Evaluation of sorption of uranium onto metakaolin using X-ray photoelectron and Raman spectroscopies

Analytica Chimica Acta 631 (2009) 69–73

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Evaluation of sorption of uranium onto metakaolin using X-ray photoelectron and Raman spectroscopies Jamil R. Memon a,∗ , Keith R. Hallam b , Muhammad I. Bhanger a , Adel El-Turki b , Geoffrey C. Allen b a b

National Center of Excellence in Analytical Chemistry, University of Sindh, Jamshoro, Sindh 76080, Pakistan Interface Analysis Centre, University of Bristol, 121 St. Michael’s Hill, Bristol BS2 8BS, UK

a r t i c l e

i n f o

Article history: Received 2 July 2008 Received in revised form 8 October 2008 Accepted 8 October 2008 Available online 18 October 2008 Keywords: Metakaolin Uranium Sorption Kinetics Isotherms

a b s t r a c t Metakaolin prepared from a natural clay mineral ore of aluminium kaolinite is a promising low cost and high activity aluminosilicate material that has been investigated for studying the sorption behavior of uranium. Here, metakaolin was characterized using X-ray photoelectron spectroscopy (XPS) and the effects of pH, contact time and initial metal ion concentration on its sorption behavior were studied. The sorption process was found to initially be rapid (∼60% at time 0 min) but became slower with time; equilibrium was established within 24 h (∼80% sorption). The data were applied to study the kinetics of the sorption process. The Langmuir and Dubinin–Radushkevich (D–R) sorption isotherms were used to describe partitioning behavior for the system at room temperature. The binding of metal ions was found to be pH dependent, with optimal sorption occurring at pH 5. The retained metal ions were eluted with 5 mL of 0.1 M HNO3 . Raman spectroscopy and XPS were used to evaluate the sorption mechanism of U(VI). © 2008 Elsevier B.V. All rights reserved.

1. Introduction Uranium is primarily used as a nuclear fuel and (in depleted form) in military ordinance. It is present naturally in air, water, food and soil. Uranium has five oxidation states, though only the 4+ and 6+ forms are stable. The 6+ uranyl ion (UO2 2+ ) forms water-soluble compounds. Naturally occurring uranium, although radioactive, presents predominantly a toxicologic rather than radiological health risk [1]. Concentrations of uranium can be accumulated during mining and processing operations, which usually involve large volumes of wastewater. The World Health Organization (WHO) has regulated the maximum concentration level of uranium in drinking water to be 15 ng mL−1 [2]. Accurate analytical determination of the presence of uranium at trace or sub-trace levels requires versatility, specificity, sensitivity and accuracy. However, in certain cases direct determination of uranium is not possible due to matrix interferences and its low concentration in analyte samples. These problems may be overcome by employing suitable treatment processes, such as solvent extraction, precipitation, coagulation–filtration, reverse osmosis, hyperfiltration, electrodialysis, ion-exchange and adsorption. Among these processes, adsorption is simple, rapid and can be used repeatedly. Removal and recovery of uranium by sorption has been studied on: date pits [3]; citrobacter freudii [4]; olive cake

∗ Corresponding author. Tel.: +92 22 2771379; fax: +92 22 2771560. E-mail address: [email protected] (J.R. Memon). 0003-2670/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2008.10.017

[5]; coffee residue [6]; activated carbon [7]; marine algae (Cystoseira indica) [8]; coir pith [9]; Ulva sp.-Na bentonite [10]; activated charcoal [11]; used tires [12]; and filamentous fungus [13]. Metakaolin prepared from a natural clay mineral ore of aluminium kaolinite has been utilized in the present initial investigation as an inexpensive adsorbent for the removal and recovery of uranium, in particular for the treatment of wastewater. The effects of relevant parameters (pH, contact time and initial metal ion concentration) are investigated and the analytical applicability of the method is demonstrated. 2. Materials and methods 2.1. Materials All the chemicals were supplied by Merck (Darmstadt, Germany) and of analytical grade. A stock standard solution of U(VI) was prepared by dissolving the appropriate amount of its acetate salt in deionized water (conductivity <0.5 ␮S cm−1 ), acidified with a small amount of nitric acid. Buffer solutions of pH 1–3, 4–6 and 7–9 were prepared by mixing appropriate ratios of 0.1 M HCl and KCl, 0.5 M acetic acid and sodium acetate and 0.5 M ammonia and NH4 Cl solutions, respectively. 2.1.1. Metakaolin Metakaolin is formed by the dehydroxylation of kaolin (Al2 (OH)4 Si2 O5 ) precursor upon heating in the ∼700–800 ◦ C tem-

