ANALYST, FEBRUARY 1992, VOL. 117 15 1 High-performance Liquid Chromatographic Study of Nickel Complexation With Humic and Fulvic Acids in an Environmental Water Peter Warwick and Tony Hall Department of Chemistry, L ough bo roug h University of Tech nolog y, Loug h borough, L eicestershire LE77 3TU, UK A high-performance liquid chromatographic method was developed to determine cation-exchange capacities and conditional association constants for metal interactions with humic and fulvic materials present in environmental waters. The method does not require prior extraction of the humic and/or fulvic compounds. A salt-gradient is used, which exploits the size-exclusion and adsorption properties of a coated porous silica stationary phase, in order t o separate and permit the measurement of the free and complexed metal concentrations.The results are subjected t o a weak and strong binding site interpretation. Keywords : Hum ic acid; fulvic acid; nickel co mplexa tion ; high -performance size-exclusion chromatograph y; water There is increasing interest in the fate and behaviour of trace metals in the environment and humic and fulvic acids, present in natural waters, play an important role. Humic and fulvic acids bind (complex) with metal pollutants and thereby affect such diverse phenomena as transport mechanisms, toxicity, bioavailability and the effectiveness of recovery and clean-up procedures.’ The acids generally occur as complex het- erogeneous mixtures of polymeric anions showing local and seasonal variations in composition.Probably every structural analytical technique, from classical elemental and functional group analysis to advanced instrumental analytical methods, including infrared (IR) , nuclear magnetic resonance (NMR) and mass spectrometry, have been applied to these com- pounds but owing to their complexity full structural elucida- tion has not been achieved. However, much information has been gained;2,3 for example, it has been discovered that humic and fulvic acids often contain aromatic backbones carrying a variety of functional groups, e.g., phthalate, salicylate and amine functions. With metals, humic and fulvic acids form anionic complexes, whereas with organic pollutants, mol- ecular association or covalently bound species can be pro- duced. One particular environmental concern is the interac- tion of humic and fulvic acids with radionuclides.Radionu- clides can enter the environment as a result of accidental or controlled releases of effluent. In addition, in the future it may be envisaged that over prolonged time periods eventual ingress of groundwater into planned radioactive waste reposi- tories is likely. Consequently, a full understanding of possible complexation and transport mechanisms is highly desirable. The speciation of a metal ( i . e . , distribution amongst the various possible physico-chemical forms) determines its over- all geochemical and biological behaviour. In order to assess the amount of ‘free’ metal species present, both the inorganic and organic complex speciation must be known, which requires a knowledge of the relevant stability constants.Also with humic and fulvic acids it is desirable to have a measure of the maximum amount of metal uptake that is likely under any given set of environmental conditions. This requires the determination of the maximum cation complexing capacity (C,) of the material. In order to explain the mechanism of metal binding with humic and fulvic acids, several models have been developed. In a review, Falck4 categorizes the models into two main types, either discrete ligand models or continuous distribution models. A survey of the literature suggests that the discrete ligand model is accepted by most workers. The model proposes that major binding sites (L), such as salicylate and phthalate, act as ligands towards the metals (M).The stoichiometry at a ligand site is 1 : 1, but of course the denticity may be higher. Hence the complexation reaction may be simply represented as M + L = M L for which the conditional association value (site binding constant) is given by where the brackets denote concentrations and in particular [L] is the concentration of free binding sites, i.e., not the concentration of humic or fulvic compounds. The number of ligand sites present determines the maxi- mum complexing capacity of the substance, i.e., (3) The mathematical interpretation of experimental data depends, of course, on the model adopted and no model has escaped criticism. A large number of analytical techniques have been applied to the determination of association and capacity values and are described in several texts.”6 However, in general, in order to simplify the chemistry involved investigations have been carried out using extracted humic and fulvic materials.Diethylaminoethyl (DEAE)-cellulose or XAD resins are commonly used as the extraction media.’ The adsorbed organics are eluted from these extractants using alkaline solutions in which they are readily soluble. However, such extraction procedures can be expected to change the proper- ties and structures of the organic compounds. States of aggregation, stereochemistry, inherent metal content and over-all purity are likely to be affected. Measured values of K and C, could well then be different to those applicable to the in situ material. Accordingly, this paper describes a high-perfor- mance liquid chromatographic technique which has been developed primarily to study in situ materials, i.e., direct investigation of the environmental water, without any prior treatment, apart from 0.45 pm filtration, which is employed to remove most of the colloidal clays and micro-organisms which may be present and mild rotary evaporation when preconcen- tration is necessary.The technique involves high-performance size-exclusion chromatography (HPSEC) and by means of152 - i --a .. . 4- c- Injection 0.05 mol I-' - NaCl Gradient ANALYST, FEBRUARY 1992, VOL. 117 A salt-gradient elution exploits cationic absorption, which is normally considered to be a disadvantage of HPSEC. A separation of anionically complexed metal from free metal is achieved, so that the relative amounts in an equilibrium mixture can be measured.During the time of the separation (<lo min), the dissociation of the previously formed complex must be negligible. The method was developed from work previously reported using Sephadex gels,8 but the gel tech- nique was rejected in favour of the higher speed and resolution of the high-performance technique. Accordingly, this investigation was conducted using nickel, a typical divalent transition metal, which is known to form complexes with humic and fulvic materials, and for which the kinetics of association and dissociation are slow , several days being required for the attainment of equilibrium. Nickel complexation has been studied extensively by various workers using extracted humic and fulvic materials.9-11 Nickel43 (t4 = 100 years, P,E,,, = 66 keV) was used to label the nickel mixtures so that the chromatographically separated com- plexed and free nickel could be assayed using liquid scintilla- tion counting.Currently the technique is being modified to study europium complexation using europium-152 (t4 = 13 years, y, 0.122 MeV, 62%) and solid-state counting. Pump Port Experimental Apparatus The HPSEC experiments were carried out using a Philips PU 4000 Series liquid chromatograph fitted with a PU 4100 gradient pumping system, a PU 4021 diode-array detector (DAD), a Rheodyne injection valve with a 100 pl loop and PU 6000 and PU 6003 integration and control software. SynChro- Pak GPC-60 guard (SO x 4.6 mm i.d.) and analytical (250 x 4.6 mm i.d.) size-exclusion columns were used, which con- tained a 5 pm porous silica stationary phase, coated with a glyceryl propyl bonded phase.This material has a stated linear relative molecular mass separation range of 300-20 000 for dextrans and 300-30 000 for proteins and is suitable for humic and fulvic acid fractionation. In exploratory experiments, the eluate emerging from the DAD was passed into a Waters Model 420 fluorescence detector, which was used to help to confirm the identity of the humic and fulvic fractions. In the complexation experiments the eluate emerging from the DAD passed through a Canberra Packard Flo-One/Beta (A140) radioactivity detec- tor, fitted with an 800 pl flow cell. A scintillation cocktail (Ecoscint A; National Diagnostics) was mixed with the column eluate before the flow-through cell.The experimental arrangement is shown schematically in Fig. 1. He . de-gassing . Environmental Water Sample A surface water was taken from moorland in the Derbyshire Peak District, near the village of