Translate this page into:
Kinetics, reactivity, initial-transition state analysis and thermodynamic parameters of base-catalyzed hydrolysis of coumalic acid in solvents with different polarities
⁎Corresponding author. Tel.: +20 1064162700; fax: +20 934601159. ahmed_benzoic@yahoo.com (Ahmed M. Abu-Dief)
-
Received: ,
Accepted: ,
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.
Peer review under responsibility of King Saud University.

Abstract
Base-catalyzed hydrolysis of coumalic acid (COU) in binary aqueous-methanol and aqueous-acetone mixtures has been studied kinetically at a temperature range from 283 to 313 K. Moreover, the change in the activation energy barrier of COU from water to water–methanol and water–acetone mixtures is estimated from the kinetic data. Solvent effects on reactivity trends have been analyzed into initial and transition state components by using transfer chemical potentials of the reactants and kinetic data. The transfer chemical potentials for COU− anion are derived from solubility data from its calcium, cerium and lanthanum salts. The decrease in rate constant of the base hydrolysis reaction of COU as the percentage of methanol or acetone increases is dominated by transition state (TS). The base hydrolysis reaction of COU follows a rate law with kobs = k2[OH−] and the reaction mechanism was suggested. The high negative values of entropy of activation support the proposal mechanism, i.e. the investigated reaction takes place via the formation of an intermediate complex. Thus, the ring opening of the intermediate complex would be the rate controlling step.
Keywords
Base-catalyzed hydrolysis
Activation energy barrier
Transfer chemical potential
Initial state
Transition state
Isokinetic temperature
1 Introduction
Coumalic acid is classified under the family of chromen-2-one derivatives. Chromen-2-ones have attracted an intense interest in recent years because of their diverse pharmacological properties. Among these properties, their cytotoxic effects were most extensively examined, and their broad range of effects on the tumors as shown by various in vitro and in vivo experiments and clinical studies were discussed (Kostova, 2005). Chromen-2-one derivatives have also found applications as fluorescent dyes (Brady and Shiek, 1984; Camur and Bulut, 2008), antitumor agents (Rosskoph et al., 1992), antioxidants (Kaneko et al., 2003), anti-inflammatory agents (Melagraki et al., 2009), antineoplastic agents (Lin et al., 2007), immunomodulant agents (Reddy et al., 2004) antifungals, anticoagulants, antibacterial, antimicrobial and insecticides. Moreover, they are used as proliferators of HIV and particular human malignant cell lines in vitro, as well as affecting tumor activity against several in vivo tumor types (El-Khatib and Nassr, 2007). Coumalic acid or its salt which is included in a tea beverage product is used to inhibit non-enzymatic browning of the tea beverage product.
Rate coefficient for the reaction between organic and inorganic compounds with hydroxide ions in aqueous solutions is sensitive to the nature and molar fraction of organic solvent. The dependence of the rate constants on the solvent comprising binary aqueous mixtures depends on the composition of the solvent (Blandamer et al., 1986; Marcus, 2007; Abu-Gharib et al., 2011a,b).
Our work treats the effect of medium on the kinetics of base hydrolysis of COU in aqueous-methanol and aqueous-acetone mixtures. The influence of solvent composition on the rate constant was analyzed in terms of initial and transition states. This approach provides complementary information on the role of the solvation and how transition state modeling can assist in interpretation. Also the activation parameters of the mentioned reactions were calculated by least squares of Arrhenius and Eyring plots.
2 Experimental
2.1 Materials
All materials, sodium hydroxide (99%), sodium chloride (98%), sodium nitrate (97%), methanol, acetone were obtained from BDH. The stock solutions of NaOH and NaCl and NaNO3 were prepared by dissolving the calculated amounts of AnalaR samples in bidistilled water. Coumalic acid (99%) was obtained from sigma.
2.2 Kinetic measurements
The kinetics of the base hydrolysis of coumalic acid was measured by following the dependence of absorbance on time using 10 mm silica cells using the thermostated cell compartment of JASCO model V-530 spectrophotometer thermally controlled at 25 °C ± 0.1 °C and at different temperatures in ultrathermostate (CRIOTERM 190). Chemical reactions were monitored in solutions held at constant ionic strength using appropriate amounts of sodium chloride for methanol and sodium nitrate for acetone over at least three half-life times. Kinetic measurements were performed so that the reaction follows pseudo-first-order kinetics in [COU], where [NaOH] is presented in large excess over [COU]. Rate constants were calculated from the dependence of absorbance on time at 319 nm. It was confirmed that there is no interference from other reagents at the selected wavelength absorption maxima for the investigated compound.
