Spectrophotometric Determination of Trace Amounts of ... - J-Stage

ANALYTICAL SCIENCES OCTOBER 2009, VOL. 25

1249

2009 © The Japan Society for Analytical Chemistry

Spectrophotometric Determination of Trace Amounts of Fluoride Using an Al-Xylenol Orange Complex as a Colored Reagent Javad ZOLGHARNEIN,*† Akram SHAHRJERDI,* Gholamhassan AZIMI,* and Jahanbakhsh GHASEMI** *Department of Chemistry, Faculty of Science, Arak University, Arak 38156-879, Iran **Department of Chemistry, Faculty of Science, Khajeh-Nasir Toosi University of Technology, Tehran, Iran

A new spectrophotometric reagent for the determination of trace amounts of fluoride has been introduced. This method is based on the decolorization of a complex of Al(III) with xylenol orange (XO) as an ultra-sensitive colored reagent. Since the Al-XO complex plays an important role in this method, the protonation and complexation of XO with Al(III) at an ionic strength of 0.1 mol L–1 at 25° C has been studied by a spectrophotometric global analysis method. The EQUISPEC program was used to evaluate the protonation constants of XO and the stability constants of the formed complexes with Al(III). The protonation and the stability constants of the major complex species such as ML, MLH and MLH2, were determined. Finally, a spectrophotometric method for the assay of fluoride based on a decrease of the color intensity of the Al-XO complex, in an aqueous solution has been designed. The effects of some important variables on the determination of fluoride based on the proposed method were investigated. The method was applied to the determination of fluoride under the optimum conditions (pH 5.2, ionic strength 0.1 mol L–1, 25° C). The determination of fluoride in the range of 0.08 – 1.4 μg mL–1 (SD = 1.2%) was successfully performed. Interferences of Fe(III) were easily eliminated by using ascorbic acid. The proposed method was applied to the determination of trace amounts of fluoride content of some real water samples. (Received July 14, 2008; Accepted November 20, 2008; Published October 10, 2009)

Introduction The determination of trace amounts of fluoride, because of its importance in health, medical and industrial processes, is of interest to analytical chemists. Fluoride is an important parameter of water quality. At low concentration it is beneficial, while at high concentration it is toxic.1 Fluoride plays a central role in the prevention of dental caries, and is regularly employed for this purpose in drinking water.1 Compounds of fluoride are very valuable and extensively utilized in human activities. The industrial effluent, combustion of coal and sewage discharged from domestic water supplies, contribute to an increase in fluoride levels in local aquatic systems. The health effect of fluoride, such as dental fluorosis and bone fracture has attracted considerable attention. This ion is considered to be essential to both plants and animals at low concentrations.2 In unpolluted fresh and seawater, it generally exists in the ranges of 0.01 – 0.3 mg L–l and 1.2 – 1.5 mg L–1, respectively.1 Thus, a continuous monitoring of the presence and movement of fluoride in the environment is of great importance. The most common techniques for fluoride assay are potentiometry using a fluorideselective electrode (ISE)3–5 and ion chromatography (IC).6 Several other techniques have been reported for the determination of fluoride in drinking water, such as capillary electrophoresis,7 solvent-extraction coupled to fluorometry.8–10 There are a few spectrophotometric reports in the literature based on a decolorizing effect of fluoride on a colored metal-dye To whom correspondence should be addressed. E-mail: [email protected] †

