Environ. Sci. Technol. 1993, 27, 2523-2532
Binding of a Fluorescent Hydrophobic Organic Probe by Dissolved Humic Substances and Organically-Coated Aluminum Oxide Surfaces Mark A. Schlautman’vt and James J. Morgan
Environmental Engineering Science, California Institute of Technology, Pasadena, California 9 1125 The binding of perylene by Suwannee River humic substances in the presence and absence of colloidal-sized aluminum oxide particles was examined using a fluorescence quenching technique. Our experiments show that binding is complete within 3 min and that the fluorescence of perylene associated with dissolved and adsorbed humic substances is fully quenched as evidenced by quantum yields which approached zero for all systems. In the absence of alumina, both humic acid and fulvic acid were able to bind perylene, and the partition coefficients decreased with increasing pH and NaCl concentrations. The presence of Ca2+had little effect on the binding of perylene by either of the dissolved humic substances. The adsorption of humic and fulvic acids onto alumina decreased their ability to bind perylene. For all solution conditions examined, the association of perylene with adsorbed fulvic acid was never detected. In NaClsolutions, partition coefficients for adsorbed humic acid at pH 4 were approximately half the values of those for dissolved humic acid; at pH 7 and 10, alumina-bound humic acid did not bind perylene in NaCl solutions. In contrast to the results observed for dissolved humic acid, the presence of Ca2+ greatly enhanced the binding of perylene by adsorbed humic acid. A major effect of solution chemistry is to alter the mechanisms by which humic substances adsorb to alumina, thereby determining how tightly the humic material is bound to the surface. The ability of weakly-adsorbed humic acid to bind perylene approaches that of the dissolved sr>ecies.
Introduction The geochemistry, and ultimately the fate, of organic compounds in natural water systems is strongly dependent on chemical reactions that occur at the solid-liquid interface of particles. These reactions can greatly affect the mobility, bioavailability, reactivity, and toxicity of organic compounds. For hydrophobic compounds, which have relatively low solubilities in water, reactions at surfaces are particularly important. The distribution (partitioning)of nonionic hydrophobic organic compounds (HOCs) between water and surface soils or sediments has been shown to depend primarily on the hydrophobicity of the compound and the fraction of organic carbon Cf,) in the sorbent (1-4). Partitioning of HOCs between organic carbon and water results primarily from the “hydrophobic interaction” (5-7). The importance of the organic carbon content of sediments and soils suggests that various components of natural organic material (NOM) bind nonionic HOCs in solution and at mineral surfaces in aquatic systems. Recently, much work has been focused on studying the + Present address: Department of Civil and Environmental Engineering, The University of Michigan, Ann Arbor, MI 48109-
2125. 0013-936X/93/0927-2523$04.00/0
0 1993 Amerlcan Chemical Society
binding of hydrophobic pollutants by dissolved organic material (DOM) in solution (e.g., refs 8-16). Collective results from these studies suggest that the binding of hydrophobic organic pollutants by DOM depends on the chemical and structural characteristics (Le., the “quality”) of the DOM and also on the aqueous chemistry of the system. In the environment, organic material adsorbs at particle surfaces and can dominate the properties of the solidliquid interface. While only a fraction of the NOM in aquatic systems may be adsorbed on particle surfaces, the solid surface may be extensively covered by organic material and thus exhibit the physical and chemical properties of the organic matter. In effect, the adsorbed species may be thought of as an organic “film” coating particle surfaces. Organic surface coatings modify particle-pollutant interactions which occur at the solid-liquid interface or on particle surfaces. Because colloidal stabilization by NOM is a common observation, the transport and fate of NOM, particles, and particle-reactive compounds in natural waters are complex, interdependent processes. Few studies have investigated the association of HOGS with mineral-bound NOM. Garbini and Lion (10) examined the binding of trichloroethene and toluene by a commercial humic acid and by alumina particles coated with the same humic acid and observed that association of the solutes with dissolved humic acid was nearly three times as much as that with humic acid adsorbed on alumina. Amy et al. (17) examined the sorption of phenanthrene by NOM collected from two different groundwater sources and by a soil-derived fulvic acid. For some experiments, the NOM was adsorbed to an aluminum oxide surface before phenanthrene binding was determined. Amy et al. (17) observed no sorption of phenanthrene by alumina-bound groundwater NOM and less sorption of phenanthrene by alumina-bound fulvic acid relative to the original fulvic acid. Also, after contacting the three