lable at ScienceDirect

Journal of Environmental Management 154 (2015) 1e7

Contents lists avai

Journal of Environmental Management

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

Removal of ammonium-nitrogen from groundwater using a fullypassive permeable reactive barrier with oxygen-releasing compoundand clinoptilolite

Guoxin Huang a, b, Fei Liu a, *, Yingzhao Yang c, Wei Deng d, Shengpin Li a,Yuanying Huang e, Xiangke Kong f

a Beijing Key Laboratory of Water Resources & Environmental Engineering, China University of Geosciences (Beijing), Beijing 100083, Chinab China Meat Research Center, Beijing Academy of Food Sciences, Beijing 100068, Chinac Hydro-Engineering Team of Sichuan Metallurgical Geology & Exploration Bureau, Chengdu 611730, Chinad College of Environmental and Energy Engineering, Beijing University of Technology, Beijing 100124, Chinae National Research Center for Geoanalysis, Beijing 100037, Chinaf Institute of Hydrogeology and Environmental Geology, CAGS, Shijiazhuang 050061, China

a r t i c l e i n f o

Article history:Received 14 November 2014Received in revised form6 February 2015Accepted 10 February 2015Available online

Keywords:Ammonium-nitrogenGroundwaterIn situ remediationPermeable reactive barrierOxygen-releasing compoundClinoptilolite

* Corresponding author. 29 Xueyuan Road, HaidianE-mail address: 21584612@qq.com (F. Liu).

http://dx.doi.org/10.1016/j.jenvman.2015.02.0120301-4797/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

A novel fully passive permeable reactive barrier (PRB) with oxygen-releasing compound (ORC) and cli-noptilolite was proposed for the removal of ammonium-nitrogen from groundwater. The PRB involves acombination of oxygen release, biological nitrification, ion exchange, and bioregeneration. A pilot-scaleperformance comparison experiment was carried out employing three parallel columns to assess theproposed PRB. The results showed that the PRB achieved nearly complete NH4

þ � N depletion (>99%).NH4

þ �N of 5.23e10.88 mg/L was removed, and NO2� � N of <1.93 mg/L and NO3

� � N of 2.03e19.67 mg/L were generated. Ion exchange and biological nitrification both contributed to NH4

þ � Nremoval, and the latter played a dominant role under the condition of sufficient oxygen. Biologicalnitrification favored a delay in sorption saturation and a release of exchange sites. The ORC could suf-ficiently, efficiently supply oxygen for approximately 120 pore volumes. The clinoptilolite ensured arobust NH4

þ � N removal in case of temporary insufficient biological activities. No external alkalinitysources had to be supplied and no inhibition of aerobic metabolism occurred. The ceramicite had anegligible effect on the biomass growth. Based on the research findings, a full-scale continuous wall PRBwas installed in Shenyang, China in 2012.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Ammonium-nitrogen contamination of groundwater hasbecome an environmental and public health issue in developingand developed countries. In Shenyang, China, NH4

þ � N of up to10 mg/L in an approximately 40 m deep aquifer beside Hun Riverhas been observed by our research group. Further, in America,Australia, England, and Korea, high NH4

þ � N has been recorded innumerous groundwater contamination plumes (Park et al., 2002;Manning and Hutcheon, 2004; Patterson et al., 2004; Miller andSmith, 2009). It is most likely that NH4

þ � N forms harmful disin-fection by-products, chlorine-related tastes and odors during

District, Beijing 10083, China.

drinking water treatment, and causes nitrifying bacterial regrowthin drinking water distribution systems (Zhang et al., 2014).

In situ permeable reactive barrier (PRB) is regarded as one of themost promising alternatives to conventional ex situ technologies(e.g., pump-and-treat) for groundwater remediation, and isreceiving increased attention in the literature (Della Rocca et al.,2007; Thiruvenkatachari et al., 2008; Moraci and Calabr�o, 2010;Hashim et al., 2011; Perego et al., 2013; Erto et al., 2014).Recently, several lab- and full-scale PRBs have been reported forremediating groundwater contaminated with NH4

þ � N, involvingaerobic biological nitrification and/or physicochemical sorption(Park et al., 2002; Patterson et al., 2004; van Nooten et al., 2010,2011). For single biological nitrification (Lahav and Green, 2000;Yusof et al., 2010; Mousavi et al., 2012), two semipassive PRBswere packedwith polymermats for delivering oxygen gas to induce

