A new flow-through reaction cell consisting of an X-ray-transparent semicrystalline thermoplastic has been developed for percolation
experiments. Core holder, tubing and all confining parts are constructed using
PEEK (polyetheretherketone) to allow concomitant surveillance of the
reaction progress by X-ray microtomography (
The combination of this flow-through reaction cell setup with a laboratory
X-ray
Fluid–rock interactions govern the critical mass transfers involved in the exchange between the oceanic crust and seawater and the formation of hydrothermal deposits in a range of settings. They also influence hydraulic and rock mechanical properties and are hence relevant to applied geosciences. Understanding these processes requires consideration of multiple factors, including changes in porosity distribution, permeability and their relationship with critical dissolution and precipitation rates. To date, the knowledge of these interdependencies is very limited, and we are hence often unable to make reliable predictions of fluid–rock behaviour in natural systems.
In numerical simulations of hydrothermal systems, fluid flow, heat- and
matter transport, and chemical reaction progress have to be considered
simultaneously (e.g., Kühn, 2009). The first step in understanding
macroscopic rock properties, such as permeability or capillary pressure, is
the correlation of rock microstructure and physical properties of the fluid
at the pore-scale. A crucial part in this is the transformation of real pore
space geometries into appropriate pore space models. In this regard, true 3-D
spatially resolved spectroscopic data (e.g., nuclear magnetic resonance;
laboratory-based and synchrotron-based computed microtomography:
In a number of studies, reaction progress in hydrothermal fluid–rock
interaction experiments was investigated by means of lab-based or
synchrotron
All former studies have in common, however, that the samples were only
scanned twice, prior to confinement within the reaction cell and at the end
of the experiments after a complete dismantling. Reaction concomitant scans
were not possible, either because non-transparent X-ray reactor materials
were used or because of impracticabilities in introducing the pressure
housing into the beam chamber. In case of 4-D
In each of these studies, the core was removed from its pressure housing
prior to scanning so that disturbance of the core during dismantling/remounting for repeated scanning was an issue. This problem can be avoided,
if the sample stays mounted for the scan. Likewise, image registration of
subsequent scans can be simplified that way. Currently, cell materials with
a low X-ray attenuation coefficient that allow the scan of a mounted sample,
are deployed to image important hydrodynamic phenomena on the pore-scale
(e.g., two-phase flow in porous media). Since high-speed tomography (even
sub-second) has become possible at synchrotron facilities, in situ drainage/imbibition or two-phase flow in sample cores have been imaged in real-time
(e.g., Myers et al., 2011: air-drainage of water-saturated Bentheim
sandstone, core
Here, we present a new flow-through reaction cell consisting of an X-ray-transparent semicrystalline thermoplastic that has been developed for
(long-term) percolation experiments to allow concomitant surveillance of the
reaction progress by X-ray microtomography without dismantling the sample.
The cell size facilitates its use in standard CT scanner models, while the
core size is large enough to permit investigation of intricate reaction
textures (e.g., in the course of serpentinization reactions) or complex
rock analogues (e.g., to study metasomatic reactions along strong contrasts
in chemical potential across different lithologies). The unique material
properties of the weakly attenuating, but mechanically strong, cell material
allow the experimental examination of rock–fluid interactions under low to
moderate temperatures (up to 200
X-ray-transparent PEEK flow-through reaction cell.
The semicrystalline thermoplastic material (polyetheretherketone: PEEK) is
commercially available (KTK Kunststofftechnik Vertriebs GmbH, Germering,
Germany). Since PEEK is machinable, withstands pressure and has a high
mechanical strength and hardness over a broad temperature range, an external
metal pressure vessel housing is not necessary in the experimental setup.
Instead, 20 mm thick PEEK walls are sufficient to safely maintain pressures up to 100 bar (10 MPa) and temperatures up to 200
The reaction cell can accommodate rock cylinders of 19 mm diameter and up to
50 mm length that are mantled with FEP (fluoroethylene propylene; Adtech
Polymer Engineering ltd., Stroud, UK) heat-shrink sleeves. PEEK capillary
tubing (Upchurch Scientific,
Scheme of the flow line. All tubing, valves etc. are made of PEEK, the pins of the needle valves are made of titanium.
The flow-through system (Fig. 1c; schematic flow diagram in Fig. 2) uses two
injection pumps (Teledyne Isco D-Series syringe pumps), which can be run
either independently of each other or in a synchronized mode. All pumps are
of a pulsation-free, positive-displacement pump type and each offers an
injection capacity of 500 cm
During the ongoing percolation experiment, the percolation cell is located inside the oven, and recharge- and discharge fluids are connected by PEEK capillary tubing (Fig. 1c). The pump system, recharge- and discharge fluid storage, and the board with the flow line are installed on a mobile rack. To facilitate the transfer of the percolation experiment from the oven laboratory to the X-ray microtomography laboratory for scanning, the mobile rack is equipped with an uninterruptible power supply.
The following samples have been chosen for initial test runs aimed at
demonstrating the unbiased X-ray observations of temporal changes in rock
samples mounted in shrink sleeves within the percolation cell. To assess
differing states of pore fillings (air vs. fluid), a quartz-dominated
sandstone (locality: Gildehaus, Romberg quarry, Germany; formation:
Valangin, Lower Cretaceous) has been partially saturated with water. To make
a sandstone sample with both water- and air-filled pore spaces, we first
flooded the sample completely with water at 50 bar fluid pressure, and then
de-pressurized and disconnected the tubing to permit evaporation towards the
state we eventually scanned. To demonstrate the sensitivity of in situ cell
X-ray observations with respect to subtle differences in mineralogy and
delicate textures, we mounted and scanned a mineralogically diverse tertiary
sandstone (“Idaho Gray sandstone”, Idaho, USA; Idaho formation, tertiary).
