DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 127–133 127 Selective electrochemical recognition of sulfate over phosphate and phosphate over sulfate using polyaza ferrocene macrocyclic receptors in aqueous solution Paul D. Beer,*a James Cadman,a José M. Lloris,b Ramón Martínez-Máñez,*b Miguel E. Padilla,b Teresa Pardo,b David K. Smith a and Juan Soto b a Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QR. E-mail: paul.beer@inorganic-chemistry.oxford.ac.uk b Departamento de Química, Universidad Politécnica de Valencia, Camino de Vera s/n, 46071 Valencia, Spain.E-mail: rmaez@qim.upv.es Received 7th September 1998, Accepted 30th October 1998 Potentiometric and electrochemical studies have been carried out with a family of ferrocene redox-functionalised polyamines (L1–L5) and have been directed towards the discrimination, using electrochemical techniques, between the two oxoanions phosphate and sulfate and the electrochemical sensing of ATP.Potentiometric titrations were carried out in THF–water (70 : 30 v/v, 0.1 mol dm23 tetrabutylammonium perchlorate, 25 8C) for L1, L2, L3, L5 and in water (0.1 mol dm23 potassium nitrate, 25 8C) for L4. Potentiometric data indicate that all receptors studied form stable complexes with sulfate, phosphate and ATP. Distribution for the ternary diagram system sulfate–phosphate–L2 shows pH dependent selectivity patterns; [L2HjSO4]j 2 2 species exist at greater than 90% in the pH range 3–4, whereas the corresponding phosphate complexes are the main species in the neutral and basic pH range.The electrochemical studies are in agreement with the speciation results. Sulfate produces in all cyclic receptors maximum cathodic shifts of the redox potential of the ferrocenyl groups around pH 3–4, whereas maximum cathodic shifts for phosphate were found between pH 7 and 8. This behaviour is not observed for the open-chain tetraamine L5.Selective quantitative electrochemical recognition of sulfate and phosphate in the presence of competing anions in aqueous solution has been achieved using the redox-active polyaza ferrocene macrocyclic L2, L3 and L4 receptors. Additionally ATP is able to cathodically shift the oxidation potential of the ferrocenyl groups of L2 and L3 receptors by up to 100 mV. The electrochemical response of L3 against ADP and AMP is also reported. Introduction Taking into account the importance of oxoanions in environmental and biological processes, the development of new oxoanion-sensing receptors is of considerable interest in fields such as environmental chemistry.In fact most of the sensors which have been developed for phosphate, sulfate, etc. do not fulfil requirements such as suYcient selectivity. With the aim of developing new chemical sensor technology, considerable interest is currently being shown in the synthesis of new receptors containing redox-active groups and binding sites for the electrochemical recognition of cationic, anionic and neutral substrates.1 This class of receptors has proved eVective in transforming host–guest interactions into measurable perturbations of the redox potential of the ligand.Examples of water soluble redox responsive receptors designed to electrochemically sense concentrations of guests in the aqueous environment are rare.2,3 This is specially so in anion-sensing where most of the studies have been carried out in non-aqueous solvents and very little is known about the potential use of ferrocene functionalised receptors as anion-sensing molecules in water.Polyamines are well known to bind anions in aqueous solution at certain pH values via favourable protonated ammonium–anion electrostatic and hydrogen bonding interactions.4 By means of incorporating the redox-active ferrocene moiety into polyamine ligand frameworks we report the study of the potential sensing behaviour against sulfate, phosphate and ATP anions of a family of ferrocene-functionalised polyamines (L1–L5) in water and THF–water mixtures.Experimental The synthesis of receptors L1, L2, L3, L4 and L5 have been published elsewhere.5–7 Physical measurements Electrochemical data were obtained with a