Crucial role of elusive isomeric η-complexes in gas-phase electrophilic aromatic alkylations

Article abstract of DOI:10.1021/jo050019q

The kinetics and stereochemistry of the protonation-induced unimolecular isomerization of (R)-1-D₁-3-(p-fluorophenyl)butane have been investigated in the gas phase at 40-100 °C and 70-760 Torr. This process leads to the formation of the relevant meta and ortho isomers with partial racemization of the migrating sec-butyl moiety. Complete racemization is observed, instead, when the isomerization reaction involves a 1,2-H shift in the moving alkyl group. These results, together with the relevant activation parameters, fully confirm the previous evidence of the occurrence in the alkyl cation/arene PES of noncovalent η-type intermediates of defined structure and stability, lying well below the classical π-complexes, as confirmed by ab initio calculations. Their crucial role in determining the positional selectivity of gas-phase electrophilic aromatic substitutions clearly emerges from the comparison of the present results with the site selectivity measured in the corresponding bimolecular arene alkylation carried out at the same temperatures and pressures.

Full text of DOI:10.1021/jo050019q

Crucial Role of Elusive Isomeric η-Complexes in Gas-Phase
Electrophilic Aromatic Alkylations
Enrico Marcantoni,† Graziella Roselli,† Laura Lucarelli,† Gabriele Renzi,† Antonello Filippi,‡
Cristiano Trionfetti,‡ and Maurizio Speranza*,‡
Dipartimento di Scienze Chimiche, Universita` degli Studi di Camerino, V. S. Agostino 1, I-62032
Camerino (Mc), Italy, and Facolta` di Farmacia, Dipartimento di Studi di Chimica e Tecnologia delle
Sostanze Biologicamente Attive (No. 64), Universita` degli Studi di Roma “La Sapienza”, P.le A. Moro 5,
I-00185 Roma, Italy
maurizio.speranza@uniroma1.it
Received January 4, 2005
The kinetics and stereochemistry of the protonation-induced unimolecular isomerization of (R)-1-
D1-3-(p-fluorophenyl)butane have been investigated in the gas phase at 40-100 °C and 70-760
Torr. This process leads to the formation of the relevant meta and ortho isomers with partial
racemization of the migrating sec-butyl moiety. Complete racemization is observed, instead, when
the isomerization reaction involves a 1,2-H shift in the moving alkyl group. These results, together
with the relevant activation parameters, fully confirm the previous evidence of the occurrence in
the alkyl cation/arene PES of noncovalent η-type intermediates of defined structure and stability,
lying well below the classical π-complexes, as confirmed by ab initio calculations. Their crucial
role in determining the positional selectivity of gas-phase electrophilic aromatic substitutions clearly
emerges from the comparison of the present results with the site selectivity measured in the
corresponding bimolecular arene alkylation carried out at the same temperatures and pressures.
Introduction
nineties, first Audier’s19 and later Gross’ group19 (A&G)
provided convincing mass spectrometric evidence that
σ-intermediates from the gas-phase protonation of alkyl-
Since Dewar’s proposal in 1946,1 the role of the
electrostatically bound π-complex in electrophilic aro-
matic substitutions has been examined thoroughly in
both the gaseous2-20 and condensed phases.21,22 In the
(7) Robin, D.; Prudhomme, P.; Audier, H. E. Adv. Mass Spectrom.
1989, 11a, 614.
(8) Holman, R. W.; Gross, M. L. J. Am. Chem. Soc. 1989, 111, 3560.
(9) Audier, H. E.; Monteiro, C.; Morgues, P.; Bartomieu, D. Org.
Mass Spectrom. 1990, 25, 245.
* To whom correpondence should be addressed. Fax: int. code +
(06)49913602.
† Universita` degli Studi di Camerino.
(10) Bartomieu, D.; Audier, H. E.; Monteiro, C.; Denhez, J. P. Org.
Mass Spectrom. 1991, 26, 271.
‡ Universita` degli Studi di Roma “La Sapienza”.
(1) Dewar, M. J. S. J. Chem. Soc. 1946, 406, 777.
(11) Bartomieu, D.; Audier, H. E.; Monteiro, C.; Denhez, J. P. Rapid
Commun. Mass Spectrom. 1991, 5, 415.
(2) Herman, J. A.; Harrison, A. G. Org. Mass Spectrom. 1981, 16,
423.
(12) Bowen, R. D. Acc. Chem. Res. 1991, 24, 364.
(13) Longevialle, P. Mass Spectrom. Rev. 1992, 11, 157.
(14) Morton, T. H. Org. Mass Spectrom. 1992, 16, 423.
(15) Berthomieu, D.; Brenner, V.; Ohanessian, G.; Denhez, J. P.;
Millie´, P.; Audier, H. E. J. Phys. Chem. 1995, 99, 712.
(16) Aschi, M.; Troiani, A.; Cacace, F. Angew. Chem., Int. Ed. Engl.
1997, 36, 83.
(3) Sen Sharma, D. K.; Ikuta, S.; Kebarle, P. Can. J. Chem. 1982,
60, 2325.
(4) Gru¨tzmacher, H. F.; Filges, U. Int. J. Mass Spectrom. 1985, 64,
193.
(5) Ba¨ther, W.; Kuck, D.; Gru¨tzmacher, H. F. Org. Mass Spectrom.
1985, 20, 572.
(6) Gru¨tzmacher, H. F.; Filges, U. Org. Mass Spectrom. 1986, 21,
(17) Fornarini, S.; Crestoni, M. E. Acc. Chem. Res. 1998, 31, 827.
(18) Aschi, M.; Attina`, M.; Cacace, F. Chem.sEur. J. 1998, 4, 1535.
673.
10.1021/jo050019q CCC: $30.25 © 2005 American Chemical Society
Published on Web 04/19/2005
J. Org. Chem. 2005, 70, 4133-4141
4133

Marcantoni et al.
CHART 1
SCHEME 1
arenes may coexist with the corresponding electrostati-
cally bonded alkyl cation/arene π-complexes. These find-
ings, coupled with the insight gained from other related
work, led A&G to formulate a refined model for gas-phase
electrophilic aromatic substitutions whose reaction co-
ordinate may involve the formation of an early π-adduct
which eventually collapses to a σ-complex. The actual
intermediacy of the π-complex depends on the π-electron
density of the aromatic substrate. The larger the π-elec-
tron donating ability of the arene, the larger the energy
gap between the π-adduct and the more stable σ-complex,
and the stronger the tendency of the first to collapse to
the latter and to undergo alkyl-group isomerization.
Although the concept of electrostatic complexes as
intermediates in electrophilic aromatic substitutions is
now consolidated, almost no information is presently
available about the mutual orientation of their compo-
nents and the nature of the forces holding them together.
This lack of information is particularly unsatisfactory
since, in principle, the structure and dynamics of non-
covalent complexes may control their evolution kinetics
and the selectivity of the substitution process. One of the
first contributions to this crucial theme came from the
study of the proton-induced isomerization kinetics of (S)-
fluorine substituent has electronic properties largely
different than those of the methyl group and (ii) the
abnormal ortho selectivity exhibited by fluorobenzene in
gas-phase alkylations.24-26 The reaction is carried out in
CH4 at 70 and 760 Torr (T ) 40-100 °C) and in the
presence of trace amounts of a radical scavenger (O2) and
a powerful base (B ) (C2H5)3N). 1FR was protonated using
the strong Brønsted acids, CnH5+ (n ) 1 or 2), which can
conveniently be generated in the gas phase by γ-radioly-
sis of CH4.
