Doi:10.1016/j.watres.2005.09.018

Removal of cosmetic ingredients and pharmaceuticals in Marta Carballa, Francisco OmilÃ, Juan M. Lema School of Engineering, Department of Chemical Engineering, University of Santiago de Compostela, Received 5 November 2004; received in revised form 12 September 2005; accepted 12 September 2005 Two physico-chemical processes, coagulation–flocculation and flotation, have been assessed for enhancing the removal of some selected pharmaceutical and personal care products (PPCPs) present in sewage. Eight compounds,representative of three main groups of PPCPs according to their physico-chemical properties, have been selected:lipophilic compounds (the synthetic musks Galaxolide and Tonalide), neutral compounds (the tranquillizer Diazepamand the antiepileptic Carbamazepine) and acidic compounds (the anti-inflammatories Ibuprofen, Naproxen andDiclofenac). During the coagulation–flocculation assays, the main parameters considered were the selection of theadditives, their doses and the temperature of operation (12 or 25 1C). Musks—which are highly lipophilic andDiclofenac—with significant sorption affinity—were removed around 50–70% at both temperatures independently ofthe dose and type of coagulant used. However, the rest of the compounds, which are more hydrophilic, were affected toa lesser degree (with maximum reductions below 25%). The exceptions to this behavior were Carbamazepine andIbuprofen, which were not removed under any condition tested. During the flotation assays, the parameters studiedwere the initial content of fat in wastewaters and temperature. Again, musks were removed to a greater degree(35–60%), followed by Diazepam (40–50%) and Diclofenac (20–45%) and, to a lesser extent, Carbamazepine(20–35%), Ibuprofen (10–25%) and Naproxen (10–30%). The best results were always obtained at 25 1C, although insome cases the operation at 12 1C gave similar results. The removal of musks and neutral compounds was higher inwastewaters with a high fat content (around 150 mg lÀ1).
r 2005 Elsevier Ltd. All rights reserved.
Keywords: Pharmaceuticals; Musks; Sewage; Coagulation–flocculation; Flotation; Adsorption; Fat; Temperature PPCPs comprise allprescription and over-the-counter drugs, diagnostic Pharmaceuticals and personal care products (PPCPs) agents, and other consumer chemicals, such as poly- constitute a diverse group of chemicals that have cyclic musk compounds frequently used as fragrances in recently been recognized as particular contaminants of perfumes and other household products.
the aquatic environment, especially in urbanized areas Due to the large amount of PPCPs consumed in developed societies, significant concentrations of these Corresponding author. Tel.: +34 981 59 44 88x16778; E-mail address: [email protected] (F. Omil).
However, conventional sewage treatment plants 0043-1354/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2005.09.018 M. Carballa et al. / Water Research 39 (2005) 4790–4796 (STPs) have been reported not to be an effective barrier dissolved in water. However, some limitations have been to these substances because of their low concentrations and specific metabolic properties (There- ) for the applicability of these coefficients to explain fore, those compounds which resist the treatment the sorption behavior of some PPCPs. Therefore, the processes commonly used in STPs or other transforma- solid–water distribution coefficient (Kd), defined as the tions which can naturally occur in the environment, can ratio between the concentrations of a substance in the end up in surface and groundwaters, as well as in solid and in the aqueous phase at equilibrium conditions (Eq. (1)), has been proposed as the most suitable Different mechanisms, such as sorption, biodegrada- tion, volatilization and fotooxidation, can be considered for PPCPs removal in STPs. Although in many cases,the differences between them cannot be easily distin- ) have concluded that only two ofthem, microbial degradation and sorption to suspended where Kd is the solid–water distribution coefficient solids, are really relevant. The effectiveness of these removal mechanisms greatly depends on the physico- (mg PPCP kg solidÀ1); and S the concentration in the chemical properties and the chemical structure of each This coefficient takes into account the two main sorption mechanisms: absorption (hydrophobic interac-tions characterized by the Kow value) and adsorption (electrostatic interactions related to the substancetendency to be ionized or dissociated in aqueous phase, The sorption of micropollutants onto solids and, which is characterized by the dissociation constant, accordingly, their behavior during the physico-chemical treatment, depends basically on their physico-chemical properties, such as lipophilicity or acidity. Two types of PPCPs can be divided into three main groups: lipophilic coefficients have been mostly used to determine the (with high Kow values), neutral (non-ionic) and acidic sorption effectiveness and the affinity of a given (hydrophilic and ionic) compounds. Substances from substance to organic matter: the octanol-water partition different therapeutical classes and representative of each coefficient (Kow) and the organic carbon partition group have been considered in this work (): two coefficient (Koc). Kow is defined as the ratio between fragrances (Galaxolide and Tonalide), one tranquillizer the equilibrium concentrations of a certain compound in (Diazepam), one antiepileptic (Carbamazepine) and octanol and water at a specific temperature and Koc three anti-inflammatories (Ibuprofen, Naproxen and relates the concentrations sorbed to organic carbon and Table 1Physico-chemical properties of the PPCPs considered in this work (water solubility in g lÀ1; Kd in l kgÀ1) a.
