Intra- and intermolecular reaction selectivities of γ-substituted adamantanylidenes

Article

pubs.acs.org/joc

Intra- and Intermolecular Reaction Selectivities of γ-Substituted

Adamantanylidenes

Wolfgang Knoll,^‡Daisuke Kaneno,^§Michael M. Bobek,^‡Lothar Brecker,^‡Murray G. Rosenberg,^‡

Shuji Tomoda,^§and Udo H. Brinker*^,‡

^‡Chair of Physical Organic and Structural Chemistry, Institute of Organic Chemistry, University of Vienna, Wahringer Strasse 38,

̈

A-1090 Vienna, Austria

^§Department of Life Sciences, Graduate School of Arts and Sciences, The University of Tokyo, 3-8-1 Komaba, Meguro-Ku, Tokyo

153-8902, Japan

S

* Supporting Information

ABSTRACT: A study of adamantanylidenes having a γ-substituent (R)

was undertaken to gauge how inductive and steric effects of remotely

positioned functional groups influence intra- and intermolecular

product selectivity. 3H-Diazirines were thermolyzed or photolyzed to

generate the corresponding carbenes. On rapid heating, the resulting

carbenes isomerized to 2,4-didehydroadamantanes by intramolecular

1,3-CH insertions. When R was an electron donor (R_D) mostly

asymmetric 1-substituted derivatives were produced but when it was an

electron acceptor (R_A) the symmetric 7-substituted ones were formed.

When solutions were exposed to UV-A light, intermolecular adducts from the carbenes and solvent predominated with lesser

amounts of intramolecular product being formed. Valence isomerization of 3H-diazirines also afforded diazo compounds. In

methanol, protonation of diazo compounds to give the corresponding 2-adamantyl cations exceeds their coupling. This diversion

was controlled with fumaronitrile by trapping the diazo compounds. The adducts possessed mostly anti configurations with

R = R_Dand syn arrangements with R = R_A. The connection between as- and anti-product formation and that of s- and

syn-products was deemed to be the consequence of a rapid equilibrium between two distinct carbene conformations. This was

qualified and quantified using ab initio calculations and NBO analyses.

stereoelectronic effects.⁶The Cieplak and Felkin−Anh models

INTRODUCTION

■

X-Functionalized adamantanylidenes¹bearing a remotely

are two competing proposals that predict opposite selectivities

based on the inductive (I) nature of R,⁷which can be either an

positioned γ-substituent (R; cf. Figure 1) have long intrigued

electron donor (R_D) or acceptor (R_A). They claim that

stereoselection is due to σ-bond I effects and antiperiplanar

MO hyperconjugation during the bond-forming transition state

(TS; cf. Figure 2).⁸Indeed, the latter effect is quite pronounced

Figure 1. X-functionalized adamantanylidenes.

physical organic chemists because an unexpected and un-

explained bias toward one of two possible isomeric addition

products is observed.²π-Facial selectivities of γ-substituted

methylideneadamantanes (X = CH₂),³adamantanones (4, X =

O),²and 2-adamantyl cations (6, X = H⁺)⁴are well

documented. It has been suggested that partially charged (δ )

R groups guide the approaching reagent, e.g., a nucleophile

(Nu), by external electrostatic interaction, but the considerable

distance between R and C−X putatively excludes Coulombic

effects and steric hindrance as factors.⁵Instead, the pendant

group is widely believed to govern reactivity through internal

Figure 2. Representation of the Felkin−Anh and the Cieplak model.

Received: October 17, 2011

Published: December 14, 2011

1340

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

with 4 despite the Felkin−Anh model being deemed

inappropriate for the adamantanyl system by its founder.⁹

However, strong arguments assert that both models demon-

strate which π-face deters Nu−C formation rather than which

one promotes it.

In any case, theoretical chemistry was used to discredit the

TS models in Figure 2,¹⁰which were also found to be in-

adequate for reactive intermediates.¹¹⁻¹³Given this conun-

drum, dismissing the outer effects of R in the ground state (GS)

may have been premature.

X-functionalized adamantanylidenes with a γ-substituent (R; X =

O, CH₂, H⁺, etc.) undergo nucleophilic addition reactions that

exhibit π-facial diastereoselectivity depending on the inductive

effect of R, cf. R_Dand R_A. However, this bias has not been

properly explained by MO theory. So, despite being far from

the CX bond, Coulombic interaction and steric hindrance by R

have been reconsidered. Note that the CX bond orbital in the

general molecule is drawn to highlight the underlying carbene

(cf. Figure 1).

Figure 3. There is a linear relationship between product selectivity,

◆

expressed as ln(isomer ratio), and the I effect of R. ( : hydride

▲

reduction of γ-substituted adamantanones; : intramolecular 1,3-CH

○

insertion reaction of γ-substituted adamantanylidenes; : reaction of

γ-substituted 2-adamantyl cations).

selectivity of 6 has been attributed to hyperconjugation,^16,17

which is maximized when σ-bond I effects stemming from R

cause the α-bridge to bend toward (proximal) or away from

(distal) the γ-substituent. Thus, access to one face of the CX-

bond is physically restricted and product ratios are concom-

itantly altered. If the difference in ΔH^⧧for each TS is negligible,

i.e., ΔΔH^⧧≈ 0, then product distribution must be dictated by

an energy difference between two separate minima inhabited by

distinct conformers,^15b,16i.e., prox-6 and dist-6.

A similar situation might exist with the isoelectronic carbene

5, which is different from carbenium ion 6 in two key ways.

First, the influence of Coulombic charge is absent with neutral

carbene 5. Second, adamantanylidenes react intramolecularly

by 1,3-CH insertion. Thus, both intra- and intermolecular

selectivities may be studied. And, even though the selectivities

of reactions with different molecularities cannot be easily com-

pared, certain conclusions may be drawn by simply comparing

the TSs and their energies in relation to the energies of the

corresponding carbene conformers. Therefore, advanced

computational methods, like density functional theory (DFT)

ab initio MO calculations, that properly reflect experimental

results should help to pinpoint the source of reaction selectivity.

Indeed, the analysis of adamantane-related intermediates should

be possible using natural bond orbital (NBO) theory,¹⁸because

it has already been used to investigate the effect of double-

hyperconjugation on 4-substituted cyclohexyl cations.¹⁹The

following is a report on the experimental intramolecular regio-

selectivities and intermolecular stereoselectivities of the bare γ-

substituted adamantanylidenes 5,²⁰⁻²²i.e., the ephemeral

carbenes without X. The dynamic origins of these reciprocal

selectivities are examined using computational methods.

As a nucleophile (Nu = O, CBr₂, H⁻, etc.) approaches the

electrophilic sp²-hybridized C-2 atom of an X-functionalized γ-

substituted adamantanylidene (X = O, CH₂, H⁺, etc.), the

incipient Nu···C-2 MO of the reactant-like activated complex

is thought to hyperconjugate with those of the antiperiplanar

C-α−C-β bonds on the opposite face of the molecule. The

linearly combined MO interaction presumably lowers the

energy of the participating electron-pair through delocalization,

so the early TS is stabilized and the rate of isomer formation

increases. Two opposing hypotheses of the TS have been pro-

posed to explain stereoselective bond formation. In the Cieplak

model, the σ*-ABMO of Nu···C-2 interacts with the better

donating C-α−C-β σ-BMO on the substituted face when R =

R_D. Thus, the addition product in which R and Nu are anti is

favored. When R = R_A, interaction of the Nu···C-2 σ*-ABMO is

greater with the higher-energy C-α−C-β′ σ-BMO on the

reference face. So, a preference for the product wherein R and

Nu are syn is expected. In the Felkin−Anh model, the σ-BMO

of Nu···C-2 interacts with the better accepting C-α−C-β′ σ*-

ABMO on the reference face if R = R_D. So, the adduct having R

and Nu configured syn predominates. When R = R_A, better

hyperconjugation of the Nu···C-2 σ-BMO with the lower

energy C-α−C-β σ*-ABMO on the substituted face leads

mostly to the product with R and Nu situated anti to one

another. Thus, these two models make opposite predictions.

The exterior frontier orbital extension (EFOE) model can be

used to evaluate how much R affects an incoming reagent.¹³It

is based on the first and last terms of the Salem−Klopman

equation,¹⁴which divides the net reaction energy (ΔE) into

three parts: (1) the exchange repulsion (steric) term, (2) the

electrostatic interaction (Coulombic) term, and (3) the donor−

acceptor MO interaction term. The π-plane divided accessible

space (PDAS; cf. Figure 1) within 2.65 Å of each molecular face

is calculated to quantify the steric effect of R on the CX-bond.

Then the EFOE electron density of each PDAS around the C−

X FMO is calculated to determine the orbital’s distortion index

(δ), which is related to the activation enthalpy (ΔH^⧧). The

difference in ΔH^⧧for each TS (ΔΔH^⧧≠ 0) usually accounts for

product selectivity, which may be plotted as ln(isomer ratio).

For example, the PDAS (cf. Figure 1) of ketone 4 is rendered

unequal when R ≠ H and the course of reaction is thereby

influenced.¹³Hydride reduction of ketone 4h with NaBH₄

yields a modest ratio (ca. 1.5:1) of the corresponding alcohols.¹⁵

However, ratios approach 9:1 for cation 6h (Figure 3).⁴The marked

RESULTS AND DISCUSSION

■

The intramolecular regioselectivity of α-substituted adamanta-

nylidenes giving 1- and 3-substituted 2,4-didehydroadaman-

tanes has been studied.²⁰In that case, the proximity of R to the

divalent C atom likely obfuscates I effects with steric ones. With

γ-substituted adamantanylidenes,²³that should not to be the

case. So, this report focuses only on such adamantanylidenes,

building upon earlier work.^12,21,24During hydride reduction of

ketones 4,^2,5athe γ-substituents exert different σ-bond inductive

effects.²⁵Therefore, various functional groups covering a wide

range of electronegativity values were sought. The Me₃Si group

served as an undisputed R_Dwhen compared with H,²⁶i.e., the

reference “group.” And F was used for its pronounced effects as

an R_A. The Me group was also chosen, because its behavior as

1341

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

a

Scheme 1. Generation of Conformationally Equilibrated γ-Substituted Adamantanylidenes and Their Subsequent Reactions

a

R = SiMe₃(a), H (b), Me (c), NH₂(d), OH (e), Br (f), Cl (g), F (h).

an R_Dor R_Ahas been in dispute for a long time.²⁷This ancillary

issue was investigated through experimental product ratios and

NBO analysis as well.

There is a broad variety of carbene precursors available,²⁸but

3H-diazirines 1 were chosen as the most appropriate ones for

carbenes 5.^23a,29,30Scheme 1 outlines the preparation of 3H-

diazirine 1, from the corresponding ketone 4, and its response

to heat (Δ) and light (hν). Labile 1 either expels N₂to directly

generate short-lived carbene 5 or it isomerizes to diazo com-

pound 2, which can be a problematic diversion. During photo-

Figure 4. Relative energies for (a) 1,2-H shift and (b) 1,3-CH

insertion calculated at the B3LYP/6-311++G(3df,2pd)//B3LYP/6-

31+G(d) level of theory.^22b

lysis,³¹3H-diazirine 1b forms an unstable mixture of carbene 5b

(40%) and diazo compound 2b (60%). Continued photolysis

of 2b does generate 5b. Depending on solution concentration

and acidity, however, dipolar 2 can couple with itself to form

inert azines 3.³²This lowers the overall yield of carbene 5,

which is shown rapidly equilibrating between distal (dist-5) and

proximal (prox-5) conformations. These two distinct conforma-

tional isomers undergo either intramolecular isomerization to

2,4-didehydroadamantanes 7 or intermolecular addition with

solvent. In aprotic cyclohexane (CyH), insertion into a C−H

bond leads to 8. In protic methanol (MeOH), however, OH-

insertion predominates and 9 is formed.

theoretically well investigated. In fact, the selectivity of substituted

derivatives has been used to support Cieplak’s TS model.^5aGas-

phase pyrolysis of γ-substituted aziadamantanes 1 were con-

ducted in this account to generate didehydroadamantanes 7.

Regioisomer ratios (except 7c , see Experimental Section) were

determined from integrated analytical GC signals (cf. Table 1).

Table 1. Ratios of Intramolecular 1,3-CH Insertion Products

Obtained by the Gas-Phase Pyrolysis of Aziadamantanes 1

Under thermal conditions,^22,33aziadamantane (1b) almost

exclusively generates the corresponding carbene 5b, which

isomerizes to 2,4-didehydroadamantane (7b). Cyclopropane

formation from 1,3-CH insertion is a signature reaction of

carbenes. Here, it is the only intramolecular transformation

possible^22bbecause the vicinal migrating hydride orbitals are

orthogonal to the p-LUMO centered on the electrophilic

divalent carbon (cf. Figure 4a). Moreover, a 1,2-H shift would

produce adamantene, which is an anti-Bredt alkene destabilized

by a twisted double-bond (E_strain= 39 kcal/mol).³⁴

reactant

R

s-7

as-7

ratio

1:2.7

1a

1c

1d

1e

1f

Me₃Si

Me

NH₂

OH

Br

27

66

81

89

90

92

73

34

19

11

10

8

1.9:1

4.3:1

8.1:1

9.0:1

11.5:1

1g

1h

Cl

F

They show that the formation of 7-substituted 2,4-didehy-

droadamantanes (s-7), i.e., carbene insertion into a reference-face

C−H bond, is strongly preferred for compounds bearing an R_A.

The most pronounced effect was observed with F-substituted

Adamantanone tosylhydrazone salts form carbenes under thermal

conditions (T = 130−140 °C).²⁰Indeed, they were used to first

generate adamantanylidene (5b).^22aThe intramolecular 1,3-CH

insertion reaction of carbene 5b is both experimentally and

1342

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Table 2. Energy of Carbene 5 Conformers and Their Intermediary TS

a

b

⧧

carbene

prox-5a

bridge tilt (deg)

GS E (hartree)

ΔE_{dist‑5→ prox‑5}(kcal/mol)

−0.44

TS E (hartree)

−798.0940

ΔE (kcal/mol)

−15.21

+13.34

15.07

−798.0950

−798.0943

−389.4101

−428.7294

−428.7304

−444.7659

−444.7675

−444.7656

−444.7676

−464.6336

−464.6359

−488.6565

−488.6595

0.59

0.19

0.52

0.18

0.81

0.08

1.14

0.00

1.24

0.00

1.44

0.00

1.85

dist-5a

5b

−389.4093

−428.7291

prox-5c

−14.0

0.63

1.00

1.26

1.44

1.88

dist-5c

+15.70

−11.86

+16.69

−13.6

+16.51

−13.38

+17.06

−12.97

+17.35

prox-5d (C_s)

dist-5d (C_s)

prox-5d (C₁)

dist-5d (C₁)

prox-5e (C_s)

dist-5e (C_s)

prox-5h

−444.7658

−444.7656

−464.6336

−488.6566

dist-5h

a

b

The dihedral angle is measured from the C-1−C-6−C-3 plane; (+) denotes distal from R and (−) signifies proximal to R. Activation energy for

conformer interconversion: ΔE^⧧= E_TS− E_GS. Calculated at the B3LYP/6-31+G(d,p) level of theory.

carbenes dist- and prox-5h, which form s- and as-7h in a ratio

of 92:8. Formation of 1-substituted 2,4-didehydroadamantanes

(as-7), i.e., insertion into a C−H bond on the substituted face,

was favored only with the R_DMe₃Si (s-7a/as-7a = 27:73). The

inverted isomeric product ratio appears to indicate a reversal in

the lower energy conformation of carbene 5, which was

investigated using computational methods.

Accurate structures of carbenes 5 were attained by geometry

optimization using high-level ab initio calculations. Carbenes 5

were found to be more stable when their C-1−C-2−C-3

α-bridges were bent toward the C-α−C-β bonds (cf. Figure 5).

a singlet ground state (S₀).³⁵But the calculated S₀−T₀energy

gap is relatively small (ca. 2.8 kcal/mol).^22b,36Those for 5a

(2.8 kcal/mol) and 5h (2.3 kcal/mol) are comparable to this.

NBO analysis of the carbene 5 conformer pairs and their

intervening TS suggests that tilting of the α-bridge has two

principal causes (cf. Figure 5). One is a ground-state Cieplak-

type interaction involving electron-donation from a collinear C-

α−C-β σ-BMO to the nominally empty p-LUMO (lp*) of the

divalent C atom.^37,38Indeed, NBO analyses of lp* show that it is

partially occupied (cf. Table 3). This is due to electron-donation

from a C-α−C-β σ-BMO, i.e., Cieplak-type interaction. The

other cause is a ground-state Felkin−Anh-type interaction, which

involves electron-donation by the filled σ-HOMO (lp) of the

divalent C atom to a synperiplanar C-α−C-β σ*-ABMO. This

contribution is somewhat less pronounced, presumably because

the Cieplak-type interaction involves electron-donation from

two filled orbitals to one empty one whereas Felkin−Anh-type

stabilization entails hyperconjugation from one filled orbital to

two empty ones (cf. Figure 5). Nevertheless, both MO interac-

tions are greater than those in the TS (5 ⇆ 5) (cf. Scheme 1).

By bending, carbene 5 better aligns its orthogonal σ- and

p-FMO with the C-α−C-β (anti)bonds thereby maximizing

hyperconjugation on both faces. Not only can these donor−

acceptor MO interactions be visualized (cf. Figure 6)they

can be quantified as well (cf. Table 3). When R ≠ H, there is a

preferred bending direction that depends on the I effect of R.

This leads to two distinct conformers with regard to R, i.e.,

dist- and prox-5. The structures of parent 5b, TMS-substituted

dist- and prox-5a, and F-substituted dist- and prox-5h were

theoretically determined and used as prime examples. Figure 7

shows the structures of dist- and prox-5a, 5b, dist- and prox-5c,

and dist- and prox-5h. It includes diagnostic parameters: E,

ZPVE, and bending angle. Table 2 lists ground state (GS) and

TS energies as well as the kinetic barriers of interconversion

between dist-5 and prox-5.

Figure 5. Stabilization of bent γ-substituted adamantanylidenes 5.

Cieplak-type (a and b) and Felkin−Anh-type (c and d) orbital

interactions reinforce one another in the GS and determine the major

carbene conformer. (Note: For the less stable minor conformers, swap

the R_Dand R_Alabels in a−d, because the donor face acts as acceptor

and the acceptor face as donor.)

