J. Chem. SOC. Faraday Trans. 1 1986,82 1135-1143 Aspects of the Optimization of Poly(viny1 chloride) Matrix Membrane Ion-selec tive Electrodes J. D. R. Thomas Department of Applied Chemistry Redwood Building U WIST P.O. Box 13 CardijTCFl 3XF PVC matrix membrane electrodes have been improved for dealing with an extended range of ions by the design of new electroactive components with appropriate plasticising solvent mediators. New approaches to calibrating the electrodes with ion buffers increases confidence in their use. Ion permeation studies in membranes demonstrate ion-selectivity and form a basis for electrochemical response mechanism of PVC matrix membranes containing trapped liquid ion-exchangers and neutral carrier complexes respectively. Immobilization of active sites on the membrane matrix has led to only limited improvements in ISE design.Rather more successful is the use of alternative solvent mediators for overcoming sample matrix interferences as in the case of samples containing anionic surfactants and certain biochemical components. PVC matrix membrane electrodes can be designed for dealing with various measuring situations such as small sample sizes flowing samples and muscle fluids. Poly(viny1 chloride) (PVC) matrix membranes have boosted the development of liquid ion-exchanger and neutral carrier ionophore ion-selective electrodes (ISE).l? These essentially liquid-based membranes may be easily prepared by the simple expedient of dissolving PVC in a solution of liquid ion-exchanger or neutral carrier ionophore systems in tetrahydrofuran and slowly evaporating at room temperat~re.l-~ The PVC matrix membrane ISE possess similar response behaviour to their liquid membrane counterparts but the much simpler design the considerable economy in sensor materials and sometimes longer lifetimes are additional desirable features that have led to their more general use.The PVC-based membranes consist of the electroactive component (ion-exchanger or ionophore) and plasticising solvent mediator trapped in the PVC matrix. Early work6 showed that ca. 30% content of PVC was appropriate since smaller proportions led to fragile membranes while larger amounts led to sluggish electrochemical response. The plasticising solvent mediator should preferably be of high viscosity for longer functional lifetimes of ele~frodes,~-~ and the nature of the solvent mediator itself can have a significant influence on the selectivity of the electroactive component.1° Choice of Electroactive Components and Calibration The single most important factor in optimizing ISE concerns the choice and development of selective electroactive components.Thus the origins of the successful commercial liquid ion-exchanger ISE lie in the use of materials with proved selectivity in solvent extraction work such as dialkylphosphates for calcium ISE and long aliphatic chain quaternary ammonium salts for nitrate and perchlorate ISE. There is a close relation between electrode selectivity coefficients kTtg and competitive ion-pair extraction constants (EQB/EQA) for Q extractantll (table 1).This and earlier studies on naturally 1135 1136 Br- k!%3.B I- 3.01 17 PVC Matrix Membrane ISE Table 1. Ratios of ion-pair extraction constants (EQ,) for the tetrabutylammonium ion compared with selectivity coefficients kp&B of a PVC matrix membrane nitrate ISE with tetradocylhexa- decylammonium nitrate sensora counter ion B NO; 1.39 log (extraction constant) (log EQx) 1 .o 1 c1- -0.11 0.005 0.032 a Data from ref. (1 1). Value for liquid membrane nitrate ISE. c10 3.48 800 1 20 1.29 O.Olb 0.79 EQB/EQN03 ~~ ~ ~ 42 occurring ionophores e.g. valinomycin for potassium ISE has been followed by the synthesis of improved materials and of organic ionophores