A multicommutation-based flow system for multi-element analysis in pharmaceutical preparations

Transcription

1 Talanta 55 (2001) A multicommutation-based flow system for multi-element analysis in pharmaceutical preparations Fábio R.P. Rocha a, Patrícia B. Martelli b, Boaventura F. Reis b, * a Departamento de Química, Uni ersidade Federal de São Carlos, São Carlos, SP, Brazil b Centro de Energia Nuclear na Agricultura, Uni ersidade de São Paulo, PO Box 96, Piracicaba, SP, Brazil Received 27 February 2001; received in revised form 2 July 2001; accepted 11 July 2001 Abstract A flow system exploiting multicommutation and multidetection is proposed for sequential determinations in pharmaceutical preparations. The feasibilities were demonstrated by the determination of zinc, iron, copper, calcium and magnesium without changing the flow set-up. The gravitational flow of the solutions was exploited for addition of different chromogenic reagents and sample aliquots, thus avoiding the use of a propulsion unit. Transient signals at different wavelengths were measured simultaneously employing a fiber-optic multichannel spectrophotometer. Coefficients of variation of 1.0, 1.5, 1.4, 2.5 and 2.0% were obtained for iron, zinc, copper, calcium and magnesium, respectively. The mean sampling rate for the five species was 60 determinations per hour. In comparison to continuous reagent addition systems, the consumption was up to 160-fold lower. Results for pharmaceutical preparations agreed with those obtained by Flame atomic absorption spectrophotometry (FAAS) at the 95% confidence level Elsevier Science B.V. All rights reserved. Keywords: Flow-injection spectrophotometry; Polyvalent system; Multicommutation; Multidetermination; Pharmaceutical preparations; Iron; Zinc; Copper; Calcium; Magnesium 1. Introduction The research and routine analysis concerning flow injection analysis have demonstrated that this technique is very suitable for mechanization and automation of analytical procedures. Although most of the proposed systems has been designed to the selective determination of a single analyte, the development of manifolds for multielement determination has been the aim of several * Corresponding author. Fax: address: reis@cena.usp.br (B.F. Reis). works and the different approaches have been revised [1,2]. Regarding the spectrophotometric detection, most of the procedures (ca. 80%) is restricted to the determination of two species [3]. The main difficulties are the need of different reagents and of detection at different wavelengths. The former drawback can be overcome by the multicommutation approach [4], i.e. by designing the flow manifold with a discrete commutation device to manage each reagent [5]. Detection at different wavelengths is usually carried out with multichannel spectrophotometers (such as photodiode arrays) [6] or photometers based on light emitting diodes and photodetectors [7] /01/$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S (01)

