Isolation and characterization of the triradical 1,3,5-trimethylene-benzene

Article abstract of DOI:10.1002/anie.201001393

The highly symmetrical quartet triradical trimethylenebenzene has been isolated for the first time (see structure). Despite its open-shell character, it was photo-chemically stable and this makes it a promising building block for magnetic materials. (Figure presented).

Full text of DOI:10.1002/anie.201001393

Angewandte
Chemie
DOI: 10.1002/anie.201001393
Radicals
Isolation and Characterization of the Triradical 1,3,5-Trimethylene-
benzene**
Patrik Neuhaus and Wolfram Sander*
Organic high-spin molecules[1,2] can serve as model com-
pounds for the design of organic magnetic materials.[3,4] The
most frequently used coupling unit to achieve ferromagnetic
coupling between p radical centers is the m-phenylene group.
Thus, 1,3-dimethylenebenzene (m-xylylene) 1 was shown to
have a robust triplet ground state.[5,6] The spectroscopic
properties, structure, and photochemistry of 1 was recently
investigated using matrix isolation spectroscopy.[7] Prerequi-
site for these studies was the observation that high yields of 1
are obtained by flash vacuum pyrolysis (FVP) of 1,3-
bis(iodomethyl)benzene 2 at 4508C and subsequent trapping
of the products in argon at 10 K (Scheme 1).
absorption in the near-UV region, they can be obtained in
high photostationary yields, depending on the wavelength of
the irradiating light.
1,3,5-Trimethylenebenzene 6 (see Scheme 2 for the struc-
ture) forms the basis of a number of organic high-spin
polyradicals, and several derivatives of 6 have been described
in literature. The first attempt to synthesize a derivative of 6
was reported in 1937 by Leo.[8] In contrast, the parent
triradical 6 has only been investigated by negative-ion
photoelectron spectroscopy in the gas phase and by theoret-
ical methods. Hammad and Wenthold measured the heat of
formation of triradical 6 by collision induced dissociation of
the biradical anion 5-chloromethyl-m-xylylene.[9] It was
shown that the interaction between the radical centers in 6
is very small, and therefore the heat of formation can be
reliably estimated by simple bond additivity. The doublet –
quartet energy gap has been calculated to 12–14 kcalmolꢀ1 by
various theoretical methods.[10,11] Herein we describe the first
isolation and spectroscopic characterization of the elusive
triradical 6 by using the matrix isolation technique.
Flash vacuum pyrolysis of 1,3,5-tris(iodomethyl)benzene
7 with subsequent trapping of the products in argon at 10 K
(Scheme 2) allowed us to obtain IR, UV/Vis, and EPR spectra
Scheme 1. Synthesis and photochemistry of m-xylylene 1.
Most notably is the photochemistry of diradical 1. Near-
UV irradiation rapidly yields the highly strained polycyclic
hydrocarbons 3–5, which are all thermodynamically less
stable than diradical 1, in photostationary equilibria with 1.
However, because these hydrocarbons show only weak or no
Scheme 2. Synthesis of 1,3,5-trimethylenebenzene 6 and by-products
through flash vacuum pyrolysis.
[*] P. Neuhaus, Prof. Dr. W. Sander
Lehrstuhl fꢀr Organische Chemie II
Ruhr-Universitꢁt Bochum
of the pyrolysis products. The IR spectrum obtained after
pyrolysis of 7 at 4808C shows three intense bands at 614.7,
727.3, and 814.7 cmꢀ1 as well as several weak bands assigned
to triradical 6 (Figure 1, Table 1). The assignment has been
confirmed by partial deuteration and by comparison with
calculations at various levels of theory (see the Supporting
Information). With a sufficiently large basis set, DFT and
MP2 calculations show only moderate spin contaminations
and predict very similar IR spectra. Therefore, here we only
44780 Bochum (Germany)
Fax: (+49)234-321-4353
E-mail: wolfram.sander@rub.de
Homepage: http://www.rub.de/oc2
[**] This work was financially supported by the Deutsche Forschungs-
gemeinschaft and the Fonds der Chemischen Industrie.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201001393.
