Publications

An entry from the Cambridge Structural Database, the world’s repository for small molecule crystal structures. The entry contains experimental data from a crystal diffraction study. The deposited dataset for this entry is freely available from the CCDC and typically includes 3D coordinates, cell parameters, space group, experimental conditions and quality measures.

We assess four DFT functionals, RTPSS, RPBE, SCAN and B97M-rV, for surface interactions. We find that B97M-rV predicts the correct site preference for CO binding on Ag and Au while RTPSS performs well for surface relaxations and binding of CO on Pt and Cu.

The principal structure possibilities for packing infinite rod-shaped building blocks are described. Some basic nets derived from linking simple rods (helices and ladders) are then enumerated. We demonstrate the usefulness of the concept of rod secondary building units in the design and synthesis of metal-organic frameworks (MOFs). Accordingly, we present the preparation, characterization, and crystal structures of 14 new MOFs (named MOF-69A-C and MOF-70-80) of 12 different structure types, belonging to rod packing motifs, and show how their structures are related to basic nets. The MOFs reported herein are of polytopic carboxylates and contain one of Zn, Pb, Co, Cd, Mn, or Tb. The inclusion properties of the most open members are presented as evidence that MOF structures with rod building blocks can indeed be designed to have permanent porosity and rigid architectures.

Microbes hold the key to life. They hold the secrets to our past (as the descendants of the earliest forms of life) and the prospects for our future (as we mine their genes for solutions to some of the planet's most pressing problems, from global warming to antibiotic resistance). However, the piecemeal approach that has defined efforts to study microbial genetic diversity for over 20 years and in over 30,000 genome projects risks squandering that promise. These efforts have covered less than 20% of the diversity of the cultured archaeal and bacterial species, which represent just 15% of the overall known prokaryotic diversity. Here we call for the funding of a systematic effort to produce a comprehensive genomic catalog of all cultured Bacteria and Archaea by sequencing, where available, the type strain of each species with a validly published name (currently∼11,000). This effort will provide an unprecedented level of coverage of our planet's genetic diversity, allow for the large-scale discovery of novel genes and functions, and lead to an improved understanding of microbial evolution and function in the environment.

Metal halide perovskites have excellent optoelectronic properties. This study aims to determine how the optoelectronic properties of a model perovskite, cesium lead bromide (CsPbBr₃), change with length and thickness in one dimension (1D). By examining the photophysics of CsPbBr₃ quantum dots (QDs), nanowires (NWs), and nanorods (NRs), we observe the influence of confinement, exciton diffusion, and trapping on their optical properties. Our findings reveal that exciton diffusion to trap states limits the photoluminescence quantum yield (PLQY) of 1D CsPbBr₃ in the weakly confined regime (8-14 nm) and explains their long-lived exciton dynamics, while enhanced radiative rates contribute to achieving near-unity PLQY in the strongly confined regime (<7 nm). Consequently, blue-emitting, 2.4 nm-thick CsPbBr₃ NRs were 3.6X more emissive than the conventional CsPbBr₃ QDs. This study underscores how structural optimization can improve the optoelectronic performance of CsPbBr₃ and provides insight into the complex interplay of radiative and nonradiative processes in 1D ionic semiconductors.

Abstract Microglial activation is emerging as an important etiologic factor and therapeutic target in neurodegenerative and neuroinflammatory diseases. Techniques have been lacking, however, for measuring the different components of microglial activation independently in vivo. We describe a method for measuring microglial proliferation rates in vivo using heavy water ( 2 H 2 O) labeling, and its application in screening for drugs that suppress neuro‐inflammation. Brain microglia were isolated by flow cytometry as F4/80 + , CD11b + , CD45 low cells, and 2 H enrichment in DNA was analyzed by gas chromatography/mass spectrometry. Basal proliferation rate was ∼1%/week and systemic administration of bacterial lipopolysaccharide (LPS) markedly increased this rate in a dose‐dependent manner. Induction of experimental autoimmune encephalomyelitis (EAE) in C57BL/6 mice by MOG 35–55 peptide stimulated proliferation of CD45 low microglia, which could be distinguished from the proliferation of CD45 high infiltrating monocytes. Minocycline (45 mg/kg/day, i.p.) inhibited resident microglial proliferation in both the LPS and EAE models. Thirteen drugs were then screened for their ability to inhibit LPS‐stimulated microglia proliferation. Female C57BL/6 mice were given LPS (1 mg/kg), and concomitant drug treatment while receiving 2 H 2 O label for 7 days. Among the drugs screened, treatment with isotretinoin dose‐dependently reduced LPS‐induced microglial proliferation, representing an action of retinoids unknown previously. Follow‐up studies in the EAE model confirmed that isotretinoin not only inhibited proliferation of microglia but also delayed the onset of clinical symptoms. In conclusion, 2 H 2 O labeling represents a relatively high‐throughput, quantitative, and highly reproducible technique for measuring microglial proliferation, and is useful for screening and discovering novel anti‐neuroinflammatory drugs. © 2007 Wiley‐Liss, Inc.

Given a video of a person in action, we can easily guess the 3D future motion of the person. In this work, we present perhaps the first approach for predicting a future 3D mesh model sequence of a person from past video input. We do this for periodic motions such as walking and also actions like bowling and squatting seen in sports or workout videos. While there has been a surge of future prediction problems in computer vision, most approaches predict 3D future from 3D past or 2D future from 2D past inputs. In this work, we focus on the problem of predicting 3D future motion from past image sequences, which has a plethora of practical applications in autonomous systems that must operate safely around people from visual inputs. Inspired by the success of autoregressive models in language modeling tasks, we learn an intermediate latent space on which we predict the future. This effectively facilitates autoregressive predictions when the input differs from the output domain. Our approach can be trained on video sequences obtained in-the-wild without 3D ground truth labels. The project website with videos can be found at https://jasonyzhang.com/phd.