Science of the Total Environment 461–462 (2013) 409–415

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Economic valuation of environmental benefits of removing pharmaceuticaland personal care products from WWTP effluents by ozonation

M. Molinos-Senante e,⁎,1, R. Reif c,d,1, M. Garrido-Baserba b,c, F. Hernández-Sancho a,F. Omil d, M. Poch b,c, R. Sala-Garrido e

a Department of Applied Economics II, Universitat de Valencia, Campus dels Tarongers, 46022 Valencia, Spainb Catalan Institute for Water Research, Scientific and Technological Park, H2O Building, Emili Grahit 101, 17003 Girona, Spainc Laboratory of Chemical and Environmental Engineering (LEQUIA), Universitat de Girona, Facultat Ciències, Campus Montilivi, 17071 Girona, Spaind Chemical Engineering Department, Universidade de Santiago de Compostela, Rua Lope Gomez de Marzoa s/n, 15782 Santiago de Compostela, Spaine Department of Mathematics for Economy, Universitat de Valencia, Campus dels Tarongers, 46022 Valencia, Spain

H I G H L I G H T S

• Environmental Benefit Analysis of PPCPs• PPCPs' removal depends on their functional group and molecular structures.• Shadow prices as a proxy of the environmental benefits from ozonation process• HHCB and AHTN have the lowest shadow prices.• The greatest environmental benefit is associated with the removal of DCF.

⁎ Corresponding author. Tel.: +34 9638283349; fax:E-mail addresses: maria.molinos@uv.es (M. Molinos

(R. Reif), mgarrido@icra.cat (M. Garrido-Baserba), franc(F. Hernández-Sancho), francisco.omil@usc.es (F. Omil)(M. Poch), ramon.sala@uv.es (R. Sala-Garrido).

1 M. Molinos-Senante and R. Reif contributed equally

0048-9697/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.scitotenv.2013.05.009

a b s t r a c t

a r t i c l e i n f o

Article history:Received 18 January 2013Received in revised form 29 April 2013Accepted 5 May 2013Available online 6 June 2013

Editor: Simon Pollard

Keywords:PPCPsEnvironmental benefitsOzonationShadow pricesWastewater treatmentPost-treatment

Continuous release of pharmaceutical and personal care products (PPCPs) present in effluents from wastewatertreatment plants (WWTPs) is nowadays leading to the adoption of specific measures within the framework ofthe Directive 2000/60/EC (Water Framework Directive). The ozonation process, normally employed for drinkingwater production, has also proven its potential to eliminate PPCPs from secondary effluents in spite of their lowconcentrations. However, there is a significant drawback related with the costs associated with its imple-mentation. This lack of studies is especially pronounced regarding the economic valuation of the environ-mental benefits associated to avoid the discharge of these pollutants into water bodies. For the first timethe shadow prices of 5 PPCPs which are ethynilestradiol, sulfamethoxazole, diclofenac, tonalide andgalaxolide from treated effluent using a pilot-scale ozonation reactor have been estimated. Fromnon-sensitive areas their values are −73.73; −34.95; −42.20; −10.98; and −8.67 respectively andexpressed in €/kg. They represent a proxy to the economic value of the environmental benefits arisenfrom undischarged pollutants. This paper contributes to value the environmental benefits of implementingpost-treatment processes aimed to achieve the quality standards required by the Priority SubstancesDirective.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

In accordance with the Water Framework Directive (WFD) 2000/60/EC, the European States should implement measures aimed toachieve good ecological status for all water bodies. Consequently

+34 963828354.-Senante), rreif@icra.catesc.hernandez@uv.es, manel@lequia.udg.edu

to this work.

rights reserved.

the European Parliament adopted the Directive 2008/105/EC whichlays down environmental quality standards for priority substancesand certain other pollutants, aimed to achieve good surface waterchemical status. Following the precautionary principle, the EuropeanCommission has included for the first time three pharmaceuticalcompounds in the proposal for amending the list of priority sub-stances, based on the potential risk they pose to the environment:the anti-inflammatory diclofenac, the sex hormone estradiol and thecontraceptive pill ingredient ethynilestradiol.2 This statement is a

