Correlation of the partitioning of dissolved organic matter fractions with the desorption of cd

Mối quan hệ của các dạng hữu cơ hòa tan đến khả năng giải hấp Cd trong môi trường đất

Correlation of the partitioning of dissolved organic matter fractions with the desorption of Cd, Cu, Ni, Pb and Zn from 18 Dutch soils Christopher A. Impellitteri a,1 , Yuefeng Lu a,2 , Jennifer K. Saxe a,3 , Herbert E. Allen a, * , Willie J.G.M. Peijnenburg b a Department of Civil and Environmental Engineering, University of Delaware, Newark, DE 19716, USA b National Institute of Public Health and the Environment, Laboratory for Ecotoxicology, PO Box 1, 3720 BA Bilthoven, The Netherlands Received 10 February 2002; accepted 12 August 2002 Abstract Eighteen Dutch soils were extracted in aqueous solutions at varying pH. Extracts were analyzed for Cd, Cu, Ni, Pb and Zn by ICP-AES. Extract dissolved organic carbon (DOC) was also concentrated onto a macroreticular resin and fractionation into three operationally defined fractions: hydrophilic acids (Hyd), humic acids (HA) and fulvic acids (FA). In this manner, change in absolute solution concentration and relative percentage for each fraction could be calculated as a function of extraction equilibrium pH. The soils were also analyzed for solid phase total organic carbon and total recoverable metals (EPA Method 3051). Partitioning coefficients were calculated for the metals and organic carbon (OC) based on solid phase concentrations (less the metal or OC removed by the extraction) divided by solution concentrations. Cu and Pb concentrations in solution as a function of extract equilibrium pH are greatest at low and high pH resulting in parabolic desorption/dissolution curves. While processes such as proton competition and proton promoted dissolution can account for high solution metal concentrations at low pH, these processes cannot account for higher Cu and Pb concentrations at high pH. DOC increases with increasing pH, concurrently with the increase in Cu and Pb solution concentrations. While the absolute concentrations of FA and HA generally increase with increasing pH, the relative proportional increase is greatest for HA. Variation in HA concentrations spans three orders of magnitude while FA concentrations vary an order of magnitude over the pH range examined. Correlation analysis strongly suggests that HA plays a major role in increasing the concentration of solution Cu and Pb with increasing pH in the 18 soils studied. The percentage of the OC that was due to FA was nearly constant over a wide pH range although the FA concentration increased with increasing pH and its concentration was greater than that of the HA fraction at lower pH values (pH = 3–5). Thus, in more acidic environments, FA may play a larger role than HA in governing organo-metallic interactions. For Cd, Ni, and Zn, the desorption/dissolution pattern shows high metal solution concentrations at low pH with slight increases in solution concentrations at extremely high pH values (pH>10). The results presented here suggest that the effects of dissolved organic carbon on the mobilization of Cd, Ni, and Zn may only occur in systems governed by very high pH. At high pH, it is difficult to distinguish in this study whether the slightly increased solution-phase concentrations of these cations is due to DOC or hydrolysis reactions. These high pH environments would rarely occur in natural settings. D 2002 Elsevier Science Ltd. All rights reserved. Keywords: Cadmium; Copper; Nickel; Lead; Zinc; pH, humic acid; Fulvic acid; Soil organic matter 1. Introduction There is little doubt that organic matter (OM) plays a significant role in metal behavior in the environment (Schnitzer and Kerndorff, 1981). Many studies have focused on the sorption of metals by solid phase soil organic matter (SPSOM) (Lion et al., 1982; Sanders, 1980; Sauve et al., 2000; Strawn and Sparks, 2000). Genera lly, SPSOM in environmental systems is implicated in retention, decreased mobility, and reduced bioavailability of trace metals. A significant amount of research has examined the role of 0160-4120/02/$ - see front matter D 2002 Elsevier Science Ltd. All rights reserved. PII: S 0160-4120(02)00065-X $ This manuscript has not been subjected to internal review by the US Environmental Protections Agency. Therefore, the research results presented herein do not, necessarily, reflect Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. * Corresponding author. Tel.: +302-831-8449; fax: +302-831-3640. E-mail address: allen@ce.udel.edu (H.E. Allen). 