Journal of Environmental Management 160 (2015) 128e138

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Research article

Intensification of ammonia removal from waste water in biologically active zeolitic ion exchange columns Azel Almutairi 1, Laurence R. Weatherley* Department of Chemical and Petroleum Engineering, The University of Kansas, Lawrence, KS 66045, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 December 2014 Received in revised form 21 May 2015 Accepted 25 May 2015 Available online xxx

The use of nitrification filters for the removal of ammonium ion from waste-water is an established technology deployed extensively in municipal water treatment, in industrial water treatment and in applications such as fish farming. The process involves the development of immobilized bacterial films on a solid packing support, which is designed to provide a suitable host for the film, and allow supply of oxygen to promote aerobic action. Removal of ammonia and nitrite is increasingly necessary to meet drinking water and discharge standards being applied in the US, Europe and other places. Ion-exchange techniques are also effective for removal of ammonia (as the ammonium ion) from waste water and have the advantage of fast start-up times compared to biological filtration which in some cases may take several weeks to be fully operational. Here we explore the performance of ion exchange columns in which nitrifying bacteria are cultivated, with the goal of a “combined” process involving simultaneous ion-exchange and nitrification, intensified by in-situ aeration with a novel membrane module. There were three experimental goals. Firstly, ion exchange zeolites were characterized and prepared for comparative column breakthrough studies for ammonia removal. Secondly effective in-situ aeration for promotion of nitrifying bacterial growth was studied using a number of different membranes including polyethersulfone (PES), polypropylene (PP), nylon, and polytetra-fluoroethylene (PTFE). Thirdly the breakthrough performance of ion exchange columns filled with zeolite in the presence of aeration and in the presence of nitrifying bacteria was determined to establish the influence of biomass, and aeration upon breakthrough during ammonium ion uptake. The methodology adopted included screening of two types of the naturally occuring zeolite clinoptilolite for effective ammonia removal in continuous ion-exchange columns. Next, the performance of fixed beds of clinoptilolite in the presence of nitrifying bacteria is compared to that in columns in which only ion exchange is occurring. The aeration performance of each of the chosen membranes was compared experimentally using a newly developed membrane support module which is also described. Comparison of ammonia removal in columns equipped with in-situ aeration using each membrane was undertaken and the breakthrough characteristics determined. The results showed that ammonia removal in the presence of the nitrifiers was significantly intensified. Column operation with membrane aeration showed further enhancement of ammonia removal. The greatest enhancement was observed in the case of the polyethersulfone membrane (PES). It is concluded that combined nitrification and ion-exchange is significantly intensified in packed columns by in-situ aeration using a novel membrane module. There is significant potential for extending the ion-exchange cycle time and thus potential cost reduction. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Ion-exchange Nitrification Water treatment Membrane aeration Intensification

1. Introduction

* Corresponding author. E-mail addresses: [email protected] (A. Almutairi), [email protected] (L.R. Weatherley). 1 Present address: Environmental Technology Management Department, College for Women, University of Kuwait, Kuwait. http://dx.doi.org/10.1016/j.jenvman.2015.05.033 0301-4797/© 2015 Elsevier Ltd. All rights reserved.

Removal of ammonia from waste-waters by ion exchange has received significant attention over the last several years. Ammonia is a toxic contaminant which is ubiquitous in domestic and municipal wastewater representing a significant component of waterborne pollution. Ammonia is also one of the most commonly occurring nitrogenous pollutant in wastewater and exists in two

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

forms, the more abundant of which is the ionized form, the ammonium ion (NHþ 4 ), and the less abundant of which is the unionized form, the ammonia molecule (NH3) depending on the pH. Free unionized ammonia is considerably harmful and more toxic than the ionic form. The latter ammonium ion form dominates typically at pH < 8 and is readily removed by ion exchange. Biological nitrification is one of the most widely used and longstanding technologies in wastewater treatment processes for removal of ammonia (Adams and Eckenfelder, 1977, Metcalf et al., 2004, Bernier et al. (2014), Lee et al., 2014). Nitrification may be defined as the microbiological oxidation of ammonia nitrogen to nitrite (NO 2 ) by ammonia oxidizing bacteria (AOB) and nitrite to nitrate (NO 3 ) by nitrite oxidizing bacteria (NOB) in the presence of oxygen. Two bacterial groups participate in each step of the oxidation process, namely Nitrosomonas and Nitrobacter (Burrel et al., 1998). Based on reaction stoichiometry a significant dissolved oxygen concentration is necessary for the nitrification reactions to proceed. A figure of 4.27 g of O2 per 1.0 g of NHþ 4 is quoted by Adams and Eckenfelder (1977). Ion exchange has also been studied for ammonia removal (as NHþ 4 ion) using naturally occurring zeolitic minerals such as clinoptilolite. Removal of ammonium ion by ion exchange from municipal wastewater and other polluted waters for example those arising in aquaculture and from landfill leachates has also been investigated (Guo, X. et al., 2008; Vassileva and Voikova, 2009; Mazeikiene et al., 2010; Beler-Baykal and Allar, 2008; Gunay, 2007; Hedstrom and Amofah, 2008; Karadag et al., 2008; Mazeikiene et al., 2008; Siljeg et al., 2010). Most of these ion exchange studies are based on a conventional approach involving packed beds of zeolite particles through which the wastewater is trickled until a specified ammonium ion breakthrough concentration is reached. The ion exchange columns may be chemically regenerated to restore capacity and thus reused for further cycles of ammonia removal. The regeneration part of the cycle involves a significant use of chemical regenerant and rinse water together with at least one duplicate column to be brought into service in order to maintain continuous operation. Ion exchange provides a technique which is responsive to shock loading and can operate at range of temperatures. Many natural ion exchange minerals are crystalline, hydrated, aluminosilicates of alkali and alkaline earth cations or anions, having a three dimensional structures formed by AlO4 and SiO4 that are connected by sharing oxygen atoms and water molecules. Clinoptilolite, mordenite, analcite, chabazite, erionite, heulandite, and laumonite are the most common types of zeolites (Bish and Ming, 2001; Mishra and Jain, 2011, Montalvo et al., 2012. Tsitsishvili et al., 1992). A nitrification process in combination with ion exchange offers some improvements, exploiting the advantages of both techniques (Sims and Little, 1973; Beler-Baykal et al., 1994; Smith, 2011) and showing some advantageous synergy. Such a combined system has the potential to be responsive to shock loading and to yield significantly longer cycle times between regenerations, and thus offer a reduction in the toxicity of the effluent arising because of the conversion of ammonia to nitrate. Earlier studies of combined nitrification and ion exchange were focused on treatment of wastewater with an already established population of nitrifying bacteria. Sims and Little (1973) studied a combined process based on the addition of ion exchanger into a suspended growth bioreactor containing active colonies of nitrifying bacteria. McVeigh and Weatherley (1999) also studied the introduction of nitrifying bacteria into packed beds of ion exchanger. Beler-Baykal et al. (1994) developed a combined ionexchange and nitrification column demonstrating the ability to