2 More information is available at: http://europa.eu/rapid/pressReleasesAction.do?reference=IP/12/88.

410 M. Molinos-Senante et al. / Science of the Total Environment 461–462 (2013) 409–415

direct consequence of the increasing concern derivative from theubiquitous presence of many different pharmaceutical and personalcare products (PPCPs) in water bodies worldwide (Camacho-Muñozet al., 2010; Chase et al., 2012; Comeau et al., 2008). These substancesfollow different pathways before reaching surface waters, being theeffluent from Wastewater Treatment Plants (WWTPs) identified astheir most relevant source (Clara et al., 2012; Reif et al., 2011). A con-sequence of the inclusion of pharmaceuticals in the priority substancelist is that member states have to ensure that their monitoring andspecific environmental quality standards should be met by 2021(deadline of the 2nd River Basin Management Plan). This leads to di-rect consequences in terms of overall cost due to the economic impactassociated with the upgrade of treatment facilities, currently unableto eliminate these and many other emerging contaminants (pesti-cides, surfactants, flame-retardants, nanomaterials, disinfectionby-products, etc.). An increasing number of papers and reviews deal-ing with the efficiency of different secondary treatment alternatives(Gabet-Giraud et al., 2010; Verlicchi et al., 2012; Petrovic et al., 2009)have been published throughout the last decade. According to them,it is obvious that biological treatments (with conventional or mod-ern configurations such as the membrane bioreactors) are inefficientto achieve the complete elimination of PPCPs from sewage, althoughthe adoption of specific operational strategiesmight contribute to in-crease their effectiveness. Therefore, research efforts are also focusedon more effective post-treatment alternatives, more concretely,technologies such as nanofiltration, reverse/forward osmosis mem-branes, activated carbon, ozonation and advanced oxidation process-es (AOPs). Such processes were able to reach eliminations up to 90%or even above for the majority of PPCPs detected in effluents (Dolaret al., 2011; Suarez et al., 2007; Lopez-Muñoz et al., 2012; De LaCruz et al., 2012). However, it is a complicated task to identify thebest post-treatment alternatives for specific scenarios, given the highnumber of substances of concern. In general, the main drawback forthe quick widespread of post-treatment technologies is based on theirhigh investment, operation &maintenance costs. Jones et al. (2007) an-ticipated these issues estimating costs of different alternatives and con-cluded that they might be economically and environmentallyundesirable. According to Owen and Jobling (2012), the upgrade of a fa-cility serving a townof about 250,000 peoplewith a post-treatment sys-tem based on granular activated carbon (GAC), efficient to cutethynilestradiol (EE2) levels, might cost about €8 million with a further€800,000/year being needed to operate the system. Despite its effec-tiveness for the elimination of EE2, GAC is not always the best optionin case a wider number of compounds have to be removed. In thissense, reverse osmosis (RO) and oxidation processes based in ozonehave proven to be suitable for removing a higher number of compounds(Radjenovic et al., 2008; Lee et al., 2012; Snyder et al., 2006). For exam-ple, Lee et al. (2012) carried out a direct comparison between ozonetreatment and RO, and found that despite their similar costs and perfor-mance in terms of PPCPs' elimination, when wider environmental im-pacts (energy consumption and waste production) were considered,ozone treatment appeared to be a more suitable technology for urbanwastewater applications. Ozonation effectiveness for PPCPs' removal isbased on two different reactive processes: direct oxidation with thehighly selective ozone and the secondary reactionswith themore pow-erful and unspecific hydroxil radical simultaneously produced duringthe ozonation process (Dodd et al., 2009). It is also necessary tomentionknown disadvantages such as the generation of intermediate productsmore toxic than the parent compound as well as the possible interfer-ences between the oxidation of bulk organic matter and target pollut-ants. However, these issues might be overcome as long as theeconomic feasibility of the overall process is granted. For example,Stalter et al. (2010) found that a further sand-filtration step reducestoxication effects of ozonated effluents.

Focusing on economic aspects, estimations carried out in previousworks have not considered the quantification of the environmental

benefits of preventing the discharge of pollutants into water bodies.Nowadays, such interest has been stirred up motivated mainly bytwo facts:

i) An increased awareness of the society on environmental protec-tion (Caballero et al., 2008). In this sense, pioneering methodolo-gies applied to wastewater management have slightly improvedthe perception of post-treatment strategies as environmentallyfriendly processes.

ii) The WFD allows temporal derogation or less stringent objectivesbased on the concept of disproportionate costs. Hence, the costsand the benefits of the measures to achieve the good ecologicalstatus should be assessed including the economic valuation ofthe environmental benefits.

In this context, the work of Hernández-Sancho et al. (2010) waspioneering since based on the methodology developed by Färe etal. (1993) it quantified the shadow prices of some wastewater pol-lutants. These shadow prices can be interpreted as the economicvalue of the environmental benefits for avoiding the discharge ofcontaminants into water bodies. The subsequent developments car-ried out (Molinos-Senante et al., 2010a, 2011a) have proven that thismethodology, named in this work as “Environmental Benefit Analy-sis” (EBA) is a very suitable approach for this purpose. Despite theusefulness of these previous works, their main limitation is thatEBA has been applied only for the main wastewater pollutants suchas chemical oxygen demand, suspended solids, nitrogen and phos-phorus. In this sense and taken into account the Directive 2008/105/CE, the aim of this research is to value the economic benefits ofremoving 5 different PPCPs from WWTP effluent by ozonation.Moreover, the performance of this post-treatment for removingsuch compounds has been also evaluated.

2. Materials and methods

2.1. PPCPs' selection and analytical methodologies

The 5 PPCPs studied in this work (diclofenac, ethynilestradiol,sulfamethoxazole, tonalide and galaxolide) were selected accordingto their different physical chemical properties being also represen-tative of different therapeutic groups and uses (Table 1). Thus,they are expected to show different behaviors when undergoingozone treatment. The five compounds are frequently found inWWTP effluents and therefore, they truly represent the vast numberof compounds which can be found in the aquatic environment.PPCPs' analysis was performed adapting methods already published(Rodríguez et al., 2003; Vanderford et al., 2003). Briefly, the contentof galaxolide, diclofenac and tonalide was determined aftersolid-phase extraction (SPE) of 300 mL samples using 60 mgOASIS HLB cartridges (Waters, Milford, MA, USA) and then quantita-tively eluted from the cartridge using 3 mL of ethyl acetate.Meclofenamic acid was spiked to the samples as surrogate standard.The extract was divided into two fractions: one of them was used forthe direct determination of tonalide and galaxolide; the second onewas employed for the determination of diclofenac, as a tert-butyldimethylsilyl derivative. Detection was carried out by GC/MSusing a Varian CP 3900 chromatograph (Walnut Creek, CA, USA)equipped with a split-splitless injector and connected to an ion-trap mass spectrometer (GC/MS). To measure ethynilestradiol andsulfamethoxazole, SPE was done as described for the previous com-pounds, although the elution step was performed with a mixture of1.5 mL of methanol and the same amount of methyl tert-butylether. Final detection and quantification were performed by LC/MS/MS, using an Agilent Liquid Chromatograph API 400 GI312A)equipped with a binary pump and autosampler HTC-PAL and with atriple quadruple mass spectrometer (ESI+ mode).

