Search

The local magnetic structure in the [FeIII (Tp)(CN)3 ]- building block was investigated by combining paramagnetic Nuclear Magnetic Resonance (pNMR) spectroscopy and polarized neutron diffraction (PND) with first-principle calculations. The use of the pNMR and PND experimental techniques revealed the extension of spin-density from the metal to the ligands, as well as the different spin mechanisms that take place in the cyanido ligands: Spin-polarization on the carbon atoms and spin-delocalization on the nitrogen atoms. The results of our combined density functional theory (DFT) and multireference calculations were found in good agreement with the PND results and the experimental NMR chemical shifts. Moreover, the ab-initio calculations allowed us to connect the experimental spin-density map characterized by PND and the suggested distribution of the spin-density on the ligands observed by NMR spectroscopy. Interestingly, significant differences were observed between the pseudo-contact contributions of the chemical shifts obtained by theoretical calculations and the values derived from NMR spectroscopy using a simple point-dipole model. These discrepancies underline the limitation of the point-dipole model and the need for more elaborate approaches to break down the experimental pNMR chemical shifts into contact and pseudo-contact contributions.

A new crystallographic method is proposed in order to refine a spin-resolved atomic orbital model against X-ray and polarized neutron diffraction data. This atomic orbital model is applied to the YTiO3 perovskite crystal, where orbital ordering has previously been observed by several techniques: X-ray diffraction, polarized neutron diffraction and nuclear magnetic resonance. This method gives the radial extension, orientation and population of outer atomic orbitals for each atom. The interaction term between Ti3+, Y3+ cations and O2− ligands has been estimated. The refinement statistics obtained by means of the orbital method are compared with those obtained by the multipole model previously published.

The anisotropy of the magnetic properties of molecular magnets is a key descriptor in the search for improved magnets. Herein, we prove how a novel analysis approach using single crystal polarized neutron diffraction (PND) provides direct access to atomic magnetic susceptibility tensors. The technique has been applied for the first time to two Dy-based single molecule magnets, showing clear axial atomic susceptibility for both Dy(III)-ions. For the triclinic system, bulk magnetization methods are not symmetry-restricted, and the experimental magnetic easy axes from both PND, angular-resolved magnetometry (ARM) and theoretical approaches all match reasonably well. ARM curves simulated from the molecular susceptibility tensor determined with PND show strong resemblance with the experimental ones. For the monoclinic compound, a comparison can only be made with the theoretically calculated magnetic anisotropy, and in this case PND yields an easy-axis direction that matches that predicted via electrostatic methods. Importantly, this technique allows the determination of all elements of the magnetic susceptibility tensor and not just the easy-axis direction as is available from electrostatic predictions. Furthermore, it has the capacity to provide each of the anisotropic magnetic susceptibility tensors for all independent magnetic ions in a molecule, thus allowing for studies of poly-nuclear complexes and compounds of higher crystalline symmetry than triclinic.

International audience; In this paper, we propose a simple cluster model with limited basis sets to reproduce the unpaired electron distributions in a YTiO3 ferromagnetic crystal. The spin-resolved one-electron-reduced density matrix is reconstructed simultaneously from theoretical magnetic structure factors and directional magnetic Compton profiles using our joint refinement algorithm. This algorithm is guided by the rescaling of basis functions and the adjustment of the spin population matrix. The resulting spin electron density in both position and momentum spaces from the joint refinement model is in agreement with theoretical and experimental results. Benefits brought from magnetic Compton profiles to the entire spin density matrix are illustrated. We studied the magnetic properties of the YTiO3 crystal along the Ti–O1–Ti bonding. We found that the basis functions are mostly rescaled by means of magnetic Compton profiles, while the molecular occupation numbers are mainly modified by the magnetic structure factors.

