Selected heavy metal biosorption by compost of Myriophyllum spicatum—A chemometric approach

Ecological Engineering 93 (2016) 112–119

Contents lists available at ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Selected heavy metal biosorption by compost of Myriophyllum spicatum—A chemometric approach Jelena Milojkovic´ a,∗ , Lato Pezo b , Mirjana Stojanovic´ a , Marija Mihajlovic´ a , Zorica Lopiˇcic´ a , Jelena Petrovic´ a , Marija Stanojevic´ a , Milan Kragovic´ a a b

Institute for Technology of Nuclear and Other Mineral Raw Materials, 86 Franchet d’Esperey St. Belgrade, Serbia Institute of General and Physical Chemistry, University of Belgrade, Studentski Trg 12 − 16, 11000 Belgrade, Serbia

a r t i c l e

i n f o

Article history: Received 7 September 2015 Received in revised form 3 March 2016 Accepted 4 May 2016 Keywords: Heavy metal removal Compost Myriophyllum spicatum Chemometric analysis

a b s t r a c t In this study adsorption characteristics of lead, copper, cadmium, nickel and zinc ions onto the compost of Myriophyllum spicatum were examined. The effects of sorbent dose, duration of sorption and solution concentration on the sorption of heavy metals have been investigated. Scanning electron microscope (SEM) and thermogravimetric and differential thermal analysis (TG-DTA) were used for the characterization of this biosorbent. Low coefficients of variation have been obtained for each applied assay, which confirmed the high accuracy of measurements. Principal component analysis (PCA) was applied for differentiation of samples. Mathematical models (form of second order polynomials) were developed for prediction of adsorption. Score analysis is being useful for accessing the effect of process parameters and the tool for determination of sorption quality. On the basic of experimental results and model parameters, it can be concluded that compost has a high biosorption capacity can be utilized for the removal of selected metals from wastewater. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Biosorption was proven to be cost-effective and eco-friendly technology, which engages the use of biological materials for the treatment of wastewater (Kiran and Thanasekaran, 2011). Application of efficient natural materials is more cost effective than artificial materials (Turan and Altundo˘gan, 2014). Different biosorbents (raw or modified) were tested for the removal of various pollutants. Most studies of biosorption were primarily focused on heavy metal and dye pollutants (Anastopoulos and Kyzas, 2015). Amongst the various technologies for removal of toxic metals from wastewaters, it really represents an inexpensive alternative, because of the application of low-cost materials as sorbents (Veglio et al., 1998). Majority of the research in the biosorption of heavy metals refers to the removal of divalent cations (Michalak et al., 2013). Divalent heavy metal cations are widespread in ground and surface waters, soils and sediments due to human activity. Furthermore, like divalent cations, heavy metals can easily enter in the food chain, producing different toxic effects on living organisms (Smiciklas

∗ Corresponding author. E-mail addresses: [email protected], [email protected] ´ (J. Milojkovic). http://dx.doi.org/10.1016/j.ecoleng.2016.05.012 0925-8574/© 2016 Elsevier B.V. All rights reserved.

et al., 2008). High solubility of heavy metals in the aquatic surroundings allows their adoption by living organisms (Babel and Kurniawan, 2004). The most important factors which affect to heavy metal mobility, toxicity, and reactivity are: pH, sorbent nature, Eh, temperature, presence and concentration of organic and inorganic ligands, etc. (Tessier et al., 1979). Chemical speciation of metal is determined by solution pH. For instance, lead is present as Pb(II) as dominant species at pH < 5.5 (Farooq et al., 2010). Metal species of selected heavy metals (Pb, Cu, Cd, Ni and Zn) are in the +2 oxidation states in aqueous solution where pH is around 5.0 (Vieira et al., 2012). The solubility of heavy metals determines their toxicity. The metals are more toxic at lower pH values, because then their solubility increases (Beˇcelic´ and Tamaˇs, 2004). Taking into account heavy metal mobility, toxicity, and reactivity for this study Pb(II), Cu(II), Cd(II), Ni(II) and Zn(II) were selected. Among different biosorbents, the researchers consider on alternative application of composts. It is well known that composts are mainly used as amendments to increase soil fertility (Anastopoulos and Kyzas, 2015). There is a constant increase in the number of papers in which compost is used as biosorbent of pollutants. Compost of M. spicatum can be successfully applied as biosorbent for Pb(II) Milojkovic´ et al. (2014a) and selected heavy metals (Pb(II), Cu(II), Cd(II), Ni(II) and Zn(II)) (Milojkovic´ et al., 2014b).

