lable at ScienceDirect

Journal of Environmental Management 149 (2015) 77e84

Contents lists avai

Journal of Environmental Management

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

Selection of odour removal technologies in wastewater treatmentplants: A guideline based on Life Cycle Assessment

Carolina Alfonsín a, *, Raquel Lebrero b, Jos�e M. Estrada b, c, Raúl Mu~noz b,N.J.R. (Bart) Kraakman d, e, Gumersindo Feijoo a, Mª Teresa Moreira a

a Department of Chemical Engineering, Institute of Technology, University of Santiago de Compostela, 15782 Santiago de Compostela, Galicia, Spainb Department of Chemical Engineering and Environmental Technology, Escuela de Ingenierías Industriales, Sede Dr. Mergelina, University of Valladolid,Dr Mergelina s/n, 47011 Valladolid, Spainc School of Engineering, London South Bank University, UKd Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlandse CH2M Hill, Level 7, 9 Help Street, Chatswood, NSW 2067, Australia

a r t i c l e i n f o

Article history:Received 22 July 2014Received in revised form9 October 2014Accepted 13 October 2014Available online

Keywords:Activated carbonBiofiltrationChemical scrubbingLife cycle assessment (LCA)Odour abatementWastewater treatment plants (WWTPs)

* Corresponding author. Tel.: þ34 881 816 739.E-mail addresses: carolina.alfonsin@usc.es (C. Alfo

es (R. Lebrero), estradaj@lsbu.ac.uk (J.M. Estrada), mBrat.Kraakman@ch2m.com.au (N.J.R.(B. Kraakman(G. Feijoo), maite.moreira@usc.es (M.T. Moreira).

http://dx.doi.org/10.1016/j.jenvman.2014.10.0110301-4797/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

This paper aims at analysing the environmental benefits and impacts associated with the treatment ofmalodorous emissions from wastewater treatment plants (WWTPs). The life cycle assessment (LCA)methodology was applied to two biological treatments, namely biofilter (BF) and biotrickling filter (BTF),two physical/chemical alternatives, namely activated carbon tower (AC) and chemical scrubber (CS), anda hybrid combination of BTF þ AC. The assessment provided consistent guidelines for technology se-lection, not only based on removal efficiencies, but also on the environmental impact associated with thetreatment of emissions. The results showed that biological alternatives entailed the lowest impacts. Onthe contrary, the use of chemicals led to the highest impacts for CS. Energy use was the main contributorto the impact related to BF and BTF, whereas the production of glass fibre used as infrastructure materialplayed an important role in BTF impact. Production of NaClO entailed the highest burdens among thechemicals used in CS, representing ~90% of the impact associated to chemicals. The frequent replacementof packing material in AC was responsible for the highest environmental impacts, granular activatedcarbon (GAC) production and its final disposal representing more than 50% of the impact in most cat-egories. Finally, the assessment of BTF þ AC showed that the hybrid technology is less recommendablethan BF and BTF, but friendlier to the environment than physical/chemical treatments.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Wastewater treatment plants (WWTPs) are considered impor-tant sources of gaseous emissions, including green house gases(GHG) and odorants (Shaw and Koh, 2013). The malodorous emis-sions associated with treatment processes are considered one ofthe major concerns of exposed population living in surroundingareas of WWTPs.

In this context, the concentration of malodours in the air is oftenmonitored and controlled with the aim of complying with odourregulations while keeping a respectable public image of the emis-sion sources.

nsín), raquel.lebrero@iq.uva.utora@iq.uva.es (R. Mu~noz),), gumersindo.feijoo@usc.es

Odour abatement technologies have beenwidely investigated ascost-efficient and reliable alternatives for the mitigation of odournuisance (Revah and Morgan-Sagastume, 2005; Schlegelmilchet al., 2005). These technologies are commonly classified intophysical/chemical and biological techniques. Physical/chemicaltechnologies have been broadly implemented as a consequence oftheir rapid start-up, low empty bed residence time (EBRT) andconsolidated know-how and experience in design and operation.These techniques are often based on the transfer of odorants fromthe gas emission to either a solid (adsorption) or liquid (absorption)phase. These pollutants can be further transformed into by-products according to their reactivity with the chemicals used.However, in the last decades biological systems have beenincreasingly used due to their ability to efficiently treat malodorousemissions at lower operating costs (Schlegelmilch et al., 2005). Themain advantages of bioprocesses compared to their physical/chemical counterparts derive from their low generation of sec-ondary wastes and low demand of resources, such as chemicals or

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e8478

adsorbent media. On the other hand, biological processes oftenrequire larger EBRTs (2e120 s vs. 1e5 s) and associated footprintthan physical/chemical alternatives at similar odour removalefficiencies.

Technologies for odour treatment have beenwidely reviewed inthe literature in order to establish their optimal range of applicationand performance for the removal of volatile organic compounds(VOCs) and volatile inorganic compounds (VICs). Nowadays, thereare enough experimental evidences regarding the capability ofbiofilters and biotrickling filters to achieve significantly highremoval efficiencies of air pollutants at both trace levels (Lebreroet al., 2014, 2012) and industrial concentrations(Balasubramanian et al., 2012; Omri et al., 2013). Estrada et al.(2012) demonstrated that biological techniques were the mostcost-efficient alternatives with lower sensitivity to design param-eters and lower operating costs than physical/chemical treatmentsat typical odour concentrations emitted in WWTPs. In terms ofsustainability, Estrada et al. (2011) assessed the performance ofdifferent physical/chemical and biological odour abatement tech-nologies based on the IChemE Metrics methodology (IChemE,2002). The analysis focused on major environmental indicatorssuch as resource use, waste production and emission impacts aswell as process economics and social impact. This preliminarystudy showed as compared to physical/chemical technologies,biological treatments presented lower demand of energy, materialand chemicals and limited production of hazardous wastes. Thesesystems were the most favourable option despite the required highinitial investment costs when analysing investment costs andoperational costs over 20 years.

