UNCLASSIFIED

AD NUMBER

AD275878

NEW LIMITATION CHANGE

TOApproved for public release, distributionunlimited

FROMDistribution authorized to U.S. Gov't.agencies and their contractors;Administrative/Operational Use; 17 MAY1962. Other requests shall be referred toUS Office of Naval Research, 800 NorthQuincy Street, Arlington, VA 22032.

AUTHORITY

ONR ltr 28 Jul 1977

THIS PAGE IS UNCLASSIFIED

UjNCL ASSIFIED

At.275 8'78

ARMED SEHlVICES TECHNICAL INFORMATION AGENCYtRLUIOGTON HALL STATIONARLINGTON 12, VIRGINIA

Best Available Copy

LA 9NC LASSIF I ED

NOTICE: When government or other drawings, speci-fications or other data are used for any purposeother than in connection with a definitely relatedgovernment procurement operation, the U. S.Government thereby incurs no responsibility, nor anyobligation whatsoever; and the fact that the Govern-ment may have formalated, furnished, or in any waysupplied the said drawings, specifications, or otherdata is not to be regarded by implication or other-wise as in any manner licensing the holder or anyother person or corporation, or conveying any rightsor permission to manufacture, use or sell anypatented invention that may in any way be relatedthereto.

FOR REFEIRENCE ONLY AT EACH 0 TH.ASTIA OFFICES. ThIs REPORT CANNOTBE SATISFACTORILY REPRODUCED; AaTZA

:= (•"DOES NOT FURNISH COPIES.

L" OwzCM OF NAVAL 10RA

rfm contract Nowr 25600V 44 m~ Task No. E055-398

Finea Tecbnica]. Reot

by

is .E LefflerI. ForobladU. Honseberg

Y. Israeli"B. Go lmey

"" - s$o RiccitielloL. ToddT. Teuno

Dopgrtnt of Cbmistry, Florida State UniversityTallahamse, Florida

aw 17, 196e

bproductiom ID whole or in Prut is permittedfor aq puWposae of the Unitead States GoverXmut

0

I NO -"r-

Contents

Part. I. Org5InOboroS Cbei try)

Part II. The Wluorenyl-Nitrogsn Syptm 23

Part MoI The Chemmistry of Phosphiuie-Azlde Co~1exes 26

Part rV. The Decocposition of a SubstitutedDsnzazides, and Benzenesul-fonazictes 36

Part V. The Deacapositioi of Benzeziesultonylazides 38

Part VI. The )ktaial Acceleration of the I)eccposi-tion of Denzenbsulforzyl Az1.de aid t-DiitylRydrcwperoxide 56

Part VIIT. TIphanyiphosphine Oxide Perow:de 76

Part VIII. Trapping of Dichiorocarbene b,TrIphenyr1pbosphine 79

Part IX. Miscellaneous Exploratory Experipients 79

PART OR)FGAIIOBORON CHMMISTPY

P ration and Decomposition of Dimesityl Boron Azide

See Technical Report No. 1 or Leffler and Todd,

Chem. and Industry L 512.

Attempted Preparation of Ar rN

See Technical Report No. 1 R

Attempted Reaction RN + R B- R A-N----> R B-N 4 N3 3 3 '2 2 ~ 2R R

See Technical Report No. 1

Attempted Preparation of 02 BCHI2 C6 H4 0CH3 (p)

See Technical Report No. 1

Reactions of Organoboron Compounds with Diazo Compounds

See Technical Report No. 1 and Leffler and Ramsey,

Proc. Chem, Soc. 196. 117. The following information

has not yet been published elsewhere.

Phenylboronic acid anhydride reacts with ] iphenyldiazcmethane in ether

to give products which on acid hydrolysis are isolable as triphenylmethane

and tripheny.carbLnol. Tolylboronic anhydride sinA.larly gives

cliphenyltolylmethane o

These.. results are consistent with the following reaction mechanism,,)rII,r,,,L3 r I?,jN• -vY-8 -c. -]-I 1

PY - I C1-9 IN-A

,- -FN - ,- C,-I t"V-

I! 1Ipll "4'liI I I I ' rl I!

I

The first step is consistent with ccuplexing of Ar3B with various

reagents having unshared pairs of electrons. (vide infra)

Triarylboron Absorption Spectra

See attached reprint, J. Chem. Phys. & 1502 (1961)

Triphenylboron in inert solvents such as methylcyclohexane has an

intense band at X max 2871 A end E 3.9 X 104 . Tri-C-naphthylboron has

intense bands at 2638 and 3526 A. These have been tentatively identified

as intraolecular charge-transfer bands. (see reprint). They vanish

when dry ammonia is bubbled through the solution, but in the case of

tri-a-naphthylboron reappear again when HCl is bubbled through.

I I

-. I-

IM N " Y --NH3

The spectra of tritolylboron and trimesitylboron are like that of

triphenylboron. The transition energy of the charge transfer band in Ar3 B

is a roughly linear function of the first ionization potential of ArK (figure 1).

A precisely linear relationship obtains between the intremolecular charge-

transfer transition energy and the intermolecular charge-transfer transition

energy for the Ar*'12 cow ex (figure 2).

•t I ' II I I I I I I • l 1 i i 1 a

0 "

1IZ13

Apal

-- i,.3 •

hi•

0 1%•

• tLmJ, _ i • I m i m; mi i-

: : m m m m m m • m rm • m mm • m • - -

S.. . •m " " " : : " -- ) -- ) ) ) -- )) .. .

• 0 " -

Fc, . ,•

k./6E c ,ohsger Air4.E.

/00.

W

A6d,3•

80 90 /00

K. CAL./ olE

Table 1

Ultraviolet C. T. Maximum in Metbylcyclohexane

Tri-l-naphthyl boron 3526 1.9 X 104

Trimesityl boron 3310 1.5 X 104

Tritolyl boron 2966 3.3 X 104

Triphenyl boron 2871 3.9 X 104

There are some anomalies in the earlier literature on the spectrum

of triphenyl boron. Thus in acetonitrile it exists only in the ccmplexed

form and the C. T. band is missing.(1)

(1) D. Geske, J. Phys. Chem. 611067 (1959)

The C. T. bands of tri-a-naphthyl boron in that solvent are weakened

(1/5 factor) and several intense new bands appear. The highly hindered

trimesityl boron has its usual spectrum even in acetonitrileo The spectrum

reported by Mikhailov(2) for triphenylboron is erroneous, and is probably

the result of air-oxidation of the triphenyl boron.

(2) B. M. Mikhailov, Optika Spectroskopiza 7644 (1959)

• .gi , I I I I I l I F ni •i i= ~ ' ,= ; ; , . .....••,,

6

Boron Radicals and Triaryl Boron Photolysis

Photolysis of tri-a-naphthyl boron in toluene, using sunlight or

ultraviolet light, gave naphthalene, a-naphthylboronic -kcid, and

di-a-naphthylborinic acid. In dry CCl 4 under the products are

hexachloroethane, binaphthyl, naphthalene, and 1-chloronaphthalene

(trace) plus readily hydrolyzable CM 3 or C1 derivatives, On hydrolysis

the reaction mixture gives Cl" (as AgCl) in large amounts plus C-naphthyl-

boronic and di-a-naphthylborinic and a-naphthoic acids. The products fume

in air. If the photolyzed solution is oxidized with air and then treated

with small amounts of water, phosgene is evolved. These products are

consistent with I 'ie following sets of reactions:

(a-n) 3 B (a-n)2 B" +

(a-n) 2B + CCl 4 - (a-n)2 B-Cl + C13C

S + co 4 _ + C13 C

(a-n) 2 B. + Cl 3 C' C (ar-n) 2 B-Ccl 3

06 + C13C -- (a-n) 2 B-Cd 3

(a-n) 2 1K c-n-B-X + etc.

Protolysis of Tetrapheniylborate Ion

This work was supported only in part under thl contract. It

constitutes the dissertation of Verne Allen Simon, lo So 1Uo 1962, and will

be published in full elsewhere. Tracel and kinetic studies have established

the protolysis mechanism in 50% aqueous dioxane as a bimolecular reaction of

protons with the phanyl substituents at the carbon attached to boron.

There is no detectable general acid catalysiuo

Experimental Section, Organoboron Chemistry

Preparation of Tri-l-naphthylboron:

Tri-l-naphtbylboron was prepared ')y the method of Brown and Sujishi(3)

from the reaction of 1-naphthylmagneshtm bromide with boron trifluoride

etherate. M. p. 203-2O0Wuncorro % B - 2.6, % B calc. 2-7

(3) H. C. Brown an S. Sujishi, JACS 7M 2793 (1948)

Lparation of Triphenyi boron:

Triphenylboron was prepared from pheny1magnes|tm b-omide and boron

trifluoride according to the procedure of Wittig,(4) or was purchased from

Aldrich Chemical Company.

The crude triphenyl boron was recrystalized to a white crystalline

substance from benzene in a dry box under prepurifielI nitrogen. M.p. 146-l147

(evacuated).

The purity of 03 B is tested by the absence of strong peaks at 1320-1400

and 1810-1820 cm-l which are characteristics of oxidation products of triphenyl

boron. The solutions were made up and the cells loaded in the nitrogen box0

(i) G. W t•i -t a,. Ann. 3 110 (-949); Ber. 88 962 (1955)

Passage of amnonia through an ether solution of triphenyl boron Lave a

white crystalline adduct. M.p. 204-207" (dec) in air and 234* in an evacuated

and Ecaled cappillary.

Preparation of Tritolylboron:

Tritolylboron was prepared in the following manner, using a procedure

adapted from Wittig° (5)

(5) G. Wittig, Ber. 88 962 (0p55)

The Grignard reagent prepared from 34.2g of parabrcumotoluene was added

under nitrogen to 9. 5g of boron trifluoride etherate in 100 ml of ether at a

rate sufficient to maintain reflux. After addition was ccuplete, the reaction

was refluxed an additional hour. The ether was removed by distillation and

replaced by 75 ml of toluene. The toluene was heated to reflux and the solution

filtered through a sintered glass disk into a flask, under a nitrogen atmosphere.

The flask was stoppered and transferred to a nitrogen box where 30-40 ml of the

toluene was evaporated off and the solut. on filtered and was allowed to stand

overnight. A white srystalline product was obtained, which was recrystallized

from toluene, and dried under N2 at reduced pressure in the air lock of the

nitrogen box. M. P. 149o5° -151, reported M. p. 1420-440. The product has-I

the following strong to medium bands in the infrared, in units of cm , in CC0 4 :

3050, 2950, 2900, 1600, 1445, 1400, 1315, 1280, 1240, 1210, 1190, 910, 895. The

spectrum was taken on an Infracord and the cells loaded under nitrogen.

Trimesityl boron:

Trimesityl boron was synthesized by the method of Brown and Dodson.o(6)

Mesityl magnesium bromide was reacted with borontrifluaride etherate, and the

product recrystallized from acetone. M. p. 1913-1920.

(6) H. C. Brown an,, V. H. Do, son, JACS 12 2304 (3.957)

VJ

RAcOtions of triphenylboron with dia~methane:

Run I

Diazcaethane (13 X 10°1 moles) in 200 .1 other vas added slovly

with vigorous stirring to 1.7 X 10.-3 moles of triphenyl boron in 100 ml

of ether, under a nitrogen atmosphere.

After addition of the diazanethane was cocplete, 100 .1 of dilute

hydrochloric acid was added and the mixture was stirred at room tmperature

for several hours.

