γ-dissociations of 3-alkylbicyclo[1.1.1]pent-1-yl radicals into [1.1.1]propellane and alkyl radicals: Verification of a theoretical prediction

Article abstract of DOI:10.1039/a808744j

The γ-dissociations of 3-substituted bicyclo[1.1.1]pent-1-yl radicals and cyclobutyl radicals were investigated by ab initio SCF MO (HF, MP2, MP3 and MP4/6-31G*) and density functional methods (B3LYP/6-31G*). The transition states were found to resemble the product alkyl radical and [1.1.1]propellane or bicyclo[1.1.0]butane. Calculated endothermicities and energy barriers were comparatively low for loss of the t-butyl radical from the 3-t-butylbicyclo[1.1.1]pent-1-yl radical, which suggested that this dissociation would be significant under laboratory conditions. The dissociation was verified experimentally by means of the reaction of 1-iodo- 3-t-butylbicyclo[1.1.1]pentane with tributyltin hydride. Arrhenius parameters for this dissociation were determined by end product analysis. The SCF MO and density functional calculations resulted in much higher endothermicities and energy barriers for γ-dissociations of cyclobutyl radicals, hence neither the formation of bicyclo[1.1.0]butane nor alkyl radical addition to this bicyclic compound was predicted to be important.

Full text of DOI:10.1039/a808744j

View Article Online / Journal Homepage / Table of Contents for this issue
c-Dissociations of 3-alkylbicyclo[1.1.1]pent-1-yl radicals into
[1.1.1]propellane and alkyl radicals: veriÐcation of a theoretical
prediction
John C. Walton*a and Lewis Whiteheadb
a School of Chemistry, University of St. Andrews, St. Andrews, Fife, UK KY 16 9ST
b Department of Chemistry, University of Southampton, Southampton, UK SO17 1BJ
Received 9th November 1998, Accepted 25th January 1999
The c-dissociations of 3-substituted bicyclo[1.1.1]pent-1-yl radicals and cyclobutyl radicals were investigated
by ab initio SCF MO (HF, MP2, MP3 and MP4/6È31G*) and density functional methods (B3LYP/6È31G*).
The transition states were found to resemble the product alkyl radical and [1.1.1]propellane or
bicyclo[1.1.0]butane. Calculated endothermicities and energy barriers were comparatively low for loss of the
t-butyl radical from the 3-t-butylbicyclo[1.1.1]pent-1-yl radical, which suggested that this dissociation would
be signiÐcant under laboratory conditions. The dissociation was veriÐed experimentally by means of the
reaction of 1-iodo-3-t-butylbicyclo[1.1.1]pentane with tributyltin hydride. Arrhenius parameters for this
dissociation were determined by end product analysis. The SCF MO and density functional calculations
resulted in much higher endothermicities and energy barriers for c-dissociations of cyclobutyl radicals, hence
neither the formation of bicyclo[1.1.0]butane nor alkyl radical addition to this bicyclic compound was
predicted to be important.
Small-ring bridgehead radicals have unique geometries and
electronic conÐgurations which signal unusual chemistry and
enhanced reactivity. Bridgehead radicals are necessarily pyr-
amidal and hence they are expected to be more reactive gener-
ally than acyclic t-alkyl radicals, such as the essentially planar
t-butyl,1 because pyramidalization reduces steric shielding of
the radical centre and reduces radical stabilisation. Pyrami-
dalization also increases the s-character of the SOMO and
hence increases nucleophilicity.2 The bicyclo[1.1.1]pent-1-yl
radical (1a) is the most extreme example of ““bridgehead
characterÏÏ known to date.3 Consequently, this radical has
attracted a lot of interest in the last few years as experimental
methods for studying it have been developed. Recent work has
shown that radical 1 is active in atom abstraction4 and addi-
tion reactions,5h7 and it has even found practical use as a
chain carrier in synthetic sequences designed to make strained
[1.1.1]propellane 2 and a hydrogen atom (Scheme 1). This
diagnosis was endorsed by our ab initio MO calculations (6È
31G* basis) which gave 27.0 kcal mol~1 as the dissociation
energy for this process.17 These theoretical estimates implied
that unimolecular loss of a C-centred radical, more thermody-
namically stabilised than the H-atom, e.g., methyl or, better
still, t-butyl, should occur from 1 at comparatively low tem-
peratures. The viability of this novel c-scission process has
been tested by ab initio MO theory for 1a–c, and experimen-
tally for 1c, and this paper reports the main Ðndings.
[1.1.1]Propellane is a unique molecule with an inverted
structure at the two bridgehead carbon atoms.18 It seemed
possible, therefore, that the occurrence of this type of c-
dissociation would be contingent on the co-production of 2
with an alkyl radical and would not be general for related
classes of radicals. To assess this prospect, we also examined a
closely analogous process, i.e., the c-dissociation of 3-
substituted cyclobutyl radicals 4, which might be expected to
generate an alkyl radical together with bicyclobutane 5, the
next lower homologue of propellane 2 (Scheme 2).
polycycles. Bicyclo[1.1.1]pent-1-yls
1 have an additional
unique feature in that the cross cage distance is very short (ca.
