Tetrahedron, 1963, Vol. 19 Suppl , I, PPo 219 to 2330Pergamon Press Ltd.

SPECTROPHOTOMETRIC METHOD

FOR THE QUANTITATIVE ANALYSIS

OF TRINITROMETHYL COMPOUNDS

D. J. GLOVER

U.S. Naval Ordnance Laboratory White Oak, Silver Spring, Maryland

Abstract - Alkaline hydrogen peroxide rapidly and quantitatively reduces trinitromethylcompounds except esters of trinitroethanol, to the corresponding I, l-dinitro anion. Esters oftrinitroethanol and 2,2-dinitroalcohols are saponified and demethylolated quantitatively totrinitromethyl and 1,I-dinitro anions, respectively.

The intense near ultraviolet absorption of these ions affords a ready method of analysisfor a variety of structures.

The mechanisms of the reactions are discussed.

Introduction

ALTHOUGH many new trinitromethyl compounds have been prepared andstudied, no general method for assaying or analyzing for these compounds inmixtures is known. The only attempt at generality seems to be reductionswith titanous chloride, but the quantity of this reagent consumed per nitrogroup appears to vary with the conditions and the compound to be reduced-: 2.

There have been successful analyses of isolated trinitro-methyl compoundsby various procedures. For example, l,l,l-trinitroethane has been quantita­tively reduced by hydroxylamine and potassium hydroxide to potassium 1,1­dinitroethane, which was then determined by weighing", Bis-(2,2,2-trinitroethyl)urea gave quantitative yields of nitro form by hydrolysis in alkaline solution,and the nitroform was determined by precipitation with tetraphenylarsoniumchloride- or nitron.

A spectrophotometric method for the determination of tetranitrornethanehas been developed at this Laboratory", This determination was based on thereaction of tetranitromethane with hydrazine in alkaline solution, which givesquantitative yields of the nitroforrn anion". Since this anion has a strong ab­sorption at 350rnj-l, spectrophotometric method was feasible.

1 W.W. Becker, Analyt, Chem. 22, 185 (1950).2 R. P. Zimmerman and E. Lieber, Analyt. Chern. 22, 1151 (1950).3 J.Meisenheimer and M.Schwarz, Ber. Dtsch. Chem, Ges. 39,2548 (1906).4 J. M. Rosen, private communication.5 D. J. Glover and S. G. Landsman, to be published.6 A. Bailie, A.K.Macbeth and N. I. Maxwell, J. Chem. Soc. 117,880 (1920).

15 Times NoR. 219

D .J.GLOVER

In view of this success with tetranitromethane, and because the dinitromethylanions corresponding to the trinitromethyl compounds also show strong ab­sorption in the near ultraviolet, an attempt was made to extend this reactionof hydrazine to the latter compounds. When l,l,l-trinitroethane, as a modeltrinitrornethyl compound, was treated with mixtures of hydrazine and sodiumhydroxide, yields of the anion of I, l-dinitroethane approaching 100% wereobtained. However, the absorption of the solution decreased with time, in­dicating further reaction.

It seemed, then, that a less powerful reducing agent was needed. Hydroxyl­amine was substituted for hydrazine, but in a few initial experiments, the yieldof the anion of l,l-dinitroethane did not approach those obtained with hydr­azine, and further use of this reagent was discontinued.

Alkaline hydrogen peroxide was tried next. This reagent is known to giveessentially quantitative yields of nitroform from tetranitromethane, and is re­ported? to react with methyl 4,4,4-trinitrobutyrate to give methyl 4,4-dinitro­butyrate. As with hydrazine, the yields of the 1,l-dinitroethane anion with thisreagent were good. By first mixing the trinitromethyl compound with hydrogenperoxide, and then adding an aliquot of this solution to sodium hydroxide,quantitative reduction, or with esters, quantititive saponification was realized.The anions so produced were determined spectrophotometrically.

EXPERIMENTAL

Apparatus. Absorbancy readings were made with a Beckman spectrophotometer, modelDU, using quartz cells with a 1 cm light path.

All volumetric glassware was calibrated.Reagents. Hydrogen peroxide (30%) was Merck "Superoxol" which was analyzed once

a week. Diluted solutions were prepared as needed from the analyzed material.Methanol, Baker and Adamson, absolute, reagent grade, was used without further puri-

fication. .Sodium hydroxide (1 Nand 0 ,] N) solutions were prepared in the usual way from a satur­

ated solution , and were stand ardized with potassium hydrogen phthalate (NBS).Sodium tetraphenylboron, "Baker Analyzed" reagent, was used without further puri­

fication.

