Using carbonized low-cost materials for removal of chemicals of environmental concern from water

Using carbonized low-cost materials for removal of chemicals of environmental concern from water Adsorption on low-cost biochars would increase the affordability and availability of water treatment in, for example, developing countries. The aim of this study was to identify the precursor materials and hydrochar surface properties that yield efficient removal of compounds of environmental concern (CEC). We determined the adsorption kinetics of a mixture containing ten CECs (octhilinone, triclosan, trimethoprim, sulfamethoxasole, ciprofloxacin, diclofenac, paracetamol, diphenhydramine, flucon- azole, and bisphenol A) to hydrochars prepared from agricultural waste (including tomato- and olive-press wastes, rice husks, and horse manure). The surface characteristics of the hydrochars were evaluated via diffuse reflectance infrared spectroscopy (DRIFTS), X-ray photoelectron spectroscopy (XPS), and N -adsorption. Kinetic adsorption tests revealed that removal efficiencies varied substantially among different materials. Similarly, surface analysis revealed differences among the studied hydrochars and the degree of changes that the materials undergo during carbonization. According to the DRIFTS data, compared with the least efficient adsorbent materials, the most efficient hydrochars underwent more substantial changes during carbonization. . . . . . Keywords Hydrochar Adsorption Hydrothermal carbonization Agro-industrial residues Organic chemicals Low-cost adsorbents Introduction and adsorption on activated carbon (Gupta and Suhas 2009; Nyenje et al. 2010). Additionally, even if current sewage treat- Water treatment poses a serious global challenge, in terms of ment plants reduce the spread of bacteria and nutrients, some providing safe sanitation and using limited water resources in chemicals pass though the treatment plant unreduced (Snyder an economical and efficient manner that ensures the availabil- et al. 2003;Lindberget al. 2014; Melvin and Leusch 2016). ity of clean water for everyone. However, high costs and tech- This is especially crucial for chemicals of environmental con- nical complexity limit the use of conventional water-treatment cern (CEC), e.g., antibiotics, pharmaceuticals, and biocides, techniques, including sewage treatment, oxidation, filtration, that are designed to have biological effects. Studies have shown that some pharmaceuticals can be absorbed by crops, and subsequently enter both humans and livestock (Franklin Responsible editor: Philippe Garrigues et al. 2016). The availability of water treatment can be in- Electronic supplementary material The online version of this article creased via simple, inexpensive techniques. For example, ad- (https://doi.org/10.1007/s11356-018-1781-0) contains supplementary sorption, which removes various contaminants even at low material, which is available to authorized users. concentrations, is easily realized, but is costly due to the lim- ited lifetime of adsorbents prepared from non-local or fossil * Stina Jansson feedstock (Gupta and Suhas 2009). By valorizing local low- stina.jansson@umu.se cost feedstocks, such as agricultural or food industry residues, Department of Chemistry, Umeå University, SE-901 adsorption becomes affordable also in countries and regions 87 Umeå, Sweden with high water stress and limited economical resources Umeå Energi AB, SE-901 05 Umeå, Sweden (Mohan et al. 2014). Biochars are carbon-rich porous materials obtained via the Industrial Doctoral School, Umeå University, SE-901 carbonization of feedstock materials, and they may be used as 87 Umeå, Sweden 15794 Environ Sci Pollut Res (2018) 25:15793–15801 adsorbent materials due to their similarity to commercial acti- subsequently carbonized at 220 °C for 2 h. Portions vated carbons. Biochars generated from wet feedstocks are (~ 600 mL) of the wet sample were carbonized during each produced preferably through hydrothermal carbonization experiment and yields of 57–66% (66, 62, 57, and 59% for (HTC), an energy-efficient wet technique performed without rice husks, manure, tomato waste, and olive waste, respective- drying of the raw materials before carbonization. The resulting ly) were realized. After carbonization, the chars were retrieved biochar, or more specifically the hydrochar, is more hydropho- via filtration and dried overnight at 105 °C. bic than the raw material and is therefore easily dried (Escala Prior to use, the chars were demineralized, i.e., dissolvable et al. 2013; Vom Eyser et al. 2015). HTC was first described in material was removed, and the surface functionalities were 1913 (Bergius 1913) and nowadays is considered one of the protonated by washing the chars in 0.1 M HCl until a clear most sustainable approaches for obtaining functional carbon- water phase was obtained (Meng et al. 2013), rinsing with based materials (Hu et al. 2010; Titirici and Antonietti 2010). ultrapure water and filtering (Munktell, Qualitative filter paper This process involves the use of organic materials (such as grade 3/< 10 μm), and drying at 105 °C for 12 h. All chars carbohydrates, manure, sludge, wood, food, and other agricul- were homogenized by grinding in a mortar, except for the tural wastes) as precursors (Hu et al. 2010; Berge et al. 2011; tomato waste chars, which were blitzed for 4 × 2-s pulses at Oliveira et al. 2013; Falco et al. 2013). Hydrochars often retain 10000 rpm (Grindomix GM 200, Retch). The dry chars were much of the surface functionalities from the raw materials then stored in a desiccator. (mostly oxygen and hydrogen-containing acidic groups, e.g., phenolic, lactonic, carboxyl, and carbonyl groups (Wiedner Preparation of model water et al. 2013)) which increase interactions with ionic or polar compounds (Liu et al. 2010). Although different types of bio- Our model water contained ten different CEC substances: two chars have gained attention recently and many potential feed- biocides: octhilinone (2-octyl-4-isothiazolin-3-one) and triclo- stocks have been identified (Mohan et al. 2014), the data on san; three antibiotics: trimethoprim, sulfamethoxasole, and these materials is inconsistent. This inconsistency results from ciprofloxacin; four pharmaceuticals: diclofenac, paracetamol, the different carbonization methods and parameters as well as diphenhydramine, and fluconazole; and the plastic additive adsorbates and adsorption test parameters employed. bisphenol A in concentrations of 10 μg/L in ultrapure water. Additionally, complex chemical mixtures and simulating real The corresponding trade names, structures, CAS (Chemical wastewater, should be investigated instead of focusing on so- Abstracts Service) registry numbers, solubility in water, log lutions containing only one compound. K , and main applications are shown in Fig. S1 OW In this screening study, the CEC-removal efficiencies of (Supplementary data). The tested substances were selected, four hydrochars prepared from low-cost feedstocks were de- owing to their widespread use in developing countries as well termined. The sorption properties were elucidated through as their ability to pass unreduced through conventional sew- surface characterization where the role of surface area and age treatment plants and being emitted into the environment surface functionalities in the removal of contaminants from through effluent water (Snyder et al. 2003; Lindberg et al. water, was evaluated. 2014; Melvin and Leusch 2016). Batch adsorption experiments Materials and methods The experiments were all performed at 20 °C and were con- Preparation of hydrochars ducted as follows: Triplicate adsorption tests were performed using 50 ± 2.5 mg hydrochar + 10 mL model water in 15 mL Four organic residues were used in this study: horse manure, plastic tubes. After weighing the hydrochar, water was added skin and seeds from tomato, olive press residues, and rice to the tube, which was then agitated for 1, 3, 5, 8, 12, 18, or husks. These materials are considered representatives of com- 25 min. Afterwards, ca 5 mL of water/hydrochar was removed mon agricultural residues in countries with high water stress. from the tube using a syringe, passed through a 0.45-μmsy- The hydrochars were prepared in a stirred high-pressure ringe filter (Filtopur, Sarstedt) into a glass vial, and weighed lab-scale reactor with an internal volume of 1 L (Zhengzhou before addition of the internal standard. The internal standard Keda Machinery and Instrument Equipment Co., Ltd., China). consisted of isotopically labeled bisphenol A, triclosan, tri- The walls of the reactor were heated by a detachable resistance methoprim, sulfamethoxasole, ciprofloxacin, diclofenac, heater. After each experiment, the heater was removed and the paracetamol, and fluconazole. Furthermore, the samples were reaction was water-cooled. The four feedstock materials were all either analyzed directly or frozen until analysis, which covered with ultrapure water, resulting in mixtures with dry- occurred within 3 weeks of the experiment. matter content ranging from 11 to 24% (rice husks 11%, ma- For each hydrochar, triplicate blank samples