Dynamics of supercooled water in various mesopore sizes

Double-quantum-filtered NMR and T(1) inversion-recovery spectroscopy were employed to exploit the temperature-dependent dynamics of D(2)O confined in MCM-41. Samples with three pore sizes of 1.58, 2.03, and 2.34 nm and two D(2)O contents were investigated. The reorientation correlation times of confined D(2)O in variously sized pores exhibit different temperature dependencies. The results reveal that the D(2)O molecules at fast motion site remain mobile below approximately 225 K and a liquid-liquid phase transition occurs around this temperature for all samples studied. This temperature is thought to be unreachable for supercooled D(2)O. Particularly, in 20 wt % D(2)O loaded samples with pore diameters of 1.58 and 2.03 nm, the reorientational correlation times of D(2)O at fast motion site exhibit Arrhenius behavior between 225 and 290 K, while other samples show power law dependency. Thus, a liquid phase of the fragile type in bigger pores changes to the strong type in samples with smaller pores.     

The fragile-to-strong dynamic crossover transition in confined water: nuclear magnetic resonance results

By means of a nuclear magnetic resonance experiment, we give evidence of the existence of a fragile-to-strong dynamic crossover transition (FST) in confined water at a temperature T(L)=223+/-2 K. We have studied the dynamics of water contained in 1D cylindrical nanoporous matrices (MCM-41-S) in the temperature range 190-280 K, where experiments on bulk water were so far hampered by crystallization. The FST is clearly inferred from the T dependence of the inverse of the self-diffusion coefficient of water (1D) as a crossover point from a non-Arrhenius to an Arrhenius behavior. The combination of the measured self-diffusion coefficient D and the average translational relaxation time tau(T), as measured by neutron scattering, shows the predicted breakdown of Stokes-Einstein relation in deeply supercooled water.     

Phase diagram and glass transition of confined benzene

We used differential scanning calorimetry, neutron scattering, and proton NMR to investigate the phase behavior, the structure, and the dynamics of benzene confined in a series of cylindrical mesoporous materials MCM-41 and SBA-15 with pore diameters, d, between 2.4 and 14 nm. With this multitechnique approach, it was possible to determine the structure and, for the first time to our knowledge, the density of confined benzene as a function of temperature and pore size. Under standard cooling rates, benzene partially crystallizes in SBA-15 matrixes (4.7 相似文献   

Vibrational relaxation of CN stretch of pseudo-halide anions (OCN-, SCN-, and SeCN-) in polar solvents

The vibrational relaxation dynamics of pseudo-halide anions XCN- (X = O, S, Se) in polar solvents were studied to understand the effect of charge on solute-to-solvent intermolecular energy transfer (IET) and solvent assisted intramolecular vibrational relaxation (IVR) pathways. The T1 relaxation times of the CN stretch in these anions were measured by IR pump/IR probe spectroscopy, in which the 0-1 transition was excited, and the 0-1 and 1-2 transitions were monitored to follow the recovery of the ground state and decay of the excited state. For these anions in five solvents, H2O, D2O, CH3OH, CH3CN, and (CH3)2SO, relaxation rates followed the trend of OCN- > SCN- > SeCN-. For these anions and isotopes of SCN-, the relaxation rate was a factor of a few (2.5-10) higher in H2O than in D2O. To further probe the solvent isotope effect, the relaxation rates of S12C14N-, S13C14N-, and S12C15N- in deuterated methanols (CH3OH, CH3OD, CH3OH, CD3OD) were compared. Relaxation rate was found to be affected by the change of solvent vibrational band at the CN- stretching mode (CD3 symmetric stretch) and lower frequency regions, suggesting the presence of both direct IET and solvent assisted IVR relaxation pathways. The possible relaxation pathways and mechanisms for the observed trends in solute and solvent dependence were discussed.     

Fragile-to-strong liquid transition in deeply supercooled confined water

Confining water in lab synthesized nanoporous silica matrices MCM-41-S with pore diameters of 18 and 14 A, we have been able to study the molecular dynamics of water in deeply supercooled states, down to 200 K. Using quasielastic neutron scattering and analyzing the data with the relaxing cage model, we determined the temperature variation of the average translational relaxation time and its Q-dependence. We find a clear evidence of an abrupt change of the relaxation time behavior at T approximately equal to 225 K, which we interpreted as the predicted fragile-to-strong liquid-liquid transition.     