Analytical procedures

(A) Stock solution. Stock solutions of all compounds were made up in methanol. Wheremethanol gave reasonably rapid trails-esterification, dioxane was used as the solvent. How­ever, the production of nitroform in such a situation does not interfere with the analysis andeither solvent may be used. The stock solution was made up to be 0'015-0·02 M in the pro­duct. For example, if the compound gave two moles of nitroforrn anion per mole, the stocksolution was made up 0'008-0 '01 M in the parent compound.

7 J. W.Copenhaver and M.H.Bigelow, Acetylene and Carbon MOl/oxide Chemistry p.24.Reinhold Publishing, (1949).

220

Spectrophotometric method for the quantitative analysis of trinitromethyl compounds

PHYSICAL PROPER TIES OF THE TRINITROMETHYL COMPOUNDS

Purity criteriaCompound

1,1, I· Trinitroethane

1,1,1-Trinitropropane

1,1,1-Trinitropentane

4,4,4-Trinitrobutyric acid

Methyl4,4,4-trinitrobutyrate

y,y,y-Trinitl'obutyrophenone

5,5,5- Trinitropentanone-Z

2,2,2-Trinitroethyl benzoate

Bis (2,2,2-tl'initroethyl) succinate

2,2,2-Trlnitroethyl 4,4.d initrohexanoate

2,2,2-Trinitroethyl 4,4,4-trinitl'obutyrate

2,2-Dinitropropyl 4,4,4-trinitrobutyrate

2,2-Dinitrobutyl 4,4,4-trinitrobutyrate

Bis (2,2,2-trinitroethyl) urea

1,1,1,3-Tetranitro-z-phenylpropane

Tetranitromethane

m.p,

28'3-29·9

78'8-79'3

43-44

76'8-77'2

125-126

63

93'0

94-95

80'9-81'8

186 (dec)

85-86

other

Sublimed

Vacuum at room temp causeddistillation of pink contamin­ant. Residue through silica gelgave colorless product.

b.p, 81'0-81,50 (6'8 film),n14 = 1·4448

Sublimed

Freezing pt= 14'1 0

(B) Hydrogen peroxide solution. An aliquot of solution (A) was pipetted into a mixture ofhydrogen peroxide, water, and methanol. This solution, when diluted to volume in a suitablevolumetric flask, was about 60 volume percent methanol, 1 M in hydrogen peroxide, and0'001-0'002 M in the compound to be analyzed.

(C) Solution for absorbancy determination. A two milliliter aliquot of solution (B) waspipetted into 2 ml of 0'1 N sodium hydroxide in a 100 ml volumetric flask, and this solutionwas diluted to volume and mixed.

The spectrophotometer cell was rinsed 3-4 times with solution (C). The absorbancy wasdetermined against water in the reference cell, and the concentration of the anion in (C) wascalculated from c = AIan,'where c is the concentration in moles per liter, A is the absorbancyof the solution, and am is the molar absorbancy index.

Bis (2,2,2-trinitroetlzyl) urea. The stock solution was prepared as in (A). An aliquot of thestock was pipetted into a mixture of sodium hydroxide, water, and methanol. This solution,when diluted to volume, was about 60 volume percent methanol, 0'01 N in sodium hydroxide,and 0'001 M in the urea. This last solution was diluted by 50 with water for the absorbancydetermination.

1,1,1,3-Tetranitro-2-phl!l1ylpropane. The stock solution was prepared as in (A). An aliquotwas then treated as in (8) and (C), or was treated like bis (2,2,2-tl'initroethyl) urea.

221

D.J.GLOVER

Effects of changes in the general analytical procedure (Table 4)

Variation ofsodium hydroxide concentration. Solutions (A) and (B) of the general analyticalprocedure were prepared as described. Two milliliter aliquots of (B) were pipetted into sodiumhydroxide solutions of different strengths in 100 ml volumetric flasks. These solutions wereimmediately diluted to volume with water (l,I,I-trinitroethane) or 60"'10 methanol-water(2,2,2-trinitroethyl benzoate) and the absorbancy determined.

Variation of hydrogen peroxide concentration. Five milliliter aliquots of solution (A) werepipetted into hydrogen peroxide solutions of different strengths containing 60 volume per­cent methanol. From these solutions, the solutions for the absorbancy determination wereprepared as in (C) of the general analytical procedure.

