Bioresource Technology 171 (2014) 442–451
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Geochemical and spectroscopic investigations of Cd and Pb sorption mechanisms on contrasting biochars: Engineering implications Lukáš Trakal a,⇑, Deniz Bingöl b, Michael Pohorˇely´ c, Miroslav Hruška a, Michael Komárek a a
Department of Environmental Geosciences, Faculty of Environmental Sciences, Czech University of Life Sciences Prague, Kamycka 129, Praha 6 Suchdol 16521, Czech Republic Department of Chemistry, Faculty of Sciences and Arts, Kocaeli University, Umuttepe Campus, 41380 Kocaeli, Turkey c Environmental Process Engineering Laboratory, Institute of Chemical Process Fundamentals, Academy of Sciences of Czech Republic, v.v.i., Rozvojova 135, Praha 6 Suchdol 16502, Czech Republic b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Metal removal efficiencies and
mechanisms depended on the parent material. Ion exchange is the predominant mechanism during metal sorption on biochars. Metals were strongly bound to the biochars. Metal removal efficiency is strongly pH-dependent.
a r t i c l e
i n f o
Article history: Received 31 July 2014 Received in revised form 22 August 2014 Accepted 24 August 2014 Available online 30 August 2014 Keywords: Biochar Metals Response Surface Methodology (RSM) Sorption mechanisms Desorption
a b s t r a c t Biochars prepared from nut shells, plum stones, wheat straws, grape stalks and grape husks were tested as potential sorbents for Cd and Pb. Mechanisms responsible for metal retention were investigated and optimal sorption conditions were evaluated using the RSM approach. Results indicated that all tested biochars can effectively remove Cd and Pb from aqueous solution (efficiency varied between 43.8% and 100%). The removal rate of both metals is the least affected by the biochar morphology and specific surface but this removal efficiency is strongly pH-dependent. Results of variable metal removal combined with different optimized conditions explain the different metal sorption mechanisms, where the predominant mechanism is ion exchange. In addition, this mechanism showed very strong binding of sorbed metals as confirmed by the post-desorption of the fully metal-loaded biochars. Finally, these biochars could thus also be applicable for metal contaminated soils to reduce mobility and bioavailability of Cd and Pb. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Among other water treatment technologies, such as precipitation, coagulation or membrane filtration, biosorption should be taken into account as a potential ‘‘low-cost candidate’’ for removing toxic metals from wastewater. Biochar is an efficient
⇑ Corresponding author. Tel.: +420 224 383 864. E-mail address: [email protected] (L. Trakal). http://dx.doi.org/10.1016/j.biortech.2014.08.108 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
biosorbent used for the removal of metals such as Cd and/or Pb (two common metal pollutants in Central Europe with contrasting geochemical behaviour) from waste waters (Inyang et al., 2011; Lu et al., 2012; Kim et al., 2013; Mohan et al., 2014a). This pyrogenic carbon-rich material is most commonly prepared from waste materials, such as woody biomass, animal waste, sludge or a variety of agricultural residues (Lehmann and Joseph, 2009). Many studies have described an effective removal of Cd and Pb using biochar originating from different agricultural wastes. For example, Kołodyn´ska et al. (2012), Xu et al. (2013a,b) reported
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
that biochars made from dairy and pig manure were able to sorb Cd and/or Pb in the range from 0.28 to 1.11 mmol g1. Inyang et al. (2011), Jiang et al. (2012), Han et al. (2013), Kim et al. (2013) and Xu et al. (2013a) described the high potential of sugarcane bagasse, rice straw, pinewood, rice husk and Miscanthus sacchariflorus to remove Cd and/or Pb from aqueous solution in the range of 0.004–0.66 mmol g1. These results thus present variable metal sorption efficiency of the tested biochars. This variability is caused by many factors, such as: (i) source and composition of the original waste material; (ii) pyrolysis process, mainly the temperature, (Kim et al., 2013); (iii) biochar post activation or modification (Xue et al., 2012; Mohan et al., 2014b); (iv) different content of various mineral components in the biochar (Xu et al., 2013a); (v) different sorption mechanisms (Lu et al., 2012); as well as (vi) the conditions used during metal sorption (such as pH value, contact time, temperature, initial metal concentration or the dose of used biochar; Kołodyn´ska et al., 2012; Bernardo et al., 2013). The metal sorption process has been previously described as a result of three different mechanisms (Sohi et al., 2010; Lu et al., 2012): (i) ion exchange (Ca2+, K+, Mg2+, Na+); (ii) metal complexation onto free and complexed carbonyl, carboxyl, alcoholic, hydroxyl or phenolic hydroxyl functional groups; and (iii) physical adsorption or surface precipitation caused by sorptive interaction involving delocalization of p electrons of organic carbon (Inyang et al., 2011; Lu et al., 2012; Xu et al., 2013b; Bernardo et al., 2013). Nevertheless, there is a gap of knowledge in comparing different mechanisms that occur during metal sorption on contrasting biochars (originating from different waste materials). Additionally, ideal conditions for metal sorption efficiency have also been tested for different biochar samples. However, these crucial experiments are time consuming and expensive. The multivariate chemometric technique was, therefore, used to identify the optimum combination of factors (conditions used during sorption) and interactions among the factors, which could not be identified using the univariate method (Bingöl et al., 2012). In general, these chemometric techniques have several important advantages over one-way optimization for analytical applications, including a relatively low cost, a reduced number of experiments, and possibilities to evaluate interactions among variables (Tarley et al., 2009). More precisely, the Response Surface Methodology (RSM; Myers et al., 2009) was employed to develop an approach for the evaluation of Cd and Pb optimal sorption on biochars originating from different waste materials. The main objective of this novel study was to evaluate the feasibility of Cd and Pb removal by various biochars. For this purpose, the sorption of Cd(II) and Pb(II) was tested in a series of batch experiments. The effects of solution pH, contact time, metal concentration and dose of used biochar were accurately investigated using the efficient and time saving RSM approach. Furthermore, in order to deeply understand the different Cd and Pb sorption mechanisms in the interaction with all tested biochars, the geochemical and spectroscopic analyses of metal-loaded biochars were investigated.
443
2.2. Biochar preparation and analyses
2. Methods
Five different waste agro-materials (nut shells, plum stones, wheat straws, grape stalks and grape husks), which are commonly found in Central Europe, were used in this study. Although all the selected materials belong to the same group titled ‘‘herbaceous and agricultural biomass (HAB)’’, they differ in the content of structural components and also vary in their chemical composition (Vassilev et al., 2010, 2012). All materials were homogenised, air dried overnight and analysed (before biochar preparation) to determine the bulk density, moisture, ash content and material composition according to TAPPI T264 (1997) and TAPPI T211 (1993). Next, pH was measured using inoLabÒ pH-meter (7310 WTW, Germany); content of C, H, N, O and S was determined using the Flash EA 1112 apparatus in the CHNS/O configuration (Thermo Fisher Scientific, USA); and Fourier transform infrared spectroscopy (FTIR Nicolet 6700 analyser connected with a Continuum microscope, Thermo-Nicolet, USA) was performed to identify the chemical functional groups presence separately for each sample. These previously homogenised and analysed agro-wastes (Table A.1; see the Appendix section available as a Supplementary Material) were then pyrolysed at 600 °C in a muffle furnace under 16.7 mL min1 nitrogen flow rate at atmospheric pressure and retention time of 30 min. Additionally, the yield of biochar (in %) was calculated as the quotient between the weight of biochar and weight of agrowaste. The resulting biochars were then cooled overnight under the same nitrogen flow rate as before. Pyrolysed products prepared in this way were then ground, homogenised, sieved (all used biochar particles were 0.25–0.50 mm in size), washed by ultra-clean water MilliQ Integral (Merck Millipore Corp., USA) and dried at 60 °C for 24 h until constant weight. Different biochars such as: nut shells biochar (NSBC); wheat straws biochar (WSBC); grape stalks biochar (GSBC); grape husks biochar (GHBC); and plum stones biochar (PSBC); were then used in all studies/experiments with a separate analysis for each. Selected initial characteristics of all five biochar samples were described before the sorption experiments started. All prepared biochar samples were analysed to determine the yield, bulk density, moisture and ash content. Surface areas were measured by N2 adsorption isotherms at 77 K using ASAP 2050 (Micrometrics Instrument Corporation, USA) surface area analyser. Specific surface area and volume of micro-pores were detected by the layered adsorption isotherm BET model.22. For each biochar sample, a single estimate analysis of the element composition in triplicate was carried out using FlashEA 1112th (Thermo Scientific Company, USA). Additionally, ash mineralogical composition of each biochar was analysed, using the XRF method, by ARL 9400 XP spectrometer (Thermo Arl., Switzerland). Surface functional groups of all five biochar samples were then analysed using FTIR (Thermo-Nicolet, USA), cation exchange capacity was determined; pH value was measured and pH at point of zero charge (pHzpc) was determined according to Fiol and Villaescusa (2009) using the immersion technique (IT). Finally, the surface structure and composition of the selected biochar samples was analysed by SEM-EDX using INDUSEM Scanning Electron Microscope (Tescan Inc., USA) equipped by the Electron Dispersive X-ray Spectroscopy (Bruker Quantax 125 eV).