complex.1,11–15 Some of these mentioned techniques have special problems in determination of fluoride.14 Because of some advantages of spectrophotometric methods, such as simplicity, facility, accuracy and reproducibility, we have been interested to design a new spectrophotometric method for fluoride assay. Spectrophotometry is a powerful technique for investigating of solution equilibria.16,17 Nowadays, spectroscopic instrumentation has the capacity to collect data in a full spectral range. Using a single or a few wavelengths discards most of the information in the collected spectra, and requires both the presence and knowledge of such a suitable wavelength. However, in many cases, the spectral responses of components overlap, and thus analysis is no longer straight-forward.17–20 To overcome this problem, using the chemometrics method we can analyze whole spectra and can thereby utilize all spectra information.17–22 EQUISPEC, which is a hard modeling in the MATAB environment, involves simultaneous, global analysis of a series of spectrophotometric titrations. Each titration is performed under different initial conditions, such as at different pH values or concentrations which has been called second order global analysis.16 This program was used to evaluate the protonation and stability formation constants of the complexation reaction between xylenol orange (XO) and Al(III) in aqueous solutions. Thus, according to a high tendency of Al(III) with fluoride ion, AlF63–, the complex of XO with Al(III) was used as a colored probe. There are few reports about spectrophotometric studies of the XO reaction with Al(III) especially using chemometrics methods described in the literature.1,14,23,24 The major complex species, recognized as the best of them, was chosen as a colored probe. The determination of fluoride based on a decrease of the color intensity of the Al-XO complex was

1250

Scheme 1


Chemical structure of xylenol orange.

planned.14,15 The optimum conditions for a trace amount of fluoride assay in some drinking water samples have been performed.

Experimental Reagents and solutions Xylenol orange (3,3′-bis[N,N′-di(carboxymethyl)aminomethyl]o-cresolsulfonphthalein) or (XO, Scheme 1) and aluminum nitrate nonahydrate were of commercial analytical grade and used without any further purification. A constant ionic strength at 0.1 mol L–1 was maintained with KNO3 as a background electrolyte. Sodium fluoride was dried at 140° C for 48 h, before using. All fluoride solutions were stored in polyethylene containers. The pH of the medium was adjusted by using 0.1 mol L–1 HNO3 and NaOH solutions. Solutions of diverse ions used for interference studies were prepared from nitrate salts of cations and the sodium salts of anions. L-Ascorbic acid solution (10%) was used for controlling the interference of Fe(III). Triply distilled water was used throughout. All chemical reagents were from Merck. A stock solution of 5 × 10–3 mol L–1 xylenol orange (XO) was prepared by the dissolution of 95 mg of XO in a 25-mL flask and adding triply distilled water. A stock solution of Al(III) is prepared by the dissolution of 46.8 mg of aluminum nitrate nonahydrate in a 25-mL flask and adding triply distilled water. A mixed solution of XO-Al (1:1) 5 × 10–5 mol L–1 was prepared by adding 55.5 mg of KNO3 into a 50-mL flask containing 0.5 mL of both Al(III) and an XO stock solution. Apparatus Absorption spectra were recorded on an Agilent (HP) 8453 spectrophotometer in the range of 360 – 610 nm equipped with a 1.0 cm quartz cell. The cell was thermostated at 25 ± 0.5° C. A Metrohm 691 digital pH meter with glass electrode calibrated on the operational pH scale with standard buffer solutions (pH 3 and 11) was used. EQUISPEC in the MATLAB (Ver. 6.5) environment was used for data analysis. Spectroscopic data analysis The EQUISPEC program was introduced by Dyson et al.16 The spectra recorded in each spectrophotometric titration (collection of spectra as a function of pH) were arranged in a data matrix Y. The matrix Y, according to Beer’s law (Y = CA + R), can be decomposed into the product of a concentration matrix C and matrix A of molar absorptivities. The matrix R of residuals is due to instrumental noise and experimental errors. Data fitting consisted of determining those unknown parameters for which the sum of squares over all elements of the matrix R of residual was minimal. Calculations were done by an iterative procedure using the Newton–Gauss–Levenberg/ Marquardt (NGL/M) algorithm of non-linear least-squares

Fig. 1 Visible absorption spectra of 5 × 10–5 mol L–1 of XO at various pH values (1.5, 1.91, 2.13, 2.34, 2.51, 2.71, 2.91, 3.13, 3.30.....11.0, 11.21, 11.43, ,11.66, 11.85, 12.0).

