You are on page 1of 8

[COXTRIBETIOX

FROM THE RESEARCH


A N D DEVELOPMENT U. S. NAVALPOWDER
DEPARTMENT, 1
FACTORY

Reaction of the Acid Chlorides of Aromatic Acids


with Bromine and Silver Oxide'
F. A. H. RICE AND W. MORGANROTH*

Received June 4, 1956

I t has been found that the acid chlorides of several aromatic acids on treatment with silver oxide and bromine in carbon
tetrachloride solution loose carbon dioxide. The acid chloride is substituted by the halide.

It has been reported3 that when the acid chlorides like the sugar acid chlorides, would react to form
of acetylated hexonic acids are dissolved in carbon carbon dioxide and the corresponding halide.
tetrachloride and heated at reflux temperature with Benzoyl chloride was found to react, with silver
silver oxide and bromine, the acid chloride loses car- oxide and bromine to yield bromobenzene. The m-
bon dioxide and forms the aldehpdo-l-bromo-penta- and p-nitrobenzoyl chlorides formed the corre-
0-acetyl-pentose. Either acetic anhydride or ace- sponding rn- and p-bromonitrobenzenes. Kaph-
tyl chloride3react with mixtures of silver oxide and thoyl chloride also reacted, although slowly, under
bromine and after respectively losing 2 moles and 1 the same conditions to form bromonaphthalene.
mole of carbon dioxide, form methyl bromide. The reaction as measured by the evolution of car-
Although the product formed by the reaction of bon dioxide was essentially quantitative (Table I).
acid chlorides with silver oxide and bromine is es- A possible mechanism for the above reaction is
sentially the same as that obtained by the decar- that the acid chloride reacts with silver oxide to
boxylation of the silver salts of organic acids by re- form the silver salt of the acid and silver chloride.
action with bromine,4 the acid chlorides, in contrast The silver salt would then react with bromine to

TABLE I
OF AROMATIC
DECARBOXYLATION ACID CHLORIDES
WITH BROMINE
A N D SILVER
OXIDE
COZevolved.
Equivalent in
Weight, ml. 0.1 N HCl Compound" w p . , OC.
Compound g. Calc'd Obsv'd Isolated Observed Literature
~~ ~ ~

Benzoyl chloride 1.0 143 140 Bromobenzene 155-1 57 155O


(b.p.)
p-Xitrotwnzovl chloride7 1 95 228 211 p-Nitrobromo- 125-127 126-1278t9
(m.p. 72-73) benzene
m-Sitrobrnzoyl chloride7eT1 2.0 216 210 m-Nitrobromo- 56 5610,ll
(m.p. 32-33) benzene
Yaphthoyl chloride12 1.87 196 116 Bromonaphthalene 59 59'3
(m.p. 52-53)
a Thr compounds gave the cor] rct elemental analyses.

to the silver salts are easily prepared in an anhy-


(5) I,. Birckenback, J. Goubeau, and E. Berninger, Rer.,
drous condition and hence are more convenient to 65, 1339 (1932); W. Bockemuller and F. W. Hoffman, iinn.,
use than the silver salts. We were not able to find 519, 165 (1935); C. C. Price, Reactions at Carbon-Carbon
that the decarboxylation of aromatic or aliphatic Double Bonds, Interscience Publisheis, New York, h'. Y.,
acids via their acid chlorides had been studied and p. 55 (1946).
hence it was considered of interest to investigate (6) J. H. Gladstone, J . Chem. SOC.,45, 241 (1884):
W. Ramsav and S. Young, J. Chem. Soc., 47, 640 (1885).
several aromatic acid chlorides to find out if they, (7) H. &lever, Monatsrh, 22, 425 (1901), Org. Syntheses,
Coll. Vol. 1, 2nd ed., 394 (1941).
(1) Published with permission of the Bureau of Ordnance, (8) J. H. CoRte and E. J. Parry, Be?., 29, 788 (1896).
Navy Department. The opinions and conclusions are those (9) o-Bromonitrobenzene has m.p. 43", F. Ullman, Ber.,
of the authors. 29, 1878 (1896).
(2) Graduate Student at the Ohio State University, (10) I. Hrilbron and H. RI. Bunbury, Dictionary of
Columbus, Ohio. Appointee for the summer of 1955 under Orgnnic Compounds, revised Ed., Vol. I, Oxford University
the summer Research program of the Research and De- Press, New York, S . Y., p. 350 (1953).
velopment Department, U. s. Naval Powder Factory. (11) L. Claison and C. M. Thompson, Ber., 12, 1942
(3) F. A. H. Rice, Abstracts 0.f the 198th Meeling of the (1879 ).
American Chemical Soczely, Minneapolis, Minn., 21D (12) P. T'ieth, Ann., 180, 305 (1876).
(1955); F. .4. H. Rice, J . Am. Chem. Soc., 78, 3171 (1956). (13) A. Michaelis, Ann., 321, 141 (1902), footnote p.
14) J. Kleinberg, C h e m Revs , 40, 381 (1947). 246.
1388
DECEMBER 1956 BROMINE A N D SILVER OXIDE ON AROMATIC ACID CHLORIDES 1389

