Application of ceramic membranes with pre-ozonation for treatment of secondary wastewater effluent

Application of ceramic membranes with pre-ozonation for treatment of secondary wastewater ef?uent

S.Geno Lehman *,Li Liu 1

MWH Americas,Inc.,National Technology Group,618Michillinda Ave.,Suite 200,Arcadia,CA 91007,USA

a r t i c l e i n f o

Article history:

Received 21August 2008Received in revised form 19December 2008Accepted 2February 2009

Published online 10February 2009Keywords:Membranes Ceramic

Wastewater reuse Pre-ozonation

a b s t r a c t

Membrane fouling is an inevitable problem when micro?ltration (MF)and ultra?ltraion (UF)are used to treat wastewater treatment plant (WWTP)ef?uent.While historically the use of MF/UF for water and wastewater treatment has been almost exclusively focused on polymeric membranes,new generation ceramic membranes were recently introduced in the market and they possess unique advantages over currently available polymeric membranes.Ceramic membranes are mechanically superior and are more resistant to severe chemical and thermal environments.Due to the robustness of ceramic membranes,strong oxidants such as ozone can be used as pretreatment to reduce the membrane fouling.This paper presents results of a pilot study designed to investigate the application of new generation ceramic membranes for WWTP ef?uent treatment.Ozonation and coagulation pretreatment were evaluated to optimize the membrane operation.The ceramic membrane demonstrated stable performance at a ?ltration ?ux of 100gfd (170LMH)at 20 C with pretreatment using PACl (1mg/L as Al)and ozone (4mg/L).To understand the effects of ozone and coagulation pretreatment on organic foulants,natural organic matter (NOM)in four waters –raw,ozone treated,coagulation treated,and ozone followed by coagulation treated wastewaters –were characterized using high performance size exclusion chromatography (HPSEC).The HPSEC analysis demonstrated that ozone treatment is effective at degrading colloidal NOMs which are likely responsible for the majority of membrane fouling.

1.Introduction

Application of micro?ltration (MF)and ultra?ltration (UF)processes in wastewater treatment is gaining popularity in the United States (US)and the world today as a result of increas-ingly stringent discharge standards and increased water reclamation https://www.360docs.net/doc/ab16252852.html,e of MF/UF to treat secondary or tertiary ef?uent provides enhanced water quality compared to the conventional treatment processes.However,an inevitable

problem with membrane processes is the loss of membrane productivity over time,i.e.membrane fouling (Adham et al.,1996;Crozes et al.,1993;Seidel and Elimelech,2002).

While numerous studies have been performed for a better understanding of the fouling phenomenon,it is generally acknowledged that the performance of MF/UF is in?uenced by the feed water characteristics,membrane type and operational conditions.Wastewater treatment plant (WWTP)ef?uent is characterized by organic foulants with a biological background,

*Corresponding author .Tel.:t16265686004;fax:t16265686015.

E-mail addresses:geno.lehman@https://www.360docs.net/doc/ab16252852.html, (S.G.Lehman),li.liu@https://www.360docs.net/doc/ab16252852.html, (L.Liu).1

Tel.:t1626568

6013.

A v a i l a b l e a t w w w.s c i e n c e d i re ct.c o m

j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /w a t r e s

w a t e r r e s e a r c h 43(2009)2020–2028

which interact with the high concentrations of colloids and micro-particles in the feed water,and the membrane surface (Tchobanoglous et al.,1998).It has been reported that polymeric MF membrane often experienced rapid fouling when treating WWTP ef?uent(Roorda et al.,2005).The operating?ux,back-wash interval(BWI),and pretreatment processes(e.g.coagula-tion)have been identi?ed as key factors for the optimization of membrane performance.

The use of MF/UF to treat WWTP ef?uent has been almost exclusively focused on polymeric membranes,while little research has been performed with ceramic membranes.It is generally known that ceramic membranes are physically superior to polymeric membranes and are more resistant to severe chemical environments(Clement,2007;Lehman et al., 2007;Lee and Cho,2004;Shanbhag et al.,1998).As a result, ceramic membranes have been found to be able to operate at high permeate?uxes,high feed water recoveries,and with less frequent chemical cleaning as compared to conventional polymeric membranes(Clement,2007;Lehman et al.,2007; Lou-Brugger et al.,2007).When treating WWTP ef?uent, ceramic membranes could achieve enhanced?ltration performance for at least two reasons.First,the ceramic membrane provides excellent backwash ef?ciency as the material itself can withstand high backwash pressure. Second,it has been reported that the incorporation of chem-ical treatment as a pretreatment step with ceramic membranes greatly mitigates fouling by natural organic matters(Karnik et al.,2005).

