Aquatic humus from an unpolluted Brazilian dark-brown stream: general characterization and size fractionation of bound heavy metals

Abstract

The main pool of dissolved organic carbon in tropical aquatic environments, notably in dark-coloured streams, is concentrated in humic substances (HS). Aquatic HS are large organic molecules formed by micro-biotic degradation of biopolymers and polymerization of smaller organic molecules. From an environmental point of view, the study of metal–humic interactions is often aimed at predicting the effect of aquatic HS on the bioavailability of heavy metal ions in the environment. In the present work the aquatic humic substances (HS) isolated from a dark-brown stream (located in an environmental protection area near Cubatão city in São Paulo-State, Brazil) by means of the collector XAD-8 were investigated. FTIR studies showed that the carboxylic carbons are probably the most important binding sites for Hg(II) ions within humic molecules. ¹³C-NMR and ¹H-NMR studies of aquatic HS showed the presence of constituents with a high degree of aromaticity (40% of carbons) and small substitution. A special five-stage tangential-flow ultrafiltration device (UF) was used for size fractionation of the aquatic HS under study and for their metal species in the molecular size range 1–100 kDa (six fractions). The fractionation patterns showed that metal traces remaining in aquatic HS after their XAD-8 isolation have different distributions. Generally, the major percentage of traces of Mn, Cd and Ni (determined by ICP-AES) was preferably complexed by molecules with relatively high molecular size. Cu was bound by fractions with low molecular size and Co showed no preferential binding site in the various humic fractions. Moreover, the species formed between aquatic HS and Hg(II), prepared by spiking (determined by CVAAS), appeared to be concentrated in the relatively high molecular size fraction F₁ (>100 kDa).

---

Introduction

The Parque Estadual da Serra do Mar is an environmental protection area located in Baixada Santista near Cubatão city in São Paulo-State, Brazil. The main industrial activities of this region are refineries, fertilizers and metal smelters, which have provoked serious environmental problems. The particular characteristic of this region is the existence of various dark-brown streams that come from the mountain and flow towards the Atlantic Ocean. A significant fraction of the dissolved organic carbon (DOC) in the surface and ground waters consists of aquatic humic substances (aquatic HS). These naturally occurring materials are formed by degradation of plants and/or by polymerization of smaller organic molecules present in the environment and studies of aquatic HS are limited due to the fact that most of the chemical structures involved are still unidentified.^1,2

Aquatic HS contain various types of phenolic and carboxylic functional (hydrophilic) groups as well as aromatic and aliphatic fragments which give hydrophobic properties to these substances. Owing to hydrophobic interactions, aquatic HS accumulated at the solid/water clay interface are adsorbed by the clay particles, modifying the chemical properties of their surfaces, especially the affinities for metal ions.³ In the studied region, the extensive weathering and leaching cycles during the major part of the year leads to the decrease of cations present in the soil. Thus, the metal ions are probably leached from mountains (Serra do Mar) to aquatic systems and bound to aquatic HS. From an environmental point of view, studies of metal–humic interactions are often aimed at predicting the effect of HS on the bioavailability of metal ions in the aquatic systems.^4,5 The characterization of functional groups by spectroscopic methods,⁶ studies about the nature of interactions between organic chemicals and aquatic HS,⁷ as well as a better knowledge of metal–aquatic HS interactions in different molecular size fractions are also important.^8–11

In the present work, the different functional groups of aquatic HS isolated from a Brazilian dark-brown stream were investigated by means of FTIR, ¹³C-NMR and ¹H-NMR and UV/VIS spectroscopy. Moreover, a sequential-stage ultrafiltration procedure⁸ served for the size-classification of the isolated aquatic HS. The distribution of metals contained in the obtained aquatic HS fractions was assessed by ICP-AES. Information about the degree of aromaticity of the different fractions could be obtained by UV/VIS spectrometry.

Experimental

Chemicals and reagents

All reagents used were high-purity grade unless otherwise stated. Diluted acid and base solutions necessary for the aquatic HS isolation were prepared by convenient dilution of 30% hydrochloric acid (Suprapur, Merck AG, Darmstadt, Germany) or sodium hydroxide-monohydrate (Suprapur, Merck AG) dissolved in high-purity water (Milli-Q system, Millipore, Bedford, USA). The XAD-8 adsorbent (Serva Feinbiochemica, Heidelberg, Germany), used for the isolation of aquatic HS, was purified before use by successive soaking with 0.5 mol l⁻¹ HCl, 0.5 mol l⁻¹ NaOH and methanol pro analysi (24 h each).

