DAEHAN HWAHAK HWOEJEE (Journal of the Korean Chemical Society) Vol. 20, No. 2, 1976 Printed in Republic of Korea NOTE

Diethyl Azodicarboxylate 에 의한 유기화합물의 수소이탈 반응기구 田 地 鎭•후 피 *터

경북대학교 공과대학 응용화학과 （1975. 9. 28 접수 ）

Mechanism of Dehydrogenation of Some Organic Compounds by Diethyl Azodicarboxylate

Moo-Jin Jun and Peter P. Fu*

Department of Applied Chemistry Kyungpook National University, Taegu, Korea (Received Sept. 28, 1975)

Diethyl azodicarboxylate (DAD) was reported below to illustrate the tentative concerted me­ to e任 ect photochemical dehydrogenation of iso­ chanism : propyl and ,1,2 and to effect 아 13-어 “9 nonphotochemical oxidation of many organic 아坏아 ! + DAD ----* H H —> CH^SHO + DHAD ( ! ) compounds for the last decade. It reacts smoo­ C가 ^。弍-心 Y3겄冲 thly with a variety of primary or secondary , mercaptans, anilines, and hydrazoben­ 部너好 zenes to form or , , + DAD 一 *- H H —►(利 ■城 0 + DHAD (2)- and azobenzenes. And also it takes place hyd­ C이니 成 )2거 '<溶才 妇 rogenation to form diethyl hydrazodicarboxylate R岫夕

(DHAD).3,4 DAD reacts also with hydroxy­ RNICH0+ DAD --- H H —-— RNCO + DHAD (3}- amines to give nitroso compounds5 and with C2H5O2CT回나 -COK간付 several 7V-monosubstituted formamides to give Based on the molecular orbital approach to corresponding isocynates (isolated as urethanes).6 chemical reactivity postulated by Fukui, ^11 It is interesting to try to reveal the thermal Mulliken,12,13 and Brown,14 etc., three require­ mechanism of this reaction. ments should be satisfied in order to confirm a At first glance, the mechanism might be sug­ concerted mechanism, namely: (1) significant gested "concerted," since the mechanism of the large coefficients of the involved atoms in highest similar reaction, Diels-Alder reaction, is con­ occupied molecular (HOMO) and/or lowest un­ certed. For instance, , phenyl hydroxyl­ occupied molecular orbital (LUMO), (2) same amine, and N—substituted formamide are used symmetry of the involved atoms in HOMO and/ or LUMO, and (3) close energy of the HOMO- * Department of Chemistry, University of Illinois at Chicago Circle Chicago, Ill. 60680, U. S. A. LUMO pair. The first requirement must be sa­

—166 Diethyl Azodicarboxylate 에 의한 유기화합물의 수소이 탈 반응기 구 167

tisfied not only to a concerted mechanism, but hydroxyamine and ^-substituted formamides, also to all the other mechaisms. If the coefficient the coefficients of all the concerned hydro용 ens of an atom in a specified molecular orbital is in HOMO and LUMO are found zero from Table zero or extremely small, it means that this atom 2 and 3. Based on this result and 요 frontier has no contribution to the molecular orbital at electron theory, " it should rule out not only all. According to the ^frontier electron theory" the tentative concerted mechanism (2) and (3), proposed by Fukui, et al. 7,8 this atom could but also all the other me사 nanisms which involve not be involved in the first step of any kinds of those concerned hydrogens. reactions. Consequently, we examined the ten­ The lowest energy conformations were used tative mechanism (1), (2), and (3) from this for the MO calculation.17 Several phenyl hy­ point of view. droxyamine conformations were chosen and com­

The molecular orbital calculation of diethyl puted in order to determine the lowest energy- azodicarboxylate was done by simple Huckel one. It was found that the conformation with Molecular Orbital Theory (HT) by Zweig, et \ al.15 The required information was given by the skeleton of------planar is the

this author. Calculation of ethanol and phenyl H hydroxy amine was done by Extended Huckel Theory (EHT) approach. Bond parameters were Table 2. Coefficients of HOMO and LUMO of the used in our choice of dimensions and geometry.16 concerned atoms in ethanol and phenyl hy droxy For easier computation, methyl, ethyl, and amine, determined by EHT method.