129

favorably enhance the response to shock loading of ammonia. Competition of other ions present in the system for the ion exchange sites was observed which reduced the overall ion exchange removal capacity for ammonium ion. Another approach involving combination of ion exchange with biological water treatment is described by Ahn et al. (2002) in which zeolitic ion exchange was used to reduce ammonia concentration as a pre-treatment prior to biological treatment. This lowered the risk of substrate toxicity during biological treatment and the risk of oxygen depletion. Oxygen depletion is an issue which can impact negatively on nitrifying filters and on combined systems (Lahav and Green, 1998; Stenstrom and Poduska (1980); Wang and Yang (2004). To overcome this limitation, mitigation of oxygen depletion using incolumn membrane aeration has been suggested. Huang et al. (2015) proposed a novel fully passive permeable reactive barrier (PRB) with oxygen-releasing compound (ORC) in the presence of clinoptilolite for the removal of ammonium-nitrogen from groundwater. This technique significantly enhanced the role of nitrification in each cycle of ammonia removal in ion exchange columns. Earlier work by Kuhlman (1987) showed that membrane based aerobic growth enhancement could be successfully performed with bubble free aeration using a silicone tubing oxygenation system in an aerobic animal cell culture bioreactor. More recent preliminary studies demonstrated scope for this approach being effectively applied to biologically active ion-exchange columns (Miladinovic and Weatherley (2008)). A tubular silicone rubber membrane located along the axis of the bed was maintained at a slight positive air pressure on the tube side to supply air to the interior of the packed bed. In both sets of experiments, longer breakthrough volumes were observed although in the case of clinoptilolite there was evidence of zeolite attrition leading to the blockage of the smaller membrane pores thus reducing aeration efficiency. The goal of the work described here was to seek better membranes for aeration, to investigate an alternative design for the membrane modules, and to determine the ammonia removal performance of clinoptilolite in combined nitrification and ion exchange columns with integrated membrane aeration. 2. Experimental There were three experimental goals. Firstly, the ion exchange zeolites were characterized and prepared for comparative column breakthrough studies for ammonia removal. Secondly, the membrane materials were evaluated to determine their relative performance for water aeration. Thirdly the breakthrough performance of ion exchange columns filled with zeolite in the presence of aeration and in the presence of nitrifying bacteria was determined to establish the influence of biomass, and aeration upon breakthrough during ammonium ion uptake. 3. The zeolites The ion exchange materials used in this research comprised two different types of clinoptilolite provided by two companies: KMI Zeolites Inc. and Boulder Innovative Technologies Inc. (BIT). Both types of clinoptilolite are referred to by abbreviated names based on the provider companies as KMI and BIT respectively. The KMI clinoptilolite was sourced from the KMI's deposits in Sandy Valley, Nevada, while BIT clinoptilolite was sourced from BIT claim deposits in Colorado. Both types have a mean particle diameter in the range of 0.420 mme1.41 mm as delivered from the factory. Sieving was carried out using a sieve shaker, (RX-29 W.S. Tyler) to ensure batches of both clinoptilolite types in the particle size of

