J. C. MARTINA N D RUSSELLG. SMITH
2252 [CONTRIBUTION FROM
THE
Vol. 86
NOYESCHEMICAL LABORATORY, UNIVERSITY OF ILLISOIS,URBANA, ILL.]
Factors Influencing the Basicities of Triarylcarbinols. The Synthesis of Sesquixanthydroll BY J. C. MARTINA N D RUSSELLG. SMITH* RECEIVEDDECEMBER 16, 1963 Synthetic routes to the novel 2,6,2‘,6’,2”,6”-hesasubstituted triphenylcarbinols ( I and 11) have been developed. Both carbinols are of considerable theoretical interest. The hexamethoxytriphenl-lcarhinol I , because its sis ortho substituents provide large steric interactions between rings, is a precursor of a carbonium ion which must esist in an exaggerated “propeller” conformation, or a t any rate a highly nonplanar conformation. Sesquisanthydrol (11) is of interest since its derivative carbonium ion (and radical) appear from models t o be capable of attaining complete coplanarity. Both carbinols yield well-characterized carbonium ions a t unusually high p H . Some of the chemistry of these carbonium ions is discussed and quantitative measurements of basicity are reported
The question of the detailed structures of the triphenylmethyl carbonium ion and radical has provided, over the past 20 years, a persistent impetus to the study of their properties. The conflicting requirements for maximizing the resonance stabilization of the ion or radical and minimizing the steric interactions between o-substituents on neighboring rings demand a compromise which has been accommodated to several different models The aryl groups may be twisted out of the plane of the central carbon atom to form a symmetrical “propeller’’ which may be converted to an unsymmetrical analog by reversing the “pitch” of one of the rings.3 Such a model, in which the angle of twist has been est i m ~ c t e dto ~ , be ~ of the order of 50-60°, must involve a loss of resonance stabilization. Much evidence concerned with the basicity of substituted triphenylcarbinols has been interpreted6-11 in terms of this model for the carbonium ion. An alternative hypothesis advanced by Deno4,12-14 explains the stabilization of the ion in terms of a model in which only one or two of the phenyl rings attains coplanarity a t any given time. In this model one of the rings, not contributing to the resonance stabilization of the ion, is regarded as essentially perpendicular to the plane of the central carbon atom. The greater basicities of triarylcarbinols, as compared to the corresponding diarylcarbinols, would be considered to be attributable primarily to the greater relief of steric strain resulting from the change from tetrahedral to trigonal geometry on going to the triarylmethylcarbonium ion. This paper reports the preparation and properties of 2,(i,2’,6’,2”,6”-hexamethoxytriphenylcarbinol (I), and 12c-hydroxy-4,S,12-trioxadibenzo [c,d,m,nIpyrene I I) Ft-om t h e P h 1) Thesis of R. G . S m i t h , University of Illinois, 1961 Presented in p a r t before t h e 138th S a t i o n a l hleeting of t h e American Chemical Society, S e w Y o r k , N.Y , S e p t . , 1!)60, p . 8 1 P of Abstracts ( 2 ) R v h m a n d Haas C o . Fellow, 19.i9-1960 ( 3 ) C: N . I,cwis, T ’r Magel, and I ) . I.ipkin, J A m . Chrm Soc., 64, 1771 (llIi2) I l e n o , J J . Jaruzelski, a n d .4. Schriesheim, J . Org Chem , 19,
0 Pritchard a n d F H S u m n e r , J (‘hem Soc., 1041 (1955). soc. chrm. f i . o , i c r , 18, ClO0 (19.51) ( 7 ) V G o l d J . [ h e m Soc., :3K ii) H
it;)
P I ) B a r t l e t t , Buii.
or sesquixanthydrol’j (11) and the corresponding carbonium ions I11 and IV.