(+)-1-D1-3-(p-tolyl)butane (1Me in Chart 1) in the gas
S
Experimental Section
phase at 70 Torr and 100-160 °C.23 The emerging
picture, confirmed by theoretical calculations, suggests
that the isomerization may proceed through the inter-
mediacy of low-energy short-lived isomeric η-type struc-
tures (lifetime is ca. 10-10 s). The presence of the η-type
structures encouraged us to reconsider the classical
model of gas-phase arene alkylations and allowed us to
understand the origin of the marked pressure depend-
ence observed in these reactions when carried out at
constant temperatures. Therefore, it would be of extreme
interest to verify if the intermediacy of the short-lived
η-complexes is restricted to the sec-butylation of toluene
or if it is a general mechanism for electrophilic aromatic
substitutions.
Synthetic Procedure. The synthesis of the enantiomeri-
cally pure (R)-1-D1-3-(p-fluorophenyl)butane (1FR) started from
commercially available p-fluoroacetophenone (2 in Scheme 1)
which underwent the Wadsworth-Emmons reaction with a
sodium salt of ethyl (diethoxyphosphoryl)acetate27,28 to produce
the R,â-unsaturated ester 3 in an 85% yield. NMR analysis
showed that the reaction led almost exclusively to the E isomer
without any appreciable formation of the Z isomer. Catalytic
hydrogenation29 smoothly led to the ester 4 (98% yield) which
was subjected to basic hydrolysis with 1 M NaOH at room
temperature to produce the pure free acid 5 without side
reactions.30 The racemic 3-(p-fluorophenyl)butyric acid (5) was
then coupled to optically pure (4S)-4-isopropyl-1,3-oxazolan-
2-one (7)31 via the formation of the mixed anhydride with
pivaloyl chloride (intermediate 6) to yield an almost equimolar
mixture of the (4S)-4-isopropyl-3-[3-(p-fluorophenyl)butanoyl]-
Toward this purpose, we proceeded in the study of the
pressure and temperature dependence of the gas-phase
proton-induced isomerization of another tailor-made
(24) Attina`, M.; Giacomello, P. J. Am. Chem. Soc. 1979, 101, 6040.
(25) Attina`, M.; De Petris, G.; Giacomello, P. Tetrahedron Lett. 1982,
23, 3525.
(26) Attina`, M.; Ricci, A. Tetrahedron Lett. 1991, 32, 6775.
(27) Cooke, M. P.; Gopal, D., Jr. J. Org. Chem. 1994, 59, 260.
(28) Snowden, R. L.; Linder, S. M.; Muller, B. L.; Schulte-Elte, K.
H. Helv. Chim. Acta 1987, 70, 1858.
(29) Durman, J.; Elliott, J.; McElroy, A. B.; Warren, S. J. Chem.
Soc., Perkin Trans. 1 1985, 1237.
(30) Hertweck, C.; Moore, B. S. Tetrahedron 2000, 56, 9115.
(31) Lewis, N.; McKillop, A.; Taylor, R. J. K.; Watson, R. J. Synth.
Commun. 1995, 25, 561.
chiral arylalkane, (R)-1-D1-3-(p-fluorophenyl)butane (1F
R
in Chart 1). This compound was chosen because (i) the
(19) Holman, R. W.; Eary, T.; Whittle, E.; Gross, M. L. J. Chem.
Soc., Perkin Trans. 2 1998, 2187.
(20) Kuck, D. Int. J. Mass Spectrom. 2002, 213, 101.
(21) Olah, G. A. Acc. Chem. Res. 1971, 4, 240.
(22) Olah, G. A.; Lin, H. C. J. Am. Chem. Soc. 1974, 96, 2892.
(23) Filippi, A.; Roselli, G.; Renzi, G.; Grandinetti, F.; Speranza,
M. Chem.sEur. J. 2003, 9, 2072.
4134 J. Org. Chem., Vol. 70, No. 10, 2005

Elusive Isomeric η-Complexes
fluorophenyl)-2-ethanone (2) (4 g, 28.96 mmol) in THF (25 mL)
was added dropwise within 20 min. The mixture was then
refluxed for 24 h and then cooled with a water bath. Then a
saturated aqueous ammonium chloride solution (20 mL) was
added dropwise to the cold mixture. The aqueous phase was
extracted with diethyl ether (4 × 100 mL), and the combined
organic phase was washed with brine (3 × 50 mL), dried over
sodium sulfate, and concentrated in vacuo. Flash chromatog-
raphy (95:5 hexanes/ethyl acetate) yielded ester 3 as a clear
oil (5.12 g, 85% yield). The product was characterized by IR
and GC-MS: IR (thin film) ν ) 3050, 2982, 1717, 1631, 1602
cm-1; MS (70 eV, EI) m/z (%) 208 (M+, 46), 179 (30), 163 (100),
133 (70), 115 (50), 109 (44), 96 (17).
SCHEME 2
Ethyl 3-(p-Fluorophenyl)butanoate (4). Unsaturated
ester 3 (4.76 g, 22.9 mmol) in absolute ethanol (25 mL) was
hydrogenated at 1 atm over 10% palladium on activated carbon
(1.724 g) for 15 h. The reaction mixture was filtered through
Celite with absolute ethanol as the eluent. The filtrate was
evaporated under reduced pressure to produce ester 4 (4.74
g, 98% yield). The GC-MS characterization data were consist-
ent with the target product: MS (70 eV, EI) m/z (%) 210 (M+,
15), 136 (71), 123 (100), 109 (27), 103 (31), 96 (12), 77 (11).
3-(p-Fluorophenyl)butanoic Acid (5). Aqueous 1 N so-
dium hydroxide (150 mL) was added to ester 4 (4.74 g, 22.54
mmol), and the emulsion was stirred overnight at room
temperature. The reaction mixture was extracted with diethyl
ether, and the ethereal solution was discarded. The aqueous
phase was acidified to pH 1 with an aqueous 1 N hydrochloride
acid solution and then extracted with diethyl ether (5 × 120
mL). The combined organic extracts were dried over sodium
sulfate and concentrated in vacuo to produce pure acid 5 (4.1
g, 82% yield).
1,3-oxazolan-2-one diastereoisomers 8 and 9.32 The relative
abundance was determined on the crude material via well-
1
resolved signals in the H NMR spectrum. Indeed, since the
1H NMR spectra of these diastereomers are distinctly differ-
ent,33 the NMR chemical shift and coupling constant (3J) of
the proton of C(4) of the oxazolidinone ring provide a reliable
guide for the assignment of the C(3) configuration of the
butyric chain. It should be noted that the use of different
substituents at C(4) of oxazolidinone does not permit good
assignment of the configuration by NMR spectra. Optically
pure stereoisomer 8 was next purified using flash chromatog-
raphy to produce a 26% yield (based on racemic butyric acid
5). The chiral auxiliary was subsequently removed using
LiBH4 in ether34 to produce (R)-3-(p-fluorophenyl)butan-1-ol
10. The transformation of alcohol 10 into the chiral 1-D1-3-
arylbutane target was achieved by the Kumada method.35 To
this end, alcohol 10 was converted into the related (R)-1-D1-
3-(p-fluorophenyl)butane (1FR) via tosyl ester derivative 11 and
substitution with LiAlD4 in ether (Scheme 2).
