Determination of phosphate in natural water employing a monosegmented flow system with simultaneous multiple injection

Talanta 62 (2004) 469–475

Determination of phosphate in natural water employing a monosegmented flow system with simultaneous multiple injection Maria Celeste Teixeira Diniz a , Orlando Fatibello Filho a , Emerson Vidal de Aquino b , Jarbas J.R. Rohwedder b,∗ a

Departamento de Qu´ımica, Universidade Federal de São Carlos, Rod. Washington Luiz, km 235, CP 676 CEP 13560-970 São Carlos, SP, Brazil b Instituto de Qu´ımica, Universidade Estadual de Campinas, CP 6154 CEP 13083-970 Campinas, SP, Brazil Received 17 March 2003; received in revised form 13 August 2003; accepted 18 August 2003

Abstract A monosegmented flow system was employed for the determination of low contents of phosphate in natural water. In this approach, sample and reagents are simultaneously injected to a PTFE reaction coil where the homogenization of the mixture occurs while the monosegment is pumped forwards the photometric detector. The proposed system was evaluated by determining phosphate ion, based on the reaction of association between molybdophosphate and malachite green. It was evaluated individually the best concentration of the reagent solution in relation to blank signal (absorbance of the blank) and the sensitivity of the method. A factorial design was proposed to explain the contribution of each component on the formation of the ion association complex, evaluating the individual contributions as well as the synergistic and antagonistic effects. With the established conditions, phosphorous is determined in the concentration range of 5.0–75 ␮g P PO4 3− l−1 (r = 0.9992), with a detection limit of 0.70 ␮g P PO4 3− l−1 and a relative standard deviation of 2% (20 ␮g P PO4 3− l−1 , n = 8). The proposed method has a sampling frequency of 72 h−1 . © 2003 Elsevier B.V. All rights reserved. Keywords: Phosphate; Monosegmented flow system; Simultaneous multiple injection

1. Introduction Since a long time ago is well known the importance of phosphorus in biological and aquatic systems [1]. Phosphorus is usually considered as a limiting nutrient for primary production, being essential for growth and maintenance of live organisms and, in aquatic atmospheres, its concentration can control the production of algae and any organism that depends on this element [2]. The excessive concentration of phosphorus in their several physico-chemical forms is a factor that causes eutrophication of both lotic and lentic waters. Furthermore, the classification of the trophic status of aquatic bodies from ultra-oligotrophic to hyper-eutrophic is still largely based on the total phosphorus concentrations rather than concentration of dissolved phosphorus. The ortophosphate concentration (phosphate ion), the form in which phosphorus can be directly assimilated by the organ-

∗

Corresponding author. Tel.: +55-1937883136; fax: +55-1937883023. E-mail address: [email protected] (J.J.R. Rohwedder).

0039-9140/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2003.08.016

isms, is usually very low. In tropical lakes, due to the high temperature, the metabolism of the organisms increases considerably, and ortophosphate is still more quickly assimilated and incorporate in its biomass. This is one of the main reasons because the ortophosphate ion concentration in these lakes is very low [3], usually below the detection limit of most of the analytical methods available. Consequently, the monitoring of ortophosphate ion becomes important by the use of robust and mainly sensitive methods. Phosphorus present in aquatic bodies is found in different forms and/or fractions. An operational classification for phosphorus depends on the filtration through a 0.45 ␮m membrane filter for the separation of phosphorus in the dissolved and particulate fractions. The fraction of total dissolved phosphorus is subdivided in dissolved inorganic phosphorus, which includes ortophosphate ion, condensed and colloidal phosphate, and dissolved organic phosphate. The particulate phosphorus is defined as the fraction that is retained by cellulose acetate filter. The spectrophotometric methods for determination of ortophosphate ion more thoroughly employed are based on