The kinetic of base-catalyzed hydrolysis was carried out on the coumalic acid but solubility measurements were carried out on its salts.
2.3 Solubility measurements
The sparingly soluble calcium, cerium and lanthanum salts of COU were prepared by double decomposition in aqueous-alcohol mixtures. Solubilities were measured by agitating a generous excess of solid with the appropriate solvent, aqueous-methanol and aqueous-acetone mixtures, in a thermostated vessel for 7 h. After allowing time for undissolved solids to settle by using centrifuge, an aliquot of saturated solution was removed and the solution diluted as necessary. Concentrations were measured spectrophotometrically at the wavelength of maximum absorbance (Blandamer et al., 1986; Abu-Gharib et al., 2011a,b). All aliquots of equilibrated solutions were diluted with the solvent methanol or acetone and the dependences of λmax on solvent composition, were found to be very small for COU, and thus ignored in calculation of solubilities. The spectra were measured at different sampling times to verify if a constant concentration has been achieved.
3 Results and discussion
The absorption band of COU at λmax = 236 nm can be attributed π–π∗ transition in the whole electronic system. Moreover, the absorption band at λmax = 290 nm is due to n–π∗.
3.1 Kinetic results and reaction mechanism
The repeated spectral scans of COU (cf. Fig. 1) in the presence of NaOH are characterized by the presence of isosbestic points which imply that the absorbance at those wavelengths remains constant throughout the whole base hydrolysis reaction. It can be realized from repeated spectral scans that the initial action on COU by NaOH takes place in one stage and leads to the rate determining opening of the pyrone ring and the formation of a salt of (2Z)-4-(hydroxymethylene)-2-pentenedioic acid at λmax = 312 nm (El-Khatib and Nassr, 2007; Abu-Gharib et al., 2011a,b). This compound is immediately converted into a salt of (2E)-4-(hydroxymethylene)-2-pentenedioic acid at λmax = 319 nm as shown in the following scheme.![Repeated spectral scan of COU in aqueous-methanol mixture at [OH−] = 0.03 M, [COU] = 1 × 10−4 M, I = 0.05 M and 298 K with interval time = 2 min.](/content/184/2017/10/1_suppl/img/10.1016_j.arabjc.2012.12.040-fig1.png)
The observed first-order rate constants (kobs) as a function on [OH−] and in the presence of different water–methanol and water–acetone ratios were calculated from the dependence of absorbance on time at λmax for the investigated compounds using Microcal Origin program version 7.5 (cf. Tables 1 and 2).
| [OH−], mol dm−3 | MeOH% (v/v) | |||||
|---|---|---|---|---|---|---|
| 0 | 10 | 20 | 40 | 50 | 60 | |
| 104 kobs, s−1 | ||||||
| 0.02 | 14.00 | 10.50 | 8.32 | 5.73 | 4.22 | 3.20 |
| 0.025 | 17.80 | 13.30 | 10.52 | 6.60 | 5.13 | 3.67 |
| 0.03 | 21.00 | 16.60 | 12.60 | 8.30 | 6.30 | 4.42 |
| 0.035 | 24.10 | 18.50 | 14.43 | 9.20 | 7.00 | 5.00 |
| 0.04 | 28.00 | 21.50 | 16.50 | 11.10 | 8.23 | 5.80 |
| 0.045 | 31.10 | 23.70 | 18.70 | 12.30 | 8.90 | 6.40 |
| k2, dm3 mol−1 s−1103 | 70.00 | 53.14 | 41.00 | 26.90 | 19.43 | 13.13 |
| δmΔG#, kJ mol−1 | 0.68 | 1.33 | 2.37 | 3.18 | 4.15 | |
Maximum error is 1%.
