Nov., 1939
DIPOLEMOMENTS OF [CONTRIBUTION FROM
THE
CERTAIN
NITROC O M P O U N D S AND
3067
AMINES
I?RICK CHEMICAL LABORATORY, PRINCETON UNIVERSITY ]
The Dipole Moments and Structures of Certain Nitro Compounds and Amines BY GEORGE L. LEWISAND CHARLES P. SMYTH The dipole moments of trinitromethane or nitroform, tetranitromethane and nitrogen pentoxide have been determined with the object of studying the molecular structures and the interactions of immediately adjacent nitro groups. A brief account is included in this paper of the determination and interpretation of the moments of cyclohexylamine and morpholine. The moments were obtained with the apparatus and methods previously described. l v 2 Preparation of Materials Benzene.-Benzene was purified as in the work previously described.' Carbon Tetrachloride.-Carbon tetrachloride was purified as in the work previously described.* Tetranitromethane.-Tetranitromethane, prepared according to the method of Chattaway3 by the action of virtually dry nitric acid on acetic anhydride, was dried over anhydrous sodium sulfate and distilled in an all-glass apparatus under reduced pressure a t 35'; nZ5D 1.43822; dZ54 1.62294. Nitroform.-This compound was prepared according t o the method of Hantzsch and Rickenberger' by the acidification with concentrated sulfuric acid of KC(NO&, which was formed by the action of alcoholic potash upon tetranitromethane; m. p. 19". Nitrogen Pentoxide.-The method of preparation was essentially that of Russ and Pokorny.6 Fuming nitric acid was mixed with an equal volume of concentrated sulfuric acid and distilled from it into phosphorus pentoxide, which was kept a t about -20" and stirred until the mixture had the consistency of a paste. The resulting nitrogen pentoxide was sublimed in a stream of ozonized oxygen and collected along with oxygen and ozone in a tube immersed in liquid air. Upon warming after removal from the liquid air, the oxygen and ozone evaporated, leaving the white solid. Cyclohexylamine.-Material from Merck and Company was distilled, dried over calcium chloride and then sodium, distilled onto fresh sodium in an atmosphere of nitrogen, and redistilled: b. p. 133.8"; P a 0.86253. Morpho1ine.-Material very kindly given us by the Carbide and Carbon Chemical Corporation was dried over calcium chloride and then over sodium, distilled onto fresh sodium in an atmosphere of nitrogen, and redistilled. The middle fraction was retained and measured without further purification; b. p. 128.7'; n % D 1.45227; d%r 0.99619.
Experimental Results The dielectric constants E and densities d of the (1) Lewis and Smyth, J . Chem. Phys., 7, Nov. (1930). ( 2 ) Lewis and Smyth, THISJOURNAL, 61, 3063 (1939). (3) Chattaway, J. Chem. S O L . 97, , 2099 (1910). (1) Hantzsch and Rickenberger, Ber., 83,628 (1899). ( . 5 ) Russ and Pokorny, Monafsh., 84, 1027 (1913).
