Diarylnitroxide diradicals: Low-temperature oxidation of diarylamines to nitroxides

Article abstract of DOI:10.1021/ja8016335

A low temperature method, in which the progress of the oxidation of secondary diarylamines with DMDO at low temperatures is monitored by magnetic resonance spectroscopy (EPR and NMR) and magnetic studies by a Superconducting Quantum Interference Device (SQUID), is developed for preparation of the first m-phenylene based diarylnitroxide diradical. EPR spectroscopy and magnetic studies (SQUID) indicate that the diradical in the dichloromethane matrix predominantly adopts anticonformation (2A-anti) and possesses triplet ground state. Similar oxidation experiments for conformationally constrained aza[1 ₄]metacyclophane provide evidence for the formation of small amounts of the corresponding diarylnitroxide diradical. Both diarylnitroxide diradicals could only be detected at low temperatures (-80°C and below).

Full text of DOI:10.1021/ja8016335

Published on Web 06/25/2008
Diarylnitroxide Diradicals: Low-Temperature Oxidation of
Diarylamines to Nitroxides
Andrzej Rajca,* Matthew Vale, and Suchada Rajca
Department of Chemistry, UniVersity of Nebraska, Lincoln, Nebraska 68588-0304
Received March 4, 2008; E-mail: arajca1@unl.edu
Abstract: A low temperature method, in which the progress of the oxidation of secondary diarylamines
with DMDO at low temperatures is monitored by magnetic resonance spectroscopy (EPR and NMR) and
magnetic studies by a Superconducting Quantum Interference Device (SQUID), is developed for preparation
of the first m-phenylene based diarylnitroxide diradical. EPR spectroscopy and magnetic studies (SQUID)
indicate that the diradical in the dichloromethane matrix predominantly adopts anticonformation (2A-anti)
and possesses triplet ground state. Similar oxidation experiments for conformationally constrained
aza[14]metacyclophane provide evidence for the formation of small amounts of the corresponding
diarylnitroxide diradical. Both diarylnitroxide diradicals could only be detected at low temperatures (-80 °C
and below).
1. Introduction
by the intrinsic weakness of intermolecular interactions between
the neutral radicals.7,8,11
Nitroxides are among the most widely used stable radicals
in chemistry and biology.1–4 One of the important applications
is in materials chemistry, in which nitroxides have been used
as building blocks for organic magnetic materials.5,6 The general
designs of such organic magnets have been based upon
intermolecular ferromagnetic interactions between the radicals
in the solid state. The first organic ferromagnet, ꢀ-polymorph
of the 4-nitrophenyl nitronyl nitroxide, was reported by Ki-
noshita and co-workers in 1991,7 and there have been several
reports of organic ferromagnets since then.8–10 It is important
to note that the nitroxide-based ferromagnets have magnetic
ordering at very low temperatures, below 2 K, which is limited
Another approach to organic materials with magnetic ordering
has been based on macromolecules with intramolecular ferro-
magnetic interactions between organic radicals (high-spin
polyradicals).12–14 In this approach, the intrinsic strength of
intramolecular ferromagnetic interactions mediated through a
cross-conjugated π-system shows promise for magnetic ordering
at relatively high temperatures, significantly exceeding the
temperature of about 10 K.15 An example of this approach is
the first organic polymer with magnetic ordering reported in
2001.16 Because m-phenylene has been shown to be one of the
most effective spin coupling units,17–19 the building blocks
consisting of nitroxides coupled through the m-phenylenes are
highly promising for preparation of organic magnetic materi-
als.20
(1) Hawker, C. J.; Bosman, A. W.; Harth, E. Chem. ReV. 2001, 101, 3661–
3688.
(2) Shibuya, M.; Tomizawa, M.; Suzuki, I.; Iwabuchi, Y. J. Am. Chem.
Soc. 2006, 128, 8412–8413.
To date, there are only relatively few known stable diarylni-
troxide radicals in which nitroxide is connected to two aryl
(3) Chernick, E. T.; Mi, Q.; Kelley, R. F.; Weiss, E. A.; Jones, B. A.;
Marks, T. J.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc.
2006, 128, 4356–4364.
(4) Soule, B. P.; Hyodo, F.; Matsumoto, K.-I.; Simone, N. L.; Cook, J. A.;
Krishna, M. C.; Mitchell, J. B. Free Radical Biol. Med. 2007, 42,
1632–1650.
(11) Intermolecular antiferromagnetic interactions, which can be relatively
strong, can lead to magnetic ordering in the presence of significant
magnetic anisotropy at temperatures as high as 36 K but providing
materials with very small magnetic moments (“weak ferromagnets”)
Banister, A. J.; Bricklebank, N.; Lavender, I.; Rawson, J. M.; Gregory,
C. I.; Tanner, B. K.; Clegg, W.; Elsegood, M. R. J.; Palacio, F. Angew.
Chem., Int. Ed. Engl. 1996, 35, 2533–2535.
(5) Kinoshita, M. Magnetism of Stable Organic Radical Crystals. In
Handbook of Organic ConductiVe Molecules and Polymers; Nalwa,
H. S., Ed.; Wiley: New York, 1997; Chapter 15, pp 781-800.
(6) Itoh, K., Kinoshita, M. Eds.; Molecular Magnetism, New Magnetic
Materials; Gordon and Breach: Amsterdam, 2000; pp 1-337.
(7) Tamura, M.; Nakazawa, Y.; Shiomi, D.; Nozawa, K.; Hosokoshi, Y.;
Ishikawa, M.; Takahashi, M.; Kinoshita, M. Chem. Phys. Lett. 1991,
186, 401–403.
(12) Rajca, A. Chem. ReV. 1994, 94, 871–893.
