3,5-Di-tert-butyl-3'-(N-tert-butyl-N-aminoxy)-4-oxybiphenyl: A heterospin diradical with temperature dependent behavior

Article abstract of DOI:10.1021/jo990317l

The title diradical was synthesized and investigated by ESR and UV-vis spectroscopy. It was found to have a lifetime of weeks even in the presence of oxygen, and even survives brief heating in toluene up to about 60°C. In the UV-vis spectrum, the diradica

Full text of DOI:10.1021/jo990317l

5176
J . Org. Chem. 1999, 64, 5176-5182
3,5-Di-ter t-bu tyl-3′-(N-ter t-bu tyl-N-a m in oxy)-4-oxybip h en yl: A
Heter osp in Dir a d ica l w ith Tem p er a tu r e Dep en d en t Beh a vior
Yi Liao,† Chunping Xie,† Paul M. Lahti,*,† Ralph T. Weber,‡ J inJ ie J iang,‡ and David P. Barr‡
Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003 and
Bruker Instruments Inc., Billerica, Massachusetts 01821
Received February 19, 1999
The title diradical was synthesized and investigated by ESR and UV-vis spectroscopy. It was found
to have a lifetime of weeks even in the presence of oxygen, and even survives brief heating in
toluene up to about 60 °C. In the UV-vis spectrum, the diradical showed reversible thermochromic
behavior in the -10 to 50 °C range. In the ESR spectrum, hyperfine analysis showed nearly isolated
behavior by the phenoxyl and nitroxide spin carrying units. Upon warming, additional, broad lines
appeared at the expense of the lower temperature lines. This temperature behavior was reversible
over the 0 to 50 °C range, so long as the raised temperatures were not maintained for long time
periods. The spectral behavior is interpreted as being due to temperature-dependent conformational
effects on the exchange coupling between the spin carrying units, i.e., J -modulated exchange
behavior.
In tr od u ction
despite the fact that a na¨ıve perturbation-based analysis
suggests 1 to have an appreciable splitting of its singly
occupied frontier molecular orbitals (MOs), a factor
typically seen as favoring singlet state behavior.
In another comparison, nondisjoint dinitroxide diradi-
cals such as 2 and 3 have been found experimentally to
have singlet ground states with low-lying excited triplet
states,12,13 in spite of the fact that qualitative parity
Conjugated diradicals have been a topic of considerable
theoretical and experimental interest since the late 19
century.1,2 During the past 30 years, advances in experi-
mental technology and theoretical understanding have
yielded a variety of qualitative structure-property rela-
tionships for diradicals and related open-shell molecules.3-9
Despite such better understanding, some facets of diradi-
cal behavior remain to be extensively probed. In particu-
lar, heterospin open shell systems have not been greatly
studied, by which in this case we specifically refer to
organic systems having hypovalent sites of different
types.
Heterospin diradicals are important probes of the
interplay between frontier orbital energy level perturba-
tion by substitution, versus the parity-based tendencies
of such systems to give high-spin ground-state multiplici-
ties. In an archetypal example, m-benzoquinomethane,
1, has been experimentally shown to have a triplet
ground state (or possibly a near-degeneracy of triplet and
singlet states).10 Recent multiconfiguration self-consistent
field computations are in agreement with the experiment,
and show a favoring of the triplet over the singlet state
by about 10 kcal/mol.11 The triplet preference is observed
models suggest high spin ground states. In these cases,
a connectivity-based preference for the high spin state
is offset by conformational dependence of intramolecular
exchange between the unpaired electrons, as well as by
heteroatom substituent effects upon the exchange. A
review has detailed a number of such so-called “violations
of Hund’s rule”.14 Overall, although widely accepted
structure-property guidelines exist for predicting the
ground-state multiplicities of open-shell molecules, com-
plex combinations of molecular connectivity and substi-
tution pattern can confound the guidelines.
† University of Massachusetts.
‡ Bruker Instruments Inc.
Relatively few heterospin open-shell molecules have
been made, by comparison to the numbers of systems
having all the same spin carrying units. Mixed carbene/
nitrene,15 carbene/nitroxide,16 nitrene/nitroxide,17 and
(1) Diradicals; Borden, W. T., Ed.; J ohn Wiley and Sons: New York,
1982.
(2) Michl, J .; Bonacic-Koutecky, V. Tetrahedron 1988, 24, 7559.
(3) Longuet-Higgins, H. C. J . Chem. Phys. 1950, 18, 265, 275, 283.
(4) Longuet-Higgins, H. C. In Theoretical Organic Chemistry, Kekule
Symposium; Butterworth: London, 1958; pp 9-19.
(5) Klein, D. J .; Nelin, C. J .; Alexander, S.; Matsen, F. E. J . Chem.
Phys. 1982, 77, 3101.
(12) Dvolaitzky, M.; Chiarelli, R.; Rassat, A. Angew. Chem., Int. Ed.
Engl. 1992, 31, 180.
(6) Klein, D. J . Pure Appl. Chem. 1983, 55, 299.
(7) Misurkin, I. A.; Ovchinnikov, A. A. Russ. Chem. Res. (Engl.)
1977, 46, 967.