Pretreatment using coagulation to mitigate NOM fouling of membrane processes has been well studied.However,few researchers have investigated the use of ozone as a pretreat-ment to membranes,due to the destruction of polymeric membranes by residual ozone(Shanbhag et al.,1998;Castro and Zander,1995;Shen et al.,1990).Short-term studies conducted using ozone-resistant polyvinylidene?uoride(PVDF)and pol-ysulfone(PS)membrane materials showed that using pre-ozonation did enhance the permeate?ux and reduce membrane fouling by the degradation of high molecular weight NOM(Lee et al.,2005;Hashino et al.,2000;Park,2002).However, only a handful of studies using small-scale ceramic membranes have been performed to show that enhanced?uxes could be achieved when ozonation pretreatment was used(Karnik et al., 2005;Schlichter et al.,2003;Wang et al.,2007).

The objectives of this study were1)to expand the industry knowledge of using ozone pretreatment with ceramic membranes;2)to elucidate the impact of coagulation and ozonation on organic foulants.Long-term pilot tests were conducted with a full-scale ceramic membrane module to investigate its performance on treating secondary WWTP ef?uent and demonstrate the impact of using pre-ozonation on ceramic membrane performance.

2.Materials and methods

2.1.Testing site

The Inland Empire Utilities Agency(IEUA)Recycling Plant No.5 (RP-5)served as the test site.This plant is located in Chino, California.The IEUA RP-5is a full-scale conventional wastewater treatment plant that treats an annual average?ow of 5.7?104m3/d(15mgd)of wastewater.The pilot testing site was located near the plant’s tertiary ef?uent structure where it had access to plant secondary ef?uent.The test site had suf?cient electrical power and drainage to conduct the testing program.

2.2.Source water

The feed water for the pilot testing was the secondary ef?uent from RP-5.Throughout the one year pilot testing program,the pH of the secondary ef?uent was consistently 6.8and temperature varied between21 C and30 C.Turbidity varied from0.4NTU to1.8NTU while the TOC values were observed between3.8and5.9mg/L.The alkalinity was approximately 140mg/L as CaCO3throughout the testing.Full nitri?cation has been consistently achieved in the secondary ef?uent,indicated by the low concentrations of nitrite(NO2-N<0.03mg/L)in the feed water to the membrane.The phosphorous concentration in the feed varied between0.3mg/L and4.2mg/L.

2.3.Ceramic membrane speci?cations

A full-scale monolith METAWATER ceramic membrane element was used.This single membrane element has dimensions of1.5m(59.1in.)in length and0.18m(7.1in.)in diameter,with a total surface area of25m2(269ft2).The ceramic membrane operates as a conventional pressurized membrane con?guration in a direct?ltration(dead-end)mode. Each ceramic membrane element utilizes2000inside/out channels(?ltration cells),in which raw water is?ltered through a thin ceramic membrane separation layer(0.1m m nominal pore size).The?ltrate out?ows from the element through water collection slits,via internal water collection cells.

2.4.Pilot system

A schematic of the ozone–ceramic micro?ltration pilot system is shown in Fig.1.The pilot system consists of three parts:pre-ozonation,coagulation/?occulation pretreatment and the ceramic membrane module with backwash facilities.

A pressurized ozone contact system(IOCS015-A24)was provided by Paci?c Ozone(Benicia,CA).Ozone was generated using an on board ozone generator,which compresses air into an oxygen separator unit,and then supplies oxygen to a corona discharge ozone module converting a percentage of the oxygen into ozone.The ozonated gas was then injected into a re-circulating water loop via a centrifugal pump driven, differential pressure,and vacuum inductor.The ozone concentration in the feed gas was controlled via a control panel ozone adjustment knob and a PID controller.The ozone concentration in the gas phase was measured with460M NEMA Ozone Monitor(Teledyne Instruments,San Diego,CA) for the feed gas and with460H NEMA Ozone Monitor(Tele-dyne Instruments,San Diego,CA)for the contactor off gas. The dissolved ozone in the contact tank was measured with an amperometric ozone sensor which was frequently cali-brated using Standard Method4500-O3B(Indigo Colorimetric) with a Hach Pocket Colorimeter.