The ether layer was separated, dried, and all but 20 ml distilled

off. The ether uas replaced by 40 ml of isooctane, (spectro grade).

This was fractionally distilled and ultraviolet spectra taken of the four

fractions obtained.

Fraction I - Identical with benzene

II - A mx-, 2422, 248, 254, 26o0, 267, 279 np

III - Amex = 248, 260, 261.5, 264.7, 268.3, 277 up

IV - Amax = 243 HP, weak 277, 298

Fraction III was redistilled through a small column (helices) into

fractions IIla, IM.b, IIIc.

The ultraviolet spectrum of fraction 1TM vas very nearly identical

with that of toluene, the only difference being an additional weak absorption

at)namx 248 up.

Run II:

An excess of diamethene (• 45 X 10-3 moles) in ether was added,

with rapid stirring under nitrogen, to (9.1 X 10"3 m, 2.21g) triphemylborom

in 100 ml ether. Thirty ml. of 30% hydrogen peroxide was then added and the

reaction stirred over night. The ether vas then extracted with 75 ml of

water and treated with lithium aluminum hydride until the ether no longer

gave a peroxide test.

iI

The ether vex then eztracted with 75 ml.of concentrated sodium

"hydroxide sboution, and dried. Rhmoval of the ether by distillation left

a residue of 0.3 g. Vacuum distillation of the residue yielded 0.201 g. of

benxyl alcohol. The infrared spectrum of the product and the spectrum of

its phenyluretban derivative, M. p. 740, were identical with spectra of

known samles.

The solid residue left from the vacuum distillation showed an

ultraviolet meximum at 246 uA. Chromtogmaphy on alumina gave a wazy

crystalline polymer whose infrared spectrum,/Amax 3050 w, 2940 s, 2870 m,

1590 v, 1480 a, and 690 a (cm 1 ) and ultraviolet spectrum, A mzx 244 up0

would seem to indicate a polymer with ( 0 - 6 ) end groups - or else

contamination with very small amounts of benzaldehyde. Benzaldehyde has a

strong band at 244 up.

Tri-l-naphthylboron and diazomethane:

Diazomethame (2 X 10-2 moles) in anhydrous ether ws added under

nitrogen to 8 grams (2 X 10-2 moles) of tri-l-naphthylboron. The reaction

proceeded with very rapid evolution of nitrogen and formation of polymer.

The products were oxidized with 30% bydrogen peroxide, the ether then freed

of peroxide and distilled off. Vacuum distillation of the residue did not

yield any of the expected 1-naphthylmetbanol. Chromatography on alumina of

the residue and elution with hexane gave a low molecular vt. hexane-soluble

polymer whose infrared spectrum in carbon tetrachloride showed only -CHE-2

absorption. The ultraviolet spectrum in metbylcyclohexane, conc. - 160

mg/liter, however, showed the presence of l-naphtbyl end groups when coupared

with the spectra of 1-tbybvinphthalene (7), l,2-di-l-naphthylethane (8) and

l-metylnaphthalene in isooctane.

(7) Rauurt-racus, ill. Soc. Chim. France I. 424, 1952.

(8) R. Mams amd E. Kirkpatrick, JAcs 60 2180 (1938)

'o,

cu

cu cu

- ~ Cu

co 0i

aj cucu CY

LA

\LJ A

o \ 0 0

,e \J u cu

ou 00

4)

-~ Ha, H,

too

H H H H1

OR 4-'0,P

H~

The insoluble polymer from the reaction of diazumethane and

.- tri-l-naphthylboron was washed with water, acetone, and dried. The infrared

spectrum in a KBr disk had bands characteristic of l-naphthyl end groups(9)

at 7U8 and 7M cm -1, other bands at 2930, 2850, 1465, 732, 720 are

characteristic of CV4" There was no other absorption. The polymer was

then placed in a Soxhlet extractor and extracted with chloroform for 24

hours. After this time the spectrum of the remaining polymer showed no

changes in position or intensity of the bands. Evaporation of the chloroform

gave a small amount of hexane-soluble polymer. The infrared spectrum of

this polymer, A max 2930, 2850, 1730, 1460, 795, 776, 730, and 720 cm-1

showed a strong relative increase in the f 9 bands as compared to the

(.cH2)4 -amu at 720 cm-1

(9) R. L. Werner et al., Australian J. Chem. 8 346 (1955)

Reaction of 0 B and 0 CN2 .

Triphenylboron (2o28g, 9.4 X i0"3 moles) and diphenyldiazomethane

in 400 ml of ether were stirred under a nitrogen atmosphere for 12 hours.

After this time approximately 200 ml of nitrogen had been evolved and the

purple color of 02CN2 had been replaced with an orange color. Twenty ml

of H202 (30%) was added to the reaction and allowed to stand overnight.

Filtration gave 2.7g of azine. The ether was removed and replaced with

30 ml of hexane. Filtration of this gave an additional gram of azine.

Total yield of azine 3.Tg, m.p. 1620. Evaporation of the hexane gave 4.5

of a black tar.

Chromatography of the black tar on alumina gave triphenylmethane,

m. p. 92" , yield o.40g, 1.6 X 10-3 moles and triphenylcarbinol, m. p. 1620,

yield O.16g, 0.6 X 1O"3 moles. The infrared spectra of these products were

identical with those of known samples. The combined yield 0 3CO + COH is

23%.

Reaction of Phenylboronic Acid Anhydride and Diphenyldiazcmethane

To a refluxing solution of 0o755g (6.52 x lO3 M) of phenylboronic

anhydride (m.p. 216.50- 2170) under nitrogen was added 1.4g of 02 CN2

dissolved in 100 ml of hot toluene.

There was a very rapid evolution of nitrogen and disappearance of

color, Concentrated hydrochloric acid was added (75 ml), and the solution

refluxed for 45 minutes. The two-phase solution was cooled and filtered,

30 mg of boric acid being collected. The toluene layer was separated, after

washing with water, and on standing overnight gave another 40 mg of boric

acid. Evaporation of the aqueous portion gave further boric acid, making the

total 0.22 g, or 3.6 X 10-3 moles.

The toluene solution was concentrated to 5-8 ml and chromutographed

on an alumina column to give triphenylmethane, m.p. 92 0, yield 0.22 g,

0.91 X iOk3m, triphenylcarbinol, m.p. 1610, yield 0.41g, 1.6 X 10-3 moles,

(combined % yield 40%) and• 2 C - 0, 0.61g.

Paratolvlboronic Anhydride and Diphenry.diazcmethane:

The product from the reaction of 1.41g of 02CN2 and 0.606g (3°8 XI lO 3 m)

of paratolylboronic anhydride In toluene under nitrogen was hydrolyzed with

HCJ and chromatographed. DiLheM3VaTato0vrmethane (0.34g, 1.3 moles, 34% yld.)

was identified by comarison of its infrared spectrum with that of an authentic

sample. Diphenyl carbinol (.05g) was also isolated. The diphenyl earbinol

probably arises from a small amount of boronic acid present.

~_

L4I

Photolyeis of Tri-l-naPhtbYlboron!

Run I

A solution of 1.1 6 g of tri-l-naphtbylboron in 20m1 of toluene was

irradiated in a "Vicor" tube with an ultraviolet lamzp for three d~ays. During

this time a. steaiy flow of nitrogen was bubbled through1 the solution.

Fractional crystallization from toluene and. hexane gave .02 4 g

(0.14~ X 10O3 moles) of riaphthylboronic acid., 0.16g recovered tri-l-naphthvl-

boron, and 0.121Lg of dinaphthyl borinic acid (m.p. 115 oll160 % c 84i.7 % H 5.33).

Chromnatography of the residue from the mo~ther liquors, 0.71g of a mixture of

na~phthalene and borinic acid, gave .0126 grams of naphthalene.

PhotoLysis in Carbontetrachloride:

RUM II

A sealed degassed amapule containing 1.07g of tri-l-naphtbylboron in

20 ml of carbon tetrachlorid~e was irradiated for 12 hours with light from an

AH-6 Hg arc lamp, using a filter of Amax 355 flA. and 0% T<~300 iiP and > 400 mp.

Exposure of the photolysized solution to moist air resulted in fuming

and production of hydrogen chloride gas.

A five ml) Eample of the carbon tetrachloride solution was evaporated

to dryness and the crude ta~rry product taken up in hexane. On standing the

mixture precipitated naphthylboronic acid,which Vas filtered off. The hexaze

zmother liquor was evaporated to a few drops and put through a silicone Grease-

on-firebrickc V.P.C. coluzon. Three peaks were obtained at 1700, flow 3Occ/uain.

These were identified by comparison of retention times and infrared spectra

as hexachloroethane,. naphthalene, and 1-chloronaphthalene.

"Run III

Spectrograde carbontetrachloride, 49 ml., was distilled frcm P2 05

into an amwule containing 3 Og of tri-l-naphthylboron.

This sample was irradiated with U. Vo light, ,m, x 355 4a, from an

AR-6 mercury arc lampo Vapor phase chrcwoatography as previously described

gave a yield of 21 mg of hexachloroethane. The presence of unreacted

trinaphthylboron obscured the detection of naphthalene or chloronaphthalene.

Ten ml of water was injected into the sample through a syringe cap and

e••. r a'aaking for 15 minutes the ampule was opened to the air. A precipitate

of (,E.. grams of phenylboronic anhydride was immediately filtered off. The

aq~u ous layer after separation was strongly acidic and gave a positive test

for chl ,ride ion.

The carbon tetrachloride solution was separately extracted with

wate r c, er a period of several days and continued to give positive tests for

chlorid, ion. The combined aqueous extracts gave 0.798 grams of silver

chlada., or 5.6 X 10-3 moles.

3xtraction of the carbon tetrachloride solution with dilute potassium

hydr:)xi,. i gave O.lOg of a dark brown precipitate (KOB [aC10V 2) and an

aqueo2us .3olution which on reacidification and chloroform extraction yielded

0.2g of L mixture of naphthalene boronic anhydride and a naphthol.

Run V

, ample identical to that used in Run II was photolyzed for two days

in s •ili hto The carbon tetrachloride solution was swept out with nitrogen.

The :Ltr gen stream then gave a positive borom flame test, and after passing

.1

through a solution of methanol# caused the methanol solution also to give

a positive flame test, attributed to the formation of methyl borate frum

boron trichloride.

Oxidation of the sample by air, followed by hydrolysis, resulted in

the formation of phosgene, detected with a Prigerwerk Lubeck Gas detector

model 19/31. The phosgene was swept with air through an Et 20 solution of

aniline with the resultant formation of phenylurea, mop. 235 .

Sodium Tetraphenylboron and NaN :

Sodium tetraphenylboron (0o5g) and sodium azide(Oollg) were refluxed

for 12 hours in 10 ml of water. At the end of this time two phases were

evident. Distillation to dryness gave a distillate whose u. v. spectrum

was identical to that of benzene in water.

The residue from the distillation did not give a green flame test,

burn, or melt, but decomposed explosively in a flame. An u. v. spectrum

was blank to 210 upo

The residue on solution in dilute hydrochloric acid and evaporation

to dryness to remove azide ion now melted to give a borax bead and on -ddition

of H2•0 4 and EtOH gave a green flame test. An uov, spectrum of the material

was blank to 210 np.L. This cl&aracterizes B(OH) 33

Control l.Og of N%04 was refluxed for a day in 4o0 m of slightly

basic water, At the end of this time the solution still gave a voluminous

white ppt. with NH40H, showing presence of (04Bo)o Distillation gave no

benzene detectable in the distillate by u.v. spectra. A second control

experiment using an equal ionic strength of NaCl in place of the NaN 3 showed

that the effect of the NaN3 is merely a salt effect.