2.1. Solutions preparation
2.3. Batch experiments
Cadmium Cd(II) and lead Pb(II), as metals with different geochemical behaviour, were the sorbates studied in all experiments. Stock solutions (1000 mg L1 of both metals) were prepared by dissolution of Cd(NO3)24H20 and Pb(NO3)2 (p.a.), (Lach-Ner, Czech Republic), respectively, in a previously prepared background electrolyte (0.01 M NaNO3). Solution pH was adjusted using 0.1 M HNO3 and NaOH, respectively.
First, the batch sorption experiment was carried out with (i) initial metal concentration (0.1–1.5 mM); (ii) initial aqueous solution pH (2.00–8.00); (iii) specific dose of biochar (1.0–10.0 g L1); and (iv) contact time (1–48 h) to optimize the maximum Cd and Pb sorption efficiency of all biochar samples. Second, in order to test the sorption maximum of all biochar samples, two consecutive equilibrium sorption experiments (in triplicates) were implemented under
444
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
equally defined and optimized conditions. Third, all such Pb- and Cdfully-loaded biochar samples were then shaken for 24 h (in triplicates) with: (i) 0.01 M NaNO3 (background electrolyte at pH = 5.00); (ii) 0.01 M CaCl2 (solution simulating ‘‘bioavailable form’’ of metals; Houba et al., 1996); and (iii) 0.43 M HNO3 (solution representing ‘‘geochemically active form’’ of metals; Tipping et al., 2003) to evaluate potential metal desorption. Suspensions of all batch experiments were shaken at 250 rpm using end-over agitator. The aqueous solution phase was then separated from the sorbent using a centrifuge at 1046 RCF (g) for 10 min and filtered through the 0.45 lm nylon filter (VWR, Germany). The residual concentrations of Pb and Cd ions in the resulting supernatant were then determined using ICP-OES (Agilent Technologies 700 series, USA).
(CCD) was implemented using Minitab software version 16. The design consisting of 20 experiments was used to assess the influence of 3 factors on 2 responses for 5 biochar samples, respectively. According to the design matrix (Table A.2), the order was randomized to avoid systematic errors in the experiments. The three factors studied were initial pH of solution (2.00–8.00), dose of applied biochar (1.00–10.0 g L1), and contact time (1–48 h). The measured responses were the sorbed amount of Pb (mg g1) and sorbed amount of Cd (mg g1). Five different biochars were used in the study with a separate analysis for each (see Subchapter 2.2). Next, another experiment was then implemented when the previously determined optimized combination of these three factors was taken into account to establish optimum initial metal concentration and to test two different biochar pHs (original vs. setting up to pH = 7.00 for all biochars).
2.4. Metal sorption mechanisms analyses The FTIR spectroscopy was used to compare potential changes in the functional groups of metal-loaded biochars with the biochar samples before the Cd and/or Pb sorption. The existence of possible precipitates of Cd and/or Pb after metal sorption was checked using the scanning electron microscopy (SEM) coupled with electron dispersive X-ray analyser (EDX). Finally, to distinguish different metal sorption mechanisms in more detail for each biochar, binding energy detection was implemented for each Cd- and Pb-loaded biochar sample using X-ray photoelectron spectroscopy (Omicron Nanotechnology, Ltd.) and Casa XPS program (for the spectra evaluation). More precisely, two positions were analysed for each Cdor Pb-loaded biochar: (i) surface metal sorption; and (ii) position inside each biochar (approx. 10 nm deep), which was implemented for 5 min by Ar ion sputtering at the intensity of 5 keV and at speed of 120 nm h1, to detect potential absorption of Cd or Pb in the biochar structure. Additionally, X-ray diffraction (XRD) analysis (PANalytical X’Pert Pro diffractometer with X’Celerator detector) was implemented to detect potential precipitated phases after metal sorption on biochar. 2.5. Response Surface Methodology The batch sorption procedures can be designed using a Response Surface Methodology (RSM) approach. One of the main objectives of RSM is to determine optimum settings of the control variables that result in a maximum (or a minimum) response over a certain region of interest (Khuri and Mukhopadhyay, 2010). Response surfaces are used to determine an optimum and to graphically illustrate the relation between different experimental variables and their responses. In order to determine an optimum it is necessary for the polynomial function to contain quadratic terms. The following quadratic model (1) is used to fit experimental data (Montgomery, 2008; Khuri and Mukhopadhyay, 2010):
y ¼ b0 þ
k k k X X X bi xi þ bii x2i þ bij xi xj þ e i¼1
i¼1
ð1Þ
16i6j
where x1, x2,. . . are numbers of associated control (or input) variables, y is a response of interest, b are constant coefficients referred to as parameters and e is a random experimental error assumed to have a zero mean. The selected independent variables Xi were coded as xi according to the following relationship (2):
xi ¼
Xi X0 Dx
ð2Þ
where X0 is the uncoded value of Xi at the centre point and Dx presents the step change. Sorption experiments were optimized to determine the sorbed amounts of Pb and Cd (q) using an RSM. A central composite design
3. Results and discussion 3.1. Characteristics of used biochars All selected raw agro-wastes (before pyrolysis) showed variable bulk density (from 0.07 to 0.37 g cm3) and material composition (e.g. ash content varied between 1.56% and 7.78%; Table A.1). It can also be noticed that all the materials presented variations in elemental composition as well as in pH value (e.g. pH of GH = 3.93 and, on the other hand, pH of WS = 6.78, see Table A.1) depending on the origin of each material. More precisely, higher pH value favour deprotonation of the biochar’s carboxylic and other acidic functions, thus the negative sites are created on the surface of all biochars (Mohan et al., 2014b). Next, all FTIR spectra of raw materials, ashes and produced biochars are presented in Fig. A.1. There is a distinct contrast between the spectra of raw materials and those of ashes and biochars. The functional group COO, which is responsible for the metal chelation (Mohan et al., 2014a) was detected (more or less) only after the pyrolysis process (biochar) but was missing after combustion (ash). Additionally, the presence of CO2 3 , responsible for the metal precipitation (Inyang et al., 2012; Xu et al., 2013a,b) was detected in both cases (biochar and ash). The main properties of all prepared biochars are presented in Table 1. First, the biochar yield from grape stalks and husks was more than 30%, while the yields from other materials varied from 18.9% (WS) to 24.7% (PS). It was also found that the volume of micropores, which is very strongly related to the BET surface (Harvey et al., 2011) of NSBC and PSBC is 6-times greater in comparison with the biochars originating from grape husks and stalks. The BET surface and micropore volume of wheat straw biochar was comparable with the nut shields and plum stone biochars. These results are in agreement with the well-developed structure of NSBC/PSBC caused by the higher content of lignin (see Table A.1; Harvey et al., 2011), partly well-developed structure of WSBC (different for the surface and internal space of straws, respectively) and poorly-developed structure of the biochars originating from grape husks and stalks. On the other hand, pH values of the studied biochars follow in a decreasing order: GSBC > GHBC WSBC NSBC > PSBC, which is not in the same sequence as the pH of raw materials. The pH values vary from 7.36 (for the PSBC) to 10.0 (for the GSBC). Together, the biochar production using pyrolysis caused the pH increase in all cases (pH from 2.16 to 6.05), which is in agreement with the study of Lu et al. (2012). Such pH increase after the pyrolysis process was caused by the alkali salts separation from the organic matrix (at pyrolysis temperature ranging from 300 °C to 600 °C; Chen et al., 2011). Furthermore, Table 1 shows the cation exchange capacity (CEC) of all tested biochars, which ranged from 84.4 mmol kg1 (NSBC) to 402 mmol kg1 (GSBC).