We report neutron-diffraction investigations of the quasi-2D Mn-II(dca)(2)(pym)(H2O) (pym = N2C4H4) compound, where high-spin Mn-II ions are bridged by dicyanamide anions, [N(CN)(2)](-) (herein abbreviated dca). Inside the layers, Mn2+ ions are connected by single or double dca bridges. The magnetic phase diagram was established by neutron diffraction on a single crystal. In the low-field phase, the Mn-II ions are antiferromagnetically ordered in the layers, with moments nearly parallel to the c axis, and the layers are antiferromagnetically coupled. The spin-flop phase corresponds to ferromagnetic coupling between the antiferromagnetic layers, in which the Mn-II moments are nearly perpendicular to the c axis. The induced spin-density distribution in the paramagnetic phase, determined by polarized neutron diffraction, visualizes the superexchange pathways through the dca ligands within the layers and through H bonding between neighboring layers. The theoretical spin density obtained by bidimensional periodic DFT calculations is compared with the experimental results. Furthermore, quantum Monte Carlo simulations have been performed to compare the DFT results with experimental susceptibility measurements.

La distribution des electrons dans un cristal peut etre reconstruite avec une grande precision au moyen de la diffraction des rayons X a haute resolution, alors que la diffraction des neutrons polarises en spin permet de retrouver la densite d’aimantation. Bien que les grandeurs ainsi mesurees soient toutes deux une traduction du comportement des electrons, ces deux types d’experience de diffusion sont de nos jours interpretees par des modeles differents. Nous retracons ici les etapes ayant conduit a la premiere determination experimentale de la densite d’electrons resolue en spin par un traitement combine des donnees de diffraction de rayons X et de neutrons polarises.

Orbital ordering below 30 K was previously observed in the ferromagnetic ${\text{YTiO}}_{3}$ compound both by polarized neutron diffraction (PND) and x-ray magnetic diffraction (XMD). In this paper we report a procedure for the joint refinement of a unique spin-density model based on both PND and XMD data. The distribution of the unpaired $3d$ electron of titanium is clearly seen on the magnetization density reconstructed by the maximum entropy method from the PND data collection at 5 K. The ${\text{Ti}}^{3+}$ $3d$ orbital populations obtained by joint model refinement are discussed in terms of the orbital ordering scheme. Small but significant magnetic moments on apical oxygen ${\text{O}}_{1}$ and yttrium atoms are found. The agreement between experimental and theoretical spin densities obtained using density functional theory is discussed.

International audience; Joint refinement of X-ray and polarized neutron diffraction data has been carried out in order to determine charge and spin density distributions simultaneously in the nitro­nyl nitroxide (NN) free radical Nit(SMe)Ph. For comparison purposes, density functional theory (DFT) and complete active-space self-consistent field (CASSCF) theoretical calculations were also performed. Experimentally derived charge and spin densities show significant differences between the two NO groups of the NN function that are not observed from DFT theoretical calculations. On the contrary, CASSCF calculations exhibit the same fine details as observed in spin-resolved joint refinement and a clear asymmetry between the two NO groups.

While it is mostly well-known for its excellent thermoelectric properties at high temperatures, Yb${}_{14}$MnSb${}_{11}$ has also come to be characterized as a ferromagnetic semiconductor. Such systems support ferromagnetic correlations similar to metallic compounds, although with semiconducting electronic configurations. As such, they are promising materials for use in spintronic devices. The authors here use neutron scattering techniques to examine the distribution of the magnetization density and the energy scale of the magnetic exchange interactions in the Yb${}_{14}$MnSb${}_{11}$ compound. They find that there is a large ferromagnetic moment localized to the Mn site in the crystal structure with a `Kondo cloud' of negative moments distributed over the other atomic sites. The results provide long sought answers to questions regarding the valence, magnetization distribution, and energy scales of the magnetic interactions in this compound. The ferromagnetic spin-wave spectrum supports this premise and is well characterized by inclusion of nearest and next-nearest neighbor exchange interactions.

A method to map spin-resolved electron distribution from combined polarized neutron and X-ray diffraction is described and applied for the first time to a molecular magnet and it is shown that spin up density is 5% more contracted than spin down density., Since the 1980s it has been possible to probe crystallized matter, thanks to X-ray or neutron scattering techniques, to obtain an accurate charge density or spin distribution at the atomic scale. Despite the description of the same physical quantity (electron density) and tremendous development of sources, detectors, data treatment software etc., these different techniques evolved separately with one model per experiment. However, a breakthrough was recently made by the development of a common model in order to combine information coming from all these different experiments. Here we report the first experimental determination of spin-resolved electron density obtained by a combined treatment of X-ray, neutron and polarized neutron diffraction data. These experimental spin up and spin down densities compare very well with density functional theory (DFT) calculations and also confirm a theoretical prediction made in 1985 which claims that majority spin electrons should have a more contracted distribution around the nucleus than minority spin electrons. Topological analysis of the resulting experimental spin-resolved electron density is also briefly discussed.

