"BUTTRESSING EFFECT"IN ELECTRONIC SPECTRA
Dec. 20, 1957
6495
mfi. T h e reaction mixture was worked up as previously hydrazine under the conditions used t o preparethederivative. described,21 the crude product being recrystallized from The only material which could be isolated was the starting hexane. The pure 2-ethyl-cis-2,3-epoxy-l,3-diphenyl-l- oxide, m.p. 81.0-82.0°, recovery 9357~~. propanol separated a s white needles, m.p. 98.8-99.S0, yield T h e infrared spectra16 of the crude reaction products ob4.56 g. (71.2%). The infrared spectrum16 of the epoxy tained from the isomerizations of a-ethyl-cis-benzalacetoalcohol contains absorption bands a t 3630 and 3520 cm.-l. phenone oxide and a-ethyl-trans-benzalacetophenone oxide The ultraviolet spectrum had a maximum a t 260 mp (e 46). were compared. The main constituent of each reaction Anal. Calcd. for C17Hl802: C, 80.28; H, 7.13. Found: mixture appeared to be the borofluoride complex of an enolized 8-dicarbonyl compound (strong, relatively sharp band a t C, 80.10; H, 7.12. 1555 ~ m . - l ) . ~ 3Since neither of the spectra exhibit absorpa-Ethyl-cis-benzalacetophenone Oxide (IIb) .-A solution of 4.0 g . (0.0158 mole) of 2-ethyl-cis-2,3-epoxy-1,3-diphenyl- tion in the 2700-2800 cm.-1 region we regard the presence 1-propanol in 30 ml. of pyridine was oxidized with the chro- of any significant quantity of a-formyl-a-phenplbutyrophenone ( X I ) , a possible rearrangement product, as very mium trioxide-pyridine complex,22prepared from 100 ml. of pyridine and 10 g. (0.10 mole) of chromium trioxide, as improbable.24 T h e spectra of the two crude products differ described by Wasserman and Aubrey.10 The pure a-ethyl- only in the intensities of three weak bands found a t 1675, cis-benzalacetophenone oxide crystallized from hexane as 1710 and 1742 cm.-1, these bands being more intense in the white needles melting a t 81.5-82.5', yield 2.90 g. (73%). spectrum of the crude rearrangement product obtained The infrared spectrum16 contains an absorption band a t from a-ethyl-trans-benzalacetophenoneoxide. The bands 1685 cm.-1 and has no band in the 3 fi region attributable of approximately equal intensity a t 1675 and 1710 cm.-' to a hydroxyl group. The ultraviolet spectrum has an ab- are judged to represent a small amount of 1,2-diplicnyl-1,3pentanedione, the product isolated from each rearrangement sorption peak a t 230 mp ( E 12,300). after complete hydrolysis of the borofluoride complex. T o Anal. Calcd. for C17H1602: C, 80.92; H, 6.39. Found: verify this hypothesis the crude rearrangement product C, 80.95: H. 6.66. from the trans-oxide was recrystallized from methanol; the Rearrangement of a-Ethyl-cis-benzalacetophenone Oxide resultant solid after successive recrystallizations from meth(IIb).-A solution of 254 mg. (0.001 mole) of the oxide in 12 anol and hexane afforded pure 1,2-diphenyl-l,3-pentanediml. of benzene was saturated with boron trifluoride gas and one, m.p. 92.5-94.5", yield 47%. T h e infrared spectrum" allowed t o stand a t room temperature for 2 hr. After the of the residue from the combined mother liquors exhibits mixture had been diluted with ether and washed with water, the bands a t 1675 and 1710 cm.-l as well as a broad band the solvents were removed and the crude solid which re- with its center a t 1600 cm.-' (enolized 8-dicarbonyl conimained was recrystallized from alcohol. An additional re- pound). The relative intensity of the band a t 1742 cm.