(8) Ovchinnikov, A. A. Theor. Chim. Acta 1978, 47, 297.
(9) Borden, W. T.; Davidson, E. R. J . Am. Chem. Soc. 1977, 99, 4587.
(10) Rule, M.; Matlin, A. R.; Seeger, D. E.; Hilinski, E. F.; Dougherty,
D. A.; Berson, J . A. Tetrahedron 1982, 38, 787.
(11) Horvat, D. A.; Murcko, M. A.; Lahti, P. M.; Borden, W. T. J .
Chem. Soc., Perkin Trans. 2 1998, 1037.
(13) Kanno, F.; Inoue, K.; Koga, N.; Iwamura, H. J . Am. Chem. Soc.
1993, 115, 847.
(14) Borden, W. T.; Iwamura, H.; Berson, J . A. Acc. Chem. Res. 1994,
27, 109.
(15) Tukada, H.; Mutai, K.; Iwamura, H. Chem. Commun. 1987,
1159.
(16) Matsuda, K.; Iwamura, H. Chem. Commun. 1996, 1131-1132.
(17) Esat, B.; Walton, R.; Lahti, P. M. J . Am. Chem. Soc. 1998, 120,
5122.
10.1021/jo990317l CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/10/1999

Temperature-Dependent Heterospin Diradical
J . Org. Chem., Vol. 64, No. 14, 1999 5177
F igu r e 1. Synthesis of diradical 4, monoradicals 13 and 14. (a) n-BuLi, Et2O, -78 °C; (b) Me2C(NO)Me, room temp; (c) ClSiMe2-
t-Bu, imidazole, DMF, 50 °C; (d) t-BuLi, Et2O, -78 °C, then B(O-iso-Pr)3, then hydrolyze; (e) 6, Pd(OAc)2, PBu4, THF; (f) Bu4NF,
THF, 50 °C; (g) PbO2, toluene, room temp; (h) Pd(OAc)2, PPh3, NaOEt, THF.
nitroxide/radical-ion combinations have been synthesized
and investigated. We aimed to synthesize systems incor-
porating phenoxyl and nitroxyl spin carrying units, both
of which have been used in polyradicals and molecular
magnetic materials. In this report, we describe the
synthesis and spectral analysis of the diradical 3,5-di-
tert-butyl-3′-(N-tert-butyl-N-aminoxy)-4-oxybiphenyl, 4.
Sterically stabilized phenoxyl radicals have also been
known for decades.19 By contrast to tBNs, phenoxyl
radicals show large spin delocalization. ESR hfc analysis
of phenoxyl itself shows that the radical has a large
contribution from carbonyl-containing resonance struc-
tures and is well represented as having its main spin
density on the position para to the hypovalent oxy site.
This tendency toward delocalization of phenoxyl radicals
has made them a much-used spin carrying unit in
molecular magnetism studies, in the anticipation that
spin delocalization will increase intramolecular exchange
values. However, the delocalization also destabilizes the
phenoxyl system, to the extent that dimerization occurs
fairly readily even in many sterically ortho/para block-
aded systems.
Our goal in this study was to combine the spin-
delocalizing capability of the 2,6-di-tert-butylphenoxyl
moiety with the stability of a phenyl tert-butylnitroxide
moiety to make a heterospin diradical system. We are
not aware of other reported examples of nondisjoint
heterospin exchange between a phenoxyl and a nitroxide
radical.
Ba ck gr ou n d
Sterically hindered nitroxide radicals have been known
for almost 40 years as stable hypovalent molecules.18
Their stability is primarily due to blockading of possible
dimerization pathways. Some degree of spin delocaliza-
tion is observable in arylnitroxide analogues, based on
the smaller nitrogen hyperfine coupling constants (hfc,
aN) that are observed experimentally by comparison to
alkylnitroxides. If one makes the usual assumption that
hfc is proportional to spin density in the nitroxides, only
a small amount of the nitroxide spin density delocalizes
onto the benzene ring in an aryl tert-butylnitroxide. This
is in part due to the two-center, three-electron nature of
the nitroxide spin carrying unit, and in part due to a
varying degree of twisting of the tert-butylnitroxide (tBN)
group relative to an attached aryl group. The torsion of
the nitroxide reduces the possibility of spin delocalization,
resulting in overall localized behavior for tBNs.
Syn th esis
The overall scheme of Figure 1 was followed to syn-
thesize 4 and related model monoradicals. The two spin-
carrying units were made separately as protected forms,
coupled, and deprotected to give the appropriate bis-
(18) Forrester, A. R.; Hay, J . M.; Thomson, R. H. Organic Chemistry
of Stable Free Radicals; Academic Press: New York, 1968.
(19) Altwicker, E. R. Chem. Rev. 1967, 67, 475.

5178 J . Org. Chem., Vol. 64, No. 14, 1999
Liao et al.
F igu r e 3. Electron spin resonance spectrum from oxidation
of 9 (ν0 ) 9.430 GHz) to 4 at room temperature in toluene
solution. The top spectrum is experimental, the lower spectrum
is line shape fitted using program BIRADG (ref 22).