Table 1Selection of PPCPs and physical–chemical properties.

PPCPs Acronym Function s log Kow pKa H

Diclofenac DCF Analgesic 2.4 4.5–4.8 4.0–4.5 1.9·10−10

Ethinylestradiol EE2 Sex hormone 11.3 2.8–4.2 10.5–10.7 3.3·10−10

Sulfamethoxazole SMX Antibiotic 610 0.5–0.9 5.6–6.0 2.6·10−11

Galaxolide HHCB Musk fragrance 1.8 5.9–6.3 – 4.5·10−3

Tonalide AHTN Musk fragrance 1.2 4.6–6.4 – 5.1·10−3

s: solubility in water (mg·L−1); H: Henry coefficient (μg·m−3 air/μg·m−3 wastewater); pKa: dissociation constant; Kow: octanol–water partition coefficient: Values taken fromSuárez et al. (2008).

411M. Molinos-Senante et al. / Science of the Total Environment 461–462 (2013) 409–415

2.2. Ozonation bench-scale reactor and membrane bioreactor

In order to simulate conditions similar to those applied in the ozon-ation process, different strategies were followed. First, a pilot-scalemembrane bioreactor (MBR) equipped with a hollow-fiber Zenonmodulewas placed at the premises of aWWTP,with the aim to producea real effluent (additional information of the MBR used can be found atReif et al., 2008). Sewage from the primary settlers of the pilot plantwaspumped to a continuously stirred tank of aluminumwhere PPCPs werespiked in concentrations close to or within the typical environmentalrange (1–10 ppb). The spiked sewage served as feeding to the MBR,whichwas operated in optimum conditions tomaximize PPCPs' remov-al during the biological treatment (extended sludge retention time).Consequently, the final permeate was suitable to carry out a furtherpost-treatment with an ozonation pilot-scale reactor, in conditionsidentical to those established in full-scale facilities. Although it mighthave been feasible to directly work with effluents from a conventionalactivated sludge plant, for the purposes of this research it was preferredto use MBR permeate due to the almost complete absence of suspendedsolids. This operational strategy permits to carry out the EBA analysis ex-clusively for PPCPs present in the liquid phase. Obviously, it is expectedthat a small amount of some PPCPs is attached to solids as well. Never-theless, such fractionmight only be relevant for the fragrances galaxolideand tonalide due to their high log Kow value (Table 1) and the consider-ation of this fraction is out of the scope of this economic assessment,since sludge management is carried out with different processes.

The ozonation reactor consisted of a feeding tank, a glass contactor of10 L connected to an ozone generator (BMT 803, Messtechnik GMBH)and two ozone analyzers in the gaseous phase (BMT 964 C, MesstechnikGMBH) placed at the inlet and outlet of the glass contactor. Additionaldevices were used to ensure the correct functioning of the pilot plant,as shown in Fig. S1. Dried air was used as a source for ozone production.Previous to the ozonation experiments, a test was carried out with tapwater, in order to adjust the production of ozone in the generator andthe air flow-rate necessary to achieve a constant concentration of 2.2 gO3/m3 at the inlet gas stream. 10 L of permeate recently collected fromthe MBR was ozonated for 10 min and 5 samples (0, 2, 4, 7 and10 min) were collected during the process.

2.3. Environmental Benefit Analysis: shadow prices' estimation

To estimate the shadow prices of wastewater pollutants, two ap-proaches can be distinguished: i) output distance function(Hernández-Sancho et al., 2010) and ii) directional distance function(Molinos-Senante et al., 2011a). Both approaches are based on thesame concept and a review of the literature (Hernández-Sancho etal., 2010; Molinos-Senante et al., 2011a,b) has illustrated that theyprovide similar results. Hence and due to its higher simplicity, theoutput distance function approach has been used.

For the purpose of deriving the shadow prices of the pollutants, theoutput distance function has several advantages. First, like the produc-tion function, it completely describes technology and is scalar-valued(Campbell, 2008). In contrast to the production function, the distancefunction models joint production of multiple outputs, being especiallyuseful when multiple inputs are used to produce multiple outputs

(Yin and Wang, 2010). The advantage of distance function over cost,profit, and revenue functions (which can also be readily used tomodel multiple output technology) is that no maintained behavioralhypothesis (cost minimum or profit/revenue maximum) is required. Adistance function only identifies the technology frontier and gives thedistance to the frontier for each observation (Kumbhakar and Lovell,2000). As a result, the different measures of economic effects are notconditional on behavioral hypothesis. Also a great practical importanceis that it can be computedwith data on quantities of inputs and outputsalone; prices are not needed (Liao et al., 2009). This fact is of particularimportance in our case because our undesirable outputs (PPCPs re-moved from wastewater) are not traded in the market and the dataon their prices/cost are not observable.

In a conventional production process, a unit employs a vector of Ninputs x � RN

þ to produce a vector of M outputs � RMþ . In the field of

thewastewater treatment, it has been considered that the vector of out-puts not only includes desirable ones (treated water), u1 � RH

þ, but alsoundesirable ones (pollutants removed from wastewater), u2 � RM−H

þ ,so that u = [u1,u2]. Hence, the distance function is defined as:

D0 x;uð Þ ¼ θ :u�

θ

� ��P xð Þg:� ð1Þ

The technology P(x) satisfies weak disposability if u � P xð Þ�½ whichimplies that θ u � P xð Þ�½ for every θ � 0;1½ � (Coggin and Swinton,1996). Weak disposability has allowed to incorporate environmentaloutputs, which are not to be freely disposed, as elements of the out-put vector (Färe et al., 1993). Färe and Grosskopf (2004) also notedthat desirable and undesirable outputs are null-jointness, whilethere are thought to be jointly produced. The null-jointness propertymeans that if desirable outputs are produced, undesirable outputs arealso produced. In our case this property is obvious, since to produceclean water, pollutants have to be removed from the wastewater.