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119

In this study, effect of sorbent dose, duration of sorption and solution concentration on sorption of selected heavy metals by compost of Myriophyllum spicatum were invesigated. Also, objectives of this study were to examine thermal stability with thermal analysis (TG-DTA) and reveal the changes in morphology after biosorption by scanning electron microscope (SEM). Principal Component Analysis (PCA) was used to discriminate different samples, processed under various process parameters. Simple regression models (second order polynomials − SOP) have been proposed for calculation of heavy metals sorption capabilities as function of proposed process parameters. In order to enable more comprehensive comparison between investigated samples, particularly the contribution of process parameters, standard score (SS), assigning equal weight to all assays applied, has been introduced. Analysis of variance (ANOVA) has been applied to show relations between applied assays.

Table 1 ANOVA calculation for biosorption capacities of selected heavy metals.

m m2 C0 C0 2 t t2 Error r2

qPb (mmol/g)

qCu (mmol/g)

qCd (mmol/g)

qNi (mmol/g)

qZn (mmol/g)

0.000141* 0.000004 0.009427* 0.000540* 0.000135* 0.000269* 0.000283 0.976

0.000032 0.000000 0.006054* 0.000485* 0.000087* 0.000112* 0.000264 0.969

0.000040* 0.000013 0.000170* 0.000042* 0.000092* 0.000036* 0.000059 0.859

0.000028 0.000005 0.001165* 0.000373* 0.000018 0.000045 0.000193 0.883

0.000042* 0.000009 0.000905* 0.000300* 0.000034* 0.000020* 0.000048 0.960

* Significant at p < 0.05 level, 95% confidence limit, error terms were found statistically insignificant.

solution (L) and M is the mass of the sorbent (g). Metal removal efficiency (R) is calculated from Eq. (2): R=

2. Materials and methods

113

Ci − Ce × 100 Ci

(2)

2.4. Biosorbent characterization

2.1. Preparation of biosorbent M. spicatum is harvested from artificial Sava Lake every year. Harvested aquatic weed (around 35 m3 per day) is disposing to the open landfill used just for that purpose. Samples of compost were taken from the surface of the landfill (1 year old). The preparation of compost was previously described in detail (Milojkovic´ et al., 2014b). The prepared compost was exposed to air and dried for a couple days at room temperature and then dried at 60 ◦ C for 6 h, crushed and sieved to give a particle size less than 0.2 mm.

2.4.1. Thermal analysis Thermal analysis of the samples was performed on a Netzsch STA 409 EP. Samples of compost were heated (20–1000 ◦ C) in an air atmosphere with a heating rate of 10 ◦ C/min. The samples were kept in a desiccator at a relative humidity of 23%, prior to analyses. 2.4.2. Scanning electron microscopy (SEM) In order to directly observe the surface morphology, Scanning electron microscope SEM JEOL JSM-6610LV model, was utilized in this study. Samples of compost were coated under vacuum with a thin layer of gold and then examined.

2.2. Reagents The heavy metal sorbates used in this study were: Pb(NO3 )2 , Cu(NO3 )2 ·3H2 O, Cd(NO3 )2 ·4H2 O, Ni(NO3 )2 ·6H2 O and Zn(NO3 )2 ·4H2 O. Stock metal solutions (10 mmol/L each metal) were prepared by dissolving above mentioned metal salts (analytical grade) in deionised water. The working solutions were obtained by diluting the stock solution. 2.3. Batch biosorption experiments Each experiment was conducted in 100 ml Erlenmeyer flasks containing 50 ml of multimetal solution. The flasks containing multimetal solutions and compost were agitated on orbital shaker Heidolph unimax 1010 at 250 r/min. pH value was regulated to the appropriate value with 0.1 M HNO3 or 0.1 M NaOH (analytical grade). Measurement of pH value was carried out with a precise pH meter (Sension MM340). Equilibrium studies were performed using different initial concentration of each metal ion (0.2 − 5 mmol/L) at respective optimum solution pH of 5.0. Kinetic of selected heavy metals biosorption by compost of M. spicatum was studied by varying the contact time from 10 to 720 min remaining other conditions constant (initial concentration 2.5 mmol/L, pH was around 5.0, biosorbent dose 1.25 g in 50 ml). The concentration of heavy metal ions in solutions was determined by Atomic absorption spectrophotometer (Perkin Elmer AAnalyst 300). The amount of metal adsorbed by the compost was calculated using Eq. (1): q=

