Cyclopropylcarbinyl-type ring openings. Reconciling the chemistry of neutral radicals and radical anions

Article abstract of DOI:10.1021/ja0041831

Cyclopropylcarbinyl → homoallyl and related rearrangements of radical ions (a) are frequently used as mechanistic "probes" to detect the occurrence of single electron transfer in chemical and biochemical processes, (b) provide the basis for mechanism-base

Full text of DOI:10.1021/ja0041831

Published on Web 03/28/2002
Cyclopropylcarbinyl-Type Ring Openings. Reconciling the
Chemistry of Neutral Radicals and Radical Anions
‡
J. Paige Stevenson, Woodward F. Jackson, and J. M. Tanko*
Contribution from the Department of Chemistry, Virginia Polytechnic Institute and
State UniVersity, Blacksburg, Virginia 24061
Received December 5, 2000. Revised Manuscript Received September 14, 2001
Abstract: Cyclopropylcarbinyl f homoallyl and related rearrangements of radical ions (a) are frequently
used as mechanistic “probes” to detect the occurrence of single electron transfer in chemical and biochemical
processes, (b) provide the basis for mechanism-based drug design, and (c) are important tools in organic
synthesis. Unfortunately, these rearrangements are poorly understood, especially with respect to the effect
of substrate structure on reactivity. Frequently, researchers assume that the same factors which govern
the reactivity of neutral free radicals also pertain to radical ions. The results reported herein demonstrate
that in some cases structure-reactivity trends in radical ion rearrangements are very different from neutral
radicals. For radical ions, delocalizations of both charge and spin are important factors governing their
reactivity.
Introduction
process (e.g., ring opening radical cations generated from
cyclopropyl arenes). In most cases, only a meager amount of
2
9
A rigorous understanding of the mechanism and kinetics
associated with the unimolecular rearrangements of neutral free
radicals has been achieved in the past two decades. Rate
constants and, more significantly, activation parameters have
become available for rearrangements of many types of neutral
radicals. As such, these processes have proven enormously
valuable as probes for radical intermediates (i.e., with ap-
propriately designed substrates, the detection of rearranged
products signals intermediacy of free radicals) and as radical
clocks (i.e., the unimolecular rearrangement is used as a
molecular level “stopwatch” to time competing, bimolecular
(
(
(
3) Maslak, P.; Chapman, W. H., Jr.; Vallombroso, T. M.; Watson, B. A. J.
Am. Chem. Soc. 1995, 117, 12380-9.
4) Maslak, P.; Narvaez, J. N.; Vallombroso, T. M., Jr. J. Am. Chem. Soc.
1
995, 117, 12373-9.
5) Maslak, P.; McGuinn, J. M. Chem. Commun. (Cambridge) 1999, 2467-
2468.
(6) Dockery, K. P.; Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.; Gould, I. R.;
Todd, W. P. J. Am. Chem. Soc. 1997, 119, 1876-1883.
(
7) Freccero, M.; Pratt, A.; Albini, A.; Long, C. J. Am. Chem. Soc. 1998, 120,
284-297.
(
8) Andersen, M. L.; Wayner, D. D. M. Acta Chem. Scand. 1999, 53, 830-
836.
(9) Carra, C.; Fiussello, F.; Tonachini, G. J. Org. Chem. 1999, 64, 3867-
3877.
(
10) Donkers, R. L.; Maran, F.; Wayner, D. D. M.; Workentin, M. S. J. Am.
1
processes of the radical). An understanding of this chemistry
has led to well-established synthetic methods based upon neutral
radical rearrangements.
Chem. Soc. 1999, 121, 7239-7248.
(
(
11) Wiest, O. J. Phys. Chem. A 1999, 103, 7907-7911.
12) Roth, H. D. Acc. Chem. Res. 1987, 20, 343-50.
(13) Roth, H. D.; Weng, H.; Herbertz, T. Tetrahedron 1997, 53, 10051-10070.
(
(
14) Dinnocenzo, J. P.; Conlon, D. A. J. Am. Chem. Soc. 1988, 110, 2324-6.
15) Dinnocenzo, J. P.; Conlon, D. A. Tetrahedron Lett. 1995, 36, 7415-18.
There have been numerous studies dealing with bond cleav-
age in radical ions. The most common and perhaps best
(16) Weng, H.; Sheik, Q.; Roth, H. D. J. Am. Chem. Soc. 1995, 117, 10655-
6
1.
17) Hasegawa, E.; Ishiyama, K.; Horaguchi, T.; Shimizu, T. J. Org. Chem.
991, 56, 1631-1635.
18) Hasegawa, E.; Ishiyama, K.; Fujita, T.; Kato, T.; Abe, T. J. Org. Chem.
997, 62, 2396-2400.
(19) Yeh, S.-R.; Falvey, D. E. J. Am. Chem. Soc. 1991, 113, 8557-8558.
20) Robbins, R. J.; Falvey, D. E. J. Org. Chem. 1993, 58, 3916-3616.
(21) Sastry, G. N.; Bally, T.; Hrouda, V.; Carsky, P. J. Am. Chem. Soc. 1998,
2
understood examples of these are dissociative processes, in
(
(
1
which two fragments, a neutral radical and an ion are generated
•
-
from the parent radical ion (e.g., for a radical anion, A-B
A + B or A + B ).
f
1
•
-
-
• 3-11
There are fewer examples of bond
(
cleavage in radical ions which can be classified as rearrange-
ments, i.e., bond breaking leads to a single species, a distonic
120, 9323-9334.
(
22) Barone, V.; Rega, N.; Bally, T.; Sastry, G. N. J. Phys. Chem. A 1999, 103,
radical ion (in which the charge and spin reside on the same
217-219.
species).12-28
(23) Maslak, P.; Varadarajan, S.; Burkey, J. D. J. Org. Chem. 1999, 64, 8201-
8209.
Compared to neutral free radicals, our understanding of the
mechanism and kinetics associated with rearrangements of
radical ions, and consequently, their use as mechanistic probes
or clocks, is in the dark ages. Sometimes the “rearrangement”
of a radical ion is more complex than a simple, first-order
(24) Oxgaard, J.; Wiest, O. J. Am. Chem. Soc. 1999, 121, 11531-11537.
(
25) Ikeda, H.; Nakamura, T.; Miyashi, T.; Goodman, J. L.; Akiyama, K.; Tero-
Kubota, S.; Houmam, A.; Wayner, D. D. M. J. Am. Chem. Soc. 1998, 120,
5832-5833.
(
26) Ikeda, H.; Minegishi, T.; Abe, H.; Konno, A.; Goodman, J. L.; Miyashi,
T. J. Am. Chem. Soc. 1998, 120, 87-95.
(27) Adam, W.; Heidenfelder, T. Chem. Soc. ReV. 1999, 28, 359-365.
(
28) Bally, T.; Bernhard, S.; Matzinger, S.; Roulin, J.-L.; Sastry, G. N.;
Truttmann, L.; Zhu, Z.; Marcinek, A.; Adamus, J.; Kaminski, R.; Gebicki,
J.; Williams, F.; Chen, G.-F.; Fulscher, M. P. Chem. Eur. J. 2000, 6, 858-
868.
*
Address correspondence to this author. E-mail: jtanko@vt.edu.
Formerly J. Paige Phillips.
‡
(
(
1) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317-323.
2) Sav e´ ant, J. M. Acc. Chem. Res. 1993, 26, 455-461.