-' crystallization from methanol followed by two recrystalli- increased indicating that the component represented by zations from petroleum ether afforded the pure keto form this band has been concentrated by the partial removal of of 1,2-diphenyl-1,3-pentanedione, m.p. 93.0-95.0", as white the 1,2-diphenyl-l,3-pentanedione.However, we were needles, yield 150 mg. (59%). -4 mixed melting point de- unablc to isolate the small amount of this second compotermination with the diketone obtained from the trans- nent present in the reaction mixturc. The location of band oxide showed no depression. 1742 cm.-l attributed t o this second component suggests In another experiment 254 mg. (0.001 mole) of the oxide that the material is not either of the two expected rearrangewas isomerized as previously described. A solution of the ment products, a-formyl-a-phenylbutprophenone ( X I ) or crude rearrangement product in alcohol was treated with 1,3-diphenyl-3-ethyl-1,2-propanedione(XII). 25 166 mg. (0.0015 mole) of phenylhydrazine and 0.4 ml. (0.007 mole) of acetic acid as previously described. The (23) L. J. Bellamy, "The Infrared Spectra of Complex Molecules," yield of 3-ethyl-l,4,5-triphenylpyrazole,m .p. 138.0-139 .OO , John Wiley a n d Sons, Inc., New York, N. Y . , 1954,p. 126. was 282 mg. (87%). A mixed melting point determination (24) Absorption in this region, attributable t o t h e C-H stretching with the pyrazole sample previously described showed no vibration of a n aldehyde function, is clearly discernible in t h e spectra depression. When the reaction time for the rearrangement of analogous compounds (see ref. 12). was 30 min. the yield of the pyrazole was 282 mg. (87%). (25) T h e absorption peaks attributable t o t h e carbonyl functions in As a control experiment the a-ethyl-cis-benzalacetophe- t h e homologous compounds, a-formyl-a-phenylpropiophenone and none oxide (254 mg., 0.001 mole) was treated with phenyl- 1.3-diphenyl-l,Z-propanedione, are found a t 1G70 and 1730 cm. and (22) G. I. Poos, G. E. Arth, R. E. Beyler and L. H. Sarett, THIS JOURNAL, 76, 422 (1053).
1630 and 1705 cm. -1, respectively.
CAMBRIDGE 39, MASSACIIUSETTS
[ C O N T R I B U T I O N 1'ROhl TIIE hlEhlORIAL UNIVERSITY O F NEWFOUNDLAND]
Electronic Spectra and Molecular Dimensions. II.1,2 The "Buttressing Effect" and Other Secondary Steric Interactions in Electronic Spectra BY W. F. FORBES AND UT.A. MUELLER RECEIVED JUNE 14, 1957 Steric interactions are known t o directly affect the main band (B-band) in ultraviolet absorption spectra. Developing the concept of steric hindrance, the so-called buttressing effect is examined, and it is shown that some other changes affecting both wave length shifts and absorption intensities can be rationalized by assuming thqoperation of more extended or indirect steric interactions.
Introduction
nance is large and the coplanarity of the niolecule is almost completely destroyed, bands occur which may be regarded as partial chromophore bands, that is, the molecule absorbs as two or more distinct entities4 If the steric interaction is less (1) Presented before t h e Division of Organic Chemistry a t the pronounced in comparison with a planar reference Miami Meeting of t h e American Chemical Society, April, 1957. molecule, a wave length shift (usually to shorter (2) Part I, Can. J. Chem., 34, 1542 (195G). (3) E. A. Braude a n d F. Sondheimer. J . Chem. S O C,~ . 3754 . (10551: . ~ ~ wave , . length) accompanied by decreased absorption
Steric effects in the B-band of ultraviolet absorption spectra have been shown3 to give rise t o various changes. If the steric inhibition of reso-
W. 1;. Forbes and W. A. Mueller, C a n . J.'Chem., 34, 1340. 1347 (19.56): 36, 488 (19373.
THISJ O U R N A L , 72,44 (1950); 68, 2'290 (4) Cf.L. W.Pickett. et d., (1936); E. Mueller and H . Neuhoff, B e y . , 72, 2063 (1939).