bcd.
efThe Kd value for Naproxen has been assumed based on its similar properties with Ibuprofen and other Kd measurements in digested M. Carballa et al. / Water Research 39 (2005) 4790–4796 Galaxolide and Tonalide are very lipophilic com- as the type and dose of coagulant, the fat content of the pounds with log Kow values around 5.5–6.0. However, wastewaters and the temperature has been studied.
while Carbamazepine and Diazepam are neutral sub-stances, as derived from their chemical structure, theanti-inflammatories are very acidic compounds (low pKa The wastewaters used in this work were collected from Coagulation–flocculation processes enhance the re- an urban STP located in Santiago de Compostela (NW moval of suspended solids and colloids, because the of Spain). The STP, which was surveyed in a previous addition of metal salts or organic compounds causes the agglomeration of these particles, thus allowing their inhabitants approximately and comprises three main elimination by decantation or filtration sections: pre-treatment (coarse/fine screening and grit/fat removal), primary treatment (sedimentation) and biolo- Lipophilic trace pollutants in water and wastewater gical treatment (conventional activated sludge). The inlet treatment systems are likely to be found associated with flow to the primary clarifier was used for coagulation– colloids because in natural systems most colloids have flocculation experiments, whereas the inlet to the fat separator was used for flotation assays. The main addition, positive charged molecules can be associated characteristics of the wastewater are: total solids (TS), to these colloids by means of low strength Van der 500–900 mg lÀ1; volatile solids (VS), 200–500 mg lÀ1; total suspended solids (TSS), 100–400 mg lÀ1; volatile sus- Literature information about the removal of PPCPs pended solids (VSS); 100–300 mg lÀ1; total chemical by coagulation–flocculation processes is scarce. When oxygen demand (CODt), 200–800 mg lÀ1; soluble chemi- some data is available, it is related to either a post- cal oxygen demand (CODs), 100–500 mg lÀ1; and fat treatment, and normally they are combined with othertechnologies, such as activated carbon or filtration The PPCPs used in this work were Galaxolide, Flotation techniques, in which finely suspended parti- Tonalide, Carbamazepine, Diazepam, Ibuprofen, Na- cles are separated by adhering to the surface of rising proxen and Diclofenac. Two solutions, one containing bubbles, have proved to be efficient, practical and musks plus the neutral pharmaceuticals and the other reliable methods for the removal of fat, as well as other one with the acidic substances, were spiked to 10 L of contaminants, such as oils, biomolecules or suspended urban wastewater in order to attain higher levels than Besides, micropollutants like lipophilic PPCPs can be Once prepared, the resulting PPCPs concentrations were removed from the wastewaters by flotation due to their solubilization in the lipid fractions or sorption onto background content (already present in sewage) and the small aggregates. For instance, associated the removal of Carbamazepine in a STP with thepresence of an unusual high content of silicone oil in the Measured concentrations of PPCPs in the spiked samples ofurban wastewater used in the coagulation–flocculation and The aim of this work is to improve the removal efficiencies of three groups of PPCPs (musks, neutral and acidic pharmaceuticals), which have differentsorption properties, during sewage primary treatment by coagulation–flocculation and flotation processes.
