Host–guest assembly of ligand systems for metal ion complexation. Synergistic solvent extraction of copper(II) and silver(I) by 1,4,8,11-tetrabenzyl-1,4,8,11-tetraazacyclodecane in combination with carboxylic acids

Abstract

Host–guest formation between 1,4,8,11-tetrabenzyl-1,4,8,11-tetraazacyclodecane (1) and lipophilic organic carboxylic acids in chloroform has been investigated and the effect of such ligand assembly on the solvent extraction of copper(II) and silver(I) has been probed. NMR titration experiments in the absence of a metal ion confirm the formation of weak 1 ∶ 1 and 1 ∶ 2 (macrocycle ∶ carboxylic acid) assemblies in CDCl₃ between 1 and palmitic (hexadecanoic) acid or 4-tert-butylbenzoic acid while difunctional salicylic acid showed a 1 ∶ 2 interaction that is somewhat stronger. The interaction between the former two acids and the tetra-N-benzylated macrocycle is significantly less than that reported previously for its non-substituted parent, cyclam; a result that likely reflects the presence of the less-basic, more sterically hindered tertiary nitrogens in 1 relative to the secondary nitrogens present in cyclam. Carboxylic acid-containing assemblies of this type have been used as extractants in a series of solvent extraction (water/chloroform) experiments. From both previous observations as well as from entropy considerations, it was anticipated that the use of a host–guest assembly of the above type for metal-ion complexation might contribute to enhanced metal ion binding (and concomitant enhanced metal ion extraction). Such behaviour is postulated to arise from the components of the coordination sphere being, at least in part, assembled for complex formation. In accord with this, the use of the ligand assembly involving palmitic acid/macrocycle 1 was found to lead to enhanced (synergistic) extraction of copper(II) at a metal ion concentration of 10⁻³ mol dm⁻³ while, for silver(I), synergism was somewhat marginal at this concentration but was clearly apparent under related conditions when the silver concentration was reduced to 10⁻⁴ mol dm⁻³. Similar behaviour towards silver was also observed when 4-tert-butylbenzoic acid was substituted for palmitic acid, while the use of salicylic acid resulted in enhanced (synergistic) extraction at both metal ion concentrations.

---

Introduction

It has been recognised for some time that the presence of supramolecular association between ligands may be employed to influence the selectivity and efficiency of separation processes.¹ For example, the use of discrete ligand assemblies for metal complexation (where the assembly exists in solution in equilibrium with its corresponding metal complex) has the potential to result in enhanced complex stability-behaviour previously referred to as the ‘assembly effect’.^2,3 This effect was proposed to reflect the presence of assembled ligand components such that the latter favourably influences formation of a corresponding metal complex. For example, it has been recently shown that the use of host–guest assemblies of type [LH₂(RX)₂] between lipophilic carboxylic or phosphinic acids [RXH] and dibenzo-substituted N₃O₂-donor macrocycles [L] incorporating secondary amine groups in a chloroform phase in each case resulted in enhanced (synergistic) solvent extraction of copper(II) from aqueous solution.⁴‡

A motivation for the present investigation was to extend the above studies to include assemblies based on a N₄-donor macrocyclic ligand incorporating only tertiary nitrogen donors, namely, 1,4,8,11-tetrabenzyl-1,4,8,11-tetraazacyclodecane (tetrabenzylcyclam) 1.

We now present the results of an investigation of the formation of host–guest assemblies between 1 and selected carboxylic acids in chloroform as well as of the solvent extraction of copper(II) and silver(I) by 1 in combination with palmitic acid for the former metal ion and palmitic, 4-tert-butylbenzoic and salicylic acid for the latter. It has been demonstrated previously that 1 forms 1 ∶ 1 complexes in solution with both copper(II) and silver(I).^5,6 Furthermore, an X-ray structure determination of the 1 ∶ 1 complex of nickel(II) benzoate with cyclam shows a close agreement of the binding pattern with the structure of the preorganized ligand package of cyclam with 4-tert-butylbenzoic acid, suggesting that an assembly effect will occur for this and related systems.⁷

Experimental

All reagents and solvents were of the highest commercial grade available and were used without further purification. The tetrabenzylated derivative 1 was synthesised and characterised as reported previously.⁸ 1,4,8,11-Tetramethyl-1,4,8,11-tetraazacyclodecane was obtained commercially. All aqueous solutions were prepared using distilled water; chloroform was presaturated with water before use in the solvent extraction experiments.

