Pharmaceutical compounds; a new challenge for wastewater treatment plants

TRITA-LWR LIC Thesis 2037 ISSN 1650-8629

ISRN KTH/LWR/LIC 2037-SE ISBN 978-91-7178-565-7

P HARMACEUTICAL COMPOUNDS ;

A NEW CHALLENGE FOR WASTEWATER TREATMENT PLANTS

Maja Długołęcka

April 2007

Maja Długołęcka TRITA LWR LIC Thesis 20:37

ii

Dedicated to Kuba for unconditional support, optimism and for keeping the faith

iii

iii

Maja Długołęcka TRITA LWR LIC Thesis 20:37

iv

Sammanfattning

Analyser har genomförts vid Himmerfjärdens avloppsverk (285.000 PE anslutna) av 70 identifie- rade farmaceutiska preparat tillhörande olika terapeutiska grupper. Dessa organiska mikroförore- ningar i halter från µg l

till något ug l

påverkade inte reningsprocesserna vid avloppsverket.

Resultat från de analytiska studierna visade på en kontinuerlig tillförsel beräknad till 1.51 kg dygn

av de mikroorganiska föroreningarna till recipienten. Metoprolol, carbamazepine och naproxen valdes för specialstudier för att utvärdera effektiviteten av olika reningsmetoder. Mätning av syre- upptagningshastighet genomfördes för att bedöma inverkan på aktiviteten hos aktivt slam från fullskaleanläggningen i närvaro av valda farmaceutiska ämnen.

En membranbioreaktor (MBR) i halv-teknisk skala behandlade utgående avloppsvatten från full- skaleverket och inympades med slam från verkets aktivslamprocess. Denna behandlingsteknik gav dock endast en svag förbättring för avlägsnandet av inkommande mikroorganiska föroreningar.

Diskontinuerliga försök med aktivt slam från membranbioreaktorn och med användning av filtre- ring med granulerat aktivt kol medförde möjligheter att bedöma den sammantagna reningseffek- ten. Metoprolol och carbamazepine är i huvudsak inte biologiskt nedbrytbara men kan i de höga använda koncentrationerna leda till cellnedbrytning hos aktivt slam. Avlägsnandet av naproxen översteg 46% med den initialt tillförda mängden 3,3 mg NAP g

MLSS. Användning av filtrering genom granulerat aktivt kol är en effektiv teknik för att avlägsna farmaceutiska ämnen från be- handlat avloppsvatten.

Användning av det statististiska programmet MODDE7 medförde betydande tidsbesparingar vid studier av igensättning. Inverkan av igensättning , som är en starkt begränsande faktor vid drift av MBR, kunde minimeras genom att anpassa driftparametrar enligt anvisningar från Modde7 pro- grammet.

v

v

Maja Długołęcka TRITA LWR LIC Thesis 20:37

vi

Acknowledgements

This licentiate work was carried out within the Snowball project financially supported by SYVAB.

SYVAB, Swedish Institute (SI) and Lars Erik Lundbergs Foundation financed my scholarship.

I would like to express my gratitude to my supervisors, especially to Associate Professor Elżbieta Płaza for her support, guidance and valuable discussions. Many thanks also to my co-supervisor Dr Józef Trela for seconding me as a candidate in Snowball Project’s membership. I wish to thank my second co-supervisor Professor Bengt Hultman for his scientific support and guidance during my research work at KTH. I admire your knowledge and never-ending ideas. Appreciations go to Profes- sor Vladimir Cvetkovic, member of supervisory committee, for the big support during the final part of my work.

I have been lucky to meet Alf-Göran Dahlberg, affiliated senior faculty KTH and former Director of Himmerfjärden WWTP. I would like to thank you for opportunity to join your interesting and stimu- lating project with membrane technology. I appreciate your encouragement and support. Your enthu- siasm and love to science are unprecedented.

Moreover, I am deeply indebt to Jan Bosander, Expert Process Engineer at the Himmerfjärden WWTP for the enormous help during 2 years of work with pilot plant operation. I am very grateful to Lars Gunnarsson, Director of Himmerfjärden WWTP, for the support during difficult period of my studies and Malin Tuvesson for your great attitude and kindness. Special thanks goes to all Staff at SYVAB, especially to Herr Olsson, for creating unforgettable working environment and friendly atmosphere. I would like to thank you all for being my best ‘teachers’ of Swedish language. STORT TACK!

Beata, without your support and friendship I would not be here today. I admire your strength, ex- traordinary attitude and extreme patience in your research carrier. Thank you for being the most valuable source of advices along hard work at KTH and in everyday life. Luiza, I appreciate your support during my master thesis work at KTH and friendship that developed during our PhD studies.

I am would like to thank PhD student Grzegorz Cema for being an excellent discussion partner over the everyday ‘cup of coffee’.

I also want to express my gratitude to Professor Roger Thunvik and Associated Professor Ann- Catrine Norrström for help and support in finalizing my work. I am very grateful to Aira Saarelainen for a wonderful assistance in administrative matters.

Great appreciation goes to Peter Ohle and Ulrich Mende from Zenon GmbH for excellent survey over the membrane full scale facilities in Germany as well as for support in pilot plant operation.

Personal communication with Annette Achtabowski from Schering AG is appreciated. Thanks to Cajsa Wahlberg from Stockholm Vatten it was possible to perform identification of PPCPs at Him- merfjärden WWTP. I am also grateful to Professor Miksch and Dr Felis for access to the analytical laboratory at Silesian University of Technology (SUT). Thanks to Associated Professor Åsa Emmer I was able to perform analytical analysis at KTH.

Professor Krystyna Mędrzycka at the Gdańsk University of Technology for keeping the faith.

I wish to thank Monica Löwén for contribution in experimental set-up. I am also grateful to Jerzy Buczak for the help with computer problems. I appreciate the contribution in experimental work from Master students Małgosia Niedziołek and Wojtek Fronczyk.

I am grateful for financial support obtained from sponsors that enabled me the attendance at IWA Conference in Prague (SYVAB), Membrane Technology Conference in Wisła (SI), Ger- many/Belgium study visit (SYVAB), Workshop in Aachen (SYVAB), analytical training course at SUT (SYVAB), IWA World Water Congress in Beijing (SYVAB/Svenskt Vatten, Knut och Alice Wallenbergs Stiftelse).