different NOM solutions with alumina, less binding was observed between phenanthrene and residual NOM (i.e., NOM not adsorbed to alumina) than was observed for phenanthrene and the original NOM solutions. Murphy et al. (18) examined the sorption of three HOCs onto hematite and kaolinite which had been coated with natural humic substances at fOc levels ranging from 0.0001to 0.005. Because the observed sorption isotherms were nonlinear, Murphy et al. (18) hypothesized that adsorption onto rather than dissolution into the surface organic phase was the sorption mechanism. Anthracene, the most hydrophobic compound, showed the greatest binding. The humic substance with the highest aromaticity, peat humic acid, was the strongest sorbent. The mineral substrate used influenced the amount of binding; however, because binding measurements were not made in the absence of minerals, the effect of mineral surfaces on the ability of humic substances to bind hydrophobic organic compounds cannot be determined directly. Envlron. Sci. Technol., Vol. 27, No. 12, 1993
2523
This study investigated the ability of adsorbed humic substances to bind a nonpolar HOC relative to "freely dissolved" humic material. A principal hypothesis was that the conformation and polarity of humic substances play important roles in their adsorption onto mineral surfaces and in their ability to bind hydrophobic pollutants both in solution and at the solid-liquid interface. Because the conformation and polarity of humic materials and the mechanisms of their adsorption onto mineral surfaces are affected by the aqueous chemistry of natural waters, a systematic study of the effects of pH, ionic strength, and the presence of bivalent cations was undertaken. The objectives of this study, therefore, were to (1)determine the amount of binding of a hydrophobic organic probe by dissolved (i.e., not adsorbed) humic material, (2) quantify the amount of binding of the probe by adsorbed humic substances, and (3) investigate the effects of different adsorption mechanisms on the ability of adsorbed humic substances to bind the probe. A central aim of this study was to distinguish between association of the hydrophobic probe with dissolved versus adsorbed humic substances in situ in order to avoid a separation and resuspension of the adsorbed humic material before a partitioning experiment (e.g., refs 10, 17, and 18). There was concern that separation of the adsorbed humic material from that remaining in solution would disrupt equilibria and lead to variable estimates of binding constants, much like those observed upon separating pollutants equilibrated between two phases (12). Experimental Section
Materials. The hydrophobic probe used in this study was perylene (Aldrich, 99+ % pure). This nonpolar, nonionogenicpolycyclicaromatic hydrocarbon (PAH) was selected in order to eliminate or minimize possible adsorption mechanisms other than the effect of hydrophobic interactions in the binding reaction. Relevant chemical and physical properties of perylene have been tabulated elsewhere (16). A conservativeprobe, rhodamine 110 (Lasergrade),was obtained from Kodak. Both probes were used without further purification (19). Fluorescence measurements were made on a Shimadzu Model RF-540 recording spectrofluorophotometer. Fluorescence intensity was monitored for perylene and rhodamine 110 at the excitation/emission wavelengths (nm/nm) of 4321470 and 496/520, respectively, with slits set for bandwidths of 10 nm on both the excitation and emission monochromators. Total extinction (i.e., absorbance plus scattering) measurements were made on a Hewlett Packard 8451A diode array spectrophotometer at the above wavelengths to correct for inner-filter effects (12, 20, 21). Correction factors for perylene and rhodamine 110 approached 1.0 for samples without colloidal particles and were always less than 1.2 for those containing particles. A combined stock solution of perylene and rhodamine 110 was prepared in methanol (16). The use of small amounts of methanol (3.3 X 103 volume fraction) as a carrier solvent did not significantly affect the partitioning results of this study (19). Well-characterized humic materials were obtained from the International Humic Substances Society (IHSS) and used without further purification. Isolation procedures and characterization of these materials have been previously reported (22). Chemical and physical properties of the humic and fulvic acids, as well as stock solution preparations, have been 2524
Envlron. Sci. Technol., Vol. 27, No. 12, 1993