This objective is based on the hypothesis that the distribution of PPCPs between the solids and the aqueous phase can be modified by the addition of some chemicals (coagulants, flocculants, tensoactives, etc.).
The influence of the main operational parameters, such M. Carballa et al. / Water Research 39 (2005) 4790–4796 ethyl acetate. This extract was then divided into twofractions: one of them being used for the direct Coagulation–flocculation assays were carried out in a determination of the soluble content of Carbamazepine, Jar-Test device, in vessels of 1 l of liquid volume. The Diazepam and fragrances; the other for the determina- influence of three additives was studied: ferric chloride tion of the soluble content of the anti-inflammatories. In (FeCl3, 50 g lÀ1), aluminum sulfate (Al2(SO4)3, 50 g lÀ1) the latter case, compounds were silylated previously to and aluminum polychloride (PAX, 17.5% w/w). The assays were conducted at two temperatures, 12 and ). In both cases, GC/MS was used to determine the 25 1C, simulating winter and summer conditions, respec- concentration of the investigated compounds in the tively. The test included an initial 3 min period of rapid SPE extract. Values given for the different samples stirring (150 rpm), after the addition of the coagulant correspond to the average value of two aliquots of and lime for neutralization, followed by 5 min of slow mixing (50 rpm) for emulsion breaking and floc forma-tion, and finally 1 h period without mixing, for flocseparation. The influence of the type and dose of coagulant and the temperature was studied. Since theobjective of the work was to enhance PPCPs removal during sewage primary treatment, all the experimentswere carried out at the neutral pH necessary for the Preliminary assays with FeCl3, Al2(SO4)3 and PAX were performed at 12 and 25 1C without PPCPs additionin order to adjust the dose range for each coagulant.
Al2(SO4)3 (100–500 mg lÀ1) and PAX (250–1250 mg lÀ1) Flotation assays were carried out in a unit consisting were tested and the parameters monitored were the TSS of a pressurized vessel of 2 l (where air was dissolved into and CODt concentrations in the supernatant. Although the wastewater) and a flotation cell of 1 l ( the differences were not significant in the range ). The pressurized cell has two inlets (for air considered, it was observed (data not shown) that the and water), and one outlet for the pressurized liquid.
higher removal efficiencies for TSS and CODt were Also, a manometer was set up in the air line to check the achieved in the following dose range: 200–300 mg lÀ1 pressure. The dissolved air was then introduced into the (FeCl3), 250–350 mg lÀ1 Al2(SO4)3 and 700–950 mg lÀ1 flotation cell where the fine air bubbles produced by depressurization helped flocs flotation. The influence of Afterwards, the influence of coagulant dose (in those the content of fat in wastewaters and the temperature ranges) and temperature on PPCPs removal was was studied. The assays were carried out in duplicate.
analyzed. From the results obtained (data not shown), Two types of wastewater with different concentrations it can be concluded that there is no significant influence of fat were used: a low fat (LF) and high fat (HF) (less than 5%) either of the coagulant dose or of the wastewater, with approximately 60 and 150 mg lÀ1, temperature (12 or 25 1C) on PPCPs removal in the respectively. While the LF wastewater was directly considered range. Because of that, the following assays taken from the STP considered, the HF wastewater were carried out at 25 1C with the following coagulant was synthetically prepared by adding fat (as liquid concentration: 250 mg FeCl3 lÀ1, 300 mg Al2(SO4)3 lÀ1 butter) to the LF wastewater, in order to evaluate and 850 mg PAX lÀ1. Besides, a blank assay (an exclusively the influence of the wastewater fat content.