70

J.R. Memon et al. / Analytica Chimica Acta 631 (2009) 69–73

perature range [14]. Metakaolin used in our study was obtained from Imerys Minerals (Cornwall, UK). Metakaolin particle size was 5 ␮m while its specific surface area determined through the BET method [15] was found to be 12 m2 g−1 . Differential thermal analysis (DTA) showed that the metakaolin decomposed at ∼985 ◦ C while X-ray diffraction (XRD) analysis revealed the metakaolin to be comprised of mica, quartz and feldspar, as previously reported in the literature [16]. 2.2. Methods 2.2.1. Chemical analysis A HORIBA Jobin Yvon (Stanmore, UK) inductively coupled plasma atomic emission spectrophotometer (ICP-OES) model Ultima was used to determine the concentration of uranium in solution. ICP-OES analysis was performed after setting of the experimental parameters (uranium,  = 385.958 nm). As per the operation manual, the instrument was allowed to warm up for about 30 min, with deionized water being aspirated all the time. Before analysis, a calibration curve was constructed for the uranium ions. All pH measurements were made with a WTW (Weilheim, Germany) digital pH meter model inolab pH level 1 equipped with a calibrated combined pH glass electrode. A Barloworld Scientific (Staffordshire, UK) Stuart rotator model SB3 was used for the batch experiments. X-ray photoelectron spectroscopy (XPS) was used to identify the major chemical composition and any minor constituents of the outer few nanometers of the surfaces. Areas of approximately 4 mm × 3 mm on each sample were analyzed using Thermo Fisher Scientific (East Grinstead, UK) an Escascope spectrometer, with an Al K␣ (1486.6 eV) X-ray source operated at 280 W (14 kV, 20 mA); the analyzer had a pass energy of 30 eV. Wide scan survey spectra were obtained between 1100 and 0 eV binding energy (BE) with a step size of 1.0 eV. Subsequent higher energy resolution scans over variable BE ranges with 0.1 eV steps were recorded for specific elements of interest: C 1s, O 1s, etc. The operating vacuum during analysis was approximately 5 × 10−9 mbar. All spectra were corrected for charging effects with respect to the adventitious hydrocarbon C 1s peak at 284.8 eV. PISCES version 2000.2 software (Dayta Systems) was used to carry out peak fitting and quantitative analysis, with the relative sensitivity factors used being those originally supplied by the instrument manufacturer. Sample materials were characterized using a Renishaw (Wotton-under-Edge, UK) Ramascope spectrometer model 2000. The system was equipped with He–Ne (633 nm) laser as an excitation source. The maximum laser power was 25 mW. The analysis was performed by focusing the laser with objective magnification 50× onto the sample surface through an Olympus BH2-UMA optical microscope, corresponding to a laser spot diameter of about 4 ␮m. The laser power at the specimen surface was of the order of 4 mW and an acquisition time of 10 s was used for each spectrum over the wave number range 100–4000 cm−1 . Prior to analysis, the spectrometer was calibrated using a monocrystalline silicon standard specimen. Peak fitting and deconvolution of Raman spectra were performed using GRAMS32 software. 2.2.2. Metal sorption experiments The metal sorption behavior of metakaolin was investigated using batch equilibrium experiments. 0.1 g of metakaolin was equilibrated with 20 mL of U(VI) solution (100 mg L−1 ) maintained at optimum pH and 20 ± 1 ◦ C and shaken at 40 rpm for the designated time (0–24 h). The mixtures were filtered and adsorbed metal ions were desorbed by agitating with 5 mL of 0.1 M HNO3 solution before being analyzed after dilution by ICP-OES. Experiments were con-

