A problem associated with the use of a calomel reference electrode in an ISE analytical system.

It was discovered during the testing of a new diluted ISE analytical system that a dark deposit formed in the vicinity of the reference electrode junction and on the junction membrane over a 1-3 month period of use. The effect of the black deposit was to shorten the life of the restricted flow reference membrane and increase the time required to maintain the system in its optimum state. Elemental analysis of the deposit revealed the presence of both mercury and sulfur. The rate of deposit build-up was found to depend on the concentrations of both the buffer and the microbicide in the system's reagents. The cause was traced in part to the generation of sulfide ions as a breakdown product of the microbicide. The disproportionation of calomel, Hg2Cl2, in saturated KCl to give Hg2+chloro-complexes and their reaction with sulfide ions at the reference junction resulted in HgS being deposited. Design changes over previous systems contributed to the effect by increasing residence time of calibrant solution at the reference junction and decreasing the frequency of reference electrolyte and membrane changes. Adding complexing agents to the reference fill solution lessened but did not eliminate the problem. The problem was solved without reformulating the reagents by using a reducing agent proximate to the calomel mercury amalgam to reduce mercuric ions to metallic mercury. This also prevents discharge of environmentally undesirable mercuric ions into the waste solution.

Frozen human serum reference material for standardization of sodium and potassium measurements in serum or plasma by ion-selective electrode analyzers.

Three interlaboratory round-robin studies (RR1, RR2, and RR3) were conducted to identify a serum-based reference material that would aid in the standardization of direct ion-selective electrode (ISE) measurements of sodium and potassium. Ultrafiltered frozen serum reference materials requiring no reconstitution reduced between-laboratory variability (the largest source of imprecision) more than did other reference materials. ISE values for RR3 were normalized by the use of two points at the extremes of the clinical range for sodium (i.e., 120 and 160 mmol/L), with values assigned by the flame atomic emission spectrometry (FAES) Reference Method. This FAES normalization of ISE raw values remarkably improved all sources of variability and unified the results from seven different direct ISE analyzers to the FAES Reference Method value. Subsequently, a three-tiered, fresh-frozen human serum reference material was produced to the specifications developed in RR1-RR3, was assigned certified values for sodium and potassium by Definitive Methods at the National Institute of Standards and Technology (NIST), and was made available in 1990 to the clinical laboratory community as a Standard Reference Material (SRM); it is now identified as SRM 956. Albeit retrospectively, we show how applying an FAES normalization step identical to that used in RR4/5 to the ISE data for SRM 956 after the NIST Definitive Method values were known, consistently moved the ISE results for RR3 closer to the true value for Na+ and K+.

Establishing the direct-potentiometric "normal" range for Na/K: residual liquid junction potential and activity coefficient effects.

The observed reference ranges for sodium and potassium as determined by direct potentiometry vary from instrument to instrument, depending on the composition of the calibration standards. To resolve the existing confusion as to which reference intervals are most appropriately considered "normal," we propose a straightforward convention (based on plasma-water concentration units) in which the difference between direct (undiluted sample) and indirect (diluted sample) methodologies is accounted for by the volume displacement effect of proteins, lipids, and other dissolved substances in a typical plasma sample. Thus, the proposed reference intervals for sodium and potassium are approximately 7% greater by direct potentiometry than by procedures involving dilution. Data consistent with this convention can be obtained with a variety of aqueous-based calibrants, provided care is taken to minimize the errors resulting from activity coefficient and liquid junction potential effects. Additional experimental results are presented to show that these effects also account for the apparent suppression of the sodium ion concentration observed in the presence of bicarbonate ion.