with highly selective complexing properties for certain ions.Investigations of the nitrate ISE based on tetra-alkylammonium nitrate in PVC have shown that the linear calibration range is extended when the alkyl chain is C or longer.12 This is significant since the measurement of nitrate is of considerable practical interest and the interfering perchlorate iodide and certain other anions are purely the concern of research laboratories. However precautions must be taken against excessive amounts of chloride in samples. Among the organophosphate calcium ion sensors the calcium bis-dialkylphosphate systems have been superseded by calcium bis-di(4-octylpheny1)phosphate used in conjunction with dioctylphenylph~sphonate~~-~~ or with certain trialkylphosphates (usually tripentyl- or trioctyl-phosphate).16 It matters but little in terms of selective calcium ISE response whether the octyl group is the n-octyl chain or the isomeric 1 ,I ,3,3-tetramethylb~tyl.~9 1 5 9 l6 This is important since the starting material in the synthesis of the sensor is 4-alkylphenol and the 1,1,3,3-tetramethylbutyI isomer is much more ac~essible.~~ 1 7 9 la The neutral carrier sensor N,N-di[( 1 1 -ethoxycarbonyl)-undecyl]N,N-4,5-tetramethyl- 3,6-dioxaoctane diamide,1g-21 is also a very effective calcium ion sensor.Despite its more difficult synthesis it frequently possesses an edge such as by rather better selectivity for calcium over magnesium.22 This neutral carrier is in a short list23 of five of the most attractive synthetic carrier molecules for sensing metal cations the other four being selective for Li+ Na+ and Ba2+ as appropriate.There is a recent review24 on neutral carrier based ISE and new ones continue to be added such as a macrocyclic polyether amide with two polyether rings for calcium ionsz5? 26 and a lipophilic crown-4-derivative for lithium.27 Apart from improved selectivity the new calcium ISE systems behave well during calibration and use with ion-buffers when calibrations are possible down to ca. lo- mol dmd3 [Caz+]. The use of ion-buffer systems for calibrating ISE stems from solutions containing various proportions of copper@) ions to EDTA or NTA [or calibrating copper electrodes.28 This has been extended to calcium ISEf2 and adopted and tested by others.139 21y 29-33 It is significant that the calcium ion-buffer calibration extrapolates to that for serially diluted calcium ion standards.Hitherto methods of obtaining calcium ion standards down to mol dm-3 have been based either on the serial dilution of solutions containing the metal ion with a single complexing ligand such as NTA EDTAl31 1 4 9 21y 34 or EGTA34 35-38 or on the variation of pH in a solution of fixed metal to ligand concentration^.^^ 1 3 9 31 Such calibrations are laborious and the conditions are not applicable to those found in uiuo since both the pH and the concentration of calcium buffering ligands vary only slightly in biological 1137 J. D. R. Thomas fluids and cells. In viuo conditions are more closely approached by a method39 where free calcium ion concentrations are adjusted by adding increments of total calcium to a mol dm-3 EGTA solution at pH 7.4.This method is essentially a titration of EGTA by Ca2+ ions and therefore lacks precision in the unbuffered region near to the equivalence point. To overcome the above shortcomings a method has recently been developed32* 33 whereby standard free calcium ion concentrations from 10-8-10-3 mol dm-3 may be obtained by means of a single calibrating titration with a calcium solution of a mixture of calcium-buffering ligands namely ethylene glycol bisv-aminoethyl ester)-N,N,N’ N’-tetra-acetic acid (EGTA) N-(2-hydroxyethyl)ethylenethiaminetetra-acetic acid (HEEDTA) and nitrilotriacetic