Ciclobenceno
Ciclobenceno
Ciclobenceno
(7) N. E. Searle. Organic Syntheses, Collect. Vol. IV, Wiley, New York,
N.Y.. 1963.D 424.
(8) (a) M. Codiiera and A. M. Trozzolo, J. Amer. Chem. SOC.,92, 1772
(1970);(b) G.L. Closs and D. R. Paulson, /bid,,92, 7229 (1970).
(9)(a) W. H. Urry and J. R. Eiszner, J. Amer. Chem. SOC.,73, 2977 (1951);
74, 5822 (1952);(b) W. H. Urry. J. R. Eiszner, and J. W. Wilt, ibid., 79,
918 (1957);(C) W. H. Urry and N. Bilow. ibid., 86, 1815 (1964).
(10)(a) R. Kaptein, F. W. Verheus, and L. J. Oosterhoff, Chem. Commun.,
877 (1971);(b) R. A. Cooper, R. G. Lawler, and H. R. Ward, J. Amer.
Chem. SOC.,94, 552 (1972).
(11)W. von E. Doering and L. H. Knox, J. Amer. Chem. SOC., 76, 4947
(1956);83, 1989 (1961).
(12)(a) M. Cocivera and H. D. Roth, J. Amer. Chem. SOC.,92, 2573 (1970);
(b) H. D. Roth, Ind. Chim. Be&, 36, 1068 (1971);(c) H. Iwamura, Y. Imahashi, and K. Kushida, J. Amer. Chem. Soc.. 96, 921 (1974).
Abstract: Displacements on benzenediazonium cation, C,jHsNz+, in solution in the absence of strong bases, reducing agents,
or light proceed by rate-determining cleavage to a singlet phenyl cation intermediate, CsH5+, rather than by the bimolecular
mechanism proposed in 1969. In aqueous solution at 2 5 O , only unrearranged products are obtained, and negligible incorporation of deuterium from solvent D2O occurs, showing that aryne processes do not occur to a significant extent. The low selectivity between nucleophiles nevertheless requires a highly reactive intermediate. In the hydrolysis of C,jHsNz+, the high entropy of activation (+10.5 cal mol- deg-) and the constancy of k I (within 4%) from H l O to D20 demonstrate that water
is not involved as a nucleophile in the rate-determining step. The constancy of k 1 (within 10%) in solutions as diverse as
14-21 M H2S.04 and 100% CH3C02H and CH2Cl2 is in accord with a common rate-determining step over this whole range
of solvents. Para-substituent effects can be dissected into nearly equal field (predominantly transition state) and resonance
(predominantly reactant) contributions by an analysis using dual substituent constants (previously published), with a correlation coefficient of 0.992, although analysis using the Hammett p u equation, i.e., with only a single set of substituent constants, yields a correlation coefficient worse than zero.
Y+-J-(/
+ N,
\
I11
I1
I11
with benzenediazonium bromide, chloride, bisulfate, or fluoborate in water a t 5-64 is first order in diazonium salt.
However, when Y is a solvent molecule (H20), the kinetics
must be first order even with a bimolecular rate-determining step. When Y is a solute anion (Br- or Cl-), a bimolecular mechanism involving Y could still give deceptive firstorder kinetics because of compensating salt effects, as
shown by excellent first-order kinetics for bimolecular reactions of trimethyl- or tribenzylsulfonium ion with chloride
ion in 90% acetone-10% water and 16 other solvents and
solvent mixtures a t 18-100 .* Analogous bimolecular possibilities for reactions of I with anions have not been eliminated in any previously published work by recourse to kinetic studies a t constant ionic strength.
Possible Mechanisms. At the outset of this work, there
appeared to be eight possibilities for the rate-determining
step, where Y may be H20, Br-, C1-, or F-. Mechanism 1
is a one-step bimolecular displacement. This simplest mech-
+Y
II
C6H,Y+
0
lv
- 0+
I-
+Y
N,
\++N2
HY+
N,
(5)
- e*+
VI
VI
v+
N,
-
(7)
VI1
anism was rejected for a long time, because it appeared difficult to reconcile with the kinetics, e.g., with first-order
Swain, et al.
