- Email: [email protected]
separation and determination using a polyurethane foam mini-column and flame atomic absorption spectrometry.
Talanta 60 (2003) 929 /936 www.elsevier.com/locate/talanta
Gallium trace on-line preconcentration/separation and determination using a polyurethane foam mini-column and flame atomic absorption spectrometry. Application in aluminum alloys, natural waters and urine Aristidis N. Anthemidis, George A. Zachariadis, John A. Stratis * Laboratory of Analytical Chemistry, Department of Chemistry, Aristotle University, Thessaloniki 54124, Greece Received 4 November 2002; received in revised form 5 February 2003; accepted 13 February 2003
Abstract A sensitive and selective flow injection time-based method for on-line preconcentration/separation and determination of gallium by flame atomic absorption spectrometry at trace levels was developed. The on-line formed gallium chloride complex is sorbed onto a polyether-type polyurethane foam mini-column, followed by on-line quantitative elution with isobutyl methyl ketone and direct introduction into the flame pneumatic nebulizer of the atomic absorption spectrometer. All chemical and flow variables of the system as well as the possible interferences were studied. The manner of strong HCl solutions propulsion was investigated and established using a combination of two displacement bottles. For 90 s preconcentration time, a sample frequency of 28 h 1, an enhancement factor of 40, a detection limit of 6 mg l 1 and a precision expressed as relative standard deviation (sr) of 3.3% (at 1.00 mg l 1) were achieved. The calibration curve is linear over the concentration range 0.02 /3.00 mg l 1. The accuracy of the developed method was sufficient and evaluated by the analysis of a silicon /aluminum alloy standard reference material. Finally, it was successfully applied to gallium determination in commercial aluminum alloys, natural waters and urine. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Gallium; Atomic absorption spectrometry; On-line preconcentration; Polyurethane foam
1. Introduction Gallium occurs at trace levels in the earth’s crust (18 mg kg 1), in silicate rocks, minerals and ores like bauxite, germanites and coal (10 /100 mg * Corresponding author. Tel.: /30-231-099-78-43; fax: /30231-099-77-19. E-mail address: [email protected] (J.A. Stratis).
kg 1). The treatment of bauxite ores, where it tends to be hidden by aluminum, is the main process for gallium recovery. Gallium and its compounds are used in the production of lowmelting alloys, intermetallic compounds used in the electronic industry for manufacturing of semiconductors, lasers, special optical glasses and thermometers. Due to the above applications, the world production is increasing and the levels of
0039-9140/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0039-9140(03)00156-5
930
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
gallium in the environment are beginning to rise, mainly around industrial areas [1]. It is a low-order toxic element and its citrate and nitrate salts are used in medicine as tumor-scanning and antitumoral agents, respectively. The role of gallium in pharmacokinetics and its toxicity have been investigated thoroughly [2,3]. Thus, there is a need for specific and precise determination of gallium traces in environmental and biological samples, as well as in aluminum alloys and gallium-containing minerals. Inductively coupled plasma mass spectrometry [4,5] and electrothermal atomic absorption spectrometry [6] are very capable for ultratrace analysis; however, they are expensive, time consuming and not readily automated techniques, requiring high operator’s skill. In addition, the application of specific matrix modifiers is necessary in electrothermal atomization, in order to avoid losses by vaporization of molecular species of gallium during the heating process [6]. On the other hand, flame atomic absorption spectrometry (FAAS) is available in almost all analytical laboratories and is easily combined and automated with on-line preconcentration techniques. Preconcentration and separation steps are necessary for trace gallium determination, in order to improve the precision and accuracy and also to overcome matrix and reagent interferences. Since Bowen [7] in 1970 investigated the excellent absorption properties of polyurethane foams (PUF), these materials have been used as solid sorbents for preconcentration and separation of a wide variety of inorganic and organic compounds in various media. Braun et al. [8 /10] and Palagyi and Braun [11] have published some extended review articles about the use of PUF in solid-phase extraction (SPE) procedures applied to various analytical methods. Gesser et al. [12] have investigated the batch extraction and recovery of gallium from hydrochloric acid solutions using polyether-type PUF without interference from greater than 1000-fold amounts of aluminum. In addition, Carvalho [13] for the same chemical system observed very high distribution coefficient for gallium (D ]/104 in 6 /9 mol l 1 HCl), while Al, In, Zn, Ni, Ti(IV) were not practically extracted (D 5/4.5), and Fe(II) and Cu were only
slightly extracted (D 5/30) from the same medium. The high retention capacity, polyether-type PUF due to its large available surface area, the resilient property, the resistance in many acids (hydrochloric acid up to 6 mol l 1), bases and organic solvents [8], and also the low cost have established PUF as a perfect sorbent material for column preconcentration/separation techniques. After de Jesus et al. [14] have employed firstly the polyether-type PUF for on-line zinc preconcentration and determination, the use of this material has been established and a number of papers have been published for on-line determination of Pb [15 /17], Ni [18], Zn [19], Cd [20], Co [21], Cu [17], Cr(VI) [17] and Al [22]. The purpose of this work is to study the feasibility of a filled mini-column with polyethertype PUF, for on-line preconcentration/separation of trace amounts of gallium, via its complexation with strong HCl and subsequent FAAS determination after elution by isobutyl methyl ketone (IBMK). One of the problems usually encountered in flow injection (FI) systems using tygon pump tubes is the manner of the strong acidic or basic solutions (more than 2 mol l 1) propulsion [23]. For this reason, a combination of two displacement bottles was introduced. The proposed method was successfully applied to the analysis of aluminum alloys, natural waters and urine samples, and its accuracy was tested by the analysis of certified reference materials.
