Equilibrium and dynamic study on hexavalent chromium adsorption onto activated carbon.

G Model

ARTICLE IN PRESS

HAZMAT-16167; No. of Pages 9

Journal of Hazardous Materials xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Equilibrium and dynamic study on hexavalent chromium adsorption onto activated carbon F. Di Natale a , A. Erto a,∗ , A. Lancia a , D. Musmarra b a b

Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Università di Napoli “Federico II”, Piazzale Tecchio 80, 80125, Napoli, Italy Dipartimento di Ingegneria Civile, Design, Edilizia e Ambiente, Seconda Università di Napoli, Via Roma 29, 81031, Aversa (CE), Italy

a r t i c l e

i n f o

Article history: Received 27 January 2014 Received in revised form 28 July 2014 Accepted 29 July 2014 Available online xxx Keywords: Hexavalent chromium Water Adsorption Equilibrium Kinetics Langmuir model

a b s t r a c t In this work, the results of equilibrium and dynamic adsorption tests of hexavalent chromium, Cr (VI), on activated carbon are presented. Adsorption isotherms were determined at different levels of pH and temperature. Dynamic tests were carried out in terms of breakthrough curves of lab-scale fixed bed column at different pH, inlet concentration and flow rate. Both the adsorption isotherms and the breakthrough curves showed non-linear and unconventional trends. The experimental results revealed that chromium speciation played a key role in the adsorption process, also for the occurrence of Cr(VI)-to-Cr(III) reduction reactions. Equilibrium tests were interpreted in light of a multi-component Langmuir model supported by ion speciation analysis. For the interpretation of the adsorption dynamic tests, a mass transfer model was proposed. Dynamic tests at pH 11 were well described considering the external mass transfer as the rate controlling step. Differently, for dynamic tests at pH 6 the same model provided a satisfying description of the experimental breakthrough curves only until a sorbent coverage around 1.6 mg g−1 . Above this level, a marked reduction of the breakthrough curve slope was observed in response to a transition to an inter-particle adsorption mechanism. © 2014 Elsevier B.V. All rights reserved.

1. Introduction In the past decades, the occurrence of hexavalent chromium, Cr(VI), in wastewater and natural water became a paradigm of industrial pollution [1]. For this reason, chromium is considered as a priority hazardous pollutant and the European Union defined severe environmental regulations to set the maximum level of hexavalent chromium allowed in industrial and civil wastewaters (200 ␮g L−1 ), as well as in superficial and underground water bodies (5 ␮g L−1 ). Moreover, because of the occurrence of redox reactions involving the harmful hexavalent chromium and the much less toxic trivalent one, Cr(III), limiting values were also defined for the total chromium concentration in waters [2]. These limits are usually ten times higher than those of hexavalent chromium [3]. Commonly adopted methods to remove chromium from industrial wastewaters include precipitation [4], membrane filtration [5], solvent extraction with amines [6], ion exchange [7], activated carbon adsorption [2,8], electro deposition [9] and biological processes [10]. Activated carbon adsorption appears as a reliable process to

∗ Corresponding author. Tel.:+39 081 7682236; fax: +39 081 5936936. E-mail address: [email protected] (A. Erto).

remove organic and inorganic pollutants from both wastewaters and groundwater and, more in general, to comply with the water quality standards [11,12]. In groundwater remediation, it actually find practical application in both pump-and-treat systems and permeable adsorbing barriers [13,14]. The adsorption of metallic ions on the surface of activated carbons is the result of the interactions between the aqueous solution and the different active sites on the carbon surface. The presence of different surface functional groups with either acid or basic properties is responsible for the surface hydrolysis of the activated carbons in aqueous solutions and for the preferential interaction with cations and anions, respectively. These structures are generically denoted as COx Hy and are mainly originated by the activation process of the raw material [15]. The presence of surface heteroatoms and the structure of the graphitic layer itself, as well as the ash properties, determine the occurrence of active sites acting either as Lewis acids or as bases, able to capture ions with base or acid character, or as nucleophilic/electrophilic active sites [15,16]. Furthermore, some of these groups, such as lactonic and phenolic, exploit reducing properties, which are also observed for the same graphitic layer. Lakatos et al. [17] pointed out the role of surface functional groups with reducing properties in determining the adsorption of chromium ions. During an adsorption experiment,

http://dx.doi.org/10.1016/j.jhazmat.2014.07.072 0304-3894/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: F. Di Natale, et al., Equilibrium and dynamic study on hexavalent chromium adsorption onto activated carbon, J. Hazard. Mater. (2014), http://dx.doi.org/10.1016/j.jhazmat.2014.07.072

G Model HAZMAT-16167; No. of Pages 9

ARTICLE IN PRESS F. Di Natale et al. / Journal of Hazardous Materials xxx (2014) xxx–xxx

2

Nomenclature [x] a BM c c0 D Dl G dp ε K kf m Q q Re s Sh T t U ω ωmax z

molar concentration of the species x (M) specific surface areas (1/m) dimensionless mass loading (–) chromium concentration (mg L−1 ) chromium concentration (mg L−1 ) chromium diffusivity (m2 s−1 ) fixed bed column dispersion coefficient (m2 s−1 ) Gibbs free energy of adsorption (kJ mol−1 ) particle diameter (m) bed void fraction (–) adsorption equilibrium constant in Langmuir model (1/M) fluid-particle mass transfer coefficient (m s−1 ) fluid viscosity (kg (ms)−1 ) sorbent mass (g) liquid flow rate (L h−1 ) adsorption capacity in the fixed bed column (mg g−1 ) particle Reynolds number, Udp / (–) fluid density (kg m−3 ) solid particle density (kg m−3 ) Sherwood number, kf dp /D (–) temperature (◦ C or K) time (s) superficial fluid velocity in the column (m s−1 ) equilibrium adsorption capacity (mg g−1 ) maximum adsorption capacity in Langmuir model (mg g−1 ) column longitudinal coordinate (m)

