Top Curr Chem (2004) 235:221–249

DOI 10.1007/b95429

Springer-Verlag 2004

Towards Spin Crossover Applications

Jean-Franois Ltard1 (*) · Philippe Guionneau1 · Laurence Goux-Capes2
1
Groupe des Sciences Molculaires,
Institut de Chimie de la Matire Condense de Bordeaux, UPR CNRS No 9048,
Universit Bordeaux 1, 87 Av. du Doc. A. Schweitzer, 33608 Pessac, France
[email protected]
2
MOTOROLA Centre de Recherche, Espace Technologique Saint Aubin,
91193 Gif/Yvette, France

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
2 Towards the Design of Molecular Switches, a Switchtronic Challenge . . . . 223
2.1 The General Context of Molecular Electronics . . . . . . . . . . . . . . . . . . 223
2.2 Molecular Switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
3 Highlights of Data Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
3.1 Current Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
3.2 Memory Property of SCO Materials . . . . . . . . . . . . . . . . . . . . . . . 230
3.3 Design of SCO Cooperative Cluster—A Challenge . . . . . . . . . . . . . . . 232
4 Devices for Data Displays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4.1 State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4.2 Towards the Design of Data Display . . . . . . . . . . . . . . . . . . . . . . . 236
4.3 Fabrication of Displays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
4.4 Specific Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
5 Recent Progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.1 Molecular Switches and Non-linear Optics. . . . . . . . . . . . . . . . . . . . 240
5.2 Optical Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

Abstract In this chapter we attempt to review the potential for the application of the spin
crossover (SCO) phenomenon in various domains, such as molecular electronics, data
storage, display devices. It is evident that SCO properties, such as room-temperature
working range, chemical stability, low addressing power, short addressing time, full re-
versibility, are of promising value in the context of the stringent limits necessary in the
future development of information technology, due to the unceasing miniaturization of
the components. Of course, many requirements must be fulfilled before any use in a gen-
uine device becomes feasible. Some of these are emphasized and discussed here. Addi-
tionally, this review reports recent progress in non-linear optics and photomagnetism of
SCO materials.

Keywords Spin crossover · Molecular bistability · Molecular devices · Nanotechnology ·

Molecular electronics · Switchtronic
222 J.-F. Ltard et al.

List of Abbreviations and Symbols

AFM Atomic force microscope
b First molecular hyperpolarizability tensor
bpp 2,6-Bis(pyrazol-3-yl)pyridine
btr 4,40 -Bitriazole
CD Compact discs
CRT Cathode-ray tube
Htrz 1,2,4-Triazole
Hyettrz 4-(20 -Hydroxyethyl)-1,2,4-triazole
ILCT Intra-ligand charge transfer
LB Langmuir-Blodgett
LCD Liquid crystal display
MLCT Metal-to-ligand charge transfer
MO Magneto-optics
NH2trz 4-Amino-1,2,4-triazole
NLO Non-linear optics
PM-Bia 20 -Pyridylmethylene 4-aminobiphenyl
PM-PEA N-20 -Pyridylmethylene-4-phenyl-ethynylaniline
PM-2PEA N-20 -Pyridylmethylene-4-phenyl-ethynylphenylethynylaniline
SONLO Second order non-linear optics
SPE Super paramagnetic effect
trz 1,2,4-Triazolate(1-) ion
tvp Bis-1,2-di(4-pyridyl)ethylene

1
Introduction
In the last 30 years, the interest in “molecular materials”, possessing a prop-
erty or a set of properties relevant for practical applications, has consider-
ably increased. This growth mainly comes from the result of the develop-
ment of communication devices and the necessity of component size reduc-
tion [1–4]. The molecular scale is increasingly being considered as the an-
swer to the miniaturization of the components used in the construction of
working devices [5–14]. The current large-downward (top-down) approach
based on silicon technology forces solid-state physicists and electronic engi-
neers to manipulate progressively smaller and smaller pieces of matter [15].
In regard to this, molecule-based devices offer prospects of sensitivity and
selectivity levels that cannot be reached with conventional solid-state mate-
rials. It is necessary, however, to reproduce traditional electronic functions,
as memories, modulators, rectifiers, transistors, switches and wires [5–14].
Today, many organic and coordination derivatives are being investigated. In
this context, coordination complexes are especially promising since a judi-
cious choice of metal ions and ligands allows one to modulate the final phys-
ical properties of the complexes, and even in some cases, to combine in a
synergistic way different physical properties.
Towards Spin Crossover Applications 223

In the present review, we focus our attention on the spin crossover (SCO)
property encountered in some coordination complexes having a dn (n=4–7)
electronic configuration. In some specific cases, iron(II) molecular materials
may exhibit a change of the spin state under external perturbations (change
of temperature, application of pressure, by light irradiation or by pulsed
magnetic field [16, 17]) combined with (i) room-temperature working
range, (ii) stability over successive cycles, (iii) low addressing power (about
5 mW/cm2), (iv) short addressing time since the transition is purely elec-
tronic in nature (nanosecond scale), and (v) a hysteresis effect when the
transition between the two states is accompanied by cooperative effects [18–
20]. Many applications can then be envisaged. Here we attempt to illustrate
the potential for application of the spin crossover (SCO) phenomenon in
molecular electronics, data storage, display devices, non-linear optics, and
photomagnetism. This treatment is certainly not exhaustive and the writers
apologize for any omissions.
This work is organized in four sections. The first introduces the general
context of molecular electronics and gives the main requirements that a mo-
lecular compound must fulfill to act as an active molecular switch. The SCO
suitability is then analyzed. The second section reports some highlights con-
cerning the current technology used to store data and the presence of a ther-
mal hysteresis loop in SCO materials. The particular challenge in the design
of cooperative clusters is emphasized. The third section deals with the fabri-
cation of display devices. The state of the art is first recalled, then the vari-
ous criteria that a molecular material has to fulfill before any use in genuine
display devices is possible, are detailed. Finally, examples of data displays
based on SCO materials are described. -The last section of this chapter fo-
cuses on the recent improvements in the design of switchable non-linear op-
tical materials as well as the possibility to address spin states through light
irradiation.

2
Towards the Design of Molecular Switches, a Switchtronic Challenge
2.1
The General Context of Molecular Electronics

For many years, the developments of telecommunication and internet net-

works have been concerned with the miniaturization of electronic devices.
As an example, Intel co-founder Gordon Moore observed in 1965 that the
number of transistors per cm2 of silicon doubled every year and predicted
that it would continue to do so for the next ten years [1]. In fact, this predic-
tion has been remarkably realized and has gained the status of a “law”. It is
now accepted that the number of transistors on a chip doubles approximate-
224 J.-F. Ltard et al.

ly every 18 months [2]. However, there are some alarming signals starting to
appear [15]. It seems that this size reduction, and thus this speed increase,
can only continue for a finite time. The fabrication of small electronic de-
vices becomes more and more difficult, and expensive, as the nanometer
scale is approached. The current technology is based on the large-downward
(top-down) approach, which means the fabrication and the manipulation of
progressively smaller and smaller pieces of silicon matter.
Nowadays, the individual features found in commercially available per-
sonal computers are less than 1 mm across and compare with the size of a
virus. There thus exist several significant challenges in the future miniatur-
ization of such components. For instance, one important limitation in the
fabrication process of a complete integrated circuit (including wires, resis-
tors, capacitors and transistors) lies in the photolithographic steps [15, 21].
The wafer with a layer of photoresist is exposed through a patterned mask.
The resolution obtained by using 248 nm irradiation is currently around
0.18 to 0.25 mm. Future plans are in mind for using 193 nm (ArF Excimer
Laser) and 157 nm (KrCl Excimer Laser) wavelengths in order to reach a
size of 0.13 and 0.1 mm [15]. Even X-ray lithography (l<1 nm) or extreme
UV lithography (l
13–4.5 nm) are being envisaged. The development of re-
liable techniques to produce conventional silicon-based devices with dimen-
sions lower than 100 nm represents an exciting technological challenge [15].
In this particular context, the molecular or the atomic level is, more and
more, considered as the size of the ultimate integrated circuit. This repre-
sents the small-upward (bottom-up) approach. Molecular-sized systems of-
fer an exciting opportunity to use ~1013 logic gates/cm2 in comparison with
the 108 gates/cm2 of present microchips. In addition, the response times of
molecular devices can be in the range of femtoseconds while the fastest pres-
ent devices operate in the nanosecond regime. The goal is then to reproduce
conventional microelectronic functions by using a single molecule, or a
small group of molecules, carbon nanotubes, or nanoscale metallic or semi-
conductor wires. There are now some remarkable examples of molecular
functions, as memories, modulators, rectifiers, transistors, switches, and
wires [5, 9, 12, 14, 22–24]. In relatively few years, the “naive” expectation of
the use of a molecule to perform electronic functions has become a wide-
spread research field, named “Molecular Electronics”. Chemists, physicists,
materials scientists, as well as electrical and chemical engineers, are working
intensely on this bottom-up approach. Of course, numerous obstacles re-
main. We shall first examine the requirements that a molecular switch,
which certainly represents the simplest device in the future of molecular-
based electronics, must fulfill.
Towards Spin Crossover Applications 225

