id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
666bf9b7c9c6a5c07a7379e5 | 9 | Cells were thawed and resuspended in 1 % Buffer B (1 g cell pellet/ 5 mL) and lysed by sonication (1 s ON, 1 s OFF, 50 % amplitude, 10 min). Cell debris was removed by high speed centrifugation at 18000 rpm for 20 min. The lysate superantant was loaded onto 5 mL of Ni-NTA Agarose (Qiagen) and the column washed with 25 mL 1 % Buffer B, 6 % Buffer B to remove untagged proteins and finally 10 mL 100 % Buffer B to elute purified protein. Fractions corresponding to the peak at 280 nm were checked for purity by SDS-PAGE and pure fractions were concentrated by spin column (Vivaspin 20, 20 kDA cutoff GE Healthcare) and buffer exchanged into 100 mM phosphate buffer, pH 7, 0. General procedure for analytical scale biotransformations in the amination direction Analytical scale biotransformations were performed using the specified ketone (2.5 mM) and the relevant enzyme (20 or 100 µM) in 100 mM KPi buffer pH 7 with 20% DMSO containing 1 mM pyridoxal phosphate and 500 mM isopropyl amine. The reaction mixture was incubated for 24 h with shaking at 800 r.p.m. The reaction was quenched with 1 volume acetonitrile with precipitated protein removed by centrifugation (14,000g for 10 minutes) with the supernatant used for UPLC analysis. For chiral analysis, reactions were quenched by the addition of 100 μL of 5 M NaOH and extracted with EtOAc (250 μL x 3). The organic phase was dried over MgSO4 and analysed by normal-phase HPLC. |
666bf9b7c9c6a5c07a7379e5 | 10 | Analytical scale biotransformations were performed using the specified amine (5 mM) and the relevant enzyme (20 µM) in 100 mM KPi buffer pH 7 with 20% DMSO containing 1 mM pyridoxal phosphate and 200 mM pyruvate. The reaction mixture was incubated for up to 24 h with shaking at 800 r.p.m. The reaction was quenched with 1 volume acetonitrile with precipitated protein removed by centrifugation (14,000g for 10 minutes) with the supernatant used for UPLC analysis. For chiral analysis, reactions were quenched by the addition of 100 μL of 5 M NaOH and extracted with EtOAc (250 μL x 3). The organic phase was dried over MgSO4 and analysed by normal-phase HPLC. |
666bf9b7c9c6a5c07a7379e5 | 11 | Libraries of transaminase mutants were generated by site saturation mutagenesis, targeting active site, 2 nd sphere and dimer interface amino acid residues or combinatorial libraries that randomly combined beneficial mutations found in a round. Saturation mutagenesis libraries were constructed by overlap extension PCR. The degenerate NNK codon was used to generate fragments that coded for all 20 amino acids. The linear fragments were then amplified using the appropriate 3' and 5' oligonucleotide primers The linear fragments and the pET28b vector were digested using Nde I and Xho I endonucleases, PCR purified and ligated using T4 DNA ligase. After each round of evolution, combinatorial libraries were generated by overlap extension PCR. Primers were designed that encoded either the parent amino acid or the identified mutation, These primers were used to generate short fragments which were gel-purified and mixed appropriately in overlap extension PCR to generate genes containing all possible combination of mutations. Genes were cloned as described above. |
666bf9b7c9c6a5c07a7379e5 | 12 | For protein expression and screening, all transfer steps were performed using Hamilton liquid-handling robots. The ligated libraries were transformed into chemically compotent E. coli BL21 (DE3) cells. Freshly transformed clones were used to inoculate 150 μL of LB media supplement with 50 μg mL -1 kanamycin in Corning®Costar®96-well microtitre flat bottom plates, covered with a BreathEasy film and incubated overnight at 30 o C, 80% humidity at 850rpm. 20 μL of overnight culture was used to inoculate 480 μL TB medium supplemented with 50 μg mL -1 kanamycin in 96-deep well plates and further incubated under the same conditions until an OD600 of 0.6 was reached and IPTG was added to a final concentration of 0.1 mM to induce protein expression. Plates were further incubated under the same conditions overnight and cells were harvested by centrifugation at 4000 rpm for 20 min. The supernatant was discarded, and the pelleted cells were stored at -80 o C. For reference, each plate contained 6 clones of the parent template, 2 clones containing RFP in pET28B as positive and negative controls. The pelleted cells were resuspended in 400 μL lysis buffer (100 mM KPi, 1 mg mL -1 lysozyme, 0.5 mg mL -1 polymixin B and 10 μg mL -1 DNase I) incubated for 2h at 30 o C, 80 % humidity at 850 rpm. Cell debris was removed by centrifugation at 4000 rpm for 30 min. 55 μL clarified lysate was transferred to 96-well microtitre plates containing, 45 μL of reaction mix. Reactions were incubated overnight at 30 o C, 80 % humidity at 850 rpm and quenched by the addition of 100 μL AcN (0.1% HCl). Samples were further incubated for 30 min and precipitated proteins were removed by centrifugation at 4000 rpm for 40 min and quenched supernatants analysed by UPLC analysis. Conversion of individual mutants were normalized to the average of the 6 parent clones of that plate. Mutants that demonstrated increased activity were rescreened as purified proteins, that were produced and purified as described above, but starter cultures were inoculated from glycerol stocks prepared from the original overnight cultures. Chemical Synthesis Synthesis of 3,6-dibromopicolinaldehyde (11) A stirred solution of 2,5-dibromo-6-methylpyridine (8 g, 31.8 mmol) in THF (80 mL) was cooled to 0 o C. To this solution tert-butyl nitrite (4.33g, 41.9 mmol) was added, followed by the dropwise addition of potassium tert-butoxide (28 mL, 20 wt% sol. In THF, 47.7 mmol). The reaction was stirred for 4 h. The reaction mixture was diluted with THF (25 mL) and quenched with NH4Cl in water (6.38g, 119 mmol). The reaction mixture was distilled under vacuum to approximately 50 mL and the resulting slurry was filtered and washed with water (2 x 50 mL). The resulting compound was added portion wise to a stirred solution of glyoxylic acid (100 mL, 50 wt%) at 80 o C and stirred for 4 h. Synthesis of N-benzhydryl-1-(3,6-dibromopyridin-2-yl)methanimine (12) 3,6-dibromopicolinaldehyde (5.0 g, 18.9 mmol) in toluene (20 mL) was heated to 50 °C and benzhydrylamine (3.47 g, 18.9 mmol) was added in one portion and stirred for 16 h. Methanol (61 mL) was added, and the reaction mixture was distilled to a volume of approximately 25 mL. Methanol (40 mL) was added, and the reaction mixture was distilled to a volume of approximately 20 mL. The resulting slurry was filtered and rinsed with two portions of methanol (15 mL each) and dried under vacuum to afford the product. Yield m = 6.5 g, 15.1 mmol, 79.9 %. Synthesis of (R,E)-N-((3,6-dibromopyridin-2-yl)methylene)-2-methylpropane-2-sulfinamide (15) 3,6-Dibromopicolinaldehyde (1 g, 3.7 mmol) and (R)-2-methylpropane-2-sulfinamide (0.48 g, 3.9 mmol) were combined in Nmethyl-2-pyrrolidone (20 mL). To the reaction mixture was added Cs2CO3 (0.55 g, 4.1 mmol). The reaction mixture was stirred for 2 h then cooled to 0 °C. Water (20 mL) was added to the reaction mixture. The resulting suspension was stirred for 1 h, solids isolated by filtration, washed with water (5 x10 mL) and dried to provide the title compound. Yield m = 0.9 g, 2.4 mmol, 62.5 %. ) To a dried round bottom flask under an inert atmosphere, (R,E)-N-( Synthesis of (R)-1-(3,6-dibromopyridin-2-yl)-2-(3,5-difluorophenyl)ethan-1-amine (R-5a) To a stirred solution of (R)-N- General procedure (A) for synthesis of 2-phenyl-1-(pyridin-2-yl)ethan-1-one and derivatives To a solution of the corresponding 2-pyridinecarbonitrile (1 equiv) in dry Et2O at 0 o C under an inert atmosphere, either benzylmagnesium bromide or (3,5-difluorobenzyl)magnesium bromide was added dropwise (1.2 equiv, 1 M in THF). |
666bf9b7c9c6a5c07a7379e5 | 13 | The reaction was stirred for 30 min and then allowed to warm to room temperature and stirred for a further 4 h. The reaction was quenched with 2 N HCl and stirred for 10 min. After separation the aqueous phase was neutralized to pH 8 with NaOH and extracted with CH2Cl2 (3 x 20 mL). The organic layers were combined, washed with water, brine and dried over MgSO4. The organic phase was concentrated in vacuo to afford the crude product. Products were purified by flash chromatography using solvent gradient of 0:100 → 50:50 cyclohexane:EtOAc over 20 min. |