the occurrence of surface reduction reactions can be observed by measuring Cr(VI) and total chromium in solution. In general, commercially available activated carbons may be expensive and the use of low cost adsorbing materials has aroused great interest in the pertinent literature [18–20]. For the same reason, the optimization of the operating conditions for adsorption processes is of the utmost importance for a correct process design. The analysis of the most recent findings in the pertinent literature showed that chromium adsorption kinetics in packed bed columns was investigated in several papers considering activated carbons [21–23], resins [24] and by-products derived materials [25–29], either for Cr(VI) or Cr(III) ions. The simultaneous presence of Cr(VI) and Cr(III) species was only considered in the packed bed tests of Vinodhini and Das [29]. The occurrence of Cr(VI) reduction reactions to Cr(III) on the surface of carbons was also observed in equilibrium studies on activated carbon (e.g. [17,19,21]). Similarly, all the tests were carried out with a constant solution pH over time. The state of the art can benefit for the development of a model able to describe equilibrium and kinetic data by taking into account for the complex ion speciation, the effect of pH and the occurrence of Cr(VI)–Cr(III) redox reactions. In this work, the results of equilibrium and dynamic tests on the adsorption of chromium on a commercial granular activated carbon (GAC) are presented. Equilibrium studies were carried out by determining the adsorption isotherms in stirred batch reactors at different levels of pH and temperature. Equilibrium data were interpreted in light of a multi-component Langmuir model supported by a ion speciation analysis, which allowed to correlate the distribution of Cr(VI) and Cr(III) species, the solution pH and the temperature, with the experimental adsorption capacity. Once the adsorption equilibrium was assessed, dynamic tests were carried out in terms of breakthrough curves of a lab-scale fixed bed

column at different levels of pH, inlet concentration and flow rate. The results of adsorption dynamics allowed determining the intercorrelation between the main process parameters (e.g. flow rate and pH). Since the equilibrium model allowed estimating the instantaneous driving force for adsorption rate, an accurate dynamic model was proposed to describe the breakthrough curves and to address the adsorption rate controlling mechanism in the investigated conditions. 2. Materials and methods 2.1. Sorbent properties The sorbent used in this work is Aquacarb 207EATM GAC, provided by Sutcliffe Carbon. The morphological analysis showed that Aquacarb 207EATM is a mesoporous material with an average pore diameter of 467 nm and an average porosity of about 17%. The density of the solid is around 500 kg m−3 and a particle range between 0.85 and 1.18 mm was selected by mechanical sieving. Sorbent characterization also included the evaluation of the pHPZC [30] and the analysis of surface functional groups by Boehm’s titration analysis [15,31]. The measurement of pHPZC (equal to 8) and the Boehm’s titration method showed that Aquacarb 207EATM has a slightly basic character and that the acidic functional groups are mainly represented by quinones, phenols and lactones. Additional properties of the GAC are reported in Di Natale et al. [32]. 2.2. Equilibrium tests The experimental runs were carried out in batch mode and were conducted at constant temperature in a proportional-integralderivative (PID) controlled thermostatic oven. The chromium solution was obtained by the dissolution of a Cr(VI) salt (K2 Cr2 O7 , reagent grade) in distilled water. The chromium concentration in the initial solution varied in the range 5–50 mg L−1 , while the sorbent concentration spanned from 2 to 10 g L−1 . The pH of the solution was adjusted in the range 2–11 by addition of either nitric acid or potassium hydroxide. Preliminary kinetic tests showed that a reaction time of 48 h was sufficient to reach equilibrium conditions in all the experimental runs. When equilibrium conditions were achieved, total and Cr(VI) concentrations in solution and total chromium uptake on the sorbent surface were measured. To this aim, the solution was filtered in a ceramic filtering Hirsch funnel using a vacuum pump to assure a very rapid and efficient separation of the granular material. The filtered solution was then analysed for pH and chromium concentration while the carbon was leached with 1 M nitric acid solution to obtain the complete extraction of sorbed chromium. Following this procedure, a direct measure of the total chromium uptake on the solid surface can be determined and a comparison with the homologous quantity as calculated by a mass balance can be made. The accuracy of the experimental runs was checked by allowing a maximum error of 5% in the chromium mass balance. Blank tests were run in parallel on chromium solutions, without addition of sorbent, for the sake of comparison. 2.3. Dynamic tests Adsorption kinetics was studied in terms of breakthrough curves in a 500 mm length and 15 mm ID glass column, fed with different masses of GAC, from 10 to 40 g, mixed with 0.9 mm inert glass beads. The liquid flow rate was varied in the range 1–11 L h−1 and three levels of inlet chromium concentration (10, 25, 50 mg L−1 ) were considered. Tests were performed for two different pH levels (i.e. 6 and 11) at room temperature (about 25 ± 1 ◦ C).


G Model

ARTICLE IN PRESS


F. Di Natale et al. / Journal of Hazardous Materials xxx (2014) xxx–xxx

12

3

pH=2

pH=2.5

pH=3

pH=6

pH=7

pH=8

ω, mg g-1

10 8 6 4 2 0 12

ω, mg g-1

10 8 6 4 2 0 0

12 pH=9.5

pH=11

ω, mg g-1

10

10

20

30

40

50

60

c, mg l-1

8 6 4 2 0 0

10

20 30 -140 c, mg l

50

60

10

20 30 -140 c, mg l

50

60

Fig. 1. Total chromium adsorption isotherms at T = 20 ◦ C and for different pH levels. Dashed lines represent model results (Eqs. (1)–(4)).