6406
W. F. FORBES AND W. A. MUELLER
Vol. 70
intensity is observed. Lastly, a type of steric and then four m-substituents contribute to a posinteraction can be detected in which the wave sible buttressing interaction. length of maximal absorption remains approxIt should be noted t h a t the value of biphenyl as imately constant, while the absorption intensity a t reported by Ross, et a1.,7a who themselves quote an maximal absorption is appreciably reduced. earlier value for this compound, is probably high Examples of this last type of steric hindrance and consequently the correct extinction coefficient have been discussed p r e v i ~ u s l y ,and ~ i t has been under identical circumstances is likely to be lower shown that the hypothesis of steric inhibition of ( ~ four . own value of Emax 15,600-16,2002). This resonance accounts satisfactorily for the observed raises an important point, since it would appear t o changes. More precisely, the hypothesis is that preclude the operation of a buttressing effect in the the individual non-hindered compounds exist in spectral data. However, as will be shown more two conformations, s-cis and s-trans, both of which fully elsewhere, the buttressing effect is of course not contribute to the observed absorption. “Hin- the only interaction operating in m-disubstituted dered” compounds listed in the papers referred to benzene derivatives, and frequently it will be opstill exist in both these conformations, but only one posed by an interaction which causes a positive of these (the less hindered conformation) contrib- wave length displacement and intensity increases. utes appreciably t o the observed maximal ab- Hence in considering the effect of m-substituents, it ~ o r p t i o n . ~Using this hypothesis it is possible to would be an oversimplification to consider excluobtain interference values of non-bonded atoms in sively the buttressing effect in causing the observed solution and this approach has been outlined in changes. In fact, a number of examples can be Part I of this series. Since the values thus ob- found, especially if the m-substituent is small, tained are larger than those previously used,2it also where an intensity increase is observed (cf., for exsuggests a rediscussion of a number oi topics which ample, ref. 7b). However, the general decrease in previously have received only scant attention. extinction coefficients, whenever the substituents In the following sections, therefore, some secondary are large, is taken to point to the operation of the stcric interactions will be discussed with special buttressing effect. reference to their possible detection in ultraviolet Further examples showing possible buttressing light absorption data. interactions are available in other benzene derivaThe Buttressing Eff ect.-The buttressing effect tives, and some of these are listed in Table I. for substituted benzene derivatives is defined as Table I illustrates how the introduction of large the interaction occurring due to a m-substituent m-substituents frequently appears to inhibit a t exerting an indirect steric effect by “buttressing” least some of the conjugation as shown by the dethe o-substituent; in this way a m-substituent may creased absorption intensity or disappearance of the alter the steric interactions of the two other vicinal B-band. For example, B-bands for spectra of m-iodo substituents. Convincing evidence for the opera- compoundshave consistently lower absorption intention of the buttressing effect is provided by the sities and frequently occur only as inflections. AIrates of racemization of certain optically active bi- though other effects may also account for these phenylsj6 and the proposed interfereiice values of changes the data correlate well with a steric intercarbon, hydrogen and oxygen as deduced in Part pretation. I n already hindered compounds, on mI2also suggest t h a t a buttressing effect may some- substitution, steric inhibition of resonance 1s entimes be discerned in the B-band of ultraviolet ab- hanced (see the first two examples in Table I) and sorption spectra. This latter suggestion is in fact the effect is generally more pronounced for larger confirmed by the spectra of a number of com- substituents. Thus it appears to be more propounds. nounced for m-nitroacetophenone than for m-nitroIn biphenyls, where a buttressing effect is known benzaldehyde as shown by a more pronounced into occur from rates of racemization data, evidence tensity decrease for m-nitroacetophenone. The infor this effect in ultraviolet absorption properties tensity decreases for the corresponding nz-chloro may be adduced from compounds like 3,3’-di- compounds are similar, and this suggests that there methoxybiphenyl which shows considerably lower the buttressing effect, if present, is similar. Howabsorption (emax 12,000 a t 250 mp) than that of bi- ever, i t is more probable t h a t in m-chloro comphenyl (emax 17,000 a t 249 mp).6 This type of ef- pounds the buttressing effect is of little importance. fect is quite generally observed with large m-sub- Not unexpectedly, with the larger bromo-substitustituents (see below and particularly Table I). ent the buttressing effect is again more evident. I t is consistent with the explanation that the m- Therefore, apparently the buttressing efTect is more methoxy substituent, by buttressing the o-hydrogen pronounced for bromine than for chlorine and is atoms, decreases the planarity of the biphenyl more pronounced for iodine than for bromiiie. system. A similar progressive lowering of absorp- This is entirely as anticipated. We may note that the case for the buttrcssing tion in the biphenyl series also has been noted for the spectra of biphenyl (Emax 20,300 a t 246 mp), effect is inherently not as definite as the evidence 3,3‘-bis-(trifluoromethy1)-biphenyl (emax 17,780 a t which indicates that the ortho-effect is primarily due 245 mp) and 3,3’,5,5’-tetrakis-(trifluoromethy1)- to steric interactions. This follows because the inbiphenyl (Emax 15,850 a t 242 m ~ ) ,where ’ ~ first two tensity changes involved in m-isomers are much smaller and may partly be caused, for example, by (5) R . Adamsand H. R. Snyder, THIS JOURNAL, 60,1411 (1938); M. Rieger and F. H. Westheimer, ibrd., ‘78, 19 (1950); S. L. Chien and R. Adams, i b i d . , 66, 1787 (1934). (0) E. A. Braude and W. F . Forbes, J Chem. SOC.,3776 (10551. (7) (a) S. D. Ross, M. Markarian and hi. Scbwarz, THIS JOURNAL.