were prepared nure 12%, tomato waste 12%, and olive waste 24%) and via agitation with ultrapure water. Leaching of the analytes by Environ Sci Pollut Res (2018) 25:15793–15801 15795 −1 −1 the tubes was investigated by agitating (for 25 min) triplicate 5000 cm was collected at a spectral resolution of 4 cm . tube blanks of 10 mL ultrapure water. Similarly, adsorption of In addition, pure KBr was used as the background, and the −1 the analytes to the tube walls was investigated through a trip- spectral range 400–3750 cm was used in the subsequent licate tube adsorption test where 10 mL of the model water multivariate analysis. Prior to multivariate analysis, the spec- was agitated (for 25 min) in the tubes. In addition, the capacity tra were baseline-corrected (asymmetric least squares, lamb- of the chars (relative to that of commercially available active da = 14,000,000, p = 0.01), smoothed (Savitzky-Golay filter- carbon) was assessed by agitating triplicate samples of ing, polynomial order = 1, frame = 5), and total-area normal- general-purpose-grade powdered activated carbon (Fisher ized, using the protocol described by Felten et al. 2015. Scientific) with the model water for 25 min. A list of samples, agitation times, and blank samples is provided in Table S1 X-ray photoelectron spectroscopy (XPS) (Supplementary data). The XPS spectra were collected with a Kratos Axis Ultra DLD LC-MS analysis electron spectrometer using a monochromatic AlKα source operated at 120 W. An analyzer pass energy of 160 eV and a The samples were all analyzed using a Thermo TSQ Quantum pass energy of 20 eV were used for acquiring wide spectra and Ultra EMR (Thermo Fisher Scientific, San Jose, CA, USA) individual photoelectron lines, respectively. The surface po- mass spectrometer coupled to a PAL HTC auto sampler (CTC tential was stabilized by the spectrometer charge neutraliza- Analytics AG, Zwingen, Switzerland). Two pumps (Surveyor tion system. The binding energy (BE) scale was referenced to and Accela, Thermo Fisher Scientific, San Jose, CA, USA) the C1s line (set at 285.0 eV) of aliphatic carbon, and the were used, and the separation was achieved with a Thermo spectra were processed with the Kratos software. Powder sam- Hypersil Gold AQ (50 × 2.1 mm, 5 μm) column. Analytes ples for the analysis were gently hand-pressed into a pellet were ionized via heated electrospray (HESI) or atmospheric (directly on a sample holder) using a clean Ni spatula. pressure photoionization (APPI) using a krypton lamp at 10.6 eV, in positive- or negative-ion mode. The settings, in- Principal component analysis (PCA) cluding the HESI/APPI ionization data, polarities, precursor/ product ions, collision energies, tube lens values, quantifica- The differences between the materials were graphically visu- tion and qualification ions, and limits of quantification (LOQ), alized via Principal Component Analysis (PCA) (Jolliffe associated with the analysis are summarized in Table S2 2002). In PCA, the variation in a dataset is determined by (Supplementary data). The analytical system is based on col- extracting orthogonal principal components from a larger umn switching using six- and ten-port valves; the basic setup number of variables. The first principal component (PC1) ac- is described by Khan et al. (2012) (Khan et al. 2012). During counts for the largest variation in the dataset, PC2 accounts for analysis, a spectral resolution corresponding to a full width the second largest, and so on. Score plots were used to display half maximum (FWHM) of 0.7 was used for both distributions of the observations (here, the carbonized and quadrupoles. non-carbonized materials) projected onto a plane. In these distributions, similar samples are closely grouped, whereas Surface-area determination dissimilar samples are separated by large distances. The SIMCA-P software package (version 13.0, Umetrics AB, The specific surface area of the hydrochars was determined Sweden) was used for the PCA modeling and all variables via N adsorption and calculated using the Brunauer– were center-scaled. Furthermore, the number of significant Emmett–Teller (BET) theory. Degassing was performed at components was determined via sevenfold full cross- 120 °C. validation (CV) (Stone 1973), and components with eigen- values lower than two were removed from the model. Diffuse reflectance infrared fourier transform spectroscopy (DRIFTS) Results and discussion Approximately 10 mg of dry sample was manually ground with ca. 390 mg Fourier transform-infrared (FT-IR) Overall adsorption efficiency spectroscopy-grade KBr (Merck, Darmstadt, Germany) in an agate mortar. The corresponding FT-IR spectra were recorded The different hydrochars exhibited differing removal proper- in diffuse-reflectance mode (Bruker IFS 66v/S; Bruker Optik ties, and CEC removal from the water differed between sub- GmbH, Ettlingen, Germany) under vacuum conditions, in ac- stance and char. Furthermore, the substances were all rapidly cordance with the protocol described by Gorzsás and removed, and the amount removed saturated after 1 to 3 min Sundberg 2014. Data in the spectral range of 400– (see Figs. S2-S5 in Supplementary data for further details of 15796 Environ Sci Pollut Res (2018) 25:15793–15801 Fig. 1 Total amount of removed CEC per gram of char. Removal was extremely rapid, thereby preventing the determination of gradients and, hence, the values displayed correspond to the average of all 21 samples. Error bars denote one standard deviation. Maximum amount removable: 2 μg/g hydrochar the substances and chars). This rapid removal enabled the use model water. Horse manure char was the most efficient mate- of a total (overall)-removal efficiency for each substance and rial for removing paracetamol, fluconazole, and sulfamethox- char, thereby improving the statistical accuracy of the removal azole, even though these compounds were generally removed results (Fig. 1). to a low degree. Moreover, the CECs in the 18 blank samples Horse manure and rice husk chars exhibited the highest were all lower than the LOQ (see Table S2, Supplementary overall removal with CEC average removal efficiencies of data), and, after 25 min, the analytes were all removed by 49 and 66%, respectively. These values are higher than the activated carbon. However, some of the analytes (especially respective values (39 and 32%) associated with olive and to- diphenhydramine, octhilinone, and bisphenol A (Fig. 1)) were mato residue chars. Diclofenac and bisphenol A were nearly adsorbed onto the tube walls. completely removed by all four chars. Similarly, rice husk Similar results have been found in previous studies. Jung char exhibited high efficiency (≥ 83%) of ciprofloxacin, di- et al. (2013) reported nearly complete adsorption of bisphenol phenhydramine, octhilinone, and triclosan removal from the A using activated carbons from loblolly pine chips, while Fig. 2 Correlation between the hydrophobicity of the adsorbate and the concentration adsorbed on each hydrochar Environ Sci Pollut Res (2018) 25:15793–15801 15797 with log K was observed (Fig. 2). Compared with their OW lower-log K counterparts, the substances with higher log OW K values were, in general, more efficiently removed by the OW chars. Removal by rice husk char exhibited the highest (R = 0.62, p = 0.0069) positive correlation with log K ,suggest- OW ing that this removal has the best fitting linear correlation with log K values of the CECs. The correlation obtained for OW carbonized olive waste and tomato waste, although lower than that of the carbonized rice husks, was still significant. Manure char exhibited the lowest overall correlation (R =0.31, p = 0.09). When the amount removed was plotted as a function of log K , the steepest slopes, 0.34, 0.34, and 0.30, of the least OW square fit were obtained for the olive waste char, rice husk char, and tomato waste char, respectively. Compared with these plots, the plot corresponding to the horse manure char exhibited a more gradual slope (0.24), suggesting that the amount removed is more independent of log K .Low-log OW K compounds, such as paracetamol and fluconazole, were OW most efficiently removed by horse manure char. This is attrib- uted to the dissimilar surface properties that may promote different removal mechanisms. Hydrophobic molecules, in Fig. 3 Baseline-corrected DRIFTS spectra of the treated materials general, attach to surfaces via hydrophobic interactions, whereas their more hydrophilic counterparts are removed diclofenac was less efficiently removed and sulfamethoxazole through other electrostatic interactions (e.g., hydrogen bond- the least adsorbed, and the adsorption followed the same order ing) (Sun et al. 2012). as in this study. Sulfamethoxazole removal by primary paper mill sludge char exceeded 50% in a study by Calisto and co- Surface characterization workers (2015), and in a study using activated sucrose hydrochar, the paracetamol removal was around 50% or less The surface properties of the carbonized material were inves- (Mestre et al. 2015). The low removal by these different tigated via three surface analysis techniques namely, BET, hydrochars suggests that these types of