X-ray crystal structures of a benzonorbornenyl cation and of a protonated benzonorbornenol

Article abstract of DOI:10.1021/ja040115t

The crystal structure of the 9-methylbenzonorbornenyl cation Me-1 ⁺ shows a relatively strong interaction between the sp ²-hybridized carbon atom C9 and the aromatic ring (C4a-C9 ≡ C8a-C9 = 1.897(10) A). The anion Sb₂F₁₁^- is refined as rotationally disordered along the Sb...Sb axis. In sharp contrast to the findings about Me-1⁺, the protonated anti-benzonorbornenol 5⁺ is essentially an oxonium ion with only weak interaction between the C9 bridge and the aromatic ring despite the fact that it is already a positively charged ion, which upon loss of a water molecule is expected to give the parent cation H-1⁺. The hydrogen atoms on the oxonium O atom are involved in strong hydrogen bonds to chlorosulfonate anions and probably partially disordered despite the large estimated pK_a differences between the corresponding acid-base pairs. The experimentally determined cation structures are compared with structures computed by DFT methods. Detailed experimental procedures are given.

Full text of DOI:10.1021/ja040115t

Published on Web 08/17/2004
X-ray Crystal Structures of a Benzonorbornenyl Cation and of
a Protonated Benzonorbornenol†
Thomas Laube
Contribution from the Cilag AG, Hochstrasse 201-209, CH-8205 Schaffhausen, Switzerland
Received April 22, 2004 ; Revised Manuscript Received June 30, 2004; E-mail: tlaube@cilch.jnj.com
Abstract: The crystal structure of the 9-methylbenzonorbornenyl cation Me-1+ shows a relatively strong
interaction between the sp2-hybridized carbon atom C9 and the aromatic ring (C4a-C9 ≡ C8a-C9 )
1.897(10) Å). The anion Sb2F11- is refined as rotationally disordered along the Sb‚‚‚Sb axis. In sharp contrast
to the findings about Me-1+, the protonated anti-benzonorbornenol 5+ is essentially an oxonium ion with
only weak interaction between the C9 bridge and the aromatic ring despite the fact that it is already a
positively charged ion, which upon loss of a water molecule is expected to give the parent cation H-1+.
The hydrogen atoms on the oxonium O atom are involved in strong hydrogen bonds to chlorosulfonate
anions and probably partially disordered despite the large estimated pKa differences between the
corresponding acid-base pairs. The experimentally determined cation structures are compared with
structures computed by DFT methods. Detailed experimental procedures are given.
Chart 1
The unsubstituted benzonorbornenyl cation with a positively
charged carbon atom C91 (H-1+) has been discussed for the
first time by Bartlett et al.2 in a solvolytic study. Their kinetic
data showed that the solvolysis of 2 is faster than that of 3 but
slower than that of 4 (X: leaving group) (Chart 1), and they
discussed an interaction between the empty 2p orbital at C9
and the π electrons at the aromatic carbon atoms C4a and C8a.
This interpretation followed the suggestions published by
Winstein et al.3 in their landmark paper of 1955. Other solvolytic
studies have been carried out to support this interaction, e.g.,
by Tanida et al.4 and Tsuno and co-workers.5 The only direct
observations of R-1+ (R: Me, p-F-phenyl) by NMR spectros-
copy have been reported by the groups of Shin6 and Volz,7 but
† Parts of this work have been presented at the 16th International
Conference on Physical Organic Chemistry in San Diego, CA (August 4-9,
attempts to observe the unsubstituted H-1+ failed due to rapid
decomposition on warming up a solution of protonated 2 (X:
OH) in superacidic media.6a Shin and co-workers assumed that
they recorded the NMR spectrum of 5+ but not of H-1+. DFT
calculations on H-1+ and 6 have been carried out by Houk and
co-workers8 which yielded for H-1+ the bond lengths C9-C4a
) C9-C8a ) 1.980 Å; i.e., they found partial bonds between
C9 and the nearest C atoms of the π system. Because the
knowledge of the structures of bridged carbocations based on
crystal structure analyses is still very limited,9 we became
interested in salts of Me-1+ and 5+, especially because of the
useful stability data published by the groups of Shin and Volz.6a,7
2002; abstract LC 3) and at the 9th European Symposium on Organic
Reactivity in Oslo, Norway (July 12-17, 2003; abstract OR 42).
(1) Two different numbering schemes for the benzonorbornene system are used.
In the older scheme, the atom in the one-carbon bridge has the number 7,
whereas, in the newer literature, this atom has the number 9. We use in
this paper only the newer numbering scheme.
(2) Bartlett, P. D.; Giddings, W. P. J. Am. Chem. Soc. 1960, 82, 1240-1246.
(3) Winstein S.; Shatavsky, M.; Norton, C.; Woodward, R. B. J. Am. Chem.
Soc. 1955, 77, 4183-4184.
Results and Discussions
(4) (a) Tanida, H.; Hata, Y.; Ikegami, S.; Ishitobi, H. J. Am. Chem. Soc. 1967,
89, 2928-2932. (b) Tanida, H. Acc. Chem. Res. 1968, 1, 239-245.
(5) Yatsugi, K.; Saeki, Y.; Fujio, M.; Tsuno, Y. Memoirs of the Faculty of
Science, Kyushu UniVersity, Ser. C 1992, 18, 223-232; Chem. Abstr. 1993,
118, 190917.
(6) (a) Lee, C.; Lee, K.; Shin, J.-H. Chayon Kwahak Taehak Nonmunjip (Soul
Taehakkyo) 1982, 7, 51-64; Chem. Abstr. 1983, 98, 160119. (b) Ahn, Y.-
S.; Lee, Y.-N.; Shin, J.-H. Bull. Korean Chem. Soc. 1989, 10, 481-482.
