Magnetism of fullerene charge-transfer complexes

farrowbrainUrban and Civil

Nov 15, 2013 (3 years and 7 months ago)

124 views

CHAPTER 00

Magnetism of fullerene charge
-
transfer
complexes

Ales Omerzu
1

and Madoka Tokumoto
2

1
Jozef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia

2
Nanotechnology Research Institute, Natl Inst. of Advanced Industrial Science and
Technology (AIST), 1
-
1
-
1 Umezono, Tsukuba, Ibaraki 305
-
8568, Japan

1 Introduction

The research on magnetic properties of fullerene charge
-
transfer (CT)
complexes was sparked i
n 1991 when Fred Wudl’s group in Santa Barbara, in their
course of studying a reduction of fullerenes with strong organic donors, discovered a
compound tetrakis(dimethylamino)ethylene
-
C
60

(TDAE
-
C
60
), which surprisingly
showed a ferromagnetic transition at
16 K
1
. That was a temperature, which exceeded
Curie temperatures of any other pure organic material known so far by more then order
of magnitude. Since 1991 a lot of experimental and theoretical work has be done
towards understanding of the ferromagnetic o
rdering in TDAE
-
C
60
. The most important
results will be presented in Section 2.

The discovery of TDAE
-
C
60

motivated researchers from several laboratories
around the world to try to synthesize new organic, fullerene
-
based ferromagnets. Their
approaches coul
d be divided into four main groups: (1) reduction of higher fullerens
with TDAE, (2) reduction of C
60

with different organic or organometallic donors, (3)
functionalization of C
60

and subsequent doping and (4) complexing C
60

with rare earth
elements. Each
of those approaches and their results will be presented in following
sections.

2 TDAE
-
C
60

In the original work
1

the authors synthesized TDAE
-
C
60

by adding TDAE
(l
iquid) to a toluene solution of C
60
. The result was black, microcrystalline powder,
which precipitated from the solution almost instantaneously after adding of TDAE. The
material was highly air sensitive and all manipulation had to be done in a glove box
w
ith an inert atmosphere. According to elemental analysis it was claimed that
stoichiometry was 1:1.16, which was later proven to be false. TDAE and C
60

react in a
simple 1:1 ratio to form a charge transfer salt TDAE
+
C
60
-
. Measurements of temperature
and fi
eld dependence of magnetisation demonstrated a clear transition to the
ferromagnetic phase below
T
c

= 16 K. The magnetisation increased abruptly below
T
c

and
T
c

increased with the measuring field. The field dependence of magnetisation below
T
c

displayed an
S

curve, characteristic to ferromagnets but with no observable
hysteresis. Magnetic susceptibility measurements at high temperatures (
T

> 30 K)
revealed almost temperature
-
independent behavior, quite distinct from the Curie
-
Weiss
behavior


=

C/(T
-

)
, which is expected for systems of localised magnetic moments.
Furthermore, the electric conductivity measured on a compressed pellet turned out to be
quite high, 10
-
2

Scm
-
1
. From those findings the authors concluded that TDAE
-
C
60

might
be an itine
rant soft ferromagnet.

Soon after discovery, the first X
-
ray diffraction study on TDAE
-
C
60

was
performed in Brookhaven National Laboratory
2
. It was demonstrated that TDAE and
C
60

crystallize in 1:1 stechiometric ratio, indeed. The structure was determined
to be
c
-
centered monoclinic (space group
C2/m
). The intramolecular C
60
-
C
60

separation is
shortest along the
c

axis (9.98 Å) and much greater in the
a
-
b

plane (10.25 Å). The two
-
fold axis of TDAE molecules also orients itself along
c

axis. From the structur
al
parameters it could be concluded that TDAE
-
C
60

has an anisotropic, low dimensional
band structure, which could account for unusual electronic and magnetic properties.

Little later, experiments of Tanaka’s group showed a slightly different picture
3
.
From

ESR

measurements they found that unpaired spins reside mostly on C
60

molecules
(the
g
-
value 2.0003 of TDAE
-
C
60

is much closer to 1.999 for electrochemically
prepared C
60
-

then 2.0036 for TDAE
+
). They also predicted the Jahn
-
Teller distortion of
C
60
-

or, i
n other words, polaron formation in C
60
. In contrast to original findings they
observed the Curie
-
Weiss behavior of the magnetic susceptibility and from the shape of
M(H)

curves at 4.5 K they calculated that magnetic moments at low temperatures form
cluste
rs with an average size of 170 spins per cluster. Evidently, TDAE
-
C
60
, at least in
it’s powder form, showed magnetic properties of a superparamagnet.