International audience; A new charge and spin density model and the corresponding refinement software were recently developed to combine X- ray and polarized neutron diffraction experiments [1]. This joint refinement procedure allows getting access to both charge and spin densitiy distributions by refining both spin up and down parameters for magnetic atoms. The paper is focused on the refinement procedure and its application to the case of an end-to-end azido double bridged copper(II) complex. The results of this joint refinement of X-ray and polarized neutron diffraction data are presented.

A new polarized neutron diffractometer has been commissioned at the ORPHEE reactor in Saclay. Hot neutron beam of 0.8 A is obtained by Heussler monochromator. 64 position sensitive detectors cover angular range of 25°×80° in vertical and horizontal directions, which allows performing measurements of flipping ratios for dozens of reflections simultaneously. Data acquisition and data treatment programs have been developed. Possibilities of multi-crystal experiments and studies of powder samples have been tested. Ensemble of these measures should boost the data acquisition rate of magnetization densities by a factor of 5-20 depending on the crystal cell parameters.

International audience; no abstract

International audience; We have determined by polarized neutron diffraction (PND) the low-temperature molecular magnetic susceptibility tensor of the anisotropic low-spin complex PPh4 [Fe(III) (Tp)(CN)3 ]⋅H2 O. We found the existence of a pronounced molecular easy magnetization axis, almost parallel to the C3 pseudo-axis of the molecule, which also corresponds to a trigonal elongation direction of the octahedral coordination sphere of the Fe(III) ion. The PND results are coherent with electron paramagnetic resonance (EPR) spectroscopy, magnetometry, and ab initio investigations. Through this particular example, we demonstrate the capabilities of PND to provide a unique, direct, and straightforward picture of the magnetic anisotropy and susceptibility tensors, offering a clear-cut way to establish magneto-structural correlations in paramagnetic molecular complexes

We report on a microscopic study of the evolution of ferromagnetism in the Ru substituted ferromagnetic superconductor FM SC UCoGe crystallizing in the orthorhombic TiNiSi type structure. For that purpose, two single crystals with composition UCo0.97Ru0.03Ge and UCo0.88Ru0.12Ge have been prepared and characterized by magnetization, AC susceptibility, specific heat and electrical resistivity measurements. Both compounds have been found to order ferromagnetically below TC 6.5 and 7.5 K, respectively, which is considerably higher than the TC 3 K of the parent compound UCoGe. The higher values of TC are accompanied by enhanced values of the spontaneous moment amp; 956;spont 0.11 amp; 956;B f.u. and amp; 956;spont 0.21 amp; 956;B f.u., respectively in comparison to the tiny spontaneous moment of UCoGe about 0.07 amp; 956;B f.u. . No sign of superconductivity was detected in either compound. The magnetic moments of the samples were investigated on the microscopic scale using polarized neutron diffraction PND and for UCo0.88Ru0.12Ge also by soft X ray magnetic circular dichroism XMCD . The analysis of the PND results indicates that the observed enhancement of ferromagnetism is mainly due to the growth of the orbital part of the uranium 5f moment amp; 956;UL, reflecting a gradual localization of the 5f electrons with Ru substitution. In addition, the parallel orientation of the U and Co moments has been established in both substituted compounds. The results are discussed and compared with related isostructural ferromagnetic UTX compounds T transition metals, X Si, Ge in the context of a varying degree of the 5f ligand hybridization