ducted in triplicate and the results presented here are the average of triplicate measurements. Precision in all cases is close to ±1%. 3. Results and discussion 3.1. X-ray photoelectron spectroscopy of untreated and uranium-treated metakaolin In order to identify the nature of components present in metakaolin, X-ray photoelectron spectroscopy was used. A typical wide scan XPS spectrum (Fig. 1) shows the presence of major peaks due to C, O, Si and Al and these were subsequently analyzed separately in more detail. The XPS peaks obtained for O 1s, Si 2p and Al 2p were centered at 531.7, 103.1 and 75.2 eV binding energies, respectively, and were assigned to the presence of Al2 O3 and SiO2 in kaolinite [17]. The C 1s peak was ascribed to the presence of adventitious hydrocarbon and set at 284.8 eV for sample charging correction purposes. The atomic concentrations of these elements, along with lesser amounts of K, were calculated from the peak areas, and application of appropriate relative sensitivity factors, and found to be: C 15.8%; O 28.0%; Al 18.8%; Si 31.7%; and K 5.8%. The metakaolin supplied by Imerys contained naturally the relatively low level of potassium, as has been previously reported in the literature [16]. U(VI)-adsorbed metakaolin was also analyzed using XPS to confirm whether U(VI) adsorbed on metakaolin or if it was reduced to U(IV). One can differentiate the various chemical oxidation states of the same element present in a sample by measuring the precise binding energies in the XPS spectra. The XPS wide scan spectrum (Fig. 2a) revealed the presence of all the above mentioned XPS peaks along with those characteristic of uranium. The corresponding U 4f peaks are presented in Fig. 2b. It should be noted that spin-orbit coupling effects in the final state lead to splitting of p-, d- and f-peaks in XPS spectra. In general, the more intense, lower binding energy component (here the U 4f7/2 ) peak is used to characterize the oxidation states present. The U 4f7/2 binding energy value was found to be 381.4 eV, which we assign to UO3 [17], indicating that the original U(VI) remained in the same oxidation state upon adsorption. The element atomic composition from XPS was measured to be: C 18.9%; O 25.0%; Al 21.0%; Si 34.8%; K 0.3%; and U 0.1%. The reduction in amount of K and O and the Si:Al ratio may be due to the dissolution as KOH into aqueous solution and masking of the metakaolin surface with adsorbed U(VI) ions, respectively. 3.2. Influence of the aqueous solution pH on the uranium sorption pH is one of the most important parameters when assessing the adsorption capacity of an adsorbent for metal ions sequestered

Fig. 1. Wide scan X-ray photoelectron spectrum of metakaolin.

J.R. Memon et al. / Analytica Chimica Acta 631 (2009) 69–73

71

Table 1 Reagents used for the elution of U(VI) ions. Reagenta

Concentration (M)

Recovery (%)

NaOH NaOH EDTA EDTA H2 SO4 H2 SO4 HCl HCl HNO3 HNO3

0.1 0.5 0.1 0.5 0.05 0.1 0.05 0.1 0.05 0.1

25 75 30 80 45 81 37 75 53 100

Bold value indicates the only one reagent whose lower concentration recovers maximum amount of uranium from the surface of metakaolin. a Volume of each reagent used = 5 mL.

main U(VI) species is UO2 2+ (70%) while at pH > 5, (UO2 )3 (OH)5 + is dominant [18]. 2. In the presence of atmospheric CO2 , uranium forms strong carbonate aqueous complexes with dissolved CO3 2− such as UO2 (CO3 ), UO2 (CO3 )2− or UO2 (CO3 )4− . These neutral and anionic complexes are repelled from any negatively charged sorbent surface. 3.3. Recovery of uranium from metakaolin surface Fig. 2. (a) Wide scan XPS of uranium adsorbed metakaolin and (b) U 4f XPS spectrum.

from aqueous solution due to its influence on the surface properties of the adsorbent and ionic forms of the metal in solution. Here, sorption experiments were carried out in the pH range 1–9 keeping all other parameters constant. The sorption of uranium was seen to be critically dependent on pH. As shown in Fig. 3, maximum sorption took place at pH 5. A value of pH above or below is unfavorable owing to the following reasons. At pH < 5, the higher concentration of H+ ions effectively leads to silica and alumina being less available for binding to metal ions. Increase in pH to 5 results in more silica and alumina being available for metal ion binding and, hence, sorption was enhanced. At pH > 5, U(VI) uptake was decreased due to two main reasons: 1. Uranium is hydrolyzed and may precipitate depending on the pH of the solution. In acidic to near neutral pH range, the four major forms of uranium are: UO2 2+ ; (UO2 )2 (OH)2 2+ (pKa = 5.62); UO2 OH+ (pKa = 5.8); and (UO2 )3 (OH)5 + (pKa = 15.63). These, together with a dissolved solid schoepite (4UO3 ·H2 O)—a hydrous uranium oxide, may all exist in the solution [18]. At pH 5, the

