Anal Bioanal Chem (2003) 375 : 1097–1100 DOI 10.1007/s00216-002-1728-6
S P E C I A L I S S U E PA P E R
L. P. C. Romão · G. R. Castro · A. H. Rosa · J. C. Rocha · P. M. Padilha · H. C. Silva
Tangential-flow ultrafiltration: a versatile methodology for determination of complexation parameters in refractory organic matter from Brazilian water and soil samples Received: 26 June 2002 / Revised: 30 October 2002 / Accepted: 27 November 2002 / Published online: 6 February 2003 © Springer-Verlag 2003
Abstract In this work the copper(II) complexation parameters of aquatic organic matter, aquatic and soil humic substances from Brazilian were determined using a new versatile approach based on a single-stage tangential-flow ultrafiltration (TF-UF) technique (cut-off 1 kDa) and sensitive atomic spectrometry methods. The results regarding the copper(II) complexation capacity and conditional stability constants obtained for humic materials were compared with those obtained using direct potentiometry with a copper-ion-selective electrode. The analytical procedure based on ultrafiltration is a good alternative to determine the complexation parameters in natural organic material from aquatic and soil systems. This approach presents additional advantages such as better sensibility, applicability for multi-element capability, and its possible to be used under natural conditions when compared with the traditional ion-selective electrode. Keywords Complexation parameters · Copper · Organic matter · Humic substances · Ultrafiltration · Ion-selective electrode
Introduction The refractory organic matter present in water and soil environments exerts strong influence on the transport, reactivity, and bioavailability of metals in the environment.
L. P. C. Romão (✉) Universidade Federal de Sergipe, Jd. Rosa Elze, s/n, 49100–000 São Cristóvão-SE, Brazil e-mail:
[email protected] L. P. C. Romão · G. R. Castro · A. H. Rosa · J. C. Rocha · H. C. Silva Instituto de Química de Araraquara-UNESP, P.O. Box 355, 14800–900 Araraquara-SP, Brazil P. M. Padilha Instituto de Biociências de Botucatu-UNESP, P.O. Box 510, 18618–000 Botucatu-SP, Brazil
Heavy metal complexation by natural organic matter (NOM), which occurs naturally, has been studied through various approaches such as ion-selective electrode (ISE), polarography and voltammetry methods, fluorescence quenching and ultrafiltration procedures [1]. However, these methods have different drawbacks; for example, poor sensitivity, and restriction to only few metals or working pH and disruption of equilibrium conditions [1, 2, 3, 4]. Potentiometric methods (ISE) are the most used methods for metal–NOM complexation studies, because of their speed and sensitivity to the free hydrated-ion activity; however, they suffers a lack of sensitivity, difficulties to obtain measurements in natural waters, and a supporting electrolyte is required [5]. Ultrafiltration (UF) methods are merely limited by detection limits of the applied determination method (e.g., atomic spectrometry) and they are applicable to a wide range of different elements. Others advantages are that the methods have no ionic strength limitations and do not disturb the complexation equilibria as is the case with ion-exchange chromatography [3, 5, 6]. However, UF suffers from disadvantages such as metal sorption on the membrane and incomplete separation. Nifanteva et al. [3] found that the main pitfall of the UF approach was probably the fact that by describing metal–HS interactions the values of the retention coefficient of metal (RM) are obtained for metal retention in the absence of ligand. If RM increases, it may be due the formation of hydroxocomplexes of the metals and their adsorption on the membrane or walls of the cell, giving rise to the formation of colloid particles. However, in the presence of NOM ligand the competitive reactions of the complexation may decrease the importance of this process and thus decrease the real RM values. In this work a versatile methodology using a singlestage tangential-flow ultrafiltration (TF-UF) technique and sensitive atomic spectrometry methods is proposed to study complexation parameters of Brazilian water and soil samples.