1.80 A)8,9 so that 3-substituents have a powerful inÑuence on
chemical properties. Attempts to provide quantitative kinetic,
thermodynamic and structural data for strained bridgehead
radicals have been meagre owing to the formidable experi-
mental difficulties.2,10,11 However, recent kinetic studies
established that bicyclo[1.1.1]pent-1-yl radicals add to alkenes
and abstract hydrogen considerably more rapidly than do
t-butyl radicals.12 Furthermore, 1 was found to react with
three-coordinate phosphorus much more rapidly than
primary alkyl radicals.13 Because of the increasing technical
importance of radicals of type 1, exploration of alternative
unimolecular options is imperative.
Experimental
EI mass spectra were obtained with 70 eV electron impact
ionisation on a VG (Winsford, Cheshire, UK) Autospec
spectrometer. GC-MS analyses were run on a Finnigan
(Sunnyvale, CA, USA) Incos 50 quadrupole instrument
Bicyclo[1.1.1]pent-1-yl radicals are usually made by addi-
tion of some initial radical to [1.1.1]propellane (2). This addi-
tion is known to be fast for carbon-centred radicals and for a
range of heteroatom-centred radicals.14,15 A theoretical study
by Feller and Davidson16 implied that comparatively little
expenditure of energy was entailed in the reverse reaction, i.e.
the c-dissociation of the C3ÈH bond of 1a to produce
Scheme 1
Phys. Chem. Chem. Phys., 1999, 1, 1399È1404
1399

View Article Online
substituent species and C3 of the cage, then allowing the rest
of the geometry to optimise. The location of the transition
states at the HartreeÈFock energy maxima was conÐrmed by
performing frequency calculations to establish one imaginary
frequency. Normal harmonic vibrational frequencies and zero
point vibrational energies (ZPVEs) were evaluated for each
species at the HF, MP2 and B3LYP levels of calculation. The
MP2-ZPVEs and thermal vibrational corrections were used in
deriving energy changes for the MP3 and MP4 data. The
transition state (TS) geometries were subsequently optimised
at the MP2 level, followed by single point MP3 and MP4
calculations. In the B3LYP calculations, TS geometries were
partly optimised starting from the MP2 data. Initial computed
SS2T values were [0.75 and \0.76 for all radicals from HF
and MP2 calculations, showing that contamination from
higher spin states was minimal. However, the initial SS2T
values of the transition states were signiÐcantly in excess of
this, partitcularly for the HF and MP2 methods, for which
values as high as 0.9 were found for TS 3c. This indicated
signiÐcant contamination of the wavefunctions with quartet
and higher spin states. However, spin projection was
employed to remove this and spin projected energies were uti-
lised for all the open shell systems. Overall energy changes
Scheme 2
coupled to a Hewlett-Packard (Avondale, PA, USA) HP 5890
chromatograph Ðtted with a 25 m HP 17 capillary column
(50% phenyl methyl silicone). For analytical GC a Philips
(Eindhoven, The Netherlands) PU 4500 chromatograph, Ðtted
with a 5 m column packed with 10% OV-101 on Chromosorb
WAW operated at 80 ¡C, was used. 1,1-Dibromo-2,2-bis(chlo-
romethyl)cyclopropane was prepared by the method of Della
and Taylor19 and [1.1.1]propellane was synthesised as
described in the literature.5,6
1-Iodo-3-t-butylbicyclo[1.1.1]pentane 7c
[1.1.1]Propellane (1.06 g, 0.016 mol) and 2-iodo-2-methyl-
propane (3.0 g, 0.016 mol) in diethyl ether (100 cm3) were irra-
diated for 6 h in a water-cooled, quartz photochemical reactor
at ambient temperature with light from a 150 W Hg lamp.
The solvent was evaporated and the product was distilled to
yield iodide 7c as pale yellow crystals (1.04 g, 26%), mp
105 ¡C; d (300 MHz, CDCl ) 0.81 (9H, s, 3 ] CH ), 2.09 (6H,
*H
and activation energy di†erences *E and *E for the
298
d
a
forward dissociation and reverse addition reactions, respec-
tively, were derived from the calculated optimum energies,
zero point vibrational energies and thermal vibrational cor-
rections (H [ H ) of the individual species.
H
3
3
s, 3 ] CH ); d (75 MHz) 26.2, 32.11, 46.7, 49.72, 56.48; m/z
2
C
250 (M`, 2%), 194 (8), 193 (14), 192 (9), 128 (18), 127 (5), 123
(18), 115 (19), 108 (22), 93 (53), 83 (93), 67 (38), 57 (100), 56 (61).
298
0
The main calculated geometric parameters of the bridge-
head radicals, the TSs for their dissociation and the product
species are listed in Table 1.
