Solvent-equilibrated ion pairs from carbene fragmentation reactions

Article abstract of DOI:10.1021/ja040146o

[R⁺ OC Cl^-] ion pairs were generated in methanol/dichloroethane solutions, with R⁺ as the 1-bicyclo[2.2.2]octyl, 1-adamantyl, or 3-homoadamantyl cation. Ion pairs were produced either by the direct fragmentation of alkoxychlorocarbenes (ROCCl), with R = 1-bicyclo[2.2.2]octyl, 1-adamantyl, or 3-homoadamantyl, or by the ring expansion-fragmentation of R′CH₂OCCl, with R′ = 1-norbornyl, 3-noradamantyl, or 1-adamantyl. Correlations of the [ROMe]/[RCl] product ratios as a function of the mole fraction of MeOH in dichloroethane showed that the homoadamantyl chloride ion pairs, produced by either the direct or ring expansion-fragmentations, were identical, solvent- and anion -equilibrated, and precursor independent. Laser flash photolysis experiments gave 20-30 ps as the time required for solvent equilibration and precursor independence. Methanol/chloride selectivities of the (less-stable) 1-adamantyl chloride and 1-bicyclo[2.2.2]octyl chloride ion pairs were not independent of their ROCCl or R′CH₂OCCl precursors. Computational studies provided transition states for the fragmentations and for the structures of the ion pairs.

Full text of DOI:10.1021/ja040146o

Published on Web 09/04/2004
Solvent-Equilibrated Ion Pairs from Carbene
Fragmentation Reactions
Robert A. Moss,* Fengmei Zheng, Jean-Marie Fed e´ , Lauren A. Johnson, and
Ronald R. Sauers*
Contribution from the Department of Chemistry and Chemical Biology, Rutgers,
The State UniVersity of New Jersey, New Brunswick, New Jersey 08903
Received May 24, 2004; E-mail: moss@rutchem.rutgers.edu
Abstract: [R+ OC Cl ] ion pairs were generated in methanol/dichloroethane solutions, with R as the
-bicyclo[2.2.2]octyl, 1-adamantyl, or 3-homoadamantyl cation. Ion pairs were produced either by the direct
fragmentation of alkoxychlorocarbenes (ROCCl), with R ) 1-bicyclo[2.2.2]octyl, 1-adamantyl, or 3-ho-
moadamantyl, or by the ring expansion-fragmentation of R′CH OCCl, with R′ ) 1-norbornyl, 3-norada-
-
+
1
2
mantyl, or 1-adamantyl. Correlations of the [ROMe]/[RCl] product ratios as a function of the mole fraction
of MeOH in dichloroethane showed that the homoadamantyl chloride ion pairs, produced by either the
direct or ring expansion-fragmentations, were identical, solvent- and anion-equilibrated, and precursor
independent. Laser flash photolysis experiments gave 20-30 ps as the time required for solvent equilibration
and precursor independence. Methanol/chloride selectivities of the (less-stable) 1-adamantyl chloride and
2
1-bicyclo[2.2.2]octyl chloride ion pairs were not independent of their ROCCl or R′CH OCCl precursors.
Computational studies provided transition states for the fragmentations and for the structures of the ion
pairs.
Winstein’s dissection1 of the intermediates of solvolysis
reactions into intimate ion pairs, solvent-separated ion pairs,
and solvated ions has stood the test of time. In some cases,
identify solvent-equilibrated ion pairs. Second, we use laser flash
photolytic (LFP) methods to examine the dynamics of the ion
pairs.
2
The fragmentations of certain alkoxychlorocarbenes (1) in
polar solvents appear to proceed via ion pairs (2). In nucleophilic
solvents such as methanol, these decay either by collapse to
picosecond laser flash photolysis, coupled with the deconvo-
lution of complex kinetic behavior, permits the assignment of
lifetimes to the several species and of rate constants to their
interconversions. For example, the intimate (or contact) diphe-
nylmethyl cation-chloride ion pair evolves to the solvent-
4
RCl or by solvolysis to ROMe (eq 1). However, the fragmenta-
tion of ROCCl is a complicated process, capable of adopting
different mechanisms depending on the structure of R and the
polarity of the solvent.
9
-1
separated ion pair in MeCN with k ) 2.87 × 10 s , in
competition with internal return (collapse) to Ph2CHCl with k
9
-1 3
)
3.81 × 10 s . The lifetime of the contact ion pair is 150
3
ps. Rate constants governing the dynamics of the solvent-
separated ion pair are also available.
3
We have encountered concerted fragmentation in nonpolar
5
5
A related problem is the time required for the initial ion pair
to become equilibrated with its solvent cage, thus permitting
the counterion and solvent molecules to compete for the cation
on the basis of comparative nucleophilicity and local concentra-
tion, and without specific ‘memory’ of the original precursor’s
structure. Here, we wish to describe a new method for the
estimation of the time required for an ion pair to become solvent-
equilibrated. Our method, which is based on carbene fragmenta-
tion chemistry, has two parts: First, we use product studies to
solvents, radical fragmentation in cryogenic matrixes or in the
6
7
gas phase, and radical-like or anion-mediated SN2 fragmenta-
tions when R cannot readily support a positive change.
8
Now, we report that analyses of the product distributions from
‘paired’ fragmentations, in which nominally identical ion pairs
form either directly or by ring expansion, allow us to identify
those ion pairs that equilibrate with their solvent shell. Lifetime
measurements then permit an estimate of the time required for
this process. These results can also be associated with lifetimes
1
,2
of the ion pairs in the Winstein Scheme.
(
1) Winstein, S.; Clippinger, E.; Fainberg, A. H.; Heck, R.; Robinson, G. C.
J. Am. Chem. Soc. 1956, 78, 328.
(4) Moss, R. A. Acc. Chem. Res. 1999, 32, 969.
(5) Moss, R. A.; Ma, Y.; Zheng, F.; Sauers, R. R.; Bally, T.; Maltsev, A.;
Toscano, J. P.; Showalter, B. M. J. Phys. Chem. A 2002, 106, 12280.
(6) Blake, M. E.; Jones, M., Jr.; Zheng, F.; Moss, R. A. Tetrahedron Lett.
2002, 43, 3069.
(
2) (a) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; New York: Harper & Row: 1987; pp 345f. (b) Carey,
F. A.; Sundberg, R. J. AdVanced Organic Chemistry, Part A: Structure
th
and Mechanisms, 4 ed.; New York: Kluwer Academic/Plenum: 2000,
(7) Moss, R. A.; Ma, Y.; Sauers, R. R. J. Am. Chem. Soc. 2002, 124, 13968.
(8) Moss, R. A.; Johnson, L. A.; Merrer, D. C.; Lee, G. E., Jr. J. Am. Chem.
Soc. 1999, 121, 5940.
pp 269f.
(
3) Peters, K. S.; Li, B. J. Phys. Chem. 1994, 98, 401.
10.1021/ja040146o CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 12421-12431
9
12421

A R T I C L E S
Moss et al.
