J. CHEM. SOC. FARADAY TRANS., 1994, 90(11), 1507-1511 Host-Guest Complexes of Cucurbituril with the 4-Methylbenzylammonium Ion, Alkali-metal Cations and NH4+ Rudiger Hoffmann, Wilhelm Knoche" and Christian Fenn Fakultat fur Chemie , Universitat Biele feld, 0-3350I Biele feld, Germany Hans-Jurgen Buschmann Deutsches Textilforschungszentrum Nord-West, 0-47798Krefeld, Germany Kinetic measurements indicate the existence of two different complexes of cucurbituril with 4-methylbenzyl- ammonium ions: (i) an association complex, where the ammonium group binds to one of the polar portals of cucurbituril and the hydrophobic part extends into the solvent, and (ii) an inclusion complex, where the hydro- phobic part extends into the cavity of cucurbituril. This complex is used as an indicator for the binding of simple monovalent cations to cucurbituril.Stability constants are reported for the 1 : 1 complexes of cucurbituril with different cations. In 1905, Behrend et al. reported on the isolation of a com- pound from the condensation of urea with glyoxal and formaldehyde1 which was later shown by Mock to be a macropolycyclic compound whose molecular structure resembles a pumpkin. Thus, Mock gave the compound the trivial name cucurbituril and then went on to study its recep- tor properties.' Cucurbituril (Fig. 1) is a nonadecacyclic cage compound of a relatively rigid structure. The two carbonyl- fringed portals at the upper and lower side of the molecule have a diameter of 4 A. The internal cavity has a diameter of ca.5.5 A, and the distance of the portals is 6 A. The interior of the molecule represents a hydrophobic region, whereas the two portals are hydrophilic. Consequently the hydrophobic organic moiety of organic ammonium ions extends into the interior, and the ionic part coordinates to one of the planes spanned through the negatively polarized carbonyl groups. Dissociation constants of complexes of cucurbituril with various organic ammonium ions have been determined from thermodynamic and kinetic mea~urements,~.~ and catalytic activity in 1,3-dipolar cycloadditions has been reported. Recently the reaction of cucurbituril with alkaline cations was studied, and a 1 : 2 complex was observed.6 In this contribution we report upon a spectrophotometric and kinetic investigation of the binding of the 4-methylbenzylammonium ion (MBAH+) to cucurbituril.When alkali-metal or ammonium salts are added to the solu- tions, we observe a competitive binding of MBAH' and the alkaline cations, which influences strongly equilibrium and '0 0 0 Fig. 1 Molecular structure of cucurbituril kinetics of the reaction. A thorough analysis of all experimen- tal results leads to a detailed understanding of the complex formation. Cucurbituril is soluble only in very acidic solvents and most studies were performed in 48 : 52 (v/v) formic acid- water mixtures. Therefore we used the same solvent in our experiments. Experimental Cucurbituril was prepared as described by Behrend et al.' and Mock et al.' 4-Methylbenzylamine was converted to the hydrochloride by dissolving the amine in ether and intro- ducing HC1 gas into the solution.The precipitate was washed with ether, dried and used without further purification, All other chemicals are commercially available (grade p.a.). All experiments were performed in 48 : 52 (v/v) mixtures of formic acid and water. In this solvent the maximum solubility of cucurbituril is 3.0 x mol dm-3. Stock solutions were prepared from cucurbituril, 4-methylbenzylamine hydro-chloride (MBAHCl) and the alkali-metal chlorides. Stock solutions of the complex were obtained by dissolving 7.0 x lop3mol dm-3 cucurbituril plus 8.0 x mol dm-3 MBAHCl. In all experiments the temperature was controlled within & 0.2 "C, and stock solutions were thermostatted before use.For recording the UV spectra and the kinetic measure- ments a Kontron Uvikon 860 spectrophotometer (double- beam operation with the same solvent in both cuvettes) was used. Absorbances are measured with an error of k0.002. Since the reactions are slow, the kinetics of the complex for- mation were studied by mixing stock solutions of cucurbituril and MBAHCl in a spectrometer cuvette. The reverse reaction was studied by mixing stock solutions of the complex and alkali-metal chlorides. The progress of the reaction was mon- itored at the wavelength 1 = 271.5 nm. The results were digi- tally averaged over at least three measurements. In most experiments one of the