AbstractsPhysics

Magnetic Resonance Pore Imaging, a new tool for porous media research

by Stefan A. Hertel




Institution: Victoria University of Wellington
Department:
Year: 2015
Keywords: Physics; Magnetic resonance; Imaging
Record ID: 1304211
Full text PDF: http://hdl.handle.net/10063/4233


Abstract

Porous media are highly prevalent in nature and span a wide range of systems including biological tissues, chemical catalysts or rocks in oil reservoirs. Imaging of the structure of the constituent pores is therefore highly desirable for life sciences and technological applications. This thesis presents the new development and application of a nuclear magnetic resonance (NMR) technique to acquire high resolution images of closed pores. The technique is a further development of diffusive-diffraction Pulsed Gradient Spin Echo (PGSE) NMR, which has been shown to image the pore auto-correlation function averaged over all pores. Until recently it was conventional wisdom that diffusive-diffraction PGSE NMR can only measure the magnitude of the form factor, due to its similarity to diffraction techniques such as x-ray and neutron scattering. In diffraction applications the loss of phase information is commonly referred to as the “phase problem”, which prevents the reconstruction of images of the pore space by inverse Fourier transform. My work is based on a recently suggested modification of the diffusive-diffraction PGSE NMR method, which creates a hybrid between Magnetic Resonance Imaging (MRI) and PGSE NMR. Therefore, we call this approach Magnetic Resonance Pore Imaging (MRPI). We provide experimental confirmation that MRPI does indeed measure the diffractive signal including its phase and thus the “phase problem” is lifted. We suggest a two-dimensional version of MRPI and obtain two-dimensional average pore images of cylindrical and triangular pores with an unprecedented resolution as compared to state of the art MRI. Utilizing a laser machined phantom sample we present images of microscopic pores with triangular shape even in the presence of wall relaxation effects. We therefore show that MRPI is able to reconstruct the pore shape without any prior knowledge or assumption about the porous system under study. Furthermore, we demonstrate that the MRPI approach integrates seamlessly with known MRI concepts. For instance we introduce “MRPI mapping” which acquires the MRPI signal for each pixel in an MRI image. This enables one to resolve pore sizes and shapes spatially, thus expanding the application of MRPI to samples with heterogeneous distributions of pores.