Influence of Heat Treatment on Microstructure, Mechanical Property, and Corrosion Behavior of Cold-Sprayed Zn Coating on Mg Alloy Substrate.

The influence of post-process heat treatment on cold-sprayed Zn coatings on the Mg alloy substrate was investigated at different temperatures (150, 250, and 350 °C) and times (2, 8, and 16 h). Phase, microstructure, microhardness, and tensile strength of Zn coatings were analyzed before and after heat treatment. Corrosion properties of Zn coatings after heat treatment were investigated in simulated body fluid by using potentiodynamic polarization and immersion testing. Results show that although the heat treatment presented little effect on phase compositions of Zn coatings, the full width at half maxima of the Zn phase decreased with the heat temperature and time. Zn coatings presented comparable microstructures before and after heat treatment in addition to the inter-diffusion layers, and the inter-diffusion layer was dependent on the heat temperature and time. Both the thickness and the microhardness of inter-diffusion layers were increased with the heat temperature and time, with the largest thickness of 704.1 ± 32.4 µm and the largest microhardness of 323.7 ± 104.1 HV0.025 at 350 °C for 2 h. The microhardness of Zn coating was significantly decreased from 70.8 ± 5.6 HV0.025 to 43.9 ± 12.5 HV0.025, with the heat temperature from the ambient temperature to 350 °C, and was slightly decreased with the heat time at 250 °C. Although the tensile strength of Zn coating was slightly increased by heat treatment, with the highest value of 40.9 ± 3.9 MPa at 150 °C for 2 h, excessive heat temperature and time were detrimental to the tensile strength, with the lowest value of 6.6 ± 1.6 MPa at 350 °C for 2 h. The heat temperature and heat time presented limited effects on the corrosion current and corrosion ratio of the Zn coatings, and Zn coatings before and after heat treatment effectively hindered the simulated body fluid from penetrating into the substrate. The corrosion behavior of Zn coatings was discussed in terms of corrosion products and microstructures after immersion.

Thermal and Catalytic Decomposition of 2-Hydroxyethylhydrazine and 2-Hydroxyethylhydrazinium Nitrate Ionic Liquid.

To develop chemical kinetics models for the combustion of ionic liquid-based monopropellants, identification of the elementary steps in the thermal and catalytic decomposition of components such as 2-hydroxyethylhydrazinium nitrate (HEHN) is needed but is currently not well understood. The first decomposition step in protic ionic liquids such as HEHN is typically the proton transfer from the cation to the anion, resulting in the formation of 2-hydroxyethylhydrazine (HEH) and HNO3. In the first part of this investigation, the high-temperature thermal decomposition of HEH is probed with flash pyrolysis (<1400 K) and vacuum ultraviolet (10.45 eV) photoionization time-of-flight mass spectrometry (VUV-PI-TOFMS). Next, the investigation into the thermal and catalytic decomposition of HEHN includes two mass spectrometric techniques: (1) tunable VUV-PI-TOFMS (7.4-15 eV) and (2) ambient ionization mass spectrometry utilizing both plasma and laser ionization techniques whereby HEHN is introduced onto a heated inert or iridium catalytic surface and the products are probed. The products can be identified by their masses, their ionization energies, and their collision-induced fragmentation patterns. Formation of product species indicates that catalytic surface recombination is an important reaction process in the decomposition mechanism of HEHN. The products and their possible elementary reaction mechanisms are discussed.

Cell-Type-Specific Metabolic Profiling Achieved by Combining Desorption Electrospray Ionization Mass Spectrometry Imaging and Immunofluorescence Staining.

Cell-type-specific metabolic profiling in tissue with heterogeneous composition has been of great interest across all mass spectrometry imaging (MSI) technologies. We report here a powerful new chemical imaging capability in desorption electrospray ionization (DESI) MSI, which enables cell-type-specific and in situ metabolic profiling in complex tissue samples. We accomplish this by combining DESI-MSI with immunofluorescence staining using specific cell-type markers. We take advantage of the variable frequency of each distinct cell type in the lateral septal nucleus (LSN) region of mouse forebrain. This allows computational deconvolution of the cell-type-specific metabolic profile in neurons and astrocytes by convex optimization-a machine learning method. Based on our approach, we observed 107 metabolites that show different distributions and intensities between astrocytes and neurons. We subsequently identified 23 metabolites using high-resolution mass spectrometry (MS) and tandem MS, which include small metabolites such as adenosine and N-acetylaspartate previously associated with astrocytes and neurons, respectively, as well as accumulation of several phospholipid species in neurons which have not been studied before. Overall, this method overcomes the relatively low spatial resolution of DESI-MSI and provides a new platform for in situ metabolic investigation at the cell-type level in complex tissue samples with heterogeneous cell-type composition.