fitting.25 EQUISPEC requires the fulfillment of a previously proposed simple chemical model. This model is defined by the stoichiometries of all species involved in the considered equilbria, and by approximate values of the equilibrium constants (Kc). Upon completion of the calculation, EQUISPEC reports refined log K values and standard deviations of these, along with the minimized square sum, which served as a single numerical guide to a goodness of fit (ssq = ∑∑R(ij)2 = f (Y, model, parameter)).20

Results and Discussion The major approach of this study was to determine trace amounts of fluoride ion in aqueous samples. The Al-XO complex was used as a new colored reagent for fluoride assay. Thus, in the first step, a spectrophotometric study of complexation reaction of XO with Al(III) was performed. Xylenol orange is one of the metallochromic indicators, which can bind to metal ions at both their amino and acidic groups. Due to the existences of more than one chelating system and various acidic and basic properties of the XO molecule, which can form various complexes, the investigation of such a system can be difficult. Although, XO is a colored ligand, most of its complexation studies have been performed by potentiometry.23,24 Potentiometry is still the standard method for the determination of stability constants, while spectrophotometry has the advantage of yielding additional information through spectral characteristics of the species. Some difficulties related to the use spectrophotometric data in the determination of equilibrium constants have been successfully overcome by chemometrics method.16–19,25,26 Determination of protonation constants of XO The spectra of the stepwise addition of standardized NaOH to an aliquot of 5 × 10–5 mol L–1 of acidic solution of XO during titration in the pH range 1.5 – 12 were recorded. A sample of the gathered spectra is shown in Fig. 1. The XO in acidic solution may be shown by H6L, which can be introducing 6 protons in 6 photolytic steps. In the absence of metal ions, depending on the pH range, there are several possible forms in which XO can exist. In addition, XO in an acidic medium can receive three protons (H+). The resulting protonation constants, which are calculated by the EQUISPEC program, are listed in Table 1. The concentration profiles of different forms


1251

Table 1 Obtained result from an EQUISPEC computation for complexation of XO with Al(III) in 0.1 mol L–1 ionic strength at 25° C Protonation constant pK

pK9 (LH) pK8 (LH2) pK7 (LH3) pK6 (LH4) pK5 (LH5) pK4 (LH6) pK3 (LH6) pK2 (LH7) pK1 (LH8)

Functional Calculated group resulta

Stability constant Species

≡NH+ 10.55 ± 0.03 ML ≡NH+ 9.36 ± 0.06 MLH –OH 7.88 ± 0.08 MLH2 –COOH 6.46 ± 0.13 ML2 –COOH 4.17 ± 0.15 –COOH 2.14 ± 0.27 — –COOH — =OH+ — –SO3H

log β

σy

23.49 ± 0.12 29.28 ± 0.11 33.58 ± 0.01 45.54 ± 0.04

0.012 0.012 0.012 0.020

Fig. 3 Pure spectra of 5 × 10–5 mol L–1 of XO in aqueous solution (0.1 mol L–1 ionic strength and 25° C).

σy: Sum of square of residual matrix. a. σy is 0.01.

Fig. 2 Concentration profile vs. pH for 5 × 10–5 mol L–1 of XO in aqueous solution (0.1 mol L–1 ionic strength and 25° C).

of XO and their corresponding spectral profiles are also shown in Figs. 2 and 3, respectively. It can be seen from the concentration profiles (Fig. 2) and Table 1 that the first three acidity constants are too low, and (–SO3H, =OH+, –COOH), and that it is not possible to measured by the presented method. As can be clearly observed from the concentration profiles, by increasing the pH (1.5 – 12), the successive dissociation of XO (H6L → L) would occur. The first three experimentally detectable deprotonation steps (H6L → H3L) relevant to three carboxylic groups, and a forth one, are attributed to the –OH group (H3L). The fifth and sixth protons (H5L, H6L) return to the protonated nitrogen atoms. It can also be seen from the spectral profiles that the spectra of the species H6L, H5L and H4L showed a high degree of overlapping. This appeared as the maximum absorbance at 435 nm (Fig. 3). By increasing the pH, a new absorbance maxima at 560 nm appeared which according to Fig. 3, is probably due to the contributions of H3L, H2L, HL and L. Complexation of XO with AL(III) The spectra of an aliquot of 5 × 10–5 mol L–1 of an Al-XO complex solution, (CM/CL = 1) and (CM/CL = 1/2), were recorded during titration with a standardized NaOH solution in the pH interval 1.5 – 12. Samples of the spectra of the absorbance–pH titration are shown in Figs. 4 and 7. The ionic strength was