acid (NTA). The effectiveness of the calibration procedure at pH 7.4 (Tris buffer adjusted with hydrochloric acid) has been demonstrated on PVC matrix membrane calcium ISE based on calcium bis-di-4-( 1,1,3,3-tetramethylbutyl) phenylphosphate electroactive material and a commercial membrane containing a neutral carrier electroactive Thus the neutral carrier calcium ISE may be used down to lob8 mol dm-3 [Ca2+] in the presence of millimolar levels of magnesium.33 The same work showed that a PVC matrix membrane organophosphate based electrode with a solvent mediator mixture of dioctylphenylphosphonate/decan-1-01 in the ratio of 1 10 (v/v) can be used to determine these higher levels of rnagne~ium.~~ So far there is no magnesium ISE.The best approaches for the direct potentiometric measurement of magnesium therefore fall on the divalent e l e c t r ~ d e ~ ~ ~ ~ ~ or on an electrode based on a neutral carrier i o n ~ p h o r e .~ ~ However the neutral carrier system still shows preference of calcium over magnesium by a factor of ca. 20 while the divalent electrode can fail to give accurate analysis in standard analyte~.~O Attempts to obtain electroactive sensors for magnesium from complexes of the metal with polyalkoxylates have 42 for example the extra methyl in polypropoxylates over polyethoxylates was sufficient in its steric influence to yield just another calcium ISE41 instead of the barium ISE of polyethoxylate-metal ion complexes.8 An interesting feature of the polypropoxylate-based calcium ISE is its high interference from lithium ions,41 which has been exploited in the development of an ISE for lithium ions.43 Of the barium and lithium polyalkoxylates investigated as electroactive components for lithium the tetraphenylborate of the barium complex with PPG- 1025 polypropoxylate used in conjunction with dioctylphenylphosphonate in PVC gave the best lithium ISE.43 Although this is not as selective as lithium electrodes previously described with electroactive components of N,N,N’,N’-tetraisobutylcyclohexane-cis- 1,2-dicarboxylic diamide44 and dodecylmethyl- 14-cr0wn-4,~~ respectively the response towards lithium ions is of sufficiently high quality for it to be a realistic ISE.Furthermore there is the added advantage that the polypropoxylate sensor is much more easily prepared in the laboratory than are the other neutral carrier sensors.Tests with lithium-containing serum samples obtained from patients under treatment for manic depression show the polypropoxylate-based lithium ISE to be more effective in the microconduit flow-injection analysis measurements studied than either of the other two ISE.4s However there is still scope for considerable improvement in lithium ion sensors. Selectivity of PVC Matrix ISE Membranes Both potentiometric and ion transport studies indicate that PVC-based ISE membranes tend to be permselective to counter ions and accordingly their electrical properties are not significantly influenced by hydrophilic co-ions. Experimental evidence for cation permselectivity has been obtained for all analytically relevant neutral carrier based membrane electrodes and a cation transference number close to 1.0 has been found in electrodialysis experiment^.