At this point it was not yet clear if TDAE
-
C
60

is a “proper” ferromagnet, but
following experiments cleare
d that issue. Suzuki et al.
4

and Dunsch et al.
5

found that
TDAE
-
C
60

could exhibit a hysteresis curve although with a very small coercive field,
H
c
~ 2 Oe and a remanent magnetic moment
M
r

~ 3x10
-
4

emu. In addition, Suzuki et al.
4

also measured the
AC

susceptibility with a nonzero imaginary part, i.e. energy losses,
related to the hysteresis. The final proof for the ferromagnetic state in TDAE
-
C
60

came
from a zero f
ield

SR

experiment by Lappas et al.
6

The authors demonstrated the
existence of an internal magnetic field of 68 Gauss with very broad distribution (48
Gauss) reflecting spatial inhomogenity.

The

SR

experiment wasn’t the only one, which indicated inhomoge
nities,
structural or magnetic, in the system. Even in the early stage of the research some
experimental results appeared to be in contradiction with the hypothesis of a long
-
range
ferromagnetic ordering in TDAE
-
C
60
. In the first place, the temperature dep
endence of
the
ESR

line
-
width showed a relatively small line broadening and no frequency shift
7

with a non
-
exponential and very slow decay of the magnetization
8
. Both of these
features are characteristic of random magnetic systems without a long
-
range orde
r.
Hence, it was suggested that TDAE
-
C
60

could be a spin glass. After experiment by
Mihailovic et al.
9
, which established a direct connection between orientational degrees
of freedom of C
60

molecules and magnetic interactions in the system, this hypothesis

seemed to be even more plausible (this connection was latter demonstrated also by
theoretical calculations
10
,
11
). By freezing C
60

molecules in random orientations one can
obtain a distribution of exchange interactions in the system, and consequently, magn
etic
disorder and frustration


two essential conditions for a spin glass. Later measurements
of linear and non
-
linear susceptibilities
12

partly confirmed the spin glass hypothesis.
The linear susceptibility

1

exhibited a broad peak centered at 10 K and t
he non
-
linear
susceptibilities

3
,

5

and

7

diverged at the same temperature. The only feature, which
deviated from the spin
-
glass behavior, was the absence of any shift of the peak position
with frequency, which is so characteristic for spin glasses. Obv
iously, TDAE
-
C
60

was
showing some characteristics of spin glasses and some of ferromagnets and it wasn’t
inconceivable that both phases coexist in a sample.

At that stage of research it was evident that lot of questions on the nature of
TDAE
-
C
60

(ferro)magnetism remained, which couldn’t be answered by experiments on
powder samples. A reproducibility of physical properties for powder samples was
unsatisfactory even for the samples from one and the same group. Due to different
sample purity, which
was mainly affected by solvent inclusion into the crystal structure
and oxygen contamination, as well as varying grain sizes in the powder samples,
TDAE
-
C
60

exhibited an inconsistent behaviour. The samples were changing their
properties even by aging, usua
lly by increasing their ferromagnetic signal. It was clear
that for making any progress in understanding of ferromagnetism in TDAE
-
C
60

monocrystals were essential. The first attempt in growing single crystals was done by
Suzuki et al.
13
, but their crystals

were rather small (0.3 mm in length, 0.05 mm in
diameter) and of poor quality. So, it was impossible to determine a crystal structure.
Finally, the crystals even didn’t show the ferromagnetic transition.

A similar approach for the single crystal growth b
y diffusion method was
adopted by one of the authors (A.O.). That time results were much better. Firstly, the
crystals showed a ferromagnetic transition at 16 K. Secondly, by improving the method,
which included a reducing of diffusion speed by smarter des
ign of a crystal
-
growing
cell and a temperature control, it was possible to obtain high quality single crystals of
millimeter size. That breakthrough paved the way for experiments, which have
followed.




Figure 1.

Three different views on the TDAE
-
C
60

c
rystal structure along
b
,
a

and
c

axis.



Having macroscopic single crystals available, one of the first questions, which
should be answered, was a mechanism of the electrical conductivity. Although
microwave conductivity measured by Schilder et al.
14

and
optical conductivity
measured by Bommeli et al.
15

confirmed an insulating behavior of TDAE
-
C
60
, their
experiments where performed on microcrystalline samples with grains typically 10


100 nm in size. Large surface to volume ration and material’s high air
sensitivity
obviously impair a clear discrimination between the insulating and the metallic intrinsic
conducting state of TDAE
-
C
60
. Omerzu et al.
16

circumvented that problem by
measuring the
AC

and the
DC

conductivity on single crystals of TDAE
-
C
60

with di
rect
electrical contacts. They found that conductivity could be decomposed into two
components: frequency
-
dependent, temperature
-
independent tunneling and temperature
-
dependent phonon
-
assisted hopping. A dynamic, rotational disorder of C
60

molecules
plays
a key role in the conductivity. The conductivity shows a crossover at
T
0

= 150 K.
It is a temperature, which separates the high
-
temperature orientationally disordered state
from the low
-
temperature ordered state as it was demonstrated by
13
C NMR
measuremen
ts
17
. The hopping mechanism prevails at
T

>
T
0

where the hopping
probabilities are higher, but at
T

<
T
0

the tunneling is a more efficient conducting
channel.