Angew. Chem. Int. Ed. 2010, 49, 7277 –7280
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7277

Communications
positions ([D6]-7), only two very strong absorptions are
observed in the spectrum (n14 and n21), whereas n16 is now
very weak.
The vibration n21 at 814.7 cmꢀ1 is described best as a
ꢀ
symmetrical ring C H oop deformation mode with only small
contributions from the methylene groups. Consequently,
deuteration of the methylene groups results in a deuterium
isotopic shift of only ꢀ5.7 cmꢀ1, in good agreement with the
calculations (ꢀ5.1 cmꢀ1).
The vibration n16 at 727.3 cmꢀ1 shows the largest deute-
rium isotopic shift (ꢀ101.6 cmꢀ1) and the largest change in
intensity on deuteration. Deuteration of the methylene
groups results in a drastic reduction of the intensity of this
band to approximately 3% of the original value. The reason
for this is the out of phase movement of the methylene
hydrogen atoms with respect to the ring hydrogen atoms. This
results in [D6]-6 in a fortuitous cancellation of the dipole
Figure 1. IR spectra obtained after FVP at 4808C and subsequent
trapping of the products in argon at 10 K. a) FVP of [D6]-7. b) FVP of 7.
c) IR spectrum of quartet [D6]-6 calculated at the UB3LYP/6-311+
G(d,p) level of theory. d) IR spectrum of quartet 6 calculated at the
UB3LYP/6-311+G(d,p) level of theory.
ꢀ
ꢀ
moment contributions of the C H and the C D movements
and thus drastically reduced IR
intensity. Both the isotopic shift
and the reduced intensity are repro-
duced by the calculations. Vibration
n14 at 614.7 cmꢀ1 shows a deuterium
shift of ꢀ43.0 cmꢀ1 (calculated shift:
ꢀ53.4 cmꢀ1).
Table 1: IR spectroscopic data of triradical 6 and [D6]-6.
Mode n Sym. 6, exp.[a]
6, calcd[b]
[D6]-6, exp.[a] [D6]-6, calcd[b] Assignment
3
ꢀ
14
16
21
22
28
35
38
41
43
47
A2’’
A2’’
A2’’
E’
E’
E’
E’
E’
E’
E’
614.7 (75)
727.3 (65)
814.7 (100) 827.9 (99)
624.1 (58)
727.4 (96)
571.7 (100)
625.7 (1)
570.7 (100)
636.2 (3)
822.8 (53)
764.8 (1)
1081.1 (3)
1270.1 (4)
1509.4 (4)
2284.0 (5)
3167.7 (15)
2416.5 (6)
C
C
C
C
C
C
C
H oop def
H2 oop def
1
3
1
1
1
1
ꢀ
ꢀ
809.0 (60)
H oop def
H2 rock
H2 scis, C H ip def
H2 scis, C C str
[c]
ꢀ
926.3 (2)
936.3 (2)
1214.9 (0)
1490.2 (7)
1512.5 (8)
–
Minor products of the FVP of 7
detected by IR spectroscopy in solid
argon are 3,5-dimethylbenzylradi-
cal 9 and triplet 5-methyl-m-xyly-
lene 8 (Scheme 2). For comparison
of the IR spectra these products
were independently synthesized
and matrix-isolated by FVP of 1,3-
bis(iodomethyl)-5-methyl-benzene
10 and 3,5-dimethylbenzyliodide 11,
respectively (see the Supporting
Information).
[c]
3
ꢀ
ꢀ
–
1058.0 (5)
1246.3 (1)
1481.3 (5)
1
2
ꢀ
=
1462.6 (5)
1478.3 (5)
3018.9 (10) 3145.3 (12) 2214.9 (5)
3038.0 (10) 3167.8 (18) 3039.4 (5)
3111.6 (10) 3241.7 (13)
2
1
=
ꢀ
C str, C H2 scis
1
ꢀ
sym C H2 str
3
ꢀ
C
H str
[c]
1
ꢀ
–
asym C H2 str
[a] Argon, 10 K. Values in parentheses are the signal intensities. [b] UB3LYP/6-311+G(d,p). Values in
parentheses are the signal intensities. [c] Not assigned owing to overlap with a strong neighboring
signal. asym=asymmetric, def=deformation, ip=in plane, oop=out of plane, rock=rocking, scis=
scissoring, str=stretch, sym=symmetric.