470

M.C.T. Diniz et al. / Talanta 62 (2004) 469–475

the ortophosphate reaction with molybdate acid, loading to the formation of the molybdophosphoric acid [4–6], a yellow specie that presents a absorption maximum at 350 nm. This method can be employed for the ortophosphate determination in the range of mg l−1 . The reduced form of the molybdophosphoric acid, forms a intensely colored blue species of molybdophosphate (phosphomolybdenum blue) [7–9]. The sensitivity of the method depends on a series of factors such as reductor, acidity and temperature. An alternative to improve the sensitivity in the determination of phosphate is based on the formation of association complexes of the molybdophosphate ion with basic dye compounds. The complex formed can be monitored by either spectrophotometry [10,11] or spectrofluorimetry [12,13]. Several dyes have been already employed for the formation of ion association complexes, such as xylenol orange, metil timol blue, crystal violet, rhodamine and malachite green. Among these dyes, malachite green presents better sensibility and stability [14]. Altmamm et al. [15] have concluded that the association complex between molybdophosphate ion and malachite green is formed in acidic solution, according to the Eq. (1). This method is more sensitive than the phosphomolybdenum blue method and it presents a molar absorptivity of 1 × 105 cm−1 mol−1 l at 650 nm. The amount of complex formed is strongly affected by the acidity, molybdate and malachite green concentrations. PO4 (MoO3 )12 3− + 3MG+
(MG)3 PO4 (MoO3 )12

(1)

The protonated species (MG+ ) of malachite green (lightly yellowish) presents maximum absorption at 440 nm, while the neutral species absorbs at 650 nm, same area of absorption of the ion association complexes with the species no reduced of molybdophosphate. The monosegmented flow analysis (MSFA) proposed by Pasquini and Oliveira [16] in 1985 became an alternative to overcome some difficulties observed in flow injection analysis. In this system, sample and reagents are injected in the analytical path between two air bubbles, which form an interface of gaseous phase between the sample and the carrier solution. When compared to flow injection analysis (FIA) a sample in a MSFA system presents a lower dispersion and, as a consequence better sensitivity can be obtained. Because of the low dispersion the sample in the MSFA even for long residence times, this system is suitable to determine species that occur in low concentrations, as manual methods can be simulated if the monosegment is compared to a volumetric flask. By this way, the sensitivity provided by a MSFA system can be compared to the manual method, as the sample dilution can be maintained as low as possible. In other words, in manual methods a small volume of a concentrated reagent is often added to a large volume of sample, allowing high sensitivity. Another advantageous aspect with respect to MSFA is the minimization or even elimination of the Schlieren effect, that appears in FIA systems due to concentration gradients that

occur when a sample of high ionic concentration is injected in a carrier fluid of low ionic strength (or vice versa). In a MSFA system, concentration gradient does not exist in the monosegment, therefore the Schlieren effect is expected to be minimized or even eliminated in these system, without the need of any experimental procedures (changing in the configuration of the system, as the detection in two wavelengths to compensate the signal due to that effect) or treatment of data. This paper describes a monosegmented flow system for the determination of low concentration of phosphorous as phosphate in natural waters, by employing the reaction of ion association between molybdophosphate and malachite green.

2. Experimental 2.1. Reagents and solutions The reagents used were of analytical grade and water was previously distilled and deionized (Milli-Q Plus-Pure Water System-Millipore). Phosphate stock solution 250 mg P PO4 3− l−1 . This solution was obtained by dissolving 0.1098 g of potassium dihydrogen phosphate (previously dry) in water for 100 ml of solution. The phosphate reference solutions from 5.0 to 75 ␮g P PO4 3− l−1 were prepared by appropriate dilutions of the stock solution. Malachite green stock solution (MG) 0.005 mol l−1 . For a volume of solution of 500 ml, 2.3175 g of malachite green (oxalate form) was dissolved. This solution was maintained in the darkness in a flask amber. 2.1.1. Reagent solution (RS) This solution was composed by the substances required for the formation of the ion association complexes between molybdophosphate and malachite green. It was composed of ammonium heptamolybdate, sulfuric acid and malachite green. For 100 ml of solution, 3.5310 g of ammonium heptamolybdate ((NH4 )6 (Mo7 O24 )4 H2 O) was previously dissolved in a small volume of water to which 11 ml of concentrated sulfuric acid were added followed of 8 ml of the stock solution of malachite green. The final solution was 0.20 mol l−1 in molybdate, 2.0 mol l−1 in sulfuric acid and 4 × 10−4 mol l−1 in malachite green. 2.2. Materials and methods A peristaltic pump (Ismatec MP-13 R) was employed for propulsion of the solutions. The monosegmented flow system was configured so that the reference solution (S) and the reagent solution (RS) were simultaneously injected in the PTFE reaction coil of 1.6 mm i.d. and 50 cm of length as can be seen in the Fig. 1. This configuration is similar to the described by Brito and Raimundo Jr. [17] for

