Evaluation of the copper anodic stripping voltammetry complexometric

A Voltammetric Study on Toxic Metals, their Speciation and Interaction with Nutrients and Organic Ligand in a South Pacific Ocean Region. H. Ida de Gr...
0 downloads 0 Views 296KB Size
2116

Anal. Chem. 1982, 54, 2116-2117

Exchange of Comments on Evaluation of the Copper Anodic Stripping Voltammetry Complexometric Titration for Complexing Capacities and Conditional Stability Constants Sir: Tuschall and Brezonik (I) in a recent paper criticize the titrametric anodic stripping voltammetry (ASV) method ( 2 , 3 )for determining total ligand concentration (CL) values and conditional stability constants (@’) for copper(I1) complexes. The authors conclude that the method is unsuitable for titration of natural water samples or natural organic matter. In contrast, our recent paper ( 4 ) concludes that the ASV method is suitable for titration of soil fulvic acid with cu2+. We contend that Tuschall and Brezonik (1)did not properly test the approach because of a poor choice of ligands and wrong application of theory (5). The measured stripping current (is) has one or more of the following components depending on the behavior of the system The current id is due to diffusion, ik is due to dissociation of the complex, and iredis due to reduction of the complex if the reduction potential of the complex is not well separated from that of metal ion present in the complex. y is the proportionality constant between plating and stripping currents. We will now examine how much each of the current components contributes to is for the titration of EDTA by Cu2+. The general reaction of interest for 1:l complexation is

where the primes designate all metal ion and ligand species not in ML. By rearrangement

pM’ = log p’

[L’I + log [MLI

A t pH 7 log p’ is approximately 13 for CU(EDTA)~-.The pM’ values during the titration are as follows: 50% titration, pM’ = 13; 90% titration, pM’ = 12; 99% titration, pM’ = 11; and 99.999% titration, pM’ = 8. Since the detection limit for ASV (5) is about lo3 M (pM’ = 9), is has virtually no id component during the titration. Therefore, is is comprised of ik and/or ired and does not correspond to free [Cu2+]. Since Shuman’s equations are derived for diffusion-controlled processes ( i d is the main component) with some ik contribution, they cannot be applied to the Cu2+-EDTA titration at pH 7 . It is unsurprising that the “measured @’ value” is far below the correct value. Previously Chau et al. (6) demonstrated the difficulty of obtaining p’ for CU(EDTA)~+ by ASV, and Eisenreich et al. (7) determined an erroneously

low value. (The latter group calculated p’ incorrectly and believed they obtained the correct value.) Consider the other results in the paper of Tuschall and Brezonik ( I ) . Experiments failed with Cu2+and a variety of ligands (Table I, ref l),because the ligands are all at least partially reducible. In contrast, they obtained the correct p’ value for Cd(EDTA)2+because it is not reduced, and the lower log p’ value of 8.5 allows measurement of id due to sufficient free Cd2+before the titration end point. Only in the Cd2+EDTA experiment is the measured is predominantly due to id.

In our Cu2+-SFA paper ( 4 ) the correct conditions for Shuman’s equations (2, 3) also exist. The Cu2+-SFA complexes are not reduced (ired = O), and p’ is about lo5 to lo6 assuring a mainly diffusion-controlledprocess. The resulting CL and p’ values are similar to those we obtained by voltammetric methods (8). For example, at pH 6, p’ is 3 X lo5 by ASV (4) and 1 X lo5 by spectrofluorometry. We conclude that Tuschall and Brezonik (1) obtained incorrect p’ values for Cu2+because they chose inappropriate ligands to test Shuman’s model (2, 3 ) . Although ASV can measure nanomolar metal ion concentrations, researchers must use it with care. Implicit and explicit assumptions are involved in determining the free metal ion concentrations from the measured is and the subsequent calculation of p’. LITERATURE CITED Tuschall, J. R., Jr.; Brezonik, P. L. Anal. Chem. 1981, 5 3 , 1986- 1989. Shuman, M. S . ; Woodward, G. P., Jr. Anal. Chem. 1973, 4 5 , 2032-2035. Shuman, M. S.; Cromer, J. L. Environ. Sci. Techno/. 1979, 13, 543-545. Bhat, G. A.; Saar, R. A.; Smart, R. B.; Weber, J. H. Anal. Chem. 1981. 53.2275-2280. Bond; A. M. “Modern Polarographic Methods in Analytical Chemistry”; Marcel Dekker: New York, 1980; Chapter 3, p 438. Chau, Y. K.; Giichter, R.; Lum-Shue-Chan, K. J. Fish. Res. Board Can. 1974, 31, 1515-1519. Eisenrelch, S.J.; Hoffmann, M. R.; Rastetter, D.; Yost, E.; Maier, W. J. Adv. Chem. Ser. 1980, No. 189, 135-176. Ryan, D. K.; Weber, J. H. Anal. Chem. 1982, 5 4 , 986-990.