6494a37924989702c2cc453f | 0 | Nitric oxide (NO) plays an important role in several biological processes, from biofilm inhibition to controlling inflammation and cancer treatment . However, NO is intrinsically unstable, and interest in stable radicals with analogous behaviour, such as (2,2,6,6tetramethylpiperidin-1-yl)oxyl (TEMPO) has rapidly grown. The TEMPO molecule, as shown in Figure , has been reported as effective in anti-biofouling applications , with its potential for use in chronic wound treatment also highlighted . In high doses, TEMPO is lethal to bacterial cells and bio-films . In low doses, it enters the cell and mimics nitric oxide behaviour as a signalling molecule, including, for instance, causing P. aeruginosa biofilm dispersion by reverting sessile cells to a planktonic state through the overstimulation of twitching motility . |
6494a37924989702c2cc453f | 1 | To effectively use TEMPO as an anti-biofouling agent, the capability to coat the surface of a wide range of materials and preserve the nitroxide moiety is required. Recently, plasma polymerisation has been shown to fulfil these requirements . Plasma polymerisation is a onestep, solvent-free, and energy-efficient process to coat surfaces with thin, functional films of organic molecules without altering the substrate material bulk properties. As a result, it is exceptionally versatile across a wide range of industrial applications such as antifouling , biomolecule immobilisation , water purification , electronics , and green energy . Although plasma polymerisation is considered a mature technology, many fundamental questions remain around the initial growth stages and the uniformity and ordering within the polymer layers. |
6494a37924989702c2cc453f | 2 | As conventionally deployed, plasma polymerisation is the cross-linking of small molecules on top of a collecting substrate due to molecule fragmentation in the energetic environment. The resulting polymers, whilst exceptionally stable, are thus disordered, and often have poor resemblance to the monomer . However, foremost low power and more latterly higher pressure can reduce the degree of monomer fragmentation . The combination of higher pressure and low power (γregime) reduces the degree of monomer fragmentation in the gas-phase and at the collecting surface due to `softlanding' of ions which improves functional group retention . These ions, comprising intact or protonated monomers and dimers, are often sufficient to account for the whole of the mass flux . Though there is still a significant debate about the contribution of ions to the film-forming species . High pressure and low power are thus optimal conditions to deposit molecules as intact as possible to study their interaction with the surface and the order of the polymer. Furthermore, recent studies challenge the typically held belief that plasma polymerisation is a surfaceindependent process. The properties in the early stages of plasma polymerisation differ from thicker films in terms of chemistry , morphology , growth rate , and adhesion because the monomers land on the pristine substrate rather than an already coated polymeric layer. |
6494a37924989702c2cc453f | 3 | Here we test the hypothesis that (1) plasma polymerisation is surface-dependent and (2) the substrate can promote ordered growth by studying the early stages of TEMPO plasma polymerisation. TEMPO thin films are prepared by plasma polymerisation on substrates with significantly differing chemistries and atomic roughness. We demonstrate, through high-resolution atomic force microscopy (AFM) and X-ray photoelectron spectroscopy (XPS), that TEMPO films are highly surface-dependent up to 30 nm, exhibiting distinct morphology and chemistry. Beyond 30 nm, the films converge to a flat, uniform structure, and are therefore considered surface-agnostic. Through a post-annealing process, we show that it is possible to grow stable single layer TEMPO coatings with an ordered arrangement on HOPG substrates. Density functional theory (DFT) provides important insights into the molecule-molecule and surface-molecule interactions, revealing an antiparallel arrangement caused by significant dipole-dipole interactions between TEMPO molecules. Moreover, the molecules are found to arrange with the NO either pointing directly upwards or downwards, perpendicular to the surface, suggesting preferential ordering of up to half of the NO groups in the molecular layer may be possible. The ability to direct and optimise the orientation of the active NO group could have significant implications for the use of TEMPO as a surface coating. |
6494a37924989702c2cc453f | 4 | Surface-dependent growth of TEMPO polymers. It has been previously reported that TEMPO deposited by plasma polymerisation results in three distinct structures as shown in Figure ; the intact molecule (NO), damaged molecule (N), and a protonated structure (NH2 + ), which are each expected to behave differently against bacteria. Moreover, the orientation and intermolecular packing of TEMPO molecules within plasma polymer films are expected to depend on the surface, so an understanding of growth and its impact on topography (Figure -e), surface area, and chemistry are essential. |
6494a37924989702c2cc453f | 5 | TEMPO growth was investigated using SiO2, HOPG, and Au(111) as substrate materials. SiO2 substrates are frequently used to characterise plasma polymerisation deposition and are therefore used as a reference substrate. In comparison, Au(111) and HOPG are wellstudied atomically flat materials known to promote good ordering in monolayer molecular films . Furthermore, whilst both are atomically flat, the chemical activity of Au(111) and HOPG are different; a property that is frequently exploited for molecular self-assembly on Au(111) . This combination of properties therefore allows us to gain insight into the effect of surface topography and chemical reactivity on the early stages of plasma polymerisation, and the potential to grow ordered monolayer films. |
6494a37924989702c2cc453f | 6 | Three stages of growth were studied with increasing layer thickness, which we label: i) sub-monolayer, ii) intermediate stage, and iii) converged film, corresponding to 5, 300 and 600 seconds of plasma deposition, respectively (see methodology for details of the plasma conditions used). The resulting films were studied using atomic force microscopy (AFM) and X-ray photoelectron spectroscopy (XPS) in order to characterise the topography and chemistry of each sample. |
6494a37924989702c2cc453f | 7 | AFM measurements of sub-monolayer TEMPO films are shown in Figure -c for SiO2, HOPG, and Au(111), respectively, each showing clearly distinct surface morphology following 5 seconds of growth. Reference images of the pristine substrates pre-deposition can be found in Figure . On SiO2 (Figure ) molecular growth appears uneven, with AFM images revealing a preference for TEMPO to collect in small molecular clusters with heights < 2 nm. In contrast, on HOPG, we observe large atomically flat single layer islands 0.6 ± 0.2 nm in height (orange in Figure ) suggesting layer by layer growth. Finally, on sputter-annealed Au(111) substrates, molecular growth again appears flat, with AFM images revealing a higher coverage of TEMPO compared to HOPG (see Figure ). By operating the AFM with a high loading force on TEMPO:Au(111), a scratch test was performed, resulting in a measured layer thickness of 3 ± 1 nm. This is supported by XPS measurements (discussed in detail below), which show a significant increase in the measured N 1s signal on Au(111), with measured peak areas of 190, 105, and 370, for SiO2, HOPG, and Au(111) respectively. We note that despite many attempts, it was not possible to prepare a sub-monolayer coverage on Au(111). We attribute this to the higher surface binding of Au(111) compared to SiO2 and HOPG, and the difficulty in preparing a stable plasma for deposition times lower than 5 seconds. |
