# LUIGI MALAVOLTI

REMIO TESI DOTTORATO FIRENZE UNIVERSITY PRESS - UNIVERSITÀ DEGLI STUDI DI FIRENZE

P

Single molecule magnets sublimated on conducting and magnetic substrates

### PREMIO TESI DI DOTTORATO

– 50 –

### PREMIO TESI DI DOTTORATO

### Commissione giudicatrice, anno 2014

Luigi Lotti, *presidente della Commissione*

Tito Arecchi, *Area scientifica* Aldo Bompani, *Area Scienze Sociali* Franco Cambi, *Area Umanistica* Paolo Felli, *Area Tecnologica* Michele Arcangelo Feo, *Area Umanistica* Roberto Genesio, *Area Tecnologica* Mario Pio Marzocchi, *Area Scientifica* Adolfo Pazzagli, *Area Biomedica* Giuliano Pinto, *Area Umanistica* Salvatore Ruggieri, *Area Biomedica* Saulo Sirigatti, *Area Biomedica* Fiorenzo Cesare Ugolini, *Area Tecnologica* Vincenzo Varano, *Area Scienze Sociali* Graziella Vescovini, *Area Umanistica*

# **Single molecule magnets sublimated on conducting and magnetic substrates**

Firenze University Press 2015

Single molecule magnets sublimated on conducting and magnetic substrates / Luigi Malavolti. – Firenze : Firenze University Press, 2015. (Premio Tesi di Dottorato; 50)

http://digital.casalini.it/9788866559702

ISBN 978-88-6655-969-6 (print) ISBN 978-88-6655-970-2 (online)

### *Peer Review Process*

All publications are submitted to an external refereeing process under the responsibility of the FUP Editorial Board and the Scientific Committees of the individual series. The works published in the FUP catalogue are evaluated and approved by the Editorial Board of the publishing house. For a more detailed description of the refereeing process we refer to the official documents published on the website and in the online catalogue of the FUP (www.fupress.com).

*Firenze University Press Editorial Board*

G. Nigro (Co-ordinator), M.T. Bartoli, M. Boddi, R. Casalbuoni, C. Ciappei, R. Del Punta, A. Dolfi, V. Fargion, S. Ferrone, M. Garzaniti, P. Guarnieri, A. Mariani, M. Marini, A. Novelli, M.C. Torricelli, M. Verga, A. Zorzi.

This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0: http://creativecommons.org/licenses/by/4.0/)

**CC** 2015 Firenze University Press Università degli Studi di Firenze Firenze University Press Borgo Albizi, 28, 50122 Firenze, Italy www.fupress.com *Printed in Italy*

A Donatella, alla mia famiglia

"Beyond the horizon of the place we lived when we were young In a world of magnets and miracles Our thoughts strayed constantly and without boundary The ringing of the division bell had begun ..."

"High Hopes", Pink Floyd

# **Index**



$$\chi = \frac{\delta \mathcal{M}}{\delta H} \qquad \text{(1 )}$$

 

$$m = -\frac{\delta E}{\delta H} \quad \text{(2 )}$$

$$M = \frac{N \, \Sigma\_n (-\delta E\_n / \delta H) \exp(-E\_n / kT)}{\Sigma\_n \exp(-E\_n / kT)} \tag{3.3}$$

The measurement of magnetisation can be carried out by employing standard magnetometry techniques while for a correct evaluation of the susceptibility the use of alternating magnetic fields is required. These techniques are described in the following paragraphs.

### 1.1 Standard magnetometry

Most magnetometers are based on the inductive detection. These instruments work with static magnetic field generated by a superconductive magnet. The magnetic flux variation induced by the presence of the sample inside a coil is measured. The flux variation produces in the coil an electromotive force proportional to - 80/8t, where o is the magnetic flux through the detection loops. In the coil circuit this generates a current that can be measured. To induce a time change of o the magnetised sample is moved inside the detection coils in a region of homogenous magnetic field. During the measurements the sample is thus assembled on a rod and is moved inside a gradiometer circuit. This circuit is sensitive only to a variation of the gradient of the magnetic field and allows to separate the contribution of the sample from the drift of the homogenous field. In particular, the second order grad ometer circuit employed in our instruments is formed by three coils, a single coil with N windings placed between two coils made of N/2 windings each and wound in the opposite direction (see fig. left). The sample is moved from the centre of the top coil to the centre of the bottom co .. The current induced in the gradiometer cir cuit changes sign between opposite coils and the difference is assumed to be propo tional to the magnetisation of the sample. The proportionality factor is determined through a calibration with standard samples with well-known magnetisation, like metallic Pt.

Fig. Schematic representation of a second order gradiometer (left): the windings of the bottom and top coil are wound in the opposite direction respect to windings of the central one. In the magnetic measure the sample is moved up and down inside the gradiometer. A first order gradiometer made by two coils wound in the opposite way (right).

The sensitivity of the system can be enhanced by inductively coupling a Superconductive Quantum Interference Device (SQUID). This device, formed by a superconductive ring with a Josephson junction (Barone & Paternò 1982; Bukel 1991), is the most sensitive detector for the magnetic flux (Clark & Braginski 2004). A comparable, although slightly lower, sensitivity can be reached employing the so called Vibrating Sample Magnetometer (VSM). In this type of magnetometers the sample is moved inside a gradiometer (fig. 1 right) at frequencies in the range of 50-100 Hz. The signal is recorded by means of a lock-in detection system without the necessity of a SQUID. This makes the magnetometer better suited for measurements under fast varying magnetic fields, which would be not compatible with SQUID devices, the latter being very sensitive to the magnetic field. The VSM technique was therefore used to record the magnetic hysteresis presented in section 3 paragraph 2.

It is important to note that these magnetometers measure the magnetisation of the sample. The molar susceptibility x can be calculated assuming a linear dependence of the magnetisation on the field as:

$$\chi = \frac{\mathsf{m}}{\mathsf{N}} \cdot H \quad \text{(4 )}$$

where M is the magnetisation of the sample and N is the number of moles. The assumption of a linear dependence of the magnetisation on the field is not correct under strong magnetic fields. To have access to the susceptibility the so called ac susceptometry was employed.

### 1.2 Ac susceptometry

In these measurements the sample is positioned inside a coil, called primary. A small oscillating magnetic field, few Oe, can be generated by applying an alternate potential to the coil's ends. Inside the coil a first order gradiometer circuit is present, called secondary. Due to the opposite wounding of the two coils no voltage is induced from the primary to the secondary. However, if a sample is inserted in the coils it magnetisation will oscillates with the oscillating field unbalancing the secondary circuit, and generating a voltage that is proportional to the change in sample magnetisation. As the oscillating field is very small, the derivative of the magnetisation vs. the oscillating field, i.e. the susceptibility, is directly measured. The sample is moved inside the gradiometer to eliminate possible spurious signal due to a nonperfect balance of the gradiometer's windings.