Reaction in a mixture of sodium hydroxide-hydrogen peroxide (Table 5)

For these experiments, a 5 ml aliquot of solution (A) was pipetted into solutions of variousratios of sodium hydroxide-hydrogen peroxide in methanol-water in a 50 ml valumetricflask. This solution, when diluted to volume with water, was 60 volume percent methanol.

Preparation ofpotassium dinitromethyl compounds. The general analytical procedure abovewas modified in order that pure samples of the potassium dinitromethyl compounds couldbe prepared. This data will be reported in a later communication. Methyl 4,4,4-trinitrobuty­rate, y,y,y-trinitrobutyrophenone, and 5,5,5-trinitropentaneone-2 did not yield the corres­ponding dinitromethyl compounds by this procedure, but did by the iodide procedures.

Spectral data and potassium determination

Spectrum. Generally, a 0-01 M solution of the potassium dinitromethyl compound wasprepared. This solution was diluted (with water containing 2 ml 0·1 N sodium hydroxidefor each 100 ml solution) to give an absorbancy reading of 0'4-0'7 at the Amax • The spectrumwas determined between 320 mfl and 400 mfl usually, readings being taken every 10 mp ex­cept near the Amax , where readings were taken every millimicron. The molar absorbancy indexat the Amax was then determined by weighing no less than 5 separate samples. The values soobtained are given in Table I.

TABLE 1. MOLAR ABSORBANCY INDEXES

Potassium salt of max mil am Sigma

1,1-Dinitroethane a 382 17,130 173I,I-Dinitropropane 382 16,440 1531,I-Dinitropentane 382 16,070 374,4-Dinitrobutyric acid b 382 16,360 77MethyI4,4-dinitrobutyrate 379 17,020 72y,y-Dinitrobutyrophenone C 379 16,370 -5,5-Dinitropentaneone-2 379 16,760 41Nitroform " 350 14,418 85

-~',.

a From chloronitroethane by Ter Meer reaction.b Dipotassium salt, monohydrate.C Acid form; ambased on 2 samples (30,0 mg and 25'3 mg); iodine was difficult to remove

from potassium salt. Attempted preparation by peroxide method gave mostly material ab­sorbing at 360 mf!.

d Private communication from A. Long, NOL.

a D.J.Glover and M.J.Kamlet, J. Org. Chem. 26, 4734 (1961).

222

Spectrophotometric method for the quantitative analysis of trinitromethyl compounds

Potassium analysis. Potassium was determined by precipitation with the complex, tetra­phenylboron anion. The method is essentially procedure B of 010ss9. A 50 ml aliquot of the0·01 M solution of the potassium dinitromethyl compound used above was pipetted into a150 ml beaker, and about 1 ml of cone hydrochloric acid was added and the solution wasmixed. The yellow color of the solution was usually discharged, and in a few cases, a smallamount of oily material floated to the top. This was the unionized form of the dinitromethylcompound, which in the case ofy,y-dinitrobutyrophenone, was a solid. Here the solution hadto be filtered, otherwise the oil did not interfere. Next, an aliquot of the sodium tetraphenyl­boron reagent solution (0'0175 M) containing no more than 1'5 times as many moles ofpotassium ions, was added with stirring. Greater quantities of the reagent gave high results,presumably due to absorption on the precipitate. After 10 min, the white precipitate was:filtered with suction onto a previously weighed, medium porosity, sintered-glass crucible.After washing the precipitate, it was dried at 100° for 1 hr, cooled, and weighed. (Found:K, 19'33,19'08,19,16; Calc. for KHC sH404 (NBS): K, 19'14 %). (Found: K, 28'60; Calc. forKH2P04 (NBS): K, 28·72 %).

ANALYTICAL RESULTS AND DISCUSSION

The analytical procedure finally adopted (described fully in the Experimentalsection), consists of three steps:

(1) A solution of the trinitromethyl compound in methanol is accuratelyprepared.

(2) An aliquot of (1) is pipetted into a mixture of water, hydrogen peroxide,and methanol.

(3)An aliquot of (2) is pipetted into 0·1 N sodium hydroxide in a suitable sizevolumetric flask. After diluting (3) to volume, its absorbancy is determined witha spectrophotometer.

In this procedure, step (2) involves the initial mixing of the trinitromethylcompound with hydrogen peroxide in the absence of base, and for generality,this order of addition was followed. However, as will be shown later, mixingthe hydrogen peroxide and sodium hydroxide first can effect the desired re­actions with 1,1,I-trinitroethane and 2,2,2-trinitroethyl benzoate.