2.2
Molecular Switches

The characteristic that defines a molecular switch is the bistability domain

with an “ON” and an “OFF” position. In its “ON” position, the molecular
switch must either perform some function or allow another device to per-
form a function while in its “OFF” position, it must remain inactive. The
two states can be isomers, an acid and its conjugate base, the oxidized and
reduced forms of a redox active molecule, or even the complexed and un-
complexed forms of a receptor [10, 25–32]. Often, these transformations are
reversible. The chemical system returns to the original state when the input
signal is turned off. In some instances, the molecular switch produces a
chemical, electrical, and/or optical output that varies in intensity with the
switching process. For example, a change in pH, redox potential, absor-
bance, or emission intensity can accompany the interconversion between the
two states of a molecular switch [10, 25–32].
In such a context, the molecular bistability associated with the SCO phe-
nomenon is very relevant. The change of the spin state in a SCO material, as
illustrated in Fig. 1, can be induced reversibly in a solid matrix under exter-
nal inputs (temperature, pressure, light irradiation, pulsed magnetic field)
[16, 17]. In response to these stimulations, the molecular SCO switches pro-
duce outputs that may be, for example, change in absorbance or in magnetic
response. It then becomes possible to associate a piece of information with
each of the LS and HS states. When the input is applied, the output changes
from a low to a high value or from a high to a low value, depending on the
logic convention positive or negative, respectively [12]. In others words, the
color and/or the (dia)paramagnetic changes of SCO complexes can be re-
garded in the molecular electronics context as a binary coding (0-1). This is
of particular importance since data processing and communication require
the encoding of all the information in the form of binary digits. In fact, the
SCO phenomenon allows the definition of a NOT gate, which is one of the
three fundamental logic operations. Let us recall that a NOT operator only
involves one input and one output while AND and OR operators combine
two inputs and one output [12].
Of course the input and the output do not need to share the same trans-
duction mechanisms. For instance, pyrazoline derivatives execute NOT oper-
ations by transducing a chemical input into an optical output [33, 34]. How-
ever, the increasing need for telecommunication and internet applications
demands the continuous enhancement of the transmission capacity and
speed of communication networks [35]. The present technology is principal-
ly based on optical signals. It becomes then crucial to identify molecular
switches able to provide an optical input/output transduction mechanism.
SCO materials, in this particular context, are promising since optical inputs
226 J.-F. Ltard et al.

Fig. 1 Molecular switch based on the Iron(II) Spin Crossover (SCO) phenomenon. The
Fe2+ ion has six valence electrons occupying the 3d orbitals. The input stimulation, i.e.,
temperature (T), light (hu), pressure (P) and magnetic field (B), induces the transforma-
tion of one state into the other and determines the intensity of the output signals, i.e.,
absorption, magnetic and volume changes

can be converted into optical outputs [16]. The last section of this review re-
ports the recent progress in the domain of photomagnetism.
Let us mention, here, that in regard to traditional photochromic switches
which also convert optical inputs into optical outputs, one of the main ad-
vantages of the SCO phenomenon is certainly the absence of any kind of fa-
tigue. At the molecular scale, the ST in an iron(II) complex corresponds to
an intra-ionic transfer of two electrons between the t2g to the eg orbitals,
centered in the immediate environment of the iron metal center (Fig. 1).
There is no through-space electron displacement. The transition can be re-
produced as many times as needed without altering its characteristics,
which is obviously crucial in a genuine device. It thus represents a major ad-
vantage but we are certainly only at the beginning of the adventure. That a
SCO molecule makes logic operations is not sufficient to make a molecular
electronic device. Today, hundreds, if not thousands, of chemical systems
are already known to execute logic operations [12]. A molecular switch must
fulfill many other requirements. For instance, it must be chemically stable
once deposited on a surface or embedded into a matrix. Moreover, any
Towards Spin Crossover Applications 227

movement or reorganization of the active components must be impeded to

ensure correct addressing and reading actions. The self-assembly of a mo-
lecular switch (on a surface or in a volume) certainly constitutes one of the
more sensitive points. In a molecule-based device, each molecular switch
must maintain a separate identity from the other switches around it. Finally,
other problems may arise from processing complexity, power-supply re-
quirements, rate of writing data, or cost.
Undoubtedly, the advances in SCO research are still far from meeting this
list of requirements. The best example of molecular SCO switches organized
on a surface is certainly given by the work of Mingotaud and Delhas [36,
37]. These authors have succeeded in preparing Langmuir-Blodgett films
with well ordered SCO complexes. The change of the spin state by light irra-
diation has been demonstrated [38]. This opens up exciting perspectives for
the future.

3
Highlights of Data Storage
3.1
Current Technology

Today, magnetic and optical disk storage is taking over the primary role of
providing storage, retrieval and world-wide distribution of information and
knowledge. We report here a brief overview of the common media storage
methods [39]. Hard drives store files as magnetically encoded areas on disks
made of metal or spin glass material. The write heads align fields of magnet-
ic particles on the surface and the read head registers a one or zero informa-
tion depending on the polarity of the area. In Compact Discs (CD) the infor-
mation is encoded by using optics. A typical optical disk is made of three
layers: a polycarbonate base through which light can pass, a layer of alumi-
num, and a protective layer of acrylic on the top. The aluminum presents
high area (pits) and flat section (land). The pits are typically 0.5 m wide.
The data read phase is performed using a laser beam. The latter passes
through the polycarbonate layer, reflects off the aluminum layer and hits an
opto-electronic device that detects changes in light intensity. The “pits” re-
flect light differently than the “lands”, and the opto-electronic sensor detects
this change in reflectivity. In a traditional CD, the write phase can be done
just once while a re-writable CD enables the storage and erasure of data. In-
stead of standard pits, the latter technology uses a phase change alloy whose
reflectivity can be modified. Using a relatively high power laser, the phase
change allows the alloy to have different refractivity indices, creating reflec-
tive and non-reflective “pits” and “lands”.
228 J.-F. Ltard et al.

In parallel with these two common magnetic and optic technologies, the
magneto-optical (MO) storage combines the two processes. MO drives write
magnetically (with thermal assist) and read optically. To record data in a
magneto-optical device, a laser heats up the medium, decreasing the medi-
um coercivity to a level that allows its magnetism to be modified by a rela-
tively weak magnetic field. Once the data are written, the area is then cooled,
and the magnetic data are not subject to modification or erasure by magnet-
ic forces encountered during the daily routine. To read the data from the
MO medium, the drive does not use magnetic current induced in the heads
by the changing magnetic flux on the disk, as a hard drive does, but instead
takes advantage of the Kerr effect. Because of the Kerr effect, the polarization
of light will change when it is shone upon an area of the disk that is magne-
tized. Shining a beam of polarized light on a magnetized surface causes the
polarization of the light to shift slightly (usually less than 0.5). If the mag-
netization is reversed, then the angle of change in polarization (the Kerr an-
gle) is changed as well. The change in direction of magnetization can be de-
tected as “1” and “0”, making MO convenient for data storage.
In the future, however, both the magnetic and conventional optical data
storage technologies, where individual bits are stored as distinct magnetic
or optical changes on the surface of a recording medium, will be approach-
ing physical limits beyond which individual bits may be too small or too dif-
ficult to store. For instance, magnetically storing of data faces a limit im-
posed by the Super Paramagnetic Effect, or SPE [40]. SPE is a physical phe-
nomenon that occurs in data storage when the energy that holds the mag-
netic spin of atoms making up a bit becomes comparable to the ambient
thermal energy. When that happens, bits become subject to random “flip-
ping” between 0s and 1s, destroying the information which they represent.
Currently extensive efforts are being made to better understand this phe-
nomenon. The energy barrier depends on the stored bit pattern and on mag-
netic interactions between adjacent particles. To now, the SPE is predicted to
happen at about 6 Gbit/cm2 but it is reasonable to expect that hard disk
magnetic recording will push the limit into the 15–30 Gbit/cm2 range [40].
Concerning the optical process, the limit is the track width and pit size
linked to the diffraction-limited optics. If more tracks are to be added, they
each must be smaller, so the ability of the recording head to accurately find
the tracks, and increase the accuracy of the position sensor and stepper mo-
tor must be improved. Moreover, the pits have to be smaller. This imposes
the use of a laser with higher frequency (shorter wavelength) to read and
write on the CD. The nanophotonics is a new frontier that describes nanos-
cale optical science, much smaller than the wavelength of optical radiation
and technology. Some examples have been already reported where nanoscale
optical storage is carried out on magneto-optical and phase-change media
with data densities up to 26 Gbit/cm2 [41]. Consequently, it is impossible to
predict the ultimate limits for magnetic and optical storage. In [42–43] are
Towards Spin Crossover Applications 229