reported elsewhere (16, 23). Commercially-available colloidal-sized aluminum oxide particles (Degussa, aluminum oxide C) were used in this study as a model for mineral surfaces. The average particle diameter from TEM was 26 nm with a range observed from 14 to 50 nm. These small particles have high specific surface areas and can form relatively stable particle suspensions. Alumina has a high zero point of charge (pH,,, -8.5) and, thus, is a good sorbent for organic acids. Because aluminum oxide C has been used for adsorption studies by previous researchers (e.g., refs 23-26), its surface characteristics are relatively well-known. Pretreatment, preparation of stock suspensions, and characterization of the particles have been reported elsewhere (23). Procedure. Adsorption isotherms at room temperature (-23 "C) of humic and fulvic acids on alumina were obtained at constant pH and ionic strength values. Initial concentrations of the humic substances ranged typically from 0 to 10 mg/L. Very low solids concentrations (20 mglL) were used in order to maximize the adsorption density (i.e., obtain high surface coverage) and minimize the effects of particle coagulation in the adsorption experiments and, particularly, in the subsequent perylene partitioning experiments. After a 24-h adsorption period, the pH of each sample was recorded and a 3-mL aliquot (designated as "heterogeneous" sample) was removed for the partitioning experiment. The remaining portion of each sample was analyzed to determine the quantity of humic material adsorbed on the aluminum oxide particles (23). Briefly, the particles were filtered, resuspended in 0.01 M HCl and then injected directly into a total organic carbon analyzer (Shimadzu Model TOC-500). f O c values ranged from 0.003 to 0.066 and were dependent on the type of humic material used, as well as solution chemistry conditions. Humic material blanks with no alumina present (designated as "homogeneous" samples) were prepared at the same initial DOM concentrations as the adsorption samples and were used in partitioning experiments. The humic blanks and sample aliquots removed for the partitioning experiments were used immediately to minimize the effects of hydrolysis of the acids (27). A fluorescence quenching method for determining perylene partition coefficients was adapted from Backhus and Gschwend (15) and has been described previously (16). Data Treatment. (A) Adsorption of Perylene to Glassware. For very hydrophobic PAH probes, decreases in observed fluorescence result not only from fluorescence quenching but also from sorption of the PAH toglassware. Assuming first-order reaction kinetics for perylene sorption to the cell walls, the observed fluorescence versus time is (15,16)
where Fo' = [PAHT] - [PAH-OCI at time zero, and k, and k-, are the first-order forward and reverse rate constants for wall adsorption. [PAHT] and [PAH-OC] are the total perylene concentration and the mass of perylene bound by NOM per unit volume of sample, respectively. By fitting eq 1to experimental data and back-extrapolating to t = 0, hypothetical fluorescence values are obtained which would have been observed had perylene not been adsorbing to the cell walls.
(B) Binding of Perylene by Dissolved Humic Substances. Most analyses using fluorescencequenching techniques to determine PAH partition coefficients rely on the assumption that only the free solute contributes to observed fluorescence. This assumption is valid only if the fluorescenceof the PAH associated with NOM is totally quenched (Le., that the fluorescence quantum yield of the PAH-NOM complex, @, is zero). However, for different types of NOM, or as chemical conditions of the system change, this assumption may not be valid and must be verified. Backhus and Gschwend (15)showed that if the fluorescence of a PAH probe associated with one type of NOM (e.g., dissolved humic material in this study) is not totally quenched, then the observed fluorescence of the probe is described by
With this formulation, the normalized fluorescence asymptotically approaches @ as [OCI ==+ m. From a plot of FIFOversus [OCI, KO,equals the value of [OCI-l at the point where FIFOequals (1 + @)/2. (C) Binding of Perylene by Dissolved and Adsorbed Humic Substances. In natural waters, NOM adsorbs at mineral surfaces; thus, a second “compartment”of organic material which may bind the PAH probe is now present in the system. In the present study, this second compartment of NOM is humic material adsorbed at and coating the aluminum oxide surface. In the presence of dissolved and adsorbed humic substances, the fluorescence intensity of perylene may now be described by:
F
a
[PAH,] (fraction dissolved + @(fractionbound-free humics) + cbAdS(fraction bound-adsorbed humics)) (3)
The extra term in eq 3 relative to Backhus and Gschwends (15) derivation represents the probe associated with the organic coating on alumina particles. Combining mass balance equations for perylene in solution, associated with dissolved humic material, and associated with humic material adsorbed at the aluminum oxide surfaces with eq 3 gives (19)
where the superscript and subscript “Ads” signifies adsorbed humic material. The normalized fluorescence intensity is now observed to be a function of two variables, [OCI and [OC]Ads. In applying eq 4 to experimental data, the values of @ and KO, were obtained from prior experiments without aluminum oxide particles present (Le., humic material blank solutions or homogeneous solutions). Utilizing these previously determined values and the measured values of [OCI and [OCIAds, @Ads and KWAd” were determined from systems having both dissolved and adsorbed humic material present.
