Synthesis and Evaluation of a Bis(Crown Ether) Ionophore with a Conformationally Constrained Bridge in Ion-Selective Electrodes

ANALYTICAL SCIENCES FEBRUARY 1998, VOL. 14 169 1998 © The Japan Society for Analytical Chemistry

Synthesis and Evaluation of a Bis(crown ether) Ionophore with a Conformationally Constrained Bridge in Ion-Selective Electrodes

Zhiren XIA, Ibrahim H. A. BADR, Shawn L. PLUMMER, Lawrence CULLEN, and Leonidas G. BACHAS†

Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506-0055, USA

A new bis(crown ether) ionophore containing two benzo-15-crown-5 moieties connected with each other via a

conformationally constrained bridge attached to a C18-lipophilic side chain was designed and synthesized. The bridge consisted of an isophthalic acid derivative, which was selected in order to enhance potassium over sodium binding. Liquid-polymeric membrane ion-selective electrodes (ISEs) based on this ionophore were prepared using different plasticizers and mole ratios of lipophilic ionic additives. Membrane electrodes with optimum composition (using o- nitrophenyloctyl ether or bis(1-butylpentyl)adipate as plasticizer and 60 mol% lipophilic borate additive) show Nernstian responses toward potassium (57 and 58 mV/decade, respectively) over a wide concentration range with a micromolar pot × Ð4 detection limit. These ISEs exhibit enhanced potassium over sodium selectivity (KK+,Na+=6 10 ). In addition, the + + electrodes show selectivities for potassium over Cs and NH4 that are better than those of valinomycin-based ISEs.

Keywords Ion-selective electrode, cation sensor, bis(crown ether), neutral carrier, liquid-polymeric membrane, lipophilic salt