Following previous works (Färe et al., 1993; Hernández-Sancho etal., 2010), the parameterization of the distance function has been car-ried out through the translog function. Taken into account that ourproblem has k units, n inputs, m desirable output and m′ undesirableoutputs the formula is:

lnD0 xk;uk� �

¼ α0 þ∑Nn¼1βn lnx

kn þ∑M

m¼1αm lnukm

þ 12∑N

n¼1∑Nn′¼1βnn′ lnxkn

� �lnxk′n

� �

þ 12∑M

m¼1∑Mm′¼1αmm′ lnuk

m

� �lnuk′

m

� �

þ∑Nn¼1∑

Mm¼1γnm lnxkn

� �lnuk′

m

� �: ð2Þ

To calculate the parameters α, β and γ the following linear pro-gram should be solved:

Min ∑K

k¼1lnD0 xk;uh

� �− ln1

h i

s.t.:

i) ln D0(xk,uh) ≥ 0

ii)∂D0 xk ;ukð Þ

∂ lnxkn≥0

0.0

0.2

0.4

0.6

0.8

1.0

0 2 4 7 10

C/C

o

t (min)

0 2 4 7 10

t (min)

Tonalide

Galaxolide

0.0

0.2

0.4

0.6

0.8

1.0

C/C

o

Diclofenac

Sulfamethoxazole

Ethinylestradiol

Fig. 1. PPCPs' depletion from permeate during ozone treatment.

412 M. Molinos-Senante et al. / Science of the Total Environment 461–462 (2013) 409–415

iii)∂D0 xk ;ukð Þ

∂ lnukn≤0; m ¼ 1ð Þ

iv)∂D0 xk ;ukð Þ∂ lnuk

m′

≥0; m′ ¼ 2;3;…ð Þ

v) ∑Mm¼1αm ¼ 1;∑N

n′¼1βnn′ ¼ ∑Nn¼1γnm ¼ 0

vi) αmm′ = αm′m;m = 1,2,…M,m′ = 1,2,…M.

βnn′ ¼ βn′n;n ¼ 1;2;…N;n′ ¼ 1;2;…N ð3Þ

The relationship of duality between the output distance functionand the revenue function (Shephard, 1970) is the key point to esti-mate the shadow prices of the undesirable outputs. On the onehand, the output distance function can be obtained from the revenuefunction through maximization over output prices. On the otherhand, the revenue function can be derived from the output distancefunction by maximization with respect to outputs. The relationshipcan be expressed as:

R x; rð Þ ¼ supu ru : D0 x;uð Þ≤1f g

D0 x;uð Þ ¼ supr

ru : R x; rð Þ≤1f gð4Þ

where R(x,r) is the revenue function and ru is the inner product of theoutput price and quantity vector. To calculate the shadow prices ofthe undesirable outputs, it is necessary to assume that the shadowprice of the desirable output coincides with its market price. Inother words, it is necessary to assign a reference price to the treatedwater. If m is a desirable output whose market price is rm, then theshadow price for each undesirable outputm′ is given by the followingexpression (Färe et al., 1993):

r′m ¼ rm

∂D0 x;uð Þ.

∂um′

∂D0 x;uð Þ.

∂um

: ð5Þ

3. Results and discussion

3.1. PPCPs' elimination

PPCPs' removal strongly depends on their functional groups andmolecular structures. Compounds with lower rates are characterizedby the absence of reactive groups and non-activated aromatic ringson their structures, as in the case of the fragrances HHCB and AHTN.On the contrary, ozone reacts faster with compounds containing dou-ble bonds, amino groups (SMX and DCF) or activated aromatic rings(EE2). Comparing the curves from Fig. 1, it is clear that PPCPs studiedin this work followed exactly that trend. On the one hand, the muskfragrances HHCB and AHTN were gradually removed during the pro-cess, although their complete elimination was never achieved, and onthe other hand, pharmaceuticals' removal was significantly fasterduring the first 2 min of ozonation.

Using different strategies, the second-order ozonation rate con-stants of several PPCPs have been determined by different authors(Table 2). A good agreement between data from Table 2 and our mea-surements is observed since, for example, the lowest KO3 value fromthe table corresponds with AHTN, closely followed by HHCB. This isin good concordance with their chemical structures as well, since nospecific site for ozone attack can be found on them.

It should be mentioned that the results include every feasible re-moval mechanism that might play a significant role during the ozon-ation process. According to Nöthe et al. (2007), stripping might berelevant for musk fragrances in a bubble column, more concretelyto AHTN due to its low Henry's law coefficients (Table 1). It is also

well-known that ozone decomposes into hydroxyl radicals which canreact nonselectively with high rate constants (109 to 1010 M−1 s−1)with ozone-recalcitrant compounds. Therefore, every parallel processthat removes PPCPs during the ozonation exerts influence on the esti-mation of environmental benefits, although it is expected that theircontribution is lower compared to the removal due to direct reactionwith ozone.