Most of all, my family deserves the biggest appreciation. To my Parents – thank you for support, encouragement and LOVE. Kuba, you are the best brother in the world! My love, Norbert - thank you for patience, support, encouragement and for being there.

vii

vii

Maja Długołęcka TRITA LWR LIC Thesis 20:37

viii

TABLE OF CONTENT

Sammanfattning... v

Acknowledgements... vii

Abbreviations ... xi

List of appended papers... xiii

ABSTRACT... 1

1. INTRODUCTION...3

2. OBJECTIVES OF THE THESIS...4

3. PHARMACEUTICAL COMPOUNDS IN THE ENVIRONMENT ...5

3.1 Occurrence and distribution pathways... 5

3.2 Classification of target compounds... 6

3.3 Environmental issues ... 7

4. METHODOLOGY...9

4.1 Membrane bioreactor... 9

4.2 Experimental tests... 9

4.3 Analytical methodology... 11

4.4 Design of Experiments with Modde 7... 11

5. DISCUSSION OF RESULTS ... 12

5.1 Pharmaceutical compounds detected at Himmerfjärden WWTP (SYVAB)... 12

5.2 Effect of spiked pharmaceuticals on bacterial activity... 15

5.2.1 Metoprolol inhibitory effect on activated sludge... 15

5.2.2 Behaviour of metoprolol, naproxen and carbamazepine in activated sludge process... 16

5.3 MBR technology for pharmaceuticals removal ... 17

5.3.1 Improvement of WWTP effluent quality... 17

5.3.2 Evaluation of system with MBR and GAC... 18

5.3.3 Assessment of fouling phenomenon in ZeeWeed10™ ... 19

6. CONCLUSIONS ... 21

7. FUTURE RESEARCH... 23

8. REFERENCES... 25

ix

ix

Maja Długołęcka TRITA LWR LIC Thesis 20:37

x

Abbreviations

BOD – biochemical oxygen demand CAR - carbamazepine

COD – chemical oxygen demand DO – dissolved oxygen

EDC – endocrine disruption compound GAC – granular activated carbon MBR – membrane bioreactor MET – metoprolol

MLSS – mixed liquor suspended solids

MLVSS – mixed liquor volatile suspended solids NAP - naproxen

OUR – oxygen uptake rate

PPCP – Pharmaceutical and personal care products TN – total nitrogen

TP – total phosphorus

WFD - Water Framework Directive WWTP – wastewater treatment plant

xi

Maja Długołęcka TRITA LWR LIC Thesis 20:37

xii

List of appended papers

This thesis in based on the following papers, which are appended at the end of the thesis and referred in the text by their Roman numbers:

I. Długołęcka M., Dahlberg A.G. & Płaza E. (2006) ‘Low concentrations of high priority’ – Pharmaceutical and Personal Care Products (PPCPs); occurrence and removal at wastewater treatment plant. Vatten, 62 (2), 139-148.

II. Długołęcka M., Płaza E., Trela J., Dahlberg A.G. & Hultman B. (2006) Application of Oxygen Uptake Rate (OUR) tests for study of Metoprolol inhibitory effects on activated sludge. In: Proceedings of IWA World Water Congress & Exhibition, Beijing, September 10-14, China (CD-ROM).

III. Długołęcka M., Płaza E. & Dahlberg A.G. (2007) Two years of experience with semi- technical scale membrane bioreactor (MBR) Vatten, 63 (1), 69-78.

IV. Długołęcka M., Płaza E. & Hultman B. (2007) Behaviour of carbamazepine, metoprolol and naproxen in membrane bioreactor (MBR) activated sludge culture and after granular activated carbon (GAC) filtration. (Submitted to: Environmental Science and Technology).

xiii

Maja Długołęcka TRITA LWR LIC Thesis 20:37

xiv

A BSTRACT

Analytical analyses conducted at the Himmerfjärden WWTP (285.000 PE connected) identified 70 pharmaceutical compounds belonging to different therapeutic classes. Such organic micropol- lutants at low detected concentration range of µg - ng l

did not affect the treatment processes at WWTP. Results from analytical studies indicated continuous discharge of organic micropollut- ants to the surface water with a calculated load amounting to 1.51 kg day

. Metoprolol, car- bamazepine and naproxen were chosen for testing different removal methods. Oxygen Uptake Rate (OUR) tests were conducted to assess the bacterial activity of an activated sludge taken from a full scale aeration plant with the presence of selected target compounds.

A semi-technical scale membrane bioreactor ZeeWeed10™, treating final effluent from the Himmerfjärden WWTP (Sweden) was seeded with activated sludge from full scale biological stage. The membrane bioreactor (MBR) system placed after the final treatment appeared to be an insufficient technology for removal of residual amounts of organic micropollutants from WWTP effluents. Batch test studies with activated sludge taken from the membrane bioreactor and with application of granular activated carbon (GAC) filtration resulted in giving an overall assessment of removal efficiency. Metoprolol and carbamazepine tend to be resistant to the biodegradation process and in the dosed high concentration lead to bacterial cell decomposition in the activated sludge. Apparently, removal efficiency for naproxen exceeded the value of 46% with the spiked initial amount of 3.3 mg NAP g

MLSS. Application of the GAC filtration proved to be an effi- cient technique for removal of pharmaceutical compounds from treated wastewater.

Application of the statistical programme Modde7 was a time saving tool in studies of fouling occurrence. The effect of fouling phenomenon, which is a highly limiting factor for MBR per- formance, was minimised by adjusting the operational parameters as predicted by the Modde7 programme.

Key words: activated sludge, biodegradation, granular activated carbon (GAC), membrane biore- actor (MBR), pharmaceutical compounds, wastewater

1

Maja Długołęcka TRITA LWR LIC Thesis 20:37

2

1. I NTRODUCTION

Nowadays we are producing tonnes of drugs, antibiotics, pain-killers, food additives, per- sonal care products and cosmetics. In devel- oped countries with advanced technology, sectors like microbiology, genetic engineering and molecular biology are contributing in new trends in designing medicines, antibiotics and specific drug-cocktails. Consequently, an increase of drug prescription and consump- tion can be observed. Detergents, cosmetics and antibacterial agents are used in larger quantities in households, recently. Very little is known about degradation and removal mechanisms of compounds after being re- leased from the human body. Some of com- pounds are degraded in almost 100% in the organisms and some are excreted in the con- siderable amount. Most of those partly me- tabolised pharmaceuticals and chemicals that originate from human and animal excretion are collected by wastewater treatment plants and enter through sewage treatment facilities the water bodies. In aquatic matrices over 250 organic substances have been detected in the WWTP effluent water, originating from different medical classes, food additives, chemicals.

Obtaining of ‘good chemical condition’ of all surface water bodies by 2015 is one of the assumptions established by Water Frame- work Directive (article 4, Directive 2000/60/EG). This is applied for all sub- stances included in the list of priority pollut- ants according to environmental directives at European level. Additionally, in the area of concern are substances that appeared in the specific water bodies in significant quantities and are selected as important for a Member State. For the priority pollutants the new limits of discharge to the surface water will be implemented by 2021 in order to obtain the ‘good status’ of European water bodies.

The goal of the WFD policy will be to estab- lish new aspects within the treatment of domestic wastewater treatment. For organic micropollutants, pesticides, hormone disrupt- ers and pharmaceutical compounds there is not enough data available in respect to wastewater treatment plant (WWTP) effluent.

Moreover, there are no legal standards im- plemented for those compounds within the WFD. However, those compounds are ex- pected to have a negative influence on aquatic environment and ecology of water bodies. Therefore, selection of the hormone disrupters and medical substances has been completed in respect of their risk assessment for water quality.

Recent research confirmed the trend of envi- ronmental contamination of pharmaceutical and personal care products (PPCPs) in wastewater and surface water samples in many courtiers, like Sweden (Paxéus &

Schröder 1996; Paxéus 2004, Lindberg 2006, Bendz et al 2005), Norway (Vogelsang et al 2006), Germany (Ternes et al 1999a, Ternes et al 1999b), Italy (Castiglioni et al 2006, Paxéus 2004), Spain (Carballa et al 2004, Paxéus 2004), Austria (Kreuzinger et al 2004) and US (Boyd et al 2003, Koplin et al 2002). The concentration of the detected pharmaceuti- cally active compounds in the aquatic sam- ples is mostly in the range of few µg l

– ng l

1

and could be assumed to have a low value.