1 Current address: USEPA-National Risk Management Research Laboratory, 26 W. Martin Luther King Drive, Cincinnati, OH 45268, USA. 2 Current address: Connecticut Agricultural Experiment Station, Department of Soil and Water, New Haven, CT 06511, USA. 3 Current address: Gradient, 238 Main Street, Cambridge, Massachu- setts 02142, USA. www.elsevier.com/locate/envint $ Environment International 28 (2002) 401 –410 dissolved organic matter (DOM) on the behavior of trace metals. Studies on humic acid (HA) metal relationships have provided information on a wide range of subjects including sorption characteristics (Spark et al., 1997a), kinetics (Boni- fazi et al., 1996), stability constants (Pandey et al., 2000), bonding mechanisms (Frenkel et al., 2000) and modeling (Liu and Gonzalez, 2000; Robertson and Leckie, 1999). Studies regarding fulvic acid (FA)–metal systems include research on binding models (Christensen et a l., 1998a; Leenheer et al., 1998), competitive effects (Mandal et al., 2000), sorption (Schnitzer and Kerndorff, 1981), binding strength (Brown et al., 1999; Sekaly et al., 1999),and stability constants (Schnitzer and Skinner, 1966, 1967). Research has also been performed on the sorption of organic molecules onto representative soil solids such as kaolin (Huang and Yang, 1995) and metal oxides (Spark et al., 1997a). Many researchers have employed increasingly com- plex systems to study the reactions between metals, organic molecules and soil solids including illite–FA–Cu 2+ (Du et al., 1999), kaolin–HA/FA– Cu 2+ (Huang and Yang, 1995), montmorillonite–HA–Cd 2+ /Cu 2+ /Pb 2+ (Liu, 1999) ,and kaolin–HA –Co 2+ /Cu 2+ /Zn 2+ (Spark et al., 1997b). Other researchers have studied these interactions in complex systems such as soils (Jordan et al., 1997; Temminghoff et al., 1997, 1998), raw sewage (Kunz and Jardim, 2000), composts (Hsu and Lo, 2000), lake waters (Xue and Sigg, 1999), biosolids (Han and Thompson, 1999), and estuaries (Alberts and Filip, 1998). Multiple studies on the effects of pH on the behavior of SOM consistently show that the solubility of SOM increases with increasing pH (Andersson et al., 2000; Erich and Trusty, 1997; Karlik, 1995; You et al., 1999). Shen (1999) examined the sorption of DOM onto soil solids and found that DOM sorption reached a max- imum at pH 4–5 with a decrease in DOM sorption with further increases in pH. Some researchers have also examined the effects of pH on the nature of DOM. Karlik (1995) found an increase in humic compounds (as defined by separation on a XAD-2 resin) wi th increasing pH. Temminghoff et al. (1994) found that the humic/fulvic ratio (as defined by size separation in 0.0033 M Ca(NO 3 ) 2 extracts of a sandy soil) increases from pH 4.4 to pH 5.7. Erich and Trusty (1997) found changes in fluorescence emissions and wavelengths in DOM from forest soil samples with increased lime applications. Ander- sson et al. (2000) found an increase in refractory hydro- phobic acids with increasing lime applications on more humus. Some researchers have applied the relationships between increasing DOM with increasing pH to the mobilization and speciation of metals in environmental systems. Much of this research focuses on Cu. Temminghoff et al. (1997) found increased Cu mobility at both low and neutral pH values in a Cu contaminated sandy soil. They found that at pH 3.9, only 30% of Cu in solution was bound by DOC, whereas 99% of the Cu was bound by DOC at pH 6.6. Strobel et al. (2000) showed that Cu mobilization in a forest soil was related to both pH and DOC whereas Cd mobilization was related solely to system pH. Naidu and Harter (1998) linked Cd mobility to organic ligands. Increasing equilibration time and temperature reduced the mobility of Cd caused by organic matter (Almas et al., 1999). Jordan et al. (1997) examined the increased mobility of Pb in the presence of natural organic matter in a sandy soil. They found that peat humic acids had a higher binding affinity for Pb than peat fulvic acids. They also illustrated the decreased binding of Pb to the sandy soil when DOM was present in column and batch sorption studies. Research on the relationship between metals and DOM show that organic molecules are, in many instances, respon- sible for the increased mobility of meta ls in soils. This has been shown for Pb (Jordan et al., 1997) and Cu (Temmingh- off et al., 1997). Hsu and Lo (2000) demonstrated an increase in solut ion Cu with increasing pH and a concurrent increase in DOM. This increase of DOM at higher pH values (pH>8) did not result in increased soluble concen- trations of Zn and Mn. These results for Cu contradict the general notion that metal sorption increases at higher pH. This result is also of critical importance for a myriad of environmental issues involving trace metal contamination including site remediation, modeling, and risk assessment. Several studies have also included the effects of Ca on the binding and dissolution behavior of OM in soils and aquatic systems. Mandal et al. (2000) examined the effects of Ca and Mg on Ni binding by FA. They concluded that bound Ni tends to be relea sed in the form of Ni 2+ ion in the presence of Ca and Mg as these