Experimental and Computational Tools for Mechanistic Study: A Modified Perkin Condensation leading to Alpha-Phenylcinnamic Acid Isomers

István Pálinkóa, Béla Töröka, Gyula Tasib and Tamás Körtvélyesic

aDepartment of Organic Chemistry, Dóm tér 8, Szeged, H-6720 Hungary bApplied Chemistry Department, Rerrich B. tér 1, Szeged, H-6720 Hungary cDepartment of Physical Chemistry, Rerrich B. tér 1, Szeged, H-6720 Hungary


A modified Perkin condensation leading to alpha-phenylcinnamic acid stereoisomers was studied with experimental and computational tools. The condensation reaction gave overwhelmingly or exclusively the E isomer. The duration and temperature of reflux and the polarity of solvent could affect the isomeric distribution only to a minor extent. However, E-Z isomerization provided with equilibrium mixture of nearly 1:1 composition. Computational results (obtained by means of molecular mechanics and semiempirical quantum chemical methods) also indicated that there was negligible difference between the stabilities of the isomers. Nevertheless, the dramatic difference in the potential energy surfaces of the isomers may account for the stereoselectivity of the condensation reaction.

Keywords: Perkin condensation, isomeric distribution, E- and Z-alpha-phenylcinnamic acid stereoisomers, computational methods, potential energy surfaces, stereoselectivity


Cinnamic acid derivatives are important building blocks in the production of lignins in higher plants [1]. They derive from the shikimic acid metabolic pathway and their mechanism of formation is complex. Nevertheless, key reactions are condensations (mostly of Claisen type [1]), just as in their laboratory scale synthesis (mostly of Perkin type [2]). The 'usual' Perkin reaction leads to the predominant or exclusive formation of the E isomer. A modified Perkin condensation (Scheme 1) [3-5], under certain conditions (low temperature, short reaction time) gives the mixture of E- and Z-alpha-phenylcinnamic acids, nonetheless, the E isomer still predominates. Early studies (see, e.g., [6]) found this fact obvious, since the E isomer was considered thermodynamically more stable. Indeed, lengthening the duration of reflux afforded exclusively this isomer. However, when we started to model the two isomers in order to find structural explanation for the large difference in the acidities of the stereoisomers, AM1 semiempirical quantum chemical calculations gave nearly identical heats of formation data for the fully optimized structures of the isolated E and Z isomers. Therefore, it seemed interesting to study the reaction from mechanistic point of view, with the aim of exploring factors, which influence isomeric distribution. The thermal and photoinduced E-Z isomerizations of the acids were also investigated. The reactions were explored experimentally as well as computationally.

Computational methods

The AM1 [7], MNDO [8], and PM3 [9]) semiempirical methods were selected for studying the isolated molecules. Routines included in the PcMolTM [10] and MOPAC [11] packages were applied for calculations. Full geometry optimization was performed by each method. In calculations performed by the MOPAC package the RHF approximation was used. The gradient norm and the energy derivatives with the internal coordinates for all the individual gradients were always less than 0.01 and 0.1 respectively. The fully optimized geometries had positive definite force matrices indicating that stationary points were found. Potential energy maps were created for both isomers using the MM2 [12] routine of the CacheTM system [13]. Both isomers were preoptimized without constraints, then the two phenyl groups were rotated by 1 degree increments and potential energies were calculated for each conformer. Thus, fairly detailed maps were created. Potential energy values in more than 14.000 points were calculated.


The condensation reaction

First, the reaction was performed following the recipe of Fieser [4]. It involved heating a mixture of benzaldehyde (2 cm3), phenylacetic acid (2.5 g), acetic anhydride (2 cm3) and triethylamine (2 cm3) for 35 min. A mixture of products was precipitated with 4 cm3 of cc. HCl. The solid material was dissolved in diethyl ether then was washed with 3x10 cm3 3% NaOH solution. The two isomers from the alkaline solution was obtained by selective precipitation. Acidifying with acetic acid to pH 5 afforded the E isomer, further lowering the pH to 1 with cc. HCl provided with the Z isomer. The crude products were crystallised from diethyl ether (m.p. (E): 446.5 K, m.p.(Z): 409.5 K).