G. Huang et al. / Journal of Environmental Management 154 (2015) 1e72

oxidation of NH4þ � N into NO3

� � N=NO2� � N (Patterson et al.,

2002). Another semipassive PRB was equipped with diffusive ox-ygen emitters to provide a nonlimiting molecular oxygen supply(van Nooten et al., 2008). Even though these PRBs achieved goodnitrification performances, they produced high residual dissolvedoxygen (DO) of 24.0 ± 9.7 mg/L (van Nooten et al., 2008), andcaused higher operation and maintenance costs as well as opera-tional complexity arising from the use of oxygen injection systems(reservoirs þ pumps þ plumbing) and electric power. To overcomethese shortcomings, oxygen-releasing compound (ORC) includingcalcium peroxide and magnesium peroxide presents a potentialchoice because it can combine with water to form oxygen gas(Cassidy and Irvine, 1999; Liu et al., 2006; Dong et al., 2009). ORCmixed well with cement, sand, and other materials has been suc-cessfully applied to aerobic biodegradation of some organic con-taminants such as BTEX (benzene, toluene, ethylbenzene, and p-xylene), MTBE (methyl tert-butyl ether), and COD (chemical oxygendemand) (Liu et al., 2006; Dong et al., 2009; Yeh et al., 2010).However, very little is known about ORC for supporting biologicalnitrification in PRBs. The corresponding ORC oxygen-releasing ca-pacity, N removal and transformation mechanisms, and inorganicgeochemical characteristics are worthy of further study. For singleion exchange (Cooney et al., 1999; Dong et al., 2009), zeolite,including clinoptilolite, mordenite, heulandite, etc. is the mostcommon ammonium ion exchanger. Zeolite has been welldemonstrated to be effective for treatment of NH4

þ � N and suit-able for in situ applications as a PRB material (Booker et al., 1996;Park et al., 2002; Du et al., 2005; Wen et al., 2006). Even so, ques-tions arise regarding sorption saturation and regeneration. As iswell known, either a periodical material replacement or chemicalregeneration after saturation renders the PRB technology lesspassive and less economically favorable (van Nooten et al., 2010).Chemical regenerants account for more than 50e60% of the totalcosts even with regenerant reuse (Njoroge and Mwamachi, 2004).Until now, a few research studies have focused on in situ biologicalregeneration of zeolite by means of injecting external molecularoxygen into PRBs (van Nooten et al., 2008, 2010, 2011). However,using ORC to accomplish bioregeneration or delay NH4

þ � Nbreakthrough is still lacking.

Here, we report a novel fully passive ORC and clinoptilolite PRBfor removing NH4

þ � N from groundwater. The concept is based ona combination of oxygen release-biological nitrification-ionexchange-bioregeneration for the efficient depletion of NH4

þ � Nand the delayed saturation of clinoptilolite. The main function ofORC in an upgradient zone is to release oxygen. Clinoptilolite in adowngradient zone is to remove residual NH4

þ � N, ensuring arobust NH4

þ � N removal. Shifts in exchange equilibria allow theoccurrence of NH4

þ � N desorption when ammonium-poorgroundwater is flowing through the saturated clinoptilolite. Thenitrifying bacteria on the reactive media utilize the released oxygento oxidize soluble NH4

þ � N inwater, increasing the N products anddelaying the NH4

þ � N breakthrough. The objectives of this studywere to: (1) evaluate the NH4

þ � N removal performance of theproposed PRB; (2) characterize the ORC oxygen-releasing capacity;(3) explore the clean up mechanisms; (4) investigate the inorganicgeochemistry; (5) determine the feasibility of bioregeneration ordelayed sorption saturation; and (6) identify the effect of ceramiciteon biomass growth.

2. Materials and methods

2.1. Materials and chemicals

River sand (~1mm)was donated by a quarry in Shenyang, China.Natural clinoptilolite (1e2 mm, purity 97%, cation exchange

capacity 130e150 mg/100 g, BrunauereEmmetteTeller area670m2/g) was obtained from Faku, China and found to contain SiO2(67.09 wt%), Al2O3 (12.44 wt%), CaO (8.81 wt%), MgO (1.22 wt%),K2O (1.20 wt%), and Fe2O3 (0.78 wt%). ORC beads (2e3 cm) wereobtained from China University of Geosciences (Beijing), China,which consisted of CaO2 (20 wt%, purity 50%, main impurityCa(OH)2), Ca-bentonite (10 wt%), portland cement (25 wt%), riversand (15wt%), and tapwater (30 wt%) (trial production). Ceramicite(1e2 mm) was purchased from Ma'anshan Huaqi EnvironmentalProtection Sci-Tech Development Co., Ltd., China. Shenyang tapwater spiked with ammonium chloride was used as syntheticgroundwater with initial NH4

þ � N of 5.0e11.0 mg/L. Unlessotherwise indicated, all chemicals used were analytical reagentgrade as received.