Desorption of uranium from the metakaolin surface was studied by using 5 mL of different concentrations of HNO3 , H2 SO4 , HCl, NaOH and EDTA. Recoveries are given in Table 1. Elution was found to be complete (100%) with 5 mL of 0.1 M HNO3 . NaOH forms precipitating hydroxides with U(VI) ions by increasing the pH. However, because of the low concentration of NaOH used, it cannot completely remove the U(VI) ions. U(VI) forms complexes with EDTA but again, because of its low concentration, it does not completely recover the U(VI) ions from the surface. H2 SO4 and HCl do not formed soluble U(VI) compounds whereas HNO3 forms soluble nitrates. Therefore, a very low concentration of HNO3 can recover the U(VI) ions from the metakaolin surface. 3.4. Kinetics of sorption for U(VI) onto metakaolin Kinetic studies were carried out under optimized conditions from 0 to 24 h, at pH 5, using 20 mL of 100 mg L−1 U(VI) solution and 0.1 g of adsorbent. Sorption of U(VI) was initially rapid and then became slower until equilibrium was established within 24 h; there was no increase in equilibrium percentage sorption. Similar results have also been previously reported in the literature [10]. Therefore, all further experiments were carried out for 24 h. The recorded data were fitted to two different sorption kinetic models, namely Morris–Weber and Lagergren. The validity of these models is based on the regression coefficient of the plots and is ≥0.99 indicating that the data follows both kinetic sorption models. The adsorbed concentration at time t, qt (␮mol g−1 ), was plotted against √ t to test the Morris–Weber equation [19] in the following form: √ (1) qt = Rd t where Rd is the rate constant of intraparticle transport. Up to 24 h, Eq. (1) held well, with a regression coefficient of 0.99, but it deviated as the agitation time increased. From the slope of the plot in the initial stage (Fig. 4), the value of Rd was estimated to be 1.681 ± 0.064 ␮mol g−1 h−1/2 for U(VI). The Lagergren equation [20]:

Fig. 3. Sorption of U(VI) ions onto metakaolin as a function of pH.

log (qe − qt ) = log qe −

kt 2.303

(2)

72

J.R. Memon et al. / Analytica Chimica Acta 631 (2009) 69–73

Fig. 4. Morris–Weber plot for the sorption of U(VI) ions onto metakaolin. Fig. 6. Raman spectra of untreated and uranium-treated metakaolin.

fore, the Langmuir and D–R isotherms were used to evaluate the sorption capacity of metakaolin for U(VI) ions. The Langmuir and D–R constants were evaluated using the slopes and intercepts of the linear plots studied at room temperature; results are listed in Table 2. The essential characteristic of the Langmuir isotherm can be explained in terms of a dimensionless constant separation factor (RL ), calculated using the equation RL = 1/(1 + bCi ) where Ci is the initial concentration of metal ions. RL describes the type of the Langmuir isotherm [21] to be irreversible (RL = 0), favorable (0 < RL > 1), linear (RL = 1) or unfavorable. The values of RL calculated were between 0.98 and 0.03, indicating highly favorable sorption of U(VI) ions onto the metakaolin surface. The value of E evaluated  from the slope (ˇ) of the D–R plot using equation (E = Fig. 5. Lagergren plot for the sorption of U(VI) ions onto metakaolin.