International audience; The spin population distribution in the ferromagnetically coupled hetero-bimetallic chain compound [MnNi(NO(2))(4)(en)(2)] (en = 1,2-ethanediamine) has been investigated by means of polarized neutron diffraction experiments, and the results compared with those from theoretical estimates obtained via calculations based on density functional theory on dinuclear molecular models of the chain. The spin distributions obtained from experiment and from theory are consistent and reflect a larger spin delocalization from the Ni atom due to the more covalent character of the Ni-N bonds compared to the Mn-O ones. Also a nearly isotropic spin distribution is observed for the more ionic d(5) Mn(2+) ion and a clearly anisotropic distribution for the d(8) Ni(2+) ion. The use of dinuclear molecular models for the calculation of the exchange coupling constant between Ni and Mn provide upper and lower limits (+17.6 and -4.2 cm(-)(1)) for the experimentally determined value (+1.3 cm(-)(1)), depending on how the missing part of the chain is simulated, but yield essentially the same spin distribution. The Mn(II)-Ni(II) weak ferromagnetic coupling in the chain is interpreted in a spin delocalization mechanism as resulting from the weakness of the overlap between the magnetic orbitals centered on nickel and those centered on manganese which are only weakly delocalized on the ligands.

International audience; The magnetic anisotropy is a prerequisite for a metal complex to behave as a single-molecule magnet (SMM). Unfortunately, today we do not fully understand the relationships between the local structural parameters and the magnetic anisotropy that results at the molecular level. This is an issue that has become recursive in this area. Out of the synthesis work which is still important, but generates a multiplication of SMMs with frustrating properties, there are various studies to understand these relationships among which most are theoretical studies. In this context, we believe that polarized neutron diffraction (PND) can provide an experimental and complementary point of view to these theoretical studies. PND is indeed well known to allow an accurate determination of the spin density in magnetic compounds and in the field of molecular magnetism it has provided unique information on the pathways and the nature of intra- or intermolecular magnetic coupling [1]. In the case of highly anisotropic paramagnetic materials, where local magnetic moments cannot be aligned by an external magnetic field, that is more tricky, but a method based on local magnetic susceptibility tensor, has been recently developed that allows now analysing the data in this case and obtaining the magnetization distribution [2]. This approach was first used for inorganic compounds. Our idea has been to use this approach to go beyond the reconstruction of spin density to study the magnetic anisotropy in molecular systems. In this paper, we present the results of such an approach applied for the first time to metal complexes that are simple mono and dinuclear cobalt(II) complexes.

International audience; Since the first works of Stewart on modelling charge density [1], huge improvements of X-ray sources, detectors and software has significantly increased the resolution and the qualities of diffraction data allowing an accurate determination of the charge density of a growing number of molecules. However, despite the technological improvement, no dramatic change of the experimental model was reached since the multipolar model of Hansen & Coppens in 1978 [2]. At the same time polarised neutron diffraction (PND) experiments were developed [3] to get access to the spin density at the molecular scale and the multipolar Hansen & Coppens model was adapted to model this quantity. These two quantities (charge and spin densities) are described by a similar multipolar atom centred model with a common parameterization, therefore a combined treatment of X-ray diffraction (XRD) and PND data, is not only possible but also useful as stated by Becker & Coppens in 1985 [4]. Recently an extended Hansen & Coppens model and the corresponding refinement program were developed [5, 6] in order to allow the joint refinement of data sets coming from three different experiments (X-ray, unpolarised and polarised neutron diffraction). By combining different data sets, the new model gives access to electron density with spin up (ρ↑) and electron density with spin down (ρ↓) separately. These two quantities (ρ↑ and ρ↓) can be observed experimentally for the first time, and this observation allows a further comparison with theoretical models. In a first part the presentation will focus on the common model and the refinement procedure. The second part will describe its application to the case of an end-to-end azido double-bridged copper(II) complex (Cu2L2(N3)2 where L=1,1,1-triuoro-7-(dimethylamino)-4-methyl-5-aza-3-hepten-2-onato) [7]. The experimental results will be presented and compared to the theoretical densities. Acknowledgments This work has been supported by l’Agence Nationale de la Recherche (CEDA project); M.D. thanks CNRS for PhD fellowship.

We have investigated the ferromagnetic (FM) phase which suddenly develops in UCo1−xRuxAl and is isolated by paramagnetic regions on both sides from the parent UCoAl and URuAl. For that purpose we have grown high quality single crystals with x = 0.62, 0.70, 0.74, 0.75, and 0.78. The properties of the FM phase have been investigated by microscopic and macroscopic methods. Polarized neutron diffraction on a single crystal with x = 0.62 revealed the gradual growth of the hybridization between U and T-site in the U–T plane with increasing x. Hybridization works here as a mediator of the strong indirect interaction, while the delocalized character of the 5f states is still conserved. As a result very weak spontaneous magnetic moments are observed for all alloys with magnitude nearly in proportion to the TC for x < 0.62, while an enormous disproportion exists between them near xcrit. The magnetization, specific heat, electrical resistivity, and Hall effect measurements confirmed that the FM transition is suppres...