NMR Titrations

The NMR titration experiments involving 1 and 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane and conducted using Bruker AM300 and AC200 NMR spectrometers at 297 K. For the NMR titration studies, a weighed amount of the required carboxylic acid was added incrementally to the respective macrocycles (at ∼0.07 mol dm⁻³) dissolved in deuterated chloroform (0.5 cm³) in the NMR tube; the amount added was determined by weight difference before and after each addition. Chemical shift changes in the ¹H and/or ¹³C NMR spectra were recorded after each addition. Induced chemical shifts were plotted as a function of the mole ratio of palmitic, benzoic, 4-tert-butylbenzoic or salicylic acid to 1 present at each titration point. Conditional step-wise formation constants in CDCl₃ were calculated from the corresponding NMR titration curves using a local version of EQNMR.⁹

Solvent extraction experiments

Unlike cyclam with its four –NH– groups, preliminary experiments confirmed that the more lipophilic derivative 1 (or its carboxylic acid adducts, see below) show little tendency to bleed from a chloroform to an aqueous phase when employed in metal ion solvent extraction experiments under the present conditions. The tendency for cyclam to be lost from the organic phase, both as its protonated form and/or as its charged metal complex, inhibited us from proceeding with solvent extraction experiments of the type described below for 1.

Two series of solvent extraction (chloroform/water) experiments involving 1 were undertaken during the present study. The first involved copper(II) and silver(I) (each with an initial aqueous phase concentration of 10⁻³ mol dm⁻³); for this series the metal ion concentration in the aqueous phase was monitored after extraction was carried out by atomic absorption spectroscopy. The second series involved only silver(I), with the silver ion concentration ([Ag⁺]_init = 10⁻⁴ mol dm⁻³) being monitored radiometrically.

The first series of extractions were performed in sealed glass vials containing a chloroform phase (5 cm³) and an aqueous phase (5 cm³). The latter consisted of a copper(II) or silver(I) solution of known concentration (1 × 10⁻³ mol dm⁻³) and was maintained at pH 5.8 for both the copper and the silver runs by the presence of the corresponding 2-(N-morpholino)ethanesulfonic acid (MES)/NaOH buffer (0.05 mol dm⁻³). The chloroform phase in each case contained 1 (10⁻³ mol dm⁻³) and a known concentration of palmitic acid for the copper extractions and palmitic, 4-tert-butylbenzoic acid or salicylic acid for the silver runs. A duplicate series of experiments were undertaken for the copper system in order to probe the influence (if any) of the buffer on the extraction behaviour; these were performed at a ratio of macrocycle to palmitic acid of 1 ∶ 4 and involved maintaining the pH at 5.8 by the careful addition of a dilute solution of NaOH or HNO₃via a micropipette, with interspersed shaking until the required pH was obtained (and maintained).

The pH of 5.8 was chosen following the results of a study of pH variation on the extraction efficiency of 1 for copper(II) in the presence of a four-fold ratio of palmitic acid. Below pH ∼5 insignificant extraction of copper was observed while above pH ∼6, hydrolysis begins to interfere with the extraction process.⁶ Further experimental details for the above (non-radiotracer) solvent extraction experiments are given as accompanying ESI.†