examples of these, which predicted maximum densities of 0.3 Mbit/cm2,

1 Mbit/cm2, and 20 Mbit/cm2. Todays best disk drives operate at nearly a
hundred times the latter limit [40].
In fact, many other approaches, using alternative materials and tech-
niques, are currently being explored by industrial and other research labora-
tories. Concerning magnetic storage, one solution is to make single grain
magnetic bits [40]. This eliminates the statistical noise associated with gran-
ular media and increases the domain size by a factor of several hundred.
This can be done by patterning a thin magnetic film into small sized islands,
e.g., with stamping or ion beam patterning. Holographic storage is also ex-
tensively studied since it provides dramatic advances in both data storage
density and transfer rates [44]. Unlike other storage methods, which record
only on the surface of a disk, holographic digital data storage allows record-
ing through the entire thickness of the material, which gives a huge increase
in storage density. Scanning probe storage is also being extensively studied
[45]. The principle is based on writing with an AFM tip on a plastic sub-
strate. Marks smaller than 0.1 micrometer can be written in this manner.
During the read process, the AFM tip scans the topography of the media
upon contacting a data pit, a physical change or a deflection occurs in the
tip and this change is sensed. The AFM process suffers, however, from the
long time necessary to write the data. In this context, it was recently demon-
strated that data-storage can be thermomechanically obtained in soft poly-
mers [45]. This was accomplished by using a modified cantilever in which
the tip was heated by applying a voltage across the legs of the cantilever. The
contact of the heated cantilever with the polymer layer, constituted of 40 nm
of polymethylmethacrylate, melts the surface and forms a pit. Each pit rep-
resents a bit of information. Multiple writing and erasing cycles have been
carried out with no degradation of the substrate. Furthermore, the use of an
array of 3232 thermomechanical tips working in parallel fashion has al-
lowed the writing of data in bit densities as high as 20–30 Gbit/cm2. A proto-
type device called “Millipede” is now close to production [46].
In conclusion to this brief overview, the use of nanotechnology appears
more and more relevant for the future development of memory devices.
There exists today a large number of ways to realize data storage in a molec-
ular system, as charge storage, conformational or positional shifting and
physical transformation [14]. The molecular memory device can be a single
molecule or even a limited number of molecules. In this latter case, we are
faced with a cluster where each active center is chemically bound or orga-
nized in a strong interacting network. Each individual center acts in a syner-
gistic way and confers on the cluster a memory property [20, 47–49]. In this
sense, a significant breakthrough has been recently obtained by Hewlett
Packard in collaboration with UCLA chemists [50, 51]. They report the
achievement of the highest density electronically addressable memory re-
ported to date. The bit density of the device is more than ten times greater
230 J.-F. Ltard et al.

than todays silicon memory chips. The electronically switchable element

seems to be an array of bistable molecules deposited on a single monolayer
between metallic electrodes in a crossbar architecture. The switch belongs to
the class of molecules called catenanes and rotaxanes [50, 51].

3.2
Memory Property of SCO Materials

In contrast to the dynamic nature of a molecular switch, a memory device

represents a static component [14]. The ON state must be switched to the
OFF state by applying an external stimulation, as a molecular switch, but
the stored information must remain. If we consider the SCO phenomenon of
an isolated complex involving only one metal center this condition is not
reached. The course of the spin transition follows a Boltzmann approach
with a temperature-dependent distribution over all HS vibronic states and
the LS state. Consequently, the SCO phenomenon is gradual (Fig. 2). The in-
put information is erased as soon as the external perturbation is stopped.
Such a molecular system exhibits no static memory property.
To obtain a memory feature in a SCO material, it is now well established
that the ST phenomenon must be accompanied by a hysteresis effect [47].

Fig. 2 Schematic drawing of a gradual spin conversion recorded for a molecular switch
involving an isolated iron(II) metal center and of a thermal hysteresis loop obtained for a
strongly interlocked polynuclear iron(II) cluster. The structure of the polymeric 1,2,4-tri-
azole iron(II) compounds is deduced from the EXAFS investigation [92]
Towards Spin Crossover Applications 231

Many general articles have been devoted to this particular behavior [16, 52–
54]. Thermal hysteresis in a SCO complex is the result of interactions be-
tween active sites within the lattice (Fig. 2) if there is no crystallographic
phase change. In iron(II) SCO materials, the spin conversion is concomitant
with a variation of the iron-to-ligand average bond distance, which is about
0.2 greater in the HS state [55]. When intermolecular interactions in the
lattice are sufficiently strong, the change in molecular size can be transmit-
ted from one site to another and this induces a cooperative effect. The extent
of the propagation defines the size of the domain.
One of the first systems in which it was shown that a thermal hysteresis
loop was associated with SCO was [Fe(4,7-(CH3)2phen)2(NCS)2] complex,
with T1/2"=121.7 K and T1/2#=118.6 K [56]. Since this report, many other
SCO complexes have been described which display thermal hysteresis. Usu-
ally, in a mononuclear complex intermolecular interactions are weak and the
width of the hysteresis loop remains below 20 K. Some rare examples, how-
ever, exhibit hysteresis widths in the 30–40 K range [57, 58]. All these exam-
ples have in common the presence of intermolecular p-p stacking of the aro-
matic ligands to generate a supramolecular organization in the crystal lat-
tice. For instance, in the particular case of the [Fe(PM-PEA)2(NCS)2] com-
pound, the crystal packing shows very short contacts between phenyl rings
belonging to adjacent molecules [57, 59]. A thermal hysteresis width of 37 K,
with T1/2"=231 K and T1/2#=194 K, has been described. By replacing the two
NCS groups bound to the iron(II) ion by the NCSe group, the thermal spin
transition was shifted (T1/2"=256 K and T1/2#=306 K) and the compound is
then bistable at 293 K [57].
Cooperativity may also be increased by covalent links between the active
sites. The first experimental result supporting this idea was provided by the
compound [Fe(btr)2(NCS)2]·H2O with btr=4,40 -bis-1,2,4-triazole [60]. In this
compound, each iron atom is surrounded by two NCS groups in trans posi-
tion and four btr ligands coordinated through nitrogen at the 1- and 10 -posi-
tions. Non-coordinated water molecules are also hydrogen bonded to the pe-
ripheral nitrogen atoms of the triazole units. The iron(II) atoms are then in-
volved in a two-dimensional network. The spin transition is extremely
abrupt and the hysteresis width is 21 K (T1/2"=144.5 K and T1/2#=123.5 K).
In addition, other two-dimensional (2D) and three-dimensional (3D) net-
works have been reported, such as [Fe(tvp)2(NCS)2] (tvp=bis-1,2-di(4-
pyridyl)ethylene) [61], [Fe(pyridine)2Ni(CN)4] [62], and [Fe(btr)3](ClO4)2
[63] systems. Typically these 3D compounds undergo strong cooperative
spin transitions with 20–40 K width hysteresis loops. Another very interest-
ing family of complexes are those involving the 1,2,4-triazole ligand. The
general formula of this family is [Fe(Rtrz)3]A2·nH2O with A being the coun-
ter-anion and nH2O the number of non-coordinated water molecules. The
structure is a linear chain of metal centers triply bridged by 1,2,4-triazole
ligands (Rtrz) through the nitrogen atoms occupying the 1 and 2 positions
232 J.-F. Ltard et al.

[64]. The first member of this family was reported by Haasnoot et al. in
1977 [65]. Larionov and Kahn have described other members of this family
with wide thermal hysteresis (20–40 K) above room temperature [19, 20, 66,
67].
It is evident that there are numerous examples of SCO materials for which
thermal hysteresis loops are observed. In addition to this, the overall num-
ber of iron(II) SCO systems is very large and will undoubtedly continue to
grow. It is thus reasonable that new systems displaying a high degree of co-
operativity will be found.

3.3
Design of SCO Cooperative Cluster—A Challenge

As already mentioned, the design of a system able to store information on

the single molecule level or assemblies of molecules is particularly impor-
tant in the development of information technology. The molecular bistability
associated with the SCO phenomenon is of particular interest since logic op-
erations can be executed. A piece of information can be associated with each
of the LS and HS spin states. The hysteresis behavior linked to cooperativity
confers a memory effect to the system. In the T1/2"–T1/2# temperature range,
the nature of the observed state depends on the history of the molecular
switch [68]. Moreover, as recently demonstrated, the input information in
the center of a thermal hysteresis loop can be addressed by pulsed light irra-
diation [69], by pulsed magnetic field [17] or by pulsed pressure effect [70].
One of the main challenges for chemists in regard to the design of molec-
ular-based memory devices is the control of the size of the SCO cluster
which exhibits thermal hysteresis. The size available to date is far from the
level of miniaturization required for the realization of nanoscale processors.
Scanning electron microscopy, for instance, of the well-known SCO 1,2,4-tri-
azole iron(II) complexes, reveals particles of around 5–10 mm (Fig. 3) [71].
However, by considering that a cluster consisting of 103 strongly interacting
metal centers supposedly is able to exhibit cooperative behavior [72], this
roughly corresponds, in the case of the 1,2,4-triazole iron(II) family, to a
cube of around 7 nm per edge. Therefore, the design of nanoscale devices
based on a SCO materials is then theoretically possible.
To reach this goal, the use of both the top-down and the bottom-up ap-
proaches can be envisaged. Concerning the former method, it is now well-
known that the preparation of insoluble compounds by using conventional
chemical reactions results in products which are heterogeneous with respect
to particle mass and hence size due to random crystal growth. Production of
fine or ultrafine particles then requires grinding and sieving steps which are
not desirable for the SCO phenomenon. When a mechanical stress is exerted
on coordination compounds in the solid state, a molecular distortion results
and the SCO behavior, e.g., residual HS fraction at low temperature, irre-
Towards Spin Crossover Applications 233