Although the aforementioned reports offered valuable informa-tion for the selection of the most suitable technology for the treat-ment of malodorous emissions, the IChemE Metrics methodology isonly focused on the target process and does not consider a holisticapproach for the assessment. For instance, despite water con-sumption or material use being considered as environmental in-dicators, the impact of their production processes is out of the scopeof the IChemE Metrics analysis. Nevertheless, the origin and man-agement of the components/consumables of the system under studycould also be major contributors to the environmental impact. Pre-vious research papers showed how a complete environmentalevaluation helps to select the most adequate treatment technology,considering not only removal efficiencies but also potential envi-ronmental impacts (Alfonsín et al., 2013; Hospido et al., 2012).

Life cycle assessment (LCA) methodology is a quantitative pro-cedure to assess the environmental burdens associated withproducts, processes and services, which is commonly used forenvironmental impact evaluation (Baumann and Tillman, 2004;ISO, 2006b). Although this methodology has been widely andsatisfactorily applied to technologies dealing with wastewatertreatment (Hospido et al., 2004; Rodriguez-Garcia et al., 2011),there is, to the best of our knowledge, only one application of LCA tothe abatement of gaseous emissions, focused on lab-scale biofilters(Alfonsín et al., 2013).

This study aimed at enriching the guidelines for the selection ofthe best technological alternative for the treatment of malodorousemissions under an environmental perspective. In this context, acomprehensive LCAwasperformed inorder todetermine theoverallenvironmental impacts of five full-scale odour abatement technol-ogies in a WWTP scenario: biofilter (BF), biotrickling filter (BTF),activated carbon filter (AC), chemical scrubber (CS) and a hybridtechnology consisting of a BTF coupled with AC (BTF þ AC). The re-sults obtained for each technology were compared to the environ-mental impact derived from the direct discharge of the malodorousemission without treatment, named non treatment scenario (NT).

2. Materials and methods

The ISO14040 (ISO, 2006a) standarddetermines fourbasic stagesfor LCA studies: goal and scope definition, life cycle inventory (LCI),life cycle impact assessment (LCIA) and finally, interpretation ofresults (Baumann and Tillman, 2004). Goal and scope definitionconstitutes the first phase of an LCA and aims at defining theboundaries of the study and the quality of data used. A functionalunit (FU), which represents the function of the system under study,must be also established in this phase. Then, LCI is performed,whichinvolves data collection and interpretation of inputs and outputs.The allocation procedure is also conducted during the LCI phase,which consists on distributing input and output flows among theproducts of the process. LCIA represents the third phase and itspurpose is to convert LCI data into potential impacts associated toproducts and processes. LCIA includes two mandatory steps (i.e.classification and characterization) and other optional elements,such as normalization and weighting. Finally, the interpretation ofthe results allows identifying the hot spots of the process as well asrecommending options to reduce the environmental burdens.

2.1. Goal and scope definition

This study was conducted to evaluate the environmental im-pacts associated with the performance of five of the mostcommonly implemented odour abatement technologies inWWTPs.The functional unit chosen in this study was 1 m3 of treated air,which is consistent with the approach used in a previous studyevaluating lab-scale gas-phase bioreactors (Alfonsín et al., 2013).

2.2. System boundaries

The system boundaries determine the units and elements of theprocess included in the analysis (ISO, 2006a). The assessment hereconducted considered the incoming polluted emissions, which aredirectly discharged to the atmosphere in the NTscenario, the outputtreated emission, material used in the infrastructure, consumables(e.g. packing material), potable or secondary plant effluent water,energy use, chemicals and transportation and final disposal of allwastes. Water input is required in all systems except in AC andpotable water instead of WWTP secondary effluent water is used inCS. The secondary effluent water of the WWTP is considered aproduct related towastewater treatment andno impact is attributedwhen used for irrigating the packing material of the biofilter or thebiotrickling filter. The leachate collected in each technology (exceptin AC) was not considered in the analysis since it is returned to theWWTPheadworkswith anegligibleflow in comparisonwith thenetflow treated in the WWTP. The transportation from themanufacturing industry to the odour treatment facility of consum-ables like packing materials and chemicals was included.

2.3. Description of odour treatment technologies

Five odour abatement technologies were assessed: two biolog-ical (BF and BTF), two physical/chemical (CS and AC) and a hybridtechnology (BTF þ AC). The model malodorous stream of50000 m3h�1, representing a typical emission from WWTPs,included 30 different VOCs (Zarra et al., 2008), methanethiol(1.97 mg m�3) and hydrogen sulphide (20.9 mg m�3) (Table 1). Theodorants were classified into groups of high, medium and low hy-drophobicity, which largely determine their removal efficiencies inthe technologies evaluated (Estrada et al., 2011). A typical release ofCO2 of 0.75 g CeCO2 (g-Coxidized)�1 due to themicrobial oxidation ofodorants and a biomass yield of 0.25 g C-biomass (g-Coxidized)�1

were considered. Water was supplied to the BF, BTF and CS in order

Table 1Summary of the inventory data for the non treatment scenario and for the five abatement technologies evaluated. The units are referred to the FU: 1 m3 of treated air.