| [OH−], mol dm−3 | Acetone% (v/v) | |||||
|---|---|---|---|---|---|---|
| 0 | 10 | 20 | 40 | 50 | 60 | |
| 104 kobs, s−1 | ||||||
| 0.02 | 14.60 | 13.20 | 11.60 | 8.25 | 6.40 | 5.30 |
| 0.025 | 18.10 | 16.70 | 14.70 | 10.10 | 8.20 | 6.00 |
| 0.03 | 21.60 | 20.30 | 17.10 | 12.30 | 9.70 | 7.25 |
| 0.035 | 25.10 | 23.40 | 20.20 | 13.90 | 11.12 | 8.44 |
| 0.04 | 29.30 | 26.10 | 22.40 | 16.20 | 12.80 | 9.50 |
| 0.045 | 33.30 | 29.50 | 25.30 | 17.40 | 14.14 | 10.50 |
| 103k2, dm3 mol−1 s−1 | 72.90 | 64.90 | 54.60 | 37.80 | 31.00 | 22.60 |
| δmΔG#, kJ mol−1 | 0.29 | 0.72 | 1.63 | 2.12 | 2.90 | |
Maximum error is 1%.
Tables 1 and 2 shows that the values of kobs decrease as the proportion of methanol or acetone increases. This may be due to the salting out of the hydrophilic OH− and stabilization of the studied compound.
The observed enhancement of the rate constant values in acetone co-solvent compared with the corresponding values in methanol co-solvent would be ascribed to the destabilization of the hydrophilic OH− ions in acetone more than in methanol.
The dependence of kobs on base concentration is linear for COU in methanol and acetone without significant intercept (cf. Figs. 2 and 3) hence the hydrolysis follows the rate law with
![Plots of the observed first order rate constants of the reaction between NaOH and COU as a function of [NaOH] in different water–methanol mixtures at [COU] = 1 × 10−4 M, I = 0.05 M and 298 K.](/content/184/2017/10/1_suppl/img/10.1016_j.arabjc.2012.12.040-fig2.png)
![Plots of the observed first order rate constants of the reaction between NaOH and COU as a function of [NaOH] in different water–acetone mixtures at [COU] = 1 × 10−4 M, I = 0.05 M and 298 K.](/content/184/2017/10/1_suppl/img/10.1016_j.arabjc.2012.12.040-fig3.png)
The change in the activation barrier δmΔG# is evaluated and reported in Tables 1–4 from the ratio of rate constants of the base hydrolysis in the aqueous-solvent (k2S) to the corresponding values in the aqueous solution (k2W) according to the following relation [11]:
| Compound | MeOH% | 104 Solubility. | δmμθ (salt) kJmol−1 | δmμθ (cation) kJmol−1 | δmμθ (anion) kJmol−1 |
|---|---|---|---|---|---|
| La(COU)3 Є = 12465 | 0 | 40.90 | |||
| Mol L−1 cm−1 | 20 | 17.93 | 8.17 | 3.70 | 1.49 |
| λmax = 234 nm | 40 | 8.40 | 15.96 | 7.29 | 2.80 |
| 60 | 4.93 | 21.00 | 10.56 | 3.48 | |
| Ca(COU)2 | 0 | 43.45 | |||
| Є = 12472 | 20 | 21.47 | 5.24 | 2.20 | 1.52 |
| Mol L−1 cm−1 | 40 | 11.43 | 9.93 | 4.29 | 2.82 |
| λmax = 235 nm | 60 | 8.20 | 12.42 | 5.42 | 3.50 |
| Average | 20 | 1.27 | |||
| δmμθ(COU-) | 40 | 2.60 | |||
| kJ mol−1 | 60 | 3.49 |
| Compound | MeOH% | 104 Solubility | δmμθ (salt) kJ mol−1 | δmμθ (cation) kJ mol−1 | δmμθ (anion) kJ mol−1 |
|---|---|---|---|---|---|
| La(COU)3 | 0 | 40.23 | |||
| Є = 12465 | 20 | 6.78 | 17.65 | −2.90(a) | 6.85 |
| Mol L−1 cm−1 | 40 | 5.40 | 19.90 | −15.80 | 11.90 |
| λmax = 234 nm | 50 | 3.76 | 23.50 | −26.60 | 16.70 |
| Ce(COU)3 | 0 | 38.25 | |||
| Є = 12470 | 20 | 6.15 | 18.15 | −2.1(a) | 6.75 |
| Mol L−1 cm−1 | 40 | 4.15 | 22.00 | −13.1 | 11.70 |
| λmax = 234 nm | 60 | 2.67 | 26.40 | −23.4 | 16.60 |
| Average | 6.80 | ||||
| δmμθ (COU−) | 11.80 | ||||
| kJ mol−1 | 16.65 |
It is observed that the values of δmΔG# increase with the increasing of methanol or acetone content and these match with the decreasing of kobs and k2 values as methanol or acetone content increases.