76, 4967 (1953); (b) cf. A. I. Bilbe and G. M. Wyman, ibid., 1 6 , 5312 (1953), for the spectra of 3,5-difluoroaniline and m-Euoroaniline; we are most grateful to a ref-ree of this paper for drawing our attention to this.
Dec. 20, 1057
“BUTTKESSIKG EFmc?’
IN
6497
ELECTRONIC SPECTRA
TABLE I XBSORPTIOS ~ ~ A X I M(B-BAKDS) A OF m-SUBSTITUTED BENZENE DERIVATIVES EXHIBITIKG A DISTINCTBUTTRESSING EFFECT AND REFERENCE COMPOUNDS (VALUES IN ITALICS REPRESENT INFLECTIONS) Solvent
Ethanol Ethanol Ethanol Ethanol
cyclohexane Cyclohexane Cyclohexane 95% ethanol Cyclohexane Ethanol
m-Substituted compound Compound L u x , mrr
2,3,4,5,6-Pentamethylaceto(216 phenorie (212 2,3,5,6-Tetramethylacetophenone wz-Sitrobenzaldehyde 265 ca. 261 nz-Kitroacetophenone 246 ?It-Iodoacetophenorie 250 244 m-Iodobenzaldehyde 249 (253 m-Iodobenzoic acid 233 m-Iodonitrobenzene ca. 262 (ca. 232 nz-Iodoaniline \ 239 m-Iodoacetanilide 246
I
Reference compound A,, mrr Compound
(ma*
Ref.
12,000)“ 2,4.6-Trimethylacetophenone
242
3,600
3,s
11,500)“ 2.6-Dimethylacetophenone
251
5,600
3,8
265 263
11,400 13,800
9 9
Acetophenone
238-239
12,500
10
Benzaldehyde
241
14.000
10
Benzoic acid Sitrobenzenc
230 260
10,000 10 8,000 11,12
Ainiline
233
(ma%
11,700 6,600 6,800 6,000 8,000 7,500 7,009 9,800 6,400 10,400 8,700 13,500
Cyclohexane wz-Bromotoacetophenone
p-Xitrobenzaldehyde p-Nitroacetophenone
.4cetanilide Acetop hcnoiie
242 238-239
{ 2 * z 1::::: Benzaldehyde Hexane 242 11,500 m-Broniobenzaldehpde 95yG ethanol m-Bromobenzoic acid Benzoic acid 225 8,500 Sitrobenzene 9570 ethanol nt-Bromonitrobenzetie C U . 239 6 ,200 Xniline Cyclohexane m-Bromoaniline 235 7,800 Acetanilide m-Bromoacetanilide Ethanol 246 14,000 Cyclohexane m-Chloroacetophenone .icetophenone 239 10,000 211-212 11,500 nz-Chlorobenzaldehyde Hexane Benzaldehyde Benzoic acid 95% ethanol m-Chlorobenzoic acid 230 8,500 Kitrobenzene 957’ ethanol rrz- Chloronitrobenzene ca. 238 7,200 Aniline Cyclohexane nz-Chloroaniline 235 8,500 Ethanol Acetanilide nz-Chloroacetanilide 245 14,900 Absorption band represents absorption mainly due to 2,3,5,6-tetratnethylbenzene( t i n f i 8500 at
association effects. Further, and partly because of this, the buttressing effect is frequently not detected in the spectra of m-substituted benzene derivatives. For reasons which will be explained elsewhere, acetophenones and nitrobenzenes provide favorable examples in the study of this effect. At this point, the choice of interference radii for halogen atoms can be discussed. If we assume t h a t the chlorine atom does not exert an appreciable buttressing effect, a siniple scale diagram shows that the effective interference radius of the chlorine atom is less than 2.0 A.,oassuming an effective interference radius of 0.95 A. for hydrogen (using a value of 1.2 A. for hydrogen, a corresponding value of 1.8 A. is obtained for the chlorine atom). The maximum error for the value of the chlorine atom, however, is probably as high as 0.4 PI., because among other uncertainties the value depends on the hydrogenradius, which has b e p estimated t o be accurate only t o within 0.1 A. Further, apart irom the other variables mentioned in Part I , 2 the possible buttressing effect of the other substituent on the ohydrogen atom has been neglected (see also below). As it is, an i n t e r f p x c e radius for the chlorine atom of less than 2.0 4 . receives support from the value ( 8 ) W. F. Forbes a n d ‘A’. A. Mueller, Caw. J . Chem., 33, 1145 (1955). (9) E. A. Walker and J. R. Young, J . Chem. Soc., 2041 (1957). (10) D a t a obtained in these laboratories. (11) H. E. U n g n a d e , . T ~ r sJ O U R N A L , 7 6 , 1601 (1954). (12) IvI. J. Kamlet and D . J. Glover, i b i d . , 77, 5696 (1955). (13) H. E. Ungnade, i b i d . , 76, 5133 (1954). (14) C.M. Moser and A. I. Kohlenberg, J . Chem. Soc., 804 (1951).
