Retention and release of hexavalent and trivalent chromium by chitosan, olive stone activated carbon, and their blend

Abstract

Shrimp shells and waste of olive stones were used as feedstock for the preparation of chitosan and activated carbon. The adsorption of Cr^VI and Cr^III species in aqueous solution by the materials prepared and their blend were studied by using the well-known kinetic and isotherm models, Fourier transform infrared spectroscopy and scanning electron microscope. It was demonstrated that the rates of adsorption were controlled by diffusion inside particles and throughout the liquid film, and adsorption occurred spontaneously (− 26 < ∆G° < − 15 kJ/mol) in the range of 298–333 K, except for that involving Cr^III species and activated carbon. The maximum amounts of Cr species retained by the composite (146 mg of Cr^VI/g and 33 mg of Cr^III/g at 298 K) were three times greater than those of the basic constituents. Adsorption was markedly affected by temperature and pH, and Cr^VI species were substantially desorbed in acid mediums, particularly in acetic acid solution. The recovery of Cr^III species varied according to the adsorbent and the solution used. The immobilization of Cr^VI species (HCrO₄⁻ and CrO₄²⁻) and Cr^III species (Cr(OH)₂⁺ and Cr₃(OH)₄⁵⁺) by chitosan was accomplished by means of amine moieties and hydroxyls of D-glucosamine units of the biopolymer. The adsorption of Cr^VI species on activated carbon involved π electrons of aromatic rings as well as oxygenated sites (C–OH, C=O, C–O–C). In such a condition, Cr^VI was partially reduced into Cr^III. For the composite, the amino functional groups of chitosan and hydroxyls of both constituents were implicated in the linkage of the biopolymer and activated carbon, and the C–O–H and C–O–C functional groups of chitosan were involved in the retention of Cr^VI species. For Cr^III species, adsorption occurred preferentially on hydroxyls of the components, and consequently, the chains of the biopolymer recovered some flexibility.

Introduction

Chitosan, which is a shellfish and crustacean originated polymer, is one of the biopolymers with versatile applications (Rinaudo 2006). It could be suitable for the uptake of various chemical pollutants (Crini et al. 2009). The chains of chitosan consist of D-glucosamine units, and the polarization of the amine groups depends on pH (Guibal 2004). In acid aqueous solutions, the macromolecular chains acquire positive charges principally due to the formation of –NH₃⁺ groups. In such a condition, anionic chemical pollutants such as oxyanions (chromate, selenate, molybdate, and arsenate) and anionic dyes might be adsorbed. In contrast, cationic chemical pollutants such as heavy metals and cationic dyes could be adsorbed in alkaline aqueous mediums. In this case, the amino groups played a key role in the process of adsorption (Gerente et al. 2007).

Biochars and activated carbons, prepared from natural lignocellulosic feedstocks, are promising ubiquitous materials for land remediation and chemical pollutant immobilization (Rajapaksha et al. 2016; Agarwal and Singh 2017). These effects are mainly linked to their porous structure, large to moderate surface areas, and to surficial moieties (–OH, –COOH, C=O, C–O–C). To activate biomass-derived chars, physical and chemical treatments are commonly performed. Physical activation is achieved by heating at temperatures in the range of 820–980 °C under controlled oxidizing atmosphere (air, stream). The chemical activation is realized by using acids (sulfuric, phosphoric, and nitric acids) or oxidizing agents (peroxide, potassium permanganate) (Rajapaksha et al. 2016). Olive stone cakes (by-products of olive oil mills mainly encountered in Mediterranean countries) were used as feedstock for activated carbon (e.g., El-Sheikh 2004; Rwayhah et al. 2017). According to some authors, activated carbons, which were prepared from olive stone wastes, can immobilize Cu²⁺ (Baccar et al. 2009), Cd²⁺ and Ni²⁺ ions (Bohli et al. 2012), and phenol (Bohli et al. 2013).

Referring to the limited published papers dealing with the adsorption ability of chitosan-activated carbon composites, the uptakes of phenol (Venault et al. 2008) and cadmium (Hydari et al. 2012) exceeded those of the basic components. On the other hand, to the best of our knowledge, no study was devoted to the adsorption processes of chemical pollutants by composites of chitosan and activated carbon prepared from olive stone wastes.

Chromate (CrO₄²⁻), dichromate (Cr₂O₇²⁻), and their derivatives (HCrO₄⁻, HCr₂O₇⁻) are commonly encountered in wastewaters. They are highly mobile and may contaminate surface water and groundwater. Because of the hexavalent chromium, known as a powerful oxidant, the latter chemical species provoke skin ulcer, liver and kidney damage, and lung carcinoid tumors (Sarkar et al. 2010; Mohan et al., 2011; Tytlak et al. 2015). Thereby, a great concern has been raised about Cr^VI-contaminated water. Referring to the World Health Organization recommendations, the amount of chromium in drinking water should be below 0.05 mg/L.

Removal of Cr cations from aqueous solutions may be achieved by adsorption, which is an effective technique and easy to be implemented. Activated carbon is widely used as adsorbent since it is suitable for removing organic and inorganic compounds from aqueous solutions. Chitosan was also considered as a suitable adsorbent for Cr^VI species. However, disparate results regarding the uptake amount, the thermodynamic data, and the adsorption isotherms were observed (Aydın and Aksoy 2009; Ngah et al. 2011; Kyzas and Bikiaris 2015). As a matter of fact, the adsorption of Cr^VI by chitosan depends on the inherent characteristics of the biopolymer (degree of deacetylation, molecular weight, and crystallinity) and the experimental factors: chitosan dose, pH, temperature, etc.