(4S)-3-[(3R)-3-(p-Fluorophenyl)butanoyl]-4-isopropyl-
1,3-oxazolan-2-one (8). Triethylamine (1.1 mL, 7.88 mmol)
and pivaloyl chloride (1 mL, 7.88 mmol) were added to a -78
°C solution of racemic acid 5 (1.35 g, 7.41 mmol) in dry THF
(33.2 mL). The resulting white suspension was stirred for 30
min at -78 °C and for 90 min at 0 °C and was cooled to -78
°C. Meanwhile, in a different flask, a solution of metalated
oxazolidinone (7) was prepared by dropwise addition of n-BuLi
(1.5 M in hexane, 5.3 mL, 8 mmol) to a -78 °C solution of
(4S)-4-isopropyl-1,3-oxazolan-2-one (1 g, 7.7 mmol) in dry THF
(29 mL), and the mixture was stirred for 30 min at -78 °C.
The preformed mixed anhydride 6 was transferred via cannula
into the reaction flask containing the lithiated chiral auxiliary
at -78 °C. Then, the reaction mixture was stirred at this
temperature for 30 min, and after being allowed to warm to
room temperature overnight, the mixture was quenched with
saturated ammonium chloride solution (20 mL) and THF and
evaporated in vacuo. The mixture was diluted with methylene
chloride (300 mL) and washed with saturated sodium hydrogen
carbonate (2 × 30 mL), water (2 × 30 mL), and brine (30 mL).
The organic phase was dried over sodium sulfate and concen-
trated in vacuo. Flash chromatography on a silica gel column
(6:3.5:0.5 hexanes/benzene/ethyl acetate) produced optically
pure 8 as a white solid (0.549 g, 26% yield): 1H NMR (300
MHz, CDCl3) δ ) 7.35-7.19 (m, 2H), 7.00-6.92 (m, 2H), 4.41-
4.35 (m, 1H), 4.26-4.11 (m, 2H), 3.57-3.32 (m, 2H), 2.94 (dd,
J ) 15.02 and 6.23 Hz, 1H), 2.18-2.11 (m, 1H), 1.30 (d, J )
6.96 Hz, 3H), 0.82 (d, J ) 6.96 Hz, 3H), 0.66 (d, J ) 6.96 Hz,
3H); 13C NMR (75 MHz, CDCl3) δ ) 14.5, 18.0, 22.3, 28.5, 35.8,
43.3, 58.5, 63.4, 115.2, 115.5, 128.6, 128.7, 141.5, 154.2, 160.0,
163.3, 171.9; MS (70 eV, EI) m/z (%) 293 (M+, 13), 278 (7), 149
(15), 136 (100), 123 (98), 109 (27), 103 (37), 77 (13), 55 (4), 41
(21). Elemental analysis calcd (%) for C16H20FNO3 (293.14):
C, 65.51; H, 6.87; N, 4.78. Found: C, 65.45; H, 6.27; N, 4.32.
Most solvents and reagents were used without purification
unless otherwise mentioned. Solvents (EtOAc and hexanes)
for flash chromatography were distilled. TLC was run on silica
gel 60 F254 plates (0.25 mm thick) and visualized by UV light,
iodine, or a Ce-Mo staining solution (phosphomolybdate, 25
g; Ce(SO4)2‚2H2O, 10 g; concentrated H2SO4, 60 mL; water,
940 mL) with heating. Flash chromatography was performed
using silica gel 60 (mesh size 0.040-0.063 mm) with a silica
gel/crude compound weight ratio of ca. 30:1. Analytical GC was
performed with a capillary-fused silica column (0.32 mm × 25
mt), stationary-phase OV1 (film thickness 0.40-0.45 µm).
Infrared spectra were taken using thin films on NaCl plates.
1
Only the characteristic peaks are noted. H NMR (300 MHz)
chemical shifts are reported in parts per million (δ) relative
to CHCl3 (7.27 ppm). Coupling constants (J) are reported in
hertz. 13C NMR spectra were obtained on a 75 MHz spectrom-
eter and are reported in δ relative to CDCl3 (77.2 ppm) as an
internal reference. Mass spectra were determined with a gas
chromatograph/mass spectrometer (GLC-MS, EI-70 eV). A
fused silica column (30 mt × 0.25 mm, cross-linked 5% Ph-
Me siloxane, 0.10 µm film thickness) was used with a helium
carrier flow rate of 1 mL/min. The temperature of the column
was varied, 3 min after the injection, from 65 to 300 °C with
a slope of 15 °C min-1
.
Ethyl 3-(p-Fluorophenyl)-2-butenoate (3). A solution of
triethyl phosphonoacetate (7.58 g, 33.8 mmol) in THF (10 mL)
was added dropwise within 30 min to a stirred slurry of NaH
(0.880 g, 36.68 mmol) in THF (90 mL) at room temperature
under nitrogen. During the addition, the temperature rose to
30-35 °C, and after a further 30 min, a solution of 1-(p-
(32) Li, G.; Patel, D.; Hruby, V. J. J. Chem. Soc., Perkin Trans. 1
1994, 3057.
(33) Dharanipragada, R.; VanHulle, K.; Bannister, A.; Bear, S.;
Kennedy, L.; Hruby, V. J. Tetrahedron 1992, 48, 4733.
(34) Penning, T. D.; Djuric, S. W.; Haack, R. A.; Kalish, V. J.;
Miyashiro, J. M.; Rowell, B. W.; Yu, S. S. Synth. Commun. 1990, 20,
307.
(35) Hayashi, T.; Konishi, M.; Fukushima, M.; Mise, T.; Kagotani,
M.; Tajika, M.; Kumada, M. J. Am. Chem. Soc. 1982, 104, 180.
(3R)-3-(p-Fluorophenyl)butan-1-ol (10). The acyloxa-
zolidinone 8 (0.53 g, 1.87 mmol) was stirred in dry ethyl ether
(38 mL) under nitrogen. Water (37.4 µL, 2.06 mmol) was
added, and the reaction mixture was cooled to 0 °C in an ice
bath. Lithium borohydride (1.03 mL, 2.06 mmol) was added
dropwise causing immediate clouding and hydrogen evolution.
J. Org. Chem, Vol. 70, No. 10, 2005 4135

Marcantoni et al.
