Magnetic Resonance Imaging of coarse sediment

https://doi.org/10.1016/j.sedgeo.2008.07.002Get rights and content

Abstract

Non-destructive observation methods for coarse sediments are usually limited to two dimensions, for instance in opened cores or at the surface. We report a trial of a promising new method for three-dimensional imaging of gravelly sediments: Magnetic Resonance Imaging (MRI). MRI maps contrasts in density and relaxation properties of protons, which are very different for sediment and water in the pores. Images of glass bead mixtures show that the local porosity varies with sorting patterns or imperfect sample mixing, illustrating the potential for this non-destructive imaging method. As a quantitative test, we verify that a macroscopic sediment property – porosity – compares well to a standard lab determination and a recent model for predicting porosity for arbitrary grain size distributions. Finally we discuss potential applications and limitations of MRI to porosity and permeability, sedimentary structures, benthic life and slow groundwater or mass flows.

Introduction

Various issues demand data of how (gravelly) sediments are sorted and packed, such as porosity and permeability of hydrocarbon-rich deposits and groundwater reservoirs, mud and sand sedimentation in tidal environments and the initial motion and sediment transport of coarse sediments in coastal, estuarine and fluvial environments. In particular, the detailed sorting patterns in river beds of sand and gravel are notoriously difficult to measure in both field and laboratory conditions. Nevertheless such sorting patterns have ramifications for local cross-stratified sorting and morphodynamics of river beds up to downstream fining along entire rivers (see Kleinhans, 2004, Frings, 2008, for reviews). Common techniques for inspection of the sedimentary structures are limited to one or two dimensions, such as coring, lacquer profiles, photography or wax sampling. Most of these methods are destructive and the three-dimensional context of grains is lost.

We present results of a state of the art imaging technique that allows non-destructive imaging of volumes: Magnetic Resonance Imaging (MRI). MRI maps the density of free protons, notably in water. This technique is widely applied in medical studies but has also been applied to study food, animals, fossils, mummies, pottery, and to the mapping of structure, porosity and permeability of rocks for groundwater and hydrocarbon flows. We will illustrate the potential of MRI for sedimentology and geomorphology.

Various phenomena relevant to sedimentology have been studied by MRI in other disciplines. Materials with water- or hydrocarbon-filled voids have been mapped in detail, such as in vugular carbonates (Borgia et al., 1996), sedimentary structure (by mapping pore structure) in seabed cores (Bortolotti et al., 2006) and sandstone cores (Marica et al., 2006), and sea-ice pore structures filled with brine (Eicken et al., 2000). Mapping of flow velocity through porous media has also been accomplished with MRI (e.g. Davies et al., 1994, Merrill and Jin, 1994, Baumann et al., 2000, Deurer et al., 2002, Swider et al., 2007). Furthermore, fluids of different natures can be distinguished such as salt- and fresh-water interfaces in porous materials (Oswald et al., 2007).

The MRI must be distinguished from computer-aided tomography (CT, or computed axial tomography, CAT) which is, like MRI, well known from medical studies and has also been applied in many other disciplines. In CT a three-dimensional image is computed from a series of X-ray images taken from various angles. CT has, for example, been used to study estuarine benthic species (Perez et al., 1999).

The Magnetic Resonance Imaging (MRI) must also be distinguished from Nuclear Magnetic Resonance Measurement (NMR), which focusses on a single location or core cross-section and can then be applied to various locations or along a core. The MRI (imaging) acquires the image as a whole and is therefore suitable for high resolutions and large volumes required for sediments. The NMR on the other hand measures a very small volume. Various portable versions are commercially available. Recently NMR microscopic imaging has been developed. Another important application of NMR is measurement of permeability and velocity of a gas or fluid. See, e.g., Blümich (2000) for details. For example, spatial variation in porosity and moisture content (or proton density) has been measured successfully (compared to destructive conventional methods) in wood (Casieri et al., 2004), in brick samples of a Greek-Roman theatre (Brai et al., 2007) and in historical paper where it was used to assess the paper degradation state non-destructively (Blümich et al., 2003). See, e.g., Blümler et al. (1998) for a variety of applications.

The objective of this paper is a proof of principle: whether and how sediments and glass beads can be mapped through imaging of pore water by the MRI and whether microscopic (structure of particles) and macroscopic properties (porosity) of these granular media can correctly be derived from the data. In particular the porosity determined from MRI data, direct measurement and a model are compared. After an introduction to the basic principles of MRI and the experimental methods, the results of the measurements and calculations are presented. The discussion focuses on potential applications and limitations of MRI. In Appendix A the detailed settings of the MRI are listed to facilitate further work.

Section snippets

Basic principles

The MRI scanner is an instrument typically used in hospitals and in medical research for imaging of the human body in general and the brain in particular. A large patient can be positioned horizontally in the tubular machine. The instrument produces a strong static magnetic field so that it must be contained in a specially designed room that shields the magnetic field from the rest of the hospital. The magnetic field is so strong that metal objects such as tools, keys, watches and glasses

Porosity definition and measurement

The porosity κ of a granular medium is defined as the volumetric portion of space taken up by the pores,κ=VporesVtotalas opposed to the volumetric portion λ taken up by the particles:λ=VparticlesVtotalwhere V = volume and κ + λ = 1. The porosity of natural sediments and glass bead mixtures has a range of about 0.2 < κ < 0.4 or perhaps less for very wide mixtures.