programmable function generator Tacussel IMT-1, connected to a Tacussel PJT 120-1 potentiostat. The working electrode was graphite with a saturated calomel reference electrode separated from the test solution by a salt bridge containing the solvent/ supporting electrolyte.The auxiliary electrode was platinum wire. Potentiometric titrations were carried out in THF–water (70 : 30 v/v, 0.1 mol dm23 tetrabutylammonium perchlorate, 25 8C) for L1, L2, L3, L5 and in water (0.1 mol dm23 potassium nitrate, 25 8C) for L4, using a reaction vessel waterthermostatted at 25.0 ± 0.1 8C under nitrogen. The titrant was added by a Crison microburette 2031. The potentiometric measurements were made using a Crison 2002 pH-meter and a combined glass electrode. The titration system was automatically controlled by a PC.The electrode was calibrated by titration of well-known amounts of HCl with CO2-free KOH solution and determining the equivalence point by Gran’s method8 which gives the standard potential E98 and the ionic product of water (K9w = [H1][OH2]). The computer program SUPERQUAD9 was used to calculate the protonation and stability constants. The titration curves for each system (ca. 250 experimental points corresponding to at least three titration128 J. Chem. Soc., Dalton Trans., 1999, 127–133 N N N N NH HN NH HN NH HN HN NH HN HN NH NH NH NH NH NH Fe Fe Fe Fe Fe Fe Fe Fe Fe Fe Fe L1 L2 L3 L4 L5 curves, pH = 2log[H], range investigated 2.5–10, concentration of the ligand and anions was ca. 1.2 × 1023 mol dm23) were treated either as a single set or as separate entities without significant variations in the values of the stability constants.Results and discussion Potentiometric anion binding studies Phosphate and sulfate complexation. Speciation studies have been carried out in THF–water (70 : 30 v/v, 0.1 mol dm23 tetrabutylammonium perchlorate, 25 8C) for L1, L2, L3 and L5 and in water (0.1 mol dm23 potassium nitrate, 25 8C) for L4. Tables 1 and 2 report the stability constants found for the L–H1–A systems (L = L1, L2, L3, L4, A = sulfate, phosphate). It is well known that macrocyclic polyamines in solution form protonated species which can interact with anions via electrostatic forces and hydrogen bonds.10 With receptors L1–L4 an additional favourable electrostatic interaction with the anionic guest will result from the oxidised ferrocenium moieties in electrochemical experiments (see below).Table 1 gives the stoichiometry of the species formed and the stability constants with phosphate. There is interaction between the receptors and the phosphate anion in a wide pH range (ca. 1–10).Despite the use of diVerent solvents (THF–water and water) the stoichiometries found in solution for the phosphate complexes formed are quite similar. In all cases 1 : 1 complexes were found. Fig. 1 shows the distribution diagram of the species for the L2–H1–phosphate system. Taking into account the complexity of the studied system the evaluation of the existing species in solution throughout the pH range studied is rather diYcult.11 Table 1 Logarithms of the stability constants for the interaction of L1, L2, L3, L4 or L5 with phosphate in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate) for L1, L2, L3, L5 and water (25 8C, 0.1 mol dm23 potassium nitrate) for L4 a Reaction L 1 2H 1 PO4 H2LPO4 b L 1 3H 1 PO4 H3LPO4 L 1 4H 1 PO4 H4LPO4 L 1 5H 1 PO4 H5LPO4 L 1 6H 1 PO4 H6LPO4 H2L 1 PO4 H2LPO4 H3L 1 PO4 H3LPO4 H4L 1 PO4 H4LPO4 H4L 1 HPO4 H5LPO4 H4L 1 H2PO4 H6LPO4 L1 41.63(2) 48.23(2) 50.28(3) 15.55 10.11 4.65 L2 31.03(5) 37.72(4) 43.96(6) 49.49(3) 9.21 11.01 5.21 2.68 L3 40.51(3) 48.58(1) 53.12(2) 12.65 8.85 5.08 L4 25.66(1) c 36.27(1) 45.13(1) 52.60(1) 59.51(1) 5.54 8.14 10.36 5.8 5.68 L5 24.88(2) 33.49(1) 41.18(1) 46.94(1) 50.07(3) 8.74 11.82 16.23 10.14 4.90 a Basicity constants for L1 ref. 13, L2 ref. 6. L3 in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate): logb1 = 9.00(1), logb2 = 16.89(1), logb3 = 24.00(1), logb4 = 27.86(1). L5 in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate): logb1 = 8.83(1), logb2 = 16.14(1), logb3 = 21.67(1), logb4 = 24.95(1), phosphate