3.1.1. Estimation of shadow pricesFollowing the methodology previously described it has been con-

sidered that the desirable output is the treated water (u1) and theundesirable outputs are the pollutants removed from wastewater:diclofenac (u2), tonalide (u3), galaxolide (u4), sulfamethoxazole(u5) and ethinylestradiol (u6). The inputs to carry out the ozonationprocess are the costs which have been grouped into four items: ener-gy (x1), staff (x2), maintenance (x3) and other (x4).

It is important to remark that most of the published studies regard-ing ozonation costswere referred only to energy for the ozone synthesiswhile other item costs such as staff or maintenance were omitted(Krichevskaya et al., 2011). However, a realistic cost assessment of theprocess must include all cost items. In this context, energy costs havebeen quantified based on the ozone dosed, the energy required for itssynthesis (28 kWh/kg O3) and the average price of the electricity forthe 27 EU member states (0.11 €/kWh) (EUROSTAT, 2012). Followingto EPA (1999) and Molinos-Senante et al. (2010b), it has been consid-ered that in the ozonation process, energy costs are around 45% of thetotal costs while staff cost is 30%, maintenance cost is 20% and othercost is 5%. This approach has enabled us to quantify the total costs asso-ciated to the ozonation process in realWWTPs. Table 3 shows themeanand the standard deviation of the parameters used to estimate the shad-ow prices.

Table 4Parameters and coefficients associated with a translog function.

α0 −95.96830 α14 4.38642 α46 0.00000 γ25 0.00000α1 0.00000 α15 0.00000 α55 0.00000 γ26 0.64676α2 0.00000 α16 0.00000 α56 0.00959 γ31 0.00000α3 0.00000 α22 0.00000 α66 −0.37500 γ32 0.00000α4 0.00000 α23 0.00000 γ11 0.00000 γ33 0.00000α5 0.00000 α24 0.00000 γ12 0.00000 γ34 0.00000α6 1.00000 α25 0.00154 γ13 0.00000 γ35 0.00000β1 0.00000 α26 −0.00073 γ14 0.00000 γ36 0.00000β2 0.00000 α33 0.00000 γ15 0.00000 γ41 0.00000β3 0.00000 α34 0.00000 γ16 0.00000 γ42 0.02082β4 0.00000 α35 −0.02424 γ21 0.00000 γ43 0.02754α11 2.40823 α36 0.04443 γ22 0.00000 γ44 5.76883α12 0.00000 α44 0.00000 γ23 0.00000 γ45 0.30378α13 0.00000 α45 0.32049 γ24 0.00000 γ46 0.00000

Table 2Reported second order rate constant (KO3) and structure of the studied PPCPs.

PPCPs KO3 (M−1 s−1) Reference KO3 Structure

Diclofenac 1.0·106 Huber et al. (2003)1.69–1.99·104 Vogna et al. (2004)

Ethinylestradiol 7.0·109 Huber et al. (2003)1.45–2.21·105 Deborde et al. (2005)

Sulfamethoxazole 4.7–6.7·105 Dodd et al. (2006)4.4·105 Dantas et al. (2008)

Tonalide 10 Janzen et al. (2011)8 Nöthe et al. (2007)

Galaxolide 67 Janzen et al. (2011)140 Nöthe et al. (2007)

413M. Molinos-Senante et al. / Science of the Total Environment 461–462 (2013) 409–415

Accordingly to the methodology, the first step to calculate theshadow prices of the selected PPCPs is to parameterize and calculatethe parameters of the distance function (Eq. (2)). In doing so, the lin-ear problem (3), involving one desirable output m (m = 1) and 5undesirable outputsm′ (m′ 2 to 6) was solved using General Algebra-ic Modeling System (GAMS) software. Table 4 shows the value of theparameters. Note that the sum of α1 to α6 values is equal to one ver-ifying that the program was solved correctly.

Because the estimation of the shadow prices requires assigning areference price for the treated water (desirable output), two scenarioshave been evaluated. Following to Hernández-Sancho et al. (2010) ref-erence prices were allocated depending on the destination of treatedwater, sensitive and non-sensitive areas. Table 5 shows the averagevalues of shadow prices and their standard deviation in parenthesis,obtained for each compound and for each scenario.

As expected, the shadow prices are negative since they are not mar-ketable outputs which could generate a direct income. However,according to previous works (Molinos-Senante et al., 2010b, 2011a),shadow prices can be interpreted positively because they representthe value of external effects that could damage the environment whenpollutants are dumped into the environment. In case the wastewatertreatment plant is placed on a sensitive area, the damage avoided willbe significantly greater, and thereby, the shadow prices estimated forthis scenario will be higher as well. The shadow prices calculated inthis work constitute a proxy of the environmental benefits gainedfrom the application of the ozonation process, i.e., they are equivalentto the positive externalities associated to avoid the discharge of PPCPsinto water bodies. Comparing the calculated shadow prices with therates observed for each PPCP, it is obvious that the economic effective-ness of the process, from the point of view of the EBA methodology, is

Table 3Sample data for estimating shadow prices.Source: Own elaboration.

Mean Standard deviation

Inputs(€/m3)

Energy 1.20E−04 2.56E−05Staff 7.98E−05 1.71E−05Maintenance 5.32E−05 1.14E−05Others 1.33E−05 2.84E−06

Desirable output(m3/year)

Wastewater treated 525.6 0.0

Undesirable outputs(mg/m3)

DCF 2.113 1.218AHTN 0.640 0.377HHCB 0.466 0.371SMX 0.851 0.429EE2 0.063 0.011

strongly influenced by its overall performance in terms of PPCPs' re-moval. In this manner, the lowest shadow prices were obtained forboth musks HHCB and AHTN, which were in turn the most recalcitrantto ozone treatment. The calculated shadow prices can be comparedwith those obtained for conventional pollutants which were, fornon-sensitive areas, −16.353 (N), −30.944 (P), −0.098 (COD) and−0.005 (SS) in €/kg (Hernández-Sancho et al., 2010). It is interestingto note that shadow prices for N and P (higher environmental damagedue to eutrophization) were within the same range of those calculatedfor PPCPs. However, for estimating the overall benefits fromwastewatertreatment, not only the value of the shadow prices (€/kg) should beconsidered but also the volume of each pollutant removed (kg/m3).Since PPCPs are eliminated in quantities much lower than traditionalpollutants, although their shadow prices are similar, the removal ofPPCPs would contribute less than the removal of the traditional pollut-ants to the global environmental benefit from wastewater.

