- Email: [email protected]
Chemical speciation of Cd, Cu, Pb, and Zn in mixed freshwater, seawater, and brine solutions
Geochimxa
et Cosmochimica
Acla.1977,
Vol. 41, pp. I183 to 1191.
Pergamon Press. Printed inGreatBritain
Chemical speciation of Cd, Cu, Pb, and Zn in mixed freshwater, seawater, and brine solutions Department
DAVID T. LONG* and ERNESTE. ANGINO of Geology, University of Kansas, Lawrence, KS 66045, U.S.A.
(Received 13 January 1916; accepted in revised form 18 April 1977)
Abstract-A theoretical model was developed to study the chemical speciation of the trace elements Zn, Cd, Cu and Pb aqueous solutions and their responses to variations in ionic strength and complexation. Two mixing solutions were investigated, a freshwater-seawater system and a freshwater-brine system. The brine was a calcium, sodium-chloride solution with a molal ionic strength of two. Trace element associations with the ligands OH-, Cl-, CO:-, SO:-, and HCO; were considered at pHs from 3.5 to 11.0 at 25°C. In general, the relative importance of the various ligand-trace element complexes can be predicted from a comparison of their stability constants. However, the effect of pH on the importance of a given complex is not readily apparent from the stability constants. Freshwaterseawater mixtures, as might be found in a totally mixed estuary, show that seawater composition is the dominant control on chemical complexing. Chloride complexing is similar for lead and zinc in the freshwater-brine mixtures. This similarity may account in part for the association of lead and zinc in strata-bound ore deposits.
INTRODUCTION VARIOUStheoretical
approaches have been proposed for determining the speciation of specific trace elements in different natural water systems. For details, one can refer to the works of WHITE et al. (1958), GARRELSand THOMPSON(1962), MOREL and MORGAN (1972), ELDER (1975), WHITFIELD (1975) and WOOD (1975). Several proposed models show the changes in chemical speciation that accompany changes in ionic
strength or in the concentrations of the individual ions. Evaluation of these models usually involves simultaneous solution of several mass balance and complexation equilibria using stability constant data. Resulting diagrams usually depict the species distribution in a particular aqueous solution. It is definitely of interest to know how the chemical speciation of elements, particularly trace elements, responds to chemical changes caused by the mixing of fresh and saline waters, such as river water and seawater., The mixtures can be studied by comparing a sequence of diagrams of species distributions resulting from proportional chemical changes between two end member solutions. In these mixing systems, changes in the trace element-ligand concentrations during the mixing process affect the trace element speciation. Since sorption reactions also compete for the trace elements, changes in speciation also affect the amount of trace elements that participate in sorption. The aim of this study was to compute changes in the distribution of aqueous species of Cu, Cd, Zn, and Pb resulting from the mixing of different aqueous solutions. Two mixing systems were studied, freshwater-seawater and freshwater-brine (brine is defined * Present address: Department of Geosciences, North Carolina State University, Raleigh, NC 27607, U.S.A.
here as a solution with greater than 100,000 ppm total dissolved solids). For this study a calcium, sodiumchloride brine was used. The freshwater-seawater system was chosen to provide insight into the chemical changes that might be expected in a totally mixed estuarine environment. Reactions in the freshwater-brine system, on the other hand, may show similarities in the chemical behaviour of certain trace elements during the formation of strata-bound ore deposits. Any similarities found among trace elements may help in the interpretation of the origin of strata-bound ore deposits. MIXING SYSTEMS Mixing systems were set up using the end member concentrations for freshwater, seawater and brine shown in Table 1. Various definitions exist for the term freshwater; consequently, its exact meaning is ambiguous. Usually the term freshwater implies an aqueous solution dilute with respect to the concentrations of the various ions (e.g. river water, ground water, rain water). In this study the term freshwater was defined as one in which the total dissolved solids content was less than 1000 mg/l. Such solutions would correspond roughly to those with ionic strength less than 0.5 mol. kg- ‘. Major and trace element speciation was calculated for the five mixtures defined as 100% seawater, 75-25 (percent seawater-percent freshwater), 50-50, 25-75, and 100% freshwater as a function of pH. The pH was varied from 3.5 to 11.0 at 25°C using 0.5 pH increments. Similar mixtures were studied for the freshwater-brine system. The calculations used to determine the chemical speciation of Cd, Cu, Pb, and Zn in the various aqueous solutions are similar to those of ZIRINO and
1183 G.C.A. 4119-A
D. T. LANGand E. E. ANGMV
1184
Table 1. Molal concentrations
of chemicai end members used in the calculations
component
Freshwater (3)
seawater (4)
Brine (5)
CalCiUU
0.000375
0.0104
0.0964
Magnesium
0.000168
0.054
0.0223
Sodium
0.000274
0.4752
1.6845
Potassium
0.000059
0.01
a.0079
Chloride
0.00022
0.554
2.0198
Ta ("
0.000955
0.00234
0.0001
Sulfate
0.000955
0.0284
0.0029
I (2)
0.0021
0.661
2.099
(1) Total alkalinity. (2) Effective molai ionic strength (after computation of species distribution). From LIVINGSTONE (1963). (4) From BERNER(1971). (5) From CARPENTER et al. (1974).
YAMAMOTO (1972) for the same trace elements. Complexes between the divalent metals and the ligands Cl-. SO:-. CO:-. HCO,. and OH- of the type “MeL,” were considered. Me and L represent the metal and the ligand, and n equals 1-4. Generally, third and fourth order associations were limited to the halides and hydroxo complexes. Since the metals were assumed to be in low concentrations ( < 10ppm) polynucfear complexes were not considered (HELGESON, 1969) As in all modeb of this type, thermodynamic equilibrium was assumed to exist. Since (1) many trace elements are normally quite undersaturated in natural aqueous systems with respect to the precipitation of minerals and (2) precipitated metal would be unavailable for any sorption reactions, the trace element speciation is computed on the basis of the percentages of the metal concentration that would be complexed and uncomplexed. In this way the absolute trace element concentrations are not needed and the systems studied would be unaffected by precipitation reactions. Competition for the trace eiements by other ligands such as clays or organic matter is not considered. The models presented and discussed are strictly inorganic systems. Two steps were used to calculate the distribution of the chemical species in each aqueous mixture. Step one involved the computation of the species distribution of the major ions by the ion association model of GARRELS and %OMP~ON (1962) using the stability constants shown in Table 2. In step two the chemical speciation of the trace elements was determined. The mass action equations used in calculating both major cations and trace element distributions are: (a) for total metal concentration, WTOTI\l_= (m) + i Ii=1
i i=*
/j+)“(rn)[L(i)]“.
E&E IlmL(i),
(1)
(3)
(b) for percent uncomplexed elemental ion, y&n)
= 1001 !‘i:
1+
i
i
n=l
i=l
J?(i).[L(i)]“.
?Zi) m
(2) 1.
(c) for percent complex r&(i), :
II I 1 i -~ 1
;‘mL(i),l
in moles; j = number of ligand associations; !I = order of association; b(i) = stability constant of ith associations nth order; [L(i)]” = concentration of ith ligand in moles for nth order association; yrn= activity coefficient of cation; yL(i)= activity coefficient of ith ligand; and ymL(i),= activity coefficient of complex. Similar equations were used to calculate the distribution of species of the major anions. Individual activity coefficients were used to calculate the activities of the chemical species and were determined in accordance with the ionic strength of the mixture. Activity coefficients of ions in concentrated solutions are difficult to determine theoretically. No method is generally accepted (TRUESDALEand JONES, 1969). Four approximations can be considered (TRUESDALEand JONES, 1969): (1) the extended Debye-Hiickel equation. where (VI) = concentration
of free cations
where: A = constant relating to solvent; B = constant relating to solvent; 6 = distance of closest approach of ith ion; Zi = charge of the ith ion; and I = true ionic strength; (2) the mean salt method (GARRELS and CHRIST, 1965). This is based on the MacInnes assumption that ?K ’ = yC1F = yKC1; (3) DAVIES’ (1962) modification
of the
Debye-Hiickel
1185
Chemical speciation of Cd, Cu, Pb, and Zn Table 2. Stability constants for trace element species used in the calculations LIGANDS
Pb: log3L1
1.6 (2)
6.2 (6)
lo*
1.78(Z)
10.9 (6)
logs 13
1.68(2)
13.9 (66)
1°gS
1.38(Z)
16.3 (6)
12
14
al: lo& l1
0.0212)
6.37(6)
logs 12
-0.71(2)
14.3 (6)
log5 13
-2.3 (2)
15.0 (6)
log@ 14
-4.6 (6)
16.0 (6)
cd: log!3l1
2.0 (6)
5.0 (6)
loge l2
2.7 (6)
10.6 (61
2.1 (6)
10.0 (6)
l@
13
lor36
14
2.9 (I.)