The adsorption of Cr^III onto chitosan and activated carbon has been the subject of few studies, likely because of its nutritional character. In aqueous solutions, Cr^III may exist as Cr(OH)²⁺, Cr(OH)₂⁺, Cr₃(OH)₄⁵⁺, Cr(OH)₃, and Cr(OH)₄⁻ depending on pH.

In this work, the dynamic and equilibrium adsorption of hexavalent and trivalent chromium onto chitosan, olive stone waste-derived activated carbon, and their blend was investigated, and the effects of some operating factors (temperature, dose of adsorbent, and pH) on the adsorption processes were assessed. Moreover, the recovery of chromium cations in different aqueous solutions was evaluated.

Materials and experimental procedures

Chitosan

Dried shells of shrimp were finely ground and demineralized with HCl solution (0.55 N). Then, they were deproteinized with NaOH solution (0.3 N). The residue (about 21 wt%) was identified to chitin.

The chitin was subjected to a deacetylation treatment using a hot (80 °C) and strong NaOH aqueous solution (40%). The optimum time of the treatment was of 18 h. The product, which manifested as white flakes, was identified to chitosan. The ratio chitosan/chitin was estimated to be 83 wt%.

The deacetylation degree (DD) and the molecular weight (MW) of the prepared chitosan, determined by adopting the experimental procedures described by Brugnerotto et al. (2001) and Tolaimate et al. (2003) and by Kumar (2000), respectively, are given in Table 1.

Table 1 Some of the characteristics of the prepared chitosan and activated carbon

Olive stone waste-derived activated carbon

The olive stone waste used was a by-product of an olive oil semi-industrial mill located nearby the city of Marrakech (Morocco). The waste was washed with hexane in order to eliminate residual oil. A portion of dried oil-free waste (30 g) was introduced in a stainless steel reactor placed in a laboratory electrical furnace and heated at 10 °C/min from room temperature to the activation temperature of 850 °C. The activation was performed by using a steam flow (0.1 mL/min) for 60 min. The activated carbon obtained was abundantly washed with distilled water, oven dried (105 °C), and grounded to pass through a 50-μm sieve. The main characteristics of the activated carbon prepared are given in Table 1. As compared to some olive stone-derived activated carbons (Baccar et al. 2009), the prepared material had a moderate BET specific surface area, and appreciable micro- and meso-porous structure.

Blend of activated carbon and chitosan (composite)

Films of blends of activated carbon and chitosan were prepared by adopting a method similar to solvent casting. For this purpose, a portion of chitosan (5 g) was dissolved in 250 mL of acetic acid solution (2% v/v) and stirred for 4 h. Then, dried activated carbon (1.25 ≤ mass ≤ 20 g; particle size < 50 μm) was introduced in the solution of chitosan. The dispersion was stirred at 200 rpm for 3 h, and poured in a Petri dish, which was kept at room temperature until total evaporation of the liquid. The film formed was abundantly washed with distilled water, dried, and stored until use. The relative mass of activated carbon in the film of the composite studied was fixed to 60 wt% based on the results given in Fig. 1.

Fig. 1

Variations of the maximum amounts of Cr^VI and Cr^III (q_e^max) retained by composites (chitosan-activated carbon) versus the relative content of activated carbon

To have an insight into the structure of the composite prepared, and the linkage between particles of activated carbon and chitosan chains, samples of the composite were analyzed with Fourier transform infrared spectroscopy and examined by using a scanning electron microscope as reported below. The mix of activated carbon with chitosan resulted in noticeable changes in the shape and the position of the infrared bands located in the range of 3700–3100 cm⁻¹ assignable to the stretching vibrations of O–H of the basic materials and N–H of chitosan (Fig. 2). Moreover, the intensities of the bands of N–H bond of chitosan, which occurred in the interval of 1760–1520 cm⁻¹, and the band at 1230–960 cm⁻¹ linked to C–O bonds of chitosan and activated carbon diminished (Fig. 2). The microscopic examinations of the composite showed that pores and particles, which formed the structure of activated carbon, were almost hidden, and a somewhat smooth surface resembling to that of chitosan appeared (Fig. 3). In view of these observations, particles of chitosan and activated carbon were bonded to each other by means of amino and oxygenated functional groups, and particles of activated carbon were coated with chitosan. Details regarding these effects are given herein.

Fig. 2

FT-IR spectra of the composite studied (40 wt% chitosan) and its basic constituents (chitosan, activated carbon)

Fig. 3

SEM micrographs of the composite prepared (a) and its basic constituents b chitosan and c activated carbon

Kinetic experiments

Dispersions composed of 10 mg of adsorbent (chitosan, activated carbon, or composite), 3.5 mL of a stock solution of potassium dichromate (2.829 g/L) or chromium (III) nitrate, 9-hydrate (3.8461 g/L), and distilled water (46.5 mL) were continuously stirred at 200 rpm, and their pH was kept constant (6.5). The temperatures tested were 293, 313, and 333 K. Samples were drawn out from solutions at regular times and filtered on a cellulose nitrate membrane (pore size 0.45 μm). The quantity of hexavalent or trivalent chromium in the supernatant (C_t) was measured with a Secomam Anthelie UV-Visible spectrophotometer operating at the wavelength 540 nm. For this goal, 1,5-diphenylcarbazide was used as a complexing agent (e.g., Basset et al. 1986), and a curve expressing the Beer-Lambert law was plotted. The retained amount of hexavalent or trivalent chromium per mass of adsorbent (q_t) was deduced according to the relation:

$$ {q}_t=\frac{\left({C}_o-{C}_t\right).V}{m} $$

(1)

C_o (70 mg/L) and C_t are the initial and the instantaneous concentrations of Cr^IV or Cr^III in the supernatant, V the volume of solution and m the mass of adsorbent.