Bridgehead carbocations can be generated directly by frag-
mentation of bridgehead ROCCl, or by ring-expanding rear-
Scheme 1
rangement-fragmentation of appropriate RCH2OCCl (cf., eqs
9
2
and 3). When the fragmentations of 1-norbornylmethoxy-
3
-homoadamantanol,14 which was converted to 3-homoadaman-
15
tyl isouronium triflate with cyanamide and triflic acid. The
latter was oxidized to 3-(3-homoadamantyloxy)-3-chlorodiaz-
irine with NaOCl. The sequence is outlined in Scheme 1, and
it was followed for all the other diazirine precursors, except
1
6
1
2
for that of carbene 7.
Each carbene was photolytically generated from the appropri-
ate 3-alkoxy-3-chlorodiazirine (λ > 320 nm, 25 °C) in MeOH/
DCE at various mole fractions of MeOH. Products were
identified by GC-MS, NMR, or GC comparisons with authentic
samples. [ROMe]/[RCl] product ratios were determined by
capillary GC.
chlorocarbene (3) and 1-bicyclo[2.2.2]octyloxychlorocarbene (4)
occur in pure methanol, the ether/chloride product ratios
9
[5-OMe]/[5-Cl] are not the same: 0.39 from 3 vs 2.5 from 4.
From 1-norbornylmethoxychlorocarbene (3), we obtained the
Clearly, identical, solvent-equilibrated ion pairs are not generated
from both carbenes. Similarly, in pure methanol, 3-noradaman-
tylmethoxychlorocarbene (6) or 1-adamantyloxychlorocarbene
9
products shown in Scheme 2. In pure MeOH, a typical product
distribution (%) was 5-OMe (10), 5-Cl (26), A (1.3), B (3.1),
C (30), D (14), and E (15); the [5-OMe]/[5-Cl] ratio was 0.39
(
6
7) afford different [8-OMe]/[8-Cl] product ratios: 0.43 from
(
0.01 (mean of 2 runs). The simplest rationale for products
vs 0.67 from 7.9
5
-OMe and 5-Cl posits intermediate ion pair 12, which arises
Can these differences be further narrowed or eliminated? The
by ring expansion-fragmentation of carbene 3. The ion pair
stability of bridgehead carbocations varies inversely with the
structurally enforced deviation from planarity at the carbocation
center.10 The relative gas-phase free energies of the 1-bicyclo-
-
then gives the products by collapse with Cl or capture by
methanol. Below, we will refine our conception of the genesis
of these products.
[
-
2.2.2.]octyl, 1-adamantyl, and 3-homoadamantyl cations are
7.4, 0.0, and 4.2 kcal/mol, respectively.10a,11 The solution
lifetimes of ion pairs 2 should be directly related to the stability
+
+
of R . Therefore, altering R from 1-adamantyl to the more
stable 3-homoadamantyl cation should extend the ion pair
lifetime. Here, we report that ion pairs with identical MeOH/
-
C1 selectivities, across the entire range of methanol-1,2-
dichlorethane (DCE) compositions, form from either 1-ada-
mantylmethoxychlorocarbene (9) or 3-homo-1-adamantyl-
methoxychlorocarbene (10) (eq 4).
Products D and E similarly arise from ion pair 13, which
stems from the other possible ring expansion-fragmentation
of 3. As previously noted, expansion of the shorter, 1-carbon
bridge of 3 exceeds expansion of the longer, 2-carbon bridge.9
Products A-C retain the structure of the initial carbene. The
products A and B represent attack by methanol or chloride at
the methylene group of 3, whereas alcohol C is most reasonably
construed as a carbene-trapping product: methanol captures
about 30% of 3, producing ROCH(Cl)OMe, which yields
orthoformate ROCH(OMe)2 and HCl by methanolysis. The
orthoformate cannot be isolated; it returns alcohol C by acid-
Results and Discussion
Precursors and Products. The diazirine precursors for
9
12
9
12
13
carbenes 3, 4, 6, 7, and 9 were available from previous
studies. The precursor for carbene 10 was prepared from
1
2,17
catalyzed hydrolysis with adventitious water.
Identification of products A-E is described by Moss et al.9
Authentic samples of key products 5-OMe and 5-Cl were
obtained as follows. Chloride 5-Cl was prepared from 1-nor-
(
9) Moss, R. A., Zheng, F.; Fed e´ , J.-M.; Sauers, R. R. Organic Lett. 2002, 4,
2
341.
(
10) (a) Abboud, J.-L. M.; Alkorta, I.; Davalos, J. Z.; M u¨ ller, P.; Quintanilla,
E.; Rossier, J.-C. J. Org. Chem. 2003, 68, 3786. (b) Abboud, J.-L. M.;
Herreros, M.; Notario, R.; Lomas, J. S.; Mareda, J.; M u¨ ller, P.; Rossier,
J.-C. J. Org. Chem. 1999, 64, 6401. (c) Abboud, J.-L. M.; Alkorta, I.;
Davalos, J. Z.; M u¨ ller, P.; Quintanilla, E. AdV. Phys. Org. Chem. 2002,
18
19
bornylmethanol, concentrated HCl, and ZnCl2, which gave
an 85:15 mixture of 5-Cl and E. Ether 5-OMe was obtained
3
7, 57.
(
11) The values correspond to the relative free energies associated with chloride
exchange between the 1-adamantyl cation and carbocation R+ in the gas
phase, and afford a quantitative ranking of the intrinsic gas-phase stabilities
(14) Farcasiu, D.; J a¨ hme, J. Ruchardt, C. J. Am. Chem. Soc. 1985, 107, 5717.
(15) Moss, R. A.; Kaczmarczyk, G. M.; Johnson, LA. Synth. Commun. 2000,
30, 3233.
(16) Graham, W. H. J. Am. Chem. Soc. 1965, 87, 4396.
(17) Smith, N. P.; Stevens, I. D. R. J. Chem. Soc., Perkin Trans. 2 1979, 1298.
See, especially, pp 1302-1303.
(18) Bixler, R. L.; Niemann, C. J. Org. Chem. 1958, 23, 742.
(19) Kostova, K.; Dimitrov, V. Synth. Commun. 1995, 25, 1575.
1
0
of the carbocations.
(
12) Moss, R. A.; Zheng. F., Fed e´ , J.-M.; Ma, Y.; Sauers, R. R.; Toscano, J. P.;
Showalter, B. M. J. Am. Chem. Soc. 2002, 124, 5258.
(13) Moss, R. A.; Johnson, L. A.; Yan, S.; Toscano, J. P.; Showalter, B. M. J.
Am. Chem. Soc. 2000, 122, 11256.
12422 J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004

Ion Pairs from Carbene Fragmentation Reactions
A R T I C L E S
Scheme 2
methanol mole fractions are collected in Table S2 of the
Supporting Information.
From 3-noradamantyl methoxychlorocarbene (6), we obtained
9
the products shown in Scheme 4, which parallel those obtained
from carbene 3 (cf., Scheme 2). In pure methanol, a typical
product distribution (%) was 8-OMe (16), 8-Cl (35), G (12),
H (9), I (1), J (25), K (2), while the [8-OMe]/[8-Cl] ratio was
0
.43 ( 0.03, (mean of 2 runs).