Moscar. The water was subjected to 0.45 pm filtration, which was commenced 3 h after collection, and this was followed by the rotary evapora- tion of a 1000 ml sample to 250 ml at 30 "C, the resulting sample henceforth being referred to as ~4 moorland water. Rotary evaporation is a mild process and the 4-fold increase in concentration was undertaken to increase the ease of detec- tion of the humic and fulvic material. The rotary evaporation lowered the pH from the original in situ value of 3.8 to 3.5. As the extent of nickel complexation increases with increase in pH, a very small amount of concentrated NaOH solution was added to adjust the pH to 6.3.The estimated concentration of the humic and fulvic acid species from both the TOC (total organic carbon) and ultraviolet (UV) absorption data was about 52 mg 1-1. The results of the analysis of the X4 moorland water are given in Table 1. Guard column Fluorescence detector - Diode t- Waste Fig. 1 HPSEC experimental arrangement a+ Analytical column array detector Table 1 Analyses of working solutions Radioactivity Parameter PH Chloride Bromide Alkalinity as CaC03 Ammonia as N Nitrite as N Calcium Magnesium Sodium Potassium Total hardness as CaC03 Sulfate Phosphate as P Silica Fluoride Total oxidized nitrogen Total organic carbon Total inorganic carbon Boron Molybdenum Uranium Lithium Strontium Iron Manganese Aluminium Vanadium Lead Chromium Copper Nickel Zinc Cadmium Barium Cobalt A Units mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg 1-1 mgl-1 mg 1-1 mg I-' mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg 1-1 mg I-' mgl-1 mg I-' mgl-1 mg 1-1 - Pg I-' MI-' Pg 1- CLg1-l I% 1-' M1-l I % - ' Pg I-' CLg I-' Pg I-' Pg 1- Pg I-' - detector - Electrical conductivity at 20 "C pS cm-1 x4 Water 6.3 86 - - 0.130 0.013 21 14 48 107 122 <0.05 40 0.33 0.5 4.1 26 51 <0.01 <0.5 <o.1 - 0.08 3200 1780 330 < 10 <5 <2 63 10.3 65 300 < 10 420 0.53 Sodium humate 6.3 33 10 - 0.090 0.057 <1 <o. 1 25 <0.1 <1 123 <0.05 <o. 1 <0.05 <0.5 25 .o - < 10 <0.01 <o. 1 <0.01 510 <10 204 < 10 <5 <2 20 <5 <lo 4 .5 (10 < 10 108 - Purified Humic Acid Sample For comparison purposes, duplicate experiments were con- ducted using humic acid (HA) prepared from the semi- reference material sodium humate purchased from Aldrich. This material is well characterized, with capacity and complex- ation data readily available.Ql3 The purified HA wasANALYST, FEBRUARY 1992, VOL. 117 1- 153 produced by lowering the pH of a solution of Aldrich sodium humate to below 1 and then filtering off the precipitated HA. After washing and drying, 57.7mg of the precipitated HA were dissolved in 1 1 of high-performance liquid chromato- graphy (HPLC)-grade water. Finally, the pH was adjusted to 6.3 by the addition of a small amount of NaOH solution. The results of the analysis are given in Table 1.Preparation of Sample Mixtures Solutions of Ni(N03)2 were prepared in the range from 1 to 1 x 10-5 moll-1 from the analytical-reagent grade salt and then either 25 or 50 pl aliquots of these solutions were added to 5 ml samples of the ~4 moorland water and the purified humic acid solution. Mixtures were produced containing total nickel concentrations ranging from about 5.0 X 10-8 to 1.0 X 10-2 moll-'. Also, to each mixture either 25 or 50 pl of nickel-63 solution were added. The larger amount of nickel-63 was necessary for the more concentrated nickel solutions to permit detection of the complex in the presence of a large excess of free nickel. The nickel-63 solution was prepared by adding 135 pl of stock NiC12 solution (37 MBq ml-1; supplied by Amersham International) to 5 ml of HPLC-grade water.The total nickel concentrations of the mixtures were corrected for dilution effects, added nickel-63 and, for the ~4 water, the original nickel content. Characterization of the Dissolved Organic Matter The presence of humic and fulvic compounds in the moorland water was established as follows. A chromatographic separation of the filtered ~4 water was carried out using the GPC-60 column and the chromatogram obtained at 230 nm is shown in Fig. 2 (A). The elution volumes of the early peaks demonstrated the presence of large organic molecules. The UV absorption spectra of peaks 1, 2 and 3 showed a gradual increase in absorption, with decreas- ing wavelength, and the absence of specific absorption bands, properties