Fig. 3A–E [Fe(NH2trz)3]Br2 particles observed by using a scanning electron microscopy:

A,B after conventional synthetic conditions; C,D prepared in H2O/Lauropal inverted wa-
ter/oil microemulsions; E the thermal hysteresis loop of a collection of well-defined parti-
cles recorded by measurement of optical reflectivity

versibly changes [73, 74]. With regard to the bottom-up method, one of the
approaches is based on the step-by-step synthesis. This method is particu-
larly promising since each step of the synthesis can be precisely controlled.
Examples of dinuclear [75], trinuclear [76] and tetranuclear [77] SCO sys-
tems, with cooperative interactions between the metal centers, have been de-
scribed in the literature. The requirement to increase the number of metal
centers involved in one molecule is not readily achieved. A considerable in-
crease in theoretical and experimental knowledge will be a prerequisite to
this being met. An alternative approach currently being developed to syn-
thesize nanoscopic materials involves the use of water-in-oil (w/o) mi-
croemulsions [78–80]. Dispersed water pools act as microreactors where
chemical reactions can be performed to generate the required product in the
form of a colloidal (nano)dispersion. Along these lines, we have shown that
by confining the synthesis to nanoscale water droplets formed in reverse mi-
croemulsions prepared from the Lauropal 205 surfactant, 1,2,4-triazole iro-
n(II) particles with a uniform shape and size can be routinely prepared [71].
The current limit size of the cluster is about 60–200 nm. Figure 3 shows the
thermal hysteresis loop recorded for a collection of individual clusters. The
234 J.-F. Ltard et al.

loop remains in the vicinity of room temperature. AFM investigation, based

on the volume change accompanying the spin transition, is in progress to
record the hysteresis loop of just one individual cluster. One can expect an
increase of the abruptness of the spin transition coupled to a sharp and
well-defined hysteresis.
Of course, these investigations are in their early stages but the trend is
clear. Today, the step-by-step approach may appear, at first sight, only inter-
esting as basic research but the growth of nanoscience and the subsequent
development of nanotechnology continuously opens news horizons. For in-
stance, the design of a linear trinuclear compound exhibiting SCO proper-
ties can be regarded as an almost linear motor with, in fact, interesting func-
tion as an artificial muscle [81] able, in a microelectronic circuit, to pass the
information from one individual element to another. Artificial molecular
machines and motors appear as one of the emerging fields of chemistry [10,
13]. A molecular-level machine has been defined as an assembly of a distinct
number of molecular components that are designed to perform machine-like
movements (output) as a result of an appropriate external stimulation (in-
put). Several systems are today known and many of them are based on
chemical, photochemical and electrochemical processes. In such a context,
the volume change associated with the SCO phenomenon and the absence of
susceptibility to fatigue is of particular interest.
Another issue that may also have important consequences in the develop-
ment of nanoscience is the design of materials able to amplify at the molecu-
lar scale an output signal. Any enhancement of the final output would re-
duce the difficulty for the reading operation. Along this line, in a coopera-
tive SCO cluster the resulting output of an appropriate external stimulation
is not only an individual response but a summation of individual contribu-
tions which lead to an amplification effect. This effect is not entirely clear
and much effort is currently being expended to evaluate, for instance, the
quantum efficiency of the photoconversion [82].

4
Devices for Data Displays
4.1
State of the Art

In the last 100 years, display technology has advanced impressively [83–85].
Information technology has moved from painstakingly slow generation and
permanent storage on bulk materials, such as stone and paper, to the almost
instantaneous production and transfer through the pervasive electronic and
optical networks. Nevertheless, such evolution is far from complete. The re-
cent progress of the mobile telecommunication networks opens new oppor-
Towards Spin Crossover Applications 235

tunities for terminal manufacturers [84]. Advances in new display technolo-

gies, such as reflective-color liquid-crystal displays, bistable displays, organ-
ic emissive displays, and microdisplays are beginning to be exploited [83].
The development of electronic displays began with the cathode-ray tube
(CRT), which was first explored in the nineteenth century [86]. The first
tube, invented by Braun, used a deflection and a fluorescent material for the
display screen. Real progress began when large scale production of mono-
chrome CRT-based television (TV) receivers commenced around 1945. The
first successful shadow mask color picture tube was developed in the 1950s.
The manufacturing expertise developed in this field was applied to the com-
puter display industry. This kind of application/technology matching has
been a hallmark of all successful display types: TV, digital watches, calcula-
tors, computer monitors, portable computers, etc.
Nevertheless, it rapidly became evident that the traditional bulky CRT
was technically impossible to convert into a flat device. This achievement
was fulfilled with the active-matrix-addressed Liquid Crystal Display (LCD)
[87, 88]. Since the discovery of liquid crystallinity in 1888 by Reinitzer, one
by one, every technical challenge posed by the CRT standard has been met
and in many cases exceeded: operating lifetime, contrast ratio, response
time, high power consumption, view-angle limitations. Today, LCD monitors
are found in most electronics retail store. The writing and the addressing
functions are accomplished by an electrooptical path. This starts with the
electrical generation of an ultraviolet-emitting plasma followed by a conver-
sion to red, green and blue wavelengths and a distribution of white light.
This complexity coupled with production costs have stimulated the search
for other technologies, such as plasma display, field emission display, organ-
ic light-emitting diode displays, or microdisplays [83–85]. These latter dis-
plays are defined as sub-1.5 in displays that are viewed indirectly either
through magnifying optics near the eye or through a projection system that
puts a large high pixel-count on a screen. The displays are built directly on a
semiconductor substrate, which allows the integration of the electronics on
to the same substrate for higher performance with lower manufacturing
costs. Without doubt, much work remains to be done since those microdis-
plays will be based on a complex interaction of optics, electronics, packaging
and chemistry.
Basically the writing and the addressing functions can be accomplished
in a flat-panel display by several nonelectronic means: thermal, chemical,
acoustic, mechanical, temporal, optical and various combinations thereof.
Acoustically addressed displays have never become commercially wide-
spread due to complexity and efficiency problems. Mechanical methods
were employed early in the development of displays and went out of favor
until the recent arrival of microelectromechanical modulators. Electronic
addressing and writing is intuitively the best, but the present state of materi-
als development demands some adaptation to the optical writing phenome-
236 J.-F. Ltard et al.

non. Thermal and chemical addressing and writing have not been successful
to date, typically due to reversibility and lifetime issues.
Concerning the SCO materials, in fact, it was only during the 1980s that
researchers realized that thermal properties could be exploited in devices.
Kahn et al. [18–20] listed a series of requirements that SCO materials must
fulfill before any use in data display devices: (i) the transition must be
abrupt, occurring within a few Kelvins, both in the warming mode at T1/2"
and in the cooling mode at T1/2#; (ii) the gHS(T) curve must exhibit a hyster-
esis effect of at least 40 K. The ideal situation is when ambient temperature
is centered in the middle of the thermal hysteresis loop; (iii) the transition
between the low-spin (LS) and high-spin (HS) states must be accompanied
by an easily detectable response, such as a change of color; (iv) the transi-
tion must be complete both at low temperature (gHS=0) and at high temper-
ature (gHS=1); (v) finally, the system must be chemically stable, and must
not be subject to fatigue over successive thermal cycles.
Not surprisingly, it quickly became evident that no pure SCO compound
met all the criteria. For instance, in the series of mononuclear compounds,
sufficiently broad hysteresis loops have been observed [57, 58] but the pres-
ence of aromatic rings connected to the FeN6 core, indispensable to assure
p-p stacking, also gives rise to intra-ligand or charge-transfer bands which
mask the d-d bands in the visible region of the spectrum. Therefore, these
materials do not exhibit a drastic visible color change with the ST and are
thus inappropriate for display devices. To date, the best hope for successful
application is provided by the iron(II)-1,2,4-triazole SCO materials. Almost
all the criteria above have gradually been addressed and are now close to be-
ing realized. The developments in the modifications of these systems are
considered below.

