Isomerization of substituted biphenyls by superacid. A remarkable confluence of experiment and theory

Article abstract of DOI:10.1021/jo0106445

Acid-catalyzed isomerization of dimethylbiphenyls is determined by the relative stability of intermediate carbocations, rather than the neutral products, and may be predicted by a simple semiempirical method (AM1). A general kinetic model for such isomeri

Full text of DOI:10.1021/jo0106445

2
034
J . Org. Chem. 2002, 67, 2034-2041
Isom er iza tion of Su bstitu ted Bip h en yls by Su p er a cid . A
Rem a r k a ble Con flu en ce of Exp er im en t a n d Th eor y
†
†
,†
‡
S. Christopher Sherman, Alexei V. Iretskii, Mark G. White,* Charles Gumienny,
,
‡
§
Laren M. Tolbert,* and David A. Schiraldi
School of Chemical Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0100,
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332-0400,
and KoSa, Spartanburg, South Carolina 29304
laren.tolbert@chemistry.gatech.edu
Received J une 25, 2001
Acid-catalyzed isomerization of dimethylbiphenyls is determined by the relative stability of
intermediate carbocations, rather than the neutral products, and may be predicted by a simple
semiempirical method (AM1). A general kinetic model for such isomerizations is suggested in which
the rearrangement of an intermediate cation is the rate-limiting step. Control of regiochemistry of
dialkylbiphenyls provides a useful entry into high-purity monomers for high-polymer synthesis.
In tr od u ction a n d Ba ck gr ou n d
Our interest in facile synthesis of disubstituted bi-
strategy of (1) oxidative dimerization of toluene to di-
methylbiphenyl and (2) oxidation of dimethylbiphenyl
(dmbp) to bibenzoic acid (BBA) or oxidation/esterification
phenyls was prompted by the use of the biphenyl moiety
as a well-known building block for high-temperature,
to dimethyl bibenzoate (DMBBA) was undertaken as a
potentially viable method for producing the desired
monomer(s) from low cost, readily available starting
materials. Unfortunately, all possible isomers were found
when toluene was coupled by conventional technology,
and the predominant species [2,X′- (X ) 2, 3, 4) and 3,3′-
1
high-modulus polymers as well as liquid crystals. While
the 4,4′-bibenzoic acid monomer has been demonstrated
2
-7
to produce thermotropic liquid crystalline polyesters,
8
copolyesteramides, and high-performance copolyester-
ether elastomers,9
-10
its commercial use has been hin-
13
dmbp] were of little commercial utility. Thus, we
dered by the absence of a rational, moderate cost syn-
thetic route to polymer purity monomer. Other isomers,
e.g., 3,4′-substituted biphenyls, for such applications have
concentrated our efforts on changing the distribution of
isomers after the coupling reaction.
14
Pd-catalyzed oxidative coupling of toluene in the
1
1
only been briefly described, but by analogy to the meta-
substituted arylenes12 could be expected to produce
materials of significant interest. Thus, the synthetic
presence of HClO
isomers. In the presence of trifluoromethanesulfonic acid
triflic acid, HOTf), the distribution of the dmbps was rich
4
favored the 3,4′- and 4,4′-dmbp
(
1
5
in 3,4′-dmbp. Variation of electronic and steric factors
ligand environment on Pd) did not improve the regio-
*
To whom correspondence should be addressed. E-mail: (M.G.W.)
(
mark.white@che.gatech.edu.
†
selectivity of the reaction, but a strong acid, such as
HOTf, immediately showed some unusual formation of
School of Chemical Engineering, Georgia Institute of Technology.
School of Chemistry and Biochemistry, Georgia Institute of
KoSa.
‡
§
3
,4′-dmbp in large excess (up to 56%). Since the most
Technology.
stable isomer among the dmbp isomers is the 4,4′ isomer,
we speculated that this augmentation of the relative
amount of 3,4′-isomer in the reaction mixture was the
result of acid-catalyzed isomerization of dmbps in which
the major product was controlled not by the stability of
the hydrocarbon, but by the most stable intermediate
cation.
(
1) Polyesters, Thermoplastic. In Kirk-Othmer Encyclopaedia of
Chemical Technology, 4th ed; Wiley-Interscience: New York, 1996; Vol.
1
9, pp 641-7.
(
(
2) Lee, D.-K.; Tsai, H. B. J . Appl. Polym. Sci. 1997, 64, 893-900.
3) J ackson, W. J .; Morris, J . C. Liquid Crystalline Polymers; ACS
Symposium Series 435; American Chemical Society: Washington, DC,
990; pp 16-32.
4) Morris, J . C.; J ackson, W. J ., J r. U.S. 4,959,450 (to Eastman
Kodak Co.), 1990.