Gajanan A. Bhat James H. Weber* Department of Chemistry Parsons Hall University of New Hampshire Durham, New Hampshire 03824

RECEIVED for review January

11, 1982. Accepted June 28, 1982. National Science Foundation Grant No. OCE 79-10571 provided partial funding for this research.

Bhat and Weber (1)use the same faulty logic expressed by Sir: The correspondence of Bhat and Weber (1)is mainly Shuman ( 3 )in stating that because Cu-EDTA produced era reiteration of discussion presented both in our original paper roneous results, the organic ligand is suspect, but not the ASV (2) and in a recent exchange of comments ( 3 , 4 ) ;Le., for copper method. Our contention is that the method has not been ASV titrations of several pure organic compounds we were validated for copper with any ligand that has a known p’ value, unable to obtain correct p’ values because of kinetic dissoand therefore the method should not be used for natural-water ciation or direct reduction of complexed copper. We are well organics until it is demonstrated that the method produces aware that if neither phenomenon were occurring, undeaccurate results. tectable to barely detectable currents would have been observed in the case of Cu-EDTA. Instead, we obtained current Contrary to Bhat and Weber’s assertion ( I ) that we are values much higher than expected, thus producing a FCuEDT*condemning Shuman and Woodward’s calculation model (5, that was lo5 lower than the predicted value. 6) to estimate @’, we instead are primarily questioning the 0003-2700/82/0354-2116$01.25/00 1982 American Chemical Soclety

Anal. Chem. 1982, 5 4 , 2117-2120

validity of the ASV titration method. We are confident that Shuman and Woodward’s mathematical treatment is sound if the stated assumptions are met. However, the validity of the assumptions has not been demonstrated for any pure ligand nor for natural water organics. On the other hand, it should be noted that the Shuman and Woodward method is simplistic in implying that one can characterize metal complexes with natural water dissolved organic matter (DOM) or soil fulvic acid (FA) in terms of a single p’. Graphical analysis of complexornetric titration data by Scatchard plots invariably has shown the presence of at least two types of ligand groups (with two characteristic clonstants) in DOM and soil FA (7-10). Scatcliard-derived stability constants, though an improvement on the idea of a single p’, are themselves a simplificationof realicy, and some have criticized this approach as ”curve-fitting”. As Langford (11) ]pointed out, with polyelectrolytes like natural water DOM and soil FA, it is more appropriate to think in terms of stability functions than stability constants. Despite Bhat and Weber’s statement that “Cu2+-SFA complexes are not relduced (ired = O),” no evidence (or even mention) of the lack of complex reduction is found in the reference they cited (their ref 4). Conversely, our work (12, 13) has demonstrated that copper-organic complexes for two humic-rich surface waters in Florida and for polyaspartic acid (mol wt 5400) are completely reducible, albeit at rates somewhat slower than that for ionic copper. Buffle and coworkers (14,15) reported similar results for Pb-humic complexes. These findings may explain the typical titration curve obtained by using the ASV method (including, perhaps, an accurate CL);however, calculations of (3’ for these samples are not accurate since one of the stated assumptions of the ASV method is that the complex be nonreducible (5, 6). It is obvious that many assumptions must be met for the ASV-titration method to yield an accurate p’ value. In general, the validity of these assumptions can be tested experimentally for defied systems containing a single, known ligand. However, for complicated mixtures of unknown ligands, such as occur in natural waisrs, the validity of the assumptions has

