Detecting anomalies in X-ray diffraction images using convolutional neural networks.

Our understanding of life is based upon the interpretation of macromolecular structures and their dynamics. Almost 90% of currently known macromolecular models originated from electron density maps constructed using X-ray diffraction images. Even though diffraction images are critical for structure determination, due to their vast amounts and noisy, non-intuitive nature, their quality is rarely inspected. In this paper, we use recent advances in machine learning to automatically detect seven types of anomalies in X-ray diffraction images. For this purpose, we utilize a novel X-ray beam center detection algorithm, propose three different image representations, and compare the predictive performance of general-purpose classifiers and deep convolutional neural networks (CNNs). In benchmark tests on a set of 6,311 X-ray diffraction images, the proposed CNN achieved between 87% and 99% accuracy depending on the type of anomaly. Experimental results show that the proposed anomaly detection system can be considered suitable for early detection of sub-optimal data collection conditions and malfunctions at X-ray experimental stations.

Photocatalytic oxidation of nitric oxide over AgNPs/TiO2-loaded carbon fiber cloths.

Series of AgNPs/TiO2-loaded carbon fiber cloth (CFC) composites were prepared by incorporation of pristine TiO2 and three AgNPs-modified TiO2 additives onto the surface of four commercial CFCs. AgNPs/TiO2 photocatalysts were synthesized by the wet impregnation method, including NaBH4 reduction of silver ions. The silver content in the modified photocatalyst was assessed by inductively coupled plasma optical emission spectrometry (ICP-OES) as well as XRD analysis. It can be indicated that silver was successfully reduced to Ag nanoparticles what was confirmed by UV-Vis/DRS as well as XRD methods. The photocatalytic activity of the AgNPs/TiO2-loaded CFCs was evaluated during the photocatalytic oxidation (PCO) tests of nitric oxide (NO) acting as a model air contaminant under UV light. It was found that the highest NO removal rate was observed for the AgNPs/TiO2-loaded CFC material containing 3.70 wt% of AgNPs. Modification of TiO2 with AgNPs stabilized the photocatalytic efficiency of the composites during 5 as well as 24 consecutive NO photooxidation cycles. It was also concluded that the presence of AgNPs was a key factor responsible for hindering NO2 formation.

Cavity ring-down absorption spectrography based on filament-generated supercontinuum light.

We performed simultaneous, multispectral CRDS measurements that for the first time use the Supercontinuum light source. We called this approach Supercontinuum Cavity Ring-Down Spectrography (SC CRDSpectrography) and successfully applied it to measuring the absorption spectrum of NO2 gas at a concentration of 2 ppm. The extrapolated sensitivity of our setup was much greater, about 5 ppb. The ppb sensitivity level is comparable to this obtainable with single wavelength dye-lasers based CRDS systems. It is, therefore, feasible to construct extremely broadband and sensitive CRDS devices basing on the SC CRDSpectrography scheme.

Investigations on the Properties and Performance of Mixed-Matrix Polyethersulfone Membranes Modified with Halloysite Nanotubes.

Ultrafiltration (UF) polyethersulfone (PES) membranes were prepared by wet phase inversion method. Commercial halloysite nanotubes (HNTs) in the amount of 0.5-4 wt % vs PES (15 wt %) were introduced into the casting solution containing the polymer and N,N-dimethylformamide as a solvent. The morphology, physicochemical properties and performance of the membranes were characterized by scanning electron microscopy (SEM) and atomic force microscopy (AFM), zeta potential, porosity and contact angle analyses, as well as permeability measurements. Moreover, the antifouling properties of the membranes were evaluated during UF of a model solution of bovine serum albumin (BSA). The research revealed a positive influence of modification with HNTs on hydrophilicity, water permeability and antifouling properties of the PES membranes. The most significant improvement of permeability was obtained in case of the membrane containing 2 wt % of HNTs, whereas the highest fouling resistance was observed for 0.5 wt % HNTs content. It was found that a good dispersion of HNTs can be obtained only at loadings below 2 wt %. Based on the results a relation between severity of membrane fouling and surface roughness was proved. Moreover, an increase of the roughness of the modified membranes was found to be accompanied by an increase of isoelectric point values.

Porous carbon material containing CaO for acidic gas capture: preparation and properties.

A one-step process for the preparation of CaO-containing porous carbons is described. Mixtures of poly(ethylene terephthalate) with natural limestone were pyrolyzed and thus hybrid sorbents could be easily obtained. The polymeric material and the mineral served as a carbon precursor and CaO delivering agent, respectively. We discuss effects of the preparation conditions and the relative amounts of the raw materials used for the preparations on the porosity of the hybrid products. The micropore areas and volumes of the obtained products tended to decrease with increasing CaO contents. Increase in the preparation temperature entailed a decrease in the micropore volume, whereas the mesopore volume increased. The pore creation mechanism is proposed on the basis of thermogravimetric and temperature-programmed desorption measurements. The prepared CaO-containing porous carbons efficiently captured SO2 and CO2 from air. Washing out of CaO from the hybrid materials was confirmed as a suitable method to obtain highly porous carbon materials.