(c) Ryu, G.-Y.; Shin, J.-H. Bull. Korean Chem. Soc. 1993, 14, 546-548.
Syntheses and Crystal Growth. The synthesis (Scheme 1)
started with the conversion of hexachlorocyclopentadiene (7)
(7) Volz, H.; Shin, J.-H.; Miess, R. J. Chem. Soc., Chem. Commun. 1993, 543-
544.
(8) Tantillo, D. J.; Hietbrink, B. N.; Merlic, C. A.; Houk, K. N. J. Am. Chem.
Soc. 2000, 122, 7136-7137; 2001, 123, 5851.
(9) Laube, T. Acc. Chem. Res. 1995, 28, 399-405.
9
10904
J. AM. CHEM. SOC. 2004, 126, 10904-10912
10.1021/ja040115t CCC: $27.50 © 2004 American Chemical Society

Benzonorbornenyl Cation and Protonated Benzonorbornenol
A R T I C L E S
Figure 1. ORTEP representation of the crystal structure of Me-1+ Sb2F11-. The cation and the anion contain a common crystallographic mirror plane
1
passing through C9, C9Me, Sb1, Sb2, F1a, F11a, F21a, F22a, F24a. The symmetry-related atoms have primed labels (′ ) x, /2 - y, z), and their chemical
names are given in parentheses (the latter ones are used in the discussion). The displacement ellipsoids are drawn at the 50% probability level, and the
hydrogen atoms are represented by spheres with a radius of 0.15 Å. The anion site with m ) Cs symmetry is drawn with dark green F ellipsoids, and the
two anion sites with 1 ) C1 symmetry (related to each other by the mirror plane) are drawn with light and middle green F ellipsoids. The three center, two
electron (3c, 2e) bond is indicated by copper bond sticks, and the shorter bonds of the aromatic19 ring are drawn as golden bond sticks.
were converted into pure 13 by catalytic hydrogenation.13
Scheme 1
Hydrolysis of 13 gave the key intermediate benzonorbornenone
(14).14 One part of 14 was converted into the tertiary anti-alcohol
15 by a Grignard reaction with MeMgI.4a The chloride 16 (one
stereoisomer of unknown configuration) was prepared in analogy
to 7-methyl-7-norbornyl chloride15 by reaction with concentrated
HCl, the fluoride 17 (one stereoisomer of unknown configura-
tion) in analogy16 by reaction with HF/pyridine. The secondary
alcohol 18 was prepared by reduction of another part of 14 with
LiAlH(OtBu)3.17 The SO2ClF for the crystal growth experiments
was prepared from SO2Cl2 and NaF18 and redistilled. The
reaction of 17 with somewhat more than 2 equiv of SbF5 and
subsequent recrystallization from methylene chloride gave
crystalline Me-1+ Sb2F11-. Slow addition of a solution of 18
dissolved in methylene chloride to a solution of ClSO3H in
methylene chloride gave upon concentration crystalline 5+
SO3Cl-‚CH2Cl2.
Crystal Structures and DFT Calculations. The structure
of Me-1+ Sb2F11- is shown in Figure 1; see also Tables S1 and
S2 (Supporting Information). The cation and the anion both
contain a crystallographic mirror plane. The anion is rotationally
disordered around the Sb1-Sb2 axis, and the fluorine atoms
of the three anion sites and their labels are represented with
different shades of green. The cation does not show disorder,
although its atomic displacement parameters are relatively large.
The most remarkable feature of the structure is the relatively
strong bending of the C9 bridge toward the aromatic19 ring
atoms C4a and C8a (C9-C4a ≡ C9-C8a ) 1.897(10) Å),
into the dimethylacetal 8,10 which was reacted with benzyne
(11) generated by thermolysis of the diazonium salt 10 of
anthranilic acid (9).11 The Diels-Alder adduct 1212 was
dehalogenated13 to a mixture of the acetals 13 and 13a, which
(14) Wilt, J. W.; Vasiliauskas, E. J. Org. Chem. 1972, 37, 1467-1472.
(15) Lustgarten, R. K.; Lhomme, J.; Winstein, S. J. Org. Chem. 1972, 37, 1075-
1078.
(16) Laube, T. HelV. Chim. Acta 1994, 77, 943-956.
(17) Okada, K.; Tomita, S.; Oda, M. Bull. Chem. Soc. Jpn. 1989, 62, 459-468.
(18) Tullock, C. W.; Coffman, D. D. J. Org. Chem. 1960, 25, 2016-2019.
(19) The term aromatic is still used for Me-1+ for simplicity, although a
significant perturbation of the π system is observed due to the (3c, 2e)
bond which is also indicated by the bond length alternation (see the bond
length pattern in 5+ and 6 in Table S2).
(10) Gassman, P. G.; Marshall, J. L. Org. Synth. Coll. Vol. 5, 424-428.
(11) Du¨rr, H.; Nickels, H.; Pacala, L. A.; Jones, M., Jr. J. Org. Chem. 1980,
45, 973-980.
(12) Wilt, J. W.; Vasiliauskas, E. J. Org. Chem. 1970, 35, 2410-2411.
(13) Ranken, P. F.; Battiste, M. A. J. Org. Chem. 1971, 36, 1996-1998.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10905

A R T I C L E S
Laube
Figure 4. Computed benzene ring C-C bond lengths of Me-1+ as functions
of the bond length C4a-C9 ) C8a-C9. The abscissa can be seen as the
reaction coordinate of the approach of the C9 bridge toward the C4a-C8a
bond retaining m ) Cs symmetry. The structures of the conformations were
optimized at the B3LYP/6-311+G(d,p) level with the constraints C4a-C9
) C8a-C9 ) 1.5(0.2)2.3 Å (x(y)z means lower bound x, increment y, upper
bound z; relaxed potential energy surface scan). The plotted functions are
interpolants. The bond lengths from the crystal structure (with error bars)
and from the fully optimized structure are drawn as points with the
corresponding color (data: see Table S2).