784
rate constants that vary less than 50% with hydrochloric
acid concentration, while the product ratio C6H5C1/
C ~ H S O Hchanges from 0.05 to 3.7c Nevertheless, it was
adopted in 1969 as the preferred mechanism for reaction of
arenediazonium ions with HzO, SCN-, Br- or CI- in aqueous solution in best accord with all the evidence in much excellent published ~ o r k . ~ gWve~too expected this mechanism
or one of its stepwise addition-elimination variants (mechanism 2 or 3)1 to be the one to survive further scrutiny when
we began our experimental work in 1962.
Mechanism 4, first proposed in 1942 by Waters,]' does
not involve Y in its transition state. This two-step mechanism proceeds through the reactive phenyl cation (IV) in
which the positive charge cannot be delocalized by conventional forms of resonance. The intermediate in mechanism 2
or 3 seems much more likely, because it avoids such concentrated charge development. Singlet phenyl cation is perhaps
the most implausible hydrocarbon carbonium ion that can
be imagined, save for various antiaromatic cations (e.g., cyclopentadienyl), since its electron deficiency is localized in
an orbital of relatively high s character (sp2). It is generally
better to give vacant orbitals maximum p character. Formally similar vinyl cation intermediates have recently been
demonstrated in several reactions of acetylenes and vinyl
halides and vinyl esters.I2 However, a b initio calculations12b indicate that the preferred structure of the vinyl
cation is linear a t the carbonium carbon, the bent sp2 structure being higher in energy by more than 30 kcal mol-'.
Experimentally,12d 1-cyclooctenyl and 2-cis-2-butenyl triflates (trifluoromethanesulfonates) solvolyze l o 5 times
faster than I-cyclopentenyl triflate and lo4 times faster
than 1-cyclohexenyl triflate in 50% ethanol a t IOO", presumably because they can give linear carbonium ions more
easily than I-cyclopentenyl and I-cyclohexenyl. I t would be
about as difficult for phenyl cation to be linear a t Ci as for
cyclohexenyl. A phenyl group is also more electronegative
than a methyl group and unlikely to be transformed into a
cation on that account also. Therefore one should be exceedingly skeptical about any interpretation requiring IV.
A variant of mechanism 4 involves not singlet I V but instead a triplet (biradical) phenyl cation, which might provide resonance stabilization and delocalization of the positive charge around the ring and even onto meta substituents, as was thought to be required to explain the peculiar
effects of substituents on rate of dediazoniation of metaand para-substituted benzenediazonium S a l k i 3
Mechanisms 5-8 all involve a benzyne intermediate.
Other possibilities seem to be only minor variants of these
eight.I4
Products. Benzyne mechanisms 5-8 are all excluded for
decomposition of I in aqueous solutions without strong
bases by the absence of isomerized products such as a r e
formed from reactions known to proceed via a benzyne intermediate under the same conditions. For example, decomposition of 0-toluenediazonium chloride in water yields O cresol but no m- cresol; likewise m - toluenediazonium ion
yields m- cresol but no 0-cresol, although as little as 0.1%
could be detected (see Experimental Section). N o rearrangements during decomposition of simple arenediazonium ions in aqueous solutions have been reported, although
the reaction has been used widely for a ~ e n t u r y . ~ ~ , ' ~
For 1 itself, the absence of benzyne mechanisms is demonstrated by negligible formation of ring-deuterated phenol