The mixture was allowed to warm to room temperature and
was stirred for 3 h at this temperature. The reaction was then
quenched with an aqueous sodium hydroxide solution (1 M,
40 mL), and the mixture was stirred until both layers were
clear. The mixture was extracted with diethyl ether (5 × 80
mL), and the organic phase was washed with water (2 × 20
mL) and brine (20 mL), dried over sodium sulfate, and
concentrated in vacuo. Flash chromatography (3.5:6.5 hexanes/
ethyl acetate) produced primary alcohol 10 (0.28 g, 90% yield)
along with the recovered chiral auxiliary. The product was
characterized by IR and 1H NMR: IR (thin film) ν´ ) 3400,
2957, 1641, 1219 cm-1; 1H NMR (300 MHz, CDCl3) δ ) 7.20-
7.11 (m, 2H), 7.04-6.94 (m, 2H), 3.55 (dt, J ) 6.85 and 6.32
Hz, 2H), 2.88 (bs, OH, 1H), 2.50-2.57 (m, 1H), 1.64-1.72 (m,
2H), 1.20 (d, J ) 7.02 Hz, 3H). Elemental analysis calcd (%)
for C10H13FO (168.21): C, 71.40; H, 7.79. Found C, 71.45; H,
7.27.
(3R)-3-(p-Fluorophenyl)butyl 4-Methyl-1-benzenesulfon-
ate (11). A solution of alcohol 10 (0.154 g, 0.916 mmol) in
dichloromethane (2 mL) was treated with tosyl chloride (0.22
g, 1.16 mmol) and pyridine (97 µL, 1.21 mmol) and stirred at
room temperature for 5 h. The reaction mixture was then
diluted with diethyl ether (100 mL) and washed with a 5%
aqueous hydrochloride solution (3 × 25 mL). The organic phase
was washed with a saturated aqueous sodium hydrogen
carbonate solution (2 × 20 mL) and brine (2 × 15 mL), dried
over sodium sulfate, and concentrated in vacuo. Flash chro-
matography (9:1 hexanes/ethyl acetate) produced 11 as a pale
yellow oil (0.26 g, 85% yield). The product was characterized
by GC-MS: MS (70 eV, EI) m/z (%) 322 (M+, 19), 150 (63),
135 (100), 123 (51), 109 (31), 91 (47), 103 (37), 77 (11), 65 (21).
(3R)-1-D1-3-(p-Fluorophenyl)butane (1FR). An ether so-
lution (5 mL) of the tosylate 11 (0.224 g, 0.694 mmol) was
added to a suspension of lithium aluminum deuteride (33.2
mg, 0.79 mmol) in ether (7 mL) at 0 °C with an ice bath. The
reaction mixture was stirred at room temperature for 10 min
and then refluxed for 5 h. After hydrolysis by successive
addition of (i) water (0.03 mL), (ii) 15% aqueous sodium
hydroxide solution (0.03 mL), and (iii) water (0.09 mL), the
precipitate formed was eventually filtered off. The filtrate (180
mL) was washed with a saturated aqueous sodium hydrogen
carbonate solution (2 × 25 mL) and brine (2 × 25 mL) and
dried over magnesium sulfate. The solvent was removed in
vacuo. Flash chromatography on a silica gel column (1:1
diethyl ether/petroleum ether) produced the target product,
1FR (57 mg, 73% yield). The product was characterized by GC-
MS: MS (70 eV, EI) m/z (%) 153 (M+, 23), 123 (100), 109 (27),
103 (39), 96 (9), 77 (11), 51 (4).
Radiolytic Procedure. Methane, n-butane, fluorobenzene,
oxygen, and triethylamine were commercially available high-
purity compounds used without further purification. The
gaseous mixtures were prepared by conventional procedures
with the use of a greaseless vacuum line. Two sets of experi-
ments were carried out. In the first series, the starting chiral
arene, 1FR, the thermal radical scavenger, O2, and the base,
(C2H5)3N (proton affinity (PA) ) 234.7 kcal mol-1),36 were
introduced into carefully outgassed 130 mL Pyrex bulbs, each
equipped with a break-seal tip. The bulbs were filled with CH4
(60 or 750 Torr), cooled to liquid-nitrogen temperature, and
sealed off. In the second set of experiments, the gaseous
mixtures were prepared containing traces of fluorobenzene 16,
O2, (C2H5)3N, and butane (60 or 750 Torr), cooled to liquid-
nitrogen temperature, and sealed off. The gaseous mixtures
were submitted to irradiation at a constant temperature (40-
100 °C) in a 60Co γ-source (dose 2 × 104 Gy, dose rate 1 × 103
Gy h-1, determined with a neopentane dosimeter). Control
experiments, carried out at doses ranging from 1 × 104 to 1 ×
105 Gy, showed that the relative yields of products are largely
independent of the dose. The radiolytic products were analyzed
by GLC-MS. The following chiral columns were used: (i) a 25
m long, i.d. 0.25 mm i.d., df ) 0.25 µm MEGADEX 5 column
operated at 40 < T < 170 °C, 4 °C min-1; (ii) a 25 m long, 0.25
i.d., df ) 0.25 µm CP-Chirasil-DEX CB column operated at 40
< T < 170 °C, 4 °C min-1; and (iii) a 25 m long, 0.25 i.d., df )
0.25 µm Dimethylpentyl â-CDX (PS 086) column operated at
40 < T < 170 °C, 4 °C min-1. The yields of the radiolytic
products were determined from the areas of the corresponding
eluted peaks using an internal standard (acetophenone) and
individual calibration factors to correct for the detector re-
sponse. Control experiments were performed to rule out the
occurrence of thermal fragmentation, isomerization, and ra-
cemization of the starting arene as well as of their isomeric
products within the temperature range investigated.
The D-content and location in the radiolytic products from
the first set of experiments were determined by GLC-MS; the
mass analyzer was set in the selected ion mode (SIM). The
ion fragments at m/z 123 ([M - C2H4D]+), 124 ([M - C2H5]+),
and 153 ([M]•+) were monitored to analyze all of the GLC-
separated isomers of 1F ()M). Fluorobenzene 16, obtained
R
from 1FR, was analyzed by monitoring its parent ion at m/z
96.
Computational Details. Quantum-chemical ab initio cal-
culations were performed using the Unix version of the
Gaussian 98 program.37 The geometries of the investigated
species have been optimized, by analytical-gradient techniques,
at the HF/6-31+G** level of theory, and the located critical
points have been unambiguously characterized as true minima
on the potential energy surface by computing, at the same
computational level, the corresponding analytical vibrational
frequencies. The latter values were used to calculate the zero-
point vibrational energies (ZPE).
Results
Radiolytic Experiments. Table 1 shows the composi-
tion of the irradiated systems as well as the identity and
the yields (Yi) of the major products, i.e., (()-1-D1-3-(m-
fluorophenyl)butane (12), (()-1-D1-2-(m-fluorophenyl)-
butane (13), (()-1-D1-3-(o-fluorophenyl)butane (14), (()-
1-D1-2-(o-fluorophenyl)butane (15), and fluorobenzene
+
(16), obtained from gas-phase CnH5 (n ) 1 or 2) proto-
nation of 1FR in the presence of trace amounts of (C2H5)3N
(0.1 Torr) (Scheme 3). The table does not include other
minor products, i.e., isomeric (()-1-D1-3-(tolyl)butanes,
which arise from the well-known addition to the methane
bulk gas of para sec-butylphenyl cations, generated by
CnH5 (n ) 1 or 2)-induced defluorination of 1F
.