On the other hand, WWTPs that were mostly designed to reduce 70-80% of nutrients and organics quantity in wastewater are not suffi- cient for elimination of pharmaceutical com- pounds. Therefore, WWTP effluents ap- peared to be a permanent source of organic micropollutants and disposing those residual amounts to the water bodies.

To prevent the occurrence of pharmaceuti- cals in the surface water bodies and reduce the possible negative effects on non-target aquatic organisms, the technology for re- moval need to be investigated and imple- mented. There is a broad number of ad- vanced removal methods for biologically active organic substances that were recently investigated, such as membrane bioreactor system with adsorption on activated carbon (Bernhard et al 2006), ozonation (Andreozzi et al 2003, Ternes et al 2003), activated sludge process (Ternes et al 1999a), sludge digestion (Ternes et al 2005) and coagulation – floccu- lation process (Carballa et al 2003). With application of activated carbon technology and ozonation process a high removal effi- ciency was achieved for all classes of studied

3

Maja Długołęcka TRITA LWR LIC Thesis 20:37

compounds. However, oxidation products are not fully examined and the toxicology test should be implemented. According to the reported research, membrane bioreactor with activated sludge process seemed to be the promising removal technique for both nutri- ents and PPCPs but only with activated car- bon adsorption an excellent removal effi- ciency for PPCPs can be achieved. Still, there is a great need for finding the applicable removal technique for the micropollutants that are included into the group of priority substances. Closer cooperation between industry, researchers and regulatory agencies would be an effective approach to achieve more and more data, standardize the ecotoxi- cological testing and form a base set for pharmaceutical compounds that are hazard- ous for wildlife organisms.

2. O BJECTIVES OF THE THESIS

Wastewater, surface water and ground water samples have been analysed in most of European countries and results proved the presence of pharmaceutically active com- pound, personal care products hormones and other commonly used chemicals. Consider- able low range of pharmaceutical contamina- tion occurred in the WWTPs’ effluents worldwide; however, it was indicated that most of the compounds are not removed during the treatment.

There is a variety of advanced methods for elimination of residual amounts of micropol-

lutants from wastewater under investigation.

However, there is still not enough data avail- able concerning the studies on behaviour and elimination of pharmaceutically active com- pounds in sewage treatment systems.

The aim of the studies was to recognize the problem of pharmaceutical compounds oc- currence in the wastewater and study the possibilities of their removal in a membrane bioreactor (MBR) technology. The main goals were:

• To analyse the fate of pharmaceuticals in the aquatic environment and ecotoxicol- ogy risk assessment in the European countries

• To identify pharmaceutical compounds at the inlet and outlet of the WWTP and membrane bioreactor (MBR) pilot plant

• To evaluate of MBR technology as a system for pharmaceutical compounds removal from treated wastewater

• To verify of possible inhibitory effect on activated sludge process by dosed se- lected target compounds

• To test the bacterial culture from mem- brane bioreactor and propose a granular activated carbon column as system for pharmaceuticals removal

• To assess the MBR process performance for studies of fouling phenomenon oc- currence.

4

3. P HARMACEUTICAL COMPOUNDS IN THE ENVIRONMENT

3.1 Occurrence and distribution pathways In 2003 the annual production of widely used anti-inflammatory diclofenac exceeded the value of 179.8 tonnes, while carbamazepine, antiepileptic drug was sold in 268.1 tonnes in Europe. Only in Germany 180 tonnes of ibuprofen was sold annually (Ferrari et al 2003). A reported by Fick (2007) 1200 phar- maceutical substances were recognized on Swedish market and the average consump- tion amounted to 1000 tonnes year . Re- search studies reported by Gómez et al (2007) confirmed the increased consumption of ibuprofen, which resulted in a high con- centration in the WWTP influent samples.

Consumption of drugs is still increasing and its availability on the market has broadened recently. It is known that a

-1

significant amount of compounds is transported as

active substance via urine as unmetabolized pharmaceutically active compounds, parent- drug conjugates, bioactive metabolites, slow- dissolving medications and collected in

WWTP. Figure 1 illustrates the pathways of pharmaceutical compounds in aquatic envi- ronment.

In early eighties pharmaceutical compounds have been detected and analysed in environ- mental samples like river water (Watts et al 1983; Richardson & Bowron 1985), sewage effluent (Waggott, 1981), sediments (Björk- lund et al 1990; 1991) and portable water (Waggott 1981; Richardson & Bowron 1985).

Based on recent research, it can be stated that drugs and their metabolites entering water supplies and the food chain may pose a threat to the ecosystem and human health (Juberg 1999; Daughton & Ternes 1999). Investiga- tions conducted in Ontario, Canada and Louisiana, USA showed the presence of pharmaceuticals in various stages of drinking water treatment plant, sewage treatment plant effluent, surface water body (Boyd et al 2003), similar to the European investigations (Daughton & Ternes 1999; Ternes et al 1999a; Heberer 2002a, Heberer 2002b). The

both for European countries and in US.

Naproxen concentration found in Mississippi River, Lake Pontchartrain and Detroit River ranging from 22-104 ng l

similar occurrence pattern can be observed

-1

corresponds to the Fig. 1. PPCP/EDCs – occurrence, fate and transport in the environment (based on Daughton &

Ternes 1999; Halling-Sørensen et al 1998).

5

Maja Długołęcka TRITA LWR LIC Thesis 20:37

one observed in other research in Germany (Ternes et al 1999a).

Most human drugs end up in the municipal

r aspect of PPCPs occurrence should

3.2 Classification of target compounds wastewater treatment plants. In order to cts

calculate the discharge rates, the documenta- tion concerning flows and technical condi- tions of WWTPs and number of recipients should be taken into account. Considering the Great Stockholm region as a suitable example of dense population (1.5 million of people) with concentrated discharges of wastewater into three discharge points in the archipelago (estuary to Baltic Sea): Käppa- laverket with 534.000 people connected, Henriksdal WWTP with 695.000 people connected and Bromma WWTP (286.000 people connected). Analysis of effluent wastewater measured at three WWTPs proved high contents for sulphamethoxazole (max 0.13 µg l

), trimethoprim (max 0.47 µg l

) and metronidazole (max 0.08 µg l

).

However, only norfloxacin, ofloxacin and trimethoprim were detectable further in the waters of Lake Mälaren and the archipelago, with maximum concentration of 0.003, 0.006 and 0.009 µg l

(Medical Products Agency 2004).

Anothe

be mentioned while discussing the detection in the environmental samples and in the WWTP effluent collectors. The effect of attenuation of pharmaceutical compounds was observed in different sites after discharge from WWTPs along the river and Mediterra- nean coast. Fate of compounds in the surface water could be affected by degradation (bio- degradation, chemical degradation or photo- degradation) and distribution (sorption onto particles or volatilization). However, many compounds are not biodegradable and tend to be persistent in the aquatic environment (carbamazepine, caffeine, furosemide, oflox- acin, atenolol, sulfomethoxazole, ciproflox- acin) (Castiglioni et al 2006). The range of discharged concentration detected in the WWTP effluents amounts to µg – ng l

and could be considered as low, however, the estimation of loads discharged into surface

water calculated globally for the each Euro- pean country is enormous.