solutions containing mole fraction c2 of the polar solute and the polarizations PZof the polar solute are given in Table I, while the molar refractions MRD, the polarizations P, obtained by extrapolating the Pz-cz curves to infinite dilution, and the dipole moments p are listed in Table 11. TABLEI DIELECTRICCONSTANTS, DENSITIESAND POLARIZATIONS AT 25' Cl
d
e
Pa
Carbon Tetrachloride-Tetranitromethane 1.58431 (Pi = 28.18) 0.00000 2.227 1.58241 42.0 .04039 2.240 1.58160 41.2 .08421 2.251 40.5 2.264 1.58268 .14291 40.4 2.275 1.58364 .18437 1.58527 40.3 ,26079 2.295 1.58639 40.7 .30898 2.317 1.00000 2.521 1.62294 40.62 Carbon Tetrachloride-Nitroform 2.308 1.57998 0.00969 2.339 1.57888 .01350 2.466 1.57345 .02944 2.546 1.56703 ,03994
169.8 167.0 160.3 157.7
Carbon Tetrachloride-Nitrogen Pentoxide 0.00000 2.223 1.5844 (Pi = 28.11) 2.368 1.5813 58.6 .05934 1.5800 57.9 .08624 2.435 1.5793 57.6 ,09361 2.452 1.5790 57.3 .lo243 2.473
0.00000 .12935 .20884 .27445 0.08588 .18193 .22967 .27621 .35608 1.00000
Benzene-Cyclohexylamine 2.276 0.87344 (Pi = 26.67) 65.7 2.559 ,87084 65.8 2.743 .86946 65.6 2.892 .86815 Benzene-Morpholine 2.574 0.88422 .89684 2.934 .90295 3.128 3.316 .go875 .91851 3.665 7.33 .99619
TABLE I1 MOLARREFRACTIONS, POLARIZATIONS MOMENTS Substance
MRD
PCV
C(N02)r CH(NOz13 Nz06 CsHiiNHz OCdHsNH
31.72 25.01 20 30.5 23.59
41.7 177 60 65.7 75.3
73.4 71.6 70.7 69.9 68.7 59.3
AND
DIPOLE p
x
101s
(0.71) 2.71 1.39 1.32 1.58
3068
GEORGEL. LEWISAND CHARLES P. SMYTH
Vol. 61.
and OttensI2 31.44 and is higher than the sum 29.4 of the refractions of four alkyl nitro groups and one carbon atom taken from Landolt-Bornstein. The value 31.7 is, therefore, to be regarded as the molar refraction. This leaves a difference between the polarization and the refraction of about 10 cc., which would give a moment 0.7. The moment of the symmetrical tetranitromethane molecule should, of course, be zero, while that of trinitromethyl nitrite would depend Discussion of Results upon rotation around the single 0-N bond of the Tetranitromethane was measured because, al- nitrite group and would vary from 3.4 to 0 with though two previous investigators had found variation in the position around this bond. Commoments for i t in solution so small as to be indis- plete freedom of rotation would give a moment 2.4, tinguishable from zero, absorption measurements7 but repulsion within the molecule should make and X-ray analysis of the crystal* had been inter- the general region of the zero moment position preted as indicating that it was not a symmetrical more probable, which would mean an average molecule but actually trinitromethyl nitrite. moment possibly less than 1.0. The possibility Mark and Noethling assumed a linear nitrite that the tetranitromethane is really trinitrostructure, which now seems highly improbable. methyl nitrite is not wholly excluded if the moAfter the measurements in the present paper had ment is as large as 0.7. However, measurements been carried out, a determination in the vapor on nitrobenzene in the vapor statel3 gave an state was published indicating a moment indistin- atomic polarization 6 for nitrobenzene as comguishable from zero. The results of these meas- pared to 1.0 for fluorobenzene and 1.8 for chlorobenzene, and similar measurements on nitrourements are summarized in Table 111. methaneI4 gave an atomic polarization 6.1 as TABLE 111 compared to 1.2 for phosgene and 0 for hydrogen DATAON TETRANITROMETHANE cyanide obtained a t the same time and small Temp, (P x Investigators 'C llKn I'm IO'S) Solvent values for the alkyl halides obtained a t other P/illiamss 25 :il 33 6 0 2 CBHR times. The large errors which commonly acWeissberger and cumulate in atomic polarization values make any 36 2 18 CCl4 Sangewaldlo 25 3-5 3 discussion of i t necessarily approximate, but it 39 Gas CoopandSuttonll 81 8 36 0 36: ti would appear that a large atomic polarization may 71 CCI, Lewis and Smyth 2,i 31 72 41 7 be associated with the nitro group. If the atomic As Coop and Sutton say that the polarization