c. K. BRADSIIERAND J.?hIGS
193s
and takes place at the 3-po~ition.~‘The result agrees with Table I11 if it assumed that under the reaction conditions (fuming sulfuric acid), pyridine1-oxide is present only in the salt form. From the correlations with reactivity it can be concluded t h a t the value of PNO for pyridine-Ioxide is between 0.53P and 0 8258 and the dipole moment calculations suggest a value nearer the higher of the two. By graphical means the charge distribution and atom localization energies niay be determined for pyridine-1-oxide a t BNO = O . T 5 P , which is probably fairly near the true value. i i i ,,,I:
0
.
0
01 i
/I,,,,,,,!
N,,,
.5>
0 -11
\\(I
I
J
11. JONES
Vol. 81
TABLE IV ATOM LOCALIZACION ENERGIES OF ( O N 0 = 0.758) Posilion
A.
A,
A.
2
2.36 2.58 2.44
2.42 2.53 2.47
2.48 2.50
3 4
@H
Calculations.-The LCAO procedure described by CoulsonZ2was employed. Pyridine.-The inductive effect of the nitrogen atom was taken into account by using for carbons 2 and 6, C Y C , , = ~ CYC 4-0.058; the other parameters were CYN = ac 0.5P and PCN = 8. For the pyridinium ion the following substitutions were made, CYN = CYC l.OP and C Y C , , ~= CYC 0.lP. The
+
+
(21) H. S.Mosher a n d F. J. Welch, T H I SJ U U R N A L , 77,2902(1955). (22) Reference 4,pp. 238-256.
Pyridine-I-oxide.-The same procedure was followed as for pyridine. The parameters were: CYO = ac 1.00, CYP; = CYC 0.6p (0.6 instead of 0.5 bec:iuse of the inductive effect of the oxygen atom), C Y C , = ~ CYC O.OSP, BCN = P , PNO = 0.5, 1.0 and 1.5. F o t the salt, a0 = CYC 1.5P, CYN= CYC f 0.658; CYC2,6 = (YC 0.065/3, P N O = 0.5P. N E W BRTJNSLVICK, N.J.
+
+
+
+
[CONTRIBUTION 1‘ROhl T l I B 1)EPARThlI’.N r O F CIIEMISTRY.
Acridizinium Ion Chemistry. 111.‘ rY
2.46
energy levels for the molecular orbitals and the coellicien ts were determined by solution of the secular equations. The atom localization energies result by subtraction of the total energy of the n-electrons of pyridine from the n-electron energy of the “localized states.” As an example, the localized state for electrophilic substitution at the 3-position is shown. For localization at each position a different set of secular equatioris was solved.
~ =~ 0 . 7) 5 ~
+
PYRIDINE-~-OXIDE
DUKEUNIVERSITY]
Reaction with Bases
c. E;.URADSIIER AND JAMES I < E C E l V E D OCTOBER
+
H. JONES*
2, 1958
I t has been deinoribtrateti t h a t phc~iylrnagncsitimbri)niidr attacks acritliziniuni bromide a t ptrsitioii 6 yielding a dihydro derivative ( V ) which can be dehydrogenated to a 6-plirii~lacridiziiiiumsalt. 111 basic mcdia substituted acridizinium bromides as well as the parent compound react w i t h potassiuiil cyanide phenylacetonitrile, acetone and acetophenone. It is suggested t h a t in these cases reaction occurs a t positiuii 6.
Since the first synthesis of acridiziniuln salts ( I ) , 3 papers have appeared concerning the photo-
IV 1,s = I3r
x
11, = OCJI,(NO*)t 111, X = OH
dinierizationh of the system as well as its participation as the diene component i l l the Dicls-Aitler reaction.’ To date it appears that no study has been made of the action of bases on the acridizinium ion or on the simple prototype of the series, (1) F o r t h e previous communication of t h i s series see TIUSJ O U R . N A L , 80,YJI( (1958). (2) T a k e n in p a r t f r o m a thesis s u b m i t t e d by J a m e s H. Jones in partial fulfillrncut of the rerliiirements f o r t h e Ph.D. degree. D u k e University. 1958. (3) C . K. Brddsher and L. E. Beavers, THISJ O U R N A L77. , 4812
(1955). (4) (a) C.K. Bradsher. I,. E. Beavers a n d J. €I. Jones, 22. 1740 (1057).
