Search for: All records

Abstract Nature's ability to create complex and functionalized organisms has long inspired engineers and scientists to develop increasingly advanced machines. Magnetotactic bacteria (MTB), a group of Gram‐negative prokaryotes that biomineralize iron and thrive in aquatic environments, have garnered significant attention from the bioengineering community. These bacteria possess chains of magnetic nanocrystals known as magnetosomes, which allow them to align with Earth's geomagnetic field and navigate through aquatic environments via magnetotaxis, enabling localization to areas rich in nutrients and optimal oxygen concentration. Their built‐in magnetic components, along with their intrinsic and/or modified biological functions, make them one of the most promising platforms for future medical microrobots. Leveraging an externally applied magnetic field, the motion of MTBs can be precisely controlled, rendering them suitable for use as a new type of biohybrid microrobotics with great promise in medicine for bioimaging, drug delivery, cancer therapy, antimicrobial treatment, and detoxification. This mini‐review provides an up‐to‐date overview of recent advancements in MTB microrobots, delineates the interaction between MTB microrobots and magnetic fields, elucidates propulsion mechanisms and motion control, and reports state‐of‐the‐art strategies for modifying and functionalizing MTB for medical applications. 

Recent advancements in Multimodal Large Language Models (LLMs) have focused primarily on scaling by increasing text-image pair data and enhancing LLMs to improve performance on multimodal tasks. However, these scaling approaches are computationally expensive and overlook the significance of efficiently improving model capabilities from the vision side. Inspired by the successful applications of Mixture-of-Experts (MoE) in LLMs, which improves model scalability during training while keeping inference costs similar to those of smaller models, we propose CuMo, which incorporates Co-upcycled Top-K sparsely-gated Mixtureof-experts blocks into both the vision encoder and the MLP connector, thereby enhancing the multimodal LLMs with neglectable additional activated parameters during inference. CuMo first pre-trains the MLP blocks and then initializes each expert in the MoE block from the pre-trained MLP block during the visual instruction tuning stage, with auxiliary losses to ensure a balanced loading of experts. CuMo outperforms state-of-the-art multimodal LLMs across various VQA and visual-instruction-following benchmarks within each model size group, all while training exclusively on open-sourced datasets. 

The Earth’s atmosphere is composed of an enormous variety of chemical species associated with trace gases and aerosol particles whose composition and chemistry have critical impacts on the Earth’s climate, air quality, and human health. Mass spectrometry analysis as a powerful and popular analytical technique has been widely developed and applied in atmospheric chemistry for decades. Mass spectrometry allows for effective detection, identification, and quantification of a broad range of organic and inorganic chemical species with high sensitivity and resolution. In this review, we summarize recently developed mass spectrometry techniques, methods, and applications in atmospheric chemistry research in the past several years. Specifically, new developments of ion-molecule reactors, various soft ionization methods, and unique coupling with separation techniques are highlighted. The new mass spectrometry applications in laboratory studies and field measurements focus on improving the detection limits for traditional and emerging volatile organic compounds, characterizing multiphase highly oxygenated molecules, and monitoring particle bulk and surface compositions. 

With an increasing prevalence of electronic cigarette (e-cigarette) use, especially among youth, there is an urgent need to better understand the biological risks and pathophysiology of health conditions related to e-cigarettes. A majority of e-cigarette aerosols are in the submicron size and would deposit in the alveolar region of the lung, where they must first interact with the endogenous pulmonary surfactant. To date, little is known whether e-cigarette aerosols have an adverse impact on the pulmonary surfactant. We have systematically studied the effect of individual e-cigarette ingredients on an animal-derived clinical surfactant preparation, bovine lipid extract surfactant, using a combination of biophysical and analytical techniques, including in vitro biophysical simulations using constrained drop surfactometry, molecular imaging with atomic force microscopy, chemical assays using carbon nuclear magnetic resonance and circular dichroism, and in silico molecular dynamics simulations. All data collectively suggest that flavorings used in e-cigarettes, especially menthol, play a predominant role in inhibiting the biophysical function of the surfactant. The mechanism of biophysical inhibition appears to involve menthol interactions with both phospholipids and hydrophobic proteins of the natural surfactant. These results provide novel insights into the understanding of the health impact of e-cigarettes and may contribute to better regulation of e-cigarette products. 

Abstract. Fires emit a substantial amount of non-methane organic gases (NMOGs), theatmospheric oxidation of which can contribute to ozone and secondaryparticulate matter formation. However, the abundance and reactivity of thesefire NMOGs are uncertain and historically not well constrained. In thiswork, we expand the representation of fire NMOGs in a global chemicaltransport model, GEOS-Chem. We update emission factors to Andreae (2019) andthe chemical mechanism to include recent aromatic and ethene and ethyne modelimprovements(Bateset al., 2021; Kwon et al., 2021). We expand the representation of NMOGs byadding lumped furans to the model (including their fire emission andoxidation chemistry) and by adding fire emissions of nine species alreadyincluded in the model, prioritized for their reactivity using data from the Fire Influence on Regional to Global Environments (FIREX) laboratory studies. Based on quantified emissions factors, we estimatethat our improved representation captures 72 % of emitted, identified NMOGcarbon mass and 49 % of OH reactivity from savanna and temperate forestfires, a substantial increase from the standard model (49 % of mass,28 % of OH reactivity). We evaluate fire NMOGs in our model withobservations from the Amazon Tall Tower Observatory (ATTO) in Brazil, Fire Influence on Regional to Global Environments and Air Quality (FIREX-AQ) and DC3 in the US, and Arctic Research of the Composition of theTroposphere from Aircraft and Satellites (ARCTAS) in boreal Canada. We show that NMOGs,including furan, are well simulated in the eastern US with someunderestimates in the western US and that adding fire emissions improves ourability to simulate ethene in boreal Canada. We estimate that fires provide15 % of annual mean simulated surface OH reactivity globally, as well as morethan 75 % over fire source regions. Over continental regions about half ofthis simulated fire reactivity comes from NMOG species. We find that furansand ethene are important globally for reactivity, while phenol is moreimportant at a local level in the boreal regions. This is the first globalestimate of the impact of fire on atmospheric reactivity. 

Abstract Agricultural and prescribed burning activities emit large amounts of trace gases and aerosols on regional to global scales. We present a compilation of emission factors (EFs) and emission ratios from the eastern portion of the Fire Influence on Regional to Global Environments and Air Quality (FIREX‐AQ) campaign in 2019 in the United States, which sampled burning of crop residues and other prescribed fire fuels. FIREX‐AQ provided comprehensive chemical characterization of 53 crop residue and 22 prescribed fires. Crop residues burned at different modified combustion efficiencies (MCE), with corn residue burning at higher MCE than other fuel types. Prescribed fires burned at lower MCE (<0.90) which is typical, while grasslands burned at lower MCE (0.90) than normally observed due to moist, green, growing season fuels. Most non‐methane volatile organic compounds (NMVOCs) were significantly anticorrelated with MCE except for ethanol and NMVOCs that were measured with less certainty. We identified 23 species where crop residue fires differed by more than 50% from prescribed fires at the same MCE. Crop residue EFs were greater for species related to agricultural chemical use and fuel composition as well as oxygenated NMVOCs possibly due to the presence of metals such as potassium. Prescribed EFs were greater for monoterpenes (5×). FIREX‐AQ crop residue average EFs generally agreed with the previous agricultural fire study in the US but had large disagreements with global compilations. FIREX‐AQ observations show the importance of regionally‐specific and fuel‐specific EFs as first steps to reduce uncertainty in modeling the air quality impacts of fire emissions.