Table 2 gives the results obtained using the standard procedure above withtrinitromethyl derivatives of aliphatic hydrocarbons, acids, esters not sub­stituted on the alcohol portion, and ketones. Quantitative reduction to thecorresponding dinitro anion was effected in each case, equation (1), except

(1)

with 1,1,1,3-tetranitro-2-phenylpropane. Here, the only instance of the reversalof the Michael reaction observed in the work occurred. This compound gavequantitative yields of nitroform anion, both by the standard analytical proced­ure and in the absence of hydrogen peroxide.

9 a.H. Gloss, Chemist-Analyst, 42, 50-55 (1953).

223

D.J.GLOVER

TABLE 2. QUANTITATIVE REDUCTION OF TRINITROMETHYL COMPOUNDS

Molarity in stock solutionCompound

Present Found" % Found Sigma

1,1,1-Trinitroethane 0'01790 0'01790 100'00·01848 0'01838 99'40'01572 0'01572 100·0

Av. 99'8 0·431,1,1-Trinitropropane 0·02120 0'02144 101·0

0'01884 0'01893 100'40-01679 0'01700 101'2

Av. 100·9 0'391,1,1-Trinitropentane 0·01812 0'01803 99'5

0'01623 0'01617 99'70'02075 0'02056 99'0

Av. 99'4 0'364,4,4-Trinitrobutyric acid 0·01621 0·01626 100'3

0'01756 0'01759 100'10·01497 0·01504 100·4

Av. 100'3 0'30Methy14,4,4-trinitrobutyrate 0'01751 0'01752 100'0

0'01609 0'01608 100'00·01463 0·01463 100'0

Av. 100'0 0'22v,Y,y-Trinitrobutyrophenone 0'01504 0'01480 98-4

0'01629 0'01601 98'30'01735 0·01696 97'8

Av. 98·2 0'395,5,5-Trinitropentanone-2 0'01366 0·01334 97'7

0·01477 0'01444 97'8Av. 97-8 0'34

Reverse Michael1,1,1,3-Tetranitro-2-pheny1propane b 0'02152 0'02156c 100'2

0'02152 0'02157d 100·2 -a Each value is the average of 3 determinations.b Nitroforrn is the species measured.C One determination only.d No peroxide present.

When trinitroethyl esters were subjected to this standard procedure, it wasanticipated that the anions of the corresponding dinitroethyl esters would beobtained. Instead, saponification with subsequent deformylation was observed,(Eq. 2), and this reaction took place quantitatively to the complete exclusionof the expected reduction. The nitroform anion,

RCOOCH2C (N0 2)3 ~~~~--.. RCOO- + CH20 + -QN02)3 (2)

which resulted from the saponification was determined spectrophotometricallyand the results obtained with several esters are given in Table 3.

224

Spectrophotometric method for the quantitative analysis of trinitromethyl compounds

TABLE 3. QUANTITATIVE SAPONIFICATION OF ESTERS

Molarity in stock solutionCompound I% FoundPresent Found " Sigma

0, 0,__,_-

2,2,2- Trinitroethyl benzoate 0'02006 0-02002 99·80'01836 0'01834 99'90'01751 0'01742 99'5

Av. 99'7 0'35Bis (2,2,2-trinitroethyl) succinate 0'009070 0'00906 99'9 I

0'009896 0'00992 100'20'008240 0'00828 100·5

Av. 100·2 0'382,2,2-Trinitroethyl 4,4-dinitrohexanoate b 0·01086 0'01098 101'0

0'009275 0'00935 100·80'008555 0'00856 100·1

Av. 100'6 0'54Saponification and reduction

2,2,2-TrinitroethyI4,4,4-trinitrobutyrate 0'01002 0'01001 99'90'01019 0'01019 100'0

I 0'01111 0'01106 99'6Av. 99·8 0·39

2,2-Dinitropropyl 4,4,4-trinitrobutyrate 0·008580 0'00859 100'10'007488 0'00749 100'00'009094 0'00908 99'9

Av, 100·0 0'272,2-Dinitrobuty14,4,4-trinitrobutyrate 0·009630 0'00965 100'2 e

0'008134 0'00817 100·4Av. 100'3

Hydrolysis

Bis (2,2,2-trinitroethyl) urea d 0'01008 0'01005 99'70'01090 0·01082 99'30'01141 0'01125 98'60'01160 0'01149 99'0

Av. 99·2 0'51~_..

a Each value is the average of three determinations.b Nitroform is the species measured.e One determination only.d No peroxide present.

