Dataset from "A vertical wide-bore magnetic resonance imaging platform enabling multiphase flow and scale-up studies in chemical and bioreactors" ================================================================================================================================================ DOI of this dataset: https://doi.org/10.15480/882.17602 DOI of corresponding publication: tbd 1. General information ---------------------- * This dataset supports the manuscript above, submitted to Nature Chemical Engineering. It contains the reconstructed magnetic resonance imaging (MRI) data, acquisition parameters, and processing/figure code underlying the figures of the manuscript. * The data were acquired on a vertical, wide-bore, 3.0 T MRI platform with its isocentre approximately 4 m above the floor, which accepts reactors up to 400 mm in outer diameter inserted from below and scanned section by section. The platform resolves phase Distribution and three-dimensional velocity inside reactors non-invasively. * The dataset spans four applications, each in its own set of folders: - a gas-liquid contactor with a structured (corrugated sheet) packing; - gas-solid fluidised beds at 94, 188, and 284 mm inner diameter; - additively manufactured triply periodic minimal surface (TPMS) gyroid packings at two scales (DN 40 and DN 100); - a commercial stirred-tank bioreactor. * Data were acquired at the Institute of Process Imaging, Hamburg University of Technology, Hamburg, Germany. * Authors: Stefan Benders, Muhammad Adrian, Till Lenczyk, Melis Ozdemir, Hannah S. Rennebaum, Joseph E. McPeak, M. Raquel Serial (Institute of Process Imaging, Hamburg University of Technology, Hamburg, Germany); Johan Overweg, Christian Stehning, Christoph Leussler, Peter Boernert (Philips Research Labs Hamburg, Hamburg, Germany); Gerrit Vissers, Paul Sanders, Hans Heesterbeek, Jurgen Mollink (Philips Healthcare, Best, The Netherlands); Hans van Oort, Dan Lovell, Lorenzo Gorini, Steve Bates, Ben Leigh (Tesla Engineering, Storrington, United Kingdom); Martijn Lunenburg, Catalina Arteaga, Martino Borgo (Tesla Dynamic Coils, Zaltbommel, The Netherlands); Martino Borgo (Futura Composites B.V., Heerhugowaard, The Netherlands); Markus Weiger, Franciszek Hennel (Institute for Biomedical Engineering, ETH Zurich and University of Zurich, Zurich, Switzerland); Marko Hoffmann, Michael Schlueter (Institute of Multiphase Flows, Hamburg University of Technology, Hamburg, Germany); Raimund Horn (Institute of Chemical Reaction Engineering, Hamburg University of Technology, Hamburg, Germany); Alexander Penn (Institute of Process Imaging, Hamburg University of Technology, Hamburg, Germany, *Corresponding author). * Corresponding author: Alexander Penn (alexander.penn@tuhh.de, ORCID 0000-0001-5596-6310). 2. Data overview ---------------- * Reconstructed MRI images are provided in HDF5 (.h5) format, prior to postprocessing, so that the reported analyses can be reproduced. Raw k-space data and the manufacturer's proprietary reconstruction are not included; image reconstruction was performed on the Philips scanner platform. * Each folder corresponds to one figure or figure group of the manuscript and contains, where applicable: a /data subfolder (reconstructed images as .h5), a /flow subfolder (velocity components and derived shear-rate fields, for the velocimetry datasets), a /params subfolder (plain-text parameter summary and the original scanner .par files), a /code subfolder (processing and figure code), and a dataset-level README. * Folder structure: fig4_packing_trickleflow/ Structured packing, Figures 4a to 4d (data + code) fig4_fluidisedbed_188mm/ Fluidised bed 188 mm, Figures 4e to 4f (data + code) fig5_GyroidTPnS_flow Gyroid DN 40 and DN 100, Figure 5 (data; code to follow) fig6_stirredtank/ Stirred tank, Figure 6 (raw data; .h5 data and code to follow) extdata_fig2_fluidisedbed_94_188_284/ Fluidised beds 94/188/284 mm, Ext. Data Fig. 2 (data + code) * Acquisition parameters are documented in the manuscript (in the Methods, or in Extended Data Table 4 for the 94/188/284 mm multi-scale set) and, in machine-readable form, in the /params subfolder of each figure folder. 3. Data collection methods --------------------------- * All data were acquired on the vertical wide-bore 3.0 T MRI platform described in the manuscript, using two-dimensional gradient-echo (GRE) sequences. Velocimetry datasets used a GRE sequence with bipolar flow-encoding gradients. Signal was detected with application-specific receive coils (a home-built birdcage coil, clip-on surface coils, and 15- or 32-channel receive arrays), as stated per dataset. * Structured packing (Figures 4a to 4d): a corrugated-sheet packing in a PMMA column, imaged fully submerged and under trickle-flow conditions; the difference of the two images gives the phase distribution map (Figure 4d). MRI-active phase: aqueous copper(II) sulfate (CuSO4) solution. * Fluidised beds (Figure 4f and Extended Data Figure 2): PMMA columns filled with poppy seeds and fluidised with air; dynamic image time series capture rising gas bubbles. Imaged at 94, 188, and 284 mm inner diameter. * Gyroid packings (Figure 5): TPMS gyroid structures (DN 40 and DN 100) with a developed inlet flow; three-dimensional liquid velocity fields were measured. MRI-active phase: aqueous CuSO4 solution. * Stirred tank (Figure 6): an MRI-compatible Ambr250 bioreactor; Velocity fields were measured. MRI-active phase: aqueous CuSO4 solution. * Full acquisition parameters and operating conditions are given in the manuscript and in the per-folder /params subfolders. 4. File formats --------------- * Reconstructed images: HDF5 (.h5), one spin-density (or complex/phase, for velocimetry) array per file, with voxel size, axis order, slice information, and temporal resolution stored as dataset attributes. The internal layout is described in each folder's README. * Velocity and shear-rate fields (velocimetry datasets): stored as labelled arrays with stated units (mm/s for velocity) and an explicit sign convention and coordinate frame, described in the corresponding README. * Parameters: plain-text summaries (.txt) plus the original scanner parameter files (.par). * Figures in the manuscript were composed in Affinity from the panels the code produces; the code reproduces the individual data panels, not the assembled plates. 5. Dependencies and tools ------------------------- * Processing and figure code is written in Python. The exact library versions are declared per code folder (requirements.txt or environment.yml), together with the Python version each script was tested with. * The finite-element model used to set the receive-array overlap is included where provided; executing it requires a commercial COMSOL Multiphysics licence. * Image reconstruction was performed with the Philips scanner platform's proprietary software, which cannot be redistributed. The images provided here are supplied after reconstruction so that the reported analyses can be reproduced without it. 6. Limitations and known issues ------------------------------- * For the multiphase datasets (structured packing, fluidised beds), the gas phase is inferred from signal absence; it is not velocity-encoded and is not quantified as a void fraction. * The mass-balance, divergence-freeness, and vorticity checks, and the detailed comparison against computational fluid dynamics, for the DN 40 gyroid data are reported in the companion publication (Merbach et al.), not in this dataset. * Each application represents a single operating point per acquisition; the dataset is not a parametric study. * Working fluids were selected for MRI compatibility (doped aqueous solutions, poppy seeds) and are surrogates for process media. * At the time of writing, the DN 40, DN 100, and stirred-tank folders contain data only; the corresponding processing code is added separately. Check the per-folder README for the current contents. 7. Usage rights and citation ---------------------------- * This dataset is published under the CC BY 4.0 licence. * The dataset can be cited by the DOI provided by TORE. * By using this data in your work, you agree to cite the corresponding publication.