The kinetic data were fitted with the nonlinear forms of the pseudo-first-order (Eq. 2) and the pseudo-second-order (Eq. 3) kinetic equations:

$$ {q}_t={q}_e\ \left(1-{e}^{-{k}_1t}\right) $$

(2)

$$ {q}_t=\frac{k_2{q_e}^2t}{1+{k}_2{q}_et} $$

(3)

k₁ and k₂ are the pseudo-first-order and the pseudo-second-order rate constants, respectively; q_e is the uptake quantity of the adsorbate at equilibrium.

For the data fitting, the solver add-in with the Microsoft Excel spreadsheet (Brown 2001; Kumar 2006) was used, and the best fit was evaluated by computing the values of R²:

$$ {R}^2=1-\frac{\sum {\left({q}_t-{q_t}^c\right)}^2}{\sum {\left({q}_t-{q_t}^a\right)}^2} $$

(4)

q_t and q_t^c are the experimental and calculated instantaneous retained quantities of adsorbate. q_t^a is the average of the values of q_t.

The rate-limiting steps were determined by using the intraparticle (Qiu et al. 2009) and the liquid film (Sağ and Aktay, 2000) diffusion models (Eqs. 5 and 6, respectively).

$$ {q}_t={k}_i{t}^{0.5}\kern0.5em \left({k}_i,\mathrm{the}\ \mathrm{intraparticle}\ \mathrm{diffusion}\ \mathrm{constant}\right) $$

(5)

$$ Ln\frac{C_t}{C_o}=-{k}_lt\kern0.5em \left({k}_{\mathrm{l}},\mathrm{constant}\ \mathrm{related}\ \mathrm{to}\ \mathrm{the}\ \mathrm{liquid}\ \mathrm{film}\ \mathrm{diffusion}\ \mathrm{coefficient}\right) $$

(6)

Measurement of adsorption isotherms

Solutions composed each of 10 mg of adsorbent and 50 mL of an aqueous solution of hexavalent or trivalent chromium (5–70 mg/L) were used for the plot of the adsorption isotherms. The solutions of pH = 6.5 were stirred (200 rpm) and maintained at 293, 313, or 333 K for 4 h. The retained amounts of chromium at equilibrium (q_e) were determined following the experimental procedure described above.

The isotherms were fitted to the nonlinear forms of the models of Langmuir (Eq. 7) (Langmuir 1918), Freundlich (Eq. 8) (Freundlich 1907), and Temkin (Eq. 9) (Temkin and Pyzhev 1940).

$$ {q}_e=\frac{K_L{q_e}^{max}{C}_e}{1+{K}_L{C}_e} $$

(7)

$$ {q}_e={K_F{C}_e}^{1/n} $$

(8)

$$ {q}_e=\frac{RT}{b}\mathit{\ln}\left({K}_T{C}_e\right) $$

(9)

q_e (mg/g), the uptake amount at equilibrium; K_L, Langmuir constant; q_e^max (mg/g), adsorption capacity; K_F (mg/g (L/mg)^1/n), Freundlich constant; K_T (L/mg), Temkin constant; b (g J/mg mol): constant; R: gas constant (8.31 JK⁻¹ mol⁻¹); T (K): temperature.

The fitting was realized by using the solver add-in with the Microsoft Excel spreadsheet, and checked by calculating R² (Eq. 4).

Regeneration of chromium cations

The release of Cr cations, which were adsorbed on chitosan and activated carbon, was followed at 25 °C in acid (HCl or CH₃COOH, 2 N), neutral (freshly distilled water), and alkaline (NaOH, 2 N) solutions. In this context, it could be noted that chitosan did not solubilize in the used solution of hydrochloric acid because of the salting out effect (Rinaudo et al. 1999). For the composite, the recovery tests were carried out in water (pH = 6.5; T = 25 °C) because of the good stability of the films in such conditions. The released amount was measured after different cycles of washing (2 h per cycle).

Regarding the regeneration procedure, a portion of 30 mg of dried Cr-saturated adsorbent was introduced in 150 mL of a solution among the aforementioned ones, and stirred by a universal shaker SM-30 Edmund Buehler GmbH for 2 h. The dispersion was filtered, and the amount of the released cations was measured by using a UV-Visible spectrophotometer, as reported previously.

Characterization techniques

Pristine and Cr-loaded adsorbent as well as samples, which were subjected to the regeneration tests, were analyzed with Fourier transform infrared spectroscopy. For this purpose, thin discs composed of dried adsorbents (1 mg) and KBr (99 mg) were shaped and analyzed with a PerkinElmer spectrophotometer operating in the range 4000–400 cm⁻¹. The deconvolution of the IR bands was carried out with the PeakFit v4.12 software (peak type: Gaussian shape). The correlation coefficient (R²), the standard error (SE), and F-statistic were used as parameters for the best fit evaluation.

The microstructure of the adsorbents was examined with a JEOL JMS 5500 scanning electron microscope equipped with an EDAX Falcon spectrophotometer. The examinations were carried out on carbon-coated samples.

