Advanced
Kinetics on the Removal of Cationic Dyes from Aqueous Solutions over Maghnia Montmorillonite Adsorbent
Kinetics on the Removal of Cationic Dyes from Aqueous Solutions over Maghnia Montmorillonite Adsorbent
Journal of the Korean Chemical Society. 2010. Oct, 54(5): 603-610
Copyright © 2010, The Korean Chemical Society
  • Received : April 27, 2010
  • Accepted : June 28, 2010
  • Published : October 20, 2010
Download
PDF
e-PUB
PubReader
PPT
Export by style
Article
Author
Metrics
Cited by
TagCloud
About the Authors
A. Elaziouti
abdelkader923@yahoo.fr
N. Laouedj
DR. Moulay Tahar University, Saida, Algeria

Abstract
The potential feasibility of Maghnia clay as adsorbent to remove three cationic dyes (methylene blue (MB), neutral red (NR) and malachite green oxalates (MG)) from aqueous solution was investigated at various operating parameters such as contact time, pH, initial dye concentration and adsorbent dose. Optimum pH values for dyes adsorption were 6 for NR and 7 for MB and MG, respectively. Maximum adsorption of dyes, i.e. ≥ 90% has been achieved in aqueous solutions in 60 minutes. The adsorption of dyes followed the pseudo-second-order rate equation.
Keywords
INTRODUCTION
Dyes effluents released (approximately 7. 10 5 tons) into the environment by technological activities pose a serious threat to the environment. Their presence in water, even at very low concentrations, may significantly affect photosynthetic activity in aquatic life due to reduced light penetration. In addition, some dyes such as azo and triphenylmethanes groups are either toxic or mutagenic and carcinogenic. 1 Due to their synthetic origin and complex aromatic molecular structures, which make them more stable, non- biodegrade, conventional wastewater treatment techniques such as oxidation-reduction and exchanging resins of ions, 2 coagulation/flocculation, 3 membrane separation, 4 adsorption, 5 biological methods, 6 and more recently the advanced processes of oxidation, 7 are ineffective. Adsorption is one of the effective methods to remove coloured textile contaminants from wastewaters.
Many low-cost adsorbents for dye removal from mineral waste, 9 agricultural wastes, microbial biomass, 10 higher plant biomass, 11 tree fern, 12 orange peel, 13 date pits 14 palm kernel fiber, sawdust, 15 peanut hull, 16 neem leaf, 17 de-oiled soya, 18 moss Rhytidiadelphus squarrosus, 19 activated carbon, 20 , 21 rice husk-based porous carbon, 22 magnetic particles 23 and paper sludge. 24
Recently, we used clay mineral for the successful adsorption of metal ions, 25 , 26 , 27 basic dyes 28 , 29 as a result of their catalytic properties and high capacity in the removal contaminants too stable and resistant for conventional methods. Adsorption of cationic organic dyes in smectite clays has been extensively studied, mainly in aqueous solution. This clay is a dioctahedral smectite with isomorphic substitutions of Al +3 by Mg +2 with structural formula unit:
  • (Si8)IV(Al4-xMgx)VIO20(OH)4
  • where x = 0.5 ‒ 0.9.
However, for Maghnia montmorillonite, part of the negative charge around 15 - 50% is due to tetrahedral substitutions of Si +4 by Al +3 . This clay has an appreciable content of iron.
The adsorption of cationic dyes is not exclusively due to the ionic exchange mechanism but other interactions, such a hydrophobic effect between guest molecules on the clay surface 30 and/or H-bonding 31 can contribute to the adsorption of organic dye on clay surface, favouring the dye agglomeration.
The aim of the present study was to assess the potential ability of locally available clay originated of Maghnia montmorillonite deposit (west of Algeria) for removal of three cationic dyes (methylene blue, neutral red and malachite green oxalates) from aqueous solutions. The adsorption of dyes was investigated in solutions of pH 3 - 13. The adsorption kinetics was determined quantitatively by the pseudofirst-and second-order reactions.
