Esters of Peroxycarbonic Acids

8/4/2019 Esters of Peroxycarbonic Acids

http://slidepdf.com/reader/full/esters-of-peroxycarbonic-acids 1/10

1254 STRAIN,ISSINGER,DIAL,RUDOFF,DEWITT,STEVENSND LANGSTON Vol. 72

5 . The isomeric methoxypropanols can be 6. Each ,of the isomeric methoxypropanolscharacterized by their 3,5-dinitrobenzoates and a-

naphthyl urethans. COLLEGEARK,MARYLAND RECEIVEDPRIL 14, 1949

forms an azeotrope with water.

[CONTRIBUTIONROM THE RESEARCH ABORATORYF TH E PITTSBURGHLATEGLASSCo., COLUMBIA HEMICALDIVISION]

Esters of Peroxycarbonic Acids

BYF. STRAIN, . E. BISSINGER,W. R. DIAL,H. RUDOFF,~. J. DEWITT,H. C. STEVENSNDJ. H. LANGSTON~

Diethyl peroxydicarbonate (I , R = C2H6),b-tained in impure form by Wieland, vom Hove andBorner, is described3 as an explosive oil of extremeinstability, decomposing at moderate tempera-tures, and possessing the sharp ozone-like odor ofvolatile acyl peroxides. The preparative method

consisted of the reaction of ethyl chloroformate inchloroform solution with powdered sodium perox-ide, promoted by a drop or two of water added tothe dry reagents. Control of the reaction was dif-ficult. The instability of the compound, under theconditions of preparation and attempted purifica-tion by distillation, probably accounted for theirfailure to obtain a product of greater than 88%

Our preliminary experiments confirmed this de-scription, and further demonstrated that the di-ethyl and other esters of peroxydicarbonic acidcould be prepared readily by the careful reactionof chloroformates with aqueous sodium peroxide

solutions, and that the products were effectivepolymerization catalysts for several ethenoid mon-omers. A series of the esters was therefore pre-pared for further observation of their properties.

While the present paper is concerned chieflywith the aforementioned esters of peroxydicar-bonic acid (Type I), illustrative preparations ofthe related esters of monoperoxycarbonic acid(Type 11) and of diperoxycarbonic acid (Type111),derived from the reactions of a chloroformate

purity.

0 0 0I1I II

RO-C-0-ORO-C-0-0-C-OR

I I1

0II

RO-0-C-0-OR

I11

and phosgene, respectively, with t-butyl hydro-peroxide in the presence of acid acceptors are alsoincluded. Although the identity of some of theinorganic sal ts of peroxycarbonic acids has beend i ~p u t e d ,~tructures corresponding to formulas

(1) Present address:

(2) Present address: Clemson College, Clemson, S.C.

(3) Wieland, vom Hove and Borner, A n n . , 446, 46 (1926).

(4) Tanatar, Ber. , 82, 1544 (1899); Wolffenstein and Peltner,

i b i d . , 41 , 28Cb297 (190 8); Le Blanc and Zellerman, 2. Eleklrochem.,

29, 179, 192 (1923); Kamtikar and Husdin, J . O sma ni o U n i v . toll.,

General Electric Company, Schenectady ,N. Y .

a, 39 (1934).

of Types I, I1 and I11 have beenPreparation of esters of these three types tends tosupport further the plausibility of the assignedstructures of these salts.

Esters of Type I.-The ready availability ofmany chloroformates from the alcohols and phos-

gene affords access to a variety of the dialkyl andsubstituted dialkyl peroxydicarbonates. Someof the chloroformates used were prepared bythe common procedure of passing phosgene in tothe cooled a1coho1,6asb ut a number were obtainedby an improved method in which the alcohol wasadded to an established pool of the chloroformatecontaining some phosgene, which refluxed andserved to cool the reaction mixture. The lattermethod enabled maintenance of an excess of phos-gene throughout the reaction without the necessityfor handling large q~ a n t i t i e s ~ ~ , ~f liquid phos-gene. This improved procedure also gave in-creased yields, and often a chloroformate of higher

purity than the simpler method. Chloroformatesemployed in this investigation included a numberpreviously unreported. These are listed in TableI, along with hitherto unreported characterizingdata upon the isopropyl and allyl compounds.

The preparative procedure for the esters ofType I consisted of careful addition of cold aque-ous sodium peroxide solutions (or slurries) to thecooled chloroformate* with efficient mixing. Dueto the high degree of thermal instability and dan-ger of violent decomposition observed for a num-ber of the esters a t temperatures only slightlyabove the temperature (0-10”) of preparation,special care was required in temperature control

in all of the operations of preparation and purifi-cation. The chief impurity present in the crudeesters was the alcohol formed by hydrolysis of thechloroformate. Since this impurity was easilyremoved from the crude lower alkyl esters bywashing with cold water to give a product of high

(5) F. Ephraim, “Inorganic Chemistry,” Fourth Edition revised,

Nordeman Publishing Co ., Inc. , New York, N. Y.,943, p. 842.

(6) (a) See Dyson, Ch cm . Revs . , 4, 149-150 (19 27) , for a review of

early literature references to the preparation of chloroformates from

alcohols and phosgene; (b) see Huntress, “Organic Chlorine Com-

pounds,” John Wiley and Sons, Inc., New York, N. Y., 948, index

pp. 1420-1421, and text indicated for references and reported proper-

ties on many of the chloroformates used.

(7) (a) Ashhurn, Collett and Lazzell, THISJOURNAL, 60, 2933

(1938); (b) Slimowicz and Degering, i b i d . , ‘71, 1043 (1949).(8) F. Strain, U. . Patent 2,370,588, Feb. 27, 1945.



March, 1950

Chloroformate

Isopropyl5

NeopentylAllyld

Tetrahydrofurfuryl

2-Carbamyloxyethyl

2-Nitrobutyl

2-Nitro-2-methylpropyl

Dodecyl

Ethylene-bis-'

2,2 -Oxydiethylene-bi~-~

ESTERS F PEROXYCARBONICCIDS

TABLE

PREPARATIONND PROPERTIESF CHLOROFORMATES

Scale ofB . P . ,

?GODprepn.,

5.0b 83 47 100 1 398lC

0.4b 83 52 27 1.409110.0 90 56 97 1.4223"

0.5 54 94-97 19 1.4522

0.5b 100 . . . . . . . 1,4640'

0 .5b 81 . . . . . . . 1.4472

1 o" g 88 . . . . . . . 1,4453

0 .2h 84 . . . . . . . 1.4408

5.0b . 77 113 25 1.4498

5. b 95 122 4 1.4542

theoret. moles Yield, 70 OC. Mm.

1255

Analyses, '% C1Calcd. Found

29.0 28.8

23.5 23.429.4 29.2

21.5 21.4

21.2 21.1

19.5 19 .1

19.5 18.8

14.3 14.1

37.9 37.9

30.7 30.7

0 Ref. 27; ref. 6b, page 1034. * Preparative Procedure B was employed. da4 1.076. Fierz-David and Muller,j dmd 1.421; mol .

Dioxane, 100cc., was employed as a solvent in the reactionEther was added as a solvent to mainta in a liquid reaction mixture and to assist in washing of the crude prod-

dZar .465; Oesper, Broker and Cook, THIS OURNAL,7, 2609 (1925) ; F. Strain, U. s.Patent 2,397,630, Apr. 2,i dZo4 .389, m. p. 5-6"; I. E. Muskat and F. Strain, U. S. Patent 2,370,568, Feb. 27, 1945.

J . Chem. SOL,125 ,26 (1924); Schving and Sabetay, Bull . SOC. chim., [4] 43,858 (1928).ref. calcd. for HzNCOOCzH40COCl: 32.22; found: 32.52.mixture.uct.1946.

e dmr 1.136.

purity, and because of the great instability ofthese esters, it was considered impracticable toat tempt distillation. The preparations of estersof Type I are summarized in Table 11. Attemptsto prepare a diary1 (diphenyl) peroxydicarbonatefailed, no compound containing active oxygenbeing obtained.