445
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
Table 1 Biochar characteristics: yield of biochar from the waste material (Y), bulk density (q), BET surface, volume of micropores (Vmicro), pH, pH of zero point charge (pHzpc), cation exchange capacity (CEC), ash content in dry sample (Ad) and elemental composition.
NSBC WSBC GSBC GHBC PSBC
Y (%)
q (g cm3)
BET (m2 g1)
Vmicro (mm3 g1)
pH (–)
pHzpc (–)
CEC (mmol kg1)
21.8 18.9 30.6 31.6 24.7
0.17 ± 0.002 0.26 ± 0.005 0.16 ± 0.004 0.21 ± 0.003 0.22 ± 0.002
465 364 72 77 443
180 130 30 32 172
8.63 ± 0.04 9.86 ± 0.05 10.0 ± 0.1 9.98 ± 0.01 7.36 ± 0.12
7.55 ± 0.11 9.67 ± 0.09 9.92 ± 0.10 9.61 ± 0.09 7.28 ± 0.09
84.4 ± 3.0 334 ± 2 402 ± 3 187 ± 4 121 ± 7
NSBC WSBC GSBC GHBC PSBC *
Ad
Elemental composition (w %)
(w %)
N
C
H
S
O
1.68 ± 0.01 21.9 ± 0.2 16.1 ± 0.1 11.8 ± 0.1 5.33 ± 0.15
0.31 ± 0.01 0.97 ± 0.02 1.45 ± 0.02 1.93 ± 0.07 0.82 ± 0.05
88.2 ± 0.10 70.4 ± 0.47 70.2 ± 0.56 74.9 ± 0.71 81.8 ± 0.76
2.10 ± 0.04 1.73 ± 0.01 1.70 ± 0.03 1.92 ± 0.07 2.01 ± 0.03
–* –* –* –* –*
7.82 ± 0.02 6.27 ± 0.05 12.5 ± 0.08 10.8 ± 0.06 10.6 ± 0.04
7.00 g L1) and longer time (18–48 h), whereas WSBC, GHBC and GSBC should be applied at lower dose (optimum at 1.00 g L1 for Pb and 3.50–6.50 g L1 for Cd) and need less time (approx. 1 h) for the sorption of both metals. This fact (concerning the optimized lower dose) was also confirmed by Kołodyn´ska et al. (2012), where the optimum dose of applied biochar was also not the highest. Additionally, two different kinetics should then be attributed with two different metal sorption mechanisms. The first mechanism should reflect a ‘‘slower’’ metal sorption, which is ‘‘pure’’
physical sorption on the surface (for NSBC and PSBC). The second mechanism should reflect a ‘‘faster’’ chemisorption, where the higher applied dose of biochar has no additional effect on the metal sorption efficiency. The pH value is generally a very important parameter during the metal sorption process. It affects not only the surface charge of the adsorbent, but also the speciation of metal ions in the solution. The pH effect of initial solution was obvious for each tested biochar (data not shown), which is in agreement with the studies of Kołodyn´ska et al. (2012), Bernardo et al. (2013) and the optimal conditions (for each biochar) are as follows: (i) pH = 6.50/8.00 for NSBC and PSBC vs. pH = 5.00 for WSBC, GHBC and GSBC during the Cd sorption; and (ii) pH = 5.00 for all biochars during the Pb sorption. Moreover, pH of the tested biochar has a significant effect on the final metal sorption efficiency as well (Fig. 1c). When setting up the biochar pH to pH = 7.00 for all tested biosorbents, the metal sorption efficiency was reduced due to strong dependency of the metal removal efficiency on the alkali pH of the utilized biochar (Inyang et al., 2012). Next, the effect of initial metal concentration on the removal efficiency was also tested. An increasing and consequently decreasing tendency in the removal efficiency (of all biochars) was observed with the increase in the initial metal concentration (Fig. 1a and b). It should be expected that all tested biochars have a limited number of binding sites for Cd and Pb, respectively (Bernardo et al., 2013). Such optimal initial concentration of both metals in the solution was up to 1 mM of sorbed Cd and Pb, respec-
Table 2 Optimized conditions and (RSM) equations for the Cd(II) and Pb(II) removal efficiency (ECd; EPb) on five biochar samples. Cd(II)
pH (–)
Dose (g L1)
Time (h)
ECd (%)
R2adj (%)
RSM equation(Cd)
NSBC
6.50
10.0
42
56.0
99.7
ENSBC ¼ 27:2361 þ 5:9701 pH þ 14:1270 D þ 0:9927 T 4:9685 pH2 1:2922 T 2 þ 1:1661 pH D
WSBC
5.00
5.00
1
98.6
95.1
EWSBC ¼ 101:859 þ 13:458 pH þ 12:428 D þ 0:458 T 12:402 pH2 7:370 D2
GSBC
5.00
3.50
1
99.3
94.5
EGSBC ¼ 95:0619 þ 10:8699 pH þ 7:3355 D 0:1944 T 8:4292 pH2 6:1144 D2 þ 4:4453 T 2
GHBC
5.00
1
>99.9
99.7
EGHBC ¼ 96:914 þ 18:011 pH þ 25:392 D þ 2:609 T 15:397 pH2 12:755 D2 þ 1:1661 pH D 10:470 pH D
PSBC
8.00
48
43.8
98.9
EPSBC ¼ 22:2360 þ 4:5462 pH þ 12:1435 D þ 0:8121:T 2:8748 pH2 1:3353 T 2 þ 1:4264 pH D
Time (h)
EPb (%)
R2adj
RSM equation(Pb)
(%)
Pb(II)
pH (–)
6.50 10.0 Dose (g L1)
NSBC
5.00
7.60
18
99.7
99.9
WSBC
5.00
1.00
1
>99.9
92.6
EWSBC ¼ 102:207 þ 6:496 pH 0:026 D þ 0:134 T 8:123 pH2
GSBC
5.00
1.60
1
99.8
98.8
EGSBC ¼ 61:7514 1:7802 pH 12:0728 D þ 0:8867 T þ 8:6329 pH2 þ 10:843 D2
GHBC
5.00
1.00
1
95.9
96.0
EGHBC ¼ 94:462 þ 5:319 pH 1:398 D þ 1:374 T 4:935 pH2 2:648 D2 þ 3:145 T 2
PSBC
5.00
8.0
18
99.8
99.9
EPSBC ¼ 88:191 þ 18:902 pH þ 28:262 D þ 6:646 T 11:820 pH2 12:073 D2 5:050 T 2
R2adj – determination coefficient.
ENSBC ¼ 92:721 þ 09:841 pH þ 24:925 D þ 3:597 T 12:763 pH2 10:316 D2 12:451 pH D
446
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
Fig. 1. Effect of initial metal concentration (a/b) and biochar pH (c) on Cd(II) and Pb(II) removal efficiency.
tively (for all tested biosorbents under previously optimized biochar dose, solution pH and contact time).
(Fig. 2b); due to the pH-buffering effect of all biochar samples (Trakal et al., 2014).