Author Correction: Optimization of Molecules via Deep Reinforcement Learning.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Proof of concept for identifying cystic fibrosis from perspiration samples.

The gold standard for cystic fibrosis (CF) diagnosis is the determination of chloride concentration in sweat. Current testing methodology takes up to 3 h to complete and has recognized shortcomings on its diagnostic accuracy. We present an alternative method for the identification of CF by combining desorption electrospray ionization mass spectrometry and a machine-learning algorithm based on gradient boosted decision trees to analyze perspiration samples. This process takes as little as 2 min, and we determined its accuracy to be 98 ± 2% by cross-validation on analyzing 277 perspiration samples. With the introduction of statistical bootstrap, our method can provide a confidence estimate of our prediction, which helps diagnosis decision-making. We also identified important peaks by the feature selection algorithm and assigned the chemical structure of the metabolites by high-resolution and/or tandem mass spectrometry. We inspected the correlation between mild and severe CFTR gene mutation types and lipid profiles, suggesting a possible way to realize personalized medicine with this noninvasive, fast, and accurate method.

Optimization of Molecules via Deep Reinforcement Learning.

We present a framework, which we call Molecule Deep Q-Networks (MolDQN), for molecule optimization by combining domain knowledge of chemistry and state-of-the-art reinforcement learning techniques (double Q-learning and randomized value functions). We directly define modifications on molecules, thereby ensuring 100% chemical validity. Further, we operate without pre-training on any dataset to avoid possible bias from the choice of that set. MolDQN achieves comparable or better performance against several other recently published algorithms for benchmark molecular optimization tasks. However, we also argue that many of these tasks are not representative of real optimization problems in drug discovery. Inspired by problems faced during medicinal chemistry lead optimization, we extend our model with multi-objective reinforcement learning, which maximizes drug-likeness while maintaining similarity to the original molecule. We further show the path through chemical space to achieve optimization for a molecule to understand how the model works.

Combining Desorption Electrospray Ionization Mass Spectrometry Imaging and Machine Learning for Molecular Recognition of Myocardial Infarction.

Lipid profile changes in heart muscle have been previously linked to cardiac ischemia and myocardial infarction, but the spatial distribution of lipids and metabolites in ischemic heart remains to be fully investigated. We performed desorption electrospray ionization mass spectrometry imaging of hearts from in vivo myocardial infarction mouse models. In these mice, myocardial ischemia was induced by blood supply restriction via a permanent ligation of left anterior descending coronary artery. We showed that applying the machine learning algorithm of gradient boosting tree ensemble to the ambient mass spectrometry imaging data allows us to distinguish segments of infarcted myocardium from normally perfused hearts on a pixel by pixel basis. The machine learning algorithm selected 62 molecular ion peaks important for classification of each 200 µm-diameter pixel of the cardiac tissue map as normally perfused or ischemic. This approach achieved very high average accuracy (97.4%), recall (95.8%), and precision (96.8%) at a spatial resolution of ∼200 µm. In addition, we determined the chemical identity of 27 species, mostly small metabolites and lipids, selected by the algorithm as the most significant for cardiac pathology classification. This molecular signature of myocardial infarction may provide new mechanistic insights into cardiac ischemia, assist with infarct size assessment, and point toward novel therapeutic interventions.

Fluorescence Polarization Anisotropy in Microdroplets.

Chemical reactions can be greatly accelerated in microdroplets, but the factors that lead to acceleration are still being elucidated. Using rhodamine 6G (R6G) as a model compound, we studied the density distribution and fluorescence polarization anisotropy of this dye in water-in-oil microdroplets. We found the density of R6G is higher on the surface of the microdroplets, and the ratio of the surface density to that of the center grows with increasing microdroplet radius or with decreasing R6G concentration. The measured fluorescence polarization anisotropy at the surface is almost the same for droplets of different sizes but becomes larger when the concentration is lowered. We also performed three-dimensional simulations by treating R6G+ and its associated anion as a dipole of fixed length and magnitude. The simulation results match quite well the experimental measurements, showing that the density distribution and fluorescence polarization anisotropy can be largely explained by a simple electrostatic model.

Personal Information from Latent Fingerprints Using Desorption Electrospray Ionization Mass Spectrometry and Machine Learning.