2.7 (1)
6.3 (3)
2.7 (6)
10.64(3)
3.47(61
5.97(3>
2.25(63
9.83(4)
2.1 (1)
4.1 (5)
2.3 (6)
2.1 (1)
5.3 (1)
2.3 (6)
10.0 (6)
zn: log5 11
0.43(2)
4.4 (6)
log6 12
0.61(2)
12.89(6)
log5 13
0.53(2)
14.0 (6)
log6 14
0 2 (2)
15.0 (6)
Values are the logarithms of the association constants. Numbers in parentheses refer to references for the constants and are listed below: (1) ZIRINOand YARDS (1972). (2) HELGE~~N (1969).(3) ERNESTet al. (1975).(4) ME~~WER and BAES(1975). (5) GARDNER (1974).(6) SILLENand MARTEL(1964).
equation. This equation is (5)
-log~i=O.5Z~(~-O.31),
where the equation parameters are the same the extended Deby~Huckel ~uation [equation (4) HELGES~N’S(1969) trace activity coefficient tion (y*) modified from the Stokes-Robinson tion This equation for 25°C is
as in (4)]. equaequa-
A."2.r112
_log$=
n.__-
1 + 6BI
+ B’I,
(6)
where the equation parameters are the same as in the Debye-Hiickel equation plus B’ which is the deviation function. A plot of activity coefficients as determined by the four methods shows that, at ionic strengths of less than about 0.1, the methods predict similar values. The Davies equation is not considered to be valid when the ionic strength is greater than 0.1, since it does not have an adjustable parameter to describe the size of the ion (TRUESDALE and JONES,1969). The mean salt method was used in this study to calculate the activities of the major ions. The method used by HELGWN (1969) was used for the trace elements. The mean salt activity coefficients for HCO; and CO:-
were taken from the data of WALKER, BRAY and J~HNKIN (1927). Recent experimental work on the activity coefficient of SO:- has shown a close correlation of values between those computed theoretically using the and those extended Deby~Huckel equation measured using lead ion electrodes over the range of ionic strengths studied (~2) (HATHAWAY,1975, personal communication). Therefore, the DebyeHuckel equation was also used for the activity coefficients of SO:-. For the individual activity coefficients of the complexes, the Debye-Huckel equation was used for ionic strengths below 0.1 when the adjustable parameters in the equation were defined for the complex. Above 0.1 or when the adjustable parameters were not defined the values were assigned based on the charge of the complex. Complexes with charges of + 1 or +2 were assigned the activity coefficients determined for (HC03)- and (CO,)‘-, respectively. Neutrally charged complexes were assigned the value of the activity coefficient (gamma) of H,CO~ as calculated from the Setchenow expression (REARDONand LANGMUIR,1974): gamma H2CO(: = antilog 0.06 (I), where I = ionic strength. Recently REARWN and LANGMUW (1976) have suggested using a different value for neu-
1186
D. T. LONGand E. E. ANGINO
tral complexes. The effect of using this new value on the theoretical distribution of chemical species will be discussed later. The data for the activity coefficients of the salts used in the mean salt method were taken from ROBINSONand STOKES(1959) for CaCl*, MgC12, NaCl and KCl. In preliminary runs activities of the ions were not adjusted for the effect of the various species on one another. Species diagrams calculated with activities corrected for salt interactions using Harned’s Rule (BERNER,1971) were compared to the same diagrams calculated using uncorrected salt activities. The results showed that for ionic strengths below 2.0 there was no significant difference between diagrams. Thus, in subsequent calculations, Hamed’s Rule was not used to correct salt activities. Since the four trace elements considered in this study have different outer electron orbital structures from the divalent ions, different chemical speciation would be expected between the elements. Zinc (2+) and cadmium (2f) have compIete 3d and 4d outer orbital structures, respectively. Copper (2 + ) has a 3d” structure, and lead (2+) a complete 6s structure with inner f orbitals. The stability constants for the chemical species can give an idea of the relative importance of a particular complex but it is not readily apparent from these constants at what concentrations or at what pHs a given complex is important.
Figures I and 4 show the results for the mixing of seawater and freshwater in different proportions. The figures are arranged with lOOY$seawater at the bottom and 100yOfreshwater at the top; pH increases to the right on the abcissa and the percentage aqueous species or association is plotted on a log scale on the ordinate. Associations are shown for percentages between 0.1 and 100’~. In order to conserve space, the results from the five mixing ratios are only shown for Zn and Cd. For the other metals only the results of the mixtures of iOOO/,freshwater, 50% freshwater-500/, seawater, and lt’@!/, seawater are shown. Zinc. Zinc speciation (Fig. la, b) in aqueous solution can be characterized by three fields-one at pHs up to about 6.5 (designated the precrossover field) a second covering the pH range 6.5-9.5 (the crossover field): and a third at pH values greater than about 9.5 (the postcrossover field). Similar fields can be recognized on the speciation diagrams of the other trace elements. The precrossover field of zinc is characterized in all the mixtures by the don~~ance of free zinc (Zn”+) and by the invariance of the associations with changes in pH. Chloride becomes an important ligand soon after the addition of small amounts of sea water to the fresh water solution. As would be expected, with greater amounts of chloride in solution, higher order associations become important. Also as the chloride concentration is increased, the difference in magnitude of the various orders of the chloride association
becomes less. The relative importance of the chloride ligands in all the zinc associations is Cl > Cl, > Cl1 > Cl,. The crossover field is characterized by an increase in the complexes of the hydroxo ligands and an increase followed by a decrease in the importance of the carbonato ligand. The carbonato ligand is less important in seawater mixtures than in freshwater. This is undoubtedly due to competition from the chloro complexes. Increasing the amount of seawater in solution moves the crossover field to higher pHs. The postcrossover field is characterized by the dominance of the second order hydroxo ligand at the expense of the other ligands. These results are similar to those of ELDER(1975) who demonstrated the importance of the hydroxo ligand in forming complexes at high pHs. Figure la shows that the speciation of zinc is affected by as little as 25% seawater mixed with freshwater. 100.0
Z”
10046Freshwater
.g loo.0 3 2 3 4” E z k
75%Freshwater 10.0 1.0 0.1 Z”
--,f-
(OH)2
50%Seawater 50%Freshwater
1.0 0.1
Fig. la. Zinc chemical speciation in seawater-fr~hwater mixtures.