The second set of (‘radiotracer’) extraction experiments were undertaken in order to probe the nature of the extraction behaviour for silver(I) more thoroughly (through systematic variation of both the macrocycle and carboxylic acid concentrations). These were carried out at 25 ± 2 °C in polypropylene microcentrifuge tubes (2 cm³) with a phase ratio V_(aq) ∶ V_(org) of 1 ∶ 1 (0.5 cm³ each). Initial experiments involved a silver nitrate solution at 10⁻⁴ mol dm⁻³ containing the ^110mAg radiotracer maintained at a series of known pH values between ∼5.2 and ∼6.7 using MES/NaOH buffers. The organic (chloroform) phase contained 1 at 1 × 10⁻³ mol dm⁻³ and carboxylic (palmitic, 4-tert-butylbenzoic or salicylic) acid at 2 × 10⁻³ mol dm⁻³. From the corresponding percentage extraction vs. pH plot, a pH of 5.2 was chosen as suitable for performing the planned experiments. For this pH, extraction experiments were performed for carboxylic acid ∶ 1 ratios varying from zero to four and log D was plotted against log[L] (where L = 1) over the range 2.5 × 10⁻⁴–2.5 × 10⁻³ dm mol⁻³; D is the distribution ratio and is given by the ratio of the total metal in the organic phase to that in the aqueous phase for each experiment (for each carboxylic acid concentration employed). In the absence of carboxylic acid, provided a ‘simple’ equilibrium of the type

is involved, the slope of the log–log plot (s) gives the stoichiometry of the extracted species directly since:
and

log D_M = log K_Ex + n log[A⁻]_(aq) + s log[L]_(org)

Control experiments in which the aqueous phase contained silver at 1 × 10⁻⁴ mol dm⁻³ and the chloroform phase contained only the respective carboxylic acids at 1 × 10⁻² mol dm⁻³ were also undertaken and in no case was silver extraction observed. All radiotracer experiments for silver involved mixing the samples for 30 min using a rotating mechanical shaking device (preliminary experiments had confirmed that extraction equilibrium is attained within this period). At the end of this time the phases were separated and then centrifuged to ensure full phase separation. The metal concentration in each phase was then determined radiometrically (gamma-irradiation from the added ^110mAg isotope) using a NaI(Tl) Scintillation Counter (Cobra II, Canberra-Packard).

Results and discussion

Formation of ligand adducts in CDCl₃

The use of NMR titrations to investigate host–guest formation between an extended range of amine-containing ligands and organic acids in solution has been reported previously.^2–4 An important result was the observation that the maximum stoichiometry of the corresponding assemblies in CDCl₃ corresponds to the number of amine sites in the host ligand that have (log) protonation constants equal to or greater than 6–7 in aqueous or aqueous/methanol media. As an example, host–guest formation between 1,4,8,11-tetraazacyclotetradecane (cyclam) and 4-tert-butylbenzoic acid was investigated using both ¹H and ¹³C NMR titrations (along with calorimetry, X-ray diffraction, neutron diffraction and semi-empirical MO calculations).² For this system, the induced shifts in both sets of NMR spectra exhibited very sharp end points corresponding to 1 ∶ 1 and 1 ∶ 2 (cyclam ∶ 4-tert-butylbenzoic acid) complex formation in CDCl₃ and a similar result was also obtained from a parallel calorimetric titration in chloroform.

In view of the above it was anticipated that parallel experiments for the tetra-N-benzylated derivative 1, which has no ring amine protons available for hydrogen bond formation, might result in less stable host–guest formation. Accordingly, step-wise addition of palmitic, 4-tert-butylbenzoic or benzoic acid to 1 under similar conditions to those used for the cyclam study yielded 1 ∶ 1 and 1 ∶ 2 inflections in each of the ¹H and ¹³C NMR spectra which were notably less well defined than for the corresponding spectra in the previous (cyclam) study; the titration plot for addition of benzoic acid to 1 in CDCl₃ is shown in Figs. 1 and 2. This result confirms that the presence of N-benzyl substituents in 1 does lead to the formation of less stable adduct species relative to cyclam. As anticipated, this seems likely to reflect both lower basicities of the amine donors and the fact that there are now four less hydrogens available for hydrogen bonding. In addition, steric interactions between the bulky benzyl groups and the carboxylic acid guests also expected.

Fig. 1 Induced shift of the benzyl methylene signal in the ¹H NMR spectrum of 1 on titration with benzoic acid in CDCl₃.

