performance of the HFBR in comparison to mechanical and filtration systems. TN removal Study

System

HFBRa Welander and Mattiasson, 2003 a,b Janoud et al., 2003 b,c Biesterfield et al., 2003 c

HRT (h)

Suspended biofilm reactor Moving bed biofilm reactor Trickling filter biofilm

Initial NO3-N concentration (mg/l)

<0.5 0.9e2.1

12.0 7

0.8 1.1e2.6

4.5e16.2

20

0.4e1

5

16e18

3.09e5.55

BOD5 (COD) Study HFBR Watasuki et al., 1993 d Rodgers et al., 2005 b Bedessem et al., 2005 b Paredes et al., 2007 b

System

RF with forced aeration St SF SF with added carbon layer Constructed wetlands

Denitrification rate (g/m2 day)

TN

Loading rate (l/m2.day)

Loading rate (g/m2.day)

418.4 150

141.5 (5.6)

97.4 73

15.7 e

61.7 e

20

22

99

2.4

27

40

8.4

83e93

1.8

67

28

e

e

8.1e14.1

55e80

St SF ¼ stratified sand filter, SF ¼ soil filter, RF ¼ recirculating filter. a At 11
C. b Synthetic wastewater used in these studies. c Municipal primary clarified effluent spiked with sodium nitrate. d Filtration media included gravel, clay loam, zeolites, iron particles and jute.

% Removal

Loading rate (g/m2.day)

% Removal

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

account (Tchobanoglous and Crites, 2003). For the HFBR, the measured potential nitrification rates indicate that the nitrification zone may require 3 times the area to achieve a similar nitrification performance at 7
C than at 16
C. However, the in situ temperature/activity correlation may be less dramatic, considering long-term acclimation effects on the biomass to different temperature regimes.

4.3. Comparison of reactor performance to other technologies Table 4 compares total nitrogen removal from filtration systems such as sand, peat and soil filters with those observed in the HFBR step-feed system. A comparison of nitrogen removal rates of the HFBR between Sheets 38 and 40 with other mechanical treatment systems is also presented in Table 4. To achieve overall higher rates of nitrogen removal the addition of more sheets and additional step-feed points are the only changes necessary. An advantage of the HFBR system when compared to sand or soil filters is the high rate of nitrogen removal that can be achieved with the HFBR without the addition of an external carbon source. In comparison to mechanical treatment systems the HFBR requires only a single pump with a timer to operate. This leads to low running costs and can reduce maintenance issues. The system can easily be adapted to households with significantly higher or lower flows simply by increasing or reducing the number of sheets or the plan surface area. The studied HFBR performed well in both removing carbon and nitrogen in comparison to these systems while employing simpler technologies that result in lower running costs compared to other mechanical systems. Noise emissions are at a minimum due to the lack of aeration devices.

5.

Conclusions

During its 418-day period of steady state operation, at a top surface plan loading rate of 418.4 l/m2.day the HFBR performed well in removing BOD5, COD, NH4-N and TN. The main findings of the study include  BOD5 and filtered COD removals averaged 97.4 and 85.7%, respectively during the study.  NH4-N removals averaged 95.7% with average total N removals of 61.7% observed.  Potential nitrification rates varied linearly with experimental temperature. FISH analyses and kinetic experiments showed increased proportions of AOB and NOB in the prokaryotic community may not lead to higher potential nitrification rates.  A unit with 60 sheets and a surface plan area of 0.5 m2/ person, based on typical STE, and operated at 11
C, could achieve almost complete BOD5 removal; full nitrification and 62% total nitrogen removal.  Kinetic data on potential nitrification rates and measurements on the effect of temperature on potential nitrification rates may be used to help achieve economic design while meeting required discharge limits.