in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate): logb1 = 11.85(1), logb2 = 20.22(1), logb3 = 24.41(1).L4 in water (25 8C, 0.1 mol dm23 potassium nitrate): logb1 = 10.67(1), logb2 = 20.12(1), logb3 = 28.13(2), logb4 = 34.77(3). b Charges have been omitted for clarity. c Values in parentheses are the standard deviations in the last significant digit.Table 2 Logarithms of the stability constants for the interaction of L1, L2, L4 or L5 with sulfate in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate) for L1, L2, L5 and water (25 8C, 0.1 mol dm23 potassium nitrate) for L4 a Reaction L 1 H 1 SO4 HLSO4 b L 1 2H 1 SO4 H2LSO4 L 1 3H 1 SO4 H3LSO4 L 1 4H 1 SO4 H4LSO4 L 1 5H 1 SO4 H5LSO4 HL 1 SO4 HLSO4 H2L 1 SO4 H2LSO4 H3L 1 SO4 H3LSO4 H4L 1 SO4 H4LSO4 H4L 1 HSO4 H5LSO4 L1 30.98(5) 35.02(5) 4.9 5.77 L2 29.80(2) 35.16(4) 3.09 5.28 L4 37.01(1) 40.83(1) 2.24 3.52 L5 12.05(3) c 20.05(2) 26.64(1) 31.89(1) 35.09(2) 3.22 3.91 4.97 6.94 6.86 a Basicity constants for sulfate in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate): logb1 = 3.28(1).b Charges have been omitted for clarity. c Values in parentheses are the standard deviations in the last significant digit.J. Chem. Soc., Dalton Trans., 1999, 127–133 129 Nevertheless bearing in mind the protonation constants of the receptors and the phosphate we have tentatively assigned the complexes H4LPO4 to the interaction of H2L21 1 H2PO4 2 and H5LPO4 to H3L31 1 H2PO4 2, taking into account that the H2PO4 2 is in greatest abundance in the pH ranges 4.06–8.08 (in THF–water) and 4.31–8.31 (in water).Assuming these interactions between species, the logarithms of the stability constants for the equilibria H2L21 1 H2PO4 2 H4LPO4 and H3L31 1 H2PO4 2 H5LPO4 (L = L1 to L4) are in the range 1.59–6.28 and 2.04–5.88, respectively. The complex H6LPO4 exists at maximum concentration at pH 4–5 and probably involves H4L41 and H2PO4 2.The nature of the remaining complexes is less clear. The stability constants corresponding to the equilibrium of L1, L2 and L4 with sulfate have also been determined by pH-metric titrations. Stability constants are reported in Table 2. For receptors L1, L2 and L4 receptor–sulfate interactions have only been found at pH lower than 7.Tentatively H4LSO4 and H5LSO4 species are attributed to H4L41 1 SO4 22 and H4L41 1 HSO4 2, respectively. Fig. 2 shows the distribution diagram for the L2–H1–sulfate system. One of our main goals in this study was the development of selective electrochemical sensing receptors able to discriminate between the oxoanions sulfate and phosphate. In order to detect selectivity and determine which are the prevailing species in solution in a mixture of sulfate and phosphate with the receptors L1, L2 and L4, we have calculated the distribution diagram Fig. 1 Distribution diagram of the species for the system L2–H1– phosphate. Fig. 2 Distribution diagram of the species for the system L2–H1– sulfate. of the ternary sulfate–phosphate–L systems by plotting the overall percentages of the free receptors and the sulfate–L and phosphate–L complexes as a function of the pH.11 These diagrams show the competition between sulfate and phosphate (equimolecular amounts) to interact with a target receptor.Fig. 3 shows the ternary diagram for the L2–sulfate–phosphate system. The figure clearly displays the pH dependent selectivity patterns. [L2HjSO4]j 2 2 species exist at greater than 90% in the pH range 3–4, whereas the corresponding phosphate complexes are the main species in the neutral and basic pH range. Similar ternary diagrams are obtained for L1–sulfate–phosphate systems, with predominant sulfate complexes at acid pH and predominant phosphate complexes at neutral and basic pH.This trend is also observed for L4 but phosphate predominates in the presence of sulfate in the pH range studied. This data strongly suggests that some receptors are able to selectively complex sulfate or phosphate by pH modulation. For the sake of comparison the protonation and formation of sulfate and phosphate complexes with the open-chain tetraamine L5 have also been determined in THF–water 