130

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

14  40 mesh (1.41 mme0.420 mm). To remove any fines and impurities, the clinoptilolite samples were washed and soaked in deionized water overnight. The washed clinoptilolite samples were soaked in a 1 M sodium chloride solution for 24 h with regular hand agitation to ensure complete conditioning into the sodium form. The sodium chloride solution was renewed on a daily basis for a period of 6 days. Then, the clinoptilolite samples were washed several times and soaked in deionized water overnight to remove all traces of free sodium chloride. This was followed by filtering and drying at 25  C. The conductivity of the water used to wash the samples was determined to ensure no residual of sodium chloride remained. 4. Membrane evaluation studies The objective of the membrane evaluation was to compare the gas transport characteristics of the membranes over a range of pressures, to establish their relative effectiveness for increasing the oxygen concentration of de-aerated water. The evaluation was conducted prior to installation into the ion exchange columns. Five different membrane materials were evaluated: polytetrafluorethylene (PTFE), polypropylene (PP), polyethersulfone (PES), nylon, and silicone. The silicone membrane differs from the other polymeric in that the porosity is very low thus influencing the mechanism by which the gas phase moves through the membrane. The apparatus used in aeration studies is shown in Fig. 1. The membrane modules comprised an array of acrylic tubes, see Fig. 2(a). Six tubes were used, each having a length of 15 cm and an outside diameter of 1.27 cm. In order to let air pass through the tubes and reach the surface of the membranes, a total of 42 equally spaced slots were made in the outer surface of the module, each with an area of 0.461 cm2. Thus, the total free area of the 42 slots amounted to 19.357 cm2. The membrane module is shown in Fig. 2. The membranes specifications are shown in Table 1. The evaluation of membrane aeration performance was conducted in a series of individual batch experiments using the module fitted with each membrane. The procedure was as flows. 2.5 L of filtered tap water were added to the 3 L glass vessel (Fig. 1) to which 167.5 mg sodium sulfite, Na2SO3, was added to achieve deoxygenation according to following equation: 2Na2 SO3 þ O2 ¼> 2Na2 SO4 The dissolved oxygen concentration was followed continuously until a zero reading was achieved. At this point the air cylinder was

Fig. 1. Aeration studies apparatus.

used to pressurize the membrane assembly to allow air to diffuse through the membrane wall at the chosen pressure to oxygenate the water. Dissolved oxygen concentration readings were then taken every 10 min until steady state was attained. Two types of ion-exchange columns were used to evaluate ammonia removal. The first column, Figs. 2(b) and 3a, was constructed to accommodate the silicon membrane module. The second column was designed on the basis of the first column, and was designed to operate with the porous membrane modules, Fig. 2(a) and Fig. 3. The latter design and a cross-sectional view are shown in Fig. 3b. The designs of these columns was informed by those developed in earlier studies (McVeigh and Weatherley (1999); Miladinovic and Weatherley (2008); Jorgensen and Weatherley (2003)). The columns were designed to enhance the aeration within the clinoptilolite bed so that the nitrification process can proceed without being limited by reduced oxygen concentration. The silicon membrane's permeability was 7961  1010 cm3 s1 cm2 cm Hg1 [Manufacturer]. The internal and external silicone tube diameters were 1.8 and 2.1 mm respectively, with the tube being wound helically around a 35 cm central module. Two tubes were assembled in the silicon membrane column with modules which were clipped on both ends. One end of the silicone tube was sealed with a metal clip to force air to diffuse through the membrane as shown in Fig. 3. The outer section of the membrane module was filled with the Clinoptilolite. Air was supplied under pressure into the inner section of the module. The columns were installed into a lab scale test loop (Fig. 4) with a maximum of four columns operated simultaneously. Columns were operated in separate experiments in downflow and up-flow mode respectively. In down-flow mode, see Fig. 4, columns were operated on a constant head principle. The flow rate was maintained constant by keeping the head in the feed tanks constant. Pre-calibrated rotameters were used to control and monitor the inlet flow to each column. In up-flow mode, Fig. 4, precalibrated peristaltic pumps, (Watson-Marlow), were used to maintain the flowrates.

5. Procedure At the commencement of the experimental runs, each column was filled with pre-conditioned clinoptilolite to a height of either 27 cm or 32 cm. The term “Bed Volume” (BV) in this study refers to the volume of the empty column, i.e., without packed clinoptilolite. Table 2 summarizes the bed volume values used. To prevent air bubbles being trapped within the column, dry pre-conditioned clinoptilolite was added to each column, which were isolated at inlet and outlet. Each column of zeolite was washed with deionized water several times after which each column was reconnected to the inlet and outlet. The water used in each experiment was supply.water filtered in a commercial PUR 3-stage faucet filter. The synthetic wastewater was made up and stored in a 120 L reservoir tank. The ammonium ion concentration and pH were measured before use. Solution was lifted up to the feed tank (90 Lt capacity) using a VESTIL hand winch lift truck and was maintained at 30  C using a bath circulator (Isotemp 3016p, Fisher Scientific). The solution was fed from the storage tank to the column and into the clinoptilolite bed, with the eluent solution collected in graduated receiving tanks which were calibrated at 4 L volume intervals. After a specified time interval, the effluent solution was sampled and analyzed for ammonium ion concentration, oxygen concentration, pH, and nitrite concentration. Experiments on biologically active clinoptilolite followed the same procedure but with aeration present via the membrane module.

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

131

Fig. 2. (a). Porous membrane module. (b). Silicone membrane module.

