Development of waste biomass based sorbent for removal of cyanotoxin microcystin-LR from aqueous phases

Accepted Manuscript Development of waste biomass based sorbent for removal of cyanotoxin microcystin-LR from aqueous phases Sok Kim, Yeoung-Sang Yun, Yoon-E Choi PII: DOI: Reference:

S0960-8524(17)31729-7 https://doi.org/10.1016/j.biortech.2017.09.164 BITE 18992

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

23 June 2017 20 September 2017 23 September 2017

Please cite this article as: Kim, S., Yun, Y-S., Choi, Y-E., Development of waste biomass based sorbent for removal of cyanotoxin microcystin-LR from aqueous phases, Bioresource Technology (2017), doi: https://doi.org/10.1016/ j.biortech.2017.09.164

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Development of waste biomass based sorbent for removal of cyanotoxin microcystin-LR from aqueous phases

Sok Kim1, Yeoung-Sang Yun2, Yoon-E Choi1, * 1

Division of Environmental Science and Ecological Engineering, Division of Environmental Science

and Ecological Engineering, College of Life Sciences and Biotechnology, Korea University, Seoul, 02841, Republic of Korea. 2

Division of Semiconductor and Chemical Engineering, Chonbuk National University, Jeonbuk,

54896, Republic of Korea.

*Corresponding author Tel: +82 2 3290 3042, Fax: +82 2 3290 3040, E-mail address: [email protected]

Abstract The purpose of this study was to establish the strategy to remove the cyanotoxin microcystin-LR (MC-LR) from aqueous solution with the use of biosorption strategy. Specifically, we focused on use of industrial waste biomass, Escherichia coli, to make efficient biosorbents for MC-LR through immobilization of the biomass with polysulfone (PS), coating the polysulfone-biomass composite with polyethylenimine (PEI), and decarboxylation of the PEI-coated composite to remove the inhibitory sites. The resulting sorbent is named in this study as decarboxylated PEI-coated polysulfone-biomass composite fiber (DC-PEI-PSBF). Various sorption experiments including isotherm, kinetics and pH effect on sorption capacity were conducted to evaluate the MC-LR adsorption performance of sorbents. As a result, the DC-PEI-PSBF could be suggested as a highly efficient sorbent able to be directly applied for MC-LR removal from aquatic natures. Keywords: Microcystin-LR, Biosorption, Biomass, Polyethylenimine (PEI), Adsorbent

1. Introduction In recent years, frequency and degree of harmful algal blooms (HABs) occurring have increased in precious water resources such as lakes and rivers due to eutrophication of water bodies (Pavagadhi et al., 2013). During widespread occurrence of HABs, various micro-pollutants including cyanotoxins are also discharged indispensably into water resources (Wei et al., 2017). Among cyanotoxins, microcystins (MCs) have been recognized as one of the typical micro-pollutants that have acute toxicity by inhibition of protein phosphatase-1 and -2A, leading to liver damage and tumor growth (Wang et al., 2015). MCs (monocyclic heptapeptides) consisting of seven amino acids, are usually produced from cyanobacteria including Microcystis, Oscillatoria, Nostoc and Anabaena species (Dawson, 1998; Hilborn et al., 2005). MCs are biosynthesized from intra-cellular of cyanobacteria species, and released into open waters during biological cell lysis and cell destruction (Pietsch et al., 2002). In the group of identified molecular MCs variants (above 150 types), microcystin-LR (MC-LR) have been recognized as one of the most commonly detected and most toxic type of MCs (Falconer, 2005; Wei et al., 2017). Due to the threat of MCs, WHO (World Health Organization) recommends 1.0 µg/L of MC-LR as a provisional safety guideline (Li et al., 2017) for the drinking water. Because of its cyclic structure, MC-LR is exceedingly stable and difficult to eliminate from aqueous phases by traditional water treatment technologies (Gao et al., 2016; Lawton & Robertson, 1999; Teng et al., 2013b). Recently, adsorption technology has attracted increasing interest as a strategy for MC-LR removal because of its suppleness, high removal efficiency and cost benefits (Kim et al., 2016c). Various sorbents including activated carbon (Donati et al., 1994), graphene oxide (Pavagadhi et al., 2013), mesoporous silica (Teng et al., 2013b), and mesoporous carbon (Teng et al., 2013a) were investigated for adsorptive MC-LR removal from aqueous phases. According to Pavagadhi et al. (Pavagadhi et al., 2013), the maximum MC-LR adsorption capacity (qm) of the graphene oxide was estimated as 1700 µg/g at pH 5. The wood-based activated carbons showed 20 ~ 280 µg/mg of maximum MC-LR adsorption capacity in the Milli-Q water (pH 5.2 ~ 6.6) (Donati et al., 1994). In case of the

mesoporous silica (SBA-15), the maximum MC-LR sorption capacities were estimated as 4.80 and 5.99 mg/g at 10 and 25 oC in Milli-Q water, respectively (Teng et al., 2013b). Although activated carbon based sorbents revealed potential as an effective adsorbent for MC-LR, it has a limitation due to relatively high operation costs and low energy-efficiency for regeneration of activated carbon after use (Kim et al., 2016c; Sathishkumar et al., 2010) Therefore, attention has shifted to development of biological methods such as biosorption (Vijayaraghavan & Yun, 2008). Various biomaterials including bacteria, fungi, algae, and agricultural/industrial bio-wastes have been applied as biosorbents for the removal of ionic pollutants, heavy metals and precious metals, since biomaterials have various functional groups which can be negatively or positively charged including carboxyl, amine, hydroxyl and sulfonate groups as electrostatic attractive binding sites for anionic and cationic materials (Kim et al., 2016c, Volesky, 2007). As a biosorbent, the peat has been applied in adsorption of MC-LR and displayed the maximum MC-LR sorption capacity as 255.71 µg/g at pH 3 (Sathishkumar et al., 2010). A large amount of biomass wastes including Escherichia coli is generated from the industrial scale fermentation process for amino acid production. It has been reported that E. coli biomass possesses amine, carboxyl, and phosphonate groups on their surface as main binding sites. Especially, since the possessed amine groups in E. coli biomass can be positively charged, it is recognized as main binding sites which can remove anionic pollutants by electrostatic interaction (Kim et al., 2015). In addition, it has been reported that MC-LR can be negatively charged above pH 2.19 by deprotonation of carboxyl groups in MC-LR molecule (Sathishkumar et al., 2010). Therefore, E. coli biomass can be potentially recyclable biosorbent for removal of MC-LR from aqueous phases. However, most of this waste biomass is treated by sea dumping, landfill, and incineration. In this study, the main purpose is to fabricate industrial waste biomass as a highly stable and highly efficient adsorbent for MC-LR removal from aqueous solution. To fabricate as a stable adsorbent, E. coli biomass was immobilized with polysulfone (PS) matrix to render it with high mechanical and chemical stability. To enhance adsorption performance of waste biomass-based sorbent, PEI-coating providing of numerous main binding sites (amine groups) for MC-LR and additional chemical modification were carried out. In addition, the MC-LR sorption performances of prepared sorbents were thoroughly compared.