2.4. Analytical procedures

3. Results and discussion

The total chromium concentration in solution was measured by means of air/acetylene flame atomic absorption spectrophotometry (AAS-F), with a Varian SpectrAA-220 spectrophotometer (Standard Method 3111B in Clesceri et al. [33]). The concentration of Cr(VI) was measured by UV–vis absorption spectrophotometry by complexation with diphenilcarbazide (Standard method 3500Cr-B, in Clesceri et al. [33]) with a PerkinElmer Lambda EZ-150 spectrophotometer. The concentration of trivalent chromium in solution was determined by the difference between total and Cr(VI) concentration. The amount of sorbed chromium was measured by leaching the solid material with 100 mL of nitric acid aqueous solution (HNO3 , 1 M) for 24 h, which assures a complete desorption of chromium from GAC surface. Then, the solution was analysed by means of AAS-F. The accuracy of the experimental runs was checked allowing a maximum error of 5% in the chromium material balance. Each test was repeated three times and the average values of the measured variables were considered.

3.1. Equilibrium test results Chromium adsorption isotherms at 20 ◦ C and different equilibrium pH are reported in Fig. 1. All the isotherms in Fig. 1 showed that equilibrium adsorption capacity, ω, was an increasing function of concentration in the liquid solution, c, and that chromium adsorption capacity reached its maximum at pH between 6 and 7. Experimental results on the fraction of trivalent chromium in solution as a function of pH and temperature are shown in Fig. 2. Experiments revealed that Cr(VI) was partially reduced to Cr(III), with a reduction degree having a decreasing sigmoid shape, that at 20 ◦ C, spanned from 100% at pH 3.5 to about 7% at pH 6 (as already shown in Di Natale et al. [16]). At 55 ◦ C no appreciable differences were observed compared with results at 20 ◦ C, while at lower temperature, the reduction at acid pH was lower. For all the temperatures, the reduction at higher pH levels was negligible. Blank tests confirmed that the initial chromium concentration, the pH and the Cr(III)/Cr(VI) ratio remained unchanged in absence


G Model

ARTICLE IN PRESS



4

Table 1 Chromium speciation reactions.

1.0 T= 7°C T= 20°C T= 55°C

[Cr(III)]/[Crtot ]

0.8

0.6

0.4

Chromic acid hydrolysis reactions [3,43] 2HCrO4 − = Cr2 O7 2− H+ + CrO4 2− = HCrO4 − H+ + HCrO4 − = H2 CrO4

G = −8724 kJ/mol G = −37,120 kJ/mol G = 1140 kJ/mol

Formation of Cr(III) complexes [3,43] Cr3+ + OH− = CrOH2+ Cr3+ + 2OH− = Cr(OH)2 + Cr3+ + 3OH− = Cr(OH)3 (aq) Cr3+ + 4OH− = Cr(OH)4 −

G = −57,021 kJ/mol G = −104,348 kJ/mol G = −136,85 kJ/mol G = −163,079 kJ/mol

Redox reactions [44] HCrO− + 6H+ + 3e− ⇔ Cr(OH)2+ + 3H2 O 4 CrO4 2− + 4H+ + 3e− ⇔ CrO2 − + 2H2 O

Eh ◦ = 1.275 V Eh = 0.945 V

0.2

0.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 pH Fig. 2. Fraction of Cr(III) in solution as a function of equilibrium pH and temperature at c = 9 ± 0.5 mg L−1 .

of activated carbon. Hence, the reduction of chromium concentration, the pH variation and the occurrence of redox reactions during adsorption tests were only related to the presence of the sorbent. Besides, in these experiments, solution was composed only by Cr(VI), K+ , NO3 − and water. Therefore, in absence of activated carbon, there were no reducing agents in solution able to promote Cr(VI)–Cr(III) reduction, even at low pH. Experimental tests at different temperature showed that chromium adsorption was slightly increasing with temperature, similarly to the case of mercury chloride and arsenic adsorption on the same GAC [12,32], but in apparent contradiction with the exothermic character of adsorption phenomena [1,34]. This effect is shown in Fig. 3 for the case of pH 6-7 and equilibrium concentration of 8 mg L−1 .

4

ω, mg/g

3

2

1 Experiments Model predictions 0 0

10

20

30

40

50

60

T, °C Fig. 3. Total chromium equilibrium adsorption capacity on Aquacarb 207EA as a function of temperature for a total chromium concentration of 8 mg L−1 and pH 6–7. Comparison between experimental data and model predictions.