M.C.T. Diniz et al. / Talanta 62 (2004) 469–475

471

Fig. 1. Schematic diagram of the monosegmented flow system with simultaneous multiple injection. B: peristaltic pump; I: injector; RC: reaction coil; S: sample or reference solution; C: carrier (water); RS: reagent solution; w: waste; D: detector; L1 and L2: air loops; L3 and L4: sample loops; L5 reagent loop.

simultaneous multiple injection, so that sample and reagent are simultaneously injected in the analysis system. The monosegment of 325 ␮l was constituted by the reference solution or sample solution (S) that fill the loops L3 and L4 (150 ␮l) each and by 25 ␮l of reagent solution (RS, loop L5) and limited by two air bubbles (L1 = L2 = 50 ␮l). Deionized water was used as carrier fluid (C) which conduct the monosegment though the reaction coil (RC), allowing the formation of the ion association complexes, that was monitored with an spectrophotometer 432 FEMTO set at 650 nm, using a flow cell with 200 ␮l of internal volume and 1 cm of optical length. The control of the system and acquisition of the data was accomplished by a micro computer through a parallel interface (Advantech 7115) and a program written in Visual Basic 3.0.

2.4. Effects of the reagent concentrations using factorial design

2.3. Effect of reagent concentration on the blank signal and the sensitivity

2.5. Effect of manifold parameters

The effect of the component concentrations in the reagent solution on the blank signal and the sensitivity was individually evaluated by changing the concentration of one of the components and maintaining constant the others according to the Table 1. In these studies, the analytical curves were obtained using phosphate reference solutions from 5.0 to 25 ␮g P PO4 3− l−1 , reaction coil of 50 cm and total flow rate of 3.4 ml min−1 . The blank signals and sensitivity were obtained from the analytical curves.

A two-level and three-factor factorial design was accomplished, establishing as variables the components of the reagent solution: (1) ammonium molybdate, (2) sulfuric acid and (3) malachite green. Eight reagent solutions were prepared, as shown in the Table 2. The reagent solution were prepared as previously described in the Section 2.1.1. Analytical curves from 5.0 to 25 ␮g P PO4 3− l−1 of phosphorous were obtained for each reagent mixture. The reference solutions were injected in a random order and in triplicate. Each reagent mixture was prepared in duplicate. Data regarding the sensitivity of the analytical curve and the blank signal were obtained for each mixture.

The interval of time of reaction for the formation of the ion association complex was evaluated by modifying the length Table 1 Study of the concentration in the reagent solution on blank signal and sensibility

[Mo]/mol l−1

[H2 SO4 ]/mol l−1 [MG]/10−4 mol l−1

Experiment 1

Experiment 2

Experiment 3

0.05–0.5 1.85 4.0

0.35 1.85–4.85 4.0

0.35 1.85 0.5–5.0

Mo: molybdate, MG: malachite green.

472

M.C.T. Diniz et al. / Talanta 62 (2004) 469–475

Table 2 Factors, levels and analytical signal for the 23 factorial design Mixture

1 2 3 4 5 6 7 8 a b c

Factors 1a

2b

3c

− + − + − + − +

− − + + − − + +

− − − − + + + +

Blank signal absorbance

Average

Sensibility (×10−3 AU ␮g−1 l−1 )

Average

0.013 0.014 0.022 0.024 0.048 0.044 0.077 0.075

0.015 0.021 0.022 0.026 0.051 0.040 0.080 0.078

1.92 1.33 2.66 2.90 3.33 3.62 3.31 5.10

1.97 1.57 2.81 2.92 3.43 3.86 3.28 4.87

0.016 0.028 0.022 0.027 0.054 0.036 0.083 0.080

2.02 1.80 2.96 2.94 3.53 4.10 3.24 4.64

1: ammonium molybdate (mol l−1 ); low level (−) 0.2; high level (+) 1.0. 2: sulfuric acid (mol l−1 ); low level (−) 1.5; high level (+) 3.5. 3: malachite green (10−4 mol l−1 ); low level (−) 1.0; high level (+) 5.0.