& ’0‘
II I
1“’
i.*
I11
Results and Discussion Carbinols I and I1 are conveniently prepared from 2,G-dimethoxyphenyllithium,which is available from resorcinol dimethyl ether by metalation with butyllithium. The synthetic sequences and reactions interrelating these compounds are outlined in Scheme 1. The additions of 2,G-dimethoxyphenyllithiumto the sterically hindered ketones or esters intermediate in this synthetic sequence do not occur in ether or tetrahydrofuran solution. I n a solvent mixture with a large percentage of benzene, however, one obtains addition to give the highly hindered carbinol, I . The use of other catalysts (X1Br3,XIC13, HBr) for the ring closure to 11 results in yields inferior to those obtained using molten pyridinium chloride as a reaction medium. The second methylation step in the conversion of VI to VII, involving the strongly hydrogen-bonded hydroxyl group, could not be effected with diazomethane, even in the presence of catalytic amounts of methanol” or boron trifluoride etherate.l*,l9 The use of trimethyl(1.5j T h e n a m e sesquixanthydrnl, suggested for this compound hy Prvfessor P 1) B a r t l e t t , will he used tht-oughorit. (16) H Cilman a n d J U‘.Slorton, Jr., in “Organic Reactions,” V v l V111, J o h n W’iley a n d Sons, I n c . , S e w ’lork, N Y , 1954. (17) A M u s t a f a a n d 0. H H i s h m a t , J . Org C h r m . , 2 2 , 1644 (I%?) (18) 51 C Casetio, J I1 R o b e r t s . hl, Neeman, a n d W S Johnson, J A m C / i f > n S n c . , 80, 2.784 (19.58) (19) E Muller a n d W R u n d e l , Angew. C h p m , 7 0 , 10.5 (1938)
2253
SYNTHESIS OF SESQUIXANTHYDROL
June 5, 1964
SCHEME I
(CdW)nCO
pyridine hydrochloride
I
~
CH307YCHa P q?& /’
CH30
I11
I1
hydrochloride
H201H+
-
NsOH
IV
A
HzO PH 5
OCH,
\ AIBrS
OH..O VI
CH30
OH
anilinium hydroxide or dimethyl sulfate on the salt of the phenol provides low, but synthetically practical, conversions to the desired product. Carbinols I and I1 react in faintly acidic media to yield the corresponding carbonium ions I11 and IV. The proton n.m.r. spectra of the carbinols and their corresponding carbonium ions, summarized in Table I ,
TABLE I1 ELECrRONIC S P E C T R A
Corn-
-Chemical Hpara
shift, P-H,,ra
Hmethoxy
J
H
~
~
~
~ Solvent , H ~
I 3.03 3.59 6.58 8 . 4 zk 0 . 2 CC1, I1 2.67 3 00 .. 9.0 f .2 Dioxane I11 2.21 3.25 6.29 8.7 f .2 CFaC02H IV 1.44 2.12 8.9 f .2 CFaC02H V 2.8SC 6.22d CFaC02H ‘ P a r t s per million relative t o the internal standard, tetramethylsilane, which is taken a s 10; see G. V. D. Tiers, J . Phys. Chem., 62, 1151 (1958). Obtained by interpolation from the tables of K. B. Wiberg and B. J . Nist, “ T h e Interpretation of N M R Spectra,” W. A. Benjamin, Inc., New York, N. Y., 1962. For the 9-aryl group, other shifts not determined. All methoxy protons of V have the same chemical shift.