Figure 2.

The
DC

conductivity of TDAE
-
C
60

single crystal as a function of
temperature.
The full squares were measured for cooling at rate 0.1 K/min, while the
open circles were measured in near
-
quench conditions, 33 K/min.




Figure 3.

Temperature dependence of the second (a) and the first (b) moment of the
13
C NMR spectra in powdered
TDAE
-
C
60
.



Single crystals, in contrast to powder samples, have reproducible physical
properties and offer a possibility for another intricate property of TDAE
-
C
60

to be also
explained. Namely, the powder samples frequently showed much lower magnetization

as expected, and even worse, it’s value for particular sample changed with aging. To
resolve that intricacy, Mrzel et al.
18

choosed TDAE
-
C
60

crystals grown at 10

C, which
when fresh show no ferromagnetic signal at low temperatures. They treated the sample
s
in several heating cycles at temperatures between 50

C and 110

C. After each heating
cycle they measured temperature dependence of the
ESR

signal. They observed a sharp
increase in the intensity of the ferromagnetic signal after the sample was treated at

70

C
or higher. The ferromagnetic signal eventually disappears when the sample was heated
above 100

C. From then on it has been clear that TDAE
-
C
60

can exist in at least two
crystal modifications: the usual or

-
TDAE
-
C
60

which has a ferromagnetic phase be
low
16 K and newly discovered


-
TDAE
-
C
60

modification without the ferromagnetic
phase. The

’ modification is the metastable one and can be irreversibly transformed
into the stable


modification by thermal treatment.


Figure 4.

The field dependence (le
ft) and the temperature dependence (right) of
magnetisation of

-
TDAE
-
C
60
.



When the existence of two different modifications of TDAE
-
C
60

was firmly
established, researchers started with experiments, which would determined the nature of
the two TDAE
-
C
60

m
odifications magnetic ground states. Arcon et al.
19

measured a
ferromagnetic resonance in

-
TDAE
-
C
60
. By using the low
-
field
ESR

technique they
showed a nonlinear variation of the resonance frequency,


with resonance field,
H

and
proved the existence of l
ong range magnetic order. From the


versus
H

dependence
they were able to rule out an antiferromagnetic behavior as well as a paramagnetic or a
spin
-
canted one. From an extremely low value of anisotropy field (29 Gauss) they
concluded that

-
TDAE
-
C
60

is

an example of an easy axis Heisenberg ferromagnet with
the easy axis along crystallographic
c
-
axis, the axis of the closest C
60
-

approach.

Another insight to the nature of the ferromagnetic transition in

-
TDAE
-
C
60

offered measurements of the critical beh
avior near the ferromagnetic phase transition
point by Omerzu et al.
20

The authors presented results of independent measurements of
the static critical exponents for susceptibility

(T) ~ (T/T
c
-
1)
-

, spontantenuos
magnetization
M
s

~ (1
-

T/T
c
)


and critical

isotherm
H ~ M


in the vicinity of the
transition temperature
T
c
. The obtained results


= 1.22


0.02,


= 0.75


0.03 and


=
2.28


0.14 differed significantly from those expected for a 3D Heisenberg ferromagnet,


= 1.38,


= 0.36 and


= 4.8. In addi
tion, the exponents didn’t obey the scaling relation


=


(


-

1)
. The authors found an explanation for such discrepancy in a reduced
effective dimensionality of the system caused by additional degrees of freedom coming
-200
-150
-100
-50
0
50
100
150
200
-0.002
-0.001
0.000
0.001
0.002
2
4
6
8
10
12
14
16
18
20
0.0000
0.0005
0.0010
0.0015


T = 5 K
M (emu)
H (Oe)


H = 10 Oe
T (K)
from C
60

molecular rotation. Those
induce an important degree of randomness into the
system and alter the nature of the ferromagnetic transition.

The presence of intrinsic randomness in

-
TDAE
-
C
60

was clearly demonstrated
in measurements of linear and non
-
linear
AC

susceptibilities by Omerz
u et al.
21

It is
know that for ferromagnetic systems with a relatively low degree of disorder in
magnetic interactions a re
-
entrant spin glass (RSG) transition follows the ferromagnetic
transition at a lower temperature
T
RSG

< T
FM
. Measurements of odd and
even harmonics
of
AC

magnetic response in TDAE
-
C
60

revealed an additional broad peak centered at 7
K, but only for odd harmonics. The reason is that at the spin glass transition the time
reversal symmetry is not broken in contrast to the ferromagnetic tran
sition. Indeed, the
measurements showed a divergence in both odd and even harmonics at
T = T
FM
. The
frequency dependence of the peak in the imaginary part of the linear susceptibility at 7
K gave an additional confirmation for the reentrant spin glass tran
sition. Thus, the riddle
of the coexistence of the long
-
range ferromagnetic order and the short
-
range spin
-
glass
disorder was resolved.