3 :- ii i i •

Pben~lborgge Acid and &drazoio Acid

Phany•boronic &aid (or more probab. the anhydride), 1.og (lO'%oles),

was dissolved in 30 .1 of CIM3 and 7 ml of cone. H2804 . To this was added

1o30g (2 X 1O"%) of sodim azide and the reaction was stirred at room

tagerature for a day.

After a day 25 ml of %20 was added, the aqueous phase saepxated,

made basic with sodium hydroxide, and extracted with ether. The ether on

evaporation gave 0.055g of an oil whose spectrum was identical to that of

aniline. Distillation of the C3C13 gave 0.61g (7.9 X l0' 3moles) of benzene,

determined by u.v. spectroscopy.

One ml of benzene (0.78g) (102 moles) in a control experiment under

the same conditions also gave aniline.

Sodium Tetraphewilboron and Fqdrasoic Acid:

Sodium tetraphenylboron (0.7g, 2 X lO 3moles) and sodium azide

(0o.65g, 0 -2moles) were dissolved in 50 ml of 0.1 N hydrochloric acid and

allowed to stand overnight at room temperature with stirring. The next day

the solution vas a clear yellow, The solution was made

basic to litmus and extracted with CHCl3.

Distillation of the CHC1 3 gave ar. nlly residue whose infrared spectrum

showed an azLde band. The oil vas redissolved in ether, washed repeatedly

with water, the ether solution dried and again evaporated to dryness. A

samll drop of the same oil was obtained* The infrared spectrum had strong

peaks at 2140 c I, 1600 cm• , 1490 cm-", and 1290 cem", corresponding to

peaks in the spectrum of phemyl azide.

The oil could not be induced to form an adduct by mixing with a drop

of dicyclopentadiene. However, it was much decomposed by this time.

The ultraviolet spectrum of the C013 distillate from the above

shovs the presence of benzene. The spectrum of the ctqueous solution after

extraction (A max = 266 w) is very sirilar to that of 0B(0H)2 (A max 266 nL )

Ultraviolet Spectra:

All spectra were done on the Cary model 11 recording spectrophotometer.

In the case of 03B and (PcH30) 3 B, the solutions were prepared under nitrogen,

and the u.v. cells loaded and sealed under nitrogen.

compound: (aC1097)3 B, m.p. 2 0 5 ,-2.06

Solvent: methylcyclohexane

conc.: 3 X 10-4 molar

Path length: 1 mm.

A max 0O, D. E-3526 X 0.566 1.9 X i0"l

288 . 0.277 9,4 X 103

2638 1 0.577 1.92 X lO4

2214 4.43 1.5 X 105

Samle.: Sam as above vith ammonia bubbled through, (ao-Co10 7)3 ' 3 "

Path A ex 0o. . 4.

1 MR 3215 .092 31 X 102

1 am 2891 0.773 2.6 x 104

0.1 mm 2282 0.498 1.7 X 105

_ _ A- -- -" -- - - -- . . .. . . . .i =

1-7

C~~ound (C 10H.7), B

Solvent: C"3 0N

Cone: 1.3 mg in 150 ml, path 1 cm.

'0\ ma 0, D.6

3775 o.31 0.4 Xo 10X

3200 0.34 o.44 X 1o0

2850 1.9 2.5 X 104

2562 1.0 1.3 X 104

Ccopound 03B, m.p. 1146-147*

Solvent: metbylcycloboxwe

Cone: 5°7 X 10-4 MOls Path- 0.1 I

mx O.D.14

2871 A 0.223 1.9 X 10

2757 1 0.198 3,5 X 104

2380 • 0.108 1.9 x 10

Same as above,, "H3 (dry) bubbled through sa~1e; va length 1 cm

A ux 0, D.e

272 o.82 1. X •o3

2630 1.08 1.9 X 103

2570 0.87 1.5 X 103

2480 0.82 1.4 x 10o3

0 B in CH3CNJ solution prepared, and cells loaded, in nitrogen box. 0nly

end absorption below 230 W.

Tritoly boron, rA.P. 149.2j ~

Solventý mnthiylcyclohexaio

'One 2 X 10- molar, path 'ami

2966 X 0.066 3.3 X 104

2820 0,,3B 19 X 104

2490 O.h:1. 20.0 x io

2350 0.,. .L5 x 101.

LAddition of~ trietbfiandne- ca-used ah disappearance~ c-P the C, T. ben~d

2966 A.

Trire c 24,, 1010.p 1,9045 l9L'i-"

Solver.t;. Mtb~y:cyclohexane

'Path: 1 zmm

331.0 A o.6,. 6S -L15 0

2830 (0h'~e~ o.2470 (shouji~er'p0,7

rphe- Spectrum, in aclatonitrile is identical.

Tri:Q-naphtllbyabron and n-b-eI.vaz~ide.

Tri-U-naphthy1bo.,.':n (4g lk-2 Moles", andi a-.& ml of n-butyl azide

(1.6 x lo"'r.,) L, 30 ml of c.L1oro~jnzi-iBe were .,'ieated in a --ealed~, idegaz3sed

ampule for 221 days at 150c,~

Addition of 5 ml of hexane to the OC1 solution and storage in a

refrigerator for 2 days gave back 3g of (OC10 ,1) 3B. Vacuum distillation of

the mother liquor gave a solid crystalline mass whose infrared spectrum was

that of a mixture of (c'1 0 E7 )3B and (*C 10o)B(0H)2 .

Tri-a-navhthylboron and Aluminum Azide:

One gram of (cOC 1037 )3 B and two moles of AI(N3 )3 generated in situ from

A1C1 3 and NW3 were refluxed in tetrabydrofuran for 24 hours.

The solid residue after distillation of the THF and washing with ether

gave only 'tri-C,-naphthylboron.

Tri-au-naphth•ylboron and Triphenymethylazide:

Tri-a-naphthylboron (2.Og) and triphenylmethbyl azide(17.7) were refluxed

overnight in tetralin under nitrogen. The tetralin was removed by vacuum

distillation and the residue chrcuatographed on alumina.

The products isolated rind identified by infrared spectra were benzo-

phenone anil (lo06g), m.p. 1320, a-naphthol (0o147g), and a-naphthylboronic

acid (1-39g9) The spectra of these coupounds were identical in every respect

to those of known sanples.

Tri-a-naphtbylboron and tDImne GYcoi

Tri-a-naphthylboron (1.00g) was refluxed in 25 ml of HBCH2 0H for

10 minutes. The ethylene glycol was diluted with 25 ml of water and extracted

twice with 30 ml of hexane and twice with 30 ml of benzene. The extracts were

combined, dried, and evaporated to dryness to give O. 8 g of white crystalline

material.

This material was chrmatographed on alumina to give naphthalene (0.o2g)),

a-naphthylboronic acid m.p. 205(vaco), and a third white crystalline material

which melted 146-1490 (crude). Repeated recrystallization of the latter

substance from EtOH gave a-nIphtbylboronic acid.

"An infrared spectrum of the material melting at 116-1490 shoved bands

at 3.4 L, 3.5 p and 6.9 p characteristic of --. The compound is probably

an ethylene glycol ester of a-naphthylboronic acid.

In a second experiment 1.0 grams of (C Y) 3B vas refluxed under N2

in 35 ml of HOCH2 CH2 ON for 6 hours. 0.7 grams of naphthalene was isolated.

Tri-c-naphthylboron -hen•~lhydrazine complex:

Tri-a-naphthylboron (2.0g) and phenylhydrazine (0.552g) (1:1 mole ratio)

were refluxed under nitrogen in 30 ml of toluene for 7 hours. The toluene

was evaporated to about 15 ml by a stream of nitrogen. White crystals were

0 0collected and umshed once with cold toluene; m.p. 122.5 -123.5o.

II

3i

Part II. The Fluorenyl-Nitrogen System

In an attempt at synthesizing the nitrogen analog (I) of Koelach's

free radical (II)((10) we prepared compound (III)(") as a precursor.

We also prepared the alleged tetrafluorenyl hydrzine (IV) by the methods

(10) Koelach, J. Am. Chm Society 72 4439 (1957)

(11) Ingold and Wilson, J. Che. Soc. I= 1493

in the literature. (12) (13) These consist of autoxidizing or photo-

Oxidising 9-fluore7yl amino*

o-

I I

CY0O

(32) Goldscbmidt and hichel, Ann. 456 152 (1927)

(13) Chu and Wei•mn, J. Am. Chb. Soc. 76 3787 (1954)

_. .--- __-II I I Ik

it

The product reported in the literature as IV proved to be inI.

Thb3 teperature and concentration-dependent color changes of the alleged

14a Infra-red,r n m r, mixed malting point, cryoscopic molecular weight

in benzene.

"IV" are apearently due to impurities. Ccomound III, prepared by eith,-r route,

contains a free radical, the sam in either case. This radical is dtected

by its ZSR signal, figure 3 . The radical is removable by crystellization,

chroumatography or by boiling in alcohol. It seems to be fairly steble in

solution at room temperature, even when exposed to air, although vs have not

been able to isolate a pure sample, Some concentration of the relical can be

achieved by extration with ether, then extracting the extracted material with

benzene. The fine structure of the ESR spectrum would seem to preclude I as the

radical, since I should be a triplet. Structure V should haw the observed

doublet with hyperfine splitting, but it is hard to see why V would not rapidly

disporportionate as in equation (M.

The Schiff's base prepared from fluorenone and benzhydryl amine also

contains a free radical by-product, but in too law a concentration to pexMit

resolution of the finestructure. It is not the same radical as that found in

the other Schiff's base, however, since its ESR signal occurs at a different

field strength. The Schiff's base prepared from ben-olhenone and benzhydrylliine

gives no signal at all.

The Koelsch radical (I1) did not show resolvable fine structure.

____NZ

_________

Figure3

ESR Spectra of the Radicel from I1I in benzene at 2f (upper curve), in

pyridine at 17c°(middle curve) and in pyridine at 1000(bottmn curve).

UUM yj j iFAT ý F

-1--- 4 j -+ -A I [I 1 7 - 1�

--- HFRT4-t-HI-iffl -A i+

_1 iF ------I -I- Vý

I I A I r A! 11 1. 1 1: t -1 1

14 IN f I t it I t AL V ýAýmvt if It f Jill I 1 4 1 ivfmu III, v I

If I I I it INFU I-Il't

IF-

-64

EE

•6

PART MI. TO D OF PoMMAM =, wM S

Ftr the acid.catal.ysed deccsosition of the tripheny1phosphine-

trityl azide celex, see Technical RePort No. 2 or 3. E. Leffler, U.

Honsberg, Y. Teuno and 1. lorablad, J. Org. Chen 26 4810 (1961). Work on

other phosphine-azide ccmlexes is described here.

Introduction

Tertiary phosphInes react with covalent azides either to give

stable complexes, isolable but unstable complexes, or nitrogen end

phosphine-iminea. (5

(15) (a) H. Staudinger and e. Hauser, Relv. Ch.1.. Acta, 4 861 (1921)

(b) H. Staudinger and 1. Meyer, Helv. CMii. Acta, 2 635 (1919)

(a) L. Horner and A. Gross, Am. 19 117 (1955)

R-N3 + RI P R (-Nf4-i.$-a' 1i 3 N2 + Rff•'3 ÷ 3• [ ' " ''3] • 2 3 "" '

Ionic azi es, for examle, biphenylene iodonium azide, do not give

the reaction.

