ANALYST, FEBRUARY 1991, VOL. 116 135 Comparative Performance of 14-Crown-4 Derivatives as Lit hiu m-selective Electrodes Ritu Kataky, Patrick E. Nicholson and David Parker* Department of Chemistry, University of Durham, South Road, Durham DHI 3LE, UK Arthur K. Covington Department of Chemistry, University of Newcastle, Newcastle-upon-Tyne, UK A series of neutral ionophore-based lithium-selective liquid-membrane electrodes have been prepared and the electrode performance compared with similar electrodes based on the lithium ionophores ETH 1810-ortho-nitrophenyl octyl ether (oNPOE) and ETH 2137-bis(l-butylpentyl) adipate (BBPA). By using a diamide substituted 14-crown-4 macrocycle, selectivities for Li+ in the presence of Na+ of log q;ha = -3.25 and -2.92 were obtained for diisobutylamide-oNPOE and di-n-butylamide-oNPOE derivatives.The di-n-butylamide-oNPOE based electrode functioned satisfactorily in serum, exhibiting a fast response time (10-15 s), an acceptable lifetime of 50 d and minimal protein interference. Keywords: Lithium; ion-selective electrode; selectivity; crown ether; serum An ideal Li+ ionophore for use in monitoring the concentra- tion of Li+ ions in blood during manic depressive psychosis therapy has yet to be found. The search is imperative as a close monitoring of lithium concentration during treatment is required in order to secure a therapeutic effect and to avoid an overdose of lithium which could lead to fatal poisoning.' There is a narrow gap between therapeutic and toxic levels (Fig. 1). A series of chiral 14-crown-4 derivatives have been systematically synthesized and characterized (Table 1).Their performance, in serum, has been compared with the perfor- mance of commercial Li sensors and with results obtained by atomic absorption measurements. A diisobutylamide and a di-n-butylamide 14-crown-4 based sensor exhibit high selectiv- ity for Li+ over Na+ with log kc>tNa values of -3.25 and -2.92, respectively. These are superior to the best Li+ in Na+ selectivities reported previously (ETH 1810, log kEtN, = -2.45).'-11 The diisobutylamide derivative, however, behaves poorly in serum. The di-n-butylamide based elec- trode functions satisfactorily in serum, exhibits a fast response time (about 10-15 s) and has an acceptable lifetime. Table 1 14-Crown-4 derivatives studied R R 2.14C4Thio 4. 14Cd Dibenz rH2OCH2Q] 3. 14C4 Monobenz (CH20CH2Ph) 5. 14C4 01 7. 14c4 Est (CHzOH) (CH2C02Me) 9. 14C4 Butam (CH~CONBU~) ( CH ;OCH zP h ) 6. 14c4 Diol (CH20H) 8. 14c4 Diest (CH2C02Me) 10. 14C4 Dibutam 11. 14c4 Diibutam (CH~CONBU~) [CHZCON(~BU)~] * To whom correspondence should be addressed. Experimental Design of the Ionophores A 14-crown-4 skeleton has been the basis of previous Li+ ionophores, as it has an optimum cavity size for incorporating Li+ ions.4-8 The substituents in these earlier sensors were chosen for their effect in improving the lipophilicity of the sensor rather than their effect in enhancing Li selectivity in m I E 2 1.5 E E . - .