Facial Selectivity of the (R)-1,3-Dimethyl-1-Cyclohexyl Cation in the Gas Phase

Article abstract of DOI:10.1002/chem.200304859

The model reaction between the (R)-1,3-dimethyl-1-cyclohexyl cation (I) and methanol has been investigated under gas-phase radiolytic conditions (750Torr; 25-120°C) with the aim of evaluating the intrinsic factors that govern the facial selectivity of biased carbocations. The peculiarity of the experimental approach allows the formation of different CH₃ ¹⁸OH·I ionic adducts. Subsequent conversion of these adducts to give the corresponding E/Z covalent products follows different reaction coordinates, which are characterized by their own activation parameters. On the grounds of density functional theory (DFT) results, several [CH ₃OH·I] structures have been located on the relevant potential-energy surface (PES). The experimental results point to a gas-phase facial selectivity, which is mainly governed by entropic factors that arise as a result of the occurrence of different noncovalent ion-molecule facial adducts (FA). The formation of FAs may also play an important role in both the reaction dynamics and the positional selectivity. The present results cannot be interpreted by any of the models based on solution-phase experiments.

Full text of DOI:10.1002/chem.200304859

FULL PAPER
Facial Selectivity of the (R)-1,3-Dimethyl-1-Cyclohexyl Cation
in the Gas Phase
Antonello Filippi*[a]
Abstract: The model reaction between
the (R)-1,3-dimethyl-1-cyclohexyl cation
(I) and methanol has been investigated
under gas-phase radiolytic conditions
(750 Torr;25 120 8C) with the aim of
evaluating the intrinsic factors that gov-
ern the facial selectivity of biased car-
bocations. The peculiarity of the exper-
imental approach allows the formation
of different CH318OH ¥ I ionic adducts.
Subsequent conversion of these adducts
to give the corresponding E/Z covalent
products follows different reaction co-
ordinates, which are characterized by
their own activation parameters. On the
grounds of density functional theory
(DFT) results, several [CH3OH ¥ I] struc-
tures have been located on the relevant
potential-energy surface (PES). The ex-
perimental results point to a gas-phase
facial selectivity, which is mainly gov-
erned by entropic factors that arise as a
result of the occurrence of different
noncovalent ion molecule ™facial ad-
ducts∫ (FA). The formation of FAs may
also play an important role in both the
reaction dynamics and the positional
selectivity. The present results cannot be
interpreted by any of the models based
on solution-phase experiments.
Keywords: diastereoselectivity
¥
facial selectivity gas-phase
¥
reactions ¥ ion molecule reactions ¥
kinetics ¥ noncovalent interactions
solvent. However, a completely different picture emerges
from the gas-phase investigation of 2-methyl-5-X-adamantyl
cations (X ˆ F, Si(CH3)3).[3] In this study, which was performed
by means of the radiolytic technique, the gas-phase facial
selectivity of the selected ions towards methanol was consid-
ered to be the result of entropy rather than enthalpy, and the
negative activation entropies obtained could be attributed to
the increased rigidity of the relevant transition structures. We
used the same gas-phase radiolytic approach to evaluate the
kinetics of the addition of methanol to a flexible cation,
namely, the (R)-1,3-dimethyl-1-cyclohexyl cation (I). It is
hoped that the results of this model reaction will provide
greater insight into the role that intrinsic factors play in
determining the facial selectivity of biased and conforma-
tionally mobile cyclohexyl cations, a class of reactive inter-
mediates recently investigated both experimentally and
theoretically.[4]
Introduction
The growing use of optically pure chiral compounds in many
areas of chemistry and biochemistry has recently sparked an
intense effort to understand the facial selectivity of planar
carbon centers. Over the years, a number of factors have been
proposed as having an influence on the facial selectivity of
reactive species towards trigonal carbon centers;[1] these
include steric hindrance, conformational and chelation effects,
and ion pairing, as well as orbital and electronic interactions.
Solvation and temperature effects are also important. Assess-
ing the relative roles of these factors is not easy. Reactions of
different 2-methyl-5-X-2-adamantyl cations with a number of
nucleophiles in condensed media have been investigated by
several research groups.[2] All the studies indicate that the
facial selectivity of these rigid cations is due to the differential
hyperconjugative stabilization of the competing syn and anti
transition states. Irrespective of its electrostatic or orbital
origin, such an intrinsic factor should play a major role in the
gas phase, owing to the lack of the mediating effects of the
Results
Gas-phase protonation of the substrate: The radiolytic
technique allowed the gas-phase protonation of 1 to be
carried out in two ways. In CH4/1/CH318OH mixtures, g-
radiolysis of the bulk gas CH4 generates stationary concen-
[a] Dr. A. Filippi
Dipartimento di Studi di Chimica e Tecnologia delle Sostanze
Biologicamente Attive Universita degli Studi di Roma
¡
‡
trations of CnH5 (n ˆ 1,2) ions. These strong Br˘nsted acids
™La Sapienza∫ P.le A. Moro5, 00185Roma (Italy)
Fax : (‡39)06-49-91-36-02
are able to protonate 1[5] (path i in Scheme 1) yielding the
chiral (R)-1,3-dimethyl-1-cyclohexyl cation (I) in the presence
of small amounts of nucleophile (CH318OH, 18O ˆ 94%), an
E-mail: antonello.filippi@uniroma1.it
Supporting information for this article is available on the WWW under
http://www.chemeurj.org/ or from the author.