which are typical of humic and fulvic materials.The narrow UV absorption spectrum of peak 4 indicated the presence of inorganic species, e.g. , NO3-, whereas the UV absorption spectrum of the low-intensity final peak (peak 5) suggested that small organic species were also present. A sample of the filtered ~4 moorland water was treated with DEAE-cellulose, which, as stated above, is known to extract humic and fulvic anions, and the chromatographic separation was repeated. The effect on the UV absorption monitored at 230 nm is shown in Fig. 2 (chromatogram B). The macromolecular organic species have been removed. Humic and fulvic compounds are generally fluorescent, hence the effect of the DEAE-cellulose treatment on the 2 I I I l l I l l I 0 2 4 6 8 10 12 14 16 18 Time/min Fig.2 UV chromatogram of x 4 moorland water monitored at 230nm: A, before and B, after treatment with DEAE-cellulose. (Peaks 1, 2 and 3 are due to large organic species, peak 4 is attributable to inorganic nitrate and peak 5 results from small organic species) fluorescence of the eluted species was determined. The results are shown in Fig. 3 (chromatograms A1 and Bl). Surprisingly, the first UV peak with a retention time of 9.3 min was not associated with fluorescence (cf., Figs. 2 and 3); however, the major UV peak exhibited fluorescence, which was removed by the DEAE-cellulose treatment. The fluorescent low relative molecular mass organic peak 5 was not completely removed. The fluorescence associated with the low relative molecular mass organic species may have indicated either fragmentation of the humic/fulvic materials or the presence of precursors. The lack of fluorescence associated with the largest molecules may be attributable to either the absence of appropriate aromatic groups or quenching caused by impurities and/or aggregation.Humic acid is by definition insoluble in very acidic solutions, i.e., pH tl. Accordingly, a sample was acidified and precipitation was observed. From measurements of the decrease in UV absorption a 70% fulvic-30% humic compo- sition was deduced (the precipitated humic material was separately tested and found to complex nickel in another series of experiments). From the above evidence, a knowledge of the source of the water, its light-brown colour and its acidity (pH = 3.8) when collected and the TOC content, it was concluded that the presence of humic and fulvic compounds had been estab- lished.Further corroborating evidence resulted from the nickel complexation experiments which, taking into account the time delay between the DAD and the radioactivity detector, 1 I I 1 1 I I I 1 I 0 2 4 6 8 10 12 14 16 18 Ti me/mi n Fig. 3 Fluorescence chromatogram of x 4 moorland water: Al, before and B1, after treatment with DEAE-cellulose. (For designa- tion of peaks see Fig. 2 and text) c-- Time -c Fig. 4 Simultaneous UV absorbance (230nm) and 63Ni activity (counts min-1) chromatograms versus time (min) for x 4 moorland water containing labelled Ni(N03)* solution154 ANALYST, FEBRUARY 1992, VOL. 117 0.40 8 0.30 C m e $j 0.20 Q 0.10 0 5 10 15 20 Ti me/m in Fig.5 UV absorbance chromatogram (230 nm) of purified Aldrich HA containing labelled Ni(N03)? solution. (Peak 1 is due to large organic species, peak 2 is due to nitrate and peak 3 results from small organic species) 45 min Run Equilibration start I I 1 35 min 15 min 5 min A (%) 100 100 100 0 B(%) 0 0 0 100 0 100 100 100 0 0 Fig. 6 Salt-gradient elution profile. Flow rate, 0.230 ml min-l; A, 0.05 moll-l NaCl; and B, 0.5 moll-' NaCl showed that nickel complexed with the species responsible for the early peaks (see Fig. 4). Chromatographic separation of the purified Aldrich HA yielded a simpler chromatogram. A typical result is shown in Fig. 5. The species producing peaks 1 and 3 are organic and fluorescent whereas peak 2 is due to added nitrate.Again, peak 3 is considered to be due to either humic fragments or precursors. Complexation Experiments After allowing 9 d for the mixtures to reach equilibrium, at 20°C, 100 yl samples of the mixtures were subjected to chromatographic analysis, using the salt-gradient technique, detailed in Fig. 6. Results In all instances the nickel activity eluted in two main fractions, as shown in Figs. 7 and 8, which show typical ~4 moorland water and purified HA chromatograms. The nickel humate/ fulvate complexes