Tmahrdron. VoI 27. pp 2191 IO 2202 Pcrgamon Preu 1971 Pnnted !n Grul ham

THREE-CHROMOPHORE COMPOUNDS-III’ STEREOCHEMICAL STUDIES OF 3,5-DISUBSTITUTED

CYCLOHEXANONES

A. Y. MEYER and J. SCHLBINGER

Department of Organic Chemistry, Hebrew University. Jerusalem

(Received in rhe UK I5 October 1970; Accepted/or publication 5 January 1971)

Ahatrrt-The stereochemistry of three 3.5-diarylcyclohexanones. prepared by catalytic reduction of the corresponding 3.5diaryL2cyclohexenona. has been derived from dipole moment measurements. as well as from NMR spectra and quantum-chemical relationships. The two phenyl groups in the 3Sdiphenyl- cyclohcxanone so prepared arc cis. occupying equatorial positions on a somewhat distorted chair-form ring. The 3-pchlorophcnyl-5-phcnylcyclohexanone obtained is rruns. with equatorial chlorophenyl and axial phenyl. 3.5~Di@chlorophenyl)yclohexanone is also frans; it adopts a ncxible structure. with an angle of +35” between the planes C6C,C, and C,C,C,. Meenvein-Ponndorf-Vcrlcy reduction converts the latter compound to a chair-form alcohol with an equatorial OH group.

IN PART I of this series* we reported the preparation of three 35diarylcyclohexanones (Is-c) by catalytic reduction of the corresponding parent olefins (1Ia-c). Compound la was obtained3 by Pd/C-catalyzed hydrogenation of Ha. whereas the chlorine-con- taining molecules, IIb and Ilc, were reduced with rhodiumchlorotris(triphenyl- phosphine) to Ib and Ic; with this catalyst. IIa produced the same saturated ketone as with Pd/C.

The olefinic compounds II are known4 to carry the 5-aryl group in a quasiequa- torial position. A way to elucidate the configuration, cis or trans. of the saturated products I is the reduction of the carbonyl to an OH group(iI-III): a cis-structure(lV) of II may lead to two epimeric alcohols (Va. b). whereas a trans-form (VI) of II can yield only one (VIIa EVIIb). Balasubramanian and D’Souza have shown3 by this method that the two phenyl groups in la are cis to each other. Among the reducing agents examined by these authors, aluminium isopropoxide in P-propanol-which provides a 1 :l mixture of the two alcohols---seems to be the least stereospecific. and therefore the most useful.

We report some experimental data concerning the stereochemistry of the three saturated ketones Iac, and also extend the theoretical analysis of the flexible cyclo- hexanone* to cover those of its derivatives that carry electron-attracting substituents at positions 3 and 5. Structural assignments are based mainly on the comparison of experimental dipole moments with the appropriate vectorial sums of group moments.5*6 Geometrical constructions follow either Corey and Sneen’ (rigid ring)

l In six-membered rings.* we distinguish the rigid (“chair”. D,, symmetry) from the flexible forms. the latter being sub-divided further into distorted boats and distorted twist forms depending on wh:ther they are closer to the regular boat (Cr.) or to the twist form (D2). The flexible forms of cyclohexanonc are characterized either by the pseudorotational phase’ or by the dihedral angle1 C,C,C&,C&; the prow-keto boat of phase 0 and dihedral angle 41’. is shown in Fig 1. For any form of non-zero phase. substituents are denoted e’ (quasiequatorial) or a’ (quasiaxialL depending on whether they become equa- torial or axial at phase 0. WC use symbols such as e; to denote an e’-substituent at position 3. etc.

2191

2192 A. Y. MEYER and J. SCHLISINCZR

or our previous data’ (flexible ring). Group moments were as follows: C-phenyl. 0*4D, C-OH, 1.65. C-pchlorophenyl and C-F, 1.9, C=O, 29 D.+

Structural studies Dipole moments, measured and computed, are recorded in Table 1. c&3,5-

Diphenylcyclohexanone (Ia) is obviously diequatorial, although some deviation from regular geometry cannot be completely excluded (difference of ca. O-2 D between the computed and experimental values). As for 3-pchlorophenyl-5-phenylcyclohexanone (Ib), the comparison indicates a trans-geometry (e,a,); indeed. the similarity of the NMR of Ia and Ib renders alternative interpretations (e.g., flexible ring) unlikely. In contrast with these compounds, the dipole moment of 3,5-di(pchlorophenyl)cyclo- hexanone (Ic) fits none of the values computed for rigid or “reversed chair”” con- formations. and its substantial difference (1 D or more) from ail computed values excludes also irregular varieties l4 of these geometries. Equally. Ic is quite dissimilar in NMR to Ia or Ib.

