We introduce the 2D counterpart of layered black phosphorus, which we call phosphorene, as an unexplored p-type semiconducting material. Same as graphene and MoS2, single-layer phosphorene is flexible and can be mechanically exfoliated. We find phosphorene to be stable and, unlike graphene, to have an inherent, direct, and appreciable band gap. Our ab initio calculations indicate that the band gap is direct, depends on the number of layers and the in-layer strain, and is significantly larger than the bulk value of 0.31-0.36 eV. The observed photoluminescence peak of single-layer phosphorene in the visible optical range confirms that the band gap is larger than that of the bulk system. Our transport studies indicate a hole mobility that reflects the structural anisotropy of phosphorene and complements n-type MoS2. At room temperature, our few-layer phosphorene field-effect transistors with 1.0 μm channel length display a high on-current of 194 mA/mm, a high hole field-effect mobility of 286 cm(2)/V·s, and an on/off ratio of up to 10(4). We demonstrate the possibility of phosphorene integration by constructing a 2D CMOS inverter consisting of phosphorene PMOS and MoS2 NMOS transistors.

Cerium and manganese ions are very effective reversible scavengers of •OH in an operating PFSA-based PEM fuel cell. The use of these ions in very small quantities can reduce the fluoride release rate by up to three orders of magnitude relative to an unmitigated sample and thereby afford extremely durable membranes. A chemically rational mechanism that accounts for the remarkable effectiveness of these chemical mitigants is presented.

An electrochemical carbon nanotube (CNT) filter has been reported to be effective for the adsorptive removal and oxidation of aqueous organic compounds. Here, we complete a detailed investigation of the aqueous dye oxidation reactive transport mechanism during electrochemical filtration. Similar to batch electrolysis, mass transfer, physical adsorption, and electron transfer are found to be three primary steps in the overall electrochemical filtration organic oxidation mechanism. Mass transfer was quantitatively examined by chronoamperometry and normal pulse voltammetry and determined to be increased 6-fold during electrochemical filtration as compared to batch electrochemistry. Convection-enhanced mass transfer to the electrode surface is determined to be the primary factor for increased current density and organic oxidation during electrochemical filtration. Physical adsorption of the organics onto the CNTs was evaluated using temperature-dependent batch adsorption and electrochemical filtration experiments. The electrochemical filtration kinetics were observed to have a minor negative temperature-dependence. Electron transfer was examined by challenging the electrochemical filter with a range of increasing dye concentrations until the mass transfer and adsorption processes were saturated. Upon surface site saturation, the electron transfer rates were determined to be 8.5 × 1015, 6.3 × 1016, and 1.3 × 1017 e– s–1 m–2 at anode potentials of 0.35, 0.77, and 1.50 V, respectively. The electron transfer mechanism was also investigated and direct electron transfer was determined to be the dominant methyl orange oxidation mechanism at all evaluated anode potentials with an increasing contribution from indirect oxidation processes at potentials ≥1.0 V. The anode potential dependent maximum electron transfer rate is also observed to be affected by the polarity of the organic charge indicating electromigration is also active. In summary, electrochemical filtration is advantageous as compared to batch electrolysis due to the liquid flow through the electrode resulting in convection-enhanced transfer of the target molecule to the electrode surface.

Machine learning (ML) regression methods are promising tools to develop models predicting the properties of materials by learning from existing databases. However, although ML models are usually good at interpolating data, they often do not offer reliable extrapolations and can violate the laws of physics. Here, to address the limitations of traditional ML, we introduce a "topologyinformed ML" paradigm-wherein some features of the network topology (rather than traditional descriptors) are used as fingerprint for ML models-and apply this method to predict the forward (stage I) dissolution rate of a series of silicate glasses. We demonstrate that relying on a topological description of the atomic network (i) increases the accuracy of the predictions, (ii) enhances the simplicity and interpretability of the predictive models, (iii) reduces the need for large training sets, and (iv) improves the ability of the models to extrapolate predictions far from their training sets. As such, topology-informed ML can overcome the limitations facing traditional ML (e.g., accuracy vs. simplicity tradeoff) and offers a promising route to predict the properties of materials in a robust fashion.

Abstract