-
Figure 2. Geometry of the cation in the crystal structure of Me-1+ Sb2F11
(Distances and pyramidalizations ∆ in Å. Blue: interplanar angles in deg.
Red: C4a-C9, C8a-C9). Bottom: approximate Newman projection along
C1-C4.
case due to the proximity of the positively charged C9. The
bond lengths in the aromatic ring of the computed structure of
benzonorbornene (6) show a different pattern; see Figure 3
(short: C4a-C5, C6-C7, C8-C8a). The C9-C9Me bond
(1.479(11) Å) is slightly shorter than a Csp2-CH3 bond (average
length:20 1.503(1) Å), presumably due to C-H hyperconjuga-
tion, although the shortening (0.024(11) Å) is not very signifi-
cant. A similar shortening has been observed in a chloro-
substituted (deloc-2,3,5)-bicyclo[2.1.1]hex-2-en-5-ylium ion.21
The gas-phase structure of Me-1+ has also been computed
on the B3LYP/6-311+G(d,p) level for comparison; see Table
S2. The interaction between C9 and the C4a-C8a bond is
somewhat stronger in the experimentally determined structure
than in the computed structure, as can be seen from the angles
between these two bridges: 95.7(6)° vs 98.6°. The aromatic
bond lengths show a similar pattern as in the crystal structure;
see Figure 4.
Figure 3. Geometry of benzonorbornene (6; distances in Å, angles in deg).
Blue: interplanar angles. Red: C4a-C9, C8a-C9. Bold: average from
three precise structures with aromatic substituents from the Cambridge file.
Plain: computed with B3LYP/6-311+G(d,p).
Another remarkable feature of Me-1+ is the nonplanarity of
the aromatic ring. The plane P(4a,5,8,8a) is bent upward (i.e.,
toward the C9 bridge) with regard to P(1,4,4a,8a) by 9.6(8)°
(see Figure 2 and Table S2), and the plane P(5,6,7,8) is bent
downward with regard to P(4a,5,8,8a) by 5.2(8)°. This leads to
leading to an angle between the C9 and the C4a/C8a bridge of
95.7(6)°; see Figure 2. This indicates a strong interaction
between the two bridges, because the corresponding distances
and the corresponding angle in benzonorbornene (as a reference
for the absence of such an interaction) are 2.35 Å and 126°;
see Figure 3.
The aromatic ring shows bond length alternation: C5-C6
≡ C7-C8 ) 1.358(9) Å are short, C6-C7 ) 1.423(15) Å and
C4a-C5 ≡ C8-C8a ) 1.395(8) Å are longer by 0.065(17) and
0.037(12) Å, respectively. The bond C4a-C8a ) 1.414(12) Å
is longer by 0.056(15) Å but must be considered as a special
the pyramidalizations22
∆C4a ≡ ∆C8a ) 0.077(5) Å. Similar ring
(20) Allen, F. H.; Watson, D. G.; Brammer, L.; Orpen, A. G.; Taylor, R. In
International Tables for Crystallography, Volume C; Prince, E., Ed.; Kluwer
Academic Publishers: Dordrecht, 2004; pp 790-811.
(21) Laube, T.; Lohse, C. J. Am. Chem. Soc. 1994, 116, 9001-9008.
(22) The pyramidalization is the distance of a trigonal atom from the plane
defined by its bond partners.
9
10906 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Benzonorbornenyl Cation and Protonated Benzonorbornenol
A R T I C L E S
1
Figure 5. Ball and stick diagram of Me-1+ and the two anions in the proximity of C9 in the crystal. The symmetry operations are as follows: ′ ) x, /2
1
- y, z; ′′ ) x, y, z + 1; ′′′ ) x, /2 - y, z + 1. The C9‚‚‚F contacts are indicated by thin red sticks; for the bond and fluorine colors, see also the legend of
Figure 1.
deformations have also been observed in the computed structure
of H-1+.8 The corresponding deformations of the aromatic ring
in the computed structure of 6 have opposite directions and are
much weaker.
The packing of the cations and anions (Figure 5) shows a
feature seen already in many other carbocation structures:23 the
positively charged carbon atom C9 has on its anti side a close
contact with F11 or the corresponding disordered sites F11a,
F11b, F11b′ (smallest C9‚‚‚F, C9‚‚‚F11a ) 3.03(2) Å; sum of
the van der Waals radii24 of C and F, 1.70 Å + 1.47 Å ) 3.17
Å). The closest anion contact on the syn side of C9 occurs with
F21 of a symmetry-related anion or its disordered sites F21a′′,
F21b′′, F21b′′′ (smallest C9‚‚‚F: C9‚‚‚F21a′′ ) 4.12(2) Å), but
this is about 1 Å more than the sum of the van der Waals radii.
The same F atom is also situated in the proximity of the aromatic
C atoms (C4a‚‚‚F21a′′ ≡ C8a‚‚‚F21a′′ ) 4.09(2) Å, C5‚‚‚F21a′′
≡ C8‚‚‚F21a′′ ) 3.84(2) Å, C6‚‚‚F21a′′ ≡ C7‚‚‚F21a′′ ) 3.65-
(2) Å). No anion is found on the other side of the aromatic ring
(or on the endo side of the C4a-C8a bond, as in the crystal
structure of a substituted norbornenyl cation25), but a methyl
group of another cation is located in its proximity; see Figure
6. The free space between the anions in the a direction illustrates
the fact that the anions have rotational degrees of freedom
leading to the observed disorder.
The arrangement of anions around a cationic C atom seems
to be reasonable for electrostatic reasons, but it can also be
interpreted in terms of the reaction coordinate26 of a nucleophilic
attack on C9 leading to anti or syn products, respectively. The
close C9‚‚‚F contact on the anti side of C9 agrees with the
observation of Volz et al.7 that quenching of Me-1+ with MeOH/
MeONa yields only the corresponding anti ether.