R
It
+
38
should be noted that, despite a specific search, no
appreciable amounts of (S)-1-D1-3-(p-fluorophenyl)butane
or (S)-1-D1-2-(p-fluorophenyl)butane were detected among
the products.
The numbers in the table represent average values
obtained from several separate irradiations carried out
under the same experimental conditions, and their
reproducibility is expressed by the uncertainty level
(37) Frish, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.;
Stratman, R. E., Jr.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels,
A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.;
Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford,
S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma,
K.; Malik, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Cioslowski, J.; Ortiz, J. V.; Raboul, A. G.; Stefaniv, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Nanayakkara, I.; Gonzales, C.; Challacombe,
M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzales, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian
98, revision A7; Gaussian, Inc.: Pittsburgh, PA, 1998.
(36) National Institute of Standards and Technology (NIST)-USA,
http://webbook.nist.gov/chemistry/.
(38) Speranza, M.; Cacace, F. J. Am. Chem. Soc. 1977, 99, 3051.
4136 J. Org. Chem., Vol. 70, No. 10, 2005

Elusive Isomeric η-Complexes
TABLE 1. Kinetics of Gas-Phase Isomerization and Dealkylation of Protonated (R)-(+)-1-D-3-(p-Fluorophenyl)butane
1F
R
rxn
rxn
e
e
e
e
e
syst compa temp
time
k12
k13
k14
k15
k16
(Torr) 1F
(°C) (τ) (×107 s)b Y12 Y13 Y14 Y15
Y16
∑
(×10-5 s-1
)
(×10-5 s-1
)
(×10-5 s-1
)
(×10-5 s-1
)
(×10-5 s-1
)
c
c
c
c
c
d
R
0.30
0.29
0.39
0.16
0.36
40
60
1.14
1.40
1.22
1.27
1.90
0.56 0.50 0.42 0.48 5.07 7.03 0.50 (4.70) 0.46 (4.66) 0.39 (4.59) 0.43 (4.64)
1.60 0.99 1.08 0.96 12.28 16.91 1.25 (5.10) 0.77 (4.89) 0.84 (4.93) 0.75 (4.87)
2.46 1.51 2.39 1.96 16.92 25.24 2.31 (5.36) 1.41 (5.15) 2.26 (5.35) 1.84 (5.27) 15.94 (6.20)
3.80 2.23 3.84 2.90 26.94 39.71 3.82 (5.58) 2.24 (5.35) 3.86 (5.59) 2.91 (5.46) 27.07 (6.43)
8.24 4.06 7.00 5.07 29.79 54.16 5.83 (5.79) 3.08 (5.49) 5.29 (5.72) 3.84 (5.58) 22.53 (6.35)
4.61 (5.66)
5.95 (5.98)
70
90
100
a CH4, 60 Torr; O2, 10 Torr; (C2H5)3N, 0.1 Torr; radiation dose, 2 × 104 Gy (dose rate 1 × 104 Gy h-1). b Reaction time, τ, calculated
from the reciprocal of the first-order collision constant between the relevant ionic adduct and (C2H5)3N (see text). c Y ) 100[G(product)/
G(CnH5+)] (G values expressed as the number of molecules of the given species per 100 eV of energy absorbed by the gaseous mixture).
The percent distribution of products is in parentheses. d ∑ ) Y(()-12 + Y(()-13 + Y(()-14 + Y(()-15 + Y16. Each value is the average of
several determinations, with an uncertainty level of ca. 5%. e Log k in parentheses.
TABLE 2. Retention vs Inversion of Configuration in the Gas-Phase Isomerization of Protonated
(R)-(+)-1-D-3-(p-Fluorophenyl)butane 1F at 70 Torra
R
syst compa
rxn temp
(°C)
12
(ret/inv)
13
(ret/inv)
14
(ret/inv)
15
(ret/inv)
(14+15)/(12+13)
(ortho/meta)
1FR (Torr)
0.30 (0.35)
0.29 (0.32)
0.39 (0.19)
0.16 (0.19)
0.36 (0.35)
40
60
1.44 (1.50)
1.86 (1.38)
1.78 (2.03)
1.86 (1.70)
2.12 (2.57)
1.17 (1.08)
1.04 (1.17)
1.04 (0.82)
1.00 (1.00)
1.04 (1.04)
2.33 (1.13)
1.27 (1.27)
1.50 (1.04)
1.44 (1.38)
1.70 (1.27)
0.85 (1.04)
0.89 (1.00)
0.96 (0.82)
1.00 (0.96)
1.00 (0.89)
0.85 (1.38)
0.79 (1.00)
1.10 (1.50)
1.12 (1.50)
0.98 (1.50)
70
90
100
a Values in parentheses refer to experiments carried out at CH4, 750 Torr; O2, 10 Torr; (C2H5)3N, 0.1 Torr; radiation dose, 2 × 104 Gy
(dose rate 1 × 104 Gy h-1).
SCHEME 3
quoted. The ionic origin of the radiolytic products is
demonstrated by the sharp decrease (more than 80%) of
their abundance as the (C2H5)3N concentration is raised
from ca. 0.1 to ca. 0.5 mol %.
The enantiomeric distribution of products 12-15 (ret/
inv) is given in Table 2. The observed product pattern
results from the isomerization of the ipso-protonated
intermediate I of Scheme 3, while fluorobenzene 16
derives from its de-sec-butylation. The long-recognized
reluctance of the aromatic fluorine to migrate around
the aromatic ring39 coupled with the extensive racem-
ization and the side-chain H-shift, observed in the isomer-
(39) Cacace, F.; Speranza, M. J. Am. Chem. Soc. 1976, 98, 7305.
J. Org. Chem, Vol. 70, No. 10, 2005 4137

Marcantoni et al.
TABLE 3. Arrhenius Parameters for the Gas-Phase Isomerization and Dealkylation of I
process
Arrhenius equation (y ) 1000/2.303RT)
corr coeff (r2)
∆H* (kcal mol-1
)
∆S* (cal mol-1 K-1
)
I f II
I f III
I f IV
I f V
log k12 ) (11.4 ( 0.3) - (9.5 ( 0.5)y
log k13 ) (9.9 ( 0.4) - (7.5 ( 0.6)y
log k14 ) (11.8 ( 0.7) - (10.4 ( 1.1)y
log k15 ) (10.7 ( 0.7) - (8.8 ( 1.0)y
log k16 ) (12.0 ( 0.6) - (6.4 ( 1.3)y
0.990
0.983
0.969
0.959
0.893
8.9 ( 0.6
6.9 ( 0.9
9.8 ( 1.1
8.2 ( 1.0
6.5 ( 1.2
-8.4 ( 1.6
-15.2 ( 1.7
-6.2 ( 3.1
-11.3 ( 3.1
-12.0 ( 3.6
I f 16
FIGURE 2. Temperature dependence of the k14 (0) and k15
(9) rate constants for the I f IV and I f V reactions,
respectively.
FIGURE 1. Temperature dependence of the k12 (O) and k13
(b) rate constants for the I f II and I f III reactions,
respectively.
TABLE 4. Isomeric Distribution of
2-(Fluorophenyl)butanes from the Gas-Phase
sec-Butylation of Fluorobenzenea
ization products, rules out any alternative isomerization
pathways.