70 : 30 v/v. Tables 1 and 2 list the stability constants found.Despite the diVerent geometric architecture of L5 (open-chain against cyclic) the stoichiometries and stability constants of the complexes are in general similar to those found for the cyclic receptors L1, L2, L3 and L4. Additionally the ternary diagram for L5–sulfate–phosphate also displays sulfate species as predominant at acid pH and phosphate complexes as the main species at neutral pH. Fig. 3 Distribution diagram for the ternary system sulfate– phosphate–L2.The sum of percentages of complexed species are plotted vs. pH. [L2] = [sulfate] = [phosphate] = 8 × 1023 mol dm23. Fig. 4 Distribution diagram of the species for the system L1–H1–ATP.130 J. Chem. Soc., Dalton Trans., 1999, 127–133 Table 3 Logarithms of the stability constants for the interaction of L1, L2, L3, L4 or L5 with ATP in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate) for L1, L2, L3, L5 and water (25 8C, 0.1 mol dm23 potassium nitrate) for L4 a Reaction L 1 H 1 ATP HLATPb L 1 2H 1 ATP H2LATP L 1 3H 1 ATP H3LATP L 1 4H 1 ATP H4LATP L 1 5H 1 ATP H5LATP L 1 6H 1 ATP H6LATP L 1 7H 1 ATP H7LATP HL 1 ATP HLATP H2L 1 ATP H2LATP H3L 1 ATP H3LATP H4L 1 ATP H4LATP H4L 1 HATP H5LATP H4L 1 H2ATP H6LATP H4L 1 H3ATP H7LATP L1 28.99(4) 36.45(5) 41.72(3) 44.84(3) 6.15 10.37 7.78 4.82 L2 33.69(5) 38.27(4) 43.67(4) 6.98 3.70 3.02 L3 12.67(7) c 22.29(5) 30.43(6) 38.59(5) 45.47(5) 50.46(7) 3.67 5.40 6.43 10.73 10.10 11.33 L4 23.32(9) 31.96(9) 39.58(8) 45.99(11) 50.40(14) 53.44(16) 3.20 3.83 4.81 4.44 4.84 5.86 L5 21.39(1) 29.34(1) 35.75(1) 39.54(1) 41.07(9) 5.25 7.67 10.80 7.08 4.60 a Basicity constants for ATP in THF–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 tetrabutylammonium perchlorate): logb1 = 7.51(1), logb2 = 11.52(1), logb3 = 14.07(3).Basicity constants for ATP in H2O (25 8C, 0.1 mol dm23 potassium nitrate): logb1 = 6.78(1), logb2 = 10.79(2), logb3 = 12.81(5).b Charges have been omitted for clarity. c Values in parentheses are the standard deviations in the last significant digit. ATP complexation. In Table 3 the stability constants of the cyclic L1, L2, L3 and L4 and the open-chain L5 polyamines with ATP are reported. Stability constants found due to the interaction of the protonated forms of the receptors with ATP are generally higher in THF–water 70 : 30 v/v than those found for L4 in water.Fig. 4 shows the distribution diagram for the L1– H1–ATP system. Receptor L4 is fully protonated (H4L4)41 at pH lower than 6.6. On the other hand the first protonation of free ATP in water is ca. 6.7. Therefore the complexes expected to exist in solution involve the interaction of HjLj1 species and ATP42 (H2L, H3L and H4L for species H2LATP, H3LATP and H4LATP species in Table 3). Further protonated species H5LATP and H6LATP are probably related to the interaction of H4L41 with HATP32 and H2ATP22, respectively. The value found for the open-chain tetraamine spermine [H2N(CH2)3- NH(CH2)4NH(CH2)3NH2] in water for its tetraprotonated form with ATP has been reported to be 3.97 which is a value close to that found for (H4L4)41 and ATP42.12 In THF–water with receptors L1, L2 and L3 the situation is more complex. The first protonation constant of ATP in THF–water is ca. 7.51. On the other hand the last protonation constants for L1, L2, L3 and L5 are ca. 3.2–4.8.The diVerence between the first protonation constant of ATP and fully protonated species L1, L2, L3 and L5 is now larger than for L4 in water and therefore several species can coexist in solution and it is more diYcult to determine the nature of the complexes taking into account only stability constant values. Electrochemical anion recognition investigations One of the most interesting features in receptors L1 to L5 is the presence near co-ordination sites of redox-active groups.These can be aVected by the presence of closely bound anionic guest species and transform the receptor–substrate interaction into a macroscopic electrochemical response. The shift of the redox potential of the ferrocenyl groups as a function of the pH in the presence and absence of