stituted for silver nitrate; however, it has the dis- benzene. The combined filtrate, on distillation, advantages that i t is acidic,6 its action is slower gave colorless 1,l-dinitroethane (18.9 g., 78%); and more difficult to reproduce, and lower yields of b.p. 55.5-56' (4.5 min.), n Z o D 1.4341-1.4346, d2'1 desired products are obtained. Other cationic 1.355,MRD (calcd.) 22.68, .URD (found) 23.01, oxidizing agents have been investigated in the neut. equiv. (calcd.) 120, neut. equiv. (found) presence of nitrites; however, none is effective in 122; lit.1b 11.p.55-57' (4 mm.), t z 2 5 1.4322. ~ yielding gem dinitro compounds. Reaction of (11) Organic Chemicals Department, Jackson Laboratories, E. I salts of 2-nitropropane with cupric ammonium d u P o u t d e Xemours a n d Co., Wilmington, Delaware. hydroxide yields 2,3-dimethyl-2,3-dinitr~butane~; DEPARTMEST OF CHEMISTRY with cuprous chloride, cuprous acetate, ammoniacal THEOHIO STATE UNIVERSITY RALPH B. KAPLAN" HAROLDSHECHTER cuprous chloride, or Fehling solution, the principal COLUMBUS10, OHIO RECEIVEDJUKE 14, 1961 product is acetone. Cupric chloride, an acidic reagent, gives propyl pseudonitrole; ferric chloride yields 2,3-dimethyl-2,3-dinitrobutaneand ferric A NEW INTRACRYSTALLINE CATALYST 2-propanenitronate. Anionic oxidants such as ammonium persulfate and sodium peroxide oxidize Sir: An acid (hydrogen) form of synthetic sodium primaryS and secondary' nitronates to vicinal dinitro compounds (oxidative dimers) and carbonyl mordenite having high intracrystalline catalytic derivatives; neutral potassium permanganate gives activity (Table I) has been prepared. Its measthe aldehydegaor ketonegbas the principal product. ured surface area is in the range 400-500 m.2/g. Since the effects of silver or mercuric ions seem (B.E.T. method). To our knowledge, this is the specific in effecting oxidative nitration of a nitro- first complete acid form of an open zeolite (Found: nate, it is possible that introduction of a second Na, 0.08) with high thermal stability (800'). The adsorption of only small molecules by other nitro group during oxidation-reduction may depend on an intermediate complex salt (Equation 5 ) various cation forms of mordenite has been deterwhose decomposition into the corresponding dinitro mined by the studies of Barrer.' In contrast to their results this new acid form mordenite has adcompound is sterically favored.'O sorption properties intermediate between zeolite 03 "A" and faujasite and catalytic properties similar 'I' tE t o zeolite 6i10x".2~334Our results are compatible with the crystal structure of mordenite as recently determined by Illeier.5 Preliminary results show that this material is a unique cation exchanger, 0' 2Ag operating over the entire pH range. Exchanges of A typical procedure for oxidative nitration is di and tri-valent cations, such as l\lg++, Ba++, illustrated for the preparation of 1,l-dinitro- A17++ can be accomplished without change in ethane. A fresh solution of nitroethane (15.0 g., crystal structure from the parent material. 