2. Experimental 2.1. Instrumentation A Perkin/Elmer model 5100 PC flame atomic absorption spectrometer equipped with a flow spoiler in the spray chamber was used. The gallium hollow cathode lamp was operated at 20 mA. The analytical wavelength was 287.4 nm. A time constant of 0.2 s was used for peak height evaluation. The flame composition was adjusted to be slightly leaner than the manufacturer’s recommendation because IBMK serves as an additional fuel. The air flow rate was set at 10.0 l min 1 and acetylene flow rate at 1.0 l min 1. FI
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
931
Fig. 1. FI manifold and its operation sequence for on-line preconcentration/separation and determination of gallium; FAAS, flame atomic absorption spectrometer; P1 and P2, peristaltic pumps; V, injection valve; DB1 and DB2, displacement bottle manifold for 8 mol l 1 HCl; DB3, displacement bottle for IBMK; C, mini-column filled with cylindrical plug of PUF; and W, waste.
manifold is shown schematically in Fig. 1. It was adapted directly to the burner-nebulizer system of the atomic absorption spectrometer and was consisted from two peristaltic pumps, P1 and P2 (Watson Marlow model 205U/BA), equipped with tygon tubes for propulsion of the aqueous streams, a two-position/six-port injection valve (Labpro, Reodyne, USA) and a mini-column, filled with polyether-type PUF for the on-line preconcentration/separation of the gallium. Two displacement bottles, DB1 and DB2 (Tecator, Hoganas, Sweden), were serially combined and used to propel the strong HCl solutions (]/2 mol l 1), due to the incompatibility of tygon tubes with the high acid concentration. For IBMK propulsion, another displacement bottle, DB3, was used. The reaction coil (RC) and all other tubes were made of PTFE (0.5 mm i.d.). The capillary between the injection valve and the flame nebulizer was short enough (ca. 10 cm/0.35 mm i.d.) to minimize the analyte dispersion. The cylindrical mini-column was constructed of PTFE (45.0 mm length/4.6 mm i.d.) in our laboratory. Commercial open-cell resilient poly-
ether-type PUF (Euroform, GmbH) with 12 /14 cells cm 1 and a density of 22/24 mg cm 3 was used as sorbent material in the mini-column, without any pretreatment or transformation of its initial structure. The open cells (distorted polyhedrons) occupy ca. 95% of the volume, thus the backpressure is kept sufficiently low, when it is used in flow systems. PUF was cut into cylindrical plug of 75 mm length/7 mm i.d., which was placed in the mini-column. An image of the used polyether-type PUF taken by scanning electron microscope has been given previously [17]. In order to remove any organic and inorganic contaminants from PUF, it was washed thoroughly by passing 1 M HNO3 solution through the column for 15 min, followed by de-ionized water until the washings were acid-free and finally flushed with IBMK. The performance of the column was stable for more than 500 cycles. 2.2. Reagents All chemicals were of analytical reagent grade and were provided by Merck (Darmstadt, Ger-
932
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
many) and doubly de-ionized water was used throughout for dilutions. Working standard solutions of gallium were prepared daily by stepwise dilution of stock standard solution (Merck Gallium ICP standard traceable to SRM from NIST Ga(NO3)3 in HNO3, 2/3%, 1000 mg l 1 Ga) by 0.01 mol l 1 HNO3. Hydrochloric acid solutions were prepared daily by appropriate dilution of 12 mol l 1 HCl in de-ionized water. IBMK was used, after saturation with de-ionized water, without any other purification. NIST SRM 87A (silicon/aluminum alloy), standard reference material, was used to verify the accuracy of the proposed method. 2.3. Procedure The operation sequence of the proposed method runs through preconcentration and elution steps as they are shown in Fig. 1. The preconcentration step (Fig. 1a), sample and HCl streams are merged together and propelled by pump P1 at the same flow rate for suitable preconcentration time. The gallium chloride complex is formed on passage of the mixture through RC and it is retained onto the mini-column C. In the elution step shown in Fig. 1b, the injection valve is switched and IBMK stream is propelled by pump P2 through the minicolumn, eluting the gallium chloride complex directly to the burner-nebulizer of FAAS. The eluent flows in reverse direction than that of the sample in order to minimize the analyte dispersion. All the signals were recorded as peak height and were proportional to the gallium concentration in samples. Five replicate measurements per sample were made in all instances. 2.4. Sample preparation The natural water samples were initially acidified for storage to pH:/1.5 /1.8 by HNO3 and filtered through 0.45 mm Millipore filters, and thus no other pH adjustment was needed. A 300 mg aliquot of aluminum alloy sample was accurately weighed in a PTFE vessel and was dissolved by 5 ml concentrated HCl (12 mol l 1). The resulting solution was diluted to 50 ml by 0.3 mol l 1 HCl solution.
Urine samples were digested in presence of concentrated nitric acid at 120 8C, in a steel pressurized autoclave, using sealed Teflon beakers in order to prevent any possible losses of the analyte. After cooling, the resulting solution was finally diluted to 50 ml by 0.3 mol l 1 HCl.