^^ Likewise cation permselectivity has been proved for organophosphate-based PVC ISE membrane^,^^ i.e.the membrane extracts and 1138 PVC Matrix Membrane ISE Table 2. d(C"/C')/dt data for solution/membrane systems [data from ref. (47)]. Membranes calcium from Orion 92-20-02 liquid ion-exchanger in PVC (exchanger batch A) [d(C"/C)/dt]/10-7 s-* series solution B(externa1) / mol dm-3 19 2.5 18 2.2 solution A(interna1) / 1 0-3 mol dm-3 solution 45CaC1 CaCI 45CaCl BeC1 45CaCl MgCI CaCl 45~a~15 BeC1 45CaCl MgCI 45CaC1 8.0 8.4 8.5 9.1 45CaCI SrCl 45CaCI s~cI 45CaCI BaCl 45CaC1 2 1 4 3 6 5 7 8 9 10 BaC1 in ISE.8.4 7.5 permeates cations from the sample solution but only negligibly small amounts of hydrophilic anions. The extent of ion permeation is related to selectivity of the ISE.479 48 Thus the data of table 2 for a calcium ISE membrane show that calcium permeates more than other cations as measured by d( C"/C)/dt data for those conditions where radiotracer flux between the initially inactive (tracer concentration C") and active (tracer concentration C ) solutions increases linearly with time. The unexpectedly low d(C"/C')/dt value for 45Ca on the active side and with beryllium chloride solution on the other side of the membrane (table 2) (on the premise that beryllium ions interfere strongly with calcium ISE response) is attributed to the high affinity of the ion-exchanger in the membrane for beryllium.This is evidenced in separate experiments by the increased concentration of beryllium in the membrane.48 Another interesting feature of ion permeation in liquid ion-exchanger based PVC matrix membranes concerns the effect of solvent mediator (table 3). Thus membranes with calcium ion-sensor and solvent mediators which give good calcium ISE properties give high values for d(C"/C)/dt data.49 Also the equal sensing of calcium and magnesium ions by the divalent ion electrode based on Orion 92-20-32 divalent exchanger (dialkyl- phosphate plus decan-1-01 solvent mediator) is reflected in the value of 32 x s-l for d(C"/C)/dt for permeation of 45Ca from calcium chloride through a membrane of the exchanger in PVC into counter solutions of calcium chloride and magnesium chloride respe~tively.~~ As mentioned above tetraphenylborates (TPB) of barium complexes with certain polyethoxylates e.g.Antarox C0880 (a nonylphenoxypolyethaneoxy unit) are good sensors for barium ISE.8 Relevant to the calcium ion permeation in the alkyl- and alkylphenyl-phosphate membranes discussed above there is very little permeation of barium ions from initially radioactive barium chloride solutions through the barium-form membranes into initially inactive barium chloride Nevertheless in counts of the membranes there is a continued uptake of 133Ba with time (table 4). This trend corresponds to observations by other that only a small proportion of available ligands in carrier type membranes participate in complex formation with cations.However in this case it is more likely to be the nature of the polyalkoxylate-metal complexation that is responsible for the small extent of permeation. Thus the confor- mation of the polyalkoxylate chain around the barium ions gives a stable entity with only small lability but sufficient for some ion transfer to produce the potentiometric response The conformation of the barium-polyalkoxylate complex is destroyed when a solvent mediator batch D of Orion 92-20-02 liquid ion-exchanger (for reference) dioctylphenylphosphonate tributylphosphate tripen tylphosphate trioctylphosphate Table 4. 133Ba uptake by eight membranes containing Ba,.- Antarox C0880 - TPB (0.04 g) + 2-nitrophenylphenyl ether (0.36 g) in PVC (0.17 g)a uptake of 133Ba time/h C"/C' Cact/Cinact by membrane (%) 9.6 5.2 12 6.0 6.5 a Data from ref. (49). Table 3. d(C"/C')/dt data for solution/membrane systems with membranes containing calcium bis-di(4- 1,1,3,3-tetramethylbutylphenyl)pho~phate and various solvent mediators in PVC mol dmd3 calcium chloride on each side of membrane one side active with 45Caa J. D. R. Thomas proportion of detection radiotracer limit of in membrane corresponding ion-selective ion-selective at end of d(C"/C') experiment /d t 18 (% ) 4.0 2.4 tri( 1,1,3,3-tetramethylb~tyl)phosphate membrane no. 4 2 - 3.2 0.01 24 48 0.01 6 - 3.1 2.1 0.03 0.02 72 0.01 144 0.01 503 a Data from ref.