.I ;

- i i i ii; ii i i i ii i i7 IV i i4.

Ihaction of Trivhsnylahosvhima with Substituted IW.maside

The rate at 25', determined by nitrogen evolution, follovs the second

order law. The second order rate constants (Table 3) are independent of

changes In the initial concentration of the reagents. The rate of decomosition

of the azide at 250in the absence of triphenyphosphine is entirely negligible.

There is a wmal solvent effect on the rate, CHIC3 being faster than

benzene and 10% acetic acid in benzene also being slightly faster then

benzene (see Table 3).

In contrast to the dec~om ition of the triphenyiphosyhIne-trityl aside

ccPlex(16) and with the decuqposition of the aszides alone,0(17) this reaction

does not seem to be acic-catalyzed.

(16) J. E. leffler, U. Honsberg, Y. Tsuno and I. Pbrsblad, J. Org. Chem 264810 (1961)

(17) !- Yukau ard Y. Tsuno JACS 81 2007 (1959)

(18)The substituent effect an the rate is correlated with ) as Shown

in Figure 4., electron-withdrawing substituents accelerating and electron-

donating substituents decelerating the reaction. The value of to is +O78.

(18) The signa values are taken from D. H. McDaniel and H. C. Brown, Z. Org.

ChMn. &140o (1955) and axe VrInsy siss values, that is, they are

based directly on benzoic acid ionization In inter.

__ _ _ _ _

__ _ . . ._,_i__i , 1 i

TI

Table 3

T!U RA22 OF T0 U • C•TMIC OF SOB 0 TTmME1 B"ZAZ,]S WIM

TM IN 33 N= AT 25.00f

Y.T. Notebook Sub•t. Initial Cone. M/L k X 'OR Li sec.page

[Azide], (Ph3P].

200 H 0.0491 0.0491 0.333 ±t 0.003

202 .0488 .0494 .345 ± .008

204 .0487 .0970 .313 ± .oo6

208 .0470 .1856 .339

206 .0472 .1886 .33C 4- oob

210 .0238 .1903 .286 + .005

212 .0983 .0493 ,330 t*.OOA

av. .332

220 p-CH3O .0493 .0494 .167

222 .o485 .1462 .120

224 .0454 .3637 .171

av. .170

230 p-4O2 .0467 .0490 1.235234 .045o .0970 1.20

232 .0459 .1441 1.20

ave 1.21

240 m*O2 .0481 .0490 0.970

238 .0479 .1473 1.018

av. 1.00

2•11 3, 5-(NO2)2 .0330 .0330 4.0o4 + .03

246 .0488 .0487 3.96 .02

av. 4.00

-- ~~ ~~~~ ~~~ ----- --- -I l - •-

Page Bubst. Initial Comoe. /L k 2 X 10 a/loseCo

(Mzid.. [Ph_?]

214 H (in CHCi 3 ) .0487 .0970 .676.

216 H * .0487 .0971 .413

226 p-C=030 • .0492 .0492 .267

SIn Acetic acid 10 ml and benzene 90 ml

I I I IIi , .., .. •i

3.0

vs,¸

47 4

*Ig

ma

/, O_______

C 430R•.-.~

0 +/1.b

There are marked contrasts between substituent effects on the

decomposition of benzazides alove and on their reaction with triphenylpbosphine.

The substituent effects in the former reactions are the net result of counter-

active and linearly colrelated changes in the enthalpy and entropy of

activation.(19), e direction of the effect on the rate is determined by

(19) Y. lukaw and 7. Twuo, JAM M 5530 (1957)

the enthalpy changes. Both neta and pare substituents give points on a

single enthalpy-entropy or isokinetic line. The sequence of meta points

on the line, and the corresponding rates, is correlated with T7 . The

value of P is -0.33. The sequence of pars points on the line, and their

rates, is not correlated with C . Apparently any pars substituent capable

of resonance decreases the rate, regardless of its electron-releasing or

electron-withdrawing nature. In the present reaction both meta and pars

substituent effects are correlated with C". The rate of the former reaction

is considerably accelerated by increases in the polarity of the solvent,

•hile the present reaction is relatively insensitive to such changes.

The behavior of pare substituents in the deccmposition of asides in the

absence of triphenW1phosphines is unexpected and difficult to explain. It

also differs frcm the behavior of substituents in the rearrangant of

diazoketones. We will not attempt to discuss the mechanisa of the azide

decumpositio•n here except to note that it obviously differs frcu that on the

present reaction.

A mechanism for the present reaction in accord with the available data

is as follows: 0 k, , ,Xp-C-( t - P -j:- X- - IV. - A1=41v- P3

" - "-MCAV- -- 3 t '" IAII

i _• '•+ ii

The rate would follow a second order l1v if step 1, forerd., were e•-

determining and k2 w.e-, fast. Since both the aside and the mi.lex (or"*

transition state of structure resembling the •o•plex) have about the ame

degree of polarity., no very large medium effect is to be expected. Electron-

withdrawing substituents should increase the rate of reaction with

triphenylphosphine by stabilizing transition state structures having a

formal negative charge on the N adjacent to the carbonyl; yam substituents

should not be exceptional in their behavior since no special resonance

interaction is foreseen. Further tests of the mechanism would be desirable,

including a nitrogen Isotope experiment and an attempt to convert to first

order kinetics, step 2 rate-determining. "he latter effect might be

expected in the presence of very strong electron-withdrawing substituents

on the azide, since such substituents should increase K and decrease k.

Examining the data for the 3,5-dinitro-substituted azide, we find that the

reaction is still quite precisely second order. Hcwever, the reaction

between benzenesulfonyl azide and triphenylphosphine does follow first

order kinetics.

The Reaction of Benzenesu •._ovl.Azide with TrMpheil2hos hine in

Benzene

The reaction of benzenesulfonylaside with triphenylphosphine in

benzene Is first order with respect to the concentration of the complex.

Since the complex is almost entirely associated in benzene solution, the

rate is simply proportional to the nominal concentration of whichever reagent

is not in excess. The products are nitrogen and the phosphine-imine,

quantitatively.

OP~~m so, IV3,] t- 41 ,

m.9, isv- 'Table 4 shows the rate constants obtained by the nitrogen evolution method.

_____---_ m --A ld IH •

Table I

IN SMEM AT 25.000

Initial Conc. /L X13sc[Azide]. ]o3 PX

.0502 .0501 1.04

oO496 .0500 i.oda)

.0500 .0500 9.0.8

.0500 .0505 slow

.0500 .iooo i.Cdb)

.1003 .0500 1.01b)

.0502 .2003 1.LA

(a) With stirrin

(b) With boiling tips

mm nn1 I ' __ ,, _ _'' '

In two runs not shown in Table 4 the reaction was considerably slover,

k1 decreased during the run, and a precipitate apeared. The cC3lex is quite

insoluble in benzene at 25; but remins in supersaturated solution in the runs

having good first-order kinetics.

Thus the reaction of benzenesulfonyl aszide with triphenlphosphine in

benzene behaved as anticipated. The reaction in more polar solvents is quite

different, however.

The Reaction of Benzenesulfonwl Azide with Tripbenylphosphine in Chloroform and

in Polar Solvents

Experiments in solvents other than benzene used both the complex and the

azide and phosphine added separately. The solid ccelex is light yellow, malts

at 87-88 with gas evolution, and has no 2130=71 azide bAnd (Nujol m2ll).

Solutions of the complex in chloroform show the 2130c1 aszide band

steadily increasing in intensity for the first 45 min. after vhich the intensity

becomes constant. During the same period, nitrogen is evolved but the yield

never becomes quantitative. The nitrogen yield can be raised from its usual

60% to about 97% by adding 1 mole excess triphenylphophinh after nitroan

evolution has sto:pd (20) The missing nitrogen is therefore present as a

covalent azide.

(20) If the 1 mole excess 0 P is present & t the yield of 12is not affected. If 2 3-oles excess F? is presmt from the beginningthe 12 yield is raised only to .

The products in chloroform are 0802;12 (which precipitates out), and

03P0 ( isolated by chromatogreaby on alumina). The infrared spectrum of the

C=C3 solution after N. evolution has ceased can be accounted for by assuming

a mixture of 2132 2 19 d 03p0* Me spctrus does not suffice to rule

A3

out the presence of 03P1N0 2 , but it in not isolated. A control experiment

abovs that 03PXW 20 (from. the reaction In benzane) is eessLy isolated and in

recovered quantitatively from an al.mina colmm.

The Initial zeroth order rates of N2 evolution in Chloroform are

reproducible within about 10 per cent, but conform to no sIZle reaction

order. Excess s eans to have no effect on the rate; excess 0802 N3 increases

it. Acetic acid Increa ses both the 1ý yield and rate, but a more polar solvent

(MW or nitrobenzene in place of the dCl03 ) decreases both the N2 yield and

the rate. At present ve have no explanation of this behavior. except that scm

side reaction is consuming 0F. W•atever that reaction my be, it is not able

to compete in the presence of 03o The decoosition of 3 13 020in the

presence Of 0&13 gives 96.7% of the N 2 expected frmthe reaction below:

Free radical reacotimsof 093' have been noted by (seay i21). A

free radical 0?P- derived from 0_? has been postulated by rlt

(21) M. Takebeyahi, private commuition. Ar6.o + 0802 N3r Ar6-F8O + N2

(22) P. D. -rtlett, E. F. Cox and R. E. Davis, JACS § 103 (1961)

V I

IV. The Decomposition of o-Substututed Denzamides and ,nzenesulfonazides

The free radical decomposition of benzoyl peroxide and of t-butyl

perbenzoates is very much accelerated by certain o-substituents.(23) (24)

This ts not the usual steric effect, but a concerted reaction involving a

cyclic transition state, and, in the case of o-iodobenzoyl peroxide, leading

to a cyclic product(23).

(23) Jo E. Leffler, R. D. Faulkner and C. C. Petrupoulos, JACS 80 5435 (1958)

(24) J. C. Martin and 3. H. Drew, JACS §j 1232 (1961)

Benzazides

Similarly, the o-bydro.yl substituent decelerates the decomposition

of benzazide in marked contrast to the acceleration by other cltho substituents.(25)

(25) Y. Yuktawa and Y. Tauno, JACS 8o 6346 (1958)

The usual accelerative effect is attributed to steric inhibition of resonance

in the ground state of the o-substituted benzazide. The deceleration by the

o-hydroyl substituent is attributed to intramolecular hydrogen bonding with

the azide group. No special effects were noted for ortho methyl, bromo, chloro

or nitro, the points for all of which fall on a single isokinetic Line.(25)

In previous work in this laboratory, we have shown that the benzazide

rearrangement still gives the usual isocyana*e in the presence of an ortho-iodo

substituent. In the present work we show that the reaction rate is also normal

and that there is no indication of iodine-nitrogen bonding in the transition state.

Table 5 *hows the rates of decomosition (by nitrogen evolution) for various

o-substituted benzazides.