HS isolation by XAD-8 resin

The aquatic HS under study were isolated from a sample collected from a tributary stream of Rio Itapanhaú within the Parque Estadual da Serra do Mar. This is an environmental protection area located in the seaboard, basin 51 of the 5th hydrographic zone of São Paulo State, Brazil (Fig. 1). For this purpose, 50 l of surface water were field filtered through 0.45 µm cellulose-based membranes and acidified with concentrated HCl solution to pH 2.0. Afterwards, the aquatic HS from the acidified sample were isolated on the XAD-8 collector¹² in the conventional way following the recommendations of Malcolm.¹³ After elution with 0.1 mol l⁻¹ NaOH solution, the 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. The determination of DOC contained in the aquatic HS concentrate was carried out by catalytic combustion in an oxygen stream and subsequent IR detection by Analyser Schimadzu TOC 2000 (Duisburg, Germany).¹⁴

Fig. 1 Site for water sampling (Rio Itapanhaú).

Complexing capacity [Cu(II)] of HS

The complexing capacity of the aquatic HS under study was determined by a copper(II) selective electrode (WTW Cu 500, Madison, WI, USA). For this purpose, 2.0 mg of aquatic HS dissolved in 100 ml of 0.1 mol l⁻¹ NaNO₃ solution (pH 5.0) were successively loaded with Cu(II) ions and the increase of the electrode potential referring to the added Cu(II) was recorded. The assessment of the Cu(II) complexation capacity (CC) was carried out according to Buffle¹⁵ and Soares and Vasconcelos¹⁶ yielding a CC of 3.4 mmol Cu(II) per g DOC of aquatic HS.

Hg(II) spiking of aquatic HS and FTIR spectroscopy

The species formed between unfractioned aquatic HS and mercury(II) ions [HS–Hg(II)] was prepared in the following way: 20.0 µg of Hg(II) were added to 10 ml of aquatic HS solution (1.0 mg ml⁻¹) buffered at pH 5.0 and mechanically stirred for 48 h. The Fourier-transform infrared spectra of unfractioned aquatic HS and their mercury(II) complex were recorded by using a Nicolet 730 SX-FT spectrometer (Madison, WI, USA). The pellets were prepared with 8 mg of aquatic HS per 100 mg of KBr.

NMR spectrometry

The ¹³C-NMR spectrum of 250 mg aquatic HS dissolved in 3.0 ml D₂O (pH 8.5), referred to acetonitrile, was measured at 100 MHz with a JEOL GX 400 NMR spectrometer (Peabody, USA) according to the following experimental conditions: acquisition time, 0.34 s; total time, 88 h; sweep width, 48 076 Hz; and resolution, 16k, where k is an experimental parameter that represents the best level for obtaining the NMR spectrum. The ¹H-NMR spectrum of the D₂O dissolved aquatic HS sample, referred to acetonitrile, was measured at 400 MHz using a JEOL GX 400 NMR spectrometer according to the following experimental conditions: acquisition time, 0.58 s; total time, 37 min; sweep width, 7002 Hz; number of scans, 1400; and resolution, 4k.

Fractionation of the aquatic HS by multistage ultrafiltration (UF)

The aquatic HS under study were fractionated on-line by means of a special tangential-flow multistage ultrafiltration (UF) unit developed by Burba and co-workers.^8,9 It consists of a cascade of up to 5 UF stages made of high-purity acrylic polymer attached by two bolts and nuts. The UF stages are equipped with appropriate 25 mm diameter UF membranes (Pall-Filtron Omega, Hamburg, Germany), leading to separation of the following molecular-size fractions: F₁, >100; F₂, 50–100; F₃, 10–50; F₄, 5–10; F₅, 1–5 and F₆ < 1 kDa. The tangential-flow UF processing was performed by means of a five-channel peristaltic pump. Accordingly, the aquatic HS solution (10 ml, 1.0 mg ml⁻¹ HS, pH 6.0) was pumped (initial pressure, 2.5 bar) through the cascade of membranes leading to a flow rate of 1.5–2 ml h⁻¹ (tangential flow, 2–3 ml min⁻¹).