- -I Y 1 .o 0.5 0 2 4 6 8 10 Consecutive controls Fig. 1 Monitoring Li levels during therapy.The first two controls were made once a week after patient A, 0, was started on 800 mg d-I of Li2C03. Side effects were detected and the patient was taken off Li therapy for 3 months. Therapy was restarted at 400 mg d-l (third control) and gradually increased to 600 mg d-l (fourth control). The fifth control showed a high Li level of 1.2 mmol dm--7 corresponding to a dosage of 1200 mg d- l . The following controls correspond to oncc every month while the dosage was maintained at 800 mg d-'. Patient B, 0, was started on 250 mg d-I of Li2C03 (first control) which was increased to 750 mg d-I (second control). The toxic level at control 3 corresponds to administration of a diuretic. The treatment was stopped and restarted after 3 months at 250 mg d-1 (fourth control), gradually increased to 500 mg d-l (controls 5 and 6) and then maintained at 800 mg d-I (controls 7-12). (Blood samples were taken 12 h after administering Li2C03 tablets.The instrument used was a Corning 405 flame photometer)136 ANALYST, FEBRUARY 1991, VOL. 116 binding. Improving the selectivity for complexing Li+ in the presence of excess of Na+ has been the central aim of our recent work. The Li+ ion may exhibit octahedral coordination. This suggested the possibility that by incorporating additional ‘axial’ donor sites on the 14-crown-4 ring, 1 : 1 complexation with Li+ would be favoured, while the formation of 2 : 1 complexes (Li+ : M+) would be suppressed. This tendency to form 2 : 1 complexes is particularly important for Na+ and K+, and is suppressed with the sterically demanding ‘axial’ substituents incorporated in the 14-crown-4 ring.The length of the side chains was chosen bearing in mind, not only purely geometric considerations for optimizing amide oxygen-Li+ interactions, but also the fact that a chelate-ring size of six intrinsically favours Li+ complexation whereas a five-mem- bered ring favours binding by the larger K+ and Na+ ions. Furthermore, Li+ is a small, ‘hard’, polarizing ion, hence ‘hard’ o-donors with large dipole moments, such as amide or phosphonate groups, should enhance Li+ in Na+ selectiv- ity. 10.11 Taking these factors into account, the 14-crown-4 derivatives shown in Table 1 were synthesized according to published procedures. 11 Cl Reagents and Chemicals Chloride salts of lithium, sodium, potassium and magnesium (BDH AnalaR) were dried at 50 “C and stored over silica gel.Calcium chloride solution (BDH AnalaR, 1 mol dm-3) was used. All standard solutions were prepared in de-ionized water and their cation concentrations checked by atomic absorption spectrometry. The 3-morpholinopropanesul- phonic acid (MOPS) and 4-(2-hydroxyethyl)piperazine-l- ethanesulphonic acid (HEPES) buffers12 and their sodium salts were obtained from Sigma. The materials for the electroactive membranes were: high relative molecular mass poly(viny1 chloride) (PVC) ; ortho- nitrophenyl octyl ether (oNPOE); bis( 1-butylpentyl) adipate (BBPA); potassium tetrakisb-chloropheny1)borate (K- TpCIPB); and lithium ionophore IV (ETH 2137), all obtained from Fluka; and a Philips Li ionophore (IS, 561).Tetrahydro- furan (THF) was of spectroscopic grade, distilled from sodium benzophenone ketyl. Serum was collected from healthy volunteers and was not stabilized by addition of heparin, but was used immediately. Apparatus Calibration and selectivity measurements A Philips IS (561) electrode body was used to mount the electroactive membranes. The filling solution was 1 x 10-3 mol dm-3 LiCI. The electrodes were fitted in a constant- volume cell made from a water-jacketed glass tube with a B19 ground-glass joint. The cell was incorporated in a flow system (Fig. 2). The reference electrode (porous plug, calomel, filled with 3.5 mol dm-3 KCI, RE1 Petiacourt) was placed, downstream, in a reference cell made from a water-jacketed glass tube from which a short capillary fitted with a ceramic plug had been drawn out.This formed a constrained diffusion