NT BF BTF AC CS BTF þ AC

Discharge of air1-Ethyl-2-methylbenzene (mg) 1.31 � 10�2 6.55 � 10�4 1.31 � 10�3 2.62 � 10�4 1.31 � 10�3 2.62 � 10�5

1,2,4-Trimethylbenzene (mg) 1.06 � 10�2 5.28 � 10�4 1.06 � 10�3 2.11 � 10�4 1.06 � 10�3 2.11 � 10�5

1,3,5-Trimethylbenzene (mg) 8.36 � 10�3 4.18 � 10�4 8.36 � 10�4 1.67 � 10�4 8.36 � 10�4 1.67 � 10�5

2-Butoxyethanol (mg) 9.63 � 10�2 9.63 � 10�4 9.63 � 10�4 9.63 � 10�3 9.63 � 10�4 9.63 � 10�5

a-Pinene (mg) 1.31 � 10�2 3.29 � 10�3 6.57 � 10�3 1.31 � 10�5 6.57 � 10�3 6.57 � 10�6

Acetic acid (mg) 1.05 � 10�1 1.05 � 10�3 1.05 � 10�3 1.05 � 10�2 1.05 � 10�3 1.05 � 10�4

Acetone (mg) 4.62 � 10�1 4.62 � 10�3 4.62 � 10�3 4.62 � 10�2 4.62 � 10�3 4.62 � 10�4

Acetophenone (mg) 5.88 � 10�1 5.88 � 10�3 5.88 � 10�3 5.88 � 10�2 5.88 � 10�3 5.88 � 10�4

Benzaldehyde (mg) 6.70 � 10�2 6.70 � 10�4 6.70 � 10�4 6.70 � 10�3 6.70 � 10�4 6.70 � 10�5

Benzene (mg) 2.19 � 10�2 1.10 � 10�3 2.19 � 10�3 4.38 � 10�4 2.19 � 10�3 4.38 � 10�5

Butanone (mg) 4.54 4.54 � 10�2 4.54 � 10�2 4.54 � 10�1 4.54 � 10�2 4.54 � 10�3

Butyric acid (mg) 2.09 � 10�2 2.09 � 10�4 2.09 � 10�4 2.09 � 10�3 2.09 � 10�4 2.09 � 10�5

D-limonene (mg) 2.30 � 10�2 5.76 � 10�3 1.15 � 10�2 2.30 � 10�5 1.15 � 10�2 1.15 � 10�5

Decane (mg) 7.08 � 10�3 1.77 � 10�3 3.54 � 10�3 7.08 � 10�6 3.54 � 10�3 3.54 � 10�6

Dimethyl disulfide (mg) 2.13 � 10�1 1.06 � 10�2 2.13 � 10�2 4.25 � 10�3 2.13 � 10�2 4.25 � 10�4

Dodecane (mg) 3.41 � 10�3 8.53 � 10�4 1.71 � 10�3 3.41 � 10�6 1.71 � 10�3 1.71 � 10�6

Ethylbenzene (mg) 1.47 � 10�2 7.36 � 10�4 1.47 � 10�3 2.94 � 10�4 1.47 � 10�3 2.94 � 10�5

Hydrogen sulfide (mg) 20.91 2.09 � 10�1 2.09 � 10�1 2.09 � 10�1 2.09 � 10�1 2.09 � 10�1

Limonene (mg) 1.15 � 10�1 2.87 � 10�2 5.73 � 10�2 1.15 � 10�4 5.73 � 10�2 5.73 � 10�5

m-Xylene (mg) 1.77 � 10�2 8.84 � 10�4 1.77 � 10�3 3.53 � 10�4 1.77 � 10�3 3.53 � 10�5

Methanethiol (mg) 1.97 1.97 � 10�2 1.97 � 10�2 1.97 � 10�1 1.97 � 10�2 1.97 � 10�3

Methylcyclohexane (mg) 1.26 � 10�2 3.14 � 10�3 6.29 � 10�3 1.26 � 10�5 6.29 � 10�3 6.29 � 10�6

Nonane (mg) 1.07 � 10�2 2.68 � 10�3 5.36 � 10�3 1.07 � 10�5 5.36 � 10�3 5.36 � 10�6

o-Xylene (mg) 1.77 � 10�2 8.84 � 10�4 1.77 � 10�3 3.53 � 10�4 1.77 � 10�3 3.53 � 10�5

Octane (mg) 1.07 � 10�2 2.68 � 10�3 5.36 � 10�3 1.07 � 10�5 5.36 � 10�3 5.36 � 10�6

p-Xylene (mg) 1.77 � 10�2 8.84 � 10�4 1.77 � 10�3 3.53 � 10�4 1.77 � 10�3 3.53 � 10�5

Propanoic acid (mg) 1.13 � 10�2 1.13 � 10�4 1.13 � 10�4 1.13 � 10�3 1.13 � 10�4 1.13 � 10�5

Tetrachloroethylene (mg) 1.07 � 10�2 2.66 � 10�3 5.33 � 10�3 1.07 � 10�5 5.33 � 10�3 5.33 � 10�6

Tetradecane (mg) 4.76 � 10�3 1.19 � 10�3 2.38 � 10�3 4.76 � 10�6 2.38 � 10�3 2.38 � 10�6

Toluene (mg) 5.09 � 10�1 2.55 � 10�2 5.09 � 10�2 1.02 � 10�2 5.09 � 10�2 1.02 � 10�3

Tridecane (mg) 4.04 � 10�3 1.01 � 10�3 2.02 � 10�3 4.04 � 10�6 2.02 � 10�3 2.02 � 10�6

Undecane (mg) 5.37 � 10�3 1.34 � 10�3 2.69 � 10�3 5.37 � 10�6 2.69 � 10�3 2.69 � 10�6

Carbon dioxide (mg) 0 14.19 13.97 0 18.63 13.97

Infrastructure materialConcrete (m3) e 9.33 � 10�9 e e e e

Glass fibre (kg) e e 1.48 � 10�5 3.23 � 10�5 2.13 � 10�5 4.83 � 10�5

Packing materialCompost (kg) e 1.31 � 10�4 e e e e

Perlite (kg) e 4.36 � 10�5 e e e e

Polyurethane foam (kg) e e 4.76 � 10�6 e e 2.85 � 10�6

Activated carbon (kg) e e e 8.18 � 10�5 e 2.05 � 10�5

Intalox saddles (kg) e e e e 1.27 � 10�6 e

ChemicalsNaOH solution 50% (kg) e e e e 9.99 � 10�5 e

NaClO solution 25% (kg) e e e e 6.35 � 10�4 e

HCl solution 5% (kg) e e e e 1.82 � 10�5 e

WaterFrom WWTP secondary effluent e 2.68 � 10�2 5.23 � 10�2 e e 5.23 � 10�2

Potable water e e e e 1.25 � 10�2 e

ElectricityFrom the grid (kWh) e 3.95 � 10�4 2.40 � 10�4 6.80 � 10�4 5.23 � 10�4 9.00 � 10�4