Results and discussion

Effect of contact time

Adsorption of hexavalent chromium

As can be deduced from the kinetic curves of the adsorption of Cr^VI species on the three materials used (Fig. 4), almost 90% of the initial amount of chromium cations was retained in about 1 h of contact, and saturation was reached in less than 2 h. Moreover, the uptake globally decreased with increasing temperature. Based on these results, the adsorption processes were fast and seemed to be exothermic. Even at this earlier stage of the study, it could be mentioned that adsorption ability of the composite exceeded largely those of the basic constituents.

Fig. 4

Kinetic curves of the adsorption of Cr^VI and Cr^III ions, at different temperatures, onto chitosan (a, a′), activated carbon (b, b′), and composite (c, c′). (a–c) Cr^VI. (a′–c′) Cr^III

Taking into consideration the values of q_e^exp, q_e^c, and R² given in Table 2, the kinetics of adsorption on chitosan as well as on the composite fitted better the pseudo-second-order equation, and the rate constant (k₂) varied in the range of (1–5.5)10⁻³ min⁻¹ mg⁻¹ g. For activated carbon, the best fitting kinetic model depended on the operating temperature (Table 2). At 313 and 333 K, the pseudo-first-order equation seemed to be the best fitting kinetic model, and the rate constants were to be 1.51 × 10⁻² and 2.11 × 10⁻² min⁻¹, respectively.

Table 2 Values of the parameters of the pseudo-first (PFKE) and pseudo-second (PSKE) order kinetic equations related to the adsorption of Cr^VI onto chitosan, olive stone-derived activated carbon, and the composite chitosan-activated carbon. R² is the nonlinear regression fitting coefficient

To get more insight in the adsorption kinetic processes, the curves q_t = f(t^0.5) and Ln(C_t/C_o) = f(t) were plotted, and the values of the parameters of their equations (Eqs. 5 and 6) are given in Table 3. The curves manifested a linear evolution (0.8 < R² ≤ 1), but they did not go through the origin (C ≠ 0; C′ ≠ 0), except in the case of the adsorption of the cations on activated carbon at 313 and 333 K. In sight of these results, the rates of the adsorption of Cr^VI species on chitosan and composite depended on the intraparticle and the liquid film diffusion. The rates of the adsorption of the cations on activated carbon at 313 and 333 K were essentially ruled by diffusion through the liquid film layer. Considering the values of C (Table 3), the effect of the thickness of the liquid boundary layer around particles of adsorbents followed the order composite > chitosan > activated carbon. It was believed that the buildup of the cations at the surfaces of the particles depended on the dimension and/or the quantity of open pores.

Table 3 Values of the parameters of the models used for the determination of rate-limiting steps and their related fitting coefficients

Adsorption of trivalent chromium

Referring to the experimental kinetic curves of the adsorption of Cr^III species on the materials studied (Fig. 4), up to about 85% of the starting amount of the cations in solution was retained in less than 1 h, and the time required for equilibrium exceeded 2 h. Moreover, the adsorption processes seemed to be endothermic. Based on the values reported in Table 4, especially those of R², q_e^exp, and q_e^c, the kinetic curves of the adsorption of Cr^III species onto chitosan and activated carbon rather followed the pseudo-second-order model, and k₂ (chitosan) ≈ 2k₂ (activated carbon). On the other hand, the kinetic curves of the retention of Cr^III species by the composite well fitted the pseudo-first-order equation, particularly at 298 and 313 K, and the kinetic constants were estimated to be 3.86 × 10⁻² and 3.39 × 10⁻² min⁻¹, respectively.

Table 4 Values of the parameters of the kinetic models used for the analysis of the kinetic data of the adsorption of Cr^III ions on the adsorbents studied. R² is the nonlinear regression fitting coefficient

Considering the fitting coefficients given in Table 3, the rates of adsorption of Cr^III species on the materials studied seemed to be controlled by diffusion within particles together with diffusion through the boundary liquid film. At lower temperature, the kinetics of the adsorption on activated carbon and composite were chiefly controlled by intraparticle diffusion. Referring once again to the data in Table 3, the ion mobility within particles and across boundary layers decreased in the following order: composite > activated carbon > chitosan. The data also indicated that for chitosan and activated carbon, the effect of the thickness of the boundary layer, represented by constant C, on the cation diffusion became appreciable with increasing temperature.

Equilibrium studies

Adsorption isotherms and thermodynamic data

Hexavalent chromium

The isotherms of the adsorption of Cr^VI species on the three materials studied manifested L shape (Fig. 5 (a–c)), which is commonly considered as an indication of a high affinity towards adsorbent sites, and the presence of micropores. Also, it may express the fact that some chemical species adsorbed in a flat position without experiencing a strong competition from solvent molecules (Calvet 1989). To know more about the interactions between Cr^VI species and adsorbents, the adsorption isotherms were fitted to the commonly used models (Eqs. 7–9). Considering for instance the isotherms at 298 K, and based on the values of the nonlinear fitting coefficient (R²) given in Table 5, the isotherms, which involved chitosan and composite as adsorbents, well correlated to the model of Temkin. In contrast, the isotherm related to the cation adsorption on activated carbon was better described by Langmuir’s model. Considering the hypotheses of the models used, the adsorption of the cations on chitosan and composite occurred on heterogeneous sites. Moreover, the heat of adsorption decreased linearly with coverage, because of the presence of lateral repulsive interactions between adsorbate species. In the case of activated carbon, the interactions between adsorbed cations were considered negligible, and adsorption occurred on homogenous sites. Moreover, the heat of adsorption was supposed to be constant.