F igu r e 2. Electron spin resonance spectra for phenoxyl
radical 14 (ν0 ) 9.762 GHz) and nitroxide radical 13 (ν0 ) 9.802
GHz) at room temperature in benzene solutions.
hydroxy precursor. Hydroxylamine 5 was made by trap-
ping 3-bromo-phenyllithium with 2-methyl-2-nitrosopro-
pane and quickly converted to the tert-butyldimethylsilanyl
ether 6 to minimize air oxidation. Tris[3,5-bis(tert-butyl)-
4-(trimethylsilyloxy)phenyl]boroxin, 7, was conveniently
made by Satoh’s method20 as the protected phenol unit.
The assembly of the two halves to make precursor 8 was
accomplished under Suzuki conditions, during which the
trimethylsiloxy group was cleaved. The protected inter-
mediate 8 was deblocked to the bis-hydroxy precursor 9
by treatment with tetrabutylammonium fluoride. Precur-
sor 9 discolored readily upon storage in atmospheric
oxygen, apparently due to air oxidation of the hydroxyl-
amine group. The oxidation was demonstrated by obser-
vation of a typical nitroxide 1:1:1 triplet spectrum in the
ESR of discolored solution samples of 9.
Figure 1 also shows the synthesis of a monoradical
model compound via Suzuki coupling of 2,6-di-tert-
butylanisole-4-boronic acid 10 with 6 to give compound
11. The silyl ether group in 11 was deprotected using
tetrabutylammonium fluoride to give hydroxylamine 12.
Precursor 12 could then be used to generate model
monoradical 13, as described below. The analogous
phenoxyl spin-site monoradical 14 was generated by
oxidation of intermediate 8.
F igu r e 4. UV-vis spectra at varying temperatures from
oxidation of 9 in toluene solution.
line shape fitting, using the WINSIM program of Dul-
ing.21 The detailed hfc data are given in the Experimental
Section.
Dihydroxy precursor 9 was oxidized in various solvents
using both aqueous sodium periodate and lead dioxide.
The latter conditions always gave the same ESR and
UV-vis spectra within 15-30 min. These spectra are
shown in Figures 3 and 4, respectively. Sodium periodate
oxidized only the nitroxide group of 9, based on the
evidence of ESR spectra that showed only nitroxide hfc.
In addition, sodium periodate oxidation gave only the
reddish color of nitroxide radicals, by comparison to the
greenish color obtained from diradical 4. When a sodium
periodate partially oxidized sample was stirred with lead
dioxide, the typical spectra of diradical 4 were obtained,
identical to those obtained by complete oxidation of 9 with
lead dioxide only.
Resu lts
The monoradicals 13 and 14 were made by heteroge-
neous phase oxidation in benzene using lead dioxide.
Both gave strong and persistent ESR spectra, as shown
in Figure 2. The hfc for these radicals were obtained by
(20) Satoh, Y.; Shi, C. Synthesis 1994, 1146.
(21) Duling, D. R. J . Magn. Reson. 1994, B104, 105.

Temperature-Dependent Heterospin Diradical
J . Org. Chem., Vol. 64, No. 14, 1999 5179
trum that appeared to be dominated by the nitroxide hfc.
During the temperature cycling, the sample varied from
a greenish color at room temperature to a darkly reddish
shade at higher temperatures. This thermochromic be-
havior was reversible and correlates with the spectral
changes shown in Figures 4 and 5.
Discu ssion
Diradical 4 is a nondisjoint system by the Borden-
Davidson9 classification scheme and is expected to have
a high-spin triplet ground state. 3,4′-Biphenyl analogues
with carbene23 and nitrene24 spin-carrying units have
been investigated and found to have high-spin ground
states. However, as mentioned in the Introduction, a
number of other systems have recently been described
that contradict simple connectivity-based ground state
predictions. Many of these latter systems incorporate
nitroxide spin-carrying units. In particular, the nondis-
joint dinitroxides 2 and 3 described in the Introduction
are both known to have singlet ground states with low-
lying excited triplet states. In the case of 2, the discrep-
ancy with theory is attributed to deconjugation of the
nitroxide spin sites by steric-induced torsion. Since
nitroxides are not greatly delocalized in the typical aryl
nitroxides, the situation grows worse when torsion is
larger than usual. For diradical 3, a crystal structure has
been obtained13 showing the nitroxide units to be twisted
by 65° and 75°, demonstrating the correlation between
nitroxide torsion and a weakening of the parity rules for
ground state prediction.
F igu r e 5. Electron spin resonance spectra at varying tem-
peratures from oxidation of 9 in toluene solution (ν0 ) 9.430
GHz).
Line shape analysis of the ESR spectrum obtained by
lead dioxide treatment of 9 was carried out using the
program BIRADG, kindly made available to us by Dr.
B. Kirste.22 BIRADG is a line shape analysis program
for weakly interacting diradicals, the input for which
includes hfc for the radical fragments that interact to give
the diradical, and the exchange coupling constant be-
tween the fragments. Figure 3 shows the simulated
spectrum obtained by us. Further details of the fitted hfc
data are given in the Experimental Section below.