2.2. Column experiment setup and operation

A pilot-scale performance comparison experiment was carriedout employing three parallel Plexiglas columns (PRBs 1e3, 20 cmi.d., 180 cm high) (Fig. 1 and Table 1). PRB 1 was packed with cli-noptilolite (150 cm high) to control for the effects of oxygen release,ion exchange, and biological nitrification. PRB 2 was packed withORC (30 cm high) and clinoptilolite (120 cm high) to simulate theproposed PRB. PRB 3 was packedwith ORC (30 cm high), ceramicite(30 cm high), and clinoptilolite (90 cm high) to identify the effect ofceramicite as supporting material on the biological growth. 8 in-termediate sampling ports were positioned along the height ofeach PRB, at 20 (port 1), 40 (port 2), 60 (port 3), 80 (port 4), 100(port 5),120 (port 6),140 (port 7), and 160 cm (port 8) from the inletend. River sand (30 cm high) in the sand zones near the inlets wasto simulate a natural aquifer and ensure flow distribution. Riversand was mixed well with ORC beads in the ORC zones in an effortto buffer the rise in pH. U-bend polyvinyl chloride tubing was uti-lized to control water table and carry effluent away for disposal.

It is important to note that the PRBs were not inoculated andtheir packing media were not replaced during operation. Each PRBwas continuously fed with synthetic groundwater in a downflowmode using a multiport peristaltic pump set (BT 100-1F drive, DG-4pump head, Baoding Longer Precision Pump Co., Ltd., China) at aflow rate of 15 mL/min (Darcy velocity ¼ 0.7 m/d). PRBs 1, 2, and 3were operated for 183, 282, and 234 pore volumes (PVs) (corre-sponding to 147, 188, and 188 days) in the dark at 13e28 �C,respectively. Water samples were collected from the inlets, outlets,and intermediate sampling ports, and stored at 4 �C until analysis.

2.3. Analytical methods, instruments and data processing

NH4þ � N was determined using a flow injection analyzer

(Lachat, Model QC 8000, USA), employing spectrophotometry withsalicylic acid. NO3

� � N and NO2� � Nwere measured using an ion

chromatograph (Dionex, Model ICS-2100, USA). The lower detec-tion limits for NH4

þ � N, NO3� � N, and NO2

� � N were 0.01, 0.10,and 0.03 mg/L, respectively. DO and water temperature weremeasured using a portable DO meter (Hach, Model HQ30d, USA).pH was monitored using a digital pH meter (Hach, Model MP-6,USA). Total hardness (TH) was determined using EDTA (ethyl-enediaminetetraacetic acid) titration method (APHA, 2005). Totalalkalinity (TA) was measured using an alkalinity test kit (Hach,Model AL-DT, USA), employing 0.16 N H2SO4 with a digital titrator(Model 16900). In each analysis, at least one in five samples wasduplicated and the deviation between the two samples was alwaysless than 5%. Total nitrogen (TN) (mg/L) was defined as the sum ofNH4

þ � N (mg/L), NO2� � N (mg/L), and NO3

� � N (mg/L). Thenumber of PVs was expressed as the ratio of accumulated watervolume (L) over time to pore volume of reactive media (L). Data are

Fig. 1. (a) Schematic and (b) photograph of the continuous flow experimental setup showing permeable reactive barriers (PRBs) 1, 2, and 3 as well as packing structures, flowdirection, U-bend tubing, and sampling ports.

Table 1Configurations of the three permeable reactive barriers (PRBs) used for study.

Column Sand zone ORC zone Clinoptilolite zone Ceramicite zone Average porosity (%)

Sand (kg) Height (cm) ORC (kg) Sand (kg) Height (cm) Clinoptilolite (kg) Height (cm) Ceramicite (kg) Height (cm)

PRB 1 12.0 30 0 0 0 42.8 150 0 0 31PRB 2 12.0 30 6.8 5.5 30 33.5 120 0 0 25PRB 3 12.0 30 6.8 5.5 30 32.8 90 6.0 30 31

G. Huang et al. / Journal of Environmental Management 154 (2015) 1e7 3

presented as means of triplicate measurements.

