Modeling on adsorption–desorption of trace metals to suspended particle matter in the Changjiang Estuary

Abstract

The uptake and release of trace metals (Cu, Ni, Zn, Cd, and Co) in estuaries are studied using river and sea end-member waters and suspended particulate matter (SPM) collected from the Changjiang Estuary, China. The kinetics of adsorption and desorption were studied in terms of environmental factors (pH, SPM loading, and salinity) and metal concentrations. The uptake of the metals studied onto SPM occurred mostly within 10 h and reached an asymptotic value within 40 h in the Changjiang Estuary. As low pH river water flows into the high pH seawater and the water become more alkaline as it approaches to the seaside of estuary, metals adsorb more on SPM in higher pH water, thus, particulate phase transport of metal become increasingly important in the seaward side of the estuary. The percentage of adsorption recovery and the distribution coefficients for trace metals remained to be relatively invariable and a significant reduction only occurred in very high concentrations of metals (>0.1 mg L⁻¹). The general effect of salinity on metal behavior was to decrease the degree of adsorption of Cu, Zn, Cd, Co, and Ni onto SPM but to increase their adsorption equilibrium pH. The adsorption–desorption kinetics of trace metals were further investigated using Kurbatov adsorption model. The model appears to be most useful for the metals showing the conservative behavior during mixing of river and seawater in the estuary. Our work demonstrates that dissolved concentration of trace metals in estuary can be modeled based on the metal concentration in SPM, pH and salinity using a Kurbatov adsorption model assuming the natural SPM as a simple surfaced molecule.

Appendix

The data interpretation can be approached using the Kurbatov adsorption model (Kurbatov et al. 1951), which was initially applied to the adsorption of metals by hydrate-oxide, but later used in the adsorption of trace metals to suspended particles from natural waters (Lead et al. 1999; Wilson et al. 2001). Adsorption of metals by suspended particles could be assumed to be an exchange of the active group with metal ions (e.g. Grantza et al. 2003). Suspended particulate materials are heterogenic in composition but there exist some active groups (i.e., –OH, –COOH and –NH₂) on the surface of the mineral and giant organic molecules (humus) in water. These can exchange with metal ions on the surface of SPM (Wilson et al. 2001), that is,

$$ {\equiv} \text{S{\hbox{--}}OH}_{x} + M\overset K \longleftrightarrow {\equiv} \text{S{\hbox{--}}OM} + x \cdot {\text{H}} $$

(14)

where ${\equiv} \text{S{\hbox{--}}OH}$ is the surface active potential of SPM in the Changjiang Estuary, M is the metal added to the simulation system, ${\equiv} \text{S{\hbox{--}}OM}$ is the metal adsorbed by SPM, and H is the hydrogen ion. In Eq. (2), x and K are adsorption parameters with x being the exchange coefficient and K the apparent adsorption equilibrium constant.

In Eq. (14), K could be expressed by

$$ K = \frac{{{\left[ { {\equiv} \text{S{\hbox{--}}OM}} \right]} \cdot {\left[ H \right]}^{x} }} {{{\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]} \cdot {\left[ M \right]}}} $$

(15)

where${\left[ { {\equiv} \text{S{\hbox{--}}OM}} \right]}$ is the adsorbed metal concentration (mg L⁻¹), [H] is the hydrogen ion concentration (mol L⁻¹), [M] is the dissolved metal concentration (mg L⁻¹), i.e., the metal concentration at adsorption equilibrium, and ${\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]}$ is the amount of surface active potential in equilibrium. The Eq. (15) can also be re-arranged as

$$ \log \frac{{[{\text{SOM}}]}} {{[M][{\text{SOH}}_{x} ]}} = x \cdot {\text{pH}} + \log K $$

(16)

When the metal concentration is very low against particle concentration, Eq. (16) can be simplified as

$$ {\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]} \approx {\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]}_{0} = k \cdot {\text{SPM}} $$

(17)

hence, ${\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]}$ can be replaced by ${\left( {k \cdot {\text{SPM}}} \right)},$ and Eq. (16) becomes

$$ \begin{aligned}{} \log \frac{{[ {\equiv} \text{S{\hbox{--}}OM}]}} {{[M]}} & = x \cdot {\text{pH}}_{{{\text{eq}}}} + \log K + \log [ {\equiv} \text{S{\hbox{--}}OH}_{x} ] \\ \; = x \cdot {\text{pH}} + \log K + \log {\left\{ {k \cdot {\text{SPM}}} \right\}} \\ \end{aligned} $$

(18)

The data of the adsorption for the five metals by SPM from the sampling site can be predicted by Eq. (18), and the adsorption parameters (e.g. x and K) can be evaluated by the slope and interception of the lines and list in Table 7.

Table 7 Linear fitting results of the five metals adsorption on to SPM

For a given SPM level, salinity, and metal concentration, Eq. (18) can be re-organized:

$$ \log \frac{{[ {\equiv} \text{S{\hbox{--}}OM}]}} {{[M]}} = x \cdot {\text{pH}} + \log K^{\prime } $$

(19)

The ratio of solid–liquid partitioning of trace metals is defined as

$$ D = \frac{{[ {\equiv} \text{S{\hbox{--}}OM}]}} {{[M]}} $$

(20)

The percentage of adsorption can be also calculated by

$$ E(\% ) = \frac{D} {{1 + D}} \times 100\% $$

(21)

In Eq. (19), ${\text{log}}\frac{{[ {\equiv} \text{S{\hbox{--}}OM}]}} {{[M]}}$ is plotted against the pH at equilibrium in the same SPM, salinity, and metal concentration systems, respectively. The slope (x) and the interception (log K′) of the plots can be obtained and the percentage adsorption E (%) can be plotted against the pH at equilibrium (Fig. 11).