International audience; New crystallographic tools were developed to access a more precise description of the spin-dependent electron density of magnetic crystals. The method combines experimental information coming from high-resolution X-ray diffraction (XRD) and polarized neutron diffraction (PND) in a unified model. A new algorithm that allows for a simultaneous refinement of the charge- and spin-density parameters against XRD and PND data is described. The resulting software MOLLYNX is based on the well known Hansen-Coppens multipolar model, and makes it possible to differentiate the electron spins. This algorithm is validated and demonstrated with a molecular crystal formed by a bimetallic chain, MnCu(pba)(H2O)3 2H2O, for which XRD and PND data are available. The joint refinement provides a more detailed description of the spin density than the refinement from PND data alone

International audience; We present the latest progress of the Convergence of spin, charge and momentum Electron Density Analysis project. In the last decade, few methods have been proposed to build a quantum model with the goal to refine some of its important features using a set of different experiments. Among these experiments, high resolution x-rays and polarized neutrons diffraction, convergent beam electron diffraction, magnetic and non magnetic X-rays Compton scattering are the most popular. The joint refinement of charge and spin electron densities in position and momentum spaces from a set of different experiments, as those previously mentioned, requires the use of a common quantity known as the “one particle reduced density matrix” (1-RDM). Its role will be briefly presented. We will then explain the model(s), the method and the strategies that were adopted in constructing a general joint refinement computer code. Results from preliminary tests will be described as well as possible future joint studies.

We present a combination of ab initio calculations, magnetic Compton scattering, and polarized neutron experiments, which elucidate the density distribution of unpaired electrons in the kagome staircase system ${\text{Co}}_{3}{\text{V}}_{2}{\text{O}}_{8}$. Ab initio wave functions were used to calculate the spin densities in real and momentum spaces, which show good agreement with the respective experiments. It has been found that the spin polarized orbitals are equally distributed between the ${t}_{2g}$ and the ${e}_{g}$ levels for the spine $(s)$ Co ions while the ${e}_{g}$ orbitals of the cross-tie $(c)$ Co ions only represent 30% of the atomic spin density. Furthermore, the results reveal that the magnetic moments of the cross-tie Co ions, which are significantly smaller than those of the spine Co ions in the zero-field ferromagnetic structure, do not saturate by applying an external magnetic field of 2 T along the easy axis $a$. In turn, the increasing bulk magnetization, which can be observed by field dependent macroscopic measurements, originates from induced magnetic moments on the O and V sites. The refined individual magnetic moments are $\ensuremath{\mu}({\text{Co}}_{c})=1.54(4){\ensuremath{\mu}}_{B}$, $\ensuremath{\mu}({\text{Co}}_{s})=2.87(3){\ensuremath{\mu}}_{B}$, $\ensuremath{\mu}(\text{V})=0.41(4){\ensuremath{\mu}}_{B}$, $\ensuremath{\mu}(\text{O}1)=0.05(5){\ensuremath{\mu}}_{B}$, $\ensuremath{\mu}(\text{O}2)=0.35(5){\ensuremath{\mu}}_{B}$, and $\ensuremath{\mu}(\text{O}3)=0.36(5){\ensuremath{\mu}}_{B}$ combining to the same macroscopic magnetization value, which was previously only attributed to the Co ions.

We present novel insight on like-spin domains (LSD) in cooperative spin transition solids by following the photo-transformation and the subsequent relaxation of a [ Fe(ptz)6] (BF4)2 single crystal in the vicinity of the light-induced instability. Self-organization under light is observed, accompanied by Barkhausen-like noise and jumps which reveal the presence of elastic interactions between LSDs. The light-induced phase separation process is discussed in terms of a dynamic potential providing spinodal instability in the corresponding temperature range. This useful concept is applicable to all types of switchable molecular solids.

