<<

This dissertation has been 62-2176 microfilmed exactly as received

KAPLAN, Ralph Benjamin, 1920- NEW REACTIONS OF NITRO COMPOUNDS.

The Ohio State University, Ph.D., 1950 Chemistry, organic

University Microfilms, Inc., Ann Arbor, Michigan NEW REACTIONS OF NITRO COMPOUNDS

DISSERTATION

Presented In Partial Fulfillment of the Requirements

for the Degree Doctor of Philosophy in the

Graduate School of the Ohio State

University

By

RALPH B. KAPLAN, B.A.,

The Ohio State University

1950

Approved by:

Adviser TABLE OF CONTENTS

Page

INTRODUCTION

Acknowledgments

Statement of the Problem

SECTION I

REACTION OF NITROALKANES WITH VARIOUS NITRATING AGENTS 1

I. Discussion 1

A. Introduction 1

B. Agents Investigated 2

1. Nltryl Chloride 2

a. Preparation 2

b. Review of thé Reactions of Nltryl Chloride With Organic Reagents U

c. Reaction of Sodium 2-Nltropropane With Nltryl Chloride 6

d. Reaction of Sodium 2-Nltropropane with Nltryl Chloride and Aluminum Chloride 11

2. Mixed Nitric and Sulfuric Acid 12

a. Reaction of Sodium 2-Nltropropane With Mixed Nitric and Sulfuric Acid 12

3. Methyl Nitrate Ih

a. Reaction of Sodium 2-Nltropropane and of Potassium Nltroethane l4

I. Alkyl Nitrates as Nitrating Agents 14

II. Discussion of Results 15 Page

II. Experimental 21

A. Agents Investigated 21

1. Nltryl Chloride 21

a. Preparation 21

1. Intermediates; Chlorosulfonlc Acid and , Anhydrous

11. Procedure for Making Nltryl-Chloride

h. Reaction of Nltryl Chloride with a Secondary Nltroalkane 2k

1. 2-Nltropropane

11. Sodium 2-Nltropropane

c. Reaction of Sodium 2-Nltropropane, Nltryl Chloride and Aluminum Chloride. 27

2. Reaction of Sodium 2-Nltropropane With Mixed Nitric and Sulfuric Acids 28

3. Methyl Nitrate 29

a. Reaction of Sodium 2-Nltropropane 29

h. Reaction of Nltroethane (Attempted Preparation of 1,1-Dlnltroethane). 52

III. Summary 35

SECTION II

REACTION OF NITRIC OXIDE WITH SODIUM NITROALKANES 34

I. Discussion 34

A. Introduction 34

B. Agents Investigated 35

w Page

II. Experimental hk

A. The Agent th

1. Kitrlc Oxide

a. Preparation kh

b. Reaction With Sodium 2-Nitropropane (Formation of a Mixed Salt (l:l) of Sodium 2-Fitropropane and ).

c. Reaction With Sodium 2-Nitrobutane (Formation of a Mixed Salt (l:l) of Sodium 2-Nitrobutane and Sodium Nitrite)

d. Reaction With Sodium Nitrocyclohexane .(Formation of a Mixed Salt (l:l) of Sodium Nitrocyclohexane and , Cyclohexanone, Cyclohexanone Oxime, and l,l-Dinitro-l,l-Dicyclohexyl

B. Chemical Proof of Constitution of the Salts Derived frcrni Sodium Secondary Nitroalkanes and Nitric Oxide 4?

1. Acidification

a. The Double Salt. Conversion to Pseudonitrole 4?

b. Disodium l-Nitrosohydroxylamino-l- and Disodium 1-Nitrosohydfoxylamino-1- nitropropane (Traube Products). 4°

2. Oxidation With Silver Ion 50

a. The Double Salt. Preparation of Secondary Gem-Dinitro Alkanes 50

1. 2,2-Dinitropropane

ii. 2,2-Dinitrobutane

iii. 1,1-Dlnitrocyclohexane

b. Traube's Products. Degradation to Water Soluble Products. 55

lU Page

5 . Oxidations With Hydrogen Peroxide and With Salts of Persulfuric Acid 5U

a. Coupling of Secondary Nltroalkane Anions to Dimeric Vicinal Dinitroalkanes 5^

i. 2,3-Dimethyl-2,5-dinitrobutane

ii. 3,4-Dimethyl-3;4-dinitrohexane

iii. 1,1'-Dinitrohicyclohexyl

iv. l,l'-Dinltrobicyclohexyl

4. Action of Various Oxidizing Agents on the Doublt Salt of Sodium 2-Nitropropane and Sodium Nitrite 56

a. Oxidation with Mercuric Nitrate. Formation of 2,%-D methyl-2,3-dinitrobutane 56

b. Oxidation With Hypobrcmite. Formation of Water Soluble Products 56

c. Oxidation with Permanganate. Fonnation of Acetone 57

d. Oxidation With Ferric Chloride. Formation of Acetone. 57

III. Summary . 58

SECTION III

PREPAEATION OF GEM-DINITEOCOMPOUNDS BY OXIDATIVE NITRATION 60

I. Discussion • 60

A . Introduction 60

B. Historical Review of Various Methods for the Preparation of Gem-Dinitro-Cr.mpounds 60

C. Discussion of Results 65

1. Reaction of Salts of Nitro Compounds With Sodium Nitrite and Silver Nitrate. Scope of the Reaction 63

a. General Procedure, Conditions for Reaction, Isolation of Products 65

\M Page

b. Identification of Products 68

c. Side Reactions 75

2. The Preparation of Potassium Rinltromethane, Homologs and Related Compounds 7^

5. The Effect of Various Oxidizing Agents in Oxidative Nitration 30

II. Experimental 02

A, Oxidative Nitration;,the Preparation of Gem-Dinitro Compounds 82

1. Preparation of 1,1-Dinitroethane 82

a. 1,1-Dinitroethane (as such) 82

h. Potassium 1,1-DJnltroethane 85

c. Silver 1,1-Dinitroethane 84

2. Preparation of 1,1-Dinitropropane 84

5 . Preparation o_ 2,2-Dinitropropane 85

4. Preparation of 2,2-Dinitrobutane 65

5 . Preparation of l,l-Dinitrodj'clohexane 66

6. Preparation of 9,9-Dinitrofluorene, 9,9'-Dinitro- 9,9'-difluorenyl and Fluorenone 87

7. Preparation of 2,2-Dimethyl-l,l,5-trinitropropane 89

a. Formation of the Intermediate, the Disodium Salt of 2,2-Dimethyl-l,5-dinitropropane 89

b. Nitration of the Intermediate Salt to Yield 2.2-Dimethyl-l,l,5-trinitropropane 89

8. Preparation of 2,2-Dinitro-l-ÿroÿanol 91

a. Sodium Salt of 2-Nitro-1-ÿropanol 9I

b. Nitration of the Intermediate Salt to Yield 2.2-Dinltro-l-propanol 91 c. Preparation of Potassium 1,1-Dinitroethane 92 V v f Page

9. Preparationof 5,3-I>initro-2-butanol 92

a. Salt Product From 2-Nltro-2-butanol 92

b. Conversion of the Intermediate Salt Into 3#5-DJnitro-2-butanol.and 1,1-Dinitroethane 93

10. Preparationof 2,2-Dlnitro-l,3-propanediol 9^

a. Sodium 2-Nitro-l,3-propanediol 9^^

b. Conversion of the Intermediate Salt into 2,2-Dlnitro-l,3-propanediol 94

c. Preparation of Prtassium 2,2-Dlnitroethanol 95

11. Preparation of Potassium Dlnitromethane from l-Nitro-2-propanol 96

a. Sodium Derivative of l-Nitro-2-propanol 96

b. Conversion of the Intermediate Salt Into l,l-Dinitro-2-propanol (crude) 96

c. Cleavage of l,l-Dinitro-2-propanol 98

i. Potassium Dlnitromethane

ii . Silver Dlnitromethane

d. Cleavage of l,l-Dinitro-2-propanol to Acetaldehyde 100

12. Preparationof Dlnitromethane from 101

B. Attempted Preparation of Potassium Dlnitromethane from Potassium 2,2-Dinitroethanol 102

1. Preparationof Dipotassium 1,1,3,3-Tetranitropropane 102

a. Dipotassium 1,1,3,5-Tetranitropropane (as such) 102

b. Disilver 1,1,3,5-Tetranitropropane 103

C . Preparation of Dipotassium 1,1,3,5-Tetranitropane from Potassium Dlnitromethane and Potassium 2,2-Dinitroethanol I05

VI Page

D. The Effect of Various Oxidizing Agents on the 2-Propanenltronate-Nltrlte Ion System 106

1. Oxidation With Mercuric Ion. Preparation of 2.2-Dlnitropropane 106

2. Oxidation With Persulfate Ion. Preparation of 2 .3-Dlmethyl-2,5-dlnltrohutane 107

3. Oxidation With Hydrogen Peroxide. Preparation of 2.3-Dimethyl-2,3-dinitrohutane 107

4. Oxidation With Cuprlc Ammonium Complex Ion. Preparation of 2,3-Dlmethyl-2,3-dinltrohutane 1C8

3. Oxidation With Cuprous Salts. Formation of Witer Soluble Products 1C9

III. Summary 109

SECTION IV

OXIDATIVE COUPLING OF SALTS OF NITROALKANES TO YIELD VICINAL DINITRO COMPOUNDS 112

I. Discussion 112

A. Introduction 112

B. Historical Review of Coupling Reactions 115

C. Discussion of Results 115

1. Nitro Compounds and Oxidizing Agents Selected for Investigation 115

a. Oxidation of Salts ofNitroalkanes With Persulfates 115

h. Reaction of Sodium 2-Nitropropane With Hydrogen Peroxide ll8

c. Reaction of Sodium 2-Nitropropane With Potassium Ferricyanide ll8

d. Reaction of Sodium 2-Nitropropane With Other Oxidizing Agents IIQ

2. Correlation of Data 120

Mil Page

11 « Experimental 12k

A. Oxidative Coupling (Dimerization) of Secondary Nitroalkanes 12U

1. Preparation of 2,3-Dimethyl-2,3-dinitrobutaneUsing Various Oxidizing Agents 12k

a. Coupling by Persulfate Ion 12k

1. Formation of Acetone by a Side Reaction

b. Couplingby Hydrogen Peroxide 126

c. Coupling by Potassium Ferricyanide 127

d. Coupling by Cupric Ammonium Complex Ion 127

2. Preparation of 3,k-Dimethyl-3,k-dinltrohexane 128

3 . Preparation of l,l_-Dinitro-l,l'-dicyclohexyl 128

k. Attempted Preparation of 2,2,3,3-Tetranitrobutane 129

3 . Attempted Preparation of 129

III. Summary 129

BIBLIOGRAPHY I3I

AUTOBIOGRAPHY

ViU ACKN0WLE3XÏMENTS

The author wishes to express his appreciation for the encouragement and counsel of Professor Harold Shechter.

The author is indebted to Mrs. Elizabeth Klotz and

Mrs. Arlene Brooks and many others for their cooperation at various times during this research.

To The Ohio State University Eesearch Foundation, the

Aerojet Engineering Corporation, Azusa, California, and to the

Office of Naval Eesearch the author expresses his sincere gratitude for the financial support that made this work possible.

IX Statement of the Problem

Advances in the chemistry of polynitronlkanes have been delayed for a long time by lack of adequate methods of synthesis. This is true because the compounds are sensitive and will not survive dras­ tic nitrations. It would be very desirable to have methods which would permit of adding nitro groups to the molecule under very mild conditions.

An investigation of the reactions of salts of nitro compounds with various potential nitrating agents has therefore been initiated for the purpose of developing general methods for converting mono- nitro compounds into polynitro compounds. SECTION I

REACTION OF SODIUli NITROALKANES WITH VARIOUS NITRATING AGENTS

I « Discussion

A, Introduction

Primary and secondary nitroalkanes are pseudo acids; neutral-

ization of these weak electrolytes (K = 10 ) yields the nitronate

anion, R-NOg (l), a resonance hybrid of the following structures:

y O - _ y) r C l o RR'-C=N( RR'-C-n ' RR'-C-N^ ^0 \ \o - '

I II III

Of the structures represented, II is of least importance (2) because

of the strongly electronegative character of the nitro group. The

overall structure of the hybrid may best be represented schematically

by structure III; however it undergoes reactions characteristic of

representations I and II, For example the alkane nitronate ion may combine with a positive reaction center to form a new bond by

which the group entering is attached to oxygen or to carbon of

structure III, Reactions producing attachment on carbon are usually

described as those of the nitro form, whereas those yielding adducts

bonded to oxygen are derivatives of the aci-nitro structure, A few

examples (3,4) which illustrate the two types of reaction are, (l)

recombination of the anion with a proton to yield the nitronic acid,

RR'-CsNOgH; the nitronic acid exists in equilibrium with the nitro -2-

form of the conjugate acid, BE'CHNOg: (2) reaction of the anion

with alkylating agents to yield either nitronic esters, EE'CsNOgR

(O-alkylated products), or nitroalkanes, EE'RWCNOg (C-alkylated

nitro Compounds); the factors which influence these competing

reactions are the structures of both the alkylating agent and the

nitroelkane anion (5, 6, f); (3) condensations with aldehydes

and ketones to yield nitro alcohols and# (4) formation of substi­

tuted halonitromethanes by reaction with halogenating agents.

Since nitroalkane anions react with many electrophilic reag­

ents to yield derivatives substituted on carbon, it was of inter­

est to examine the action, on salts of nitroalkanes, of reagents

capable of furnishing the , This offers the possibility of preparing gem dinitroalkanes:

1. EB'-C=N02 + HOgT * BE'CCKOgïg + Y"

Of many potential nitrating agents of the type NOgY, it was exped­

ient to investigate nitryl chloride (NO^Cl), mixed nitric acid and

sulfuric acid (EOg-S^SO^), and alkyl nitrites (RO-NOj) on salts of nitro compounds.

B. Agents Investigated

1, , Nitryl Chloride

a. Preparation

When this investigation was undertaken, two methods

of preparation of nitryl chloride were known: (l) the oxidation

of with Ojone (Equation 2) (6), and (2) the -3- reaction of nitric acid with chioresulfonic acid (Equation 3) (?)•

Recently, pure nitryl chloride was obtained from pentoxide and . (8)

2 . NOCl + 0^ -- . NOgCl + Og

3 . HONO„ + ClSO H HO Cl + H.SO. 2 3 2 2 4 Meiny other methods for preparation of nitryl chloride have been reported; however, critical survey of the literature revealed that these methods are inefficient and yield impure products

(8, 9i 10, 11, 12), The ozonization of nitrosyl chloride yields pure nitryl chloride and its physical constants have been adequate­ ly described; however, the method is not adaptable for prepara­ tion of the quantities necessary for the present investigation,

The reaction of chlorosulfonic acid with nitric acid is reported to give nitryl chloride in excellent yield; however, neither the physical constants nor the purity of the product has been deter­ mined. Since the reaction of nitric acid with chlorosulfonic acid is relatively simple, it was investigated as a method for prepara­ tion of nitryl chloride on a large scale, The laboratory procedure thus developed, a modification of the method of Dachlauer (9), gives nitryl chloride of 98-99 per cent purity in yields of 8O-9O per cent. The product is pale yellow, since it is contaminated with small amounts of chlorine and nitrogen dioxide; its boiling point, -17 — “15° and its freezing point, -145° ± 2°, agree with the constants reported for the colorless product (6) prepared from -4- ozone and nitrosyl chloride. Rectification of the product at at­ mospheric pressure (using a packed colump ^6 inches long) did not increase its purity. On storage, even at -80®, nitryl chloride darkens in color. It is probable that its decomposition is also accelerated by light.

b. Review of the Reaction of Nitryl Chloride with Organic

Reagents

Zuskine (ll) claims to have prepared nitryl chloride

(b.p. 4*'6°C,) from phosphorus oxychloride and nitric acid. Reac­ tion of his "nitryl chloride" with phenyl magnesium bromide yielded biphenyl and chlorobenzene; with ethyl magnesium bromide no was obtained. Steinkopf and Kuhnel (13),using nitryl chloride prepared from chlorosulfonlc acid and nitric acid (9), report the following addition reactions of nitryl chloride. Ethyl- one eund nitryl chloride yield ethylene dichloride; cyclohexene yields l-chloro-2-nitrocyolohexane; styrene in benzene yields l-chloro-2-nitro-2-phenylethane; styrene in ether yields styrene pseudonitrosite, Stilbene in benzene yields i-ohloro-l-phenyl-2- phenylothane; stilbene in ether yields stilbene dichloride. Cin­ namic acid gives 2-chloro-3“Uitro-3“?henylpropanoic acid; phenyl- at —8o°.yields 1—chioro-2-nitro—2-phenylethylene. Ke- tene in ether at -8o° produces chloroacetyl chloride and some ni- troacetyl chloride. Methyl, ethyl, and phenyl magnesium bromides give only chlorinated products, Diazomethane yields chloronitro- “5“ methane; diazoacetic ester yields chloronltroacetic ester and

chloronltromethane* Only Isocyanide dichloridos are obtained

from methyl isocyanide and phenyl isocyanide. Benzene yields the

adduct l-chloro-2-nitro-3*4~cyclohexadiene; decomposition of

this addiict yields nitrobenzene. Sodium nitromethane in carbon

disulfide exploded violently with nitryl chloride.

The substitution reactions of nitryl chloride have not been

studied extensively. Phenol, at low temperatures, yields 2-ni-

trophenol, or, at room temperature 2,4~dichloro-6-nitrophenol, A

mixture of 1-nitronaphthalene and 1-chloronaphthalone is obtained

from naphthalene, Anisole and nitryl chloride form a Itl mole­

cular compound from which the ether can be regenerated (13).

Recently Brintzinger and Pfannstiel (14) reported that nitryl

chloride reacts with methyl acrylate and with acrylonitrile to yield methyl_ 2-chloro-3~nitropropanoate and 2-chloro-3“nitropropanni-

trile, respectively; no evidence except analyses was presented

to substantiate the structures assigned to these products.

Evaluation of the various addition reactions reported for nitryl chloride leads to no consistent pattern of either hemoly­

tic or heterolytic fission of the acid chloride in nonpolar sol­ vents, It appears likely that both ionic and free radical type

addition reactions may occur with nitryl chloride, the type de­ pending upon the unsaturated compound selected. Similarly, since both nitrated and chlorinated products are produced, no definite —6— conclusion can be dratm concerning the mode of attack of nitryl chloride on aromatic systems*

c. Reaction of sodium 2-Fropanenitronate with Nitryl

Chloride

2-Nitropropane was chosen as a typical model for the study of reactions of secondary nitroalkanes with nitryl chloride; moreover the desired nitration product, 2,2-dinitropropane, can be easily handled since it is a stable solid (m.p. 54°). In benzene solution, 2-nitropropano does not react with nitryl chloride; no hydrogen chloride is evolved from the mixture, a trace of nitro­ benzene is obtained but 2-nitropropano is almost completely recov­ ered, Sodium 2-nitropropane, however, readily reacts with nitryl chloride; the anhydrous salt, when dusted into ligtuid nitryl chlor­ ide at 80°, reacts violently— sometimes explosively— to give 2,3“ dimethyl-2,3“dinitrobutane in low yield. Smoother reaction is ob­ tained when a diluent (benzene, ethyl bromide, pentane, ) is present at temperatures ranging from -80 to 0° and either the nitryl chloride is distilled into a slurry of sodium 2-nitropro­ pane and solvent, or the dry salt of 2-nitropropane is dusted into a solution of nitryl chloride in the diluent. The inorganic pro­ ducts from the reaction include , sodium nitrite and sodium nitrate. The main organic products are 2 ,2-dinitropropane

(3-115^) » 2-chloro-2-nitropropane (up to 26%), and 2,3-d.imethyl-2,3- dinitrobutano (6%); in many reactions some 2-nitroso-2-nitropropane is formed, much of the sodium 2-nitropropane is also degraded to -7-

acetone» Since 2,2-dinitropropane was the product most desired, no attempt was made to investigate the effect of solvent on rela­

tive yields of the three products. No conclusions can be drawn

concerning the effect of temperature, solvent, order of addition,

or mode of addition of reagents on the yield of 2,2-dinitropropane; however, the yields of 2,2-dinitropropane were, in general, lower

than yields of 2-chloro-2-nitropropane and 2,3-dimethyl?2,^-dini-

trobutane.

A major difficulty in conducting the reaction of nitryl chlor­

ide with sodium 2-nitropropane in nonpolar solvents is the hetero­

geneous nature of the mixture. However, sodium 2-nitropropahe is

somewhat soluble in absolute ethanol and it is possible to obviate the difficulty by distilling nitryl chloride into a dilute solu­ tion of sodium 2-nitropropano in absolute ethanol at -8o°. Little

change in the nature of the reaction occurs in this solvent; the major organic products are 2,2-dinitropropane (6?G), and 2,3”dimethyl-

2,3-dinitrobutane (20#).

It is apparent from the results of these experiments, that nitryl chloride functions as a nitrating agent, a chlorinating agent, and an oxidizing agent. Since it is unlikely that nitryl

chloride, in a nonpolar solvent, cleaves simultaneously into a ni­

tronium ion (NO^), a chloride ion (Cl), a chloronium ion (Cl), and

a nitrite ion (NO^) (Equations 1 and 4) and since interpretation

of those data from Steinkopf and Kuhnel suggests that nitryl chlor- -8-

Ide exhibits reactivity expected of a mixture of nitrogen dioxide and chlorine, an alternate series of reactions (Equations $-11) is postulated to account for the results obtained with salts of secondary nitroalkanes* It is believed that these reactions of nitryl chloride are essentially oxidation-reduction processes

(reactions in which one electron is transferred) leading to inter­ mediate radicals and chains which either complement or eliminate the ionic bimolecular processes (Equations 1 and 4)*

4 . ' (CH^igCClNOg + HO"

5* (CHj)2C=N02 + NO2CI ---- » (CH^)2C-N02 + 'NO2 + cT

6. (CB^igCNOg + NO2C1 ------, (CH^)2CC1N02 + 'BO2

7. (05^)2^02 ' (CH^)2^(1102)2 + 'Cl

8»,b.(CH^)2C=a02 + *01 (or -NO^) (CH^i^CNO^ + Cf (or N^)

9. (CH^)^CNO^ + .NO2 ► (02^)20(^02)2

10. (OH ) CHO ► (CH ) C=0 + .NO 3 2 2 3 2

11. (CH^)20=N02 + N2Û^(or NgO^) ► (CH^)20(N0)(NO2) + NO" (or NOp

An alternative to the reaction represented by Equation 9 is a reaction which produces nitrite ions and chlorine atoms; however it is less likely to occur because of the greater stability of ni­ trogen dioxide (an "odd molecule" of long half-life) over that of the chlorine atom. For the sane reason it is predicted that Equa­ tion 7 Is less probable than Equation 6 and that Equation 8a is of greater significance than 8b. The formation of nitrogen dioxide and of chlorine atoms either by Equations 6 or 7 mny promote -9- decompositlon of nitryl chloride into chlorine and oxides of nitro­ gen (8), those products may then react further with sodium 2-nitro­ propane» Reaction of sodium nitrite and nitryl chloride does not yield sodium chloride; this eliminates that reaction as a possible route for the decomposition of nltryl chloride»

It has been reported that 2-nitroso-2-nitropropano is the ini­ tial product (15) from the reaction of nitrogen tetroxide and sodium

2-nitropropano (Equation ll); the pseudonitrole is also oxidized by nitrogen tetroxide (16) to 2,2-dinitropropane and 2,3-diraethyl-

2,3**dinitrobutane» The coupled product may arise either by the

Hass-Seiglo reaction (Equation 12) (l?), of by dimerization of two nitropropyl radicals (Equation I3)» The stability of nitro- propyl radicals may be attributed to marked resonance interaction in the electronegative, unsaturated nitro group»

12» (CH^)2C=N0" + (CH^)2CC1N02 (CH^)2C(N02)C (N02) (CH^)2

13. 2(CH^)2C=N02 2(^2)261102 ----» (CH^)2C(N02)C(N02)(CH^)2

The coupling of secondary nitroalkanes in alkaline medium has been effected with various oxidants (these results are discussed in Sec­ tion 17); a principal product of this reaction is also acetone»

Both reactions may be readily interpreted by Equations 10 and I3»

Further evidence supporting the postulate that nitroalkyl radicals are involved in the oxidative coupling and degradation reactions may be derived from the facts that anodic oxidation (18) of an al­ kaline solution of 2-nitropropane yields 2,3-dimethyl-2,3~dinitrobu- -10-

tane, acetone, and nitric oxide (see Equation lO).

In summation, the reaction of nitryl chloride with secondary nitroalkane anions is interpreted primarily as an oxidation-reduc­ tion reaction in which the anion (a base, a reducing agent) under­

goes rapid oxidation at the negatively charged center (probably at the least hindered, electron-rich oxygen atoms in the hybrid) with

transfer of one electron to nitryl chloride (an acid, an oxidizing

agent) or any oxidizing agent derived therefrom, to produce the

2-nitro-2-propyl radical. Secondly, a consistent conception of

the reactivity of nitryl chloride in nonpolar solvents is based on its homolytic fission rather than on dissociation into the various

ions*

The results of these experiments do not preclude direct form­ ation of 2,2-dinitropropane and 2-chloro-2-nitropropane by reaction of the nitroalkane ion with nitryl chloride. It may be that the product produced depends only on whether collision occurs at the chloro or at the nitro portion of the nitryl chloride molecule*

Since these possibilities imply heterolytic bond fission of nitryl chloride, a series of experiments was conducted in which polariza­ tion of the nitryl chloride was attempted with aluminum chloride— an attempt to favor the formation of 2,2-dinitropropane (Equation

14).

14* {CE^)^C=^0~ + nSj + AlCl------(CH^) 20(1102)2 + AlCl- -11-

d» Reaction of Sodium 2-Ritropropane With Nitryl Chloride

and Aluminum Chloride

A qualitative study of the type and degree of polar­ ization of nitryl chloride with aluminum chloride was made. Nitryl chloride at its boiling point (-l6°) does not form a stable complex with aluminum chloride. No appreciable increase was observed in the weight of aluminum chloride after it had been exposed to nitryl chloride in an inert atmosphere; however, evidence was obtained for formation of a complex, since the aluminum chloride residue produced a deep red color in benzene, A mixture of aluminum chlor­ ide and nitryl chloride produces nitrobenzene from benzene. It has already been observed that reaction of nitryl chloride with benzene readily yields nitrobenzene (l]i).

The reaction of sodium 2-nitropropano, nitryl chloride and molar quantities or excess aluminum chloride results, depending on the temperature, in formation of 2-chloro-2-nitropropane or formation of degraded products. Hfhen the mole ratio of aluminum chloride to nitryl chloride is 4*1 at -80 to 40° (i.e., in reflux- ing propane) a bright yellow solid is formed. The solid, after removal of the propane, hydrolyzos on ice to 2,7-dimethyl-2 ,7- dinitrobutane in low yield. When the temperature of reaction ranges from -20 to 20° (i.e., in pentane) nitryl chloride refluxes and the mixture changes color from blue-green to red. The marked color change at the higher temperature is interpreted to mean that -12- decompositlon of nitryl chloride occurs as the reaction with sodium

2-nitropropane prooeeds* The main products from this reaction are

2-chloro-2-nitropropane (35/^) 2,3~dimothyl-2,3“dinitrohutane

(6#); 2,2-dinitropropane is produced in low yield*

From the experiments with aluminum chloride, it can be in­ ferred that polarization of nitryl chloride (Equation 14) was either aiegligible at the temperature of reaction or that under these conditions nitration does not occur by an ionic mechanism. Polar­ ization of nitryl chloride into Cl and Cl^AlNO" (Equation I5) fol­ lowed by reaction with sodium 2-nitropropane will lead to 2-chloro-

2-nitropropane.