(51). Active solution Inactive solution potential is applied to electrodes placed in solutions on each side of the membrane.52 Thus the membranes were unable to maintain stable current flows for applied potentials of 2 V for the currents of 15 pA decayed to 7 pA in 4 h recovered on reversal of the polarity and fell to 5 pA in the next 4 h with no further evidence of recovery of current upon further potential reversals.52 Organophosphate-based membranes on the other hand are characterized by stable current flows over prolonged periods and over successive polarity reversals.52 Immobilization of Electroactive Sites Functional lifetimes of ISE based on polymer matrix membranes with trapped liquid ion-exchanger electroactive components are normally considerably shorter than for solid-state membrane electrodes.Among the reasons for this is the leaching of active components from the membranes especially at the membrane-solution interface. mol dmP3 133BaC12. mol dmV3 BaCl,. Not counted. electrode electrode /mol dm-3 Ca2+ lifetime 32 2.6 59 37 25 7.5 x 3.0 x 1.2 x 6.0 x 1.1 x 6.0 x 1 .o 1.8 2.4 4.0 1.6 1.4 1.8 1.4 5.6 6.6 b - 6.6 1139 3 months 3 months 1 week 3 months 4 weeks 3 months 1140 PVC Matrix Membrane ISE Various attempts have been made to lessen this detrimental effect by the covalent bonding immobilization of active components to the polymer matrix membrane upp port.^^-^* For anionic surfactant ISE a tertiary amine was bound to the ends of PVC chains and converted into a positively charged quaternary ammonium site by reaction with an alkyl Electrode membranes cast from tetrahydrofuran solution were conditioned in aqueous solutions of sodium dodecylsulphate in order to exchange the bromide ions with dodecylsulphate.Grafted cationic surfactant-sensitive electrode membranes were made by the low- temperature polymerization of vinyl chloride using the SO; radical anion.53 This gave a polymer of relative molecular mass ~ 7 7 0 0 0 in which ca. one-third of the polymer chains terminated with sulphonated end groups. Membranes were conditioned by exchanging the associated hydrogen ions for the desired surfactant cation.53 Among the approaches to immobilizing the ion-exchange site of calcium ISE is the binding of the styrene-b-butadiene-b-styrene (SBS) triblock e l a ~ t o m e r .~ ~ - ~ ~ Thus membranes were produced by firstly cross-linking SBS with triallyl phosphate followed by alkaline hydrolysis of the resulting covalently bound trialkylphosphate grouping to yield a pendant dialkylphosphate capable of acting as calcium ion sensor.54 Good calcium ISE of lifetimes in excess of 6 months were obtained although of limited selectivity towards calcium over important ions such as sodium and magnesium. Use of triundec- 10-enyl phosphate and of diallyl phenylphosphonate instead of triallylphosphate gave robust cross-linked membranes but without significant improvements in ~electivity.~~ An alternative approach for immobilizing the organophosphate sensor is to condense monodecyl dihydrogen phosphate with hydroxy groups of a partially hydrolysed vinyl chloride-vinyl acetate copolymer (copolymer VAGH).57 The resultant product when fabricated into membranes incorporating dioctylphenylphosphonate yielded good calcium ISE but without the expected advantage of extended lifetime over the PVC matrix membrane electrodes with a physically trapped sensor.57 Grafting of mono- octylphenylphosphonate to copolymer VAGH in PVC with calcium ion-sensor led to only a slightly lengthened lifetime while phosphonated polystyrene by Friedel-Crafts and free-radical processes was rather less succe~sfu1.