Optimization of Flow Injection On-Line Microcolumn Preconcentration of Ultratrace Elements in Environmental Samples prior to Their Spectrochemical Determination

MICROCHEMICAL JOURNAL ARTICLE NO.

54, 391–401 (1996)

0116

Optimization of Flow Injection On-Line Microcolumn Preconcentration of Ultratrace Elements in Environmental Samples prior to Their Spectrochemical Determination1 ZS. HORVA´TH,2 A. LA´SZTITY, K. ZIH-PERE´NYI,

AND

´ . LE´VAI A

Institute of Inorganic and Analytical Chemistry, Eo¨tvo¨s L. University, P.O. Box 32, Budapest 112, H-1518 Hungary The preconcentration of some elements such as Cd, Co, Ni, and V(IV) was modeled in the presence of complexing agents such as citrate and oxalate at high Ca, Mg, and sulfate concentrations on iminodiacetic acid/ethyl cellulose (IDAEC), a chelating cellulose. The effect of the species present in the solution was studied after construction of the species distribution curves using critical, estimated, and measured stability constants. The stability constants of the IDAEC chelates were determined potentiometrically. The constants were calculated or estimated using computer programs. The diagrams calculated in homogeneous media were used for optimization of the flow injection on-line preconcentration for analysis of ultratrace metals in the highly mineralized water ‘‘Hunyadi.’’ q 1996 Academic Press, Inc.

INTRODUCTION

Chelating ion exchangers including chelating celluloses are widely used in flow injection (FI) on-line systems as microcolumns for preconcentration of trace elements from environmental samples (1). The stability of metal chelates in solution and on the chelating celluloses depends on the experimental conditions, namely, on the presence of buffering and masking substances. The matrix elements of the sample also influence the recovery of the trace elements. Previous results showed that iminodiacetic acid/ethyl cellulose (IDAEC) chelating ion exchanger possessed ideal properties for the preconcentration of trace metals in on-line FI inductively coupled plasma atomic emission spectrometry (ICP–AES) systems (2). IDAEC readily forms chelates, mostly with heavy metals. The exchange rate on the iminodiacetic group is high enough for determination of the stability constants using dynamic potentiometric titrations (3). At preconcentration, the metals can be easily eluted from the microcolumn with acid by injection. The idea was to foretell the behavior of trace metal ions in the presence of complexing agents and high salt content in heterogeneous media when the reaction does not attain equilibrium by using equilibrium calculations in solution. The aim of the work was the optimization of determination of ultratrace metals in a model mineral water of high salt content called ‘‘Hunyadi.’’ It is a hard sample containing high concentrations of Ca, Mg, and sulfate. The direct determination of Cd using matrix modification 1 Submitted in conjunction with the Seventh Italian – Hungarian Symposium on Spectrochemistry: Innovative Methodologies for Health and Environmental Protection, Rome, Italy, November 27 – December 1, 1995. 2 To whom correspondence should be addressed.

391 0026-265X/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

´ TH ET AL. HORVA

392

and Zeeman background correction could not be achieved. The Cd concentration was less than 2 mg/liter (4). The preconcentration of Cd, Co, Ni, and V(IV) was modeled in the presence of complexing agents like citrate and oxalate and at a high Ca, Mg, and sulfate concentrations on IDAEC. Citrate was chosen as the stabilities of citrate complexes are similar to those of humic acids. The effect of the species present in the solution was studied after construction of the species distribution diagrams. For the computation of species distributions from measured, estimated, and critical stability constants, program SPEA, a FORTRAN computer program, and for computation of potentiometric stability constants, program BESTA were used (5). For unknown stability constants an estimation was needed. Predicted stability constants were estimated by the use of stability constants of the metal complexes of analogous ligands compiled in Critical Stability Constants (6) using the method reported by Martell and Motekaitis (5). MATERIALS AND METHODS

Throughout the study high-purity water (distilled water purified on ion-exchange celluloses) was employed. Ammonium acetate was purified on an IDAEC column. All reagents were of analytical grade. Atomic absorption spectrometry standards supplied by Merck (Darmstadt, Germany) were used to prepare calibration solutions. Instrumentation A Perkin–Elmer Model 3110 atomic absorption spectrometer with HGA-600 graphite furnace and AS-60 autosampler was used. Radelkis Type OP 930/1 automatic titrator (made in Hungary) and Type OP 208/1 precision digital pH meter (resolution, 0.001 pH; accuracy, { 0.005 pH) were also used. Operating conditions and analytical parameters were from Perkin–Elmer’s Analytical Techniques for Graphite Furnace Atomic Absorption Spectrometry. Potentiometric Titration Potentiometric titration of 200 mmol IDAEC in the acid form was carried out dynamically in an automated titration system in a 50-ml volume at 22 { 17C in 1 M NaClO4 solution under an atmosphere of CO2-free N2 with a glass–calomel electrode pair in the absence and in the presence of 200 mmol citric acid and 100 mmol of the metal ion being investigated. The heterogeneous equilibria were reached within some minutes. Preconcentration Multielement preconcentration of the elements was performed by on-line flow injection on an IDAEC microcolumn. The capacity of IDAEC was 0.8 mmol/g. Into the column was placed 40 mg of IDAEC in NH4 form. The sample was buffered with ammonium acetate. The optimum pH range for the on-line FI system was 3–6. A dual four-way valve assembly was used for the preconcentration and elution of the trace elements (2). The eluant was 2 M HNO3 ; the flow rate 2.5 ml/min. Eluates were collected into the vessels of the autosampler of the spectrometer. The sample volume loaded was