1098
Experimental Reagents All reagents used were of high-purity grade unless otherwise stated. Diluted acids and bases were prepared by diluting 30% hydrochloric acid (suprapur, Merck AG), 65% nitric acid (p.a. Merck AG, pre-purified by sub-boiling distillation), and sodium hydroxide-monohydrate (suprapur, Merck AG) with high-purity water (Milli-Q system, Millipore). For metal determinations and their calibration synthetic standards were used (Merck AG). Isolation and purification of aquatic humic substances (AHS) and soil humic substances (SHS) The AHS and SHS were extracted from natural samples collected at different Brazilian sites according to procedures recommended by the International Humic Substances Society [7]. AHS isolation About 50 L of surface water was field filtered through 0.45-µm cellulose-based membranes and acidified with concentrated HCl solution to pH 2.0. The AHS from the acidified sample was then isolated using the XAD 8 collector in the conventional way following the recommendations of Malcolm [8]. After elution with 0.1 mol L–1 NaOH solution, the concentrate obtained (4.5 mg mL–1 DOC equivalent to 9.0 mg mL–1 aquatic HS) was neutralized to pH 7.0 with 0.1 mol L–1 HCl solution, stored in high-density polyethylene containers, and maintained at 4 °C. SHS extraction The dried soil samples were ground and passed through a 2-mm sieve. Five grams of the resulting material were extracted with a volume of 0.1 mol L–1 HCl equal to ten times the weight of the sample. The pH of the solution was adjusted to between 1.0 and 2.0 with 1.0 mol L–1 HCl. The soil/HCl mixture was shaken for 1 h and the suspension was allowed settle. The mixture was centrifuged at 1,478 g for 10 min, and the supernatant was separated from the sediment. The sediment was neutralized with 1.0 mol L–1 NaOH (to pH 7.0), and a volume of 0.1 mol L–1 NaOH equal to ten times the weight of the sample was added under nitrogen atmosphere. The mixture was shaken for 4 h and then allowed to settle
overnight. The supernatant was separated from the sediment by centrifugation at 1,478 g for 10 min, and the sediment discarded. The supernatant (HS) was evaporated in a vacuum and dried in a circulating-air oven at 55 °C [9]. The purification of the AHS and SHS was done by dialysis tubing and drying by lyophilization using a Savant model E-C. The NOM, AHS, and SHS samples investigated are described in Table 1. Cu(II) complexation capacity Ultrafiltration procedure The complexation capacity (CC) of the samples was determined using an alternative approach based on a single-stage tangential-flow ultrafiltration (Sartorius Ultrasart X) technique equipped with a membrane filter (polyethersulfone, Gelman Pall–Filtron OMEGA, 1-kDa cut-off and 47-mm diameter) and atomic spectrometry methods (ET-AAS). The samples (200 mL, pH 5.0, and I = 0.1 mol L–1 NaNO3) were pumped through the TF-UF system, and a small volume of a 9.45×10–3 mol L–1 Cu(II) solution was added step by step up to 4.72×10–4 mol L–1 Cu(II). After establishment of each equilibrium, in practice reached within 10 min, small volumes (approximately 2 mL) of samples fractions, containing free Cu(II), were isolated by filtration through the UF device, acidified with dilute HNO3, and then analyzed by ET-AAS according to [10]. Ion selective electrode (ISE) procedure The Orion model 94–29 Cupric electrode was used to measure free cupric ions during titration. It was connected to a Tecnal model Tec-2 pH meter. The Orion model 90–02 double-junction reference electrode was also connected to a pH meter. The reference electrode was a sleeve-type Ag/AgCl with an outer chamber filled with 10% KNO3 solution isolating the inner reference element and filling solution (saturated AgCl solution) from the sample. Values of pH were measured during each titration using an Orion combination pH electrode connected to a model 250A Orion pH meter. Before each experiment a calibration curve was performed to test the ion-selective electrode slope which was influenced by temperature (at 25 °C the slope is 29.6 mV). Experiments were performed under constant light conditions because a Cu(II) ISE is photosensitive. [11] Titrations were performed in a Pyrex cell (capacity 100 mL) covered with a Teflon cover with five holes for insertion of electrodes, a nitrogen bubbler, and microburette tips. A Radelkis OP 936 microburette supporting a 10-mL glass syringe was used to
Table 1 Location and properties of the natural aquatic organic matter, aquatic and soil humic substances studied Label
Location
Isolation
pH
TOC (mg L–1)
Total metal content (mg kg–1 DOC) Al
IS
Water of Itapitanguí Stream/ São Paulo State
IS-XAD 8 IR
Water of Iguape River/ São Paulo State
IR-XAD 8 RN-XAD 8 RNS-1 RNS-2
Water of Rio Negro/ Amazonas State Flooded soil of Rio Negro Basin/ Amazonas State Unflooded soil of Rio Negro Basin/ Amazonas State
Fe
Ni
Cd
Original sample
4.92
38.7±1.2
79.97
28.23
0.16
0.04
XAD-8 procedure Original sample
4.40
34.2±0.7
35.49
39.16
0.05
0.02
5.25
41.6±0.5
12.06
42.72
0.11
0.04
XAD-8 procedure XAD-8 procedure IHSS procedure
3.75
36.1±1.5
82.87
26.41
0.09
0.04
3.84
46.1±0.7
84.97
4.86
0.10
0.04
6.29
20.9±0.6
166.08
76.81
0.23
0.08
IHSS procedure
5.84
27.0±0.8
184.19
140.05
0.14
0.07
1099 dispense copper titrant. The dissolved oxygen was previously removed by bubbling ultrapure nitrogen (99.99%) through the solution for 15 min and a nitrogen atmosphere was maintained above the solution during the experiments [12]. Almost all experiments were performed at 25.0±0.1 °C and ionic strength adjustment of 0.1 mol L–1 NaNO3. Cu(II) binding capacity was measured by incremental addition of 25–125 µL of standard Cu(II) solution (9.45×10–3 mol L–1) at pH 5.0 to obtain a total Cu(II) concentration in the range 0.0–1.5×10–4 mol L–1. The pH was maintained constant by the addition of standard KOH to neutralize protons released in the coordination reaction and those added via acidic Cu(II) solution [11, 13]. Determination of metals Al, Fe, Ni, and Cd determinations were carried out by means of atomic absorption spectrometry (flame AAS) using a Shimadzu model AA-6800 with flame model AA-6800F, according to recommendations of the manufacturer. Synthetic standard solutions were used for calibration of the flame AAS.