Pharmaceutical and personal care produ (PPCPs), cosmetics, fragrances and chemical additives detected in wastewater samples can be divided in different classes, depending on chemical properties, therapeutic classes, removal rates in treatment processes and level of occurrence in aquatic environment.

Among the pharmaceutical compounds anti- biotics, anti-inflammatory drugs, β-blockers, antineoplastics, antiepileptic drugs, diuretic, anxiolitic drugs, lipid regulators, antidepres- sants and antiasthmatic can be distinguished.

Concerning personal care products mostly musk fragrances and antibacterial agents occurred in aquatic environment. Analysed compounds reported in different investiga- tions were mostly selected in respect of the annual production statistics, consumption pattern and statistical prescription database available in the particular region or country.

Detailed analysis results reported in the litera- ture in different wastewater treatment plants, changeable climate conditions and a large variety of experimental trials could create the map of ‘pharmaceutical contamination’ in an aquatic environment in a global perspective.

One group of pharmaceutical compounds known as 'endocrine disruptor compounds' (EDCs) is nowadays a new area of concern as these hormonally active chemicals with pos- sible adversely can affect human health. It is noted that also adverse effect on wildlife species could be associated with concentra- tion of EDCs introduced to water bodies.

EDCs are defined by Environmental Protec- tion Agency as ' an exogenous agent that interfaces with the synthesis, secretion, transport, binding, action or elimination of natural hormones in the body that is responsible for the maintenance of homeostasis, repro- duction or behaviour' (EPA 1997). Not only hormones but also other chemicals that re- veal the disruption activity belong to this group (bisphenol A).

6

In order to select the compounds needed for further investigation due to the potential risk, different aspects were taken into considera- tion. In Italy, the amount of discharged PPCPs to the water bodies was the major selection criteria. The priority pollutants selected in Italy were identified as antibiotics, anti-inflammatory, diuretics, lipid regulators, cardiovascular, gastrointestinal, discharged to the surface water in amount ranging between 60 – 180 kg day

(Castiglioni et al 2006). In Netherlands the hormone disrupters and medical substances were recommended for selection to List of Priority substances. The extract from this list is presented in Table 1.

According to Castiglioni et al (2006), pharma- ceutical compounds could be divided into classes due to the different removal rates among the WWTPs; with higher degradation during the summer than winter, with the constant removal rates all the year and not

eliminated in the treatment processes. Major- ity of studied antibiotics were detected in two to five times higher concentration in different year periods, like ofloxacin with 373 and 84 mg day

per 1000 inhabitants in the winter and summer, respectively. This varying pharmaceutical loading express the pattern of usage of different pharmaceutical classes, where diuretics and beta-blockers are used constantly and consumption of antibiotics and anti-inflammatory drugs increases over the winter period.

It is impossible to implement the routine analysis and discharge limits for all detected compounds in WWTP effluents. Therefore,

the criteria for an adequate selection must be described carefully taking into account occur- rence frequency, range of detected concentra- tion, possible environmental risk assessment and removal efficiency among available treatment technologies. Chemicals that reveal disruption activity, antibiotics, antineoplastics and elimination-resistant compounds should be prioritized in respect of inclusion for List of Pollutants. Above all, the priority should consider the areas and regions where the wastewater is discharged directly to the sur- face water reservoir or source for drinking water production.

According to the report by Medical Products Agency (2004) the pharmaceuticals sub- stances were divided into classes as environ- mentally harmful (assessed by Swedish Chemicals Inspectorate), not environmentally harmful and where insufficient information concerning the risk assessment was provided

in the literature. Based on the information in the literature and delivered by Association of Swedish Chemical Industries Environmental Protection Sheet, diclofenac, 17α- etinylestradiol, ivermectin, ibuprofen, nore- thisterone, metoprolol, oxytetracycline, ty- losin and paracetamol were included to the list as environmentally harmful substances for aquatic organisms, while only terbutaline, furosemide and hydrochlorothiazide were found to be not harmful for environment.

Table 1. Selected compounds included in the List of Substances (de Jong et al 2005).

Concentration Compound Presence in WWTP effluent

min µg l^-1 max µg l^-1

17α-ethinyloestradiol + 0.00 <0.01 bisphenol A ++ 0.04 4.09

Hormone

disruptors Oestrone ++ 0.00 0.01 Ibuprofen +++ 0.12 0.76 Anhydro-erythromecine +++ 0.15 0.52

Sulfomethoxazole +++ 0.06 0.13 Carbamazepine +++ 0.33 1.00

Sotalol +++ 0.97 1.6 Medical

substances

Amidotrizoic acid +++ 0.23 1.2 + found in >5% of effluents < 50%, ++ > 50% analyzed effluents < 95%; +++ analyzed in > 95% of all effluents

3.3 Environmental issues

Small amounts of PPCPs accumulating over the years in water basins may cause harmful

7

Maja Długołęcka TRITA LWR LIC Thesis 20:37

effects not only on aquatic organisms. It also pointed out the arising problem of pharma- ceutical contamination and a possible impact on natural environment. Furthermore, an increasing tendency of the production and consumption of drugs is observed Gómez et al (2007). The concentration of natural and synthetic hormones is rather low, ranging around a few ng l

. However, this low range of concentration reported by Khanal et al (2006) has been recognised to disrupt the endocrine system of many species including humans by this estrogen hormones like 17ß- estradiol. Ternes et al (1999a) detected the natural hormones (17β-estradiol, estrone) and contraceptives (17α-ethinylestradiol) in the effluents of 16 WWTPs (mainly treatment of household discharges) in Germany and 10 in Canada. Analysis performed by Alcock et al (1998) indicate that natural estrogens were found in all analyzed effluent samples, in the range of 1.4-76 ng l

, and ethinyloestradiol in 3 out of 7 samples in the range of 0.2-7 ng l

. Those concentrations are at very low level;

however were found in biologically active form. It suggests that oestrogens were not biodegradable during the treatment process and re-activated in sewage system. Except the WWTP effluents, 139 streams in US have been analysed taking into account the con- tamination of wastewater organics including steroids and hormones. 57.3 µg l

was found as maximum total concentration of com- pounds that exhibit hormonal activity and disrupt the endocrine functions (Koplin et al 2002).

As presented in Table 1, except of hormones, mostly antibiotics are included into the list of priority substances. It is connected to resis- tance to bacterial activity and this attribute is not reduced during the treatment process in WWTP. Resistance to antimicrobial agents can depend on three factors: mutation in common genes, transfer of resistance gene among the microorganisms and/or en- hancement of development of resistant or- ganisms by increased pressures. What is

more, antibiotics were found to be toxic to algae with EC

of 12.5 mg l

and 1.6 mg l

for blue-green algae (Stuer-Lauridsen et al 2000).

There is growing interest in the genotoxi- cological effects connected to the genotoxins continuously introduced to the environment.

Analyzing the genotoxic potential, 800 wastewater samples from hospitals were collected. 13% of samples showed genotoxic activity, with the highest amount in the morning hours (Halling-Sørensen et al 1998).