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

393

OPTIMIZATION OF PRECONCENTRATION TABLE 1 Stability Constants Used for Calculations of the Species Distribution Curves Log b M Element H

Cd

Co

Ni

V(IV)

Ca Mg

a

L Ligand

ML

IDAEC Citrate Oxalate Sulfate IDAEC Citrate Oxalate Sulfate IDAEC Citrate Oxalate Sulfate IDAEC Citrate Oxalate Sulfate IDAEC Citrate Oxalate Sulfate IDAEC Citrate IDAEC Citrate

9.10 5.33 3.55 1.35 8.73 3.15 2.75 1.08 8.65 5.40 3.84 0.57 8.86 5.40 4.16 0.57 10.88 8.84 6.45 2.00 3.90 3.55 3.75 3.40

ML2

MHL

MH2L

11.75 9.41 4.59

12.21

15.40 7.69

7.53

15.46 8.35 5.16 16.05 8.70

18.40 12.34 11.78

5.60a

MOHL.

20.0 ml; the eluted volume was 0.80–0.95 ml. Usually at about 20- to 30-fold, preconcentration was achieved. The loaded sample contained 0.002 M citric or oxalic acid or 0.02 M citric acid and was buffered with ammonium acetate to pH 3–5.5 { 0.1. For the calibration curves the concentrations (mg/liter) of the elements before preconcentration were: Cd, 0.025 – 0.2; Co, 0.5 – 10; Ni, 2.5 – 10; V(IV), 0.6 – 5. For the sorption study the concentrations (mg/liter) were: Cd, 0.05; Co, 1.0; Ni, 2.5; V(IV), 5.0. For the analysis of NIST 1643c Trace Elements in Water Standard Reference Material was used after 40 times dilution for preconcentration. RESULTS AND DISCUSSION

The stability constants of Ca–, Cd–, Co–, Ni–, and Zn–IDAEC complexes were determined by potentiometric titration and calculated using program BESTA (5). Table 1 shows the calculated and critical stability constants used for calculation of the species distribution curves.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

´ TH ET AL. HORVA

394

TABLE 2 Estimated and Measured Stability Constants of IDAEC Chelates Log b Element

Predicted ML

ML2

Measured ML2

V(IV) Cu Pb Zn

10.88

18.40 18.42 16.24 15.24

15.17

Because of the difficulties in potentiometric determination of the stability constants of Cu, Pb, and V(IV), the stability constants were estimated. In Table 2 are listed the predicted stability constants of Cu–, Pb–, and V(IV)–IDAEC, as calculated from the stability constants of metal complexes of analogous ligands. The ratio of stability constants for a pair of metal ion complexes is nearly constant for a large number of analogous ligands (5). For the prediction of IDAEC complex stabilities, iminodiacetic acid derivates were used as analogous ligands and Zn or Co was used as metal ion pair. The reliability of the estimated values was checked by the stability constant of Zn–IDAEC. The difference between the estimated and measured constants is not more than one-tenth of a log unit, which is satisfactory. The predicted and measured stability constants were used for calculation of the species distribution curves for metal species in the presence of complexing agents and at high Ca, Mg, and sulfate concentrations. The concentrations of Ca, Mg, and sulfate were as high as in the model ‘‘Hunyadi’’ mineral water, whereas the concentrations of the metal ions were several orders of magnitude lower than the salt content. Table 3 lists the major constituents of ‘‘Hunyadi’’ mineral water. Figure 1 shows the species distribution diagrams of Ni species in the presence of complexing agent and high Ca, Mg, and sulfate contents. It can be seen that sulfate has no effect above pH 3. When the citrate concentration is 0.02 M at pH 4 and 5, a very low percentage of Ni is in the form of IDAEC chelates (Fig. 1A). At the same citrate concentration, when the concentrations of Ca, Mg, and sulfate are as high as TABLE 3 Major Constituents of ‘‘Hunyadi’’ Mineral Water Examined

ah04$$1363

Constituent

Concentration (g/liter)

