Water, Air, & Soil Pollution

, 224:1599

Valuation of Unmodified Rice Husk Waste as an Eco-Friendly Sorbent to Remove Mercury: a Study Using Environmental Realistic Concentrations

Authors

    • Department of Chemistry/CESAMUniversity of Aveiro
  • Cláudia B. Lopes
    • Department of Chemistry/CESAMUniversity of Aveiro
  • J. A. Borges
    • Álvaro Alves Borges, LdaAlto Brenha-Brenha
  • A. C. Duarte
    • Department of Chemistry/CESAMUniversity of Aveiro
  • E. Pereira
    • Department of Chemistry/CESAMUniversity of Aveiro
Article

DOI: 10.1007/s11270-013-1599-9

Cite this article as:
Rocha, L.S., Lopes, C.B., Borges, J.A. et al. Water Air Soil Pollut (2013) 224: 1599. doi:10.1007/s11270-013-1599-9
  • 229 Views

Abstract

The present work explores the sorption capacity of an inexpensive and highly available agricultural waste, rice husk, to remove mercury using realistic concentrations of this metal. The efficiency of the process was evaluated for two initial Hg(II) concentrations, one representing the maximum value for Hg discharges from industrial sectors (0.05 mg L−1), and the other ten times higher. A very small amount of rice husk (0.25 and 0.50 g L−1) was able to reduce the Hg(II) levels in more than 80 % for an initial concentration of 0.05 mg L−1 and in more than 90 % for 0.50 mg L−1, corresponding to residual concentrations of Hg(II) of 0.048 and 0.009 mg L−1, respectively. The biosorvent was reused in further cleaning treatments, maintaining the efficiency and high performance. The sorption kinetics of the Hg–rice husk system is well fitted by the Elovich model and the diffusion models suggested that, depending on the initial Hg(II) concentrations, the sorption process can be controlled by intraparticle diffusion or by both film and intraparticle diffusion. The equilibrium data are well described by the linear isotherm and the distribution coefficient found was 36.1 L g−1.

Keywords

MercuryRice huskWasteRemediationContaminated waters

1 Introduction

The presence of metals in aquatic systems has become a serious threat, of environmental extreme concern, due to the non-biodegradable and persistent character of these elements (Krishnani et al. 2008). Mercury (Hg) is regarded as one of the most harmful metals found in the environment not only due to its high toxicity, volatility, and persistence both in the environment and biota, but also because Hg bioaccumulates in living organisms and bioamplifies along the food chain (Coelho et al. 2008). Moreover, according to the European Parliament and the Council of European Union, Hg (and its compounds) is considered a priority hazardous substance (EU 2008), and its emissions, discharges, or losses should be phased out.

The nature and magnitude of the environmental problems are frequently changing and with legal constraints being imposed, there is a constant need for developing newer and more appropriate technologies for water treatment. In this context, biosorption, an eco-friendly and cost-effective treatment technology presents a tremendous potential to provide the needs and holds hope for environmental protection, sustainability, and management (Krishnani et al. 2008; Juwarkar et al. 2010). In the last two decades, efficient and cost-effective approaches have been developed to remove Hg(II) and other metal ions from wastewaters based on the use of non-conventional sorbents. These materials are available in nature or are either a by-product or waste material from the industry, and since they require less prior processing and are available in large quantities, they can be classified as low-cost sorbents. As a result, these materials could represent an excellent alternative not only to expensive treatment processes (Meena et al. 2004), but also to effective but costly materials, such as activated carbon and titano- and zirconico-silicates (Krishnani et al. 2008; Lopes et al. 2008; Lopes et al. 2009).

Rice husk (RH) is a by-product of rice processing and it comprises 23 % of the rice grain, being considered a significant waste disposal problem (Jeon, 2011). This material, like all vegetable biomasses, is composed essentially of cellulose (29 to 34 %), hemicellulose (21 to 29 %), and lignin (19 to 30 %) (Ahmaruzzaman and Gupta, 2011). The presence of carbon and silica in the rice husk composition and the existence of functional groups with binding capabilities (such as carboxyl and silanol) on its surface makes the sorption processes possible, though it has been reported that this material was used to remove various pollutants from waters and wastewaters (Ahmaruzzaman and Gupta, 2011; Jeon, 2011).

Despite the fact that the potential of rice-husk-based sorbents to remove Hg(II) from waters has already been investigated by some authors, the studies developed in this field used initial concentrations of Hg(II) considerably high, i.e., between 8 to 2,000 mg L−1 (Tiwari et al. 1995; Khalid et al. 1999; Feng et al. 2004; Krishnani et al. 2008; El-Said et al. 2010; El-Shafey 2010), values that largely exceed the values found in contaminated environments. As a result, the removal capacity of rice-husk-based materials in those studies is being overestimated and even when high removal values were achieved, the residual concentration of Hg(II) at the end of the process, largely exceeds (at least ten times) the guideline limits for effluents discharges (0.05 mg L−1, Council Directive 84/156/EEC). For those reasons, it is very important to evaluate the sorption capacity of rice husk toward Hg(II), for concentrations close to the ones found in contaminated environments, so that, the real value of this material as a cleaning agent is recognized. Besides that, it should also be highlighted that in most of the studies, the authors performed several pre-treatments to the sorbent and, frequently the physical (calcination) and/or chemical (using acidic, alkaline, among other solutions) modified forms of rice husk are used (Tiwari et al. 1995; Feng et al. 2004; Krishnani et al. 2008; El-Said et al. 2010; El-Shafey 2010). Although the modification of the rice husk usually increases the sorption efficiency of the material, it is also time consuming, increases the cost of the material and consequently of the cleaning process, and also wipes out one of the initial concepts behind biosorption, the revaluation of agricultural wastes. Additionally, in those studies, a substantial amount of rice husk per volume of solution (m/VRH) is usually used (between 2 and 100 g L−1), which may cause secondary problems of waste disposal (Tiwari et al. 1995; Feng et al. 2004; Krishnani et al. 2008; El-Said et al. 2010; El-Shafey 2010).