2. Materials and methods

2.1. Materials The waste biomass, E. coli was obtained as spray dried from industrial amino acid, L-phenylalanine fermentation process, Daesang Co. (Gunsan, Korea). Polysulfone (PS, [OC6H4OC6H4SO2C6H4]n, average Mn ~ 22,000 by MO, beads) and branched polyethylenimine (PEI, 50 % in H2O, M.W. ~ 750,000) were purchased from Sigma Aldrich Co., Ltd. N,N-dimethylformamide (DMF, C3H7NO, 99.8 %) and glutaraldehyde (GA, C5H8O2, 25 wt %) were supplied from Daejung Chemical & Metals Co., Ltd. (Siheung, Korea) and Junsei Chemical Co., Ltd. (Tokyo, Japan), respectively. The target cyanotoxin, microcystin-LR (MC-LR, isolated from Microcystis aeruginosa, >95 %) was purchased from Enzo Life Sciences, Inc. (New York, USA). Trifluoroacetic acid (CF3COOH, TFA, >99.5 %) was purchased from Alfa Aesar (Thermo Fisher Scientific Inc.). Methyl alcohol (CH3OH, 99.9 %, HPLC grade) and acetonitrile (CH3CN, 99.9 %, HPLC grade) were obtained from TEDIA. All other chemical reagents were analytical grade.

2.2. Preparation of waste biomass based sorbents To fabricate pristine PS fiber (PSF), the 9 %w/v of PS solution was prepared by dissolving PS in DMF and stirring overnight at room temperature. In addition, to fabricate the polysulfone-E. coli biomass composite fibers (PSBF), 22.2 g of the E. coli biomass was suspended into the 200 g of 9 %w/v PS solution for preparation of 10 %w/w biomass containing PS-biomass mixture. Subsequently, it was stirred for 10 hours at room temperature for uniform dispersion of biomass in the PS-biomass composite solution. Then, these solutions were extruded into deionized water through a plastic hub needle (TAEHA Co., Namyangju, Korea) to form PSF and PSBF. Prepared fibers were

washed several times with distilled water to remove residual DMF. For the PEI-coating on PSBF, PEI was immobilized on the surface of PSBF through the agitating 100 gwet of prepared PSBF into 1 L of 23 g/L PEI solution for six hours. After that, the PEIimmobilized PSBF was washed one time using distilled water to remove residual PEI. Then, crosslinking was allowed using GA. For the cross-linking, PEI-immobilized PSBF was agitated into the 1 L of GA solution (0.7 mL/L) at pH 10.32 for four hours at room temperature. After cross-linking step, PEI-coated PSBF (PEI-PSBF) were separated from the solution and washed three times with deionized water. The prepared fibers were freeze dried for 24 hours. To fabricate the decarboxylated PEI-PSBF (DC-PEI-PSBF), the carboxyl groups on the PEI-PSF was modified by esterification reaction. Two grams of dried PEI-PSBF was suspended in 200 mL of methanol, and 8.6 mL of concentrated hydrochloric acid (HCl, 35 %) was added to the suspension. The mixture was stirred for 6 hours at 25oC. After completion of the esterification, DC-PEI-PSBF was separated by filtration, and then washed three times using deionized water. The esterification reaction processed for DC-PEI-PSBF can be represented as follow. 

H Biomass - COOH  CH3OH  Biomass - COOCH3  H 2 O

(1)

2.3. Determination of functional groups characteristics on sorbents Characteristics regarding the functional groups of E. coli biomass, PSF, PSBF, PEI-PSBF, and DCPEI-PSBF were determined using Fourier transform infrared spectrometer (FT-IR, Agilent Cary 630 FTIR, Agilent Technology, USA). These samples were prepared as KBr pallet. Their FT-IR spectra were recorded within the range of 700 ~ 4,000 cm-1.

2.4. X-ray photoelectron spectroscopy (XPS) analysis

The surface of the PSF, PSBF, PEI-PSBF, and DC-PEI-PSBF were analyzed by XPS (Micro-XPS) to determine the change of elements signal (C1s, O1s, and N1s) related the functional groups in the sorbents. The XPS instrument was calibrated at the Au 4f7/2 peak (binding energy: 84.0 eV). The Xray source was operated at 180 W. During XPS measurement, the pressure in the analysis chamber was maintained at less than 7x10-9 torr. To determine main element peaks of the sorbents, all binding energies were referenced by the neutral C1s peak at 284.6 eV.

2.5. Adsorption experiments To evaluate MC-LR sorption performances of prepared sorbents, adsorption experiments were conducted in a batch system. For adsorption experiments, 100 mg/L of MC-LR stock solution was prepared by dissolving 1 mg of MC-LR standard in 1 mL of methanol. To prepare the MC-LR solution of sorption experiments, the MC-LR stock solution was diluted with deionized water to be desired MC-LR concentrations. To determine the pH effect for MC-LR adsorption capacity of sorbents, MC-LR adsorption was conducted in different pH (pH edge experiment). Particularly, 0.01 g of E. coli, PSBF, PEI-PSBF, and DC-PEI-PSBF were suspended in 20 mL of MC-LR solution (initial concentration: 1,000 µg/L) and stirred at 190 rpm and 25oC for 24 hours in a shaking incubator. In the case of isotherm experiment, 0.01 g of PEI-PSBF and DC-PEI-PSBF were soaked into 30 mL of MCLR solutions with different initial MC-LR concentration (100, 200, 300, 400, 500, 700, and 1,000 µg/L), and stirred for 24 hours at the same condition with pH edge experiment. During isotherm experiments, the pH of samples was controlled to pH 5 and pH 7 by using HCl and NaOH solutions. To evaluate the sorption equilibrium time for MC-LR, kinetic tests were conducted at pH 7 and 25oC until equilibrium states were achieved. 0.06 g of PEI-PSBF and DC-PEI-PSBF were suspended in 200 mL of MC-LR solution (initial concentration: 566.07 µg/L), respectively. Kinetic samples were collected at the predetermined time. The amount of MC-LR adsorbed on the sorbents was estimated using the following equation:

(2) where q (µg/g) is the MC-LR uptake, and Vi and Vf (L) are the initial and final (after the addition of acid or base solution to adjust the pH) working volume. Ci and Cf (µg/L) are the initial and final concentration of MC-LR, respectively. M (g) is the weight of sorbents.