Antineoplastics, drugs for treating cancer, were also found in the influent and effluent wastewater treatment samples and low re- moval rates can be observed. Those pharma- ceuticals by nature are toxic, behave as non- specific alkylating agents and therefore they could act as acute or (and) long term stress- ors for any species of living organism (Daughton & Ternes, 1999).

Antiepileptic drug, detected in more than 95% of all analysed WWTP effluents, is a compound of great concern. Behaviour of carbamazepine was observed on the labora- tory-scale treatment plants, soil passage and within groundwater. Detectable decrease of concentration obtained in groundwater sam- ples can be only explained due to dilution factor. It was confirmed that carbamazepine is an extremely persistent compound even during wastewater infiltration and in polish- ing lagoons (Tixier et al 2003). The behaviour of carbamazepine allows considering it as suitable biomarker of wastewater (Clara et al 2004). Additional experiments with spiked carbamazepine in laboratory sand column were artificially illustrating the track of wastewater effluent target compounds in the surface waters and irrigation fields. Not sur- prisingly, carbamazepine was not eliminated during this investigation, what could indicate and finally prove it persistence in groundwa- ter samples and surface water (Scheytt et al 2006).

8

4. M ETHODOLOGY

The membrane technology for removal of persistent organic pollutants residues was investigated at semi-technical scale at Him- merfjärden Wastewater Treatment Plant (SYVAB WWTP) (Grödinge, Sweden). In this chapter, a summary of the used experi- mental tools and procedures are presented.

Additional information and detailed descrip- tion are given in Papers II-IV. Comprehen- sive description of Modde 7 designing of experiments tool is introduced.

4.1 Membrane bioreactor

Membrane bioreactor pilot plant (MBR) ZeeWeed10™ was constructed by Zenon GmbH. It consisted of a main process tank (0.07 m

), a backwash tank (0.04 m

) and a permeation process pump and is presented in Figure 2a. The submerged membrane module with a surface area of 0.93 m

is located in a process tank and consists of polymeric hol- low fibers as illustrated in Figure 2b. In the semi-technical scale as well as the full-scale, hollow fibres are bound together into mod- ules using a unique dual-potting system. By applying a low vacuum to the inside of the hollow fibers, the fully oxidized water is filtered through the membranes. Meanwhile, mixed liquor and air are pumped continu- ously across each membrane module fiber bundle. The resulting flow across the mem- brane (cross-flow) continuously scours the membrane surface, preventing solids build up on the membrane surface. Figure 2c shows the cut-off of the single membrane fibre.

Membrane hollow fibres modules used in the full scale application are packed into larger cassettes, as presented in Figure 3.

Semi-technical pilot plant membrane bioreac- tor was connected and operated to the end of full-scale technological process at WWTP.

ZeeWeed10™ ultrafiltration membrane system as an experimental tool was supplied with the WWTP effluent and operated since

2004. Figure 4 illustrates the scheme of wastewater treatment processes and MBR location within the technological system. The regular membrane cleaning procedures al- lowed a process performance with average flux value of 35 l m

h

(Paper III).

4.2 Experimental tests

• Identification of pharmaceutical and personal care products in the influent and effluent samples collected at full-scale plant and MBR pilot plant (Paper IV)

• Verification of possible inhibitory effect on activated sludge was performed by measurement of bacterial respiratory ac- tivity with application of Oxygen Uptake Rate (OUR) test series. Those short term tests were conducted in BOD bottles of 150 ml working volume with activated sludge samples according to a specified procedure. Samples were collected from full-scale biological reactors and spiked with increasing concentrations of phar- maceutical compound, metoprolol (MET) (Paper II).

• Additional OUR tests were conducted occasionally for controlling the bacterial activity in MBR process performance (Paper III).

• To investigate the removal rates of pharmaceutical compounds: car- bamazepine (CAR), metoprolol (MET) and naproxen (NAP), the batch test series were conducted with additional applica- tion of column with granular activated carbon (GAC) (Paper IV).

• Assessment of fouling occurrence during the MBR operation performance was evaluated according to the calculation procedures (Paper III) and evaluated with application of Design of Experiments by Modde 7 program.

9

Maja Długołęcka TRITA LWR LIC Thesis 20:37

Fig. 2. MBR, ZeeWeed10™ semi-technical scale pilot plant (a) with submerged hollow fibre membrane module (b – membrane structure; c – fibre cut-off)(Zenon webpage).

Fig.3. Ultrafiltration membrane module cassette in the full scale application (a) and sub- merged in the biological reactor of WWTP (b)(Zenon webpage).

10

Fig. 4. Technological scheme of treatment process at Himmerfjärden WWTP (SYVAB) with placement of MBR pilot plant (sludge handling is not included).

4.3 Analytical methodology

• Application of HPLC-UV method for quantification of MET, CAR and NAP was developed and tested (Paper IV). The methodology was adapted for determina- tion of target compounds in wastewater samples collected during batch test per- formance within the concentration range of 0.5 – 10 mg l

.

• Analysis of WWTP composite influent and effluent samples were estimated with LC-MS-MS method developed and per- formed by Karlsson (2007) (Paper IV).

• Among analysis of wastewater samples at inlet and outlet to membrane bioreactor, MLSS, MLVSS, TN, TP, TOC, COD were performed once a week (Paper III).

4.4 Design of Experiments with Modde 7 Modde 7 program included in the SIMCA- P+ package is a statistical based programme for designing the experiments for industrial, developing and research applications. Range of parameters, factors, conditions and vari- ables influencing the tested technology, proc- ess or system is not limited. The robustness testing of method and products is possible.

The purpose of employing the Modde 7

program is screening for the most influential parameters, named factors, and its ranges;

and finding the optimum of process, method.

The experimental work concerned testing the efficiency of desired system or method and its optimization. Traditionally, it is performed by changing one variable at a time until the response in determined. Combining the re- sults form separate trails the optimum could be reached. However, it is probably ex- tremely difficult to achieve when there are some interaction between the influential factors within the process. Depending on the application, it is usually time consuming or/and an expensive approach. Modde 7 was the solution that constructed the set of ex- perimental trials with the simultaneous vari- ability of all selected factors at a time. The collected results, called responses, from the designed trials were tailored into the response contour plots illustrated as maps with the clear picture of optimal selection of ranges of parameters (Eriksson et al 2001).

The efficiency of the MBR pilot plant per- formance in respect of the fouling phenome- non occurrence was tested with application of Modde 7 program. Experimental trials were designed and tested with two membrane units delivered by Zenon GmbH (Paper III).

11

Maja Długołęcka TRITA LWR LIC Thesis 20:37

5. D ISCUSSION OF RESULTS

5.1 Pharmaceutical compounds detected at Himmerfjärden WWTP (SYVAB) Composite samples of influent and effluent to Himmerfjärden WWTP (SYVAB) were analysed. Overall, 71 pharmaceutical com- pounds have been detected within the range of 0.8 -9350 and 0.8 - 1900 ng l

in WWTP influent and effluent, respectively.

Some of pharmaceuticals like: febantel, praziquantel, pyrantel, ketoconazole, cetiriz- ine, loratadine, thioridazine, omeprazole, bromocriptine, risperidone, clozapine, fluni- trazepam, zolpidem, zopiclone were not detected in investigations reported in the literature studies (based on the available data).