Density measurement of 1-d confined water by small angle neutron scattering method: pore size and hydration level dependences

Small angle neutron scattering (SANS) is used to measure the absolute density of water contained in 1-D cylindrical pores of a silica material MCM-41-S with pore diameters of 19 and 15 A. By being able to suppress the homogeneous nucleation process inside the narrow pore, one can keep water in the liquid state down to at least 160 K. From a combined analysis of SANS data from both H(2)O and D(2)O hydrated samples, we determined the absolute value of the density of 1-D confined water. We found that the average density of water inside the fully hydrated 19 A pore is 8% higher than that of the bulk water at room temperature. The temperature derivative of the density shows a pronounced peak at T(L) = 235 K signaling the crossing of the Widom line at ambient pressure and confirming the existence of a liquid-liquid phase transition at an elevated pressure. Pore size and hydration level dependences of the density are also studied.     

Structure of methanol confined in MCM-41 investigated by large-angle X-ray scattering technique

Large-angle X-ray scattering (LAXS) measurements over a temperature range from 223 to 298 K have been made on methanol confined in mesoporous silica MCM-41 with two different pore diameters, 28 A (C14) and 21 A (C10), under both monolayer and capillary-condensed adsorption conditions. To compare the structure of methanol in the MCM-41 pores with that of bulk methanol, X-ray scattering intensities for bulk methanol in the same temperature range have also been measured. The radial distribution functions (RDFs) for the monolayer methanol samples showed that methanol molecules are strongly hydrogen bonded to the silanol groups on the MCM-41 surface, resulting in no significant change in the structure of adsorbed methanol with respect to the pore size and temperature. On the other hand, the RDFs for the capillary-condensed methanol samples showed that hydrogen-bonded chains of methanol molecules are formed in both pores. However, the distance and number of hydrogen bonds estimated from the RDFs suggested that hydrogen bonds between methanol molecules in the pores are significantly distorted or partly disrupted. It has been found that the hydrogen bonds are more distorted in the smaller pores of MCM-41. With decreasing temperature, however, the hydrogen-bonded chains of methanol in the pores were gradually ordered. A comparison of the present results on methanol in MCM-41 pores with those on water in the same pores revealed that the structural change with temperature is less significant for confined methanol than for confined water.     

Combined structural refinement of Bi3.5La0.5Ti3O12 using neutron and X-ray powder diffraction data

A combined structural refinement of Bi3.5La0.5Ti3O12 against both neutron and X-ray diffraction data was performed at 298 K on the basis of the Raman study. The upshift of Raman peaks suggested that the substitution sites of La atoms in Bi3.5La0.5Ti3O12 were only the Bi sites in the perovskite units. Of the two crystal structural models (orthorhombic and monoclinic systems) considered for the crystal structural system of Bi3.5La0.5Ti3O12, the weighted R factor, Rwp, and goodness-of-fit indicator, S (=Rwp/Re), of the monoclinic system were lower than those of the orthorhombic one. The final Rwp and S values based on the monoclinic system were 7.04% (6.34 and 7.76% for the neutron data and the X-ray data, respectively) and 1.45, respectively. The lattice parameters obtained from the combined structural refinement were a = 5.4321(1) A, b = 5.4161(1) A, and c = 32.8614(3) A. The beta angle was 89.95(4) degrees . Spontaneous polarizations calculated from the refined structural parameters were 27.0 microC/cm2 for the monoclinic system and 1.8 microC/cm2 for the orthorhombic one.     

Detection of homogeneous distribution of functional groups in mesoporous silica by small angle neutron scattering and in situ adsorption of nitrogen or water

The distribution of SO(3)H-functional groups attached to the ordered inner pore walls of mesoporous Si-MCM-41 materials based on SiO(2) was investigated by gas adsorption combined with in situ small angle neutron scattering (SANS). The functionalization was performed by two different methods, (i) grafting and (ii) co-condensation. The adsorbates N(2) at 77 K or a H(2)O/D(2)O mixture of 42:58 at 298 K possess neutron scattering length densities (SLD) similar to that of SiO(2) and therefore quench the diffraction signals of the nonmodified silica. SANS measurements show that N(2) matches completely not only with the pristine mesoporous Si-MCM-41 but also with Si-MCM-41-SO(3)H functionalized by grafting. Thus, full access of adsorbate into the entire length of the pores is proven. For the analysis of the distribution of functional groups within the pores in dependence on the used functionalization method, grafting or co-condensation, however, the more specific adsorbate H(2)O/D(2)O (42:58) is necessary, because it reacts more sensitively toward small changes in the SLD of the host material. For grafted Si-MCM-41-SO(3)H materials, an incomplete quenching was observed, indicating that only some regions, probably the pore mouths, have been modified. For a sample functionalized by co-condensation, almost no quenching of the neutron diffraction was found, indicating a very homogeneous distribution of the functional groups along the entire pores.     