There has been a recent rapid growth in the develop- the ionophore precludes its leaching from the polymer ment of chemical sensors because of the demand for membrane phase to the aqueous phase.11Ð13 Direct simple analytical techniques that can provide continu- attachment of a bis(crown ether) to a polymer back- ous monitoring of analytes in a variety of matrices.1Ð3 bone, however, may restrict the free movement of the In the area of membrane-based ion-selective electrodes ionophore or pose serious steric problems with the (ISEs), special emphasis has been placed on the com- backbone of the polymer, which could lead to worsen- position of the membrane phase and on the develop- ing of the response.7,14 A long flexible chain connect- ment of new ionophores, aiming at enhancing the ing the bis(crown ether) and the polymer could provide potentiometric response characteristics (e.g., selectivity, a solution to this problem. In this paper, we report on sensitivity, linear range and lifetime) of the ISEs. the synthesis and evaluation of a new bis(crown ether) The majority of cation-selective electrodes are based (BCE) ionophore with a long and flexible side chain, on neutral carrier ionophores. In particular, the anti- which can serve as a precursor for the chemical immo- biotic valinomycin has proved to be an excellent bilization to a polymer backbone. Another feature of ionophore in potassium-selective liquid-polymeric the designed ionophore is the incorporation of a confor- membrane electrodes.4 In addition to valinomycin, sev- mationally constrained bridge (an isophthalic acid eral other ionophores have been used in the preparation derivative) between the two crown ether moieties. This of potassium ISEs.5Ð8 Among these, bis(crown ethers) isophthalic acid derivative allows each of the crown with two 15-crown-5 ethers attached via a bridge7Ð10 ether macrocycles to be in a parallel alignment with exhibit a favorable potassium binding over sodium due each other during complexation, which helps maximize to the formation of sandwich-type complexes with their interaction with the analyte ion. potassium.9,10 The observed selectivities of ISEs prepared with bis(15-crown-5 ethers) are sometimes comparable to those obtained by using valinomycin as the Experimental ionophore.5 Bis(crown ethers) can be prepared with functional Reagents groups that allow their covalent attachment on polymer 5-Nitroisophthalic acid, 4-nitrobenzo-15-crown-5, matrices. This is an advantage when designing ISEs 10% Pd/C, 97% thionyl chloride, triethylamine, alu- with extended lifetimes because covalent attachment of minum oxide (neutral, Brockmann I), and stearoyl chloride were all obtained from Aldrich (Milwaukee, † To whom correspondence should be addressed. WI). Selectophore-grade tetrahydrofuran (THF), 2- I. H.A.B.: On leave from the Department of Chemistry, Faculty nitrophenyloctyl ether (NPOE), tris(2-ethylhexyl)phos- of Science, Ain-Shams University, Cairo, Egypt. phate (TOP), bis(2-ethylhexyl) sebacate (DOS), bis(1- 170 ANALYTICAL SCIENCES FEBRUARY 1998, VOL. 14 butylpentyl) adipate (BBPA), poly(vinyl chloride) meric membrane electrodes were calculated by using (PVC), and potassium tetrakis(4-chlorophenyl)borate the matched-potential method at 2.5×10Ð6 M of the pri- (KTCPB) were obtained from Fluka (Ronkonkoma, mary ion concentration.15 NY). Dichloromethane, pyridine, and dimethylform- amide (DMF) were freshly distilled from bulk solvents. Synthesis of the ionophore All metal ion salts used were of analytical grade. A 16- 5-Nitroisophthaloyl chloride. An amount of 0.3039 MΩ deionized distilled water was used for the prepara- g (2.9 mmol) of 5-nitroisophthalic acid was mixed with tion of solutions. 