3. Results and discussion

3.1. Overall performances of PRBs 1, 2, and 3

Nearly complete NH4þ � N depletion (percent removal >99%)

was achieved in PRBs 1, 2, and 3 (Fig. 2), demonstrating their idealremoval capacities. Nevertheless, different spatial and temporaltrends in NH4

þ � N distribution were observed: for PRB 1, at theinlet NH4

þ � N of 5.29e10.80 mg/L, low NH4þ � N (�0.11 mg/L)

was constantly found at ports 3, 6 and the outlet during operationover the 183 PVs (Fig. 2a); for PRB 2, at the inlet NH4

þ � N of5.29e10.90 mg/L, NH4

þ � N fluctuated between 0.01 and 7.67 mg/Lat port 3, and afterward was maintained at a low, stable level of�0.10 mg/L at port 6 and the outlet during operation over the 282PVs (Fig. 2b); for PRB 3, at the inlet NH4

þ � N of 5.10e10.70 mg/L,NH4

þ � N gradually decreased to a value below 0.10 mg/L betweenthe inlet and port 6, and was maintained invariably constant at theoutlet during operation over the 234 PVs (Fig. 2c). The differencesin NH4

þ � N distribution among the three PRBs were probablyattributed to different nitrogen removal and transformationmechanisms.

3.2. ORC oxygen-releasing capacity and DO distribution

The DO content at ports 3 of PRBs 2 (5.50e22.00 mg/L) and 3(8.47e22.00 mg/L) was much higher than the level at port 3 of PRB1 (�2.00 mg/L) (Fig. 3) and gas bubbles, presumably of oxygen,

were seen in the samples from ports 3 of PRBs 2 and 3 whensampling during the first half of operation (before PV 120 for PRB 2;before PV 100 for PRB 3). These suggested that the ORC beads had acapacity to provide oxygen (Eq. (1)). The spatial and temporaltrends in DO distribution for PRBs 2 and 3 were similar to eachother. The DO content was the highest at ports 3, and was followedby the inlets, outlets, and other intermediate ports during the firsthalf of operation due to the facts that ports 3 were downgradientandmost close to the ORC zones (Fig. 3b and c). However, like PRB 1,the DO concentration at the intermediate ports and the outlets waslower than that at the inlets during the second half of operation(after PV 120 for PRB 2; after PV 100 for PRB 3), demonstrating theexhaustion of the ORC beads (Fig. 3b and c). It is thereforeconcluded that the ORC medium could sufficiently, efficientlysupply oxygen for approximately 120 PVs for PRB 2 and 100 PVs forPRB 3.

CaO2 þ H2 O / 0:5O2 þ CaðOHÞ2 (1)

3.3. N removal and transformation

For PRB 1, NH4þ � N of 5.25e10.77 mg/L was reduced, and

NO2� � N of <1.19 mg/L and NO3

� � N of <2.13 mg/L were formedduring the 183 PVs (Fig. 4a). Besides, the sum of NO2

� � N andNO3

� � N formedwasmuch lower than NH4þ � N reduced, and TN

removed was a little lower than NH4þ � N removed (Fig. 4a). These

results indicated that only a proportion of NH4þ � N removed

(<26%) was biologically transformed to NO2� � N and NO3

� � N

Fig. 2. Changes in spatial and temporal distribution of NH4þ � N and its percent

removal in permeable reactive barriers (PRBs) (a) 1, (b) 2, and (c) 3 at a flow rate of15 mL/min in a long-term study. Distances of the sampling ports from inlet are given inthe legend, together with their names.

Fig. 3. Changes in spatial and temporal distribution of dissolved oxygen (DO) inpermeable reactive barriers (PRBs) (a) 1, (b) 2, and (c) 3 at a flow rate of 15 mL/min in along-term study. Vertical dashed lines represent the times when the oxygen-releasingcompound beads are exhausted. Distances of the sampling ports from inlet are given inthe legend, together with their names.

G. Huang et al. / Journal of Environmental Management 154 (2015) 1e74

(Eqs. (2)e(4)) due to the deficiency of oxygen (Fig. 3a). Taking intoaccount the usually very poor DO (<3 mg/L) in real groundwater(Yeh et al., 2010), biological nitrification will be further weaken inPRB 1. By combining the results of Figs. 2a and 4a, it is evident thatabiotic ion exchange by the clinoptilolite had a good ability toremove the remaining NH4

þ � N (>74%) (Eq. (5)) in the absence ofoxygen gas.

NHþ4 þ 1:5O2 ���������������!Ammonia oxidizing bacteria

NO�2 þ 2Hþ þ H2O

(2)

NO�2 þ 0:5O2���������������!