(29) Dinnocenzo, J. P.; Simpson, T. R.; Zuilhof, H.; Todd, W. P.; Heinrich, T.
J. Am. Chem. Soc. 1997, 119, 987-993.
10.1021/ja0041831 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 4271-4281
9
4271

A R T I C L E S
Stevenson et al.
Experimental Section
Materials. N,N-Dimethylformamide (DMF, EM Science, 98%) was
stirred over copper(II) sulfate (Aldrich, 98%) and activated alumina
(Aldrich, neutral, Brockman activity I) for >3 days and vacuum distilled
immediately before use. Alumina was flame dried under vacuum (until
evolution of water vapor ceased) prior to use. 1-Methyl-2-pyrrolidinone
(NMP) was vacuum distilled and stored over molecular sieves. Dimethyl
Figure 1. Ring opening of neutral radicals vs radical anions.
2
sulfoxide (DMSO) was stirred over CaH for several days and vacuum
distilled just prior to use. Tetra-n-butylammonium perchlorate (TBAP)
was prepared by the method of House37 and recrystallized 4× from
ethyl acetate/hexane and vacuum oven dried before use. 1-Acetyl-2,2-
dimethyl cyclopropane3 and cyclobutyl cyclopropyl ketone were
prepared according to published synthesis. The following compounds
were obtained from Aldrich and used as received: Cyclopropyl methyl
ketone, cyclobutyl methyl ketone, and dicyclopropyl ketone. All
mediators used in this study except fluoranthene (Agros Organics,
>98%) and anthracene (Matheson, Coleman & Bell, >98%) were
obtained from Aldrich and used as received.
General. GC/MS was performed on a Hewlett-Packard HP 5890
gas chromatograph interfaced to a HP 5970 low-resolution mass
spectrometer and a HP series computer. High-resolution mass spectral
data were obtained from a VG Analytical model 7070 E-HF double-
focusing magnetic sector high-resolution spectrometer with use of
electron impact (70 eV) ionization. GC analysis was performed on a
Hewlett-Packard 5890A gas chromatograph equipped with an FID
detector and an HP 3393A reporting integrator.
Electrochemistry. Electrochemical measurements were performed
on an EG&G Princeton Applied Research (EG&G/PAR) model 273
potentiostat/galvanostat interfaced to an MS-DOS computer. The details
of this system were described earlier.33 Voltammetric measurements
were performed on solutions which contained 0.5 M tetra-n-butylam-
monium perchlorate (TBAP) in DMF. Solutions were prepared by
weighing out 1.71 g of TBAP into a 10 mL volumetric flask. The 10
mL flask along with all other voltammetric cell pieces (with the
exception of the working electrode) were treated in a Baxter DP-22
vacuum-drying oven (30-40 mmHg) at 40° C for 1.5 h. Voltammetric
cell pieces were then placed immediately in a desiccator to cool. The
voltammetric cell was assembled under argon/nitrogen flow. The 10
mL flask containing electrolyte was diluted with purified DMF and
added to the voltammetric cell. The completed cell assembly was argon
kinetic information is available, yet such rearrangements are
critically important for both mechanistic and synthetic purposes.
Rearrangements of radical anions derived from single electron
reduction of ketones (e.g., using reagents such as SmI2) have
proven extremely useful synthetic methods.30 Radical ion
rearrangements are often used to diagnose single electron
transfer and the intermediacy of paramagnetic species in
numerous chemical processes. (For example, the most compel-
ling arguments that the enzyme monoamine oxidase B processes
substrates via a single electron-transfer pathway are based upon
the chemistry of a rearrangeable radical cation probe, benzyl
8,39
40
31
cyclopropylamine. )
Our studies have focused on the mechanism and kinetics of
rearrangements of radical anions derived from cyclopropyl
ketones. These species undergo a ring opening process directly
analogous to the cyclopropylcarbinyl f homoallyl neutral free
radical rearrangement (Figure 1), leading to a distonic radical
anion. In terms of substituent effects, both systems (Figure 1)
behave similarly: (1) Radical-stabilizing groups at the R-carbon
(R) decelerate the ring-opening process and (2) radical-stabiliz-
ing substituents (e.g., alkyl, phenyl, and vinyl) on the cyclo-
propyl group (R′) accelerate the reaction, leading to the more
substituted (and more stable) radical/radical ion. For both types
of process, relief of cyclopropyl ring strain is not the sole
mitigating factor associated with the rate of these rearrange-
ments; resonance energy (spin delocalization) in the ring-closed
vs ring-opened forms must also be considered.3
2-35
In a recent Communication,36 we reported that as important
as spin delocalization is in determining the rate of radical ion
rearrangements, charge delocalization is at least as (and likely
a more) important a factor. In our present paper, we test critical
assumptions which led to this conclusion, and expand the
discussion of structure/reactivity trends in radical anion rear-
rangements. Experimentally, radical anions generated from
aliphatic ketones 1-5 were studied by using direct (cyclic
voltammetry and preparative electrolysis) and indirect (homo-
geneous redox catalysis) electrochemical techniques. The results
of these experiments were augmented with molecular orbital
calculations.
(dry, deoxygenated) purged for approximately 15 min prior to use. The
electroactive substance was added to the cell only after clean
backgrounds were obtained at each sweep rate used (a maximum cell
current <10 uA at V ) 100 mV/s over the potential range -2.7 to
-3.2 V was considered acceptable).
A three-electrode voltammetry cell was used. The GCE working
electrode (5 mm diameter) was prepared for use by polishing with an
alumina slurry to a mirror finish and stored in a desiccator. The Ag/
+
Ag (0.1 M in CH CN, 0.337 V vs SCE) reference electrode was freshly
3
prepared. A Pt wire coil was used as the auxiliary electrode. Mechanical
stirring was performed between voltammetric runs to clean the working
electrode surface. Positive-feedback iR compensation was set by
monitoring the current response. IR compensation was increased until
oscillation and then backed off to 90% of that value. All experiments
were performed at ambient temperature.
The GCE working electrodes were prepared as follows. A glassy
carbon rod (Alfa Aesar) was cut into several 4-5 mm plugs. Carbon
plugs were secured into glass rods with Torrseal-Varian vacuum epoxy
resin (Varian vacuum products) being careful not to get epoxy on the
reverse of the carbon plug where the electrical connection must be made.
Rods were left to cure in air for 2 days. A small amount of silver 2-part
(
30) Berger, D. J.; Tanko, J. M. Radical anions and radical cations deriVed
from compounds containing CdC, CdO, or CdN groups; Patai, S., Ed.;
John Wiley & Sons, Ltd.: New York, 1997; pp 1281-1354.
(
(
(
(
31) Silverman, R. B. Acc. Chem. Res. 1995, 28, 335-342.
32) Tanko, J. M.; Drumright, R. E. J. Am. Chem. Soc. 1990, 112, 5362-5363.
33) Tanko, J. M.; Drumright, R. E. J. Am. Chem. Soc. 1992, 114, 1844-1854.
34) Tanko, J. M.; Drumright, R. E.; Suleman, N. K.; Brammer, L. E., Jr. J.
Am. Chem. Soc. 1994, 116, 1785-1791.
(37) House, H. O.; Feng, E.; Pert, N. P. J. Org. Chem. 1971, 36, 2371-2375.
(38) Dauben, W. G.; Wolf, R. E. J. Org. Chem. 1970, 35, 374-379.