UV/VIS and ICP-AES

The UV/VIS spectra of the ultrafiltrated fractions were registered in the 200–465 nm range with a two beam Varian Cary 1/3 spectrometer (Palo, Alto, USA). The metals bound to the fractions (F₁ to F₆) were determined by ICP-AES, using a Thermo Jarrel Ash-CID-DUO spectrometer (Franklin, USA), according to the experimental conditions given in Table 1.

Table 1 Operating parameters for the ICP-AES determinations

Incident power	1.5 kW
Plasma air flow-rate	15 l min^−1
Nebulizer air flow-rate	1.2 l min^−1
Auxiliary air flow-rate	0.5 l min^−1
Sample flow-rate	2.4 ml min^−1
Analytical lines	Cu 327.396 {079} (I); Co 228.616 {114} (II); Ni 221.647 {117} (II); Cd 228.802 {113} (I); Mn 257.610 {100} (II) nm

---

Determination of Hg(II)

Hg(II) determinations in solutions were carried out by means of conventional cold vapour atomic absorption spectrometry (CVAAS), using the sodium borohydride method.¹⁷ The manufacturer's recommendations were taken into account for the Perkin Elmer Analyst 300 spectrometer (Norwalk, USA) used. The detection limit (3s) was 0.1 ng ml⁻¹ Hg(II) and the relative standard deviation was 5% (2 ng ml⁻¹ Hg(II), n = 10).

Results and discussion

Some information about the studied aquatic HS is summarized in Tables 2 and 3.

Table 2 Complementary information related to the humic substances from Rio Itapanhaú (May, 1999)

Origin	Rio Itapanhaú
pH	5.0
Conductivity	58 µS cm^−1
Temperature	25 °C
DOC	4.5 mg ml^−1
Original aquatic humic substance	9.0 mg ml^−1
Complexation capacity	3.4 mmol Cu(II) per g DOC

---

Table 3 Concentration of the trace metals in water and in the aquatic humic substances isolated from Rio Itapanhaú (n = 5)

	Concentration
Metals	Water^a/µg l^−1	Isolated HS^b/mg l^−1
^a Total metals in the 0.45 µm membrane-filtered original sample. ^bAfter isolation by the XAD-8 resin procedure (concentration 250×). ^cDetection limit.
Cu	79.9 ± 0.59	20.2 ± 1.61
Co	≤0.5^c	0.69 ± 0.04
Ni	11.6 ± 0.08	3.12 ± 0.24
Cd	≤0.25^c	5.53 ± 0.42
Mn	18.7 ± 0.11	4.71 ± 0.16
Hg	≤0.5^c	0.04 ± 0.0002

---

FTIR studies

A variety of spectroscopic methods have been employed to characterize the functional groups and behaviour of humic substances and one of the most widely used methods is Fourier-transform infrared spectroscopy.^18,19 FTIR absorption spectra in the region between 400 cm⁻¹ and 4000 cm⁻¹ of unfractioned aquatic HS are shown in Fig. 2. The strong and broad absorption band at 3400 cm⁻¹ arises from H-bonded OH groups, including those of COOH. The contribution of N–H groups to absorption in this region is insignificant. The absorption band in the 2900 cm⁻¹ region (aliphatic CH stretching) is relatively weak, indicating the absence of large amounts of CH₂ and CH₃ groups. The strong absorption band at 1721 cm⁻¹ is due to C=O stretching vibrations of carboxyl and/or ketonic carbonyl groups. The band at 1624 cm⁻¹ which arises together with one at 1721 cm⁻¹ is attributed to C=C vibrations in aromatic rings. The weak band at 1380 cm⁻¹ is probably due to C–H deformation of aliphatic groups. The broad band near 1200 cm⁻¹ is due to C–O stretching vibrations and/or OH deformation of COOH. When aquatic HS were treated with Hg(II) ions there was a decrease in the intensity of the band at 1720 cm⁻¹ and the extinction of the band at 1200 cm^-1. The results indicate that the majority of hydrogen ions on the carboxyl groups may be exchanged for Hg(II) ions. The carboxylate groups are primarily responsible for binding metals under natural conditions.²⁰

Fig. 2 The FTIR spectra of the unfractioned aquatic HS from Itapanhaú River and aquatic HS–Hg(II) complexes.