Fig. 2 Induced shift of the signal for the ring methylene (NCH₂CH₂N) groups in the ¹³C NMR spectrum of 1 on titration with palmitic acid in CDCl₃.

Despite the observed weak binding, the evidence for 1 ∶ 1 and 1 ∶ 2 host–guest formation (Figs. 1 and 2) is in accordance with two of the four (log) protonation constants of 1 being higher than 6–7, even after steric factors and the slight electron-withdrawing nature of the N-benzyl groups in 1 are taken into account. Unfortunately low solubility of 1 in a suitable solvent prevented determination of its protonation constants by potentiometric (pH) titration as part of the present study; however, a calculated value for the highest log protonation constant is 8.8 ± 0.7.¹⁰ For comparison, the reported log protonation constants for cyclam are 11.29, 10.19, 1.61 and 1.91 at I = 0.1; KCl, 25 °C.¹¹

It proved possible to analyse the plot of the induced chemical shift of the benzyl methylene signal from the ¹³C NMR titration of 1 with palmitic acid to obtain conditional step-wise association constants for the corresponding 1 ∶ 1 and 1 ∶ 2 complexes. The determined log K values for the 1 ∶ 1 and 1 ∶ 2 species are both quite low at 2.1 ± 0.1 and 0.3 ± 0.1, respectively. For comparison, a parallel titration was also performed in which benzoic acid was substituted for palmitic acid. In this case the calculated (log) step-wise constants were again similar, within experimental error, at 2.1 ± 0.1 and 0.3 ± 0.1.

The reported protonation constants for the tetra-N-methyl derivative of cyclam are (log) 9.34, 8.99, 2,59 and 2.25 (I = 0.1; NaNO₃; 25 °C)].¹² In a parallel study to the above, this tetramethyl derivative was also shown to give similar (weak) 1 ∶ 1 and 1 ∶ 2 host–guest formation on titration with benzoic acid; the step-wise (log) constants obtained being 2.5 ± 0.1 and 1.6 ± 0.1. Clearly, collectively these studies confirm that relatively weak association occurs when both tertiary amine macrocycles, 1 and the tetramethyl-substituted analogue, interact with palmitic or benzoic acid under the conditions outlined in the Experimental.

In an extension of these studies, an investigation of the interaction of difunctional salicylic acid and 1 was undertaken in order to investigate whether the presence of the phenol hydroxy group might influence host–guest formation relative to the behaviour observed for benzoic acid. A ¹H NMR titration of 1 (in CDCl₃) with salicylic acid yielded a titration plot (Fig. 3) which showed a single 1 ∶ 2 inflection which was significantly sharper than any of the inflections found for the above tertiary amine macrocyclic systems, in accord with the presence of enhanced host–guest binding in the latter case.

Fig. 3 Induced shift of the benzyl methylene signal in the ¹H NMR spectrum of 1 on titration with salicylic acid in CDCl₃.

Solvent extraction experiments

The results of an investigation of the effect of ligand assembly on the solvent extraction of copper(II) and silver(I) are now reported.

Copper(II). The effect of incremental addition of palmitic acid on the degree of copper(II) extraction by 1 is shown in Fig. 4; the pH for each palmitic acid ∶ macrocycle ratio investigated was either adjusted manually to 5.8 or maintained at this pH with the aid of buffer (see Experimental section). Both procedures resulted in similar extraction plots within experimental error, indicating that the presence of buffer does significantly influence the extraction behaviour. Clearly, from the results presented in Fig. 4, the percentage of copper(II) extraction increases on addition of palmitic acid to the organic phase (for a constant concentration of 1), confirming that the macrocycle/palmitic acid combination leads to synergistic extraction of copper(II). Maximum extraction appears to have been reached at approximately a 4 ∶ 1 ratio of palmitic acid ∶ macrocycle 1.