MATERIALS AND METHODS
Natural clay from Maghnia (Algeria) was used in a purified form. A thiazin group cation, methylene blue, MB (MW = 319,5 g.mol -1 , λ max = 665 nm, ε = 95000 cm -1 mole -1 dm 3 ), a phenazin group cation, Neutral red dye, NR (MW = 319.50 g.mol -1 , λ max = 520 nm, ε = 25000 cm -1 mole -1 dm 3 ) and a triphenylemethane group cation, Malachite green oxalates, MG (MW = 507. 51 g/mol, λ max = 617 nm, ε = 42000 dm 3 /cm mole) from Across product for microbiological analysis and used without any further purification. Molecular structures of dyes is shown in . 1 .
Natural clay was supplied by the Ceramics Society of Maghnia (west of Algeria). The sodium form of clay with a particle size < 0.2 μm was prepared in order to obtain good clay colloidal dispersions, as follows: clay was saturated with Na + by repeated exchange with NaCl (1 M) aqueous solution. The < 0.2 particle size fraction was separated by centrifugation and it was washed several times with distilled water until free Cl was not detected in the suspension (AgNO 3 test). The < 0.2 fraction was collected, freeze dried at 60 ℃ for 24 h and crushed. The material resulting from different treatments will be indicated by clay-Na.
PPT Slide
Lager Image
Molecular structures of dyes (a) Methylene blue (MB) (b) Neutral red (NR) and (c) Malachite green oxalates (MG).
Aqueous dye solution stock was prepared by dissolving accurately weighed neat dye in distilled water to the concentration of 1 g/L. Experimental solutions were obtained by successive dilutions.
Adsorption kinetics experiments were carried out as a single-stage batch test using a magnetically stirrer at 298 K using 250 mL beaker containing 100 mL different concentrations and initial pH values of dye solutions. The initial pH of dye solutions was adjusted by HCl/NaOH solutions using digital pH-meter (HANNA HI 83141). Different doses of sorbent were added to each beaker. The samples were withdrawn from the beaker at the predetermined time intervals, and adsorbent was separated from the solution by centrifugation (EBA-Hetlich) at 3500 rpm for 15 min. Dye concentrations in the supernatant solutions were subsequently estimated by measuring absorbance at maximum wavelengths of dyes with UV mini-1240 spectrophotometer (UV-Vis Spectrometer SHIMAHIDZU) and computing from the calibration curves. The experiments were done by varying the initial pH (3 - 13); amount of adsorbent (0.1 - 0.5 g/L) for MB and (0.07 - 0.2 g/L) for NR and MG respectively and concentration of dye solution (30 - 70 mg/L) at different time intervals.
The efficiency of color removal was expressed as the percentage ratio of decolorized dye concentration to that of initial one (Eq. (1) and (2)):
PPT Slide
Lager Image
PPT Slide
Lager Image
Where C o and C e are the initial and equilibrium dye concentrations (mg/L) respectively. V is the volume of the solutions (L), and m is the weight of the clay-Na (g).
RESULTS AND DISCUSSION
1 displays the results of chemical composition (Oxford spectrometer) in the clay-Na sample. The major elements of the clay-Na, expressed in terms of oxide, are primarily of silica, and alumina with appreciable iron content. The CaO and Na 2 O (K 2 O) reflect the contents of interlayer cationic in clay. The other oxides present in traces amounts such. In the perfect montmorillonite structure, the total theoretical content of SiO 2 and Al 2 O 3 is 92% 32 and the content ratio of SiO 2 to Al 2 O 3 is 2.6. In the Maghnia clay, the total content of SiO 2 and Al 2 O 3 is 84.90% and the corresponding ratio of SiO 2 to Al 2 O 3 is 2.52, so this type of clay belongs to the Na-montmorillonite category. The cation capacity of exchange and total specific surface of the clay-Na were estimated by the methylene blue (MB) method 33 are of 101.72 meq/100 gr of clay-Na and 673.93 m 2 /g respectively.
XRF analysis of the clay-Na sample
PPT Slide
Lager Image
LOI: Loss on ignition
The effects of initial pH solution on the adsorption percentages of NR and MB dyes were investigated over a range of pH values from of 3 to 13, but for MG, the experiments were only conducted from pH 3to 9 for avoiding dye degradation. As shown in . 2 , when the initial pH for both MB and MG dyes was increased from 3 to 7, the percentage of dyes adsorbed increased from 81.47 to 90. 81% for MB and from 59.84 to 80.40% for MG, then the dye adsorption was unaffected by pH in the range of 7 - 9. At higher pH values of 9 - 13 the dye adsorption decreased to 80. 46 % at pH of 13. However, for NR, the percentage of dye increased from 62.52 to 87.19% as the pH value was increased from 3 to 6.5, which decreased to 52.39 % at pH of 12. The maximum adsorptions of dyes were 90.81, 87.19 and 80.40 % for MB, NR and MG were achieved at pH 7, 6 and 7 respectively. For this reason, the pH 7 for MG, 6 for NR and 7 for MG are selected for subsequent experiments.
