Linear free energy relationships in the Bergman cyclization of 4-substituted-1,2-diethynylbenzenes

Article abstract of DOI:10.1016/S0040-4039(00)01181-3

A series of 4-substituted-1,2-diethylnylbenzenes were prepared (1a-f) and subjected to Bergman cyclization. Bulk cyclization proceeded to produce the corresponding 2-substituted naphthalenes (2a-f) in good yields (59-78%). Kinetic experiments show a linear free energy relationship between the cyclization rate and the Hammett σ(m) substituent coefficient and the Swain-Lupton field and resonance parameters. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01181-3

Tetrahedron Letters 41 (2000) 6955±6958
Linear free energy relationships in the Bergman cyclization
of 4-substituted-1,2-diethynylbenzenes
Nakyen Choy, Chang-Sik Kim, Cynthia Ballestero, Leanne Artigas, Christian Diez,
Frank Lichtenberger, Jed Shapiro and K. C. Russell*
University of Miami, Department of Chemistry, 1301 Memorial Dr., Coral Gables, FL 33124-0431, USA
Received 14 July 2000
Abstract
A series of 4-substituted-1,2-diethylnylbenzenes were prepared (1a±f) and subjected to Bergman cyclization.
Bulk cyclization proceeded to produce the corresponding 2-substituted naphthalenes (2a±f) in good yields
(59±78%). Kinetic experiments show a linear free energy relationship between the cyclization rate and the
Hammett sm substituent coecient and the Swain±Lupton ®eld and resonance parameters. # 2000
Elsevier Science Ltd. All rights reserved.
Ever since the discovery of the naturally occurring enediyne anticancer antibiotics in the mid
1980s, interest in the Bergman cyclization reaction has been rejuvenated.1 Numerous studies have
explored ways to activate cyclization in analogues of the natural products. Evidence has resulted
in the recognition of two primary contributing factors which e€ect cyclization. First is the strain
energy. As the triple bond termini are drawn closer together, the strain energy of the molecule is
increased, facilitating cyclization. This can be accomplished in structurally simple enediynes by
incorporating the enediyne into a 10-membered ring, lowering the cyclization onset temperature.2 The
second factor is an electronic e€ect. This e€ect can be seen through manipulation of several dif-
ferent variables, such as solvent,3 incorporation of the double bond into an aromatic ring,4 or the
use of electron-withdrawing or -releasing groups.5
Recently, we have been involved in the investigation of enediynes where the double bond is
incorporated into an electron-de®cient heterocycle.6 Our data strongly supports the notion that
when the double bond is electron-de®cient, Bergman cyclization is accelerated. Based on these
results, we wished to determine if a linear free energy relationship exists for the Bergman
cyclization of 4-substituted-1,2-diethynyl benzenes (Scheme 1).
The Bergman cyclization is unusual in that it generates a neutral s,s-diradical intermediate.
Since currently no parameters are available for diradical forming reactions of this type, an
* Corresponding author.
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01181-3

6956
Scheme 1.
important issue is to establish which parameters or combinations of parameters, if any, are
appropriate to describe this relationship. The various types of Hammett parameters are usually
involved with charges or radicals which are associated with p-systems. The fact that the radicals
are orthogonal to the p-system, although they are in benzylic positions brought into question
whether any of the currently available Hammett-type parameters could be used to draw reasonable
linear free energy relationships in this case.
Table 1 shows the Hammett parameters used in this study. The standard sm and sp values
were examined, in spite of the fact that they are generally used for polar reactions.7 The sjj and
smb parameters were developed speci®cally to deal with radical systems.8 The constants F and R
are for ®eld and resonance contributions, respectively, for the Swain±Lupton model.9
Table 1
Hammett-type parameters for compounds used in this studya
Compounds 1a±f were synthesized using palladium coupling strategies from the corresponding
4-substituted dihalobenzenes.10 In order to be certain the kinetics being measured were related to
Bergman cyclization, bulk cyclization of each material was performed under conditions nearly
identical to those used for the kinetics runs. The yields of isolated Bergman cyclized products
(2a,c±f) ranged from 58±78% and are summarized in Table 2. The cyclization yield of 2b was not
determined since the cyclization of this compound has been carefully studied previously.11
Kinetics were measured at 170ꢀC using chlorobenzene as the solvent and 100 equiv. 1,4-cyclo-
hexadiene (1,4-CHD) as the radical trap.12 Each sample was measured over a minimum of two
half-lives. The kinetics were measured as a function of the disappearance of starting material
under the pseudo-®rst order conditions using anthracene as an internal standard. The kinetic data
are summarized in Table 2. In each case the data gave an excellent linear correlation (R ꢁ0.990).
As expected, the nitro compound (1f) had the shortest half-life (4.3 min) and the aniline derivative