Detailed data of the results from analytical survey is presented in Paper IV. Comparison of detected compounds in Himmerfjärden and two WWTPs in the Stockholm region (Bromma and Henriksdal) with literature findings is presented in Table 2. Concentra- tion of some antibiotics like ciprofloxacin, erythromycin and metronidazole were higher in the influent water to SYVAB WWTP in comparison to Bromma and Henriksdal.

Sulfomethoxazole, tetracycline and trimeto- prim contamination of raw wastewater slightly variated between WWTPs. It should be noticed that the average values of detected pharmaceutical compounds in this study is based on two sampling surveys.

Investigations conducted by Lindberg (2006) indicated that antibiotics compounds in five different wastewater treatment plants in Sweden showed relatively small variability in concentration between geographical differ- ences and seasonal periods. Additionally, in reported studies analysed sludge samples revealed the presence of norfloxacin and ciprofloxacin in all investigated WWTPs.

Active forms of those compounds remain unchanged during sludge handling in the range of 40% and 60%, respectively.

Diuretic substances, furosemide and hydro- chlorothiazide, were detected at investigated influent wastewater within the range of 1.3 - 2.6 µg l

and in effluent 1.0 – 1.9 µg l

. Tak-

ing into account the excretion mechanisms of hydrochlorothiazid, more than 50% of com- pounds are excreted by patients in the origi- nal form. Therefore a considerable amount could be detected in the wastewater samples of influent water. The low removal efficiency calculated for studied WWTP could explain the persistent character of those compounds.

As reported in studies from Italy, Germany, France and Spain the most commonly de- tected pharmaceutical compounds in waste- water samples are β-blockers and anti- inflammatory drugs (Paper I). Among the selected compounds in these investigations, naproxen and ketoprofen were detected with in the range of 2.7 – 3.6 µg l

and 0.08 – 1.5 µg l

, respectively. High removal rates for naproxen were observed in studies at SYVAB WWTP as well as at Bromma and Henriksdal WWTPs (63-94%). Acidic compounds such as ibuprofen and naproxen with low solid- liquid partition coefficient are not signifi- cantly removed during primary treatment and sedimentation step (Carballa et al 2004). Anti- inflammatory drugs are biodegradable mostly during biological treatment (Ternes et al 2005). Furthermore, atenolol (1.4 µg l

) metoprolol (1.1 µg l

), codeine (0.69 µg l

) in the effluent made a substantial part, while other compounds were detected in the range below 100 ng l

. Metoprolol, most com- monly used β-blocker analysed at highest concentration of 1.5 µg l

at influent to SYVAB was thereby eliminated with 30% of efficiency. Interestingly, removal efficiency for metoprolol compared to Bromma and Henriksdal amounted to 0% and only 4%.

Detection of hormones in the survey re- vealed a low contamination of those com- pounds in wastewater. Concentration of estriol exceeded the value of max 353 ng l

. Those concentrations are at very low level however, were found in biologically active form. It suggests that oestrogens were re- activated in sewage treatment. Studies per- formed by Suidan et al (2005) complemented the picture of sex hormones behavior along the treatment system. Surprisingly, it could be stated that significant amount of hormones were absorbed to the primary sludge and no detection in activated sludge was observed.

12

Summarizing, the reported wastewater sys- tem was able to remove sex hormones with efficiency of 95, 99, and 83% for estrone, estradiol and estriol, respectively. The elimi- nation of estrogens along the treatment facili- ties might be influenced by parameters such as microbial activity and rain events, although the long term study would be appropriate to confirm which factors are responsible for those changes (Ternes et al 1999a). Neverthe- less, the several hundreds of nanograms still remain in the effluent and are discharged.

The concentration range of ng l

in the SYVAB WWTP effluent tends to be consid- ered as extremely low. However, after sum- ming the values of every studied compound the concentration of 11.6 µg l

of PPCPs and ECDs detected in effluent was obtained.

Having in mind the average flow of each treatment plant as Henriksdal Treatment Plant 241.000 m

day

, Bromma Treatment Plant 123.000 m

day

and SYVAB 130 000 m

day

, the discharging treated wastewater contains of 1.83 kg day

, 1.06 kg day

and 1.51 kg day

, respectively.

Heberer & Feldmann (2005) carried out similar investigations in Germany. Municipal WWTP in Berlin-Ruhleben, that serves popu- lation of approximately 1 million people including various hospitals (12.060 hospital beds), is discharging 2.0 kg week

(105 kg annually) of carbamazepine, 4.4 kg week

(226 kg annually) of diclofenac to surface water.

On the other hand, after discharging the treated wastewater to the surface water, the

degradation in the water basins might take place and attenuation effect is observed.

River and WWTPs effluent concentrations of carbamazepine, clofibric acid, diclofenac, ibuprofen, ketoprofen and naproxen were calculated and presented by Tixier et al (2003) as average daily loading to Lake Greifensee amounted to 29.2, 2.85, 18.9, 6.07, 0.41 and 11.7 g day

, for each compound respectively.

Decreasing concentrations of target com- pound measured in the Lake Greifensee (as the pharmaceutical’s catchments basin) showed the attenuation effect and indicated the elimination processes as the sum of chemical degradation, sedimentation, direct and indirect photodegradation and biodegra- dation. As carbamazepine and clofibric acid were considered as persistent in the surface water, for diclofenac, ketoprofen and naproxen the direct photodegradation could be possible as elimination mechanism. Be- cause ibuprofen does not absorb the sunlight, sedimentation and biodegradation processes could be responsible for elimination in this case.

Those values of organic micropollutants occurrence are rather significant when calcu- lated as a daily or weekly loading. On the other hand in the surface water effect of attenuation can be observed and detection of pharmaceutical compounds has been ob- served in the range of ng l

(Boyd et al 2003).

However, the continuous exposure of aquatic organisms has taken into consideration in the surface water basins where discharged points from WWTP are located.

13

Maja Długołęcka TRITA LWR LIC Thesis 20:37

TT aa bb ll ee 22 .. LL ii ss tt oo ff dd ee tt ee cc tt ee dd pp hh aa rr mm aa cc ee uu tt ii cc aa ll cc oo mm pp oo uu nn dd ss aa tt SS YY VV AA BB WW WW TT PP ii nn cc oo mm pp aa rr ii ss oo nn ww ii tt hh oo tt hh ee rr WW WW TT PP ii nn SS tt oo cc kk hh oo ll mm (( cc oo nn cc ee nn tt rr aa tt ii oo nn ii nn nn gg ll

)) ..

14

5.2 Effect of spiked pharmaceuticals on bacterial activity

5.2.1 Metoprolol inhibitory effect on activated sludge

Pharmaceutical compound, metoprolol, a commonly used beta-blocker which is pro- duced in the Astra Zeneca situated 30 km north from Himmerfjärden WWTP (SYVAB), was found at the influent and effluent to the plant in concentration of 1.5 and 1.05 µg l

, respectively (Table 2). More- over, slight differences in concentrations between influent and effluent indicate that metoprolol is passing the technological sys- tem without any trace of degradation process.

This rather low amount of metoprolol does not affect the activated sludge process effi- ciency in the full scale biological reactors. In this low range of concentration the inhibitory effect of pharmaceutical compound should not be taken into account. However, the higher amounts of metoprolol content in the wastewater can affect the activated sludge process.