The lower aliphatic esters of Type I were ob-tained as oily liquids or low-melting solids possess-ing a characteristic odor siniilar to that of thelower acyl peroxides. As the size of the aliphaticradical is increased, especially if polar groups arecontained therein, solid compounds tend to result

and the characteristic odor is diminished. The

esters were found to be soluble in a wide range oforganic solvents, but only. slightly or negligiblysoluble in water. They exhibit the common re-actions of organic peroxides such as liberation ofiodine from acidified solutions of potassium iodide,decomposition on contact with concentrated sul-furic acid, deflagration on contact with flame, andin some cases explosive decomposition on beingsubjected to heat, friction or shock. With few ex-ceptions the peroxydicarbonates prepared tendedto decompose spontaneously a t normal room tem-peratures. The heating resulting from this de-composition texded to produce an autoacceler-

ated explosive decomposition of the lower alipha-

TABLE1

PREPARATIONND PROPERTIESF PEROXYDICARBONATES,YPE , ROCOOCOR

0 0II I/

K

Scale of Gross decompn., temp., OC., A n a l . , active 0,

moles" % ? i%b T C Calcd. Foundpreparation, Yield, methodo %

Methyl 0.10 6 . . . . . 5h 60 , expl. 10.66 10.50

Ethyl 5.0 81 1.4017 28-35 8. 98 8.88

n-Propyl 0 . 5 81 . . . . . 32-36 7.76 7.67

Isopropyl 0.6-6.0 81-89 1.4034d 35-38 47-51 7.76 7.6 8

Isobutyl 1.3 75 1,4148" 40-45 6.83 6. 79

9 51-56 6.10 6.03

65-70 5.59 5.5 7

Neopentyl 0. 08 67/

Cyclohexyl .1 5 72'Allyl .05 46 1.434 Expl. 7.91 7.13

Tetrahydrofurfuryl .05 . . . . . . . . 10-20 5. 51 5.51

j3-Methoxyethyl . 1 28 1.4250 34, expl. 6.72 6.52

a-Carbethoxyethyl .2 5 27 1.4266 . . . . 4.96 4.81

2-Chloroethyl .10 79' 1.4582 4c-45 6. 48 6.44

2-Csrbamyloxyethyl .0 5 58 i 99-100 5. 40 5.25

2-Nitrobutyl . l o 85' 94 4.93 4.89

2-Nitro-2-methylpropyl .4 78' 101 101 4.93 4.90n

h

iBenzyl .03 90 101-102 5.29 5.22

k

m

a Theoretical yield of peroxydicarbonate. Melting points indicated by footnotes d, g- k an d m in this column areuncorrected. Minimum temperature ranges required for steady gas evolution, observed in 4-ml. test-tubes (T) orm. p.capillaries (C) see Experimental. M. p. 8-10'; mol. wt. calcd. for CsHl4Os: 206.18; found (cryoscopic in ethylene di-bromide), 206; d15.64 16080. A M. p.46'. 6 M. p. 101-102 . Compound was a solid atdo om temperature. A M. p. 50'. Ethyl acetate was employed

as a solvent in this preparation. M. p. 100.5-101 .11 A n al . Calcd. for C10H16010N2: C, 37.0; H, 4.97; N, 8.64.Found (by Dr. C. Tiedcke, Laboratory of Microchemistry, Teaneck, K . J.) : C, 36.4; H, 4.72; Pi,7.52.

e d44 1.060. Ether was employed as a solvent in th is preparation. 0 M. p . 45".



1256 STRAIN,ISSINGER,DIAL,RUDOFF, EWITT,STEVENSND LANGSTON Vol. 72

tic esters. Quantities of more than a few gramsof the pure diethyl and diisopropyl esters, for ex-ample, were found to be hazardous for this reasona t temperatures above 10") although kilogram orlarger quantities were kept with negligible decom-position when adequately refrigerated. In smaller

masses with highly efficient heat transfer, thetendency for the decomposition to become violentcould be suppressed up to temperatures of 40-50".AS the size of the aliphatic radical is increased,the tendency for spontaneous heating with violentdecomposition is moderated a t normal tempera-tures, resulting in a sudden-to-mild effervescenceon exposure to slightly elevated temperatures.When subjected to impact or explosive shock, thedimethyl and diethyl esters appeared more sensi-tive than benzoyl peroxide. Several of th e otherlower esters also were exploded easily, but as thetransition was made to t he higher aliphatic esters,these became much less sensitive to such shocksthan benzoyl peroxide. The peroxydicarbonatesprepared f rom the bis-chloroformates of ethyleneand diethylene glycols were polymeric solids orgums possessing stability characteristics similar tothose of the monomeric esters.

Decomposition of most of the higher aliphaticand substituted aliphatic esters which were solids(Table 11,footnotes g-k and m) roceeded a t ratesof 1-2% per day a t 25-30', but was retarded byrefrigeration. Th e bis-(2-nitrobutyl) and bis-(2-nitro-2-methylpropyl) esters were notable ex-ceptions (see following discussion) in that thesecompounds could be kept for weeks or months a t

ordinary temperatures with but little (1-2010 orless) decomposition.Evidence that chain reactions of the radical

type are involved in the decompositions of the es-ters of Type I was provided by the observationthat the tendencies for autoaccelerated decompo-sition are susceptible to moderation or inhibitionby small additions (1-5%) of a variety.of com-pounds, the most effective being known inhibitorsof chain reactions, including polyhydroxy- andnitroaromatics, certain unsaturated compounds,oxygen (or air) and i~ d in e . ~s shown by the d atagiven in Table 111, the inhibiting compoundslargely suppressed the normal autoaccelerated de-

composition of diisopropyl peroxydicarbonate oc-curring a t room temperature, bu t some decompo-sition continued a t different rates with the differ-ent additives. An induction period was often ob-served in the presence of these inhibitors, thelength of which depended upon the amount ofthe inhibitor. Apparently the consumption ofthe inhibitor terminated this induction period af-ter which the autoaccelerated decomposition oc-curred.

The decomposition rates of the esters of TypeI in solutions were strongly influenced by the na-ture of th e solvent. This was illustrated by thelosses in peroxydicarbonate assay values of ap-

(9) H. . Stevens, U. S. Patent 2,415,971, eb. 18, 1947.

TABLE11

INHIBITIONF DECOMPOSITIONF DIISOPROPYL

PEROXYDICARBONATEPeroxy- Assaydicar- loss,

bonate, Time, Temp., av.g . Inhibitor % hr. OC. per br .

3 0 0 . 0 Iodine 1 288* 8-16 0.04"

5 0 . 0 Iodine 1 72b 25-30 - 3 25 0 . 0 Iodine 1 25Sb 25-30 .12d

5 . 0 Trinitroben-

1 . 0 Trinitroben-zene 5 23 30 . 3 2

zene 1 23 30 .33

2 0 . 0 Acetanilide 1 18 30 .53*

10.0 Nitromethane 5 45 25-30 .541 . 0 Ethyl aceto-

acetate 1 24 30 .60

5 . 0 Phenol 5 27 30 .63

5 . 0 Phenol 1 27 30 .8 65 . 0 Tetralin 5 18 25-30 .91

1. 0 None . . 1.00 2 5 . 0 - 2 7 . 3 6 . 8

(I Av. per cent. of original assay lost per hr. b Sampleswere stored in darkness. e A loss in wt. of 23.2 g . occurredduring this tes t. Extensively decomposed solutionscontaining iodine exploded violently on several occasions,especially if stored a t higher temperatures ( 3 0 4 5 " ) .e Violent decomposition occurred af te r forty -two hours.

proximately 30% solutions of the diisopropyl esterin solvents containing different constituent (ethe-noid, aromatic, ester, carbinol) groups. For ex-ample, losses of the original assay values in diethylmaleate, tricresyl phosphate and dibutyl phthal-ate afte r seven days (one hundred and sixty-eighthours) at 25 * 1' amounted to 28.5, 34.8 and91.20/,, respectively. In isopropyl alcohol as sol-vent, the diisopropyl ester was highly unstable,decomposing with violent ebullition of the solutionafter twenty-five minutes. The order of activ ityof a number of the compounds tested in retardingor promoting the decomposition of the diisopropylester (including acceleration by some amines, seeExperimental), when added in small amounts orwhen used as solvents, was similar to the order ofactivity of th e same or like compounds in in-fluencing the rate of decomposition of benzoylperoxide when these are added to its solutions orused as solvents for this peroxide.loa-C,lla Someexceptionsllb were noted, however.

