Determination of complex formation constants of neutral cation

complex cases, when charges of the ions and the stoichiometries of the complexes are different. The stronger the interaction with the primary ion is, ...
1 downloads 0 Views 665KB Size
Anal. Chem. 1994,66, 516-521

Determination of Complex Formation Constants of Neutral Cation-Selective Ionophores in Solvent Polymeric Membranes Eric Bakker,’it Michael Wilier, Markus Lerchl, Kurt Seller, and Ern8 Pretsch’ Department of Organic Chemistry, Swiss Federal Institute of Technology (ETH), Universitatstrsse 16, CH-8092 Zurich, Switzerland

Formal complexformationconstants between lipophilic ligands and cations have been determined within the solvent polymeric membrane phase. The method is based on spectrophotometric measurements on a 1-2-bm thin membrane phase (optode) containing a H+-selective chromoionophore. Ion exchange equilibria between the membrane and the aqueous phase are traced for membranes with and without ionophore. A theoretical discussion is given as well as experimental results with the ionophoresvalinomycin and BME-44 (for K+),ETH 4120 (for Na+), ETH 1001 and ETH 129 (for Ca2+), and ETH 7025 (for Mg2+). Complexes with various stoichiometriesare formed for the Mg2+and Ca2+complexes of ETH 7025, which make the realization of a selective magnesium optode with this ionophore not possible. With this novel method, the key parameters determining the performance of ionophore-based ion-selectiveoptodes and electrodesbecome directly accessible. Solvent polymeric membranes are used in a great variety of chemical sensors with potentiometric’ and optical2 transduction. In ion-selective electrodes (ISEs) and optodes, an incorporated ionophore favors the extraction of the primary ion by selective complexation. For ions of the same charge, forming complexes of equal stoichiometry with the ionophore, the selectivity is proportional to the ratio of the complex formation constants of the primary to interfering ion.3 The complex formation constants also govern the ion selectivities in more complex cases, when charges of the ions and the stoichiometries of the complexes are different. The stronger the interaction with the primary ion is, the more selective the electrode will be for it, if other parameters are unchanged. There is, however, an upper limit of allowed complex formation constants. If the complexes formed are too stable (or if lipophilic anions are present in the sample solution), so that a coextraction of sample cations and anions occurs, the electrode response is no longer a function of the primary ion a ~ t i v i t y . ~Thus, , ~ knowledge of the stability constants in the membrane phase is of great importance for understanding the phase-transfer equilibria involved and for planning and modeling novel ionophores. + Present address: Department of Chemistry, The University of Michigan, Ann Arbor, MI 48109-1055. (1) Umezawa, Y. Handbookoflon-Selectiue Electrodes: Selectivity Coefficients; CRC Press: Boca Raton, Ann Arbor, Boston, 1990. (2) Arnold, M. A. Anal. Chem. 1992, 64, IO15A. (3) Morf, W. E. The Principles of Ion-Selectiue Electrodes and of Membrane Transport; Akaddmiai Kiad6: Budapest, 1981. (4) Morf, W. E.; Ammann, D.; Simon, W. Chimia 1974, 28, 65. (5) Buchi, R.; Pretsch, E.; Morf, W. E.; Simon, W. Helv. Chim. Acta 1976, 59, 2407.