In previous papers (ref 9, Figure 7; ref 25, Figure 9), we
discussed the possibility that the observed bridged structures
in the crystal could be a static or dynamic superposition of
classical carbocations. We found that the expected distances
between disordered atoms of the hypothetical classical cations
were too large to remain undetected in a thermal motion analysis
or that the disordered atoms should even have resolvable
Figure 6. Space-filling diagram showing the surrounding of Me-1+ in the
crystal. The unit cell is shown as parallelepiped (axes: a, red; b, green; c,
blue). The arrangement from Figure 5 is drawn twice differing by the
translation vector (1,0,0). For the fluorine colors, see also the legend of
Figure 1.
positions. This is also the case for Me-1+ if one uses suitable
neutral molecules as models27 for classical ions. We now also
analyze the potential energy surface of the carbocation optimized
at a somewhat lower level of theory if compared with its fully
optimized structure (see Table S2). The potential energy surface
(Figure 7; the conformation of the cation in the crystal lies 0.6
kcal/mol above the minimum of the depicted function) shows
only one minimum for the bridged structure. The potential
energy function is not symmetrical along the diagonal (C4a-
C9 ) C8a-C9) indicating that weakening the (3c, 2e) bond is
easier than strengthening it. There is no hint that classical
isomers, which should have C4a-C9 ≈ 1.63 Å and C8a‚‚‚C9
≈ 2.09 Å (or vice versa), are stable species. They are expected
to be ∼9 kcal/mol higher in energy if compared with the bridged
structure.
The structure of 5+ in the crystal is shown in Figures 8 and
9. The most remarkable feature is the fact that the carbon
skeleton of this ion is structurally more similar to the hydro-
carbon 6 or to its precursor 1827 than to a bridged cation like
Me-1+. The angle between the C9 bridge and the C4a,C8a
bridge is 122.38(10)° if compared with 126.1° for 6 and 95.7(6)°
for Me-1+. The C-O bond length (1.480(2) Å) is elongated
(23) Laube, T. Chem. ReV. 1998, 98, 1277-1312.
(24) Bondi, A. J. Phys. Chem. 1964, 68, 441-451.
(25) Laube, T. J. Am. Chem. Soc. 1989, 111, 9224-9232.
(26) Bu¨rgi, H. B.; Dunitz, J. D. Acc. Chem. Res. 1983, 16, 153-161.
(27) DFT results are given in the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10907

A R T I C L E S
Laube
Figure 7. Contour line diagram of the calculated potential energy surface
of Me-1+ in the gas phase depending on the lengths of the weak bonds
(C4a-C9 and C8a-C9) of the assumed (3c, 2e) bond involving C4a, C8a,
C9. The potential energies were computed by optimizing the structure at
the B3LYP/6-311++G(3df,3pd)// B3LYP/6-31G(d) level with the con-
straints C4a-C9 ) 1.5(0.2)2.3 Å and C8a-C9 ) 1.5(0.2)2.3 Å (relaxed
potential energy surface scan). The plotted function is a least-squares 4th
order polynomial in two variables, which is symmetrical with regard to an
exchange of the independent variables. Emin is the energy at the minimum
of the adjusted polynomial.
Figure 9. Geometry of the cation in the crystal structure of 5+
SO3Cl-‚CH2Cl2 (Distances in Å. Blue: interplanar angles in deg). The
hydrogen atom numbers (not drawn) correspond to their carbon atoms
(x ) exo, n ) endo). Bottom: approximate Newman projection along
C1-C4.
crystal structures (C-O ) 1.413(5) Å).20 The observed C-O
bond in 5+ is not as long as the C-O bonds in the computed
structures of protonated tert-BuOH (C-O ) 1.663 Å)29 or of
protonated 2-norbornanols (C-Oexo ) 1.624 Å; C-Oendo
)
1.615 Å)30 or of other protonated alcohols.31 The results of
Uggerud et al.31 show that the C-O bonds of protonated
alcohols become shorter upon complexation with a water
molecule (i-PrOH2+, C-O ) 1.579 Å; i-PrOH2+‚OH2, C-O
) 1.532 Å).32 Therefore we computed also the structure of 5+
in the gas phase and in water33 (Table S2). The experimental
structure of 5+ is more similar to the computed structure of 5+
in water (C-O ) 1.526 Å) than in the gas phase (C-O ) 1.704
Å). The unusual C-O bond length in the computed gas-phase
structure of 5+ made an investigation of its potential energy
surface necessary: the potential energy as a function of C4a-
C9 ) C8a-C9 is shown in Figure 10.
One sees a double-minimum system, where 5+ is about 1.5
kcal/mol higher in energy if compared with the absolute
Figure 8. ORTEP representation of the crystal structure of 5+ SO3Cl-‚
CH2Cl2. The additional anion is shown to depict the hydrogen bonds. Two
sites are drawn for each of the disordered oxonium H atoms (H1o and
H2o): the major oxonium H sites (H1oa, sof ) 0.76(4); H2oa, sof )
0.71(4)) are drawn as weakly transparent spheres, and the corresponding
minor sites (H1ob, H2ob), as more transparent spheres. The symmetry-
related atoms have primed labels (′ ) x, y - 1, z). The displacement
ellipsoids and the hydrogen spheres are drawn at the 50% probability level.
(28) Minkwitz, R.; Schneider, S. Z. Anorg. Allg. Chem. 1998, 624, 1989-1993.
(29) Jorgensen, W. L.; Buckner, J. K.; Huston, S. E.; Rossky, P. J. J. Am. Chem.
Soc. 1987, 109, 1891-1899.
(30) Schreiner, P. R.; Schleyer, P. v. R.; Schaefer, H. F., III. J. Org. Chem.