To answer to some of the drawbacks of the studies reported in the literature, the main goal of this work was to evaluate the efficiency of unmodified rice husk to uptake Hg(II) from aqueous solutions under more realistic conditions and to optimize the sorption process in terms of cost (type and amount of rice husk). In order to achieve this, only metal concentrations similar to those that can be possibly found in the environment and the natural/unmodified form of rice husk were used. The study comprised the assessment of the effect of some physico-chemical parameters such as contact time, initial Hg(II) concentration, and mass of rice husk, together with the kinetic and equilibrium behavior of the Hg(II)–rice husk system.

2 Materials and Methods

2.1 Material and Chemicals

All chemicals were of analytical reagent grade and solutions were prepared with ultra-pure water (18.2 MΩ cm, Milli-Q system). The standard stock solution of mercury (1001 ± 2 mg L−1) and the nitric acid 65 % (Suprapur) were purchased from Merck. Biohit Proline pipettes equipped with disposable tips were used for appropriate dilutions. Experiments were carried out at room temperature (21 ± 1 °C). All glassware material used in the experiments was acid-washed prior to use.

2.2 Biosorbent Material Preparation and Characterization

The rice husk was kindly supplied by the industries of Álvaro Alves Borges Lda. from Figueira da Foz, Portugal, whose activity is related with the peeling, bleaching, and other treatments of rice. The rice husk was washed with distilled water and dried at 60 °C. Afterwards, the material was triturated using a coffee mill (model Taurus aromatic), sieved to obtain a fraction with particles of size ≤500 μm and preserved at room temperature. The procedure used for the preparation of this sorbent was simple, less time consuming, and inexpensive.

The structural characteristics of the rice husk were determined by powder X-ray diffraction of the dried samples and at room temperature. The spectrum was obtained using a Philips X’Pert X-ray diffractometer equipped with a Cu Ka monochromatic radiation source. Intensity data were collected by the step counting method (step 0.02° and time 5 s) in the range of 2θ = 3.5°–70°. The morphological characteristics of rice husk were evaluated by scanning electron microscopy (SEM) using an S-4100 HITACHI equipment. This system was coupled with an electron dispersive spectroscopy (EDS) allowing the identification of the most abundant inorganic elements in the rice husk. Prior to the SEM analysis, the rice husk samples were covered with a thin layer of carbon and an electron acceleration voltage of 20 kV was applied.

Total Hg concentration in the rice husk was evaluated by pyrolysis atomic absorption spectrometry with gold amalgamation (LECO® model AMA-254). No sample digestion was involved in this process.

2.3 Sorption Kinetic Experiments

Batch sorption kinetic experiments were performed at room temperature (21 ± 1 °C), by contact of a certain amount of the unmodified rice husk samples (particle size <500 μm) with a known concentration of Hg(II) solution (0.05 and 0.5 mg L−1). The experimental trials were carried out in volumetric flasks (2 L) under constant stirring conditions (1,400 rpm). The two initial Hg(II) concentrations selected aimed to represent the maximum value for Hg discharges from industrial sectors (0.05 mg L−1) and to simulate an eventual situation of an accidental spill (0.5 mg L−1). The amount of rice used to perform the experiments was 0.50 and 1.00 g for both initial Hg(II) concentrations, corresponding to a rice husk mass-to-volume ratio of (m/VRH) of 0.25 and 0.50 g L−1, respectively. The pH of the solution was kept between 5.7 and 6.0. According to several authors (Khalid et al. 1999; Chuah et al. 2005; El-Said et al. 2010), pH values around six favors and maximizes the sorption of Hg(II) by the rice husk. Two replicates of each experiment were performed and the results were always expressed as the mean value obtained.

Hg(II) solutions were prepared by diluting the standard stock solution to the desired concentration in ultra-pure water. Experiments started when a known mass of the biosorbent was added to Hg(II) solutions and stirring was initiated. Aliquots (10 ml) were collected at increasing times, filtered through an acid-washed 0.45-μm Millipore membrane and then the filtrate was adjusted to pH < 2 with Suprapur nitric acid. Each experiment was maintained until the Hg(II) concentration in the solution remained constant. An experiment without rice husk was always run as a control to assess the loss of Hg(II) to the glass vessels and in the sample filtration procedure.

Hg(II) concentration in the samples was evaluated by cold vapor atomic fluorescence spectroscopy (CV-AFS), on a PSA cold vapor generator, model 10.003, associated with a Merlin PSA detector, model 10.023, using SnCl2 as reducing agent.

2.4 Sorbent Reuse

Kinetic experiments were performed as described previously, and whenever the equilibrium was attained (Ce), the solution was spiked with Hg(II). Each spike (spk) corresponded to an increase of 0.50 mg L−1 in the concentration of Hg(II). This procedure was repeated every time a new equilibrium was achieved (Ce_spk1; Ce_spk2, and so on), until the amount of Hg(II) sorbed by the rice husk remained constant. Two m/VRH were used: 0.25 and 0.50 g L−1. An experiment without rice husk was run as a control, to assess the concentration of Hg(II) in the solution after each spike, and the losses of Hg(II) to the glass vessels and in the filtration procedure.

2.5 Sorption Isotherms Experiments

Sorption equilibrium experiments were conducted by maintaining the initial Hg(II) concentration constant at 0.5 mg L−1 and increasing the amount of rice husk. The m/VRH used for this study was: 0.005, 0.012, 0.025, 0.050, 0.12, 0.25, 0.50, and 1.0 g L−1. The experiments were performed at room temperature (21 ± 1 °C) and at a pH range of 5.7–6.0. The concentration of Hg(II) in the samples was quantified as described previously.

2.6 Evaluation of the Experimental Results

The amount of Hg(II) removed by the rice husk at a given time t (qt, mg g−1) was deduced from the mass balance between the initial Hg(II) concentration in the solution (C0, mg L−1) and the concentration after a particular period of time t (Ct, mg L−1):
$$ {q}_t=\frac{\left({C}_0-{C}_t\right)\;V}{m} $$
(1)
where V (L) is the volume of the solution and m (g) is the dry weight of rice husk. Upon Hg(II) removal, qt increases and Ct decreases along time until equilibrium values (qe and Ce) are attained. The results were also compared using the removal percentage (R), which at time t is defined by:
$$ {R}_t\;\left(\%\right)=\frac{\left({C}_0-{C}_t\right)}{C_0}\times 100 $$
(2)

The C0 and Ct values were always corrected by means of a recovery factor that accounts for the Hg(II) losses from the controls.