516

Analytical Chemistry, Vol. 66, No. 4, February 15, 1994

This fact has been realized since the early days of the development of ISES.~However, the calibration curve and selectivity coefficients of such electrodes give no reliable information about stability constants, since selectivities only reflect their ratios. Under the assumption that a very lipophilic and bulky interfering ion is not complexed by the ionophore, some data about the minimum possible complex formation constant and the stoichiometry were, however, ~ b t a i n e d . ~ Approaches for measuring complex formation constants in aqueous, methanol, or ethanol solution have been reportedus The constants obtained were usually small, increasing with decreasing polarity of the solution, and in addition, the complex stoichiometries showed a dependence on the solvent used. Therefore, a correlation between ion selectivities of sensors and measured complex formation constants in bulk solutions could only be observed in special cases such as valin~mycin.~ With coextraction equilibrium measurements of picrate salts from the aqueous phase into an apolar organic phase with and without ionophore, much larger complex formation constants have been found.1° These results might not be directly comparable to the ones obtained in bulk solutions, because the anionic dye may form strong ion pairs and shift the equilibrium considerably.11J2 None of the methods described so far provides reliable data about the formation contants of ionophores in the membrane phase. In recent years, a novel type of sensors with optical transduction based on solvent polymeric membranes (bulk optodes) has been introduced,13J4which is well understood t h e o r e t i ~ a l l y . ~The ~ J ~degree of protonation of an incorporated H+-selective ionophore, changing its spectral properties upon protonation (chromoionophore), is measured spectrophotometrically. The thin membrane, cast onto a glass plate, contains in addition an ionophore and a lipophilic anionic additive. The optode signal depends on the cation exchange (6) Kirsch, N. N. L.; Simon, W. Helv. Cfiim.Acta 1976, 59, 357. (7) Bliggensdorfer, R. Dissertation, ETH Ziirich, 1990; No. 9190. (8) Bliggensdorfer, R.; Suter, G.;Simon, W. Helu. Cfiim. Acta 1989, 72, 1164. (9) Morf, W. E.; Simon, W. Ion-Selective Electrodes Based on Neutral Carriers. In Ion-Selective Electrodes in Analytical Chemistry; Freiser, M., Ed.; Plenum Press: New York, London, 1978; p 211. (10) Rakhman’ko, E. M.; Yegorov, V. V.; Gulevich, A. L.; Lushchik, Y.F. Sel. Electrode Reu. 1991, 13, 5 . (1 1) Kirsch, N. N. L.; Funck, R. J. J.; Simon, W. Helu. Cfiim.Acta 1978,61,2019. (12) Oggenfuss, P. Dissertation, ETH ZBrich, 1984; No. 7619. (13) Gantzer, M. L.; Hemmes, P. R.; Wong, D. Eur. Pat. Appl. EP 153.641 (CI. G01N33/00), September 4, 1985. (14) Morf, W.E.;Seiler,K.;Lehmann,B.;Behringer,Ch.;Tan,S.S.S.;Hartmann, K.; Sorensen, P. R.; Simon, W. In Ion-Selective Electrodes; hngor, E., Ed.; Akademiai Kiado: Budapest, 1989; Vol. 5, p 115. (15) Seiler, K.; Simon W. Anal. Cfiim.Acta 1992, 266, 73. (16) Bakker, E.; Simon, W. Anal. Chem. 1992, 64, 1805.