3. Results and discussion 3.1. The propulsion of strong HCl solution Preliminary experiments on the effect of strong HCl solution on the tygon peristaltic pump tubes showed that the flow rate of the above tubes decreases continuously during a period of 2 h and a white film is formed on the inner surface due to tube’s deterioration. In addition, when Acidflex peristaltic pump tubes were used, the relative standard deviation of the recorded flow rate was 7/8%. Thus, the above type of tubes cannot be used safely for ‘‘time-based’’ loading systems, where the forwarded volume depends significantly on the flow rate. Based on these assumptions it was attempted to apply two serial displacement bottles as they are shown in Fig. 1. The main problem usually encountered in this combination is the intermediary liquid. This liquid must be an organic solvent immiscible with water and also immiscible and unaffected from strong HCl aqueous solutions. Hexane (C6H14) is practically immiscible with water (solubility in water 0.001%, w/w) and, as it has proved from extended experiments, it was unaffected from HCl aqueous solutions up to 10 mol l 1 and vice versa. In addition, no increased backpressure or decrease on the flow rate with and without the use of the pair of the displacement bottles was observed; thus, this combination was adapted to the on-line preconcentration manifold for the following study. 3.2. Optimization of chemical and FI variables Chemical and FI variables of the manifold shown in Fig. 1 were optimized using a standard solution of 2.0 mg l 1 of Ga(III) for 90 s preconcentration time. Equal flow rates for sample
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
and HCl streams were used throughout the whole study. As already pointed out in batch SPE [12,13], the gallium extraction efficiency of polyether-type PUF is increasing by increasing the hydrochloric acid concentration, reaching a maximum value in the range 6/9 mol l 1 HCl. Gallium sorption, as H GaCl4 complex, is considered to be an etherlike solvent extraction process and PUF can be regarded as a ‘‘viscous liquid’’ solvent of moderate dielectric constant [12]. On the other hand, gallium is extracted as chloride complex from strong HCl solutions with various organic solvents like diethyl ether, di-isopropyl ether, methyl ethyl ketone and IBMK with more than 95% efficiency [24 /26]. Thus, the effect of HCl concentration on Ga(III) complex preconcentration was studied within the range from 2.0 to 10.0 mol l 1 HCl. Maximum absorption signals were recorded within the range 7.5 /9.0 mol l 1 HCl (Fig. 2), and thus 8 mol l 1 HCl was used throughout, which corresponds to 4 mol l 1 in the final merged stream. The decrease of the preconcentration efficiency at higher and lower pH values is probably due to non-complete complex formation or sorption. Water or weakly basic solutions are effective in removing gallium chloro-complex from PUF columns [12]. On the other hand, IBMK has excellent extraction properties for gallium chloro-complexes
933
and it is almost completely immiscible with water and less-polar solvent. In addition, the use of it as eluent in on-line PUF preconcentration systems may enhance the sensitivity of FI-FAAS increasing the flame temperature and decreasing the analyte dispersion. Various agents like chloroform, IBMK, NH4OH/NH4CI buffer and H2O were tested for the elution of the gallium complex. IBMK produced higher and the sharper signals with stable baseline against the other eluents, and consequently it was chosen as eluent. The effect of PUF structure on the preconcentration efficiency was studied comparatively for PUF in normal form cut in cylindrical plug and PUF blended into small particles of the same mass. As it was proved, the normal form of PUF is much more effective because the blended form increases the backpressure which probably affects the two steps of the procedure. The influence of RC (Fig. 1) length was investigated in the range 20 /100 cm. The length of RC determines the time given for the complex formation reaction to proceed, prior to the preconcentration. The efficiency of the preconcentration remained constant within the above range, and thus, 20 cm RC was fixed for further study. The loading flow rate (sample and HCl streams) affects significantly the analyte sorption and was investigated in the range from 1.6 to 13.8 ml
Fig. 2. Effect of HCl solution concentration on the absorbance of 2.0 mg l 1 Ga(III) for 90 s preconcentration time. All other parameters are as in Fig. 1.
934
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
min 1. The acidity of the merged stream was maintained stable because the same flow rates for sample and HCl streams were used in any case. The absorption increased linearly by increasing the loading flow rate up to 10.0 ml min 1 and then gradually leveled-off. This shows that the complexation is complete and the contact time is sufficient. Therefore, 10.0 ml min 1 was selected as a compromise between high sensitivity and low sample consumption. The eluent flow rate effect was studied in the range 1.3 /3.8 ml min 1. The absorbance was increased with the increase of the flow rate up to 1.7 ml min 1 and was leveled-off between 1.7 and 2.1 ml min 1. Above this range, the absorbance was slightly decreased due to the higher analyte dispersion. Finally, an elution flow rate of 2.1 ml min 1 was selected and, at this flow rate, 30 s was found to be sufficient time for quantitative elution of the retained gallium complex. The preconcentration time was varied from 30 to 180 s (Fig. 3). The absorbance increased linearly up to 90 s, indicating that practically no leaching of the retained complex occurs. At prolonged preconcentration times, the increasing rate of the analytical signal becomes lower. Finally, a 90 s preconcentration time was used in this study in order to achieve a reasonably high throughput and medium sample consumption; however, longer
periods may be used when higher enhancement factors are required, with some sacrifice in sample frequency. 3.3. Interference studies An inherent advantage of FAAS is the high selectivity of element determination. In spite of several potential interferents encountered in aluminum alloys, natural waters and biological samples were tested in order to evaluate the selectivity of the proposed on-line preconcentration method under the optimum conditions described above. The recoveries obtained from 1 mg l 1 Ga in presence of various interfering elements are summarized in Table 1. Only Fe(III) was found to interfere when present at high concentrations. 3.4. Performance of the on-line preconcentration system The characteristic performance data for the proposed method under the optimum conditions are presented in Table 2 for a preconcentration time of 90 s. The sampling frequency was 28 h 1 and the enhancement factor was 40 (as calculated by the ratio of the slopes of the calibration curves obtained with and without preconcentration using FAAS). The detection limit was calculated by the
Fig. 3. Effect of the preconcentration time on the absorbance of 2.0 mg l 1 Ga(III). All other parameters are as in Fig. 1.
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
Table 1 Effect of interferents on the recovery of 1 mg l 1 Ga(III) by the proposed on-line column preconcentration FAAS method Interferent
Concentration (mg l 1)
Recovery (%)
Al(III)
100 500 10 100 10 100 10 100 10 50 10 100 10 100 100 200 100 200
98 95 99 97 95 78 97 99 98 94 96 94 97 95 102 100 97 99
Fe(II) Fe(III) As(III) Ti(IV) Zn(II) Mn(II) Ca(II) Mg(II)
used for manufacturing containers for food transportation and beverage cans, and also to the analysis of river water and human urine. The results are given in Table 3. The obtained recovery from the alloys varied from 94 to 97% and from the liquid samples from 93 to 102%.