Table 1 Membrane specifications. Parameter

Pore size (mm) Thickness (mm) Length (mm) Maximum pressure (psi) Air Flow Shape Type Provider (company)

Membrane type Polytetrafluorethylene (PTFE), (or Teflon)

Polypropylene, (PP)

Polyethersulfone, (PES), (or Supor-R200)

Nylon

0.2 175 300 12e23 e Flat sheet Porous STERLITECH

0.1 75e110 200 30 e Flat sheet Porous STERLITECH

0.2 119.4e215.9 250 20e57 LPM/3.7 cm2 @ 13.5 psi Flat sheet Porous PALL

0.1

6. Bacteria immobilization One of the challenges facing the combination of ion exchange and nitrification as a treatment method is establishment of biofilm on the ion exchange material without significantly reducing adsorption capacity. Bacterial immobilization was by means of an adsorption method which was developed in earlier work by McVeigh and Weatherley (1999); Miladinovic and Weatherley (2008). 50 ml of bacteria rich culture solution prepared as described below, were added to 2 L of ammonium ion solution of low concentration (0.5 mg N  NHþ 4 =l), and the combined active solution was manually introduced to the packed bed column in the presence of aeration. After approximately 4 h the effluent was recycled to the column. This procedure was repeated over a 24 h period. The biomass concentration in the initial feed water was compared with that at the end of the uptake cycle in order to determine the net uptake of bacteria. Since Nitrosomonas europea is very sensitive to light, the biologically active column was coated with aluminum foil during each experiment. For the column fitted with the silicon tube membrane, air was supplied at a pressure of 25 psig. For the columns fitted with the porous membranes, the air supply pressure was maintained at constant pressure in the range 1e8 psig.

Silicon

200

Flat sheet Porous Osmonics Inc.

Tube Dense Cole-Parmer

The bacterial cultures (ATCC 19718) were obtained from the American Type Culture Collection (ATCC). The culture medium was formulated as follows:

-

Step 1: 900 ml DI water were added to a 2 L Erlenmeyer flask and the following materials were added in sequence: 3.3 g (NH4)2SO4 (50 mM) 0.41 g KH2PO4 0.75 ml 1 M MgSO4 0.2 ml 1 M CaCl2 0.33 ml 30 mM FeSO4/50 mM EDTA 0.01 ml 50 mM CuSO4 The flask was then sterilized using an autoclave.

Step 2: 400 ml of DI water were added into a 500 ml baker after which the following materials were added: - 27.22 g KH2PO4 - 2.4 g NaH2PO4 The pH was adjusted to pH8.0 with using 10 N NaOH solution. The final volume was made up to 500 ml with deionized water in a 500 mL bottle and sterilized.

132

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

a

b

Fig. 3. a. Cross sectional view of the silicon membrane module column. b. Cross-sectional view of porous membrane module column.

Step 3: 500 ml of 5% (w/v) Na2CO3 was prepared and sterilized. To prepare a 1 L of inoculum, the following steps were made: - 100 ml of solution prepared in Step 2 were added to the flask prepared in Step 1. - 8 ml of the solution made in Step 3 were added to the flask prepared in Step 1. - 10 ml of 3-day old culture were added to the flask prepared in Step 1, and incubated on a rotary shaker (@150 rpm) at 30  C.

After 7 days of incubation, the culture was centrifuged in a SORVAL Evolution centrifuge at 5000 RCF (relative centrifugal force) for 20 min. The biomass was added to 50 ml DI water, referred to as “bacteria rich” solution. 7. Analytical methods 7.1. Ammonia Ammonium ion concentration measurements were carried out

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

133

Fig. 4. Flow arrangement for continuous column studies.

7.4. pH

Table 2 Column dimensions and volumes. Column design

Bed height, h (cm)

BV

Silicon membrane column

32 27 27

1.605 L 1.355 L 1.455 L

Porous membrane column

Determination of pH was carried out by using either an Orion 4 Star pH meter along with Gel-Filled pH electrode or using an OAKTON pH tester 30. Both meters have a built-in temperature probe. Both meters were calibrated using Thermo Scientific buffer solutions pH ¼ 4.01, pH ¼ 7.00, and pH ¼ 10.01.

using a Thermo Scientific Orion ISE meter 4-Star along with a Thermo Scientific Orion ammonium ion selective electrode (No. 9512BNWP). Prior to measurement, pH buffer (ISA Orion 951211) was added to the sample in a ratio 50:1. To prevent any contamination, the electrode was washed before and after measuring the sample with DI water and dried with a lint-free tissue. All ammonium ion solutions, including calibration solutions, were made up from ammonium chloride (NH4Cl).

Conductivity was measured to check for the presence of sodium hydroxide in the solution. An Orion 4 Star conductivity meter along with an Orion conductivity electrode (DuraProbe conductivity Cells 013005MD) were used. These were calibrated using 100 mS/cm and 1413 mS/cm Orion conductivity standards.

7.2. Nitrite

7.6. Protein

Nitrite (NO 2 ) was accurately determined using a UVevisible spectrophotometer following the standard AWWA procedure. In order to establish a photometric calibration curve, a color reagent was prepared by adding 100 ml of 85% phosphoric acid (H3PO4) solution and 10 g sulfanilamide (C6H8N2S) to 800 ml deionized water. After the sulfanilamide was dissolved completely, 1 g of NED dihydrochloride was added, and the solution diluted to 1 L. Standard nitrite solution was prepared by adding 1.232 g sodium nitrite (NaNO2) to 1 L deionized water, and dilution to the required concentrations. After removing any suspended solids by filtering through a 0.45-mm diameter pore membrane filter, the pH was adjusted to the range of 5e9 with 1N hydrochloric acid (HCl) or 1N ammonium hydroxide (NH4OH) as needed. To develop the color of the sample or standard, 2 ml color reagent were added to a 50 ml portion of sample. For different known concentrations, the absorbance was measured at 543 nm to achieve an accurate standard.

Protein measurements were carried out using either a UVevisible spectrophotometer or a NanoDrop 1000 (Thermo Scientific). The samples were measured at OD600 nm (OD ¼ optical density).