junction with the sample. The temperature of the system was maintained at 37 “C by using a Techne Tempette Junior TE-85 thermostatically controlled bath. The solution was drawn at a constant rate (about 3 ml min-1) using an RS 330-812 peristaltic pump. The ion-selective and reference electrodes were connected to a digital multimeter (Keithley 197, Autor- anging Microvolt DMM) via a buffer amplifier. A flat-bed Linseis 17100 chart recorder, provided with back-off facilities, was used for monitoring potential differences. A suitable capacitance was connected across the input of the chart recorder to smooth out residual noise. I l l Initial Diluent Constant solution solution volume cell mstant volume cell Reference cell - Waste Fig.2 Constant volume cell and reference cell ( a ) Flow system in which the cell was incorporated. (b) Blood plasma studies The electroactive membranes were mounted on poly- carbonate stems, according to published procedures. 13 The cells were connected to a multiple switch (Fig. 3). The solution was injected into the system by use of disposable 1 ml syringes. The signal monitoring system was similar to that described above. Measurements were made at ambient temperature. Atomic absorption measurements were carried out on a Perkin-Elmer (5000) instrument. A Corning 654 (Na+, K+, Li+) analyser was used for comparative studies. The determinations of refractive indices of serum for plasma water were made on a Goldberg AO, TS meter and concentrimeter.Membrane preparation The oNPOE-based membranes were composed of 1.2% ionophore, 65.6% oNPOE, 32.8% PVC and 0.4% KTpClPB in 6 cm3 of THF. The ETH 2137 and a 14-C-4 di-n-butylamide sensor were also made into membranes consisting of 2.0% ionophore, 65.6% BBPA and 32.4% PVC in 6 cm3 of THF. The membranes were cast according to published proce- dures. 14 Procedure Calibration and selectivity measurements The ion-selective electrodes (ISEs) were calibrated using a constant dilution technique.15 A fixed interference method was used for selectivity coefficient measurements. In order toANALYST, FEBRUARY 1991, VOL. 116 300 280 260 240 220 200 $180 € ail60 140 120 100 80 137 - - - - - - - - - - - - Chart recorder Syringe Reference I I Waste Flow diagram showing the cells connected to the multiple Fig.3 switch establish the behaviour of the electrodes in the ionic concen- trations present in blood, a background interferent solution of 150 mmol dm-3 NaCl, 4.3 mmol dm-3 KCI and 1.26 mmol dm-3 LiCI was used and all of the measurements were made at 37°C. Blood plasmu studies Initial studies were carried out using three different dilutions of serum (serum + diluent: 1 + 1, 1 + 4 and 1 + 9) each containing a range of LiCl concentrations, 0.25,0.50,0.75 and 1.00 mmol dm-3. The diluent was 145.0 mmol dm-3 NaCl, 5.0 mmol dm-3 NaMOPS, 6.7 mmol dm-3 MOPS and 4.0 mmol dm-3 KCl (pH = 6.86). The Corning 654 (Na+, K+, Li+) analyser was used, according to the manufacturer’s instructions, for comparative studies.16 Further tests were carried out with the best ionophore and ETH 2137 (the ionophore used in the Corning analyser). An aqueous solution of 140.0 mmol dm-3 NaCl, 4.0 mmol dm-3 KCl, 1.2 mmol dm-3 CaCI2 (20 ml) and a serum sample (20 ml) were dosed with 0.1 mol dm-3 LiCl solution to give Li+ concentrations ranging from 0.5-5.0 mmol dm-3 (AQ1-7 and PLl-7, Tablc 4). The calibration solutions used were: Cal 1 [ 135.0 mmol dm-3 NaCl, 3.6 mmol dm-3 KCI, 1 .0 mmol dm-3 LiCI, 5.0 mmol dm-3 NaMOPS, and 6.7 mmol dm-3 MOPS, pH = 6.86, ionic