25%Freshwater
3.5 4.5 5.5 6.5 7.5 8.5 9.510.511.5 PH
Fig. lb. Zinc chemical speciation in seawater-freshwater mixtures, continued,
1187
Chemical speciation of Cd, Cu. Pb, and Zn Chemicaf speciation of zinc and the other trace elements in mixtures containing smaller amounts of seawater was also studied. Species distribution was affected by additions of as little as loo/, seawater. This behavior would be expected from the mixing of high and low ionic strength solutions. The speciation is controlled by both the absolute and relative concentrations of ligands in the solution. In a near shore environm~t such as an estuary or delta front the pH values expected might be in the range 7.5-8.5. The speciation of zinc to be expected in these environments is demonstrated by the crossover field. The dominant species would be free zinc ion, second-order hydroxo complexes, and first-order chloride complexes. Figures la, b show that slight changes in the pH will affect the species distribution to a significant degree and therefore might affect the ability of the sediments to concentrate zinc. ~~drniurn, Cadmium (2+ ) has au outer orbital structure similar to that of zinc. Figures 2a, b show
lN.OW 2 & $ a E 8 a$
10.0 1.0 0.1
the speciation of cadmium is also similar to zinc, as might be expected. Cadmium, however, is more highly complexed to Cl-. Because of the chloride associations, the effect of adding seawater in small amounts is readily apparent by the rapid increased in Cl complexes and the decrease in all other species. The crossover field narrows with increased additions of seawater. For Cd, the postcrossover field is minor. The complexation of cad~um with chloride further demonstrates the trends noted for zinc; i.e. the increase in complexing order and the decrease in the difference of magnitudes between the orders, with increased chloride concentration. For cadmium, this trend is continued to the point where a reversal in the relative importance of the complexes occurs. This reversal takes place at a mixture of 25% freshwhich seawater in water and 75% CdCl(: > CdCl; > CdCl;. Zn a near shore environm~t, the speciation of cadmium should correspond to that showu in the precrossover and crossover fields. Cadmium’s species distribution would be much simpler than that of zinc. 100% Freshwater One would expect it to be dominated by chloride and relatively unaffected by small changes in pH. The species change, however, during the change from a freshwater to seawater, such as a river entering an ocean, would affect Cd more than Zn. Many of the freshwater zinc species including Zn2+ retain a similar distribution in 100% seawater. For example, 2556Seawater 75% Freshw~er little change can be noted for the ZnCOz and ZnHCO; species during the mixing process This type of chemical behavior may affect the manner in which Cd and Zn are concentrated in sediments. Copper. Copper has a d9 outer orbital structure and its complexes are characterized by Jahn-Teller 50% Seawater distortion. Copper’s crystal field stabilization energy 50% Freshwater
ltXJ.0096 Freshwater 10.0
0.1 Fig. 2a. Cadmium chemical speciation in seawater-fresh-
1.0 )3
water mixtures. XT
a. 1
75%Seawater
50%Seawater
25% Freshwater
50% Freshwater
100% Seawater
0.1 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.511.5 PH Fig.2b. Cadmium speciationin seawater-freshwater
tures, continued.
Seawater
PH
mix-
Fig. 3. Copper chemical speciation in seawater-freshwater mixtures.
D. T.
I188
LONG
and E. E. ANGINO
and chemical speciation would be expected to differ compared to zinc and cadmium. Figure 3 shows that the crossover field embraces a wider pH range at all salinities than was found for either Zn or Cd. This reflects the increased importance of the carbonato and hydroxo ligands and the decreased importance of the chloride ligand relative to Cd and Zn. The major species for all mixtures are free Cu’+, the bicarbonate compIex and the second order hydroxo complex. There is only a minor shift of the crossover field toward higher pHs with increased amounts of seawater. The relative importance of the copper complexes present in a near shore environment will evidently be little affected by additions of seawater and be largely a function of changes in pH. Lead. Figure 4 shows that the complexation of lead is much more complicated than for the previous elements discussed. Clearly, the speciation of lead is distributed among many more ligands over the pH range studied. Association with the carbonato ligand is dominant in all systems studied. Increasing the amount of seawater in the mixtures shifts the large crossover field only slightly toward higher pH values (Fig. 4). Complexing with Cl- rapidly increases with minor additions of seawater. No reversal occurs, but at seawater concentrations the chloride complexes compete with the carbonato complexes in a pH range that might be expected in an estuarine environment. In the near-shore systems, Cl- and CO:- are by far the most important complexes of lead. The percentage of CO:- complexing does not change appreciably with changes in the other ligand concentrations. The major speciation changes upon the addition of seawater to freshwater are the disappearance
of free Pb” + and the appearance of the chloride complexes. Freshwutrr-brine
mixirtg
system
The next series of species distribution diagrams (Figs. 558) summarize the results obtained for the changes in complexation resulting from the mixtures of freshwater and Ca, Na-Cl brine. These figures are arranged similarly to those previously shown except that only the mixtures 257: brine and 75% freshwater, 507; brine and 50% freshwater and 100% brine are shown. Zinc. The brine soeciation of zinc is characteristically simple, the result of the dominance of the chloro ligand. A rather large shift of the crossover field to higher pHs is apparent. Zinc behaves similarly to cadmium in yielding higher order chloro complexes. At Xl”/, freshwater and 50?/;, brine in the Cd system a chloride concentration is reached in which the percentage concentrations of the chloro associations start to approach one another. Further increasing the chloro concentration, a reversal occurs in which the higher order ligands become more important. Following the reversal. continued increases in the ligand concentration cause the higher order complexes to become dominant. This type of trend has been found to occur for zinc, cadmium and lead. The relative order is Cl1 > Cl, > Cl > Cl, for pure brine for Zn and Pb. One should also note that the difference in magnitude among the orders of the chloro complexes has increased. It is also apparent that the speciation in the mixture of the end members is more complicated that it is for the pure end members. Cadmium. The species distribution of cadmium (Fig. 6) shows the affinity of the cation for the chloride
Freshwater 75%
Freshwater
50%Freshwater
Brine
1.0 0.1 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5 PH Fig. 4. Lead chemical speciation in seawater-freshwater mrxtures.
PH Fig.
5.
Zinc chemical speciation in brine-freshwater mixtures.
Chemical speciation of Cd, Cu, Pb, and Zn Brine Freshwater
50% Brine 50% Freshwater
'l:lr-&
1189
75%Freshwater
50% Freshwater
100%Brine
Fig. 6. Cadmium chemical speciation in brinefreshwater mixtures. ligand. Speciation in the brine is again simple. As the concentration of a ligand is increased in solution (in this case, Cl-) higher order complexes occur. The difference between the absolute magnitudes of the percentage concentration of the chloride complexes decreases with an increase in the chloro ligand concentration. The second order chloro complex remains relatively dominant, however, even in the 100% brine solution (ionic strength equals two). Copper. Figure 7 shows that the distribution of the copper species remains fairly complicated even at brine concentrations. Even with the total alkalinity in the brine system at less than SOppm, copper still complexes the carbonato ligand. It is clear that for
3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.511.5 PH
Fig. 8. Lead chemical speciation in brine-freshwater mixtures. a reversal to occur with chloride and the chloro ligand to dominate the speciation, Cl- would have to be present in much larger concentrations. The percentage of free copper is still high even at brine concentrations. Competition from other complexing components like organic material or clays, would have much free copper available. Lead. Figure 8 shows the speciation for lead. It is immediately apparent that lead, like Cd and Zn, has a simple chemical species distribution in the brine solution. Lead also complexes with CO:- in a manner similar to copper and shows a reversal in the relative importance of the chloride complexes.
APPLICATION OF BRINE-FRESH WATER MODEL TO ORES One concern in studies of the genesis of sedimentary ore deposits is the common occurrence of some base metals to the exclusion of others. For example, 3 present observations show that copper is not commonly associated with lead and zinc strata-bound ore 50% Brine deposits. Similarly, many copper deposits are found 50% Freshwater that are low in zinc and lead concentrations. Why this selective occurrence of trace elements occurs is open to speculation. One must, of course, consider source as a factor. In the segregation, however, if the various trace elements were transported as complexes to the site of deposition in a low-tem100%Brine perature (< 200°C) brine and deposited as sulfides, two other factors deserve consideration. First is the chemical behavior of the trace elements with the various complexing ligands and the other is the behavior of the trace elements with S- or S2-. 3.5 4.5 5.5 6.5 7.5 8.5 9.510.511.5 RENFRO(1974) demonstrated how the behavior of PH Fig. 7. Copper chemical speciation in brine-freshwater trace elements with sulfide in a sabkha could account for the formation of some strata-bound ore deposits. mixtures. 25% Brine 75% Freshwater
1190
D. T.
LONG
and E. E. ANGINO
Results of this theoretical study demonstrate that the behavior of the base metals with ligands present in the transporting medium must also influence the base metal distribution in sedimentary ore deposits. Compare Figs. 5, 7 and 8 for the 100% brine chemical speciation of zinc, copper and lead respectively. Unfortunately, these diagrams are calculated for 25”C, whereas most strata-bound ores are deposited at higher temperatures. Nevertheless, the figures do show the similarity in the chemical speciation of zinc and lead in the brine solution. Both metals are dominated by the chloro ligand; both show a reversal: and both are dominated (r 9Oy:,) by the fourth-order chloride. Significant contribution (> 1%) of other ligands to the complexing do not occur until the pH is greater than about 9.5 for zinc and about I 1.Ofor lead. It might be speculated that fourth-order chloride complexes dominate the speciation of zinc and lead in ore-forming solutions. Copper speciation in the brine, on the other hand. is more varied (Fig. 7). The speciation is not dominated by one order of ligand as it was for zinc and lead. Free copper is a major species (>30%) up to a pH of about 1.5, and no reversal occurs. Thus, it might be anticipated that in an ore forming solution copper would be distributed among more ligands and its chemical behavior at the depositional site would be expected to be different compared with that of zinc or lead. The difference in chemical behavior among copper, lead and zinc could be studied in terms of (1) the consequences of the availability of free copper at the depositional site versus the non availabiiity of free zinc or lead, and (2) the consequences of fourth order ligand associations of lead and zinc versus the lower order associations of copper and the ability of the metal to form sulfide bonds.