PPT Slide
Lager Image
Effect of initial pH solution on the adsorption percentages of dyes onto clay-Na for [Clay-Na] = 0.1 g/L (NR and MG) and 0.3 g/L (MB), [MB] = [NR] = [MG] = 50 mg/L, T = 298 K, and contact time of 60 min.
The adsorption of these charged dye groups onto the adsorbent surface is primarily influenced by the surface charge on the adsorbent, which is in turn influenced by the solution.
For clay minerals the potential determining ions are H + and OH and complex ions formed by bonding with H + and OH . The broken Si-O bonds and Al-OH bonds along the surfaces of the clay crystals result in hydrolysis, is an amphoteric surface hydroxyl group that can associate (Rec. 1) and dissociate (Rec. 2) proton.
To explain the relationship between the layer charge density and the adsorption, so-colled models of surface complexation (SCM) was developed. 34 In the 2-pK approach we assume two reactions for surface protonation. 35
At low pH the reaction might be:
PPT Slide
Lager Image
Lower adsorption rate of dyes at low pH is probably due to the presence of excess H + ions competing with the cations groups on the dyes for the adsorption sites.
However, the high removal percentage observed for NR at pH 6 (low pH) is attributed to the formation of surface hydrogen bonds between the surface of the hydrogen bonds of the hydroxyl groups on the clay-Na surface and the nitrogen atoms of neutral red cations. 36
At high pH the reaction is:
PPT Slide
Lager Image
PPT Slide
Lager Image
Adsorption kinetic for MB adsorption at various initial concentrations for [Clay-Na] = 0.3 g/L, pH = 7 and T = 298 K.
PPT Slide
Lager Image
Adsorption kinetic for NR adsorption at various initial concentrations for [Clay-Na] = 0.1 g/L, pH = 6 and T = 298 K.
PPT Slide
Lager Image
Adsorption kinetic for MG adsorption at various initial concentrations for [Clay-Na] = 0.1 g/L, pH = 7 and T = 298 K.
PPT Slide
Lager Image
Pseudo-second order kinetic for MB adsorption at various initial concentrations for [Clay-Na] = 0.3 g/L, pH = 7 and T = 298 K.
PPT Slide
Lager Image
Pseudo-second order kinetic for NR adsorption at various initial concentrations for [Clay-Na] = 0.1 g/L, pH = 6 and T = 298 K.
PPT Slide
Lager Image
Pseudo-second order kinetic for MG adsorption at various initial concentrations for [Clay-Na] = 0.1 g/L, pH = 7 and T = 298 K.
In our studies the adsorption maximum for both MB and MG dyes were achieved at pH 7 respectively. So at neutral and high pH, the solution in contact with the basal oxygen surface of the tetrahedral sheet will contain excess hydroxyls. The surface will then exhibit a cation exchange capacity (Rec. 3):
PPT Slide
Lager Image
The adsorption percentage of MB decreased over the pH range of 9 - 13(above pH 9), then remained constant up to pH 13. Such decrease is assigned to the presence of dimers in the clay-Na environment competing with the monomers cations (MB + ) on the dyes for the active sites (Supporting results).
PPT Slide
Lager Image
Adsorption kinetic for MB adsorption at various adsorbent doses for [MB] = 50 mg/L, pH = 7 and T = 298 K.
PPT Slide
Lager Image
Adsorption kinetic for NR adsorption at various adsorbent doses for [NR] = 50 mg/L, [Clay-Na] = 0.1 g/L, pH = 6 and T = 298 K.
PPT Slide
Lager Image
Adsorption kinetic for MG adsorption onto clay-Na at various adsorbent doses for [NR] = 50 mg/L, pH = 7 and T = 298 K.