ions out-compete Ni for binding sites. Curtin et al. (1998) identified soil organic matter as the major source of Ca preferring sites in smectitic Canadian prairie soils. Romkens and Dolfing (1998) found that Ca additions precipitated high molecular weight acids and that Cu co-precipitated with these acids. Temminghoff et al. (1998) found that Ca as well as pH affects Cu mobility in a Cu contaminated sandy soil. The research presented here is unique because of the number and variety of soils involved in the desorption/ partitioning experiments. The met hodology employed for fractionation and characterization of the organic matter allows comparisons to be made between soils both rich and poor in solid phase organic matter. We have not made additions to the soils; neither trace metal concentrations nor organic matter contents have been altered. This work con- tributes to the mounting evidence that illustrates the impor- tance of the role of organic matter in metal mobility in the environment. This research focuses on desorption of organic carbon (DOC) and trace metals from the solid phase to solution in 18 Dutch soils. The main objectives of this study are to (1) fractionate the DOC in water extractions of the soils, (2) examine the concentrations of the operationally defined DOC fraction s in the extract s as a function of systempH,and(3)correlateCd,Cu,Ni,PbandZn concentrations in the extracts with the DOC fractions, as a function of system pH. The experimental findings presented C.A. Impellitteri et al. / Environment International 28 (2002) 401–410402 in this work build on existing knowledge by further eluci- dating more precisely the relations between OM fractions and metal mobility in a wide variety of unspiked soils. 2. Materials and methods 2.1. Definitions Soluble throughout this work refers to the constituents passing through a 0.45-Am cellulose fiber membrane filter (Fisher Scientific, Fairlawn, NJ). DOC fractions are opera- tionally defined as hydrophilic acids (Hyd), humic acids (HA), and fulvic acids (FA). DOC fractions were analyzed for their carbon (C) content by a TOC analyzer (DC 190, Rosemount Analytical, Dohrmann Division, Santa Clara, CA). Solid phase organic carbon (SPOC) was calculated from total carbon measurements using a boat sampler for the analysis of solids. An unamended portion of each soil sample to be assessed for SPOC was analyzed for total carbon (TC unacidified ). Another portion of the sample was then thoroughly mixed with 0.5 M HCl (1:1 w/v), equili- brated for 24 h, and the liquid evaporated under N 2 . The dried sample was then re-homogenized and analyzed for total carbon (TC acidified ). The SPOC was then calculated as: SPOC ¼ TC unacidified À TC acidified ð1Þ where all units are mg C/kg soil. 2.2. Soil metal extractions Eighteen soils from the Netherlands were employed in this study. Table 1 provides summary statistics for some key parameters for all of the soils and information for two of the soils for which detailed results will be presented. Further information regarding the soil characteristics can be found elsewhere (Janssen et al., 1997). Each soil was extracted in a 35-ml polypropylene round-bottom centrifuge tube. Mix- tures (3 g:30 ml) of soil and de-ionized (DI), ultrapure (18.3 MV cm) H 2 O were equilibrated at a minimum of 4 pH levels (pH here is defined as the pH after the 24-h extraction period). Trace metal grade HNO 3 and NaOH (Fisher Scien- tific) were used to adjust the pH of the mixture. The samples were then place d on an orbital shaker (60 rpm) and equilibrated for 24 h. The samples were then centrifuged, filtered (0.45 Am), tested for pH (equilibrium pH values ranged from 3 to 9), and analyzed for metals by inductively coupled argon plasma-optical emission spectrometry (ICP, Spectro Analytical Instruments, Kleve, Germany). We used EPA Method 3051 (USEPA, 1997) for the microwave digestion of soil using HNO 3 for the deter mi- nation of total recoverable metals. The digestates were analyzed by ICP. 2.3. Soil organic carbon extraction and fractionation The 18 soils were then extra cted exactly as they were for metal analyses at various pH values in glass centrifuge tubes at a ratio of 1 g soil/10 ml H 2 O. All soils were extracted at three different pH values. Two soils (Budel and Callant- soog) were extracted at five different pH values. These particular soils were chosen for more in depth study because of the relative ease of pH manipulation. Budel is a sandy humic soil from land around a Zn factory and Callantsoog is a sandy, humus poor soil from a shooting range (this sample contains a high concentration of Pb originating from leaded bullets). After filtration and centrifugation, the solution samples were analyzed for DOC and prepared for fractio- nation into operationally defined Hyd, HA, and FA frac- tions. The fractionation procedure is modeled