ANALYST, JANUARY 1989, VOL. 114 15 Studies on Bis(crown ether)-based Ion-selective Electrodes for the Potentiometric Determination of Sodium and Potassium in Serum G. J. Moody, Bahruddin B. Saad and J. D. R. Thomas* School of Chemistry and Applied Chemistry, University of Wales College of Cardiff, PO Box 972, Cardiff CFI 3TB, UK Bis(crown ether)-based ion-selective electrodes for sodium and potassium are described, based on the bis[( 12-crown-4)-2-ylmethyI]-2-dodecyl-2-methyl malonate sensor(1) for sodium and the bis[( benzo-I 5- crown-5)-15-ylmethyl] pimelate sensor(l1) for potassium. The best results were obtained when the sensors were used in association with 2-nitrophenyl octyl ether as plasticising solvent mediator and potassium tetrakis(4-chlorophenyl)borate as anion excluder in poly(viny1 chloride) matrices.Electrode slopes were near-Nernstian, with detection limits of <lO-5 M. The electrode features are compared with those of a sodium glass membrane electrode, for sensor I, and with a valinomycin-based potassium electrode, for sensor II. The electrodes are also discussed in relation to others reported for sensors I and II and are shown to be superior. However, although the electrodes described offer promising alternatives to glass electrodes for sodium and valinomycin electrodes for potassium, data for sodium and potassium measurements in blood serum indicate a need for further research in order to improve the correlation with flame photometric measurements. Keywords : ion-selective electrodes; blood serum measurements of sodium and potassium; analate addition; flow injection analysis; bis(crown ether) ion-selective electrodes for sodium and potassium A recent review1 has indicated that an average sized district general hospital in the UK, serving a population of 250000, would carry out 150-200 analyses of sodium and potassium on each working day.Until recently, such analyses were carried out solely by flame photometry, but ion-selective electrodes (ISEs) are being used increasingly for the determination of sodium and potassium in body fluids.l-4 All the potassium ISEs in clinical analysers use valinomycin as the electroactive component, whereas sodium glass mem- branes are the most widely used for the determination of sodium. However, the use of glass membranes presents several difficulties, such as contamination of the glass mem- brane surface by proteins, high resistance, interferences from hydrogen ions and technical problems in the design of various configurations and shapes of cell assemblies.3 With regard to these drawbacks, solvent polymeric membranes are superior.The observation by Mallinson and Truters that potassium forms a 1 : 2 complex with benzo-15-crown-5 has resulted in the synthesis by Kimura and co-workers61" of bis(crown ethers) in which the two crown ether rings are linked together by a single bridge. Owing to the co-operative effect of the two crown ether rings, bis(crown ethers) exhibit remarkable selectivity for alkali metal ions, forming 1 : 1 cation : bis(crown ether) intramolecular sandwich complexes.10 These bis(crown ethers) and their derivatives have been used in neutral carrier type ISEs, for example, a bis(l2-crown-4) [(I), Fig.11 for sodium7-8 and a bis(benzo-15-crown-5) [(11), Fig. 13 for potassium.6.s Earlier work68 has been restricted mainly to poly(viny1 chloride) (PVC) membranes containing a sensor plasticised with 2-nitrophenyl octyl ether. This paper is concerned with studies of the effects of a wider range of plasticising solvent mediators with sensors I and 11, in the presence and absence of an anion excluder, in PVC for the potentiometric sensing of sodium and potassium ions, respectively. The best systems were evaluated for the determination of sodium and potas- sium in serum using analate addition and flow injection analysis (FIA) techniques.To whom correspondence should be addressed. Experimental Materials The reagents and materials used were obtained from the following sources: bis[ (12-crown-4)-2-ylmethyl]-2-dodecyl-2- methyl malonate (I), bis[(benzo-l5-crown-5)-15-ylmethyl] pimelate (11), valinomycin, bis(2-ethylhexyl) adipate (BEHA), glucose and urea from Aldrich (Gillingham, Dor- set, UK); dioctyl sebacate (DOS), a-, b- and y-globulins and I 0 I CH2 I 0 I CH2 W W It Fig. 1. Structures of bis(crown ether) compounds: (I) bis (12-crown- 4)-2-ylmethyl]-2-dodecyl-2-methyl malonate; and (11) bis[(henzo- 15- crown-5)-15-ylmethyl] pimelate16 ANALYST, JANUARY 1989, VOL. 114 human albumin from Sigma (Poole, Dorset, UK); 2-nitro- phenyl octyl ether (NPOE) and bis( 1-butylpentyl) adipate (BBPA) from Fluka (Buchs, Switzerland); and poly(viny1 chloride) (PVC) and Breon resin I11 EP from BP Chemicals (Barry, South Glamorgan, UK).AnalaR grade chlorides of calcium, lithium, magnesium, potassium and sodium were supplied by BDH (Poole, Dorset, UK). The Kent EIL Model 1048-2 sodium glass electrodes were obtained from Kent Industrial Measurements (Chertsey, Surrey, UK). PVC Membrane Electrode Fabrication and Measurements The PVC immobilised sensors were prepared as membranes and assembled into conventional ISEs using established procedures. 11 Unless stated otherwise, the master membranes contained 5 mg of sensor, 360 mg of solvent mediator and 170 mg of PVC. In some instances, a 50% mole ratio of potassium tetrakis(4-chloropheny1)borate (KTCIPB) anion excluder, relative to sensors I and 11.was also added to the membrane. Electrochemical measurements were made with respect to a Corning Model 003 11 602H 4M calomel reference electrode with a Beckman Model 4500 digital pH - millivoltmeter coupled to a Servoscribe 1s potentiometric recorder. Elec- trode calibrations were carried out by spiking with successive aliquots of metal chlorides of known concentration into doubly de-ionised water thermostated at 25 2 0.5 "C. The electrodes were conditioned before use by immersing them for 24 h in a 0.1 M chloride solution of the appropriate primary ion. Between calibrations, the electrodes were stored in the same 0.1 M chloride solutions. The pH - e.m.f. profiles of the best systems for sensors I and 11, and for an EIL sodium glass electrode, were obtained by adding ammonia solution (ca.1 M) to give a pH of about 10.5 and then adding appropriate amounts of concentrated hydrochloric acid to reduce the pH. The membrane resistances of the assembled electrodes were measured using a Bruker E130M potentio/galvanostat. These measurements were performed on electrodes that had been stored for 24 h in a 10-1 M chloride solution of the appropriate primary ion. The separate solution method, at a 10-2 M concentration of the cation, was used to calculate the selectivity coefficients. Determination of Sodium and Potassium in Serum This was performed by analate addition and flow injection analysis (FIA) on blood serum samples stored at 40°C and obtained from the University Hospital of Wales, Cardiff, who also supplied the flame photometric data.Immediately prior to carrying out measurements, the samples were allowed to equilibrate to room temperature over a period of about 1 h. The analate addition procedure is based on the following equation : where A E is the e.m.f. difference on addition of an analyte of volume V,, and concentration C,, to a larger volume of a standard V,, but of lower concentration, C,, and S is the calibration slope of the electrode for a graph of e.m.f. versus log[A], where [A] is the concentration of the primary ion. For FIA, a flow-through sandwich potentiometric detector, in which a PVC-based sensor cocktail, with components in the proportions mentioned above for master membrane construc- tion. was attached to a conductive epoxy support as described by Alegret et al.12 was used [Fig. 2(a)]. The manifold of the flow system is outlined in Fig. 2(b). Solutions were propelled by a multi-channel peristaltic pump (Ismatec Model 1P4) through PTFE tubing (0.8 mm i.d.). Samples were injected into an Omnifit injection valve and the distance between the injection valve and the flow cell was about 5 cm. (a) -5 U Fig. 2. (a) Details of the flow-through ISE detector (a gift from Professor A. A. S. C. Machado, Oporto); and ( b ) the FIA apparatus. A, Block with sensing chamber; B, cover; C, assembled sandwich detector; 1, conductive epoxy; 2, PVC-based ISE membrane; 3, connecting lead; 4, clamping screws; 5 , inlet channel; 6, Corning Model 003 11 602H calomel reference electrode; 7, effluent fluid pool; 8, outlet from effluent pool; 9, peristaltic pump; 10, pulse suppressor; 11, injection valve; 12, sandwich flow-through cell with reference electrode; 13, potentiometer; and 14, recorder For studies relating to serum measurements, all standard sodium and potassium solutions were prepared in mock serum A (consisting of 5 mM potassium chloride, 1 mM calcium chloride and 1 mM magnesium chloride in 0.05 M Trizma base buffer) and mock serum B (consisting of 140 mM sodium chloride, 1 mM calcium chloride and 1 mM magnesium chloride, also in 0.05 M Trizma base buffer), respectively, with the pH adjusted to 7.5 with hydrochloric acid.A direct potentiometric approach (i. e., without prior dilution) was used for the determination of sodium and potassium levels in serum in the analate addition method.Electrode 5, based on bis(l2-crown-4) and NPOE plus anion excluder, was considered to be the best because of its freedom from noise and its general quality. It was used in conjunction with a Corning ceramic junction saturated calomel reference electrode and was placed in mock serum electrolyte A (15 cm3) containing 1 mM sodium chloride, spiked with undiluted serum (0.15 cm3) and the e.m.f. readings were taken. For the determination of potassium in serum, the valinomycin or bis( benzo-15-crown-5) macro-ISE, as appro- priate, was immersed in 0.5 mM potassium in mock serum electrolyte B (15 cm3), spiked with serum (0.15 cm3) and the e.m.f. readings were recorded. For sodium measurements in serum with the FIA mode, an indirect potentiometric approach was used in which serum (0.15 cm3) was diluted to 5 cm3 with electrolyte A. Samples were not diluted for the potassium measurements.For the FIA measurements, the carrier stream consisted of mock serum A and B for the determination of sodium and potassium, respectively, using a flow-rate of 1.98 cm3 min-1 and sample volumes of 100 mm3.ANALYST, JANUARY 1989, VOL. 114 17 Interferences from Biochemicals The bis( 12-crown-4) and bis(benzo-15-crown-5) electrodes were tested for possible interferences from some biochemicals that are normally present13 in blood serum and plasma. Standard amounts of the biochemicals at the upper concentra- tion range (Table 1), e.g., 0.03% urea, were dissolved in 5 mM sodium or potassium chloride, as appropriate, in background mock serum electrolytes A and B and injected into the FIA apparatus while running the corresponding electrolyte (A or B) carrier stream.The mean peak heights for five replicate injections of the same sample containing the same concentra- tion of sodium or potassium ( 5 mM) with, and without, biochemicals were compared. Lifetime Studies The lifetimes of the bis( 12-crown-4) and bis(benzo-15-crown- 5 ) electrodes were studied using membranes plasticised with NPOE and with no anion excluder. For the assessment, both types of conventional electrodes, in conjunction with a Corning Model 003 11 6024 calomel reference electrode, were immersed in doubly de-ionised water (15 cm3) and spiked with non-diluted, pooled serum samples (150 mm3). After expo- sure of the electrodes to the solution for about 90 s, fresh charges of doubly de-ionised water and spikes of pooled serum samples were taken.This procedure was repeated until the supply of pooled serum samples was exhausted, i.e. , after 89 Table 1. Content of biochemicals in plasma and their effect on optimised bis(crown ether) electrodes Normal range in AE*lmV human Relative plasma13 Bis(benzo- molecular lg per Bis( 12-crown-4) 15-crown-5) Constituent mass 100 cm3 (electrode 5 ) (electrode 13) Glucose . . 180 0.065-0.09 O(0.005) O( 0.004) Albumin . . 