25 ml of thionyl chloride and allowed to reflux to yield 0.35 g of 5-nitroisophthaloyl chloride (~100% yield) Apparatus that was confirmed by 1H-NMR by comparison to liter- Cell potentials were measured using a Macintosh ature data.16 Quadra 700 equipped with a MacADIOS ABO ana- 4-Aminobenzo-15-crown-5. Closely following the log/digital input/output board (GW Instruments, procedure described by Ungaro et al.17, 1 g (3.2 mmol) Somerville, MA) and a custom-built electrode interface of 4-nitrobenzo-15-crown-5 was dissolved in 50 ml of module controlled by SuperScope II software (GW anhydrous DMF. An amount of 0.10 g of 10% Pd/C Instruments). A Parr hydrogenator was used for the was then added under nitrogen. The reagents were hydrogenation reactions. allowed to react within a Parr hydrogenator under a hydrogen pressure of 30 psi to yield 0.83 g of light Membranes and cell assembly brown, viscous liquid (93% yield). The synthesis of 4- The membranes were prepared by dissolving 1 Ð 2 aminobenzo-15-crown-5 was confirmed by 1H-NMR. mg of bis(crown ether) (BCE; ionophore 2, Fig. 1), 33 Nitro-bis(15-crown-5 ether). 4-Aminobenzo-15- mg of PVC, varied amounts of KTCPB, and 66 mg of crown-5 (838 mg, 2.9 mmol) was dissolved in 40 ml of plasticizer in 800 µl of THF. The mixture was shaken dichloromethane (anhydrous) with one equivalent of until all components were dissolved. Then, the result- pyridine and transferred to a stirred flask under a nitro- ing homogenous solution was poured into a 16-mm gen atmosphere. Using another 50 ml of solvent, 5- glass ring sitting on a glass slide, and the solvent was nitroisophthaloyl chloride (358 mg, 2.9 mmol) was dis- allowed to evaporate at room temperature. The result- solved and kept under nitrogen. This solution was ing membrane was cut into disks with a 7-mm i.d. cork added dropwise to the reaction flask over a 1-h period. porer. A single disk was mounted on the tip of a The reaction was allowed to continue for another 5 h. Philips electrode body (Philips IS-561; Glasblaserei The product was precipitated by methanol, filtered, and Möller, Zurich). The potentiometric measurements washed first with methanol and then with chloroform. were conducted using the following setup: The product was dried in a phosphorus pentoxide desic- cator under reduced pressure for a period of one day. Ag/AgCl0.10 M lithium acetatesample solution The yield was 700 mg (85%) and the m.p. was 350 Ð ISE-membrane0.10 M KClAgCl/Ag 352ûC. 1H-NMR (200 MHz, DMSO): δ 3.2 Ð 4.1 (m, 32H), 6.95 (d, 2H), 7.80 (d, 2H), 7.95 (s, 2H), 8.92 (s, All measurements of potentials were made in beakers 2H), 8.95 (s, 1H), 10.6 (s, 2H). EI-mass spectrum: that were thermostated at 25ûC. Standard potassium molecular ion peak (741). FAB-mass spectrum: chloride solutions and the other cation solutions used M+(741), MH+(742), MNa+(764), MK+(780). for the calibration plots were obtained by serial dilution Amino-bis(15-crown-5 ether). An amount of 300 mg of 1.0 M chloride salt solutions (lithium standards were (0.40 mmol) of nitro-bis(15-crown-5 ether) was dis- made from lithium acetate). The activities of the ions solved in 50 ml of anhydrous DMF. Then, 75 mg of were calculated using the Debye–Hückel equation. 10% Pd/C was added to this solution. The nitro-bis(15- Selectivity coefficients for the BCE-based liquid poly- crown-5 ether) was hydrogenated in a Parr hydrogenator (25 psi) overnight. The solid catalyst was then filtered out, and the solvent was evaporated. The product was taken up in enough anhydrous DMF to yield a clear, golden solution, that was treated with MgSO4 and filtered. The filtrate was left under nitrogen over 4 Å molecular sieves for a week. After filtering off the sieves, the solvent was eliminated using a rotary evaporator. The remaining brown residue was taken up in 70 ml of anhydrous chloroform and filtered through a celite layer in a Buchner funnel. The solvent was then eliminated in a rotary evaporator yielding 184 mg of dry product (compound 1, Fig. 1). 1H-NMR (200 MHz, Fig. 1 Structure of the bis(crown ether) ionophore 2 (R= DMSO): δ 3.2 Ð 4.1 (m, 32H), 6.85 Ð 7.98 (m, 11H),