2.7 Kinetic and Equilibrium Models

Pseudo-first-order (Eq. 3), pseudo-second-order (Eq. 4), and Elovich (Eq. 5) reaction-based models were used to fit experimental kinetic data:
$$ {q}_t={q}_e\left(1-{e}^{-{k}_1t}\right) $$
(3)
$$ {q}_t=\frac{q_e^2{k}_2t}{1+{q}_e{k}_2t} $$
(4)
$$ {q}_t=\frac{1}{\beta } \ln \left(1+\alpha \beta \right) $$
(5)
where k1 (h−1) is the pseudo-first-order rate constant, k2 (g mg−1 h−1) the pseudo-second-order rate constant, α (g mg-1 h−1) is the initial adsorption rate, and β (g mg−1) is the desorption constant.
Additionally Boyd’s film diffusion (Eq. 6) and Webber’s intraparticle diffusion (Eqs. 7 and 8) models were adjusted to the kinetic data:
$$ {q}_t={k}_{\mathrm{id}}\;{t}^{1/2} $$
(6)
$$ Bt=-0.4977- \ln \left(1-F\right),\mathrm{for}\ F\;\mathrm{values}>0.85 $$
(7)
$$ Bt={\left(\sqrt{\pi }-\sqrt{\pi -\frac{\pi^2F}{3}}\right)}^2,\mathrm{for}\;F\;\mathrm{values}<0.85 $$
(8)
where kid is the internal diffusion rate constant (mg g−1 h−1/2), F is the fractional attainment of equilibrium at different times t and Bt is a function of F.
To fit experimental equilibrium data, linear (Eq. 9), Langmuir (Eq. 10), and Freundlich (Eq. 11) models were used:
$$ {q}_e={K}_l{C}_e $$
(9)
$$ {q}_e=\frac{q_m{K}_L}{1+{K}_L{C}_e} $$
(10)
$$ {q}_e={K}_f{C}_e{}^{1/n} $$
(11)
where Kl (L g−1) is the distribution coefficient, qm (mg g−1) the maximum sorption capacity of the sorbent, KL (L mg−1) is a constant related to the affinity of the binding sites, Kf (mg1−1/n L1/n g−1) is the sorption capacity constant and n the sorption intensity constant.

Additional information regarding the kinetic and equilibrium models is described in Appendix A.

2.8 Error Analysis

The parameters of the kinetic and equilibrium models here considered were obtained by nonlinear regression analysis using GraphPad Prism 5 program (trial version), which uses the least-squares as fitting method and the method of Marquardt and Levenberg, which blends two other methods, the method of linear descent and the method of Gauss-Newton for adjusting the variables. The goodness of the fittings of the kinetic and equilibrium models to the experimental data was confirmed by calculating the R square (R2) and the sum of squares (SS). Additionally, the Akaike’s information criteria (AIC) and the evidence ratio were used to compare equilibrium models. Lower AIC values (on a scale form –∞ to +∞) suggests that the corresponding model is more likely to occur than the alternative models and the evidence ratio is a numerical value representative of the number of times that the model with a lower AIC is more likely to be correct (Malash and El-Khaiary 2010).

The analyses of the experimental sorption dynamic results and the determination of parameters from the diffusion models (film and intraparticle diffusion) were obtained by a Microsoft® ExcelTM worksheet developed by Malash and El-Khaiary (2010). The use of this method allows the definition of the linear segments and provides a much more accurate prediction of the diffusion coefficients and the diffusion mechanisms involved. The definition of the numbers of linear segments that described the diffusion process (two, three, or four steps) was assessed by comparing the AIC obtained for each model and the evidence ratio. The goodness of the fittings was also evaluated in terms of the correlation coefficient (r) of linear segments. The standard deviations (SD) of all fitted parameters are presented.

3 Results and Discussion

3.1 Rice Husk Characterization

The morphologic and structural characterization of the unmodified rice husk is shown in Fig. 1. According to the scanning electron micrograph obtained at a lower magnification (50 times; Fig. 1, A1), the rice husk material presents an inner and an outer epidermis. The inner (Fig. 1, A2) and the outer (Fig. 1, A3 and A4) epidermis is characterized by having different morphologic aspects and levels of organization. Typically, the morphology of the outer epidermis of rice husk presents a well-organized profile and is characterized by having an arrangement of linear ridges and furrows and the ridges are punctuated with globular protuberances. The inner epidermis is composed of rectangular tissues disposed in a parallel way (Park et al. 2003; Liou and Wu, 2009). The EDS analysis (Fig. 1) indicates the presence of high contents of carbon, oxygen, and silica in both epidermis of rice husk and small amounts of potassium and calcium, respectively, in the outer and inner epidermis. The levels of silica found in the outer epidermis of this material were higher than those from the inner epidermis and conversely the levels detected for carbon were lower. The results from the X-ray diffraction analysis of rice husk (Fig. 1) show a dominant broad peak from 15° to 35° 2θ diffraction angles, indicating the presence of amorphous silica. The irregular morphology and the high amounts of silica in its surface (in particular in the outer epidermis tissue), give a suitable morphological profile to rice husk for retaining metals (Tarley and Arruda, 2004; Rocha et al. 2009). The outer epidermis of rice husk was also analyzed after the sorption studies and the morphologic features observed were quite analogous, although the surface of this material became smoother like a sort of surrounding layer (Fig. 1, A4). The data obtained from the EDS spectrum (data nor showed) corroborate these results and reveal the similar composition of the rice husk surface before and after being used in water treatment.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1599-9/MediaObjects/11270_2013_1599_Fig1_HTML.gif
Fig. 1

Characterization of rice husk particles with size ≤500 μm: EDS and X-ray spectra; SEM micrographs for a magnification of ×50 (A1) and ×300 (A2 to A4), before (A1 to A3) and after (A4) the use of the material in the sorption experiments

The analysis of the total concentration of mercury in the rice husk indicates that this material has a very low Hg(II) content (3.9 ng g−1, mean value of three replicates, with a relative standard deviation less than 8 %).