Recently REARWN and LANGMUIR(1976) demonstrated that the activity coefficients of MgCOz and CaSOi determined using ~t~tiornet~~ data did not equal computed values based on the Setchenow expression. At 25°C they found that the activity co&cients (y) for these ion pairs fit the equation, logy = -BI, where B equals 0.63 and 0.45 for MgCO$ and CaSOz, respectively; and I equals the ionic strength. Therefore, assigning values to neutral ion pairs based on the Setchenow expression for the y of H,COz may be invalid. In order to examine the effect this new value would have in the computation of species ~str~bu~ons, the species distributions of the elements Cd, Cu, Pb, and using the equation Zn were recalculated log y = -0.5 I, (LANGMUIR,1976, personal communication) for the 7 of neutral species. The result in all cases was to increase the dominance of the neutral species at the expense of the charged species. For example, the relative order of chloride importance in the brine solution changed for both lead and zinc
from Cl, > CIL)> Cl > Clz to Cl, > Cl, > Cl, > CI using the new constant. The speciation of copper still differed from both Pb and Zn and speculations on the behavior of lead and zinc versus copper in sedimentary ore deposits based on chemical speciation differences would be the same as discussed earlier. Use of the expression log ;’ = -0.5 I for neutral species clearly changes the species distribution. This suggests that past dete~inations and interpretations of the theoretical distributjon of chemical species in aqueous solutions may need to be reinvestigated. CONCLUSIONS (1) The distribution of chemical species of Cd, Cu, Zn, and Pb was determined for freshwater-seawater mixtures. As one might have predicted from their outer orbital structures, cadmium and zinc show similar chemical species distributions. In an estuarine system the dist~bution of the chemical species of both Cd and Zn would be constantly readjusting to new Cl- concentrations as the freshwater and seawater mixed. The distribution of the complexing species of copper, on the other hand, was found to change only slightly in the mixtures from freshwater to seawater. (2) Lead was highly complexed to the carbonato ligand in the freshwater-seawater mixtures. This means that lead is likely to be highly available for sorption onto suspended materials as suggested by LELAND and SHIMP (1973) and as demonstrated by ANGINCIet al. (1972) for freshwater. (3) In the freshwater-s~water mixtures the amount follows the order of free ions in solution Cu > Zn >>Pb > Cd. (4) The major controls on the amount and type of inorganic complexing that takes place in various natural water solutions are the absolute and relative concentrations of the competing inorganic ligands. GARDNER(1974) theoretically demonstrated that free amino acids and hydrocarboxylic acids probably are not effective complexers of trace metals. Therefore conclusion number 4 would not change by incorporating organic ligands into the model. The effect of polymerized organic matter. humic and fulvic acids, on the model is unknown. (5) The distribution of the chemical species of Cd, Cu, Zn, and Pb was also determined for freshwater-brine mixtures. The brine was a Ca, Na-Cl solution with an ionic strength of two. The chemical speciation of Zn. Cd and Pb was very similar. All were highly complexed by chloride. The chemical speciation of copper remained fairly complex (no ligand was dominant except for hydroxide at high pHs) and a significant amount of copper remained as the free ion even in the pure brine. Theoretical investigations such as this have limitations. Two of these are the methods used in the calculation of the activities of the ions and the choice of the stability constants. We chose to use individual ionic activity coefficients and feel that our interpre-
Chemical speciation of Cd, Cu, Pb, and Zn
1191
GARRELSR. M. and THOMPSONM. E. (1962) A chemical model for seawater at 25°C and 1 atmosphere total pressure. Aht. J. Sci. 260. 5766. HELGE~N H. C. (1969) Thermodynamics of hydrothermal systems at elevated temperatures and pressures. Am. J. Sci. 266, 729-804. LIVINGSTONE D. A. (1963) Chemical composition of rivers and lakes. U. S. Geof. Surv. Prd Paner 440-G. MESMERR. E. and BAES C. F. JR. (1675) The hydrolysis of cations. CJRNLNSF-EATC-3, Part III, Oak Ridge National Labs. MORELF. and MORGANJ. J. (1972) A numerical method for computing equilibria in aqueous chemical systems. Etlvirorl. Sci. Techol. 6. 58-67. REARDONE. J. and LANGMUIRD. (1974) Thermodynamic properties of the ion pairs MgCOz and CaCOz from 10 to 50°C. Am j. Sci. 274, 599-612. REARDONE. J. and LANGMUIRD. (1976) Activity coeffiAcknowledgements-This study was partially funded by cients of M&O! and CaSOi ion pairs as a function grants from the University of Kansas General Research of ionic strength. Geoclzinzico Cosmochim. Acta 40. Fund and the Mobil Foundation. Computer time was 549-554. made available by the Computation Center of the Univer- RENFROA. R. (1974) Genesis of evaporite-associate stratisity of Kansas. We wish to thank Drs. D. LANGMUIRand form me~iliferous deposits-a sabkha process. Econ. G. HEL~Z for their constructive comments and criticism. Geof. 69. 3345. ROBINSON R. A. and STOKESR. H. (1959) Electrolyte So/utions, 2nd ed., Butterworths. REFERENCES SILLENL. C. and MARTELLA. E. (1964) Stability constants of metal-ion complexes I: Inorganic ligands. Chem. Sot. ANGINOE. E., MAGNIJ~ON L. M. and STEWART6. F. (1972) London Spec. Publ., V. 17. Effects of urbanization on storm water runoff quality: TRUESDELLA. H. and JONESB. F. (1969) Ion association a limited experiment, Naismith Ditch, Lawrence, in natural brines. Chem. Geol. 4, 51-62. Kansas. Water &sour. Res. 8, 135-140. WALKERA. C., BRAYV. B. and JOHNZQNJ. (1927) EquilibBERNERR. A. (1971)Principles of Chemicaf Se~ime~to~~y, McGraw-Hill. rium in solutions of alkali carbonates. Am. Chem. Sot. d. 49, 1235-1256. CARPEN~R A. B., TROUTM. L. and P~CKETTE. E. (1974) WHITEW. B., JOHNXINJ. M. and DANTIZIGG. B. (1958) Preliminary report on the origin and chemical evolution Chemical equilibrium in complex mixtures. J. Chem. of lead- and zinc-rich oil field brines in central Mississippi. Econ. Geol. 69. 1191-1206. Phys. 28, 751-755. WHI~ELD M. (1975) The extension of chemical models ELDERJ. F. (1975) Complexation side reactions involving trace metals in natural water systems. ~~~~~. Oceanugr. for seawater to include trace components at 25°C and 20.96102.’ 1 ATM pressure. Geochim. Cosmochim. Acta 39, ERNESTR., ALLENH. E. and MANCY K. H. (1975) Charac1545-l 557. terization of trace metal species and measurement of Wool J. R. (1975) Thermodynamics of brine-salt equilitrace metal stability constants by electrochemical techbria-1. The systems NaCl-KCl-MgC12-HZ0 and NaCl-MgS04H,0 at 25°C. Geochinz. Cosmochim. Acta niques. Water Res. 9. 969-979. GARDNERL. R. (1974) Organic vs inorganic trace metal 39, 1147-l 163. complexes in sulfidic waters-some speculative calculaZIRINOA. and YA~~o~ S. (1972) A pH-de~ndent model from the chemical speciation of copper, zinc, cadmium tions based on available stability constants. Geo~~~~. and lead in seawater. Limnot, Oceanogr. 17, 661-671. Cosmochim. Acta 38, 1297-1302.
tations at the ionic strengths considered in this paper would not be significantly different if another method were used. The stability constants were picked according to their current nature and availability. Except for the bicarbonate complexes, the stability constants used were values determined by experimentation. Values for the bi~rbonato complexes were taken from ZIRINO and YAMAMOTO (1971) and these values probably overestimate the stability of HCO; complexes (LANGMUIR, personal communication, 1975). De-emphasizing the bicarbonate complexes by choosing another stability constant would change the distribution diagrams but would not change the conclusions of this paper.
et Cosmochimica
Acla.1977,
Vol. 41, pp. I183 to 1191.
