1:1 Complexes of 5-(4-[N-tert-butyl-N-aminoxyl]phenyl)pyrimidine with manganese(II) and copper(II) hexafluoroacetonylacetonate

Article abstract of DOI:10.1039/b111295n

Mn(hfac)₂ and Cu(hfac)₂ form 1:1 complexes with 5-(4-[N-tert-butyl-N-aminoxyl]phenyl)pyrimidine that exhibit strong metal-nitroxide exchange; spin polarization models do not explain the antiferromagnetic exchange behavior between spi

Full text of DOI:10.1039/b111295n

1+1 Complexes of 5-(4-[N-tert-butyl-N-aminoxyl]phenyl)pyrimidine with
manganese(II) and copper(II) hexafluoroacetonylacetonate†
Lora M. Field,a Paul M. Lahti*a and Fernando Palaciob
a Department of Chemistry, University of Massachusetts, Amherst MA 01003, USA.
E-mail: lahti@chem.umass.edu
b Instituto de Ciencia de Materiales de Aragón, CSIC – Universidad de Zaragoza, 50009, Zaragoza, Spain.
E-mail: palacio@posta.unizar.es
Received (in Purdue, IN, USA) 7th December 2001, Accepted 24th January 2002
First published as an Advance Article on the web 25th February 2002
Mn(hfac)2 and Cu(hfac)2 form 1+1 complexes with 5-(4-[N-
tert-butyl-N-aminoxyl]phenyl)pyrimidine that exhibit
strong metal–nitroxide exchange; spin polarization models
do not explain the antiferromagnetic exchange behavior
between spin sites in these complexes.
Cu–O bond lengths, 2.10–2.30 Å. The nitroxide Cu–ON bond is
identifiably axial in 3, and we consider the Mn–ON bond in 2 to
be analogous, although the lower molecular symmetry of 2
makes harder a clear geometric differentiation of the equatorial
and axial orientations. The interannular torsional angle in 2 is
29–30°, while that in 3 is 25°. There is substantial torsion about
the nitroxide to phenyl N–C bond in both 2 (40°, 45°) and 3
(27°). The metal–metal distances within the complexes are
fairly long: r(Mn–Mn) = 10.4 Å in 2, r(Cu–Cu) = 10.5 Å in 3.
The inter-dimer metal–metal distances are shorter: the closest
intermolecular r(Mn–Mn) = 6.1 Å in 2, r(Cu–Cu) = 8.2 Å in
3.
Dc magnetic susceptibility measurements were carried out on
samples of 2 and 3 in a Quantum Design MPMS SQUID
magnetometer over a range of 1.8–300 K at 1000 oersted (0.1
tesla). Fig. 2 shows plots of the corrected paramagnetic
susceptibility c as a function of absolute temperature T for each
complex, as well as 1/c = f(T) (Curie–Weiss) plots of the same
data.
Although 2 shows nearly linear Curie–Weiss behavior, 3
shows substantial deviation indicative of AFM interactions
between spin sites. The high temperature limiting slopes of the
Curie–Weiss plots are C = 0.667 and 0.977 emu K Oe21
mol21, corresponding to S = 2.12 and 0.98 for 2 and 3,
respectively. In the case of 2, this corresponds to S  (5/2) 2
(1/2), or strong AFM exchange between the manganese ion and
the nitroxide radical spin. For 3, S  (1/2) + (1/2), correspond-
ing to a strong FM exchange between the copper ion and the
nitroxide. The T > 100 K regions of the Curie–Weiss plots yield
An important and promising strategy for the development of
hybrid organic–inorganic molecule-based magnetic materials is
the coordination of paramagnetic ions with organic open-shell
radicals and polyradicals.1 In such systems, the exchange
between the ion and the radical may be either ferromagnetic
(FM) or antiferromagnetic (AFM), and still yield materials with
net magnetic moments, so long as the spin quantum numbers of
the building blocks are unequal. Here we report the crystallo-
graphic and preliminary magnetic studies of the coordination
complexes formed between the stable radical 5-(4-[N-tert-
butyl-N-aminoxyl]phenyl)pyrimidine (1) and Mn(hfac)2 and
Cu(hfac)2.
Stable red radical 1 was synthesized as shown in Scheme 1,
and reacted separately with one equivalent each of Mn(hfac)2
and Cu(hfac)2 in mixed solvent systems under slow crystalliza-
tion conditions to give the 1+1 complexes 2 and 3 with good
elemental analysis. The structures of these complexes were
confirmed by X-ray crystallography.†‡ Over a period of
months, there was no apparent decomposition of 2 or 3 when
these were stored in the absence of strong light.
Fig. 1 shows ORTEP diagrams of 2 and 3. Both are cyclic
dimer type complexes in which the nitroxide spin site and one
pyrimidine coordination site are complexed. Multiple efforts to
make different complexes with molar ratios of 1+M(hfac)2
<
1.0 have failed to yield anything other than 1+1 complexes. In
2, all of the Mn–O bond lengths are 2.10–2.30 Å; in 3, the Cu–
ON bond length is 2.40 Å, substantially longer than the other
Scheme 1
† Electronic supplementary information (ESI) available: Spectroscopic data
and crystallography for 1. Fig. S1: ESR spectrum of 1. See http://
www.rsc.org/suppdata/cc/b1/b111295n/
Fig. 1 ORTEP diagrams of 2 and 3. Hydrogen and fluorine atoms are
omitted for clarity.
636
CHEM. COMMUN., 2002, 636–637
This journal is © The Royal Society of Chemistry 2002

