Intramolecular nucleophilic capture of radical cations by tethered hydroxy functions

Tetrahedron 62 (2006) 6471–6489

Intramolecular nucleophilic capture of radical cations by

tethered hydroxy functions

*

Heinz D. Roth, Torsten Herbertz, Ronald R. Sauers and Hengxin Weng

*

Department of Chemistry and Chemical Biology, Rutgers University, Wright-Rieman Laboratories,

New Brunswick, NJ 08854-8087, USA

Received 28 September 2005; accepted 6 December 2005

Available online 24 April 2006

Abstract—A range of systems bearing hydroxy functions tethered to the molecular framework gives rise to a family of interesting radical

cations, 5^ꢀ⁺–11^ꢀ⁺, upon electron transfer to photo-excited cyanoaromatics. Geraniol (5), nerol (6), citronellol (7), chrysanthemol (8), homo-

chrysanthemol (9), trans-1-o-hydroxyphenyl-2-phenylcyclopropane (10), and endo-5-hydroxymethylnorbornene (11), generate a series of

mono-, bi-, or tricyclic ethers via a series of four- to seven-membered transition states. Two of the radical cations, 5^ꢀ⁺and 6^ꢀ⁺, undergo tandem

cyclizations where 1,5- and/or 1,6-C–C cyclizations precede nucleophilic capture.

1. Introduction

The various factors governing regio- and stereochemistry of

nucleophilic substitution on radical cations have been evalu-

ated in a series of substrates that offer regiochemical compe-

tition within the target molecule. Not surprisingly, molecular

orbital (MO) factors, particularly the nature of the singly oc-

cupied molecular orbital (SOMO) and the lowest unoccupied

molecular orbital (LUMO), are of major importance. If MO

factors allow more than one reaction, the thermodynamic

stability of the resulting free radicals determines the out-

come.^7b,c,15Steric factors typically do not play a major

role; in fact, several radical cations are captured by attack

on highly congested centers.^7,13However, at least one radical

cation is captured selectively at the less encumbered site.¹⁵

We illustrate the major factors in the following examples,

which were chosen to differentiate unambiguously between

potentially governing factors.

The structures of organic radical cations and their reactions

have attracted much attention during recent decades and con-

tinue to be the focus of intense interest.^1–13Because of their

dual nature, radical cations contain an unpaired spin and a

positive charge, they may undergo a wide range of diverse re-

actions, including rearrangements^3–6and sigmatropic shifts,⁷

nucleophilic substitution or ‘capture’,^8,9cycloadditions,^2d,10

as well as fragmentations¹¹and cycloreversions.¹²Compared

to the analogous reactions of the parent molecules, many rad-

ical cation reactions show a dramatic decrease in activation

barriers,^11,13one of the most striking aspects of radical cation

chemistry.³Among the plethora of reactions the capture by

nucleophiles has been investigated in detail.

As a result of their electron deﬁcient nature radical cations

are strong acids and are excellent targets for nucleophilic

substitution or capture. Awide range of strained ring radical

cations undergo nucleophilic substitution, i.e., backside at-

tack with inversion of conﬁguration.^8,9The stereochemical

requirements for this reaction are considered rigorous;

molecular orbital calculations indicate that a trajectory

close to 180^ꢁis critical.¹⁴Many of these substitutions occur

at sterically encumbered carbons;^7,13apparently, relief of

ring strain signiﬁcantly increases the driving force; this

increase more than compensates for the steric hindrance.

A simple system illustrating the role of molecular orbital

effects is the radical cation of trans-1,2-dimethylcyclo-

propane. The unpaired spin density, SOMO and LUMO of

this species are shown in Figure 1; both HOMO and

LUMO have a node at C-3. Accordingly, this radical cation

and likewise other 1,2-disubstituted cyclopropane radical

cations are unreactive at C-3.^3b,16

The radical cation of sabinene, 1^ꢀ⁺, contains a rigidly ar-

ranged vinylcyclopropane system; unpaired electron spin

and charge are delocalized over four atoms, C-5, C-1, C-4,

and C-4⁰; however, only two centers, C-1 and C-4⁰have sig-

niﬁcant orbital coefﬁcients in both HOMOand LUMO. Thus,

nucleophiles should be able to attack 1^ꢀ⁺either at the exocy-

clic alkene position or at the cyclopropane ring. In fact, nu-

cleophilic attack occurs exclusively at the highly congested

Keywords: Electron transfer; Radical cation; Photochemistry; Intramolecu-

lar nucleophilic substitution; Tandem cyclization.

* Corresponding authors. Tel.: +1 732 445 5664; fax: +1 732 445 5312

(H.D.R.); tel.: +1 732 445 2626 (R.R.S.); e-mail addresses: roth@

rutchem.rutgers.edu ; sauers@rutchem.rutgers.edu

doi:10.1016/j.tet.2006.03.064

6472

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

.

OCH₃

.

a

OCH₃

2-B-OCH₃

b

.

c

2-A-OCH₃

.

+

[∆H = –18.7 kcal/mol]

[∆H = –13.9 kcal/mol]

c

d

b

H₃CO

d

.

2 ⁺

.

H CO

.

3

2-C-OCH₃

2-D-OCH₃

[∆H = 0.0 kcal/mol]

[∆H = 0.4 kcal/mol]

H₃CO

Ar

Ar = C₆H₄-CN

H₃CO

Scheme 1. Potential regioisomeric free radicals generated by nucleophilic

addition/capture of tricyclo[4.3.1.0^1,6]deca-2,4-triene radical cation, 2^ꢀ⁺,

by methanol. The enthalpies of the corresponding truncated norcaradiene

system are given in brackets.

tertiary radicals, ^ꢀ3-A-OCH₃and ^ꢀ4-A-OCH₃. The symmetri-

cal tricyclane, 3, does not offer an unambiguous conclusion:

attack at C-1/C-6 is unencumbered form either face. On the

other hand, the chiral isomer, 4, clearly illustrates the role

of steric hindrance. One of the tertiary cyclopropane centers

Figure 1. Stereoview of unpaired spin density (bottom), singly occupied

molecular orbital (SOMO, center), and lowest unoccupied molecular orbital

(LUMO, top) for the radical cation of trans-1,2-dimethylcyclopropane.¹⁶

quaternary cyclopropane carbon; this reaction is favored

because (1) it gives rise to a more stable allylic free radical

(/^ꢀ1-A-OCH₃) and (2) it occurs with complete release of

ring strain.^7b

Ar

.

b

OCH₃

1-A-OCH₃

.

+

4'

1

.

4

5

H₃CO

.

a

.

1 ⁺

In contrast, tricyclo[4.3.1.0^1,6]deca-2,4-triene radical cation,

2^ꢀ⁺, which formally has two pathways of nucleophilic substi-

tution with relief of ring strain (a,b), fails to utilize either.

The species has spin and charge density at C-2, C-5, and

C-10 but is attacked exclusively at C-2/C-5 (c,d), retaining

the signiﬁcant ring strain in the products. The course of

this substitution is determined by the large LUMO coefﬁ-

cients at C-2, C-5 in contrast to the absence of major

LUMO coefﬁcients at C-10 or at the bridgehead carbons,

C-1 and C-6 (Scheme 1; Fig. 2).¹⁷

The role of steric factors is illustrated unambiguously by

comparison of two tricyclane isomers, 2,3,3-trimethyl-

tricyclo[2.2.1.0^2,6]heptane,¹⁸3, and 1,3,3-trimethyltri-

cyclo[2.2.1.0^2,6]heptane,¹⁵4. Nucleophiles attack 3^ꢀ⁺and

4^ꢀ⁺exclusively at tertiary carbons, not the quaternary ones;

this ﬁnding is ascribed to the formation of the more stable

Figure 2. Stereoview of unpaired spin density (bottom), SOMO (center) and

LUMO (top) for the radical cation of norcaradiene.¹⁷

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6473

of 4, C-2, is a neopentyl type carbon whereas C-6 is sterically

unencumbered. The result, exclusive attack at C-6, clearly

establishes a signiﬁcant role of steric hindrance in this

In order to probe various features of intramolecular reactions

of radical cations with nucleophiles, we selected a series of

substrates designed to cover a range of cases that would

allow us to elucidate the principles governing this reaction

type. For several of these species the tethered nucleophile

has only a single possibility to react whereas in others it

has the ‘option’ of attacking two or three different centers

via different transition states with different driving forces

and different barriers.

ꢀ

system.¹⁵The intermediates of both reactions, 3-A-OCH₃

and ^ꢀ3-A-OCH₃respectively, form the ﬁnal product by hydro-

gen, respectively, form the ﬁnal product by hydrogen atom

transfer to an adventitious acceptor.

OCH₃

.

+

3

2

H₃C

The systems chosen provide the opportunity to evaluate reac-

tions proceeding via four- to eight-membered transition

states, in several cases with intra-species competition. A

four-membered transition state is actually realized in one

system. Not surprisingly, several substrates utilize ﬁve- and

six-membered transition states; both substitution and capture

readily proceed via these common transition states. Three of

our systems have the potential of reacting via seven-

membered transition states; two radical cations react in this

fashion, one actually in competition with other cyclizations.

Finally, two radical cations offer the possibility of reaction

via an eight-membered transition state without, however,

using this pathway. Three of the systems studied contain

three-membered rings; they were chosen for the additional

driving force which relief of ring strain might provide.

OCH₃

3-A-OCH₃

b

.

a

1

.

3

+

OCH₃

.

3-B-OCH₃

H₃CO

.

3

+

CH₃

H

b

.

4-A-OCH₃

2

1

a

.

4

+

H₃CO

.

The seven target molecules chosen for our investigation are

listed in Chart 1; geraniol (5), nerol (6), citronellol (7),

chrysanthemol (8), homochrysanthemol (9), trans-1-o-

hydroxyphenyl-2-phenylcyclopropane (10), and endo-5-

hydroxymethylnorbornene (11), generate a family of mono-,

bi-, or tricyclic ethers via a series of four- to seven-

membered transition states.

CH₃

H

.

4-B-OCH₃

Three of the reactions discussed above, those of 1^ꢀ⁺, 3^ꢀ⁺, and

4^ꢀ⁺, are ‘genuine’ substitution reactions, akin to an S_N2 reac-

tion, albeit on substrates that are one electron shy of the dia-

magnetic prototype. These reactions involve backside attack

with release of a leaving group, in each case a free radical that

remains attached to the molecule. For these reactions, the tra-

jectory of approach may be important.¹⁴The reaction of 2^ꢀ⁺is

different as the nucleophile adds to, or is ‘captured’ by an

‘empty p-orbital’ or an electron deﬁcient p-system. This

reaction is akin to the second, product-forming step of an

S_N1 reaction. The stereochemical requirements for this

type of reaction may be less demanding.

CH₂OH

7

5

6

CH₂OH

We have an interest in the intramolecular variant of these

reactions, i.e., nucleophilic substitution or capture of radical

cations by hydroxy groups tethered to the centers bearing

spin and charge. Compared to the more general case of cap-

ture by an external nucleophile the intramolecular case offers

some intriguing features. Steric features, including confor-

mational preferences, may play a signiﬁcant role in the sub-

stitution by or capture of an intramolecular nucleophile.

Particularly, introduction of ring strain is seen as a major

potential impediment for the cyclization of such systems.

In general, the factors may be similar to those governing

other ring-forming reactions such as the Dieckman conden-

sation¹⁹or intramolecular nucleophilic substitution of

bifunctional compounds.²⁰For example, the Dieckman con-

densation works well for ﬁve-, six-, and seven-membered

rings, whereas larger rings require high-dilution techniques,

a direct reﬂection of the conformationally encumbered tran-

sition states of such ring systems. Similarly, intramolecular

S_N2 reactions proceed optimally via ﬁve- and six-membered

transition states.