V (Ci − Ce ) M

2.5. Statistical analyses The experimental data used for the study of experimental results were obtained with three sets of experiment in which only one process parameter was variable (sorbent dose, duration of sorption, solution concentration on sorption of heavy metals), while the other two were constant (Table 1). These experiments were performed to test the sorption quantity of heavy metals (qPb , qCd , qCu , qNi and qZn ), considering these three factors (Brlek et al., 2013; Madamba, 2002; Montgomery, 1984). Descriptive statistical analyses of all the obtained results were expressed as the mean ± standard deviation (SD). The evaluation of ANOVA of the obtained results was performed using Statistica software version 12 (STATISTICA, 2012). 2.6. Principal component analysis (PCA) The algorithm of PCA can be found in standard chemometric material (Otto, 1999; Kaiser and Rice, 1974). In summary, PCA decomposes the original matrix into several products of multiplication into loading (different samples) and score (measured assays) matrices. The different samples were taken as variables (column of the input matrix) and measured data of qPb , qCu , qCd , qNi and qZn as mathematical-statistical cases (rows of the matrix). The number of factors retained in the model for proper classification of measuring data, in original matrix into loading and score matrices were determined by application of Kaiser and Rice’s rule. This criterion retains only principal components with eigenvalues>1.

(1)

where sorption values q is the amount of metal adsorbed by biosorbent at any time (mmol/g), Ci and Ce the initial and equilibrium metal concentrations (mmol/L), V the volume of multimetal

2.7. Determination of normalized standard scores (SS) A standard score is one of the most widely used technique to compare various characteristics of various samples determined

114

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119

using multiple measurements, where samples are ranked based on the ratio of raw data and extreme values of the measurement used. Since the scale (and sometimes the units) of the data acquired from various samples, using measuring methods are different, the data in each dataset should be transformed into normalized scores, dimensionless quantity derived by subtracting the minimum value from the raw data, and divided by the subtraction of maximum and minimum value, according to the following equation (Brlek et al., 2013): maxxi − xi xi = 1 −

i

maxxi − minxi i

, ∀i

(3)

i

where xi represents the raw data. The normalized scores of a sample of different measurements when averaged give a single unitless value termed as SS, which is a specific combination of data from different measuring methods with no unit limitation. This approach also enables the ease of employing some others set of data to this elaboration. Standard scores for different samples investigated in this article were calculated and written in Table 1. 3. Results and discussion 3.1. Characterization compost of M. spicatum 3.1.1. Thermal analysis Thermogravimetric (TG) analysis, Derivative thermogravimetry analysis (DTG), Differential thermal analysis (DTA) The chemical composition of M. spicatum compost showed that this material contains 66.85% neutral detergent fiber (NDF): cellulose, hemicellulose and lignin. Microbiological degradation of this aquatic weeds leads to an increase of aromatic compaunds by forming fulvic 3.44% and humic 0.38% acids which can be noted in chemical composition of M. spicatum compost (Milojkovic´ et al., 2014a). Thermal decomposition of a lignocellulosic material may be a superposition of the decomposition behaviour of its individual components. Carbohydrate polymers (cellulose, hemicellulose), break down faster and provide volatile products. Aromatic polymer (lignin) goes through a slow charring process (Varhegyi et al., 1989). The TG, DTG and DTA curves, which display the thermal degradation characteristics for the compost M. spicatum before and after the biosorption of selected heavy metals, are presented on Fig. 1. TG diagram (Fig. 1a) of compost before and after biosorption, showed non-continuous weight loss. The greatest loss of mass more than 10% is in the fourth region of the temperature 600–900 ◦ C. Total mass loss in the temperature range 25–900 ◦ C is slightly higher in compost after biosorption 16.08% compared to untreated compost 15.21%. On Fig. 1b are presented results of differential thermogravimetric analisis DTG. The thermal decomposition of compost M. spicatum exhibited DTG peaks at 85, 323, 500, 600, 804 ◦ C. DTG endo peak, that occurs in the second temperature range, 150–375 ◦ C, may be pointed to the decomposition of cellulose and hemicellulose, and in the third temperature range is clear lignin peak at 500 ◦ C, organic compounds, which are very present in the compost. The tested compost contains: 20.23% cellulose, 1.72% hemicellulose and 43.20% lignin (Milojkovic´ et al., 2014a). Thermal stability of the compost could be modified with humic-like colloids formed in the process of composting. Also on the DTG curves of the combustion process, different inorganic salts have an effect. Salts like ammonium carbonate, sodium bicarbonate or calcium carbonate are able to transfer peaks toward lower temperatures (Blanco and Almendros, 1994). On the DTG curve after the biosorption of heavy metal ions there is a small shifts of the peaks which are from loss of water and combustion cellulose and hemicellulose. However, after binding of metal ions to compost there is a shift of the peak from