(39) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353-1364.
(40) Hanack, M.; Ensslin, H. M. Justus Liebigs Ann. Chem. 1966, 697, 100-
110.
(35) Phillips, J. P.; Gillmore, J. G.; Schwartz, P.; Brammer, L. E., Jr.; Berger,
D. J.; Tanko, J. M. J. Am. Chem. Soc. 1998, 120, 195-202.
(36) Tanko, J. M.; Phillips, J. P. J. Am. Chem. Soc. 1999, 121, 6078-6079.
4272 J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002

Cyclopropylcarbinyl-Type Ring Openings
A R T I C L E S
Scheme 1
conductive adhesive (Alfa Aesar) was placed into the glass rods and a
short piece of Cu brazing rod was inserted and allowed to cure for 2
days. The electrode surface was polished with alumina slurry (Buehler)
starting with 1.0 um grit and decreasing to 0.3 and finally 0.05 um
until a mirror finish was obtained. The surface was maintained by
polishing routinely with 0.05 um grit polish, rinsing with methanol
and water, and wiping lightly.
Preparative electrolyses were performed on solutions which contained
0
.2 M TBAP in DMF. Solution preparation and cell handling were
performed similarly to that for voltammetry. A conventional H-cell
with two compartments separated by a medium glass frit was utilized.
Fifty milliliters of electrolyte solution was divided equally between
the two compartments of the vacuum oven dried H-cell, and the
resulting system was purged with dry, deoxygenated argon for >15
min before use. The electroactive substrate was placed in the working
compartment exclusively and electrolysis was conducted as specified
in the specific experiments. For constant potential experiments, iR
compensation was set as in voltammetric experiments. A gold foil
working electrode was utilized. All electrolysis experiments were
performed at ambient temperature. Reaction progress was monitored
by GC where necessary. Solution workup consisted of quenching the
Figure 2. Cyclic voltammogram of methyl cyclopropyl ketone (1, ν )
100 mV/s, 0.003 M).
constant ko. Either of these steps may be rate limiting, depending
on the structure of the substrate.
These techniques typically examine the effect of sweeprate
and substrate concentration on the electrochemical response.
When the chemical step is rate limiting, it is possible to obtain
information such as the formal reduction potential of the
substrate, the rate law for the chemical step (usually first order
in these systems), and the rate constant ko. When heterogeneous
electron transfer is rate limiting, it is possible to measure kS
cathodic compartment with ca. 1 mL of 5% H
of water, and extracting with 4 × 50 mL of ether. Ether layers were
combined, washed with water and saturated NaCl, dried over MgSO
2 4
SO , adding ca. 50 mL
4
,
and concentrated. Products of bulk electrolysis were all known
compounds. Yields were determined by GC vs an appropriate internal
standard.
4
6,47
and the transfer coefficient R.
Molecular Orbital Calculations. MO calculations were performed
4
1,42
Very quickly, it became obvious that cyclic and linear sweep
voltammetry are not effective tools for studying the radical
anions generated from 1-5. This point was graphically il-
lustrated by examining the cyclic voltammogram of cyclopropyl
methyl ketone (Figure 2). The small wave attributable to the
reduction of substrate was superimposed on a large background
current attributable to solvent/electrolyte decomposition. Only
at low concentrations and slow sweep rates was it possible to
generate the small voltammetric response shown. As concentra-
tion or sweep rate increased, the peak potential was quickly
shifted into the background response. Because of these com-
plications, a thorough linear sweep voltammetry study (i.e.,
examining peak potential as a function of sweeprate and
substrate concentration) was not feasible.
by using the density functional (B3LYP)
and/or the Hartree-Fock
43
theory as implemented through Titan (Wave function Inc., Irvine, CA,
and Schr o¨ dinger Inc., Portland, OR). Geometry optimizations were
4
4,45
performed with the 6-31+G*
basis set; single point calculations
4
4,45
were subsequently performed at the 6-311+G*
level to obtain the
pertinent energies and charge/spin densities. Vibrational frequencies
were calculated to verify that transition states were successfully located.
Electrochemical Simulations. Digital simulations of cyclic volta-
mmograms were performed with Digisim 2.1 (Bioanalytical Systems
Inc., 2701 Kent Ave. W. Lafayette, IN 47906) and working curves
were generated from simulated responses with use of TableCurve 2D
(Jandel Scientific Software: 2591 Kerner Blvd., San Rafael, CA 94901).
Results
Cyclic and Linear Sweep Voltammetry. Direct electro-
Crudely, the broadness of the peak (Ep/2 - Ep ∼106 mV,
where Ep and Ep/2 are the peak- and half-peak potentials,
respectively) suggests that the homogeneous electron-transfer
chemical techniques such as cyclic and linear sweep voltam-
metry (CV and LSV, respectively) are incredible tools for
studying the chemistry of radical ions.4
6,47
As applied to this
step (kS) is rate limiting. On this basis, the transfer coefficient
study, a cyclopropyl ketone (A, Scheme 1) is reduced at an
47
(
R) can be estimated according to eq 1 , yielding R ≈ 0.45.
•
-
electrode surface generating its radical anion A ; kS is the
heterogeneous rate constant for this step. This radical anion
undergoes subsequent ring opening yielding B with the rate
RT 1.85
F Ep/2 - Ep
•
-
R )
(1)
(
(
(
41) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
The transfer coefficient is a measure of transition state location.
For systems in which electron transfer and bond-breaking occur,
R can be used as a diagnostic tool to determine whether these
two processes occur in a manner that is stepwise or concerted
42) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5651.
43) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular
Orbital Theory; Wiley: New York, 1986.
(
44) Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett 1972, 66, 217.
45) Clark, T.; Chandrasekhar, J.; Spitznagel, K.; Schleyer, P. v. R. J. Comput.
Chem. 1983, 4, 294.
(
2
(
46) Andrieux, C. P.; Sav e´ ant, J. M. Electrochemical reactions,; 4th ed.;
Bernasconi, C., Ed.; Wiley: New York, 1986; pp 305-390.
47) Parker, V. D. Linear Sweep and Cyclic Voltammetry; Bamford, C. H.,
Compton, R. G., Eds.; Elsevier: Amsterdam, 1986; Vol. 26, pp 145-202.
(i.e., dissociated electron transfer). The observed value of R
near 0.5 is supportive of a stepwise pathway, suggesting that
(
•
-
A
exists as a discrete intermediate.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002 4273

A R T I C L E S
Stevenson et al.
Table 1. Products Arising from the Electrochemical Reduction of 1 f 5
Scheme 2
Scheme 3
Preparative-Scale Electrolyses/Product Analysis. In all
cases, preparative-scale electrolyses of 1-5 gave ring-opened
products (Table 1), consistent with the ECE mechanism depicted
in Scheme 2. Depending on the specific substrate, the elec-
trolyses were conducted either directly at a gold electrode or
indirectly with biphenyl as an electron-transfer mediator.
In the mediated approach, byproducts resulting from addition
of the ring-opened, distonic radical anions to the radical anion
Scheme 4
•
-
•-
-2
of the mediator may result (e.g., R + M f R-M ). For
t
•
•-
radical anion; 70% of the reaction of Bu + Ph-Ph involves
example, in the mediated reduction of primary alkyl halides
48
addition. Because the radical anion generated from 1-acyl-
-
•
-
(
which produce primary radicals via R-X + e f R + X ),
•-
2
,2-dimethylcyclopropane (2 ) can open either to the primary
only 5% of the reaction with biphenyl radical anion involves
addition.48 Hence, for substrates expected to produce a 1°
distonic radical anion (e.g., 1, 3, 4, and 5), biphenyl is used as
an electron-transfer mediator with the expectation that products
arising from addition to the mediator would be minor.