Pergamon Press. Printed inGreatBritain
Chemical speciation of Cd, Cu, Pb, and Zn in mixed freshwater, seawater, and brine solutions Department
DAVID T. LONG* and ERNESTE. ANGINO of Geology, University of Kansas, Lawrence, KS 66045, U.S.A.
(Received 13 January 1916; accepted in revised form 18 April 1977)
Abstract-A theoretical model was developed to study the chemical speciation of the trace elements Zn, Cd, Cu and Pb aqueous solutions and their responses to variations in ionic strength and complexation. Two mixing solutions were investigated, a freshwater-seawater system and a freshwater-brine system. The brine was a calcium, sodium-chloride solution with a molal ionic strength of two. Trace element associations with the ligands OH-, Cl-, CO:-, SO:-, and HCO; were considered at pHs from 3.5 to 11.0 at 25°C. In general, the relative importance of the various ligand-trace element complexes can be predicted from a comparison of their stability constants. However, the effect of pH on the importance of a given complex is not readily apparent from the stability constants. Freshwaterseawater mixtures, as might be found in a totally mixed estuary, show that seawater composition is the dominant control on chemical complexing. Chloride complexing is similar for lead and zinc in the freshwater-brine mixtures. This similarity may account in part for the association of lead and zinc in strata-bound ore deposits.
INTRODUCTION VARIOUStheoretical
approaches have been proposed for determining the speciation of specific trace elements in different natural water systems. For details, one can refer to the works of WHITE et al. (1958), GARRELSand THOMPSON(1962), MOREL and MORGAN (1972), ELDER (1975), WHITFIELD (1975) and WOOD (1975). Several proposed models show the changes in chemical speciation that accompany changes in ionic
strength or in the concentrations of the individual ions. Evaluation of these models usually involves simultaneous solution of several mass balance and complexation equilibria using stability constant data. Resulting diagrams usually depict the species distribution in a particular aqueous solution. It is definitely of interest to know how the chemical speciation of elements, particularly trace elements, responds to chemical changes caused by the mixing of fresh and saline waters, such as river water and seawater., The mixtures can be studied by comparing a sequence of diagrams of species distributions resulting from proportional chemical changes between two end member solutions. In these mixing systems, changes in the trace element-ligand concentrations during the mixing process affect the trace element speciation. Since sorption reactions also compete for the trace elements, changes in speciation also affect the amount of trace elements that participate in sorption. The aim of this study was to compute changes in the distribution of aqueous species of Cu, Cd, Zn, and Pb resulting from the mixing of different aqueous solutions. Two mixing systems were studied, freshwater-seawater and freshwater-brine (brine is defined * Present address: Department of Geosciences, North Carolina State University, Raleigh, NC 27607, U.S.A.
here as a solution with greater than 100,000 ppm total dissolved solids). For this study a calcium, sodiumchloride brine was used. The freshwater-seawater system was chosen to provide insight into the chemical changes that might be expected in a totally mixed estuarine environment. Reactions in the freshwater-brine system, on the other hand, may show similarities in the chemical behaviour of certain trace elements during the formation of strata-bound ore deposits. Any similarities found among trace elements may help in the interpretation of the origin of strata-bound ore deposits. MIXING SYSTEMS Mixing systems were set up using the end member concentrations for freshwater, seawater and brine shown in Table 1. Various definitions exist for the term freshwater; consequently, its exact meaning is ambiguous. Usually the term freshwater implies an aqueous solution dilute with respect to the concentrations of the various ions (e.g. river water, ground water, rain water). In this study the term freshwater was defined as one in which the total dissolved solids content was less than 1000 mg/l. Such solutions would correspond roughly to those with ionic strength less than 0.5 mol. kg- ‘. Major and trace element speciation was calculated for the five mixtures defined as 100% seawater, 75-25 (percent seawater-percent freshwater), 50-50, 25-75, and 100% freshwater as a function of pH. The pH was varied from 3.5 to 11.0 at 25°C using 0.5 pH increments. Similar mixtures were studied for the freshwater-brine system. The calculations used to determine the chemical speciation of Cd, Cu, Pb, and Zn in the various aqueous solutions are similar to those of ZIRINO and
1183 G.C.A. 4119-A
D. T. LANGand E. E. ANGMV
1184
Table 1. Molal concentrations
of chemicai end members used in the calculations
component
Freshwater (3)
seawater (4)
Brine (5)
CalCiUU
0.000375
0.0104
0.0964
Magnesium
0.000168
0.054
0.0223
Sodium
0.000274
0.4752
1.6845
Potassium
0.000059
0.01
a.0079
Chloride
0.00022
0.554
2.0198
Ta ("
0.000955
0.00234
0.0001
Sulfate
0.000955
0.0284
0.0029
I (2)
0.0021
0.661
2.099
(1) Total alkalinity. (2) Effective molai ionic strength (after computation of species distribution). From LIVINGSTONE (1963). (4) From BERNER(1971). (5) From CARPENTER et al. (1974).
YAMAMOTO (1972) for the same trace elements. Complexes between the divalent metals and the ligands Cl-. SO:-. CO:-. HCO,. and OH- of the type “MeL,” were considered. Me and L represent the metal and the ligand, and n equals 1-4. Generally, third and fourth order associations were limited to the halides and hydroxo complexes. Since the metals were assumed to be in low concentrations ( < 10ppm) polynucfear complexes were not considered (HELGESON, 1969) As in all modeb of this type, thermodynamic equilibrium was assumed to exist. Since (1) many trace elements are normally quite undersaturated in natural aqueous systems with respect to the precipitation of minerals and (2) precipitated metal would be unavailable for any sorption reactions, the trace element speciation is computed on the basis of the percentages of the metal concentration that would be complexed and uncomplexed. In this way the absolute trace element concentrations are not needed and the systems studied would be unaffected by precipitation reactions. Competition for the trace eiements by other ligands such as clays or organic matter is not considered. The models presented and discussed are strictly inorganic systems. Two steps were used to calculate the distribution of the chemical species in each aqueous mixture. Step one involved the computation of the species distribution of the major ions by the ion association model of GARRELS and %OMP~ON (1962) using the stability constants shown in Table 2. In step two the chemical speciation of the trace elements was determined. The mass action equations used in calculating both major cations and trace element distributions are: (a) for total metal concentration, WTOTI\l_= (m) + i Ii=1
i i=*
/j+)“(rn)[L(i)]“.
E&E IlmL(i),
(1)
(3)
(b) for percent uncomplexed elemental ion, y&n)
= 1001 !‘i:
1+
i
i
n=l
i=l
J?(i).[L(i)]“.
?Zi) m
(2) 1.
(c) for percent complex r&(i), :
II I 1 i -~ 1
;‘mL(i),l
in moles; j = number of ligand associations; !I = order of association; b(i) = stability constant of ith associations nth order; [L(i)]” = concentration of ith ligand in moles for nth order association; yrn= activity coefficient of cation; yL(i)= activity coefficient of ith ligand; and ymL(i),= activity coefficient of complex. Similar equations were used to calculate the distribution of species of the major anions. Individual activity coefficients were used to calculate the activities of the chemical species and were determined in accordance with the ionic strength of the mixture. Activity coefficients of ions in concentrated solutions are difficult to determine theoretically. No method is generally accepted (TRUESDALEand JONES, 1969). Four approximations can be considered (TRUESDALEand JONES, 1969): (1) the extended Debye-Hiickel equation. where (VI) = concentration
of free cations
where: A = constant relating to solvent; B = constant relating to solvent; 6 = distance of closest approach of ith ion; Zi = charge of the ith ion; and I = true ionic strength; (2) the mean salt method (GARRELS and CHRIST, 1965). This is based on the MacInnes assumption that ?K ’ = yC1F = yKC1; (3) DAVIES’ (1962) modification
of the
Debye-Hiickel
1185
Chemical speciation of Cd, Cu, Pb, and Zn Table 2. Stability constants for trace element species used in the calculations LIGANDS
Pb: log3L1
1.6 (2)
6.2 (6)
lo*
1.78(Z)
10.9 (6)
logs 13
1.68(2)
13.9 (66)
1°gS
1.38(Z)
16.3 (6)
12
14
al: lo& l1
0.0212)
6.37(6)
logs 12
-0.71(2)
14.3 (6)
log5 13
-2.3 (2)
15.0 (6)
log@ 14
-4.6 (6)
16.0 (6)
cd: log!3l1
2.0 (6)
5.0 (6)
loge l2
2.7 (6)
10.6 (61
2.1 (6)
10.0 (6)
l@
13
lor36
14
2.9 (I.)