The UV/Vis spectrum of an
discuss the results from calculations at the UB3LYP/6-311 +
G(d,p) level of theory.
argon matrix at 10 K obtained after FVP of 7 shows
absorptions from several species (see the Supporting Infor-
mation). A weak band at 292.3 nm and parts of a broad and
intense band around 232.4 nm are assigned to triplet diradical
8, while the weak transition at 260.9 nm and the very weak
transition at 318.5 nm are assigned to radical 9—these results
are in good accordance with the fluorescence excitation and
absorption bands of benzylradicals.[12–15] Two remaining weak
bands at 367 nm and 254.9 nm might belong to 6, however, we
cannot exclude the possibility that these bands are caused by
an unknown impurity. Nguyen and co-workers calculated for
6 the first electronic transition to 354 nm by MS-CASPT2
theory, which would be in good agreement with the 367 nm
band.[16]
The three most intense bands of 6 are the A2’’ symmetrical
ꢀ
out of plane (oop) C H deformation modes n14, n16, and n21 of
the methylene groups and the ring hydrogen atoms (Figure 2,
Table 1). If the precursor is deuterated at all methylene
Further evidence for the formation of triradical 6 comes
from EPR spectroscopy where 6 could be generated by either
FVP of 7 with subsequent trapping of the products in argon at
5 K or by photolysis of matrix-isolated 7 at 5 K (Figure 3).
Matrices containing 6 show a centrosymmetric five-line EPR
spectrum centered at 3427 G. At 5 K the spectrum is
persistent for days without any change. A half-field signal at
Figure 2. A2’’ symmetrical out of plane vibrations of 6.
7278
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7277 –7280

Angewandte
Chemie
Experimental Section
Synthesis: 1,3,5-tris(bromomethyl-d2)benzene was synthesized
according to the literature starting from trimethyl-1,3,5-benzenetri-
carboxylate.[25]
Synthesis of 1,3,5-tris(iodomethyl-d2)benzene [D6]-7: 1,3,5-tris-
(bromomethyl-d2)benzene (0.363 g, 1 mmol) was dissolved in 10 mL
of dry acetone. To this solution NaI (2.248 g, 15 mmol) was added and
the mixture stirred for 3 h at RT. The suspension was filtered and after
removal of the solvent a yellow-white solid remained. This solid was
further purified by column chromatography on silica gel using n-
~
hexane as the eluent to give the product in 85% yield. IR (KBr): n =
2271, 2263, 1593, 1436, 1183, 1070, 951, 909, 852, 830, 694, 512,
499 cmꢀ1; HRMS(EI): m/z calcd for [C9H3D6]: 503.8215; found:
503.8211.
Matrix isolation: Matrix isolation experiments were performed
by standard techniques[26] using a closed-cycle helium cryostat and a
CsI spectroscopic window cooled to 10 K. FTIR spectra were
recorded with a standard resolution of 0.5 cmꢀ1, using a liquid-
nitrogen-cooled MCT detector in the range 400–4000 cmꢀ1. X-band
EPR spectra were recorded from a sample deposited on an oxygen-
free high-conductivity copper rod (75 mm length, 2 mm diameter)
cooled with a closed-cycle cryostat to 4 K. UV/vis spectra were
recorded in the spectroscopic range of 800 to 200 nm with a standard
resolution of 0.1 nm, using a Varian UV/Vis NIR spectrophotometer
from a sample deposited on a sapphire window cooled to 10 K by a
closed-cycle cryostat. Flash vacuum pyrolysis was carried out by
slowly subliming 7 through a 7 cm quartz tube heated electrically with
a tantalum wire.