(13) Rajca, A.; Wongsriratanakul, J.; Rajca, S. J. Am. Chem. Soc. 2004,
126, 6608–6626.
(14) Fukuzaki, E.; Nishide, H. J. Am. Chem. Soc. 2006, 128, 996–1001.
(15) Rajca, A. Chem.-Eur. J. 2002, 8, 4834–4841.
(16) Rajca, A.; Wongsriratanakul, J.; Rajca, S. Science 2001, 294, 1503–
1505.
(8) Chiarelli, R.; Novak, M. A.; Rassat, A.; Tholence, J. L. Nature 1993,
363, 147–149.
(9) (a) Fullerene radical cation salt is a ferromagnet with magnetic ordering
at 16 K. Allemand, P.-M.; Khemani, K. C.; Koch, A.; Wudl, F.;
Holczer, K.; Donovan, S.; Gruner, G.; Thompson, J. D. Science 1991,
253, 301–303. (b) Narymbetov, B.; Omerzu, A.; Kabanov, V. V.;
Tokumoto, M.; Kobayashi, H.; Mihailovic, D. Nature 2000, 407, 883–
885.
(17) Fang, S.; Lee, M.-S.; Hrovat, D. A.; Borden, W. A. J. Am. Chem.
Soc. 1995, 117, 6727–6731.
(18) Wenthold, P. G.; Kim, J. B.; Lineberger, W. C. J. Am. Chem. Soc.
1997, 119, 1354–1359.
(19) Wang, T.; Krylov, A. I. J. Chem. Phys. 2005, 123, 104304-1104304-
6.
(10) Alberola, A.; Less, R. J.; Pask, C. M.; Rawson, J. M.; Palacio, F.;
Oliete, P.; Paulsen, C.; Yamaguchi, A.; Farley, R. D.; Murphy, D. M.
Angew. Chem., Int. Ed. 2003, 42, 4782–4785.
(20) Rajca, A.; Takahashi, M.; Pink, M.; Spagnol, G.; Rajca, S. J. Am.
Chem. Soc. 2007, 129, 10159–10170.
9
10.1021/ja8016335 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 9099–9105 9099

A R T I C L E S
Rajca et al.
Figure 2. Diarylnitroxide radicals.
Figure 1. Diarylnitroxide radicals and secondary diarylamines.
groups.21–24 Delocalization of spin density to aryl groups allows,
in principle, for the design of extended high-spin polyradicals
based upon cross-conjugation of nitroxide radicals.15 On the
basis of this design, the diarylnitroxide moiety was connected
to two tert-butylnitroxides (Figure 1), nitronyl nitroxides, and
imino nitroxides, to provide quartet (S ) 3/2) ground-state
nitroxide triradicals.25,26 Although steric shielding of spin density
by the tert-butyl groups at the meta positions with respect to
the nitroxides is sufficient for stability of triradicals at ambient
conditions, there is no report of polyradicals based on this
approach.25,26 A polymer of diarylnitroxide radicals was re-
ported, though the reported concentration of radicals was modest
(Figure 1).27 In this polymer, nitroxides were cross-conjugated
by terphenylene groups, which provided a long exchange pathway
leading to a very weak ferromagnetic exchange coupling.27
Figure 3. SQUID magnetometry for polycrystalline nitroxide radical 1,
crystallized from ethanol. Main plot: M/Msat vs H/T at set temperatures of
1.8, 3, and 5 K, corresponding to S ) 0.85, 0.80, and 0.67, respectively.
Inset plot: ꢁT vs T in the warming and cooling modes at set magnetic fields
of 5000 and 500 Oe.
successful preparation of such diarylnitroxide diradicals or
polyradicals.30
Recently, we reported the synthesis of a series of
aza[1n]metacyclophanes that are macrocyclic oligomers of
poly(m-aniline) (Figure 1).31 These structures were designed as
potential building blocks for robust high-spin nitroxide polyradi-
cals, in which we expected that the tert-butyl groups at the meta
positions with respect to the nitroxides would provide adequate
shielding to the nitroxides, analogously to those in the diarylni-
troxide moiety of stable S ) 3/2 nitroxide triradicals. Assessment
of the effectiveness of this structure design will require an
effective methodology for the generation of nitroxide polyradi-
cals derived from secondary diarylpolyamines.
Herein, we report the development of a methodology for the
generation of diarylnitroxide radicals from secondary diary-
lamines. We first prepared nitroxide radical 122 from the
corresponding diarylamine following known procedures for
oxidation of secondary amines with meta-chloroperbenzoic acid
(m-CPBA)21 and for oxidation of secondary dialkylamines with
dimethyldioxirane (DMDO) to the corresponding nitroxides.32
Radical 1 is isolated as crystalline solid, stable at ambient
conditions, and showing intermolecular interactions at low
temperatures. However, these oxidation methods failed to
It is well-established that cross-conjugation of nitroxides
through m-phenylene could lead to a strong ferromagnetic
coupling, exceeding RT at room temperature.17,20,28 Furthermore,
it may be postulated that such nitroxide polyradicals could be
prepared from poly(m-aniline) or its oligomers as their oxidation
products (Figure 1);29 however, there is no report to date of the
(21) Forrester, A. R. , Hay, J. M. , Thomson, R. H. Organic Chemistry of
Stable Free Radicals; Academic Press: London, 1968; Chapter 5, pp
180-246.
(22) (a) Ivanov, Y. A.; Kokorin, A. I.; Shapiro, A. B.; Rozantsev, E. G.
IzV. Akad. Nauk SSSR, Ser. Khim. 1976, 10, 2217–2222. (b) In the
original synthesis of radical 1, the yield of 70% from the corresponding
amine, using Na2WO4/H2O2, was reported, though purity of 1 was
not established.