experiment without any additive) was carried out tomonitor the removal of these compounds merely associated with the sedimentation of solids in thebeakers. The results obtained are summarized in TS, VS, TSS, VSS, COD and fat were analyzed Except for Carbamazepine and Ibuprofen, which were not affected by the addition of any coagulant, determined using a selective electrode and temperature shows that the use of an additive increased the removal with a digital thermometer. The soluble content of the efficiencies of all PPCPs tested. In the case of musks, fragrances, anti-inflammatories, Carbamazepine and while ferric chloride and aluminum sulfate lead to Diazepam was determined after a solid-phase extraction similar eliminations of both substances (around 50%), (SPE) of 500 ml samples using 60 mg OASIS HLB the use of aluminum polychloride improved the removal cartridges (Waters, Milford, MA, USA). Meclofenamic efficiencies of each: 63% for Galaxolide and 71% for acid and dihydrocarbamazepine were added to the Tonalide. Conversely, the elimination of Diclofenac was samples as surrogate standards. All compounds were higher with ferric chloride and aluminum sulfate quantitatively eluted from the cartridge using 3 ml of (around 70%), although PAX also gave a significant M. Carballa et al. / Water Research 39 (2005) 4790–4796 coagulant enhances the binding of Diclofenac to thesuspended solids throughout the trivalent cations, thus allowing a further removal from the water phase.
Diazepam and Naproxen removal is also improved by the action of coagulants (20–25%), although in a lower extent than Diclofenac, which can be explained by their Finally, Carbamazepine and Ibuprofen were not eliminated at any tested conditions, which is in accordance with their very low Kd values.
Preliminary assays were carried out to determine the pressurized liquid flow necessary to produce fat separa-tion in the flotation cell. This value was adjusted to 200 ml operating inside the pressurized cell at 6.4 atm.
These conditions implied the following air–solid ratios Fig. 1. Removal efficiencies from the aqueous phase obtained (A/S): 0.07 (12 1C) and 0.01 (25 1C).
during the coagulation–flocculation assays.
The effect of the initial content of fat in wastewaters and temperature (12 and 25 1C) was studied. Two types reduction (around 50%). The concentrations of Diaze- of wastewaters with different concentrations of fat were pam and Naproxen were reduced by 20–25%. While for used: a LF and HF wastewater, with approximately 60 Diazepam there were no significant differences between and 150 mg lÀ1. The assays were carried out in duplicate.
ferric chloride and aluminum sulfate, Naproxen was shows the results obtained for the different only removed with ferric chloride. In both cases, PAX PPCPs considered when LF wastewaters were used. A was the less effective additive (below 5%).
similar behavior between both musks can be observed: The different behavior obtained in coagulation–floc- culation assays for each compound can be explained by substantially reduced at both temperatures (35–45%), their different physico-chemical properties. In this way, with the highest removal by efficiencies being obtained the good removal of musks is concordant with their high at 25 1C. The elimination of Diazepam was similar to ability to attach to solid particles (log Kd values between that obtained for musks (40–45%), although no 3.3 and 3.7), mainly due to hydrophobic interactions significant difference was observed between both tem- with the lipid fractions of the sludge cells membrane peratures. However, according to its lower lipophilicity (log Kow around 2.4), Carbamazepine was removed The maximum removal efficiency expected for a given to a lesser extent (around 20%) independently of the compound can be estimated from its distributionbetween the solid (Eq. (2)) and the aqueous (Eq. (3)) phases, using the Kd value and the PPCP concentrationin the aqueous (S) and solid (X) phase. For musks, these maximum values ranged from 60% to 85%, very close to those obtained in these experiments (50–70%): When applying the same methodology for calculating the elimination of Diclofenac, which could be obtained after a coagulation–flocculation treatment, just con- sidering its Kd value (log Kd ranged from 1.2 to 2.7), theforeseen figures (15–40%) appear to be quite lower that those obtained in the experiments (50–70%). This can bedue to the acidic nature of this compound (pKa$4), Fig. 2. Removal efficiencies from the aqueous phase obtained which in aqueous phase remains partially ionized. The during the flotation assays with low fat (60 mg lÀ1) wastewaters.
M. Carballa et al. / Water Research 39 (2005) 4790–4796 temperature. The anti-inflammatories were also affected pine and Diazepam, and regardless of the initial fat by flotation, the highest removals being those obtained content, no effect was observed. However, for anti- for Diclofenac (20–40%). For these three compounds, inflammatories, the best results were obtained at 25 1C temperature influenced removal significantly and, as for with both LF and HF wastewaters. In the case of musks, musks, the highest values were obtained at 25 1C.