5) Morris, J . C.; J ackson, W. J ., J r. U.S. 5,011,878 (to Eastman
Kodak Co.), 1990.
6) Morris, J . C.; J ackson, W. J ., J r. U.S. 5,037,946 (to Eastman
Kodak Co.), 1990.
7) Perez, E.; Benevente, R.; Bello, A.; Perena, J . M.; Van der Hart,
D. L. Macromolecules 1995, 28, 6211-19.
8) Gupta, B.; Halke, W.; Shepherd, J . P.; Sawyer, L. C. World
Patent 95/17446 (to Hoechst Celanese Corp.).
9) Lee, H.-B.; Liu, J .-L.; Tsai, Y.-S. J . Appl. Polym. Sci. 1996, 59,
1
(
(
Olah and Molnar discussed the use of superacids in
23
hydrocarbon isomerizations. They commented that
(
superacids permitted alkane isomerizations at lower
temperatures and thus the products showed much skel-
(
(
(13) Iataaki, H.; Yoshimoto, H. J . Org. Chem. 1973, 38, 76. Yoshi-
moto, H.; Iataaki, H. Bull. Chem. Soc. J pn. 1973, 46, 2490. US
Patents: 4,294,976, 1981; 3,895,055, 1975.
(14) van Helden, R.; Verberg, G. Recl. Trav. Chemi. Pays-Bas 1965,
84, 1263. Davidson, J . M.; Triggs, C. Chem. Ind. (London) 1966, 457;
J . Chem. Soc. A 1968, 1324. Ferrers, R. S. C.; Norman, R. O. C.;
Thomas, C. B.; Wilson, J . S. J . Chem. Soc., Perkin Trans. 1 1974, 1289-
94.
(
1
027-31.
(10) Lee, H.-B.; Lee, D.-K.; Liu, J .-L.; You, J .-W. J . Polym. Res. 1995,
2
, 109-14.
(11) Mang, M. N.; Brewbaker, J . L. U.S. 5,138,022 (to Dow Chemical
Co.).
(12) Polyamides, Aromatic. In Encyclopaedia of Polymer Science and
Engineering, 2nd ed; Wiley-Interscience: New York, 1988; Vol. 11, pp
(15) Xu, B. Q.; Sood, D.; Iretskii, A. V.; White, M. G. J . Catal. 1999,
187, 358-66.
3
81-409.
1
0.1021/jo0106445 CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/08/2002

Isomerization of Substituted Biphenyls by Superacid
J . Org. Chem., Vol. 67, No. 7, 2002 2035
etal branching. The product distribution was dictated by
the thermodynamic equilibrium of the carbocations when
excess acid was used and when the system was quenched.
In this chapter, the authors described a mechanism for
isomerization that uses a carbocation intermediate.
Reaction mechanisms are described for isomerization of
cycloalkanes, side chain isomerization in arylalkanes and
positional isomerization of substituted arenes. McCaulay
and Lien16 showed that the equilibrium distribution of
xylenes could be varied from (15%/64%/22%: o-/m-/p-
xylene) to exclusively m-xylene at 80 °C by introducing
solvents (benzene, toluene, octane, o-, m-, p-xylene, cumene,
p-cymene, and mesitylene), which were used without further
purification. Other dimethylbiphenyls were prepared by in-
dependent synthesis in small quantities for the purpose of their
identification by GC/MS according to published methods.
P r oced u r es: Ch em ica l Con ver sion s. All reactions were
carried out under Ar using standard Schlenk techniques. Care
was exercised in transferring the triflic acid to minimize
exposure to the atmosphere. Typically, 5-10 mL of triflic acid
was taken with the required volume of substrate in a 25 mL
glass vial affixed with a rubber stopper. The contents were
stirred so that the observed concentration versus time data
were not a function of the stirring speed, thus ensuring that
the kinetic reaction rates were controlling the overall reaction
process. An inert diluent, e.g., hexane, was added so that the
substrate or acid amounts could be changed individually. The
reactors were immersed in the appropriate bath to control the
reaction temperature. Small aliquots of the reaction mixture
were withdrawn at frequent intervals (15-30 min) to establish
the reaction kinetics. Samples were introduced into ice-water
to quench the reaction.
The effect of acid/substrate ratio was demonstrated for the
reaction of 4,4′-dmbp at room temperature using the following
proportions of acid/substrate: 10, 20, and 50 mol/mol. No
added solvent was used in these experiments. The reaction
temperature for all tests was 25 °C. The reactions were
completed in a Pyrex glass reactor of 25-50 mL in volume.
Products were withdrawn at 2, 18, 43, and 67 h for analysis
by GC/MS. The products were prepared for analysis by pouring
into water, separating the organic layer and treating with
aqueous sodium bicarbonate.