Figure 4 also shows the results of experiments in which
precursor 9 was oxidized with lead dioxide in toluene and
subjected to variable temperature analysis of its UV-
vis spectrum. The observed variation of the spectrum was
reversible over multiple cycles between -10 and 60 °C.
Upon storage at -30 °C, the same sample showed
essentially no change in the intensity or line shape of its
UV-vis spectrum for at least two months, despite not
being under inert atmosphere.
Precursor 9 was also oxidized with lead dioxide in
toluene, subjected to a 3-fold freeze-pump-thaw degas-
sing, and sealed under vacuum for variable temperature
ESR spectral analysis. The results are shown in Figure
5. As the sample was heated above room temperature,
the highly hyperfine-split spectrum broadened and be-
came complicated by the appearance of broad peaks
within the original spectral extent. Over a temperature
range of 0-60 °C, the line shape behavior was reversible
so long as the sample was not subjected to the upper
temperature range for extended times. But, if the tem-
perature was raised above 70 °C the original spectrum
could not be recovered, instead yielding a simpler spec-
The room-temperature solution phase ESR spectrum
of 4 is essentially that expected for a summation of
isolated, noninteracting spin sites. Simulation of the
spectrum (Figure 2) by BIRADG as a pair of weakly
interacting nonequivalent radicals gives a very good fit
to the experimental spectrum with an intramolecular
exchange constant J < 0.1 G. If the spectrum is fitted as
a summation of isolated radicals (WINSIM), another very
good fit is obtained, for which the phenoxyl and nitroxide
radical components are found each to comprise 50 ( 3%
of the spectrum. All the room-temperature simulations
are consistent with 4 having extremely weak exchange
interaction between its spin units. Solution spectra
obtained at temperatures below room temperature down
to 0 °C show no further change of the narrow line
features.
However, even at room temperature there was a weak,
broad spectral feature that rises above the baseline
between the low field nitroxide features and the central
phenoxyl features. This peak and a corresponding one
(23) Teki, Y.; Fujita, I.; Takui, T.; Kinoshita, T.; Itoh, K. J . Am.
Chem. Soc. 1994, 116, 11499.
(22) Kirste, B. Anal. Chim. Acta 1992, 265, 191.
(24) Minato, M.; Lahti, P. M. J . Phys. Org. Chem. 1991, 4, 459.

5180 J . Org. Chem., Vol. 64, No. 14, 1999
Liao et al.
superimposed on the region between the central phenoxyl
features and the high-field nitroxide features strength-
ened steadily as the temperature was raised to the 50-
70 °C range, while the rest of the narrow line features
broadened and weakened. When the sample was cooled
again, the line shape returned to the original appearance.
The cycling of the ESR spectrum could be repeated
multiple times reproducibly, so long as the sample was
not subjected to the higher temperature ranges for more
time than was required to run a spectrum. A sample that
was kept in a heated ESR spectral cavity for over 2 h
and heated to >60 °C underwent irreversible changes,
giving a spectrum that was essentially that of an isolated
nitroxide unit. This sample also stopped showing ther-
mochromic behavior. We did not further investigate this
result, but postulate that the more reactive phenoxyl site
in 4 undergoes irreversible chemical reaction at elevated
temperatures.
The thermochromic behavior of 4 that is shown in
Figure 4 correlates with the ESR behavior. One peak at
about 490 nm increased as the sample was warmed, and
decreased as the sample was cooled again. The visual
effect was to change the greenish room temperature color
to a darker green plus red color at elevated temperatures.
Although one should not push qualitative reasoning too
far, it is worth noting that many phenoxyl radicals are
blue to green in color, while aryl nitroxides are commonly
reddish. For this reason, we suspected that an increase
in temperature led to an increase in nitroxide-like
character in 4.
The combination of ESR and UV-vis spectral results
strongly supports the presence of conformationally con-
trolled electronic variation in 4. Both rotation of the
nitroxide unit and phenyl-phenyl torsion within the
biphenyl moiety would be expected to lead to changes in
overall conjugation and intramolecular exchange behav-
ior. It is not obvious from the experimental behavior
which rotationsor mix of rotational behaviorssis most
likely in this case.
We carried out computational investigation of 4 to
probe the variation of exchange coupling as functions of
the two major torsional modes, using the semiempirical
molecular orbital plus configuration interaction (CI)
methodology in MOPAC9325 that has previously been
established26,27 to give good reproduction of experimental
results for such diradical systems. Geometries were
constrained at the multielectron CI (MECI) level with a
CI ) (6,2) active space. The tert-butyl groups on the
phenoxyl unit were replaced with hydrogen atoms, and
the tert-butyl group on the nitroxide group was replaced
with a methyl group. Three cases were considered: (1)
free optimization, (2) bisected biphenyl constraint, (3)
bisected nitroxide constraint. At the PM3-CI28 level of
theory, the fully planar and bisected biphenyl conformers
are nearly isoenergetic, with the bisected biphenyl con-
former being 0.3 kcal/mol lower in energy. The bisected
nitroxide geometry was found to be 2.7 kcal/mol higher
in energy than the bisected biphenyl conformer.
