Isolation of the triplet ground state aminyl diradical

Article abstract of DOI:10.1002/anie.201002811

Spin cycle: A high-spin nitrogen-centered (aminyl) diradical with a triplet (S = 1) ground state (see picture) has a large singlet-triplet energy gap (ΔE_ST) Effective shielding of the reactive aminyl radical centers provides stability at room t

Full text of DOI:10.1002/anie.201002811

Angewandte
Chemie
DOI: 10.1002/anie.201002811
Reactive Intermediates
Isolation of the Triplet Ground State Aminyl Diradical**
´
Przemysław J. Boratynski, Maren Pink, Suchada Rajca, and Andrzej Rajca*
Nitrogen-centered (aminyl) radicals are typically short-lived,
and have been detected as intermediates in a variety of
chemical and biological processes.[1,2] To the best of our
knowledge, only two known stable aminyl radicals have been
reported (more than 30 years ago),[3,4] while a few others have
recently been isolated as aminyl metal complexes and as
heteroatom (e.g., O, S) stabilized aminyls.[5–7]
effective 2pp–2pp overlap that enhances electron spin–spin
interactions (exchange coupling), and the adequate protec-
tion of the carbon and nitrogen atoms with significant spin
densities within the diazapentacene backbone in order to
achieve the stability required for isolation. This result is
significant as the 2pp–2pp overlap is generally severely
perturbed by the increased steric bulk that is required for
achieving radical stability, and thus a combination of a large
DEST value and persistence at room temperature is not
commonly accomplished. In addition, we detected an unex-
pected self-associated dimer of the S = 1 aminyl diradical 2 in
2-MeTHF solution; this species appears to resemble a
p dimer. Notably, characterization of the dimer is facilitated
by its unique zero-field splitting tensor (D). p Dimers of
neutral organic radicals in solution are relatively rare and the
Aminyl diradicals, in which the radical centers are
connected through m-phenylene to form a planar cross-
conjugated p system devoid of stabilizing heteroatoms, are
predicted by computational studies to possess a strong
preference for a high-spin, triplet (S = 1) ground state, that
is, a singlet–triplet energy gap (DEST) that is an order of
magnitude greater than the thermal energy (RT) at room
temperature.[8,9] Such a nitrogen-centered diradical with large
DEST values and persistence at room temperature may be
relevant to the design of strongly paramagnetic relaxation
reagents[10,11] as well as the development of magnetic materi-
als and devices.[12,13] To date, the two known examples of
aminyl diradicals have been examined only at very low
temperature. Platz and coworkers first reported an aminyl
diradical in a 2-methyltetrahydrofuran (2-MeTHF) matrix at
about ꢀ1958C.[14] Recently, we prepared an aminyl diradical
with an S = 1 ground state that was persistent in solution at
about ꢀ1008C.[9]
reports of such phenomena are limited to those of S = 1=
2
radicals.[15]
Diradicals 1 and 2 were prepared from the corresponding
diamines 5 and 6, which were obtained from diamine 3
(Scheme 1).[16] The structure of diamine 6 was confirmed by
X-ray crystallography, and shows overcrowding of the bulky
ꢀ
tert-undecyl groups so that the C C bond axes that connect
the two outer pendant benzene rings to the backbone bend
outward (Figure 1).[17] This bending creates more space for
the pendant benzene rings in order to attain moderate torsion
angles of ꢀ52.0(4), 61.0(4), and ꢀ68.2(3)8 with respect to the
diazapentacene backbone. Nevertheless, both nitrogen atoms
are well-shielded and the diazapentacene backbone remains
planar in spite of the steric congestion. The structure of
diradical 2, which differs from its diamine precursor by the
absence of two hydrogen atoms, is expected to be similar to
that of diamine 6. Such a sterically shielded, planar structure
should have enhanced stability and a strong preference for the
S = 1 ground state.
Herein we report the isolation of an S = 1 ground state
aminyl diradical with a large DEST value and persistence at
room temperature. The synthesis of the aminyl diradicals 1
and 2 demonstrates the balance between maintaining an
In spite of hindered access to the NH moieties in the
diazapentacene backbone, double deprotonation of diamines
5 and 6 in THF/hexane at about ꢀ308C, followed by oxidation
of the resultant dianions with iodine between ꢀ90 and
ꢀ1158C provided the corresponding diradicals 1 and 2,
respectively. Generation of 1 is difficult because of its low
solubility, and characterization of 1 is possible only by EPR
spectroscopy, which gave spectra that are typical of the aminyl
diradical in the S = 1 state (Figure S3 in the Supporting
Information).[9] In contrast, the generation of 2 is very
efficient, and is facilitated by the good solubility of the
diradical and its diamine precursor.