Sodium Calcium Magnesium Chloride Sulfate Bicarbonate Total solids

5.20 0.54 3.24 0.79 27.0 1.01 39.5

10-01-96 05:24:11

mica

AP: MICROCHEM

OPTIMIZATION OF PRECONCENTRATION

395

FIG. 1. Species distribution diagrams of Ni species as a function of pH. Component concentration in mol/liter: (A) L Å IDAEC, 0.0012; Y Å citrate, 0.02; Ni, 1.7 1 1005. (B) L Å IDAEC, 0.0012; Y Å citrate, 0.02; S Å sulfate, 0.225; Ni, 5 1 1007; Ca, 0.0135; Mg, 0.135. (C) L Å IDAEC, 0.0012; S Å sulfate, 0.225; Ni, 5 1 1007; Ca, 0.0135; Mg, 0.135.

in ‘‘Hunyadi’’ mineral water (Table 3), the free citrate concentration was decreased because of the formation of citrato complexes of Ca and Mg. In this case the estimated sorption of Ni at pH 5 is 87% (Fig. 1B). In this matrix in the absence of citrate (Fig. 1C) the sorption of Ni is 100%.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

´ TH ET AL. HORVA

396

FIG. 1—Continued

When the concentration of the matrix is the same as in the case of Ni, the computed curves of Cd and Co show a species distribution similar to that of Ni. Figure 2 shows the species distribution diagrams of V(IV) species as a function of pH only in the range pH 2 to 7, because above pH 7 vanadate(IV) ion (V4O20 9 ) can be formed. It can be seen from Figs. 2A and B that V(IV) is almost quantitatively in citrato or oxalato complexes between pH 4 and 5. The stabilities of vanadium(IV) citrato and oxalato chelates are higher than those of the other elements studied (Table 1). Between pH 5 and 7, formation of the V(IV)–IDAEC chelate is observed. As in the case of Ni in the presence of high concentrations of Ca and Mg, the effect of citrate is suppressed because of the formation of Ca– and Mg–citrato complexes (Fig. 2C). In the absence of citrate, as Fig. 2D shows, V(IV) is almost quantitatively in V(IV)–IDAEC form from pH 4 to 7. Table 4 gives the sorption of elements in ultratrace concentration on IDAEC at pH 4 and 5 in the presence and absence of 0.002 M citrate and oxalate estimated from the species distribution curves and found experimentally. In the cases of Cd, Co, and Ni, the experimental values are in accordance with the estimations. The sorption is high even in the presence of citrate and oxalate. Formation of V(IV)–IDAEC was highly influenced by these ligands. Table 5 presents the recovery of the elements on IDAEC from ‘‘Hunyadi’’ mineral water at pH 5 at different citrate concentrations. From the species distribution curves, pH 5 was found to be optimum for multielement preconcentration. At this pH the sorption of Ca and Mg was negligible. The experimentally measured values for Ni are in good agreement with the values read from Figs. 1B and C. Recovery is near 100%. The same was found for Cd and Co. The estimated V(IV) sorption from Fig. 2C in the presence of 0.002 M citrate at pH 5 was 90%, but the measured sorption

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

OPTIMIZATION OF PRECONCENTRATION

397

FIG. 2. Species distribution diagrams of V(IV) species as a function of pH. Component concentration in mol/liter: (A) L Å IDAEC, 0.0012; Y Å citrate, 0.002; VO Å vanadyl ion, 5 1 1008. (B) L Å IDAEC, 0.0012; X Å oxalate, 0.002; VO Å vanadyl ion, 5 1 1008. (C) L Å IDAEC, 0.0012; Y Å citrate, 0.002; S Å sulfate, 0.225; VO Å vanadyl ion, 5 1 1007; Ca, 0.0135; Mg, 0.135. (D) L Å IDAEC, 0.0012; S Å sulfate, 0.225; VO Å vanadyl ion, 5 1 1007; Ca, 0.0135; Mg, 0.135.

was only 27%, probably because of the inertness of the V(IV) complexes. This means that when V(IV) is in very low concentration, it should not be preconcentrated in the presence of citrate. From Table 5 it can also be read that if citrate is not present,

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

´ TH ET AL. HORVA

398

FIG. 2—Continued

the sorption is 89% for V(IV). The same percentage was found when V(V) was preconcentrated. To test the performance of the on-line preconcentration method on IDAEC, NIST 1643c Trace Elements in Water Standard Reference Material was analyzed for Cd, Co, Ni, and V. Appropriate agreement was obtained with the certified values, as shown in Table 6.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

399

OPTIMIZATION OF PRECONCENTRATION TABLE 4 Sorption of Elements on IDAEC Estimated and Measured Cd

pH 4

pH 5

Water 0.002 M 0.002 M Water 0.002 M 0.002 M

citrate oxalate citrate oxalate

Co

Ni

V

Ea

Mb

E

M

E

M

E

M

100 98 — 100 86 96

109 91 — 112 102 94

100 87 — 100 94 79

99 97 — 83 97 98

100 75 — 100 68 82

93 82 — 99 86 104

100 10 0 100 8 1

78 35 11 96 27 29

Note. Concentrations in mg/liter: Cd, 0.05; Co, 1.0; Ni, 2.5; V(IV), 5.0. a E Å % sorption estimated from the species distribution diagrams. b M Å % sorption measured (average of two separate determinations).