after OM fractionation procedures using macroreticular resins (Aiken and Leenheer, 1993; Christensen et al., 1998b; Leenheer and Huffman, 1976; Malcom et al., 1994; Thurman and Mal- colm, 1981). The procedure presented in this work modifies existing procedures by employing the resin as a tool to concentrate soluble organic molecules as well as separate them. Known volumes of acidified (pH = 2) soil extracts were passed through a Supelite DAX-8 resin (Sigma- Aldrich, St. Louis, MO) contained in a low-pressure liquid chromatography column (Sigma-Aldrich). The resin bed volume (BV) was 8 ml and the loading rate for the extracts was 8 –10 BV/h. Sample passing through the resin con- tained the operationally defined Hyd fraction. After loading, the column was back eluted with 0.1 M NaOH at a rate of 2 BV/h. The eluate was collected in 10 or 25 ml volumetric Table 1 Selected soil characteristics for 18 Dutch soils examined in this study Total Cd (mg/kg) Total Cu (mg/kg) Total Ni (mg/kg) Total Pb (mg/kg) Total Zn (mg/kg) pH Solid phase OC (mg/kg) Average 5.4 34.5 14.0 120 446 6.2 28,500 Median 2.4 22.6 14.2 64.0 137 6.6 20,600 Standard deviation 9.8 40.4 12.0 154 757 1.3 19,200 Minimum 0.1 0.6 0.7 12.1 7.3 4.2 10,400 Maximum 42.6 136 40.9 679 3120 7.7 70,700 Budel 0.6 3.2 0.9 13.0 14.9 4.3 16,500 Callanstsoog 0.1 0.6 0.7 105 7.3 5.1 11,800 Total metals are from HNO 3 digestions (EPA Method 3051; USEPA, 1997). The pH values are from 1:1 (w/v) soil-deionized H 2 O slurries. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410 403 flasks, sub-sampled and acidified to pH < 1. Samples wer e then refrigerated for 24 h with intermittent agitation. At the end of the 24-h period, samples were centrifuged (at 25 jC). The supernatant was sampled and analyzed for TOC as the operation ally defined FA fraction. HA-C was calculated from the equation: HA À C ¼ DOC total ÀðFA À C þ Hyd À CÞð2Þ where DOC total refers to the original, unfractionated sample. All units are related to the mass of solid extracted (mg/kg). Analyses were performed on solutions with defined amounts of each operationally defined fraction in order to assess recoveries. 2.4. Data analysis Soluble metal (as a function of pH) was modeled by means of parabolic equations. This allowed estimates of soluble metals at the exact pH values of the OM extractions. Direct metal analyses in each of the operationally defined fractions are of limited value because of the manipulations (especially pH) that are required in the fractionation proce- dure. The concentration of organic carbon and the percent- age of the TOC contained in each fraction of the samples were calculated as a function of pH. Correlative analyses were then performed between the various fractions and the concentrations of soluble Cd, Cu, Ni, Pb, and Zn. For some analyses (Figs. 6b,d and 7), data from water extractio ns with Fig. 1. H 2 O extractable metal as a function of pH for the Budel soil. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410404 equilibrium pH between 4 and 9 were exclusively used, as pH values outside of this range are uncommon in most environmental settings. 3. Results and discussion 3.1. Metals in variable pH extractions Fig. 1 shows the results of metal solubilization as a function of pH for the Budel soil. Similar results are observed for the remaining 17 soils. It is noteworthy that approx- imately 160 times more Zn than Cu is extracted at the lowest pH in the Budel soil despite the fact that total Zn and total Cu differ by no more than a factor of 4.5. Six times more Cd than Cu is extracted at low pH though total Cd concentration is substantially less than six times the total Cu concentration. The percentages of each metal extracted in the Budel soil at low pH are Cd-80%, Cu-2.5%, Ni-22%, Pb-9.2%, and Zn- 87%. This suggests that even at lower pH values, stronger binding of Cu and Pb occurs relative to the binding of Cd, Ni, and Zn by solid constituents. Our results indicate that proton competition and/or proton promoted dissolution greatly affect the amount of Cd, Ni, and Zn in solution, while the effect on Cu and Pb is less significant. Cu and Pb in these soils may be strongly bound to solid forms of OM and, in low pH environments, have a greater association with the solid phase relative to Cd, Ni, and Zn. A significant increase in extractable Cu and Pb occurs as the pH increases from 5 to 9. This increase in water extractable metal as a function of pH is much less significant for Cd, Ni, and Zn. Similar results have been published elsewhere (Hsu and Lo, 2000) for Cu, Mn, and Zn in compost extracts and for Cd and Cu in soils (Salam and Helmke, 1998). The increase in solubility of Cu and Pb could be due to the pH-induced solubilization of organic matter. This would have important implications for many natural and engineered syst ems. For example, increasing pH in wastewater