0.2 mole), sodium nitrite (4.0 g., 97% assay) and The catalytic cracking properties of the acid aqueous sodium hydroxide (8.5 g., 80 ml.) was rnordenite are similar to those described for "10X"4 poured into a stirred mixture of aqueous silver including a higher paraffin-olefin ratio than that nitrate (70.5 g., 0.41 mole in 120 ml.), sodium observed with a silica-alumina catalyst in our hydroxide (2-3 drops until silver oxide appeared) experiments. TABLE I and ether (130 ml.) at 0-5'. A cream-colored solid formed immediately and the temperature rose to Cracking of n-decane; 1 hr., 450°, L.H.S.V.0.5 10'. The solid decomposed rapidly with blackenRatios % IsobutaneParaffining and reduction of volume; the temperature rose Catalyst Conversion n-butane olefin to 20'. After a few minutes, the cooling bath was Acid mordenite 36 1.3 4.6 removed and the mixture was stirred for 30 minutes. Silica alumina 19 3.3 3.3 The silver was filtered and washed with etherIn the 450" temperature range, a somewhat ( 6 ) T h e desired nitration reaction does n o t occur in acid solution, higher catalyst deposit observed with decane t h e nitrous acid produced reacts with t h e primary or secondary nitro compounds t o yield t h e corresponding nitrolic acids or pseudonitroles. cracking over mordenite as compared with that (7) See also H. Shechtet a n d R.B. Kaplan, J . A m . Chem. S o c . , 76, from silica-alumina and silica-magnesia in our 3980 (1953) work, suggested possible catalytic activity a t lower (8) A. L. Pagano, P h . D . Dissertation, T h e Ohio S t a t e University, temperatures. Indeed this was observed, when 1960. (9) (a) F. T. Williams, P h . D . Dissertation, T h e Ohio S t a t e Univercracking of n-hexadecane occurred a t temperatures sity, 1958; ( b ) S. S a m e t k i n and E. Posdnjakova, J . Rirss. Phys. C h e m . as low as 300". Ikrhen n-hexadecane was cracked S o c . , 46, 1420 (1913). a t 350" over acid mordenite and silica-alumina (10) A similar mechanism m a y be involved in electrolysis of nitrounder nearly identical conditions, 6 times more nates a n d nitrite ions a t platinum anodes t o give gem dinitro compounds along with vicinal oxidative dimers.Id (b) gem Dinitro comlight hydrocarbon (up to C,) was obtained from the pounds also are formed upon precipitating silver salts of mononitroformer. nates in t h e presence of preformed silver nitrite; gem dinitro compounds also are obtained b y related processes in t h e Victor Meyer reaction for preparing mononitro compounds from alkyl halides a n d silver nitrite. (c) Silver salts of primary nitronates decompose in aqueous suspension t o gem dinitroalkanes, sicinal dinitro compounds, aldehydes, t h e p a r e n t nitroalkane, silver, a n d nitrous acid, t h e nitrite renerated competes so effectively t h a t g e m dinitro compnuads are t h e major t ~ r r , d u t . t s - f i v r n d In nuch N y d t r m r i
(1) R . 11. Barrer, T r a n s . F a r a d a y Soc., 40, 555 (1944). (2) D . W. Breck, W. G. Eversole, R. M . Milton, T . B. Reed and T . L . Thomas, J . A m . Chem. Soc., 7 8 , 5963 (1956). (3) R . AM. Barrer, W. Buser a n d W, F. G r u t t e r , Chimio, 9, 118 (1955). ( 4 ) P. €3. Weisz and V. J . F r i l e t t e , J. P h y s . C h a , n . . 64. 382 (1960). (5) 1 %'. 21. Lreier, private communication,
Aug. 20, 1961
COMMUNICATIONS TO THE EDITOR
Further evidence of low temperature activity was revealed by micro-scale dehydration of ethanol, in which an activated alumina a t 340’ gave about equal quantities of ethylene and ether, while our catalyst gave ethylene exclusively a t 250’. We have observed cracking with a magnesium exchanged acid mordenite, and isomerization of cyclohexene over acid mordenite and its aluminum exchanged derivative. Further work is in progress on several of the more interesting aspects of this material.