being partially excluded were eluted first with the 0.05 moll-1 NaCl and were identifiable with the early peaks in the corresponding DAD spectra, whereas the free Ni2+(aq) which suffered adsorption was eluted much later by the 0.5 moll-' NaCI.The nickel-63 chromatographic peak areas were used in conjunction with the known total nickel concentration to calculate the amounts of free and complexed nickel in each mixture. The assumption was made that the contribution of other nickel species to either peak was negligible. The results are given in Tables 2 and 3. A control experiment employing quench correction was conducted. Sample quenching was found to be insignificant, as peak area calculations employing disintegrations min-1 instead of counts min-1 gave identical results. Maximum Complexing Capacities In order to determine the maximum complexing capacities (C, values) of the ~4 moorland water and the purified HA, logarithmic plots of complexed versus free nickel concentra- tions were constructed (Figs.9 and 10). By using curve-fitting software (Macintosh Plus computer; Cricket Graph, polynomial order 2), best fit equations were 0 5 10 15 20 25 30 35 40 45 50 55 60 Ti melmin Fig. 7 Activity chromatogram of a typical x4 moorland water sample containing labelled Ni(N03)2 solution 0 5 10 15 20 25 30 35 40 45 50 55 60 Time/m in Fig. 8 Activity chromatogram of a typical HA sample containing labelled Ni(N03)2 solution obtained and differentiated to calculate the maximum com- plex concentrations theoretically achievable under the con- ditions used if precipitation is ignored. In this way the X4 moorland water was determined to have a C, value of 3.70 X 10-6mol 1-1 at a free nickel concentration of 3.18 X 10-3 moll-1, whereas the purified Aldrich HA yielded a C, value of 4.20 x 10-6 moll-1 at a free nickel concentration of 5.14 x 10-3 moll-'. In both instances precipitation occurred at about 1 x 10-3 moll-' of added nickel, which precluded precise experimental location of the maxima.Conditional Association Constant Determinations By using the appropriate maximum complexing capacity and eqn. (3), the individual values of the conditional association constants were determined for each mixture. The log Kvalues are included in Tables 2 and 3. For the moorland water the average log K was 4.40 [standard deviation (SD) = 0.41 (n = 9)] and for the purified Aldrich HA log K = 4.1 [SD = 0.25 (n = lo)]. Discussion This investigation was conducted at pH 6.3 to ensure a significant degree of complexation, as the exact nickel speciation of the original water was not an objective of this study.The results given in Table 4 cover a range of pH values. Generally, stability constants and maximum complexing capacities are found to decrease with decreasing pH, which is attributable to increased competition from H+ ions for binding sites. However, these experiments were conducted below pH 7 to avoid complications arising from the formation of hydroxy species. It can be seen that the log K values found are similar t o those reported by other workers for similar systems. Preconcentration results in pH changes, hence assessment of the nickel speciation in an original in situ water would need to take this into account. However, the inherent sensitivity ofANALYST, FEBRUARY 1992, VOL. 117 155 Table 2 Nickel complexation with x4 moorland water (C, = 3.72 x rnol 1-1) [Ni complex]/ [Nil/ moll-' mol 1-1 [Ni total]/ [Ni complex] [Nil moll-' (YO area) (% area) 2.619 x 10-7 16.01 83.99 4.193 x 10-8 2.200 x 3.107 x 10-7 15.69 84.31 4.875 x 10-8 2.620 x 10-7 7.075 x 10-7 13.41 86.59 9.488 x 10-8 6.126 x 1.197 x 10-6 16.05 83.95 1.921 x 10-6 1.005 x 5.163 x 10-6 10.26 89.74 5.297 X 10-7 4.633 x 10-6 1.006 x 10-5 5.85 94.15 5.885 x 10-7 9.472 x 10-6 4.973 x 10-5 2.86 97.14 1.422 X 10-6 4.831 x 10-5 9.829 x 10-5 2.67 97.33 2.624 x 10-6 9.567 x 10-5 4.927 x 10-4 0.443 99.557 2.183 x 10-6 4.905 x 9.805 x 10-4 0.443 99.557 4.344 X 10-6 9.762 X * Average log K = 4.40 [SD = 0.41 (n = 9)].Log [Ni complex] -7.378 -7.312 -7.023 -6.716 -6.276 -6.230 -5.847 -5.581 -5.661 -5.362 Log "il -6.658 -6.582 -6.213 -5.998 -5.334 -5.024 -4.316 -4.019 -3.309 -3.01 li Log K* 4.72 4.71 4.63 4.74 4.56 4.30 4.11 4.41 3.46 - Table 3 Nickel complexation with purified Aldrich HA (C, = 4.2 X mol I-l) [Ni complex]/ [Nil/ moll-1 moll-' [Ni total]/ [Ni