2.7 (1)
6.3 (3)
2.7 (6)
10.64(3)
3.47(61
5.97(3>
2.25(63
9.83(4)
2.1 (1)
4.1 (5)
2.3 (6)
2.1 (1)
5.3 (1)
2.3 (6)
10.0 (6)
zn: log5 11
0.43(2)
4.4 (6)
log6 12
0.61(2)
12.89(6)
log5 13
0.53(2)
14.0 (6)
log6 14
0 2 (2)
15.0 (6)
Values are the logarithms of the association constants. Numbers in parentheses refer to references for the constants and are listed below: (1) ZIRINOand YARDS (1972). (2) HELGE~~N (1969).(3) ERNESTet al. (1975).(4) ME~~WER and BAES(1975). (5) GARDNER (1974).(6) SILLENand MARTEL(1964).
equation. This equation is (5)
-log~i=O.5Z~(~-O.31),
where the equation parameters are the same the extended Deby~Huckel ~uation [equation (4) HELGES~N’S(1969) trace activity coefficient tion (y*) modified from the Stokes-Robinson tion This equation for 25°C is
as in (4)]. equaequa-
A."2.r112
_log$=
n.__-
1 + 6BI
+ B’I,
(6)
where the equation parameters are the same as in the Debye-Hiickel equation plus B’ which is the deviation function. A plot of activity coefficients as determined by the four methods shows that, at ionic strengths of less than about 0.1, the methods predict similar values. The Davies equation is not considered to be valid when the ionic strength is greater than 0.1, since it does not have an adjustable parameter to describe the size of the ion (TRUESDALE and JONES,1969). The mean salt method was used in this study to calculate the activities of the major ions. The method used by HELGWN (1969) was used for the trace elements. The mean salt activity coefficients for HCO; and CO:-
were taken from the data of WALKER, BRAY and J~HNKIN (1927). Recent experimental work on the activity coefficient of SO:- has shown a close correlation of values between those computed theoretically using the and those extended Deby~Huckel equation measured using lead ion electrodes over the range of ionic strengths studied (~2) (HATHAWAY,1975, personal communication). Therefore, the DebyeHuckel equation was also used for the activity coefficients of SO:-. For the individual activity coefficients of the complexes, the Debye-Huckel equation was used for ionic strengths below 0.1 when the adjustable parameters in the equation were defined for the complex. Above 0.1 or when the adjustable parameters were not defined the values were assigned based on the charge of the complex. Complexes with charges of + 1 or +2 were assigned the activity coefficients determined for (HC03)- and (CO,)‘-, respectively. Neutrally charged complexes were assigned the value of the activity coefficient (gamma) of H,CO~ as calculated from the Setchenow expression (REARDONand LANGMUIR,1974): gamma H2CO(: = antilog 0.06 (I), where I = ionic strength. Recently REARWN and LANGMUW (1976) have suggested using a different value for neu-
1186
D. T. LONGand E. E. ANGINO
tral complexes. The effect of using this new value on the theoretical distribution of chemical species will be discussed later. The data for the activity coefficients of the salts used in the mean salt method were taken from ROBINSONand STOKES(1959) for CaCl*, MgC12, NaCl and KCl. In preliminary runs activities of the ions were not adjusted for the effect of the various species on one another. Species diagrams calculated with activities corrected for salt interactions using Harned’s Rule (BERNER,1971) were compared to the same diagrams calculated using uncorrected salt activities. The results showed that for ionic strengths below 2.0 there was no significant difference between diagrams. Thus, in subsequent calculations, Hamed’s Rule was not used to correct salt activities. Since the four trace elements considered in this study have different outer electron orbital structures from the divalent ions, different chemical speciation would be expected between the elements. Zinc (2+) and cadmium (2f) have compIete 3d and 4d outer orbital structures, respectively. Copper (2 + ) has a 3d” structure, and lead (2+) a complete 6s structure with inner f orbitals. The stability constants for the chemical species can give an idea of the relative importance of a particular complex but it is not readily apparent from these constants at what concentrations or at what pHs a given complex is important.
Figures I and 4 show the results for the mixing of seawater and freshwater in different proportions. The figures are arranged with lOOY$seawater at the bottom and 100yOfreshwater at the top; pH increases to the right on the abcissa and the percentage aqueous species or association is plotted on a log scale on the ordinate. Associations are shown for percentages between 0.1 and 100’~. In order to conserve space, the results from the five mixing ratios are only shown for Zn and Cd. For the other metals only the results of the mixtures of iOOO/,freshwater, 50% freshwater-500/, seawater, and lt’@!/, seawater are shown. Zinc. Zinc speciation (Fig. la, b) in aqueous solution can be characterized by three fields-one at pHs up to about 6.5 (designated the precrossover field) a second covering the pH range 6.5-9.5 (the crossover field): and a third at pH values greater than about 9.5 (the postcrossover field). Similar fields can be recognized on the speciation diagrams of the other trace elements. The precrossover field of zinc is characterized in all the mixtures by the don~~ance of free zinc (Zn”+) and by the invariance of the associations with changes in pH. Chloride becomes an important ligand soon after the addition of small amounts of sea water to the fresh water solution. As would be expected, with greater amounts of chloride in solution, higher order associations become important. Also as the chloride concentration is increased, the difference in magnitude of the various orders of the chloride association
becomes less. The relative importance of the chloride ligands in all the zinc associations is Cl > Cl, > Cl1 > Cl,. The crossover field is characterized by an increase in the complexes of the hydroxo ligands and an increase followed by a decrease in the importance of the carbonato ligand. The carbonato ligand is less important in seawater mixtures than in freshwater. This is undoubtedly due to competition from the chloro complexes. Increasing the amount of seawater in solution moves the crossover field to higher pHs. The postcrossover field is characterized by the dominance of the second order hydroxo ligand at the expense of the other ligands. These results are similar to those of ELDER(1975) who demonstrated the importance of the hydroxo ligand in forming complexes at high pHs. Figure la shows that the speciation of zinc is affected by as little as 25% seawater mixed with freshwater. 100.0
Z”
10046Freshwater
.g loo.0 3 2 3 4” E z k
75%Freshwater 10.0 1.0 0.1 Z”
--,f-
(OH)2
50%Seawater 50%Freshwater
1.0 0.1
Fig. la. Zinc chemical speciation in seawater-fr~hwater mixtures.
25%Freshwater
3.5 4.5 5.5 6.5 7.5 8.5 9.510.511.5 PH
Fig. lb. Zinc chemical speciation in seawater-freshwater mixtures, continued,
1187
Chemical speciation of Cd, Cu. Pb, and Zn Chemicaf speciation of zinc and the other trace elements in mixtures containing smaller amounts of seawater was also studied. Species distribution was affected by additions of as little as loo/, seawater. This behavior would be expected from the mixing of high and low ionic strength solutions. The speciation is controlled by both the absolute and relative concentrations of ligands in the solution. In a near shore environm~t such as an estuary or delta front the pH values expected might be in the range 7.5-8.5. The speciation of zinc to be expected in these environments is demonstrated by the crossover field. The dominant species would be free zinc ion, second-order hydroxo complexes, and first-order chloride complexes. Figures la, b show that slight changes in the pH will affect the species distribution to a significant degree and therefore might affect the ability of the sediments to concentrate zinc. ~~drniurn, Cadmium (2+ ) has au outer orbital structure similar to that of zinc. Figures 2a, b show
lN.OW 2 & $ a E 8 a$
10.0 1.0 0.1
the speciation of cadmium is also similar to zinc, as might be expected. Cadmium, however, is more highly complexed to Cl-. Because of the chloride associations, the effect of adding seawater in small amounts is readily apparent by the rapid increased in Cl complexes and the decrease in all other species. The crossover field narrows with increased additions of seawater. For Cd, the postcrossover field is minor. The complexation of cad~um with chloride further demonstrates the trends noted for zinc; i.e. the increase in complexing order and the decrease in the difference of magnitudes between the orders, with increased chloride concentration. For cadmium, this trend is continued to the point where a reversal in the relative importance of the complexes occurs. This reversal takes place at a mixture of 25% freshwhich seawater in water and 75% CdCl(: > CdCl; > CdCl;. Zn a near shore environm~t, the speciation of cadmium should correspond to that showu in the precrossover and crossover fields. Cadmium’s species distribution would be much simpler than that of zinc. 100% Freshwater One would expect it to be dominated by chloride and relatively unaffected by small changes in pH. The species change, however, during the change from a freshwater to seawater, such as a river entering an ocean, would affect Cd more than Zn. Many of the freshwater zinc species including Zn2+ retain a similar distribution in 100% seawater. For example, 2556Seawater 75% Freshw~er little change can be noted for the ZnCOz and ZnHCO; species during the mixing process This type of chemical behavior may affect the manner in which Cd and Zn are concentrated in sediments. Copper. Copper has a d9 outer orbital structure and its complexes are characterized by Jahn-Teller 50% Seawater distortion. Copper’s crystal field stabilization energy 50% Freshwater
ltXJ.0096 Freshwater 10.0
0.1 Fig. 2a. Cadmium chemical speciation in seawater-fresh-
1.0 )3
water mixtures. XT
a. 1
75%Seawater
50%Seawater
25% Freshwater
50% Freshwater
100% Seawater
0.1 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.511.5 PH Fig.2b. Cadmium speciationin seawater-freshwater
tures, continued.