(23) Delen, Z.; Lahti, P. M. J. Org. Chem. 2006, 71, 9341–9347.
(24) Several examples of diarylnitroxides with ferromagnetic exchange
coupling in the solid state were reported, and the strength of coupling
was generally weak. Seino, M.; Akui, Y.; Ishida, T.; Nogami, T. Synth.
Met. 2003, 133-134, 581–583.
(25) Ishida, T.; Iwamura, H. J. Am. Chem. Soc. 1991, 113, 4238–4241.
(26) Tanaka, M.; Matsuda, K.; Itoh, T.; Iwamura, H. J. Am. Chem. Soc.
1998, 120, 7168–7173.
(27) Oka, H.; Kouno, H.; Tanaka, H. J. Mater. Chem. 2007, 17, 1209–
1215.
(30) Miura, Y. Synthesis and Properties of Organic Conjugated Polyradicals.
In Magnetic Properties of Organic Materials; Lahti, P. M., Ed.; Marcel
Dekker: New York, 1999; pp 267-284.
(28) Sato, H.; Kathirvelu, V.; Spagnol, G.; Rajca, S.; Rajca, A.; Eaton,
S. S.; Eaton, G. R. J. Phys. Chem. B 2008, 112, 2818–2828.
(29) Baumgarten, M.; Mu¨llen, K.; Tyutyulkov, N.; Madjarova, G. Chem.
Phys. 1993, 169, 81–84.
(31) Vale, M.; Pink, M.; Rajca, S.; Rajca, A. J. Org. Chem. 2008, 73, 27–
35.
(32) Murray, R. W.; Singh, M. Tetrahedron Lett. 1988, 29, 4677–4680.
9
9100 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Diarylnitroxide Diradicals
A R T I C L E S
generate diarylnitroxide diradical 2. To solve this problem, we
developed an oxidation method in which the progress of the
oxidation of secondary diarylamines with DMDO is monitored
1
by magnetic resonance spectroscopy (EPR and H NMR) and
by magnetic studies in solution using a Superconducting
Quantum Interference Device (SQUID), with the reaction
mixture strictly maintained at low temperature (-80 °C or
below). This allowed us to optimize the reaction conditions.
Using this methodology, we detected for the first time an
m-phenylene-based diarylnitroxide diradical such as 2, and
establish that this diarylnitroxide is persistent only at low
temperatures (-80 °C or below). However, when this method
was tested for the oxidation of aza[1n]metacyclophane (n ) 4)
to tetraradical 3 (Figure 2), only the corresponding nitroxide
diradical was detected.
2. Results and Discussion
2.1. Oxidation of Diarylmonoamine. Oxidation of diarylamine
4 with meta-chloroperbenzoic acid (m-CPBA) or dimethyldiox-
irane (DMDO) produces diarylnitroxide radical 1 in 40-60%
isolated yields (eq 1).22b
Figure 4. SQUID magnetometry for ∼0.1 M nitroxide radical 1 in
2-MeTHF. Main plot: M/Msat vs H/T at set temperatures of 1.8, 3, and 5 K.
Inset plot: ꢁT vs T in the warming and cooling modes at set magnetic field
of 5000 Oe.
For polycrystalline diarylnitroxide radical 1, the plot of the
product of magnetic susceptibility and temperature (ꢁT) versus
T is flat at higher temperatures (T > 25 K) with ꢁT ) 0.35 emu
K mol-1, which is nearly identical to the value of S(S + 1)/2
) 0.375 emu K mol-1 for an S ) 1/2 paramagnetic radical
(Figure 3). Interestingly, the ꢁT versus T plot shows an upward
turn at low temperature, with ꢁT ) 0.42 emu K mol-1 at 1.8
K, that is assigned to intermolecular ferromagnetic interactions.24
The ferromagnetic nature of these interactions is confirmed by
the increasing values of S > 1/2 for the Brillouin plots, M/Msat
versus H/T, at decreasing T ) 5, 3, and 1.8 K.
For 1 in 2-methyltetrahydrofuran (2-MeTHF), a near perfect
S ) 1/2 paramagnetic behavior is found, thus confirming
intermolecular nature of the ferromagnetic interaction of poly-
crystalline radical. The ꢁT versus T plot is flat in the T ) 2-200
K range and the M/Msat versus H/T plot coincides with the S )
1/2 Brillouin curve. Both values of ꢁT ) 0.37 emu K mol-1
and Msat ) 0.97 µB are nearly identical to those expected for
an S ) 1/2 radical (Figure 4).
2.2. Oxidation of Diaryldiamine and Aza[14]metacyclophane.
Oxidations of diaryldiamine 5 and aza[14]metacyclophane 6
(Chart 1) with m-CPBA or DMDO, using similar conditions to
those that gave good yields of diarylnitroxide radical 1, resulted
in complex mixtures, with no evidence for the corresponding
diradical or the tetraradical. To solve this problem, we developed
a method for oxidation of diarylamines to nitroxides at low
temperature, with the EPR, EPR/SQUID/EPR,37,38 and 1H
NMR/EPR monitoring of the reaction mixtures.