We suggest that products 8-OMe and 8-Cl come from ion
pair 14, formed by ring expansion-fragmentation of carbene
. The ion pair either collapses with its chloride counterion or
6
is captured by methanol. Protoadamantyl products G and H can
be similarly rationalized, invoking ion pair 15, which is formed
by the alternative ring expansion-fragmentation of carbene 6.
As in the case of carbene 3, ring expansion of the shorter
(
0-carbon) bridge of 6 exceeds expansion of its longer (1-carbon)
Figure 1. Product ratio [5-OMe]/[5-Cl] from the fragmentations of carbenes
9
bridge.
3
(Scheme 2) and 4 (Scheme 3) in MeOH/DCE vs øMeOH (cf.; Tables S1
and S2 in the Supporting Information). Key: 0 from carbene 3, [ from
carbene 4.
Scheme 3
Product J represents the methanol trapping product of carbene
, ROCH(Cl)OMe, with subsequent methanolysis and hydroly-
6
by the AgClO4-mediated methanolysis of 1-norbornylmethyl
iodide.20 Comparisons by GC, NMR, and MS verified product
identities.
The dependence of the product ratio [5-OMe]/[5-Cl] on the
mole fraction of MeOH in DCE appears in Figure 1. Table S1
in the Supporting Information collects distributions for all of
the products in Scheme 2 at each of 11 MeOH mole fractions.
From 1-bicylo[2.2.2]octyloxychlorocarbene (4), we obtained
the products shown in Scheme 3.12 Products 5-OMe and 5-Cl
sis. The formation of J from carbene 6 parallels the formation
of alcohol C from carbene 3 (above). Identification of products
22
9
H, I, and J is described by Moss et al. The identity of the
protoadamantyl methyl ether (G) was assumed based on its MS
molecular ion. An authentic sample of 8-OMe was obtained
from the AgClO4-mediated methanolysis of 1-adamantyl bro-
20,23
mide,
whereas authentic 8-Cl was purchased from Aldrich.
The product assignments in Scheme 4 were verified by GC-
MS and GC spiking experiments.
stem from ion pair 12, which here arises directly by the
The dependence of the product ratio [8-OMe]/[8-Cl] on the
mole fraction of MeOH appears in Figure 2. Table S3 in the
Supporting Information details the distribution of all the products
of Scheme 4 at each of nine MeOH mole fractions in DCE.
From 1-adamantyloxychlorocarbene (7), we obtained the
fragmentation of carbene 4. The bicyclooctanol F1
8,19,21
may
be regarded as the methanol capture product of 4, which
furnishes the alcohol after methanolysis and hydrolysis (see
above, product C). In pure methanol, a typical distribution (%)
is 5-OMe (48), 5-Cl (19), and F (32), with [5-OMe]/[5-Cl] )
12
products illustrated in Scheme 5. Products 8-OMe and 8-Cl
2
.52 ( 0.04 (mean of 2 runs). The dependence of the product
ratio on the mole fraction of MeOH (in DCE) is shown in Figure
. Distributions of the products in Scheme 3 at each of eight
come from ion pair 14, here formed directly by fragmentation
of carbene 7. The 1-adamantanol (K) represents the methanol/
1
(
22) (a) Tae, E. L.; Zhu, Z.; Platz, M. S. J. Phys. Chem. A 2001, 105, 3803. (b)
Vogt, R. B.; Hoover, J. R. E. Tetrahedron Lett. 1967, 8, 2841. (c) Stoelting,
D. T.; Shiner, V. J., Jr. J. Am. Chem. Soc. 1993, 115, 1695.
(
20) Kropp, P. J.; Poindexter, G. S.; Pienta, N. J.; Hamilton, D. C. J. Am. Chem.
Soc. 1976, 98, 8135.
(21) Wiberg, K. B.; Lowry, B. R. J. Am. Chem. Soc. 1963, 85, 3188.
(23) Poindexter, G. S.; Kropp, P. J. J. Org. Chem. 1976, 41, 1215.
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12423
9

A R T I C L E S
Scheme 4
Moss et al.
expansion-fragmentation of carbene 9. Products M and N
represent capture of the carbene by methanol or water,24
respectively. Product L stems from chloride attack on carbene
2
4
9
at its a-carbon.
The identity of M was secured by comparison with an
authentic (commercial) sample. Chloride L was obtained from
2
5
alcohol M and SOCl2, while formate N was prepared from
M and formic acid, using BF3(MeOH)2 as a catalyst.26 For the
key products, 11-OMe and 11-Cl, 3-homoadamantanol was
first converted to 11-Cl with SOCl2 (reflux, 4 h).
1-Cl could then be converted to 11-OMe by refluxing with
NaOMe/MeOH for 10 h. Both 11-Cl and 11-OMe were
characterized by GC-MS, as well as H and C NMR
spectroscopy. The reaction products in Scheme 6 were then
verified by GC and GC-MS comparisons with the authentic
samples.
14
2
5,27
Chloride
1
1
13
The dependence of the product ratio [11-OMe]/[11-Cl] on
the mole fraction of methanol is illustrated in Figure 3, while
Table S5 in the Supporting Information collects the distributions
of all of the products of Scheme 6 at each of 11 methanol mole
fractions.
Figure 2. Product ratio [8-OMe]/[8-Cl] from the fragmentations of carbenes
6
(Scheme 4) and 7 (Scheme 5) in MeOH/DCE vs øMeOH (cf., Tables S3
and S4 in the Supporting Information). Key: 0 from carbene 6, [ from
carbene 7.
From 3-homoadamantyloxychlorocarbene (10), we obtained
the products shown in Scheme 7. Products 11-OMe and 11-Cl
arise from ion pair 16, formed here by the direct fragmentation
carbene trapping-methanolysis/hydrolysis sequence that has
been described above.
In pure methanol, a typical product distribution (%) is 8-OMe
14
of carbene 10. The 3-homoadamantanol (O) derives from the
(36), 8-Cl (55), and K (9), with [8-OMe]/[8-Cl] ) 0.67 ( 0.01
methanol trapping/methanolysis/hydrolysis of carbene 10.
In pure methanol, a typical product distribution (%) is 11-
OMe (27), 11-Cl (36) and O (37), with [11-OMe]/[11-Cl] )
(mean of 2 runs). The dependence of the product ratio on the
mole fraction of MeOH (in DCE) is shown in Figure 2;
distributions of all of the products in Scheme 5, at each of 11
MeOH mole fractions, are collected in Table S4 of the
Supporting Information.
From 1-adamantylmethoxychlorocarbene (9)13 we obtained
the products shown in Scheme 6. In pure methanol, a typical
product distribution (%) was 11-OMe (23), 11-Cl (35), L (not
detected), M (42), and N (not detected). In pure DCE, we
observed a distribution (%) of L (7.6), N (5.9), and 11-Cl (86.5).
In MeOH, the ratio [11-OMe]/[11-Cl] ) 0.67 ( 0.01 (mean of
0
.75 ( 0.01 (mean of 2 runs). The dependence of the product
ratio on the mole fraction of methanol is shown in Figure 3,
while distributions of the products in Scheme 7 at each of eight
methanol mole fractions appear in Table S6 of the Supporting
Information.
