Production of adsorbents by pyrolysis of paper mill sludge and application on the removal of citalopram from water

Bioresource Technology 166 (2014) 335–344

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Production of adsorbents by pyrolysis of paper mill sludge and application on the removal of citalopram from water Vânia Calisto a,⇑, Catarina I.A. Ferreira a, Sérgio M. Santos b, María Victoria Gil c, Marta Otero d, Valdemar I. Esteves a a

Department of Chemistry and CESAM (Centre for Environmental and Marine Studies), University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal Department of Chemistry and CICECO (Centre for Research in Ceramics and Composite Materials), University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal Instituto Nacional del Carbón, INCAR-CSIC, Apartado 73, 33080 Oviedo, Spain d Department of Applied Chemistry and Physics, IMARENABIO, University of Léon, Campus de Vegazana, Léon, Spain b c

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Pyrolysis was applied for adsorbent

production for paper mill residues valorization.  Pyrolysed paper mill sludge was tested for the removal of an antidepressant from water.  Adsorbents were produced by environmentally friendly methods without activation steps.  Primary sludge resulted in higher porous and efficient adsorbents than biological sludge.  Best results were obtained for primary sludge pyrolysed at 800 °C for 150 min.

a r t i c l e

i n f o

Article history: Received 11 April 2014 Received in revised form 13 May 2014 Accepted 14 May 2014 Available online 23 May 2014 Keywords: Biochar Pharmaceuticals Environment Remediation Industrial residues

a b s t r a c t This work describes the production of alternative adsorbents from industrial residues and their application for the removal of a highly consumed antidepressant (citalopram) from water. The adsorbents were produced by pyrolysis of both primary and biological paper mill sludge at different temperatures and residence times. The original sludge and the produced chars were fully characterized by elemental and proximate analyses, total organic carbon, specific surface area (BET), N2 isotherms, FTIR, 13C and 1H solid state NMR and SEM. Batch kinetic and equilibrium experiments were carried out to describe the adsorption of citalopram onto the produced materials. The fastest kinetics and the highest adsorption capacity were obtained using primary sludge pyrolysed at 800 °C during 150 min. The use of pyrolysed paper mill sludge for the remediation of contaminated waters might constitute an interesting application for the valorization of those wastes. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction

⇑ Corresponding author. Tel.: +351 234401408. E-mail address: [email protected] (V. Calisto). http://dx.doi.org/10.1016/j.biortech.2014.05.047 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

Paper mill sludge is produced in the order of eleven million tons per year only by European mills (Monte et al., 2009). The use of pulp and paper production sub-products is quite dependable on the legislation applied in each country; however, in general, main applications include energy recovery, disposal on landfills and

336

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

composting (CEPI, 2004). Taking into account that some of those traditional disposal options are being progressively restricted, the management of these wastes in an economically and environmentally acceptable manner is a critical issue for the mills. Hence, from a sustainable point of view, it is necessary to develop innovative handling strategies that maximize recovery of useful materials and/or energy and that also allow the minimization of such wastes (Stoica et al., 2009). Taking into account that these residues are derived from natural resources (wood) and are a result of controlled and well known processes, their use as raw materials in other industries has been increasing (such as cement, brick and tile manufacturing), in opposition to the most conventional applications referred above (CEPI, 2004). The contamination of sewage treatment plant’s (STPs) effluents with pharmaceuticals is well documented and considered to be generalized and, consequently, the discharge of those effluents is pointed out to be the primary source of pharmaceutically active ingredients into the environment (Kinney et al., 2006; Jelic et al., 2011). Taking into account the scarcity of potable water resources that the world has been facing and the need to maintain an unpolluted environment, it is of the utmost importance to develop effective strategies that mitigate this continuous contamination. Considering the need of applying those removal strategies at large scales, economically feasible and high-efficient low-cost strategies ought to be a priority (Chong et al., 2010). For pharmaceuticals, the most promising results are usually obtained with advanced treatment methods such as advanced oxidation processes (ozonation, Fenton, photo-Fenton), reverse osmosis and membrane bio-reactors (Silva et al., 2012). These methods have the major drawback of being cost consuming and, in some cases, resulting on the formation of undesirable by-products (such as nitrosamines, formaldehyde, chlorinated products) (Chong et al., 2010; Margot et al., 2013; Rivera-Utrilla et al., 2013). In this context, removal by adsorption onto solid matrices is a very promising option due to its effectiveness and versatility. Activated carbons (in powdered and granular forms) are the most frequently used adsorbents and, in general, result in very satisfactory removal rates (Aksu and Tunç, 2005; Yu et al., 2008, 2009). However, using commercial activated carbons is quite expensive (Aksu and Tunç, 2005). As a consequence, there are various works reported in the literature describing the production (and subsequent application) of alternative low-cost adsorbents produced by pyrolysis, generally combined with chemical or physical activation of industrial wastes (Yao et al., 2012), sewage sludge (Smith et al., 2009) and agricultural residues (Mohan et al., 2014; Ioannidou and Zabaniotou, 2007; Antunes et al., 2012), to mention a few. Yet, until now, very few studies reported the production of adsorbents using paper mill sludge (Devi and Saroha, 2014; Khalili et al., 2002; Li et al., 2011) and there are no data describing the use of primary sludge, in particular. In this work, primary and biological paper mill sludge is used for the production of alternative adsorbents. The main goal is to obtain a carbon with high adsorption capacity using cheap and environmentally friendly production methods (without employing chemical or physical activation) and, simultaneously, to propose a new way to valorize this industrial sub-product. The produced adsorbents are here applied on the removal of the antidepressant citalopram from water which is, to the best of our knowledge, the first description of an application of paper sludge derived adsorbents to the adsorption of a pharmaceutical. The antidepressant citalopram was selected for this study due to its high consuming patterns (caused by the large prevalence of stress related conditions and mental diseases (OECD, 2011)), the frequency of occurrence in the environment and its ability to interfere in the regulation of behavior and neuro-endocrine signaling of aquatic non-target organisms (Calisto and Esteves, 2009).