efficient radical scavenger (O2), and
a powerful base
5396
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304859
Chem. Eur. J. 2003, 9, 5396 5403

5396 5403
plex route), the products
(1R,3R)-1-methoxy-1,3-dimeth-
ylcyclohexane
(3E)
and
(1S,3R)-1-methoxy-1,3-dimeth-
ylcyclohexane (3Z) arise when
external CH318OH molecules
present in the gaseous mixture
react with preformed I to firstly
give the encounter complexes
IIsyn and IIanti. On the other
hand, in route ii (henceforth
called the intracomplex route),
the same products arise from
the intimate IIsyn and IIanti com-
plexes characterized by an ini-
tial specific orientation be-
tween the incipient ion I and
the incipient CH318OH mole-
cule, which is internally gener-
ated by proton transfer from
‡
CH318OH2 to 1. The high pres-
sure of the inert gaseous me-
dium (CH4 or CH3F;750 Torr)
ensures effective collisional
thermalization
collisionssÀ1)[8] of the ion mo-
lecule complexes before their
Scheme 1. Reaction pathways for the extracomplex (path i), intracomplex (path ii), and epimerization (path iii)
experiments.
(7 50 Â 109
[(N(C2H5)3)]. The base ensures that the oxonium-ion inter-
mediates are efficiently deprotonated (kb and k'b in
Scheme 1). In CH3F/1/H218O mixtures, g-radiolysis of CH3F
conversion to IIIZ and IIIE (vide infra).
To evaluate the extent of oxonium-ion epimerization
(IIIE > IIIZ, kE!Z and kZ!E in Scheme 1) prior to neutral-
ization, a third set of experiments was performed with
(1R,3R)-1,3-dimethylcyclohexanol (2E) or (1S,3R)-1,3-dime-
thylcyclohexanol (2Z) (enantiomeric excess (ee) ca. 99%) as
the starting substrates, CH3Cl as the bulk gas, and traces of
H218O, O2, and N(C2H5)3. g-Irradiation of such gaseous
mixtures generates stationary concentrations of the Lewis
‡
generates stationary concentrations of (CH3)2F , whose
‡
reaction with H218O yields CH318OH2 ions, in the absence
of any neutral CH318OH molecules.[6] The so formed
‡
CH318OH2 then acts both as a Br˘nsted acid catalyst[7]
(path ii in Scheme 1) and as a nucleophile (CH318OH)
generator. The substantial difference between the two ap-
proaches is that, in route i (henceforth called the extracom-
acid (CH3)2Cl , which barely reacts with H218O,[9] but which is
‡
strong enough to methylate the alcoholic chiral substrates to
yield the corresponding oxonium intermediates IIIE and IIIZ
(path iii in Scheme 1).[10] All the irradiations were performed
at a constant temperature in the range 25 1208C.
Abstract in Italian: Al fine di valutare il ruolo dei fattori
¡
intrinseci che determinano la selettivita facciale di carbocationi
non simmetrici, la cinetica della reazione fra il metanolo e il
¡
catione chirale (R)-1,3-dimetil-1-cicloesile (I) e stata studiata in
Radiolytic experiments: The upper part of Table 1 reports the
relative distribution and the total absolute yield of the 18O-
labeled ethereal products (3E and 3Z) obtained from the g-
irradiated CH4/1/CH318OH gaseous mixtures (the extracom-
plex reaction). The product yields from the CH3F/1/H218O
mixtures (the intracomplex reaction) are reported in the
lower part of Table 1. Over 95% of the 18O incorporated into
3E and 3Z occurs through the extracomplex reaction;this is
in agreement with the isotopic composition of the nucleophile
(CH318OH, 18O ˆ 94%). For the same products in the intra-
complex reaction (CH3F/1/H218O systems), this percentage
decreases to 50%, due to the presence of ubiquitous H216O[11]
in the gaseous mixtures. However, the yield of both labeled or
unlabeled 3E and 3Z decreases substantially (by over 80%)
when the molar fraction of the strong base N(C2H5)3 is
quintupled in the starting mixtures;this is evidence that the
ethereal products in Table 1 are ionic in origin.
fase gassosa utilizzando la tecnica della radiolisi (750 torr; 25
1208C). La peculiarita del metodo impiegato consente la
formazione di diversi addotti CH3 OH I, i quali evolvono
¡
18
À
verso i corrispondenti prodotti covalenti E/Z seguendo diverse
coordinate di reazione, ciascuna caratterizzata da parametri di
attivazione specifici. Alcune strutture [CH3OH ¥ I] sono state
localizzate mediante uno studio teorico (DFT) della superficie
di energia potenziale. I risultati sperimentali evidenziano che,
in fase gassosa, la selettivita facciale e determinata principal-
mente da fattori entropici, i quali riflettono la dinamica di
reazione di diversi ™addotti facciali∫ ione-molecola non
covalenti (FA). La loro formazione puo avere un ruolo
importante nel determinare sia la dinamica delle reazioni sia
la selettivita posizionale. I risultati ottenuti non possono essere
razionalizzati utilizzando i modelli elaborati sulla base di studi
condotti in fase condensata.
