Enantiomeric discrimination of chiral crown ether ionophores containing phenazine subcyclic unit by ion-selective potentiometry

In this paper the enatiomeric selectivity of two chiral phenazino-18-crown-6 ether hosts ((R, R)-1 and (R, R)-2) is quantified. These hosts were incorporated into plasticized PVC membranes and used as recognition elements of ion-selective electrodes. The potentiometric response towards the two enantiomers of 1-phenylethylammonium ions (PEA) was measured. Potentiometric selectivity coefficients were calculated which reflect the ratio of the stability constants of the diastereomeric complexes. Ligand (R, R)-1 does not show enantiomeric recognition, while ligand (R, R)-2 has a slight preference for the (S)-(-) enantiomer over the (R)-(+) enantiomer manifested by a selectivity coefficient of 0.77. The results were compared to enantioselectivity patterns of the ligands towards α(1-naphthyl)ethyl ammonium perchlorate (NEA+ClO−4 ) enantiomers measured by circular dichroism and by 1H NMR titrations.


Introduction
Enantiomeric recognition is of great importance nowadays, when the pharmaceutical industry has adopted the new strategy of patenting enantiopure forms of certain optically active drugs. A widely used approach to distinguish between enantiomers is the application of high performance separation or electromigration techniques using chiral separation phases or forming diastereomers with chiral selectors prior to separation [1]. Another possibility is the use of enantioselective sensors or biosensors which allow the determination of the enantiomers without a preceding separation step [2,3]. These sensors comprise potentiometric enantioselective membrane electrodes, amperometric biosensors and immunosensors.
Chiral selectors used in these systems, among others, are crown ethers, natural polysaccharides, cyclodextrins [4], maltodextrins [5], polyether and macrocyclic type antibiotics, antibodies and molecularly imprinted polymers [6]. It is of utmost importance to find or synthetize chiral selectors with as high selectivity as possible. An important step of this process is testing the selectivity of the resulting compounds in the targeted chiral system.
Enantiomeric selectivity can be assessed in numerous ways like calorimetric, UV-visible and NMR titrations, solvent extraction, transport through different membrane systems, mass spectrometry [7], circular dichroism (CD) measurements [8,9] and chromatography [10]. Chiral selectors can be incorporated into a plasticized PVC based electrode membrane serving as an ionophore. Their enantioselectivity can be easily established by immersing the electrode into the solutions of the pure enantiomers separately, and measuring the potential difference formed at the membranesolution interface. From this experiment the potentiometric selectvity coefficient can be easily calculated. The latter corresponds to the ratio of the stability constants of the two ionophore host-enantiomer guest complexes. This simple procedure has been elaborated and first applied for the determination of enantioselectivity of chiral crown ether compounds in the laboratory of W. Simon [11]. Later Horvath et al. demonstrated the feasibility of the method, i.e. that the enantioselectivity coefficient obtained by this method does not depend on experimental variables, like the type of plasticizer or the addition of lipophilic salts to the membrane [12]. The potentiometric method has several advantages in the evaluation of enantiomeric selectivity. It requires very small amount of the chiral selector, typically in the order of 1 milligram. The procedure and the instrumentation is very simple and cheap and requires little time. Therefore we selected this option for the determination of the enantioselectivity of two phenazino-18-crown-6 ether ligands. These ligands were synthetized earlier [8,13] and their enantioselectivity for α-(1-naphthyl)ethylammonium perchlorate (NEA + ClO − 4 ) was probed by circular dichroism spectroscopy, resulting in qualitative information, and by 1 H-NMR titration that provided stability constants of the host-guest complexes [9].
Solutions were prepared with double-distilled water.

Electrode preparation
PVC based ion-selective electrode membranes were prepared by weighing the appropriate amount of PVC, plasticizer, lipophilic salt and chiral crown ether ionophore into a glass vial and dissolving them in tetrahydrofuran. The solution was cast into a teflon mould. After evaporation of the solvent a translucent membrane was obtained with an approximate thickness of 150-200 µm and a diameter of 21 mm. 7 mm circular disks were cut from the membranes and mounted into a Philips electrode body. 0.1 M racemic PEA + Cl − (pH=5.20) was used as an internal solution and Ag/AgCl as an internal reference electrode.
Different membrane compositions used throughout the study are shown in Table 1.