Dependence of Product Distributions on øMeOH. Inspection
of Figures 1-3 reveals a dramatic evolution in correlations of
the product ratio [ROMe]/[RCl] with the mole fraction of
methanol in the DCE solvent. Consider first eq 2, Schemes 2
and 3, and Figure 1. Here, [ROMe]/[RCl] is strongly precursor
dependent. From the fragmentation of carbene 4 (Scheme 3),
2
runs).
Products 11-Cl and 11-OMe are attributed to collapse or
methanol capture of ion pair 16, which arises from ring
[5-OMe]/[5-Cl] increases with øMeOH (Figure 1). This is in
keeping with an ion pair intermediate, such as 12, which can
select between MeOH capture or chloride collapse, and which
is responsive to methanol concentration. However, from the ring
(
24) Moss, R. A.; Ge, C.-S.; Maksimovic, L. J. Am. Chem. Soc. 1996, 118,
9792.
(25) Margosian, D.; Speier, J.; Kovacic, P. J. Org. Chem. 1981, 46, 136.
(26) Dymicky, M. Org. Prep. Proced. Int. 1982, 14, 177.
(27) Stetter, H.; Goebel, P. Chem. Ber. 1963, 96, 550.
12424 J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004

Ion Pairs from Carbene Fragmentation Reactions
A R T I C L E S
Scheme 5
Scheme 6
al.29 provided experimental and computational evidence for an
SNi-ion pair continuum in the decomposition of alkyl chloro-
sufites, a reaction analogous to the fragmentation of alkoxy-
chlorocarbenes.
In the 1-adamantyl ion pairs (14) obtained in the fragmenta-
tions of carbenes 6 and 7 (Schemes 4 and 5), the cation is more
1
0,11
stable
than the bicyclo[2.2.2]octyl cation obtained from
carbenes 3 and 4. Now, the ring expansion-fragmentation of
carbene 6 more closely approaches the direct fragmentation of
carbene 7 (Figure 2): [8-OMe]/[8-Cl] exhibits roughly parallel
dependences on øMeOH from either carbene, and the mean final
ratios in pure methanol, 0.43 from 6 and 0.67 from 7, are not
very widely separated.
Here, ion pair 14 is a reasonable qualitative representation
of the product-forming intermediate from either carbene, and it
is supported by computational studies (see below). The residual
differences in [8-OMe]/[8-Cl] as a function of precursor suggest
that the ion pairs, represented by 14, approach, but do not quite
achieve, equivalence. They cannot both be solvent equilibrated.
Figure 3. Product ratio [11-OMe]/[11-Cl] from the fragmentations of
carbenes 9 (Scheme 6) and 10 (Scheme 7) in MeOH/DCE vs øMeOH (cf.,
Tables S5 and S6 in the Supporting Information). Key: 0 from carbene 9,
[
from carbene 10.
Scheme 7
(It is also possible that one ion pair, presumably from direct
fragmentation of carbene 7, becomes solvent equilibrated,
whereas the ion pair from ring expansion-fragmentation of
carbene 6 does not reach that state.)
The even-more stable10,11 ion pairs 16, however, achieve
equivalence: within experimental error ((0.04), the [11-OMe]/
expansion-fragmentation of carbene 3 (Scheme 2), relatively
little 5-OMe forms; the dependence of [5-OMe]/[5-Cl] on øMeOH
is nearly flat when øMeOH > 0.1. At øMeOH ) 1.0, [5-OMe]/[5-
Cl] is only ∼0.4, compared with ∼2.5 from the direct
fragmentation of carbene 4. Clearly, the same ion pair (12) does
not arise from both carbenes 3 and 4, and the product-forming
intermediate from 3 is certainly not a solvent-equilibrated ion
pair.
In fact, computational studies (below) show that the conver-
sion of carbene 3 to 5-Cl transits a “tight” transition state that
approximates the SNi mechanism. It is distinct from the ion pair
arising in the fragmentation of carbene 4, which more closely
resembles the qualitative structure 12.
[11-Cl] ratios are very similar across the range of methanol mole
fractions, and they are independent of their origin in either ring
expansion-fragmentation of carbene 9 or direct fragmentation
of carbene 10, that is, “precursor amnesia” (cf., eq 4, Schemes
6
and 7, and Figure 3). These results constitute very strong
evidence for counterion- and solvent-equilibrated ion pair 16
as the product determining intermediate in eq 4.
Computational Studies. A more-detailed structure for ion
30
pair 16 is provided by DFT methodology. We first computed
the transition state for the fragmentation of 1-AdCH2OCCl (9)
(
cf., Figure 4a).
We suggested,13,24 that the 1-AdCH2OCCl fragmentation
involved concerted ring expansion of the 1-AdCH2 unit directly
to the homoadamantyl cation. The B3LYP/6-31G(d) transition
state (Figure 4a) supports this idea. Comparison of the ground-
state bond lengths (in parentheses) with the corresponding
transition-state bond lengths shows that the 1,2-carbon migration
is in progress as the carbene fragments. Thus, the separation
Only a small component of the ring expansion-fragmentation
of carbene 3 accesses ion pair 12 and ultimately yields 5-OMe.
Most of the fragmentation leads to 5-Cl by a lower-energy, more
“
concerted” SNi-like transition state (see Figure 8a, below). We
28
note that Jones et al. suggested that certain alkoxychlorocar-
benes fragment via SNi-like transition states, while Schleyer et
(
29) (a) Schreiner, P. R.; Schleyer, P.v. R.; Hill, R. K. J. Org. Chem. 1993, 58,
2822. (b) Schreiner, P. R.; Schleyer, P.v. R.; Hill, R. K. J. Org. Chem.
1994, 59 1849.
(
28) Likhotvorik, I. R.; Jones, M., Jr.; Yurchenko, A. G.; Krasutsky, P.
Tetrahedron Lett. 1989, 30, 5089.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004 12425

A R T I C L E S
Moss et al.
Figure 4. (a) B3LYP/6-31G(d) transition state for the ring expansion-fragmentation of carbene 9. Bond lengths are shown for the TS and the carbene (in
parentheses). ZPE and thermal corrections to 298 K have been applied. (b) Optimized ion pair from an intrinsic reaction coordinate calculation on the TS
shown in (a), using B3LYP/6-31G(d) in simulated methanol with the polarized contiunuum model (PCM).30 See the Computational Methodology section in
the Experimental Section.
between the migrating CH2 unit and the CH2 migration terminus
decreases from 2.479 to 2.143 Å, while the (breaking) bond
between the migrant CH2 and the tertiary carbon (C-3), which
will become the cationic carbon in the homoadamantyl cation,
increases from 1.552 to 1.658 Å. At the same time, the
fragmenting C-O and C-Cl bonds increase from 1.489 to 2.234
Å and from 1.888 to 2.727 Å, respectively.
We believe this species to be a good approximation of the
solvent-equilibrated ion pair formed by the ring expansion-
fragmentation of carbene 9, which was represented qualitatively
3
1
as structure 16 (above). Moreover, given the small separation
of the cation and chloride anion (4.18 Å in Figure 4b), we
associate this ion pair with the intimate or contact ion pair of
1
,2
the Winstein solvolysis scheme.