Included in this group is 2,2,2-trinitroethyl 4,4-dinitrohexanoate. This com­pound was tried in order to see if there was any side reaction, such as a reverseMichael reaction, caused by the presence of an internal gem-dinitro group else­where in the molecule. Again, only saponification occurred.

Since quantitative reduction was observed with 4,4,4-trinitrobutyric acid andquantitative saponification with 2,2,2-trinitroethyl esters, it was of interest tosee what would happen with a compound containing both groups. For example,with 2,2,2-trinitroethyl 4,4,4-trinitrobutyrate one might hope to get an equi­molar mixture of-the anions of nitroform and 4,4-dinitrobutyric acid. Figure (1)

225

D.l.GLOVER

shows the spectrum of each of these ions, curves I and IT, and the curve ob­tained by adding the two individual spectra, curve III, solid line. That bothquantitative reduction and saponification did indeed occur by subjecting thiscompound to the standard procedure can be seen by the plot of the experi­mental data superimposed on curve III (circled points). Similarly preparedcurves for 2,2-dinitropropyl 4,4,4-trinitro butyrate and its products, the anionsof 1,I-dinitroethane and 4,4-dinitrobutyric acid, are shown in Fig. (2). Againexperimental points obtained fell quite closely on the theoretical curve, III. Themolar absorbancy index at the Am Dx of the mixture was obtained by adding themolar absorbancy indexes for each of the products at this wavelength. Usingthe molar absorbancy index so obtained, the analytical results for both of these

2~ ...----.------,----..-------.-------..,.-------..,.------,

x

'"o'"~ 15 f--++----\----t------\;,-£----t---"<;I-"""--I'a;aa:o<f1CD«

..., 20 f---t----7!--\----t-----t-----t-o>"'...----I---j'2

400390380360 370WAVELENGTH. »»

350340

DL.-__-'--__--'--__--'-__--'-__--L.__---'__----'

330

FIG. 1. Spectra of the anions of nitroform and 4,4-dinitrobutyric acid. Curve III isthe theoretical curve formed by the addition of curves I and II. The circles representexperimental points o.btained with 2,2,2-trinitroethyl 4,4,4-trinitrobutYrate

(TNETB) using the standard analytical procedure.

compounds and also for 2,2-dinitrobutyl 4,4,4-trinitrobutyrate are given 111

Table 3.From both Tables 2 and 3, it can be seen that recoveries of 100 ± 1% were

realized. For greater accuracy and precision, a differential spectrophotometricanalysis may be used 10.

The formation of the nitroform and dinitromethyl anions apparently is notfollowed by any further reaction, the absorbancy of the solution remainingconstant over a period of several hours.

10 M.Beroza, Analyt, Chern. 25, 112 (1953).

226

Spectrophotometric method for the quantitative analysis of trinitromethyl compounds

The analysis of three compounds could not be effected by the standard pro­cedure. The analytical results given in Table 3 for bis (2,2,2-trinitroethyl) ureawere based on the formation of nitroform anion produced in the absence ofperoxide by sodium hydroxide. With a mixture of peroxide and hydroxide,

35r---r----.---..-~-...,.....----r---_._--_,

301----+---+---+---~--+---_I<_--_I

251----+---+---+--I---I----+----f---\---1

~ 2of----\----I---I-+-----j---+----+-----'L:;>cz

"~~ 151----+---++---I---7i"7~--+------''''<f'';,_--j

<r

"..Jo

"101-----+-f---I--ti4---+----l----+---~

330 340 350 360 370WAVEI.ENGTH [,"/-,I

380 390 400

FIG. 2. Spectra of the anions of l,l-dinitroethane and 4,4-dinitrobutyric acid.Curve III is the theoretical curve formed by.the addition of curves I and II. Thecircles represent experimental points obtained with 2,2-dinitropropyl 4,4,4-tri.

nitrobutyrate (DNPTB) using the standard analytical procedure

yields of 90-95 % of nitroform from tetranitromethane were realized. Nosatisfactory procedure could be found for bis(2,2,2-trinitroethyl) nitramine.

MECHANISM

Possible reaction paths

The attack of the nucleophilic reagents, hydroxide ion and hydroperoxideion, on trinitromethyl compounds can take place at several sites. Using 1,1,1­trinitroethane and 2,2,2-trinitroethyl benzoate as examples of compounds where

227

Di J, GLOVER

reduction and saponification are the desired reactions, the outline below (3)shows some of the possible reaction paths. Reactions following these routes areknown to take place with analogous compounds under similar conditions.

