CHEMISTRY

E. Ya. GREN and Academician of the Academy of Sciences of the Latvian SSR G. Ya. VANAG

2-PHENYL-4,5,6,7-TETRAHYDROINDANDIONE-1,3

Continuing the study of the relationship between structure and properties in the indandione-1,3 series, we recently reported investigations of the prototype of the indandione system—cyclopentene-4-dione-1,3 \((^{1,2})\). The present work is devoted to a new analogue of indandione-1,3, namely, 2-phenyl-4,5,6,7-tetrahydroindandione-1,3 (I).

This compound had already been obtained in 1935 by Berlingozzi and Senatori \((^3)\), but they did not study it more closely and assigned to it an incorrect structure. We obtained I by a method analogous to that used for the synthesis of 2-arylindandiones-1,3, namely, by thermal condensation of \(\Delta^{1,2}\)-tetrahydrophthalic anhydride (III) with phenylacetic acid, obtaining the corresponding benzal-tetrahydrophthalide (II), which was rearranged with sodium methylate into 2-phenyl-4,5,6,7-tetrahydroindandione-1,3 (I). II is also obtained by condensation of \(\Delta^{2,3}\)-tetrahydrophthalic anhydride (IV) with phenylacetic acid, and here, evidently, the long-known thermal isomerization of IV into III first takes place \((^{4,5})\). This circumstance was not taken into account by Berlingozzi and Senatori, who carried out the condensation precisely with IV \((^{3,6})\) and thus arrived at an incorrect conclusion concerning the position of the double bond both in II and in I.

\[ \begin{array}{ccccc} \text{(IV)} & \xrightarrow{\,t^\circ\,} & \text{(III)} & \xrightarrow[\,-\mathrm{H_2O};\ -\mathrm{CO_2}\,]{+\mathrm{C_6H_5CH_2COOH}} & \text{(II)} \xrightarrow{\ \mathrm{CH_3ONa}\ } \text{(I)} \end{array} \]

2-Phenyl-4,5,6,7-tetrahydroindandione-1,3 (I) is a light-yellow substance, dissolves in alkalis with a dark-violet coloration; it gives a colorless dioxime. Owing to the stability of I toward bases, we were able readily to carry out alkylation reactions of it in an alkaline medium, which was impossible with cyclopentene-4-dione-1,3 \((^{1,2})\). Heating I with an equivalent amount of sodium methylate and methyl iodide gave, in nearly quantitative yield, the 2-methyl derivative (V). This reaction proceeds smoothly also at room temperature and even on shaking I with methyl iodide in the presence of calcium oxide. Reduction of I with zinc in acetic acid or catalytic hydrogenation in the presence of Raney nickel gives 2-phenylhexahydroindandione-1,3 (VI). Along with VI, a thick oil was also isolated, evidently a product of further reduction.

\[ \begin{array}{ccc} \text{(VI)} & \xleftarrow{\ +\mathrm{H}\ } & \text{(I)} \xrightarrow[\mathrm{CH_3ONa}]{+\mathrm{CH_3J}} \text{(V)} \end{array} \]

The IR spectra of the compounds obtained were recorded (see Table 1). In the spectrum of I only normal carbonyl bands and a double-bond band were found. A similar spectrum was obtained for V. It is interesting to note that the carbonyl frequencies of I and V are appreciably lower than those found for 2-phenylindan-1,3-dione and 2-methyl-2-phenylindan-1,3-dione, which is a general regularity when an aromatic ring is replaced by a double bond. The similarity of the IR spectra of I and V, as well as the absence of absorption in the region of the stretching vibrations of O—H bonds, excludes the possibility of the existence of I in the enol form and confirms its diketone structure. In the spectra of solid I and V a weak band appears near 3400 cm\(^{-1}\), which is evidently an overtone of the carbonyl vibrations, as in (1).

2-Phenylhexahydroindan-1,3-dione (VI), on the other hand, differs greatly from I both in chemical properties and in structure. The IR spectrum of VI is characteristic of completely enolized \(\beta\)-diketones and does not contain a single band that could be assigned to \(\nu\) CO of the dicarbonyl form VI\(^{(7,8)}\). At the same time, in the spectrum of solid VI a broad band is observed at 2640–2660 cm\(^{-1}\), found for many enolized \(\beta\)-diketones with a strong hydrogen bond (for example, dimedone \(^{(9)}\)). This confirms the existence of solid VI exclusively in a strongly associated enol form. It was likewise not possible to detect the existence of the dicarbonyl form of VI in a solution of dichloroethane with alcohol. Unfortunately, we were unable to study VI in a less polar solvent because of its poor solubility.

The IR spectra show that compounds of the cyclopentene-4-dione-1,3 \(^{(1)}\) and 4,5,6,7-tetrahydroindan-1,3-dione systems, like indan-1,3-diones \(^{(10)}\), are not tautomeric and exist exclusively in the dicarbonyl form. By contrast, cyclopentane-1,3-diones \(^{(2)}\), hexahydroindan-1,3-diones \(^{(11)}\), and \(\Delta^{5,6}\)-tetrahydroindan-1,3-diones \(^{(2)}\) are tautomeric substances, with characteristic existence in the enol form (as in six-membered cyclic \(\beta\)-diketones).

To confirm the structure of II, the IR spectra of II and 3-benzalphthalide were recorded. In addition to the high frequency of the lactone carbonyl \(^{(12)}\), two double-bond bands also appear in the spectrum of II. The higher of them (1660 cm\(^{-1}\)), also found in the spectrum of 3-benzalphthalide, evidently belongs to the vibrations of the exocyclic double bond; and the lower, observed only in the spectrum of II, to the vibrations of the endocyclic bond.

