Search

Extensive efforts to gather materials data have largely overlooked potential data redundancy. In this study, we present evidence of a significant degree of redundancy across multiple large datasets for various material properties, by revealing that up to 95% of data can be safely removed from machine learning training with little impact on in-distribution prediction performance. The redundant data is related to over-represented material types and does not mitigate the severe performance degradation on out-of-distribution samples. In addition, we show that uncertainty-based active learning algorithms can construct much smaller but equally informative datasets. We discuss the effectiveness of informative data in improving prediction performance and robustness and provide insights into efficient data acquisition and machine learning training. This work challenges the "bigger is better" mentality and calls for attention to the information richness of materials data rather than a narrow emphasis on data volume., (© 2023. The Author(s).)

The pressure to use sustainable materials and adopt practices reducing the carbon footprint of the construction industry has risen. Such materials include recycled concrete aggregates (RCA) made from waste concrete. However, concrete made with RCA often presents poor fresh and hardened properties along with a decrease in its durability performance, especially when using its fine fraction (i.e., FRCA). Most studies involving FRCA use direct replacement methods (DRM) to proportion concrete although other techniques are available such as the Equivalent Volume (EV) and Particle Packing Models (PPMs); yet their impact on the durability performance, especially its performance against freezing and thawing (F/T), remains unknown. This work, therefore, appraises the F/T resistance of FRCA mixtures proportioned through various mix proportioning techniques (i.e., DRM, EV and PPMs), produced with distinct crushing processes (i.e., crusher's fines vs. finely ground). The results show that the mix design technique has a significant influence on the FRCA mixture's F/T resistance where PPM-proportioned mixtures demonstrate the best overall performance, exceeding the specified requirements while DRM-proportioned mixtures failed F/T resistance requirements. Moreover, the crushing process plays an important role in the recycled mixtures' cracking behavior under F/T cycles, where less processing leads to fewer cracks while remaining the most sustainable option overall.

Thermal energy storage is an important component in energy units to decrease the gap between energy supply and demand. Free convection and the locations of the tubes carrying the heat-transfer fluid (HTF) have a significant influence on both the energy discharging potential and the buoyancy effect during the solidification mode. In the present study, the impact of the tube position was examined during the discharging process. Liquid-fraction evolution and energy removal rate with thermo-fluid contour profiles were used to examine the performance of the unit. Heat exchanger tubes are proposed with different numbers and positions in the unit for various cases including uniform and non-uniform tubes distribution. The results show that moving the HTF tubes to medium positions along the vertical direction is relatively better for enhancing the solidification of PCM with multiple HTF tubes. Repositioning of the HTF tubes on the left side of the unit can slightly improve the heat removal rate by about 0.2 in the case of p5-u-1 and decreases by 1.6% in the case of p5-u-2. It was found also that increasing the distance between the tubes in the vertical direction has a detrimental effect on the PCM solidification mode. Replacing the HTF tubes on the left side of the unit negatively reduces the heat removal rate by about 1.2 and 4.4%, respectively. Further, decreasing the HTF temperature from 15 °C to 10 and 5 °C can increase the heat removal rate by around 7 and 16%, respectively. This paper indicates that the specific concern to the HTF tube arrangement should be made to improve the discharging process attending free convection impact in phase change heat storage.

A wavy shape was used to enhance the thermal heat transfer in a shell-tube latent heat thermal energy storage (LHTES) unit. The thermal storage unit was filled with CuO-coconut oil nano-enhanced phase change material (NePCM). The enthalpy-porosity approach was employed to model the phase change heat transfer in the presence of natural convection effects in the molten NePCM. The finite element method was applied to integrate the governing equations for fluid motion and phase change heat transfer. The impact of wave amplitude and wave number of the heated tube, as well as the volume concertation of nanoparticles on the full-charging time of the LHTES unit, was addressed. The Taguchi optimization method was used to find an optimum design of the LHTES unit. The results showed that an increase in the volume fraction of nanoparticles reduces the charging time. Moreover, the waviness of the tube resists the natural convection flow circulation in the phase change domain and could increase the charging time.