Reaction of 7c with tributyltin hydride
1-Iodo-3-t-butylbicyclo[1.1.1]pentane (20 mg, 0.8 mmol) and
tributyltin hydride (30 ll, 1.6 mmol) in C D (0.5 cm3) in an
NMR tube were photolysed at 80 ¡C for 4 h with light from a
The MP2 bond lengths and angles of the bridgehead mol-
ecules 2 and 5 are essentially identical with those reported
previously9,26 and in close agreement with the B3LYP data;
congruence with the experimental structures (Table 1) is also
impressive. The MP2 calculated geometries of radicals 1a–c
and 4a,b are generally close to the B3LYP calculated struc-
tures, the main di†erence being that the latter method predicts
slightly longer CÈC and CÈH bonds. As is normally found, the
radical structures displayed a shortening of their C2ÈC1~
bonds and a lengthening of their C3ÈC2 bonds in comparison
with the hydrocarbons from which they were derived. The
cross-cage distances between the bridgeheads [R(C1ÈC3)] in
radicals 1a–c were calculated to be short, as expected, and all
three methods indicated a slight lengthening of this bond as
Me and t-Bu replaced H as the 3-substituent. Noteworthy fea-
tures of TS geometries for dissociation of radicals 1a–c were
Ðrst, the comparatively long ([2 A) distances between the 3-
substituent and the cage. Second, in all three cases, the
geometry of the cage part of the molecule had closed up dra-
matically and was approaching that of [1.1.1]propellane.
Simultaneously, the geometries of the Me and t-Bu substit-
uents closely resembled those of the free Me~ and t-Bu~ rad-
icals. We can conclude that the TSs for the dissociations are
““lateÏÏ on the reaction coordinate and that the TSs of the
matching reverse addition reactions are ““earlyÏÏ. Analogous
structural behaviour was calculated for the dissociations of
the cyclobutyl radicals 4a,b (see Table 1).
6
6
150 W Hg arc. The 1H NMR spectrum of the product mixture
showed the presence of [1.1.1]propellane (2) (d 2.02)
H
together with 3-t-butylbicyclo[1.1.1]pentane (8), the spectrum
of the latter [d 0.82 (9H, s), 1.59 (6H, s), 2.48 (1H, s)] being in
H
good agreement with the literature.20 GC-MS analysis veriÐed
the presence of isobutane (conÐrmed by retention time com-
parisons with authentic material) and 8 [m/z 109 (11%), 91 (7)
84 (73) 83 (100), 82 (22), 67 (65), 56 (32), 55 (95), 42 (23), 41 (72),
40 (43)].
Kinetics of dissociation of radical 1c
Photolytic reactions were carried out as described above with
7c and excess tributyltin hydride (0.216 mol dm~3) in C D or
6
6
t-butylbenzene as solvent. The following product ratios,
[8]/[t-BuOH], were determined by GC analysis: 323 K, 781;
338 K, 725; 353 K, 249; 368 K, 42.7; 392 K, 105; and 423 K,
25.7.
Results and discussion
Ab initio SCF and density functional calculations
Ab initio SCF computations were carried out using the Gauss-
ian 94 suite of programs21 (revisions D.4 and E.1) employing a
6È31G* basis set and electron correlation up to MP4. Struc-
tures were fully optimised at the HF and MP2-fc levels, start-
The calculated energies (H ) of the radicals, TSs and disso-
0
ciation products are listed in Table 2. The HF, MP2 and
ing in each case from a C point group geometry and using
the spin-restricted (RHF) and spin-unrestricted (UHF)
B3LYP data are for fully optimised geometries but the MP3
and MP4 data refer to single point calculations obtained with
the MP2 geometry. Quantitative calculations of thermochemi-
cal parameters are difficult to achieve and require large basis
sets. However, with MP4/6È31G* and similar basis sets, good
agreement with experiment has been achieved, provided
ZPVEs and thermal vibrational enthalpies are taken into
account.27,28 We calculated ZPVEs and thermal vibrational
enthalpies at the HF, MP2 and B3LYP levels. Our energies
have not been scaled, Ðrst, because this introduces an empiri-
3v
methods for closed shell and open shell species, respectively;