The transition state (TS) was then subjected to an intrinsic
reaction coordinate (IRC) calculation affording an incipient ion
pair (associated with CO) that was optimized in simulated
methanol using the polarized continuum model (PCM). The
resulting ion pair, shown in Figure 4b, displays apparent
H-bonding between the chloride anion and protons on the carbon
atoms that flank the cationic center. This feature undoubtedly
contributes to the ion pair’s existence in simulated methanol.
Analogous computational studies of the fragmentation of
3-homoadamantyloxychlorocarbene (10) led to the TS depicted
in Figure 5a. An IRC calculation afforded a structure that, when
optimized in vacuo, gave homoadamantene by net elimination
of HCl. However, no homoadamantene was detected in our
solution experiments, and optimization of the computed transi-
tion structure in simulated methanol by the PCM procedure gave
an ion pair (Figure 5b) that was essentially identical with that
obtained from carbene 9 (Figure 4b). The computational studies
thus support a mechanism in which solvent-equilibrated ion pairs
(
30) (a) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pommelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.;
Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; 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. G.; Chen, W.; Wong, M.
W.; Andres, J. L.; Gonzales, C.; Head-Gordon, M.; Repogle, E. S.; Pople,
J. A. Gaussian98, Revision A9; Gaussian, Inc.: Pittsburgh, 1998. 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. Gaussian03, Revision
B.02; Gaussian, Inc.: Pittsburgh, 2003. (b)Becke, A. D. J. Chem. Phys.
1
6 form from either carbene 9 or 10, and display the identical
-
Cl /MeOH selectivity illustrated in Figure 3.
Computational analyses of the fragmentations of 3-norada-
mantylmethoxychlorocarbene (6) and 1-adamantyloxychloro-
carbene (7) were similarly performed at the B3LYP/6-31G(d)
level, affording the transition states depicted in Figure S1
(Supporting Information). For either carbene, IRC calculations,
followed by optimization in simulated methanol (PCM), gave
the same 1-adamantyl chloride ion pair (Figure 6). This species,
rendered qualitatively as 14 (see above), resembles the 3-ho-
moadamantyl chloride ion pair (16) obtained from carbenes 9
+
-
or 10 (Figures 4b and 5b); both have similar C /Cl separations,
and both feature H-bonding of the chloride ion and ring protons.
Given the similarity of the ion pairs formed from carbenes 9
-
and 10, or 6 and 7, why do the former display identical Cl /
MeOH selectivity (Figure 3), whereas the latter do not (Figure
2
)? ion pair lifetime must play the key role. Ion pair 16 from
(31) The computed E
a
for the fragmentation of 9 in simulated MeCN at the
1
993, 98, 5648. (c)Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37,
B3LYP/6-31G* level is 5.4 kcal/mol, in decent agreement with the
1
3
7
85.
measured value of 3.6 kcal/mol.
12426 J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004

Ion Pairs from Carbene Fragmentation Reactions
A R T I C L E S
Figure 5. (a) B3LYP/6-31G(d) transition state for the fragmentation of carbene 10. (b) Optimized ion pair from an IRC calculation on the TS shown in (a),
using B3LYP/6-31G(d) in simulated methanol with the PCM.30 See the Computational Methodology section in the Experimental Section.
Computational studies of 1-norbornylmethoxychlorocarbene
(3) and 1-bicyclo[2.2.2]oxychlorocarbene (4) are particularly
interesting. The TS for fragmentation of carbene 4 (Figure 7a)
leads, after IRC calculation and optimization in simulated
methanol (PCM), to the ion pair (12) shown in Figure 7b.
However, ring expansion-fragmentation of carbene 3 accesses
two transition states (cf., Figure 8). The lower-energy TS (Figure
8
a, ∆H‡ ) 12.0 kcal/mol in vacuo) leads by IRC to chlorides
5
-Cl and B (see Scheme 2) in processes that approximate SNi
reactions. The higher-energy TS (Figure 8b, ∆H‡ ) 19.6 kcal/
mol in vacuo), after IRC calculation, affords the same ion pair
obtained from carbene 4 (Figure 7b).
The large energy gap (7.6 kcal/mol) between these two
transition structures does not necessarily preclude competition
from the higher-energy form because these results relate to
computations carried out in vacuo. Attempts to determine the
effects of solvation are limited to single-point calculations using
simulated solvation PCM methodology. Upon application of
single-point PCM calculations in simulated methanol, both
transition-state energies were lowered, but the energy gap was
not significantly altered. This method evaluates the overall effect
of dielectric media on energies, but not specific solvation effects,
e.g., hydrogen bonding, so it is likely that important differences
could be overlooked.
Figure 6. Optimized ion pair resulting from the ring expansion-
fragmentation of carbene 6 or the fragmentation of carbene 7, deriving from
IRC calculations on the transition states shown in Figure S1 (Supporting
Information), and using B3LYP/6-31G(d) in simulated methanol with the
PCM. See the Computational Methodology section in the Experimental
Section.
3
0
Now, we can understand the disparity between the [5-OMe]/
[5-Cl] product ratios from carbenes 3 and 4 (Figure 1). From
carbene 4, we obtain ion pair 12. (Figure 7b, ∆H‡ ) 10.3 kcal/
mol in vacuo), which gives 5-OMe and 5-Cl by competitive
reaction with either methanol or chloride. Carbene 3, in contrast,
carbenes 9 or 10 contains the more stable10 homoadamantyl
cation and lives longer than the adamantyl chloride ion pair 14
from carbenes 6 or 7, which permits equilibration of the former
ion pair relative to its solvent cage. Ion pair 16 thus becomes
independent of its precursor carbene. This is not true of ion
pair 14. With its less-stable cation and shorter lifetime, it
(
32) Note that the polarized continuum model employed to optimize the ion
pairs in simulated methanol does not use solvent molecules, and the model
cannot speak to the molecular details of the solvent sheath surrounding
the ion pair.
3
2,33
evidently retains some “memory” of its precursor carbene.
Below, we will estimate the time required for ion pair-solvent
equilibration.
(33) For examples of “memory effects” in carbocation rearrangements, see
Berson, J. A. Angew. Chem., Int. Ed. Engl. 1968, 7, 799.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004 12427

A R T I C L E S
Moss et al.
Figure 7. (a) B3LYP/6-31G(d) transition state for the fragmentation of carbene 4. (b) Optimized ion pair from an IRC calculation on the TS shown in (a),
using B3LYP/631G(d) in simulated methanol with the PCM.30 See the Computational Methodology section in the Experimental Section.
Figure 8. (a) Lower-energy B3LYP/6-31G(d) transition state for the ring expansion-fragmentation of carbene 3, which leads to chlorides 5-Cl and B (see
Scheme 2). (b) Higher-energy TS for the ring expansion-fragmentation of carbene 3, which leads to the ion pair of Figure 7b. See text and the Computational
Methodology section of the Experimental Section.
mainly fragments to 5-Cl by an SNi-like reaction via the TS of
Figure 8a; little of this carbene fragments to ion pair 12 (Figure
8
b). The computational study accords with the [5-OMe]/[5-
Cl] dependences on øMeOH shown in Figure 1. Carbene 4, via
ion pair 12, leads to 5-OMe or 5-Cl in a ratio that is responsive
to øMeOH, but carbene 3 mainly gives 5-Cl by SNi fragmentation,
no matter how much methanol is available.
