Bridgehead carbocations via carbene fragmentation: Erasing a 1010 kinetic preference

Article abstract of DOI:10.1021/ja020002e

1-Norbornyloxychlorocarbene (1-NorOCCl), 1-bicyclo[2.2.2]octyloxychlorocarbene (1-BcoOCCl), and 1-adamantyloxychlorocarbene (1-AdOCCl) were generated in dichloroethane (DCE) by photolysis of the appropriate diazirines. The exclusive product in each case was the bridgehead alkyl chloride formed by fragmentation of the carbene to [R⁺ OC Cl^-] ion pairs, loss of CO, and cation-anion collapse. In mixtures of methanol and DCE, each carbene gave three products: RCl, ROH, and ROMe. RCl and ROMe resulted from competition between ion pair collapse and methanol capture of the cation. ROH resulted from methanol capture of the carbene (before fragmentation), followed by eventual methanolysis and hydrolysis of ROCH(Cl)OMe. The ratios of carbene capture to carbene fragmentation fell in the order 1-NorOCCl > BcoOCCl > 1-AdOCCl; 1-Nor+ was the least stable cation and the slowest to form by fragmentation, so that this carbene was the most readily captured. This trend was accentuated in methanol-pentane mixtures, where ionic fragmentation was further slowed in the less polar solvent. Laser flash photolysis with either UV or time-resolved infrared (TRIR) monitoring permitted the determination of the absolute rate constants for fragmentations of the carbenes in DCE at 25 °C. The rate constants (s^-1) were: 1-NorOCCl (3.3 × 10⁴), 1-BcoOCCl (1.5 × 10⁵), and 1-AdOCCl (5.9 × 10⁵). The rate constants decreased in the order of increasing strain in the resulting bridgehead carbocation, but the range of rate constants was compressed to a factor of only ～18. This constrasts with the factor of 10¹⁰ by which the acetolysis of 1-AdOTs at 70 °C exceeded that of 1-NorOTs. The fragmentation of 1-NorOCCl to the ion pair was 3 × 10¹⁵ times faster than the acetolysis of 1-NorOTs. The activation energies were measured as 9.0 kcal/mol (log A = 11.2 s^-1) for the fragmentation of 1-NorOCCl and 4.4 kcal/mol (log A = 8.44 s^-1) for that of 1-BcoOCCl both in DCE. B3LYP/6-31G* computed activation energies in simulated DCE were 14.6 and 2.7 kcal/mol, respectively. Copyright

Full text of DOI:10.1021/ja020002e

Published on Web 04/18/2002
Bridgehead Carbocations via Carbene Fragmentation: Erasing a 1010 Kinetic
Preference
Robert A. Moss,*,† Fengmei Zheng,† Jean-Marie Fede´,† Yan Ma,† Ronald R. Sauers,*,†
John P. Toscano,‡ and Brett M. Showalter‡
Department of Chemistry and Chemical Biology, Rutgers, the State UniVersity of New Jersey,
New Brunswick, New Jersey 08903 and Department of Chemistry, Johns Hopkins UniVersity,
Baltimore, Maryland 21218
Received January 1, 2002; Revised Manuscript Received February 25, 2002
Typical bridgehead substrates form carbocation/anion ion pairs
very slowly due to the high energetic costs of increased strain
imposed on the cations which cannot become planar. The more
the structurally imposed departure from planarity, the greater the
activation energy associated with cation formation and the energy
of the resultant cation. Consequently, the relative solvolysis rates
of 1-substituted adamantyl, bicyclo[2.2.2]octyl, and norbornyl
tosylates (HOAc, 70 °C)1a or bromides (80% EtOH, 25 °C)1b span
an interval of 1010; cf.; 1-3.2-4
Precursors 8 and 9 were then converted to alkoxychlorodiazirines
4 by oxidation with 12% aqueous NaOCl.14 The diazirines were
characterized by NMR and UV spectroscopy.15
Photolysis of diazirines 4 in dichloroethane (DCE) at 25 °C (λ
> 320 nm) afforded only chlorides 7, which were identified
spectroscopically or by GC and GC-MS comparisons to authentic
samples.10,11,16 That chlorides 7 stem from ion pairs 6 is supported
by the competitive interception of the carbocations by methanol.6,17
Photolyses of diazirines 4 in MeOH-DCE mixtures led to three
products derived from ROCCl: RCl, ROH, and ROMe; compare
eq 2.