then single point calculations were performed at MP3 and
MP4 levels. Selected species were also examined using density
functional theory with no geometry constraints (B3LYP/6È
31G*).22 The dissociations and reverse radical addition reac-
tions were explored by means of reaction coordinate scans
performed at the HartreeÈFock 6È31G* level of theory. These
scans involved incrementing the bond distance between the
1400
Phys. Chem. Chem. Phys., 1999, 1, 1399È1404

View Article Online
Table 1 Calculated structures of radicals and transition states
Species
Unita
HF/6È31G*
MP2/6È31G*
B3LYP/6È31G*
Obs.b
Bicyclo[1.1.1]pentyl
1a
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈC3)
R(C3ÈH)
1.535
1.814
1.546
1.086
1.082
89.3
1.541
1.800
1.548
1.101
1.093
89.7
1.549
1.806
1.555
1.102
1.093
89.8
R(C2ÈH)
n(C2C1C4)
n(C1C2C3)
n(C2C3H)
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈC3)
R(C3ÈH)
72.1
71.3
71.1
126.4
1.508
1.587
1.508
2.090
1.077
94.9
125.8
1.508
1.540
1.508
2.090
1.087
96.3
125.7
1.521
1.598
1.521
2.090
1.088
94.9
TS 3a
R(C2ÈH)
n(C2C1C2@)
n(C1C2C3)
n(C2C3H)
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈC3)
R(C3ÈC6)
R(C2ÈH)
63.5
61.4
63.4
121.8
1.532
1.823
1.553
1.515
1.083
1.086
89.4
120.7
1.538
1.809
1.554
1.516
1.093
1.094
89.8
121.7
1.544
1.818
1.563
1.524
1.094
1.096
89.9
3-Methylbicyclo[1.1.1]pentyl
1b
R(C6ÈH)
n(C2C1C4)
n(C1C2C3)
n(C2C3C6)
n(HC6H)
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈC3)
R(C3ÈC6)
R(C2ÈH)
71.6
71.6
71.6
126.7
108.2
1.515
1.640
1.515
2.342
1.078
1.075
93.5
126.2
108.5
1.509
1.556
1.509
2.342
1.087
1.081
95.8
126.3
108.3
1.525
1.636
1.525
2.342
1.089
1.085
93.9
TS 3b
R(C6ÈH)
n(C2C1C4)
n(C1C2C3)
n(C2C3C6)
n(HC6H)
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈC3)
R(C3ÈC6)
R(C2ÈH)
65.6
62.1
64.8
122.8
117.0
1.530
1.824
1.557
1.535
1.083
1.535
89.7
121.0
118.3
1.537
1.806
1.558
1.532
1.093
1.529
90.1
122.4
117.2
1.543
1.816
1.566
1.550
1.093
1.539
90.2
3-t-Butylbicyclo[1.1.1]pentyl
1c
R(C6ÈC7)
n(C2C1C4)
n(C1C2C3)
n(C2C3C6)
n(C7@C6C7)
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈC3)
R(C3ÈC6)
R(C2ÈH)
R(C6ÈC7)
n(C2C1C4)
n(C1C2C3)
n(C2C3C6)
n(C7@C6C7)
R(C1ÈC2)
R(C1ÈC3)
R(C2ÈH)
72.4
71.4
71.5
126.9
109.4
1.517
1.650
1.517
2.302
1.078
1.508
93.3
126.3
109.8
1.511
1.566
1.511
2.302
1.087
1.495
95.6
126.3
109.7
1.528
1.647
1.528
2.302
1.089
1.504
93.7
TS 3c
65.9
62.4
65.2
123.0
115.7
1.503
1.544
1.076
90.6
121.2
117.1
1.516
1.596
1.088
94.8
122.6
116.4
1.518
1.580
1.088
95.4
[1.1.1]Propellane 2
Bicyclo[1.1.0]butane 5
Cyclobutyl 4a
1.525
1.596
1.106
95.1
n(C2C1C4)
n(C1C2C3)
n(C1C3H)
R(C1ÈC2)
R(C1ÈC3)
R(C1ÈH)
61.8
63.5
62.7
63.1
117.6
1.488
1.467
1.070
1.078
98.4
130.4
1.510
1.555
2.127
1.076
1.088
1.083
93.6
117.2
1.494
1.500
1.080
1.089
98.7
129.5
1.503
1.556
2.114
1.086
1.097
1.093
93.5
117.5
1.499
1.494
1.081
1090
98.8
129.8
1.503
1.565
2.127
1.086
1.101
1.094
94.6
116.9
1.489
1.497
1.071
1.093
R(C2ÈH)
n(C2C3C4)
n(C2C3H)
R(C1ÈC2)
R(C2ÈC3)
R(C1ÈC3)
R(C1ÈH)
128.4
R(C2ÈH)
R(C3ÈH)
n(C2C1C4)
n(C1C2C3)
n(C2C3C4)
n(C2C1H)
n(C1C3H)
87.9
90.1
131.4
113.4
87.4
89.5
131.5
116.4
87.8
89.8
132.7
114.5
Phys. Chem. Chem. Phys., 1999, 1, 1399È1404
1401

View Article Online
Table 1ÈContinued
Species
Unita
HF/6È31G*
MP2/6È31G*
B3LYP/6È31G*
Obs.b
TS 6a
R(C1ÈC2)
R(C2ÈC3)
R(C1ÈC3)
R(C1ÈH)
1.522
1.522
1.887
1.080
1.080
1.645
85.6
1.495
1.495
1.555
1.096
1.091
1.645
95.4
1.503
1.503
1.595
1.092
1.092
1.645
94.4
64.1
94.4
154.9
127.5
R(C2ÈH)
R(C3ÈH)
n(C2C1C4)
n(C1C2C3)
n(C2C3C4)
n(C2C1H)
n(C1C3H)
R(C1ÈC2)
R(C2ÈC3)
R(C1ÈC3)
R(C1ÈH)
76.6
85.6
62.7
95.4
156.5
127.7
1.508
1.559
2.133
1.075
1.088
1.521
93.8
88.1
89.9
131.5
115.3
154.6
127.2
1.502
1.559
2.120
1.085
1.099
1.519
93.8
87.7
89.3
131.8
113.9
3-Methylcyclobutyl
4b
R(C2ÈH)
R(C3ÈC5)
n(C2C1C4)
n(C1C2C3)
n(C2C3C4)
n(C2C1H)
n(C2C3C5)
a Angles in degrees; bond lengths in angstroms; see Schemes 1 and 2 for key to numbering. b [1.1.1]Propellane: refs. 23 and 24.