On applying the steady-state approximation for the concentration of the intermediates B, C and D on the suggested reaction mechanism of COU, the total concentration of the compound: [A]T = [B] + [C] + [D], where [A] is very small and neglectable because it reacts very fast with the base to form the intermediate B. Thus, the rate equation can be formulated as:
3.2 Michaelis–Menten kinetics
The base hydrolysis reaction of COU exhibits Michaelis–Menten kinetics (cf. Fig. 4). This suggests that the reaction occurs via the formation of intermediates as shown in the suggested mechanism (cf. Scheme 1). From Lineweaver–Burk plots, Michaelis–Menten constants (KM) were calculated and found to be 0.43 mol dm−3 s−1 for COU. The large values of KM indicate the formation of intermediates during the reaction progress (Michaelis and Menten, 1918; Nassr, 2010; Awad et al., 2008).![Lineweaver–Burk plot of the base hydrolysis of COU in aqueous solution with [compound] = 1 × 10−4 mol dm−3, I = 0.05 M and 298 K.](/content/184/2017/10/1_suppl/img/10.1016_j.arabjc.2012.12.040-fig4.png)

3.3 Initial state-transition state analysis
Solvent effects on the rate constants reflect solvation changes on the initial and transition states as these are transferred from water into mixed aqueous methanol or aqueous acetone media. If solvent effects on the initial state can be established from thermodynamic data (solubility data), then solvent effects on the transition state can be calculated from initial state and kinetic data assuming constancy of mechanism and validity of transition state theory.
The transfer chemical potentials for the constituent ions of a salt between two solvents are given by the following equation (Blandamer et al., 1986; Abu-Gharib et al., 2011a,b).
In order to establish solvation effects on the chemical potential of COU− anion, the solubilities of a suitable salt in the water–methanol and water–acetone mixtures are needed. Such a salt should be only sparingly soluble, to minimize activity corrections in calculating transfer chemical potentials. These salts are Ca(COU)2, Ce(COU)3 and La(COU)3. Thus, the transfer chemical potentials of single cations: Ca2+, Ce3+ and La3+ are required.
The transfer chemical potentials of these cations were obtained from reference (). Then the mean of transfer chemical potentials of COU− anion is deduced from the values of the transfer chemical potentials of sparingly soluble salts Ca(COU)2, La(COU)3 and Ce(COU)3 in aqueous-methanol and aqueous-acetone mixtures (cf. Tables 3 and 4). Indeed the values of (COU−) which are calculated from its calcium and lanthanum salts in aqueous-methanol mixtures are in good agreement. Also the values of (COU−) which are calculated from its lanthanum and cerium salts in aqueous-acetone mixtures are the same.
The values of the transfer chemical potentials of OH− anion in methanol are obtained from references (Abu-Gharib et al., 1984) and for OH− in acetone from references Marcus (2007). Also the values of transfer chemical potentials of La3+, Ca2+, and Ce3+ are obtained from reference (Abu-Gharib et al., 2011a,b).
The change in second order rate constant on going from water to medium (aqueous -methanol or aqueous-acetone mixtures) of mole fraction x2, δmΔG# will be equal to ΔG#(x2) minus ΔG#(x2 = 0).
The free energies of activation and transfer chemical potentials are now used in the analysis of reactivity trends with changing solvent composition into initial state and transition state components, i.e., combination of kinetic data, solubility data and transfer chemical potentials yield the effect of solvent on the transition state according to equation (5).