Results and Discussion Methods Verification. (A) Kinetics of Perylene Binding by Dissolved and Adsorbed Humic Material. Using the fluorescence quenching method of Backhus and Gschwend (15), the rate at which perylene partitions between water and NOM is observed indirectly because
of the competing reactions (association with NOM versus adsorption to the fluorescence cell walls). In a sense, the rate of perylene binding by NOM must be known a priori because a key assumption of the technique is that perylene binding by NOM is fast relative to the rate of adsorption to the cell walls, By allowing the partitioning reaction to reach equilibrium before the initial measurement is taken, subsequent decreases in fluorescence are then attributed to adsorption of the PAH to cell walls. In the presence and absence of NOM, measurements of the perylene fluorescence intensity decreased exponentially with time, in accordance with eq 1 (data not shown). The binding of perylene by dissolved and adsorbed humic material was observed to be complete within 3 min, the time allotted before recording the first fluorescence measurement. Fast equilibration times for the association of PAHs with DOM have been reported previously (9,12,15, 16). No kinetic information has been reported previously, however, for the binding of hydrophobic organic compounds by wellcharacterized, humic-coated hydrous oxide surfaces. It should be noted that the fluorescence quenching technique utilized in this study to investigate the binding of perylene by alumina-bound humic substances does not measure adsorption of the fluorescent probe by the bare aluminum oxide surface. Fluorescence quenching only determines the amount of the probe associated with NOM, because inorganic surfaces do not quench the fluorescence of organic molecules which are physically adsorbed onto them (19,28). The use of inorganic materials as substrates for studying the photochemistry of physically adsorbed organic molecules and the use of fluorescent organic molecules to probe the solid-liquid interfacial region of inorganic colloids have been extensively documented (29). The rates reported here are indicative only of perylene binding by NOM. It is realistic to believe that any adsorption of perylene that was occurring on the bare surfaces of the aluminum oxide particles was much slower, based on the removal of the organic probe from solution to the fluorescence cell walls. For example, an apparent equilibrium for perylene adsorption to the cell walls was reached only after a contact time of a few hours (30). (B) Quenching Efficiency of Dissolved and Adsorbed Humic Substances. On the basis of diffusion calculations and temperature studies, fluorescence quenching of PAH compounds by NOM has been hypothesized to be a static quenching process (12, 14). In a study of fluorescence quenching and lifetimes of naphthalene in humic acid solutions, Morra et al. (31)observed that both static and dynamic processes were contributing to observed quenching; however, the dynamic quenching contribution was 2 orders of magnitude smaller than static quenching. Mechanisms of energy transfer from an excited organic molecule (donor) to a static quenching species (acceptor) have been discussed by Turro (28),and these concepts have been summarized for systems in which humic substances bind PAHs (19). Few published results are available for the quantum yields of PAH-NOM complexes. Schlautman and Morgan (16) observed that the quantum yields of anthracene, pyrene, and perylene all approached zero when the PAHs were associated with dissolved Suwannee River humic acid or fulvic acid. Backhus and Gschwend (15) found that Aldrich humic acid fully quenched the fluorescence of perylene associated with it (@ = 01,while bovine serum albumin quenched only 42% of the associated perylene Environ. Scl. Technol., Vol. 27, No. 12, 1993
2525
0 0
Heterogeneous Homogeneous
0.8-
0.6-
= 8.87 x I O 5 mL/g-C
9
= -3 x IO-”
7
fluorescence (4 = 0.58). The observations of Gauthier et al. (12,131 suggest that values of zero for 4 are found for a wide variety of humic and fulvic acids. However, as chemical conditions of a system change, the assumption of complete fluorescence quenching may not be valid and must be verified. Partition coefficients and quantum yields for peryleneDOM complexes were determined by analyzing the initial perylene fluorescence intensities (after accounting for wall losses) of homogeneous solutions using eq 2. Results for a typical experiment are shown in Figure 1,in which the normalized fluorescence intensities obtained from backextrapolation to t = 0 are plotted as a function of humic acid concentration. The curve asymptotically approaches a quantum yield of zero as [OC] * a, The quantum yield of perylene approached zero for all solution chemistry systems when it was associated with dissolved humic acid or fulvic acid. Heterogeneous samples containing both dissolved and