- i - _ _- -- _-_-_- - - -' -' ' -'• " I I • • l

Bssaaslds Dsoaqiisitian In folufte

Bubstituent Tooi k, X 0*Anl

-AS i44.50 1.87

o-3r& 15.00 2.72

0-I 46.o 3.10

112.2 2.32

Ie30 .05

Ref~.erence (25)

kThijs vork

o-,Jodobenzazide v" o~bained by the reaCtion Of the acid chloride with

N03 In 50% aqueou aoVtow- (26

(26) P. A, S. fuithp Organic DeActions, vol Ins, wiley, (19116)

It melted at 29-30 0and gave at, least go or the theoretical amount of N. on

decceosition. The rate of deccaposition vas determnined by the nitrogen

evolution intbod. (5

A L

As part& of ow research an the InrIleua nte'rection of Iodo-

substituzants; with free radicals of variono t~ypeex ve have found acce2*~wicato

and the formation, of acyclic proftct in the decomposItIon of o'siodobenzoy.

peroxides 27 but not 1n the decoaqosition of o-lodopheracetA2 perozidoe(2 8 )

jr c± o..iodobenzasids. (aee Part IV)

(27) (a) W. Honeberg and Jr. E. Loffler., J. Ora. Obem. 26 733 (1961)

(b) J. It. tafflr,, R. D. 700awbm and C. C. Petropoptaoss, ZU 8054~5 (1958)

(28) j.x.1 raf2.r ank,V, .Wilsmoa J. Org. Chem. pIj 42 (1.96)

law Outwt1us ramnsntof bonmasites probably binvoles a cownoeed

rewation, and a cyclic trunsition staft rather than a nifreng Intermadifts(32)

and thssf owe formtIoui of a cycl.ic iodine comound is possrsW only if the

iodo msatitwnt attacks the aside popIn a reaction -hich to faster than the

normls attack of the mlpating phany2. on the aziA. ouap. Since the &eceqosition

of banmulon2 atslss -iwjff%..7re4t3-Y Inolves an actual iutemdHaft iaitaece

,wr nitrogen diredIcal vhich selectively attacks solvent wlio(0 we still

ii~nsflLtrsd It like~ly that UiOr.O11 odformtion might take p1ace at

sawc step in the ducompoeltion or o.idoezzeeslfonyl azift. To t-,si this

bYPýBLGi it 15 necessary to COMIn tli, prodUCt (Uhich PrOvet to b&. roMMSl

althmuji certain ocher experim~ents shoved that cyclic nitrogen-iodlie structures

mre caMabe or euioteuce) and to ecapare the rate iith those of other

brnnnamnerAonyl 82: des. %arate also prornd to be noor'ml,

39

Denmensmltnyl az~de vas tirt prepwad by O.rtiue()ag) the

(2•) T. Curtiu. , Z. Prakt. Cha. M 303 (1930)

decomposition products in various aromatic solvents were studind by lermer and

idmison (30) •owever, no kinetic stiAdy of the decomposition has been avalable,

(30) 0. C. Dlrwr and M. T. dmison, J. Am. Mum, So. U 70 (1955)

in contrast to numerous studies of the decoposition of benzazides. '1") (•)

(33) M. S. Nevman, S. R. Lee,, Jr. andA.D. Oarett, 1_i.J. & 113 (19147)

It. S. Nevzon and R. V. Gildenhorn,. ibid, 10p 317 (19W.).

R. A. Colmn, M. S. Nemmannd A. D. Garrett, Ibid. (1954).

(32) Y. Yukakma and Y. buno,, Ibid. 5530 (1957); ibid. ~ 3. 15)ibid. 812007 (1959). 64 15)

Recently, ve have learned that akebayashi and Sh±vki have studied the

decomposition kinetically and that the deoapsition was found to be accelerated

by thiophenol. and also by radical s rees .(3).

(33) M. Tbayashi and T. Shingaki, Private eommunication. The authorsare indebted to Prof. A who kinly mad hi data availableto us prior to publication.

Orr own study has be=n exAteded to the daeompositioU in the preaSn

of peroxides. (Part VI)

/4-0

Benzen.sulfonyl azide The method of preparation of benzenesulfonyl azide

was exsentially the same as that reported by Deorer and Edmison(29)

Beuzenesulfonyl chloride .was alloyed to react with sodium azide in an aqueous

alcohol o: aqueous acetone solution. The product was taken up in ether, washed

well with ice water and dried over sodium sulfate. (When the solution vas

dried over calcium chloride, the solution bubbled and the drying agent turned

pink or red, with evolution of heat.) After the solvent vms completely remoyed

at reduced pressure, the remaining oil solidified on cooling in an tee-salt

bath. The product was recrystallized from ether-petrol ether mixture twice and

the 3olvent contained in crystals was removed as completely as possible at

30--40under reduced pressure. The azide melted at 13-I&, and decomposed with

bubbling at about 135%

o-Nitrobenzenesunlonyl azide To an aqueous acetone solution of sodium aszide

(5g in 75 ml), 15g of o-nitrobenzenesulfonyl chloride dissolved in 75m1 acetone

was added dropvise vith stirring at -10'. The stirring was then continued for

one hour at this temperature and another hour at room temperature. The solution

was filtered and diluted with 5O0wi icevater to precipitate a yellow solid,

which was collected on a filter, washed with cold water and dried. The product

was dissolved in 1M1 of warm alcohol and. insoluble material was removed.

12a of yellow needles having mp•p 68-71 was obtalned on cooling. Reerystalli-

zations from alcohol gave big needles, m.p. 71-73.

Anl Coed for C6N)4N14OJS3 C. 31.58; R. 1.76; N. 241.56

Found: C. 32.15; 3. 1.99; N. 24.T5

Infrared spectre in a Nujol mulls had a characteristic azide band at 2170 cm"1,

_o-.odobensnuuslfonyl asie Seven grms of o-iod obeznw ltoM chloride in

20 ml ethaol was added in svrlpwfn!to the colld soL-Ition oJf Ovalkiau

stide in 70 ml aqueous ethanol (30%). the solution vas stirred for 2 houns at

"this tAmerature., and then gradually warmed to room tepesue. mixture

vas diluted with an equal amumnt of cold water and extracted with ether. the

ether solution was washed well with cold water and dried over adb•droue sodium

sulfate. After removing the solvent In vo, 4.55g of an almost colorless oil

remained. Oe oil solidifiled at -30, and the solid was recrystallized from

ether-petrol ether solution, giving colorless needles which had a m.p. 30-326

and decomposed above 1,j : frared spectra in chloroform solution showed-1

characteristic bands at 2120 and 2330 am

p-Methylbenzenesulfony. azide was prepared in the sam manner in 60% yieldj

colorless needles, n.p. 19-20.

MI:xed --7lene which had been purified for kiietio purposes was used

without further purification.

p- was fr*sly prepared by reluxing the sulfur-free xyles

with sodium for 20 hrs, foilved by distillation, b.p. 1 3780.

Chlorobenzene vas dried over aslc•m chlori4e and distilled, b.p. 130.5-1.0°

Nitrobenzene, dried over calcium chloride, was distilled under redueed

pressure, b.p. 12e/32 m.I

Kinetic rperimet 2e rates of decomposition of benzsenefufonl aszides vere

estinmted by folloving the rate of evolution of nitrogen. 2he procedure Ma d

apparatus we the sam as those describd peosly(30)(3L) o temperature

* of the bath was maintained at 126.68 t 0.030. One humbed ml, of solvent In a

124& m1 longnack* flask was lImmrsed In the constant bam;erature bath and

allowed to reach the equilibrium temAMertre and pressm.e Several boiling

chips were added before immrsing the flask Into the baths in order to avoid

"the effect ot supersaturation.

gho Velged amount of the asides 0.01-0.005 moles, vas poued into the

flask In IOWl of the solvent. So tr3ak was comoeted to the azaoAtater

throu&h a smll condenser and a stop-cook and the reading of volume wus adjusted

to zero. The St'op•cock vas closed, the solution was shaken wall and the

• ~um sesnt was started i tey. w vol.um change of Vas vas fo.loved

over a period of 75-80% reaction. Six or more readings were mad. during the

initial 20% reaction and at least 20 readings during subseqment two half-lives

of the reaction. The infinity volume of gas vas determined by the reading at

the time af•ar 8-9 half-lives of the reaction proceeded and an additional

reading vas made after another half-live in order to asure that no leakage

of gas ocaurTed.

Decomposition of benzenesulfopyl aszde Several preliminary experiments vere

Made and the products of rate runs vere treated Ina sa mlar manner.

aenes.fonyl aside, 2.54 ,vas decomposed under ref•iu in 251a

p-xylem for 20 hrs. goh resulting dark-colored solution mas coacentrated

and the rm Idue h tMOItog-aphed throu&i eLmina with bewn=n, bensene-ethers

ether and alcohol, successively. Each fraction vas treated vith chareoal

in ether. From the first three solvent fractions, 2.799 of yellov needles

having m.p. within 129-38°was obtalid. On rearystallisation from alcohol,

the mbstanoe Save 2..IO of colorless needles (m.p. 13-8), Identified as

bensnesnlfony2_-p-xylIdide. From the alcohol elvent o.38S of slid,, a.p.

135147scointaining same oily matter vas obtained. The mixed malting point

vith the sdids as depressed and shoved 138&52 •vIth be=mnsulfonMl ad.

the product from 2.36g aszide in ohlorobenzene vas separated by the

chwoaato api method Into O0.9gt of oily cwytals., 1.1; of needles having

m.p. 123-8*and 0.36g of plates, r.p. 11h-20'. 2he first product gave crystals

having m.p. 116-20P by the rearysta2izaeion from ether. fte second wa

recrystallized from alcohol,, giving needles m.p. 12M-9%. From the third

Vprofct, colorless plates having m.p. 119-21 were obtained. Ohe first product

vas identified as the _-chloroanllide, M second es Identified as

o-chlosnilide and the third as thberloranlide, by mixed melting points.

toac zperiment, oxygen vas passed into g-xylene for 22 hre. sat

room tamperature and 0.9232S of the szide wse decomposed in this solvent.

The -homtogrsapb of the product gave 0.58g of benzenesulfonyl-p-xylidide

and 0.07 of oil from the benzene eluent, and a large amount of black mterial

from the alcohol eluent.

Decomposition in nitrobenzene gave a lag amount of black tar, black

solid •whch vas insoluble In organic solvents and had no m.p., and a s*ll

amount of needles. This reaction resulted in about 60% excess amount of gas.

The gs liberated iodine from potassium Iodide solutiont and probably contains

nitrogen oxide derived from the n1trobdnsene.(29)

Mw decomposition of benzenesulfonyl azide in the presence of t-b7ty1

hydperoide in p-xylene solution proceeded quantitatively and gave a colorless

solution. On removing the solvent, benzenesulfonyl-p-xylidide vas obtained in

quantitative yield. (Bee Part vi)

Dac0Woition of o-Io•. o MbnO, aside . te product of the

decomposition of 2.72S of the aside in mixed xylene vasn chzmmotoaphed and

gav O.&%g of 0-ioobnunaufony amid., x.p. 161-f and s1101l amount Of

crystals , m.p. 81-116° vhtch vould be a mixtore of the *-Iodoben w.nesulfonyl-

xylidides. A large amount of black said was also obtained.

Decomposition of o-Nitrobensenelfon aszide Two grams of o-nitro-

benzenesulfonyl azide us decomposed at 1 30 -4 0 1n 25m1 of -xylene. ahe

gas evolved iws mixed with air and passed into a potassium iodide solution

containing a smeallmount of acetic acid. After about one hour, the

solution gradually became red. 2he amount of iodie liberated corresponded

to 14.26a of 1-N sodium thlosulfate solution. This may be due to nitrogen

oxide.