reactants is in large excess, and the reaction proceeds according to pseudo-first-order kinetics.In those cases where the reaction proceeds under second-order conditions, the change of absorbance was fitted to the corresponding equation a exp(-t/z) (1)A=A,+ 1 + b exp(-t/7) A, is the absorbance at equilibrium (t p 7).The time constant z of the reaction is called relaxation time. If the reaction pro- ceeds according to pseudo-first-order kinetics, the term 'b exp(-t/z)' in eqn. (1) may be neglected, a represents the amplitude of the relaxation effect and l/z = /'cobs. That means, z is obtained in different ways from the change of absorbance, depending on the order of the reaction studied. Close to equi- librium all reactions proceed according to first-order kinetics, and therefore z has the same meaning (and the same depen- dence on concentration) for first- and second-order reactions.A,, z and the amplitude a/(l + b) of the relaxation effect will be discussed in this contribution. Results Fig. 2 shows spectra of solutions of cucurbituril and 4- methylbenzylamine hydrochloride in 48 : 52 formic acid- water mixtures, where the same solvent is used as reference. The measurements are restricted to the range R > 250 nm, since below this wavelength the solvent absorbs too strongly. For cucurbituril a broad absorption band is observed at R < 300 nm [spectrum 01. This absorbance does not appear for other compounds with the same -N-CO-N-structural unit (e.g.tetramethylurea or 1,3-dimethylimidazolidin-2-one),and we cannot attribute it to a certain electronic transition. The 4-methylbenzylammonium ion (MBAH') shows the weakly structured a-band of the aromatic ring [spectrum (a)]. When cucurbituril is added to the solution, this absorbance increases and the band is better resolved. At high concentrations of cucurbituril we observe the well resolved spectrum of a complex with a narrow peak t 250 290 qnrn Fig. 2 Spectra of solutions containing cucurbituril and Cmethyl- benzylamine. Lo = 8.0 x mol dm-3 for (a)-@) and C, = 0 (a), 1.6 x (b),9.7x (c),7.0 x (4, 1.6 x mol dm-3 (e).cf)is the spectrum of 7.0 x mol dm-3 cucurbituril. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 at R = 271.5 nm.The absorbances of 4-methylbenzylamine and the complex differ most strongly at this wavelength, which therefore is used to determine the equilibrium constant of the complex. For the evaluation we use the abbreviations C and L+ for cucurbituril and the ligand MBAH', respectively. Assuming the simple complexation reaction k C+L+ LCL+ (2)k' the concentrations at equilibrium are determined by the total concentrations Co and Lo and by the stability constant K: [C] + [CL'] = c, (3) [L'] + [CL'] = Lo (4) The absorbance of the solution (compensating the absorb- ance of the solvent) is given by A7= ECCCI + EJ-L+] + ECL[CL+] K and cCL are evaluated by fitting these equations to the experimentally obtained absorbances. In Fig. 3 the contribution of the free ligand L+ and the complex CL+ to the absorbance is plotted against the logarithm of the concentration of cucurbituril at constant concentration of 4-methylbenzylamine, and the expected sigmoidal curve is obtained.The spectrum of the complex (insert in Fig. 3) is calculated by subtracting the absorbance due to free cucurbituril and free ligand from the observed spectrum. At two different concentrations of ligand the same value of K is obtained, guaranteeing the 1 :1 stoichiometry of the complex. Finally K was determined at different tem- peratures. All results are summarized in Table 1. 250 2700.2 l-------I -4 -3 -2 log(Co/mol dm-3) Fig. 3 Absorbance of L+ and CL" us. the logarithm of the total concentration of cucurbituril.The inserts show the spectra of the L+ (left-hand side) and CL+ (right-hand side). [C] is calculated from eqn. (3)-(5). Table 1 Constants evaluated for the binding of MBAH+ to cucurbituril obtained from spectrophotometric measurements at different tem- peratures with Lo = 8.0 x mol dm-' -~ EC EL TpC /dm3iol-' /dm3 mol-' cm /dm3 mol-' cm-' ECL K. /mol dm-3 /dm3 mol-' K,/dm3 mol-' /dm3 mo";' cm-' E, /dm3 mol-' cm-' k, s-l ki /W3s-' 25 1200 21.4 148 400 25' 1400 21.4 148 380 15@ 1150b 41@ 169 l.lb 1.6* 35 1700 21.3 148 355 230 1490 385 169 2.2 3.3 45 1500 21.4 148 error +200 *0.2 335 310 & 20 & 50 1160 f200 380*20 169 f1 3.2 +20% 8.6 *20% a Lo = 1.6 x 10.' rnol dm-'. Evaluated for both 8 x lO-*and 16 x lo-' rnol dm-3.J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 