According to Scheme 3, and with the assumption that
the CnH5+ protonation of 1FR leads predominantly to the
products of Table 1 and that the deprotonation of the
relevant ionic precursor by (C2H5)3N is fast, the rate
constants for de-sec-butylation of I (k16), its isomerization
without side-chain rearrangement (k12 and k14), and its
isomerization with side-chain rearrangement (k13 and k15)
can be calculated by inserting the appropriate Yi values
in eq 1, where Yi ()100 × [Gi/G(CnH5+)])40 is the absolute
percent yield of the ith product relative to the starting I,
∑ ) Y12 + Y13 + Y14 + Y15 + Y16, and τ is the collision
time between the arenium ions and B ) (C2H5)3N (τ )
(kcoll[B])-1) at the given temperature.41
rxn
syst comp n-butane temp ortho meta para ortho/ overall
16 (Torr)
(Torr)
(°C)
(%)
(%)
(%) meta
G(M)
0.99
0.74
0.70
0.83
0.79
0.77
70
70
400
760
760
760
40
100
40
43
45
72
76
68
55
21
29
5
36
26
2.05
1.55
3.69
2.89
0.53
0.95
1.26
0.32
23 14.40
20 19.00
26 11.33
40
4
70
100
6
11
34
5.00
a O2, 10 Torr; (C2H5)3N, 0.1 Torr; radiation dose, 2 × 104 Gy
(dose rate 1 × 104 Gy h-1).
I. Thus, the ionic sec-butylation of fluorobenzene was
carried out at temperatures from 40 to 100 °C using free
sec-butyl cations generated by the γ-radiolysis of butane
(70-760 Torr). At any temperature, 2-(o-fluorophenyl)-
butane is the dominant product which is accompanied
by minor amounts of the para and meta regioisomers
(Table 4). The abundance of these latter products in-
creases when the pressure decreases and temperature
increases, in agreement with the behavior of strictly
related gas-phase electrophilic alkylations on the same
substrate.24-26 The striking difference in the [2-(o-fluoro-
phenyl)butane]/[2-(m-fluorophenyl)butane] yield ratio from
the sec-butylation of fluorobenzene (ortho/meta: 2.0 (40
°C); 1.5 (100 °C); Table 4) and that obtained from I
isomerization under the same conditions (ortho/meta: 0.8
(40 °C); 1.0 (100 °C); Table 1) excludes the possibility that
the two reactions proceed through the intermediacy of
the same loosely bound alkyl cation/arene π-complexes.
Theoretical Calculations. The stable intermediates
and transition structures involved in I isomerization were
investigated by quantum-chemical ab initio calculations
at the HF/6-31+G** level of theory. The exploration of
the potential energy surface (PES) of I around conceiv-
Y ln[100/(100 - )]
∑
i
ki )
(1)
τ
∑
The Arrhenius plots of k12 and k13 over the 40-100 °C
temperature range are reported in Figure 1. Figure 2
illustrates the Arrhenius plots of k14 and k15 over the
same temperature range. A similar linear plot is obtained
for the de-sec-butylation of I under the same conditions
(k16). The relevant linear curves obey the equations listed
in Table 3, which also shows the corresponding activation
parameters, as calculated from the transition state
theory.
Ancillary experiments were carried out to shed some
light on the specific mechanism of the isomerization of
(40) G values are expressed as the number of molecules produced
per 100 eV of energy absorbed by the gaseous mixture. G(CnH5+) )
1.9 (n ) 1), 0.9 (n ) 2). See: Ausloos, P.; Lias, S. G.; Gorden, R., Jr. J.
Chem. Phys. 1963, 39, 3341.
(41) The collision constant, kcoll, between the arenium ions of Scheme
3 and (C2H5)3N is calculated according to the following: Su, T.;
Chesnavitch, W. J. J. Chem. Phys. 1982, 76, 5183.
4138 J. Org. Chem., Vol. 70, No. 10, 2005

Elusive Isomeric η-Complexes
FIGURE 3. Geometrical parameters of the main HF/6-31+G**-optimized critical structures located on the potential energy surface
(PES) of I (bond lengths in Å). See also Supporting Information for structures I-V and for computational details.
able regions led to the location of several distinct critical
structures which were unambiguously characterized as
true minima or first-order saddle points by analytical
computation of the corresponding vibrational frequencies.
The connectivity and the main geometrical parameters
of these structures are shown in Figure 3.
Discussion
Energetics. Gas-phase protonation of 1F by the
R
strong Brønsted acids, CnH5+ (n ) 1 or 2), is an exother-
mic process yielding the corresponding arenium ion
protomers. Formation of the ipso-protonated intermediate
I is estimated to be exothermic by 53 (n ) 1) and 20 kcal
mol-1 (n ) 2).42 The excess energy imparted to ion I and
the CnH4 fragments by the exothermicity of their forma-
tion process is efficiently dissipated by multiple unreac-
tive collisions with the bulk CH4 gas. As illustrated in
Scheme 3, ipso-intermediate I isomerizes to the meta and
ortho structures II and IV, respectively, with no H-
scrambling and partial racemization of the sec-butyl
group (Table 2), or to the meta and ortho structures III
and V, respectively, with complete rearrangement and
racemization of the sec-butyl group (Table 2). Alterna-
tively, ipso-intermediate I can undergo de-sec-butylation
to eventually produce fluorobenzene 16 (Table 1).
FIGURE 4. Schematic potential energy reaction coordinate
diagram for the gas-phase sec-butylation of fluorobenzene.
23 kcal mol-1 endothermic (Figure 4), in good agreement
with previous calculations carried out at the AM1 semi-
empirical level.19
The Isomerization and Dealkylation Reactions.
As pointed out above, the formation of 1-D1-3-(m-fluoro-
phenyl)butane (12) from 1F must be the result of the I
R
With the reasonable assumption that the proton af-
f II isomerization (path (ii) in Scheme 3) involving the
partial racemization of the chiral sec-butyl moiety (Table
2). This isomerization process is accompanied by a I f
IV isomerization (path (iv) in Scheme 3), almost as
extensive as the I f II isomerization, leading to a
finity (PA) of 1F at C(1) is comparable to that of
R
p-fluorotoluene (PA ) 182.6 kcal mol-1),42 the heat of
formation of I can be estimated as 132 kcal mol-1
.
According to HF/6-31+G** calculations, the heat of
formation of II (and III) can be estimated as 140 kcal
mol-1, whereas that of IV (and V) is ca. 135 kcal mol-1 43
.
(43) According to ref 42, the proton affinity of the para position of
fluorobenzene is identical to that of p-fluoro-tert-butylbenzene at C(1)
(PA ) 182.6 kcal mol-1). It is therefore plausible to take the same
proton affinity for the C(1) center of p-fluoro-sec-butylbenzene. Given
the vanishingly small effect of the alkyl group on the PA of fluoro-
alkylbenzenes at the C-alkyl site, the PA of isomeric fluoro-sec-
butylbenzenes is taken to be equal to that of the corresponding ortho,
meta, and para positions of fluorobenzene.