69000 2.8-4.5 +9.0(0.7) -3.0(0.4) a-Globulin 41 000- 0.3-0.6 +10.2(0.76) 0( 0.006) P-Globulin 5-20 X 106 0.6-1.1 -0.5(0.01) y-Globulin 150 000 0.7-1.5 +8.0(1.2) +3.0(0.8) Mixture of the above six com- ponents - - +10.0(1.6) -2.2(0.05) * A E = difference in e.m.f.of 5 mM solutions of the primary ions with and without the biochemicals (standard deviation given in parentheses for n = 5 ) Urea . . 60 0.02-0.03 t-O.j(O.2) O(0.02) 54 000 r P 50 0 20 40 60 80 ” 0 10 20 30 No. of serum spikes Electrode ageid Fig. 3. during lifetime studies on (H) sodium and (a) potassium ISEs Electrode slope and membrane resistance data obtained spikes. After exposure to the serum, the ISEs were kept in a 0.1 M sodium or potassium chloride solution, as appropriate, and calibrations and resistance measurements were made at various intervals (Fig. 3). Results and Discussion In optimising the performances of the ISEs based on sensors I and 11, the effect of several plasticising solvent mediators was investigated (Tables 2 and 3).The use of NPOE as solvent mediator, together with a 50% mole ratio of KTClPB anion excluder to sensor, yields the best over-all electrode for both sensors (electrodes 5 and 13). These conditions u’ere therefore used in all subsequent studies, unless stated otherwise. The electrodes exhibited fast dynamic response times (<10 s) and a wide pH range (Fig. 4). The electrodes (5 and 13) are superior to those reported pre~iously6~7.14 in that they have improved calibration slopes and wider e.m.f. - log[A] linear ranges, e.g., slopes of 61.0 and 59.2 mV decade-1 (Tables 2 and 3) compared with 53 and 57 mV decade-1. For comparison purposes, a commercial sodium glass electrode (EIL Model 1048-2) was evaluated after condition- ing it in a 10-1 M sodium chloride solution for 2 d before use. This electrode showed poor discrimination for sodium over potassium, lithium and hydrogen ions (Table 2 and Fig.4) and was inferior to the bis(12-crown-4) (sensor I) electrode with respect to dynamic response times (1-2 min as opposed to <10 s). The discrimination of the electrode based on sensor I over other physiologically important cations is generally better than for the ionophores ETH 157 and ETH 227.15 The best bis(benzo-15-crown-5) (sensor 11) potassium electrode possessed properties comparable to those of the valinomycin electrode (Table 3). However, its lifetime was limited. Hence, its slope decreased (Fig. 3) and the membrane resistance increased, relative to the sodium bis( 12-crown-4) electrode, after 70 exposures to serum in a conventional ISE mode (Fig. 3).The sodium bis(l2-crown-4) electrode, on the other hand, showed no significant differences in slope or resistance over the full range of 89 contact periods in serum and continued to function well for more than 1 month (Fig. 3). The results from five replicate analyses on each sample by analate addition using the conventional type electrode and by the FIA approach are depicted in Figs. 5 and 6 together with the flame photometric data obtained at the University of Wales Hospital. Typical FIA peaks for the calibration of electrodes 5 and 13 (constructed from sensors I and 11, respectively, plus NPOE mediator and KTClPB anion excluder) are shown in Fig. 7 together with some data for the serum samples. For the FIA method, the electrodes were recalibrated after every 30 injections of serum.For analate addition, the electrodes were calibrated twice, both before and after each set of serum runs, and it was found that the over-all electrode characteristics (slopes of 58.0 and 59.0 mV decade-1 for sensors I and 11, respectively) did not change under such conditions. However, with respect to precision, it should be borne in mind that for ISE measurements4 with blood serum, precisions of k0.10 and k0.63 mV are required for the sodium and potassium readings, respectively, as the normal range in samples is 135-145 mM (1.9 mV e.m.f. spread) for sodium and 3.5-5.0 mM (9.5 mV e.m.f. spread) for potassium. The correlation between the analate addition and FIA techniques and flame photometry is low (Fig.5), particularly for the determination of sodium using the FIA technique. Similarly, poor correlation is evident on inspection of the Eilood serum data obtained by Tamura et al. ,8 particularly for sodium. However, these data8 were extended greatly by artificial controls, as a result of which correlations of 0.97 or better were obtained. Correlations for the blood serum samples alone were not calculated by these workers.8 For the determination of sodium using the FIA approach, and despite18 ANALYST, JANUARY 1989, VOL. 114 I Table 2. Some characteristics of sodium bis(l2-crown-4) (sensor I) electrodes compared with those of a commercial glass membrane sodium electrode 0 . 