strength ( I ) = 144.61 and Cal2 (135.0 mmol dm-3 NaCl, 3.6 mmol dm-3 KCI, 2.5 mmol dm-3 LiCI, 5.0 mmol dm-3 NaMOPS, and 6.7 mmol dm-3 MOPS, pH = 6.86,I = 146.1). The cells were flushed with deionized water and air and calibrated with Cal 1 and Cal 2 prior to each sample injection.Atomic absorption measurements were performed in parallel. Results and Discussion The results of the calibrations of the various 14-crown-4 based ionophores in comparison with those used by Philips and Corning both in aqueous LiCl and in LiCl in the presence of interferents (150.0 mmol dm-3 Na+, 4.3 mmol dm-3 K + , and Table 2 Characteristics of Li ionophores. Theoretical slope at 37 “C = 61.54 mV decade-’ LiCl in a solution of LiCl in (ISONaCI, 4.3 KCI, de-ionizcd 1.26 CaC12)/ water mmol dm-3 Philips 14-crown-4-thio I4-crown-4-monobenz 14-crown-4-dibenz 14-crown-4-01 14-crown-4-diol 14-crown-4-est 14-crown-4-diest 14-crown-4-nbutam 14-crown-4-dinbut am 14-crown-4-diibutam ETH 2137 Slope 62.0 60.0 53.1 60.0 37.3 32.0 54.5 62.0 56.0 60.0 50.0 62.0 * LD = Log (limit of detection).LD* -4.5 -5.1 -4.9 -4.6 -4.6 -3.2 -3.9 -5.0 -5.0 -4.4 Slope LD” Logky’ 47.0 -2.6 -1.89 44.0 -2.0 -1.14 45.0 -2.2 -1.35 62.0 -2.6 -1.77 - - - - 25.0 - 36.0 - 61.0 -3.8 -2.92 61.0 -4.1 -3.25 60.0 -2.7 -1.90 - 6o i Therapeutic range 1.2-0.7 mmol dm-3 0 1.0 2 0 3.0 4.0 5.0 6.0 -Log [Li] Fig. 4 Li ISE slopes in the presence of interferents: Na+, 150; K + . 4.3; and Gal+, 1.26 mmol dm-3. A, Ideal; B, diisobutylamide C. di-n-butylamide; D, ETH 2137; E, dibenzyl: and F, Philips 1.26 mmol dm-3 Ca’+) are given, (Table 2). These prelimi- nary tests revealed that the alcohol derivatives behaved very poorly. The di-substituted derivatives were better behaved than their mono-substituted analogues.The dibenzyl, di-n- butylamide and diisobutylamide derivatives performed well even in aqueous solutions containing serum levels of sodium, potassium and calcium, the last two showing a significant improvement over the ETH 2137 and Philips (IS 561) Li+ sensors (Fig. 4). The selectivity coefficients of the most promising iono- phores were determined in interferent concentrations of 0.1 mol dm-3. The only exception was proton interference, in which a 1 X 10-3 mol dm-3 HCl solution was used. The values obtained are shown in Table 3. The target selectivity coefficients were calculated using the Nikolsky-Eisenman equation based on a contribution of less than 1%, by the activity of the interferent ion, in comparison with the activity of the primary ion.” The values for the ‘best’ lithium ionophore, to date, (ETH 1810) are included for comparison.The diisobutylamide and the di-n-butylamide ionophores meet the target selectivities for K+, Ca’+ and Mg2+. As the pH of plasma is 7.3-7.4, the proton interference measured at pH 3 is not important.The selectivity observed with respect to Na+ is superior to ETH 1810 and ETH 2137.138 ANALYST, FEBRUARY 1991, VOL. 116 Preliminary tests with the two amide ionophores exhibiting the best selectivities in vitro, were performed with diluted serum to give an indication of their stability in biological media. The results were compared with those obtained using a Corning 654 (Na+, K+, Li+) analyser in both the concentra- Table 3 Selectivity coefficients of Li ionophores in 0.1 mol dm-3 interferent background Log k r t Na K Ca Mg H* Target values (-4.3 <-2.8 <-3.0 (-3.5 <-2.1 Philips -1.3 -2.15 -3.22 -3.7 -1.0 14-crown-4-dibenz -1.4 -2.3 -4.5 -5.8 -3.5 14-crown-4-dinbutam -3.0 -3.5 -4.2 -5.7 -0.9 14-crown-4-diibutam -2.9 -4.3 -4.3 -5.3 1.1 ETH 1810? -2.45 -2.6 -2.7 -4.0 -1.0 ETH2137 (Corning) -1.9 - - - - * Proton interference in 0.001 mol dm-3 HCl (pH = 3).? Values from the Fluka Selectophore catalogue. Table 4 Protein interference. Error ( Y o ) = { [C(actual) - C(expected)]/ [ c ( , , ~ ~ ~ ~ ~ ~ J } ~ 1 0 0 . kYt = 1.2 x 10-2 Corning, 1.2 x 10-3 di-n- butylamide, 6.3 X 10-4 diisobutylamide Error (YO) Plasma : water [Li]/ ratio mol dm-3 Corning* Corning? 