In the pertinent literature, experiments on the effect of solution temperature on chromium adsorption reported different results. Some authors found an increase in adsorption with temperature [35–39], while some others observed the opposite trend, for both Cr(VI) and Cr(III) at acid pH [40,41]. In many cases, the increase in adsorption capacity with temperature was related to either kinetic effects due to increased ion diffusivities (which however, imply that equilibrium conditions were not achieved in the referred experiments) or to the “activation” of new adsorption sites on the solid surface or to the evolution of sorbent porosity at higher temperature. However, these justifications were not clarified and the overall adsorption process was thus referred as apparently endothermic. In this work, experiments were interpreted in light of the methodological approach and the thermodynamic model developed in the past by our research group. This was successfully applied to the case of arsenic, cadmium and mercury adsorption [12,32,42] and described in details in Di Natale et al. [16]. The starting point of this model is that adsorption of metallic ions onto activated carbon occurs through parallel and consecutive reactions that involve the active sites on carbon surface and the ionic species of the dissolved metal, while the adsorption of molecular soluble species is negligible [1,3,16]. This approach involves the analysis of chromium speciation at equilibrium based on mass and electric charge balances coupled with the equations representative of chemical equilibria of the single chromium species (Table 1). Davies’ procedure [1,3] is used for the evaluation of the activity coefficients of the ionic species. The occurrence of reaction between K+ and Cr(VI) were neglected [43]. Cr(III) and Cr(VI) speciation in solution are shown in Figs. 4 and 5. The ion speciation in all the investigated conditions shows that the main ionic species in solution at pH < 3 is by far the Cr+3 (Fig. 4) while for pH > 7 the CrO4 −2 anion is prevalent (Fig. 5). The intermediate values are difficult to analyze because of the simultaneous presence of Cr(III) and Cr(VI) chromium in experimental conditions which cannot be defined of pure equilibrium, due to the presence of slow reduction reactions on the carbon surface [17]. The increase in temperature leads to an increasing dissociation grade of ionic complexes. Thus, the concentration of ions with higher net charge, such as CrO4 −2 and Cr+3 , increases with temperature. The activated carbon surface functional groups can be divided into four categories: acid, basic, nucleophilic and electrophilic. We are assuming that each of these categories is characterized by a unique value of the adsorption energy. Theoretically, this allows applying a Langmuir model to describe the adsorption of each ion on each active site category. In fact, the description of adsorption phenomena is carried out adopting a multicomponent Langmuir model to take into account the competitive effects between solvent ions (H+ , OH− , Cl− , etc.) and of the additive effects of the dissolved chromium ions (CrO4 −2 , HCrO−4 , Cr2 O7 −2 , Cr+3 ,CrOH+2 , Cr(OH)+2 , etc.), also involving the effect of temperature on their complexation reactions. However, due to the ion speciation in the investigated


G Model

ARTICLE IN PRESS



5

10-3 7°C 20°C 55°C

10-4

+

10-6

Cr

+3

,M

Cr(OH)2 , M

10-5

10-7

10-8

10-9 10-3

10-4

Cr(OH)4-, M

CrOH+2, M

10-5

10-6

10-7

10-8

10-9 2

4

6

8

10

12

2

4

6

pH

8

10

12

pH

Fig. 4. Cr(III) ion speciation as a function of pH and temperature for a total chromium concentration of 20 mg L−1 .

conditions, CrO4−2 was used to describe Cr(VI) adsorption, while Cr+3 was used to describe Cr(III) adsorption. The equilibrium model proposes a correlation between total adsorbed chromium and dissolved Cr(III) and Cr(VI) ions experimentally determined. The model equations are [16]:

ωCr(VI) =

max ωCr(VI)

ωCr(III) =

max ωCr(III)

K1 CrO4 −2

1 + K1 CrO4 −2 + KOH [OH− ]

K2 CR+3

1 + K2 Cr+3 + KH H+

ω = ωCr(III) + ωCr(VI)

(1)

(2) (3)

negligible compared to Hi . Therefore, since Gi < 0, it follows that Hi < 0, thus indicating that the adsorption process is exothermic. The model results suggest that Cr(VI) adsorption increased with temperature due to a higher dissociation of Cr(VI) ions, which determines a higher concentration of CrO4 −2 (Fig. 5), the final dissociation product of chromic acid (H2 CrO4 ). This ion has a higher net charge, which is expected to lead to a higher adsorption capacity on the GAC, especially at pH below the pHPZC . The model also allows a correct description of the effect of solution pH: the optimal pH conditions to maximize chromium adsorption capacity occurring in the range 6–7 is related to the compensation between higher Cr(VI) ionization, favored at higher pH, and lower competition with OH− ions, favored at lower pH.

Where the Langmuir equilibrium constants are expressed as:

G i

Ki = exp −

RT

3.2. Dynamic test results (4)

The good accordance between experimental and model results, reported in Figs. 1 and 3, testifies the model accuracy in the description of the effects of both pH and temperature on chromium adsorption. The values of the model parameters and the results of statistical analysis are shown in Table 2, and their calculation was detailed in Di Natale et al. [19,32]. The analysis of experimental data revealed that Gi is a constant independent of temperature. This means that the term (TSi ) is

Breakthrough curves of Cr(VI) were carried out at initial pH of 11 and 6, at different chromium concentrations (c0 = 10–50 mg L−1 ) and liquid flow rates (Q = 1–11 L h−1 ). These values were chosen as the typical composition and operational parameters for a fixedbed treatment of polluted water, either industrial or natural (by pump and treat operation). It is worth noticing that typical superficial velocity in packed bed experiments ranges between 0.2 and 3 mm s−1 [21–29]. In this paper, we adopted a higher superficial velocity (i.e. 1.5–17 mm s−1 ) to operate in a regime of fast external


G Model

ARTICLE IN PRESS



6

10-3

10-4

-

HCrO4 , M

H2 CrO4 , M

10-5

10-6

10-7

10-8

10-9 10-3

10-4

CrO4 2-, M

Cr2 O7 2-, M

10-5

10-6

10-7 7°C 20°C 55°C

10-8

10-9 2

4

6

8

10

12

2

4

pH

6

8

10

12

pH

Fig. 5. Cr(VI) ion speciation as a function of pH and temperature for a total chromium concentration of 20 mg L−1 . Table 2 Mean values of the adsorption model parameters (Eqs. (1)–(4)), their standard errors and a synoptic of the statistical data analysis. Mean Trivalent chromium max , mg g−1 ωCr(III)

G2 , kJ mol−1 GH , kJ mol−1

Hexavalent chromium max , mg g−1 ωCr(VI)