9,000
10
14,500 13 12,500 10
211 14,000 228 10,000 8,000 260 9,000 233 242 14,500 12,500 239 14,000 241 228 10,000 8,000 260 9,000 233 242 14,500 214 mp).
10 14 11,12 10 13 10 10
14 11,12 10 13
a.
of 1.73 obtained for the chlorine atom by Bastiansen‘j and from the gccepted van der N7aals radius for chlorine of 1.8A.I6 For bromine, the buttressing effect is inore pronounced. It is detectable in most m-isomers, occasionally causing the B-band t o appear as an inflection only (cJ 14). consequently, a? effective interference radius of about 2.0 + 0.4 A. is indicated, which may beo compared with the van der \Vaals radius of 1.95A.16 and with the value of 1.81 A. obtained by BastiansenIb from electron diffraction data. M7hatever the exact value may be, the data even a t this stage suggest that careful spectral analysis, given favorable circumstances, may detect the interactions involving the fringe of the van der Waals interference radii. For iodine, as expected, the buttressing effect is yet more pronounced and the B-band of m-isomers occurs more frequently only as an inflection (and hence incidentally is less reliable for the accurate determination of absorption intensities). This suggests t h a t the maximum effective interfereqce radius of iodine is slightly greater than 2.0 A. arid the value of 2.1 + 0.4 A. may be suggeste4 for iodine (cf. the van der Waalg radius of ‘7.13 A.16 and Bastiansen’s value of 1.91 A.’5 for iodine). Further evidence for a buttressing effect is provided by the observation that the combined ap(15) 0. Bastiansen, Acta Chem. S c n n d . , 4 , 926 (1950). (16) L. Pauling, “ T h e N a t u r e of t h e Chemical Bond,” 2 n d . ed., Cornell University Press, Ithaca, N. Y., 1948, p. 189.
W. F. FORBES AXD W. ,I.MUELLER
6498
Vol. 79
TABLE I1 ABSORPTIONBANDSO F 2,5-I)INITRO-1,4-DICHLOROBENZENEAND REFERENCE COMPOUNDS
,--------. "X
Compound
Solvent
2,5-Dinitro-1,4-dichlorobenzene Water m-Chloronitrobenzene m'ater o-Chloronitrobenzene
tvater
-A m .,
mp
emax
mp
221 208.5 209
23,900 17,300 13,700
236 224 228
enlax ,
.
6800 4400
aX" X.n*x, 7
i
-
7
mp
fmnr
w
tm,,
252.5 264 260
6250 7100 4000
322.5 313 310
2720 1300 1400
Ref
18 19 19
parent effect of two o-substituents is sometimes very much greater than t h a t of one. This previously has been ascribed by Hedden and Brown t o the buttressing effect,l' and we accept this interpretation with only a slight modification. Thus, the fact t h a t two o-groups sometimes have more than twice the effect of one o-group does not necessarily imply a buttressing effect. It may be explained by the existence of s-cis and s-tvnns configurational isomers. It may also be shown that identification of the buttressing effect is not by any means limited t o simple molecules and actually becomes frequently more important in spectral analyses of more complex molecules. This is illustrated by the following examples. 2-Carboxy-4,5-dimethoxy-2'-nitrobiphenylabsorbs as follows (see Fig. 1): Xmax 223 mp, 6 38,000; Amax 256 mp, e 16,000; Xina 296 mp, E 6000. The band a t 256 nip is ascribed t o nitrobenzene absorption (nitrobenzene in ethanolic solution absorbs with Emax SO00 a t 237 mp) . The increased absorption in the biphenyl derivative is presumably due t o additional absorption because of the other benzene ring, as well as due t o interactions involving both rings. The inflection a t 296 mp presumably represents C-band absorption and is associated with the second benzene ring. It may be compared t o a similar absorption band for m-anisic acid in ethanol (ema, 2450 at 293 mpl?). The increased absorption intensity is again ascribed to the other benzene ring and biphenyl conjugation. On introducing an m-methyl group the spectrum obtained (see Fig. 1) is Xmax 218 mp, E 33,000; k i n e 256 mp, E 10,700; Xinfl 294 mp, E 4500. Both bands are evidently affected and the simplest explanation t o account for the changes is t h a t the methyl group dislodges the nitro group by means of the buttressing effect. This in turn increases the interplanar angle of the biphenyl system and consequently the decreased absorption intensity is observed. I n concluding this section, it should be einphasized t h a t in a relatively complex spectral analysis, great care must be taken in the assignment of absorption bands. For example, i t has been suggested that an inflection a t 236 m p in the spectrum of 2,sdinitro-1,4-dichlorobenzene may be due t o the buttressing effect.I8 However, the spectrum of this tetrasubstituted compound, largely for steric reasons, will be comprised of the spectra of the disubI n this way, since stituted parent 0- and m-chloronitrobenzenes absorb maximally a t four positions, it seems unnecessary t o associate the inflection a t 236 mp with a buttressing effect