A series of B3LYP/6-311+G(d,p) calculations of the affinity of monodentate ligands for [Cd(H 2 O) 3 ] 2+ are performed. Three types of ligands containing O (phosphine oxide, lactam, amide, carboxylic acid, ester, ketone, aldehyde, ether, halohydrin, enol, furan), N (thiocyanate, amine, ammonia, azide), and S (thioester, thioketone, thiol, thiophene, disulfide) interacting atoms are investigated. The results show that phosphine oxide has the largest affinity for the cadmium cation due to the polarization of the P=O bond. As the P atom has a large atomic radius, the O atom can polarize the electronic cloud enhancing its amount of electronic charge and favoring the interaction with Cd 2+ . The affinity order found is phosphine oxide > thioester > lactam > amide > carboxylic acid > ester > thioketone > ketone > thiocyanate > amine > ammonia > aldehyde > ether > thiol > thiophene > enol > halohydrin > disulfide > azide > furan ligands. These results were also corroborated by the functional M06-2X. The electronic effects (resonance and induction) of neighboring groups of the interacting atom modulate the strength of metal-ligand binding. For almost all the O-donor ligands the electrostatic component has the same magnitude as the covalent term, while for the N- and S-donor ligands the covalent term is predominant. The polarization term accounts for twice the exchange term as part of the covalent component. The dispersion term varies less than 2 kcal mol -1 for the complexes analyzed. The Pauli repulsion term is correlated with the metal ligand distance, increasing in the compounds with decreased metal-ligand bond length. The charge between the interacting atoms is also strongly correlated with both the interacting strength and the electrostatic interaction component. The natural bond orbital analysis highlights correlations of the bond order, and S and P contributions of the interacting metal-ligand orbital with the coordination strength. Graphical abstract The affinity of 20 monodentate ligands with different functional groups for the [Cd(H 2 O) 3 ] 2+ cation is calculated based on the interaction enthalpy and Gibbs free energy for the substitution of one water molecule from the fully hydrated cation. The affinity is correlated with geometric, electronic, and energetic parameters of the ligands and the complexes as well as with energy decomposition and natural bond order analyses results.

Calcium complexes with bidentate carbonyl ligands are important in biological systems, medicine and industry, where the concentration of Ca 2+ is controlled using chelating ligands. The exchange of two water molecules of [Ca(H 2 O) 6 ] 2+ for one bidentate monosubstituted and homo disubstituted dicarbonyl ligand was investigated using the B3LYP/6-311++G(d,p) method. The ligand substituents NH 2 , OCH 3 , OH, CH 3 , H, F, Cl, CN and NO 2 are functional groups with distinct electron-donating and -withdrawing effects that bond directly to the sp 2 C atom of the carbonyl group. The geometry, charge and energy characteristics of the complexes were analyzed to help understand the effects of substituents, spacer length and chelation. Coordination strength was quantified in terms of the enthalpy and free energy of the exchange reaction. The most negative enthalpies were calculated for the coordination of bidentate ligands containing three to five methylene group spacers between carbonyls. The chelate effect contribution was analyzed based on the thermochemistry. The electronic character of the substituent modulates the strength of binding to the metal cation, as ligands containing electron-donor substituents coordinate stronger than those with electron-acceptor substituents. This is reflected in the geometric (bond length and chelating angle), electronic (atomic charges) and energetic (components of the total interacting energy) characteristics of the complexes. Energy decomposition analysis (EDA)-an approach for partitioning of the energy into its chemical origins-shows that the electrostatic component of the coordination is predominant, and yields relevant contribution of the covalent term, especially for the electron-withdrawing substituted ligands. The chelate effect of the bidentate ligands was noticeable when compared with substitution by two monodentate ligands. Graphical abstract The affinity of 18 bidentate carbonyl ligands toward the [Ca(H 2 O) 4 ] 2+ cation is evaluated in terms of energetic, geometric and electronic parameters of the isolated ligands and the substituted aqua complexes. The electronic effects-inductive and mesomeric-intrinsic to the molecular structure of each ligand are found to modulate the strength of the metal-ligand interaction. The effects of polysubstitution, chelation and the length of the alkyl spacers between the anchor points of the ligand are also analyzed.