0003-2700/94/0366-05 16$04.50/0 0 1994 American Chemical Society

equilibrium of the primary ion and H+ between the aqueous and membrane phase. Ion activities of the primary ion are accessible, if the sample pH is measured or kept constant by buffering. In this work, we present a novel approach for determining formation constants of ionophore complexes directly within the membrane phase. They can be evaluated from measurements on optode membranes with and without ionophore. Theoretical considerations, model assumptions, and results with ionophores, selective for different cations and forming complexes of various stoichiometries, are presented. EXPERIMENTAL SECTION Reagents. All aqueous solutions were prepared with doubly quartz-distilled water. Salts and acids of highest purity available were used throughout. For membrane preparation, poly(viny1 chloride) (PVC, high molecular weight), potassium tetrakis[3,5-bis(trifluoromethy1)phenyll borate (K(TFPB)), bis(2-ethylhexyl) sebacate (DOS), o-nitrophenyl octyl ether (0-NPOE), bis( 1butylpentyl) adipate (BBPA), and tetrahydrofuran (THF, freshly destilled prior to use) were obtained from Fluka AG (Buchs, Switzerland). The ionophores used were N-heptylN’,N’-bis[8-[ [ 3-(heptylmethy1amino)-1,3-dioxopropyl]amino]octyll-N-methylpropanediamide(ETH 7025, Fluka), valinomycin (Moller AG, Zurich, Switzerland), 2,2’-(4-octadecanoyloxymethyl- 1,2-phenylenedioxy)-N,N,N’,N’tetracyclohexyldiacetamide (ETH 4 120, synthesis described in ref 17). The synthesis of the chromoionophores have been described previously: 4-(octadecy1amino)azobenzene (ETH 5315),18 4-[ [9-dimethylamino)-5H-benzo[a]phenoxazin-5ylidene]amino]benzoic acid 1 1-[( l-butylpentyl)oxy]-l1oxoundecyl ester (ETH 5418),18 N,N-diethyl-5-(octadecanoylimino)-5H-benzo[a]phenoxazin-9-amine(ETH 5294),19 and N,N-diethyl-5-[(2-octyleicosanoyl)imino]-5Hbenzo[a]phenoxazin-9-amine(ETH 2458).20 Apparatus. pH values were determined with a glass electrode (Orion Ross Model 81-02) and pH meter (Orion Model 920A, Orion Research AG, Uetikon am See, Switzerland). Spectrophotometric studies on optode membranes were performed with a specially designed measuring cell19 mounted in a conventional double-beam spectrophotometer (constant bandwidth, 2 nm; Uvikon 8 10, Kontron AG, Zurich, Switzerland). Absorption spectra were recorded between 800 and 400 nm at 25 f 1 OC. Optode Membranes. Optode membranes were prepared and measured according to ref 16. Concentrations in the organic phase are given as moles per kilogram. The K+selective membrane consisted of 1.9 mg of K(TFPB), 3.3 mg of the ionophore valinomycin, and 1.2 mg of the chromoionophore ETH 5294 in a DOS/PVC (2:l) membrane with a total weight of 250 mg. The corresponding optode without ionophore was prepared with an equal amount of components, but without valinomycin. The membrane without ionophore for determining the exchange constant for Na+ contained 4.13 (17) Gehrig, P.; Rusterholz, B.; Simon, W . Anal. Chim. Acta 1990, 233, 295. (18) Lerchi, M.; Bakker, E.;Rusterholz, B.; Simon, W. Anal. Chem. 1992, 64, 1534. (19) Morf,W. E.;Seiler,K.;Rusterholz, B.;Simon, W.Ana1. Chem. 1990,62,738. (20) Bakker, E.;Lerchi, M.; Rosatzin, T.; Rusterholz, B.; Simon, W. Anal. Chim. Acta 1993, 278, 21 1.

mg of K(TFPB) and 3.03 mg of ETH 2458 in 79.9 mg of PVC and 160.5 mg of BBPA. The Mg2+-selectivemembranes 1-111 consisted of 4.80, 10.89, and 19.59 mg of ionophore ETH 7025; 6.00,6.01, and 6.05 mg of chronoionophore ETH 5418; 7.51,7.53, and 7.65 mg of K(TFPB), in o-NPOE/PVC (2:l) with a total weight of 257,250, and 25 1 mg, respectively. The optodes without ionophore contained 3.77 and 3.89 mg of chromoionophore ETH 5315 and 7.57 and 10.46 mg of K(TFPB) ino-NPOE/PVC (2:l) and DOS/PVC (2:1), with a total weight of 250 mg. Buffers. The K+-selectiveoptode membrane was measured with 104-10-2 M KCl in 0.01 M NaOAc/HOAc buffers at pH 5.0, prepared according to ref 21. The optode membrane without valinomycin was equilibrated with different KOH solutions as described in ref 18. The membrane without ionophore for determining the exchange constant for Na+ was measured accordingly with different NaOH solutions. The Mg2+-selective optodes were equilibrated with 104-10-2 M CaCl2 and MgClz solutions in 0.1 M 3-morpholinopropanesulfonic acid (MOPS), adjusted to pH 7.19 and 6.57 with 0.01 M NaOH, giving a total Na+ background of 0.005 and 0.002 M, respectively. The Mg2+/Ca2+optodes with ETH 53 15, without ionophore, were equilibrated with 0.1 M MgC12 and 0.1 M CaC12 solutions. Activity Coefficients. Activity coefficients were calculated according to the Debye-Hiickel theory.22 The pH of all buffered solutions were measured with a pH glass electrode. The pH values of KOH and NaOH solutions were calculated. THEORETICAL CONSIDERATIONS The model is based on the following assumptions: , (1) The solvent polymeric membrane phase behaves like a homogeneous organic phase and is in equilibrium with the contacting aqueous solution. The solid support of the organic phase, usually glass, is inert and does not participate in the equilibrium. (2) The incorporated chromoionophore is ideally H+selective,20and Lambert-Beer’s law holds for all absorbance measurements. (3) Anionic additives are confined to the membrane phase, and ion pairs in the organic phase are neglected or association constants of any complexed and uncomplexed cations with the anionic additive are equal. (4) Ionophores form stable complexes of not more than two different stoichiometries at a time. (5) The equilibrium constants determined are related to concentrations of the species in the organic phase and to activities in the aqueous phase. Assumptions 1-3 have been shown to be justified for a large series of ion-selective optodes described in the past, (4) is not absolutely necessary, but simplifies the evaluation, and (5) is an assumption that might be justified for optode membranes containing an ionic additive.16 However, it has been shown that certain types of PVC membranes can take up a substantial amount of water,23which is likely to influence (21) Perrin, D. D.; Dempsey, B. Bu//ers/orpHand Metal Ion Control; Chapman and Hall: London, New York, 1983; p 134. (22) Meier, P. C. Anal. Chim. Acta 1982, 236, 363. (23) Chan, A. D. C.; Harrison, D. J. Anal. Chem. 1993, 65, 32.