More recent studies afford gas phase ∆G differences of ∼27 and
∼8 kcal/mol, respectively, for the formation of the 1-norbornyl (1-
Nor) and 1-bicyclo[2.2.2]octyl (1-Bco) cations, relative to the
1-adamantyl (1-Ad) cation.5 Moreover, ∆G is proportional to log
ksolv for these and other bridgehead substrates.5
Products RCl and ROMe represent competition between chloride
return to R+ and MeOH capture of R+ in ion pair 6; both products
represent fragmentation of 5 (eq 2a). ROH, on the other hand,
represents MeOH capture of the carbene, ultimately followed by
acid-catalyzed hydrolysis with adventitious water18,19 (eq 2b).
Product distributions for eq 2 at several MeOH concentrations
appear in Tables S1-S3 (Supporting Information). We note the
following trends. (a) In pure MeOH, carbene trapping (ROH
formation) is in the order R ) 1-Nor (42%) > R ) 1-Bco (32%)
> R ) 1-Ad (9%); that is, the order of R+ energies.1,5 The
competition between carbene capture (eq 2b) and carbene frag-
mentation (eq 2a) increasingly favors capture as the alternative
fragmentation generates a less stable carbocation. (b) At very low
mol fractions (0.01-0.02) of MeOH in DCE, ROH formation is
already substantial and remains relatively constant over the remain-
ing [MeOH] range. Presumably, the likelihood of increased ROH
yield due to increased carbene capture is balanced by an increasing
carbene fragmentation rate as the solvent polarity rises with
increasing [MeOH].20 (c) In pentane, where fragmentation is further
retarded by low solvent polarity,19 low concentrations of MeOH
elicit much larger yields of the ROH trapping products. Again, the
We reported that the fragmentation of alkoxychlorocarbenes
generates alkyl cations as components of ion pairs; eq 1.6
Experimentally7 and computationally,8,9 Ea’s for these reactions are
quite low (<10 kcal/mol). Here, we demonstrate that the low
activation energy of carbene fragmentation permits the ready
generation of bridgehead carbocations within ion pairs and obviates
the classical 1010 kinetic difference between the solvolysis of
1-Nor-X and 1-Ad-X. Most impressively, the fragmentation of
1-NorOCCl to 6 (R+ ) 1-Nor+) at 25 °C occurs 3 × 1015 times
faster than the acetolysis of 1-NorOTs at 70 °C.3
1-Norbornanol10 and 1-bicyclo[2.2.2]octanol11 were converted12
to isouronium triflates 8, whereas 1-adamantanol was transformed
to the N-isourea mesylate 9 by sequential reaction with KH/BrCN
(to 1-AdOCN), hydroxylamine, and MeSO2Cl.13
yields of ROH follow the “instability” order of R+: at øMeOH
)
0.02, 1-Nor (82%) > 1-Bco (43%) > 1-Ad (19%). (d) The product
ratios of ROMe/RCl increase with increasing [MeOH] for each
carbene, but even in pure MeOH, substantial quantities of RCl form
via ion pair return within ion pair 6.21
* To whom correspondence should be addressed. E-mail: moss@
rutchem.rutgers.edu.
† Rutgers University.
‡ Johns Hopkins University.
9
5258
J. AM. CHEM. SOC. 2002, 124, 5258-5259
10.1021/ja020002e CCC: $22.00 © 2002 American Chemical Society

C O M M U N I C A T I O N S
For ambiphilic ROCCl, ktrap is probably ∼104 s-1 in pure
methanol.22 The corresponding (ROH)/(RCl + ROMe)23 distribu-
tions, which reflect ktrap/kfrag, indicate that the kfrag values for 5 fall
in the anticipated order, 1-Nor < 1-Bco < 1-Ad, but the differences
between kfrag must be rather small. To quantitate this conclusion,
we measured kfrag directly by laser flash photolysis (LFP).6,7,9
LFP at 351 nm and 25 °C of diazirine 4 (R ) 1-Nor) in DCE
(A356 ) 0.5) in the presence of pyridine produced an ylide absorption
at 416 nm due to pyridine trapping of ROCCl. A correlation of the
apparent rate constants for ylide formation (kobs ) 1.6-5.0 × 105
s-1) versus [pyridine] (1.65-7.42 M) was linear (8 points, r )
0.997) with a slope of 5.47 × 104 M-1 s-1, equivalent to the rate
constant for ylide formation (ky), and a Y-intercept of 2.91 × 104
s-1, equivalent to kfrag for 1-NorOCCl;7,24 see Supporting Informa-
tion, Figure S-1. Repetition afforded kfrag ) 3.78 × 104 s-1 (ky )
5.46 × 104 M-1 s-1), leading to an average kfrag ) 3.3 ( 0.4 × 104
s-1. Similarly, we measured kfrag ) 1.5 ( 0.2 × 105 s-1 for
1-BcoOCCl (two runs). One example (kfrag ) 1.34 × 105 s-1, r )
0.998 for seven points) is included in the Supporting Information
as Figure S-2.
kinetic range for 1-norbornyl, 1-bicylo[2.2.2]octyl, and 1-adaman-
tyloxychlorocarbene fragmentations.