3.2 Sorption Studies

3.2.1 Influence of the Initial Hg(II) Concentrations and Rice Husk Amount on the Sorption Kinetic Process

Figure 2 presents the kinetic curves (Ct/C0 vs t) for two initial Hg(II) concentrations (0.05 and 0.5 mg L−1) and for two m/VRH (0.25 and 0.50 g L−1). The time profile was characterized by an sudden and noteworthy decrease in the Hg(II) concentration within the first hours of contact (ca. 6 h), representing a rapid uptake of this metal by the rice husk, followed by a subtle decrease in Ct/C0 (low uptake of Hg(II)). This behavior is due to the high driving force at the beginning of the sorption process, since initially the rice husk particles are Hg-free. The initial Hg(II) concentration and the rice husk m/V ratio did not have a strong effect on the contact time needed to attain the equilibrium and for all tested conditions, the full equilibrium was attained after a contact period of 96 h.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1599-9/MediaObjects/11270_2013_1599_Fig2_HTML.gif
Fig. 2

Variation of the normalized concentration (Ct/C0) with time (t, h) for the following experimental conditions: C0 Hg of 0.05 mg L−1 (a and b) and 0.5 mg L−1 (c and d), and m/VRH of 0.25 g L−1 (black circle, black diamond) and 0.50 g L−1 (white circle, white diamond)

The efficiency of the process was evaluated at different periods of contact (6 and 24 h and at the equilibrium) and the values were expressed as: percentage (Rt), amount of Hg(II) sorbed per gram of rice husk (qt, mg g−1) and concentration of Hg(II) remaining in the solution (CHg, mg L−1; Table 1). The results reveal that after a period of contact of just 6 h, the removal efficiency of Hg(II) by rice husk achieved values between 42 and 66 % for C0 Hg of 0.05 mg L−1 and between 68 and 72 % for C0 Hg of 0.50 mg L−1. In both cases, the highest percentages were obtained for a m/VRH of 0.50 g L−1. By extending the period of contact, the removal efficiency increased and the values obtained, ranged between 61 and 73 % after 24 h and between 82 and 92 % after 96 h. The highest percentages of Hg(II) removal were achieved for the highest C0 Hg (0.50 mg L−1), which can be explained by the fact that the use of a higher initial concentration provides the required driving force to overcome sorbate mass-transfer processes between the aqueous and solid phases (Ahmaruzzaman and Gupta, 2011).
Table 1

Removal of Hg(II) (Rt, %), residual concentration of Hg(II) in solution (CHg,, mg/L), and amount of Hg(II) sorbed per gram of rice husk qt (mg g−1), at different times: 6 and 24 h and at equilibrium. The values are expressed as the mean of two replicate experiments

C0 Hg (mg L−1)

m/VRH (g L−1)

t (h)

Rt (%) (%)

CHg (×10−3 mg L−1)

qt (mg g−1)

0.05

 

6

42

27.9

0.088

 

0.25

24

61

19.7

0.120

  

eq

82

8.8

0.164

  

6

66

17.0

0.066

 

0.50

24

72

14.2

0.071

  

eq

84

8.2

0.083

0.50

 

6

68

160

1.35

 

0.25

24

73

135

1.45

  

eq

91

47.5

1.80

  

6

72

138

0.72

 

0.50

24

73

135

0.76

  

eq

92

38.1

0.92

For the highest initial concentration and for the entire range of time, no relevant differences were obtained on the removal of Hg(II) when different rice husk amounts were used (m/VRH of 0.25 and 0.50 g L−1). However, when starting with C0 Hg of 0.05 mg L−1 the time removal profiles are quite distinct for the initial hours but tend to converge for the same removal value (ca. 80 %), at equilibrium. In terms of the amount of Hg(II) sorbed at equilibrium per gram of material, the values varied between 0.16 and 1.80 mg g−1 (m/VRH of 0.25 g L−1) and 0.083 and 0.92 mg g−1 (m/VRH of 0.50 g L−1), for an initial Hg(II) concentration of 0.05 and 0.50 mg L−1, respectively (Table 1), increasing with the increasing of initial Hg concentration and with the decreasing of m/VRH. Considering the levels of Hg(II) remaining in solution after the sorption process with unmodified rice husk, it must be highlighted that for an initial concentration of 0.5 mg L−1 and for both m/VRH, the residual concentrations achieved were lower than the guideline limits established in the Portuguese legislation for Hg discharges from industrial sectors (CHg < 0.05 mg L−1, Council Directive 84/156/EEC). When the initial Hg(II) concentration was the actual maximum value for Hg discharges from industrial sectors, itself, the equilibrium concentrations achieved for both m/V ratios (0.25 and 0.50 g L−1) were lower than 9 × 10−3 mg L−1, a value relatively close to the guideline value for drinking water quality (1 × 10−3 mg L−1, Council Directive 98/83/EC).