Analytical Chemistry, Vol. 66, No. 4, February 15, 1994

517

the activity of the dissolved species in the organic phase.24 Since the formal complex formation constants determined by the proposed method only refer to concentrations, they may not be equal to the respective thermodynamic constants. Deviations due to different concentrations of components in the membrane, osmolality of the aqueous phase, and nature of the plasticizer used have therefore to be encountered. However, more experiments will be necessary to fully understand the influence of these parameters on the phasetransfer equilibrium. Determination of Complex Stability Constants in the Membrane Phase. A neutral I”+-complexingionophore L and a H+-selective chromoionophore C, incorporated in a bulk optode membrane, are assumed to form the complexes IL,’+ (with the stoichiometry p ) and CH+. For electroneutrality reasons, lipophilic anionic additives such as tetraphenylborate derivativesmust also be present. The following phase-transfer equilibrium takes place with the contacting aqueous solution (“aq” denotes species in the aqueous phase, “org” in the organic phase): I”(aq)

+ pL(org) + vCH+(org) + IL,”+(org) + vC(org) + vH+(aq)

(1)

with the exchange constant

complex have to be considered, the electroneutrality condition and the mass balance are extended and, together with the ratio of the two formation contants, inserted in eq 2. The equilibrium activity of the primary ion I”+for a given a in the presence of complexes of 1:l together with 1:2 stoichiometry is then given as aH+(y a,+ = (Kexc;L,)-I(

G)

[LIZ

(5)

where v[L]’

+ a[L] + b = 0

(6)

with

and for the presence of complexes with 1:2 together with 1:3 stoichiometry = (KexchlL3)-’

with

1

(aH+a)”(2/V){(1- .)CT-

RT] + 4 - IL1

[LI~

(9)

where

kI, = exp(bI0(aq) - qo(org))/RT)

v[L]’

kH+ = eXp(b‘HO(aq)- pHo(org)]/RT) Brackets designate concentrations in the organic phase, and activity symbols stand for the respective activities in the aqueous phase. The stability constant of the ionophore complex, 814,and the acidity constant of thechromoionophore, Ka, refer to the organic phase. The symbols and represent the chemical standard potentials of the ions in the respective solvent, and R and T have their usual significance. Introducing the electroneutrality condition, the mass balance for the ionophore L and the chromoionophore C in the membrane phase, and the degree of protonation (1 - a) as the ratio of protonated to total chromoionophore concentration in the membrane (1 - a = [CH+]/CT),eq 2 is rewritten as16

+ a[L] + b = 0

(10)

with

PIL,

b = -(2(R,

- (1 - a)CT)- v&)

PIL,

(12)

The cation exchange constant for a membrane without cationselective ionophore, but of otherwise equal composition, is described as a [C] ” [I”+] KcxchI= L, [CH’] -= K,)”k,, k ~ +

(

(

(1 3)

Considering eq 4 and introducing the mass balances for the two membrane components C and R- leads to where RT, CT, and LT represent the total concentrations of anionic additive, chromoionophore, and ionophore, respectively, in the organic phase. For absorption measurements, a is a function of the absorbance values measured at equilibrium ( A ) , at tokl pwbas~tiw(A?), and complete deprotonation (AD) of the chromoionophore: cy=-

Ap-A AD

KexchlLP/Kcxchl

(4)

If two different stoichiometries of the ionophore/cation (24) Janata, J. Anal. Chem. 1992, 64, 921A.