treatment to precipitate compounds could actually increase Cu in solution by increasing the solubility of OM. Released organic molecules could transport metals through soils systems to ground and surface waters. Depend- ing on the chemistr y of the receiving waters, a significant portion of the metals could disassociate from the organic molecules and become more biologically active. Fig. 2. Dissolved soil organic carbon (DOC) in the fractions (FA, HA and Hyd) as a function of pH (a—Budel, c—Callantsoog). The log scale on the y-axis for (a) and (c) emphasizes the increase in HA as a function of pH. (b) (Budel) and (d) (Callantsoog) show percent distribution of the three fractions as a function of the DOC solubilized at the 24-h equilibrium pH. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410 405 3.2. Hyd, HA, and FA in solution We tested the DOC fractionation procedure with isolated FA, HA and Hyd from well-characterized solutions of organic matter isolates using the procedure developed in this laboratory (Impellitteri, 2000; Lu, 2000). Recoveries of the isolated FA, HA, and Hyd fractions were 100%, 107%, and 102%, respectively. Fig. 2 shows the total concentrations and proportions of each of the DOC fractions in solution as a function of pH for two of the soils (Budel and Callantsoog). For seven of the soils extracted at only three pH values, HA was undetected at the lowest pH value for the conditions used in this study. Both HA and FA concentrations increase with increasing solution pH. These results have been shown elsewher e for US coastal plain soils (You et al., 1999). The concentration of the Hyd fraction remains relatively constant as a function of pH. The percentage of the total DOC that is HA increases markedly with increasing pH, while the FA fraction per- centage remains more constant as a function of pH, and the relative percentage of the Hyd fraction decreases with increasing pH. The y-axes in Fig. 2a and c are in log form and thus show the striking increase in soluble HA as a function of pH. For the 18 soils studied, the increase in soluble HA in the pH range of 3 to 9 was between two and three orders of magnitude. The increase in soluble FA typically remained within an order of magnitude. Fig. 3 shows the partitioning of each fraction as a function of pH. Fig. 3. Partitioning of each OC fraction as a function of pH for the (a) Budel and (b) Callantsoog soils. K d-OC is based on SPOC (less the TOC in the fraction) divided by TOC in each fraction at a particular pH. Fig. 4. Partitioning of each OC fraction as a function of pH for all soils. K d-OC is SPOC (less the fraction TOC) divided by TOC in each fraction at a particular pH. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410406 The partitioning values , K d-OC , are calculated based on the equation: K dÀOC ¼ðSPOC À TOC all fractions Þ=TOC fraction ð3Þ where SPOC (mg/kg) represents the solid phase organic carbon, TOC all f raction s (mg/kg) is the sum of the OC removed during the extraction and TOC fraction (mg/kg) is the TOC in each operationally defined fraction. The results show a linear dependency of the log K d-OC values for HA with equilibrium pH with a strong correlation. Strong correlations also exist for the partitioning of FA; however, there is little dependency of the partitioning of Hyd on pH. The results for these two soils are representative of the other 16 soils tested. For log K d-OC Hyd values vs. pH in all 16 soils, the average R 2 = 0.44 (s = 0.42), for log K d-OC FA average R 2 = 0.91 (s = 0.19), and for log K d-OC HA average R 2 = 0.89 (s = 0.20). Thus, within a particular soil, partition- ing of FA and HA correlates well with pH. This is in contrast to the data presented in Fig. 4. Here, all of the data are combined for all of the soils. The results in Fig. 4 illustrate poor correlation between partitioning of all of the operationally defined fracti ons and pH. This may be explained by the fact that the slope values of the regression lines for partitioning of log K d-OC values vs. pH vary among individual soils. This suggests that the factors governing the dissolution/desorption of organic carbon from solid to solution vary in these soils. This should be expected in this Fig. 5. Partitioning of metal in the Dutch soils vs. partitioning of soil organic carbon. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410 407 sample set of 18 soils, factors such as clay (type and concentration), metal-oxide content, and presence/absence of Ca species will all affect the behavior of organic carbon partitioning. It may also be possible that variations in solid phase carbon source material (e.g. soot vs. leaf litter) may play a role in governing the nature of DOC in soils. 