Bicyclo[1.1.0]butane: ref. 25.
cal element into the results, and second, because a recent
study has shown that the scale factors for MP2 and B3LYP
data are very close to unity.29 Only the HF vibrational modes
are signiÐcantly overestimated, but the errors that this intro-
duces into the Ðnal energy di†erences are very small. Individ-
ual energies were therefore estimated using the following
expression:
and alkyl radicals to [1.1.1]propellane would be exothermic.
The B3LYP reaction coordinate indicated essentially no
energy maximum for radical 1a. Reaction coordinate scans
were not successful for radical 4b because of sudden discontin-
uities, possibly because algorithms were unable to cope with
the symmetry change. The c-dissociation of each cyclobutyl
radical 4 was calculated by all the methods to be 14È20 kcal
mol~1 more endothermic than the dissociation of the homolo-
gous radical 1. Because the dissociation of each radical 1 pro-
H
\ H (level/6È31G*) ] ZPVE(level/6È31G*)
298
0
duces [1.1.1]propellane the enthalpy change *H , and to a
] (H [ H ) (level/6È31G*)
298
large extent *E , should be controlled by the thermodynamic
d
298 0 vib
except for the H
298
values at MP3 and MP4 where the
stabilisation of the released alkyl radical. The experimentally
ZPVEs and (H [ H ) values calculated at MP2 were
determined stabilisation energy of t-Bu~ is ca. 9 kcal mol~1
greater than that of Me~.30 Similarly, Me~ is 10È15 kcal
mol~1 more stable than the H-atom (depending on how this is
298
0 vib
used. Using these data, the reaction enthalpies (*H ) and
298
the energy di†erence between the TS and the initial bridge-
head radical (*E ) and the energy di†erence between the TS
assessed).31 Hence the calculated *H
values are expected
d
298
and the product species (*E ) were calculated (Table 3).
to decrease by 10È15 kcal mol~1 on introducing the 3-Me
substituent and by about a further 9 kcal mol~1 on changing
this to t-Bu. Provided that the Hammond postulate holds for
a
Not surprisingly, large swings between the various methods
of calculation were observed in the *H
and *E values.
298
Usually the HF and B3LYP methods were in good agreement
except that the B3LYP calculations indicated lower activation
energies for the reverse radical addition reactions. All methods
agreed that the c-dissociations of the bicyclo[1.1.1]pent-1-yl
radicals would be endothermic and that the additions of H~
this series of dissociations, corresponding decreases in the *E
d
values are expected. Reference to Table 3 shows that only the
HF and B3LYP data roughly follow this trend but that the
MP2 and MP4 methods predict decreases in *E only from
d
1b to 1c, while the MP3 data show only small variations.
Table 2 Calculated energies H (hartree) for radicals 1 and 4, TSs and productsa
0
HF/6È31G*
Geometry optimised
MP2/6È31G*
Geometry optimised
MP3/6È31G*
Single pointb
MP4/6È31G*
Single pointb
B3LYP/6È31G*
Geometry optimised
Species
1a
[193.26378
[232.30489
[349.40676
[193.18522
[232.23756
[349.35211
[192.69107
[154.87177
[155.46726
[194.50398
[155.28210
[39.55899
[156.67501
[193.89417
[233.06774
[350.57519
[193.84244
[233.01298
[350.52916
[193.35028
[155.38997
[155.96192
[195.13148
[155.81956
[39.67074
[157.17911
[193.92895
[233.11061
[350.64221
[193.864223
[233.04823
[350.58198
[193.37039
[155.41770
[155.99835
[195.17813
[155.84614
[39.68585
[193.96181
[233.14416
[350.69623
[193.89986
[233.07887
[350.64939
[193.40612
[155.44389
[156.02168
[195.20665
[195.87341
[39.68936
[157.24550
[194.58419
[233.90380
[351.84512
[194.51299
[233.84488
[351.80234
[194.00898
[155.94805
[156.54532
[195.86192
[156.39284
[39.83829
[157.79833
1b
1c
TS 3a
TS 3b
TS 3c
2
5
4a
4b
TS 6a
CH ~
3
t-Bu~
[157.22232
a H for the H-atom \[0.49823 (HF, MP2, etc.) and [0.50027 (B3LYP). b With MP2 geometry.