This is true even for unsubstituted 5b,^22bwhich is most stable

when its divalent-carbon bridge is bent 15.07° from cen-

ter (cf. Table 2). Thus, the C_2v-symmetric TS(5b ⇆ 5b)

(cf. Scheme 1) represents the unperturbed carbene bridge,

which swings between two turning points. This rapid dynamic

entails two equivalent GS minima separated by a TS that is

0.50 kcal/mol higher in energy. For γ-substituted adamantany-

lidenes, both minima are not isoenergetic and the difference

reveals the influence of R. There is some concern about how

this thermodynamic interchange might affect the multiplicity of

carbene 5. The parent 5b was assumed to have a triplet ground

state (T₀), like other dialkylcarbenes, until it was found to have

There are two conformers of carbene 5, i.e., proximal and

distal with regard to R, because α-bridge bending can occur in

either direction. The major conformation is lower in energy

than the minor one; the amount of interaction between the

p- and σ-FMOs centered on C-2 and the C-α−C-β BMO and

ABMO, respectively, determines which is which. With an R_D,

such as Me₃Si, prox-5 is more stable than dist-5. There are

two reasons. First, the lp* achieves better overlap with the

better donating C-α−C-β σ-BMO on the substituted face

1343

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Figure 6. Contour plots of the (a) HOMO and (b) LUMO of carbene 5b. The arrows indicate donor−acceptor electron-delocalization. FMO

energies of carbenes 5 are listed in Table 4.

Table 3. NBO Stabilization Energies for Carbenes 5

a

b

c

d

e

−

structure

prox-5a

ΔE_{dist‑5 → ‑5}(kcal/mol) lp (e )

lp* (e )

Cieplak-type ΔE (kcal/mol)

Felkin−Anh-type ΔE (kcal/mol)

total ΔE (kcal/mol)

−0.44

1.9099

0.20

12.3

11.0

11.9

5.1

6.6

6.3

6.5

2.9

5.7

7.1

37.8

34.6

36.9

31.8

40.4

dist-5a

1.9108

1.91

0.19

5b

0.186

0.16

TS(5b → 5b)

prox-5h

dist-5h

1.92

1.88

1.917

1.911

0.159

0.187

10.2

13.1

a

b

Electron occupancy of carbene 5 HOMO or corresponding C atom in the TS. Electron occupancy of carbene 5 LUMO or corresponding C atom

c

in the TS. Interaction energies between lp* and one of two C-α−C-β σ-BMO donors. These are C-α−C-β (C-1−C-9 and C-3−C-4) for prox-5 and

d

C-α−C-β′ (C-1−C-8 and C-3−C-10) for dist-5. Interaction energies between lp and one of two C-α−C-β σ*-ABMO acceptors. These are C-α−C-

e

β′ (C-1−C-8 and C-3−C-10) for prox-5 and C-α−C-β (C-1−C-9 and C-3−C-4) for dist-5. The sum of all relevant ΔEs.

(cf. Cieplak-type interaction; Figure 5a). Second, the lp

overlaps better with the better accepting C-α−C-β′ σ*-

ABMO on the reference face (cf. Felkin−Anh-type interaction;

Figure 5c). With an R_A, like F, dist-5 is lower in energy than

prox-5. In this case, stabilization is better achieved when the lp*

overlaps with the higher-energy C-α−C-β′ σ-BMO on the

reference face (cf. Cieplak-type interaction; Figure 5b) and the

lp overlaps with the lower energy C-α−C-β σ*-ABMO on the

substituted face (cf. Felkin−Anh-type interaction; Figure 5d).

Thus, in contrast to nucleophilic addition of X-functionalized

γ-substituted adamantanylidenes (cf. Figure 2), the ground-state

Cieplak- and Felkin−Anh-type interactions are collaborative.

The dependence on R is revealed in Figure 8 and Figure 9.

the energy values for prox- and dist-5 are approximately the

same. On the other hand, the LUMO levels are influenced by

R and the conformation. Indeed, the proximal carbenes are af-

fected more than the distal ones. These findings are in accordance

with the effects of hyperconjugation, which affect the p-LUMO,

cf. Cieplak-type interaction, more than the σ-HOMO, cf. Felkin−

Anh-type. Thus, the major carbene 5 conformer is mainly

determined by the amount of electron-donation to its LUMO.

Thus far, all of the results indicate that the Me group is an R_A

when compared with the H atom reference group. Moreover,

NBO analyses of 5-methyladamantan-2-ylidenes dist- and prox-

5c agree with the findings. They were used to estimate inter-

action energies between C-γ−CH₃and the relevant C-α−C-β

bond. The interaction energy with C-γ−CH₃as acceptor and

one of the C-α−C-β bonds on the substituted face (2.37 kcal/mol

for dist-5c and 2.31 kcal/mol for prox-5c) was greater than

that with C-γ−CH₃as donor (1.79 kcal/mol for the major

conformer dist-5c and 1.93 kcal/mol for the minor conformer

prox-5c). The C-γ′−H bond on the reference face behaves

contrarily. As an acceptor, the interaction energies are

1.92 kcal/mol for dist-5c and 1.97 kcal/mol for prox-5c.

But they are greater as a donor: 2.92 kcal/mol for dist-5c and

2.73 kcal/mol for prox-5c.

The carbene 5 conformational energy difference (ΔE_{dist‑5→prox‑5}

)

depends on the I effect of R. When R = R_D, prox-5 is the major

conformer. When R = R_A, dist-5 predominates. Note that there

is a reversal in the preferred conformation when R is an R_D

versus when it is an R_A. Carbenes 5d and 5e can exist as dif-

ferent rotational isomers. The C₁- and C_s-symmetric rotamers

lead to two data sets (cf. Table 2 and Table 3). Carbenes 5f and

5g exhibit deviations, because their halogen substituents have

similar I effects but different valence shells.

The γ-substituent of carbene 5 affects its FMO energies

(cf. Figure 9 and Table 4). The general appearance and shape

of the HOMO and LUMO are shown in Figure 6. An R_Agroup

withdraws electron density from the remote divalent C

atom thereby enhancing its electrophilicity. The HOMO and

LUMO energy levels are lowered, i.e., become more negative

(cf. Figure 9). The opposite may be said for R_Dgroups. Although

the HOMO levels of 5 are affected by the I effect of R, they are

barely influenced by which direction the α-bridge is bending;

When compared with that for typical alkanylidenes, the ΔE^⧧

for intramolecular 1,3-CH insertion of carbene 5 to dide-

hydroadamantane 7 is higher (13 kcal/mol; cf. Figure 4). This

considerable TS barrier confers two experimental advantages:

(1) compounds s-7 and as-7 are conveniently obtained as the

main products of thermal gas-phase reactions, and (2) they are

formed only as minor byproducts during liquid-phase solu-

tion photolyses. The calculated activation entropy (ΔS^⧧) for

1344

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Figure 7. Structures of carbenes dist-5a, prox-5a, 5b, dist-5c, prox-5c, dist-5h, and prox-5h with selected parameters: E (hartree), bending angle (deg),

ZPVE (kcal/mol), ΔE_{dist‑5→prox‑5}(kcal/mol; (−) exothermic vs (+) endothermic.

TS(5 → 7) is negative (from −2 to −3 cal/(mol·K)), which

indicates a particularly ordered unimolecular TS. The energy

differences between TS(prox-5 → as-7) and TS(dist-5 → s-7)

were calculated and found to roughly parallel ΔE_{dist‑5→prox‑5}

(cf. Table 2 and Table 5), although slightly increasing with

increasing substituent electronegativity. This implies that the

ΔE^⧧for the 1,3-CH insertion in carbene 5 depends somewhat

on its conformation.

The ratio of intramolecular products 7 from carbenes 5

depends on the activation free energy (ΔG^⧧) values quali-

tatively depicted in Figure 10. For a rapid equilibrium, ΔG^⧧(b)

and ΔG^⧧(c) are small. Therefore, their difference is small and

usually ignored. Indeed, the Curtin−Hammett principle³⁹states

that the product ratio depends only on ΔG^⧧(d) − ΔG^⧧(a), i.e.,

k_d/k_a. However, the solution to the rate equations reveals greater

flexibility. Here, the position of the carbene 5 conformational

equilibrium, which is reflected by the sign and magnitude of

ΔG°(f), would be the main source of 1,3-CH insertion regio-

selectivity.

As noted, Figure 10 illustrates transient carbene 5 in rapid

conformational equilibrium between distal and proximal forms

that interconvert through a TS that is C_2v-symmetric in 5b, i.e.,

R = H. The lower energy conformer depends on the direction

of inductive effects of γ-substituent R. Each conformer isom-

erizes to a preferred intramolecular regioisomer 7 and the

product ratio mostly reflects the relative populations of the

carbene conformers prox-5 and dist-5. In solution, intermo-

lecular reaction (not shown) predominates exhibiting stereo-

selectivity that is also R-dependent. The energies of the lowest

vibrational energy levels within each minimum were deter-

mined in order to make zero-point vibrational energy (ZPVE)

corrections to ΔE values.

1345

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Table 5. Effect of the γ-Substituent on the Energy of TS

(5 → 7)

a

TS

ZPVE (kcal/mol)

E (hartree)

ΔE (kcal/mol)

−1.08

prox-5a → as-7a

dist-5a → s-7a

5b → 7b

197.9771

198.0467

134.8502

152.0577

152.0114

128.8329

128.8852

129.5381

129.6684

−798.0752

−798.0736

−389.3896

−428.7092

−428.7099

−848.9880

−848.9912

−488.6356

−488.6392

prox-5c → as-7c

dist-5c → s-7c

prox-5g → as-7g

dist-5g → s-7g

prox-5h → as-7h

dist-5h → s-7h

0.51

1.97

2.14

Figure 8. Conformational energy difference of substituted adamanta-

nylidenes 5 vs the electronegativity of the substituent.

a

ΔE = E_{TS(prox‑5→as‑7)}− E_{TS(dist‑5→s‑7)}

.

Figure 9. Energy of the frontier orbitals of substituted adamantany-

lidenes 5 vs the electronegativity of the substituent. The I effect of R

■

within carbene 5 affects the energy of frontier molecular orbitals ( :

◆

▲

prox-5 HOMO; : prox-5 LUMO; ×: dist-5 HOMO; : dist-5

LUMO).

Figure 10. Curtin−Hammett-type energy profile (not to scale)

showing saddle points where the GS surface meets orthogonal TS

surfaces.

Table 4. Frontier Molecular Orbital Energies

carbene

prox-5a

E_HOMO(eV)

E_LUMO(eV)

−5.0407

−5.0485

−5.1098

−5.1321

−5.1419

−5.2243

−5.2324

−5.1998

−5.2189

−5.3114

−5.3277

−5.3525

−5.3468

−5.4794

−5.5087

−1.8406

−1.9229

−1.9199

−1.9882

−1.9243

−2.1087

−1.9863

−2.1179

−1.9672

−2.2680

−2.0608

−2.2550

−2.0821

−2.3973

−2.2171

dist-5a

5b

prox-5c

dist-5c

prox-5d (C₁)

dist-5d (C₁)

prox-5d (C_s)

dist-5d (C_s)

prox-5e (C₁)

dist-5e (C₁)

prox-5e (C_s)

dist-5e (C_s)

prox-5h

Figure 11. Magnitude of the bending angle in substituted carbenes 5

vs the difference in energy of both conformers.

dist-5h

(ΔE_{dist‑5→prox‑5}) and the corresponding intramolecular TS(5 → 7)

(cf. Figure 12). This correlation is linear and has a slope slightly

greater than unity. Assuming there is only one factor determining

this line slope, two relationships are thus demonstrated: (1)

the insertion ΔE^⧧is slightly affected by conformation, and (2)

1,3-CH insertion within the major conformers of 5 becomes a

bit more facile as the I effect of R grows more withdrawing.

This notion is reasonable considering that the major conformer

is always bent to a higher degree than is the minor one and

should be better poised for intramolecular 1,3-CH insertion

(cf. Figure 13).

Ab initio computations do confirm that the insertion ΔΔE^⧧

for prox- and dist-5 is not zero. So, although facial selectivity is

mainly determined by the relative populations of the carbene

conformers prox-5 and dist-5, it is partially influenced by

each conformer’s intrinsic ability to react. The magnitude of α-

bridge-bending and resulting stabilization for the major carbene

5 conformer is always marginally greater than that for the

minor one (cf. Figure 11). And as the I effect of R strengthens

with respect to the H atom reference, there is a slight increase

in energy difference between the carbene 5 conformers

1346

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

calculated for the intermolecular C−H insertion of CyH (vide

infra).

Table 6. NBO Electronic Parameters of the TS for 1,3-CH

Intramolecular Insertion Reactions of Carbene 5h

Conformers

R

F

carbene

prox-5h

1.917

0.159

as-7h

8

dist-5h

1.911

0.187

s-7h

a

lp (e⁻)

b

−

lp* (e )

Product

Relative Yield (%)

92

TS(5h → 7h):

Figure 12. Correlation of the energy difference between carbene 5

−

◆

σ(C−H) (e )

1.7069

0.1109

1.850

0.393

22.9

1.7098

0.1086

1.849

0.401

22.7

conformers in the GS and at the pertinent TS ( : intramolecular

■

▲

insertion; : insertion into CyH; : formal insertion into MeOH).

σ*(C−H) (e⁻)

a

lp (e⁻)

b

−

lp* (e )

The C-1−C-2−C-3 α-bridge of carbene 5 bends from the

C-1−C-6−C-3 plane to stabilize 5. The more the major conformer

(cf. Figure 5) is bent toward (proximal; −) or away from (distal; +)

the γ-substituent (R), the greater is the net energy difference

between the conformers prox-5 and dist-5 (ΔE_{dist‑5→prox‑5}; cf.

Table 2). Clearly, the magnitude of bending in the major

conformer (prox-5 with R_Dand dist-5 with R_A) is always greater

than that in the minor one (dist-5 with R_Dand prox-5 with R_A).

This is because the extent of Cieplak- and Felkin−Anh-type

MO interactions in 5 is greater with the major conformers than

with the minor ones. In this study, the linear dependence of

distal bending toward the reference face is of special interest,

because dist-5 is the major conformer for the several carbenes

with an R_A(◆: distal bending toward the reference face; ■:

proximal bending toward the substituted face).

The transition states for the formation of fluorodehydroada-

mantane 7h represented in Figure 13 show that the incipient

C-2−H and C-2−C-β bond lengths in (a) and (b), respectively,

are essentially the same. Thus, regioselectivity must be determined

before the TS. The relative importance of the Cieplak- and

Felkin−Anh-type interactions is complicated. For dist-5h, 1,3-CH

insertion of a C-β′−H bond takes place while the collinear C-

α−C-β′ bond (C-3−C-10) donates electron density to lp*. This

reduces the carbene’s electrophilicity and perhaps its ability to

insert. Note that the incipient C-2−C-β bond lengths are shorter

(ca. 1.9 Å) than the incipient C-2−C bond length (ca. 2.4 Å)

lp···σ*(C−H) (kcal/mol)

lp*···σ(C−H) (kcal/mol)

Cieplak-type ΔE (kcal/mol)

116.3

34.00

113.3

c

33.72

d

3.55

5.37

3.79

5.55

e

Felkin−Anh-type ΔE (kcal/mol)

f

3.62

3.84

a

Electron occupancy of carbene 5h HOMO or corresponding C

atom in the TS. Electron occupancy of carbene 5h LUMO or

corresponding C atom in the TS. Interaction energy between lp* and

b

c

the C-α−C-β σ-BMO donorC-α−C-β (C-3−C-4) for prox-5h and

C-α−C-β′ (C-3−C-10) for dist-5hof the inserted C-β−H.

d

Interaction energy between lp* and the C-α−C-β σ-BMO donor

C-α−C-β (C-1−C-9) for prox-5h and C-α−C-β′ (C-1−C-8) for dist-

e

5hof the uninserted C-β−H. Interaction energy between lp and the

C-α−C-β σ*-ABMO acceptorC-α−C-β′ (C-3−C-10) for prox-5h

and C-α−C-β (C-3−C-4) for dist-5hof the inserted C-β−H.

f

Interaction energy between lp and the C-α−C-β σ*-ABMO

acceptorC-α−C-β′ (C-1−C-8) for prox-5h and C-α−C-β (C-1−

C-9) for dist-5hof the uninserted C-β−H.

Table 6 lists the NBO electron populations of the lp and lp*

for both carbene 5h conformers and their respective intra-

molecular 1,3-CH insertion TSs.⁴⁰The analysis shows that lp*

occupancy (lp*_TS− lp*_5h) increases more than three times lp

vacancy (lp_TS− lp_5h) upon going from the GS to the TS. This

indicates that the ground-state carbene behaves as an electro-

phile and receives electron density from the C−H bond into

Figure 13. Representations of the 5h-derived transition structures during intramolecular 1,3-CH insertion: (a) TS(prox-5h → as-7h) and

(b) TS(dist-5h → s-7h). Distances in Å.

1347

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

may even be trimolecular in some cases (vide infra). In liquid-

phase solution, both molecules must diffuse through the solvent

to react. If they are inclined to do so then the reaction rate is

limited only by viscosity. Indeed, the reaction rate may even

exceed the diffusion limit, because carbene 5 reacts with the

ubiquitous solvent itself. The activated complex at the TS is formed,

however, by the collision of two particles into one and is

expected to require a particular arrangement. Thus, the activation

entropy is expected to be negative, i.e., ΔS^⧧< 0 cal/(mol·K),

thereby increasing ΔG^⧧(a,d) (cf. Figure 10). Nevertheless,

reactive carbenes are known to undergo “barrierless” reactions

wherein the entropic limitations are offset by negative

activation enthalpies, i.e., ΔH^⧧< 0 kcal/mol). This favorable

attribute is not seen with intramolecular insertion reactions.

Therefore, the more facile intermolecular C−H insertions are

expected to be less selective than intramolecular ones.