0.1 kcal/mol at both the bisected biphenyl geometry and
the bisected nitroxide geometry, with the triplet favored
in both cases. The triplet-singlet gap at the planar
geometry is consistent with the usual parity rule predic-
tions for systems with effective intramolecular exchange.
The very small gap at the bisected geometries is consis-
tent with the experimental results and fits earlier
computations showing that nitroxide delocalization is
vitiated by torsion. Overall, the computations predict that
twisting about the central bond in the biphenyl ring is
easy at room temperature, and that the twisting would
virtually eliminate effective exchange interaction be-
tween the spin-carrying sites.
The ESR spectra of the two monoradical models for 4,
systems 13 and 14, are useful in demonstrating the
different spin delocalization behaviors of the phenoxyl
and nitroxide spin units in the diradical. The phenoxyl
unit (14) showed strong spin delocalization onto the
second phenyl ring, with hyperfine coupling being ob-
served for all of the hydrogen atoms on the biphenyl
system. Significantly, there was no resolvable hfc from
the nitrogen atom in the spectrum of 14. By contrast,
the spectrum for 13 showed no resolvable hfc beyond that
on the ring to which the nitroxide is attached. The
nitrogen hfc aN ) 12.5 G at room temperature, which is
in the upper range of hfc for aryl tert-butylnitroxides,
consistent with modest spin delocalization at best.
Monoradical 13 was also warmed in toluene solution
to 67 °C in an effort to gauge the effect of temperature
on the nitrogen hfc. Although some line broadening was
observed, aN remained essentially unchanged. It seemed
likely that, if nitroxide rotation became sufficiently eased
at elevated temperatures to improve the exchange cou-
pling in 4, then a similar change in nitroxide rotation
would lead to a decrease (or at least a visible change) in
aN for 13. Because temperature-dependent behavior of
aN was not observed for 13, we tend to favor biphenyl
central bond rotation as the conformational mode affect-
ing the ESR spectrum for 4. This is also in agreement
with the greater stability of the bisected biphenyl geom-
etry over the bisected nitroxide geometry in the semiem-
pirical computations for 4. We assign the spectral be-
havior shown in Figure 5 to modulation of the intramolec-
ular exchange energy (J -modulation) by conformational
mobility. As the temperature increases, phenyl-phenyl
rotation becomes easier, and the biphenyl group becomes
more planar as a time average. J -modulated line shape
behavior can be complex,29,30 and further experiments are
needed to understand more fully the spectral behavior
of Figure 5. However, the observed changes are not
consistent with equilibration between two structurally
different species, or with irreversible reactions of any
sort. The spectrum cannot be interpreted in terms of a
biradical for which the exchange and hyperfine constants
are nearly equal, or for which the exchange is appreciably
larger than the hyperfine. The remaining choice of a
biradical having exchange much smaller than the hy-
perfine interaction presumably is correct, with the added
complexity of conformationally induced J -modulation of
the line shape behavior.
The computed triplet-singlet state energy gap ∆E(T
- S) varied as a function of geometry. At the fully planar
geometry, ∆E(T - S) ) 3.1 kcal/mol, but ∆E(T - S) ∼
We were concerned with the possibility of incomplete
oxidation of precursor 9. Indeed, we found that use of
(25) Stewart, J . J . P. J . Comput.-Aided Mol. Des. 1990, 4, 1.
(26) Lahti, P. M.; Ichimura, A. S.; Berson, J . A. J . Org. Chem. 1989,
54, 958.
(27) Lahti, P. M.; Ichimura, A. S. J . Org. Chem. 1991, 56, 3030.
(28) Stewart, J . J . P. J . Comput. Chem. 1989, 10, 209, 221.
(29) Avdievich, N. I.; Forbes, M. D. E. J . Phys. Chem. 1996, 100,
1993.
(30) Avdievich, N. I.; Forbes, M. D. E. J . Phys. Chem. 1995, 99, 9660.

Temperature-Dependent Heterospin Diradical
J . Org. Chem., Vol. 64, No. 14, 1999 5181
freshly prepared lead dioxide as opposed to commercial
lead dioxide greatly speeded the oxidation process. If
fresh oxidant was used, the ESR and UV-vis spectra
failed to show further changes if stirring was continued
for >30 min. Oxidation of 9 in carbon tetrachloride gave
the usual ESR spectrum and in addition showed complete
consumption of the distinctive phenolic OH peak in the
room-temperature solution FTIR spectrum. In these and
other typical cases, both 4-substituted-2,6-di-tert-bu-
tylphenols and aryl tert-butylhydroxylamines appear to
undergo complete conversion to the corresponding radi-
cals within 30 min or so of heterogeneous phase (fresh)
lead dioxide treatment. Given the loss of the OH peak in
the FTIR and the constancy of the ESR spectral appear-
ance, we feel that oxidation of 9 is essentially complete
under our conditions to give 4.
We were not able to observe any dipolar splitting of
the spectrum of 4 in frozen benzene or toluene at 77 K.