8

9

OH

HO

11

Ph

10

Chart 1. Target systems chosen for this study.

In all these systems the hydroxyl function serves as the

nucleophile. This moiety is not a potent nucleophile in S_N2

reactions; for example, OH or COOH functions only undergo

intramolecular substitution after being deprotonated.

However, in reactions with radical cations, as well as carbo-

cations, which are superacids as well as ‘super electrophiles’,

the OH function is an efﬁcient nucleophile. This is borne out

by rates of nucleophilic capture, e.g., k¼w10⁸l mol^ꢂ1s^ꢂ1

,

for the capture of phenylcyclopropane radical cation by

methanol.^13d

6474

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

In selected cases, these experimental ﬁndings are viewed in

the light of molecular orbital calculations with ab initio or

density functional theory (DFT) methods (vide infra). In ad-

dition to the geometries of selected intermediates, especially

key bond lengths, we have generated pictorial representa-

tions of unpaired spin densities (r), the singly occupied

molecular orbitals (SOMOs), and the lowest unoccupied

(LU) MOs. The unpaired spin densities on carbon determine

the hyperﬁne coupling constants (hfcc) of free radicals

and radical ions; in most cases the signs and magnitudes

of hfccs provide the only available experimental evidence

for the structures of these intermediates.

of the Rehm–Weller equation (Eq. 10), [2.6 eV/3ꢂ0.13 eV]

is an empirical term accounting for solvents of different po-

larity,^21awhen the excitation energy and redox potentials are

measured in acetonitrile; in polar solvents, e²/3a has a value

of w0.06 eV. Typically, a driving force, ꢂDG_ETꢃ0.5 eV, is

sufﬁcient to generate solvent separated radical ion pairs

(SSRIP).

ꢂDG⁰¼ E_ð0;0Þꢂ E_ð⁰_D=D_þ_Þþ E_ð⁰_A_ꢂ_=AÞꢂ e²=3a;

ð9Þ

ꢂDG_S⁰_SRIP¼ E_ð0;0Þꢂ E_ð⁰_D=D_þ_Þþ E_ð⁰_A_ꢂ_=AÞ

ꢂ ½2:6 eV=3 ꢂ 0:13 eVꢄ

ð10Þ

2. Method of generation

The oxidative power of an acceptor/sensitizer can be gauged

by its excited state reduction potential,

The radical cations of the target molecules, 5–11, were gen-

erated by electron transfer to a photo-excited sensitizer or

a sensitizer/co-sensitizer system (e.g., Scheme 2, Eq. 1–3).

The intramolecular nucleophilic substitution or capture of

radical cations, 5^ꢀ⁺–11^ꢀ⁺, give rise initially to bifunctional

radical cations containing an oxonium ion and a localized

or delocalized free radical. Rapid deprotonation of the oxo-

nium ions leaves free radicals (Eq. 4), which have several

potential reactions available, including aromatic substitution

on the sensitizer radical anion (Eq. 6); reduction by return

electron transfer (RET) from the sensitizer radical anion, fol-

lowed by protonation (Eq. 5); or hydrogen abstraction from

an adventitious hydrogen atom donor (Eq. 7). In some cases,

the neutral radicals may form alkenes by hydrogen transfer

to a suitable hydrogen atom acceptor (Eq. 8).

^ꢅE_ð⁰_A_ꢂ_=AÞ¼ E_ð⁰_0;0Þþ E_ð⁰_A_ꢂ_=AÞ

ð11Þ

Photosensitizers used in this study include 1,4-dicyanobenz-

ꢂ

0

ꢂ

ene (DCB; E_(0,0)¼4.29 eV, E_{(A /A)}¼ꢂ1.6 V; *E_{(A /A)}

¼

2.7 V), 9,10-dicyanoanthrazene (DCA; E⁰_(0,0)¼2.88 eV,

0

ꢂ

E_{(A /A)}¼ꢂ0.9 V; *E_{(A /A)}¼2.0 V), and triphenylpyrylium

0

ꢂ

E_{(A /A)}

tetraﬂuoroborate

(TPP;

¹E₍⁰_0,0)¼2.8 eV,

¼

1

0

3 0

ꢂ

ꢂ0.29 V; E_{(A /A)}¼2.5 V; E_{(A /A)}¼2.0 V). Their oxidative

strength should be sufﬁcient to oxidize all electron donors

studied. Accordingly, all photoreactions discussed here can

be viewed as radical cation reactions. Details will be given

as appropriate.

¹Sens*

³Sens*

Sens

(1)

(2)

(3)

.

–

+

^1,3Sens*

+

co-sens

Sens

+

co-sens

3. Results

.

Photo-induced one-electron transfer causes the terpene gera-

niol, 5, to undergo an interesting ﬁve-center C–C cyclization.

With 1,4-dicyanoanthracene/biphenyl (DCA/BP) as sensi-

tizer/co-sensitizer, a primary stereoselective C–C cyclization

is followed by intramolecular hydrogen transfer yielding 12.

With 9,10-dicyanobenzene/phenanthrene (DCB/Ph) the

C–C cyclization is less selective and is followed by a second

ﬁve-center cyclization, capture of C₇by the alcohol function.

Overall, this reaction generates a series of cis- and trans-

fused 3-oxabicyclo[3.3.0]octanes, 13.^22a

–

+

–

+

Sens + co-sens

+

D

Sens

co-sens +D

H

+

H

.

D

.

D

O

_D.+

(4)

.

–

O

.

O

–

D

Ar-CN

B-H

–

–B

D

D H

(5)

(6)

– Ar-CN

O

.

–

D

Ar

H

D

+

CN

+

Ar-CN

R

H₃C

CH₂OH

O

.

D

(7)

(8)

O

+

R

Y

H

+

R

O

.

D

12

5

13

R = H,C₆H₄-CN

.

D

Y–H

Scheme 2. Photo-induced electron transfer reactions of donor molecules

bearing a tethered hydroxyl function.

The Z-isomer of 5, nerol, 6, undergoes a range of signiﬁ-

cantly different reactions. Irradiation of DCA/BP in the

presence of 6 gives rise to two monocyclic and two bicyclic

ring systems, 14–17. The formation of two products, 1,2,2-

cyclopentylacetaldehyde, 14, and 4-cyanophenyl-5,7,7-tri-

methyl-2-oxabicyclo[3.2.1]octane, 15, appears compatible

with a ﬁve-center C–C cyclization; one product, 2,2,6-tri-

methyl-7-oxabicyclo[4.2.0]octane, 16, arises by a ‘tandem’

The free energy of formation of radical ion pairs generated

by electron transfer between a donor and an acceptor

(Scheme 2; Eq. 3) is determined by the excited state energy,

(E_0,0), the reduction potential of the acceptor, E₍⁰_A^ꢂ_/A), the ox-

0

+

idation potential of the donor, E_{(D/D )}, and a term accounting

for ion pairing (Eqs. 9 and 10).²¹In a modiﬁed formulation

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6475

cyclization, initiated by a six-center C–C ring closure, which

is followed by a four-center nucleophilic capture. Finally,

the oxepene, 4-methyl-7-isopropyl-3-oxepene, 17, requires

a seven-center intramolecular nucleophilic capture. With

the DCB/Ph pair, 6 undergoes Z-to-E isomerization, and

forms aldehyde, 14, in addition to traces of 17 and tandem

cyclization product, 15.^22b

The electron transfer photochemistry of trans-10 was carried

out with triphenylpyrylium tetraﬂuoroborate (TPT). Sensi-

tized irradiation of trans-1 through Pyrex in methylene chlo-

ride under argon gave rise to 2-phenyl-2H-benzopyran, 21,

as the sole product (15% conversion after 5 h, 97% material

balance).²⁵

Ph

O

OH

CHO

R

Ph

21

10

O

15

14

CH₂OH

Finally, irradiation of DCA/BP as sensitizer/co-sensitizer in

acetonitrile in the presence of endo-5-hydroxymethylnor-

bornene, 11, generated the tricyclic ether, 4-oxatricyclo-

[4.2.1.03,7]nonane, 22, in moderate yield.²²

O

6

17

R

16

R = H, C₆H₄-CN

Irradiation of DCA/BP as sensitizer/co-sensitizer in aceto-

nitrile in the presence of citronellol, 7, which has only one

oleﬁnic function (i.e., dihydro-6), gives rise to the oxepane

derivative 18 (i.e., dihydro-17) in good yield, conﬁrming

the dimethylethene function as a suitable electron donor

under these conditions. This reaction requires a seven-center

intramolecular nucleophilic capture.^22b

O

HO

11

22

4. Discussion

We will discuss the radical cations whose reactions are being

treated here in three groups. First we consider the terpene

radical cations, 5^ꢀ⁺, 6^ꢀ⁺, and 7^ꢀ⁺; they are of particular interest

because 5^ꢀ⁺and 6^ꢀ⁺provide a competition between C–C and

C–O bond formations. The second group, comprised of

chrysanthemol and homochrysanthemol radical cations, 8^ꢀ⁺

and 9^ꢀ⁺, offers two different competing pathways involving

nucleophilic substitution and capture. Finally, we discuss

two radical cations that undergo 1,5-cyclizations, at least

formally. The diarylcyclopropane radical cation, 10^ꢀ⁺, poses

an intriguing structure/reactivity problem whereas 11^ꢀ⁺

probes the interesting aspect of endo attack on a norbornene

radical cation.

O

CH₂OH

R

18

7

R = H, C₆H₄-CN

The OH function of chrysanthemol [1-(2-hydroxy-ethyl)-

2,2-dimethyl-3-(2-methyl-1-propenyl)-cyclopropane] radi-

cal cation, 8^ꢀ⁺, could attack the three-membered ring by

nucleophilic substitution or the oleﬁnic side chain by nucleo-

philic capture. Only the capture reaction is observed forming

ring-expanded aryl-substituted products of type 19.²³

4.1. Geraniol and nerol radical cations—competing C–C

and C–O cyclizations

Ar

HO

O

The radical ions of geraniol and nerol undergo both C–C and

C–O cyclizations. Both processes generate bifunctional rad-

ical cations. The difference lies in the fact that the C–O clo-

sures generate a ‘product’ containing an oxonium ion, which

is readily deprotonated, and is not suitable for further cycli-

zation. In contrast, the species resulting from C–C closure

contain a carbocationic site; intermediates of this type are

less easily deprotonated and are still suitable/susceptible

for nucleophilic capture. These species have the potential

for tandem cyclization: both 5^ꢀ⁺and 6^ꢀ⁺undergo 1,5- and/

or 1,6-C–C cyclizations, which are followed by nucleophilic

capture.

19

8

Homochrysanthemol [1-(3-hydroxy-propyl)-2,2-dimethyl-

3-(2-methyl-1-propenyl)-cyclopropane] radical cation, 9^ꢀ⁺,

like the lower homolog 8^ꢀ⁺, has two possible pathways for

the OH function to react with the electron deﬁcient moiety.

The extended tether diverts the point of attack completely

to the three-membered ring, giving rise to aryl-substituted

products of type 20, as well as a free radical dimer.²⁴

Ar

Because of the E-geometry of geraniol the nucleophilic

center of its radical cation, 5^ꢀ⁺, can only serve in this capacity

after having been ‘released’ or ‘unlocked’, for example,

by an addition at C-2. Accordingly, it was not surprising

that 5^ꢀ⁺reacted by bond formation between C-2 and C-6,

i.e., via a ﬁve-membered transition state; the resulting

O

HO

20

)

2

9

6476

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

bifunctional radical cation, cis-5-C–Ccy-5^ꢀ⁺, is a ditertiary

species and should be relatively stable. Indeed, with DCA/

BP as sensitizer/co-sensitizer, the reaction was arrested at

the stage of the monocyclic intermediate. The high stereo-

speciﬁcity observed with DCA/BP was unexpected: only

the cis-fused ring was formed and only one stereoisomer of

12 was generated.^22a

into a challenge and has precluded synthetic utility, at least

to date.^22a

CH₃

R

CH₂OH

.