International audience; A new charge and spin density model and the corresponding refinement software were recently developed to combine X-ray and polarised neutron diffraction experiments [1,2]. This joint refinement procedure allows for an access to both the charge and spin densities but also to spin up ( ) and spin down ( ) electron distributions. These two quantities ( and ) were thus separately modelled and for the first time it was possible to compare them with theoretical results. The first part of the presentation will introduce the refinement procedure and describe its application to the case of an end-to-end azido double bridged copper(II) complex[3]. The results of this joint refinement of X-ray and polarized neutron diffraction data will be compared to theoretical calculations. The second part will be devoted to recent applications to other materials including a purely organic radical.

International audience; Although analysis of joint experimental data as diverse as x-rays structure factors, polarized neutron flipping ratios, neutron structure factors, CBED measurements, x-ray Compton magnetic (and non magnetic) profiles, among others, is theoretically feasible and desirable, to this day only few attempts have been made. We will remind some of the important strategies that have been elaborated in the past and we will propose a new possible way of combining and exploiting the richness of the diversity of experimental methods. We will show that as long as the data are issued from elastic coherent scattering experiments only marginal changes have to be made to the usual pseudo-atoms model. This will be illustrated with recent results obtained on magnetic compounds. We will finally address a critical discussion concerning the difficulties occurring in combining real and momentum space data.

The topology of the electrostatic potential has been studied for single molecules using geometries and electron distributions determined from high-resolution singlecrystal X-ray diffraction experiments. The electrostatic potential gradient, which is the negative of the electric field, has been represented, revealing the position of zeroflux surfaces and critical points [1]. Through the representation of the electric field lines, the relationship between the topology of the electrostatic potential and the reactivity of the molecule is explored. Local maxima, corresponding to the nuclei, are associated to electrophilic sites, while local minima, which are related to local accumulations of electron density (i.e. in lone pairs) are identified as nucleophilic sites. Both kind of reactive sites present influence zones delimited by zeroflux surfaces. Space can be partitioned in disjoint volumes (primary bundles) [2], each of them corresponding to the intersection of the influence zones of at most one nucleophilic and one electrophilic sites, which are both situated on the surface of the primary bundle. As a result of the interaction with the whole molecule, a probe charge inside one primary bundle is directed to, or away from, the reactive sites on the surface of this volume. The bundles can be added to generate two separate partitions in nucleophilic and electrophilic influence zones, the completeness of these last partitions depending on either the neutral or ionic character of the molecule. Analogous to the bond and ring critical points in the electron density topology, electrostatic potential saddle critical points appear on the zero-flux surfaces. They are related to the limits of the electrophilic and nucleophilic influence zones, which can extend to infinite or present a finite volume. In this last case, a saddle point reveals the limit of the influence zone and the path for preferred attack on the reactive site.

The spin density of the free radical DPPH (diphenylpicrylhydrazyl) has been studied in the crystallized form DPPH C6H6 with the paramagnetic moments orientated, at low temperature, by an external magnetic field. As the crystal structure is not centro-symmetrical, it is not possible to obtain directly the magnetic structure factors from the experimental flipping ratios measured by polarized neutron diffraction. For this reason we have represented the spin density by a multipole expansion on the different atoms, and determined directly the parameters of this expansion by comparing with the observed flipping ratios. A large amount of the spin density is shared between the 2 nitrogen atoms (Nα and Nβ) of the hydrazyl group. The remaining part is delocalized on the 2 phenyl and the picryl rings. These results are compared to magnetic resonance data.

The spin density of the free radical DPPH (diphenylpicrylhydrazyl) has been studied in the crys ta l l ized form DPPH CgHg with the paramagnetic moments orientated, a t low temperature, by an external magnetic f i e ld . As the crys ta l s t ructure i s not centro-symmetrical, i t is not possible to obtain d i rec t ly the magnetic s t ruc ture factors from the experimental f l ipping r a t io s measured by polarized neutron di f f rac t ion. For t h i s reason we have represented the spin density by a multipole expansion on the different atons, and determined d i rec t ly the parameters of this expansion by comparing with the observed flipping r a t i o s . A large amount of the spin density i s shared between the 2 nitrogen atoms (% and Ng) of the hydrazyl group. The remaining part i s delocalized on the 2 phenyl and the picryl r ings . These r e su l t s are compared to magnetic resonance data.