15, NO^Cl + AlCl^ --- > ct + Cl^AlNO"

Formation of an aluminum 2-propanenitronate, followed by reaction with nitryl chloride would be expected to give similar results*

2* Mixed Nitric and Sulfuric Acids

a* Reaction of Sodium 2-Nitropropane with Mixed Nitric

and Sulfuric Acids

The nitronium ion (19), derived from mixed nitric acid and sulfuric acid (Equation 16), is the active reagent

16* EONOg + EgSO^ " ' H^o + N$g + 2HS0^ in the nitration (20) of aromatic compounds* Therefore an attempt was made to nitrate sodium 2-nitropropane with a mixture of nitric acid (100 per cent) and concentrated sulfuric acid (Equation l)*

Addition of small portions of anhydrous sodium 2-propanenitronate -13- to the nitrating mixture at -20 to -10° results in the evolution of gases. The gaseous product, vonted through aqueous barium hy­ droxide and a cold trap (-80®) supports combustion and is identi­ fied as nitrous oxide (probably formed by the Hef reaction (2l),

The acidic mixture from the reaction, poured on ice, neutralized with , and steam-distilled, yields 2,3-dinethyl-

2.3-dinitrobutane as the only water insoluble product; apparently the principal reaction of this nitrating agent with sodium 2-ni­ tropropane is production of acetone and nitrous oxide (the Nef reaction (21)).

The coupling of a pair of 2-propanenitronate anions to form

2.3-dinethyl-2,3-dinitrobutane must involve the loss of two elec­ trons; simultaneously the mixed acid functions as an oxidizing agent (Equations I7 and 18) in addition to decomposing sodium 2- nitropropane by the Nef reaction (2l).

17. 2(CH3)2C=N0" ----- (Ch ^)2C(N02)C(N02)(CH^)2 + 2e

18. 2NO2 + 2e ----- 2NO2

It is of interest that Ben ford and Ingold (22) propose that in the nitration of aromatics with nitric acid in nitromethane, the effective nitrating agent is CH2=N0(0N02), formed by isomeriza­ tion of nitromethane, followed by condensation with nitric acid.

It is possible that a similar intermediate could be formed by coordination of 2-propanenitronate and nitronium ions (Equation

19), This adduct, isomeric with 2,2-dinitropropane might -14“

19. (083)20=^02 + 802 (OH,) 20 = / 080•'2' rearrange to yield 2,2-dinitropropano, since such a conversion, of the adduct to 2,2-dinitropropaneS

OH 0 OH, 80^ \ J ^ V '

CHj ''oKOj “ 3 is analogous to the conversion of aci-2-nitropropane to 2-nitro- propane

OH, 0 OH, H \ / \ / / ' / ° \ OH, OH OH 80 3 3 2 however this change did not occur (perhaps because of the rapidity of the 8ef reaction). 2,2-^initropropane is preparodcommercially by reaction of 2-nitropropane with nitric acid at elevated temper­ atures and pressures (23). From the results of the present inves­ tigation, it may be inferred that the nitration does not occur by an ionic mechanism involving aci-2-nitropropane and nitronium ions,

3. Methyl 8itrate

a. Reaction of Sodium 2-8itropropane and Potassium Nitro-

ethane with Methyl 8itrate

i. Alkyl 8itrates as 8itrating Agents

Alkyl nitrates in basic medium are efficient nitrating agents for compounds containing active methylene groups

(24). The general reaction may be represented by Equations 20 and -15"

21, in which G' and 0” represent acidifying groups,

20, G'GwCHg + oi — » G*G"di * EOH

21, G'G"CE + NO^OB <— > G'GwCHNOg + OR -- , G'G"C=N02

Compounds which have teen nitrated by this reagent in the presence of alkoxides are phenylacetonitrile (25), malononitrile, aceto- acetic ester (26), fluorene (27)* (28) and diben­ zyl ketone (28). In general these nitrations are rapid, exother­ mic reactions which occur below room temperature.

Primary nitroalkanes and secondary nitroalkanos, because of the acidifying effect of the nitro group, have active hydrogen atoms. Since the failure to nitrate sodium 2-nitropropane with mixed acid could be attributed to the Nef reaction— a reaction which takes place only in acidic medium— nitration of sodium 2- nitropropane was attempted in basic medium with methyl nitrate as the nitrating agent. This constitutes a further test of the postulate that a nitroalkane ion does not accept a positive nitro group by a bimolecular reaction at the alpha carbon atom (Equation

1).

ii. Discussion of Results

In contrast to other compounds containing active hydrogen, 2-nitropropane was not nitrated by methyl nitrate in methanolic sodium methoxide. Reaction of equimolecular amounts of methyl nitrate and sodium 2-nitropropane in methanol is very slow at 0-20°, After I6 hours in refluxing methanol, 59 per cent —16— of the methyl nitrate is converted to sodium nitrate. The products isolated are acetone oxime (l4?^) * 2-methyl-2~nitro-l-propanol (20^), and 2,3“dimethyl-2 ,3~dinitrobutano (9%); no 2 ,2-dinitropropane could be detected.

The results of these experiments may be interpreted by assum­ ing alkylation of the nitroalkane ion by methyl nitrate (the methyl ester of nitric acid) in the same manner that nitroalkane salts are alkylated by alkyl esters of other inorganic acids, e.g., methyl iodide (3 »5t^)i dimethyl sulfate (29,30)* Products alkylated on oxygen (i.e. nitronic esters, HB'CsNO^H) resulting from these reac­ tions, are usually thermally instable and decompose to yield alde­ hyde or ketone (derived from the alkylating agent) and an oxime

(derived from the nitroalkane) (30). The following reactions may account for the formation of acetone oxime and 2-methyl-2-nitro-l- propanol: OCH 22. (CH^)2C=N02 + CE^ONOg » (CH^)2C=N^

OCE 23* (0Hj )2C=N^ — — * (CE^^gCsNOE + CE^O

yCE„0” 24. (CB ) C=NO: + CE_0 ^ (CH,)2CT ^ ^ NO2

The low yield of acetone oxime (l4î^) and 2-methyl-2-nltro-l-pro- panol (20?^) may be attributed to the inefficient extraction of oxime from the solution, and to the ease with which the nitro al­ cohol is cleaved into 2-nitropropane and (the ease -17- of reversal of such aldol-type condensations was demonstrated ty acidifying an alkaline solution of 2-methyl~2-nitro-l-propanol and

sodium nitrite; 2-nitroso-2-nitropropane was immediately formed in 80-85?i yield).

Evidently methyl nitrate is a bifunctional reagent capable either of nitrating or of methylating various anions. Alkyl ni­ trates react with alcoholates to yield either ethers and nitrate ion (Equation 25) (3I1 32, 33^ or aldehydes and nitrite ion (Equa­ tion 26) depending on the structure of the alkyl nitrate,

25 , HO + H'CHgOHOg--- - ROCHgH' + NTJ^

26. BO + B'CEgOHOg----► R'CHO + BOH +

Hef (c3) reported that reaction of alcoholic potassium hydroxide and methyl nitrate yield ethyl methyl ether and potassium nitrate; no nitrite ion was formed (similar observations were also made in the present research; reaction of sodium methoxide and methyl nitrate in refluxing methanol yields dimethyl ether). Ag the com­ plexity of the-alkyl nitrate increases, the yield of nitrite ion

(Equation 26) increases (32, 33): benzyl nitrate reacts almost completely by the second scheme, (Formation of ether and nitrate ion may be considered to be an reaction, the formation of al­ dehyde and nitrite ions an E^ reaction.) Since methyl nitrate is not converted into sodium nitrite and formaldehyde by sodium methox­ ide, it is likely that 2-methyl-2-nitro-l-propanol arises from con­ densation of 0-nitropropane anion with formaldehyde (the latter —18— formed by decomposition of the intermediate nitronic ester (Equa­ tion 23).

The formation of 2,3~dimethyi-2,3”dinitrobutane may be ex­ plained by the oxidizing action of methyl nitrate. In addition to its alkylating action on sodium 2-nitropropane, methyl nitrate can conceivably nitrate the aci anion on oxygen to produce the hypothetical intermediate (Ch^)2C=N0(0N02)• It is believed that this intermediate does not rearrange to 2,2-dinitropropane but instead decomposes into nitrogen dioxide and 2-nitro-2-propyl radical. Pairing of the 2-nitro-2-propyl radicals yields 2,3-di- methyl-2,3-dinitrobutane. Nitrogen dioxide in basic medium is converted into nitrite and nitrate ions. Reaction of the inter­ mediate, (CH^^gCzNOfONOg), with 2-propanenitronato ion may also lead to 2,3-dimethyl-2,3-dinitrobutane and nitrite ion.

Since the reaction of methyl nitrate and sodium 2-nitropro­ pane is slow at ordinary temperatures, a mixture was allowed to stand 12 hours. Saturation of the solution with carbon dioxide yields sodium methyl carbonate, and then sodium 2-nitropropane,

On concentrating the filtrate and adding ether, a small quantity of salt was precipitated. Upon acidification of the salt, like sodium 2-nitropropane (34), it yields 2-nitroso-2-nitropropane,

The salt instantly reduces silver nitrate to metallic silver and yields 2,2-dinitropropane, In contrast, reaction of pure sodium

2-nitropropane with silver nitrate yields 2,3~dimethyl-2,3“dini- -19- trobutane. It is significant, in the reactions of this new salt with acids or with silver nitrate, that neither protons nbr silver ions become part of the final molecule, (These observations led to the discovery of a new reaction, an excellent method for pre­ paring gem dinitro compounds; this reaction is more fully discussed in Sections II and III). An analysis of this salt indicates that it is almost completely inorganic and contains approximately 15 per cent sodium 2-nitropropane. The structure of a component of the salt was tentatively designated as { C E ^ ) , where Z is a combination of nitrogen and oxygen of unknown composition and charge. Later research (Sections II and III) established that sodium nitroalkane and sodium nitrite in 1:1 ratio are oxidized with silver nitrate to the gem dinitro compound. Therefore, by inference, component Z is the nitrite ion. It is known that pri­ mary and secondary nitroalkanes in alkali undergo decomposition and yield nitrite ions (35i

Reaction of potassium nitroethnne and methyl nitrate was in­ vestigated to determine whether the nitroalkane necessarily has to have two al-pha-hydrogen atoms for nitration to occur (Equations

20 and 2l)« It is possible that preparation of 2 ,2-dinitropropane from sodium 2-nitropropane by this reaction may have failed because the 2-propanenitronate ion is too weak a base to displace methoxide ion from combination with the nitronium ion. Assuming that in the reaction of 2-propanenitronate ion and methyl nitrate the transition -20- stato leading to 2 ,2-dinitropropano is:

_ + _ (CH )2-C...H02...0CH ) ^ NO ^ 2

If the displacement is reversible it nay he necessary for it to he followed hy a faster reaction, that of salt formation, (Equa­ tion 2l) to effect nitration of the nitroalkane. The proposed nitration of nitroethane with methyl nitrate and potassium methox­ ide is represented hy Equation 27. — + 2 7 , CR^CHgNOg + CH^ONO^ + CH^OK ---- CH^C(HÏÏ2 )2 K + 2CH^0H

Since potassium 1 ,1-dinitroethane, a yellow salt, is the only component in this system which is insoluble in methanol, its form­ ation should he readily followed. When this experiment was carried out, only small amounts of and potassium nitrate separated during 24 hours at 20-65°; it was concluded that potas­ sium 1 ,1-dinitroethane does not form. This experiment further indicates that the nitroalkane ion does not react with a positive nitro group at the carhanion center, possibly because of the elec­ tronegativity of the nitro group, i.e. nitration of a nitroalkane does not occur by the ionic mechanism represented by Equation 1. -21-

II. Experimental

A, Agents Investigated

1. Nitryl Chloride

a. Preparation of **itryl Chloride (9)

i. Intermediates

Chiorosulfonic Apia* Chlorosulfonic acid was purified from the technical product (Eastman Kodak Compeuoy) hy distillation (3?) at atmospheric pressure in a glass apparatus protected from the atmosphere by à drying tube containing anhy­ drous calcium chloride. The fraction distilling at 149-151^* (?45 nm.) was collected and stored in a glass-stoppered bottle.

Nitric AçiôL. Anhydrous: One hundred per cent nitric acid, prepared by distillation of a mixture of concentrated nitric acid and excess fuming sulfuric acid in a glass apparatus at atmospheric or reduced pressure was used efficiently; however, it was more convenient and less time consuming to use fuming nitric acid to which had been added the theoretical amount of fuming sulfuric acid necessary to combine with the water present.

ii. Procedure

The operations were performed in a hood.

All connections in the apparatus were of ground glass; openings in the system were fitted with drying tubes of anhydrous calcium chloride.

*All temperature readings recorded in the text are in degrees Cen­ tigrade and are uncorrected values. -22-

Ftuning sulfuric acid (123 g.$ 3®^ sulfur trioxide) was added dropwise to fuming nitric acid (lOO g., 1.47 “oles, sp, gr, I.50,

91*6^1 acid content), cooled to 0® in a 3OO ml. throe—necked, round bottom flask equipped with a dropping funnel (Note l), a motor driven glass stirrer (Note 2) sealed with sulfuric acid, and a

Dry Ice cooled distilling head (Note 3) to ^riiioh a test-tube re­ ceiver (123-150 ml. capacity) was connected (Note 4)« test- tube receiver (which also serves as a storage vessel) was immersed in a freezing mixture (-80®). The mixed acids were stirred vigor­ ously at 0® while chlorosulfonic acid (170,5 g.» 97*5 “!•* 1.47 moles) was added slowly from the dropping funnel over a three to four hour period, or at such a rate that brown "nitrous fumes" did not appear above the reaction mixture. With the addition of each drop of chlorosulfonic acid, almost colorless gaseous nitryl chlor­ ide (Note 6) was evolved. After the chlorosulfonic acid had been introduced, the cooling bath was removed and the mixture was stirred for an additional one-half hour.

The product was a dense, pale-yellow liquid freezing at -145 i 2° and boiling from -I7 to -15° (Note 7). Yield: 95-IO8 g,

(85-90%). Anal.* Calcd. for NO^Cli Acidity in water, 24,55 “ • eq, gm, percent. Found: 24,72. Calcd. for NO^Cl: Cl, 43»^

*The analytical data for nitryl chloride was supplied by Castarina and Tomlinson who checked this procedure at the Chemical Research

Section, Pioatinny Arsenal, Dover, New Jersey, “23-

Foundl 44*9 * On the basis of these results the compound has been identified as nitryl chloride of 98-99 percent purity (98# based on chlorine content, 99?* on acidity).

Notes;

1. It has been suggested (9) that the chlorosulfonic acid be introduced below the surface of the mixed acids. In this procedure it vas dropped directly into the reaction mixture.

2, Since the heat of reaction is large and decomposition of ni­ tryl chloride is accelerated at higher temperatures (8), efficient agitation vas necessary to eliminate localized reaction. A glass stirrer vas used having a vide blade mounted on a swivel so that it could be readily inserted through the flask opening* The stirrer shaft vas sealed vith mercury (attacked slovly) or vith concentra-, ted sulfuric acid in which nitryl chloride is insoluble*

3* If a lov temperature distillation head is not available, an efficient reflux condenser through which cold water is circulated may be used as a dephlegmator, and the nitryl chloride is condensed directly into the receiver cooled vith Dry Ice,

4 , The receiver is not opened to the system until the nitric and sulfuric acids have been mixed; otherwise sulfur trioxide fumes pass into it during the operation,

5 * The temperature range of previously described (9) is less easy to control,

6* In the original procedure (9) the nitryl chloride is washed -24—

In a sulfurlo aold toner to remove entrained acids and oxides of

nitrogen. This nas found to te unnecessary; in one experiment

an inferior product was obtained (deep red, nide boiling range)

nhen this technique nas employed.

7 * The yollon color of the product may be attributed to the

presence of traces of chlorine and nitrogen dioxide, although the

freezing point curve emid analysis indicates almost a complete ab­

sence of impurities. In a simple distillation from one test tube

into another, nith a calibrated alcohol thermometer (nhose bulb

nas nrapped nith glass nool to promote ebullition) immersed in

the liquid, less than tno per cent distilled belon -l6°. Pure

nitryl chloride is a colorless liquid, f.p. -145°, b.p. -1$°. (8)

b. Reaction of Nitryl Chloride nith a Secondary Nitro­

alkane

la Reaction of nitryl chloride nith' 2-nitropropane

Nitryl chloride (35 g«, 0»43 mole) nas dis­

tilled in three hours into a mixture of 2-nitropropane (lO g,,

0.112 mole) in anhydrous benzene (lOO ml,) at 20-25°. The nitryl

chloride was completely absorbed and imparted a yallon color to r,' the solution; no hydrogen chloride was evolved (a gas scrubber,

attached to a Dry Ice trap directly connected to the reaction flask,

remained neutral and gave a negative test for chloride ion at the

end of three hours). The mixture nas flushed nith air to sweep

out unreacted nitryl chloride and then distilled. Denzene (99 ml.) -25- and 2—nitropropano (a total of 9,1 g,, distilling between 9^ and

125°) were recovered. ^ small residue of nitrobenzene (13) remained.

The recovered benzene contained nitryl chloride and its decomposition products.

ii. Reaction of Nitryl Chloride with Sodium

2-Ni tropropane

Sodium 2-nitropropane (38) was prepared by adding a slight excess of 2-nitropropane in ether to a 10 per cent

solution of sodium ethylato in ethanol. Tho product was cooled, filtered, washed with ether and dried in vacuum over anhydrous calcium chloride (Yield; 80-85%).

1. Addition of sodium 2-nitropropane to nitryl chloride at

-80° resulted in vigorous reaction ( hat became explosive if the addition was rapid) that evolved gases. After excess nitryl chlor­ ide had evaporated and the solid residue had been washed with water, a small quantity of 2,3“dimethyl-2,3-dinitrobutane remained. The white solid, recrystallized from methanol, melted at 211°; lit.

210-211° (17•39*40)• Addition of an authentic sample did not de­ press the melting point of the product.

2. Nitryl chloride (35 g., O.4 mole) was distilled into a stirred slurry of sodium 2-nitropropane (28.0 g,, 0.25 ®ole) in benzene (lOO ml.) cooled in an ice bath. The liquid phase became deep green. At the end of an hour, the mixture was filtered from the residue of sodium chloride, sodium nitrite and sodium nitrate. —26— and tbondistillad. When the volume had been reduced to approximate­ ly 25 ml,, 2 ,3“dimethyl-2 ,3''dinitro butane, began to crystallise

(1.4 g., 6 .4*), m.p. 205-210°, 210-211° (from methanol). The fil­ trate and mother liquors from the solid product was distilled.

2 .2-Dinitropropane (l,2 g,, 3 *5^) contaminated with 2-chloro-2- nitropropano distilled at YY-8o°/l5 mm., at 90° (ca. 2 mm.),

2.3-dimethyl—2,3-dinitrobutane sublimed to the cooler parts of the apparatus* The crude 2,2-dinitropropane was sublimed at 20 mm.

(®*P» 47-49 ), and then was recrystallized from petroleum ether, m.p, 52-53°, 2 ,2-Dinitropropane is a white waxy solid, m.p. 54°i b,p, 185.50(41).

3» Hitryl chloride (15 g,, 0.2 mole) was distilled into a stirred suspension of sodium 2-nitropropane (7*5 g», O.067 mole) in dry ethyl bromide (lOO ml.) at —10°, After one hour the mixture was filtered, and the deep green filtrate was distilled to remove ethyl bromide, excess nitryl chloride and oxides of nitrogen. The higher boiling residue, a viscous oil, was distilled through a short column to yield: (l) 1,0 g,, b.p. 48-130°; (2) 2.0 g., b.p, 130-140°; (3) 1,5 g»t b,p, 140-180°; and (4) & fraction which was distilling at l80-l85° when a sudden decomposition of the mater­ ial in the still occurred and resulted in the loss of most of this fraction.

Fractions (l) and (2) were combined and distilled to yield

2-chloro-2-nitropropane (2,1 g,, 26*), b.p, 134-136°; lit, 133-134°

(4Z). -27-

Fractions (3) and (4) were combined and distilled at 175“

185°. The distillate, washed with alkali and chilled, yielded

2 ,2-dinitropropane (l.O g,, 11^), m.p, 44“4 &°. It was recry­ stallized from methanol and sublimed in vacuum, m,p, 53-53»5°»

4 * Propane (5OO ml.) was condensed into a flask equipped with a reflux condenser cooled with Dry Ice, a stirrer and a gas inlet tube. Sodium 2-nitropropane (28,0 g,, 0,25 mole) was added, and the slurry was cooled to -80®, Nitryl chloride (@2,0 g,, 0,27 mole) was distilled into the stirred mixture in refluxing propane

(”40°)• Solvent end excess nitryl chloride were evaporated. The solid residue, upon steam distillation, yielded a blue oil. Dis­ tillation of the blue oil gave 2-chloro-2-nitropropane (5*5 g«* l8?i), b,p, 134-135°» and 2,2-dinitropropane, an oil (2.0 g,, 6?&) b.p, 65“75°/l2 mm., which solidified to a white, waxy solid, m.p,

44-46°, 2 ,2-Dinitropropane was purified by recrystallization from petroleum ether, m,p, 52-53°*

5, % e n a homogeneous solution of sodium 2-nitropropane in ethanol was treated with nitryl chloride at -80°, using a procedure identical with those described for this reaction in benzene, ethyl bromide and propane, the yields of 2,2-dinitropropane and 2,3- dimethyl-2,3-dinitrobutane were 6 and 20 per cent respectively.

c, Boaction of Sodium 2-Nitropropane, Nitryl Chloride

and Aluminum Chloride

Sodium 2-nitropropane (22,0 g , , 0,2 mole) was added in two hours to a cold (—80 ), stirred mixture of pentane -28-

(300 ni»), anhydrous aluminum chloride (33 g«, 0,25 mole) and nitryl chloride (16 g,, 0,2 mole; b.p, -I7 to @15°), After one additional hour the cooling bath vas removed. As the reaction mixture warmed, the liquid phase changed from green to red, and a red liquid rofluxed in the condenser cooled with Dry Ice, After refluxing ceased the mixture was poured onto ice (iQOO g,) and concentrated hydrochloric acid (80 ml,), then steam distilled. An ether extract of the distillate was distilled to yield 2-chloro-

2-nitropropano (8,5 g., 35^) b.p. 131-136°. From the residue, which was recrystallized from methanol, 2,3"dimethyl-2,3-dinitro­ butane (1,0 g,, 6^) was obtained, m.p, 210-211°,

When propane (b.p, —40° was the solvent in this reaction a bright yellow solid was formed. After propane had evaporated the yellow solid decomposed spontaneously to a brown tar and brown fumes were evolved. In one experiment tho solid was successfully hydrol­ yzed by adding it to ice. Steam distillation of the hydrolysate yielded a very small amount of 2,3-&imethyl-2,3-diaitrobutane,

2. Reaction of Sodium 2-Nitropropane with Mixed Nitric and

Sulfuric Acids.

Sodium 2-nitropropane (28,0 g,, 0,25 mole) was slowly dusted into a stirred mixture of nitric acid (l6,0 g., lOOji, 0,25 mole) and concentrated sulfuric acid (25O ml.) at -10 to 15°, Dur­ ing the addition a colorless gas was evolved which was vented through a Dry Ice trap, through aqueous barium hydroxide, and then -29“ collected over water* The gas, possibly nitrous oxide, supported dombustion; however, it was not further investigated. The acidic mixture was poured onto ice, neutralized with sodium carbonate and steam distilled, 2 ,3“Dimethyl42,3“dinitrobutano, m.p. 210-211°

(from methanol), in very low yield, co-distilled with steam; 2 ,2- dinitropropane, a more volatile substance, was not found*

3» Reaction of tl^thyl Nitrate with Sodium 2-Ni tr opropane

a. 2-Nitropropane (44»5 £•» 0.50 mole) was added at tempera­ tures below 20° to a stirred solution of sodium (l2.0 g., 0.F2 gram atom) in absolute methanol (3OO ml.). After ten minutes methyl nitrate (43) (3^.5 g . , O.5O mole) was added, and the mixture was heated to its boiling point. A vrhi'te voluminous, pasty solid soon filled the flask; after being refluxed for sixteen hours it had changed to a dense white crystalline solid. The product was collected, washed with methanol and ether, dried, and identified as sodium nitrate containing traces of sodium nitrite (25.0 g., corresponding to 59^ of the methyl nitrate). The red filtrate was saturated with carbon dioxide until carbonates no longer pre­ cipitated, filtered, acidified at 0° with dilute hydrochloric acid, and concentrated by distillation to approximately y^-lOO ml. On diluting the concentrate with water (5OO ml.), crude 2 ,3“dimethyl-

2,3“&initrobutane (4.1 g . , 9 »3^) separated; it was recrystallized from methanol, m.p. 209“210°. The remaining aqueous filtrate was extracted with an ether-benzene mixture (8 *2 , 9 % 100 ml.), and -30-

tbe dried extract distilled* After solvents had been removed the

products were) (l) Acetone oxime (3.O g,, 1^% based on the methyl nitrate consumed), b.p, 6l-63°/35 mm*, m*p* 59»5“^0°; mixed melt­

ing point with authentic acetone oxime, not depressed* (2) 2-

Methyl-2-nitro-l-propanol (7»2 g., crude, 20^ based on the methyl nitrate consumed), b*p* yO-SO^/^-? mm,, which crystallized in the

receiver as fine needles contaminated with a yellow liquid, m.p*

70-80°* A portion was recrystallized twice from petroleum ether- benzene and dried at 2'ÿ mm* for two hours; m*p* and mixed m*p* with

an authentic specimen 89-90° (44) • !I^he crude 2-methyl-2-nitro-l- proponol was characterized further by conversion into 2-nitro-2-

nitrosopropane; an aqueous solution of the nitroalcohol (O.^O g*

in 4 ml*) was mixed with dilute alkali (slightly in excess of the

molar equivalent), sodium nitrite (0*3-0*4 g*) and crushed ice; the mixture was immersed in an ice bath and acidified by adding as

rapidly as possible 1:1 hydrochloric acid (lO ml*). The pseudoni-

trole (0*40 g., 80^) was obtained as the white dimer, o*p* 74~75°

(blue melt, dec*)* An authentic specimen of 2-nitro-l-propanol under the same conditions yielded 2-nitroso-2-nitropropane (0,43

g*» 86^), m*p* 74-75^ (blue melt, dec*) (45)! the melting points were unchanged on mixing the samples, (3) A yellow viscous liquid

(2*6 g#), b*p* 80-92^/2 mm*, insoluble in water, dilutealkali and

acid, which was not further investigated* (4) Residue, 1.6 g*

In a similar experiment, the solid which separated during the -31- first two hours at the boiling point of the mixture or after twelve hours at 25-30° w^s identified principally as sodium 2-nitropropane

(l4«7 g«, 26,5%). The salt burned explosively, gave a blood red

Color with ferric chloride (46) and yielded 2-nitrosor-2-nxtropropane

on acidification (34»45)« Distillation of the mixture resulting from reaction of the solid with aqueous silver nitrate produced

2 ,3“dimothyl-2 ,3~

When methyl nitrate was refluxed with an equivalent of sodium methylate in methanol for two hours, sodium nitrate (quantitative ) and dimethyl other were produced. Dimethyl ether was identified by its boiling point, -22 to -l8°; lit. -23.7° (47).