of the reaction coil and the total flow rate. The length of the PTFE reaction coil (1.6 mm i.d.) was modified from 50 to 150 cm and the total flow rate from 2.0 to 4.0 ml min−1 . In both studies, the composition of the reagent solution was 0.20 mol l−1 in ammonium molybdate, 2.0 mol l−1 in sulfuric acid and 4 × 10−4 mol l−1 in malachite green.

obtained for a concentration of malachite green in the range from 4.0 × 10−4 to 5.0 × 10−4 mol l−1 . It was not observed an alteration in the sensitivity when the sulfuric acid and ammonium molybdate concentrations were changed in the reagent solution. Although the

A slight increase was verified in the value of the blank signal when the concentration of the ammonium molybdate in the reagent solution was increased from 0.05 to 0.5 mol l−1 , as can be seen in Fig. 2a. It was also observed that there is a more significant increase in the blank signal when the sulfuric acid concentration is increase (Fig. 2b). For reagent solutions for concentration sulfuric acid lower than 1.5 mol l−1 it was observed that the resulting mixture turns the monosegment unstable, breaking the second air bubble and as a consequence, losing the integrity of the reacting mixture. The component of the reagent solution that showed a most pronounced effect on the intensity of the blank signal was the malachite green as can be seen in Fig. 2c. The blank signal increase with the concentration of malachite green, obeying the Beer’s Law. This behavior originates the fact that the neutral molecule of malachite green absorbs in the same wavelength that the ion association complex does. This high absorption could impair the linear analytical range, with consequent loss of resolution of the measurement. Another disadvantage with respect to the high blank signal was the adsorption of the malachite green on the internal wall of the tube and in flow cell when a 5.0 × 10−4 mol l−1 solution malachite green was used, resulting in a long cleaning time, decreasing therefore the analytical frequency. The increase in the concentration of the malachite green in the reagent solution favored an increase in the sensitivity, as can be observed in Fig. 3. A maximum sensitivity was

0.06 0.04 0.02

(a) 0.00 0.0

0.1

0.2

0.3

0.4

Molybdate concentration / mol l

0.5

0.6

-1

0.08

Absorbance

3.1. Effect of reagent concentration on the blank signal and the sensitivity

0.06 0.04 0.02

(b) 0.00 1

2

3

4

5

Sulfuric acid concentration / mol l

6

-1

0.08 0.06

Absorbance

3. Results and discussion

Absorbance

0.08

0.04 0.02

(c) 0.00 0.000

0.001

0.002

0.003

0.004

0.005

Malachite green concentration / mol l

0.006

-1

Fig. 2. Effect of the component concentrations in the reagent solution on the blank signal: (a) ammonium molybdate, (b) sulfuric acid and (c) malachite green.

M.C.T. Diniz et al. / Talanta 62 (2004) 469–475

0.10

473

-1

[malachite green] / mol l -4

0.5x10

-4

1.0x10

0.08

-4

Absorbance

2.0x10

-4

4.0x10

0.06

-4

5.0x10 0.04

0.02

0.00 0

5

10

15

20

25 3- -1

Phosphorous concentration / µg P-PO4 l Fig. 3. Effect of the malachite green concentration on the sensitivity.

literature shows that the formation of the ion association complexes is influenced by the sulfuric acid and molybdate, this fact was not confirmed as an increase in the sensitivity was not observed when the concentrations of these species were individually modified. 3.2. Effects of the reagent concentrations using factorial design A factorial design was proposed to evaluate the individual effects and the possible interactions among the components of the reagent solution on the sensitivity and the blank signal. The experiments were carried out in duplicate in agreement with the conditions established in Table 2, producing two groups of 16 responses as also presented in Table 2. The effects of each variable and the possible interactions are presented in the Table 3 confirming that malachite green Table 3 Principal effect and interaction values for 23 factorial Sensibility (×10−3 AU ␮g−1 l−1 ) Principal effects 1: Ammonium molybdate 2: Sulfuric acid 3: Malachite green Interaction of two factors 12 13 23 Interaction of three factors 123