was tested by plotting log (qe − qt ) against time t (Fig. 5), where qe is the adsorbed concentration of uranium on metakaolin (3.365 × 10−5 mol g−1 ) at equilibrium. The overall value of rate constant (k) was estimated to be 0.121 ± 0.002 h−1 from the slope of the plot with a regression coefficient of 0.99. 3.5. Sorption isotherms The sorption of U(VI) ions was also investigated as a function of its concentration at room temperature in the range 0.5–1000 mg L−1 using 0.1 g of adsorbent, 20 mL of adsorbate solution and 24 h shaking time. The metal ion uptake increased with increasing concentration up to 10 mg L−1 but then a decreasing trend was observed. These results reflect the efficiency of the adsorbent for the removal of uranium ions from aqueous solution. The effect of the tested metal ion concentration on metakaolin was analyzed in terms of the Freundlich (log Cads = log A + (1/n) log Ce ), Langmuir ((Ce /Cads ) = (1/Qb) + (Ce /Q)), and Dubinin–Radushkevich (D–R) (ln Cads = ln Xm − ˇε2 ) and ε = RT ln (1 + (1/Ce )) equations, where Cads is the concentration of metal ions sorbed per unit mass of sorbent, Ce is the concentration of metal ions in the liquid phase at equilibrium and A and n, Q and b, and Xm and ˇ are the Freundlich, Langmuir and D–R constants, respectively. The sorption data was seen to follow the Langmuir and D–R isotherms. There-

1/ −2ˇ) was 11.6 ± 0.3 kJ mol−1 , which is in the range expected for chemisorption or ion-exchange (9–16 kJ mol−1 ) [22]. Hence, it was very likely that the metal ions were sorbed onto metakaolin predominantly by ion-exchange. The calculated values of sorption capacity were different and in the range of 0.77–1.4 mmol g−1 . This difference in sorption capacity can be interpreted in terms of the assumptions taken into consideration while deriving these sorption models. Comparison of surface area and sorption capacity of U(VI) on metakaolin with data from previously reported sorbents (Table 3) highlights the efficiency of this cheaper material. 3.6. Reusability of metakaolin for U(VI) In order to check the reusability of the adsorbent, metakaolin was subjected to several loading and elution experiments. The capacity of the adsorbent was found to be practically constant (variation of 1–3%) after 10 times repeated use; thus multiple use of the adsorbent was seen to be feasible. 3.7. Evaluation of mechanism for sorption of U(VI) ions onto metakaolin using Raman spectroscopy

In order to confirm the species responsible for the sorption of uranium ions, Raman spectroscopy was used. Raman spectra (Fig. 6) of as-received metakaolin depicted a low intensity peak at 513.3 cm−1 , assigned to SiO2 , while the same was diminished further in intensity after the uranium treatment due to masking,

Table 2 Langmuir and D–R constants of U(VI) ions onto metakaolin at room temperature. Langmuir

Dubinin–Radushkevich −1

Q (mmol g

0.77 ± 0.03

)

−1

b × 10 (L g 3

8.9 ± 4.3

)

RL

R

Xm (mmol g−1 )

ˇ × 10−3 (mol2 kJ−2 )

E (kJ mol−1 )

R2

0.98–0.03

0.99

1.4 ± 0.1

−(3.7 ± 0.1)

11.6 ± 0.3

0.99

R2 : regression coefficient; RL : dimensionless constant.

2

J.R. Memon et al. / Analytica Chimica Acta 631 (2009) 69–73

73

Table 3 Comparison of surface area and sorption capacity of U(VI) onto metakaolin with previously reported sorbents. Sorbent

Surface area (m2 g−1 )

Sorption capacity (mg g−1 )

Reference

Date pits Citrobacter freudii Olive cake Coffee residue Activated carbon Ulva sp.-Na bentonite Activated charcoal Char produced from used tires Filamentous fungus

Not reported Not reported Not reported 3.28 965–1200 Not reported 364.17 67.1–75.3 Not reported

10 34.43 71.4 40.5 10.47 54 28.49 22.13 116.5

[3] [4] [5] [6] [7] [10] [11] [12] [13]

Metakaolin

12

183.3

This work is based on the Langmuir isotherm

Table 4 Determination, removal and recovery of U(VI) from uranium ore samples. Reference number of the ore samples

Found (␮g mL−1 )

38834 38835 38842 38849 38850

106 626.6 423 1273.7 1718

Removal (%)

tively be used to remove and recover U(VI) ions from uranium ore samples.

Recovery (%)

Acknowledgements 80 75 78 70 68

100 99.5 100 99 98.8

indicating that SiO2 is the main species responsible for uranium adsorption.

The authors acknowledge the financial support to Jamil R. Memon received from Higher Education Commission (HEC), Islamabad, Pakistan under their International Research Support Initiative Program (IRSIP). The Interface Analysis Centre, University of Bristol, Bristol, UK and National Center of Excellence in Analytical Chemistry, University of Sindh, Jamshoro, Pakistan are acknowledged for providing access to instrument facilities for this work.

