Kinetics and Catalysis, Vol. 43, No. 1, 2002, pp. 51–55. Translated from Kinetika i Kataliz, Vol. 43, No. 1, 2002, pp. 56–60. Original Russian Text Copyright © 2002 by Librovich, Kislina.
Catalytic Properties of Acid Solutions and the Structure of Acid–Base Complexes with Strong Hydrogen Bonds N. B. Librovich and I. S. Kislina Semenov Institute of Chemical Physics, Russian Academy of Sciences, Moscow, 117977 Russia Received January 4, 2001
Abstract—The catalytic action of acid solutions involves an equilibrium step in the formation of a reactive complex of reactant molecules with catalyst entities. The relative concentrations of these complexes are determined by using thermodynamic parameters (acidity functions and activities of catalyst components). Data on the structure of complexes with strong H-bonds can be obtained from the vibrational spectra of these species. These data are required for establishing the mechanisms of catalytic reactions and for predicting the catalytic properties of acid solutions containing these complexes.
INTRODUCTION As a rule, the acid-catalyzed reactions of organic compounds in solutions proceed via an equilibrium step in the formation of intermediate complexes of reactant molecules (usually, with the participation of a molecule with the highest basicity) with catalyst entities in an acid solution. Data on both the amount and the structure of these complexes are required for determining reaction mechanisms. Because the concentration of reactive transient species in the course of a reaction is usually extremely low, it cannot be measured directly. Therefore, indirect methods are used, which consist of establishing analogies between the equilibrium steps of acid–reactant and acid–indicator (a stable compound that does not undergo further transformations) interactions. This approach can be used under the following necessary conditions: (1) the reactant and the indicator should form complexes with the catalyst by the same mechanism; (2) the reactant and indicator molecules and complexes should interact with the medium in a similar manner; and (3) the mechanisms of ionization and interactions between reactant and indicator species should be retained when in the acid concentration and the ratio between solvent components change. These strict requirements result from the fact that, because of strong interparticle interactions in electrolyte solutions, the equilibrium formation constants of complexes or ions should be expressed in terms of the thermodynamic activities of species that participate in the equilibrium. Thus, if a complex K is formed by the interaction of an organic base B with an acid HA, the ratios fB fHA/fK between the activity coefficients of both the reactant and the indicator should be equally dependent on the composition of the medium and on temperature; this is provided by the above conditions.
A typical mistake made by many researchers is their certainty that they obtain concentration ratios between the unionized and protonated forms of an indicator in the measurements of acidity functions by the indicator method. Meanwhile, the concentration of only one (unionized) species is usually measured, whereas the concentration of the other species is calculated from the balance equation I = C B / ( C 0 – C B ), where C0 is the analytical (stoichiometric) concentration of the indicator in an acid solution, and CB is the experimental equilibrium concentration of an unionized (neutral) form of the indicator (it is usually found from UV spectra). Therefore, this approach a priori suggests that the difference (C0 – CB) is equal to the equilibrium concentration of a protonated form of the indicator. This is usually substantiated by the fact that either cryoscopic or conductometric data indicate the formation of only a protonated species in 100% sulfuric acid [1]. However, in more highly dilute solutions of sulfuric acid, as well as in the solutions of other strong mineral acids, ionization may consist of the formation of not only a protonated form of the indicator but also other charged and uncharged species. These species can be molecular complexes B · HA (where B is an indicator molecule) and ion pairs BH+ · A– (where BH+ is a protonated form of the indicator and A– is an acid anion). Proton transfer from an acid solution to an organic base molecule occurs in accordance with the classical Shatenshtein–Izmailov scheme [2] B + HA ⇔ B ⋅ HA ⇔ BH ⋅ A ⇔ BH + A . +
–
+
–
This interaction is referred to as complete if the above equilibrium system is completely shifted to the right; that is, the acid–base interaction results in the formation of ions. This is the case when the measured indicator ratio I corresponds to the process of protonation.