4
6
8
10
12
14
16
18
20
22
0
2
4
6
8
10
4
6
8
10
12
14
16
18
20
-1
0
1
2
3
4
5
6
7
8
9

'(a.u.)
T(K)


''(a.u.)
T(K)
Figure 5.

The temperature dependence of the real (left) and the imaginary (left) parts of
the

linear
AC

susceptibility of

-
TDAE
-
C
60

measured at difrent frequencies between 33
Hz and 3 kHz.


Magnetic properties of


-
TDAE
-
C
60

are much simpler. Measuring the
macroscopic magnetic properties i.e. the temperature and the field dependence of the
magnet
isation, Omerzu et al.

22

showed that

’TDAE
-
C
60

is a paramagnet. The field
-
temperature dependence of the magnetisation exactly follows the Brillouin formula

M = N


tanh (

H / k
B
T
),




where
N

is the number of spins in a sample and


is the magnetic mom
ent ( in the case
of
S

= ½,


=

B



the Bohr magneton). From the formula and the measured
magnetisation and the mass of the sample it was possible to calculate an effective
number of spins per formula unit
N
eff
. It turned out that
N
eff

equals the number o
f C
60
-

ions in the sample. That notion immediately posed a question on missing contribution of
TDAE
+

spins. Additional measurements at higher temperatures showed that
N
eff

increases from 1 to 2 per formula unit as temperature approaches 100 K. A mechanism
which could account for such behavior might be an antiferromagetic correlation among
TDAE
+

spins, which causes the TDAE
+

subsystem of spins to “freeze out” from the
bulk magnetization at low temperatures. It could also explain why antiferromagnetic
correla
tions were frequently observed in measurements of high
-
temperature
susceptibility. However, the role of the TDAE
+

spins in the TDAE
-
C
60

magnetism
remains provocative until now.

0
10
20
30
40
50
0.000
0.005
0.010
0.015
0.020
0.025
T = 2 K
M (emu)
H (kOe)

Figure 6.
Magnetisation of


-
TDAE
-
C
60

as a functi
on of an applied field at
T

= 2 K.
The solid line is the Brillouin function,
M = N


tanh (

H / k
B
T
).

0
50
100
150
200
250
300
0.0
2.0x10
-9
4.0x10
-9
6.0x10
-9
8.0x10
-9
1.0x10
-8
1.2x10
-8
1.4x10
-8



(emu/Oe)
T (K)

Figure 7.
Magnetisation of


-
TDAE
-
C
60

as a function of temperature measured in an
external field of 10 kOe. The solid line i
s the Curie
-
Weiss function,


= C /(T
-

).



The circumstance that TDAE
-
C
60

appears in two modifications with completely
different magnetic properties was a clue for the microscopic understanding of its
magnetism. An irreversible transition from the metasta
ble, nonferromagnetic form


-
TDAE
-
C
60

into the stable, ferromagnetic form

-
TDAE
-
C
60

can be performed in a
controlled way. Usually, crystals of


-
TDAE
-
C
60

are sealed into glass or quartz
capillaries under He. The transformation takes place at 70ºC. It ne
eds 6 hours for
completion and any excess heating can gradually degrade the samples. The whole
procedure can be controlled by measuring magnetization curves at low temperatures
before and after the annealing.

-50
-40
-30
-20
-10
0
10
20
30
40
50
-0.008
-0.006
-0.004
-0.002
0.000
0.002
0.004
0.006
0.008
T = 2 K


M (emu)
H (kOe)

Figure 8.

The field

dependence of the magnetisation of TDAE
-
C
60

before (open
squares) and after (filled squares) annealing. The measurements were performed on the
same single crystal.



Since the transformation occurs at mild conditions one would suppose only
minor structura
l differences between

’and


TDAE
-
C
60
. As it was expected,
Narymbetov and co
-
workers
23

found the two modifications to be structurally
indistinguishable at room temperature. Differences appeared at temperatures below 50
K as additional diffuse lines in a d
iffraction pattern. By further cooling down to 7 K
those lines disappeared in the case of paramagnetic (PM)


-
TDAE
-
C
60

and in the case
of ferromagnetic (FM)

-
TDAE
-
C
60

they coalesced into additional sharp diffraction
spots. Those additional diffraction sp
ots for

-
TDAE
-
C
60

correspond to a primitive unit
cell, indicating that the crystal transformed from a
C
-
centered structure to a primitive
one. A refinement of the crystal structure was possible only after introducing additional
degree of freedom


a relat
ive rotation of C
60

molecules around their three
-
fold axis by
±60º with 50% occupancy. In the PM sample, the relative C
60

orientations are similar to
those encountered in other C
60

solids: the 6
-
6 double bond faces the center of the
hexagon on the neighbou
ring molecule. In the FM sample on the other hand, a new
orientation appears (±60º), which leads to three possible relative configurations of the
C
60
s. However, only configuration in which the neighbouring C
60

are rotated relative to
each other by ±60º is
compatible with 50% occupancy of two rotations determined from
the structural refinement. In that configuration, the double bond on one molecule faces
the center of the pentagon of its neighbour, leading to C
60

molecules ordered along the
c
-
axis with alte
rnating orientations.