(
k1) or tertiary (k3) distonic radical anions shown in Scheme 3,
the reduction of 2 is carried out directly (in the absence of a
mediator) to avoid the formation of addition products. On the
basis of the observed regiochemistry of the products (Table 1),
ring opening occurs with only slight preference for the more-
substituted (stable) distonic radical anion (k3/k1 = 2.5).
Indirect Electrochemistry. Homogeneous redox catalysis is
a powerful technique for studying the chemistry of highly
reactive intermediates produced via electron transfer.
sider the reactions depicted in Scheme 4. Rather than substrate
Accordingly, for 1, after passage of 1 equiv of electrons in
the presence of biphenyl and acidification, 2-pentanone was
obtained in 34% yield, with 60% unreacted starting material.
1
A tiny quantity of addition products was detected by H NMR
49-53
Con-
1
and GC/MS but they were not individually isolated. H NMR
indicates a substituted biphenyl and a cleaved cyclopropane ring.
Although potential structures such as these are suggested, no
effort was made to discern the regiochemistry.
A, an electron-transfer mediator or catalyst M is reduced at the
electrode surface. For this to occur, the mediator must be more
easily reduced than the substrate, (i.e., E°A/A•- < E°M/M•-).
Reduction of the substrate occurs in solution (homogeneous)
via electron transfer from the reduced form of the mediator
•
-
(
M ).
(
49) Andrieux, C. P.; Hapiot, P.; Sav e´ ant, J.-M. Chem. ReV. 1990, 90, 723-
738.
In contrast to 1° radicals, addition is more competitive with
electron transfer in the reaction of 3° radicals with biphenyl
(50) Andrieux, C. P.; Blocman, C.; Dumas-Bouchiat, J. M.; M’Halla, F.; Sav e´ ant,
J. M. J. Electroanal. Chem. 1980, 113, 19-40.
(
51) Sav e´ ant, J. M.; Su, K. B. J. Electroanal. Chem. 1985, 196, 1-22.
52) Andrieux, C. P.; Sav e´ ant, J.-M. J. Electroanal. Chem. 1986, 205, 43-59.
(
(
48) Andrieux, C. P.; Gallardo, I.; Sav e´ ant, J.-M.; Su, K.-B. J. Am. Chem. Soc.
(53) Nadjo, L.; Sav e´ ant, J. M.; Su, K. B. J. Electroanal. Chem. 1985, 196, 23-
1
986, 108, 638-647.
34.
4274 J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002

Cyclopropylcarbinyl-Type Ring Openings
A R T I C L E S
as a rapid preequilibrium. Under these conditions the composite
rate constant k0k1/k-1 can be determined. Because log(K1) )
•
-
•-
-
F(E°M/M - E°A/A )/(2.303RT), k0 can be extracted if the
reduction potential of the substrate is known.
Though similar in appearance, different working curves
pertain to these two conditions, and it is critical to accurately
assess whether the kinetics are governed by the electron transfer
or chemical step. For rate-limiting electron transfer, the current
ratio is a function of γ and λ1, the latter of which is related to
the mediator concentration (eq 2). For the rate-limiting chemical
step, ip/ipd is a function of γ and λ0λ1/λ-1, and is concentration
independent at constant γ (eqs 2-4). Thus, the distinguishing
characteristic between these two rate-limiting conditions is the
o
effect of mediator concentration (C M) on ip/ipd at constant γ
and ν. The peak current ratio varies as a function of mediator
concentration only when the electron-transfer step is rate
5
0
limiting.
The reduction of 1-5 by several mediators was studied.
Because of the limited quantity of these substrates available, it
was more economical to examine the current ratio ip/ipd as a
function of sweep rate and mediator concentration at constant
excess factor γ (published working curves present ip/ipd as a
function of γ at constant sweeprate). As revealed in eqs 2-4,
Figure 3. Cyclic voltamogram of biphenyl in the absence and presence of
methyl cyclopropyl ketone (1, ν ) 100 mV/s).
o
In this manner, the reference is taken away from the electrode
at constant γ, ip/ipd is a function of log(C /ν) (when electron
M
-
and placed on the reversible 1e reduction of a compound with
transfer is rate limiting) or ip/ipd is a function of log(1/ν) when
the chemical step is rate limiting. Our approach was to obtain
the voltammograms of several mediators in the absence and
presence of 1-5. By comparing plots of [i /i vs log(1/ν) ]
a known E°. Effects of substrate addition on this reversible
electron transfer are manifested experimentally by an increase
in peak current and a loss of reversibility (if catalysis is
occurring). The key experimental observable is the current ratio
ip/ipd, where ip and ipd are the voltammetric peak currents of the
mediator in the presence and absence of the substrate, respec-
tively, at a particular value of γ (the ratio of the substrate to
p
pd
o
and [ ip/ipd vs log(C M/ν) ] obtained at different concentrations
of mediator (γ constant), any concentration dependence was
readily apparent.
For 1 (using naphthalene as a mediator), the plot of ip/ipd vs
log(1/ν) (Figure 4a) at various substrate concentrations (γ )
o
o
mediator concentrations, C A/C M). Figure 3 depicts the effect
of added cyclopropyl methyl ketone (1) on the cyclic voltam-
mogram of biphenyl.
1
.00) appears as three discrete lines. When substrate concentra-
o
tion is factored out (i.e., a plot of ip/ipd vs log(C M/ν)), the data
converge onto a single curve (Figure 4b) clearly showing ip/ipd
is concentration dependent, and thus establishing that electron
transfer is rate limiting. For 1, 2, 4, and 5, over the range of
mediators examined, ip/ipd is found to vary as a function of
mediator concentration (at constant γ), indicating that electron
transfer is rate limiting for all these substrates with all mediators.
Sav e´ ant et al. introduce the dimensionless rate constants λ1,
λ-1, and λ0 defined below (eqs 2-4, where ν is the sweep rate
-1
in V s and R, T, and F are the ideal gas constant, temperature,
and Faraday’s constant, respectively). Published working curves
are available which depict the current ratio ip/ipd (a) as a function
of log(λ1) when the electron-transfer step (k1) is rate limiting
or (b) as a function of log(λ0λ1/λ-1) when the chemical step
(All the pertinent plots are available in the Supporting Informa-
(
k0) is rate limiting.50
tion.)
The kinetics of these systems are further complicated by a
o
M
k1C
RT
F
•-
competing bimolecular reaction between M and the ring-
λ1 )
(2)
•-
ν
opened product B (i.e., addition, vide infra). Coupling
reactions between alkyl radicals and aromatic anion radicals are
known to be fast and nearly diffusion-controlled. This competi-
tion (Scheme 5) between addition to the mediator (kadd) and
the second electron reduction (ket) must be considered in the
overall reaction profile and introduces a new kinetic parameter
F, where F ) ket/(kadd + ket). The parameter F reflects the fraction
o
M
k C
-1
RT
F
λ-1
)
(3)
(4)
ν
k0
RT
ν F
λ0 )
•
-
of B which adds to the mediator. A treatment of this problem
and the appropriate theoretical working curves have been
published by Sav e´ ant.