Fig. 4 Visible absorption spectra of 5 × 10–5 mol L–1 of Al-XO (CM/CL = 1) complex vs. both the wavelength and the pH in aqueous solution (0.1 mol L–1 KNO3 ionic strength and 25° C).

adjusted at 0.1 mol L–1 by adding KNO3. The results for the stability constants of XO with Al(III) are listed in Table 1. It is evident that the small three valent metal ions, such as Al(III) can strongly react with XO as a multidentate ligand with two different stoichiometries. The concentration and spectral profiles for the 1:1 (CM/CL = 1) stochiometry of Al-XO complex are shown in Figs. 5 and 6. As can be seen from Fig. 5, the Al-XO (ML) complexes should be protonated in the pH range 2 – 8. The major species are ML, MLH and MLH2 since the pH decreases from pH 8 to 2. This is quite expected due to competition of H+ with Al(III). Figure 6 shows the spectral behavior of the species forms in the solution; it illustrates the spectrum of ML, and MLH overlaps and shows a maximum at around 550 nm. However, MLH2 shows a blue shift (440 nm) due to influences of a lower pH on the chromospheres group of XO. In the case of CM/CL = 1/2, by increasing the pH, the ML2 species formed in solution. Each of the iminiodiactate groups and the nitrogen atom in one arm of XO with the corresponding groups of another XO complex with Al(III) and a stable chelate structure was formed. Determination of fluoride In order to assay fluoride based on a reaction with the Al-XO complex, a proper form of the Al-XO complex should be chosen. There are some restrictions in selecting the pH. At a lower pH, there is greater competence between H+ and Al(III) to react with

1252

Fig. 5 Concentration profile vs. pH for 5 × 10–5 mol L–1 of the Al-XO (CM/CL = 1) complex in aqueous solution (0.1 mol L–1 ionic strength and 25° C).

Fig. 6 Pure spectra of 5 × 10–5 mol L–1 of Al-XO (CM/CL = 1) complex in aqueous solution (0.1 mol L–1 ionic strength and 25° C).

the XO, and it also has a tendency to react with fluoride.27,28 For high pH values, there are serious interferences due to OH– and HCO3– (in drinking water), which could react with Al(III) and decrease the accuracy of the measurement. Thus, according to Figs. 5 and 6, at intermediate pH (4 – 8) values, the most predominate species in solution is MLH, which has a higher molar absorptivity at 550 nm and pH 5.2. This pH is close to the pH of fresh-water samples (drinking water). Thus, the best condition was chosen as pH 5.2 and a wavelength is 555 nm, where MLH is the major species in solution, and has a proper condition for reacting with fluoride. Addition of fluoride to an aliquot sample solution of MLH causes the formation of a more stable and known complex with Al-XO. This reaction leads to the formation of a colorless complex, [AlF6]–3 (log K = 21.4),29 and releases of XO in solution. Therefore, the stepwise addition of fluoride to this reagent solution causes decrease of the absorbance at 555 nm (Fig. 7). According to these optimum condition, a series of standard solutions containing fluoride (0.0, 0.1, 0.05, 0.1, 0.3, 0.5, 0.7, 0.9, 1.1, 1.3, 1.5 and 2.0 µg mL–1) separately was added to a 5 × 10–5 mol L–1 solution of Al-XO (1:1) at pH 5.2 and ionic strength 0.1 mol L–1 respectively. The resulting solutions remained sufficient before measuring of their spectra to achieve equilibrium. Their spectrum from 380 – 680 nm was recorded, and is shown in Fig. 7. It can be seen that the absorbance of the Al-XO complex (CM/CL = 1) (A0) decreased as the amount of