The kinetic data given in Table IV, and repre-sented in Figs. 1 and 2 show th at esters of Type I,even though differing greatly in gross stability inpure form, decompose in dilute solutions at nearlyidentical rates conforming quite closely to first or-der kinetics. These rates are approximatelythirty times the rates of decomposition reportedfor benzoyl peroxide in many solutions in the

(10) (a) Nozaki and Bartlett, THIS OU R N A L , 68, 1686 (1946);

(b) Bartlctt and Nozaki, ibid. , 69, 2299 (1947); (c) Cass, i b i d . , 68,

1976 (1946).

(11) (a) Barnett and Vaughan, J . P h y s . Coll . C h e m . , 61, 926, 942

(1947); (b) for example phenol tended t o inhibit th e decomposition

of the diisopropyl ester (see Table I11 and Experimental) bu t appears

to promote th e decomposition of benzoyl peroxide when used as sol-

vent (ref. lob).



March, 19.50 ESTERS F PEROXYCARBONICCIDS 1257

TABLEV

ATES IX DILUTESOLUTIOKS

RATES F DECOMPOSITIOXF DIALKE'L P E R O X Y D I C A R B O N -

Initialferoxydi- concn., Apparent k , hr . -Icarbonate m,/l , Solvent 40.O00 50 . 0 ° c 60 O o c

Diethyl 0.210" , 0.025 0.103 0 . 4 6

Diisopropyl ,144 Toluene . . . . lo9 . .

Dibenzyl .Os3 Toluene . . . . 10513 . .Bis-(2-N02-2-

MePr) ,118 Toluene . . . .080 .a Concentration, rn./kgd b 2,2'-Oxydiethylene his-

.152" * 0,023 ,082 0.34

(allyl carbonate ). +0.2 .

concentration range represented, at temperaturesof 40-60' . l z a a b , l a The over-all activation energies

for the decomposition of the diethyl and diisopro-pyl esters in the dilute solutions, 30.4 and 28.1

kcal. per mole, respectively, are sufficiently closeto the values of 29-31 kcal. reported for benzoylperoxide12a n solutions of similar concentration toindicate that the primary rate controlling step isthe same for the peroxydicarbonates as for benzoylperoxide, "Jiz., dissociation into acyloxy radi-

0 2 4 6 8 10

Time, hours.

Fig. 1 -Decornposition of esters of peroxydicarhonic

acid diethyl ester: O,W,D , 0.210 mole/kg. in 23'-o xy-

diethylene bis-(allyl carbona te) a t 40.0, 50.0 and 60.0",

respectively (a= 2) ; diisopropyl ester: 0 ,0.144 mole/l. in

toluene a t 50.0' ( a = 3 ) ; O,.',,., 0.152 mole/kg. in 2,2'-

oxydiethylene his-(allyl carbonate) a t 40,0, 50.0 and 60.0°,

respectively (a = 3) ; bis-(2-nitro-2-methylpropyl)ster:

8 , 0.118 mole/l. in toluene a t 50.0" (a= 2).

(12) (a) Refs. loa , b, c, l l a and earlier references cited therein;

(13) Prof.P. D.

Bartlett and Dr. R. Altschul, private communica-

(b) Bartlett and Nozaki, THIS OURNAL, 68, 1497 (1946).

tion.

For the esters of peroxydicarbonic acidthis step is represented by equation 1 (R = H oralkyl, R' = alkyl or subst ituted alkyl)

0 0

il ki I1

Acceleration or inhibition effects upon the de-compositions by accumulated decomposition prod-ucts, such as have been observed in the decompo-sition of benzoyl peroxide in solvents,loa~bppar-ently were largely avoided in these rate measure-ments, probably due to the limited extents (15-60y0) o which the decompositions were allowedto proceed. The conformity of the data to firstorder kinetics, when considered in conjunctionwith the inordinately high rates of decompositionof the lower alkyl peroxydicarbonates in pureformI5and in some of their concentrated solutions,

(RR'CH-O-C-Ojs + RR'CH-0-C-0- (1)

2.6

2.2

.&2 .80-+?:

1.4

1.0

0.6

2.80 3.00 3.20

I / T (OK.)x 103.

Fig. 2.-Decomposition of esters of peroxydicarbonic

acid and benzoyl peroxide in solvents, effect of temperature

0 = diethyl ester, 0 =i diisopropyl ester, 0.210 and 0.152

mole/kg., respectively, in 2,2'-oxydiethylene bis-(allyl

carbonate); 0 = benzoyl peroxide, 0.18-0.19 mole/kg. in

allyl acetatelZb; O = benzoyl peroxide, 0.07-0.10 mole/kg.

in benzene11a.

(14) (a) See refs. lo a, b, c and 1 1s; (b) Bartle tt and Altschul,

THIS OURNAL, 7 , 813,819 (1945); (c) Price and Kel!, ib id . , 63, 2800

(1941).

(15) Extrapolation of the curve for diisopropy! peroxydicarbonate

i n Fig. 1 indicates values for k a t 25 and 30°, espectively, of 0.003

an d 0.006 hr.- l . The decomposition data for the pure ester at

25.0-27.3O (last entr y, Tabl e 111) show the decomposition in the

latter case to be at least twenty times as rapid as would be predicted

on he basis of these extrapolated values of k.



1258 STRAIN,ISSINGER,IAL,UDOFF,DEWITT,STEVENSND LANGSTON Vol. 72

shows that reactions of higher than first order areinvolved in the chain decompositions. The nearlyidentical over-all decomposition rates of the estersof widely differing gross stabilities in pure form(Table IV) further show that the differing specificchain decomposition reactions responsible for

these differing gross stabilities have been reducedto minor roles in the decompositions due to dilu-tion in the solventsloa-C~llased. It is probablethat a t the apparent over-all first order decompo-sition rates (Table IV), containing the contribu-tions from the true unimolecular dissociation(kl, quation 1) still include appreciable contribu-tions from the higher order chain reaction (whichmay involve the solvent) analogous to the higherorder contribution ( K i ) lo a reported for the chaindecomposition of benzoyl peroxide in solvents.

The products recovered from the thermal de-composition of pure diethyl and diisopropyl per-oxydicarbonates by slow passage of the estersthrough a steam-heated tube contributed furtherevidence that these decompositions involve freeradicals. Ethyl alcohol and acetaldehyde (recov-ered as paraldehyde) were identified as decompo-sition products of the ethyl ester, while a morecomplete analysis of the liquid and gaseous prod-ucts recovered from decomposition of the diiso-propyl ester gave the results summarized in TableV.

TABLE

PEROXYDICARBONATETHERMALECOMPOSITIONR OD U C TS OF DIISOPROPYL

Moles p w mole

Product of peroxydicarbonateAcetaldehyde

Acetone 0.42

Isopropyl alcohol 0. 91

Carbon dioxide 1.61

Ethaneb 0.41

Other gases‘ 0. 11

5

a Since some loss of acetaldehyde occurred, the initialexperiment was in part duplicated giving 0.45 mole ofacetaldehyde, 0.43 mole of acetone and 0.61 mole of iso-propyl alcohol per mole of peroxydicarbonate. * Satu-rated hydrocarbon , mol. wt. 26.4; ethane qualitative lyconiirmed by chlorination to hexachloroethane. Oxygen,carbon monoxide and olefins.

Possible modes of chain decomposition for thedialkyl peroxydicarbonates capable of accountingfor the available data, with initiation by radicalsresulting from the initial dissociation (equation 1)and analogous to formulations proposed for thechain decomposition of benzoyl peroxide,‘oa-Care

I / It

0 0

RR’CH-0-C-0- + (RR’CH-O-C-O)2 +

0

ItRR’CHOH + RR’C=O + 2C02 + RR’CH-0-C-0-

(2)0

IIRR’CH-0-C-0- d R’CH-0- + COP (3)

The alkoxy radicals resultirig from reaction 3,lacking a capacity for resonance stabilization incontrast to the strong resonance possible in thealkylcarbonate radicals, may be especially reac-tive in attacking the peroxydicarbonate (equation4) or other molecules formed as decomposition

0/I

RR’CH-0- + (RR’CHO-C-O)* +RR’CHOH + RR’C=O + 2C02 + RR’CH-0- (4)

products, or may decompose further, e. g. , byloss of an alkyl fragmentlBaSbor hydrogen atom,to form a carbonyl compound, i. e .

RR’CHO- + - + R’CH=O ( 5 )

The recovery of acetaldehyde along with thesaturated hydrocarbon gas consisting largely ofethane (Table V) as products of the decompositionof diisopropyl peroxydicarbonate constitutes ‘evi-dence for the formation of isopropoxy radicals

(equation 3) followed by their decomposition ac-cording to equation 5, as well as for the possiblechain terminating combination of the (methyl)radicals,16b . e.