(see Table 11). More probably, the inflection corresponds to the 228 mp band in o-chloronitrobenzene slightly moved to longer wave length on account of the additional substituents. Steric Interactions in Molecules Possessing a High Degree of Rigidity.-Examples where a t least in theory ultraviolet spectra inay provide an unusually sensitive index of steric inhibition of resonance (orbital delocalization) are rigid molecules which cannot readily take up alternate configurations. This is illustrated in Fig. 2. I n Fig. 21 solid lines indicate the potential energy curves of a hypothetical nun-hindered molecule --l-B, where electronic interaction across the central bond is possible and planarity represents the energetically most favorable configuration. The dotted lines indicate potential energy curves if steric interactions are present, but are s1n:ill. These steric interactions may raise the enerfi?; levels of both ground and electronic excited states. 11sshowii in Fig. 2.1, the transition energy of the absorption mill hardly be affected because the steric interaction is largely hidden on account of the permitted free rotation as indicated by shallow energy curves. Figure 2B indicates a similar system -I--Bin which the molecule is approximately coplanar because of molecular environment. Free rotation is now considerably more difficult, and consequently tlic slightly sterically hindered coplanar coiiforniation will exert a greater effect on the absorption hand. This follows because even a small interplaiiar angle appreciably raises the energy level of the ground state and hence more transitions will be coiifined to the coplanar conformation where the stcric cifect will be apparent. I n this way, it is conceivable that a steric interaction of say 3 kcal./mole or less (ix., steric hindrance which will not norinally be detectable if there is free rotation) may cause a wave length shift in a molecule when the atoniic environment does not permit free rotation. Further, on altering the interplanar angle 8, a considerable energy increase will occur in both ground arid electronic excited states, and therefore the over-all transition energy in a more rigid systeiii nay be considerably altered. This variation may causc either a positive or a negative wave length shift, depending on the relative positions of ground and excited states energy levels. This increased steric interference, presumed to be due t o greater rigidity, has been noted Ix-eviously for polycyclic systems by Braude and Solidheimer20 (and references cited there), and in all probability this type of steric interference accounts also for many of the "anomalous" spectra of other cyclic To illustrate the iinportance of rigitlsystems.21~22
(17) G. D. Hedden and W. G. Brown, THISJ O U R N A L , 76, 3744 (1993). (18) G. S. Hammond a n d 17. J. Modic, i b i d . . 76, 138.5 (1953). (19) I.. D o u b a n d J. M. Vandenbelt, i b i d . , 71, 2414 (194Y).
( 2 0 ) E . A . Braude a n d F. Sondheimer, J . C l r c i i r . SOL.,3773 (19.55). (21) R . S . Moore and G. S . I'isher, Ttrrs JOIIRNAI., 78, 4362 (19.50); 11. C. Cookson, J . CIIzcnr. S n c . , 2 8 2 (1'364); 0.H. XVheeler, THIS J O U R N A L , 78, 3216 (195ii); W. A I . Scbubert and W. A . Sweeney,
Dec. 20, 1957
6499
“BUTTRESSING EFFECT”I N ELECTRONIC SPECTRA
30
c”
I
X
20
VI
-8-
-4
B
Fig. 2.-Schematic representation of electronic transitions between ground and excited states for non-rigid (.4) and more-rigid (B) systems in the absence (solid lines) or presence (dotted lines) of small steric interactions.