518

The ratio of the two exchange constants for membranes with and without cation-selective ionophore is given by combining eqs 2 and 13:

Analytical Chemistry, Vol. 66, No. 4, February 15, 1994

= PIL,

(15)

As stated above, the result is based on the assumption that ion pairs between oppositely charged species are neglected. If strong ion pair formation must be assumed, the respective ion and KIR.,have to be considered: pair formation constants, KIL&”

In this work however, the ion pair formation constants are assumed to be equal for the complexed and uncomplexed ion I"+, and eq 15 is used throughout. for membranes without The exchange constants ionophore might be too small to be determined directly. In that case, exchange constants measured with chromoionophores C of lower basicity can be used and normalized by ApKa as described elsewhere.'*

"'1

0.2 -

VALINOMYCIN ETH 5294 KTFPB DOS PVC

DOS

pvc

0.0 -

RESULTS AND DISCUSSION Cation-selective optodes based on various neutral ionophores, H+ chromoionophores, and anionic additives have been described in the past. The measurements were in fact thermodynamic determinations of exchange constants KexchILP of H+ and the primary cation I"+ between the aqueous and organic phase according to eq 3. Such data have been given for ions including Ca2+,19K+,25Na+,26NH4+,27and Pb2+.18 For the evaluation of these constants, a given stoichiometric factor was assumed for all complexes. In addition to KexchILJ', the ion exchangeconstant for a similar membrane, but without ionophore, must be known for the evaluation of the complex formation constant of the ionophore/cation complex 014. This is demonstrated in the present work with the neutral ionophore valinomycin in combination with the H+-selective chromoionophore ETH 5294 and K(TFPB) as anionic additive, dissolved in a DOS/PVC (2:l) membrane. The optode membrane has been equilibrated with sodium acetate buffer (pH 5.0) and different concentrations of KC1. Assuming a stoichiometric factor p = 1, the theoretical curve was fitted to the experimental points according to eq 3, giving the exchange constant log KexchKL= -2.3. A similar optode membrane, but without ionophore, was equilibrated with KOH solutions to measure the exchange constant log KexchK according to eq 13, which was found to be -1 1.6. Figure 1 shows the degree of protonation (1 - CY)of the chromoionophore in the organic phase as a function of the ratio OK+/ ( I H t in the sample solution for both experiments. It can be clearly seen that the cation exchange equilibrium of such bulk optodes is strongly shifted by the incorporation of the ionophore. The horizontal distance of the two curves at a given 1 - CY value (ratio of eqs 2 and 13) represents the complexing capability of the ionophore valinomycin, log UK+/ UHt(K) -log aK+/UH+(KL)= OKL[L]. At half protonation (CY = OS), log (PKL[L]) has been determined as 7.2 (see Figure 1). This correspondes to a complex formation constant of log PKL= 9.3 (cf. Table 1). The constants log PKLdetermined in bulk solutions have been smaller (aqueous solution 0.37,28 0.09;29methanol solution 4.90,304.48;31ethanol solution 6.30,32 ~

( 2 5 ) Wang, K.; Seiler, K.; Morf, W. E.; Spichiger, U. E.; Simon, W.; Lindner, E.; Pungor, E. Anal. Sei. 1990, 6, 715. (26) Seiler, K. Dissertation, ETH Ztrich, 1990; No. 9221. (27) Sciler, K.; Morf, W. E.; Rusterholz, B.; Simon, W. Anal. Sci. 1989, 5, 557. (28) Feinstcin, M. B.; Felscnfeld, H. Proc. Natl. Acad. Sci. U.S.A. 1971,68,2037. L29) Eyal, E.; Rechnitz, G.A. Anal. Chem. 1971, 43, 1090. (30) Grell, E.; Funck, Th.; Eggers, F. Dynamic properties of membrane activity of ion specific antibiotics. In Mechanisms of Antibiotic Action on Protein Biosynthesis and Membranes; Mufioz, E., Garcia-Ferrindiz, F., Vazquez, D., Eds.; Elsevier Scientific Publishing Co.: Amsterdam, London, New York, 1972; p 646. (31) Funck, Th.; Eggers, F.; Grell, E. Chimia 1972, 26, 637.