The dissolved oxygen (DO) measurements performed in the series of tests with addi- tional metoprolol dosage allowed the evalua- tion of OUR profiles of bacterial groups

present in the activated sludge process (Paper II). Purposely, the tested dosage of pharma- ceutical compounds was 10

-10

higher that the range detected in the treatment process.

Only by this extreme amount of added

metoprolol it was possible to obtain an in- hibitory effect on respiratory activity of nitri- fying and heterotrophic bacterial groups.

Additionally, estimation of the inhibitory concentration of metoprolol allowed the further selection target compound dosage in batch test performance with MBR and GAC filtration column (described in the next sec- tion). Interesting differences in behaviour and oxygen consumption by bacterial groups were achieved and are presented in Figure 5.

The stable oxygen consumption of nitrifying bacterial group (Fig. 5b) was not disturbed even with the higher, 416 mg MET g

MLVSS, tested dosage of metoprolol, and thereby the metabolism was not affected.

However, with the presence of 832 mg MET g

MLVSS some inhibition was observed. In comparison, the increasing oxygen consump- tion by heterotrophic bacteria observed with the initial dosages (Fig. 5a) indicated the influence of pharmaceutical compound on metabolism. Further higher metoprolol dos- age caused a drop of respiratory activity and probable inhibition took place. Figures 5a and 5b clearly illustrates the differences be- tween the behaviour of bacterial groups, indicating the higher sensitivity of heterotro- phic bacteria that originally are responsible

for organic compounds degradation and removal. However, the short time of expo- sure during test performance should be pointed out and necessity for long term test

O UR of Nitrifying bacteria

5 10 15 20 25 30 35 40

0,00 0,04 0,42 4,16 41,61 416,1 832,2 mg MET g^-1 MLVSS

mg O2 g-1 MLVSS h-1

O UR of Heterotrophic bacteria

5 10 15 20 25 30 35 40

0,00 0,04 0,42 4,16 41,61 416,1 832,2 mg MET g^-1 MLVSS

mg O2 g-1 MLVSS h-1

S1 S2 S3

b

S1 S2 S3

a

Fig. 5. Oxygen uptake rate profiles in correlation with metoprolol dosage (MET) calculated over the MLVSS concentration for heterotrophic (a) and nitrification (b) bacterial groups (S – test series).

15

Maja Długołęcka TRITA LWR LIC Thesis 20:37

are required for better understanding the metoprolol character and interactions.

5.2.2 Behaviour of metoprolol, naproxen and carbamazepine in activated sludge process Chemical conversions observed in activated sludge process have been found as the most essential during the biodegradation of organic compounds. Biological degradation rates differ between the target compounds and are influenced by the length of sludge age that was found as the major factor in this step.

Degree of transformation depends on the physical-chemical properties of organic com- pounds but also on medium source charac- teristics, easiness of access to micro- and macro-elements and to necessary substrates.

Additionally, temperature and oxygen-rich conditions have influence on the process performance effectiveness. Therefore, the biodegradation process is rather individual for each compound in every biological reac- tor depending on the characteristics of the sludge and need to be estimated experimen- tally (Ternes et al 2005).

Three target compounds chosen for investi- gation were selected taking into consideration degradation rates reported in the literature, behaviour in the treatment processes and occurrence in environmental samples.

Membrane bioreactor situated at the end of Himmerfjärden WWTP technological proc- ess was seeded with activated sludge from biological reactor and continuously supplied with treated wastewater. During batch tests series, activated sludge from MBR was spiked with target compounds in order to observe an effect of its addition. Developed analytical methodology with HPLC-UV allowed the determination of selected target compounds in the wastewater samples in the range of 0.1 – 10 mg l

. In order to estimate the critical loading of pharmaceuticals that can nega- tively affect the activated sludge process performance and bacterial activity, the sludge samples have been diluted in one of the conducted series and amounted to 1.8 and 3.1 g l

MLVSS, respectively.

Carbamazepine (CAR) known as popular antiepileptic drug that is unchangeable during the treatment processes was detected in most

of the WWTP effluents worldwide and tend to be persistent in surface waters. The char- acter of this target compound and its high resistance on degradation process was proved by the low removal rates obtained in these studies in the batch test experiment (8.6 - 17.6 %). In the studies conducted by Bern- hard et al 2006 and Ternes et al 2005, up to 13% of degradation was observed. Other authors classified this compound as a suitable biomarker in respect of its non-biodegradable character. Second compound selected for the investigation conducted in batch test series, metoprolol (MET), is most frequently pre- scribed beta-blocker, found in WWTP efflu- ents in Sweden (see Table 2). Insignificant degradation efficiency was reported in the literature up to 10%. However, 22.6 % of degradation achieved in the batch test was significantly higher in comparison to the full scale plant where only 4% (Henriksdal WWTP) was eliminated. An interesting ob- servation could be noticed at Bromma WWTP (Table 2), where higher concentra- tion of metoprolol was detected at the efflu- ent than in the influent wastewater. This could confirm that is not easily biodegradable among the treatment processes. For naproxen (NAP), an easily biodegradable non-steroidal anti-inflammatory drug, the biodegradation efficiency were obtained in the range of 0.2 – 46.6 % (Paper IV). Satis- factory (80%), degradation rates for naproxen in the conventional and membrane bioreac- tor activated sludge process were reported also by Ternes et al (2005). Additionally, 94%

of degradation in Henriksdal WWTP and 63% at Bromma WWTP were observed.

During batch test performance the additional measurements of DO, conductivity, TN, COD analysis were carried out in the begin- ning and end of each test in order to observe the changes in the medium composition and probable traces of degradation reactions.

With external addition (10 mg l

) of metoprolol or carbamazepine, a 0.52 - 1.19 mg l

increase of TN could occur due to degradation of pharmaceuticals that contain the nitrogen atom in the structure (metoprolol and carbamazepine, respec- tively). However, detected higher increment

16

in TN after batch tests performance was probably caused by activated sludge decom- position. The correlation between the degra- dation rates of pharmaceutical compounds and ΔTN is presented in Figure 6. It can be noticed that for higher carbamazepine con- centration in the batch test series (3.19 mg CAR g

MLSS), the greater increment in TN concentration was determined than for the test with 0.38 mg CAR g

MLSS and amounted to 16.4 mg. Taking into account low degradation rates in those two tests (re- spectively 12 and 18%) such a total nitrogen increase could only indicated the inhibitory effect caused by spiked carbamazepine. Addi- tionally, the degradation process was ob- served in the first four hours of test perform- ance, thereafter that only a slight change in degree of removal was obtained (Paper IV).

Its extremely high concentration, detected even after 24h of batch test, led to bacteria decomposition and nitrogen release.

Similarly, the addition of metoprolol at the level of 3.25 mg g

MLSS could be the rea- son of the increased TN concentration of 11- 12 mg and thereby the bacteria activity was inhibited. For easily biodegradable naproxen with attained degradation rates at 28.7 and 46.6 %, the significantly low TN production was observed. According to this theory, the biodegradable compound that was added in

the high concentration, after 24h of batch test performance did not cause the inhibition of bacterial activity that might lead to de-

composition of activated sludge. Further- more, the degree of NAP degradation was the highest among all studied compounds.