We then carried out LFP experiments on solutions of 3-(1-
adamantylmethoxy)-3-chlorodiazirine (A356 ) 1.0) in MeCN
containing 0.67-2.68 M TMB. In each case, we measured the
Kinetics. Ion pair 16 (Figure 4b), formed from carbenes 9
or 10, resembles an intimate or contact ion pair in the Winstein
formulation.1 In MeOH/DCE, it can collapse to homoadamantyl
chloride 11-Cl, or evolve to a methanol-separated ion pair, from
which solvent trapping generates ether 11-OMe. However, at
the solvent-separated ion pair stage, other trapping agents can
compete for the homoadamantyl cation. For example, in the
presence of 2.68 M trimethoxybenzene (TMB), laser flash
photolytic (LFP) generation of carbene 9 in MeCN gives a UV
signal for adduct 17 (R ) 3-homoadamantyl) at 385 nm.13,34
,2
absorbance of 17 at 385 nm. A Stern-Volmer analysis of 1/A17
vs 1/[TMB] was linear, and is shown in Figure 9. Assuming
diffusion control for the reaction of the homoadamantyl cation
(in 16) with TMB, we can extract the cation lifetime from the
intercept and slope of the Stern-Volmer correlation.34,35 The
precision of the method, as applied to cation 16, is not especially
good: over nine experiments, τ ranged from (0.72-5.87) ×
-11
-11 36
10 s, with an average value of 2.55 × 10 s. We estimate
12428 J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004

Ion Pairs from Carbene Fragmentation Reactions
A R T I C L E S
decrease in cation stability and lifetime (e.g., to the adamantyl
cation) leads to the precursor-dependent, nonequilibrated se-
lectivity between chloride and methanol illustrated in Figure 2.
On the other hand, we would predict that solvent-equilibrated
ion pairs should be formed with cations that are as stable or
more stable than the homoadamantyl cation (i.e., when the cation
has ∆∆G g 4 kcal/mol relative to the 1-adamantyl cation in
1
0a
the gas phase ).
+
-
Conclusions. [R OC Cl ] ion pairs 12, 14, and 16 were
+
generated, with R as the 1-bicyclo[2.2.2]octyl, 1-adamantyl,
or 3-homoadamantyl cation. The ion pairs were produced by
either the direct fragmentation of alkoxychlorocarbenes ROCCl,
with R ) 1-bicyclo[2.2.2]octyl, 1-adamantyl, or 3-homoada-
mantyl, or by the ring expansion-fragmentation of R′CH2OCCl,
with R′ ) 1-norbornyl, 3-noradamantyl, or 1-adamantyl. Ion
pairs were generated in methanol/dichloroethane solutions
containing varying quantities of methanol, and gave ROMe and
RCl products by methanol capture or chloride return. From
correlations of the [ROMe]/[RCl] product ratios as a function
of methanol mole fraction, the homoadamantyl chloride ion pairs
produced either by the direct fragmentation of 3-homoadaman-
tyloxychlorocarbene or by the ring expansion-fragmentation
of 1-adamantylmethoxychlorocarbene were identical. These ion
pairs were solvent- and anion-equilibrated, and they were
independent of precursor identity. LFP experiments involving
a trimethoxybenzene cation trap afforded an estimate of 20-
30 ps as the time required for solvent equilibration and
“precursor amnesia.” This interval was also associated with the
evolution of intimate to solvent-separated ion pairs in the
Winstein solvolysis scenario.
In contrast, the methanol/chloride selectivities of the less-
stable 1-adamantyl chloride and 1-bicyclo[2.2.2]octyl chloride
ion pairs were not independent of their ROCCl or R′CH2OCCl
carbene precursors. These ion pairs (especially when generated
by ring expansion-fragmentation) were not solvent- and anion-
equilibrated. Computational studies provided transition states
for the carbene fragmentations and structures of the ion pairs.
They also suggested that ring expansion-fragmentation of
Figure 9. Stern-Volmer correlation of 1/Absorbance of cation 17 vs
-
1
1
/[TMB] (M ); r ) 0.995 for 8 points.
that the lifetime of the 3-homoadamantyl cation trappable by
TMB in MeCN is ∼20-30 ps.
It seems reasonable to identify the trappable cation as part
of a solvent-separated ion pair derived from an intimate ion
pair that resembles 16. We note that the lifetime of the solvent-
3
separated diphenylmethyl chloride ion pair is ∼4.8 ns in MeCN,
which is about 200 times longer than that of our homoadamantyl
+
-
chloride ion pair. The longer lifetime of [Ph2CH //Cl ] can be
associated with the much greater stability of the diphenylmethyl
cation, relative to the 3-homoadamantyl cation (9.2 kcal/mol
1
0a
in the gas phase).
The polarities of MeCN (ꢀ ) 36.6) and methanol (ꢀ ) 32.6),
as reflected by their dielectric constants, are similar. Therefore,
we estimate that the time required for the evolution of the
intimate homoadamantyl chloride ion pair to its solvent-
separated form in MeOH/DCE is also on the order of 20 to 30
ps. This would be an upper limit to the time needed for solvent
equilibration of the initial ion pairs generated from carbenes 9
or 10 (cf., Figures 4b and 5b). The 20-30-ps time interval is
similar to the principal time constants for the main solvent
relaxation processes of pure methanol (1, 39 ps)3 and to the
time constant (30.5 ps) for the reaction of diphenylmethyl cation
with a methanol solvent shell.3
1-norbornylmethoxychlorocarbene to 1-bicyclo[2.2.2]octyl chlo-
ride occurred via an SNi-like transition state, and largely
bypassed the 1-bicyclo[2.2.2]octyl chloride ion pair.
Experimental Section
7a
Isouronium salts. A. 1-Norbornylmethylisouronium Triflate. The
general method of Moss et al.15 was employed. In a 50-mL round-
7b
bottom flask, equipped with a stirring bar and a reflux condenser
protected with a calcium chloride tube, were placed 1.0 g (23.8 mmol)
Solvent equilibration of the homoadamantyl chloride ion pair
in methanol is in balance with ion pair decay, so that a modest
1
8
2
of cyanamide (NH CN), 3.0 g (23.8 mmol) of 1-norbornylmethanol,
and 12 mL of anhydrous THF. To this solution was added 3.57 g (23.8
mmol) of trifluomethanesulfonic acid. The mixture was magnetically
stirred at 55 °C (oil bath) for 30 h. After the reaction mixture was
cooled to room temperature, it was diluted with ether (200 mL) and
kept in a refrigerator. White crystals of the isouronium salt (2.5 g, 70%)
(
34) For the visualization of carbocations by capture with TMB, see Pezacki, J.