Broadband irradiation was carried out with mercury high-
pressure arc lamps in housings equipped with quartz optics and
dichroic mirrors in combination with cutoff filters (50% transmission
at the wavelength specified). For 254 nm irradiation a low-pressure
mercury arc lamp was used. An excimer laser (XeCl) was used for
308 nm photolysis.
Figure 3. a) EPR spectrum obtained after irradiation of matrix isolated
7 with 308 nm. b) Simulated quartet EPR spectrum using jD/
hcj=0.0128 cmꢀ1, jE/hcj=0 cmꢀ1, g=2.0023, m=9.60067 GHz.
Inset: half-field signal.
1708 G and a very weak signal at 1130 G corresponding to a
Dms = ꢁ 3 transition appear together with the main signals at
g = 2. The five lines in the spectrum can be assigned to the z
and xy signals of a randomly oriented quartet molecule with a
threefold or higher axis of symmetry. Similar EPR spectra
have been observed for other quartet molecules with high
symmetry and have been predicted by theory.[17–24] The
simulation of a quartet state (S = 3/2) with the zero-field-
splitting (zfs) parameters j D/hc j = 0.0128 cmꢀ1 and j E/hc j
= 0 cmꢀ1 gives the best agreement with the experimental
spectrum. The experimental spectrum cannot be simulated
with a triplet state, which rules out the possibility that the
half-field signal belongs to a triplet species. Triplet diradical 8
has been independently synthesized and characterized by
EPR spectroscopy. As expected, its zfs parameters j D/hc j =
0.0111 cmꢀ1 and j E/hc j ꢂ 0.0009 cmꢀ1 are very similar to that
of triplet m-xylylene 1 (see the Supporting Information).
UV photolysis (308 nm) of matrix-isolated 7 produces a
clean EPR spectrum of 6 (Figure 3). Interestingly, other
paramagnetic intermediates (radicals or diradicals formed by
loss of one or two of the three iodine atoms) are not formed in
detectable amounts during initial photolysis of 7. Under the
same conditions the IR spectrum shows no apparent photo-
chemistry of 7. Obviously, the recombination of the radical
Computational Methods: Optimized geometries and vibrational
frequencies of all species were calculated at the B3LYP[27–29] level of
theory employing the 6-311 + G(d,p) polarized valence-triple-x basis
set.[30,31] Tight convergence criteria were used throughout. A spin-
unrestricted formalism was used for all high-spin systems and for
singlet biradicals, whenever instability[32] was observed. All calcula-
tion were carried out with Gaussian 03.[33]
Received: March 9, 2010
Published online: August 19, 2010
Keywords: EPR spectroscopy · IR spectroscopy ·
.
matrix isolation · 1,3,5-trimethylenebenzene · triradicals
ꢀ
pairs produced by cleavage of the C I bonds is very efficient,
and thus the yield of 6 obtained by photolysis of 7 is too low
for detection by IR spectroscopy.
[1] W. Sander, D. Grote, S. Kossmann, F. Neese, J. Am. Chem. Soc.
2008, 130, 4396.
[2] H. Quast, W. Nudling, G. Klemm, A. Kirschfeld, P. Neuhaus, W.
Sander, D. A. Hrovat, W. T. Borden, J. Org. Chem. 2008, 73,
4956.
[3] K. Sato, D. Shiomi, T. Takui, M. Hattori, K. Hirai, H. Tomioka,
Mol. Cryst. Liq. Cryst. 2002, 376, 549.
[4] P. M. Lahti, Magnetic Properties of Organic Materials, Marcel
Dekker, New York, 1999, p. 661.
[5] B. B. Wright, M. S. Platz, J. Am. Chem. Soc. 1983, 105, 628.
[6] L. Catala, J. Le Moigne, N. Gruber, J. J. Novoa, P. Rabu, E.
Belorizky, P. Turek, Chem. Eur. J. 2005, 11, 2440.
[7] P. Neuhaus, D. Grote, W. Sander, J. Am. Chem. Soc. 2008, 130,
2993.
[8] M. Leo, Ber. Dtsch. Chem. Ges. B 1937, 70, 1691.