Base metal smelters may be a source of particulates containing metals of environmental concern released to the atmosphere. Knowledge of the quantitative chemical speciation of particulate releases from base metal smelters will be of value in smelter emission fingerprinting, site-specific risk assessments, predictions of the behaviour of smelter stack particulates released to the environment and in resolving liability issues related to current and historic releases. Accordingly, we have developed an innovative approach comprising bulk chemical analysis, a leaching procedure, X-ray diffraction analysis and scanning electron microscopy/electron probe microanalysis characterisation in a step-wise apportioning procedure to derive the quantitative speciation of particulate samples from the stacks of three copper smelters designated as A, B and C. For the A smelter stack particulates, the major calculated percentages were 29 CuSO(4), 20 ZnSO(4).H(2)O, 13 (Cu(0.94)Zn(0.06))(2)(AsO(4))(OH), 11 PbSO(4) and four As(2)O(3). For the B smelter stack particulates, the primary calculated percentages were 20 ZnSO(4).H(2)O, 20 PbSO(4), 12 CuSO(4) and nine As(2)O(3). Finally, we calculated that the C smelter stack particulates mostly comprised 34 ZnSO(4).H(2)O, 19 (Cu(0.84)Zn(0.16))(AsO(3)OH), 11 PbSO(4), 10 As(2)O(3) and nine Zn(3)(AsO(4))(2). Between 56% and 67% by weight of the smelter stack particulates, including the As, was soluble in water. For these and other operations, the data and approach may be useful in estimating metals partitioning among water, soil and sediment, as well as predictions of the effects of the stack particulates released to the environment.

A biotic ligand model (BLM) to predict chronic Cu toxicity to Ceriodaphnia dubia was developed and tested. The effect of cationic competition, pH and natural organic matter complexation of Cu was examined to develop the model. There was no effect of cationic competition using increasing Ca and Na concentrations in our exposures. However, we did see a significant regression of decreasing toxicity (measured as the IC25; concentration at which there was a 25% inhibition of reproduction) as Mg concentration increased. However, taking into account the actual variability of the IC25 and since the relative increase in IC25 due to additional Mg was small (1.5-fold) Mg competition was not included in the model. Changes in pH had a significant effect on Cu IC25, which is consistent with proton competition as often suggested for acute BLMs. Finally, natural organic matter (NOM) was added to exposures resulting in significant decreases in toxicity. Therefore, our predictive model for chronic Cu toxicity to C. dubia includes the effect of pH and NOM complexation. The model was validated with Cu IC25 data generated in six natural surface waters collected from across Canada. Using WHAM VI, we calculated Cu speciation in each natural water and using our model, we generated "predicted" IC25 data. We successfully predicted all Cu IC25 within a factor of 3 for the six waters used for validation.

Destruction of polychlorinated biphenyls (PCBs) in contaminated soils and wastes using circulating fluidized bed combustion (CFBC) technology was studied using a pilot plant and simulated waste material. The results show that the technology is effective and particularly promising for treatment of PCB-containing materials like the toxic sludge from a large contaminated site. Destruction of the toxics in the gas phase appears to be very fast, and over 99.9999% destruction and removal efficiency can be achieved in the temperature range 875-880 degrees C. Heat transfer in the fluidized bed also appears adequate. Toxic residues in treated soil can be reduced to very low levels. Rate-controlling factors of the decontamination process are analyzed, and key issues for determination of the process conditions are discussed.