-
TDAE
-
C
60





-
TDAE
-
C
60




Figure 9
. A schematic diagram of the C
60

molecular orientations in the
a
-
b

plane for
the PM (left) and the FM (right) structures. The corresponding
C
-
centered and primitive
unit cells in the
a
-
b

plane are shown.


a)





b)




c)





d)




Figure 10.

Projections of two neighbouring C
6
0

units along the
c
-
direction. a) The PM
phase. b) to d) Three possibilities of mutual orientations in the FM phase.

Kambe et al.
24

explored a temperature dependence of

-
TDAE
-
C
60

structure in
more detail. They found that additional diffraction spots, whic
h correspond to the
primitive lattice start to appear at 180 K. They followed an increase of the new Bragg
reflection as the samples temperature decreased. From the smooth increase of the
intensity they concluded that the
C
-
centered to primitive latti
ce structural transition in

-
TDAE
-
C
60

is of the second order.

When it seemed that the relation between TDAE
-
C
60

structure and its magnetic
properties was satisfactory resolved a new discovery appeared. In their investigation of
pressure effect in TDAE
-
C
60

Mizoguchi et al.
25

found not only that
T
c

of the
ferromagnetic transition in

-
TDAE
-
C
60

decreased with increasing pressure and
eventually disappeared at 9 kbar but also that above 10 kbar at 300 K

-
TDAE
-
C
60

polymerised. The new, polymer

-
TDAE
-
C
60

phase
consist of 1D C
60

chains covalently
interconnected by [2+2] cycloaddition in a similar way as in polymer
o
-
Rb
1
C
60
. The
new phase was stable even after pressure release. Garaj et al.
26

measured the
temperature dependence of the

-
TDAE
-
C
60

ESR

signal above 3
00 K and found that

-
TDAE
-
C
60

depolymerised at 520 K. This process was irreversible and the
depolymerized samples showed magnetic properties similar to the ferromagnetic

-
TDAE
-
C
60
. There is another interesting property of

-
TDAE
-
C
60
: it is a paramagnet
w
ith the magnetic susceptibility showing the Curie
-
Weiss temperature dependence but
with twice as many spins as in

-
TDAE
-
C
60
. From a shift in the
ESR

g
-
factor from
2.0005 in

-
TDAE
-
C
60

to 2.0028 in

-
TDAE
-
C
60
, which is much closer to 2.0036 in
TDAE
+

cation

radical, the authors concluded that the missing TDAE
+

spins revived in

-
TDAE
-
C
60
. TDAE
+

spins, which are mutually cancelled in

’ and

-
TDAE
-
C
60

(the
physical origin of that is still unknown) appeared to be localized and noninteracting in
the polymer

-
T
DAE
-
C
60
.


3 Higher fullerens reduced with TDAE

In the early stage of research on TDAE
-
C
60

magnetism it was interesting to
compare it with higher fullerenes reduced with TDAE. It turned out that C
70
, C
84
, C
90

and C
96

readily form charge
-
transfer (CT) comple
xes with TDAE
27
. Their magnetic
properties were characterised mainly with
ESR
. The
g
-
factor and line
-
width of the
ESR

line for all of TDAE
-
higher fullerene samples were almost temperature independent and
the intensity of the
ESR

line, which is proportional

to the spin susceptibility, followed
the Curie low
I

~
C/T
. Hence, the TDAE
-
higher fullerenes CT complexes are simple
paramagnets. From the
g

value the authors concluded that the unpaired spins reside
mainly on fullerene units. Later, Oshima et al.
28

were

able to crystalize TDAE
-
C
70
-
toluene complex and to obtain its crystal structure. In their samples C
70

molecules
formed singly bonded dimers. Magnetically, the crystals were paramagnets down to 1
K. It was supposed that spins on C
70

dimers form spin single
ts, so that magnetic signal
of TDAE
-
C
70

could only originate from TDAE
+

cation radicals. That would suggest that
also in other TDAE
-
higher fullerenes complexes TDAE
+

spins are not silent.

Tanaka et al.
29

succeeded to synthesise molecular alloys TDAE
-
(C
60
)
1
-
x
(C
70
)
x

in
a broad
x

range from 0.1 to 0.9. The low temperature magnetic properties of the alloys
were monotonically changing from a ferromagnetic for TDAE
-
C
60

to a paramagnetic for
TDAE
-
C
70
. Interestingly, the Curie temperature,
T
c

also linearly decrease
d as the
content of C
70

increased. This is consistent with the mean field result for
T
c
: T
c

=
2JzS(S+1)/3k
B
, where
z

is the effective number of nearest neighbours. This result is
important because if the magnetic interactions in TDAE
-
C
60

would be only alon
g chains
in the
c

direction, any amount of impurities (C
70

substitutions) would brake the
ferromagnetic order.