In an other set of experiments duration and temperature of reflux (this latter by varying the solvent) were altered. Solvents were chosen to provide with different boiling points, polarities and hydrogen bonding capabilities. Usually 100 cm3 of solvent was used to enforce the properties of the solvent on the reaction mixture (the quantities of the other components remained unchanged). Duration of reflux was varied from 35 min to 9 hours in some cases. Products were isolated and separated.

Thermal and photoinduced isomerization reactions

Thermal isomerization was performed in the Fieser mixture, i.e., 0.5 g of E-alpha-phenylcinnamic acid was placed to a 1:1 mixture acetic anhydride and triethyl amine (2 cm3 each) mixture and was kept under reflux for 168 hours. For determining the isomeric composition, samples were withdrawn for GC-MS analysis at certain time intervals [14].

Photoinduced isomerization was performed with the help of an immersion mercury vapour lamp (intensity maximum at lambda = 253.1 nm). The reaction was performed at room temperature in a jacketed flask (water was used for controlling the temperature) under nitrogen atmosphere, placing 1 g of E-alpha-phenylcinnamic acid in 500 cm3 diethyl ether. Duration of illumination was 30 hours. Analysis was performed the GC-MS method.


The condensation and the double bond isomerization reactions

The modified Perkin reaction provided an isomeric mixture with the predominance of the E isomer. The maximum amount of the Z isomer was 21.1%. It was obtained when the Fieser mixture was kept under reflux (about 423 K) with 35 min reaction time. A compilation of literature results revealed (Table 1) that any modification in the composition of reactants, temperature and duration of reflux resulted in the preferential or in most cases exclusive formation of the E isomer.

Literature survey of isomeric distribution under various condititons Table 1

Isomeric distribution at various temperatures and reaction times based on literature results


T/oC t/min benz/mola Ac2O/molb Phac/molc Et3N/mold E/Z __________________________________________________________________________________

301 90 0.74 0.3 0.2 0.2 100/0

301 90 0.74 0.3 0.2 0.2* 100/0

651 180 0.74 0.3 0.2 0.4 100/0

651 180 0.74 0.3 0.2 0.4* 100/0

1402 35 0.18 2.1 0.11 1.5 96/4

1503 35 0.3 0.21 0.18 0.15 79/21

1504 330 0.4 0.84 0.4 0.29 100/0


a benzaldehyde

b acetic anhydride

c phenylacetic acid

d triethylamine

* added dropwise during the reaction

1 ref. [17]

2 ref. [6]

3 ref. [4]

The amount of Z isomer did not increase when solvents of various properties (dielectric constant, hydrogen bonding capability) were used (Table 2). Moreover, the exclusive formation of the E isomer can be considered typical. The Z isomer was formed (in small amounts even under these circumstances) when high temperature, long reaction time and apolar solvent was used or the solvent was capable of hydrogen bonding (Table 2).

Effects of solvents and length of heating

Table 2

Isomeric distribution in solvents of different polarities (mu = dielectric constant) and boiling points at various reaction times (reactions were carried out under reflux)


t/min E/mol% Z/mol%


Fieser mixture* (b.p.: app. 150 oC)

35 79.9 21.1 _______________________________________________________________________________

H2O (b.p.: 100 oC, mu(25): 1.82)

180 94.4 5.6 _______________________________________________________________________________

diethyl ether (b.p.: 35 oC, mu(25): 1.87)

180 100 0 _______________________________________________________________________________

dibutyl ether (b.p.: 142-143 oC, mu(20): 1.15)

40 100 0 _______________________________________________________________________________

diglyme (b.p.: 162 oC, mu(25): 1.97)