G1 , kJ mol−1 GOH , kJ mol−1

23

Stand. error

T-test

Dependency

3.2

10

0.98

−15.63 −16.74

1.2 0.25

14.04 −19.53 −29.70

8.49 7.66

0.96 0.96

1.54

7.21

0.95

0.68 0.50

28.66 36.88

0.96 0.7

mass transfer, thus allowing the exploitation of intra-particle mass transfer rate. A constant sorbent loading in the column (m = 10 g) was adopted, but some tests were carried out with different sorbent loadings (i.e. 25 and 40 g). In all the tests, the total chromium and the Cr(VI) concentrations were measured at the column outflow, as well as the solution pH. The experiments at pH 11 are shown in Fig. 6 in terms of dimensionless concentration c/c0 as a function of the dimensionless mass loading in the column, BM, defined as: BM =

Q · c0 · t m

(5)

F-test

49.2

244

R2

2 Radj

Normality test, P

0.70

0.69

>0.01

0.85

0.845

>0.01

where t is the test time expressed in hours. Fig. 6 shows that all the experimental data almost collapse on a single master curve in the plane (BM, c/c0 ), with slight differences between the tests at different flow rates and inlet chromium concentrations: the higher Q and c0 , the steeper is the breakthrough curve. The solution pH remained almost unchanged during these tests and, as expected, no reduction of Cr(VI) to Cr(III) was observed. Dynamic tests at initial pH of 6 are shown in Fig. 7, in which the effects of chromium concentration, liquid flow rate and sorbent loading are reported.


G Model

ARTICLE IN PRESS



7

1.0

1.0

-A0.8

0.6

0.6

c/c0

c/c0

0.8

0.4

0.4

1 L h-1 - 10 mg L-1 4 L h-1 - 10 mg L-1 7 L h-1 - 10 mg L-1

0.2

0.2

11 L h-1 - 10 mg L-1 -1

6.5 l/h, 10 mg/l, 10 g 6.5 l/h, 25 mg/l, 10 g 6.5 l/h, 50 mg/l, 10 g 7.5 l/h, 50 mg/l, 25 g 1.9 l/h, 50 mg/l, 25 g

-1

4 L h - 25 mg L Model prediction

0.0

0.0 0

2

4

6

8

0

10

15

30

45

60

75

90

3

BMx 10 , -

BM, 1.0 Fig. 6. Experimental data (symbols) and model results (lines) of Cr(VI) breakthrough curve at pH 11. T = 25 ± 1 ◦ C.

-B0.8

0.6

c/c0

The experiments showed that the solution pH slightly varied between 7 and 6 with the exception of the very first samples, which had a pH around 8. Moreover, experiments indicate that Cr(VI) reduction to Cr(III) while occurring during equilibrium tests (contact time >48 h) was not observed during dynamic tests (contact time < 8 h), as already observed in the pertinent literature [17,19,21–29]. This result can be explained considering that Cr(III) formation on the carbon surface is a process slower than adsorption. Nothing can be said concerning the oxidation state of adsorbed chromium, which is expected to include both Cr(III) and Cr(VI). The experiments showed that the effect of chromium concentration can be well described considering the dependence of c/c0 as a function of BM, but data at different sorbent mass provided significant deviation from a hypothetic master curve (Fig. 7A). Besides, the shape of the curves reveals a marked difference with the experiments carried out at pH 11. In fact, all the breakthrough curves showed a marked change in their slope for a certain value of BM (see also the zoomed diagram in Fig. 7B) corresponding to an average sorbent loading of about 1.6 ± 0.15 mg g−1 of chromium, for all the experiments. This loading corresponds to an equilibrium concentration of chromium in solution of about 1 mg L−1 . This change in slope led to a slower adsorption rate which gave rise to values of BM at saturation almost ten times higher than those observed at pH 11. A similar result was already observed in the literature. Indeed, available experimental data demonstrated that an adsorption model based on a single rate controlling mechanism, usually addressed by a mass transfer coefficient, is able to describe Cr(III) adsorption on granular activated carbon [22,23]. A deviation from a single mass transfer rate controlling mechanism, similarly to that observed in this study for pH 6, was observed by Malkoc et al. [26] and by Sa˘g and Aktay [28] for Cr(VI) adsorption. In order to analyze the results of dynamic tests, the breakthrough curves are interpreted assuming a plug flow with axial dispersion velocity profile in the column, and an ion adsorption rate controlled by external mass transfer rate [34,45,46]. To this end, the mass balance equations of Cr(VI) ions were taken into account.

0.4 6.5 l/h, 10 mg/l, 10 g 6.5 l/h, 25 mg/l, 10 g 6.5 l/h, 50 mg/l, 10 g 7.5 l/h, 50 mg/l, 25 g 1.9 l/h, 50 mg/l, 25 g

0.2

0.0 0

2

4

6

8

10

3

BMx 10 , Fig. 7. Experimental data of Cr(VI) breakthrough curves at T = 25 ± 1 ◦ C and pH 6 (A). Zoom of the experimental data at small BM values and application of the adsorption model to the experimental data (B).