Fig. 5

Isotherms of the adsorption of Cr^VI and Cr^III ions onto chitosan (a, a′), activated carbon (b, b′), and composite (c, c′). (a–c) Cr^VI. (a′–c′) Cr^III

Table 5 Values of the parameters of the adsorption isotherm models used, and the nonlinear correlation coefficient (T = 298 K)

The variation of the maximum amount of Cr^VI species (q_e^max) retained by chitosan as a function of temperature showed a nonlinear evolution, and the higher value (about 50 mg/g) was determined at 298 K. For activated carbon and composite, the variations of q_e^max versus temperature evolved linearly according to the following equations:

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{activated}\ \mathrm{carbon}\right)=187.71\hbox{--} 0.49\ T $$

(10)

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{composite}\right)=162.2\hbox{--} 0.05\ T $$

(11)

Globally, q_e^max decreased with increasing temperature, and it increased in the order q_e^max (activated carbon) < q_e^max (chitosan) < q_e^max (composite). The adverse effect of the increase of T on q_e^max was linked to the exothermic character of the adsorption processes (see hereafter).

The Gibbs free energies (ΔG^°_T) of the adsorption processes, computed by using the thermodynamic relation:

$$ \Delta {G^{{}^{\circ}}}_T=-{RTLnK}_e $$

(12)

(R: gas constant (8.31 JK⁻¹ mol⁻¹); T: operating temperature (K); K_e: equilibrium constant; K_e = 10³q_e^maxK_L; q_e^max and K_L were determined by using Eq. 7 (Moura et al. 2016)) showed that Cr^VI species adsorbed spontaneously on the three materials used, and Cr cations manifested a relative high affinity for chitosan and composite (Table 6). The heat (ΔH^°_T) and entropy (ΔS^°_T) (Table 6), determined according to the relation:

$$ Ln{K}_e=-\frac{\Delta {H}^{{}^{\circ}}}{R\ T}+\frac{\Delta {S}^{{}^{\circ}}}{R} $$

(13)

indicated that the adsorption processes on chitosan and composite were exothermic (ΔH^°_T < 0), and the dispersal energies, represented by ΔS^°_T, were different.

Table 6 Thermodynamic data of the adsorption of Cr^VI and Cr^III ions on the materials studied

Trivalent chromium

The isotherms of the adsorption of Cr^III species on chitosan and composite displayed L shape (Fig. 5 (a′ and c′), but those plotted for activated carbon (Fig. 5 (b′)) slightly deviated from this shape. For the latter case, substantial active sites became available as the concentration of Cr^III increased. Considering the fitting coefficients listed in Table 5, the adsorption isotherms of chitosan at 298 K better fitted the model of Freundlich. However, those obtained for activated carbon and composite well followed the equation of Temkin. These results suggested that adsorption took place on heterogeneous sites, and interactions were developed between adsorbed species.

For all studied materials, the increase of temperature favored cation adsorption, and the maximum uptake varied with temperature according to these relations:

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{chitosan}\right)=-69.38+0.26\times T $$

(14)

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{activated}\ \mathrm{carbon}\right)=-13.63+0.08\times T $$

(15)

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{composite}\right)=-223.55+0.86\times T $$

(16)

These results showed that the maximum uptakes of the cations adsorbed on chitosan and activated carbon were somewhat similar, but that related to the adsorption on the composite was much higher (q_e^max (composite) ≈ 3q_e^max (chitosan/activated carbon).

Taking into consideration the thermodynamic data given in Table 6, the adsorption of Cr^III species on chitosan and composite occurred spontaneously (ΔG^°_T < 0). Moreover, the adsorption processes were endothermic (ΔH^° > 0), and the dispersal energy increased as a result of the adsorption on chitosan and composite.

Effects of pH and dose of adsorbents

Up to about pH = 8, the maximum amounts of Cr^VI retained by chitosan and activated carbon declined with increasing pH (Fig. 6):

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{chitosan}\right)=135.5-13.1\mathrm{pH} $$

(17)

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{activated}\ \mathrm{carbon}\right)=80.9-8.3\mathrm{pH} $$

(18)

Fig. 6

Variations of the maximum uptakes of Cr^VI and Cr^III ions by chitosan, activated carbon, and composite versus pH

Beyond pH = 8, q_e^max still decreased linearly, but the slope was relatively low.

For the composite, the change of q_e^max versus pH (Fig. 6) obeyed the relation:

$$ {q_e}^{\mathrm{max}}\left(\mathrm{mg}/\mathrm{g}\ \mathrm{of}\ \mathrm{composite}\right)=121.8-8.2\mathrm{pH} $$

(19)

Based on the speciation diagram shown in Fig. 7, the relative amounts of Cr₂O₇²⁻ and HCrO₄⁻ in solution regressed drastically as pH exceeded 5, when that of CrO₄²⁻ increases substantially. As pH > 8, only CrO₄²⁻ ions are present. Thus, the deflexion of the curves q_e^max = f(pH) was presumably linked to the change in the nature and the amount of adsorbate species. Taking into consideration the pKa (6.5) of chitosan, the protonated amine form (–NH₃⁺), which is considered as the main potential site for the retention of the anions cited above, should be predominant in acid solution. But, the amount of such sites decreased with increasing pH, thereby the uptake of Cr^VI species declined. Given the similarities between the curves of chitosan, activated carbon, and composite, it could be believed that the main active sites of activated carbon and composite consisted of positively polarized moieties, which amounts regressed with increasing pH. In fact, it was shown by using X-ray photoelectron spectroscopy that phenol, anhydride, and carbonyl/quinone were the main functional groups found at the surface of the activated carbon prepared. Concerning the active sites of the composite, a detailed study done by FT-IR spectroscopy is reported below.