It is possible to apply a constant field (Bo) along the same direction of the ac field. The total field experienced by the sample is thus:

0 + ℎ cos ሺͷሻ

ℎ 0

$$M(t) = M\_0 + \chi(t)h \qquad \qquad (6)$$

′ ′′ ( )

$$\chi(t) = \chi' \cos \omega t + \chi'' \sin \omega t \tag{7}$$

$$M(t) = M\_0 + (\chi' \cos \omega t + \chi'' \sin \omega t)h \qquad \text{(\(\otimes\))}$$

 −1 ≪ 1 −1 ≫ 1 −1

$$
\chi' = \frac{\chi\_{T-\chi\_S}}{1+\omega^2 \tau^2} + \chi\_S \qquad \text{(9 )}
$$

$$
\chi\_{\prime\prime} = \frac{(\chi\_T - \chi\_S)\omega\tau}{1+\omega^2 \tau^2} \qquad \text{(10 )}
$$

$$
\chi'(\omega) = \chi\_S + (\chi\_T - \chi\_S) \frac{1 + (\omega \tau)^{1 - a} \sin(\pi a/2)}{1 + 2(\omega \tau)^{1 - a} \sin(\pi a/2) + (\omega \tau)^{2 - 2a}} \tag{111}
$$

$$
\chi\_{\prime\prime}(\omega) = (\chi\_T - \chi\_S) \frac{1 + (\omega \tau)^{1 - a} \cos(\pi a/2)}{1 + 2(\omega \tau)^{1 - a} \sin(\pi a/2) + (\omega \tau)^{2 - 2a}} \tag{122}
$$

 ͳʹ 

$$\Delta \mathbf{s} = \frac{\kappa}{\rho \cdot x \cdot f\_c} \tan^{-1} \left\{ \mathbf{z} \tan \left[ \frac{\pi (f\_q - f\_c)}{f\_q} \right] \right\} \tag{13}$$

### 4 PhotoElectron Spectroscopy (PES)

The chemical characterisation of the sublimated films on conductive substrates can be performed by employing the photoemission technique. The technique is based on the photoemission effect, which was discovered by H. Hertz in the 1887. The effect was then rationalized in the following years by Einstein (1905) (Einstein 1905) and its application in spectroscopic studies was developed only in the '50s and '60s. The technique was refined during the last 60 years to achieve the actual state of the art.

The photoemission effect takes place when a photon transfers enough energy to an electron to overcome the potential barrier (binding energy) holding it to its atom. The exceeding energy imparted by the photon manifests as kinetic energy of the electron. Thus knowing the photon energy (hu) and the kinetic energy (Ekin) of the photoelectron it's possible to calculate the binding energy (EB) using the following formula:

$$E\_B = h\upsilon - E\_{Kin} + \phi\_{samp.} - \phi\_{spec.} \quad \text{(14 )}$$

where \$samp. and \$spec. are respectively the work functions of the sample and the spectrometer (fig. 3). The energy level diagram of the process is depicted in fig. 3 considering the case of a metallic sample electrically connected to the spectrometer and both grounded. This allows for minimising the charging effect maintaining a fixed potential during the photoemission process. The scheme considers a system at 0 K where the Fermi level is the highest occupied energy level. At the equilibrium the sample and the spectrometer have the same Fermi energy due to the electrical contact. When the photoelectron goes from the sample into the spectrometer it feels an accelerating or decelerating potential due to the difference in work function of the two materials (\$samp. - \$spec.). The work function is defined as the energy difference between the Fermi and the vacuum level and it depends on the material. Although the Fermi levels of the sample and spectrometer are the same they have different work functions. This introduces a shift in the detected kinetic energy of the electrons that must be corrected to be able to compare spectra acquired with different spectrometers.

The PES spectrum is acquired recording the number of electrons detected in function of their kinetic energy. However, since the kinetic energy depends on the radiation energy, the use of the binding energy abscissa is generally preferred. It's important to note that in PES experiments the energy level of the final state is measured, which is lacking one electron, while energy of the initial state can be only obtained by theoretical considerations.

Fig. 3. Sample levels scheme (right) and spectrometer levels (left). The sample and the spectrometer are in direct contact, thus their Fermi energies are equal. When photoelectrons go from the sample to the spectrometer they experience a potential due to the difference of the work functions of the two materials.

Fig. Scheme of the levels in the description of the shake up satellite peaks generation; the one electron process (left) and the two electron process (right).

In some cases the photoemission process involves also a transition of a core electron into the valence band. Thus, the final ionised atom is not in its ground state, and the kinetic energy of photon resulting from this two electrons process is lower than the expected one (fig. 4) and also depends on the energy of the excited state. It can be calculated as respect to the Exp of the one electron process as:

$$E\_{Kin}' = E\_{Kin} - \Delta E \quad \text{(15 )}$$

which corresponds in binding energy:

Ep = EB + ΔE (16)

where ΔE is the energy involved in the transition of the core electron to the valence band. This process leads to discrete structures at higher binding energy side of the main photoelectron peak (shake-up satellite(s)).

Fig. 5: The photon (hu) excites the electrons of the sample which are emitted as photoelectrons. The photoelectrons are then separated in energy by means of a hemispherical analyser.

Strong satellite peaks are generally observed in many transition metal and rare earth based compounds. In particular in the XPS spectrum of compounds containing Fe3+ ions show strong shake-up satellites (see section 4 paragraphs 3 and 4) and this is an additional feature that can be used to recognise it. Moreover, emitted photoelectrons can leave the sample's surface directly or they can interact with the atoms of the sample before escaping from the surface. The primary spectrum is formed by photoelectrons which leave the surface without exchanging energy with the surrounding atoms. The secondary spectrum is formed by the photoelectrons which have exchanged energy by means of inelastic scattering processes and it contributes to form a background. The latter ones are labelled as secondary electrons and their energy cannot be calculated with equation 14. The scattering processes limit the escape length of the emitted photoelectrons giving rise to one of the more interesting properties of the photoelectron spectroscopy. Photoelectrons with energy >50 eV allow to investigate a thickness of a maximum 50 Å. This feature makes the technique surface sensitive. The short escape length is also the reason for the necessity to work in UHV conditions. The surface sensitivity of the PES technique requires an atomically clean surface. Thus 10-10 mbar base pressure is recommended for PES investigations. A schematic representation of the PES experiment is reported in fig. 5.

According to the radiation energy used the photoelectron spectroscopies are classified as X-ray Photoemission Spectroscopy (XPS) and Ultraviolet Photoemission Spectroscopy (UPS).

### 4.1 X-ray Photoemission Spectroscopy (XPS)

In a XPS experiment the incident photons have energies 100 < hv < 1500 eV; the most common X-ray sources used in a conventional laboratory are the Al Kα (1486.6 eV) and Mg Kα (1253.6 eV). The photon energy is thus enough to promote core electrons into the vacuum continuous level of the investigated samples. As already mentioned, the binding energy of the photoelectrons can be used as fingerprint of the element allowing a chemical analysis of the surface. Each peak is labelled with the corresponding level of the initial state. The photoemission peak associated with an electron ejected from an orbital characterised by l > 0 can show a well-defined splitting in two components. The two states derive from the spin-orbit coupling. Accounting for the spin-orbit coupling with the j-j coupling scheme the total angular momentum of each electron is given by j = l + s and the total angular momentum of the whole atom is calculated as J = ∑j. Two possible states arise when an electron is ejected from an orbital characterised by l > 0. The difference in energy of the two states is due to the parallel or antiparallel alignment of the spin electron and its orbital angular momentum i.e. j = l ± 1/2. This separation for the core shells can be many electron-volts and is another fingerprint of a specific element. The relative intensity of the peaks is given by the ratio of their respective degeneracies (2j + 1). An example of the spin-orbit splitting is reported in fig. 6 for the Au 4f peak.