O

+

.

Z-C-Ccy-5 ⁺

Z,Z- and E,Z-12

R = H, C₆H₄-CN

Z-tan-5

In order to explain the (unexpected) stereospeciﬁcity of ring

closure, we note the limited driving force for electron trans-

CH₃

R

CH₂OH

ꢂ

.

fer. The reduction potential of DCA (E_{(A /A)}¼ꢂ0.9 V), and

O

its low excited state energy (E_0,0¼2.88 eV) render the oxida-

1

+

tion of 5 by DCB* mildly exergonic (DG_ETwꢂ0.1 eV) in

.

E-C-Ccy-5 ⁺

E,Z- and E,E-12

R = H, C₆H₄-CN

CH₃CN and slightly endergonic (DG_ETw+0.2 eV) in

CH₂Cl₂. Accordingly, the radical ions DCA^ꢀ^ꢂand 5^ꢀ⁺can

be generated only as tight (‘sandwich’) ion pairs, causing

the substituents at the developing bonds to move ‘outward’,

away from the counter ion, resulting in the observed cis-

stereochemistry. The orientation of the methyl group trans

to the isopropenyl function was explained via triplet return

electron transfer, generating a biradical, cis-5-C–Ccy-5^ꢀꢀ;

a 1,5-intramolecular (suprafacial) hydrogen shift readily

accounts for the trans-stereochemistry.^22a

E-tan-5

The structural features of the three key radical cations, 5^ꢀ⁺,

cis-5-C–Ccy-5^ꢀ⁺, and 5,5-tan-5^ꢀ⁺, were calculated using den-

sity functional theory (DFT) methods (Fig. 3); in order to al-

low a comparison of their calculated enthalpies the proton on

the oxonium ion, 5,5-tan-5^ꢀ⁺was not allowed to dissociate.

The ﬁrst (C–C) cyclization is endergonic by 2.5 kcal mol^ꢂ1

,

whereas the second (O–C) cyclization is exergonic by

4.3 kcal mol^ꢂ1. The deprotonation of the oxonium function

lowers the enthalpy further, rendering the tandem cycliza-

tion irreversible.

Return electron transfer in triplet radical ion pairs has been

of great interest for several decades.²⁶Of special interest

are systems that undergo rearrangement during the consecu-

tive lifetimes of radical ions and triplet states or biradicals.²⁷

Geraniol belongs to a group of systems where the radical cat-

ion undergoes a major structure change, 5^ꢀ⁺to cis-5-C–Ccy-

5^ꢀ⁺, a species corresponding to a non-Kekule structure in the

ground state.²⁷The reorganized structure is retained upon

electron return (cis-5-C–Ccy-5^ꢀ⁺to cis-5-C–C-cy-5^ꢀꢀ), then

reverts to a Kekule structure with a minor change in struc-

ture, a 1,5-H shift forming cis,trans-12.²⁷Details of this

interesting topic go beyond the scope of this article.

A comparison of the spin and charge density distributions in

the three species is illuminating. The parent radical cation,

5^ꢀ⁺, has unpaired spin density in both alkene groups; the

dimethylethylene carbons bear similar spin densities (r₆¼

0.212, r₇¼0.213), whereas the hydroxyl group apparently

polarizes the partial double bond between C-2 and C-3

(r₂¼0.272, r₃¼0.096). In the ﬁrst cyclization product, cis-

5-C–Ccy-5^ꢀ⁺, the spin is distributed between C-3 (r¼0.461)

and C-7 (r¼0.542). Upon closing the second ring, forming

5,5-tan-5^ꢀ⁺, the unpaired spin becomes localized on C-7

(r¼0.958) whereas the charge is placed on the oxygen, polar-

izing the bonds to the adjacent atoms (C₁¼0.339, C₇¼0.265,

O–H¼0.474).

H

CH₂OH

+

CH₂OH

.

RET

1,5-H

.

..

+

cis,trans-12

cis-C-Ccy-5

Interestingly, under different reaction conditions, using 1,4-

DCB/Ph in anionic micellar solution, the acetate of 5 (23) as

well as the sesquiterpene and diterpene acetates, farnesyl

(24) and geranylgeraniol acetate (25), undergo 1,6-cycliza-

tions, generating six-membered mono-, bi-, and tricyclic

For the reaction of 5 with the DCB/Ph pair, electron transfer

from the donor substrates to ¹DCB* dicyanobenzene

0

ꢂ

(E_(0,0)¼4.29 eV, E_{(A /A)}¼ꢂ1.6 V; *E_{(A /A)}¼2.7 V) should

be efﬁcient, regardless of the solvent. Under these conditions

the reaction takes a different course in two respects: (a) the

C–C ring closure is much less stereospeciﬁc and (b) the re-

action proceeds past the stage of 5-C–Ccy-5^ꢀ⁺, as the partial

positive charge at C-7 is captured by the hydroxy function,

completing a tandem cyclization.

DCB / Ph

anionic

micelle

OAc

HO

23

24

25

OAc

DCB / Ph

anionic

micelle

Because the electron transfer reaction is comfortably exer-

gonic (ꢂDG_ETw0.8 eV), the ion pair, DCA^ꢀ^ꢂ–5^ꢀ⁺, is gener-

ated as a solvent separated radical ion pair without

intra-pair interactions that would affect the stereochemistry

of C–C closure. The resulting bicyclic free radicals, Z- or

E-5^ꢀ, react by aromatic substitution²⁸(Scheme 2, Eq. 6) as

well as by hydrogen abstraction (Scheme 2, Eq. 7)—neither

reaction has any pronounced stereochemical preference. The

lack of speciﬁcity in this reaction turned product separation

HO

H

OAc

DCB / Ph

anionic

micelle

H

HO

H

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6477

2.47

.

5-C-Ccy-5 ⁺

0

.

5 ⁺

–4.3

.

5,5-tan-5 ⁺

Figure 3. Chem-3D models of geraniol radical cation, 5^ꢀ⁺, and two isomers generated by C–C cyclization, cis-5-C–Ccy-5^ꢀ⁺, and tandem cyclization, 5,5-tan-5^ꢀ⁺,

respectively. The geometries were calculated by DFT methods using the 6-31G* basis set; relative enthalpies (kcal mol^ꢂ1) are indicated.

products, respectively.²⁹The bi- and tricyclic products are

trans-fused and all are hydroxylated in anti-Markovnikov

fashion; apparently the radical ions are terminated by water.

These reactions are the ﬁrst examples of photochemically

initiated biomimetic terpenoid cyclizations.

7-O–Ccy-6^ꢀ⁺, respectively; the latter would be deprotonated

readily to 7-O–Ccy-6^ꢀ.^22b

.

(+)

CH₂OH

+

.

( )

+

O

H

CH₂OH

.

(+)

.

The Z-isomer of 5^ꢀ⁺, nerol radical cation, 6^ꢀ⁺, cannot be ex-

pected to undergo cyclization between C-2 and C-6 because

the conformer required for this conversion appears to be se-

riously hindered by the steric repulsion between the hydroxy-

methyl group and the terminal dimethylethylene moiety. For

5^ꢀ⁺, a helical arrangement can relieve the less severe steric re-

pulsion between the hydrogens at C₂and C₆without compro-

mising their required proximity. However, the helical

conformer of 6^ꢀ⁺is still severely hindered, so that alternative

reactions must be expected. Indeed, 6 does not generate any

product derived by bond formation between C-2 and C-6.^22b

( )

.

+

.

+

5-C-Ccy-6 ⁺

7-O-Ccy-6

6-C-Ccy-6

With DCB/Ph as sensitizer/co-sensitizer, 6 forms aldehyde

14 as the major product and undergoes Z-to-E isomerization.

This cyclization is not only sterically disadvantaged, but it

places spin and charge on two secondary carbons in the pu-

tative intermediate. For the conversion of 5-C–C-cy-6^ꢀ⁺to 14

we consider triplet electron return and an intramolecular

(suprafacial) 1,5-hydrogen shift from the b-carbon of the

side chain (C₂) to the cyclopentyl ring (C₆) in the resulting

biradical, 5-C–Ccy-6^ꢀꢀ; this species also explains the

observed Z-to-E isomerization.^26,27The bicyclic ether, 15,

suggests that 5-C–Ccy-6^ꢀ⁺, in analogy to 5-C–Ccy-5^ꢀ⁺,

undergoes intramolecular nucleophilic capture completing

a tandem cyclization. The resulting 5,6-tan-6^ꢀgenerates 15

by hydrogen abstraction (cf., Scheme 2, Eq. 7).^22b

However, other cyclizations also have steric challenges. For

example, products 14 and 15 result from ﬁve-carbon C–C

cyclization between C-3 and C-7, yielding 5-C–Ccy-6^ꢀ⁺. Al-

though this approach avoids the C-2–C-6 crowding, the C-3

onto C-7 approach does not appear signiﬁcantly less

crowded. Perhaps as a result of this steric impediment 6^ꢀ⁺

undergoes the most varied ring formations of any system

discussed here. The various reactions of 6^ꢀ⁺result in the for-

mation of ﬁve-, six-, and seven-membered rings and support

nucleophilic capture via four-, six-, and seven-membered

transition states. These products are compatible with pri-

mary cyclizations yielding 5-C–Ccy-6^ꢀ⁺, 6-C–Ccy-6^ꢀ⁺, and

.

(+)

CHO

CH₂OH

1,6-capt

.

RET

1,5-H

O

.

( )

+

.

5-C-Ccy-6 ⁺

14

5,6-tan-6

6478

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

With the DCA/BP as sensitizer/co-sensitizer, 6 is converted

to 14, 16, and 17. The intermediate leading to 16 is a cyclo-

hexane-1,4-diyl system, 6-C–Ccy-6^ꢀ⁺, formed by bond for-

mation between C₂and C₇; spin and charge are localized

on a secondary and tertiary carbon, respectively. This type

of system contains spin and charge in two parallel p-orbitals;

this arrangement allows ready delocalization, thereby stabi-

lizing the species.³⁰Interestingly, 6-C–Ccy-6^ꢀ⁺undergoes

nucleophilic capture of the tertiary carbon by the alcohol

function via a four-membered transition state. The conver-

sion of the 7-oxabicyclo[4.2.0]octan-4-yl radical, 6,4-tan-

6^ꢀ, to 16 by hydrogen abstraction (cf., Scheme 2, Eq. 7) is

unexceptional. The formation of 6,4-tan-6^ꢀfrom 6-C–C-

cy-6^ꢀ⁺is signiﬁcant as the only example of oxetane forma-

tion from a radical cation.²²The reason for the general

lack of speciﬁcity observed for 6^ꢀ⁺may well lie in the steri-

cally less demanding nature of nucleophilic capture and C–C

bond formations between two trisubstituted, sp²hybridized

carbons.^22b

carbon atoms C-3 and C-7 is connected with a steep rise in

energy due to the interference of the alkyl groups, rendering

the formation of 5-C–Ccy-6^ꢀ⁺highly unfavorable. Accord-

ingly, we searched for additional pathways leading from

6^ꢀ⁺to 14.