¡
¡
¡
¡
Chem. Eur. J. 2003, 9, 5396 5403
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5397

FULL PAPER
n
o
o
eeq
t
eeq
Table 1. Relative distribution of the 18O-labeled ethereal products from
attack of methanol on I in the gas phase.[a]
kZ!E
ˆ
ln
(1a)
(1b)
(1c)
(1d)
eeq À e
T
CH318OH
[8C] [Torr]
B
t[c]
Keq
eeq
ˆ
[Torr] [Â108 s] 3E[d]
3Z[d]
G(M)
[e]
1 ‡ Keq
path i[b]
n
zeq
zeq
kE!Z
ˆ
ln
25
40
40
40
50
60
60
70
80
1.05
1.06
1.17
0.53
0.92
1.04
0.97
1.14
1.03
1.03
0.56
1.12
0.70
0.67
0.69
0.59
0.71
0.71
0.71
0.75
0.69
0.70
0.79
0.75
3.61
4.01
3.88
4.49
3.89
4.02
4.02
3.94
4.41
4.36
3.34
4.28
0.463 0.537 0.99
0.485 0.515 0.92
0.473 0.527 0.67
0.482 0.518 0.52
0.510 0.490 0.60
0.502 0.498 0.53
0.477 0.523 0.84
0.510 0.490 0.48
0.509 0.491 0.59
0.487 0.513 0.37
0.511 0.489 0.43
0.498 0.502 0.83
t
zeq À z
1
zeq
ˆ
1 ‡ Keq
whereby e and z indicate the extent of epimerization of IIIZ
and IIIE, respectively, during their lifetime t. These epimer-
ization factors are evaluated as the molar fraction of
unlabeled ethers, 3E and 3Z, obtained from the chiral
alcohols 2Z and 2E, respectively.
At any given temperature, the IIIZ > IIIE equilibrium
constant (Keq) can be calculated from Equation (2). The
results are given in Table 2.
80
100
100
path ii[f]
25
40
50
50
60
70
70
80
100
3.10
2.89
3.02
3.02
3.12
3.01
3.13
2.90
2.97
0.70
0.71
0.72
0.73
0.71
0.68
0.68
0.65
0.73
3.61
3.76
3.81
3.77
4.02
4.31
4.32
4.69
4.43
0.450 0.550 0.06
0.436 0.564 0.11
0.424 0.576 0.12
0.419 0.581 0.14
0.420 0.580 0.13
0.400 0.600 0.07
0.402 0.598 0.10
0.412 0.588 0.11
0.409 0.591 0.08
e
Keq
ˆ
(2)
z
Table 2. Kinetics of the gas-phase epimerization of IIIE and IIIZ.[a]
[d]
Substrate
T
B
t[b]
kE!Z
[a] Substrate 1 (0.5 Torr), O2 (5 Torr), B ˆ N(C2H5)3; radiation dose 1 Â
104 Gy (dose rate ˆ 5  103 GyhÀ1). [b] The extracomplex pathway;bulk
gas CH4 (750 Torr). [c] Reaction time, calculated as the reciprocal of the
pseudo-first-order collision constant between intermediates IIIE and IIIZ
and B (ref. [8]). [d] Mean values from repeated experiments;uncertainty
level ca. 5%. [e] G(M) ˆ number of molecules M produced per 100 eV of
absorbed energy. [f] The intracomplex pathway;bulk gas CH 3F (750 Torr).
[d]
[8C] [Torr] [Â108 s] z[c]
[Â10À6 sÀ1
]
Keq
2E
2E
2E
2E
2E
2E
2E
2E
25
40
50
60
70
80
100
120
0.74
0.71
0.71
0.69
0.71
0.66
0.64
0.63
3.40
3.75
3.88
4.14
4.11
4.57
5.00
5.38
0.047
0.060
0.106
0.135
0.168
0.238
0.281
1.44 (6.16) 0.58 (À0.24)
1.68 (6.23) 0.62 (À0.21)
3.01 (6.48) 0.67 (À0.17)
3.69 (6.57) 0.67 (À0.17)
4.87 (6.69) 0.80 (À0.09)
6.81 (6.83) 0.80 (À0.10)
8.23 (6.92) 0.99 (0.00)
In the CH4/1/CH318OH experiments, besides the ethereal
products, appreciable amounts of unlabeled 2E and 2Z, as
well as (R)-1,3-dimethylcyclohexene and (R)-1,5-dimethylcy-
clohexene are formed. In the CH3F/1/H218O mixtures, the
ethereal products are accompanied by 18O-labeled 2E and 2Z
in yields that are comparable to those obtained for the
unlabeled analogues. The alcoholic products 2E and 2Z are
thought to arise from attack of the free I ion by labeled and/or
unlabeled[11] water molecules followed by N(C2H5)3-neutral-
ization of the oxonium intermediates so formed. Direct
N(C2H5)3-deprotonation of I is the most probable route to
the formation of (R)-1,3-dimethylcyclohexene and (R)-1,5-
dimethylcyclohexene. The fact that the diastereomeric 18O-
labeled products are not accompanied by their enantiomers
indicates that the tertiary cation I does not undergo extensive
prototropic rearrangement. Therefore, the neutral 18O-la-
beled products 3E and 3Z do indeed arise from the reactions
depicted in Scheme 1. If the efficiency of the deprotonation
steps (kb and k'b) depicted in Scheme 1 are taken to be equal,
the relative quantities of neutral 18O-labeled products 3E and
3Z reflect those of the corresponding oxonium intermediates
IIIE and IIIZ.[12] The ksyn/kanti ratio can then be inferred from
the relative concentration of IIIE and IIIZ, once the extent of
their interconversion, under the same experimental condi-
tions, is considered. To this end, intermediates IIIE and IIIZ
were generated by path iii in Scheme 1, and their epimeriza-
tion constants (kE!Z and kZ!E) were calculated using the
Equations (1a) (1d):[3]
0.464 17.07 (7.23) 0.70 (À0.15)
Substrate
T
B
t[b]
kZ!E
[d]
[d]
[8C] [Torr] [Â108 s] e[c]
[Â10À6 sÀ1
]
Keq
2Z
2Z
2Z
2Z
2Z
2Z
2Z
2Z
25
40
50
60
70
80
100
120
0.75
0.74
0.71
0.67
0.71
0.67
0.65
0.65
3.37
3.61
3.88
4.24
4.11
4.51
4.94
5.23
0.027
0.037
0.071
0.091
0.135
0.191
0.279
0.83 (5.92) 0.58 (À0.24)
1.08 (6.03) 0.62 (À0.21)
2.01 (6.30) 0.67 (À0.17)
2.43 (6.39) 0.67 (À0.17)
3.92 (6.59) 0.80 (À0.09)
5.53 (6.74) 0.80 (À0.10)
8.29 (6.92) 0.99 (0.00)
0.327 12.37 (7.09) 0.70 (À0.15)
[a] Substrate (0.5 Torr), bulk gas CH3Cl (750 Torr), O2 (5 Torr), H218O (ca.