The binding energy of the peaks depends also on the electronic density of each atom that is related to the chemical environment, i.e. oxidation state and functional groups present (chemical shift). For instance the Fe2+, Fe++ and Feº 2p peaks have not the same binding energy: the core electrons of atoms with low electron density (like Fe3+) feel stronger Coulomb interaction respect to a more reduced species (like Fe2+), therefore the energy required for the ionisation process is shifted to higher values. This phenomenon allows for discriminating the oxidation state of each element on the surface. In order to compare the intensity of different XPS peaks the relative atomic photoemission cross-section (o) must be taken into account. This parameter depends on the atomic initial state and on the photon energy. A complete table of the element photoemission cross-section can be obtained from the web address: http://ulisse.elettra.trieste.it/services/elements/WebElements.html. (Yeh & Lindau 1985; Yeh 1993). The tabulated values have been used in this work for the quantitative interpretation of the XPS spectra. The intensity of each peak can be evaluated by integrating the corresponding area after the subtraction of the inelastic background. This process can be performed by using one of the many available programs which allow the deconvolution of the XPS signal and the background subtraction. The calculated area is then divided by its relative cross section allowing the comparison be-

 

$$\frac{I\_A'}{I\_B'} = \frac{N\_A \left\{ 1 - \exp\left[ {-d}\_{\lambda\_{A,\mathcal{A}}} \cos \theta \right] \right\}}{N\_B \left\{ \exp\left[ {-d}\_{\lambda\_{B,\mathcal{A}}} \cos \theta \right] \right\}} \quad \text{(17 )}$$

 A ~ = d

$$\mathbf{d} = \lambda \cos \theta \ln \left( \frac{\mathbf{l}\_{\mathbf{A}}^{\prime} \mathbf{N}\_{\mathbf{B}}}{\mathbf{l}\_{\mathbf{B}}^{\prime} \mathbf{N}\_{\mathbf{A}}} + 1 \right) = \lambda \cos \theta \ln \left( \frac{\mathbf{l}\_{\mathbf{A}}^{\prime} \mathbf{l}\_{\mathbf{B} \infty}}{\mathbf{l}\_{\mathbf{B}}^{\prime} \mathbf{l}\_{\mathbf{A} \infty}} + 1 \right) \tag{18.1}$$

IA∞ IB∞ d NB⁄NA IB∞⁄IA∞ IB∞ IA∞ IA ′ IB ′ ͳͺ d

~ =

 → → →

emission of an electron from the level 2p1/2,3/2 (L2,3). The energy of the emitted Auger electrons is not correlated to the X-ray excitation energy and it depends only on the energy levels of the system. In some cases the emitted Auger electrons of an element present on the surface sample have the same energy of the XPS peak of another element. In such a situation a correct evaluation of the XPS signal is hamper, as in the case presented in sections 4 paragraphs 5 and 6.

### 4.2 Ultraviolet Photoemission Spectroscopy (UPS)

In the UPS technique the incident photons have energies 10 < hv < 50 which are enough to remove the electrons from the valence band. The most common laboratory source for ultraviolet photoemission spectroscopy is the helium discharge lamp, which can provide photons with energy of 21.2 eV He(I) and 40.8 eV He(II). UPS, allowing the study of the valence band, provides useful information about the electronic structure of the surface and of molecular films. By comparison of the UPS spectrum with the calculated density of state (DOS) of the isolated molecule it is possible to evaluate the intactness of the molecules and their interaction with the substrate. An example of UPS spectrum acquired on the clean Au(111) surface is reported in fig. 9.

Fig. 9: UPS spectrum of the Au(111) valence band; the zero of the energy scale is set to the Fermi energy of the gold.

### 5 X-ray Absorption Spectroscopy (XAS)

The synchrotron based X-ray absorption spectroscopy is a fundamental tool for the magnetic and structural investigation of thin films (Wende 2004). In fact, the magnetic characterisation of monolayer and submonolayer molecular films cannot be carried out using the standard magnetometry techniques. These are not able to measure the signal of such small amounts of magnetic materials (Cornia et al. 2011).


sertion devices are divided in two classes the wigglers and the undulators The main differences are related to the range of photon energies emitted and the collimation of the light. The wigglers produce low collimated high energy photon (10-20 KeV) spread on a wide range of energies. The undulators generate high collimated energy photon with a narrow spectral emission range. The maximum of the spectral emission can be tuned by changing the distance of the two rows of magnets above and below the electrons path (fig. 10). In both insertion devices the emitted photons are linearly polarised. However, by using two or more set of undulators it is also possible to obtain the circularly polarised light needed for the XMCD measurements.

Fig. 10: Schematic representation of an insertion device. The grey and black elements are permanent magnets aligned in two rows, one above and one below the electron beam. In the undulator the gap distance between the two rows can be modified to tune the spectral emission.

### 5.2 X ray absorption spectroscopy principles

The simplest way to describe X ray absorption process is the so called one electron approximation (Stohr & Siegmann 2006). In this description the photon transfers its energy to a single core electron which is promoted from the core level to the valence band neglecting all the other electrons in the process. This view is clearly an oversimplification of the physics involved in the process and it does not allow a quantitative interpretation of the spectra. A more correct way to describe the phenomenon is the configuration picture. In this interpretation an atom is excited from its initial to the final excited configuration. Both configurations are described by theirs total angular momentum () which can be calculated with the (j-) coupling scheme, as described in section 2 paragraph 4 1, or by the so called Russell Saunders (RS) coupling:

$$L = \Sigma \, l\_i \qquad ; \qquad \mathcal{S} = \Sigma \, \mathbf{s}\_i \qquad \qquad ; \qquad \qquad f = L + \mathcal{S}$$

where L and S are the total angular orbital momentum and the total spin angular momentum respectively, the and s are the angular and spin momentum of the single electrons, respectively. This description allows to take in account the angular


$$T\_{if} = \frac{2\pi}{\hbar} \left| \langle f | H\_{int} | i \rangle + \sum\_{n} \frac{\langle f | H\_{int} | n \rangle \langle n | H\_{int} | i \rangle}{\varepsilon\_{i} - \varepsilon\_{n}} \right|^{2} \rho \left( \varepsilon\_{f} \right) d(\varepsilon\_{i} - \varepsilon\_{f}) \tag{19}$$