Coagulation and?occulation was performed via inline direct coagulation using a static mixer positioned after the feed

w a t e r r e s e a r c h43(2009)2020–20282021

pump to evenly distribute the dosed coagulant in the feed pipe.The average retention time in the feed pipe was 40s when the membrane was operated at 100gfd.Poly aluminium chloride (PACl)was used as the coagulant as per manufacturer’s recommendation,as it helps to destabilize the micro-particles in the water and promote the aggregation of natural organic matters inside of the ?ow channels that can be effectively removed by high pressure backwash (Yonekawa et al.,2004).

The ceramic membrane pilot operated with a single module and has two primary phases of operation:?ltration and backwash.When in the ?ltration phase,feed water is pumped to the bottom of the membrane module and enters the membrane channels.Filtration is conducted in a dead-end con?guration.At the completion of each ?ltration phase,the membrane is hydraulically backwashed with portion of the ?ltrate.During the backwash,high pressure with a maximum value of 5bar (72.5psi)is applied for 20s.The high pressure loosens accumulated solids on the membrane surface,and is followed by an air blow process that ?ushes the solids from the membrane channels.Following the air blow,the feed side and ?ltrate side of the membrane are rinsed before resuming ?ltration.The backwash phase lasts a total of 3min.Acid or sodium hypochlorite can be added to the backwash tank when a chemical enhanced backwash (CEB)is required.CEB is conducted less frequently than normal backwash and involves soaking of the membranes in solutions of sodium hypochlorite or acid for a period of time (typically 15–30min).The CEB can enhance the backwash ef?ciency and extend the time interval between full chemical cleanings.

2.5.Flux calculation

The ?ltrate ?ux (J )was calculated as the ?ow of product water (Q p )divided by the surface area of the membrane (S ).Speci?c ?ux (J sp )referred to the ?ltrate ?ux (J )that has been normalized by the transmembrane pressure (TMP).Normalized speci?c ?ux (J sp /J sp,0)is calculated as the ratio of speci?c ?ux (J sp )to the initial speci?c ?ux at the beginning of the ?ltration (J sp,0).

Flux data was corrected to a standard temperature of 20 C using the following formula to account for the variation of water viscosity with temperature.J t eat 20 C T?h Q p ?1:024e20àT Ti

=S where

J t ??ltrate ?ux (gfd,L/(h m 2))

Q p ?measured ?ltrate ?ow (gpd,L/h)T ?measured temperature ( C)

S ?membrane surface area (ft 2,m 2).

2.6.High performance size exclusion chromatography (HPSEC)

The size distribution of organic materials present in water samples was characterized using a Shimadzu VP-series High Performance Liquid Chromatograph (Shimadzu Corp.,Kyoto,Japan)with ultraviolet absorbance detection from 200to 300nm.A size exclusion column with a molecular weight separation range of 2–80kDa (Protein-Pak 125,Waters,Mil-ford,MA)was operated at 25 C.The eluent was reagent-grade water buffered at pH 6.75with 2mM K 2HPO 4and 2mM KH 2PO 4;the ionic strength was adjusted to 0.1M with NaCl.The instrument was calibrated with polystyrene sulfonate standards (Polysciences,Inc,Warrington,PA)with nominal molecular weight values of 1.8,4.6,8,18,35,67kDa.Samples were pre?ltered using 0.22-m m syringe membrane ?lters.

Results and discussion

3.1.

Membrane performance with increasing ?ux

The METAWATER ceramic membrane features high ?ltration ?ux with extended backwash interval compared to traditional

Ozone Membrane

Fig.1–Schematic of the ceramic membrane pilot unit.

w a t e r r e s e a r c h 43(2009)2020–2028

2022

polymeric membranes.To demonstrate the ?ltration property of ceramic membranes,pilot testing initially focused on obtaining long-term membrane performance data at increasing ?ux conditions,while maintaining a constant backwash interval of 60min and coagulation dose of 3.5mg-Al/L.

Fig.2summarizes the long-term performance data of the METAWATER ceramic membrane treating secondary ef?uent under three different ?ux conditions.The membrane was continuously operated for 950h.An increase of trans-membrane pressure (TMP)was recorded over time,which yielded the speci?c ?ux decline indicating the membrane fouling.

The pilot testing began at a ?ux of 85L/m 2h (50gfd)to establish the baseline comparison with polymeric membranes typically operated at a similar ?ux and recovery for water reuse applications.Typical membrane operating conditions used in wastewater treatment (Metcalf &Eddy,Inc.,2003)are ?ux of 17–68L/m 2h (10–40gfd),recovery of 90–98%,backwash every 15–60min,and maintenance clean every 1–2days.Utilizing one Cl 2CEB every two days,no fouling was observed over the 200h of operation (approximately 9days).Both the transmembrane pressure and the speci?c ?ux remained constant at this condition,demonstrating the improved performance of ceramic membrane when treating secondary wastewater at ?uxes typical to commercial poly-meric membranes.