0
1402
Phys. Chem. Chem. Phys., 1999, 1, 1399È1404

View Article Online
Table 3 Calculated dissociation enthalpies (*H ) and activation energies for c-dissociations and additions to 2 and 5a
298
Reaction
Parameter
HF
MP2
MP3
MP4
B3LYP
1a ] 2 ] H~
*H
298
*E
d
31.8
35.7
3.9
14.9
20.8
5.9
24.1
28.9
4.8
22.3
27.2
4.9
33.8
33.5
[0.3
*E
a
1b ] 2 ] Me~
1c ] 2 ] t-Bu~
4a ] 5 ] H~
*H
298
*E
d
22.4
33.6
11.2
17.0
27.0
10.0
21.8
31.8
10.0
18.2
33.7
15.5
23.5
30.3
6.7
*E
a
*H
298
*E
d
15.5
27.7
12.2
20.3
23.7
3.4
22.6
32.6
10.0
19.6
24.2
4.6
15.0
21.4
6.4
*E
a
*H
298
*E
d
48.4
102.1
53.7
33.1
78.1
45.0
38.6
84.3
45.7
36.8
81.8
45.0
49.1
85.0
35.9
*E
a
4b ] 5 ] Me~
*H
36.1
32.9
35.3
34.7
36.9
298
a Including ZPVEs and thermal vibrational corrections (see text); energies in kcal mol~1 (1 cal \ 4.184 J).
Hence the trends and absolute values of the HF and B3LYP
reaction enthalpies and activation energies are more credible.
3). The 1H NMR spectrum of the product mixture showed
evidence of the formation of [1.1.1]propellane (d 2.02, s) but,
H
Of particular note are the low *H
and *E predicted for
not surprisingly, the amount was extremely small, because 2
298
d
loss of t-Bu~ from 1c (Table 3). These calculated energy di†er-
ences imply that loss of the t-Bu~ radical, and hence other
radicals with appreciable stabilisation energies such as benzyl
and allyl, should occur comparatively rapidly and should be
detectable within the normal temperature range of organic
laboratory experiments (25È150 ¡C). On the other hand, all
will rapidly polymerise under the reaction conditions.
Radical clock methodology was employed to determine rate
data for the c-scission reaction (k ). Photochemical reductions
d
were carried out in C D or t-butylbenzene with excess tribu-
6
6
tyltin hydride (0.22 M) and the product ratio was measured by
GC at a series of temperatures (50È150 ¡C). It can easily be
shown that k \ k [Me CH][Bu SnH]/[8], where k is the
Ðve computational methods yielded much greater *H
and
298
d
H
3
3
H
*E values for the cyclobutyl radical dissociations, which sug-
rate constant for hydrogen abstraction from tributyltin
hydride by the intermediate radicals. This rate constant is well
d
gested that this process would be unimportant in the normal
temperature range.
established for the t-Bu~ radical [log(k /dm3 mol~1
H
The calculated *E values are fairly small (\6 for addition
s~1 \ 8.43 [ 2.95/2.3RT ]36 but is unknown for 1c. Recent
a
of H~ and \D10 kcal mol~1 for addition of alkyl radicals)
research has shown that bicyclo[1.1.1]pentyl radicals abstract
hydrogen from dienes signiÐcantly faster (ca. 50-fold) than do
t-Bu~ radicals.12 Rate constants for hydrogen transfer from
and this agrees well with the known easy addition of a variety
of radicals to 2. Quantitative experimental data for compari-
son with the calculated *E s are sparse. Scaiano and co-
Bu SnH are nearly the same for primary and tertiary alkyl
a
3
workers15,32 reported that the rate constant for addition of
radicals34 but do increase signiÐcantly for r-radicals. Hence
Et Si~ to 2 was 6 ] 108 dm3 mol~1 s~1 at 21 ¡C. With a
the
k
data measured for the cyclopropyl radical
3
H
““normalÏÏ Arrhenius pre-exponential factor of ca. 1010 dm3
[log(k /s~1) \ 9.3 [ 1.9/2.3RT ],37 which has a r-structure
H
mol~1 s~1, this corresponds to an activation energy of ca. 2
more similar to that of radical 1c than a tertiary alkyl radical,
kcal mol~1, which compares very well with the calculated *E
will be the most appropriate. A plot of the derived k values
a
values. In contrast, the large calculated *E values for addi-
a
d
vs. 1/T gave a satisfactory straight line (r \ 0.93) from which
the following Arrhenius parameters were derived:
tion of Me~ and H~ to bicyclo[1.1.1]butane suggest that this
reaction will be considerably more difficult. Quantitative
kinetic data are not available for radical addition to 5, but the
theoretical prediction is in good agreement with the few qual-
itative observations which indicate that 5 is much inferior to 2
in this aspect of reactivity.33 For instance, radical addition of
log(k /s~1) \ 12.1 ^ 2.0/s~1 [ (11.4 ^ 3.0 kcal mol~1/2.3RT )
d
The measured pre-exponential factor is close (within the error
limits) to the ““normalÏÏ value of 1013 s~1 for a unimolecular
decomposition, which lends credence to the results. These
CCl to 5 was reported5 to be about seven times slower than
4
that to 2.