3
increasing amounts of BF to a mixture of HF and the
substrate (HF/substrate ) 6). The calculated1 equilib-
7
rium compositions of these three isomers were 18%/58%/
2
4%: o-/m-/p-xylene. They explained these results by a
theory in which a complex equilibrium existed among
neutral species in the organic phase and carbocation
species in the acid phase. When the ratio of BF /substrate
3
exceeded unity, the acid-phase equilibrium dominated the
system. This same phenomenon was observed for other
substrates such as trimethylbenzenes and tetramethyl-
benzenes.
The thermodynamic mixture of isomeric diisopropyl-
benzenes following AlCl catalyzed isomerization con-
3
tained only about 66% of the meta isomer, as was shown
by Olah, Meyer, and Overchuck.18 Later, Olah19 showed
that a mixture of diisopropylbenzene isomers equilibrated
to form the meta isomer exclusively in the presence of
excess HF or a perfluorinated alkanesulfonic acid (e.g.,
trifluoromethane sulfonic acid) and a Lewis acid fluoride
An a lysis. All samples were diluted with acetone (100 µL
per 1.75 mL of acetone) and analyzed on a Hewlett-Packard
HP 5890 Series II GC/HP 5972 MS. The partitioning agent
was a Supelco SPB-5 column (30 m × 0.25 mm × 0.5 µm).
3 5
(BF , PF , etc.). The result was the formation of practi-
cally pure (98-100%) meta-diisopropylbenzene, con-
sistent with control of the reaction by the most stable
Wheland intermediate, the 1,5-diisopropylphenonium
ion.
Resu lts: Isom er iza tion of d m bp
Effect of Acid /Su bstr a te Ra tio. A. Room Tem p er -
a tu r e. The data, Figure 1, showed that the approach to
chemical equilibrium depended on the amount of excess
acid starting with either neat 3,3′- or 4,4′-dmbp. In Figure
Carbocation intermediates have been proposed in the
2
4
isomerization of hexanes and the coupling of methyl-
cyclopentane25 to form the isomers of dimethylspiro-
1
, we show the concentration versus time behavior when
the acid/substrate ratio was 10, 20, or 50 mol/mol for the
,4′-dmbp and when the acid/substrate ratio was 10 mol/
decane and dimethyldecahydronaphthalene. The reactiv-
ity of these alkanes and cycloalkanes in trifluoromethane-
sulfonic acid is relevant in considering their use as
solvents for other organic substrates.
With these results in mind, we explored the superacid-
catalyzed isomerization of related type of substituted
aromatic compoundssdimethylbiphenyls. Pure samples
of 3,3′-dmbp and 4,4′-dmbp were treated with HOTf in
different reaction conditions at room temperature and
elevated temperatures (100 °C) to demonstrate the effect
of changing acid/substrate ratio without a solvent. In
subsequent tests, the effect of several solvents upon the
reactivity and isomer distribution were examined.
4
mol for 3,3′-dmbp as the reactant. The equilibrium
distribution of isomers contained no 2,X′-dmbp isomers
(
X ) 2′, 3′, or 4′), and the 3,4′-dmbp composition was the
greatest of the remaining three isomers. At the highest
acid/substrate ratio, the yield of 4,4′-dmbp was very low
(
(
1-3%) with the 3,4′-dmbp showing the highest yield
67%) with the balance 3,3′-dmbp (30%). In contrast to
these results, the catalytic coupling of toluene to di-
methylbiphenyl in a weakly acidic system (HOAc) showed
a distribution of all six isomers.13 The combined yield of
the 2,X′-dmbp isomers was 37% whereas the 3,3′- and
3
,4′-dmbp isomers showed a combined yield of 53%.
Increasing the amount of acid/substrate changes the
Exp er im en ta l Section
Ca lcu la tion s. AM1 calculations were performed on a
personal computer with Titan software suite (Wavefunction,
Inc., Schr o¨ dinger, Inc.). Isomer distributions were obtained by
calculation of the heats of formation of all cations leading to
each isomer and statistically corrected to free energies.
Ma ter ia ls. Trifluoromethanesulfonic acid and 3,3′- and 4,4′-
dimethylbiphenyl were obtained from Aldrich, as well as the
trajectory of concentration versus time so that high
concentrations of 3,4′-dmbp are formed early in the
reaction trajectory when the acid/substrate ratio in-
creases from 10 to 50 mol/mol. The conversion rate also
increases with increasing ratio of acid/substrate.
B. 100 °C. The effect of changing the acid-to-substrate
ratio was elaborated further for the isomerization of 3,3′-
dmbp at 100 °C using the protocol described above. The
acid/substrate ratio in these studies was 10 and 25 mol/
mol. The results showed that the concentrations of
isomers did not change for reaction times greater than
10 min at either acid/substrate ratio. The distribution of
(
0.