The effect of the amount of rice husk on the sorption process of Hg(II) was widely studied for an initial concentration of Hg(II) of 0.50 g L−1. Figure 3 presents the values obtained for the removal of Hg(II) expressed in percentage (Re, %) and as amount of Hg(II) sorbed per gram of rice husk, at the equilibrium (qe, mg g−1). High percentages of Hg(II) removal and low values of Hg(II) sorbed per gram of material (qe) were obtained by increasing the m/VRH. The minimum qe value was 0.47 mg g−1 (m/VRH of 1.0 g L−1) and the maximum was 24.1 mg g−1 (m/VRH of 0.005 g L−1). In absolute terms, the removal of Hg(II) increased from 24 to 94 %, with the increase of the m/VRH from 0.005 to 1.0 g L−1. This behavior is attributed to an increase on surface area and consequently on the number of available sorption sites. However, for m/VRH higher than 0.25 g L−1 (Fig. 3), there are no relevant improvements on the removal of Hg(II), which remained practically unchanged, reaching a “plateau” of 92 ± 2 %, while the amount of Hg(II) sorbed decreased continually reaching a very small value. These results suggest that although for m/VRH higher than 0.25 g L−1 the large majority of the active sorption sites of the rice husk remain free, the removal efficiency is no longer dependent of the number of sorption sites available. Contrary, for values lower than 0.25 g L−1, the surface area and consequently the number of available sites for sorption takes place, clearly play an important role on the removal efficiency, although there are still a high number of active sorption sites free (because qe values increase continuously with decreasing m/VRH). This fact may be related with more intense electrostatic interactions when the number of sorption sites decreases. As a result an m/VRH of 0.25 g L−1 represents the optimal amount that can be effectively used for an initial concentration of Hg(II) of 0.50 g L−1, assuring the efficiency on the removal process and minimizing the amount of sorbent required, reducing issues concerning with secondary waste disposal problems.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1599-9/MediaObjects/11270_2013_1599_Fig3_HTML.gif
Fig. 3

Variation of the removal of Hg(II) (Re, %, bars) and the amount of Hg(II) sorbed per gram of rice husk (qe, mg g−1, white square) for different m/VRH

3.2.2 Sorbent Reuse

The results obtained in the foregoing section revealed that under the experimental conditions used, the amount of Hg(II) sorbed did not reach a constant value, suggesting that the maximum sorption capacity (saturation) of the rice husk was not achieved, and also indicating the presence of available binding sites in the rice husk and the possibility of reusing the same material in further treatments. The capacity of the unmodified rice husk to remove Hg(II) continuously was evaluated, by increasing the concentration of Hg(II) in the solution (C0_spk, mg L−1) in contact with the sorbent, through consecutive mercury spikes, added after the system reached equilibrium. Figure 4 presents both the amount of Hg(II) sorbed per gram of rice husk (qe_spk, mg g−1; Fig. 4a) and the residual concentration of Hg(II) in solution (Ce_spk, mg L−1; Fig. 4b), afterward reached a new equilibrium, after each spike.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1599-9/MediaObjects/11270_2013_1599_Fig4_HTML.gif
Fig. 4

Effect of increase mercury levels (C0_spk, mg L−1) on: a amount of Hg(II) sorbed per gram of material at equilibrium (qe_spk, mg L−1) and b residual concentration of Hg(II) in the solution (Ce_spk, mg L−1), for a m/VRH of 0.25 g L−1 (black) and 0.50 g L−1 (white). The dash line in (b) corresponds to the guideline value of Hg discharges from industrial sectors, 0.05 mg L−1

According to Fig. 4a, the amount of Hg(II) sorbed per gram of rice husk at equilibrium is directly proportional to the increase of Hg(II) concentration in solution for a narrow range of concentrations. The strong correlation, given by the linearity between qe_spk and C0_spk, for the range of concentrations 0–1.0 mg L−1 in the case of 0.25 g L−1 of m/VRH (r = 0.999) and 0–2.0 mg L−1 for m/VRH of 0.50 g L−1 (r = 0.997), is indicative of the unsaturation of the rice husk and of the existence of the available binding sites on the material. Although qe_spk increases continuously with C0_spk after several spikes, the deviation from linearity started to occur for concentrations of Hg(II) (C0_spk) higher than 1.0 and 2.0 mg L−1, respectively, for m/VRH of 0.25 and 0.50 g L−1, suggesting the beginning of the saturation of this sorbent. The “plateau” value on qe_spk was only reached for Hg(II) concentrations higher than 3.2 mg L−1 (Fig. 4a). Moreover, the results suggest that with the same rice husk, it was still possible to achieve concentrations lower than the maximum value for Hg discharges from industrial sectors (0.05 mg L−1, value represented by a dashed line in Fig. 4b) after two and three consecutive spikes, respectively, for m/VRH of 0.25 and 0.50 g L−1, which corresponded to Hg(II) concentrations in the range of 0.5–1.6 mg L−1 (Fig. 4b). For concentrations of Hg(II) higher than 1.0 mg L−1 (third spike) for m/VRH = 0.25 g L−1 and higher than 1.6 (fourth spike) for m/VRH = 0.50 g L−1, the removal process continues to run with removal percentages that varied between 69 and 64 % for a m/VRH of 0.25 g L−1 and between 87 and 77 % for a m/VRH of 0.50 g L−1.

3.2.3 Kinetic Modeling of the Results

The sorption kinetic data obtained for the Hg(II)–rice husk system were fitted to the non-linear forms of the following reaction-based models: pseudo-first-order reaction, pseudo-second-order reaction and Elovich model. The fittings of the experimental results and the estimated parameters values are presented in Fig. 5 (a to d) and Table 2, respectively. It can be seen by the R square (R2) and the sum squares (SS) that under the experimental conditions used in this work, the Elovich model provides the best fit, with R2 values ranging from 0.961 to 0.989 and a SS values between 0.0004 and 0.0539. The kinetic data fitted by the pseudo-first-order and pseudo-second model order were less satisfactory than the fitting obtained with Elovich model, which was reflected by the values of R2 (between 0.874 and 0.937 for pseudo-first-order and 0.910–0.974 for pseudo-second-order) and SS (between 0.0006 and 0.55 for pseudo-first-order and 0.0003–0.28 for pseudo-second-order). Additionally, the qe values predicted (qefit) from those models were poorly estimated and different from the experimental values (qeexp). While the qe1fit values obtained by the pseudo-first-order equation were always underestimated, in the case of the pseudo-second-order equation, the qe2fit values were overestimated for the highest C0 Hg and underestimated for the lowest. The modeling of the kinetic process by the Elovich equation indicates that high initial sorption rates α were obtained for the highest initial concentration of Hg(II) and for high m/VRH. As for the desorption constant β, the values estimated increased with a decrease of the initial concentration of Hg(II) and with an increase of the amount of sorbent used.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1599-9/MediaObjects/11270_2013_1599_Fig5_HTML.gif
Fig. 5