5.3 MBR technology for pharmaceuticals removal

5.3.1 Improvement of WWTP effluent quality Combination of activated sludge process and fine membrane filtration has been investi- gated as the promising replacement and up- grading of existing biological step in waste- water treatment plants. Full scale plants for municipal and industrial wastewater treat- ment with designed membrane bioreactor technology are already in operation in several countries in Europe, US and Asia (Stephen- son et al 2000; van der Graaf & Roorda 2000;

van der Roest et al 2002).

The ZeeWeed10™ semi-technical scale pilot plant was in operation as the polishing step in the end of WWTP technological line since 2004. Results of TOC, COD, TN, TP, MLSS and MLVSS analysis demonstrated the capa- bility of improvement the WWTP effluent quality with 99% for MLSS and 75% for TP (Paper III). In research reported by Rosen- berger et al (2002) during 535 days of experi- ence with MBR for aerobic treatment for municipal WWTP, the results of COD and nitrogen analysis revealed that the treatment performance of the membrane bioreactor

was better in comparison to conventional sewage works. COD reduction amounted to 95% at low feed to microorganisms F/M

3.25 MET 3.16 NAP

0.34 NAP

3.19 CAR 0.38 CAR

3.25 MET

0 5 10 15 20 25 30 35 40 45 50

0 5 10 15 20

increase in TN [mg l^-1]

% of degradation [MET, CAR, NAP]

Fig.6. Correlation

between a degrada- tion rates of target compounds and Δ TN in batch test series (pharmaceuti- cal concentration values are calculated as mg [CAR, NAP or MET] g-1 MLSS).

17

Maja Długołęcka TRITA LWR LIC Thesis 20:37

ratio was stable. After seeding the bioreactor with activated sludge, the nitrification was complete after 30d, however denitrification only after 100d. The increase of biomass concentration up to 18-20 g MLSS l

has been observed, but there was no decline in membrane performance noticed.

Membrane bioreactor process was investi- gated in these studies as the technology for removal of pharmaceuticals. Most of organic compounds are partly degraded and several are passing the treatment process of full scale plants without any change. As it can be no- ticed in Table 2, residual amounts of this organic material are continuously discharged with WWTPs effluent. Taking into considera- tion the result of analytical investigations (Paper IV), MBR was capable for efficient removal of only some compounds such as metronidazole, erythromycin, atenolol, ser- tralin, clozapine, losartan and codeine in comparison to the full scale plant. It could be caused by very long sludge age of MBR acti- vated sludge that was supplied only with treated wastewater containing not sufficient nutrient content for bacterial growth.

5.3.2 Evaluation of system with MBR and GAC

Experimental assessment of membrane tech- nology with additional filtration on granular

activated carbon (GAC) was investigated by application of batch test series (Paper IV).

Table 3 presents the comparison of removal

efficiency for MBR and combined MBR+GAC for studied compounds.

Activated sludge was spiked with significantly high concentration of target compounds for accessing the biodegradation not influenced by other indirect elimination mechanisms.

Furthermore, high concentration of pharma- ceutical dosage can exclude assumption of tolerance of the studied compound by bacte- rial metabolism. As it was discussed in chap- ter 3, detected pharmaceuticals were origi- nated from large variety of therapeutic classes with different physical-chemical properties and behaviour in the wastewater. Mixture of organic compounds with specially designed active functions cannot be re- moved/eliminated efficiently by a simple single stage technology. Therefore, the appli- cation of membrane bioreactor system fol- lowed by granular activated carbon filtration is proposed as the efficient technology for removal of pharmaceutically active com- pounds. This approach can combine biologi- cal degradation of easy degradable com- pounds by activated sludge process performed in MBR and elimination of ‘diffi- cult’ compounds (diclofenac, carbamazepine, metoprolol) on activated carbon. Addition- ally, fine membrane filtration can assure the effluent quality that no suspended solids

presented studies (see Table 3), removal efficiency for all studied compounds ex- ceeded 96%. For carbamazepine, which is known as extremely persistent in aquatic affect the GAC filtration efficiency. In the

Target compound Concentration

[mg g^-1 MLSS] % of removal MBR % of removal MBR + GAC

Table 3. Removal efficiency for metoprolol, carbamazepine and naproxen with application of MBR and GAC.

MET 6.11 7.4 99.5 MET 5.67 5.2 98.9 NAP 5.61 0.2 96.8 NAP 0.61 15.4 96.7 Series 1

[1.8 MLSS]

CAR 5.67 8.6 99.8 CAR 0.67 13.3 99.8 MET 3.25 22.6 100 MET 3.25 20.3 100 NAP 3.16 28.7 99.7 NAP 0.34 46.6 97.0 CAR 3.19 11.9 100 Series 2

[3.2 MLSS]

CAR 0.38 17.6 100

18

environment and resistant on conventional treatment technologies, the elimination val- ues amounted to 99.8 - 100%. Similar results were reported by Snyder et al (2007) in inves- tigations conducted with powdered and granular activated carbon where the endo- crine disrupting chemicals and pharmaceuti- cal compounds were removed with greater than 90% efficiency. The studies confirm that membrane technology with filtration on activated carbon assure the efficient removal of most compounds. However, several can still be detected in the effluent water.

5.3.3 Assessment of fouling phenomenon in Apart from the experiments with pharmaceu-

er, results concerning the prediction

Fig. 7. Correlation between predicted and observed results for estimation MBR flow parameters in 8h trial series.

ZeeWeed10™

tical compounds, the operation of the MBR as a technological process was studied.

Mainly, aspects were investigated concerning operational parameters estimation, adjust- ment of optimal conditions and eliminating the factors causing the reduction of process efficiency (Paper III). Fouling phenomenon that occurred during long-term membrane operation was mostly affecting the process efficiency. A tool, Design of experiment Modde 7, was employed in two experimental periods for testing the membrane capacity in respect of fouling phenomenon occurrence.

Firstly, with 8 h trials series the principal parameters for efficient and stable membrane operation were established. The correlation of observed and predicted responses gener- ated by Modde 7 is presented in Figure 7 with the satisfactory 0.97 correlation coeffi- cient.

Howev

of fouling rates were improved in the second series, where longer period (6 days trials)

appeared to be more sufficient for parameter estimation. Figures 8a and 8b illustrates the comparison of predicted values in both series according to the Modde7 model with fouling as a main factor. Correlation coefficient val- ues of observed and predicted values in two trial series were estimated for shorter trial and longer at 0.59 and 0.85, respectively. It was assumed that 6 days trials were more sufficient for optimal parameter assessment taking into consideration fouling occurrence.

By adjusting the operational MBR parameters at levels predicted by Modde 7, diminishing of fouling occurrence was observed in com- parison to 8 h trials.

0,42 0,44 0,46 0,48 0,50 0,52 0,54 0,56 0,58 0,60 0,62

0,42 0,44 0,46 0,48 0,50 0,52 0,54 0,56 0,58 0,60 0,62

Observed

Predicted 1

2

3

4

5

6

7

8

9

10

11 12

13 14

15 16 17 y= 1*x-2 ,802e- 007

R2= 0, 9718

Y=1*x-2.802e-007 R²=0.9718

flow

19