Table 7 summarizes the analysis of ‘‘Hunyadi’’ mineral water. According to the results there is good agreement between external calibration and standard addition. There is no need for standard addition for the analysis of ‘‘Hunyadi’’ mineral water and most probably for analysis of other highly mineralized waters. Table 8 gives the detection limits. The measured concentrations in ‘‘Hunyadi’’ are at least 10 times higher. CONCLUSIONS

The results of the preconcentration experiments of ultratrace elements on IDAEC chelating cellulose ion exchanger in the presence of complexing agents and high salt content were in accordance with the results expected from species distribution diagrams calculated in homogeneous media for Cd, Co, and Ni. In some cases, V(IV) behavior was different probably because of the inertness of V(IV)-complexes. The analysis of the highly mineralized water ‘‘Hunyadi’’ for ultratrace elements Cd, Co, Ni, and V(IV) could be optimized.

TABLE 5 Recovery of Elements on IDAEC from ‘‘Hunyadi’’ Mineral Water at pH 5: Estimated and Measured Cd

Co

Ni

V

Citrate concentration

Ea

Mb

E

M

E

M

E

M

— 0.002 M 0.02 M

100 100 79

99 111 114

100 100 75

88 89 87

100 100 87

89 98 79

100 90 —

89 27 —

Note. The results given are the averages of three separate determinations. a E Å % sorption estimated from the species distribution diagrams. b M Å % sorption measured by spike recovery. Spike concentrations in mg/liter: Cd, 0.05–0.15; Co, 0.25–7.5; Ni, 2.5–10; V(IV), 0.5–5.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

´ TH ET AL. HORVA

400

TABLE 6 Recovery from NIST 1643c on IDAEC Concentration (mg/liter) Element

Certifieda

Founda

Cd Co Ni V

12.2 { 1.0 23.5 { 0.8 60.6 { 7.3 31.4 { 2.8

11.2 { 0.4 24.0 { 1.3 66.8 { 1.3 29.7 { 4.3

a The standard was diluted 40 times before the preconcentration.

TABLE 7 Results of Analysis of ‘‘Hunyadi’’ Mineral Water by Flow Injection Graphite Furnace Atomic Absorption Spectrometry after Preconcentration on IDAEC at pH 5 Element Citrate concentration

Ce

Co

External calibration — 0.082 4.70 0.097 5.29 2 1 1003 M 0.101 5.32 2 1 1002 M Average 0.093 5.10 Standard addition — 0.085 — 0.089 5.83 2 1 1003 M 0.085 5.70 2 1 1002 M Average 0.086 5.76 Concentration of ultratrace elements mg/liter 0.089 5.4 (0.010) (0.4) ({SD)

Ni

V

5.95 5.83 6.7 6.16

1.27 — — —

6.55 6.05 — 6.30

1.14 — — —

6.2 (0.3)

1.2

a

Results in mg/liter. Each measurement was repeated three times.

TABLE 8 Detection Limitsa Element

DL (mg/liter)

Cd Co Ni V

0.009 0.083 0.25 0.081

a

Refers to the original solution, 3s criterion, seven blanks.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM

OPTIMIZATION OF PRECONCENTRATION

401

ACKNOWLEDGMENTS We thank To´th Ka´rolyne´ for her technical assistance. The research was sponsored by the National Foundation of Scientific Research (OTKA No. A196/95/450).

REFERENCES 1. 2. 3. 4. 5.

Fang, Z. Spectrochim. Acta Rev., 1991, 14, 235. Caroli, S.; Alimonti, A.; Petrucci, F.; Horva´th, Zs. Anal. Chim. Acta, 1991, 248, 241. Horva´th, Zs.; Nagydio´si, Gy. J. Inorg. Nucl. Chem., 1975, 37, 767. Bozsai, G.; Schlemmer, G.; Grobenski, Z. Talanta, 1990, 37, 545. Martell, A. E.; Motekaitis, R. J. Determination and Use of Stability Constants. VCH, Weinheim/New York, 1992. 6. Smith, R. M.; Martell, A. E. Critical Stability Constants, Vols. 1–5. Plenum, New York, 1974, 1975, 1976, 1977, 1982.

ah04$$1363

10-01-96 05:24:11

mica

AP: MICROCHEM