1 : 10 0.22 0 -58.8 0.40 0 -6.7 0.60 -1.3 -6.0 0.80 -2.1 -5.0 1 : 5 0.20 0 -46.7 0.35 -0.3 -11.1 0.54 -2.8 -8.9 0.72 -4.8 -5.2 1 : 2 0.125 -100.0 -52.6 0.22 -17.4 -9.1 0.34 -18.1 -9.1 0.42 -18.9 -7.9 * Concentration mode.? Millivolt mode. Di-n- butyl- amide 0 0 0 -1.2 0 0 -3.0 -4.2 -44.0 -6.8 -7.8 -7.4 Diiso- amide butyl- -30.6 -24.0 +5.4 +12.9 -100.0 - 100.0 -65.0 -30.7 - 100.0 -91.3 -82.8 -52.2 tion and millivolt mode. Voltage readings were converted to concentrations using the equation where E,, cLi(u) and cNa(,) are the voltage reading, Li+ concentration and sodium concentration, respectively, in the unknown sample and E c a l 1 , ~ ( ~ ~ 1 1 ) and C N a ( C a l 1 ) are the 5.0 - ( a ) 4.5 - 4.0 - 3.5 - 3.0 - 2.5 - 2.0 - 4.0 3.5 2.5 3.0 i *.O 1 & 0 1.5 1.0 1 !i 0.51 1 , , , , , 0 1 2 3 4 5 6 7 Sample no. Fig. 5 Comparison of atomic absorption spectrometry, ETH 2137- BBPA and di-n-butylamide-oNPOE ISEs in (a) a ueous solution and (b) plasma.0, ETH 2137; ., di-n-butylamide; $. flame; A , ETH 2137; and 0, di-n-butylamide. 0 and H, Using concentrations; and A , and 0, using activities Table 5 Comparisonsf flame photometry, ETH 2137-BBPA ISE and di-n-butylamide-oNPOE ISE, using concentrations directly. Concentration units, mmol dm-3; cun, uncorrected concentrations; cCrr concentrations corrected for plasma water; refractive index for plasma sample = 1.354; plasma water = 91.5%; “a] in plasma determined by atomic absorption spectrometry = 148 mmol dm-3; error (%) = (cISE - cFlame)/ CFlame ETH 2137* Di-n-butylamide” Flame? Solution C Error (YO) C Error(%) c AQ1 0.40 +1.1 0.40 +1.1 0.39 AQ2 0.86 +6.2 0.78 -3.7 0.81 AQ3 1.24 -13.9 1.39 -3.5 1.44 AQ4 1.58 -9.2 1.70 -2.3 1.74 AQ5 2.19 -14.0 2.41 -5.5 2.55 AQ6 3.28 -15.0 3.67 -4.9 3.86 AQ7 3.79 -14.4 4.22 -4.7 4.43 Cun PL1 0.57 PL2 0.97 PL3 1.43 PL4 1.69 PL5 2.40 PL6 3.34 PL7 4.29 CC, 0.62 1.06 1.56 1.85 2.62 3.65 4.69 Error (% ) + 19.2 +0.9 +3.3 +0.5 -7.1 -6.6 -6.6 Cun 0.46 0.99 1.37 1.65 2.43 3.39 4.37 CC, 0.50 1.08 1.50 1.81 2.66 3.71 4.78 Error (YO) -3.8 0.52 +2.8 1.05 -0.7 1.51 -1.6 1.84 -5.6 2.82 -5.1 3.91 -4.8 5.02 * Error on ISE values, about 1% (equivalent to k0.3 mV).? Error on flame values, about 3%.ANALYST, FEBRUARY 1991, VOL. 116 139 corresponding values in calibration solution, Cal 1. kC.tNa is the appropriate selectivity coefficient and S is the slope of the electrode at ambient temperature.Concentrations were cor- rected18 for 93% plasma water. Expected values were based on atomic absorption measurements of Li+ in the diluents (Table 4). The diisobutylamide based derivative behaved very poorly, whereas the performance of the di-n-butylamide based electrode is better than that of the Corning ETH 2137 electrode. Errors are larger in solutions with higher serum and lower lithium concentrations. The approximate 18% error (Table 4) observed in the Corning concentration mode, which is the commonly used mode, is disconcerting. To substantiate these results, further tests were performed on the di-n-butylamide-oNPOE based electrode and an ETH 2137-BBPA based electrode