The spectrum under these conditions was just a single,
featureless peak. This is by contrast to the observable
triplet dipolar splitting in analogue 15, which was
recently reported by Shultz and Farmer.31 For 15, the
dipolar zero field splitting (zfs) |D/hc| ) 0.0084 cm-1 in
cryogenic matrix (with a sodium counterion). The cryo-
genic spectrum also showed a half-field transition that
followed the Curie Law at varying temperatures, indicat-
ing that the triplet state of 15 is either the ground state,
or is virtually degenerate with the singlet state. The
room-temperature ESR spectrum of 15 was estimated to
have hfc of aN ) 12.5 G and aH ) 1.8 G (3H), all on the
nitroxide-bearing ring. It is interesting to note that 15
shows low-temperature zfs, while 4 does not. This sug-
gests that the semiquinone-bearing system has more
effective exchange than the phenoxyl bearing system. It
not clear whether this is may be due to inherently
stronger exchange between the spin sites, or perhaps a
more planar conformation in 15.
Exp er im en ta l Section
Gen er a l P r oced u r es. 2-Methyl-2-nitrosopropane was pre-
pared by the procedure of Stowell.32 Diethyl ether and tet-
rahydrofuran (THF) were distilled under argon after drying
over sodium. N,N-Dimethylformamide (DMF) was dried over
anhydrous magnesium sulfate. Toluene was dried over sodium.
All the other starting materials were obtained from Aldrich
Chemical Corp. and used without further purification. El-
emental analyses were carried out by the University of
Massachusetts Microanalytical Laboratory.
ter t-Bu tyl(3-br om op h en yl)-h yd r oxyla m in e (5). To a
solution of 2.0 g (8.5 mmol) of 1,3-dibromobenzene in 35 mL
of diethyl ether was added 3.6 mL (8.9 mmol) of a 2.5 M
solution of n-butyllithium in hexane at -78 °C under nitrogen.
After being stirred for 30 min, the solution was warmed to
room temperature and then cooled to -78 °C. A solution of
1.11 g (12.7 mmol) of 2-methyl-2-nitrosopropane32 in diethyl
ether was added, and the solution was stirred overnight at
ambient temperature. Next, 10 mL of saturated ammonium
chloride aqueous solution was added. The organic layer was
separated and washed with water. The aqueous layer was
extracted with diethyl ether (30 mL × 3), and the organic
layers were combined. Evaporation of the organic solvent layer
under reduced pressure gave 1.6 g (80% yield) of pale yellow
crystals. This product was used as soon as possible in the next
step, to avoid oxidation by ambient air. 1H NMR (CDCl3) δ
1.13 (s, 9 H), 5.70 (br s, 1 H), 7.10-7.45 (m, 4 H). IR (KBr,
cm-1) 3195 (N-OH str).
ter t-Bu tyl(3-br om op h en yl)(ter t-bu tyld im eth ylsiloxy)-
a m in e (6). To a solution of 1.0 g of 5 (4.10 mmol) in 5.0 mL of
DMF were added 1.23 g (8.20 mmol) of tert-butyldimethyl-
chlorosilane and 0.74 g (12.30 mmol) of imidazole. The solution
was stirred for 24 h under nitrogen at about 50 °C. The
solution was extracted with hexane (40 mL × 3), and the
organic layers were combined and evaporated under reduced
pressure. The resulting residue was chromotographed on silica
gel with hexane to give 0.94 g of 2 (64% yield) as a colorless
oil. H NMR (CDCl3) δ -0.11 (s, 6 H), 0.91 (s, 9 H), 1.10 (s, 9
H), 7.11-7.43 (m, 4 H).
Tr is[3,5-b is(ter t-b u t yl)-4-(t r im et h ylsilyloxy)p h en yl]-
bor oxin (7). This compound was synthesized by the method
of Satoh and Shi.20
1
3,5-Di-ter t-b u t yl-3′-[N-ter t-b u t yl-N-(ter t-b u t yld im et h -
ylsiloxy)a m in o]-4-h yd r oxybip h en yl (8). To a stirred sus-
pension of 0.0157 g (0.07 mmol) of Pd(OAc)2 in 0.5 mL of THF
was added 0.0142 g (0.07 mmol) of tributylphosphine in 0.5
mL of THF under nitrogen. After being stirred for 10 min, this
solution was added to a solution of 0.5 g (1.4 mmol) of 6 plus
0.5 g (1.55 mmol) of boroxin 7 in 10 mL of THF. To this solution
was added a sodium methoxide solution prepared by reaction
of 0.113 g (5 mmol) of sodium with 2 mL of methanol. The
resulting reaction mixture was stirred for 24 h under nitrogen
at room temperature. The solution was washed with water.
The organic solvent was evaporated under reduced pressure
to give a crude brownish oil that was chromotographed on
silica gel with hexane to give 0.08 g of 8 as colorless crystals
Con clu sion s
1
(12%, mp 99-101 °C). H NMR (CDCl3) δ -0.08 (s, 6 H), 0.94
Diradical 4 was found not to be a robustly exchange-
coupled system. Instead, it showed an unusual thermal
variation in exchange behavior, apparently governed by
conformational effects on exchange strength. The sum-
mary of evidence seems most consistent with torsion of
the tert-butylnitroxide group as causing the exchange
variation. The unusual stability of the diradical and its
reversible spectral line shape behavior is consistent with
4 exhibiting conformationally controlled exchange effects.