2.6. Detection of MC-LR MC-LR were measured by high-performance liquid chromatography with a photodiode array detection system (HPLC-PDA, Waters 515, Waters, USA). Forty-two percent of acetonitrile solution containing 0.1 percent of TFA was prepared as a mobile phase. To prevent clogging of HPLC line and column by tiny dust, prepared mobile phase was filtrated using membrane filter prior to HPLC analysis. The flow rate of mobile phase was set to be 1.5 mL/min. HPLC samples for MC-LR detection were prepared by filtration of aquatic samples (1 mL) using syringe filter unit (PTFE, 0.2 µm) to remove the remaining sorbent particles and/or dusts in the collected samples . Then, 180 µL of filtrated samples were injected into HPLC column (YMC-Triart C18). The separated MCLR peak from HPLC column was detected at 238 nm by UV detector, and MC-LR quantitative information was calculated using software (Waters Empower 2).

3. Results and discussion

3.1. Properties of functional groups in sorbents The characteristics of functional groups on sorbents are important factors affecting to sorption performance against target materials. Therefore, to determine the properties of functional groups on sorbents (E. coli biomass, PSF, PSBF, PEI-PSBF and DC-PEI-PSBF), FT-IR and XPS analyses were

carried out. In the FT-IR peaks of E. coli biomass, various peaks related to the various functional groups including amine, hydroxyl, carboxyl, and phosphonate groups located on the E. coli biomass were recorded. The observed peaks in the range of 3,600 ~ 3,200 cm-1 contributed to the overlapping of OH bonding of hydroxyl groups and –NH asymmetric stretching of amine groups (Fang et al., 2009; Liu et al., 2013). Strong peaks at around 2,923 cm-1 were related to the C–H band for symmetric stretching of methylene groups (–CH2) and deformation vibration of methyl groups (–CH3) (Mona et al., 2011). The 1,538 and 1,648 cm-1 peaks were derived from C–N stretching and secondary amide C=O bond or C=O chelate stretching of the carboxyl groups (Choi & Yun, 2006; Leone et al., 2007; Liu et al., 2009; Ngah & Fatinathan, 2010). Peaks at approximately 1,232 cm-1 indicate C-O stretching vibration of ketones, aldehydes and lactones or carboxyl groups. In addition, peaks at 1,080 cm-1 were related to P–OH stretching of phosphonate groups in the E. coli biomass (Kim et al., 2015; Vijayaraghavan et al., 2008). In PSF, various spectrum peaks corrsponding to C-O (1,108 cm-1), C-OC and C-SO2-C (1,150 cm-1), S=O (1,245 and 1,585 cm-1), aromatic groups (1,489 cm-1), and aliphatic C-H bond (2,965 cm-1) in molecular structure of PS were observed (Ficai et al., 2010; Singh et al., 2012). In the case of PSBF, PEI-PSBF, and DC-PEI-PSBF, E. coli biomass was immobilized in the PS matrix of sorbents thereby resulting in FT-IR peak characteristics similar to broad peaks of E. coli; broad peak (in range of 3600 ~ 3200, 1648, 1538, and 1400 cm-1) and peak matching to PS (1585, 1489, 1245, 1150, 1108, 1081, and 1014 cm-1) were simultaneously observed. Interestingly, after PEIcoating on the PSBF, the changes of peaks were observed accordingly. The peak at 3,433 cm-1 associated with amine groups was enhanced compared to that of PSBF. Since the PEI coated on the PEI-PSBF has numerous amount of amine groups in their structure, it might indicate that the amount of amine groups was introduced on the PEI-PSBF through PEI-coating on the PSBF. In addition, during cross-linking using GA, imine groups (=N-) can be formed. The FT-IR peak for imine groups have been observed at around 1,650 cm-1 (Ghoul et al., 2003). Therefore, as the imine groups were formed on the PEI-PSBF, the peak at 1,648 cm-1 of PSBF was shifted to 1,655 cm-1 and changed sharply after PEI-coating due to overlapping the peak of imine groups. The similar FT-IR peak changes after PEI-coating were reported by Deng and Ting (Deng & Ting, 2005a). After esterification

of carboxyl groups on the PEI-PSBF, FT-IR peak changes became obvious at approximately 2,965 cm-1 and 1,740 cm-1. A newly appeared shoulder peak at 1,740 cm-1 is an indicator for carbonyl stretching of un-ionized carboxylates (Kapoor & Viraraghavan, 1997) attributed by esterified carboxyl groups (-COOCH3) of DC-PEI-PSBF. According to Kim et al. (Kim et al., 2016c), the appearance of shoulder peak was observed at similar FT-IR peak position after esterification of carboxyl groups on E. coli biomass. In addition, the enhancement of peaks at 2,965 cm-1 assigned to aliphatic C-H bond of CH3 (Won et al., 2013) might be attributed to esterified carboxyl groups on the DC-PEI-PSBF. To further determine characteristics of functional groups on the sorbents, XPS analyses were carried out using PSF, PSBF, PEI-PSBF and DC-PEI-PSBF. To determine the introduction of E. coli composition in PS matrix and coated PEI on the PSBF, the N1s spectrum of PSF, PSBF, PEI-PSBF, and DC-PEI-PSBF was compared. In addition, to confirm esterification of carboxyl groups, C1s and O1s XPS spectrum of PEI-PSBF and DC-PEI-PSBF were compared, respectively. The obtained XPS signals of elements were decomposed by using XPS peak processing software (XPSPEAK 4.1) for detail determination of chemical bonding information. The parameters of peaks estimated from XPS peak processing for N1s, O1s, and C1s XPS signals of sorbents were summarized in table 1 (N1s peaks) and table 2 (C1s and O1s peaks). In XPS results, detail information of changed characteristics of functional groups along with biomass immobilization, PEI-coating and esterification of carboxyl groups were observed and demonstrated with FT-IR analysis. As results of N1s XPS analyses, PSF, constructed only PS matrix, did not reveal any N1s signal, since PS has not any nitrogen atom-based chemical bonds such as amine groups in the molecular structure. However, the N1s XPS signal of PSBF exhibited a weak and broad main N1s peak at approximately 399.1 eV. The N1s signal of PSBF might be derived from the amine groups of E. coli biomass which composited in PS matrix. According to table 1, the N1s signal of PSBF could be decomposed to two peaks at 399.1 (peak area: 2273.28) and 401.3 eV (peak area: 110.78) contributed by N-H bonding of amine groups (-NH2) and protonated amine groups (-NH3+), respectively (Deng & Ting, 2005a; Yu et al., 2012). The main N1s signal of PEI-PSBF was observed in range from 396 eV to