2. Methods 2.1. Chemicals Citalopram hydrochloride (>98%) was purchased from TCI Europe. All chemicals used for capillary electrophoresis were of analytical grade: sodium dodecylsulphate (SDS, 99%, for electrophoresis, Sigma Aldrich), hexadimethrine bromide (polybrene, Sigma Aldrich), sodium chloride, ethylvanillin (99%, Sigma Aldrich), sodium tetraborate (Riedel-de Haën), sodium hydroxide (Fluka). All solutions were prepared using ultra-pure water, obtained from a Milli-Q Millipore system (Milli-Q plus 185). 2.2. Adsorbent materials Primary (PS) and biological paper mill sludge (BS) were provided by a mill which employs the kraft ECD (elemental chlorine free) pulp production process. The mill operates exclusively with eucalyptus wood (Eucalyptus globulus). On average, PS and BS are produced at a rate of 20 and 10 kg per ton of air dried pulp, respectively. The PS results from fibers rejected after the cooking/digestion pulping step and losses of fibers and other solids which occur when liquid effluents are involved (for example, washing and bleaching); the composition of the PS is very similar to the pulp, consisting essentially of organic matter (fibers). After primary treatment, the effluent is then submitted to biological treatment: the BS is essentially biomass (after dehydration) that results from the action of microorganisms, under aerobic conditions, which are meant to reduce the organic matter content of the effluent. After collection, PS and BS were dried at room temperature for several days, followed by a 24 h period at 60 °C in an oven. BS was then grinded with a mortar grinder and separated by grain size (<0.18 mm and between 0.18 and 0.5 mm, referred as 0.18 mm and 0.5 mm, respectively). The same procedure was not applied to PS due to its particular physical characteristics: it was not effectively grinded by a mortar grinder; instead, a blade mill was used, resulting in an extremely light net of fibrous material, impossible to sieve. PS and BS were then pyrolysed into porcelain crucibles using a muffle (Nüve, series MF 106, Turkey). The pyrolysis was carried out at a heating rate of 10 °C min1, under N2 saturated atmosphere (N2 flow of 0.5 dm3 min1). The final pyrolysis temperature and residence time was varied in order to evaluate the influence of those parameters in the adsorption capacity of the material. The temperatures were chosen according to the mass losses observed in the thermogravimetric analysis (see Sections 2.3.1 and 3.1 and Fig. S1 of Supporting Information (SI)). Accordingly, PS and BS were pyrolysed at 315 °C for 150 min (PS315-150 and BS315-150), at 600 °C for 10 min (PS600-10 and BS600-10), at 800 °C for 10 min (PS800-10 and BS800-10) and finally PS was also pyrolysed at 800 °C for 150 min (PS800-150). After the pyrolysis, the pyrolysed sludge was maintained inside the muffle until the temperature reached room temperature; the nitrogen stream was continuously applied during the cooling step. Apart from the paper mill sludge based adsorbents, and given that, to the best of our knowledge, there are no data in the literature on the adsorptive removal of citalopram, the commercial powdered activated carbon PULSORB FG4 (PBFG4), provided by Chemviron Carbon, was used for comparison purposes. 2.3. Materials characterization 2.3.1. Thermogravimetric analysis Thermal characterization of PS and BS was carried out by thermogravimetric (TG) analysis and derivative thermogravimetric

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

analysis (DTG). Non-isothermal TG analysis was performed using a TAG analyser (Setaram Instrumentation, TAG24, France). The analyses were carried out under a 50 cm3 min1 N2 flow at a heating rate of 10 °C min1 from room temperature to 1000 °C. Approximately 5 mg of sample was used for each experiment. The mass loss (TG) and derivative curves (DTG) of the samples were represented as a function of temperature, to allow the selection of the pyrolysis temperatures. 2.3.2. Proximate and ultimate analyses and heating value The proximate and ultimate analyses, as well as the heating value determination, were applied to PS, BS and all the pyrolysed materials. Proximate analysis was conducted in a TGA-601 automatic analyzer (LECO, model TGA601, Spain). Standard methods were employed to determine the moisture (UNE 32002) (AENOR, 1995), volatile matter (UNE 32019) (AENOR, 1985) and ash contents (UNE 32004) (AENOR, 1984). The fixed carbon content was determined as the remaining fraction after ash and volatile matter (both at a dry basis) determination. Ultimate analysis, involving the determination of the sample content in C, H, N, S and O, was performed in a CHNS-932 analyzer (LECO, model CHNS-932, Spain). The oxygen content was calculated by difference. The results were corrected considering the water content of the samples (presented at a dry basis); the contribution of water for the percentage of O and H was also taken into account. The higher heating value (HHV) of the samples was measured in a calorimetric pump IKA 4000 according to the guideline ISO 1928 (ISO, 2009). The lower heating value (LHV) was determined according to LHV = HHV – 49  Hdb (with Hdb corresponding to the hydrogen mass percentage on a dry basis). 2.3.3. Total organic carbon (TOC) Total carbon (TC) and inorganic carbon (IC) analyses were performed using a TOC analyzer (Shimadzu, model TOC-VCPH, SSM5000A, Japan) on PS, BS, all the pyrolysed materials and PBFG4. All the materials were analyzed in triplicate. Total organic carbon was calculated by difference. 2.3.4. Fourier transform infra-red spectroscopy (FTIR) PS, BS, all the corresponding pyrolysed materials and PBFG4 were characterized by FTIR. The spectra were performed in a FTIR spectrophotometer, using an attenuated total reflectance (ATR) module (Shimadzu, IRaffinity-1, Japan), with nitrogen purge. FTIR measurements were recorded in the range of 600–4000 cm1, 4.0 of resolution, 128 scans and were corrected against background and atmosphere. 2.3.5. 13C CP-MAS NMR and 1H MAS NMR NMR analyses were performed on PS, BS and all the pyrolysed sludges. All spectra were acquired on a Bruker Avance III 400 spectrometer operating at a B0 field of 9.4 T, corresponding to Larmor frequencies of 400.1 (1H) and 100.6 MHz (13C), using a spinning rate of 10 kHz. The 1H and 13C 90° pulses were set to 3.75 and 3.80 ls (RF field strength of 66.7 and 65.8 kHz), respectively. The CPMAS experiments were acquired with the 50–100% RAMP-CP shape using a contact time of 3.0 ms and a 1H RF field strength of 114 kHz. The time between scans was set to 60 s and 5 s for the MAS and CPMAS experiments, respectively. All spectra were referenced with respect to glycine (NH+3 at 8.35 ppm and C = O at 176.03 ppm). 2.3.6. Physical characterization The physical characterization, comprising porosimetry and N2 adsorption, was carried out with all the pyrolysed materials and PBFG4. The apparent density (qHg) was determined with mercury at 0.1 MPa in a mercury porosimeter (Micromeritics, Autopore IV