Seawater
PH
mix-
Fig. 3. Copper chemical speciation in seawater-freshwater mixtures.
D. T.
I188
LONG
and E. E. ANGINO
and chemical speciation would be expected to differ compared to zinc and cadmium. Figure 3 shows that the crossover field embraces a wider pH range at all salinities than was found for either Zn or Cd. This reflects the increased importance of the carbonato and hydroxo ligands and the decreased importance of the chloride ligand relative to Cd and Zn. The major species for all mixtures are free Cu’+, the bicarbonate compIex and the second order hydroxo complex. There is only a minor shift of the crossover field toward higher pHs with increased amounts of seawater. The relative importance of the copper complexes present in a near shore environment will evidently be little affected by additions of seawater and be largely a function of changes in pH. Lead. Figure 4 shows that the complexation of lead is much more complicated than for the previous elements discussed. Clearly, the speciation of lead is distributed among many more ligands over the pH range studied. Association with the carbonato ligand is dominant in all systems studied. Increasing the amount of seawater in the mixtures shifts the large crossover field only slightly toward higher pH values (Fig. 4). Complexing with Cl- rapidly increases with minor additions of seawater. No reversal occurs, but at seawater concentrations the chloride complexes compete with the carbonato complexes in a pH range that might be expected in an estuarine environment. In the near-shore systems, Cl- and CO:- are by far the most important complexes of lead. The percentage of CO:- complexing does not change appreciably with changes in the other ligand concentrations. The major speciation changes upon the addition of seawater to freshwater are the disappearance
of free Pb” + and the appearance of the chloride complexes. Freshwutrr-brine
mixirtg
system
The next series of species distribution diagrams (Figs. 558) summarize the results obtained for the changes in complexation resulting from the mixtures of freshwater and Ca, Na-Cl brine. These figures are arranged similarly to those previously shown except that only the mixtures 257: brine and 75% freshwater, 507; brine and 50% freshwater and 100% brine are shown. Zinc. The brine soeciation of zinc is characteristically simple, the result of the dominance of the chloro ligand. A rather large shift of the crossover field to higher pHs is apparent. Zinc behaves similarly to cadmium in yielding higher order chloro complexes. At Xl”/, freshwater and 50?/;, brine in the Cd system a chloride concentration is reached in which the percentage concentrations of the chloro associations start to approach one another. Further increasing the chloro concentration, a reversal occurs in which the higher order ligands become more important. Following the reversal. continued increases in the ligand concentration cause the higher order complexes to become dominant. This type of trend has been found to occur for zinc, cadmium and lead. The relative order is Cl1 > Cl, > Cl > Cl, for pure brine for Zn and Pb. One should also note that the difference in magnitude among the orders of the chloro complexes has increased. It is also apparent that the speciation in the mixture of the end members is more complicated that it is for the pure end members. Cadmium. The species distribution of cadmium (Fig. 6) shows the affinity of the cation for the chloride
Freshwater 75%
Freshwater
50%Freshwater
Brine
1.0 0.1 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5 PH Fig. 4. Lead chemical speciation in seawater-freshwater mrxtures.
PH Fig.
5.
Zinc chemical speciation in brine-freshwater mixtures.
Chemical speciation of Cd, Cu, Pb, and Zn Brine Freshwater
50% Brine 50% Freshwater
'l:lr-&
1189
75%Freshwater
50% Freshwater
100%Brine
Fig. 6. Cadmium chemical speciation in brinefreshwater mixtures. ligand. Speciation in the brine is again simple. As the concentration of a ligand is increased in solution (in this case, Cl-) higher order complexes occur. The difference between the absolute magnitudes of the percentage concentration of the chloride complexes decreases with an increase in the chloro ligand concentration. The second order chloro complex remains relatively dominant, however, even in the 100% brine solution (ionic strength equals two). Copper. Figure 7 shows that the distribution of the copper species remains fairly complicated even at brine concentrations. Even with the total alkalinity in the brine system at less than SOppm, copper still complexes the carbonato ligand. It is clear that for
3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.511.5 PH
Fig. 8. Lead chemical speciation in brine-freshwater mixtures. a reversal to occur with chloride and the chloro ligand to dominate the speciation, Cl- would have to be present in much larger concentrations. The percentage of free copper is still high even at brine concentrations. Competition from other complexing components like organic material or clays, would have much free copper available. Lead. Figure 8 shows the speciation for lead. It is immediately apparent that lead, like Cd and Zn, has a simple chemical species distribution in the brine solution. Lead also complexes with CO:- in a manner similar to copper and shows a reversal in the relative importance of the chloride complexes.
APPLICATION OF BRINE-FRESH WATER MODEL TO ORES One concern in studies of the genesis of sedimentary ore deposits is the common occurrence of some base metals to the exclusion of others. For example, 3 present observations show that copper is not commonly associated with lead and zinc strata-bound ore 50% Brine deposits. Similarly, many copper deposits are found 50% Freshwater that are low in zinc and lead concentrations. Why this selective occurrence of trace elements occurs is open to speculation. One must, of course, consider source as a factor. In the segregation, however, if the various trace elements were transported as complexes to the site of deposition in a low-tem100%Brine perature (< 200°C) brine and deposited as sulfides, two other factors deserve consideration. First is the chemical behavior of the trace elements with the various complexing ligands and the other is the behavior of the trace elements with S- or S2-. 3.5 4.5 5.5 6.5 7.5 8.5 9.510.511.5 RENFRO(1974) demonstrated how the behavior of PH Fig. 7. Copper chemical speciation in brine-freshwater trace elements with sulfide in a sabkha could account for the formation of some strata-bound ore deposits. mixtures. 25% Brine 75% Freshwater
1190
D. T.
LONG
and E. E. ANGINO
Results of this theoretical study demonstrate that the behavior of the base metals with ligands present in the transporting medium must also influence the base metal distribution in sedimentary ore deposits. Compare Figs. 5, 7 and 8 for the 100% brine chemical speciation of zinc, copper and lead respectively. Unfortunately, these diagrams are calculated for 25”C, whereas most strata-bound ores are deposited at higher temperatures. Nevertheless, the figures do show the similarity in the chemical speciation of zinc and lead in the brine solution. Both metals are dominated by the chloro ligand; both show a reversal: and both are dominated (r 9Oy:,) by the fourth-order chloride. Significant contribution (> 1%) of other ligands to the complexing do not occur until the pH is greater than about 9.5 for zinc and about I 1.Ofor lead. It might be speculated that fourth-order chloride complexes dominate the speciation of zinc and lead in ore-forming solutions. Copper speciation in the brine, on the other hand. is more varied (Fig. 7). The speciation is not dominated by one order of ligand as it was for zinc and lead. Free copper is a major species (>30%) up to a pH of about 1.5, and no reversal occurs. Thus, it might be anticipated that in an ore forming solution copper would be distributed among more ligands and its chemical behavior at the depositional site would be expected to be different compared with that of zinc or lead. The difference in chemical behavior among copper, lead and zinc could be studied in terms of (1) the consequences of the availability of free copper at the depositional site versus the non availabiiity of free zinc or lead, and (2) the consequences of fourth order ligand associations of lead and zinc versus the lower order associations of copper and the ability of the metal to form sulfide bonds.