Nitrite oxidizing bacteriaNO�

3 (3)

NHþ4 þ 2O2�����������!

Nitrifying bacteriaNO�

3 þ 2Hþ þ H2O (4)

Z�Mþ þ NHþ4 / Z�NHþ

4 þ Mþ (5)

where Z� represents the zeolite, and Mþ represents theexchangeable cations.

For PRB 2, NH4þ � N of 5.23e10.88mg/L was lost, and NO2

� � Nof <1.93 mg/L and NO3

� � N of 2.03e19.67 mg/L were generatedduring the 221 PVs (Fig. 4b). Clearly, NH4

þ � N reduced was mainlyconverted to NO3

� � N, and to a minor extent to NO2� � N, sug-

gesting that biological nitrification played a dominant role inNH4

þ � N removal. Sometimes (e.g., PVs 9, 99, and 194), similar toPRB 1, the sum of NO2

� � N and NO3� � N formed was less than

NH4þ � N reduced and simultaneously TN removedwas lower than

NH4þ � N removed (Fig. 4b). Consequently, it can be deduced that

the rest of the NH4þ � N reduced was removed by ion exchange,

Fig. 4. Temporal changes in N concentrations for permeable reactive barriers (PRBs)(a) 1, (b) 2, and (c) 3 at a flow rate of 15 mL/min in a long-term study. Total nitrogen(TN) is the sum of NO3

� � N, NO2� �N, and NH4

þ �N.

Fig. 5. Scanning electronic microscope images of the clinoptilolite: (a) a biofilm-uncoated original surface; (b) a biofilm-coated surface.

G. Huang et al. / Journal of Environmental Management 154 (2015) 1e7 5

which was consistent with the literature assessing a combinedbiological nitrification-ion exchange approach in a sand tankexperiment (Li et al., 2014). Apparently, biotic nitrification andabiotic ion exchange both contributed to NH4

þ � N removal. Theclinoptilolite used ensured a robust NH4

þ � N removal via ion ex-change in case of temporary insufficient biological activities (e.g.,during the start-up (<18 PVs) or the ORC exhaustion phase (>120PVs)), as evidenced by data on NH4

þ � N percent removal, DO andNH4

þ � N concentrations, and NO2� � N and NO3

� � N formations(Figs. 2e4). In addition, the clinoptilolite served as a solid carrier forthe nitrifying bacterial growth, as shown by comparing the varia-tion in the clinoptilolite's surface (Fig. 5a and b).

It is interesting to note that NO2� � N þ NO3

� � N generatedexceeded NH4

þ � N removed and negative TN removal was foundat some points in time (such as PVs 18, 68, 78, and 120) (Fig. 4b),suggesting that the partial NH4

þ � N adsorbed onto the clinopti-lolite was released as a result of shifts in exchange equilibria andconcomitantly biologically oxidized by the nitrifying bacteria. Thesimultaneous desorption and nitrification could be expressed bythe following simplified reaction (Eq. (6)). Apparently, biologicalnitrification was favorable for a delay in sorption saturation and arelease of exchange sites, which threw a new light on sustainableuse of sorption materials for in situ groundwater remediation.

Z�NHþ4 þ 2O2 �������!Nitrifying bacteria

Z�Hþ þ Hþ þ NO�3 þ H2O (6)

where Z� represents the zeolite.For PRB 3, NH4

þ � N of 5.07e10.66 mg/L was eliminated, andNO2

� � N of <2.77 mg/L and NO3� � N of 2.42e27.56 mg/L were

produced during the 183 PVs (Fig. 4b). Likewise, PRB 3 also relied onbiological nitrification and abiotic sorption to deplete NH4

þ � N,and a nitrifier-mediated release of exchange sites also occurred.

3.4. Inorganic geochemistry

Like PRB 1, the significant increase in pH was noted in PRBs 2and 3 in the early period of operation, as demonstrated by the inletand outlet values (Fig. 6). In a previous study, a similar phenome-non was also found in a PRB packed with clinoptilolite: the pHincreased from 6.8 to 7.9 (Park et al., 2002). Hydration of clinopti-lolite was mainly responsible for the pH increase. As was apparentfrom Fig. 6b and c, there were close, positive correlations betweenthe changes in pH and alkalinity in PRBs 2 and 3. Hþ can begenerated when NH4