Blank signal absorbance

0.43 ± 0.09 0.76 ± 0.09 1.54 ± 0.09

−0.001 ± 0.002 0.020 ± 0.002 0.041 ± 0.002

0.42 ± 0.09 0.58 ± 0.09 −0.33 ± 0.09

0.002 ± 0.002 −0.006 ± 0.002 0.014 ± 0.002

0.16 ± 0.09

0.003 ± 0.002

is the component whose main effect is more significant for a larger increase at the sensitivity. This can be verified by the largest value (1.54×10−3 AU ␮g−1 l−1 ) indicated in Table 3. The main effects of sulfuric acid and of ammonium molybdate show a synergism in respect to the sensitivity, although the results found are less significant when compared to the main effect of malachite green. The results also show that the interactions between molybdate and sulfuric acid (1 and 2) as well as between molybdate and malachite green solutions (1 and 3) produce a synergic effect on the sensitivity. Therefore, both values (Table 3) have the same importance. On the other hand, the interaction between sulfuric acid and malachite green solutions showed an antagonistic effect, as the combination of these two components did not contribute for sensitivity gain. This antagonistic interaction can justify the results obtained in the previous experiments when the concentration of each component of the solution was individually altered, that is an increase in the sensitivity was not observed for these two components. The values found for the interaction among the three components of the reagent solution do not contribute for a sensitivity gain. It is not considered a statistically significant effect, because its value is as lower as the error values of the measurements. As far as the increase of the blank signal in concerned, the most significant effect is due to the malachite green. The sulfuric acid also presents a significant effect however it is less pronounced. These results are in agreement with the results discussed previously for the individual experiments. The interactions between molybdate and sulfuric acid (1 and 2), as well as the (1 and 3) and (1, 2 and 3) interactions are not statistically significant in relation to the increase of the blank signal only the interaction between sulfuric acid and

474

M.C.T. Diniz et al. / Talanta 62 (2004) 469–475

malachite green (2 and 3) that contributes in smaller degree when compared with the malachite green main effect. According to the individual studies and the factorial design, the best composition of the reagent solution was defined as 0.20 mol l−1 in ammonium molybdate, 2.0 mol l−1 in sulfuric acid and 4.0 × 10−4 mol l−1 in malachite green. This composition reflects a compromise between high sensitivity and low blank signal. 3.3. Effect of the length of the reaction coil It was not observed alteration on the analytical signal being due to the variation of the length of the reaction coil, indicating that the reaction of formation of the ion association complexes is fast. In this sense, the length of the reactor was fixed in 50 cm, which was enough for the reaction reach the equilibrium. 3.4. Effect of the total flow rate There were no significant differences between the sensibility (slopes of the analytical curves) for the flow rate varying from 1.0 to 4.0 ml min−1 . However, for a flow rate higher than 2.0 ml min−1 it was observed a distortion of the analytical signal, making the precision of the measurements. Besides, it was also observed that high flow rate (4 ml min−1 ) causes the breaking of the monosegment. Therefore, the flow rate was set at 2.0 ml min−1 . 3.5. Effect of concomitant ions The effect of different ions on the analytical signal was evaluated for a phosphate reference solution of 50 ␮g P PO4 3− l−1 . Table 4 presents the values of maximum concentration allowed for the ions studied. Higher concentrations than the values presented in Table 4 caused interference of 2% or higher.

Table 5 Determination of phosphate in water of lakes by proposed method and standard method Sample

Proposed method (␮g P PO4 3− l−1 )

Standard method (␮g P PO4 3− l−1 )

Relative error (%)

Zara lake Maria Bonita lake Lara lake

25.32 ± 0.4 9.47 ± 0.2 38.65 ± 0.6

24.15 ± 0.9 9.53 ± 0.5 39.15 ± 0.9

4.84 −0.63 1.28

10−4 mol l−1 in malachite green. These conditions provided a linear analytical range from 5 to 75 ␮g l−1 of phosphorous, that fits the equation A = 0.021 + 0.0029C (r = 0.9992), where A is the absorbance peak height and C is the concentration of phosphorous in the solution in ␮g P PO4 3− l−1 . A detection limit of 0.70 ␮g P PO4 3− l−1 was achieved, with an analytical frequency of 72 determinations per hour. The repeatability of the system was evaluated by successive injections of a 20 ␮g P PO4 3− l−1 phosphate reference solution, presenting a relative standard deviation of 2% (n = 8). The method was applied to the phosphate determination in waters of lakes collected in the Amazonian region. These samples were provided by the Laboratory of Environmental Chemistry of IQ/UNICAMP. The results obtained by the proposed and standard methods are presented in Table 5. Before doing the determinations of phosphate, samples were filtered though acetate cellulose membranes (45 ␮m) for elimination of particulate material and preserved in nitric acid. The standard procedure described in the literature [18] was taken for reference, which is based on the phosphomolybdenum blue reaction using ascorbic acid as reducer and potassium tartarate and antimony as catalysts. A cell with an optical length of 5 cm was also employed. These two methods are in agreement at a 95% confidence level.