Table 7. Geometric Changes between the GS and TS of the

Two Confomers of 5-Chloroadamantan-2-ylidene (prox- and

dist-5g)

r(C-2−H)

r(C-2−C-β)

structure

prox-5g

(Å)

Δr (Å)

(Å)

Δr (Å)

2.654

1.513

2.578

1.516

1.141

2.427

1.922

2.369

1.930

0.505

TS(prox-5g → as-7g)

dist-5g

1.062

0.439

TS(dist-5g → s-7g)

which it is inserting. This is highlighted by the development of

a slightly negative charge on the divalent C atom, i.e., −0.08

NBO charge for TS(dist-5h → s-7h). NBO analyses of carbenes

dist- and prox-5h also give interaction energies between C-β−H

and C-2 as a donor (lp) vs acceptor (lp*). The divalent C atom

is a much better acceptor than donor (cf. Table 6). This finding

also supports an electrophilic mechanism. Because the divalent

C atom inserts into a C-β−H bond of C-α−C-β, the Cieplak-

and Felkin−Anh-type interactions within the TS can be gauged.

The Cieplak-type interaction of lp* with the C-α−C-β σ-BMO

was remarkably high for the one with the C-β−H being

inserted. This strong interaction is not nearly the case for the

corresponding Felkin−Anh-type interaction.

The TS-geometry is not easily achieved for the minor isomer

as-7g (cf. Table 7). The degree of bridge-bending again plays a

crucial role. So, it is necessary to analyze the intramolecular

distances closely. Table 7 shows that the major conformer dist-

5g must change its relative geometry, i.e., C-2−H, by ca. 8%

less than the minor conformer prox-5g. Carbenes prox- and

dist-5g were used as an example, because carbene 5g has been

closely investigated.²⁰

It is common to perform intermolecular carbene C−H

insertion reactions using cyclohexane,⁴¹because its carbon

atoms are chemically equivalent; it simplifies product analysis.

So, each 3H-diazirine 1 was dissolved in cyclohexane and

photolyzed. As mentioned, up to 60% of 1 isomerizes to diazo

compound 2 (Scheme 1). Because 2 does not react with aprotic

cyclohexane, it has ample opportunity to couple with itself

thereby forming substantial amounts of inert azine 3. Also,

varying amounts of ketone 4 were observed, because the low-

lying T₁state becomes populated in the presence of trace

amounts of oxygen. Notwithstanding, the cyclohexyl adducts 8

were obtained. They were configured in two diastereomeric

forms, i.e., syn- and anti-8. The stereoselectivity clearly depends

on R, as the relative ratios in Table 8 indicate.

Ab initio computations for the intermolecular insertion reactions

of carbenes prox- and dist-5 with solvent suggest that product

stereoselectivity is at least partially influenced by a nonzero ΔΔE^‡,

as the insertion line slopes in Figure 12 indicate. But in contrast to

intramolecular insertion, the less-than-unity line slopes mean that

intermolecular insertions by the major conformers of 5 become a

little less facile as the I effect of R grows stronger. This notion is

reasonable considering that the major conformer is always bent

to a higher degree than is the minor one. As can be shown by

EFOE considerations (herein not listed), the reactive part of the

LUMO is less accessible and should be less poised for inter-

molecular insertion. Regardless, the experimental isomer ratios for

intra- (cf. Table 1) and intermolecular insertions (cf. Table 8 and

Table 11, vide infra) are essentially the same. So, as with intra-

molecular bond insertion, intermolecular selectivity is dominated

by the conformational ΔE rather than the insertion ΔΔE^⧧and is

readily explained using the Curtin−Hammett principle (cf.

Figure 10). Thus, the stereoselective formation of 8 (and 9, vide

infra) also stems mainly from equilibrated carbene 5 conformers.

The insertion of a carbene into a C−H bond is considered to

be concerted and electrophilic.⁴²So, the p-LUMO of the

electron-deficient carbene may be envisioned approaching the

target C atom and interacting with a filled C−H MO to form

two new σ-bonds. The Bach model for intermolecular insertion

of a carbene into a C−H bond posits two possible orientations

at the TS: σ- and π-approach (cf. Figure 14).⁴³A partial positive

charge is supposed to develop on the solvent C atom.⁴⁴And, as

it approaches the TS, the divalent C atom in carbene 5b be-

comes more negatively charged, i.e., from +0.14 to −0.1 NBO

charge. Therefore, the TS for intermolecular C−H bond

insertion is somewhat polar. The kinetic barrier between the

carbene reaction intermediate and the product depends on the

net contributions of ΔH^⧧and ΔS^⧧. Again, the activation entropy is

assumed to be negative, i.e., ΔS^⧧< 0 cal/(mol·K),^43,45because the

TS has a particular orientation. But this is expected to be offset

by a negative activation enthalpy, i.e., ΔH^⧧< 0 kcal/mol).

Figure 15b shows the HOMO of the TS(5b → 8b). The

approach of carbene 5b toward cyclohexane is symmetric and it

inserts preferably into an equatorial C−H bond, because this

TS is 0.63 kcal/mol more stable than that for axial C−H

insertion. An inspection of the HOMO of cyclohexane shows

that it excludes the axial H atoms (cf. Figure 15a).⁴⁶

Table 8. Ratios of Intermolecular C−H Insertion Products

Obtained by the Photolysis of Aziadamantanes 1 in

Cyclohexane

R

syn-8

anti-8

Ratio

1:1.7

Me₃Si

Me

NH₂

OH

Br

37

67

79

88

87

89

92

63

33

21

12

13

11

8

2.0:1

3.8:1

7.3:1

6.7:1

8.1:1

11.5:1

Cl

F

The activated complexes between each carbene 5a conformer

and cyclohexane were also analyzed by the NBO method

(cf. Figure 16a,b). The Cieplak- and Felkin−Anh-type interac-

tions depicted in Figure 5 were quantified from the electron

populations of lp, σ(C−H), lp*, and σ*(C−H). At the TS,

Intermolecular insertion reactions of carbenes into C−H

bonds differ from intramolecular ones in various ways. First,

they have different molecularities. Intramolecular reactions are

unimolecular, but intermolecular reactions are bimolecular and

1348

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Table 9. Structural and Electronic Parameters of the TS for

the Intermolecular Reaction of Carbenes 5 and Cyclohexane

donor (R_D)

Me₃Si

acceptor (R_A)

F

R

carbene

prox-5a

dist-5a

prox-5h

dist-5h

1.911

0.187

syn-8h

a

−

lp (e )

1.917

0.159

anti-8h

b

−

lp* (e )

0.20

0.18

product

anti-8a

(major)

syn-8a

(minor)

(major)

relative yield (%)

TS(5 → 8):

63

37

8

92

Δr(C−H) (Å)

1.294

1.577

0.249

1.70

1.288

1.591

0.240

1.71

1.250

1.624

0.205

1.74

0.49

95

1.261

1.607

0.214

1.73

0.51

99

σ(C−H) (e⁻)

−

σ*(C−H) (e )

a

−

lp (e )

b

−

lp* (e )

0.54

0.53

lp···σ*(C−H)

136

131

(kcal/mol)

Figure 14. Frontier molecular orbital interaction of carbene C−H

insertion into a methylene group depicting the (a) π_CH2−p and (c)

σ_CH2−p approach. Overlap of the carbene’s LUMO with the filled π

CH₂(b₁) group orbital is preferred for intramolecular insertion (b) and

with the filled σ CH₂(a₁) group orbital for intermolecular insertion (d).

lp*···σ(C−H)

153.8

47.3

9.65

7.49

144.52

44.9

125

37

136.22

40

(kcal/mol)

lp*···σ*(C−H)

(kcal/mol)

c

Cieplak-type ΔE

9.47

9.71

7.18

9.95

7.61

(kcal/mol)

d

the donor−acceptor lp···σ*(C−H) and acceptor−donor

lp*···σ(C−H) interactions leading to C−H insertion are

comparable in magnitude (cf. Table 9). Interaction between

the carbene lp* and the CyH σ*(C−H) is possible, because the

lp* at C-2 is not empty. So, the lp*···σ*(C−H) interaction was

also analyzed (cf. Table 9). The major carbene 5 conformers

were found to always have a later TS than the minor ones and

be further advanced in the reaction. The results of the NBO

analyses do support a concerted electrophilic mechanism for

C−H insertion. This is made evident by the donor−acceptor

interaction energies and the stretching of the inserted C−H

bond. With an R_D, the divalent C atom in carbene 5 becomes

less electron-deficient through electron donation. Though still

electrophilic, the difference in lp···σ*(C−H) and lp*···σ(C−H)

interactions is smaller than that with an R_A. Electron-donation

to the σ*(C−H) and from the σ(C−H) of cyclohexane is

enhanced. Stabilization by the Felkin−Anh-type interaction is

also augmented, because the TS leading to the main product is

relatively tight, i.e., cyclohexane is close to the divalent C atom

(cf. Figure 16a,b and Figure 14d). The interactions between

each carbene FMO and the inserted C−H bond of cyclohexane

were analyzed for the F-substituted carbenes dist- and prox-5h

as well (cf. Figure 16d,e). The importance of Felkin−Anh-type

stabilization increases if just the adamantanylidene moiety of

the TS is considered, but the Cieplak-type stabilization is still

greater. Although the magnitudes of the latter type are smaller

than in ground-state carbenes 5 (cf. Table 3), the relative

Felkin−Anh-type ΔE

7.37

(kcal/mol)

a

Electron occupancy of carbene 5 HOMO or corresponding C atom

b

in the TS. Electron occupancy of carbene 5 LUMO or corresponding

c

C atom in the TS. Interaction energies between lp* and one of two

C-α−C-β σ-BMO donors. These are C-α−C-β (C-1−C-9 and C-3−C-

4) for prox-5 and C-α−C-β′ (C-1−C-8 and C-3−C-10) for dist-5.

d

Note, the approach is symmetric. Interaction energies between lp and

one of two C-α−C-β σ*-ABMO acceptors. These are C-α−C-β′ (C-

1−C-8 and C-3−C-10) for prox-5 and C-α−C-β (C-1−C-9 and C-3−

C-4) for dist-5. Note: the approach is symmetric.

importance between both types of stabilization differs only

slightly from Me₃Si to F.

In addition to electron populations and interaction energies

of selected orbitals in TS(dist-5 → syn-8) (cf. Figure 16),

noticeable differences are found by comparing interatomic

distances. The C-α−C-β and C-α−C-β′ bonds at the TS

become shorter, because of diminished electron-donation to

the lp* of the divalent C atom (cf. Table 10), when compared

with those in ground-state carbenes 5. Moreover, the bending

angle of the carbene α-bridge at the TS is reduced (cf. Figure 7

and Figure 16). This lessens the distance between the divalent

C atom and the inserted C−H bond. The major conformers

benefit from this more than the minor ones, because the

cyclohexane molecule approaches from the more exposed face.

Thus, the I effects of R determine the preferred conformation

Figure 15. Contour plots for the insertion of carbene 5b into CyH: (a) HOMO of CyH, (b) HOMO of TS(5b → 8b).

1349

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Figure 16. Structure and characteristics of the activated complex between carbenes 5 and cyclohexane: E (hartree), bending angle (deg), ZPVE

(kcal/mol)) ΔE (kcal/mol; (−) exothermic vs (+) endothermic. Note: a and b were calculated at the B3LYP/6-31+G(d) level.

Table 10. Comparison of the Stabilizing Interactions in

Carbene 5b and at the TS with Cyclohexane

Photolysis of 3H-diazirines 1c−h in MeOH mainly gave

methyl ethers 9c−h, which could be distinguished as either the

syn- or anti-stereoisomer (cf. Table 11). It is clear that the

relative amount of syn-9 increases as the I effect of R increases.

Complications arise when 3H-diazirine 1 is photolyzed in

MeOH. As mentioned, up to 60% of 1 isomerizes to diazo

compound 2 (Scheme 1). In aprotic media like cyclohexane,

unstable 2 either expels N₂and simply goes on to generate

carbene 5 or it couples to give inert azine 3. But in protic

solvents, like water and alcohol, a diazo compound decays even

faster, because it is readily protonated even by weak acids.⁴⁷

The resulting diazonium cation (not shown) releases N₂and

the ensuing cation 6 can also react with MeOH to yield 9. So,

isomeric product ratios may be skewed by dissimilar

a

electron population

C-α−C-β bond

bond length (Å)

σ-BMO

σ*-ABMO

5b:

(C-3−C-4)

(C-3−C-10)

TS(5b → 8b):

(C-3−C-4)

(C-3−C-10)

1.580

1.552

1.924

1.970

0.022

0.036

1.565

1.551

1.943

1.974

0.0192

0.0347

a

cf. Figure 16.

of 5 at the GS and they affect the reaction by controlling the

degree of electron deficiency of the divalent C atom at the TS.

1350

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

rearrangement product 7-substituted tricyclo[3.3.1.1^2,7]decan-

3-yl methyl ether (rearr-9; 1.6%).⁵⁰

Table 11. Ratios of Intermolecular O−H Insertion Products

Obtained by the Photolysis of Aziadamantanes 1 in MeOH

There is some ambiguity regarding the formal insertion of a

carbene into an O−H bond.^48,51Several mechanisms for the

intermolecular reaction of carbene 5 with MeOH are presented

in Scheme 2: (1) a proton transfer followed by ion-pair recom-

bination; (2) a concerted interaction between reactants; (3) a

concerted trimolecular step with a “dimeric” reactant; and (4)

oxonium ylide formation with subsequent Stevens rearrangement.⁵²

Because it is not possible to assign an isomeric product ratio to

each mechanism, other evidence is needed.

with fumaronitrile

without fumaronitrile

R

syn-9

anti-9

ratio

1:2.2

1.8:1

syn-9

anti-9

ratio

1:2.2

1.5:1

Me₃Si

Me

NH₂

OH

Br

31

64

81

85

83

85

91

69

36

19

15

17

15

9

31

60

73

75

77

79

84

69

40

27

25

23

21

16

4.3:1

5.7:1

2.7:1

3.0:1

3.3:1

3.8:1

5.3:1

4.9:1

Cl

5.7:1

F

10.1:1

No C−H bond insertion by 5 was observed. This strict

chemoselectivity of the reaction casts doubt on mechanism 2. It

is further unlikely, because of the highly dipolar nature of the

O−H bond and the strong influence of polarity on the TS.^48a

In fact, the charge separation that arises in mechanisms 1 and 4

should be facilitated by a polar solvent, like MeOH (dielectric

constant (κ) = 32.7). Further inspection of these pathways

reveals that they are complementary; carbene 5 is an electro-

phile in mechanism 4, but it is a nucleophile in mechanism 1.

Ostensibly, one might rule out mechanism 1 in favor of mechanism

4, because it seems unlikely that an electron-deficient carbene

could act as an electron-donating nucleophile. However, there

is growing evidence that this can occur. The simple FMO

energies listed in Table 12 suggest that carbene 5b is at least

ambiphilic.⁵³So, if MeOH does oligomerize then the cyclic TS

in mechanism 3 becomes quite plausible,⁵⁴because electron-

Scheme 2. Proposed Mechanisms for the Formal Insertion of

Carbene 5 into the O−H Bond of MeOH

Table 12. Comparison of the Energies of the Frontier

a

Molecular Orbitals of Select Carbenes

carbene

E_LUMO(eV)

E_HOMO(eV)

π-bond reactivity

Cl−C−Cl

−3.75

−2.86

−2.01

−1.90

−1.03

−0.49

−7.54

−6.50

−6.94

−5.09

−5.71

−6.48

electrophilic

ambiphilic

Me−C−Cl

MeO−C−Cl

5b

ambiphilic

MeO−C−Me

MeO−C−OMe

nucleophilic

a

B3LYP/6-31+G(d,p)//B3LYP/6-31+G(d) calculations.

donation from the O atom of one MeOH moiety enhances the

nucleophilicity of the divalent carbon and induces proton

transfer from the other. Here, the O−H insertion truly is formal,

because the O and H atoms come from different molecules.

The propensity of carbene 5 to accept H⁺from MeOH and

form cation 6, as in mechanism 1, is a simple matter of acid−

base chemistry. The gas-phase proton affinity (PA) of a car-

bene, which is the net energy required to protolyze its carbe-

nium ion conjugate acid, is a thermodynamic quantity that

parallels its Brønsted−Lowry basicity (K_b). This, in turn, often

correlates with its nucleophilicitya kinetic term. Although the

K_bvalues for carbenes can be determined by comparison with

those of known indicators, the PA values are usually calculated

using theoretical methods, like ab initio B3LYP/6-31G(d).⁵⁵

Hence, the abilities of different carbenes to act as electron donors

during an elementary step can be estimated and arranged. For

example, the PA values of norbornanylidene and norborneny-

lidene (254.7 and 263.0 kcal/mol, respectively) are higher than

that of dichlorocarbene (209.0 kcal/mol), which suggests the

former are nucleophiles that react with electrophiles, like H⁺,

and the latter is an electrophile that reacts with nucleophiles,

like π-rich alkenes. Indeed, the PA values for hydrocarbon carbenes

can rival or even exceed those of the celebrated nucleophilic

selectivities of intermediates 5 and 6. Therefore, 2 must be

rapidly trapped by dipolar addition with the efficient scavenger

fumaronitrile (FN).^24,48Indeed, different product ratios were

observed in its absence. And cation 6 left other clues. When

3H-diazirine 1b was photolyzed in methanol-d₄, typical amounts

of 9b-d₄(95%) and 7b-d₁(5%) were formed. The absence of

unlabeled 7b, however, rules out direct isomerization of

carbene 5b. Exclusive label incorporation suggests another

mechanism involving cation 6b or other suitable precursors of

isotopomer 7b-d₁.⁴⁹Another sign that route 2 → 6 is operative

and must be thwarted by FN was formation of the skeletal

1351

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Figure 17. Structure and characteristics of the activated complex between carbenes 5 and MeOH: E (hartree), bending angle (deg), ZPVE (kcal/mol),

ΔE (kcal/mol; (−) exothermic vs (+) endothermic).

carbenes imidazolylidene and 3,4-dimethylthiazolylidene (258.2

and 252.5 kcal/mol, respectively). And because the pK_avalue of

the latter’s conjugate acid was estimated to be 16.5 in DMSO

and 18.9 in H₂O,⁵⁶one may infer that 3,4-dimethylthiazoly-

lidene and alkylcarbenes are readily protonated in relatively acidic

hydroxylic solvents, like H₂O (pK_a= 15.7) and MeOH (pK_a=

15.5). So, the participation of carbene 5 in mechanism 1 deserves

further consideration, because the calculated PA value of 5b

(268.9 kcal/mol; B3LYP/6-31G(d)) is even higher than that of

norbornenylidene. Its greater basicity and nucleophilicity, which is

also supported by FMO analysis, likely comes from the absence

of ring strain in carbenes 5. Finally, if cation 6 can rearrange faster

than it recombines with its counterion then observing products

stemming from 6 is another way to support mechanism 1. Of

course, one would need to exclude all other channels leading to

cation 6, e.g., 2 → 6.