log (aK+/aH+)

Figure 1. Cation exchange equilibria measured with optodes containing the H+ chromoionophore ETH 5294 and the anionic additive K(TFPB) with (left Curve) and wRhout (right curve) the K+-selective ionophore valinomycin. Theoretical curves were calculated accordlng to eqs 3 and 14, respectively. The horizontaldistanceof !he two curvesreflects the complexation capability of the Ionophore (see text). Table 1. Stability Constants for Various Catlon-Selectlve Ionophom, Calculated from Exchange Cofutants of Optodes Determined Previously

chromo- plasti-

I"+" ionophore L p b ionophore cizer K+ valinomycin 1 ETH5294 DOS Na+ valinomycin 1 ETH5294 DOS K* BME-44 1 ETH5294 DOS Na+ BME44 1 ETH5294 DOS Nat ETH4120 2 ETH2439 BBPA Na+ ETH4120 2 ETH5350 BBPA Ca2+ ETH129 2 ETH5350 DOS Ca2+ ETHlOOl 2 ETH2439 DOS Ca2+ ETHlOOl 2 ETH5294 DOS

lo Kexch'P'

-2.fiZbe -6.P6 -3.9% -6.9% -3.026 -6.02s

-5.6n -3.1% -7.119

log

log

Kexd*

BL,~

-1l.g

9.3 6.4 7.9 5.5 7.4 7.6 23.8 19.9 19.5

-12.48

-11.g -12.48 -10.48 -13.68 -29.9 -23.oh -26.6h

*

Measuring ion; bold, primary ion. Stoichiometry of the ionionophwe complex. Exchange constant according to e 2; superscript, reference cited. According to eq 15. e Correspon%n to the value found in this work. f See ref 20.8 Measured with E T h 2458 and corrected to the respective chromoionophore (column 4) with basicity difference given in ref 20. Measured with ETH 5315 and corrected to the respective chromoionophore with basicity difference given in ref 20.

6.0833). The increasing complex stability constants with decreasing polarity are well understood with the different free energies of solvation of the potassium ion in the respective solvents. Larger stability constants have only been reported from coextraction experiments of picrate salts into a binary solution of chloroform and nitrobenzene (4:l) (log PKL= 12.4). In Table 1, exchange constants for different ionophores reported in the literature are listed together with their stability constants, calculated according to eq 15. The ionophores are usually complexing the interfering ions considerably, though more weakly than the primary ion. The K+-selective ionophores valinomycin and BME-44 both show similar selectivities in ISE membranes,34v35but the complex stabilities (32)Shcmyakin, M. M.; Ovchinnikov, Y. A.; Ivanov, V. T.; Antonov, V. K.; Vinogradova, E. I.; Shkrob, A. M.; Malenkov, G.G.;Evstratov, A. V.; Laine, I. T.; Melnik, E. I.; Ryabova, I. D. J. Membr. Biol. 1969, I , 402. (33) Maschler, H. J.; Weder, H.-G.; Schwyzer, R. Helu. Chim. Acta 1971, 54,

1437.

(34) Jenny, H. B.; Riess, C.; Ammann, D.; Magyar, B.; Asper, R.; Simon, W. Mikrochim. Acta 1980, 2, 309. (35) T6th, K.; Lindner, E.; HorvBth, M.; Jeney, J.; Bitter, I.; Agai, B.; Meisel, T.; T6ke, L. Anal. Lett. 1989, 22, 1185.

Analytical Chemistty, Vol. 66,No. 4, February 15, 1994

519

Table 2. Exchango Constant8 for M$+ and Ca2+ Solutions 0MaIn.d with DMerenl Optode Membranes and pH Values' ion RTI log log log 1% I"+ optode LT pH KexchU K,,h& Kex& KexchN*c