B.1-

e -+ CH 3-C(N02) 3 f- A

C,j.

Ph-CO-O-CH2-C(N02h f- A't tD B'

(3)

Reduction. A and A' lead to reduction producing the dinitro compound.Where hydroxide ion is the attacking agent, the mechanism probably involvesan SN2 or SN2' displacement of l,l-dinitroalkane anion on nitrite by hydroxideto give nitrate ion:

RC(N02) 3 + OH- -+ [RCCN0 2) 2 ••• N02 ••• OH]- -+ RC(N02) 2 +

+. HON02 (4)

There is no evidence to show whether attack is on nitrogen (SN2) or oxygen(SN2') of the nitro group. With hydroperoxide ion, a colorless gas was evolved,and in a reaction with tetranitromethane in excess, one mole of nitrite ion wasformed for each mole of peroxide present. Thus, the mechanism is probably

(5)

as shown ill equation (5).Hydrolysis. Band B', which involve attack on the terminal carbon atom, lead

to hydrolysis to the carboxylic acid. Although steric considerations would in­dicate that access of nucleophlic reagents to this crowded carbon atom mightbe difficult, the formation of carbonate from the reaction of tetranitromethanewith hydroxide ion-', equation (6),

(6)

and succinic acid from 4,4,4-trinitrobutric acid-s, would indicate that under cer­tain conditions this reaction can take place.

11 E.Schmidt, Bel'. Dtsch. Chern. Ges. 52, 400 (1919).12 H.Feuer, E.H.White and 8. M. Pier, J. Org. Chern. 26,1639 (1961).

228

Spectrophotometric method for the quantitative analysis of trinitrornethyl compounds

Unsaturation. C and C/, involving abstraction of a hydrogen atom from thecarbon atom vicinal to the trinitromethyl group, would be expected to give anunsaturated dinitro compound with the elimination of HN02 by the EN2 route:

The instability of dinitro olefins has been recorded-".The course of the further decomposition is not known but possible routes

include:

(9)

Thus, a potential product of route C or C' is dinitromethane. If the terminaldouble. bond initially formed can shift into conjugation with another group(such as carboxyl), then the decomposition reaction will not occur. That thisshift is more rapid than the decomposition reaction is indicated by the fact thatdipotassium 4,4-dinitrobutenoate is readily formed by the reaction of methyl4,4,4-trinitrobutyrate with alkali under preparative conditions with no tem­perature control.

The formation of the hydroxy dinitro anion (Eq. 9) as a possible intermediateis further indicated by the finding of Kaplan and Kamlet'<, that dimethyl 3­hydroxy-4,4-dinitropimelate is formed as a by-product in the reaction of nitro­form with methyl acrylate under alkaline conditions.

Saponification. D gives saponification, and in the case of 2,2,2-trinitroethylbenzoate the 2,2,2-trinitroethanol demethylolates giving nitroform. The ef­ficacy of hydroperoxide ion as a saponification catalyst compared to hydroxideion is discussed below.

Experimental results and mechanisms

That the conditions of the standard analytical procedure give preference to Aover Band C, and Dover A', B', and C, is evident by the quantitative for­mation of 1,I-dinitroethane and nitroform from 1,1.l-trinitroethane and 2,2,2­trinitroethyl benzoate, respectively. The reaction conditions were varied inorder to find the range over which the reactions at A and D were solely orprimarily operative.

13 Dr. W.F.Sager, George Washington University, private communication.14 L. A. Kaplan and M. J. Kamlet, private communication.

229

D.l.GLOVER

The effects of changes in the general analytical procedure are shown inTable 4. Note that the conditions may be varied rather widely with 1,1,1­trinitroethane, but not with 2,2,2-trinitroethyl benzoate. With the latter com­pound, using the standard hydrogen peroxide concentration, the sodium hydrox­ide concentration can only be increased by a factor of two. Above this con­centration, the apparent yield of nitroform is greater than I00%. Such an ob­servation can only be accounted for by a species having a greater molar ab-

TABLE 4. EFFECTS OF CHANGES IN THE GENERAL ANALYTICAL PROCEDURE

1,1,I-Trinitroethane = 3'574 x 10-5 M; 2,2,2-Trinitroethyl Benzoate = 4·000 X 10-5 M

Molarity x lOsa % Reaction b

HzOs NaOH TNEt TNEBz

0'229 2'070 c 84 93·60'457 2'070· 97'2 -0'686 2'070· 99'8 -0'914 2'070· 99·7 -1'14 2'070· - 97'51'83 2'070· - 98·64'57 2'070· - 99'7