Apparatus
A Radelkis OP-208/1 type precision digital pH meter was used in the potentiometric measurements. All e.m.f. measurements were carried out in stirred solutions. A double-junction    It is known from earlier experiments that dummy membranes containing no ionophore can also respond to lipophilic cations even without lipophilic salt additives [12]. Their linear range for PEA + , however, is relatively narrow (down to 5 ·10 −3 M). They behave as low capacity ion-exchangers and a suitably lipophilic cation can enter into the membrane, creating an interfacial potential. The addition of lipophilic salts to the membrane creates liquid ion-exchanger type ion-selective electrodes that can measure PEA + cations down to 10 −5 -5·10 −6 M concentration.  Membranes prepared from ligand (R, R)-1, however, have poorer performance characteristics. They show a sub-Nernstian behavior and a relatively large drift during calibration. The course of the calibration curves of membranes 1 and 4 i.e. the ones without lipophilic salt additive is quite different from the others, but still have much lower detection limit, than dummy membranes. Membranes 2 and 3 have similar characteristics, but this more reproducible behavior is probably due to the added lipophilic salts. can be attributed to the relatively low lipophilicity of ligand (R, R)-1. This can cause the slow leaching of this ionophore from the membrane phase into the aqueous phase resulting in non-reproducible response of the electrode. Electrode membranes prepared from ligand (R, R)-2 show much better calibration behavior. They show Nernstian response with response times below 1 minute and the drift is negligible. All four membranes have very similar calibration curves. The type of plasticizer or the different added lipophilic salts do not have an influence on the shape of the curves, except for membrane 5, prepared with DOS without lipophilic salt, the linear range of which is somewhat smaller. e.m.f. is the potential difference in the measuring cell, E 0 is the standard potential of the cell s is the Nernst factor or slope of the calibration line a (S)−(−) is the activity of the measured ion and a (R)−(+) is the activity of the interfering ion (the other enantiomer). If a membrane cannot differentiate between the enatiomers there is no difference in the electrode potentials obtained in the two solutions. Enantioselective membranes show different e.m.f. responses in the solutions of the two enantiomers, the difference in e.m.f.s being larger with higher enantioselectivity. A typical measurement record is shown in Fig. 4 for membrane 7. Potentiometric selectivity coefficients calculated for all the electrodes studied are shown in Table 3.
All electrode membranes containing ligand (R, R)-2 show higher potential values in (S)-(-)-PEA + Cl − than in (R)-(+)-PEA + Cl − . It can be seen that the potentiometric selectivity coefficients obtained with different membrane compositions i.e. with different plasticizers or different or no lipophilic salt additives are not differing within the limits of the standard error. The average potentiometric selectivity coefficient is 0.77. This conforms to the ratio of the stability constants of the two complexes  [9]. They explained this phenomenon with the overcompensation of the steric repulsion between the small methyl group of the host and the naphthalene hydrogens of the guest by the strong π − π interaction.

Determination of selectivity coefficients in the solutions of hydrophilic interfering cations
Selectivity coefficients for PEA + over potentially interfering common cations were determined by the separate solution method. These data can provide useful information when enantiomer selectivity is determined in buffered solutions. Table 4 shows the results obtained for H + , Na + , K + and NH + 4 -ions. Among all the cations studied, H + -ion is the only one showing substantial interference in the membranes containing ligand (R, R)-1 as ionophore, and only in those compositions, that do not have any added lipophilic cation (Membrane 1 and 4). The other electrode membranes measure PEA + selectively over hydrophilic cations with a selectivity coefficient ranging from 6.8·10 −2 to 1.3·10 −3 . This implies that dilute buffer solutions containing the above cations do not interfere with the enan-  tiomer selectivity determination.

Conclusion
Two chiral phenazino-18-crown-6 ligands ((R, R)-1 and (R, R)-2) were incorporated into solvent polymeric electrode membranes in order to measure their enantiomeric selectivity towards (S)-(-) and (R)-(+) PEA + ions. Different membrane compositions were used changing the type of plasticizer and the type of lipophilic salt additive. After verifying that the potentiometric membrane electrodes are measuring PEA + ion, the electrode responses were recorded in (S)-(-)-PEA + Cl − and in (R)-(+)-PEA + Cl − solutions, successively. The potential differences showed in the solution of the two enantiomeric forms were used to calculate the potentiometric selectivity coefficients, which in turn correspond to the ratio of the complex stability constants of the two diastereomeric crown ether-organic ammonium salt complexes. Ligand (R, R)-1 did not show enantiomeric differentiating ability, while ligand (R, R)-2 had a slight preference for the (S)-(-) enantiomeric form over the (R)-(+) one. The average potentiometric selectivity coefficient that approximates the ratio of the two stability constants was 0.77 with a 4.16% CV. This is in good agreement with the earlier results obtained by circular dichroism and 1 H NMR measurements, i.e. ligand (R,R)-2 preferably forms heterochiral complexes (host-guest complexes with opposite configurations) with enantiomers of protonated primary ammonium salts.