2 862 F.R.P. Rocha et al. / Talanta 55 (2001) In this work, a versatile flow system based on multicommutation was associated to a multichannel fiber-optic spectrophotometer to implement multi-element determinations without changing the manifold. Sample and reagent aliquots were alternately introduced in the analytical path. The feasibilities were demonstrated by determination of iron, zinc, copper, calcium and magnesium in pharmaceutical preparations. 2. Experimental 2.1. Apparatus The flow network was assembled with eight three-way solenoid valves (NResearch, 161T031). A Pentium 166 MHz microcomputer equipped with an electronic interface (Advantech Corporation, PCL-711S) was employed to switch the commutation devices. Control signals were generated at TTL pattern and an electronic interface analogous to that previously described [4] was employed to generate the electric potential and current required to switch the valves (12 V, ca. 100 ma). Signals were measured with a multichannel fiber-optic spectrophotometer (Ocean Optics) equipped with an 80- l Hellma flow cell, by means of an interface package Dynamic Link Library (Ocean Optics). The software for controlling and data acquisition was developed in Visual Basic 3.0. Mixing coils and transmission lines were made of 0.8-mm i.d. polyethylene tubing Reagents and solutions All solutions were prepared with deionized water and analytical grade chemicals. Stock 1000 mg l 1 solutions were prepared by dissolving appropriated amounts of Zn, Fe, CaCO 3 or MgO in 10 ml of 6 mol l 1 HCl and Cu in 5 ml concentrated HNO 3 and making the volume up to 500 ml. Working reference solutions within 1.00 and 12.0 mg l 1 Zn 2+ ; 2.00 and 15.0 mg l 1 Fe 3+ ; and 2.00 mg l 1 Cu 2+ ; 20.0 and 150 mg l 1 Ca 2+ as well as 10.0 and 100 mg l 1 Mg 2+ were prepared in 0.10 mol l 1 HClO 4 in order to match the acid concentration with the mean expected acidity of the sample digests. Reagents for catalytic determination of copper were 0.10 mol l 1 Fe(III) plus 0.10 mol l 1 HCl (R 1 ) and 0.10 mol l 1 Na 2 S 2 O 3 (R 2 ). For iron determination, it was employed a 0.20 mol l 1 acetate buffer solution at ph 5.0 also containing 1.0% (m/v) ascorbic acid (R 3 ) and a 0.25%(m/v) 1,10-phenantroline solution (R 4 ). For zinc determination, it was prepared a buffer-masking solution (R 5 ) containing 3.0% (m/v) H 3 BO 3, 1.0% (m/v) sodium citrate, 0.40% (m/v) Na 3 PO 4 and 2.0% (m/v) thiourea with ph adjusted to 9.6 with NaOH. As chromogenic reagent, it was employed a 0.030% (m/v) zincon (2-{[ -hydroxy-5-sulfophenylazo)-benzylidene]-hydrazino}-benzoic acid, monosodium salt) solution (R 6 ). For the above mentioned species, water was used as carrier stream (C). The chromogenic reagent for calcium and magnesium was 0.050% (m/v) 3,3 -bis[n,n-bis(carboxymethyl)aminometyl]-o-cresolphatalein (CPC) dissolved in 0.05 mol l 1 HCl (R 7 ). Solutions 0.20% (m/v) 8-hydroxyquinoline plus 0.05 mol l 1 HCl (M 1 ) and 4.5 mmol l 1 ethyleneglycol-2- (2-aminoethyl)-tetracetic acid (EGTA) plus mol l 1 NaOH (M 2 ) were used as masking agents for determination of calcium and magnesium, respectively. A mol l 1 H 3 BO 3 buffer solution at ph 10.5 was employed as carrier stream (C). This solution contained also 0.75% (m/v) triethanolamine for iron masking Sample preparation Samples of pharmaceutical preparations were dissolved by a nitric perchloric acid procedure [3]. Ten tablets of each sample were powdered and a mass of 200 mg was placed in graduated digestion tubes with 2 ml of 65% HNO 3. After about 1 h, the tubes were heated at 160 C for 30 min. Then, the vessels were cooled and 1.0 ml 70% HClO 4 was added to each tube, which were further warmed at 210 C until formation of fumes of perchloric acid (about 15 min). After cooling, the volumes were made up to 100 ml with water.

3 F.R.P. Rocha et al. / Talanta 55 (2001) Flow diagram and procedure Fig. 1. Flow diagram of the system. C, carrier stream; S, sample; R i, reagents; V i, three-way solenoid valves; B 1, B 2, coiled reactors (50 and 100 cm, respectively); D, flow-cell of the spectrophotometric detection system. Solutions were placed at h=1.6 m above the waste vessel (W). Dashed lines represent the flow paths when the valves are switched on. The flow diagram of the proposed system is shown in Fig. 1. In the initial status, all valves were switched off and only the carrier stream flowed through the analytical path. Each solution can be introduced by switching the corresponding valve and the sampled volume is proportional to the time that the valve is maintained switched on. During sample and reagent selection, the carrier stream is interrupted (by switching the valve V 1 on) in order to assure the precision of the sampled volumes. The solution vessels were placed at 1.6 m above the waste, enabling the solutions to flow through the analytical path by action of the gravitational Table 1 Valve schedule for sequential determinations with the proposed flow system. Numbers 0 or 1 correspond to the valves switched OFF or ON Species V 1 V 2 V 3 V 4 V 5 V 6 V 7 V 8 t (s) Sampling cycles (nm) Fe(III) Zn(II) Cu(II) Ca(II) Mg(II)