TransportLorry 3.5e7.5 t (kg � km) e 3.72 � 10�2 3.52 � 10�3 6.05 � 10�2 2.74 � 10�4 6.27 � 10�2

Final disposalComposting (kg) e 1.74 � 10�4 e e e e

Landfill deposits (kg) e e 4.76 � 10�6 8.18 � 10�5 8.25 � 10�6 2.33 � 10�5

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e84 79

to maintain the moisture conditions required and to compensatefor evaporation losses. The pressure drop associated with eachtechnology determined the energy required for gas circulation ac-cording to the Eq. (1):

E ¼ Qgas$DP$BE�1 (1)

where E is the energy needed (kWh), Qgas is the gas flow (m3s�1),DP is the pressure drop in the reactor (kPa) and BE is the blower

efficiency, inwhich case a typical value of 0.7 was assumed (Estradaet al., 2011).

2.3.1. Biofilter (BF)This configuration consisted of a reactor packed with a mixture

of organic and inorganic materials: 75% compost and 25% perlite (v/v) (Lebrero et al., 2010). The BF was built in standard concrete witha volume of 694 m3 (1 m height) and operated at an EBRT of 50 s.Irrigation water (129 kg h�1) from the secondary effluent of theWWTP was used. A standard pressure drop of 1000 Pa was

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e8480

considered. The malodorous air was moistened in a humidifierprior to biofiltration with additional water (1211 kg h�1). Theleachate produced from the BF and the excess water from the hu-midifier were returned to the WWTP headworks for furthertreatment. The mixture of compost and perlite was replaced after 2years of operation and processed in a composting facility, since itscharacteristics are similar to those of an average organic waste likesludge.

2.3.2. Biotrickling filter (BTF)A two-stage biotrickling filter of 208 m3 with a total EBRT of 15 s

was selected. The glass fibre BTF was packed with inert plasticsupports, which exhibited a density of 100 kg m�3 and a totalpressure drop of 500 Pa (Dorado et al., 2009). In the first stage, thepH was maintained at 2 for the efficient removal of H2S and in asecond stage a neutral pH was kept to favour the removal of theother odorants. Water was supplied by spray nozzles and recircu-lated at a rate of 7.2 L m�3 min�1 using a renewal ratio of 2.5 L ofwater added per gram of H2S removed. Landfills were selected asthe disposal scenario for the packed-bed after 10 years of lifespan.

2.3.3. Chemical scrubber (CS)A two-stage NaOHeNaClO scrubber (2 m) packed with Intalox

Saddles (density of 100 kg m�3) was considered, operated with atotal pressure drop of 500 Pa. The glass fibre reactor was operatedat an EBRT of 4 s (2 s per stage), which entailed a reactor volume of55.6 m3. The liquid phase was recirculated at a rate of180 L m�3 min�1 via spray nozzles (with a pressure drop of 50 kPa)(Gabriel and Deshusses, 2004) while 15040 L of additional waterwere added daily. The pH was maintained at 10e12 in the firststage, whereas the second stagewas operated at a slightly lower pH(8.5e9.5). NaClO (25% w/w) at a flow rate of 28.85 L h�1 and NaOH(50% w/w) at a flow rate of 3.26 L h�1 were continuously supplied.The packing material was washed once a year with HCl (5% w/w) inorder to avoid the accumulation of precipitates in the medium. Thepacking material was disposed of in a landfill after 10 years ofoperation.

Table 2Summary of data sources considered in this study.

Inventory inputs Data sources

Energy Electricity (Spanishelectricity profile)

Ecoinvent database (Dones et al.,2007)

Infrastructurematerial

Concrete Ecoinvent database (Kellenbergeret al., 2007)

Glass fibre Ecoinvent database (Kellenbergeret al., 2007)

Packingmaterial

Compost Ecoinvent database (Nemecekand K€agi, 2007)

Expanded perlite Ecoinvent database (Kellenbergeret al., 2007)

2.3.4. Activated carbon filter (AC)Two glass fibre towers of 42 m3 packed with granular impreg-

nated activated carbon (density of 430 kg m�3) and operating at anEBRT of 3 s and a pressure drop of 1750 Pa were used as a modeladsorption filter. One filter was in operation while the second onewas in stand-by for the periodic renewal of the activated carbonevery 6 months, which was later disposed of in landfills. Regener-ation of the activated carbon was not considered according to thecurrent operation of most full scale facilities (Estrada et al., 2011,2012).

Granular activated carbon Mu~noz (2006)Polyurethane, rigid foam Ecoinvent database (Hischier,

2007)PVC Ecoinvent database (Hischier,

2007)Chemicals Sodium hydroxide (50%) Ecoinvent database (Althaus

et al., 2007)Sodium hypochlorite (15%) Ecoinvent database (Althaus

et al., 2007)Hydrochloric acid (30%) Ecoinvent database (Althaus

et al., 2007)Water Tap water Ecoinvent database (Althaus

et al., 2007)Transport Truck 3.5e7.5 t, EURO 4 Ecoinvent database (Spielmann

et al., 2007)Waste

treatmentLandfill deposits Ecoinvent database (Doka, 2007)

2.3.5. Biotrickling filter coupled with an activated carbon tower(BTF þ AC)

The hybrid technology consisted of a biotrickling filter coupledwith an activated carbon unit. The biotrickling filter, packed with2 m of inert plastic support (density of 100 kg m�3 and pressuredrop of 500 Pa), was operated at an EBRTof 9 s. The 42 m3 AC towerwas operated at an EBRT of 3 s and pressure drop of 1750 Pa. Themain difference between the AC coupled to a biotrickling filter andthe AC operated stand-alone was the lifetime of the activated car-bon in the combined configuration: 2 years, due to the lowerodorant concentrations in the inlet stream. The packing materialswere identical to the previous configurations and consequently,landfills were considered for disposal.