337

9500, USA). The textural properties were studied by physical adsorption of N2 at 196 °C on a ASAP 2420 apparatus (Micromeritics, ASAP 2420, USA). The samples were outgassed overnight at 100 °C under vacuum prior to adsorption measurements. The specific surface area (SBET) was calculated from the Brunauer– Emmett–Teller equation (Brunauer et al., 1938) in the relative pressure range 0.01–0.1. The total pore volume (Vp) was estimated from the amount of nitrogen adsorbed at a relative pressure of 0.99. The total volume of micropores (W0) was determined by applying the Dubinin–Radushkevich equation (Dubinin, 1966) to the lower relative pressure zone of the N2 adsorption isotherm. The average micropore width (L) was calculated by means of the Stoeckli–Ballerini equation (Stoeckli and Ballerini, 1991). The average pore diameter (D) was calculated as D = 2Vp/SBET. Furthermore, the surface morphology of PS, BS, PS800-10, PS800-150 and BS800-10 was observed by scanning electron microscopy (SEM) (Hitachi, SU-70, Japan) at 500, 3,000, 10,000 and 30,000. 2.4. Micellar electrokinetic chromatography (MEKC) analyses The quantification of citalopram in water was carried out by MEKC analyses using a P/ACE MDQ capillary electrophoresis (Beckman, P/ACE MDQ, USA), equipped with a UV–vis photodiode array detection system. A dynamically coated silica capillary was used as described in previous work (Calisto et al., 2011) (see SI for a brief description). 2.5. Batch adsorption experiments All the adsorption studies (kinetic and equilibrium experiments) were performed using a batch experimental approach. Citalopram test solutions and adsorbents were put in contact in 15 or 45 mL polypropylene tubes. Batch experiments were carried out in an overhead shaker (Heidolph, Reax 2) at 50 rpm and under controlled temperature (25 °C). Each experiment included a control where the pharmaceutical solution was shaken without the presence of the adsorbent. All tests were performed in triplicate. Prior to kinetic and equilibrium experiments, preliminary tests were performed in order to guarantee that no adsorption of citalopram onto the polypropylene tubes occurs and that the MEKC analyses is not affected by PS and BS adsorbents matrix effects. For this purpose, the adsorbents were shaken with ultra-pure water (in the absence of the pharmaceutical) and the matrix obtained after filtration was spiked with a known concentration of citalopram and analyzed by MEKC. 2.5.1. Adsorption kinetics of citalopram In order to determine the time needed to attain the adsorption equilibrium, a 5 mg L1 citalopram solution was shaken during increasing time intervals (between 5 and 240 min) with each pyrolysed PS and BS, at a final adsorbent concentration of 0.5 and 5 g L1, respectively. The same procedure was repeated with PBFG4 but with an adsorbent concentration of 0.05 g L1. The remaining concentration of citalopram in solution after each contact time was then determined by MEKC. Experimental data was then fitted to pseudo-first order (1) and pseudo-second order kinetic models (2), according to:

qt ¼ qe ð1  ek1 t Þ qt ¼

q2e k2 t 1 þ qe k2 t

ð1Þ ð2Þ

with t (min) representing the adsorbent/solution contact time, qt (mg g1) the amount of solute adsorbed by mass unit of adsorbent

338

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

at time t, qe the amount of solute adsorbed when the equilibrium is attained (mg g1) and k1 (min1) and k2 (mg g1 min) the pseudofirst and pseudo-second order rate constant, respectively. 2.5.2. Adsorption equilibrium of citalopram After determining the equilibrium contact time for each adsorbent, 5 mg L1 citalopram solutions were shaken with different adsorbent concentrations, varying from 0.1 to 2.5 g L1 for PS adsorbents, from 0.5 to 50 g L1 for BS adsorbents and from 0.02 to 0.07 g L1 for PBFG4. After reaching the equilibrium time, the concentration of citalopram for each adsorbent concentration was determined by MEKC. The obtained data were fitted to three non-linear models, commonly used to describe adsorption processes: the Freundlich Eq. (3), an empirical model which is adequate to describe heterogeneous adsorption of the solute onto the adsorbent surface; the Langmuir Eq. (4) which assumes homogeneous monolayer adsorption and the Langmuir–Freundlich Eq. (5) which is a combined approach between the two former models.

qe ¼ K F C e1=N qe ¼

qe ¼

q2m kL C e 1 þ K LCe qm K LF C 1=N e 1 þ K LF C 1=N e

ð3Þ ð4Þ

ð5Þ

with qe representing the amount of solute adsorbed at equilibrium (mg g1), Ce the amount of solute in the aqueous phase at equilibrium (mg L1), KF the Freundlich adsorption constant (mg g1 (mg L1)-N), N the degree of non-linearity, qm the maximum adsorption capacity (mg g1) and KL (L mg1) and KLF the Langmuir and Langmuir–Freundlich affinity coefficients, respectively. 3. Results and discussion 3.1. Paper mill sludge pyrolysis The thermogravimetric characterization of PS and BS was performed in order to select the optimum temperatures for the production of the pyrolysed adsorbents. The results, presented in Fig. S1 of SI, show that PS and BS have a distinct behavior: PS has two main mass losses corresponding to DTG peaks at approximately 335 °C (decomposition of the most thermo-labile fraction of organic matter) and at 755 °C (possibly the decomposition of the most thermo-resistant organic matter and carbonates); BS has four significant mass losses corresponding to DTG peaks at 50 and 310 °C (water evaporation and decomposition of organic matter) and 720 and 950 °C (decomposition of thermo-resistant organic matter along with decomposition of carbonates). The PS major mass loss corresponds to the peak at 755 °C while the major mass loss of BS occurs during the decomposition of organic matter at 310 °C. In addition, PS and BS also differ in the total mass loss which is approximately 42% and 74% at 900 °C, respectively. In order to compare the effect of pyrolysis temperature on the adsorption capacity of the adsorbents, three temperatures were selected: 315, 600 and 800 °C. Despite the differences observed in the thermogravimetric analyses, the chosen temperatures were used for both PS and BS to allow a direct comparison of the produced adsorbents. The pyrolysis yield varied from 58% at 315 °C to 34% at 800 °C for both sludges. 3.2. Materials characterization The proximate and ultimate analyses of PS, BS and all the resulting adsorbents after pyrolysis are summarized in Table 1. The