1 ( b ) .. r. Electrode to solvent relative to Slope/ Detection Resistance type according KTCIPB Log kEit, B t Electrode No.mediator sensor/mol-% mV decade-' limit/l0-6M /MR B = K B = Li B = Ca B = Mg 1 BBPA" 0 53.0 4.0 23 - 1.43 Not determined owing 2 BEHA* 0 52.0 1.8 44 -0.81 to poor slopes 3 DOS* 0 60.0 1.3 60 -1.38 -2.93 -4.06 -3.96 4 NPOE* 0 60.8 6.3 4 -1.74 -2.40 -3.88 -3.94 5 NPOE* 50 61 .0 6.0 0.08 -1.85 -1.80 -3.68 -3.15 6 EIL glass membrane 60.5 3.2 - -1.58 -1.41 -4.17 -4.21 electrode (Model 1048-2) i Separate solution method, [MI = 10-2 M. Sensor I . Table 3. Some characteristics of potassium bis(benzo-15-crown-5) (sensor 11) and valinomycin electrodes KTCIPB Electrode type relative Detection Log krtBT No. solvent mediator /mol-% mV decade 1 110-6 M /MR B = N a B = L i B = C a B = M g Electrode according to to sensor Slope/ limit Resistance 7 8 9 10 11 12 13 14 BBPA* 0 BEHA* 0 DOS* 0 NPOE* 0 BEHA* 50 DOS* 50 NPOE* 50 NPOE 0 (Valinomycin) Sensor 11.-t Separate solution method, [MI = 10 - ? M. 52.0 60.0 60.5 61 .O 45.5 57.5 59.2 59.6 7.6 7.5 2.5 3.2 5.5 3.5 7.5 8.0 34 43 110 5 3 7 1 8 -3.16 -2.72 -2.53 -2.58 -2.67 -3.05 -3.08 -3.02 Not determined owing to poor slopes -3.23 -4.21 -4.18 -3.25 -4.20 -4.08 -3.28 -4.00 -4.04 Not determined owing to poor slopes -3.16 -3.94 -4.09 -3.14 -3.88 -3.92 -2.88 -3.80 -3.96 -40 I - 80 2 4 6 8 10 PH Fig. 4. E.m.f. responses of various ISEs at various pH values. ( W ) Sodium ISE based on sensor I; (0) EIL Kent Model 1048-2 glass membrane sodium ISE; and (0) potassium ISE based on sensor I1 the convenience and speed of the technique (about 40-60 samples h-1 can be analysed), the relative standard deviation 130 1 .8% a I 2. 120 LL 140 150 120 130 140 150120 130 [ N a * ] / m ~ (flame photometry) Fig. 5. Correlation of mean sodium ISE (sensor I) potentiometric (n = 5 ) and flame photometric ( n = 2) data for (a) analate addition; and ( b ) FIA potentiometry. Equations: (a) y = 1 . 4 3 ~ - 63.4, r = 0.68; and ( h ) y = - 0 . 3 0 ~ + 181.0, r = 0.012 is greater (mean, 3.7%) than for analate addition (mean, For FIA the low correlation with the flame photometric data may be due to poisoning of the membrane with serum components. However, studies of the effect of interferences from various biochemicals on the electrode proved to be inconclusive (Table 1), as sodium (2-6%) is frequently present in the formulation used.16 In commercial ISE analysers, the usual method of eliminating interferences from biochemicals is to place an exclusion membrane over the TSE sensor surface. However, the use of such membranes sometimes fails to remove the interferences from unidentified low relative molecular mass serum components.17918 The optimised potassium electrode based on sensor I113 was found to suffer negligible interference from the biochemicals 2.6%).ANALYST, JANUARY 1989, VOL. 114 - 19 Serum sample Serum sample 0 t 0 a ’ a 0 i __L_ n t 0 L o 0 0 0 0 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 [ K + ] / ~ M (flame photometry) Fig. 6. Correlation of mean potentiometric ( n = S j and flame (n = 2) photometric data for ( a ) potassium ISE (sensor 11): ( b ) potassium ISE (valinomycin); and (c) FIA (ISE of sensor 11) potentiometry.Equations: ( a ) y = 0 . 8 6 ~ + 1.32, r = 0.89; ( b ) = 0 . 8 9 ~ + 1.42, r = 0.88; and (c) y = 0 . 