500 to 551 ◦ C, which may indicate that the content of inorganic salts in the compost reduced after biosorption. After biosorption, exothermic maximums and endothermic minimums of DTA curve are moved (Fig. 1c). This obtained data are similar to the data presented by Som et al. (2009). DTA curve of compost show endothermic peak with minimum at 111 ◦ C (Fig. 1c) which is located in the first region of the temperature (25–150 ◦ C). The water release profile was found in that region (Nikfarjam et al., 2015). The weight loss of 0.93% of the compost is mostly related to the loss of water. The following two exothermic phenomena are observed in the range 200–550 ◦ C, and they correspond to the oxidation of organic compounds. Oxidation takes place in 2 stages. The first exothermic peak between 200 and 375 ◦ C (336 ◦ C), corresponding to decomposition of carbohydrates, cellulosic and lignocellulosic substance (Otero et al., 2002). The second exothermic peak between 400 and 550 ◦ C (511 ◦ C), is associated with the degradation of complex aromatic structure (Geyer et al., 2000; Peuravuori et al., 1999). Endothermic peak with minimum at 827 ◦ C indicates complex oxidation of thermostable carbon compounds, and the degradation of the mineral and mineral salts, such as carbonates (Atanasow and Rustschev, 1985; Baffi et al., 2007). After biosorption peaks had no significant shifts, except for a significant shift of the peak (from 511 ◦ C to 425 ◦ C). Observed change may indicate the binding of metal ions on aromatic structure of compost that originates from the presence of lignin, humic and fulvic acids. 3.2. Biosorption mechanisms Diferent instrumental techniques were used to explain biosorption mechanisms of binding selected metal ions on compost. Surface of biosorbent was characterized by scanning electron microscopy (SEM). Thermal analysis (TG/DTG/DTA) (Fig. 1) showed possible binding of metal ions on aromatic structure of compost (lignin, humic and fulvic acids) and that the content of inorganic salts in the compost reduced after biosorption. The previously obtained results of FTIR instrumental technique Milojkovic´ et al. (2014b), showed that carbonyl, carboxyl, hydroxyl and phenyl groups are main binding sites for those heavy metal ions. Also in that investigations, EDS analysis showed that ion exchange between divalent cations Ca(II) and selected metals takes place. Finally, based on results gained with all mentioned instrumental techniques, presumed mechanism of biosorption of selected metals on M. spicatum compost which include ion exchange and complexing is presented on Fig. 2. Based on chemical composition, calcium is present with the highest percentage of all metals, both in the plant M. spicatum (5%) and in compost (30%) (Milojkovic´ et al., 2014a). Such a large proportion of calcium can be explained by the fact that leaves of submerged plants can be covered with a whitish scum which originated from participated calcium carbonate. During the process of photosynthesis, aquatic plants are supplied, not only with free carbon dioxide, but also with one from aqueous solution of calcium bicarbonate. After that, the following phenomena occur like precipitation of insoluble calcium carbonate and its deposition on the ´ 2001). surface of submerged leaves (Stevanovic´ and Jankovic, The precipitation of calcite in natural waters can be summarized by the following reaction: Ca2+ +2HCO− 3 ⇔ CaCO3 ↓ +CO2 ↑ +H2 O

(4)

To saturation of the solution takes primarily by removing carbon dioxide from the aqueous solution, so that the equilibrium hydrogen carbonate moves in the direction of the formation of carbonate. Water from the Sava Lake is moderately alkaline (pH = 8.2 to 8.8) and that pH value favors the formation of CaCO3 . The harvested M.

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119

115

Fig. 1. TG (a), DTG (b) and DTA (c) curves compost M. spicatum before and after the biosorption of selected heavy metals.

Fig. 2. Possible mechanisms of biosorption of selected heavy metal ions − Me2+ with compost of aquatic weed M. spicatum.

spicatum is disposed of in the open landfill where decomposition of organic matter leads to concentration of CaCO3 in the compost.