while higher removal efficiencies were attained at 25 1C shows the results obtained for the different PPCPs studied when HF wastewaters were used. It can differences between both temperatures were observed be observed that the elimination of musks is higher (around 60%) under these conditions and that tempera-ture did not significantly influence removal. Thisbehavior was also observed for Carbamazepine andDiazepam, with removals increasing to 35% and 50%, respectively. Once again, these rates were uninfluencedby temperature. Since the soluble content of the anti- Compounds with high sorption properties (high inflammatories was independent on the fat content in log Kd values), such as musks (Galaxolide and Tonalide) the wastewaters, their removal patterns were similar to and Diclofenac, are significantly removed during coa- those observed in the assays with LF wastewaters: gulation–flocculation with efficiencies of 70% in the 20–45% for Diclofenac, 10–30% for Naproxen and temperature range of 12–25 1C. Lipophilic compounds, 10–20% for Ibuprofen. Temperature clearly influence like musks, are mainly absorbed on the lipid fractions of removal efficiencies, since the best results were obtained the sludge, while acidic compounds, like Diclofenac, are mainly adsorbed due to electrostatic interactions.
The different affinities of PPCPs for organics can be Compounds with lower Kd values, such as Diazepam, clearly seen when HF wastewaters were used. While the Carbamazepine, Ibuprofen and Naproxen, were reduced removal of lipophilic substances, such as musks, is to a lesser extent (Diazepam and Naproxen), up to 25%, enhanced, the elimination of more polar compounds or not affected at any condition tested (Carbamazepine remains at the same level. Furthermore, there are no and Ibuprofen). Although PAX gives the best results for significant differences in the removal of these substances musks, the option of ferric chloride appears to be the when LF or HF wastewaters are used. This shows that most suitable since the concentration of PAX required is not only the physico-chemical properties of the PPCPs has to be considered, but also the presence of the other All substances were removed during flotation assays substances in the medium, such as the fat globules, the with higher efficiencies when HF wastewaters were used colloidal matter or the flocs formed during coagula- (around 60% for musks, 35% for Carbamazepine, 50% for Diazepam and 20–45% for the anti-inflammatories).
The influence of temperature in the various flotation For musks, Carbamazepine and Diazepam, temperature assays depends on the type of PPCP. For Carbamaze- was not very important. However, in the case ofDiclofenac, Naproxen and Ibuprofen, the best resultswere attained at 25 1C, independently of wastewater fat content. Although the results obtained are quitesatisfactory for all PPCPs, it is possible that for some of them removal could be improved by adding some chemical which modifies surface properties.
Taking into account that some PPCPs, as well as other micropollutants present in sewage, appear to be not readily biodegradable, enhancing their removal in the sewage primary treatment could be an interesting strategy for minimizing costs in the biological and This work was supported by the EU (POSEIDON project, EVK1-CT-2000-00047) and the Spanish Minis- Fig. 3. Removal efficiencies from the aqueous phase obtainedduring the flotation assays with high fat (150 mg lÀ1) waste- ter of Education and Science (FARMEDAR project, M. Carballa et al. / Water Research 39 (2005) 4790–4796 Paxeus, N., 2004. Removal of selected non-steroideal anti- inflammatory drugs (NSAIDs), gemfibrozil, carbamazepine, APHA–AWWA–WEF, 1999. In: Clesceri, L.S., Greenberg, b-blockers, trimethoprim and triclosan in conventional A.E., Eaton, A.D. (Eds.), Standard Methods for the wastewater treatment plants in five EU countries and their Examination of Water and Wastewater, 20th ed. American discharge to the aquatic environment. Water Sci. Technol.
Public Health Association, American Water Work Associa- tion, Water Environment Federation, Washington, DC.
POSEIDON Project (EVK1-CT-2000-00047), 2005. Assessment Boyd, G.R., Reemtsma, H., Grimm, D.A., Mitra, S., 2003.
of technologies for the removal of pharmaceuticals and Pharmaceuticals and personal care products (PPCPs) in personal care products in sewage and drinking water surface and treated waters of Louisiana, USA and Ontario, facilities to improve the indirect potable water reuse. Final Canada. Sci. Total Environ. 311, 135–149.
Carballa, M., Omil, F., Lema, J.M., Llompart, M., Garcı´a- Rodriguez, I., Quintana, J.B., Carpinteiro, J., Carro, A.M., Jares, C., Rodrı´guez, I., Go´mez, M., Ternes, T., 2004.