6494a37924989702c2cc453f | 8 | Variations in TEMPO film topography are observed at layer thicknesses up to 30 nm. Figure -g shows AFM images collected at what we term the intermediate stage, corresponding to 300 seconds of plasma polymer deposition. The topography images share similar characteristics to the single layer images, suggesting that the observed variations in initial layer formation influence growth for several layers during the early stages of plasma polymer deposition. After 600s, AFM topography appears to converge across samples, with comparable RMS roughness of 280 pm, 330 pm and 290 pm for SiO2, HOPG and Au(111), respectively (Figure ). Molecular thicknesses of 30 ± 2 nm, 28 ± 2 nm and 35 ± 2 nm were measured on SiO2, HOPG and Au(111), respectively, determined by an AFM scratch test (Figure ). |
6494a37924989702c2cc453f | 9 | The chemistry of TEMPO layers was studied with XPS, providing insight into changes in chemical structure and attachment resulting from surface interactions. Overlaid XPS spectra collected in the N 1s region are shown in Figure , h, and i for deposition times of 5s, 300s, and 600s, respectively (see Figure for the fitted spectra). As a guide, we also show dotted lines indicating three common nitrogen environments (N, NO, NH2+) corresponding to the amine groups (primary, secondary and tertiary), the nitroxide group, and protonated monomer at 399, 400, and 401.5 eV, respectively. For samples exposed to 5 seconds of TEMPO plasma, we see that the N 1s environment on SiO2 and HOPG is centred at 399.2 eV due to a dominating N component and exhibit a tail at higher binding energies attributed to the NH2 + environment (Figure ), consistent with previous observations . In comparison, the N 1s spectra on Au(111) is significantly wider, and is centred at 400 eV, which is indicative of multiple chemical environments. The precise origin of this broadening is difficult to determine, and could be attributed to either an increased relative abundance of NO in the TEMPO film, (i.e. reduced molecular damage and higher functional group retention), or potentially the presence of nitrogen interaction with the Au surface . |
6494a37924989702c2cc453f | 10 | Based on the assumption that the methyl groups provide a steric barrier for N-Au binding, we propose that accelerated growth on Au(111) limits the time TEMPO spends in the plasma environment, therefore reducing molecular damage and increasing NO retention. As the film forms, the surface sticking coefficient will reduce, resulting in similar growth rates across all samples. This appears to be confirmed by XPS spectra measured on thicker TEMPO films deposited at 300 s and 600 s, which show similar structure in the N 1s region where peaks are found to converge to a single energy of 399.2 eV, identical to that of the 5 second TEMPO film on SiO2 and HOPG. Overview spectra for each sample discussed here can be found in Figure . We note that control experiments, where the plasma was not ignited, result in virtually no deposition (Figure ), highlighting the necessity of the plasma environment. We also note that XPS could not be collected on the TEMPO molecular powder, which is not stable under ultra-high vacuum environments, and sublimates even at room temperature. To briefly summarise these results, we find that plasma polymer growth of TEMPO films strongly depend on the surface material in the first 30 nm of growth. XPS reveals that the TEMPO chemistry is most affected in the first layer of growth, beyond which molecules exhibit a consistent chemical structure. AFM measurements, however, indicate significant variation in topography up to 30 nm thickness. Beyond 30 nm thickness, the coatings studied become 'surface agnostic', converging to a flat, uniform layer of comparable roughness. |
6494a37924989702c2cc453f | 11 | Temperature induced ordering of TEMPO thin films. Motivated by the consistent single layer growth observed on HOPG at low deposition times, we explored the ability to improve ordering by post-annealing single layer TEMPO:HOPG samples at 50 °C under atmospheric conditions. We once again note that the sublimation temperature of is relatively low, with temperatures in excess of 50 °C likely to remove the surface coating. Post-annealing at 50 °C is therefore expected to provide sufficient energy to the molecular system for the TEMPO layer to reach its minimum energy configuration. Before and after AFM images of annealed single layer TEMPO:HOPG and their corresponding height histograms are shown in Figure . Prior to annealing, the molecular layer height is measured as 0.6 ± 0.2 nm. Post-annealing, the height is found to increase to 0.9 ± 0.1 nm. We note that measured changes in height are consistent across a range of samples, and comfortably within the resolution of the AFM used (the measured difference of 0.3 nm is comparable, for instance, to the step height of HOPG). XPS data collected on the same samples are shown in Figure , and confirm that there is no significant difference in the binding energy or area of spectra measured in the N1s region, indicating that the amount of TEMPO on the surface is unchanged. These observations are now discussed within the context of computational modelling, from which we propose that the measured increase in height is a result of improved molecular ordering. |
6494a37924989702c2cc453f | 12 | Assuming that monomers account for the majority of the flux to the surface 10 , we first model the electronic properties of single TEMPO monomers in gas phase and then adsorbed on graphite using a combination of DFT and the counterpoise method . As a first step, we explore two different orientations for two TEMPO molecules in gas phase (Figure ): parallel NO groups (denoted up-up) or anti-parallel (denoted up-down). Figure demonstrates that the latter is the most energetically favourable with an energy difference of approximately 0.10 eV. This is a result of the electrical dipole moments preferring an anti-aligned orientation, which forces the oxygen atoms in the two neighbouring TEMPO molecules into the up-down orientations (for more details see TEMPO-TEMPO gas phase simulations in the SI and Figures ). |
6494a37924989702c2cc453f | 13 | Initially we modelled a single TEMPO molecule on a graphite surface, which has many possible binding geometries . The process of identifying preferred adsorption sites was therefore divided into two stages. First, we modelled a single TEMPO molecule 'scanned' across the graphite surface in a series of 1000 steps with different locations and orientations using a GGA + van der Waals functional (Figure ) . The strongestbinding configurations from this process with the corresponding binding energies are shown in Figure . This reveals that there are four favoured configurations corresponding to four energy minima. We find that the three most stable structures (1-3) correspond to geometries with upwards facing NO groups, with a slight preference for a tilted orientation. The two strongest binding configurations (1 and 2) correspond to TEMPOgraphite binding energies of approximately 0.6 eV, while the third and fourth minima (3 and 4) have a weaker binding energy of 0.4 eV. Simulations were also carried with the more conventional GGA functional as shown in Figures (the difference between binding energies obtained using vdW and GGA functionals is approximately 0.2 eV). |
6494a37924989702c2cc453f | 14 | As an approximation for the complete molecular layer, we simulated two TEMPO molecules (dimer) on the graphite surface. Our calculations reveal that the TEMPO-TEMPO interaction is substantially higher compared to the TEMPO-graphite interaction (Figures ). By computing the total energy of a TEMPO-TEMPO dimer as a function of their separation, we find that antiparallel orientations are always the lowest energy arrangement (Figures and), and that an upright antiparallel geometry is the most favourable. This is due to the upright arrangement maximising the dipole-dipole interaction between molecules. This was tested by artificially increasing the TEMPO-TEMPO spacing such that the molecules become well separated and the dipoledipole interaction is reduced. In this case, the molecules return to a tilted orientation similar to the individual monomers. Simulations for the TEMPO-TEMPO dimer on graphite are shown in Figure , involving calculations in multiple locations with different positions and orientations. This shows that the upright configuration is the most stable, with a film thickness (i.e., the distance between HOPG and the highest atom of the TEMPO molecule) of 0.87 nm. In comparison, the tilted configuration is the third most stable configuration, resulting in a calculated film thickness of 0.65 nm. |