compounds are not DRIFTS, and XPS. These analyses showed that the surface easily adsorbed from the solution. Nevertheless, comparisons properties of rice husk and manure chars differ significantly should made with caution, because the concentrations in the from those of the tomato- and olive-residue chars. above-mentioned studies were in the mg range. This is many In addition, BET-determined specific surface areas of 16.92 orders of magnitude higher than the concentrations used in and 4.62m /g were obtained for rice husk char and horse ma- this study. Thus, it is likely that the solute concentration will nure char, The surface area of the tomato and olive waste chars affect the removal efficiency. This was indicated by Li et al. 2 2 were below 1 m /g (0.74 and 0.65 m /g respectively), and 2018, where ciprofloxacin removal efficiency was less for the these values were associated with high levels of uncertainty lowest studied concentration compared to higher concentra- due to both the small surface area and the large fraction of tions in the range of 150–500 mg/l. volatiles on the char surfaces. This may have resulted from the Correlations among the amount adsorbed, water solubility relatively low carbonization temperature (220 °C). Moreover, of each substance, and log K were determined. No corre- OW considering that rice husk char had nearly fourfold higher lation was found between the removed amount of CEC and surface area compared to horse manure char (16.92 vs. their solubility; however, in some cases, a positive correlation 4.62 m /g), the overall removal for horse manure char (49%) was noteworthy. This indicates that the surface area consti- Table 1 Elemental composition (atomic-%) of the surface tutes only one factor that drives the adsorption. The DRIFTS spectra of the char materials and the raw Element Rice husks Horse manure Olive waste Tomato waste materials are shown in Fig. 3 and Fig. S6 (Supplementary data), respectively. Broad O-H stretching bands from hydrox- C 54.30 74.52 87.84 83.71 −1 yl groups (Franca et al. 2010) occurred at 3500–3300 cm for O 37.69 23.57 11.51 14.71 all samples, but occurred with higher intensity in the horse N – 1.52 0.64 1.58 manure and rice husk chars. Furthermore, the materials all Si 8.01 0.38 –– consist of aliphatic CH , as indicated by the occurrence of 2 15798 Environ Sci Pollut Res (2018) 25:15793–15801 Fig. 4 Deconvoluted C1s lines of the horse manure char, rice husk char, tomato waste char, and olive waste char −1 C-H stretching vibrations at 2925 and 2853 cm , and small functionalities. Additionally, a large band at 1100– −1 −1 bands at 1450–1430 and 1370 cm (Chen and Chen 2009). 1000 cm arising from the rice husk sample results probably The bands associated with the olive and tomato waste were from Si-O-Si in-plane vibrations, as suggested by XPS analy- more intense than those associated with manure and rise husks sis results that revealed the presence of silicon (Table 1). The before and after carbonization. Moreover, the sharp C = O spectra of the treated materials are less diverse than those of −1 stretching band occurring at 1750–1700 cm results from the untreated materials, i.e., the intensities of C-H and C = O, carboxylic acids, esters, ketones, lactones, or aldehydes and C-O bands vary more within untreated materials com- (Esteves et al. 2013). The band associated with the olive and pared to the treated ones. tomato wastes is sharper than that corresponding to the ma- The elemental composition of the biochar surface is deter- nure and rice husks. Furthermore, the occurrence of bands at mined in further detail via XPS analysis (see Table 1). As the −1 1600–1450 cm is indicative of aromatic ring C = C table shows, the carbon content of the olive and tomato waste stretching in the samples (Chen and Chen 2009;Lietal. chars is higher than that of the rice husk and horse manure −1 2014), and small sharp bands at 900–700 cm may have chars. This result concurs with the DRIFT spectra, which ex- originated from aromatic C-H out-of-plane vibrations (Fang hibits high intensities for aliphatic carbon vibrations. In addi- −1 et al. 2014). Several narrow bands at 1300–1000 cm are tion, silicon (bound as SiO ) occurs at a rather high concen- attributed to O-H bending and C-O stretching in ethers, alco- tration in the rice husk char, but constitutes only trace amounts hols, phenols, lactones or carboxyl acids, and anhydrides of the horse manure char. Small amounts of carbon-bound (Bustin and Guo 1999; Shuttleworth et al. 2015). The en- nitrogen occur in all samples, except for the rice husk chars hancement of these bands after hydrothermal carbonization (Yang and Jiang 2014), and other elements occur at levels of the materials is indicative of changes in the surface lower than the limit of detection (~ 0.1 atomic-%). Environ Sci Pollut Res (2018) 25:15793–15801 15799 Fig. 5 Plot showing the first two principal components of the PCA model of the DRIFTS spectra Figure 4 shows the deconvoluted C1s lines of the the removal of more polar molecules via H-bonding (Sun et al. hydrochars. The peaks occurring at 284.9–285.0 eV corre- 2012), than the other chars considered. spond to aliphatic and/or graphitic carbon, the dominant spe- cies in all the chars. Compared with the olive and tomato waste chars, rice husk and manure chars contain higher frac- Multivariate data modeling tions of carbon species with a single bond to oxygen, i.e., hydroxyls and ethers (at 286.5–286.8 eV), which support the The differences between the materials were elucidated by previously presented DRIFTS data. Carbonyls at 288.1– performing a PCA using the baseline-corrected DRIFTS spec- 288.6 eV and esters or carboxyl acids at 289.2–289.6 eV oc- tra. The resulting model consists of four principal components curred in only small concentrations. Additionally, a minor line that account for 86% of the variability (R2) with 69% predic- at 291.2 eV associated with the horse manure char, results tive ability (Q2; see Fig. 5 for the score plot of the first two from either a carbonate or π-π* shakeup satellite (Yu et al. components). The PCA reveals that the treated materials are 2012). more homogeneous, as evidenced by the close grouping of the Surface analyses revealed substantial differences between char replicates, than the untreated material. The PCA score the studied biochars. Although BET-determined surface areas plot also shows that the first principal component accounts varied between the best- (manure and rice husk chars) and for 57% of the variability (R2) and 42% of the predictive worst-performing (tomato waste and olive waste chars) mate- ability (Q2) (Fig. 5, t[1]). Furthermore, compared with that rials, this factor constituted only one of the several important occurring for materials with poor removal capacity (i.e., toma- features affecting this performance. DRIFTS and XPS re- to and olive waste), the variability in the DRIFTS spectra is vealed differences in the carbon content and O-functionalities, higher for materials with good removal capacity (i.e., rice especially the OH content. Furthermore, the SiO and OH husks and manure). The inter-surface differences between content of the rice husk and horse manure chars is higher than high- and low-adsorption materials are larger than those be- those of the tomato waste and olive waste chars, whereas the tween carbonized and non-carbonized materials (which are aliphatic carbon content is lower. Therefore, the rice husk and distinguished by the second principal component, t[2] with, horse manure chars are more polar, and thereby may promote variability (R2): 17%, predictive ability (Q2): 4%). The PCA 15800 Environ Sci Pollut Res (2018) 25:15793–15801 and waste-based carbons. J Environ Manag 152:83–90. https://doi. model also revealed that the olive and tomato chars are very org/10.1016/j.jenvman.2015.01.019 similar. Chen B, Chen Z (2009) Sorption of naphthalene and 1-naphthol by bio- The overall results of the PCA model of the DRIFTS spec- chars of orange peels with different pyrolytic temperatures. tra may allow screening of untreated materials (based on sur- Chemosphere 76:127–133. https://doi.org/10.1016/j.chemosphere. 