This sandstone was previously characterized by M. Halisch and S. Kaufhold (personal communication, 2015)
as coarse grained sandstone with more than 75 % of the
grains ranging between 0.35 and 1.1 mm in size. It consists mostly of quartz
(
As a test bed for a 4-D study, the conversion of gypsum single crystals
(selenite, Morocco; Mineraliengrosshandel Hausen GmbH, Telfs, Austria) to
anhydrite has been investigated by concomitant
The X-ray microtomography scans aimed at assessing the applicability of the
percolation cell were performed using a CT-ALPHA system (ProCon, Germany) at
the Department of Geosciences, University of Bremen, Germany. The sandstone
micro-cores mounted in shrink sleeves inside the percolation cell (Fig. 1)
were scanned with a beam intensity of 100 kV, an energy flux of 300
Our results demonstrate that the placement of the micro-cores in the PEEK
flow-through cell setup, mounted in shrink sleeves, does not bias the
results of X-ray investigation and allows the concomitant surveillance of
hydrothermal mineral–fluid interactions by repeated
Reconstructed images of a partially saturated, quartz-rich
sandstone sample (Gildehaus, Romberg quarry;
The distinction between air-filled and fluid-filled pores of the partially
water-saturated quartz-rich Gildehaus sandstone (Fig. 3) is straightforward.
Air-filled pores (black in Fig. 3) can easily be distinguished from
fluid-filled pores (dark gray) in the unfiltered image (Fig. 3a, voxel size
is 8.95
The reconstructed images of Idaho Gray sandstone sample (
Reconstructed images of a micro-core of Idaho Gray sandstone (
To assess the experimental performance of the flow-through system in terms
of the feasibility of repeated, concomitant scans of long-term percolation
experiments, we monitored the gypsum/anhydrite transition as a case study
of a 4-D fabric evolution. Two percolation experiments were performed
(experiment FTGy-A for 77 days and experiment FTGy-B for 35 days) using
artificially fractured selenite crystals that were subjected to a partially
gypsum-saturated fluid (1.5 g L
4-D series of the conversion from gypsum to anhydrite (110
By concomitant
A comparison of fabric and mineralogy produced in both selenite percolation experiments provides additional information on the kinetic impact of different cooling rates on the mineralogy of the final run product. While experiment FTGy-A underwent fast cooling (< 10 min) from oven to room temperature before scanning, run FTGy-B was subjected to a slow cooling process (to room temperature within ca. 60 to 80 min) before each scan. The final fabric in both experiments features densely intergrown anhydrite needles separated by fluid-filled intergranular space. The picture obtained after cooling depends on the cooling rate: the newly formed anhydrite is apparently preserved after cooling in the fast-cooling run FTGy-A. In contrast, the slow cooling of run FTGy-B resulted in an extensive secondary conversion of the newly formed anhydrite needles to gypsum.
The X-ray-transparent reaction cell setup presented in this work holds great
potential for fostering the knowledge about the interdependencies between
changes in porosity distribution, permeability and their feedback
relationship in the course of dissolution and precipitation reactions.
the consideration of true 3-D spatially resolved the identification of preferred growth- or dissolution sites can be
correlated to preferential fluid pathways; analysis of changing fluid compositions, changing mineral chemistry and
volume proportions, can be used for the parametrization of coupled
reaction–transport models; the 4-D evolution of rock fabric and mineralogy, in particular the
dissolution or precipitation rate spectra derived from the experiment by
The results presented in this communication suggest that PEEK as the cell
construction material is easily processable, mechanically stable, and
generally well-suited for X-ray imaging purposes. The straightforward cell
design presented here is capable of manifold adaptations to specific
experimental requirements in order to mimic a plethora of environmental and
geotechnical conditions.
The following additional figures and table have been submitted as
supplementary material with this manuscript: Figure S1 – Collage of photographs
featuring the components of the cell. The labels correspond to the positions in Table S1; Figure S2 – Documentation of formation and evolution of flow paths
in the course of gypsum dissolution and anhydrite growth by concomitant
We are grateful to the people that have contributed to this development: first of all to Georg Nover for his qualified and cordial discussions concerning cell construction details. We thank Norbert Schleifer & Volker Fendrich and Robert Hinkes & Volker Feeser for helpful insights into their laboratory materials and methods. We are greatly indebted to Elke Sorgenicht, forewoman of the mechanical workshop, and her staff for their brilliant technical realization. We will remember Matthias Lange and highly value his input on flow path issues and Labview programming. Thanks also go to Martin Kölling and Katja Beier for discussions in the early phase of this project. Matthias Halisch and Stephan Kaufhold are acknowledged for providing both samples and characterizations of Idaho sandstone, Norbert Schleifer is thanked for providing an aliquot of Gildehaus sandstone. We gratefully acknowledge the reviews of G. Nover and H. Steeb.
This study was funded through a grant of the DFG to WB within a Reinhart-Koselleck Project BA 1605/10-1. The article processing charges for this open-access publication were covered by the University of Bremen. Edited by: M. Halisch