, ) = ( ⁄ ) ⃗ ⃗ ⃗⃗ = − ⃗ ⃗⃗⃗⃗⃗⁄ 0

$$
\sigma = \frac{r\_{if}}{\phi\_0} \tag{20.1}
$$

→

→

<sup>+</sup> <sup>−</sup> <sup>−</sup> − <sup>+</sup>

+ −

$$X \\ MCD = \sigma^- - \sigma^+ \tag{21}$$

→ 

$$m\_{orb} = -\frac{4\int\_{L\_2 + L\_3} (\sigma^- - \sigma^+) d\omega}{3\int\_{L\_2 + L\_3} (\sigma^+ + \sigma^-) d\omega} \cdot (n\_{3d}) \quad \text{(22 )}$$

 <sup>−</sup> − <sup>+</sup> 3 

$$m\_{\rm spin} = \frac{6 \int\_{L\_3} (\sigma^- - \sigma^+) d\omega - 4 \int\_{L\_2 + L\_3} (\sigma^- - \sigma^+) d\omega}{\int\_{L\_2 + L\_3} (\sigma^+ + \sigma^-) d\omega} \cdot (n\_{3d}) \cdot \left(1 + \frac{7 \langle T\_x \rangle}{2 \langle S\_x \rangle} \right)^{-1} \tag{23.1}$$

⟨⟩ ⟨⟩ ⟨⟩ ⟨⟩⁄ (<sup>−</sup> + +)

$$m\_{orb} = \frac{-4 \mathbf{q} \cdot n\_{3d}}{2r} \tag{24 \text{ }}$$

$$m\_{spin} = \frac{-(6 \mathbf{p} - 4 \mathbf{q}) \cdot n\_{3d}}{6r} \tag{25 \text{ }}$$

Fig. 13: Integration of the cobalt XMCD spectrum (left) and the integration of the XAS spectrum calculated as (σ- + σ\*)/2 after the subtraction of the two steps background.

### 5.4 X-ray Natural Linear Dichroism (XNLD)

The XNLD experiment is defined as the difference between the XAS spectra acquired with vertical and horizontal polarised light. The XNLD technique provides information about the anisotropy of the charge distribution around the absorbing atom. This information can be useful for individuating preferential orientation in SMM hybrid surfaces (Cornia et al. 2011). A simple way to visualize the phenomenon originating the linear dichroism is to invoke the so called "search light effect" introduced by Stöhr "X-ray absorption is governed by electric dipole transition and the photoelectrons are preferentially excited into the direction of the electric field" (Stöhr 1995). The dichroism derives from the anisotropy in the valence level of the investigated atom. Considering a transition 2p -> 3d its intensity is related to the sum over all the degenerate 2p states and all the 3d states. The core levels leads to a spherical symmetry contribution. The 3d orbital contribution is strictly related, in mononuclear molecular complexes, to the symmetry of the complex. In cubic symmetry the 3d degeneracy is broken by the ligand field obtaining the t20 and exirreducible representations. The sum of each representation gives rise to a spherical contribution. In this situation no dependence on the photon electrical field orientation has to be expected. However lowering the symmetry of the complex the contribution is no more spherically symmetric and a dependence on the electrical field orientation is found. The intensity of the transition is then at its maximum when the electric field points to the empty states and it will be zero on nodal plane. This allows to extract information about the local symmetry of the absorber.

### 5.5 X-ray absorption detection

As reported in the previous sections the XAS absorption leads to a core hole final state. This state is not stable and it evolves relaxing in two principal ways: by the

emission of Auger electrons or by emission of fluorescence photons. Both emissions are related to the probability of the XAS absorption and thus they can be measured in function of the photon energy to provide the XAS spectrum. Different detection methods can be employed for the two emissions. In this work the Total Electron Yield (TEY) mode was used.

The Auger emission leads to ionised atoms. If the sample is grounded an electron current flows to the sample to restore the neutrality. This current can be read by means of a picoammeter (fig. 14). Since the process is related to the emission of the Auger electron the TEY mode is surface sensitive. In fact the escape length of the Auger electrons is ~2 nm and no information on processes occurring deeper in the sample are accessible. This detection method allows the study of the superficial properties of the samples.

Fig. 14: Schematic representation of the Total Electron Y eld (TEY) detection system.

### 6 DEIMOS beamline

Most of the XAS experiments carried out during this thesis work have been performed on the DEIMOS beamline, which will be shortly described here. DEIMOS beamline experimental setup was designed to perform XAS and its derivate techniques. The beamline's photon energy range is 250-2500 eV. The insertion device allows the study with linearly and circular polarised light. A cryo-magnet allows measuring the XAS at temperature as low as 1.5 K with tuneable magnetic field along the photon direction up to 70 kOe.

Two preparation chambers connected to the beamline allows the in situ preparation of the sample (fig. 15). In particular the MBE chamber was dedicated to the surface preparation and provides sputtering/annealing and metal evaporation facilities. The other (RAOUL) was employed for the molecular sublimation. The UHV system offers also the possibility to characterise the investigated surfaces by STM, Low Electron Energy Diffraction (LEED) and Auger Electron Spectroscopy (AES).

In Our experiments (see section 3 paragraph 4) a EFM3 Omicron evaporator was assembled in the MBE chamber for the evaporation of the cobalt while our home-made molecules sublimation system was installed in the RAOUL chamber.

 Ͳ 

Ͳ θ 

$$E\_i/E\_0 = \{ \pm [(M\_t^2 - M\_1^2 \sin^2 \theta)^{1/2} + M\_i \cos \theta]/(M\_1 + M\_t) \}^2 \tag{26.1}$$

0 < 1000

()

$$\left[-\frac{\hbar^{\mathbb{I}}}{2m}\frac{d^{\mathbb{I}}}{dz^{\mathbb{I}}} + V(\mathbf{z})\right]\psi(\mathbf{z}) = E\psi(\mathbf{z})\qquad(2.277)$$

() = 0 () =

() = (0) ሺʹͺሻ

$$k = \sqrt[2]{\frac{2mE}{\hbar^1}}\qquad\qquad\qquad\text{(29 )}$$

$$
\psi(\mathbf{z}) = \psi(0)e^{\cdot \mathbf{k} \cdot \mathbf{z}} \qquad \text{(30 }).
$$

$$k = \sqrt[2^2]{\frac{2m(V\_\circ - E)}{\hbar^2}} \qquad E < V\_0 \tag{31}$$


$$I = \frac{2\pi e}{\hbar} \Sigma\_{\mu,\nu} \{ f(E\_{\mathbb{T}\mu})[1 - f(E\_{\mathbb{S},\mathbb{V}} + \mathbf{eV})] - f(E\_{\mathbb{S},\mathbb{V}} + \mathbf{eV})[1 - f(E\_{\mathbb{T}\mu})] \} \cdot |\mathcal{M}\_{\mu,\nu}| \, \tag{32}$$
  $\delta(E\_{\mathbb{T}\mu} - E\_{\mathbb{S},\mathbb{V}}) \tag{32}$ 

 − ) 

$$M\_{\mu\upsilon} = -\frac{\hbar^2}{2m\_\upsilon} \int \left[ (\psi\_\mu)^\star \vec{\mathcal{V}} \psi\_\upsilon - \,\psi\_\upsilon \vec{\mathcal{V}} (\psi\_\mu)^\star \right] d\vec{s} \tag{2.33 \,\,\,} $$

$$I \ll \mathcal{V} \cdot \rho\_{\Gamma}(E\_{\mathcal{F}}) \cdot \rho\_{\mathcal{S}}\left(E\_{\mathcal{F}}, \vec{r}\_{0}\right) \cdot e^{2k\mathcal{R}} \tag{34}$$

 ( ) ( , ) √(2 ) <sup>2</sup> ⁄ħ

$$
\rho\_{\mathbb{S}}(E\_{\mathbb{H}}, \vec{r}\_0) = \sum\_{\circ} |\psi\_{\circ}(\overrightarrow{r\_0})|^2 |\,\delta(E\_{\mathbb{V}} - E\_{\mathbb{F}}) \tag{35.1}
$$


͵Ͷ ͵ͷ

ing constant current mode and small bias, the islands are showed as depression (fig. 20). A full monolayer of Cu2N grown on Cu(100) will be used as substrate for the deposition of the Fe4Ph molecule in section 4 .