complex] [Nil moll-' (YO area) (YO area) 8.692 x 10-8 7.40 92.60 6.432 x 10-9 8,049 x 10-8 1.357 x 8.99 91.01 1.220 X 10-8 1.235 X 5.325 x 10-7 7.32 92.68 3.898 x 10-8 4.935 x 1.022 x 10-6 7.11 92.89 7.266 X 10-8 9.493 X 4.988 x 6.46 93.54 3.222 x 4.666 x 9.889 x lo-" 5.61 94.39 5.548 x 10-7 9.334 x 10-6 4.955 x 10-5 2.60 97.40 1.288 X 10-6 4.826 X 9.811 x 10-5 2.15 97.85 2.109 x 10-6 9.600 X 4.927 x 10-4 0.58 99.43 2.833 x 10-6 4.899 x 9.805 x 0.37 99.63 3.628 X 10-6 9.769 X * Average log K = 4.10 [SD = 0.25 (n = lo)].Log [Ni complex] -8.191 -7.914 -7.409 -7.139 -6.492 -6.256 -5.890 -5.676 -5.548 -5.440 Log [Nil -7.094 -6.908 -6.307 -6.023 -5.331 -5.030 -4.316 -4.018 -3.310 -3.010 Log K* 4.28 4.37 4.28 4.27 4.25 4.21 3.96 4.02 3.60 3.81 Fig. 9 Ni-X4 moorland water binding. y = -6.1289 - 0.55856~ - 0.11186~~ and r2 = 0.986. Shaded area shows precipitation region Fig. 10 Ni-humate binding. y = -6.0161 - 0.55820~ - 0.12192~~ and r2 = 0.998. Shaded area shows precipitation region the technique can be exploited for direct measurements on in situ waters when the level of dissolved organic material is appropriately high, i.e., about 10 mg 1-1 TOC or higher. It is worth noting that the high SD of the current results is due in part to the wide concentration range studied, i.e., six orders of magnitude (see later), and the use of individual experimental results rather than the averaging of the means of several replicate series of experiments.Table 4 Over-all log K values Sample PH Log K Ref. Soil fulvic acid 5.0 4.20 2 Ground water fulvic acid 6.5 5.2 13 Stream fulvic acid 7.0 4.63 3 Lake water fulvic acid 8.0 5.14 4 x4 Moorland water 6.3 4.40 This work Purified Aldrich HA 6.3 4.10 This work ~~~~~~ ~~~~~ The apparent difference between the average log K values obtained for the moorland water and purified HA is not statistically significant. The capacities show about a 10% difference. The increased capacity of the purified material is arguably due to the freeing of metal sites during the dissolution and re-acidification stages. The values of K and C, obtained must be regarded only as operational values.They are not thermodynamic constants and cannot even be described as stoichiometric constants. The reasons are easily demonstrated by returning to eqn. (3). Rearrangement of this equation gives Hence a graph of [ML]/[M] against [ML] should be a straight line with a slope of - K and an intercept on the abscissa of The graphs obtained by treating the data in this way are presented in Figs. 11 and 12. The lack of linearity is immediately apparent. However, the C, values obtained from the intercepts on the abscissa are comparable to the previously produced values, i.e., 3.9 x 10-6mol 1-1 for the Ni-HA (previous value 4.20 x 10-6 moll-1) and 4.8 X 10-6 moll-1 for the ~4 moorland water (previous value 3.70 X 10-6 mol 1-1).It should be noted that Perdue14 stated that maximum capacities determined by adding excess of metal ion may be in error, especially at low ligand concentrations, and suggested that the H+ ion capacity should be used as an upper limit for the site capacity. This approach is precluded with an in situ water sample, but Kim et al. ,*5 for example, used one third of the H+ ion capacity as a measure of humic concentration in Am3+ complexation studies. However, the [MLlrnax.156 ANALYST, FEBRUARY 1992, VOL. 117 2000 rf 0 X 7 , log K = 3.8 1 0 1000 2000 3000 4000 5000 [ML]/10-9 mot I-' Fig. 11 Ni-x4 moorland water complexation loo0 ~ 800 K = 4.8 2 i 4 0 0 1 1 1 200 \ - - log K = 3.9 \ m 0 1000 2000 3000 4000 [ML]/10-9 mot 1-1 Fig. 12 Ni-humate complexation lack of linearity is because within humic and fulvic materials, there are a variety of ligand sites with different binding strengths.In more refined treatments of the ligand site model, capacities of each type of site are determined in order to derive individual binding site constants. Methods usually involve the Scatchard analysis technique, an approach much used by biochemists in drug-receptor binding studies.16 Figs. 11 and 12 can be interpreted in this way. If the simplifying assumption is made that only two types of site are present, A-sites, which are strongly binding, and B-sites, weakly binding, then from eqn. (4) Now, during the early stages of metal addition [MLIB = 0 because the added metal binds preferentially to the A-sites, and