4.2
Towards the Design of Data Display

As already mentioned, many groups have investigated metal triazole com-

plexes. The reasons for this interest are numerous, such as the modeling of
enzymatic reactions, use as pharmaceutical or agricultural components, ap-
plications in materials science. One of the properties of triazole relevant to
the last point is its ligand field strength which is in the vicinity of the SCO
region for iron(II). Depending on the substituents within the triazole ring,
the transition temperatures span the range 100 to 400 K [65–67, 89].
In fact, the features of the first thermal spin transitions recorded with ir-
on(II)-1,2,4-triazole derivatives were not suitable for any of the applications
mentioned above. Indeed, most of these complexes displayed SCO phenom-
ena above room temperature, with relatively gradual ST and with hysteresis
loops narrower than 50 K. However, their ease of synthesis, the chemical sta-
bility of the final iron(II) complexes, and the pronounced color change ac-
Towards Spin Crossover Applications 237

companying the ST strongly indicated that these materials may present some
potential in the design of display devices. The absorption spectrum of the
LS state corresponds to a broad band centered around 520 nm (1A1g!1T1g
transition) while the HS state does not show any band in the visible region,
the 5T2g!5Eg transition being located around 850 nm. In addition to this, it
was found in some instances that very abrupt ST with well-shaped thermal
hysteresis loops perfectly reproducible over successive thermal cycles, can
be obtained. This was found for [Fe(NH2trz)3](NO3)2 with T1/2"=342 K and
T1/2#=310 K [90], and for [Fe(Htrz)2(trz)](BF4) with T1/2"=383 K and
T1/2#=345 K [19, 20, 91–93]. Several approaches have since been used to dis-
place the thermal hysteresis loop so as to encompass room temperature.
One of the first approaches was to replace to some extent iron(II) ions by
non-spin crossover atoms. This method was originally inspired by the work
of Sorai and G
tlich on the tris(2-picolylamine)iron chloride system [94].
These authors reported a linear change of the ST temperature with dilution
by zinc. Similar effects were effectively observed in the 1,2,4-triazole series
[95–97]. However, it soon became evident that this method resulted in a dra-
matic reduction of the color contrast between LS and HS states, relative to
that of the pure compounds. Moreover, the abruptness of the thermal transi-
tion was found to decrease with the metal dilution.
A much more successful approach consisted of the design of “molecular
alloys” by combining different ligands in the preparation of salts. The gener-
al formula of an alloy complex, combining R1trz and R2trz triazole ligands,
was [Fe(R1trz)33x(R2trz)3x]A2·nH2O. The T1/2" and T1/2# of, for instance,
[Fe(Htrz)3–3x(NH2trz)3x](ClO4)2·H2O, was found to vary linearly as a func-
tion of the concentration x [20]. The two ligands were chosen in such a way
that the transition temperatures for one of the pure compounds are above
room temperature, and for the other pure compound are below room tem-
perature. For a given composition, room temperature has been found to fall
in the middle of the hysteresis loop. Similar results were obtained by mixing
counter-anions instead of mixing triazole ligands. The latter approach was
based on the fact that the thermal spin transition for a given type of cationic
chains, [Fe(Rtrz)3], depends on the radius of the counter-anion, A [98]. The
spin transition regime can then be fine-tuned through the composition
of alloys, [Fe(Rtrz)3]A1A2·nH2O. In the particular case of [Fe(NH2trz)3]
(NO3)1.7(BF4)0.3, the T1/2" and T1/2# were found on either side of room tem-
perature and the width of the thermal hysteresis was around 60 K [20].

4.3
Fabrication of Displays

The first principles of the application of iron-triazole systems in a display

device were reported by Kahn et al. [20, 99]. The substrate consists of an
aluminum plate, on which resistive dots and connecting electrodes were pre-
 238 J.-F. Ltard et al.

Fig. 4 Laboratory scale display. ST screens printed layer before writing (left) and after
thermal writing (right). Schematic illustration of the display device involving the ST lay-
er, the aluminum plate containing the ICMCB logo and the heat dissipater resistance
(bottom). When the temperature exceeds the critical “up temperature” the material tran-
sits from purple to white and the ICMCB information is addressed

viously screen printed. The dots are addressed through columns and rows. A
layer of the active SCO material is deposited as a printed ink on the sub-
strate. The criteria required for the ink are (i) to adhere to the substrate, (ii)
to be chemically inert to the spin transition materials, (iii) to protect the
material from environmental influences, and (iv) to be mechanically com-
patible with a printing procedure involving heat. The dots, when electrically
addressed, act as heat dissipaters. When the temperature is above T1/2", the
material passes from violet to white. The information is stored as long as the
system remains at a temperature within the hysteresis loop. To erase the in-
formation, the temperature has to be lowered below T1/2#. This can be
achieved with a Peltier element. Various related displays can be derived from
this principle.
Figure 4 shows a display device based on the displacement of a heat alu-
minum plate [100]. When the heat plate is in contact with the polymer sur-
face, the temperature increases above T1/2" and the thermochromic effect is
observed. The ICMCB logo results from the design of the heat aluminum
plate. The white color corresponds to the region where the polymer is in
contact with the heat plate. In this latter system, the requirement of thermal
hysteresis is not necessary. The presence of a well-defined abrupt thermal
spin transition with a very low residual of HS and LS fractions is, however,
Towards Spin Crossover Applications 239

more crucial. In this context, triazole derivatives are obviously very good
candidates.

4.4
Specific Aspects

In this particular section, our target is to present other perspectives which

can be, in some aspects, related to the display arena. Let us first consider the
very high sensitivity of the SCO materials to any temperature or pressure
changes. This property may have some important applications in the sensor
domain. For instance, in some cases the spin state is governed by the syner-
gy between the spin transition and a dehydration-rehydration process. This
happens, for example, for polymeric systems such as [Fe(NH2trz)3](tosy-
late)2·3H2O [101] and [Fe(hyettrz)3]A2·3H2O (hyettrz=4-(20 -hydroxyethyl)-
1,2,4-triazole and A=3-nitrophenyl-sulfonate) [102]. As the temperature in-
creases, the non-coordinated water molecules are released, followed by an
abrupt spin transition at 370 K from the metastable low spin state to the sta-
ble high-spin state. This phenomenon is accompanied by a pronounced
thermochromism. The color changes from violet to white. The transition
temperatures for the dehydrated material, [Fe(hyettrz)3]A2, are T1/2"=110 K
and T1/2#=100 K. In other words, once the starting violet material has
turned white by heating, it remains white (in the HS state) when brought to
room temperature. The materials cannot return to the hydrated LS state at
room temperature without some decomposition involving, for example, oxi-
dation of the iron(II) ions.
Such materials may be of use in applications in which detection of a specif-
ic temperature is required in a very simple, irreversible, and accurate way.
The SCO material then acts as a thermal sensor giving an optical response
only when T1/2" is reached, for instance as an alert mechanism when the tem-
perature exceeds an upper limit. More interestingly, this thermal addressing
accompanied by an abrupt and stable optical response may be used in a sin-
gle-use display. Incorporation of SCO materials in a prepaid stored-value card
has been envisaged [103]. Similar effects can be anticipated in many other
SCO compounds showing dehydration-rehydration processes [104], e.g.,
[Fe(bpp)2]X2·3H2O (bpp=2,6-bis(pyrazol-3-yl)pyridine and X=BF4 and I).
The above applications arise from the thermochromism associated with
the spin transition. In some particular applications, it may be interesting to
design temperature sensors which exploit a change of magnetic properties
[68]. In this context, applications in magnetic resonance imaging are partic-
ularly promising. At the critical temperature, the spin transition causes a
pronounced change of magnetic image, since the LS form is diamagnetic
and the HS form is paramagnetic. Note that in this case the occurrence of a
hysteresis loop is not important. More crucial is the presence of an abrupt
spin transition at a desired temperature. For instance, human cancer cells
240 J.-F. Ltard et al.

are more sensitive to heat than healthy cells. In the treatment, it is important
to reach the temperature not higher than that at which damage of both heal-
thy and diseased cells would take place. Work is underway to design materi-
als with transition temperatures at this upper limit.