7.3. Dissolved oxygen Dissolved oxygen concentrations were determined using an oxygen probe/meter, DO200, (Yellow Spring Instruments), with automatic temperature compensation.

7.5. Conductivity

7.7. Temperature Temperature thermometer.

was

recorded

using

a

Traceable®

digital

8. Results and discussion The first set of experimental results compare the relative effectiveness of each of the porous membrane types for increasing the oxygen concentration in de-aerated water. Figs. 5 and 6 show the comparisons for membrane side air pressure of 1.0 and 4.0 psig respectively, indicating the superior performance of the PTFE membrane compared with the polypropylene, nylon, and polyethersulfone membranes. The PTFE showed the best aeration performance at the pressure levels used. At 1.0 and 4.0 psi, the polypropylene exhibited greater aeration rates compared with the polyethersulfone and with the Nylon. Nylon gave the worst performance, compared to the other

134

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138 Table 3 Breakthrough BV and uptake capacities based on Fig. 7. KMI

5% Breakthrough point (BV) Uptake column capacity (meq/g)

Fig. 5. Porous membrane oxygenation of de-aerated water e comparison of polyethersulfone, PTFE, nylon and polypropylene membranes (Temperature 20C, air feed at transmembrane pressure of 1.0 psi).

Fig. 6. Porous membrane oxygenation of de-aerated water e comparison of polyethersulfone, PTFE, nylon and polypropylene membranes (Temperature 20C, air feed at transmembrane pressure of 4.0 psi).

porous membranes, taking more than 60 min to reach 7.5 mg DO/l. Despite having a similar pore size as the PES membrane, aeration with PTFE reached 100% saturation within 10 min while with PES it took 60 min to reach 90% saturation at 4.0 psi. Fig. 7 compares the breakthrough performance for the ion-

Fig. 7. The uptake values for column packed with KMI clinoptilolite fitted with the silicon aeration tube. Downflow regime, initial ammonia concentration: 20 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate: 22 L/hr, bed height: 32 cm, transmembrane pressure ¼ 25.0 psi.

No aeration

Aeration

240 0.38

260 0.41

exchange column fitted with the silicon-tube membrane (nonporous) in the presence and in the absence of aeration. This comparison is an important “control” experiment which examined the possibility that aeration to the column per se may have some effect upon the breakthrough performance. It is possible that the introduction of air via the membrane may have some minor influence on the internal hydrodynamics affecting mass transfer and dispersion. There is some evidence for this when the breakthrough data are integrated to yield the bed volumes processed to breakthrough as shown in Table 3 though the difference in breakthrough bed volumes treated is less than 10%. It was observed during the experiment that large air bubbles were not formed which may have an impact on the mass transfer within the column. However, air is still introduced to the aerated column and it might enhance the ion exchange process. On the other hand if there is build up of gas hold-up in the column this may inhibit contact between the water and zeolite. Also the effective residence time for the liquid flow through the column may be reduced in such cases. The previous study (Miladinovic and Weatherley, 2008) showed negligible impact of aeration on the packed column performance which is consistent with the observations here. Fig. 8 shows the column breakthrough comparing the biologically inoculated column and non-active column fitted with the silicon tube aeration membrane. In this case no improvement was observed in the inoculated column when compared to the non-active column. In fact, the uptake capacity of the active column is lower than that of the non active column, and the BV breakthrough was reached at 115BV (Table 4). A possible explanation for is that the silicon tube membrane was not introducing enough oxygen to the column and therefore the biological activity is not only reduced but the membrane itself impedes contact between the liquid phase and the zeolite. As shown in Fig. 9, the non-active porous membrane column showed breakthrough after 110 BV of synthetic wastewater

Fig. 8. Column breakthrough in the biologically active and non active silicon tube column. Downflow, initial ammonia concentration: 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 22 L/hr, bed height: 27 cm, trans-membrane pressure ¼ 25 psi.

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

135

Table 5 The breakthrough BV and the uptake capacities based on Fig. 9. KMI, PES membrane

5% Breakthrough point (BV) Uptake column capacity (meq/g)

No bacteria

Bacteria

110 0.38

165 0.56

Table 6 The breakthrough BV and the uptake capacities based on Fig. 10. KMI, PTFE membrane

5% Breakthrough point (BV) Uptake column capacity (meq/g)

No bacteria

Bacteria

105 0.35

140 0.47

Fig. 9. Biologically active and non active PES membrane column. Upflow, initial ammonia concentration: 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, trans-membrane pressure ¼ 1.5 psi.

Fig. 11. Biologically active and non active PP membrane column. Upflow, initial ammonia concentration: 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, transmembrane pressure ¼ 1.5 psi, flow rate ¼ 4 BV/h.

(Table 5) had passed through the bed, while the biologically active porous column treated a further 55BV before breakthrough. This is a significant improvement, and the presence of bacteria in the column significantly enhanced the columns' performance. The effective breakthrough uptake capacity was increased from 0.38 to 0.56 meq/g. Miladinovic and Weatherley (2008) observed a similar increase in the breakthrough uptake capacity from 0.15 to 0.22 meq/g when the bacteria were introduced to the column they used. In Fig. 9, both curves behaved differently, and they had different slopes, where the active column shows a sharper breakthrough curve. The PTFE membrane column inoculated with bacteria performed better than the one without bacteria, see Fig. 10 and Table 6.