Fig. 7 Visible absorption spectra of the addition of the standard fluoride solution (0.0, 0.05, 0.10, 0.30, 0.50, 0.70, 0.90, 1.10, 1.30, 1.50, 2.0 mg L–1 of F–) to 5 × 10–5 mol L–1 of Al-XO (CM/CL = 1) complex in aqueous solution (0.1 mol L–1 ionic strength and 25° C).

Fig. 8 Calibration plot for addition (0.0, 0.1, 0.05, 0.1, 0.3, 0.5, 0.7, 0.9, 1.1, 1.3, 1.5 and 2.0 µg mL–1) of the standard fluoride solution to 5 × 10–5 mol L–1 of Al-XO (CM/CL = 1) at pH 5.2 and ionic strength 0.1 mol L–1, respectively.

fluoride ion increased. Figure 8 shows a plot of the absorbance decreasing (A0 – Ai) versus the added fluoride, which has a good correlation with the concentration of fluoride. The empirical data of (A0 – Ai) versus the concentration of fluoride was well fitted to a least-squares equation (Y = 0.9778 – 0.3849X) with a coefficient of determination R2 = 0.9998. The best dynamic range was from 0.08 to 1.40 mg L–1. The limit of detection (LOD) for the proposed method was 0.07 µg mL–1 fluoride, given by the equation YLOD = Ybl + 3σbl.30 The reproducibility of the method was checked for 10 measurements of a 5 × 10–5 mol L–1 solution of Al-XO (CM/CL = 1) complex and 0.5 µg mL–1 fluoride at the optimum condition. The relative standard deviation was 1.6% (n = 10). Interference effects and applicability The effect of diverse cationic and anionic species with commonly interference ions with fluoride in water samples (Cl–, Br–, SO4–, PO43–, HCO3–, Ca2+, Mg2+, Al(III), Fe(III)) was examined. The only serious interference was due to Fe(III), which was removed by the addition of a portion of L-ascorbic acid solution (10%). In order to test the applicability of the proposed method for drinking water, five samples (Table 2) under the optimum conditions (pH 5.2, ionic strength 0.1 mol L–1,

ANALYTICAL SCIENCES OCTOBER 2009, VOL. 25 Table 2

1253

Determination of fluoride in different water samples using proposed procedures F– found/µg mL–1

F– added/µg mL–1

F– (spiked + sample water) found/µg mL–1

Recovery, %

Refined water by reverse osmosis

0.120

Tap water 1: University region

0.500

Tap water 2: Urban region

0.330

Bottle water 1: Damavand mineral watera

0.190

Bottle water 2: Nava mineral watera

0.510

0.100 0.200 0.100 0.200 0.100 0.200 0.100 0.200 0.100 0.200

0.230 0.320 0.600 0.690 0.450 0.540 0.300 0.380 0.600 0.720

104 100 100 98 102 101 103 97 98 101

Source

Conditions: pH, 5.2; ionic strength, 0.1 mol L–1; λmax, 557 nm. a. Iranian mineral water.

25° C, wavelength 555 nm) were measured by using calibration plot (four sample measurements). A good agreement between spiked water and real samples was seen (Table 2).

13. 14.

Conclusion

15.

A new spectrophotometric reagent for determination of trace amounts of fluoride was established. Al-XO complex at pH 5.2 and ionic strength 0.1 mol L–l at 25° C was a suitable probe for this assay. The only serious interference of Fe(III) was successfully eliminated. The linear range and a good LOD was found. Some real samples were analyzed and a spiked sample emphasized the validity and accuracy of the method.

16.

References

17. 18. 19. 20. 21.