10
220
240 260 280 300 \ T a w length, nit. Fig. 1.--Ultraviolet absorption spectra in ethanol: -, 2-carboxy-4,5-dimethoxy-2‘-nitrobiphenyl;------ , 2-carboxy-4,5-dimethoxy-2’-nitro-3 ’-methylbiphenyl.
ity in a system we may consider the recently reported data for some azulenes, which are listed in Table 111. The effect observed for 5,B-benzazuTABLE 111 MAIN ULTRAVIOLET LENGTHS)O F
A4BSORPTIO?;
-4ZULENES I N
Compound
1,2-Benzazulene 5,6-Benzazulene Azulene
MAXIXA (LOSC \ v A V E CYCLOHEXANEz3
Amax, mir
emax, mu
383 355.5 338.5
4600 3550
4000
lene, which appears to be steric in origin but cannot be explained in terms of conventional interference radii,3q23may now tentatively be ascribed to this type of steric interaction. It may be noted that in this example molecular models indicate an interference radius for hydrogen of ca. 1.0 A., approaching the value usually accepted for the van der Waals radiGs and slightly in excess of the value of 0.95 0.1 A.2,;.e., the value which normally can be detected in ultraviolet spectra.
*
Conclusions It has been shown in previous communications that the ortho-effect iii the B-band of ultraviolet absorption spectra can best be explained by assuming the operation of steric interactions associated with repulsions between vicinal non-bonded atoms. ibid., 77, 2297 (1965); E . A. Braude, Chemistry & Industuy, 1557 (1954): R. Huisgen, W. Rapp, I . Ugi, H. Walz a n d E. Mergenthaler, A n n . , 686,1 (1954). ( 2 2 ) E. Heilbronner and R. Gerdil, Helv. Chim. Acta, 39, 1996 (1956); W.F.Forbes, ibid., 40, in press. (23) E . Kloster-Jensen, E . KovLts, A . Eschenmoser and E. Heilbronner, ibid., 39, 1051 (1956).
Extending this hypothesis we can conclude that the buttressing effect must also be taken into account, particularly when analyzing the spectra of more complex molecules. However, care is required in the spectral analysis of complex molecules, because only infrequently can the buttressing effect be identified with certainty. The limit of the interference radius, that is the smallest steric interaction which may be detected in the B-band of ultraviolet spectra, probably varies with the compound under investigation. An interference radius larger than the normal value can sometimes be detected in molecules where free rotation is restricted, since this can be shown t o increase the sensitivity to steric interactions in certain examples. Unfortunately, in molecules of this type, other indirect interactions frequently so modify both energy curves that no simple generalizations are justified and each set of examples must be examined separately. All the data discussed indicate that interactions involving interference radii tending toward van der Waals radii often play their part in determining ultraviolet spectra. Previous suggestions that much smaller interference radii generally should be used in estimating steric effects are ascribed to inadequate spectral analyses. Experimental 2-Carboxy-4,5-dimethoxy-2’-nitrobiphenyl and 2-Carboxy-4,5-dimethoxy-2’-nitro-3’-methylbiphenyl.-~Ve are grateful t o Dr. 2 . Valenta of the University of S e w Brunswick for generous samples of these compounds. The melting points of the substances were 207 and 2 1 6 O , respectively.24 The preparation of all the other compounds has prcviously been described,3 or is well known. All spectra were determined under the same conditions as described in the first paper of this series.2 T h e spectral data are recorded in Table I and Fig. 1. It is seen from Fig. 1, t h a t for this example the band shape for the nitrobenzene B-band remains reasonably similar, and this general similarity of band shape is also observed for most of the pairs of compounds listed in Table I. Consequently, since no quantitative comparisons are intended at present, the molecular extinction coefficients (24) K. Wieser, Z . Valenta, A. 77, 675 (19%).
JOURNAL,
I.Manson and
F. W. Stonner, THIS
CHESTERW. MUTII,JOHN C. ELLEKSAND 0. FREDFOIAIUR
6500
tensity changes. This statement is not intended t o preclude minor changes in band shape on m-substitution. However, it may be noted t h a t frequently emax and oscillation strengths are directly related one t o another (cf. ref. 2 5 ) . (25) B M. Wepster, Rec. Irav. c h i m . , 76, 336 (1957).