give the p r o d ~ c t RCOOBr


~,~ which when heated approximately 50 ml. of carbon tetrachloride in a three-
would give R X +
C02.4,6That the above mecha- neck 250-ml. flask. The carbon tetrachloride had been previ-
ously dried over phosphorus pentoxide and then redistilled.
nism may not be the true one, is suggested by the ob- An excess amount of silver oxide (2.5 g. dried over phos-
servation that when either the acid chloride of D- phorus pentoxide in a high vacuum a t 100' for 24 hours)
gluconic acid penta-O-acetate3 or an aromatic acid and 2 g. of redistilled bromine in 20 ml. of dry carbon tetra-
chloride is added slowly to a boiling mixture of bro- chloride, was added and the mixture was stirred and heated
under reflux. A stream of nitrogen from which traces of
mine and silver oxide in carbon tetrachloride, car- carbon dioxide had been removed by bubbling the gas
bon dioxide is immediately and continuously through saturated aqueous sodium hydroxide and from which
evolved. On the other hand, refluxing the acid moisture had been removed by bubbling the gas through
chloride with silver oxide does not lead to the im- conc'd sulfuric acid was used to sweep out the carbon
mediate formation of the silver salt of the acid. On dioxide as it was formed. After the gases had passed through
two traps immersed in Dry Ice-Acetone, to remove any
the contrary, after four hours heating only traces of trace of bromine, the carbon dioxide was absorbed in 0.2 N
silver salt can be found. aqueous barium hydroxide. The amount of carbon dioxide
evolved was determined by titrating the barium hydroxide
solution with 0.1 il' hydrochloric acid and subtracting the
EXPERIMENTAL hlank which had been run without the addition of the acid
chloride, but otherwise under the same conditions. Table I
Preparation of the acid chlorides f r o m the free acid. The acid shows the amount of carbon dioxide evolved from a number
chlorides were prepared by heating the corresponding acid of acid chlorides.
with thionyl chloride a t reflux temperature and, after con- The decarboxylated product was isolated by filtering the
centrating the solution t o dryness under reduced pressure carbon tetrachloride solution, concentrating it to dryness and
a t room temperature, crystallizing the resulting product from distilling or crystallizing the product from ethanol.
petroleum ether (b.p. 30-60'). Before use the acid chlorides
were dried under high vacuum at room temperature. Re- Acknowledgment. We wish to thank Mrs. P. P.
agent grade benzoyl chloride was used. Wheeler for performing the microanalyses.
Reaction of acid chloride with bromine and silver oxide.
Approximately one gram of the acid chloride was dissolved in INDIANHEAD,MARYLAND
280 COMMUNICATIONS VOL. 26

A Convenient Synthesis of Alkyl Halides from mercuric oxide, and reaction in light gives poly-
Carboxylic Acids bromination.
When either exo- (I) or endo-5,6,7,8,9,9-hexa-
Sir: ch10ro-1,2,3,4,4a15,8,8a-octahydro-exo-endo 1,4,5,8- -
dimethano-2-naphthoic acid (11)2was treated under
The salts of R variety of metals and carboxylic these conditions, a mixture of 71% em-(111) and
acids upon treatment with halogens lead to the 29% endo-bromides (IV)z was formed. This is
halogenative decarboxylation reaction RCOOM +
precisely the same mixt,ure of bromides obtained
+ +
X,+ RX MX CO,.' Of these reactions, that from the silver salt of either acid with bromine.2
of the silver salt (generally termed the Hunsdiecker This suggests that both procedures lead to the same
reaction) appears to be the most useful.' This intermediate RCOOX. The reaction may involve
reaction has the practical disadvantage that the the existence of the mercuric salt although (a) it has
silver salt must be relatively pure and must be been reportedla that mercury salts of acids such as
scrupulously dry in order to obtain satisfactory stearic acid do not give good yields following the
yields. In the course of our studies2 on the mecha- normal Hunsdiecker reaction, and (b) glutaric
nism of the decomposition of the acyl hypobromite acid does not. give ybutyrolactone, whereas the
intermediate' in the Hunsdiecker reaction, we a b silver (and presumably the mercury) salt gives the
tempted to prepare this intermediate in another lactone. It seems most likely that the mercuric oxide
fashion. This has led to what appears to us to be a gives a positive halogen species: which reacts with
useful preparative method, as reagents and opera- the carboxylic acid to give the acyl hypohalite.
tions are much more convenient than the silver If this is the case, and if equation (1) represents the
salt-bromine reaction. stoichiometry of the reaction, we do not under-
Treatment of a slurry of excess red mercuric stand why the water formed does not interfere
oxide in a refluxing solution of an aliphatic car- with the reaction as it apparently does in the Huns-
boxylic acid in carbon tetrachloride with approxi- diecker procedure where glassware and chemicals
mately one equivalent of bromine in the dark led to need to be scrupulously dry.' I n any case, the pre-
excellent yields of the corresponding alkyl bromide. parative value of this procedure lies in its conven-
For example, treatment of 0.25 mole of stearic acid ience and simplicity in comparison to those proce-
with 0.25 mole of bromine and 0.19 mole of red dures involving preparation of silver or mercury
mercuric oxide in 150 ml. of carbon tetrachloride for salts.
one hour gave a 93% yield of crude heptadecyl
bromide, m.p. 22" ( l k 3m.p. 32" for pure material).
The presumable stoichiometric equation is as "3

follows:

2 RCOOH + HgO + 2 Brp+ c1 COzH


2 RBr + HgBrt + H20 + 2 Cot (1) / Cl
c1 I
Since our original observation, the reaction has
been found to go in fair to excellent yields with
cyclopropanecarboxylic acid and bromine14 with
9,10-dihydro-9,10-ethano-9-anthroic acid with bro-
mine in carbon tetrachloride and with iodine in I
I11 IV Br
cyclohexane,5 with lauric acid and bromine,e with
stearic acid and iodine,6 and in poor yield to give
1,3dibromopropane from glutaric acid.6 No 7-bu- Acknmledgment. The authors are indebted to
tyrolactone was found in the latter case. Professor John S. Meek, D. T. Osuga, P. W. Jen-
The reaction gave poor yields with benzoic acid ningsand n/IissJ. S. Nelson for permission to mention
and bromine. Tetrachloroethane may also be used their work previous to publication. W. C. F. is
as solvent.4Silver oxide may not be substituted for further indebted to the Shell Development Com-
pany and the University of Colorado for research
(1) For leading references see: (a) H. Hunsdiecker and fellowships.
C. Hunsdiecker, Ber ,7J, 291 (1942);(b) R.G Johnson and
R. K. Ingham, Chem. Revs., 56,219 (1956); (c) C.V. Wilson, DEPARTMENT OF CREM~STFU STANLEYJ. CRISTOL
Org. Reactions, 9,332 (1957).
C. FIRTH,JR.
UNIVEESITT OF COLORADO WILLIAM
(2) S. J. Cristol, J. R. Douglass, W. C Firth, Jr ,and R. BOULDER, Cow).
E. Krall, J . Am. Chem. Suc., 82, 1928 (1960). Received October 21, 1960
(3) For comparable results with the silver salt see: J. W.
H.'Oldham, J.-CChem. Soc., 100 (1950).
(4) J. S. Meek and D T. Osuga, unpublished work.
(5) J. S. Meek and P. W. Jennings, unpublished work. (7) W. Brenshede and H.J.Schumacher, 2. physikal.
( 6 ) S. J. Cristol and J. S. Nelson, unpublished work. Chem., 29B, 356 (1935); Z.anorg. Chem., 226, 370 (1936).
FEBRUARY
1965 OF THE HUNSDIECKER
MODIFICATIONS REACTION 415

the crude product produced recovered alcohol, 198 g. of 1,1,7-tri- form containing 0.02 mole of titanium tetrachloride for 5 hr. a t
hydroperfluoroheptyl benzoate, b.p. 155"(0.5mm.), and 295 g. of 50'. After work-up similar to above experiments, fractionation
1,1,7-trihydroperfluoroheptyl orthebenzoate, b.p. 187" (0.1 of the product yielded 24.2 g. (80%) of the methyl fluoro alcohol
mm.), 62ycyield based on unrecovered alcohol. carbonate, b.p. 46" (0.1 mm.).
Anal. Calcd. for C28H14F3803: C, 31.05; H, 1.29. Found: The two methyl fluoro alcohol carbonates from these experi-
C, 31.21; H, 1.36. ments were found to be identical by infrared and n.m.r. analyses
l,l,S-TrihydroperfluoropentylCh1orocarbonate.-A solution of and determination of the same retention times by g.1.p.c.
69.6 g. (0.3mole) of 1,1,5-trihydroperfluoropentanoland 5.4 g. 1,1,7-Trihydroperfluoroheptylchlorocarbonate was prepared in
(0.3mole) of water was added intermittently over a period of 1.25 a similar manner except the alcohol and water were added por-
hr. to a mixture of 26.4 g. of ferric chloride in 60 ml. of carbon tionwise simultaneously. The chlorocarbonate was obtained
tetrachloride a t 76". Large amounts of phosgene and hydrogen in 40% yield, b.p. 73" (14mm.).
chloride were evolved over a period of 5 hr. After the addition of Anal. Calcd. for C8H3ClF1202: C, 27.70; H, 1.54. Found:
2.5 ml. more of water, the reaction was allowed to continue an C, 27.81; H, 1.44.
additional 8 hr. After cooling, the entire reaction mixture was Phosgene was bubbled through a mixture of 22.7g. (0.1mole) of
dissolved in ether, washed very quickly with cold water, dried, 1,1,5-trihydroperfluoropentanoland 8.1 g. (0.05 mole) of ferric
and distilled. The fractionation gave recovered alcohol, 40 g. chloride for 6 hr. a t 50". Less than 10% of chlorocarbonate was
(6097,)of the chlorocarbonate, b.p. 53" (13 mm.), and 15 g. obtained.
(217,) of the bis carbonate.
A part of the 1,1,5-trihydroperfluoropentyl chlorocarbonate Acknowledgment.-This paper is related to an obser-
(6.0 g., 0.02 mole) was added to excess methanol in ether a t 50". vation by one of us (M. E. H.) at the U. S. Naval
Distillation after work-up yielded 4.3 g. (70y0of methyl 1,1,5- Ordnance Laboratory, Silver Spring, Maryland. The
trihydroperfluoropentyl carbonate, b.p. 50' (0.15mm.).
Methyl chlorocarbonate (10.4 g., 0.11 mole) was stirred with many stimulating discussions with Dr. D. V. Sickman
23.2 g. (0.1 mole) of 1,1,5-trihydroperfluoropentanolin chloro- and Dr. 0. H. Johnson are gratefully acknowledged.

Modifications of the Hunsdiecker Reaction'"


JOHN A. DAVIS,JAMES , ~ ~CAR ROLL,'^ JIMB U N D SAND
H E R Y N KSAM JOHN SON^^
, ~ ~DOUGLAS
Department of Chemistry, Washburn University of Topeka, Topeka, Kansas
Received August 18, 196'4

Bromodecarboxylation of organic acids by the Cristol-Firth modification of the Hundsiecker reaction using
the free acid and red mercuric oxide was found to be much more effective than the modification using the acid
chloride and silver oxide. A study has been made of the possible use of oxides other than mercuric oxide and
of the effectiveness of various solvents.