The initial state–transition state analysis of reactivity trends for COU− anion in aqueous–methanol and aqueous–acetone mixtures is set out in Table 5 and depicted in Figs. 5 and 6.
| Solvent | Solvent% | δmμθ(COU−) kJ mol−1 | δmμθ(OH−) kJ mol−1 | δmμθ(IS) kJ mol−1 | δmΔG# kJ mol−1 | δmμθ(TS) kJ mol−1 |
|---|---|---|---|---|---|---|
| MeOH | 0 | |||||
| 20 | 1.51 | −0.20 | 1.31 | 1.33 | 2.64 | |
| 40 | 2.81 | 0.10 | 2.91 | 2.37 | 5.28 | |
| 60 | 3.49 | 1.60 | 5.10 | 4.15 | 9.25 | |
| Acetone | 0 | |||||
| 20 | 6.80 | 1.64 | 8.44 | 0.72 | 9.16 | |
| 40 | 11.80 | 2.45 | 14.25 | 1.63 | 15.88 | |
| 50 | 16.65 | 2.73 | 19.38 | 2.90 | 22.28 |


Table 5 shows that COU- anion and the hydroxide ion are destabilized, resulting in the destabilization of the initial state as the proportion of methanol or acetone increases. However the transition state is significantly destabilized on increasing the methanol or acetone content (cf. Figs. 5 and 6). These results affect the decrease of the rate constant of COU− anion i.e. the dominant effect of (TS) on the rate constant of the base hydrolysis of COU− anion as the co-solvent methanol or acetone increases.
3.4 Determination of the activation parameters
The activation parameters were calculated (cf. Tables 6 and 7) by the least squares of Arrhenius plot (Fig. 7) and Eyring plot (Fig. 8).(see Fig. 9).
| MeOH% | T (K) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 283 | 288 | 293 | 298 | 303 | 308 | 313 | Ea kJ mol−1 | ΔH≠k J mol−1 | ΔG≠ kJ mol−1 | ΔS≠ J mol−1 K−1 | A mol−1 dm3 s−1 | |
| COU | 103 k2, mol−1 dm3 s−1 | |||||||||||
| 0 | 24.77 | 34 | 49.5 | 70.0 | 101.3 | .... | .... | 39.25 | 37.8 | 67.1 | −98.19 | 5.11 × 108 |
| 10 | 18.35 | 25.1 | 37.2 | 53.14 | 75.8 | 104.6 | 153.6 | 40.42 | 38.9 | 67.35 | −95.48 | 4.15 × 108 |
| 20 | 15.36 | 20.0 | 30.7 | 41.0 | 62.54 | 81.0 | 123 | 42.18 | 40.6 | 68.3 | −92.93 | 3.9 × 108 |
| 40 | 10.43 | 13.7 | 20.74 | 26.90 | 41.49 | 55.16 | 82.95 | 44.73 | 43.4 | 69.7 | −88.17 | 8.45 × 107 |
| 50 | 7.39 | 9.5 | 14.43 | 19.43 | 28.65 | 33.61 | 57.29 | 46.19 | 44.7 | 70.15 | −85.39 | 5.76 × 107 |
| 60 | 4.21 | 6.45 | 8.87 | 13.13 | 17.76 | 26.63 | 35.22 | 47.45 | 46.1 | 70.76 | −82.75 | 1.97 × 107 |
| T (K) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Acetone % | 283 | 288 | 293 | 298 | 303 | 308 | 313 | Ea kJ mol−1 | ΔH≠ kJ mol−1 | ΔG≠ kJ mol−1 | ΔS≠ J mol−1 K−1 | A mol−1 dm3 s−1 |
| COU | 103 k2, mol−1 dm3 s−1 | |||||||||||
| 0 | 24.4 | 36.9 | 49.0 | 72.9 | .... | .... | ... | 36.81 | 35.3 | 65.3 | −100.6 | 2.1 × 109 |
| 10 | 22.2 | 32.4 | 44.0 | 64.9 | 89.0 | .... | ... | 38.12 | 36.5 | 65.73 | −98.1 | 1.7 × 109 |
| 20 | 18.23 | 27.0 | 36.0 | 54.6 | 73.0 | 110 | 144 | 39.78 | 38.2 | 66.6 | −95.2 | 1.32 × 109 |
| 40 | 12.90 | 19.0 | 25.4 | 37.8 | 50.0 | 76.0 | 106 | 41.33 | 41.1 | 68.1 | −90.6 | 7.31 × 108 |
| 50 | 10.60 | 15.4 | 20.0 | 31.0 | 41.0 | 61.7 | 80 | 43.79 | 42.3 | 68.5 | −87.8 | 5.53 × 108 |
| 60 | 7.44 | 11.1 | 14.7 | 22.6 | 30.0 | 45.0 | 59 | 44.55 | 43.5 | 68.92 | −85.3 | 3.8 × 108 |



It is worth mentioning that the activation parameters are important to determine the rate of reaction and reaction mechanism. The high negative values of entropy of activation (Tables 6 and 7) support the proposal mechanism, that is the investigated reaction takes place via the formation of an intermediate complex (El-Khatib and Nassr, 2007; Awad et al., 2008; Abu-Gharib et al., 2011a,b), cf. the mechanism. Moreover, these values refer to the rigidity of the complex (B) and its high stability. Thus, the ring opening of the intermediate complex would be the rate controlling step (RCS) and proceeds with an inner-sphere electron transfer step.