16) McCaulay, D. A.; Lien, A. P. J . Am. Chem. Soc. 1952, 74, 6246-
5
(17) Taylor, W. J .; Wagman, D. D.; Williams, M. G.; Pitzer, K. S.;
Rossini, F. D. J . Res. Natl. Bur. Standards 1946, 37, 95.
(
18) Olah, G. A.; Meyer and Overchuck J . Org. Chem. 1964, 29, 2315.
(19) Olah, G. A. U.S. Patent 4,547,606, 1985.

2
036 J . Org. Chem., Vol. 67, No. 7, 2002
Sherman et al.
F igu r e 1. Isomerization of 3,3′- and 4,4′-dmbp at room temperature.
F igu r e 2. Effect of temperature upon the isomerization of 3,3′- and 4,4′-dmbp at acid/substrate ratio ) 10.
isomers was 2,X′-dmbp ) 0; 3,3′-dmbp ) 39(38)%; 3,4′-
dmbp ) 49(55)%; and 4,4′-dmbp ) 9(5)% (the numbers
in parentheses indicate data for acid/substrate ratio )
Effect of Tem p er a tu r e a t Con sta n t Acid /Su b-
str a te Ra tio. The effect of temperature was demon-
strated for the isomerization of 3,3′-dmbp using an acid/
substrate ratio of 10 mol/mol (Figure 2). The protocol was
as before and the temperatures were 30, 50, and 100 °C.
The reaction time could be reduced to less than 10 min
by using a temperature of 100 °C. The distribution of
isomers did not appear to change significantly with
increasing temperatures: 2,X′-dmbp ) 0%; 3,3′-dmbp )
2
5; the numbers without parenthesis indicate data for
acid/substrate ratio ) 10). This distribution of isomers
obtained by reacting 3,3′-dmbp was very similar to the
distribution (36%/56%/7% for 3,3′-/3,4′-/4,4′-dmbp) ob-
tained from reacting the 4,4′-dmbp at a similar acid/
substrate ratio and 25 °C for 67 h.

Isomerization of Substituted Biphenyls by Superacid
J . Org. Chem., Vol. 67, No. 7, 2002 2037
F igu r e 3. Effect of solvent upon the isomerization of 3,3′-dmbp at 25 °C.
3
1-39%; 3,4′-dmbp ) 49-52%; 4,4′-dmbp ) 8-9%. The
products of the Pd-catalyzed coupling of toluene.13 This
product mixture was purified to remove the unreacted
toluene and the Pd catalyst. Thereafter, the mixture of
dmbp isomers was handled as described above.
effect of temperature was confirmed for the isomerization
of 4,4′-dmbp using the same protocol and an acid/
substrate ratio of 10. The reaction temperatures were 25,
3
3
0, and 50 °C. The final concentration of isomers were
1-37%/52-67%/2-8% for 3,3′-/3,4′-/4,4′-dmbp at these
Effect of Acid /Su bstr a te Ra tio. A. 25 °C. The effect
of varying the acid/substrate ratio from 3.6 to 183 mol/
mol in the presence of toluene as a solvent is shown in
Table S1 (Supporting Information). The reaction tem-
perature was 25 °C, and samples were withdrawn at 20
and 140 h. At low values of the acid/substrate ratio (3.6
mol/mol), the isomerized product distribution changed
only slightly from that of the substrate, even after 140 h
of reaction. However, for very large values of the acid/
substrate ratio (91 and 183 mol/mol) the isomerized
products contained only the 3,3′- and 3,4′-dmbp isomers
conditions. The distribution of isomers (8%/37%/52%) at
the longest reaction time, 67 h, and 50 °C was similar to
that observed when 3,3′-dmbp was isomerized at 100 °C
and for 2 h (9%/39%/49%).
Solven t Effects. The effect of solvent was examined
for its impact on the reactivity and composition of the
final mixture (Figure 3). For these tests, 3,3′-dmbp was
isomerized at 25 °C in a batch reactor. Samples were
withdrawn at times of 20 and 94 h. The acid/solvent ratio
was 1 mol/mol whereas the substrate/solvent ratio was
(26.3 and 73.8%, respectively). Quite remarkable was the
0
.05 mol/mol. Use of substituted aromatics as solvents
complete absence of the 2,X′-dmbp isomers where X′ )
such as o-, m-, p-xylene, cumene, cymene, and mesitylene
appeared to slow the isomerization rate so that little
conversion of substrate was observed after 20 h. Other
data for reaction in solvents such as benzene, toluene,
and octane showed reactant conversion at 20 h that was
similar to that observed for the neat substrate. The data
at 94 h showed a definite trend in isomerization activity
in these different solvents. The product mixtures of the
runs using cumene and p-cymene showed evidence of
disproportionation reactions between the solvents. This
conclusion is based on the fact that di- and triisopropyl-
benzene were found in the reaction with cumene as the
solvent and diisopropyltoluene was found in the reaction
with p-cymene as the solvent. None of the other experi-
ments showed evidence that the solvent reacted during
the isomerization.
