theoretical study of spin-alignment control in molecular magnets
TRANSCRIPT
Current Applied Physics 4 (2004) 539–542
www.elsevier.com/locate/cap
Theoretical study of spin-alignment control in molecular magnets q
Ping Huai *, Yukihiro Shimoi, Shuji Abe
Research Consortium for Synthetic Nano-Function Materials Project (SYNAF) and Nanotechnology Research Institute (NRI),
National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568, Japan
Received 20 November 2003; accepted 30 January 2004
Available online 24 March 2004
Abstract
We investigated theoretically the electronic control of spin-alignment in p-conjugated molecular magnets, stimulated by the
recent successful observation of the spin control in novel diradical molecules. To clarify the mechanism of the spin alignment, a
microscopic model was designed for a p-electron system interacting with two stable radical groups. We demonstrated that electronic
doping induces a transition from low-spin to high-spin states, depending on topological structures of molecules and the strength of
interaction. In the doped case, spins align in a different way from that predicted by the topological rule established in the half-filled
case. These theoretical results provide useful insights into molecular design of organic magnets controllable by external stimuli.
� 2004 Elsevier B.V. All rights reserved.
PACS: 75.50.Xx; 71.10.Fd; 75.20.Hr
Keywords: Molecular magnets; Spin-alignment control; Kondo–Peierls–Hubbard model; Exchange interaction
1. Introduction
Molecular magnetism has attracted considerable
interests for decades since the discovery of high spin
ground states in dicarbene [1,2]. Recently a new class of
purely organic molecular magnets has received increas-
ing attention as their spin alignment is controllable byexternal stimuli such as charge doping or photoexcita-
tion [3,4], as schematically shown in Fig. 1. Among these
pioneering studies, Izuoka and his co-workers have
succeeded in controlling intramolecular spin alignment
by charge doping in a newly designed organic molecule,
thianthrene bis(nitronyl nitroxide) [3]. This molecule,
consisting of p-conjugated moiety and two stable radi-
cals, is spin singlet (S¼ 0) in its ground state and be-comes spin quartet (S¼ 3/2) upon one-electron
oxidation. The control of spin alignment by photoexci-
tation has also been demonstrated in a similar type of
molecules by Teki et al. [4].
qOriginal version presented at QTSM&QFS 2003 (Quantum
Transport Synthetic Metals & Quantum Functional Semiconductors),
Seoul National University, Seoul, Korea, 20–22 November 2003.* Corresponding author. Fax: +81-298-61-5400.
E-mail address: [email protected] (P. Huai).
1567-1739/$ - see front matter � 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.cap.2004.01.019
The dominant mechanism of spin alignment in the
ground state of magnetic molecules has been well
established as a topological rule [5–7], based on the
dynamical antiferromagnetic spin polarization effect of
p electrons with on-site Coulomb repulsion. However,
the topological rule cannot be applied to doped or ex-
cited molecules, and most discussions on the spin-alignment control so far have been given on the basis of
individual molecular orbitals [8–12].
On the other hand, to design such novel functional
molecular magnets, it is necessary to clarify the mech-
anism of spin alignment not only in the ground state,
but also in the doped or excited states. Therefore in the
present study, we tried to shed light on the spin-align-
ment control by charge doping from the general point ofview.
2. Model and theoretical treatment
A microscopic model Hamiltonian approach [13] is
applied to study novel molecular magnets: a Peierls–
Hubbard model with N sites and Ne electrons, which is
coupled with two localized spins. That model is given in
the following Hamiltonian:
High Spin
Charge Doping Photoexcitation
Low SpinJ J
J J
π
hνe
Fig. 1. Schematic picture of spin-alignment control by charge doping
or photoexcitation.
(a)
S =0
S =3/2
R
RNe = N
Ne = N − 1
R
R
(b)
S =0
S =1/2
R
RNe = N
Ne = N − 1
R
R
Fig. 2. Schematic picture of spin-alignment control by hole doping in
polyene-based molecular magnets.