6494a37924989702c2cc453f | 15 | The calculated molecular layer heights are in good agreement with the two experimental heights of 0.6 ± 0.2 nm and 0.9 ± 0.1 nm, before and after annealing, respectively. The height measurement of the as-deposited TEMPO layer in Figure , agrees with that predicted for a tilted configuration, suggesting that prior to annealing, the molecular arrangement is more disordered and the TEMPO-TEMPO interaction weaker. In comparison, the measured height of the postannealed sample in Figure compares well to the calculated upright antiparallel arrangement (structures 1 and 2). These results are consistent with the expectation that postannealing will provide sufficient energy to the molecular system to promote reorganisation, therefore resulting in the most stable molecular configuration consisting of a 50/50 arrangement of upwards and downwards facing functional NO groups. |
6494a37924989702c2cc453f | 16 | It is important to note that in reality the plasma polymerisation process results in a mixture of molecular monomers and a mixture of macromolecular chains of varying length and varying degrees of molecular fragmentation. As discussed in detail elsewhere , mass spectrometry confirms that the plasma conditions primarily produce intact monomers, however, there will also be a distribution of fragmented monomers with varying degrees of damage in the NO and methyl groups (which facilitate molecular polymerisation). It is therefore clear that the modelling presented here will not fully capture the complexity of the TEMPO polymer system, nevertheless, we believe it serves as a good approximation considering the expected monomer dominated composition of the film. |
6494a37924989702c2cc453f | 17 | We have shown that plasma polymerisation can be used as a method for the ordered growth of monolayer films that are otherwise challenging to deposit. An analysis of the early stages of plasma polymer film growth confirms a significant surface dependence in the first 30 nm of growth, beyond which molecular films converge to a flat uniform coating. Through an investigation of the stable free radical TEMPO molecule, we have found that plasma polymerisation followed by a sample annealing process results in a stable, ordered monolayer film where TEMPO molecules are predicted to adopt an antiparallel arrangement with up to 50% of the functional NO groups aligned perpendicular to the surface. These results provide new insights into the importance of the early stages of plasma polymer film formation and provide new routes for the development of anti-biofouling polymer coatings. |
6494a37924989702c2cc453f | 18 | Plasma polymerisation. Plasma polymerisation of TEMPO (Alfa Aesar, purity ≥ 98%) was performed with a custom-built glass barrelled plasma reactor. The glass barrel of the reactor was a QVF process pipe (De Dietrich, UK) -dimensions 500 mm length, 100 mm diameter, clamped between two custom-made steel end plates which serve as ground electrodes. The system was evacuated using an Edwards RV3 rotary vane pump (base pressure: c. 2 x 10 -3 mbar). The pressure within the system was monitored by a Thermovac TTR 96N SC Pirani gauge and Display One Controller (Leybold UK Ltd). An in-line liquid nitrogen cold trap was installed to enhance the system base pressure and prevent monomer vapours from reaching the rotary pump. Plasma was ignited using a RFG-C-100-13 power generator and automatic matching network (Coaxial Power System Ltd, Eastbourne, UK), operated at a frequency of 13.56 MHz, connected to a 1 cm thick copper braid (Tranect Ltd, Liverpool, UK) wrapped around the glass barrel three times. Sample wafers include SiO2 (Inseto, UK), HOPG (Mickromasch), and Au(111):mica (Georg Albert PVD) placed 12 cm away from the ground electrode through which the monomer enters the reactor. SiO2 was precleaned using ultrasonication, HOPG was exfoliated with scotch-tape, and Au(111) was cleaned via standard sputter-annealing cycles under UHV conditions. After the chamber was evacuated the monomer valve was opened and the pressure adjusted to reach 10 Pa by partially closing the valve to the pump and keeping the monomer flask in a water bath at 25 °C. The polymerisation was performed at a nominal power of 5 W for 5, 300, and 600 s after the plasma ignition. |
6494a37924989702c2cc453f | 19 | X-ray photoelectron spectroscopy. XPS was performed using a Kratos Analytical AXIS Supra spectrometer with a monochromatic Al Kα 1486.7 eV Xray source, operating at 15 kV, 15 mA, and equipped with an electron gun for charge neutralisation. The binding energies were referenced to the C 1s peak at 285.0 eV. For the 5 s set of samples on HOPG, the C 1s peak was referenced to 284.5 eV due to the pure sp2 hybridisation of the substrate dominating the polymer signal. All the spectra were analysed using CASAXPS (Casa Software Ltd, UK). Atomic force microscopy. AFM images were collected using a Bruker Multimode 8 AFM equipped with a Nanoscope V controller and operated in the PeakForce mode. NuNano Scout 70 probes were used with a nominal resonant frequency of 70 kHz, 2 N/m cantilever stiffness, and less than 10 nm probe radius. To measure the polymer thickness, the coating was first scratched in contact mode in a 2 µm x 2 µm area and then scanned in peak force mode to check that the film was completely removed. The thickness was computed from peaks of height distribution corresponding to the area inside and outside the scratch. The uncertainty associated with the measurement depends on the peaks FWHM. The RMS roughness was computed using Gwyddion by analysing 500 nm x 500 nm scans. Density Functional Theory. Using the density functional code SIESTA the optimum geometries of isolated TEMPO molecules were obtained by relaxing the molecules until all forces on the atoms were less than 0.01 eV / Å. A double-zeta plus polarization orbital basis set, norm-conserving pseudopotentials, and an energy cut-off of 250 Rydbergs defining the real space grid were used. Both GGA and van der Waals functionals were employed to find the most energetically favourable structures. |
6494a37924989702c2cc453f | 20 | Binding Energy Simulations. To calculate the optimum molecule-molecule and molecule-surface binding distances, we used DFT and the counterpoise method, which removes basis set superposition errors (BSSE). The optimum binding distance z is defined as either the optimum TEMPO-TEMPO or TEMPO-HOPG distance. If one of the entities (TEMPO or HOPG) is defined as entity A and the TEMPO or graphite as entity B, then the ground-state energy of the total system is calculated using SIESTA and is denoted E AB AB . The energy of each entity is then calculated on a fixed basis, which is achieved in SIESTA using ghost atoms, (basis set functions which have no electrons or protons). Hence, the energy of the individual TEMPO or graphite in the case of the fixed basis is defined as E A AB and for the TEMPO or graphite as E B AB . The binding energy is then the difference between the isolated entities and their total energy when placed a distance 𝑧 apart, which is calculated using the following equation: Binding Energy = 𝐸 𝐴𝐵 𝐴𝐵 -𝐸 𝐴 𝐴𝐵 -𝐸 𝐵 𝐴𝐵 , as shown in Figs. |
62088fce0c0bf03b8fe7cb21 | 0 | High explosives (HEs) have contributed enormously to the prosperity of humankind since the invention of black powder in China and the era of nitroglycerin brought by Alfred Nobel . In an HE, the chemical energy stored in the bonds of the explosive material is converted into kinetic energy of the gaseous products via chemical reactions. A quantitative as well as qualitative understanding of the complex reaction network is critical to the application of HE. Such a reaction network involves thousands of elementary reactions that take place over a wide range of time scales. In addition, the reactions usually occur in extreme conditions (i.e., > 10 GPa and > 3000 K ), which makes it very difficult or even impossible to obtain the detailed reaction mechanisms of HE materials using experimental or conventional simulation approaches. |