Acknowledgment. We are grateful to the National Science
Foundation for financial support, and to the center for Computa-
tional Neuroscience of Rutgers University (Newark) for computa-
tional support. J.P.T. acknowledges NSF Faculty Early Career
Development and Camille Dreyfus Teacher-Scholar Awards.
Supporting Information Available: Tables of product distribu-
tions; kinetics and Arrhenius correlations; ground- and transition-state
geometries for 1-NorOCCl; thermodynamic data for ground states and
transition states of 1-3 (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) (a) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper and Row: New York, 1987; pp 391-393. (b)
Carey, F. A.; Sundberg, R. J. AdVanced Organic Chemistry, Part A:
Structure and Mechanism, 4th ed.; Kluwer Academic/Plenum: New York,
2000; pp 287-289.
(2) krel (X ) OTs):1a 1 (10-10), 2 (10-4), 3 (1.0); (X ) Br):1b 1 (10-10), 2
(10-3), 3 (1.0).
kfrag was determined for 1-AdOCl by LFP of 4 (R ) 1-Ad, A355
(3) Bingham, R. C.; Schleyer, P.v. R. J. Am. Chem. Soc. 1971, 93, 3189.
(4) Review: Fort, R. C., Jr. In Carbonium Ions; Olah, G. A., Schleyer, P. v.
R., Eds.; Wiley: New York, 1973; Vol. IV, Chapter 32.
(5) Abboud, J.-L. M.; Herreros, M.; Notario, R.; Lomas, J. S.; Mareda, J.;
Mu¨ller, P.; Rossier, J.-C. J. Org. Chem. 1999, 64, 6401. Abboud, J.-L.
M.; Castano, O.; Della, E. W.; Herreros, M.; Mu¨ller, P.; Notario, R.;
Rossier, J. C. J. Am. Chem. Soc. 1997, 119, 2262.
) 0.4) with time-resolved infrared (TRIR) monitoring of the CO
formed as the carbene fragmented.7,25 We thus obtained kfrag ) 5.9
× 105 s-1 in DCE. The time dependence of CO formation at 2132
cm-1 is illustrated in the Supporting Information, Figure S-3.
kfrag for ROCCl does increase with decreasing strain energy of
the derived carbocation, but the dependence is very compressed:
1-NorOCCl, 3.3 × 104 s-1 < 1-BcoOCCl, 1.5 × 105 s-1 < 1-Ad-
OCCl (5.9 × 105 s-1). One must be wary of small differences
between kfrag values obtained by variant methodologies,7,9 but there
is no doubt that the range of kfrag values for these bridgehead ROCCl
is small. In particular, the 1010 spread of ksolv for 1-3 (X ) OTs)
is reduced to ∼18 for kfrag when X ) OCCl.
(6) Moss, R. A. Acc. Chem. Res. 1999, 32, 969.
(7) Moss, R. A.; Johnson, L. A.; Yan, S.; Toscano, J. P.; Showalter, B. M. J.
Am. Chem. Soc. 2000, 122, 11256.
(8) Yan, S.; Sauers, R. R.; Moss, R. A. Org. Lett. 1999, 1, 1603.
(9) Moss, R. A.; Zheng, F.; Sauers, R. R.; Toscano, J. P. J. Am. Chem. Soc.
2001, 123, 8109.
(10) Bixler, R. L.; Niemann, C. J. Org. Chem. 1958, 23, 742. Della, E. W.;
Cotsaris, E.; Hine, P. T. J. Am. Chem. Soc. 1981, 103, 4131.
(11) Kostova, K.; Dimitrov, V. Synth. Commun. 1995, 25, 1575. Bixler and
Niemann, ref 10. Wiberg, K. B.; Lowry, B. R. J. Am. Chem. Soc. 1963,
85, 3188.
This enormous compression of kfrag must reflect low activation
energies and early transition states for the 5 f 6 transformation;
eq 1. We determined Ea ) 9.0 ( 0.2 kcal/mol, log A ) 11.2 ( 0.1
s-1 (two experiments) for the fragmentation of 1-NorOCCl in DCE.
Rate constants were measured from 0 to 50 °C (r ) 0.996 for nine
points); an example appears as Figure S-4 in the Supporting
Information. Similarly, Ea for the fragmentation of 1-BcoOCCl in
DCE was determined as 4.4 kcal/mol, log A ) 8.44 s-1 (r ) 0.987,
five points).
The 9 kcal/mol activation energy for the fragmentation of
1-NorOCCl is the highest yet measured for this type of reaction.6,7
Ea for the fragmentation of 1-NorOCCl is, as expected, larger than
that for 1-BcoOCCl, but the difference in log A (which favors
1-NorOCCl) narrows the difference in kfrag to a factor of only 4.5
at 25 °C.