Fig. 11 shows the breakthrough curves for both the biologically activated and non-activated polypropylene membrane column. The breakthrough BV and the uptake capacities for these columns are presented in Table 7 and a significant improvement in ammonia removal is observed the biologically active column. The performance of the active and non-active Nylon membrane columns is presented Fig. 12, and the breakthrough BV and the uptake capacities based on Fig. 12 are presented in Table 8. Unlike the other columns fitted with porous membranes, the column fitted with non activated nylon membrane column performed better than the activated column. This might be on account of the relatively poor performance of the nylon membrane as described in the earlier evaluation of aeration performance, Fig. 6. A comparison of ammonia breakthrough in the columns fitted with the porous membranes columns with and without biologically active material present is shown in Fig. 14. Figs. 13 and 14 shows the performance of the porous membranes columns used in this work. Table 10 shows the calculated breakthrough capacities and breakthrough BVs.

Table 4 The breakthrough BV and the uptake capacities based on Fig. 8.

Table 7 The breakthrough BV and the uptake capacities based on Fig. 11.

Fig. 10. Biologically active and non active PTFE membrane column. Upflow, initial ammonia concentration: 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, transmembrane pressure ¼ 1.5 psi.

KMI

5% Breakthrough point (BV) Uptake column capacity (meq/g)

KMI, PP membrane

No bacteria

Bacteria

140 0.45

115 0.37

5% Breakthrough point (BV) Uptake column capacity (meq/g)

No bacteria

Bacteria

130 0.43

200 0.67

136

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

Fig. 12. Biologically active and non active Nylon membrane column. Upflow, initial ammonia concentration: 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, trans-membrane pressure ¼ 1.5 psi.

Fig. 14. Ammonia breakthrough performance of biologically active columns fitted with porous membranes:40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, trans-membrane pressure ¼ 1.5 psi.

Table 9 The breakthrough BV and the uptake capacities based on Fig. 13. Table 8 The breakthrough BV and the uptake capacities for the nylon membrane equipped column based on Fig. 12. KMI, Nylon membrane

5% Breakthrough point (BV) Uptake column capacity (meq/g)

No bacteria

Bacteria

150 0.51

110 0.37

KMI, No bacteria

5% Breakthrough point (BV) Uptake column capacity (meq/g)

PES

PTFE

PP

Nylon

110 0.38

105 0.35

130 0.43

150 0.51

Table 10 The ammonia breakthrough BV and the uptake capacities for Fig. 14. KMI, Bacteria

5% Breakthrough point (BV) Uptake column capacity (meq/g)

Fig. 13. Biologically free porous membrane columns 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, trans-membrane pressure ¼ 1.5 psi.

The packed beds were set up with the same bed volume and the same particle size so the residence times were similar. Although the breakthrough curves shown in Fig. 13 show slightly different breakthrough points, within the limits of experimental error they show similar breakthrough curve profiles. Considering the results of the non-biologically active column experiments, the columns fitted with the polypropylene and nylon membranes showed the greatest breakthrough capacity. The column fitted with the nylon membrane had the highest breakthrough capacity of 0.51 meq/g (Table 9). Since the permeability of the porous membrane is not the same as shown in the aeration studies, the amount of air existing in each column during the cycle was also different. Thus, the presence of air could affect the column hydrodynamics and thus the

PES

PTFE

PP

Nylon

165 0.56

140 0.47

200 0.67

110 0.37

breakthrough capacity of each column. During the experiments, bubbles were formed and this observation might support the effect of aeration on the columns hydrodynamics. Review of the literature shows that the type of the feed solution has a great impact on the column capacity. Miladinovic and Weatherley (2008) showed significant differences between feed solutions based on natural creek waters compared with deionized water. In one set of experiments using columns in the absence of significant biological activity it was shown that the column capacity could be reduced by as much as 30% on account of competing uptake of other ions alongside NHþ 4. Fig. 14 shows the comparison between the biologically activated porous membrane columns, and the breakthrough BV points are given in Table 10. The feed solution used for the biologically active and non-active porous membrane columns was the same. This was to quantify the impact of presence of nitrifying bacteria on the removal of ammonium ion. The ammonium ion concentration in the inlet feed þ was chosen to be 40 mg N  NHþ 4 =l instead of 20 mg N  NH4 =l, since the nitrification kinetics would be improved due to the higher substrate concentration within the ion exchange column. Initially 50 ml of rich nitrifying bacteria were added to a solution of very low ammonia ion concentration (1e0.5 mg N  NHþ 4 =l). The bacteria concentration of the combined solution was measured at OD600 by spectrophotometer and then it was circulated in the column to loaded the bacteria on the column. Bacteria concentration of the effluent solution was re-measured at OD600. Time was allowed for