during the decomposition of I BF4- in DCI-D20. As shown
in Table I, the amount of phen-dl-ol formed is only 0.05%
during solvolysis for 16 hr a t 25'. A t the acidities employed, phenol and phenoxide both undergo significant
acid-catalyzed exchange.I9 O n the basis of control experiJournal of the American Chemical Society
Reactant
I BF4-*
I BF4-b
C6HjOH
CsHsOH
H+, M
Time,
hr
mle 95/(m/e 95 f
--m/e
94)-Dz0
Hz0
X D
0.01
0.01
0.0087
0,0434
16
16
44.5
44.0
0.0635
0.0646
0.0640
0.0642
0.06
0.04
0.04
0.06
0.0629
0,0642
0.0636
0.0636
25
25
25
50
50
50
13.4
11.9
9.7
12.0
10.9
10.1
0.85
0.88
0.89
0.92
0.82
0.76
785
Table 111. First-Order Rate Constants for Dediazoniation
0.0003-0.0015 M I BF,- at 25.0"
Run
Solvent
no.2
25
24
66
20
15
62
22
121
60
59
67
40
43
42
73
122
71
72
100% CH,CO*H
+ 0 . 1 3 MLiCl
+ 1 . 0 MLiCl
100% CH3C02H
100% CH3C02H
CHiCIz
3-Methylsulfolane
97 Z sulfolane-3 (C2H&0
Dioxane
105kl, sec-1
4.59 i 0 . 0 3
4.55 i 0 . 0 3
4 . 1 2 i 0.01
3.56 i 0.02
2.68 i 0 . 0 1
2.13 i 0.02
2.11 i.0.01
2.15 i 0.01
4.76 f 0.04
4.71 f 0.04
4.11 i 0.07
2.26 i0.01
3.71 i 0.07
4 . 5 1 =t0 . 0 2
2.20 i 0.05
1 . 3 6 -I 0 . 0 3
1 . 2 4 i 0.01
1 . 1 5 i. 0.02
(9)
+ 3H20
* 3H7O'
K
12DZO
786
for halfway breakage of the C-N bond (1.02, basea on a
calculated N isotope effect of 1.04-1.045 for complete
breakage of a C-N bond). If mechanism 2 or 3 were correct, the transition state should be close to the unstable intermediate 11, and the a-N isotope effect should be close to
unity. In the second paper following this
we present
data that show an a-N isotope effect of 1.038. This large kinetic isotope effect is inconsistent with mechanisms 1-3 but
in accord with mechanism 4 since its transition state should
resemble the high-energy phenyl cation IV more closely
than I, and hence the C-N bond should be extensively broken a t its transition state.
The large kinetic hydrogen secondary isotope effect
k H / k D of 1.22 for each ortho hydrogen (see following
paper4) indicates an extremely electron-deficient transition
state, as one close to the carbonium ion (IV, c6H5+) in
mechanism 4 would be, but one close to the resonance-stabilized diazonium ion (I, structure 9) in mechanism 1 [or to
the anionic intermediate (11) in mechanism 2 or 31 would
not. Hyperconjugation is expected to become this important
only when better possibilities for stabilization by resonance,
solvation, or other covalent participation by leaving, entering, or neighboring groups a r e all absent.
Other Solvent Effects. Products vary widely in different
In
solvents, but rates vary remarkably little (Table
glacial acetic acid with 0-1 M LiCl added,
M I BF4yields 2 6 9 % phenyl acetate, 2-3% phenol, 29-0% fluorobenzene, and 0-27% chlorobenzene; in methanol, the
products are 9 1-93% anisole, 5 4 % fluorobenzene; in methylene chloride, they are 66% chlorobenzene and 34% fluorobenzene; in 3-methylsulfolane, they are 54% phenol and
17% f l u o r o b e n ~ e n e . ~ ~
The similarity of rate constants shown in Table I11 for
three solvents of widely different polarity, CH2C12 (2.2),
CH3COrH (2.3), and fuming H2S04 (2.2), is striking. I t
seems to us to imply that the rate-determining step does not
change, and that these solvents solvate I* and I only very
feebly. The twofold decrease in rate constant in the sulfolanes and dioxane is puzzling, but it may result from tighter
association into large ionic aggregates approaching the stability of solid I BF4- in these solvents of especially low
anion-solvating ability.