´
[*] Dr. P. J. Boratynski, Prof. Dr. S. Rajca, Prof. Dr. A. Rajca
Department of Chemistry, University of Nebraska
Lincoln, NE 68588-0304 (USA)
Fax: (+1)402-472-9402
E-mail: arajca1@unl.edu
Homepage: http://www.chem.unl.edu/rajca/rajcahome.html
Dr. M. Pink
Diradical 2 was prepared on a 1 mg (1 mmol) scale in
custom-designed sample tubes, thus enabling sequential
measurements by EPR spectroscopy and superconducting
quantum interference device (SQUID) magnetometry at low
temperatures. The EPR spectra of concentrated purple-blue
solutions of 2 in 2-MeTHF at 132 K are strikingly complex,
Department of Chemistry, Indiana University
Bloomington, IN 47405-7102 (USA)
[**] We thank the National Science Foundation for support of this work
through grant CHE-0718117.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201002811.
Angew. Chem. Int. Ed. 2010, 49, 5459 –5462
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5459

Communications
about 130 K; Figure 2a) because of weak
exchange coupling (j J/k j ! 130 K). The two out-
ermost EPR bands, which are separated by 6 j D2 j
and are not present in the spectra of poorly soluble
1, are assigned to the S = 2 state for the dimer with
j D2/hc j = 5.49 ꢀ 10ꢀ3 cmꢀ1. This assignment is
confirmed by dilution experiments, that is, when
the concentration of 2 is decreased, the relative
intensity of the two outer peaks decreases (Fig-
ure 2a) and the EPR spectra of 2 become increas-
ingly similar to those of 1 with j Dd/hc j = 8.67 ꢀ
10ꢀ3 cmꢀ1 (Figure S3 in the Supporting Informa-
tion).
Scheme 1. Synthesis of aminyl diradicals 1 and 2. a) NBS in chloroform, NaHCO3,
Na2SO4, ꢀ408C for 1 h, then 08C for 2 h, 35%. b) 4-tert-butylphenylboronic acid or
4-tert-undecylphenylboronate, [Pd(PPh3)4], Na2CO3, benzene/EtOH/water, 908C (5:
62% yield, 6: 43% yield). c) nBuLi (2.4 equiv), THF/hexane, ca. ꢀ308C for 1 h, then
solvent change to THF or 2-MeTHF, and I2 (1.06 equiv) at ꢀ908C or titration with I2
at ꢀ90 to ꢀ1158C monitored by EPR spectroscopy. NBS=N-bromosuccinimide.
Figure 1. X-ray structure of diamine 6. Disorder is omitted for clarity.
C gray, H white, N blue.
(Figure 2a ; 4 mm solution of 2). A similar EPR spectrum was
obtained for a 10 mm solution of 2 (Figures S6–S8 in the
Supporting Information), which was also examined by
SQUID magnetometry (Figure 2b). In spite of this complex-
ity, SQUID magnetometry unequivocally indicates an
approximate S = 1 paramagnetic behavior for both magnet-
ization (M) versus magnetic field (H), and magnetic suscept-
ibility (c) versus temperature (T). The magnetization curves
at 1.8–5 K follow the Brillouin plot for S = 1 (Figure 2b),
though after correction by a mean-field parameter q ꢁ
ꢀ0.3 K. This value, which is similar for both 10 mm and
50 mm 2 in 2-MeTHF, implies a weak antiferromagnetic
exchange coupling between the S = 1 ground state diradicals.