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

bacteria to be attached and immobilized. For example, the biomass in the PP membrane column was measured before the bacteria solution was introduced to the column and the absorbance was found at OD600 equal to 0.062 and after the bacteria solution circulated it was found that its absorbance was 0.023 when measured at OD600. The binding efficiency of bacteria was thus estimated at 63%. In all the columns, the attachment efficiency for the bacteria was estimated at greater than 60%. The remaining bacteria were either washed out or were unable to attach to the clinoptilolite. The biologically active column equipped with the polypropylene membrane was shown to have the best breakthrough performance with a breakthrough capacity for ammonium ion of 0.67 meq/g. The column equipped with the polyethylsulfone membrane showed the second-best performance with a breakthrough uptake capacity of 0.56 meq/g. There was a significant difference in the performance of the columns equipped with polypropylene and nylon membranes columns performance in spite of very similar pore sizes, see Table 1. There are some differences in the shapes of the breakthrough curves. The breakthrough curve for the column equipped with the polypropylene membrane exhibited a sharp S-shape curve in contrast to the columns fitted with the other porous membranes. The columns fitted with PTFE and Nylon membranes displayed breakthrough curves which although S-shaped were flatter overall. Overall the presence of biological activity improved the ammonia removal process before breakthrough was reached since all curves had the same general S-shape. KMI clinoptilolite, apparently, has the ability to act as a support for immobilized bacteria. These results also confirmed that the adsorption method as opposed to insitu cultivation, as an immobilization technique was found to be effective. Nitrite concentrations in the outlet streams from each column were also measured in the case of the columns fitted with the porous membranes using the procedure described earlier. Figs. 15 and 16 show the production of nitrite during the uptake runs, comparing columns fitted with each of the four different porous membranes. The nitrite concentration of the eluents from the PES and PTFE columns ranged between 5 and 20 mg NO 2 =l, see Fig. 15, until a sudden exponential increment occurred at approximately 450BV. The same observation was noticed for the columns fitted with polypropylene and nylon membranes, Fig. 16, with a sharp spike occurring in the later stages of the experiment. It is concluded that more nitrite was produced in the biologically active column fitted with the polyethersulfone membrane compared with the column fitted with the PTFE membrane. This agrees with the breakthrough capacity calculations as the column fitted with the polyethersulfone membrane showed higher breakthrough

137

Fig. 16. Nitrite production of biologically enhanced PP and Nylon membrane column. Initial feed 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, trans-membrane pressure ¼ 1.5 psi.

capacity. On the other hand more nitrite was produced in the columns fitted with polypropylene membranes and nylon membranes compared with the columns fitted with the polyethersulfone or PTFE membranes. For example at bed volumes treated value of 300 BV as shown in Fig. 16, the column fitted with the polypropylene membrane shows a higher nitrite concentration in the eluent compared with the column fitted with the nylon membrane. This is consistent with results for ammonia removal and also the superior aeration performance of the PP membrane which would maintain higher oxygen availability for ammonia oxidation to nitrite. However at volumes of water fed beyond 600BV, the nitrite concentration in the effluent from the column fitted with the nylon membrane appears to be higher than that for column fitted with the polypropylene membrane, see Fig. 16. The apparent late spike in the nitrite concentrations in both cases is most likely explained by the slower kinetics of nitrite oxidation relative to ammonia oxidation. This would be especially true in the case of the nylon membrane case where the aeration efficiency of the membrane is much lower than the PP membrane thus leading to lower oxygen availability for nitrite oxidation. This was less so in the case of the PES/PTFE columns, since both columns reached breakthrough closely and yielding much lower nitrite concentrations overall. 9. Conclusions

Fig. 15. Nitrite production of biologically enhanced PES and PTFE membrane columns. Initial feed 40 mg N  NHþ 4 =l, particle size: 0.42e1.41 mm, flowrate 5.6 Lt/hr, bed height: 27 cm, trans-membrane pressure ¼ 1.5 psi.

Comparison of membranes for aeration of the water showed that PTFE showed the best aeration performance at the pressure levels used. The polypropylene membrane exhibited the second most effective aeration rates and was more effective compared with the polyethersulfone membrane. The nylon membrane was the least effective for aeration. In-situ aeration of the ion exchange columns equipped with a non-porous silicone membrane, in the absence of nitrification resulted in a very small increase in breakthrough performance and was consistent with the findings of earlier work. In the case of the biologically active ion exchange columns equipped with membrane aeration through the four porous membranes, significant increases in breakthrough capacity were observed in three cases in the order: polypropylene > polyethersulfone > polytetrafluorethylene (PTFE). The biologically active ion-exchange column fitted with the nylon