The small (twofold) increase in rate from 14 M (80%)
H 2 S 0 4 to
M H2S04 may a t first appear to suggest a
change or partial change to a new mechanism in
M
H2S04 in which water is involved as a nucleophile, as in
mechanism 1 , 2, or 3. However, this deduction is unjustified
because water activity or vapor pressure increases 30-fold
over this range, and this is a region already noted for large
and highly specific changes in ionic activity coefficients
(and consequent inconstancy of rate constants for ionic
reactions), in contrast to the relative constancy of activity
coefficients in 83-100% H2S04.26This small change (doubling) of rate is in fact completely reasonable for mechanism 4, as a consequence of differing effects of water solvation on I and I*. The activity coefficient of I should be lowered less than that of I* by increasing water activity, even
though there may be the same number of water molecules
in the innermost shells of I and I* a t any given concentration of HzS04. Because the charge in I is better delocalized
by resonance (eq 9), it benefits less from replacement of
neighboring H2S04 molecules by H20 molecules. In short,
no water is involved in the rate-determining step from 21 M
H2SO4 down to 14 M H2SO4. It is weakly involved in
M H2S04 but only to an extent that is reasonably attributed to a more reduced activity coefficient for I*, owing to
differentially stronger solvation of I* a t the highest water
c ~ n c e n t r a t i o nI . ~
A similar (twofold) increase in rate occurs in 100% acetic
Journal of the American Chemical Society
97.4
>*,(Pi
+ i,
(- 3.18
and
log ( / ~ / k=~j , S
) ~+
J*$
+ i,
(5.08
= (- 2.74
* 0.36)R
= (- 2.60
-C
* 0.20)3 +
(0.27 * 0.12)
* 0.1613 +
* 0.10)
+ 0.37m + (-0.25
using 5 and (R (field and resonance) substituent constants defined and tabulated p r e v i o u ~ l y , without
~~
any
change. These J r, and i values were fitted by leastsquares, and the f numbers are their standard deviation^.^^
The f m and f P reaction constants represent the sensitivities of these reactions to the field constants 5 of meta and
para substituents (representing all influences except those
transmitted by resonance or x bonds). They are comparable
(-2.74 and -2.60) and of the expected sign and magnitude,
because electron-supplying (negative 5) substituents should
facilitate departure of N2 with its previously bonding electron pair, by stabilizing the transition state I* (which is
close to the electron-deficient phenyl cation IV in structure)
much more strongly than the reactant I.
The r ,,,and r P reaction constants represent the sensitivities of these reactions to the resonance constants CK of meta
and para substituents. For the ten meta substituents, resonance with electron-supplying (negative (R) substituents significantly stabilizes the transition state leading to phenyl
cation by supplying negative charge to the ortho carbons.
Since meta substituents are poor at stabilizing the reactant,
they do not increase the double bond character of its C-N
bond. Therefore the sign of this sensitivity to resonance r ,,,
due to meta substituents (-3.18) is the same as that off,,,
and f p (nonresonance) sensitivities to both meta and para
substituents. The % (R (average relative importance of resonance) is 42.0%, which may be compared with 0% for d,
22% for urn,53% for up, and 66% for ap+.38
For para substituents, however, resonance with electronsupplying substituents does increase the double-bond character of the C-N bond in the initial reactant diazonium cat-
787
Table IV.
Substituent
OCHa
c1
Br
Nos
n
---
--Substituent constant3
a?
0.000
-0.052
-0.065
-0.104
0.139
0,487
0.413
0.690
0.727
1.109
O.OO0
O.OO0
-0.141
-0.114
-0.138
0.088
-0.643
-0.500
-0.161
-0.176
-0.155
0.650
0.930
1.250
0.398
1.088
0.708
-1.465
-1.205
-3.063
0.272
0.864
0.814
0.997
0.171
0.981
0.729
-1.109
-1.163
-3.265
0.0o0
-0.930
-0.251
-0.832
-0.710
-1.252
-0.682
-1,060
-3.842
-2.773
-3.866
-2.865
-2.401
-2.351
0-
I-I
OH
C(CH3)3
CH3
C6H5
0 OCH,
-4
-2
-I
Calcd log ( k / k , )
-3
Figure 1. Dual substituent constant cf5 r(R i ) plot of para-substituent effects on rate of dediazoniation of I CI- in 0. I M HCI at 25'.