4. Conclusions
Table 2 Analytical performance data of the on-line preconcentration/ separation FAAS method for gallium determination Enhancement factor Preconcentration time (s) Sample volume (ml) Sampling frequency (h 1) Linear range (mg l 1) Regression equation (12 standards, n /5; [Ga] in mg l 1) Correlation coefficient Detection limit (3s ) (mg l 1) Precision (RSD, n /12) (%)
935
40 90 7.5 28 0.02 /3.00 A /0.0015/ 0.067[Ga] r /0.9992 6 3.3% (1.00 mg l 1)
The use of polyether-type PUF as sorbent material into mini-columns has been successfully demonstrated for the on-line preconcentration/ separation and determination of gallium by FAAS, for the first time. It was proved that the polyether-type PUF, without any transformation of its normal structure, improves significantly the performance of the on-line preconcentration method because it permits the implementation of high loading flow rates resulting to better enhancement factors. The obtained detection limits are very low, allowing trace gallium determination by FAAS. The analysis of alloys, water and urine
Table 3 Recovery of gallium in aluminum alloys, natural waters and urine by the proposed on-line FAAS method Samples Aluminum alloy (food container) (mg g 1)c
Added 0 100 200
Aluminum alloy (beverage can) (mg g 1)c 1
3s criterion and was found to be cL /6 mg l . The quantification limit was calculated by the 10s criterion and the lower limit of the linear calibration range (0.02 mg l 1) was defined. The accuracy of the proposed method was tested by analyzing a standard reference aluminum alloy (NIST 87A) with a certified content of Ga, 2009/ 10 mg g 1, and the recovery obtained was 97% (1949/8 mg g 1, n /5). Finally, the method was applied to the analysis of aluminum-based alloys
River water (mg l 1)
Urine (mg l 1)
0
Founda
Recoveryb
829/5 1769/6 2769/18
94 97
169/2
100 200
1129/8 2049/8
96 94
0 500 1000
B/cL 5109/20 9829/36
102 98
0 500 1000
B/cL 4659/21 9489/58
93 94
a Mean value9/standard deviation based on five replicates (n /5) determinations. b Recovery obtained from spiked samples. c After 1:5 dilution.
936
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
samples revealed the reliability, selectivity and robustness of the proposed method.
References [1] J.E. Fergusson, The Heavy Elements: Chemistry, Environmental Impact and Health Effects, Pergamon Press, Oxford, 1991. [2] W.R. Harris, L. Messori, Coord. Chem. Rev. 228 (2002) 237. [3] D.P. Kelsen, N. Alcock, S. Yeh, J. Brown, C. Young, Cancer 46 (1980) 2009. [4] B. Wen, X.-Q. Shan, R.-X. Liu, H.-X. Tang, Fresenius’ J. Anal. Chem. 363 (1999) 251. [5] K.J. Orians, E.A. Boyle, Anal. Chim. Acta 282 (1993) 73. [6] D. Ma, Y. Okamoto, T. Kumamaru, E. Iwamoto, Anal. Chim. Acta 390 (1999) 201. [7] H.T.M. Bowen, J. Chem. Soc. A (1970) 1082. [8] T. Braun, A.B. Farag, Anal. Chim. Acta 99 (1978) 1. [9] T. Braun, Fresenius’ J. Anal. Chem. 314 (1983) 652. [10] T. Braun, J.D. Navratil, A.B. Farag, Polyurethane Foam Sorbents in Separation Science, CRC Press, Boca Raton, FL, 1985. [11] S. Palagyi, T. Braun, Separation and preconcentration of trace elements and inorganic species on solid polyurethane foam sorbents, in: Z.B. Alfassi, C.M. Wai (Eds.), Preconcentration Techniques for Trace Elements, CRC Press, Boca Raton, FL, 1992. [12] H.D. Gesser, E. Bock, W.G. Baldwin, A. Chow, D.W. McBride, W. Lipinsky, Sep. Sci. 11 (1976) 317.
[13] M.S. Carvalho, D.Sc. Thesis, Pontificia Universidade Catolica, Rio de Janeiro, 1993. [14] D.S. de Jesus, R.J. Cassella, S.L.C. Ferreira, A.C.S. Costa, M.S. de Carvalho, R.E. Santelli, Anal. Chim. Acta 366 (1998) 263. [15] S.P. Quinaia, J.B.B. da Silva, M.C.E. Rollemberg, A.J. Curtius, Talanta 54 (2001) 687. [16] V.A. Lemos, S.L.C. Ferreira, Anal. Chim. Acta 441 (2001) 281. [17] A.N. Anthemidis, G.A. Zachariadis, J.A. Stratis, Talanta 58 (2002) 831. [18] S.L.C. Ferreira, D.S. de Jesus, R.J. Cassella, A.C.S. Costa, M.S. de Carvalho, R.E. Santelli, Anal. Chim. Acta 378 (1999) 287. [19] R.J. Cassella, D.T. Bitencourt, A.G. Branco, S.L.C. Ferreira, D.S. de Jesus, M.S. de Carvalho, R.E. Santelli, J. Anal. At. Spectrom. 14 (1999) 1749. [20] V.A. Lemos, R.E. Santelli, M.S. de Carvalho, S.L.C. Ferreira, Spectrochim. Acta B 55 (2000) 1497. [21] R.J. Cassella, V.A. Salim, L.S. Jesuino, R.E. Santelli, S.L.C. Ferreira, M.S. de Carvalho, Talanta 54 (2001) 61. [22] R.J. Cassella, R.E. Santelli, A.G. Branco, V.A. Lemos, S.L.C. Ferreira, M.S. Carvalho, Analyst 124 (1999) 805. [23] B. Karlberg, G.E. Pacey, Flow Injection Analysis, a Practical Guide, Elsevier, Amsterdam, 1989. [24] Z. Marczenko, M. Balcerzak, Separation Preconcentration and Spectrophotometry in Inorganic Analysis, Elsevier, Amsterdam, 2000. [25] J.A. Stratis, P. Spathis, D. Bikiaris, G.C. Papanikadros, Anal. Lett. 22 (1989) 1021. [26] K. Venkaji, P.P. Naidu, T.J.P. Rao, Talanta 41 (1994) 1281.