In a previous communication6 it was shown that two alkali-metal cations can bind to a cucurbituril molecule. In order to study this complexation we added alkali-metal chlo- rides and ammonium chloride to solutions of cucurbituril, but no change in absorbance was observed. This means that the complexes formed have the same absorbance as the free cucurbituril molecule. Therefore we used reaction (2) as indi- cator for the formation of an alkali-metal complex, i.e. we added the salts to solutions containing both cucurbituril and 4-methylbenzylamine. As an example Fig. 4 shows the titra- tion curve for NaC1. Now we have competing equilibria, which are described by the reaction scheme CL' + 2Me+=C + L+ + 2Me+ eCMef + L' + Me'eCMe,'' + L+ (7) where Me' stands for the monovalent cations.The absorb- ance of the solutions is given by A -= &C{[C] + [CMe'] + [CMe,"]} + E~[L+]+ cCL[CL+]a (8) since the absorption coefficients of the complexes are equal to that of cucurbituril, as stated above. All absorption coeffi- cients and the association constant of CL+ are already known, and by fitting this equation to the experimental data we may evaluate the stability constants of the two alkali- metal cation complexes: [CMe'] K1 = [C][Me'] (9) [CMe,, '1K2 = [CMe '3 [Me '3 Assuming that the concentration of the 1 :1 complex CMe+ is negligible ([CMe+] + C0),the best fit yields the dashed curve in Fig.4 with strong deviations from the experimental points. On the other hand, neglecting the 1 :2 complex CMe2,+ in the fitting procedure ([CMe,2' J 4 C,) leads to the dotted curve in Fig. 4, where the deviations are much Table 2 Association constants of complexes of cucurbituril with monovalent cations at 25 "C and ionic radii of Me' Kl/dm3 mol-' K,/dm3 mol-' rlA Li + 170 <3 0.76" Na' 1450 60 1.02' K' 560 <20 1.38' Rb+ NH4 + 410 170 <1 t3 1.52' 1.70' Ref. 7; ref. 8; ref. 9. smaller, but they are still systematic. The residual plots demonstrate that for the dashed and the dotted curve the deviations are in opposite directions. Therefore the fitting has to take into account both complexes. This leads to the full curve in Fig.4, which agrees well with the experimental points. The dashed curve shows the largest deviations, i.e. the 1 : 1 complex contributes more strongly to the absorbance than the 1 :2 complex, and correspondingly the association constant K2 has a large error. For the other cations an acceptable fit is obtained even if the 1 : 2 complex is neglected, and therefore for those ions only a lower limit of K, can be estimated (see Table 2). The kinetics of the complexation between cucurbituril and MBAH' were studied by mixing different solutions of cucurbituril with solutions containing 8.0 x rnol dm-3 MBAH', and the progress of the reaction was observed at 2 = 271.5 nm. In all experiments the absorbance changed in a single relaxation effect according to a first- or a second-order reaction, i.e.the relaxation time is obtained by fitting eqn. (1) to the absorbance. For equilibrium (2) the relaxation time z is given by [C] and [L'] are the concentrations at equilibrium, which are calculated from eqn. (3)-(5). According to eqn. (11) the plot of 2-' us. ([C] + [L']) yields a straight line, with inter- cept k' and slope k'K. The straight line in Fig. 5 shows the best fit of the experimental results to eqn. (ll), taking into account that the ratio between slope and intercept has to be equal to K. The disagreement is obvious, and thus the reac- tion scheme has to be expanded. The stoichiometry of the complex is unambiguously determined by the spectrophoto- metric measurements at equilibrium, and thus the simplest possibility to expand equilibrium (2) is to assume two differ- ent 1 :1 complexes: an association complex CL,+, which is formed very quickly without a change in absorbance, and an k :-I t1 I-3 -2 -1 log( [NaCl],/mol dm-3) Fig.4 (a)Absorbance of solutions containing Lo = 8.0 x lop4mol dm-3 and C, = 7.0 x lop3 mol dm-3 us. the logarithm of the con- centration of NaCI. The curves are obtained by fitting eqn. (8) to the experimental points considering both complexes (full line) and with the assumption that complex CNa+ (dashed line) or CNa,'+ (dotted line) may be neglected. (b) Deviations between experimental points and the best fit assuming that complex CNa+ (0)or CNaZ2+ (A) may be neglected.01 I I 0 0.5 1.o ([C] + [L+])/l 0-2 mol dm-3 Fig. 5 t-l us. ([C] + [L']) at 25°C. The full curve is calculated with the constants given in Table 1. The dashed curve shows the best fit of eqn. (1 1) to the experimental points with the restriction : slope/ intercept = K. - 3. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 inclusion complex CLi*, the formation of which is observed in the rate-determining step. This is described in the equi- o.olllibrium C+L+ e,CL,+ CL, + for which the relaxation time is given by: Now the equilibrium is determined by two stability constants K and ECL relate to the corresponding constants of the associ- ation and inclusion complexes by the equations : K = K,+ Ki (16) E, K, + E~Ki ECL = K,+ Ki Furthermore, we may assume that the absorption coefficient of CL,' is equal to E~ + EL, since we do not observe a fast change in the absorbance when the two solutions are mixed.In eqn. (13) K, can be calculated from K and Ki, and thus the two constants kf and Ki have to be obtained by fitting the equation to the relaxation times. The curve in Fig. 5 shows that the experimental results are well described by eqn. (13). The intercept of the curve and the limiting slope for small concentrations yield ki and kfK,, respectively, and K, deter-mines the curvature. The kinetic measurements have been performed at different temperatures, and all constants evalu- ated are included in Table 1. When salts are added to the solutions of cucurbituril and 4-methylbenzylamine, the relaxation time changes in a char- acteristic way, as shown in Fig. 6 for KCl.To describe this behaviour, we have to include the complex formation of cucurbituril with Me+ ions, which leads to the scheme C + L+ + Me+ CL,+ + Me' The measurements are restricted to relatively low concentra- tions of KCl, where according to the values of K, the concen- tration of the 1 :2 complex CMe22+ may be neglected. Assuming again that the formation of the inclusion complex CLi+ is the rate-determining step of the overall reaction, scheme (18) leads to the very complicated equation Xi -[CL,'] *-[Me']] Xi I c +--I v)--. ye 0.005--I I I -3 -2 -1 log( [KCI],/mol dm-3) Fig.6 Reciprocal relaxation time of the reaction of cucurbituril with MBAH' in the presence of KC1 us. total concentration of KCI. C, = 7.0 x lo-' mol dm-3;Lo = 8.0 x mol dm-3; T = 25°C. for the relaxation time. In this expression xj, xk are defined as the difference between the actual concentrations and the con- centrations at equilibrium of speciesj, k. The ratios xj/xk are calculated from the stability constants. The rate constant kf has already been determined by eqn. (12). Thus k& is the only parameter, which can be varied to fit eqn. (19) to the experimental value of z. The best fit leads to the curve in Fig. 6, which describes the experiments well. The derivation of eqn. (19) and the analytical expressions for xj/xkas well as all experimental results are given in ref.10. Discussion The spectrophotometric titrations clearly indicate the exis- tence of a 1 : 1 complex of cucurbituril with MBAH'. According to the results summarized in Table 1 the stability constant of this complex does not depend on the temperature within experimental error. However, we observe a systematic change in its absorption coefficient cCL,which strongly hints at the existence of two isomeric structures of the complex. Two different complexes are also necessary to explain the concentration dependence of the relaxation time: one complex, which is in fast pre-equilibrium with cucurbituril and MBAH', and a second slow one, the formation of which can be followed spectrophotometrically. No fast change of absorbance is observed when solutions of cucurbituril and MBAH are mixed, and therefore the absorption coefficient + of the first complex has to be equal to the sum of the absorp- tion coefficients of cucurbituril and MBAH'.The formation of the second complex is connected to an increase of absorb- ance and leads to a well resolved spectrum of the chromo- phore of MBAH'. This indicates that in the first complex the environment of the chromophore is not changed compared to that of the free MBAH', whereas the chromophore is desol- vated in the second complex. These arguments lead to the identification of the first complex as an association complex, where the charged ammonium group of the ligand binds to the six carbonyl groups in one of the portals of cucurbituril and where the chromophoric aromatic group still extends into the solvent.The second complex is an inclusion complex, where again the ammonium group binds to six