We noted that in the lowest energy conformer of 6^ꢀ⁺one of

the H atoms attached to C-1 is pointing inward, quasi poised

for a hydride shift to C-6. This shift would lead to an inter-

mediate, 6,1-H-6^ꢀ⁺, containing an allylic free radical teth-

ered to a tertiary carbocation. Rotational reorganization of

the tether would allow cyclization between C-3 and C-7,

yielding enol-14^ꢀ⁺. Return electron transfer from the counter

ion followed by tautomerization would complete a pathway

to the major product, 14, which potentially is lower in

energy. In contrast to the conversion of 6^ꢀ⁺to 5-C–Ccy-6^ꢀ⁺,

the hydride shift precedes the cyclization step in the newly

considered pathway.

+

O

.

( )

+

OH

CH₂OH

H

CH₂OH

.

(+)

+

.

6

1,6-H-6

6-C-Ccy-6 ⁺

6,4-tan-6

Finally, the hydroxymethyl group of 6^ꢀ⁺interacts with the di-

methylethylene function by forming the oxepene system, ap-

parently by nucleophilic capture of C₆. The Z-arrangement

of 6^ꢀ⁺causes the hydroxy function to reside in the general

vicinity of the molecule’s ‘tail end’, allowing nucleophilic

capture with formation of 7-O–Ccy-6^ꢀ⁺. The transition state

appears sterically congested, but it may be possible to

arrange the seven centers C₆through O in a quasi-boat that

will reduce steric repulsion yet facilitate capture of C₆by

the OH function.²²Formally, 6^ꢀ⁺also has the potential to

form an eight-membered ring system, but no product of

this structure type was observed.^22b

OH

+

OH

+

enol-14 ⁺

An examination of the proposed pathway by density func-

tional calculations revealed the hydride shift to be exergonic;

the corresponding transition state lies only 2.7 kcal mol^ꢂ1

above 6^ꢀ⁺. Furthermore, the cyclization of 1,6-H-6^ꢀ⁺to

enol-14^ꢀ⁺is endergonic by only 0.13 kcal mol^ꢂ1. These ener-

getic features readily explain why 14 is obtained as the major

product. The pertinent structures delineating the hydride

shift are shown in Figure 4.

These results show that the potential hypersurface of nerol

radical cation is signiﬁcantly more complicated than that of

5^ꢀ⁺. The products suggest three primary cyclized species, 5-

C–Ccy-6^ꢀ⁺, 6-C–Ccy-6^ꢀ⁺, and 7-O–Ccy-6^ꢀ⁺, and two species

resulting from tandem cyclization, 5,5-tan-6^ꢀ⁺and 6,4-tan-

6^ꢀ⁺, in addition to the parent radical cation, 6^ꢀ⁺. In order to

gain additional insight into this potential surface, the key

radical cations were probed using density functional theory

(DFT) methods. As was the case for the intermediates derived

from 5, the oxonium protons were not ‘allowed’ to dissociate.

The hydride shift revealed by the calculations is interesting

because it converts a bifunctional (‘distonic’) species in

which spin and charge occupy separate regions of the molec-

ular framework into a species in which spin and charge share

the same pi system. In general, hydride shifts in radical cat-

ions are not without precedent; several rigid radical cations

undergo stereospeciﬁc sigmatropic shifts. For example, the

puckered ions, anti- and syn-5-methyl-26^ꢀ⁺, undergo stereo-

speciﬁc hydride or methyl migration, respectively, forming

1-methylcyclopentene radical cation, 27^ꢀ⁺, as well as the

3-methyl isomer, 28^ꢀ⁺.³¹

The calculations yielded ﬁve of the radical cations con-

sidered as likely intermediates. The cyclized species,

6-C–Ccy-6^ꢀ⁺and 7-O–Ccy-6^ꢀ⁺, lie 9.96 kcal mol^ꢂ1and

8.85 kcal mol^ꢂ1, respectively, above 6^ꢀ⁺(DH¼0) and, thus,

should be accessible during its lifetime. The products of tan-

dem cyclization, 5,5-tan-6^ꢀ⁺(DH¼7.97 kcal mol^ꢂ1) and 6,4-

tan-6^ꢀ⁺(DH¼11.70 kcal mol^ꢂ1), appear likewise accessible;

deprotonation of the oxonium functions should lower the

free enthalpies further, rendering the second cyclizations

essentially irreversible. However, the calculations revealed

one irreconcilable discrepancy with the simple mechanism

considered: the putative precursor, 5-C–Ccy-6^ꢀ⁺, for the

major product (14) proved to be elusive. The approach of

R_syn

CH₃

R_syn= H

R_syn= CH₃

R_anti

+

CH₃

+

•

+

• +

28

• +

27

5-methyl-26

Similarly, sabinene radical cation, 1^ꢀ⁺, undergoes a stereo-

speciﬁc [1,3] shift to b-phellandrene radical cation, 29^ꢀ⁺,

with high retention of optical purity whereas a-thujene

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6479

2.72

.

TS-1,6-H-6 ⁺

0

.

6 ⁺

–6.0

.

1,6-H-6 ⁺

Figure 4. Chem-3D models of nerol radical cation, 6^ꢀ⁺, an isomer generated by H-shift, 1,6-H-6^ꢀ⁺, and the transition state for the H-shift, TS-1,6-H-6^ꢀ⁺. The

geometries were calculated by DFT methods (6-31G* basis set); relative enthalpies (kcal mol^ꢂ1) are indicated.

radical cation (not shown) undergoes competing [1,3] and

home-[1,5] shifts to a-phellandrene radical cation.^7cIn these

examples radical cations containing spin and charge in

a lengthened cyclopropane ‘sigma’ bond are converted

into species in which spin and charge share the same pi

system.

the spin density to a lesser degree than for 5^ꢀ⁺. The product

generated by hydride shift, 1,6-H-6^ꢀ⁺, has an unusual distri-

bution of spin and charge; the spin is located mainly on C-7

(r₇¼0.632) and to a lesser extent on C-1 (r₁¼0.214) and C-3

(r₃¼0.210) whereas the charge is distributed between C-1

(0.156) and C-3 (0.205) and C-7 (0.211). Cyclization prod-

uct 6-C–Ccy-6^ꢀ⁺bears spin density mainly at C-6

(r₆¼0.758) and C-3 (r₃¼0.322) whereas the charge has its

greatest density at C-3 (0.273), which interacts with the

hydrogen atoms at C-2, C-4, and the adjacent methyl group;

ﬁnally, 7-O–Ccy-6^ꢀ⁺has the unpaired spin essentially local-

ized at C-7 (r₇¼0.667) and the charge is placed on the oxy-

gen, polarizing the bonds to the adjacent atoms, C-6 (0.175),

C-1 (0.266), C-3 (0.197), and O–H (0.447). Tandem cycliza-

tion product 5,6-tan-6^ꢀ⁺has the unpaired spin localized on

C-2 (r₂¼0.944) whereas the charge on the oxygen atom

polarizes the bonds to the adjacent atoms, C-1 (0.323), C-6

(0.265), and O–H (0.473). Finally, 6,4-tan-6^ꢀ⁺has the un-

paired spin localized on C-6 (r₆¼1.019) whereas the charge

on the oxygen atom polarizes the bonds to the adjacent

atoms, C-1 (0.398), C-3 (0.277) and the adjacent methyl car-

bon (0.142), and O–H (0.471). The relative enthalpies of the

key intermediates are summarized in Figure 5.

H

+

29

1

The high barrier connected with the putative ring closure of

6^ꢀ⁺to 5-C–Ccy-6^ꢀ⁺affects the overall mechanism further as

it eliminates the ‘logical’ precursor for the tandem cycliza-

tion product, 5,6-tan-6^ꢀ⁺. The newly uncovered species is

not bifunctional and, therefore, cannot undergo a second

ring closure. The possibility of yet another hydride shift,

generating 5,6-tan-6^ꢀ⁺from enol-14^ꢀ⁺, is remote because of

the prohibitive enthalpy (DH¼+13.84 kcal mol^ꢂ1) of this

reaction.

A comparison of the spin and charge density distributions in

6^ꢀ⁺and the diverse radical cations derived from it provides

additional insights. The parent radical cation, 6^ꢀ⁺, has un-

paired spin density in both the alkene groups; the spin den-

sities of both the groups are slightly polarized (r₆¼0.167,

r₇¼0.212; r₂¼0.232, r₃¼0.137); the hydroxyl group affects

Our interest in citronellol, 7, a terpene with only one double

bond, not likely to be an excellent electron donor, arose from

the unusual reaction of 6^ꢀ⁺. This reaction might suggest

that the terminal dimethylethylene group serves as the pri-

mary electron donor. This possibility can be probed by

6480

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

11.7

.

6,4-tan-6 ⁺

9.96

.

6-C-Ccy-6 ⁺

8.85

.

7.97

7-O-Ccy-6 ⁺

.

5,6-tan-6 ⁺

0

.

+

6

–5.87

.

+

–6.0

enol-14

.

1,6-H-6 ⁺

Figure 5. Chem-3D models of nerol radical cation, 6^ꢀ⁺, and isomers: (i) generated by C–C or O–C cyclization, 6-C–Ccy-6^ꢀ⁺and 7-C–Ocy-6^ꢀ⁺, (ii) resulting from

tandem cyclization, 5,6-tan-6^ꢀ⁺and 6,4-tan-6^ꢀ⁺, and (iii) formed by a H-shift, 1,6-H-6^ꢀ⁺, and subsequent cyclization, enol-14^ꢀ⁺, respectively. The geometries were

calculated by DFT methods with the 6-31G* basis set; relative enthalpies (kcal mol^ꢂ1) are indicated.

investigating the electron transfer photochemistry of citro-

nellol, 7, in essence the dihydro-derivative of 6. Irradiation

of DCA/BP in the presence of 7 gives rise to oxepane 18

(i.e., dihydro-17) in good yield, conﬁrming the dimethyl-

ethene function as a suitable electron donor.²²

tive groups of 7^ꢀ⁺have a full complement of conformers

accessible and have to meet by conformational diffusion.

Again, 7^ꢀ⁺has the potential to form, in addition, an eight-

membered ring system. The fact that this was not realized

has thermodynamic as well as kinetic reasons. Energetic rea-

sons favor the product formed via 7-O–C-cy-7, because it

has a tertiary free radical site compared to the secondary

site of 8-O–Ccy-7^ꢀ. Conformational/kinetic reasons, such

as the precedent of the Dieckmann condensation,¹⁹further

argue against the eight-membered transition state as a viable

option.

However, 7^ꢀ⁺is of interest also because it offers a less biased

test for a seven-membered transition state than the formation

of 17 from 6^ꢀ⁺. Radical ion 7^ꢀ⁺lacks the Z-arrangement of

6^ꢀ⁺, which limits the hydroxy function to the same hemi-

sphere as the dimethylethylene target. In contrast, the reac-

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6481

O

CH₂OH

X

O

.

8-O-Ccy-7

7

7-O-Ccy-7

We note, however, that a nucleophilic substitution via an

eight-membered transition state has been achieved by

Floreancig and co-workers. In this reaction a benzyl group is

being replaced by a tethered hydroxy function (vide infra).³²

The intramolecular substitution in the system studied does

not have the option of a competing pathway, which may

be more favorable for either energetic or kinetic reasons.