2009.02.004 face functionalities) and prediction of removal capacities. Escala M, Zumbühl T, Koller C, Junge R, Krebs R (2013) Hydrothermal However, a large database and multiple runs would be needed carbonization as an energy-efficient alternative to established drying for highly heterogeneous untreated material (for example, ma- technologies for sewage sludge: a feasibility study on a laboratory nure or tomato waste). scale. Energy Fuels 27:454–460. https://doi.org/10.1021/ef3015266 Esteves B, Velez Marques A, Domingos I, Pereira H (2013) Chemical changes of heat treated pine and eucalypt wood monitored by FTIR. Maderas Cienc y Tecnol 15. https://doi.org/10.4067/S0718- 221X2013005000020 Conclusions Falco C, Marco-Lozar JP, Salinas-Torres D, Morallón E, Cazorla-Amorós D, Titirici MM, Lozano-Castelló D (2013) Tailoring the porosity of These findings demonstrate that carbonized low-value mate- chemically activated hydrothermal carbons: influence of the precur- sor and hydrothermal carbonization temperature. Carbon 62:346– rials can remove CEC from water, but the removal efficiency 355. https://doi.org/10.1016/j.carbon.2013.06.017 varies with the feedstock. Furthermore, DRIFTS and XPS Fang Q, Chen B, Lin Y, Guan Y (2014) Aromatic and hydrophobic analyses revealed significant differences in the elemental com- surfaces of wood-derived biochar enhance perchlorate adsorption position and functionalities of the hydrochars. Multivariate via hydrogen bonding to oxygen-containing organic groups. Environ Sci Technol 48:279–288. https://doi.org/10.1021/ analysis based on the DRIFTS data showed that, compared es403711y with those corresponding to the worst-performing hydrochars, Felten J, Hall H, Jaumot J, Tauler R, de Juan A, Gorzsás A (2015) larger differences between the treated and untreated materials Vibrational spectroscopic image analysis of biological material occurred for the most promising hydrochars. using multivariate curve resolution–alternating least squares (MCR-ALS). Nat Protoc 10:217–240. https://doi.org/10.1038/ nprot.2015.008 Acknowledgements The DRIFTS analysis was performed, with the as- Franca AS, Oliveira LS, Nunes AA, Alves CCO (2010) Microwave sistance of András Gorzsás, at the Vibrational Spectroscopy Core Facility assisted thermal treatment of defective coffee beans press cake for at the KBC, Umeå University, and the XPS analysis and interpretation the production of adsorbents. Bioresour Technol 101:1068–1074. were conducted by Andrey Shchukarev at Umeå University. The authors https://doi.org/10.1016/j.biortech.2009.08.102 also gratefully acknowledge Professor Andres Fullana from the Franklin AM, Williams CF, Andrews DM, Woodward EE, Watson JE University of Alicante for use of their HTC facility. Professor Jean- (2016) Uptake of three antibiotics and an antiepileptic drug by wheat François Boily at Umeå University is gratefully acknowledged for FT- crops spray irrigated with wastewater treatment plant effluent. J IR interpretation. Part of the study was financed by grants from Environ Qual 45:546–554. https://doi.org/10.2134/jeq2015.05.0257 Ångpanneföreningen’s Foundation for Research and Development and Gorzsás A, Sundberg B (2014) Chemical fingerprinting of arabidopsis The J. Gust. Richert Memorial Fund. We also thank The Swedish using fourier transform infrared (FT-IR) spectroscopic approaches. Research Council and Bio4Energy, a strategic research environment In: Sanchez-Serrano J, Salinas J (eds) Arabidopsis Protocols. appointed by the Swedish government, for supporting this work. Humana Press, Totowa, pp 317–352 Gupta VK, Suhas (2009) Application of low-cost adsorbents for dye Open Access This article is distributed under the terms of the Creative removal—a review. J Environ Manag 90:2313–2342. https://doi. Commons Attribution 4.0 International License (http:// org/10.1016/j.jenvman.2008.11.017 creativecommons.org/licenses/by/4.0/), which permits unrestricted use, Hu B, Wang K, Wu L, Yu SH, Antonietti M, Titirici MM (2010) distribution, and reproduction in any medium, provided you give Engineering carbon materials from the hydrothermal carbonization appropriate credit to the original author(s) and the source, provide a link process of biomass. Adv Mater 22:813–828. https://doi.org/10. to the Creative Commons license, and indicate if changes were made. 1002/adma.200902812 Jolliffe IT (2002) Principal component analysis. Springer series in statis- tics, 2nd edn. Springer, New York Jung C, Park J, Lim KH, Park S, Heo J, Her N, Oh J, Yun S, Yoon Y (2013) Adsorption of selected endocrine disrupting compounds and References pharmaceuticals on activated biochars. J Hazard Mater 263(Pt 2): 702–710. https://doi.org/10.1016/j.jhazmat.2013.10.033 Berge ND, Ro KS, Mao J, Flora JRV, Chappell MA, Bae S (2011) Khan GA, Lindberg R, Grabic R, Fick J (2012) The development and Hydrothermal carbonization of municipal waste streams. Environ pplication of a system for simultaneously determining anti- Sci Technol 45:5696–5703. https://doi.org/10.1021/es2004528 infectives and nasal decongestants using on-line solid-phase extrac- Bergius F (1913) Die Anwendung hoher Drucke bei chemischen tion and liquid chromatography-tandem mass spectrometry. J Pharm Vorgängen und eine Nachbildung des Entstehungsprozesses der Biomed Anal 66:24–32. https://doi.org/10.1016/j.jpba.2012.02.011 Steinkohle. Halle a. S.: Knapp Li Y, Shao J, Wang X, Deng Y, Yang H, Chen H (2014) Characterization Bustin RM, Guo Y (1999) Abrupt changes (jumps) in reflectance values of modified biochars derived from bamboo pyrolysis and their uti- and chemical compositions of artificial charcoals and inertinite in lization for target component (furfural) adsorption. Energy Fuel 28: coals. Int J Coal Geol 38:237–260. https://doi.org/10.1016/S0166- 5119–5127. https://doi.org/10.1021/ef500725c 5162(98)00025-1 Li J, Yu G, Pan L, Li C, You F, Xie S, Wang Y, Ma J, Shang X (2018) Calisto V, Ferreira CIA, Oliveira JABP, Otero M, Esteves VI (2015) Study of ciprofloxacin removal by biochar obtained from used tea Adsorptive removal of pharmaceuticals from water by commercial leaves. J Environ Sci. https://doi.org/10.1016/j.jes.2017.12.024 Environ Sci Pollut Res (2018) 25:15793–15801 15801 Lindberg RH, Östman M, Olofsson U, Grabic R, Fick J (2014) Shuttleworth PS, Baccile N, White RJ, Nectoux E, Budarin VL (2015) Bulk and surface analysis of carbonaceous materials. In: White RJ Occurrence and behaviour of 105 active pharmaceutical ingredients in sewage waters of a municipal sewer collection system. Water Res (ed) Porous carbon materials from sustainable precursors. The Royal 58:221–229. https://doi.org/10.1016/j.watres.2014.03.076 Society of Chemistry, Cambridge, pp 311–354 Liu Z, Zhang F-S, Wu J (2010) Characterization and application of chars Snyder S, Westerhoff P, Yoon Y, Sedlak D (2003) Disruptors in water: produced from pinewood pyrolysis and hydrothermal treatment. implications for the water industry. Environ Eng Sci 20:449–469. Fuel 89:510–514. https://doi.org/10.1016/j.fuel.2009.08.042 https://doi.org/10.1089/109287503768335931 Melvin SD, Leusch FDL (2016) Removal of trace organic contaminants Stone M (1973) Cross-validatory choice and assessment of statistical from domestic wastewater: a meta-analysis comparison of sewage predictions. J R Stat Soc Ser B 36:111–147 treatment technologies. Environ Int 92–93:183–188. https://doi.org/ Sun K, Jin J, Keiluweit M, Kleber M, Wang Z, Pan Z, Xing B (2012) 10.1016/j.envint.2016.03.031 Polar and aliphatic domains regulate sorption of phthalic acid esters Meng J, Wang L, Liu X, Wu J, Brookes PC, Xu J (2013) Physicochemical (PAEs) to biochars. Bioresour Technol 118:120–127. https://doi.org/ properties of biochar produced from aerobically composted swine 10.1016/j.biortech.2012.05.008 manure and its potential use as an environmental amendment. Titirici M-M, Antonietti M (2010) Chemistry and materials options of Bioresour Technol 142:641–646. https://doi.org/10.1016/j.biortech. sustainable carbon materials made by hydrothermal carbonization. 2013.05.086 Chem Soc Rev 39:103–116. https://doi.org/10.1039/b819318p Mestre AS, Tyszko E, Andrade MA, Galhetas M, Freire C, Carvalho AP Vom Eyser C, Palmu K, Schmidt TC, Tuerk J (2015) Pharmaceutical load (2015) Sustainable activated carbons prepared from a sucrose- in sewage sludge and biochar produced by hydrothermal carboniza- derived hydrochar: remarkable adsorbents for pharmaceutical com- tion. Sci Total Environ 537:180–186. https://doi.org/10.1016/j. pounds. RSC Adv 5:19696–19707. https://doi.org/10.1039/ scitotenv.2015.08.021 C4RA14495C Wiedner K, Rumpel C, Steiner C, Pozzi A, Maas R, Glaser B (2013) Mohan D, Sarswat A, Ok YS, Pittman CU (2014) Organic and inorganic Chemical evaluation of chars produced by thermochemical conver- contaminants removal from water with biochar, a renewable, low sion (gasification, pyrolysis and hydrothermal carbonization) of agro- cost and sustainable adsorbent—a critical review. Bioresour Technol industrial biomass on a commercial scale. Biomass Bioenergy 59: 160:191–202. https://doi.org/10.1016/j.biortech.2014.01.120 264–278. https://doi.org/10.1016/j.biombioe.2013.08.026 Nyenje PM, Foppen JW, Uhlenbrook S, Kulabako R, Muwanga A (2010) Yang GX, Jiang H (2014) Amino modification of biochar for enhanced Eutrophication and nutrient release in urban areas of sub-Saharan adsorption of copper ions from synthetic wastewater. Water Res 48: Africa—a review. Sci Total Environ 408:447–455. https://doi.org/ 396–405. https://doi.org/10.1016/j.watres.2013.09.050 10.1016/j.scitotenv.2009.10.020 Yu L, Falco C, Weber J, White RJ, Howe JY, Titirici MM (2012) Oliveira I, Blöhse D, Ramke HG (2013) Hydrothermal carbonization of Carbohydrate-derived hydrothermal carbons: a thorough character- agricultural residues. Bioresour Technol 142:138–146. https://doi. ization study. Langmuir 28:12373–12383. https://doi.org/10.1021/ org/10.1016/j.biortech.2013.04.125 la3024277 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Environmental Science and Pollution Research Springer Journals

Using carbonized low-cost materials for removal of chemicals of environmental concern from water

Free
9 pages
Loading next page...
