compared favourably to corresponding alternating pulsed field gradient (APFG) propagator measurements, featuring a stimulated echo, for xanthan solution flowing through the Ballotini packings. The flow, and subsequent mixing, of water and a glycerol solution through the two convergent entry ports of a Y-piece microchannel was also imaged using MRI, such that spatial distributions of the glycerol concentration and the velocity field were produced in 3D. These were compared with the simulation methodology, which featured a novel adaptation in that local viscosity was now dependent on local glycerol concentration. The simulation thus included coupled mass transfer and hydrodynamics. Agreement was excellent. [1] Sullivan SP, Gladden LF, Johns ML. J Non-Newtonian Fluid Mech 2006;133:91–8. [2] Harris RJ, Sederman AJ, Mantle MD, Crawshaw J, Johns ML. Magn Reson Imag 2005;23:355–7. doi:10.1016/j.mri.2007.01.057 Freezing transition in porous materials revised A. Khokhlov, R. Valiullin, J. Ka ¨rger Department of Interface Physics, University of Leipzig, Germany Phase transitions in confined space as a rule exhibit size-dependent shift in transitions [1]. As one of the applications, this is used for characterization of porous materials, exploiting, for example, melting or freezing temperature suppression. From these two, preference is usually given to the melting transition because it is often found that the freezing transition temperature is not reproducible [2]. This results from the fact that freezing is intrinsically a nucleation-driven process, making it to be dependent on particular conditions for nucleation (homogeneous nucleation or preexisting nucleus). In the present work, the conditions under which freezing proceeds via a solid-front propagation initiated by the frozen bulk phase surrounding porous particles are experimentally addressed. This is relevant for verification of the pore-blocking effect in freezing transition. The success of the experiments was based on the use of porous silicon samples with well-defined structures. For the comparative studies, samples (i) with nearly cylindrical pore morphology with the uniform diameters d 1 = 6 nm and d 2 = 8 nm, and (ii) with cross sections alternating between d 1 and d 2 have been utilized. NMR cryoporometry [3], exploiting the size- dependent melting point suppression, revealed two distinct melting transitions resulting in a bimodal pore-size distribution with the mean pore diameters of about 6 and 8 nm (black points in figure) for the porous silicon with the modulated structure. The thus obtained structural information has been further confirmed by SEM micrography. In contrary to the obtained two melting transitions, for the samples with the frozen bulk phase only one freezing transition has been detected. The location of the freezing temperature well corresponded to that measured in the porous silicon sample with the uniform cross section (d 1 ). This observation points out that the freezing in bigger pores is delayed until the corresponding thermodynamical conditions for a solid-front penetration through the narrower pore throats are fulfilled. The obtained experimental results unveil the significance of a pore- blocking effect for, at least, materials with the characteristic dimensions of the pore units used in this work. Additional experiments with no excess bulk phase have shown that the freezing initiated by a homogeneous nucleation in the bigger pores occurs at lower temper- atures. Thus, in our case the pore-blocking effect cannot be bypassed by a homogeneous nucleation. It may be of relevance, however, for porous materials with a wider pore-size distribution. Interestingly, a similar conclusion has been made for the evaporation transition, where the pore- blocking effect is found to be significant only in a certain range of the pore throat/pore size ratios [4]. Our experimental findings yield that combined analysis of the melting and freezing behavior of fluids in pores may provide more comprehensive information on the details of porous structure. [1] Gelb LD, et al. Rep Prog Phys 1999;62:1573. [2] Landry MR. Thermochim Acta 2005;433:27. [3] Strange JH, Rahman M, Smith EG. Phys Rev Lett 1993;71:3589. [4] Libby B, Monson PA. Langmuir 2004;20:4289. doi:10.1016/j.mri.2007.01.058 Investigation of self-diffusion processes in ionic hydrogels with a single-sided sensor J. Kolz a,b , K. Hunger b , S. Rath a,b , T. Mang a , B. Blu ¨mich b a Institute for Applied Polymer Science of Aachen University for Applied Science, Germany, b Institute for Technical and Macromolecular Chemistry of RWTH Aachen University, Germany Hydrogels consist of polymer networks that are able to absorb water to a multiple of their own weight, without dissolving. Typically, they are used as absorbers in a variety of products, like, for example, diapers or sealing materials in the building industry. In the last few years they have also attracted the attention of fields like biomedicine and sensor technology due to their use as smart hydrogels. These special types of hydrogels can change their shape in response to an external stimulus, like changes in pH value, temperature or ionic strength. Nevertheless, processes like swelling, water binding or diffusion taking place in hydrogels are not fully understood. In this work, we present measurements of self-diffusion processes in ionic Na-acrylate hydrogels conducted with a profile NMR-MOUSE. The use of a single-sided sensor allows easy sample preparation and free swelling of the hydrogels, without being restricted by the sample tube. Self-diffusion measurements were carried out using a Hahn-echo sequence followed by a CPMG train used to increase the signal-to-noise ratio. Short diffusion times were chosen to ensure that the diffusion of the water molecules was not restricted by the polymer chains. A series of hydrogels with increasing cross-linker content were synthesized to investigate the dependence of the diffusion coefficient on both the swelling degree and the cross-link density. Fig. 1 shows three diffusion curves for the sample with the lowest cross-linking with different amounts of water. The slope of the diffusion curves and therefore the diffusion coefficient increases with increasing amount of water. Fig. 2 shows the dependence of the diffusion coefficient on the swelling degree for samples with different cross-link densities (in mole cross-linker per mole of monomer). The measured diffusion coefficients are lower than the one of free water (2.21 m 2 /s). This can be explained by a fraction of water bound to the polar carboxylate groups of the polyacrylate gels. With higher swelling degree, the measured diffusion coefficient increases, due to an increasing amount of water that is not bound to the polymer chains. At high swelling degrees, the measured diffusion coefficient gets close to the one of free water proving that the fraction of bound water is relatively small. The samples with higher cross-link density did not reach the free water-like diffusion coefficient due to a lower degree of equilibrium swelling. Neglecting surface diffusion and setting the diffusion coefficient of the Abstracts / Magnetic Resonance Imaging 25 (2007) 544 – 591 566