Experimental study of the c-dissociation of the
3-t-butylbicyclo[1.1.1]pent-1-yl radical 1c
We generated radical 1c and examined its behaviour by end
product analysis over a range of temperatures to test our
theoretical prediction. 1-Iodo-3-t-butylbicyclo[1.1.1]pentane
7c was chosen as the radical precursor and prepared from 1,1-
bis(chloromethyl)-2,2-dibromocyclopropane, via 2,34,35 as
shown in Scheme 3. Radical 1c was generated from 7c by
treatment with tributyltin hydride in hydrocarbon solution at
various temperatures using photochemical initiation. Products
from the reaction in benzene at 80 ¡C were shown by a com-
bination of GC, NMR and MS to be 1-t-butylbicyclo[1.1.1]-
pentane (8) and isobutane (ratio 1 : 0.004). The main process is
transfer of hydrogen from the organotin hydride to the 3-t-
butylbicyclo[1.1.1]pent-1-yl radical 1c, formed by abstraction
of iodine from 3. However, the formation of isobutane estab-
lished that radical 1c does dissociate with loss of a t-Bu~
radical, which picks up hydrogen, even at 80 ¡C (see Scheme
Scheme 3
Phys. Chem. Chem. Phys., 1999, 1, 1399È1404
1403

View Article Online
6
P. Kaszynski, N. D. McMurdie and J. Michl, J. Org. Chem., 1991,
56, 307; P. Kaszynski, A. C. Friedli and J. Michl, J. Am. Chem.
Soc., 1992, 114, 601.
experimental data conÐrmed our theoretical prediction that
c-dissociation would be a signiÐcant process for loss of t-Bu~
from radical 1c. The rate of dissociation of 1c is not large at
7
8
9
E. W. Della, P. E. Pigou, C. H. Schiesser and D. K. Taylor, J.
Org. Chem., 1991, 56, 4659.
room temperature [k (25 ¡C) \ 6 ] 103 s~1], but the reaction
d
gains in importance at higher temperatures. The experimental
A. Almenningen, B. Andersen and B. A. Nyhus, Acta Chem.
Scand., 1971, 25, 1217.
K. B. Wiberg, C. M. Hadad, S. Sieber and P. v. R. Schleyer, J.
Am. Chem. Soc., 1992, 114, 5820.
Arrhenius activation energy refers to the dissociation in solu-
tion for T [ 273 K and is therefore not directly comparable
with the computed *E values (Table 3) which refer to iso-
d
10 See, for example, R. C. Fort and P. v. R. Schleyer, J. Am. Chem.
Soc., 1971, 93, 3189; C. Ruchardt, Angew. Chem., Int. Ed. Engl.,
1970, 9, 830; P. S. Engel, Chem. Rev., 1980, 80, 99.
11 C. Ruchardt, K. Herwig and S. Eichler, T etrahedron L ett., 1969,
421; B. Giese, T etrahedron L ett., 1979, 857; B. Giese and J. Stell-
mach, Chem. Ber., 1980, 113, 3294.
12 J. T. Banks, K. U. Ingold, E. W. Della and J. C. Walton, T etra-
hedron L ett., 1996, 37, 8059.
13 K. P. Dockery and W. G. Bentrude, J. Am. Chem. Soc., 1997, 119,
lated molecules. Normally, however, experimental activation
energies for neutral free radical reactions in hydrocarbon solu-
tions are very similar to the corresponding gas phase values.
There is a signiÐcant discrepancy of about a factor of two
between the calculated *E s and the experimental activation
a
energy. Highly strained open shell species such as 1 and 2 are
a severe test of theory and this is shown by the considerable
swings from level to level in the calculated energy di†erences;
perhaps a more sophisticated basis set is needed. The experi-
mental pre-exponential factor is on the low side. An increase
in this would lead to a correspondingly higher activation
energy, perhaps as high as 15 kcal mol~1. Taking the poten-
tial error limits on both calculated and experimental data into
account, the di†erence between theory and experiment is not
inordinate.
1388.
14 U. Bunz, K. Polborn, H.-U. Wagner and G. Szeimies, Chem. Ber.,
1988, 121, 1785.
15 P. F. McGarry, L. J. Johnston and J. C. Scaiano, J. Org. Chem.,
1989, 54, 6133.
16 D. Feller and E. R. Davidson, J. Am. Chem. Soc., 1987, 109, 4133.
17 W. Adcock, G. T. Binmore, A. R Krstic, J. C. Walton and J.
Wilkie, J. Am. Chem. Soc., 1995, 117, 2758.
18 K. B. Wiberg and F. H. Walker, J. Am. Chem. Soc., 1982, 104,
5239; K. B. Wiberg, T etrahedron L ett., 1985, 26, 599.
19 E. W. Della and D. K. Taylor, J. Org. Chem., 1994, 59, 2986.
20 E. W. Della, D. K. Taylor and J. Tsanaktsidis, T etrahedron L ett.,
1990, 36, 5219.
21 M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G.
Johnson, M. A. Robb, J. R. Cheeseman, T. A. Keith, G. A.