 
/lp/springer_journal/using-carbonized-low-cost-materials-for-removal-of-chemicals-of-0ANhK4CMqw
Publisher
Springer Berlin Heidelberg
Copyright
Copyright © 2018 by The Author(s)
Subject
Environment; Environment, general; Environmental Chemistry; Ecotoxicology; Environmental Health; Atmospheric Protection/Air Quality Control/Air Pollution; Waste Water Technology / Water Pollution Control / Water Management / Aquatic Pollution
ISSN
0944-1344
eISSN
1614-7499
D.O.I.
10.1007/s11356-018-1781-0
Publisher site
See Article on Publisher Site

Abstract

Adsorption on low-cost biochars would increase the affordability and availability of water treatment in, for example, developing countries. The aim of this study was to identify the precursor materials and hydrochar surface properties that yield efficient removal of compounds of environmental concern (CEC). We determined the adsorption kinetics of a mixture containing ten CECs (octhilinone, triclosan, trimethoprim, sulfamethoxasole, ciprofloxacin, diclofenac, paracetamol, diphenhydramine, flucon- azole, and bisphenol A) to hydrochars prepared from agricultural waste (including tomato- and olive-press wastes, rice husks, and horse manure). The surface characteristics of the hydrochars were evaluated via diffuse reflectance infrared spectroscopy (DRIFTS), X-ray photoelectron spectroscopy (XPS), and N -adsorption. Kinetic adsorption tests revealed that removal efficiencies varied substantially among different materials. Similarly, surface analysis revealed differences among the studied hydrochars and the degree of changes that the materials undergo during carbonization. According to the DRIFTS data, compared with the least efficient adsorbent materials, the most efficient hydrochars underwent more substantial changes during carbonization. . . . . . Keywords Hydrochar Adsorption Hydrothermal carbonization Agro-industrial residues Organic chemicals Low-cost adsorbents Introduction and adsorption on activated carbon (Gupta and Suhas 2009; Nyenje et al. 2010). Additionally, even if current sewage treat- Water treatment poses a serious global challenge, in terms of ment plants reduce the spread of bacteria and nutrients, some providing safe sanitation and using limited water resources in chemicals pass though the treatment plant unreduced (Snyder an economical and efficient manner that ensures the availabil- et al. 2003;Lindberget al. 2014; Melvin and Leusch 2016). ity of clean water for everyone. However, high costs and tech- This is especially crucial for chemicals of environmental con- nical complexity limit the use of conventional water-treatment cern (CEC), e.g., antibiotics, pharmaceuticals, and biocides, techniques, including sewage treatment, oxidation, filtration, that are designed to have biological effects. Studies have shown that some pharmaceuticals can be absorbed by crops, and subsequently enter both humans and livestock (Franklin Responsible editor: Philippe Garrigues et al. 2016). The availability of water treatment can be in- Electronic supplementary material The online version of this article creased via simple, inexpensive techniques. For example, ad- (https://doi.org/10.1007/s11356-018-1781-0) contains supplementary sorption, which removes various contaminants even at low material, which is available to authorized users. concentrations, is easily realized, but is costly due to the lim- ited lifetime of adsorbents prepared from non-local or fossil * Stina Jansson feedstock (Gupta and Suhas 2009). By valorizing local low- stina.jansson@umu.se cost feedstocks, such as agricultural or food industry residues, Department of Chemistry, Umeå University, SE-901 adsorption becomes affordable also in countries and regions 87 Umeå, Sweden with high water stress and limited economical resources Umeå Energi AB, SE-901 05 Umeå, Sweden (Mohan et al. 2014). Biochars are carbon-rich porous materials obtained via the Industrial Doctoral School, Umeå University, SE-901 carbonization of feedstock materials, and they may be used as 87 Umeå, Sweden 15794 Environ Sci Pollut Res (2018) 25:15793–15801 adsorbent materials due to their similarity to commercial acti- subsequently carbonized at 220 °C for 2 h. Portions vated carbons. Biochars generated from wet feedstocks are (~ 600 mL) of the wet sample were carbonized during each produced preferably through hydrothermal carbonization experiment and yields of 57–66% (66, 62, 57, and 59% for (HTC), an energy-efficient wet technique performed without rice husks, manure, tomato waste, and olive waste, respective- drying of the raw materials before carbonization. The resulting ly) were realized. After carbonization, the chars were retrieved biochar, or more specifically the hydrochar, is more hydropho- via filtration and dried overnight at 105 °C. bic than the raw material and is therefore easily dried (Escala Prior to use, the chars were demineralized, i.e., dissolvable et al. 2013; Vom Eyser et al. 2015). HTC was first described in material was removed, and the surface functionalities were 1913 (Bergius 1913) and nowadays is considered one of the protonated by washing the chars in 0.1 M HCl until a clear most sustainable approaches for obtaining functional carbon- water phase was obtained (Meng et al. 2013), rinsing with based materials (Hu et al. 2010; Titirici and Antonietti 2010). ultrapure water and filtering (Munktell, Qualitative filter paper This process involves the use of organic materials (such as grade 3/< 10 μm), and drying at 105 °C for 12 h. All chars carbohydrates, manure, sludge, wood, food, and other agricul- were homogenized by grinding in a mortar, except for the tural wastes) as precursors (Hu et al. 2010; Berge et al. 2011; tomato waste chars, which were blitzed for 4 × 2-s pulses at Oliveira et al. 2013; Falco et al. 2013). Hydrochars often retain 10000 rpm (Grindomix GM 200, Retch). The dry chars were much of the surface functionalities from the raw materials then stored in a desiccator. (mostly oxygen and hydrogen-containing acidic groups, e.g., phenolic, lactonic, carboxyl, and carbonyl groups (Wiedner Preparation of model water et al. 2013)) which increase interactions with ionic or polar compounds (Liu et al. 2010). Although different types of bio- Our model water contained ten different CEC substances: two chars have gained attention recently and many potential feed- biocides: octhilinone (2-octyl-4-isothiazolin-3-one) and triclo- stocks have been identified (Mohan et al. 2014), the data on san; three antibiotics: trimethoprim, sulfamethoxasole, and these materials is inconsistent. This inconsistency results from ciprofloxacin; four pharmaceuticals: diclofenac, paracetamol, the different carbonization methods and parameters as well as diphenhydramine, and fluconazole; and the plastic additive adsorbates and adsorption test parameters employed. bisphenol A in concentrations of 10 μg/L in ultrapure water. Additionally, complex chemical mixtures and simulating real The corresponding trade names, structures, CAS (Chemical wastewater, should be investigated instead of focusing on so- Abstracts Service) registry numbers, solubility in water, log lutions containing only one compound. K , and main applications are shown in Fig. S1 OW In this screening study, the CEC-removal efficiencies of (Supplementary data). The tested substances were selected, four hydrochars prepared from low-cost feedstocks were de- owing to their widespread use in developing countries as well termined. The sorption properties were elucidated through as their ability to pass unreduced through conventional sew- surface characterization where the role of surface area and age treatment plants and being emitted into the environment surface functionalities in the removal of contaminants from through effluent water (Snyder et al. 2003; Lindberg et al. water, was evaluated. 2014; Melvin and Leusch 2016). Batch adsorption experiments Materials and methods The experiments were all performed at 20 °C and were con- Preparation of hydrochars ducted as follows: Triplicate adsorption tests were performed using 50 ± 2.5 mg hydrochar + 10 mL model water in 15 mL Four organic residues were used in this study: horse manure, plastic tubes. After weighing the hydrochar, water was added skin and seeds from tomato, olive press residues, and rice to the tube, which was then agitated for 1, 3, 5, 8, 12, 18, or husks. These materials are considered representatives of com- 25 min. Afterwards, ca 5 mL of water/hydrochar was removed mon agricultural residues in countries with high water stress. from the tube using a syringe, passed through a 0.45-μmsy- The hydrochars were prepared in a stirred high-pressure ringe filter (Filtopur, Sarstedt) into a glass vial, and weighed lab-scale reactor with an internal volume of 1 L (Zhengzhou before addition of the internal standard. The internal standard Keda Machinery and Instrument Equipment Co., Ltd., China). consisted of isotopically labeled bisphenol A, triclosan, tri- The walls of the reactor were heated by a detachable resistance methoprim, sulfamethoxasole, ciprofloxacin, diclofenac, heater. After each experiment, the heater was removed and the paracetamol, and fluconazole. Furthermore, the samples were reaction was water-cooled. The four