4 864 F.R.P. Rocha et al. / Talanta 55 (2001) Fig. 2. Schematic representation of the sampling patterns for determination of (a) iron, zinc and copper and (b) calcium and magnesium. C, carrier stream; S, sample; R, R a,r b, reagents; M, masking solution. The number of sequences of sample and reagents is defined by the number of sampling cycles (Table 1). force at ca. 3.5 ml min 1. The reagents R 1 R 6 were employed for sequential determination of zinc, iron and copper, using water as carrier stream. For determining calcium and magnesium, the solutions R 1,R 2 and R 3 were changed by M 1,M 2 and R 7. The borate buffer solution was employed as carrier. The flow network was operated according to the valves switching course shown in Table 1, yielding the sampling patterns represented in Fig. 2. For determination of iron, zinc or copper, sample aliquots were sandwiched with aliquots of the reagents (Fig. 2a). Calcium and magnesium were determined with the same chromogenic reagent (CPC) by introducing the sample between aliquots of the masking solutions M 1 or M 2 (Fig. 2b). Thus, potential interfering species were masked before the overlap between the sample and the chromogenic reagent zones. The analytical measurements were ever based on the difference of the signals obtained at the absorption maximum (Table 1) and at 800 nm, aiming the compensation of drawbacks caused by Schlieren effect [8,9]. 3. Results and discussion 3.1. Flow system The proposed flow system (Fig. 1) exploited the action of the gravitational force for fluid propelling. This allowed operating the flow manifold without a peristaltic pump. Moreover, sample and reagent aliquots were precisely selected by eliminating the flow pulsation during the sampling steps. In experiments with a dye solution, coefficients of variation of 0.50 and 3.3% were achieved by employing gravitational flow and fluid propelling through a Fig. 3. Schlieren effect by the insertion of solutions, (1) 0.10 mol l 1 NaCl; (2) 10% (v/v) ethanol; and (3) 0.10 mol l 1 HClO 4 in water carrier stream. Measurements at 540 nm performed with (a) the optical-fiber multichannel spectrophotometer and (b) a conventional spectrophotometer. (c) Difference of the signals simultaneously measured at 540 and 800 nm with the multichannel spectrophotometer.

5 F.R.P. Rocha et al. / Talanta 55 (2001) Table 2 Reagent amounts consumed by the proposed method and by a flow system with reagent addition by confluence [13] Reagent Reagent amount ( g per determination) Proposed method Confluent system 1,10-Phenantroline Zincon Na 2 S 2 O CPC (calcium determination) CPC (magnesium determination) peristaltic pump, respectively. The variation in the flow rates during the analysis is minimized due to the low reagent volumes ( l) consumed per determination. Gravitational flow is difficult to be implemented in conventional confluence manifolds because the physical characteristics of the solutions define their volumetric partition, hindering the introduction of solutions with higher viscosity, for example. The independent control of the commutation devices was exploited for sampling one solution at a time. The reagents were introduced in the confluence points a or b (Fig. 1) in order to implement different residence times of the generated sample zones. This configuration also allowed the dispersion of sample aliquot in the reactor B 1 before reagent addition in the confluence point b. Moreover, the portion of sample zone that receives the reagent can be properly selected, avoiding its interaction with the most concentrated region of the sample zone. This strategy allowed sensitivity adjustment for zinc determination (valve schedule in Table 1). The software was designed to vary systematically the sampling times, the volumetric fractions and the number of sampling cycles [4] for self-optimization of the sensitivity for determination of each species [10]. The best results were obtained with the parameters defined in Table 1. The software makes feasible the development of other applications, after a suitable choice of the chromogenic reagents and masking agents Chemical ariables Usual chromogenic reagents were selected for determination of iron, copper and zinc, considering the required sensitivity and selectivity. For calcium and magnesium, the possibility of using the same reagent for sequential determination was also taken into account. Initial reagent concentrations were the same as in previous studies [11 13]. However, the reduction of the reagent concentration was generally possible, without affecting sensitivity. Table 3 Maximum tolerable concentrations of metallic ions in the determination of 5 mg l 1 of each analyte Species Maximum tolerable concentration a /mg l 1 Iron Zinc Copper Calcium Magnesium Iron(III) Zinc(II) Copper(II) * Calcium(II) 500* * 150 Magnesium(II) 250* * 100 Nickel(II) Manganese(II) 50* 5 50* * Cromium(III) * 100* 100* Cromium(VI) * 100* Molibdenum(VI) 50* 20 50* 50* 50* Vanadium(V) * 50* 50* *, Maximum evaluated concentration. a Concentration that causes changes in the signal lower than 5%.