4 Reduction of C
60

with different organic or organometallic donors

Although C
60

is a weak electron acceptor it combines with many organic or
org
anometallic donors to form charge
-
transfer salts. Here we will mention only those,
which are relevant for magnetism of fullerene
-
based charge
-
transfer compounds.

Klos et al.
30

were stimulated by the discovery of ferromagnetism in TDAE
-
C
60

to try reduction

of C
60

with other amines similar to tertiary amine TDAE. For that
purpose they synthesised tertiary amines diazobicyclononene (DBN) and
diazobicycloundecene (DBU). In contrast to TDAE, where its eight methyl groups
sterically hinder direct reaction with C
60
, DBN and DBU are not so well protected.
DBN
+

and DBU
+

reacted with C
60
-

forming covalent bonds and only a few percent of
nominal spin survived. Nevertheless, those residual spins showed in the case of DBU
-
C
60

a short range magnetic order, which evolves
below 70 K.

In 1994 Wang and Zhou
31

published results of magnetic measurements on the
charge transfer complex [1, 1’, 3, 3’
-
tetramethyl
-

2,2’
-
bi(imidazolidine)]
+
-
C
60
-

(TMBI
-
C
60
). The complex showed ferromagnetic behavior up to 140 K with large hysteresis
l
oop (coercive field 1000 Oe!). However, very soon Schilder et al.
32

showed that it was
false. In fact, TMBI even hardly made a CT complex with C
60

and resulting product was
mainly diamagnetic (the
ESR

signal came only from impurities).

Otsuka et al.
33

use
d a variety of electron donors including aromatic amines,
phenothiazines, phenazines, tetratianofulvalene derivatives and metallocenes to form
charge
-
transfer complexes with C
60
. Among them, only CT complexes with
metalocenes: decamethyferrocene (Cp
2
*
Fe),
cobaltocene (Cp
2
Co) and nicklocene
(Cp
2
Ni), showed ferromagnetic characteristics. All three complexes exhibited an
S
-
shaped
M(H)

curve even at room temperature with a narrow hysteresis. Similarly, the
charge transfer complex with 1,1’
-
biferrocene
34

showed
signs of ferromagnetism at 20
K. Unfortunately, synthesis and magnetic properties of charge
-
transfer complexes of C
60

with the metallocenes were not reproducible.


5 Charge transfer complexes of C
60

derivatives

Another approach in the synthesis of novel fu
llerene
-
based molecular magnets
was based on functionalizing of C
60

by covalently attaching different adducts to the
fullerene cage. Subsequently, such derivatives of C
60

would be combined with organic
or organometallic donors to form charge
-
transfer compl
exes. The idea was to slightly
alter the fullerene electronic properties, e.g. electronic affinity, as well as to hinder C
60
s
rotational degrees of freedom.

In 1994 Venturini et al.
35

presented the first successful synthesis of doped
fullerene derivative
ferromagnet. It was dinitro
-
spiromethanofullerene (C
61
”No
2
”)
doped with bicyclopentadiene cobalt (Cp
2
Co or cobaltocene). It showed paramagnetic
to ferromagnetic transition at 8 K. Transition temperature was lower then in TDAE
-
C
60

but it was encouraging sig
n for a further exploration. The second breakthrough
happened in 1998 when Mrzel et al.
36

reported ferromagnetic transition in a cobaltocene
doped C
60

derivative at 19 K
-

significantly higher then in TDAE
-
C
60
. The derivative
was 1
-
(3
-
aminophenyl)
-
1
H
-
metha
nofullerene[C
60
]. The important feature in both
compounds was that cobaltocene in its oxidised state Cp
2
Co
+

has spin
S

= 0 and by no
means could contribute to the magnetic signal. Only spins
S

= ½ on fullerene moieties
contributed to the magnetic ordering.

That was essential discovery having in mind that
the role of TDAE
+

spin
S

= ½ in TDAE
-
C
60

ferromagnetic ordering was unknown.





Figure 11.

Fullerene derivative 1
-
(3
-
aminophenyl)
-
1
H
-
methanofullerene[C
60
] (left)
and cobaltocene Cp
2
Co (right).



For cob
altocene
-
doped fullerene derivatives it was found that the temperature at
which the doping was performed plays a crucial role in determining the low
-
temperature
magnetic properties of these materials. A detailed study
37

revealed the optimum
conditions, par
ticularly the temperature for the synthesis of ferromagnetic material. The
magnetisation of samples differed markedly both in a magnitude and in a critical
temperature
T
c
. The magnetisation was highest when the synthesis was performed in the
vicinity of 45

C and fell off rapidly on either side of that temperature. The critical
temperatures ranged from 13 K to 17 K. The low
-
temperature magnetisation in a weak
external field (the spontaneous magnetisation) could vary approximately by a factor of
three among
different samples. The samples in ref. 36 showed also a hysteretic behavior
in their magnetic curves below
T
c

with a coercive field
H
c

~ 100 Oe and a remanent
magnetisation
M
r

which is about 0.1 percent of the expected saturation magnetisation
M
s
. The magn
etisation did not show saturation in fields up to 1 kOe.