### 8.3 Scanning Tunnelling Spectroscopy (STS)

It has been demonstrated that the STM technique is sensitive to the local density of states. This allows to get information on the LDOS at atomic scale. In order to achieve information on the energy dependence of the LDOS the scanning tunnelling spectroscopy can be employed. In this measure the tip is positioned on top the interesting point and the feedback loop is turned off. A ramp of potential is applied on the tunnelling junction while measuring the tunnelling current. In this way a single point scanning tunnelling spectroscopy is recorded. The data obtained will be the convolution of the LDOS of the tip and the sample. Assuming a flat LDOS for the tip the current is described as:

I x LDOSsampleV (36)

Thus the d(1)/d(V) curves provide direct information of the LDOS of the sample. In order to be able to measure the d(I)/d(V) a lock-in detection of an alternate bias is commonly employed. The alternate bias is superimposed to the potential ramp and the lock-in allows to discard the capacitive current signal. In this way it is possible to record the d(I)/d(V) spectra directly. In order to take in account the variation of the transmission coefficient respect to the applied bias a further normalization of the data is generally employed (Selloni et al. 1985). This dependence can be eliminated by using the (d(I)/d(V))/(I/V) as proposed by Stroscio and Feenstra (Stroscio et al. 1986; Feenstra & Mårtensson 1988).

It's important to note that the STS experiment allows a direct measure of the conductance of the investigated point. This is crucial point for the study of inelastic tunnelling process (see section 4 paragraph 7).

### 9 Multiplatform Ultra-High Vacuum system

In this work mostly of the surface characterisations were carried out exploiting a multiplatform UHV system at the Center for Scanning Probe Characterisation Techniques (CeTeCS) of the Dept. of Chemistry at Florence University (fig. 21). The system offers the possibility to prepare the samples and their investigation in situ employing XPS, UPS, LEIS and STM techniques.

# Section 3 The terbium bis(phthalocyaninato) complex

### 1 ntroduction

Terbium bis(phthalocyaninato) complex (TbPc2) is one of the most investigated SMM (Ishikawa et al. 2003). Due to its magnetic properties the TbPc, molecule, and its derivatives, is a promising candidates for spintronic purposes (Bogani & Wernsdorfer 2008; Urdampilleta et al. 2011; Katoh, Isshiki, et al. 2012). The TbPc. molecule belongs to the metal double decker class in which a Tb+ ion is bound in the centre of two phthalocyanine molecules staggered 45° by each other, as show in fig. 22.

Fig. 22 Terbium bis(phtahalocyaninato) complex (TbPc2) structure.

TbPc can be easily synthesised as neutral [TbPc] and cationic [TbPc] | powder, as described in the next paragraph (Katoh et al. 2009; De Cian et al. 1985) By chemical oxidation of [TbPc2] (Takamatsu et al. 2007) or by electrochemical synthesis (Gonidec et al. 2010; Zhu et al. 2004) is also possible to achieve the [TbPc2] + compound which is not stable, be aving as strong oxidative molecule. It is important to note that the oxidation number of the metal ion does not change from the oxidised [TbPc] to the reduced [TbPc] molecule because the oxidation/reduction process involves only the electronic structure of the ligands (Zhu et al. 2004).

±

### 

′′

$$\chi\_{\prime\prime}^{\prime\prime} = (\chi\_{\,T} - \chi\_{\,S}) \frac{(\omega\tau)^{1-a} \cos(\pi a/2)}{1 + 2(\omega\tau)^{1-a} \sin\left(\pi \frac{a}{2}\right) + (\omega\tau)^{2-2a}} \tag{37.1}$$

 

 


$$\chi = \frac{c}{r - \theta} \tag{38}$$

 

1 2⁄ (2) ⁄

$$
\mu^+ \to e^+ + \upsilon\_e + \bar{\upsilon}\_\mu \qquad\qquad(39)
$$

this decay includes the emission of an electron neutrinos Up, a muon antineutrinos Uu and a positron e +.

Positrons are preferentially emitted along the muons' spin polarisation at the moment of the decay (fig. 37). In fact, the probability of the positrons emission in a direction forming a 0 angle with the muon polarisation can be calculate as:

$$W^{+}(\theta,\varepsilon) = 1 + A^{+}(\varepsilon)\cos\theta \qquad (40)$$

where & is the energy of the positron and A+ is the asymmetry. Thus, counting the spatial distribution of the emitted positrons allows to follow the time evolution of the muons polarisation. From this data is thus possible to get information on the local field experienced by the muons.

In these experiments, fully spin-polarised muons are implanted into the sample and used as a local probe to detect dipolar fields from the surrounding molecules. They provide the direct observation of the spin dynamics of individual SMMs.

Fig. 37: Preferential emission of the positron along the polarisation axis of the muon.

In this study three samples were investigated: (1) a microcrystalline powder sample of TbPc2^CH2Cl2, (2) a thick (ca. 1 um) and (3) a thinner film sample (ca. 100 nm). The films were evaporated onto 200 nm polycrystalline gold films grown on freshly cleaved Muscovite mica substrates. The thickness of the film was estimated by Atomic Force Microscopy (AFM) with the standard scratch method and crosschecked by magnetometry measurements. The sublimation rate was estimate using the following procedure. A TbPc2 film was sublimated on quartz for at list one hour. The sample was then undergone to scratch and AFM investigations. The scratch procedure was carried out by using a thin needle in order to selectively remove in the scratched area (ca 40um per several mm) the TbPc2 deposit from the quartz substrate. The AFM was then used to trace the step profile and get the thickness information. Thus the thickness allows the calculation of the deposition rate. It's important to note that for film deposited on Au(111) grown on mica this procedure cannot be employed because the substrate is not hard enough and it can be damaged

 

$$Mass = \frac{M\_{sample}}{M} \cdot M\_W \tag{41}$$

 ⁄ ) ⁄ ≪ ≫

$$A(t) = \frac{A\_0}{3} \left[ 1 + 2(1 - \gamma \delta)e^{-\gamma \delta t} \right] e^{-\sqrt{\lambda t}} \tag{42}$$

where Ao is the initial asymmetry and is the relaxation rate, which contains information regarding the dynamics of the local field. The square root relaxation reflects the averaging of the relaxation behaviour of muons stopping in many nonequivalent sites (Uemura et al. 1985; Lascialfari et al. 1998; Salman et al. 2002; Blundell et al. 2003; Branzoli et al. 2009; Branzoli et al. 2010). By fitting the data with the above mentioned equation the parameters and o were obtained and their temperature dependence is shown fig. 39

Fig. 38 Typical muon spin relaxation curves in the (a) bulk and (b) thick film samples measured in zero applied field and at various temperatures. The insets show the early time relaxation, where the dip in the relaxation can be clearly seen at low temperatures. The lines are fits to equation 42. Reprinted with the permission from (Hofmann et al. 2012) Copyright (2012) American Chemical Society.