3827

 The system performed well in comparison to other technologies in removing organic carbon and nitrogen. Future development of wastewater treatment technologies for single houses may require systems with reduced mechanical and electrical components that are reliable, energy efficient and low maintenance, and can meet discharge limits. The simple, robust, flexible technology may provide a low maintenance, cost effective alternative for the removal of carbon and nitrogen from wastewater from single houses.

references

Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., Stahl, D.A., 1990. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Applied Environmental Microbiology 56, 1919e1925. APHA, 1995. Standard Methods for the Examination of Water and Wastewater, 19th ed., American Public Health Association, Washington, DC. Barth, E.F., 1968. Chemicalebiological control of nitrogen and phosphorus in wastewater effluent. Journal of the Water Pollution Control Federation 40, 2040e2054. Bedessem, M.E., Edgar, T.V., Roll, R., 2005. Nitrogen removal in laboratory model leachfields with organic-rich layers. Journal of Environmental Quality 24, 936e942. Biesterfield, S., Farmer, D., Figueroa, L., Parker, D., Russell, P., 2003. Quantification of denitrification potential in carbonaceous trickling filters. Water Research 37 (16), 4011e4017. Copp, J.B., Murphy, K.L., 1995. Estimation of the active nitrifying biomass in activated sludge. Water Research 29 (8), 1855e1862. Corley, M., Rodgers, M., Mulqueen, J., Clifford, E., 2006. The performance of fibrous peat biofilters in treating domestic strength wastewater. Journal of Environmental Science and Health: Part A 41 (5), 811e824. Council Directive of 16 June 1975 concerning quality required of surface water intended for the abstraction of drinking water (75/440/EEC). Official Journal L 194. Council Directive (2000/60/EC) of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in the field of water policy. Official Journal L 327/1. Council Directive (2006/118/EC) of the European Parliament and of the Council of 12 December 2006 on the protection. Crites, R., Tchobanoglous, G., 1998. Small and Decentralized Wastewater Treatment Systems. Boston/London, WCB/ McGraw-Hill. Daims, H., Bru¨hl, A., Amann, R., Schleifer, K.H., Wagner, M., 1999. The domain-specific probe EUB338 is insufficient for the detection of all bacteria: development and evaluation of a more comprehensive probe set. Systematic and Applied Microbiology 22, 434e444. EPA, 2007. Wastewater Treatment Systems for Single Houses: (PE < 10). Consultation Draft. EPA, Wexford, Ireland. Ferguson, G., Dakers, A., Gunn, I., 2003. Sustainable Wastewater Management: A Handbook for Smaller Communities. New Zealand Ministry for the Environment, Weillington, New Zealand. ISBN: 0-478-24095-5. Gill, L.W., O’Su´lleabha´in, C., Misstear, B.D.R., Johnston, P.J., 2007. The treatment performance of different subsoils in Ireland