Lorenzo, R.A., Cela, R., 2003. Determination of acidic Behavior of pharmaceuticals, cosmetics and hormones in a drugs in sewage water by gas chromatography-mass sewage treatment plant. Water Res. 38, 2918–2926.
spectrometry as tert.-butyldimethylsilyl derivates. J. Chro- Cargoue¨t, M., Perdiz, D., Mouatassim-Souali, A., Tamisier- Karolak, S., Levi, Y., 2004. Assessment of river contamina- Romero, J., Ventura, F., Folch, M., Salgot, M., Torrens, A., tion by estrogenic compounds in Paris area (France). Sci.
2003. Characterisation of volatile organic contaminants after different pretreatment systems in reclaimed waste- Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and water. Water Sci. Technol. 3 (4), 139–143.
personal care products in the environment: agents of subtle Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 2003.
change? Environ. Health Persp. 107, 907–938.
Environmental Organic Chemistry. Wiley, New York, Holbrook, R.D., Love, N.G., Novak, J.T., 2004. Sorption of Chichester, Brisbane, Toronto, Singapore.
17b-estradiol and 17a-ethinylestradiol by colloidal organic Stackelberg, P.E., Furlong, E.T., Meyer, M.T., Zaugg, S.D., carbon derived from biological wastewater treatment Henderson, A.K., Reissman, D.B., 2004. Persistence of systems. Environ. Sci. Technol. 38 (12), 3322–3329.
pharmaceutical compounds and other organic wastewater Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2002. Aquatic contaminants in a conventional drinking water treatment environmental assessment of the top 25 English prescription plant. Sci. Tot. Environ. 329, 99–113.
pharmaceuticals. Water Res. 36, 5013–5022.
Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry—Chemical Kanda, R., Griffin, P., James, H.A., Fothergill, J., 2003.
Equilibria and Rates in Natural Waters, third ed. Wiley- Pharmaceuticals and personal care products in sewage treatment works. J. Environ. Monit. 5, 823–830.
Syracuse Research Corporation, 2004. Physprop database.
Kolpin, D.W., Skopec, M., Meyer, M.T., Furlong, E.T., Zaugg, S.D., 2004. Urban contribution of pharmaceuticals and other organic wastewater contaminants to streams during sewage treatment plants and rivers. Water Res. 32 (11), differing flow conditions. Sci. Tot. Environ. 328, 119–130.
Kummerer, K., 2000. Pharmaceuticals in the Environment: Ternes, T.A., Meisenheimer, M., McDowell, D., Sacher, F., Sources, Fate, Effects and Risks, second ed. Springer, Germany.
Brauch, H.J., Haiste-Gulde, B., Preuss, G., Wilme, U., Lai, K.M., Johnson, K.L., Scrimshaw, M.D., Lester, J.N., Zulei-Seibert, N., 2002. Removal of pharmaceuticals during 2000. Binding of waterborne steroid estrogens to solid drinking water treatment. Environ. Sci. Technol. 36, phases in river and estuarine systems. Environ. Sci. Technol.
Ternes, T.A., Herrmann, N., Bonerz, M., Knacker, T., Siegrist, Li, G., Gregory, J., 1991. Flocculation and sedimentation of H., Joss, A., 2004. A rapid method to measure the solid- high turbidity waters. Water Res. 25, 1137–1143.
water distribution coefficient (Kd) for pharmaceuticals and Liebig, M., 2004. Cooperation in POSEIDON project (EVK1- musk fragrances in sewage sludge. Water Res. 38 (19), CT-2000-00047). ECT Oekotoxikologie GmbH (personal Zouboulis, A.I., Avranas, A., 2000. Treatment of oil-in-water Metcalf Eddy, 1991. Wastewater Engineering: Treatment, Dis- emulsions by coagulation and dissolved-air flotation.
posal and Reuse, third ed. McGraw-Hill, New York, USA.

Source: http://www.totagua.es/pdf/tecnologias/Removal_of_cosmetic_ingredients_and_pharmaceuticals_in_sewage_primary_treatment.pdf