results revealed that PS and BS have a very distinct chemical composition which is then reflected in the composition of the pyrolysed sludges. PS possesses more than twice the amount of ashes of BS. However, with increasing pyrolysis temperature, the percentage of ash does not increase significantly for PS but doubles in the case of BS (for BS800-10); consequently, the several PS800 and BS800 adsorbents have similar percentages of ash, approximately between 50% and 60%. As expected, the percentage of volatile matter decreases significantly with increasing pyrolysis temperature; nevertheless, BS suffers a much more accentuated release of volatile matter during pyrolysis (reducing 4 times the volatile percentage from the starting material to BS800-10) while the release of volatiles for PS is only significant for PS800-150. The same conclusion is valid for the fixed carbon content: PS and BS have very similar percentages but the increase of its percentage after pyrolysis is much more evident for the BS. Following the evolution of the ratio between the volatile matter and fixed carbon percentages, it is possible to conclude that this ratio evolved more favorably for the BS adsorbents, decreasing from 6.0 to 0.5, indicating good volatile matter release and the increase of non-volatile carbon which are two factors that positively contribute for the formation of a carbon rich and highly porous adsorbent. Concerning the ultimate analysis, the results point out that PS has lower carbon content than BS which also results in lower carbon content materials after the pyrolysis. Both PS and BS have a very significant content in oxygen (in the case of PS it corresponds to twice the carbon percentage). The oxygen percentage suffers a constant decrease with increasing pyrolysis temperature; this fact is much more obvious for the BS pyrolysis remaining only 2% of oxygen for BS800-10 which is in good agreement with a more accentuated release of volatiles with the loss of oxygen functional groups. Moreover, a reduction in the hydrogen content is also observed which is consistent with the increase of the aromaticity of the produced adsorbent. The total organic carbon analyses (Table S1, SI) are in accordance with the ultimate analysis: BS has a higher content of carbon than PS. These analyses allowed to conclude that BS content in inorganic carbon is negligible while in the case of PS it accounts for fifty percent of the total carbon. The PS percentage in inorganic carbon is significantly reduced after pyrolysis resulting in PS related adsorbents with approximately 2–6% of inorganic carbon. The total organic carbon content of PBFG4 was also determined, showing that the produced pyrolysed sludge has a considerably lower content in carbon than the commercial material. The FTIR spectra of PS, BS and pyrolysed materials are presented in Fig. S2 of SI and the spectrum of PBFG4 is shown in Fig. S3 of SI. The spectrum of PS presents several typical bands of cellulose (Méndez et al., 2009). The broad band between 3200 and 3600 cm1 can be attributed to OAH stretching from hydroxyl groups of cellulose and water; bands at 1110 and 1160 cm1 correspond to the cellulosic ethers (CAOAC bonds) and the small band at 1055 cm1 results from CAOH stretch of primary alcohols and carbohydrates (Coates, 2000). The narrow band at 1315 cm1 can be attributed to alcohols OAH stretch (Méndez et al., 2009) or to aromatic ring stretches. Around 1630 cm1 it is possible to identify the characteristic deformation band of water OH groups; however, this band might also be due to carbonyl groups of hemicelluloses. No other characteristic peaks of hemicelluloses or lignin are evident indicating that, if present in PS, these natural polymers occur in relatively small amounts. Finally, the broad band at 1415 cm1 combined with a very sharp band at 870 cm1 is characteristic of carbonates (Coates, 2000). Despite approximately 50% of the PS carbon content being inorganic carbon (Table S1 of SI), which is in agreement with the presence of carbonates, these two peaks remain present in the FTIR spectrum of PS after pyrolysis at 800 °C, temperature at which carbonates should have already

339

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344 Table 1 Chemical characterization of PS, BS and pyrolysed sludge.a PS Proximate analysis (wt%) Moisture content 1.57 Ash 55.31 Volatile matter 36.09 (VM) Fixed carbon (FC) 8.60 VM/FC 4.2 Ultimate analysis (wt%) C 14.83 H 1.26 N 0.40 S 0.29 O 27.91 HHV (kcal kg1) 594 LHV (kcal kg1) 532

PS315150

PS60010

PS80010

PS800150

3.73 45.77 35.72

2.30 47.06 35.39

3.15 50.50 31.40

3.16 61.25 20.77

18.51 1.9

17.55 2.0

18.10 1.7

30.84 2.14 0.43 0.50 20.32 – –

30.69 0.96 0.32 0.56 20.41 – –

28.81 0.47 0.33 0.60 19.29 – –

BS

BS315-150 0.18 mm

BS600-10 0.5 mm

BS800-10 0.18 mm

BS800-10 0.5 mm

16.83 24.82 64.43

4.14 32.26 47.83

4.95 45.63 27.19

5.11 56.58 20.42

5.56 50.18 16.89

17.98 1.2

10.75 6.0

19.91 2.4

27.18 1.0

23.00 0.9

32.93 0.5

27.05 0.82 0.33 0.82 9.73 – –

39.07 4.68 6.00 2.10 23.33 3921 3692

43.62 3.73 6.30 1.93 12.16 – –

39.18 1.63 4.38 2.19 6.99 – –

35.35 0.72 2.47 2.87 2.01 – –

42.07 0.70 3.10 2.27 1.68 – –

a Proximate analysis, ultimate analysis, higher heating value (HHV) and lower heating value (LHV) are presented on a dry basis (with the exception of the moisture content). Fixed carbon (proximate analysis) and oxygen (ultimate analysis) were calculated by difference.