sulfate, phosphate, ATP and nitrate anions was monitored in THF–water (70 : 30 v/v, 0.1 mol dm23 tetrabutylammonium perchlorate, 25 8C) for L1, L2, L3, L5 and in water (0.1 mol dm23 potassium nitrate, 25 8C) for L4.A unique oxidation potential wave was observed for all the receptors throughout the pH range, except for L4 at neutral pH in which two unresolved waves were observed. Plots of E1/2 vs. pH show for all receptors that a steady anodic shift of the redox potential occurs when the solution is acidified. The diVerence found between the oxidation potential at basic pH (pH = 12) and acidic pH (pH = 0) (obtained by extrapolation of the curves E1/2 vs. pH because of the instability of ferrocenyl groups at pH lower than 2) was 100, 260, 250, 326 and 110 mV, for L1–L5 respectively.As a general rule the fewer the number of ferrocenyl centres and the closer the N-donor atoms are to the redox-centres, the larger is DE1/2.13 Electrochemical response towards sulfate and phosphate. The electrochemical response of sulfate, phosphate and nitrate anions was monitored as a function of pH range. Plots of E1/2 vs. pH for the systems L–H1–A, (L = L1 to L5; A = sulfate, phosphate, nitrate) with a ligand-to-anion molar ratio of 1 : 1 have been determined.Fig. 5 graphically displays the electrochemical anion response found for receptors L1, L2, L3, L4 and L5 as a function of the pH [DE1/2 defined as E1/2 (receptor) 2 E1/2 (anion–receptor)]. Nitrate does not produce any significant redox potential shift at any pH value. Sulfate produces in all receptors maximum cathodic shifts of the redox potential of the ferrocenyl groups around pH 3–5, whereas maximum cathodic shifts for phosphate were found between pH 6 and 8.Maximum selective redox potential shifts (DE1/2) of 54 and 50 mV were observed for sulfate and phosphate, using receptors L2 and L4 at pH 4 and 7, respectively (see Fig. 5). If we compare the potentiometric data and the electrochemical response as a function of the pH the results appear to suggest that the contributions to DE1/2 of the diVerent species found in solution are not the same.For example from Fig. 1 and Fig. 5 it can be observed that although phosphate interacts with L2 in the range pH 2 to 9, the maximum electrochemical response was found in the pH range 5–7 suggesting that only the [H5L2PO4]21 and [H4L2PO4]1 species are able to significantly perturb the oxidation potential of the ferrocenyl moiety, whereas the [H6L2PO4]31 and [H3L2PO4] complexes are not capable of doing so. This is also observed for the remaining receptors L1, L3 and L4, for which the maximum phosphate–receptor interaction always coincides with the pH range of existence of the [H5LPO4]21 and [H4LPO4]1 species.Assuming that [H5LPO4]21 and [H4LPO4]1 species are associated with the interaction of the H2PO4 2 anion with H2L21 and H3L31 species it can be concluded that receptors L1 to L4 are able to selectively detect the presence of the H2PO4 2 anion. From our point of view this is of importance because the data suggests for the first time, to the best our knowledge, that there is a selective electrochemical speciation in the sense that not all the HjPO4 j 2 3 species produce the same oxidation potential shift of the ferrocene groups.For all the L1, L2, and L4 receptors sulfate produces oxidation potential shifts at pH values lower than 7, where the species [H4L2SO4]21 and [H5L2SO4]31 exist. In order to demonstrate the potential use of redoxfunctionalised receptors as practical sensors we have carried out studies on the selective quantitative determination of sulfateJ. Chem.Soc., Dalton Trans., 1999, 127–133 131 Fig. 5 Redox potential shift (DE1/2) for L1, L2, L3, L4 and L5 in the presence of phosphate and sulfate as a function of the pH. and phosphate using receptors L2, L3 and L4. Although the following studies, from a practical point of view, can probably not be applied to a real analytical problem they point out the selective nature of the interaction and reinforce the arguments stated above.For example Fig. 6 shows DE1/2 at pH = 4.0 versus sulfate-to- L2 ratios in the presence and