180 100 0 _______________________________________________________________________________

benzene (b.p.: 80 oC, mu(20): 0.00)

180 100 0 _______________________________________________________________________________

xylene (isomeric mixture) (b.p.: 138-142 oC, mu(20): 0.30)

30 100 0

180 98.1 1.9

360 97.9 2.1 _______________________________________________________________________________

decalin (isomeric mixture) (b.p.: 192 oC, mu(25): 0.00)

35 94.2 5.8

360 87.4 12.6

540 90.5 9.5


*Fieser mixture: benzaldehyde, phenylacetic acid, acetic anhydride, triethylamine

The results of thermal as well as photoinduced isomerization reactions revealed that starting from the E isomer an equilibrium mixture of nearly 1:1 composition could be achieved (Table 3). It is true that reaching this composition requires considerable amount of time.

E-Z isomerization

Table 3

Double bond isomerization of E-alpha-phenylcinnamic acid under thermal conditions (reflux in a mixture of acetic anhydride and triethylamine) and upon UV irradiation (in diethyl ether)


   time/h              thermal                          UV light
                  ___________________              ____________________

E Z E Z _______________________________________________________________________

0 100 0 100 0

22 - - 61.1 38.9

24 96.7 3.3 - -

30 - - 61.1 38.9

48 83.3 16.7 - -

96 66.2 33.6 - -

120 54.7 45.3 - -

168 54.7 45.3 - -


Computational results

Semiempirical calculations on the isolated molecules gave similar stability data for the two isomers. However, calculations also revealed that the dipole moments were different (Table 4).

Standard heat of formation and dipole moment data on alpha-phenylcinnamic acid isomers calculated by semiempirical methods

Table 4

The standard heat of formation (in kJ/mol) and dipole moment (D in Debye) data for the alpha-phenylcinnamic acid isomers calculated by semiempirical quantum chemical methods

                  MNDO                    AM1                     PM3
           ________________       _________________       __________________

deltaHf,298 D deltaHf,298 D deltaHf,298 D


E -109.7 2.13 -99.2 2.53 -97.7 2.04

Z -114.0 1.80 -97.5 1.88 -97.2 1.80


Potential energy maps also differed considerably. The quasi three-dimensional potential energy surface was table-like for the E isomer with ridges of very small heights between the shallow valley-like minima. There were sharp maxima when the hydrogens on the phenyl groups approached each other. The the map upside down). The two sets of almost symmetrical minima network were connected by relatively high energy hills, which corresponded to a near parallel arrangement of the phenyl groups.


Based on published information [6, 15-17] and common sense, a detailed mechanism of this modified Perkin condensation can be put together (Scheme 2). It consists of two parts: (1) nucleophilic addition of the mixed anhydride formed from acetic anhydride and phenylacetic acid on the carbonyl group of benzaldehyde via base catalysis (Scheme 3); (2) the elimination of acetic acid and the hydrolysis of the product anhydride (Scheme 4). Zimmermann et al. have shown that the condensation reaction is elimination controlled [6]. There is no reason to assume that preferential formation of any of the four intermediates (Table 5) takes place.

Configuration of the intermediates

Table 5

The absolute configurations of the diastereomers and those of the products after elimination (for graphical view)


compounds chiral centers products after elimination ____________________________________

a b _________________________________________________________________________________

1 S R Z

2 R R E

3 S S E

4 R S Z


According to AM1 semiempirical quantum chemical calculations, the heats of formation are very close to each other, their dipole moments are not too different either (Table 6). Also there is no need to assume other than the usual anti elimination [18], it can account for the formation of both isomeric acids.