I. The Isolation of the Mixed Salt of Sodium 2-Nitropropane and Sodium Nitrite. The Silver Ion Oxidation of the Mixed Salt to 2,2-Dinitropropane.

A solution of 2-nitropropane (22,6 g., O.25 mole) methyl nitrate (19.3 g., O.25 mole) and sodium methylate (14.O g., 0.26 mole) in methanol (3OO ml.) was allowed to stand at 0° for twelve hours, then saturated with carbon dioxide until sodium methyl car­ bonate no longer precipitated. The filtrate was concentrated by atmospheric distillation until it became turbid. The distillation until it became turbid. The distillation residue was cooled and diluted with ether to precipitate a white powdery salt (6,3 g,).

Anal. Calcd. for the double salt of sodium 2-nitropropane and

sodium nitrite, C^EgO^NgNag* C, 20 ,00; H, 3.36; N, I5.56; -32-

Na, 25.54. Found* C, 2,98; H, 0 ,4 9 ; N, 13.3 3 : Na, 29.18. The ratio of per sent carbon found to per cent carbon calculated is

0 ,149; the ratio of the percentages of hydrogen, 0,148, Calcd, from the carbon-carbon ratio* H, (l4 ,9)(3»3^^~®*51. Found* H,

0,49, %ie >fe.jor constituent of the salt product is probably sod­ ium nitrate

Acidification of an aqueous solution of this salt yielded

2-nitroso-2-nitropropane, The salt product reduced silver nitrate in aqueous medium and was converted into 2,2-dinitropropane in low yield, (See Experimental Section II and for details of this reaction),

b. Attempted Preparation of Potassium 1,1-Dinitroethane (48)

A mixture of nitroethane (7,5 g,, 0,10 mole) and methyl nitrate (7.? g« « 0,10 mole) was added to a solution of potassium

(4,0 g,, 0,10 gram atom) in methanol (lOO ml,) under nitrogen.

During four hours at 25-30° the clear solution became yellow.

After the mixture had been refluxed for one hour, a small quantity of white crystalline solid separated. The mixture was left for sixteen hours at 25-30°, then reheated to boiling, cooled and filtered. The white product was identified as a mixture of po­ tassium nitrate and potassium nitrite. It was concluded that potassium 1,1-dinitroethane, a yellow salt insoluble in methanol, had not formed. “ 33“

III. SUMMABY

The reaction of sodium 2-nitropropane with nitryl chloride yields a mixture of products. 2,2-Dinitropropane is produced in low yields (^-11%);2-chloro-2-nitropropane and 2,3”dinethyl-2,3~ dinitrobutane are also formed (up to 26^),

An attempt to nitrate sodium 2-nitropropane to give 2,2-di­ nitropropane, by use of mixed nitric and sulfuric acids, was un­ successful. Sodium 2-nitropropane is almost completely decomposed by the mixture of acids; 2,3~dimothyl-2,3~dinitrobutane is formed in traces.

Methyl nitrate apparently does not convert sodium 2-nitropro­ pane into 2,2-dinitropropane. The products of this reaction are acetone oxime, 2-metbyl-2-nitro-l-propanol and 2,3-dimethyl-2,3-dini­ trobutane. Similarly, potassium 1,1-dinitroethane does not result when potassium nitroethane is treated with methyl nitrate.

In the course of these investigations a new reaction of ali­ phatic nitro compounds was discovered. This portion of the sub­ ject is treated more fully in Section III. -34-

SECTION II

HEACTIOH OF NITRIC OXIDE WITH SODIUM NITROAI&ANES

I. DlBeusaion

Introduction

As a result of attempts to develop a method for converting mono- nitro compounds to gem dinitro derivatives (Section l) it was postu­ lated that the nitroalkane anion does not accept a positive nitro group. The difficulty experienced in introducing a second nitro group at the alpha carbon atom is at variance with the observed ease with which ions of primary and secondary nitroalkanes are nitrosated (Equa­ tion l) with nitrous acid (34t 43)t nitrosyl chloride (ig), and nitrogen dioxide (15) •

1. RR'CaNOg + E*HONO (or HOCl, or N2O4) — ^ RR'CCwOÏNOg +

HgO (or Cîp or HO^ )

Traube(49) has found that primary nitroalkanes react with nitric oxide slowly in the presence of one mole of sodium ethoxide, anA rapidly when two moles of the base are used, thus producing disodium l-nitrosohydroxylamino-l-nitroalkanes (Equation 2). (Traube named these salts isonitramino derivatives; for the structure of the ^2^2 group of. 50.)

2. RCH„HO„ + 2ND + 2C^qONa . R-C^ -35-

Upon acidification, these salts (Equation 3) yield the correspond­ ing nitrolic acid and nitrous oxide.

3. H C^ + 2H » EC* + W5O + ^ N O H

A number of stable metallic derivatives of l-nitrosohydroxylamino-l- nitroalkanas are known. The silver salt is unstable and deposits metallic silver; tho other products of its decomposition were not determined (49).

Since the preparation of nitrosohydroxylamlno derivatives of

secondary nitroalkanes has not bean reported, it was decided to investigate the reaction of nitric oxide with salts of secondary nitroalkanes. A salt (which forms in the reaction of methyl nitrate

on sodium 2-nitropropane in basic medium ) has been mentioned in

Section I. The salt yields 2-nitroso—2-nitropropane upon acidifica­ tion; if it is treated with silver nitrate the products are 2,2-di- nitropropane and metallic silver. Similarities between the salt in

question and disodium nitrosohydroxylaminonitroalkanes also served to promote interest in these reactions of nitric oxide.

B. Agents Investigated

Investigation of the reaction of nitric oxide with salts of

secondary nitroalkanes has led to the discovery of a new reaction for the preparation of gem dinitro compounds. Unlike the sodium

salts of primary nitroalkanes, salts of secondary nitroalkanes do not combine with nitric oxide to yield nitrosohydroxylamino deriva­ tives. The products from secondary nitroalkanes are a double salt -36- of sodium secondary nitroalkane and sodium nitrite, a ketone and Its oxime (toth derived from the nitroalkane), and the corresponding coupled product (a vicinal dinitro compound containing twice as many carton atoms as the original nitroalkane). Oxidation of the mixed salt with silver nitrate produces a gem dinitro derivative with the same number of carbon atoms as the original nitroalkane.

Reaction of three secondary nitro compounds, 2-nitropropane, 2- nltrobutane, and nltrocyclohexane, with nitric oxide is conducted as follows; The nitroalkane is dissolved in ethanollc sodium ethoxide containing 10 per cent to $0 per cent more base than is required for neutralization. Dilute alcoholic solutions of sodium 2-nltropropane or sodium 2-nltrobutane containing 10 per cent excess of sodium eth­ oxide are homogeneous. Sodium nltrocyclohexane is relatively insol­ uble in ethanol; the mixture was heterogeneous throughout the reaction.

The solutions are flushed with nitrogen, cooled to 0-10° and treated with nitric oxide (approximately 4 moles to 1 of nitro compound) for two to three hours. Precipitation of a salt begins soon after intro­ duction of nitric oxide. After the reaction is complete, the mixture is swept with nitrogen and then filtered to isolate the precipitate.

Yields of the salt are not affected by varying the concentration of nitroalkane nor by increasing the concentration of ethoxide ion

(lO to 100# excess). The yield of salt usually ranges from 45 to 50 per cent based on sodium nitroalkane (Equation 4)« the yield of salt does not exceed 50 per cent in any experiment. -37-

- + y'^2^2 4, RB'Cs NOjNb + 2H0 > RB'C

For example, BH*C = C^Sg (l); C^Eg (ll); CgE^^ (ill),

The empirical formula of the salt obtained from nitric oxide and sodium 2-nitrobutane (based on euaalyees) is C^E30^E2Ha2 ; whereas

the empirical formula for the corresponding nitrosohydroxylamino derivative (II, Equation 4) lo C^EgO^E^Ea. The salts prepared from

sodium 2-nltropropane and sodium nltrocyclohexane in heterogeneous alcoholic suspensions are much less pure ; however analyses indicate that the empirical formulas for these derivatives are also in the range ER'CO^EgEag, On the basis of analytical data (Table l) and empirical formulas, it is apparent that the salt products are not monosodium nitrosohydroxylaminonitroalkanes (I, II, III, Equation 4) but mixed salts (ill) of sodium secondary nitroalkane and sodium nitrite;

RE'C=N02Ea»EaE02

For example, EH'C = C^Eg (l'); C^Eg (ll'); and CgEj^Q (III'). -56-

Table I

Analyses of Nitric Oxide Derivatives of Sodium Salts of 2-Nitro-

propane (I) 2-Nitrobutane (II) and Nitroeyelohexane (III)

Prod( % 0 $ H $ N $ Na

I Oalcd. for CydgO^N^Na 21.06 5.55 24.56 15.44

I» , Calcd. for 20.00 5.56 15.56 25.54

Pound ( a) 16.6 2.65 14.9 26.6

II Oalcd. for 04HeP^NjNa 25.90 4.56 22.70 12.45

II' Calcd. for 24.74 4.15 14.45 25.70

Pound 25.51 4.21 15.44 25.92

Pound 25.55 4.17 15.72 24.15

III Calcd. for 06HloO4^)Na 54.10 4.74 19.90 10.60

III' Calcd. for ^6^10®4^2^*2 52.75 4.56 12.75 20.69 Found (b) 25.54 5.76 10.61 20.64

Pound 25.59 5.49 10.91 20.54

(a) 65»7?6 Purity, the remainder sodium nitrite.

(b) 78.1$ Purity. -39“

Sisodium nltrosobydroxylamlno salts of nitroethano and 1-nitro- propane (4 9 » Equation 2) wore prepsrod for conparison with the salts derived from nitric oxide and secondary nitroalkanes, The alkyl- nitrosohydroxylamino salts, when acidified (as reported by Iraube) yield nitrolic acids (43* 49» 31* 32, 33^ and large quantities of gases (impure nitrous oxide» 8o-liy# yield. Equation 3)* Acidifica­ tion of the salts derived from nitric oxide and sodium secondary nitro- alkanes yields pseudonitroles; however only very small amounts of gases are evolved. The behavior of the salts obtained from secondary nitre compounds is not the behavior expected of the nitrosohydroxylamino group (Equation 3 ) » instead the behavior is characteristic of salts of secondary nitroalkanes with nitrous acid (43)(Equation l).

The salts derived from secondary nitroalkanes and nitric oxide are oxidized with silver nitrate to the corresponding gen dinitro compounds; 2 »2-dinitropropane (70?^), 2 »2-dinitrobutane (79?"), and

1 ,1-dinitrocyclohexane (60#). In contrast, the salts derived from nitroethano and 1-nitropropane and nitric oxide are oxidized rapidly with silver nitrate to water soluble products.

The salts derived from nitric oxide and the three secondary nitro compounds (2-nitropropane, 2-nitrobutane, and nitrccyclohexane) are oxidized with acidic hydrogen peroxide, or with salts of per- sulfuric acid, to the vicinal dinitro compounds ( 2 ,3~dimethyl-2 »3“ dinitrobutane, 3 »4“Aimethyl-3 »4'*dinitrohexane, and l,l*-dinitro- bioyclchexyl). Beaction of the salt obtained from sodium 2-nitro- -40- propane and nitric oxide with mercuric nitrate produces 2 ,7-dimethyl-

2,3-dinitrohutane; the same salt yields acetone idien it is oxidized with sodium hypohromite, potassium permanganate, or ferrio chloride.

(These oxidation reactions also occur with sodium 2-nitropropane or a mixture of sodium 2-nitropropane and sodium nitrite. Of, Sections

III and IV,)

It is prohahle that the salts produced in the reaction of nitric oxide with sodium secondary nitroalkanes are double salts (ill) of sodium secondary nitroalkanes and sodium nitrite, since they exhibit chemical properties identical with those of an equimolar mixture of sodium secondary nltroalkane and sodium nitrite. Reaction of two equivalents of silver ion with a mixture of nltroalkane and nitrite ions ( 1 equivalent of each) yields gem dinitroalkanes in 70-92 per cent yield (Equation 3) •

5. RR'CaNOg + HÔ2 + 2Ag ---► RH»C(N02)2 + 2Ag

Since the yields of the double salt from the reaction of nitric oxide with sodium secondary nitro compounds are 45*50 per cent based on complete precipitation of sodium nltroalkane, it appears that the nitrite ion is produced from the nltroalkane ion. The maximum yield of mixed (ill) salt would be only 50 per cent by this process. On this basis, the yields, of double salt are 30-93 pe-? cent.

In addition to the mixed salt of sodium nitrocyolohexane and sodium nitrite (68^), the reaction of nitric oxide with the sodium salt of nitrocyolohexane yields oyclohexanone and cyclohexanone oxime —41—

(28%) and l,l*-dinitro'bicyolohexyl (3#). The following reactions are postulated to account for the formation of these products!

/ N 0 “ 6, HE*C=HO, + *N=0 i®'C'

• - /NO 7. HH»C^ RH'C=HO" + "NO, 'NOg

8a. HH»C=NO~ + CjH^OH HR'C=NOH + CgH^o"

8b. NOg + NO ï=£ NgOj

8c. CgE^O" + NgOg --- » CgE^ONO + NOg

By these processes both oxime and nitrite ion are produced from second­ ary nltroalkane ion. The exchange reaction of a nitroso group for a nitro group (Equations 6 and 7) has been observed previously (34)I esters of al-pha-nitrooarboxvlic acids have been converted into the nitroso derivatives by reaction with oxides of nitrogen (Equation lO).

10. aca(N02)C02C2Eg » RCH(N0)C02C2H5

The reverse reactions of those represented by Equations 6 and J are similar to those of nitrogen dioxide with aliphatic ketoximes (33, 36,

37) (Equations 11 and 12).

/NOH 11. HH»C=NOH + «NGj HB'C ^NOg -42-

y&OH /NO 12. Bs'c; .HO, --- » ER'C + HHO, NOg

The intermediate product postulated in Equation 11 is the conjugate acid of the intermediate proposed in Equation 6, No pseudonitroles are formed in the reaction of nitric oxide with salts of secondary nitro compounds; presumahly, therefore, the intermediate proposed in Equation 6 is not oxidized readily by nitric oxide. Since nitric oxide is a stable free radical, the intermediate produced in Equation

6 would be a negatively charged free radical. Reaction of this free radical with a second molecule of nitric oxide would be ejqpected to yield the nitrosohydroxylamino derivative. A possible explanation for the fact that primary nitroalkanes yield disodium 1-nitroso- hydroxylamino-1—nitroalkanes (Equation 2) whereas secondary nitroalk­ anes yield mixed salts of secondary nitroalkanes and sodium nitrite, may be that the intermediate formed by the combination of two nitric oxide molecules with sodium primary nitroalkane can be stabilized by reaction with a second molecule of base (Equation 13. cf. Equation

2).

/^2®2 - /®2®2 13. RCH + CgHrO ---- ► EC + CgHcOH \NOg ^^N02

The compound that would be produced (by Equation 4) from the secondary nitroalkane lacks an alpha hydrogen atom and thus may be unstable, decomposing into the conjugate base of the oxime and oxides of nitrogen (Equation 14) • -43“

/ ^2®2 - 14* RR*C *'■ ■ ■< RR'C=IfO + NnO?

Cyclohexanone may have been derived from the oxime, or in part from a process that yields the vicinal dinitro compound, l,l'-dinitro- bicyclohexyl (Equations 15 and l6),

15. HR'CsHo”------> HH'C-NOg ----» EE'C=0 + 'HO

coupling I

16. IIH'-C(N02)C(H02)EE'

In an Investigation of the effect of oxidizing agents on salts of secondary nitroalkanes (Section IT) it was found that the ketone and the corresponding vicinal dinitro compound are produced(Equations

15 and 16); it is possible that nitric oxide or any oxidizing agent derived from it during reaction effects "oxidative dimerization" (58) of sodium nitrocyolohexane.

From the reactions which have been postulated, it is to be expected that eventually all of the secondary nitroalkane should be converted into its carbonyl derivative and nitrite ion. This was demonstrated qualitatively in the reaction of sodium nitrocyolohexane with excess nitric oxide. The yield of mixed salt diminished when the reaction time was lengthened. Apparently this degradation of the sodium nitroalkane does not go to completion rapidly, because of the insolubility of the sodium nitrite-sodium nitroalkane precipitate. -44-

II. Expérimental

A. Ihe Agent

1. Nitric Oxide (Reaction of Sodium Nitroalkanes)

a. Preparation

The method employed was essentially that of Johnston and

Criauque (59) .

Sulfuric acid (^O?^) was added slowly to a solution of sodium nitrite (4 molar) and potassium iodide (l molar). The nitric oxide was passed through sulfuric acid (gO^O, (5Oλ) and a Dry Ice trap, before being used,

b* Reaction With Sodium 2-Nitropropane (Formation of a Mixed Salt (l:l) of Sodium 2-Nitropropane and Sodium Nitrite).

Nitric oxide (approx. 1 mole) was passed into a stirred

solution of 2-nitropropane (22.3 B»» 0.25 mole), absolute ethanol

(350 ml.) and sodium ethylate (0.3 mole) for three hours at 0-10°.

(Before nitric oxide was added, the reaction mixture was flushed with nitrogen.) After the introduction of nitric oxide was completed, the mixture was flushed with nitrogen. The insoluble product, a mix­ ture of sodium 2-nitropropane and sodium nitrite, was filtered, wash­ ed with alcohol and ether, and dried to constant weight in vacuum

over calcium chloride. Yield: 21,2 g. (94^", based on a 1:1 mixture

of sodium 2-nitropropane and sodium nitrite).

Anal. Calcd. for the double salt of sodium 2-nitropropane and sodium

nitrite, C^EgO^NgNsg: C, 20.00; H, 3.36; N, 15.56; Na, 25.54.

Found: C, 16.8; H, 2.65; N, 14.9 : Ea, 26,6. (Calcd, for the -45-

produots expected from the Irauhe reaction (49)> sodium 2-nitroso-

hydro%ylamino-2-nitropropane, CjHgO^HjNai C, 21 ,08; H, 5 *5 5 l N, 24*53;

Na, 13,44*) ^Ne main contaminant in the crude product iras sodium nitr­

ite, On the basis of analytical data, the product contains 83,7 psr

cent of the double salt and 16*3 per cent of sodium nitrite,

c. Reaction With Sodium 2-Nitrohutane (Formation of a Mixed Salt (ill) of Sodium 2-Nitrohutane and Sodium Nitrite,)

A solution of sodium 2-nitrohutane in absolute alcohol

TOS prepared by adding 2-nitrobutane (25,8 g,, 0,25 mole) to ethyl

alcohol (350 ml.) containing sodium ( 7,0 g,, 0,3 g* atom). The mixture

was cooled to 0°, stirred, flushed with nitrogen and then treated with

nitric oxide (approx, 1 mole) for 100 minutes. The precipitated salt

was collected, washed with alcohol and ether, and dried to constant

weight in vacuum. Yield: 22,0 g, (9I# based on a 1:1 mixture of

sodium 2-nitrobutane and sodium nitrite),

Anal. Calcd, for the double salt of sodium 2-nitrobutane and sodium

nitrite, G^Eg0^N2Na2: C, 24,74: H, 4 .15; N, 14,43: Na, 23,70. Found:

C, 25,31, 25,53; H, 4,21. 4,17: N, 13,44, 13,72; Na, 23,92, 24,13.

(Calcd, for the product expected from the Traube reaction (49) t sodium

2-nitrosohydroxylamino-2-nitrobutano, C4EgO^N^Na: C, 25,90: N, 4.36,

N, 22,70; Na, 12,42),

d. Reaction With Sodium Nitrocyolohexane (Formation of a Mixed Salt (ill) of Sodium Nitrocyolohexane and Sodium Nitrite, Cyclohexanone, Cyclohexanone Oxime and 1 ,1- Dinitro-1,1*-Dicyclohexyl)

1, Nitrocyolohexane (25,8 g,, 0,20 mole) in absolute

ethanol (50 ml.) was added to sodium (8,9 g,, 0,39 g, atom) dissolved -46- in ethanol (25O ml*) under nitrogen. The suspension of sodium nitrocyolohexane was stirred vigorously while nitric oxide (approx,

1 mole) was added in IO5 minutes’ at 0-20°, The mixture was treat­ ed as described for 2-nitropropane and 2-nitrohutane, A light tan powder (2^,0 g,) was obtained (theoretical yield based on the double salt of sodium nitrocyolohexane and sodium nitrite, 22,0 g,),

Anal, Calcd, for CgHioO^N2Na2: C, 32,73: 4 .58; N, 12*73 ;

Na, 20,89,

Found: C, 25*54. 25*59: H, 3,78, 3,49; N, 10*8l, 10,91;

Na, 20,84, 20*54; volatile solvents, 12*6, The ratio of per cent carbon found to the calculated value is 0,781; 78.1 per cent of the calculated value for hydrogen is 3*58; the exper­ imentally determined analytical values for hydrogen are 3.78, 3'49«

On this basis, the purity of the dried double salt is 78 per cent.

2 , Nitric oxide (approx, 1 mole) was passed into a mix­ ture (prepared under nitrogen) of nitrocyclohaxane (25.8 g,, 0,20 mole) and sodium ethylate (0,22 mole) in absolute ethanol (4OO ml*) for three hours at 0-10°* The mixture, after being flushed with nitrogen and filtered, yielded the double salt of sodium nitrocyclo- hexane and sodium nitrite (15 g,, 68%)*

The filtrate was saturated with carbon dioxide until precipi­ tation of carbonates no longer occurred, then filtered, and dis­ tilled at atmospheric pressure to one-third of its original volume.

The distillate, ethanol, was discarded* Distillation of the pro­ -47- duct was continued at 40° (40 mm). The high toiling residue was triturated with ether, filtered free of inorganic salts and recom­ bined with the distillate. Rectification of the product through a short column yielded the following:

(l) Cyclohexanone (3.5 2 «, 17»^^)» §4~57° (20 mm,), n^®

1,4487» (2) Cyclohexanone and cyclohexanone oxime (0,7 g,, 3 *1"

3 *6^) t,p, 57*97° (20 mm,), n^^ 1 ,4514» (3) Cyclohexanone oxime

(l»5 g«, 6 ,6^), t,p, 109-111° (20 mm). The oxime solidified in the condenser and receiver, and was recrystallized from petroleum other, (1,3 g») m.p* 89-89*5°* (4) l,l'-Dinitro-bieyclohexyl

(0,81 g,, 3*l^'i from the concentrated hot methanolic extract of the undistillatle residue); recrystallized from methanol-acetone, m,p, 214° (darkening) -220° (decomposition), (5) Residue (7 *4 /»)» undistillatle and uncrystallizatle, Oxines prepared from frac­ tions (1) and (2), m,p, 89-90°, did not depress the melting point of fraction (3) nor that of an authentic sample of cyclohexanone oxime, Fraction (4) did not depress the melting point of an au­ thentic sample of 1,1'-dinitro-ticyclohexyl, m,p, 214-220® (See

Section 17), Nitrocyolohexane was not detected in any of the fractions,

B, Chemical Proof of Constitution of the Salts Derived from

Sodium Secondary Nitroalkanes and Nitric Oxide,

1, Acidification

a. The Double Salt Conversion to Pseudonitrole -48-

i. An aqueous solution of the double salt of sodium 2-nitropropane and sodium nitrite (l.yO g., 0,0094 mole in 14 ml.) at 0° was gradually acidified with 6N acetic acid (lO ml,). After one-half hour the mixture was warmed to 25® and 3N hydrochloric acid (5 ml,) was added. After one hour, only 10 ml, of gas had been evolved (collected over water). The blue-white solid that was formed, 2-nitroso-2-nitropropane, was filtered, washed with water and air-dried. Yield: 0,90 g, (8o^), m.p,

75-76° (blue melt, dec,); lit, 75-76° (45). The melting point of the product was not depressed by an authentic sample of 2- nitroso-2-nHropropane,

If the solution was acidified at 25° the yield of pseudoni­ trole was lowered (0,64 g,, 5^/») and the volume of gas evolved was increased (30 ml,). Rapid acidification with 3^ hydrochloric acid gave similar results.

ii. Acidification of an aqueous solution of the salt derived from sodium 2-nitrobutane (0,92 g,, 0,0048 mole) evolved no gas and yielded 2-nitroso-2-nitrobutane (0,51 g., 8lÿ),

®»?* 52“54° (blue melt, dec,). The pseudonitrole was recrystal­ lized from chloroform, m.p, 57-58° ('blue melt, dec.); lit, 58°

(blue melt, dec.) (53)*

iii. Addition of acetic acid to an a^iueous solu­ tion of the double salt of sodium nitrocyolohexane and sodium ni­ trite (0,50 g., 0,0023 “ole in 10 ml.) resulted in the formation of

1-nitroso-l-nitrocyolohexane (O.23 g., 64%), m.p. 69-71° (blue melt, dec,); gas evolution was negligible (5 ml,). -49-

When dilute hydrochloric acid was added at 25°, the useudoni- trole was obtained in much lower yield. The pseudonitrole of cy- clohexane is reported (60) to melt at samples with lower melting points were recrystallized from a mixture of chloroform and petroleum ether.

The product derived from sodium nitrocyolohexane and nitric oxide analyzed for 12,o per cent of solvent (water-alcohol), It is likely, therefore, that the yield of pseudonitrole is 75 per cent,

6, Disodium 1-IIitrosohydroxylamino-l-nitroethane and

Disodium 1-Nitrosohydroxylamino-l-nitropropane

(Traube Products)

The disodium salts of 1-nitrosohydroxylamino-1- nitroethane and l-nitrosohydroxylamino-l-nitropropane were pre­ pared (49), in yields of ^8 and 85 per .cent respectively, from nitric oxide and the corresponding nitroalkane in ethanolic sodium ethylate (lOO# over the theoretical quantity required to form the sodium nitroalkane). The conditions employed were identical with those described for the reaction between nitric oxide and secon­ dary nitroalkanes,

i, Disodium 1-nitrosohydroxylamino-l-nitroethane, derived from sodium nitroethano, (l,BO g,, 0,01 mole) was dissolved in water (15 ml,) and acidified dropwise with 6N acetic acid at 0°,

The volume of gas that was evolved (slowly but continuously at this -50- temperature; more rapidly at 25°) was 85 ml, (corrected to 0°

760 mm,). Calcd, for the disodium nitrosohydroxylaminoethane structure: 112 ml., 0,005 mole of nitrous oxide. The reaction mixture, if made alkaline, became deep rod, this action is charac­ teristic of nitrolic acids (45» 49» 51» 52» 53)*

ii. Under the same conditions, disodium-l-ni- trosohydroxylamino-l-nitropropane, (1.94 g., 0.01 mole), yielded

124 ml. of gas; the calculated volume of nitrous oxide is 112 ml. (0.005 mole). The mixture gave a red color with alkali (test for a nitrolic acid).