provide good evidence for the proposed structures.20 In each case the region of the spectrum assigned to the aromatic proteins shows the expected AB2 multiplet structure. Electronic spectral data for these species are summarized in Table 11. Basicities of Carbinols.-The ease of formation of a carbonium ion from a carbinol in acidic medium may be considered to reflect what has been termed the “secondary basicity”21of the carbinol. The expression of this basicity in terms of values of ~ K Rprovides + ~ ~a basis for comparisons of carbinols (Table 111). A modification of conventional method^,^ using spectrophotometric measurements of carbonium ion concentrations, in a series of borate buffer solutions, was used to obtain the ~ K R for + 11. This carbinol (11) is 50yoconverted to its carbonium ion IV in solutions of pH 9.05 (corresponding to a ~ K R + value of 9.05). Its aqueous solutions may be titrated potentiometrically with hydrochloric acid, giving typical weak-base titration curves. I t is particularly instructive to note that the analogous tris-p-anisylcarbinol, which gives a carbon(20) For a discussion of these spectra see J C. M a r t i n , J . Chent. E d i r c . , 80, 286 (lY6l). T h e values of T a b l e I a r e more accurate t h a n those reported previously. (21) V . Gold a n d B. W. V. Hawes, J. Chem. SOL.,2102 (1951). (22) T h e negative logarithm of t h e equilibrium c o n s t a n t for t h e reaction
4 280 ( 3 . 8 6 ) 288 ( 3 . 9 5 ) 240 ( 4 . 6 8 ) Carbonium ion IV’ 282 ( 4 , 4 7 ) 330 (4 52) 452 ( 3 . 9 5 ) 475 ( 3 . 9 3 ) ( s h ) Carbinol V” 215 ( 4 . 7 7 ) ~ ~ ~ 276 ( 4 . 0 3 ) 285 (shoulder) Carbonium ion“ from V 288 ( 4 . 6 2 ) 357 ( 4 . 3 4 ) 450 ( 4 . 0 0 ) 525 ( 3 . 6 2 ) ( s h ) Carbinol I ” 277 ( 3 . 6 6 ) 284 ( 3 . 6 5 ) Carbonium ion I I I d 272 ( 4 . 0 4 ) 315 ( s h ) 522 ( 4 . 2 5 ) Carbonium ion I I I * 522 ( 4 . 3 0 ) Carbonium ion 111’ 522 ( 4 . 2 9 ) Carbonium ion 111’ 322 ( 4 . 2 8 ) a In ethanol. I n 0.1 M aqueous C H 3 C 0 2 H . In 0.1 M aqueous HC1. In 0.5 M aqueous HC1. e In 0.5 M ethanolic I n 0.5 M methanolic HC1. HC1. In 90% aqueous dioxane containing 0.5 M HCI. Compound
h a x , mp (log
Carbinol 11”
TABLE I NUCLEAR MAGNETIC RESONANCE SPECTRA pound
0 OCH3 VI1
~
’
TABLE I11 BASICITIESOF TRIARYLCARBINOLS Compound
*
PKR+
4,4’,4”-Tri( dimethy1arnino)triphenylcarbinol” 9.36 Sesquixanthydrol (11) 9.05 4,4’,4 “-Triaminotriphenylcarbinol” 7.57 2,6,2’,6‘,2”,6”-Hexamethosytriphenylcarbinol >6.5 9-(2,6-Diniethoxyphenyl)-1,8diniethoxyxanthydrol ( V ) 2.5 4,4’,4”-Trimethoxytriphenylcarbinol 0.82 Triphenylcarbinolh - 6.63 4,4 ’,4”-Trinitrotriphenylcarbinolb -16.27 R. J. Goldacre and J. N . Phillips, .I. Chem. SOC.,1724 (1949). Reference 12.
ium ion stabilization by oxonium-type resonance structures related to those important in the sesquixanthylium ion (structure V I I I ) , has a ~ K R of+ only 0.82. This difference of 106in the equilibrium constants for the ionizations of the two carbinols reflects an unusual stabilization of IV which deserves comment.