Fig. 7

Speciation diagrams, showing the variations of the relative fractions of stable chromium species versus pH

Referring once again to Fig. 6, the maximum uptakes (q_e^max) of Cr^III by chitosan and activated carbon increased linearly with increasing pH. The relations followed at pH > 6.5 are

$$ {q_e}^{\mathrm{max}}=-5.97+1.64\mathrm{pH} $$

(20)

$$ {q_e}^{\mathrm{max}}=-6.91+1.72\mathrm{pH} $$

(21)

In the case of the composite, q_e^max increased with pH according to the equation:

$$ {q_e}^{\mathrm{max}}=3.12+2.23\mathrm{pH} $$

(22)

In Cr^III aqueous solutions, CrOH²⁺ and Cr₃(OH)₄⁵⁺ ions almost vanished at around pH 7, and the relative amount of Cr(OH)₂⁺ was at its maximum (Fig. 7). Beyond pH 7, the amount of the latter species diminished and that of Cr(OH)₃ intensified. In view of these observations, it was believed that at 6.5 < pH < 7.5, Cr^III attached to hydroxyls and amine moieties of chitosan/composite, and to the oxygenated sites of activated carbon/composite mainly as Cr₃(OH)₄⁵⁺ and Cr(OH)₂⁺. In the pH range of 7.5–9, a part of Cr^III absorbed as Cr(OH)₂⁺, and a part was involved in the formation of Cr(OH)₃. In strong alkaline solutions (pH > 9), the amount of Cr(OH)₃ regressed, whereas that of the complex Cr(OH)₄⁻ intensified. The continual increase of q_e^max at high pH could be taken as an indication of the retention of Cr(OH)₃ and Cr(OH)₄⁻ species by the three adsorbents used.

As can be deduced from the curves, shown in Fig. 8, the uptake of chromium cations increased significantly up to about 200 mg/L of chitosan or composite, and 150 mg/L of activated carbon. For further additions of adsorbents, the uptake slightly augmented. It was believed that, due to the quantitative increase of the dose of adsorbents, a crowding effect took place. Thus, Van der Waals attractive forces developed between adsorbent particles, and consequently, some of the active sites were unreachable by chromium cations.

Fig. 8

Changes of the maximum retained amounts of Cr^VI and Cr^III ions versus the dose of the adsorbents used. a Chitosan. b Activated carbon. c Composite

Regeneration of Cr^VI and Cr^III ions

The results of the regeneration of Cr^III species retained by chitosan (Fig. 9) showed that only 5% of the ions were desorbed in alkaline aqueous solution. In contrast, the release did not happen in neutral and acid aqueous solutions. Thus, Cr^III species were tightly bound to the active sites of chitosan, likely by complexation. Regarding the regeneration of hexavalent chromium, the results showed that 78 and 10% of the uptake amount were released in acid aqueous solution and distilled water, respectively (Fig. 9). Nonetheless, desorption did not occur in alkaline aqueous solution. It was believed that in acid solution, CrO₄²⁻ ions detached from chitosan to react according to these reactions characterized by high equilibrium constants:

$$ {\displaystyle \begin{array}{cc}{{\mathrm{Cr}\mathrm{O}}_4}^{2-}+{\mathrm{H}}^{+\kern1.5em }==={{\mathrm{H}\mathrm{CrO}}_4}^{-}& K=3.16\times 1{0}^6\\ {}2{{\mathrm{Cr}\mathrm{O}}_4}^{2-}+2{\mathrm{H}}^{+\kern0.5em }==={\mathrm{Cr}}_2{{\mathrm{O}}_7}^{2-}+{\mathrm{H}}_2\mathrm{O}& K=1.3\times 1{0}^{16}\end{array}} $$

Fig. 9

Amounts of Cr^VI and Cr^III ions released by chitosan, activated carbon, and composite in different operating conditions

It may be noticed that HCrO₄⁻ and Cr₂O₇²⁻ are the main stable species at the operating pH (Fig. 7). The release of Cr^VI species could be considered as an indication of the presence of weak bonds between these species and chitosan sites.

As can be deduced from Fig. 9, the quantity of Cr^VI species, which released by activated carbon in the tested aqueous solutions, decreased as follows: AcOH > NaOH > H₂O ≈ HCl. In contrast, the removed amount of Cr^III species evolved in the order AcOH > HCl > H₂O ≈ NaOH. These results showed that both chromium cations were quantitatively desorbed in acetic acid solution, and disparities were observed for the remaining aqueous solutions. These differences were presumably linked to the nature of the interactions between desorbed Cr species and solute derivative species (AcO⁻, OH⁻, H⁺, Na⁺, and Cl⁻) and/or to the presence of competing adsorption. Considering once again Fig. 9, the release of Cr^III and Cr^VI species by the composite in aqueous solution of pH = 6.5 intensified with increasing the number of washing cycle. The desorbed amount increased by about 1.5 times per cycle. In consideration of these results, adsorption of Cr species on composite seemed to be a reversible process which extent possibly depended on the law of mass action. Details related to the interactions between Cr cations and composite are given hereafter.