2.2.1. EPR Monitoring. First, we determined that oxidation
of diarylamine 4 with DMDO (2-2.5 equiv) in dichloromethane
at -95 or -78 °C gives intense EPR spectra for nitroxide
monoradical at 140 K (Figures S4A-C, Supporting Informa-
tion). Using similar conditions (2-2.5 equiv of DMDO per NH
moiety), oxidation of diaryldiamine 5 and aza[14]meta-
These yields are significantly lower than the nearly quantita-
tive yields reported for oxidation of sterically hindered secondary
dialkylamines with DMDO to the corresponding dialkylnitroxide
radicals such as TEMPO.32
EPR spectrum of dilute solution of 1 in diethyl ether or in
toluene shows the expected hyperfine splittings, including 14N-
hyperfine splitting, |aN| ) 0.97 mT (Figures S1A and S1B,
Supporting Information).22a 1H NMR spectrum of concentrated
solution of 1 in chloroform-d shows a broad singlet at about
6.2 ppm, which is assigned to the tert-butyl groups (Figure S2,
Supporting Information). This chemical shift corresponds to a
paramagnetic shift, ∆δp ) +4.9 ppm for protons of the tert-
butyl groups.33 The positive sign of ∆δp may be ascribed to
the conjugative mechanism (“homohyperconjugation”) for the
transfer of positive spin density from para positions of the
phenylnitroxide to the tert-butyl hydrogens, as observed for
4-alkylphenyl-tert-butylnitroxide radicals; the value of isotropic
electron-proton hyperfine splitting (aH in mT) for tert-butyl
hydrogens, aH ≈ +0.006 mT,20,34 is similar to those reported
in 4-tert-butylphenyl-tert-butylnitroxide and 4-tert-butylphenyl-
substituted radicals.35,36 The integration indicates about 10 mol%
content of solvent used for crystallization and negligible amount
of other diamagnetic byproduct. The purity of radical 1 is
confirmed by magnetic studies using a SQUID.
(33) As a small correction for bulk paramagnetic susceptibility of the order
of 0.5 ppm is not included, the actual value of ∆δp is somewhat lower.
(34) For S ) 1/2 nitroxide radical (g ) 2.005) at room temperature (T )
295 K): aH (mT) ) 0.00133 ∆δp.
(35) (a) The values of aH ) +0.0054 mT and aH ) +0.0059 mT for the
protons of the tert-butyl group were obtained from the 1H NMR spectra
of 4-tert-butylphenyl-tert-butylnitroxide radical in chloroform-d and
carbon tetrachloride. (b) Torsell, K.; Goldman, J.; Petersen, T. E.
Liebigs Ann. Chem. 1973, 231–240. (c) Forrester, A. R.; Hepburn,
S. P.; McConnachie, G. J. Chem. Soc., Perkin Trans. I 1974, 2213–
2219.
(36) In polyarylmethyl triradical, the 1H-resonance at +10 ppm was
assigned to the t-Bu groups in the 4-tert-butylphenyls adjacent to the
radicals, suggesting analogous homohyperconjugation: Rajca, S.; Rajca,
A. J. Am. Chem. Soc., 1995, 117, 9172-9179.
(37) Rajca, A.; Shiraishi, K.; Pink, M.; Rajca, S. J. Am. Chem. Soc. 2007,
129, 7232–7233.
(38) Rajca, A.; Shiraishi, K.; Vale, M.; Han, H.; Rajca, S. J. Am. Chem.
Soc. 2005, 127, 9014–9020.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 9101

A R T I C L E S
Rajca et al.
Chart 2. Structure of alkylarylnitroxide diradical 7
appears as a pentuplet with significant splitting of |Ayy/2hc| ≈
8.5 × 10-4 cm-1 and the corresponding value of gy ) 2.0028
is similar to the free-electron value of 2.0023. These spectral
features suggest that the nitroxide moieties are approximately
coplanar with the m-phenylene, or perhaps only slightly distorted
from coplanarity.20
For the minor conformer, only the outermost bands may be
discerned, thus allowing only for determination of approximate
value of |D/hc| ≈ 11 × 10-3 cm-1. As mentioned in the
preceding paragraphs, this relatively large value of |2D/hc|
suggests smaller interspin distance in the minor conformer.
The side bands in EPR spectra for the reaction mixture
obtained by oxidation of aza[14]metacyclophane 6 are very
weak, and thus it is most likely that they would correspond an
S ) 1 diradical. The value of 2D is similar to that for the major
conformer of 2. Spectral simulations of the |∆ms| ) 1 region,
which is not overlapped with the intense S ) 1/2 resonances,
reproduce the experimental spectra with |D/hc| ) 8.9 × 10-3
cm-1 and |E/hc| ) 1.15 × 10-3 cm-1 (Figures S6A and S7,
Supporting Information).
Figure 5. EPR (X-band, ν ) 9.4875 GHz, dichloromethane, 140 K) spectra
for the reaction mixture obtained by oxidation of diaryldiamine 5 with
DMDO at -90 °C. The simulation parameters for the S ) 1 state are: |D/
hc| ) 8.6 × 10-3 cm-1, |E/hc| ) 1.3 × 10-3 cm-1, |Ayy/2hc| ) 8.5 × 10-4
cm-1, gx ≈ 2.005, gy ) 2.0028, gz ) 2.0072, Gaussian line (Lx ) 0.8, Ly
) 0.65, Lz ) 1.0 mT). The center lines correspond to an S ) 1/2 product.
Chart 1. Structures of diaryldiamine 5 and aza[14]metacyclophane 6
In diradical 2, with approximately planar conformation, the
spin density is likely to be significantly delocalized from the
nitroxide moieties onto the m-phenylene and 4-tert-butylphenyls.