Gallium trace on-line preconcentration/separation and determination using a polyurethane foam mini-column and flame atomic absorption spectrometry. Application in aluminum alloys, natural waters and urine Aristidis N. Anthemidis, George A. Zachariadis, John A. Stratis * Laboratory of Analytical Chemistry, Department of Chemistry, Aristotle University, Thessaloniki 54124, Greece Received 4 November 2002; received in revised form 5 February 2003; accepted 13 February 2003
Abstract A sensitive and selective flow injection time-based method for on-line preconcentration/separation and determination of gallium by flame atomic absorption spectrometry at trace levels was developed. The on-line formed gallium chloride complex is sorbed onto a polyether-type polyurethane foam mini-column, followed by on-line quantitative elution with isobutyl methyl ketone and direct introduction into the flame pneumatic nebulizer of the atomic absorption spectrometer. All chemical and flow variables of the system as well as the possible interferences were studied. The manner of strong HCl solutions propulsion was investigated and established using a combination of two displacement bottles. For 90 s preconcentration time, a sample frequency of 28 h 1, an enhancement factor of 40, a detection limit of 6 mg l 1 and a precision expressed as relative standard deviation (sr) of 3.3% (at 1.00 mg l 1) were achieved. The calibration curve is linear over the concentration range 0.02 /3.00 mg l 1. The accuracy of the developed method was sufficient and evaluated by the analysis of a silicon /aluminum alloy standard reference material. Finally, it was successfully applied to gallium determination in commercial aluminum alloys, natural waters and urine. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Gallium; Atomic absorption spectrometry; On-line preconcentration; Polyurethane foam
1. Introduction Gallium occurs at trace levels in the earth’s crust (18 mg kg 1), in silicate rocks, minerals and ores like bauxite, germanites and coal (10 /100 mg * Corresponding author. Tel.: /30-231-099-78-43; fax: /30231-099-77-19. E-mail address: [email protected] (J.A. Stratis).
kg 1). The treatment of bauxite ores, where it tends to be hidden by aluminum, is the main process for gallium recovery. Gallium and its compounds are used in the production of lowmelting alloys, intermetallic compounds used in the electronic industry for manufacturing of semiconductors, lasers, special optical glasses and thermometers. Due to the above applications, the world production is increasing and the levels of
0039-9140/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0039-9140(03)00156-5
930
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
gallium in the environment are beginning to rise, mainly around industrial areas [1]. It is a low-order toxic element and its citrate and nitrate salts are used in medicine as tumor-scanning and antitumoral agents, respectively. The role of gallium in pharmacokinetics and its toxicity have been investigated thoroughly [2,3]. Thus, there is a need for specific and precise determination of gallium traces in environmental and biological samples, as well as in aluminum alloys and gallium-containing minerals. Inductively coupled plasma mass spectrometry [4,5] and electrothermal atomic absorption spectrometry [6] are very capable for ultratrace analysis; however, they are expensive, time consuming and not readily automated techniques, requiring high operator’s skill. In addition, the application of specific matrix modifiers is necessary in electrothermal atomization, in order to avoid losses by vaporization of molecular species of gallium during the heating process [6]. On the other hand, flame atomic absorption spectrometry (FAAS) is available in almost all analytical laboratories and is easily combined and automated with on-line preconcentration techniques. Preconcentration and separation steps are necessary for trace gallium determination, in order to improve the precision and accuracy and also to overcome matrix and reagent interferences. Since Bowen [7] in 1970 investigated the excellent absorption properties of polyurethane foams (PUF), these materials have been used as solid sorbents for preconcentration and separation of a wide variety of inorganic and organic compounds in various media. Braun et al. [8 /10] and Palagyi and Braun [11] have published some extended review articles about the use of PUF in solid-phase extraction (SPE) procedures applied to various analytical methods. Gesser et al. [12] have investigated the batch extraction and recovery of gallium from hydrochloric acid solutions using polyether-type PUF without interference from greater than 1000-fold amounts of aluminum. In addition, Carvalho [13] for the same chemical system observed very high distribution coefficient for gallium (D ]/104 in 6 /9 mol l 1 HCl), while Al, In, Zn, Ni, Ti(IV) were not practically extracted (D 5/4.5), and Fe(II) and Cu were only
slightly extracted (D 5/30) from the same medium. The high retention capacity, polyether-type PUF due to its large available surface area, the resilient property, the resistance in many acids (hydrochloric acid up to 6 mol l 1), bases and organic solvents [8], and also the low cost have established PUF as a perfect sorbent material for column preconcentration/separation techniques. After de Jesus et al. [14] have employed firstly the polyether-type PUF for on-line zinc preconcentration and determination, the use of this material has been established and a number of papers have been published for on-line determination of Pb [15 /17], Ni [18], Zn [19], Cd [20], Co [21], Cu [17], Cr(VI) [17] and Al [22]. The purpose of this work is to study the feasibility of a filled mini-column with polyethertype PUF, for on-line preconcentration/separation of trace amounts of gallium, via its complexation with strong HCl and subsequent FAAS determination after elution by isobutyl methyl ketone (IBMK). One of the problems usually encountered in flow injection (FI) systems using tygon pump tubes is the manner of the strong acidic or basic solutions (more than 2 mol l 1) propulsion [23]. For this reason, a combination of two displacement bottles was introduced. The proposed method was successfully applied to the analysis of aluminum alloys, natural waters and urine samples, and its accuracy was tested by the analysis of certified reference materials.