0023-1584/02/4301-0051$27.00 © 2002 MAIK “Nauka /Interperiodica”
52
LIBROVICH, KISLINA
Otherwise, real equilibria of the formation of either molecular complexes or ion pairs should be used as a basis. Hence, the catalytic activity of an acid solution, that is, its ability to convert the neutral form of substrate B into the reactive species B · HA or BH+ · A–, depends on the thermodynamic activity of the acid rather than the acidity function H0 [1]. Note that the indicator method for evaluating the catalytic activity of acid solutions simulates only a part of acid-catalyzed reactions, namely, the formation of an ionized substrate species. This is true only when the ionized species occurs in a thermodynamic equilibrium with parent substrate and acid molecules rather than in a transition state. In this case, the reaction rate depends on the transformation of the ionized species into reaction products. The considerations presented in this paper may be incorrect if the ionized species appears as an activated state rather than as a real molecular entity. It is important to keep in mind that the acidity function H0 is a thermodynamic quantity determined to a constant term. Thus, in each particular case, the acidity function is reduced to a standard state, which is usually taken to be infinitely dilute acid solution. A typical mistake made by a number of researchers is that they make a quantitative comparison between the acidity functions H0 of a mineral acid in different solvents without considering the difference of standard states. In actual practice, this is usually reduced to measuring indicator ratios I and calculating acidity functions by the equation H 0 = pK BH+ + log I. In doing this, the value of the indicator basicity constant K BH+ is taken from another acid–solvent system (usually, from the acid–water system). In this case, the difference in the standard states, which manifests itself in different values of p K BH+ , is ignored. Presently, we can state with certainly [3, 4] that, in many cases, the protonation of an indicator B is the reaction of substituting B for a fragment in the simplest stable proton solvate: B + ( C···H···D ) ⇔ ( B···H···D ) + C +
+
(I)
(where C and D are the components of a solvent in which an acid HA is dissolved) rather than the simple formation of the BH+ ion, in which the added proton occupies a position equivalent to the positions of other hydrogen atoms in an acceptor functional group (for example, in an amino group). In this case, the acidity function H0 is expressed by the equation H 0 = – log ( a H+ f B / f BH+ )
(where a H+ is the thermodynamic activity of the solvent component fB and f BH+ is the activity coefficient of the ion f BHD+ , which is a protonated species formed by a strong symmetrical hydrogen bond (B···H···D)+) rather than by the traditional equation H 0 = – log ( a H+ a D f B / f BHD+ ), where aD and f BHD+ are the activity coefficients of unionized and protonated forms of the indicator and a H+ is the thermodynamic activity of protons. This mechanism for the formation of the protonated form of an organic base explains a number of special features of the reactions of organic bases with acid solutions and, most importantly, determines the practical limits of using the Hammett acidity method to describe the protonation of weak organic bases. Nevertheless, it was found [5] that the step of protonation may consist of the formation of both BH+ ions and ions with a strong symmetrical hydrogen bond ((B···H···D)+ or (B···H···OH2)+ in an aqueous acid solution). Hammett’s hypothesis, which draws an analogy between the protonation steps of structurally different organic compounds (this analogy is expressed as equal fB/ f BH+ ratios between the activity coefficients of the unionized and protonated forms of structurally different compounds B), has not been proven. This resulted in numerous speculations on the applicability or inapplicability of the function H0 to describe the protonation of a particular class of compounds. Experimental and theoretical studies of ions formed by strong hydrogen bonds [6] demonstrated that the geometry and energy parameters of the central bridge A···H···B are almost independent of the structures of ligands A and B. This property is responsible for maintaining Hammett’s hypothesis [5]. At the same time, Hammett’s hypothesis cannot be maintained in general when BH+ ions are formed in the course of protonation. Thus, the use of H0 is here unjustified, and it is apparently limited by the description of protonation within a particular class of structurally similar compounds [5]. If a protonated species is the BH+ ion, in principle, its structure is independent of the proton-donor entity in an acid solution (an acid molecule, a proton solvate, or a complex that possesses acid properties). However, if protonation occurs as a substitution reaction (I), it is important to know which equilibrium species in the acid–solvent system participates in this process, because it affects the structure of the protonated form. Data on the mechanism of ionization and on the structure of resulting ions or complexes cannot be obtained by measuring the acidity function. These data can be obtained by studying the structure of molecular entities responsible for the catalytic activity of an acid solution and by comparing the results with the acidity of the KINETICS AND CATALYSIS
Vol. 43
No. 1
2002
CATALYTIC PROPERTIES OF ACID SOLUTIONS