absence of phosphate ([L2] = 50 × 1025 mol dm23; [phosphate] = 52 × 1025 mol dm23). Apart from the selectivity exhibited for sulfate in the presence of phosphate, Fig. 6 indicates that 1 : 1 complexes are formed. This is in agreement with the ternary diagram in Fig. 3 which indicates that in a mixture of sulfate and phosphate at pH 4 the L2 receptor selectively forms complexes with sulfate.We have also determined DE1/2 vs. phosphate-to-L ratios for receptor L3 and L4 at pH 8 and 7, respectively. The linear range of the curve in Fig. 6 (sulfate anion-to-receptor ratios < 0.9 : 1) can be used132 J. Chem. Soc., Dalton Trans., 1999, 127–133 Table 4 Determination of the concentration of sulfate in the presence of phosphate, nitrate, chloride or acetate with receptor L2 in THF–water (70 : 30 v/v) at pH 4.0 by using electrochemical methods a [sulfate] × 105 14.3(8) a [15.2] b 29(1) [29] 39(2) [42] [sulfate] × 105 22.0(7) c [15.0] b 31(1) [29] 41(1) [42] [sulfate] × 105 12.8(6) d [13.4] b 25(1) [26] 33(2) [37] [sulfate] × 105 11.1(3) e [11.1] b 23(1) [21] 29(1) [31] [sulfate] × 105 12.2(8) f [11.3] b 23(1) [22] 31(2) [32] a Concentration (mol dm23) determined by electrochemical methods.Values in parentheses are the standard deviations in the last significant digit. b Sulfate concentration (mol dm23). c [sulfate] determined in the presence of phosphate, [phosphate] = 52 × 1025 mol dm23. d [sulfate] determined in the presence of nitrate, [nitrate] = 46 × 1025 mol dm23.e [sulfate] determined in the presence of chloride, [chloride] = 38 × 1025 mol dm23. f [sulfate] determined in the presence of acetate, [acetate] = 38 × 1025 mol dm23. as a calibration curve for the quantitative determination of sulfate, whereas linear ranges in DE1/2 vs. phosphate-to-L ratio curves for receptor L3 and L4 have been used for the quantitative determination of phosphate. Table 4 shows the selective determination of sulfate in the presence of phosphate, nitrate, chloride, or acetate.The presence of chloride or acetate, which are able to interact with protonated polyamines, does not appear to significantly aVect the sulfate determination indicating that sulfate can be selectively determined in the presence of these competing anions. Table 5 gives the results found in the selective quantitative determination of phosphate using receptor L3 employing electrochemical methods, whereas Table 6 reports the selective determination of phosphate using L4 in water in the presence of sulfate. Sulfate is not able to perturb the electrochemical response against phosphate at pH 7 in agreement with the tertiary diagram of the L4–sulfate– phosphate system which shows predominant L4–phosphate versus L4–sulfate species.Of particular note is the selective quantitative phosphate determination in water even in the presence of other anions such as sulfate and nitrate (often present in water), at the environmentally typical neutral pH.The importance of the molecular architecture is noteworthy when the comparison is drawn between the electrochemical response found for the cyclic receptors L1, L2, L3 and L4 and the open-chain tetraamine L5. The half-wave potential of the open-chain tetraamine L5 is also pH dependent, but neither the presence of nitrate nor phosphate produce any significant change in the oxidation potential of the ferrocenyl groups in clear contrast with that found for the corresponding cyclic tetraamines L1 to L4.On the contrary at acid pH L5 is able to electrochemically recognise sulfate. Bearing in mind that both cyclic and acyclic tetraamines form stable complexes with Fig. 6 Redox potential shift (DE1/2) of L2 vs. sulfate-to-L2 ratios in the absence (s) and presence of phosphate (r). sulfate and phosphate (see above), the diVerent electrochemical response can only be attributed to a diVerent molecular architecture (cyclic versus acyclic).In considering the electrochemically observed behaviour one should be aware of the nature of the interaction process between the ferrocene/ferrocenium groups and the anion. In a first step for a determined pH the anion interacts with the poly-amine/-ammonium