CO(CH2)16CH3) prepared from the corresponding amino- 10.2 (s, 2H). EI-mass spectrum: molecular ion peak bis(15-crown-5 ether) 1 (R=H). (711). FAB-mass spectrum: M+(711), MH+(712), ANALYTICAL SCIENCES FEBRUARY 1998, VOL. 14 171

MNa+(734), MK+(750). of attachment of a long lipophilic chain in order to N-Stearoyl amide of 1. An amount of 178 mg (0.25 improve the lipophilicity of the ionophore and subse- mmol) of amino-bis(15-crown-5 ether) (1) was dis- quently evaluate its feasibility for use in ion-selective solved in 4 ml of anhydrous DMF and 38 ml (0.27 electrodes. mmol) of triethylamine under nitrogen and stirred for 5 It is well established that the selectivity of neutral min. Then, 75 mg of stearoyl chloride (0.25 mmol) in carrier-based liquid-polymeric membrane cation-selec- 1.5 ml of anhydrous DMF was added to the stirred tive electrodes can be optimized by the addition of solution. An immediate color change was observed lipophilic anionic additives (e.g., lipophilic tetraphenyl- and the stirring was allowed to continue for about borate derivatives) in the membrane.18,19 These another 10 min. The solvent was eliminated by evapo- lipophilic anionic sites can help reduce membrane ration under reduced pressure to yield 203 mg of light resistance and limit the interference from anions at high brown product. Flash chromatography on neutal Al2O3 sample activities. In addition, by influencing the con- (8 mm×300 mm; graduate elution from 1% of methanol centration of free carrier available for complexing in chloroform 8% of methanol in chloroform) was used cations, they can also improve the selectivities of the to purify the product (ionophore 2 in Fig. 1) that had a ISEs.18 The optimum concentration of such lipophilic 1 m.p. of 149 Ð 152ûC. H-NMR (200 MHz, CDCl3) additives in the membrane phase is dependent in part δ 0.83 (t, 3H) 1.28 (s, 30H), 2.25 (t, 2H), 3.6 Ð 4.2 (m, on the charge of the primary ion and its complexation 32H), 6.6 Ð 9.4 (m, 12H). FAB-mass spectrum: stoichiometry with the carrier as compared to that of M+(977.5), MH+(978.5), MNa+(1000.5), MK+(1016.5). the interfering ion.19 To optimize the potentiometric selectivity of BCE-based liquid-polymeric membrane Results and Discussion electrodes toward the potassium ion, different membrane compositions were prepared. In that regard, A new bis(crown ether) ionophore (ionophore 2 in membranes doped with different mole percentages of Fig. 1) was designed and synthesized. This ionophore potassium tetrakis(4-chlorophenyl)borate relative to the incorporates a conformationally constrained bridge amount of the ionophore were evaluated. As shown in between two 15-crown-5 ether moieties. This bridge Fig. 2, the selectivity of BCE-based liquid-polymeric was designed to preorganize the bis(15-crown-5 ether) membrane electrodes toward potassium was enhanced cavity to allow selective complexation of cations. An by the addition of KTCPB and reached its optimal amino group was positioned at the meta position rela- value at 60 mol% KTCPB relative to the amount of the tive to the carboxylate functionalities in the 5-amino- ionophore. In addition, the slope of the calibration plot isophthalate bridge. This amino group allows, if neces- for potassium changed from sub-Nernstian (41 mV/ sary, for attachment of the ionophore, either directly or decade) in the case of membranes without additives to through a spacer molecule to polymer matrices. In the a Nernstian response (58 mV/decade) in the case of present study, the amino group was used as point membranes doped with 60 mol% KTCPB relative to