4.2. Chrysanthemol and homochrysanthemol radical

cations—nucleophilic substitution versus nucleophilic

capture

Figure 6. Stereoview of unpaired spin density (bottom), SOMO (center) and

The second group of target molecules is comprised of chrys-

anthemol and homochrysanthemol, whose radical cations,

8^ꢀ⁺and 9^ꢀ⁺, offer two different cases of competing pathways

between nucleophilic substitution and capture. As deriva-

tives of vinylcyclopropane, the distribution of spin and

charge in these species is of special interest. The radical

cation of the parent system is of a very special structure

type: spin and charge are delocalized between the vinyl

group and the tertiary cyclopropane carbon (Fig. 6).³³This

radical cation is one of the few cyclopropane species with

two lengthened ring bonds.^3a,c,d,17,33

LUMO (top) for the radical cation of vinylcyclopropane.³²

radicals and radical ions from enhanced NMR spectra, ob-

served in emission and/or absorption, during radical (ion)

pair reactions.³⁴The CIDNP spectrum of the electron trans-

fer reaction from 8 to photo-excited chloranil (Fig. 7) shows

that the spin density of 8^ꢀ⁺is extended to C-1. This assign-

ment is conﬁrmed by the HOMO and LUMO coefﬁcients

of 8^ꢀ⁺(Fig. 8).²³

The SOMO and LUMO coefﬁcients of 8^ꢀ⁺and the distribu-

tion of spin and charge are such that nucleophilic substitu-

tion is possible at two cyclopropane carbons, C-2 and C-3.

In addition, nucleophilic capture of the b-carbon in the

2-methylpropenyl side chain is feasible. We evaluate the

probability of the potential pathways by considering steric

The additional substituents in 8^ꢀ⁺and 9^ꢀ⁺distort the symme-

try and affect the distribution of spin and charge. The nature

of 8^ꢀ⁺was probed by chemically induced dynamic nuclear

polarization (CIDNP), an NMR technique that allows one

to derive patterns of hyperﬁne coupling constants of free

Figure 7. ¹H CIDNP spectra observed during irradiation of chloranil solutions in acetonitrile in the presence of cis- (left) and trans-chrysanthemol, 8 (right).²³

The spectra support radical cations of very similar spin density distributions. The signals representing the two pairs of non-equivalent methyl groups atw1.6 and

at 1.1 and 1.0 ppm, respectively, show strong emission; the single oleﬁnic resonance appears in weak emission (not shown). The allylic cyclopropane proton

(1.25 ppm) and the proton adjacent to the hydroxymethyl group show negligible polarization.

6482

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

opening with formation of a tertiary free radical as an intra-

molecular leaving group. Formally, the overall mechanism

belongs to the elusive S_N2⁰type. In view of these consider-

ations, it is hardly surprising that 8^ꢀ⁺, exclusively forms

6-NC-8^ꢀ.²³The ﬁnal product, 19, once again is formed by

aromatic substitution (cf., Scheme 2, Eq. 7).

.

+

c

HO

b

a

H

a

c

H

O

b

.

O

H

.

H

4-NS-8

6-NC-8

H

We illustrate the conformational challenges to intramolecu-

lar nucleophilic substitution by exploring the conforma-

tional hypersurface of 8^ꢀ⁺using density functional theory

(DFT) methods. The lowest energy conformer of radical

cation 8^ꢀ⁺is one in which the hydroxyl function is deployed

far from either electron deﬁcient center, C-2 or Cb (Fig. 9).

Rotation of the 2-methylpropenyl function around the C-3–

C-1⁰bond results in a complex conformational proﬁle; for

comparison, the distance between the oxygen atom and

C-2 is shown for selected conformers (Fig. 10).

Figure 8. Stereoview of unpaired spin density (bottom), SOMO (center) and

LUMO (top) for cis-chrysanthemol radical cation, 8^ꢀ⁺.²³

and energetic factors. Substitution at C-2 generates 4-NuS-

8^ꢀ; substitution at C-3 generates 4⁰-NuS-8^ꢀwith inversion

of conﬁguration at C-3; ﬁnally, nucleophilic capture at the

b-carbon forms 6-NuC-8^ꢀ; the conﬁguration at C-1 is re-

tained in all three reactions.²³

The two pathways leading to oxetane formation do not

appear favorable, because they involve replacing a three-

membered ring by a four-membered one with negligible

change in strain energy. The pathway leading to 4⁰-NuS-8^ꢀ

is particularly unfavorable, because it generates an isolated

radical site, albeit a tertiary one, and a double bond rather

than an allylic radical, as in 4-NuS-8^ꢀ. For both reactions,

the trajectory of approach is far from the ideal 180^ꢁ. On the

other hand, the nucleophilic capture at the b-carbon offers an

unencumbered six-membered transition state, which has an

additional opportunity of stabilization by subsequent ring

The case of homochrysanthemol radical cation, 9^ꢀ⁺, is differ-

ent because the two reaction types, substitution and capture,

now proceed via ﬁve- and seven-membered transition states,

respectively. This changes the energetic and conformational

features for both. Nucleophilic substitution now becomes fa-

vorable because of a ﬁve-membered transition state in which

an allyl radical is an intramolecular leaving group.²⁴On the

other hand, capture of the b-carbon, generating 7-NuC-9^ꢀ,

now is encumbered by the presumably less favorable

seven-membered transition state.²⁴As a consequence of

Figure 9. cis,syn,syn-Conformer (left) and cis,syn,anti-conformer of cis-chrysanthemol radical cation, cis-8^ꢀ⁺(right) calculated by DFT methods (6-31G* basis

set).

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6483

+

O

S

Tol

S

Tol

OH

HO

5-NS-30 ⁺

O

.

31

30

S

Tol

O

OH

6-NS-30 ⁺

6-NS-30

Scheme 3.

substitution in general) are further illustrated by the failure of

33^ꢀ⁺to yield products that could arise via the seven-mem-

bered transition state, 7-NuS-33^ꢀ⁺(Scheme 4).

Figure 10. Conformational proﬁle of cis,syn-chrysanthemol, cis,syn-8^ꢀ⁺, for

rotation around the bond between C-3 and C-1⁰(—, solid curve) and C-2⁰–O

distance as a function of dihedral angle (- - -, dashed curve). The geometries

of seven individual conformers were calculated by DFT methods (6-31G*

basis set).

+

N

O

OH

HO

34

32

6-NS-33 ⁺

the changed energetics compared to 8^ꢀ⁺, the reaction of 9^ꢀ⁺

takes a different course: it proceeds exclusively via the

ﬁve-membered transition state, producing 5-NuS-9^ꢀ. The

ﬁnal product, 20, once again is formed by aromatic substitu-

tion (Scheme 2, Eq. 7).

+

HO

7-NS-33 ⁺

7-NS-33

+

Scheme 4.

+

HO

A system similar to 32, but with a side chain extended by one

methylene group, might be useful to probe the feasibility of

nucleophilic substitution via a seven-membered transition

state, although the approach of the OH function generating

the bridgehead free radical 7-NuS-35^ꢀmight have a trajectory

deviating from the ideal 180^ꢁ. A reaction of this type was

realized by Floreancig (vide infra).

+

9

O

7-NC-9

5-NS-

9

The intramolecular nucleophilic substitution of 9^ꢀ⁺was not

without precedent. A bicyclic system, 1-(3-hydroxypropyl)-

bicyclo[4.1.0]heptane radical cation, 30^ꢀ⁺, was known to

form the spiro-fused ether, 31. This reaction also involves

a ﬁve-membered transition state, 5-NuS-30^ꢀ⁺; it proceeds

by backside attack with inversion of conﬁguration and reten-

tion of chirality.³⁵In principle, 30^ꢀ⁺could also react via a six-

membered transition state, viz. 6-NuS-30^ꢀ⁺, but the approach

of the OH function would have to follow a trajectory far from

the ideal 180^ꢁ. Furthermore, this pathway does not beneﬁt

from the favorable loss of benzyl radical (Scheme 3).

+

N

HO

O

HO

7-NS-35 ⁺

7-NS-35(–H)

Arnold and co-workers evaluated the intramolecular nucleo-

philic capture of 6-methyl-5-hepten-2-ol, 36^ꢀ⁺, 6-methyl-6-

hepten-2-ol, 37^ꢀ⁺, and terpineol [4-(1-hydroxy-1-methyl-

ethyl)cyclohexene] radical cations, 38^ꢀ⁺.³⁷Their results

were published essentially simultaneously³⁸with our work

on geraniol^22aand chrysanthemol.²³The systems studied

can form tetrahydrofuran or pyran rings via ﬁve- or six-

membered transition states, respectively. The ﬁve-membered

transition state is favored signiﬁcantly: the internal alkene,

36^ꢀ⁺, exclusively forms the tetrahydrofuran; the terminal

alkene, 37^ꢀ⁺, preferentially forms the tetrahydrofuran ring,

eventhoughthisreaction requiresformation of a primary rad-

ical (!); ﬁnally, the terpene, 38^ꢀ⁺, prefers the ﬁve-membered

transition state by a ratio of 10:1.

Results observed in the photo-induced electron transfer

reaction of 1-(4-hydroxypentyl)-4-methyl-2,3-diazabicyclo-

[2.2.1]hept-2-ene, 32, provide an interesting complement

to those observed for 30. Radical cation, 32^ꢀ⁺, yields

a [6.4]spiro-fused six-membered ether, 34, via the deazetized

radical cation 33^ꢀ⁺.³⁶This product arises by backside attack,

via a six-membered transition state, 6-NuS-33^ꢀ⁺, with a tra-

jectory near 180^ꢁ. The rigid steric requirements for intramo-

lecular nucleophilic substitution (as well as for nucleophilic

6484

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

(E_0,0¼65 kcal mol^ꢂ1; DGwꢂ1.0 eV), and the triplet state,

+

³TPT* (E_T¼53 kcal mol^ꢂ1; DGwꢂ0.5 eV).⁴¹In addition,

the formal product of intramolecular capture, the benzylic

free radical, 5-NuC-10^ꢀ, is the likely immediate precursor

for 21.

O

OH

36

+

This leaves the question whether trans-10^ꢀ⁺is the direct pre-

cursor for 5-NuC-10^ꢀ, or whether the conversion of trans-

10^ꢀ⁺into 5-NuC-10^ꢀproceeds via an additional intermediate

and, thus, involves a mechanism other than intramolecular

nucleophilic substitution. The electronic structure of trans-

10^ꢀ⁺is of major importance in determining its reactivity.

Typically, disubstituted cyclopropane radical cations adopt a

‘trimethylene’ structure in which signiﬁcant spin and charge

reside in one lengthened C–C bond.^2e,3a,dHowever, the pres-

ence of the o-OH function in conjugation with charge and

unpaired spin, may change the structure of trans-10^ꢀ⁺to a

bifunctional one, e.g., bf-10^ꢀ⁺. The details of this interesting

conversion go beyond the scope of this paper. Approaches

to gain further insight into the mechanism of this reaction

will be reported in a separate paper, including the results

of density functional theory (DFT) calculations, laser ﬂash

photolysis (LFP), and an evaluation of the reactivity of a

related phenoxyl radical, trans-10^ꢀ(–H^ꢀ).

O

37

+

O

HO

6-O-Ccy-38(–

5-O-Ccy-38(–H)

H)

38

More recently, O–C cyclizations yielding ﬁve- through

eight-membered rings have been accomplished, in which

benzyl groups, the primary seat of spin and charge, act as

leaving groups for the tethered hydroxyl function; an alkoxy

group at the electrophilic center serves to further weaken the

benzylic bond.³²These studies included one case of an intra-

molecular competition between ﬁve- and six-membered

transition states. Substrate 39, which has hydroxyl functions

in both g- and d-positions, clearly prefers the formation of

the tetrahydrofuran derivative, 40, compared to the cycliza-

tion leading to the pyran derivative. In these reactions the

promise of synthetic utility is fulﬁlled.³²

H

+

O

OH

Ph

.

Ph

+

.