3 Torr), B ˆ N(C2H5)3;radiation dose
1
 104 Gy (dose rate ˆ 5 Â
103 GyhÀ1). [b] Reaction time, calculated as the reciprocal of the pseudo-
first-order collision constant between intermediates IIIE and IIIZ and B
(ref. [8]). [c] z ˆ [3Z]/([3E] ‡ [3Z]) and e ˆ [3E]/([3E] ‡ [3Z]) are mean
values from repeated experiments;uncertainty level ca. 5%; 18O-ethers
< 3%. [d] See text for
parentheses.
k and Keq definitions;log k and logKeq in
Linear correlations were found between logkE!Z (and
logkZ!E) and TÀ1. A similar correlation was also obtained for
logKeq (logKeq ˆ (0.5 Æ 0.3) À (0.9 Æ 0.4)1000/(2.303RT); r2 ˆ
0.533), as depicted in Figure 1.
Table 3 gives the relevant Arrhenius equations and the
corresponding activation parameters calculated according to
the transition-state theory. Activation enthalpies indicate that
IIIZ is more stable than IIIE by 0.9 kcalmolÀ1.
The Arrhenius equations in Table 3 allow the e and z terms
for each t value in Table 1 to be evaluated. In turn, this data,
5398
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5396 5403

Ion Molecule Reactions
5396 5403
the slope and y axis intercept of the straight lines obtained by
linear regression analysis (Table 4). Surprisingly, opposite
differential activation parameters are observed. The entropic
contributions are large enough to invert the enthalpic facial
selectivity in both the extra- (T> 438C) and the intracomplex
(T> 58C) reaction.
Table 4. Differential Arrhenius equations and relevant differential activation
parameters for extracomplex (i) and intracomplex (ii) attack of methanol on I
in the gas phase.
Path Arrhenius equation[a]
corr. coeff. DDH=
r2 [kcalmolÀ1
DDS=
[calmolÀ1 KÀ1
]
]
i
log(ksyn/kanti) ˆ (À0.7 Æ 0.1) 0.773
À (À1.0 Æ 0.2)x
À 1.0 Æ 0.2
1.2 Æ 0.1
À 3.1 Æ 0.5
4.4 Æ 0.4
ii
log(ksyn/kanti) ˆ (1.0 Æ 0.1)
À (1.2 Æ 0.1)x
0.933
Figure 1. Temperature dependence of logkE!Z (full circles), logkZ!E (open
circles), and logKeq (full triangles) for the epimerization reaction IIIE >
IIIZ.
[a] x ˆ 1000/2.303RT.
Table 3. Kinetics of the epimerization of IIIZ and IIIE in the gas phase:
Arrhenius equations and relevant activation parameters.
Theoretical calculations: A comprehensive and high-level ab
initio investigation of the PES of the reaction network
depicted in Scheme 1 is practically inaccessible either intrinsi-
cally, because of the conformational mobility of the system,
and computationally, because of the need for an extended
basis set. However, several critical points have been located at
the B3LYP/6 31G* level of theory;[13] their energies are
reported in Table 5, while the relevant optimized structures
are shown in Figure 3. It is worth noting that the structural
Arrhenius equation[a]
corr. Coeff DH=
r2 [kcalmolÀ1
DS=
]
[calmolÀ1 KÀ1
]
logkE!Z ˆ (10.6 Æ 0.3) À (6.1 Æ 0.4)x 0.974
logkZ!E ˆ (11.0 Æ 0.2) À (7.0 Æ 0.4)x 0.985
5.5 Æ 0.4
6.4 Æ 0.4
À 12.1 Æ 1.2
À 9.9 Æ 1.1
[a] x ˆ 1000/2.303RT.
along with the molar concentrations of 3E and 3Z given in
Table 2, have been used to calculate the ksyn/kanti ratio by using
Equation (3).[3]
Table 5. B3LYP/6-31G*//B3LYP/6-31G* results.
E[a]
[Hartree]
ZPE[b]
[Hartree]
E ‡ ZPE
E ‡ ZPE
ksyn
kanti
‰3 ZŠ À z
‰3 EŠ À e
[kcalmolÀ1
]
ˆ
(3)
[Hartree]
CH3OH
CH3OH2
1eq
1ax
Ieq
À 115.714406
À 116.015745
À 313.278724
À 313.276204
À 313.620403
À 313.616850
À 429.364812
À 429.364040
À 429.353500
À 429.353348
À 429.352705
À 429.350565
À 429.334810
À 429.294469
0.051458
0.064381
0.203706
0.203943
0.214103
0.214188
0.271732
0.271748
0.267317
0.267151
0.267255
0.266908
0.265561
0.268087
À 115.662948
À 115.951364
À 313.075018
À 313.072261
À 313.406300
À 313.402662
À 429.093080
À 429.092292
À 429.086183
À 429.086197
À 429.085450
À 429.083657
À 429.069249
À 429.026382
The results for both the extracomplex and intracomplex
reactions are plotted as log(ksyn/kanti) against TÀ1 in Figure 2.