2
′, 3′, or 4′.
B. 100 °C. The reaction was completed in 10 min for
acid/substrate ratios of 5 or 10 mol/mol. The isomerized
product distribution (39%/52%/8% ) 3-3′-/3,4′-/4,4′-
dmbp) was nearly the same as the results (39%/49%/9%)
obtained from a pure substrate when the acid/substrate
ratio was 5 or 10 mol/mol (Tables S2 and S3, Supporting
Information). At these reaction conditions, the combined
yields of 2,X′-dmbp isomers was less than 3%, whereas
the substrate showed a combined yield of 37% for the
2
,X′-dmbp isomers.
All these data suggest that all possible isomers, in the
presence of HOTf, were converted into a mixture of 3,3′-,
,4′-, and 4,4′-dmbps. These three isomers define the
3
thermodynamic equilibrium that depends only slightly
upon temperature and acid/substrate ratio provided that
a sufficient amount of superacid is present initially. The
low concentration of 2,X′-species is remarkable.
Resu lts: Isom er iza tion of Mixtu r es
The remaining reactions were completed using, as the
substrate, a mixture of the dmbp isomers that were the
Deu ter iu m -Hyd r ogen Exch a n ge. We described
recently a fast H-D exchange (at NMR time scale)

2
038 J . Org. Chem., Vol. 67, No. 7, 2002
and HOTf.15 The speculation was that
Sherman et al.
between toluene-d
8
a protonated analogue of Wheland complex could facili-
tate the formation on Pd-C bond through a transmeta-
+
lation reaction with consecutive elimination of H :
The same type of carbocations could be responsible for
isomerization of dmbp. The tentative isomerization scheme
could be as follows:
F igu r e 4. Data of 3,4′-dmbp isomer versus time at 35 °C.
confirm the shifting-order kinetics (Figure 6), to deter-
mine the order of the reaction, and to extract the
coefficients of the curve-fits. The appropriate fitting
function was as follows:
R ) R[dmbp]/[γ + â[dmbp]]
Where R was the reaction rate, M/min; R, â, and γ were
fitting coefficients. The solid curves in Figure 5 were
produced using this fitting equation and the coefficients.
These coefficients may be related to kinetic constants
governing the reaction, provided that a mechanism can
be established to explain all of the data. This mechanism
is discussed in a following section.
Resu lts: Ca lcu la ted Equ ilibr iu m Com p osition s
The rate-determining step may be an isomerization of
carbocationic species because all other transformations
The observation of shifting-order kinetics does not
necessarily contradict the hypothesis that the rate-
determining step of the isomerization of dmbp is the
transformation of a protonated dmbp. To elaborate this
consideration, we calculated the differences in free energy
of formation with the most stable carbocations and
consequently the composition of the reaction mixture at
the equilibrium using the AM1 method (Table 1). The
data of the neutrals were that obtained for coupling of
(proton exchanges) are fast.
Rea ction Kin etics. The acid-catalyzed isomerization
of 3,3′-dmbp was examined to determine the reaction
kinetics at the following temperatures: 5, 20 and 35 °C.
The data were collected at low conversion so that the
reverse reaction was not a factor when the data were
extrapolated to zero conversion. The raw data for isomer-
ization of 3,3′-dmbp at 35 °C are shown in Figure 4. The
concentration of the initial product (3,4′-dmbp) was
plotted as a function of time for five, different, initial
concentrations of 3,3′-dmbp (0.05-0.2 M) denoted here
as tests 1 f 5. The conversion of the reactant was kept
low (0.005-0.06) so that the data could be treated by the
toluene with Pd(OAc)
2
in pentanedione at long reaction
times.13 We showed that the isomer distribution did not
1
5
change significantly in absence of a strong acid. These
predictions suggest that the 3,3′- and 3,4′-dmbp isomers
are major products in superacid media, whereas a
distribution of all six isomers is predicted for reactions
in weak acids. This approach was extended to predict the
equilibrium distribution of isomers among others sub-
stituted arenes in the presence of strong acids. These
predictions matched well the reported experimental data
of these same equilibria for xylenes, trimethylbenzenes,
tetramethylbenzenes and the observations in our labo-
ratories for dimethylbiphenyl. The most stable isomer
predicted for the neutral species was not the most stable
isomer predicted for the corresponding carbocation. The
final distribution of isomers was achieved by quenching
the reaction with water and thus preserving the distribu-
tion established in the carbocationic equilibria.