6 Intercalation of magnetic ions

Because various atom and molecules can be intercalated into C
60

crystal, it was
expected that new magnetic C
60

compounds could be synthesised by intercalation in
whic
h magnetic moments would be carried by intercalants. For that purpose Eu was an
obvious choice
38
. Europium has a magnetic moment 7

B

in the divalent state, while it is
nonmagnetic in the trivalent state. In a fulleride Eu
6
C
60

europium ions are in the
dival
ent state and they order ferromagnetically below 12 K
39
. Substitution of Eu with
nonmagnetic Sr ions in Eu
6
-
x
Sr
x
C
60
, (
x

= 1

5) had little effect on the transition
temperature,
T
c
. In addition, Eu
6
C
60

showed a huge negative magnotoresistance at and
below
T
c
. Evidently, there exist a strong interaction between conduction carriers and
localized magnetic moments; namely, the strong

-
f
interaction in Eu
6
C
60
. This fact
indicates that the ferromagnetism in Eu
6
C
60

comes from the indirect exchange
interaction via
C
60

molecules, which is quite in contrast with the case of magnetic
semiconductor EuO.

Second fullerid with rare earth intercalated ions, which showed a magnetic
behavior was Ce
x
C
60
40
. Cerium has outermost electronic configuration 4f
1
5d
1
6s
2
. In the
case of

Ce
x
C
60

cerium ion is in a trivalent state Ce
3+

with unpaired 4f
1

electron.
Magnetic properties of Ce
x
C
60

were rather controversial. When cooled in a zero
magnetic field the system showed transition to the superconducting state below 13.5 K.
In contrary, c
ooling even in a very low magnetic field (2 Oe) destroyed
superconductivity and the system exhibited a ferromagnetic transition at 15 K. The
ferromagnetic state was also characterised by a hysteresis loop in a
M(H)

curve.
Although proximity of two differen
t ground states in Ce
x
C
60

hasn’t been explained yet,
we could roughly ascribe superconductivity to the doped C
60

subsystem and
ferromagnetism to superexchange interactions between Ce
+

ions.




References


1

Allemand, P. M.,Khemani, K. C., Koch, A., Wudl, F., Holczer, K., Donovan, S., Gruner, G., and
Thompson, J. D., Science 253, 301
-
303 (1991).

2

Stephens, P. W., Cox, D., Lauher, J. W., Mihaly, L., Wiley, J. B., Allemand, P.
-
M., Hirs
ch, A., Holczer,
K., Li, Q., Thompson, J. D., and Wudl, F., Nature 355, 331
-
332 (1992).

3

Tanaka, K., Zakhidov, A. A., Yoshizawa, K., Okahara, K., Yamabe, T., Yakushi, K., Kikuchi, K.,
Suzuki, S., Ikemoto, I, and Achiba, Y., Phys Rev. B 47, 7554
-
7559 (1993
).

4

Suzuki, A.,Suzuki, T., Whitehead, R. J., and Maruyama, Y., Chem. Phys. Lett. 223, 517
-
520 (1994).

5

Dunsch, L., Eckert, D., Frohner, J., Bartl, A., and Muller, K.
-
H., J. Appl. Phys. 81, 4611
-
4613 (1997).

6

Lappas, A., Prassides, K., Vavekis, K., Arcon
, D., Blinc, R., Cevc, P., Amato, A., Feyerherm, R.,
Gygax, F. N., and Schenck, A., Science 267, 1799
-
1802 (1995).

7

Venturini, P., Mihailovic, D., Blinc, R., Cevc, P., Dolinsek, J., Abramic, D., Zalar, B., Oshio, H.,
Allemand, P.
-
M., Hirsch, A., and Wudl,

F., Int. J. Mod. Phys. B 6, 3947
-
3951 (1992).

8

Blinc, R., Cevc, P., Arcon, D., Mihailovic, D., and Venturini, P., Phys. Rev. B 50, 1
-
3 (1994).

9

Mihailovic, D., Arcon, D., Venturini, P., Blinc, R., Omerzu, A., and Cevc, P., Science 268, 400
-
402
(1995).

1
0

Tanaka, K., Asai, Y., Sato, T., Kuga, T., Yamabe, T., and Tokumoto, M., Chem. Phys. Lett. 259, 574
-
578 (1996).

11

Sato, T., Saito, T., Yamabe, T., and Tanaka, K., Phys. Rev. B 55, 11052
-
11055 (1997).

12

Tanaka, K., Sato, T., Yoshizawa, K., Okahara, K., Y
amabe, T., and Tokumoto, M., Chem. Phys. Lett.
237, 123
-
126 (1995).






13

Suzuki, A., Suzuki, T., and Maruyama, Y., Solid State Commun. 96, 253
-
257 (1995).