‡
[c]
0.0
1.7
0.0
2.3
0.0
0.5
4.3
4.3
4.8
Iax
IIIE
IIIZ
IIanti
IIasyn
IIbsyn
[d]
TSinv
5.9
15.0
41.8
Ieq ‡ CH3OH
‡
1eq ‡ MeOH2
[a] Total energy. [b] Zero-point energy. [c] Calculated B3LYP/6-31G* proton
affinity: 208.9 kcalmolÀ1. [d] Imaginary frequency: i64.92 cmÀ1
.
À
deformation of both Ieq and Iax is similar to that of the C C
hyperconjomer of the 1-methylcyclohexyl cation.[4a] In partic-
Figure 2. Differential Arrhenius plots (log(ksyn/kanti) vs TÀ1) for intra-
complex (full circles) and extracomplex (open circles) attack of ion I by
methanol in the gas phase.
À
À
ular, there is: 1) bond elongation of C2 C3 and C5 C6;2)
À
C H bond elongation in the methyl group, which eclipses the
À
À
empty p orbital;3) C1 C2 C3 angle contraction;and 4) slight
equatorial pyramidalization (Ieq ˆ 172.398; Iax ˆ 173.378) of
the charged C1 atom. Also, at least at the B3LYP/6 31G*
The relevant differential Arrhenius equations and the
corresponding activation parameters were calculated from
À
level of theory, a C H hyperconjomer of I (the carbocation
Chem. Eur. J. 2003, 9, 5396 5403
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5399

FULL PAPER
while diastereomer IIIZ is
0.5 kcalmolÀ1 higher in energy.
Epimerization of the two spe-
cies involves the transition
structure TSinv (5.9 kcalmolÀ1
higher than IIIE). Taking into
account the uncertainty of
the
method
used
(ca.
2 kcalmolÀ1),[15] the theoretical
results correspond closely to
those obtained experimen-
tally
6.4 kcalmolÀ1
(Table 3:
DH=
DH8(IIIZ >
ˆ
Z!E
;
IIIE) ˆ 0.9 kcalmolÀ1). Only a
single IIanti structure (4.3
kcalmolÀ1) has been located,
but two different syn adducts
have been found; IIasyn
(4.3 kcalmolÀ1), which has a
À
single methyl C H interaction,
and the less stable IIbsyn
(4.8 kcalmolÀ1), in which the
oxygen atom is coordinated to
both the axial C2 and C6 hydro-
gen atoms. At the same level of
theory, the proton affinity of
3-methyl-1-methylenecyclohex-
ane (208.9 kcalmolÀ1) has also
been calculated.
Discussion
The addition reaction: In the
intracomplex reaction, enthal-
py favors anti rather than syn
attack of
I
by methanol
(1.2 kcalmolÀ1) (Table 4). How-
ever, as a result of entropy
(4.4 calmolÀ1 KÀ1), at T> 58C
syn attack predominates.
A
completely different result aris-
es in the extracomplex expe-
riments: although enthalpy
(À1.0 kcalmolÀ1) favours syn
Figure 3. Optimized structures and relevant geometrical parameters of the critical points located on the
[CH3OH ¥ I] PES at the B3LYP/6 31G* level of theory. Bond lengths are given in ä, angles are italicized and
given in degrees.
attack,
entropy
(À3.1
calmolÀ1 KÀ1) reverses the se-
lectivity at T> 438C. Such divergent results obtained for the
same reaction (Figure 2) cannot be interpreted in terms of any
of the theories so far proposed to explain the facial selectivity
of trigonal carbon centers,[1, 16, 17] but must be related to the
different behavior of the reactive intermediates in the extra-
and intracomplex reactions.
As reported in the previous section, the vastly different
proton affinities of the involved species ensures that ion
molecule adducts like IIsyn and IIanti (Scheme 1) are stable
intermediates in both the extra- and intracomplex reaction.
The hypothesis that IIsyn and IIanti reach a rapid equilibrium
prior to being converted into IIIZ and IIIE is not supported
À
À
axially oriented by C2 Hax and C6 Hax hyperconjugation)
does not exist. This could be due to the fact that I contains just
one 3-methyl group.[4b]
According to the E‡ZPE column of Table 5, the 1eq and Ieq
chair structures, in which the methyl group is equatorial, are
1.7 and 2.3 kcalmolÀ1 more stable than the corresponding
chair conformers, 1ax and Iax, respectively. This is in agreement
with the experimental value obtained for the conformational
energy of the methyl group (1.9 2.1 kcalmolÀ1).[4b, 14] There-
fore, 1eq and Ieq are, by far, the most abundant conformers of 1
and I. The oxonium ion IIIE has the lowest E‡ZPE value on
À1
À
the [I ¥ CH3OH] PES (C1 O binding energy: 15.0 kcalmol ),
5400
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5396 5403

Ion Molecule Reactions
5396 5403
by the present results. In fact, the plots in Figure 2 clearly
indicate that in the extra- and intracomplex attack of I by
methanol, facial selectivity has a different temperature
dependence, which is the result of opposite DDH= and DDS=
contributions to DDG=. The alternative hypothesis that the
IIsyn and IIanti intermediates react through the same species in
both the extra- and intracomplex reaction is also not
supported by the current results. In fact, because different
reaction partners are involved (steps i and ii in Scheme 1), it is
likely that the relative populations of IIsyn and IIanti formed are
not the same, contributing to the difference in DDS= values
seen in Table 4. Moreover, the presence of the same IIsyn and
IIanti reactive structures in the two processes should lead to the
same DDH= values, in contrast to the experimental evidence.
From this analysis, the experimental results clearly indicate
that: 1) along the extra- and intracomplex reaction pathways,
the IIsyn and/or IIanti structures are different and their relative
abundance depends on the formation process (i or ii in
Scheme 1);and 2) interconversion among different IIsyn
structures must be slow relative to their conversion into IIIZ.