6 866 F.R.P. Rocha et al. / Talanta 55 (2001) Fig. 4. Transient signals for reference solutions of (a) iron(iii), (b) zinc(ii), (c) copper(ii), (d) calcium and (e) magnesium. Numbers on signals represent analyte concentrations in mg l 1. The zincon reagent was evaluated for determination of both copper and zinc, performing measurements at ph 6.0 and 9.6. However, copper determination was hindered by high blank values because of the strong absorption caused by the reagent at the lower ph. Moreover, sensitivity was unsuitable, considering that copper concentration in the sample digests was expected to be lower than 2.0 mg l 1. Suitable results were obtained for zinc when a solution containing citrate, phosphate and thiourea was employed for both ph adjustment and masking of interferents, namely iron(iii), copper(ii) and nickel(ii). Copper was determined exploiting its catalytic effect on iron(iii) reduction by thiosulfate. The colored [Fe(S 2 O 3 ) 2 ] complex is formed fast and

7 Table 4 Determination of iron, zinc, copper, calcium and magnesium in pharmaceutical preparations (mean values and uncertainties, n=3) by the proposed method (FIA) and by FAAS Sample Fe mg/g Zn mg/g Cu mg/g Ca mg/g Mg mg/g FIA FAAS FIA FAAS FIA FAAS FIA FAAS FIA FAAS Centrum Centrum silver Gerimix Supradyn Vitabase Vitasay F.R.P. Rocha et al. / Talanta 55 (2001)

8 868 F.R.P. Rocha et al. / Talanta 55 (2001) it is slowly decomposed by excess of thiosulfate. However, copper catalyzes the kinetic limiting step. In the proposed system, samples and reagents were processed at a constant residence time, yielding signal peak heights inversely proportional to copper concentration. The gravitational flow and the alternate introduction of sample and reagent aliquots allowed avoiding oscillation in baseline verified when the solutions were introduced continuously through a peristaltic pump [13]. As the formation constants of the CPC complexes with calcium and magnesium are similar (K= and for calcium and magnesium, respectively [12]), a careful selection of the masking agents and their concentration was necessary. In this sense, two strategies were evaluated. The former was based on sequential measurements in the presence or absence of a masking agent for calcium (8-hydroxyquinoline) or magnesium (EGTA), and mathematical data processing. However, the accuracy was affected by the lack of absorbance additivity and relative errors up to 30% were observed in the presence of the highest concentration of the other analyte. Moreover, the increase in the concentration of the masking agents strongly affected the sensitivity. The accuracy was improved by the second strategy, which consisted of sequential measurements in the presence of different masking solutions. The addition of barium(ii) as a masking buffer in the magnesium determination, as previously suggested [13], resulted in high blank values, due to the formation of the Ba 2+ /CPC complex, with absorption maximum at 570 nm Schlieren effect compensation The spectrophotometer was furnished with 0.5- mm core optical fibers to transport the radiation from the light source to the flow cell and from this device to the detector array. The light beam was focused by lenses placed at the support of the flow cell in order to minimize light scattering. Spectrophotometric measurements with this configuration were strongly affected by refractive index differences in the sample zone (Schlieren effect). This can be observed in Fig. 3 by comparing the results obtained with the optical-fiber spectrophotometer (Fig. 3a) with those achieved with an equipment in which the radiation strikes the whole observation volume (Fig. 3b). In addition, the proposed procedure is more susceptible to drawbacks caused by Schlieren effect due to the insertion of sample and reagent aliquots with chemical and physical characteristics different from those of the carrier. As shown in Fig. 3c, the Schlieren effect was compensated by the dual wavelength strategy [8,9]. By using the multi-channel spectrophotometer, measurements were simultaneously performed at the absorption maxima and at one wavelength in which the absorption of radiation by the products is negligible (800 nm) Analytical features and effect of foreign ions Under the optimized conditions, coefficients of variation of the recorded peak heights were estimated (n=20) as 1.0, 1.5, 1.4, 2.5 and 2.0% for iron, zinc, copper, calcium and magnesium, respectively. The detection limits were estimated (99.7% confidence level) as 0.2, 0.2 and 0.05 mg l 1 for iron, zinc and copper, respectively. Detection limits as low as 10 g l 1 can be attained for calcium and magnesium [14] by minimizing the sample dispersion. For determining the five species, the average sampling rate was 60 measurements per hour. As can be seen in Table 2, the reagent consumption was reduced from 4 to 160-fold in comparison to a flow system for multidetermination with continuous reagent addition [13]. Transient signals obtained by processing reference solutions containing all species according to the valve schedule in Table 1 are shown in Fig. 4. The maximum tolerable concentrations of several ions in the determination of each species are shown in Table 3. According to the adopted criteria (signal variation lower than 5% in the presence of the foreign ion), all species are tolerable in concentrations higher than the expected in the sample digests. Anionic species (nitrate, sulfate, chloride and phosphate) did not interfere even in concentrations of 500 mg l 1.