On the other hand, at neutral and higher pH (above 10), most NR dye occurred in the undissociated form. Due to hydrophobic interactions between aromatic rings of dye molecules, Neutral red tends to aggregate in aqueous solution. This aggregation phenomenon is more noticed at high alkali medium. 37
PPT Slide
Lager Image
Pseudo-second order kinetic for MB adsorption at various adsorbent doses for [MB] = 50 mg/L, pH = 7 and T = 298 K.
PPT Slide
Lager Image
Pseudo-second order kinetic for NR adsorption at various adsorbent doses for [NR] = 50 mg/L, pH = 6 and T = 298 K.
PPT Slide
Lager Image
Pseudo-second order kinetic for MG adsorption onto clay-Na at various adsorbent doses for [NR] = 50 mg/L, pH = 7 and T = 298 K.
The effects of initial dye concentration ( . 3 , 4 and 5 ) and adsorbent dose ( . 9 , 10 and 11 ) as a function of contact time were conducted in the range of 0 to 60 min. The results indicated that the process is found to be very rapid initially, and the equilibrium is achieved within 10 min.
As illustrated in . 3 , 4 and 5 , the adsorption capacityincreased from 96.971 to 228.900 mg/g for MB, from 239.223 to 440.506 mg/g for NR and from 251.71 to 498.13 mg/g for MG when the initial concentration of dyes was increased from 30 to 70 mg/L respectively. Initial concentration provides an important driving force to overcome all mass transfer resistances of the dye between the aqueous and solid phases. Hence a higher initial concentration of dye may enhance the process.
Adsorption kinetic parameters for dyes adsorption onto clay-Na at various initial concentrations for [Clay-Na] = 0.1 g/L (NR and MG), 0.3 g/L (MB), pH = 6 (NR), pH = 7 (MB and MG) and T = 298 K.
PPT Slide
Lager Image
R2: Regression coefficient
. 9 showed that the adsorption capacity of MB decreased from 367.515 to 99.631 mg/g when the adsorbent dose was increased from 0.1 to 0.5 g/L. As results showed in . 10 and 11 , the adsorption capacity of dyes decreased from 537.010 to 227.217 mg/g for NR and from 460.44 to 230.05 mg/g for MG as the adsorbent dose was increased from 0.07 to 0.2 g/L respectively.
This may be attributed to overlapping of adsorption sites as a result of overcrowding of adsorbent. 38
The adsorption kinetics of dyes at different initial concentration, illustrated in . 3 , 4 and 5 , were treated with a pseudo-first order and pseudo-second order models.
The pseudo-first order equation of Lagergren 39 is generally expressed as follows (Eq. (3)):
PPT Slide
Lager Image
where Qe and Qt (both in mg/g) are the amount of dye adsorbed per unit weight of adsorbent at equilibrium and at any time t , respectively and K 1 (min -1 ) is the rate constant for adsorption of dye. At given boundary conditions for t = 0, Qt = 0, the equation Eq. (3) can be integrated to give (Eq. (4)):
PPT Slide
Lager Image
The values of K 1 were calculated from the slopes of the respective linear plots of log (Qe - Qt) versus t. The regression coefficients, R 2 , (given in 2 ) for the pseudo-first order model at different initial concentration (30 - 70 mg/L) did not exceed the values of 0.925, 0.968 and 0.988 for MB, NR and MG respectively. The calculated Qe values obtained from pseudo-first order kinetic model for dyes were much different compared with experimental Qexp values. These results suggest that the process does not follow the pseudofirst order rate equation of Lagergren.
The adsorption of dyes onto clay-Na at different adsorbent doses were also analyzed with pseudo-first-order model, the regression coefficients, R 2 (given in 3 ) result as 0.539, 0.921 and 0.966 for MB, NR and MG respectively. This finding indicates that the adsorption of dyes onto clay-Na does not follow the pseudo-first order model of Lagergren.
The adsorption kinetics following pseudo-second order model given by Ho 40 is represented in the form (Eq. (5)):
PPT Slide
Lager Image
where Q and Qe represent the amount of dye adsorbed (mg/g) at any time t; K 2 is the rate of sorption (g/mg.min) and Qe the amount of dye adsorbed onto clay-Na at quilibrium (mg/g). Separating (Eq. (6)) gives:
PPT Slide
Lager Image
Integrating Eq. (5) with respect to the boundary conditions Q = 0 at t = 0 and Q = Q at t = t, the linearised form of pseudo-second order expression can be obtained as (Eq. (7)):
Adsorption kinetic parameters for dyes adsorption onto clay-Na at various adsorbent doses for [MB] = [NR] = [MG] = 50 mg/L pH = 6 (NR); pH = 7 (MB and MG) and T = 298 K.