J. Org Chem.,
the quinoliziniuni ion.4h Excellent models for such an investigation are provided by earlier work6-’’ with pyridinium, quinolinium and isoquinolinium salts. An aqueous solution of acridizinium bromide (I) ( 4 ) ( b ) ADDEDI N PRooP.-In a recent article (J.Chem. SOC.,3067 (1958)) A. Richards a n d T . S. S t e v e n s h a v e studied t h e behavior of bases on b o t h quinolizium a n d acridizinium salts. I t is interesting t h a t under t h e conditions which t h e y used t h e v were unable t o eITect t h e condensation of acridizinium s a l t s with “reactive methylene compuunds” in t h e presence o( babes. ( 5 ) J. 0. Aston. THIS]ouaNAr., 64. 5254 (1030): 53, 1 4 4 8 (1931). ( 6 ) A. Hantzsch a n d M . Kalb. Brr.. 33, 3109 (1899). (7) J. G . Aston a n d P. A . 1,asselle. ’Trlrs J O U R N A L . 66, 4 2 6 (1934). (8) J . G. Astun a n d C. W. MontKumery, i b t d . , 63,4298 (1931). (9) J. Mcisenheimer a n d M . Schiiltze. & r , , 56B, 1353 ( 1 9 2 3 ) . (IO) J . hlrisenlieimer a n d E. S t o t z and R. Bauer. i h r d . . 58B,2320 (lY25). ( 1 1 ) \\’. Bradley a n d S. Jcffrie. J. C h e m . SOC.. 2770 (1954). 112) A . Kaufmann. Bcr., 61, 116 (1018). ( 1 3 ) N . J. Leonard a n d G . W. L r u b n c r . ‘l“19 ] < J U K N A l , 71, 3405
(1949). ( 1 4 ) N. J. 1,eonard (15) N . J. Leonard, 3325 (1951). ( 1 6 ) N J. Leonard (17) K ,J . Leonard
a n d G . It’. I x u b n e r , fbicl., 71, 3 l W i ( I ! J I ! j ) . H. A. D r w a l t . J r . , and C . W. Lciibner i b i , i . , 73, a n d R . L. Foster. i b i d . . 1 4 , 3ci71 (1052). a n d K. 1.. Foetcr. r b i d . , 74, 2 1 1 0 (1952).
April 20, 1959
L\CRIDIZINIUM
when made basic by addition of sodium hydroxide solution produced a red, water-insoluble preci pitate. The precipitate appeared unstable, aiid was not obtained in a state of analytical purity. The freshly prepared precipitate dissolved in ether or benzene and the resulting solution yielded acridizinium picrate (11) upon addition of picric acid. Since it would be very remarkable if ionic acridizinium hydroxide (111) should prove to be soluble in benzene, the most likely explanation is that the pseudo base IV18 is formed instead. Aston and Montgomery8 pointed out that in aqueous solution aromatic quaternary hydroxides are largely ionic, and in equilibrium with only a small amount of the pseudo base, but if the pseudo base is sufficiently insoluble in water nearly all of the hydroxide can precipitate as the pseudo base. The nucleophilic attack of Grignard reagents on aromatic quaternary salts has been the subject of considerable s t ~ d y . ~ - l lJust as had been observed with 1-methylquinolinium saltslo we found that acridizinium bromide (I) reacted with phenylm a p e s i u m bromide to yield a dihydrobase (V) which on refluxing with ethanolic picric acid yielded the picrate of the aromatic quaternary salt VI On oxidation, the dihydro base V yielded a color-
VI1
ION
1930
WITH BASES
6-cyano-7-metliylacridizinium salts. Although the reaction of quinolinium salts occurs a t the o-position with Grignard reagents and para with the CN R
&h“k
u
d
B-
I X , R = H,B = ClOi X, R = CHj, B = OCsH2(h’Oi)s
cyanide ion, it does not seem likely that the cyanide ion would add a t position 2 of the acridizinium ion since the resulting dihydro compound XIA unlike that X I formed by addition a t position 6, would have no benzenoid rings.