9 4 ~ - 0.41, r = 0.58 - Scan 100 mM -- 10 min - Scan Fig. 7. Typical FIA peaks obtained for (a) the sodium (sensor I) ISE (0.15-cmi serum samples were diluted to 5 cm’ with electrolyte A before taking a 100-mm3 aliquot for FIA); and ( h ) the potassium (sensor 11) ISE studied (Table 1). Also, there was better correlation between the direct (analate addition) ISE and the flame photometric methods (Fig. 6) than for the sodium measurements. Again, the analate addition technique yielded better correlation with flame photometry. Conclusion In the optimisation of ISEs based on the bis(l2-crown-4) (sensor I) and bis(benzo-15-crown-5) (sensor In) compounds, the best PVC matrix membrane electrodes are those based on NPOE as plasticising solvent mediator and which contain a 50% mole ratio of KTCIPB anion excluder relative to the sensor.Both the optimised bis( 12-crown-4) sodium ISE, which has a number of similar, desirable characteristics to the sodium glass electrode, and the bis(benzo-15-crown-5) potas- sium ISE, which has comparable characteristics to the valinomycin electrode, should, therefore, be suitable for monitoring sodium and potassium in body fluids. However, the FIA assessments are disappointing, although the analate addition approach seems promising, particularly for the determination of potassium. Nevertheless, the relative eco- nomical price of these sensors [the bis(12-crown-4) sensor (I) is only one-sixth of the cost of another neutral carrier type sodium sensor (ETH 227)] provides sufficient encouragement for continued research in this area, particularly to improve the ISE - flame photometric correlation by directing attention to the calibrant systems as in the various activities of the International Federation of Clinical Chemists,lq such as those of the European Working Group on Ion-selective Elec- t r odes.20 Financial support and leave of absence from the Universiti Sains Malaysia to one of us (B. B. S.) is gratefully acknow- ledged. Also, Dr. Keith Davies of the University Hospital of Wales is thanked for providing the serum samples and the flame photometric data, and Professor A. A. S. C. Machado of the University of Oporto is thanked for discussions on flow injection analysis, made possible by NATO Grant No.0069. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. References Worth, H. G. J . , Analyst, 1988, 113, 373. Savory, J., Bertholf, R. L., Boyd, C. J . , Bruns, D. E., Felder, R. A . , Lovell, M., Snipe, J . R . , Wills, M. R . , Czaban, J . D., Coffey, K . F., and O’Connell, K. M., Anal. Chrm. Acta, 1986, 180. 99. Oesch, U., Ammann, D., and Simon, W., Clin. Chem., 1986, 32, 1448. Byrne, T. P., Zon-Sef. Electrode Rev., 1988, 10, 107. Mallinson, P. R . , and Truter, M. R., J . Chem. Soc., Perkin Tram. ZZ, 1972, 1818. Kimura, K., Maeda, T., Tamura, H., and Shono, T.. J . Electroanal. Chem., 1979, 95, 91. Shono, T . , Okahara, M., Ikeda, I., Kimura. K., and Tamura, T . , J . Electroanaf. Chem., 1982, 132, 99. Tamura, H., Kumami, K., Kimura. K., and Shono, T., Mikrochim. Acta, 1983, 11, 287. Kimura, K . , Tamura, H., avd Shono, T . , J . Chem. Soc., Chem. Commun., 1983, 492. Kimura, K., and Shono, T., Anal. Chem. Symp. Ser., 1985,22, 155. Craggs, A., Moody, G. J . , and Thomas, J . D. R.. J. Chem. Educ., 1974, 51, 541. Alegret, S . , Alonso, J . , Bartroli, J . , Lima, J. L. F. C., Machado, A. A. S . C . , and Paulis, J. M., Anal. Leu.. 1985, 18, 2291, White, A . , Handler, P., and Smith, E . L., “Principles of Biochemistry,” Fourth Edition, McGraw-Hill, New York, 1972, p. 705.20 14. 15. 16. 17. 18. Tamura, H., Kimura, K., and Shono, T., Anal. Chem., 1982, 54, 1224. Anker, P., Jenny, H-B., Wuthier, U., Asper, R., Ammann, D., and Simon, W., Clin. Chem., 1983, 29, 1508. Davies, 0. G., Moody, G. J., and Thomas, J. D. R., Analyst, 1988, 113, 497. Gadzekpo, V. P. Y . , Moody, G. J., and Thomas, J. D. R., Analyst, 1986, 111, 567. Ladenson, J. H., and Koch, D. D., Anal. Chem., 1983, 55, 1807. ANALYST, JANUARY 1989, VOL. 114 19. 20. Maas, A. H. J., Siggaard-Andersen, O., Weinsberg, H. A . , and Ziglotra, W. G., lnt. Fed. Clin. Chem. News, 1982, 31, 5. Maas, A. H. J., Editor, “Proceedings of the International Federation of Clinical Chemistry, 1st-8th, European Working Group on ISEs,” Private Press, Copenhagen, 1979-1987. Paper 8103378A Received August 22nd, I988 Accepted October 4th, 1988