Standard heat of formation and dipole moment data for the intermediates obtained by the AM1 semiempirical method

Table 6

Standard heat of formation data (in kJ/mol) and dipole moments (D in Debye) for the isolated intermediates obtained by the AM1 method


compounds deltaHf,298 D _________________________________________________________________

1 -657.8 5.48

2 -640.2 5.51

3 -652.8 5.24

4 -640.8 5.45


As concerns the stability data of the intermediates and the alpha-phenylcinnamic acid stereoisomers it is safe to state that the elimination reaction is endothermic. (It is strictly true for gas-phase reactions, nevertheless, it should be approximately valid in our system as well.) Thus, the transition state is product-like, therefore, the potential energy surfaces of the products can be used in explaining stereoselectivity towards the E isomer. As it has been already mentioned the potential energy surface for the E isomer contains extensive low-energy regions close to the absolute minimum throughout. The potential energy surface of the other isomers has several local minima but considerable amount of the surface is in the high energy region (saddles, ridges, local maxima). Consequently, the reason for the preferential formation of the E isomer is of kinetic origin, its formation is much faster than that of the Z. When reaction conditions are more severe this isomer also forms, but never in excess.


Using the modified Perkin condensation leading to alpha-phenylcinnamic acid isomers as model reaction, it has been demonstrated that experimental and computational methods together are able to give a more complete account of the observed stereochemical features than any of them could be alone.


This work was supported by the National Science Foundation of Hungary through grant F4297/1992. The financial help is highly appreciated.


[1] Mann, J., in Secondary Metabolism Ch. 4, p. 173, Clarendon Press, Oxford, 1987.

[2] Johnson, J., in Organic Reactions (Eds. Adams, R., Bachmann, W.F., Johnson, J.R., Fieser, L.F. and Snyder, H.R.) p. 210, John Wiley & Sons, Inc., Chapman & Hall, Ltd., New York and London, 1947.

[3] Buckles, R.E., J. Chem. Ed. 27 (1950) 210.

[4] Fieser, L.F., in Experiments in Organic Chemistry p. 182, Heath and Co., Boston, 1955.

[5] Buckles, R.E. and Bremer, K., in Org. Synt. Coll. Vol. 4, p.777, John Wiley & Sons, Inc., New York, London, Sydney, 1967.,

[6] Zimmermann, H. and Ahramjian, L., J. Am. Chem. Soc. 81 (1959) 2086.

[7] Dewar, M.J.S., Zoebisch, E.G., Healy, E.F. and Stewart, J.J.P., J. Am. Chem. Soc. 107 (1985) 3902.

[8] Dewar, M.J.S. and Thiel, W. J. Am. Chem. Soc. 99 (1977) 4899.

[9] Stewart, J.J.P., J. Comput. Chem. 10 (1989) 209, 221.

[10] Tasi, G., Pálinkó, I., Náray-Szabó, G., Halász, J., PcMolTM3.0, semiempirical Quantum Chemical Calculations on Microcomputers, Chemicro Ltd., Budapest, 1992.

[11] Stewart, J.J.P., QCPE Bull. 6 (1983) 43; MOPAC 6.0, QCPE program 455.

[12] Allinger, N.L., J. Am. Chem. Soc. 99 (1977) 8127.

[13] CacheTM, version 2.2, Tektronix Inc., licensed to Prof. Seddon, K.R., The Queen's University of Belfast, permission for use is highly appreciated.

[14] Török, B., Pálinkó, I., Tasi, Gy., Nyerges, L. and Bogár, F., J. Chrom. A 668 (1994) 353.

[15] Ogata, Y. and Tsuchida, M., J. Org. Chem. 24 (1959) 78.

[16] Ketcham, R. and Jambotkar, D., J. Org. Chem. 28 (1963) 1034.

[17] Buckles, R.E. and Cooper, J.A., J. Org. Chem. 30 (1965) 1588.

[18] Isaacs, N.S., in Physical Organic Chemistry p. 524, Longman Scientific & Technical, John Wiley & Sons, Inc., 1987; Carey, F. and Sundberg, R.J., in Advanced Organic Chemistry, Part A p. 377, Plenum Press, New York and London, 1993.

keyword search Home page