22-9 2'070· 100'6 100'791-4 2'070 c 100'1 100-222-9 0'1035 d 74·2 72·622-9 1.035 d 100'0 100'322-9 2'070 d - 100'522-9 10'35 d - 101'322-9 20'70 d - 102'122-9 98'8 d - 102·622'9 lO1-4d 100'0 -22-9 1 X 103 e 100·7 65'9

a Calculated from the amount put into the final diluted solution.b Absorbancy found at the Amnx of the desired product divided by the calculated value.c 2 ml of trinitromethyl compound-hydrogen peroxide solution into 2 1111 0·1035 N sodium

hydroxide, then diluted to 100 ml to give this value.d 2 ml trinitromethyl compound-hydrogen peroxide solution into 95 ml sodium hydroxide­

water, then diluted to 100 ml to give this value.• 2 ml trinitromethyl compound-hydrogen peroxide solution into 5 ml saturated (20 M)

sodium hydroxide, then diluted to 100 ml to give this value.

sorbancy than nitroform, such as a dinitro species produced by reaction A/(equation 12) and/or C/. The latter route would give dinitromethane (A mnx

= 361m/-l, Am = 20,90015, 16.

(12)

15 H. Shechter, Ohio State University, private communication.16 Aerojet Engineering Corporation, Azusa, California, private communication.

230

Spectrophotometric method for the quantitative analysis of trinitromethyl compounds

From the results with 1,1,I-trinitroethane (Table 4), it is felt that A' is probablythe predominant reaction site.

Quantitative reduction and saponification can also be effected by mixingthe hydroxide and peroxide before addition of the compound to be analyzed.The results obtained by varying the hydroxide and peroxide concentrations with1,1,I-trinitroethane and 2,2,2-tdnitroethyl benzoate as model compounds aregiven in Table 5. With both compounds, the optimum conditions exist when theperoxide-base ratio is 84. When this ratio is decreased to about 8, the yield ofdesired product is decreased by about 5-6 %, and is further decreased as thisratio becomes smaller.

TABLE 5. REACTION IN A MIXTURE OF SODIUM HYDROXIDE-HYDROGEN PEROXIDE"

I, I, I-trinitroethane = 1'982 x 10-3 M b ; 2,2,2-trinitroethyl benzoate = 2'018 x 10-3 M b

-- ---_.__..".,----

Molarity Ratio 1,1,1-Trlnitroethane 2,2,2-Trinitroethyl benzoate------

H 202 NaOH H 202!NaOH'}~C Amnxs mu %C Am• x, mu

Reaction Product Reaction Product

0'858 0·01026 83'6 99'3 382 100'3 3500'0858 0'01026 8'36 95 379 94 3500-0858 0·1015 0'845 95 379 86 3510'0858 0'3045 0·282 90 377 81 3560 0·002052 - - - 80 357

" Solvent is 60 volume percent methanol.b Solution was diluted by 50 to find absorbancy.

C Absorbancy found at the Amnx, of the desired product divided by the calculated value.

From the data in Table 5, it can be seen that with l,l,l-trinitroethane, BandC take an increased part when both hydroxide and peroxide are present initially.The shift in I'max is hypsochromic (from 382m.u) and the species is probably theanion of dinitromethane produced as described above. The decreased yieldcan be accounted for by the complete destruction of the nitro species by reactionat B.

With 2,2,2-tdnitroethyl benzoate, similar conclusions can be reached fromthe data in Table 5. When both hydroxide and peroxide are present initially,or when hydroxide only is present, there is a bathochromic shift in Amux (from350m.u) which reflects the formation of the dinitro species as described pre­viously, [equations (8), (9), and (12)]. The lower than 100% yield shows theincrease of B' participation.

The failure of bis (2,2,2-trinitroethyl) nitramine and bis (2,2,2-trinitroethyl)urea to give quantitative yields of nitro form anion in the presence of hydro­peroxide ion, can be explained with the help of these reaction schemes. Fromthe urea, nitroform is produced quantitatively by hydroxide ion in the absence

231

D.l.GLOYER

of hydroperoxide ion. Nitroforrn is probably formed by an attack on the amidehydrogen.

(N02)aCCH2NHCONHCH2C(N02)3~ (N02)3CCH2NCONHCH2C(N02)a

--> (N02)3C- + CH2= NCONHCH2C(N02h

(13)

In the presence of hydroperoxide ion, a bathochromic effect is observed in thespectrum of the reaction products. The shift is to 357m.u from 350m.u, in­dicating that, although the rate of attack on amide nitrogen is probably notsubstantially increased, the reduction reaction is greatly enhanced by hydro­peroxide ion and a mixture is formed.