þ � N is biologically converted into NO3� � N

and/or NO2� � N (Eqs. (2) and (4)) by nitrifiers in an aerobic

environment, resulting in the drops in pH and alkalinity. Incontrast, Ca(OH)2 can be produced when CaO2 releases oxygen gas(Eq. (1)), leading to the rises in pH and alkalinity. The river sand inthe ORC zones can act as a pH buffer, and control the drastical pHand alkalinity changes (Xie et al., 2010). Undoubtedly, consideringthe hydration of clinoptilolite, the generation of Hþ and Ca(OH)2,and the pH buffering of river sand, their net result was the increasesin pH and alkalinity in PRBs 2 and 3. This suggested that the alka-linity production by both the hydration of clinoptilolite and thedecomposition of CaO2 exceeded the alkalinity consumption by thenitrification of NH4

þ � N. Therefore, no external alkalinity sources

Fig. 6. Temporal changes in total alkalinity (TA), total hardness (TH), and pH forpermeable reactive barriers (PRBs) (a) 1, (b) 2, and (c) 3 at a flow rate of 15 mL/min in along-term study.

G. Huang et al. / Journal of Environmental Management 154 (2015) 1e76

had to be supplied for the nitrifying bacterial growth.On the other hand, in combination with the changes in

NH4þ � N, NO2

� � N, NO3� � N, and pH (Figs. 4 and 6), the pH

(7.32e7.94 for PRB 2; 7.32e8.21 for PRB 3) did not appear to bedetrimental to the enzymatic activity of the nitrifying bacteria.Researchers have reported that a pH value helping for microbialgrowth should be kept in the range of 6.5e8.5 (Liu et al., 2006;Yusof et al., 2010; Mousavi et al., 2012). Kong et al. (2014) foundthe high pH (>8.5) in the effluent of an aerobic reaction columnpacked with zeolite during the first 15 days. To ensure a robust pHcontrol system in future applications, the combination of K2HPO4and KH2PO4 can be used as the basic components of the ORC beads(Yeh et al., 2010).

There was a decrease in hardness in the outlet of PRB 1 (Fig. 6a),suggesting that the clinoptilolite was able to sorb the hardness(Dong et al., 2009). Similarly, a diminished hardness was alsoobserved in the outlets of PRBs 2 and 3 during the first half ofoperation (Fig. 6b and c), indicating that the clinoptilolite was alsoable to sorb the hardness caused by CaO2, Ca(OH)2, and Ca-bentonite from the ORC material. However, the sorption capacity

thereafter decreased with time increasing, leading to the outlethardness being higher than the inlet value during the second half ofoperation (Fig. 6b and c).

3.5. Implications of pilot-scale performances to full-scaleapplication

PRB 1 significantly depended on ion exchange using the cli-noptilolite to eliminate NH4

þ � N under the condition of insuffi-cient oxygen (Figs. 3a and 4a). Nevertheless, the clinoptilolite willeventually become saturated in the presence of NH4

þ and othercompeting cations (Naþ, Kþ, Ca2þ, etc.), making ion exchange un-dependable for groundwater treatment. Also, periodical chemicalregeneration or even a medium replacement will render this PRBless economically favorable.

On a basis of observed NH4þ � N at port 4 of PRB 3 (Fig. 2c), the

ceramicite as a biofilm carrier had a negligible effect on the biomassgrowth. There were a slight difference in NH4

þ � N removal per-formance (Fig. 2b and c) and a similar change in N transformationbetween PRBs 2 and 3 (Fig. 4b and c). Furthermore, ceramicite isapproximately 10 times higher than clinoptilolite. For these rea-sons, the ceramicite is not recommend as a reactive medium forfuture field applications.

The pH of 7.32e8.21 and the maximum hardness of 400 mgCaCO3/L in the outlets from PRBs 2 and 3 (Fig. 6b and c) werelower than their respective limitation values (pH 6.5e8.5 andhardness 450 mg CaCO3/L for drinking use) of quality standard forgroundwater of China (SBTS, 1993). Hence, there were minornegative effects of pH and hardness on groundwaterhydrochemistry.

From the above analysis, ORC and clinoptilolite were chosen asthe reactive media for the full-scale groundwater plume remedia-tion at a site beside Hun River in Shenyang, China. Following design,a continuous wall PRB, consisting of an upgradient ORC zone (5oxygen source wells, 400 mm i.d., 40 m deep) and a downgradientclinoptilolite zone (15 m long, 0.6e1 m wide, 40 m deep) wasinstalled across the plume to remove NH4

þ � N via ion exchangeand nitrification in July 2012. The changes in hydrogeological pa-rameters, mineral precipitaiton, and microorganism identificationare being investigated.

