3D deep convolutional neural network segmentation model for precipitate and porosity identification in synchrotron X-ray tomograms.

New developments at synchrotron beamlines and the ongoing upgrades of synchrotron facilities allow the possibility to study complex structures with a much better spatial and temporal resolution than ever before. However, the downside is that the data collected are also significantly larger (more than several terabytes) than ever before, and post-processing and analyzing these data is very challenging to perform manually. This issue can be solved by employing automated methods such as machine learning, which show significantly improved performance in data processing and image segmentation than manual methods. In this work, a 3D U-net deep convolutional neural network (DCNN) model with four layers and base-8 characteristic features has been developed to segment precipitates and porosities in synchrotron transmission X-ray micrograms. Transmission X-ray microscopy experiments were conducted on micropillars prepared from additively manufactured 316L steel to evaluate precipitate information. After training the 3D U-net DCNN model, it was used on unseen data and the prediction was compared with manual segmentation. A good agreement was found between both segmentations. An ablation study was performed and revealed that the proposed model showed better statistics than other models with lower numbers of layers and/or characteristic features. The proposed model is able to segment several hundreds of gigabytes of data in a few minutes and could be applied to other materials and tomography techniques. The code and the fitted weights are made available with this paper for any interested researcher to use for their needs (https://github.com/manasvupadhyay/erc-gamma-3D-DCNN).

Achromatic Non-Interferometric Single Grating Neutron Dark-Field Imaging.

We demonstrate a simple single grating beam modulation technique, which enables the use of a highly intense neutron beam for differential phase and dark-field contrast imaging and thus spatially resolved structural correlation measurements in full analogy to interferometric methods. In contrast to these interferometric approaches our method is intrinsically achromatic and provides unprecedented flexibility in the choice of experimental parameters. In particular the method enables straight forward application of quantitative dark-field contrast imaging in time-of-flight mode at pulsed neutron sources. Utilizing merely a macroscopic absorption mask unparalleled length scales become accessible. We present results of quantitative dark-field contrast imaging combining microstructural small angle scattering analyses with real space imaging for a variety of materials.

Chasing quantitative biases in neutron imaging with scintillator-camera detectors: a practical method with black body grids.

We propose a method for improving the quantification of neutron imaging measurements with scintillator-camera based detectors by correcting for systematic biases introduced by scattered neutrons and other sources such as light reflections in the detector system. This method is fully experimental, using reference measurements with a grid of small black bodies (BB) to measure the bias contributions directly. Using two test samples, one made of lead alloy and having a moderate (20%) neutron transmission and one made of stainless-steel and having a very low (1%) transmission, we evaluated the improvement brought by this method in reducing both the average quantification bias and the uncertainty around this average bias after tomographic reconstruction. The results show that a reduction of the quantification bias of up to one order of magnitude can be obtained. For moderately transparent samples, little sensitivity is observed to the parameters used for the correction. For the more challenging sample with very low transmission, a correct placement of the BB grid is of utmost importance for a successful correction.

Characterizing pearls structures using X-ray phase-contrast and neutron imaging: a pilot study.

Some cultured and natural pearls can be reliably distinguished by visual inspection and by the use of lens and microscope. However, assessing the origin of the pearls could be not straightforward since many different production techniques can now be found in the pearl market, for example in salt or freshwater environments, with or without a rigid nucleus. This wide range of products requires the use of new effective scientific techniques. Indeed, X-ray radiography has been used by gemologists since last century as the only safe and non-destructive way to visually inspect the interior of a pearl, and recently, also X-ray computed micro-tomography was used to better visualize the inner parts of the gems. In this study we analyzed samples of natural and cultured pearls by means of two non-destructive techniques: the X-ray Phase-Contrast Imaging (PCI) and the Neutron Imaging (NI). PCI and NI results will be combined for the first time, to better visualize the pearls internal morphology, thus giving relevant indications on the pearl formation process.

A Monte Carlo approach for scattering correction towards quantitative neutron imaging of polycrystals.

The development of neutron imaging from a qualitative inspection tool towards a quantitative technique in materials science has increased the requirements for accuracy significantly. Quantifying the thickness or the density of polycrystalline samples with high accuracy using neutron imaging has two main problems: (i) the scattering from the sample creates artefacts on the image and (ii) there is a lack of specific reference attenuation coefficients. This work presents experimental and simulation results to explain and approach these problems. Firstly, a series of neutron radiography and tomography experiments of iron, copper and vanadium are performed and serve as a reference. These materials were selected because they attenuate neutrons mainly through coherent (Fe and Cu) and incoherent (V) scattering. Secondly, an ad hoc Monte Carlo model was developed, based on beamline, sample and detector parameters, in order to simulate experiments, understand the physics involved and interpret the experimental data. The model, developed in the McStas framework, uses a priori information about the sample geometry and crystalline structure, as well as beamline settings, such as spectrum, geometry and detector type. The validity of the simulations is then verified with experimental results for the two problems that motivated this work: (i) the scattering distribution in transmission imaging and (ii) the calculated attenuation coefficients.