NMR studies

Nuclear magnetic resonance spectroscopy is another technique suitable for functional groups characterization of humic substances.²¹ In general, both ¹H NMR and ¹³C NMR are used to compare the difference in concentration of the main functional groups among samples of humic substances.²² The various peaks in the spectrum can be assigned to specific carbon functional groups, although some overlaps of resonance assignments, which were observed, may be due to the large variety of different groups present in complex material, like humic substances. Thus, direct measurements of the contents of carboxyl, aromatic and aliphatic carbons can be made by integration of specific regions, taking into account any possible overlap of the regions.²³

Fig. 3 shows the ¹³C-NMR spectrum of the unfractioned aquatic HS. Chemical shifts and their assignments are presented in Table 4. The aliphatic region 0–50 ppm presents a broad unresolved band, which indicates that the aliphatic components are probably of short chain length or are highly branched. The integration of the signals in this region shows that 17% of the carbon present has bonding of the kind sp³. The oxygenated aliphatic carbon region, which extends from ≈50 to 100 ppm, exhibits a relatively small amount of signal. The 55 ppm peak suggests the presence of methoxy carbon (–OCH₃) while the broader, lower field resonance group from 60 to 80 ppm is representative of a variety of alcohol, ether and ester carbons. The integration within this region yielded only 4% of the carbon. The aromatic and olefinic region (100–165 ppm) contains a significant signal (44%). The region from 165 to 190 ppm is due to carboxyl carbon. The region of amide and ester carbons accounts for 32% of these carbons present in the aquatic HS, but the region of ketone and aldehyde carbons is low in the peaks, only 3%.

Fig. 3 ¹³C-NMR spectrum (100 MHz, D₂O) of the unfractioned aquatic HS from Rio Itapanhaú: chemical shifts (ppm) referred to acetonitrile. (a) = blanks of acrylesther dissolved from XAD-8.

Table 4 Peak assignments for the ¹³C NMR spectrum of the studied aquatic HS from Rio Itapanhaú and their integrated areas (%)

ppm	Assignments	Integration (%)
65–0	Paraffin	17
100–65	Hydroxyl, Ether	4
165–100	Olefinic, Aromatic	44
190–165	Carboxyl	32
230–190	Ketone	3

---

Some of the first NMR spectroscopic applications in humic substances research involved the use of ¹H NMR for the examination of soluble humic material. The information obtained by using this technique has been useful in characterizing the aromatic and aliphatic structures of these complex materials.²¹

Fig. 4 shows the ¹H-NMR spectrum of the unfractioned aquatic HS. Chemical shifts and their assignments are presented in Table 5. The signals in the 0–1.6 ppm region are assigned to the methyl proton of hydrocarbon, methylene proton and aliphatic proton, β-substituted. The region from 1.6 to 3.0 ppm is relative to α-monosubstituted aliphatic proton. The peaks between 3.3–4.5 ppm are typical of protons on carbons attached to O or N heteroatoms. It seems from the signals around 3.6 ppm that a large proportion (37%) of the aliphatic protons is bound to carbon in the α position to oxygen. The predominance of such aliphatic oxygen-bound carbon is assigned to a major proportion of sugar-like components or other polyhydroxy or polyether materials. This is supported by the ¹³C-NMR spectrum where the signals between 60–80 ppm are characteristic of oxygenated carbon, suggesting the presence of sugars or polyhydroxy material which also contain COOH, aromatic and alkyl groups. The very broad peak which appears at 5.0 ppm can be attributed to the traces of H₂O in the D₂O. The peaks from 5.0 to 9.0 ppm are assigned to olefinic and aromatic protons. These protons contribute 22% in the integration of signals.

Fig. 4 ¹H-NMR spectrum (400 MHz, D₂O) of the unfractioned aquatic HS from Rio Itapanhaú: chemical shifts (ppm) referred to acetonitrile. (a) XAD-8 acrylesther peak; (b) water; (c) acetonitrile.