No attempt was made either to identify the gaseous product, reported (49) to he nitrous oxide, or to absorb other oxides of nitrogen and carbon dioxide,

2, Oxidation with Silver ion

a. The Double Salt. Preparation of Secondary Gem

Dinitroalkanes.

i, 2,2-Dinitropropane, An aqueous solution of the double salt of sodium 2-nitropropane and sodium nitrite (2.0 g,, 0.011 mole in 20 ml.) vreis gradually added to a stirred solution of silver nitrate (4.O g., 0,0235 mole in 25 ml,). Silver was deposited almost instantly. Whan addition of reagent was complete, the mixture was distilled into 10 per cent alkali. Crude 2,2- dinitropropane (l,0-1,1 g, 70-73%), m.p, 47-49®, was isolated from the cooled distillate by filtration or by extraction with ether — 51- and evaporating the dried ether extract. After recrystallization from petroleum other the product melted at 53”53»5°; lit, 54°

(41), ^he melting point of the product was not depressed by the addition of an authentic sample of 2,2-dinitropropane,

ii, 2,2-Dinitrobutane, An aqueous solution of the product derived from sodium 2-nitrobutane and nitric oxide

(10,0 g,, 0,052 mole in 25 ml.) was added to aqueous silver nitrate

(12.0 g,, 0,071 mole in 25 ml,). The mixture was continuously extracted with ether for six hours. Distillation of the extract yielded a yellow liquid (4.O g., 79#), 'b.p. 190-19%° ’«’ith slight decomposition at 745 °®»* n^^ 1,434^; lit. b.p. 199° (ol).

The product was redistilled at reduced pressure to yield a colorless liquid, b.p, 78-80° (9-IO mm.), n^° 1.4353. dj°l,212; C 4 20 Mjj calcd. 31.92. Mg found 31.90. A small forerun, n^ 1,4290, was discarded. Calculation of yield is based on moles of silver nitrate used; the stoichiometric ratio for the complete reaction is two moles of silver ion to one of the double salt of sodium

2-nitrobutane and sodium nitrite, (of, 2,2-dinitrobutane, Experi­ mental Section III).

iii, 1,1-Dknitrocyclohexane. A solution of the product derived from sodium nitrocyolohexane and nitric oxide

(10,0 g., 0,045 mole) in water (50 ml,) was gradually added to a stirred solution of aqueous silver nitrate (15.O g,, 0,088 mole in 50 ml,) at 0-5°, The mixture was distilled to half-volume; -52- silver nitrate solution (4*0 g*» iu 50 ml.) was again added and distillation was continued till oil droplets ceased co-distilling with water. The yellow two-phase distillate was extracted with ether. Distillation of the dried ether extract gave crude 1,1- dinitrocyclohexane (5*2 g,), m.p. ca, 0-10°, h,p, 120-125° (35 mm,), n^^ 1,4690-1,4744* On standing at 0-5° for twenty-four hours the crude product partially solidified. The liquid phase was pipetted emd distilled (fractions (l), (2)); the solid phase melted on warming, n^® 1,4727, and was also distilled (fraction

(3))* Total; 4*7 8* (60#) of colorless product, (l) h.p. 69-71°

(1-1*5 mm.), n^O 1.4742 (2) b,p, 80-83° (Z mm.), n^° 1*4742,

,20, (3) b.p. 82-83° (2 mm,), n^^l.4742. Each fraction solidified, on D cooling, to a white waxy solid, m,p, > 20®; therefore, the samplet used for measurement of the refractive indices must have been super? cooled, A portion of the product was redistilled, b,p. 67.5-68®

(0,75 mm,), m,p, 33“35°; recrystallized from petroleum ether, m,p, 35*5“3^°* ^ second portion was recrystallized twice from petroleum ether, m,p, 35*5~3^°* 1,1-Dinitrocyclohexane is repor­ ted (60) as a yellow liquid, b,p, 142-143° (35 mm,), n^^ 1,4732, d^^ 1,2452, 4 The melting point of a mixture of the sample with 1,1-dini­ trocyclohexane which had been prepared by the oxidation of sodium nitrocyolohexane and sodium nitrite with silver ion (see Experimen­ tal Section III), was not depressed, m,p. 35*5“3^°* -53-

Traube'B ^roducts. Degradation to ^ater-Soluble

Products,

i « Di sodium 1-nitrosohydroxylamino-l-nitroethano

(49) was prepared from sodium nitroethano and nitric oxide. An aqueous solution of the product (l8,0 g,, 0,1 mole in ^0 ml.)

Tfas added, at 0°, to a stirred solution of silver nitrate (20,0 g., 0,118 mole) in water (5O ml,). Silver was deposited imme­ diately, gas was evolved continuously during the slow addition.

After one hour the mixture was filtered through charcoal; the filtrate, acid to litmus and having a strong odor of acetaldehyde, was made alkaline and treated with dilute aqueous silver nitrate,

A small amount of silver was formed. The mixture was filtered, re-acidified at 0° with sulfuric acid and extracted with an ether-benzene mixture (9*1), The dried extract was distilled in vacuum. An acidic water-soluble forerun (acetic acid odor, oxides of nitrogen) was discarded, A small amount of colorless liquid was obtained, b,p, 50-60° (5 mm,), n^^ 1,4358 which was immiscible with and more dense than water, and yielded a yellow solution with alkali. Identification of the product was not investigated further,

1,1-Dinitroethano has the following constants: b,p, 55“5^° (4»5 mm,), n 1,4346 (see Experimental Section III),

ii. Aqueous silver nitrate (18.O g,, 0,10$ mole in ^0 ml,) was added slowly at 0° to disodium 1-nitrosohy- droxylamino-l-nitropropane (49), the product derived from sodium -54- l-nitropropane and nitric oxide, (l9»4 g., 0,10 mole), in dilute sodium hydroxide (2.5#, ^5 ml*)* Silver was deposited and gas was evolved, The alkaline mixture was filtered through charcoal, and warmed at 200 mm, until a few milliliters distilled; the dis­ tillate gave a positive test for a carbonyl group with 2,4-dinitro- phenylhydrazino, The contents of the still pot were cooled to 0°, covered with ether and acidified slowly with 6N sulfuric acid.

The aqueous layer was extracted with ether (3 x 30 ml,); the com­ bined ether extracts were distilled in vacuum. Oxides of nitrogen were evolved throughout the dfetillation. Two fractions totalling

2,5-3.0 g, wore obtained! (l) b,p. 44"54° (20 mm,), n^° 1,3985;

(2) b,p, 54— 55° (20 mm,), n^^ 1*3995* Since both fractions were miscible with water they were not investigated. It was concluded that 1,1-dinitropropane had not formed in this reaction, .

3* Oxidation with Hydrogen Peroxide and with Salts of

Persulfuric Acid,

a. Coupling of Secondary Nitroalkane Anions to

Dimeric Vicinal Dinitroalkanes,

i . 2 ,3-Pimethyl-2 ,3-dinitrobutane, The double salt of sodium 2-nitropropane and sodium nitrite, obtained from the action of nitric oxide on sodium 2-nitropropane (l.O g., 0,0056 mole), was dissolved in water (30 ml,). Acetic acid (6n) was added dropwise, until a light blue color was produced, %ite cry­ stals of 2,3-d-lmethyl-2,3-

(60/«), m.p, 207-209*^; recrystallized from methanol, shining plates, m.p. and m.p, mixed with an authentic sample 210-211° (17). (The presence of sodium nitrite is not essential to this reaction) see

Experimental Section IV).

ii. 3 «4“Sioethyl-3,4-'iinitrohexane. The pro­ duct derived from sodium 2-nitrobutane and nitric oxide was oxi­ dized to 3»4“dimethyl-3,4“dinitrohexane (l7,40) (lO-12^J, plates, m.p. 7^~7®«5°» with the hydrogen peroxide-acetic acid reagent by the procedure given for the preparation of 2,3“dimethyl-2,3-dini- trobutane,

iii. l,l'-Dinitro-bicyclohexyl. Employing the procedure given for the oxidation of the double salt of so­ dium 2-nitropropane and sodium nitrite, the product obtained from sodium nitrocyolohexane and nitric oxide was converted into

1,1'-dinitro-bicyclohexyl (lO#i) by reaction with hydrogen peroxide and dilute acetic or hydrochloric acids. 1,1'-Dinitro-bicyclohexyl crystallizes from methanol, ether, or methanol-acetone in shining plates, m.p. 214° (darkening) -220° (dec.); lit. 216.5-217° (60),

208-209° (58).

iv. 1,1'-Dinitro-bicyclohexyl. Ammonium persul­ fate (2.3 g., 0,01 mole) was dissolved in water (5 ml.). The sol­ ution was made alkaline with dilute alkali and added at 0-10° to eui aqueous solution of the product prepared from sodium nitrocyclo- hexane and nitric oxide (2.20 g., 0.01 mole in 5 ml.). At the end — ^6—

of four hours, l.l'-dinitro-hicyclohexyl (0.21 g., l6#) was col­

lected and recrystallized from acetone, m.p, 215° (darkening)

-220° (dec.)

4« Action of Various Oxidizing Agents on the Double

*alt of Sodium 2—Hitropropane and Sodium Nitrite,

a. Oxidation with Mercuric Nitrate, Formation of

2,3-Nimethyl-2,3“

oxide (l.O g,, 0,005° mole in 15 ml.) was added to mercuric ni­

trate hemihydrate (2,0 g,, O.OO6 mole) dissolved in water (3O ml.).

The mixture became cloudy, then blackened; after heating to boil­ ing, it was extracted with ether. Evaporation of the dried ether

extract left 2,3-dinethyl-2,3-dinitrobutane (0.30 g., 5°/^) m.p, and m.p. with authentic material 210-211° (from methanol).

b. Oxidation with . Formation of Water

Soluble Products. Bromine water (5O ml, contain­ ing 1,6 g., 0,01 mole, saturated at 25°) was made alkaline with

solid sodium carbonate, and then added to an aqueous solution of the product obtained from sodium 2-nitropropane and nitric oxide

(l.O g,, 0.0056 mole in 2 ml,). After twenty-four hours in a

stoppered flask, the clear yellow solution was acidified with di­ lute hydrochloric acid, A small quantity of 2-nitroso-2-nitropro- pane separated (detected by its odor, blue solution in chloroform, blue melt on warming) which when volatilized left a trace of high “ 57“

melting yollo\T-v/hite solid (possibly 2,3'-dimothyl-2 , 3~dinitrobutano) ,

c. Oxidation with Potassium Permanganate. Formation

of Acetone, The double salt of sodium 2-nitropropane and sodium

nitrite in aqueous solution (l.O g., O.OO^o mole in 25 ml.) was

added dropwise to aqueous potassium permanganate (2.0 g., O.OI3

mole in 25 ml.). Heat and gas were evolved. Manganese dioxide

separated. After the mixture had cooled it was acidified with

hydrochloric acid and decolorized with sodium hydrogen sulfite.

The clear yellow solution was distilled to yield aqueous acetone;

the acetone was identified as its 2,/(-dinitrophenylhydrazone, m.p.

125-1260. -

d. Oxidation with Ferric Chloride. Formation of

Acetone. Mixtures of aqueous solutions of ferric chloride and

the mixed salt of sodium 2-nitropropane and sodium nitrite produced

deep red-colored homogeneous solutions* which were unchanged after

twenty-four hours at 25°, TOien the mixture was refluxed 2-nitroso-

2-nitropropane (blue condensate solidifying to the white dimer)

appeared in the condenser; the reaction mixture became brown.

Concentrated hydrochloric acid we.s added to the cooled mixture to

dissolve the brown solids. Distillation gave a homogeneous dis­ tillate; the fraction boiling between 30“98° contained acetone,

identified as its 2,4-dinitrophenylhydrazono,

*The red color produced with ferric chloride is a characteristic

of aci-nitro compounds and salts of nitro compounds (46). — 58“

III. Summary

Nitric oxide reacts vriuh sodium salts of secondary nitroalkanes

(e, g. 2-nitropropane, 2-nitrobutane, nitrocyolohexane) to produce a mixed salt (isl) of sodium secondary nitroalkane and sodium ni­ trite, Sodium nitrite is produced in this reaction by degradation of the salt of the secondary nitroalkane to yield sodium nitrite, the corresponding ketone and its oxime. The salt products are

shown to be mixed salts of sodium secondary nitroalkane and sodium nitrite rather than nitrosohydroxylaminonitroalkanes by quantita­ tive analysis and by their chemical reactivity. Upon acidifica­ tion of those salts obtained from 2-nitropropane, 2-nitrobutane and nitrocyolohexane, the corresponding pseudonitroles are obtained.

Oxidation of the double salts by silver nitrate (a new reaction) yields 2,2-dinitropropane, 2,2-dinitrobutane or 1,1-dinitrocyclo­ hexane, respectively. Reaction of the mixed salts with acidified hydrogen peroxide or with salts of persulfuric acid yields the vici­ nal dintro compounds 2,])-dimethyl-2,^-dinitrobutane, 3»4-dimethyl-?

3,4“dinitrohexane, or 1,1'-dinitrobicyclohexyl, respectively. The mixed salts were degraded to water soluble products (acetone in the case of the salt from 2-nitropropane) by potassium permanganate, sodium hypobromite, or ferric chloride.

All the reactions which have been investigated indicate that the properties of the salt products derived from nitric oxide and sodium secondary nitroalkanes are identical with those of a mixture -59-

(Ul) of sodiim secondary nitroalkane and sodium nitrite. -6o-

SSCTION III

PHEPARATION OF GEM DINITRO COMPOUNDS BY OXIDATIVE

NITRATION

I. Discussion

A, Introduction

As mentioned in Sections I and II n new method has been found by which salts of primary and secondary nitroalkanes are converted into gem dinitro compounds by the action of silver ni­ trate and inorganic nitrite in alkaline or neutral medium. This reaction proceeds in excellent yields; accordingly it has been investigated further, with the purpose of developing a general metliod for preparing substituted gem dinitro compounds.

B, Historical Review of Various Methods for the Preparation

of Gem Dinitro Compounds,

Pseudonitroles have been oxidized with chromic oxide in acetic acid (l6, 4I» 5^) to the corresponding secondary gem dini­ troalkanes; oxidation of 2-nitroso-2-nitropropane with oxygen

(16), hydrogen peroxide (16), nitrogen dioxide (16,62) or 100 per cent nitric acid (16) yields 2,2-dinitropropane in low yields (up to 20/4), 2,2-Dinitrobutane has been prepared by thermal decompo­ sition of 2-nitroso-2-nitrobutane (6l), However, the method gives poor results with other pseudonitroles (56), Ethane nitrolic acid yields 1,1-dinitroethane (lO^J when oxidized with nitric acid (63),

In general, ketoximes are oxidized to pseudonitroles with -6l- nitrogen dioxide (55*5^) or nitrous acid (^y); however, some oxi- aino compounds have been converted directly into gem dinitro com­ pounds by nitrogen dioxide (64, 65) (i.e. phenyldinitromethane is obtained from benzaldoxime) or nitric acid (00). Nitration of

2-chloro-2-methyl-3~nitrosobutane has been reported to yield 2- methyl-2,3»3~^rinitrobutano (Ô7).

Liquid-phase nitration of aliphatic hydrocarbons (3,4) has been reported to yield polynitroalkones containing the gem dini­ tro structure. Aikylbenzenos have been,nitrated on the side chain to yield mixture containing phenyl substituted gem dinitro­ alkanes (6 8 ). It is claimed (0 9 ) that in the vapor phase nitra­ tion of propane, 2,2-dinitropropane is produced along with mono- nitroalkanes. 2-Nitropropane has been nitrated to 2,2-dinitronro- pane with nitric acid or nitrogen dioxide at high pressures and temperatures (16,23); direct nitration of primary nitroalkanes failed. The formation of 2,2-dinitropropane by the nitration of cholesterol with nitric acid has been used e-s evidence for the presence of a gem dimethyl group in the sterol (yo). 2,2-Dinitro­ propane has been prepared by reaction of isobutyric and isovaleric acids, respectively, with nitric, acid (yi). Reaction of nitric acid with aliphatic ketones containing the propionyl or biityryl group yields 1,1-dinitroethane or 1,1-dinitropropano, respectively

(63,72,73); however the method is considered poor for preparing

1,1-dinitroalkanes (74). Ethyl acetate has been nitrated with -62- acetyl nitrate to give ethyl dinitroacetate (75) in low yield,

9-Diazofluorene and ethyl diazoacetate yield 9,9-dinitrofluorene and ethyl dinitroacetate, respectively, when treated with nitro­ gen dioxide (76).

Potassium has been prepared by reducing di- broaodinitronethane (77) in alkçline medium with arsenites (78) or thiosulfate (79). Dinitromethane has also been prepared by hydrolyzing 2,2-dinitro-l,1-diphenylethylene (80).

1.1—Dinitroethane and 1,1-dinitropropano may be prepared by reaction of the 1-bromo or the 1,1-Sibromo derivatives of 1-nitro- ethane or 1-nitropropane with potassium nitrite and potassium hydroxide (48, 2l); similarly, 1-chloro-l-nitroothane and 1-chloro-’

1-nitropropane (16) give 1,1-dinitroothane and 1,1-dinitropropane respectively (in about 5O per cent yield), Reaction of 1-halo- l-nitroethane with alkali (in the absence of added nitrites) also yields salts of 1,1-dinitroethane (16, 81)

Reaction of silver dinitromethane with methyl iodide gives

1,1-dinitroethane in low yield (78); 2,2-dinitropropane has been obtained (also in low yield) from silver 1,1-dinitroethane and methyl iodide (2l). This type of alkylation is of limited value,

1.1.1-Trinitroethane has been reduced to potassium 1,1-di­ nitroethane with 50 per cent aqueous potassium hydroxide (82) or with alkaline hydroxylamine (83), Reaction of 1,1,1-trinitroethane with methanolic potassium methoxide (or ethanolic potassium ethoxide) J

. —63“ yields the methyl (or ethyl) ether of potassium 2,2-dinitroethanol

(83), Potassium 3 i3"&in^"fopfopBuenitrile is produced from 1,1 ,1- trinitroethane aud potassium cyanide (83).

Condensations of the aldol—type have been effected with salts of dinitromethane, i.e., reaction of potassium dinitromethane with formaldehyde, acetaldehyde and propionaldéhyde yield the potassium salts of 2,2-dinitroethanol, 1,lpdinitro-2-propanol, and 1,1-dini- tro-2-butanol, respectively (84). 2,2-Dinitro-l,3~propanediol has been prepared from 2,2-dinitroethanol and formaldehyde (85). 1,1-

Sinitroethane and 1,1-dinitropropane condense with formaldehyde to yield 2,2-dinitro-l-propanol and 2,2-dinitro-l-butanol, respectively; reaction of those primary gem dinitroalkanes with acetaldehyde failed (16).

C. Discussion of Bosults

1. Beaction of Salts of Nitro Compounds with Sodium

Nitrite and Silver Nitrate. Scope of the Reaction.

A series of primary and secondary gem dinitro com­ pounds has been prepared by reaction of the salt of the corres­ ponding mononitro compound, silver nitrate, and a soluble metallic nitrite. The general reaction for the nitration of either pri­ mary or secondary nitro compounds is illustrated by Equation 1.

1. EHC=IIO~ + Noj + 2Ag+ -- » BBCCNOgig + 2Ag

H = alkyl, aryl, hydroxylalkyl, hydrogen.

The compounds directly prepared by this reaction (Table l) -64“ are dinitromethane (aS the potassium salt), 1,1-dinitroethane,

1.1-dinitropropano, 2,2-dinitropropane, 2,2-dinitrobutouie, 1,1- dinitrocyclohexane, ^y^-dinitrofluorene, 2,2-dinethyl-l,l,3~tri- nitropropane, l,l-dinitro-2-propanol, 2,2-dinitro-l-propanol, 3»3~ dinitro-2-butanol and 2,2-dinitro~l,3~propanediol. The average yield of product ' with the exception of dinitromethane (3^) and 9 ,g-dinitrofluorene {"J is 80 per cent* 3«3"Dinitro-

2-butanol and 2,2-diaethyl-l,l,3“trinitropropane are new compounds.

1.1-Dinitro-2-proponol could not be isolated in pure fora, Gem dinitro compounds containing an adjacent hydroxyl group serve as intermediates for 1,1-dinitro compounds since they are readily

split by alkali (Equation 2) to a dinitro compound and an aldehyde or ketone,

2, HECOH-CRCNOp^ + OH" ^ RHCO + + HgO.

Preparation of a gem dinitro compound in this way is illustrated by formation of potassium dinitromethane (47 to 505^ yield) from 1- nitro-2-propanol ^Equations 1 and 2),

Similarly, potassium 1,1-dinitroethane (yield, 1005^) results either from 2,2-dinitro-l-propanol or from 3,3"&lBÎtro-2-butanol; potassium 2,2-dinitroethanol is obtained in quantitative yield from

2.2-dinitro-l,3-propanediol.

Potassium dinitromethane could not be prepared by splitting potassium 2,2-dinitroethanol.

Although it is probable that the cleavage yields potassium “05“

dinitromethane, it is not isolated; instead the product is dipo­

tassium 1 ,1 ,2^^-tetranitropropane (yield, 47^^* Dipotassium

l,l,^,3"tetranitropropane can be synthesized in 77 P®* cent yield

by reaction of potassium dinitromethane with potassium 2,2-dini­

troethanol,

a. General Procedure, Conditions for Reaction» Isola­

tion of Products

Prior to nitration, salts of primary and secon­

dary nitro compounds are prepared, either as insoluble precipi­

tates by reaction with alcoholic alkali, followed by isolation,

or as solutions by reaction of the nitro compound with aqueous al­ kali, (in general, sodium salts of nitroalkanes are more insoluble

in alcohol than the potassium salts; potassium salts of 1,1-di­

nitro compounds are less soluble in water and alcohol than the

corresponding sodium derivatives.) The sodium salts of vicinal

nitro alcohols were precipitated from alcohol and then stored;

aqueous solutions of these salts are instable and therefore must

be used guickly.

An equivalent amount of potassium nitrite or sodium nitrite

is added to an aqueous solution of the salt to be nitrated and

the resulting solution is rapidly mixed with two equivalents of

aqueous silver nitrate. The oxidative-nitration proceeds in homo­

geneous solution at temperatures below 30° to form a light-colored

solid which decomposes rapidly into gem dinitro compound and finely m m

TABLE I

PHYSICAL COIÎSTAUTS^ OF POLYHITHO COMPOUNDS^

Conpoimd^ n.p.,°C t.p.,°C Calcd, Found Yield, lo 1 .1-Dinitroethano 55-56/4.3-4.5 1.4346 1.361 22.68 23.01 80

1,IpDini tropropane 57-58/4.0-4.5 1.4340 1.267 27.30 27.55 80

2 .2-Dinitropropane 53.5-54 ------—— ------— 92

2.2-Dinitro'butane 196-198/754 1.4354 1.212 31.92 31.90 81 77.8-78/10

1.1-Dinitrocyolohexane 35»5~3^ 82-83/2 — — ----— ----— ----- 75 06-67/0.7

1.1-Dinitro-2-propanol® ------—— ------50-70 I cr> I 2.2-Dinitro-l—propanol 89-9O ---—— ------— ---— — 70-85

3,3“Diaitro—2-butanol^ ------73-75/ca.2 1.4586 1.332 33.44 33.69 13 1.4550 2 .2-Dlnitro-l,3-propane- 139~140 ----— ------70 diol

2.2-Dimethyl-l,1,3-^ 122-123 ------— —— —--- 60-65 trinitropropane 9 ,9-Dinitrofluorene 130T131 ------7.8

a. Literature values appear in the experimental section

h. See Equation 1.

c. Dinitromethane -was prepared from sodium methanenitronate in low yields,

d. Obtained as a solid for the first time; previously reported as a liquid.

e. Was not isolated pure; yield estimated on weight basis and on degradations to potassium dinitromethane and acetaldehyde.

f. Hew compounds. -6y- divided metallic silver. The reaction is only slightly exothermic and is essentially complete in less than five minutes. Bulky mole­ cules, di sodium 2,2-dimethyl-l, 3“d.initropropane and potassium 9” nitrofluorene, react more slowly, side reactions occur (discussed later) and the yield of the corresponding gem-dinitro compound is lowered.

Optimum conditions for carrying out the reaction vary with the nature of the reactants and the products. In the preparation of

2.2-dinitro-l,3“propanediol, fO per cent yields are obtained if the silver nitrate is added rapidly; however, if the silver nitrate is added dropwise, the yield of dinitrodiol is lowered to 3 to 10 per cent. The low yield is attributed to the decomposition of so­ dium 2-nitro-l,3“propanediol into salts 2-nitroethanol and nitro- methane, and formaldehyde (Equation 3): each product is rapidly

3. HOCH2C(;N02)CH20H CHjO + CEq OHCHziHOÔ -— » CE^O + CEgsNOg oxidized by silver ion. It is apparent that the yield of dinitro­ diol is determined by rate of cleavage of the nitro alcohol and rapidity of oxidation of the products of decomposition. Similarly,

3,3"&iBitro-2-butanol is obtained in only 13 per cent yield; the principal product is 1,1-dinitroethane, Sodium 2-nitro-l-propanol is more stable in solution and therefore can be converted into

2.2-dinitro-l-propanol in yO-83 per cent yield.

In general, in the preparation of dinitro alcohols, silver nitrate solution is added (as rapidly as possible and with stirring) -68-

to a fresh solution containing the salt of the nitroalcohol and.

sodium nitrite cooled to 0°, However, the order of addition of

reagents is immaterial if the solutions are rapidly mixed.

Secondary gem dinitro compounds such as 2,2-dinitropropane,

2,2-dinitrobutane and 1,1-dinitrocyclohexene are relatively stable,

immiscible with water, and volatile in steam, therefore these nro-

ducts may be obtained from the mixture by steam distillation; ex­

traction of the mixture with solvents (, methylene

chloride or benzene) may also bo used to isolate the nitrated pro­

duct, 1,1-Dinitroalkanes are sensitive to steam, and therefore are

usually isolated by extraction. Often it is convenient to effect

the nitration in mixed solvent systems of water with ether, benzene

or methylene chloride; after reaction is complete, the mixture is

filtered from the silver residue and the organic layer in the fil­

trate is separated. All of the dinitroalcohols made in this way, with the exception of ]^?-dinitro-2-butanol, are very soluble in water, and are separated by extraction. Occasionally it is expedient

to isolate 1,1-dinitroalkanes as their potassium salts,

b. Identification of Products

The compounds thus made (Tables I and II) were identi­ fied by various combinations of the following methods! (a) compar­

ison of chemical and physical properties with those of an authentic

sample, (b) determination of molar refractivity, (c) quantitative

analysis, (d) determination of neutralization equivalent, (e) de- -69- gradation; preparation and analysis of derivatives, and (f) compar­ ison by infrared and ultraviolet absorption spectroscopy.