It is important to note that the results obtained for the microcrystalline powder are consistent with previous muon spin relaxation measurements (Branzoli et al. 2010). The qualitative similarity between bulk and films behaviour is an indication that SMM nature of the TbPc2 in the film is retained. The static and the dynamic behaviour in all the three samples are in fact characterised by three different temperature regimes. At high temperature is small and o is ca. zero. As the temperature is decreased increases sharply while o remains zero. At ca. 100 K, 1 peaks and o be-

 

 ± ± 

small difference in the value between the microcrystalline powder and the films in this temperature regime implies that the energy gap between the ground and the first excited state does not change significantly. This is a direct evidence that there is almost no change in the crystal field of the Tb ions between microcrystalline powder and films. In contrast to this behaviour, the saturation of at low temperature is a consequence of persistent spin dynamics at temperatures far below the energy gap. This is attributed to quantum tunnelling between the two quasi-degenerate J = ±6 ground states (Branzoli et al. 2009; Branzoli et al. 2010), which is particularly eff cient in zero applied magnetic field, corresponding to the condition of our experment. Interestingly, in fig. 39 one can observe a clear difference in the saturation value of depending on the sample: bulk, thick or 100 nm film. Moreover, in the 100 nm film it also depends on the muons' implantation energy/depth.

Fig. 40 Correlation time as a function of temperature measured in bulk and TbPc films. The lines are fits to equation 44. Reprinted with the permission from (Hofmann et al. 2012) Cooyright (2012) American Chemical Society.

In what follows we extract the correlation time of the molecular spin dynamics as a function of temperature. At low temperatures where vo>>,A(t) is almost identical to the dynamic Lorentzian Kubo-Toyabe function, and hence 2 =(2/3t) (Uemura et al. 1985; Hayano et al. 1979), where is the correlation time of the local magnetic field experienced by the muon, which in our case is that generated by the TbPc SMMs. However, at high temperatures, where is ca. 0, the relaxation rate can be written as A = 2 (yoo)2 (Uemura et al. 1985; Hayano et al. 1979), where o jis the size of the fluctuating field at the measured temperature. Note that the SMMs are in their ground J = 6 manifold throughout the measured temperature range. Therefore, do can be evaluated from the low temperature saturation value, 00 = 0(T > 0), which reflects the size of the dipolar field from the magnetic moment of a single TbPc2 molecule. Thus, we can readily extract as a function of temperature in the high and low-temperature ranges, as shown in fig. 40.

± ±

$$\frac{1}{\tau\_{sp}} = \mathcal{C} \Delta^3 \exp\left(\frac{\Delta}{\tau}\right) \tag{43.1}$$

 ∆ ∆ 

$$\frac{1}{\tau} = \frac{1}{\tau\_{sp}} + \frac{1}{\tau\_q} \tag{44}$$

∆ ∆ 


() 

 ± ∼ ∼ ∼

 

∼

gradual variation of the molecular packing from the lie down geometry in proximity of the metal substrate to the standing up characterising the top layers of the thick film. It is important to note that this interpretation is in agreement with what shown in the previous paragraph about pristine and heated powder.

### 4 Toward TbPc2 spintronic devices

### 4.1 Introduction

TbPc2, as well as all SMMs, have been a workbench for the study of quantum effects in magnetism (Gatteschi et al. 2006); more recently these systems have attracted a growing interest as active elements in organic spintronic devices (OSPDs) (Dediu et al. 2009). These generally consist of a semiconductor organic film located between two ferromagnetic electrodes, one acting as spin injector and the other as spin analyser (Dediu et al. 2009) (fig. 43).

Fig. 43: Schematic representation of a vertical spin valve device. The organic film (OF) is stratified between the two ferromagnetic electrodes of cobalt (spin injector) and Lanthanum Strontium Manganite Oxide (LSMO)(spin analyser).

Recent studies, employing tris(8-hydroxyquinolinato) aluminium (Alq3) organic film, have shown how OSPDs can be designed to behave as spin valves or "memristor" (Prezioso et al. 2013) where the non-volatile resistence of the device can be tuned and depends on the hystoric current and bias voltage applied. Recent electric transport studies involving SMMs are based on single molecules transistor (Heersche et al. 2006; Jo et al. 2006) where the molecule is placed between two nonmagnetic electrodes. On the other hand Urdampilleta et al. have recently obtained a supramolecular spin valve (Urdampilleta et al. 2011) employing TbPc2 molecules deposited on a carbon nanotube. At best of our knowledge, SMMs have not yet been employed as organic film in macroscopic spin valve devices.

In OSPDs the molecules-electrode interface plays a key role to define the final device properties (Sanvito 2011). Therefore, the knowledge of the magnetic

$$d = \lambda \cos \theta \ln \left( \frac{l\_A' \sigma\_B}{l\_B' \sigma\_A} + 1 \right) \quad \text{(45 )}$$

where is the escape length of the photoelectrons, I , are the areas and oA op are the cross section of the Cu 2p3/2 and Co 2p3/2 peaks, respectively. The calcular ed thickness of the 42 min evaporated cobalt film was 2.1 ML. Same investigations were repeated increasing the evaporation time obtaining the calibration curves reported in fig. 51c. The data could be fitted by a linear relation, as expected for constant Co evaporation flux.

Before the XPS investigation the 2.1 ML cobalt film sample was also character ised by STM. In fig. 52 are reported the STM images for two different nominal thicknesses, 2.1 and 4 ML of Co, as estimated from the previous calibration. Both samples show a quasi layer-by-layer growth (Allmers & Donath 2011) with the formation of small cobalt islands. They present rounded edges preferentially oriented along the Cu [±110] directions; moreover the cobalt films do not exactly reproduce the underling Cu atomic step edges looking meandering (Ramsperger et a 1996). All these features confirm the good quality of the cobalt films.

It is well known that in thin films an intermixing of the two metals can take place even at room temperature due to an atomic exchange process 26,28]. This phenomenon could be responsible of the formation of a layer of copper on top of the cobalt film.

Fig. 52 STM image of 2.1ML (400 pA; -0.4 V ; 75x75 nm2) (a) and 4ML of Co on Cu(100) (900 pA; 1.4 V; 100x100 nm2) (b). They show the typical cobalt rounded edge islands preferentially oriented along the Cu [±110] directions.