1997, 62, 4216-4228.
(31) Uggerud, E.; Bache-Andreassen, L. Chem.sEur. J. 1999, 5, 1917-1930.
(32) A probably related shortening of a B-N bond is observed in H3B-NH3.
(a) Bu¨hl, M.; Steinke, T.; Schleyer, P. v. R.; Boese, R. Angew. Chem., Int.
Ed. Engl. 1991, 30, 1160-1161. (b) Dillen, J.; Verhoeven, P. J. Phys. Chem.
A 2003, 107, 2570-2577. (c) Leopold, K. R.; Canagaratna, M.; Phillips,
J. A. Acc. Chem. Res. 1997, 30, 57-64.
(33) We choose water as solvent because it is the solvent with the highest
dielectric constant for which predefined parameters are available, see
computational section. We assume that the environment of 5+ in the crystal
is also very polar due to the cation-anion contacts. The dielectric constant
of ClSO3H is 60 ( 10 at 15° C according to: Cremlyn, R. J. Chlorosulfonic
Acid; Royal Society of Chemistry: Cambridge, 2002; p 2.
by 0.048(2) Å if compared with the average C-O bond length
in secondary alcohols (1.432(1) Å).20 A similar C-O elongation
upon protonation (0.04(1) Å) is observed in the crystal structure
of MeOH2+ (C-O ) 1.45(1) Å)28 if compared with MeOH in
9
10908 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Benzonorbornenyl Cation and Protonated Benzonorbornenol
A R T I C L E S
Figure 11. Computed benzene ring C-C bond lengths (solid lines; left
scale) and C9-O bond length (dashed line; right scale) of 5+ in the gas
phase as functions of the bond length C4a-C9 ) C8a-C9. The abscissa
can be seen as the reaction coordinate of the approach of the C9 bridge
toward the C4a-C8a bond retaining m ) Cs symmetry. The structures of
the conformations were optimized at the B3LYP/6-311+G(d,p) level with
the constraints C4a-C9 ) C8a-C9 ) 1.7(0.1)2.4 Å (relaxed potential
energy surface scan). The plotted functions are interpolants. The bond
lengths of the crystal structure are drawn as filled circles, and those of
various computed structures as empty symbols with the corresponding color
(data: see Table S2 and Supporting Information).
Figure 10. Potential energy of 5+ and of the water complex of H-1+ in
the gas phase calculated at various levels of theory as a function of the
bond length C4a-C9 ) C8a-C9. The abscissa can be seen as the reaction
coordinate of the approach of the C9 bridge toward the C4a-C8a bond
retaining m ) Cs symmetry. The structures of the conformations were
optimized with the constraints C4a-C9 ) C8a-C9 ) 1.7(0.1)2.4 Å. The
plotted functions are interpolants. Blue: lower level geometry. Red: higher
level geometry. Solid functions represent energies computed at a higher
level of theory if compared with the dashed ones. The boxes with arrows
indicate qualitatively the ranges of existence of the various ions.
minimum, a H-1+/water complex.34 In Figure 11 various bond
lengths of the experimental structure of 5+ and of some
computed structures are drawn along the reaction coordinate
of the bridging. Based on the C-O and on the C4a-C9 ) C8a-
C9 bond lengths, one can say that there is only a very weak
interaction or bridging between the π system and C9 in the
crystal; i.e., the experimental structure of 5+ must be seen as
an early point on the reaction coordinate of the dissociation into
H-1+ and H2O.
containing a chlorosulfonate anion). There is a significant
shortening of the S-Cl bond (2.0506(6) Å) if compared with
the structure computed by Steudel et al. in the gas phase39 (S-
Cl ) 2.242 Å).40 Each CH2Cl2 molecule is interacting with
another CH2Cl2 molecule and the Cl atom of an anion in a very
characteristic angle pattern (Figure 14).
These Cl‚‚‚Cl interactions are also observed in the crystal
41
42
structures of CH2Cl2 (Cl‚‚‚Cl ) 3.492 Å) and of Cl2
(Cl‚‚‚Cl ) 3.258 Å). They may be interpreted as nucleophile-
electrophile interactions as suggested by Bent;43 i.e., the CH2Cl2
molecule is not simply situated in a cavity of suitable size.
Several details of the structure of 5+ SO3Cl-‚CH2Cl2 attract
attention. The site occupation factors (sof) of the H atoms on
the oxonium O atom (H1oa, H2oa) are significantly smaller than
1 if they are refined. If we assume that this is not an artifact
resulting from the usual difficulties in refining H atoms in X-ray
crystal structures, the missing H atom populations can be refined
with geometrical and Uiso constraints on the anion O atoms
closest to the major H sites: O1 and O2 (H1ob and H2ob).
The crystal packing (Figure 12) shows infinite chains
consisting of cations and anions in alternating order, which are
connected via H bonds in the b direction. The geometrical
parameters (O‚‚‚O ≈ 2.55 Å; see Figure 13) indicate that the
H bonds are strong.35 The interaction between O and O3′′ may
be interpreted as an additional bifurcated H bond.36,37 The
chlorosulfonate anion in our structure is not significantly
different from the anion in NO+ SO3Cl- reported Ho¨hle and
Mijlhoff38 (the only other crystal structure in the literature
(34) It was obviously a fortunate coindidence that the optimization on the
B3LYP/6-311+G(d,p) level lead to the local minimum 5+. The barrier
between 5+ and the H-1+/water complex increases only with the large basis
set; see Figure 10.
(38) Ho¨hle, T.; Mijlhoff, F. C. Recl. TraV. Chim. Pays-Bas 1967, 86, 1153-
1158.
(39) Steudel, R.; Otto, A. H. Eur. J. Inorg. Chem. 2000, 2379-2386.