“differential reactor” technique.
Thus, the slope of the concentration versus time data
extrapolated to zero time gave the initial reaction rate
at the corresponding initial concentration of reactant. A
curve-fit of the data, Figure 4, showed the slopes and
intercepts for all five tests at 35 °C. We now cross-plot
the slopes of these five tests versus the initial concentra-
tion of each test to develop Figure 5 that gives the initial,
reaction rate versus initial concentration for T ) 5, 20,
and 35 °C.
The data of Figure 5 suggest shifting-order kinetics.
These rate data were fit to a reciprocal rate plot to

Isomerization of Substituted Biphenyls by Superacid
J . Org. Chem., Vol. 67, No. 7, 2002 2039
F igu r e 5. Initial rates plot.
Ta ble 2. Isom er iza tion of Alk ylben zen esa
equilibrium composition, mol %
cation
isomer
neutrals14 obsd16 predictions
1
1
1
1
1
1
1
1
1
1
1
1
,2-dimethylbenzene
,3-dimethylbenzene
,4-dimethylbenzene
18
58
24
9
60
31
0
70
30
0
66
34
0
100
0
0
0
100
0
0
100
0
99
1
0.2
99.5
0.3
0.0
0.3
99.7
0.5
2.4
97.1
0.0
99.2
0.8
, 2, 3-trimethylbenzene
, 2, 4-trimethylbezene
, 3, 5-trimethylbenzene
, 2, 3, 4-tetramethylbenzene
, 2, 4, 5-tetramethylbezene
, 2, 3, 5-tetramethylbezene
,2-diisopropylbenzene
,3-diisopropylbenzene
,4-diisopropylbenzene
a
Note: calculations and data for dimethylbenzenes were for a
temperature of 80 °C and for tri- and tetramethylbenzenes of
00 °C.
1
results.16 The results of our calculations appear to predict
well the distribution of protonated xylenes at equilibrium
as rationalized by McCaulay and Lien.16 For comparison,
we repeated these calculations for the isomerization of
trimethyl- and tetramethylbenzene. We used this same
approach to model the results of Olah for the isomeriza-
tion of diisopropylbenzene in superacids at room tem-
perature.19
The semiempirical method apparently showed a re-
markable agreement with the experimental results for
isomerization of the substrates by superacids to produce
a mixture that is rich in the meta isomer. In each case,
this observed equilibrium composition is much different
from what could be expected considering the free energy
of formation for the neutral species.
F igu r e 6. Shifting order kinetics plot.
Ta ble 1. Isom er iza tion of Dim eth ylbip h en yls a t 25 °C
predicted and observed
equilibrium composition, mol %
cation
neutralsa
predictedb
isomer
observed
2
2
2
3
3
4
, 2′-dmbp
, 3′-dmbp
, 4′-dmbp
, 3′-dmbp
, 4′-dmbp
, 4′-dmbp
2.0
13.0
10.0
27.0
35.0
13.0
0.0
0.1
0.0
21.6
77.5
0.8
0.0
0.0
0.0
26.3
73.7
0.0
a
Data of ref 13. b Determined from AM1 calculations.
Ca lcu la tion s: Su bstitu ted Ben zen es
Discu ssion
AM1 calculations of the free energies of formation were
used to predict the distribution of isomeric xylenes
carbocations. These results are shown (Table 2) for a
reaction temperature of 80 °C along with the literature
The data for isomerization of either 3,3′- or 4,4′-
dimethylbiphenyl showed that only three (3,3′-/3,4′-/4,4′-
) ∼40/∼55/∼5%) out of the six dmbp isomers were

2
040 J . Org. Chem., Vol. 67, No. 7, 2002
Sherman et al.
produced when the catalyst was triflic acid. No 2,X′-dmbp
isomers (X ) 2, 3, or 4) were observed even after very
long reaction times. Changing the reaction temperature
and or the ratio of acid/substrate altered the rate of
approach to this equilibrium mixture. The same equilib-
rium distribution was observed with a single reactant,
such as 3,3′- or 4,4′-dmbp or a mixture of all six dmbp’s.
Thus, we conclude that the composition observed at long
reaction times was the equilibrium composition. This
distribution of isomers was different from that observed
from the catalytic coupling of toluene in weak acids over
Sch em e 1
a Pd2 catalyst. However, recent results for the coupling
+
15
2
+
of toluene in strong acids over the same Pd catalyst
show distributions of isomers that are rich in the 3,4′-
and 4,4′-dmbp isomers. Apparently, the strong acid
influenced the distribution of isomers either during and/
or after the coupling reaction.
groups. The kinetic data are consistent with the following
reaction mechanism:
+
-
+
-
isomer + H A T [isomer -H] A , reversible (2)
1
1
Different solvents also influenced the rate of isomer-
ization. The more basic methylated aromatic solvents
slowed the isomerization rate, whereas nonaromatic or
weakly basic solvents did not. In the extreme case of
mesitylene, the solvent was protonated preferentially to
the substrate and the isomerization ceased.