Table 5 Peak assignments for the ¹H NMR spectrum of the studied aquatic humic HS from Rio Itapanhaú and their integrated areas (%)

ppm	Assignments	Integration (%)
0–1.6	C–CHn (n = 2;3)	17
1.6–3.0	(O)–CO–CHn	24
	Aliphatic protons attached C atom α to a benzene ring
3.3-4.5	O–CHn	37
	N–, CHn (n = 1;2)
5.5–9.0	Olefinic and aromatic	22

---

Ultrafiltration studies

Fig. 5 shows the molecular-size distribution of the studied aquatic HS in the ultrafiltration fractions from F₁ to F₆ and their original acid-inert metal trace contents. The loss in the recovery, about 5%, might be attributed to adsorption onto the inner surface of the UF device.^8,24 The fractionation of the aquatic HS isolated from Rio Itapanhaú indicates that 65% of the molecules are concentrated in the fraction with a molecular size around 50–100 kDa (F₂), while F₅ (1–5 kDa) contains a minor quantity (3.3%). The remaining molecules are distributed in other fractions. Aster et al.⁹ reported that the molecular size distribution of the original sample of the Venner Moor bog was characterized by remarkably big fractions (34%) of molecular size >100 kDa. According to these authors, the isolation of the aquatic HS by the XAD-8 adsorbent causes a systematic shift of all fractions compared to the original sample, presumably due to extreme pH changes applied during the XAD-8 isolation (sorption at pH 2, elution at pH 13). It is worth noting that the distribution patterns of different samples containing aquatic HS can be compared only if the separation is carried out under the same conditions. The findings, accordingly,²⁵ reveal a strong influence of pH values and the sample origin on the distribution patterns of metals. The fractionation pattern of aquatic HS demonstrates that acid-inert metal traces still retained in this HS after its XAD-8 isolation from the acidified water sample (pH 2.0) have different distributions between various fractions, as shown in Fig. 5.

Fig. 5 Molecular distribution pattern of the aquatic HS from Rio Itapanhaú attained after a five-stage on line ultrafiltration (F₁, >100 kDa; F₂, 50–100 kDa; F₃, 10–50 kDa; F₄, 5–10 kDa; F₅, 1–5 kDa; F₆, <1 kDa) and distribution of Mn (□), Co (△), Ni (•), Cu (■) and Cd (○) originally loaded. For these studies 20.0 µg Hg(II) (▲) were added to 10 ml of aquatic HS before the ultrafiltration procedure. Conditions: 1.0 mg ml⁻¹ aquatic HS; pH 6.0.

Beyond the classification of HS according to their size, Fig. 5 shows the distribution of the original acid-inert HS metal species as a function of their molecular size. None of the metals studied exhibits a fractionation pattern comparable to the distribution of HS molecules which are preferably found in the high molecular-size fraction (about 65% in F₂). The Cd, Mn and Ni distributions appear to be similar. The major part of these metals is concentrated in the fractions F₂ and F₃ whereas Co presents a distinctive distribution pattern. This element does not occupy specific sites in the aquatic HS fractions, unlike the case reported by Shkinev et al.²⁵ in which 60% of total Co content in a river water sample remained in the high molecular-size fraction. The Cu ion (Fig. 5) is enriched in the low molecular-size fractions (about 30% in both F₅ and F₆). Rocha et al.¹¹ using aquatic HS isolated from Rio Negro Amazon-State Brazil waters by sorbent XAD-8, reported that the major amount of Cu ion was complexed with large molecular-size fractions. However, the same metal ions exhibit different finger prints for samples of different origins.¹⁰

Mercury concentration in the original sample was below the detection limit of CVAAS. Therefore, for investigation of the distribution pattern, this environmentally relevant metal ion was added (2 µg ml⁻¹). Fig. 5 shows that 50% of HS–Hg(II) is in the fraction with the large molecular-size (F₁ > 100 kDa) with a continuous decrease in the other fractions. This suggests that, for the aquatic HS studied, Hg(II) ions occupy specific complexation sites in this fraction.