1. J. E. Harwood and D. J. Huyser, Water Res., 1968, 2, 637. 2. Position of the American Dieteic Association: The Impact of Fluoride on Health (ADA Reports), J. Am. Diet. Assoc., 2000, 100, 1208. 3. M. S. Frant and J. W. Ross, Science, 1966, 154, 1553. 4. “Standard Test Method for Fluoride Ion in Water”, ASTM Book of Standard, 2003, Vol. 11.01, ASTM, International, West Conshohocken, PA. 5. H. Hara and C. C. Huang, Anal. Chim. Acta, 1997, 53, 141. 6. L. N. Moskvin, A. N. Kataruzov, and T. G. J. Nikitana, Anal. Chem., 1998, 53, 173. 7. R. Bodor, V. Madajova, D. Kaniansky, M. Masar, M. Johnck, and B. J. Stanislawski, J. Chromatogr., A, 2001, 916, 155. 8. M. Garrido, A. G. Lista, M. Palomeque, and B. S. Fernandez Band, Talanta, 2002, 58, 849. 9. T. L. Har and T. S. West, Anal. Chem., 1971, 43, 136. 10. J. Nishimoto, T. Yamada, and M. Tabata, Anal. Chim. Acta, 2001, 428, 201. 11. “Standard Methods for the Examination of Water and Wastewater”, 20th ed., 1998, Method 4500 FD, American Public Health Association, Washington, 4 – 62. 12. “EPA Methods for Chemical Analysis of Water and Waste”,

22. 23. 24. 25. 26. 27. 28. 29. 30.

1978, Method 340.1, US Environment Protection Agency, Washington. E. Bellack and P. Schouboe, Anal. Chem., 1958, 30, 2032. C. Q. Zhu, J. L. Chen, H. Zheng, Y. Q. Wu, and J. G. Xu, Anal. Chim. Acta, 2005, 539, 311. M. E. Khalifa and M. A. H. Hafez, Talanta, 1998, 47, 547. M. R. Dyson, S. Kaderli, A. Lawrance, M. Maeder, and A. D. Zuberbuhler, Anal. Chim. Acta, 1997, 353, 381. M. Kubista, R. Sjoback, and B. Albinsson, Anal. Chem., 1993, 65, 994. K. Y. Tam, M. Hadley, and W. Palterson, Talanta, 1999, 49, 539. J. Ghasemi, Sh. Ahmadi, M. Kubista, and A. Elbergali, J. Chem. Eng. Data, 2003, 48, 1178. J. Ghasemi, A. Niazi, and M. Maeder, J. Braz. Chem. Soc., 2007, 18, 267. J. M. Bosque-Sendra, E. Almansa-Lopez, A. M GarciaCampana, and L. Cuadros-Rodriguez, Anal. Sci., 2003, 19, 1431. G. Azimi, A. Khoobi, Kh. Zamani, and J. Zolgharnein, J. Chem. Eng. Data, 2008, 53, 1862. A. E. Maretell and R. J. Motekaities, “Determination and Use of Stability Constants”, 1988, Verlag Chemie, Weinheim. M. B. Gholivand, F. Bamdad, and J. Ghasemi, Talanta, 1998, 46, 875. M. Maeder and A. D. Zuberbuhler, Anal. Chem., 1990, 62, 2220. H, Gampp, D, Haspra, M. Maeder, and A. D. Zuberbuheler, Inorg. Chem., 1984, 23, 3724. J. J. Lagowski, “Modern Inorganic Chemistry”, 1973, Marcel Dekker, New York, 535. R. B. Heslop and K. Jones, “Inorganic Chemistry”, 1976, Elsevier Scientific Publishing Company, 287. J. Laurie, “Handbook of Analytical Chemistry”, 1987, Mir, Moscow. D. L. Massart, B. G. M. Vandeginste, L. M. C. Buydens, S. D. E. Jong, P. J. Lewi, and J. Smeyers Verbeke, “Handbook of Chemometrics and Qualimetrics”, 1997, Part A, Elsevier, Amsterdam.