pyrimidine ring of > 0.1 G. Still, we hoped that coordination of
the paramagnetic cations at the pyrimidine ring would induce
spin polarization in the pyrimidine. If this is occurring, it is
insufficient to induce the parity-expected FM exchange be-
tween the M–NIT spin sites within the dimer. As a result, 2 and
3 do not fit a regiospecific, intramolecular spin polarization
model of exchange between M–NIT sites in the dimer. Also, the
strength of exchange in 2 and 3 is not simply a function of the
distance between M–NIT spin sites—the intramolecular dis-
tance is nearly the same, the intermolecular distance is smaller
in 2, but 3 shows the largest exchange effect. We think it likely
that intermolecular mechanisms—or lack thereof—dominate
exchange between the M–NIT spin sites in 2 and 3, and we hope
to report full details in the near future. Overall, despite the allure
of spin-parity models to explain and predict exchange and
magnetic behavior in molecular magnetic systems, the present
work shows that such connectivity models do not invariably
have predictive values. The models are useful guides, but not
iron-clad rules.
This work was supported by the National Science Foundation
(NSF CHE-9809548 and CHE-0109094), the Comision Inter-
ministerial de Ciencia y Tecnologia (CICYT-MAT2000-
1388-C03-03), and the Fullbright España commission. The
authors thank Dr. A. Chandrasekaran of the University of
Massachusetts X-ray Structural Characterization Center (NSF
CHE-9974648) for assistance with crystallographic analyses,
and Dr. G. Dabkowski of the Microanalytical Laboratory for
elemental analyses.
Fig. 2 Paramagnetic susceptibility (c) vs. temperature plots for 2 and 3, a
and b, respectively. Inset figures show Curie–Weiss plots of 1/c vs. T for the
same data.
Weiss constants of q = 20.15 K and 20.35 K for 2 and 3,
respectively.
It has been shown elsewhere that Mn–radical exchange is
typically AFM in nature, while Cu–nitroxide exchange is
typically FM when the coordination is axial.1b,2 The strength of
the M–NIT exchange interactions in 2 and 3 is notable. Based
on the plots in Fig. 2, J/k(Mn–NIT) < 2300 K and J/k(Cu–
NIT) > 300 K. However, only for 3 are there significant
interactions between the M–NIT spin sites.
Based on the crystal structures, it was assumed that the
magnetic behaviors of 2 and 3 could be fitted to a Bleaney–
Bowers dimer type model.3 For 2, we used the model described
by Ishimaru et al.4 for interactions between S = 2 spin sites. For
3, we used Eqn. 1,
Notes and references
‡
For (2): mp 142–143 °C. Anal. Calcd. for C24H18N3O5F12Mn: C, 40.04;
H, 2.52; N, 5.84. Found: C, 40.28; H, 2.63; N, 5.64%. Dark red needles from
ether/hexane. Crystal data for 2: 0.50 3 0.15 3 0.05 mm, formula =
C24H18N3O5F12Mn, M = 1422.71, monoclinic, space group P21/c (#14), Z
= 4, a = 9.888(9), b = 28.922(5), c = 21.273(1) Å, b = 78.339(7)°, V =
5958.8(2) Å3, Dcalc = 1.585 g cm23, T = 293 K, l(Mo-Ka) = 0.7107 Å,
N0g1g2m2B
exp(-2J / kT)
1 + 3exp(-2J / kT) T -q
1
(1)
.
c =
¥
k
m(Mo-Ka)
=
0.557 mm21. 12102 Reflections were measured at an
where J is the exchange coupling between M–NIT spin sites, g1
and g2 are g-factors of the different spin components, k is the
Boltzmann constant, mB is the Bohr magneton constant, and q is
a mean field term. For 2, g1 = g2 = 2.00 and q = 0.0 K was
assumed, leading to an excellent fit for c = f(T) with 2J/k =
20.45 K. For 3, satisfactory fitting required a non-zero mean-