In another experiment, ig of the azide was decomposed in 20m

r-xylene, and the product was chromatographed. The benzene eluent Save

a small amount of brown crystals, m.p. 177-80 (mixed m.p. with

o-nitrobenzenesulfonaiddej, 167-7-T ). The alcohol eluent. gav O.Wig of

colorless plates of o-nitrobenzenesulfonyl-p-xylidide, m.p. 141-2,5% on

recrystallization from ether. A considerable amount of black solid remained

in the column. This material was insoluble in the usual organic solvents

and In wat•r but seemed to be soluble to some extent in cone. alkali.

fTe decomposition of 1.559 of the azide in mimed xgdene av a

large amount of black solid, a small amount of tar and 0.6 8 g of crystalline

product. Mhe latter was treated with charcoal in ethereal solution and

recrystallized from ether-petrol ether solution, m.p. 100-le11 . This

product is presumably a mixture of isomeric o-nitrobenzenesultonyl-xyliditdes.

Results

knwsnesulfonylazide and the substituted Lamunesulaonyl asides all

dewxmosed at moderate rates at 126.68 in evey solvent.

In mixed xylenem, the decomposition occmured queantitatively giving

nearly the theoretical amount of nitrogen, except for o-nitrobensenesulfonyl

aside. Tft Geacqositiao of beem am y aside followed the first

order kinetic law except for initial 10-1% period of the reaction in

wne seampl of mixe xyleme, doring thick the rafts vas 100rhdlY biPsr than

the sWmbqumt first order part, as show In I1FZr 5. 90h 0 hihr rate

at the Initial step was also observed In the oases ao the o-9o04 and

r-03 derivatives In that sample of mixed xy2.nee. (igUre 6)

On the other hand, oowpletel first order rate constants Ve0"

obtained to freshly prepared -3bme alnoroad crude xylem.

from a different source. *in the presence or molecular iodine, the

decomposition did not show the Initial fast reaction but SNv a straight

first order plot eve in the miued xyle, AlthUMh the rate ws lowered

to some extent. The deccapovition In the product solution also did not

indicate such an irrePlaritr (Fig. 7). Me pecularities of the reaction

In the one mixed zylene sample lead to the investigtion of the reaction

with peroxides described in part VI.

In all cases, the first order rate constants obtained were

iee n of the Initial concentration of the asides. These results

are summarized in TWbe 6.

The decomposition of o-nitrobenuenesulfonyl aszide In xylene

yielded a gas in about 166 of the theoretical amount of nitrogen. The

change at ges volume followed the first order k:,notic law and the rate was

higher than that of bensenesulfonyl &side. Changes of the Initial

concentration showed no effect on either the excessive gas yield or the

rate constant (Flomr 8). no addition of iodine also had no eftect. The

decomposition of benzenesulfonl aside in nitrobenzene •am' the eameresults as the decomposition of the o-nltrom except that the reacion

followed the first order kinetics only after the first 1% of reactio. The

rate constant us eqnel to that in mixed xylene (Figure 9).

00

w 04 a 0 *-4,0 X: 0 H 0 ob

H4 WN N 0

04

r-4 o H 4 H 0 H

to r-4 0 0 m S S 0% 0 I

H- H m'. 4'. 0 rW-4q0 X 44 0 0'. 0 0~ 0

-r0 4 0 0 0 0 * 0 a.

0 ON C) 000.0 ~ 0

0 f-44 05G

0 0 0

00 0 C

C)0 1 O

a 444

Ol 0m r- A 0 9 a ll 0 %4 44 )

(a 0 4 4 4.440 4)

0 0 %U)

H ~ F-44~ 4 04)

04 4. 4-3 44W40

0m

UIA%

r4 0 00 00 1- 0 lS

5 . Ul% r4 H- Nl0 0 0 M: 0 44 > 10

$3 53 0 r-4 0 HeIV4 r44 0% N

* r r- 0 0) -44 r44 TI cw. 0 0 0 9 r r wr4 C r.4

C; 4 )*~ 4.)~ 44 1b;;5 WC) 0 r H

4 e4 IDc'J m 4> r,ON + + H4 IA 0 E4

0 r-40 %0 N 04 WU C0 0 0 H- %O H4 () NV He * 0 04 n 0- NO0 0 N 0

0 000 . 0 v 0 * 0

0 ' 0H 4 NC H C H'4H 0 '.0 OD CO 0% W H r- 0 H

44 ; %r 1 04 r4 r-444 4' 14 4 r4

0

) m~ 0 0 co- ~W . ~C oO0 4 0 0 H 0 0 0)W' 0 WO, % 0%

0 0 0 r. 0 00 0

00

00

.4-

a 0 0 0 0

N O 0'ý %0 % 0 (~0

00

1- 00S H

XO

C% cNCIH (% 0 00- 1

H 0% 0 0' 0R C; C; 4. 0 00

0 (V' 4 (~ ~ .- H 0

.40 co t0 w N c- % %0 C* 0 0 0 IN00 0 0 0 0 0 * r 0- N r-

C4 H 4 4 4 4 4

m o N NH- U 0' 0 cp 0% M. m' %cU - 4 U- 8~ n N" CH 4 " (V' .H N NN 4 H

00 0 0 0 0 0 0 * S S 0 0 02

00

.00

94OH c. ' .0 H N C t' 0 0 % ) 4 3'

As cam be seen from ti data in Table 6, the deooeposition of

benzenesulfonyl atide is rathir insensitive both to the nature of the

solvent or of the substituent. For example, the rates differ very

little in p-xylene, chlorobenzene and nitrobenzene. A pars methyl

group seems to decelerate onl4 ve7 s.ightly (at 126.68"). mh

o-nitro substituent acceleratets the gs evolution moderately, but

obviously leads to a different reaction since more than one mole of gas

is evolved. Tf the rate of disappearance of azide is measured(34)

rather than the rate of appeamice of gas, the o-nitro group is found

to decelerate slightly. So do *;he o-methyl and o-iodo substituents.

(34) BY the triPhenYlPhosPhine method, Part VI

In contrast to the benzenexulfonyl azide decomposition, the

decomposition of benzoyl azdes(32) is accelerated by one or two orders

of magnitude by ortho substituents. Par substituents decelerate the

decomposition of benzoyl azide(32) but do not follow the P 6 relation-

ship established by mets substituents in the same reaction.

The behavior of the o-iodo substituents was of special intereit

because of the possibility of attack of the nitrogen diradical on the

iodine atom either during the rate-determining step or later. 8ince the

rate and product Are nowml, this does not occur. In this respect the

decomposition of benzenesulfonyl sitde reseubles the decomposition of

bensoyt a&ide rather than the decomposition of bensoyl pe ldes.(2)(35)

(35) Chlorination of o-iodobenzenesulfonSmide gives a yellov iodo-diebloride m.p. UOvhich decomposes to the original am6e.

Treatment of the iodo-dibhloride vith dilute alkali, followed byneutralization gives a yellow rwswhtme., probably

melting at 130-13*with deocmosition. fte latter cooretwith acetic acid to give a substance decoMposing t 103-at 06 andhaving the usual amid. bands In the infra-red. It is probably

We plan to Inveotipte the deco ositIon prodcts of

o-tolenegulfonyl aside for signs of intramoleculer bhdrogen transfer.

..00aea e o• e ccltiob productv\

appeais to be amunkn.

'-2

oN

600

B, ~oIn. oM, c

a

0

a-

a

a

L I

F:: � �, Pc :.om posA � a. Z0 t, � �fo�Io�;d�

%S-3

L I

AlN ped .4 V,

g-

Q -3

a

- - M . . . .. b . . f

)--ii i1 It. I Soo x )0

10 I.._

• ,, i i i j •j '1

(40

Y-':+yo 0

n ýevl~ncý N p~c.-mf0

Part VI. The Mutual Acceleration of the Decompositi of

In part V we reported that in certain samples of xylene the rate of

decomposition of benzene sulfoniflazide wes accelerated by some impurity. The

impurity was soon consumed, and the rate became accurately f- rst order.

Because we suspected that the Impurity mi 3ht be a hydroperoxide36 we did some

36See H. Heck and S. Lary, Ber. 76B, 169 (1943) for x-qlene autoxidationo

experiments on the simultaneous decomposition of benzenesulfonazide and t-

butyl hydroperoxide. We have been able to show the existence of several

distinct reactions in this system. However, even though we carried out more

than one hundred kinetic runs, the behavior of the system proved to be too

complicated for a complete anajlsis of the reaction mechanisms. Intermediates

frcm the azide accelerate the doecumposition of the hydroperoxide, and

intermediates from the hydroperoxide accelerate the decomposition of the

azide in chlorobenzene, but not in p-xylene. Both ccmpounds are known to be

subject to decomposition induced by certain radicals; sulfur radicals in the

case of the azide37 and carbon radicals in the case of the hydroperoxide.38

In addition, the decomposing azide is known to initiate vinyl polymerization. 3 9

371(. Takebysshi. and T. Shingaki, private coiunication.

38(a) D. K. Morse, JACSp, 3375 (1957). (b) V. Btannett and R. B.,

Nesrobian, JACS,. 7& 4.125 (19a0

39j.. 1. ,eacock and N. T. Ndmisoi, JACS, 2, 33 6o (1960)

The peroxide by itself can decompose by two paths, one leading to t-

butyl alcohol and oxygen and the other, a 0-cleavage reaction, leading to

acetome, ithanol, and onlY traces of gas. 3 8p04 The 1-cleavage reaction is

|I

J.A. Nil". and Do 14. Surguzior,, JACS, 68, 205 (19146).

the usual one in the gas phase and in dilute solutions, the oY4gen-generating

reaction being favored by high peroxLde concentration, for example in the

neat peroxide. It is the rxygen-gentrating reaction which is induced by the

intermediates from the deccoposing a t:de.

The products of the deccIpositicii of the azide are changed only slightly

by the presence of t-butyl hydroperoxLde. In 2- 1 the yield of benzene-

sulfonyl xylidide becomes nearly qtarwtitative at the expense of the benzene

sulfoamaide. The solution also rezains colorless rather than developing the

usual dark color. In chlorobenzene plus t-butyl bydroperoxide the main

product is still a mixture of substiýuted benzene sulfonanLlides, but the

yield is lowered, and there is sawe benzenesulfonamide as well as considerable

tar.

Reaction Plte~s in P-MVln

The decomposition of the aside plus t-butyl bydraperoxide in p-.xylene

solution gives nitrogen plus oxygen. The oxygen can be removed by meane of

alkaline pyrogallol, and the volume of the remaining gas corresponds to a

theoretical yield of nitrogen. The initial rate of nitrogen evolution is

very closely equal to the initial rate of total gas evolution, since the

deccmposition of the peroxide is Initially very slow, as determined by

titration. In xylege solution the presence of the peroxide has little or

no effect on the initial rate of nitrogen evolution and certainly no large

effect on the nitrogen evolution rate at any time. Using the first order

rate constant 1.44 x lOmin 1, obtained for the decomposition of the aside

in p-xVlenealone at 126.70, the rate of oxygen evolution my be estimnted as

a difference betwen total gas evolved and nitrogen evolved. The results of

58

such a calculation indicate coniderably less net evolution of o.ren than

expected on the basis of the hhWdreroxide titer and sugaest that oxygen is

being consmed by solvent radicals. With excess peroxide the rate of total

Ss evolution steadily increases during the run, eventually becoming too

rapid to measure.

The rate of peroxide decomposition as determined by iodcostric titration

is slow and nonreproducible for the firt five hws or go. SU cOmtexaM4

to acut 5% de 0o a stio of the peroxde .d about 39% decaomoetim of the

.314. he rates in the latter put of the er a four or five times an

feet mw4 ane repozle (tUOl 7). SeY appeOr to be a 1ine0w funotiom of

the initial aside eenaentzaem. foe point for zra asid ewoeentratla tfals

wU below the line, howver.

