Umeå universitets logga

To optimize contrast agent in MRI scan region, e.g. to enhance paramagnetic relaxation in the MRI scan fields(0.1T-3T), one possible way is to slow down the tumbling of the paramagnetic complex. The effect of slowing down the reorientational motion of the complex to increase relaxivity is obvious and this strategy has already been employed in producing MRI contrast agent that can bind to specific proteins. An example is MS-325 binds to human serum albumin(HSA). The slow down effects on the ligands around paramagnetic ion, and on the zero field splitting(ZFS) interaction are under studies and the physics behind is still not clear. In this thesis, a generalized Solomon-Bloembergen-Morgan(GSBM) theory together with stochastic Liouville approach(SLA), is applied to investigate the mechanism behind the slow down effects. Two gadolinium complexes, MS-325+HSA and Gd(H₂O)₈³⁺+glycerol are studied by means of NMRD and ESR experiments.

GSBM is a second order perturbation theory with closed analytical form. The computation based on this theory is fast, but it has its limitation and in the case of Gd(S=7/2) the ZFS strength times its correlation time(Δt.τ_ƒ) should be less than 0.1. In comparison, the SLA is an "exact" theory that can evaluate the validity of GSBM calculation. However, the calculation in SLA is time consuming due to the large matrix it constructed. The major model used in GSBM is a two dynamic model, characterized by transient ZFS Δt and static ZFS Δs and their corresponding correlation time τ_ƒ and τR, while in SLA the model is only described by Δt and τ_ƒ. A combined NMRD and ESR analysis is used to understand the details of ZFS interaction. Both models can reproduce experimental NMRD profiles and model parameters are similar; for ESR linewidths the model parameters are quite different. The fitting results indicate the NMRD profiles are less sensitive to the detail expression of ZFS correlation function. In order to interpret both NMRD and ESR experiments with identical parameters, a more complex ZFS interaction model should be developed.