A series of anion-selective electrodes of the liquid membrane type has been rigorously examined with respect to selectivity characteristics using three different experimental methods. This evaluation gives, for the first time, an understanding of how selectivity ratios depend on concentration levels and other variables and yields a recommended procedure for the consistent numerical determination of selectivity ratios. The resulting selectivity data are examined in terms of a theory of liquid membrane electrodes.Since the introduction of the liquid membrane electrode selective for calcium in 1967 (/), many new electrodes of this type with selective response to both cations and anions have become available. Evaluative studies (2-6) of these electrodes have demonstrated the Nernstian response and dynamic range of these electrodes, but have also revealed their gradual loss of selectivity and general deterioration with time under extended use.Generally speaking, the selectivities of liquid membrane electrodes for the ion of primary interest with respect to common interfering ions are only moderate (7-8). Never-theles^gfhe electrodes have been successfully employed analytical! < direct potentiometry or potentiometric titrations (3, 5, c , 9-13) under suitable conditions where interfering ions are absent or present in low concentrations.If the full usefulness of ion-selective electrodes of the liquid membrane type is to be realized, reliable methods of establishing their selectivity characteristics and accurate numerical data on selectivities must be made available. Unfortunately, there is little agreement in the literature regarding optimal methods of determining selectivities or, even, on the selectivity ratios of a specified electrode. In part, this difficulty arises from a lack of systematic study and from the tendency to report selectivities under a single, and often arbitrary, set of conditions.In this paper we, therefore, set out to evaluate the selectivities of a series of anion-selective, liquid membrane electrodes

No abstract

slowly changing signals. In Table II results are presented for rapidly changing synthetic signals. Resistor R¡ was changed to 15K and integrator times of 50 msec were used for measuring these slopes. Relative errors are again within 0.1% and relative standard deviations are about 0.2%.Results for the determinations of phosphate by the reaction rate procedure are shown in Table III. These results were based on a single phosphate standard. In Table IV results of automatic determinations of glucose in the 5-20 ppm range are presented. In this case adjustments were made to obtain a direct digital readout of the glucose concentration. Rela-tive errors and standard deviations of about 1% were obtained.Many of the rate curves recorded simultaneously indicated that inputs to the rate measuring system were often quite noisy. The results obtained here show the high noise immunity of the integration procedure. In addition, because of its versatility, it is possible to use the same readout system for reactions whose initial rates vary over a wide range, as described in the Instrumentation section.

No abstract

No abstract