Fig. 4 Extraction studies using water/chloroform. Aqueous phase contained copper(II) at 10⁻³ mol dm⁻³. The organic phase contained 1 at 10⁻³ M and varying ratios of palmitic acid, shaken at 130 cycles min⁻¹ for 24 h in a temperature-controlled water-bath at 25 °C. The pH was maintained at 5.8 by two methods: manual adjustment with HNO₃/NaOH and the use of MES/NaOH buffer.

While a 1 ∶ 2 (macrocycle ∶ palmitic acid) ratio might be considered to be ‘ideal’ for achieving maximum extraction of this divalent metal ion, the observed present higher ratio accords with the results from our previous studies⁴ under similar conditions where higher ratios than 1 ∶ 2 were also necessary to maximise the extraction of copper(II). Especially for the present system where acid/macrocycle host–guest formation has been demonstrated to be weak, such behaviour appears likely to reflect that the excess acid acts to drive the equilibrium towards the formation of the ‘ideal’ 1 ∶ 2 assembly; alternatively, it has been proposed that neutral carboxylic acid molecules may also contribute to extraction by additional solvation of the extracted complex.¹³ In the latter case lipophilicity of the system has been postulated to be enhanced through outer-sphere complexation by acid molecules.

Silver(I). In previous studies by us⁵ and others¹⁴ it has been demonstrated that 1 shows enhanced metal ion discrimination towards silver(I) relative to a range of other metal ions. In part based on this documented affinity of 1 for silver and the observation that it forms a discrete 1 ∶ 1 complex with this macrocycle in solution,⁵ we carried out extraction experiments involving this monovalent ion and 1 in the presence of palmitic, 4-tert-butylbenzoic or salicylic acids.

As illustrated in Fig. 5, the control experiment involving palmitic acid in the absence of 1 gave 16% extraction of silver whereas the other two acids yielded no extraction under the present conditions. Clearly, at best, evidence for synergism is marginal in the case of the 4-tert-butylbenzoic acid system as well as for the palmitic acid system (when the degree of extraction by both the latter acid and 1 alone are considered). However, in contrast, clear evidence for significant synergism is apparent for the salicylic acid system, paralleling the enhanced host–guest behaviour of this acid with 1 as discussed above.

Fig. 5 Extraction of silver(I) employing 1 and t-butylbenzoic, palmitic and salicylic acid. Aqueous phase: pH 5.8 (MES/NaOH buffer), [AgNO₃] = 10⁻³ mol dm⁻³; organic phase: [1] = 10⁻³ mol dm⁻³, [acid] = 5 × 10⁻⁴–4 × 10⁻³ mol dm⁻³ in chloroform.

Radiotracer experiments

Owing to the nature of the detection employed, ‘micro extraction’ experiments were performed using a lower metal ion concentration (1 × 10⁻⁴ mol dm⁻³); namely, an order of magnitude less than employed for the above (non-tracer) studies. Furthermore, the equilibrium pH was lowered by 0.6 pH units (from 5.8 to 5.2) for the latter experiments. In the initial series of extractions discussed above, the extraction of silver (with [Ag⁺] = 1 × 10⁻³ mol dm⁻³ and [1] = 1 × 10⁻³ mol dm⁻³, in the absence of carboxylic acid) was ∼40%, see Fig. 5. In the present series 1 alone at a similar concentration was observed to again extract silver but, as expected, to a much lesser degree (15%, see Fig. 6) when the concentration of the macrocycle was 1 × 10⁻³ mol dm⁻³ (rising to 35% when the macrocycle concentration was 2.5 × 10⁻³ mol dm⁻³). However, when 1 was employed in conjunction with palmitic, 4-tert-butylbenzoic or salicylic acid, in each case silver extraction was significantly higher than for 1 alone for each concentration of 1 employed (Fig. 6). No extraction of silver was observed for these acids alone (in the case of palmitic acid this is undoubtedly a reflection of the lower pH (5.2) employed for these extractions compared with the pH of 5.8 used for the non-tracer experiments where 16% extraction was observed). Collectively, these results clearly confirm that synergistic extraction by these acid/macrocycle combinations occurs under the conditions employed.