Unimolecular dissociation of I to produce either the
separated fluorobenzene/sec-butyl cation or the p-fluoro-
benzenium ion/2-butene pairs can be calculated to be ca.
(42) Hunter, E. P.; Lias, S. G. J. Phys. Chem. Ref. Data 1998, 27,
413.
J. Org. Chem, Vol. 70, No. 10, 2005 4139

Marcantoni et al.
partially racemized 1-D1-3-(o-fluorophenyl)butane (14).
The good linearity of the relevant Arrhenius plots
(Figures 1 and 2) supports the view that isomerization
processes (ii) and (iv) proceed independently without any
appreciable II T IV interconversion. Similarly independ-
ent reactions involve a 1,2-H shift in the moving sec-butyl
moiety, i.e., I f III and I f V (paths (iii) and (v) in
Scheme 3, respectively). They lead to completely racem-
ized (()-1-D1-2-(m-fluorophenyl)butane (13) and (()-1-
D1-2-(o-fluorophenyl)butane (15), respectively. It may be
argued, in this case, that the partial racemization
observed in the formation of 12 and 14 from I is to be
ascribed to a fast reversible 1,2-H shift in the moving
sec-butyl moiety. This process would lead to the formation
of [(()-12]/[(()-13] ) [(()-14]/[(()-15] ) ca. 1 and,
therefore, contribute to the racemization of 12 and 14.
However, a fast reversible 1,2-H shift in the side chain
of I is also expected to produce appreciable amounts of
both (S)-1-D1-3-(p-fluorophenyl)butane and (S)-1-D1-2-(p-
fluorophenyl)butane which instead are completely absent
from the products. On these grounds, it is concluded that
the formation of 13 and 15 proceeds through just a single
irreversible 1,2-H shift in the side chain of I leading to
transient species which rapidly convert to the corre-
sponding σ-bonded intermediates.
The incomplete sec-butyl group racemization in the I
f II and I f IV reactions points to transition structures
in which the C(1)-C(R) bond (Chart 1) is sufficiently
elongated to allow positional shifting of the sec-butyl
moiety but not so elongated to allow full rotation around
its C(R)-C(â) bond.8,15,44 Exploration of the PES govern-
ing the I f II and I f IV reactions led to the location of
several distinct critical points, identified as true minima
(i.e., A and B of Figure 3) and as transition structures
(i.e., C, D, and E of Figure 3).
simultaneously faces both the C(2) and C(5) ring atoms
thus pointing to a facile meta T ortho′ isomerization,
when allowed by the experimental conditions.
The activation parameters measured for the I f IV
isomerization (iv) (∆Hq ) 9.8 ( 1.1 kcal mol-1 and ∆Sq )
-6.2 ( 3.1 cal mol-1 K-1) conform to a I f B transition
structure in which the C(R)+‚‚‚C(1) interaction is suf-
ficiently elongated to allow sliding of the sec-butyl group
over the aromatic ring to establish C(R)+‚‚‚C(3), C(â)-
H‚‚‚F, and C(γ)-H‚‚‚F. The smaller activation enthalpy
(∆Hq ) 8.9 ( 0.6 kcal mol-1) and less favorable activation
entropy (∆Sq ) -8.4 ( 1.6 cal mol-1 K-1) measured for
the competing I f II process (ii) reflect, instead, slightly
tighter transition structures C and D in which the C(R)+
center of the sec-butyl moiety still partially interacts with
C(1) of fluorobenzene while moving over its C(2) and C(6)
sites, respectively.
This mechanistic picture shows an analogy close to that
proposed in our previous kinetic investigation on gas-
phase isomerization of protonated 1Me (henceforth de-
S
noted as IMe).23 In this study, only the para f meta
isomerization was observed with complete racemization
and partial H-scrambling in the moving sec-butyl moiety.
The activation parameters of the isomerization without
side-chain hydrogen scrambling (∆Hq ) 10.3 ( 1.2 kcal
mol-1 and ∆Sq ) -5.3 ( 3.6 cal mol-1 K-1) are compa-
rable with those measured for the analogous I f II and
I f IV reactions. This suggests that the two sets of
reactions proceed through similar transition structures
characterized by extensive C(R)+‚‚‚C(1) elongation. The
slightly smaller activation enthalpies and less favored
activation entropies measured for the I f II and I f IV
reactions likely reflect the effects of the second C(γ)-H‚‚‚F
interaction in the relevant transition structures (Figure
3) which is obviously absent in the reaction with IMe
.
The first consideration is the nature of the quasi-
degenerate intermediates A and B. The geometry of
complex A reflects specific C(R)‚‚‚C(1) (3.39 Å) and C(â)-
H‚‚‚C(4) (2.77 Å) electrostatic interactions. Simple sliding
of the sec-butyl group in A over the aromatic ring leads
to intermediate B. This complex presents C(γ)-H‚‚‚F (2.67
Å) and C(â)-H‚‚‚F (3.13 Å) hydrogen bonds coupled with
a C(R)‚‚‚C(3) interaction (3.26 Å) which is suitably
predisposed for σ-bonding to the more stable intermedi-
ate, IV (path (iv) in Figure 4). The arrangements of the
specific noncovalent forces in A and B do not allow us to
classify them into the well-established η1- and η2-
categories,45 and therefore, we coined the symbol η2′ to
denote the chelating interactions operating in these
noncovalent structures. In competition with the A f B
motion, the sec-butyl group in η2′-chelate A may slide over
the adjacent meta positions eventually yielding σ-inter-
mediate II. Since it takes place without any C(R)-C(â)
bond rotation, the A f II process may involve two
different although almost degenerate transition struc-
tures, i.e., C and D of Figure 3 (path (ii) in Figure 4).
Besides the A f B and A f II sliding, η2′ chelate A can
undergo C(R)-C(â) bond rotation through transition
structure E of Figure 3. It is interesting to note that the
formerly empty p orbital of the sec-butyl moiety in E
In addition to rotational transition structure E of
Figure 3, the racemization of the sec-butyl moiety of I
may simply involve a C(â) f C(R) hydrogen shift. The
exploration of the relevant PES led to the location of the
F transition structure for the 1,2-H shift (Figure 3) which
is thought to connect the stable complexes A and B′ (or
B′′). Compared to the I f II and I f IV processes (Table
3), both the competing I f III and I f V isomerizations
(paths (iii) and (v) of Figure 4, respectively) are charac-
terized by smaller activation enthalpies (∆Hq ) 6.9 ( 0.9
and 8.2 ( 1.0 kcal mol-1, respectively) and much less
favorable activation entropies (∆Sq ) -15.2 ( 1.7 and
-11.3 ( 3.1 cal mol-1 K-1, respectively). These findings
are consistent with the tight transition structure F,
wherein a C(â) f C(R) hydrogen shift participates in the
C(R)+‚‚‚C(1) bond cleavage in I. This sort of hydride ion
anchimeric assistance seems to be much less effective in
the isomerization of IMe 23
Indeed, the para f meta
.
isomerization of IMe involves activation parameters (∆Hq
) 16.9 ( 3.1 kcal mol-1 and ∆Sq ) +9.9 ( 7.0 cal mol-1
K-1) which suggest that the sec-butyl moiety undergoes
a H-shift only after the almost complete C(R)-C(1) bond
cleavage. This behavior may be explained by the hypoth-
esis that a significant amount of positive charge must
develop at the C(R) site of the sec-butyl moiety to promote
the 1,2-H shift. This requirement can easily be fulfilled
in I by the poor π-electron donating character of fluoro-
(44) Denhez, J. P.; Audier, H. E.; Berthomieu, D. Org. Mass
Spectrom. 1993, 28, 704.