PPT Slide
Lager Image
R2: Regression coefficient
PPT Slide
Lager Image
Equation (7) can be further linearised to (Eq. (8)):
PPT Slide
Lager Image
The linearity of the plots of t/Qt versus t for adsorption of dyes onto clay-Na at different initial concentration ( . 6 , 7 and 8 , 2 ) and various adsorbent dose ( . 12 , 13 and 14 , 3 ) suggest that all adsorption data of dyes are satisfactorily described by pseudo-second order model with extremely higher regression coefficient (R 2 > 0.998), based on the assumption that the rate-limiting step may be chemisorptions involving valency forces through sharing or exchange of electrons between the hydrophilic edge sites of clay-Na and polar dye ions. 41 The calculated Qe values obtained from pseudo-second-order kinetic model were close to the experimental Qexp values for all dyes.
CONCLUSION
The potential ability of Maghnia clay to remove three cationic dyes (methylene blue, neutral red and malachite green oxalates) from aqueous solution was investigated at various parameters (contact time, pH, initial dye concentration and adsorbent dose). Optimum pH values for dyes adsorption were 6 for NR and 7 for MB and MG respectively. Maximum adsorption of dyes, i.e. ≥ 90% has been achieved in aqueous solutions by using 0.1 and 0.3 g/L of clay at pH 7 (for both NR and MG) and 6 for MB respectively. The adsorption of dyes followed the pseudo-secondorder rate equation.
Acknowledgements
The authors thank U.S.T.MB and DR T.M Universities for their material supports.
References
Srivastava S. , Sinha R. , Roy D. 2004 Aquatic Toxicology 66 (3) 319 - 329    DOI : 10.1016/j.aquatox.2003.09.008
Dusart O. , Serpaud B. 1991 La tribune de l’eau 44 (554) 15 - 22
Linsheng Z. , Dobias B. 1992 Water Treatment 7 221 - 232
Ciardelli G. , Corsi L. , Marucci M. 2000 Resour. Conserv. Recy. 31 189 - 197    DOI : 10.1016/S0921-3449(00)00079-3
Paprowicz J. , Slodezyk S. 1988 Env. Tech. Let. 9 271 - 280    DOI : 10.1080/09593338809384567
Milano J. C. , Loste-Berdot P. , Vernet J. L. 1994 Environ. Techn. 16 329 - 341    DOI : 10.1080/09593330.1995.9618234
Yener J. , Kopac T. , Dogu T. 2006 Journal of Colloid and Interface Science 294 (2) 255 - 264    DOI : 10.1016/j.jcis.2005.07.040
Aksu Zümriye 2005 Process Biochemistry 40 (3-4) 997 - 1026    DOI : 10.1016/j.procbio.2004.04.008
Ho Y. S. , Mckay G. A. 1999 Resources, Conservation and Recycling 25 (3) 171 - 193
HO Yuh-Shan , Chiang Tzu-Hsuan , Hsueh Yu-Mei 2005 Process Biochemistry 40 (1) 119 - 124    DOI : 10.1016/j.procbio.2003.11.035
Arami M. , Yousefi L. , Nargess M. , Niyaz M. , Tabrizi N. S. 2005 Journal of Colloid and Interface Science 288 (2) 371 - 376    DOI : 10.1016/j.jcis.2005.03.020
Banat F. , Al-asheh S. , Al-makhadmeh L. 2003 Process Biochemistry 39 (2) 193 - 202
Garg V. K. , Kumar R. , Gupta R. 2004 Dyes and Pigments 62 (1) 1 - 10    DOI : 10.1016/j.dyepig.2003.10.016
Gong Renmin , Sun Yingzhi , Chen Jian , Liu Huijun , Yang Chao 2005 Dyes and Pigments 67 (3) 175 - 181    DOI : 10.1016/j.dyepig.2004.12.003