suffered degradation. For this reason, these peaks might also be attributed to the functional groups of the most thermo-resistant fraction of PS organic matter. In the FTIR spectra of the pyrolysed PS (PS315-150, PS600-10, PS800-10 and PS800-150), the typical bands of cellulose gradually disappear with the increase of pyrolysis temperature. In fact, PS800-10 and PS800-150 only have bands at 870 and 1415 cm1 showing that, at this temperature, there was a considerable degradation of the main functional groups of PS. Concerning the BS spectrum, it is possible to observe several bands which were identified in PS as cellulose characteristic, such as the band at 1110 cm1 and the broad band between approximately 3100 and 3600 cm1. Nevertheless, some of the bands found in PS are absent and, simultaneously, the presence of absorption at 1015, 1220 and 1530 cm1 could be indicative of aromatic groups. However, the clear identification of the provenience of these absorption bands is difficult: for instance, 1530 cm1 can be a typical absorption band of lignin or could be attributed to amide groups of cellular proteins from the microorganisms present in the biological sludge reactor. Absorption at 1220 cm1 could also be tentatively attributed to phenol CAO stretch or to aryl-O stretch in aromatic ethers (Coates, 2000). Similarly to PS, the two peaks at 870 and 1415 cm1 are clearly visible, though, in this case, the relative intensity of those peaks toward the organic functional groups is much lower (contrarily to PS). The disappearance of the functional groups with the pyrolysis is also observed and, in BS80010, there are no perceptible bands. 13 C CPMAS solid state NMR spectra of PS, BS and the pyrolysed adsorbents are presented in Fig. S4(a) and (b) of SI. The PS spectrum strongly resembles the published spectrum of cellulose, leading to the same conclusions as those taken from FTIR (McBrierty and Packer, 2006). The peak at 65 ppm is due to cellulose carbon C6, the doublet at 72 and 75 ppm corresponds to carbons C2, C3 and C5, signals at 89 and 83 ppm are due to carbon C4 and finally the signal at 105 ppm can be assigned to C1. Moreover, the spectrum also indicates the presence of little or no lignin due to the absence of a peak at 56 ppm attributed to lignin methoxyl groups and the absence of signal in the aromatic region. Also, the inexistence of a shoulder in the peak at 105 ppm (hemicellulose carbons of xylose, mannose and arabinose) and the absence of signals at 22 and 174 ppm (methyl and carbonyl carbons of acetate groups) could be an indication that PS does not have hemicelluloses (Jackson and Line, 1997). After pyrolysis, the most clear difference is the emergence of a wide band in the aromatic region (between 110 and 160 ppm) and the disappearance of the cellulose typical peaks (first, partially at 315 °C and then completely at 600 and

800 °C), showing that the pyrolysis resulted, as expected, in the increase of the aromaticity of the structure along with major functional group losses. Similarly to FTIR, 13C NMR analysis also evidenced clear differences between PS and BS. The BS spectrum does not resemble that of pure cellulose, despite having some of the typical peaks (such as 72 and 105 ppm). The appearance of several peaks in the high shielded region (20, 24, 30, 33 and 40 ppm) is also observable and can be derived from alkyl carbons which resulted from cellulose degradation by microorganisms during the biological treatment. There are a number of peaks that can be indicative of the presence of lignin (such as 56, 105, 130 and 137 ppm) (Beyer et al., 1997) and a dominant peak centered at 175 ppm which points out to the presence of carboxyl and/or carbonyl groups. The occurrence of lignin typical peaks in BS and not in PS could be explained by the resistance of lignin to biodegradation (increasing its concentration during the biological treatment) and by the easiness of biodegradability of cellulose by microorganisms. Similarly to PS, BS peaks gradually disappear with pyrolysis increasing temperatures with the simultaneous emergence of a wide band in the aromatic region that fully dominates the spectra of BS600-10 and BS800-10. The 1H MAS solid state NMR spectra of the raw sludges and produced adsorbents are shown in Fig. S4(c) and (d) of SI. Due to very high dipolar coupling between 1H spins in the solid-state, the typical resolution of 1H solid-state spectra is drastically diminished when compared to those obtained in liquids, mainly because of peak broadening. Nevertheless, some important results can be withdrawn from these spectra profiles. The results are consistent with the conclusions taken from the 13C CPMAS NMR spectra. PS presents two broad peaks centered at 1.2 and 4.4 ppm which might correspond to alkyl protons and ether/alcohol protons of the cellulose, respectively. With the increase of pyrolysis temperature (PS315-150 and PS600-10) these bands are replaced for broader bands shifted toward the aromatic region of the spectra. Finally, PS800-10 presents a narrower band than the previous adsorbents centered at 0.6 ppm. Taking into account that FTIR and 13C CPMAS NMR point out to the formation of an aromatic material with little or no aliphatic groups, the occurrence of this peak might result from highly shielded protons due to very intense aromatic electronic currents in a highly aromatic structure. The same conclusions can be applied to BS and corresponding adsorbents. The main difference relies on BS800 spectrum that presents an even broader band, including shielded and non-shielded regions, probably caused by a considerable dispersion of the protons chemical environments in the produced material.

340

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

The results of the physical characterization of the pyrolysed sludges and PBFG4 are detailed in Table 2. Despite knowing, from the ultimate analysis, that the pyrolysis of BS resulted in the largest release of volatiles, physical characterization indicates that PS related adsorbents developed a more porous surface. The apparent density of PS adsorbents is around 0.5 g cm3, similarly to PBFG4, while BS adsorbents have a very high apparent density, suggesting low pore formation. Concerning the specific surface area (SBET), there is a steep increase of this parameter for PS adsorbents with the increase of pyrolysis temperature, reaching 209.12 m2 g1 for PS800-150 (which is a very satisfactory value for a pyrolysed non-activated carbon), while BS adsorbents’ surface area did not evolved significantly. The total pore and micropore volumes are also much higher for PS pyrolysed sludges as well as the average micropore width which is 1.02 nm for PS800-150 (very close to PBFG4, 1.30 nm). The same conclusions can be taken from the average pore diameter: it reduces considerably with the increase of pyrolysis temperature of PS, reaching 1.26 nm for PS800-150, compatible with a microporous structure and again very close to that of the commercial carbon. On the other hand, BS sludges did not significantly developed micropores: their characterization is consistent with a mesoporous material. Overall, the physical characterization of the produced adsorbents indicates that the pyrolysis of PS results in the most interesting characteristics: low apparent density, high surface area and higher total pore and micropore volumes. These results also show that the residence time of 150 min (PS800-150) in comparison with 10 min (PS80010) resulted in an adsorbent that should have the most promising adsorption capacity. The morphological characterization of the PS, BS and adsorbents produced at 800 °C (chosen due to their higher surface areas and thus expected to have the most satisfactory adsorption capacities) was evaluated by SEM (PS and BS micrographs are shown in Fig. S5 of SI). All the samples were visualized at 500, 3000, 10,000 and 30,000 to allow a direct comparison of different detail levels between them. PS micrographs show long and homogeneous cellulose fibers with a relatively smooth surface. After pyrolysis at 800 °C for 10 and 150 min (PS800-10 and PS800-150) it is clearly observed a high fragmentation of those fibers and an enormous increase in the roughness of the surface. When comparing the micrographs taken at 10,000 and 30,000, the increase on surface area is evidently shown; this fact is particularly clear for PS800150 where the existence of several layers of adsorbent and pores can be easily observed. These results fully support the conclusions taken from the remaining characterization techniques, showing a very satisfactory increase in the surface area and development of a porous adsorbent. The micrographs of BS and BS800-10 evidence that, despite the formation of a roughest surface after pyrolysis, BS are in the form of unevenly organized large aggregates (see the scale at 500 micrograph in comparison with PS) which should be the main reason for the low surface areas obtained in the physical characterization.