~~ Membrane Modifications to Meet Sample Matrix Interferences Incorporation of a phenyl group between alkyl and the phosphorus oxygen of the dialkylphosphate calcium ion-sensors leads to less interference from pH.13-15 However freedom from pH interference at pH < 5 is best for a calcium ISE based on the macrocyclic siloxane tetracosamethylcyclododecasiloxane.59 A significant interference of calcium ISE response is that by anionic surfactants e.g.the addition of such a small amount as 2 x mol dm-3 sodium dodecylsulphate (SDS) lowers the e.m.f. of calcium ISE in many circumstances.60 This observation also applies to electrodes made from a commercial (Philips IS 561/SP) PVC matrix membrane for calciums1 (table 5). Replacement of some of the dioctylphenylphosphonate (DOPP) plasticising solvent mediator by decan- 1-01 in the organophosphate membrane electrodes reduces the interference by dodecylsulphate but at the expense of some loss of calcium ion-selectivity.60 Mechanistic studies by X-ray fluorescence and chromatography on calcium ISE membranes of bis-di[4-( 1,1,3,3-tetramethylb~tyl)-phenyl]phosphate sensors and DOPP solvent mediator in PVC show that SDS is an effective agent for leaching membrane components especially DOPP.61 Electrodes of membranes with the same sensor but with trioctyl phosphate as plasticising solvent mediator instead of DOPP are much less susceptibles1 to interference by SDS (table 5).The leaching of lipophilic membrane components raises concern for the clinical/bio- chemical field where natural surfactants are frequently present in samples.This concern 1141 J. D. R. Thomas Table 5. AE Caused by mol dm-3 SDS on calcium PVC ISE responsea - 3 CaC12/mV 1 0-2 mol dmP3 1 0-4 mol dm-3 solvent mediator component CaCl,/mV - 68 - 70 - 61 -2 - 65 - 70 -113 dioct ylphen ylphosp hona te-PVC trioctylphosphate-PVC tripen t ylp hosp hate-PVC Philips IS 56 1 /SP Ca2+ (PVC) membrane a Data from ref. (61). All electrodes based on calcium bis-di[4- (1,1,3,3-tetramethylbutyl)phenyl] phosphate except for Philips IS 561/SP Ca2- membrane. is justified since there has been much debate on which measuring technique provides the most accurate values of ion concentration for clinical use.s2 Related to these are studies on the effect of protein concentration on ISE measurements of ionized calcium.63 To complement such observations and studies on the interferences of calcium ISE by anionic surfactants several biochemical components in calcium-containing solutions have been studied for their effect on various forms of calcium ISEs4 (table 6).Here also the PVC electrodes with trioctylphosphate solvent mediator and calcium bis-di[4-( 1,1,3,3- tetramethylbutyl)phenyl]phosphate sensor are the most resistant.s4 However in this case contrary to the above observations for added anionic surfactant the commercial Philips IS 56 1 /SP membrane is reassuringly relatively resistant to the effect of added biochemical component (table 6). Architectural Optimization PVC matrix membrane electrodes can be modified to suit various measuring situations.For example samples down to 1 mm3 can be handled by sandwiching between a specially prepared flat surface of a reference electrode fabricated from a glass cone and socket and a PVC membrane electrode of conventional design.65 Flow injection analysis samples can be analysed by suitably positioned conventional ISE or by ISE based on tubular membranes.6s Fluids inside biological cells can be measured with specially designed microelectrodes with PVC matrix membranes and conventional inner filling ~ o l u t i o n s . ~ ~ ~ 24 The above architectural forms are either too fragile or otherwise unsuitable for plunging into muscle flesh for measurements of ions in muscle fluids. For this and other applications there is the design of all-solid-state PVC matrix