Kinetic modeling of the experimental data obtained from the sorption process of Hg(II) onto rice husk. The following reaction based models were used: pseudo-first order (gray line), pseudo-second (dashed line) order, and Elovich (black line); and the diffusion model applied was the intraparticle diffusion. Experimental conditions: C0 Hg of 0.05 mg L−1 (A, B, E, and F) and 0.5 mg L−1 (C, D, G, and H) and m/VRH of 0.25 (black circle, black diamond) and 0.50 g L−1 (white circle, white diamond)

Table 2

Experimental qeexp and data obtained from the fitting of the experimental results with: first-order (qefit and k1), pseudo-second-order (qefit and k2) and Elovich (a and α) reaction models, and Boyd’s film diffusion (intercept and its corresponding interval, Di and r) and Weber’s pore diffusion (breakpoint, kid and r) models

   

Reaction models

   

Pseudo-1st-order

Pseudo-2nd-order

Elovich

C0 Hg(II)

m/VRH

qeexp ± SDa

qe1fit ± SDb

k1 ± SDb

qe2fit ± SDb

k2 ± SDb

β

α

(mg L−1)

(g L−1)

(mg g−1)

(mg g−1)

(h−1)

(mg g−1)

(g mg−1 h−1)

(g mg−1)

(g mg−1 h−1)

0.05

0.25

0.163 ± 0.003

0.130 ± 0.008

1.27 ± 0.37

0.135 ± 0.007

11.4 ± 3.9

57.2 ± 3.8

0.669 ± 0.241

 

0.50

0.083 ± 0.003

0.074 ± 0.003

2.50 ± 0.43

0.076 ± 0.002

48.3 ± 7.2

128 ± 9

2.72 ± 1.38

0.50

0.25

1.80 ± 0.01

1.55 ± 0.07

1.26 ± 0.30

1.61 ± 0.06

1.13 ± 0.26

5.51 ± 0.20

22.6 ± 5.0

 

0.50

0.921 ± 0.004

0.808 ± 0.039

2.67 ± 0.61

0.825 ± 0.028

5.29 ± 1.15

12.8 ± 0.6

61.0 ± 20.8

   

Diffusion models

   

Boyd’s film diffusion

Webber’s diffusion model

C0 Hg(II)

m/VRH

Stage

Intercept

Di

r

Breakpoint

kid ± SD

r

(mg L−1)

(g L−1)

  

(cm2 h−1)

 

(h)

(×10−2 mg g−1 h−1/2)

 

0.05

0.25

1st

−0.022 [−0.049–0.005]

0.997

1.66

5.77 ± 6.31

0.967 (n = 4)

  

2nd

 

0.66 ± 0.09

0.994 (n = 8)

 

0.50

1st

−0.091 [−0.287–0.105]

0.998

1.66

4.00 ± 4.53

0.952 (n = 4)

  

2nd

 

0.17 ± 0.02

0.992 (n = 8)

0.50

0.25

1st

0.061 [0.035–0.088]

5.13 × 10−4

0.997

3.70

53.2 ± 6.33

0.996 (n = 6)

  

2nd

 

4.41 ± 0.62

0.993 (n = 7)

 

0.50

1st

0.113 [0.050–0.176]

1.14 × 10−3

0.989

1.66

35.0 ± 9.6

0.988 (n = 5)

  

2nd

 

2.05 ± 0.16

0.997 (n = 8)

The experimental conditions used were: initial concentration of Hg(II) of 0.25 and 0.50 mg L−1 and m/VRH of 0.25 and 0.50 g L−1

aStandard deviation (SD) calculated by means of 6 values (n = 6)

bStandard deviation (SD) calculated by means of 14 values (n = 14)

The works reported in the literature for rice-husk-based materials, generally uses the pseudo-first-order and pseudo-second-order models to describe the sorption kinetic behavior of Hg(II)–rice husk systems, and in most of these studies, the pseudo-second-order model was better adjusted to the experimental data (Wong et al. 2003; Al-Degs et al. 2006; Kumar and Bandyopadhyay, 2006; El-Said et al. 2010; El-Shafey, 2010; Feng et al. 2004). In the present work, besides the pseudo-first-order and pseudo-second-order, the Elovich model was used, and the results obtained indicate that this model provides a better adjustment to experimental data than the other models.

In order to gain insight into the mechanism and rate-controlling steps affecting the kinetic of the sorption process of Hg(II) onto unmodified rice husk, the Webber’s pore-diffusion and Boyd’s film-diffusion models were used. The piecewise linear regression (PLR) was applied to the experimental data and the numerical results are presented in Table 2. By applying the PLR treatment to the data obtained using a C0 Hg of 0.05 mg L−1 and for both m/VRH, it was found that the first linear segment in the Boyd’s plot (data not showed) is linear and that the confidence interval of the intercept, for a 95 % confidence level, includes zero. This strongly suggests that film diffusion is not the rate controlling during this stage and that intraparticle diffusion controls the entire sorption process (Hameed and El-Khaiary 2008; Malash and El-Khaiary, 2010). The Weber Morris plots (Fig. 5e and f) indicate the existence of multi-stages for both m/VRH. The definition of the numbers of linear segments was assessed by comparing the Akaike’s information criteria obtained for each model and by calculating the evidence ratio. Lower AIC were obtained considering two linear segments and the evidence ratio calculated, indicate that the two stages approach is 145 (m/VRH of 0.25 g L−1) and 413 (m/VRH of 0.50 g L−1) times better than the three stages. As result, the Weber Morris plots for both m/VRH were defined by two operational stages, with a break point (time corresponding to point where the two linear segments meet) of 1.66 h. The first stage corresponds to the steep-sloped portion (from 0.08 to 1.66 h) and the second stage (from 1.66 to 144 h) corresponds to the linear gentle-sloped portion of the qt vs t1/2 plot. The regression results from the first linear segment of Weber Morris plots estimated intercept values that are not significantly different from zero, corroborating the conclusion obtained from the Boyd plots, i.e., that pore diffusion controls the overall rate sorption process, even in the early stage. The intraparticle diffusion rate constants kid were calculated from the slope of the two corresponding lines (Table 2) and the values obtained were higher for the first than for the second stage. According to some authors, this fact can be explained by the diffusion of Hg(II) in pores of rice husk of distinct sizes and progressively smaller (macropores and mesopores) (Hameed and El-Khaiary 2008; El-Khaiary and Malash, 2011). Since the available path for the diffusion of Hg(II) in pores of smaller sizes is reduced, this explains the lower kid values obtained in the second linear segment. For a C0 Hg of 0.05 mg L−1, the amount of sorbent used had a little effect on the diffusion of Hg(II) through macropores of rice husk (kid for the lowest m/VRH was 1.4 times higher than kid for the highest one), however the diffusion through mesopores was more affected by the m/VRH, resulting in a decrease in kid ca. four times by doubling the m/VRH.