feedstock materials were all either analyzed directly or frozen until analysis, which covered with ultrapure water, resulting in mixtures with dry- occurred within 3 weeks of the experiment. matter content ranging from 11 to 24% (rice husks 11%, ma- For each hydrochar, triplicate blank samples were prepared nure 12%, tomato waste 12%, and olive waste 24%) and via agitation with ultrapure water. Leaching of the analytes by Environ Sci Pollut Res (2018) 25:15793–15801 15795 −1 −1 the tubes was investigated by agitating (for 25 min) triplicate 5000 cm was collected at a spectral resolution of 4 cm . tube blanks of 10 mL ultrapure water. Similarly, adsorption of In addition, pure KBr was used as the background, and the −1 the analytes to the tube walls was investigated through a trip- spectral range 400–3750 cm was used in the subsequent licate tube adsorption test where 10 mL of the model water multivariate analysis. Prior to multivariate analysis, the spec- was agitated (for 25 min) in the tubes. In addition, the capacity tra were baseline-corrected (asymmetric least squares, lamb- of the chars (relative to that of commercially available active da = 14,000,000, p = 0.01), smoothed (Savitzky-Golay filter- carbon) was assessed by agitating triplicate samples of ing, polynomial order = 1, frame = 5), and total-area normal- general-purpose-grade powdered activated carbon (Fisher ized, using the protocol described by Felten et al. 2015. Scientific) with the model water for 25 min. A list of samples, agitation times, and blank samples is provided in Table S1 X-ray photoelectron spectroscopy (XPS) (Supplementary data). The XPS spectra were collected with a Kratos Axis Ultra DLD LC-MS analysis electron spectrometer using a monochromatic AlKα source operated at 120 W. An analyzer pass energy of 160 eV and a The samples were all analyzed using a Thermo TSQ Quantum pass energy of 20 eV were used for acquiring wide spectra and Ultra EMR (Thermo Fisher Scientific, San Jose, CA, USA) individual photoelectron lines, respectively. The surface po- mass spectrometer coupled to a PAL HTC auto sampler (CTC tential was stabilized by the spectrometer charge neutraliza- Analytics AG, Zwingen, Switzerland). Two pumps (Surveyor tion system. The binding energy (BE) scale was referenced to and Accela, Thermo Fisher Scientific, San Jose, CA, USA) the C1s line (set at 285.0 eV) of aliphatic carbon, and the were used, and the separation was achieved with a Thermo spectra were processed with the Kratos software. Powder sam- Hypersil Gold AQ (50 × 2.1 mm, 5 μm) column. Analytes ples for the analysis were gently hand-pressed into a pellet were ionized via heated electrospray (HESI) or atmospheric (directly on a sample holder) using a clean Ni spatula. pressure photoionization (APPI) using a krypton lamp at 10.6 eV, in positive- or negative-ion mode. The settings, in- Principal component analysis (PCA) cluding the HESI/APPI ionization data, polarities, precursor/ product ions, collision energies, tube lens values, quantifica- The differences between the materials were graphically visu- tion and qualification ions, and limits of quantification (LOQ), alized via Principal Component Analysis (PCA) (Jolliffe associated with the analysis are summarized in Table S2 2002). In PCA, the variation in a dataset is determined by (Supplementary data). The analytical system is based on col- extracting orthogonal principal components from a larger umn switching using six- and ten-port valves; the basic setup number of variables. The first principal component (PC1) ac- is described by Khan et al. (2012) (Khan et al. 2012). During counts for the largest variation in the dataset, PC2 accounts for analysis, a spectral resolution corresponding to a full width the second largest, and so on. Score plots were used to display half maximum (FWHM) of 0.7 was used for both distributions of the observations (here, the carbonized and quadrupoles. non-carbonized materials) projected onto a plane. In these distributions, similar samples are closely grouped, whereas Surface-area determination dissimilar samples are separated by large distances. The SIMCA-P software package (version 13.0, Umetrics AB, The specific surface area of the hydrochars was determined Sweden) was used for the PCA modeling and all variables via N adsorption and calculated using the Brunauer– were center-scaled. Furthermore, the number of significant Emmett–Teller (BET) theory. Degassing was performed at components was determined via sevenfold full cross- 120 °C. validation (CV) (Stone 1973), and components with eigen- values lower than two were removed from the model. Diffuse reflectance infrared fourier transform spectroscopy (DRIFTS) Results and discussion Approximately 10 mg of dry sample was manually ground with ca. 390 mg Fourier transform-infrared (FT-IR) Overall adsorption efficiency spectroscopy-grade KBr (Merck, Darmstadt, Germany) in an agate mortar. The corresponding FT-IR spectra were recorded The different hydrochars exhibited differing removal proper- in diffuse-reflectance mode (Bruker IFS 66v/S; Bruker Optik ties, and CEC removal from the water differed between sub- GmbH, Ettlingen, Germany) under vacuum conditions, in ac- stance and char. Furthermore, the substances were all rapidly cordance with the protocol described by Gorzsás and removed, and the amount removed saturated after 1 to 3 min Sundberg 2014. Data in the spectral range of 400– (see Figs. S2-S5 in Supplementary data for further details of 15796 Environ Sci Pollut Res (2018) 25:15793–15801 Fig. 1 Total amount of removed CEC per gram of char. Removal was extremely rapid, thereby preventing the determination of gradients and, hence, the values displayed correspond to the average of all 21 samples. Error bars denote one standard deviation. Maximum amount removable: 2 μg/g hydrochar the substances and chars). This rapid removal enabled the use model water. Horse manure char was the most efficient mate- of a total (overall)-removal efficiency for each substance and rial for removing paracetamol, fluconazole, and sulfamethox- char, thereby improving the statistical accuracy of the removal azole, even though these compounds were generally removed results (Fig. 1). to a low degree. Moreover, the CECs in the 18 blank samples Horse manure and rice husk chars exhibited the highest were all lower than the LOQ (see Table S2, Supplementary overall removal with CEC average removal efficiencies of data), and, after 25 min, the analytes were all removed by 49 and 66%, respectively. These values are higher than the activated carbon. However, some of the analytes (especially respective values (39 and 32%) associated with olive and to- diphenhydramine, octhilinone, and bisphenol A (Fig. 1)) were mato residue chars. Diclofenac and bisphenol A were nearly adsorbed onto the tube walls. completely removed by all four chars. Similarly, rice husk Similar results have been found in previous studies. Jung char exhibited high efficiency (≥ 83%) of ciprofloxacin, di- et al. (2013) reported nearly complete adsorption of bisphenol phenhydramine, octhilinone, and triclosan removal from the A using activated carbons from loblolly pine chips, while Fig. 2 Correlation between the hydrophobicity of the adsorbate and the concentration adsorbed on each hydrochar Environ Sci Pollut Res (2018) 25:15793–15801 15797 with log K was observed (Fig. 2). Compared with their OW lower-log K counterparts, the substances with higher log OW K values were, in general, more efficiently removed by the OW chars. Removal by rice husk char exhibited the highest (R = 0.62, p = 0.0069) positive correlation with log K ,suggest- OW ing that this removal has the best fitting linear correlation with log K values of the CECs. The correlation obtained for OW carbonized olive waste and tomato waste, although lower than that of the carbonized rice husks, was still significant. Manure char exhibited the lowest overall correlation (R =0.31, p = 0.09). When the amount removed was plotted as a function of log K , the steepest slopes, 0.34, 0.34, and 0.30, of the least OW square fit were obtained for the olive waste char, rice husk char, and tomato waste char, respectively. Compared with these plots, the plot corresponding to the horse manure char exhibited a more gradual slope (0.24), suggesting that the amount removed is more independent of log K .Low-log OW K compounds, such as paracetamol and fluconazole, were OW most efficiently removed by horse manure char. This is attrib- uted to the dissimilar surface properties that may promote different removal mechanisms. Hydrophobic molecules, in Fig. 3 Baseline-corrected DRIFTS spectra of the treated materials general, attach to surfaces via hydrophobic interactions, whereas their more hydrophilic counterparts are removed diclofenac was less efficiently removed and sulfamethoxazole through other electrostatic interactions (e.g., hydrogen bond- the least adsorbed, and the adsorption followed the same order ing) (Sun et al. 2012). as in this study. Sulfamethoxazole removal by primary paper mill sludge char exceeded 50% in a study by Calisto and co- Surface characterization workers (2015), and in a study using activated sucrose hydrochar, the paracetamol removal was around 