made from the parent iono- phore. (Calibrations with electrodes based on the opposite 0.005 0.004 0.003 0.002 r t 0.001 0 28 32 36 40 44 48 52 Days Lifetime of a 14-crown-4 di-n-butylamide ISE used in serum Fig.6 for 4 weeks combination, viz., di-n-butylamide-BBPA and ETH 2137- oNPOE in aqueous solutions containing Na+, K+, Ca2+ and Li+ at levels normally present in blood, showed that these combinations were unsuitable for use as clinical sensors.) Aqueous samples AQ1-AQ7 and plasma samples PL1-PL7 were injected into the flow cells and readings were taken as before. Atomic absorption measurements, on each sample, were performed in parallel. The results, shown in Table 5 , were corrected for the presence of protein using refractive index measurements. The di-n-butylamide electrodes, again, appeared to be superior to the ETH 2137 electrode. The errors are higher for plasma samples containing >2.0 mmol dm-3 Li+.Activity coefficients for Cal 1, Cal 2 and solutions AQ1- AQ7 were calculated using the Pitzer equation19 (Table 6). Concentrations were calculated using E, - E c a l 1 = S log{[cLi(u) yLi(u) + kc:Na C N ~ ( U ) ~ ~ a ( u ) l / [cLi(Call) YLi(Cal1) -k kE,t~a CNa(Ca1 1 ) YNa(Cal I)]> (2) where cy = a , 'a' being the activity of the appropriate ion and y the corresponding activity coefficient. The other symbols have the same significance as in equation (1). The results given in Table 7 show marked improvement when activity coefficient corrections are made for the solutions of higher lithium concentrations (AQ3-AQ7 and PL4-PL7). The behaviour of the di-n-butylamide electrode is significantly improved by using activities [Fig.5(a) and ( b ) ] . The stability of a di-n-butylamide-oNPOE based ionophore used for 4 weeks in serum was monitored for a further 3 weeks with Cal 1 and Cal 2. The results are shown in Fig. 6. The selectivity coefficient was 1.26 x 10-3 for the fresh membrane. After 4 weeks it decreased to 2.0 x 10-3 and after a further 2.5 weeks to 3.96 X 10-3. The fresh ETH 2237-BBPA membrane has a selectivity coefficient of 1.26 x 10-2, much higher than that of the di-n-butylamide sensor. Table 6 Activity coefficients based on the Pitzer equation. Solutions AQI-AQ7 contain 0.140 rnol dm-3 N a f . 0.004 rnol dm-3 K+ and 0.0012 rnol dm-3 Ca7+; Cal 1 and Cal2 contain 0.140 rnol dm-3 Na+ and 0.004 rnol dm-3 K+ [Li]/ Solution rnol dm-3 I YLi YN a AQ 1 AQ2 AQ3 AQ4 A 0 5 AQ6 AQ7 Cal I Cal2 0.0005 0.001 0.0015 0.002 0.003 0.004 0.005 0.001 0.0025 0.1457 0.1462 0.1467 0.1472 0.1482 0.1492 0.1502 0.1 440 0.1455 0.7758 0.7757 0.7756 0.7754 0.7752 0.7749 0.7747 0.7765 0.7761 0.7542 0.7540 0.7537 0.7535 0.7533 0.7529 0.7526 0.775 1 0.7546 Conclusion The best clinically relevant Li ionophore reported, to date, is ETH 1810-oNPOE, with a log ky2y3tNa = -2.45; well below the target value of -4.3 required for less than 1% interference. This ionophore is reported to have a slow response and limited lifetime when used in serum.20.21 The preferred ionophore, ETH 2137-BBPA has log kEtNa = -1.9.The target ionophore with kE:Na = 5.0 X 10-5 has a potential difference ( A E ) of 24.5 mV at 37°C when trans- ferred from Cal 1 to Cal 2.The ETH 2137-BBPA electrode has a AE of 11.9 mV (kstNa = 1.2 x 10-2) corresponding to a correction factor of 12.6' mV whereas the di-n-butylamide- Table 7 Comparison of flame photometry. ETH 2137-BBPA ISE and di-n-butylamide-oNPOE ISE, making activity coefficient corrections ETH 21 37 Di-n-butylamide Flame Solution AQ I AQ2 AQ3 AQ4 AQ5 AQ6 AQ7 c u n PL 1 