4. Conclusions

Based on the research findings the main conclusions were asfollows:

(1) PRBs 1, 2, and 3 all had ideal NH4þ � N removal capacities

(percent removal >99%). PRB 1 significantly depended on ionexchange to eliminate NH4

þ � N due to a lack of oxygen.PRBs 2 and 3 dominantly relied on biological nitrification toremove NH4

þ � N under the condition of sufficient oxygen.(2) The ORCmedium could sufficiently, efficiently supply oxygen

for approximately 120 and 100 PVs for PRBs 2 and 3,respectively.

(3) Biological nitrification was favorable for a delay in sorptionsaturation and a release of exchange sites, and the clinopti-lolite ensured a robust NH4

þ � N removal in case of tem-porary insufficient biological activities and served as a solidcarrier for the nitrifying bacterial growth.

(4) No external alkalinity sources had to be supplied for the ni-trifying bacterial growth in PRBs 2 and 3, and the pH(7.32e7.94 for PRB 2; 7.32e8.21 for PRB 3) did not appear tobe detrimental to the enzymatic activity of the nitrifyingbacteria.

G. Huang et al. / Journal of Environmental Management 154 (2015) 1e7 7

(5) The ceramicite had a negligible effect on the biomass growth.It is not recommend as a reactive medium for future fieldapplications.

Acknowledgments

This study is financially supported jointly by NSFC (41172226),National Program of Control and Treatment of Water Pollution(2009ZX07424-002-002), China Postdoctoral Science Foundation(2013M541111) and Beijing Excellent Talents Program(2012D001055000001).

References

APHA, 2005. Standard Methods for the Examination of Water and Wastewater,twenty-first ed. American Public Health Association, Washington, DC.

Booker, N.A., Cooney, E.L., Priestley, A.J., 1996. Ammonia removal from sewage usingnatural Australian zeolite. Water Sci. Technol. 34, 17e24.

Cassidy, D.P., Irvine, R.L., 1999. Use of calcium peroxide to provide oxygen forcontaminant biodegradation in a saturated soil. J. Hazard. Mater. B69, 25e39.

Cooney, E.L., Booker, N.A., Shallcross, D.C., Stevens, G.W., 1999. Ammonia removalfrom wastewaters using natural australian zeolite. I. characterization of thezeolite. Sep. Sci. Technol. 34, 2307e2327.

Della Rocca, C., Belgiorno, V., Meriç, S., 2007. Overview of in-situ applicable nitrateremoval processes. Desalination 204, 46e62.

Dong, J., Zhao, Y., Zhang, W., Hong, M., 2009. Laboratory study on sequencedpermeable reactive barrier remediation for landfill leachate-contaminatedgroundwater. J. Hazard. Mater. 161, 224e230.

Du, Q., Liu, S., Cao, Z., Wang, Y., 2005. Ammonia removal from aqueous solutionusing natural Chinese clinoptilolite. Sep. Purif. Technol. 44, 229e234.

Erto, A., Bortone, I., Di Nardo, A., Di Natale, M., Musmarra, D., 2014. Permeableadsorptive barrier (PAB) for the remediation of groundwater simultaneouslycontaminated by some chlorinated organic compounds. J. Environ. Manag. 140,111e119.

Hashim, M.A., Mukhopadhyay, S., Sahu, J.N., Sengupta, B., 2011. Remediation tech-nologies for heavy metal contaminated groundwater. J. Environ. Manag. 92,2355e2388.

Kong, X., Bi, E., Liu, F., Huang, G., Ma, J., 2014. Laboratory column study for evalu-ating a multimedia permeable reactive barrier for the remediation of ammo-nium contaminated groundwater. Environ. Technol. 22, 1e8.

Lahav, O., Green, M., 2000. Bioregenerated ion-exchange process: the effect of thebiofilm on the ion-exchange capacity and kinetics. Water SA 26, 51e57.

Li, S., Huang, G., Kong, X., Yang, Y., Liu, F., Hou, G., Chen, H., 2014. Ammoniumremoval from groundwater using a zeolite permeable reactive barrier: a pilot-scale demonstration. Water Sci. Technol. 70, 1540e1547.

Liu, S.-J., Jiang, B., Huang, G.-Q., Li, X.-G., 2006. Laboratory column study forremediation of MTBE-contaminated groundwater using a biological two-layerpermeable barrier. Water Res. 40, 3401e3408.