After 200h,the ?ux was increased to 170L/m 2h (100gfd),while the CEB frequency was maintained at once every two days.As shown in Fig.2,initial fouling was observed in the ?rst 48h after changing the condition.However,CEB (100mg/L Cl 2,30min soaking)was able to recover most of the speci?c ?ux except for the ?rst cycle.From hour 250to 650,the transmembrane pressure (TMP)was maintained between 0.28bar and 0.37bar (4.1–5.4psi)with no signi?cant

050(85)100(170)150(255)200(340)250(425)

Time (min)

F l u x @ 20°C - g f d (L m 2h )

(0.34)

10(0.69)

15(1.03)20(1.38)25(1.72)TMP - psi (bar)

Flux @ 20°C (L/m 2

TMP (psi)

F l u x @ 20°C (g f d )

085

170

255

340

TMP (bar)

Fig.3–Flux–TMP pro?le for ?ltration of secondary ef?uent with the METAWATER ceramic membrane (dotted line indicates the linear reference line for determining the critical

?ux).

T M P (p s i )

TMP (bar)

0.000.340.691.031.38

1.72

0200400600800

1000

Time (hrs)

S p e c i f i c F l u x @ 20°C (g f d /p s i )

Specific Flux @ 20°C (L/m 2h-bar)

02214426638841105

Fig.2–Effect of increasing ?ux on METAWATER ceramic membrane performance on secondary ef?uent.

w a t e r r e s e a r c h 43(2009)2020–2028

2023

irreversible fouling.For the450h of operation(approximately 18days),the ceramic membrane demonstrated stable performance at170L/m2h(100gfd),as the Cl2CEB every two days fully controlled the fouling tendency with TMP below 8psi.

At hour660,a CIP(3000mg/L Cl2with overnight soaking) was performed to restore the initial permeability of the membrane.Speci?c?ux increased to1105L/m2h bar(45gfd/ psi)after the cleaning.Flux was increased to255L/m2h (150gfd)and the CEB frequency was set to once every two days.As expected,the membrane experienced more fouling at high?ux.Moreover,fouling signi?cantly increased once TMP exceeded7.5psi at255L/m2h(150gfd).These results showed that the CEB with a2-day frequency was insuf?cient to stabilize the membrane performance at this high?ux and irreversible fouling was observed.

3.2.Filtrate water quality

Filtrate samples were collected throughout the pilot testing for water quality analysis and focused on evaluating the suspended solids removal capability of the ceramic membrane.Turbidity of the?ltrate was less than0.1NTU throughout the testing period,and the TOC values were between3.1and4.1mg/L(reduced from5mg/L measured in the feed).The TSS(total suspended solids)in the?ltrate were consistently below detection limits(10mg/L).Silt density index(SDI)measurements were also performed on the feed and?ltrate water to predict their particulate fouling potential on RO membranes.The SDI-15values were reduced from6.2 in the secondary ef?uent to less than1.4in the membrane ?ltrate throughout the testing.The ceramic MF?ltrate with such a low SDI value is suitable for sequential RO treatment.

3.3.Critical?ux testing

The critical?ux can be generally de?ned as the‘‘threshold’’permeate?ux below which fouling does not occur,and above which fouling becomes noticeable.This concept might be of use in understanding and improving the operation of membrane?ltration systems(Bacchin et al.,2006).During pilot testing,?ux stepping experiments were performed in30-min?ltration cycle to establish the critical?ux for the oper-ation.In each cycle,the?ux was kept constant while an increase in TMP was recorded if the membrane fouled.A normal backwash was performed between each cycle before stepping up to higher?ux.As shown in Fig.3,the TMP started to increase within the?ltration cycle at?ux of148L/m2h (87gfd,temperature corrected to20 C).The rate of TMP increase versus time,i.e.,the slope of TMP curve in Fig.3(a) started to increase dramatically once?ux exceeded255L/m2h (150gfd,temperature corrected to20 C).Fig.3(b)shows the ?ux–TMP gradient of the30-min?ltration cycle.The critical ?ux,by general de?nition,is the point where the?ux–TMP relationship becomes non-linear.As shown in Fig.3(b),if extrapolated from the?ux–TMP(averaged within the30-min cycle)pro?le,the relationship became non-linear at TMP of 122L/m2h(72gfd,at20 C).These results justify operating the ceramic membrane at a design?ux of170L/m2h(100gfd),as the slight fouling can be well controlled by normal and chemical enhanced backwashes.