isopropyl formamides were used by CNDO me­ Compounds Concernec atoms HOMO LUMO thod. The useful parts of the HT, EHT, and H (8) H⑸ S 一 . 1143 .0382 CNDO res니 ts are shown in Table 1, 2, and 3. 1 / CH3 — C—0 ⑼ H⑹ S . 0413 The coefncients of the dehydrogenated hydro­ 1 \ -.6957 gens were examined first. It is found from Table H H C⑻ S . 0519 .1465 (5) (6) 2 that one hydrogen in an ethanol molecule has 已 4275 .4001 I s .0849 .7940 significantly large coefficient while another pos­ 1 .0479 .0888 sesses extremely small one, both in HOMO and 0⑼ S -.0248 一 .0758 LUMO. However, both values are required to 已 .5464 .4263 be significantly large in order to meet the ten­ % 1689 .5678 tative concerted mechanism (1). As to phenyl 已 -.2089 .3801

宜 (6) S . 0000 .0000 (14) C6H5-N-O(15) ⑺ Table 1. Coe伍 cients of HOMO and LUM。of the 1 \ H s . 0000 .0000 nitrogens in diethyl azodicarboxylate, determined by H H N(14) s . 0000 .0000 HT method. (6) (7) Px . 0000 .0000 八 1 1 Concerned Compound 丨 atoms HOMO LUMO Py . 0000 .0000 -.7020 .0238 C2H5O2CN = N⑴4 .378 一 .526 (1) 0(15) S . 0000 .0000 NCO2C2H5 N(2)“ .378 .526 (2) 已 .0000 .0000 a. the numbers were used to indicate the concerned Py ・ 0000 .0000 atoms of the compound on the first column. Same 已 .3476 -.0170 indication was also used in Table 2 and 3.

Vol. 2, No. 20, 1976 168 田地鎭 •후 피 터

Table 3. Coefficients of HOMO and LUMO of the Hoffman, et al.18 concerned atoms in methyl, ethyl, and isopropyl As it is w사 1 known, no matter which MO formamides, determined by CNDO method. calculation method (such as HT, EHT, CNDO, Compounds Concerned atoms HOMO LUMO IMDO, etc. ) is used to compute the compounds O H⑻ S .0000 .0000 concerned in this paper, the HOMO and LUMO ⑶z are contributed only by conjugated ^-electron CH3------C(2) H⑼ s .0000 .0000 I \ system (composed from p orbitals). Thus, if a H H C⑵ s 0000 .0000 ⑻ (9) hydrogen atom is in the plane of this conjugated Px .0000 .0000 % 一 .0000 .0000 system, the contribution of the hydrogen atom 已 .1498 一 .7669 to the positive sign lobe of the conju응 afed p

N⑶ s -.0000 -.0000 orbitals is equal to the negative sign lobe. The Px .0000 .0000 net contributions are then cancelled out each p, .0000 .0000 other, and the coefficients of this hydrogen in ps 一 6959 . 2904 HOMO and LUMO 사 lould be zero. This theo­ H(8) S .0000 0000 retical consideration is consistent with the result O H⑼ s .0000 .0000 of our calculation. If the planar skeleton of a (3n — secondary formamide is distorted, such as rota­ C(2) s .0000 -.0000 h (8 pr 0000 -.0000 tion of the - ----bond of the formyl 후 roup, one pv 一 .0000 0000 hydrogen will gain a very small coefficient value 一 .1442 .7633 in HOMO and/or LUMO, but another hydrogen