The same holds true for the IIanti adducts.
one considers, in the absence of solvent-mediating properties,
that any type of interaction is better than none, the existence
of FAs in gas-phase ion-neutral reactions should be rather
common. In fact, when an ion and a neutral molecule
approach each other, more than one reciprocal stabilizing
interaction can be sought. FAs that interconvert quickly will
not be kinetically distinguishable, but if their conversion to
the corresponding s-bonded intermediates is faster than their
interconversion, they will dictate the reaction troposelectiv-
ity.[19] Indeed, in the latter case, the relative populations of
different FAs that operate along parallel reaction coordinates
will reflect the positional selectivity.[18] However, if converging
reaction coordinates are involved, these will contribute to the
overall kinetic constant (as in the syn extracomplex reaction:
IIasyn ! IIIZ IIbsyn).
As similar investigations in solution have not been con-
ducted, a direct comparison of the present gas-phase results
cannot be made. Facial selectivity observed in condensed
media has often been interpreted on the grounds of intrinsic
stereoelectronic factors, which favor anti attack by nucleo-
philic reagents of both charged and neutral unsaturated
carbon atoms of methylenecyclohexane derivatives and their
carbonyl analogues.[1, 16, 17] This enthalpic approach assumes
that entropic factors in the competing syn and anti processes
are compensating or negligible, which can be accepted only in
low-temperate investigations. The present investigation, along
with a previous report in which the rigid 2-methyl-5-X-2-
adamantyl cation (X ˆ F, Si(CH3)3) was used, suggests that
facial selectivity of charged trigonal carbon centers in the gas
phase is strongly influenced by entropic rather than enthalpic
factors. A comprehensive analysis of the geometric parame-
ters for the structures in Figure 3 indicates that hyperconju-
gation may play a role, but, above all, it suggests that facial
selectivity in the gas phase is strongly dependent upon the
intrinsic ability of both the ionic substrate and the nucleophile
to make different ion molecule adducts, and on their
evolution dynamics. In the condensed phase, the solvent
cage not only modifies the intrinsic hyperconjugative features
of ions like I,[4b] but it also mediates the electrostatic
interaction between the ion and the neutral nucleophile.
Thus, the pronounced solvent effect on the facial selectivity of
biased substrates in the condensed phase[1] may be primarily
ascribed to the polarizability and dielectric properties of the
solvent, and to differential face solvation.[4b]
The occurrence of different ™facial adducts∫ (FAs) between
methanol and I, that is, noncovalent ion molecule adducts in
which the neutral molecule is coordinated to different points
on the same face of the ion, is not unexpected. Indeed, proton
‡
transfer from CH318OH2 to 1 (path ii in Scheme 1) is thought
to lead to the IIasyn and IIanti adducts depicted in Figure 3, in
which the CH318OH molecule is primarily coordinated to a
methyl hydrogen atom of Ieq. In contrast, more than one
center of interaction on both faces of I is kinetically accessible
to CH318OH (path i in Scheme 1). The theoretical results
(Table 5 and Figure 3) support the above view, and suggest
that a single adduct (IIasyn) is generated in the syn intra-
complex reaction, while two different FAs (IIasyn and IIbsyn
)
are involved in the syn extracomplex process. On the other
hand, the IIanti adduct is favored in both extra- and intra-
complex anti attack.
Facial selectivity: Conversion of II to the corresponding
oxonium ions IIIE and IIIZ must involve rather small
activation barriers. This point is supported by: 1) the relatively
low experimental activation enthalpy for IIIE > IIIZ epimer-
ization (6.4 kcalmolÀ1), 2) the poor dependence of the
absolute yields of the ethereal products 3E and 3Z upon
temperature (G(M) in Table 1), and 3) the computational
evidence. Moreover, since the theoretical results indicate that
an identical IIanti adduct is involved in the anti attack of I by
methanol in both the extra- and intracomplex reaction, the
DDH= values in Table 4 suggest that DH= for the syn
extracomplex reaction (IIbsyn ! IIIZ) is approximately
2.2 kcalmolÀ1 lower in energy than DH= for the syn intra-
complex reaction (IIasyn ! IIIZ). It should be noted, that this
value is 4.4 times greater than the theoretical IIbsyn À IIasyn
relative energy (Table 5: 4.8 À 4.3 ˆ 0.5 kcalmolÀ1);this sug-
gests that the syn transition-state structures involved in the
conversion of IIbsyn and IIasyn to IIIZ are different.
Conclusion
The kinetics of the reaction between methanol and the (R)-
1,3-dimethyl-1-cyclohexyl cation (I) have been investigated in
the gas phase using the radiolytic technique at a pressure of
750 Torr and temperatures ranging from 25 to 1208C. The
results have revealed that entropy contributes to the free
activation energy of the competing syn and anti processes, and
plays a major role in determining the facial selectivity by
increasing the temperature. The key role of a number of FAs,
as suggested by experimental differential activation parame-
ters, is further supported by DFT calculations. The combined
results indicate that gas-phase facial selectivity of biased
The occurrence of more than one FA has already been
reported by Speranza et al. in their paper on the isomerization
of protonated para-sec-butyltoluenes in the gas phase.[18] If
Chem. Eur. J. 2003, 9, 5396 5403
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5401

FULL PAPER
The extent of 18O incorporation into the radiolytic products was deter-
mined by GLC MS, whereby the mass analyzer was set on the selected-ion
carbocations is mainly dependent upon the intrinsic aptitude
of both the ionic substrate and the neutral nucleophile to form
different FAs, and upon the reaction dynamics whereby these
FAs interconvert or form the products. FAs are common
intermediates in gas-phase ion chemistry, while their occur-
rence in condensed media may be mediated by solvation.