+
-
+
-
[
isomer -H] A f [isomer -H] A , slow
(3)
1
2
+
-
+
-
[
isomer -H] A + H O f isomer + [H O] A , fast
2
2
2
3
(4)
Earlier researchers16 speculated that methyl-substi-
Step 1 is the reversible protonation of the arene isomer,
tuted arenes were protonated by HF/BF
3
to form a cation
by triflic acid, and step 2, the “methyl shift”, is the slow,
rate-determining step of this process. The methyl shift
can take place by one of two processes. The first is ipso
protonation to the methyl group, followed by methyl
migration. The second is ipso protonation to the phenyl
group, followed by phenyl migration (see Scheme 1).
Although our studies do not allow us to distinguish
kinetically between the two pathways, we note that the
protonation ipso to phenyl produces an intermediate
having ca. 5.4 kcal/mol higher energy than the first.
Thus, we conclude that the first process predominates.
After the reaction was completed, then water was
added to quench the reaction, step 3. The reaction of
water with the products was very fast and produces a
weaker acid than triflic acid. We showed how small
amounts of water compromised the reactivity of another
reaction catalyzed by triflic acid.20 Therefore, we believe
complex, and the product distribution was determined
by the free energies of this carbocation complex. The
conventional models of arene, cation-complexes were used
to argue that the meta-substituted, arene, carbocation
was the most stable species for the cases of xylene and
trimethyl- and tetramethylbenzene. We calculated the
free energies of formation for neutral and carbo-cation
complexes at the reaction conditions described in the
literature using a semiempirical molecular orbital cal-
culation (AM1) to determine the enthalpy of formation.
From these calculations we predicted the equilibrium
isomer distributions for the carbocations that agreed with
the literature data and expanded it to the isomerization
of di-substituted biphenyls. The agreement was good
between the calculated equilibrium compositions with
experimental data for these substituted arenes. Thus the
success of the AM1 algorithm to predict these equilibrium
distributions represents both a refinement and a gener-
alization of the previous rationalization based upon
qualitative resonance predictions.
that the addition of water (1) releases isomer
acid and (2) prevents the back-isomerization of isomer
into isomer . This kinetic trapping is the key to synthesis
2
from the
2
1
of the thermodynamically unfavored neutral isomer. With
this mechanism we may write a rate law assuming that
the amount of triflic acid initially is related to the amount
of unbound acid and bound acid:
The kinetics of the isomerization were correlated by
shifting-order kinetics:
R ) R[dmbp]/[γ + â[dmbp]]
(1)
+
o
+
+
[
H ] ) [isomer
- H] + [H ]
(5)
total
That this shifting-order equation can be used to cor-
relate the data suggests that an equilibrium step is in
sequence with a slow, kinetic step. The following discus-
sion elaborates this suggestion in more detail.
+
+
[
isomertotal - H] ) [isomer - H] +
1
+
Σi)2,6[isomer - H] (6)
i
1
15
H NMR and mass spectra data clearly demonstrated
that toluene was protonated by triflic acid on the ring to
form an equilibrium mixture at room temperature. It is
logical to suppose that triflic acid also protonates di-
methylbiphenyl in a similar manner. These data together
+
+
d[isomer - H] /dt ) k [isomer ][H ] -
1
1
+
1
+
k-1[isomer - H] - k [isomer - H] ) 0 (7)
1
2
1
1
6
with the data of McCaulay and Lien and the data of
Olah19 for isomerization of substituted benzenes in
superacid add credibility to the reaction mechanism
that we postulate here. The AM1 calculations show that
the distribution of dmbp isomers can be predicted well if
it is assumed that carbocations are formed with energies
that depend on location of the proton vis- a` -vis the methyl
Equations similar to eq 7 may be written for all six
protonated isomers, and these equations may be summed
to give the following expression for initial rate as a
function of initial concentrations.
(20) Xu, Bo-Qing; Sood, D. S.; Gelbaum, L. T.; White, M. G. J . Catal.
1999, 186, 345-52.