As noted above, kinetic control may be governed by either
the homogeneous electron-transfer step (k1) or the chemical step
k0, Scheme 4). If the rate of the chemical step is faster than
back electron transfer (k0 > k-1 [M]), then the electron-transfer
step is rate limiting and k1 can be determined. If the chemical
step is slow relative to back electron transfer (k0 < k-1 [M]),
the chemical step is rate limiting with the electron-transfer step
5
1
(
As noted, published working curves dealing with addition to
the catalyst expressed ip/ipd as a function of γ.51 In these
experiments, because ip/ipd was measured at various sweep rates
at constant γ, it was necessary to derive the appropriate working
curves (42 plots of ip/ipd vs log(λ1) at γ ) 1.00 and γ ) 10 for
J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002 4275

A R T I C L E S
Stevenson et al.
function of sweep rate and mediator concentration at constant
excess factor, γ ) 1.
A plot of [ip/ipd vs log(1/ V)] and [ip/ipd vs log(C M/ V)] for
o
the reduction of 3 by naphthalene radical anion is presented in
Figure 5, parts a and b. Comparing these two, it is apparent
that there is no concentration effect on the peak current ratio,
and that the kinetics are governed by the chemical step (ring
opening) rather than electron transfer.
The composite rate constant (kobs ) K1k0) was determined
from theoretical working curves. Though similar in appearance,
different working curves pertained to the two limiting conditions.
Again, it was necessary to derive the appropriate working curves
as described earlier. A representative fit of the experimental
data is provided in Figure 5a.
Discussion
Estimates of the Reduction Potentials of 1, 2, 4, and 5
with Marcus Theory. In accordance with Marcus theory, the
relationship between k1 and the free energy of electron transfer
∆G° ) F(EM/M•- - EA/A )) is described by eq 5 . For this
•-
54
(
1
1
1
1
)
+
+
2
)
k1 kd
-λ
kd exp(-∆G°/RT)
K Z exp
(1 + ∆G° /λ)
d
(
4
RT
(5)
analysis, it is assumed kd ) 1 × 10 M-1 s-1 (the diffusion-
10
-
1
controlled rate constant in DMF), Kd ) 0.16 M , and the
1
1 -1
frequency factor Z ) 6 × 10 s . The rate constants in Table
o
2
are fit to eq 5 via nonlinear regression analysis, with E A/B
Figure 4. Mediated reduction of methyl cyclopropyl ketone (1) by
naphthalene (ν ) 0.1-1.0 V s , γ ) 1.00).
-
1
and λ as the only adjustable parameters. An excellent fit is
achieved in all cases (pertinent plots are included in the Sup-
porting Information). The results are summarized in Table 3.
Values derived for the reorganization energy (λ) of these
compounds are consistent with an electron-transfer reaction from
an aromatic hydrocarbon to an aliphatic ketone, suggesting that
there is not an additional contributor (such as bond lengthening
or bond angle deformation) to the overall reorganization energy.
The λ values summarized in Table 3 are somewhat higher than
those which we observed previously for the reduction of 6-8
by similar aromatic hydrocarbons (λ ≈ 15-20 kcal/mol) and
merit further comment.
Scheme 5
F ) 0.00 to 1.00 in 0.05 increments) via digital simulation.
These working curves were subsequently fit to a polynomial of
2
3
4
2
3
the form y ) (a + cx + ex + gx + ix )/(1 + bx + dx + fx
+
4
5
hx + jx ), where the coefficients a-j were determined for
each working curve. Via nonlinear regression, the experimental
o
data [ip/ipd vs log(C M/ν)] were fit to the polynomial form of
the working curves, y ) f(x + x′), and the adjustable parameter
x′ ) log(k1RT/F) was determined. The parameter F was
determined by the working curve that gave the best fit to the
experimental data, and k1 was determined from x′. A representa-
tive fit of the experimental data is provided in Figure 4b.
For 1, 2, 4, and 5, derived values of k1 and F for the mediators
studied are summarized in Table 2. Kinetically, the fact that
electron transfer is rate-limiting for these four substrates means
k0 > k-1[M]. A lower limit on the rate constant k0 can be
established by estimating the quantity k2[M]. The lowest
concentration of mediator used in these experiments is 0.001
M; assuming the reverse electron-transfer rate constant k-1 is
approximately diffusion-controlled, k0 is estimated to be greater
Eberson has summarized the structural and environmental
5
4
effects on λ. Spiro compounds 6-8 are highly conjugated
substrates, and the charge gained in producing the radical anion
can be accommodated over a large volume. This results in little
reorganization energy to reach the transition state. In the
aliphatic cyclopropyl substrates, charge is more highly localized
in the radical anion than the neutral ketones, and a higher solvent
reorganization energy would be required.
7
-1
10
-1 -1
7
-1
than 10 s (k-1 [M] ) (10
s ).
M
s )(0.001 M) ) 10 M
-1
Structural effects on the Reduction Potential of Aliphatic
Ketones. Within experimental error, the reduction potentials
The indirect reduction of cyclobutyl methyl ketone (3) was
examined with naphthalene and biphenyl as mediators. As
described above, the current ratio ip/ipd was examined as a
(
54) Eberson, L. Electron-transfer reactions in organic chemistry; Springer-
Verlag: Berlin, 1987; Vol. 25.
4276 J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002

Cyclopropylcarbinyl-Type Ring Openings
A R T I C L E S
Table 2. Rate Constants for Homogeneous Electron Transfer (k1) Between the Reduced Form of the Mediator (M•-) and Cyclopropyl
Ketones 1, 2, 4, and 5
1
2
4
5
E° (V)a
k
1
(M-1
s
-1 b
)
Fc
k
1
(M-1
s
-1 b
)
Fc
k
1
(M-1
s
-1 b
)
Fc
k
1
(M s-1)b
-1
Fc
mediator
naphthalene
2
3
3
3
3
3
4
2
2
3
3
3
3
4
3
3
3
3
4
2
3
3
3
3
3
-2.901
-2.937
-2.971
-2.977
-2.988
-3.027
-3.086
4.2(2) × 10
1.3(1) × 10
3.5(2) × 10
2.1(1) × 10
5.2(2) × 10
9.6(3) × 10
3.4(1) × 10
0.60
1.00
0.95
0.95
1.00
1.00
0.95
5.9(2) × 10
8.1(6) × 10
2.2(1) × 10
2.4(2) × 10
2.8(2) × 10
8.0(3) × 10
3.6(1) × 10
0.45
0.90
0.55
0.65
0.80
0.30
0.50
1.3(1) × 10
4.2(4) × 10
7.4(9) × 10
6.5(2) × 10
1.5(1) × 10
.70
1.00
1.00
0.95
1.00
4.0(1) × 10
1.5(2) × 10
2.4(1) × 10
2.6(2) × 10
5.2(4) × 10
6.7(5) × 10
0.75
1.00
1.00
1.00
1.00
1.00
3
1
,6-dimethylphenanthrene
,3-dimethylnaphthalene
biphenyl
1
2
-methoxynaphthalene
,7-dimethoxynaphthalene
o-methoxybiphenyl
a
vs 0.1 M AgNO3/Ag. b Numeral in parentheses is the error in the last digit. c (( 0.05) for all F.