404 eV. From the N1s signal of PEI-PSBF, along with observed peaks (peaks at 399.1 and 401.3 eV) in N1s signal for PSBF, the new peaks at 400.0 and 398.5 eV were appeared. In addition, according to table 1, the total area of N1s signal for PEI-PSBF (total signal area: 9493.17) estimated as 3.98 times higher than that of PSBF (total signal area: 2384.06) due to appearance of new peaks at 400.0 and 398.5 eV and enhancement of peaks at 399.1 and 401.3 eV. These enhanced peak intensities and peak generation might be originated from amine groups of coated PEI on the PEI-PSBF. The new peak at 400.0 eV (peak area: 2099.29) assigned to the N-H band (Kim et al., 2016a) could be attributed from amine groups of coated PEI on the PEI-coated sorbents. In addition, the peak at 398.5 eV (peak area: 1053.32) represent the imine group (=N-), which was generated by GA cross-linking reaction between amine groups of E. coli biomass and PEI (Deng & Ting, 2005b). In addition, the areas of peaks at 399.1 (peak area: 4633.74) and 401.3 (peak area: 1706.82) of PEI-PSBF were enhanced as 2.03 and 15.4 times compared those of PSBF. In the case of DC-PEI-PSBF, the main N1s signal was recorded at similar binding energy range with PEI-PSBF. Also, the total area of N1s signal for DC-PEI-PSBF (9996.73) was estimated as similar with that of PEI-PSBF. It indicates that PEI molecules were maintained on the DC-PEI-PSBF after the chemical modification. From the N1s signal of DC-PEIPSBF, not only peaks at 398.5 (peak area: 528.62), 399.1 (peak area: 3841.66), 400.0 (peak area: 2362.86), and 401.3 eV (peak area: 2528.34) which observed in N1s signal of PEI-PSBF, but also new peak at 402.1 eV (peak area: 735.25) assigned to protonated imine groups (-N+=) (Kovtyukhova & Mallouk, 2005) were observed. According to the information of decomposed N1s peaks for DC-PEIPSBF, with appearance of new peak at 402.1 eV, the portion of 401.3 eV peak assigned to protonated amine groups (-NH3+) was increased from 17.9 % to 25.3 % after esterification of carboxyl group on the PEI-PSBF. These results might indicate that the amine and imine groups on PEI-PSBF were positively charged by protonation during the esterification reaction under high acidic condition (1M HCl in methanol). As some parts of the amine and imine groups of PEI-PSBF were ionized during chemical modification, the portion of unionized amine (399.1 eV) and imine (398.5 eV) groups in DC-PEI-PSBF were decreased from 48.8 % and 11.1 % to 38.4 % and 5.3 %, respectively. The esterification of carboxyl groups on the PEI-PSBF surface could be determined by the

comparison of C1s and O1s XPS spectra for PEI-PSBF and DC-PEI-PSBF. According to table 2, the C1s XPS signals of PEI-PSBF and DC-PEI-PSBF showed the decomposed peaks at 284.6, 285.4, 286.0, and 287.3 eV which indicate C=C or C-C (Dong et al., 2010; Tien et al., 2012), C-N or C-C (Ahmed et al., 2013; Shan et al., 2013), C=N or C-O (Dong et al., 2010; Zhou et al., 2014), and C-O (Ahmed et al., 2013) bonding, respectively. It was noted that the area ratio of C-O (286.0 and 287.3 eV), corresponding C-O-C bond of esterified carboxyl group on the DC-PEI-PSBF, was increased from 18.5 and 15.1 % to 22.7 and 23.2 %, respectively. Also, alteration of peak related C-O bonding could be identified in O1s XPS spectra of PEI-PSBF and DC-PEI-PSBF. The O1s XPS signals of PEIPSBF and DC-PEI-PSBF were observed in range of binding energy from 526 eV to 538 eV with three decomposed peaks at 531.0, 532.1 and 533 eV. These 531.0 and 532.1 eV peaks are attributed to C=O/O-C=O of deprotonated carboxyl groups (Cossaro et al., 2012; Zhou et al., 2014) and anhydride (O=C-O-C=O) or hydroxyl groups (C-O-H) (Kang et al., 2016), respectively. The peak at 533 eV assigned to the C-O bond of ester (O-C=O) or ether (C-O-C) groups (Kang et al., 2016; Mosch et al., 2015; Zhou et al., 2014). According to table 2, after the esterification of PEI-PSBF, area ratio peak at 533 eV was enhanced from 22.6 to 26.9 % in O1s XPS signal. It indicates that the number of esterified carboxyl groups (-COOCH3) are increased on DC-PEI-PSBF by chemical modification. These difference of C1s and O1s spectrum for PEI-PSBF and DC-PEI-PSBF could introduce the change of FT-IR characteristics for PEI-PSBF and DC-PEI-PSBF like FT-IR peaks enhancement of C-H bonding (2,965 cm-1) and unionized carbonyl groups (1,740 cm-1) after chemical modification of PEIPSBF.

3.2. pH effect on MC-LR sorption efficiency of sorbents From these FT-IR and XPS results, functional groups characteristics of sorbents can be summarized as all sorbents have carboxyl, phosphonate, and amine groups, derived from E. coli biomass in the sorbents. In addition, it was recognized that the PEI was well-coated on the PEI-PSBF and DC-PEIPSBF, and the carboxyl groups of DC-PEI-PSBF was blocked by esterification. Since located

functional groups such as carboxyl and amine groups can be negatively or positively charged by deprotonation depending on their pKa property, solution pH has been recognized as a critical factor that can determine sorption capacity of sorbent for ionic materials (Volesky, 2007). Therefore, to understand the influence of pH on the sorption capacities of sorbents for MC-LR, MC-LR sorption tests were evaluated using raw E. coli biomass, PSBF, PEI-PSBF, and DC-PEI-PSBF under different pH condition of range of pH 3 ~ pH 7. The effect of solution pH on MC-LR adsorption is presented in Fig. 1. According to results of pH effect, the MC-LR sorption capacities of sorbents were significantly affected by pH. The significant pH dependence of MC-LR sorption capacities can be explained by the electrostatic interactive sorption mechanism between binding site of sorbents and target material, MC-LR. According to Fig. 1, the MC-LR uptake of E. coli biomass as a raw material recorded a maximum value (65.5 ± 0.84 µg/g) at approximately pH 3, and that was decreased as pH increased to pH 4.5. In addition, from pH 4.5, E. coli biomass could not adsorb MC-LR. It has been reported that MC-LR is negatively charged above pH 2.19, because the anionic species of MC-LR, [(COO-)2(NH2+)] and [(COO-)2(NH)] are activated (de Maagd et al., 1999; Sathishkumar et al., 2010). As the raw material, E. coli possesses the carboxyl, phosphonate group, and amine groups which have pKa values of 3.5 ~ 5.0, 6.1 ~ 6.8, and 8 ~ 10, respectively (Kim et al., 2016b). Therefore, amine groups of E. coli biomass have positive charge under acidic condition (pH 3), and could bind negatively charged MC-LR by electrostatic attraction. However, as pH increased, carboxyl and phosphonate groups can be negatively charged by deprotonation at pH higher than their pKa values. Therefore, binding of anionic MC-LR to the amine group might be interfered by electrostatic repulsion between negatively charged functional group like carboxyl groups and MC-LR. The PSBF revealed similar pH dependence with E. coli biomass. The MC-LR sorption capacity of PSBF was estimated as maximum value (40.24 ± 4.30 µg/g) at pH 4, and that decreased along with increasing pH. It can be explained as the same manner with the pH dependent MC-LR uptake of the E. coli biomass, because the functional groups of PSBF were derivenen from composited E. coli biomass in PSBF.