Now, even if protonation of carbene 5 by an O−H group is

an energetically favored step, it could be a kinetically stymied

one. In addition to basicity, other factors that determine

nucleophilicity, like charge and polarizability, may need to be

considered. It is conceivable that 5 reacts with MeOH to form a

methoxonium ylide, as in mechanism 4. Such routes are widely

supported, because the formation of ylides from electrophilic

carbenes has been directly observed.⁵⁷However, neutral O

atoms are not strong nucleophiles and the ability of Et₂O and

MeOH to form oxonium ylides is weak.⁵⁸Thus, it seemed most

worthwhile to explore mechanism 1.

1352

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Table 13. Structural and Electronic Parameters of the TS for

the Intermolecular Reaction of Carbene 5h and MeOH

Table 14. Distances between the Divalent Carbon (C-2) and

the O−H Bond of MeOH at the TS for Intermolecular O−H

Insertion

acceptor (R_A)

activated complex

r(C-2−O) (Å)

r(C-2−H) (Å)

R

F

TS(dist-5a → syn-9a)

TS(5b → 9b)

2.269

2.256

2.209

1.358

1.364

1.380

carbene

prox-5h

dist-5h

1.911

a

lp (e⁻)

1.917

b

−

TS(dist-5h → syn-9h)

lp* (e )

0.159

0.187

product

anti-9h (minor)

syn-9h (major)

91

NBO analyses were performed on the methanolysis TS of

the following carbenes: dist- and prox-5a, dist- and prox-5c, dist-

and prox-5h, and parent carbene 5b. They predict that MeOH

approaches from the more exposed face of carbene 5 (cf. Figure 17).

This means that anti-9a will predominate, because carbene 5a

spends more time in its proximal conformation. Although the

TS(5h → 9h) was found to be earlier than the others, likewise

predictions were made. Because dist-5h outnumbers prox-5h,

syn-9h is expected to abound. The inversion in diastereoselec-

tivity reflects the different I effects of R, which reverse from an

R_Dto an R_A. It is not unreasonable that a nucleophilic carbene

would be attracted to the protic H atom of MeOH, because it

bears a sizable partial positive charge (+0.45).⁶¹

relative yield (%)

TS(5 → 9):

9

Δr(O−H) (Å)

1.11

1.13

−

σ(O−H) (e )

1.833

1.822

σ*(O−H) (e⁻)

0.267

0.27

a

lp (e⁻)

1.706

1.702

b

−

lp* (e )

0.363

0.385

lp···σ*(O−H) (kcal/mol)

lp*···σ(O−H) (kcal/mol)

lp*···σ*(O−H) (kcal/mol)

Cieplak-type ΔE (kcal/mol)

123.2

123.8

39.0

43.55

9.33

10.32

c

9.91, 10.11

5.12, 5.21

11.43, 11.28

5.78, 5.9

d

Felkin−Anh-type ΔE (kcal/mol)

a

Electron occupancy of carbene 5 HOMO or corresponding C atom

in the TS. Electron occupancy of carbene 5 LUMO or corresponding

b

Indeed, NBO analysis found that the donor−acceptor

lp···σ*(O−H) interaction between carbene 5 and MeOH at

the TS was much greater than the acceptor−donor lp*···σ(C−

H) interaction leading to O−H insertion (cf. Table 13).

Therefore, unlike electrophilic insertion of 5 into the C−H

bond of CyH (cf. Table 9), formal insertion into the O−H

bond of MeOH is made by a nucleophilic carbene 5. As with

cyclohexane, there is an interaction between both “vacant”

orbitals. Because electron-density was donated into the lp*, it is

not empty and interaction between the carbene 5 lp* and the

MeOH σ*(O−H) is possible. So, the lp*···σ*(O−H)

interaction was analyzed as well (cf. Table 13). Evaluation of

the Cieplak- and Felkin−Anh-type interactions gave typical

values. As with cyclohexane (cf. Table 9), the Cieplak-type were

stronger. The major carbene conformer is more stabilized by

these interactions than is the minor one. The conclusion is that

the carbene insertion into MeOH is nucleophilic and involves a

protic mechanism.⁶²Compared with the more electron-rich

carbenes 5a and 5b, the distance between C-2 and the O atom

of MeOH for carbene 5h is shorter at the TS (cf. Table 14). In

contrast to the unsubstituted carbene 5b, the electron-deficient

carbenes dist- and prox-5h prefer the approach of the O atom’s

lone-pair of electrons, i.e., mechanism 4, slightly more than

protonation, i.e., mechanism 1. Regardless, the distance

between the divalent C atom and the O atom seems long

(2.256 Å). One should keep in mind, however, that these are

gas-phase calculations with inherent disadvantages. We will

show only the results for compound 9h where the NBO

analysis of the TS did not attribute the H atom to the divalent

carbon in 5 anymore, because of the reasons listed in note 61.

The protonation of carbenes 5 and resulting formation of

cations 6 cannot be excluded. But one ensures that the presence

of cations 6 is due only to the protonation of carbene 5 by

capturing diazo compounds 2 before their protonation. Unfor-

tunately, it is not possible to differentiate between reaction

intermediates 5 and 6 based solely on product diastereose-

lectivities, because they differ only slightly (cf. Figure 3).^4bThis

might be expected, because the ΔE values for the rapid

equilibrium between prox- and dist-6 are similar to those for the

corresponding carbene 5 conformers.⁶³A comparison of

carbene 5a and cation 6a presents an interesting counter-

c

C atom in the TS. Interaction energies between lp* and both C-α−C-

β σ-BMO donors are given. These are C-α−C-β (C-1−C-9 and C-3−

C-4) for prox-5 and C-α−C-β′ (C-1−C-8 and C-3−C-10) for dist-5.

d

Note, the approach is not symmetric. Interaction energies between lp

and both C-α−C-β σ*-ABMO acceptors are given. These are C-α−C-

β′ (C-1−C-8 and C-3−C-10) for prox-5 and C-α−C-β (C-1−C-9 and

C-3−C-4) for dist-5. Note, the approach is not symmetric.

Although the reaction kinetics of transient carbene 5b with

cyclohexane are fast (t_1/2= 20 ns; pseudo-first-order rate

constant k_obs< 5 × 10⁷s⁻¹),^36athe measurement of those for

“invisible” S₀carbene 5 with MeOH are limited only by

indicator diffusion.^24b,59The reaction of carbene 5 with MeOH

is almost instantaneous. So, if protonation mechanism 1 is

operative then it is not hampered at all. The TS of the insertion

of carbene 5 into MeOH is computed to indeed be 6.3 kcal/

mol more stable than the isolated reactants.⁶⁰Therefore, here is

another example of a “barrierless” reaction with ΔS^⧧< 0 cal/

(mol·K) and ΔH^⧧< 0 kcal/mol (vide supra).

As discussed,^4b,11the protonation of carbene 5b is predicted

to be favored both thermodynamically and kinetically. Indeed,

protonation of carbenes 5 might be a straightforward way to

generate “free” 2-adamantanylium cations 6, whose reactions in

MeOH have been well examined.^4a,11Regardless, protonation

mechanism 1 is fundamentally different from the others out-

lined in Scheme 2, because nucleophilic behavior of a carbene is

unconventional. A proton would interact first with the HOMO

of 5. But the stabilizing Cieplak- and Felkin−Anh-type inter-

actions in the resulting cation 6 should mimic those in the

isoelectronic carbene 5. Indeed, the ΔE_dist→proxvalues between

their respective conformers are similar and structural response

to the I effect of R appears to be the same for both intermediates 5

and 6. So, similar stereoselectivities might be expected if it were

not for the fact that the activation energies for 5 → 9 and 6 →

9 are different. The computed ΔΔE^⧧values, however, have been

found herein to be minimal and that product stereoselectivity

is based mainly on the ΔE_dist→proxbetween the conformers

(cf. Figure 10). Thus, if protonation of 5 does occur then it is

likely to have only a small effect on the isomeric product ratios.

1353

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

example, however, because the observed isomeric product

ratios and computed ΔE values^11,63differ significantly. Never-

theless, product selectivities are usually correlated with the I

effect of R.^64aThus, the insertion product ratios may be

predicted from the GS of the reaction intermediates alone and

Cieplak’s model of electron-donation into the σ*-ABMO of the

incipient bond during the TS is not needed.

adamantanylidenes. Similar conclusions could be made for

γ-substituted 2-adamantylium cations.¹¹

COMPUTATIONAL METHODS

■

Density functional theory (DFT) ab initio MO calculations were

performed using the Gaussian 03 program.⁶⁵The hybrid method used

combines Becke’s three-parameter exchange functional with the

correlation functional of Lee, Yang, and Parr (B3LYP).⁶⁶Unless

otherwise stated, the B3LYP method^66,67was used with a 6-31G(d)

basis set for geometry optimizations and a larger 6-31+G(d,p) basis set

for single-point energies. Vibrational frequency analyses were per-

formed on each stationary point within energy minima, in order to

make zero-point vibrational energy (ZPVE) corrections. The values

discussed in the text represent internal energies (E) at 0 K. Quanti-

fication of donor−acceptor interactions was made by second-order

perturbation analysis as incorporated in the natural bond orbital

(NBO) method.

CONCLUSION

■

3H-Diazirines 1 are labile compounds that generate carbene 5

reaction intermediates when exposed to heat or light. The α-

bridges of the reaction intermediates bend to maximize the

stabilizing effects of Cieplak- and Felkin−Anh-type orbital

interactions, which are atypically cooperative here. Because the

transients have substituted and unsubstituted faces, this leads to

proximal and distal conformational isomers that are in rapid

equilibrium with each other. The energy and relative population of

each conformer depends on the σ-bond inductive effect of the

γ-substituent. The proximal conformer is favored when it is an

electron-donor and the distal one when it is an electron-

acceptor. Each conformer leads to a unique TS and the ratio of

isomeric products reflects the ratio of carbene conformers.

Selectivity is observed for both intra- and intermolecular

insertion reactions. The warped structure of each carbene con-

former is responsible for the reciprocity between intramolecular

regioselectivity and intermolecular stereoselectivity. Because

the proximal carbene is the major conformer when the γ-

substituent is an electron-donor, it is simultaneously poised to

favorably produce the as- and anti-products. The C−H bond of

the former is already within reach of the bent carbene, whose

exposed reference face is more apt to receive an incoming

solvent molecule. And because the distal carbene is the major

conformer when the γ-substituent is an electron-acceptor, it is

ready to proliferate the s- and syn-products. Again, the C−H

bond of the former is already within reach of the bent carbene,

whose exposed substituted face is more receptive to an incom-

ing solvent molecule.

EXPERIMENTAL SECTION

■

General Information. FT NMR spectra were recorded at T = 300 K

on spectrometers at the following Larmor frequencies: ω_L(¹H) =

250.1, 400.13, and 600.13 MHz, respectively, and ω_L(¹³C) = 62.9,

100.58, and 150.86 MHz, respectively. Although δ_Hand δ_Cvalues for

chemical shifts were measured relative to tetramethylsilane (TMS), the

deuterated solvents were not doped with TMS but rather their residual

peaks were used to calibrate the ¹H and ¹³C NMR spectra, e.g.,

δ_H(CHCl₃) = 7.24 ppm and δ_C(CDCl₃) = 77.02 ppm. Coupling

constants (J) are reported in Hz. Structural assignments are made on

the basis of 2-D COSY, NOESY, HMQC, and HMBC. IR absorption

spectra were recorded on a FT IR spectrometer using KBr pellets,

except otherwise noted. UV/vis absorption spectra were measured on

a double-beam spectrometer. Melting points were measured on a

Kofler-type melting point microscope and are uncorrected. Photolyses

were conducted using a 700-W medium pressure Hg-arc lamp doped

with FeI₂and equipped with a Pyrex filter. Electron impact (EI)

spectra were obtained at 70 eV. Preparative GC was performed using

an instrument equipped with SE-30 or Carbowax/KOH glass columns.

Tandem GC-MS analyses were conducted on an instrument outfitted

with a column and a mass selective detector (70 eV) using helium as

carrier gas. Analytical gas chromatograms were obtained on a GC

instrument equipped with a flame-ionization detector and a capillary

column in order to determine product ratios, when applicable.

Preparation and Reactions of 3H-Diazirines. Ketone deriva-

tives 4 were usually synthesized from 5-hydroxyadamantan-2-one (4e),

but 5-methyladamantan-2-one (4c) was obtained in concentrated

H₂SO₄either by isomerization of 2-methyladamantan-2-ol (10) or by

oxidation of methyladamantane (12) (Scheme 3).⁶⁸

Photolyis of 3H-diazirines 1 produces a substantial amount

of diazo compounds 2, which may also be a source of carbenes

5. This appears to be unlikely, however, because their coupling

was probably responsible for the extensive formation of inert

azines 3 in aprotic CyH. In protic MeOH, the diazo com-

pounds were protonated faster than they could couple and no

azines 3 were formed. This side reaction had the potential of

producing cations 6, which were not under investigation. So,

the potent dipolarophile fumaronitrile was used to capture

dipolar 2 and permit the isomeric product ratios to be

attributed to carbene 5 alone. This does not preclude 5 from

being protonated to form 6, because the electron-deficient

carbene is predicted to behave as a nucleophile in hydroxylic

solvents.

These syntheses are well described and are to be found in the

literature. The yields of the oxidation were 12% for the first and 7% for

Scheme 3. Syntheses of 5-Methyladamantan-2-one (4c)

The differences in energy between proximal and distal

carbenes 5 decrease toward the intermolecular TS.⁶⁴This

b

implies that neither TS stabilization nor a TS model is

necessary to explain carbene 5 selectivity. Ground-state hyper-

conjugation alone is sufficient to explain the observed ratios,

although this might be different for a late TS.¹⁰For the cases

studied here, this is a natural consequence of the reaction and

increased interaction with the solvent and therefore the reduced

steering influence of the substituent within the adamantanty-

lidene moiety. It should be emphasized that both the Cieplak-

and Felkin−Anh-type stabilizations favor the same conformer,

but the Felkin−Anh type has less relevance for γ-substituted

the lower pathway. Note, however, that the upper pathway affords the

cleaner product, because byproducts are not easily removed in the

second pathway.

Via 2-methyl-2-adamantanol (10):^68gδ_H/ppm (2501 MHz,

CDCl₃) 1.30 (CH₃), 1.53 (2 H, d J 4.5), 1.55 (1 H), 1.65−2 (10 H,

1354

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

m), 2.2 (2 H, d J 12.26); δ_C/ppm (62.9 MHz, CDCl₃) 27.0, 27.4,

27.52, 32.9, 35.1, 38.3, 39.1, 73.8.; m/z (EI) 166 (0.08), 152 (10), 151

(100), 148 (11), 133 (6).

1-Methyladamantane (12): m/z (EI) 150 (14), 136 (10), 135

(100), 107 (6), 93 (20), 79 (11).

5-Methyladamantan-2-one (4c): mp 117−119 °C; δ_H/ppm

(250.1 MHz, CDCl₃) 0.92 (3 H, s), 1.7−1.82 (6 H, m), 1.95−2.01 (4

H, m), 2.1−2.16 (1 H, m), 2.53 (2 H, s); δ_C/ppm (62.9 MHz, CDCl₃)

28.1, 29.2, 30.0, 38.6, 43.4, 45.6, 46.6 (CO was not recorded); m/z

(EI) (13.16 min): 165 (12), 164 (100), 131 (20), 118 (12), 107 (12),

106 (16), 95 (17), 94 (20), 93 (76), 92 (20), 91 (20), 80 (14),

79 (29).

2,4-Didehydro-1-(trimethylsilyl)adamantane (as-7a): The

pyrolysis of 3H-diazirine 1a yielded intramolecular insertion products

s-7a (4%) and as-7a (10%) (isolated ratio =26:74). The separation was

performed by preparative GC (T_oven= 140 °C, T_injector= 260 °C):^21b

δ_H/ppm (250.1 MHz, CDCl₃) −0.02 (9 H, s), 1.19−1.29 (2 H, m),

1.31−1.37 (3 H, m), 1.41−1.45 (2 H, m), 1.47−1.51 (1 H, m), 1.80−

1.88 (3 H, m), 2.00−2.10 (1 H, m), 2.34−2.36 (1 H, m); δ_C/ppm

(62.9 MHz, CDCl₃) −4.2 (−Si(CH₃)₃), 21.4 (C-3), 25.3 (C-4), 25.6

(C-2), 26.8 (C-7), 27.7 (C-1), 28.7 (C-10), 33.1 (C-5), 33.3 (C-6),

33.6 (C-8), 53.6 (C-9); ν_max/cm⁻¹(CCl₄) 3022, 3003, 2954, 2915,

̅

2850, 1453, 1441, 1247, 1056, 1043, 930, 891, 869, 858, 848, 685; m/z

(EI) 208 ([(³⁰Si)M and (²⁹Si)M + H]⁺, 7%), 207 ([(²⁹Si)M and M +

H]⁺, 28), 206 (M⁺, 100), 191 (39), 132 (47), 117 (11), 92 (20), 91

(21), 79 (18), 74 (33), 73 (85); found M⁺206.1492, C₁₃H₂₂Si requires

206.1491.