0.63 PL2 1.04 PL3 1.52 PL4 I .80 PL5 2.54 PL6 3.52 PL7 4.51 c 0.45 0.92 1.33 1.69 2.30 3.45 3.97 Error (YO ) +15.4 + 13.6 -7.6 -2.9 -9.8 - 10.6 - 10.4 c,, Error (YO) 0.69 +32.7 1.14 +8.6 1.66 +9.9 1.97 +7.1 2.77 -1.8 3.85 - 1.5 4.93 -1.8 C 0.38 0.78 1.42 1.75 2.50 3.83 4.41 C" n 0.44 1 .oo 1.40 1.70 2.52 3.53 4.56 c,, 0.48 1.09 1.53 1.85 2.75 3.86 4.99 Error(%) c -2.6 0.39 -3.7 0.81 -1.4 1.44 +0.6 1.74 -2.0 2.55 -0.8 3.86 -0.5 4.43 Error (YO) -7.7 0.52 +3.8 1 .05 -1.3 1.51 +0.5 1.84 -2.5 2.82 -1.3 3.91 -0.6 5.02140 ANALYST, FEBRUARY 1991, VOL.116 oNPOE electrode has a AE of 18.5 mV (kg:Na = 2.4 X 10-3) requiring a correction factor of 6.0 mV. The sensor is stable in serum, has a short response time (about 10-15 s) and a lifetime of at least 50 d. These factors demonstrate that this ionophore is reliable for the assay of Li. We thank SERC for support, Dr. C. T. G. Flear (Royal Victoria Infirmary, Newcastle-upon-Tyne) and Dr. C. J. Fischer (County Hospital, Durham) for useful discussions and Robert Coult for carrying out the atomic absorption measure- ments. References Amdisen, A., in Handbook of Lithium Therapy, ed. Johnson, F. N., MTP Press, Lancaster, 1986, ch.2. Metzger, E . , Dohner, R., and Simon, W., Anal. Chem., 1987, 59. 1600. Oggenfuss, P., Mod, W. E., Oesch, U . , Ammann, D., Pretsch, E., and Simon, W., Anal. Chim. Acta, 1986, 180, 299. Kimura, K., Kitazawa, S . , and Shono, T., Chem. Lett., 1984, 639. Metzger, E., Aeschmann, R., Egli, M., Suter, G., Dohner, R., Ammann, D., Dobber, M., and Simon, W., Helv. Chim. Acta, 1986, 69, 1821. Attiyat, A. S . , Christian, G. D., Xie, R. Y . , Wen, X . , and Bartsch, R. A., Anal. Chem., 1988, 60, 2561. Gadzekpo, V. P. Y., Moody, G. J . , and Thomas, J . D. R., Analyst, 1986, 111, 567. Kimura, K., Yano, H., Kitazawa, S . , and Shono, T., J. Chem. SOC., Perkin Trans. 2, 1986, 1945. Metzger, E., Ammann, D., Schefer, U., Pretsch, E . , Simon, W., Chimia, 1984, 38, 440. 10 11 12 13 14 15 16 17 18 19 20 21 Kataky, R., Nicholson, P. E., and Parker, D., Tetrahedron Lett., 1989, 30,4559. Kataky, R., Nicholson, P. E., and Parkcr, D., J . Chem. SOC., Perkin Trans. 2, 1990, 321. Good, N. E., Winget, G. D., Winter, W., Connolly, T. N., Izawa. A.. and Singh, R. M. M., Biochemistry, 1966. 5 , 467. Covington, A. K., Kelly, P. M., and Maas, A. H. J . , in Methodology and Clinical Applications of ion-Selective Elec- trodes. eds. Maas, A. H. J . , Buckley, B. M., Manzoni, A . , Moran, R. F., Siggaard Andersen, O., and Sprokholt, R., Davies Printing Co., Rochester, 1989, vol. 10, p. 4. Craggs, A., Moody, G. J . , and Thomas, J . D. R., J. Chem. Educ., 1974,51, 541, Horvai, G., Toth, K., and Pungor, E., Anal. Chim. Acta, 1976, 82, 45. Corning 654 (Naf, K f , Lif) Analyser Reference Manual, Corning, 1988. Ammann, D., Anker, P., Metzger, E., Oesch, U., and Simon, W., in Ion Measurements in Physiology and Medicine, eds. Kessler, M., Harrison, D. K., and Hoper, J . , Springer Verlag, Berlin, 1985, p. 102. Czaban, J . D., and Legg, K. D., in Proceedings of the Workshop on Direct Potentiometric Measurements in Blood, ed. Koch, W. F., Gaithersburg, MD, 1983, p. 63. Covington, A. K., and Ferra, M. I. A., Scand. J. Clin. Lab. Invest., 1989, 49, 667. Metzger, E., Dohner, R., Simon. W., Voncherschmitt, D. J., and Gautschi, K., Anal. Chem., 1987, 59, 1600. Gadzekpo, V. P. Y., Hungerford, J . M., Kadry, A. M., Ibrahim, Y. A., Xie, R. Y., and Christian. G. D., Anal. Chem., 1986. 58, 1948. Paper 0l03204B Received July I7th, 1990 Accepted August 20th, 1990