Real-Time Neutron Imaging to Detect Origin of Blocking in Drug Injection Devices.

Nondestructive testing is a common method for root cause investigations of malfunction of mechanical devices, for example, medical devices for drug dose delivery. Radiography is a method that has the advantage that it is possible to see through the sample. In this work we are using neutron radiography to observe drug distribution in drug injection devices during the injection process and as post-injection examination. Using neutrons it is possible to show small amounts of liquid in capillaries, and foam bubbles are shown with great contrast compared to metal and glass. The investigation has two parts optimized for high spatial and high temporal resolution, respectively. Using high spatial resolution it is possible to resolve the thin films of drug product in foam and even to detect the drug residues in the injection needle. Switching to high temporal resolution we demonstrate that it is possible to follow the injection process. Spatio-temporal data sets of the injection process were acquired using remotely triggered injection devices and a camera allowing sub-second frame rates.The motion analysis required the application of an edge-preserving spatio-temporal denoising filter to improve the signal-to-noise ratio. After filtering it is possible to detect relevant edges and extract motion curves from the spatio-temporal data. LAY ABSTRACT: Neutron imaging is a nondestructive method based on radiography using neutrons and is suitable for detecting small amounts of aqueous liquids even in metallic casing/sheath/tubing. This property has here been used to visualize the distribution of a drug product in a syringe needle of a drug injection device. In the static case the method clearly showed the difference between needles that were empty, full, or contained a mix of gas and liquid. A second investigation was aimed to visualize the dynamic behavior of an auto-injector device. In this experiment the imaging method was capable of following the injection phase of the device. By analyzing the acquired images in time and space it is possible to measure the injection velocity curves of the piston and drug, respectively.

Visualizing the morphology of vortex lattice domains in a bulk type-II superconductor.

Alike materials in the solid state, the phase diagram of type-II superconductors exhibit crystalline, amorphous, liquid and spatially inhomogeneous phases. The multitude of different phases of vortex matter has thence proven to act as almost ideal model system for the study of both the underlying properties of superconductivity but also of general phenomena such as domain nucleation and morphology. Here we show how neutron grating interferometry yields detailed information on the vortex lattice and its domain structure in the intermediate mixed state of a type-II niobium superconductor. In particular, we identify the nucleation regions, how the intermediate mixed state expands, and where it finally evolves into the Shubnikov phase. Moreover, we complement the results obtained from neutron grating interferometry by small-angle neutron scattering that confirm the spatially resolved morphology found in the intermediate mixed state, and very small-angle neutron scattering that confirm the domain structure of the vortex lattice.

Quantification of the sensitivity range in neutron dark-field imaging.

In neutron grating interferometry, the dark-field image visualizes the scattering properties of samples in the small-angle and ultra-small-angle scattering range. These angles correspond to correlation lengths from several hundred nanometers up to several tens of micrometers. In this article, we present an experimental study that demonstrates the potential of quantitative neutron dark-field imaging. The dark-field signal for scattering from different particle sizes and concentrations of mono-dispersive polystyrene particles in aqueous solution is compared to theoretical predictions and the good agreement between measurements and calculations underlines the quantitative nature of the measured values and reliability of the technique with neutrons.

Dual spectrum neutron radiography: identification of phase transitions between frozen and liquid water.

In this Letter, a new approach to distinguish liquid water and ice based on dual spectrum neutron radiography is presented. The distinction is based on arising differences between the cross section of water and ice in the cold energy range. As a significant portion of the energy spectrum of the ICON beam line at Paul Scherrer Institut is in the thermal energy range, no differences can be observed with the entire beam. Introducing a polycrystalline neutron filter (beryllium) inside the beam, neutrons above its cutoff energy are filtered out and the cold energy region is emphasized. Finally, a contrast of about 1.6% is obtained with our imaging setup between liquid water and ice. Based on this measurement concept, the temporal evolution of the aggregate state of water can be investigated without any prior knowledge of its thickness. Using this technique, we could unambiguously prove the production of supercooled water inside fuel cells with a direct measurement method.

Development of a cold-neutron imaging detector based on thick gaseous electron multiplier.