3.2.2. Sorption kinetics and isotherms The kinetics of Cd and Pb sorption on all tested biochars (Fig. A.3) was solved in the previous experimental step by using the RSM approach in order to optimize contact time (Table 2). According to these results, the sorption time was established after 24 h for all biochar samples to ensure the steady state (changes after 1 h were minimal for all BCs; Fig. A.3). Next, according to other previously optimized conditions (dose of applied biochar and initial pH of solution), these parameters were set up equally for all biochars as follows: pH = 5.00 for Cd, Pb and mixture of Cd + Pb; and applied dose of biochar at 1 g L1 (according to well-sorbed samples). The equilibrium sorption isotherms of Cd and Pb including their mixture for each tested biochar (presented in Fig. 2a) fitted the Langmuir (A.2) model better than the Freundlich (A.1) one (according to model efficiency E; Table A.3). The presented Langmuir isotherms (Fig. 2a) significantly varied not only between the two tested metals, but also between the biochars. Furthermore, in case of multi-metal sorption, there was an obvious competition effect of Pb during Cd sorption, whereas the effect of Cd on the Pb sorption was negligible, as expected Xu et al. (2013a). In more detail, the GSBC was effective for sorption of Cd and Pb in quantities of 0.45 and 2.87 mmol g1, respectively (Table A.3; Fig. 2a). On the other hand, the PSBC removed only 0.04 mmol g1 of Cd and 0.11 mmol g1 of Pb, respectively (Table A.3; Fig. 2a). Such difference in metal sorption efficiency is caused by variable biochar characteristics (e.g. pH, CEC, BET surface or functional groups presence) and different metal sorption mechanisms as well. Furthermore, the pH of all biochar suspensions after equilibrium decreased gradually with increasing of metal(s) concentration in the initial solution during sorption
3.2.3. Cd and Pb sorption mechanisms As mentioned before, the sorption efficiency varied not only between the two tested metals, but mostly between all the tested biochars. This could be caused by different sorption mechanisms of each metal on tested biochars. As described by Sohi et al. (2010), Lu et al. (2012), Ahmad et al. (2014), ion exchange, complexation and physical adsorption are responsible for the metal sorption to biochar. Nevertheless, these metal sorption mechanisms vary according to the type of biochar (group of biochars originating from similar materials) as presented in this study. In more detail, the following post-sorption analyses (FTIR, SEM-EDX, XPS and XRD) were implemented to specify these mechanisms of both metals for each tested biochar. First, FTIR spectra showed that the sorption of Cd and Pb induced the peak of CO2 at 1397/1408 cm1 in the GSBC/GHBC 3 shift to 1413 cm1 for Cd (for both biochars) and the shift to 1398/1397 cm1 for Pb, respectively (Fig. A.4). The presence of the CO2 3 peaks only in case of GSBC and GHBC is probably reflecting the higher ash content of these two original materials (Table A.1; Fig. A.2). The shift of these peaks was, therefore, attributed to ion exchange (obvious from XPS analyses; Fig. A.5) or surface precipitation of metal-(hydro)carbonates (confirmed by the XRD analyses; Fig. A.7), which play an important role in the sorption capacity and the removal mechanism of both sorbates (Xu et al., 2013a,b). Additionally, in the studies of Xu et al. (2013a,b) the metal precipitates with PO3 4 have also been reported. Nevertheless, in this study the metal-phosphates have not been found (confirmed by EDX; Fig. A.6). Subsequently, the band for the surface complexation with carboxylic groups (ACOO-metal chelate) shows the following results: (i) the peak of COO at 1571 cm1
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
447
Fig. 2. Cd(II) and Pb(II): (a) sorption isotherms; and (b) pH changes during single- and multi-metal sorption. pHinitial = 5.00; T = 22 ± 1 °C; mBC = 1 g L1; t = 24 h; and c0 = 0.1– 1.5 mM.
in the original WSBC shifted to 1562 cm1 in Cd- and Pb-loaded biochars, respectively; (ii) the peak of the carboxyl at 1571 cm1 in the GSBC/GHBC shifted to 1562/1569 cm1 for Cd and to 1551/ 1564 cm1 for Pb, respectively; and (iii) comparison of pre- and post-sorption COO spectra for NSBC and PSBC, respectively, showed similarity, which provides no evidence of Cd and Pb chela-
tion carboxylic complex on/in these two biochars. In summary, these results did not show a significant shift caused by the Cd and Pb sorption. This explains the complexation with carboxylic groups as a non-significant mechanism in metal sorption. This fact is in agreement with the studies of Inyang et al. (2012) and Lu et al. (2012), but contrasts with the studies of Xu et al. (2013a,b). These
448
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
differences can be explained by the different temperature during material pyrolysis (200–350 °C studies of Xu et al. (2013a,b)) and (550–600 °C studies of Inyang et al. (2012), Lu et al., (2012); and this study). Furthermore, hydroxyl groups (AOH), responsible for the coordination of a metal d-electron and producing of AO-M bond (Inyang et al., 2011), were not detected in our case. This should be probably caused by the dehydration (cleavage of hydroxyl group) during high temperature of pyrolysis as well. As another approach, XPS analyses were implemented to specify sorption mechanisms using the results of different binding energies of both sorbed metals (Fig. A.5). Analyses were performed on the surface and at approximately 10 nm depth of each metal-loaded biochar. Each tested biochar showed a different quantity of loaded metal (Cd and Pb) on the surface and at the 10 nm depth, respectively. More precisely, the amount of loaded metals decreased in the following order (Fig. A.5a,c): (i) GHBC > WSBC GSBC NSBC > PSBC for Cd on the surface; (ii) GHBC > GSBC > WSBC PSBC > NSBC for Cd at 10 nm depth; and (iii) GSBC > GHBC > WSBC PSBC > NSBC for Pb both on the surface and 10 nm depth. Furthermore, from the results of the two different positions (surface vs. 10 nm depth) it is obvious that the amount of sorbed metals decreased inwards for (Cd-, PbNSBC; Cd-, Pb-PSBC; and Cd-WSBC) while it decreased outwards for (Cd-, Pb-GSBC; Cd-, Pb-GHBC; and Pb-WSBC). These results suggest different sorption mechanisms, where NSBC and PSBC with well-developed structure sorbed both metals predominantly on the surface. By contrast, in the other three biochars the metals are not sorbed solely on the surface, but are also bonded inside the biochar structure (at higher concentrations). In more detail, Fig. A.5b shows exact sorption mechanisms of each metal on/in studied biochars. The weak p-bindings of the two metals with poly-organic chains (bonding with electron-rich domains on graphene-like structures; Harvey et al., 2011) demonstrated physical adsorption (Lu et al., 2012). This finding is supported by the binding energy results which were 139 and 406 eV for Pb and Cd, respectively and were detected in all tested biochars. The ion exchange as another significant sorption mechanism, was detected in all tested biochars as well (mainly inside the biochar structure; Fig. A.5b). More specifically, such cation release of both metals is limited predominantly to internal positions, whereas the confirmed presence of precipitated metal-(hydro)carbonates were limited solely to the biochar surface (Xu et al. (2013a,b)). Additionally, these binding energies of Pb with the biochar are 137 and 142 eV and 408 eV in case of Cd, respectively. The complexation with carboxylic functional groups (as the third sorption mechanism) is demonstrated in Fig. A.5b by the binding energy at 141 eV for the Pb and by the binding energy at 407 eV for the Cd, respectively. In more detail, this sorption mechanism (metal-chelate creation) is, nevertheless, obvious only in GSBC biochar for both metals and in WSBC solely for Pb. Additionally, these metal-chelates are created not only on the biochar surface, but also inside, which is in agreement with the study of Lu et al. (2012). These results confirmed the previously mentioned fact (by the FTIR analyses), that this chemical process is a non-significant mechanism (limited to GSBC and WSBC) during metal sorption (explained earlier in this Subchapter). Finally, to confirm the different removal efficiency and to see the contrasting structure of all tested biochars during metal sorption, the metal loaded biochars were also scanned by SEM and selected areas were also analysed by EDX (Fig. A.6). The SEM images show very different surface morphology, which reflects the results of BET surface and volume of micro-pores as well (Table 1). Fig. A.6 shows the surface morphology of all biochars at lower (Fig. A.6a) and higher (Fig. A.6b) resolutions. The ‘‘high lignin biomass/biochars’’ (NSBC an PSBC) is showing well-developed structure with a high macro- and micro-porosity, as also shown
by the studies of Chen et al. (2011) and Mohan et al. (2014b). On the other hand, poorly-developed structure of GSBC and GHBC shows limited presence of these pores. This is in agreement with the lower BET surface and lower volume of micro-pores in these cases. Additionally, the structure of the WSBC is very heterogeneous. The surface is poorly-developed (Fig. A.6), nevertheless the structure of this biochar is microporous (Fig. A.8) and, therefore, the value of the BET surface of WSBC is also high. Moreover, the morphology of biochar is not a crucial factor which would affect the metal sorption efficiency (discussed before). This fact is supported by this study, where the biochars with poorly-developed structure (such as GSBS and GHBC) are very efficient in removing both the studied metals from aqueous solution. This metal loading is also confirmed by the EDX analyses coupled to the SEM images. As visible from Fig. A.6, the ‘‘high lignin’’ NSBC and PSBC loaded both metals in very low amounts (
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Geochemical and spectroscopic investigations of Cd and Pb sorption mechanisms on contrasting biochars: Engineering implications Lukáš Trakal a,⇑, Deniz Bingöl b, Michael Pohorˇely´ c, Miroslav Hruška a, Michael Komárek a a
Department of Environmental Geosciences, Faculty of Environmental Sciences, Czech University of Life Sciences Prague, Kamycka 129, Praha 6 Suchdol 16521, Czech Republic Department of Chemistry, Faculty of Sciences and Arts, Kocaeli University, Umuttepe Campus, 41380 Kocaeli, Turkey c Environmental Process Engineering Laboratory, Institute of Chemical Process Fundamentals, Academy of Sciences of Czech Republic, v.v.i., Rozvojova 135, Praha 6 Suchdol 16502, Czech Republic b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Metal removal efficiencies and
mechanisms depended on the parent material. Ion exchange is the predominant mechanism during metal sorption on biochars. Metals were strongly bound to the biochars. Metal removal efficiency is strongly pH-dependent.