50% or less The surface properties of the carbonized material were inves- (Mestre et al. 2015). The low removal by these different tigated via three surface analysis techniques namely, BET, hydrochars suggests that these types of compounds are not DRIFTS, and XPS. These analyses showed that the surface easily adsorbed from the solution. Nevertheless, comparisons properties of rice husk and manure chars differ significantly should made with caution, because the concentrations in the from those of the tomato- and olive-residue chars. above-mentioned studies were in the mg range. This is many In addition, BET-determined specific surface areas of 16.92 orders of magnitude higher than the concentrations used in and 4.62m /g were obtained for rice husk char and horse ma- this study. Thus, it is likely that the solute concentration will nure char, The surface area of the tomato and olive waste chars affect the removal efficiency. This was indicated by Li et al. 2 2 were below 1 m /g (0.74 and 0.65 m /g respectively), and 2018, where ciprofloxacin removal efficiency was less for the these values were associated with high levels of uncertainty lowest studied concentration compared to higher concentra- due to both the small surface area and the large fraction of tions in the range of 150–500 mg/l. volatiles on the char surfaces. This may have resulted from the Correlations among the amount adsorbed, water solubility relatively low carbonization temperature (220 °C). Moreover, of each substance, and log K were determined. No corre- OW considering that rice husk char had nearly fourfold higher lation was found between the removed amount of CEC and surface area compared to horse manure char (16.92 vs. their solubility; however, in some cases, a positive correlation 4.62 m /g), the overall removal for horse manure char (49%) was noteworthy. This indicates that the surface area consti- Table 1 Elemental composition (atomic-%) of the surface tutes only one factor that drives the adsorption. The DRIFTS spectra of the char materials and the raw Element Rice husks Horse manure Olive waste Tomato waste materials are shown in Fig. 3 and Fig. S6 (Supplementary data), respectively. Broad O-H stretching bands from hydrox- C 54.30 74.52 87.84 83.71 −1 yl groups (Franca et al. 2010) occurred at 3500–3300 cm for O 37.69 23.57 11.51 14.71 all samples, but occurred with higher intensity in the horse N – 1.52 0.64 1.58 manure and rice husk chars. Furthermore, the materials all Si 8.01 0.38 –– consist of aliphatic CH , as indicated by the occurrence of 2 15798 Environ Sci Pollut Res (2018) 25:15793–15801 Fig. 4 Deconvoluted C1s lines of the horse manure char, rice husk char, tomato waste char, and olive waste char −1 C-H stretching vibrations at 2925 and 2853 cm , and small functionalities. Additionally, a large band at 1100– −1 −1 bands at 1450–1430 and 1370 cm (Chen and Chen 2009). 1000 cm arising from the rice husk sample results probably The bands associated with the olive and tomato waste were from Si-O-Si in-plane vibrations, as suggested by XPS analy- more intense than those associated with manure and rise husks sis results that revealed the presence of silicon (Table 1). The before and after carbonization. Moreover, the sharp C = O spectra of the treated materials are less diverse than those of −1 stretching band occurring at 1750–1700 cm results from the untreated materials, i.e., the intensities of C-H and C = O, carboxylic acids, esters, ketones, lactones, or aldehydes and C-O bands vary more within untreated materials com- (Esteves et al. 2013). The band associated with the olive and pared to the treated ones. tomato wastes is sharper than that corresponding to the ma- The elemental composition of the biochar surface is deter- nure and rice husks. Furthermore, the occurrence of bands at mined in further detail via XPS analysis (see Table 1). As the −1 1600–1450 cm is indicative of aromatic ring C = C table shows, the carbon content of the olive and tomato waste stretching in the samples (Chen and Chen 2009;Lietal. chars is higher than that of the rice husk and horse manure −1 2014), and small sharp bands at 900–700 cm may have chars. This result concurs with the DRIFT spectra, which ex- originated from aromatic C-H out-of-plane vibrations (Fang hibits high intensities for aliphatic carbon vibrations. In addi- −1 et al. 2014). Several narrow bands at 1300–1000 cm are tion, silicon (bound as SiO ) occurs at a rather high concen- attributed to O-H bending and C-O stretching in ethers, alco- tration in the rice husk char, but constitutes only trace amounts hols, phenols, lactones or carboxyl acids, and anhydrides of the horse manure char. Small amounts of carbon-bound (Bustin and Guo 1999; Shuttleworth et al. 2015). The en- nitrogen occur in all samples, except for the rice husk chars hancement of these bands after hydrothermal carbonization (Yang and Jiang 2014), and other elements occur at levels of the materials is indicative of changes in the surface lower than the limit of detection (~ 0.1 atomic-%). Environ Sci Pollut Res (2018) 25:15793–15801 15799 Fig. 5 Plot showing the first two principal components of the PCA model of the DRIFTS spectra Figure 4 shows the deconvoluted C1s lines of the the removal of more polar molecules via H-bonding (Sun et al. hydrochars. The peaks occurring at 284.9–285.0 eV corre- 2012), than the other chars considered. spond to aliphatic and/or graphitic carbon, the dominant spe- cies in all the chars. Compared with the olive and tomato waste chars, rice husk and manure chars contain higher frac- Multivariate data modeling tions of carbon species with a single bond to oxygen, i.e., hydroxyls and ethers (at 286.5–286.8 eV), which support the The differences between the materials were elucidated by previously presented DRIFTS data. Carbonyls at 288.1– performing a PCA using the baseline-corrected DRIFTS spec- 288.6 eV and esters or carboxyl acids at 289.2–289.6 eV oc- tra. The resulting model consists of four principal components curred in only small concentrations. Additionally, a minor line that account for 86% of the variability (R2) with 69% predic- at 291.2 eV associated with the horse manure char, results tive ability (Q2; see Fig. 5 for the score plot of the first two from either a carbonate or π-π* shakeup satellite (Yu et al. components). The PCA reveals that the treated materials are 2012). more homogeneous, as evidenced by the close grouping of the Surface analyses revealed substantial differences between char replicates, than the untreated material. The PCA score the studied biochars. Although BET-determined surface areas plot also shows that the first principal component accounts varied between the best- (manure and rice husk chars) and for 57% of the variability (R2) and 42% of the predictive worst-performing (tomato waste and olive waste chars) mate- ability (Q2) (Fig. 5, t[1]). Furthermore, compared with that rials, this factor constituted only one of the several important occurring for materials with poor removal capacity (i.e., toma- features affecting this performance. DRIFTS and XPS re- to and olive waste), the variability in the DRIFTS spectra is vealed differences in the carbon content and O-functionalities, higher for materials with good removal capacity (i.e., rice especially the OH content. Furthermore, the SiO and OH husks and manure). The inter-surface differences between content of the rice husk and horse manure chars is higher than high- and low-adsorption materials are larger than those be- those of the tomato waste and olive waste chars, whereas the tween carbonized and non-carbonized materials (which are aliphatic carbon content is lower. Therefore, the rice husk and distinguished by the second principal component, t[2] with, horse manure chars are more polar, and thereby may promote variability (R2): 17%, predictive ability (Q2): 4%). The PCA 15800 Environ Sci Pollut Res (2018) 25:15793–15801 and waste-based carbons. J Environ Manag 152:83–90. https://doi. model also revealed that the olive and tomato chars are very org/10.1016/j.jenvman.2015.01.019 similar. Chen B, Chen Z (2009) Sorption of naphthalene and 1-naphthol by bio- The overall results of the PCA model of the DRIFTS spec- chars of orange peels with different pyrolytic temperatures. tra may allow screening of untreated materials (based on sur- Chemosphere 76:127–133. https://doi.org/10.1016/j.chemosphere. 