This delocalization should lead to the relatively low value of
14N-hyperfine coupling for 2. In particular, the value of |Ayy/
hc| ≈ 0.0017 cm-1 obtained from the spectral simulations in
Figure 5 may be viewed as the lower limit for the largest
principal value of the 14N-hyperfine tensor.42 The value of |Ayy/
hc| ≈ 0.0017 cm-1 for delocalized diarylnitroxide diradical 2
is significantly lower than the largest principal values of the
14N-hyperfine tensor for planar alkylarylnitroxide diradical 7
(0.0024 cm-1) (Chart 2),20,43 tert-butylarylnitroxide diradical
(∼0.0026 cm-1),44 and for localized di-tert-butylnitroxide
radical (0.0030 cm-1).45
cyclophane 6 is monitored by EPR spectroscopy. The EPR
spectra of frozen reaction mixtures derived from 5 and 6 show
triplet-like center resonances of nitroxide monoradical that are
flanked by weak side bands (Figure 5 and Figure S7, Supporting
Information).
The EPR spectra for the reaction mixture obtained by
oxidation of diaryldiamine 5 are better resolved. The side bands
in EPR spectra at 140 K could be assigned to dipolar coupling
patterns of two S ) 1 diradicals 2, presumably one major and
one minor conformer, with smaller and larger spectral widths
(|2D/hc|), respectively (Figure 5);28,39,40 based upon the relative
values of |2D/hc|, the major and minor conformer are expected
to possess longer and shorter interspin distances between
nitroxides, respectively.41
These EPR spectral analyses establish the presence of triplet
state for diradical 2. To determine whether the triplet state is
the ground-state or the excited state, and to estimate concentra-
tion of the paramagnetic radicals, SQUID magnetic measure-
ments were carried out.
2.2.2. EPR/SQUID/EPR Monitoring. Another set of oxida-
tions of diaryldiamine 5 and aza[14]metacyclophane 6 was
carried out in homemade tubes that allow for handling of
samples at strictly low temperature and direct EPR/SQUID/EPR
monitoring of the reaction mixtures (Figures S10A-C and
S11A-C, Supporting Information). DMDO was reacted with
5 and 6 at about -80 and -90 °C, respectively; EPR spectra
For the major conformer of diradical 2, the part of the
experimental spectrum in the |∆ms| ) 1 region that is not
obstructed by the intense S ) 1/2 resonances can be reproduced
by the simulated powder pattern for electron-electron dipolar
couplingand14N-hyperfinecoupling(Figure5).Theelectron-electron
dipolar coupling gives rise to two zero-field splitting (zfs)
parameters: |D/hc| ) 8.6 × 10-3 cm-1 and |E/hc| ) 1.3 × 10-3
cm-1; the value of |E| is a significant fraction of |2D|, thus
indicating large deviation of spin density distribution from axial
symmetry. The yy-component of the 14N-hyperfine coupling
(42) In the spectral simulation of 2 in Figure 5, it is assumed that the
magnetic dipole-dipole inter-spin vector (principal axis of the diradical
D-tensor) is along a principal axis of the nitroxide g- and A-tensor.
Out-of-plane distortion in 2 may misalign the principal axes, and thus
the value of “|Ayy/hc|” with the y-axis defined by the D-tensor in the
spectral simulation would be lower, underestimating the actual value
of |Ayy/hc| with the y-axis parallel to the 2pN orbital axis.
(43) Rassat, A.; Sieveking, U. Angew. Chem., Int. Ed. Engl. 1972, 11, 303–
304.
(39) Calder, A.; Forrester, A. R.; James, P. G.; Luckhurst, G. R. J. Am.
Chem. Soc. 1969, 91, 3724–3727.
(40) Spagnol, G.; Shiraishi, K.; Rajca, S.; Rajca, A. Chem. Commun. 2005,
5047–5049.
(41) (a) The dependence of the spectral width (|2D|) on the inter-dipole
distance (r) is calculated according to the following equation: |2D| )
3gꢀ/r3 ) 2.78 × 103 (g/r3), where |D| is in mT and r is in Å. (b)
Eaton, S. S.; More, K. M.; Sawant, B. M.; Eaton, G. R. J. Am. Chem.
Soc. 1983, 105, 6560–6567.
(44) Griffith, O. H.; Cornell, D. W.; McConnell, H. M. J. Chem. Phys.
1965, 43, 2909–2910.
(45) Rajca, A.; Mukherjee, S.; Pink, M.; Rajca, S. J. Am. Chem. Soc. 2006,
118, 13497–13507.
9
9102 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Diarylnitroxide Diradicals
A R T I C L E S
similar to those obtained in the previous EPR monitoring
experiments were obtained. SQUID magnetometry showed that
the values of ꢁT and Msat are very low, indicating that the content
of paramagnetic radicals is on the order of 1%, based upon the
weight of the starting amines. The magnetization data plotted
as M/Msat vs H/T follow Brillouin functions with the value of S
) 1/2 - 1, as expected for a mixture of S ) 1/2 monoradical
and S ) 1 ground-state diradical. These magnetic data are
consistent with the EPR spectra, which showed dominant center
peaks (S ) 1/2 monoradical) and much smaller side bands (S
> 1/2 species, most likely, S ) 1 diradical). For diradical 2,
the relative intensities of the side bands and the center peaks in
the EPR spectra, taken before and after the SQUID magnetom-
etry, are in qualitative agreement with the average values S
obtained from the curvature of the M/Msat vs H/T Brillouin
plots.46 In magnetic studies, the measurement of 1% content of
paramagnetic radicals with the overwhelming diamagnetic
component and organic solvent matrix is very difficult. There-
fore, the similar molar ratio of the S ) 1 diradical to S ) 1/2
monoradical determined by EPR spectra and SQUID magne-
tometry is significant, as it precludes a possible contribution
from paramagnetic metal impurities. Thus, diarylnitroxide
diradical 2 possesses triplet ground state. However, the ground-
state for the diradicals derived from aza[14]metacyclophane 6
could not be established, due to a relatively weak EPR signal
for side bands.