23) s -.0000 .0000 is still zero. Therefore, it is neither possible to 已 .0000 .0000 go through a concerted mechanism. Conse­ p, .0000 .0000 quently, the tentative concerted mechanism 巴 .6899 -.2869 (1), (2), and (3) are complexly ruled out. O H(8) S -.0183 -.0324 (3) Z A free radical mechanism was then taken into (CH3)2CH-N-O(2) H(9) S 0300 .0030 consideration. No obvious information of MO i \ c⑵ s .0025 一 .0064 calculation could be obtained to propose a free (8) (2) 已 -.0001 -.0035 radical mechanism at this stage of our work. .0188 一 .0047 However, based upon our experimental results 已 . 1326 .7639 obtained, we successfully applied thermal de­ 0⑶ s .0110 .0307 hydrogenation of DAD into the corresponding 巴 —.0166 .0075 isocyanates. When the initiator, 2, 2z-azobis-2~ Py —.0225 .0098 methylpropiontrile was separately added to the Pz -.6766 -.2958 mixture of A^-phenyl formamide with DAD and of TV-cyclohexylformamide with DAD, the reac­ most stable (lowest energy) one. The most stable tion rate was not increased at all. As the results conformation of TV-substituted formamides was of these experiments, a free radical mechanism is not likely. the skeleton of 一 了一C〈 planar, cal­ The ionic mechanism was then considered. The “overlap and orientationM principle proposed culated by EHT and CNDO/2 methods by by Mulliken12 has been considered only the

Journal of the Korean Chemical Society Diethyl Azodicarboxylate 에 의 한 유기 화합믈의 수소이 달 반웅기 구 169 overlap interaction between the HOMO of the MO, the oxygen of ethanol, the nitrogen of donor and the LU MO of the acceptor since it phenyl hydroxyamine, and the nitrogen of me­ oxidizes a solution of in glacial thyl, ethyl, and isopropyl formamides have the to form iodine quantitatively, 4, i9,20 largest coefficient with the value 丨.54641, and reacted with alkali metal to give a dianion.15 1.7020b 1.69591, |.6899|, and 676이 , resp­ The HT clculation indicates that DAD is un­ ectively. Accordingly, the ionic mechanism usual in possessing a vacant bonding orbital (LU- is seem to be plausible and is illustrated as MO). 15 From these facts, it would be expected follows： that DAD might have a high tendency to abs­ tract electron (s) while ethanol, phenyl hydro­ Aromatic and aliphatic amines, which are xyamine, or N^secondary formamides etc. should electron donor, reacted with DAD to give ad- be an electron (s) donor. In this view, the LU- ducts21^26 similar to the intermediate of (B) in MO of DAD and the HOMO of the others were mechanism (5) and of (C) in (6). By compa­ chosen. According to the ^frontier electron the­ rison with these experimental results, it is rea­ ory, "the element of DAD which has the largest sonable to propose the intermediates of (B) and coefficient (absolute value) in LUMO, and the (C) (in) the ionic mechanism. The intermediate element of the other compounds which has the of (C) is an isoesters of carbamate. It is well largest coefficient (absolute) in HOMO should known that the products of thermal decomposi­ be taken into consideration. Both nitrogens of tion of monosubstituted carbamates are isocyan­ DAD have the largest value (. 526) in LUMO. (The coefficients of other elements of DAD are ates and alcohols. 27 Therefore, the production not shown in Table 1 due to the limit of space. of isocynate and DHAD obtaining from the in­ Same reason is applied to other compounds in termediate (C) in mechanism (6) is also very

Table 2 and 3. logical. From Table 2 and 3, it is found that in 너 0-

Conclusion

尸 ----- ___ ch3-ch2-?)^) 1. The mechanism of thermal dehydrogenation 아 3釦 2-어너 + C2H5O2c'N=N-CO2C2H5 N-N C2H5O2C CO2C2H5 of organic compounds by diethyl azodicarboxyl- H ate seems not likely to be a concerted one, CH3- H N、 -- a Ch-tCHO + DHAD (4) c C02C2H5 $ C2H58C 5 which is proved by the molecular orbital ap­ - (A) e >0 proach. 甘 5시 潮旳 ' --- * - 4. (J + C기一 5。2히斗 =矣 2対5 一 . N-N® 2. It is neither a free radical mechanism based 钏02# 必离 시 K으 H on the experimental results. N-N CeH5NO + DHAD (5) CSH502C/ XC02C2H5 3. Instead, an ionic mechanism is plausible. (B) • C夕 It could be fully explained by molecular orbital

