Appropriate linking of Lewis acidic and Lewis basic components—in this case boronic acid and tertiary amine functionalities—allows the construction of a robust organometallic receptor capable of selectively binding hydrogen fluoride. This system can differentiate between exposure to HF and to related acids, such as HCl, by providing opposing electrochemical responses (see picture).

A systematic investigation of fluoride anion binding properties as a function of chelate backbone has been carried out for ferrocene functionalised boronic esters of the types FcB(OR)2 and fc[B(OR)2]2 [Fc = ferrocenyl = (eta5-C5H5)Fe(eta5-C5H4); fc = ferrocendiyl = Fe(eta5-C5H4)2]. Cyclic boronic esters containing a saturated five- or six-membered chelate ring are readily synthesized from ferrocene, and selectively bind fluoride via Lewis acid/base chemistry in chloroform solution. The resulting complexes are characterized by relatively weak fluoride binding (e.g.K = 35.8 +/- 9.8 M(-1) for FcBO2C2H2Ph2-S,S), and by cathodic shifts in the ferrocene oxidation potential that form the basis for electrochemical or colorimetric fluoride detection. The fluoride selectivity of these systems is attributed to relatively weak Lewis acidity, resulting in weak F- binding, and essentially no binding of potentially competitive anions. By contrast, more elaborate Lewis acid frameworks based on calix[4]arene (calixH4), such as (FcB)2calix or fcB2calix, do not survive intact exposure to standard fluoride sources (e.g. [nBu4N]F.xH2O solutions in chloroform or acetonitrile). Instead B-O bond cleavage occurs yielding the parent calixarene; the differences between alkoxo- and aryloxo-functionalised derivatives can be rationalised, at least in part, by consideration of the differences in electron donating capabilities of RO- (R = alkyl, aryl).

Synthetic approaches based on the direct borylation of ferrocene by BBr(3), followed by boryl substituent modification, or on the lithiation of ferrocene derivatives and subsequent quenching with the electrophile FBMes(2), have given access to a range of ferrocene derivatized Lewis acids with which to conduct a systematic study of fluoride and cyanide binding. In particular, the effects of borane electrophilicity, net charge, and ancillary ligand electronics/cooperativity on the binding affinities for these anions have been probed by a combination of NMR, IR, mass spectrometric, electrochemical, crystallographic, and UV-vis titration measurements. In this respect, modifications made at the para position of the boron-bound aromatic substituents exert a relatively minor influence on the binding constants for both fluoride and cyanide, as do the electronic properties of peripheral substituents at the 1'- position (even for cationic groups). By contrast, the influence of a CH(2)NMe(3)(+) substituent in the 2- position is found to be much more pronounced (by >3 orders of magnitude), reflecting, at least in part, the possibility in solution for an additional binding component utilizing the hydrogen bond donor capabilities of the methylene CH(2) group. While none of the systems examined in the current study display any great differentiation between the binding of F(-) and CN(-) (and indeed some, such as FcBMes(2), bind both anions with equal affinity within experimental error), much weaker boronic ester Lewis acids will bind fluoride (but give a negative response for cyanide). Thus, by the incorporation of an irreversible redox-matched organic dye, a two-component [BMes(2)/B(OR)(2)] dosimeter system can be developed capable of colorimetrically signaling the presence of fluoride and cyanide in organic solution by Boolean AND/NOT logic.

The thermodynamic factors underlying the use of ferroceneboronic esters as electrochemical or colorimetric fluoride ion sensors have been investigated through the synthesis of a range of systematically related derivatives differing in the number/nature of the boronic ester substituents and in the nature of ancillary ligands. Thus, if the shift in electrochemical potential associated with the conversion of one (or more) boronic ester group(s) to anionic boronate(s) on fluoride binding is sufficient to allow oxidation of the resulting host/guest complex by dioxygen, colorimetric sensing is possible. In practice, while monofunctional systems of the type CpFe[eta(5)-C(5)H(4)B(OR)(2)] offer selectivity in fluoride binding, electrochemical shifts in chloroform solution are insufficient to allow for a colorimetric response. Two chemical modification strategies have been shown to be successful in realizing a colorimetric sensor: (i) the use of the more strongly electron-donating Cp(*) ancillary ligand (which shifts the oxidation potential of both the free receptor and the resulting fluoride adduct cathodically by ca. -400 mV) and (ii) receptors featuring two or more binding sites and consequently a larger fluoride-induced electrochemical shift. Thus, in the case of [eta(5)-C(5)H(4)B(OR)(2)](2)Fe [(OR)(2) = OC(H)PhC(H)PhO, 2(s)], the binding of 2 equiv of fluoride gives an electrochemical shift (in chloroform) of -960 mV (cf. -530 mV for the corresponding monofunctional analogue, 1(s)). Related tris- and tetrakis-functionalized systems are also shown to be oxidized as the bis(fluoride) adducts, presumably because of fast oxidation kinetics, relative to the rate of the (electrostatically unfavorable) binding of a third equivalent of fluoride. Furthermore, the rate of sensor response (as measured by UV/vis spectroscopy) is found to be strongly enhanced by the presence of pendant (uncomplexed) three-coordinate boronic ester functions (e.g., a rate enhancement of 1-2 orders of magnitude for 3(s)/4(s) with respect to 2(s)) and/or delocalized aromatic substituents.