Figure 6. Structures for the nitroxide monoradicals 2-m and 3-m (drawing
in black); the remaining parts of the structure, derived from diaryldiamine
5 and aza[14]metacyclophane 6 (drawn in green), are not detected in the
EPR spectra and thus details of their structure are not known. The 14N-
and H-hyperfine splittings (mT) are shown in red and blue, respectively.
EPR spectra and spectral simulations are reported in Figures S8H and S9E
in the Supporting Information.
1
of three limiting conformers of diradical 2 are fully optimized
at the UB3LYP/6-31G(d) level (Table 1).47 Vibrational fre-
quency calculations indicate that all three conformers are local
minima on the potential energy surface. In all three conformers,
the torsional angles (R and ꢀ, or 180 - R and 180 - ꢀ) between
the nitroxide and m-phenylene moieties are in the 25-29° range.
All three conformers are triplet ground states with the
singlet-triplet energy gaps of 1.5-1.8 kcal/mol, as expected
for a moderately twisted m-phenylene-based nitroxide diradi-
cals.17
Because the diamagnetic component is dominant in oxidations
starting from 5 and 6, the 1H NMR/EPR monitoring experiments
of the reaction mixtures were carried out.
The triplet ground states of 2A-anti, 2B-endo, and 2C-exo
are within only 1.1 kcal mol-1, thus not allowing for a
conclusive assignment of the major conformer observed in the
EPR spectra in the condensed phase. For example, the relative
energies of 2A-anti, 2B-endo, and 2C-exo in polar matrices,
such as dichloromethane, may change significantly, as the
conformers with the higher energies possess considerably larger
dipole moments.39,40
A tentative assignment of the major conformer and the minor
conformer for diradical 2 is possible using the experimental
values of |2D/hc| (Figure 5) and their approximate relationship
with the interspin distance.41 Conformationally constrained
alkylarylnitroxide diradical 7 (Chart 2),20,43 which possesses
planar structure with an interspin distance of 4.8 Å between
the midpoints of the NO bonds, provides a well-characterized
reference structure for conformer 2B-endo. The measured value
of |2D/hc| ) 2.7 × 10-2 cm-1 for 7 corresponds to the interspin
distance of 5.8 Å; as expected for a diradical with delocalized
spin density, the point-dipole approximation overestimates
interspin distance, defined as average of N. . .N and O. . .O
distances. This overestimate is expected to exceed 1 Å for more
delocalized diarylnitroxide diradical 2. The observed values of
|2D/hc| ) 1.72 × 10-2 cm-1 and |2D/hc| ) 2.2 × 10-2 cm-1
for major and minor conformers correspond to the interspin
distances of 6.7 Å and 6.2 Å, respectively. These distances are
overestimated by 1.1 - 1.3 Å compared to the calculated
averages of N. . .N and O. . .O distances in 2A-anti and 2B-
2.2.3. 1H NMR/EPR Monitoring. Oxidations of diaryldiamine
5 and aza[14]metacyclophane 6 with DMDO were carried out
in flame sealed NMR sample tubes while the reaction progress
1
was monitored by H NMR and EPR spectroscopy (Figures
S12A-E and Figures S13A-H, Supporting Information). For
the reaction mixtures that are strictly handled at low temperature
(-78 °C or lower temperature) for 1-4 h, resonances that may
be assigned to the diamagnetic byproduct are observed. EPR
spectra of such reaction mixtures at 140 K are similar to those
observed for the reaction mixtures monitored by EPR and EPR/
SQUID/EPR, including the presence of the side-bands, as
described in the preceding paragraphs. When the reaction
mixtures are allowed to attain room temperature, the integrations
of the resonances that may be assigned to diamagnetic byproduct
are significantly increased. The EPR spectra of such mixtures
at 140 K show no side bands and only the center resonances
assigned to nitroxide monoradicals are detectable. The EPR
spectra at room temperature show resolved hyperfine patterns
that are in qualitative agreement with the 14N- and 1H-hyperfine
splittings that are expected for the nitroxide monoradicals 2-m
and 3-m, formed by oxidation of diaryldiamine 5 and
aza[14]metacyclophane 6, respectively (Figure 6).
2.2.4. Conformers of Nitroxide Diradical 2: DFT Cal-
culations and EPR Spectral Widths. Geometries for triplet states
(46) (a) For oxidations of diaryldiamine 5, the relative EPR signal intensities
from double integration of the side bands and the center peaks are
about 0.3: 0.7, respectively (Figures S5A and S5C, Supporting
Information). Because the side-bands underneath the center peaks could
not be integrated and the EPR intensities are related by a factor of
S(S + 1), the ratio of 0.3: 0.7 is probably not much different from the
molar ratio of S ) 1 diradical to S ) 1/2 monoradical. This ratio
would give spin-average of S close to Ss ≈ 0.7, which is in agreement
with S ) 1/2-1 obtained from the curvature of Brillouin plots (Figure
S5B, Supporting Information). (b) Ss is approximately related to the
curvature of Brillouin plots. Rajca, A.; Rajca, S.; Wongsriratanakul,
J. J. Am. Chem. Soc. 1999, 121, 6308–6309.
(47) Frisch, M. J.; et al. Gaussian 03; Gaussian, Inc.: Wallingford CT,
2004.
(48) Full geometry optimization of the broken-symmetry singlet for 2A-
anti lowers the energy by 0.08 kcal mol-1 only, thus the singlet
geometries of were not optimized. This may introduce an additional,
though small, source of overestimate of the singlet-triplet gaps.
(49) Zhang, D. Y.; Hrovat, D. A.; Abe, M.; Borden, W. T. J. Am. Chem.
Soc. 2003, 125, 12823–12828.