cavity via electrostatic forces and/or hydrogen bonds. In a second step when the ferrocene groups are oxidised to ferrocenium an additional cation(ferrocenium)– anion interaction would occur.This ferrocenium–anion interaction would probably be the factor having the largest contribution to the oxidation potential shift found using electrochemical techniques. This interaction would be favoured if the ferrocene groups are fixed and are in close proximity to the anion bound within the cavity. By considering the molecular architecture of cyclic and acyclic receptors it seems clear that most of these factors can be better accommodated by receptors L1, L2, L3 and L4 than by the open-chain molecule L5 and in general one would expect to obtain a greater degree of selectivity and larger DE1/2 shifts in the presence of anions in cyclic rather than in acyclic receptors.Electrochemical response towards ATP. The electrochemical response of receptors L1, L2, L3, L4 and L5 towards ATP in THF–water (70 : 30 v/v) has also been monitored as a function of the pH.Fig. 7 shows DE1/2 [DE1/2 defined as E1/2 (receptor) 2 E1/2 (anion–receptor)] for the L–H1–ATP systems. Although all the receptors L1, L2, L3, L4 and L5 have been found to form stable complexes with ATP their electrochemical response is quite diVerent. First it is interesting to point out that Table 5 Determination of the concentration of phosphate in the presence of sulfate and nitrate with receptor L3 in THF–water (70 : 30 v/v) at pH 8.0 by using electrochemical methods a [PO4 23] × 105 14.8(8) a [11.8] b 21(1) [23] 33(2) [33] [PO4 23] × 105 11(2) c [12] b 21(1) [23] 33(2) [33] [PO4 23] × 105 14(1) d [15] b 28(3) [28] 39(2) [41] a Concentration (mol dm23) determined by electrochemical methods.Values in parentheses are the standard deviations in the last significant digit. b Sulfate concentration (mol dm23). c [PO4 23] determined in the presence of sulfate, [SO4 22] = 42 × 1025 mol dm23. d [PO4 23] determined in the presence of nitrate, [NO3 2] = 50 × 1025 mol dm23.Table 6 Determination of the concentration of phosphate in the presence of sulfate with receptor L4 in water at pH 7.0 by using electrochemical methods a [phosphate] × 105 14(2) a [14] b 27(2) [27] 39.1(9) [39.0] [phosphate] × 105 15(1) c [15] b 27(2) [29] 41(2) [42] a Concentration (mol dm23) determined by electrochemical methods. Values in parentheses are the standard deviations in the last significant digit. b Phosphate concentration (mol dm23).c [phosphate] determined in the presence of sulfate, [sulfate] = 52 × 1025 mol dm23.J. Chem. Soc., Dalton Trans., 1999, 127–133 133 ATP is able to cathodically shift the oxidation potential of the ferrocenyl groups of receptors L2 and L3 by up to 100 mV. Thus DE1/2 found in aqueous solutions for ATP is quite large and is even larger than some of the DE1/2 values found for the interaction of polyazaalkanes with metal ions. In general for the same receptor transition metal ions form more stable complexes than anions, however the large DE1/2 found for L2 and L3 with ATP suggest that there is no direct relation between stability constants and oxidation potential shift.The L5 receptor displays the lowest oxidation potential shift (DE1/2 lower than 20 mV) in the presence of ATP. This appears to reinforce the fact that macrocyclic receptors compared to acyclic structures generally exhibit an enhanced electrochemical recognition eVect.There is also a contrast between the electrochemical response of receptors L2, L3, L4 and L1 which could be explained by taking into account the smaller cyclic cavity in L1 when compared with L2, L3 and L4. Additionally we have also carried out preliminary studies on the electrochemical recognition of ADP and AMP. ATP, ADP and AMP are a series of anions where the charge and the size is steadily reduced from ATP to AMP. Fig. 8 shows DE1/2 [DE1/2 defined as E1/2 (receptor) 2 E1/2 (anion–receptor)] for the L3–H1–A (A = ATP, ADP, AMP) systems as a function of the pH.Maximum oxidation potential shift was found about pH 6–7, where the anions are in their deprotonated form