Fig. 6 Extraction of silver(I) employing 1 and palmitic (a), 4-tert-butylbenzoic (b) and salicylic (c) acid. Aqueous phase: pH 5.2 (MES/NaOH buffer), [AgNO₃] = 10⁻⁴ mol dm⁻³; organic phase: [1] = 2.5 × 10⁻⁴–2.5 × 10⁻³ mol dm⁻³, [acid] = 1.25 × 10⁻⁴–1 × 10⁻² mol dm⁻³ in chloroform.

An initial log–log plot of the distribution ratio for silver(I) against macrocycle concentration, in the absence of carboxylic acid and at constant ionic strength [I = 0.01, NaNO₃], was linear with a slope that was close to one. This result serves to confirm that 1 ∶ 1 (macrocycle ∶ silver ion) complexation occurs under the present conditions.

Clearly relative to the above situation, the presence of carboxylic acid will alter the equilibria involved in both complex formation and extraction. For example, the latter will include an additional equilibrium in the organic phase between the respective free carboxylic acids and the corresponding host–guest ligand assembly. For these latter runs, it is difficult to assign significance to the slopes of the respective log–log plots. As a consequence of this added complexity, constant ionic strength was not maintained for these runs in which the effect of added acid on the extraction behaviour was investigated. Nevertheless, all the slopes of the log–log plots are observed to be similar, ranging from nearly 1 for the macrocycle alone to 1.3 for the mixture (Fig. 7).

As discussed earlier, ¹H NMR studies indicated that weak 1 ∶ 1 and 1 ∶ 2 adducts are formed by 1 in CDCl₃ for the systems incorporating palmitic or 4-tert-butylbenzoic acid, with a stronger interaction (1 ∶ 2) being evident for the assembly involving salicylic acid. In parallel to this stronger 1 ∶ 2 host–guest formation in this latter case, silver(I) extraction is significantly higher for the system containing this acid for each concentration of macrocycle employed compared to the corresponding systems incorporating 4-tert-butylbenzoic or palmitic acids (Fig. 6).

The following conclusions can be drawn from the above behaviour. First, the interaction of silver(I) with 1 is 1 ∶ 1 in the absence of added carboxylic acid. Secondly, each of the plots in Fig. 7 lies above that for the system for which no carboxylic acid is present, clearly demonstrating the presence of synergistic extraction for each acid ∶ macrocycle ratio employed. Thirdly, the individual plots for the 1 ∶ 1 to 4 ∶ 1 ratios of acid to macrocycle each fall in close proximity for each system; this strongly suggests that maximum extraction had been reached by an acid to macrocycle ratio that approximates 1 ∶ 1. Such behaviour is in accord with the extraction process involving a 1 ∶ 1 ∶ 1 (silver ∶ macrocycle ∶ carboxylate) species in each case and this corresponds to the simplest stoichiometry for formation of a neutral complex with silver(I).

Fig. 7 Dependence of log D_Ag on concentration of 1 at different acid ∶ macrocycle ratios for 0–4. Aqueous phase: pH 5.2 (MES/NaOH buffer), [AgNO₃] = 10⁻⁴ mol dm⁻³; organic phase: [1] = 2.5 × 10⁻⁴–2.5 × 10⁻³ mol dm⁻³, [palmitic, 4-tert-butylbenzoic and salicylic acid] = 2.5 × 10⁻³–2 × 10⁻² mol dm⁻³ in chloroform.

Concluding remarks

In the present study a range of solvent extraction experiments (water/chloroform) involving ligand assemblies between 1 and selected carboxylic acids in the respective organic phases and copper(II) or silver(I) have been investigated. In most instances metal ion extraction was associated with synergism relative to extraction by macrocycle or carboxylic acid alone. While the situation is not necessarily straight forward due to the likely presence of back equilibria involving the ligand assemblies, the observed extraction behaviour is consistent with the operation of an assembly effect. The origins of the latter are postulated to include a favourable entropic term associated with the involvement of a (partially) assembled ligand package, as well as a variable contribution reflecting overall lipophilicity considerations.