TABLE I. EXPERIMENTAL AhP COMPUTED DIPOLE MOMENTS

Computed values (Dcbye)

ws a3a3 e3a) a3e5 Exptl

Ia. chair 2.46 3.21 2.86 269

Ib. chair (3-Cl-p. 5-q) 1.99 3.18 2.44 3.79 2.44

Ic. chair 084 5.54 3.28 2.32

Ic. “reversed chair” 0.42 6.00 3.23

Ilk. hydroxyl equatorial 0.76 4.62 2.48 2.35

Ilk. hydroxyl axial 1.83 5.45 3.42

The carbonyl-reduction of Ic with aluminium isopropoxide in 2-propanol (as pointed out, a non-stereospecific reagent; see Experimental) provides only one alcohol (111~) with equatorial OH.?

The trans-Ar. Ar relationship, thus suggested, is confirmed by the dipole moment which identifies 111~ as cis, truns-3e,5a-di(pchlorophenyl)-le-cyclohexanol, so that the ketone Ic is trans-3.5-di(p-chlorophenyl)cyclohexanone. The nearest alternative (still off by ca. @5 D) is a,e,e,, not in line with the equatorial assignment for the OH. Moreover, an a,e,e,-alcohol implies that Ic is cis and therefore expected3 to furnish 1 :l mixture of two alcohols, a,e3e, and e,e,e,. Let us add that the cis,trans-alcohol is obtainable directly from IIc, but its low yield and the large bulk of by-products in the process (Experimental) rob this reaction of much of its stereochemical interest.

Having discarded, for Ic. chair-like conformations, we examine now the flexible

l Reduced moments”’ of the following one-dipole molecules: to1ucoe.l’ @43 D. McOH.‘~ I.70 D.

pchlorophenylcyclohexane.” I.96 D. cyclohexanonc.” 3.08 D.

7 The assignment is based upon the NMR spectrum of the proton. geminal to the OH: 6 (CDCI,.

60 MH,) 4.3S3.66 ppm from TMS. width at half-height 22c/s. It rests upon the observation” that the

coupling constant between neighbouring axial hydrogens is larger than between hydrogens in other

orientations. from which it follows 16. *’ that multiplets corresponding to highly-couplbd’axial protons are

wider than others. Relevant referencesI are the two isomers of 5-phenyl-2_cyclobexe-l-ol: the compound

of C-OH and a-H (geminal to OH) has 6 4.35 (Ccl,). width at half-height 18 c/s. against 4.20 and 8 for the

other.

Thrcc-chromophorc compounds-III 2193

forms characterizing each’ by the angle 0 between the planes C6C1Cz and CICzCJ (Fig 1). Fig 2 (full curves) displays the dependence of the computed dipole moments* upon 8. and selects a;e; at 8 z 35” (a somewhat distorted “prow-keto”’ form) as the most probable structure for Ic. The nearest alternative (u z 2G D) is e;a; at 8 x 41”. A choice between these two, based upon quantum-chemical grounds. is deferred to the next section.

FIG 1. Geometry of the flexiblccyclohcxanonc structure: coordinate system [xyz] and the

distortional angle 8 (tl = 41” shown)

Ib and Ic have substituents of similar steric bulk so that their conformational differ- ence stems most probably from dipole-dipole repulsion.“’ A rough estimate of this factor may be obtained by approximating the total interaction in the three-dipole molecules as sum of the three two-dipole contributions (CO-Ar,, CSAr,. Ar-Ar,). given each by19

E = plpz (cos 1 - 3 cos aIaJr3

Here, r is the length of a vector R connecting the dipoles uI and u2, 1 the dipolar angle. a, and a2 the angles (t.tl. R) and (u2. R). respectively. It is seen immediately

l Owing to the particular geometrical construction a;e; and c;a; are identical only at 0 = 41”