3.4.Membrane performance with pre-ozonation

Following the evaluation of membrane performance on secondary ef?uent,optimization studies were conducted by utilizing ozone pretreatment before the ceramic membrane. With ozone pretreatment,it was expected that the membrane performance could be stabilized at an optimal coagulant dose lower than what was established in the baseline runs.A series of dosing conditions were investigated and the optimal combination was selected for the long-term performance evaluation,including varying the PACl doses between1and 3.5mg/L and the ozone dose between0and6mg/L.

3.4.1.Performance of Train1–?xed PACl(3.5mg/L)and increased ozone(0to6mg/L)doses

The membrane experienced less fouling with increased pre-ozonation and fouling was fully controlled at an ozone dose of 6mg/L,when the PACl dosed was maintained at3.5mg/L-Al. As shown in Fig.4,membrane fouling at each condition was evaluated via comparing the trend of normalized speci?c?ux decline(each data point was selected right after backwash).It shows that with3.5mg/L PACl and no ozone,the membrane speci?c?ux decreased to42%of the initial speci?c?ux after3 days of operation,and to29%after5days.Fouling of the membrane was reduced with6mg/L ozone and3.5mg/L PACl as the membrane speci?c?ux over six days was stabilized at 1231–1280L/m2h bar(50–52gfd/psi,adjusted to20 C),which was97–100%of the initial speci?c?ux.Since membrane fouling was fully controlled at this condition,the PACl dose was further optimized in Train2while the ozone dose was maintained at6mg/L.

3.4.2.Performance of Train2–?xed ozone(6mg/L)and decreased PACl(3.5to1mg/L)doses

The membrane continued to experience minimum fouling with6mg/L ozone and1mg/L PACl.As shown in Fig.5,when PACl dose was reduced from3.5mg/L to1mg/L and ozone remained at6mg/L,the membrane speci?c?ux was stabilized at1132–1280L/m2h bar(46–52gfd/psi,adjusted to20 C),

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Fig.4–Fouling of METAWATER membrane with3.5mg-Al/ L PACl and increasing ozone(0to6mg/L).

w a t e r r e s e a r c h43(2009)2020–2028 2024

which was 88–100%of the initial speci?c ?ux.The temporary speci?c ?ux declines at Day 0.5,Day 4.3and Day 5.3were probably caused by the variation of feed water quality,but the water quality was not measured on these speci?c times to quantify this variation.Because the membrane now relied on much less coagulant (i.e.,almost 1/3of the previous 3.5mg/L-Al)for stable operation,it was possibly more ‘‘vulnerable’’to water quality change.However,the speci?c ?ux decline was minimum (<12%)and membrane permeability was quickly recovered in less than 10h.Since fouling was still well controlled under this condition,the ozone dose was further reduced to 2.7mg/L in the sequential Train 3,while the PACl dose remained at 1mg/L.

3.4.3.Performance of Train 3–?xed PACl (1mg/L)and decreased ozone (6to 2.7mg/L)doses

The membrane experienced increased fouling with decreased pre-ozonation,when the PACl dose was maintained at 1mg/L-Al.Results showed that the membrane operation could be stabilized at an ozone dose as low as 3.5mg/L.Fig.6shows that different trends of speci?c ?ux decline were developed when the ozone dose was decreased stepwise from 6mg/L to 2.7mg/L.With 6mg/L ozone and 1mg/L PACl,almost no fouling was observed after 5days of operation.With 5mg/L ozone,the membrane speci?c ?ux decreased to about 84%of

the initial speci?c ?ux after 2days of operation,and stabilized at 82–88%of the initial speci?c ?ux (i.e.,1058–1132L/m 2h bar,or 43–46gfd/psi,adjusted to 20 C)in the next 3days.With 3.5mg/L ozone,the membrane speci?c ?ux further decreased to 76%of the initial speci?c ?ux after 2days,and stabilized at 70–82%of the initial speci?c ?ux (i.e.,36–43gfd/psi adjusted to 20 C)in the next 3days.When the ozone dose was further decreased to 2.7mg/L,however,the membrane started to foul rapidly.The membrane speci?c ?ux decreased to 42%of the initial speci?c ?ux after 3days,and to 36%after 5days.No signi?cant improvement of the membrane operation was observed compared to the baseline condition of 3.5mg/L PACl and no ozone.Therefore,the ozone dose was backed to 3.5mg/L,since it appeared to be a critical dose above which a stable membrane operation could be maintained.