Fig. 1 Complexation capacity curves for aquatic humic substances by copper(II) using ultrafiltration technique. Conditions: samples IR-XAD 8, RN-XAD 8, and IS-XAD 8; concentration 100 mg L–1; pH 5.0; I = 0.10 mol L–1 NaNO3
TOC determination The determination of total organic carbon (TOC) contained in the aquatic HS concentrate was carried out by catalytic combustion in an oxygen stream and subsequent IR detection by an Analyser Schimadzu TOC 5000-A with detection limit (3 s) of 0.1 mg L–1 TOC.
Results and discussion The complexation between dissolved organic matter and metal species in solution can be represented in the simplest fashion by the following equilibrium expression: M+L ML
K =
[ML] [M][L]
in which charges have been omitted for simplicity. For EIS, the free metal concentration [M] is obtained by potential measurements according to the Nernst equation (E = E0 + plog[M]), while for the UF-TF technique [M] is determined in the UF filtrate assuming that these metal ions completely penetrate (i.e., retention coefficient R=0) the membrane whereas NOM-bound metal species are completely retained (R = 1) [14, 15]. Then, based on the mass balance equation ([ML] = [M]t–[M]), it is possible to estimate the conditional stability constant (K). Figure 1 shows the typical curves of loaded and free Cu(II) concentrations for AHS. The complexation capacity curves for both methods were obtained by plotting free metal (mmol L–1) versus total added metal (mmol L–1). The curve has a slope change in the final portion. The CC was obtained from the tangent intersections of the two linear graph sections, which corresponds to the appearance of the free metal ion [4, 16]. At least three titrations were performed for each sample. In order to estimate the conditional stability constant (K) of the AOM–Cu, AHS–Cu, and SHS–Cu complexes, the experimental data was handled using the Scatchard method [17] as illustrated in Figs. 2 and 3. This method is a linearization for the 1:1 complex formation model but it can provide information about two different types of coordination sites. The mean values of CC, K1, and K2, ob-
Fig. 2 Typical results from the Scatchard plot for ultrafiltration technique at an RNS-1 concentration of 100 mg L–1, pH 5.0, and 25 °C
Fig. 3 Typical results from the Scatchard plot for ultrafiltration technique at an IR-XAD 8 concentration of 100 mg L–1, pH 5.0, and 25 °C
1100 Table 2 Complexation capacity (mmol Cu(II) g–1 TOC) and conditional stability constant (K) of natural aquatic organic matter, aquatic humic substances, and soil humic substancesa Samples
IS IS-XAD 8 IR IR-XAD 8 RN-XAD 8 RNS-1 RNS-2
Ion-selective electrode (ISE)
Tangential-flow ultrafiltration (TF-UF)
CCb
logK1c
logK2c
CCb
logK1c
logK2c
1.09±0.20 1.12±0.18 1.19±0.02 1.21±0.04 1.27±0.04 1.50±0.30 1.44±0.25
3.05 3.80 2.42 3.45 4.22 4.37 4.13
1.50 1.92 1.17 1.85 2.15 2.26 2.53
1.25±0.07 1.23±0.11 1.18±0.05 1.12±0.05 1.44±0.04 1.99±0.18 1.52±0.24
3.13 4.10 2.30 4.20 4.21 4.35 3.00
1.59 2.13 1.79 1.36 1.38 2.60 1.49
aConditions:
25 °C, pH 5.0, I = 0.1 mol L–1 NaNO3 by intersection of the tangents of the two linear sections of the graph [Cu]free = f([Cu]total) for three replicates [16]
cObtained
bObtained
[16]