2117

been essentially a matter of faith among users of the ASVtitration method. To restate our conclusion, the ASV method should not be used with environmental samples until the procedure has been shown to be accurate when using copper as the titrant. To date, no validation of the procedure has been reported. LITERATURE CITED Bhat, 0. A.; Weber, J. H. Anal. Chem., preceding paper in this issue. Tuschall, J. R., Jr.; Brezonik, P. L. Anal. Chem. 1981, 53, 1986-1989. Shuman, M. S. Anal. Chem. 1982, 5 4 , 998-1000. Brezonlk, P. L; Tuschall, J. R., Jr. Anal. Chem. 1982, 54, 1000-1001. Shuman, M. S.;Woodward, G. P., Jr. Anal. Chem. 1973, 45, 2032-2035. - - - - -. .- . Shuman, M. S.;Woodward, G. P., Jr. Environ. Sci. Techno/. 1977, I f . 809-813. Mantoura, F. F. C.; Riley, J. P. Anal. Chim. Acta 1975, 78, 193-200. Guy, R. D.; Chakrabarti, C. L. Can. J. Chem. 1976, 5 4 , 2600-2811. Breshnahan, W. T.; Grant, C. L.; Weber, J. H. Anal. Chem. 1978, 50, 1874- 1679. Sposito, G.; Holtzclaw, K. M.; LeVesque-Madore, C. S. Soil Sci. SOC. Am. J . 1979, 4 3 , 1148-1155. Langford, C. H. Paper presented at the Symposium on Terrestrial and Aquatic Humic Materials; University of North Carollna, Chapel Hill, NC, Nov 1981. Tuschall, J. R., Jr. Ph.D. Dlssertatin, Universlty of Florlda, Gainesvilk, FL, 1981. Tuschall, J. R., Jr.; Brezonlk, P. L. I n “Terrestrial and Aquatic Humic Materials”; Christman, R. F., Gjessing, E., Eds.; Ann Arbor Science: Ann Arbor, MI, 1982. Buffle, J.; Greter, F. L.; Nenbrini, G.; Paul, J. P.; Haerdi, W. Z . Anal. Chem. 1976. 882.339-350. Greter. F. L.: Buffle. J.: Haerdl. W. J. Nectroanal. Chem. Interfacial Nectrochem’. 1979, f O f , 2111229.

J. R. Tuschall, Jr. Illinois State Water Survey P.O. Box 5050, Station A Champaign, Illinois 61820

P.L. Brezonik* Department of Civil and Mineral Engineering University of Minnesota Minneapolis, Minnesota 55455 RECEIVED for review May 17,1982. Accepted June 28,1982.

Isotopic Analysis of Iodine by Multiphoton Ionization Sir: The problem of monitoring ultralow levels of various radioisotopes in the environment is of current interest. In particular, lBI with a hdf-life of 17 million years, can be used as a tracer of radioactive contamination in the environment over an extended peri(odof time (1-8). lz9Iis produced as a major product (-1%) in the fission process and soon mixes with the natural isotope, 12’1, in the environment. Thus monitoring lz9Ican be accomplished through the detection of the mixed molecule 12711BI. The I2samples for analysis are collected from the products of burning vegetation (9). The sensitivity desired for analyzing these gaseous samples is on the order of lo7 atoms/cm3 of lz9Iin 1015 atoms/cm3 of lz7I, Le., an isotope detecticlri ratio of 1:108. Radioactive counting techniques are not practical for detection of lZgIbecause of its low specific activity and low-energy decay products. The present method of anallysis is based upon neutron activation. This technique is typicidly capable of a discrimination of only between 1:106 to l:lOLi of 12gI:1271 under real experimental conditions (9) although it may be extended with great care. This is due to background radionuclides generated in the neutron activation process. The background contaminants can be chemically removed before activation to produce a discrimination on the order of 1:109 (10-14). However, the 0003-2700/82/0354-2117$01.25/0

time scale of this analysis including the purification process and the decay counting procedure is on the order of several weeks. This method is therefore extremely time-consuming and cannot be used for real-time measurements. It also requires access to a high flux reactor and poses the danger to personnel of handling hazardous radioactive samples. Thus, a real-time or rapid measurement technique which can handle the isotopic discrimination requirements would be preferable. We demonstrate a detection scheme using two-color laser multiphoton ionization which is able to discriminate 2-3 parts of lmIin lo4parts of I2’I in a bulb at room temperature. The use of mass spectrometry in conjunction with this technique promises to reduce this limit to better than 1:108. Multiphoton ionization occurs when ionization follows the absorption of more than one photon by a molecule in the presence of an intense visible or UV light source. As the laser source is tuned to an allowed n-photon transition, this process is greatly enhanced. This is referred to as REMPI, i.e., resonance enhanced multiphoton ionization. In the case of REMPI, ionization occurs via a real intermediate state. Since the density of states above the lowest energy state populated is usually quite high, subsequent absorptions are often resonant or nearly resonant. When the laser is not tuned to a real 0 1982 American Chemical Society