With the nitramine, hydroxide ion mixed with hydroperoxide ion gives amixture of products having an initial Amnx of 362m.u. During the reaction, thereis a hypsochromic shift and the final Amax is 358m«. Similarly, with hydroxideonly, a mixture is formed. Like the urea, the nitramine in the presence ofhydroperoxide ion probably gives nitroform and a dinitro species. Here, ap­parently, the dinitro species is formed initially in greater amounts and, as thenitroform forms, there is a hypsochromic shift. This indicates that with eitherreagent, the attack on the nitramino group is comparable in rate with the re­duction.

Function of hydrogen peroxide

Hydrogen peroxide under the conditions of the analytical procedure performsthree major functions: (1) it effects reduction, A and A', and saponification, D,as effectively, or about as effectively, as hydroxide ion; (2) it is not as effectiveas hydroxide ion in the side reactions, Band B', C and C'; (3) it is a strongeracid than water (k' = 2·4 X 10-12 at 26°)17, and by its buffer action depressesthe concentration of hydroxide ion, thereby lessening the ability of hydroxideion to take part in the undesired side reactions. Under the conditions of thestandard procedure, the hydroperoxide ion is about 0·05 M while the hydroxideion has been depressed from 0·05 N to about 3·3 x 10-4 M. This explains whyin the presence of excess base (insufficient hydrogen peroxide), where de­pression of the hydroxide ion is incomplete, the undesired side reactions be­come significant and quantitative yields are not obtained.

Hydroperoxide ion as a saponification catalyst

The quantitative hydrolysis of 2,2,2-trinitroethyl benzoate and quantitativelack of hydrolysis of methyl 4,4,4-trinitrobutyrate were initially surprising. Onfurther consideration, however, it was seen that these results tie in quite nicely

17 R.A.loyner, Z. Allor. Chem. 77, 103 (1912); Chern. Abstr., 6,2880 (1912).

232

Spectrophotometric method for the quantitative analysis of trinitromethyl compounds

with a theory postulated by Wiberg" to explain the extreme difference in therate of saponification of ethyl acetate and phenyl benzoate. He attributed thefact that benzonitrile reacted 104 times as fast with hydroperoxideion ashydrox­ide ion to the increased nucleophilic character of the former [Eq. (14)].

PhCN + OOR~..... PhC=N- H.O...... PhCONH2 + O2 + OH- (14)I

OOH

In this case the rate determining step involved k1 , the addition across the mul­tiple bond. In the case of the saponification of esters [Eq. (15)],

0­I

RCOY + x- /'..... R-C-Y -.!<!-,. RCOX + Y-k : , I

x

(15)

the rate depends on k 1 , but also, to a greater extent on the ratio k 2/k_1

Wiberg concludes that the ratio k2/L1 depends on the relative basicity of thegroups, X- and Y-. The weaker the basicity of the group being split off, thegreater the rate of the step leading to this ion. With ethyl acetate, k- 1 wouldlead to splitting off of hydroperoxide ion, while k 2 would lead to the loss ofethoxide ion. Since ethoxide ion is a substantially stronger base than hydro­peroxide ion, the ratio k2/k- 1 is very small and reaction does not proceed ata measurable rate, despite the .fact that k, is probably substantially greaterthan it would be in the case of saponification by hydroxide ion. With phenylbenzoate, phenoxide ion is a weaker base than hydroperoxide ion. Here, theratio k 2/k_ 1 is appreciable, saponification proceeds, and at a more rapid ratethan with hydroxide ion because of the increased magnitude of k1 • The situationinvolving methyl 4,4,4-trini~robutyrate and 2,2,2-trinitroethyl benzoate is iden­tically analogous. Here again, methoxide ion is a stronger base than hydro­peroxide ion, and saponification does not occur, while the 2,2,2-trinitroeth­oxide ion is a substantially weaker base than hydroperoxide ion, and saponi­fication is rapid and quantitative.

Acknowledgment- The help of the members of the Organic Chemistry Division in preparingand in purifying the starting material is greatly acknowledged, particularly the help ofMr. Francis Taylor, Jr. and Dr. Rip Rice. I am indebted to Dr. Mortimer J. Kamlet for themany helpful discussions during the preparation of this report, with special emphasis on themechanism section.

18 K.B.Wiberg, J. Amer. Chem. Soc. 77,2519 (1955).

233