1,1-Dinitroethane, 2,2-dinitropropane and 2,2-dinitro-l,3“ propanediol wore compared with authentic samples. Determination of molar refractivities of liquid gen dinitro compounds serves as an excellent indication of the molecular composition of the compounds.

Acidic gem dinitroalkanes are converted into potassium salts and then analyzed for potassium. Potassium salts of primary gem dinitro compounds are converted into silver salts in quantitative yield by reaction with aqueous silver nitrate; analysis of these derivatives for silver is made by titrating the salts with standardized ammonium thiocyanate in dilute nitric acid using a ferric alum indicator (86),

Jacobson (l6) reported that 1,1-dinitroethane and 1,1-dinitropro- ponc can bo titrated with alkali using phenolphthalein as an indica­ tor (Equation 4).

4 . aCECwOgjp + RC(N02)2 ^ 2°

In the present research, it is shown that secondary gem dinitroal- kanes containing a hydroxyl group adjacent to the nitro groups be­ have as monobasic acids and can readily be titrated with alkali (Equa­ tion 2). Since the yellow color of the aci-dinitro group masks color changes of acid-base indicators, the neutralizations are best ' followed potentiometrically. Neutralization equivalents were deter­ mined for 1,1-dinitroethane, 1,1-dinitropropane, 2,2-dimethyl-l,l,3~ trinitropropane, 2,2-dinitro-l-propanol, 3#?"dinitro-2-butanol, and -70-

2 .2-d.initro-l,3*'P^opanediol (Table III),

It has been reported (l6) that 2,2r-dinitro-l-propanol is cleaved by alcoholic potnssiun hydroxide to form potassium 1,1-dinitroethano

(Equation 2), The reaction has been used in this research to identify

2.2-dinitro-l-propanol (prepared by oxidative nitration), ^\7-dini- tro-2-butanol (which yields potassium 1,1-dinitroethane) and 2,2- dinitro-l,3~propanediol (which yields potassium 2,2-dinitroethanol).

1,l-Dinitro-2-propanol could not be isolated either as such or as its potassium salt and therefore it was identified by conversion into potassium dinitromethane (50/<>). Gem dinitroalcohols yield yellow crystalline precipitates by reaction with 2,4-dinitrophenylhydrazine in acid solution, reaction of l,l-dinitro-2-propanol with acidic

2,4-dinitrophenylhydrazine yields acetaldehyde 2,4-d.iuitrophenylhy- drazone in 6l/^ yield (based on Equation 2) .

Ultraviolet absorption spectrograms were obtained for potassium salts of dinitromethane, 1,1-dinitroethane, 2,2-dinitroethancl, and l,l,l,3"tetranitropropane (Table II), The regions of maximum and minimum absorption by these compounds, based on data by Kortum (87), qualitatively indicate presence of the aci-dinitro group. The spec­ trogram for potassium dinitromethane was identical with that repor­ ted by Kortu# (87). -71-

TABLE II

PROPERTIES OF SALTS OF 1,1-DINITRO COI.IPOUNDS*

Com-oound Decomp. Range Solubility,25° Ultraviolet Absorption^ Ml.HgO for 0.10 g, M u log Ê

max. 362 4.32 Potassium min. 292 3.00 Dinitromethane 210, expl. 4* max. 247 3.63 SiIver 120, expl, Dinitromethane 100-135® insol.

Potassium 1,1-Dinitroethane 155t expl. il max. 377 4.19 min. 303 2.94 plateau 250 3.6 max. 220 3.78 Silver 1 .1-Dinitroethane 120-155 insol.

Potassium max. 365 4.28 2.2-Dinitroethanol^ljO-230 O.gi min. 295 3.00 max. 229 3.81

Dipotassium max. 387 4.27k l,l,3*3"totra- min. 297 3.15 nitropropano^ 220-250 11 max. 221 4.04 Disilver 1,1,3,3- tetranitropropane IIO-I5O insol.

a. See Equation 2. b. Does not yield a silver salt.

c. See Equations 9 end 10. d. Reported values range from 204-208° . e. 135°, expl. (78) f. ca, 150°, expl. (81) g. 1 part in 43 15° h. 1 part in 29 at 0° (72) i. 1 part in 6 (84).

je Aqueous solution, 10 ^ and 4 % 10 ^ moles/liter. The values for potassium dinitromethane are in good agreement with those taken from the spectrogram reported by Kortum (87). k. Log 6 values may be in doubt. Aqueous solution are not too stable. -72-

t a b l e III

NEUTRALIZATION EQUIVALENTS OF GEM-DINITRO COMPOUNDS

Compound* Indicator^ Calcd Found

1,1-Dinitroethane Potentiometri 0 120 122

1,1-Dinitropropane Potentiometric 134 137 Thymolphthalein 135,137 2,2-Dinitro-l-propanol Potentiometric 150 152

3,l-Dinitro-2—butanol Potentiometric 164 168 2 ,2-Dinitr0-1,3-pro­ Potentiometric 166 168 panediol

Thymolphthalein 165 2,2-Dimethyl-l,1,3“ trinitropropane Potentiometric 207 207

a. Solutions of weighed samples (o.001-0,002 molar quantities) of the compounds listed, (all of which were prepared by the silver in oxidation method), were titrated with O.IOOON sodium hydroxide. All exàept 2,2-dinitro-l-propanol and 2,2-dinitro-l,1-propanediol (which are water soluble) were dissolved in 5-IO ml. of methanol and 45*50 ml. of water.

b. All behaved as monobasic acids. Sharper endpoints (between pH9 and lO) were obtained potentiometrically than when the acid- base indicator, thymolphthalein (pH 9 »3“^0.5i colorless-blue) was used. At the endpoint observed with the acid-base indicator the color was green. -73-

c. Side Reactions

The reaction of disodiun 2,2-dimethyl-1,3—propane, silver nitrate and sodium nitrite vas studied in an effort to pre­ pare 2,2-dimethyl-l,l,3,3-tetranitropropane. If tv;o nitro groups could he introduced into dinitroneopentane it vould indicate that

steric factors do not play an important role in this fast reaction and that the less hindered molecules in which primary or secondary nitro groups are separated hy at least one carbon atom would read­ ily undergo polynitration. The oxidative nitration of dinitro­ neopentane occurs slowly and yields 2,2-dimethyl-l,l,3“trinitro- propane (6O#0 * No tetranitro derivative could be obtained.

The failure of 2,2-dimethyl-l,3~dinitropropane to undergo polynitration may be attributed to steric hindrance and acid-base reactions. If nitration is slow and stepwise with hindered mole­ cules, the newly formed primary gen dinitro group (a stronger acid than the mononitro group) can neutralize the unreacted primary nitronate and thus prevent it from undergoing nitration. Also, introduction of the first nitro group produces a more bulky mole­ cule in which the remaining mononitronate becomes less susceptible to approach. An important factor in the nitration reaction may be that competing oxidation reactions become more important when ni­ tration is slowed. It is known that silver salts of primary nitroalkanes are instable and decompose into coupled products

(93*94)• Also, in the coupling of sodium 2-nitropropane by per- -74- sulfate ion to yield 2,3-dimethyl-2,3-dinitro'butane, acetone is the principal hy-product (Section IV),

Competing reactions are observed in nitration of potassium

9~nitrofluorene. Nitration of potassium 9-nitrofluorene occurs slowly and yields 9 *9-&initrofluorene 9«9 '-dinitro-9 ,9 '- bifluorene and fluorenone (7*5/') « ^he relative importance

06 steric or electrical factors in this slow reaction cannot be assessed. However it is possible that the competing reactions, oxidation of 9-nitrofluorene ion by silver ion, may proceed through the 9-nitfofluorenyl radical (95) to the corresponding dimer or to fluorenone; these oxidation processes should be practically inde­ pendent of steric factors,

Further studies will be necessary before any definite conclu­ sions can be drawn concerning the electrical or steric effects of aromatic nuclei on the oxidative-nitration reaction,

2, The Preparation of Potassium Dinitromethane, Homologs and

Belated Compounds

a. Cleavage of Gem Dinitroalcohols with Alkali

Secondary gem dinitro compounds containing a hydroxyl group adjacent to the nitro group are readily cleaved by alkali into 1,1-dinitro compounds and the carbonyl derivative (Equation 2,

Table II), A modification of the reaction is used for preparation of potassium 2,2-dinitroethanol as follows: sodium 2-nitro-l,3~ propanodiol (prepared from nitromethane and formaldehyde in 94 to “75“

97 per cent yield) is nitrated to 2,2-dinitro-l,3-propanedioI; the dinitrodiol, upon reaction with potassium hydroxide, yields potassium 2,2-dinitroethanol (9^-100%). The overall yield from nitromethane is "JO per cent. Older methods for preparation of potassium 2,2-dinitroethanol involve the condensation of potassium dinitromethane and formaldehyde (84).

Reversibility of the reaction between salts of primary gem dinitro compounds and formaldehyde has been demonstrated by syn­ theses of 2,2-dinitro-l,3“propanodiol from potassium 2,2-dinitro- ethanol and formaldehyde (85) and of 2,2-dinitro-l-propanol from

1,1-dinitroethane and formaldehyde (lb), ^owever, it is reported

(16), that 1,1-dinibroethane and acetaldehyde do not condense to yield 3i3“&initro-2-butanol. Preparation of 3«3“&iaitro-2-butanol from 3“BÎtro-2-butanol was therefore attempted in an effort to de­ termine whether the dinitro alcohol can exist. Reaction of 3“B^tro-

2-butanol with othanolic sodium hydroxide yields a salt which reacts with silver nitrate and sodium nitrite and yields 3»3~di- nitro-2-butanol (l3^) and 1,1-dinitroethane (40-65^'). The 1,1- dinitroethane is probably formed by decomposition of %,3-&initro-2- butanol (Equation 5) or by the nitration of sodium nitroethane E 0 5. CH^CHOHC (NO_)_C&., CH^CHO + CB^CHfNOgyg produced cleavage of sodium 3“&itro-2-butanol.

Once it is isolated, 3*3“^^^^^ro-2-butanol can be distilled without decomposition. On the basis of these results and the fact -j6 - that 3,3-dinitro-2-tutanol is not obtained on acidifying a solution of potassium 1,1-dinitroethane and acetaldehyde (l6), it can be con­ cluded that at equilibrium the reaction represented by Equation 5 is in favor of the reactants. It is interesting to comment on the great difference between the behaviors of T^3"&initro-2-butanol and

2,2-dinitro—l—propanol in solution. Since both are secondary gem dinitroalcohols, attack by alkali must occur at the hydroxyl group to form the conjugate base, an alcoholate ion (Equation 6),

6, RCH0HCH(H02)2 + 0E~ --- » RCH-CEfNOgjg + 0_ which is in equilibrium (Equation j) with aldehyde and the conju­ gate (weak) base of a 1,1-dinitroalkane

7, RCH-CR(MO^)^ ; " RCHO + RC(N02)2

When a solution of potassium 1,1-dinitroethane and formaldehyde is acidified, 2,2-dinitro-l-propanol is obtained: under the same con­ ditions potassium 1,1-dinitroethane and acetaldehyde do not con­ dense (l6). This marked difference in behavior can be attributed to the differences in structure of the corresponding alcoholate ions

(Structures I and II):

O - C R g - C ( N O g ) Ô-CH-C(H0_) CH I 2 3 5

II

It appears that substitution of the methyl group for hydrogen leads to an internal strain greater in II than in I because of the larger spatial requirements for a methyl group. This internal steric ef­ -77- fect may influence the ease of decomposition of the alcoholate ion and the eguilihrium position between the aldehyde and 1,1-dinitro

compound. The concept of internal strain can also explain the great

difference between the yields of 2,2-dinitro-l-propanol and 3»3“

dinitro-2-butanol obtained in the nitration of the nitroalcohols.

5 , The Relationship of Potassium Dinitromethane to Potas­

sium 2,2-Dinitroethanol and Dipotassium 1,1,3,

nitropropane.

Oxidative nitration of nitromethane (Equation l) gives

low yields of potassium dinitromethane. Although a mixture of one equivalent each of nitromethane, aqueous alkali and sodium nitrite

is rapidly oxidized by silver nitrate (two equivalents), only traces

of potassium dinitromethane are obtained when an ethereal extract from the mixture is neutralized with aqueous potassium hydroxide.

Also, potassium dinitromethane is not obtained if two equivalents of potassium hydroxide are used during nitration in an attempt to

convert dinitromethane into potassium salt as it is formed.

Since alkyl-substituted nitromethanes such as nitroethane and

1-nitropropano are readily nitrated to give primary gem dinitroal- kanes (80-83^) it appears that nitration of vicinal hydrosyalkyl- nitromethanes should provide an efficient method for preparing po­ tassium dinitromethane. Cleavage of dinitroalcohols of this type into potassium dinitromethane end carbonyl compound was first noted by Duden and Ponndorf (84) and later by Lipp (3o). A saturated -y8- solution of potassium 1,l-dinitro-2-propanol slowly deposits potas­ sium dinitromethane, the more insoluble of the two salts (84) (Equa­ tion 8) .

8. CByCHOECfNOgjgE " ^ CH^CHO + CH(N0^)^K

In the present investigation, potassium dinitromethane was prepared in 47-50 per cent yield from l-nitro-2-propanol by a two step pro­ cess involving oxidative nitration (Equation l) of the sodium salt and cleavage of the crude product with potassium hydroxide (Equa­ tion 2), Attempts to isolate 1,1—dinitro—2—propanol were unsuc­ cessful because of instability of the crude product. The product of nitration (after being stored in ether for one week) yields an explosive salt when treated with alcoholic potassium hydroxide; precipitation of the instable salt is avoided by decomposing the aged product with potassium hydroxide in aqueous methanol. No po­ tassium 1,l-dinitro-2-propanol is obtained by these procedures.

Crude l,l-dinitr6-2-propanol is identified by cleaving it into po­ tassium dinitromethane (in basic medium) and acetaldehyde (in acid medium); the acetaldehyde was identified as its 2,4-d.initrophenyl- hydrazone.

The instability of crude 1,l-dinitro-2-propanol, when aged, may limit this process as a large scale method for preparing potas­ sium dinitromethane. Since potassium dinitromethane decomposes on storage, its is desirable to prepare a stable compound that can be readily converted into potassium dinitromethane. It has been - 79 - reported that potnssiun 2,2-dinitroothanol loses formaldehyde when heated in aqueous solution; the other products of decomposition were not determined (84). Potassium 2 ,2-dinitroethanol (prepared in 70^ yield from nitromethane, as described before) was investi­ gated as a possible source of potassium dinitromethane (Equation

9).

9 . H0CH2C(N02)2K ' ' CH^O + ^(1^02)2^

It is shown that potassium 2,2-dinitroethano cannot be cleaved into potassium dinitromethane. The methods investigated, in efforts to remove formaldehyde, were heating (a) aqueous potassium 2,2-dini­ troethanol and (b) aqueous mixtures of potassium hydroxide and po­ tassium 2 ,2-dinitroothanol. Upon heating aqueous potassium 2 ,2- dinitroothanol, dipotassium 1,1,3,3-tGtranitropropane (91) is ob­ tained in 47 P®- cent yield; after being heated under vacuum at temperatures up to or with potassium hydroxide for twenty min­ utes at 90-95°, potassium 2,2-dinitroethanol is recovered unchanged.

Upon heating pure samples of potassium dinitromethane and potassium

2,2-dinitroethanol in aqueous solution at 85-90° for four and one half hours, dipotassium l,l,l,T^tetranitropropane is formed in

77 per cent yield. The reactions postulated for the conversion of potassium 2,2-dinitroethanol into dipotassiu%l,l,T,^-Tetranitro- propane (Equations 9 and lO) are*

9. H0CH2C(N02)2 ~ " CHgO + ^(NOg)^

10. CH(N02)i + H0CH2C(N02)2 . H 2 O + "(02N)2C-CH2-C(N02) 2 — 80“

Only part of the formaldehyde Is expelled (odor) on heating potas­ sium 2 ,2-dinitroethanol. The process hy which the tetranitro deriva­ tive is formed is formally analogous to a Mannioh reaction (^2).

The samples prepared hy this series of experiments, potassium dinitromethane, potassium 2,2-dinitroethanol and dipotassium 1,1,3»3“ tetranitropropane, are distinguished from each other hy their (a) solubilities in water, (h) decomposition points, (c) reactions with silver nitrate, (d) decomposition points of silver salts, and (e)

quantitative analysis. Ultraviolet spectrograms of these salts show small differences in the range, '^oO-yjO mu, of maximum absorption of the aci-dinitro structure (87). Properties of these potassium salts and their silver derivatives are given in Table II.

3 » The Effect of Various Oxidizing Agents in the Oxidative "i-

tration.

It was found (Section II) that oxidation of the mixed salt (ill) of sodium 2-nitropropane and sodium nitrite (obtained from nitric oxide and sodium 2-nitropropane) with either mercuric, persulfate, hypobromite, or permanganate ions does not yield 2,2-dinitropropane.

2,3-Dlmethyl-2,3-dinitrobutane is formed by reaction of the salt with mercuric nitrate, ammonimm persulfate, , or acidic hydrogen peroxide; only acetone is obtained by reaction \-dth sodium hypobromite or potassium permanganate.

Investigation of the effect of oxidizing agents on a mixture of sodium 2-nitropropane and sodium nitrite was continued in the —Si*" hope of finding an economical substitute for silver nitrate in the

nitration. Preparation of 2 ,2-dinitropropane from sodium 2-nitro­ propane and sodium nitrite was attempted with mercuric, cuprous and

cupric ions, salts of persulfuric acid, or acidic hydrogen peroxide

as the oxidizing agent. Of these reagents, only mercuric nitrate

is effective; 2,2-dinitropropane is obtained in 25 per cent yield.

However, this result is difficult to reproduce; in many experiments

the principal product is 2 ,3~dimethyl-2 ,3“iiuitrobutane. On the basis of this brief study, mercuric ion is much inferior to silver

ion for effecting oxidative-nitration.

Oxidation of mixtures of sodium 2-nitropropane and sodium ni­

trite with persulfate ion or acidic hydrogen peroxide proceeds as before (Section II) to yield 2 ,3”diwethyl-2 ,3“iiinitrobutano (ll-

12.5^). Aqueous cupric chloride, an acidic reagent, rapidly decom­ poses mixtures of sodium 2-nitropropane and sodium nitrite to yield

2-*nitroso-2-nitropropane. Ammoniacal cupric chloride slowly oxidizes

a mixture of the salts to form 2,3~

Oxidation of the 2-propanenitronate-nitrite ion system whether with

Tehling's solution, cuprous chloride or cuprous acetate, fails to yield either 2,2-dinitropropane or 2,3~

It appears from these results that introduction of the second nitro group during oxidation depends on an intermediate complex whose decomposition into the dinitro group is sterically favored

(Equations 11 and 12) . - 82-

11. RECzrHOj + +NO2 +2Ag EEC=l/ Ag

Ag

12. (EECsNO^AgONO)” Ag"*" -- ► BacCüOg)^ + 2Ag

The structure of the intermediate silver complex is analogous to structures of other complex silver salts, e.g. HAgCl^, AgAg(CN)^,

Ag(i-'H^) 2CI, in which silver ion is coordinated with two electron- donating groups. Extension of this concept may explain why the tetrammino copper (ll) ion (in which cupric ion is already strongly coordinated with electron-donating groups) does not oxidize 2-pro- paaenitronate-nitrite ions to 2,2-dinitropropane, and why it slowly undergoes reduction at higher temperatures to yield 2,3“diaethyl-

2,3-dinitrobutane. This concept is parallel with the fact that the tetrammine copper (ll) ion has à lower oxidation potential than cupric ion (96).

II. Experimental

A. Oxidative Nitration; the Preparation of &em Dinitro

Compounds

1. Preparation of 1,1-Dinitroethano

a. 1,1-Dinitroethane. A solution of rectified nitro­ ethane (15.0 g., 0.2 mole) and sodium nitrite (4.O g., 97^ assay) in aqueous sodium hydroxide (8.5 g., 80 ml.) was added as rapidly as possible to a stirred mixture of aqueous silver nitate (70.9 g.,

0.41 mole in 120 ml.), sodium hydroxide, (2-3 drops until silver - 83- oxiàe appeared) and ether (15O ml.) at 0-3°. (The solution of nitro­ ethane in alkali was effected below 20®), A cream-colored solid formed immediately which almost filled the flask. The temperature of the mixture rose to 10®; when the solid began to decompose rap­ idly (with blackening, large reduction in volume) the temperature rose to 20®, After a few minutes, the temperature began to fall; the cooling bath was removed and the mixture was stirred for thirty minutes at 20®, The silver deposit was filtered and washed with othor-bonr.cnc (3 i5*li 2 x 10 ml.). The etheral layer of the filtrate, after being washed with a saturated sodium chloride solution and filtered slowly through anhydrous , yielded, on dis­ tillation, colorless 1,1-dinitroethane (18,5 g,, y8#); b.p. 55*5“

56® (4.25-4.5 mm.), n^Ol.4341-1,4346» dj° 1,355. found 23,11,

Neut, oquiv. calcd. 120, nout, equiv. found, 122,

b. Potassium 1,1-dinitroethane. Beaction of 1,1-dini­ troethane with methanolic potassium hydroxide yielded potassium

1.1-dinitoethane, a yellow salt, which when recrystallized from hot water, as shining yellow plates melts at 155° (expl). Thé potassium

1.1-dinitroethane developed a characteristic red tinge (81) when exposed to air and light.

Anal, Calcd. for ('2^3^4^2^' 24»72, Pound: K, 24.37» 24.41.

1,1-Dinitroethane prepared from 1-chloro-l-nitroethane and potassium nitrite has the following physical constants: b.p, 5^“ 57°

(4.5 mm,), n^ 1,4346, d^ 1,361, found 23.01. The ultraviolet -84“

spectrum (Beckman Model DU quartz spectrophotometer; 1 on, cell)

of aqueous potassium 1 ,1-dinitroethane (lO x 10 ^ molar) has

an absorption maximum at 377 log ^ max. 4*19 (Table II). The

infrared spectrograms of both stunnies of 1 ,1-dinitroethane are iden­

tical.

c. Silver 1 ,1-Dinitroethane. An aqueous solution of

potassium 1 ,1-dinitroethane (0 .l6 g . , 0.001 mole in 4 ml.) was

added to an aqueous silver nitrate solution (O.34 g . , 0.002 mole,

100% excess, in 1 ml.) at 3-10°. Filtration of the mixture yiel­

ded bright yellow plates of silver 1 ,1-dinitroethane (O.23 g., 100%,

washed with methanol and ether), üfhe silver salt, stable at room

temperatures, darkesn progressively at temperatures above 120° and

decomposes suddenly at 135°.

An^.* Calcd. for CgB^O^NgAg: 47*?4. Found: 47*55*

2. Preparation of 1,1-Dinitropropane

A solution of 1-nitropropano (8.9 g., 0.1 mole) and

sodium nitrite (7*1 g., 97^ assay) in aqueous sodium hydroxide j (4*2 g . , 1.05 mole in 45 al.) was added rapidly to a mixture of

aqueous silver nitrate (34*0 g., 0.2 mole) and ether (13O ml.) at

0-3°. The procedure followed was identical with that described for

the preparation of 1,1-dinitroethane. Distillation of the product

*Tho content of silver was determined by titrating the salt (O.OOO3

moles) in dilute nitric acid with standard ammonium thiocyanate

(0.05 N) using a ferric alum indicator (86). — 85“

yields l.l-dinitropropane (1O.8 g., 8o^i, no fore-run or residue),

■b.p. 56-58° (4-4.5 mm.), n^*^ 1.434°, d^° 1.26?; found 27.55;

neut. equiv, calcd. 134. neut. equiv. found 1^6. (The neutralization

equivalent was determined hy potentiometric titration of the product

with standard alkali; when thymolphthalein (pH 9 *3"^®*5. colorless-

hlue) was used as indicator, values for the neutralization equivalent

of 135 nnd 137 wens obtained; however, the end point (green in this

system) was difficult to see.)

3. Preparation of 2,2-Dinitropropane.

2-Hitropropane (8.9 g., 0,1 mole) and sodium nitrite

(7.5 £•• 97^ assay, 0.11 mole) were dissolved in a slight excess

of 10 per cent aqueous sodium hydroxide (4.5 g . , 0.11 mole in 45

ml.) Aqueous silver nitrate (34.° g., 0.2 mole in 100 ml.) was then

added dropwise to the stirred mixture. The light colored salt that

formed initially blackened rapidly, evolving heat; during addition

of the silver nitrate solution the mixture was cooled periodically.

The mixture on distillation (till two-phase droplets no longer

appeared) yielded 2,2-dinitropropane (l2.7 g., 92«7^). “ »p. 48“5°°.

When recrystallized from petroleum ether, the dinitro compound melts

(alone or when mixed with an authentic specimen) at 53«5~54°; lit.

(41) m.p. 54°. (of. Preparation of 2,2-dinitropropane, Experimental

Section II).

4. Preparation of 2,2-Dinitrobutane.

2-Nitrobutane (IO.3 g., 0.1 mole) and sodium nitrite - 86—

(7*5 g*« 97^ assay) were dissolved in. aq,ueons sodinm hydroxide (4 »& g., 0,115 mole in 45 ml,). Aqueous silver nitrate (34*0 g., 0,2 mole) was added dropwise to the stirred solution at 0-20°, After ten minutes the mixture was distilled, The two—phase distillate was extracted with ether. Distillation of the ether extract (dried with anhydrous calcium sulfate) yielded pure (there was no fore-run and no residue), colorless 2 ,2-dinitrobutane (l2,0 g., 8l^), b.p.

77,8-78° (10 mm,), n^O 1,4353, d|° 1.2124: calcd. 3I.92, found 31«90" (cf. Preparation of 2,2-dinitrobutane, Experimental

Section II).

Anal, Calcd. for 32.43: H, 5.44; H, 18.92.

Pound: C, 31.88; H, 5.37; N, 18.16.

5. Preparation of l.l-^initrocyclohoxane

Nitrocyclohexane (l2«9 g., 0,1 mole) was refluxed for ten minutes with aqueous sodium hydroxide (6,5 g., O.I37 mole in

65 ml.) until the solution was homogeneous. The yellow solution was cooled to 0° and treated with aqueous silver nitrate (34*0 g.,

0,2 mole in 80 ml,). On steam distillation of the aqueous mixture, a product was obtained first as a yellow insoluble liquid, then as a white, waxy solid (m.p, 33-35°), The distillate was extracted with ether. Distillation of the ether extract (dried with anhydrous sodium sulfate) yielded a colorless product (l2,2 g., fO^o), b.p.

65-67° (0.7 mm,), m.p. ca. 30°. When redistilled (b.p. 67,5-68°

(0.75 mm)) or recrystallized from petroleum ether, 1,1-dinitrooyclo- - 87-

hexane melts at 35*5“3^»

Anal. Calcd. for C, 41.38; », 5.79: N, I6.O9.

Found* C, 41.23; H, 5.6 4 ; N, I5.85.