The general synthesis of 3H-diazirines from their corresponding

ketones is well documented.²⁹The syntheses of γ-substituted azia-

damantanes 1a, 1d, 1e, 1f, 1g, and 1h are reported elsewhere.^24,69The

reactions with hydroxylamine-O-sulfonic acid (HOSA) in methanolic

ammonia were performed at T = −15 °C. The putative diaziridines

were not isolated, but directly oxidized. Compounds 1e, 1f, 1g, and 1h

are described elsewhere. Likewise, data for the MeOH-insertion pro-

2,4-Didehydro-7-(trimethylsilyl)adamantane (s-7a): δ_H/ppm

(250.1 MHz, CDCl₃) −0.10 (9 H, s), 1.15−1.25 (1 H, m), 1.28−1.32

(4 H, m), 1.38 (1 H, dm J 9.5), 1.45−1.50 (2 H, m), 1.70−1.77 (2 H,

m), 2.06−2.16 (1 H, m), 2.18−2.27 (2 H, m); δ_C/ppm (62.9 MHz,

CDCl₃) −5.7 (−Si(CH₃)₃), 18.0 (C-7), 19.5 (C-3), 23.5 (C-2/C-4),

ducts from carbenes 5a, 5c, 5e, and 5h,¹⁶as well as some of the

a

28.1 (C-10), 31.2 (C-1/C-5), 32.8 (C-6/C-8), 52.4 (C-9); ν_max/cm⁻¹

intramolecular insertion products may be found elsewhere.^20,23All

data, except for 1h and derivatives of 1e, are listed in the Supporting

Information. Photolyses were performed in anhydrous cyclohexane or

anhydrous MeOH. Argon gas was bubbled through the solutions for

10 min, and the diazirines were photolyzed for 2.5 h at T = 12−14 °C.

The mixtures were analyzed by analytical GC and afterward separated

or purified either by liquid-chromatography or by preparative GC

(vide infra).

̅

(CCl₄) 3026, 2988, 2954, 2923, 2846, 1247, 934, 875, 858, 846; m/z

(EI) 207 ([(²⁹Si)M and M + H]⁺, 5%), 206 (M⁺, 25), 79 (27), 73

(100); found M⁺206.1488, C₁₃H₂₂Si requires 206.1491.

anti-4-Cyclohexyl-1-(trimethylsilyl)adamantane (anti-8a):

Photolysis of 3H-diazirine 1a in CyH afforded a mixture of syn- and

anti-4-cyclohexyl-1-(trimethylsilyl)adamantane syn-8a/anti-8a (ratio

=38:62) in 64% yield. In addition, 18% of azines 3a were formed.

The insertion products were separated by preparative GC on a 15%

SE-30 column (T_oven= 250 °C, T_injector= 220 °C): mp 86.5−88.5 °C;

δ_H/ppm (600.13 MHz, C₆D₆)⁷⁰0.08 (9 H, s), 0.88−0.90 (2 H, m),

1.25−1.45 (4 H, m), 1.57 (2 H, d J 12.5), 1.60−1.66 (1 H, m), 1.68 (2

H, d J 12), 1.74 (2 H, s), 1.78−1.83 (1 H, m), 1.84−1.90 (5 H, m),

1.93−2.01 (4 H, m), 2.03 (2 H, s); δ_C/ppm (150.9 MHz, C₆D₆)⁷¹

−5.6 (−Si(CH₃)₃), 20.9 (C-1), 26.8 (C-3′/C-5′), 27.0 (C-4′), 27.5

(C-7), 28.1 (C-3/C-5), 30.8 (C-2′/C-6′), 32.0 (C-6/C-10), 36.6

Gas-Phase Pyrolysis Experiments. The gas-phase pyrolysis

experiments were performed either in a classical manner (method 1)

or in the injector unit (T = 270 °C) of a preparative GC (method 2).

Method 1. A three-necked 100-mL round-bottom flask equipped

with a twistable solids addition flask, a stopper, and a trap immersed in

an ice−water bath was evacuated to 0.5 Torr. The solids addition flask

was charged with the diazirine and the round-bottom flask was heated

to T = 250−300 °C. The solid diazirine was slowly released from the

solids addition flask onto the heated surface of the round-bottom flask

in small portions. The products from the decomposition were col-

lected in a trap and in the glass tube connecting the trap with the

round-bottom flask.

Method 2. An aliquot containing 40 mg of diazirine in 80 μL of

CHCl₃was slowly injected into the injector unit (column 15% SE-30

at 180 °C, gas flow at 80 mL/min). Working on a preparative GC can

cause a substantial loss of substance for two reasons: (1) the volatility

of the formed adamantane derivatives and (2) the suspected formation

of azine in the injector block. When using a lower gas flow and a lower

column temperature, the number of byproducts increased. With the

exception of compounds s- and as-7a and s- and as-7d, pyrolytic

insertion products could not be separated.

(C-1′), 37.9 (C-8), 39.1 (C-2/C-9), 49.7 (C-4); ν_max/cm⁻¹(KBr)

̅

2920, 2850, 1632, 1445, 1342, 1316, 1255, 1240, 941, 867, 832, 741,

684, 623; m/z (EI) 292 ([(³⁰Si)M and (²⁹Si)M + H]⁺, 6%), 291

([(²⁹Si)M and M + H]⁺, 24), 290 (M⁺, 100), 217 (15), 165 (25), 162

(57), 149 (11), 135 (11), 79 (14), 73 (86); found M⁺290.2434,

C₁₉H₃₄Si requires 290.2430. Anal. Calcd for C₁₉H₃₄Si: C, 78.54; H,

11.79. Found: C, 78.60; H, 11.48.

syn-4-Cyclohexyl-1-(trimethylsilyl)adamantane (syn-8a): mp

83.5−85.0 °C; δ_H/ppm (250.1 MHz, C₆D₆)⁷⁰0.07 (9 H, s), 0.80−1.02

(2 H, d J 13), 1.25−1.35 (4 H, m), 1.4−1.5 (2 H, d J 14), 1.55−1.70

(1 H, m), 1.74 (2 H, s), 1.81−2.13 (14 H, m); δ_C/ppm (62.9 MHz,

C₆D₆)⁷¹−5.5 (−Si(CH₃)₃), 20.6 (C-1), 26.9 (C-3′/ C-5′), 27.1 (C-4′),

28.0 (C-7), 28.3 (C-3/C-5), 31.1 (C-2/C-9), 31.6 (C-2′/ C-6′), 36.9

(C-1′), 38.1 (C-8), 39.7 (C-6/C-10), 49.8 (C-4); ν_max/cm⁻¹(KBr)

̅

The azines were synthesized as reference compounds with the

following method: 0.11 mmol of the corresponding diazirine was

photolyzed in a rotating flask cooled to 5 °C for 2 h. The crude

product was dissolved in chloroform and analyzed by GC and GC−

MS. For the NMR analyses the crude product was washed with hot

petroleum ether and dried in vacuo.

2932, 2902, 2841, 1447, 1342, 1254, 1243, 976, 910, 890, 868, 826,

740; m/z (EI) 292 ([(³⁰Si)M and (²⁹Si)M + H]⁺, 7%), 291 ([(²⁹Si)M

and M + H]⁺, 27), 290 (M⁺, 100), 217 (18), 165 (39), 149 (16), 135

(11), 83 (10), 73 (80); found M⁺290.2425, C₁₉H₃₄Si requires

290.2430. Anal. Calcd for C₁₉H₃₄Si: C, 78.54; H, 11.79. Found: C,

78.47; H, 11.57.

Preparation and Reactions of 4-Azi-1-(trimethylsilyl)-

adamantane (1a):^21bThe synthesis of 3H-diazirine 1a, as described

elsewhere,^21bwas achieved in a yield of 43%: mp 38−39.5 °C; δ_H/ppm

(250.1 MHz, CDCl₃) −0.04 (9 H, s), 0.64 (2 H, s), 1.60−1.71 (4 H,

m), 1.74 (2 H, d J 11), 1.92−2.15 (5 H, m); δ_C/ppm (62.9 MHz,

CDCl₃) −5.4 (−Si(CH₃)₃), 20.9 (C-1), 27.1 (C-7), 34.1 (C-3/C-5),

4-Methoxy-1-(trimethylsilyl)adamantane (9a): Photolysis of

3H-diazirine 1a in MeOH afforded a mixture of syn- and anti-4-

methoxy-1-(trimethylsilyl)adamantane (oil) syn-9a/anti-9a in 93%

yield (ratio =31:69). In addition, 3% of the intramolecular insertion

products s-7a and as-7a were formed. These were separated by

preparative GC on a 15% SE-30 column (T_oven= 180 °C, T_injector

=

35.0 (C-2/C-9), 35.2 (C-6/C-10), 36.4 (C-8), (C-4 not obsd); ν_max

/

220 °C). Samples with purities of 100% (anti-9a) and 96% (syn-9a)

were obtained.

̅

cm⁻¹(KBr) 2921, 2849, 1571, 1450, 1243, 1132, 1080, 1056, 997, 950,

866, 831; λ_max/nm (C₆H₁₂) 356 (ε/(10⁴M⁻¹cm⁻¹) 0.17), 362 (ε/(10⁴

M⁻¹cm⁻¹) 0.14), 376 (ε/(10⁴M⁻¹cm⁻¹) 0.23); m/z (EI) 219 ([M −

CH₃]⁺, 0.6%), 206 ([M − N₂]⁺, 2), 132 (9), 91 (9), 74 (9), 73 (100);

found [M − CH₃]⁺219.1319, C₁₂H₁₉N₂Si requires 219.1318). Anal.

Calcd for C₁₃H₂₂N₂Si: C, 66.61; H, 9.46; N, 11.95. Found: C, 66.92;

H, 9.59; N, 12.14.

anti-4-Methoxy-1-(trimethylsilyl)adamantane (anti-9a): δ_H/ppm

(250.1 MHz, CDCl₃) −0.08 (9 H, s), 1.38−1.6 (6 H, m), 1.68−1.77

(3 H, m), 1.94−2.08 (4 H, m), 3.3 (1 H, s), 3.35 (3 H, s); δ_C/ppm

(62.9 MHz, CDCl₃) −5.4 (−Si(CH₃)₃), 20.3 (C-1), 26.9 (C-7), 30.8

(C-3/C-5), 31.4 (C-6/C-10), 36.4 (C-2/C-9), 37.0 (C-8), 55.2

(−OCH₃), 83.4 (C-4); ν_max/cm⁻¹(CCl₄) 2896, 2844, 1447, 1383,

̅

1355

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

1255, 1196, 1152, 1099, 1085, 1073, 934, 914, 866, 837; m/z (EI)

239 (M+1, 19), 238 (M⁺, 100), 223 (17), 206 (16), 165 (12), 134

(11), 133 (20), 110 (20), 92 (13), 91 (15), 89 (35), 79 (23), 78

(10), 73 (92); found M⁺238.1749, C₁₄H₂₆OSi requires 238.1753.

Anal. Calcd for C₁₄H₂₆OSi: C, 70.52; H, 10.99. Found: C, 70.53;

H, 10.84.

mixture: syn-4-cyclohexyl-1-methyladamantane (syn-8c) (major) and

anti-4-cyclohexyl-1-methyl-adamantane (anti-8c) (minor). The dia-

stereomers were differentiated using 2-D NMR techniques: δ_H/ppm

(600.1 MHz, CDCl₃) 0.73−0.74 (5 H, m), 1.07−1.18 (4 H, m), 1.20

(2 H, d J 13), 1.37−1.38 (3 H, m), 1.49 (2 H, d J 12), 1.58 (2 H, d J

11.5), 1.61−1.75 (5 H, m), 1.77−1.83 (2 H, m), 1.88−1.90 (1 H, m),

1.95 (2 H, s); δ_C/ppm (150.9 MHz, CDCl₃) 26.6 (C-3′/ C-5′), 26.8

(C-4′), 28.8 (C-7), 29.2 (C-3/C-5), 29.4 (C-1), 30.8 (C-2′/ C-6′), 31.5

(−CH₃), 36.5 (C-1′), 38.6 (C-6/C-10), 39.0 (C-2/C-9), 45.4 (C-8),

syn-4-Methoxy-1-(trimethylsilyl)adamantane (syn-9a): δ_H/

ppm (250.1 MHz, CDCl₃) −0.13 (9 H, s), 1.31 (2 H, d J 12.5),

1.56 (1 H, s), 1.62−1.70 (2 H, d J 11), 1.72−1.93 (5 H, m), 1.98 (2 H, s),

3.32 (4 H, s); δ_C/ppm (62.9 MHz, CDCl₃) −5.5 (−Si(CH₃)₃), 20.2

(C-1), 26.9 (C-7), 30.8 (C-3,C-5), 30.8 (C-2,C-9), 36.4 (C-6,C-10),

48.8 (C-4); ν_max/cm⁻¹(CCl₄) 2920, 2849, 1552, 1455, 1358, 1260,

̅

1219, 1110, 1030, 982; m/z (EI) 232 (M⁺, 14), 217 (25), 150 (15),

149 (100), 93 (17), 82 (21); found M⁺232.2188, C₁₇H₂₈requires

232.2191.

37.0 (C-8), 55.3 (−OCH₃), 83.2 (C-4); ν_max/cm⁻¹(CCl₄) 2905, 2849,

̅

2822, 1463, 1447, 1382, 1248, 1190, 1106, 1073, 902, 867, 841; m/z

(EI) 239 (M+1, 17), 238 (M⁺, 81), 223 (84), 206 (16), 165 (16), 134

(16), 133 (16),110 (15), 92 (20), 91 (23), 89 (37), 79 (34), 78 (16),

73 (100); found M⁺238.1747, C₁₄H₂₆OSi requires 238.1753.

anti-4-Cyclohexyl-1-methyladamantane (anti-8c): δ_H/ppm

(600.1 MHz, CDCl₃) 0.75−0.77 (5 H, m), 1.08−1.09 (1 H, m),

1.18−1.22 (3 H, m), 1.31 (2 H, d J 12.6), 1.38−1.42 (4 H, m), 1.44−

1.46 (1 H, m), 1.52 (2 H, s), 1.62−1.64 (1 H, m), 1.70−1.74 (4 H, m),

1.81−1.85 (3 H, m), 1.93 (2 H, s); δ_C/ppm (150.9 MHz, CDCl₃) 26.7

(C-3′/ C-5′), 26.8 (C-4′), 28.5 (C-7), 29.1 (C-3/C-5), 29.7 (C-1), 31.0

(CH₃, C-2′/ C-6′), 31.3 (C-6/C-10), 36.4 (C-1′), 45.3 (C-8), 46.3 (C-

2/C-9), 49.1 (C-4); m/z (EI) 232 (M⁺, 15), 217 (27), 150 (13), 149

(100), 93 (17), 82 (23).

The products from the insertion reaction of 5-methyladamantan-2-

ylidene into MeOH (syn-9c/anti-9c) were isolated in a total yield of

84%. In addition, (according to GC analysis) 5% of the pyrolytic

insertion products syn-9c and anti-9c were formed. After chromatog-

raphy with pentane/CH₂Cl₂(9:1) on silica gel 60 (230−400 mesh)

the mixture was analyzed: syn-4-methoxy-1-methyladamantane (syn-

9c) and anti-4-methoxy-1-methyladamantane (anti-9c) were formed in

a ratio of 3:2

syn-4-Methoxy-1-methyladamantane (syn-9c): δ_H/ppm

(600.1, CDCl₃) 0.72 (3 H, s), 1.18 (2 H, d J 12.4), 1.39 (2 H, s),

1.58 (2 H, d J 12), 1.69−1.72 (4 H, m), 1.85 (1 H, s), 2.06 (2 H, s),

3.22 (1 H, s), 3.30 (3 H, s): δ_C/ppm (150.9 MHz, CDCl₃) 28.4 (C-7),

29.5 (C-1), 31.4 (−CH₃), 32.3 (C-3/C-5), 36.1 (C-6/C-10), 38.7

(C-2/C-9), 44.9 (C-8), 55.7 (−OCH₃), 82.8 (C-4); m/z (EI) 180

(M⁺, 15), 165 (5), 149 (18), 148 (100), 133 (23), 107 (13), 106 (24),

105 (20), 93 (31), 92 (26), 91 (20), 79 (18). From the mixture: found

M⁺180.1517, C₁₂H₂₀O requires 180.1514.

5-(Trimethylsilyl)adamantan-2-one azine (3a):³²mp 188−

190 °C; δ_H/ppm (40013 MHz, CDCl₃) −0.13 (9 H, s), 1.64−1.94 (11

H, m), 2.58 (1 H, s), 3.23 (1 H, s); δ_C/ppm (100.6 MHz, CDCl₃)

−5.4 (-Si(CH₃)₃), 21.4 (C-5), 31.4, 39.2 (C-1/C-3), 36.2 (C-6), 37.7,

37.9 (C-4/C-4′), 38.2, 38.3 (C-10/C-10′), 39.04, 39.07 (C-9/C-9′),

27.3 (C-7), 39.62, 39.64 (C-8/C-8′), 171.4, 171.6 (C-2/C-2′); ν_max

/

̅

cm⁻¹(KBr) 2922, 2840, 1645, 1455, 1342,1250, 1077, 995, 961, 865,

832, 738, 626; m/z (EI) 440 (M⁺, 12%), 275 (10), 268 (14), 238 (10),

237 (40), 236 (32), 147 (12), 145 (12), 131 (12), 86 (22), 84 (34), 73

(100); found M⁺440.3055, C₂₆H₄₄N₂Si₂requires 440.3043.

Preparation and Reactions of 4-Azi-1-methyladamantane

(1c): 3H-Diazirine 1c was synthesized from ketone 4c (52% yield)⁷²

and purified by chromatography (silica gel 60, 230−400 mesh) using

pentane as eluant to give a transparent, low-melting solid: mp 41−42 °C;

δ_H/ppm (400.13 MHz, CDCl₃) 0.64 (2 H, s), 0.86 (3 H, s), 1.44

(2 H, dm J 13), 1.52 (2 H, s), 1.62 (2 H, dm J 12.5), 1.79 (2 H, d J

11.6), 1.99 (2 H, d J 11.6), 2.07−2.11 (1 H, m); δ_C/ppm (100.6 MHz,

CDCl₃) 28.2 (C-7), 30.0 (C-1), 30.5 (−CH₃), 34.5 (C-6/C-10), 34.8

(C-3/C-5), 35.3 (C-4), 41.9 (C-2/C-9), 43.9 (C-8); ν_max/cm⁻¹(KBr)

̅

2920, 2849, 1574, 1538, 1455, 1376, 1358, 1098, 984, 934, 879, 842,

626; λ/_max/nm (c-C₆H₁₂) 353 (ε/(10⁴M⁻¹cm⁻¹) 0.16), 359 (ε/(10⁴

M⁻¹cm⁻¹) 0.13), 372 (ε/(10⁴M⁻¹cm⁻¹) 0.22); m/z (EI) 148 ([M −

N₂]⁺, 19), 133 (27), 107 (29), 106 (42), 105 (48), 94 (19), 93 (69),

92 (89), 91 (100), 81 (12), 80 (28), 79 (60); found [M − N₂]⁺

148.1249, C₁₁H₁₆requires 148.1252). Anal. Calcd for C₁₁H₁₆N₂: C,

74.96; H, 9.15; N, 15.89. Found: C, 75.06; H, 9.07; N, 16.04.