Plasticizer effect on the dynamics of polyvinylchloride studied by dielectric spectroscopy and quasielastic neutron scattering

We have studied the influence of plasticization on the microscopic dynamics of a glass-forming polymer. For this purpose we studied polyvinylchloride (PVC) with and without the commercially used plasticizer dioctylphthalate (DOP). We used dielectric spectroscopy and inelastic neutron scattering employing the neutron spin echo (NSE) technique. For both kinds of spectra the alpha relaxation could be consistently described by a model involving a distribution of individual relaxations of the Kohlrausch type. In contrast to earlier studies it turned out that an asymmetric distribution is necessary to fit the data at the lower temperatures investigated here. The shape parameters of the distribution (width, skewness) for PVC and PVC/DOP turned out to coincide when the characteristic relaxation times were the same. This means that the plasticizer only induces a remapping of the temperature dependence of the alpha relaxation. Comparison of NSE spectra S(Q,t)S(Q) at different scattering vectors Q gave the result that the slowing down at the structure factor peak Q(max) is surprisingly small for PVC while it is in the normal range for PVC/DOP.     

Variable temperature neutron diffraction and X-ray charge density studies of tetraacetylethane

Single crystal neutron diffraction data have been collected on a sample of enolized 3,4-diacetyl-2,5-hexanedione (tetraacetylethane, TAE) at five temperatures between 20 and 298 K to characterize the temperature-dependent behavior of the short, strong, intramolecular hydrogen bond. Upon decreasing the temperature from 298 K to 20 K, the O2-H1 distance decreases from 1.171(11) to 1.081(2) A and the O1...H1 distance increases from 1.327(10) to 1.416(6) A. The convergence of the C-O bond lengths from inequivalent distances at low temperature to identical values (1.285(4) A) at 298 K is consistent with a resonance-assisted hydrogen bond. However, a rigid bond analysis indicates that the structure at 298 K is disordered. The disorder vanishes at lower temperatures. Short intermolecular C-H...O contacts may be responsible for the ordering at low temperature. The intramolecular O...O distance (2.432 +/- 0.006 A) does not change with temperature. X-ray data at 20 K were measured to analyze the charge density and to gain additional insight into the nature of the strong hydrogen bond. Quantum mechanical calculations demonstrate that periodic boundary conditions provide significant enhancement over gas phase models in that superior agreement with the experimental structure is achieved when applying periodicity. One-dimensional potential energy calculations followed by quantum treatment of the proton reproduce the location of the proton nearer to the O2 site reasonably well, although they overestimate the O-H distance at low temperatures. The choice of the single-point energy calculation strategy for the proton potential is justified by the fact that the proton is preferably located nearer to O2 rather than being equally distant to O1 and O2 or evenly distributed (disordered) between them.     

Hygrothermal degradation of (3-glycidoxypropyl)trimethoxysilane films studied by neutron and X-ray reflectivity and attenuated total reflection infrared spectroscopy