1,1-Dinitrocyclohexane is reported (60) as a yellow liquid,

b.p. 142-143° (35 ““»)* 1.4732» 1,2452. (of. Preparation of

1,1-Dinitrocyclohexano, Experimental Section II).

6. Preparation of 9 ,9~Dinitrofluorene, 9 »9 '“Dinitro-9 »9 '~

Difluorenyl and Fluorenone.

a. Potassium 9-nitrofluorene was prepared in oO per

cent yield by nitrating (27) fluorene with methyl nitrate (43) and

potassium athylato,

b. Aqueous silver nitrate (6.8 g., O.O4 mole in 20 ml.)

was poured into a stirred mixture of potassium 9"Bltrofluorene (5.O

g., 0.02 mole), sodium nitrite (2,1 g., 97^ assay, O.03 mole), water

(50 ml,) and benzene (5O ml.), A yellow solid formed and immediately

agglomerated and turned rod. In two hours the lumpy solid, while

being vigorously stirred, darkened slowly, turning orown then black.

After forty-eight hours at 25-30°, the mixture was filtered; the

black residue was washed several times with hot benzene. The benzene r,' extract was washed with a saturated sodium chloride solution and

distilled in vacuo until a yellow solid crystallized. After petro­

leum ether (25 ml.) was added, filtration of the mixture yielded

crude 9,9'-dinitro-9,9'-bifluorene (l,60 g., 76%), pink powder, m.p,

160-166°. After several recrystallizations from benzone-petroleum -88- othor, iriiite crystals of 9 i5'“dinitro-9,9 '“tifluorene were obtained, m.p, l80-l°, 183-183.9°. a small portion had a pale blue cast and melted at 182-183°. The melting point of the white solid was not depressed by the addition of the blue product nor by an authentic specimen of the dimer.*

The benzene-petroleum other filtrate was concentrated; upon addition of petroleum ether, a peach colored solid, crude 9»9“di~ nitrofluorene, precipitated (0.4 g., y.8#J. The amorphous solid began to decompose at 120° and finally melted with evolution of gas at 126°. After recrystallization from ether-petroleum ether, 9 ,9“ dinitrofluorene, peach colored, melts with decomposition at 130-131.5°,

The reported melting point of 9 ,9“&i&itrofluorene is

128° (dec.) (7Ô).

■^nal. Calcd. for C^^BgO^Ng* 10.94» found: N, 10.54

The filtrate that remained after the 9 »9“&initrofluorene frac­ tion had been isolated, was concentrated and, on cooling, a yellow solid separated, m.p. 70-75°, Becrystallization of this product from a mixture of ethanol and water yielded fluorenone (O.16 g.,

4.1^), m.p. 82-83°; lit.(97)» The ketone gave the characteristic

*An authentic specimen of 9 ,9 '“&initro-9 ,9 '“bifluorene was pre­ pared (95) by reaction of potassium 9“uitrofluorene with iodine and then decomposing the 9“lodo-9~nitrofluorene in boiling methanol or in hot acetic acid. The crude product, white and pink, amorphous, after several recrystallization from benzene-petroleum ether, was obtained as a white crystalline solid, m.p. l80-l8l°. - 89- violet color (Sj) in concentrated sulfuric acid.

Evaporation of the mother liquors left a yellow oil which did not crystallize from mixtures of benzene, ether and petroleum ether.

Two-thirds of this material, when treated with hydroxyleuaine reagent

(98) yielded fluorenone oxi::e (0.08 g., 3*1^) ! after recrystalli­

zation from chloroform-petroleum other, the oxime, yellow needles, melts at 193-194°: lit. 193-194° (97.98).

7. Preparation of 2 ,2-^imethyl-l, 1 ,3-i^rinitropropane.

a. Formation of the Intermediate Disodium Salt of

2.2-dimethyl-l,3-dinitropropane.

2.2-Dinethyl-l,3“dinitropropane (4.2 g . , 0.025 wole) in absolute ethyl alcohol (50 ml.) was added at 0° to a stirred solution of sodium (1.3 g., O.O6 mole) in alcohol (50 ml.). The jelled white mass was filtered; the product was washed with a little alcohol and then liberally with ether. After being dried in vacuo in a desiccator, the white, powdery disodium dinitroneopenO tane weighed 6.1 g. (theoretical, 5*4 g.). A portion (O.l g.) was dried over phosphoric anhydride (2 hours at 2 mm.) and analyzed for sodium (as sodium sulfate).

Anal. Calcd. for C^HgO^N^Na: Ka, 12.49. Calcd. for C^HgO^N^Na^:

Na, 22.31. Calcd. for C^H0O4N2Na2*C2H^OHi Na, 18.3O.

Pound: Na, 17.7 »

b. Nitration of the Intermediate Salt to 2,2-Dimethyl-

1,1,3-trinitropropane. —90—

The salt product prepared from dluitroneopentane was dissolved, in portions, in ice water (50 ml,); sodium nitrite

(5,0 g., 97^ assay, 0,07 “ole) was added, and the cold solution was poured rapidly into a stirred mixture of aqueous silver nitrate

(18,5 g,, 0,11 mole in 5O ml,) and methylene chloride (lOO ml,).

After one hour the mixture was filtered; the silver residue was washed with methylene chloride. Evaporation of the organic layer from the filtrate left a yellow waxy solid (3*4 g.) which, when tri­ turated with cold Skollysolve F (25 ml,) and filtered, yielded 2,^-

dimethyl-1,1,3-trinitropropane (3.2 g., bl# crude), m,p, 106-109°.

Purification of trinitroneopentane hy fractional crystallization from ether—petroleum ether gave a white product, m,p, 122.5-123°, which was soluble in ether, methanol, chloroform, benzene and aqueous

alkali (deep yellow solution) and insoluble in water and cold petroleum other. The melting points (between II4 and 122°) of less pure fractions were not depressed when pure 2,2-dimethyl-l,1,3- trinitropropane was added in varying amounts. The probablg contam­ inant, 2,3“dimethyl-l,3-dinitropropane (m.p, 93° (99.100)) could not be easily eliminated by fractional crystallization. However, when crude material was treated with 0,5 H sodium hydroxide for ten minutes and the solution filtered and reacidified at 0°, 2,2-dimethyl-

l«l,3"trinitropropane was regenerated; recrystallized from ether- petroloum ether, m,p, 121-122°, or sublimed at 0O-70® (5 mm,), m.p,

121-122°, -91-

Ihe neutralization equivalent, 20J, of 2,2-dimethjrl-l,1,3- trinitropropane (m.p. 122-123°) agreed with the theoretical value,

207, for the monobasic acid (titrated potentiometrically with stan­ dard alkali).

Anal. Calcd. for C, 29.00 ; H, 4.38; N, 20.29.

Found; C, 3O.37; H, 4.2I; N, I9.O5.

8. Preparation of 2 ,2-Dinitro-l-Propanol

a. Sodium Salt of 2-Nitro-l-propanol

A solution of 2-nitro-l-propanol (44) (IO.9 g.,

0.1 mole) in ether (25 ml.) was added dropwise at 0° to a stirred solution of sodium (2.3 g., 0.1 gram atom) in absolute ethanol (5O ml.). Ether (25 ml.) was added and, after one-half hour, the mix­ ture was filtered. White, powdery sodium 2-nitro-l-propanol (l2.2-

12.7 g.» 9^-100%) was obtained after the solid had been washed and stored for twenty-four hours over anhydrous calcium chloride.

b. Nitration of the Salt to 2 ,2-Dinitro-l-propanol.

Sodium 2-nitro-l-propanol (6.0 g . , O.O4B mole) and sodium nitrite (4.O g., 97^ assay) were dissolved in water (30 ml.).

Aqueous silver nitrate (17.O g., 0.1 mole in 50 ml.) was added (all at once) at 0°, and the cream-colored solid that formed was shaken

(or stirred) for two to three minutes until it had decomposed to metallic silver. After fifteen minutes, the reaction mixture was filtered; the black residue was washed with hot water (3 x 10 ml.) and ether (50 ml,). The aqueous filtrate and washings were extracted “ 92- "■ - with ether-henzene (9 :1 , 4 % &0 ml.); the combinai extracts were washed with saturated sodium chloride solution and concentrated to approximately 30 ml. by distillation. On slowly adding the cooled concentrate to Skellysolve F (19O ml) at 0-9°, 2 ,2-dinitro-l-propa- nol separated as a white, waxy solid. Additional product was obtained by concentrating the mother liquor and adding Skellysolve F, The crude product obtained (thrice) in yields of 7O-89 per cent, melted at 89-90° (after twenty-four hours over phosphoric anhydride); re­ crystallization from petroleum ether-benzene did not alter its melt­ ing point (lit. 90° (16)).

Anal. Calcd. for C^E^O^Ng: C, 24.00; H, 4 .03! 18.67. Found:

C, 24.14, 24.14; H, 5.74, 3.73; N, 17,85, 17.81.

2,2-Dinitro-l-propanol behaved as a monobasic acid idien titra­ ted (potontiometrically) with standard alkali; neut. equiv. calcd.

150, neut. equiv, found 152. The end point (pH lO) was not as sharp as those observed for 1,1-dinitroethane, 1,1-dinitropropano, 3»3” dinitro-2-butanol and 2,2-dinitro-l,3~propanediol.

c. Potassium dinitroethane.

When a methanolic solution of 2 ,2-dinitro-l-propanol was addod to methanolic potassium hydroxide, potassium 1,1-dinitro- ethano, m.p. 150° (expl.), precipitated immediately in quantitative yield. (See 1,1-dinitroethane, Experimental, this section).

9. Preparation of 3 ,3"Dinitro-2-butanol.

a. Salt Product from 3-aitro-2-butanol. - 93-

An ether solution of 3-nitro-2-hutanol (44) ( H «9 g . , 0,10 mole in 50 ml,) was slowly added to a stirred solution of sodium (2.3 g., 0,1 mole) in absolute ethanol (50 ml,% at 0®, After addition was completed and the mixture filtered the yellow salt

(11,2 g ., 80fo) was washed with ether and stored over anhydrous cal­ cium chloride,

b « Conversion of the Intermediate Salt into 3 *3“

Dinitro-2-butanol and 1 ,1-Dinitroethane,

The sodium derivative of 3-mitro-2-butanol (ll,2 g,) and potassium nitrite (y,0 g,, assay, 0,03 mole) were dissolved in water (40 ml,) at 0-5°. Aqueous silver nitrate (27,5 g«, l,o2 mole in 90 ml,) was added rapidly; a cream colored solid was formed that darkened rapidly. The temperature was kept below 20® for fif­ teen minutes and then allowed to rise to 25®, The silver deposit was filtered and washed with ether; the aqueous filtrate was ex­ tracted with ether. Distillation of the combined and dried ether extracts yielded the following colorless liquid fractions: (l) b,p, 38-460/2 mm,. 3,6 g., n^° 1,4379: (2) b.p. 48-60°/2 mm, . 1,8 g., n^° 1.4358; (3) b.p. 73-75®/2 mm, , 2,1 g., n^° 1,4586-1,4590, d^® 1,332, Fraction (3) was identified as 3»3-&lnltro-2-butanol 4 (yield, 13% from 3-aitro-2-butanol): calcd., 33,.14; found,

33,69, Anal. Calcd. for C^HgO^Nj*. C, 29,27; H, 4.9I; H, I7.O7; neut, equiv, (monobasic acid), 164, Found: C, 29,10; H, 4 ,25;

N, 17,02; neut, equiv,, 168, Methanolic potassium hydroxide -94-

conrorto 3i3"&ÎBltro-2-butanol into potassium 1,1-dimitroethane,

expl, 156°, in quantitative yield.

Fractions (l) and (2), a mixture of 1,1-dinitroethane and

3,3-&iBitro-2-hutanol, ■were redistilled to give 1,1-dinitroethane,

^•P* 57“5®° (5 1,4352# identified as potassium 1,1-

dinitroethane, expl, 1$6°, (See 1,1-dinitroethano and its potassium

derivative, this section). The yield of 1 ,1-dinitroethane was 40#"*

In other experiments, using the procedure described for preparations

of 1,1-dinitroethane and 1,1-dinitropropane, the yield of ^^^-dini-

tro-2-butanol was not improved; however, 1,1-dinitroethane was ob­

tained in yields of Ô5 per cent from ]^nitro-2-butanol,

10, Preparation of ■2 ,2-Dinitro-l,3“propanediol ,

a. Sodium 2-nitro-l,3-propanediol (94“97<® yield) was

prepared by the reaction of nitromethane, paraformaldehyde and sodium

methoxide in methanol. The dried product contains two molecules of

methanol of crystallization (88,89).

b, 2,2-Dinitro-l,3-propanediol. Aqueous silver nitrate

(17,0 g., 0,10 mole in 50 ml,) was added all at once* to a solution

of sodium 2-nitro-l,3“propanediol (7»2 g,, 0,035 mole) and sodium

nitrite (4.O g., 97^ assay, 0 ,05b mole) in water (40 ml,). The

reaction flask was swirled vigonously in an ice bath so that the

temperature did not exceed 30°, After thirty minutes at 25-30°

the mixture was filtered and the silver deposit was washed with hot

*In two experiments, dropwise addition resulted in low yields (lO%)

of the diol. -95-

Tdth hot water (3 x 10 al.). The combined filtrate and washings were extracted with ether (6 x 50 ml.); benzene (3O to 40 ml.) was added to the ether extract and the mixture was distilled at atmospheric pressure: ether, benzene-water azeotropo, an- benzene were removed until crystallization occurred. The product was cooled, filtered, and washed with chloroform. Approximately 4*0 g. (average yield) of 2 ,2-dinitro-l,3"propanediol was obtained, white needles, m.p,

-34-137°: a second crop of crystals weighed Q.3 g., yellow solid, m.p. 110-116°, Yield of crude product: 72~75 P®- cent. Vfhen recry­ stallized from hot benzene and dried over phosphoric anhydride,

2,2-dinitro-l,3-propanediol (alone or when mixed with an authentic specimen**) melts at 139-140°. Yield: yO per cent.

Anal. Calcd. for CyHgO^Ng: C, 21,09; H, 3,64; N, lb,87;

neut. equiv. (monobasic acid), 166, Found: C, 21.93:

H, 3,64; N, 16,12; neut. equiv., 168.

2 ,2-Dinitro-13,-propanediol was further characterized by con­ version into potassium 2,2-dinitroethanol in quantitative yield

(see below).

c. Preparation of Potassium 2 ,2-Dinitroethanol.

2,2-iJinitro-1,3-propanediol (1.66 g., 0,010 mole) in methanol (15 ml,) was added dropwise at 0-5° to a solution of potassium hydroxide (O.7 5 g., 85% assay, 0,011 mole) in water

**An authentic specimen of 2 ,2-dinitro-l,3-propanediol (85) was supplied by Herman Roy end Frank Conrad, The General Tire and

Rubber Co., Akron, Ohio. -gÉ-

(5 ml.) and methanol (20 ml,). On filtering and washing the yellow crystals with methanol and ether, potassium 2 ,2-dinitroethanol (l,73~

1,74 g., 99“1005t) was obtained; decomposition range 130-230°,

Anal. Calcd, for C^H^O^N^Kl C, 13,79; H, I.74; N, 16.O9

Found: C, 13,57? 1,3b; N, 15,96. Ultraviolet absorption spec­ trograms (Beckman Model DU quartz spectrophotometer, 1 cm. cell) of aqueous solutions of potassium 2,2-dinitroethanol (lO~4 ^ 4 * 10~^ molar) show a band peaked at 3^5 mu, log E max. 4*28; a minimum at

295 mu, log E min, 3.00? and a band at 229 mu, log E max. 3.81).

The solubility of potassium 2 ,2-dinitroethanol in water at 20-25° was 0.1 g., in 0,8 ml,; the reported solubility is one part in six

(84). When treated with aqueous silver nitrate in concentrated aqueous solutions, potassium 2,2-dinitroethanol did not yield a water insoluble silver salt; the mixtures deposited a silver mirror, slow­ ly at 25° and rapidly when warmed,

11, Preparation of Potassium Dinitromethane from l-Hltro-2-

Propanol,

a. Sodium Derivative of l-Nitro-2-propanol,

An ether solution of l-nitro-2-propanol (44) (IO.5 g., 0,10 mole in 4O ml.) was added to a stirred solution of sodium

(2,3 g . , 0,1 mole) in alcohol (5O ml.) at 0°. The product was fil­ tered, washed with ether, air dried (l5*0 g,; theoretical, 12,7) and immediately used for nitration,

b. Conversion of the Intermediate Salt into 1,1-Dini-

tro 2-propanol (Crude) -97-

The sodium salt was dissolved in ice water (75 ml*) and sodium nitrite (7*5 g*, 97^ assay, 0.11 mole) was added. The solution was poured into a stirred mixture of aqueous silver nitrate

(34*0 g.» 0,20 mole in 75 ml*) and ether (lOO ml.) at 0 °. After five minutes the cooling bath was removed; stirring was continued for one half hour. The mixture was filtered and the silver deposit was washed with ether. Saturated sodium chloride solution (a few milliliters) was added to the combined filtrate and washings, and the suspension was filtered free of silver chloride. The ether layer was separated, and the aqueous layer was extracted with ether (4 x

$0 ml,). The combined ether extracts were washed with a saturated sodium chloride solution (2 x 10 ml,) and slowly filtered through anhydrous sodium sulfate. After ether had been removed under vac­ uum, a mobile yellow liquid (approx. I5 g; theoretical, 15.O g.) of disagreeable odor remained. The crude product slowly decom­ posed at 25°, evolving heat and gas; decomposition was arrested

(i.e. no gas evolution) at 0° or in cold other solution. The ether solution, after several days at 0-10°, served as source material for the preparation of potassium dinitromethane if adequate precautions were taken (see below),

l,l-Dinitro-2-propanol, prepared from potassium dinitromethane 20 and Bcetaldehyde is reported (84) to have the constants, n^ 1,449• d^^ 1.33. An attempt to distill a portion of the crude oxidation 4 product was unsuccessful; decomposition occurred even at low pres- — 58— sures (3~4 mm*), and, although atout 3O par cent was collected as distillate (74“90*/3“4 mm,), the product was instable and decom­ posed in a short time with evolution of gas and heat. If the crude oxidation product is hot distilled but merely stored for a time in vacuum to remove solvent, it has a refractive index of n^ 1,452.

c. Cleavage of 1 ,l-Dinitro-2-propanol, Potassium

Dinitromethane and Silver dinitromethane,

1, The freshly prepared nitration product (l.FO g., O.Ol mole based on 1 ,l-dinitro-2-propanol) in methanol (5 ml,) was added at 0-5° to aqueous potassium hydroxide (0.9 g., 83^ assay,

0,014 mole in 5 ml,). Light yellow, powdery potassium dinitro­ methane precipitated immediately and was filtered and washed with methanol and ether. Yield: 0.68 g. (47%). When heated in a cap­ illary tube the product exploded at 210®, Various samples of po­ tassium dinitromethane decomposed explosively between 2O4 and 210®; lit, 204—207® (78,79(80,90), Four milliliters of water were re­ quired to dissolve 0,10 g., the solubility reported (78) for potas­ sium dinitromethane in water is 1 part in 43 13®. When recry­ stallized from water, the product was obtained as pale yellow prisms, exploding at 210®. Anal, Calcd, for CHO^NgE: K, 27,13,

Found: K, 26,9, 26,5. The ultraviolet spectrograms (Beckman

Model DU quartz spectrophotometer) of aqueous solutions (lO”^j 4 x

10 ^ molar) of potassium dinitromethane show maximum absorption at 381 mu, log E max, 4 .32 ' ^ minimum at 290 mu, log E min, 3*00 ; -99- and a band at 247 log ®niax, 3*&3 * Corresponding data calculated from the ultraviolet spectrogram of potassium dinitromethane (87) are* an absorption maximum at 2'J,J00 cm,^ (361 mu) , log 4*35» & min­ imum at 290 mu, log E , 2,97 * and a band peaked at 247 mu, log Z BIB# fll&X# 3.60.

Silver dinitromethane (78) vas prepared from aqueous silver nitr­ ate (0*34 g*, 0.002 mole)in 1 ml.) and potassium dinitromethane (O.14 g., 0.001 mole in 4-5 ml.). Ihe yellow ( or yellow-green) salt (0.20 g., theoretical, 0,21 g.) was washed with methanol and ether. The decomposition point of this product varied with the preparation: dark­ ening ocourred from 100 to 130°, decomposition with gassing at 133°: or explosive decomposition at II7, 120°, Silver dinitromethane (78) is reported to explode at 135* Silver dinitromethane is photo­ sensitive and in twenty-four hours it decomposed extensively.

Anal. Calcd. for CHO^K^ig: Ig, 50.67. Found* $0.1, 5O.3.

2 . The crude nitration product was recovered from a three day old ether solution that had been stored at 0-$°, by removing ether under vacuum. It was dissolved in methanol (1.5O g., in 9 ml.) and added to methanolic potassium hydroxide (1.4 g.» 89I6 assay, in 5 ml.) at 0-10°. A yellow salt was filtered and washed with methanol and ether.

One to two minutes later it detonated.

3. A portion (1.9O g.) of the crude nitration prod­ uct that had been stored in ether, was dissolved in 90 per cent aqueous methanol (lO ml.) and added to aqueous potassium hydroxide -100-

(l*4 g, in 3 ml.) at 0°. The potassium dinitromethane which formed

(0,90 g., 30^Loverall yield from l-nitro-2-propanol) was identified hy (1) explosive decomposition at 20J, 210^: (2) when heated for one hour at 5^° (2-3mm.) a sample weighing 0,1172 g, lost O.OOO6 g.(0 ,3#);

a sample weighing 0,1092 g, heated for one hour at 92-95® in a drying

oven lost 3 oent of its weight (no further decrease after an addition­

al one and one-half hours at 90°): (3) the decomposition point,after heating,was 210®. Potassium 1 ,1-dinitro-2-propanol loses one mole of

acetaldehyda (30.6% of its weight) on heing heated at 90-95® snd

explodes at 205° (84)* silver derivative (yellow-green leaves;

photosensitive), prepared as before, explodes at 120®,

Anal. Calcd, for CHO^H^Agi Ag, 50,67. Pound* Ag, 50,7 » 50,7 *

After potassium dinitromethane had been isolated, the filtrate deposited a yellow salt. Precipitation of this second crop of crystals was complete on addition of alcohol, A minute after the solid had been collected and washed with methanol and ether, it suddenly decom­ posed, leaving no residue.

d. Cleavage of l,l-Dinitro-2-propanol to Acetaldehyda

A solution of the crude product (0 ,5 0 g,, 0,0033 mole based on l,l-dinitro-2-propanol) in methanol (5 ml,) was treated with

a solution of 2,4-&lnitrophenylhydrazine (0,70 g,, 0,0035 mole) in 2# hydrochloric acid. Precipitation of a yellow solid, needles, began

after a few minutes. After three days, 0,45 8» (61I0 of acetaldehyda

2,4-dinitrophenylhydrazona was obtained, m.p. 143-146°, After several -101- xeorystallizations from ethamol-irater and from obloroform-ethanol,the derivative melts at 155-156°. In authentic sample of acetaldehyde 2t4" o dinitrophenylhydrazone melts at 155“15® • X-ray diffraction patterns of both 2t4~ÈLinitrophenylhydrazones were identical.

Potassium dinitromethane, prepared from the crude nitration prod­ uct, did not yield a solid derivative on being treated with a solution of 2,4-dinitrophenylhydrazine in 2N hydrochloric acid. From these data it is concluded that the crude product from oxidative nitration is principally l,l-dinitro-2—propanol (50# yield, as determined by con­ version into potassium dinitromethane; 6l{6 by the determination of acetaldehyda as the 2,4-dinitrophenylhydrazone).

12.. Preparation of Potassium Dinitromethane from Fitromethano

Fitromethane (0,6l g., 0.01 mole) was dissolved slowly in agueous sodium hydroxide (0 «4 l g.* 0.01 mole) at 0-3°. After sodium nitrite (0«75 assay, 0.011 mole) had been added, the solution was poured into a mixture of aqueous silver nitrate (3.41 g., 0.02 mole in 50 ml.) and ether ( 25 ml.) at 0°, A idiite solid formed and rapidly turned black. Twenty minutes later, the black deposit was filtered and washed with ether. The ethereal extracts were washed with a small volume of water and then slowly added to a stirred solution of aqueous potassium hydroxide (l.O g. in 10 ml.). The ether layer lost its color, the alkaline layer became deep red, and pale yellow crystals of potassium dinitromethane separated (O.O5 g., 3%)» the crystals explode at 210°. The filtrate was diluted with alcohol and cooled for twenty- four hours; however no additional precipitate was obtained. -102-

An aqueous solution (2$ ml*) of nitromethane (O.Og mole), po­

tassium hydroxide (O.IO mole) and potassium nitrite (0 .0 $ mole) was

added to aqueous silver nitrate (O.IO mole in 2^ ml*) at 0-10°* After

ten minutes potassium chloride (0*03 mole) was added; the mixture was

0 heated to 90 and filtered* On cooling the filtrate, potassium nitrate

(l*8 g*) separated* More potassium nitrate (2*5 g*, 1*5 g*) crystall­

ized on addition of alcohol to the deep red filtrate, (Calculated

amount of potassium nitrate, 10*1 g*) It was concluded that potassi­

um dinitromethane (solubility in water, 1 part in 43 15° (jS)) was

not present in the remaining filtrate.

B* Attempted Preparation of Potassium Dinitromethane from Potassi­

um 2,2-Dinitroethanol

1. Preparation of Dipotassium l,l,3 ,3-%*tronitropropane

a* Dipotassium 1,1,3,3-Petranitropropane (As Suoh)(91)

Potassium 2 ,2-dinitroethanol (l.OO g*, 0*0052 mole) in water (lO ml*) was heated on a steam bath (70-80°) for three hours

(evaporated to approximately half volume): the odor of formaldehyde was

still noticeable* On cooling the deep red solution, an orange salt

(prisms) separated. On filtering and washing the solid with methanol,

dipotassium l«l,3»3"t*tranitropropane (0*41 g«« 475^) obtained*

Addition of the methanol-ether washings to the filtrate precipitated a

yellow-red solid (0*13 g«)! the solid was identified as unreacted

potassium 2,2-dinitroethanol by its solubility in water (completely

soluble in 2 ml*) and by its failure to form a water-insoluble silver

derivative (see preparation and properties of potassium 2,2-dinitro- -103- ethauol» Experimental, Section III)*

Sipotassinm l,l,3*3"t*tranitropropane is much less soluble in water

(O.IO g,, requires 11 ml*) than potassium dinitromethane (0*10 g* re­ quires 4 ml*)* Ifhen heated in a free flame the salt explodes; in a capillary tube, heating decomposes it over the range 210 to 25 O***

■Anal. Calcd* for potassium 2 ,2-dinitroethanol, C2E^0^E2^ ! for potassium dinitromethane, CEO^E^E; and for dipotassium 1,1 ,3 ,3-tetra- nitropropane, ^'^2^8^4^2 ' 22»4 5 : 27*13; 20*04, respectively.