1

ETH 7025 ETH 5418 KTFPB o-NPOE

o.41

u-. RT /LT= 0.38

.',

0.2

I

-8

1

a

"''''''''h

-6

-4

pH I6.57

-2

I

0

log a

Figure 2. Catlon exchange equilibria for the Mg*+-selective optodes I and I I I (see ExperimentalSectlon)at pH 7.19 and 6.57, respectively, containing total addltive/ionophoreratiosof 1.52 and 0.38,respectively. 0 and solid lines: experimental pointsand theoretical curves for Mg2+; right curve, p = 1 and 2 (eq 5); left curve, p = 2 (eq 9). 0 and dotted lines: experimental points and theoretical curvesfor Ca*; right curve, p = 2 (eq 3); left curve p = 2 and 3 (eq 9). Since the horizontal distance of the curves of primary and interfering ion represents the selectivity coefficient at a given degree of protonation,lothe selectivity is changing with the degree of protonatlon of the chromoionophore.

differ significantly. As expected, divalent ions are usually forming complexes with higher stability constants than monovalent ions. The Ca2+ ionophores ETH 129 and ETH 1001 are both highly selective carriers, but the latter forms complexes that are 4 orders of magnitude weaker. These results show clearly that membrane selectivities do not give information about the absolute stability of the respective cation/ionophore complexes. Stability Constants of the Mg*+-SelectiveIonophore ETH 7025. In the field of ISEs, the development of Mg2+-selective ionophores was a major challenge. When the malonate diamide derivative ETH 7025 was synthesized in our group, ISEs were obtained that could fulfill the selectivity requirements for extra- and intracellular measurement^.^^ From potentiometric measurements, it was concluded that the Mg2+selective ionophore ETH 7025 is forming complexesof variable stoichiometry with the same ion. Since the ionophore had been designed to form stable 1:l complexes with Mg2+, 155 mol 9% anionic additive (relative to the ionophore) was chosen to discriminate the interfering ion Ca2+,which was assumed to form preferably complexes of higher stoichiometry. A clearer insight into the various equilibria involved could now be gained from measurements with optodes. Studies on membranes of different compositions were necessary to evaluate the various equilibrium constants. Three optode membranes with different amounts of ionophore ETH 7025 in combination with a constant concentration of H+-selective chromoionophore ETH 5418 and NaTFPB in o-NPOE/PVC (2:l) wereequilibrated with 10"M MgC12 and CaC12 solutions in 0.1 M MOPS buffer (pH6.57-7.19). The background bufferionNa+wasassumed to form a complex of 1:1 stoichiometry with ETH 7025. Figure 2 shows the ion activities log UM* and log uta, respectively, vs 1 - a,together with the theoretical curves for Ca2+and Mg2+ by assuming the formation of complexes of different stoichiometries. The concentration of uncomplexed ionophore, and (36) Rouilly, M. Dissertation, ETH Zurich, 1990; No. 9081.

520

AnalyticalChemistry, Vol. 00, No. 4, February 15, 1994

Mg2+ Mg2+ Mg2+ Mg2+ Ca2+ Ca2+ Ca2+ Ca2+

I I1 I1 I11

I I1

I1 I11

1.52 0.64 0.64 0.38 1.52 0.64 0.64 0.38

7.19 -11.5 7.18 6.57 6.57 7.19 7.18 6.57 6.57

-8.4 -8.3 -8.3 -8.5 -8.1 -7.9 -8.0

-7.3 -5.9 -6.0

-5.8 -5.3 -5.3 -5.1 -5.8 -5.3 -5.2 -5.1

a The uncertainty of the experimental ointa was estimated to be not larger than 0.1 logarithmic unit. b &change constant for the ionophore complex of respective stoichiometry (eqs 3, 5, and 9). c Exchange constant for the back ound ion Na+,assumingformation of 1:1 complexes with the ionopTore.