Arthrobacter species strain FR-3, isolated from sediments of a swamp, produced a novel serine-type sulfide oxidase. The production of sulfide oxidase was maximal at pH 7.5 and 30 degrees C. Among various carbon and nitrogen sources tested, glucose and yeast extract were found to be the most effective substrates for the secretion of sulfide oxidase. The sulfide oxidase was purified to homogeneity and the molecular weight of the purified enzyme was 43 kDa when estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified sulfide oxidase can be effectively immobilized in DEAE (diethylaminoethyl)-cellulose matrix with a yield of 66%. The purified free and immobilized enzyme had optimum activity at pH 7.5 and 6.0, respectively. Immobilization increases the stability of the enzyme with respect to temperature. The half-life of the immobilized enzyme was 30 min at 45 degrees C, longer than that of the free enzyme (10 min). The purified free sulfide oxidase activity was completely inhibited by 1 mM Co2+ and Zn2+ and sulfhydryl group reagents (para-chloromercuribenzoic acid and iodoacetic acid). Catalytic activity was not affected by 1 mM Ca2+, Mg2+, Na+ and metal-chelating agent (EDTA).

Nanophase aluminum powder was characterized in a field-emission-gun transmission electron microscope (TEM). Different techniques were used to investigate the structure of the particles, including conventional bright-field and dark-field imaging, scanning transmission electron microscopy (STEM), high-resolution lattice imaging, diffraction studies, energy dispersive X-ray spectroscopy (EDS) analysis and mapping, and electron energy loss spectroscopy (EELS) analysis and mapping. It has been established that the particle cores consist of aluminum single crystals that sometimes contain crystal lattice defects. The core is covered by a passivating layer of aluminum oxide a few nanometers thick. The alumina is mostly amorphous, but evidences of partial crystallinity of the oxide were also found. The thickness of this layer was measured using different techniques, and the results are in good agreement with each other. The particles are agglomerated in two distinct ways. Some particles were apparently bonded together during processing before oxidation. These mostly form dumbbells covered by a joint oxide layer. Also, oxidized particles are loosely assembled into relatively large clusters.

Polychlorinated dibenzodioxins/furans (PCDD/F) were formed in substantial quantities in a pilot-scale fluidized bed combustor burning salt-laden waste wood, a common fuel for Canadian coastal pulp and paper mills. Formation of PCDD/F increased with increasing chloride content in the feed, and appeared to correlate with the chlorine content in the fly ash. It took a very long time for the ash chlorine content to stabilize, suggesting that chlorine transferred slowly from the flue gas to the ash. The baghouse may contribute largely to formation of the PCDD/F, owing to its temperature range and the potentially long residence time for ash particles. Controlling the baghouse temperature to reduce the PCDD/F formation in the baghouse should be effective in reducing the total emission level. While sulphur addition was found to reduce the emission level by as much as 90%, the emission level was still above the regulated level for the mills burning salt-laden wood under the conditions of the present study. No relation between the emission level and CO concentration in the flue gas was observed.

To minimize the disposal of highly reactive spent sorbent from a fluidized bed combustor, a new method for reactivation has been developed. The method consists of grinding the spent ash in a rotary mill, hydrating the ash with an excess of water, and mixing the wet ground ash with dry solids to absorb the excess water. The mixing process eliminates the formation of a concrete-like product that normally results as wet fluidized bed combustor ash ages. Pilot-scale combustion trials proved to be successful, and the process was scaled up using a 35MWt utility boiler at Purdue University. The test lasted for 3 days and resulted in net reduction of limestone sorbent use of 18%. The results generated in this work have been used to develop an economic evaluation for a 165MWe circulating fluidized bed (CFB) boiler, which projects significant savings due to reduction of limestone supply and ash disposal costs. The evaluation also suggests that the process is cost competitive with other processes, albeit that those processes have not been demonstrated at industrial scale. Furthermore, it also has the potential to make a small net reduction in CO(2) emissions, due to reduced limestone usage.