We present the results of our recent studies on a cold-neutron imaging detector prototype based on THick Gaseous Electron Multiplier (THGEM). The detector consists of a thin Boron layer, for neutron-to-charged particle conversion, coupled to two THGEM electrodes in cascade for charge amplification and a position-sensitive charge-readout anode. The detector operates in Ne∕(5%)CF4, at atmospheric pressure, in a stable condition at a gain of around 10(4). Due to the geometrical structure of the detector elements (THGEM geometry and charge read-out anode), the image of detector active area shows a large inhomogeneity, corrected using a dedicated flat-filed correction algorithm. The prototype provides a detection efficiency of 5% and an effective spatial resolution of the order of 1.3 mm.

The study of bronze statuettes with the help of neutron-imaging techniques.

Until recently fabrication techniques of Renaissance bronzes have been studied only with the naked eye, microscopically, videoscopically and with X-radiography. These techniques provide information on production techniques, yet much important detail remains unclear. As part of an interdisciplinary study of Renaissance bronzes undertaken by the Rijksmuseum Amsterdam, neutron-imaging techniques have been applied with the aim of obtaining a better understanding of bronze workmanship during the Renaissance period. Therefore, an explanation of the fabrication techniques is given to better understand the data collected by these neutron-imaging techniques. The data was used for tomography studies, which reveal hidden aspects that could not at all or scarcely be seen using X-radiography. For this specific study, the representative bronze 'Hercules Pomarius' of Willem van Tetrode (ca 1520-1588) has been examined, along with 20 other Renaissance bronzes from the Rijksmuseum collection.

Hydraulic contacts controlling water flow across porous grains.

Water flow between porous grains varies widely depending on the water distribution in contacts between grains. The hydraulic behavior of contacts varies from highly conductive when water fills the contacts to a bottleneck to flow as water pressure drops and contact asperities rapidly drain. Such changes greatly impact the hydraulic conductivity of porous grain packs such as aggregated soil. The dominant driving force of water flow across contacts is capillarity, often quantified relative to gravity and viscous forces using the capillary and Bond numbers. For fast water infiltration, viscous forces dominate. For simplicity we modeled the water distribution between spherical porous grains whose surfaces are covered by spherical bumps of much smaller radii. We provide experimental evidence obtained by neutron radiography and synchrotron-based x-ray tomographic microscopy documenting transitions in the flow behavior across contacts.

3D imaging of microstructure of spruce wood.

Synchrotron radiation phase-contrast X-ray tomographic microscopy (srPCXTM) was applied to observation and identification of the features of spruce anatomy at the cellular lengthscale. The pilot experiments presented in the paper clearly revealed the features of the heartwood of Spruce (Picea abies [L.] Karst.), such as lumina and pits connecting the lumina, with a theoretical voxel size of 0.7 x 0.7 x 0.7 microm(3). The experiments were carried out on microspecimens of heartwood, measuring approximately 200 by 200 micrometers in cross-section. The technique for production and preparation of wood microsamples was developed within the framework of this investigation. The total porosity of the samples was derived and the values of the microstructural parameters, such as the diameters of tracheid, cell wall thicknesses and pit diameters were assessed non-invasively. Microstructural features as thin/small as approximately 1.5 microm were revealed and reconstructed in 3D. It is suggested that the position of sub-voxel-sized features (such as position of tori in the bordered pit pairs) can be determined indirectly using watershed segmentation. Moreover, the paper discusses the practical issues connected with a pipelined phase-contrast synchrotron-based microtomography experiment and the possible future potentials of this technique in the domain of wood science.

Transcatheter closure of atrial septal defects under echocardiographic guidance without X-ray: initial experiences.

BACKGROUND: Transcatheter closure of atrial septal defects is performed under fluoroscopy, but echocardiography has gained an important role in the procedure. With the new Amplatzer Septal Occluder a device has become available which is easy to implant with minimal fluoroscopy time. We developed an interventional procedure with this device under transesophageal echocardiography alone without fluoroscopy. METHODS AND RESULTS: Four patients (3 to 16 years of age, bodyweight 14 to 60 kg) with atrial septal defects centrally located in the oval fossa were elected for transcatheter closure. After sedation with midazolam and propofol a diagnostic and interventional catheterization was performed in all cases without fluoroscopy. Oxymetric shunt was Qp: Qs = 1.7 (1.5 to 2.1). Under transesophageal echocardiography, the defects were sized over the wire with a balloon catheter. Mean balloon stretched diameter was 10 mm (7 to 14 mm). Under transesophageal echocardiography an Amplatzer Septal Occluder was placed into the defect. In two patients this was achieved with a 5 MHz monoplane pediatric transducer, in two patients a 10mm 5 MHz multiplane probe was used. Complete closure was achieved in all patients and no complications were encountered. CONCLUSION: We conclude that in selected cases with an atrial septal defect located in the oval fossa and clear-cut echocardiographic findings, an Amplatzer Septal Occluder can be safely deployed under echocardiographic guidance alone.