SEM micrographs of a surface of compost (1000 times magnified) are presented on Fig. 2. Compost particles are irregular in

116

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119

shape and their surfaces are diverse multi-layered and lumpy. On the surface of the material metal aggregates are not present, so there isn’t visible microprecipitation. Metals − Me2+ (Pb(II), Cu(II), Cd(II), Ni(II) and Zn(II)) are uniformly distributed over the surface of compost. The release of calcium, primarily fixed onto the compost of M. spicatum (CMS) has been simultaneously accompanied with adsorption of metals. Similar observation was reported in findings of sorption mechanisms (Ahmady-Asbchin et al., 2008). This release depends on the initial metal concentration of the solution. Ion exchange as a surface reaction happened under certain conditions, where ions are attracted to a solid surface and may be exchanged with other ions in an aqueous solution. Cation exchange is the dominant process in some natural material such as compost. In case that metal bonding onto CMS takes place by ionic exchange, the involved reaction could be:



CMS − Ca2+ + Me2+



K Me

solution

Selectivity constant K Me K Me

2+ /Ca2+

=

CMS − Me2+ CMS − Ca2+

2+

/Ca2+

↔

2+ /Ca2+



CMS − Me2+ + Ca2+

 solution

(5)

Fig. 3. Biplot of different samples of heavy metals biosorption by compost of Myriophyllum spicatum.

can be formulated as:

 2+  Ca  × Me2+

(6)

If the constant is high, the equilibrium changes to the right side, implying that the cation Ca2+ is easily released or that the ion of metal Me2+ is better bound (Ahmady-Asbchin et al., 2008). 3.3. ANOVA and RSM analysis The amount of metal adsorbent under different processing conditions is presented in Table TS1 in Supplementary material, and statistically significant differences in quantity of sorption data were found in almost all samples. As predicted, the amount of metal adsorbed by the compost increased with longer duration of sorption and larger solution concentration, while the increasing of sorbent dose pronouns the decrease of metal adsorption. Investigated samples are characterized by a relatively high sorption quantity for heavy metals, and the largest values of qPb , qCu , qCd , qNi and qZn (0.1126, 0.0976, 0.0229, 0.0477 and 0.0424 (mmol/g), respectively) were observed for m = 1.25 g, C0 = 5 mmol/L and t = 120 min, which leads to conclusion that C0 is the most influential variable for optimizing the adsorption process. Standard scores for the evaluation of sorption quality under different processing conditions with t (1–24 h), m (0.5-1.5 g) and Co (0.2–5 mmol/L) have been calculated and written in Table 1. As seen, qPb , qCd , qCu , qNi and qZn strongly influence the final score result. Best scores have been obtained for sample processed under processing parameters: m = 1.25 g; Co = 5 mmol/L and t = 2 h (SS = 1.00). 3.4. Principal component analysis (PCA) The PCA allows a considerable reduction in a number of variables and the detection of structure in the relationship between measuring parameters and different varieties of processing parameters that give complimentary information (Otto, 1999; Kaiser and Rice, 1974). All samples have different m, C0 and t, as predicted by PCA score plot (Fig. 3). The full autoscaled data matrix consisting of measured values of qPb , qCu , qCd , qNi and qZn are submitted to the PCA. For visualizing the data trends and the discriminating efficiency of the used descriptors a scatter plot of samples using the first two principal components (PCs) issued from PCA of the data matrix is obtained (Fig. 3). As can be seen, there is a neat separation of the 22 samples with differentiation of processing parameters, according

to m, C0 and t. Quality results show that the first two principal components, accounting for 94.83% of the total variability can be considered sufficient for data representation. The variables qPb (which contributed 19.76% of the total variance, calculated based on the correlation), qCu (15.87%), qCd (17.52%), qNi (23.22%) and qZn (18.04%) negatively influenced the first principal component. The variables qPb and qCu showed the positive influence on second principal component calculation (showing 22.50% and 39.86% of the total variance, respectively), while qCd showed the negative impact on second principal component calculation (with 23.06% total variance explained). The points shown in the PCA graphics, which are geometrically close to each other indicate the similarity of patterns that represent these points. The orientation of the vector describing the variable in factor space indicates an increasing trend of these variables, and the length of the vector is proportional to the square of the correlation values between the fitting value for the variable and the variable itself. The angles between corresponding variables indicate the degree of their correlations (small angles corresponding to high correlations). The influence of different parameters that describes the observed samples could be evaluated from the scatter plot, Fig. 3, in which the samples with higher qPb , qCu , qCd , qNi and qZn values are located at the left side of the graphic (samples 21 and 22), showing the best biosorption quality of the observed samples. Analysis of variance and the following post-hoc Tukey’s HSD test were evaluated for comparison of heavy metal sorbent characteristics for different process parameters. According to ANOVA (Table 1), all response variables (qPb , qCu , qCd , qNi and qZn ) are mostly affected by a linear term of C0 , in SOP model, statistically significant (for all assays), at p < 0.01 level. The quadratic term of C0 is also very influential (p < 0.01). The linear and the quadratic term for t is very influential for qPb , qCu , qCd and qZn calculation. The linear term of m in SOP models for qPb , qCd and qZn were also important for calculation. The other non/linear and interchange terms were found statistically insignificant or negligible. The average error between the predicted values and experimental values were below 10%. Values of average error below 10% indicate an adequate fit for practical purposes. To verify the significance of the models, analysis of variance was conducted and the results indicate that all models were significant with minor lack of fit, suggesting they adequately represented the relationship between responses and factors. All SOP models had an insignificant lack of fit tests, which means that all the models represented the data satisfactorily. The