intensity threshold of 2s(I). 6164 Independent reflections (Rint = 0.1341)
were analyzed with 511 parameters using the program SHELXL-97. The
final wR(F2) was 0.3377. The major source of disorder is in the CF3 groups
of the hfac ligands.
For (3): mp 132–134 °C. Anal. Calcd. for C24H18N3O5F12Cu: C, 40.52;
H, 2.55; N, 5.91. Found: C, 40.65; H, 2.48; N, 5.88%. Dark red plate from
ether/hexane. Crystal data for 3. 0.50 3 0.15 3 0.05 mm, formula =
C24H18N3O5F12Cu, M = 1439.91, monoclinic, space group C2/c (#15), Z =
4, a = 13.5050(2), b = 14.5352(2), c = 33.6543(5) Å, b = 100.9537(5)°,
V = 6485.91(16) Å3, Dcalc = 1.475 g cm23, T = 293 K, l(Mo-Ka) =
0.7107 Å, m(Mo-Ka) = 0.778 mm21. 10456 Reflections were measured at
field field term. Fig. 2 shows the fit where g1 = 2.0 and g2
2.2, q = 21.1 K and 2J/k = 29.9 K.
=
Despite an intermolecular approach of 6.7 Å between the Mn
atoms in 2, there is only a weak, AFM interaction between the
Mn–NIT spin sites. By comparison, a twenty-fold larger AFM
interaction is found in 3 as readily seen by the maximization and
downturn of a c vs. T plot at about 7 K. In an effort to rationalize
the observed magnetic behavior, we compared our results to
those for the structural analogue 4, previously investigated by
Ishimaru et al.4 System 4 exhibits an upturn in magnetization,
corresponding to a modest FM exchange interaction of 2J/k =
+1.18 K between strongly exchange linked, S = 2 Mn–NIT
sites. By contrast, meta-linked 5 exhibited a downturn in
magnetization corresponding to AFM exchange coupling
between the Mn–NIT sites of 2J/k = 20.45 K. The authors
attributed the differing behavior to regiospecific intramolecular
spin-polarization exchange effects transmitted through the
organic radical p-systems, as in M–pyridyl–nitroxide com-
plexes described elsewhere.5 Although the exchange inter-
actions in 4 and 5 could be attributed to through space effects,
it is less obvious how to analyze these in structure–property
terms.
an intensity threshold of 2s(I). 5710 Independent reflections (Rint
=
0.1280) were analyzed with 382 parameters using the program SHELXL-
97. The final wR(F2) was 0.3634. The major source of disorder is in the CF3
groups of the hfac ligands. CCDC reference numbers are 178242–178244
for 1–3. See http://www.rsc.org/suppdata/cc/b1/b111295n/ for crystallo-
graphic data in CIF or other electronic format.
1 (a) O. Kahn, Molecular Magnetism, VCH, New York, NY, 1993; (b) D.
Gatteschi, in Magnetic Properties of Organic Materials, ed. P. M. Lahti,
Marcel Dekker, New York, NY, 1999, p. 601ff; (c) Molecular
Magnetism: New Magnetic Material, ed. K. Itoh and M. Kinoshita,
Gordon & Breach, Amsterdam, The Netherlands, 2000, pp. 304–337; (d)
H. Iwamura, K. Inoue and T. Hayamizu, Pure Appl. Chem., 1996, 68,
243.
2 H. Iwamura, in Magnetic Properties of Organic Materials, ed. P. M.
Lahti, Marcel Dekker, New York, NY, 1999, pp. 641–644 and references
therein.
3 B. Bleaney and K. D. Bowers, Proc. R. Soc. London, Ser. A, 1952,
214.
4 Y. Ishimaru, K. Inoue, N. Koga and H. Iwamura, Chem. Lett., 1994,
1693.
5 (a) H. Iwamura and N. Koga, Mol. Cryst. Liq. Cryst., 1999, 334, 437; (b)
M. Kitano, Y. Ishimaru, K. Inoue, N. Koga and H. Iwamura, Inorg.
Chem., 1994, 33, 6012.
Although systems 2 and 3 are connectivity analogues of 4,
they qualitatively show AFM instead of FM exchange. Possibly,
spin polarization is stronger in 4 and 5 than in 2 and 3. Solution
ESR spectral analysis of 1† shows no hyperfine coupling in the
CHEM. COMMUN., 2002, 636–637
637