With the aim to estimate environmental benefits expressed inmon-etary termsderived from treating the effluentwith ozone, an exemplarycase study can be presented. Verlicchi et al. (2012) reviewedmore than30 papers to report about the concentration of 118 pharmaceuticals insecondary effluents. For DCF, SMX and EE2, effluent concentrationswere 0.8, 0.28 and 0.003 μg/L respectively. Regarding the twomusk fra-grances, the typical concentrations after the secondary treatment havebeen estimated based on the data provided by Artola-Garicano et al.(2003), Horii et al. (2007) and Chase et al. (2012). All of them moni-tored AHTN and HHCB in secondary effluents from real WWTPs.Hence, the mean effluent concentrations were 0.686 μg/L for AHTNand 3.675 μg/L for HHCB. Considering that ozone treatment might beable to reduce those loads down to 95%, and the mean shadow prices(Table 5), the environmental benefits obtained to avoid the dischargeof each PPCP can be evaluated. For example, in aWWTPof 100,000 pop-ulation equivalent treating 50,000 m3/d of sewage and discharging thefinal effluent in a sensitive area, the total benefit associated with theelimination of the 5 PPCPs would be of 1833.3 €/year. The greatest en-vironmental benefit is associated with the removal of DCF since it rep-resents 40.4% of the total benefit (741.6 €/year). In spite of its lowshadow price, HHCB is the next most important pollutant since it is re-moved inmajor quantity, accounting for 38.4% (705 €/year). The contri-bution of SMX and AHTN to the total benefit is similar and around 10%(216.0 and 166.3 €/year respectively). EE2 is the PPCP with the highestshadow price, but due to its low concentration in effluents, its contribu-tion to the total benefit is almost negligible (0.3% or 4.9 €/year) sincethe mass of compound removed from wastewater is low comparedwith other PPCPs.

Perhaps the value of 1833.3 €/year may seem low; however, wemust not forget that the ozonation process removes moremicropollutants than the 5 PPCPs evaluated in this example. There-fore, to estimate the total environmental benefit associated to theozonation post-treatment, the shadow price of the rest of pollutantsremoved should be considered. Otherwise, there would be an

Table 5Average values of shadow prices for undesirable outputs (€/kg) and their standard deviation in parenthesis.

Scenarios DCF AHTN HHCB SMX EE2

Non-sensitive −42.20 (−4.63) −10.98 (4.33) −8.67 (−3.97) −34.95 (−17.76) −73.73 (−24.13)Sensitive −53.47 (−5.21) −13.98 (5.88) −11.06 (−4.85) −44.46 (−23.06) −93.76 (−28.57)

414 M. Molinos-Senante et al. / Science of the Total Environment 461–462 (2013) 409–415

underestimation of the environmental benefits. In this sense, in arough estimation, and considering that the number of PPCPsreported by Verlicchi et al. (2012) was 118, the average environmen-tal benefit might be around 20 twice higher than our estimation.

4. Conclusions

The importance of preventing the discharge of PPCPs into surfacewater has been confirmed by the European Commission which hasrecently adopted the inclusion of three pharmaceutical compoundsin the list of priority substances. While technical aspects about the re-moval of micropollutants from wastewater have been widely studied,the truth is that the assessment of the economics has been morelimited.

In this context, and as it has been reported by the WFD, it is impor-tant to evaluate not only the cost of the measures or processes but alsothe benefits associated to them, including the environmental ones.Hence, this work presents an economic valuation of the environmentalbenefits derived from eliminating DCF, AHTN, HHCB, SMX and EE2 fromwastewater. In doing so, their shadow prices have been estimated forboth sensitive and non-sensitive areas. In the first scenario, the valuesobtained, expressed in €/kg, are −53.47; −13.98; −11.06; −44.46and −93.76 respectively. On the other hand, for non-sensitive areasthe shadow prices are lower being −42.20; −10.98; −8.67; −34.95and −73.73 (also expressed in €/kg).

Ozonation experiments were carried out to determine the concen-tration of PPCPs in the effluent. It has been verified that the elimina-tion rates depend on the functional groups and molecular structuresof the compounds. Thereby the two musk fragrances studied, HHCBand AHTN, present lower rates than the pharmaceuticals — SMX,DCF and EE2.

The value of the shadow prices illustrates that the removal ofpharmaceutical compounds involves higher environmental benefitsthan the elimination of musk fragrances. This result is consistentwith the current policy and management measures aimed to protectEuropean water bodies. However, to estimate the total environmentalbenefit not only the shadow prices have to be considered but also thequantity of each pollutant removed from wastewater. In this sense, ithas been verified that the DFC is the PPCP whose elimination contrib-utes more to the total environmental benefit.

To conclude, we would like to emphasize that additional effortsare needed aimed to estimate the shadow prices of more PPCPs.Hence, the true environmental benefits of implementing tertiarytreatments might be assessed. Nevertheless, this paper contributesto this challenge since for the first time the environmental benefitgained for avoiding the discharge of some PPCPs into surface waterhas been valuated from an economic point of view.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2013.05.009.