tained by both TF-UF and ISE methods are presented in Table 2. Based on TOC results (Table 1) and CC values (Table 2) it was not possible to observe good correlation in the 21–42 mg L–1 TOC range, such that R=0.725 and R=0.482 for ISE and TF-UF, respectively. These results are similar to those obtained by Pardo et al. [18] and Silva [19], who studied natural aquatic organic matter. However, a discrepancy was observed with the correlations obtained by Soares and Vasconcelos [17], who studied aquatic fulvic acids. These different conclusions are probably associated with the characteristics of the organic matter (NOM and fractionated organic matter in fulvic acids). The results shown in Table 2 indicate that CC, K1, and K2 values obtained by the TF-UF method are slightly more highly binding than those compared to the ISE. The discrepancy between the values obtained has been attributed to the intrinsic problems of determining free metal ion concentration. For the ISE, the magnitude of logK1 depends on the availability of data at low values of [Cu]bound. Then, many values would appear to be low because of the failure of the ISE to measure the free form of the metal ion at low concentrations [20]. Since the sensitivity approach used to determine free metal concentration is important in studies of metal complexation, we also observed that soil humic substances have higher CC values in relation to aquatic organic matter and aquatic humic substances. Comparing the results between CC values in the natural organic matter and aquatic humic substances, it is possible to verify similarities, indicating that the extraction procedure does not change the original complexation characteristic of the samples.
tages of this TF-UF technique compared with conventional UF using closed stirred cells for the assessment of complexation parameters are that it: (i) is fast with “open” samples, (ii) is simple to scale-up, (iii) has in situ capability, and (iv) presents reduction of membrane clogging and fouling due to tangential flow.
Conclusions The analytical procedure based on ultrafiltration is a good alternative to determine the complexation parameters in organic material from aquatic and soil systems. This approach presents additional advantages in relation to the traditional ion-selective electrode due to better sensitivity, applicability for multi-element capability, and its possible use under natural conditions. Moreover, the main advan-
by the Scatchard method [Cu]bound/[Cu]free = f([Cu]bound)
Acknowledgements The financial support of this work by CAPES, FAPESP, PROPP, FUNDUNESP, and CNPq is gratefully acknowledged.
References 1. Buffle J, Deladoey P, Greter FL, Haerdi W (1980) Anal Chim Acta 116:255–274 2. Buffle J (1988) Complexation reactions in aquatic systems an analytical approach. Ellis Horwood, New York, pp 304–380 3. Nifanteva TI, Shkinev VM, Spivakov B Ya, Burba P (1999) Talanta 48:257–267 4. Tuschall JR Jr, Brezonik PL (1983) Anal Chim Acta 149: 47–58 5. Thurman EM (1986) Organic geochemistry of natural waters. Martinus Nijhoff, Dordrecht, pp 365–423 6. Staub C, Buffle J, Haerdi W (1984) Anal Chem 56:2843–2849 7. Aiken GR, Mcknight DM, Wershaw RL (1985) (eds) Humic substances in soil, sediment, and water. Geochemistry, isolation, and characterization. Wiley, New York, pp 329–385 8. Malcolm RL (1990) Anal Chim Acta 232:19–30 9. Rosa AH, Vicente AA, Rocha JC, Trevisan HC (2000) Fresenius J Anal Chem 368:730–733 10. Burba P, Van den Bergh J, Klockow D (2001) Fresenius J Anal Chem 371:660–669 11. Town RM, Powell HKJ (1993) Anal Chim Acta 279:221–233 12. Abate G, Masini JC (1999) Química Nova 22(5):661–665 13. Antonelli ML, Calace N, Centioli D, Petronio BM, Pietroletti M (2001) Anal Lett 34:989–1002 14. Van den Bergh J, Jakubowski B, Burba P (2001) Talanta 55: 587–593 15. Rocha JC, Rosa AH, Van den Bergh J, Burba P (2002) J Environ Monit 4:799–802 16. Morrison GMP, Florense TM (1989) Elektroanalysis 1:107– 112 17. Soares HMVM, Vasconcelos MTSD (1994) Anal Chim Acta 293:261–270 18. Pardo R, Barrado E, Veja M, Deban L, Tascón ML (1990) Wat Res 28:2139–2146 19. Silva ST (1996) Masters dissertation. Instituto de Química, Unicamp, São Paulo 20. Stevenson FJ (1994) Humus chemistry. Wiley, New York, pp 405–428