a r t i c l e
i n f o
Article history: Received 31 July 2014 Received in revised form 22 August 2014 Accepted 24 August 2014 Available online 30 August 2014 Keywords: Biochar Metals Response Surface Methodology (RSM) Sorption mechanisms Desorption
a b s t r a c t Biochars prepared from nut shells, plum stones, wheat straws, grape stalks and grape husks were tested as potential sorbents for Cd and Pb. Mechanisms responsible for metal retention were investigated and optimal sorption conditions were evaluated using the RSM approach. Results indicated that all tested biochars can effectively remove Cd and Pb from aqueous solution (efficiency varied between 43.8% and 100%). The removal rate of both metals is the least affected by the biochar morphology and specific surface but this removal efficiency is strongly pH-dependent. Results of variable metal removal combined with different optimized conditions explain the different metal sorption mechanisms, where the predominant mechanism is ion exchange. In addition, this mechanism showed very strong binding of sorbed metals as confirmed by the post-desorption of the fully metal-loaded biochars. Finally, these biochars could thus also be applicable for metal contaminated soils to reduce mobility and bioavailability of Cd and Pb. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Among other water treatment technologies, such as precipitation, coagulation or membrane filtration, biosorption should be taken into account as a potential ‘‘low-cost candidate’’ for removing toxic metals from wastewater. Biochar is an efficient
⇑ Corresponding author. Tel.: +420 224 383 864. E-mail address: [email protected] (L. Trakal). http://dx.doi.org/10.1016/j.biortech.2014.08.108 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
biosorbent used for the removal of metals such as Cd and/or Pb (two common metal pollutants in Central Europe with contrasting geochemical behaviour) from waste waters (Inyang et al., 2011; Lu et al., 2012; Kim et al., 2013; Mohan et al., 2014a). This pyrogenic carbon-rich material is most commonly prepared from waste materials, such as woody biomass, animal waste, sludge or a variety of agricultural residues (Lehmann and Joseph, 2009). Many studies have described an effective removal of Cd and Pb using biochar originating from different agricultural wastes. For example, Kołodyn´ska et al. (2012), Xu et al. (2013a,b) reported
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
that biochars made from dairy and pig manure were able to sorb Cd and/or Pb in the range from 0.28 to 1.11 mmol g1. Inyang et al. (2011), Jiang et al. (2012), Han et al. (2013), Kim et al. (2013) and Xu et al. (2013a) described the high potential of sugarcane bagasse, rice straw, pinewood, rice husk and Miscanthus sacchariflorus to remove Cd and/or Pb from aqueous solution in the range of 0.004–0.66 mmol g1. These results thus present variable metal sorption efficiency of the tested biochars. This variability is caused by many factors, such as: (i) source and composition of the original waste material; (ii) pyrolysis process, mainly the temperature, (Kim et al., 2013); (iii) biochar post activation or modification (Xue et al., 2012; Mohan et al., 2014b); (iv) different content of various mineral components in the biochar (Xu et al., 2013a); (v) different sorption mechanisms (Lu et al., 2012); as well as (vi) the conditions used during metal sorption (such as pH value, contact time, temperature, initial metal concentration or the dose of used biochar; Kołodyn´ska et al., 2012; Bernardo et al., 2013). The metal sorption process has been previously described as a result of three different mechanisms (Sohi et al., 2010; Lu et al., 2012): (i) ion exchange (Ca2+, K+, Mg2+, Na+); (ii) metal complexation onto free and complexed carbonyl, carboxyl, alcoholic, hydroxyl or phenolic hydroxyl functional groups; and (iii) physical adsorption or surface precipitation caused by sorptive interaction involving delocalization of p electrons of organic carbon (Inyang et al., 2011; Lu et al., 2012; Xu et al., 2013b; Bernardo et al., 2013). Nevertheless, there is a gap of knowledge in comparing different mechanisms that occur during metal sorption on contrasting biochars (originating from different waste materials). Additionally, ideal conditions for metal sorption efficiency have also been tested for different biochar samples. However, these crucial experiments are time consuming and expensive. The multivariate chemometric technique was, therefore, used to identify the optimum combination of factors (conditions used during sorption) and interactions among the factors, which could not be identified using the univariate method (Bingöl et al., 2012). In general, these chemometric techniques have several important advantages over one-way optimization for analytical applications, including a relatively low cost, a reduced number of experiments, and possibilities to evaluate interactions among variables (Tarley et al., 2009). More precisely, the Response Surface Methodology (RSM; Myers et al., 2009) was employed to develop an approach for the evaluation of Cd and Pb optimal sorption on biochars originating from different waste materials. The main objective of this novel study was to evaluate the feasibility of Cd and Pb removal by various biochars. For this purpose, the sorption of Cd(II) and Pb(II) was tested in a series of batch experiments. The effects of solution pH, contact time, metal concentration and dose of used biochar were accurately investigated using the efficient and time saving RSM approach. Furthermore, in order to deeply understand the different Cd and Pb sorption mechanisms in the interaction with all tested biochars, the geochemical and spectroscopic analyses of metal-loaded biochars were investigated.
443
2.2. Biochar preparation and analyses
2. Methods
Five different waste agro-materials (nut shells, plum stones, wheat straws, grape stalks and grape husks), which are commonly found in Central Europe, were used in this study. Although all the selected materials belong to the same group titled ‘‘herbaceous and agricultural biomass (HAB)’’, they differ in the content of structural components and also vary in their chemical composition (Vassilev et al., 2010, 2012). All materials were homogenised, air dried overnight and analysed (before biochar preparation) to determine the bulk density, moisture, ash content and material composition according to TAPPI T264 (1997) and TAPPI T211 (1993). Next, pH was measured using inoLabÒ pH-meter (7310 WTW, Germany); content of C, H, N, O and S was determined using the Flash EA 1112 apparatus in the CHNS/O configuration (Thermo Fisher Scientific, USA); and Fourier transform infrared spectroscopy (FTIR Nicolet 6700 analyser connected with a Continuum microscope, Thermo-Nicolet, USA) was performed to identify the chemical functional groups presence separately for each sample. These previously homogenised and analysed agro-wastes (Table A.1; see the Appendix section available as a Supplementary Material) were then pyrolysed at 600 °C in a muffle furnace under 16.7 mL min1 nitrogen flow rate at atmospheric pressure and retention time of 30 min. Additionally, the yield of biochar (in %) was calculated as the quotient between the weight of biochar and weight of agrowaste. The resulting biochars were then cooled overnight under the same nitrogen flow rate as before. Pyrolysed products prepared in this way were then ground, homogenised, sieved (all used biochar particles were 0.25–0.50 mm in size), washed by ultra-clean water MilliQ Integral (Merck Millipore Corp., USA) and dried at 60 °C for 24 h until constant weight. Different biochars such as: nut shells biochar (NSBC); wheat straws biochar (WSBC); grape stalks biochar (GSBC); grape husks biochar (GHBC); and plum stones biochar (PSBC); were then used in all studies/experiments with a separate analysis for each. Selected initial characteristics of all five biochar samples were described before the sorption experiments started. All prepared biochar samples were analysed to determine the yield, bulk density, moisture and ash content. Surface areas were measured by N2 adsorption isotherms at 77 K using ASAP 2050 (Micrometrics Instrument Corporation, USA) surface area analyser. Specific surface area and volume of micro-pores were detected by the layered adsorption isotherm BET model.22. For each biochar sample, a single estimate analysis of the element composition in triplicate was carried out using FlashEA 1112th (Thermo Scientific Company, USA). Additionally, ash mineralogical composition of each biochar was analysed, using the XRF method, by ARL 9400 XP spectrometer (Thermo Arl., Switzerland). Surface functional groups of all five biochar samples were then analysed using FTIR (Thermo-Nicolet, USA), cation exchange capacity was determined; pH value was measured and pH at point of zero charge (pHzpc) was determined according to Fiol and Villaescusa (2009) using the immersion technique (IT). Finally, the surface structure and composition of the selected biochar samples was analysed by SEM-EDX using INDUSEM Scanning Electron Microscope (Tescan Inc., USA) equipped by the Electron Dispersive X-ray Spectroscopy (Bruker Quantax 125 eV).