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119

Fig. 4. Observed responses qPb , qCd , qCu , qNi and qZn , affected by the sorbent dose, duration of sorption and solution concentration.

117

118

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119

Table 2 Separation factor or equilibrium parameter RL values for the adsorption of Pb(II), Cu(II), Cd(II), Ni(II), Zn(II) onto compost. Initial concentration (mmol/L)

Pb Cu Cd Ni Zn

0.2

0.4

1

1.5

2

2.5

3

4

5

0.301 0.628 0.0504 0.0159 0.991

0.169 0.467 0.0310 0.00801 0.982

0.0794 0.284 0.0115 0.00309 0.956

0.0574 0.196 0.00849 0.00200 0.934

0.0434 0.161 0.00583 0.00155 0.920

0.0395 0.132 0.00462 0.00116 0.888

0.0306 0.110 0.00388 0.00103 0.871

0.0222 0.0878 0.00303 0.000619 0.838

0.0175 0.0689 0.00248 0.000571 0.806

coefficient of determination, r2 , is defined as the ratio of the explained variation to the total variation and is explained by its magnitude. A high r2 is indicative that the variation was accounted and that the data fitted satisfactorily to the proposed model (SOP in this case). The r2 values for qPb , qCu , qCd , qNi and qZn (0.976, 0.969, 0.859, 0.883 and 0.960 mmol/g, respectively), were found very satisfactory and showed the good fitting of the model to experimental results. The three-dimensional graphic have been plotted for experiment data visualization (white colored points) and for the purpose of observation the fitting of regression models (qPb , qCu , qCd , qNi and qZn ) to experimental data, Fig. 4. All plots showed “rising ridge” configuration. Experienced values raised, with the increase of C0 and t, while the increase of m leads to lower adsorption of metals (Fig. 4). Like in outcomes of Podstawczyk et al. (2015) models were developed according to experimental results on the basis of multimetal-containing wastewater synthetic solutions, while the experiments were performed to best reflect the real ecological situation. The duration of sorption, the solution concentration, and the sorbent dose are the most important parameters in multimetalcontaining wastewater treatment. The appropriate design, scale-up and optimization of biosorption process in the industrial scale can be realized by the predictive mathematical model. The models proposed in this study predict the efficiency of biosorption in batch mode with high accuracy for varying operational conditions characteristic for industrial multimetal-containing wastewater, thus they have potential applicability in wastewater industry. Industrial implementation of the models can improve process monitoring and controlling and in turn save time and reduce costs. 3.5. Separation factor The comparison of maximum biosorption capacities of selected five metals demonstrated that Pb(II) has the highest sorption capacity. From Table 2 it can be seen that lead has the highest value of KL which showed that compost has higher affinity for Pb(II). While 1/n is in the range of 0.1<(1/n)<1, it characterizes a heterogeneous surface structure for the adsorbent with an exponential distribution of the energy of the surface active sites (Oo et al., 2009): RL =

1 1 + KL Ci

(7)