Replacement of the carboxyl group in organic acids oyl chloride, p-nitrobenzoyl chloride, and p-chloro-
by a halogen atom, especially bromine, by treatment of benzoyl chloride.
the anhydrous silver salt of the acid with bromine, The silver oxide was carefully dried as was all the
chlorine, or iodine in an inert solvent is commonly glassware and the carbon tetrachloride, bromine, and
known as the Hunsdiecker r e a c t i ~ n . ~ -Yields
~ are acid chlorides were redistilled before use. Although
generally quite good (usually better than 60%), but numerous variations of the reaction conditions were
the reaction is extremely sensitive to trace amounts of tried, only trace amounts of the desire(. bromides
water, tlle presence of which leads to the recovery of were obtained by fractional distillation of the reaction
unchanged acid. Unfortunately, the preparation of mixtures. However, fair yields of bromobenzene were
dry silver salts of carboxylic acids is frequently very obtained from silver benzoate by the original Huns-
difficult and, because such salts are usually quite sensi- diecker reaction.
tive to heat also, they are &n quite hard to dry Fund, Cristol, and Firth have reported' that treat-
thoroughly. ment of a slurry of excess red mercuric oxide in a re-
Several years ago there appeared a report5 that the fluxing solution of stearic acid in carbon tetrachloride
acid chloride of penta-0-acetyl-D-gluconic acid under- with approximately an equivalent amount of bromine
goes smooth decarboxylation when heated with bro- in the dark gave a 93% yield of crude heptadecyl
mine and silver oxide in carbon tetrachloride, furnish- bromide. The reaction also was reported to give
ing aldehydo-1-bromo-D-arabinose penta-0-acetate in poor yields with benzoic acid arid glutaric acids, and
an excellent yield. A second report6 described the fair to excellent yields with cyclopropanecarboxylic
reactions of benzoyl chloride and three other aromatic acid, 9,lO-dihydro-9,lO-ethane-9-ant hronic acid, and
acid chlorides with bromine and silver oxide in carbon lauric acid. As soon as we learned of the Cristol-Firth
tetrachloride. modification, its apparent convenience and simplicity
An investigation of this modification of the Huns- in comparison to those procedures involving prepara-
diecker reaction was started using the following acid tion of silver salts or acid chlorides immediately sug-
chlorides: benzoyl chloride, n-butyryl chloride, phthal- gested a broad study of this modified procedure using
different classes of carboxylic acids and the use of
(1) (a) This work was made possible b y Grant R897-B from the Petroleum solvents other than carbon tetrachloride. Since mer-
Research Fund, administered by the American Chemical Society: (b)
American Chemical Society Petroleum Research Fund Scholar.
curic oxide is relatively expensive and mercury com-
(2) R. G. Johnson and R . K . Ingham, Chem. Rev., S6, 219 (1956). pounds are tokic, a study also was made of the possi-
(3) C. V. Wilson, Org. Reactions, 9 , 332 (1957). bility of using other oxides for this reaction.
(4) C. W. Shoppee and R . J. Stephenson, J . Chem. Soc., 2230 (1954).
(5) F. A. H . Rice, .I. A m . Chem. Soc., 7 8 , 3173 (1956).
(6) F. A. H. Rice and W. Morganroth, J . O w . Chem., 11, 1388 (1956). (7) Stanley J. Cristol and William C. Firth, Jr., i b i d . , 96, 280 (1961)
416 DAVIS,HERYNK,
CARROLL,
BUNDS,AND JOHNSON VOL.30
TABLEI
PERCENTYIELDSBASEDO N WEIQHTOF BROMIDE
RECOVERED
Acid Solvent Product Yield, 7
'0 B.p., OC." nDOC.a

Chloroacetic CzHzCli Bromochloromethane 38 68-73 (68-69)b 1.4795%(1. 4832)b


n-Butyric C2HzC1, 1-Bromopropane 53 70-74 (72 )" 1.43512' (1.4348y
Isobutyric CzH zC14 2-Bromopropane 21 59-62 (59) 1.425026( 1,4254)
2-Methylbutyric C2HzClr 2-Bromobutane 61 90-94 (91) 1 .430126(1.4366)
Hexanoic CCl4 1-Bromopentane 31 125-130 (128) 1.444031(1.4450)
Heptanoic cc4 1-Bromohexane 37 150-158 (159) 1.447028(1.4478)
Octanoic cc4 1-Bromoheptane 31 175-180 (177.5) 1.446231(1.4505)
Dodecanoic cc14 1-Bromoundecane 46 55-60 (114)d 1.456826(1.4569)
a Numbers in parentheses refer to literature values. Literature values of refractive index are given for 20". C. L. Stevens, T. K.
Mukherjee, and V. J. Traynelis, J. Am. Chem. Soe., 78, 2264 (1956). The literature values for 1-bromopropane and for all the com-
pounds below it in the table are taken from A. I. Vogel, J. Chem. SOC.,636 (1943). Value given for 0.05 mm.; liturature value
given for 5 mm.

TABLEI1 As the isolation by fractional distillation of a small


PERCENTYIELDSBASEDON AMOUNT
OF CARBON
DIOXIDE amount of liquid product from a relatively large volume
PRODUCED~ of solvent is a time-consuming operation and normally
Acid8 CChb CiHnClrb is not quantitative, a large number of determinations
Chloroacetic 71 (38) was made in which the extent of reaction was deter-
Trichloroacetic 82 (13) mined by measuring the amount of carbon dioxide
Propionic 94 evolved. In these determinations 0.01 mole of acid
n-Butyric 99 (53) was mixed with 5 g. (an excess) of red mercuric oxide
Isobutyric 90 (21)
77 (61)
(or an equivalent amount of other oxide) and 50 ml.
2-Methylbutyric
2-Ethylbutyric 80 87 (17) of solvent. The mixture was warmed and 2 ml. of
Trimethylacetic 85 (13) bromine was slowly added. The carbon dioxide was
Hexanoic 88 (31) swept out of the apparatus with nitrogen and absorbed
Cyclohexanecarboxylic 88 in a standard barium hydroxide solution. The barium
Heptanoic 90 (37) carbonate precipitate was removed by filtration and the
Octanoic 59 (31) excess barium hydroxide was titrated with standard
2-Ethylhexanoic 76 92 acid. See Table I1 for yields obtained when carbon
Decanoic 69 tetrachloride and 1,1,2,2-tetrachloroethanewere used
Lauric 67 (46) as solvents.
Myristic 91
Of the oxides that were tested as replacements for
Palmitic 90 81
Stearic 95 red mercuric oxide, only the five listed in Table I11
Phenylacetic 41 71 showed any reaction. Propionic acid dissolved in
Diphenylacetic 51 tetrachloroethane and treated with bromine in the
Benzoic 48 83 presence of an oxide of each of the following gave no
p-Toluic 40 76 reaction: aluminum, antimony, barium, bismuth,
Anisic 54 calcium, cobaltic, cupric, cuprous, ferric, magnesium,
p-Nitrobenzoic 0 70 nickelic, stannic, zinc, and sodium and titanium di-
Phthalic 85 oxides.
Terephthalic 20
Malonic 19 33
TABLE I11
Glutaric 85
Hexanedioic 53 PERCENTYIELDSBASEDO N AMOUNT
OF CARBON
DIOXIDE
Octanedioic 29 75 PRODUCEDO