3537
The tentative assignment of a 1,2-dihydropyridine structure to the complex is based on the previously observed orientation in closely-related nucleophilic additions of organometallics and metal hydrides to N-heteroaromatics. 2,
RESEARCH AND DEVELOPMENT DEPARTMENT NORTONCOMPANY ALLENH. KEOUGH ~ ‘ O R C E S T E R , MASSACHUSETTS L. B. SAND RECEIVED JUNE 28, 1961
The specific reducing properties of I are illustrated by the facile reductions of benzophenone, 2,4‘-dichlorobenzophenone and hexachloroacetone LITHIUM N-DIHYDROPYRIDYLALUMINUM HYDRIDE-A SELECTIVE REDUCING AGENT FOR to the carbinols, whereas 1-naphthoic acid and HIGHLY ELECTROPHILIC CARBONYL COMPOUNDS methyl 1-naphthoate are recovered unchanged Sir: even after overnight exposure to excess complex. We have reported previously that pyridine is an On the other hand, the reduction of such acids and excellent solvent for the reduction of ketones and esters in pyridine solution proceeds normally to the metalation of certain weakly acidic hydro- alcohols when lithium aluminum hydride is added. carbons by lithium aluminum hydride.’ Car- This selective action of I may be used advanboxylic acids and esters also are cleanly reduced to tageously for reduction of ketone groups in certain alcohols in this medium. In contrast to the above polyfunctional compounds. Several examples, in results, which are obtained when lithium aluminum which the simplicity of execution is illustrated, are hydride is added to pyridine solutions of substrate, recorded: To a 60-ml. serum bottle containing 50 a highly selective reducing agent is formed when ml. of anhydrous pyridine is added 0.50 g. of the hydride and pyridine are allowed to interact powdered lithium aluminum hydride. After the before addition of the substrate. This species,2 initial exothermic reaction, during which the lithium N-dihydropyridylaluminum hydride (I), solution acquires a dull orange coloration, the acts as a weak hydride donor, reacting only with bottle is stoppered tightly and aged for a t least one highly electrophilic carbonyl groups. Its rate of day, after which the lithium aluminum hydride formation is quite rapid (but not as fast as the is essentially consumed (see above). A ten millireduction of a ketone when one is present1), as liter aliquot (containing ca. 2.6 mmoles of I) of this indicated by the results of a series of benzophenone solution4 is withdrawn by a hypodermic syringe reductions carried out by aliquots taken periodi- and transferred to 1-1.5 mmoles of substrate, discally from a solution of lithium aluminum deu- solved in pyridine, the solution then being kept teride in pyridine and analyzing the resultant mix- overnight before hydrolysis and work-up in the ture of benzhydrol and benzhydrol-a-D by in- usual manner.’ Under such conditions of prolonged contact with excess reducing agent, 0-(afrared spectroscopy. naphthoy1)-benzoic acid gave 76% 3-a-naphthylTABLE I phthalide, whereas the methyl ester gave a 78% REDUCTIONS OF BESZOPHENONE B Y LiA1D4 IN PYRIDINEyield of this lactone. In both cases, infrared ex“Age” of solution % C-H amination of the crude products showed only lac6 min. 2 tone absorption (1745 cm.-’) in the carbonyl 10 7 regi~n.~ Furthermore, o-benzoylbenzoic acid and 30 21 and its methyl ester were reduced to 3-phenyl2 hr. 50 phthalide in 93% and 62% yields, respectively. 24 71 As above, the crude reduction products consisted 50 essentially only of the lactone (carbonyl absorption 77 cm.-l). The postulate that hydride transfer from the di- a t If1740 a ratio of lithium aluminum hydride to 0-(ahydropyridine ring is the path involved in carbonyl naphthoy1)-benzoic acid comparable t o that used reduction is strengthened by the above data, as well as by the novel behavior of the complex (see in the above reductions involving I is employed, below). This latter factor rules out the possibility according to the earlier procedure, the conversion that the results shown in Table I are due merely to to diol is complete within 15 minutes. This conreversal of the nucleophilic addition of the hydride (3) (a) K. Ziegler and R . Zeiser, Ber., 68, 1847 (1930): (b) E. A. to the pyridine ring. Such isotope scrambling Braude, J. Hannah and R . P. Linstead, J . Chem. Soc., 3249 (1960), and references cited therein; (c) H. S. Mosher, “Heterocyclic Comcould account for the observed C-H/C-D ratios, pounds,” R. C. Elderfield, editor, John Wiley and Sons, Inc., New but the selectivity of I, as compared with the power- York, N.Y., 1950, Volume 1. Chap. 8. (4) Solutions aged for up to ten days still retain their reducing ful reducing action of lithium aluminum hydride, is not explainable in terms of such an hypothesis. powder. (1) P. T . Lansbury, J . A m . Chem. SOL,88, 429 (1961). (2) F. Bohlmann, Chem. Ber., 86, 390 (1952).
(5) This method of reducing this ketoacid appears to be far superior to sodium borohydride in methanol, from which the starting material is recovered unchanged after five hours at room temperature,