2.4. Life cycle inventory, data quality and simplifications

The LCI included the emission data reported by Zarra et al.(2008) as well as theoretical and literature data for each technol-ogy concerning the odorant concentrations in the outlet emission,infrastructure material, packing bedmaterial, energy requirements,chemicals use, cleaning agents, water consumption, transport dis-tances and final disposal of wastes (Table 1).

The inventory data for auxiliary processes were taken from theEcoinvent database and completed with recent bibliography(Table 2). The Spanish profile for energy was adapted with data ofthe electricity production/importation system for Spain in 2011(RedElectrica, 2011). The electricity transmission network, emis-sions of sulphur hexafluoride to air, and losses during the mediumvoltage transmission and transformation from high to mediumvoltage were also considered (Dones et al., 2007).

The long term stability of the technologies was indirectly esti-mated by the lifespan of the packing materials throughout theoperation of the systems, which was applied to calculate packingmaterial requirements per FU. The production of Intalox Saddles forthe CS configuration was modelled according to the process ofpolyvinyl chloride (PVC) production (Hischier, 2007). Granularactivated carbon (GAC) from coal was considered in the AC system(Mu~noz, 2006). The amount of infrastructure material (concrete orglass fibre) was calculated assuming a wall thickness of 15 cm ofconcrete and 15 mm of glass fibre. The production processes ofNaOH, NaClO and HCl were taken from the Ecoinvent database andadapted to the concentrations and consumptions used in this casestudy (Althaus et al., 2007). Diaphragm cells were considered forthe production of NaOH (50%w/w), since it is the most commonprocess applied (Althaus et al., 2007). The production processes ofNaClO (25%w/w) and HCl (5%w/w) were estimated from processdata of 15%w/w and 30%w/w solutions, respectively (Althaus et al.,2007). Finally, the production of 1 kg of compost from 1 kg ofpacking material was assumed within the composting process.

Inventory data considered in the LCI were related to the steady-state operation of the systems. Therefore, the evaluation of the

Fig. 2. Cumulative energy demand (CED) for the five technologies evaluated and NTscenario.

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e84 81

robustness of the technologies towards process fluctuations andoperational upsets was out of the scope of the manuscript. For adetailed evaluation of the robustness of the most common odourtreatment technologies see Estrada et al. (2012).

2.5. Impact categories

In the LCIA phase, the role of each element of the LCI wasquantitatively assessed in terms of its contribution to nine impactcategories: i) Cumulative Energy Demand (CED), which accountsfor direct and indirect energy use throughout the entire life cycle ofa product or process (Huijbregts et al., 2010); ii) malodorous air,defined as the quantity of contaminated air exceeding the odourthreshold value established for each compound (Heijungs et al.,1992) and evaluated using the methodology developed by theCentre of Environmental Science of Leiden University (CML, 2001)(Guin�ee et al., 2002), and; iii) climate change, freshwater eutro-phication, photochemical oxidant formation, human toxicity, andmarine, terrestrial and freshwater ecotoxicity assessed with theReCiPe Midpoint method using the Hierarchist Perspective (H)version 1.05 (Goedkoop et al., 2009). The software SimaPro 8 wasused for the computational implementation of all the inventories(http://www.pre.nl/simapro).

3. Results

3.1. Malodorous air

The impact category of malodorous air is of major relevance inthis research study. The impact of malodorous air in the NT scenariowas very high: 55-times higher than the impact related to theworstcase scenario (BF): 48728 m3 m�3 for NT vs. 894 m3 m�3 for BF. Thetreatment of packing material in composting facilities entailed thehighest impacts for BF in this category. At this point, it must bestressed that the potential of malodorous air was reduced by 98% byBF and 99% with other alternative technologies.

Only the results related to the five technological alternativeswere plotted in Fig. 1 based on the large differences between the NTscenario and the impacts associated with the different technolo-gies. The main impact associated with the technology performancewas that corresponding to the treated effluent released to the at-mosphere. The treated gas accounted for values of malodorous air~95% in BTF, AC, CS and BTFþAC, whereas odorants associated withthe packing material production and disposal contributed to ~50%of the impact in BF (Fig. 1). Among the odorants released to theatmosphere, H2S was found the main responsible of themalodorous air impact related to direct emissions.

Fig. 1. Malodours air impact for the five technologies evaluated. Malodorous air of NT:48728 m3 per FU.

3.2. Cumulative energy demand (CED)

No cumulative energy demand was computed in the NT sce-nario. Therefore, the potential information that this indicator pro-vides is only useful when identifying the most energy efficienttechnology. The lowest CED impacts were obtained for BF and BTF,where this parameter was mainly associated with the use of energyfor air and water pumping (Fig. 2). The impact associated with ACwas mainly due to the packing material production, accounting for76% of the overall impact. Both GAC production and energy userepresented the impact associated with the hybrid technology,higher than those of biological techniques but ~50% lower than ACpotential impact. Finally, the production of chemicals used in the CScontributed with 72% to the global CED impact.

3.3. Climate change

The odorants present in the target malodorous emission do notcontribute directly to this category and, consequently, climatechange impact in the NT scenario was considered negligible.Moreover, although the oxidation of carbon-based odorants impliesthe emission of CO2, this contributionwas deemed negligible whencompared to other CO2 sources such as the indirect emissionsderived from packing material or chemicals use (Fig. 3). Biologicaltreatments exhibited the lowest climate change impacts. In fact,BTF and BF presented impacts ~91% and 84% lower than the impactassociated with AC (Fig. 3). Process energy requirements in BTF andBF were the main contributors to climate change, while GAC pro-duction and transportation was the main responsible of its climatechange potential in AC. The use of chemicals in CS accounted for

Fig. 3. Climate change impact for the five technologies evaluated and NT scenario.