Thin films of organosilanes have great technological importance in the areas of adhesion promotion, durability, and corrosion resistance. However, it is well-known that water can degrade organosilane films, particularly at elevated temperatures. In this work, X-ray and neutron reflectivity (XR and NR) were combined with attenuated total reflection infrared (ATR-IR) spectroscopy to study the chemical and structural changes within thin films of (3-glycidoxypropyl)trimethoxysilane (GPS) after exposure for various periods of time to air saturated with either D2O or H2O at 80 degrees C. For NR and XR, ultrathin (approximately 100 A) films were prepared by spin-coating. Both D2O and H2O provide neutron scattering contrast with GPS. Variations in the neutron scattering length density (SLD) profiles (a function of mass density and atomic composition) with conditioning time were measured after drying the samples out and also swelled with H2O or D2O vapor at room temperature. For samples that were dried out prior to measurement, little or no change was observed for H2O conditioning up to 3.5 days, but large changes were observed after 30 days of conditioning. The range of conditioning time for this structural change was narrowed to between 4 and 10 days with XR. The SLD profiles indicated that the top portion of the GPS film was transformed into a thick low-density layer after conditioning, but the bottom portion showed little structural change. A previous NR study of as-prepared GPS films involving swelling with deuterated nitrobenzene showed that the central portion of the film has much lower cross-link density than the region nearest the substrate. The present data show that the central portion also swells to a much greater extent with water and hydrolyzes more rapidly. The chemical degradation mechanism was identified by IR as hydrolysis of siloxane bonds. For ATR-IR, GPS films were prepared by dip-coating, which resulted in a greater and more variable thickness than for the spin-coated samples. The IR spectra revealed an increase in vicinal silanol generation over the first 3 days of conditioning followed by geminal silanol generation. Thus, the structural change detected by NR and XR roughly coincided with the onset of geminal silanol generation. Finally, little change in the reflectivity data was observed for films conditioned with D2O at 80 degrees C for 1 month. This indicates that hydrolysis of Si-O-Si is much slower with D2O than with H2O.     

A new crystal phase of barium nitroprusside trihydrate studied by neutron diffraction at 20 K

The crystal of barium penta­cyano­nitro­syl­ferrate trihydrate {barium nitro­prusside trihydrate, Ba[Fe(CN)₅(NO)]·3H₂O} has been studied by neutron diffraction at 20 K. The study was performed to characterize the structural phase generated by the phase transition undergone by the crystals at 80 K, at which temperature the unit‐cell volume doubles. This crystal phase still exists at 20 K. The crystal structure, in space group P1, is completely ordered. The positional changes of the water mol­ecules in the present structure with respect to those of the compound at 105 K are presented.     

An ordered low‐temperature phase of barium nitro­prusside trihydrate studied by neutron diffraction

Crystals of barium penta­cyano­nitro­syl­ferrate trihydrate (barium nitro­prusside trihydrate), Ba[Fe(CN)₅(NO)]·3H₂O, have been studied by neutron diffraction in order to ex­amine the structural behaviour of the compound in the 20–120 K temperature range and to determine the structure at 105 K. The results show the existence of a new crystal phase of the compound at 80 K (with a duplicated a parameter), which still exists at 20 K. The crystal structure at 105 K shows a rearrangement of the water mol­ecules, which results in an ordered structure with P1 symmetry. Two of the four independent nitro­prusside cations are rotated by 4.5° around the [100] direction.     

Structure and dynamics of halogenoethanol-water mixtures studied by large-angle X-ray scattering, small-angle neutron scattering, and NMR relaxation

To clarify the structure of solvent clusters formed in halogenoethanol-water mixtures at the molecular level, large-angle X-ray scattering (LAXS) measurements have been made at 298 K on 2,2,2-trifluoroethanol (TFE), 2,2,2-trichloroethanol (TCE), and their aqueous mixtures in the TFE and TCE mole fraction ranges of 0.002 < or = x(TFE) < or = 0.9 and 0.5 < or = x(TCE) < or = 0.9, respectively. The radial distribution functions (RDFs) for TFE-water mixtures have shown that the structural transition from inherent TFE structure to the tetrahedral-like structure of water takes place at x(TFE) approximately 0.2. In the TCE-water mixtures inherent TCE structure remains in the range of 0.5 < or = x(TCE) < or = 1. Small-angle neutron scattering (SANS) experiments have been performed on CF(3)CH(2)OD- (TFE-d(1)-) D(2)O and CF(3)CD(2)OH- (TFE-d(2)-) H(2)O mixtures in the TFE mole fraction range of 0.05 < or = x(TFE) < or = 0.8. The SANS results in terms of the Ornstein-Zernike correlation length have revealed that TFE and water molecules are most heterogeneously mixed with each other in the TFE-water mixture at x(TFE) approximately 0.15, i.e., both TFE clusters and water clusters are most enhanced in the mixture. To evaluate the dynamics of TFE and ethanol (EtOH) molecules in TFE-water and ethanol-water mixtures, respectively, (1)H NMR relaxation rates for the methylene group within alcohol molecules have been measured by using an inversion-recovery method. The alcohol concentration dependence of the relaxation rates for the TFE-water and ethanol-water mixtures has shown a break point at x(TFE) approximately 0.15 and x(EtOH) approximately 0.2, respectively, where the structural transition from alcohol clusters to the tetrahedral-like structure of water takes place. On the basis of the present results, the most likely structure models of solvent clusters predominantly formed in TFE-water and TCE-water mixtures are proposed. In addition, effects of halogenation of the hydrophobic groups on clustering of alcohol molecules are discussed from the present results, together with the previous ones for ethanol-water and 1,1,1,3,3,3-hexafluoro-2-propanol- (HFIP-) water mixtures.     