The value of RL provides information about the adsorption and it is irreversible (RL = 0), favourable (0 < RL < 1), linear favourable (RL = 1) or unfavourable (RL > 1). According to the value of equilibrium parameter RL presented in Table 2 values between 0 and 1 imply favourable adsorption. 4. Conclusion The biosorption of Pb(II), Cu(II), Cd(II), Ni(II) and Zn(II) onto M. spicatum compost was investigated in multimetal aqueous solutions. Scanning electron microscope (SEM) confirmed that the surface morphology of compost changed apparently after metal binding. It can be concluded that the usefulness of sorbent mate-

rial under different processing conditions with sorbent weight (0.5–1.5 g), duration of sorption (10–720 min) and solution concentration (0.2–5 mmol/L) exerted better results for larger sorbent weights, longer process duration and increased solution concentrations. The quality of sorption is affected by sorbent dose and solution concentration the most, which is confirmed by ANOVA calculation and standard score evaluation. The developed models for prediction of qPb , qCu , qCd , qNi and qZn show high r2 values (0.976, 0.969, 0.859, 0.883 and 0.960, respectively), which can be considered very satisfactory and the good fit of the model to experimental results can be expected. Principal component analysis enabled better visualization of discrimination and differentiation among the samples. The results showed that preference of compost was following the order Pb(II) > Cu(II) > Cd(II) > Zn(II) > Ni(II) at optimal conditions of pH 5.0. The presence of other metal ions reduces the biosorption capacity metal so that the capacity is less than when the individual metals are present in solution. This experiment showed that M. spicatum compost was an appropriate biosorbent for removal of heavy metals from wastewater because it is natural, low-cost and abundantly available. Acknowledgements These results are part of the projects supported by the Ministry of Education and Science of the Republic of Serbia, TR 31003, TR 31055 and TR34013. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ecoleng.2016.05. 012. References Ahmady-Asbchin, S., Andrès, Y., Gérente, C., Cloirec, P.L., 2008. Biosorption of Cu(II) from aqueous solution by Fucus serratus: surface characterization and sorption mechanisms. Bioresour. Technol. 99, 6150–6155. Anastopoulos, I., Kyzas, G.Z., 2015. Composts as biosorbents for decontamination of various pollutants: a review. Water Air Soil Pollut. 226, 61. Atanasow, O., Rustschev, D., 1985. Thermal analysis of peat and peat soils. Thermochim. Acta 90, 373–377. Babel, S., Kurniawan, T.A., 2004. Cr(VI) removal from synthetic wastewater using coconut shell charcoal and commercial activated carbon modified with oxidizing agents and/or chitosan. Chemosphere 54 (7), 951–967. Baffi, C., Dell’Abate, M.T., Nassisi, A., Silva, S., Benedetti, A., Genevini, P.L., Adani, F., 2007. Determination of biological stability in compost: a comparison of methodologies. Soil Biol. Biochem. 39 (6), 1284–1293. ´ M., Tamaˇs, Z., 2004. Analiza i kontrola kvaliteta fiziˇcko-hemijskih Beˇcelic, parametara. In: Dalmacija, B., Ivanˇcev-Tumbas, I. (Eds.), Analiza Vode − Kontrola Kvaliteta, Tumaˇcenje Rezultata. Prirodno-matematiˇcki Fakultet. Departman za hemiju, Novi Sad, pp. 51–79. Blanco, M.J., Almendros, G., 1994. Maturity assessment of wheat straw composts by thermogravimetric analysis. J. Agr. Food Chem. 42, 2454–2459. ˇ ´ N., Kriˇcka, T., Vukmirovic, ´ Ð., Colovi ´ R., Bodroˇza-Solarov, Brlek, T., Pezo, L., Voca, c, M., 2013. Chemometric approach for assessing the quality of olive cake pellets. Fuel Process. Technol. 116, 250. Farooq, U., Kozinski, J.A., Khan, M.A., Athar, M., 2010. Biosorption of heavy metal ions using wheat based biosorbents-A review of the recent literature. Bioresour. Technol. 101, 5043–5053.