Fig. 5. Photochemical oxidant formation of the five technologies evaluated and NTscenario.

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e8482

approximately 80% of the total impact, with NaClO presenting thelargest contribution. The hybrid technology presented lowerimpact than physical/chemical technologies but higher than bio-logical alternatives.

3.4. Freshwater eutrophication

This impact category is not directly applicable to the NT scenarioas the odorants present in the raw emission do not contribute towater eutrophication. CS presented the highest eutrophicationimpact due to the eutrophying emissions related to the productionof chemicals required (Fig. 4). In this particular case, NaClO pro-duction accounted for 88% of the impact associated with chemicals.Biological treatments presented the lowest impacts, with electricityuse as the main contributor in BF (91%) and BTF (55%). Similarly,energy use accounted for up to 50% of the eutrophication impact inthe hybrid system. Finally, GAC production in AC represented 63% ofthe eutrophication potential.

3.5. Photochemical oxidant formation

A total impact of 3.78� 10�6 kg NMVOC per FUwas calculated inthe NT scenario as a result of the direct release of VOCs to the at-mosphere (Fig. 5). The most important impacts were observed forphysical/chemical alternatives, followed by BTF þ AC and finally,biological technologies. Packing material accounted for 56% of thetotal impact in AC whereas chemicals production entailedremarkable impact in CS. Significantly lower impacts compared tothe NT scenario of 38%, 77% and 80% were estimated in BTFþAC, BFand BTF, respectively. The photochemical oxidant formation impactrelated to the production of energy was considerable in all tech-nologies evaluated. Finally, it is worth noting that the use of glassfibre as a building material resulted in higher impact compared tothe use of concrete.

3.6. Ecotoxicity and human toxicity

The toxicity associated with the treatment of the targetmalodorous emission was assessed in freshwater, marine andterrestrial environments and humans (Fig. 6 aed). Biologicaltreatments presented the most environmentally friendly perfor-mance, with energy use appearing as the main responsible of theimpact associated with BF. Production of glass fibre and disposal inlandfills entailed the largest impacts in BTF. The chemicals added toCS (especially NaClO) induced the largest impacts regardless of thetoxicity category. The contributors to toxicity-associated impacts in

Fig. 4. Freshwater eutrophication impact for the five technologies evaluated and NTscenario.

AC and BTF þ AC depended to a great extent on the categoriesanalyzed, disposal of packing material, energy use and glass fibreproduction contributing significantly to freshwater and marineecotoxicity, terrestrial ecotoxicity and human toxicity, respectively.Thus, although direct odorant emissions caused a negative impacton the receiving environment and on human health, this impactwas considered negligible when compared to the indirect impactsrelated to energy consumption, the production of the packingmaterial or the bed material disposal in landfills.

4. Discussion

Overall, biological processes exhibited an improved environ-mental performance, with the exception of the malodorous aircategory, where emissions from the composting process led tomore significant impacts in BF. In this context, it must be stressedthat although the conversion of organic wastes to compost reducestheir potential impacts to the environment, composting entailsnegative environmental impacts, such as the emission of H2S(Nemecek and K€agi, 2007). The main contributor to malodorous airassociated with the technologies were direct emissions, H2Sappearing as the most significant odorant compound. Firstly, theelevated H2S concentration in the raw emission(2.09 � 101 mg H2S m�3) resulted in a significantly high concen-tration of H2S in the outlet stream despite the great H2S removalefficiency (99%) achieved in all the technologies assessed. Secondly,this compound presents a characterisation factor of 2.33 � 109 m3

per kg of H2S, which is calculated as the inverse of the OdourThreshold Value (OTV). H2S exhibits a significantly low OTV(4.3� 10�4 mgm�3), which entailed negative impacts even at tracelevel concentrations. Treatment of odorants entailed a reduction of98e99% in the malodorous impact category.

The two biological treatments were found to be the friendliestalternatives to the environment, with diverse outcomes in thedifferent impact categories. BF was the most suitable technology interms of freshwater eutrophication, human toxicity, freshwaterecotoxicity and marine ecotoxicity, whereas the BTF exhibited thelowest impacts in terms of CED, climate change, photochemicaloxidant formation and terrestrial ecotoxicity. Energy use was foundto be the largest contributor to the impact in the BF scenario. Morespecifically, this consumption was directly related to the pressuredrop associated with the packing material. On the other hand,waste disposal in landfills and glass fibre used in BTF significantlyinfluenced its overall environmental impact, whereas concreteproduction was almost negligible in BF.

CS presented the highest impacts in freshwater eutrophication,photochemical oxidant formation, human toxicity and ecotoxicity

Fig. 6. Ecotoxicity in (a) freshwater, (b) marine and (c) terrestrial environments and (d) human toxicity associated with the five technologies evaluated and NT scenario.

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e84 83

due to the use of large amounts of chemicals. The significantamount of NaClO required for odorant oxidation had a key role inthe impacts associated with the use of chemicals. The negativeburdens derived from the use of chemicals in odour abatementtechnologies were also pointed out by Alfonsín et al. (2013) whenassessing the environmental consequences for human toxicity andecotoxicity of conventional biofilters at laboratory scale. Likewise,this high impact related to the use of chemicals in CS was inaccordance with the results obtained by Estrada et al. (2011).Nevertheless, the authors considered the amount of reagents usedper year, chemicals production being out of the analysis. The impactassociated with AC was similar to that related to NT scenario inphotochemical oxidant formation mainly due to packing materialproduction, whereas the use of chemicals caused the highestimpact in CS. Therefore, indirect impacts were found more impor-tant than the reduction of butanone directly released to the at-mosphere, which represented ~75% of the impact associated withthe NT scenario.