14

Schilder, A., Klos, H., Rystau, I., Schultz, W., and Gotschy, B., Phys. Rev. Lett. 73, 1299
-
1302 (19
94).

15

Bommeli, F., Degiorgi, L., Wachter, P., Mihailovic, D., Hassanien, A., Venturini, P., Schreiber, M.,
and Diedrich, F., Phys. Rev B 51, 1366
-
1369 (1995).

16

Omerzu, A.,Mihailovic, D., Tomic, S., Milat, O., Biskup, N., Phys. Rev. Lett. 77, 2045
-
2048
(1996).

17

Arcon, D., Dolinsek, J., Blinc, R., Phys. Rev. B 53, 9137
-
9143 (1996).

18

Mrzel, A., Cevc, P., Omerzu, A., and Mihailovic, D., Phys. Rev. B 53, R2922
-
R2925 (1996).

19

Arcon, D., Cevc, P., Omerzu, A., and Blic, R., Phys. Rev. Lett. 80, 1529
-
1532 (1998).

20

Omerzu, A., Tokumoto, M., Tadic, B., and Mihailovic, D., Phys. Rev. Lett. 87, 177205
-
1
-
177205
-
4
(2001).

21

Omerzu, A., Mijatovic, D., and Mihailovic, D., Synth. Metal
s 121, 1155
-
1156 (2001).

22

Omerzu, A., Mihailovic, D., and Tokumoto, M., Phys. Rev. B 61, R11883
-
R11885 (2000).

23

Narymbetov, B., Omerzu, A., Kabanov, V. V., Tokumoto, M., Kobayashi, H., and Mihailovic, D.,
Nature 407, 883
-
885 (2000).

24

Kambe, T., Nogami, Y., and Oshima, K., Phys. Rev. B 61, R862
-
R865 (2000).

25

Mizoguch, K., Machino, M., Sakamoto, H., Kawamoto, T., Tokumoto, M., Omerzu, A., and
Mihailovic, D., Phys. Rev. B 63, 140417
-
1
-
140417
-
4 (2001).

26

Garaj, S., Kambe, T., Forro, L.
, Sienkiewicz, A., Fujiwara, M., and, Oshima, K., Phys. Rev. B 68,
144430
-
1
-
144430
-
7 (2003).

27

Tanaka, K., Zakhidov, A. A., Yoshizawa, K., Okahara, K., Yamabe, T. Kikuchi, K., Suzuki, S.,
Ikemoto, I, and Achiba, Y., Solid State Comunn. 85, 69
-
72 (1993).

2
8

Oshima, K., Kambe, T., Fujiwara, M., and Nogami, Y., Synth. Met. 133
-
134, 699
-
701 (2003).

29

Tanaka, K., Sato, T., Yamabe, T., Yoshizawa, K., Okahara, K., and Zakhidov, A. A., Phys. Rev. B 51,
990
-
995 (1994).

30

Klos, H, Rystau, I, Schutz, W., Gotschy, B
., Skiebe, A., and Hirsch, A., Chem. Phys. Lett. 224, 333
-
337 (1994).

31

Wang, H, and Zhu, D., J. Phys. Chem. Solids 55, 437
-
439 (1994).

32

Schilder, A., Gotschy, B., Seidl, A., and Gompper, R., Chem. Phys. 193, 321
-
326 (1995).

33

Otsuka, A., Teramoto, T.,

Sugita, Y., Ban, T., and Saito, G., Synth. Met. 70, 1423
-
1424 (1995).

34

Wang, W.
-
J., Lay, Y. L., Chang, C.
-
S., Chiu, H.
-
S., Chuang, K.
-
S., and Wang, B.
-
C., Synth. Met. 86,
2293
-
2294 (1997).

35

Venturini, P., Krasevec, V., Mihailovic, D., Eiermann, M., Sr
danov, G., Sariciftci, N. S., Li, C., and
Wudl, F., Proceedings of International winterschool on Electronic Properties of Novel Materials,
Singapore: World Scientific, 1994, pp. 514
-
517.

36

Mrzel, A., Omerzu, A., Umek, P., Mihailovic, D., Jaglicic, Z, Tron
telj, Z., Chem. Phys. Lett. 298, 329
-
334 (1998).

37

Umek, P., omerzu, A., Mihailovic, D., and Tokumoto, M., Chem. Phys. 253, 361
-
366 (2000).

38

Ksari
-
Habiles, Y., Claves, D., Chouteau,G., Touzain, P., Jrandey, C., Oddou, J. L., and Stepanov, A., J.
Phys. C
hem. Solids 58, 1771
-
1778 (1997).

39

Ishii, K., Fujiwara, A., Suematsu, H., and Kubozono, Y., Phys. Rev. B 65, 134431
-
1
-
134431
-
6 (2002).

40

Maruyama, Y., Motohashi, S., Sakai, N., Watanabe, suzuki, K., Ogata, H., and Kubozono, Y., Solid
State Commun. 123,
229
-
233 92002).