5
Recent Progress
5.1
Molecular Switches and Non-linear Optics

Second order non-linear optical (SONLO) materials have recently attracted

much interest because of their potential applications in emerging optoelec-
tronic technologies [105]. Investigations were focused initially on purely in-
organic systems which were the first solids to demonstrate SONLO proper-
ties [106]. Attention was later directed towards organic materials. On the ba-
sis of a two-state model [107], donor-acceptor substituted p-systems have
been synthesized and provided large SONLO responses. However, one major
drawback of these push/pull substituted molecules is that strong non-linear-
ities are often antagonized by a strong electronic absorption in the visible
region. In an alternative way, octupolar molecules belonging to Td, D3h, or
C2v symmetry groups have attracted attention since significant SONLO mo-
lecular b values could be achieved with non-dipolar molecules [108]. Coor-
dination compounds are extremely promising since NLO active ligands
around the metal ion give highly symmetrical molecules and display metal-
to-ligand (MLCT) or ligand-to-metal (LMCT) or metal centered (MCCT)
charge transfer transitions in addition to the traditional p-p* transitions of
the conjugated system. Moreover, the symmetry properties of d-orbitals in
coordination compounds are different from the s- and p-orbitals usually
found in organic molecules [109].
In addition, coordination compounds are also of interest since they may
be used to design switchable non-linear optical materials. The concept of
such molecular switches has recently opened a large number of perspectives
[110, 111]. Up to now, switchable SONLO responses of organic compounds
have been obtained by photoisomerization [112], phototautomerization
[113], or photocyclization [114]. The first example of a coordination com-
pounds-based molecular switch was reported by Sakaguchi for a rutheni-
um(II) derivative in Langmuir-Blodgett (LB) films [115]. Coe et al. later used
the redox-switching of the Ru-metal center to design various switchable
SONLO compounds [116].
To achieve a pronounced SONLO switching effect, a molecule must be
stable in two states that exhibit very different b responses. Complete re-
versibility and high switching speed are also highly desirable for practical
Towards Spin Crossover Applications 241

applications. Basically, such fundamental requirements are exhibited by the

SCO phenomenon. An open question, however, is whether the spin transi-
tion between a paired and an unpaired electronic configuration is able to
switch the SONLO response. The work previously reported by Di Bella et al.
[116, 118] suggests that the NLO response is dependent on the electronic
configuration of the metal center. The b responses of (N,N0 -disalicylidene-
1,2-phenylene-diaminato)M(II) compounds in planar geometry are found
larger for paramagnetic Cu- and Co-derivatives than for the diamagnetic Ni-
complex [117, 118]. In parallel with this work, we have recently examined
the role of the metal electronic configuration in octahedral geometry by sys-
tematically varying the metal ions (MII=Mn, Fe, Co, Ni, and Zn) in the
two series of coordination compounds [M(PM-BiA)2(NCS)2] and [M(PM-
2PEA)2(NCS)2] [119].
Direct hyper-Rayleigh scattering measurements have shown an enhance-
ment of the molecular hyperpolarizabilities (b) from the d8 to the d5 metal
electronic configuration. Highest SONLO responses were recorded for Mn
compounds with full open-shell configuration (five unpaired electrons) and
lowest values were obtained for Ni compounds with two unpaired electrons
[119]. In fact, by taking into account the inverted vectorial direction of the
metal-to-ligand charge transfer (MLCT) transition and the intra-ligand
charge transfer (ILCT) transition (see Fig. 5), a direct enhancement of the
bMLCT along the d8-d5 metal electronic configuration was observed. The
ILCT transition is characterized by an intra-ligand charge transfer directed
towards the pyridyl group, while in MLCT transition with the metal as donor
the charge transfer occurs from the metal to the pyridyl group. The sum of
the contributions of the MLCT and ILCT transitions leads to destructive in-
terference. Assuming that the SONLO response of the Zn complexes directly
reflects the ILCT contribution, the b values of the other compounds have
been corrected to give the pure bMLCT contributions (see Fig. 5). In both
families of [M(PM-BiA)2(NCS)2] and [M(PM-2PEA)2(NCS)2], the highest
bMLCT values were found for the Mn compounds which posses five unpaired
electrons. The b values of the Fe compounds in the paramagnetic HS state
with four unpaired electrons (t2g)4(eg)2 were found to be larger than those of
the diamagnetic Zn compounds. This finding indirectly suggested that if the
b value of the diamagnetic Fe compound (LS state) with six paired electrons
is modeled by the response of the diamagnetic Zn ion, a drastic NLO change
accompanying the SCO phenomenon can be expected [119].
Unfortunately, to the best of our knowledge, there exists no direct experi-
mental proof of a spin crossover effect on the quadratic NLO response. The
only evidence was obtained via a theoretical investigation performed by
Lacroix [120] on the [Fe(5-NO2-Sal-N(1,4,7,10))] complex which undergoes
a two step spin transition. An increase of about 25% of the NLO signal was
found between the LS and the HS forms, mainly due to the structural chang-
es occurring in the coordination sphere upon spin-state transition. Never-
242 J.-F. Ltard et al.

Fig. 5 Left: molecular hyperpolarizabilities associated to the metal-to-ligand charge

transfer (MLCT) transition (bMLCT) vs the d10–d5 metal electronic configuration of the
metal complexes, [M(PM-BiA)2(NCS)2] (1) and [M(PM-2PEA)2(NCS)2] (2); right vectori-
al direction of the MLCT and the intra-ligand charge transfer (ILCT) transitions

theless, the recent application of Langmuir-Blodgett films [36, 37] may yield
some new insights. The design of switchable SONLO materials based on the
iron(II) SCO phenomenon may then become achievable.

5.2
Optical Addressing

Another interesting feature of the SCO phenomenon is that it offers the pos-
sibility to address spin states by light irradiation [16]. At the present time,
as mentioned previously, numerous processes in communication networks
involve optical connections. The information, for instance, is transferred
over long distances in the form of optical signals propagating through inter-
connected optical fibers. The design of materials allowing optical manipula-
tion is then of major importance. In the domain of transition metal com-
plexes, there are several possibilities to induce light sensitive electronic
changes accompanied by a drastic modification of magnetic and/or optical
properties [121]. Among the best known and extensively studied systems,
we cite metal-centered thermal spin transition (observed in iron(II) SCO
materials), valence tautomerism (catecholate complexes of Co(II)), the met-
Towards Spin Crossover Applications 243

al-to-metal charge transfer (Prussian blue analogues), metal-to-ligand

charge transfer (nitroprusside complexes), and finally, all ligand modifica-
tions which induce change of the spin state, such as isomerization (stil-
benoid complexes) and open/close (diarylethene type ligands) processes.
The first photomagnetic effect on an iron(II) SC compound in solid state
was reported by Decurtins et al. in 1984 [122]. The authors demonstrated
the possibility of converting a LS state into a metastable HS state at low tem-
peratures (50 K) by using green light irradiation. This phenomenon, called
Light-Induced Excited Spin-State Trapping (LIESST), was extended, later by
Hauser who showed that red light switches the system back to the LS state
(reverse-LIESST) [123]. At present, most of the photomagnetic studies on
iron(II) complexes involve the LIESST and reverse-LIESST processes with
the ultimate goal of obtaining suitable materials for technical applications
such as displays, information storage and sensors. In regard to these pro-
spects, a deficiency of the LIESST effect is its temperature domain. The
metastable HS state usually decays within minutes above 50 K. The photoin-
duced HS information is consequently quickly erased, except at very low
temperatures.
The HS!LS relaxation process in iron(II) metal complexes has been in-
vestigated in considerable detail [124–128]. It is now established that the ki-
netics follow the non-adiabatic multi-phonon theory with a temperature in-
dependent relaxation below ~50 K and a thermally activated process above
~50 K which can be regarded as a tunneling from thermally populated vibra-
tional levels of the HS state [129]. Nevertheless, in some unexplained cases,
long-lived metastable HS states have been reported above 80–100 K [130–
132]. It seems that some unknown secondary features in the nearby lattice
surroundings of a light-generated HS complex molecule are able to stabilize
the metastable HS state.
In the last four years, in collaboration with the Motorola company, we
have introduced the notion of critical LIESST temperature, T(LIESST),
defined as the temperature for which the light-induced HS information is
erased [133]. Our idea was to constitute a database which collects the photo-
magnetic properties of a large number of iron(II) SCO compounds. It is evi-
dent that such T(LIESST) values are somewhat dependent on the detection
method. This approach, however, is useful as it allows the rapid comparison
of the photomagnetic properties of a series of SCO compounds, and the de-
termination of the factors which play a role in the light-induced conversion
effect.
The procedure adopted to record the T(LIESST) value of SCO compounds
is the following (see Fig. 6): the sample is first slowly cooled to 10 K. The
sample in the LS state is then irradiated into the metal-to-ligand charge-
transfer band (1MLCT) or into the 1A1!1T absorption band. The magnetic
response of the compound is found to change rapidly, and usually after 1 h
a photo-stationary limit is reached. The light irradiation is then switched
244 J.-F. Ltard et al.