Isomerization of Substituted Biphenyls by Superacid
J . Org. Chem., Vol. 67, No. 7, 2002 2041
o
+ o
o
demonstrate a negative entropy change since the proton
and substrate become associated and thus the entropy
is expected to decrease.22
R ) (k /k )k [I ] [H ] /[(1 + k /k ) + (k /k )[I ] ]
1
-1
2
1
2
-1
1
-1
1
(8)
The form of this equation is the same as the equation
Con clu sion s
1
used to fit the reaction rate data (eq 1) where R ) (k /
+
o
k
-1)k
2
[H ] , â ) (k
1
/k-1) ) K, and γ ) (1 + k
2
/k-1).
Triflic acid is an efficient catalyst for isomerization of
dimethylbiphenyls. The study of this process allowed us
to propose the general model of alkyl benzenes isomer-
ization in superacid media. It consists of protonation of
aromatic nucleus with consecutive isomerization of the
initially-formed Wheland intermediate is the rate-limit-
ing step. The thermodynamically unfavored neutral
isomer can be kinetically trapped by quenching the
reaction with water. The semiempirical method AM1
provides an extraordinarily accurate tool for prediction
of the isomer distribution for substituted arenes based
on the free energy of formation for corresponding carbo-
cations.
Equation 8 may be inverted to find the fitting coefficients
using the rate data at 5, 20, and 35 °C (Figure 3).
+
o
o
+ o
(
1/R) ) (1 + k /k-1)/(k K[H ] )(1/[I ] ) + 1/(k [H ] )
2 2 1 2
(9)
The slopes (s) and intercepts (i) of the lines in Figure
3
may be related to the rate constant (k ) and equilibrium
2
constants of protonation (K) through eq 9 as follows:
+
o
k [H ] ) 1/i; K ) (i/s)(1 + k /k-1)
(10)
2
2
The kinetic data were reduced by this equation to give
Figure 5 assuming that the initial concentrations of
protons between each test were nearly the same. Thus,
the rate constant is proportional to 1/i and the equilib-
Abbr evia tion s. R, â, γ, fitting constants for reaction
+
rate equation, eq 1; [H ], concentration of hydrogen ion
+
at time t (M); [H ]°, concentration of hydrogen ion at time
t (M); i, intercepts of lines in Figure 2 (time/M); [I
concentration of isomer n at time t (M); [I ]°, concentra-
tion of isomer n at time t ) 0 (M); k , forward rate
constant for eq 2 (1/M-time); k-1, reverse rate constants
for eq 2 (1/time); k , forward rate constant for eq 3 (1/
time); K, k /k-1; R, reaction rate (M/time); s, slopes of lines
n
],
rium constant equals (i/s) assuming that k-1 . k
2
.
n
Therefore, we may relate the fitting coefficients estab-
lished in Figure 6 with the rate constant of the kinetically
1
slow step, k
2
, and the equilibrium constant of the arene
2
protonation, K, for each temperature. These constants
as a function of temperature were used to determine the
activation energy (22.5 kcal/mol), preexponential factor
1
in Figure 2 (time); T, temperature (K).
1
4
-1
Ack n ow led gm en t. We acknowledge the support of
the KoSa Co. and the thoughtful discussions with
J effrey C. Kenvin (Micromeretics Corp., Norcross, GA).
(1.56 × 10
h ), enthalpy (3.2 kcal/mol), and entropy of
protonation (-14.2 cal/mol-K). The activation energy is
characteristic of a kinetically controlled process as is the
preexponential factor.21 The enthalpy of protonation
suggests that the process is endothermic whereas the
entropy of protonation reflects a large decrease in the
entropy that associates with the protonation. These
results are related to the equilibrium of protonation and
suggest that the solubility of the hydrocarbon in the acid
phase may be responsible for the endothermic enthalpy
of protonation, as it was observed that the apparent two-
phase system disappears when the reaction temperature
is 100 °C. The protonation of the arene is expected to
Su p p or tin g In for m a tion Ava ila ble: Tables S1-S6 and
data from AM1 calculations. This material is available free of
charge via the Internet at http://pubs.acs.org.
J O0106445
(22) Espenson, J . H. Chemical Kinetics and Reaction Mechanisms;
McGraw-Hill Book Co.: New York, 1981; p 122.
(23) Molnar, A.; Olah, G. Hydrocarbon Chemistry; J ohn Wiley and
Sons: New York, 1995.
(24) Farcasiu, D.; Lukinskas, P. J . Chem. Soc., Perkin Trans. 2 2000,
2
295-2301.
(25) Iretskii, A. V.; Sherman, S. C.; White, Mark G. Novel coupling
and isomerization reactions catalyzed by a protic superacid. In
Catalysis of Organic Reactions; Ford, M., Ed.; Marcel-Dekker, Inc.:
New York, 2000; Chapter 15, pp 181-192.
(
21) Gardiner, W. C. Rates and Mechanisms of Chemical Reactions;
W. A. Benjamin, Inc.: Reading, MA, 1969; pp 156-7.