-0.5
Figure 2. Hamrnett ( p a ) plot of para-substituent effects on rate of
dediazoniation of I CI- in 0.1 M HCI at 2 5 O .
Experimental Section
I (Benzenediazonium) BF4- was prepared by diazotization of aniline in HBF4 solution in an ice-salt bath,41 recrystallizing twice
from saturated 25' solutions in 5% HBF4 by chilling to O o , and
twice more by dissolving in (CH3)2CO at 25O, adding CHC13 u n t i l
a few crystals appeared, then chilling to - 2 O O . After being dried at
25O ( I mm) for 2 hr, it could be stored in a vacuum desiccator at
Oo in the dark for several months with no sign of decomposition: ir
urnax(2% in KBr) 3100-3000, 2298, 1570, 1470, 1310, 1200-1000
Swain, et al.
7 88
(BFd-), 770, 760, 665, 535, 523, 512, and 460 cm-I; uv A,
(HzO) 260 nm ( t 12,000) and 295 (1880); nmr complex multiplet
~ T V S(20% in CH3CN) 7.2-8.8; decomposes slowly above 100'
with evolution of BF3 and melts into boiling CsHsF at 119-121
(lit.42 121- 1 22').
Anal. Calcd for C6HsN2BF4: C , 37.55; H , 2.63; N, 14.60.
Found: C, 37.70; H, 2.97; N, 14.68.
Solvents. Distilled water was passed through a 2-ft Pyrex column filled with Fisher analytical grade Amberlite MB-1 resin (a
mixture of sulfonic acid and quarternary ammonium ion resin) and
sealed with a Teflon stopcock. Before use each day, at least 1 I. of
water was flushed through the column. The conductivity of the
water, measured in a Pyrex conductivity cell with Pt electrodes by
a General Radio 1650A impedance bridge, was 2.5 X IO-' mho
cm-I. This deionized H2O was used to rinse all reaction vessels
and as the solvent for all kinetic runs and product studies in aqueous or partially aqueous solutions.
HzS04 used for kinetic runs and syntheses was analytical reagent 96%, diluted with the deionized HzO. Fuming H2S04
(Baker and Adamson reagent) showed a uv A,,
279-283 nm (ab(HzO) 276 nm ( t 525).43
sorbance A , 0.8), attributed to SO2, A,,
Fuming H ~ S O A100
, ml, and 30% HzOz (1 ml of Baker reagent
grade) were refluxed 4 days in an all-glass apparatus sealed with a
tube containing Drierite (CaS04). Destruction of SO2 and Hz02
was quantitative: a 1-ml portion of the acid failed to decolorize 100
ml of
M KMn04. The A of the purified acid was 0.030 a t
280 nm vs. 96% H2S04, and its concentration, by titration with
standard N a O H , was 105.4 f 0.2% H2S04 (28% SO3 by weight).
DzS04 was prepared by Jones44 by reaction of SO3 (Baker and
Adamson reagent) with 99% DzO. Titration with standard N a O H
corresponded to 97.0% D2S04.
Acetic acid (reagent glacial) was titrated with standard N a O H ,
indicating 99.7 f 0.3% HOAc. This material was used for most
purposes without further purification. The CH3C02H used in
some of the kinetic runs and for the product studies using LiCl was
refluxed for 10 hr over CrOi and distilled through a 30-cm Vigreux column. The middle 80% was redistilled from triacetylborate,45 bp l17', mp 16.2'.
Cyclohexane (Baker reagent, 2.5 I.) was stirred with four 400ml portions of 96% H 2 S 0 4 for a t least 2 hr each. The last two portions remained colorless. The cyclohexane was washed with 1 I. of
HzO, I 1. of 0.1 M NaOH. and three 1-1. portions of H 2 0 , dried
overnight over Drierite, and distilled through a 30-cm Vigreux column, The middle 80%, bp 80.3', n DzS 1.4232 was used. Methanol
(Fisher reagent) was used without further purification.
