Poly(l-cysteine) (PLC) was successfully immobilized on controlled-pore glass (CPG) and used in a flow injection system incorporating a microcolumn for separation of Cd from synthetic solutions. On-line breakthrough curves were used to study the effect of pH, concentration, and influent flow rate on Cd adsorption. The Cd breakthrough experiments revealed a significant strong-site capacity as well as weaker binding sites that showed a dependence on influent concentration. In addition, increased capacity was observed at more alkaline pH values. The flow rate studies showed nonequilibrium conditions existed for Cd binding at flow rates of 0.55−3.75 mL/min, with decreased capacity for faster flow rates. Elemental analysis of immobilized PLC for sulfur estimated 6.8 × 10-6 mol of PLC/g of CPG and the calculated mole ratio of Cd to PLC was 6.2:1 at pH 9. Stability constants governing the Cd adsorption by PLC were obtained by a nonlinear least-squares analysis of the Cd binding data and revealed at least four classes of binding sites were present and that the stable Cd complexation observed for the free molecule was retained by immobilized PLC. Stability constants for the most stable sites were estimated using EDTA and en as competing ligands and resulted in two sites:  K 1 = 1 × 1013 and K 2 = 109−1011 with capacities of n 1 = 1 μmol/g and and n 2 = 6 μmol/g, respectively. Two weaker binding sites were also identified and represented by stability constants of K 3 = 1 × 106 and K 5 = 2 × 104 with site capacities of n 4 = 20 μmol/g and n 5 = 10 μmol/g, respectively.

Poly(-cysteine) (PLC ) and 8-hydroxyquinoline (8HQ) were immobilized on controlled-pore glass and used in a flow injection system for the separation of Cd, Pb and Cu from synthetic sea-water, Co and Ni matrices as well as CRM sea-water. Both resins allowed for the quantitative recovery of 50 mg L−1 Cd and Pb in synthetic sea-water. However, low recoveries of 2-4% and 40-50% were observed using 8HQ for the separation of 50 mg L−1 Cd and Pb, respectively, from a 10 000-fold excess of Co and Ni, while PLC maintained quantitative recoveries. Neither 8HQ nor PLC showed reproducible or complete recoveries of Cu2+ from the columns using the typical means for stripping (1 M HNO 3 ). On-line breakthrough experiments showed that 8HQ had a significant strong binding site capacity for Cd, Pb, Cu, Co and Ni. PLC also had strong sites for Cd, Pb and Cu but showed only weak binding of Co and Ni. The selectivity of PLC against these harder acid metals allowed for quantitative recovery of Cd, Pb and Cu in Co and Ni matrices. Extracting low level spikes of Cd and Pb from CRM sea-water (CASS-1 and NASS-2) tested the application to 'real' samples. Recovery efficiencies of Cd were high for both CRM matrices studied. Pb recovery was good for CASS-1; however, recovery from NASS-2 was unexpectedly low. Mass transfer limitations were observed for both resins in the flow system, resulting in apparent decreased capacities at faster flow rates. Stability constants governing Cd adsorption by PLC and 8HQ were obtained by a non-linear least-squares regression analysis of the Cd binding data and revealed that at least four classes of binding site were present on both resins. Stability constants for the most stable sites were estimated using EDTA or ethylenediamine (en) as competing ligands. 8HQ had no sites that were competitive with EDTA, whereas PLC had an EDTA-competitive site with a stability constant of 1×1013 and a capacity of 1 mmol g−1. Both PLC and 8HQ had sites that were stronger than Cd(en) 2 with estimated stability constants ranging from 109 to 1011. Weaker sites on the resins had stability constants that ranged from 104 to 106. Cd was used to demonstrate the viability of this method for stability constant determination as it is well characterized for both 8HQ and PLC.

The on-line reversible oxidation of poly(l-cysteine) immobilized on controlled-pore glass (PLC−CPG) was demonstrated using 1% hydrogen peroxide, aerated buffer solution, and Cu(II) as oxidizing agents and dithiothreitol (DTT) as a reducing agent. Cd breakthrough curves on the oxidized forms of PLC−CPG, used as an ion exchanger, revealed that hydrogen peroxide was an effective oxidizing agent for the sulfhydryl groups and eliminated all of the strong binding sites for Cd, while oxidation by aerated water removed only 33% of the total Cd capacity after an 18-h reaction. A Cu(II) breakthrough curve on a partially reduced form of PLC−CPG showed that strong Cu binding sites were present and that the Cu(II) oxidized PLC−CPG, most likely through the binding of Cu(II) to sulfhydryl binding sites, followed by their oxidation to disulfides with the production and likely binding of Cu(I). At least 10% of the bound Cu had to be removed using KCN, which suggested Cu(I) was strongly complexed. Oxidation of PLC−CPG using Cu(II) eliminated fewer Cd binding sites than hydrogen peroxide, which may, in addition to forming disulfides, oxidize lone sulfhydryl groups to sulfenites or sulfonates. A competitive binding experiment with Cd and Cu(II) suggested Cd was able to compete with Cu(II) for binding sites on PLC−CPG and that the complexed Cd may inhibit Cu(II) oxidation of those sites. PLC−CPG could be fully reduced in 10 min from its peroxide-oxidized form with pH 8.1 DTT, while reduction with pH 4.5 DTT was much slower and allowed for intermediate (partially reduced) oxidation states. Oxidized forms of PLC−CPG were used successfully for on-line Cd, Zn, and Pb chelation and preconcentration at pH 7.0 and influent flow rates of 1.0 mL/min. Oxidized PLC−CPG also retained enough strong binding sites for Cd to permit its quantitative recovery from a 400-fold excess Cu(II) matrix and average recoveries of 93+% of the Cd in alkali and alkaline earth matrixes.