OH

trans-10 (–H )

+

trans-10

bf-10

Ph

OH

X

RO

Ph

RO

O

HO

OR

Laser ﬂash photolysis has proved to be an exceedingly

valuable tool to probe fast photo-induced conversions. This

technique has yielded a wealth of information about free

radicals, carbocations, carbenes, or nitrenes.⁴²Molecular

orbital calculations, either using ab initio or density func-

tional theory (DFT) methods are well suited to elucidate

this problem. Structural features such as spin and charge

density distribution will delineate the preferred structure

type and provide the key to the reactivity of the prevailing

species.

OH

40

39

4.3. Electron transfer photosensitized cyclization of

trans-1-(o-hydroxyphenyl)-2-phenylcylopropane

In the light of the preceding discussion, the formation of the

dehydrogenated cyclic ether, 21, upon triphenylpyrylium

tetraﬂuoroborate (TPT) sensitized irradiation of trans-10

poses a highly interesting mechanistic problem. The poten-

tial intramolecular capture of trans-10^ꢀ⁺is of special interest

because it would amount to an unprecedented front-side

substitution with retention of conﬁguration. Formally, the

required ﬁve-membered transition state has precedence

(vide supra), but the trajectory is far from the suggested ideal

one for nucleophilic substitution.¹⁴

4.4. Electron transfer photosensitized cyclization of

endo-2-hydroxymethylbicyclo[2.2.1]heptene and related

compounds

The ﬁnal intramolecular reaction to be discussed is the intra-

molecular capture of radical cation, 11^ꢀ⁺, derived from endo-

5-norbornene-2-methanol, which forms the tricyclic ether

22. Compared to the complex structural and mechanistic

problem posed by the reaction of trans-10 this system

may appear trivial, but it does command some interest in

its own right. We selected this target because of its relation-

ship with norbornadiene and norbornene radical cations,

41^ꢀ⁺and 42^ꢀ⁺, respectively, and to the radical cation, 43^ꢀ⁺,

OH

O

Ph

O

.

Ph

.

trans-10

21

5-NuS-10

The available evidence supports the intermediacy of the radi-

cal cation, trans-10^ꢀ⁺. Thus, the oxidation potential of trans-

10 is assumed to lie near that of 1,2-diphenylcyclopropane,

E_ox¼1.17 V versus Ag/Ag⁺;³⁹given the reduction potential

of TPT (E_red¼ꢂ0.29 V vs SCE),⁴⁰electron transfer is ener-

OH

1

getically feasible to both the excited singlet state, TPT*

43

41

42

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6485

1) –H⁺

2) Ar-CN ⁺

derived from 6,6-dimethylbicyclo[3.1.1]hept-2-ene-2-etha-

nol (nopol).

Norbornadiene radical cation, 41^ꢀ⁺, and its valence isomer,

quadricyclane radical cation, undergo a rich variety of re-

actions with nucleophiles.⁴³Interestingly, 41^ꢀ⁺is captured

exclusively from the exo face,^43a,bbut forms a molecular

complex with cyanoaromatic radical anions from the endo

side.^43c,dFor example, addition of methanol forms a free

radical, 41-A^ꢀ, which is in equilibrium with two isomeric

radicals, 41-B^ꢀand 41-C^ꢀ, by cyclopropylcarbinyl–butenyl-

carbinyl rearrangements.^43a,bThe free radicals form products

by reduction/protonation^43b(Eq. 5, Scheme 2), hydrogen

abstraction^43b(Eq. 7, Scheme 2), or aromatic substitution

(Eq. 6, Scheme 2).^43b

+

–

C₆H₄CN

–

3) CN

°

[α]₅₄₆= 170

46

1) –H⁺

OCH₃

45

44 ⁺

[α]₅₈₉= 48

2) CA

–Η

+

°

[α]₅₈₉= –86

47

Scheme 5.

Nopol radical cation, 43^ꢀ⁺, has two potential pathways for

intramolecular nucleophilic reaction, capture at C-3

(/49^ꢀ) and substitution at C-1 (/48^ꢀ). Given the fact that

signiﬁcant charge density is removed from the alkene func-

tion, it may not be too surprising that 43^ꢀ⁺fails to undergo

capture. In addition, this reaction will reintroduce any ring

strain partially relieved by delocalizing spin and charge

into the C-1–C-6 bond. The failure to undergo intramolecu-

lar substitution may have steric reasons; the required transi-

tion state is different from 5-NuS-30^ꢀ⁺(vide supra), because

three atoms are ﬁxed leaving only rotations around two

bonds to align the nucleophile for the required trajectory.

Given the relatively high energy calculated for intermediate

5-NuS-10^ꢀ⁺in the analogous reaction of 10, it is understand-

able that 43^ꢀ⁺only reacts with intermolecular nucleophiles

(/52^ꢀ; Scheme 6).¹⁸

OCH₃

H₃CO

OCH₃

H₃CO

.

41-A

41-B

41-C

OCH₃

Ar

OC

H₃

H₃CO

Ar

H₃CO

Ar

+

O

OH

Since norbornene, 42, on the other hand, is attacked from

endo and exo face,^8c11^ꢀ⁺was expected to readily undergo

nucleophilic capture. An interesting question concerned

the regiochemistry of capture, whether a six-membered tran-

sition state might give rise to a tricyclic pyran derivative. The

only product isolated from this reaction, 22, shows that

nucleophilic capture via a ﬁve-membered transition state

prevails; product 22 arises from the tricyclic radical inter-

mediate 5-O–Ccy-11(–H)^ꢀeither by hydrogen abstrac-

tion (Eq. 7, Scheme 2) or reduction/protonation (Eq. 5,

Scheme 2).^22b

O

+

43

48

49

O

OH

O

H₃CO

.

50

51

52

+

Scheme 6.

X

Concerning the relationship of 11^ꢀ⁺to 43^ꢀ⁺, the analogy is but

a formal one, a hydroxyl function attached in a fashion that

would allow a ﬁve-membered transition state for nucleo-

philic capture or substitution. The difference in reactivity

can be ascribed to the steric and electronic features that set

the reactions of 43^ꢀ⁺apart from those of 11^ꢀ⁺. One signiﬁcant

difference that might contribute to the eventual outcome of

the reactions, lies in the fact that 11^ꢀ⁺reacts by a Baldwin-

5-exo-trig process, whereas the conversion of 43^ꢀ⁺to 49^ꢀpro-

ceeds via a 5-endo-trig process. We ignore the 4-exo-trig

closure, which would generate a spiro-fused system. There

is only one analog for this reaction, the formation of 6,4-

tan-6^ꢀfrom 6-C–C-cy-6^ꢀ⁺(vide supra), where the OH func-

tion is captured by an empty p orbital. However, the analogy

is remote, as it does not involve a nucleophilic substitution.

O

HO

11

+

5-O-Ccy-11 (–H)

6-O-Ccy-11 (–H)

Nopol is one of several vinylcyclobutane systems whose

radical cations have attracted attention.^18,44As was demon-

strated particularly clearly for a-pinene radical cation, 44^ꢀ⁺,

these intermediates delocalize spin and charge between the

vinyl group and the cyclobutane ring while retaining their

chirality.⁴⁵They are captured by nucleophilic attack on C-6

(/45^ꢀ) or give rise to unusual ‘substitution’ products, e.g.,

46, which are initiated by deprotonation.^44,45The dehydroge-

nation product, verbenene, 47, also can be rationalized via

deprotonation (Scheme 5).^44,45

6486

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

5. Conclusion

of the resulting chalcone with hydrazine hydrate generated

a pyrazoline,^49,50which was deazetized under basic condi-

tions.⁵¹A mixture of endo- and exo-5-norbornene-2-metha-

nol (Aldrich) was used without separation because the

presence of the exo-isomer was not expected to interfere

with the chemistry of endo-11^ꢀ⁺.

The seven systems, 5–11, investigated in our laboratory and

additional systems, 30, 32, 36–39, and 43, reported in the lit-

erature, provide a consistent framework for the understand-

ing of intramolecular nucleophilic substitution and capture

of alkene and strained ring systems and offer guidelines

for further studies. The target molecules form mono-, di-,

and tricyclic ethers via four to seven-membered transition

states. Not surprisingly, the majority of cyclizations proceed

via ﬁve- and six-membered transition states for both substi-

tution and capture. A four-membered transition state is real-

ized only in one system: the bicyclic oxetane, 16, is formed

following the C–C cyclization of nerol radical cation, 6^ꢀ⁺,

which places a hydroxy group in close proximity to a carbo-

cationic site with no alternative cyclization accessible.

O

OH

C₆H₅CHO

Ph

Ex-4

Ex-5

N

NH

Ph

OH

Ph

trans-10

Ex-6

Two of the systems studied react via seven-membered tran-

sition states; in both cases competing reactions via eight-

membered transition states are avoided. The outcome of

the competition does not reﬂect kinetic factors alone; the

formation of the eight-membered ethers is also unfavorable

thermodynamically. Nerol radical cation, 6^ꢀ⁺, forms a seven-

membered ether in competition with C–C cyclizations form-

ing ﬁve- and six-membered rings. For homochrysanthemol

radical cation, 8^ꢀ⁺, nucleophilic substitution via a ﬁve-

membered transition state is preferred over nucleophilic

capture via a seven-membered transition state. Finally the

formation of dihydropyran, 21, from trans-1-(o-hydroxy-

phenyl)-2-phenylcyclopropane, trans-10, poses an intrigu-

ing mechanistic puzzle.

The electron acceptor/sensitizers, 1,4-dicyanobenzene (Al-

drich; 98%) and phenanthrene (Aldrich; 98%) were puriﬁed

by recrystallization. 9,10-Dicyanoanthracene (Eastman

Kodak) was puriﬁed by recrystallization from acetonitrile.

Acetonitrile (Fischer), methanol (Fischer), and methylene

chloride (Fischer; Spectranalyzed^Ò) were distilled from cal-

cium hydride and stored over 4A molecular sieves in brown

bottles under argon atmosphere.

6.2. Electron transfer photosensitized

reactions—irradiation procedures

Solutions containing 0.1 M of donors 1–5 or 7 and either

0.1 M of 1,4-dicyanobenzene/0.02 M phenanthrene or

0.1 M of 9,10-dicyanoanthracene/0.02 M biphenyl as sensi-

tizer/co-sensitizer in acetonitrile or methylene chloride were

deoxygenated by purging with argon for 15 min and irradi-

ated in a Rayonet RPR-100 photoreactor equipped with 16

RPR-3500 lamps. The progress of the reaction was moni-

tored by gas chromatography on a GC/MS system (HP

5890 series II GC interfaced with an HP 5971 mass selective

detector), using a 12 mꢆ0.2 mmꢆ0.33 mm HP-1 capillary

column (cross-linked methyl silicone on fused silica).

Exploratory runs were carried out in 4-mm ID NMR tubes

capped with latex stoppers, preparative runs in 30-mm ID

tubes with central cooling ﬁngers (water-cooling).

6. Experimental

6.1. Materials

Three hydroxy-substituted substrates, 5–7, are available

commercially (Aldrich). Chrysanthemol, 8, was prepared

by LiAlH reduction of the methyl chrysanthemate prepared

from a commercially available mixture of cis- and trans-

chrysanthemic acid (Aldrich). Homochrysanthemol, 9, was

prepared by LiAlH reduction of the methyl ester of homo-

chrysanthemic acid, which was prepared by Arndt–Eistert

homologation of chrysanthemic acid, Ex-1.⁴⁶The crude di-

azoketone, Ex-2, obtained by treatment of Ex-3 with thionyl

chloride, followed by reaction with diazomethane⁴⁷was

heated in methanol solution in the presence of silver benzo-

ate.⁴⁸

For donor 6 exploratory experiments were carried out by ir-

radiating solutions of 0.02 g substrate in 20 mL methylene

chloride with triphenylpyrilium tetraﬂuoroborate in 10%

molar ratio under argon for 1 h in Pyrex tubes surrounding

a central quartz cooling jacket with a 125-W medium-pres-

sure mercury lamp. For preparative runs solutions of 1.0 g

of the substrate in 400 mL freshly distilled methylene chlo-

ride were irradiated at ambient temperature with a 125-W

medium-pressure mercury lamp inside a quartz immersion

well.