(“regular boat”)

2194 A. Y. MEYER and J. SCHLESINGER

that the interaction in Ic, which carries two strong dipoles, is more pronounced than in Ib. containing only one. In order to verity that the prow-keto boat represents less repulsion than the chair, we disregard E(Ar,-Ar,) and the term in (a,a,)-these being almost identical in the two conformations-and note that the cosines of the dipolar angles ,r(CCB-Ar) are lower in the first, than in the second possibility: boat. X. 150” 20’ (cos 1, = - 0.87). xe 89“ 39’ (O*Ol), chair xe 121” 35’ (- 0*52), 1. 68” 58’ (0.36).

30- (@-

26-

24-

FIG 2. Calculated dipole moment of 3.5-di@chlorophenyl)yclohexanone (in Dcbyc units)

as function of the distortional angle. Bold curves: group-moment summation. Dashed curves:

quantum-chemical evaluation

For later reference, we show in Fig 3 the trends of the three-dipole mutual interac- tion, approximated as sum of three two-dipole contributions. A hypothetical flexible form of 3,5-difluorocyclohexanone serves as a model, with vector R assumed to connect the mid-points of CJ-F and C5-F. Let us remark, in anticipation of the subsequent discussion, that the trans-structure represent&deed less interaction than the cis. although the conformation of least dipole-dipole repulsion is not the

Threechromophore compounds111 2195

one we suggest for Ic. but rather a;e; at 0 = - 20” (we shall show that this latter form is destabilized by severe steric interaction). An interesting result, concerning the cis-configuration. is that the repulsive forces are lower in the diaxial, than in the diequatorial conformation.

-25 -15‘ -5 5 I5 25 35

e

FIG 3. Computed dipole4ipole repulsion as function of the distortional angle. in four stereo-

chemical varieties (hypothetical) of 3.5difluorocyclohexanonc

All valence-electron results

In view of the unusual shape assigned to 3.5di(p-chlorophenyl)cyclohexanone (1~). and its configurational dissimilarity with the diphenyl and phenyl-pchlorophenyl compounds (Ia, b), it was of some interest to carry out “all valence-electron” analyses of flexible cyclohexanones that carry electron-attracting substituents at positions 3 and 5. Because of computer-capacity limitations, we had to content ourselves with the simplest model, namely, 3,5-difhiorocyclohexanone, which represents correctly the effects of the two C-Hal bonds in Ic, but not the steric requirements of its two aryl

1IG

267 O--307.8

26&7--38&l

286~

-3

88

4

ET

26&l--3&T7--12267

206.8--3tWO--122.70

2136.&-380.3--122.73

213&2--38&S-122.70

284%-3a0%-122

7s

2eM---3902--12282

III”“

“““”

-25

-15

-5

5

I5

25

36

e

FIG. 4

EN

5

- 266.2- -3644

2648--386~0

ET

2142--3akc-12207

263,&--386

2--122.70

203.&-3388-S-122.73

262C-307C-12276

261.2--38%G-122.82

Ro.5

E,

E,

1-m0--387-6

271 2--3M.4

2wc-3902

20&E--3008

Ir

2S&2--381

C-122.67

267.lL-3392.&-1227:

207 0--392.b-122.73

2lWC-393.2--122.70

266-a --3s3a--122

78

2652---3W.C-122.82

-25

-15

-5

5

I5

25

35

e

FIO

. 6

EN

-

E,

263%--3844

263+--384.7

263.2--386.0

ET

262 EL--386.>-122

67

262.&-386b-122.70

262.3--386(c-122

73

2620--386.2--122-76

261 CL--3tW&-122.79

213,

&

--3W

~E

--(2

282

-25

-15

-5

5

15

25

35

9

FIG.~

FIG

S 4

-7.