(50) (a) Cramer, C. J.; Smith, B. A. J. Phys. Chem. 1996, 100, 9664–
9670. (b) Winter, A. H.; Falvey, D. E.; Cramer, C. J.; Gherman, B. F. J.
Am. Chem. Soc. 2007, 129, 10113–10119.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 9103

A R T I C L E S
Rajca et al.
Table 1. UB3LYP/6-31G(d) Calculations for Conformers of Diradical 2
a Dipole moment in Debeye. b In Hartree/molecule. c In kcal mol-1
.
d Broken-symmetry singlet with S(S + 1) ≈ 1.0 at the geometry optimized for
the triplet.48–50 e Full geometry optimization of the broken-symmetry singlet for 2A-anti lowers the energy by 0.08 kcal mol-1 only. The UB3BLYP
(broken symmetry with S(S + 1) ≈ 1.0 singlets and S(S + 1) ≈ 2.0 triplets) level of theory usually overestimates the singlet-triplet gaps, 2(SEBS
-
TE),
for the triplet ground state organic diradicals, and undestimates such gaps for the singlet ground state diradicals.49,50
repeat reaction was carried was in acetone (5 mL), starting from
15.7 mg of amine 4. The combined crude orange solids (31.1 mg)
from the two reactions were purified by preparative TLC, followed
by recrystallization (as described above), to give deep orange
crystals (13.7 mg, 43%) of nitroxide radical 1. In another reaction,
57.9 mg (60%) of nitroxide radical 1 was obtained from 90.0 mg
of amine 4 in dichloromethane (10 mL) and 2.9 equiv of DMDO
at -20 °C, after column chromatography and recrystallization from
ethanol at 35 °C with cooling to either -20 or -78 °C. On the
basis of 1H NMR integration, the crystals of 1 contained about 10
mol % of solvent of crystallization (methanol or ethanol).
endo, respectively. An alternative assignment of the minor
conformers to 2A-anti and the major conformer to 2C-exo
would lead to an overestimate of about 0.6 - 0.7 Å, too low
for delocalized m-phenylene diarylnitroxide diradicals. There-
fore, we conclude that 2A-anti corresponds to the major
conformer of 2.
3. Experimental Section
3.1. Materials. Amine 4 and diamine 5 were prepared using
Pd-catalyzed aminations as described in the Supporting Information.
Aza[14]metacyclophane 6 was prepared using recently reported
procedure.30 Amines were oxidized to nitroxide radicals using
dimethyldioxirane (DMDO) or m-chloroperbenzoic acid (m-CPBA)
using the following procedures.
3.2. 3.2. 4,4-Di-tert-butyl-diphenylnitroxide Radical 1. m-
CPBA (22.4 mg, 0.130 mmol) in dry chloroform-d (1 mL) was
added to a stirred solution of amine 4(18.2 mg, 64.1 µmol) in dry
chloroform-d (2 mL) at 0 °C over 5 min under nitrogen atmosphere.
The reaction mixture turned orange immediately and was stirred
for a further 30 min at 0 °C, and then was concentrated in vacuo
to give an orange solid (35.0 mg). Preparative TLC (chloroform/
hexanes, 60:40) produced 1 as an orange solid (18.0 mg, 74%).
From two other reactions, 51.0 mg (63%) of 1 was obtained from
76.4 mg of amine 4; the orange solid was combined and recrystal-
lized from methanol (30 °C) to give two crops of deep orange
crystals (46 mg, 57%) of nitroxide radical 1.
In an alternative procedure, DMDO (0.06 M, 2 mL, 0.121 mmol,
2.2 equiv) in acetone at -20 °C was added to a solution of amine
4 (15.4 mg, 0.055 mmol, 1.0 equiv) in dichloromethane (5 mL) at
-78 °C. The reaction mixture turned orange immediately and it
was allowed to reach -20 °C over 1 h. The solvent and excess
oxidant was removed in vacuo. The crude was extracted with
chloroform, the organics were separated and dried over MgSO4,
and concentrated in vacuo to give an orange solid (17.8 mg). A
1H NMR (300 MHz, chloroform-d): δ ) 6.20 ppm (br s, 18 H),
methanol solvent peak at 3.88 (0.25 H), impurities at 0.489 (0.22
H). EPR (X-band, 9.4229 GHz, 8 × 10-5 M, diethyl ether):
hyperfine splitting in mT (number of nuclei), aN ) 0.97 (1), aH )
0.19 (4), aH ) 0.08 (4), 100% Lorentzian line width ) 0.035 mT,
g ≈ 2.005, R ) 0.998. IR (ZnSe, cm-1): 2960, 2866, 1493, 1361,
1
1266, 830. EPR, H NMR, and IR spectra are shown in Figures
S1-S3 (Supporting Information).
3.3. General Procedure for EPR/SQUID/EPR and EPR Monitor-
ing of Oxidation of Diarylamines at Low Temperature. Diarylamine
4, diaryldiamine 5, or aza[14]metacyclophane 6 (1-3 mg) was
placed in the homemade SQUID tube (5-mm O.D. EPR quartz tube
with a thin bottom ∼6 cm from the end of the tube)47,48 or in the
4-mm O.D. EPR quartz tube equipped with high-vac PTFE
stopcock, and then evacuated. Dichloromethane, followed by
DMDO, were vacuum transferred to the vessel. The reaction mixture
was stirred at specified temperature between -80 and -95 °C for
specified periods, and then EPR spectra were obtained. After
SQUID data were obtained, another set of EPR spectra were
obtained. The EPR spectra that were obtained before and after the
magnetic measurements were practically identical. The samples for
magnetic measurements and EPR spectroscopy were handled at
strictly low temperatures, using similar procedures that were
previously developed for polyarylmethyl polyradicals.38
9
9104 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Diarylnitroxide Diradicals
A R T I C L E S
3.4. General Procedure for 1H NMR/EPR Monitoring of
Oxidation of Diarylamines at Low Temperature. Diaryldiamine
5 or aza[14]metacyclophane 6 (3-6 mg) was placed in a 5-mm
NMR tube (or custom-made 5-mm NMR tube transitioning in the
lower part a 3-mm NMR tube) equipped with high-vac PTFE
stopcock, and then evacuated. Dichloromethane-d2, followed by
DMDO, were vacuum transferred to the vessel, and then the NMR
tubes were flame-sealed. The reaction mixture was stirred at
specified temperature between -80 and -95 °C for specified
periods, and then the NMR tube was inserted to an NMR probe at
upon high temperature linear extrapolation of the M vs 1/T plots,
using numerical fits in 200-70 K range (R2 ) 1.0000, cooling
mode).