Experimental part

3-Benzal-4,5,6,7-tetrahydrophthalide (II). a) In a flask equipped with a thermometer and a descending tube for distillation of the water liberated in the reaction, 45 g of \(\Delta^{1,2}\)-tetrahydrophthalic anhydride (III) \(^{(13)}\), 2.5 g of anhydrous sodium acetate, and 45 g of phenylacetic acid are heated on a metal bath first to 180°, then gradually to 200° (in the flask). During the reaction, two additional portions totaling 2.5 g of sodium acetate are added. When about half of the calculated amount of water has distilled off, the temperature is gradually raised to 210°. After the calculated amount of water has been evolved (about 5.5 ml), the orange-red mass is cooled to 100–110°, poured into 200 ml of methanol, and left overnight in a refrigerator. The orange-yellow crystals are washed with a small amount of cold methanol. Yield 30.5 g (46%) of crude II, quite suitable for obtaining I. A little more material was isolated from the filtrate, so that the total yield increases to 55%. Mp 117°.

b) Similarly, from 0.43 g of \(\Delta^{2,3}\)-tetrahydrophthalic anhydride (IV) \(^{(5)}\), 0.5 g of phenylacetic acid, and 0.1 g of anhydrous sodium acetate, by heating at 180–190° for 20 min, the same product II was obtained. Mp 117–118°. For purification, crude II is shaken with a warm dilute solution of sodium methylate, and the residue is crystallized from methanol or petroleum ether.

White crystals, m.p. 118°.

Found, %: C 79.72; H 6.30
$\mathrm{C_{15}H_{14}O_2}$. Calculated, %: C 79.64; H 6.23

Dioxime was obtained from 0.5 g of substance I, 0.4 g of $\mathrm{NH_2OH\cdot HCl}$, and 0.5 g of $\mathrm{CH_3COONa}$ in dilute methanol. White crystals, 0.25 g, m.p. 260° (from methanol).

Found, %: N 11.01
$\mathrm{C_{15}H_{16}O_2N_2}$. Calculated, %: N 10.93

2-Methyl-2-phenyl-4,5,6,7-tetrahydroindandione-1,3 (V). 3 g of I are boiled for 3 h in 50 ml of methanol with sodium methylate (from 0.31 g of sodium) and 1.2 ml of $\mathrm{CH_3I}$. After only a few minutes the dark-violet color of the solution changes to orange. The cooled solution is poured into water and left in a refrigerator. The crystals that separate (3 g; 94%) are recrystallized from methanol (or $\mathrm{CH_3OH}$ + water). Yellow crystals, m.p. 94–95°.

Found, %: C 80.29; H 7.00
$\mathrm{C_{16}H_{16}O_2}$. Calculated, %: C 79.98; H 6.71

2-Phenylhexahydroindandione-1,3 (VI). a) 1.5 g of I in glacial acetic acid are heated on a boiling water bath with zinc dust until the solution no longer gives a violet color with alkali (characteristic of I), and poured into water. The white precipitate crystall-

Table 1

Infrared absorption spectra

Note. a — weak band near 1600 cm$^{-1}$; b — weak band near 3400 cm$^{-1}$; c — recorded in the interval 2500–3700 cm$^{-1}$; d — because of solvent absorption this interval was not accessible.

All spectra were recorded on a single-beam IKS-12 instrument with a NaCl prism, in most cases in the intervals 1500–1760 and 3000–3700 cm$^{-1}$. Spectra of solid substances were obtained in paraffin oil. The band values are given in reciprocal centimeters; in parentheses is their relative intensity in absorption percent.

crystallized from dilute alcohol or hexane ether. Yield 0.3 g (20%), m.p. 229–230° (from CHCl₃).

b) 0.1 g of I in alcoholic solution is hydrogenated in the presence of Raney nickel and a small amount of KJ. The alcohol is distilled off in vacuo and the residue is crystallized from CHCl₃. M.p. 229°; with the product obtained by method a), it gives no depression of the melting point.

Found, %: C 78.79; H 7.23
C₁₅H₁₆O₂. Calculated, %: C 78.94; H 7.06

Institute of Organic Synthesis
Academy of Sciences of the Latvian SSR

Received
7 IV 1961

CITED LITERATURE

¹ E. Ya. Gren, G. Ya. Vanag, DAN, 133, 588 (1960).
² E. Ya. Gren, G. Ya. Vanag, Izv. AN LatvSSR, No. 5 (1961).
³ S. Berlingozzi, V. Senatori, Gazz. Chim. Ital., 65, 424 (1935); Chem. Abstr., 92 (1936).
⁴ R. Bruckles, Chem. Rev., 57, 641 (1957).
⁵ A. Baeyer, Lieb. Ann., 258, 145 (1890).
⁶ S. Berlingozzi, Gazz. Chim. Ital., 61, 886 (1931); Chem. Abstr., 2728 (1932).
⁷ P. Wieland, K. Miescher, Helv. Chim. Acta, 33, 2220 (1950).
⁸ S. Bernstein, M. Heller et al., J. Am. Chem. Soc., 82, 3685 (1960).
⁹ W. Chan et al., J. Chem. Soc., 1956, 3495.
¹⁰ D. G. O’Sullivan, J. Chem. Soc., 1960, 3278.
¹¹ B. F. Kucherov, L. N. Ivanova, DAN, 131, 1077 (1960).
¹² J. F. Grove, H. A. Willis, J. Chem. Soc., 1951, 877.
¹³ J. C. Godfrey, R. A. Barnes, J. Am. Chem. Soc., 80, 3902 (1958).
¹⁴ Ya. F. Freimanis, G. Ya. Vanag, ZhOKh, 20, 3369 (1960).
¹⁵ M. Bailey, J. Am. Chem. Soc., 78, 3828 (1956).