2.1. Solutions preparation
2.3. Batch experiments
Cadmium Cd(II) and lead Pb(II), as metals with different geochemical behaviour, were the sorbates studied in all experiments. Stock solutions (1000 mg L1 of both metals) were prepared by dissolution of Cd(NO3)24H20 and Pb(NO3)2 (p.a.), (Lach-Ner, Czech Republic), respectively, in a previously prepared background electrolyte (0.01 M NaNO3). Solution pH was adjusted using 0.1 M HNO3 and NaOH, respectively.
First, the batch sorption experiment was carried out with (i) initial metal concentration (0.1–1.5 mM); (ii) initial aqueous solution pH (2.00–8.00); (iii) specific dose of biochar (1.0–10.0 g L1); and (iv) contact time (1–48 h) to optimize the maximum Cd and Pb sorption efficiency of all biochar samples. Second, in order to test the sorption maximum of all biochar samples, two consecutive equilibrium sorption experiments (in triplicates) were implemented under
444
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
equally defined and optimized conditions. Third, all such Pb- and Cdfully-loaded biochar samples were then shaken for 24 h (in triplicates) with: (i) 0.01 M NaNO3 (background electrolyte at pH = 5.00); (ii) 0.01 M CaCl2 (solution simulating ‘‘bioavailable form’’ of metals; Houba et al., 1996); and (iii) 0.43 M HNO3 (solution representing ‘‘geochemically active form’’ of metals; Tipping et al., 2003) to evaluate potential metal desorption. Suspensions of all batch experiments were shaken at 250 rpm using end-over agitator. The aqueous solution phase was then separated from the sorbent using a centrifuge at 1046 RCF (g) for 10 min and filtered through the 0.45 lm nylon filter (VWR, Germany). The residual concentrations of Pb and Cd ions in the resulting supernatant were then determined using ICP-OES (Agilent Technologies 700 series, USA).
(CCD) was implemented using Minitab software version 16. The design consisting of 20 experiments was used to assess the influence of 3 factors on 2 responses for 5 biochar samples, respectively. According to the design matrix (Table A.2), the order was randomized to avoid systematic errors in the experiments. The three factors studied were initial pH of solution (2.00–8.00), dose of applied biochar (1.00–10.0 g L1), and contact time (1–48 h). The measured responses were the sorbed amount of Pb (mg g1) and sorbed amount of Cd (mg g1). Five different biochars were used in the study with a separate analysis for each (see Subchapter 2.2). Next, another experiment was then implemented when the previously determined optimized combination of these three factors was taken into account to establish optimum initial metal concentration and to test two different biochar pHs (original vs. setting up to pH = 7.00 for all biochars).
2.4. Metal sorption mechanisms analyses The FTIR spectroscopy was used to compare potential changes in the functional groups of metal-loaded biochars with the biochar samples before the Cd and/or Pb sorption. The existence of possible precipitates of Cd and/or Pb after metal sorption was checked using the scanning electron microscopy (SEM) coupled with electron dispersive X-ray analyser (EDX). Finally, to distinguish different metal sorption mechanisms in more detail for each biochar, binding energy detection was implemented for each Cd- and Pb-loaded biochar sample using X-ray photoelectron spectroscopy (Omicron Nanotechnology, Ltd.) and Casa XPS program (for the spectra evaluation). More precisely, two positions were analysed for each Cdor Pb-loaded biochar: (i) surface metal sorption; and (ii) position inside each biochar (approx. 10 nm deep), which was implemented for 5 min by Ar ion sputtering at the intensity of 5 keV and at speed of 120 nm h1, to detect potential absorption of Cd or Pb in the biochar structure. Additionally, X-ray diffraction (XRD) analysis (PANalytical X’Pert Pro diffractometer with X’Celerator detector) was implemented to detect potential precipitated phases after metal sorption on biochar. 2.5. Response Surface Methodology The batch sorption procedures can be designed using a Response Surface Methodology (RSM) approach. One of the main objectives of RSM is to determine optimum settings of the control variables that result in a maximum (or a minimum) response over a certain region of interest (Khuri and Mukhopadhyay, 2010). Response surfaces are used to determine an optimum and to graphically illustrate the relation between different experimental variables and their responses. In order to determine an optimum it is necessary for the polynomial function to contain quadratic terms. The following quadratic model (1) is used to fit experimental data (Montgomery, 2008; Khuri and Mukhopadhyay, 2010):
y ¼ b0 þ
k k k X X X bi xi þ bii x2i þ bij xi xj þ e i¼1
i¼1
ð1Þ
16i6j
where x1, x2,. . . are numbers of associated control (or input) variables, y is a response of interest, b are constant coefficients referred to as parameters and e is a random experimental error assumed to have a zero mean. The selected independent variables Xi were coded as xi according to the following relationship (2):
xi ¼
Xi X0 Dx
ð2Þ
where X0 is the uncoded value of Xi at the centre point and Dx presents the step change. Sorption experiments were optimized to determine the sorbed amounts of Pb and Cd (q) using an RSM. A central composite design
3. Results and discussion 3.1. Characteristics of used biochars All selected raw agro-wastes (before pyrolysis) showed variable bulk density (from 0.07 to 0.37 g cm3) and material composition (e.g. ash content varied between 1.56% and 7.78%; Table A.1). It can also be noticed that all the materials presented variations in elemental composition as well as in pH value (e.g. pH of GH = 3.93 and, on the other hand, pH of WS = 6.78, see Table A.1) depending on the origin of each material. More precisely, higher pH value favour deprotonation of the biochar’s carboxylic and other acidic functions, thus the negative sites are created on the surface of all biochars (Mohan et al., 2014b). Next, all FTIR spectra of raw materials, ashes and produced biochars are presented in Fig. A.1. There is a distinct contrast between the spectra of raw materials and those of ashes and biochars. The functional group COO, which is responsible for the metal chelation (Mohan et al., 2014a) was detected (more or less) only after the pyrolysis process (biochar) but was missing after combustion (ash). Additionally, the presence of CO2 3 , responsible for the metal precipitation (Inyang et al., 2012; Xu et al., 2013a,b) was detected in both cases (biochar and ash). The main properties of all prepared biochars are presented in Table 1. First, the biochar yield from grape stalks and husks was more than 30%, while the yields from other materials varied from 18.9% (WS) to 24.7% (PS). It was also found that the volume of micropores, which is very strongly related to the BET surface (Harvey et al., 2011) of NSBC and PSBC is 6-times greater in comparison with the biochars originating from grape husks and stalks. The BET surface and micropore volume of wheat straw biochar was comparable with the nut shields and plum stone biochars. These results are in agreement with the well-developed structure of NSBC/PSBC caused by the higher content of lignin (see Table A.1; Harvey et al., 2011), partly well-developed structure of WSBC (different for the surface and internal space of straws, respectively) and poorly-developed structure of the biochars originating from grape husks and stalks. On the other hand, pH values of the studied biochars follow in a decreasing order: GSBC > GHBC WSBC NSBC > PSBC, which is not in the same sequence as the pH of raw materials. The pH values vary from 7.36 (for the PSBC) to 10.0 (for the GSBC). Together, the biochar production using pyrolysis caused the pH increase in all cases (pH from 2.16 to 6.05), which is in agreement with the study of Lu et al. (2012). Such pH increase after the pyrolysis process was caused by the alkali salts separation from the organic matrix (at pyrolysis temperature ranging from 300 °C to 600 °C; Chen et al., 2011). Furthermore, Table 1 shows the cation exchange capacity (CEC) of all tested biochars, which ranged from 84.4 mmol kg1 (NSBC) to 402 mmol kg1 (GSBC).