These findings demonstrate that fine-tuning is possible
for exchange behavior in open-shell systems with small
ground-to-excited state energy splittings, especially when
conformational effects are important relative to molecular
connectivity effects.
(s, 9 H), 1.12 (s, 18 H), 1.49 (s, 18 H), 5.21 (s, 1 H), 7.22-7.48
(m, 6 H). IR (KBr, cm-1) 3643 (phenolic OH). Anal. Calcd for
C
30H49NO2Si: C, 74.48; H, 10.21; N, 2.90. Found: C, 74.78; H,
10.16; N, 2.81.
3,5-Di-ter t-b u t yl-3′-(N-ter t-b u t yl-N-h yd r oxya m in o)-4-
h yd r oxybip h en yl (9). To a stirred solution of 15 mg (0.031
mmol) of 8 in 1 mL of THF was added 0.15 mL of a 1.0 M
solution of tetrabutylammonium fluoride in THF. The solution
was stirred for 24 h under argon at about 50 °C, following
which an additional 0.15 mL of the tetrabutylammonium
fluoride solution was added and the reaction stirred for another
24 h of stirring. Next, an aqueous saturated ammonium
chloride solution was added. The organic layer was separated
and washed with water. The aqueous layer was extracted with
diethyl ether (5 mL × 3), and the organic layers were
(31) Shultz, D. A.; Farmer, G. T. J . Org. Chem. 1998, 63, 6254.
(32) Stowell, J . C. J . Org. Chem. 1971, 36, 3055.

5182 J . Org. Chem., Vol. 64, No. 14, 1999
Liao et al.
combined. The organic solvent was evaporated under reduced
pressure, and the residue was purified by preparative thin
layer chromotography to give 9 mg of 9 as colorless crystals
(78%, mp 104-105 °C) that turned slightly orange if allowed
to stand in air. 1H NMR (CDCl3) δ 1.19 (s, 9 H), 1.49 (s, 18 H),
5.22 (s, 1 H), 5.61 (s, 1 H), 7.20-7.40 (m, 6 H). IR (KBr disk,
cm-1) 3440 (N-OH), 3630 (phenolic OH). Anal. Calcd for
into the reaction vessel. Next, 80 mg of sodium was carefully
added to 2 mL of anhydrous ethanol (CAUTION: FIRE
HAZARD), and the resultant sodium ethoxide solution was
cooled to room temperature and added to the reaction vessel.
The reaction was stirred at room temperature under nitrogen
overnight (turns black). The precipitates were filtered away,
and the filtrate was concentrated by rotary evaporation to give
a black oil, which was purified by column chromatography
(silica gel, 9:1 hexane:ethyl acetate). The resultant yellow oil
was allowed to stand overnight and crystallized. Recrystalli-
zation from hexane gave colorless needles (mp 126-128 °C,
C
24H35NO2: C, 78.00; H, 9.55; N, 3.78. Found: C, 78.01; H,
9.54; N, 3.67.
2,6-Di-ter t-bu tyla n isole-4-bor on ic Acid (10). 4-Bromo-
2,6-di-tert-butylanisole33 (6.0 g, 20 mmol) was dissolved in 40
mL of dry THF under argon and cooled to -78 °C. Next, 23.5
mL (40 mmol, 1.7 M in hexanes) of tert-butyllithium was added
dropwise by syringe, and the reaction stirred for 2 h at -78
°C. Triisopropyl borate (4.4 mL, 20 mmol) was added by
syringe, and the solution was allowed to warm slowly to room
temperature. Dilute aq HCl was added until the aqueous layer
was acidic. The layers were separated, and the aqueous layer
was extracted with 2 × 20 mL of dichloromethane. The
combined organic layers were extracted with 8 × 10 mL of 1
N aq NaOH. The combined aqueous layers were then acidified
by slow addition of 2 N aq HCl. A white precipitate formed
and was compound 10 was collected by filtration. Sometimes
this product was obtained as a mixture of the acid and the
cyclic “anhydride” boroxin form20ssuch mixtures are accept-
able for use in the next step, synthesis of 4 by Suzuki coupling.
Yield: 3.6 g (68%), mp 295-297 °C. 1H NMR (200 MHz,
CDCl3): δ 1.52 (s, 18 H), 3.76 (s, 3 H), 8.16 (s, 2 H).
3,5-Di-ter t-b u t yl-3′-(N-ter t-b u t yl-N-a m in oxy)-4-oxyb i-
p h en yl (4). To a stirred solution of 5.0 mg (0.0135 mmol) of 9
in 1 mL of degassed toluene was added 9.3 mg (0.0675 mmol)
of freshly prepared lead dioxide34 at room temperature. After
40 min of stirring under argon, the brownish-green suspension
was allowed to settle, and aliquots could be removed by pipet
for spectroscopic study. UV-vis (toluene, room temp, nm): 390
(strong), 490, 576 (broad). ESR (toluene, 9.430 GHz, modula-
tion frequency ) 100 kHz, hfc derived by simulation with
WINSIM): nitroxide-derived portion of spectrum, aN ) 11.98
G, a1 ) 1.97, 2.09, 2.09 (protons ortho, para to nitroxide), a2 )
0.87 (proton meta to nitroxide), simple linewideth ) 0.33 G,
spectrum is 47.8% of total simulated area; phenoxyl-derived
portion of spectrum, a1 ) 1.73, (2 H, 3,5-protons on phenoxyl
ring), a2 ) 1.74, 1.80, 1.81 (phenoxyl hfc to 2′, 4′, 6′ protons on
nitroxide ring), a3 ) 0.73 (phenoxyl hfc to 5′ proton on nitroxide
ring), simple line width ) 0.36 G, spectrum is 52.2% of total
simulated area. See Figure 5 for comparison of experiment and
simulation by the alternate BIRADG method.22 Both the UV-
vis and ESR spectra were persistent for at least two months
of storage at -30 °C, with almost no loss of spectral intensity
in the UV-vis.