polarization of two mononitro compounds is A, may be "somewhat greater than the above value, i t would not be surprising for a molecule con hut not by more than :3 cc.," and, as our value tairiing four nitro groups to have an atomic may be as much as I cc. high, the agreement is polarization of 10. I t seems reasonable, theresatisfactory. The value of Weissberger and fore, to attribute the difference between thr Sangewald is also low by probably iio more than polarization and the refraction to atomic polarithe experimental error. Coop and Sutton's value zation and report the moment as zero and thc of the molar refraction was taken as the average molecule, therefore, as the symmetrical tetraof the value of Weissberger and Sangewald, nitromethane and not trinitromethyl nitrite. A which may have a considerable error because of determination of structure by the electron diffracbeing determined in dilute solution, and of a value tion method just published15 obtains the interprivately communicated to them by LM.W. atomic distances to be expected on the assumption Lister. The value 31.72 determined in the pres- of a symmetrical tetranitromethane. Before ent work agrees closeIy with that of von Auwers leaving the subject of tetranitromethane, atten(6) Von Auwers and Harres, Rev., 64B,2287 (l(129) tion should be directed to the somewhat unusual (7) Harper and Macbeth, J . Chem Soc., lM,87 (1915).
The molar refractions were calculated for tetranitromethane and morpholine from the refractive indices and densities herein reported and that for nitrogen pentoxide as the refraction of n'~O4plus half that of O2 obtained from data in LandoltBornstein. The molar refraction of cyclohexylamine was calculated from similar data in LandoltBornstein and that of nitroform was the value of von Auwers and Harres.G
(8)Mark and Noethling, Z.Kyrst., 65, 435 (1927) (9) Williams, Physik. Z.,B , 2 7 (1928). (10) Weissberger and Sangewald, Be?., 65, 701 (1932) (11) C'oop and Sutton .r. Chem Soc , 1269 (l'J38).
(12) Von Auwers and Ottens, Ber , 51B,446 (1924). (13) McAlpine and Smyth, . I Chem. P h y s , 3, 55 (1935). (14) Smyth and McAlpine, THISJOURNAL, 56, 1697 (1934) (15) atosick, THIS J O U R N A L 61, 1127 (1339)
Nov., 1939
DPOLE MOMENTS OF CERTAINNITROCOMPOUNDS AND AMINES
3069
occurrence of a minimum in the density-concen- bent and a very rough value of the oxygen valence tration curves in the very dilute region of its solu- angle may be calculated from the group and bond tions in carbon tetrachloride. moments involved. If from the solution value of No account of atomic polarization has been nitromethane 3.2, the approximate moment of taken in the calculation of the moment of nitro- the H-C bond 0.3 is subtracted, a moment 2.9 form in Table 11. The addition of 8 cc. to the is obtained for the nitro group. If induction rerefraction, a maximum value for atomic polariza- duces this by 10% as compared to the 15% retion, would lower the moment by only 0.07. The duction in nitroform, the group moment is 2.6. solvent effect has presumably already lowered it This moment in each half of the molecule is opby more than this from its true gas value. The posed by the N-0 bond moment, which is about ratio of the value 2.71 to the solution value of 0.3.l' The moment of each half of the molecule nitromethane is 0.85, while the ratio of the is, therefore, 2.3, acting approximately in the chloroform moment to that of methyl chloride is direction of the oxygen valence bonds. The 0.56. If there were no inductive action between moment of the molecule as a whole should, thereneighboring dipoles, the moments of the mono- fore, be 2 X 2.3 cos (8/2) = 1.39, where 8 is the and the tri-substituted molecules should be equal oxygen valence angle. This gives 8 = 145O, as required by their tetrahedral structures. The which seems rather large. However, electron lowering of the moment by induction is much diffraction's gives an oxygen valence angle 100' in less in nitroform than in chloroform, for not only OF2 and 115' in OCl2, in which latter molecule