membrane micro-electrodes using glass and Perspex capillaries back-filled with silver-conducting epoxy surfaced with a PVC matrix membrane for interfacing with the test fluides7 Conclusion Ion-selective electrodes have tremendous scope in terms of practical applications.However their use as in-dwelling sensors imposes harsh criteria for dealing with samples of very diverse character. PVC matrix membrane electrodes can be optimized for meeting many of these demands but there continues to be the need for further investigations in order to recognize and elucidate problems so that the ISE can be improved for use in new applications. ~~ Philips membrane electrode /mol dm-3 1142 component (0.05 mol dm-3) - 1.6 -6.8 - 0.6 - 10.5 1.7 0.2 -1.0 -6.5 0.8 - 2.8 0.2 -0.1 PVC Matrix Membrane ISE Table 6.AE Caused by biochemical components in calcium chloride in 0.15 mol dm-3 sodium chloride solutions on calcium ISE responsea AE caused by added components to solutions/mV TOPd /mol dm-3 TPPC /mol dm-3 DOPPb /mol dm-3 - 1.1 ~~ - 3.5 - 9.8 - 7.0 -0.1 - 11.3 - 85.7 - 40.0 - 15.0 - 9.7 -41.5 - 24.0 - 17.4 2.5 - 6.5 ~ -2.8 -7.2 -1.5 -3.8 0.5 1.9 0.2 0.2 2.8 -0.2 0.2 -0.3 -0.1 ~ ~ ~ 0.4 - 1.8 - 26.5 - 25.1 - 39.8 - 59.8 ~~ 3.3 1.7 - 0.8 -0.5 ~ _ _ 0.4 ~ 0.0 0.4 5.7 -0.1 - 0.5 ~~ ~~ lo- SDS (for comparison) -85 DBSS (for comparison) - 12.5 cholic acid cholesterol lecithin vitamin D urea glucose - 13.1 4.1 - 5.6 - 3.0 -0.9 -0.2 a Data from ref.(64). All electrodes based on calcium bis-di[C( 1,1,3,3-tetramethylb~tyl)phenyl]- phosphate except for Philips IS 561 /SP Ca2+ membrane in PVC with solvent mediator indicated. Dioctyl phenylphosphonate. Tripentyl phosphate. Trioctyl phosphate. The author thanks the S.E.R.C. for supporting this work by grants and studentships through the CASE scheme in association with Unilever Research (Port Sunlight Laboratory) and the Central Electricity Research Laboratories. Thanks are also extended to the British Council for Visiting Fellowships and the University of Wales UWIST the University of Technology Baghdad and the Foundation of Technical Institutes Baghdad for studentships.Also the many coworkers are thanked for their dedication. References 1 G. J. Moody R. B. Oke and J. D. R. Thomas Analyst (London) 1970,95,910. 2 G. J. Moody and J. D. R. Thomas in Ion-Selective Electrodes in Analytical Chemistry ed. H. Freiser (Plenum Press New York 1978) vol. 1 p. 287. 3 G. J. Moody and J. D. R. Thomas in Ion-Selective Electrode Methodology ed. A. K. Covington (CRC Press Boca Raton Florida 1979) vol. 1 p. 11 1. 4 A. Craggs G. J. Moody and J. D. R. Thomas J. Chem. Educ. 1974,51,451. 5 G. J. Moody and J. D. R. Thomas Ion-Selective Electrode Reviews 1979 1 3. 6 G. H. Griffiths G. J. Moody and J. D. R. Thomas Analyst (London) 1972,97,420.7 J. E. W. Davies G. J. Moody and J. D. R. Thomas Analyst (London) 1972,97 87. 8 A. M. Y. Jaber G. J. Moody and J. D. R. Thomas Analyst (London) 1976,101 179. 9 A. Hulanicki M. Maj-Zurawaska and R. Lewandowski Anal. Chim. Acta 1978 98 151. 10 A. Craggs L. Keil G. J. Moody and J. D. R. Thomas Talanta 1975 22,907. 11 G. J. Moody and J. D. R. Thomas Chem. Ind. (London) 1974,644. 12 H. J. Neilsen and E. H. Jansen Anal. Chim. Acta 1976 86 1. 13 J. Ruzicka E. H. Hansen and J. C. Tjell Anal. Chim. Acta 1973,67 155. 14 H. M. Brown J. P. Pemberton and J. D. Owen Anal. Chim. Acta 1976,85 261. 15 L. Keil G. J. Moody and J. D. R. Thomas Anal. Chim. Acta 1978,% 171. 16 G. J. Moody N. S. Nassory and J. D. R. Thomas Analyst (London) 1978 103 68. 17 J.D. R. Thomas Lab. Pract. 1978 27 857. 18 A. Craggs P. G. Delduca L. Keil B. J. Key G. J. Moody and J. D. R. Thomas J. Inorg. Nucl. Chem. 1978,40 1483. 19 D. Ammann E. Pretsch and W. Simon Helv. Chim. Acta 1973,56 1780. 