In order to take in account this phenomenon the LEIS techniques, sensitive to the top-most layer chemical composition, was employed (see section 2 paragraph 7). The LEIS spectra, presented in fig. 53, were acquired using a He beam accelerated to 1000 eV on samples with increasing thickness of the Co film. The data are relative to

m m m m

m m

Section 4 The Fe4 SMM class

### 1 ntroduction

The slow magnetic relaxation properties of the family of Fe4 molecules, conjugated with their chemical stability, made them one of the most interesting SMMs. Despite the low blocking temperature (hysteresis opening only below 1 K) Fe, mole cules have been used as model systems to shed some light on the complex and fasc nating behaviour of SMMs (Gatteschi et al. 2006). Fe SMMs have the propeller like shape shown in fig. 62 In the archetypal Fe system the three peripheral iron(III) ions are coordinated by two diketonate ligands, i.e. dipivaloylmethanate (dpm), while six -- methoxide (OMe) anions bridge them to the central iron(III) ion to give the final formula [Fe (OMe) (dpm)o] (Barra et al. 1999) The molecule is character ised by the presence of a C symmetry axis passing through Fel and Fe2, thus the peripheral ions are arranged at the vertices of an isosceles triangle (Barra et al. 1999) The antiferro nagnetic coupling of the central ion (s=>/2) to the peripheral ones (s=3x5/2) causes the total spin of the ground state to be S=5. The ground state can be considered selectively populated for temperature below 5 K.

Fig. 62. Molecular structure of the archetypal Fe system, the [Fea(OMe)s(dpm)}]

$$H\_{\rm zfd} = D \left[ \mathcal{S}\_{\rm z}^2 - \frac{\imath}{3} \mathcal{S} (\mathcal{S} + 1) \right] \tag{46}$$

̂ ̂ 

∆ =

$$Fe\% = \frac{A\_{Fa}/\sigma\_{Fe}}{A\_{Fe}/\sigma\_{Fe} + A\_C/\sigma\_C + A\_O/\sigma\_O} \tag{47}$$

$$C\% = \frac{A\_{\mathcal{C}} / \sigma\_{\mathcal{C}}}{A\_{\text{Fe}} / \sigma\_{\text{Fe}} + A\_{\mathcal{C}} / \sigma\_{\mathcal{C}} + A\_{\mathcal{O}} / \sigma\_{\mathcal{O}}} \tag{48}$$

$$O\% = \frac{A\_O/\sigma\_O}{A\_{Fe}/\sigma\_{Fe} + A\_C/\sigma\_C + A\_O/\sigma\_O} \tag{49}$$

Δ


(⁄) (⁄⁄ ) 

(⁄) (⁄⁄ )


(⁄) (⁄⁄ )

Mannini et al. 2008) it is possible to hypothesize different origins for the decomposition process. It could be generated during the sublimation process or simply be due to a surface-mediated fragmentation of the cluster.

Before trying to improve the morphological quality of the images of the Bdomains, a test was performed in order to have some insight on the decomposition process.

### 4.1 Back exposure sublimation test

In order to shed more light on the process involved in the formation of A domains, a simple test was performed by using the sublimation geometry shown in fig. 72. The clean Au(111) crystal was kept in the preparation chamber for 80 minutes with the crucible at the sublimation temperature (215 ℃ but without exposing the sample to the direct flux of the molecules. The surface was then investigated by STM. In large scale images (fig. 73a) the surface appears covered by a densely packed layer through which the corrugation of the herring bone reconstruction of the Au(111) surface can be still detected.

Fig. 72 Back exposure geometry. The sample is not directly exposed to the molecular flux coming from the crucible.

A close inspection by STM of the surface (see the enlargement in fig. 73b) reveals a textured layer of small objects similar to that observed in the A- ype domains of the 0.25 nm Fe. Ph sample (see fig. 71c). No trace of the B-type (FeyPh) islands is revealed. This observation suggests the presence of very volatile species that can be stuck on the surface even if the sample is not directly exposed to the flux coming from the crucible. The probability that Fe Ph molecule, as big as they are, can bounce on the chamber walls then be able to reach the sample surface is expected to be lower than for smaller, more volatile fragments.

A partial decomposition of the Fe, molecules during the sublimation process seems more likely than a surface induced fragmentation. By looking at the shape of

the A objects we can speculate that they are formed by dpm-containing species. Images with objects with similar shape have been indeed reported in the study of the dissociation of a chromium tris-diketonato complex on Cu(100) surface (Grillo et al. 2002)

Fig. 73 Large scale STM image of the 80 min back exposed sample (3 pA; 2 V; 110x110 nm') (a), in the enlargement (3 pA; 2 V; 15x7.5 nm2) (b) the textured layer formed by A-type mo ecule.

### 5 STM investigation on Cu(100) surface

Assuming that only domains of type B are compatible with intact Fe. Ph molecules, it is quite evident that the molecules tend to aggregate into islands, hampering the collection of high resolution images. It would be therefore advisable to reduce the mobility of the molecules. This could be achieved simply sublimating the molecules on cooled surfaces. Unfortunately our UHV system does not allow such a procedure. An alternative way is the deposition on more reactive surfaces like Cu(100). This surface is more reactive than the Au(111) one and could allow the onset of stronger molecule-surface interactions, thus reducing molecules mobility.

Again a sub-monolayer sample with similar coverage as the previous one, 0.25 nm, was prepared. The sample was cooled down to 35 K before being investigated. The STM image reported in fig. 74 clearly shows islands formed by well-defined spherical objects and, underneath, a wetting layer arranged in dendritic structures. As in the case of Au(11), the height of the islands measures 0.72±0.02 nm while the dendritic structures are height 0.19±0.01 nm (see profile of fig. 74b). The two structures were labelled as A' (0.19 nm height) and B' (0.72 nm height) in analogy to the A and B islands on Au(111) surface. It is important to point out that the B' height is referred to the A' under (fig 74b) while in the Au(111) surface is referred to the bare gold. Among the A' layer spherical objects (labelled as C') are present with a

'' ''



Δ ± Δ ±


1st Born approximation. The transition intensities were computed using an interaction Hamiltonian:

H = ¿ · Š + u ( 50)

where o is the electron spin, S is the spin of the Fe atoms and u is a spin independent scattering term (Loth, Lutz, et al. 2010). This operator acts on the product states of the tunneling electron and molecule spin. In this way the selection rules are imposed by energy and angular momentum conservation; spin-dependent conduction channels can be accounted for.

Although only a preliminary characterisation of the spin excitation of FeaPh molecules through IESTS has been possible during my stay in Hamburg, the results are rather encouraging, confirming that objects consistent with intact and isolated Fe4 molecules can be detected and addressed at the Cu2N surface. Moreover, the observation of spectroscopic features in agreement with the ones predicted assuming the Spin Hamiltonian parameters of Fe molecules in the crystalline phase suggests that the magnetic features of this SMM are very robust. These novel results corroborate the observation of magnetic bistability in self-assembled monolayers by XMCD, or the detection of spectroscopic features in the conductivity inside nano-gap electrodes (Heersche et al. 2006; Zyazin et al. 2010; Zyazin et al. 2011) and open the fascinating perspective of detecting the magnetic bistability, and related memory effect, at the single molecule level by using scanning-probe microscopy techniques.