We propose that the S = 1 diradical 2 assembles to form a
dimer 22 in 2-MeTHF. A numerical fit to the variable-
temperature magnetic data, that is, cT versus T, gave a weak
antiferromagnetic exchange coupling J/k ꢁ ꢀ0.2 K (singlet–
quintet energy gap, 6 j J j ꢁ 2 calmolꢀ1), between two S = 1
diradicals within the dimer 22 (Figure 2b, inset).[18] This
exchange coupling implies the presence of two EPR-active
states, S = 2 and S = 1, in addition to the S = 0 ground state,
for 22. If 22 were a centrosymmetric structure with the
diradicals oriented similarly to those in a p dimer, the
properties of the zero-field splitting tensors (D) would be
similar to those in exchange-coupled, symmetric metal ion
pairs.[19] Therefore, the largest components of D tensors for
the S = 1 state (D1) and S = 2 state (D2), as well as for
monomeric 2 (Dd) and exchange contribution (Dexch), would
be collinear and the corresponding zero-field splitting param-
eters may be approximately related by Equations (1a and
1b):[19] The S = 2 and S = 1 states of the dimer are expected to
be significantly populated in the EPR spectrum (obtained at
Figure 2. a) EPR (X band) spectra of 4.0 mm (top) 0.5 mm 2 (bottom)
in 2-MeTHF at 132 K, and direction of the z axis that defines the
orientation of the largest components of the D tensor for the aminyl
diradical and its dimer. Black line: experimental values, red line:
simulated values. b) SQUID magnetometry of ca. 10 mm 2 in 2-
MeTHF. Main plot: M/Msat versus H/(Tꢀq) with qꢁꢀ0.3 K, showing
curvature corresponding to the value of S=0.98, which is independent
of the spin concentration of the sample (Msat =magnetization at
saturation). Inset: cT versus T, showing numerical fit to dimer of S=1
diradicals with J/kꢁꢀ0.2 K (R2 =0.995).
5460 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5459 –5462

Angewandte
Chemie
D1 ¼ ꢀDd þ Dexch
ð1aÞ
ð1bÞ
ꢀ268C and ꢀ788C, respectively, resulted in almost complete
disappearance of the diradical. In the absence of iodine and
oxygen, EPR spectra of 2 in 2-MeTHF show only slight
changes in intensity upon sequential warming to ꢀ26, 0, and
228C. This remarkable persistence of 2 prompted us to
explore the isolation of the diradical.
D2 ¼ Dd=3 þ Dexch=3
For dilute solutions of 2, which behave largely like a
monomeric S = 1 diradical, another pair of outermost EPR
bands separated by 2 j Dd j appears as pentuplets (Figure 2a),
which arise from 14N-hyperfine coupling with two nitrogen
atoms in the S = 1 diradical.[9] Therefore, the largest compo-
nents of the zero-field splitting tensor (D) and hyperfine
tensor (A) are parallel to each other, and to the 2pp orbitals on
the nitrogen atoms (2pz in Figure 2a), thus indicating that
diradical 2 has an approximately planar diazapentacene
backbone, as inferred from the X-ray crystal structure of
diamine 6. This parallel orientation of the largest components
for the D and A tensors in 2 is qualitatively different
compared to the perpendicular orientation in the recently
reported S = 1 planar nitroxide diradical derived from
diamine 3,[20] thus precluding the possibility that 2 is a
nitroxide diradical.
The orientation of the largest component of the D tensor
in 2 indicates that the magnetic dipole–dipole interactions are
strongest along the z axis,[21] with the strength qualitatively
related to the spectral width 2 j Dd j (Figure 2a). This
observation suggests that the spin density is most pronounced
along the z axis in 2, and even more so in the S = 2 state of 22,
which has an approximately centrosymmetric p-dimer-like
structure (Figure S54 in the Supporting Information). Thus,
the magnetic dipole–dipole interactions in the S = 2 state of 22
are expected to be even stronger along the z axis, therefore
leading to the relatively large spectral width 6 j D2 j (Fig-
ure 2a), which both enables the spectroscopic detection of the
S = 2 state and provides evidence for a structure of the dimer.
Based upon values of D2 and Dd, and assuming that these
values have the same signs, we can estimate from Equa-
tion (1) that j D1/hc j is on the order of 1 ꢀ 10ꢀ3 cmꢀ1, which is
smaller than j Dd/hc j , as qualitatively expected for the more
spin-dilute S = 1 state of the dimer. This approximate value is
reasonably consistent with the value of j D1/hc j ꢁ 1.8 ꢀ
10ꢀ3 cmꢀ1, which was obtained from numerical simulations
of EPR spectra (Figure 2a and Figure S13 in the Supporting
Information). These results provide further support for the
structure of 22. The association constant of 22 is Kassoc ꢁ 3 ꢀ
102 mꢀ1 at 132 K (DG ꢁ ꢀ1.5 kcalmolꢀ1),[22] estimated from
numerical simulations of EPR spectra for three samples of
0.5–4.0 mm 2 in 2-MeTHF.
We devised a stringent procedure for the isolation of
diradical 2. The procedure involves generation of 30–65 mg of
diradical, followed by precipitation and filtration at low
temperature in a Schlenk vessel under an inert atmosphere.
After diradical 2 was generated from the diamine as described
above, the solvent was removed under vacuum at ꢀ788C.