2. Experimental 2.1. Instrumentation A Perkin/Elmer model 5100 PC flame atomic absorption spectrometer equipped with a flow spoiler in the spray chamber was used. The gallium hollow cathode lamp was operated at 20 mA. The analytical wavelength was 287.4 nm. A time constant of 0.2 s was used for peak height evaluation. The flame composition was adjusted to be slightly leaner than the manufacturer’s recommendation because IBMK serves as an additional fuel. The air flow rate was set at 10.0 l min 1 and acetylene flow rate at 1.0 l min 1. FI
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
931
Fig. 1. FI manifold and its operation sequence for on-line preconcentration/separation and determination of gallium; FAAS, flame atomic absorption spectrometer; P1 and P2, peristaltic pumps; V, injection valve; DB1 and DB2, displacement bottle manifold for 8 mol l 1 HCl; DB3, displacement bottle for IBMK; C, mini-column filled with cylindrical plug of PUF; and W, waste.
manifold is shown schematically in Fig. 1. It was adapted directly to the burner-nebulizer system of the atomic absorption spectrometer and was consisted from two peristaltic pumps, P1 and P2 (Watson Marlow model 205U/BA), equipped with tygon tubes for propulsion of the aqueous streams, a two-position/six-port injection valve (Labpro, Reodyne, USA) and a mini-column, filled with polyether-type PUF for the on-line preconcentration/separation of the gallium. Two displacement bottles, DB1 and DB2 (Tecator, Hoganas, Sweden), were serially combined and used to propel the strong HCl solutions (]/2 mol l 1), due to the incompatibility of tygon tubes with the high acid concentration. For IBMK propulsion, another displacement bottle, DB3, was used. The reaction coil (RC) and all other tubes were made of PTFE (0.5 mm i.d.). The capillary between the injection valve and the flame nebulizer was short enough (ca. 10 cm/0.35 mm i.d.) to minimize the analyte dispersion. The cylindrical mini-column was constructed of PTFE (45.0 mm length/4.6 mm i.d.) in our laboratory. Commercial open-cell resilient poly-
ether-type PUF (Euroform, GmbH) with 12 /14 cells cm 1 and a density of 22/24 mg cm 3 was used as sorbent material in the mini-column, without any pretreatment or transformation of its initial structure. The open cells (distorted polyhedrons) occupy ca. 95% of the volume, thus the backpressure is kept sufficiently low, when it is used in flow systems. PUF was cut into cylindrical plug of 75 mm length/7 mm i.d., which was placed in the mini-column. An image of the used polyether-type PUF taken by scanning electron microscope has been given previously [17]. In order to remove any organic and inorganic contaminants from PUF, it was washed thoroughly by passing 1 M HNO3 solution through the column for 15 min, followed by de-ionized water until the washings were acid-free and finally flushed with IBMK. The performance of the column was stable for more than 500 cycles. 2.2. Reagents All chemicals were of analytical reagent grade and were provided by Merck (Darmstadt, Ger-
932
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
many) and doubly de-ionized water was used throughout for dilutions. Working standard solutions of gallium were prepared daily by stepwise dilution of stock standard solution (Merck Gallium ICP standard traceable to SRM from NIST Ga(NO3)3 in HNO3, 2/3%, 1000 mg l 1 Ga) by 0.01 mol l 1 HNO3. Hydrochloric acid solutions were prepared daily by appropriate dilution of 12 mol l 1 HCl in de-ionized water. IBMK was used, after saturation with de-ionized water, without any other purification. NIST SRM 87A (silicon/aluminum alloy), standard reference material, was used to verify the accuracy of the proposed method. 2.3. Procedure The operation sequence of the proposed method runs through preconcentration and elution steps as they are shown in Fig. 1. The preconcentration step (Fig. 1a), sample and HCl streams are merged together and propelled by pump P1 at the same flow rate for suitable preconcentration time. The gallium chloride complex is formed on passage of the mixture through RC and it is retained onto the mini-column C. In the elution step shown in Fig. 1b, the injection valve is switched and IBMK stream is propelled by pump P2 through the minicolumn, eluting the gallium chloride complex directly to the burner-nebulizer of FAAS. The eluent flows in reverse direction than that of the sample in order to minimize the analyte dispersion. All the signals were recorded as peak height and were proportional to the gallium concentration in samples. Five replicate measurements per sample were made in all instances. 2.4. Sample preparation The natural water samples were initially acidified for storage to pH:/1.5 /1.8 by HNO3 and filtered through 0.45 mm Millipore filters, and thus no other pH adjustment was needed. A 300 mg aliquot of aluminum alloy sample was accurately weighed in a PTFE vessel and was dissolved by 5 ml concentrated HCl (12 mol l 1). The resulting solution was diluted to 50 ml by 0.3 mol l 1 HCl solution.
Urine samples were digested in presence of concentrated nitric acid at 120 8C, in a steel pressurized autoclave, using sealed Teflon beakers in order to prevent any possible losses of the analyte. After cooling, the resulting solution was finally diluted to 50 ml by 0.3 mol l 1 HCl.