After PEI-coating on the PSBF, the MC-LR uptake was significantly increased in all ranges of solution pH, due to the introduction of numerous amounts of amine groups from PEI. According to the pKa property of amine group, since the large amount of amine groups on the PEI-PSBF can be fully positively charged under in acidic pH, MC-LR uptake at pH 3 should be recorded as the maximum uptake in experimental pH range. However, unexpectedly, the PEI-PSBF revealed maximum MC-LR uptake (1,781.12 ± 17.62 µg/g) at approximately pH 5. The reason of this observation is not clearly explained, but it may be suggested that the electrostatic force might be involved. According to the chemical structure of MC-LR, two sites of carboxyl groups are possessed in MC-LR molecule. These two sites of carboxyl groups can be negatively charged at above pH 2.17 according to their pKa properties (pKa1: 2.17 and pKa2: 3.96) (Klein et al., 2013). Therefore, the electrostatic force between PEI-PSBF and MC-LR might be maximized at approximately pH 5, because negativity of MC-LR might be enhanced as increased pH by deprotonation of carboxyl groups in MC-LR molecule. In addition, the partial of negatively charged functional groups like carboxyl groups which have a pKa value in range of 3 ~ 5 of PSBF were blocked by electrostatic ion paring with positively charged amine groups of PEI during PEI-coating step (Won et al., 2013). Hence, the optimum pH of PEI-PSBF for MC-LR adsorption shifted to around pH 5 compared with that of PSBF. Above pH 5, the MC-LR adsorption capacity of PEI-PSBF decreased as pH increased. According to Willner et al. (Willner et al., 1993) and Kim et al. (Kim et al., 2016b), the pKa values of amine groups in PEI (primary, secondary, and tertiary) were reported as 4.5, 6.7, and 11.6 respectively. Therefore, at higher pH than pH 5, the MC-LR uptake of PEI-PSBF decreased, because the amine group of coated PEI on the PEI-PSBF lost positivity as pH increased. Another reason for decreasing MC-LR uptake may be attributed to increasing electrostatic repulsion between negatively charged inhibition sites, residual carboxyl and phosphonate groups, of the PEI-PSBF and negatively charged MC-LR from pH 5 of solution pH. After esterification of PEI-PSBF, the MC-LR uptake was enhanced in a whole range of pH. The DC-PEI-PSBF revealed optimum uptake value (1,823.48 ± 1.76 µg/g) at approximately pH 5.3. Although the uptake of MC-LR for DC-PEI-PSBF was decreased above pH 6.5, that remained

significantly higher MC-LR uptake than that of PEI-PSBF up to neutral pH (around pH 7) due to the removal of interfering site, carboxyl groups, by esterification, as expected. Based on the pH effect of MC-LR adsorption of DC-PEI-PSBF, we found a potential that DC-PEI-PSBF can be applied as an efficient waste biomass-based adsorbent for removal MC-LR from nature which has pH range from 6 to 7.5.

3.3. Comparison of affinity toward MC-LR In the natural aquatic system, although harmful algae occur in water bodies, cyanotoxin MC-LR is usually exists as a micro-pollutant in extremely low concentration. For efficient adsorptive removal of micro-pollutant, sorbents should have high affinity toward pollutants. Therefore, to determine affinity of sorbent toward MC-LR, isotherm experiments were conducted using PEI-PSBF and DC-PEI-PSBF at pH 5 and pH 7. The Fig. 2 presents linear isotherm results of PEI-PSBF and DC-PEI-PSBF at pH 5 and pH 7. Although initial concentration of MC-LR in isotherm experimental use, were too low to reveal saturated uptakes for PEI-PSBF and DC-PEI-PSBF, MC-LR sorption capacities of PEI-PSBF and DC-PEI-PSBF were increased as initial concentration of MC-LR was increased. In addition, increasing rates of MC-LR uptake, indicating affinity toward MC-LR were significantly different depending on solution pH and sorbent. According to Fig. 2, MC-LR uptakes of PEI-PSBF and DCPEI-PSBF recorded higher values at pH 5 compared with those of pH 7. In addition, after modification (DC-PEI-PSBF), uptakes of PEI-PSBF were enhanced at both pH. For detailed comparison of affinities toward MC-LR, isotherm results of PEI-PSBF and DC-PEI-PSBF were linearly regressed using the follow equation. (3) where q (µg/g) is the experimental uptake of sorbents for MC-LR at the specific pH, and Ce is the equilibrium concentration of MC-LR (µg/L). a and b are the slope and constant values for linear

equation. Estimated parameters from linear equation are summerized in Table 3. The slope of PEI-PSBF and DC-PEI-PSBF were calculated at pH 5 as 19.34 ± 0.49 and 59.22 ± 3.57 L/g, respectively (Table 3). It was 17.9 and 9.3 times higher than that of PEI-PSBF (1.08 ± 0.09 L/g) and DC-PEI-PSBF (6.38 ± 0.33 L/g) at pH 7, respectively. It clearly indicates that affinity towards MC-LR of sorbents could be significantly affected by solution pH due to pH dependence of functional groups on the sorbents. As discussed in pH effect on MC-LR sorption capacity, positively charged amine groups can bind with negatively charged MC-LR by electrostatic interaction. However, as increasing pH above pH 5, carboxyl groups on the sorbents can be negatively charged by deprotonation as the pKa property of carboxyl groups. In addition, the negatively charged functional groups can interfere with the MC-LR binding on amine groups by electrostatic repulsion. Therefore, the affinities of sorbents toward MC-LR were decreased at higher pH. Although the affinities of sorbents toward MC-LR were pH dependent, after blocking of sorption inhibition site (carboxyl groups) in PEI-PSBF, the affinities of modified sorbent (DC-PEI-PSBF) toward MC-LR were estimated as improved values at both pH. It can be explained by the fact that the carboxyl groups which interfering bind MC-LR to amine groups by electrostatic repulsion were blocked by chemical modification, esterification. The esterified carboxyl groups (-COOCH3) do not be converted to negatively charged groups (-COO-). Therefore, it might lead to the increasing affinity of sorbent toward MC-LR by decreasing of electrostatic repulsion between sorbent and MC-LR.