Then, methanol was added by vacuum transfer in order to
dissolve salts and any excess iodine, and to precipitate the
diradical. Because of its reactivity towards methanol above
ꢀ788C, the dark blue solid aminyl diradical was filtered at low
temperature to minimize decomposition. The solid sample
(7:3 diradical/monoradical) could be stored in a glovebox at
ꢀ208C for seven months, during which time only slight
decomposition occurred. At room temperature, the concen-
tration of the S = 1 diradical decreased to 60% after five days
(Figure S21 in the Supporting Information).
The properties of 2, that is, a solid diradical that is
persistent at room temperature, facilitate the magnetic
measurements that provide an estimated DEST value with
significantly greater precision compared to low-temperature
measurements in solution. In particular, the cT versus T plot
for solid 2 is flat from 30 K to 290 K, which is the highest
temperature of the SQUID measurement (Figure S22 in the
Supporting Information), thus indicating that DEST ꢂ 2 kcal
molꢀ1. This limiting value is in qualitative agreement with the
value of DEST ꢁ 7 kcalmolꢀ1 calculated for diradical 1 using
the UB3LYP/6-31(d) level of theory.[23,24]
The isolation of diradical 2 also allows a more in-depth
study of kinetic stability of the aminyl diradical and the
derived monoradical. The stability of 2 in 2-MeTHF was
monitored by EPR spectroscopy at room temperature. The
first-order decay kinetics of diradical and monoradical
corresponds to a half-life of about 3 and 7 hours, respectively
1
(Figure 3); after several days, H NMR spectra showed that
the isolated sample was predominantly diamine 6 (Figur-
es S23–S29 in the Supporting Information). These results
suggest that the process of conversion of 2 to 6 proceeds by a
hydrogen abstraction mechanism, which is facilitated by the
One of the possible structures of 22 is modeled using a
simplified centrosymmetric dimer without long alkyl chains,
that is, 12, for which the geometry of S = 2 state is optimized
using the standard density functional level of theory
(UB3LYP/6-31G(d)).[23] This model structure possesses a
relatively long plane-to-plane distance (5.73 ꢁ) between the
diazapentacene backbones of bulky S = 1 diradicals; this
distance is consistent with the weak exchange coupling
between the diradicals (Figure S53 in the Supporting Infor-
mation).
Figure 3. Decay kinetics of diradical and monoradical in 2-MeTHF at
*
294 K ( : experimental values, c: numerical fit). The four-parameter
fit to the first-order decay kinetics (R2 =0.9999) gives rate constants
k1 =6.8ꢀ10ꢀ5 and k2 =2.9ꢀ10ꢀ5 sꢀ1 (additional details are given in
Figures S23–S31 in the Supporting Information).
EPR spectroscopy of 2 in 2-MeTHF shows that the
diradical is inert toward an excess of iodine and oxygen at
temperatures as low as ꢀ1088C, however a brief warming to
Angew. Chem. Int. Ed. 2010, 49, 5459 –5462
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5461

Communications
[25]
ꢀ
[7] Y. Miura, T. Tomimura, Chem. Commun. 2001, 627 – 628.
[8] S. Amiri, P. R. Schreiner, J. Phys. Chem. A 2009, 113, 11750 –
11757.
[9] A. Rajca, K. Shiraishi, M. Pink, S. Rajca, J. Am. Chem. Soc. 2007,
129, 7232 – 7233.
[10] G. M. Clore, J. Iwahara, Chem. Rev. 2009, 109, 4108 – 4139.
[11] A. J. Maliakal, N. J. Turro, A. W. Bosman, J. Cornel, E. W.
Meijer, J. Phys. Chem. A 2003, 107, 8467 – 8475.
[12] A. Rajca, J. Wongsriratanakul, S. Rajca, Science 2001, 294, 1503 –
1505.
low C H bond dissociation energy of solvents such as THF.
This mechanism is analogous to that observed for matrix-
isolated dialkyl aminyl radicals.[26]
The diradical 2 reacts with oxygen at room temperature
both in the solid state and in solution to result in a new
diamagnetic product as well as variable amounts of diamine 6
(see the Supporting Information). This reactivity is in contrast
to that of other aminyl radicals, which were reported to react
ꢀ
with oxygen to form nitroxide radicals (N OC) via peroxy
[13] C. Herrmann, G. C. Solomon, M. A. Ratner, J. Am. Chem. Soc.
2010, 132, 3682 – 3684.
[26,27]
ꢀ ꢀ
radicals (N O OC).