6957
Table 2
Kinetic dataa and bulk cyclization yieldsb for the thermal cyclization of 1a±f
(1a) had the longest half-life (15.6 min). The only apparently anomalous data came from ketone
1d and ester 1c. Although comparison of the parameters examined in this study show the electron-
withdrawing potential of the ketone and ester are similar, in each case the ketone is the more
electron-withdrawing of the two. However, the kinetics clearly show that while the half-lives of 1c
and 1d were the closest of any two compounds, ketone 1d, with a half life of 6.5 min cyclized
slightly more slowly than ester 1c, whose half-life was 5.9 min.
Individual Hammett plots were constructed for the sm, sp, sjj and smb parameters. The
coecients for the ®eld and resonance parameters of the Swain±Lupton model were determined
using a two parameter regression algorithm.13 The results are summarized in Table 3. The best
correlation was achieved using the Hammett sm value. This gave a Hammett constant of 0.654.
None of the other parameters gave nearly the same correlation. In addition, since the H value for
each parameter is 0.0, with the exception of F, where the value is 0.3, the intercept of the
Hammett plot is expected to be 0.0. Again, the Hammett plot using sm has the intercept closest to
its expected value. The Hammett constants and intercepts for sp, and smb are quite similar, but
still not as good as those for sm. The Swain±Lupton model gave a relationship of 0.662F+0.227R,
with correlation comparable to sm.
Table 3
Slope (r), intercept and correlation from Hammett plots of 1a±f versus various parameters
The results from the Hammett study indicate a linear free energy relationship between the rate
of Bergman cyclization and the Hammett sm value as well as appropriate contributions from the
®eld and resonance e€ects. The value of ꢀ from sm (0.654) is reasonable since the reaction goes

6958
through uncharged transition states and intermediates. Also, the larger ®eld parameter (0.662)
versus the resonance parameter (0.227) from the Swain±Lupton model is reasonable since the
developing diradical is orthogonal to the p-system. These results give a clearer insight into the
factors which provide electronic control over the Bergman cyclization, and may provide a
valuable design tool in the development of enediyne based drugs.
Acknowledgements
C.D. would like to thank the Howard Hughes Foundation for an undergraduate fellowship
(#71199514104). We also thank Mr. X. Calyptratus for technical assistance.
References
1. Borders, D. B.; Doyle, T. W. Enediyne Antibiotics as Antitumor Agents; Marcel Dekker: New York, 1995.
2. (a) Nicolaou, K. C.; Dai, W.-M. Angew. Chem., Int. Ed. Engl. 1991, 30, 1387±1416. (b) Grissom, J. W.;
Gunawardena, G. U.; Klingberg, D.; Huang, D. H. Tetrahedron 1996, 52, 6453±6518.
3. Yoshida, K.-I.; Minami, Y.; Otani, T. Tetrahedron 1994, 35, 5253±5256.
4. (a) Semmelhack, M. F.; Neu, T.; Foubelo, F. J. Org. Chem. 1994, 59, 5038±5047. (b) Semmelhack, M. F.; Neu, T.;
Foubelo, F. Tetrahedron Lett. 1992, 33, 3277±3280.
5. (a) Maier, M. E.; Greiner, B. Liebigs. Ann. Chem. 1992, 855±861. (b) Schmittel, M.; Kiau, S. Chem. Lett. 1995, 55,
953±954.
6. (a) Kim, C.-S.; Diez, C.; Russell, K. C. Chem. Eur. J. 2000, 6, 1555±1558. (b) Choy, N.; Russell, K.; Alvarez, J.;
Fider, A. Tetrahedron Lett. 2000, 41, 1515±1518. (c) Choy, N.; Russell, K. C. Heterocycles 1999, 51, 13±16.
(d) Kim, C.-S.; Russell, K. C. J. Org. Chem. 1998, 63, 8229±8234. (e) Kim, C.-S.; Russell, K. C. Tetrahedron Lett.
1999, 40, 3835±3838.
7. Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165±195.
8. Jiang, X. K. Acc. Chem. Res. 1997, 30, 283±289.
9. Swain, C. G.; Lupton, E. C. J. Am. Chem. Soc. 1968, 90, 4328±4337.
10. Compounds 1a±f and 2a±f all gave the expected spectral data.
11. Grissom, J. W.; Calkins, T. L.; McMillen, H. A.; Jiang, Y. J. Org. Chem. 1994, 59, 5833±5835.
12. General procedure for kinetic studies:
A solution was made by dissolving the enediyne, methyl-3,4-
diethynylybenzoate (1c; 2.88 mg, 0.012 mmol), 1,4-cyclohexadiene (1.1 mL, 1.2 mmol, 100 equiv.) and an
anthrathene internal standard (3.1 mg, 0.012 mmol) in chlorobenzene in a 2 mL volumetric ¯ask. In each of 10
capillary melting point tubes was placed 40 mL of the solution. The tubes were cooled to ^78ꢀC and degassed
under high vacuum and sealed at same dead volume heights. Each tube was placed in an oil bath maintained at
170ꢀC and removed at proper time intervals to obtain a suitable half-life for our studies. The content of each tube
was analyzed by analytical HPLC (water:acetonitrile, isocratic 30:70) at 315 nm. The percent starting material
remaining was plotted versus time and the resulting curve was ®tted to a single exponential decay rate equation
using the KaleidaGraph program on a Macintosh PC.
13. Psi Plot version 6.0. Poly Science International, Salt Lake City, UT, 1999.