(40) S-Cl bond lengths computed by the B3LYP/6-311+G(d,p) method in the
gas phase: SO3Cl- 2.290 Å; 18‚ClSO3H 2.108 Å. See Supporting
Information.
(41) Kawaguchi, T.; Tanaka, K.; Takeuchi, T.; Watanabe´, T. Bull. Chem. Soc.
Jpn. 1973, 46, 62-66.
(42) Powell, B. M.; Heal, K. M.; Torrie, B. H. Mol. Phys. 1984, 53, 929-939.
(43) Bent, H. A. Chem. ReV. 1968, 68, 587-648.
(35) Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford University
Press: Oxford, 1997; pp 33-55.
(36) Taylor, R.; Kennard, O.; Versichel, W. J. Am. Chem. Soc. 1984, 106, 244-
248. Rozas, I.; Alkorta, I.; Elguero, J. J. Phys. Chem. A 1998, 102, 9925-
9932.
+
(37) A similar H bond network is observed in the crystal structure of MeOH2
- 28
AsF6
.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10909

A R T I C L E S
Laube
Figure 14. Geometry of the Cl‚‚‚Cl interactions in the crystal structure of
5+ SO3Cl-‚CH2Cl2 (symmetry operations omitted for clarity). The three
anions (drawn with the minor hydrogen sites) and the three CH2Cl2
molecules are symmetry-related by the translation vector (0,1,0), see Figure
12. Dashed light blue lines mark the contacts, the arrows point from the
nucleophilic to the electrophilic atom. Red: solvent-solvent interactions;
blue: solvent-anion interactions.
pKa for protonated secondary alcohols relative to water in the
condensed phase is about -2.48 An estimated pKa value for
chlorosulfonic acid of -6 is reported by Guthrie,49 and from
50
the Hammett acidity function H0 it could be cautiously
estimated that the pKa lies around the range from -9
(approximate pKa for H2SO4)48a to -12 (approximate pKa for
FSO3H).48a Such a pKa difference should guarantee a practically
complete protonation of the alcohol. On the other hand, DFT
calculations of the energy profile of the proton transfer in an
isolated ion pair constrained to a geometry as in the asymmetric
unit of the crystal structure indicates that the proton is
completely transferred to the anion; i.e., a molecular complex
18‚ClSO3H is present (see Figure 15).
Figure 12. Crystal packing of 5+ SO3Cl-‚CH2Cl2. The unit cell is shown
as parallelepiped (axes: a, red; b, green; c, blue). The interactions in chains
along the b axis consisting of cations and anions (violet ellipse) and in
chains consisting of anions and CH2Cl2 molecules (orange ellipse) are
depicted in Figures 13 and 14.
A second minimum describing an ion pair 5+ SO3Cl- does
not exist; only a shoulder is recognizable. We optimized also
the structure of an unconstrained ion pair and obtained also only
a hydrogen-bonded complex 18‚ClSO3H.27 These results show
that an isolated ion pair in the gas phase is not a suitable model
for 5+ SO3Cl- in the crystal. In view of the discrepancy between
the pKa value prediction (5+ SO3Cl- much more stable) and
the gas-phase DFT calculations (18‚ClSO3H much more stable),
we assume that the experimentally observed energetic similarity
of the two sites of each oxonium proton is the result of a
cooperative effect in the infinite cation-anion chains in the
crystal.
The SO3Cl- anion does not have local 3m ) C3V symmetry.
The Oanion atoms engaged in the strong H bonds (O1 and O2)
have longer S-O bonds and smaller Cl-S-O angles than O3,
and O1-S-O2 is smaller than the other O-S-O angles (Figure
13); i.e., the anion has only approximate m ) Cs symmetry with
the mirror plane passing through O3, S, and Cl. Such structural
deviations are usually ascribed to packing interactions, but in
this case they could at least partially be the result of an averaging
of ClSO3H and SO3Cl- structures due to the unresolved disorder
mentioned above. Therefore the structural parameters of 5+ in
the crystal must be interpreted carefully. The observed C9-O
Figure 13. Geometry of the H bonds and of the anion in the crystal structure
of 5+ SO3Cl-‚CH2Cl2 (symmetry operations: ′ ) x, y - 1, z; ′′ ) -x +
1
1
1/2, y - /2, -z + /2; see also Figure 12). The H bonds are described by
the following parameters: dashed light blue lines ) O‚‚‚O distances, solid
dark blue lines ) H‚‚‚O distances, italic ) O-H‚‚‚O and H‚‚‚O-S angles.
Only the major H sites are drawn. Anion: red ) O-S-O angles, green )
Cl-S-O angles.
This would mean that the measured crystal structure is an
averaged or disordered structure consisting of superimposed
cations 5+ and alcohol molecules 18 on the cation site, and of
superimposed SO3Cl- anions and ClSO3H molecules on the
anion site. While H bond disorder is well-known for certain
forms of ice and other structures,44-47 it is more difficult to
understand in the structure of 5+ SO3Cl-‚CH2Cl2. The estimated
(47) Selected papers about double minimum hydrogen bonds: Novak, A. Struct.
Bonding (Berlin) 1974, 18, 177-216. Bernstein, J.; Etter, M. C.; Leiser-
owitz, L. The Role of Hydrogen Bonding in Molecular Assemblies. In
Structure Correlation, Volume 2; Bu¨rgi, H.-B., Dunitz, J. D., Eds.; VCH:
Weinheim, 1994; pp 431-507.
(48) (a) Smith, M. B.; March, J. March’s AdVanced Organic Chemistry, 5th
ed.; Wiley: New York, 2001; pp 329-331. (b) Literature values of the
pKa of 2-propanol between -0.35 and -5.2 (depending on the analytical
method) are reported in: Arnett, E. M.; Quirk, R. P.; Burke, J. J. J. Am.
Chem. Soc. 1970, 92, 1260-1266.