2009.02.004 face functionalities) and prediction of removal capacities. Escala M, Zumbühl T, Koller C, Junge R, Krebs R (2013) Hydrothermal However, a large database and multiple runs would be needed carbonization as an energy-efficient alternative to established drying for highly heterogeneous untreated material (for example, ma- technologies for sewage sludge: a feasibility study on a laboratory nure or tomato waste). scale. Energy Fuels 27:454–460. https://doi.org/10.1021/ef3015266 Esteves B, Velez Marques A, Domingos I, Pereira H (2013) Chemical changes of heat treated pine and eucalypt wood monitored by FTIR. Maderas Cienc y Tecnol 15. https://doi.org/10.4067/S0718- 221X2013005000020 Conclusions Falco C, Marco-Lozar JP, Salinas-Torres D, Morallón E, Cazorla-Amorós D, Titirici MM, Lozano-Castelló D (2013) Tailoring the porosity of These findings demonstrate that carbonized low-value mate- chemically activated hydrothermal carbons: influence of the precur- sor and hydrothermal carbonization temperature. Carbon 62:346– rials can remove CEC from water, but the removal efficiency 355. https://doi.org/10.1016/j.carbon.2013.06.017 varies with the feedstock. Furthermore, DRIFTS and XPS Fang Q, Chen B, Lin Y, Guan Y (2014) Aromatic and hydrophobic analyses revealed significant differences in the elemental com- surfaces of wood-derived biochar enhance perchlorate adsorption position and functionalities of the hydrochars. Multivariate via hydrogen bonding to oxygen-containing organic groups. Environ Sci Technol 48:279–288. https://doi.org/10.1021/ analysis based on the DRIFTS data showed that, compared es403711y with those corresponding to the worst-performing hydrochars, Felten J, Hall H, Jaumot J, Tauler R, de Juan A, Gorzsás A (2015) larger differences between the treated and untreated materials Vibrational spectroscopic image analysis of biological material occurred for the most promising hydrochars. using multivariate curve resolution–alternating least squares (MCR-ALS). Nat Protoc 10:217–240. https://doi.org/10.1038/ nprot.2015.008 Acknowledgements The DRIFTS analysis was performed, with the as- Franca AS, Oliveira LS, Nunes AA, Alves CCO (2010) Microwave sistance of András Gorzsás, at the Vibrational Spectroscopy Core Facility assisted thermal treatment of defective coffee beans press cake for at the KBC, Umeå University, and the XPS analysis and interpretation the production of adsorbents. Bioresour Technol 101:1068–1074. were conducted by Andrey Shchukarev at Umeå University. The authors https://doi.org/10.1016/j.biortech.2009.08.102 also gratefully acknowledge Professor Andres Fullana from the Franklin AM, Williams CF, Andrews DM, Woodward EE, Watson JE University of Alicante for use of their HTC facility. Professor Jean- (2016) Uptake of three antibiotics and an antiepileptic drug by wheat François Boily at Umeå University is gratefully acknowledged for FT- crops spray irrigated with wastewater treatment plant effluent. J IR interpretation. Part of the study was financed by grants from Environ Qual 45:546–554. https://doi.org/10.2134/jeq2015.05.0257 Ångpanneföreningen’s Foundation for Research and Development and Gorzsás A, Sundberg B (2014) Chemical fingerprinting of arabidopsis The J. Gust. Richert Memorial Fund. We also thank The Swedish using fourier transform infrared (FT-IR) spectroscopic approaches. Research Council and Bio4Energy, a strategic research environment In: Sanchez-Serrano J, Salinas J (eds) Arabidopsis Protocols. appointed by the Swedish government, for supporting this work. Humana Press, Totowa, pp 317–352 Gupta VK, Suhas (2009) Application of low-cost adsorbents for dye Open Access This article is distributed under the terms of the Creative removal—a review. J Environ Manag 90:2313–2342. https://doi. Commons Attribution 4.0 International License (http:// org/10.1016/j.jenvman.2008.11.017 creativecommons.org/licenses/by/4.0/), which permits unrestricted use, Hu B, Wang K, Wu L, Yu SH, Antonietti M, Titirici MM (2010) distribution, and reproduction in any medium, provided you give Engineering carbon materials from the hydrothermal carbonization appropriate credit to the original author(s) and the source, provide a link process of biomass. Adv Mater 22:813–828. https://doi.org/10. to the Creative Commons license, and indicate if changes were made. 1002/adma.200902812 Jolliffe IT (2002) Principal component analysis. Springer series in statis- tics, 2nd edn. Springer, New York Jung C, Park J, Lim KH, Park S, Heo J, Her N, Oh J, Yun S, Yoon Y (2013) Adsorption of selected endocrine disrupting compounds and References pharmaceuticals on activated biochars. J Hazard Mater 263(Pt 2): 702–710. https://doi.org/10.1016/j.jhazmat.2013.10.033 Berge ND, Ro KS, Mao J, Flora JRV, Chappell MA, Bae S (2011) Khan GA, Lindberg R, Grabic R, Fick J (2012) The development and Hydrothermal carbonization of municipal waste streams. Environ pplication of a system for simultaneously determining anti- Sci Technol 45:5696–5703. https://doi.org/10.1021/es2004528 infectives and nasal decongestants using on-line solid-phase extrac- Bergius F (1913) Die Anwendung hoher Drucke bei chemischen tion and liquid chromatography-tandem mass spectrometry. J Pharm Vorgängen und eine Nachbildung des Entstehungsprozesses der Biomed Anal 66:24–32. https://doi.org/10.1016/j.jpba.2012.02.011 Steinkohle. Halle a. S.: Knapp Li Y, Shao J, Wang X, Deng Y, Yang H, Chen H (2014) Characterization Bustin RM, Guo Y (1999) Abrupt changes (jumps) in reflectance values of modified biochars derived from bamboo pyrolysis and their uti- and chemical compositions of artificial charcoals and inertinite in lization for target component (furfural) adsorption. Energy Fuel 28: coals. Int J Coal Geol 38:237–260. https://doi.org/10.1016/S0166- 5119–5127. https://doi.org/10.1021/ef500725c 5162(98)00025-1 Li J, Yu G, Pan L, Li C, You F, Xie S, Wang Y, Ma J, Shang X (2018) Calisto V, Ferreira CIA, Oliveira JABP, Otero M, Esteves VI (2015) Study of ciprofloxacin removal by biochar obtained from used tea Adsorptive removal of pharmaceuticals from water by commercial leaves. J Environ Sci. https://doi.org/10.1016/j.jes.2017.12.024 Environ Sci Pollut Res (2018) 25:15793–15801 15801 Lindberg RH, Östman M, Olofsson U, Grabic R, Fick J (2014) Shuttleworth PS, Baccile N, White RJ, Nectoux E, Budarin VL (2015) Bulk and surface analysis of carbonaceous materials. In: White RJ Occurrence and behaviour of 105 active pharmaceutical ingredients in sewage waters of a municipal sewer collection system. Water Res (ed) Porous carbon materials from sustainable precursors. The Royal 58:221–229. https://doi.org/10.1016/j.watres.2014.03.076 Society of Chemistry, Cambridge, pp 311–354 Liu Z, Zhang F-S, Wu J (2010) Characterization and application of chars Snyder S, Westerhoff P, Yoon Y, Sedlak D (2003) Disruptors in water: produced from pinewood pyrolysis and hydrothermal treatment. implications for the water industry. Environ Eng Sci 20:449–469. Fuel 89:510–514. https://doi.org/10.1016/j.fuel.2009.08.042 https://doi.org/10.1089/109287503768335931 Melvin SD, Leusch FDL (2016) Removal of trace organic contaminants Stone M (1973) Cross-validatory choice and assessment of statistical from domestic wastewater: a meta-analysis comparison of sewage predictions. J R Stat Soc Ser B 36:111–147 treatment technologies. Environ Int 92–93:183–188. https://doi.org/ Sun K, Jin J, Keiluweit M, Kleber M, Wang Z, Pan Z, Xing B (2012) 10.1016/j.envint.2016.03.031 Polar and aliphatic domains regulate sorption of phthalic acid esters Meng J, Wang L, Liu X, Wu J, Brookes PC, Xu J (2013) Physicochemical (PAEs) to biochars. Bioresour Technol 118:120–127. https://doi.org/ properties of biochar produced from aerobically composted swine 10.1016/j.biortech.2012.05.008 manure and its potential use as an environmental amendment. Titirici M-M, Antonietti M (2010) Chemistry and materials options of Bioresour Technol 142:641–646. https://doi.org/10.1016/j.biortech. sustainable carbon materials made by hydrothermal carbonization. 2013.05.086 Chem Soc Rev 39:103–116. https://doi.org/10.1039/b819318p Mestre AS, Tyszko E, Andrade MA, Galhetas M, Freire C, Carvalho AP Vom Eyser C, Palmu K, Schmidt TC, Tuerk J (2015) Pharmaceutical load (2015) Sustainable activated carbons prepared from a sucrose- in sewage sludge and biochar produced by hydrothermal carboniza- derived hydrochar: remarkable adsorbents for pharmaceutical com- tion. Sci Total Environ 537:180–186. https://doi.org/10.1016/j. pounds. RSC Adv 5:19696–19707. https://doi.org/10.1039/ scitotenv.2015.08.021 C4RA14495C Wiedner K, Rumpel C, Steiner C, Pozzi A, Maas R, Glaser B (2013) Mohan D, Sarswat A, Ok YS, Pittman CU (2014) Organic and inorganic Chemical evaluation of chars produced by thermochemical conver- contaminants removal from water with biochar, a renewable, low sion (gasification, pyrolysis and hydrothermal carbonization) of agro- cost and sustainable adsorbent—a critical review. Bioresour Technol industrial biomass on a commercial scale. Biomass Bioenergy 59: 160:191–202. https://doi.org/10.1016/j.biortech.2014.01.120 264–278. https://doi.org/10.1016/j.biombioe.2013.08.026 Nyenje PM, Foppen JW, Uhlenbrook S, Kulabako R, Muwanga A (2010) Yang GX, Jiang H (2014) Amino modification of biochar for enhanced Eutrophication and nutrient release in urban areas of sub-Saharan adsorption of copper ions from synthetic wastewater. Water Res 48: Africa—a review. Sci Total Environ 408:447–455. https://doi.org/ 396–405. https://doi.org/10.1016/j.watres.2013.09.050 10.1016/j.scitotenv.2009.10.020 Yu L, Falco C, Weber J, White RJ, Howe JY, Titirici MM (2012) Oliveira I, Blöhse D, Ramke HG (2013) Hydrothermal carbonization of Carbohydrate-derived hydrothermal carbons: a thorough character- agricultural residues. Bioresour Technol 142:138–146. https://doi. ization study. Langmuir 28:12373–12383. https://doi.org/10.1021/ org/10.1016/j.biortech.2013.04.125 la3024277

Journal

Environmental Science and Pollution ResearchSpringer Journals

Published: Mar 26, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off