2R- + -R (6)

The products of the decompositions of both thediethyl and diisopropyl esters suggest, further-more, that, as an additional chain terminatingstep, disproportionation of the alkoxy radicals(equation 3) may occur as follows

2RR’CH-0- + R’CHOH + RR’C=O (7 )

The wide differences n stability observed among

the different esters of Type I in pure form are thusclearly due to the different extents to which thespecific radicals produced in the respective de-compositions attack the peroxydicarbonate mole-cules (equations 2, 4 and other similar possiblereactions of radicals derived from the decomposi-tion products with the peroxydicarbonate), andthereby propagate the chain decomposition. Whiledilution with solvents, i. e., reduction of the con-centration of the peroxydicarbonate, would be ex-pected to reduce the proportion of the decomposi-tion occurring by the higher order chain mechanism(see preceding text), the profound influence of sol-vent character on decomposition rate must be at-

tributed to participation of the solvent in the chainreaction. This may involve transfer, propagationor termination reactions, in a manner analogous tothose observed upon decomposition of benzoyl per-oxide in different solvents.loa-C~llaIn the sensethat these influences associated with solvents areprovided by and vary with the differing structuresof the alkyl groups of the peroxydicarbonates, t hecompounds themselves may be regarded as theirown solvents. It is of interest in this connection to

(16) (a) This has been suggested for ethoxy radicals, see Steacie.

”Atomic and Free Radical Reactions,” Reinhold Publishing Corpora-

tion, New York, N. Y., 948, p. 141; (b) compare similar decomposi-

tions found for tertiary alkoxy radicals: Milas and Surgenor, THIS

JOURNAL, 68, 205, 64 3 (1946); Raley, Rust and Vaughan, i b i d . , 70,

88, 1336 (1948); Rust, Seubold and Vaughan, ibid., 70, 95 (1948).



March, 1950 ESTERSF PEROXYCARBONIC ACIDS 1259

note the increasing gross stability with increasingsize of the alkyl group above ethyl (Table II), aswell as the correlation of the high stabilities of thebis-(2-nitro-2-methylpropyl) nd dibenzyl esterswith the improved stability of the unstable diiso-propyl ester in the presence of small amounts ofnitromethane, toluene and the bis-(2-nitro-2-methylpropyl) ester itself (Experimental).

As illustrated by isopropyl peroxydicarbonate(Experimental), many of the esters listed in TableI1 were effective catalysts for polymerization oftypical ethenoid monomers at moderate tempera-t u r e ~ . " ~ ' ~ ome of the esters, notably the bis-nitroalkyl and dibenzyl esters, although activecatalysts, produced somewhat less polymerizationin allyl esters1*per active oxygen equivalent con-sumed than the others, indicating either somewastagelg in reactions not producing polymeriza-tion or some concurrent inhibiting action resulting

from use of these particular peroxydicarbonates.Esters of Types I1 and 111.-Reaction of iso-propyl chloroformate and phosgene, respectively,with t-butyl hydroperoxide in the presence ofacid acceptors readily produced examples of theesters of Types 1120and 111, namely, 00-t-butyl0-isopropyl peroxycarbonate in the first case anddi-t-butyl diperoxycarbonate n the second. Thesecompounds were stable a t room temperature] ap-pearing to be comparable in stability to the simi-lar reported t-butyl peroxyesters of organic car-boxylic acids.21 The ester of Type I1 was muchless sensitive to impact than that of Type 111.

The gross thermal stabilities of the pure esters

of Types I1and 111,as indicated by minimum tem-peratures required for steady gas evolution (104-106" and 86-88', respectively)22 were muchgreater than for the pure unsubstituted alkyl es-ters of Type I (see Table 11). Dilution with ni-trobenzene (concentrations 1.00 M ) also affected(increased) these minimum temperatures for thecompounds of Types I1 and I11 less than for anunsubstituted dialkyl ester of Type 1 (respectivetemperatures 117-118", 85-86' and 63-65' for di-

isopropyl peroxydicarbonate) u indicating thatthe tendencies for chain decomposition in the purecompounds of Types I1 and I11 were less than inthe pure unsubstituted ester of Type I.

The compounds of both Types I1 and I11 were(17) A. Pechukas, U. S. Patent 2,464,056, Mar. 8, 1949; F.

Strain, U.S. Patent 2,464,062, Mar. 8, 1949.

(18)Unpublished observations, this Laboratory. Experiments

show that in many cases polymerization in the presence of equimolar

concentrations of peroxydicarbonate or benzoyl peroxide proceeds

at approximately equal rates if the temperature is about 25O lower

in the peroxydicarbonate solution than in the benzoyl peroxide solu-

tion, i. ., if the rates of decomposition of the peroxy catalysts are

approximately equal (compara data given in Fig. 2).

(19) Re f. 14b p. 820; ref. 10a p. 1691; ref. 10b p. 2300; Cohen,

THISJOURNAL, 67, 19 (1945); Redington, J . Polymer Sci., 8: 504

(1948).

(20) F. Strain, U. . Patent 2,374,789, May 1 , 1945.

(21) Milas and Surgenor, THISOURRAL, 68,642 (1946).

(22) Observed in 4-ml. test-tubes; see Experimental.

(23) The Corresponding temperature observed for a 1.00 &f

In theohtion of benzoyl peroxide in nitrobenzene waa 87-89..sbmnc. d rdvmat ~mntalline easRJ1 pwaridt mxplodrd at 1Wo.

also active polymerization catalysts for typicalethenoid monomers.24

Experimental

Preparation of Chloroformates.-The alcohols em-ployed for the preparation of t he corresponding chlorofor-

mates, except for neopentyl alcohol and 2-hydroxyethylcarbamate described below, were Eastman Kodak Co.white or yellow label chemicals, or technical products ifavailable; they were used directly if a check of physicalconstants indicated high purity ; if not they were redis-tilled before use.

Neopentyl alcohol was prepared by the action of form-aldehyde upon t-butylmagnesium chloride.*s Th e re-action mixture was poured into ice-water, the resultingsolution acidified with dilute sulfuric acid, and extractedwith ether. After washing and drying over sodium sulfatethe extract was distilled, giving 2?% of th e theoreticalof neopentyl alcohol, b. p . 109-114 .

2-Hydroxyethyl carbamate was obtained by allowing amixture of 88 g. (1.0 mole) of ethylene carbonate*' and80 ml. (1.2 moles) of concentrated aqueous ammonia tostand overnight at room temperature. After heating to

constant weight on a steam-bath a t 3-mm. pressure, thereremained a 98-g. residue (93% yield) of liquid product;n% 1.4626, azo,1.289; mol. refraction: calcd. for 2-hy-droxyethyl carbamate, CaHlOsN: MZOD 21.5; observed:M*OD 22.4.

Methyl, ethyl, n-propyl and isobutyl chloroformateswere the Eastman Kodak Co. white or yellow label chemi-cals. Procedure A, following, was employed for prepara-tion of cyclohexyl,gbbenzyl,6bmethoxyethyl,'b acarbeth-oxye th yP and 2-chloroethyleb chloroformates and for t heremaining chloroformates listed in Table I, except forthose in which the somewhat more efficient Procedure Bwas followed, as noted in th e Table.

Procedure A.-An excess (10-20% over th e theoretical)of phosgene was introduced with vigorous stirring into thealcohol, maintaining the temperature somewhat above th eb. p. of phosgene (a t 10-15") by adjustment of th e ra te of

phosgene introduction and by use of an ice-salt-bath, withsome cooling by reflux of phosgene from a Dry Ice-acetonecooled reflux condenser. The cooling bath was then re-moved and s tirring continued for one-half t o one and one-half hours while allowing the temperature to rise to that ofthe room. Hydrogen chloride and phosgene remainingwere removed by blowing dry air through the mixture, exitgases being absorbed in methanol-sodium hydroxide traps.With the less volatile chloroformates this was followed byheating to 65' a t 9-15 mm. pressure for approximatelyfifteen minutes. The products were usually washedseveral times with cold water to remove alcohol, and driedover calcium chloride. They were then distilled for finalpurification unless analysis showed high purity, or unlessa product of lower purity tended t o result due to some de-composition on distillation. Boiling point da ta are in-cluded in Table I.