445
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
Table 1 Biochar characteristics: yield of biochar from the waste material (Y), bulk density (q), BET surface, volume of micropores (Vmicro), pH, pH of zero point charge (pHzpc), cation exchange capacity (CEC), ash content in dry sample (Ad) and elemental composition.
NSBC WSBC GSBC GHBC PSBC
Y (%)
q (g cm3)
BET (m2 g1)
Vmicro (mm3 g1)
pH (–)
pHzpc (–)
CEC (mmol kg1)
21.8 18.9 30.6 31.6 24.7
0.17 ± 0.002 0.26 ± 0.005 0.16 ± 0.004 0.21 ± 0.003 0.22 ± 0.002
465 364 72 77 443
180 130 30 32 172
8.63 ± 0.04 9.86 ± 0.05 10.0 ± 0.1 9.98 ± 0.01 7.36 ± 0.12
7.55 ± 0.11 9.67 ± 0.09 9.92 ± 0.10 9.61 ± 0.09 7.28 ± 0.09
84.4 ± 3.0 334 ± 2 402 ± 3 187 ± 4 121 ± 7
NSBC WSBC GSBC GHBC PSBC *
Ad
Elemental composition (w %)
(w %)
N
C
H
S
O
1.68 ± 0.01 21.9 ± 0.2 16.1 ± 0.1 11.8 ± 0.1 5.33 ± 0.15
0.31 ± 0.01 0.97 ± 0.02 1.45 ± 0.02 1.93 ± 0.07 0.82 ± 0.05
88.2 ± 0.10 70.4 ± 0.47 70.2 ± 0.56 74.9 ± 0.71 81.8 ± 0.76
2.10 ± 0.04 1.73 ± 0.01 1.70 ± 0.03 1.92 ± 0.07 2.01 ± 0.03
–* –* –* –* –*
7.82 ± 0.02 6.27 ± 0.05 12.5 ± 0.08 10.8 ± 0.06 10.6 ± 0.04
7.00 g L1) and longer time (18–48 h), whereas WSBC, GHBC and GSBC should be applied at lower dose (optimum at 1.00 g L1 for Pb and 3.50–6.50 g L1 for Cd) and need less time (approx. 1 h) for the sorption of both metals. This fact (concerning the optimized lower dose) was also confirmed by Kołodyn´ska et al. (2012), where the optimum dose of applied biochar was also not the highest. Additionally, two different kinetics should then be attributed with two different metal sorption mechanisms. The first mechanism should reflect a ‘‘slower’’ metal sorption, which is ‘‘pure’’
physical sorption on the surface (for NSBC and PSBC). The second mechanism should reflect a ‘‘faster’’ chemisorption, where the higher applied dose of biochar has no additional effect on the metal sorption efficiency. The pH value is generally a very important parameter during the metal sorption process. It affects not only the surface charge of the adsorbent, but also the speciation of metal ions in the solution. The pH effect of initial solution was obvious for each tested biochar (data not shown), which is in agreement with the studies of Kołodyn´ska et al. (2012), Bernardo et al. (2013) and the optimal conditions (for each biochar) are as follows: (i) pH = 6.50/8.00 for NSBC and PSBC vs. pH = 5.00 for WSBC, GHBC and GSBC during the Cd sorption; and (ii) pH = 5.00 for all biochars during the Pb sorption. Moreover, pH of the tested biochar has a significant effect on the final metal sorption efficiency as well (Fig. 1c). When setting up the biochar pH to pH = 7.00 for all tested biosorbents, the metal sorption efficiency was reduced due to strong dependency of the metal removal efficiency on the alkali pH of the utilized biochar (Inyang et al., 2012). Next, the effect of initial metal concentration on the removal efficiency was also tested. An increasing and consequently decreasing tendency in the removal efficiency (of all biochars) was observed with the increase in the initial metal concentration (Fig. 1a and b). It should be expected that all tested biochars have a limited number of binding sites for Cd and Pb, respectively (Bernardo et al., 2013). Such optimal initial concentration of both metals in the solution was up to 1 mM of sorbed Cd and Pb, respec-
Table 2 Optimized conditions and (RSM) equations for the Cd(II) and Pb(II) removal efficiency (ECd; EPb) on five biochar samples. Cd(II)
pH (–)
Dose (g L1)
Time (h)
ECd (%)
R2adj (%)
RSM equation(Cd)
NSBC
6.50
10.0
42
56.0
99.7
ENSBC ¼ 27:2361 þ 5:9701 pH þ 14:1270 D þ 0:9927 T 4:9685 pH2 1:2922 T 2 þ 1:1661 pH D
WSBC
5.00
5.00
1
98.6
95.1
EWSBC ¼ 101:859 þ 13:458 pH þ 12:428 D þ 0:458 T 12:402 pH2 7:370 D2
GSBC
5.00
3.50
1
99.3
94.5
EGSBC ¼ 95:0619 þ 10:8699 pH þ 7:3355 D 0:1944 T 8:4292 pH2 6:1144 D2 þ 4:4453 T 2
GHBC
5.00
1
>99.9
99.7
EGHBC ¼ 96:914 þ 18:011 pH þ 25:392 D þ 2:609 T 15:397 pH2 12:755 D2 þ 1:1661 pH D 10:470 pH D
PSBC
8.00
48
43.8
98.9
EPSBC ¼ 22:2360 þ 4:5462 pH þ 12:1435 D þ 0:8121:T 2:8748 pH2 1:3353 T 2 þ 1:4264 pH D
Time (h)
EPb (%)
R2adj
RSM equation(Pb)
(%)
Pb(II)
pH (–)
6.50 10.0 Dose (g L1)
NSBC
5.00
7.60
18
99.7
99.9
WSBC
5.00
1.00
1
>99.9
92.6
EWSBC ¼ 102:207 þ 6:496 pH 0:026 D þ 0:134 T 8:123 pH2
GSBC
5.00
1.60
1
99.8
98.8
EGSBC ¼ 61:7514 1:7802 pH 12:0728 D þ 0:8867 T þ 8:6329 pH2 þ 10:843 D2
GHBC
5.00
1.00
1
95.9
96.0
EGHBC ¼ 94:462 þ 5:319 pH 1:398 D þ 1:374 T 4:935 pH2 2:648 D2 þ 3:145 T 2
PSBC
5.00
8.0
18
99.8
99.9
EPSBC ¼ 88:191 þ 18:902 pH þ 28:262 D þ 6:646 T 11:820 pH2 12:073 D2 5:050 T 2
R2adj – determination coefficient.
ENSBC ¼ 92:721 þ 09:841 pH þ 24:925 D þ 3:597 T 12:763 pH2 10:316 D2 12:451 pH D
446
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
Fig. 1. Effect of initial metal concentration (a/b) and biochar pH (c) on Cd(II) and Pb(II) removal efficiency.
tively (for all tested biosorbents under previously optimized biochar dose, solution pH and contact time).
(Fig. 2b); due to the pH-buffering effect of all biochar samples (Trakal et al., 2014).