J. Milojkovi´c et al. / Ecological Engineering 93 (2016) 112–119 Geyer, W., Hemidi, F.A.-H., Brüggemann, L., Hanschmann, G., 2000. Investigation of soil humic substances from different environments using TG–FTIR and multivariate data analysis. Thermochim. Acta 361 (1–3), 139–146. Kaiser, H.F., Rice, J., 1974. Little jiffy mark IV. Educ. Psychol. Meas. 34, 111–117. Kiran, B., Thanasekaran, K., 2011. Copper biosorption on Lyngbya putealis: Application of response surface methodology (RSM). Int. Biodeterior. Biodegrad. 65, 840–845. Madamba, P.S., 2002. The response surface methodology: an application to optimize dehydration operations of selected agricultural crops. LWT-Food Sci. Technol. 35 (7), 584–592. Michalak, I., Chojnacka, K., Witek-Krowiak, A., 2013. State of the art for the biosorption process-a Review. Appl. Biochem. Biotechnol. 170, 1389–1416. ´ J.V., Mihajlovic, ´ M.L., Stojanovic, ´ M.D., Lopiˇcic, ´ Z.R., Petrovic, ´ M.S., Milojkovic, ˇ staric, ´ T.D., Ristic, ´ M.Ð., 2014a. Pb(II) removal from aqueous solution by Soˇ Myriophyllum spicatum and its compost: equilibrium, kinetic and thermodynamic study. J. Chem. Technol. Biotechnol. 89, 662–670. ´ J.V., Stojanovic, ´ M.Ð., Mihajlovic, ´ M.L., Lopiˇcic, ´ Z.R., Petrovic, ´ M.S., Milojkovic, ˇ staric, ´ T.D., Ristic, ´ M.Ð., 2014b. Compost of aquatic weed myriophyllum Soˇ spicatum as low-Cost biosorbent for selected heavy metal ions. Water Air Soil Pollut. 225 (1927). Montgomery, D.C., 1984. Design and Analysis of Experiments. John Wiley and Sons Inc, New York. Nikfarjam, N., Sabzi, M., Deng, Y., Qazvini, N.T., 2015. Water-Expandable polystyrene using cross-Linked starch nanoparticle as water-Stabilizing agent. Ind. Eng. Chem. Res. 54, 6627–6633. Oo, C.W., Kassim, M.J., Pizzi, A., 2009. Characterization and performance of Rhizophora apiculata mangrove polyflavonoid tannins in the adsorption of copper (II) and lead (II). Ind. Crops. Prod. 30, 152–161. Otero, M., Calvo, L.F., Estrada, B., García, A.I., Morán, A., 2002. Thermogravimetry as a technique for establishing the stabilization progress of sludge from wastewater treatment plants. Thermochem. Acta 389, 121–132. Otto, M., 1999. Chemometrics Statistics and Computer Application in Analytical Chemistry. Wiley-VCH, Weinheim.

119

Peuravuori, J., Paaso, N., Pihlaja, K., 1999. Kinetic study of the thermal degradation of lake aquatic humic matter by thermogravimetric analysis. Thermochim. Acta 325, 181–193. Podstawczyk, D., Witek-krowiak, A., Dawiec, A., Bhatnagar, A., 2015. Biosorption of copper (II) ions by flax meal: empirical modeling and process optimization by response surface methodology (RSM) and artificial neural network (ANN) simulation. Ecol. Eng. 83, 364–379. STATISTICA, 2012. (Data Analysis Software System) v. 12., Stat-Soft, Inc., USA (accessed 24. 1. 15.) www.statsoft.com. ´ S., Janackovi ´ ´ D., Mitric, ´ M., 2008. Factors Smiciklas, I., Onjia, A., Raicevic, c, influencing the removal of divalent cations by hydroxyapatite. J. Hazard. Mater. 152, 876–884. Som, M.-P., Lemée, L., Amblès, A., 2009. Stability and maturity of a green waste and biowaste compost assessed on the basis of a molecular study using spectroscopy, thermal analysis, thermodesorption and thermochemolysis. Bioresour. Technol. 100 (19), 4404–4416. ´ B.M., Jankovic, ´ M., 2001. Ekologija Biljaka Sa Osnovama Fizioloˇske Stevanovic, Ekologije Biljaka. NNK international, Belgrade. Tessier, A., Campbell, P.G.C., Bisson, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem. 51, 844–851. Turan, M.D., Altundo˘gan, H.S., 2014. A study on Cr(VI) reduction from aqueous solutions by bauxite. J. Cent. South Univ. 21, 1961–1967. Varhegyi, G., Antal, M.J., Szekely, T., Szabo, P., 1989. Kinetics of the thermal decomposition of cellulose, hemicellulose, and sugarcane bagasse. Energy Fuels 3, 329–335. Veglio, F., Beolchini, F., Toro, L., 1998. Kinetic modeling of copper biosorption by immobilized biomass. Ind. Eng. Chem. Res. 5885, 1107–1111. Vieira, M.G.A., de Almeida Neto, A.F., Carlos da Silva, M.G., Nóbrega, C.C., Melo Filho, A.A., 2012. Characterization and use of in natura and calcined rice husks for biosorption of heavy metals ions from aqueous effluents. Brazilian J. Chem. Eng. 29, 619–633.