3828

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 8 1 9 e3 8 2 8

receiving on-site wastewater effluent. Journal of Environmental Quality 36, 1843e1855. Gray, N.F., 1999. Water Technology: An Introduction for Environmental Scientists and Engineers. John Wiley & Sons, New York, USA. Helmer, C., Kunst, S., 1998. Simultaneous nitrification/ denitrification in an aerobic biofilm system. Water Science and Technology 37, 183e187. Henze, M., Harremoes, P., La Cour Jansen, J., Arvin, E., 1997. Wastewater Treatment: Biological and Chemical Processes, third ed. Springer-Verlag. Hovanec, T.A., Taylor, L.T., Blakis, A., Delong, E.F., 1998. Nitrospira-like bacteria associated with nitrite oxidation in freshwater aquaria. Applied Environmental Microbiology 64, 258e264. Jonoud, S., Vosoughi, M., Khalili Daylami, N., 2003. Study on nitrification and denitrification of high nitrogen and COD load wastewater in moving bed biofilm reactor. Iranian Journal of Biotechnology 1 (2), 115e120. Johnson, K.R., Atwater, J.W., 1988. Role of saturated and unsaturated zones in soil disposal of septic tank effluent. Canadian Journal of Civil Engineering 15 (4), 709e716. Lowe, K.S., Van Cuyk, S.M., Siegrist, R.L., Drewes, J.E., 2008. Field evaluation of the performance of engineered on-site wastewater treatment units. Journal of Hydrologic Engineering 13 (8), 735e743. Lydmark, P., Lind, M., So¨rensson, F., Hermansson, M., 2006. Veritcal distribution of nitrifying populations in bacterial biofilms from a full-scale nitrifying trickling filter. Environmental Microbiology 8 (11), 2036e2049. Mattila, K., Zaitsev, G., Langwaldt, J., 2007. Biological Removal of Nutrients from Mine Waters. Finnish Forest Research Unit. ISBN: 978-951-40-2060-5. Mobarry, B.K., Wagner, M., Urbain, V., Rittmann, B.E., Stahl, D.A., 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. Applied Environmental Microbiology 62, 2156e2162. Nielsen, M., Revsbech, N.P., Larsen, L.H., Lynggaard, A., 2002. Online determination of nitrite in wastewater treatment by use of a biosensor. Water Science and Technology 45, 69e76. Odegaard, H., Rusten, B., 1980. Nitrogen Removal in Rotating Biological Contactors without the Use of an External Carbon Source. Proceedings of the First National Symposium/Workshop on Rotating Biological Contactor Technology vol. 2, 1301e1317.

Paredes, D., Kuschk, P., Stange, F., Mu¨ller, R.A., Ko¨ser, H., 2007. Model experiments on improving nitrogen removal in laboratory scale subsurface constructed wetlands by enhancing the anaerobic ammonia oxidation. Water Science and Technology 56 (3), 145e150. Randall, C.W., 1992. Introduction and principles of biological nutrient removal. In: Randall, C.W., Barnard, J.L., Stensel, H.B. (Eds.), Design and Retrofit of Wastewater Treatment Plants for Biological Nutrient Removal. Technomic Publishing Co. Inc., Lancaster, USA. Rodgers, M., Mulqueen, J., Lambe, A., Xiao, L., Prendergast, J., Clifford, E., 2004. A Biofilm Reactor. Patent Application No. S2004/0461. Rodgers, M., Healy, M., Mulqueen, J., 2005. Organic carbon removal and nitrification of high strength wastewaters using stratified sand filters. Water Research 39 (14), 3279e3286. Rodgers, M., Clifford, E., 2009. Horizontal-flow biofilm reactors for the removal of carbon and nitrogen from domestic strength wastewaters. Water Environment Research 81 (4), 339e397. Tchobanoglous, G., Crites, R., 2003. In: Tchobanoglous, G., Crites, R. (Eds.), Wastewater Engineering (Treatment Disposal Reuse), fourth ed. Metcalf & Eddy, Inc, McGraw-Hill, New York, USA. USEPA, 1980. Design Manual e Onsite Wastewater Treatment and Disposal Systems. US EPA, Washington, USA. 625/1-80012. Von Mu¨nch, E., Lant, P., Keller, J., 1996. Simultaneous nitrification and denitrification in bench-scale sequencing batch reactors. Water Research 30 (2), 277e284. Wagner, M., Rath, G., Koops, H.P., Flood, J., Amann, R., 1996. In situ analysis of nitrifying bacteria in sewage treatment plants. Water Science and Technology 34, 237e244. Watasuki, T., Esumi, H., Omura, S., 1993. High performance N and P-removal on-site domestic wastewater treatment system by multi-soil-layering method. Water Science and Technology 21, 31e40. Welander, U., Mattiasson, B., 2003. Denitrification at low temperatures using a suspended carrier biofilm process. Water Research 37 (10), 2394e2398. Witzig, R., Manz, W., Rosenberger, S., Kru¨ger, U., Kraume, M., Szewzyk, U., 2002. Microbiological aspects of a bioreactor with submerged membranes for aerobic treatment of municipal wastewater. Water Research 36, 394e402.