3.3. Adsorption kinetics The amount of citalopram adsorbed onto the paper mill sludge based materials and onto PBFG4 (qt, mg g1) is represented versus agitation time (t) in Fig. 1 together with the fittings to the considered kinetic models (Eqs. (1) and (2)). The parameters obtained from the fittings of experimental results are summarized in Table 3. In general, the pseudo-second order model is the most adequate to fit the experimental data with correlation coefficients ranging from 0.9559 to 0.9993. The equilibrium is more quickly attained for the adsorbents produced at the highest temperatures and longest residence times. This effect is clearly observed for PS adsorbents with equilibrium times varying from 120, 60, 30 to 10 min (approximately) for PS315-150, PS600-10, PS800-10 and PS800-150, respectively. The fastest kinetics was obtained for BS800-10 with the lowest granulometry (<0.18 mm), which achieved the equilibrium almost instantaneously (<5 min). On the other hand it is to highlight that PBFG4 displayed the slowest kinetics, attaining the adsorption equilibrium after 120 min. 3.4. Adsorption isotherms The adsorption isotherms were represented as the amount of citalopram adsorbed onto the produced adsorbents and PBFG4 (qe, mg g1), at equilibrium, versus the amount of citalopram that remained in solution (Ce, mg L1). The experimental data were fitted to three equilibrium adsorption models: Freundlich, Langmuir and Langmuir–Freundlich (see Eqs. 3–5). The combined Langmuir– Freundlich model (data not shown) revealed not to be a good option to estimate the adsorption capacity: the standard deviation associated to the resulting parameters was, in general, higher than the value itself which implies the lack of physical significance of the determined values. The Freundlich and Langmuir fittings are summarized in Table 3 and illustrated in Fig. 2. Overall, the Freundlich model has a better correlation with the experimental data (with r2 ranging from 0.9357 to 0.9967) than the Langmuir model. This is explained by the absence of a defined plateau in the graphical representation of quantity of citalopram adsorbed versus quantity in solution and might mean that the adsorbent has a heterogeneous surface without a fixed or limited available adsorption sites. The shape of the adsorption isotherms of citalopram onto the produced adsorbents revealed that the process is favorable (with N > 1, from the Freundlich model (see Eq. (3))), which means that the adsorbents are efficient not only removing high but also low concentrations of the contaminant. Taking into account that pharmaceutically active contaminants usually occur in surface and waste waters at very low concentrations, this characteristic is a major advantage of the produced adsorbent. A direct comparison between PS and BS adsorbents shows that the adsorption capacity of BS (with the highest KF value of 0.59 ± 0.01 mg g1 (mg L1)N obtained for BS800-10 with the lowest granulometry, 0.18 mm) is considerably lower than PS (with the highest KF value of

Table 2 Physical characterization of the pyrolysed sludge and commercial activated carbon: surface area (SBET), total pore volume (Vp), total micropore volume (W0), average micropore width (L) and average pore diameter (D). Sample

Apparent density/(g cm3)

SBET/(m2 g1)

Vp/(cm3 g1)

W0/(cm3 g1)

L/nm

D/nm

PBFG4 PS315-150 PS600-10 PS800-10 PS800-150

0.55 0.50 0.50 0.51 0.52

848.22 3.43 94.39 120.86 209.12

0.36 0.02 0.06 0.08 0.13

0.295 0.0014 0.039 0.047 0.078

1.30 – – 0.88 1.02

0.84 9.82 1.37 1.41 1.26

BS315-150 0.18 mm BS600-10 0.5 mm BS800-10 0.5 mm BS800-10 0.18 mm

0.83 1.34 0.99 1.06

0.77 1.01 10.82 2.36

0.0043 0.0053 0.020 0.018

0.0003 0.0004 0.0014 0.0010

– – – –

11.15 10.54 3.69 15.46

341

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

Fig. 1. Kinetic study of the adsorption of citalopram onto pyrolysed paper mill sludge and commercial activated carbon (PBFG4). The results were fitted to pseudo-first (dashed lines) and pseudo-second (solid lines) order kinetic models. Each point (±standard deviation) is the average of three replicates. Note that x and y-axis scales are not the same in all graphs to allow a better visualization of the results.

Table 3 Fitting parameters of pseudo-first and pseudo-second order kinetic models and of Freundlich and Langmuir equilibrium models to the experimental data.

Kinetic models Pseudo 1st k1 (min1) qe (mg g1) order r2 n* Pseudo 2nd k2 (mg g1 min) order qe (mg g1) r2 n* Equilibrium adsorption models Freunlich KF (mg g1 (mg L1)N) N r2 n* Langmuir qm (mg g1) KL (L mg1) r2 n* *

PS315-150

PS600-10

PS800-10

PS800-150 BS315-150 0.18 mm

BS600-10 0.5 mm

BS800-10 0.5 mm

BS800-10 0.18 mm

PBFG4

0.08 ± 0.01 5.1 ± 0.2 0.9694 7 0.018 ± 0.002 5.69 ± 0.09 0.997 7

0.22 ± 0.03 3.9 ± 0.1 0.9815 7 0.10 ± 0.03 4.1 ± 0.1 0.9801 7

0.32 ± 0.04 7.2 ± 0.1 0.9925 7 0.105 ± 0.008 7.42 ± 0.04 0.9993 7

0.46 ± 0.07 16.6 ± 0.2 0.9940 7 0.09 ± 0.02 16.9 ± 0.2 0.9968 7

0.010 ± 0.002 0.57 ± 0.04 0.9535 9 0.020 ± 0.006 0.65 ± 0.04 0.9662 9

0.22 ± 0.02 0.362 ± 0.006 0.9938 7 1.2 ± 0.4 0.37 ± 0.01 0.9738 7

0.17 ± 0.05 0.38 ± 0.02 0.9049 7 0.6 ± 0.2 0.41 ± 0.02 0.9559 7

56 0.60 ± 0.01 0.9883 7 5  1015 0.60 ± 0.01 0.9883 7

0.05 ± 0.01 68 ± 4 0.9412 7 0.0009 ± 0.0002 76 ± 3 0.9825 7

2.81 ± 0.07

3.4 ± 0.1

6.7 ± 0.2

14.3 ± 0.2

–

–

0.165 ± 0.008 0.59 ± 0.01 75 ± 1

3.4 ± 0.3 0.9789 11 4.4 ± 0.3 2.1 ± 0.7 0.9088 11

6±1 0.9357 9 3.8 ± 0.3 19 ± 12 0.8523 9

4.9 ± 0.6 0.9793 9 8.5 ± 0.5 6±2 0.9358 9

4.6 ± 0.5 0.9967 6 19.6 ± 0.5 2.9 ± 0.4 0.9976 6

– – – – – – –

– – – – – – –

1.10 ± 0.05 0.9851 14 4±3 0.04 ± 0.02 0.9838 14

1.58 ± 0.06 0.9909 12 2.2 ± 0.2 0.37 ± 0.06 0.9859 12

4.8 ± 0.4 0.9940 9 103 ± 3 2.9 ± 0.4 0.9899 9

n is the number of data points used in the fitting.