3.3. DOC partitioning and metal partitioning Log K d values for all metals (with the exception of Zn) do not correlate with extraction equilibrium pH when the data for all soils are combined (Cd-R 2 = 0.14, Cu-R 2 = 0.07, Ni-R 2 = 0.05, Pb-R 2 = 0.01, Zn-R 2 = 0.30). Fig. 5 shows the correlations between log K d values for all the metals and log K d-OC where K d-OC is defined as in Eq. (3). This data shows a relationship between the partitioning of Cu and OC in these soils. Evidence also exists for a relationship between the partitioning of Pb and OC. This relationship has been shown previously for Cu (Temminghoff et al., 1994, 1997, 1998; Yin et al., 2002). Dissolved natural organic matter has also been implicated in preventing the sorption of Pb onto a sandy soil (Jordan et al., 1997). Further analysis of the data reveal that the presence of Cu and Pb in the water extracts can be more precisely asso- ciated with the operationally defined HA fraction. Table 2 shows R 2 values for log K d of all metals vs. log K d-OC for all of the DOC fractions. The data set was further restricted to extracts with equilibrium pH values between 4 and 9. These pH values are more environmentally realistic, though in some situations with significant anthropogenic disturbances, pH extremes may occur. The R 2 values for this restricted range of pH are shown in parentheses in Table 2. Data for Cu and Pb are shown graphically in Fig. 6. Fig. 7 shows the same data as in Fig. 6b and d with the K d-OC values normalized by total Ca concentrations in the soils. With all else being equal (pH, cation concentrations, etc.), increased Ca concentrations in soils will aid in the floccu- lation of HA from solution causing increased K d-OC values Table 2 R 2 values for log K d values for all metals vs. log K d-OC values for the three operationally defined DOC fractions across the entire pH range Log K d-OC À Hyd Log K d-OC À FA Log K d-OC À HA Log K d À Cd 0.0068 (0.00058) 0.020 (0.0027) 0.0022 (0.02) Log K d À Cu 0.028 (0.12) 0.0025 (0.20) 0.21 (0.62) Log K d À Ni 0.0023 (0.037) 0.0017 (0.02) 0.071 (0.14) Log K d À Pb 0.0029 (0.13) 0.014 (0.16) 0.30 (0.46) Log K d À Zn 0.023 (0.017) 0.044 (0.044) 0.021 (0.018) The data sets for each comparison were also restricted by eliminating extracts with equilibrium pH values less than 4 or greater than 9. R 2 values for the restricted data sets are in parentheses. Fig. 6. Partitioning of Cu and Pb vs. partitioning of HA for all water extracts from all soils where HA was quantified in the extracts. (b) and (d) show the same data for water extracts where 4 < pH < 9. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410408 for HA. The partitioning of Cu and Pb in the water ex- tractions is clearly related to the partitioning of HA in these soils. For Cu, these results build on results generated by other researchers (Temminghoff et al., 1994, 1997, 1998). The results for Pb shown in Fig. 6 are in agreement with work performed by Jordan et al. (1997) . They concluded that HA (peat derived) had a higher affinity than FA for Pb and prevented the binding of Pb by a sandy soil. Here, the data strongly suggest that the incre ased desorption/dissolu- tion of HA is strongly correlated to the increase of the concentration of Pb in solution. As both Cu and Pb form strong complexes with HA (Tipping, 1994), it is likely that the correlation of increased soluble Cu and Pb with increased desorption of HA is a consequence of the com- plexation of these metals by HA. 4. Conclusions DOC increases with increasing pH. The fractionation of DOC desorbed/dissolved from 18 Dutc h soils shows that the largest relative increase occurs for the operationally defined HA fraction. The Hyd fraction percentage generally de- creases with increasing extraction equilibrium pH. The percent FA tends to remain constant over the range of pH values. Within soils, the partitioning of FA and HA corre- lates with the system equilibrium pH. The partitioning of FA and HA as a function of pH varies widely for the soils studied. The K d-OC values for FA and HA tend to decrease (more OC associated with solution phase) with increasing system pH. Correlative studies provide indication that HA is capable of transporting Cu and Pb into solution upon desorption/dissolution from the solid phase. Ca may antag- onize the solubilization of metals seques tered by HA by flocculation of the HA –metal complex. The partitioning of Cd, Ni, and Zn could be affected by the partitioning of OC in these soils at very high pH (>10) though in this study, it is impossible to distinguish the effect of OC partitioning and hydrolysis reactions. This study simulates a situation in the environment where soil systems are ‘‘titrated’’ with acid or base in the form of atmospheric precipitation or soil amend ment. The soil solution is considered to be in a state of pseudo- equilibrium for this study and thus the pH of the soil – solution mixture in nature is of critical importance. We presume that the pseudo-equilibrium pH is