Table 1summarizes the ozone residual in the contactor ef?uent measured for each ozone dose.The residual data shows that the ozone demand of the feed water,which is the ozone dose generating a minimum ozone residual,was approximately 3.3mg/L.This was close to the critical ozone dose of 3.5mg/L as identi?ed in the above section.Therefore,an interesting observation was made when relating the ozone residual to the membrane performance.It appears that the membrane operation could be stabilized for long-term oper-ation when ozone was applied above the demand.However,it is important to note that the critical ozone dose for membrane operation is related to both the feed water quality and the coagulation condition.

3.5.Long-term operation with pre-ozonation

Based on the optimization results,a combination of 4mg/L ozone and 1mg/L PACl was selected as the pretreatment dosing condition,under which membrane operation was stabilized for long-term evaluation.The ozone dose was slightly higher than the aforementioned critical ozone dose of 3.5mg/L to provide a measurable ozone residual,usually in the range of 0.4–0.6mg/L.

Fig.7summarizes the long-term performance data of the METAWATER ceramic membrane ?ltering secondary ef?uent with optimized pretreatment conditions.The membrane unit had been continuously operated for 4weeks,with a pretreat-ment dosing of 4mg/L ozone and 1mg/L PACl.The opera-tional condition of the membrane was kept at a ?ux of 170L/m 2h (100gfd)at 20 C,2-h backwash interval and no CEB.An increase of transmembrane pressure (TMP)was recorded over time,which yielded the speci?c ?ux decline indicating the membrane fouling.

0.0

0.20.40.60.81.0

Time (days)

N o r m a l i z e d S p e c i f i c F l u x

Fig.6–Fouling of METAWATER membrane with 1mg-Al/L PACl and decreasing ozone (6

to 2.7mg/L).

0.0

0.20.40.60.81.0

Time (days)

N o r m a l i z e d S p e c i f i c F l u x

Fig.5–Fouling of METAWATER membrane with 6mg/L ozone and decreasing PACl (3.5to 1mg-Al/L).

w a t e r r e s e a r c h 43(2009)2020–2028

2025

As shown in Fig.7,the ceramic membrane demonstrated stable performance through the ?rst three weeks of operation,as the TMP was well controlled in the range of 0.13–0.17bar (1.9–2.5psi)and the speci?c ?ux between 984and 1354L/m 2h bar (40and 55gfd/psi)at 20 C.The break of data after 400h of operation (numbered as ‘‘’’in Fig.7)was caused by the temporary loss of data logging.

In Week 4,a TMP spike and sudden speci?c ?ux drop was observed after 560h of operation (numbered as ‘‘’’in Fig.7)when a strong rain storm occurred at the testing site.It was discovered that the cover of the coagulant tank was blown away in the storm and the coagulant in the tank was diluted by the rain.As a result,the membrane rapidly fouled to 20psi in 24h when running with diluted coagulant.The next day,the operator restored the coagulant feed and resumed the membrane operation without any cleaning.Interestingly,the membrane ‘‘self-recovered’’with the restoration of 1mg/L PACl in the feed line.The TMP quickly recovered to 0.28bar (4psi)in 8h and further to 0.17bar (2.5psi)by the end of Week 4.Accordingly,the speci?c ?ux gradually restored to 984L/m 2h bar (40gfd/psi)at 20 C by the closure of the testing after 680h of operation.

Overall the membrane was able to show a stable perfor-mance through the long-term operation with 4mg/L ozone and 1mg/L PACl.Although testing was interrupted due to the loss of the coagulant in Week 4,the membrane quickly recovered itself without any cleaning after the restoration of coagulant feed.

3.6.Discussion of fouling mechanisms

To understand the effects of ozone and coagulation pretreat-ment on organic foulants,natural organic matters (NOMs)in four waters –raw,ozone treated,coagulation treated,and ozone followed by coagulation treated wastewaters –were characterized using HPSEC.Fig.8compares the molecular

0481216

20T M P (p s i )

0.000.280.560.841.12

1.40TMP (bar)

0102030405060

Time (hrs)

S p e c i f i c F l u x @ 20°C (g f d /p s i )

024649273898412301476

Specific Flux @ 20°C (L/m 2hr-bar)

Fig.7–Long-term operation of METAWATER ceramic membrane with 4mg/L ozone and 1mg-Al/L

PACl.