carbonyl groups but where the chromophore is desolvated and extends into the cavity of cucurbituril. This is shown schematically in Fig. 7, which includes three water molecules. According to X-ray crystallographic studies these molecules are located in the cavity of cucurbituril." Since the two complexes are of the same stoichiometry, the titration yields only the overall stability constant K, as defined in eqn. (16). The concentration dependence of the relaxation time in Fig. 5 is characterized by the limiting value J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 + CH3 Fig. 7 Formation of the association and the inclusion complex of cucurbituril and MBAH+ of l/z at low concentrations, the limiting slope and the curva- ture. The good agreement between the calculated curve and the experimental values confirms the reaction scheme pro- posed, since we fit only the two constants k' and Ki to those three features. (K, is determined by K and Ki.) The measure- ments are evaluated for different temperatures, and the absorption coefficients E~ of the inclusion complex are calcu- lated with the values of Ki and E, = cC + E~ as stated above. A further confirmation of the existence of two different com- plexes is that E~ does not depend on temperature, in contrast to cCL,which is evaluated assuming the single-step reaction, eqn.(2). The complex formation between cucurbituril and MBAH + is reduced, when an alkali-metal salt or NH4C1 is added to the solution. Obviously the added cations form complexes with cucurbituril, and thus they compete with MBAH' for the binding sites in the carbonyl planes. The absorbance of cucurbituril is not altered by the binding of the cations, and their binding constants have to be determined using MBAH' as an indicator. The results show that a 1 :1 complex is favoured. Owing to the repulsion of two equally charged ions through the nearly empty cavity of cucurbituril a 1 : 2 complex is formed only at very high concentrations of salt. The binding constants relate well to the crystal radii of the ions: Na' binds much stronger than Li', and for ions of larger size the complexation decreases monotonically (see Table 2).The proposed mechanism indicates that the simple cations bind to cucurbituril in the same way, as MBAH' is bound in the association complex. This is confirmed by the fact that the stability constant of the association complex of MBAH' is equal to that of the NH4+ complex. The reaction rate of the formation of the MBAH' inclu-sion complex is reduced when small amounts of salt are added. This dependence is expected, since MBAH+ can form an inclusion complex only with the free cucurbituril molecule. However, the reaction rate increases on further addition of salt, as shown in Fig. 6. Eqn. (19) allows for this increase by considering a pathway in scheme (18), where the ligand reacts with CMe' to form the inclusion complex.Thus also the dependence of the reaction rate on the salt concentration can well be described. Summarizing, it may be said that in this contribution results are presented which allowed us to determine stability constants and to deduce the mechanism for the formation of association and inclusion complexes of cucurbituril. These studies were performed in water-formic acid mixtures at a constant composition. For a more detailed discussion the specific solvation of cucurbituril in this solvent has to be con- sidered. Therefore the studies are extended to aqueous solu- tions of acid at relatively low concentration. The authors are indebted to the Fonds der Chemischen Industrie for the financial support of this work. References 1 R. Behrend, E. Meyer and F. Rusche, Liebigs Ann. Chem., 1905, 339, 1. 2 W. L. Mock, W. A. Freeman and N-Y. Shih, J. Am. Chem. SOC., 1981,103,7367. 3 W. L. Mock and N-Y. Shih, J. Org. Chem., 1986,51,4440. 4 W. L. Mock and N-Y. Shih, J. Am. Chem. SOC., 1989,111,2697. 5 W. L. Mock, T. A. Irra, J. P. Wepsiec and M. Adhya, J. Org. Chem., 1989,54,5302. 6 H-J. Buschmann, E. Cleve and E. Schollmeyer, Inorg. Chim. Acta, 1992,193, 9. 7 R. M. Izatt, J. S. Bradshaw, S. A. Nielsen, J. D. Lamb and J. J. Christensen, Chem. Rev., 1985,85,271. 8 R. D. Shannon and T. C. Prewitt, Acta Crystullogr., Sect. B, 1969, 25,925. 9 D. H. Aue, H. M. Webb and M. T. Bowers, J. Am. Chem. SOC., 1976,98,318. 10 R. Hoffmann, Ph.D. Thesis, Universitat Bielefeld, 1993. 11 W. A. Freeman, Acta Crystallogr., Sect. B, 1984,40,382. Paper 3/07417J; Received 16th December, 1993