Bhattacharyya Krishna, G. , Sarma Arunima 2003 Dyes and Pigments 57 (3) 211 - 222    DOI : 10.1016/S0143-7208(03)00009-3
Mittal A. , Krishnan L. , Gupta V. K. 2005 Separation and Purification Technology 43 (2) 125 - 133    DOI : 10.1016/j.seppur.2004.10.010
Lucia R. , Martin P. , Miroslav H. , Jozef A. 2009 Nova Biotechnologica 9 - 1
Santhi T. , Manonmani S. , Smitha T. , Mahalakshmi K. Journal of Applied Sciences in Environmental Sanitation 4 (3) 263 - 271
Arivoli S. 2009 The Arabian Journal for Science and Engineering 34 (2A)
Guo Tao N. N. , Y. P. Zhang H. , Liu Y. H. , Qi J. R. , Wang Z. C. , Xu H. D. 2003 Mater. Chem. Phys. 82 (1) 107 - 115    DOI : 10.1016/S0254-0584(03)00191-3
Safaric I. , Safarikova M. , Vrchotova N. 1995 Collect. Czech. Chem. Commun. 60 34 - 42    DOI : 10.1135/cccc19950034
Elaziouti A. , Laouedj N. , Segheir A. A. 2010 Journal of the Korean Chemical Society 54 (1) 115 - 119    DOI : 10.5012/jkcs.2010.54.01.115
Rauf N. , Ikram M. , Tahir S. S. 1999 Adsorp. Sc. Technol. 17 431 - 440
Rauf M. A. , Ikram M. , Rauf N. 2003 J. Trace Microprobe Teh. 20 119 - 125
Rauf M. A. , Iqbal J. , Ikram M. , Rauf N. 2003 J. Trace Microprobe Teh. 21 337 - 342    DOI : 10.1081/TMA-120020267
Lopez A. F. , Arbeloa T. , Martinez V. M. , Banuelos P. J. , Lopez A. I. 2005 Clay science 12 Supplement 1 31 - 37
Elaziouti A. , Derriche Z. , Bouberka Z. , Laouedj N. 2010 Journal of the Korean Chemical Society 54 (1) 110 - 114    DOI : 10.5012/jkcs.2010.54.01.110
Garg V. K. , Kumar R. , Gupta R. 2004 Dyes and Pigments 62 (1) 1 - 10    DOI : 10.1016/j.dyepig.2003.10.016
Elaziouti A. , Derriche Z. , Bouberka Z. , Laouedj N. 2010 Journal of the Korean Chemical Society 54 (1) 110 - 114    DOI : 10.5012/jkcs.2010.54.01.110
Schoonheydt R. 1995 A. Mineral surfaces; Vanghen, D. J., Pattrick, R. A. D., Eds. Chpman-hall London 303 -
Meifang H. , Beiyan M. , Fangbai L. 2003 Geology-Geochemistry 30 (1) 70 - 73
Bergman K. , O’Konski C. T. 1963 J. phys. Chem. 67 2169 -    DOI : 10.1021/j100804a048
Rytwo G. , Nir S. , Crespin M. , Margulies L. 2000 Journal of Colloid and Interface Science 222 12 - 19    DOI : 10.1006/jcis.1999.6595
Yates D. E. , Levine S. , Healy T. W. 1974 J. Chem. Soc. Faraday Trans. 70 1807 - 1818    DOI : 10.1039/f19747001807
Davis A. , James R. O. , Leckie J. O. 1978 J. Colloid Interface Sci. 36 480 - 499    DOI : 10.1016/S0021-9797(78)80009-5
Allen S. J. , Koumanova B. 2005 J of the university of chemical Technology and metallurgy 40 (3) 175 - 195
Renmin Gong , Mei Li , Chao Yang , Yingzhi Sun , Jian Chen 2005 Journal of Hazardous Materials B 121 247 - 250    DOI : 10.1016/j.jhazmat.2005.01.029
Namasivayam C. , Prabba D. , Kumutha M. 1998 Bioresour. Technol. 48 79 - 81    DOI : 10.1016/0960-8524(94)90141-4
Lagergren S. 1989 Kungl. Sv. Vetensk. Handl. 24 1 - 39
Ho Y. S. , Mckay G. A. 1999 Resources, Conservation and Recycling 25 (3) 171 - 193
Aslıhan Gücek , Sava¸s Sener , Sedat Bilgen , Ali Mazmancı M. 2005 Journal of Colloid and Interface Science 286 53 - 60    DOI : 10.1016/j.jcis.2005.01.012