138

A. Almutairi, L.R. Weatherley / Journal of Environmental Management 160 (2015) 128e138

column produced an anomalous result showing a reduction in breakthrough capacity compared with the non-biologically active column. This remains unexplained. The tracking of the nitrite concentration in the effluent from each column showed much lower levels of nitrite present in the case of the columns fitted with PES and PTFE membranes, compared with the columns fitted with nylon membranes and polypropylene membranes. Acknowledgment The authors gratefully acknowledge a Kuwaiti Government Scholarship for Dr Almutairi. References Adams, C.E., Eckenfelder, W.W., 1977. Nitrification design approach for high strength ammonia wastewaters. J. Water Pollut. Control Fed. 49 (3), 413e421. Ahn, D.H., Chung, Y.C., Chang, W.S., 2002. Use of coagulant and zeolite to enhance the biological treatment efficiency of high ammonia leachate. J. Environ. Sci. Health Part A 37 (2), 163e173. Beler-Baykal, B., Oldenburg, M., Sekoulov, L., 1994. Post equalization of ammonia peaks. Water Res. 28 (9), 2039e2042. Beler-Baykal, B., Allar, A.D., 2008. Upgrading fertilizer production wastewater effluent quality for ammonium discharges through ion exchange with clinoptilolite. Environ. Technol. 29 (6), 665e672. Bernier, J., Rocher, V., Guerin, S., Lessard, P., 2014. Modelling the nitrification in a full-scale tertiary biological aerated filter unit. Bioprocess Biosyst. Eng. 37, 289e300. Bish, D.L., Ming, D.W. (Eds.), 2001. Natural Zeolites, Occurrence, Properties, Application. Reviews in Mineralogy and Geochemistry, vol. 45. Burrel, P.C., Keller, J., Blackall, L.L., 1998. Microbiology of a nitrite oxidizing bioreactor. May Appl. Environ. Microbiol. 1878e1883. Gunay, A., 2007. Application of nonlinear regression analysis for ammonium exchange by natural (Bigadic) clinoptilolite. J. Hazard. Mater. 148 (3), 708e713. Guo, X., Zeng, L., Li, X., Park, H., 2008. Ammonium and potassium removal for anaerobically digested wastewater using natural clinoptilolite followed by membrane pre-treatment. J. Hazard. Mater. 151 (1), 125e133. Hedstrom, A., Amofah, L.R., 2008. Adsorption and desorption of ammonium by clinoptilolite adsorbent in municipal wastewater treatment systems. J. Environ. Eng. Sci. 7 (1), 53e61. Huang, G., Liu, F., Yang, Y., Deng, W., Li, S., Huang, Y., Kong, X., 2015. Removal of ammonium-nitrogen from groundwater using a fully passive permeable reactive barrier with oxygen-releasing compound and clinoptilolite. J. Environ. Manag. 154, 1e7. Jorgensen, T.C., Weatherley, L.R., 2003. Ammonia removal from wastewater by ion

exchange in the presence of organic contaminants. Water Res. 37 (8), 1723e1728. Karadag, D., Akkaya, E., Demir, A., Saral, A., Turan, M., Ozturk, M., 2008. Ammonium removal from municipal landfill leachate by clinoptilolite bed columns: breakthrough modeling and error analysis. Ind. Eng. Chem. Res. 47 (23), 9552e9557. Kuhlman, W., 1987. Optimization of a membrane oxygenation system for cell culture in stirred tank reactors. Dev. Biol. Stand. 66, 263e268. Lahav, O., Green, M., 1998. Ammonium removal using ion exchange and biological regeneration. Water Res. 32 (7), 2019e2028. Lee, C.O., Boe-Hansen, R., Musovic, S., Smets, B., Albrechtsen, H., Binning, P., 2014. Effects of dynamic operating conditions o nitrification in biological rapid sand filters for drinking water treatment. Water Res. 64, 226e236. McVeigh, R.J., Weatherley, L.R., 1999. Ammonium ion (NHþ 4 ) removal from secondary effluent through ion-exchange: the effect of biological activity and the presence of other cations. Water Sci. Technol. 40 (2), 143e149. Mazeikiene, A., Valentukeviciene, M., Rimeika, M., 2008. Removal of nitrates and ammonium ions from water using natural sorbent zeolite (clinoptilolite). J. Environ. Eng. Landsc. Manag. 16 (1), 38e44. Mazeikiene, A., Valentukeviciene, M., Jankauskas, J., Matuzevcius, A.B., Dauknys, R., 2010. Laboratory study of ammonium ion removal by using zeolite (Clinoptiolite) to treat drinking water. J. Environ. Eng. Landsc. Manag. 18 (1), 54e61. Metcalf, L., Eddy, H.P., Tchobanoglous, G., 2004. Effects of Dynamic Operating Conditions on Nitrification in Biological Rapid Sand Filters for Drinking Water Treatment Wastewater Engineering: Treatment, Disposal, and Reuse McGrawHill Series in Water Resources and Environmental Engineering. McGraw-Hill, New York. Miladinovic, N., Weatherley, L.R., 2008. Intensification of ammonia removal in a combined ion-exchange and nitrification column. Chem. Eng. J. 135 (1e2), 15e24. Mishra, M., Jain, S.K., 2011. Properties and applications of zeolites: a review. Proc. Natl. Acad. Sci. India, Sect. B-Biol. Sci. 81 (3), 250e259. Part: 3. nchez, E., Mila n, Z., Corte s, I., Angeles de la la Montalvo, S., Guerrero, L., Borja, R., Sa Rubia, M., 2012. Application of natural zeolites in anaerobic digestion processes: a review. Appl. Clay Sci. 58, 125e133. Siljeg, M., Foglar, L., Kukucka, M., 2010. The ground water ammonium sorption onto Croatian and Serbian clinoptilolite. J. Hazard. Mater. 178 (1e3), 572e577. Sims, R.C., Little, L.W., 1973. Enhanced nitrification by addition of clinoptilolite to tertiary activated sludge. Environ. Lett. 4 (1), 27e34. Smith, D.P., 2011. Chabazite biofilter for enhanced stormwater nitrogen removal. Water Environ. Res. 83 (4), 373e384. Stenstrom, M.K., Poduska, R.A., 1980. The effect of dissolved oxygen concentration on nitrification. Water Res. 14 (6), 643e649. Tsitsishvili, G.V., Andronikashvili, T.G., Kirov, G.N., Filizova, L.D., 1992. Natural Zeolites, first ed. Ellis Horwood, London. Vassileva, P., Voikova, D., 2009. Investigation on natural and pretreated Bulgarian clinoptilolite for ammonium ions removal from aqueous solutions. J. Hazard. Mater. 170 (2e3), 948e953. Wang, J.L., Yang, N., 2004. Partial nitrification under limited dissolved oxygen conditions. Process Biochem. 39 (10), 1223e1229.