As discussed previously, the assembly concept, which spans the areas of supramolecular host–guest chemistry and classical metal coordination chemistry, has implications for the design of new reagents for use in solvent extraction together with the prospect of rationalising a selection of the many examples of ‘synergism’ reported in the solvent extraction literature.

Acknowledgements

We thank the Australian Research Council, and the Deutsche Forschungsgemeinschaft for funding to support the collaboration between the research groups involved in this work. We acknowledge Professor P. A. Tasker, University of Edinburgh, for many helpful discussions, Assoc. Professor I. A. Atkinson of James Cook University for the determination of the conditional stability constants and Ms. U. Stöckgen of TU Dresden for assistance during the radiotracer studies.

References

1. (a) B. A. Moyer, in Comprehensive Supramolecular Chemistry, ed. J.-M. Lehn, J. L. Atwood, J. E. O. Davies, D. D. McNichol and F. Vögtle, Pergamon Press, Oxford, 1996, vol. 1, pp. 377–416; (b) M. K. Beklemishev, S. G. Dmitrienko, N. V. Isakova, in Macrocyclic Compounds in Analytical Chemistry, ed. Yu. A. Zolotov, Wiley, New York, 1997, pp. 63–208; (c) A. H. Bond, M. L. Dietz and R. Chiarizia, Ind. Eng. Chem. Res., 2000, 39, 3442.
2. K. R. Adam, M. Antolovich, I. M. Atkinson, A. J. Leong, L. F. Lindoy, B. J. McCool, R. L. Davis, C. H. L. Kennard and P. A. Tasker, J. Chem. Soc., Chem. Commun., 1994, 1539.
3. K. R. Adam, I. M. Atkinson, S. Farquhar, A. J. Leong, L. F. Lindoy, M. S. Mahinay, P. A. Tasker and D. Thorp, Pure Appl. Chem., 1998, 70, 2345.
4. K. A. Byriel, V. Gasperov, K. Gloe, C. H. L. Kennard, A. J. Leong, L. F. Lindoy, M. S. Mahinay, H. T. Pham, P.A. Tasker and D. Thorp, Dalton Trans., 2003, 3034.
5. Y. Dong, S. Farquhar, K. Gloe, L. F. Lindoy, B. R. Rumbel, P. Turner and K. Wichmann, Dalton Trans., 2003, 1558.
6. Y. Dong, G. A. Lawrance, L. F. Lindoy and P. Turner, Dalton Trans., 2003, 1567.
7. L. F. Lindoy, M. S. Mahinay, B. W. Skelton and A. H. White, J. Coord. Chem., 2003, 56, 1203.
8. Y. Dong and L. F. Lindoy, Aust. J. Chem., 2001, 54, 291.
9. M. J. Hynes, J. Chem. Soc., Dalton Trans., 1993, 311.
10. Calculated using Advanced Chemistry Development (ACD) software, Solaris V4.76 (194-2004, ADC).
11. R. D. Hancock, R. J. Motekaitis, J. Mashishi, I. Cukrowski, J. H. Reibenspies and A. E. Martell, J. Chem. Soc., Perkin Trans. 2, 1996, 1925.
12. M. Micheloni, A. Sabatini and P. Paoletti, J. Chem. Soc., Perkin Trans. 2, 1978, 828.
13. H. Yamada and M. Tanaka, Adv. Inorg. Chem. Radiochem., 1985, 29, 143.
14. H. Tsukube, K. Tagaki, T. Higashiyama, T. Iwachido and N. Hayama, J. Chem. Soc., Perkin Trans. 1, 1986, 1033.

---

Footnotes

† Electronic supplementary information (ESI) available: Further experimental details for the (non-tracer) solvent extraction studies. See http://www.rsc.org/suppdata/dt/b4/b412255k/

‡ The situation may be broadly compared with the observed enhancement of coordination when pendant arms incorporating additional donors are appended covalently to a macrocyclic core. In the present situation the additional donors are associated with the macrocyclic component, not by covalent bonding but in a host–guest manner via hydrogen bonding.

---