Fig. 11

Relationships between pH1/2 and SPM, salinity, and T _m (metal concentration). a S = 0.0, T _Cu = 1.0 mg L⁻¹, b SPM = 300 mg L⁻¹, T _Cu = 1.5 mg L⁻¹, c S = 0.0, SPM = 448 mg L⁻¹

The Eq. (18) can be reorganized using the ratio of solid–liquid partitioning D.

$$ {\text{pH}}_{{{\text{eq}}}} = - \frac{1} {x} \cdot \log {\left\{ {\frac{{K \cdot {\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]}}} {D}} \right\}} $$

(22)

When the percentage adsorption [E (%)] is equal to 50%, the pH (i.e. pH of 50% uptake) can be calculated by Eq. (22):

$$ {\text{pH}}_{{{\text{1/2}}}} = - \frac{1} {x} \cdot \log {\left\{ {K \cdot {\left[ { {\equiv} \text{S{\hbox{--}}OH}_{x} } \right]}} \right\}} $$

(23)

Namely,

$$ {\text{log}}\,D = x \cdot {\left( {{\text{pH}}_{{{\text{eq}}}} - {\text{pH}}_{{{\text{1/2}}}} } \right)} $$

(24)

$$ E(\% ) = \frac{{10^{{{\left\{ {x \cdot {\left[ {{\text{pH}}_{{{\text{eq}}}} - {\text{pH}}_{{{\text{1/2}}}} } \right]}} \right\}}}} }} {{1 + 10^{{{\left\{ {x \cdot {\left[ {{\text{pH}}_{{{\text{eq}}}} - {\text{pH}}_{{{\text{1/2}}}} } \right]}} \right\}}}} }} \times 100\% $$

(25)

In Fig. 11, pH_1/2 increases with higher levels of SPM and salinity and the availability of dissolved metals.

The linear relationship of x and pH_1/2 against SPM, salinity, and metal concentrations are further examined in Fig. 12. In Fig. 12, the expression of SPM, salinity (S), and metal concentrations (T) for a given system x and pH_1/2 of Cu adsorbed by SPM from Xuliujing can be written as

$$ \begin{aligned}{} x & = X_{{{\text{SPM}}}} \cdot \log ({\text{SPM}}) + X_{S} \cdot \log (S + 1) + X_{{T_{{{\text{Cu}}}} }} \cdot \log (T_{{{\text{Cu}}}} ) + X \\ {\text{pH}}_{{1/2}} & = P_{{{\text{SPM}}}} \cdot \log ({\text{SPM}}) + P_{S} \cdot \log (S + 1) + P_{{T_{{{\text{Cu}}}} }} \cdot \log (T_{{{\text{Cu}}}} ) + P \\ \end{aligned} $$

(26)

where experiential parameters X _SPM, X _S , and $X_{{T_{{{\text{Cu}}}} }}$ are slopes of data x compared against log(SPM), log(S + 1), and log(T _Cu) respectively (Table 8). P _SPM, P _S , and $P_{{T_{{{\text{Cu}}}} }}$ represent slopes of pH_1/2 as affected by log(SPM), log(S + 1), and log(T _Cu).

Fig. 12

Relationships between the fitted slope, pH_1/2, and SPM, salinity, and T _m (metal concentration). a S = 0.0, T _Cu = 1.0 mg L⁻¹, b SPM = 300 mg L⁻¹, T _Cu = 1.5 mg L⁻¹, c S = 0.0, SPM = 448 mg L⁻¹

Table 8 The simulation parameters for five metals adsorption onto suspended sediments in the Changjiang Estuary

Equation (27) is employed to calculate X and P:

$$ \begin{aligned}{} X & = x - {\left[ {X_{{{\text{SPM}}}} \cdot \log ({\text{SPM}}) + X_{S} \cdot \log (S + 1) + X_{{T_{{{\text{Cu}}}} }} \cdot \log (T_{{{\text{Cu}}}} )} \right]} \\ P & = {\text{pH}}_{{1/2}} - {\left[ {P_{{{\text{SPM}}}} \cdot \log ({\text{SPM}}) + P_{S} \cdot \log (S + 1) + P_{{T_{{{\text{Cu}}}} }} \cdot \log (T_{{{\text{Cu}}}} )} \right]} \\ \end{aligned} $$

(27)

All the parameters (i.e. x, pH_1/2, X _SPM, X _S , $X_{{T_{{{\text{Cu}}}} }}$, P _SPM, P _S , and $P_{{T_{{{\text{Cu}}}} }}$) of experiment can be calculated from Figs. 11 and 12. When these experiential parameters are substituted into Eq. (27), X and P can be calculated and listed in Tables 7 and 8. The average of X and P are considered as the parameters of Eq. (27).

The other metals can also be calculated as Cu is, and the experiential parameters of other metals can also be calculated according to the above equations. The values are listed in Table 8. A good fit is achieved when we input the parameters of SPM, Salinity, pH, and concentration of the metals from the batch experiments to the model. The correlation coefficients of the fitted curve γ² are from 0.95 to 0.99.