Found» K, 25*5»

The orange colored product, when recrystallized from water, forms yellow prisms, decomposing between 220 and 2 $0°* During storage the yellow crystals gradually acquire a rod tinge. The product (recrystalliz­ ed) yields a silver derivative identical with that obtained from the orange colored salt (see the next paragraph).

b* Disilver l#l,3«3"^*tranltropropane

Disilver l,l,3 »3-tetranitropropane was prepared from aqueous solutions of the dipotassium salt (O.IO g*, 0*00033 mole in

15 ml.) and silver nitrate (0*55 g*, 0*0032 mole in 2 ml*)* After five minutes at $-10°, the product was filtered and washed with methanol and ether. The disilver salt (0*13 g*, 1007^) was obtained as close- packed, fluffy yellow needles* The freshly prepared product decomposes at 110-150°. Unlike silver dinitromethane, it cannot be recrystallized from water without extensive decomposition, and it is not photosensitive; however, the samples did darken slightly on long storage* The silver content of freshly prepared samples was determined as before* — 104-

Anal. Calcd. for C^E20gN^ig2 * Ag, 49»28. Found; Ag, 49*3 » 49*1 *

A solution of potassium 2 ,2-dinitroethanol (0 .8y g«, 0.003 mole) in water (13 ml.) was heated to 45°(t&th temperature) under vacuum.

Tihen approximately one-third of the water had been removed, yellow

crystals separated. Water (lO ml.) was added and distillation was

resumed at 63° until 7 ml. of solution remained. This solution was

rapidly heated to 80** at atmospheric pressure; after cooling to 0°, potassium 2 ,2-dinitroethanol crystallized (0.49 g*)* ^be salt was wash­

ed with methanol and ether* the washings were added to tha filtrate,

whereupon potassium 2 ,2-dinitroethanol (0.22 g.) separated. When the

remaining filtrate was evaporated in vacuo, impure (red) potassium

2 ,2-dinitroethanol (O.ll g.) remained. The product (O.82 g., 94?^ re­

covery) was identified as the initial material by its solubility in

water (O.IO g. soluble in 1 ml.), its behavior toward silver nitrate

(producing a silver mirror rather than a water insoluble salt),and by

its decomposition range (123-I3O to 2^0°).

Potassium 2 ,2-dinitroethanol (I.30 g., O.OO86 mole) and potassi­

um hydroxide (O.60 g., 83?^ assay, O.OO9I mole) in water ( 6 ml.) were heated at 90-93° for twenty minutes. On cooling the deep yellow

solution to 0°, yellow needles of potassium 2,2-dinitroethanol (1.16

g.) separated. The filtrate was washed with methanol and ether.

Addition of the methanol-ether washings to the filtrate resulted in

the precipitation of potassium 2 ,2-dinitroethanol (O.16 g.); similarly,

additional salt (0,13 g,) was precipitated from the residual filtrate. -105“

The total weight of product recovered was 1*45 6* the theoretic­ al weights expected if potassium dinitromethane and dipotassium 1,1,3 *3 " tetranitropropane were produced are 1 ,2 4 g, and 1.29 g«, respectively.

Each fraction of potassium 2,2-dinitroethanol was identified hy the extent of its solubility in water, its failure to yield a silver deriva­ tive, and the temperature at which decomposition was initiated (l25 ~

130°)* The first crop of crystals was analyzed for potassium.

Anal. Calcd. for 22.45. Found! K, 22 .97, 22.29.

C. Preparation of Dipotassium l,l,3,3“2atranitropropane from

Potassium Dinitromethane and Potassium 2,2-Dinitroethanol.

A solution of potassium dinitromethane (O.1 4 g., 0,0001 mole) and potassium 2,2-dinitroethanol (0 .I7 g., 0.001 mole) in water (lO ml., representing a 100^ excess over that amount needed for complete solubil­ ity) was heated on a steam bath (85-9O ) for two and one-half hours.

After 16 hours at 0-10°, the mixture was filtered to yield dipotassium l»l*3*3"tetranitropropane (0.12 g., yellow prisms). After the filtrate was heated for two hours on a steam bath and cooled, additional dipotassi­ um l»l%3*3"t*tranitropropane (o.ll g.) was obtained. Total yield! O.23 g.

(77#). Each of the two fractions (before or after being reorystallized from water) was found to be identical with potassium l,l,3*3"tetra- nitropropane obtained directly from potassium 2,2-dinitroethanol by observing the solubility in water (O.ll g. required 13 ml. for solution) and the decomposition range (initial at 2l8-220°, final at 245 -250 °).

Anal. Calcd. for ^^^2^8^4^2^ ^ * 12.00; H, 0.675 H, 18.66. Found: -io6-

C, 11.77; H, 0.67; N, 18.11,

On standing, tho yellow prisms gradually acquire an orange-red tinge. The ultraviolet absorption spectrogram (Beckman Model DU

quartz spectrophotometer; 1 cm, cell) of an aqueous solution of di­ potassium 1,1,3,3-tetranitropropane ( 4 X 10"5 Bolar) shows an absorption maximum at ^67 mu, log 4»27' ® minimum at 297 log 3*15' and a band at 222 mu, log 4*04.

Disilver 1 ,1 ,3 ,3-totranitropropane was prepared in quantitative yield.

AnalŸ Calcd. for G^H20gE^Agg* 0 , 8.23; H, O.46; Ag, 49*28.

Found: C, 7.61; H, 0 .42 ; Ag, 49 *1*

* The silver determination was performed on freshly prepared material. Approximately ten days elapsed before determinations of carbon and hydrogen were made. During this period, the salt underwent slight decomposition, as evidenced by darkening of the yellow needles.

D. The Effect of Various Oxidizing Agents on the 2-Propanenitron-

ate-Hitrite Ion System.

1 . Oxidation With Mercuric Ion. Preparation of 2 ,2-Dlnltro-

propane

Aqueous mercuric nitrate hemihydrate (17,0 g., O.O3 mole in 25 ml. water) was added rapidly to a stirred solution of 2-nltro- propane (4*45 g*» O.O5 mole), sodium hydroxide (2,25 S*)» sodinua nitrite

(3*75 g*» 97^ assay) and water (25 ml.). After one-half hour the mixture was heated to its boiling point (a blue condensate of propyl pseudonitrole formed on the condenser) and allowed to cool. The mixture was extract­ - 107- ed irlth ether and filtered. She ether layer from the filtrate was wash­ ed with saturated sodium chloride solution and then distilled, leaving an oily blue residue. The crude residue was dissolved in boiling petrol­ eum ether (60-63°) and then crystallized to yield a low melting, waxy solid (2.3 g», pale blue in color). After being triturated with hot thirty per cent hydrogen peroxide for a few minutes, the solid was re- crystallized from petroleum ether, yielding 2,2-dinitropropane (1.7O g., 23%), m.p. 32-53°» melting point of the product was not altered by admixture with an authentic sample of 2,2-dinitropropane.

2 . Oxidation With Persulfate Ion. Preparation of 2 ,3-Dimethyl-

2,3-dinitrobut ane.

2-Witropropone (0 .8g g., 0.01 mole) and sodium nitrite

(0.70 g., 0.01 mole) were dissolved in dilute sodium hydroxide (approx­ imately 3 ml» of a 10^ aqueous solution). The solution was cooled to

0° and aqueous ammonium persulfate (4,6 g., 0.02 mole, in 10 ml.) was

added. After three hours, 2 ,3-4 imethyl-2 ,3-dinltrobutano (O.ll g., 12.3

%) was obtained, m.p. 210-211° (crystals from methanol).

In a previous experiment, persulfate ion oxidized nitrite slowly at 0° in slightly alkaline medium; when the mixture became acidic

(after about ten minutes) oxides of nitrbgen were rapidly evolved.

3« Oxidation With Hydrogen Peroxide. Preparation of 2,3-Di­

methyl— 2 ,3-iinitrobutane

2-Hitropropane (0.S9 g.» 0.01 mole) and sodium nitrite

(0.75 g** 97^ assay, 0.01 mole) ware dissolved in aqueous sodium - i o 8 - hydroxido (O.^O g#, 0.012 mole in 4 ml# water). Hydrogen peroxide (13 ml. of ^0% aqueous solution) was added at 0°. Dilute acetic acid was added until a faint green color appeared, ^he appearance of a blue- green color was rapidly followed by formation of a white solid. After six hours in melting ice, the mixture was filtered to yield 2 ,3~dimethyl-

2,3-dinltrobutane (0.10 g,, ll?t) containing traces of 2-nitroso“2-nitro- propane (odor, faint blue color in chloroform). After one recrystalliza­ tion from methanol, 2,3-dimethyl—2,3-dinitrobutane melts at 210-211°

(Cf. the oxidation with hydrogen peroxide of the mixed salts obtained from the action of nitric oxide on the sodium salts of secondary nitroalkanes, Sxperimental, Section II).

4# Oxidation With Cupric Ammonium Complex Ion. Preparation

of 2,3“Dimethyl-2,3-dinitrobutane.

A solution of cupric chloride dihydrate (8.6 g., O.O3 mole), water (30 ml.) and concentrated ammonia ( 5-10 ml.) was added to a mixture of 2-nitropropane (4#45 g#, O.O5 mole), sodium hydroxide

(2.3 g # , 0.06 mole) smi sodium nitrite ( 3*8 g., assay, O.06 mole) in water (30 ml.). Ho observable change occurred in one to three hours at 25°. When the blue homogeneous mixture was refluxed, a red-brown solid was deposited on the walls of the flask.,After one hour (long­ er periods of refluxing did not alter the result), the mixture was cooled and extracted with ether; evaporation of the ethereal extract yields 2,3-dimethyl-2,3~dinitrobutane (O.37 g«* 8^i), (Cf. the oxida­ tion of sodium 2-nitropropane with cupric ion. Experimental, Section IV.) - 109-

Ylhen the previous procedure was repeated, hut with an excess of sodium hydroxide (up to 100# excess), the cupric ion-ammouia solution was converted into cupric oxide at the hoiling point of tho mixture.

Ho products soluble in water and volatile with steam are obtained.

Oxidation with Fehling's solution was unsuccessful. Neither 2,2- dinitropropane nor 2,3-dimethyl-2,3“dinitrobutane was obtained when an aqueous solution of 2-nitropropane, sodium nitrite, sodium hydroxide, copper sulfate and sodium-potassium tartrate was steam distilled,

Beaction of sodium 2-nitropropane and sodium nitrite with aqueous cupric chloride results in the formation of 2-nitro-2-nitrosopropane,

9. Oxidation With Cuprous Salts, Formation of Water Soluble

Products,

When ammoniacal cuprous chloride is the oxidant, under conditions identical with those described for the oxidation of aqueous sodium-2-nitropropane and sodium nitrite with ammoniacal cupric chloride, neither 2,2-dinitropropane nor 2,3-diaethyl-2,3-dinitrobutane is obtain­ ed, The cuprous salt is converted to a blue-green solid, insoluble im water; the distillate is homogeneous.

When an aqueous slurry of cuprous acetate and an alkaline solution of 2-nitropropane and sodium nitrite is steam distilled, the distillate is homogeneous,

III, Summary

A new reaction has been described, by which salts of primary and secondary nitro compounds are converted into the corresponding gem - 110- dlnitro oompounds by tho aotlon of silver nitrate and soluble inorganic nitrites in an alkaline or neutral aqueous medium. The reaction pro­ ceeds in excellent yield, A series of gem dinitro compounds was pre­ pared that includes primary and secondary gem dinitroalkanes and primary and secondary gem dinitro alcohols in which the hydroxyl group is adjacent to the nitro groups. A dinitro alcohol is cleaved by alcohol­ ic potassium hydroxide into a potassium 1,1-dinitroalkane and a carbonyl compound. By the latter reaction potassium dinitromethane is prepared in 47“50 per cent yield, whereas the nitration of nitro­ methane itself gives potassium dinitromethane in low yield.

Attempts to prepare potassium dinitromethane by cleaving potassi­ um 2,2-dinitroethanol into potassium dinitromethane and formaldehyde were unsuccessful; potassium 2,2-dinitroethanol is converted into l#l»3*3"t*tranitropropane on heating its aqueous solution, Reaction of potassium dinitromethane with potassium 2,2-dinitroethanol also yields dipotassium l,l,3»3"tGtranitropropane.

A series of oxidizing agents was investigated in an effort to find a substituto for silver ion in the nitration. Mercuric nitrate also oxidizes a mixture of sodium 2-nitropropane and sodium nitrite to 2,2-dinitropropane. However, the action of mercuric nitrate is less easy to control than that of silver nitrate; in many experiments 2,3** dimethyl-2,3-dinitrobutane is the main product. Other oxidizing agents

(salts of persulfuric acid, acidified hydrogen peroxide, or cupric ammonium complex ion) effect the coupling of sodium 2-nitropropane to 2,3'*Aimethyl-2,3'~dinitrobutane ; cuprous salts merely degrade a mixture of sodium 2-nitropropane and sodium nitrite to water-soluble - 111- products. - 112-

SECTION IV

THE OXIDATIVE COUPLING OF SALTS OF NITROALKANES TO VICINAL DINITRO

COMPOUNDS

I* Discussion

A, Introduction

In this research it has been found that sodium 2-nitropropane is in part converted to 2,3“dimethyl-2,3“dinitrobutane by reaction with either nitryl chloride, nitric and sulfuric acids, or methyl nitrate (Section l). Also the reaction of nitric oxide with sodium nitrocyolohexane results in formation of some 1,1'-dinitrobicyclo- hexyl (Section II). Oxidation of mixtures of sodium secondary nitro­ alkanes and sodium nitrite with acidified hydrogen peroxide, per­ sulfates, mercuric or tetrammine copper (ll) ion, yields vicinal dinitro compounds (Sections II and III). Oxidative nitration (with - silver ion. Section III) of potassium g-nitrofluorene produces

9 ,9-dinitrofluorene, fluorenone, and 9 »9'"&ioitro-9 ,9'"bifluorene.

Since such a wide variety of oxidizing agents produces dimer­ ic compounds, it appears that nitroalkyl radicals are formed, then conplod to yield vicinal dinitro compounds. Nenitzescu and Isacescu

(95) Rave proposed nitroalkyl radicals as intermediates in the various dimerization reactions of aci-9-nitrofluoreno end phenyl- nitromethane and their derivatives. Since reactions between salts of nitro compounds and halonitro compounds (ly, 95) may occur by - 115-

'bimolecular ionlo processes, an investigation of the action of oxi­ dizing agents on salts of nitroalkanes vas made in an effort to de­ termine whether or not nitroalkyl radicals are formed.

B, Historical Review of Coupling Reactions.

A variety of methods for the coupling of nitroalkanes to form vicinal dinitro compounds have heen reported. The action of silver on 2-hromo-2-nitropropane, 2-hromo-2-nitrohutane (40), 1-hromo-l- nitrocyclohexane (oO) and phenylhromonitromethane (lOl), respec­ tively, has heen used to prepare 2,3~dimethyl-2,3~dinitrohutane,

3,4”

2,3-dimethyl-2,3-dinitrohutane. It has been observed (60) that

1,1'-dinitrobicyolohexyl is produced, along with nitrocyolohexane, by liquid phase nitration of cyclohexane.

The most extensive study of reactions of arylalkylnitroalkanes which leeid to coupled products has been made by Nenitzescu and

Isacescu (95)* Their experiments illustrate that potassium 9-nitro*- fluorene and sodium phenylnitromethane may be converted into "dimers" by a number of methods; thermal decomposition of the iodo derivatives of these nitro compounds yields 9;9'"&iuitro-9,9'-bifl*orene and - 114- l«2-dlnitTo-l,2-dipheayIetheuae, respectively; reaction of $-iodo-

(or bromo-) g-nitroflnorene irith potassium 9-nitrofluorene or vith sodium phenylnitromethane produces only 9,9'-dinitro-g,9 '-bifluorene; similarly, phenyliodonitromethana and sodium phenylnitromethane yield l,2-dinitro-l,2-diphenylehtane; aoi-9-nitrofluorene dispro- portionates in boiling ethanol to 9,9'-dinitro9,9'-bifluorene and fluorenone oxime; electrolysis of potassium 9"aitrofluorene also gives the vicinal dinitro compound.

The reaction of halonitro compounds has been used for prepara­ tion of aliphatic vioinal dinltroalkanes (17, 102). In general this reaction is restricted to salts of secondary nitroalkanes and secondary halonltroalkanes. Although small amounts of 2,3~ dimethyl-2,3-dlnitrobutane are obtained from mixtures made either from sodium 2-nitropropane with 1-bromo-l-nitroethane or from sodium nitroethane with 2-bromo-2-nitropropane (IO2) this result is proba­ bly brought about by an interchange of halogen which forms the mis­ sing secondary intermediate in the mixture. Similarly it has also been found that sodium 2-nitropropane is converted into 2,3-dimethyl-

2,3-dinltrobutane by either tetraiodomethane, diethyl bromomalonate malonate, or 1,1-dichloro-l^nitroethane (103); 2-bromo-2-nitropro- pane and sodium diethyl malonate yield 2,3-dimethyl-2,3-dinitro- butane (104).

Agpieous solutions of salts of nitro compounds have been oxi­ dized cleetrolytically (18, 95, 102, IO5, I06, IO7) to vicinal dinitro compounds; it has also been found (18) that in the anodic - 115-

oxldation of sodium 2-nitropropaae, acetone and nitric oxide are produced along with 2,3-dimethyl-2,3-dinitrohntane« Eecently it has heen shown that salts of secondary nitro compounds may he dimer- ized hy the action of air or hydrogen peroxide (58).

C« Discussion of Results

1* Mitro Compounds and Oxidizing Agents Selected for Investi­ gation.

Nitro compounds chosen for study in the oxidative coupling reac­ tion were nitroethane, 2-nitropropane, 2-nitrohutane, nitrocyclohex- ane, 1,1-dinitroethane and nitroform. Solutions of the nitro com­ pounds in an amount of aqueous alkali necessary for neutralization, or aqueous solutions of the sodium salts of the nitro compounds, are treated at 0-25° with aqueous solutions of one of the following oxi­ dizing agents; ammonium persulfate, sodium persulfate, acidified hydrogen peroxide, potassium ferricyanide, ferric chloride, ammoniacal cupric chloride or ammoniacal cuprous chloride.

Vicinal polynitroalkanes were obtained only from secondary nitroalkanes; these products are 2,3‘~

3,4-&lm=thyl-3 ,4-d,im@thylhexane and l,l'-dinitrobicyclohexyl. Ac­ companying the vicinal dinitro compound is the ketone corresponding in structure to the nitroalkane from which it is derived; sometimes the ketone is the major or sole product. She effective coupling reagents were salts of persulfurio aoid, acidified hydrogen peroxide, potassium ferricyanide and ammoniacal cupric chloride.

A, Oxidation of Salts of Nitroalkanes With Persulfates. - 116-

NitToethane, the only primary nitroalkane investigated, under­ goes a complex reaction with persulfate ion in alkaline medium to yield, ultimately, 3>4 ,5~trimethylisoxazole. Hot direct evidence was obtained for the formation of 2,3~dinitrobutane,

Failure to obtain 2,3-dinitrobutane in these experiments cannot be used as evidence for or against the existence of the nitroethyl radical since vicinal dinitro compounds in which the nitro groups are attached to secondary carbon atoms yield 3,4 i5""trisubstituted isoxazoles in alkali (94,95)* !Che reaction is further complicated because nitroethane is converted into 3*4,5-trimethylisoxazole by alkali (36). However, anodic oxidation of aqueous solutions of sodium nitroethane yields 2,3-dinitrobutahe, 2-nitro-2-butene, and

3-nitro-2-butanol (18, IO7),

Oxidation of alkaline solutions of secondary nitroalkanes with persulfate ion occurs readily to yield ketones and vicinal dinitro alkanes (Equations 1 and 2); 2,3“dimethyl-2,3~dinitrobutane (26-58#),

3,4“dimethyl-3,4**'iiaitrohexano (37#) and l,l'-dinitrobicyolohezyl

(26—30#) were prepared from salts of 2-nitropropane, 2-nitrobutano and nitroeyclohexane,

1. 2BH‘C=H0“ * SgOg- --- » aB'C(HOg)C(HOg)BH' + 2 S0°

2. HR’CsHOg + SiSgOg + ZHgO --- » BR'C=0 + 4H'^ + HO^ + 480=

The reaction proceeds rapidly at 5-10° (crystalline product begins tp separate almost immediately upon mixing the reactants) and is essen­ tially complete in four hours. Highest yields were obtained with - 117- ooncentrated reagents; excess persulfate did not signifleantIjr in­ crease the efficiency of the reaction. The pH of the mixtures de­ creases, and unless alkali is added during the oxidation the solu­ tion finally becomes acidic. Under acidic conditions the nitroal­ kane is partially regenerated or decomposed into ketone and pseudo- nitrole. In stronglytlkaline solutions the dimerization process occurs slowly and the yields of coupled product are lower theui aver­ age. However, >rhen solutions of sodium 2-nitropropane and sodium persulfate are buffered with yield of 2,3*'difflathyl-

2.3-di ni trobut ane consistently range from ^-58 per cent.

The reactions accompanying the coupling process produce ketone, hydrogen ion and presumably nitrite ion (which is also oxidized by persulfate ion in alkaline solution, Zgnation 2). Both 2,3-dimethyl-

2.3-dinltrobutane (20#) and acetone (8.5#, as its 2,4~dinitrophenyl- hydrazone) were isolated from alkaline mixtures of sodium 2-nitro­ propane and sodium persulfate* The presence of unreacted sodium 2- nitropropano was detected by acidifying the mixture; 2-nitroso-2- nitropropane was isolated in 11 percent yield. Under identical con­ ditions an alkaline solution of 2-nitropropane in the absence of oxidizing agents yields no acetone. These results indicate that the ketone is formed in alkaline medium by an oxidation reaction.

Attempts to prepare 2,2,3»3"tetranitrobutane and hexanitro- ethane by reaction of sodium 1,1-dinitroethane and sodium trinitro- methane with persulfates were unsuccessful; both substances are decomposed to water-soluble products in 24 hours. - 118-

"b. Reaction of Sodium 2-Nitropropane With Hydrogen Peroxide

2,3-Dinethyl-2,3-diaitrohutane is formed very slowly and in poor yields hy reaction of sodium 2-nitropropane with excess hydrogen per­ oxide; upon the addition of a few drops of dilute acetic or hydro­ chloric acids the coupled product begins to separate rapidly* The yields from these reactions are of the same order (IO-15#) as those obtained by oxidation of the mixed salts (itl) of sodium 2-nitropro­ pane and sodium nitrite with hydrogen peroxide (Sections II and III),

In all of the experiments with hydrogen peroxide the coupling reac­ tion had to be initiated by acids; alkaline solutions of the reac­ tants were still homogeneous after twenty four hours* Evidently the oxidative coupling with hydrogen peroxide is more dependent on pH than the reaction involving persulfate ions*

T T + 3* 2RRC=H02 + HgOg — » HHC(H02)C(H02)RR + 20H“

c* Reaction of Sodium 2-Hltropropane With Potassium Perricyanide

Sodium 2-nitropropane reacts with potassium ferricyanide in slightly alkaline solution to form 2 ,3"^^'^ ethyl-2,3"^^ ^i tr obut ane

(15?^) • Separation of the coupled product was not as rapid as when persulfate ion was the oxidant* The yield of coupled product was not improved by reaction times longer than five hours at 0-3° nor at higher temperatures (up to 30°)*

4. 2(CH^)2C=H02 + 2/Pe(CN)g/ --- > (CH^)2C(H02)C(H02) (CH^)^ + 2/Pe(CN)J* - 1 1 9 -

d* Reaction of Sodium 2-Hitropropane idth Other Oxidizing Agents

It is reported in Section III that an ammoniacal solution of

cupric chloride slowly oxidizes a mixture of sodium 2-nitropropane and sodium nitrite to 2,^-dimethyl-2,1-dinitrotutane (8.3^), The reaction was repeated in the absence-of nitrite ion. As before, re­ duction of the ouprio salt to copper required heating the mixture to boiling; the yield of 2:,^-dimethyl-2,]^-dinitrobutane is 4»5 par cent. Like silver ion (33) and mercuric ion (Section III), cupric ion (or its ammoniated complex) is a cationic oxidizing agent. Other cations capable of undergoing change to a lower valence number should also be able to effect the coupling of nitroalkanes to vicinal dinir tro compounds.

Ferric chloride and aqueous sodium 2-nitropropane produce a deep red solution (46) which remains homogeneous for at least 24 hours at 23-30°; on heating it decomposes into 2-nitroso-2-nitro- propane and acetone. Failure of ferric chloride to effect the coup­ ling reaction is attributed to the stability of the complex,

/ (CE^)gC=R02/^e, at moderate temperatures. Neither is any coupled product obtained when mixtures of sodium 2-nitropropane and Fehling's solution or ammoniacal cupric chloride are distilled.

It is reported (108) that sodium nitroeyclohexane is oxidized with potassium permanganate to give cyclohexaneone in excellent yield; potassium permanganate or air (27) degrade potassium 3-nitrofluorene to fluorenone and potassium nitrite. Similarly, in this research ”• 120-

(Seotions II and III) it vas found that only acetone results from the oxidation of the mixed salt of sodium 2-nitropropane and sodium nitrite either with potassium permanganate or with alkaline sodium hypohromite.

2. Correlation of Data

There are two possible reactions which may account for the coupling process; (l) Oxidation of salt of nitroalkane hy a trans­ fer of one electron to the oxidant to form a free nitroalkyl radical

(18, 95)» Dimerization of the nitroalkyl radical yields the vicinal dinitro compound. (2) Coordination of the oxidant with two anions of the salt of the nitro compound to form a complex ion which decom­ poses directly to produce a coupled product hy an internal process.

Oxidation involving the transfer of one electron is represen­ ted (18, 95) as;

5a BBC=NOg + Oxidant ' HEC-NO^ + Oxidant”^

5h BBCsHOg + eT..Oxidant — * SHC-NO^ + H—Oxidant 6 2 aac-BOg-- » BacfNOgicfNOgyaE

7 EEC-EOg — — * RHC=0 + "BO

Oxidants that may initiate these reactions (Equations w e per­ sulfate ion, hydrogen peroxide, ferricyanide ion, the anode in an electrolytic cell, oxygen, silver ion, and tetrammine copper (ll) lOW.

It is likly that coupling of nitroalkyl radicals (pairing of electrons, Equation 5) is favored at lower temperatures and decompo­ sition (splitting of a hond. Equation 7) predominated at higher - 121- temperatures. (Best yields of coupled product were obtained below

20°).

In this research it was found that in oxidation with persulfate ion or with hydrogen peroxide there is an optimum pH range (close to neutral) for the coupling of salts of secondary nitroalkanes to vicinal dinitro compounds. Since, in alkaline solution, both the reductant and the oxidant (persulfate ion, hydrogen peroxide, ferrky- anide ion) are negatively charged, the slowness of the reaction can in part be attributed to repulsion between particles bearing like charges. Another factor, one idiich cannot be completely evaluated, is the difference in the oxidizing potential of an oxidizing agent in alkaline and acid medium. However, the acid-initiated coupling with hydrogen peroxide and the buffering effect of sodium acetate on the oxidation with persulfate suggests that the efficiency of the coupling process depends on the formation of a hydrogen bond between the reactants (Equation 5^)• Pearson and Evans (l8) showed that in electrolysis of an alkaline solution of 2-nitropropane, 2,3-dimethyl-

2,3-dinitrobutane, acetone, and nitric oxide appear at the anode.