‡
mode (SIM). The ion fragments at m/z ˆ 127 (16O-[M À CH3] ) and m/z ˆ
‡
129 (18O-[M À CH3] ) were monitored in order to analyze the 3E and 3Z
ethers. The corresponding alcohols 2E and 2Z were examined by using the
‡
‡
fragments at m/z ˆ 113 (16O-[M À CH3] ) and m/z ˆ 115 (18O-[M À CH3] ).
‡
The ion fragment at m/z ˆ 110 ([M] ) was monitored in order to analyze
substrate 1 and any possible isomers.
Computational details: Quantum-chemical calculations were performed by
using the GAUSSIAN 98 set of programs.[21] The geometry of the
investigated species was optimized at the B3LYP/6 31G* level of
theory.[13] The structures corresponding to the critical points were
submitted to frequency calculations, at the same level of theory, in order
to ascertain their real-minimum or transition-structure character on the
PES, and to evaluate the corresponding ZPE.[22]
Experimental Section
Materials: Methane, methyl fluoride, methyl chloride, and oxygen were
high-purity gases, which were purchased from UCAR Specialty Gases N.V.,
and used without further purification. H218O (18O ˆ 94.6%) and CH318OH
(18O ˆ 94%) were purchased from ICON Services. Research grade
N(C2H5)3 and (R)-3-methylcyclohexanone (98%, ee ˆ 99%) were supplied
by Aldrich.
Acknowledgements
¡
This work was supported by the Ministero dell×Istruzione, dell×Universita e
Synthesis of the substrates: (R)-3-Methyl-1-methylenecyclohexane (1) was
synthesized from (R)-3-methylcyclohexanone (98%, ee ˆ 99%) by using
Corey×s procedure[20] for the Wittig reaction. The same starting chiral
ketone was treated with CH3MgBr in dry diethyl ether to give, after
hydrolysis, both (1R,3R)-1,3-dimethylcyclohexanol (2E) and (1S,3R)-1,3-
dimethylcyclohexanol (2Z). Each diastereomer was initially assigned on
the basis of the product ratio (E/Z ˆ 68:32) reported in the literature,[16] and
was further confirmed by NMR spectroscopic analysis of the corresponding
methyl ethers (see below). NaNH2 and subsequently CH3I were added to a
stirred sample of the latter crude reaction mixture to afford the
diastereomers (1R,3R)-1-methoxy-1,3-dimethylcyclohexane (3E) and
(1S,3R)-1-methoxy-1,3-dimethylcyclohexane (3Z). All the synthesized
compounds were purified by preparative gas liquid chromatography
(GLC) on either a 5 m  4 mm [inside diameter (i.d.)] stainless steel
column packed with 10% Carbowax 20M on 80 100 mesh Chromosorb
WAW at 1108C (Chrompack), or on a 3 m  4 mm (i.d.) stainless steel
column packed with 10% OV-17 on 80 100 mesh Chromosorb WAW at
808C (Chrompack). The chemical (>99.9%) and optical (ee ca. 99%)
purity of 1, 2E, 2Z, 3E, and 3Z was verified by analytical GLC by using the
same chiral columns that were used for the analysis of the g-irradiated
gaseous mixtures (MEGADEX DACTBS-b (30% 2,3-di-O-acetyl-6-O-
tert-butyldimethylsilyl-b-cyclodextrin in OV 1701, 25 m long, 0.25 mm i.d.,
df ˆ 0.25 mm), 40 < T< 1708C, 28CminÀ1;CP-Chirasil-DEX CB, 25 m long,
0.25 mm i.d., df ˆ 0.25 mm), 40 < T< 1808C, 58CminÀ1). A sample of the
della Ricerca (MIUR) and the Consiglio Nazionale delle Ricerche (CNR).
I wish to express my gratitude to Maurizio Speranza for helpful discussions.
¬
I would also like to thank Luisa Mannina, Stephane Viel, and Annalaura
Segre for NMR spectroscopic analysis.
[1] Special Thematic Issue on Diastereoselection; Chem. Rev. 1999, 95,
whole issue.
[2] a) C. K. Cheung, L. T. Tseng, M. H. Lin, S. Srivastava, W. J. le Noble,
J. Am. Chem. Soc. 1986, 108, 1598 1605;corrigendum: C. K. Cheung,
L. T. Tseng, M. H. Lin, S. Srivastava, W. J. le Noble, J. Am. Chem. Soc.
1987, 109, 7239;b) S. Srivastava, W. J. le Noble, J. Am. Chem. Soc.
1987, 109, 5874 5875;c) W. Adcock, J. Cotton, N. A. Trout, J. Org.
Chem. 1994, 59, 1867 1876;d) R. Herrmann, W. Kirmse, Liebigs
Ann. Chem. 1995, 699 702;e) W. Adcock, N. J. Head, N. R. Lokan,
N. A. Trout, J. Org. Chem. 1997, 62, 6177 6182.
[3] A. Filippi, N. A. Trout, P. Brunelle, W. Adcock, T. S. Sorensen, M.
Speranza, J. Am. Chem. Soc. 2001, 123, 6396 6403.
[4] a) A. Rauk, T. S. Sorensen, C. Maerker, J. W. de M. Carneiro, S.
Sieber, P. von R. Schleyer, J. Am. Chem. Soc. 1996, 118, 3761 3762;
b) A. Rauk, T. S. Sorensen, P. von R. Schleyer, J. Chem. Soc. Perkin
Trans. 2 2001, 869 874.
[5] a) http://webbook.nist.gov/chemistry–Proton
affinity:
CH4 ˆ
1
ethers 3E and 3Z was analyzed by 13C and H NMR spectroscopy.