is the ratio of the tri- to the mono-substituted the repulsion between the two attached atoms compound 0.85 as compared to 0.56, but the ab- should be larger. Calculation of the dimensions solute lowering is only 0.5 as compared to 0.8. of the OClp and O(N02)2 molecules shows that, for The total polarizability of the nitro group is a given oxygen valence angle, steric repulsion slightly greater than that of chlorine as shown by between the two nitro groups must be much its refraction of 6.7 as compared to 5.97, but its greater for certain positions in the rotation of the polarizability per atom and per displaceable elec- groups about their bonds to the central oxygen tron is only a little more than a third of that of than the repulsion between the two chlorine chlorine and that for the chlorine electrons. In atoms. Thermal oscillation around these bonds some positions of the nitro groups as they rotate should, therefore, tend to widen the oxygen around the C-N bonds, an oxygen of one nitro valence angle and may be sufficient to account for group may come very close to that of another with the large value 145' estimated approximately. a considerable resultant lowering of the N-0 di- It should be remarked that the bond moment pole moments, but if, as i t was found necessary to calculated from the moment and valence angle assume in the electron diffraction investigation of of the OCl2 molecule was consistent with other tetranitromethane16 and as should be equally indications of its size.lg Neglect of atomic polariprobable here, the nitro groups are undergoing zation in calculating the moment of nitrogen pentrotatory oscillation about the C-N bonds in such oxide presumably raises its value somewhat, a way as to keep the oxygen-oxygen repulsions but the presence of impurity due to decomposiminimized as far as possible, the oxygens are tion of the unstable substance tends to compensomewhat removed from neighboring dipoles and sate for this. It is thus clearly established that the inductive effects should be smaller than those the molecule is bent, but the angle of bending is in chloroform. The smaller reduction in dipole only very approximately estimated. The moment of cyclohexylamine 1.32 is 0.2moment is thus accounted for. The nitrogen pentoxide molecule should con- 0.3 higher than those of secondary alkyl amines, sist of two nitro groups attached to a central about the same size as those of the primary alkyl oxygen. A rather early electron diffraction in- amines, and smaller than those of the aryl privestigation16 indicated a symmetrical molecule mary amines as would be expected. Its behavior with a 180' valence angle for the central oxygen. is thus normal. The moment of morpholine 1.58 in Table I1 is As this structure would have zero moment, the slightly higher than the value 1.48 recently pubconsiderable moment 1.39 given in Table I1 shows (17) Smyth, THISJOURNAL, 60, 183 (1938). it to be impossible. Evidently, the molecule is (16) Maxwell, Mosleg and Deming, J . Chcm. Phys., I, 331 ( l W 4 ) .
(18) Brockway, Rev. Modern Phys.. 8,231 (1936). ( I D ) Smyth, J , Pkrr, Chew,, 4L POQ (19a7).
3070
WILLIAM G.
YOUNG,
LAWRENCE RICHARDS AND
lished by Partington and CoomberlMwhich does not affect its interpretation. The dioxane molecule, which has two possible forms, a Z-shape and a U-shape, has been shown by its moment21s22 to exist almost wholly in the 2-form. Similarly, the moment found for morpholine, which has two possible 2-forms and four possible U-forms, agrees well with the identical values calculated for the two 2-forms. Our reasoning has agreed with that of Partington and Coomber in leading to the conclusion that the morpholine molecule exists primarily in its 2-forms, although the presence of U-forms is not entirely excluded. Summary The dipole moments of tetranitromethane, nitroform and nitrogen pentoxide have been de(20) Partington and Coomber, Nolure, 141,918 (1938). (21) Hunter and Partington, J . Chem. SOC.,87 (1933). (22) Smyth and Walls, T~rrsJ O U R N A L , 63,2115 (1931).