One-electron changes were postulated (18) to account for formation of the products (Equations 5a, 6 and %). Still the possibility exists that the salt of the nitroalkane can be degraded to its car­ bonyl derivative by oxidation involving two-electron changes. One such sequence (equations 8 and 9) would also yield aldehyde or ketone, hydrogen ion, and nitrous acid# - 122-

8, HHC=HO" + HgO ^ RRC(OH)NHo “

9, HHC(0H)1IH0“ . , RKC(0H)N02H---- » HBC=0 + H ’*' + HNOg

It is known (2%, 108) that ketones are produced in good yields hy

oxidation of salts of nitro compounds with potassium permanganate.

Perhaps oxygenated molecules that lose oxygen in going to the reduced

state degrade the salt of the nitroalkane to ketone hy direct trans­

fer of oxygen. It an oxidizing agent can he reduced only hy a two-

electron transfer, then in order to effect the coupling reaction the

oxidant must coordinate with two anions of the salt of the nitro

compound. However, when the oxidant is negatively charged (e.g. permanganate, hypohromite ion) the probability of its colliding with

two anions simultaneously is small, A positively charged oxidant

should he a more effective coupling agent.

Results from this research support the postulate (l8, 95) that nitroalkyl radicals are involved in the coupling process. Oxidation

of the 2-propanenitronate ion to 2,3~diaethyl-2,3~dinitrohutane hy ferricyanide ion is readily interprétable only if it is assumed that ferricyanide ion is reduced to ferrocyanide ion; this requires only

one electron (1O9). Coupling effected hy oxygen (98) can he similar­

ly explained, (in this investigation it was observed that the dry

sodium salt of 2-nitropropane is slowly oxidized to 2,3-dimethyl-

2,3-dimethyl-2,3-dinitrohutane in storage.) Since 2,3“dimethyl-2,3-

dinitrohutane has heen produced* hy decomposing acetyl peroxide in

*Private communication from Harold Barker, The Ohio State University. - 125-

2-nitropropane, it is probable that secondary nitroalkyl radicals are sufficiently long-lived to couple to form vicinal dinitroalkanes.

The postulate that nitroalkane ion can function as a reducing agent by loss of one electron seems logical, since reactions involv­ ing transfer of pne electron are rapid and have low energies of act­ ivation (no). Liberated by such a process, the secondary nitroalkyl radical should be relatively stable because of resonance interaction of the odd electron with the strongly electronegative and unsatura­ ted nitro group. Since the nitro group is more electronegative than the carbethozy group the radical (CH^ÏgCNOg should be more stable and hence less reactive than the radical (CH^jgCCOgCgE^ (derived from ethyl isobutyrate) which is reported (ill) to dimerize to diethyl tetramethylsuccinate. Oxidation of an alkaline solution of ethyl dimethylmalonate, ^O^CC{CH^)with persulfate ion also yields a dimeric product, diethyl tetramethylsuccinate, C2H^02CC(CH^)^C(CH^)^-

(112), It follows that the secondary nitroalkyl radical is not reactive enough to initiate or continue a free radical chain reac­ tion (e.g. peroxide-initiated polymerization of an olefin (115)).

Since active free radicals are oxidizing agents, the nitroalkane ion

(reacting by Equation $) should function as an inhibitor of free rad­ ical chain processes. Likewise, nitroalkyl radical (or its decom­ position product, nitric oxide) can behave as retarders in the same way that nitric oxide and some nitro compounds are known to do (II3)

114). - 124-

II. Experimental

A. Oxidative Coupling (Dimerization) of Secondary Nitroalkanes

1* Preparation of 2,3“^nietliyl-2 ,3“dinitrobutane Using

Various Oxidizing Agents.

a. Coupling by Persulfate Ion,

Sodium 2-nitropropane (0«55 g., O.OO^ mole) was dis­ solved in water ($ ml,). An aqueous solution of ammonium persul­ fate (l,b g,, 0,004 mole in 2 ml,) was added, A white crystalline solid began to precipitate almost immediately and heat was evolved.

The mixture was allowed to stand at 0-10® for 12 hours and then fil­ tered; the white solid, 2,3~dimethyl-2,3~dinitrobutane, was washed with water and dried in air (0,25 g,, 57#), m,p, 206-209°, when re- crystalligsd from methanol, m,p, and m.p, mixed with an authentic sample, 210-211° (17),

Alkaline* solutions of 2-nitropropane (0,89 g., 0,01 mole, dis­ solved in 3*~3 ml, of 10^ sodium hydroxide) and ammonium persulfate**

*Concentrations employed were 10% and 23# sodium hydroxide. The yields of product were slightly better in the more concentrated sol­ utions,

**Sodium persulfate may be used. Aqueous solutions of the commercial product are acidic and were made slightly alkaline before being added to the mixture. Excess of alkali is to be avoided. Best yields (30-

36%) were obtained if the amount of alkali used was just sufficient to prevent formation of pseudonitrole when the reagents were mixed.

The 111 mole ratio of reactants corresponds to 100 per cent excess of persulfate ion; when the ratio of nitro compound to persulfate ion was 2ll only slightly lower yields than average were obtained. - 125- (2.30 g., 0.001 mole, in 4 ml. of water containing a few drops ot dilute alkali) were mixed at 0-10®. After 12 hours, 2,3-dimethyl-

2,3-dinitrohutane was obtained in yields of 26-33?»; average 359^»

A solution of ammonium persulfate (2.3O g., 0.91 mole) and sodium acetate (l.O g., 0.012 mole) in water (b ml.) was added at Q-10® to

2-nitropropane (0.89 g., 0.01 mole) dissolved in 10 per cent sodium hydroxide (3 ml.). After 2-6 hours at O-lO®, filtration of the mix­ ture yields 2,3-^.imethyl-2,3-dinitrobutane (0.48-O.3I g., 34-58?^), m.p. 209-210°,

Four experiments were conducted under similar conditions (O-IO®,

6 hours), but with the quantity of sodium acetate varied from 1.0 to 4*0 g. Yields of coupled product ranged from 3Y-^8?^ per cent,

i. Formation of Acetone By a Side-Reaction.

2-Nitropropane (1.76 g., 0,02 mole), dissolved in sodium hy­ droxide (12 ml. of 10?6 solution) and cooled to O-3®, was treated with ammonium persulfate (2,30 g,, 0.01 mole) in sodium hydroxide

(11 ml. of 1# solution). After 24 hours at 0-10° the mixture was still alkaline. Filtration yielded 2,3-dimethyl-2,3-dinitrobutane, which was washed with water and dried in air (yield, 0,33 g,, 20#; m.p. 208-209®). An aliquot (lO ml.) from the combined filtrate and washings (60 ml.) on acidification with 6R acetic acid yields 2- nitroso-2-nitropropane (0.02 g,, 11#), m.p, 75“7^° (blue malt, dec.)

(45)* I!he remainder of the filtrate-washings was distilled. Acetone was isolated from the first 3"10 ml, of distillate as its 2,4~Aini- - 126- trophenÿlhyàraïone, (0»40 g., , m.p. 125-126°; the melting point of the derivative mixed with an authentic sample of acetone

2,4-dinitrophenylhydrazono vrao unaltered.

Under the same conditions, the distillate obtained from an al­ kaline solution of 2-nitropropane that had heen stored for 24 hours at 25-30® produced only a faint precipitate with 2,4-dinitrophonyl- hydrazine hydrochloride.

h. Coupling by Hydrogen Peroxide

2-Hitropropane (0,89 g., 0.01 mole) was dissolved in alkali

(0,50 g. of sodium hydroxide in 4 ml. of water). Aqueous hydrogen peroxide (3O ml. of 30?^) was added at 0°, After 24 hours at 0-20°, the solution was still homogeneous.

A solution of 2-nitropropane in alkali and hydrogen peroxide was prepared as before and treated at 0® with dilute acetic acid

(2-3 drops of 6H acid) until a faint green color was produced. The mixture was allowed to stand in a bath of melting ice. During the first hour no solid separated. After 6 hours, white crystalline

2,3-dimethyl-2,3-dinitrobutano (0.09 g.) was obtained; additional pure product (O.Ol g.) separated from the filtrate at the end of 24 hours. Yield* 0.10 g. (ll%), m.p. 209-210°; a melting point of a mixture of product and authentic dimer was unaltered.

In Experimental, Section III, the 2-propanenitronate-nitrite ion system was oxidized with acidified hydrogen peroxide under condi­ tions identical with those just described for the coupling of 2-nitro- " 127- propane, 2,]|-Dimethyl-2,3-&iaitrotutane was obtained in 11 per cent yield; however it was formed more rapidly in the presence of nitrous aoid,

c. Coupling by Ferricyanide Ion,

Aqueous potassium ferricyanide (3«3® 8»» 0*01 mole in 10 ml,) was added to an alkaline solution of 2-nitropropane (0,8g g., 0,01 mole, in 4 ml, of 10% sodium hydroxide solution) at 0-3°* Slowly, the mixture darkened and solid separated. After 3**3 hours the mix­ ture was filtered; recrystallization of the solid from methanol yields 2,3“dimethyl-2,3-dinitrobutane (0,13 g., 13%)» m.p. and m.p, after mixing with an authentic sample, 210-211°, The yield was not improved by longer reaction time at 0-3°, nor at higher temperatures,

(up to 30°)*

d. Coupling by Ammoniacal Cupric Chloride

2-Nitropropane (4*3 g** 0,03 mole) was dissolved in aqueous so­ dium hydroxide* (2,3 g., 0,037 mole in 20 ml,) of cupric chloride dihydrate (4,23 g,, 0,023 mole) dissolved in 3^ ammonium hydroxide

(15 ml,) was added, and the mixture was left at 23-30° for 12 hours.

The homogeneous blue solution was refluxed for 2 hours (a red coating of copper deposited on the walls of the flask), cooled and extracted

*Zxces3 of alkali reduced the yield, %hen 100 per cent excess was present and the reaction was refluxed or steam distilled, all of the soluble cupric ion was converted into water-insoluble black cupric oxide and no water-insoluble organic products were obtained. - 128- with ether. Evaporation of the ethereal extract left white crystals of 2,3“iiimethyl-2,3“cLinitrobutane (0,20 g,, 4*5^) •

2, Preparation of 3»4“®l“othyl-3,4“^.initrohexane,

2-Nitro'butane (l.O g., 0,01 mole) and sodium hydroxide (6 ml, of 10% solution) were warmed in a stoppered flask till solution was effected, then cooled to 0—5° amd treated with faintly alkaline aqueous sodium persulfate (2,38 g», 0,01 mole in 5 nl.)« The mix\ .re was stored for 8 hours at 0-10°, during vdiich time a few drops of dilute alkali was added every two hours. Filtration yields white crystals of 3»4“ii®cbhyl-3,4“4initrohexane (0,37 g., 37^)» shining plates of the dimer were obtained on recrystallization from petroleum ether, m.p, 79,5-80°; lit, 79-80° (17, 40),

Longer reaction times give ioly products; as the reaction pro­ ceeds the pH of the mixture decreases and, when the mixture becomes acidic, the initially-formed idiite solid becomes contaminated with pseudonitrole and is liquefied,

3, Preparation of l,l'-Dinitrobicyclohexyl

Kibrocyclohexane (l.O g., O.OO78 moles) and aqueous sodium hy­ droxide (7 ml, of 10% solution) were warmed in a stoppered flask until a homogeneous solution resulted, then cooled to 0-5° and treated with aqueous sodium persulfate (2,40 g., 0,01 mole in 5 ml, water con­ taining enough dilute alkali to produce alkalinity in the resulting reaction), After 8-12 hours at 0-10°, the crude dimer was collected, washed with water and dried in air (0,26-0,29 g., 26-30%). 1,1*- -129-

Dinltrobicyolohezyl crystallizes from acetone or acetone-methanol as shining plates, m.p, 214° (darkens) to 220° (decomposes); lit*

216.5-217® (60), or 208-209® (58).

4* Attempted Preparation of 2,2,3,3-Tetranitrohutane,

Interaction of aqueous alkaline solutions of sodium or potas­ sium 1,1-dinitroethane and ammonium persulfate solutions yield homo­ geneous, colorless, strongly acidic mixtures on standing at 0-25° for 24 hours. I^o water-insoluble products are formed.

5. Attempted Preparation of Hexanitroethane (115).

Aqueous ammonium persulfate (1.I5 g., O.OO5 mole in 4 ml.) was mixed at 0® with a solution of (l.O g., O.OO66 mole) in sodium hydroxide (3 ml. of 101& solution). A very small quantity of crystalline material separated after a few minutes. If it is placed in an ice chest (O—10®) the mixture slowly decomposed (gas evolution); no increase in the quantity of solid was observed. After

24 hours, the temperature of the mixture was allowed to rise to 25°; further gassing occurred, heat was evolved, the original yellow color disappeared and the solution became strongly acidic. The solid had disappeared.

III. Summary

Nitroethane and sodium persulfate undergo a complex reaction in alkaline solution. The coupled product, 2,3-dinitrobutane, could not be isolated; the end-product of the reaction is 3*4»5”trimethyl- isoxazole. - 1) 0-

Alkaline solutions of secondary nitroalkanes (2-nitropropane;

2-nitrobutane, and nitrooyclohexane) have heen coupled to give the corresponding vicimal dinitro compounds (2,)-dimethyl-2,3-dlnitro- hutano, 3»4-&lBethyl-3,4-&ialtrohe%ane, and l,l'-dinitrohicyclohexyl, respectively) reaction with salts of persulfuric acid, acidified hydrogen peroxide, potassium ferricyanide or ammoniacal cupric chlor­ ide. Highest yields (26-)8^) of coupled products are obtained when salts of persulfuric acid are the oxidants in aqueous solutions; best yields (54“5®î^) of 2,^-dimethyl-2,)-dinitrobutane result idian the aqueous solution of sodium 2-nitropropane and persulfate ion is buffered with sodium acetate.

In addition to being coupled to give the corresponding vicinal dinitro compounds, the salt of the secondary nitroalkane is also oxidized to the corresponding ketone*

The results of this investigation support the postulate (l8,

9))that nitroalkyl radicals are intermediates in the coupling pro­ cess*

. Attempts to convert sodium 1,1-dinitroethane and sodium trini­ tromethane to 2,2,),)-tetrBinitrobutEine and hexanitroethane, respec­ tively, by reaction with persulfate ion, were unsuccemful. - 1) 1-

BIBI.I0aa&FH7

1« Koxnblvtnt B., llchtin, B,, Batton, J», and Iff land, D., J, ia. Cham.

Boo., 307 (1947)' Borntlum, B., Patton, J,, and Bordmann, J,,

i M i . 22, 746 (1948).

2. Whsland, G,W., AdTanoed Organic Cheaistry, Second Edition, Bow York,

John Wiley and Sons, Inc,, 1949# PP» 83I-32#

3» Hass, H,B, and Biley, E*F*, Chen, Bey..'^2. 373"4)0 (l943)«

4, Iisvy, B,, and Bose, J*B«, Quarterly Be?,, i, 358 (l947) #

3, Weisler, L,, Ph.D. Thesis, University of Bochester, 1939*

6, Thurston, J,T, and Sbriner, B,L,, J, Org, Chen,, 2. » I83 (l937)«

7» Hass, H,B, and Bender, M,, J, An, Chen, Soc,, 21# 3482 (l949)«

8, Schunacher, H,J. and Sprenger, C,, Z. anorg, allgen, Chen,, 182.

139 (1929)? Z, Electrochen,, 35, 653 (1929)5 Z, physik, Chen,,B 12.

115 (1931).

9, Daohlauer, K«, Ger, 509»405# (August 30» 1929).

10, Schneisser, M, and Gregor-Haschke, E,, Z, anorg, Chen,, 255. 33

(1947).

11, Zuskine, B,, Bull, soc, chin. [4], 31 » 187 (192$),

12, Noyes, ÏÏ.A», J, An, Chen, Soc,, 34» 3^15 (l932).

13, Stoinkopf, W, and Kilhnel, U,, Ber,, 75B . I323 (1942),

14, Brintzinger, H, and Pfanstiel, E., Z, anorg, Chen,, 256. 324 (1948),

13, Meitner, J,, Ph.D, Thesis, Purdue University, 194^,

16, Jacohscn, H,, Ph, D« Thesis, Purdue University, 1942,

17..Seigle, L,, and Hass, H.B«, J, Org. Chen,, 3» 100 (l947)* - 1 5 2 -

l8, Pearson, R. and Rvans, W»V«, Irons. Sleetroohem. Soc., 173

(1943).

19* Bennett, Gr., Brand, J., and Williams, Cr«, J. Chem. Soc. 869, 873

880 (1946).

20. Gillespie, R.J. and Miller, D*J*, Quarterly Rev., 277 (l948).

21. He£, J.U., Ann., ^80, 263 (1894).

22. Benford, Gr.A., and ingold, G.E., J. Chem. Soc., 93^ (l93^) *

23. Denton, W.I., Bishop, R.?., Rygaard, E.M., and Roland, I.I.,

Ind. Sng. Chem. 40, 381 (1948).

24. Rouben, J., Die Methoden Der Organischen Chemie, Vol. 17, Third

Edition, Leipzig; Georg Ihieme, 1941» pp. 211-14.

25. Black, A.P. and Barters, P.H., Organic Syntheses, Coll. Vol. II,

New York; John Wiley and Sons, Ind., 1943» P» 512.

26. Wislicenas, W. and Bndres, A., Bar.. 19. 1735 (1902),

27. Wislicenus, W., and Waldmtiller, M., Ber., 41. 3334 (1908).

28. Thiele, J.» Ber., 33, 666 (19OO).

29. V . Anvers, K., and Ottons, B., Ber., 44^ (1924).

30. Arndt, P., and Rose, J.D., J. Chem. Soc. 1 (l935)•

31. Berthelot, M., Ann., 111. 80 (i860).

32. Ref, J.U., Ann., 3Qi, 175, 181 (1899).

33* Klason, P., and Carlson, T., Bar., 42.» 4183 (19O7); Carlson, T.,

Bar. 40, 4191 (1907).

34* Rygaard, E.M., McCracken, J.H», and Roland, T.T., U*S. 2,370,185,

Pet. 27, 1943: 2,401,267 and 2,401,269, May 28, 1946.

33* Mayor, V., and Jacobson, P., Lehrbuch Der Organischen Chemie,

Second Edition, Vol. I, Part 1, Leipzig; Von Veit, I907, pp. 403- 07. - 1 5 5 -

36, Dunstan, W,E. and Goulding, E,, J, Chem, Soo,,2 Z* 1262 (19OO);

Cf, Forster, M,0«, J, Chem, Soc..77. 255 (19OO),

37» Sanger, C,, and Roigel, E,, Z, anorg, Chem,, 76. 79 (1912)* Froo,

Am, Acad., 4 2 , 673.

38, Sulc, 0,, Z, phys, Chem,, 72 . 627 (19OO).

39, Schmidt, J,, Ber,, 36, 1776 (19O3),

40, Beyad, I,, and Pirinsky, A,, Ber,, 32.* I23I (1906).

41» Meyer, V,, and locher, J,, Ann,, 180. 167 (1876),

42. Henry, L,, Bull, acad, roy, Belg,, [3], 34* 547 (1898),

43» Black, A.P, and Barters, F.H., Organic Syntheses, Coll. Toi, II,

Hew York; John Wiley and Sons, Ind., 1943» P» 412.

44» Vanderbilt, B.M., and Hass, H.B., Ind, Eng, Chem,, 72. 34 (l240),

45» Meyer, V., Ann,, JJâ» 120 (1875).

46, Konovaloff, M,, Bor,, j8, 185O (1895).

47» Heilbron, I.M., and Bunbury, E.U., Dictionary of OrganicCompounds,

Vol. I, Oxford University Press, 1943» P» 912.

48. ter Meer, ÏÏ., Ann., ^§1, 1 (1876),

49» Yraube, W., Ann., 700. 8l (1898),

50. Carmack, M., and Leavitt, J.J., J. Am. Chem. Soc., 1221 (l949)»

51» Meyer, V., and Constam, E. , Ann. 214. 328 (1882).

52, Tsoherniac, J,, Ann., 180. 166 (1876),

53» Meyer, V., and Locher, J., Ann,, 18O. 136 (1876),

54» Schmidt, J,, and Dieterle, H,, Ann,, 777. 34 (19IO),

55» Schcll, H,, Ber., 2 1 » 508 (1888).

56, B o m , G,, Ber,, 22,» 90 (1896),

57» SchBfer, A,, Ber,, 34, 191O (19OI). - 1 5 4 -

58, Coe, C.S,, Attane, B,C,, and Doumaao, I.F#, U.S. 2, 485» 974 (l949)

59* Johnston, H.L*, and Gianque, W*9 ., J* Am. Chem. Soo*, 3^94

(1929).

60. Nametkin, S., J. Russ. phys.-chem.Soo.,.42, 586; Chem. Zentr.

II, 1376, 1377 (1910).

61. Meyer, V., Ber., 7OI (1876).

62. McCracken,J.and Crandall, U.S. 2,489,398» May 10, 1949*

63. Behrend, H. and Tryller, H., Ann., 283. 209-245 (1894).

84. Ponzio, G., J. prakt. Chem. [ 2 J , 2 2 » 494 (19O8); Gazz. chlm. ital,,

3 1 . II, 133 (1901)5 ibid. 36, II, 287 (1906),

85. Feiser, L.F. and Doerning, 2 ., J. Am, Chem. Spc., 2252 (1948).

86. Charlton, W,, Earl, J . , Kenner, J., and Luoiano, A., J. Chem. Soc.

30 (1932).

67. Schmidt, J., Ber., 33, 3727 (1902). (Cf. Reference 16).

88. Titov, A.I., J, Gen. Chem. (U.É.S.S.), 16. 1898 (1948).

89. Levy, U., Scaife, W., and Bryson, U.S. 2 ,455,425» Dec, 7, 1948;

Brit, 606, 898, Aug. 23, 1948,

70. ffindaus. A., Chem, Zentr. I, 814 (1903).

71. Bredt, J., Ber,, 1 3 , 2319 (1882).

72. Chancel, G,, Bull, soc. chin. [ 2j f 3 1 , 503 (1879)1 Compt.rend,,

• 3â.t 1488 (1883); Chem. Zentr. 3 3 , 484 (1883).

73. Fileti, M. and Ponzio, G,, J. prakt. Chem.||^2j, 33,186 (1897),

74. See Reference 24 for discussion and bibliography.

75. Pictet, A,, and Khotinsky, 2,, Ber,, JO, I163 (1907).

78, Wieland, H,, and Reisenegger, C,, Ann,, 401. 244 (1913)•

77, Losanitsch, S.M,, Ber,, 1 3 , 471 (1882), -1)5-

78, Duden, P., Bor., 26, 3003 (1B93).

79* Ootts, H,, and Hunter, L,, J, Chem. Soc,, 12S. 442 (1924)*

80, lipp. P., Ann,, 44a. 23 (1926),

81, Wieland, H,, and Sakollarios, P., Her. ,^2 , 9^1 (l919)*

82, Hantzsch, A,, and Hinckenherger, A,, Her,, 3 2 » ^37 (l899)*

83, Ueisenhelmer, J,, and Schwartz, M,, Bor,, 32.» 2548 (1908),

84, Duden, P, and Ponndorf, G,, Ber,, 18. 203I (1903).

83, Private Communication from Henry Peuer, Purdue University.

86, Kolthoff, I.M», and Sandell, D.B,, Texthook of Quantitative

Inorganic Analysis, ^he MacMillan Co., Hew York, 1947» P* 572*

87, Kortflm, C,, Z, physik, Chem., B. dl. 279 (l939)*

88, Schmidt, E,, and Wilkendorf, E., fier,, 3 2 , 389 (1919).

89, Darzens, G,, Compt, rend,, 259. 942 (l947)*

90, Wislicenus, W,, and Schlichenmaier, H,, Ann,, d60. 278 (1928).

91* Private Communication from the Aerojet Engineering Corporation,

Azusa, California.

92, Elicke, P.P., Organic Enactions, Hoi, I, Hew York; John Wiley

and Sons, Inc., 1942, p, 3^ 3»

93* Angeli, A,, and Alessandri, L,, Atti, accad. Lincei, 22., I, 784

(1910).

94* Shriner, E,L,, and Brown, G.B., J. Org, Chem,, 2 » 37^ (193?)*

95, Henitzescu, C,D,, and Isacescu, D,A,, Ber,, 6?. 2484 (l930)*

96, Latimer, W,M,, The Oxidation States of the Elements and their

Potentials in Aqueous Solutions,, Hew tork; Prentice Hall, 1938

P* 175* - 156-

97* Huntress» E,M., and Mullikin, S*P,, Identification of Pure

Organic Compounds, New ^ork, John Wiley and Sons, Inc., 1941•

pp. 609-610,

98. Shriner, H,L«, and Fuson, B.C., The Systematic Identification of

Organic Compounds, Second Edition, Hew Fork, John Wiley.and Sons,

Inc., 1940» P» 167 (Procedure 34» Method B.)

99* Fraser, E,, and Kon, 0 ., J. Chem. Soc., 608 (1934)•

100. Hass, H,B», and Bourland, J*F., U.S, 2 ,343 *256, March 7» 1944»

101. Wieland, H., and Beindel, 7 , Ann., /L2A. 74 (l92l).

102. Seigle, L,W., Ph.D, Thesis, Purdue University, 1939»

103» Hudgin, D.E., M.S. Thesis, Purdue University, 1940.

104» Van Tamelen, S., and Van Zyl, 0 ., J. Am. Chem. Soc., , 835

(1949)»

105. Zelinsky, H,J«, J. Rues. phys.-chem. Soo. j6, 6#0 (l§94).

106. Ulpiani, C. and Gasparini, 0., Gazz. chim. ital., II, 242 (19O2)

107. Bahner, C.Ï., U.S. 2,485,803 , Oct. 25, 1949»

108. Hametkin, S., and Posdnjakowa, E., J. Buss, phys-chem. Soc. 4 5 .

1420; Chem. Zentr. 1914. I, 757*

109. Reference 9^ P» 412

110. Wilson, E»H., and Bremaer, J., Quarterly. Rev. 2 * 1 (l94^)«

111. Eharasch, M. , and Gladstone, M. , J. Am. Chem. Soc., 65. I5 (l943)«

112. Fichter, F, and Beer, J., Helv. chim. Acta. ,28, I278 (l935)»

113» Price, C.C., Mechanisms of Reactions at Carhon-Carhon Double Bonds.

Hew York, Interscience Publishers, Inc., pp. 87-89.

114» Waters, W.A., The Chemistry of Free Radicals, Second Edition. - 137“

il§. Will, W., Ber,, Ü , 961 (1914). -PS-

AUTOBIOGRAPHY

I, Ralph B. Kaplan» waa born in Haw York, New York. My seoond-

axy school education was obtained in the public schools of New York

Oity. In 1942 I received the degree Bachelor of Arts from Brooklyn

College. I served in the United States Am y Air Forces in the

United States and in the Southwest Pacific from 1942 to 1946. In

1946 I entered The Ohio State University. I received an appoint­ ment as Research Fellow in 1947, and held this position for three years while completing the requirements for the degree Doctor of

Philosophy.