(44) Structure of ice Ih: Leadbetter, A. J.; Ward, R. C.; Clark, J. W.; Tucker,
P. A.; Matsuo, T.; Suga, H. J. Chem. Phys. 1985, 82, 424-428.
(45) Structure of malonaldehyde: Rowe, W. F., Jr.; Duerst, R. W.; Wilson, E.
B. J. Am. Chem. Soc. 1976, 98, 4021-4023. Perrin, C. L.; Thoburn, J. D.
J. Am. Chem. Soc. 1989, 111, 8010-8012.
(46) Proton transfer in acid base complexes: Lethonen, O.; Hartikainen, J.;
Rissanen, K.; Ikkala, O.; Pietila¨, L.-O. J. Chem. Phys. 2002, 116, 2417-
2424. Chipot, C.; Gorb, L. G.; Rivail, J.-L. J. Phys. Chem. 1994, 98, 1601-
1607.
(49) Guthrie, J. P. Can. J. Chem. 1978, 56, 2342-2354.
(50) H0(ClSO3H) ) -13.80: Gillespie, R. J.; Peel, T. E.; Robinson, E. A. J.
Am. Chem. Soc. 1971, 93, 5083-5087. H0(H2SO4) ) -11.93, H0(FSO3H)
) -15.07: Gillespie, R. J.; Peel, T. E. J. Am. Chem. Soc. 1973, 95, 5173-
5178.
9
10910 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Benzonorbornenyl Cation and Protonated Benzonorbornenol
A R T I C L E S
Chart 2. Interpretation of the Structure of Me-1+
the MeOH2+ structure (0.04(1) Å),28 for which a proton disorder
-
is not expected due to the weakly coordinating AsF6 anion.
One would expect a significantly larger C-O elongation for
5+ because it is probably the precursor52 of the short-lived
carbocation H-1+ and could advance on the reaction coordinate
+
of the H2O elimination, whereas MeOH2 is not a cation
+
precursor, although MeOH2 could loose H2O in an SN2
reaction. Thus the similar estimated C-O bond elongations upon
protonation also support our 5+/18 disorder hypothesis.53 Olah
et al.54 observed in NMR experiments a decomposition tem-
perature of MeOH2+ around +50° C, whereas other protonated
alcohols decompose at a much lower temperature; this demon-
strates the higher stability of MeOH2+. Additional support for
the disorder hypothesis of 5+ may be seen in the similarity
between the crystal structures of 5+ and of the rather stable,
neutral 2 (X ) brosylate).55 Comparing these two potential
precursors of H-1+, one would expect 5+ to be more advanced
on the reaction coordinate toward H-1+, because the brosylate
dissociation involves a charge separation.56 Unfortunately the
problem of this kind of potential disorder has not been discussed
in related crystal structures containing H2SO4 or FSO3H (see
Supporting Information).
Figure 15. Potential energy of an ion pair 5+ ClSO3- and of the molecular
complex 18‚ClSO3H in the gas phase calculated at various levels of theory
as function of the bond length O2-H2o (blue functions). The abscissa can
be seen as the reaction coordinate of the proton transfer. The structures of
the conformations were optimized with several constraints to retain a mutual
spatial orientation of the ions or molecules as found in the asymmetric unit
of the crystal structure (see Supporting Information). The dipole moment
(green) increases upon formation of the ion pair. The distance O-H2o
(orange) is a nearly linear function of O2-H2o up to O2-H2o ≈ 1.55 Å
because O2-H2o-O g 160°. The boxes with arrows indicate qualitatively
the ranges of existence of the various complexes.
Shin et al.6a assumed that they observed 5+ in two different
superacidic media by 1H NMR spectroscopy. We recorded the
1H NMR spectrum of a mixture of 18 with an excess of ClSO3H
in CD2Cl2 and obtained a spectrum similar to the spectrum in
FSO3H/SO2ClF reported by Shin et al.. GIAO calculations of
5+ SO3Cl- and 18‚ClSO3H (crystal geometry or fully optimized)
at the B3LYP/6-311++G(2d,2p) level27 show a better agree-
ment with our experimental chemical shifts than those of free
5+. We cannot decide on the basis of the GIAO calculations
whether complexes containing 5+ or 18 or mixtures of them
are present in a ClSO3H/CD2Cl2 solution.
Figure 16. Superposition of 5+ (DFT, water; red) on the two enantiomers
of 18 (DFT, gas phase; green and blue). The slight asymmetry is the result
of the least squares adjustment (only heavy atoms) of 5+ on a weighted
average of 18 and ent-18 (the weights were derived from the sofs).
Conclusion
The crystal structure of Me-1+ can be interpreted as shown
in Chart 2. The π electron cloud is partially delocalized into
the empty p orbital at C9. This leads to a mutual approach of
the C9 and the C4a-C8a bridges. From the bond lengths in
bond length (1.480(2) Å) could be seen as a lower boundary
for its true value. Assuming the correctness of the H site
occupation factors, the structure is a superposition of 47(5)%
5+ and of 53(5)% 18 (24% of one and 29% of the other
enantiomeric conformers). A simple superposition of 5+ (DFT,
H2O) and 18 (DFT, gas phase) shows that such a disorder would
be impossible to detect by investigation of the atomic displace-
ment ellipsoids because of the very similar structures (see Figure
16).
The weighted average of the C-O bond lengths of 5+ (DFT,
H2O 1.526 Å; see Table S2) and of 18 (DFT, gas phase 1.420
Å) is 1.469 Å and thus in sufficient agreement with the observed
value.51 The C-O elongation upon protonation in our structure
(0.048(2) Å) is comparable to the corresponding elongation in
(51) Semiempirical calculations (MOPAC, AM1) yield the following C-O bond
lengths: 5+ 1.512 Å, 18 1.406 Å.