Hence, the mass balance over the column height z = [0; H] and time t = [0; ∞] is: 2

ε

∂c ∂c ∂ C ∂q − (1 − ε) s + εU = εDI ∂t ∂Z ∂Z 2 ∂t

(6)

where c(z, t) and q(z, t) are the concentration in the liquid and in the solid phases of each Cr(VI) ion present, respectively, U is the empty column liquid velocity in the absorber; s is the solid particle density; ε is the bed void fraction; Dl is the axial dispersion coefficient. It is worth to observe that at pH 11 only CrO4 −2 species is present, hence it coincides with total Cr(VI) concentration. The adsorption rate (∂q/∂t) for an external liquid-solid mass transfer rate controlling mechanism is expressed by the equation: (1 − ε) s

∂q = kf a · c − ceq (q) ∂t

(7)

where kf [m s−1 ] is the mass transfer coefficient between the liquid and a single adsorbent particle; a = 6/dp is the specific area of the


G Model

ARTICLE IN PRESS



8

Table 3 Values of the mass transfer coefficient, kf , and of the Sherwood number, Sh, calculated by best fitting of experimental data at pH 11 with the adopted adsorption model. Q (L h−1 )

U (mm s−1 )

Re

kf (m s−1 )

Sh

1 4 7 11

1.49 6.47 11.6 18.2

1.74 7.52 13.43 21.1

7.419 × 10−6 1.462 × 10−5 2.028 × 10−5 2.602 × 10−5

5.98 11.8 16.3 21.0

sorbent; and ceq (q) is the concentration in the liquid phase that should be in equilibrium with the actual sorbent uptake q at any time and position of the sorbent particle in the column. Therefore, the value of ceq (q) was calculated using the equilibrium model (1) assuming q = ω at the actual pH at time t of the tests. Initial and boundary conditions for this system are: t = 0 c (z, t) = 0 q (z, t) = 0 z=H

ce (z, t) = c (z, t)

z = 0 c (z, t) = c0 − εDI

(8)

∂q (z, t) =0 ∂z ∂c (z, t) ∂z

q (z, t) = 0

(9) (10)

The mass transfer coefficient was estimated through a bestfitting analysis of experimental data. The values of kf were assumed to be constant for any chromium concentration and solution pH, but not with liquid flow rate. For tests at pH 11, they are resumed in Table 3 in terms of Sherwood number Sh = kf dp /D, where dp is the particle average diameter and D is the chromium diffusivity in water assumed equal 1.44 × 10−9 m2 s−1 [47,48]. The values of kf obtained by best fitting of data are consistent with former models found in the available literature (e.g. [45,46]). This data can be eventually fitted by the following expression: Sh = 2 + 0.376 · Re0.55 p

(11)

where Rep is the particle Reynolds number. Moreover, a comparison between experiments and model results is reported in Fig. 6, which clearly indicates that the assumption of external mass transfer rate controlling mechanism can proficiently describe the experimental data at pH 11 in the entire investigated range and explains both the effects of liquid flow rate and sorbent mass. For the experimental data at pH 6, the same adsorption model with external mass transfer velocity assumed as the rate controlling mechanism allowed an effective description of the experiments only for the initial part of the breakthrough curves (Fig. 7B). However, it is worth observing that the model allows a correct determination of the breakpoint of the curves, which represents a key design parameter for packed bed column. In these calculations, we successfully used the same expression of the external mass transfer coefficient obtained from tests at pH 11 and reported in Eq. (11). Differently, the model failed to describe data above the critical time when the change in the curve slope was observed. This slope change can be related to a transition to a inter-particle control mechanism as diffusion in macro/micropores or to surface reaction kinetic. To date, the experiments are not able to provide enough information to address this behavior with the due accuracy. 4. Conclusions Chromium adsorption on a granular activated carbon was studied in terms of process equilibrium and kinetics. Batch experiments were carried out to characterize equilibrium adsorption capacity at different pH and temperature of the solution. Experimental results suggest that Cr(VI) adsorption increases with temperature

due to higher dissociation of chromic acid. Furthermore, optimal pH conditions were derived from the compensation between higher Cr(VI) ionization, favored at higher pH, and lower competition with OH− ions, favored at lower pH. An additive-competitive Langmuir model was used to describe the experiments allowing a proper description of the effects of both pH and temperature on chromium adsorption capacity, also accounting for the chromium ion speciation. Dedicated dynamic tests were carried out to assess the effect of chromium concentration and flow rate in a fixed-bed column configuration. Dynamic tests at initial pH 11 were properly described by a mass transfer model in which an external mass transfer was hypothesized as rate controlling mechanism. However, at initial pH 6 the same model failed to describe the experimental breakthrough curves obtained at a sorbent coverage higher than 1.6 mg/g circa. Above this level, a marked reduction of the breakthrough curve slope was observed. This slope change can be related to a transition to an inter-particle control mechanism such as diffusion in macro/micropores or to surface reaction kinetic. To date, the experiments do not provide enough information to address this behavior with the due accuracy. However, this study suggests that future developments should include new experiments at different pH levels and with GAC particle of different sizes, in order to assess the nature of the intra-particle adsorption rate mechanism. The development of a proper modelling to take into account for the change of controlling mechanism should be also considered. References [1] M.M. Benjamin, Water Chemistry, McGraw-Hill, New York, 2002. [2] D. Mohan, C.U. Pittman Jr., Activated carbons and low cost adsorbents for remediation of tri- and hexavalent chromium from water, J. Hazard. Mater. 137 (2) (2006) 762–811. [3] W. Stumm, J.J. Morgan, Aquatic Chemistry, 3rd ed., Wiley & Sons, New York, 1996. [4] M. Gheju, I. Balcu, Removal of chromium from Cr(VI) polluted wastewaters by reduction with scrap iron and subsequent precipitation of resulted cations, J. Hazard. Mater. 196 (2011) 131–138. [5] P. Religa, A. Kowalik-Klimczak, P. Gierycz, Study on the behavior of nanofiltration membranes using for chromium(III) recovery from salt mixture solution, Desalination 315 (2013) 115–123. [6] A. Senol, Amine extraction of chromium(VI) from aqueous acidic solutions, Sep. Purif. Technol. 36 (1) (2004) 63–75. [7] S. Mustafa, K.H. Shah, A. Naeem, M. Waseem, M. Tahir, Chromium(III) removal by weak acid exchanger Amberlite IRC-50 (Na), J. Hazard. Mater. 160 (1) (2008) 1–5. [8] S. Yacoumi, C. Tien, Kinetics of Metal Ion Adsorption from Aqueous Solutions, Kluwer Academic Publisher, Boston, 1995. [9] P. Rana, N. Mohan, C. Rajagopal, Electrochemical removal of chromium from wastewater by using carbon aerogel electrodes, Water Res. 38 (12) (2004) 2811–2820. [10] Y. Chen, G. Gu, Preliminary studies on continuous chromium(VI) biological removal from wastewater by anaerobic–aerobic activated sludge process, Bioresour. Technol. 96 (15) (2005) 1713–1721. [11] A. Erto, R. Andreozzi, A. Lancia, D. Musmarra, Factors affecting the adsorption of trichloroethylene onto activated carbons, Appl. Surf. Sci. 256 (2010) 5237–5242. [12] F. Di Natale, A. Erto, A. Lancia, D. Musmarra, Mercury adsorption on granular activated carbon in aqueous solutions containing nitrates and chlorides, J. Hazard. Mater. 192 (2011) 1842–1850. [13] A. Erto, I. Bortone, A. Di Nardo, M. Di Natale, D. Musmarra, Permeable Adsorptive Barrier (PAB) for the remediation of groundwater simultaneously contaminated by some chlorinated organic compounds, J. Environ. Manage. 140 (2014) 111–119. [14] I. Bortone, A. Di Nardo, M. Di Natale, A. Erto, D. Musmarra, G.F. Santonastaso, Remediation of an aquifer polluted with dissolved tetrachloroethylene by an array of wells filled with activated carbon, J. Hazard. Mater. 260 (2013) 914–920. [15] H.P. Boehm, Surface oxides on carbon and their analysis: a critical assessment, Carbon 40 (2002) 145–149. [16] F. Di Natale, A. Erto, A. Lancia, D. Musmarra, A descriptive model for metallic ions adsorption from aqueous solutions onto activated carbons, J. Hazard. Mater. 169 (2009) 360–369. [17] J. Lakatos, S.D. Brown, C.E. Snape, Coals as sorbents for the removal and reduction of hexavalent chromium from aqueous waste streams, Fuel 81 (5) (2002) 691–698.