3. Results and discussion 3.1. The propulsion of strong HCl solution Preliminary experiments on the effect of strong HCl solution on the tygon peristaltic pump tubes showed that the flow rate of the above tubes decreases continuously during a period of 2 h and a white film is formed on the inner surface due to tube’s deterioration. In addition, when Acidflex peristaltic pump tubes were used, the relative standard deviation of the recorded flow rate was 7/8%. Thus, the above type of tubes cannot be used safely for ‘‘time-based’’ loading systems, where the forwarded volume depends significantly on the flow rate. Based on these assumptions it was attempted to apply two serial displacement bottles as they are shown in Fig. 1. The main problem usually encountered in this combination is the intermediary liquid. This liquid must be an organic solvent immiscible with water and also immiscible and unaffected from strong HCl aqueous solutions. Hexane (C6H14) is practically immiscible with water (solubility in water 0.001%, w/w) and, as it has proved from extended experiments, it was unaffected from HCl aqueous solutions up to 10 mol l 1 and vice versa. In addition, no increased backpressure or decrease on the flow rate with and without the use of the pair of the displacement bottles was observed; thus, this combination was adapted to the on-line preconcentration manifold for the following study. 3.2. Optimization of chemical and FI variables Chemical and FI variables of the manifold shown in Fig. 1 were optimized using a standard solution of 2.0 mg l 1 of Ga(III) for 90 s preconcentration time. Equal flow rates for sample
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
and HCl streams were used throughout the whole study. As already pointed out in batch SPE [12,13], the gallium extraction efficiency of polyether-type PUF is increasing by increasing the hydrochloric acid concentration, reaching a maximum value in the range 6/9 mol l 1 HCl. Gallium sorption, as H GaCl4 complex, is considered to be an etherlike solvent extraction process and PUF can be regarded as a ‘‘viscous liquid’’ solvent of moderate dielectric constant [12]. On the other hand, gallium is extracted as chloride complex from strong HCl solutions with various organic solvents like diethyl ether, di-isopropyl ether, methyl ethyl ketone and IBMK with more than 95% efficiency [24 /26]. Thus, the effect of HCl concentration on Ga(III) complex preconcentration was studied within the range from 2.0 to 10.0 mol l 1 HCl. Maximum absorption signals were recorded within the range 7.5 /9.0 mol l 1 HCl (Fig. 2), and thus 8 mol l 1 HCl was used throughout, which corresponds to 4 mol l 1 in the final merged stream. The decrease of the preconcentration efficiency at higher and lower pH values is probably due to non-complete complex formation or sorption. Water or weakly basic solutions are effective in removing gallium chloro-complex from PUF columns [12]. On the other hand, IBMK has excellent extraction properties for gallium chloro-complexes
933
and it is almost completely immiscible with water and less-polar solvent. In addition, the use of it as eluent in on-line PUF preconcentration systems may enhance the sensitivity of FI-FAAS increasing the flame temperature and decreasing the analyte dispersion. Various agents like chloroform, IBMK, NH4OH/NH4CI buffer and H2O were tested for the elution of the gallium complex. IBMK produced higher and the sharper signals with stable baseline against the other eluents, and consequently it was chosen as eluent. The effect of PUF structure on the preconcentration efficiency was studied comparatively for PUF in normal form cut in cylindrical plug and PUF blended into small particles of the same mass. As it was proved, the normal form of PUF is much more effective because the blended form increases the backpressure which probably affects the two steps of the procedure. The influence of RC (Fig. 1) length was investigated in the range 20 /100 cm. The length of RC determines the time given for the complex formation reaction to proceed, prior to the preconcentration. The efficiency of the preconcentration remained constant within the above range, and thus, 20 cm RC was fixed for further study. The loading flow rate (sample and HCl streams) affects significantly the analyte sorption and was investigated in the range from 1.6 to 13.8 ml
Fig. 2. Effect of HCl solution concentration on the absorbance of 2.0 mg l 1 Ga(III) for 90 s preconcentration time. All other parameters are as in Fig. 1.
934
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
min 1. The acidity of the merged stream was maintained stable because the same flow rates for sample and HCl streams were used in any case. The absorption increased linearly by increasing the loading flow rate up to 10.0 ml min 1 and then gradually leveled-off. This shows that the complexation is complete and the contact time is sufficient. Therefore, 10.0 ml min 1 was selected as a compromise between high sensitivity and low sample consumption. The eluent flow rate effect was studied in the range 1.3 /3.8 ml min 1. The absorbance was increased with the increase of the flow rate up to 1.7 ml min 1 and was leveled-off between 1.7 and 2.1 ml min 1. Above this range, the absorbance was slightly decreased due to the higher analyte dispersion. Finally, an elution flow rate of 2.1 ml min 1 was selected and, at this flow rate, 30 s was found to be sufficient time for quantitative elution of the retained gallium complex. The preconcentration time was varied from 30 to 180 s (Fig. 3). The absorbance increased linearly up to 90 s, indicating that practically no leaching of the retained complex occurs. At prolonged preconcentration times, the increasing rate of the analytical signal becomes lower. Finally, a 90 s preconcentration time was used in this study in order to achieve a reasonably high throughput and medium sample consumption; however, longer
periods may be used when higher enhancement factors are required, with some sacrifice in sample frequency. 3.3. Interference studies An inherent advantage of FAAS is the high selectivity of element determination. In spite of several potential interferents encountered in aluminum alloys, natural waters and biological samples were tested in order to evaluate the selectivity of the proposed on-line preconcentration method under the optimum conditions described above. The recoveries obtained from 1 mg l 1 Ga in presence of various interfering elements are summarized in Table 1. Only Fe(III) was found to interfere when present at high concentrations. 3.4. Performance of the on-line preconcentration system The characteristic performance data for the proposed method under the optimum conditions are presented in Table 2 for a preconcentration time of 90 s. The sampling frequency was 28 h 1 and the enhancement factor was 40 (as calculated by the ratio of the slopes of the calibration curves obtained with and without preconcentration using FAAS). The detection limit was calculated by the
Fig. 3. Effect of the preconcentration time on the absorbance of 2.0 mg l 1 Ga(III). All other parameters are as in Fig. 1.
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
Table 1 Effect of interferents on the recovery of 1 mg l 1 Ga(III) by the proposed on-line column preconcentration FAAS method Interferent
Concentration (mg l 1)
Recovery (%)
Al(III)
100 500 10 100 10 100 10 100 10 50 10 100 10 100 100 200 100 200
98 95 99 97 95 78 97 99 98 94 96 94 97 95 102 100 97 99
Fe(II) Fe(III) As(III) Ti(IV) Zn(II) Mn(II) Ca(II) Mg(II)
used for manufacturing containers for food transportation and beverage cans, and also to the analysis of river water and human urine. The results are given in Table 3. The obtained recovery from the alloys varied from 94 to 97% and from the liquid samples from 93 to 102%.