a master variable governing the desorption/dissolution of organic molecules. Metals transported by organic molecules into surface and/or groundwaters will undergo further reactions depending on the chemistry of the receiving waters. Acknowledgements The International Copper Association, the International Lead–Zinc Research Organization, and the United States Environmental Protection Agency funded this work. References Aiken G, Leenheer J. Isolation and chemical characterization of dissolved and colloidal organic matter. Chem Ecol 1993;8:135– 51. Alberts JJ, Filip Z. Metal binding in estuarine humic and fulvic acids: FTIR analysis of humic acid– metal complexes. Environ Technol 1998;19: 923 –31. Almas A, Singh BR, Salbu B. Mobility of cadmium-109 and zinc-65 in soil inf luenced by equilibration time, temperature and organic matter. J Environ Qual 1999;28:1742 –50. Andersson S, Nilsson SI, Saetre P. Leaching of dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) in more humus as af- fected by temperature and pH. Soil Biol Biochem 2000;32:1 –10. Bonifazi M, Pant BC, Langford CH. Kinetic study of the speciation of copper(II) bound to humic acid. Environ Technol 1996;17:885– 90. Brown GK, MacCarthy P, Leenheer JA. Simultaneous determination of Ca, Cu, Ni, Zn and Cd binding strengths with fulvic acid fracti ons by Schubert’s method. Anal Chim Acta 1999;402:169– 81. Christensen JB, Tipping E, Kinniburgh DG, Gron C, Christensen TH. Pro- ton binding by groundwater fulvic acids of different age, origins, and structure modeled with the Model V and NICA-Donnan Model. Environ Sci Technol 1998a;32:3346– 55. Fig. 7. Partitioning of Cu and Pb as a function of the Ca-normalized partitioning of HA. If the two outliers shown as filled-in symbols are removed in (b), R 2 = 0.85. Data are for water extracts where 4 < pH < 9. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410 409 Christensen JB, Jensen DL, Gron C, Filip Z, Christensen TH. Character- ization of the dissolved organic carbon in landfill leachate polluted groundwater. Water Res 1998b;32:125– 35. Curtin D, Selles F, Steppuhn H. Estimating calcium– magnesium selectivity in smectitic soils from organic matter and texture. Soil Sci Soc Am J 1998;62:1280– 5. Du Q, Sun Z, Forsling W, Tang H. Complexation in illite-fulvic acid-Cu 2+ systems. Water Res 1999;33:693 –706. Erich MS, Trusty GM. Chemical characterization of dissolved organic mat- ter released by limed and unlimed forest soil horizons. Can J Soil Sci 1997;77:405 – 13. Frenkel AI, Korshin GV, Ankudinov AL. XANES study of Cu 2+ -bind- ing sites in aquatic humic substances. Environ Sci Technol 2000;34: 2138 –42. Han N, Thompson ML. Copper-binding ability of dissolved organic matter derived from anaerobically digested biosolids. J Environ Qual 1999;28: 939 –44. Hsu JH, Lo SL. Characterization and extractability of copper, manganese, and zinc in swine manure composts. J Environ Qual 2000;29:447 –53. Huang C, Yang YL. Adsorption characteristics of Cu(II) on humus– kaolin complexes. Water Res 1995;29:2455– 60. Impellitteri CA. Predicting partitioning of cadmium, copper, lead, nickel, zinc, and soil organic matter in soil solution from soil parameters. PhD Thesis, University of Delaware, Newark; 2000. Janssen RPT, Peijnenburg WJGM, Posthuma L, Hoop MAGTvd. Equili- brium partitioning of heavy metals in Dutch field soils: I. Relationship between metal partition coefficients and soil characteristic s. Environ Toxicol Chem 1997;16:2470 – 8. Jordan RN, Yonge DR, Hathhorn WE. Enhanced mobility of Pb in the presence of dissolved natural organic matter. J Contam Hydrol 1997; 29:59 –80. Karlik B. Liming effect on dissolved organic matter leaching. Water Air Soil Pollut 1995;85:949 –54. Kunz A, Jardim WF. Complexation and adsorption of copper in raw sew- age. Water Res 2000;34:2061 – 8. Leenheer JA, Huffman EWDJ. Classification of organic solutes in water by using macroreticular resins. J Res US Geol Surv 1976;4:737 –51. Leenheer JA, Brown GK, Maccarth y P, Cabaniss SE. Models of me tal binding structures in fulvic acid from the Suwanee River, Georgia . Environ Sci Technol 1998;32:2410 –6. Lion LW, Altmann RS, Leckie JO. Trace metal adsorption characteristics of estuarine particulate matter: evaluation of contributions of Fe/Mn oxide and organic surface coatings. Environ Sci Technol 1982;16: 660 –6. Liu A. Adsorption/desorption in a system consisting of humic acid, heavy metals, and clay minerals. J Colloid Interface Sci 1999;218:225 –32. Liu A, Gonzalez RD. Modeling adsorption of copper(II), cadmium(II) and lead(II) on purified humic acid. Langmuir 2000;16:3902 –9. Lu Y. Copper complexation with natural dissolved organic matter and partitioning onto suspended particulate matter in river waters. PhD Thesis, University of Delaware, Newark; 2000. Malcom RL, Aiken GR, Bowles EC, Malcolm JD. Isolation of fulvic and humic acids