# Section 5 Conclusion

In this work we have presented two molecular systems: the TbPC2 and the Fe4Ph SMMs. Both of them can be sublimated allowing the preparation of hybrid surfaces employing the ultimate cleanliness of the UHV environment.

TbPc2 is considered the archetypal of sublimable SMMs and the ideal candidate for investigations with scanning probe techniques. However, we have demonstrated that its magnetic properties are strongly influenced by the surrounding environment. We have indeed proven that its slow relaxation of the magnetization depends on the molecular packing. Despite the many experiments performed by us and others a rationalization of its properties and a full comprehension of the role played by the exchange interactions is lacking. Nevertheless, TbPc2, thanks to its relative high blocking temperature, is an interesting candidate for OSPDs. Moving toward this type of applications we have investigated the role played by the presence of a magnetic conducting substrate in the hysteretic behaviour of the TbPc2. Our study has concerned metallic cobalt and LSMO, two of the most employed ferromagnetic electrodes in molecular spin valves. The results we have presented here suggest that no significant polarisation of the TbPc2 magnet moments is induced by the magnetic Co and LSMO substrates. However, this does not diminish the interest in this type of molecules, and magneto-transport experiments on devices embedding TbPc2 molecules are likely to be the focus of this research in a near future.

On the other hand, Fe4Ph molecules show a more predictable magnetic behaviour and, albeit the low blocking temperature of the system that makes them less appealing for OSPDs, the investigation of its electron transport properties could allow to have a reliable picture of the role played by magnetic molecules at the interface between the ferromagnetic electrodes and the organic semiconductor. Within this framework we have exploited the sublimation of FeaPh in order to investigate the hybrid surfaces by means of STM and STS techniques.

Our characterisation has allowed to address the spin excitation signal of an isolated Fe2Ph molecule. This first important step opens new possibility for the investi-

### References



Eckstein, W., (1991). Computer Simulation of Ion-Solid Interactions. Springer: Berlin.



Kramers, H.A.Z., (1926) Phys., 39, 826-840.






## Acknowledgments

Ringrazio la Prof.ssa Sessoli per la grande passione che dedica alla ricerca e per la sua pazienza e disponibilità. Ringrazio i miei colleghi per il loro fondamentale contributo: Dr. Mannini, Dr.ssa Lanzilotto, Dr. Poggini, Cortigiani, Dr. Totti e Dr.ssa Ninova. Ringrazio il Prof. Caneschi per avermi dato la possibilità di far parte del LAboratorio di Magnetismo Molecolare (LAMM) e tutti coloro che del LAMM fanno parte.

I wish to thank Dr. Zaher and Dr. Hofmann of the Paul Scherrer Institute (PSI) for the muon spin relaxation measurements. I am also in debt to those who helped me during beamtimes at the SOLEIL synchrotron: Dr. Sainctavit, Dr. Ohresser, Dr. Otero and Dr. Choueikani. Thanks to Dr. Loth and Dr. Burgess (Max Planck Research Group - Dynamics of Nanoelectronic Systems in Hamburg) for the spin excitation measurements.

Ringrazio tutti gli amici che mi hanno supportato e sopportato in questi anni.

Infine un ringraziamento particolare a mia moglie e alla mia famiglia.

### premio tesi di dottorato

### anno 2007

Bracardi M., *La Materia e lo Spirito. Mario Ridolfi nel paesaggio umbro*


### anno 2008

Bemporad F., *Folding and Aggregation Studies in the Acylphosphatase-Like Family*


Colica G., *Use of Microorganisms in the Removal of Pollutants from the Wastewater*

Gabbiani C., *Proteins as Possible Targets for Antitumor Metal Complexes: Biophysical Studies of their Interactions*

### anno 2009

Decorosi F., *Studio di ceppi batterici per il biorisanamento di suoli contaminati da Cr(VI)*


Pace R., *Identità e diritti delle donne. Per una cittadinanza di genere nella formazione*

Vignolini S., *Sub-Wavelength Probing and Modification of Complex Photonic Structures*

### anno 2010

Fedi M., *«Tuo lumine». L'accademia dei Risvegliati e lo spettacolo a Pistoia tra Sei e Settecento*

Fondi M., *Bioinformatics of genome evolution: from ancestral to modern metabolism. Phylogenomics and comparative genomics to understand microbial evolution*

Marino E., *An Integrated Nonlinear Wind-Waves Model for Offshore Wind Turbines*

Orsi V., *Crisi e Rigenerazione nella valle dell'Alto Khabur (Siria). La produzione ceramica nel passaggio dal Bronzo Antico al Bronzo Medio*

Polito C., *Molecular imaging in Parkinson's disease*

Romano R., *Smart Skin Envelope. Integrazione architettonica di tecnologie dinamiche e innovative per il risparmio energetico*

### anno 2011

Acciaioli S., *Il* trompe-l'*œil letterario, ovvero il sorriso ironico nell'opera di Wilhelm Hauff*

Bernacchioni C., *Sfingolipidi bioattivi e loro ruolo nell'azione biologica di fattori di crescita e citochine*

Fabbri N., *Bragg spectroscopy of quantum gases: Exploring physics in one dimension*

Gordillo Hervás R., *La construcción religiosa de la Hélade imperial: El Panhelenion*

Mugelli C., *Indipendenza e professionalità del giudice in Cina*

Pollastri S., *Il ruolo di TAF12B e UVR3 nel ciclo circadiano dei vegetali*

Salizzoni E., *Paesaggi Protetti. Laboratori di sperimentazione per il paesaggio costiero euro-mediterraneo*

### anno 2012


### anno 2013


Cucinotta E., *Produzione poetica e storia nella prassi e nella teoria greca di età classica*

Pellegrini L., *Circostanze del reato: trasformazioni in atto e prospettive di riforma*

Locatelli M., *Mid infrared digital holography and terahertz imaging*

Muniz Miranda F., *Modelling of spectroscopic and structural properties using molecular dynamics*

Bacci M., *Coarse-grained molecular dynamics and continuum models for the transport of protein molecules*

Martelli R., *Characteristics of raw and cooked fillets in species of actual and potential interest for Italian aquaculture: rainbow trout (Oncorhynchus mykiss) and meagre (Argyrosomus regius)*

### anno 2014


Pastorelli G., *L'immagine del cane in Franz Kafka*

Bussoletti A., *L'età berlusconiana. Il centro-destra dai poli alla Casa della Libertà 1994-2001*

Malavolti L., *Single molecule magnets sublimated on conducting and magnetic substrates*

Belingardi C., *Comunanze urbane. Autogestione e cura dei luoghi*

Guzzo E., *Il tombeau di Jean-Jacques Rousseau al Pantheon. Dalla capanna rustica di Vitruvio alla sua fortuna iconografica nella trattatistica fra XV e XVIII Secolo*