(45) Smith, W. B. J. Phys. Org. Chem. 2002, 15, 347 and references
therein.
benzene (PA ) 180.7 kcal mol-1 33
development of a sufficient amount of positive charge on
)
which allows for the
4140 J. Org. Chem., Vol. 70, No. 10, 2005

Elusive Isomeric η-Complexes
the C(R) center of the sec-butyl moiety, even in the
presence of a substantial C(R)+‚‚‚C(1) interaction. The
same condition does not apply with IMe whose aromatic
moiety, i.e., toluene (PA ) 187.4 kcal mol-1),33 is a better
π-electron donor. In this case, an almost complete C(R)-
C(1) bond cleavage is necessary to promote relocation of
the positive charge from the ring to the C(R) center of
the sec-butyl moiety.
result of the extent of the one-way reaction pathways,
(ii)-(v), from A before the quenching of the formed
noncovalent transients into the corresponding σ-bonded
intermediates. The product distribution from the sec-
butylation of fluorobenzene reflects, instead, the inter-
conversion among all of the simultaneously populated η2′-
complexes, before the relevant σ-bonding.
This model suitably accounts for the dramatic pressure
dependence of the isomeric distribution of 2-arylbutanes
from gas-phase arene sec-butylation (Table 4) at any
given temperature. The limited extent of the proton-
Similar arguments apply to the I f 16 dealkylation
reaction. According to Figure 4, the energy required for
the unimolecular de-sec-butylation of I (23 kcal mol-1
)
largely exceeds the activation enthalpy measured for the
formation of fluorobenzene 16 from I (∆Hq ) 6.5 ( 1.2
kcal mol-1). This discrepancy, coupled with the negative
activation entropy value (∆Sq ) -12.0 ( 3.6 cal mol-1
K-1), excludes any conceivable unimolecular I fragmenta-
tion to 16 while supporting the direct formation of
fluorobenzene by the exothermic bimolecular â-elimina-
tion process induced by a (C2H5)3N attack at the side
chain of I (∆Helim ) -31 kcal mol-1).33 The same base-
induced elimination process on IMe involves a larger
activation energy (∆Hq ) 9.8 ( 0.9 kcal mol-1) and a more
favorable activation entropy (∆Sq ) -8.7 ( 2.7 cal mol-1
K-1). This difference can be explained by the greater
acidic character of the sec-butyl hydrogens of I, relative
to those of IMe, which places the relevant base-induced
elimination transition structure much earlier along the
reaction coordinate.
induced isomerization of 1F (Table 1) excludes facile
R
interconversion among the corresponding isomeric σ-
intermediates as a result of the observed pressure
dependence. Indeed, at 70 Torr and 40-100 °C, the
σ-bonded arenium ion I is found to be long-lived. Its
limited isomerization rate constants (ca. 1 × 105 s-1 (at
40 °C) and ca. 9 × 105 s-1 (at 100 °C), Table 1) indicate
that I undergoes from ca. 2000 (100 °C) to ca. 21400 (40
°C) collisions with CH4 before rearranging to the ortho
and meta isomers.41 Thus, at 70 Torr and T e 100 °C,
the I arenium ion lives long enough to be in thermal
equilibrium with the bulk gas before reacting. The same
conclusion is valid a fortiori when I and its regioisomers
are generated under the same conditions by direct sec-
butylation of fluorobenzene. Therefore, their relative
population should depend only on the temperature and
not on the bulk gas pressure. As a consequence, the large
pressure effect on the isomeric distribution of products
from gas-phase sec-butylation of fluorobenzene (Table 4)
can be explained only by acknowledging the intermediacy
of isomeric η2′-type chelates with lifetimes comparable
to their collision time with the bulk gas at 70-760 Torr
(2 × 10-9 ÷ 2 × 10-10 s).
Role of η2′-Chelates in Gas-Phase Arene Alkyl-
ations. From the above experimental and computational
results, it appears that the I isomerization is governed
by a relatively flat PES region located more than 13 kcal
mol-1 below the I f 16 + sC4H9+ and I f 16H+ + 2-C4H8
dissociation limits and more than 8 kcal mol-1 below
A&G’s loosely bound π-complex (Figure 4). This low-
energy region is characterized by the presence of several
shallow potential energy wells corresponding to η2′-
chelates A, B, B′, and B′′ separated from the correspond-
ing σ-bonded intermediates, I-V, by very limited energy
barriers.23,45,46 As suggested in a previous study,23 these
low-energy η2′-chelates may play a crucial role in gas-
phase aromatic alkylation reactions as well. The pro-
nounced difference in the ortho/meta ratios arising from
I isomerization (Table 1) and from the direct sec-butyl-
ation of fluorobenzene (Table 4), carried out under
comparable experimental conditions, excludes the pos-
sibility that the first process proceeds through the same
loosely bound π-complex proposed by A&G as an inter-
mediate of the latter reaction.19 Rather, such a difference
is due to the different phase space experienced in the
isomerization of I (or IMe) and in the gas-phase sec-
butylation of fluorobenzene (or toluene). The first process
necessarily proceeds through the initial exclusive forma-
tion of the corresponding A-type η2′-complexes (Figure 4),
whereas the latter involves the simultaneous (although
not necessarily equal) population of all of the accessible
isomeric η2′-chelates from A&G’s loosely bound π-com-
plex. The product pattern from the I isomerization is the
The existence of relatively stable η2′-chelates B, B′, and
B′′ on the I PES accounts for the exceptional ortho
selectivity exhibited by fluorobenzene in gas-phase
alkylations.24-26,38 At high gas pressures, when B under-
goes several cooling collisions with the bulk gas, it tends
to collapse preferentially to the ortho σ-bonded interme-
diate. At lower pressures, i.e., at lower cooling collision
frequencies, the η2′-complexes B-B′′ remain sufficiently
hot to promote entropy-driven motions over the fluoro-
benzene ring (e.g., through the E transition structure)
to allow the formation of other isomeric η2′-type com-
plexes (e.g., A) eventually collapsing to para and meta
σ-bound intermediates, the latter being the thermody-
namic sink of the entire process.
Acknowledgment. Financial support from the Uni-
versity of Rome “La Sapienza”, the University Cam-
erino, and the Italian National Research Council (CNR)
is gratefully acknowledged. The authors express their
gratitude to F. Grandinetti for his support in the
theoretical calculations.
Supporting Information Available: Computational de-
tails for all of the structures shown in Figure 3 and for the
ipso-protonated intermediates I-V of Scheme 3. This material
is available free of charge via the Internet at http://pubs.acs.org.
(46) Heidrich, D. Angew. Chem., Int. Ed. 2002, 41, 3208.
JO050019Q
J. Org. Chem, Vol. 70, No. 10, 2005 4141