Search for: All records

Carbon nitride materials paired with Co atoms have been shown to be effective for CO₂photoreduction. 

Photoactive single-atom catalysts (SACs) are among the most exciting catalytic materials for solar fuel production. Different SACs, including our own Co SACs, have been prepared on graphitic carbon nitride (C3N4) for use in photocatalysis. Building on our prior success, we report here doped C3N4 using various supplemental carbon dopants as the support for Co SACs. The Co SAC on a dianhydride doped C3N4 showed the highest activity in photocatalytic CO2 reduction. Catalyst characterization was carried out to explore the origin of the enhanced activity of this particular Co SAC. The dianhydride doped C3N4 possesses unique microstructural features, including large inter-layer space and fibrous morphology, that could contribute to the enhanced photocatalytic activity. Our results further indicate that the dianhydride is the most effective dopant to incorporate aromatic moieties in C3N4, which resulted in improved charge separation and enhanced activity in photocatalysis. 

Two-dimensional, 2D, niobium carbide MXene, Nb2CTx, has attracted attention due to its extraordinarily high photothermal conversion efficiency that has applications ranging from medicine, for tumor ablation, to solar energy conversion. Here, we characterize its electronic properties and investigate the ultrafast dynamics of its photoexcitations with a goal of shedding light onto the origins of its unique properties. Through density functional theory, DFT, calculations, we find that Nb2CTx is metallic, with a small but finite DOS at the Fermi level for all experimentally relevant terminations that can be achieved using HF or molten salt etching of the parent MAX phase, including –OH, –O, –F, –Cl, –Br, –I. In agreement with this prediction, THz spectroscopy reveals an intrinsic long-range conductivity of ∼60 Ω−1 cm−1, with significant charge carrier localization and a charge carrier density (∼1020 cm−3) comparable to Mo-based MXenes. Excitation with 800 nm pulses results in a rapid enhancement in photoconductivity, which decays to less than 25% of its peak value within several picoseconds, underlying efficient photothermal conversion. At the same time, a small fraction of photoinjected excess carriers persists for hundreds of picoseconds and can potentially be utilized in photocatalysis or other energy conversion applications. 

Garnering attention for high conductivity, nonlinear optical properties, and more, MXenes are water-processable 2D materials that are considered candidates for applications in electromagnetic interference shielding, optoelectronic and photonic devices among others. Herein we investigate the intrinsic and photoexcited conductivity in Nb 2 CT x, a MXene with reported high photothermal conversion efficiency. DFT calculations show that hydroxyl and/or fluorine-terminated or is metallic, in agreement with THz spectroscopy, which reveals the presence of free charge carriers that are highly localized over mesoscopic length scales. Photoexcitation of Nb 2 CT x, known to result in rapid heating of the crystal lattice, is found to produce additional free carriers and a transient enhancement of photoconductivity. Most photoexcited carriers decay over the sub-picosecond time scales while a small fraction remain for much longer, sub-nanoseconds, times. 

Regression ensembles consisting of a collection of base regression models are often used to improve the estimation/prediction performance of a single regression model. It has been shown that the individual accuracy of the base models and the ensemble diversity are the two key factors affecting the performance of an ensemble. In this paper, we derive a theory for regression ensembles that illustrates the subtle trade-off between individual accuracy and ensemble diversity from the perspective of statistical correlations. Then, inspired by our derived theory, we further propose a novel loss function and a training algorithm for deep learning regression ensembles. We then demonstrate the advantage of our training approach over standard regression ensemble methods including random forest and gradient boosting regressors with both benchmark regression problems and chemical sensor problems involving analysis of Raman spectroscopy. Our key contribution is that our loss function and training algorithm is able to manage diversity explicitly in an ensemble, rather than merely allowing diversity to occur by happenstance.