[CONTRIBUTION FROM THE CHEMISTRY
JULJAN kZORLOSA
Vol. 61
termined in carbon tetrachloride solution and those of cyclohexylamine and morpholine have been measured in benzene solution. Tetranitromethane is concluded to have a large atomic polarization but zero dipole moment, which means that it is not trinitromethyl nitrite as sometimes suggested. The moment of nitroform is consistent with its tetrahedral structure, inductive effects in the molecule being smaller than those in chloroform, probably because of the lower polarizability of the atoms and of a staggering of the nitro groups in their rotation around their bonds to the central carbon atom. The moment of cyclohexylamine is of the same size as those of the primary alkyl amines, as it should be. The moment of morpholine shows that it exists primarily in the 2- or chair-shaped forms of the molecule. PRINCETON, NEWJERSEY
RECEIVED SEPTEMBER 5, 1939
DEPARTMENT OF THE UNIVERSITY OF CALIFORNIA
AT L O S ANGELES]
Allylic Rearrangements. M. The Isolation and Rearrangement of Primary and Secondary Pentenyl, Hexenyl and Heptenyl BromideslJ BY WILLIAMG. YOUNG,LAWRENCE RICHARDS AND In the first paper3of this series i t was recognized for the first time that the so-called crotyl bromide was actually a mixture of primary and secondary butenyl bromides which were far more mobile than previously suspected. Although they could be separated by low-temperature fractional distillation] they rearranged even a t room temperature to an equilibrium mixture. Later papers4 have pointed out that studies involving the formation or reaction of these butenyl bromides must take into consideration the possibility of an allylic rearrangement during the reaction] and also a thermal rearrangement after the products are formed. Since other alkenyl bromides have been (1) This work was accomplished with the aid of a grant from the Board of Research of the University of California. (2) The terms primary alkenyl and secondary alkenyl bromides are used t o designate the two allylic isomers, R-CH=CHCHzBr and R-CHBrCH=CHn respectively. Accordingly, the terms primary pentenyl, hexenyl and heptenyl bromides refer to 1-bromo-2pentene, l-bromo-2-hexene, and l-bromo-2-heptene, respectively. Similarly, secondary pentenyl, hexenyl and heptenyl bromides refer to 3-bromo-1-pentene, 3-hromo-1-hexene, and !&bromo-l-heptene, respectively. (3) Winstein and Young, THIS JOURNAL. 68, 104 (1936). (4) (a) Young, Winstein and Prater, ibid., 68, 289 (1036); (b) Youngand Winstein, ibid., 68,441 (1936); (c) Youngand Lane, ibid., 69,2081 (1937); (d) 60, 847 (1938): (e) Young, Lane, Loshokoff and Winstein, ibid., 69, 2441 (1937); (f) Young, Kaufman, Loshokoffand h b 9 n i a H ; ( b i d . , 94, 9OO (IiW8).
JULIAN
AZORLOSA
prepared and used in the synthesis of many types of compounds without recognition by investigators of the need for carefully controlled conditions, there exists, therefore, a need for a careful and extended study of the alkenyl bromides, their equilibria and reactions. It is the purpose of this paper to discuss the separation and thermal rearrangement of the primary and secondary pentenyl, hexenyl and heptenyl bromides. A separation of the hexenyl bromides6" and the synthesis of several pure primary alkenyl bromides5bhave been reported, but it will be shown that sufficient precautions were not used in the isolation of the pure compounds. Experimental Part Preparation of Materials
Alkylvinylcarbinols.-The ethyl, propyl and butylvinylcarbinols used as a source of the desired vinyl bromides were prepared in 60% yields by the action of the appropriate Grignard reagent on acrolein according to the method of Delabye as modified by Prevost.' (5) (a) Van Riueghem and Gredy, Compl. rend., 4033, 489 (1036); (h) Delaby and Lecomte, Bull. SOC. chim., [ 5 ] 4, 740 (1937). (6) Delaby, Cornfir. rend., 176,967 (1922). ( 7 ) Prbvost. A x n . c h i - . , 1101 10, 113, 147 (1928).
