ANALYST FULL PAPER THE www.rsc.org/analyst Assessment of the cross-reactivity and binding sites characterisation of a propazine-imprinted polymer using the Langmuir-Freundlich isotherm Esther Turiel†, a Concepcion Perez-Conde a and Antonio Martin-Esteban†* b a Departamento de Quimica Analitica, Facultad de Ciencias Quimicas, Universidad Complutense de Madrid, 28040 Madrid, Spain b Departamento de Quimica y Materiales, Universidad Europea de Madrid, Villaviciosa de Odon, 28670 Madrid, Spain. E-mail: antonio.martin-esteban@irmm.jrc.be; Fax: +32-14-571 787; Tel: +32-14-571 735 Received 30th October 2002, Accepted 16th January 2003 First published as an Advance Article on the web 22nd January 2003 In this paper, the Langmuir-Freundlich isotherm is used to model the interaction of several triazines (desethylatrazine, desisopropylatrazine, simazine, atrazine, propazine and prometryn) with a propazine-imprinted polymer and to explain the observed cross-reactivity. Different rebinding experiments (each herbicide alone or all together in a mixture) were carried out and the experimental binding isotherms were fitted to the Langmuir-Freundlich isotherm. The fitting coefficients obtained (total number of binding sites, mean binding affinity and heterogeneity index) allowed the description of the kind of binding sites present in the imprinted polymer under study. It was concluded that the recognition mechanism was mainly governed by the molecular size although slight differences in the molecular structure may also play an important role. The obtained results suggest that the use of this new methodology can open new pathways for understanding how molecular recognition in imprinted polymers takes place. Introduction During the last years, molecular imprinting technology has received much attention from the scientific community. This technique allows the preparation of tailor-made polymers, namely molecularly imprinted polymers (MIPs), in an easy and rapid manner. Briefly, a pre-polymerisation mixture is prepared with a template molecule and a suitable monomer (in the presence of an appropriate solvent) to form a covalent or non- covalent complex. Subsequently, polymerisation is carried out in the presence of an excess of cross-linker, and finally, the template is removed from the polymer leaving cavities complementary in shape and size to the template molecule. MIPs were introduced by Wulff 1 (covalent approach) and by Mosbach 2 (non-covalent approach) and have been employed in different fields, such as sensors, 3 sorbents, 4 and chromato- graphic stationary phases, 5 among others and applied to a wide array of template (analyte) molecules. Unfortunately, especially when the non-covalent approach is used, the pre-polymerisation step is a non-well defined process leading to the formation of complexes with different template- :monomer stoichiometry. 6,7 In addition, several authors have reported that the process of crushing and sieving the polymer after polymerisation can break the imprinted sites and, also, a substantial number of the cavities may shrink after the template is removed with polar organic solventes. 8 As a consequence of these effects, the obtained MIPs present a non-homogeneous binding site distribution, limiting their applicability range (i.e. broad peaks in chromatography, non-linear response in sensors, etc.) and their selectivity. In fact, MIPs have been compared to polyclonal antibodies owing to the fact that they have different binding sites. However, the observed cross-reactivity cannot be considered as a drawback but as an advantage for certain applications, especially in the development of solid-phase extraction processes for the simultaneous determination of structurally related compounds (i.e. pesticide families, parent compound and its degradation products, etc.). 9–12 On the other hand, although only in a qualitative manner, a detailed study of the observed cross-reactivity, through binding and capacity studies, may suggest the kind of binding sites present in imprinted polymers and help to understand how molecular recognition takes place. Several mathematical models have been used for explaining the heterogeneity in MIPs. The bi-Langmuir (using the limiting slopes analysis of curved Scatchard plots) 13,14 and Freundlich 15 isotherms are the most commonly used although both present some limitations to model different (covalent or non-covalent) MIPs at different concentration ranges. Recently, Umpleby et al. 16 demonstrated that the Langmuir-Freundlich (LF) isotherm is the more suitable model. The LF isotherm describes a relationship between the concentration of bound (B) and free (F) guests in heterogeneous systems with three different coefficients according to the following equation: (1) where N t is the total number of binding sites, a is related to the median binding affinity constant K 0 (K 0 = a 1/m ), and m is the heterogeneity index. For a homogeneous material, m is equal to 1 whereas when m is between 0 and 1 the material is heterogeneous. For homogeneous material (m = 1), the LF isotherm reduces to the Langmuir isotherm (eqn. (2)) and, on the other hand, as either F or a approaches 0, the LF isotherm reduces to the Freundlich isotherm (eqn. (3)). (2) † Present address: Analytical Chemistry Unit, Institute for Reference Materials and Measurements, Joint Research Centre, European Commis- sion, Retieseweg, 2440 Geel, Belgium. This journal is © The Royal Society of Chemistry 2003 DOI: 10.1039/b210712k Analyst, 2003, 128, 137–141 137