H, ,CN
H
XI.4
XI
A final decision concerning the structure of these products as well as of those t o be described subsequently will probably have to wait further advances in our knowledge concerning the synthesis of acridizinium derivatives. Leonard and Foster16 have shown that phenylacetonitrile will react with 1-methylquinolinium iodide in the presence of sodium etlioxide t o yield 1 - methyl - 4 - (a - cyanobenzyl) - 1,4 - dihydroquinoline. When a solution containing 9-methylacridizinium bromide and phenylacetonitrile in ethanol was made basic with sodium hydroxide and refluxed for two hours, an addition product was formed which was isolated as the picrate (XII). Unlike X I the dihydro base XI1 was not affected
VI I I
less crystalline compound having the composition expected for 2-(2-picolinyl) -benzophenone (VII) . This structure was suggested further by the infrared absorption spectrum which showed the presence of the carbonyl group, but indicated that carboxyl was absent. The new diketone (VI) was synthesized from o-benzylbenzonitrile (VIII) by reaction with o-pyridyllithium followed by oxidation. This demonstrates beyond any doubt that phenylmagnesium bromide attacks the acridizinium nucleus at position 6. Kaufman12 showed that 1-alkylquinolinium salts react with the cyanide ion to yield cyano dihydro bases. These bases may be aromatized by the action of iodine yielding 4-cyano- l-alkylquinolinium salts. When an aqueous solution of potassium cyanide is added t o an aqueous solution of acridiziniurn or 7-rnethylacridizinium bromide a dark red precipitate is formed. The precipitate was taken up in methylene chloride and oxidized by boiling the solution with bromine. From the ultraviolet and infrared absorption spectra of the perchlorate salts, i t is quite clear that cyanoacridizinium salts have been obtained. These we have formulated as the 6-cyanoacridizinium (IX) and (18) While this is considered the most likely structure for the pseudo base, at the present time no proof can he offered for its corrcctness.
XII, R‘ = CHa, R H X I I I , R’ = H, R = CHI XIV, R’G R = H
by heating with bromine in an inert solvent. \I-hen the dihydro base XI1 was refluxed for 12 hours with ethanolic picric acid it was not dehydrogenated to yield a 6-substituted acridizinium derivative, but instead yielded some 9-methylacridizinium picrate. Leonard and FosterI6 s h o m d that l-methyl-4-(acyanobcnzy1)-I ,4-dihyclroquinoliiie was so sensitive to the action of picric acid that attempts to prepare a picrate led to the formation of l-methylquinolinium picrate. With phenylncetonitrile, acridizinium bromide and T-methylacridiziniuin bromide gave products believed to be XIV and XIII. I n basic acetone solution 8,9-inethylcnedioxyacridiziniurn bromide affords a mixture from which, by addition of picric acid, a compound believed to be 6-acetonyl-~,~-methylenediosy-S,G-dihydro~cridinizinium picrate (XV) was isolated in 33% yield. The addition of acetophenone to acridizinium bromide appears to give an analogous product (XVI). Under the same conditions, 9-methylacridizinium bromide or 8,9-methylenedioxyacridi-
1930
S V ,K = CIIa, R ' = R ' I T T I , R = C61I5, R' =c €1
OCIIzO
ziniuni bromide yield products the composition of which indicates that tzuo moles of acetophenone have reacted with each mole of the acridiziniuni halide. .It present there is insufficient evidence available to make a decision concerning the structure of these two interesting derivatives.