Using B3LYP/6-31G* methodology,26a in analogy to our pre-
vious studies,7-9 we computed Ea’s for the fragmentations of
carbenes 1 - 3 (X ) OCCl) as differences between ground- and
transition-state energies.26b We obtained Ea (kcal/mol, gas phase):
1-NorOCCl, 25.2; 1-BcoOCCl, 10.8, 1-AdOCCl, 7.4; and in
simulated DCE: 1-NorOCCl, 14.6; 1-BcoOCCl, 2.2; 1-AdOCCl,
-0.95.
The computed Ea’s are in the appropriate order, and the DCE
values are in reasonable agreement with experiment for 1-NorOCCl
and 1-BcoOCCl. Computed ground-state and transition-state ge-
ometries for 1-NorOCCl appear as Figure S-5 in the Supporting
Information.
(12) Moss, R. A.; Kaczmarczyk, G. M.; Johnson, L. A. Synth. Commun. 2000,
30, 3233.
(13) Moss, R. A.; Perez, L. A.; Wlostowska, J.; Guo, W.; Krogh-Jespersen,
K. J. Org. Chem. 1982, 47, 4177. Grigat, E.; Pu¨tter, R.; Konig, C. Chem.
Ber. 1965, 98, 144.
(14) Graham, W. H. J. Am. Chem. Soc. 1965, 87, 4396.
(15) λmax (nm) for 4: R ) 1-Nor, 350, 364 (DCE); R ) 1-Bco, 353, 366 (DCE);
R ) 1-Ad, 353, 370 (pentane).
(16) About 10% of 1-bicyclo[3.2.1]octyl chloride was also formed in the
experiments involving 4 and 5 (R ) 1-Bco) due to contamination.11
(17) Moss, R. A.; Balcerzak, P. J. Am. Chem. Soc. 1992, 114, 9386. Moss, R.
A.; Ho, G. J.; Wilk, B. K. Tetrahedron Lett. 1989, 30, 2473.
(18) Smith, N. P.; Stevens, I. D. R. J. Chem. Soc. Perkin Trans. II 1979, 1298.
(19) Moss, R. A.; Wilk, B. K.; Hadel, L. M. Tetrahedron Lett. 1987, 28, 1969.
(20) The dielectric constants of MeOH and DCE are 32.6 and 10.36,
respectively. MeOH may also catalyze the fragmentation of 5 by specific
solvation of the chloride leaving group; cf.; the solvolysis of 1-AdCl:
Gajewski, J. J. J. Am. Chem. Soc. 2001, 123, 10877.
(21) Adamantene is not a precursor of 1-AdOMe in the fragmentation and
methanolysis of 1-AdOCCl; photolysis of 4 (R ) 1-Ad) in MeOD
produces 1-AdOMe containing no deuterium (GC-MS). 1-Norbornene
is similarly excluded as an intermediate in the fragmentation/methanolysis
of 1-NorOCCl.
(22) Du, X.-M.; Fan, H.; Goodman, J. L.; Kesselmayer, M.A.; Krogh-Jespersen,
K.; LaVilla, J. A.; Moss, R. A.; Shen, S.; Sheridan, R. S. J. Am. Chem.
Soc. 1990, 112, 1920.
(23) (ROH)/(RCl + ROMe) ) 0.71 (1-Nor), 0.48 (1-Bco), 0.096 (1-Ad).
(24) For this methodology, see: (a) Jackson, J. E.; Soundararajan, N.; Platz,
M. S.; Liu, M. T. H. J. Am. Chem.. Soc. 1988, 110, 5595. (b) Moss, R.
A.; Ge, C.-S.; Maksimovic, L. J. Am. Chem. Soc. 1996, 118, 9792.
(25) Wang, Y.; Yuzawa, T.; Hamaguchi, H.; Toscano, J. P. J. Am. Chem. Soc.
1999, 121, 2875 and references therein. Toscano, J. P. AdV. Photochem.
2000, 26, 41. For comparisons of TRIR and LFP-UV kfrag data, see ref 9.
(26) (a) Gaussian 98, revision A.7; Gaussian, Inc.: Pittsburgh, PA, 1998. DFT
calculations used Becke’s three-parameter hybrid method with the LYP
correlation functional: Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b)
Single-point calculations were based on fully optimized gas-phase
geometries with ZPE and thermal corrections from B3LYP/6-31G* results.
Simulated DCE solvent calculations employed ∈ ) 10.36, and the pcm
and scipcm protocols.
In conclusion, fragmentations of ROCCl readily afford bridge-
head carbocations as [R+ OC Cl-] ion pairs. The low activation
energies associated with these reactions dramatically compress the
JA020002E
9
J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5259