Emulsification requiring very little input energy can be induced at an oil-water interface that is initially in a state of equilibrium. The process involves destabilization, through contraction, of local interfacial regions. For emulsification to occur, it is necessary for the interfacial structure to have no resistance to surface shearing. Such a mechanism of emulsification may have important implications for the approach to solving emulsion problems in the petroleum industry. Copyright 1999 Academic Press.

CsBr- and CsI-kaolinite intercalation complexes were synthesized by gradually heating caesium halide disks of the DMSO-kaolinite intermediate up to 330 degreesC. Infrared spectroscopy revealed two types of complexes with the caesium salts: almost nonhydrous, obtained during thermal treatment of the DMSO complex, and hydrated, produced by regrinding the disk in air. Comparison of band positions for CsBr-kaolinite and CsI-kaolinite with those for the CsCl complex (observed in a previous study) shows that the strength of the hydrogen bond between the intercalated halide and the inner surface hydroxyl decreases on the order CsCl > CsBr > CsI. The nonreactivity of CsI in mechanochemical intercalation may arise from weak interaction between I- and inner surface hydroxyl groups, resulting from the fact that caesium is a very soft acid and iodide is a very soft base. Consequently, the very strong interaction between the two ions in the crystal is not disrupted during mechanochemical treatment. Copyright 1998 Academic Press and Minister of Natural Resources, Canada.

A Pseudomonas stutzeri isolate rapidly reduced both selenite and selenate ions to elemental selenium at initial concentrations of both anions of up to 48.1 mM. Optimal selenium reduction occurred under aerobic conditions between pH 7.0 and 9.0 and at temperatures of 25 to 35 degrees C. Reduction of both selenite and selenate was unaffected by a number of anions except for sulfite, chromate, and tungstate ions, which inhibited both growth and reduction.

During the 1983 field season, the rhizobacteria (including organisms from rhizosphere soil and the root rhizoplane) of cotton plants at one location in Mississippi were inventoried at different plant growth stages. Isolates (1,000) were identified to the genus level and characterized for repression of Pythium ultimum and Rhizoctonia solani. Cotton seedlings were initially colonized by bacteria of many different genera, and populations quickly reached 10 CFU/g of root tissue. As the season progressed, the bacterial populations declined as root mass increased and the roots became more woodlike in consistency. Fluorescent pseudomonads were the most numerous gram-negative rhizobacterial isolates of those that were randomly collected and identified, and they provided the largest number of isolates with fungal repressive activity. Several other gram-negative bacterial genera were recovered throughout the growing season, and some gram-positive bacteria were also isolated routinely, but at lower numbers. There was no correlation between the proportion of rhizobacterial isolates that possessed fungal repressive activity and the plant growth stage from which the isolates were obtained. Approximately twice as many bacterial isolates demonstrated fungal repression in the agar assay compared with the inplanta assay, and isolates were found more frequently with fungal repressive activity against P. ultimum than against R. solai.

A technical evaluation of an automated, programmable, grab-sampler manufactured by Pylon Electronic Development under the commercial name WL-1000C has been conducted. Six different methods are implemented in the instrument for analyzing radiation data. Any one method can be used on command and easily selected by means of a keyboard. Available radiation data that can be retrieved on command include radon (thoron) daughter concentrations and radon (thoron) Working Levels. Measurements were carried out under laboratory-controlled conditions in a large (26 m(3)) radon/thoron test facility, designed for calibration purposes, and at an underground uranium mine. Data obtained with the WL-1000C have been compared with conventional grab-sampling (e.g., Kusnetz, Thomas-Tsivoglou and Markov methods) and with other automated radiation instrumentation previously tested at our laboratories. Tests were done under constant radiation conditions and also under rapidly fluctuating conditions in order to determine the response of the instruments and methods in these two cases of practial interest. The Working Level used in these experiments was in the approximate range of 0.01 to 10 WL. Tests were conducted under a variety of environmental conditions. Good agreement with grab-sampling data was found for radon daughters. Some discrepancies with grab-sampling data were found for radon daughter/thoron daughter mixtures. Disagreement in the latter case is to be expected because of α-energy overlap between RaA and ThC.