20 D. Ammann R. Bissig M. Guggi E. Pretsch W. Simon I. J. Borowitzand L. Weiss Helv. Chim. Acta 1975 58 1535. 1143 J. D. R. Thomas 21 D. Ammann M. Guggi E. Pretsch and W. Simon Anal. Lett. 1975 8 709. 22 T. S. Tsieng personal communication. 23 W. E. Morf and W. Simon Hung. Sci. Instrum. 1977 41 1. 24 D. Ammann W. E. Morf P. Anker P. C. Meier E. Pretsch and W. Simon Ion-Selective Electrode Revs. 1983 5 3. 25 K. Kimura K. Kumani S. Kitazawa and T. Shono J. Chem. Soc. Chem. Commun. 1984,442. 26 K. Kimura K. Kumami S. Kitazawa and T.Shono Anal. Chem. 1984,56 2369. 27 S. Kitawawa K. Kimura H. Yano and T. Shono J. Am. Chem. Soc. 1984 106,6978. 28 R. Blum and H. M. Fog J. Electroanal. Chem. 1972,34 485. 29 A. Craggs G. J. Moody and J. D. R. Thomas Analyst (London) 1979 104,412. 30 A. Craggs G. J. Moody and J. D. R. Thomas Analyst (London) 1979 104 961. 31 0. Scharff Anal. Chim. Acta 1979 109 291. 32 M. Otto and J. D. R. Thomas Anal. Proc. 1984,21 369. 33 M. Otto P. M. May K. Murray and J. D. R. Thomas Anal. Chem. 1985,57 151 1. 34 Y. S. Kim and G. M. Padilla Anal. Biochim. 1978 89 521. 35 A. Coray C. H. Fry P. Hess J. A. S. McGuigan and R. Weingart J. Physiol. 1980 305 60P. 36 C. 0. Lee D. Y. Uhm and K. Dresdner Science 1980,209 699. 37 E. Marban T. S. Rink R. W. Tsien and R. Y.Tsien Nature (London) 1980 286 845. 38 B. A. Bulas and B. Sacktor Anal. Biochem. 1979,95 62. 39 D. M. Bers Am. J. Physiol. 1982 242 C404. 40 S. K. A. G. Hassan G. J. Moody and J. D. R. Thomas Analyst (London) 1980 105 147. 41 A. M. Y. Jaber G. J. Moody and J. D. R. Thomas Analyst (London) 1977,102,943. 42 P. H. V. Alexander G. J. Moody and J. D. R. Thomas to be published. 43 V. P. Y. Gadzekpo G. J. Moody and J. D. R. Thomas Analyst (London) 1985 110 1381. 44 A. F. Zhukov D. Erne D. Ammann M. Guggi E. Pretsch and W. Simon Anal. Chim. Acta 1981 131 117. 45 S. Kitzazwa K. Kimura H. Yano and T. Shono Analyst (London) 1985,110,295. 46 V. P. Y. Gadzekpo G. J. Moody and J. D. R. Thomas Analyst (London) in press. 47 A. Craggs G. J. Moody J. D. R. Thomas and Anne Willcox Talanta 1976 23 799.48 A. M. Y. Jaber G. J. Moody J. D. R. Thomas and Anne Willcox Talanta 1977,24,655. 49 G. J. Moody and J. D. R. Thomas J. Power Sources 1983,9 137. 50 A. Craggs B. Doyle S . K. A. G. Hassan G. J. Moody and J. D. R. Thomas Talanta 1980 27 277. 51 B. Doyle G. J. Moody and J. D. R. Thomas Talanta 1982 29 257. 52 B. Doyle G. J. Moody and J. D. R. Thomas Talanta 1982,29,608. 62 J. H. Ladenson Anal. Proc. 1983 20 554. 63 R. B. Payne RSC International Symposium on Electroanalysis in Biomedical Environmental and Industrial Sciences UWIST Cardiff 5-8 April 1983 Paper 13. 64 S. A. H. Khalil G. J. Moody and J. D. R. Thomas Analyst (London) 1985 110 353. 65 J. D. R. Thomas in Ion-Selective Electrodes ed. E. Pungor (Akademiai Kiad6 Budapest 1978) p. 175. 66 A. J. Frend G. J. Moody J. D. R. Thomas and B. J. Birch Analyst (London) 1983,108 1357. 67 S. A. H. Khalil G. J. Moody J. D. R. Thomas and J. L. F. C. Lima Analyst (London) in press. 53 S . G. Cutler and P. Meares J. Electroanal. Chem. 1977 85 145. 54 L. Ebdon A. T. Ellis and G. C. Corfield Analyst (London) 1979 104 730. 55 G. C. Corfield L. Ebdon and A. T. Ellis Anal. Proc. 1981 18 1 12. 56 L. Ebdon A. T. Ellis and G. C. Corfield Analyst (London) 1982 107 288. 57 L. Keil G. J. Moody and J. D. R. Thomas Analyst (London) 1977 102 274. 58 P. C. Hobby G. J. Moody and J. D. R. Thomas Analyst (London) 1983,108 581. 59 C. J. Olliff and G. R. Pickering British Patent 1558553 (C1 GOlN27/30) 03 January 1980. 60 A. Craggs G. J. Moody J. D. R. Thomas and B. J. Birch Analyst (London) 1980 105,426. 61 A. J. Frend G. J. Moody J. D. R. Thomas and B. J. Birch Analyst (London) 1983 108 1072. Paper 511886; Received 16th September 1985