In summary, we isolated a triplet ground state aminyl
diradical that is persistent at room temperature and possesses
a singlet–triplet energy gap that is significantly greater than
the thermal energy at room temperature. The diradical
assembles in solution to form a weakly exchange coupled p-
dimer-like structure that was detected spectroscopically in its
thermally populated S = 2 state. Both the magnitude of the
exchange coupling (j J/k j ꢁ 0.1–1 K)[28] and relatively large
zero-field splitting in the dimer suggest that electron spin
relaxation rates could be enhanced by an order of magni-
tude.[29] Such diradicals may have potential in the design of
paramagnetic relaxation reagents and spin-transport devi-
ces.[10,11,13]
[14] K. Haider, N. Soundararajan, M. Shaffer, M. S. Platz, Tetrahe-
dron Lett. 1989, 30, 1225 – 1228.
[15] D. Small, V. Zaitsev, Y. Jung, S. V. Rosokha, M. Head-Gordon,
J. K. Kochi, J. Am. Chem. Soc. 2004, 126, 13850 – 13858.
[16] K. Shiraishi, A. Rajca, M. Pink, S. Rajca, J. Am. Chem. Soc. 2005,
127, 9312 – 9313.
[17] CCDC 758072 (6) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
[18] A. Rajca, M. Takahashi, M. Pink, G. Spagnol, S. Rajca, J. Am.
Chem. Soc. 2007, 129, 10159 – 10170.
[19] A. Bencini, D. Gatteschi, EPR of Exchange Coupled Systems,
Springer, Berlin, 1990, pp. 48 – 85.
[20] A. Rajca, K. Shiraishi, S. Rajca, Chem. Commun. 2009, 4372 –
4374.
Received: May 10, 2010
Published online: July 2, 2010
[21] The spin–orbit coupling contribution to the D tensor may be
´
neglected in the N-centered diradicals: Z. Havlas, M. Kyvala, J.
Michl, Mol. Phys. 2005, 103, 407 – 411.
Keywords: amines · dimerization · p interactions · radicals ·
reactive intermediates
[22] J. Wu, A. Fechtenkꢃtter, J. Gauss, M. D. Watson, M. Kastler, C.
Fechtenkꢃtter, M. Wagner, K. Mꢂllen, J. Am. Chem. Soc. 2004,
126, 11311 – 11321.
[23] M. J. Frisch et al., Gaussian 03, revision E.01, Gaussian, Wall-
ingford, CT, 2004 (full citation: see Supporting Information).
[24] This level of theory, which is based upon the broken-symmetry
wave function, tends to overestimate the value of DEST for triplet
ground state molecules: see refs [2] and [8].
.
[1] K. Piech, T. Bally, A. Sikora, A. Marcinek, J. Am. Chem. Soc.
2007, 129, 3211 – 3217.
[2] H. Quast, W. Nꢂdling, G. Klemm, A. Kirschfeld, P. Neuhaus, W.
Sander, D. A. Hrovat, W. T. Borden, J. Org. Chem. 2008, 73,
4956 – 4961.
[25] S. J. Blanksby, G. B. Ellison, Acc. Chem. Res. 2003, 36, 255 – 263.
[26] J. R. Roberts, K. U. Ingold, J. Am. Chem. Soc. 1973, 95, 3228 –
3235.
[3] F. A. Neugebauer, H. Fischer, S. Bamberger, H. O. Smith, Chem.
Ber. 1972, 105, 2694 – 2713.
[27] S. F. Nelsen, R. T. Landis, L. H. Kiehle, T. H. Leung, J. Am.
Chem. Soc. 1972, 94, 1610 – 1614.
[4] N. Ballester, J. Castaner, S. Olivella, Tetrahedron Lett. 1974, 15,
615 – 616.
[28] A. Rajca, S. Mukherjee, M. Pink, S. Rajca, J. Am. Chem. Soc.
2006, 128, 13497 – 13507.
[29] H. Sato, V. Kathirvelu, G. Spagnol, S. Rajca, A. Rajca, S. S.
Eaton, G. R. Eaton, J. Phys. Chem. B 2008, 112, 2818 – 2828.
[5] T. Bꢂttner, J. Geier, G. Frison, J. Harmer, C. Calle, A. Schweiger,
H. Schꢃnberg, H. Grꢂtzmacher, Science 2005, 307, 235 – 238.
[6] N. P. Mankad, W. E. Antholine, R. K. Szilagyi, J. C. Peters, J.
Am. Chem. Soc. 2009, 131, 3878 – 3880.
5462
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5459 –5462