3.8. Analytical application of the developed method

References

The analytical applicability of metakaolin was tested with five uranium ore samples obtained from Atomic Energy Mineral Centre, Lahore, Pakistan. After acid digestion, the samples were analyzed using ICP-OES (Table 4, found values). A 20 mL aliquot of each sample was adjusted to pH 5 and then agitated with 1 g of metakaolin for 24 h (Table 4, removal values). The metal ions were then eluted with 5 mL of 0.1 M HNO3 and determined by ICP-OES (Table 4, recovery values). The R.S.D. was always within 2%, thus clearly showing the efficiency of metakaolin for the removal and recovery of U(VI) ions from these ore samples. 4. Conclusions Removal and recovery of U(VI) ions was successfully accomplished using a cheaper adsorbent – metakaolin. Low cost of analysis, rapid attainment of phase equilibrium and high sorption capacity values were some of the significant features. X-ray photoelectron spectroscopy showed that uranium remained in the same oxidation state upon adsorption. Raman spectroscopy of uraniumtreated metakaolin confirmed that silica was the main species responsible for sorption of uranium ions. The kinetics of sorption for uranium followed a pseudo first order rate equation. Sorption of U(VI) on metakaolin followed the Langmuir and D–R isotherms. The energy value obtained from the D–R isotherm showed that the sorption was ion-exchange in nature. Metakaolin can effec-

[1] C.D. Klaassen, Toxicology the basic science of poisons, McGraw-Hill, New York, 2001. [2] WHO, Guidelines for Drinking Water Quality, 3rd ed., 2004, p. 454. [3] E.M. Saad, R.A. Mansour, A. El-Asmy, M.S. El-Shahawi, Talanta 76 (2008) 1041. [4] S. Xie, J. Yang, C. Chen, X. Zhang, Q. Wang, C. Zhang, J. Environ. Radioact. 99 (2008) 126. [5] M. Konstantinou, I. Pashalidis, J. Radioanal. Nucl. Chem. 273 (2007) 549. [6] M. Mahramanlioglu, I.O. Bicer, T. Misirli, A. Kilislioglu, J. Radioanal. Nucl. Chem. 273 (2007) 621. [7] A. Mellah, S. Chegrouche, M. Barkat, J. Colloid Interface Sci. 296 (2006) 434. [8] M.H. Khani, A.R. Keshtkar, B. Meysami, M.F. Zarea, R. Jalali, Electron. J. Biotech. 9 (2006) 100. [9] H. Parab, S. Joshi, N. Shenoy, R. Verma, A. Lali, M. Sudersanan, Bioresour. Technol. 96 (2005) 1241. [10] R. Donat, S. Aytas, J. Radioanal. Nucl. Chem. 265 (2005) 107. [11] C. Kütahyali, M. Eral, Sep. Purif. Technol. 40 (2004) 109. [12] M. Mahramanlioglu, J. Radioanal. Nucl. Chem. 256 (2003) 99. [13] E. Guibal, C. Roulph, P. Le Cloirec, Water Res. 26 (1992) 1139. [14] J. Ambroise, S. Maximilien, J. Pera, Adv. Cem. Based Mater. 1 (1992) 161. [15] P.C. Hiemens, R. Rajagopalan, Principles of Colloid and Surface Chemistry, Marcel Dekker Inc., New York, 1997. [16] A. Bakolas, E. Aggelakpoulou, A. Moropoulou, S. Anagnostopoulou, J. Therm. Anal. Calorim. 84 (2006) 157. [17] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer Corporation, USA, 1992, p. 213–241. [18] C.F. Baes Jr., R.E. Mesmer, The Hydrolysis of Cations, Wiley-Interscience, New York, 1976. [19] W.J. Morris, C. Weber, J. Saint. Eng. Div. ASCE 89 (1963) 31. [20] Y. Ho, Scientometrics 59 (2004) 171. [21] K. Kadirvelu, K. Thamaraiselvi, C. Namasivayam, Sep. Purif. Technol. 24 (2001) 497. [22] S.M. Hasany, M.M. Saeed, M. Ahmed, Talanta 54 (2001) 89.