Fig. 2 Influence of the concentration of lipophilic anionic additive (in mol% relative to the amount of the ionophore) on the potentiometric selectivity (determined by matched-potential method) of liquid-polymeric membrane electrodes based on BCE. 172 ANALYTICAL SCIENCES FEBRUARY 1998, VOL. 14 the amount of the ionophore. However, when the confor potassium over cesium was superior to that of vali- centration of the anionic sites in the membrane exceed- nomycin. ed 90 mol% (e.g., 90 mol% to 120 mol%) the potassi- Quite interestingly, we found that liquid-polymeric um selectivity of the polymeric membrane electrodes membrane electrodes based on BCE and prepared with dropped dramatically, and the selectivity pattern revert- TOP as plasticizer induce a completely different selec- ed to the Hofmeister pattern; i.e., the ISEs were more tivity pattern, compared with those prepared using o- selective toward more lipophilic cations: Cs+ > K+ > NPOE, DOS or BBPA. Membranes plasticized with Na+ > Li+. This is because at higher mol%, the poten- TOP showed high selectivity toward lithium compared tial at the membrane/sample solution interface is con- to ammonium, sodium, and cesium ions, while the trolled by the tetrakis(4-chlorophenyl)borate ion- response to potassium was similar to that of lithium exchanger. Consequently, all ISEs used in the remain- (Fig. 4, column IV). It is plausible that the enhanced der of our studies were prepared with membranes selectivity for lithium obtained with TOP as plasticizer doped with 60 mol% of KTCPB relative to the amount is due to the interaction between this plasticizer and of the ionophore. lithium ions. To support this notion, a blank membrane It has been reported in the literature that plasticizers was prepared without the bis(crown ether) ionophore; have a considerable effect on the potentiometric rather, the membrane contained only 34%(w/w) PVC response of both cation and anion-selective elec- and 66%(w/w) TOP. The selectivity of this membrane trodes.20,21 Since there is no simple rule that can electrode was measured toward different cations. account for the improvement in the potentiometric Interestingly, we found that the blank membrane elec- selectivity upon using a particular plasticizer22, a series trode gave a near-Nernstian response toward lithium of plasticizers had to be screened to select the optimal ions, and the potentiometric selectivity toward lithium membrane composition. Thus, we studied the effect of was the highest compared to that of the other cations different plasticizers, such as o-NPOE, DOS, BBPA (Fig. 4, column V), indicating that TOP preferentially and TOP, on the potentiometric response of polymeric binds to lithium. This is consistent with the selectivity membrane electrodes based on BCE in order to further data obtained using TOP as the plasticizer in the BCE- optimize their response characteristics. As can be seen based membrane electrodes. Although the mechanism in Fig. 3, membrane electrodes prepared with o-NPOE, of Li+-TOP interaction is not yet fully understood, it is BBPA or DOS as plasticizers show similar detection not without precedent among lithium-selective elec- limits and slopes with better linear range in the case of BBPA and DOS (see also data in Table 1). In terms of selectivity, it was found that among the tested plasticiz- Table 1 Respnse of the bis-crown ether based ISEs to potassium ers membrane electrodes prepared using o-NPOE as Slope Detection Response Plasticizer plasticizer had the best potassium over sodium selectiv- (mV/decade) limit/M timea/s pot × Ð4 ity (KK+,Na+= 6 10 ) (Fig. 4), which is close to that of o-NPOE 58 7×10Ð6 15 13 ISEs based on valinomycin. However, the selectivity DOS 54 6×10Ð6 25 BBPA 57 3×10Ð6 15 TOP 52 6×10Ð5 75 a. Response time refers to the concentration change from zero to 5×10Ð4 M.