Characterization and mechanisms of adsorption

Chitosan

As a result of the adsorption of Cr^VI and Cr^III species, the broad and intense infrared band at about 3440 cm⁻¹ due to the overlapping of the stretch of O–H and N–H of chitosan splitted into two bands at 3482 and 3413 cm⁻¹ (Fig. 10). In addition, a shoulder appeared at about 3545 cm⁻¹. These observations were taken as an indication of the involvement of hydroxyls and amino functional groups in the fixation of chromium ions. The band at about 2983 cm⁻¹ assignable to the stretch of the C–H bond of CH₂ was not disturbed. As a result of the adsorption of the ions, intense and sharp bands manifested at about 1640 and 1618 cm⁻¹. These frequencies were associated to the vibration of the N–H bond of the amino groups. It is worth noting that due to the adsorption of Cr^VI species, the spectrum displayed a medium band centered at about 1543 cm⁻¹ related to the protonated amine group. The latter frequency shifted towards 1523 cm⁻¹ as Cr^VI species were desorbed. This fact supported the implication of the protonated amine groups in the adsorption process of Cr^VI. As far as the infrared analysis was concerned, the bands of chitosan in the range 1500–1000 cm⁻¹ were not the subject of any shift.

Fig. 10

Infrared spectra of chitosan after adsorption (Ads) and desorption (Des) of Cr^VI and Cr^III ions

Referring to Fig. 10, the infrared spectrum of Cr^VI-loaded chitosan (chitosan-Cr^VI (Ads)) showed the presence of well-resolved bands at about 936 and 765 cm⁻¹, which were assigned to the vibrations of Cr=O and Cr–O bonds, respectively (Kyzas et al. 2009). As expected, these bands vanished after desorption of Cr^VI species (chitosan-Cr^VI (Des)). On the other hand, it should be noted that the spectrum of Cr^III-loaded chitosan (chitosan-Cr^III) displayed, in addition to the frequencies related to chitosan, an extra weak band at 827 cm⁻¹ considered as the main fingerprint of the interaction between Cr^III and chitosan.

Considering that a macromolecular chain of the studied chitosan (MW = 325,000 g/mol) consisted on average of about 1800 units of D-glucosamine, and the amount of the protonated amine groups at pH = 6.5 (about 50% of the initial amount of NH₂ groups), the loading for Cr^VI, determined on the basis of the Langmuir maximum retained amounts, varied between 15 and 35%. In the case of Cr^III, the estimated loading was in the range 3–6%. These results showed that a quantitative amount of the potential reactive groups of glucosamine were not directly involved in the adsorption process, presumably because of the aggregation of the chains of chitosan, as mentioned above.

Activated carbon

The analysis of the FT-IR spectra shown in Fig. 11a indicated that due to the adsorption of Cr^VI species on activated carbon and to their desorption, (i) the intensity of the prominent band at about 3429 cm⁻¹ linked to the OH stretching vibration substantially regressed; (ii) the band at 1844 cm⁻¹, which is assigned to the stretch of the C=O bond, vanished; (iii) the intensity of the band at 1568 cm⁻¹ associated to the vibration of C=C bond of the aromatic rings drastically diminished; (iv) the band at 1458 cm⁻¹ due to the vibration of C–O–H disappeared; (v) The intensity of the band at 1088 cm⁻¹ assignable to the vibration of C–O–C regressed; and (vi) the bands at 879 and 783 cm⁻¹, which are attributable to the C–H bond vibration of the aromatic ring, vanished. In sight of these observations, Cr^VI species, which were mainly present as HCrO₄⁻ (Fig. 7), adsorbed on activated carbon by developing interactions with π electrons of aromatic rings on one hand and oxygenated sites (C–OH, C=O, C–O–C) on the other hand. In line with some authors (Tran et al. 2017), we believed that Cr^VI was partially reduced into Cr^III as a result of the interaction of HCrO₄⁻ with π electrons, and it was also adsorbed as HCrO₄⁻ on positively polarized oxygenated sites. The formed Cr^III ions were released in aqueous solution and/or immobilized at the surface of the particles of activated carbon. Considering the aforementioned desorption graph related to the activated carbon (Fig. 9), the quantitative desorption of chromate in acetic acid solution was probably due to the displacement of immobilized Cr^III ions by acetate. Regarding the substantial release of chromium in sodium hydroxide solution, it seemed that OH⁻ ions interacted with adsorbed HCrO₄⁻ for forming CrO₄²⁻, the stable species in such a solution (Fig. 7). Hydroxyls in the alkaline solution could also react with immobilized Cr^III for forming Cr(OH)₄⁻, which is the stable species at the used pH (Fig. 7).

Fig. 11

Infrared spectra of activated carbon after adsorption (Ads) and desorption (Des) of Cr-cations. a Cr^VI. b Cr^III

The adsorption of Cr^III species on activated carbon as well as the cation desorption had a noticeable effect on the intensity of the infrared band at about 3433 cm⁻¹, which is assigned to the stretching vibration of O–H (Fig. 11b). This fact indicated that hydroxyls of activated carbon were involved in the retention of Cr^III species, likely by means of attractive electrostatic forces. Electrostatic attractions also manifested between Cr^III species and oxygenated moieties of the framework of activated carbon. They were revealed by the marked changes of the intensities of the bands of C=O (1836 cm⁻¹), C–O–H (1462 cm⁻¹), and C–O–C (1080 cm⁻¹). Given the appreciable release of Cr^III ions in acid solutions (Fig. 9), it appeared that attractive forces between Cr^III species and the widespread protonated sites of activated carbon formed in strong acid solutions become ineffective. Considering the change of the intensity of the IR band at about 1578 cm⁻¹ related to the C=C bond (Fig. 11b), π electrons of the aromatic rings were also involved in the retention of Cr^III species. The implication of π electrons in the electrostatic interaction of Cr cations may be supported by the drastic reduced of the band at 879 cm⁻¹ associated to the C–H bond of aromatic rings.