62088fce0c0bf03b8fe7cb21 | 1 | In recent decades, ab initio molecular dynamics (AIMD) simulations have been applied to gain atomic insights into HE materials . In an AIMD simulation, an HE material is represented by an atomic model with interatomic forces determined by electronic structure calculations, and the Newtonian equations of motion are solved to obtain the dynamic trajectories. Thus, AIMD allows chemical reactions (i.e. bond breaking and forming events) to occur and accounts for electronic polarization effects. |
62088fce0c0bf03b8fe7cb21 | 2 | AIMD has been applied to investigate the shock-induced mechanical response and initiation of a detonation in numerous studies, providing fundamental insight into the microscopic mechanisms of such complex phenomena. However, the computational cost of AIMD is so high that the accessible system size and time scale in simulations are limited to several hundred atoms and dozens of picoseconds, respectively. |
62088fce0c0bf03b8fe7cb21 | 3 | Recently, artificial neural networks (NNs) have been applied to construct potential energy surfaces (PESs) in a fully data-driven manner, where the PES is abstracted from a well-selected training dataset using suitable functional expressions . Several formulations, such as DeepMD , GAP , sGDML , and SchNet , have been proposed to develop NN potentials and have achieved success in the modeling of water , small organic molecules , and metal materials . The performance of NN potential depends on the completeness of the training dataset . In other words, NN does well in finding solutions in the function space of the training dataset but might fail in configurations outside the dataset. Therefore, the quality of the dataset is critical to developing NN potentials, and it is recommended to include all the critical configurations during potential reaction processes . MD sampling is the most straightforward way to construct the training dataset, and the evolution of energies and forces are recorded as training datasets . Various bond dissociations and recombinations exist in a reactive system, e.g., the decomposition of HE materials. Such processes cannot be sampled appropriately in classical MD sampling due to the high energy barriers. Enhanced sampling techniques, such as metadynamics , , could improve the sampling quality in the potential reaction coordinates. The concept of interactive molecular dynamics in virtual reality (VRMD) was first proposed by Glowacki et al. and applied to sample the PES of a simple hydrogen abstraction reaction . This method has been proven to accelerate the intelligent curation of high-quality datasets. In a VRMD simulation, VR forces are implemented on atoms to accelerate the reaction processes, similar to metadynamics. |
62088fce0c0bf03b8fe7cb21 | 4 | In this paper, we develop a neural network-based molecular dynamics (NNMD) method for investigation into the reaction dynamics of a novel high explosive (ICM-102) . The training dataset is derived from both AIMD and VRMD simulations. The configuration spaces of datasets from different sampling methods are compared to yield a good NN potential. A set of MD simulations is performed to explore the elementary reactions in the decomposition of ICM-102 molecules. A panoramic visualization of the complex reaction networks is abstracted, and new pathways are identified from atomic trajectories for the first time. |
62088fce0c0bf03b8fe7cb21 | 5 | DFT calculations were performed using the CP2K package . Core electrons were treated using Goedecker-Teter-Hutter (GTH) pseudopotentials and the Perdew Burke Ernzerhof generalized gradient approximation method . The Grimme DFT-D3 method was used to account for dispersion interactions. A double-zeta Gaussian basis set plus polarization (DZVP-MOLOPT) was considered. AIMD calculations were performed for an ICM-102 system of 160 atoms using the Quickstep module in CP2K. In both cases, simulations were carried out at constant volume and temperature conditions with periodic boundary conditions. An integration time step of 0.5 fs was used. |
62088fce0c0bf03b8fe7cb21 | 6 | We trained NN potentials using the DeePMD-kit package . The smooth version of the deep potential model is adopted with a cut-off radius of 6.0 Å . To remove the discontinuity introduced by the cut-off, the 1/r term in the network construction is smoothly switched off by a cosine shape function from 1.0 Å to 6.0 Å. |
62088fce0c0bf03b8fe7cb21 | 7 | The filter (embedding) network has three layers with (25, 50, 100) nodes/layer, and the fitting net is composed of three layers, with 240 nodes each. The network is trained with the ADAM optimizer, with an exponentially decaying learning rate from 1.0 × 10 - 3 to 5.0× 10 -8 . During the optimization process, the pre-factors in the loss function change from 1 to 10 and 1000 to 1 for the energy and force terms, respectively. The final NN model used for the production run was trained for 1.0 × 10 6 steps. Additional configurations were obtained using an active learning sampling procedure as implemented in the DP-GEN package . Four NN potentials were trained on the same dataset with different initializations of weights and biases. NVT-MD simulations were performed at multiple temperatures (300, 3000, and 4000 K) using the LAMMPS package . By comparing the structures from MD simulations, the agreement on the force predictions made by these potentials is used to select new configurations. When the deviation of NN prediction for one configuration was in the range of [0.4, 0.8] eV/Å, the corresponding structure was labeled as a candidate for the training set. The upper limit is used to filter nonphysical configurations . |
62088fce0c0bf03b8fe7cb21 | 8 | The reaction network was postprocessed using ReacNetGenerator by Zeng et al. from the NNMD trajectories. All the species in the reaction network are clustered by their fingerprints using the scikit-learn package . The fingerprints are defined by the molecular structure and calculated by RDKit . The importance of each species is calculated according to the observed number that it occurs and is represented by the size of the dot. |
62088fce0c0bf03b8fe7cb21 | 9 | The performance of NN potential depends on the quality of the training dataset, which describes the chemical space of PES . The whole training dataset of bulk ICM-102 molecules is constructed using both AIMD and VRMD sampling methods, and the detailed configurations are listed in Table . The structural landscapes of ICM-102 decomposition are shown in Fig. , where the molecular configurations are projected onto their first two principal components (PCs). The full configurations can be divided into crystal, partial decomposition, and gas species. The most common structure is the crystalline phase of ICM-102, represented by a sheet-like structure. As the decomposition reaction's starting point, the crystal group has the lowest potential energies, as expected. Along with the 1 st principal component (PC1), crystalline ICM-102 gradually turns into states undergoing partial decomposition. The molecules become irregular under thermal stimulation, and the loss of hydrogen atoms is observed. |
62088fce0c0bf03b8fe7cb21 | 10 | decomposed. Figure shows the average molecular masses along PC1. When PC1 < -2, the average molecular mass is 202 a.u., corresponding to the mass of the ICM-102 molecule. As PC1 increases, the molecular mass gradually decreases, indicating the start of the decomposition reaction. Therefore, we conclude that PC1 refers to the overall reaction coordinate, where the initial stable reactants gain energy to overcome energy barriers and reach a stable product state. Then, the distributions of subsets are compared. The first subset is, in fact, obtained from AIMD simulations at 300-4000 K, combined with an active learning sampling strategy . An additional dataset of VRMD is constructed from the intuition of the first author using an in-house VRMD simulator (Manta). It allows researchers to explore the configuration space of chemical reactions and performs expert-biased enhanced sampling on the targeted PES. In the VRMD dataset, reaction pathways include the transfer of hydrogen atoms between ICM-102 molecules, ring-opening reactions and the formation of gas molecules (Fig. ). A detailed distribution of each subset along PC1 and energy is included in The distribution of AIMD and VRMD datasets. |