1
0.13 g, 19%). H NMR (CDCl3) δ -0.08 (s, 6 H), 0.95 (s, 9 H),
1.12 (s, 9 H), 1.48 (s, 18 H), 3.74 (s, 3 H), 7.26 (m, 4 H), 7.45
(s, 2 H).
3′-[N-ter t-Bu t yl-N-a m in oxy]-3,5-d i-ter t-b u t yl-4-m et h -
oxybip h en yl (13). To a solution of 3,5-di-tert-butyl-3′-[N-tert-
butyl-N-(tert-butyldimethylsiloxy)amino]-4-methoxybiphenyl
(11, 4.1 mg, 0.008 mmol) in 1.5 mL of benzene was added 0.4
mL of 1 N tetrabutylammonium fluoride in tetrahydrofuran
(Aldrich). The reaction was stirred at room temperature for
48 h, until no starting material was detected by thin-layer
chromatography. The solution of 12 was passed through a plug
of silica gel and immediately used for the subsequent oxidation.
The solution of 12 in 1 mL of degassed benzene was added
excess freshly prepared lead dioxide at room temperature.
Within 30 min, the orange-red suspension was allowed to settle
and was used as is for spectroscopic study by withdrawing
aliquots by pipet. UV-vis (benzene) shows a long, tailing
absorption from 200 nm out into the visible region. ESR
(toluene, 9.802 GHz, modulation amplitude ) 0.031, modula-
tion frequency ) 12.5 kHz, hfc derived by simulation with
WINSIM): aN ) 12.48 G, a1 ) 1.79, 1.81, 1.96 (protons ortho,
para to nitroxide), a2 ) 0.80 (proton meta to nitroxide), simple
line width ) 0.49 G, correlation coefficient of fit is 0.993.
3,5-Di-ter t-b u t yl-3′-[N-ter t-b u t yl-N-(ter t-b u t yld im et h -
ylsiloxy)a m in o]-4-oxybip h en yl (14). To a stirred solution
of 6 mg (0.01 mmol) of 8 in 1 mL of degassed benzene was
added an excess of recently prepared lead dioxide at room
temperature. Almost at once a purple suspension formed and
persisted. The suspension was allowed to settle and was used
as is for spectroscopic studied by withdrawing aliquots by
pipet. UV-vis (benzene, nm, λmax): 481, 514, 554. ESR
(toluene, 9.762 GHz, modulation amplitude ) 0.031 G, modu-
lation frequency ) 12.5 kHz, hfc derived by simulation with
WINSIM): a1 ) 1.74, (2 H, 3,5-protons on phenoxyl ring), a2
) 1.62, 1.63, 1.87 (phenoxyl hfc to 2′, 4′, 6′ protons on nitroxide
ring), a3 ) 0.65 (phenoxyl hfc to 5′ proton on nitroxide ring),
simple line width ) 0.37 G, correlation coefficient of fit is 0.996.
Ack n ow led gm en t. Support of this work by the
National Science Foundation (CHE-9521594 and CHE-
9809548) is gratefully acknowledged. The opinions
expressed in this paper are solely those of the authors
and not necessarily those of the Foundation.
3,5-Di-ter t-b u t yl-3′-[N-ter t-b u t yl-N-a m in oxy]-4-m et h -
oxybip h en yl (11). A solution of 0.37 g (1.4 mmol) of 10 and
0.5 g (1.4 mmol) of 6 was prepared in 2.5 mL of degassed THF.
A catalytic solution of 15.7 mg (0.07 mmol) of Pd(OAc)2 plus
12 mg (0.06 mmol) of tributylphosphine was separately
prepared in a similar amount of THF and then transferred
Su p p or tin g In for m a tion Ava ila ble: General spectral
1
procedures, H NMR spectra for compounds 5, 6, and 10, and
(33) For a preparation of 4-bromo-2,6-di-tert-butylanisole, see Dhami,
K. S.; Stothers, J .B. Can. J . Chem. 1966, 44, 2855.
(34) Fresh PbO2 was prepared by addition of lead tetraacetate to
distilled water, filtration, washing of the product with distilled water
to neutral pH, and drying in air to give a very dark brown powder of
suitable quality for oxidation.
descriptions of the hyperfine coupling data for radicals 13-
14 and biradical 4. This material is available free of charge
via the Internet at http://pubs.acs.org.
J O990317L