3.2.2. Sorption kinetics and isotherms The kinetics of Cd and Pb sorption on all tested biochars (Fig. A.3) was solved in the previous experimental step by using the RSM approach in order to optimize contact time (Table 2). According to these results, the sorption time was established after 24 h for all biochar samples to ensure the steady state (changes after 1 h were minimal for all BCs; Fig. A.3). Next, according to other previously optimized conditions (dose of applied biochar and initial pH of solution), these parameters were set up equally for all biochars as follows: pH = 5.00 for Cd, Pb and mixture of Cd + Pb; and applied dose of biochar at 1 g L1 (according to well-sorbed samples). The equilibrium sorption isotherms of Cd and Pb including their mixture for each tested biochar (presented in Fig. 2a) fitted the Langmuir (A.2) model better than the Freundlich (A.1) one (according to model efficiency E; Table A.3). The presented Langmuir isotherms (Fig. 2a) significantly varied not only between the two tested metals, but also between the biochars. Furthermore, in case of multi-metal sorption, there was an obvious competition effect of Pb during Cd sorption, whereas the effect of Cd on the Pb sorption was negligible, as expected Xu et al. (2013a). In more detail, the GSBC was effective for sorption of Cd and Pb in quantities of 0.45 and 2.87 mmol g1, respectively (Table A.3; Fig. 2a). On the other hand, the PSBC removed only 0.04 mmol g1 of Cd and 0.11 mmol g1 of Pb, respectively (Table A.3; Fig. 2a). Such difference in metal sorption efficiency is caused by variable biochar characteristics (e.g. pH, CEC, BET surface or functional groups presence) and different metal sorption mechanisms as well. Furthermore, the pH of all biochar suspensions after equilibrium decreased gradually with increasing of metal(s) concentration in the initial solution during sorption
3.2.3. Cd and Pb sorption mechanisms As mentioned before, the sorption efficiency varied not only between the two tested metals, but mostly between all the tested biochars. This could be caused by different sorption mechanisms of each metal on tested biochars. As described by Sohi et al. (2010), Lu et al. (2012), Ahmad et al. (2014), ion exchange, complexation and physical adsorption are responsible for the metal sorption to biochar. Nevertheless, these metal sorption mechanisms vary according to the type of biochar (group of biochars originating from similar materials) as presented in this study. In more detail, the following post-sorption analyses (FTIR, SEM-EDX, XPS and XRD) were implemented to specify these mechanisms of both metals for each tested biochar. First, FTIR spectra showed that the sorption of Cd and Pb induced the peak of CO2 at 1397/1408 cm1 in the GSBC/GHBC 3 shift to 1413 cm1 for Cd (for both biochars) and the shift to 1398/1397 cm1 for Pb, respectively (Fig. A.4). The presence of the CO2 3 peaks only in case of GSBC and GHBC is probably reflecting the higher ash content of these two original materials (Table A.1; Fig. A.2). The shift of these peaks was, therefore, attributed to ion exchange (obvious from XPS analyses; Fig. A.5) or surface precipitation of metal-(hydro)carbonates (confirmed by the XRD analyses; Fig. A.7), which play an important role in the sorption capacity and the removal mechanism of both sorbates (Xu et al., 2013a,b). Additionally, in the studies of Xu et al. (2013a,b) the metal precipitates with PO3 4 have also been reported. Nevertheless, in this study the metal-phosphates have not been found (confirmed by EDX; Fig. A.6). Subsequently, the band for the surface complexation with carboxylic groups (ACOO-metal chelate) shows the following results: (i) the peak of COO at 1571 cm1
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
447
Fig. 2. Cd(II) and Pb(II): (a) sorption isotherms; and (b) pH changes during single- and multi-metal sorption. pHinitial = 5.00; T = 22 ± 1 °C; mBC = 1 g L1; t = 24 h; and c0 = 0.1– 1.5 mM.
in the original WSBC shifted to 1562 cm1 in Cd- and Pb-loaded biochars, respectively; (ii) the peak of the carboxyl at 1571 cm1 in the GSBC/GHBC shifted to 1562/1569 cm1 for Cd and to 1551/ 1564 cm1 for Pb, respectively; and (iii) comparison of pre- and post-sorption COO spectra for NSBC and PSBC, respectively, showed similarity, which provides no evidence of Cd and Pb chela-
tion carboxylic complex on/in these two biochars. In summary, these results did not show a significant shift caused by the Cd and Pb sorption. This explains the complexation with carboxylic groups as a non-significant mechanism in metal sorption. This fact is in agreement with the studies of Inyang et al. (2012) and Lu et al. (2012), but contrasts with the studies of Xu et al. (2013a,b). These
448
L. Trakal et al. / Bioresource Technology 171 (2014) 442–451
differences can be explained by the different temperature during material pyrolysis (200–350 °C studies of Xu et al. (2013a,b)) and (550–600 °C studies of Inyang et al. (2012), Lu et al., (2012); and this study). Furthermore, hydroxyl groups (AOH), responsible for the coordination of a metal d-electron and producing of AO-M bond (Inyang et al., 2011), were not detected in our case. This should be probably caused by the dehydration (cleavage of hydroxyl group) during high temperature of pyrolysis as well. As another approach, XPS analyses were implemented to specify sorption mechanisms using the results of different binding energies of both sorbed metals (Fig. A.5). Analyses were performed on the surface and at approximately 10 nm depth of each metal-loaded biochar. Each tested biochar showed a different quantity of loaded metal (Cd and Pb) on the surface and at the 10 nm depth, respectively. More precisely, the amount of loaded metals decreased in the following order (Fig. A.5a,c): (i) GHBC > WSBC GSBC NSBC > PSBC for Cd on the surface; (ii) GHBC > GSBC > WSBC PSBC > NSBC for Cd at 10 nm depth; and (iii) GSBC > GHBC > WSBC PSBC > NSBC for Pb both on the surface and 10 nm depth. Furthermore, from the results of the two different positions (surface vs. 10 nm depth) it is obvious that the amount of sorbed metals decreased inwards for (Cd-, PbNSBC; Cd-, Pb-PSBC; and Cd-WSBC) while it decreased outwards for (Cd-, Pb-GSBC; Cd-, Pb-GHBC; and Pb-WSBC). These results suggest different sorption mechanisms, where NSBC and PSBC with well-developed structure sorbed both metals predominantly on the surface. By contrast, in the other three biochars the metals are not sorbed solely on the surface, but are also bonded inside the biochar structure (at higher concentrations). In more detail, Fig. A.5b shows exact sorption mechanisms of each metal on/in studied biochars. The weak p-bindings of the two metals with poly-organic chains (bonding with electron-rich domains on graphene-like structures; Harvey et al., 2011) demonstrated physical adsorption (Lu et al., 2012). This finding is supported by the binding energy results which were 139 and 406 eV for Pb and Cd, respectively and were detected in all tested biochars. The ion exchange as another significant sorption mechanism, was detected in all tested biochars as well (mainly inside the biochar structure; Fig. A.5b). More specifically, such cation release of both metals is limited predominantly to internal positions, whereas the confirmed presence of precipitated metal-(hydro)carbonates were limited solely to the biochar surface (Xu et al. (2013a,b)). Additionally, these binding energies of Pb with the biochar are 137 and 142 eV and 408 eV in case of Cd, respectively. The complexation with carboxylic functional groups (as the third sorption mechanism) is demonstrated in Fig. A.5b by the binding energy at 141 eV for the Pb and by the binding energy at 407 eV for the Cd, respectively. In more detail, this sorption mechanism (metal-chelate creation) is, nevertheless, obvious only in GSBC biochar for both metals and in WSBC solely for Pb. Additionally, these metal-chelates are created not only on the biochar surface, but also inside, which is in agreement with the study of Lu et al. (2012). These results confirmed the previously mentioned fact (by the FTIR analyses), that this chemical process is a non-significant mechanism (limited to GSBC and WSBC) during metal sorption (explained earlier in this Subchapter). Finally, to confirm the different removal efficiency and to see the contrasting structure of all tested biochars during metal sorption, the metal loaded biochars were also scanned by SEM and selected areas were also analysed by EDX (Fig. A.6). The SEM images show very different surface morphology, which reflects the results of BET surface and volume of micro-pores as well (Table 1). Fig. A.6 shows the surface morphology of all biochars at lower (Fig. A.6a) and higher (Fig. A.6b) resolutions. The ‘‘high lignin biomass/biochars’’ (NSBC an PSBC) is showing well-developed structure with a high macro- and micro-porosity, as also shown
by the studies of Chen et al. (2011) and Mohan et al. (2014b). On the other hand, poorly-developed structure of GSBC and GHBC shows limited presence of these pores. This is in agreement with the lower BET surface and lower volume of micro-pores in these cases. Additionally, the structure of the WSBC is very heterogeneous. The surface is poorly-developed (Fig. A.6), nevertheless the structure of this biochar is microporous (Fig. A.8) and, therefore, the value of the BET surface of WSBC is also high. Moreover, the morphology of biochar is not a crucial factor which would affect the metal sorption efficiency (discussed before). This fact is supported by this study, where the biochars with poorly-developed structure (such as GSBS and GHBC) are very efficient in removing both the studied metals from aqueous solution. This metal loading is also confirmed by the EDX analyses coupled to the SEM images. As visible from Fig. A.6, the ‘‘high lignin’’ NSBC and PSBC loaded both metals in very low amounts (