Co

mp

ute

d

ener

gy

and

en

erg

y-co

mp

on

ents

(i

n

ato

mic

u

nit

s)

as f

un

ctio

n

of

the

dis

tort

ion

al

ang

le.

in

fou

r st

ereo

chem

ical

va

riet

ies

(hyp

oth

etic

al)

of

3Sd

iflu

oro

cycl

o-

hex

ano

ne

q E s

2198 A. Y. Meveu and J. SCHLISINGW

X

FIG 8a

Thrccchromophorc compounds111 2199

e’

24- 0,

Z 06-

-0.6-

FIG. 8b

2290 A. Y. MEYER and J. SCMLBINOER

FIG Bc

Ftcs 8a-c. Projections on the XL yz and xy plans of a flexible cyclohexanone. distorted at

0 = i-35”

Threwhromophorc compound+111 2201

substitucnts. Still, we do not consider the latter factor as decisive, in view of the observed conformational differences between Ib and Ic.

We used the CNDO/2 technique’O and the program authored by Segal;” geo- metries were defined in terms of 8, calculations being performed in steps of 5”.

Dipole moments, computed for e;a; and a;e; (Fig 2, dashed lines) as well as for a&a; and e;e; are close to the group-moment values (full curves). In Table 2 we com- pare the charge distribution of flexible cyclohexanone and 3.5-difluorocyclohexanone. both deformed at 8 = + 35”. The reported values indicate that the dipoles are almost fully localized at C=O, C,-F and C,-F, the perturbation at other centres being small. This obviously provides an a posteriori justification for the group-moment approach.

TABLE 2. Cwl~ce DLSTRIBUTON AT DISTORTION 0 = 35”

C, 0

C* H,(e’) Wa’)

C3 H,. F,(d) HI@‘)

G We’) IUa’)

C5 H ,. F&c’) H ,(a’)

G We’) H&f)

cyclohexanone electron units

a,e,-difluorwyclohexanone electron units

+ @250 -0256

-0.006

-0043 + 0.020 - 0003 + 0.020

+01)21 -0033 +0012 -0006 1

+0419 -0.015 +OOOl -0.003 1

+0.031 -oGt-J9 COO15 -0Ou7 1

-0085 +0032 -0W6 +0047 1

+ @247 -0208 +013 -0026 1

-0057 +OQl5

1 -0018

+ @024

+@257 -0211 +0023 - OQ23 1

- 0084 +0045 - 0.007 + OQ32

The &dependence of the nuclear-repulsion (EN), electronic (E,J and total energies (E, = E, + Ee) for the four possible patterns of substitution (a;e;, e;a;, a;a;, e;e;) are displayed in Figs 47 in atomic energy units (1 a.u. = 627 kcal/mol). We first note that, whereas EN and EE depend strongly upon the pattern, the total energy is almost unaffected, achieving its minimum, for all cases, at 8 = + 35”. The minimal-energy values are as follows: a;e;,-122.8122 a.u., e;a;,-122.8119, a;a;,-122.8116, e;e;.-122.8107 a.u. True, the differences are rather small, but it is significant that the ordering, at 8 = 35”. parallels exactly the trend in dipole-dipole repulsion energies (Fig 3): a;e;, 0234 eV, e;a;, O-268, a;a;, 0.310, and e;e;, 0371 eV.

As to relative stabilities, we now note that the trans-configuration is more stable than the cis, and that the form a;e; represents, at 8 = 35”, the lowest attainable energy. This may settle the ambiguity regarding the dipole moment of Ic. which

2202 Y. MEYER and J. SCHLESINGER

could indicate both a;e; at 35” and regular e;a; (Fig 2); the former of the two possi- bilities is more plausible.

For completeness, we show in Figs 8a-c the projections of the lowest-energy structure (aje;. 35”) upon the planes xz yz and xy ; the coordinate system was given in Fig 1.

EXPERIMENTAL

Dipole moments were calculated according to Halverstadt and Kumler.” from measurements carried out at 300 = 0.1 in Q-i6 (An&R. dried over Na). Dielectric constants were determined by the heterodyne

beat method (500 kHz). specific volumes-- with an Oswald-Sprengel type pycnometer. and the refractive

indices with a sodium-lamp Bellingham and Stanley Pulfrich refractometer. We are indebted to Dr. H.

Weiler-Feilchenfeld for these measurements.

Meerwein-Ponndorf- Verley reduction of 3.5di(p-chlorophcnyl)cyclolrpxano~ (Ic). Compound Ic (1 g)

was treated with the reducing agent. prepared ” from 037 g aluminium and 13.5 ml 2-propanol. for 3 hr.z4

The solvent was evaporated and the residue kept overnight with a soln ofconc HCI (10 ml) in water (35 ml).