adsorbed humic material were analyzed similarly using eq 1to determine fluorescencevalues at time zero. The values of KO,and 4 obtained from homogeneous solutions and the respective fractions of free and adsorbed humic substances (Le., [OCI and [OCIAds) obtained from TOC analysis were then utilized in eq 4 in order to determine partition coefficients and quantum yields for the humic substances adsorbed to aluminum oxide particles. Figure 1shows the results for a typical study. For the same total amount of humic acid, the presence of aluminum oxide particles diminished the ability of humic acid to bind perylene. For this particular system, the association of the probe with dissolved humic acid was nearly two times that with the humic acid adsorbed on alumina surfaces. All quantum yields for perylene associated with adsorbed humic acid or fulvic acid were not significantly different from zero. Most quantum yields for the adsorbed humic substances, however, may not be credible because of the absence of significant binding of perylene. Reliable values of $Ads are those for humic acid at pH 4 and for humic acid in systems in which Ca2+was present. 2526
Envlron. Scl. Technol., Vol. 27, No. 12, 1993
(C)Fractionation of Humic Material. The method described above for determining partition coefficients for adsorbed humic substances assumes that the humic substances which are not adsorbed have the same properties and charqcteristics as the original humic material; thus, partition coefficients and quantum yields of the dissolved humic material can be utilized to calculate the partition coefficients and quantum yields of the adsorbed material. The major criterion for using this method, therefore, is that the adsorption process does not fractionate the humic substances. In a study of fractionation, Davis and Gloor (32)found that higher molecular weight DOM was preferentially adsorbed to an aluminum oxide surface. Thus, the process of adsorption fractionated the DOM, with the more hydrophobic compounds adsorbed to the alumina surface and the more hydrophilic compounds remaining in solution. Similar effects may be inferred from the results of Amy et al. (17). They observed that after contacting three different NOM solutions with alumina, less binding was observed between phenanthrene and the residual NOM than was observed for the PAH and original NOM solutions. Although the humic substances used in this research were isolated fractions (33),they still are not homogeneous materials. The ratio of weight-averageto number-average molecular weights, a measure of sample dispersity (19), ranges from 1.66 to 1.83 for Suwannee River fulvic acid and from 2.78 to 3.87 for Suwannee River humic acid (3436). The method used by Beckett et al. (35)to determine molecular weights (field flow fractionation), in fact, utilizes the ability to fractionate mixtures based on the different diffusion coefficients of sample species. An empirical size fractionation of Suwannee River humic acid has also been accomplished with gel permeation chromatography (37). If significant fractionation of the humic substances by adsorption onto alumina occurred in this study, the method utilized to determine KocAda and 4Ads would not be valid. Murphy et al. (18)examined whether the fractionation of Suwannee River humic acid could be detected in an adsorption/partitioning experiment. In their experiment, Suwannee River humic acid was sequentially reacted with hematite to determine if the characteristics of the humic coating would change upon sequential adsorption to clean mineral surfaces; the nonsorbed supernatant in each batch reaction was used for the subsequent reaction with clean hematite. Murphy et al. (18) postulated that if fractionation occurred, both the extent of humic acid adsorption onto hematite and the cosorption of a hydrophobic compound would not remain constant for the sequential reactions. However, they found that neither the fractional adsorption of humic acid nor the relative amount of the hydrophobic compound bound by the humic coating changed significantly between successive batch experiments. These results show that Suwannee River humic acid does not fractionate appreciably upon adsorption to a mineral surface and that the method utilized in this study to determine KocAdsand 4Ads is valid. Although Murphy et al. (18)did not examine Suwannee River fulvic acid, it should not fractionate in the present study because it is less polydisperse than Suwannee River humic acid. (D) Fluorescence Quenching for Unstable Particle Suspensions. The fluorescence quenching method used in this study was originally intended for use with stable colloid suspensions because the background fluorescence intensity from solution components (principally light