Starting with an initial concentration of Hg(II) of 0.50 mg L−1 and for both m/VRH, the Boyd’s plots are linear (correlation r ≥ 0.989) in the initial sorption period and presents an intercept, for a 95 % confidence level, statistically different from zero. These results strongly suggest that under these experimental conditions, film diffusion is the rate-controlling process in the period between 0.08 and 1 h. The Webber’s pore-diffusion plots to the same set of experimental data (Fig. 5g and h) indicate the existence of multi-linearity, and according to the AIC and the evidence ratio, two linear segments can be defined with a break point of 3.70 and 1.66 h, for m/VRH of 0.25 and 0.50 g L−1, respectively. The information obtained from Boyd’s plots and the intercept values from the Weber–Morris plots (statistically different from zero for a 95 % confidence level), indicate that for the period corresponded to the first segment, the sorption process is controlled by film diffusion of Hg(II) from the external surface into the pores (Hameed and El-Khaiary 2008; Malash and El-Khaiary, 2010). The period that corresponds to the second segment is controlled by the diffusion of Hg(II) on the internal surface of the pores. The kid calculated from the second segment and for a m/VRH of 0.50 g L−1 was ca. 2.4 times higher than the one obtained for 0.25 g L−1. The modeling of the kinetic process by the intraparticle diffusion model also indicates that kid (from the second stage) increased ca. one order of magnitude, increasing the initial Hg(II) concentration in solution. According to Ofomaja (2010)), this phenomenon can be explained by the fact that a high initial concentration will give rise to a higher concentration gradient, being responsible to eventually cause a faster diffusion and a rapid sorption process.

3.3 Sorption Isotherms Studies

In order to investigate the sorption capacity of the rice husk, three isotherms were used in the present work: linear, Freundlich (F), and Langmuir (L). In the range of equilibrium Hg(II) concentrations studied, the fitting of the equilibrium data using the Langmuir isotherm failed. At low sorbate concentrations, the Langmuir isotherm effectively reduces to a linear isotherm (Ho et al. 2002) and the non-linear convex shape curve (“plateau”), which is characteristic of this model, it was not achieved (Voudrias et al. 2002). Contrary, there is a good agreement between the experimental data and the fittings of the linear and the Freundlich isotherms (Fig. 6), as confirmed by the sum of squares (SSlinear = 4.2 and SSF = 3.1) and the R square (R2linear = 0.986 and R2F = 0.990). According to the previous parameters, the Freundlich isotherm seemed to be the model that better adjusts the experimental data. However, linear and Freundlich have different degrees of freedom, and as a result, the goodness of the fitting cannot be based solely on the SS and R2 values and other parameters, such as the AIC and the evidence ratio must be calculated. The smaller AIC value obtained for linear isotherm (AIClinear = −6.13 and AIClinear = −5.71) suggests that this model is more likely to be adjusted to the experimental data. The calculus of the evidence ratio indicates that the linear model is 314 times more likely to be correct than the Freundlich model. According to the data found in the literature for the sorption of Hg(II) using the unmodified rice husk and its carbonized form (rice husk ash), both Langmuir and Freundlich isotherms are suitable to fit the experimental values (Tiwari et al. 1995; Khalid et al. 1999; Feng et al. 2004; Chuah et al. 2005; Krishnani et al. 2008; El-Said, 2010). However, the range of concentrations used on those studies are considerably higher than the used in this study, explaining the adjustment of the linear isotherm to the experimental data.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1599-9/MediaObjects/11270_2013_1599_Fig6_HTML.gif
Fig. 6

Experimental equilibrium data (square) for the Hg(II)–rice husk system, together with fittings corresponding to linear (bold line) and Freundlich (dash line) isotherms (21 ± 1 °C)

The distribution coefficient Kl obtained from the linear isotherm, which is indicative of the distribution of Hg ions between the liquid and the solid phase, was 36.1 ± 1.0 L g−1.

3.4 Sorption Efficiency of Rice Husk to Remove Hg(II)

The sorption capacity of the unmodified rice husk to remove Hg(II) was compared with other sorbent materials, by evaluating the qe values, that is, the amount of Hg(II) per unit mass of sorbent at equilibrium (Table 3). The results reveal that the sorption capacity of the rice husk toward Hg(II) is, in general, of the same order of magnitude than the values found in the literature for other sorbents, including the eucalyptus bark, the plant Coriandrum sativum, the Fraxinus tree leaves, and the leaves of castor tree. Furthermore, the qe values are in agreement with the ones obtained with derived sorbents from rice husk (unmodified/natural rice husk, rice husk ash and biomatrix from rice husk). Only the rice husk chemically treated with sulfuric acid and the white rot fungus present qe values which are exceptionally higher (ca. 10 to 20 times higher, respectively) than the maximal qe value obtained for the unmodified rice husk, in the present study. However, it must be highlighted that majority of these studies were performed for Hg(II) concentrations, which in most of the cases are extremely high, and do not represent the existing environmental conditions or the ones that may be provoked by effluents or wastewaters discharges. Consequently, the outputs of these studies cannot be translated into the real world, since the conditions studied are not representative of the real problems. On the contrary, this study not only used realistic mercury levels, but also sorbent amounts considerably lower than the majority of the ones, used in other studies, reducing this way the production of residues. The results obtained using unmodified rice husk showed that this biosorbent has potential to remove Hg(II) from aqueous media, under environmental realistic conditions.
Table 3

Sorption capacity found in the literature (qe, mg g−1) for some sorbents used in the removal of Hg(II). The sorbent mass-to-volume ratio (m/Vsorbent, g L−1), pH, and initial Hg(II) concentration (C0Hg(II), mg L−1) are also mentioned

Material

m/Vsorbent (g L−1)

pH

C0Hg(II) (mg L−1)

qe (mg g−1)

Reference

Microorganism d (Bacillus sp.)