62088fce0c0bf03b8fe7cb21 | 11 | To capture the reactive dynamics of ICM-102 molecules, 100 ps MD simulations were performed using the new NN potential (e.g., case 4 in Table ) with 64 ICM-102 molecules at 3000 K. The species evolution during ICM-102 decomposition predicted by NN and AIMD are compared in Fig. . It is clear that NN agrees well with the AIMD results, where ICM-102 molecules are rapidly consumed, and H2O is the main product. With the application of NN potential, the system size could be extended to thousands or even millions of atoms . Compared to the large oscillation in the predicted species evolution of AIMD, NNMD simulates an eight-times larger system, resulting in smooth species evolution. Such a large system allows studies on the macroscopic nature of a statistical ensemble, which is beyond the ability of the traditional AIMD method. In total, 5799 products are identified from the NNMD trajectories. Direct derivation of the reaction mechanism from thousands of products can be quite challenging. We propose an interactive reaction network to illustrate the detailed reaction mechanism. In Fig. , the full reaction network contains detailed pathways and intermediates. These intermediates are represented by fingerprints and clustered into eight groups using the k-means clustering algorithm . Groups 1 and 2 represent ICM-102 and molecules produced by releasing H, O and OH radicals, respectively. |
62088fce0c0bf03b8fe7cb21 | 12 | Groups 3 and 4 share similar elemental compositions as groups 1 and 2 but with broken C-N rings. The decomposition of groups 5, 6, and 7 produces intermediates with a low H/O ratio (see Figure ). Group 8 represents small gas products, including CHNO, H2O, OH, NO, and N2. These species are frequently involved in the reaction network as the product of many reaction pathways. There are two pathways for reactants to release the final gas products: group 1 => 2 => 8; group 1 => 3 and 4 => 5 and 6 => 7 Derived from the overall ICM-102 decomposition (Fig. ), the primary reaction pathways of ICM-102 and the formation of H2O are constructed in Fig. . ICM-102 decomposition starts with intermolecular and intramolecular H transfer to form R-OH species (R1 and R7). These species further decompose into radicals such as OH and H (R2 and R10). The NO2 molecule also abstracts a hydrogen atom from ICM-102 to form HNO2 (R8), which further dissociates as an OH radical (R9). The combination of OH and H radials forms water molecules (R6). The R-OH structures could also undergo ring-opening reactions by C-N bond scission (R3), and this reaction is not preferred due to the high bond energy (305 kJ/mol for C-N bonds ). From the above discussion, the NNMD method breaks through the limitations of AIMD. On the one hand, it extends the timescale of simulation to allow observation of the reaction pathways related to N2, CHNO, CO2 and many other intermediates of ringopening reactions; on the other hand, it expands the system size to thousands of atoms to obtain statistically significant results on the branching of detailed reaction pathways. |
62088fce0c0bf03b8fe7cb21 | 13 | In summary, ICM-102 decomposition reactions are explored using NN-based molecular dynamics simulations. This is the first study to achieve a panoramic view of the complex explosive reaction process with an ab initio level of accuracy. NNMD extends simulation of high explosives to the sub-nanosecond scale, which helps the development of high accuracy kinetics models and promotes the design and application of high explosives. |
62088fce0c0bf03b8fe7cb21 | 14 | AIMD and VR enhanced sampling. with a small number of additional configurations. We believe such a reduced dataset is extremely valuable to the application of supervised learning in the field of computational chemistry because dataset labeling is the most time-consuming step in supervised learning, corresponding to electronic structure calculations in NN potential training. A small improvement in accuracy requires a great effort in computational resources. Recently, it has been found that the computational efforts at high-level ab initio calculations can be minimized by combining datasets of molecular forces from different levels of theory . However, such a strategy is highly dependent on the wellselected training set of high-level ab initio calculations. We are working on integrating the VR-enhanced sampling algorithms to develop NN potentials with a high level of accuracy. |
62088fce0c0bf03b8fe7cb21 | 15 | Gaps between the MD simulations and experiments still exist in explosive decomposition reactions. Although the landscape of ICM-102 molecules has been drawn via AIMD and VR-enhanced sampling, further work is required to assemble reaction networks into kinetics models to interpret macro-phenomena in diverse fields, such as combustion, catalysis, and atmospheric chemistry. There are many wellestablished methods, such as model reduction and sensitivity analysis, to construct kinetic models and describe the macroproperties. Importantly, the present work constructs a novel VR-enhanced sampling method to train NN potentials and enables MD simulations of a novel high explosive (i.e. ICM-102) with complex reaction networks at the level of ab initio calculations, providing atomic insights into the detailed reaction pathways that are otherwise difficult to uncover, such as ring-opening reactions and the formation of N2, CHNO, and CO2. Thus, the study opens up new possibilities to build reaction kinetics models based on high-fidelity and low-cost AI algorithms. |
62088fce0c0bf03b8fe7cb21 | 16 | We develop a workflow combining ab initio molecular dynamics and interactive molecular dynamics in virtual reality (VRMD) to construct a neural network potential of a novel high explosive. The VRMD helps sample the reactive process and significantly improves the model performance. Our potential enables the large-scale MD simulation with an ab initio level of accuracy, achieving a panoramic view of the complex explosive reaction process. |
63d42e751125960c4f7cc083 | 0 | Biomass burning events emit a large amount of volatile organic compounds (VOCs), which can be oxidized in the atmosphere, leading to secondary organic aerosol (SOA) formation by gas-toparticle partitioning of the oxidized products. Phenolic compounds such as catechol, guaiacol, and syringol (Fig. ) are major types of VOCs from biomass burning events that lead to the formation of biomass burning SOA (BB-SOA) in the atmosphere. BB-SOA and other particulate matter emitted from biomass burning play an important role in climate and air quality. They alter climate directly by absorbing and scattering radiation, as well as indirectly by acting as nuclei for clouds. BB-SOA also impacts air quality, since they penetrate into the respiratory system, causing adverse health outcomes. Furthermore, the atmospheric load of BB-SOA is expected to increase in the future due to the growing frequency and size of forest fires as a result of climate change. The viscosity of BB-SOA is needed to predict important atmospheric processes related to climate and air quality. For example, the long-range transport of pollutants depends on SOA viscosity. In addition, predictions of gas-particle partitioning (and hence, mass and size distributions) depend on the viscosity of SOA. Similarly, heterogeneous chemical reactions and photochemical reactions in SOA depend on the mobility and uptake of reactants and products, which can be inhibited at high viscosity. Furthermore, the heterogeneous ice nucleation potential of SOA particles may depend on whether it is in a glassy phase state (η > 10 12 Pa s). In addition to the gaseous components that can lead to SOA formation, biomass burning also directly emits organic aerosols into the atmosphere, which are referred to as biomass burning primary organic aerosol (BB-POA). The viscosity of BB-POA from the pyrolysis of pine wood has been studied previously and was found to depend strongly on relative humidity (RH) and temperature. Similarly, the diffusion rates of organic molecules in BB-POA, which are inversely related to particle viscosity, were also found to depend on RH. In contrast to BB-POA, much less is known about the viscosity of BB-SOA. Only one study has reported a viscosity for BB-SOA: catechol + O3 at 0% RH and room temperature was found to have a viscosity (η) greater than 3 × 10 8 Pa s. With only one reported measurement, considerable uncertainty still exists concerning the viscosity of BB-SOA under relevant atmospheric conditions. |