SOCl₂

CH₂N₂

COOH

COCHN₂

Ex-2

Ex-1

LiAlH₄

(C₂H₅)O

CH₃OH

C₆H₅COOAg

CH₂COOCH₃

CH₂OH

6.3. Isolation of reaction products

9

Ex-3

Diarylcyclopropane trans-10 was prepared by a sequence of

reactions, initiated by Claisen–Schmidt condensation of

benzaldehyde with o-hydroxy-acetophenone; condensation

Reaction products obtained in yields>5% were isolated by

chromatography on columns, 1 cm<ID<5 cm, packed with

w15-cm of TLC standard grade silica gel (Aldrich; without

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6487

binder), and eluted with solvent gradients, usually from light

petroleum ether (bp<65 ^ꢁC) to mixtures with either methyl-

ene chloride or ethyl acetate. Several passes were required to

isolate the products.

SPARTAN, an HF/6-31G* single point calculation was car-

ried out followed by a surface analysis for spin density,

SOMO and LUMO.

The product resulting from trans-10 was isolated and puriﬁed

by conventional column chromatography on silica gel Merck

60 (0.063–0.200 mm), by preparative layer chromatography

on silica gel Merck 60 PF₂₅₄, using dichloromethane as

eluent, or by means of isocratic HPLC equipment ﬁtted

with a semi-preparative Microporasil column, using hexane/

ethyl acetate as eluent.

Acknowledgements

H. D. Roth thanks NSF (Grant NSF-97-14850) and grate-

fully acknowledges an IBERDROLA Award. R. R. Sauers

thanks the National Center for Supercomputer Applications

for an allocation of time on the IBM P Series 690 (Grant

CHE030060).

6.4. Characterization/identiﬁcation of products

References and notes

Structure assignments of isolated products rest on MS and

NMR data, including DEPT, two-dimensional COSY, and

HETCOR experiments, where appropriate. NOE difference

spectra were recorded to elucidate the substituent stereo-

chemistry and the spatial relationship between the various

functional groups. ¹H NMR spectra (CDCl₃; d, ppm down-

ﬁeld of TMS) were recorded on a Varian XL-400, a

300 MHz Varian Gemini instrument, or a Varian VXR-200

spectrometer. ¹³C and HETCOR spectra were recorded on

the Varian VXR-200 spectrometer operating at 50.3 MHz.

IR spectra were recorded on a GC–FTIR Hewlett-Packard

5965; the major bands are characterized by their n_max

(cm^ꢂ1). Mass spectra were obtained using a Hewlett-Packard

5988 A spectrometer.

1. (a) Radical Ions; Kaiser, E. T., Kevan, L., Eds.; Interscience:

New York, NY, 1968; (b) Forrester, R. A.; Ishizu, K.; Kothe, G.;

Nelsen, S. F.; Ohya-Nishiguchi, H.; Watanabe, K.; Wilker, W.

Organic Cation Radicals and Polyradicals. In Landolt-

Bornstein, Numerical Data and Functional Relationships in

¨

Science and Technology; Springer: Heidelberg, 1980; Vol. IX,

Part d2; (c) Yoshida, K. Electrooxidation in Organic Chemis-

try: The Role of Cation Radicals as Synthetic Intermediates;

Wiley: New York, NY, 1984; (d) Shida, T. Electronic Absorp-

tion Spectra of Radical Ions; Elsevier: Amsterdam, 1988; (e)

Radical Ionic Systems; Lund, A., Shiotani, M., Eds.; Kluwer

Academics: Dordrecht, The Netherlands, 1991; In addition,

SciFinder Scholar retrieves 2352 papers on the chemistry, spec-

troscopy, and physics of radical cations in the ﬁrst four years of

the current millennium alone.

6.5. Computational details

2. (a) Ledwith, A. Acc. Chem. Res. 1972, 5, 133–139; (b) Shida,

T.; Haselbach, E.; Bally, T. Acc. Chem. Res. 1984, 17, 180–

186; (c) Nelsen, S. F. Acc. Chem. Res. 1987, 20, 269–276; (d)

Bauld, N.; Bellville, D. J.; Harirchian, B.; Lorenz, K. T.; Pabon,

R. A., Jr.; Reynolds, D. W.; Wirth, D. D.; Chiou, H.-S.; Marsh,

B. K. Acc. Chem. Res. 1987, 20, 180–186; (e) Roth, H. D. Acc.

Chem. Res. 1987, 20, 343–370; (f) Mangione, D.; Arnold, D. R.

Acc. Chem. Res. 2002, 35, 297–304.

3. (a) Makosza, M. Pure Appl. Chem. 1997, 69, 559; (b) Makosza,

M.; Wojciechowski, K. Liebigs Ann.-Rcl. 1997, 9, 1805; (c)

Makosza, M.; Wojciechowski, K. Heterocycles 2001, 54,

445; (d) Makosza, M.; Wojciechowski, K. Chem. Rev. 2004,

104, 2631.

4. (a) Takahashi, Y.; Mukai, T.; Miyashi, T. J. Am. Chem. Soc.

1983, 105, 6511–6513; (b) Miyashi, T.; Takahashi, Y.; Mukai,

T.; Roth, H. D.; Schilling, M. L. M. J. Am. Chem. Soc. 1985,

107, 1079–1080; (c) Gassman, P. G.; Hay, B. A. J. Am.

Chem. Soc. 1985, 107, 4075–4076; (d) Gassman, P. G.; Hay,

B. A. J. Am. Chem. Soc. 1986, 108, 4227–4228; (e) Arnold,

A.; Burger, U.; Gerson, F.; Kloster-Jensen, E.; Schmidlin,

S. P. J. Am. Chem. Soc. 1993, 115, 4271–4281.

5. (a) Gassman, P. G.; Hay, B. A. J. Am. Chem. Soc. 1985, 107,

4075; (b) Gassman, P. G.; Hay, B. A. J. Am. Chem. Soc.

1986, 108, 4227.

6. Arnold, A.; Burger, U.; Gerson, F.; Kloster-Jensen, E.; Schmi-

dlin, S. P. J. Am. Chem. Soc. 1993, 115, 4271–4281.

7. (a) Adam, W.; Walter, H.; Chen, G.-F.; Williams, F. J. Am.

Chem. Soc. 1992, 114, 3007–3014; (b) Weng, H.; Sethuraman,

V.; Roth, H. D. J. Am. Chem. Soc. 1994, 116, 7021–7025; (c)

Weng, H.; Sheik, Q.; Roth, H. D. J. Am. Chem. Soc. 1995,

117, 10655–10661.

Density functional theory (DFT) and/or ab initio calcula-

tions⁵²were carried out with the GAUSSIAN 03 series of

electronic structure programs,⁵³in some cases with earlier

versions, using extended basis sets, including p-type polari-

zation functions on carbon (6-31G*). The geometries of the

neutral parent molecules and radical cations were optimized

at the unrestricted Hartree-Fock (UHF/6-31G*//UHF/6-

31G*) and UB3LYP/6-31G* levels, respectively. Previous

experience suggests that this level of theory will reproduce

the major geometric features of the systems under study.

Some radical cations were also calculated to include higher

degrees of electron correlation at the MP2 level of theory

(MP2/6-31G*//MP2/6-31G*). Wavefunction analyses for

charge and spin density distributions used the conventional

Mulliken partitioning scheme.⁵²

Møller–Plesset perturbation theory (MP2) reproduces posi-

tive ¹H hyperﬁne coupling constants satisfactorily, but over-

estimates spin densities on carbon and negative hfcs

signiﬁcantly, often by factors>2.^54–57On the other hand,

density functional theory methods⁵⁹give satisfactory agree-

ment with experimental results.^58–60Indeed, positive and

negative hfcs of norbornadiene, quadricyclane, and bicyclo-

butane radical cations are reproduced accurately with either

the (B3LYP/6-31G*//MP2/6-31G*) or the (B3LYP/6-31G*//

B3LYP/6-31G*) method.^58–60In selected cases, including

the radical cations of 1,2-dimethylcyclopropane, norcara-

diene, vinylcyclopropane, and chrysanthemol, pictorial rep-

resentations of spin density, SOMO and LUMO were

derived with the program SPARTAN.⁶¹The previously

optimized MP2/6-31G* geometries were imported into

6488

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

8. (a) Neunteufel, R. A.; Arnold, D. R. J. Am. Chem. Soc. 1973,

95, 4080–4081; (b) Maroulis, A. J.; Shigemitsu, Y.; Arnold,

D. R. J. Am. Chem. Soc. 1978, 100, 535–541; (c) Arnold,

D. R.; Snow, M. S. Can. J. Chem. 1988, 66, 3012–3026; (d) Ar-

nold, D. R.; Du, X. J. Am. Chem. Soc. 1989, 111, 7666–7667.

9. (a) Klett, M.; Johnson, R. P. J. Am. Chem. Soc. 1985, 107,

6615–6620; (b) Rao, V. R.; Hixson, S. S. J. Am. Chem. Soc.

1979, 101, 6458–6459; (c) Mizuno, K.; Ogawa, J.; Otsuji, Y.

Chem. Lett. 1981, 742–744; (d) Hixson, S. S.; Xing, Y. Tetra-

hedron Lett. 1991, 32, 173–174.

10. (a) Mattes, S. L.; Farid, S. Org. Photochem. 1983, 6, 233–326;

(b) Lewis, F. D. Photoinduced Electron Transfer; Fox, M. A.,

Chanon, M., Eds.; Elsevier: Amsterdam, 1988; Part C, Chapter

4.1; (c) Bauld, N. L. Adv. Electron Transfer Chem. 1992, 2, 1.

11. (a) Maslak, P.; Asel, S. A. J. Am. Chem. Soc. 1988, 110, 8260–

8261; (b) Maslak, P.; Asel, S. A. Angew. Chem., Int. Ed. Engl.

1990, 29, 8260–8261.

12. (a) Lorenz, K.; Bauld, N. L. J. Catal. 1985, 95, 613–616; (b)

Roth, H. D.; Schilling, M. L. M. J. Am. Chem. Soc. 1985,

107, 716–718; (c) Roth, H. D.; Schilling, M. L. M.; Abelt,

C. J. Tetrahedron 1986, 42, 6157–6166; (d) Roth, H. D.; Schil-

ling, M. L. M.; Abelt, C. J. J. Am. Chem. Soc. 1986, 108, 6098–

6099; (e) Miyashi, T.; Konno, A.; Takahashi, Y. J. Am. Chem.

Soc. 1988, 110, 3676–3677.

13. (a) Dinnocenzo, J. P.; Schmittel, M. J. Am. Chem. Soc. 1987,

109, 1561–1562; (b) Dinnocenzo, J. P.; Conlon, D. A. J. Am.

Chem. Soc. 1988, 110, 2324–2326; (c) Dinnocenzo, J. P.;

Todd, W. P.; Simpson, T. R.; Gould, I. R. J. Am. Chem. Soc.

1987, 109, 1561–1562; (d) Dinnocenzo, J. P.; Todd, W. P.;

Simpson, T. R.; Gould, I. R. J. Am. Chem. Soc. 1990, 112,

2462–2464.