Fig. 3 Potassium calibration curves (average response of three Fig. 4 Influence of different plasticizers on the selectivity of electrodes) for ISEs prepared with different plasticizers. 1: ISEs based on ionophore 2. The last column reflects the BBPA; 2: DOS; 3: o-NPOE; 4: TOP. selectivity of a blank membrane (no ionophore present). ANALYTICAL SCIENCES FEBRUARY 1998, VOL. 14 173 trodes in which TOP has been found to be the plasticiz- ration, and the National Science Foundation for financial sup- er that results in ISEs with best response toward lithi- port of this work. um.5 In addition, organophosphorous compounds with phosphine oxide functional groups show enhanced References lithium selectivity23, indicating that the selectivity pattern obtained using TOP as a plasticizer is due to its 1. V. J. Wotring, D. M. Johnson, S. Daunert and L. G. Bachas, in "Immunochemical Assays and Biosensor interaction with the lithium cation. Technology for the 1990’s", ed. R. M. Nakamura, Y. The effect of pH on the response of the BCE-based Kasahara and G. Rechnitz, p. 355, American Society of liquid-polymeric membrane electrode toward potassi- Microbiology, Washington, D. C., 1992. um was also evaluated. As demonstrated in Fig. 5, the 2. D. M. Pranitis, M. Telting-Diaz and M. E. Meyerhoff, Crit. response of BCE-based liquid-polymeric membrane Rev. Anal. Chem., 23, 163 (1992). electrodes toward 10 mM or 1 mM potassium does not 3. V. V. Cosofret and R. P. Buck, Crit. Rev. Anal. Chem., 24, practically change over the pH range 2 to 6.5. The 1 (1993). experiments were performed in a phosphoric acid solu- 4. D. Ammann, W. E. Morf, P. Anker, P. C. Meier, E. Pretsch tion, and the pH was adjusted by addition of sodium and W. Simon, Ion-Sel. Electr. Rev., 5, 17 (1983). hydroxide. 5. Y. Umezawa, "Handbook of ISEs: Selectivity Coefficients", CRC Press, Boca Raton, 1990. The response time of the ISEs was evaluated accord- 6. A. Casnati, A. Pochini, R. Ungaro, D. Bocchi, F. Ugozzoli, ing to IUPAC recommendations as the time between R. J. M. Egberink, H. Struijk, R. Lugtenberg, F. de Jong addition of analyte to the sample solution and the time and D. N. Reinhoudt, Chem. Eur. J., 2, 436 (1996). when a steady-state potential with less than 0.1 mV/ 7. K. Kimura, T. Maeda, H. Tamura and T. Shono, J. Electro- min change has been achieved.24 It was found that the anal. Chem., 95, 91 (1979). response time of the ISEs was less than 25 s as shown 8. Z. J. Shi, W. D. Hu, F. Zeng, Y. Wang and Z. F. Huang, in Table 1, which ensures the compatibility of the sen- Acta Chim. Sin., 47, 495 (1989). sor with commercial analyzers. 9. E. Lindner, K. Tòth, M. Horváh, E. Pungor, I. Bitter, B. In conclusion, the performance characteristics of liq- Ágai, L. Töke and Z. Hell, Fresenius’ Z. Anal. Chem., 322, uid-polymeric membrane ISEs based on the bis(crown 157 (1985). ether) ionophore 2 was comparable to other currently 10. E. Lindner, K. Tòth, J. Jenny, M. Horváh, E. Pungor, I. Bitter, B. Ágai and L. Töke, Mikrochim. Acta [Wien], 1990 used potassium-selective ISEs in terms of sensitivity, I, 157. detection limit, and response time. As a synthetic com- 11. S. Daunert and L. G. Bachas, Anal. Chem., 62, 1428 pound, this ionophore is also structurally modifiable. (1990). Thus, it is possible to attach this compound, either 12. O. Dinten, U. E. Spichiger, N. Chaniotakis, P. Gehrig, B. directly or through a spacer, to polymer matrices, to Rusterholz, W. E. Morf, and W. Simon, Anal. Chem., 63, yield ISEs with improved stabilities and lifetimes. 596 (1991). 13. P. B. Ross, A. J. Johonston and A. K. Judd, US Patent The authors wish to thank the Kentucky Space Grant 4,973,394, 1990. Consortium, the National Aeronautics and Space Administ- 14. K. Kimura and T. Shono, in "Ion-Selective Electrodes, 4", ed. E. Pungor, p. 175, Elsevier, Amsterdam, 1984. 15. Y. Umezawa, K. Umezawa and H. Sato, Pure Appl. Chem., 67, 508 (1995). 16. K. F. Jennings, Chem. Soc. 1957, 1172. 17. R. Ungaro, B. E. Haj and J. Smid, J. Am. Chem. Soc., 98, 5198 (1976). 18. P. C. Meier, W. E. Morf, M. Läubli and W. Simon, Anal. Chim. Acta, 156, 1 (1984). 19. R. Eugster, P. M. Gehrig, W. E. Morf, U. E. Spichiger and W. Simon, Anal. Chem., 63, 2285 (1991). 20. R. Eugster, T. Rosatzin, B. Rusterholz, B. Aebersold, U. Pedrazza, D. Rüegg, A. Schmid, U. E. Spichiger and W. Simon, Anal. Chim. Acta, 289, 1 (1994). 21. R. S. Hutchins, P. Molina, M. Alajarín, A. Vidal and L. G. Bachas, Anal. Chem., 66, 3188 (1994). 22. W. P. R. V. Stauthamer, J. F. J. Engberson, W. Verboom and D. N. Reinhoudt, Sens. Actuators B, 17, 197 (1994). 23. O. M. Petrukhin, E. N. Avdeeva, A. F. Zhukov, I. B. Polosukhina and G. V. Bodrin, in "Ion-Selective Elec- trodes, 5", ed. E. Pungor, p. 529, Pergamon Press, Oxford, 1989. 24. R. P. Buck and E. Lindner, Pure Appl. Chem., 67, 507 Fig. 5 Potentiometric response (average of three electrodes) of (1995). the bis-crown ether based-ISE to pH in the presence of 1 mM (Received November 19, 1997) (∆), and 10 mM (O) potassium ion concentrations. (Accepted December 19, 1977)