UV spectrometric studies

Traina et al.²⁶ correlated the percentage of aromatic carbon determined by NMR with UV-absorbance measurements. It was suggested that the UV-absorbance data can be used to provide a rapid and quantitative estimate of the aromatic nature of the dissolved HS. Peuravuori and Pihlaja²⁷ found that the correlation between the E₂/E₃ ratio (absorbances at 250 and 365 nm) and the molar absorptivity (ε) at 280 nm was quite moderate (r² = 0.81). The authors suggested that the relationship between the aromaticity and the E₂/E₃ ratio for humic materials could be obtained according to: This implies that when the quotient E₂/E₃ increases, the aromaticity and the molecular-size of aquatic HS decreases.

Using eqn. (1), Fig. 6 shows that the fractions with larger molecular size (F₁ > 100 and F₂ = 50–100 kDa) contain, as a first approximation, a majority of aromatic structures and the presence of gradually lower proportions of aromatic structures in the fractions with smaller molecular size. This behaviour can be associated with a continuous decrease of the molecular size from fractions F₃ to F₆ and the presence of relatively larger proportions of aliphatic structures. Recent studies¹¹ with other Brazilian aquatic HS reports a similar aromaticity distribution in the different fractions.

Fig. 6 Molecular distribution pattern of the aquatic HS from the Rio Itapanhaú attained by five-stage ultrafiltration (F₁, >100 kDa; F₂, 50–100 kDa; F₃, 10–50 kDa; F₄, 5–10 kDa; F₅, 1–5 kDa; F₆, <1 kDa) and aromaticity. Conditions: 1.0 mg ml⁻¹ aquatic HS; pH 6.0.

Conclusions

In general, HS in aquatic systems and in soils exhibit considerable complexing capacities for metal ions due to their great variety of binding sites. Thus HS are important natural “buffers" for heavy metals in the environment. Transport, deposition and biovailability of heavy metals in natural waters strongly depend on stability of the metal–HS species formed.

The tropical weather and regular rain periods in the studied area lead to a great leaching of metal ions from soil to aquatic systems and these metals are probably complexed by the aquatic humic substances. In this study, the FTIR spectra showed that the carboxylic carbons are probably the most important binding sites for Hg(II) ions within humic molecules. The molecular spectroscopic studies of unfractioned aquatic HS showed the presence of constituents with a high degree of aromaticity (40% of carbons) and small substitution. The carboxyl content was around 32% of total carbons present. UV/VIS absorption ratios E₂/E₃ suggest a high aromaticity for the F₁ and F₂ fractions, which decreases towards the lowest molecular-size fractions. The findings from both ¹³C NMR and UV/VIS studies indicated that aquatic HS are materials containing highly aromatic structures.

The fractionation patterns indicate that metal traces originally complexed by aquatic HS have different distributions. Generally, the major percentage of traces of Mn, Cd and Ni was preferably complexed by molecules with relatively high molecular size. However, Cu was bound by fractions with low molecular size and Co showed no preferential binding site in the various humic fractions. On the other hand, the Hg(II) ion appeared to be concentrated in the relatively high molecular size fraction F₁. In general, the investigated sample showed a considerable diversity with regard to the rank of binding preference for metal ions, which apparently results from different binding strengths for the metal ions studied. Moreover, it can be suggested that the complexation behaviour of the molecules in humic substances towards trace metals in aquatic environments is considerably affected by their molecular size. From this point of view, this multimethodological study of aquatic HS isolated from a Brazilian dark-brown stream is an additional contribution to characterize the potential of environmental chelators (e.g., humic substances, nitritotriacetic acid) contained in organic-rich water samples.

Acknowledgements

This work was financially supported by FAPESP-CNPq-CAPES (Brazil) and DAAD (Germany). The authors are indebted to Dr Peter Burba, Helmut Herzog and Miss Brit Aster from ISAS (Germany).