3.4. Determination of sorption equilibrium time To determine the time for MC-LR sorption equilibrium in nature, kinetic experiments were conducted at pH 7 using PEI-PSBF and DC-PEI-PSBF. Fig. 3 shows plots of MC-LR sorption by PEI-PSBF and DC-PEI-PSBF versus contact time at pH 7. Almost complete adsorption equilibrium was attained within 60 minutes for each sorbent. It was revealed that the equilibrium MC-LR uptake by DC-PEI-PSBF was higher than that of PEI-PSBF, due to removal of the sorption inhibition site,

carboxyl group by esterification. To estimate the sorption kinetic parameters, the experimental MCLR kinetic data of PEI-PSBF and DC-PEI-PSBF were fitted by kinetic models, pseudo-1st-order and pseudo-2nd-order kinetic models which have been widely used in adsorption kinetics (Ho & McKay, 1999). Pseudo-1st-order and pseudo-2nd-order kinetic models can be represented as follows: Pseudo-1st-order model:

(4)

Pseudo-2nd-order model:

(5)

where qe is the amount of MC-LR adsorbed at equilibrium (µg /g), qt the amount of MC-LR adsorbed at time t (µg /g), k1 and k2 are the pseudo-1st-order rate constant (1/min) and pseudo-2nd-order rate constant (g/µg min), respectively. Estimated kinetic parameters from kinetic models are summarized in Table 3. Although the predicted curves by kinetic models showed good agreement with the experimental MCLR kinetic data of PEI-PSBF and DC-PEI-PSBF (Fig 3), based on the calculated correlation coefficient (R2) from kinetic models, the pseudo-2nd-order model better described the experimental kinetic data of PEI-PSBF and DC-PEI-PSBF compared to the pseudo-1st-order model. This may indicate that the MC-LR adsorption on the sorbents follows the second-order reaction. According to table 3, estimated R2 values from pseudo-1st-order model and pseudo-2nd-order model for PEI-PSBF were 0.9135 (R2

1st, PEI)

and 0.9200 (R22nd,

PEI),

respectively. The qe values for PEI-PSBF were

estimated as 279.69 ± 8.39 µg /g (qe1, PEI) and 286.48 ± 9.22 µg /g (qe2, PEI) from pseudo-1st-order and pseudo-2nd-order model, respectively. In addition, k1 and k2 for PEI-PSBF toward MC-LR were calculated as 1.339 ± 0.37 1/min (k1 PEI) and 0.009 ± 0.004 g/ µg min (k2 PEI), respectively. In the case of DC-PEI-PSBF, R2, qe, and k1 values were estimated as 0.8549 (R2 1st, DC), 1,142.47 ± 54.31 µg /g (qe1, DC), and 0.147 ± 0.031 1/min (k1 DC), respectively, from pseudo-1st-order model. From pseudo-2ndorder model, the R2, qe, and k2 values for DC-PEI-PSBF toward MC-LR at pH 7 were recorded as 0.9260 (R2 2nd, DC), 1,229.61 ± 49.22 µg /g (qe2 DC), and 0.0002 ± 0.00004 g/µg min (k2 DC), respectively.

4. Conclusion In this study, waste biomass was successfully valorized to high stable and efficient sorbent (DCPEI-PSBF) for MC-LR removal through immobilization, PEI-coating and sorption-inhibition site blocking. After modification, characteristics change of functional groups could be identified by FT-IR and XPS analyses. Although the DC-PEI-PSBF revealed the highest uptake and affinity toward MCLR, these were significantly pH dependent. Developed sorbents exhibited the equilibrium state within 60 minutes for MC-LR adsorption at nature pH. Our results provide a suitable method for regeneration of waste biomasses to efficient sorbents for removal of micro-pollutants such as MC-LR from aquatic systems.

Acknowledgement This work was supported by the Government of South Korea through the National Research Foundation of Korea (NRF-2016R1A6A3A01012976 and NRF-2016R1D1A1B03932773) and Korea Basic Science Institute under the R&D program (Project No. C36703), supervised by the Ministry of Science, ICT and Future Planning. This research was also supported by the research program of KAERI, Republic of Korea and Korea University grants.

References 1.

Ahmed, M.H., Byrne, J.A., McLaughlin, J., Ahmed, W. 2013. Study of human serum albumin adsorption and conformational change on DLC and silicon doped DLC using XPS and FTIR spectroscopy. Journal of Biomaterials and Nanobiotechnology, 4(2), 194-203.

2.

Choi, S.B., Yun, Y.-S. 2006. Biosorption of cadmium by various types of dried sludge: an equilibrium study and investigation of mechanisms. Journal of hazardous materials, 138(2), 378-383.

3.

Cossaro, A., Cvetko, D., Floreano, L. 2012. Amino–carboxylic recognition on surfaces: from 2D to 2D+ 1 nano-architectures. Physical Chemistry Chemical Physics, 14(38), 13154-13162.

4.

Dawson, R. 1998. The toxicology of microcystins. Toxicon, 36(7), 953-962.

5.

de Maagd, P.G.-J., Hendriks, A.J., Seinen, W., Sijm, D.T. 1999. pH-dependent hydrophobicity of the cyanobacteria toxin microcystin-LR. Water Research, 33(3), 677-680.

6.

Deng, S., Ting, Y.-P. 2005a. Characterization of PEI-modified biomass and biosorption of Cu(II), Pb(II) and Ni(II). Water Research, 39(10), 2167-2177.

7.

Deng, S., Ting, Y.P. 2005b. Polyethylenimine-modified fungal biomass as a high-capacity biosorbent for Cr (VI) anions: sorption capacity and uptake mechanisms. Environmental Science & Technology, 39(21), 8490-8496.

8.

Donati, C., Drikas, M., Hayes, R., Newcombe, G. 1994. Microcystin-LR adsorption by powdered activated carbon. Water Research, 28(8), 1735-1742.

9.

Dong, X., Su, C.-Y., Zhang, W., Zhao, J., Ling, Q., Huang, W., Chen, P., Li, L.-J. 2010. Ultra-large single-layer graphene obtained from solution chemical reduction and its electrical properties. Physical Chemistry Chemical Physics, 12(9), 2164-2169.

10. Falconer, I.R. 2005. Is there a human health hazard from microcystins in the drinking water supply? CLEAN–Soil, Air, Water, 33(1), 64-71. 11. Fang, L., Cai, P., Chen, W., Liang, W., Hong, Z., Huang, Q. 2009. Impact of cell wall structure on the behavior of bacterial cells in the binding of copper and cadmium. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 347(1), 50-55. 12. Ficai, D., Ficai, A., Voicu, G., Vasile, B.S., Guran, C., Andronescu, E. 2010. Polysulfone based membranes with desired pores characteristics. Materiale Plastice, 47(1), 24-27. 13. Gao, Y., Wang, F., Wu, Y., Naidu, R., Chen, Z. 2016. Comparison of degradation mechanisms of microcystin-LR using nanoscale zero-valent iron (nZVI) and bimetallic Fe/Ni and Fe/Pd nanoparticles. Chemical Engineering Journal, 285, 459-466. 14. Ghoul, M., Bacquet, M., Morcellet, M. 2003. Uptake of heavy metals from synthetic aqueous solutions using modified PEI—silica gels. Water Research, 37(4), 729-734.