Adsorption of beta-hairpin peptides on the surface of water: a neutron reflection study

Neutron reflectivity has been used to determine the thickness and surface coverage of monolayers of two 14-residue beta-hairpin peptides adsorbed at the air/water interface. The peptides differed only in that one was labeled with a fluorophore, while the other was not. The neutron reflection measurements were mainly made in null reflecting water, NRW, containing 8.1% D(2)O. Under this isotopic contrast the water is invisible to neutrons and the specular signal was then only from the peptide layer. At the highest concentration of ca. 4 microg/mL studied, the area per peptide molecule (A) was found to be 230 +/- 10 and 210 +/- 10 A(2) for the peptides with and without a BODIPY-based fluorophore, respectively. The thickness of the peptide layers was about 10 A for a Gaussian distribution. With decreasing bulk peptide concentration, both surface excess and layer thickness showed a steady trend of decrease. While the neutron results clearly indicate structural changes within the peptide monolayers with increasing bulk concentration, the outstanding structural feature is the formation of rather uniform peptide layers, consistent with the structural characteristics typical of beta-strand peptide conformations. These structural features are well supported by the parallel measurements of the adsorbed layers in D(2)O. With this isotopic contrast the neutron reflectivity provides an estimate about the extent of immersion of the peptide layers into water. The results strongly suggest that the 14-mer peptide monolayers were fully afloat on the surface of water, with only the carboxy groups on Glu residues hydrated.     

Inelastic neutron scattering study of electron reduction in Mn12 derivatives

We report inelastic neutron scattering (INS) studies on a series of Mn(12) derivatives, [Mn(12)O(12)(O2CC6F5)16(H2O)4]z, in which the number of unpaired electrons in the cluster is varied. We investigated three oxidation levels: z = 0 for the neutral complex, z = -1 for the one-electron reduced species and z = -2 for the two-electron reduced complex. For z = 0, the ground state is S = 10 as in the prototypical Mn12-acetate. For z = -1, we have S = 19/2, and for z = - 2, an S = 10 ground state is retrieved. INS studies show that the axial zero-field splitting parameter D is strongly suppressed upon successive electron reduction: D = -0.45 cm(-1) (z = 0), D = -0.35 cm(-1) (z = -1), and D approximately -0.26 cm(-1) (z = -2). Each electron reduction step is directly correlated to the conversion of one anisotropic (Jahn-Teller distorted) Mn3+ (S = 2) to one nearly isotropic Mn2+ (S = 5/2).     

Dynamics of water confined in single- and double-wall carbon nanotubes

Using high-resolution quasielastic neutron scattering, we investigated the temperature dependence of single-particle dynamics of water confined in single- and double-wall carbon nanotubes with the inner diameters of 14+/-1 and 16+/-3 A, respectively. The temperature dependence of the alpha relaxation time for water in the 14 A nanotubes measured on cooling down from 260 to 190 K exhibits a crossover at 218 K from a Vogel-Fulcher-Tammann law behavior to an Arrhenius law behavior, indicating a fragile-to-strong dynamic transition in the confined water. This transition may be associated with a structural transition from a high-temperature, low-density (<1.02 gcm(3)) liquid to a low-temperature, high-density (>1.14 gcm(3)) liquid found in molecular dynamics simulation at about 200 K. However, no such dynamic transition in the investigated temperature range of 240-195 K was detected for water in the 16 A nanotubes. In the latter case, the dynamics of water simply follows a Vogel-Fulcher-Tammann law. This suggests that the fragile-to-strong crossover for water in the 16 A nanotubes may be shifted to a lower temperature.     