4. Conclusions
Table 2 Analytical performance data of the on-line preconcentration/ separation FAAS method for gallium determination Enhancement factor Preconcentration time (s) Sample volume (ml) Sampling frequency (h 1) Linear range (mg l 1) Regression equation (12 standards, n /5; [Ga] in mg l 1) Correlation coefficient Detection limit (3s ) (mg l 1) Precision (RSD, n /12) (%)
935
40 90 7.5 28 0.02 /3.00 A /0.0015/ 0.067[Ga] r /0.9992 6 3.3% (1.00 mg l 1)
The use of polyether-type PUF as sorbent material into mini-columns has been successfully demonstrated for the on-line preconcentration/ separation and determination of gallium by FAAS, for the first time. It was proved that the polyether-type PUF, without any transformation of its normal structure, improves significantly the performance of the on-line preconcentration method because it permits the implementation of high loading flow rates resulting to better enhancement factors. The obtained detection limits are very low, allowing trace gallium determination by FAAS. The analysis of alloys, water and urine
Table 3 Recovery of gallium in aluminum alloys, natural waters and urine by the proposed on-line FAAS method Samples Aluminum alloy (food container) (mg g 1)c
Added 0 100 200
Aluminum alloy (beverage can) (mg g 1)c 1
3s criterion and was found to be cL /6 mg l . The quantification limit was calculated by the 10s criterion and the lower limit of the linear calibration range (0.02 mg l 1) was defined. The accuracy of the proposed method was tested by analyzing a standard reference aluminum alloy (NIST 87A) with a certified content of Ga, 2009/ 10 mg g 1, and the recovery obtained was 97% (1949/8 mg g 1, n /5). Finally, the method was applied to the analysis of aluminum-based alloys
River water (mg l 1)
Urine (mg l 1)
0
Founda
Recoveryb
829/5 1769/6 2769/18
94 97
169/2
100 200
1129/8 2049/8
96 94
0 500 1000
B/cL 5109/20 9829/36
102 98
0 500 1000
B/cL 4659/21 9489/58
93 94
a Mean value9/standard deviation based on five replicates (n /5) determinations. b Recovery obtained from spiked samples. c After 1:5 dilution.
936
A.N. Anthemidis et al. / Talanta 60 (2003) 929 /936
samples revealed the reliability, selectivity and robustness of the proposed method.
References [1] J.E. Fergusson, The Heavy Elements: Chemistry, Environmental Impact and Health Effects, Pergamon Press, Oxford, 1991. [2] W.R. Harris, L. Messori, Coord. Chem. Rev. 228 (2002) 237. [3] D.P. Kelsen, N. Alcock, S. Yeh, J. Brown, C. Young, Cancer 46 (1980) 2009. [4] B. Wen, X.-Q. Shan, R.-X. Liu, H.-X. Tang, Fresenius’ J. Anal. Chem. 363 (1999) 251. [5] K.J. Orians, E.A. Boyle, Anal. Chim. Acta 282 (1993) 73. [6] D. Ma, Y. Okamoto, T. Kumamaru, E. Iwamoto, Anal. Chim. Acta 390 (1999) 201. [7] H.T.M. Bowen, J. Chem. Soc. A (1970) 1082. [8] T. Braun, A.B. Farag, Anal. Chim. Acta 99 (1978) 1. [9] T. Braun, Fresenius’ J. Anal. Chem. 314 (1983) 652. [10] T. Braun, J.D. Navratil, A.B. Farag, Polyurethane Foam Sorbents in Separation Science, CRC Press, Boca Raton, FL, 1985. [11] S. Palagyi, T. Braun, Separation and preconcentration of trace elements and inorganic species on solid polyurethane foam sorbents, in: Z.B. Alfassi, C.M. Wai (Eds.), Preconcentration Techniques for Trace Elements, CRC Press, Boca Raton, FL, 1992. [12] H.D. Gesser, E. Bock, W.G. Baldwin, A. Chow, D.W. McBride, W. Lipinsky, Sep. Sci. 11 (1976) 317.
[13] M.S. Carvalho, D.Sc. Thesis, Pontificia Universidade Catolica, Rio de Janeiro, 1993. [14] D.S. de Jesus, R.J. Cassella, S.L.C. Ferreira, A.C.S. Costa, M.S. de Carvalho, R.E. Santelli, Anal. Chim. Acta 366 (1998) 263. [15] S.P. Quinaia, J.B.B. da Silva, M.C.E. Rollemberg, A.J. Curtius, Talanta 54 (2001) 687. [16] V.A. Lemos, S.L.C. Ferreira, Anal. Chim. Acta 441 (2001) 281. [17] A.N. Anthemidis, G.A. Zachariadis, J.A. Stratis, Talanta 58 (2002) 831. [18] S.L.C. Ferreira, D.S. de Jesus, R.J. Cassella, A.C.S. Costa, M.S. de Carvalho, R.E. Santelli, Anal. Chim. Acta 378 (1999) 287. [19] R.J. Cassella, D.T. Bitencourt, A.G. Branco, S.L.C. Ferreira, D.S. de Jesus, M.S. de Carvalho, R.E. Santelli, J. Anal. At. Spectrom. 14 (1999) 1749. [20] V.A. Lemos, R.E. Santelli, M.S. de Carvalho, S.L.C. Ferreira, Spectrochim. Acta B 55 (2000) 1497. [21] R.J. Cassella, V.A. Salim, L.S. Jesuino, R.E. Santelli, S.L.C. Ferreira, M.S. de Carvalho, Talanta 54 (2001) 61. [22] R.J. Cassella, R.E. Santelli, A.G. Branco, V.A. Lemos, S.L.C. Ferreira, M.S. Carvalho, Analyst 124 (1999) 805. [23] B. Karlberg, G.E. Pacey, Flow Injection Analysis, a Practical Guide, Elsevier, Amsterdam, 1989. [24] Z. Marczenko, M. Balcerzak, Separation Preconcentration and Spectrophotometry in Inorganic Analysis, Elsevier, Amsterdam, 2000. [25] J.A. Stratis, P. Spathis, D. Bikiaris, G.C. Papanikadros, Anal. Lett. 22 (1989) 1021. [26] K. Venkaji, P.P. Naidu, T.J.P. Rao, Talanta 41 (1994) 1281.