Decanedioic 66 79 -Yield, %-
HgO
Using red mercuric oxide and either carbon tetrachloride or b-1-
lll,2,2-tetrachloroethaneaa solvent. Numbers in parentheses Acid low) AgnO PbO Pbso4 CdO
are per cent yields baaed upon the amount of organic bromide Propionic 87 78 42 45 8
recovered by fractional distillation. Palmitic 76 12 5 (CClr)
Acid chlorides
After trying a number of variations of the method, Benzoyl chloride 7 (CCla)
it was found that best yields were obtained when the Palmitoyl chloride 8 (cc4)
acid, the red mercuric oxide, and the solvent were Butyryl chloride 45
mixed and stirred with a magnetic stirrer. The mix- a Using 1,1,2,2-tetrachloroethaneaa solvent and oxides other
ture was then heated and a small fraction of the bro- than red mercuric oxide.
mine was added. If a reaction did not occur spon-
taneously, heating was continued until the reaction For the effectiveness of solvents other than carbon
started; then the remainder of the bromine was slowly tetrachloride and tetrachloroethane see Table IV.
added. The mixture was kept warm for 1 hr., then
filtered and washed with 5y0sodium hydroxide solution, Discussion
then with water, and dried over anhydrous magnesium
sulfate. It was then filtered and the bromide was In this investigation it was found that the Cristol-
separated by fractional distillation and identified by its Firth method using red mercuric oxide and the free acid
boiling point and refractive index (see Table I). gave much more decarboxylation than did the method
FEBRUARY
1965 p31-Hl COUPLING
FOR STRUCTURE
DETERMINATION 417

TABLEIV least 1 hr., then filtered, washed, and dried, and the organic
bromide was separated by fractional distillation. The prepara-
OF CARBON
PERCENTYIELDSBASEDON AMOUNT DIOXIDE
tion of 1-bromohexane is typical of the method used.
PRODUCED' Preparation of I-Bromohexane from Heptanoic Acid.-A few
Yield, $' & drops of a solution containing 16 g. (0.1 mole) of bromine in 50
Nitro- Bromo- Chloro-
ml. of dry carbon tetrachloride produced an immediate reaction
Ben- ben- ben- ben- CzClr- C&la- Hex-
Acid zene nene zene nene Fz F3 ane
when added to a warm, stirred mixture of 13.0 g. (0.1 mole) of
heptanoic acid and 22 g. of red mercurir oxide in 150 ml. of dry
Palmitic 69 82 76 67 71 58 40 carbon tetrachloride. The remainder of the bromine solution was
Phthalic 0 0 32 added slowly and the reaction mixture was refluxed for 1 hr.
Terephthalic 26 The mixture was then filtered, washed with 5y0 sodium hy-
Propionic 79 61 2 40 droxide solution and then with water, and dried with anhydrous
Benzoic 6 48 41 0 magnesium sulfate. Careful fractional distillation of the solution
Stearic 81 gave 6 g. of I-bromohexane, b.p. 150-159", 12% 1.4470.
2-Ethylbutyric 60 43 To determine the per cent yield based on the amount of
carbon dioxide produced, 2 ml. of bromine was slowly added to a
a Using red mercuric oxide and different solvents.
warm, stirred mixture containing 50 ml. of solvent, 0.01 mole of
acid, and 5 g. of red mercuric oxide in a 500-ml., three-neck
using silver oxide and the acid chloride. The Cristol- flask. The flask was fitted with a dry nitrogen inlet, separatory
funnel for addition of the bromine, and a reflux condenser. The
Firth method is simpler than the original Hunsdiecker carbon dioxide produced was swept out of the flask by nitrogen
method in that it does not require the preparation of the and passed through three Dry Ice traps to remove all bromine
dry silver salt. In many cases higher yields were ob- vapor and then absorbed in standard barium hydroxide solution.
tained with tetrachloroethane as solvent than with car- The barium carbonate precipitate was filtered off and the excess
barium hydroxide was titrated with standard acid. The follow-
bon tetrachloride. This probably was due to the higher ing is typical of the method used.
boiling point of the tetrachloroethane which permitted a Bromodecarboxylation of Palmitic Acid.-Bromine (2 ml.)
higher reaction temperature. Bromobenzene and tet- was added slowly to a warm, stirred mixture containing 50 ml.
rachlorodifluoroethane seem to offer soine promise as of carbon tetrachloride, 2.564 g. (0.01 mole) of palmitic acid,
other solvents for this reaction. Yellow mercuric oxide and 5 g. of red mercuric oxide. The carbon dioxide pro-
duced was swept out of the flask by nitrogen and absorbed in
was found to be almost as effective as red mercuric 125.0 ml. of standard 0.2 N barium hydroxide solution. The
oxide. As a general rule aromatic and dicarboxylic barium carbonate precipitate was filtered off and the excess
acids underwent less bromodecarboxylation than did barium hydroxide titrated with standard hydrochloric acid,
aliphatic monocarboxylic acids. yielding 0.0090 (goyo) mole of carbon dioxide.
Acknowledgment.-Acknowledgment is made to the
Experimental donors of The Petroleum Research Fund, administered
by the American Chemical Society, for support of this
For the preparation and isolation of the organic bromides, 0.1 research and to the E. 1. du Pant de Nelllours and
mole of bromine was slowly added to a warm, stirred mixture of
0.1 mole of acid and 0.1 mole of red mercuric oxide in 150 to 200 Company for furnishing samples of tetrachlorodifluoro-
ml. of ,solvent. The reaction mixture was then refluxed for a t ethane and trichlorotrifluoroethane.