Manning, D.A.C., Hutcheon, I.E., 2004. Distribution and mineralogical controls onammonium in deep groundwaters. Appl. Geochem 19, 1495e1503.

Miller, D.N., Smith, R.L., 2009. Microbial characterization of nitrification in ashallow, nitrogen contaminated aquifer, Cape Cod, Massachusetts and detectionof a novel cluster associated with nitrifying Betaproteobacteria. J. Contam.Hydrol. 103, 182e193.

Mousavi, S., Ibrahim, S., Aroua, M.K., 2012. Sequential nitrification and denitrifica-tion in a novel palm shell granular activated carbon twin-chamber upflow bio-electrochemical reactor for treating ammonium-rich wastewater. Bioresour.Technol. 125, 256e266.

Moraci, N., Calabr�o, P.S., 2010. Heavy metals removal and hydraulic performance inzero-valent iron/pumice permeable reactive barrier. J. Environ. Manag. 91,2336e2341.

Njoroge, B.N.K., Mwamachi, S.G., 2004. Ammonia removal from an aqueous solutionby the use of a natural zeolite. J. Environ. Eng. Sci. 3, 147e154.

Park, J.-B., Lee, S.-H., Lee, J.-W., Lee, C.-Y., 2002. Lab scale experiments for permeablereactive barriers against contaminated groundwater with ammonium andheavy metals using clinoptilolite (01-29B). J. Hazard. Mater. B95, 65e79.

Patterson, B.M., Grassi, M.E., Davis, G.B., Robertson, B.S., Mckinley, A.J., 2002. Use ofpolymer mats in series for sequential reactive barrier remediation ofammonium-contaminated groundwater: laboratory column evaluation. Envi-ron. Sci. Technol. 36, 3439e3445.

Patterson, B.M., Grassi, M.E., Robertson, B.S., Davis, G.B., Smith, A.J., Mckinley, A.J.,2004. Use of polymer mats in series for sequential reactive barrier remediationof ammonium-contaminated groundwater: field evaluation. Environ. Sci.Technol. 38, 6846e6854.

Perego, C., Bagatin, R., Tagliabue, M., Vignola, R., 2013. Zeolites and related meso-porous materials for multi-talented environmental solutions. MicroporousMesoporous Mater. 166, 37e49.

SBTS (State Bureau of Technical Supervision), 1993. Quality Standard for GroundWater (GB/T 14848-93). Beijing.

Thiruvenkatachari, R., Vigneswaran, S., Naidu, R., 2008. Permeable reactive barrierfor groundwater remediation. J. Ind. Eng. Chem. 14, 145e156.

van Nooten, T., Diels, L., Bastiaens, L., 2011. A laboratory-scale mixed multibarrier forremoval of ammonium from landfill leachate contamination. Int. J. Environ. Eng.3, 240e252.

van Nooten, T., Diels, L., Bastiaens, L., 2008. Design of a multifunctional permeablereactive barrier for the treatment of landfill leachate contamination, laboratorycolumn evaluation. Environ. Sci. Technol. 42, 8890e8895.

van Nooten, T., Diels, L., Bastiaens, L., 2010. Microbially mediated clinoptiloliteregeneration in a multifunctional permeable reactive barrier used to removeammonium from landfill leachate contamination: laboratory column evalua-tion. Environ. Sci. Technol. 44, 3486e3492.

Wen, D., Ho, Y.-S., Tang, X., 2006. Comparative sorption kinetic studies of ammo-nium onto zeolite. J. Hazard. Mater. B133, 252e256.

Xie, L., Liu, F., Liu, Y., 2010. Improving property of filler in oxygen-releasingpermeable reactive barrier. Environ. Sci. Technol. 33, 44e48 (in Chinese).

Yeh, C.-H., Lin, C.-W., Wu, C.-H., 2010. A permeable reactive barrier for the biore-mediation of BTEX- contaminated groundwater: microbial community distri-bution and removal efficiencies. J. Hazard. Mater. 178, 74e80.

Yusof, N., Hassan, M.A., Phang, L.Y., Tabatabaei, M., Othman, M.R., Mori, M.,Wakisaka, M., Sakai, K., Shirai, Y., 2010. Nitrification of ammonium-rich sanitarylandfill leachate. Waste Manag. 30, 100e109.

Zhang, S., Wang, Y., He, W., Wu, M., Xing, M., Yang, J., Gao, N., Pan, M., 2014. Impactsof temperature and nitrifying community on nitrification kinetics in a moving-bed biofilm reactor treating polluted raw water. Chem. Eng. J. 236, 242e250.