63d42e751125960c4f7cc083 | 1 | In the current study, the RH-dependent viscosity of four different types of BB-SOA was measured using the poke-flow technique. BB-SOA was generated by the reaction of catechol, guaiacol, and syringol with the hydroxyl radical (OH) in an environmental chamber. In addition, BB-SOA was generated by the dark ozonolysis of catechol. We also developed a novel parameterization for predicting the viscosity of BB-SOA as a function of RH and temperature following a previous approach. Using this parameterization along with RH and temperature conditions in the troposphere, we constructed global viscosity profiles for BB-SOA. Using this information, we also made predictions of the occurrence of glassy BB-SOA particles and the mixing time of organics within them. |
63d42e751125960c4f7cc083 | 2 | BB-SOA was produced in a continuous-flow environmental chamber from the oxidation of catechol, syringol, or guaiacol with OH, and the oxidation of catechol with O3. SOA from O3 oxidation of guaiacol and syringol was not studied because these reactions are much less important than OH oxidation of guaiacol and syringol in the atmosphere (Table ). All of the SOA studied here was generated via OH or O3 oxidation under NOx-free conditions. While high-NOx conditions are relevant for high temperature wildfires, low NOx conditions are relevant for lower temperature wildfires. In addition, low temperature wildfires produce more phenolic VOCs than high temperature wildfires. The environmental chamber used to generate the SOA consisted of a 1.8 m 3 Teflon bag housed in an aluminum enclosure as described in detail elsewhere. The chamber was operated in continuous flow mode, where a pure air generator was used to supply air for the chamber at a flow rate of ~18 L min -1 , yielding a residence time of ~1.7 h. At the chamber inlet, the flow from the pure air generator was split between two lines; one line for the VOC and one line for the oxidant. |
63d42e751125960c4f7cc083 | 3 | For the VOC line, the pure air flowed through a heated (T = ~373 K) round bottom flask and into the chamber. A solution of 2 wt% VOC (catechol, syringol, or guaiacol) was continuously injected from a glass syringe into the flask at a rate of 30-75 µL h -1 using a syringe pump. For experiments with OH oxidation, the solvent was Milli-Q water. For experiments with O3 oxidation, the solvent was 2-butanol, which acted as a scavenger for OH radicals, which can be formed during ozonolysis reactions. The oxidant supply line was set up differently depending on the oxidant being used. For OH oxidation, a 30 wt% solution of H2O2 in Milli-Q water was continuously injected (120 µL h -1 ) into a second heated (T = 303 K) round bottom flask, and the vapors were carried into the chamber with the pure air flow (2 L min -1 ). To induce H2O2 photolysis in the environmental chamber, 24 UV-lights (40W, λmax ~ 360 nm) surrounding the Teflon bag were turned on for the duration of the experiment to generate OH radicals. For O3 oxidation, the pure air instead flowed (1.25 L min -1 ) through an ozone generator which then entered the chamber. For all our ozonolysis experiments, excess O3 concentrations were measured at the chamber outlet to be 360-390 ppb using an O3 monitor (Thermo Scientific, 49i). |
63d42e751125960c4f7cc083 | 4 | The aerosol mass loading at the exit of the chamber was measured using an optical particle counter (GRIMM, 11-S OPC) at the start and end of each experiment. The mass loading ranged from ~5-50 µg m -3 across all experiments. During SOA collection, the chamber was operated at room temperature and the RH was < 1.5%. |
63d42e751125960c4f7cc083 | 5 | At the outlet of the chamber, SOA material was collected onto fluorinated glass slides (12 mm slides; coated with CYTONIX, FluoroPel 800) using either a multi-orifice single stage impactor (MSP Corporation) or a slit impactor (Sioutas Cascade Impactor, SKC Inc.). Each sample was collected continuously for ~7-24 h. During this time, the individual SOA particles coagulated on the glass slide forming larger SOA particles with diameters of ~30-200 µm. At the end of the collection, the sample slide was taken out of the impactor, sealed, and stored at -18 °C until the viscosity measurements. |
63d42e751125960c4f7cc083 | 6 | The room-temperature viscosity of the SOA was determined with the poke-flow technique (Fig. ) that has been described in detail elsewhere. Briefly, the sample slides were placed in a flow cell mounted above an inverted optical microscope. The RH within the flow cell was controlled by adjusting the RH of the carrier gas, using a bubbler system. The RH was calculated from measurements of the temperature and the dewpoint temperature of the carrier gas. After RHconditioning, an ultra-fine tungsten needle with an oleophobic coating was used to poke one of the phenolic BB-SOA particles while recording images with a digital camera coupled to the microscope. At RH values of approximately 20% and greater, poking the particle changed the geometry of the particle from a spherical cap to a half-torus (Fig. , top row). Following the poke, the particle material flowed and returned to its energetically preferred spherical cap geometry. The experimental flow time (exp,flow) was assigned as the time for the area of the hole (A0) to reach ¼A0. The experimental flow time has previously been shown to be proportional to the particle's viscosity. In some experiments with very long recovery times, the experiment was stopped before the inner area could reach 1/4 th its initial value. In these cases, the time from the initial poke until the last image captured was used as a lower limit to exp,flow. In experiments with RH > 40%, exp,flow was shorter than the camera's frame rate (3 fps) and the time required to remove the needle from the center of the particle. When this fast flow occurred, we used a time of 0.5 s (the approximate time to remove the needle from the particle) as an upper limit to exp,flow. In experiments with RH ~ 0%, poking caused the particle to crack as opposed to forming a half-torus geometry (Fig. , bottom row). In this case, no flow was observed even after ~20 h. |
63d42e751125960c4f7cc083 | 7 | The mixing time of water in the BB-SOA particles, τmix,H2O, was calculated following a method used previously. For all experiments reported here, the calculated τmix,H2O was shorter than the conditioning times (Table ). Consistent with these calculations, the measured viscosity was not sensitive to the conditioning time (Section S3 and Fig. ). Based on these criteria, we assume the water activity in the SOA was at near-equilibrium with the RH of the carrier gas during the viscosity measurement. |
63d42e751125960c4f7cc083 | 8 | During the poke-flow experiments, the particles were exposed to a constant flow of gas, which could lead to the evaporation of semi-volatile organic compounds (SVOCs). A loss of SVOCs by evaporation would lead to an increase in the viscosity of the SOA particles. To estimate SVOC loss, the cross-sectional area of several particles was measured over ~24 h (Fig. ). In all cases, the change in cross-sectional area was less than the uncertainty in the measurement except for experiments with guaiacol + OH, which had a change in area of -8.3 ± 4.2% and -9.9 ± 9.1% after 24 h of conditioning (Fig. ). Considering all the evaporation experiments together and the uncertainties in measuring the cross-sectional area, we conclude that the evaporation of SVOCs during conditioning was relatively small. This is consistent with expectations based on the volume of air the particles were exposed to in the environmental chamber (~2.5 × 10 4 L) compared to during conditioning (conservatively < 3.0 × 10 3 L). |