Synthesis and Reactions of Some Triphenylphosphazines. The Use of


Long-Range P3I-H1 Coupling for Structure Determination
SINGHAND HANSZIMMER~
GURDIAL
Department of Chemistry, University of Cincinnati, Cincinnati d l , Ohio
Received July 17, 1964

Some 8-N-alkylidene- and -arylidenehydrazinotriphenylphosphoniumbromides have been synthetized by


reacting triphenylphosphine dibromide with appropriate hydrazones. Dehydrobromination of these phos-
phonium bromides with sodamide in liquid ammonia led to the corresponding triphenylphosphazines. These
triphenylphosphazines are nucleophiles as is shown by their reactions with methyl and ethyl iodides t o yield
a-N-alkyl-8-N-alkylidene- and -arylidenehydrazinotriphenylphosphoniumiodides. Thes tructures of these
phosphonium iodides have been verified through n.m.r. spectroscopy. The ultraviolet and visible absorp-
tion data of triphenylphosphazines are compared with those of the corresponding hydrazones.

Recently we became interested in the synthesis and drolysis of dialkylaminotriphenylphosphonium iodides


reactions of organophosphorus compounds containing and a-N-alkyl-p-N-disubstituted hydrazinotriphenyl-
a -N=P 6 linkage, such as triphenylphosphinalkyl- phosphonium iodides furnished the corresponding
imines2 and triphenylphosphin- and disubstituted secondary amines and 1,I ,Ztrisubstituted hydrazines,
aminoimines. 3 , 4 These compounds reacted readily respectively, in excellent yields. Thus, these com-
with methyl and ethyl iodides to give dialkylamino- pounds, containing a phosphorus-nitrogen double
triphenylphosphonium iodides2 and a-N-alkylhy- bond, represent a convenient tool for the synthesis of
drazino- and a-S-alkyl-P-N-disubstituted hydrazino- certain, especially sterically hindered, secondary amines
triphenylphosphonium iodides14 respectively. Hy- and 1,l,Btrisubstituted hydrazines.
(1) Author to whom inquiries should be addressed. The structures of all organophosphorus compounds,
(2) H. Ziinmerand G . Singh, J . O w . Chem.. 46,483 (1963); Anoew. Chem., mentioned above, were proved by elemental analyses,
11, 574 (1983): Angew. Chem., Intern. Ed. Eng., '2, 395 (1963).
(3) F. Kaplan, G . Singh, and H. Zimrner, J . Phys. Chem., 61, 2509 (1963).
chemical behavior, and ultraviolet and n.m.r. spectro-
(4) H. Ziminer and G. Singh, J . 0 7 0 . Chem., 28, 1579 (1964). scopic studies. The latter method proved to be of
Notes J. Org. Chem., Vol. 44, ~ V O19,
. 1979 3405
Table I. Titration of Organolithium Reagents Using N-Benzylidenebenzylamine ( 1 ) as Indicatora

titration conditions Watson and


EasthamC
organolithiumb solvent method A method B method total alkalid
n-BuLi in hexane
sample 3. Et,O 1.62e 1.60 1.76
THF 1.61e
sample 2 THF 1.6sesf 1.69
sec-BuLi in cyclohexane
sample 1 Et,O 1,50e 1.49 1.64
THF 1.50e
sample 2 THF 1 .15e8f 1.14
t-BuLi in pentane
sample 1 Et,O 1.26e 1.25 1.18
THF 1.24e
sample 2 THF 1.46e,f 1.46
PhLi in benzene-diethyl ether
Et,O l.Ole 1.00 1.96
THF 0.9ge
benzene 1.0oe
hexane l.Ooe
a Each value is an average of two or more titrations. The organolithium solutions were purchased from Aldrich. The
concentrations indicated by Aldrich were 1 . 6 M for n-BuLi, 1 . 4 M for sec-BuLi, 1.6 M for t-BuLi, 1.67 M for PhLi. 1,lO-
Phenanthroline as indicator, benzene or hexane as solvent, sec-butyl alcohol/xylene as acid solution, according t o ref 2.
Obtained by titration by a standard acid using phenolphthalein as indicator, after hydrolysis of a 5-mL aliquot of the or-
ganolithium solution by 10 mL of distilled water. e sec-Butyl alcohol/xylene as acid solution. f Benzoic acid/THF as acid
solution,