P.; Shukla, D.; Lusztyk, J.; Warkentin, J. J. Am. Chem. Soc. 1999, 121,
6
o
589. The operative form of the Stern-Volmer equation is 1/Acat ) k /
(A∞cat
k
TMB[TMB]) + 1/A∞cat, where Acat is the absorbance of adduct 17 in
the presence of various amounts of added TMB, A∞cat is the absorbance
of 17 at infinite [TMB], k is the sum of all the rate constants that lead to
o
were filtered off, washed with cold ether (3 × 15 mL), and finally
the disappearance of the homoadamantyl cation in the absence of TMB,
and kTMB is the bimolecular rate constant for the reaction of the homoada-
mantyl cation with TMB. Division of the intercept of the equation by its
1
crystallized from acetone, mp 175-176 °C. H NMR (δ, DMSO-d
6
):
8
.30 (br s, 4H, 2NH
2
), 4.36 (s, 2H, CH
2
O), 1.20-2.31 (m, 11H,
slope gives the value of kTMB/k
o
. Taking kTMB as diffusion controlled gives
k
o
, and 1/k is τ, the lifetime of the homoadamantyl cation in ion pair 16.
o
norbornyl H).
(
(
(
35) Platz, M. S.; Modarelli, D. A.; Morgan, S.; White, W. R.; Mullins, M.;
Celebi, S.; Toscano, J. P. Prog. React. Kinet. 1994, 19, 93.
36) Johnson, L. A. Ph.D. Dissertation, Rutgers University, New Brunswick,
NJ, 2001.
Anal. Calcd for C10H N F O S: C, 37.7; H, 5.4; N, 8.8. Found: C,
17 2 3 4
3
7.6; H, 5.1; N, 9.0.
B. 1-Bicyclo[2.2.2]octylisouronium Triflate. This material was
37) (a) Asaki, M. L. T.; Redondo, A.; Zawodzinski, T. A.; Taylor, A. J. J.
Chem. Phys. 2002, 116, 8469. (b) Peon, J.; Polshakov, D.; Kohler, B. J.
Am. Chem. Soc. 2002, 124, 6428.
prepared from 1-bicyclo[2.2.2]octanol1
8,19,21
using our standard method;
15
see the detailed example above. We obtained 9-10% of the isouronium
J. AM. CHEM. SOC. VOL. 126, NO. 39, 2004 12429
9

A R T I C L E S
Moss et al.
1
salt as an oily solid. H NMR (δ, DMSO-d
each, 2NH
), 1.40-1.76 (m, 13H, bicyclooctyl H). 1 C NMR (δ,
DMSO-d ): 160.9, 120.6 (q, J ) 319.7 Hz, CF ), 74.1, 33.7, 26.8,
4.6. A satisfactory analysis could not be obtained. The derived diazirine
see below) had an appropriate UV maximum at 352 nm (pentane) or
6
): 8.51 and 8.32 (br s, 2H
via an addition funnel. The mixture was protected from moisture and
stirred for 12 h at 25 °C. It was then washed with 2 × 5 mL of water,
5 mL of 1 M phosphoric acid, and finally with 5 mL of brine. The
organic phase was dried over anhydrous sodium sulfate, filtered, and
stripped to give the title compound as a yellow solid, 1.86 (86%). A
3
2
6
3
2
(
3
53 nm (DCE).
small portion was recrystallized from petroleum ether/acetone, mp 110
1
C. 3-Noradamantylmethylisouronium Triflate. This material was
°C. H NMR (δ, DMSO-d
6
): 6.54 (s, 2H, 2NH), 3.12 (s, 3H, CH
3
),
2
2
15
prepared from 3-noradamantylmethanol using our standard method;
see the detailed example above. We obtained 60% of the isouronium
1.57-2.12 (m, 15H, adamantyl).
Anal. Calcd for C H N SO : C, 50.0; H, 6.9; N, 9.7. Found: C,
12
20
2
4
1
salt, mp 156-157 °C. H NMR (δ, DMSO-d
.22 (s, 2H, CH
O), 1.5-2.30 (m, 13H, noradamantyl H). 13C NMR
δ, DMSO-d ): 162.1, 75.9, 48.4, 45.3, 40.6, 36.8, 34.8.
Anal. Calcd for C12 S: C, 41.9; H, 5.5. Found: C, 42.0;
H, 5.5.
6
): 8.41 (br s, 4H, 2NH
2
),
49.5; H, 7.0; N, 9.6.
4
(
2
Diazirines. The 3-alkoxy-3-chlorodiazirines were prepared from the
6
isouronium or isourea precursors by oxidation with NaOCl (Graham’s
H
19
N F
2 3
O
4
16
reaction). The preparation of 3-chloro-3-(1-norbornylmethyoxy)-
diazirine is presented as a representative example.
D. 1-Adamantylmethylisouronium Tosylate. This material was
The 1-norbornylmethylisouronium triflate (1.0 g, 3.3 mmol), 3.5 g
of LiCl in 50 mL of DMSO, and 50 mL of pentane were cooled to 20
prepared from a mixture of commercial 1-adamantylcarbinol (20 g,
0
.12 mol), cyanamide (5.0 g, 0.12 mol), and anhydrous p-toluene-
°
C and stirred magnetically in a 500-mL flask. Then 200 mL of 12%
sulfonic acid (from 22.8 g of the monohydrate, dried at 90 °C for 12
h under high vacuum) in 250 mL of freshly dried THF. The mixture
was stirred and heated at 60-65 °C for 5 days under a nitrogen
atmosphere. After cooling, the solid residue was filtered, washed with
THF, air-dried, and chromatographed over silica gel with 10:1 CHCl
MeOH as eluent. We obtained 63% of the desired isouronium salt; mp
aqueous sodium hypochlorite (“pool chlorine”), saturated with NaCl,
was slowly added. Stirring was continued for 15 min at 15 °C after
addition was complete. The reaction mixture was transferred to a
separatory funnel containing 150 mL of ice water, the aqueous phase
was removed, and the pentane phase was washed twice with ∼75 mL
3
/
2
of ice water, and then dried for 2 h over CaCl at 0 °C. The diazirine/
1
2
7
29-231 °C. H NMR (δ, DMSO-d
6
): 8.40 (br s, 4H, 2NH
2
); 7.09-
), 1.53-1.96
pentane solution was purified by chromatography over silica gel, eluted
with pentane. The pentane eluate was concentrated by rotary evaporation
and the pentane was replaced by DCE. Residual pentane was then
removed by rotary evaporation at 0 °C.
.49 (q, 4H, Ar), 3.79 (s, 2H, CH O), 2.27 (s, 3H, CH
2
3
(m, 15H, adamantyl).
Anal. Calcd for C19
28 2 4
H N O S: C, 60.0; H, 7.4; N, 7.4. Found: C,
5
9.6; H, 7.4; N, 7.4.
E. 3-Homoadamantylisouronium Triflate. This material was
1H NMR (δ, CDCl
3 2
): 4.00 (s, 2H, CH O), 2.25 (m, 1H, H4), 1.16-
2
.10 (m, 10 H, norbornyl). UV (DCE): 351, 364 nm.
1
4
15
prepared from 3-homoadamantanol using our standard method; see
above. We obtained 10% of white crystals (recrystallized from ether);
mp 198-200 °C. H NMR (δ, DMSO-d
1
The other alkoxychlorodiazirines were synthesized similarly. They
1
all exhibited UV maxima in the 350-370 nm region. Their DCE
solutions were adjusted so that A ) 0.5-1.0 at λmax
6
): 8.30 (br s, 4H, 2NH
2
),
.