solution. Important information can be derived from a comparison between the structure of catalytically active ions and complexes and the acidity of solutions in different solvents. However, as mentioned above, the difference between standard states should be taken into account for a quantitative comparison between the values of H0. Figure 1 demonstrates the acidity functions (as comparable normalized values) of methanesulfonic acid (MSA) in the following three solvents: water [7], ethyl acetate [8], and dimethylformamide [7]. Aqueous solutions exhibited the highest acidity, whereas dimethylformamide solutions exhibited the lowest acidity (more negative values of H0 correspond to a higher acidity of solution). It was found by spectroscopy that + H5 O 2 ions are formed in equilibrium in aqueous MSA solutions [9], molecular complexes of the 1 : 1 composition are formed in ethyl acetate solutions [10], and quasi-ion pairs (neutral complexes with incomplete proton transfer from the acid to the solvent molecule (B···H···A)) are formed in dimethylformamide solutions [11]. Thus, various situations can occur in acid solutions. Evidently, the structures and equilibrium concentrations of ionizing agents, which are responsible for the catalytic properties of the acid–solvent system, should be known in order to describe correctly the mechanism of ionization and to predict the acid properties. These data can be obtained by IR and Raman spectroscopy. If ions with strong symmetrical hydrogen bonds are formed in the course of the acid–base interaction, the IR spectrum exhibits intense continuous absorption as an intense broad band that extends from 500–700 to 3000–3500 cm–1. Individual and combination bands due to the vibrations of an ion or uncharged complex with a strong symmetrical hydrogen bond are present on the background of the continuous absorption. Detailed studies on the nature of continuous absorption resulted in the development of a theoretical model that described this phenomenon [6]. Until now, this has been the only model that adequately describes experimental data without the use of empirical parameters. According to Yukhnevich et al. [6], continuous absorption results from the strong interaction of the vibrations of a central bridge with lateral ligands. A variety of intense combination bands appears because of the strong electrooptical anharmonism of the vibrations of a central bridge in the above interaction. The superposition of these bands forms the total spectrum. Thus, an important conclusion can be drawn: the position of a proton in the central bridge and the structure of lateral ligands affect the shape and intensity of a continuous absorption band. The idea that this phenomenon could be used for the identification of ions and complexes with strong symmetrical hydrogen bonds in acid solutions has already arisen. Detailed spectroscopic studies demonstrated the productivity of this idea [12]. It was found that ions with a strong symmetrical H-bond and KINETICS AND CATALYSIS
Vol. 43
No. 1
2002
53
H0 2 [MSA], mol % 0
25
50
75
100
3 –2 2
–4 1 –6 –8
Fig. 1. Concentration dependence of the acidity function H0 of methanesulfonic acid (MSA) solutions in (1) water, (2) ethyl acetate, and (3) dimethylformamide at 25°C.
+
a positive charge (H5 O 2 , alcoholic proton disolvates, etc.) exhibit the most intense signals in IR spectra, molecular complexes exhibited signals of the lowest intensity, and BH+ · A– ion pairs exhibited no continuous absorption [12]. Quasi-ion pairs with incomplete proton transfer produce a continuous absorption band of medium intensity. In general, depending on the proton position in a central bridge, the intensity of continuous absorption is minimal if a proton occupies one of the extreme positions, the molecular complex B · HA or the ion pair BH+ · A–. The central position of a proton, for example, in the (H2O···H···OH2)+ ion, is responsible for the most intense absorption. The quasi-ion pairs (B···H···A) exhibit somewhat lower absorption. A consideration of the IR spectra and acidity function of the boron trifluoride–ethanol system readily illustrates the possibility of predicting the mechanism of the catalytic action of an acid solution. Data on the acidity of this system were obtained by Burya [13]; however, they remained uninterpreted until recent times. Figure 2 demonstrates the IR spectra of the BF3– C2H5OH system. Based on an analysis of changes in the IR spectra in the frequency ranges 700–1600 and 2800– 3700 cm–1 with changes in the composition of the boron trifluoride–ethanol system, Burya [13] found that the complex BF3 · 2C2H5OH is formed in an excess of ethanol. Complexes of composition 1 : 1 are formed when the ratio between the analytical concentrations of BF3 and C2H5OH is higher than 0.5. However, the nature of the complexes was not explained by Burya [13]. This can be done by invoking data on the relation between continuous absorption parameters and the structures of ions and complexes formed by strong H-bonds [12]. The spectra indicate that continuous absorption occurs
54
LIBROVICH, KISLINA
Transmittance, %
H0 2
(‡)
20 40 60 80
0
0.5
1.0 q
2 –2
(b)
20 40 60 80
–4 1
–6 20 40 60 80
(c)
20 40 60 80
(d)
20 40 60 80
(e)
100 700
–8
1100
1500
1900
Fig. 3. Acidity functions H0 of the solutions of (1) BF3 and (2) HCl in ethanol at (1) 25 and (2) 20°C plotted against q, where q is the ratio between the analytical concentrations of BF3 (HCl) and C2H5OH.