The porosity was directly determined for the samples in containers with a known volume that were also used in the MRI. The grains plus pore

MR Imaging results

The basic experimental results are given in Table 1, and example slices are presented in Fig. 3, Fig. 5.

The marbles and 8 mm glass beads of series A are nearly circular in the images. There is some ‘spearhead’ distortion visible of the marbles, which rotates with the read-out direction of the MRI. This distortion is less if the read-out gradient is increased in accordance with theory (Bakker et al., 1993, Haacke et al., 1999). The orientation of the paramagnetic artifacts (visible as very

Proof of principle

The images clearly demonstrate that individual particles in mixtures with different sizes can well be imaged in three dimensions. The image resolution feasible for fast experiments is at present sufficient for the coarser sands and expected to improve further in the coming decade. Higher resolution images of smaller volumes can already be obtained in longer experiments. The quantitative data analysis in this paper is limited to a macroscopic property, namely porosity, which is accurately

Conclusions

MRI was demonstrated to be useful for three-dimensional mapping of sorting and particles in natural sediments. Several methods to determine porosity for mixtures of glass beads and gravel compared well in general and discrepancies were explainable. This method has potential for non-destructive imaging and quantitative analysis of stratification and sorting and for the determination of configuration of grains at a surface such as in armour layers of rivers as well as many other sedimentological

Acknowledgements

The comments by two reviewers and editor Gert Jan Weltje are acknowledged. The authors contributed in the following proportions to conception and design, data collection, data analysis and conclusions, and manuscript preparation: MK60,40,50,80%, CJ20,40,20,0%, CB10,20,20,20% and RF10,0,10,0%.

References (36)

  • KleinhansM.

    The key role of fluvial dunes in transport and deposition of sand-gravel mixtures, a preliminary note

    Sedimentary Geology

    (2001)
  • KleinhansM.

    Sorting in grain flows at the lee side of dunes

    Earth-Science Reviews

    (2004)
  • MaricaF. et al.

    Spatially resolved measurement of rock core porosity

    Journal of Magnetic Resonance

    (2006)
  • MerrillM. et al.

    Velocity measurements in natural porous rocks

    Magnetic Resonance Imaging

    (1994)
  • OswaldS.E. et al.

    Three-dimensional saltwater-freshwater fingering in porous media: contrast agent MRI as basis for numerical simulations

    Magnetic Resonance Imaging

    (2007)
  • SwiderP. et al.

    Use of high-resolution MRI for investigation of fluid flow and global permeability in a material with interconnected porosity

    Journal of Biomechanics

    (2007)
  • YuA. et al.

    An analytical-parametric theory of the random packing of particles

    Powder Technology

    (1988)
  • AberleJ. et al.

    Statistical properties of armored gravel bed surfaces

    Water Resources Research

    (2006)
  • Cited by (28)

    • Quantitative analysis on distribution of microcosmic residual oil in reservoirs by frozen phase and nuclear magnetic resonance (NMR) technology

      2020, Journal of Petroleum Science and Engineering
      Citation Excerpt :

      Therefore, investigating on the distribution law and occurrence characteristics of remaining oil is of great significance for EOR (Amber et al., 2013). As to the analysis of residual oil of sealed cores and displaced cores indoor, some methods are employed including of microscopic physical models (Oren and Pinczewski, 1994; Hu and Li, 2000; Wang et al., 2006; Benyamin and Koorosh, 2011), imaging analysis methods (fluorescence analysis) (Liu et al., 2003, 2007; Amber et al., 2013), CT imaging (Schembre and Kovscek, 2003; Beckingham et al., 2013; Liu et al., 2017), nuclear magnetic imaging (Kleinhans et al., 2008; Rezaee et al., 2012; Xiao et al., 2013; Gao and Li, 2015; Meng et al., 2016; Santamarina et al., 2019a,b), scanning electron microscopy (Sun et al., 2005; Desbois et al., 2011; He et al., 2016; Wang et al., 2017), network simulation (Yao et al., 2007). Among them, the CT imaging and fluorescence analysis, especially for the laser confocal technology (Salim et al., 2008; Seifried et al., 2013; Li et al., 2014a; Su et al., 2016), are widely used to characterize the structure of cores and the distribution of oil and water inside the pores.

    • Validation of a stochastic digital packing algorithm for porosity prediction in fluvial gravel deposits

      2015, Sedimentary Geology
      Citation Excerpt :

      Compared with vector-based approaches and for complex shapes, this digital approach is advantageous in several respects. First, there is no conversion or parameterization required, since objects digitized by modern imaging devices, such as X-ray tomography (e.g., Richard et al., 2003) or nuclear magnetic resonance imaging (e.g., Kleinhans et al., 2008), are already in the digital volumetric format required by the packing algorithm. Secondly, collision and overlap detection (normally the most computationally expensive part of packing simulations) are much easier to implement as computer code, and usually faster to execute for complex shapes.

    • Study of the fluid flow characteristics in a porous medium for CO<inf>2</inf> geological storage using MRI

      2014, Magnetic Resonance Imaging
      Citation Excerpt :

      With the intensity method, two-dimensional and three-dimensional image data were evaluated to quantify the porosity profiles and the gas/liquid distributions in packed beds. The concept of the relative permeability for two-phase flow is not new [12–18]. These quantities arise from a generalization of Darcy's law, originally defined for single-phase flow.

    View all citing articles on Scopus
    1

    Present address: Elkerliek Hospital, PO-Box 98, 5700 AB Helmond, The Netherlands.

    View full text