.20-2.31 (m, 17H, homoadamantyl).
Photolysis of Diazirines. Preparative photolysis solutions of diaz-
irines in DCE or in MeOH/DCE were photolyzed at 25 °C for 1 h
with a focused Oriel UV lamp, λ > 320 nm (uranium glass filter). The
products were analyzed by capillary GC, GC-MS, and NMR. The
capillary GC analysis used a 30 m × 0.25 mm (o.d.) × 0.25 µ (i.d.)
CP-Sil 5CB (100% dimethyl polysiloxane) column at 50 °C, pro-
grammed to 250 °C at 20 deg/min. Product identities were confirmed
by GC and GC-MS comparisons to authentic samples.
Anal. Calcd for C13
3.6; H, 5.9; N, 7.4.
N-Methanesulfonyloxy-O-1-adamantylisourea. A. 1-Adamantyl
H N F O S: C, 43.6; H, 5.9; N, 7.8. Found: C,
21 2 3 4
4
Cyanate. Potassium hydride (20.0 mL of a well-shaken 30-wt-%
solution in mineral oil) and 75 mL of dry ether were cooled to 0 °C in
a 500-mL flask. 1-Adamantanol (4.0 g, 27 mmol) was then added in
small portions, with stirring. The flask was protected with a CaCl
and the mixture was stirred for 12 h at 25 °C. After recooling to 0 °C,
.9 g (37 mmol) of cyanogen bromide in 10 mL of ether was added
dropwise via an addition funnel, and stirring was continued for 3 h at
5 °C. The resulting mixture was filtered through a fritted glass funnel.
2
tube,
3
Laser flash photolytic experiments employed a Lambda Physik
2
COMPex model 120 XeF excimer laser. For a description of the LFP
system, see Moss et al.8 The system has been upgraded by the
incorporation of a 150-W pulsed Xe monitoring lamp, replacing the
previously used 1000-W Xe lamp.
2
38
-1
The filtrate displayed cyanate IR absorptions at 2258 and 2262 cm
B. N-Hydroxy-O-1-adamantylisourea. Crystalline hydroxylamine39
.
(
1.0 g, 30 mmol) was dissolved in 10 mL of dry ether in a 250-mL
Products. The products enumerated in Schemes 2-7 were identified
by capillary GC, GC-MS, and NMR comparisons to authentic samples.
References to literature preparations of these materials are given in
the text. Exceptions include protoadamantyl methyl ether (Scheme 4,
G), whose identity is assumed based on its GC-MS molecular ion;
adamantylmethyl formate (Scheme 6, N), which was prepared from
flask. The ethereal adamantyl cyanate solution (prepared above) was
added dropwise, with stirring, via an addition funnel, while maintaining
a reaction temperature of 0 °C. Stirring was continued for 3 h at 25
°
C. Filtration and a wash with cold ether gave 1.43 g (26% based on
1
1
8
-adamantanol) of the desired N-hydroxy-O-1-adamantylisourea; mp
1
26
66-168 °C (recrystallized from MeOH). H NMR (δ, DMSO-d
6
):
1-adamantylmethanol and formic acid; and 3-homoadamantyl methyl
41
.21 (s, 1H, OH), 5.20 (s, 2H, 2NH), 1.56-2.09 (m, 15H, adamantyl).
ether (Schemes 6 and 7, 11-OMe) whose preparation is described here.
Anal. Calcd for C11
3.0, H, 8.4; N, 13.2.
C. N-Methanesulfonyloxy-O-adamantylisourea. The procedure is
based on related methods in the literature. To a stirred suspension of
N-hydroxy-O-1-adamantylisourea (1.43 g, 6.8 mmol) in 20 mL of
18 2 2
H N O : C, 62.8; H, 8.6; N, 13.3. Found: C,
Sodium methoxide (0.15 g, 3.4 mmol) was added to a solution of
.30 g (1.6 mmol) of 3-homoadamantyl chloride (11-Cl) in 10 mL
of methanol. The solution was refluxed for 12 h, then cooled, diluted
6
25
0
4
0
with 30 mL of ether, and washed with 3 × 30 mL of water. The organic
4
phase was dried over anhydrous MgSO , concentrated, and chromato-
2 2
CH Cl at 0 °C, was added 0.6 g (7.6 mmol) of pyridine. After 10
graphed over silica gel using 80:20 hexane/ether as eluent. 3-Homoada-
mantyl methyl ether (11-OMe) was obtained as a colorless oil (0.25 g,
min, 0.86 g (7.6 mmol) of methanesulfonyl chloride was added dropwise
8
5%).
1H NMR (δ, C
(
38) Kauer, J. C.; Henderson, W. W. J. Am. Chem. Soc. 1964, 86, 4732.
39) Hurd, C. D. Inorganic Synthesis; New York: McGraw-Hill: 1939; Vol.
(
6
6 3
D ): 3.08 (s, 3H, CH ), 1.94-1.96 (m, 1H), 1.81-
1
, p 87.
1.85 (m, 6H), 1.70-1.74 (m, 4H), 1.55-1.60 (m, 2H), 1.35-1.40 (m,
(
40) (a) Grigat, E.; Putter, R.; K o¨ nig, C. Chem. Ber. 1965, 98, 144. (b) Stafford,
J. A.; Gonzales, S. S.; Barret, D. G.; Suh, E. M.; Feldman, P. L. J. Org.
Chem. 1998, 63, 10040.
(41) We thank Dr. Jingzhi Tian for this preparation.
12430 J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004

Ion Pairs from Carbene Fragmentation Reactions
A R T I C L E S
4
2
H). 13C NMR (δ, C
8.3. MS (m/e): 180 (M ), 149, 109, 79, 67.
Computational Methodology. All structures were fully optimized
6 6
D
): 76.7, 48.4, 43.8, 38.2, 36.7, 36.6, 32.0, 30.5,
on the IBM p series 690 (Grant CHE 030060). We thank Dr.
X. Fu for technical assistance.
+
by analytical gradient methods using the Gaussian98 and Gaussian03
30a
Supporting Information Available: Tables S1-S6 of product
distributions for carbenes 3, 4, 6, 7, 9, and 10 in MeOH/DCE,
transition states for the fragmentations of carbenes 6 and 7 at
B3LYP/6-31G(d) (Figure S1), and Cartesian coordinates, en-
thalpies, and energies, and (where appropriate) imaginary
frequencies for all ground and transition states, and intermediates
(ion pairs) computed in this study (23 pages, print/PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
suites, as well as density functional (DFT) calculations at the 6-31G-
3
0a,b
(
d) level, with the exchange potentials of Becke,
and the correlation
30a,c
functional of Lee, Yang, and Parr.
Activation energies were corrected
for zero-point energy differences (ZPVE, unscaled) and thermal effects
at 298.150 K. Vibrational analyses established the nature of all
stationary points as either energy minima (no imaginary frequencies)
or first-order saddle points (one imaginary frequency). Solvation
3
0a
simulations used PCM methodology.
Acknowledgment. We are grateful to the National Science
Foundation for financial support and to the National Center for
Supercomputing Applications for allocation of computer time
JA040146O
J. AM. CHEM. SOC.
9
VOL. 126, NO. 39, 2004 12431