benzylidenebenzylamine (1) are known to produce a red of imine 1 (in excess) in 10 mL of solvent (see Table I). A strong
purple color? and this coloration remains intense as long crimson color appeared immediately with the addition of or-
as carbanion is present in the solution. We used this ganolithium; the solution was then titrated by a 1 M solution of
property for the analysis of organolithium reagents by two sec-butyl alcohol in xylene or a 1 M solution of benzoic acid in
methods (see Scheme I). tetrahydrofuran.
Method B. A 5-mL aliquot of the solution to be analyzed was
In method A, the organolithium reacts with a solution added at room temperature under nitrogen t,o a solution of 2 mL
of an excess of the Schiff base 1 by addition and meta- of diisopropylamine (in excess) in 10 mL of solvent (see Table
lation. The colored anion 3 is obtained by metalation of I). Imine 1 (2-3 drops) was added to this mixture, and the crimson
the Schiff base by means of RLi or the lithium amide 2 color appeared immediately. The solution was then titrated as
formed by addition. In method B, the organolithium is in method A.
converted to lithium diisopropylamide (4), which produces In the two methods, the end point was reached when the color
anion 3 from a few drops of Schiff base 1. The lithium of the solution became a persistent yellow.
compound RLi is quantitatively transformed into 2 and Registry No. 1, 780-25-6;n-BuLi, 109-72-8;sec-BuLi, 598-30-1;
3 (method A) or 4 and 3 (method B). Titration is made t-BuLi, 594-19-4;PhLi, 591-51-5.
by addition of an acid solution which reacts both with the
lithium amide 2 or 4, and with the colored anion 3. At the
end point, the added acid is equal to the initial quantity
of RLi. Results are summarized in Table I.
These methods have the advantage of using organo- Photoassisted Cristol-Firth-Hunsdiecker
lithium compounds in the common conditions of use of Reaction
these reagents in various solvents such as diethyl ether,
tetrahydrofuran, benzene, and hexane. These are single A. I. Meyers* and Michael P. Fleming‘
titrations, simple and rapid to realize in the laboratory,
and present a sharp and easy to observe end point. The Department of Chemistry, Colorado State University,
Fort Collins, Colorado 80523
Schiff base 1, easily prepared from benzylamine and
b e n ~ a l d e h y d e ,is~a, ~liquid at room temperature and can Received March 28, 1979
be stored as a solid at -30 OC. Finally, these methods can
be used to control the quality of the solvents.8 A well-known route to aryl and alkyl bromides is the
Hunsdiecker reaction2 or its more recent modification by
Experimental Section Cristol and Firth.3 The latter workers found that mercuric
Solvents were drit?d on molecular sieves, then distilled from salts of carboxylic acids could replace the more tediously
LiAIHl prior to use. prepared and sensitive silver salts in the key bromo-
Method A. A 5-mL aliquot of the solution to be analyzed was decarboxylation step upon treatment with bromine.
added a t room temperature under nitrogen to a solution of 2 g Recent studies on the scope and mechanism4* of the
Cristol-Firth modification indicate that the carboxylic
(6) (a) T. Kauffmanr,, K. Habersaat, and E. Koppelmann, Chem. Ber.,
110,638 (1977). (b) L. I. Sivova, N. A. Sivov, R. A. Gracheva, and V. M.
Potapov, Zh. Org. Khin., 14, 791 (1978). (1) National Service Award (NIH) Postdoctoral Fellow, IF32CA
(7) For a review see for example: R. W. Layer, Chem. Reu., 63, 489 05946-02.
(1963). (2) Wilson, C. V. Org. React. 1957, 9, 341.
(8) For example, titrations (method A) by an identical acid solution (3) Cristol, S. J.: Firth. W. C.. Jr. J. Ore. Chem. 1961. 26. 280
of an identical solution of n-BuLilhexane in THF freshly distilled from (4) Davis, J. A.; Hernyk, J.;Carroll, S.; Bunds, J.; Johnson, D. J . Org.
LAM4and THF distilled some months prior to the experimentationyielded Chem. 1965,30, 415.
N = 1.69 and N = 1.59, respectively. The difference between these two ( 5 ) Bunce, N . J. J . Org. Chem. 1972, 37, 664.
values is attributed to impurities contained in “old” THF. (6) Cason, J.; Walba, D. J . Org. Chem. 1972, 37, 669.

0022-326317911944-3405$01.00/0 0 1979 American Chemical Society


3406 J. Org. Chem., Vol. 44, No. 19, 1979 Notes
Table I. Conversion of RC0,H t o RBr (HaO-Br,-liahtP 579-75-9; o-phthalic acid, 88-99-3; 1-naphthoic acid, 86-55-5; n-
C13HnC02H, 544-63-8;n-C&1C02H, 57-10-3;n-C17H&02H, 57-11-4;
RBr p-C1CsH4Br,106-39-8;p-N02C6H4Br,586-78-7; 3,5-(N02)2CBH3Br,
% yield 18242-39-2; 3-N02-4-C1-5-MeOC6H2Br,63603-16-7; m-BrC6H4Br,
% yield (without 108-36-1;p-BrC6H8r,106-37-6;o-ClC6H4Br,694-80-4;m-ClC6H4Br,
RC0,H (with light)b light) 108-37-2;o-BrC6H4Br,583-53-9; n-C13H2,Br,765-09-3; n-C15HB1Br,
p-CIC,H,CO,II
p-NO,C, H,CO,H
80
95
Oe
42, 0:
- 70f
629-72-1; n-C17H35Br,3508-00-7.