Recently REARWN and LANGMUIR(1976) demonstrated that the activity coefficients of MgCOz and CaSOi determined using ~t~tiornet~~ data did not equal computed values based on the Setchenow expression. At 25°C they found that the activity co&cients (y) for these ion pairs fit the equation, logy = -BI, where B equals 0.63 and 0.45 for MgCO$ and CaSOz, respectively; and I equals the ionic strength. Therefore, assigning values to neutral ion pairs based on the Setchenow expression for the y of H,COz may be invalid. In order to examine the effect this new value would have in the computation of species ~str~bu~ons, the species distributions of the elements Cd, Cu, Pb, and using the equation Zn were recalculated log y = -0.5 I, (LANGMUIR,1976, personal communication) for the 7 of neutral species. The result in all cases was to increase the dominance of the neutral species at the expense of the charged species. For example, the relative order of chloride importance in the brine solution changed for both lead and zinc
from Cl, > CIL)> Cl > Clz to Cl, > Cl, > Cl, > CI using the new constant. The speciation of copper still differed from both Pb and Zn and speculations on the behavior of lead and zinc versus copper in sedimentary ore deposits based on chemical speciation differences would be the same as discussed earlier. Use of the expression log ;’ = -0.5 I for neutral species clearly changes the species distribution. This suggests that past dete~inations and interpretations of the theoretical distributjon of chemical species in aqueous solutions may need to be reinvestigated. CONCLUSIONS (1) The distribution of chemical species of Cd, Cu, Zn, and Pb was determined for freshwater-seawater mixtures. As one might have predicted from their outer orbital structures, cadmium and zinc show similar chemical species distributions. In an estuarine system the dist~bution of the chemical species of both Cd and Zn would be constantly readjusting to new Cl- concentrations as the freshwater and seawater mixed. The distribution of the complexing species of copper, on the other hand, was found to change only slightly in the mixtures from freshwater to seawater. (2) Lead was highly complexed to the carbonato ligand in the freshwater-seawater mixtures. This means that lead is likely to be highly available for sorption onto suspended materials as suggested by LELAND and SHIMP (1973) and as demonstrated by ANGINCIet al. (1972) for freshwater. (3) In the freshwater-s~water mixtures the amount follows the order of free ions in solution Cu > Zn >>Pb > Cd. (4) The major controls on the amount and type of inorganic complexing that takes place in various natural water solutions are the absolute and relative concentrations of the competing inorganic ligands. GARDNER(1974) theoretically demonstrated that free amino acids and hydrocarboxylic acids probably are not effective complexers of trace metals. Therefore conclusion number 4 would not change by incorporating organic ligands into the model. The effect of polymerized organic matter. humic and fulvic acids, on the model is unknown. (5) The distribution of the chemical species of Cd, Cu, Zn, and Pb was also determined for freshwater-brine mixtures. The brine was a Ca, Na-Cl solution with an ionic strength of two. The chemical speciation of Zn. Cd and Pb was very similar. All were highly complexed by chloride. The chemical speciation of copper remained fairly complex (no ligand was dominant except for hydroxide at high pHs) and a significant amount of copper remained as the free ion even in the pure brine. Theoretical investigations such as this have limitations. Two of these are the methods used in the calculation of the activities of the ions and the choice of the stability constants. We chose to use individual ionic activity coefficients and feel that our interpre-
Chemical speciation of Cd, Cu, Pb, and Zn
1191
GARRELSR. M. and THOMPSONM. E. (1962) A chemical model for seawater at 25°C and 1 atmosphere total pressure. Aht. J. Sci. 260. 5766. HELGE~N H. C. (1969) Thermodynamics of hydrothermal systems at elevated temperatures and pressures. Am. J. Sci. 266, 729-804. LIVINGSTONE D. A. (1963) Chemical composition of rivers and lakes. U. S. Geof. Surv. Prd Paner 440-G. MESMERR. E. and BAES C. F. JR. (1675) The hydrolysis of cations. CJRNLNSF-EATC-3, Part III, Oak Ridge National Labs. MORELF. and MORGANJ. J. (1972) A numerical method for computing equilibria in aqueous chemical systems. Etlvirorl. Sci. Techol. 6. 58-67. REARDONE. J. and LANGMUIRD. (1974) Thermodynamic properties of the ion pairs MgCOz and CaCOz from 10 to 50°C. Am j. Sci. 274, 599-612. REARDONE. J. and LANGMUIRD. (1976) Activity coeffiAcknowledgements-This study was partially funded by cients of M&O! and CaSOi ion pairs as a function grants from the University of Kansas General Research of ionic strength. Geoclzinzico Cosmochim. Acta 40. Fund and the Mobil Foundation. Computer time was 549-554. made available by the Computation Center of the Univer- RENFROA. R. (1974) Genesis of evaporite-associate stratisity of Kansas. We wish to thank Drs. D. LANGMUIRand form me~iliferous deposits-a sabkha process. Econ. G. HEL~Z for their constructive comments and criticism. Geof. 69. 3345. ROBINSON R. A. and STOKESR. H. (1959) Electrolyte So/utions, 2nd ed., Butterworths. REFERENCES SILLENL. C. and MARTELLA. E. (1964) Stability constants of metal-ion complexes I: Inorganic ligands. Chem. Sot. ANGINOE. E., MAGNIJ~ON L. M. and STEWART6. F. (1972) London Spec. Publ., V. 17. Effects of urbanization on storm water runoff quality: TRUESDELLA. H. and JONESB. F. (1969) Ion association a limited experiment, Naismith Ditch, Lawrence, in natural brines. Chem. Geol. 4, 51-62. Kansas. Water &sour. Res. 8, 135-140. WALKERA. C., BRAYV. B. and JOHNZQNJ. (1927) EquilibBERNERR. A. (1971)Principles of Chemicaf Se~ime~to~~y, McGraw-Hill. rium in solutions of alkali carbonates. Am. Chem. Sot. d. 49, 1235-1256. CARPEN~R A. B., TROUTM. L. and P~CKETTE. E. (1974) WHITEW. B., JOHNXINJ. M. and DANTIZIGG. B. (1958) Preliminary report on the origin and chemical evolution Chemical equilibrium in complex mixtures. J. Chem. of lead- and zinc-rich oil field brines in central Mississippi. Econ. Geol. 69. 1191-1206. Phys. 28, 751-755. WHI~ELD M. (1975) The extension of chemical models ELDERJ. F. (1975) Complexation side reactions involving trace metals in natural water systems. ~~~~~. Oceanugr. for seawater to include trace components at 25°C and 20.96102.’ 1 ATM pressure. Geochim. Cosmochim. Acta 39, ERNESTR., ALLENH. E. and MANCY K. H. (1975) Charac1545-l 557. terization of trace metal species and measurement of Wool J. R. (1975) Thermodynamics of brine-salt equilitrace metal stability constants by electrochemical techbria-1. The systems NaCl-KCl-MgC12-HZ0 and NaCl-MgS04H,0 at 25°C. Geochinz. Cosmochim. Acta niques. Water Res. 9. 969-979. GARDNERL. R. (1974) Organic vs inorganic trace metal 39, 1147-l 163. complexes in sulfidic waters-some speculative calculaZIRINOA. and YA~~o~ S. (1972) A pH-de~ndent model from the chemical speciation of copper, zinc, cadmium tions based on available stability constants. Geo~~~~. and lead in seawater. Limnot, Oceanogr. 17, 661-671. Cosmochim. Acta 38, 1297-1302.
tations at the ionic strengths considered in this paper would not be significantly different if another method were used. The stability constants were picked according to their current nature and availability. Except for the bicarbonate complexes, the stability constants used were values determined by experimentation. Values for the bi~rbonato complexes were taken from ZIRINO and YAMAMOTO (1971) and these values probably overestimate the stability of HCO; complexes (LANGMUIR, personal communication, 1975). De-emphasizing the bicarbonate complexes by choosing another stability constant would change the distribution diagrams but would not change the conclusions of this paper.