Acknowledgments

The authorswould like to thank all NOVEDARprojectmembers. Thisstudy has been partially financed by the Spanish Ministry of Educationand Science (Consolider Project-NOVEDAR) (CSD2007-00055), the Eu-ropean Commission (FP7-ENV-2010-265213; LIFE10-ENV/ES000520),

the Generalitat Valenciana (Project ACOMP/2012/259) and the Xuntade Galicia (Project 09MDS010262PR). H2O Building, headquarters ofthe ICRA, has been fundedby theMinistry of Economyand Competitive-ness (MINECO) and the European Regional Development Fund (ERDF)under the ERDF Operational Programme 2007–2013 in Catalonia.

References

Artola-Garicano E, Borkent I, Hermens JLM, Vaes WHJ. Removal of two polycyclicmusks in sewage treatment plants: freely dissolved and total concentrations. Envi-ron Sci Technol 2003;37:3111–6.

Caballero R, Gil A, Fernandez-Santos X. An experts' survey on sustainability acrosstwenty-seven extensive European systems of grassland management. EnvironManag 2008;42(2):190–9.

Camacho-Muñoz MD, Santos JL, Aparicio I, Alonso E. Presence of pharmaceutically activecompounds in Doñana Park (Spain) main watersheds. J Hazard Mater 2010;177:1159–62.

Campbell HF. The shadow-price of labour and employment benefits in a developingeconomy. Aust Econ Pap 2008;47(4):311–9.

Chase DA, Karnjanapiboonwong A, Fang Y, Cobb GP, Morse AN, Anderson TA. Occurrenceof synthetic musk fragrances in effluent and non-effluent impacted environments.Sci Total Environ 2012;416:253–60.

Clara M, Windhofer G, Weilgony P, Gans O, Denner M, Chovanec A, et al. Identificationof relevant micropollutants in Austrian municipal wastewater and their behaviourduring wastewater treatment. Chemosphere 2012;87(11):1265–72.

Coggin JS, Swinton JR. The price of pollution: a dual approach to valuing SO2 allow-ances. J Environ Econ Manag 1996;30(5):58–72.

Comeau F, Surette C, Brun GL, Losier R. The occurrence of acidic drugs and caffeine insewage effluents and receiving waters from three coastal watersheds in AtlanticCanada. Sci Total Environ 2008;396(2–3):132–46.

Dantas RF, Contreras S, Sans C, Esplugas S. Sulfamethoxazole abatement by means ofozonation. J Hazard Mater 2008;150(3):790–4.

de la Cruz N, Gimenez J, Esplugas S, Grandjean D, De Alencastro LF, Pulgarin C. Degra-dation of 32 emergent contaminants by UV and neutral photo-Fenton in domesticwastewater effluent previously treated by activated sludge. Water Res 2012;46(6):1947–57.

Deborde M, Rabouan S, Duguet JP, Legube B. Kinetics of aqueous ozone-induced oxida-tion of some endocrine disruptors. Environ Sci Technol 2005;39(16):6086–92.

Dodd MC, Buffle MO, von Gunten U. Oxidation of antibacterial molecules by aqueousozone: moiety-specific reaction kinetics and application to ozone-based wastewa-ter treatment. Environ Sci Technol 2006;40(6):1969–77.

Dodd MC, Kohler HPE, von Gunten U. Oxidation of antibacterial compounds by ozoneand hydroxyl radical: elimination of biological activity during aqueous ozonationprocesses. Environ Sci Technol 2009;43(7):2498–504.

Dolar D, Vukovic A, Asperger D, Kosutic K. Effect of water matrices on removal of vet-erinary pharmaceuticals by nanofiltration and reverse osmosis membranes. J Envi-ron Sci 2011;23(8):1299–307.

EPA. Wastewater technology fact sheet. Ozone disinfection. EPA 832-F-99-063. EPA;1999.

EUROSTAT. Energy price statistics. Available at:http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Energy_price_statistics#Further_Eurostat_information.

Färe R, Grosskopf S. Modeling undesirable factors in efficiency evaluation: comment.Eur J Oper Res 2004;157(1):242–5.

Färe R, Grosskopf S, Lovell CAK, Yaisawarng S. Derivation of shadow prices for undesir-able outputs: a distance function approach. Rev Econ Stat 1993;75(2):374–80.

Gabet-Giraud V, Miège C, Choubert JM, Ruel SM, Coquery M. Occurrence and removalof estrogens and beta blockers by various processes in wastewater treatmentplants. Sci Total Environ 2010;408(19):4257–69.

Hernández-Sancho F, Molinos-Senante M, Sala-Garrido R. Economic valuation of envi-ronmental benefits from wastewater treatment processes: an empirical approachfor Spain. Sci Total Environ 2010;408:953–7.

Horii Y, Reiner JL, Loganathan BG, Kumar KS, Sajwan K, Kannan K. Occurrence and fateof polycyclic musks in wastewater treatment plants in Kentucky and Georgia, USA.Chemosphere 2007;68:2011–20.

Huber MM, Canonica S, Park GY, von Gunten U. Oxidation of pharmaceuticals during ozon-ation and advanced oxidation processes. Environ Sci Technol 2003;37(5):1016–24.

Janzen N, Dopp E, Hesse J, Richards J, Türk J, Bester D. Transformation products and re-action kinetics of fragrances in advanced wastewater treatment with ozone.Chemosphere 2011;85:1481–6.

Jones OAH, Green PG, Voulvoulis N, Lester JN. Questioning the excessive use of ad-vanced treatments to remove organic micropollutants from wastewater. EnvironSci Technol 2007;41(14):5085–9.