On the other hand, AC showed the worst results for CED andclimate change potentials followed by CS in both categories. Asalreadymentioned, themost significant parameter determining theAC impact was the GAC production process (Mu~noz, 2006), whichwas mainly attributed to a large demand of natural gas and itsreplacement every 6 months. The contribution of the packingmaterial to the environmental impact showed differences with thepreliminary analysis conducted by Estrada et al. (2011), who re-ported that BF was the technology presenting the highest annualpacked bed-material requirements. Nevertheless, LCA demon-strated that GAC production associated impacts weremore relevantthan those related to BF packing materials.

Finally, the hybrid technology (BTF þ AC) showed a lowerenvironmental performance than its biological counterparts, but

better results than physical/chemical technologies. BTF þ ACexhibited an improved environmental performance as compared toAC filtration due to the fact that the replacement frequency of GACwas reduced to 2 years instead of 6 months for stand-alone AC.Estrada et al. (2011), using the IChemE metrics, concluded thatBTF þ AC was the less favourable option in terms of energy con-sumption. In this study, energy use in BTF þ AC was found to besignificant in categories such as terrestrial ecotoxicity, but in nocase was this contribution comparable to the global impact ofphysical/chemical technologies. Therefore, BTF þ AC could beranked second after its biological counterparts, its implementationbeing highly recommended in odour sensitive scenarios due to itsmore reliable performance.

5. Conclusions

Biological treatments exhibited the best environmental perfor-mance among the technologies evaluated, except in terms ofmalodorous air impact where BF showed the highest impact as aresult of the contribution of compost production and disposal.Nonetheless, BF achieved a reduction in the malodour air impact of98% compared to the NT scenario, whichwas slightly lower than thereduction achieved by the other technologies (99%). Evenwhen theuse of a technology implied more significant impacts than the NTscenario, BF and BTF showed an improved environmental perfor-mance compared to physical/chemical technologies. In this context,AC and CS presented the highest impacts due to the use of GAC aspacking material, chemical use and waste disposal in landfills.Finally, the operation of BTF þ AC was more recommendable froman environmental sustainability perspective than physical/chemi-cal treatments, but slightly worse than biotechnologies. In otherwords, based on the fact that odour treatment is either compulsory

C. Alfonsín et al. / Journal of Environmental Management 149 (2015) 77e8484

to comply with, based on the legislative thresholds, or highly rec-ommended, in order to avoid public complaints, biotechnologiesare the best treatment options from an environmental point ofview.

Acknowledgements

The authors acknowledge the financial support of the SpanishMinistry of Economy and Competitiveness (project CTQ2012-34949and contract BES-2010-030994) and the European Union (LIVE-WASTE project: LIFE12 ENV/CY/544). The authors from the Uni-versity of Santiago de Compostela belong to the GalicianCompetitive Research Group GRC 2013-032, programme co-fundedby FEDER.

References

Alfonsín, C., Hernandez, J., Omil, F., Prado, O.J., Gabriel, D., Feijoo, G., Moreira, M.T.,2013. Environmental assessment of different biofilters for the treatment ofgaseous streams. J. Environ. Manage. 129, 463e470.

Althaus, H.J., Chudacoff, M., Hischier, R., Jungbluth, N., Osses, M., Primas, A., 2007.Life Cycle Inventories of Chemicals. Ecoinvent Report no. 8. Swiss Centre for LifeCycle Assessment, Dübendorf.

Balasubramanian, P., Philip, L., Murty Bhallamudi, S., 2012. Biotrickling filtration ofVOC emissions from pharmaceutical industries. Chem. Eng. J. 209, 102e112.

Baumann, H., Tillman, A.M., 2004. The Hitch Hiker's Guide to LCA : an Orientation inLife Cycle Assessment Methodology and Application. Studentlitteratur, Lund.

Doka, G., 2007. Life Cycle Inventories of Waste Treatment Services. Ecoinvent reportNo. 13. Swiss Centre for Life Cycle Inventories, Dübendorf.

Dones, R., Bauer, C., Bolliger, R., Burger, B., Faist Emmenegger, M., Frischknecht, R.,Heck, T., Jungbluth, N., R€oder, A., Tuchschmid, M., 2007. Life Cycle Inventories ofEnergy Systems: Results for Current Systems in Switzerland and Other UCTECountries. Ecoinvent report no. 5, Paul Scherrer Institut Villigen. Swiss Centrefor Life Cycle Assessment, Dübendorf, CH.

Dorado, A.D., Rodríguez, G., Ribera, G., Bonsfills, A., Gabriel, D., Lafuente, J.,Gamisans, X., 2009. Evaluation of mass transfer coefficients in biotrickling fil-ters: experimental determination and comparison to correlations. Chem. Eng.Technol. 32, 1941e1950.

Estrada, J.M., Kraakman, N.J., Munoz, R., Lebrero, R., 2011. A comparative analysis ofodour treatment technologies in wastewater treatment plants. Environ. Sci.Technol. 45, 1100e1106.

Estrada, J.M., Kraakman, N.J.R., Lebrero, R., Mu~noz, R., 2012. A sensitivity analysis ofprocess design parameters, commodity prices and robustness on the economicsof odour abatement technologies. Biotechnol. Adv. 30, 1354e1363.

Gabriel, D., Deshusses, M.A., 2004. Technical and economical analysis of the con-version of a full-scale scrubber to a biotrickling filter for odor control. Water Sci.Technol. 50, 309e318.

Goedkoop, M., Heijungs, R., Huijbregts, M., De Schryver, A., Struijs, J., van Zelm, R.,2009. ReCiPe 2008 e a Life Cycle Impact Assessment Method Which comprisesharmonised Category Indicators at the Midpoint and the Endpoint Level.Ministry of Housing, Spatial Planning and Environment (VROM), TheNetherlands.