Petersson, J. A. Montgomery, K. Raghavachari, M. A. Allaham,
V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B.
Stefanov, A. Nanayakkara, M. Challacombe, C. Y. Peng, P. Y.
Ayala, W. Chen, M. W. Wong, J. L. Anders, E. S. Replogle, R.
Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees, J.
Baker, J. J. P. Stewart, M. Head-Gordon, C. Gonzales and J. A.
Pople, Gaussian 94, Revisions D.4 and E.1, Gaussian, Pittsburg,
PA, 1995.
Conclusions
Ab initio computations predict that radical additions to
[1.1.1]propellane are exothermic and should occur with low
activation energies. This is in good agreement with qualitative
experimental evidence. Calculated enthalpies and activation
energies indicated that the reverse c-dissociations of 3-
alkylbicyclo[1.1.1]pent-1-yl radicals could be important for
extrusion of stabilised radicals. Experimentally, loss of t-Bu~
from 1c was conÐrmed and the kinetic study indicated a
remarkably low activation energy for a CÈC bond Ðssion. This
result implies that similar loss of other thermodynamically
stabilised radicals such as benzyl and allyl should become
22 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
23 L. Hedberg and K. Hedberg, J. Am. Chem. Soc., 1985, 107, 7257.
24 K. B. Wiberg, W. P. Dailey, F. H. Walker, S. T. Waddell, L. S.
Crocker and M. Newton, J. Am. Chem. Soc., 1985, 107, 7247.
25 K. W. Cox, M. D. Harmony, G. Nelson and K. B. Wiberg, J.
Chem. Phys., 1968, 50, 1976; J. Am. Chem. Soc., 1968, 90, 3395.
26 J. Pacansky and M. Yoshimine, J. Phys. Chem., 1985, 89, 1880.
27 L. A. Curtis and J. A. Pople, J. Chem. Phys., 1988, 88, 7405.
28 L. A. Curtis and J. A. Pople, J. Chem. Phys., 1989, 91, 2420.
29 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16502; W. J.
Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab Initio
Molecular Orbital T heory, Wiley, New York, 1986, ch. 6, p. 133.
30 J. Berkowitz, C. B. Ellison and D. Gutman, J. Phys. Chem., 1994,
98, 2744.
competitive
at
temperatures
above
ambient.
3-
Benzylbicyclo[1.1.1]pentane derivatives have been prepared
via 3-benzylbicyclo[1.1.1]pent-1-yl radicals.13,38 However, the
addition of benzyl bromide to 2 did not proceed smoothly38
and a possible cause of this is obviously competition from the
c-dissociation. The formation of [1.1.1]propellane substan-
tially promotes CÈC bond Ðssion, which occurs much more
readily in 1 than in analogous c-dissociations of the next
lower homologue i.e. 3-substituted-cyclobutyl radicals 4. c-
Dissociation of 1 is also more facile than analogous b-
scissions of alkyl radicals to yield alkenes.
31 D. F. McMillen and D. M. Golden, Annu. Rev. Phys. Chem.,
1982, 33, 493.
32 J. C. Scaiano and P. F. McGarry, T etrahedron L ett., 1993, 34,
The authors thank the UK Computational Chemistry Facility
for the allocation of time on the DEC 8400 superscalar service
(Columbus), Drs. K. U. Ingold, J. W. Essex and O. Parchment
for helpful discussions and NATO for a travel grant.
1243.
33 K. U. Ingold and J. C. Walton, Acc. Chem. Res., 1986, 19, 72, and
references cited therein.
34 K. Semmler, G. Szeimies and G. Belzner, J. Am. Chem. Soc., 1985,
107, 6410; J. Belzner, U. Bunz, K. Semmler, G. Szeimies, K. Opitz
and A. Schluter, Chem. Ber., 1989, 112, 397.
35 P. Kaszynski and J. Michl, J. Org. Chem., 1988, 53, 4593.
36 C. Chatgilialoglu, K. U. Ingold and J. C. Scaiano, J. Am. Chem.
Soc., 1981, 103, 7739.
References
1
D. Griller, K. U. Ingold, P. J. Krusic and H. Fischer, J. Am.
Chem. Soc., 1978, 100, 6750.
J. C. Walton, Chem. Soc. Rev., 1992, 21, 105.
The tetrahedryl radical is even more ““tied backÏÏ but will be
unstable and has not been reported.
2
3
37 L. J. Johnston, J. Lusztyk, D. D. M. Wayner, A. N. Abeywick-
reyma, A. L. J. Beckwith, J. C. Scaiano and K. U. Ingold, J. Am.
Chem. Soc., 1985, 107, 4594.
4
5
E. W. Della and I. J. Lochert, Org. Prep. Proced. Int., 1996, 28,
38 P. Kaszynski and J. Michl, J. Am. Chem. Soc., 1988, 110, 5225.
411.
K. B. Wiberg and S. T. Waddell, J. Am. Chem. Soc., 1990, 112,
2194.
Paper 8/08744J
1404
Phys. Chem. Chem. Phys., 1999, 1, 1399È1404