3.6. System characteristics and application

4. Conclusions

The best work conditions were defined as: length of the reaction coil of 50 cm, total flow rate of 2.0 ml min−1 , sample volume of 300 ␮l, volume of reagent solution of 25 ␮l and composition of the reagent solution of 0.2 mol l−1 in ammonium molybdate, 2.0 mol l−1 in sulfuric acid and 4.0×

The proposed method which employs a monosegmented system with simultaneous multiple injection, was proved to be efficient for phosphate determination in waters from lakes, where the concentration of this species is very low. A single reagent solution was used, constituted by all the reactants necessary for formation of the ion association complex, being eliminated the addition of another reagent through a confluence point in the system. When compared to other methods described in the literature based on the same chemical reaction, it can be concluded that the proposed system presents a better sensitivity, because the sample dispersion in the MSFA system is smaller than that obtained in FIA systems. The factorial design explained the importance of all the components of the reagent solution for the sensitivity and the blank signal, revealing that the malachite green is the

Table 4 Tolerable concentrations of co-existing species Species

Maximum tolerable concentration

NaCl NaNO3 K+ , Fe3+ , Ca2+ , Cu2+ , Mg2+s SiO3 2− HCO3 − AsO4 3−

0.05 mol l−1 0.02 mol l−1 50 mg l−1 1000 ␮g l−1 50 ␮g l−1 20 ␮g l−1

M.C.T. Diniz et al. / Talanta 62 (2004) 469–475

main component that shows the most pronounced effect on these parameters. Acknowledgements The authors are grateful to Dr. Ivo M. Raimundo J. for manuscript revision and to CNPq for the scholarship to M.C.T.D. References [1] D.E.C. Corbridge, Phosphorus, an Outline of its Chemistry, Biochemistry and Uses, Elsevier, Amsterdam, 1995. [2] W. Stumm, J.J. Morgan, Aquatic Chemistry: An Introduction Emphasizing Chemical Equilibria in Natural Waters, Wiley, New York, 1981.

475

[3] F.A. Esteves, Fundamentos de Limnologia, Interciˆencia, Rio de Janeiro, 1988. [4] J.B. Martin, D.M. Doty, Anal. Chem. 21 (1949) 965. [5] J.F. Kennedy, D.A. Weetman, Anal. Chim. Acta 55 (1971) 448. [6] O. Røtset, Anal. Chim. Acta 178 (1985) 217. [7] J. Murphy, J.P. Riley, Anal. Chim. Acta 27 (1962) 31. [8] A. Henriksen, Analyst 88 (1963) 88. [9] A. Henriksen, Analyst 90 (1965) 29. [10] T. Nasu, S. Kaneko, Anal. Sci. 5 (1989) 449. [11] D.T. Burns, D. Chimpalee, N. Chimpalee, S. Ittipornkul, Anal. Chim. Acta 254 (1991) 197. [12] W. Fusheng, W. Zhongxiang, T. Enjiang, Anal. Lett. 22 (1989) 3081. [13] M. Kan, T. Nasu, M. Taga, Anal. Sci. 7 (1991) 87. [14] E.B. Cogan, G.B. Birrell, H. Griffith, Anal. Biochem. 271 (1999) 29. [15] H.J. Altmann, E. Furstenau, A. Gielewski, L. Scholz, Anal. Chem. 256 (1971) 274. [16] C. Pasquini, W.A. Oliveira, Anal. Chem. 57 (1985) 2575. [17] V.O. Brito, I.M. Raimundo Jr., Anal. Chim. Acta 371 (1998) 317. [18] A.D. Eaton, L.S. Clesceri, A.E. Greenberg, Standard Methods for the Examination of Water and Wastewater, Washington, 1995.