Fig. 6 T(LIESST) Procedure. Temperature dependence of cMT recorded in the cooling

mode without irradiation (filled diamonds), during 1 h of irradiation at 10 K (empty cir-
cles) and in the warming mode after the light irradiation was applied for 1 h, then turned
off (filled squares). The insert graph shows the derivative d(cMT)/dT plot as function of
the temperature

off, and the temperature is increased at the rate of 0.3 K/min [133]. At low
temperature, the HS!LS relaxation is controlled by the tunneling process
up to ca. 40 K, and therefore the kinetics is very slow with respect to the
time scale of the experiment. As a matter of fact, the time required to record
the data up to 40 K is approximately three hours. Above ca. 40 K, the system
reaches the thermally activated domain, and the HS!LS relaxation becomes
increasingly more rapid. As a result of this behavior, cMT remains nearly
constant as T increases from 10 K up to 40 K, then drops rapidly. In fact, a
slight increase of cMT is usually observed between 10 to ca. 20 K, owing to
the zero-field splitting within the spin quintet HS state. The dcMT/dT vs T
plot shows a minimum at a temperature that we define as the T(LIESST) val-
ue; at this temperature, the light-induced information is erased in the
SQUID setup. The behavior of [Fe(PM-BiA)2(NCS)2] (PM-BiA=N-(2-pyridyl-
methylene)aminobiphenyl) is typical and is illustrated in Fig. 6.
To date, more than 60 SCO compounds have been characterized by the
same T(LIESST) procedure. It is becoming increasingly evident that a direct
Towards Spin Crossover Applications 245

Fig. 7 Variation of T(LIESST) vs T1/2 for spin crossover compounds. Data for the 120 and
100T0 lines are reported in [133]. The 150T0 line is the result of the linear regression ob-
tained with the [Fe(bpp)2]X2·nH2O family [134] and the 200T0 line is based on the work
of Hashimoto [135]

relation exists between the T(LIESST) value and the thermal spin transition
temperature, (T1/2); i.e., T(LIESST)=T00.3 T1/2 with T0 estimated at T!0
(Fig. 7) [133, 134]. The 150T0 line followed by the [Fe(bpp)2]X2·nH2O com-
plexes gives for the first time measurable T(LIESST) values with T1/2 around
room temperature [134]. The 200T0 line was recently reported by Hashimoto
et al. [135] for the Co-Fe Prussian blue series where a charge-transfer-in-
duced spin transition occurs between the FeII(LS, S=0)-CN-CoIII(LS, S=0)
site (LT phase) and the FeIII(LS, S=1/2)-CN-CoII(HS, S=3/2) site (HT phase).
All these findings open up exciting prospects and extensive research ef-
forts will be required to identify the key factors involved in the stabilization
of the metastable HS state. We are currently pursuing this by investigating
the influence of the coordination sphere (nature of the organic core, of the
anion, of the hydration degree) and of the nature of the cooperativity on the
magnitude the T(LIESST) [136].

6
Conclusions
Extensive progress has been made over the last 40 years in the spin cross-
over field and, nowadays, there are many reasons to believe in the use of
such molecular materials in a genuine device in the not too distant future.
The dimensions of synthetic molecular systems are now close to overlap
with the levels of miniaturization required in microelectronics. At the same
246 J.-F. Ltard et al.

time, the growing need for faster computers and internet applications is a
stimulus to the exploration of innovative materials and operating principles
for digital processing and communications. Within this context, the investi-
gations of compounds exhibiting in a synergistic way multiple physical
properties, as illustrated by the SCO phenomenon, are of particular interest.
Of course, at the present early stages, only rudimentary examples of practi-
cal applications are known and, in certain cases, only a presentation of likely
concepts has been given. Considerable fundamental studies are thus needed
for a further understanding and to apply the many facets of this broad re-
search field. It is for many of us an exciting and stimulating challenge.

Acknowledgements We thank the European Commission for granting the TMR-Network

“Thermal and Optical Switching of Spin States (TOSS)”, Contract No. ERB-FMRX-CT98-
0199 and the COST D14 program “Non-linear magneto-optical properties of organic and
metallo-organic compounds: potential molecule-based switches”. The financial support
from the CNRS-ACO “Mmoire holographique dans des disques photopolymres dops”
and DFG/CNRS bilateral project “Towards the Design of Molecular Switches”, University
and city of Bordeaux, is greatly appreciated.
Most of the work discussed here was initiated by the late Olivier Kahn to whom the au-
thors dedicate this review.

References

1. Moore GE (1965) Electronics 38:114

2. Moore GE (1975) In: Digest of the 1975 International Electron Devices meeting,
IEEE, New York, p 1113
3. Feynman RP (1960) Eng Sci 23:22
4. Feynman RP (1960) Saturday Rev 43:45
5. Aviram A, Ratner MA (1974) Chem Phys Lett 29:277
6. Day P (1996) Chem Br 32:29
7. Day P (1998) Proceeding of the Royal Institution of Great Britain, vol 69. Oxford
University Press
8. Chen J, Reed MA, Rawlett AM, Tour JM (1999) Science 286:1550
9. Metzger RM (1999) Acc Chem Res 32:950
10. Balzani V, Credi A, Raymo FM, Stoddart JF (2000) Angew Chem Int Ed 39:3348
11. Huang Y, Duan X, Cui Y, Lauhon LJ, Kim K-H, Lieber CM (2001) Science 294:1313
12. Raymo FM (2002) Adv Mater 14:401
13. Balzani V, Credi A, Venturi M (2002) Chem Eur J 8:5525
14. Caroll RL, Gorman CB (2002) Angew Chem Int Ed Eng 41:4378
15. The international technology roadmap for semiconductors (2001) Update (http://
public.itrs.net/)
16. G
tlich P, Hauser A, Spiering H (1994) Angew Chem Int Ed Engl 33:2024
17. Bousseksou A, Negre N, Goiran M, Salmon L, Tuchagues J-P, Boillot M-L, Boukhed-
daden K, Varret F (2000) Eur Phys JB 13:451
18. Zarembowitch J, Kahn O (1991) New J Chem 15:181
19. Kahn O, Krber J, Jay C (1992) Adv Mater 4:718
20. Kahn O, Jay-Martinez C (1998) Science 279:44
Towards Spin Crossover Applications 247

21. Service RF (1996) Science 274:1834

22. Gimzewski JK, Joachim C (1999) Science 283:1683
23. Joachim C, Gimzewski JK, Aviram A (2000) Nature 408:541
24. Tour JM, Kozaki M, Seminario JM (1998) J Am Chem Soc 120:8486
25. Fabbrizzi I, Licchelli M, Pallavicini P (1999) Acc Chem Res 32:846
26. Willner I, Katz E (2000) Angew Chem Int Ed 39:1180
27. Irie M (2000) Chem Rev 100:1685
28. Yokoyama Y (2000) Chem Rev 100:1717
29. Berkovic G, Krongauz V, Weiss V (2000) Chem Rev 100:1741
30. Feringa BL, Van Delden RA, Koumura N, Geertsema EM (2000) Chem Rev 100:1789
31. Feringa BL (2001) Acc Chem Soc 101:504
32. Diederich F (2001) Chem Commun 219
33. De Silva AP, Gunaratne ON, MCCoy CP (1993) Nature 42:364
34. De Silva AP, De Silva AA, Dissanavake AS, Sandanayake KRAS (1989) J Chem Soc
Chem Commun 1054
35. Special Issue on “Optical networking” (1994) Bell Labs Tech J 4:3
36. Soyer H, Mingotaud C, Boillot M-L, Delhas P (1998) Thin Solids Films 327/329:435
37. Soyer H, Mingotaud C, Boillot M-L, Delhas P (1998) Langmuir 14:5890
38. Ltard J-F, Nguyen O, Soyer H, Mingotaud C, Delhas P, Kahn O (1999) Inorg Chem
38:3020
39. Kugler J (2002) Computer Science, 321 and references cited therein
40. Thompson DA, Best JS (2000) IBM J Res Dev 44:311
41. Shen Y, Friend CS, Jiang Y, Jakubczyk D, Swiatkiewicz J, Prasad PN (2000) J Phys
Chem B 104:7577
42. Eschenfelder AH (1970) J Appl Phys 42:1372
43. Wildmann M (1974) IEEE Trans Magn 10:509
44. Ashley J et al. (2000) IBM J Res Develop 44:341
45. Binnig G, Despont M, Drechsler U, H berle W, Lutwyche M, Vettiger P, Mamin HJ,
Chui BW, Kenny TW (1999) Appl Phys Lett 74:1329
46. Vettiger P et al. (2000) IBM J Res Develop 44
47. Kahn O, Launay JP (1988) Chemtronics 3:140
48. Fujita W, Awaga K (1999) Science 286:261
49. Itkis ME, Chi X, Cordes AW, Haddon (2002) Science, 296:1443
50. Stoddard JF (2001) Acc Chem Res 34:410
51. Stoddard JF, Heath JR (2000) Science 289:1172
52. Spiering H, Kohlhaas H, Romstedt H, Hauser A, Bruns-Yilmaz C, Kusz J, G
tlich P
(1999) Coord Chem Rev 190/192:629
53. Hauser A, Jeftic J, Romstedt H, Hinek R, Spiering (1999) Coord Chem Rev 109/
192:471
54. Real JA, Gaspar AB, Niel V, Muoz MC (2003) Coord Chem Rev 236:121
55. Knig E (1987) Prog Inorg Chem 35:527
56. Knig E, Ritter G (1976) Solid State Comm 18:279
57. Ltard J-F, Guionneau P, Codjovi E, Lavastre O, Bravic G, Chasseau D, Kahn O (1997)
J Am Chem Soc 119:10,861
58. Zhong ZJ, Tao J-Q, Yu Z, Dun C-Y, Liu Y-J, You X-Z (1998) J Chem Soc Dalton Trans
327
59. Guionneau P, Ltard J-F, Yufit DS, Chasseau D, Bravic G, Goeta AE, Howard JAK,
Kahn O (1999) J Mater Chem 9:985
60. Vreugdenhil W, Van Diemen JH, De Graff AG, Haasnoot JG, Reedijk J, Van der Kraan
AM, Kahn O, Zarembowitch J (1990) Polyhedron 9:2971
248 J.-F. Ltard et al.