14.3 ± 0.2 mg g1 (mg L1)N obtained for PS800-150). Apart from having a 25 times lower adsorption capacity, the BS adsorbents presented a series of other disadvantages. Firstly, it was observed

that BS adsorbents color the citalopram solution after very short contact times, especially for BS315-150 and BS600-10, which should be a consequence of the extraction of the remaining organic

342

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

Fig. 2. Equilibrium data of the adsorption of citalopram onto pyrolysed paper mill sludge and commercial activated carbon (PBFG4). The results were fitted to Langmuir (dashed lines) and Freundlich (solid lines) equilibrium models. Each point (±standard deviation) is the average of three replicates. Note that x and y-axis scales are not the same in all graphs to allow a better visualization of the results.

matter into the solution. In addition, increasing quantities of BS315-150 and BS600-10 did not resulted in higher removal percentages of citalopram despite the high concentrations of the pharmaceutical that still remain in the aqueous phase. This is why it was impossible to present isotherms for these adsorbents and is tentatively explained by the aggregation of BS315-150 and BS600-10 particles in solution when their concentration is increased. Taking into consideration all the described results, the BS adsorbents did not present advantageous characteristics for the removal of this pharmaceutical from contaminated waters. On the other hand, PS adsorbents had shown, as referred before, a reasonable adsorption capacity with KF between 2.81 ± 0.07 mg g1 (mg L1)N (for PS315-150) and 14.3 ± 0.2 mg g1 (mg L1)N (for PS800-150). The increase of the pyrolysis temperature resulted in adsorbents with better adsorption capacity (5 times more adsorption for the highest temperature). The effect of the residence time at 800 °C (PS800-10 and PS800-150) was also

tested and the highest residence time doubled the adsorption capacity. In general, the best results were obtained for the adsorbent with highest surface area and microporous volume (PS800-150) and not for the adsorbents with the highest carbon content and the most efficient volatiles release (as it is the case of the BS related adsorbents). Also, the oxygen content is a major difference between PS and BS adsorbents (under comparable conditions, PS pyrolysis resulted in materials which have approximately 2–5 times higher oxygen percentage) which might indicate that, in this case, the presence of remaining oxygen functional groups on the char surface might have a relevant role on the adsorption of citalopram. When comparing with the commercial activated carbon, although both PS800-150 and PBFG4 displayed favorable adsorption isotherms (with N between 4.6 and 4.8), the adsorption capacity of PBFG4 was five times higher than that of PS800-150. As it has been remarked above, to the best of our knowledge, this is the first study

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

describing the adsorptive removal of citalopram from water; consequently it has not been possible to compare the results here reported using alternative and commercial adsorbents with any results obtained with other adsorbents. 3.5. Potential applications of the produced adsorbents Amongst the produced pyrolysed materials, the best performance for the removal of citalopram was obtained for PS800150, which has several interesting properties for the application on the remediation of contaminated waters. Despite not having, for the tested compound, an adsorption capacity as high as the commercial activated carbon, the obtained result is very promising taking into account that it is a non-activated carbon produced without applying expensive and environmentally aggressive methodologies (such as chemical or physical activation or washing steps with strong acids or bases). Moreover, the adsorption kinetics is very fast which constitutes a relevant advantage of the produced material. On the other hand, considering the nature of the precursor used for the production of the adsorbents, this application also constitutes a novel alternative for the valorization of paper mill sludge, produced at very large scale all over the world. Naturally, further research is needed in order to improve the performance of the adsorbent, evaluate its adsorption capacity for other environmentally relevant compounds and to understand its behavior under realistic conditions (possible effects of pH, temperature and presence of dissolved organic matter). 4. Conclusions Physical and chemical characterization of the adsorbents revealed that pyrolysis effectively resulted in the disappearance of PS and BS main functional groups, generating highly aromatic structures. However, contrarily to BS, only PS resulted in the production of a microporous char with a larger surface area. The application of the BS and PS pyrolysed materials for the removal of citalopram showed that PS derived adsorbents are the most effective, which are in line with the conclusions taken from the adsorbent’s characterization. The best results were obtained by pyrolysis of PS at 800 °C for 150 min (KF = 14.3 ± 0.2 mg g1 (mg L1)N); also, the adsorption equilibrium is very quickly attained (approximately 10 min). Acknowledgements This work was supported by European Funds through COMPETE and by National Funds through the Portuguese Science Foundation (FCT) within project PEst-C/MAR/LA0017/2013. Vânia Calisto and Catarina I.A. Ferreira thank (FCT) for their postdoctoral (SFRH/ BPD/78645/2011) and PhD grants (SFRH/BD/88965/2012), respectively. Also, Marta Otero acknowledges financial support from the Spanish Ministry of Science and Innovation (RYC-201005634). The authors also thank the kind collaboration of Eng° Luís Machado and Engª Ana Reis from RAIZ – Instituto de Investigação da Floresta e do Papel. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.05. 047. References AENOR, 1984. Solid mineral fuels. Determination of ash. Asociación Española de Normalización y Certificación.