3.5-(NO., LC,H,CO,H 97 Oe
3~NO,-4%l-~-MeOd6H,CO,H 9 6c -40
m-BrC,H,CO,H 83
p - Br C, H,CO, 13 86 Maximum Optical Rotation of 2-Fluorooctane?
o-CIC,H,CO,H 81 Survey of Fluorinating Reagents
m-ClC, H, CO, H 75
o-MeOC,H,CO,H d Jacques Leroy,la Eric Hebert,lb and Claude Wakselman'"*
o-phthalic acid 42 8 5f
1-naphthoic acid d C.N.R.S., 94320 Thiais, France
n-C, ,H,,CO,H 99 9lf
n-C, ,H,,CO,H 90 90f Received March 28, 1979
11-C,,H,,CO,H 85 95f
a Reactions run in carbon tetrachloride a t reflux, molar Optically active 2-alkyl halides are of particular interest
ratios of RC0,H-HgO-Br, are 1.00:1.50:1.50, respective- as they can serve as models in the study of numerous
ly. All yields are for pure (distilled o r crystallized) pro- reactions of organic chemistry. 2-Chloro-, 2-bromo-, or
ducts and were identified by comparison with published 2-iodooctaneis commonly used2 and their optical purities
IR and NMR data (Aldrich or Sadtler). J. M. Kane and
A. I. Meyers, Tetrahedron Lett., 771 (1977). Gave only are determined without ambig~ity.~ By contrast, however,
polybrominated products. e See ref 5. Based on car- the maximum optical rotation of 2-fluorooctane remains
bon dioxide produced. Actual isolated yields were much unknown, as no chemical correlation has been established,
lower in those instances where isolation was attempted owing to the poor leaving-group character of fluorine.
(see ref 4). Thus, only an estimate of optical purity has been made
for 2-fluorooctane (1) based on a correlation between
acids form a diacyloxy mercury salt 1, which then leads molecular rotation and bond refra~tion.~
to the acyl hypobromite 2 followed by thermal decom- Recent results obtained with methyltri-n-butylfluoro-
position to the alkyl bromide. During the course of other phosphorane5 (2) leading to 2-fluorooctane of higher optical
2RCOzH ---
HgO
(RCOZ),Hg
1
-ZBQ
2RC02Br 2RBr
2
- rotation than that of San Filippo and Romano4 (stated as
about 100% optical purity) have prompted us to check
other fluorinating reagents.6 Phenyltetrafluoro-
studies, we had occasion to investigate the CFH reaction phosphorane7 (3), N - (2-chloro-1,1,2-trifluoroethy1)di-
with aromatic acids. The observed trend@ is that aromatic ethylamine8 (4) (FAR), and (diethy1amino)sulfur tri-
acids form rather insoluble mercuric salts and give lower fluorideg ( 5 ) (DAST) were used according to slightly
yields than aliphatic acids. We now describe a further modified procedures.
modification of the CFH reaction which simply involves CGH,,CH(OTs)CH3 + (n-C4Hg)3PFCH, +

irradiating the reaction mixture with a 100-W bulb. In this 2


fashion excellent and reproducible yields of aryl and alkyl C6H13CHFCH3+ octenes + ( ~ Z - C , H ~ ) ~ P ( O T ~(1)
)CH~
bromides were obtained (Table I). For those aromatic
acids containing electron-rich substituents (e.g.,anisic acid)
the reaction gave only polybrominated products.
1
+ C6HbPF, 1 +
C6H,3CH[OSi(CH3)3JCH3
3
-
C5H,,CHFCH2CH3+ octenes + FSi(CH3)3+C6HSPOF2
Experimental Section
(2)
General Procedure for Light-Assisted Bromodecarb-
oxlyation. Into a dry, nitrogen-purged, three-neck, 100-mL flask CGH13CHOHCH3 + (C2Hh)ZNCF2CHFCl-+
containing a spin bar and fitted with a reflex condenser were 4
placed the carboxylic acid (10.0 mmol), carbon tetrachloride 1 + octenes + (C,H,),NCOCHFCl + H F (3)
(reagent grade, 50 mL), and red mercuric oxide (B & A, 15.0
mmol). The flask and an adjacent 100-W bulb were enclosed in Cp,H&HOHCH3 + (CzHS)zNSF3 -+

aluminum foil. The mixture was heated to reflux in an oil bath 5


while being irradiated, and bromine (Baker, "purified", 15.0 mmol) 1 + octenes + (CzH,),NSOF + HF (4)
was cautiously added via syringe. The mixture was heated a t
reflux and irradiated for 3 h, and cooled to room temperature.
Saturated aqueous NaHC03 (30 mL) was added and the mixture (1) (a) Centre d'Etudes et de Recherches de Chimie Organique
was vigorously stirred for 15 min. The two-phase mixture was Appliquge; (b) Groupe de Recherche No. 12.
vacuum filtered through a Celite pad and the pad was washed (2) M. Malissard, J. P. Mazaleyrat, and Z. Welwart, J . Am. Chem. SOC.,
99, 6933 (1977); E. Hebert, J. P. Mazaleyrat, and Z. Welvart, J . Chem.
several times with chloroform. The organic phase of the filtrate Soc., Chem. Commun., 877 (1977).
was washed with saturated NaHC03 (30 mL) and brine (30 mL) (3) H. M. R. Hoffman, J . Chem. SOC.,1249 (1964).
and dried (MgS04). Removal of solvent and distillation gave the (4) J. S. Filippo, Jr., and L. J. Romano, J. Org. Chem., 40, 1514 (1975).
bromide. (5) Preliminary communication: J. Bensoam, J. Leroy, F. Mathey, and
C. Wakselman, Tetrahedron Lett., 353 (1979).
Acknowledgment. Financial assistance for this study (6) G. Gelbard and co-workers have recently prepared (+)-(S)-2-
fluorooctane, [ a I m ~ 910.2' (neat), in 60% yield from the mesylate of
came from the National Institutes of Health (CA-16051). (-)-(R)-2-octanolby anion exchange resins. Private communication on
a further publication in J. Chem. SOC.,Perkin Trans. I.
Registry No.p-C1C6H,C02H,74-11-3;p-NO2C6H4CO2H, 62-23-7; ( 7 ) D.U. Robert, G.N. Flatau, A. Cambon, and J. G. Rim. Tetrahedron,
3,5-(N02)2C6H3C02H, 99-34-3; 3-N02-4-C1-5-MeOC6H2c02H, 29. 1877 (1973).
I ~~ ~~-- I

71001-78-0; m-BrC6H4Co2H,585-76-2; p-BrC6H,CO2H, 586-76-5; (8) C. M. Sharts and W. A. Sheppard, Org. React., 21, 125 (1974).
O-CIC&~CO~H, 118-91-2;m-ClC&C02H, 535-80-8; o-Me0C&I4CO2H, (9) W. J. Middleton, J . Org. Chem., 40, 5'74 (1975).

0022-326317911944-3406$01.00/0 0 1979 American Chemical Society

You might also like