61. Real JA, Andrs E, Muoz MC, Julve M, Granier T, Bousseksou A, Varret F (1995)
Science 268:265
62. Kitazawa T, Gomi Y, Takahashi M, Takeda M, Enomoto M, Miyazaki A, Enoki T
(1996) J Mater Chem 6:119
63. Garcia Y, Kahn O, Rabardel L, Chansou B, Salmon L, Tuchagues JP (1999) Inorg
Chem 38:4663
64. Garcia Y, van Koningbsruggen PJ, Bravic G, Guionneau P, Chasseau D, Cascarano
GL, Moscovici J, Lambert K, Michalowicz A, Kahn O (1997) Inorg Chem 36:6357
65. Haasnoot JG, Vos G, Groeneveld WL (1977) Z Naturforsch B 32:1421
66. Lavrenova LG, Ikorskii NV, Varnek VA, Oglezneva IM, Larionov SV (1986) Koord
Khim 12:207
67. Lavrenova LG, Ikorskii NV, Varnek VA, Oglezneva IM, Larionov SV (1990) Koord
Khim 16:654
68. Kahn O (1999) Chem Br 24
69. Montant S, Ltard S, Freysz E, Chastanet G, Marcen S, Ltard J-F (2002) 5 TMR-TOSS
Meeting, 7th Spin Crossover Family Meeting, Mainz, Germany
70. Bousseksou A, Molnr G, Tuchagues JP, Menndez N, Codjovi E, Varret F (2003) C
R—Chim 6:329
71. Ltard J-F (2002) SFC Eurochem Toulouse
72. Bolvin H (1993) PhD thesis, University of Paris-Sud
73. M
ller EW, Spiering H, G
tlich P (1982) Chem Phys Lett 93:567
74. Tsuchiya N, Tsukamoto A, Ohshita T, Isobe T, Senna M, Yoshioka N, Inoue H (2001)
Solid State Sci 3:705
75. Real JA, Zarembowitch J, Kahn O, Solans X (1987) Inorg Chem 26:2939
76. Vos G, Le FÞbre RA, De Graaff RAG, Haasnoot JG, Reedijk J (1983) J Am Chem Soc
105:1682
77. Breuning E, Ruben M, Lehn J-M, Renz F, Garcia Y, Ksenofontov V, G
tlich P,
Wegelius E, Rissanen K (2000) Angew Chem Int Ed 39:2504
78. Pillai V, Kumar P, Hou MJ, Ayyub P, Shah DO (1995) Adv Colloid Interface Sci 55:241
79. Moulik SP, De Panda GC, Bhowmik BB, Das AR (2001) Langmuir 17:1811
80. Vaucher S, Li M, Mann S (2000) Angew Chem Int Ed 39:1793
81. Collin JP, Dietrich-Buchecker C, Gavia P, Jimenez-Molero MC, Sauvage J-P (2001)
Acc Chem Res 34:477
82. Enachescu C, Oetliker U, Hauser A (2002) J Phys Chem B Condens Matter 106:9540
83. Mentley DE (2002) Proceedings of the IEEE 90:453
84. Kimmel J, Hautanen J, Levola T (2002) Proceedings of the IEEE 90:581
85. Wisnieff RL, Ritsko JJ (2000) IBM J Res Dev 44:409
86. Tannas LE (1985) Flat panel displays and CRTs. Van Nostrand, New York, pp 1–158
87. Marconi Wireless Telegraph Company (1936) British Patent 441 274
88. Williams RJ (1963) J Chem Phys 39:384
89. Haasnoot JG (2000) Coord Chem Rev 200/202:131
90. Bausk NV, Erenburg SB, Mazalov LN, Lavrenova LG, Ikorskii VN (1994) J Struct
Chem 35:509
91. Krber J, Audire JP, Rene C, Codjovi E, Kahn O, Haasnoot JG, Grolire F, Jay C,
Bousseksou A, Linars J, Varret F, Gonthier-Vassal A (1994) Chem Mater 6:1404
92. Kahn O, Codjovi E (1996) Phil Trans R Soc Lond A 354:359
93. Sugiyarto KH, Goodwin HA (1994) Aust J Chem 47:263
94. Sorai M, Ensling J, G
tlich P (1976) Chem Phys 18:199
95. Lavrenova LG, Ikorskii NV, Varnek VA (1993) Zh Strukt Khim 34:145
96. Varnek VA, Lavrenova LG (1994) J Struct Chem 35:842
Towards Spin Crossover Applications 249

97. Varnek VA, Lavrenova LG (1995) J Struct Chem 36:97, 104

98. Lavrenova LG, Yudina NG, Ikorskii NV, Varnek VA, Oglezneva IM, Larionov SV
(1995) Polyhedron 14:1333
99. Jay C, Groliere F, Kahn O, Krber J (1993) Mol Cryst Liq Cryst 234:255
100. Ltard J-F et al. (2000) Third TMR-TOSS Meeting, 5th Spin Crossover Family Meet-
ing, Oegstgeest, Netherlands
101. Codjovi E, Sommier L, Kahn O (1996) New J Chem 20:503
102. Garcia Y, Van Koningsbruggen P, Codjovi E, Lapouyade R, Kahn O, Rabardel L
(1997) J Mater Chem 7:857
103. Kahn O, Jay-Martinez C, Garcia Y, Codjovi E (1998) Patents EP0842988
104. Sugiyarto KH, Craig DC, Rae AD, Goodwin HA (1994) Aust J Chem 47:869
105. Chemla DS, Zyss J (1987) Non-linear optical properties of organic molecules and
crystals. Academic Press, Orlando, FL
106. Franken PA, Hill AE, Peters CW (1961) Phys Rev Lett 7:118
107. Oudar JL (1977) J Chem Phys 67:446
108. Zyss J (1991) Non-linear Opt 1:3
109. Long NJ (1995) Angew Chem Int Ed Engl 34:21
110. Lehn JM (1995) Supramolecular chemistry-concepts and perspectives. VCH, Weinheim
111. Coe BJ (1999) Chem Eur J 5:2464
112. Loucif-Sabi R, Nakatani K, Delaire JA, Dumont M, Sekkat Z (1993) Chem Mater 5:229
113. Nakatani K, Delaire JA (1997) Chem Mater 9:2682
114. Gilat SL, Kawai SH, Lehn JM (1995) Chem Eur J 1:275
115. Nagumura T, Sakaguchi H, Matsuo T (1992) Thin Solid Films 210:160
116. Coe BJ, Chamberlain MC, Essex-Lopresti JP, Gaines S, Jeffery JC, Houbrechts S, Per-
soons A (1997) Inorg Chem 36:3284
117. Di Bella S, Fragal I, Ledoux I, Marks TJ (1995) J Am Chem Soc 117:9481
118. Di Bella S, Fragal I, Marks TJ, Ratner MA (1996) J Am Chem Soc 118:12,747
119. Gaudry JB, Capes L, Langot P, Marcn S, Kollmannsberger M, Lavastre O, Freysz E,
Ltard J-F, Kahn O (2000) Chem Phys Lett 324:321
120. Averseng F, Lepetit C, Lacroix PG, Tuchagues J-P (2000) Chem Mater 12:2225
121. G
tlich P, Garcia Y, Woike T (2001) Coord Chem Rev 219/221:839
122. Decurtins S, G
tlich P, Khler CP, Spiering H, Hauser A (1984) Chem Phys Lett 105:1
123. Hauser A (1986) Chem Phys Lett 124:543
124. Xie C-L, Hendrickson DN (1987) J Am Chem Soc 109:6981
125. Hauser A (1991) J Phys Chem 94:2741
126. Hauser A, Vef A, Adler P (1991) J Chem Phys 95:8710
127. Hauser A (1991) Coord Chem Rev 111:275
128. A Hauser (1995) Comm Inorg Chem 17:17
129. Buhks E, Navon G, Bixon M, Jortner J (1980) J Am Chem Soc 102:2918
130. Wu CH, Jung J, Gantzel PK, G
tlich P, Hendrickson DN (1997) Inorg Chem 36:5339
131. Ltard J-F, Guionneau P, Rabardel L, Howard JA, Goeta AE, Chasseau D, Kahn O
(1998) Inorg Chem 37:4432
132. Renz F, Oshio H, Ksenofontov V, Waldeck M, Spiering H, G
tlich P (2000) Angew
Chem 39:3699
133. Ltard J-F, Capes L, Chastanet G, Moliner N, Ltard S, Real J-A, Kahn O (1999) Chem
Phys Lett 313:115
134. Marcen S, Lecren L, Capes L, Goodwin HA, Ltard J-F (2002) Chem Phys Lett 358:87
135. Shimamoto N, Ohkoshi S-S, Sato O, Hashimoto K (2002) Inorg Chem 41:678
136. Marcen S, Lecren L, Capes L, Guionneau P, Ltard J-F (2002) Final TMR-TOSS
Meeting, 7th Spin Crossover Family Meeting, Seeheim, Germany