15. Hilborn, E.D., Carmichael, W.W., Yuan, M., Azevedo, S.M. 2005. A simple colorimetric method to detect biological evidence of human exposure to microcystins. Toxicon, 46(2), 218-221. 16. Ho, Y.S., McKay, G. 1999. Pseudo-second order model for sorption processes. Process Biochemistry, 34(5), 451-465. 17. Kang, Y., Huang, F., You, L., Li, X., Gao, B. 2016. Impact of fracturing fluid on multi-scale mass transport in coalbed methane reservoirs. International Journal of Coal Geology, 154, 123-135. 18. Kapoor, A., Viraraghavan, T. 1997. Heavy metal biosorption sites in Aspergillus niger. Bioresource Technology, 61(3), 221-227. 19. Kim, S., Choi, Y.-E., Yun, Y.-S. 2016a. Ruthenium recovery from acetic acid industrial effluent using chemically stable and high-performance polyethylenimine-coated polysulfoneEscherichia coli biomass composite fibers. Journal of Hazardous Materials, 313, 29-36. 20. Kim, S., Kumar Reddy, D.H., Choi, Y.-E., Yun, Y.-S. 2016b. Importance of the coating pH in fabrication of polyethylenimine-coated polysulfone-Escherichia coli composite fiber sorbent. Journal of the Taiwan Institute of Chemical Engineers, 66, 379-385. 21. Kim, S., Song, M.-H., Wei, W., Yun, Y.-S. 2015. Selective biosorption behavior of Escherichia coli biomass toward Pd(II) in Pt(IV)–Pd(II) binary solution. Journal of Hazardous Materials, 283(0), 657-662. 22. Kim, S., Won, S.W., Cho, C.-W., Yun, Y.-S. 2016c. Valorization of Escherichia coli waste biomass as a biosorbent for removing reactive dyes from aqueous solutions. Desalination and Water Treatment, 57(42), 20084-20090. 23. Klein, A.R., Baldwin, D.S., Silvester, E. 2013. Proton and iron binding by the cyanobacterial toxin microcystin-LR. Environmental science & technology, 47(10), 5178-5184. 24. Kovtyukhova, N.I., Mallouk, T.E. 2005. Ultrathin anisotropic films assembled from individual single-walled carbon nanotubes and amine polymers. The Journal of Physical Chemistry B, 109(7), 2540-2545.

25. Lawton, L., Robertson, P.J. 1999. Physico-chemical treatment methods for the removal of microcystins (cyanobacterial hepatotoxins) from potable waters. Chemical Society Reviews, 28(4), 217-224. 26. Leone, L., Ferri, D., Manfredi, C., Persson, P., Shchukarev, A., Sjöberg, S., Loring, J. 2007. Modeling the acid-base properties of bacterial surfaces: a combined spectroscopic and potentiometric study of the gram-positive bacterium Bacillus subtilis. Environmental science & technology, 41(18), 6465-6471. 27. Li, J., Li, R., Li, J. 2017. Current research scenario for microcystins biodegradation – A review on fundamental knowledge, application prospects and challenges. Science of The Total Environment, 595, 615-632. 28. Liu, Y.-g., Liao, T., He, Z.-b., Li, T.-t., Wang, H., Hu, X.-j., Guo, Y.-m., He, Y. 2013. Biosorption of copper(II) from aqueous solution by Bacillus subtilis cells immobilized into chitosan beads. Transactions of Nonferrous Metals Society of China, 23(6), 1804-1814. 29. Liu, Y., Cao, Q., Luo, F., Chen, J. 2009. Biosorption of Cd2+, Cu2+, Ni2+ and Zn2+ ions from aqueous solutions by pretreated biomass of brown algae. Journal of Hazardous Materials, 163(2–3), 931-938. 30. Mona, S., Kaushik, A., Kaushik, C. 2011. Biosorption of reactive dye by waste biomass of Nostoc linckia. Ecological Engineering, 37(10), 1589-1594. 31. Mosch, H.L., Höppener, S., Paulus, R.M., Schröter, B., Schubert, U.S., Ignaszak, A. 2015. The correlation of the binding mechanism of the polypyrrole–carbon capacitive interphase with electrochemical stability of the composite electrode. Physical Chemistry Chemical Physics, 17(20), 13323-13332. 32. Ngah, W.S.W., Fatinathan, S. 2010. Pb(II) biosorption using chitosan and chitosan derivatives beads: Equilibrium, ion exchange and mechanism studies. Journal of Environmental Sciences, 22(3), 338-346. 33. Pavagadhi, S., Tang, A.L.L., Sathishkumar, M., Loh, K.P., Balasubramanian, R. 2013. Removal of microcystin-LR and microcystin-RR by graphene oxide: Adsorption and kinetic

experiments. Water Research, 47(13), 4621-4629. 34. Pietsch, J., Bornmann, K., Schmidt, W. 2002. Relevance of intra-and extracellular cyanotoxins for drinking water treatment. Acta hydrochimica et hydrobiologica, 30(1), 7-15. 35. Sathishkumar, M., Pavagadhi, S., Vijayaraghavan, K., Balasubramanian, R., Ong, S. 2010. Experimental studies on removal of microcystin-LR by peat. Journal of Hazardous Materials, 184(1), 417-424. 36. Shan, C., Wang, L., Han, D., Li, F., Zhang, Q., Zhang, X., Niu, L. 2013. Polyethyleneiminefunctionalized graphene and its layer-by-layer assembly with Prussian blue. Thin Solid Films, 534, 572-576. 37. Singh, K., Ingole, P.G., Bajaj, H.C., Gupta, H. 2012. Preparation, characterization and application of β-cyclodextrin-glutaraldehyde crosslinked membrane for the enantiomeric separation of amino acids. Desalination, 298, 13-21. 38. Teng, W., Wu, Z., Fan, J., Chen, H., Feng, D., Lv, Y., Wang, J., Asiri, A.M., Zhao, D. 2013a. Ordered mesoporous carbons and their corresponding column for highly efficient removal of microcystin-LR. Energy & Environmental Science, 6(9), 2765-2776. 39. Teng, W., Wu, Z., Feng, D., Fan, J., Wang, J., Wei, H., Song, M., Zhao, D. 2013b. Rapid and efficient removal of microcystins by ordered mesoporous silica. Environmental science & technology, 47(15), 8633-8641. 40. Tien, H.-W., Huang, Y.-L., Yang, S.-Y., Hsiao, S.-T., Liao, W.-H., Li, H.-M., Wang, Y.-S., Wang, J.-Y., Ma, C.-C.M. 2012. Preparation of transparent, conductive films by graphene nanosheet deposition on hydrophilic or hydrophobic surfaces through control of the pH value. Journal of Materials Chemistry, 22(6), 2545-2552. 41. Vijayaraghavan, K., Lee, M.W., Yun, Y.-S. 2008. A new approach to study the decolorization of complex reactive dye bath effluent by biosorption technique. Bioresource technology, 99(13), 5778-5785. 42. Vijayaraghavan, K., Yun, Y.-S. 2008. Bacterial biosorbents and biosorption. Biotechnology Advances, 26(3), 266-291.