E~perirnental'~ Absorption Spectra.-The ultraviolet absorption spectra were measured in 95% ethanol solution using a Warren Spectracord spectrophotometer and 1-cm. quartz cells. The itifmred spectra were obtained with a Perkin-Elmer model 21 spectropliotomcter and were measured by the pot:issium bromide pcllet technique. 4-Phenylbenzo[b]-4H-quinolizine(V).-To a solution cnntaining slightly over two equivalents of phenylmagnesiuin broniidc in 75 nil. of ether, 0.5 g. of acridizinium bromide3 was added and the mixture refluxed with rapid stirring for 1.5 hours. T h e inorganic snlts were precipitated b y adtiition of saturated ammonium chloride followed by sodium hydroxide solutions and the solid washed with an acetone-cthcr mixture until the wash solutions were colorless. T h e :icetone-ether wash solution was Concentrated t o 50 inl. and a small amount of water added. On standing the solution afforded 0.23 g . (50.5%) of dark red needles, n1.p. 110-115'. The melting point was not affccted by recrystallization of the compound froin dilute acetone. :lnnl. Cnlcd. fiir C19HlsN: C, 88.70; €€, 5.88; N, 5.41. I'ountl: C , 88.61; 13, 5.93; N , 5.46. 6-Phenylacridizinium Picrate (VI).-By refluxing t h e free base V for several hours in alcoholic picric ncitl, brownish-yellow flnkes were ubtained, m . p . 188-190.5". .Inol. Ciilcd. for C ? ~ T I 1 & ~ O ; ~ I I ~ CO , :59.80; H, 3.61; S ,11.11. Fouiicl: C , 59.96; €1, 3 . 5 5 ; X, 10.99. 2-(2-Picolonyl)-benzophenone (VII). (a) By Oxidation of the Benzoquinolizine Base (V).-To 0.3 g. of the base V in 100 ml.of acetone n 5yo solution of potxssiutii perinangannte was added dropwise until the purple color persisted. The excess permanganate was destroyed by addition of alcohol, the manganese dioxide removed by filtration and i r a s h e d with acetone. From the combined acetone snlutions 0.06 g. of a brown solid melting a t about 106" \vas obtained. On rccrystallization from niethanol it melted a t 132- 133.,cIT3 -
R3 R,
/
I ; Z\ \ C+K1-
RL
/
C\CH3 0
\
-----)
+
Br-
&R
'
I
R I
R;
Rl
I1
I11 VI IV V R, = C H I IS VI11 RZ = CH3 VI1 XI SI1 X RS = CHs S V (picrate) R2 = OCHS XI11 XIT' X V I I (bromide) STIII X\'I R, = OH All R groups not otherwise specified are ussuincd to be hydrogen. (1) F o r t h e previous communication of this series, see J . O Y ~Chem., . 23, 430 (1958). (2) Abstracted f r o m p a r t of a dissertation t o lie presented lis. K . B.
hloser in partial fulfillment of t h e requirements for t h e degree of Doctor of Philosophy at D u k e University. T h i s research was sul,ported by a g r a n t (NSF-G?BGI) from t h e S a t i o n a l Science Foundation. ( 3 ) C . K . Rradsher a n d I.. E . Beavers, THISJOCRXAI.,77, 453 (1955). ( 4 ) T h e n a m e phen3ntliridizinii~m a s a s u h 7 t i t u t e for benzojalquinolizinium ion offers t h e a d v a n t a g e of simlilicity in t h e naming of derivatives, :rnd is consistent with those of t h e quinoliziniiim ancl acridizinium ions. T h e numbering riseil is t h a t recommended I,y Chemical Abstracts.
TABLE I PIIESA?iTfIRIDIZISIC~C s.\I-lS
:~>
Suhctitiiei~t
Yield. qiiateriiization
Cyclizntirin time, h r .
Ri = CH3
(i,j
.io, 5
R? =
..
1; r,
ar,
y i e l d , mc cycliiation
71 64* 0
= CIIj S3 ,-)O 3 R* = o c H 3 00 0 n;, .jO RI OH 62 3n 83 a Yield not deteriiiined, but must be a t least equ:tl t o over-all yield (645%). Ovcr-all yield from arylpyridine. R3
,
The salt XI formed by reaction of iodoacetone with 2-(2-t(ilyl)-p!-ridine (X) proved inuch more difficult to cyclizc than did its isomers l' and T'III. salt (SI) affortlctl only The 2-(~-tol~l)-l)vritliiiiuni