(f) Roth, H. D.; Schilling, M. L. M. J. Am. Chem. Soc. 1980,

102, 6805–6811; (g) Karki, S. B.; Dinnocenzo, J. P.; Farid,

S.; Goodman, J. C.; Gould, I.; Zona, T. A. J. Am. Chem. Soc.

1997, 119, 431–432.

27. (a) Roth, H. D. J. Photochem. Photobiol., C: Photochem. Rev.

2001, 2, 93–116; (b) Roth, H. D. Pure Appl. Chem. 2005, 77,

1075–1085.

28. The aromatic substitution products, E,E-, E,Z-, Z,E-, and Z,Z-

12 (R¼C₆H₄–CN), are the ﬁrst of several examples for this

product type in this paper. For simple systems, the overall reac-

tion sequence generating them has been dubbed NOCAS,

for Nucleophile-Oleﬁn-Combination-Aromatic-Substitution.

In French NOCAS is converted to its mirror image, SACON,

´

for Substitution Aromatique avec Combinaison d’une Oleﬁn

et d’une Nucleophile. In the spirit of this acronym, the above

´

reaction isaTandem-Cyclization-Intramolecular-Nucleophilic-

Capture-Aromatic-Substitution or TCINCAS (a creative mind

has no limits designing awkward acronyms).

29. Hoffmann, U.; Gao, Y.; Pandey, B.; Klinge, S.; Warzecha,

K. D.; Krueger, C.; Roth, H. D.; Demuth, M. J. Am. Chem.

Soc. 1993, 115, 10358–10359.

30. (a) Guo, Q.-X.; Qin, X.-Z.; Wang, J. T.; Williams, F. J. Am.

Chem. Soc. 1988, 110, 1974–1976; (b) Ikeda, H.; Minegishi,

T.; Miyashi, T.; Lakkaraju, P. S.; Sauers, R. R.; Roth, H. D.

J. Phys. Chem. B 2005, 109, 2504–2511.

31. Williams, F.; Guo, Q.-X.; Kolb, T. M.; Nelsen, S. F. J. Chem.

Soc., Chem. Commun. 1989, 1835–1836.

32. (a) Kumar, V. S.; Floreancig, P. E. J. Am. Chem. Soc. 2001, 123,

3842–3843; (b) Wang, L.; Seiders, J. R., II; Floreancig, P. E.

J. Am. Chem. Soc. 2004, 126, 12596–12603.

33. Herbertz, T.; Roth, H. D. J. Am. Chem. Soc. 1998, 120, 11904–

11911.

14. (a) Shaik, S. S.; Dinnocenzo, J. P. J. Org. Chem. 1990, 55,

3434–3436; (b) Shaik, S. S.; Reddy, A. C.; Ioffe, A.; Dinno-

cenzo, J. P.; Danovich, D.; Cho, J. K. J. Am. Chem. Soc.

1995, 117, 3205–3222.

15. Wlostowski, M.; Roth, H. D. Unpublished results, cited in

Roth, H.D., Ref. 3d.

16. Herbertz, T.; Roth, H. D. Unpublished results, cited in Roth,

H.D., Ref. 3d.

17. Herbertz, T.; Blume, F.; Roth, H. D. J. Am. Chem. Soc. 1998,

120, 4591–4599.

34. (a) Closs, G. L. Adv. Magn. Reson. 1974, 7, 157–229; (b) Kap-

tein, R. Adv. Free Radical Chem. 1975, 5, 319–380; (c) Roth,

H. D. Encyclopedia of Nuclear Magnetic Resonance, Vol. 2;

Grant, D. M., Harris, R. K., Eds.; 1996, pp 1337–1350.

35. Takemoto, Y.; Ohra, T.; Koike, H.; Furuse, S.-i.; Iwata, Ch.;

Ohishi, H. J. Org. Chem. 1994, 59, 4727–4729.

36. Adam, W.; Sendelbach, J. J. Org. Chem. 1993, 58, 5316–5322.

37. (a) McManus, K. A.; Arnold, D. R. Can. J. Chem. 1995, 73,

2158–2169; (b) Arnold, D. R.; Connor, D. A.; McManus,

K. A.; Bakshi, P. K.; Cameron, T. S. Can. J. Chem. 1996, 74,

602–612.

38. The independent pursuit of these topics is borne out by the

dates of submission of the respective publications, June 20

(Ref. 37a) and November 29, 1995 (Ref. 37b), and July 31

(Ref. 22) and October 26, 1995 (Ref. 23). Of course, all four

papers were submitted long after Ref. 29 (May 24, 1993).

39. Wong, P. C.; Arnold, D. R. Tetrahedron Lett. 1979, 2101–2104.

40. Saeva, F. D.; Olin, G. R. J. Am. Chem. Soc. 1980, 102, 299–303.

18. Arnold, D. R.; Du, X. Can. J. Chem. 1994, 72, 403–414.

19. Schaefer, J. P.; Bloomﬁeld, J. J. Org. React. 1967, 15, 1–203.

20. Makosza, M.; Owczarczyk, Z. J. Org. Chem. 1989, 54, 5094.

´

21. (a) Makosza, M.; Golinski, J.; Baran, J. J. Org. Chem. 1984, 49,

1488; (b) RajanBabu, T. V.; Reddy, G. S.; Fukunaga, T. J. Am.

Chem. Soc. 1985, 107, 5473; (c) Makosza, M.; Bia1ecki, M.

J. Org. Chem. 1998, 63, 4878.

22. (a) Weng, H.; Scarlata, C.; Roth, H. D. J. Am. Chem. Soc. 1996,

118, 10947–10953; (b) Weng, H. Ph.D. Thesis, Rutgers

University, 1996.

´

41. Garcıa, H.; Miranda, M. A. Chem. Rev. 1994, 94, 1063–1089.

23. Herbertz, T.; Roth, H. D. J. Am. Chem. Soc. 1996, 118, 10954–

10962.

24. Herbertz, T.; Roth, H. D. J. Org. Chem. 1999, 64, 3708–3713.

42. (a) Johnston, L. J.; Scaiano, J. C.; Sheppard, J. W.; Bays, J. P.

Chem. Phys. Lett. 1986, 124, 493–498; (b) Das, P. K.

Chem. Rev. 1993, 93, 119–144; (c) Hadel, L. M.; Maloney,

V. M.; Platz, M. S.; McGimpsey, W. G.; Scaiano, J. C. J.

Phys. Chem. 1986, 90, 2488–2491; (d) Ford, F.; Yazawa, T.;

´

25. Delgado, J.; Espinos, A.; Jimenez, M. C.; Miranda, M. A.;

Roth, H. D.; Tormos, R., unpublished results.

26. (a) Goldschmidt, C. R.; Potashnik, R.; Ottolenghi, M. J. Phys.

Chem. 1971, 75, 1025–1031; (b) Ottolenghi, M. Acc. Chem.

Res. 1973, 6, 153–160; (c) Schomburg, H.; Staerk, H.; Weller,

A. Chem. Phys. Lett. 1973, 21, 433; (d) Taylor, G. N. Unpub-

lished results, cited in Roth H.D. Mol. Photochem. 1973, 5,

91–126. (e) Schulten, K.; Staerk, H.; Weller, A.; Werner,

H. J.; Nickel, H. Z. Phys. Chem. NF 1976, 101, 371–390;

Platz, M. S.; Ratzinger, S.; Fulscher, M. J. Am. Chem. Soc.

¨

1998, 120, 4430–4438; (e) Pandurangi, R. S.; Lustak, P.; Kuntz,

R. R.; Volkert, W. A.; Rogowski, J.; Platz, M. S. J. Org. Chem.

1998, 63, 9019–9030.

43. (a) Gassman, P. G.; Olsen, K. D. Tetrahedron Lett. 1983, 1, 19–

22; (b) Weng, H.; Roth, H. D. J. Org. Chem. 1995, 60, 4136–

4145; (c) Weng, H.; Roth, H. D. Tetrahedron Lett. 1996, 37,

H. D. Roth et al. / Tetrahedron 62 (2006) 6471–6489

6489

4895–4898; (d) Weng, H.; Roth, H. D. J. Phys. Org. Chem.

1998, 11, 101–108.

44. Zhou, D.; Sheik, M.; Roth, H. D. Tetrahedron Lett. 1996, 37,

2385–2388.

45. (a) Roth, H. D.; Weng, H.; Zhou, D.; Lakkaraju, P. S. Acta

Chem. Scand. 1997, 51, 626–635; (b) Zhou, D.; Chou, J.;

Roth, H. D. J. Phys. Org. Chem. 1999, 12, 867–874.

46. Crombie, L.; Crossley, J.; Mitchard, D. A. J. Chem. Soc. 1963,

4957–4969.

47. Black, T. H. Aldrichim. Acta 1983, 16, 3–10.

48. Wiberg, K. B.; Hutton, T. W. J. Am. Chem. Soc. 1956, 78,

1640–1645.

P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg,

J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain,

M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari,

K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;

Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko,

A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.;

Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe,

M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.;

Gonzalez, C.; Pople, J. A. Gaussian 03, Revision B.02; Gauss-

ian: Pittsburgh, PA, 2003.

54. Raghavachari, K.; Haddon, R. C.; Roth, H. D. J. Am. Chem.

Soc. 1983, 105, 3110–3114.

49. Chattree, N. S.; Sharma, T. C. Asian J. Chem. 1994, 6, 405–407.

50. Iksanova, S. V. Geterosikl. Soedin. 1991, 1131–1136; Chem.

Abstr. 1992, 116, 194239s.

51. Gao, C.; Hay, A. S. Polymer 1995, 36, 5051.

52. Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Ini-

tio Molecular Orbital Theory; Wiley Interscience: New York,

NY, 1986; presents a detailed description of the theoretical

methods used in this work.

53. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;

Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven,

T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.;

Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani,

G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara,

M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.;

Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.;

Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.;

Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;

Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala,

55. Raghavachari, K.; Roth, H. D. J. Am. Chem. Soc. 1989, 111,

7132–7136.

56. Roth, H. D.; Schilling, M. L. M.; Raghavachari, K. J. Am.

Chem. Soc. 1984, 106, 253–255.

57. Krogh-Jespersen, K.; Roth, H. D. J. Am. Chem. Soc. 1992, 114,

8388–8394.

58. (a) Eriksson, L. A.; Malkin, V. G.; Malkina, O. L.; Salahub,

D. R. J. Chem. Phys. 1993, 99, 9756–9763; (b) Eriksson,

L. A.; Malkin, V. G.; Malkina, O. L.; Salahub, D. R. Int. J.

Quant. Chem. 1994, 52, 879–901.

59. Batra, R.; Giese, B.; Spichty, M.; Gescheidt, G.; Houk, K. N.

J. Phys. Chem. 1996, 100, 18371–18379.

60. Roth, H. D.; Herbertz, T. 1996, unpublished results.

61. SPARTAN, SGI Version 4.0.4 GL; Wavefunction: Irvine, CA,

1991–1995; Deppmeier, B. J.; Driessen, A. J.; Hehre, W. J.;

Johnson, H. C.; Leonard, J. M.; Yu, J.; Lou, L., Development

Staff; Baker, J.; Carpenter, J. E.; Dixon, R. W.; Fielder, S. S.;

Kahn, S. D.; Pietro, W. J., Contributors.

Article Doi

Intramolecular nucleophilic capture of radical cations by tethered hydroxy functions

DOI: 10.1016/j.tet.2006.03.064

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.tet.2006.03.064

Full text of DOI:10.1016/j.tet.2006.03.064

Products guided by the article

R&D Labs maybe for 5617-92-5

Relevant to this article

Hot Product