G Model HAZMAT-16167; No. of Pages 9

ARTICLE IN PRESS F. Di Natale et al. / Journal of Hazardous Materials xxx (2014) xxx–xxx

[18] D. Mohan, K.P. Singh, V.K. Singh, Trivalent chromium removal from wastewater using low cost activated carbon derived from agricultural waste material and activated carbon fabric cloth, J. Hazard. Mater. 135 (1–3) (2006) 280–295. [19] F. Di Natale, A. Lancia, A. Molino, D. Musmarra, Removal of chromium ions from aqueous solutions by adsorption on activated carbon and char, J. Hazard. Mater. 145 (2007) 381–390. [20] A. Molino, A. Erto, F. Di Natale, A. Donatelli, P. Iovane, D. Musmarra, Gasification of granulated scrap tires for the production of syn-gas and a low cost adsorbent for Cd(II) removal from waste waters, Ind. Eng. Chem. Res. 52 (34) (2013) 12154–12160. [21] G. Huang, J.X. Shi, T.A.G. Langrish, Removal of Cr(VI) from aqueous solution using activated carbon modified with nitric acid, Chem. Eng. J. 152 (2009) 434–439. [22] R.M. Schneider, C.F. Cavalin, M.A.S.D. Barros, C.R.G. Tavares, Adsorption of chromium ions in activated carbon, Chem. Eng. J. 132 (2007) 355–362. [23] A.H. Sulaymon, B.A. Abid, J.A. Al-Najar, Removal of lead copper chromium and cobalt ions onto granular activated carbon in batch and fixed-bed adsorbers, Chem. Eng. J. 155 (2009) 647–653. [24] Z. Aksu, F. Gönen, Z. Demircan, Biosorption of chromium(VI) ions by Mowital®B30H resin immobilized activated sludge in a packed bed: comparison with granular activated carbon, Process Biochem. 38 (2002) 175–186. [25] S. Chen, Q. Yue, B. Gao, Q. Li, X. Xu, K. Fu, Adsorption of hexavalent chromium from aqueous solution by modified corn stalk: a fixed-bed column study, Bioresour. Technol. 113 (2012) 114–120. [26] Malkoc, Y. Nuhoglu, M. Dundar, Adsorption of chromium(VI) on pomace—an olive oil industry waste: batch and column studies, J. Hazard. Mater. 138 (2006) 142–151. [27] M. Pereira, P. Arroyo, M. de Barros, V. Sanches, E. da Silva, I. Fonseca, R. Lovera, Chromium adsorption in olive stone activated carbon, Adsorption 12 (2006) 155–162. [28] Y. Sa˘g, Y. Aktay, Application of equilibrium and mass transfer models to dynamic removal of Cr(VI) ions by Chitin in packed column reactor, Process Biochem. 36 (2001) 1187–1197. [29] V. Vinodhini, N. Das, Packed bed column studies on Cr (VI) removal from tannery wastewater by neem sawdust, Desalination 264 (2010) 9–14. [30] J.S. Noh, J.A. Schwarz, Estimation of surface ionization constants for amphoteric solids, J. Colloid Interface Sci. 139(1) 1990 139–148. [31] T. Bandosz, J. Jagiello, J.A. Schwarz, Comparison of methods to assess surface acidic groups on activated carbons, Anal. Chem. 64 (1992) 891–989. [32] F. Di Natale, A. Erto, A. Lancia, D. Musmarra, Experimental, modelling analysis of As(V) ions adsorption on granular activated carbon, Water Res. 42 (2008) 2007–2016.