Procedure B.-A small pool of liquid phosgene was con-densed in a reaction flask after which the alcohol was in-troduced slowly with vigorous stirring while continuing thephosgene addition a t such a ra te t ha t some excess phosgenecontinued to reflux from a Dry Ice-acetone cooled refluxcondenser through which hydrogen chloride was vented.Cooling was thus effected by the refluxing phosgene, thetemperature of the reaction mixture soon rising above theboiling point of pure phosgene and being maintained in

This catalytic

activity was developed at somewhat higher temperatures than for

benzoyl peroxide as suggested by the thermal stability data (see

also ref. 23) . The ester of Type I1possessed high catalytic efficiency

in the polymerization of allyl esters but that of Type I11 wan less

efficient; see refs. 18 and 19.

( 2 5 ) Conant, Webb and Mendum, THIS JOURNAL, 61, 1846

(26) Nemirowsky, 1.@&. Chcm., [2 ] 48,439 ( l8W .(27) Tbit le nnd Pen*, A+-, , 199, 245 (1898).

(24) Unpublished observations, this Laboratory.

(1929).



1260 Vol. 72

th e range of 10-20'. Stirring was continued in the pres- uct was then ~ep ara ted ,~Ond washed with cold distilledence of some excess (refluxing) phosgene for one-half to water a t 5 to 10" until free from chloride ion,31 then driedone and one-half hours while gradually allowing this ex- by addit ion of cold anhydrous sodium sulfate. Cooling wascess of phosgene to escape and the reaction mixture to applied as necessary to maintain the temperature in thiswarm to room temperature. The reaction mixture was range. After filtration from the drying agent, the diiso-then degassed by blowing with dry ai r and heating under propyl peroxydicarbonate was obtained as a clear, colorlessreduced pressure as in Procedure A, after which the crude oil weighing 83-92 g. (81-89c1,). The liquid diisopropyl

products were purified by distillation, or used directly as ester froze readily on cooling (m. p. 9-10' when pure ),in Procedure A. and was stored conveniently in the frozen state (D ry Ice ).Except where noted in Table I the chloroformates were When melted carefully, in the presence of some solid, it

analyzed for chloroformate chlorine by dissolving a weighed could be handled readily without appreciable thermal de-sample (0.2-0.5 8.) contained in a small vial into 20 ml. of composition. Spontaneous heating could be observed in aa freshly prepared aqueous reagent solution containing mass of the liquid ester if allowed to warm toward roonl2.5% sodium hydroxide and 10% pyridine by weight. temperature, and unless strong cooling was applied dangerAfter agitation of the solution for approximately twenty of an uncontrollable, violent, exothermic decompositiollminutes, or until all of the chloroformate phase had dis- was present.appeared, the solution was acidified with dilute nitri c acid In cases in which the chloroformate or percarbonate wasand the chloride ion determined by the Volhard method. a solid, a volatile organic solvent, preferably one which

Phenyl chloroformate, formed readily by th e reaction could be washed out of the product by cold water or re-of phenol with phosgene only a t elevated temperatures,z88 moved by vacuum evaporation, such as ether or ethyl ace-was prepared more conveniently by the addition of di- ta te , was added as necessary to the reaction mixture beforemethylaniline,a*b12.6 g. (0.10 mole), to a stirred solution or during th e preparations to maintain it in the liquid state.of 9.4 g. (0.10 mole) of phenol and 9;9 g. (0.10 mole) of Following the above indicated procedure 3.1 g. (0.02phosgene in 40 g. of benzene a t 5 to 10 . The oily product mole) of phenyl chloroformate gave a solid product, 0.15

obtained by dilution of the reaction mixture with cold g., m. p. 74-75", which contained no active oxygen butwater was washed with dilute hydrochloric acid, water, was diphenyl carbonate ; mixed m. p. with an authenticand dried over calcium chloride. Distillation gave 9.0 g. sample of diphenyl carbonate (m. p. 78") was 74'. A(58% yield) of phenyl chloroformate, b. p. 74-75' (13 similar preparation from dodecyl chloroformate gave anmm.), n% 1.5131. impure waxy peroxydicarbonate containing dodecanol

Anal . Calcd. for C7H&)zCl: C1, 22.6. Found: Cl, which could not be removed by washing. The wax lique-21.4. fied 0: warming and decomposed with effervescence at

Esters of Peroxydicarbonic Acid.-The esters listed in 60-65 *

Table I1 were prepared by a general procedure consisting The peroxydicarbonates in most cases were analyzed forof careful addit ion of a cold solution of aqueous sodium active oxygen by solution of 0.1-0.4 g. in 50 ml. of absolute

peroxide to the well-stirred and cooled chloroformate. acetone, addit ion Of 5 ml. of 20% aqueous acetic acid con-Initial preparations w s e usually conducted behind safety taining 2 g. Of dissolved potassium iodide, .and after fiveScreens on a scale equivalent to theoretical yields of 0.02- minu tes standing with occasional shaking, tit rat ion of the0.10 mole of the peroxydicarbonate in order to observe liberated iodine with standard sodium thiosulfate solution.properties such as stability of the product to heat and Some samples were also analyzed employing a different

shock, without attempting to obtain optimum yields, andin order to evaluate xp~osionhazards which might be Peroxydicarbonates of Bis-chloroformates.-Following

presented by the larger scale preparations. Details of the the general Procedure employed fo r Preparation of th e es-preparative procedure, essentially the Same for the smaller ters listed in Table I, 2,2 -oxydiethylene bis-chloroformatescale as for the larger preparations, are illustrated by the (i. e .? diethylene glycol bis-chloroformate) gave a whitefollowing preparation of diisopropyl peroxydicarbonate : Polymeric solid peroxydicarbonate which was only par-

A sodium peroxide solution was prepared by adding 68.2 tially soluble in acetone and which decomposed slowly

g. (0.55 mole) of 27.4% hydrogen peroxide (Albone, ana- forming a gummy mass on warming to room temperature.

lyzed iodometrically immediately before use) to a cooled A similar product was also obtained from ethylene bis-solution of 44.0 g. (1.10 moles) of C. p. sodium hydroxide chloroformate. A solid polymeric peroxydicarbonate pre-

in 300 ml. of water, keeping the temperature in the range pared as fOllOWS from a mixture of ethylene bis-chlorofor-

of 10-15'. Some solid sodium peroxide octahy&ate sepa- mate and ethyl chloroformate was mostly, but not com-rated ou t bu t was easily suspended by stirring the solution. pletely$This procedure for preparation of the sodium peroxide To a mixture of 15.0 g. (0.080 mole) of ethylene bis-solution superseded one employed in th e earlier prepara- chloroformate and 5.00 g. (0.046 mole) of e thyl chloro-tions involving a solution of A. R. odium peroxide in formate was added with vigorous stirring a cold solutionwater, and avoided the greater care required in the latter of 9.0 g. (0.11 mole) Of sodium peroxide dn 105 ml. of cold

method to prevent overheating, some decomposition and water, keeping the temperature near 0 . The resulting

formation of large crystal masses of difficultly dispersible suspension was then st irred for one-half hour, th e solid

octahydrate. filtered of f and washed several times with cold distilled

The sodium peroxide solutionz9 sus- water. The whit; solid after drying overnight in air atpended octahydrate was then added slowly through a approximately 20 , weighed 6.9 g., equivalent t o a 43%dropping funnel to 122.6 g. (1.0 mole) of vigorously yield of product on the basis of the formula CzHsOCOO-stirred isopropyl chloroformate, cooling with an ice-salt ( ° C O O ~ H ~ O C O O ) ~ ~ C ~ O C ~ ~ ~ssuming the averagemixture and adjusting the rate of addition so that the re- value of x to be 3 . 5 , the theoretical valuawhich would re-action temperature was maintained at 6 to 100,but never sult from complete reaction of the two chloroformates.allowed to rise higher. Lower temperatures tended to Anal . Calcd. for CnoHzrOzr: active 0 , 10.34. Found:cause solidification of the oily product suspended in the active 0, 0.42.reaction mixture while danger of thermal decompositionof this product was present at higher temperatures. (30) It was frequently desirable to make certain of complete re-

Stirring was continued for one-half hour after a l ~f the action of the chloroformate before making this separation by addi-

aqueous pyridine followed by acidification with nitric acid and quali-

(28) (a) Kempf, J . p r o k f . Chcm., [ 2 ] 1, 403 (1870); (b) Oesper, tative test for chloride ion.