43. Volesky, B. 2007. Biosorption and me. Water Research, 41(18), 4017-4029. 44. Wang, S.-l., Wang, L.-l., Ma, W.-h., Johnson, D.M., Fang, Y.-f., Jia, M.-k., Huang, Y.-p. 2015. Moderate valence band of bismuth oxyhalides (BiOXs, X=Cl, Br, I) for the best photocatalytic degradation efficiency of MC-LR. Chemical Engineering Journal, 259, 410416. 45. Wei, G., Quan, X., Fan, X., Chen, S., Zhang, Y. 2017. Carbon-nanotube-based sandwich-like hollow fiber membranes for expanded microcystin-LR removal applications. Chemical Engineering Journal, 319, 212-218. 46. Willner, I., Eichen, Y., Frank, A.J., Fox, M.A. 1993. Photoinduced electron-transfer processes using organized redox-functionalized bipyridinium-polyethylenimine-titania colloids and particulate assemblies. The Journal of Physical Chemistry, 97(28), 7264-7271. 47. Won, S.W., Kim, S., Kotte, P., Lim, A., Yun, Y.-S. 2013. Cationic polymer-immobilized polysulfone-based fibers as high performance sorbents for Pt(IV) recovery from acidic solutions. Journal of Hazardous Materials, 263, Part 2(0), 391-397. 48. Yu, J., Guan, M., Li, F., Zhang, Z., Wang, C., Shu, C., Wei, H., Zhang, X.-E. 2012. Effects of fullerene derivatives on bioluminescence and application for protease detection. Chemical Communications, 48(89), 11011-11013. 49. Zhou, J.S., Li, J.M., Liu, K.H., Lan, L., Song, H.H., Chen, X.H. 2014. Free-standing cobalt hydroxide nanoplatelet array formed by growth of preferential-orientation on graphene nanosheets as anode material for lithium-ion batteries. Journal of Materials Chemistry A, 2(48), 20706-20713.

Figure Captions Figure 1. Effect of pH on the MC-LR sorption capacity of E. coli, PSBF, PEI-PSBF, and DC-PEIPSBF. Figure 2. The isotherm results of PEI-PSBF (coated by M.W. 750,000 of PEI) and DC-PEI-PSBF for MC-LR at pH 5 and pH 7. Figure 3. Adsorption equilibrium of MC-LR onto PEI-PSBF and DC-PEI-PSBF at pH 7.

Figure 1.

PSBF PEI-PSBF DC-PEI-PSBF E. coli biomass

MC-LR uptake (g/g)

2000

1500

1000

500

0 3

4

5

pH (-)

6

7

8

Figure 2.

MC-LR uptake (g/g)

4000 pH 5, PEI-PSBF pH 5, DC-PEI-PSBF pH 7, PEI-PSBF pH 7, DC-PEI-PSBF Linear fit

3000

2000

1000

0 0

200

400

600

800

Equilibrium concentration of MC-LR (g/L)

1000

Figure 3.

1800 PEI-PSBF (pH 7) DC-PEI-PSBF (pH 7) Pseudo-second order fit Pseudo-first order fit

MC-LR uptake (mg/g)

1600 1400 1200 1000 800 600 400 200 0 0

100

200

300

Time (min)

400

500

Table 1. N1s XPS spectral bands based on the binding energies for PSF, PSBF, PEI-PSBF, and DCPEI-PSBF Peak area Peak position (eV)

Assignment

398.5

399.1

>N=

-NH2

400.0

N-H

PSF

PSBF

PEI-PSBF

DC-PEI-PSBF

-

-

1053.32

528.62

(1.15)

(1.1)

2273.28

4633.74

3841.66

(1.52)

(1.4)

(1.4)

2099.29

2362.86

(1.4)

(1.35)

110.78

1706.82

2528.34

(1.5)

(1.5)

(1.45)

-

-

735.25

-

-

401.3

402.1

-NH3+

>N+=

-

-

(2.99) Total area

-

2384.06

9493.17

9996.73

*Values in the bracket indicate the value for full width at half maximum (FWHM) of decomposed peaks at specific position

Table 2. C1s and O1s XPS spectral bands based on the binding energies for PEI-PSBF and DC-PEIPSBF Peak area (%) Signal

C1s

O1s

Position

Assignment PEI-PSBF

DC-PEI-PSBF

284.6

C-C/C=C

36.1 (1.30)

32.0 (1.36)

285.3

C-N/C-C

30.3 (1.25)

22.1 (1.25)

286.0

C=N/C-O

18.5 (1.16)

22.7 (1.16)

287.3

C-O (C-O-C)

15.1 (2.01)

23.2 (2.10)

531.0

C=O/O-C=O

55.3 (1.65)

52.0 (1.50)

532.1

O=C-O-C=O/C-O-H

22.1 (1.40)

21.1 (1.30)

533.0

C-O (C-O-C)

22.6 (2.10)

26.9 (2.30)

*Values in the bracket indicate the value for full width at half maximum (FWHM) of decomposed peaks at specific position

Table 3. The parameters calculated from linear equation, pseudo-1st-order, and pseudo-2nd-order model Equation

pH

PEI-PSBF a

DC-PEI-PSBF

b

a

b

R2

Linear

(L/g)

(µg/g)

19.34

-82.97

5

(L/g)

(µg/g)

59.22

-167.87

(3.57)

(104.14)

6.38

-54.07

(0.33)

(60.65)

qe1

k1

(µg/g)

(1/min)

1,142.47

0.147

0.9961 (0.49)

(38.66)

1.08

-79.86

7

Pseudo-1st-order

R2

0.9787

0.9629 (0.09)

(36.92)

qe1

k1

(µg/g)

(1/min)

279.69

1.339

R2

0.9842

7 0.9135

Pseudo-2nd-order

R2

(8.39)

(0.37)

qe2

k2

(µg/g)

(g/µg min)

286.48

0.009

R2

0.8549 (54.31)

(0.031)

qe2

k2

(µg/g)

(g/µg min)

1,229.61

0.0002

(49.22)

(0.00004)

R2

7 0.9200 (9.22)

(0.004)

0.9260

*The value in the bracket is the standard error of parameters calculated from equations.

Highlights  Waste E. coli was valorized as a stable and high-performance sorbent for MC-LR.  MC-LR adsorption equilibrium was reached within 60 min at natural pH.  The MC-LR uptakes and affinity were dependent upon the solution pH.  The sorption performance was enhanced by PEI-coating and decarboxylation.