Mosat of o e xerasetts wz done in alobenasne in order to avoid the

omp.lioatlio Of sawivnt Mtamiistlon. M resit for the loMo--nAe--

ewpe-iments wre edmerfte in the follovIn sections.

MK m&%To Wr - gnT MAI W=I to-D.oIN p4T1V kA! 226.70s, 1 M __

Initial Conc. N/L k 0 x 0

(Asidel ]I t-DUO0E] NfL. Xim

0 (.05) .o3

.0504 .0503 .82

.05M7 .0498 .78

.0507 .0499 .86

.0177 .0503 .50

.0333 .0698 .62

InUitS~al rate. noe apparent rate-law in absence of azide isfirst order. It a used to estite tOn initial ko for a run ofinitial Concentration .05 Molar.

Pipre 10 is a vat-c r pviot of the sas evolution rate in the

presence of Vario" sMnMs of hAfO. ter a period of about 30

minutes during ifbih tba ips evolution rate is about that evpeated for

the aside aoNe, te is a period of about )5 minutes ring ida• h a"

evolution Is very rapid, f..mvde by •as evoution at the e2pM~ Vat.

for the azide alone during tL rest a n rm e period of mpid gas

e"voluton averlaps the period %ar-ing vhick the hy wopezid Is Gscem-

poing, MnA N gas evaved dw'Ing this perlod I* a mixtWr of nl-agen

and oxye. fwovev, It an easo be sheitbat tbw has been a

tfmpoa. soaleration of tIh 11side doup"ltIM. te! tI Of the

linear portion of the curv In figne 10 back to zero time gives an azide

concentration considerably less than the actual initial azide concentration.

We were also able to con:'ilrm the rapid azide decomposition rate by

direct analysis for azide. Th.i method used is reaction with triphenyl

phosphine and is described in the experimental section° Again we find a

decceioeition at the normal sl,sv rate for the azide alone, followed, first,

by a period of rapid deccoposit.ion, and then by a period of decomposition

at the norml rate. A typical run is shown in figure UI, which also shows

the decrease in iodowetric titer of the peroxide. It can be seen that the

beginning of the period of accelerated azide decomposition coincides with

an acceleration of the peroxide decomposition. The peroxide deccoposition

rate in typical runs remains constant from the time of the sharp break

in the curve to the end of the reactiono

The decomposition of approximately equ!l conecatratione of the azide

and peroxide in chlorobenzene in the presence of iodine gives the usual

mixture or benzene sulfonanilides, but without the black tar. The rate of

60

ngur 10 Deo~ostio ofthO azide ini cblorobazieml

In heprsnce of t-butyl hbrdrayerozide.

A. Msidej, 0.05M; b~rdz'OPeO=id*, 0.02%1

N. Azidep 0.0,5U bydroper'oxide, 0.05 X1.

0. Mside, 0.033313 hydropelucidd., 0.0,.M

ei

EAI( 0 037? Al

Y 10 Oqo A4

44

AI 7. o -A//

A-Aide

c~l,,/ -e • / ) .7

/ 003

/0 -, e t,,

. ./S--

o 1£ .. . . • ...... ... •-

62

total gas evolution fits two simaltaveotz4 first-order processes, one for

the aside and one for 1-he peroxide. 'he first-order rate oonstant for

peroxide decomposition under these ccriditions can be obtained independently

by iodcaetric titration, The kA for the aside deccosition used to

calculate the curve shown in figure 10 for the azide-peroxide-iodine system

is the se as that observed for the aside or the aside plus iodine. The

kip used for the peroxide, howover, is a function of the initial concentra-

tion of azide. At very low aside comnentration the rate of peroxide

disappearance is initially viry fast, bscoming first order only at a later

stage of the reaction.

Figure 13 shows a first-order plot for the disappearance of peroxide

in a system containing iodine and also aside ir- concentration about equal

to that of the peroxide.

In auury we oan say that the aceleration of the aside dei2gosition

by the peroxide a!pears to be due to attack by radicals which can be

diverted either by reaction with p-xylen (solve-at) or by reaction with

iodine in chlorobenzene. In the next section we will oxmine the effects

of the aside on the peroxide decoposition rates.

Mhe Effect of the Aside on the Peroxide Deeosaition

t-Butyl bidroperoxide in chlorobenzene solution in the absence of aside

decc~oses by a first-order process into products that are largely non-

gaseous. The first-order rate constant at 126.70 is about i0.& x 10-3 Mini,

for runs of initial peroxide concentration near .065 nolar.o1 In the presence

41the rate ca probably be changed by scavenging of 0 2 from the systemby mens of a nitropn stream. Bee reference 38a.

of the aside the gas evolved is nitrogen in the theoretical swunt plus

o0qp in 90% or better of the theoretical mount. The s aixture is

63

liguw. 12

2.0

ea c-4- v/'4!/o -~ 0ei # 0/: ' oq id a-A

?9Alr~ A ~~e

g, pK~d~

8. 5a~-w 0a 3 &-k'~i

64

0 -00

0 2 0 N

U . .... .. 1.. . .. 0 ID

Figure 13. Change of the concentration of t-butyl hytdroperoxidf!in the reaction of the azide [.05 molar] and thehydroperoxide .05 molar] in the presence of iodinr:[.0O1 molar).

evolved rqidly at first, then decreases to the first-order rate of

nitrooen evolution characteristic of the azide by itself.

Iodcmtric titration of the peroxide in chlorobenzene in the presence

of the azide shovs dectposition occurring by a process.of nea:.'ly zeroth

order, especially at iov azide concentration. A typical run is shown in

figure 14. However, at high concentration of the azide the rato of

peroxide decomposition (k OP) increases with tine, and we have used both the

initial and final zeroth order rate constants. The zeroth order rate

constants axe related to the initial azide concentration (the aside

concentration does not change drastically during the peroxide deccupositicm)

65

½

C.O)

0\0

Scoo

0

4o , o

Fgare 14,. Zeroth order plck; of the decouposition Qf t-butylbydroperoxide ir chlorobenzene at 126.7.o

Akzde- .0060 molarPeroxi•e--°050 molar

as shown in figure 15 for runs o * initial peroxide concentration equal to

.05 molar. The upper branch at high azide concentration shows the rates

late in the run; the lover brat ch shows the initial rates.

The effect of changing initial peroxide concentration at a constant

initial aside concentmtion of 01 molar is shown in ftlure 16.

The ocobined effect of cbties in azide and peroxide concentrations

on the peroxide deccmosition rite can be approximted equally vell by

66/.0

M /1 .*it

0A~CAI MA " 0

pe ox,'de )-•' ci' o .x.zide comc1

several equations. One of the-e is equation (1),, which uses only rates

from runs of low enough azide ,-oncentration so that the runs are fairly

accurately zero order (I.e., (A] 0 .017).

M .&5 x lo"3 [) + 26 x lo"3 [P•]AI1/2 + 3.1 x 10o3[AI(Squ 1)

The constant 3.1 x 10-3 is approuimately twice the first-order rate con-

stant, 1.5 x 10-3 mun-I, for the decoosition of the aside alone.

By, using the steady-state qproxlmtion and assuui tnterplay between

14.

050.-"

07 67

00

0

deelplojt o/

0 /o5

as mell as the aside and the peroxide, it is posaible to derive several

kinetic equations vhich feature a cross term containing both the aside and

peroxide concentrations and a 2klA[A] term. The reaction scheme shown

below is an example of such a echnism. It ewlains the term proportional

to [P] in equation (i) as a uniwlecular rate oa peroxide decomposition,

The observed kI for the peroxide deccmposition In the absence of aside

is about tvwie as great as the proportionality cnmstant in equation (1),

however.4 Like the other simle mechanisms it is not ccaletely satisfactory.

42~s mwy be an effect due to sweeping out of 02 kr the Sas evolvedfr the azide. See table 6.

S• .

68

An addAi~tLCWJ defect Is that this mechanis neGlcs3ts the Wnuced doom-

position of the &aide; howvver, this is lesss Impiortant at low ooscentraftons.

§Stead Stt Derivation for a 8Sllified X~~d

(Specal symbols ar vrittes, under the corz'spzadIng opeais* in th. chemicalequations.)

~k4#3 4So1y -V'+ AA N,

I _

Aoo' y3-4Poo R24 ~OO'- M3 --

d r3 [P LW LPLA#

C)~ cfrA3 Ao

7r9pd ZV.-7 ~ 1A, 1 r~lbV

-~ ~ dIJ I ~3[fbJ4L#L#7 J L",73~

-ic-

dl-

69

The Effect of the Aside on the Peroxide Deoqosition Mte

The rate of decomposition of t-butyl hydroperoxide by itself in

dodecane 3a depends on the rate at which the xqrgen produced is removed

from the system, apparently because oxygen diverts radicals which are

particularly effective in inducing the decomosition of the peroxide. The

effect of sweeping with helium is an acceleration by a factor of 4O to

50. We find that the effect of nitrogen-sveeping on the rate of

decomposition of the peroxide in the presence of azide is very much less,

as can be seen from Table 8.

Allowing the azide (GO06 Molar) to deccsvose for one hour (about

7o5% decomposition) before adding the peroxide (.0507 molar) gave a

peroxide deacuposition rate about twice that of a cco~arable run in which

both aside and peroxide were added at the same time. On the other hand,

the me'or aside deccmposition products, benzenesulfonamide (.003 molar)

and benzenesulfonyl-p-chloroanLlide (.0009 molar), had no effect.

We next studied the decoqposition of the peroxide and aside in the

presence of iodine as a trap for radicals, in the hope that the peroxide

deccmposition kinetics vwtild be simplified.

Table 8

Effect of It. Sweeping on Rate of Decomposition of t-Butyl Hjdro-pero~ide in Chlorobensene with Benenesulfonyl Aside

initial COmi. k x 103 Moles/Lo~n.PeoieAieUnder N'~ Swptb

.05 .001o .08 .24 to .34

.05 .00o 4 .13 .21

.05 .0o2-.013 .90 .35

.05A .36

TO

The Effect of tha Azide on the Peroxide Do b1mTte-n -n Presae o lane-

The decumposition rate of the peroxide in the absence of azide is

dependent, as we have seen, on various factors. For exzple, it is such

slower in p-xylane than in chlorobenzene and is somewhat faster in nitrogen-

swept chlorobenzene than in chlorobonzene in the presence of the evolved

oxygen. The decmposition in ahlorobenzene, unswept, in the presence of

.01 molar iodine is a first-order reaction with kip3.4 x lO3in"I, about

three times as fast as the decomposition of the peroxide alone. However,

this reaction of the peroxide alone is attended by little or no Gas evolution

and is not the same as the apparent first-order deccmposition of the peroxide

in the presence of iodine plus the azide.

Again, we find a change in the kinetics of peroxide within a run at a

critical azide concentration (about .0124 molar). This time, however, the

kinetics are more complex at low azide concentrations rather than at high

azide concentrations. At low azide concentrations there is a very fast

initial peroxide decumposition, not leading to gas, which quickly subsides

and is followel by a slower, gas-evolving, first-order peroxide decomposition.

A typical run of this type is shown in figure 17.