The solid precipitate was filtered. dissolved in a sm;tll quantity of C,H, and chromatographed3 on a column

of 35 g alumina. Light petroleum eluted some impure starting material (ti neat 1720 cm _ ‘): further elution.

with a mixture of light petroleum and C,H,. yielded 0.6g of 3.5di(p-chlorophenyl)-I-cyclohexanol (111~).

m.p. 131 (from cyclohexane), ; Ccl, (dilute solution) 3605 cm- ’ (unassociated see OH”). (Found: C.

67.39; H. 5.73: Cl. 22.20. C,sH,,C120 requires: C. 67.30: H. 5.61 : Cl. 22.12,:;,).

Cataiyric reduction of 3.5-di(p-ch~orophenyl)-2-c~clohexen-1-0~ (11~). To 2.5 g of Ifc. dissolved in

CbH6 (lOOmI). 0.3g Pd/C (lOA) was added and the mixture hydrogenated for 5 hr at 3.4atm. Work-up

and three recrystallizations from cyclohexane left 06 g of 111~. m.p. 131’.

REFERENCES

’ Part Il. A. Y. Meyer. J. Schlesinger and E. D. Bergmann. J. Chim. Phys. 68. 668 (1971)

’ Part I. A. Y. Meyer and J. Schlesinger. Israel 1. Chem. 8.671 (1970) s M. Balasubramanian and A. DSouza. ~errahedro~ 24.5399 (I 968) a A: Y. Meyer and E. D. Bergmann. Israel J. Chem. 6.735 (1968)

’ E. D. Bergmann and A. Weizmann, Chem. Rev. 29.553 (1941)

’ C. P. Smyth. Dielectric Behavior and Structure. McGraw-Hill. New York (1955)

’ E. J. Corey and R. A. Sneen. J. Am. Chem. Sot. 77.2505 (1955)

8 J. D. Dunitz J. Chem.%d. 47.488 (1970) ’ N. L. Allinger. J. Allinger and M. A. DaRooge. J. Am. Chem. Sot. 86.4061 (1964)

to J. Cantacuztne. J. Chim. Phys. 59. 186 (1962)

’ ’ A. L. McClellan. Tables of Experimental Dipole Moments. W. H. Freeman. San Francisco (I 963)

I2 A. Y. Meyer. unpublished measurement (benzene. 30”)

” L. C. G. Goaman and D. F. Grant. Terrahedron 19. 1531 (1963)

” J. B. Lambert. R. E. Carhart and P. W. R. Corfield. J. Am. Chem. Soc.91.3567 (1969)

I5 R. U. Lemieux. R. K. Kullnig. H. J. Bernstein and W. G. Schneider. J. Am. Ckem’Soc. 80.6098 f 1958)

lb R. Levene. M.Sc. Thesis. Jerusalem (1969); cf. J. Klein and R. Levene. Tetrahedron Letters 4833 (1969) ” J. Cantacur.&e and R. Jantzen. Tephedron 26.2429 (1970)

‘e E L Eliel. N. L. Allinger. S. 1. Angyal and G. A. Morrison. Conformational Analysis. Interscience.

New’ York (1965)

I’ J. M. Lehn and G. Ourisson. Bull. Sot. Chim. Fr. I1 13 (1963)

,zo J. A. Pople. D. P. Santry and G. A. Segal. J. Chem. Phys. 43.5129 (1965); J. A. Pople and G. A. Segal.

‘3 Ibid. 43.5136 (1965): eid.. Ibid. U. 3289 (1966)

i G. A. Segal. CNDO II. QCPE 91 (1966)

22 1. F. Haiverstadt and W. D. Kumler. J. Am. Chem. Sot. 64.2988 (1942)

23 A. L. Wilds. in R. Adams. Ed., Organic Reactions, Vol. II. p. 198. Wiley. New York (1944)

2* M. Balasubramanian and N. Padma. Tetrahedron 19. 2135 (1963)

” C. N. R. Rao. Chemical Applications of infrared Spectroscopy. Academic Press. New York (1963)