from the Suwanee River. Humic substances in the Suwanee river, Georgia: interactions, proper and proposed structures. Water Sup- ply Pap 1994;2373:13 –9. Mandal R, Salam MSA, Murimboh J, Hassan NM, Chakrabarti CL, Back MH, et al. Competition of Ca(II) and Mg(II) with Ni(II) for binding by a well-characterized fulvic acid in model solutions. Environ Sci Technol 2000;34:2201– 8. Naidu R, Harter RD. Effect of different organic ligands on cadmium sorption by and extractability from soils. Soil Sci Soc Am J 1998;62:644 –50. Pandey AK, Pandey SD, Misra V. Stability constants of metal– humic acid complexes and its role in environmental detoxification. Ecotoxicol En- viron Saf 2000;47:195 –200. Robertson AP, Leckie JO. Acid/base, copper binding, and Cu 2+ /H + ex- change properties of a soil humic acid, and experimental and modeling study. Environ Sci Technol 1999;33:786 –95. Romkens PFAM, Dolfing J. Effect of Ca on the solubility and molecular size distr ibution of DOC and Cu binding in so il solution samples. Environ Sci Technol 1998;32:363 –9. Salam AK, Helmke PA. The pH dependence of free ionic activities and total dissolved concentrations of copper and cadmium in soil solution. Geo- derma 1998;83:281 –91. Sanders JR. The use of adsorption equations to describe copper complexing by humified organic matter. J Soil Sci 1980;31:633 –41. Sauve S, Hendershot W, Allen HE. Solid-solution partitioning of metals in contaminated soils: dependence on pH, total metal burden, and organic matter. Environ Sci Technol 2000;34:1125– 31. Schnitzer M, Kerndorff H. Reactions of fulvic acids with metal ions. Water Air Soil Pollut 1981;15:97 –108. Schnitzer M, Skinner SIM. Organo-metallic interactions in soils: 5. Stability constants of Cu, Fe, and Zn– fulvic acid complexes. Soil Sci 1966;102: 361 –5. Schnitzer M, Skinner SIM. Organo-metallic interactions in soils: 7. Stability constants of Pb, Ni, Mn, Co, Ca, and Mg –fulvic acid complexes. Soil Sci 1967;103:247 –52. Sekaly ALR, Mandal R, Hassan NM, Murimboh J, Chakrabarti CL, Back MH, et al. Effect of metal/fulvic acid mole ratios on the binding of Ni(II) , Pb(II), Cu(II), Cd(II), and Al(III) by two well-c haracterized fulvic acids in aqueous model solutions. Anal Chim Acta 1999;402: 211 –21. Shen YH. Sorption of natural dissolved organic matter on soil. Chemo- sphere 1999;38:1505 –15. Spark KM, Wells JD, Johnson BB. Characteristics of the sorption of humic acid by soil minerals. Aust J Soil Res 1997a;35:103– 12. Spark KM, Wells JD, Johnson BB. Sorption of heavy metals by mineral– humic acid substrates. Aust J Soil Res 1997b;35:113– 22. Strawn DG, Sparks DL. Effects of soil organic matter on the kinetics and mechanisms of Pb(II) sorption and desorption in soil. Soil Sci Soc Am J 2000;64:144 – 56. Strobel BW, Hansen HCB, Borggaard OK, Andersen MK, Raulund-Ras- mussen K. Cadmium and copper release kinetics in relation to affores- tation of cultivated soil. Geochim Cosmochim Acta 2000;65:1233 –42. Temminghoff EJM, Zee SEATMVD, Keizer MG. The influence of pH on the desorption and speciation of copper in a sandy soil. Soil Sci 1994; 158:398 –408. Temminghoff EJM, Zee SEATMVD, Haan FAMD. Copper mobility in a copper-contaminated sandy soil as affected by pH and solid and dis- solved organic matter. Environ Sci Technol 1997;31:1109– 15. Temminghoff EJM, Zee SEATMVD, Haan FAMD. Effects of dissolved organic matter on the mobility of copper in a contaminated sandy soil. Eur J Soil Sci 1998;49:617– 28. Thurman EM, Malcolm RL. Preparative isolation of aquatic humic sub- stances. Environ Sci Technol 1981;15:463 –6. Tipping E. WHAM—a chemical equilibrium model and computer code for waters, sediments and soils incorporating a discrete-site/electrostatic model of ion binding by humic substances. Comput Geosci 1994;20: 973 –1023. USEPA. Method 3051: microwave assisted acid dissolution of sediments, sludges, soils, and oils. SW846 Test Methods for Evaluating Soli d Waste, Physical/Chemical Methods 2nd ed. U.S. Environmental Protec- tion Agency, Washington, DC, 1997. Xue H, Sigg L. Comparison of the complexation of Cu and Cd by humic or fulvic acids and by ligands observed in lake waters. Aquat Geochem 1999;5:313 –35. Yin Y, Impellitteri CA, You SJ, Allen HE. The importance of organic matter distribution and extract soil:solution ratio on the desorption of heavy metals from soils. Sci Total Environ 2002;287:107 –19. You SJ, Yin Y, Allen HE. Partitioning of organic matter in soils: effects of pH and water/soil ratio. Sci Total Environ 1999;227:155 –60. C.A. Impellitteri et al. / Environment International 28 (2002) 401–410410 . Correlation of the partitioning of dissolved organic matter fractions with the desorption of Cd, Cu, Ni, Pb and Zn from 18. constant over the range of pH values. Within soils, the partitioning of FA and HA corre- lates with the system equilibrium pH. The partitioning of FA and HA