Gas-phase pyrolysis of 3H-diazirine 1c afforded s-7c and as-7c in a

yield of 52%. Separation of the products s-7c and as-7c (2:1,

respectively by NMR) was not possible. Even analytical GC was not

able to separate the peaks at 50 °C, where they only started to get

slightly asymmetrical. The mixture of the insertion products, therefore,

had to be analyzed by 2-D NMR and proved to be identical with

published results.²⁰The IR shows the characteristic bands for the

three-membered rings at 3027 cm⁻¹.

anti-4-Methoxy-1-methyladamantane (anti-9c): δ_H/ppm

(600.1, CDCl₃) 0.77 (3 H, s), 1.32 (2 H, d J 12), 1.36−1.38 (2 H,

m), 1.44 (2 H, s), 1.49 (2 H, d, J 13.4), 1.83 (1 H, s), 1.94 (2 H,

J 11.3), 2.01 (2 H, s), 3.26 (1 H, s), 3.31 (3 H, s); δ_C/ppm (150.9

MHz, CDCl₃) 28.4 (C-7), 29.6 (C-1), 30.8 (−CH₃), 31.1 (C-6/C-10),

32.0 (C-3/C-5), 43.6 (C-2/C-9), 44.9 (C-8), 55.8 (−OCH₃), 83.4

(C-4); ν_max/cm⁻¹(CCl₄) from the mixture (anti-9c /syn-9c) 2915,

̅

2849, 1454, 1381, 1197, 1104, 1086; m/z (EI) 180 (M⁺, 14), 165 (6),

149 (18), 148 (100), 133 (23), 107 (11), 106 (24), 105 (21), 93 (30),

92 (25), 91(20).

5-Methyladamantan-2-one azine (3c): δ_H/ppm (400.13 MHz,

CDCl₃) 0.76−0.81 (3 H, s), 1.50−2.10 (11 H, m), 2.59 (1 H, s), 3.24

(1 H, s); δ_C/ppm (100.6 MHz, CDCl₃) 39.6, 31.6 (C-1/C-3); 43.3

(C-6); 44.50, 44.53, 45.7 (C-4/C-9); 38.6, 37.3 (C-8/C-10), 28.5 (C-7),

30.0 (C-5); 170.9, 170.8 (C-2, C-2′); IR (KBr) 2908, 2844, 1647,

1453, 1375, 1356, 1182, 1104, 1084, 939 cm⁻¹; (EI) m/z 325 (M+1,

17), 324 (M⁺, 76), 323 (15), 281 (22), 267 (43), 253 (15), 229 (15),

240 (25), 190 (15), 189 (54), 164 (100), 163 (28), 162 (35); found

M⁺324.2572, C₂₂H₃₂N₂requires 324.2565.

2,4-Didehydro-7-methyladamantane (s-7c): δ_H/ppm (600.1,

CDCl₃) 0.73 (3 H, s), 1.05 (2 H, s), 1.12 (2 H, d J 12.5), 1.31−1.33 (2

H, m), 1.39 (2 H, s), 1.49 (2 H, s), 2.04 (1 H, s), 2.29 (2 H, s); δ_C/ppm

(150.9 MHz, CDCl₃) 23.8 (C-2.C-4), 24.0 (C-3), 30.0 (C-7), 31.4

(−CH₃), 33.5 (C-1, C-5), 36.5 (C-10), 41.0 (C-6,C-8), 52.5 (C-9).

2,4-Didehydro-1-methyladamantane (as-7c): δ_H/ppm (600.1,

CDCl₃) 1.07−1.08 (4 H, m), 1.12 (1 H, s), 1.23 (1 H, d J 9.8), 1.26−

1.38 (4 H, m), 1.49 (1 H, s), 1.79 (2 H, s), 1.90 (1 H, s), 2.02 (1 H,

s_b), 2.22 (1H, s); δ_C/ppm (150.9 MHz, CDCl₃) 20.5 (C-3), 24.9 (C-

4), 27.2 (C-7), 27.3 (−CH₃), 28.8 (C-10), 29.1 (C-2), 32.5 (C-6),

Preparation and Reactions of 4-Aziadamantan-1-amine

(1d):^24a3H-Diazirine 1d was synthesized in 26% yield from ketone

4d. The crude product, which contained 7% of 4-aziadamantan-1-ol

byproduct, was chromatographed (silica gel 60, 230−400 mesh) using

a mixture of methanolic NH₃(0.7 M) and EtOAc (7:11) as eluant and

a solid, which sublimed at reduced pressure, was obtained: mp 130 °C

dec; δ_H/ppm (250.1 MHz, CDCl₃) 0.76 (2 H, s), 1.61 (2 H, d J 13.6),

1.67 (2 H, d J 14.8), 1.70 (2 H, s), 1.96 (2 H, d J 12.3), 2.01 (2 H, d J

11.3), 2.25 (1 H, s), (NH₂-group not obsd); δ_C/ppm (62.9 MHz,

CDCl₃) 29.1 (C-7), 34.0 (C-6/C-10), 35.5 (C-3/C-5), 43.6 (C-2/C-

32.6 (C-5), 37.1 (C-1), 41.0 (C-8), 59.0 (C-9); ν_max/cm⁻¹(CCl₄)

̅

3027, 2995, 2924, 2847, 1454, 1372, 1358, 1336, 1034, 907; m/z (EI)

149 (M+1, 7), 148 (M⁺, 69), 133 (32), 119 (15), 107 (34), 106 (48),

105 (40), 94 (31), 93 (61), 92 (63), 91 (71), 80 (18), 79 (100), 77

(35); found M⁺148.1255, C₁₁H₁₆requires 148.1252.

syn-4-Cyclohexyl-1-methyladamantane (syn-8c): The crude

product mixture was chromatographed (silica gel 60, 230−400 mesh)

using hexane as eluant and gave syn-8c/anti-8c (ratio =3:2, respec-

tively) in 65% (35% isolated yield). In addition, azines 3c (19−29%)

and ketone 4c (10−19%) were formed. Analyses were done upon the

9), 45.4 (C-8), 47.4 (C-1), (C-4 not obsd); ν_max/cm⁻¹(KBr) 3355,

̅

3265, 2922, 2854, 1575, 1449, 1350, 1293, 1132, 1108, 1067, 926;

1356

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

λ_max/nm (c-C₆H₁₂) 352 (ε/(10⁴M⁻¹cm⁻¹) 0.26), 356 (ε/(10⁴

M⁻¹cm⁻¹) 0.24), 370 (ε/(10⁴M⁻¹cm⁻¹) 0.31); m/z (CI) 178 ([M +

H]⁺, 100), 165 (64), 149 (92). Anal. Calcd for C₁₀H₁₅N₃: C, 67.76; H,

8.53; N, 23.71. Found: C, 67.61; H, 8.50; N, 23.45.

2,4-Didehydroadamantan-7-amine (s-7d): Pyrolysis of 3H-

diazirine 1d gave 1,3-CH insertion products s-7d and as-7d in 89%

5-Aminoadamantan-2-one azine (3d): mp 249−253 °C; δ_H/ppm

(400.13 MHz, CDCl₃) 1.50−1.90 (10 H, m), 2.20 (1 H, s), 2.59

(1 H, s), 3.24 (1 H, s), (NH₂-group not obsd); δ_C/ppm (100.6 MHz,

CDCl₃) 29.5 (C-7), 32.1 (C-3), 36.93, 36.96 (C-10/C-10′), 38.2 (C-8),

40.1 (C-1), 45.24, 45.27 (C-6/C-6′), 45.8, 45.87 (C-4/C-4′), 46.93,

46.94 (C-9/C-9′), 47.4 (C-NH₂), 170.1 (C-2); ν_max/cm⁻¹(KBr) 3442,

̅

yield (ratio = 4:1). By using preparative GC (Carbowax/KOH, T_oven

=

2924, 2850, 1644, 1445, 1357, 1141, 1089, 1074, 955; m/z (EI) 327

([M + H]⁺, 15.5), 326 (M⁺, 100), 205 (10), 165 (11), 163 (52), 94

(22); found M⁺326.2463, C₂₀H₃₀N₄requires 326.2470.

140 °C, T_injector= 260 °C) a sample of the major product was obtained

(the minor product was enriched, vide infra): δ_H/ppm (600.13 MHz,

CDCl₃) 1.16 (2 H, d J 12.1), 1.27−1.31 (2 H, m), 1.34 (1 H, dm J 10),

1.40 (2 H, m), 1.5 (1 H, m), 1.61−1.64 (2 H, m), 2.02−2.07 (1 H, m),

2.41 (2 H, s), (NH₂-group not obsd); δ_C/ppm (150.9 MHz, CDCl₃)

23.0 (C-2/C-4), 25.6 (C-3), 34.4 (C-1/C-5), 37.5 (C-10), 42.6 (C-6/

Preparation and Reactions of 4-Aziadamantan-1-ol (1e):^21a

3H-Diazirine 1e was synthesized in 46% yield from ketone 4e.⁷³

Purification by chromatography with Et₂O/ligroin (8:2) on silica gel

60 (230−400 mesh) gave a readily crystallizing white solid (mp

140 °C dec). The pyrolytic reactions of 1e afforded intramolecular

insertion products s-7e and as-7e in 94% yield (ratio =9:1). Analyses

were done on the mixture. The products syn-8e and anti-8e (ratio =

89:11) from the insertion reaction of 5-hydroxyadamantan-2-ylidene

into cyclohexane were isolated in a total yield of 92%. In addition,

traces of azines 3e were formed and 6% of ketone 4e. After chro-

matography on silica gel 60 (230−400 mesh) with ligroin/EtOAc

(85:15), syn-cyclohexyladamantan-1-ol (syn-8e) was isolated in 21%

yield. Photolysis of 1e in MeOH afforded syn-9e and anti-9e in 97%

yield (ratio = 3:1). After chromatography on silica gel 60 (230−400

mesh) using ligroin/EtOAc (8:2), syn-9e was isolated in 21% yield and

anti-9e in 7% yield.

C-8), 48.2 (C-7), 51.5 (C-9); ν_max/cm⁻¹of mixture (CDCl₃) 3356,

̅

3032, 2934, 2850, 1442, 1356, 1334, 1141; m/z (EI) 150 ([M + H]⁺,

11), 149 (M⁺, 100), 148 (37), 134 (19), 117 (16), 108 (16), 107 (66),

106 (97), 94 (56), 93 (13), 91 (17), 77 (13), 61 (14); found M⁺

149.1199, C₁₀H₁₅N requires 149.1204.

2,4-Didehydroadamantan-1-amine (as-7d): δ_H/ppm (600.13

MHz, CDCl₃) 1.18−1.30 (3 H, m), 1.31−1.35 (2 H, m), 1.46 (1 H, d

J 12), 1.53−1.59 (2 H, m), 1.76−1.85 (2 H, m), 2.02 (1 H, s), 2.21−

2.23 (2 H, m), (NH₂-group not obsd); δ_C/ppm (150.9 MHz, CDCl₃)

20.2 (C-3), 24.3 (C-4), 27.4 (C-7), 28.4 (C-10), 30.2 (C-2), 31.6 (C-

5), 31.9 (C-6), 42.0 (C-8), 55.2 (C-1), 59.5 (C-9); m/z (EI) 149 (M⁺,

6.6), 94 (12), 93 (13), 61 (100), 60 (10); found M⁺149.1209,

C₁₀H₁₅N requires 149.1204.

5-Hydroxyadamantan-2-one azine (3e): mp 250 °C (sublimes),

305 °C dec; δ_H/ppm (250.13 MHz, CDCl₃) 3.47 (1 H, “s”), 2.82 (1

H, “s”), 2.30 (1 H, “s”), 1.60−1.95 (11 H, m); δ_C/ppm (62.9 MHz,

syn-4-Cyclohexyladamantan-1-amine (syn-8d): Product mix-

ture syn-8d/anti-8d (ratio = 89:11, respectively) was isolated in 48%

yield. In addition, 32% of ketone 4d^24awas formed. Separation using

preparative GC (Carbowax/KOH) failed. Only the major product

could be enriched (>90%): δ_H/ppm (600.13 MHz, CDCl₃) 0.60−0.75

(2 H, m), 1.05−1.18 (4 H, m), 1.29 (2 H, d J 11.7) 1.33−1.39 (1 H,

m), 1.48−1.58 (4 H, m), 1.58−1.65 (5 H, m), 1.66 (2 H, m), 1.78 (2

H, d J 11.7), 2.01 (1 H, s), 2.07 (2 H, s), (NH₂-group not obsd); δ_C/

ppm (150.9 MHz, CDCl₃) 26.6 (C-3′/ C-5′), 26.7 (C-4′), 29.8 (C-7),

30.3 (C-3/C-5), 30.7 (C-2′/ C-6′), 36.5 (C-1′), 37.9 (C-6/C-10), 40.6

CDCl₃) 30.7, 33.0, 37.2, 38.5, 41.0, 44.7, 45.0, 46.1, 68.2, 170.1; ν_max

/

̅

cm⁻¹(KBr) 3386, 2923, 2853, 1646, 1451, 1352, 1291, 1115, 1093,

1067, 971, 926; m/z (EI) 329 ([M + H]⁺, 21), 328 (M⁺, 88), 282

(11), 279 (14), 167 (24), 166 (55), 165 (11), 164 (27), 151 (13), 150

(17), 149 (59), 148 (12), 109 (17), 108 (21), 107 (17), 97 (25), 96

(24), 95 (100), 94 (26), 83 (20), 71 (29), 70 (20), 69 (30), 67 (25),

57 (45), 55 (40); found M⁺328.2157, C₂₀H₂₈N₂O₂requires 328.2151.

Reactions of 4-Azi-1-bromoadamantane (1f) and 4-Azi-1-

chloroadamantane (1g): The preparation and reaction of 3H-

diazirines 1f and 1g were performed in the usual manner (vide

supra).^24b,68Their respective reaction products were studied using GC

and GC−MS. The insertion products with MeOH are described

(C-2/C-9) 47.0 (C-8), 48.4 (C-4), (C-1 not obsd); ν_max/cm⁻¹

̅

(CDCl₃) 3355, 2921, 2851, 1449, 1359, 1304, 1182, 1115, 1098,

1048, 990, 952, 932, 912, 734, 530; m/z (EI) 234 ([M + H]⁺, 7), 233

(M⁺, 43), 176 (15), 95 (10), 94 (100), 61 (22); found M⁺233.2149,

C₁₆H₂₇N requires 233.2143.

elsewhere.¹⁷

a

Intramolecular Insertion Reaction Products: The intramolec-

ular insertion products are formed in the injector unit of the analytic

GC and GC−MS, respectively: s-7f m/z (EI) 214, 212 (M⁺, 6), 134

(10), 133 (91), 117 (11), 105 (18), 91 (65), 79 (100), 78 (62); as-7f

m/z (EI) 214, 212 (M⁺, 12,10), 172, 170 (12), 157 (14), 133 (100),

117 (28), 115 (12), 105 (19), 103 (12), 92 (18), 91 (60), 79 (33), 77

(19); s-7g m/z (EI), 168 (M⁺, 16), 133 (11), 127 (10), 126 (20), 117

(15), 105 (14), 92 (25), 91 (76), 79 (100), 78 (74), 77 (25).

Insertion into Cyclohexane: syn-8g m/z (EI) 252 (M⁺, 5) 218

(18), 217 (100), 169 (22), 135 (20), 134 (22), 133 (23), 93 (14), 92

(14), 91 (34), 83 (25), 82 (37), 79 (31); anti-8g m/z (EI) 252 (M⁺, 9)

218 (19), 217 (100), 169 (20), 135 (19), 133 (27), 115 (10), 105

(11), 93 (14), 92 (10), 91 (33), 83 (22), 82 (32), 79 (31); syn-8f m/z

(EI) 218 (M⁺-Br, 20), 217 (100), 161 (7), 149 (8), 135 (29), 121

(16). 93 (15), 91 (14), 81 (10), 79 (20); anti-8f m/z (EI) 218 (M⁺−

Br, 14), 217 (100), 161(7), 135 (32), 121 (13), 93 (11), 91 (14), 81

(13), 79 (17).