Fig.8–NOM characterization of raw,coagulation treated,ozone treated,and ozone followed by coagulation treated wastewaters by HPLC–SEC (Fig.(a)presents absorbance (milli-absorbance unit)at 254nm for MW range from 67kDa to 1.8kDa;Fig.(b)presents absorbance (milli-absorbance unit)at 205nm for MW range from 67kDa to 4.6kDa.)

w a t e r r e s e a r c h 43(2009)2020–2028

2026

weight(MW)distributions of organic materials after various levels of treatment.In the case of pretreatment with coagu-lation only,majority of the organic materials remains the same as in the raw(Fig.8(a));the most signi?cant reduction of NOM was observed in the colloidal portion that is greater than 67kDa(Fig.8(b)).Conversely,with ozone pretreatment,NOM of the entire MW range(from greater than67kDa to less than 1.8kDa)decreased as a result of oxidation.Particularly for the colloidal portion(greater than67kDa),the NOM reduction with ozone pretreatment was observed to a slightly greater extent compared to with coagulation only.Moreover, a synergistic effect was observed when ozone and coagulation were combined since the peak of the colloidal materials (greater than67kDa)was almost completely removed.Our study has con?rmed that ozone greatly reduced membrane fouling(Fig.4)through degradation of organic foulants, especially the colloidal portion,that have been reported as essential foulants to membranes(Howe and Clark,2002).

Table2also summarizes the water quality of raw secondary ef?uent,after ozone pretreatment,and?ltrate after ozone,coagulation and membrane.Samples were collected during the long-term performance evaluation when the average ozone residual was averagely0.44mg/L in the con-tactor ef?uent and not detected before the membrane.The reduction of UV absorbance was substantial throughout the ozonation process.Additionally,17–25%of the phosphate reduction was achieved by coagulation and membrane?ltra-tion,leaving the?nished water of an average of2.1mg/L phosphate-P.DOC concentration slightly increased through the ozone oxidation treatment.This indicated that large organics could have been degraded to lower MW organics and measured as DOC.The degradation of organic compounds is likely to lead to signi?cant membrane fouling reduction.

4.Conclusions

New generation ceramic membranes have been recently introduced which have unique advancements and properties over currently available polymeric membranes.Ceramic membranes are physically superior to polymeric membranes and more resistant to severe chemical environments,which allow them to operate successfully under more rigorous conditions,such as higher?ltration?ux,higher feed water recoveries,extended backwash intervals,low chemical cleaning requirements,and long membrane life without breakage.Pilot-scale testing on secondary wastewater ef?u-ents demonstrated that the ceramic membrane could operate with minimal fouling at?uxes as high as125gfd.However, pre-coagulation doses of up to4mg/L as PACl are required as well as the use of100mg/L free chlorine chemical enhanced backwashes to stabilize permeate?uxes.Signi?cant fouling was observed when the ceramic membrane was operated at 150gfd.Due to the ozone-resistant nature of the ceramic membranes,ozone pretreatment allows the ceramic membranes to operate at even further optimized conditions. Pilot test results showed that even when operated under rigorous conditions(i.e.,100gfd at20 C,120min backwash interval and no CEB),membrane fouling could be stabilized at a lower PACl dose of1mg/L and ozone dose as low as3.5mg/L. Interestingly,the optimal pre-ozone dose corresponded to the ozone demand of the source water,at which point a majority of high molecular weight NOM was degraded.Long-term studies performed by operating the ceramic membrane at 120gfd using1mg/L PACl and4mg/L ozone pretreatment demonstrated that over a4-week period,little to no fouling occurred and a stable TMP was maintained at3psi.The HPSEC analysis has demonstrated that ozone treatment is effective at degrading colloidal NOMs which are likely responsible for the majority of membrane fouling.

Acknowledgement

The authors would like to gratefully acknowledge the contri-butions of the following individuals:Patrick Shields,Chris Berch,Jeff Noelte,Randy Lee,and Mike Hoover from Inland Empire Utilities Agency for providing the pilot test site and operation support;Eric Bruce from MWH’s National Tech-nology Group for providing project assistance;David Ladner from University of Illinois at Urbana-Champaign for HPSEC analysis;and?nally the entire METAWATER global ceramic membrane team from Japan for providing a pilot ceramic and ozonation systems and?nancial support.

r e f e r e n c e s

Adham,S.S.,Jacangelo,J.G.,Laine,J.M.,1996.Characterization and cost of MF and UF plants.Journal AWWA88,22–31. Bacchin,P.,Aimar,P.,Field,R.W.,2006.Critical and sustainable ?uxes:theory,experiments and applications.Journal of

Membrane Science281(1–2),70–87.