therefore the selectivity coefficient,16is strongly changing with varying degree of protonation of the chromoionophore. The presented optode is therefore, in contrast to the respective ISE, not suitable for a practical application. The exchange constants obtained for theexperiments made with these optode membranes are listed in Table 2. The primary ion Mg2+ obviously forms 1:l complexes for membranes with a high amount of anionic additive (membrane I), but both ions Mg2+ and Ca2+also form 1:2 and even 1:3 complexes, if sufficient uncomplexed ionophore is present at a certain equilibrium. For Ca2+,no exchange constant could be measured for the 1:1 complex. The relatively large experimental error for the exchange constants obtained may be attributed to the complexity of the equilibria involved and to the assumption that the background ion Na+, which is considerably extracted into the membrane, is only forming 1:l complexes with the ionophore. A similar membrane, but without the ionophore ETH 7025, showed no response to the same solutions and the chromoionophore remained fully protonated. It is concluded that Mg2+,Ca2+,and Na+ are all significantly stabilized by ETH 7025. The exchange constants for a membrane containing no ionophore were evaluated with the less basic chromoionophore ETH 53 15. Since, as expected, monovalent background ions interfered too much, unbuffered MgCl2 and CaC12 solutions had to be used. The optode signal was stable and reproducible, but the determination of the pH value with a pH glass electrode was not very accurate, giving an error of f O . l pH. The exchange constants determined by this method were used to calculate the repsective values for the more basic chromoionophore ETH 5418 according to ref 18. In Table 3, the means of the exchange constants shown in Table 2 are given together with the stability constants of Mg2+ and Ca2+ of various stoichiometries with ETH 7025 in o-NPOE/PVC (2: 1) determined according to eq 7. As can be seen in Tables 1 and 3, the Ca2+-selectiveionophores ETH 129 and ETH 1001 are forming much more stable complexes than ETH 7025, which is in agreement with assumptions made previ0us1y.~~ The stability constants are increasing with increasing stoichiometry for both ions, but Ca2+ has the tendency to form complexes of higher stoichiometry than Mg2+. It can (37) Rosatzin, T.; Bakker, E.; Suzuki, K.; Simon, W. Anal. Chim. Acfa 1993,280, 197.

Table 3. Moan Exchange Constants Detwmlned for Optodes (o-NPOE/PVC, 2 1 ) wHh Chromolonophore ETH 5418 and wlthout the Ion-Selective Ionophore ETH 7025, Together with the Corresponding StaMllty Constants I"+ log &&IL ' log &I&% ' log Kexchh ' log KeXcb1(5315)* log K,,,I(5418)' log @Ld log @Gd log flhd

Mg2+ Ca2+

-11.5 nde

-8.31 f 0.08 -7.88 f 0.07

-7.30 -5.95 f 0.15

-14.8 f 0.2 -13.7 f 0.2

-24.2 f 0.2 -23.1 f 0.2

12.7 f 0.3 nde

15.9 i 0.3 15.2 f 0.3

16.9 f 0.3 17.2 f 0.4

a Exchange constant for the membrane with ionophore of respective stoichiometry; error values, standard deviation. * Experimental exchange constants for an optode with the weakly basic chromoionophore ETH 5315 without ion0 hore ETH 7025; the error value was estimated from the uncertainty of the H determination. Calculated exchange constant for an opto& with the more basic chromoionophore ETH 5418 without ionophore ETIf7025, with log Ke1&'(5315) and the relative basicity of both chromoionophores given in ref 20. d Stability constants for complexes of various stoichiometry according to eqs 5 and 9. Concentrations in the organic phase, (mol kgl). e Not determined.

Mg2+. It can therefore be understood that maximum selectivity is achieved with membranes of high additive/ ionophore ratio.38

CONCLUSIONS Formal complex formation constants of ionophores with cations may be determined by measurements on two optode membranes of analogous composition, one with and one without ionophore. By variation of the ionophore concentration, formation constants with various complex stoichiometries can be evaluated. The constants found for the primary ions are 107.9-109.3kg mol-' (1:l stoichiometry) for monovalent ions and 1012.7 kg mol-' (1: 1 stoichiometry) and 101s.2-1029.4 kg2 mol-2 (1:2 stoichiometry) for divalent ions. (38) Eugster, R.; Gehrig, P. M.; Morf, W. E.;Spichiger, U. E.; Simon, W. Anal. Chem. 1991.63.2285,

On the basis of the ion selectivities observed, it can be stated that most interfering ions also form complexes of appreciable strength in the membrane. The accessibility of stability constants may be a chance to understand more fully the correlation of structure and selectivity of hosts and does allow a more precise design of ionophores.

ACKNOWLEDGMENT This work was partly supported by the Swiss National Science Foundation, by AVL GmbH, and by Ciba Corning Diagnostic Corp. Received for review May 18, 1993.

Accepted November 19,

1993.' e

Abstract published in Aduunce ACS Abstmcts, January 1 , 1994.

AmMicaiChemistry, Vol. 66, No. 4, February 15, 1994

521