Insertion into Methanol: syn-9f m/z (EI) 166 (M + 1 − Br, 11),

165 (M⁺− Br, 100), 133 (12), 109 (9), 105 (10), 91 (11); anti-9f m/z

(EI) 166 (M + 1 − Br, 12), 165 (M⁺− Br, 100), 133 (45), 105 (12),

91 (27), 79 (13); rearr-9f m/z (EI) 214, 212 (M⁺, OCH₃, 20), 165

(38), 133 (100), 109 (13), 105 (14), 91 (38), 79 (44), 78 (26); syn-9g

m/z (EI) 200 (M⁺, 13), 170 (31), 168 (100), 165 (23), 133 (40), 126

(17), 105 (11), 92 (16), 91 (43), 79 (28), 78 (17); anti-9g m/z (EI)

200 (M⁺, 21), 170 (36), 168 (100), 165 (21), 133 (42), 126 (21), 105

(11), 92 (17), 91 (48), 79 (24), 78 (17).

anti-4-Cyclohexyladamantan-1-amine (anti-8d): δ_H/ppm

(600.13 MHz, CDCl₃) 0.64−0.78 (2 H, m), 1.02−1.18 (4 H, m),

1.23−1.27 (2 H, m), 1.35−1.41 (1 H, m), 1.43−1.57 (4 H, m), 1.58−

1.68 (5 H, m), 1.69−1.73 (2 H, m), 1.77−1.79 (2 H, m), 1.96 (1 H, s),

2.03 (2 H, s), (NH₂-group not obsd); δ_C/ppm (150.9 MHz, CDCl₃)

26.6 (C-3′/ C-5′), 26.7 (C-4′), 29.3 (C-7), 30.0 (C-3/C-5), 30.6 (C-2′/

C-6′), 30.9 (C-6/C-10), 36.3 (C-1′), 47.2 (C-8), 47.5 (C-2/C-9), 48.4

(C-4); (C-1 not obsd); m/z (EI) 233 (M⁺, 30), 176 (20), 94 (100);

found M⁺233.2151, C₁₆H₂₇N requires 233.2143.

syn-4-Methoxyadamantan-1-amine (syn-9d): Photolysis of

3H-diazirine 1d in MeOH afforded syn-9d and anti-9d in a combined

yield of 93% (ratio = 3:1). Separation using preparative GC

(Carbowax/KOH) failed. However, enriched fractions fractions of

the major (98%) and minor (70%) products were analyzed: δ_H/ppm

(400.13 MHz, CDCl₃) 1.15 (2 H, s), 1.24 (2 H, d J 12), 1.45−1.53 (4

H, m), 1.62 (2 H, d J 11.8), 1.77 (2 H, d J 11.1), 1.91 (1 H, s), 2.12 (2

H, s), 3.10−3.14 (1 H, m), 3.24 (3 H, s); δ_C/ppm (100.6 MHz,

CDCl₃) 28.9 (C-7), 33.0 (C-3/C-5), 35.1 (C-6/C-10), 40.4 (C-2/C-

9), 45.6 (C-8), 55.2 (−OCH₃), 81.5 (C-4), (C-1 not obsd); ν_max/cm⁻¹

̅

of mixture (CCl₄) 3367, 2912, 2853, 1615, 1450, 1360, 1106; m/z (EI)

182 ([M + H]⁺, 8), 181 (M⁺, 61), 124 (24), 95 (10), 94 (100); found

M⁺181.1472, C₁₁H₁₉NO requires 181.1467.

anti-4-Methoxyadamantan-1-amine (anti-9d): δ_H/ppm

(400.13 MHz, CDCl₃) 1.20−1.27 (4 H, m), 1.50−1.55 (6 H, m),

1.85−1.92 (3 H, m), 2.05 (2 H, s), 3.21−3.27 (4 H, m); δ_C/ppm

(100.61 MHz, CDCl₃) 29.0 (C-7), 30.2 (C-6/C-10), 32.4 (C-3/C-5),

44.7 (C-2/C-9), 46.5 (C-8), 56.5 (−OCH₃), 82.4 (C-4), (C-1 not

obsd); m/z (EI) 182 ([M + H]⁺, 6), 181 (M⁺, 54), 124 (48), 109 (10),

95 (11), 94 (100); found M⁺181.1471, C₁₁H₁₉NO requires 181.1467.

The preparation of 4-azi-1-fluoroadamantane (1h) has been

discussed thoroughly in the Supporting Information of a recent

publication.^24bOnly the products from (1h) are described here.

1357

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

Preparation and Reactions of 4-Azi-1-fluoroadamantane

(1h):^48e3H-Diazirine 1h was synthesized from ketone 4h. The crude

product was chromatographed (silica gel 60, 230−400 mesh) using

pentane as eluant and a readily crystallizing white solid (mp 60 °C

(sublimes)) was obtained in 44% yield. The pyrolytic intramolecular

insertion reactions of 1h afforded s-7h and as-7h in 43% in yield (ratio

=92:8); separation of these products was not possible. The s-7h/as-7h

mixture was analyzed using 2-D NMR.

C-6/C-10), 34.7 (J_CF9.9; C-3/C-5), 36.7 (J_CF18; C-2/C-9), 42.6 (J_CF

17.5; C-8), 55.3 (−OCH₃), 80.8 (J_CF2.3 C-4), 91.7 (J_CF184 C-1);

ν_max/cm⁻¹(CCl₄) 2922, 2860, 1451, 1381, 1355, 1193, 1106, 1072,

̅

992, 924, 909 ; m/z (EI) 184 (M⁺, 15), 153 (13), 152 (100), 137 (10),

123 (21), 111 (13), 110 (22), 109 (15), 97 (24); found M⁺184.1260,

C₁₁H₁₇OF requires 184.1262. Anal. Calcd for C₁₆H₂: C, 71.71; H,

9.31. Found: C, 71.86; H, 9.60.

anti-1-Fluoro-4-methoxyadamantane (anti-9h): δ_H/ppm

(600.1 MHz, CDCl₃) 1.33 (2 H, d J 12.7), 1.84−1.88 (6 H, m),

1.93 (2 H, m), 2.13 (1 H, s), 2.23 (2 H, s), 3.31 (3 H, s) 3.32 (1 H, s);

δ_C/ppm (150.9 MHz, CDCl₃) 29.8 (J_CF2 ; C-6/C-10), 30.6 (J_CF9.7;

C-7), 33.6 (J_CF10.4; C-3/C-5), 40.8 (J_CF18.4; C-2/C-9), 42.7 (J_CF

16.6;C-8), 55.8 (−OCH₃), 81.5 (J_CF= 1.6 ; C-4), 91.9 (J_CF183.6;

C-1); m/z (EI)184 (M⁺, 20), 153 (15), 152 (100), 110 (27), 109 (16),

97 (25), 92 (20), 91 (17).

It was possible to determine the structure of the protic rearran-

gement product rearr-9h by 2-D NMR (see ref 29 in Supporting

Information) because it was enriched to 12% by chromatography.

7-Fluorotricyclo[3.3.1.1^2,7]decan-3-yl methyl ether (rearr-9h)

(Table 15): m/z (EI) 184 (26), 153 (14), 152 (100), 137 (10), 123

(19), 111 (14), 110 (20), 109 (22), 98 (11), 97 (22), 93 (13), 92 (17),

91 (18), 84 (13), 79 (22).

2,4-Didehydro-7-fluoroadamantane (s-7h): δ_H/ppm (600.13

MHz, CDCl₃) 1.43 (3 H, m), 1.50 (2 H, d J 10.3), 1.61−1.66 (3 H,

m), 1.97−1.99 (2 H, m), 2.07 (1 H, s), 2.59 (2 H, s); δ_C/ppm (150.9

MHz, CDCl₃) 23.4 (J_CF1.6; C-2/C-4), 28.5 (J_CF11.3; C-3), 34.8 (J_CF

19.8; C-10), 36.6 (J_CF9.3; C-1/C-5), 39.1 (J_CF17.6; C-6/C-8), 51.2

(J_CF2.2; C-9), 91.4 (J_CF180.3; C-7); ν_max/cm⁻¹(CCl₄) 3034, 2942,

̅

2864, 1458, 1448, 1352, 1333, 1247, 1086, 1054, 975, 926, 891; m/z

(EI) 152 (M⁺, 78), 137 (14), 117 (10), 111 (35), 110 (55), 109 (49),

98 (27), 97 (100), 93 (15), 92 (53), 91 (60), 80 (11), 79 (26), 77

(28), found M⁺152.0998, C₁₀H₁₃F requires 152.1001.

2,4-Didehydro-1-fluoroadamantane (as-7h): Obtained from a

concentrated solution containing 8% of as-7h: δ_H/ppm (600.1 MHz,

CDCl₃) 1.22 (1 H, s), 1.25 (1 H, s), 1.32 (1 H, s), 1.57−1.58 (2 H,

m), 1.6 (1 H, br s), 1.75−1.80 (3 H, m), 1.86 (1 H, m), 2.14 (1 H, s),

2.30 (1 H, s), 2.57 (1 H, s); δ_C/ppm (150.9 MHz, CDCl₃) 19.7 (C-3),

22.3 (J_CF1.2; C-4), 26.4 (J_CF31; C-2), 28.0 (J_CF1.4; C-7), 28.6 (J_CF

11; C-10), 30.3 (J_CF9.1; C-5), 33.7 (J_CF2.7; C-6), 36.5 (J_CF18.7; C-

8), 55.6 (J_CF17.7; C-9), 92.3 (J_CF180; C-1); m/z (EI) 152 (M⁺, 89),

137 (18), 117 (12), 111 (17), 110 (86), 109 (73), 98 (23), 97 (100),

93 (27), 92 (83), 91 (68), 80 (15), 79 (72), 78 (15), 77 (36).

syn-4-Cyclohexyl-1-fluoroadamantane (syn-8h): The insertion

reaction of 5-fluoroadamantan-2-ylidene into cyclohexane produced a

crude mixture that consisted of mostly syn-8h/anti-8h (76%, ratio

=92:8) with lesser amounts of azines 3h (11%), ketone 4h (4%), and

unreacted starting material (7%). It was chromatographed (silica gel

60, 230−400 mesh) using pentane as eluant and the readily

crystallizing solid syn-8h (mp 54−56 °C) was isolated in 46% yield:

δ_H/ppm (600.1 MHz, CDCl₃) 0.73 (2 H, m), 1.00−1.20 (4 H, m),

1.30−1.40 (1 H, m), 1.54 (2 H, d J 14.9), 1.55−1.72 (7 H, m), 1.80−

1.83 (4 H, m), 1.98 (2 H, s), 2.19 (1 H, s), 2.26 (2 H, s); δ_C/ppm

(150.9 MHz, CDCl₃) 26.5 (C-3′/C-5′), 26.6 (C-4′), 30.6 (C-2′/C-6′

C), 31.5 (J_CF9.8; C-7), 32.1 (J_CF9.1; C-3/C-5), 36.5 (C-1′), 37.4 (J_CF

15.4; C-2/C-9), 37.5 (J_CF3.8; C-6/C-10), 43.4 (J_CF16.9; C-8), 48.2

Table 15. Chemical Shifts of 7-Fluorotricyclo[3.3.1.1^2,7]decan-

3-yl Methyl Ether (rearr-9h) Extracted from 2-D NMR

rearr-9h

δ_C(ppm)

H₁

H₂

J_CF(Hz)

1

91.6

43.8

45.6

37.7

78.6

33.9

30.9

26.9

33.8

31.1

55.3

179.4

19.5

16.2

19.8

9/10*

2.03

2.11

1.984

3.48

1.92

2.14

2.265

2.615

3.27

1.735

1.52

2

1.434

4

5

1.84

1.2

7

2

6

11.9

9.7

7.5

8

3

OCH₃

5-Fluoroadamantan-2-one azine^24b,74(3h): 3H-Diazirine 1h

(20 mg, 0.11 mmol) was photolyzed in a cooled, rotating flask

(T = 5 °C, 2 h). For GC and GC−MS analyses, the crude product was

dissolved in CHCl₃. For NMR analyses, the crude product was washed

with hot ligroin and dried in vacuo: δ_H/ppm (400.1 MHz, CDCl₃)

1.68 (2 H, s), 1.70−1.85 (6 H, m), 1.91−1.97 (4 H, m), 1.98−2.05 (8

H, m), 2.36 (2 H, s), 2.83 (2 H, s), 3.53 (2 H, s); δ_C/ppm (100.6

MHz, CDCl₃) 32.9 (J_CF9.6; C-3); 31.0 (J_CF9.2; C-7), 36.7, 37.9

(C-8/C-10), 41.0 (J_CF9.9; C-1), 41.7 (J_CF= 19.7; C-6), 41.9, 42.9 (J_CF

20; C-4/C-9); 91.1, 91.2 (J_CF185.5; C-5), 169.4 (C-2); λ_max/nm

(C₆H₁₄) 211 (ε/(10⁴M⁻¹cm⁻¹) 25), 232 (ε/(10⁴M⁻¹cm⁻¹) 7.7).

(J_CF1.7 Hz; C-4), 97.8 (J_CF183.4 Hz; C-1); ν_max/cm⁻¹(CCl₄) 2932,

̅

2852, 1452, 1355, 1304, 1184, 1110, 1085, 1048, 1030, 946, 908; m/z

(EI) 237 ([M + H]⁺, 17), 236 (M⁺, 100), 153 (57), 152 (11), 135

(19), 134 (30), 133 (18), 97 (98), 93 (12), 91 (24), 88 (21); found

M⁺236.1946, C₁₆H₂₅F requires 236.1940. Anal. Calcd for C₁₆H₂₅F: C,

81.30; H, 10.66. Found: C, 81.17; H, 10.90.

anti-4-Cyclohexyl-1-fluoroadamantane (anti-8h): The syn-8h/

anti-8h mixture could be enriched in epimer anti-8h, which was

characterized: δ_H/ppm (600.13 MHz, CDCl₃) 0.73 (2 H, mc), 1.11−

1.18, (4 H, m), 1.32 (2 H, d J 12.7), 1.39−1.41 (1 H, m), 1.62−1.72 (5

H, m), 1.78−1.84 (8 H, m), 2.14 (1 H, s), 2.20 (2 H, s); δ_C/ppm

(150.9 MHz, CDCl₃) 26.5 (C-4′), 26.6 (C-3′/ C-5′), 30.3 (J_CF2.3; C-

6/C-10), 30.9 (C-2′/C-6′), 31.2 (J_CF9.2; C-7), 31.5 (J_CF9.7; C-3/C-

5), 36.4 (J_CF1.9; C-1′), 43.2 (J_CF16.9; C-8), 43.4 (J_CF17.5; C-2/C-9),

48.4 (J_CF1.6; C-4), 92.3 (J_CF186; C-1); m/z (EI) 237 ([M + H]⁺, 16),

236 (M⁺, 91), 154 (12), 153 (100), 152 (16), 135 (32), 134 (48), 133

(28), 97 (18), 93 (18), 92 (16), 91 (43), 83 (45), 82 (85), 80 (23).

1-Fluoro-4-methoxyadamantane (9h): The insertion reaction

of 5-fluoroadamantan-2-ylidene into MeOH produced a crude mixture

that consisted of mostly syn-9h/anti-9h (85%, ratio = 84:16) with

lesser amounts of intramolecular insertion products (5%) and a

rearrangement product (2%) that was further investigated from an

enriched mixture obtained after chromatography (vide infra). It was

chromatographed (silica gel 60, 230−400 mesh) using pentane/

CH₂Cl₂(9:1) as eluant and syn-9h was isolated in 50% overall yield.

syn-1-Fluoro-4-methoxyadamantane (syn-9h): δ_H/ppm (600.1

MHz, CDCl₃) 1.52 (2 H, d J 13.8), 1.59 (2 H, s), 1.68 (2 H, d J 12),

1.82 (2 H, s), 2.13−2.14 (3 H, m), 2.33 (2 H, s), 3.14 (1 H, s), 3.30 (3

H, s); δ_C/ppm (150.9 MHz, CDCl₃) 30.6 (J_CF9.5;C-7,), 34.6 (J_CF2.2;

ASSOCIATED CONTENT

■

S

* Supporting Information

NMR, IR, UV spectra, GC−MS data of new compounds, and

computational data. This material is available free of charge via

the Internet at http://pubs.acs.org.

AUTHOR INFORMATION

■

Corresponding Author

*Phone: +43-1-4277-52121. Fax: + 43-1-4277-52140. E-mail:

udo.brinker@univie.ac.at.

ACKNOWLEDGMENTS

■

We thank Mag. P. Billing for discussions, Dozent Dr. J.-L.

Mieusset for his valuable help and support, and Dr. G. Wagner

1358

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

J. Chem. 1998, 76, 851. Level of theory: UB3LYP/6-311+

+G(3df,2pd)//UB3LYP/6-31+G(d)+ 0.98 × ΔZPVE.

of the Institute of Organic Chemistry of the University of

Vienna for the preparation of compounds 1f and 1g.

(23) (a) Isaev, S. D.; Yurchenko, A. G.; Stepanov, F. N.; Kolyada, G.

G.; Novikov, S. S.; Karpenko, N. F. Zh. Org. Khim. 1973, 9, 724.

(b) Isaev, S. D.; Kulik, N. I.; Yurchenko, A. G. Theor. Exp. Chem.

(Transl. Teor. Eksp. Khim.) 1993, 29, 80. (c) Kadentsev, V. I.; Chizhov,

O. S.; Kolotyrkina, N. G.; Isaev, S. D.; Kozlov, O. F.; Yurchenko, A. G.

Izv. Akad. Nauk., Ser. Khim. 1987, 9, 2010. (d) Yurchenko, A. G.; Isaev,

S. D.; Novoselov, E. F. Zh. Org. Khim. 1984, 20, 222.

(24) (a) Knoll, W.; Bobek, M. M.; Giester, G.; Brinker, U. H.

Tetrahedron Lett. 2001, 42, 9161. (b) Wagner, G.; Knoll, W.; Bobek,

M. M.; Brecker, L.; van Herwijnen, H. W. G.; Brinker, U. H. Org. Lett.

2010, 12, 332.

REFERENCES

■

(1) Carbene Rearrangements 84. For part 83, see: Gupta, S.;

Choudhury, R.; Krois, D.; Wagner, G.; Brinker, U. H.; Ramamurthy, V.

Org. Lett. 2011, 13, 6074.

(2) See the special issue devoted entirely to this topic:

Diastereoselection; le Noble, W. J.; Gung, B. W., Eds. Chem. Rev.

1999, 99, 1067−1480 and references therein.

(3) Johnson, C. R.; Tait, B. D.; Cieplak, A. S. J. Am. Chem. Soc. 1987,

109, 5875.

(25) Adcock, W.; Cotton, J.; Trout, N. A. J. Org. Chem. 1994, 59,

1867.

(4) (a) Filippi, A.; Trout, N. A.; Brunelle, P.; Adcock, W.; Sorensen,

T. S.; Speranza, M. J. Org. Chem. 2004, 69, 5537. (b) Adcock, W.;

Trout, N. A. J. Org. Chem. 1991, 56, 3229. (c) Laube, T.; Stilz, H. U.

J. Am. Chem. Soc. 1987, 109, 5876.

(5) (a) Kaselj, M.; Chung, W.-S.; le Noble, W. J. Chem. Rev. 1999, 99,

1387. (b) According to the IUPAC, steric effect refers to “the effect

on a chemical or physical property (structure, rate or equilibrium

constant) upon introduction of substituents having different steric

requirements. Steric effects arise from contributions ascribed to strain

as the sum of (1) non-bonded repulsions, (2) bond angle strain, and

(3) bond stretches or compressions. A “secondary steric effect”

involves the differential moderation of electron delocalization by non-

bonded compressions.

(6) Stereoelectronic effects arise from the different alignment of

electronic orbitals in different arrangements of nuclear geometry.

Moss, G. P. Pure Appl. Chem. 1996, 68, 2193 (basic terminology of

stereochemistry (IUPAC Recommendations 1996), p 2218).