It is interesting that, the different thermodynamic functions are consistent in their trends. As entropy of activation (ΔS≠) increases, i.e., changes to less negative value, the rate constant decreases and activation energy increases. This behavior can be ascribed to enhancing stability of activated complex intermediate. Again the relatively high free energy change would assume that the slow step (RCS) is the ring opening of the established intermediate (B) and many vibrational degrees of freedom have been transformed into translations (Abu-Gharib et al., 2011a,b; Jagan, 2009). The large frequency factor would suggest synergistic evidence for the proposal pathway. Furthermore, the positive free energy of transfer from water to methanol or acetone reported in Tables 6 and 7, assumes that the transient species in hand are polar entities [El-Khatib and Nassr, 2007; Abu-Gharib et al., 2011a,b].
The values of activation enthalpies (ΔH≠) and activation entropies (ΔS≠) are important in controlling the reaction rate. The values of ΔH≠ are plotted versus the values of ΔS≠ (cf. Fig. 8) and a straight line is obtained. The isokinetic temperature (β) was determined by least square of the slope and found to be 305 K, which is greater than the average experimental temperature (Texp = 298 K) indicating that the reaction rate is enthalpy controlled (Karunakaran and Chidambaranathan, 2000; Karunakaran and Chidambaranathan, 2001a,b). The straight line obtained, show that the base hydrolysis of COU and other chromen-2-one derivatives followed one mechanism, i.e. the same rate determining step (R. D. S.).
4 Conclusion
Base-catalyzed hydrolysis of coumalic acid follows a rate law with kobs = k2[OH−]. The decrease in the rate constants of coumalic acid as the proportion of methanol or acetone increases is due to the destabilization of OH- ion. The values of rate constants (kobs and k2) decrease in the following order water > acetone > methanol with increasing the methanol or acetone content. The dominant effect on the rate constant of the base hydrolysis of COU as the co-solvent methanol or acetone increases is the initial state. The high negative values of entropy of activation support the proposal mechanism, i.e. the investigated reaction takes place via the formation of an intermediate complex. Moreover, these values refer to the rigidity and stability of the intermediate complex .Thus, the ring opening of the intermediate complex would be the rate controlling step.
References
- J. Trans. Metal Chem.. 1984;9:306.
- Z. Phys. Chem.. 2011;225:1.
- J. Korean chem. Soc.. 2011;50(3):346.
- Spectrochimica Acta. 2008;71:921.
- J. chem. Soc. Fraday Trans.. 1986;82:1471. 1
- J. Heter. Chem.. 1984;21:1337.
- Dyes and Pigments. 2008;7:165.
- Spectrochimica Acta. 2007;67:643.
- Jagan, V., 2009. Osmania University, India, 18, 4, 89.
- Chem. Biol. Interact.. 2003;142:239.
- Mon. für Chemie. 2001;131:1123.
- J. coroatica chemica acta. 2001;74:51.
- Curr. Med. Chem., Anti-Cancer Agents. 2005;5:29.
- J. Bioorg. Med. Chem. Lett.. 2007;17:4557.
- Chem. Rev.. 2007;107(9):3880.
- Euro. J. Med. Chem.. 2009;44:3020.
- Biochem. Z.. 1918;49:333.
- Int. J. Chem. Kinet.. 2010;42:372.
- J. Bioorg. Med. Chem. Lett.. 2004;14:4093.
- Pharmazie. 1992;47:139.