Scheme 6
Scheme 7
dienones 6-8, where it was observed that increased methyl
substitution on the cyclopropyl group slightly stabilizes the
35
radical anion. For the spirocyclohexadienones, the substituent
effect is readily explained on the basis of an aromatic resonance
form, which is expected to be an important contributor to the
resonance hybrid. For radical anions generated from aliphatic
cyclopropyl ketones, the contribution of the nonbonded reso-
nance form is apparently minor (Scheme 6).
Rate Constants for Ring Opening of Radical Anions
Derived from Aliphatic Cyclopropyl and Cyclobutyl Ke-
tones. The rate-limiting step in the electron transfer-mediated
reduction of cyclopropyl ketones 1, 2, 4, and 5 is the electron-
transfer step (k1). These observations permit us to assign a lower
limit to the rate constant for ring opening of radical anions
Figure 5. Mediated reduction of cyclobutyl methyl ketone (3) by
naphthalene (DMF, GCE, 0.5 M TBAP, ν ) 0.05-1.0 V s , γ ) 1.00).
-
1
Table 3. Reduction Potentials and Reorganization Energies for 1,
2, 4, and 5 Derived from Marcus Theory
compd
E° (V)a
λ (kcal/mol)
7 -1
derived from aliphatic cyclopropyl ketones (k0) of 10 s (vide
supra). There is good reason to suspect that ring opening may
be considerably faster than the cyclopropylcarbinyl neutral
radical.
1
2
4
5
-3.215 ((0.068)
-3.196 ((0.067)
-3.205 ((0.106)
-3.188 ((0.145)
28 ( 4
30 ( 4
26 ( 6
31 ( 7
•
-
For 2 , ring opening occurs with only slight preference for
the more-substituted (stable) distonic radical anion (Scheme 3).
On the basis of the yield of products observed in the preparative-
scale electrolysis, k3/k1 = 2.5. (Dauben and Wolf obtained
similar results (k3/k1 = 3.1) in 1970 with Li/NH3 as the reducing
of aliphatic ketones 1, 2, 4, and 5 are identical. Consistent with
this conclusion is the fact that rate constants for electron transfer
•-
between these substrates and M are also nearly identical (Table
). These results suggest that (a) alkyl substituents on the
2
38
agent. ) For comparison, the dimethyl-substituted neutral radical
cyclopropyl group do not perturb E° and (b) that the difference
in E° between methyl-, cyclopropyl-, and cyclobutyl-sub-
stituted aliphatic ketones is (within experimental error) undetect-
able.
This conclusion offers an interesting comparison to the
substituent effect on the reduction potential of spirocyclohexa-
9
7
leads to 3° and 1° radicals 10 and 11 in a 6.7:1 ratio (Scheme
). The low selectivity of both these processes suggests an
55
early transition state for ring opening. The extremely low
•-
selectivity observed for ring opening of 2 compared to neutral
(55) Beckwith, A. L. J.; Bowry, V. W. J. Org. Chem. 1989, 54, 2681-2688.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002 4277

A R T I C L E S
Stevenson et al.
Table 4. Composite Rate Constant for Ring Opening of
Scheme 8
Cyclobutyl Methyl Ketone
mediator (M)
EoM/M (V)a
•−
Fb
k
obs (s-1
)
o
k (s )
-1 c
-
1
4
4
naphthalene
biphenyl
-2.901
-2.977
0.50 2.1((0.1) × 10
2.5((1.4) × 10
2.5((1.4) × 10
0
0.90 4.0 (( 0.2) × 10
a
vs 0.1 M AgNO3/Ag. b (0.05. c Calculated assuming E°3/3•- ) -3.201
(
0.012 V vs 0.1 M AgNO3/Ag.
radical 9 suggests the former may be more reactive. If this
reactivity/selectivity argument is valid, the rate constant for ring
opening of these radical anions is expected to be of similar
magnitude, perhaps even greater, than the structurally related
neutral free radicals. Convincing evidence that this is the case
comes from studying a slower rearrangement (e.g., a cyclobutyl
ketyl anion).
After observing that for radical anions generated from
aliphatic cyclopropyl ketones ring opening was apparently so
rapid that electron transfer was rate limiting (vide supra), we
decided to study cyclobutyl ketone 3 in the hope that ring
opening would be slower and, hopefully, rate limiting. For
neutral free radicals, the cyclobutylcarbinyl radical ring opens
respectively. Thus, the calculations support the proposition that
the cyclopropyl group does not significantly stabilize these
radical anions. With the assumption that the reduction potential
of 3 is -3.201 ( 0.012 V (the average of compounds 1, 2, 4,
•-
5
orders of magnitude slower than the cyclopropyl carbinyl
and 5), the derived rate constant for ring opening of 3 is 2.5
56-58
4
-1
radical.
(The proposition that three-membered-ring cleavage
× 10 s .
in these radical anions is substantially faster than four-
membered-ring cleavage is confirmed, qualitatively at least, by
the preparative-scale electrolysis of 4sonly products arising
from cleavage of the cyclopropane ring are detected, Table 1).
For the mediated reduction of 3, ring opening was indeed
the rate-limiting step, and composite rate constants were
determined with use of naphthalene and biphenyl as electron-
transfer mediators, kobs ) (k1/k-1)k0 ) K1k0 (Table 4). The
equilibrium constant for electron transfer was related to the
difference in reduction potential of the mediator and the substrate
Cyclopropane Ring Opening of Radical Anions: Ther-
mochemical Considerations. The free energy of ring opening
of several radical anions pertinent to this study were determined
by using eq 6, which is derived from the thermochemical cycle
depicted in Scheme 8. The quantities needed to solve for ∆G°
o
•-
o
+
•
are (1) E A/A , the reduction potential of ketone 6, (2) E H /H ,
+
the reduction potential of H (-2.38 V vs NHE DMSOs
corrected to 0.1 M Ag /Ag reference, and assumed to be
unchanged in DMF), (3) the pKa of 7 (assumed to be the
same as the corresponding ketone in DMSO), and (4) ∆HH•,add
(the enthalpy of H addition to ketone 6. ∆H
lated (B3LYP/6-31G ) based upon the isodesmotic reaction in
+
5
9
60
o
•-
•
(
log(K1) ) -F(E°M/M•- - E A/A )/(2.303RT)); unfortunately,
H•,add
was calcu-
3
6
*
the latter value was unknown. In our Communication, we
assumed that the reduction potentials of 3 and 1 were similar.
We can now address this assumption more confidently. In
analogy to 1 and 2, electron transfer is rate-limiting in the
mediated reduction of 4 and 5. As a consequence, rate constants
for electron transfer with several mediators are determined
eq 7 and referenced to the experimental value of ∆H° for the
•
addition of H to cyclopropane (derived from experimental heats
of formation).
(Table 2), and reduction potentials are estimated by using
Marcus theory (Table 3). Compared to 1, introduction of a
second cyclopropyl group (5) or a cyclobutyl group (4) has no
detectable effect on the kinetics or thermodynamics (i.e., E°)
of the electron transfer. With any given mediator, the rate
constants for electron transfer for all of these compounds are
identical, as are the derived reduction potentials. Thus, a
cyclopropyl (or cyclobutyl) group has no measurable effect on
the stability of the radical anion.
Of course, the best comparison would be the reduction
potential of 1 vs that of acetone. While this cannot be achieved
experimentally, this issue can be addressed computationally.