(7) Cieplak, A. S.; Tait, B. D.; Johnson, C. R. J. Am. Chem. Soc. 1989,

111, 8447.

(8) (a) Anh, N. T. Top. Curr. Chem. 1980, 88, 145. (b) Anh, N. T.;

Eisenstein, O. Nouv. J. Chim. 1977, 1, 61. (c) Anh, N. T.; Eisenstein,

O. Tetrahedron Lett. 1976, 3, 155. (d) Cieplak, A. S. J. Am. Chem. Soc.

1981, 103, 4540.

(26) (a) Burgess, K.; Van der Donk, W. A.; Jarstfer, M. B.; Ohlmeyer,

M. J. J. Am. Chem. Soc. 1991, 113, 6139. Adcock (see ref 4b) re-

evaluated the results of Xie and le Noble, who found a ratio of 45:55 in

favor of the syn-product: (b) Xie, M.; le Noble, W. J. J. Org. Chem.

1989, 54, 3836.

(27) (a) Dannenberg, J. J. Chem. Rev. 1999, 99, 1225. (b) Fort, R. C.

Jr.; Schleyer, P. v. R. Chem. Rev. 1964, 64, 277. (c) Bartmess, J. E.;

McIver, R. J. In Gas Phase Ion Chemistry; Bowers, M. T., Ed.;

Academic: New York, 1979; Vol. 2, p 101.

(28) (a) Thamattoor, D. M.; Jones, M. Jr.; Pan, W; Shevlin, P. B.

Tetrahedron Lett. 1996, 37, 8333. (b) Buron, C.; Tippmann, E. M.;

Platz, M. S. J. Phys. Chem. A 2004, 108, 1033. (c) Lemal, D. M.;

Gosselink, E. P.; McGregor, S. D. J. Am. Chem. Soc. 1966, 88, 582.

(d) Hartwig, J. F.; Jones, M. Jr.; Moss, R. A.; Lawrynowicz, W.

Tetrahedron Lett. 1986, 27, 5907. (e) Doering, W. v. E.; Hoffmann,

A. K. J. Am. Chem. Soc. 1954, 76, 6162. (f) Hine, J. J. Am. Chem. Soc.

1950, 72, 2438. (g) Hine, J.; Dowell, A. M. Jr. J. Am. Chem. Soc. 1954,

76, 2688. (h) Brinker, U. H.; Ritzer, J. J. Am. Chem. Soc. 1981, 103,

2116. (i) Brinker, U. H.; Miebach, T. J. Org. Chem. 1999, 64, 8000.

(j) Nordvik, T.; Mieusset, J.-L.; Brinker, U. H. Org. Lett. 2004, 6, 715.

(29) (a) Schmitz, E. In Chemistry of Diazirines; Liu, M. T. H., Ed.;

CRC: Boca Raton, 1987; Vol. 1, p 57. (b) Schmitz, E.; Ohme, R.

Chem. Ber. 1962, 95, 795. (c) Schmitz, E.; Ohme, R. Chem. Ber. 1961,

94, 2166.

(9) Anh, N. T. Frontier Orbitals: A Practical Manual; Wiley:

Chichester, 2007.

(10) Yadav, V. K.; Gupta, A.; Balamurugan, R.; Sriramurthy, V.;

Kumar, N. V. J. Org. Chem. 2006, 71, 4178.

(30) Moss, R. A. Acc. Chem. Res. 2006, 39, 267.

(31) (a) Maeda, Y.; Matsunaga, Y.; Wakahara, T.; Takahashi, S.;

Tsuchiya, T.; Ishitsuka, M. O.; Hasegawa, T.; Akasaka, T.; Liu,

M. T. H.; Kokura, K.; Horn, E.; Yoza, K.; Kato, T.; Okubo, S.;

Kobayashi, K.; Nagase, S.; Yamamoto, K. J. Am. Chem. Soc. 2004, 126,

6858. (b) Niino, Y.; Wakahara, T.; Akasaka, T; Liu, M. T. H. ITE Lett.

Batteries, New Technol. Med. 2002, 3, 82. (c) Akasaka, T.; Liu,

M. T. H.; Niino, Y.; Maeda, Y.; Wakahara, T.; Okamura, M.;

Kobayashi, K.; Nagase, S. J. Am. Chem. Soc. 2000, 122, 7134.

(32) (a) Shustov, G. V.; Liu, M. T. H.; Houk, K. N. Can. J. Chem.

1999, 77, 540. (b) Bonneau, R.; Liu, M. T. H. J. Phys. Chem. A 2000,

104, 4115. (c) Liu, M. T. H.; Ramakrishnan, K. Tetrahedron Lett. 1977,

3139. (d) Note: the substituted adamantane-derived azines consist of

two isomers each.

(11) Kaneno, D.; Tomoda, S. Tetrahedron Lett. 2004, 45, 4559.

(12) Kaneno, D.; Tomoda, S. Org. Lett. 2003, 5, 2947.

(13) (a) Tomoda, S. Chem. Rev. 1999, 99, 1243. (b) Tomoda, S.;

Senju, T. Tetrahedron 1999, 55, 5303.

(14) (a) Klopman, G.; Hudson, R. F. Theor. Chim. Acta 1967, 8, 165.

(b) Klopman, G. J. Am. Chem. Soc. 1968, 90, 223. (c) Salem, L. J. Am.

Chem. Soc. 1968, 90, 543. (d) Fleming, I. Frontier Orbitals and Organic

Chemical Reactions; Wiley: London, 1976.

(15) (a) Suzuki, Y.; Kaneno, D.; Tomoda, S. J. Phys. Chem. A 2009,

113, 2578. (b) Cheung, C. K.; Tseng, L. T.; Lin, M.-H.; Srivastava, S.;

le Noble, W. J. J. Am. Chem. Soc. 1986, 108, 1598.

(16) Alabugin, I. V.; Gilmore, K. M.; Peterson, P. W. WIREs Comput.

Mol. Sci. 2011, 1, 109−141.

(33) Liu, M. T. H.; Choe, Y.-K.; Kimura, M.; Kobayashi, K.; Nagase,

S.; Wakahara, T.; Niino, Y.; Ishitsuka, M. O.; Maeda, Y.; Akasaka, T.

J. Org. Chem. 2003, 68, 7471.

(17) (a) Adcock, W.; Head, N. J.; Lokan, N. R.; Trout, N. A. J. Org.

Chem. 1997, 62, 6177. (b) Adcock, W.; Trout, N. A. Chem. Rev. 1999,

99, 1415. (c) Adcock, W.; Trout, N. A. J. Phys. Org. Chem. 2007, 20,

791.

(34) Maier, W. F.; Schleyer, P. v. R. J. Am. Chem. Soc. 1981, 103,

1891.

(18) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,

(35) Bally, T.; Matzinger, S.; Truttmann, L.; Platz, M. S.; Morgan, S.

Angew. Chem., Int. Ed. Engl. 1994, 33, 1964.

83, 735.

(19) (a) Alabugin, I. V.; Manoharan, M. J. Org. Chem. 2004, 69, 9011.

(b) Duddeck, H. Tetrahedron 1978, 34, 247. (c) Muller, P. Pure Appl.

Chem. 1994, 66, 1077 (glossary of terms used in physical organic

chemistry (IUPAC Recommendations 1994) p 1123).

(36) (a) Bonneau, R.; Hellrung, B.; Liu, M. T. H.; Wirz, J.

J. Photochem. Photobiol., A 1998, 116, 9. A singlet GS is predicted for

adamantanylidene, with an S−T energy gap of 2.82 kcal/mol at the

1

RHF/6-31G(d) and B3LYP/6-31+G(d) levels for 5b and UHF/6-

̌ ̌ ́

(20) Hirsl-Starcevic, S.; Majerski, Z. J. Org. Chem. 1982, 47, 2520.

(21) (a) Bobek, M. M.; Brinker, U. H. J. Am. Chem. Soc. 2000, 122,

7430. (b) Knoll, W.; Bobek, M. M.; Kalchhauser, H.; Rosenberg, M. G.;

Brinker, U. H. Org. Lett. 2003, 5, 2943.

31G(d), ROHF/6-31G(d) and UB3LYP/6-31G(d) levels for ³5b.

Geometries were optimized, and vibrational frequencies were

determined at the DFT (B3LYP/6-31+G(d)) level to make the

zero-point corrections using a scale factor of 0.95. (b) Bonneau, R.;

Liu, M. T. H. J. Am. Chem. Soc. 1996, 118, 7229.

(22) (a) Udding, A. C.; Strating, J.; Wynberg, H.; Schlatmann, J. L.

M. A. Chem. Commun. 1966, 657. (b) Shustov, G.; Liu, M. T. H. Can.

1359

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

The Journal of Organic Chemistry

Article

(37) Here, the term antiperiplanar always refers to a hypothetical

intermolecular “σ_inc” bond.

(60) Initially, a more stable complex is formed. The following TS is

still more stable than the energy of the reactants combined.

(61) Except for carbene 5h, NBO analysis of each MeO···H···5 TS

attributed the H atom to the C atoms of the carbene 5 moieties.

According to gas-phase calculations, the H atom of the inserted O−H

bond is protic and the interaction with the carbene 5 lp is dominant.

No Felkin−Anh-type interaction is found whenever the NBO analysis

attributes the H atom from MeOH onto the adamantanylidene moiety,

which confers a partial positive charge onto the formerly divalent C atom.

However, the usefulness of gas-phase computations for TS(5 → 9) is

questionable, because they overemphasize concerted mechanism 2.

Therefore, further calculations including solvent effects are welcome.

(62) Solution dynamics are expected to be more complex.

(63) Kaneno, D. Ph.D. Dissertation, The University of Tokyo, Japan,

2001.

(38) The vacant carbene p-orbital is called lp* in the output file of

NBO analyses. It is relatedbut not identical withthe LUMO of 5.

(39) (a) Carey, F. A.; Sundberg, R. J. In Advanced Organic Chemistry,

Part A, Structure and Mechanisms, 3rd ed.; Plenum Press: New York,

1990; pp 215−216. (b) IUPAC, Compendium of Chemical

Terminology, 2nd ed. (the “Gold Book”) 1997. Online corrected

version: 1994, “Curtin−Hammett principle”.

(40) The FMO at the TS shows the influence of hyperconjugation on

the calculated orbitals.

(41) Jones, W. M.; Brinker, U. H. In Pericyclic Reactions; Marchand,

A. P., Lehr, R. E., Eds.; Academic: New York, 1977; Vol. 1, pp 109−

198.

(42) (a) Doering, W. v. E.; Prinzbach, H. Tetrahedron 1959, 6, 24.

(b) Ramalingam, M.; Ramasami, K.; Venuvanalingam, P.; Sethuraman, V.

J. Mol. Struct. (Theochem.) 2005, 755, 169.

(64) (a) When ΔE_cat.and ΔE_carb. values (differences of the

conformer energies) are plotted instead of electronegativities, the same

substituent is not at the same point of the x-axis. (b) Probably also in

case of the corresponding cations 6.

(65) Gaussian 03: Frisch, M. J. et al. Gaussian, Inc., Pittsburgh, PA,

2003.

(43) Bach, R. D.; Su, M.-D.; Aldabbagh, E.; Andres

J. Am. Chem. Soc. 1993, 115, 10237.

́

, J. L.; Schlegel, H. B.

(44) (a) Arduengo, A. J. III; Calabrese, J. C.; Davidson, F.; Dias, H.

V. R.; Goerlich, J. R.; Krafczyk, R.; Marshall, W. J.; Tamm, M.;

Schmutzler, R. Helv. Chim. Acta 1999, 82, 2348. (b) Mieusset, J.-L.;

Brinker, U. H. J. Org. Chem. 2007, 72, 10211.

(66) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.;

Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.

(45) The calculated TS for 5b···CyH is ca. 10⁻²hartree (∼6 kcal/mol)

above their combined energies.

(67) (a) Mendez, F.; Garcia-Garibay, M. A. J. Org. Chem. 1999, 64,

7061. (b) Schreiner, P. R. Angew. Chem., Int. Ed. 2007, 46, 4217.

(68) (a) Kira, M.; Akiyama, M.; Ichinose, M.; Sakurai, H. J. Am.

Chem. Soc. 1989, 111, 8256. (b) Schleyer, P. v. R.; Nicholas, R. D.

J. Am. Chem. Soc. 1961, 83, 182. (c) Cuddy, B. D.; Grant, D.; Karim,

A.; McKervey, M. A.; Rea, E. J. F. J. Chem. Soc., Perkin Trans. I 1972,

2701. (d) Geluk, H. W.; Schlatmann, J. L. M. A. Recl. Trav. Chim. Pays-

Bas 1969, 88, 13. (e) Bone, J. A.; Whiting, M. C. J. Chem. Soc., Chem.

Comm. 1970, 115. (f) Molle, G.; Dubois, J. E.; Bauer, P. Can. J. Chem.

1987, 65, 2428. (g) Loomes, D. J.; Robinson, M. J. T. Tetrahedron

1977, 33, 1149.

(69) (a) Bobek, M. M. Ph.D. Dissertation, University of Vienna,

Austria, 2000. (b) van Herwijnen, H. W. G. Ph.D. Dissertation,

University of Vienna, Austria, 2002.

(70) Calibrated to solvent residual peak at δ_H= 7.28 ppm.

(71) Calibrated to solvent residual peak at δ_C= 127.77 ppm.

(72) For ketone 4c, see refs 68b and 68c. Ketone 4c was synthesized

either from alcohol 10 (12% yield) or from adamantane 12 (7% yield)

and was contaminated with up to 3% of 4b and dimethyladamanta-

none. Hence, these byproducts were carried along during the synthesis

and reactions of 3H-diazirine 1c.

(46) According to NBO analysis, the equatorial H atoms are slightly

more positive and the axial C−H bonds are more stretched. This is

reflected in the electron population of the C−H bonds σ(C−H)_eq

σ(C−H)_axand σ*(C−H)_eq< σ*(C−H)_ax.

>

(47) (a) Pezacki, J. P.; Wood, P. D.; Gadosy, T. A.; Lusztyk, J.;

Warkentin, J. J. Am. Chem. Soc. 1998, 120, 8681. (b) Pezacki, J. P.;

Shukla, D.; Lusztyk, J.; Warkentin, J. J. Am. Chem. Soc. 1999, 121,

6589. (c) Mieusset, J.-L.; Bespokoev, A.; Pacar, M.; Abraham, M.;

Arion, V. B.; Brinker, U. H. J. Org. Chem. 2008, 73, 6551.

(48) (a) Kirmse, W. In Advances in Carbene Chemistry; Brinker, U. H.,

Ed.; JAI Press: Greenwich, 1994, Vol. 1, pp 1−57. (b) Kirmse, W.;

Meinert, T. J. Chem. Soc., Chem. Commun. 1994, 1065. (c) Kirmse, W.;

Meinert, T.; Modarelli, D. A.; Platz, M. S. J. Am. Chem. Soc. 1993, 115,

8918. (d) Holm, K. H.; Skattebøl, L. J. Am. Chem. Soc. 1977, 99, 5480.

(e) Knoll, W.; Mieusset, J.-L.; Arion, V. B.; Brecker, L.; Brinker, U. H.

Org. Lett. 2010, 12, 2366.

(49) Intramolecular insertion product 7 should be formed by a protic

mechanism, because its formation is a thermal reaction with a high

activation barrier.

(50) Structures determined using 2-D NMR; cf. Lenoir, D.; Hall, R. E.;

Schleyer, P. v. R. J. Am. Chem. Soc. 1974, 96, 2138.

(51) (a) Steenken, S. Pure Appl. Chem. 1998, 70, 2031. (b) Bethell,

D.; Newall, A. R.; Stevens, G.; Whittaker, D. J. Chem. Soc., B 1969, 749.

(c) Bethell, D.; Newall, A. R.; Stevens, G.; Whittaker, D. J. Chem. Soc.,

B 1971, 23.

(73) Bobek, M. M.; Brinker, U. H. Synth. Commun. 1999, 29, 3221.

(74) (a) Nelsen, S. F.; Klein, S. J.; Trieber, D. A. II; Ismagilov, R. F.;

Powell, D. R. J. Org. Chem. 1997, 62, 6539. (b) UV(C₆H₁₄): λ_max(ε ×

10⁴): 211 (25), 232 (7.7); melting was not observed up to 260 °C;

compound 3h seems to melt and to sublime at this temperature

simultaneously.

(52) A radical recombination mechanism (not shown) may be

dismissed for singlet reactions of carbene 5.

(53) Thus, one could portray such carbenes as CR₂.

(54) Pliego, J. R. Jr.; De Almeida, W. B. J. Phys. Chem. A 1999, 103,

3904.

(55) (a) Mieusset, J.-L.; Brinker, U. H. J. Am. Chem. Soc. 2006, 128,

15843. (b) Mieusset, J.-L.; Brinker, U. H. J. Org. Chem. 2006, 71, 6975.

(56) (a) Bordwell, F. G.; Satish, A. V. J. Am. Chem. Soc. 1991, 113,

985. (b) Washabaugh, M. W.; Jencks, W. P. Biochemistry 1988, 27,

5044.

(57) (a) Liu, M. T. H.; Subramanian, R. J. Chem. Soc., Perkin Trans. 2

1986, 1233. (b) Kirmse, W.; Lelgemann, R.; Friedrich, K. Chem. Ber.

1991, 124, 1853. (c) Warner, P. M.; Chu, I.-S. J. Am. Chem. Soc. 1984,

106, 5366. (d) Turro, N. J.; Cha, Y.; Gould, I. R. Tetrahedron Lett.

1985, 5951.

(58) Mieusset, J.-L.; Brinker, U. H. Eur. J. Org. Chem. 2008, 3363.

(59) Pezacki, J. P.; Warkentin, J.; Wood, P. D.; Lusztyk, J.; Yuzawa,

T.; Gudmundsdottir, A. D.; Morgan, S.; Platz, M. S. J. Photochem.

Photobiol., A 1998, 116, 1.

1360

dx.doi.org/10.1021/jo202132c | J. Org. Chem. 2012, 77, 1340−1360

Article Doi

Intra- and intermolecular reaction selectivities of γ-substituted adamantanylidenes

DOI: 10.1021/jo202132c

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo202132c

Full text of DOI:10.1021/jo202132c

Products guided by the article

R&D Labs maybe for 1353745-32-0

Relevant to this article

Hot Product