Methylene chloride (Fisher reagent) was purified by the same
procedure as cyclohexane, except that after drying over Drierite, it
was dried over Fisher reagent grade molecular sieves (3-mm pellets
of a sodium calcium aluminum silicate with a pore size of 4 %.)for
at least 30 min, then distilled through a 60-cm column packed with
glass helices. The middle 80% of the distillate, bp 40.3', nD2'
1.4209, was stored in the dark and used for kinetic runs within a
week.
Sulfolane (tetrahydrothiophene 1, I -dioxide, Shell Development)
was distilled through a 60-cm column packed with 3-mm Pyrex
beads. The middle 80% was redistilled twice without the fractionating column. Redistilled sulfolane was colorless and odorless, bp
90' ( 1 mm), n D2s I .4825. Analysis by glc on a 6-ft 20% Apiezon L
on Chromosorb W column at 25-300' and 14 psi of H e showed
0.1% volatile impurity. Since sulfolane solidifies at 28O, 3.3% Et20
(Mallinckrodt reagent) was added to depress the melting point
below 25'. 3-Methylsulfolane was purified by the procedure of
Alder and Whiting,46 with the exception that C a O was used instead of PlOs as the drying agent. The purified material, bp 83'
(0.2 mm), when analyzed by glc on the Apiezon L column. showed
less than 0.1% volatile impurities. The A of a 1-cm cell of 3-methylsulfolane vs. air was 0.124 at 315 nm, 0.155 at 295 nm, 0.233 at
270 nm and increased gradually a t lower wavelengths.
Determination of Yields of Halobenzenes from Dediazoniation of
I BFd-. Three 80-mg samples of I BF4- were dissolved in 10 ml of
HBr or HCI solution in 25-1111 volumetric flasks. Cyclohexane (10
mi) was added, and the flasks were shaken, sealed, and placed in a
constant-temperature bath at 25 or 50' for at least 3 days a t 25'
or 1-2 hr at 50' ( I O half-lives). The acid layer was separated and
extracted with two more 5-ml portions of cyclohexane. The combined cyclohexane solutions were extracted thrice with 0.1 M
N a O H and once with 1 M H2S04, then diluted to 25 ml. The absorbance A of the C6HsBr solutions was measured a t 265.6 and
261.5 nm. The yields were calculated from the extinction coefficients of C6H5Br [265.6 nm ( t 211), 261.5 (182)] and C ~ H S F
[265.6 nm ( t 613), 261.5 (SOO)]. The yield of CsHsCI was calculated from A at 264.6 nm. Since A was measured a t only one
wavelength, the yield of CsHsF could not be calculated but was assumed to be 1%, in agreement with the ChHsBr solutions. The
yields of C&isF'. determined from A a t 266 nm ( 6 1070), for decomposition of 0.03-0.05 M salt in l .O and 7.0 M H ~ S O with
J
no
added halide were 1.2-1.4 and 0.8-1.096, respectively. When solutions were extracted with cyclohexane immediately after dissolution, the yield was 0.2-0.5%. Part of the ChHsF therefore arises
from decomposition of the salt in the solid phase.
Products in Methanol. Solutions containing 0.016-0.028 M salt
in C H 3 0 H were decomposed at 25' in the dark for 42 hr (20 halflives). The solutions showed absorption a t 260-400 nm equivalent
to formation of 0.01-0.02% of p - H O C & ~ N ~ C ~ HThe
S . solutions
were diluted 25-50-fold and A measured at 277.9, 266.5, 260.5,
and 254.5 nm. The t values for ChH50CH3 (1578, 1340, 883,480)
and CsHsF (2, 800, 900, 600) were measured at these wavelengths. The yields of ChHsOCH3, 91.5 f 0.5%, and C6H5F, 5.5 f
1%. were calculated from A at 277.9 and 266.5 nm. The values
calculated for A at 260.5 and 254.5 nm agreed with those observed
within 1%. The rate constant was 9.80 f 0.04 X
sec-' with
6.8 X
M I BF4- a t 25'.