3.7. EPR Spectral Simulations and Numerical Curve
Fitting for SQUID Magnetic Data. WinSIM program (Public EPR
Software Tools, D. A. O’Brien, D. R. Duling, Y. C. Fann) and
WINEPR SimFonia program (Version 1.25) were used for EPR
spectral simulations of solution phase spectra and rigid matrix
spectra, respectively. The SigmaPlot software package was used
for numerical curve fitting of magnetic data. The reliability of fit
is measured by the parameter dependence, which is defined as
follows: dependence ) 1 – ((Variance of the parameter, other
parameters constant)/(Variance of the parameter, other parameters
changing)). Values close to 1 indicate overparametrized fit.
1
about -87 °C. After obtaining H NMR spectra, the sample was
ejected cold from the probe, and then EPR spectra were obtained
at low temperatures. Subsequently, the samples were allowed to
attain room temperature, and 1H NMR spectra at room temperature
were obtained, as well as EPR spectra at low temperatures and at
room temperature.
4. Conclusions
3.5. EPR Spectroscopy. CW X-band EPR spectra were acquired
on an instrument, equipped with a frequency counter and nitrogen
flow temperature control (130-300 K). The samples were contained
in either the 5-mm quartz SQUID sample tubes or 4-mm quartz
EPR sample tubes, which were either flame-sealed or closed with
high-vac PTFE stopcocks. Typically, the tubes were inserted to the
precooled cavity at 140 K. All spectra were obtained with the
oscillating magnetic field perpendicular (TE102) to the swept
magnetic field. The g values were referenced using DPPH (g )
2.0037, powder).
In polycrystalline diarylnitroxide radical 1, weak intermo-
lecular ferromagnetic interactions are determined by magnetic
susceptibility and magnetization studies. Our low-temperature
procedure for oxidation of diarylamine allowed us to detect for
the first time the m-phenylene-based diarylnitroxide diradical.
EPR spectroscopy and magnetic studies (SQUID) indicate that
diarylnitroxide diradical 2 in frozen dichloromethane is observed
in predominantly anti conformation (2A-anti) and it possesses
triplet ground state. Similar oxidation experiments for
aza[14]metacyclophane provide evidence for the formation of
small amounts of the corresponding diarylnitroxide diradical.
It appears that diarylnitroxide diradicals are much less stable
than their corresponding alkylarylnitroxide diradicals, as well
as diarylnitroxide monoradicals. This may imply that steric
shielding of m-phenylene by a 5-tert-butyl group is insufficient
and steric shielding of all para positions with respect to
nitroxides may be required for stable m-phenylene-based
diarylnitroxide diradicals.
3.6. SQUID Magnetometry. Quantum Design MPMS5S (with
continuous temperature control) was used. All samples were
contained in homemade SQUID tubes.51,52
For solid state samples, polycrystalline 1 was obtained by
crystallization from ethanol at 35 °C with cooling to -20 °C.
Radical 1 was loaded to the SQUID tube, placed under vacuum,
and then flame sealed under partial pressure of helium gas.
Correction for diamagnetism was based upon high-temperature
linear extrapolation of the M vs 1/T plots at 5000 Oe; the numerical
fits in 270-90 K range (R2 ) 1.0000) gave intercept (Mdia), which
was added to the magnetization at field H as a factor, (MdiaH)/
5000. Analogous procedure for correction for diamagnetism was
used for a sample in gelatine capsule. For solution sample of
nitroxide radical 1 in 2-MeTHF, solid 1 (2.04 mg) was loaded to
the tube, 2-MeTHF (0.05-0.07 mL) was vacuum transferred, and
then the tube was flame-sealed under vacuum. The tube was inserted
to the sample chamber of the magnetometer at 290 K; the initial
sequence of measurements was in the cooling mode to allow for
condensation of 2-MeTHF. Correction for diamagnetism was based
Acknowledgment. This research was supported by the National
Science Foundation (CHE-0107241, CHE-0414936, CHE-0718117),
including the purchase of the Electron Paramagnetic Resonance
(EPR) spectrometer used in this work (DMR-0216788).
Supporting Information Available: Materials and general
procedures, experimental details (routine analyses, EPR, EPR/
SQUID/EPR, EPR/1H NMR monitoring data for preparation of
nitroxide radicals, NMR spectra of synthetic intermediates),
computational details, and complete ref 47. This material is
available free of charge via the Internet at http://pubs.acs.org.
(51) Rajca, A.; Pink, M.; Rojsajjakul, T.; Lu, K.; Wang, H.; Rajca, S. J. Am.
Chem. Soc. 2003, 125, 8534–8538.
(52) Rajca, S.; Rajca, A.; Wongsriratanakul, J.; Butler, P.; Choi, S. J. Am.
Chem. Soc. 2004, 126, 6972–6986.
JA8016335
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 9105