2254
J . C. MARTINA N D RUSSELLG . SMITH
Vol. 86
pH 6.0-6.5. I t is therefore conservative to say that the ~ K R has + a value greater than (3.5, probably considerably greater. The failure to regenerate carbinol I from I11 can be understood by considering the geometry of 111. The steric interactions of the six o-methoxy substituents + IX must result in an exaggerated angle of twist of the aryl VI11 groups in the propeller-like ion. This places three An examination of molecular models (Stuart-Briemethoxyl groups above and three below the central gleb) indicates that ion IV may exist in a completely carbon, effectively shielding it from nucleophilic a t copolanar configuration with little if any more strain tack. than is present in the tetrahedral carbinol. By conThis twisting of the aryl groups out of the plane of the trast the analogous ion I X , having only two ether ring must result in a decrease in resonance stabilization bridges, which was reportedz3during the course of this and charge delocalization in III.*j I t is therefore work, cannot be completely coplanar. I t is forced into necessary to look elsewhere for an explanation of the a helical configuration by hydrogen-hydrogen interacgreat basicity of I. tions. The steric repulsion between methoxyl groups on difThe usual basicity of both carbinols is reflected in the ferent rings would be expected to be less unfavorable in ionic nature of the corresponding chlorides. The the sp2-hybridized, trigonal carbonium ion than in the chloride hydrochloride of IX is a red solid, describedz3 tetrahedral carbinol.26 It is impossible to construct a as being insoluble in most organic solvents but disStuart-Briegleb model of this molecule. Relief of solving in water with decomposition. Sesquixanthydryl strain in the ionization step provides an obvious drivchloride was isolated as an orange crystalline dihydrate ingforce in a manifestation of steric effects reminiscent of which yields, on heating, its yellow unhydrated form. the rapid ionizations observed with derivatives of triBoth forms are hygroscopic, insoluble in ether or bent-butylcarbinol.6 zene, and soluble in water, methanol, ethanol, or aceThe 9-arylxanthydrol derivatives V in which this tone. An aqueous solution shows an electronic specsource of driving force for ionization is obviously much trum identical with that obtained when I1 is dissolved of 2 , s . The establishment less important, has a ~ K R + in aqueous acid. of one of the ether bridges forming the xanthydryl The evidence a t hand does not allow a quantitative nucleus would be expected to have a very favorable assessment of the importance of the possible coplanarity electronic effect on the stability of the carbonium ion. of IV as a factor in explaining its stability with respect Even so, the comparison of V with I ( ~ K K > + 6.3) to 11. I t is interesting to compare I1 with the closely clearly indicates the importance of an unusual factor, analogous carbinols I and V, which also bear oxygen probably steric, favoring the ionization of I. substituents a t each o-position, but have quite different Electronic Absorption Spectra.-The unusual, very geometry. Formation of the intensely purple carbonintense purple color of solutions of I11 suggests a ium ion I11 from I is rapid a t pH 6.0. I t is not, howpossible effect of the pronounced lack of coplanarity ever, possible to obtain an accurate value of the ~ K R + , of the conjugated systems in this hexamethoxytriphenylsince this reaction does not give a true equilibrium. carbonium ion. The substitution of methoxy groups The addition of excess sodium hydroxide to an aqueinto the p-positions of the triphenylcarbonium ion ous solution of 111 gives no regeneration of the carbinol produces’l a shift of the low energy bond of the parent I , as evidenced by the failure to get any of the characion (431 mp, log E 4.60) to longer wave lengths and teristic color of the carbonium ion on treatment of the higher intensity; for the -l,-l’,l”-trimethoxy derivative solid reaction product with acid. The product mixture this bond is shifted to 483 mp (log 6 5.02). The exdoes contain carbinol V, in addition to a complex mixpected**effect of the twisting of the aryl rings out of the ture of colored products giving infrared absorption in central plane of the ion is to produce a bathochromic the region characteristic of quinone-like carbonyl shift accompanied by a lowering of the extinction coefgroups. Quinonoid products might be expected from ficient. the nucleophilic attacks by hydroxide ion (or water) a t ( 2 5 ) T h i s effect is mil-roied in t h e n m r spectra nf l’ahle I which show t h e the o-position of I I I . 2 4 downfield s h i f t . as one goes f r o m carhinnl tn carbonium ion, f n r t h e {>r‘rt