9 F.R.P. Rocha et al. / Talanta 55 (2001) Applications The proposed procedure was applied for the determination of iron, zinc, copper, calcium and magnesium in pharmaceutical preparations (Table 4). The results agreed with those obtained by Flame atomic absorption spectrometry (FAAS) at the 95% confidence level. 4. Conclusions The multicommuted manifold showed to be versatile and robust. The intermittent addition of reagents and multidetection through the multichannel spectrophotometer allowed expanding the possibilities for sequential determinations in flow systems. Procedures for determination of other species can be readily developed by proper selection of chromogenic reagents. Sampling strategies adopted also allowed minimizing the reagent consumption and the generation of wastes. Acknowledgements The authors acknowledge the financial support of the Brazilian agencies FAPESP (process no. 97/ ), CAPES and CNPq/PRONEX. References [1] V. Kubán, Crit. Rev. Anal. Chem. 23 (1992) 15. [2] M.D. Luque de Castro, M. Valcárcel, Analyst 109 (1984) 413. [3] F.R.P. Rocha, B.F. Reis, J.J.R. Rohwedder, Fresenius J. Anal. Chem. 370 (2001) 22. [4] B.F. Reis, M.F. Giné, E.A.G. Zagatto, J.L.F.C. Lima, R.A. Lapa, Anal. Chim. Acta 293 (1994) 129. [5] F.R.P. Rocha, B.F. Reis, Anal. Chim. Acta 409 (2000) 227. [6] F. Lázaro, A. Ríos, M.D. Luque de Castro, M. Valcárcel, Analusis 14 (1986) 378. [7] P.K. Dasgupta, H.S. Bellamy, H. Liu, J.L. Lopez, E.L. Loree, K. Morris, K. Petersen, K.A. Mir, Talanta 40 (1993) 53. [8] E.A.G. Zagatto, M.A.Z. Arruda, A.O. Jacintho, I.L. Mattos, Anal. Chim. Acta 234 (1990) 153. [9] F.R.P. Rocha, P.B. Martelli, B.F. Reis, Anal. Chim. Acta, 438 (2001) 11. [10] M.F. Giné, R.L. Tuon, A.A. Cesta, A.P. Packer, B.F. Reis, Anal. Chim. Acta 366 (1998) 313. [11] Z. Marczenko, Separation and Spectrophotometric Determination of Elements, second ed., Ellis Harwood, Chichester, [12] K.L. Cheng, K. Ueno, T. Imamura, Handbook of Organic Analytical Reagents, CRC Press, Boca Raton, FL, [13] F.V. Silva, A.R. Nogueira, G.B. Souza, E.A.G. Zagatto, Anal. Chim. Acta 370 (1998) 39. [14] F.R.P. Rocha, P.B. Martelli, R.M. Frizzarin, B.F. Reis, Anal. Chim. Acta 366 (1998) 45.