References

1. P. MacCarthy I. H. Suffet, in Aquatic humic substances. Influence on fate and treatment of pollutants, ed. P. MacCarthy and I. H. Suffet, ACS, Washington, 1989, vol. 219, p. 384..
2. J. A. Leenheer, in Environmental chemistry of lakes and reservoirs, ed. L. A. Baker, ACS, Washington, 1994, vol. 237, pp. 195–221..
3. R. L. Wershaw, J. Contam. Hydrol., 1986, 1, 29.
4. P. Burba, J. C. Rocha and D. Klockow, Fresenius' J. Anal. Chem., 1994, 349, 800.
5. J. C. Rocha, I. A. S. Toscano and P. Burba, Talanta, 1997, 44, 69.
6. R. Swift, in Methods of Soil Analysis. Part 3. Chemicals Methods. Soil Science of America and American Society of Agronomy, 1996, 677 S. SSSA Book Series no. 5, pp. 1011–1069..
7. N. Senesi, in Organic substances in soil and water: Natural constituents and their influences on contaminant behaviour, ed. A. J. Beck, K. C. Jones, M. H. B. Hayes and U. Mingelgrin, Royal Society of Chemistry, Cambridge, 1993, pp. 73–101..
8. P. Burba, V. Shkinev and B. Y. Spivakov, Fresenius' J. Anal. Chem., 1995, 351, 74.
9. B. Aster, P. Burba and J. A. C. Broekaert, Fresenius' J. Anal. Chem., 1996, 354, 722.
10. B. Aster, A. Bohlen and P. Burba, Spectrochim. Acta, Part B, 1997, 52, 1009.
11. J. C. Rocha, É. Sargentini, Jr, I. A. S. Toscano, A. H. Rosa and P. Burba, J. Braz. Chem. Soc., 1999, 10, 169.
12. G. R. Aiken, in Humic substances in soil, sediment and water. Geochemistry, isolation and characterization, ed. G. R. Aiken, D. M. MacKnight, R. L. Wershaw and P. MacCarthy, John Wiley & Sons, New York, 1985, pp. 363–385..
13. R. L. Malcolm, in Humic Substances in the aquatic and terrestrial environment, ed. B. Allard, H. Boren and A. Grimvall, Springer-Verlag, Berlin, 1989, pp. 9–19..
14. P. J. Wangersky, Mar. Chem., 1993, 41, 61.
15. E. H. Buffle, in Complexation reactions in aquatic systems—an analytical approach, Ellis Horwood, Chichester, UK, 1988, p. 427..
16. H. M. V. M. Soares and M. T. S. D. Vasconcelos, Anal. Chim. Acta, 1994, 293, 261.
17. E. Jackwerth, P. G. Willmer, R. Höhn and H. Berndt, At. Absorpt. Newsl., 1979, 18, 66.
18. L. Campanella, B. M. Petronio, C. Braguglia, R. Cini and D. N. Innocenti, Int. J. Environ. Anal. Chem., 1994, 60, 49.
19. Y-Ping Chin, G. Aiken and E. O'Loughlin, Environ. Sci. Technol., 1994, 28, 1853.
20. J. S. Gaffney N. A. Marley S. B. Clark, in Humic and fulvic acids. Isolation, structure, and environmental role, ed. J. S. Gaffney, N. A. Marley and S. B. Clark, ACS Symposium Series, Washington, 1996, vol. 651, pp. 2–17..
21. M. A. Wilson, in Humic substances II, ed. M. H. B. Hayes, P. MacCarthy, R. L. Malcolm and R. S. Swift, John Wiley & Sons, Chichester, 1989, pp. 310–335..
22. R. L. Wershaw, D. J. Pinckney, E. C. Llaguno and V. Vicente-Beckett, Anal. Chim. Acta, 1990, 232, 31.
23. J. M. Bortiatynski P. G. Hatcher H. Knicker, in Humic and fulvic acids. Isolation, structure, and environmental role, ed. J. S. Gaffney, N. A. Marley and S. B. Clark, ACS Symposium Series, 1996, vol. 651, pp. 57–77..
24. J. C. Rocha L. F. Zara A. H. Rosa E. Sargentini, Jr. P. Burba, Quim. Nova, in the press..
25. V. M. Shkinev, O. M. Fedorova, B. Ya. Spivakov, J. Mattusch, R. Wennrich and M. Lohse, Anal. Chim. Acta, 1996, 327, 167.
26. S. J. Traina, J. Novak and N. E. Sneck, J. Environ. Qual., 1990, 19, 151.
27. J. Peuravuori and K. Pihlaja, Anal. Chim. Acta, 1997, 337, 133.

---

Footnote

† Dedicated to the 60th birthday of Dr Peter Burba.

---