(52) We do not have any experimental information that 5+ is a precursor of
H-1+. The energy profile depicted in Figure 10 shows that this reaction
may easily take place in the gas phase.
(53) The C-O bond length in the MeOH2+ crystal structure may be influenced
by weak C‚‚‚F contacts anti to the C-O bond, which could advance the
cation somewhat towards a SN2 transition state.
(54) Olah, G. A.; Sommer, J.; Namanworth, E. J. Am. Chem. Soc. 1967, 89,
3576-3581.
(55) Koyama, H.; Okada, K. J. Chem. Soc. B 1969, 940-950.
(56) If one compares the crystal structures of benzonorbornenes with 9-syn and
9-anti leaving groups, the 9-anti derivatives have slightly shorter C4a-C9
and C8a-C9 contacts than the 9-syn derivatives. See Supporting Informa-
tion.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10911

A R T I C L E S
Laube
from the Cambridge Structural Database.63 Semiempirical calculations
were carried out with SYBYL 6.064/MOPAC,65 DFT calculations with
Gaussian 03 W66 with GaussView 3.09.67 Molecular modeling was
carried out with SYBYL 6.064 and ViewerPro.68 ADP, ball-and-stick,
and packing diagrams were produced with Ortep.69,70 Photorealistic
graphics were generated with POV-Ray.71
the benzene-like ring, we conclude that a (3c, 2e) bond
interacting with a butadienyl system is a suitable description of
Me-1+ as suggested by Bartlett et al..2 The experimentally
determined structure shows a higher degree of bridging than
the computed DFT gas-phase structure.
The crystal structure of 5+ is probably disordered with 18
and can thus only cautiously be interpreted. The C skeleton
indicates only a weak interaction between the C9 and the C4a-
C8a bridges. The observed elongation of the C-O bond upon
protonation is 0.048(2) Å or more. DFT calculations of 5+ in
the gas phase are not appropriate for a description in the crystal
because of the high tendency to dissociate into a H-1+/water
complex observed in the calculations. If the polar environment
of 5+ in the crystal is simulated by a DFT calculation in water,
a relatively good description with the exception of the C-O
bond length is achieved. With regard to the C-O bond length,
the ion 5+ seems to be related to H3B-NH3 and other “partially
bonded” molecules.32c The disordered oxonium proton positions
cannot be explained by DFT calculations on isolated ion pairs
in the gas phase, which lead to a practically complete proton
transfer to the anion, i.e., the hydrogen-bonded complex 18‚
ClSO3H. Therefore we assume that the observed equilibrium
between the oxonium ion 5+ and the alcohol 18 in the crystal
occurs due to cooperative hydrogen bonds and cation-anion
interactions in the infinite chains.
Acknowledgment. We thank Dr. Anthony Linden (University
of Zurich, Zurich, Switzerland) for carrying out the X-ray
measurements and Mrs. Sandra Weber (Cilag AG) for help with
the syntheses.
Supporting Information Available: Tables S1 and S2,
Experimental Section. Tables and additional graphics of the
crystal structure data, NMR spectra, results of quantum chemical
calculations, data about various structures from the literature
mentioned in the discussion. Two X-ray crystallographic files
in CIF format. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA040115T
(62) Laube, T. FOTOPLOT, version 6; Schaffhausen, Switzerland, 2001.
(63) Allen, F. H. Acta Crystallogr., Sect. B 2002, B58, 380-388. Bruno, I. J.;
Cole, J. C.; Edgington, P. R.; Kessler, M. K.; Macrae, C. F.; McCabe, P.;
Pearson J.; Taylor, R. Acta Crystallogr., Sect. B 2002, B58, 389-397.
(64) SYBYL 6.0; Tripos Inc.: St. Louis, MO, 1992.
(65) Stewart, J. J. P. MOPAC, version 5.0 (integrated in SYBYL 6.064); QCPE
#455.
(66) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa,
J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.;
Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo,
J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;
Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels,
A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03W,
revision B.03; Gaussian, Inc.: Pittsburgh, PA, 2003.
Computational Section
The X-ray data reductions were carried out with teXsan,57 the
absorption corrections (based on multiscans from symmetry-
related measurements) were carried out with Sortav.58 The crystal
structures were solved with SHELXS59 and refined with SHELXL.60
Analyses of the refined structures were carried out with PLATON61
and analyses of the reciprocal space data have been displayed also with
FOTOPLOT.62 Experimental reference structures have been retrieved
(57) teXsan 1.10 (Unix version); Molecular Structure Corporation: The
Woodlands, TX.
(58) Blessing, R. H. Acta Crystallogr., Sect. A 1995, A51, 33-38.
(59) Sheldrick, G. M. SHELXS-97 - CRYSTAL STRUCTURE SOLUTION -
UNIX VERSION, release 97-2; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(60) Sheldrick, G. M. SHELXL-97 - CRYSTAL STRUCTURE REFINEMENT
- UNIX VERSION, release 97-2; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(61) Spek, A. L. PLATON: A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 2002. Farrugia, L. J. Platon for
Windows; University of Glasgow: Glasgow, U.K., 2002.
(67) GaussView 3.09; Gaussian, Inc.: Pittsburgh, PA, 2003.
(68) DiscoVery Studio ViewerPro; Accelrys Inc.: San Diego, CA, 2002.
(69) Burnett, M. N.; Johnson, C. K. ORTEP-III: Oak Ridge Thermal Ellipsoid
Plot Program for Crystal Structure Illustrations; Oak Ridge National
Laboratory Report ORNL-6895, 1996.
(70) Farrugia, L. J. Ortep-3 for Windows Version 1.07; University of Glasgow:
Glasgow, U.K., 2001.
(71) POV-Team POV-Ray, version 3.1g; c/o Hallam Oaks P/L: P.O. Box 407,
Williamstown, Victoria 3016, Australia, 1999.
9
10912 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004