Hartree-Fock calculations reveal (a) that the energies of the
LUMO of 1 and acetone are nearly identical (acetone, 1.75 eV;
This procedure is attractive because the calculations are based
partly on experimental solution-phase measurements for the
charged species, and they should adequately account for the
effect of solvent and electrolyte. The results are summarized in
Table 5.
Ring Opening: Cyclopropyl vs Cyclobutyl. The results
summarized in Table 5 indicate that the driving force for ring
•
-
opening of the cyclopropyl group of 1 is nearly the same as
•-
that for the ring opening of the cyclobutyl group of 3 .
•
-
1
, 1.74 eV). Likewise, the ionization potentials (HOMO energy)
Kinetically, however, ring opening of 1 is at least 3 orders of
•
-
•-
•-
of 1 and Me2CdO are very close, 1.60 and 1.56 eV,
magnitude faster than that of 3 . A similar situation pertains
to ring opening of the neutral free radicals: Ring opening of
the cyclopropylcarbinyl radical is exothermic by 2-4 kcal/mol
(
(
(
56) Beckwith, A. L. J.; Moad, G. J. Chem. Soc. Perkin Trans. 2 1980, 1083-
092.
57) Ingold, K. U.; Maillard, B.; Walton, J. C. J. Chem. Soc., Perkin Trans. 2
981, 970-974.
58) Walton, J. C. J. Chem. Soc., Perkin Trans. 2 1989, 173-177.
1
1
(59) Parker, V. D. J. Am. Chem. Soc. 1992, 114, 7458-7462.
(60) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463.
4
278 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002
9

Cyclopropylcarbinyl-Type Ring Openings
A R T I C L E S
Table 5. Kinetic and Thermodynamic Data for Rearrangement of Structurally similar Neutral Free Radicals and Radical Anions
a
Calculated (B3LYP/6-31G ) this work. Literature (th ) theory, B3LYP/6-31G , ex ) experiment, ad ) additivity rules). Reference 66. d See ref 67
*
b
*
c
e
f
g
h
i
for a summary and discussion. References 68-70. Reference 55. Reference 58. See also references 56 and 57. ∆G°, estimated from forward and
reverse rate constants in ref 71. References 71 and 72. References 32 and 33. Reference 73.
Unpublished results at VPI&SU).
j
k
l
73 m
Reference 34. n Reference 74. o M. Chahma, J. Tanko
(
8
1
mol measured by EPR);64 for cyclobutyl, the difference is only
1.1 kcal/mol.
and occurs with a rate constant of 1.0 × 10 s . The driving
force for ring opening of the cyclobutylcarbinyl is similar, but
the rate constant is 5 orders of magnitude lower. Thus, it would
be incorrect to conclude that the sluggish rates of ring opening
of the cyclobutyl species are attributable to differences in the
relief of ring strain. Indeed, the strain energiesy of cyclopropane
Scheme 9
6
1
and cyclobutane are nearly identical, 27.5 vs 26.5 kcal/mol,
respectively.
The calculated barriers and transition state geometries for ring
opening of the cyclopropylcarbinyl and cyclobutylcarbinyl
radicals obtained in this study compare favorably to previously
reported values. For cyclopropylcarbinyl and cyclobutylcarbinyl
ring opening, the calculated barriers are 8.6 and 14.2 kcal/mol,
respectively. Furthermore, as noted by Dolbier, bond breaking
is considerably more advanced in the transition state for ring
opening of the cyclobutylcarbinyl radical, where the length of
the rupturing C-C bond (2.05 Å) is significantly greater than
that for the cyclopropylcarbinyl radical (1.91 Å). This difference
is also manifested in the calculated spin densities: There is
considerably greater spin density at the developing radical center
in the transition state of the cyclobutylcarbinyl radical (0.520)
compared to the cyclopropylcarbinyl (0.469).
The rapid rate of ring opening associated with cyclopropyl-
carbinyl radicals can be attributed to the unique conjugative
properties of the cyclopropyl vs cyclobutyl groups and the
resulting differences in the extent of spin delocalization in the
radical anions and transition states for ring opening.
6
2
In the case of the neutral radicals, EPR studies reveal that
63-65
57
both the cyclopropylcarbinyl
and cyclobutylcarbinyl
systems adopt the bisected conformation vs perpendicular
conformation. The reactive conformation for ring opening is
the bisected conformation, because the overlap between the
cycloalkyl HOMO and the radical center is maximal (Scheme
9). DFT calculations show that for these radicals, a greater
amount of spin density is transferred to the cyclopropyl group
compared to the cyclobutyl group. (The spin densities at the
CH2 are 0.860 and 0.905 for the cyclopropylcarbinyl and
cyclobutylcarbinyl radicals, respectively.) This difference in spin
delocalization is reflected in the energetics: For cyclopropyl,
the bisected conformation is favored over the perpendicular by
In the case of the three- vs four-membered ring radical anions
•-
(
1•- vs 3 ), the situation though similar, is not identical.
Consistent with the observed reduction potentials, DFT calcula-
tions provide no indication that the cyclopropyl group signifi-
•
-
cantly stabilizes ketyl radical anions. For 1 , the bisected and
perpendicular conformations are nearly degenerate (e0.2 kcal/
3.3 kcal/mol (a value similar to the rotational barrier of 2.7 kcal/
(
61) Dewar, M. J. S. J. Am. Chem. Soc. 1984, 106, 669-682.
62) Tidwell, T. T. ConjugatiVe and substitutent properties of the cyclopropyl
group; Rappoport, Z., Ed.; John Wiley & Sons: Ltd.: New York, 1987;
pp 565-632.
•-
mol difference in energy). For 3 , the perpendicular conforma-
(
tion is lower in energy by about 1.0 kcal/mol. In the bisected
•-
conformation, the spin densities at the CdO are nearly
(63) Kochi, J. K.; Krusic, P. J.; Eaton, D. R. J. Am. Chem. Soc. 1969, 91, 1879-
•
-
•-
1
881.
identical for 1 and 3 , 0.703 and 0.698, respectively.
As expected, the calculated barrier for ring opening of 3
(7.5 kcal/mol) is higher than that of 1 (3.3 kcal/mol), though
(
64) Walton, J. C. Magn. Reson. Chem. 1987, 25, 998-1000.
•-
(
65) Bauld, N. L.; McDermed, J. D.; Hudson, C. E.; Rim, Y. S.; Zoeller, J., Jr.;
Gordon, R. D.; Hyde, J. S. J. Am. Chem. Soc. 1969, 91, 6666-6676.
•
-
J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002 4279

A R T I C L E S
Stevenson et al.
it must be stressed that these values pertain to a gas-phase
calculation and the experimental results are from a study in
solution. As is the case for neutral radicals, the extent of bond
cleavage is greater in the transition state for ring opening of
the four- vs three-membered ring: The lengths of the rupturing
C-C bond and spin densities (at the developing radical center)
Scheme 10
•
-
are 2.04 Å and 0.398 for 3 , compared to 1.83 Å and 0.333
•
-
for 1 . The bond length differences and spin densities suggest
a more delocalized, resonance-stabilized transition state for
rupture of the three- vs four-membered ring.
Consequently, we suggest that the reason cleavage of a
cyclopropyl group is so much faster than that of a cyclobutyl
group (in both radicals and radical anions) is that for the
cyclopropyl systems (a) the more reactive bisected conformation
is preferred and (b) because the transition state for rupture of
the three-membered ring is stabilized to a greater extent (via
spin delocalization).