[COSTRIBUTION FROM THE
\'ol. 79
knowledge t h i receipt of research gr&ts fro& the Research Of Canada and the Research Corporation of New York. ST. JOHS'S, N E W F O U N D L 4 X D , CAS14DA
DEPARTMEST O F CHEMISTRY, \brEST
\JIRGISIA
CNIVERSITY]
A Novel Ring Closure Involving a Nitro Group ; Preparation of Phenanthridine-5-oxidel BY CHESTERW.MUTH,JOHX C. ELLERS AND 0. FRED FOLXER RECEIVED JULY 3, 1957 I n the biphenyl series a nitro group in the 2-position will react with an activated methylene group in the 2'-position in the presence of sodium hydroxide or sodium methoxide to form substituted phenanthridine-5-oxides. Cyano, carbomethoxy and carbamyl groups are activating groups; hydrogen, hydroxyl, bromo aiid carboxyl do not serve as activators.
Discussion An unexpected reaction was observed when methyl 2-(2'-nitropheny1)-phenylacetate (I)2 was warmed with methanolic sodium hydroxide in an effort to saponify it. The product obtained was not the parent acid of I, but was compound A, an acid, which had no nitro group and which decomposed with gas evolution to yield compound B. Compound B could be treated with zinc and hydrochloric acid to produce compound C.
A
4J
I
Compound C was identified a s phenanthridine by comparison of its picrate with that of an authentic sample. Compound B was identified as phenanthridine-5-oxide (111) by m.m.p., infrared spectrum and by its picrate. Since the gas given off in the decomposition of compound X was carbon dioxide, the structure postulated for compound A is 6-carboxyphenanthridine-%oxide (11). M-ith this information the foregoing reactions may be rewritten
1
+
i
4
--+
C, phenanthridine
0 A or I1
Since the reaction was very rapid under mild coliditions and since the literature did not disclosc such a reaction, a further investigation appeared profit(1) Supported by t h e S a t i o n a l Science Foundation. Research Grant G-1581, whose help we wish t o gratefully acknowledge. From the P h . D . dissertation of 0 . F. F . ,1!157, and l I . S . thesis of J , C. Is.. 1058, both from Tl'est Virginia University. Presented in part a t t h e 130th hIeeting of t h e A.C.S., Atlantic City, X . J . , September, 1LI36 (2) C. 1%'. Liuth, W. L. Sung and Z . B. Papanastassiou, THIS J O U R K A L . 77,3393 (195.5).
able. The further study of this reaction was approached along two lines. First, the reaction of ester I was investigated under a variety of conditions to learn more about the yields, the structure of the acid material and the mechanism of the rcaction. Secondly, the carbomethoxy group was replaced by other groups and the resulting conipounds were tested for this cyclization. The bases used which effected the cyclization of ester I were sodium hydroxide in methanol or water and sodium methoxide in methanol. The results are summarized in Table 11. 115th low base concentration ester I yielded G carbomethoxyphenanthridine-%oxide (IV), whereas ester I with high base concentration and longer reaction time produced acid -1aiid I11 or acid I1 and 111. Acid A evolved a gas a t 120-138" and melted a t about 200°, whereas acid I1 inclted a t about 200" with no previous gas evolution. Both acids X and I1 reacted with diazomethane to give the same ester; both yielded phenanthridine-5oxide when warmed with dimethylformamide and their infrared spectra were nearly identical. The use of a nitrogen atmosphere3 or an equal molar quantity of hydroquinone had no effect on the reaction, hence it is assumed that the cyclization is not an oxidation-reduction process. Since the cyclization is not affected by acids (concentrated or dilute sulfuric acids) or by dehydrating agents (acetic anhydride or thionyl chloride) nor by ammonium hydroxide, but only by strong bases (sodium hydroxide and sodium methoxide) it seems likely t h a t the cyclization is a base-catalyzed reaction similar to an aldol condensation. T h a t hydrolysis of the ester group is not a necessary condition for the cyclization is shown by the production of the cyclized ester, 6-carbonicthoxyphenanthridine-3-oxide (IV), This is also shown by the fact that the parent acid, ~-(2'-nitrophenyl)phenylacetic acid (I7),of cster I does not cyclize. Seven model compounds were subjected t o basic conditions to determine whether Cyclization would oc.cur. These compouricls were the same as methyl 2-(2'-riitrophenyl)-phenylacetate (I) except that the carboinethoxy group was replaced (see V-XI). If3th I< as carboxyl (IT),hydrogen (\:I), hydroxyl (IT) and bromine (VIII) no cyclization (3; T. Tsuruta, T. Fueno and J. Furukawa, ibid., 77, 3205 (1953J.