415M. Molinos-Senante et al. / Science of the Total Environment 461–462 (2013) 409–415

Krichevskaya M, Klauson D, Portjanskaja E, Preis S. The cost evaluation of advanced oxida-tion processes in laboratory and pilot-scale experiments. Ozone Sci Eng 2011;33(3):211–23.

Kumbhakar SC, Lovell CAK. Stochastic frontier analysis. Cambridge: Cambridge Univer-sity Press; 2000.

Lee CO, Howe KJ, Thomson BM. Ozone and biofiltration as an alternative to reverse os-mosis for removing PPCPs and micropollutants from treated wastewater. WaterRes 2012;46(4):1005–14.

Liao CN, Önal H, Chen MH. Average shadow price and equilibrium price: a case study oftradable pollution permit markets. Eur J Oper Res 2009;196(3):1207–13.

Lopez-Muñoz MJ, Sotto A, Arsuaga JM. Nanofiltration removal of pharmaceutically ac-tive compounds. Desalin Water Treat 2012;42(1–3):138–43.

Molinos-Senante M, Hernández-Sancho F, Sala-Garrido R. Estudio de viabilidad delproceso de recuperación del fósforo contenido en las aguas residuales: unaaproximación económica Rect@, 11. ; 2010a. p. 179–95. [In Spanish]

Molinos-Senante M, Hernández-Sancho F, Sala-Garrido R. Economic feasibility studyfor wastewater treatment: a cost–benefit analysis. Sci Total Environ 2010b;408:4396–402.

Molinos-Senante M, Hernández-Sancho F, Sala-Garrido R. Cost–benefit analysis ofwater-reuse projects for environmental purposes: a case study for Spanish waste-water treatment plants. J Environ Manage 2011a;92:3091–7.

Molinos-Senante M, Hernández-Sancho F, Sala-Garrido R. Assessing disproportionatecosts to achieve good ecological status of water bodies in a Mediterranean riverbasin. J Environ Monit 2011b;13(8):2091–101.

Nöthe T, Hartmann D, von Sonntag J, von Sonntag C, Fahlenkamp H. Elimination of themusk fragrances galaxolide and tonalide from wastewater by ozonation and con-comitant stripping. Water Sci Technol 2007;55(12):287–92.

Owen R, Jobling S. Environmental science: the hidden costs of flexible fertility. Nature2012;485:441.

Petrovic M, De Alda MJL, Diaz-Cruz S, Postigo C, Radjenovic J, Gros M, et al. Fate and re-moval of pharmaceuticals and illicit drugs in conventional and membrane bioreactorwastewater treatment plants and by riverbank filtration. Philos Trans R Soc Lond2009;367(1904):3979–4003.

Radjenovic J, Petrovic M, Ventura F, Barceló D. Rejection of pharmaceuticals innanofiltration and reverse osmosis membrane drinking water treatment. WaterRes 2008;42(14):3601–10.

Reif R, Suarez S, Omil F, Lema JM. Fate of pharmaceuticals and cosmetic ingredientsduring the operation of a MBR treating sewage. Desalination 2008;221:511–7.

Reif R, Santos A, Judd SJ, Lema JM, Omil F. Occurrence and fate of pharmaceutical and per-sonal care products in a sewage treatment works. J Environ Monit 2011;13:137–44.

Rodríguez I, Quintana JB, Carpinteiro J, Carro AM, Lorenzo RA, Cela R. Determination ofacidic drugs in sewage water by gas chromatography–mass spectrometry astert-butyldimethylsilyl derivates. J Chromatogr A 2003;985:265–74.

Shephard RW. Theory of cost and production functions. Princeton: Princeton UniversityPress; 1970.

Snyder SA, Wert EC, Rexing DJ, Zegers RE, Drury DD. Ozone oxidation of endocrinedisruptors and pharmaceuticals in surface water and wastewater. Ozone Sci Eng2006;28(6):445–60.

Stalter D, Magdeburg A, Weil M, Knacker T, Oehlmann J. Toxication or detoxication?In-vivo toxicity assessment of ozonation as advanced wastewater treatment withthe rainbow trout. Water Res 2010;44(2):439–48.

Suarez S, Dodd MC, Omil F, von Gunten U. Kinetics of triclosan oxidation by aqueousozone and consequent loss of antibacterial activity: relevance to municipal wastewa-ter ozonation. Water Res 2007;41(12):2481–90.

Suárez S, Carballa M, Omil F, Lema J. How are pharmaceutical and personal care prod-ucts (PPCPs) removed from urban wastewaters? Rev Environ Sci Biotechnol2008;7(2):125–38.

Vanderford BJ, Pearson RA, Rexing DJ, Snyder SA. Analysis of endocrine disruptors, pharma-ceuticals, and personal care products in water using liquid chromatography/tandemmass spectrometry. Anal Chem 2003;75:6265–74.

Verlicchi P, Al Aukidy M, Zambello E. Occurrence of pharmaceutical compounds inurban wastewater: removal, mass load and environmental risk after a secondarytreatment—a review. Sci Total Environ 2012;429:123–55.

Vogna D, Marotta R, Napolitano A, Andreozzi R, d'Ischia M. Advanced oxidation of the phar-maceutical drug diclofenac with UV/H202 and ozone. Water Res 2004;38(2):414–22.

Yin Z, Wang X. The study on the effect of shadow price in emission trading 2010. ICAMS2010 — Proceedings of 2010 IEEE International Conference on Advanced Manage-ment Science, 2. ; 2010. p. 514–6. [art. No. 552975].