Scheme 11
Comparison of Neutral Radicals and Structurally Related
Radical Anions. Table 5 summarizes the pertinent kinetic and
thermodynamic data for ring opening of several neutral free
radicals and radical anions. Our new results dispel the prior
belief that ring opening of the radical anions was considerably
slower because of the thermodynamic stability of ketyl radicals
compared to structurally related alkyl radicals.
The data in Table 5 reveal that as a result of R-phenyl
substitution, ring opening of both the neutral radicals and radical
anions becomes less favorable kinetically and thermodynami-
cally. Placement of a radical-stabilizing substituent on the
cyclopropane ring has the opposite effect: Ring opening
becomes more favorable thermodynamically and kinetically.
These observations are readily explicable on the basis of
resonance: R-substitution stabilizes the ring-closed from; sub-
stituents on the cyclopropyl ring stabilize the ring-opened form
negative charge can be delocalized into the aromatic ring. Upon
ring opening, this stabilization is lost because the charge
becomes more localized. The numbers in Scheme 11 are the
Mulliken charges associated with various regions of the
molecule (obtained from Hartree-Fock calculations).
(or more appropriately, the transition state leading to the ring-
The lower rate of ring opening of radical anions derived from
aromatic ketones is not attributable to the intrinsic stability of
ketyl radicals (in general). Rather, it is the fact that aromatic
ketyls are especially stable, and the important consideration is
the delocalization of negative charge (rather than spin) into the
aromatic ring. Our results suggest that aliphatic ketyls are at
least (and likely more) as reactive as alkyl radicals in â-scission-
type reactions (vide infra). At first glance, this conclusion may
seem surprising because radicals are generally stabilized by
electron-donating groups (e.g., based upon the C-H bond
strengths of methane vs methanol, an OH group stabilizes a
radical to the tune of about 7 kcal/mol). Why is it that a
opened form, Scheme 10).
For radical anions generated from aromatic ketones, ring
opening of the neutral radical is considerably faster than that
-
of the radical anions (i.e., replacement of H by O makes ring
opening less favorable and diminishes the rate.)
However, for radical anions derived from aliphatic ketones,
the opposite scenario pertains. The energetics of ring opening
for the neutral radicals and radical anions are quite similar (Table
5). Where numbers are available (e.g., the cyclobutyl system),
the rate constant for ring opening of the radical anion is
comparable to, or slightly faster than, that of the structurally
related neutral free radical. Unquestionably, the difference
between radical anions generated from aliphatic vs aromatic
ketones is the ability of the aromatic ring to stabilize the negative
charge via resonance (Scheme 11). In the ring-closed form, the
-
deprotonated OH group (i.e., O ), which is even more electron
donating, not more able to stabilize radicals?
Indeed, considering the isodesmotic reaction depicted in eq
o
8
(∆E obtained from B3LYP/6-311+G* calculations on the
pertinent species), this supposition appears reasonable: A radical
(
(
(
66) Halgren, T. A.; Roberts, J. D.; Horner, J. H.; Martinez, F. N.; Tronche, C.;
stabilization energy (RSE) of 33 kcal/mol can be attributed to
Newcomb, M. J. Am. Chem. Soc. 2000, 122, 2988-2994.
67) Smith, D. M.; Nicolaides, A.; Golding, B. T.; Radom, L. J. Am. Chem.
Soc. 1998, 120, 10223-10233.
68) Effio, A.; Griller, D.; Ingold, K. U.; Beckwith, A. L. J.; Serelis, A. K. J.
Am. Chem. Soc. 1980, 102, 1734-1736.
-
the O group relative to hydrogen.
(69) Mathew, L.; Warkentin, J. J. Am. Chem. Soc. 1986, 108, 7981-7984.
(70) Newcomb, M.; Glenn, A. G. J. Am. Chem. Soc. 1989, 111, 275-277.
(71) Beckwith, A. L. J.; Bowry, V. W. J. Am. Chem. Soc. 1994, 116, 2710-
2
716.
(
(
(
72) Bowry, V. W.; Lusztyk, J.; Ingold, K. U. J. Chem. Soc., Chem. Commun.
990, 923-925.
73) Hollis, R.; Hughes, L.; Bowry, V. W.; Ingold, K. U. J. Org. Chem. 1992,
7, 4284-4287.
74) Tanner, D. D.; Chen, J. J.; Luelo, C.; Peters, P. M. J. Am. Chem. Soc.
992, 114, 713-717.
However, “stabilization” depends on the reference reaction
used for comparison. Consider eq 9, which depicts an isodes-
motic reaction to evaluate and compare the ability of ethyl
radical vs acetaldehyde radical anion to undergo â-scission
1
5
1
4280 J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002

Cyclopropylcarbinyl-Type Ring Openings
A R T I C L E S
•
(
O
through loss of H ). In this case, the results suggest that the
for the structurally related neutral free radical (and possibly more
favored thermodynamically). The high reactivity of aliphatic
ketyl anions toward â-scission arises because the resonance
-
group actually destabilizes a radical center (relative to
hydrogen) by nearly 10 kcal/mol.
-
stabilization between an O group and a double bond (i.e., the
-
enolate in the products) is greater than that of an O group
•
-
with the radical center (i.e., CdO in the reactants).
Although this work has focused specifically on radical anion
rearrangements, there is little doubt that many of the same
considerations pertain to rearrangements of radical cations. The
chemistry of radical ions is perhaps more complex than is
generally appreciated; however, it is important to emphasize
that they do not appear to defy established principles of organic
reactivity. The complexity arises from the fact that these species
are both radicals and ions, and their chemistry is reminiscent
of both types of reactive intermediates. Indeed, as more systems
are studied and patterns regarding the role played by stabilization
The difference in these two analyses is that eq 8 considers
-
only the interaction of O with a radical center (vs a methyl
-
group), where eq 9 compares the interaction of O with a radical
center vs a carbon-carbon double bond. The implication is that
-
O stabilizes the CdC to a greater extent than a radical center,
and for that reason, â-scission is thermodynamically more
favorable for aliphatic ketyl anions compared to alkyl radicals.
(
An equivalent, and we suspect indistinguishable, argument is
-
that it is actually O which is stabilized more by a CdC than
an adjacent radical center.) In either case, the final result is
identica The reason â-scission is more favorable for aliphatic
ketyl anions (compared to alkyl radicals) is that introduction of
(or destabilization) of charge and spin are revealed, it is likely
that an excellent understanding of the chemistry of this unique
class of reactive intermediates will be achieved.
-
the O group leads to more resonance stabilization in the
products than in the reactants.
Acknowledgment. Financial support from the National Sci-
ence Foundation (CHE-9732490) is gratefully acknowledged.
Closing Remarks
In summary, stabilization of both charge and spin are
important factors pertaining to substituent effects on the rates
of radical anion rearrangements. Ring opening of radical anions
generated from aromatic ketones is slower than that of the
corresponding neutral free radicals because the aromatic ring
stabilizes the negative charge in the ring-closed, but not the
ring-opened form. In contrast, for radical anions generated from
aliphatic ketones, ring opening is significantly faster than that
Supporting Information Available: Plots of ip/ipd vs log(1/
ν) and vs log(CM /ν) for all substrate/mediator combinations
studied and results of MO calculations pertaining to cyclopro-
pylcarbinyl and cyclobutylcarbinyl radicals and related radical
anions (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
o
JA0041831
J. AM. CHEM. SOC.
9
VOL. 124, NO. 16, 2002 4281