Guin�ee, J.B., Gorr�ee, M., Heijungs, R., Huppes, G., Kleijn, R., Koning, A.d., Oers, L.v.,Wegener Sleeswijk, A., Suh, S., Udo de Haes, H.A., Brujin, H.d., va., D.R.,Huijbregts, M.A.J., 2002. Handbook on Life Cycle Assessment. Operational Guideto the ISO Standards. I: LCA in Perspective. IIa: Guide. IIb: Operational Annex. III:Scientific Background. Kluwer Academic Publishers, Dordrecht, ISBN 1-4020-0228-9.

Heijungs, R., Guin�ee, J.B., Huppes, G., Lankreijer, R.M., Udo de Haes, H.A., WegenerSleeswijk, A., Ansems, A.M.M., Eggels, P.G., Duin, R.v., Goede, H.P.d., 1992.

Environmental life cycle assessment of products. I: Guide - October 1992. II:Backgrounds - October 1992. CML (NOH Report 9266 þ 9267), ISBN 90-5191-064-9 (Leiden).

Hischier, R., 2007. Life Cycle Inventories of Packaging and Graphical Papers.Ecoinvent-Report No. 11. Swiss Centre for Life Cycle Inventories, Dübendorf.

Hospido, A., Moreira, M., Fern�andez-Couto, M., Feijoo, G., 2004. Environmentalperformance of a municipal wastewater treatment plant. Int. J. Life Cycle Assess.9, 261e271.

Hospido, A., Sanchez, I., Rodriguez-Garcia, G., Iglesias, A., Buntner, D., Reif, R.,Moreira, M.T., Feijoo, G., 2012. Are all membrane reactors equal from an envi-ronmental point of view? Desalination 285, 263e270.

Huijbregts, M.A., Hellweg, S., Frischknecht, R., Hendriks, H.W., Hungerbuhler, K.,Hendriks, A.J., 2010. Cumulative energy demand as predictor for the environ-mental burden of commodity production. Environ. Sci. Technol. 44, 2189e2196.

IChemE, 2002. The Sustainability Metrics. The Institution of Chemical Engineers,Rugby, UK.

ISO, 2006a. ISO 14040: 2006. Environmental Management e Life Cycle Assessmente Principles and Framework.

ISO, 2006b. ISO 14044: 2006. Environmental Management - Life Cycle Assessment -Requirements and Guidelines.

Kellenberger, D., Althaus, H.J., Jungbluth, N., Künniger, T., 2007. Life Cycle In-ventories of Building Products. Final report Ecoinvent data v2.0 No. 7. SwissCentre for Life Cycle Inventories, Dübendorf, Switzerland.

Kellenberger, D., Althaus, H.J., Jungbluth, N., Künniger, T., 2007. Life Cycle In-ventories of Building Products. Final report Ecoinvent data v2.0 No. 7. SwissCentre for Life Cycle Inventories, Dübendorf, Switzerland.Lebrero, R.,Gondim, A.C., P�erez, R., García-Encina, P.A., Mu~noz, R., 2014. Comparativeassessment of a biofilter, a biotrickling filter and a hollow fiber membranebioreactor for odour treatment in wastewater treatment plants. Water Res. 49,339e350.

Lebrero, R., Rodríguez, E., Estrada, J.M., García-Encina, P.A., Mu~noz, R., 2012. Odorabatement in biotrickling filters: effect of the EBRT on methyl mercaptan andhydrophobic VOCs removal. Bioresour. Technol. 109, 38e45.

Lebrero, R., Rodríguez, E., Martin, M., García-Encina, P.A., Mu~noz, R., 2010. H2S andVOCs abatement robustness in biofilters and air diffusion bioreactors: acomparative study. Water Res. 44, 3905e3914.

Mu~noz, I., 2006. Life Cycle Assessment as a Tool for Green Chemistry: Application toDifferent Advanced Oxidation Processes for Wastewater Treatment. DoctoralThesis. Autonomic University of Barcelona.

Nemecek, T., K€agi, T., 2007. Life Cycle Inventories of Swiss and European Agricul-tural Production Systems. Final report ecoinvent V2.0 No. 15a. AgroscopeReckenholz-Taenikon Research Station ART, Swiss Centre for Life Cycle In-ventories, Zurich and Dübendorf, CH.

Omri, I., Aouidi, F., Bouallagui, H., Godon, J.-J., Hamdi, M., 2013. Performance studyof biofilter developed to treat H2S from wastewater odour. Saudi J. Biol. Sci. 20,169e176.

RedElectrica, 2011. Red El�ectrica de Espa~na. El sistema el�ectrico espa~nol (inSpanish).

Revah, S., Morgan-Sagastume, J., 2005. Methods of odor and VOC control. In:Shareefdeen, Z., Singh, A. (Eds.), Biotechnology for Odor and Air PollutionControl. Springer, Berlin Heidelberg, pp. 29e63.

Rodriguez-Garcia, G., Molinos-Senante, M., Hospido, A., Hernandez-Sancho, F.,Moreira, M.T., Feijoo, G., 2011. Environmental and economic profile of six ty-pologies of wastewater treatment plants. Water Res. 45, 5997e6010.

Schlegelmilch, M., Streese, J., Stegmann, R., 2005. Odour management and treat-ment technologies: an overview. Waste Manage. 25, 928e939.

Shaw, A.R., Koh, S.-H., 2013. Gaseous emissions from wastewater facilities. WaterEnviron. Res. 85, 1298e1307.

Spielmann, M., Bauer, C., Dones, R., Tuchschmid, M., 2007. Transport Services.Ecoinvent report No. 14. Swiss Centre for Life Cycle Inventories, Dübendorf.

Zarra, T., Naddeo, V., Belgiorno, V., Reiser, M., Kranert, M., 2008. Odour monitoringof small wastewater treatment plant located in sensitive environment. WaterSci. Technol. 58, 89e94.