A
team of physicists headed by Prof. Rudolf Hilfer at the Institute for
Computational Physics (ICP) of the University Stuttgart has established a
world record in the field of three-dimensional imaging of porous
materials. The scientists have generated the largest and most precise 3D
image of the pore structure of sandstone. The image was generated
within a project of the Simulation Technology Cluster of Excellence, and
contains more than 35 trillion (a number with thirteen digits) voxels.
It allows researchers to study the relationship between microstructure
and physical properties of porous rocks with unprecedented accuracy.
Sandstones and porous rocks are of paramount importance for applications
such as enhanced oil recovery, carbon dioxide sequestration or
groundwater management.
In
three-dimensional imaging one discretizes spatial structures similar to
digital photographs. Three-dimensional image elements are called
voxels—analogous to pixels for two-dimensional digital photos. The 3D
ICP images systematically resolve the microstructure of a cubic sample
of Fontainebleau sandstone over three decades from submillimeter to
submicron scales. The microstructure of sandstones is important for the
hydraulic properties of many oil reservoirs and thus for efficient
production of hydrocarbons. The largest three-dimensional image, that
the physicists around Hilfer have generated, contains 32768 cubed, or
35184372088832, voxels.
For
comparison: Medical magnetic resonance images of the human contain
roughly 720 million voxel. Even state of the art 3D images in science
and engineering contain only up to 20 billion voxels. Expressed in
digital photos a medical image thus corresponds to only 72 photos. The
largest ICP image, however, with 35 trilion voxels amounts to a stack of
35 million such digital photographs.
“This
world record is important for the physics of porous materials, because
it allows for the first time to investigate extremely complex
microstructures as a function of resolution,” says Hilfer.
The
microstructure of a porous material determines its elastic, plastic,
mechanical, electrical, magnetic, thermal, rheological and hydraulic
properties. Inversely, physicists can infer information about the
microstructure from measuring such physical properties.
Until now it was not possible to image a sample of several centimetres with a resolution of several hundred nanometers.
“To
achieve this size and accuracy would require several years of beam time
at a particle accelerator such as the European Synchrotron Radiation
Facility in Grenoble,” explains Hilfer.
His
team has therefore chosen a different approach. Firstly, the scientists
developed theories and methods that allow to compare and to calibrate
microstructures. Then they invented algorithms and data structures that
allow generating computer models of sufficient size and accuracy. These
models were finally digitized and carefully calibrated against real rock
samples.