Broker and Cook, THISOURNAL, 47, 2609 (1925); German Patent (31) In some preparations traces of unreacted chloroformate

261,805 (1912). made this di5c ult; one or more washes with cold, dilute (e. p. . 5%)

(29) A slight (e. I . , -10%0) excess of sodium peroxide was desirable aqueoua pyridine were then employed for removal followed by addi-

in the preparation in order to assure complete rrwtion of the chloro- tional r ar h u with cold distilled water.

formats.

STRAIN, ISSINGER,DIAL,RUDOFF , EWITT,STEVENSND LANGSTON

without Significant alteration of the results.

in acetone:

sodium peroxide solution had been added, The oily prod- tion Of a Sample Of the filtered and washed Oily product to Cold 20%

(82) Noraki, Znd. Bng. Chrm. , A n d . E d . , 18, 688 (1940).



March, 19.50 ESTERSF PEROXYCARBONICCIDS 1261

00-&Butyl 0-Isopropyl Peroxycarb0nate.-A mixtureof 63.8 g. of 58% t-butyl hydroperoxide (0.41 mole, analy-sis by titanous chloride procedure outlined below) (UnionBay S ta te Co.) and 50.9 g. (00.41 mole) of isopropyl chloro-formate was cooled to -5 . A cold solution of 17.2 g.(0.43 mole) of sodium hydroxide in 160 cc. of water wasthen added dropwise with vigorous s t p n g and cooling to

maintain the temperature at -5 to 0 . After stirring themixture for thirty minutes the upper oily layer was sepa-rated , washed with distilled water unt il free from chlorideion, then dried over sodium sulfate. Distillation in threeportions through a 1 X 8-cm. Vigreux column gave 58.5 g.(81%) of colorless 00-t- butyl-0- iso prqyl peroxycarbon-ate, b. p. 52-55' (1 mm.), T Z ~ O D 1.4050, 0.966. Theproduct was analyzed for active oxygen using standard0.03 N itanous chloride reagent, employing glacial aceticacid to maintain homogeneity of the solution of peroxidein the aqueous reagent, and heating at 60-65' in a carbondioxide atmosphere for ten t o twenty minutes before backtitra tion of unused reagent with standard ferric alumsolution.

Anal. Calcd. for 00-t-butyl 0-isopropyl peroxyclr-bonate, CeH1e04: active 0, 9.08; M a o ~peroxide 0 =2.1933),44.65. Found: act ive 0,9 .05 ; M m ~ ,4.52.

Di-t-butyl Diperoxycarb0nate.-Phosgene, 0.080 mole,was introduced a t a ra te of six millimoles per minute in toa sti rred solution of 18.0 g. of 71 % t-butyl hydroperoxide(0.14 mole) and 16.6 g. (0.21 mole) of pyridine in 40 ml.of ether, cooling to keep the temperature at 3-5". Theresulting mixture was allowed to stand in the cold forforty-five minutes after the phosgene addition was com-plete, then diluted with 100 ml. of ice-water. The etherlayer was separated, washed once with cold water, twicewith cold 6% aqueous hydrochloric acid, then again withcold water until free from chloride ion. The ether solutionwas then dried over sodium sulfate, filtered, and evaporatedto constant weight a t 20-mm. pressure. The yieldof colorless di-t -butyl diperoxycarbonate remaining was11.0 g., 76%; T Z ~ O D 1.4106. The b. p. of the product wasabove 95" (4 mm.); decomposition with gas evolutionoccurred under these conditions which prevented satis-

factory distillation. In one experiment a partially de-composed sample being heated a t 135-140' exploded withviolence.

Anal. Calcd. for CsHlsOs: active 0, 15.5. Found:active 0 (titanous chloride method; cf. preceding prepa-ration), 14.9.

Stability of Peroxycarbonate Esters.-The gross the r-mal stability of the liquid peroxydicarbonate esters wasobserved by raising the tempera ture of 1ml. of the esterin a 4-ml. (10 X 75-mm.) test-tube (behind a protectivescreen) at a ra te of approximately one degree per minuteusing a water-bath with a thermometer immersed in theester. The minimum tempera ture range in which bubblingof the liquid with gas evolution became readily observableand steady was determined. This temperature rangedepended upon the efficiency of hea t transfer and ra te ofheating, nevertheless, was quite narrow and reproduciblefor the different compounds under these standardized con-ditions. This temperature range is recorded as the grossdecomposition temperature in Table 11. In many casesthe autoaccelerated exothermic decomposition soon raisedthe temperature above the minimum values given; forexample, in one case the temperature recorded by thethermometer quickly reached 124' in a test using diethylperoxydicarbonate. The decompositions, especially ofthe lower aliphatic esters, thus tended to become explosive.In a few cases, indicated in the table, sharp explosionsoccurred at the temperatures noted.

Decomposition temperatures for crystalline peroxydi-carbonates melting above room temperature (25') andrecorded in Table I1 were observed in the m. p. capillarytubes (Kimble Glass2No. 34505) on continuing t o raisethe temperature at a ra te of apptoximately one degree perminute after melting had occurred. Decomposition

(88) Milaa, Surgenor and Perry, THIS OVRNAL, 88, 1617 (1946).

temperature ranges observed in this manner were usuallyhigher, by up to 10-15", th an those observed in 4-ml. test-tubes as described above. the temDerature observed in acapillary for diisopropyl peroxydi&bonate being for ex-ample 47-51' (Table 11).

The extent of decomposition on standing of samples ofperoxydicarbonates which were solids at room temperature

was determined by iodometric assay for active oxygen.Samples, a few grams in size, of each of th e following per;oxydicarbonates were stored at 25-30' (except a t 10where indicated) in loosely-closed glass bott les; initialassay values, in per cent. of t he theoretical active oxygenfor each ester, are followed by th e assay values found afterthe different respective periods of standing: dineopentyl98.4, one day 96.0; dicyclohexyl99.5, three weeks 40-50,one week (10") 98.0 ; dibenzyl 98.0, one day 96.513;bis-(2-~arbamyloxyethyl)a) 88.9, four weeks 51.3, (b)92.4, eight weeks (10') 88.9; bis-(2-nitrobutyl) 99.1,twelve weeks a t 10" followed by two weeks 97.4; bis-(2-nitro-2-methylpropyl) (a) 98.6, five weeks 98.6, (b) 99.4,one year (10)' 7.0.

A preliminary indication of th e sensitivity of the per-oxycarbonate esters to impact was obtained by placing afew drops or crystals of the compound on a clean flat steel

plate followed by striking blows of increasing force with asteel hammer. Dimethyl peroxydicarbonate was easilyexploded in this manner, and the diethyl compound withmore difficulty, but the higher esters gave less evidence ofsensitivity, and were much less sensitive t o such impactsthan benzoyl peroxide. 00 -t- Butyl 0-isopropyl peroxy-carbonate likewise exhibited little sensitivity to suchshocks, but di-t-butyl diperoxycarbonate could be ex-ploded easily in this manner.

The sensitivity to explosive shock of d iethyl and diiso-propyl peroxydicarbonates was observed34 by explodingno. 6 electric detonating caps (containing 1.5g. of mercuryfulminate) in hollow chambers on the interior of frozen(estimated temperatures -20 to -50" ) 0.25-lb. chargesof these esters contained in open metal cans, 26/8 in. indiameter by 4 in. high. The report produced by the di-ethyl ester was audibly sharper than that produced by anequal amount of benzoyl peroxide (Lucidol) (tempera tureapprox. 10") ,while the report produced by the diisopropylester, which was only partially exploded was milder thanthat produced by benzoyl peroxide.

Inhibition and Acceleration of Decomposition of Diiso-propyl Peroxydicarbonate.-When enough liquid pure di-isopropyl peroxydicarbonate was placed in small (4 to15-ml.) test-tubes t o occupy about one-third of the volumeof t he test-tube and allowed to stand a t 25-30' free fromdrafts, the ester heated spontaneously and decomposed withviolent effervescence after a period of ten to thi rty minutes.Various added chemicals, in the amounts indicated asfollows based on the ester, (a) affected but little, (b), (c)lengthened, or (d) shortened this period:

(a) Chemicals affecting period before decompositionbut little (five to fifty minutes) : water 1%, acetone l% ,methyl alcohol 1%, ethyl alcohol 1%, isopropyl alcohol1%, t-butyl alcohol 1%, acetic acid 1% , ropionaldehyde1%, benzyl alcohol 1%, ethyl acetate 1% ,acetamide 1% ,diisopropyl ether 1yo,dioxane 1%, diethyl malonate 1yo,acetylacetone 1% , riacetin I%, benzene 5% , pyridine 1% ,benzoyl peroxide 1%, 10% aqueous hydrochloric acid 170,nitrogen bubbled continuously through liquid, carbondioxide bubbled continuously through liquid.