129.9 kcalmolÀ1; C2H4 ˆ 162.6 kcalmolÀ1;b) Calculated proton affin-
Procedure: A greaseless vacuum line was used to prepare the gaseous
mixtures according to conventional procedures. The starting chiral
substrate and the labeled nucleophile (1 and CH318OH or H218O; 2E or
2Z and H218O), as well as a thermal radical scavenger (O2), and a powerful
ity of 1 (this work) ˆ 208.9 kcalmolÀ1
.
[6] R. J. Blint, T. B. McMahon, J. L. Beauchamp, J. Am. Chem. Soc. 1974,
96, 1269 1278.
[7] Proton affinity: CH3OH ˆ 180.3 kcalmolÀ1 (ref. [5a]).
[8] Estimated according to: T. Su, W. J. Chesnavitch, J. Chem. Phys. 1982,
76, 5183 5185.
base [N(C2H5)3] (proton affinity ˆ 234.7 kcalmolÀ1
)
[5a] were introduced into
carefully outgassed 130 mL Pyrex bulbs, each equipped with a break-seal
tip. The bulbs were filled with 750 Torr of CH4 (1/CH318OH mixtures),
CH3F (1/H218O mixtures), or CH3Cl (2E or 2Z/H218O mixture), cooled to
liquid-nitrogen temperature, and sealed off. The gaseous mixtures were
then submitted to irradiation at a constant temperature (25 1208C) in a
60Co source (dose: 1 Â 104 Gy;dose rate: 5 Â 103 GyhÀ1, determined with a
neopentane dosimeter). Control experiments, carried out at doses ranging
from 1 Â 103 to 1 Â 105 Gy, showed that the relative yields of products are
largely independent of the dose. The radiolytic products were analyzed on
the above-mentioned chiral columns by GLC using a Chrompack-9002 gas
chromatograph equipped with a flame-ionization detector. The products
were identified by comparing their retention volumes with those of
authentic standard compounds, and were further confirmed by GLC MS
by using a Hewlett-Packard 5890A gas chromatograph in sequence with a
HP-5970B mass spectrometer. The yields were determined from the areas
of the corresponding eluted peaks by using the internal standard (i.e.,
benzyl alcohol) method, and individual calibration factors were used to
correct the detector response. Blank experiments were carried out in order
to ensure that thermal decomposition of the starting substrates and their
ethereal products, as well as the thermal racemization of 1, and the
epimerization of 2E and 2Z was not occurring within the temperature
range investigated.
[9] M. Speranza, A. Troiani, J. Org. Chem. 1998, 63, 1020 1026.
[10] Direct evidence of the inertness of (CH3)2Cl ions toward water arose
‡
from the observation that ethereal products recovered from the
CH3Cl/2E or 2Z/H218O mixtures contained less than 3% of the 18O
label.
[11] H216O is an ubiquitous impurity present in the gaseous systems, which
is either introduced with the bulk component or formed during
radiolysis. As pointed out by A. Troiani, F. Gasparrini, F. Grandinetti,
M. Speranza, J. Am. Chem. Soc. 1997, 119, 4525–4534, the average
stationary concentration of H216O in the irradiated systems is
estimated to approach that of the added H218O (ca. 3 Torr).
[12] This assumption appears even more reasonable if one considers that
unit efficiency is expected for the deprotonation of oxonium
intermediates by a very strong base such as N(C2H5)3.
[13] a) A. D. Becke, J. Chem. Phys. 1993, 98, 1372 1377, 5648 5652;b) C.
Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785 789.
[14] a) E. L. Eliel, T. J. Brett, J. Am. Chem. Soc. 1965, 87, 5039 5043, and
references therein;b) F. R. Jensen, B. H. Beck, J. Am. Chem. Soc.
1968, 90, 1066 1067.
[15] W. Koch, M. C. Holthausen, A Chemist×s Guide to Density Functional
Theory, 2nd ed., Wiley-VCH, Weinheim, 1999.
5402
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5396 5403

Ion Molecule Reactions
5396 5403
[16] E. C. Ashby, J. T. Laemmle, Chem. Rev. 1975, 75, 521 546.
[17] a) A. S. Cieplack, B. D. Tait, C. R. Johnson, J. Am. Chem. Soc. 1989,
111, 8447 8462;b) J. Hudec, J. Huke, J. W. Liebeschuetz, J. Chem.
Soc. Perkin Trans. 2, 1998, 1129 1138.
[18] M. Speranza, A. Filippi, G. Renzi, G. Roselli, F. Grandinetti, Chem.
Eur. J. 2003, 9, 2072 2078.
[19] The term troposelectivity has been recently coined (A. Filippi, M.
Speranza, J. Am. Chem. Soc. 2001, 123, 6077 82). Strictly speaking it
stands for the selectivity of a reagent toward the chiral or prochiral
face of a substrate on which it has been generated. In light of the
present work, troposelective reactions are as common as the
occurrence of slowly interconverting FAs in the gas phase.
[20] R. Greenwald, M. Chaykovsky, E. J. Corey, J. Org. Chem. 1963, 28,
1128 1129.
Stratman, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N.
Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R.
Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski,
G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malik, A. D.
Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
A. G. Raboul, B. B. Stefaniv, G. Liu, A. Liashenko, P. Piskorz, I.
Nanayakkara, C. Gonzales, M. Challacombe, P. M. W. Gill, B.
Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzales, M.
Head-Gordon, E. S. Replogle, J. A. Pople, GAUSSIAN 98, Revisio-
n A7, Gaussian Inc., Pittsburg, PA, 1998.
[22] A. P. Scott, L. Radom, J. Phys. Chem. 1996, 100, 16502 16513.
[21] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E.
Received: February 18, 2003
Revised: May 28, 2003 [F4859]
Chem. Eur. J. 2003, 9, 5396 5403
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5403