RNHCHO + C才切 2圳枭乂啓才 0 theory. The proposed intermediate (an adduct) C2H5O2C/ 、C02기니 5 R食 could be assumed logically by comparison with • N RNCC + DHAD (6) some reported reactions,28 and also the inter­ C2H502Cz、C0zC간切 (c) mediate could be accounted for the formation of the product.

Vol. 20, No. 2, 1976 170 田 琪 鎭 •후 피 터

13. R. S. Mulliken, Rec. Trav. Chim., 75, 845 REFERENCES (1956). 1. G. 0. Schwenck and H. Formancek, Angew. 14. R. D. Brown, J. Chem, Soc., 2232 (1959). Chem., 70, 505 (1958). 15- A. Zweig and A. K. Hoffman, J. Amer. Chem. 2. R. C. Cookson, I. D. Stevens and C. T. Watt, Soc., 85, 2736 (1963). Chem. Comm., 259 (1965). 16. "Table of Interatomic Distances and Configuration 3. F. Yoneda, K. Suzuki, and Y. Nitta, J. Amer. in Molecules and Ions," Chemical Society, London Chem. Sot— 88, 2328 (1966). (1958). 4. F. Yoneda, K. Suzuki and Y. Nitta, J. Org. 17- The computer program obtained from the Quantum- Chem., 32, 727 (1967). Chemistry Program Exchange, Program 141 (P. 5. a) E. C. Taylor and F. Yoneda, Chem. Comm., A. Bobosh, Indiana Univ., Bloomington, Indiana) 199 (1967)： b) J. H. Boyer and P. Frints, J. was used for our calculation. Org. Chem., 33, 4554 (1968). 18. J. F. Yan, F. A. Momany, R. Hoffman and H. 6. P. P. Fu and J. H. Boyer, J. Chem. Soc., Perkin A. Scheraga, J. Phys. Chem., 74, 420 (1970). I, 2246 (1974). 19. R. Stolle, Ber.. 45, 273 (L912). 7. K. Fukui. "Molecular Orbitals in Chemistry, 20. R. Huisgen and F. Jakob, Ann., 590, 37 (1954) Physics and Biology," ed. by P. O. Lowdin and 21. O. Diels and P. Fritsche, Ber., 44, 3018(1911). B. Pullman, Academic Press, New York, 1964. 22. O. Diels and M. Paquin, ibid., 46, 2000 (1913). 8. K. Fukui, T. Yonezawz, and H. Shingu, J, 23- O. Diels, Ann., 429, 1 (1922). Chem. Phys., 20, 722 (1952). 24. O. Diels and G. Behnen, Ber., 56, 561 (1923). 9- K. Fukui, T. Yonezawa, C. Nagata, and H. 25. O. Diels, ibid., 56, 1933 (1923). Shingu, ibid., 22, 1433 (1954). 26. G. W. Kenner and R. J. Steduan, J. Chem. Soc.，- 10. K. Fukui and H. Fujimoto, Bull. Chem. Soc. 2089 (1952). Japan, 41, 1989 (1962). 27. P. Adams and F. A. Baron, Chem. Rev., 65, 11. K. Fukui and H. Fujimoto, ibid., 42, 3399 567 (1965). (1969). 28. G. H. Kerr, O. M. Cohn, E. H. Pullock and 12. R. S. Mulliken, J. Amer. Chem. Soc., 74, H. Suschitzky, J. Chem. Soc., Perkin I, 1614 811 (1952). (1974).

Journal of the Korean Chemical Society