Desorption electrospray ionization-mass spectrometry imaging (DESI-MSI) was applied to latent fingerprints to obtain not only spatial patterns but also chemical maps. Samples with similar lipid compositions as those of the fingerprints were collected by swiping a glass slide across the forehead of consenting adults. A machine learning model called gradient boosting tree ensemble (GDBT) was applied to the samples that allowed us to distinguish between different genders, ethnicities, and ages (within 10 years). The results from 194 samples showed accuracies of 89.2%, 82.4%, and 84.3%, respectively. Specific chemical species that were determined by the feature selection of GDBT were identified by tandem mass spectrometry. As a proof-of-concept, the machine learning model trained on the sample data was applied to overlaid latent fingerprints from different individuals, giving accurate gender and ethnicity information from those fingerprints. The results suggest that DESI-MSI imaging of fingerprints with GDBT analysis might offer a significant advance in forensic science.

Optimizing Chemical Reactions with Deep Reinforcement Learning.

Deep reinforcement learning was employed to optimize chemical reactions. Our model iteratively records the results of a chemical reaction and chooses new experimental conditions to improve the reaction outcome. This model outperformed a state-of-the-art blackbox optimization algorithm by using 71% fewer steps on both simulations and real reactions. Furthermore, we introduced an efficient exploration strategy by drawing the reaction conditions from certain probability distributions, which resulted in an improvement on regret from 0.062 to 0.039 compared with a deterministic policy. Combining the efficient exploration policy with accelerated microdroplet reactions, optimal reaction conditions were determined in 30 min for the four reactions considered, and a better understanding of the factors that control microdroplet reactions was reached. Moreover, our model showed a better performance after training on reactions with similar or even dissimilar underlying mechanisms, which demonstrates its learning ability.

Nanotip Ambient Ionization Mass Spectrometry.

A method called nanotip ambient ionization mass spectrometry (NAIMS) is described, which applies high voltage between a tungsten nanotip and a metal plate to generate a plasma in which ionized analytes on the surface of the metal plate are directed to the inlet and analyzed by a mass spectrometer. The dependence of signal intensity is investigated as a function of the tip-to-plate distance, the tip size, the voltage applied at the tip, and the current. These parameters are separately optimized to achieve sensitivity or high spatial resolution. A partially observable Markov decision process is used to achieve a stabilized plasma as well as high ionization efficiency. As a proof of concept, the NAIMS technique has been applied to phenanthrene and caffeine samples for which the limits of detection were determined to be 0.14 fmol for phenanthrene and 4 amol for caffeine and to a printed caffeine pattern for which a spatial resolution of 8 ± 2 µm, and the best resolution of 5 µm, was demonstrated. The limitations of NAIMS are also discussed.

A dual amplification strategy for DNA detection combining bio-barcode assay and metal-enhanced fluorescence modality.

Silver-enhanced fluorescence was coupled with a bio-barcode assay to facilitate a dual amplification assay to demonstrate a non-enzymatic approach for simple and sensitive detection of DNA. In the assay design, magnetic nanoparticles seeded with silver nanoparticles were modified with the capture DNA, and silver nanoparticles were modified with the binding of ssDNA and the fluorescently labeled barcode dsDNA. Upon introduction of the target DNA, a sandwich structure was formed because of the hybridization reaction. By simple magnetic separation, silver-enhanced fluorescence of barcode DNAs could be readily measured without the need of a further step to liberate barcode DNAs from silver nanoparticles, endowing the method with simplicity and high sensitivity with a detection limit of 1 pM.

A distance-dependent metal-enhanced fluorescence sensing platform based on molecular beacon design.

A new metal-enhanced fluorescence (MEF) based platform was developed on the basis of distance-dependent fluorescence quenching-enhancement effect, which combined the easiness of Ag-thiol chemistry with the MEF property of noble-metal structures as well as the molecular beacon design. For the given sized AgNPs, the fluorescence enhancement factor was found to increase with a d(6) dependency in agreement with fluorescence resonance energy transfer mechanism at shorter distance and decrease with a d(-3) dependency in agreement with plasmonic enhancement mechanism at longer distance between the fluorophore and the AgNP surface. As a proof of concept, the platform was demonstrated by a sensitive detection of mercuric ions, using thymine-containing molecular beacon to tune silver nanoparticle (AgNP)-enhanced fluorescence. Mercuric ions were detected via formation of a thymine-mercuric-thymine structure to open the hairpin, facilitating fluorescence recovery and AgNP enhancement to yield a limit of detection of 1 nM, which is well below the U.S. Environmental Protection Agency regulation of the Maximum Contaminant Level Goal (10nM) in drinking water. Since the AgNP functioned as not only a quencher to reduce the reagent blank signal but also an enhancement substrate to increase fluorescence of the open hairpin when target mercuric ions were present, the quenching-enhancement strategy can greatly improve the detection sensitivity and can in principle be a universal approach for various targets when combined with molecular beacon design.