scattering by the particles) is determined for each sample prior to spiking with a fluorescent probe. However, for a few heterogeneous samples it was observed that sedimentation of unstable, organically-coated particles was occurring in the fluorescence cells during the quenching portion of the experiments. Sedimentation was identified by the following observations: (i) corrected perylene fluorescence intensities of the suspect samples at time zero (obtained from eq 1)were higher than expected and were sometimes even higher than the corrected fluorescence intensity of the blank sample, (ii) values of the firstorder forward rate constants for loss of perylene to wall adsorption (k, in eq 1)were abnormally high relative to other samples, and (iii) net fluorescence intensities for the conservative dye rhodamine 110 decreased systematically with time. Upon checking the extinction values (used for the innerfilter corrections) of these particular samples, an exponential decay in total extinction versus time was observed (Figure 2b). For all of the stable particle suspensions as well as the homogeneous samples, extinction values were scattered around a constant value (Figure 2a,c). Because the low concentrations of perylene (0.4 pg1L) and rhodamine 110 (0.1NglL) in the samplescannot be detected by absorbance (19),the observed extinction values reflect only the background solution components. The turbidity of colloidal suspensions results primarily from the light scattered by the dispersed particles, and the amount of scattering depends on the size and shape of the particles as well as the wavelength of the incident light. Because coagulation increases the particle size distribution, the intensity of light scattering increases. Thus, in Figure 2 the initial extinction values show that coagulation was appreciable in the presence of 0.86 mg/L humic acid. The particle suspensions undergoing sedimentation all had low amounts of NOM present. Adsorption of small amounts of humic material modified the particle surface charges in these suspensions close to a region where they had a low stability; further increases in NOM adsorption restabilized the particles. Liang and Morgan (38, 39) observed similar phenomena for submicron hematite particles. Upon particle sedimentation, the background fluorescence decreased for both perylene and rhodamine 110. In using the initial background fluorescence intensities to correct all data, the anomalies above (i-iii) were introduced. To a first approximation, the rate of sedimentation can be modeled by the expression (40) dh41dt = J i n - Jout (5) where M is the mass of aluminum oxide particles in the system, and J i n and Joutare the input and removal fluxes, respectively, of theparticles. For systems with no material added, the input flux is zero. Assuming the removal flux is proportional to the total amount of material present and that the observed extinction values and fluorescence intensities are proportional to the particle mass present, eq 5 may be solved to give:
EIE, = FIFO= e-kt (6) where E and F denote the total extinction and fluorescence, respectively, and k is the rate constant of sedimentation. The validity of the assumption that total extinction and fluorescence are proportional to particle mass can be observed in Figure 2b. Because the four extinction data sets are tracking the same process, the curves are parallel
0.030 Q-Q 432 nm W 470 nm
(0)
0.01 5
0
1
.
0
0 10
1
20
0
30
1 40
I 50
Time (rnin)
".""" 0.025-
I
@€I 432
nm
W 470 nm
A-A 496
:(
nm
W 520 nm
0.020-
0.015-
0.010
I
I
I
I
I
0
10
20
30
40
Time (rnin) 0.030 Q-Q 432 nm
(c)
W 470 nm
0.025
.r; W -0
0.020
Y
0
I-
0.0151
0.0104
0
I
I
I
I
10
20
30
40
1
Time (rnin) Figure 2. Light extinction as a function of time for suspensions of aluminum oxide particles with varying total concentrations of humic acid in 0.001 M NaCi at pH 4. The alumina solids concentrations are 20 mg/L. (a) 0.43 mg/L humic acid. (b) 0.88 mg/L humic acid. (c) 1.72 mg/L humic acid.
with the same decay time constant. Fluorescence decay curves similarly will have the same decay time constant. The approach used to correct the background fluorescence intensities for unstable particle suspensions was to (1)fit exponential decay curves to extinction data for the four different wavelengths, (2) use average values from these fits to generate new background fluorescence intensities (now functions of time), and (3) subtract these new background values from measured values. Corrections for sedimentation were made only to those particle suspensions for which an obvious decrease in total exEnviron. Sci. Technol., Val. 27, No. 12, 1093
2527
Table I. Experimental Partition Coefficients &SD) for Perylene and Humic Substances8
dissolved humic substance, K, X
adsorbed humic substance, x 10-5
Humic Acid PH 4 0.001M NaCl 0.01M NaCl 0.1M NaCl 0.001 M Ca2+ b PH 7 0.001M NaCl 0.01M NaCl 0.1M NaCl 0.001 M Ca2+ pH 10 0.001M NaCl 0.01M NaCl 0.1M NaCl 0.001 M Ca2+ b
11.28 (0.65) 10.31 (0.62) 8.87 (0.46) 9.29 (0.11)
6.68 (0.58) 4.47 (0.33) 4.72 (0.39) 5.47 (0.19)
9.58 (0.30) 8.99 (0.28) 4.69 (0.14) 4.10 (0.09)