(b) Chemicals extending period before decompositionto one to ten hours: toluene lY0, cyclohexene 2%, allylalcohol 1%, dicyclopentadiene 570, 8-phenylethyl alcoholl%, phenetole 1% , salicylic acid 1%, naphthalene 570,30% aqueous hydrogen peroxide 1%, lead tetraethyl 0.470,nickel carbonyl 5%, sulfur 1%, sulfur dioxide bubbledthrough liquid for thirty minutes, bis-(2-nitro-2-methyl-propyl) peroxydicarbonate 2%.

(c) Chemicals extending period before decompositionto twenty hours or longer:, phenol 1%, hydroquinone 1% ,resorcinol 1% ,pyrogallol 1%, tetralin 1% ,diphenyl ether

(84) Teat8 conducted w i t h the amristaac@ f Mr. . . Fredunburg.



Vol. 72282 STRAIN, ISSINGER,IAL,RUDOFF, EWITT,STEVENSND LANGSTON

5% , ethyl acetoacetate l% , acetanilide 1%, phenetole570,salicylic acid 5% , n itromethane 2%, trinitrobenzeneI%, iodine 1%, air or oxygen bubbled continuouslythrough liquid.

(d) Chemicals shortening period before decomposition(to respective times indicated) : aniline 1%, 5-10 sec.;dimethylaniline 1%, instant (explosion); ethylenedi-

amine 1%, nstant; potassium iodide lye, instant; 10%aqueous potassium hydroxide, less than 60 sec.To quantities of liquid diisopropyl peroxydicarbonate

(assay 98.0-99.8%) were added small amounts of some ofthe more efficient inhibiting compounds listed above, afterwhich the solutions were stored in loosely stoppered glassbottles in diffuse light at room temperature, unless otherwiseindicated (Table 111). Samples were withdrawn after dif-ferent periods of time for assay of th e liquid for peroxydi-carbonate by iodometric analysis for active oxygen. A1.00-g. sample of the diisopropyl ester containing no in-hibitor was allowed to decompose in a 9-ml. (13 X 100-mm;) test-tube immersed in a water-bath kept at 25 f0.1 , Air was displaced from thi s tube using carbon diox-ide, and the tube was closed by a clean cork stopper car-rying a thermometer with bulb immersed in the ester. Asmall groove was filed in th e cork for release of pressure de-veloping in the test-tube. The temperature of th e inhibi-tor-free ester, brought rapidly from 10 to 25.0' a t the startof the experiment as the tube was immersed in the bath,reached 25.5" within four minutes , and rose gradually to27.3" a t the end of 1.00 hr. Small bubbles of gas wereevolved slowly after approximately fifteen minutes. Theperoxydicarbonate conten t of the liquid was determined byiodometric analysis a t th e beginning and end of t he experi-men t; respectiveassays were 98.0 and 91.3%. Table I11summarizes the results.

Decomposition of Peroxycarbonates in Solvents.-Thegross stabilities of the peroxycarbonate esters of t he threedifferent types in 1.00 M nitrobenzene solutions were ob-served by raising th e temperature of 1.0-ml. samples in4-ml. test-tubes, observing the minimum temperaturerange in which bubbling of th e solutions due to gas evolu-tion resulting from decomposition became steady in thesame manner as described previously for observation of thegross stabilities of the pure esters of peroxydicarbonicacid. The corresponding temperature ranges for pure00-t-butyl 0-isopropyl peroxycarbonate and di-t-butyldiperoxycarbonate were observed in the same manner.

Ten-gram quantities of solutions containing approxi-mately 30% by weight of diisopropyl peroxydicarbonatein several solvents were stored in loosely stoppered glasscontainers in air a t 25 * 1 . The peroxydicarbonate con-te nt s of th e solutions were determined by withdrawal ofweighed samples for iodometric assay a t the beginning ofth e storage period, and af ter periods of twenty-four andone hundred and sixty-eight hours. Init ial, twenty-four-hour a nd one hundred and sixty-eight-hour assay values,respectively, for the different solvents were as follows:diethyl maleate, 27.7,27.5,20 .0; tricresyl phosphate 27.6,25.6, 18.0; tetralin 29.4, 23.7, 8.6; tetrachloroethylene,30.6, 28.5, 7.0; dibutyl phthalate 28.8, 24.6, 2.5; bis-

2-(2-butoxyethoxy)-ethyl maleate , 26.2, 15.1, 0.4. Al-though a solution in isopropyl alcohol decomposed vio-lently after twenty-five minutes, a solution in tricresylphosphate-isopropyl alcohol (1 1) mixture gave thefollowing respective assay values: 29.8, 13.2, 0.4. Addi-tion of 1% of (a ) iodine or (b) hydroquinone to the diethylmaleate solution gave as respective assay values: (a )29.2, 25.0, 17.9; (b) 27.0, 21.1, 12.6.

For the decomposition rates of the diethyl, diisopropylan d bis-(2-nitro-2-methylpropyl) esters in more dilutesolqtions (Tab le IV, Figs. 1 and 2 ) , redistilled toluene and2,2 -0xydiethylene bis-(allyl carbonate) prepared as de-scribed in the following section were used as solvents.Two-ml. samples of the toluene solutions or 2-g. samplesof th e bis*(allyl carbonate) solutions of the peroxydicar-bonates were heated in a nitrogen atmosphere in individual,

(36) Experiment# Y D L I ~ ~ bio E9lt-m perfmrncd bt Ad; Jehn 0 ,

Pur.leromm,

closed, Pyrex tubes (10 mm. X 75 mm.) using a water-bath controlled a t 50.0' by a thermostat. Separate testsusing a thermocouple showed th at the temperature withinthe tubes did not differ from tha t of the ba th by more than0.1'. Tubes were removed after different lengths of t ime,cooled, opened and the total peroxydicarbonate remain-ing in each determined iodometrically by the same pro-

cedure as used in analysis in preparation of the esters.Insoluble polymer gels from the diallyl compound werecrushed and extracted with 50 ml. of absolute acetone fora t least two hours in the dark a t 10'. Undissolved gelparticles were then filtered off and washed with a smallamount of acetone which was combined with the filteredextract for analysis.

Thermal Decomposition Products of Peroxydicarbon-ates.-Diethyl and diisopropyl peroxydicarbonates weredecomposed by dropping slowly from an ice-water-cooled,jacketed dropping funnel into an adapting tube attachedto an inclined p t - t y p e condenser steam-heated at ap-proximately 100 . A collection train attached to the endof the heated condenser for collecting the decompositionprpducts consisted of an ice-water-cooled trap , a Dry Ice-acetoneqooled trap and a gasometer for collecting uncon-densed gases over saturated aqueous sodium sulfate solu-tion. Liquid products were separated by fractional dis-tillation, while gaseous products were analyzed by theOrsat method.

Decomposition of 5.0 g. (0.028 mole) of die thy l peroxy-dicarbonate in the above indicated manner, without col-lection of uncondensed gases, gave 2.3 g. of distilled liquidproduct which consisted of approximately equal parts ofethy l alcohol (b. p . 78-80"2 and paraldehyde (dinitro-phenylhydrazone m. p. 143 , mixed m. p. with an au-thentic sample of the hydrazone from paraldehyde of m. p.145' remained at 145").