2900
3300
3700 ν, cm–1
Fig. 2. IR spectra of boron trifluoride solutions in ethanol at analytical concentration ratios between BF3 and C2H5OH equal to (a) 0, (b) 0.255, (c) 0.514, (d) 0.764, and (e) 1.02 at 20°C.
in the system, and its intensity is maximum at the ratio BF3 : C2H5OH = 1 : 2. Based on data presented in [12], we can suggest that, in an excess of ethanol, a boron trifluoride molecule forms a complex with two ethanol molecules, and this complex has the strong central symmetrical (or quasi-symmetrical) hydrogen bond F3B : O(R)···H···O(H)R. At BF3 : C2H5OH ratios lower than 0.5, the system primarily behaves as a Brønsted acid, and the above complex is a proton donor (an ionizing agent). Complexes of composition 1 : 1 are formed in more concentrated solutions of BF3, and the system may behave as a Lewis acid transferring a boron trifluoride molecule to an indicator or reagent. Thus, the mechanism of ionization is dramatically changed depending on the concentration ratio between the components. Indicator measurements cannot detect this difference because, as mentioned above, only a decrease
in the concentration of unionized species is followed in these measurements. Figure 3 demonstrates the concentration dependence of H0 in the BF3–C2H5OH [13] and HCl–C2H5OH [14] systems in comparable units. Both of the functions were measured from the same standard state of pure ethanol. It can be seen in Fig. 3 that the curves intersect at a BF3(HCl) : C2H5OH ratio of ~0.5. This is due to a change in the ionization mechanism of indicators in the BF3–C2H5OH system, namely, a change from Br∅nsted to Lewis acidity, whereas the HCl–C2H5OH system is a Brønsted acid over the entire range of concentrations. In conclusion, it is imperative to note that the classical Shatenshtein–Izmailov scheme of acid–base interactions should be supplemented with the step of formation of quasi-ion pairs as follows: B + HA ⇔ B ⋅ HA ⇔ B···H···A ⇔ BH ⋅ A ⇔ BH + A . +
–
+
–
Moreover, BH+ ions can form (B···H···AH)+ and (B···H···S)+ ions with strong symmetrical H-bonds by the interaction with solvent (S) or acid molecules. Under certain conditions, A– anions can also form (A···H···A)– ions. ACKNOWLEDGMENTS This work was supported by the Russian Foundation for Basic Research (project no. 00-03-32004). REFERENCES 1. Vinnik, M.I., Usp. Khim., 1966, vol. 35, no. 11, p. 1922. KINETICS AND CATALYSIS
Vol. 43
No. 1
2002
CATALYTIC PROPERTIES OF ACID SOLUTIONS 2. Izmailov, N.A., Elektrokhimiya rastvorov (The Electrochemistry of Solutions), Moscow: Khimiya, 1966. 3. Vinnik, M.I., Kislina, I.S., and Librovich, N.B., Dokl. Akad. Nauk SSSR, 1980, vol. 251, no. 1, p. 138. 4. Kislina, I.S., Sysoeva, S.G., and Librovich, N.B., Khim. Fiz., 1999, vol. 18, no. 2, p. 51. 5. Librovich, N.B., Khim. Fiz., 1992, vol. 11, no. 5, p. 627. 6. Yukhnevich, G.V., Tarakanova, E.G., Maiorov, V.D., and Librovich, N.B., Usp. Khim., 1995, vol. 64, no. 10, p. 963. 7. Kislina, I.S. and Sysoeva, S.G., Izv. Akad. Nauk, Ser. Khim., 1999, no. 10, p. 1940. 8. Kislina, I.S. and Sysoeva, S.G., Izv. Akad. Nauk, Ser. Khim., 2001, no. 6, p. 961.
KINETICS AND CATALYSIS
Vol. 43
No. 1
2002
55
9. Kirilova, A.P., Maiorov, V.D., Serebryanskaya, A.I., Librovich, N.B., and Gur’yanova, E.N., Izv. Akad. Nauk, Ser. Khim., 1985, no. 7, p. 1493. 10. Burdin, V.V., Maiorov, V.D., and Librovich, N.B., Izv. Akad. Nauk, Ser. Khim., 2000, no. 2, p. 292. 11. Burdin, V.V., Kislina, I.S., Maiorov, V.D., Sysoeva, S.G., and Librovich, N.B., Izv. Akad. Nauk, Ser. Khim., 1998, no. 12, p. 2484. 12. Librovich, N.B., Burdin, V.V., Maiorov, V.V., and Kislina, I.S., Khim. Fiz., 2000, vol. 19, no. 4, p. 41. 13. Burya, G.F., Cand. Sci. (Chem.) Dissertation, Moscow: Inst. of Chemical Physics, 1971. 14. Nahlovsky, B. and Chvalovsky, V., Collect. Czech. Chem. Commun., 1968, vol. 33, no. 10, p. 3122.