One of the major issues in the field of molecular magnetism is to better understand and predict the correlations between the structural properties of molecule-based compounds and their magnetic properties, all of which may be tunable using “bottom-up” synthesis methods. In particular, the understanding and control of the magnetic anisotropy at the atomic scale is essential, especially with the aim to design Single-Molecule Magnets (SMM) with higher blocking temperatures. In this context, this thesis work is focused on two mains subjects. The first part deals with the determination and the characterization of the local magnetic anisotropy in low-nuclearity molecular complexes based on transition ions. Polarised neutron diffraction (PND) allows us, for the first time, to directly access the local susceptibility tensor in a Low-Spin Fe(3+) mononuclear complex as well as in two, mononuclear and dinuclear, High-Spin Co(2+) complexes. This innovative approach leads to the establishment of unique and direct magneto-structural correlations, by relating the local magnetic principal directions with the coordination environment of the metallic ions and, in particular, with the local distortion axes. We have also carried out an original investigation by inelastic neutron scattering (INS) of a Mn(3+) thermo-induced spin-transition compound in both High-Spin (HS) and Low-Spin (LS) states. On the basis of this study, we were able to propose an anisotropic spin-Hamiltonian model in both HS and LS phases, and their relationships with the structure of the molecule are discussed. In a second more exploratory part of the thesis, we have performed by small-angle neutron scattering (SANS) a complete study of the structural and magnetic properties of Prussian blue analogues (PBA) ferromagnetic nanoparticles CsNiCr. The effects of size, organization and concentration on their superparamagnetic properties have been clearly highlighted. In particular, a strong magnetic contribution has been observed for the smallest particles (5 nm diameter) which results from the manifestation of a collective process, while the biggest (28 nm diameter) appear to be in a multi-domain state., Un des enjeux majeurs dans le domaine du magnétisme moléculaire est de mieux comprendre et prévoir, dans les composés à base moléculaire, les corrélations qui existent entre les propriétés structurales (modulables à partir de méthodes de synthèse de type « bottom-up ») et les propriétés magnétiques. En particulier, la compréhension et la maîtrise de l’anisotropie magnétique à l’échelle locale est primordiale, notamment en vue de concevoir des molécules-aimants avec de plus hautes températures de blocage. Dans ce contexte, ce travail de thèse s’organise autour de deux grands axes. La première partie se concentre sur la détermination et la caractérisation de l’anisotropie magnétique locale dans des complexes moléculaires d’ions de transition de faible nucléarité. La diffraction de neutrons polarisés (PND) nous a permis, pour la première fois, de mettre clairement en évidence le tenseur de susceptibilité magnétique locale dans un complexe moléculaire mononucléaire de Fe(3+) Bas-Spin ainsi que dans deux complexes, mononucléaire et dinucléaire, de Co(2+) Haut-Spin. Cette approche novatrice mène à l’établissement de relations magnéto-structurales claires et directes, en reliant les directions magnétiques locales propres à l’environnement de coordination des ions métalliques et en particulier aux axes locaux de distorsion. Nous avons également mené l’étude originale d’un complexe à transition de spin thermo-induite de Mn(3+) par diffusion inélastique de neutrons (INS) dans les deux phases Haut-Spin (HS) et Bas-Spin (BS). Cette étude nous a conduits à la proposition d’un modèle d’hamiltonien de spin anisotrope dans les deux états HS et BS, en relation avec la structure du complexe. Dans une seconde partie plus exploratoire de la thèse, nous avons mené une étude complète des propriétés structurales et magnétiques de nanoparticules ferromagnétiques d’analogue du bleu de Prusse CsNiCr, par diffusion de neutrons aux petits angles (SANS). Les effets de taille, d’organisation et de concentration sur leurs propriétés superparamagnétiques ont ainsi été clairement mis en évidence. En particulier, nous avons mis en exergue, pour les particules de plus petite taille (5 nm de diamètre), une contribution magnétique qui résulte de la manifestation d’un phénomène collectif, tandis que celles de plus grande taille (28 nm de diamètre) apparaissent être dans un état complètement multidomaine.