The Initial fast peroxide decomosition (figure ) has rate constants

of the order 10 x 10-3 to 60 x 1.0'3 Muin"I as ccmpared to the later rate

constants of the order of 2 x ý.0-3 to 6 x 10-3 Mln."I The initial fast

peroxide decomposition requires the presence of sam azide but is absent

at hiUier concentrations of azide, Tables 9 and .0 show the rate of the

fast initial reaction, if any, expressed as a first-order rate constant.

It can be seen from the tables that the initial fast reaction is regularly

ob'erved whenever the initial azide concentration is below about .015 molar.

It is absent or slower than usual, however, if benzenesulfonyl p-chloroenilide

71

Pigurs 17

Peroxide Deccoaition in the Presence of .01 X 2 an.0016 olar initial aside concentration

iin I WI I0 20o 4c0o

(a product of the azide decomposition) is added at the beginning of the run.

We therefore have an induced deccoosition of the peroxide dependent on

iodine and azide but inhibited by the anilideo The second stage of the

reaction in the presence of iodine exhibits fairly precise first-order rates.

Figure 18 shows how these rate constants are re3 ated to the square

root of the initial azide concentration at constant initial iodine and

peroxide concentrations. Neglecting the rates for runs without azide

(which are too high, see the figure) the plot passes through the origin.

Figure 19 is a plot of initial zeroth order rate constants for the

peroxide obt&ined by multiplying the observed first-order rate constants

by the initial peroxide concentration for runs of constant initial concen-

tration of iodine (.01 X) and a-ide (.o164M).

72

abls 9

Ic4smtrio taes of Peroxide Decmopoitimat-P A A23•X IN

S126.70; IN YUm I ilI oli n z M

Initial ewo. N/L k, xoe 1 n. 4 /A]

First Order initialtiON 0 0 112 o Part aset Reaction

11-74 .o05o .0000 .0100 .33 absent .15

n1-75 .05OO .0000 .0101 .35 absent .16

11-61 .0501 .0o2.4 ,0101 .375 2.0 .27

11-14 .0503 .o06k .0101 .415 1.0 .24

11-65 .0502 .e8 .0100 .506 absent

n-17 .0500 .oe2 .00•6 .518 absent .23

n1-64 .0501 .0287 .0101 .562 absent .20

II--.5 .0505 .0348 .01ce .68% absent .20

n1-63 .0501 .0364 .0100 .665 absent .20

260 .0502 .0500 .0101 .766 absent .21

n-6e .o05 .0514 .0101 .761 absent .20

n-72 .oell .0164 .0100 • 537 absent .28

I-T7 .oe63 .o064 .0100 .493 absent .27

n1-69 .03417 .0164 .0100 .44T absent .30

11-66 .0507 .o164 .0300 .375 absent .28

11-67 .o564 .0164 .0101 .32 absent .30

n1-68 .0680 .o164 .0101 .364 1.8 .30

11-70 .0832 .0164 .0100 .350 2.5 .38

11-17b .0o97 ,A .019o .607 absent

Ir-6o .0W9 .0016 .0101 .1814 3.5 .,45

11-08 .0501 .0031 .0100 .265 2.8 .454

n-18 .0501 .00•9 .0099 .2c 14.0 .40

hz-16 .0302 .0091 .0100 .297 2.4 .23

n1-59 .02 .o0031 .0100 .a8 5.8 .40

I

iifit

opal 40

74&

Figure 18

First-order rate constant for peroxide deccposition in chlorobensene at126.70 in the presence of .01 K 12 end .050 d. initial peroxide concentra-tion. Mhe solid points reprsent runs baving an initial fast part ieoserat s not plotted.

/0

00

0 6I,

.. Figuvre 19

Initial zeroth order rate constants and peroxide concentratior for runs havingconstant initial concentration of iodine (.0l x) and azide (.o164 x).

.3

00

.11

c)N i - I

75

The data of fi•,-res 18 and 19 can be sumrized b equation (2):

ko e . x 10o31(A] ÷ +23 P]o[JAo• (Uqu. 2)

Since equation (2) differs from equation (1) in having a balf-order tern In

aside In place of a first-order teorm, we tested the fit of the data to an

equation wre nearly like (1) and found a fair degree of fit using the

coefficient@ In equation (3).

k(N/L min) SW (0 to ox1"3) x [P] + (15xo"3 to gtlo'3)[P](AJ* 4

3.1 x l0- 3[A] (1Nquo 3)

7n view of the experimental uncertainty in the rates at low azide concentra-

tion in the presence of iodine, we believe that equation (3) merits Just

as much consideration as (2)0 If so, the main effect of iodine (except for

the initial fast reaction) Is to reduce the importance of the term in

[P][A3ý, perhaps by diversion of t-butyloq free radicals. The roason for

the shift from zeroth to first-order kinetics within a given run remains

obscure.

190rime Pa frt

Trtphenyiphosphlne Method for the Aside

A five 1l. aliquort of the azide solution is added to 10 ml. of benzene

in an azotcmeter equipped vith a magnetic stirrer and a rotatable sidear&

for adding a mixture of 5 ml. of triphenylphosphine reagent and 3 Ai. of

acetic acid. The triphenyl phosphine reagent consists of 13.120 of ? in

50 al. of CW13, The gas b rette is adjusted to zero, the sidearm is rotated

so as to add the reagent, and the volume of nitrogen from the azide is read.

Nitrogen is evolved quantitatively. A considerable excess of 03P is used in

order to destroy any unreacted peroxide.

P?+ 0 802N3-~* :0N 3*'W

+P4C/-4%. C ow -4 0.1 Do -f-I1) C O/

76

A 10 al. aliquot is diluted with 25 ml. isopropyl alcohol, 5 ml.

acetic acid and 5 ml. of aqueous sat. KI are added, aw the xixtue

is shaken and Vaned for 20 ain. at 60-70o. The solution L then diluted

with w•ter and titrated with .05 N thiosulfate.

For free iodine, the samle in isopro]Wl alcohol is titrated with

thiosulfate before adding the KX and acetic acid and vining.

Pert VII

Tripbeny~l2phoph~ne Oxide Peroxide

Attelpted oxidation of the phosphlne imines fr• triphenyliphosphine

and benzoyl azide or benzenesulfonyl azide43 with peracids or hjdroperoxides

4 3Analogous oxidation of ylides gives alkenes. SeeD. Z. Bestmann, c Ca, , 72 34 (1960)jD. B.Dewney and-L 7. S 1 &M., §&2396 (1960).

gave unchanged starting material frcm the sulfonyl phosphine-iine and

benzadde, benzonitrile, and triphezylphosphine oxide from the benzoyl

phosphine-imine. However, treatment of the bwyl pbho.%hbne-Imine with

hydrogen peroxide in mus acetic acid gave a new substance which ve will

tentatively call the tripbenylphosphineAperoxide. It vas subsequently

found that the new peroxide could be prepared directly from triphenyl

phosphine oxide. I

"?.srstIon and Prprisof 1 aenypoehn Oxide Peroxide

To 4.0 g. of 0PO dissolved in 30 aIl. of dioxene is added 15 al. of

30% 3202 cooling in ice vater. The mixture is then a~llowed to stand for

24 houis and diluted with ice vatero The peroxide which precipitates melts

at 127-130 with gas evolution. It my be dried by treatment with Na&SO4

in hot benzene; it crystallizes on cooling the filtered solution. Yield,

3.5 g-, a.p. 130-1310 with gas evolution. The reaction my also be carried

out in acetic acid, but t-butyl hydroperoxide my not be used in place of

%202'

M M

77

311

281

or [ ÷10

quivalent weight by Oe evolved from the &ecMosition in UC1 is 282.

The 03PO obtaied on removal of the OM melted at 154.50 without any further

purification. Oxygen was identified by the gloving splint test and

absorption in alkallne Wrosllol.

CNP An~e M glark Miraav~a Laboratoq

C H P

73.73 6.01 10.5373.85 5.82 10.50

73-895.03Av. 73.79 5.62 10.52

Theory for 40

C H P Equ. t.

73.21 5.47 10.119 301

Theory for 3 ^

C H P Nqu. wt.

75.51 5.29 lo.82 286

LDfrsred Spectrum

The infrared spectrum (in ftjol) is very much like that of 3PO,

except for a band at 3.2 L (- OH dlaer) end the following slight shifts:

Bund In 03PO Bund in (03P0)2 N2o2

absent 3.2

10.85 lo.813.2-13.4 13.1

13.35

Reactions Of ~ p yp~pieOxide Peroxide

Altho•gh a quantitative evolution of oxygen is observed In chloro-

benzene at 126.70, a s=Vle of the peroxide added to a solution of anthracene

in Vlene at the sma temerature evolves no gas at all. The products of

the decnaosition in the presence of anthracene have not yet been Investigated.

The peroxide (.10 g.) dissolved in 15 al. of styrene and wxied at 600

for 15 hours gives considerable nthbanol-insoluble polymr. A control

experlmnt without the peroxide gave no polymer.

Befluxing the peroxide (2.32 g.) with styrene (0.40 go.) In 50 al1

0313 for three days gave an oil having a carbonyl peak in the infroaed at

1700 CM. 1 It does not appear to be acetophenoneo

The peroxide dissolves only very slowly in refluxing cyclohexene.

Besides 0?0 , the product is a m amount of an oil having a carbo•yl

absorption at 1670 cam-. (Cyclohexenone absorbs at 1710 ca'1.) V.P.C.

with a silicon@e-firebrick colum shows only one mior fraction, probably

cyclohexenone.

The peroxide (6.00 g.) and benzhydrol (1.85 g.) in 25 al. of benzene

ms refluxed for three days, at the end of vhdch the peroxide bad been

al2ost caletely consumed. After rmoval of the bensene and extraction

of the residue with ether an Insoluble residue of 4.05 g. of 3PO MAs

obtained. Me other-soluble material was chrmatogrehed on alumina, giving

1.1 g. of be•nzophmone and a wall further mount of 0o.

The peroxide In benzene does not appear to oxidize ep, benzil,

or 1-onthol.

79

Part VflZ.

The generation of dichlorocarbene (fram CKC1 3 ) in the presence of

triphenYiphosphine results in the formation of the corresponding ylide.

This substance is readily hydrolyzed to triphenylphosphine oxide, It will

react with carbonyl compounds to form tertnal alkenes. Since parallel

results have in the meantime been published by Seyferth, Grim, and Read 44

"DSeyf"rth, S. 0. Grim, and T. 0. Read,, JAJ, 2& 1511 (1960).

we discontinued this part of our program.

Part IX.

Miscellaneous Mmloret=r Inrmet

(1) Reactions of silyl azides and aso ecmpounds were discontinued,

(2) Negative results were obtained in attepts at trapping anthracene

triplet states or diredicals with tripheny1phosphine,

(3) The electron-deficient intermediate frcm the photolysis of •-

naphthalenesulfonyl azide in iodobenzene is not trapped by the iodobenzene

with fora~tion of trivalent iodine-nitrogen ccupounds.

(4) The reaction of tripbhenylphosphIne with sodium trichloroacetate

in ethyleneglycol diwethyl ether gives A product which appears to be a

sixtureof o 0r-Pd;

(5) An attempt at inducing the decomposition of a sulfonyl aside by

means of diablorocarbene (rcam sodium trichloroacetate) had negative

results.

(6) The reaction of diphenyiphosphonyl chloride with diazamethane

appears to give the chloromtbyl derivative rather than the desired

80

di~s~tbderivative.

(7) The dao.oopsitica of diphenyllodonium aside does not proceed by

the dlpbe3Wliodlzae-nitrens path, but oilt by my of the iodossidobipheny1.