ATP41, ADP31 and AMP21. Fig. 7 Redox potential shift (DE1/2) for L1, L2, L3, L4 and L5 in the presence of ATP. Fig. 8 Redox potential shift (DE1/2) for L3 in the presence of ATP, ADP and AMP anions. Conclusions In summary we have shown that redox-active ferrocene polyazamacrocyclic receptors L1–L4 can, through an electrochemical response, selectively detect at certain pH values sulfate and phosphate in the presence of competing anions in the aqueous environment.A diVerent electrochemical response has been found for open-chain receptor L5 pointing out the importance of the molecular architecture in the electrochemical recognition process. Maximum selective redox potential shifts (DE1/2) of 54 and 50 mV were observed for sulfate and phosphate, using receptors L2 and L4 at pH 4 and 7, respectively.Larger cathodic DE1/2 shifts of up to 100 mV have been found for ATP and L2 and L3. Both the selectivity and the large redox potential shift found for some anions strongly suggest the potential use of these receptors as transducers in amperometric sensor devices in the near future. Of particular note is the selective quantitative phosphate determination in water in the presence of competing anions at the environmentally common neutral pH.Acknowledgements We should like to thank the DGICYT (proyecto PB95-1121- C02-02) for support. We also thank the EPSRC and British Petroleum for studentships (J. C., D. K. S.) and the EPSRC for use of the mass spectrometry service at University College Swansea. References 1 See for example, P. D. Beer, M. G. B. Drew and R. Jagessar, J. Chem. Soc., Dalton Trans., 1997, 881; P. D. Beer, A. R. Graydon, A. O. M. Johnson and D. K. Smith, Inorg. Chem., 1997, 36, 2112; P.D. Beer, Chem. Commun., 1996, 689; P. D. Beer, Chem. Soc. Rev., 1989, 18, 409; P. D. Beer, M. G. B. Drew, D. Hesek, J. Kingston, D. K. Smith and S. E. Stokes, Organometallics, 1995, 14, 3288; P. D. Beer, Z. Chen, M. G. B. Drew and P. A. Gale, J. Chem. Soc., Chem. Commun., 1995, 1851. 2 M. E. Padilla-Tosta, R. Martínez-Máñez, T. Pardo, J. Soto and M. J. L. Tendero, Chem. Commun., 1997, 887; J. M. Lloris, R. Martínez-Máñez, T. Pardo, J. Soto and M. E. Padilla-Tosta, Chem. Commun., 1998, 837. 3 P. D. Beer, Z. Chen, M. G. B. Drew, J. Kingston, M. Ogden and P. Spencer, J. Chem. Soc., Chem. Commun., 1993, 1046. 4 A. Bianchi, K. Bowman-James and E. García-España (Editors), Supramolecular Chemistry of Anions, Wiley-VCH, New York, 1997. 5 P. D. Beer, J. E. Nation, S. L. W. McWhinnie, M. E. Harman, M. B. Hursthouse, M. I. Ogden and A. H. White, J. Chem. Soc., Dalton Trans., 1991, 2485. 6 M. J. L. Tendero, A. Benito, R. Martínez-Máñez, J. Soto, J. Paya, A. J. Edwards and P. R. Raithby, J. Chem. Soc., Dalton Trans., 1996, 343; J. M. Lloris, R. Martínez-Máñez, M. E. Padilla-Tosta, T. Pardo and J. Soto, J. Chem. Soc., Dalton Trans., 1998, 3657. 7 P. D. Beer, Z. Chen, M. G. B. Drew, A. O. M. Johnson, D. K. Smith and P. Spencer, Inorg. Chim. Acta, 1996, 246,143. 8 G. Gran, Analyst (London), 1952, 77, 661. F. J. Rossotti and H. J. Rossotti, J. Chem. Educ., 1965, 42, 375. 9 P. Gans, A. Sabatini and A. Vacca, J. Chem. Soc., Dalton Trans., 1985, 1195. 10 B. Dietrich, M. W. Hosseini and J. M. Lehn, J. Am. Chem. Soc., 1981, 103, 1282; A. Bencini, A. Bianchi, C. Giorgi, P. Paoletti and B. Valtancoli, Inorg. Chem., 1996, 35, 1114. 11 B. Dietrich, J. Guilhem, J. M. Lehn, C. Pascard and E. Sonveaux, Inorg. Chim. Acta, 1984, 67, 91; A. Andres, J. Aragó, A. Bencini, A. Bianchi, A. Domenech, V. Fusi, E. Garcia-España, P. Paoletti and J. A. Ramírez, Inorg. Chem., 1993, 32, 3418; A. Andres, C. Bazzicalupi, A. Bencini, A. Bianchi, V. Fusi, E. Garcia-España, N. Nardi, P. Paoletti and J. A. Ramírez, J. Chem. Soc., Perkin Trans. 2, 1994, 2367. 12 B. Dietrich, L. Fyles, T. M. Fyles and J. M. Lehn, Helv. Chim. Acta, 1979, 2763; A. Bencini, A. Bianchi, M. I. Burguete, A. Domenech, E. García-España, S. V. Luis, M. A. Niño and J. A. Ramírez, J. Chem. Soc., Perkin Trans. 2, 1991, 1445. 13 A. Benito, R. Martínez-Máñez, J. Soto and M. J. L. Tendero, J. Chem. Soc., Faraday Trans., 1997, 93, 2175. Paper 8/06944A