2.0

6.0

0.25–10

0.12–3.41

Green-Ruiz 2006

White rot fungus (Lentinus edodes)

1.0

6.0

500

403

Bayramoğlu and Arica 2008

Fraxinus tree leaves

5.0

4.4

10–300

9.37–25.4

Zolgharnein and Shahmoradi 2010

Eucalyptus bark

5

5.0

25–200

9.37–21.7

Ghodbane and Hamdaoui 2008

Leaves of castor tree (Ricinus communis L.)

2.5

5.5

5–100

1.71–26.5

Shaban et al. 2008

Plant (Coriandrum sativum)

13

[28]

25

24.0

Karunasagar et al. 2005

Fly ash

2.0

5.8

20–40

5.60–10.1

Banerjee et al. 2004

Rubber from tyre wastes (residual rubber)

3.2

6.1

802

∼40

Manchón-Vizuete et al. 2005

Rice husk ash

10

6.0

10–100

0.81–3.40

El-Said et al. 2010

Rice husk ash

2

5.6–5.8

8–335

∼0.6–7.0

Feng et al. 2004

Rice husk treated with H2SO4

0.75

6.0

100–1,500

∼40–250

El-Shafey 2010

Biomatrix from rice husk

3

6.0

200

33.1

Krishnani et al. 2008

Unmodified rice husk

100

6.0

130

1.26

Khalid et al. 1999

Unmodified rice husk

0.005–1.0

[5.7–6.0]

0.50

0.47–24.1

Present study

4 Conclusions

The removal of Hg(II) by unmodified rice husk was investigated in batch mode, and for the first time realistic environmental concentration levels of Hg(II) (μg L−1 range) were used. The performed experiments revealed that small amounts of rice husk (0.25 and 0.50 g L−1) were able to reduce Hg(II) levels in ca. 80 % when the initial concentration (C0Hg) was the maximum value allowed for Hg discharges (0.05 mg L−1) and ca. 90 % when the initial concentration was ten times higher (0.50 mg L−1). The saturation capacity of the rice husk was not reached, which allowed the reuse of the same material in some additional treatments, maintaining the efficiency in the removal process (>90 %) and the residual concentrations of Hg(II) with levels lower than the established value for Hg discharges from industrial sectors (0.05 mg L−1).

The kinetics of the Hg(II)/rice husk system was studied in this work and both reaction and diffusion models were used. The experimental data was well predicted by the Elovich reaction model and the diffusion models suggested the occurrence of two operational stages in the sorption process of Hg(II) onto the rice husk. The results obtained from the statistical piecewise linear regression indicate that intraparticle diffusion controlled the entire sorption process for the lowest levels of Hg(II) (C0Hg = 0.05 mg L−1), and for the highest levels (C0Hg = 0.50 mg L−1), both film diffusion and pore diffusion mechanisms had an important role in controlling the sorption process. The equilibrium results indicate that under the experimental conditions used in this work, the experimental data was better described by the linear isotherm model.

As a main conclusion of this work, the unmodified rice husk material proved to be very efficient for the removal of Hg(II), even from aqueous solution with relative low levels of contamination (0.05 and 0.5 mg L−1). Although the removal of pollutants from water is an abundant field in literature, the capacity of biosorbents to remove metals under realistic conditions is still largely unknown, and so, this study is an important contribution to understand and to evaluate the effective capacity of the unmodified rice husk, under realistic conditions.

Acknowledgments

Thanks are due to University of Aveiro/CESAM and Fundacão para a Ciência e a Tecnologia (FCT). Luciana Rocha and Cláudia B. Lopes acknowledge their Post-doc grants to FCT (SFRH/BPD/47166/2008 and SFRH/BPD/45156/2008, respectively).

Supplementary material

$$ \frac{d{q}_t}{ dt}={k}_1\left({q}_e-{q}_t\right) $$$$ {q}_t={q}_e\left(1-{e}^{-{k}_1t}\right) $$$$ \frac{d{q}_t}{ dt}={k}_2{\left({q}_e-{q}_t\right)}^2 $$$$ {q}_t=\frac{q_e^2{k}_2t}{1+{q}_e{k}_2t} $$$$ h={k}_2{q}_e{}^2 $$$$ \frac{d{q}_t}{ dt}=\beta \exp \left(-\alpha {q}_t\right) $$$$ {q}_t=\frac{1}{\beta } \ln \left(1+\alpha \beta \right) $$$$ {q}_t={k}_{\mathrm{id}}\;{t}^{1/2} $$$$ F=1-\frac{6}{\pi^2}{\displaystyle \sum {{}_n^{\infty}}_{=1}\frac{1}{n^2} \exp}\left(-{n}^2 Bt\right) $$$$ F=\frac{q_t}{q_e} $$$$ \mathrm{For}\;F\;\mathrm{values}>0.85\kern0.5em  Bt=-0.4977- \ln \left(1-F\right) $$$$ \mathrm{For}\;F\;\mathrm{values}>0.85\kern0.5em  Bt={\left(\sqrt{\pi }-\sqrt{\pi -\frac{\pi^2F}{3}}\right)}^2 $$$$ B=\frac{\pi^2{D}_i}{r^2} $$$$ {q}_e={K}_l{C}_e $$$$ {q}_e=\frac{q_m{K}_L}{1+{K}_L{C}_e} $$$$ {q}_e={K}_f{C}_e^{1/n} $$

Copyright information

© Springer Science+Business Media Dordrecht 2013