540 P. Huai et al. / Current Applied Physics 4 (2004) 539–542
H ¼ �X
i;s
t0ð1þ DiÞðCyi;sCiþ1;s þ h:c:Þ þ
X
i
Uni;"ni;#
þX
i
t02k
ðDi � Diþ1Þ2 �J2ðCy
i1rCi1 ST1
þ Cyi2rCi2 ST2Þ: ð1Þ
The Peierls–Hubbard model corresponds to the p-con-jugated moiety of an organic molecule, while the local-
ized spins (ST1 and ST2) correspond to the unpaired
electrons of stable radical groups (R) attached to it.
In the p electron system, the nearest-neighbor transfer
integral t0 is modified by the dimensionless lattice
deformation Di in the form of Su–Schrieff–Hegger (SSH)coupling. The electron–electron correlation is taken into
account by using on-site Coulomb repulsion energy U .
The lattice deformation is treated classically and its
elastic constant is given by the inverse of the dimen-
sionless SSH electron–lattice coupling constant k.The two localized spins interact indirectly to each
other through the exchange interaction (J ) with p elec-
trons, while a direct spin–spin interaction is neglected inthis model. We exactly diagonalized the electronic and
spin parts of the Hamiltonian by the Lanczos algorithm,
taking into account all of the correlation effect. The
lattice deformation is optimized by means of the Hell-
mann–Feynman force equilibrium condition.
Due to the electron–hole symmetry in this model,
electron doping (Ne ¼ N þ 1) gives the same results as
hole doping (Ne ¼ N � 1). Thus only the results for holedoping will be discussed in this paper.
3. Results and discussion
The main target of this study is to determine the
intramolecular spin alignment before (Ne ¼ N ) and after
(Ne ¼ N � 1) one electron oxidation in the p electronsystem. Two kinds of p-conjugated moieties are studied:
polyene with 10 carbons and anthracene with 14 car-
bons. Typical results are shown below with the follow-
ing parameters (in the unit of t0): U ¼ 2:5, J ¼ 0:2(ferromagnetic) while k ¼ 0:2 for polyene and k ¼ 0:0for anthracene.
Fig. 2 schematically shows the change in spin align-
ment in polyene-based molecular magnets obtained bythe exact-diagonalization calculations. Two molecules
with different radical positions are chosen in Fig. 2(a)
and (b) to examine the topological effect. In Fig. 2(a),
the half-filled ground state (Ne ¼ N ) is a spin singlet
(S¼ 0) with antiparallel alignment of the radical spins,
and bond alternation appears in the lattice. If an elec-
tron is removed from the p-conjugated moiety, the total
spin turns to be quartet (S¼ 3/2) with parallel alignmentof radical spins, leading to controllable spin alignment
between the low- and high-spin configurations by charge
doping.
In Fig. 2(b), the spin alignment of the half-filled
ground state is quite similar to that of the case (a), as
predicted by the topological rule. However, the doped
molecule is a spin doublet (S¼ 1/2) quite different from
that of case (a). Although the radical spins are in parallelalignment, they are antiferromagnetically coupled to the
net p spin, resulting in the spin doublet. Such doping-
induced spin alignment with different p topology cannot
be explained based on the topological rule.
Next we proceed to the spin-alignment control in
anthracene-based molecular magnets. To investigate the
topological effect, we change the position of one of the
radicals. Fig. 3(a) shows that the half-filled ground stateis a spin singlet (S¼ 0) with antiparallel alignment of
radical spins. Single hole doping into the p-electronsystem induces the spin alignment into parallel, resulting
in a transition from the low pin state (S¼ 0) to the high
spin one (S¼ 3/2).
Fig. 3(b) shows that the half-filled ground state is a
spin triplet (S¼ 1), while the doped molecule is a spin
quartet (S¼ 3/2). The spin alignment of the half-filledground state can be easily understood in terms of the
topological rule. However it is very interesting that the
Fig. 4. Spin density in doped Hubbard model for (a) polyene (b)
anthracene. The radius of circles is proportional to the absolute value
of spin density. Filled (open) circles correspond to positive (negative)
values.
S i1 S i2
J > 0 J > 0
S= 3/2
S i1 S i2
J > 0 J > 0
S=1/2
(a)
(b)
Fig. 5. Schematic picture of intramolecular spin alignment of doped
system with ferromagnetic coupling (J > 0): (a) S¼ 3/2 (b) S¼ 1/2.