Synthesis, magnetic properties, and magnetic structure of a natrochalcite structural variant, KM(II)2D3O2(MoO4)2 (M = Mn, Fe, or Co)

We report the syntheses, crystal structures, and magnetic properties of KMn(2)(H(3)O(2))(MoO(4))(2) (MnH), KMn(2)(D(3)O(2))(MoO(4))(2) (MnD), KFe(2)(H(3)O(2))(MoO(4))(2) (FeH), KFe(2)(D(3)O(2))(MoO(4))(2) (FeD), KCo(2)(H(3)O(2))(MoO(4))(2) (CoH), and KCo(2)(D(3)O(2))(MoO(4))(2) (CoD), and the magnetic structures of MnD and FeD. They belong to the structural variant (space group I2/m) of the mineral natrochalcite NaCu(2)(H(3)O(2))(SO(4))(2) (space group C2/m) where the diagonal within the ac-plane of the latter become one axis of the former. The structure of MnD, obtained from Rietveld refinement of a high-resolution neutron pattern taken at 300 K, consists of chains of edge-sharing octahedra bridged by MoO(4) and D(3)O(2) to form layers, which are connected to K through the oxygen atoms to form the three-dimensional (3D)-network. The X-ray powder diffraction patterns of the other two compounds were found to belong to the same space group with similar parameters. The magnetic susceptibilities of MnH and FeH exhibit long-range ordering of the moments at a Ne?el temperature of 8 and 11 K, respectively, which are accompanied by additional strong Bragg reflections in the neutron diffraction in the ordered state, consistent with antiferromagnetism. Analyses of the neutron data for MnD and FeD reveal the presence of both long- and short-range orderings and commensurate magnetic structures with a propagation vector of (?, 0, ?). The moments are antiferromagnetically ordered within the chains with alternation between chains to generate four nonequivalent nuclear unit cells. For MnD the moments are perpendicular to the chain axis (b-axis) while for FeD they are parallel to the b-axis. The overall total is a fully compensated magnetic structure with zero moment in each case. Surprisingly, for KCo(2)(D(3)O(2))(MoO(4))(2) neither additional peaks nor increase of the nuclear peaks' intensities were observed in the neutron diffraction patterns below the magnetic anomaly at 12 K which was identified to originate from a small quantity of a ferromagnetic compound, Co(2)(OH)(2)MoO(4).     

Studies of finite molecular chains: synthesis, structural, magnetic and inelastic neutron scattering studies of hexa- and heptanuclear chromium horseshoes

We report the synthesis and structural characterisation of a family of finite molecular chains, specifically [{[R(2)NH(2)](3)[Cr(6)F(11)(O(2)CCMe(3))(10)]}(2)] (in which R=nPr 1, Et 2, nBu 3), [{Et(2)NH}(2){[Et(2)NH(2)](3)[Cr(7)F(12)(O(2)CCMe(3))(12)][HO(2)CCMe(3)](2)}(2)] (4), [{[Me(2)NH(2)](3)[Cr(6)F(11)(O(2)CCMe(3))(10)]2.5 H(2)O}(4)] (5) and [{[iPr(2)NH(2)](3)[Cr(7)F(12)(O(2)CCMe(3))(12)]}(2)] (6). The structures all contain horseshoes of chromium centres, with each Cr...Cr contact within the horseshoe bridged by a fluoride and two pivalates. The horseshoes are linked through hydrogen bonds to the secondary ammonium cations in the structure, leading to di- and tetra-horseshoe structures. Through magnetic measurements and inelastic neutron scattering studies we have determined the exchange coupling constants in 1 and 6. In 1 it is possible to distinguish two exchange interactions, J(A)=-1.1 meV and J(B)=-1.4 meV; J(A) is the exchange interactions at the tips of the horseshoe and J(B) is the exchange within the body of the horseshoe (1 meV=8.066 cm(-1)). For 6 only one interaction was needed to model the data: J=-1.18 meV. The single-ion anisotropy parameters for Cr(III) were also derived for the two compounds as: for 1, D(Cr)=-0.028 meV and |E(Cr)|=0.005 meV; for 6, D(Cr)=-0.031 meV. Magnetic-field-dependent inelastic neutron scattering experiments on 1 allowed the Zeeman splitting of the first two excited states and level crossings to be observed. For the tetramer of horseshoes (5), quantum Monte Carlo calculations were used to fit the magnetic susceptibility behaviour, giving two exchange interactions within the horseshoe (-1.32 and -1.65 meV) and a weak inter-horseshoe coupling of +0.12 meV. Multi-frequency variable-temperature EPR studies on 1, 2 and 6 have also been performed, allowing further characterisation of the spin Hamiltonian parameters of these chains.