343

AENOR, 1985. Hard coal and coke. Determination of volatile matter content. Asociación Española de Normalización y Certificación. AENOR, 1995. Solid mineral fuels. Determination of moisture in the analysis sample. Asociación Española de Normalización y Certificación. Aksu, Z., Tunç, Ö., 2005. Application of biosorption for penicillin G removal: comparison with activated carbon. Process Biochem. 40, 831–847. Antunes, M., Esteves, V.I., Guégan, R., Crespo, J.S., Fernandes, A.N., Giovanela, M., 2012. Removal of diclofenac sodium from aqueous solution by Isabel grape bagasse. Chem. Eng. J. 192, 114–121. Beyer, L., Fründ, R., Mueller, K., 1997. Short-term effects of a secondary paper mill sludge application on soil properties in a Psammentic Haplumbrept under cultivation. Sci. Total Environ. 197, 127–137. Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 60, 309–319. Calisto, V., Domingues, M.R.M., Erny, G.L., Esteves, V.I., 2011. Direct photodegradation of carbamazepine followed by micellar electrokinetic chromatography and mass spectrometry. Water Res. 45, 1095–1104. Calisto, V., Esteves, V.I., 2009. Psychiatric pharmaceuticals in the environment. Chemosphere 77, 1257–1274. CEPI, 2004. Discovering the high potential of Pulp and Paper Production Residues. Confederation of European Paper Industries. Chong, M.N., Jin, B., Chow, C.W.K., Saint, C., 2010. Recent developments in photocatalytic water treatment technology: a review. Water Res. 44, 2997– 3027. Coates, J., 2000. Interpretation of infrared spectra – a practical approach. In: Meyers, R.A. (Ed.), Encyclopedia of Analytical Chemistry. John Wiley & Sons Ltd., Chichester, pp. 10815–10837. Devi, P., Saroha, A.K., 2014. Risk analysis of pyrolyzed biochar made from paper mill effluent treatment plant sludge for bioavailability and eco-toxicity of heavy metals. Bioresour. Technol. 162, 308–315. Dubinin, M.M., 1966. Properties of active carbons. In: Chemistry and Physics of Carbon. Marcel Dekker Inc., New York, pp. 51–120. Ioannidou, O., Zabaniotou, A., 2007. Agricultural residues as precursors for activated carbon production – a review. Renewable Sustainable Energy Rev. 11, 1966–2005. ISO, 2009. Determination of Gross Calorific Value by the Bomb Calorimetric Method and Calculation of Net Calorific Value. International Organization for Standardization. Jackson, M.J., Line, M.A., 1997. Organic composition of a pulp and paper mill sludge determined by FTIR, 13C CP MAS NMR, and chemical extraction techniques. J. Agric. Food Chem. 45, 2354–2358. Jelic, A., Gros, M., Ginebreda, A., Cespedes-Sánchez, R., Ventura, F., Petrovic, M., Barcelo, D., 2011. Occurrence, partition and removal of pharmaceuticals in sewage water and sludge during wastewater treatment. Water Res. 45, 1165– 1176. Khalili, N.R., Vyas, J.D., Weangkaew, W., Westfall, S.J., Parulekar, S.J., Sherwood, R., 2002. Synthesis and characterization of activated carbon and bioactive adsorbent produced from paper mill sludge. Sep. Purif. Technol. 26, 295–304. Kinney, C.A., Furlong, E.T., Werner, S.L., Cahill, J.D., 2006. Presence and distribution of wastewater-derived pharmaceuticals in soil irrigated with reclaimed water. Environ. Toxicol. Chem. 25, 317–326. Li, W.-H., Yue, Q.-Y., Gao, B.-Y., Wang, X.-J., Qi, Y.-F., Zhao, Y.-Q., Li, Y.-J., 2011. Preparation of sludge-based activated carbon made from paper mill sewage sludge by steam activation for dye wastewater treatment. Desalination 278, 179–185. Margot, J., Kienle, C., Magnet, A., Weil, M., Rossi, L., de Alencastro, L.F., Abegglen, C., Thonney, D., Chèvre, N., Schärer, M., Barry, D.A., 2013. Treatment of micropollutants in municipal wastewater: ozone or powdered activated carbon? Sci. Total Environ. 461–462, 480–498. McBrierty, V.J., Packer, K.J., 2006. Nuclear Magnetic Resonance in Solid Polymers. Cambridge University Press, New York. Méndez, A., Fidalgo, J.M., Guerrero, F., Gascó, G., 2009. Characterization and pyrolysis behaviour of different paper mill waste materials. J. Anal. Appl. Pyrolysis 86, 66–73. Mohan, D., Sarswat, A., Ok, Y.S., Pittman Jr., C.U., 2014. Organic and inorganic contaminants removal from water with biochar, a renewable, low cost and sustainable adsorbent – a critical review. Bioresour. Technol. 160, 191–202. Monte, M.C., Fuente, E., Blanco, A., Negro, C., 2009. Waste management from pulp and paper production in the European Union. Waste Manage. 29, 293–308. OECD, 2011. Health at a Glance 2011. OECD Publishing, http://dx.doi.org/10.1787/ health_glance-2011-en. Rivera-Utrilla, J., Sánchez-Polo, M., Ferro-García, M.Á., Prados-Joya, G., OcampoPérez, R., 2013. Pharmaceuticals as emerging contaminants and their removal from water. A review. Chemosphere 93, 1268–1287. Silva, C.P., Otero, M., Esteves, V., 2012. Processes for the elimination of estrogenic steroid hormones from water: a review. Environ. Pollut. 165, 38–58. Smith, K.M., Fowler, G.D., Pullket, S., Graham, N.J.D., 2009. Sewage sludge-based adsorbents: a review of their production, properties and use in water treatment applications. Water Res. 43, 2569–2594. Stoeckli, F., Ballerini, L., 1991. Evolution of microporosity during activation of carbon. Fuel 70, 557–559. Stoica, A., Sandberg, M., Holby, O., 2009. Energy use and recovery strategies within wastewater treatment and sludge handling at pulp and paper mills. Bioresour. Technol. 100, 3497–3505. Yao, Y., Gao, B., Chen, H., Jiang, L., Inyang, M., Zimmerman, A.R., Cao, X., Yang, L., Xue, Y., Li, H., 2012. Adsorption of sulfamethoxazole on biochar and its impact on reclaimed water irrigation. J. Hazard. Mater. 209–210, 408–413.

344

V. Calisto et al. / Bioresource Technology 166 (2014) 335–344

Yu, Z., Peldszus, S., Huck, P.M., 2009. Adsorption of selected pharmaceuticals and an endocrine disrupting compound by granular activated carbon. 1. Adsorption capacity and kinetics. Environ. Sci. Technol. 43, 1467–1473.

Yu, Z.R., Peldszus, S., Huck, P.M., 2008. Adsorption characteristics of selected pharmaceuticals and an endocrine disrupting compound – naproxen, carbamazepine and nonylphenol – on activated carbon. Water Res. 42, 2873–2882.