[9] L. A. Hammad, P. G. Wenthold, J. Am. Chem. Soc. 2001, 123,
12311.
In contrast to the highly photolabile diradical 1, triradical
6 is rather photostable. Prolonged UV irradiation results in a
very slow degradation of 6. The IR spectrum of the photo-
products of 6 show a number of broad, low intensity bands
that could not be assigned. The slow photochemistry of 6 is
remarkable since under the same conditions 1 rearranges in
clean reactions to the strained hydrocarbons 3, 4, and 5.[7]
In summary, we were able to synthesize trimethyleneben-
zene 6 by FVP of triiodide 7. The matrix IR and EPR spectra
clearly prove that 6 is a highly symmetrical triradical with a
quartet ground state. This is in accordance with earlier
theoretical studies that predicted a large quartet–doublet
splitting. The photostability makes derivatives of 6 interesting
building blocks of organic magnets.
Angew. Chem. Int. Ed. 2010, 49, 7277 –7280
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7279

Communications
[10] C. R. Kemnitz, R. R. Squires, W. T. Borden, J. Am. Chem. Soc.
[21] G. Kothe, E. Ohmes, J. Brickmann, H. Zimmermann, Angew.
Chem. 1971, 83, 1015; Angew. Chem. Int. Ed. Engl. 1971, 10, 938.
[22] J. Brickmann, G. Kothe, J. Chem. Phys. 1973, 59, 2807.
[23] S. I. Weissman, G. Kothe, J. Am. Chem. Soc. 1975, 97, 2537.
[24] A. Rajca, S. Utamapanya, J. Am. Chem. Soc. 1993, 115, 2396.
[25] M. Nakazaki, K. Yamamoto, T. Toya, J. Org. Chem. 1981, 46,
1611.
[26] I. R. Dunkin, Matrix-Isolation Techniques, Oxford University
Press, Oxford, 1998.
[27] A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
[28] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.
[29] B. Miehlich, A. Savin, H. Stoll, H. Preuss, Chem. Phys. Lett. 1989,
157, 200.
[30] A. D. McLean, G. S. Chandler, J. Chem. Phys. 1980, 72, 5639.
[31] R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J. Chem. Phys.
1980, 72, 650.
1997, 119, 6564.
[11] H. M. T. Nguyen, A. Dutta, K. Morokuma, M. T. Nguyen, J.
Chem. Phys. 2005, 122, 154308/1.
[12] V. LeJeune, A. Despres, E. Migirdicyan, J. Phys. Chem. 1984, 88,
2719.
[13] V. Lejeune, A. Despres, E. Migirdicyan, J. Baudet, G. Berthier, J.
Am. Chem. Soc. 1986, 108, 1853.
[14] D. M. Friedrich, A. C. Albrecht, Chem. Phys. 1974, 6, 366.
[15] T. Izumida, T. Ichikawa, H. Yoshida, J. Phys. Chem. 1980, 84, 60.
[16] H. M. T. Nguyen, T. T. Hue, M. T. Nguyen, Chem. Phys. Lett.
2005, 411, 450.
[17] T. D. Selby, K. R. Stickley, S. C. Blackstock, Org. Lett. 2000, 2,
171.
[18] K. R. Stickley, T. D. Selby, S. C. Blackstock, J. Org. Chem. 1997,
62, 448.
[19] K. Yoshizawa, A. Chano, A. Ito, K. Tanaka, T. Yamabe, H.
Fujita, J. Yamauchi, Chem. Lett. 1992, 369.
[20] K. Yoshizawa, A. Chano, A. Ito, K. Tanaka, T. Yamabe, H.
Fujita, J. Yamauchi, M. Shiro, J. Am. Chem. Soc. 1992, 114, 5994.
[32] R. Bauernschmitt, R. Ahlrichs, J. Chem. Phys. 1996, 104, 9047.
[33] G. W. T. M. J. Frisch, et al. Gaussian, Inc., Pittsburgh PA, 2003.
See the Supporting Information for the full citation.
7280 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7277 –7280