therefore the initial slope is -KA, and in the latter stages of the titration [MLIA = [ML]A,max, and therefore the final slope is -KB.The log K values thus determined are presented in Table 5 and again over-all similarity is observed. However, this approach can be criticized: the location of the linear sections is arbitrary and co-operativity and stereochemical effects may, in reality, cause the binding strengths and capacities to change continually as complexation progresses, i.e. , the tenets of the continuous distribution model.3 In yet another approach, Klotzl7 recommended the use of so-called stepwise stoi- chiometric constants. Accordingly, it is emphasized that the values given in Tables 2 and 3 are based on total capacities whereas the values in Table 5 are based on the weak and strong site approach.Conclusion Within the 'analytical window' of the technique and under the conditions used, only minor differences have been discovered between the nickel complexation properties of the dissolved humic and fulvic components of the ~4 moorland water and Table 5 Approximate strong and weak site log K values x 4 Moorland Purified Site water Aldrich HA Strong 5.5 4.8 Weak 3.8 3.9 the purified HA material. However, the ~4 water apparently possesses a proportion of stronger sites than the purified HA. The main objective of the study was achieved, namely to demonstrate that, when the rate of dissociation is slow, this form of HPSEC can be used in complexation studies to measure both complexed and free metal concentrations in environmental waters without prior extraction of the humic and fulvic materials. The technique is now being applied to compare extracted humic and fulvic materials from the same moorland water.A full account of the comparison will be reported elsewhere. 18 The authors thank the UK Department of Environment for funding this work. The results of this work will be used in the formulation of government policy, but the views expressed in this paper do not necessarily represent government policy. The Analytical Laboratories of the National Rivers Author- ity, Meadow Lane, Nottingham, are also thanked for the water analysis, as is Christine Bartrop for typing the manu- References script. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Patterson, J. W., and Passino, R., Metals Speciation, Separation and Recovery, Lewis, Chelsea, MI, 1987.Flaig, W., Beutelspacher, H., and Rietz, E., in Soil Com- ponents, ed. Gieseking, J. E., Springer-Verlag, New York, Humic Substances in Soil, Sediment and Water, eds. Aiken, G. R., McKnight, D. M., Wershaw, R. L., and MacCarthy, P., Wiley, New York, 1985. Falck, W. E., A Review of Modelling the Interaction Between Natural Organic Matter and Metal Cations, British Geological Technical Report WE/88/49, British Geological Survey, Not- tingham, 1988. Buffle, J., Complexation Reactions in Aquatic Systems, Ellis Horwood, Chichester, 1988. Christman, R. F., and Gjessing, E. T., Aquatic and Terrestrial Humic Materials, Ann Arbor Science Publishers, Ann Arbor, MI, 1983. Miles, C. J., Tuschall, J. R., Jr., and Brezonk, P. L., Anal Chem., 1983,55, 410. Warwick, P., Shaw, P., Williams, G. M., and Hooker, J. P., Radiochim. Acta, 1988, 44/45, 59. Schnitzer, M., and Hansen, E. H., Soil Sci., 1970, 109,333. Mantoura, R. F. C., and Riley, J. P., Anal. Chim. Acta, 1975, 78, 193. Haworth, D. T., Pitluck, M. R., and Pollard, B. P., J. Liq. Chromatogr., 1987, 10, 2877. Kim, J. I., Buckau, G., Li, G. H., Duschner, H., and Psarros, N., Fresenius J. Anal. Chem., 1990, 338, 245. Smith, B., Higgo, J. J. W., Moodie, P., Davis, J., Williams, G. M., and Warwick, P., Comparative Study of Humic Sub- stances in Groundwaters: I . The Extraction of Humic Material from Drigg Groundwater and a Study of its Ability to Form Complexes with Cobalt and Nickel, DOE Report No. DOE/ HMIP/R/90/087. DOE, London, 1990. Perdue, M. E., in Metal Speciation: Theory, Analysis and Application, eds. Kramer, J. R., and Allen, E. H., Lewis, Chelsea, MI, 1988. Kim, J. I., Rhee, D. S., and Buckau, G., Radiochim. Acta, 1991, 52/53, 49. Scatchard, G., Ann. N.Y. Acad. Sci., 1949, 51, 660. Klotz, I. M., Acc. Chem. Res., 1974, 7, 162. Warwick, P., and Hall, A,, Radiochim. Acta, submitted for publication. Paper 1 I04583 K Received September 4, 1991 Accepted September 30, 1991 1975, VOI. 1, pp. 1-211.