Castro,K.,Zander,A.K.,1995.Membrane air-stripping effects of pretreatment.Journal AWWA87,50–61.

Clement,J.,2007.Ceramic membrane and coagulation for TOC removal.In:Proceeding of2007American Water Works

Association Membrane Technology Conference,Tampa,

Florida.

Crozes,G.,Anselme,C.,Mallevialle,J.,1993.Effect of adsorption of organic matter on fouling of ultra?ltration membranes.

Journal of Membrane Science84,61–77.

Hashino,M.,Mori,Y.,Fujii,Y.,Motoyama,N.,Kadokawa,N., Hoshikawa,H.,Nishijima,W.,Okada,M.,2000.Pilot plant

evaluation of an ozone-micro?ltration system for drinking water treatment.Water Science and Technology41(10–11), 17–23.

Howe,K.J.,Clark,M.M.,2002.Fouling of micro?ltration and ultra?ltration membranes by natural waters.Environmental Science and Technology36(16),3571–3576.

Karnik,B.S.,Davies,S.H.R.,Chen,K.C.,Jaglowski,D.R., Baumann,M.J.,Masten,S.J.,2005.Effects of ozonation on the

w a t e r r e s e a r c h43(2009)2020–20282027

permeate?ux of nanocrystalline ceramic membranes.Water Research39(4),728–734.

Lehman,S.G.,Adham,S.,Liu,L.,Fujiura,S.,Kanaya,S.,2007.

Performance of new generation ceramic membranes using coagulation and ozonation pretreatment.In:Proceeding of 2007American Water Works Association Membrane

Technology Conference,Tampa,Florida.

Lee,S.,Cho,J.,https://www.360docs.net/doc/ab16252852.html,parison of ceramic and polymeric membranes for natural organic matter(NOM)removal.

Desalination160(3),223–232.

Lee,S.,Lee,K.,Wan,W.M.,Choi,Y.,https://www.360docs.net/doc/ab16252852.html,parison of membrane permeability and a fouling mechanism by pre-ozonation followed by membrane?ltration and residual

ozone in membrane cells.Desalination178(1–3),287–294. Lou-Brugger,A.,Panglisch,S.,Hattori,K.,Yonekawa,H., Tomita,Y.,Gimbel,R.,2007.Open up new doors in water

treatment with ceramic membranes.In:Proceeding of2007 American Water Works Association Membrane Technology Conference,Tampa,Florida.

Park,Y.G.,2002.Effect of ozonation for reducing membrane-fouling in the UF membrane.Desalination147(1–3),43–48. Roorda,J.H.,Wortel,N.C.,Dalen,R.,2005.New process for treatment of organically fouled water:experiences with

WWTP ef?uent.Desalination178(1–3),141–148.Schlichter,B.,Mavrov,V.,Chmiel,H.,2003.Study of hybrid process combining ozonation and membrane?ltration–?ltration of model solutions.Desalination156(1–3),

257–265.

Seidel,A.,Elimelech,M.,2002.Coupling between chemical and physical interactions in natural organic matter(NOM),fouling of nano?ltration membranes:implications for fouling control.

Journal of Membrane Science203(1–2),245–255. Shanbhag,P.V.,Guha,A.K.,Sirkar,K.K.,1998.Membrane-based ozonation of organic compounds.Industrial&Engineering Chemistry Research37(11),4388–4398.

Shen,Z.S.,Semmens,M.J.,Collins,A.G.,1990.A novel approach to ozone–water mass transfer using hollow-?ber reactors.

Environmental Technology11(7),597–608. Tchobanoglous,G.,Darby,J.,Bourgeous,K.,McArdle,J.,Genest,P., Tylla,M.,1998.Ultra?ltration as an advanced tertiary

treatment process for municipal wastewater.Desalination119 (1–3),315–321.

Wang,X.,Wang,L.,Liu,Y.,Duan,W.,2007.Ozonation pretreatment for ultra?ltration of the secondary ef?uent.

Journal of Membrane Science287(2),187–191.

Yonekawa,H.,Tomita,Y.,Watanabe,Y.,2004.Behavior of micro-particles in monolith ceramic membrane?ltration with pre-coagulation.Water Science&Technology50(12),317–325.

w a t e r r e s e a r c h43(2009)2020–2028 2028