Decomposition of 46.1 g. (0.224 mole) of diisopropylperoxydicarbonate gave a total of 22.7 g. of liquid collectedin the two t raps and 12.3 liters of uncondensed gas, meas-ured at 34" and 741 mm. Most of t he liquid product(19.2 g.) was distilled a t 741 mm. using a 0.7 X 33-cm.column packed with 6-mesh carborundum to give thefollowing fractions: (a) 0.5 g ., plus some loss, bb p. to35', ~ Z ~ D.3383-1.3434; (b ) 4.6 g., b. t . 47-58 , nZoD

1.3776; (d) 1.7 g. of tarry residue. Fraction ( a) gavea dinitrophenylhydrazone, m. p. 143-144'; mixed m. p.with an authentic derivative preparoed from acetaldehyde(of m. p. 146-147') w y 144-145 . Although most offraction (b) boiled a t 56 , preparation of a pure dinitro-phenylhydrazone proved difficult, probably due to thepresence of some acetaldehyde (positive fuchsin-aldehydetest) ; the dinitrophenylhydrazone was finally obtained inpure form from a corresponding fraction in a duplicationof part of this experiment (see below), however, by threerecrystallizations f rom absolute alcohol coontaining a t raceof hydrochloric acid, m. p. 124.5-125.5 ; a mixed m. p.with an authentic derivative prepared from acetone ofm. p. 124.5-125.5' exhibited no depression. Fraction(c) gave a 3,5dinitrobenzoate, m. p. 120"; mixed m. p.

with a known sample of isopropyl 3,5din itrobenzoate(m. p. 120') was not depressed. An Orsat analysis ofthe uncondensed gaseous product of th e decompositionshowed this to be composed of th e following constituents(yoby volume) : carbon dioxide 75.9y0, .oxygen 2.8%,carbon monoxide 1.4%, ethylene O.O%, higher olefins aspropylene O.8y0, atura ted hydrocarbon (apparent mol.wt. 26.4) 19.1%. In the aforementioned duplication ofthe experiment ,36 the acetaldehyde, acetone and isopropylalcohol fractions from the to tal product weighed 4.46 g.,5.61 g. and 8.23 g . , respectively. Also, the presence ofsubstantial quan tities of ethane in the saturated hydro-carbon fraction of the gaseous product was c on kne d bychlorination to hexachloroethane at 480-500 ' sing excesschlorine. The product, after recrystallization from abso-lute ethano1;sublimed entirely in a melting point capillaryat 186-187 . The sublimation tempereture WB)(1 W*

1.3584-1.3620; (c) 10.3 g., b. p. 75-81 , nzoD 1.3709-

*

( PO ) Pufmmed br Dr. W, A. S t r e U E .



March, 19.50 HEATSOF COMBUSTIONF ISOMERICITROSTILBENES 1263

changed on admixture with an authent ic sample of hexa-chloroethane.

Anal . Calcd. for C2Cle: C1, 89.9. Found: C1,89.4.

Catalysis of Polymerization of Ethenoid Monomers byPeroxydicarb0nates.-Monomeric styrene, vinyl acetateand methyl methacrylate from commercial sources wereredistilled before use.

2,2'-Oxydiethylene bis-(allyl carbonate), i. e. , the bis-(allyl carbonate) of diethylene glyc01,~' was prepared38by the addition of 1326 g. (11.0 moles, 10% excess) ofallyl chloroformate to an efficiently stirred mixture of 531g. (5.0 moles) of diethylene glycol and 949 g. (12.0 mole;)of pyridine, cooling to keep th e reaction mixture a t 5-10 .After the addition was complete, the reaction mixture wasallowed to warm to room temperature, diluted with anequal volume of water and neutralized t o the methylorange end-point using dilute hydrochloric acid. Thecrude oily ester was separated, washed with an equalvolume of cold 2% aqueous hydrochloric acid, once withan equal volume of cold 2y0 aqueous sodium hydroxidethen twice further with cold water. The crude productwas dried over sodium sulfate, and purified by distillationwith a small stream of carbon dioxide passing through theester. The yield was 1030 g. (75%) of colorless 2,2'-oxy-diethylene bis-(allyl carbonate), b. p. 160' (2 mm.) ,nmD1.4503, dZo41.143.

Anal. Calcd. for (C~H~OCOOCzH&O, IZ HI~ O~ :,52.5; iodine number ,sg 185.1. Found: C, 52.2; iodinenumber, 185.1.

A loss of abou t 15% due to polymerization during dis-tillation was avoided in other preparations by heating thecrude ester a t 150" in vacuo (5-10 mm.) for removal ofvolatile impurities, the product obtained being 97-99%pure by unsaturation analysis.

Diisopropyl peroxydicarbonate was dissolved in thesemonomers in the following percentages by weight, basedon the monomer: styrene o.570, e n y l acetate O.l%,methyl methacrylate 0.2y0, 2,2'-oxydiethylene bis-(allyl-carbonate) 3.0%. Two-ml. samples of the solutionswere heated under nitrogen in closed test -tubes (10 mm. X75 mm.) for three days at 45" then cooled to room tem-perature. Ha rd , colorless, solid polymers were obtainedin all cases.

(37) Muskat and Strain, U. . Patent 2,370,565, Feb. 27, 1945.

(38) Muskat and Strain, U. . Patent 2,370,567, Feb. 27, 1945.

(39) Hanus method, Jamieson, "Vegetable Fats and Oils," A. C. S.

Monograph No. 58, Second Edition, (The Chemical Catalog Co.,

Inc.) Reinhold Publ. Corp., New York, N. Y.,943, p. 394.

Summary

Dialkyl and substituted dialkyl peroxydi-carbonates (Type I peroxycarbonates) may beprepared readily, often in good yields and highpurity, from the corresponding chloroformates andsodium peroxide solutions. Most of these esters in

pure form are unstable, decomposing, in somecases explosively, a t normal temperatures. Ap-propriate temperature control in preparation andhandling is accordingly required.

Wide differences in thermal stability in pureform are observed among esters of Type I, depend-ing upon the alkyl or substituted alkyl group.Similarly wide differences in stability are observedin solutions of a given ester depending upon thenature of the solvent used. The thermal decom-position of the more unstable esters is retarded byknown inhibitors of chain reactions, and yields acomplex mixture of products characteristic ofchain decompositions. In dilute solutions in se-

lected solvents the wide differences in decomposi-tion rates are no longer observed, and the decom-positions conform closely to first order kinetics.The thermal decomposition of the different estersis thus shown to consist of a spontaneous unimo-lecular decomposition, with the formation of freeradicals which induce concurrent chain decompo-sition reactions of the esters to different extentsdepending upon t he particular ester and solvent in-volved. These chain decompositions are similar tothose known to occur in benzoyl peroxide solutions.

3 . Examples of the related esters of monoper-oxycarbonic acid (Type11)and of diperoxycarbonicacid (Type 111) have also been prepared; theseesters are more stable than the esters of Type I .

Many of the esters of Type I are effective freeradical type catalysts for polymerization of typi-cal ethenoid monomers at moderate temperatures.

BARBERTON,HIO RECEIVED AY20, 1949

1.

2.

4.

[CONTRIBUTION FROM TH E DEPARTMENT O F CHEMISTRY, OREGON STATE COLLEGE]

Heats of Combustion of Some Isomeric Nitrostilbenesl

B Y CARLM. A N D E R S O N , ~ ~ELAND . COLE^^ AND E. C. GILBERT

For a number of years this Laboratory has beeninterested in energy relations of isomeric organiccompounds as obtained from heats of combustion.

Through the courtesy of Prof. Melvin Calvin,i t has now been possible to obtain two additionalpairs of carefully purified isomers, cis- and trans-

(1) Published with the approval of the Monographs Publication

Committee, Oregon State College, as Research PaperNo. 143, School

of Science, Department of Chemistry.

(2) (a) Present address: Department of Chemistry, Linfield Col-lege, McMinnville, Oregon; (b) Present address: Chemistry Sec-

tion, Jet Propulsion Laboratory, Calif. Inst. of Technology, Pasa-

dena.

( 8 ) Corruccini and Gilbert, Tars JOURNAL, 61, 2926 (1939);

DavIw and Gilbert, Ibid., 68, 1686, 2780 (1041); Aaderwn m d Gil-

bwt, ik 'd . , M, 889 (lQ42).

4,4'-dinitrostilbene, and cis- and trans-4-mono-nitrostilbene, (unsym. p-nitrostilbene) . Measure-ments of the heat of combustion of these fourcompounds are reported here, and the energiesof isomerization in the solid state are indicated.

Experimental

The adiabatic calorimeter has been previously de-scribed.3 The energy equivalent was determined by th ecombustion of benzoic acid Standard Sample, 39e, from theNational Bureau of Standards, by Anderson who obtained2607.30 cal. deg.-l with a maximum deviation of 1.45 cal.from t he mean, a mean deviation of 10 .5 6 cal., and a root-mean-square deviation of a0.28 cal. After approxi-mately a year a redetermination by Cole using Standard

3Qf, gave 260734 m l , deg.-l with a mean deviation of

Documents

Esters of Peroxycarbonic Acids