9

[33] L.S. Clesceri, A.E. Greenberg, A.D. Eaton (Eds.), Standard Methods for the Examination of Water and Wastewater, 20th ed., Am. Public Health Ass., Washington, 1998. [34] D. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley & Sons, 1984. [35] M. Barkat, D. Nibou, S. Chegrouche, A. Mellah, Kinetics and thermodynamics studies of chromium(VI) ions adsorption onto activated carbon from aqueous solutions, Chem. Eng. Process.: Process Intensif. 48 (2009) 38–47. [36] F. Granados-Correa, J. Jiménez-Becerril, Chromium (VI) adsorption on boehmite, J. Hazard. Mater. 162 (2009) 1178–1184. [37] H. Demiral, I˙ . Demiral, F. Tümsek, B. Karabacako˘glu, Adsorption of chromium(VI) from aqueous solution by activated carbon derived from olive bagasse and applicability of different adsorption models, Chem. Eng. J. 144 (2008) 188–196. [38] D.D. Das, R. Mahapatra, J. Pradhan, S.N. Das, R.S. Thakur, Removal of Cr(VI) from aqueous solution using activated cow dung carbon, J. Colloid Interface Sci. 232 (2000) 235–240. [39] G.S. Agarwal, H.K. Bhuptawat, S. Chaudhari, Biosorption of aqueous chromium(VI) by Tamarindus indica seeds, Bioresour. Technol. 97 (2006) 949–956. [40] A.B. Albadarin, C. Mangwandi, A. a.H. Al-Muhtaseb, G.M. Walker, S.J. Allen, M.N.M. Ahmad, Kinetic and thermodynamics of chromium ions adsorption onto low-cost dolomite adsorbent, Chem. Eng. J. 179 (2012) 193–202. [41] J. Hu, G. Chen, I.M.C. Lo, Removal and recovery of Cr(VI) from wastewater by maghemite nanoparticles, Water Res. 39 (2005) 4528–4536. [42] F. Di Natale, M. Di Natale, R. Greco, A. Lancia, C. Laudante, D. Musmarra, Groundwater protection from cadmium contamination by permeable adsorbing barrier, J. Hazard. Mater. 160 (2008) 428–434. [43] A.E. Martell, R.M. Smith, Critical Stability Constants, Vol. 4. Inorganic Complexes, Plenum Press, 1976. [44] M. Pourbaix, Atlas d’Equilibres Electrochimiques, Gauthier-Villars & C., Paris, 1963. [45] D. Green, R. Perry, Perry’s Chemical Engineers’ Handbook, 8th ed., McGraw-Hill, 2008. ˜ J.C. [46] M. Vocciante, M. Trofa, T. D’Auria, L. Giraldo, P. Rodriguez-Estupinan, Moreno Pirajan, A. Erto, Experimental and simulation-based studies of cadmium and nickel adsorption from wastewater, J. Clean. Prod. 77 (2014) 35–46. [47] N. Iadicicco, L. Paduano, V. Vitagliano, Diffusion coefficients for the system potassium chromate–water at 25 ◦ C, J. Chem. Eng. Data 41 (1996) 529–533. [48] B.E. Poling, J.M. Prausnitz, J.P. O’Connell, Properties of Gases and Liquids, 5th ed., McGraw-Hill Education, Milan, 2001.


Adsorption of hexavalent chromium onto organic bentonite modified by the use of iron(III) chloride.

Effective removal of hexavalent chromium from aqueous solutions by adsorption on mesoporous carbon microspheres.

surfactant: process optimization, kinetics and equilibrium.

Kinetics, equilibrium, and thermodynamics investigation on the adsorption of lead(II) by coal-based activated carbon.

Efficient Adsorption of Sulfamethazine onto Modified Activated Carbon: A Plausible Adsorption Mechanism.

Adsorption equilibrium and dynamics of gasoline vapors onto polymeric adsorbents.

Adsorption behavior comparison of trivalent and hexavalent chromium on biochar derived from municipal sludge.

Characterizing o- and p-nitrophenols adsorption onto innovative activated carbon prepared from date pits.

Adsorption of local anesthetics on activated carbon: Freundlich adsorption isotherms.

Adsorption characteristics of trichloroethylene and 1,1,1-trichloroethane onto activated carbon fiber in gaseous phase.

activated carbon composite.

Enhancing the adsorption of ionic liquids onto activated carbon by the addition of inorganic salts.

activated carbon magnetic composite.

Investigation of dye adsorption onto activated carbon from the shells of Macoré fruit.

A DFT study on carbon monoxide adsorption onto hydroxylated α-Al₂O₃(0001) surfaces.

Studies on the adsorption of chromium(VI) onto 3-Mercaptopropionic acid coated superparamagnetic iron oxide nanoparticles.

Toxicity and mutagenicity of hexavalent chromium on Salmonella typhimurium.

Effect of calcium on adsorption capacity of powdered activated carbon.

Localized opto-mechanical control of protein adsorption onto carbon nanotubes.

2,4-D adsorption to biochars: effect of preparation conditions on equilibrium adsorption capacity and comparison with commercial activated carbon literature data.

Potential of Live Spirulina platensis on Biosorption of Hexavalent Chromium and Its Conversion to Trivalent Chromium.

Adsorption of Selected Pharmaceutical Compounds onto Activated Carbon in Dilute Aqueous Solutions Exemplified by Acetaminophen, Diclofenac, and Sulfamethoxazole.

Arsenic (III) adsorption on iron acetate coated activated alumina: thermodynamic, kinetics and equilibrium approach.

Multi-component adsorption of copper, nickel and zinc from aqueous solutions onto activated carbon prepared from date stones.