R R
Ne =N−1R R
Ne =N
S=1
S=3/2
(b)
R
R
Ne =N−1R
RNe =N
S=0
S=3/2
(a)
Fig. 3. Schematic picture of spin-alignment control by hole doping in
anthracene-based molecular magnets.
P. Huai et al. / Current Applied Physics 4 (2004) 539–542 541
high spin also appears in the doped molecule similar to
the case in Fig. 3(a). In fact, we found more doping-
induced high spin (S¼ 3/2) in many other topological
structures in the anthracene case.
To understand the spin alignment in doped case, we
investigated spin density in p-conjugated moieties using
the Hubbard model. Fig. 4 shows the spin density in the
doped states (N" ¼ N=2, N# ¼ N=2� 1) for both poly-ene and anthracene with U=t0 ¼ 2:5. The electron–lat-
tice coupling in polyene is omitted as the dimerization is
fairly weakened in the doped case.
Since a down-spin electron is removed from the
p-conjugated moiety, all the spin density should be po-
sitive if the on-site Coulomb interaction is absent.
Nevertheless with the finite U , negative spin density
appears in polyene shown in Fig. 4(a). In terms of thespin density, the spin alignment can be justified in a
perturbative way shown in Fig. 5. The radical spin
prefers to be parallel to the p spin at the site to which the
radical is attached. If both attached sites have positive
spin density, the radical spins are in the parallel align-
ment with the net p spin, as shown in Fig. 5(a). On the
other hand, if the attached sites have negative spin
densities, the radical spins are in the antiparallel align-
ment with the net p spin, as shown in Fig. 5(b).
The spin alignment in anthracene can be clarified in
the similar way. If we take a closer look at the spindensity in Fig. 4(b), negative values appear only in the
sites that are impossible for radical to attach. Therefore
the radical spins always prefer parallel alignment with
the net spin in p-conjugated moiety, unless the exchange
coupling J is strong, leading to spin quartet (S¼ 3/2).
4. Summary
In summary, the spin-alignment control by charge
doping in p-conjugated molecular magnets has beenstudied in the theoretical model of the Peierls–Hubbard
model coupling with two radicals spins. Two kinds of p-conjugated moieties, polyene and anthracene, are
investigated to clarify the structural dependence of
intramolecular spin alignment. It is demonstrated that
electronic doping induces low-spin to high-spin transi-
tion, which depends on topological structures of mole-
cules. The spin density provides useful information todetermine the manner of spin alignment in the doped
case where the topological rule cannot be applied.
References
[1] K. Itoh, Chem. Phys. Lett. 1 (1967) 235.
[2] E. Wasserman, R.W. Murray, W.A. Yager, A.M. Trozzolo,
G. Smolinsky, J. Am. Chem. Soc. 89 (1967) 5076.
[3] A. Izuoka, M. Hiraishi, T. Abe, T. Sugawara, K. Sato, T. Takui,
J. Am. Chem. Soc. 122 (2000) 3234.
[4] Y. Teki, S. Miyamoto, M. Nakatsuji, Y. Miura, J. Am. Chem.
Soc. 123 (2001) 294.
[5] N. Mataga, Theoret. Chim. Acta 10 (1968) 372.
[6] A.A. Ovchinnikov, Theoret. Chim. Acta 47 (1978) 297.
[7] E.H. Lieb, Phys. Rev. Lett. 62 (1989) 1201.
542 P. Huai et al. / Current Applied Physics 4 (2004) 539–542
[8] H. Sakurai, A. Izuoka, T. Sugawara, J. Am. Chem. Soc. 122
(2000) 9723.
[9] Y. Teki, S. Miyamoto, M. Nakatsuji, Y. Miura, J. Am. Chem.
Soc. 123 (2001) 294.
[10] M. Matsushita, T. Nakamura, T. Momose, T. Shida, Y. Teki, T.
Takui, T. Kinoshita, K. Itoh, J. Am. Chem. Soc. 114 (1992) 7470.
[11] S. Yamanaka, T. Kawakami, M. Okumura, K. Yamaguchi,
Chem. Phys. Lett. 233 (1995) 257.
[12] H. Mizouchi, A. Ikawa, H. Fukutome, J. Phys. Soc. Jpn. 66 (1997)
881.
[13] P. Huai, Y. Shimoi, S. Abe, Phys. Rev. Lett. 90 (2003)
207203.