Composite

Considering the FT-IR spectra shown in Fig. 12a, the shape of the prominent band in the range of 3670–3100 cm⁻¹, linked to the stretching vibrations of O–H and N–H, has remained unchanged following the adsorption and desorption of Cr^VI species. Nevertheless, the intensity of the band slightly diminished. So, only a small fraction of the aforementioned functional groups took part in the retention of Cr^VI species. The main part of the active sites, i.e., amino groups and hydroxyls, were involved in the linkage of chitosan chains to the activated carbon particles, as previously mentioned. Referring once again to Fig. 12a, the intensity of the band in the range of 1300–1000 cm⁻¹, composed of frequencies of C–O–H and C–O–C bonds, drastically decreased after adsorption/desorption of Cr^VI species. Thus, hexavalent chromium, mainly present as HCrO₄⁻ and CrO₄²⁻, was essentially adsorbed on polarized C–O–H and C–O–C sites possibly by developing electrostatic attractive forces. The abundance of these sites at the surfaces of particles was responsible for the quantitative retention of Cr^VI species by the composite. In view of these results, it could be postulated that because of the presence of activated carbon particles, the chains of the biopolymer disaggregated, and changed their conformations in such a way that the C–O–H and C–O–C bonds were oriented towards the outside of the particles. Likely for this reason as well as for the links between chitosan chains and activated carbon via amino and hydroxyl functional groups, chitosan-coated particles of the composite manifested a spheroidal shape such as seen in Fig. 13a. In the presence of Cr^VI species, the spheroidal grains drastically disappeared, and coarse particles with different morphologies took place (Fig. 13b). It was believed that Cr^VI species (HCrO₄⁻, CrO₄²⁻) mainly placed at the surfaces were involved in the growth process of the grains.

Fig. 12

Infrared spectra of the composite after adsorption (Ads) and desorption (Des) of Cr-cations. a Cr^VI. b Cr^III

Fig. 13

SEM micrographs and EDAX spectra of the composite before (a) and after adsorption of Cr^VI (b) and Cr^III (c)

As a result of the adsorption of Cr^III species on the composite, the bands at about 3476, 1636, 1386, and 631 cm⁻¹ intensified, and the shape of those placed in the range of 750–450 cm⁻¹ was modified (Fig. 12b). A close examination of the main bands in the domain of 3700–3100 cm⁻¹ of the spectra obtained before and after adsorption (Fig. 14) showed that the frequencies associated to the stretching vibrations of O–H of chitosan (3554 cm⁻¹) and activated carbon (about 3425 cm⁻¹) experienced a shift of about 30 cm⁻¹. The band linked to the stretch of N–H (about 3488 cm⁻¹) remained almost unchanged, and its intensity increased. The microscopic examination of Cr^III-loaded composite revealed a featureless structure (Fig. 13c) looking like that of chitosan. Based on these investigations, Cr^III species preferentially adsorbed on hydroxyls groups, which were mainly involved in the assemblage of the composite particles. Therefore, the chains of chitosan were more flexible, and the original spheroidal shape of the particles disappeared.

Fig. 14

Deconvoluted bands of the infrared spectra of the composite and Cr^III-loaded composite, and curve fitting parameters

Conclusion

The results of this study showed that (i) attractive electrostatic forces were established between chitosan chains and particles of activated carbon, involving the amino functional groups of the biopolymer, the hydroxyls, and oxygenated moieties of both constituents. (ii) The best fitting kinetic model (pseudo-first or pseudo-second order) for the kinetic data of the adsorption of Cr^VI and Cr^III species on chitosan, activated carbon, and their blend depended on the nature of the adsorbent used as well as on the operating temperature. The adsorption rates were limited by diffusion within particles and/or through the liquid layer. (iii) Adsorption of chromium cations occurred spontaneously. A part from the adsorption of Cr^III species on activated carbon, the adsorption of Cr cations took place on heterogeneous sites. The Langmuir maximum adsorption capacities of Cr^VI species were as follows: composite > chitosan > activated carbon. In the case of Cr^III species, the maximum amount retained by the composite was about three times higher than that of chitosan or activated carbon. (iv) The increase of the operating temperature or pH had an adverse effect on the adsorption of Cr^VI species, whereas it had a positive impact on the retention of Cr^III species. These facts were related to the nature of the electrostatic interactions between the active sites of the adsorbent and Cr species, identified on the basis of the speciation diagrams. (v) Cr^VI species (mainly HCrO₄⁻) and Cr^III, mostly present as Cr(OH)₂⁺ and Cr(OH)₄⁵⁺, were adsorbed on amino groups and hydroxyls of D-glucosamine units of chitosan. Protonated amines of the biopolymer contributed to the retention of Cr^VI species. The release of Cr^III species was inappreciable in the solutions used. However, the recovery of Cr^VI species was quantitative (78%) in acid solution. (vi) Cr^VI and Cr^III species adsorbed on oxygenated sites of activated carbon (C–OH, C=O, C–O–C), and interacted with π electrons of aromatic rings. Due to the latter effect, the hexavalent chromium was partially reduced to trivalent chromium. The recovery of both species was relatively important in acetic acid solution probably because of the involvement of the acetate in the desorption process. (vii) Hexavalent chromium species were retained by the composite by means of electrostatic forces developed between Cr^VI species and C–O–H and C–O–C sites of chitosan-coated solid particles. The adsorption of Cr^III species on the composite took place preferentially on hydroxyl groups, which were no more involved in the linkage of chitosan and particles of activated carbon. Cr^III or Cr^VI species, which were immobilized by the composite, were partly released in aqueous solution of pH = 6.5, and the released amount increased by about 1.5 times/cycle.