Products in Acetic Acid. Solutions containing approximately
M salt and 0, 0.0100, 0.100, and 1 .OO M LiCl were decomposed a t 25' for 4 days (10 half-lives). The yields of C6H6,
ChH5C1, C ~ H S FCsHsOH,
,
and ChHsOAc were calculated from A
of the solutions and the measured t values at 280, 271, 265.8,
263.5, 260.5, and 259.5 nm. As a check, 76 mg (4 X I O T 4 mol) of
salt was dissolved in 10 ml of glacial HOAc containing 1.00 M
LiC1. After 4 days at 25'. the solution was diluted to 100 ml and
shaken with three 10-ml portions of cyclohexane. The cyclohexane
solution was analyzed for C6HsCl as previously described. The
yield of CsHsCl was 29 vs. 27% calculated directly from A of the
reaction mixture.
Products in Methylene Chloride. When 11.6 mg of salt was
shaken with 100 ml of CH2C12 and allowed to stand in a dry box at
25', the salt dissolved gradually over 2 days as the decomposition
progressed. After 7 days ( I O half-lives after complete solution), 90
ml of the solution was concentrated to about 1 ml by fractional distillation on a 60-cm Nester-Faust Pt spinning-band column at a reflux ratio of 7.1. The concentrate showed glc peaks for C6HjCI
and ChHsF in a ratio 3/2. The remainder of the original solution
was used for spectrophotometric measurements. The products of
decomposition of saturated solutions in CHzCl2 were also analyzed
spectrophotometrically. The yields of C6Hs (2 f 2%). C6HsF (33
f 3%), and C6HsCI (65 i 3%) were calculated from the A and t
at 270, 266.5, 264.5, 264, 262.5, 261, 260, 258, and 255 nm. Presumably CH2CIF was formed equivalent to the C ~ H S C Ibut
, it was
not isolated because of its volatility.
Products in Acetone. When 0.493 g, in 25 ml reagent (CH3)2C O was decomposed a t 25' for 4 days, the resulting solution was
deep purple, with an unpleasant odor. Analysis of the solution by
glc on a 6-ft 20% Apiezon L on Chromosorb W S-60-80 column at
130' and 20 psi of H e showed 13 f 2% ChHsF and 100 f 15%
C6H50H. Estimation of the area of the C6H50H peak was less accurate because the peak tailed badly. 4-Hydroxy-4-methyl-2-pentanone (2.73 mol per mol of salt) and other higher molecular
weight condensation products were also formed.
Products in 3-Methylsulfolane. I (0.194 g, 1.01 X
mol) as
decomposed in 10 ml of 3-methylsulfolane at 25.00 & 0.02' for 1 I
days (16 half-lives). The solution analyzed by glc showed four
(17% yield); unknown, 0.24: unknown, 0.24;
peaks, C ~ H S F0.31,
,
unknown, 0.16; CsHsOH, 1.0 (57% yield). When 90% of the solvent was distilled at 95' (1 mm), no phenol appeared in the distillate and the residue showed the same peaks by glc. The uv of the
266 nm ( 6 1.78), 260 (258), and 255 (251) but no
solution had A,
peaks for phenol at 280 nm ( t 1500) or 273 (1800). The product
was postulated to be a sulfoxonium salt which decomposed to yield
phenol during glc analysis. Attempts to isolate the salt were unsuccessful. Chalkley, et a1.,4' have subsequently confirmed the existence of the sulfoxonium salt and prepared samples of it by decomposition of I in sulfolane a t 70-80".
789
Table V. Yields of Brornobenzene and p-Dibromobenzene from
Dediazoniation of 0.1 M I BF4- in 75% HOAc Containing Brn and
0.25 M H 2 S 0 4at 25'
0.44
1.1
2.2
3.0
4.0
7.7
12.2
12.1
0.8
0
0
0.1
5.9
21.5
28.0
7.7
12.2
18.0
22.3
28.0
Swain, et al.
790
E. Sheats. and K. G. Harbison, J. Amer. Chem. SOC.,97, 796, (1975).
(6)We have concentrated our study on the mechanism of the homoge-
-.
-+ +
+
97:4