63d42e751125960c4f7cc083 | 9 | Observations from poke-flow experiments were modeled with fluid dynamics simulations to determine the BB-SOA viscosity. The simulations were performed using the micro-fluidics module in COMSOL Multiphysics (v5.4). The following parameters were required as inputs: the particle's surface tension, the Navier slip length, and the contact angle. Conservative upper and lower limits for these parameters were used as inputs (Table ). The combination of these input parameters that yielded the highest and lowest viscosity outputs was used to obtain an upper and lower limit of viscosity. |
63d42e751125960c4f7cc083 | 10 | For experiments where the particles formed a half-torus geometry after poking, the particle's half-torus geometry was imported into COMSOL. In addition, the observed experimental flow time (exp,flow) was used as the time input for the fluid dynamics simulations (sim,flow). The viscosity was then varied until the simulated inner area matched the observation at exp,flow. For experiments where the particles cracked, we used a quarter sphere with a sharp edge as the geometry after the initial poke. The simulation viscosity was varied until the sharp edge of the moved by 0.5 µm (the resolution of the microscope). Since no flow was observed experimentally, the true viscosity of the particle must have been higher than the simulated viscosity, hence the simulated viscosity is a lower limit. |
63d42e751125960c4f7cc083 | 11 | The oxygen-to-carbon (O/C) ratios of each BB-SOA type were determined with a high-resolution time-of-flight aerosol mass spectrometer (HR-ToF-AMS, Aerodyne Inc.) which directly sampled BB-SOA-containing air masses from the environmental chamber. SOA particles were impacted onto a heated (~600 °C) filament, causing the organic particles to vaporize. The vapors were ionized by electron impact ionization (~70 eV). The bulk chemical composition was determined by collecting mass spectra using the MS mode. Air interferences were corrected by particle-filtered air from the chamber, which was sampled with the AMS. To determine the average elemental O/C ratios, mass spectra were collected in V-mode (V-shaped ion flight paths, single ion reflection; resolution of ~2400 m/Δm) and processed by the improved ambient method 69 using the PIKA v.1.25 and SQUIRREL 1.65 software packages in Igor Pro 8 (Wavemetrics Inc.). 200 |
63d42e751125960c4f7cc083 | 12 | Using the poke-flow technique, we determined the viscosity of SOA material generated from catechol + OH, catechol + O3, syringol + OH, and guaiacol + OH (Fig. ). These viscosities were converted from the measured exp,flow values (Fig. ) using fluid dynamics simulations. For each 205 type of BB-SOA, the viscosity increased as the RH decreased, which is expected since water acts as a plasticizer when mixed with highly viscous material. 75 Viscosity measurements at similar RH values were combined to determine overall upper and lower limits (Fig. ). Across all BB-SOA types, the viscosity increased by at least 4-5 orders of magnitude as the RH decreased from ~40 (or 50% for catechol + O3) to ≲ 3%. At RH ≳ 40%, the viscosity was < 3 × 10 3 Pa s, while at RH ≲ 3% the viscosity was > 2 × 10 ) for primary organic aerosol from biomass burning. The purple and orange bands are drawn to help guide the eye to the difference in viscosity between phenolic BB-SOA and BB-POA. For comparison, the viscosity of peanut butter is ~10 3 Pa s and the viscosity of tar pitch is ~10 Pa s. The viscosity of catechol + O3, catechol + OH, and syringol + OH were the same within the uncertainty of our measurements. For guaiacol + OH, the viscosity appears to be lower than the other three SOA types, at least at RH ~ 25%. To attempt to explain this difference, we determined the O/C ratio of each BB-SOA using an AMS, and the results are listed in Table . Viscosity is expected to depend on the O/C ratio, molecular weight, and types of functional groups of the SOA components. Table . Elemental oxygen-to-carbon (O/C) ratio for phenolic BB-SOA from AMS data. The associated uncertainty in the calculated average O/C ratio is 12% (average absolute value of relative error), based on a comparison to previously reported standards. |
63d42e751125960c4f7cc083 | 13 | The O/C ratio of each type was similar, suggesting that the difference between guaiacol SOA and the other types of phenolic SOA is not due to just a difference in this parameter. To understand why the viscosity of guaiacol + OH is less than that of the other three phenolic SOA types at RH ~ 25% requires information on the average molecular weight and types of functional groups of the SOA types. |
63d42e751125960c4f7cc083 | 14 | Figure also shows a comparison between the viscosity of the phenolic BB-SOA and the viscosity of biomass burning primary organic aerosol (BB-POA) from Schnitzler et al. (2022). The BB-POA was generated by smoldering pine wood chips in a tube furnace at 400 °C. For RH values of 0-10%, the viscosity of the phenolic BB-SOA is at least ~2 orders of magnitude higher than the viscosity of BB-POA. |
63d42e751125960c4f7cc083 | 15 | Temperature and RH span wide ranges in the troposphere. To extrapolate our room-temperature viscosity results to other temperatures and RH values, we developed a parameterization to predict BB-SOA viscosity as a function of temperature and RH based on a previous approach (details are provided in Section S7). The resulting viscosity parameterization for catechol + OH SOA as a function of RH and temperature is shown in Fig. . Using the data from catechol + O3 or syringol |
63d42e751125960c4f7cc083 | 16 | Catechol + OH 1.06 0.13 Syringol + OH 1.18 0.14 Guaiacol + OH 1.12 0.13 + OH SOA instead will yield similar results since the parameterization is based on roomtemperature viscosity measurements and our results for these types of phenolic BB-SOA were similar. Since guaiacol + OH SOA had lower viscosity at RH ~ 25%, we expect the results to tend toward lower viscosity compared with catechol + OH SOA. |
63d42e751125960c4f7cc083 | 17 | Based on Fig. , the viscosity of phenolic BB-SOA is strongly dependent on both RH and temperature. Under dry conditions, phenolic BB-SOA is highly viscous at room temperature (~10 Pa s) and becomes a glass (viscosity greater than 10 12 Pa s) when the temperature is only slightly below room temperature (< 280 K). The viscosity of the BB-SOA particles observed here has important implications for heterogeneous chemical reactions and interactions with other compounds co-emitted during biomass burning. An important species co-emitted during biomass burning is BB-POA, which often contains organic molecules that strongly absorb solar radiation, termed brown carbon. This brown carbon component is important for the radiative properties of biomass emissions, i.e. |
63d42e751125960c4f7cc083 | 18 | their direct climate effects. While transported in the atmosphere, BB-POA can undergo various aging processes, such as heterogeneous oxidation reactions that can impact the chemical composition of the brown carbon. Ozonolysis of BB-POA causes a decrease in the absorbance of brown carbon (i.e., whitening) of the BB-POA, thus changing the particle's radiative properties from a warming to a cooling climate effect. showed that the whitening of BB-POA depends on the viscosity of the BB-POA, with whitening rates strongly decreasing with increasing particle viscosity. The strong contrast in viscosity between BB-POA and phenolic BB-SOA (Fig. ) could have considerable implications for the whitening rates of BB-POA in the atmosphere. Since BB-POA and phenolic BB-SOA coexist in biomass burning plumes, BB-POA particles could become coated with or internally mixed with highly viscous phenolic BB-SOA material. In this scenario, the high-viscosity phenolic BB-SOA would slow the uptake of O3, thereby slowing down the rate of whitening (Fig. ). |
63d42e751125960c4f7cc083 | 19 | Combing the viscosity data from Fig. with annual average temperature and RH data in the troposphere following the approach used previously, yields predictions of BB-SOA viscosity as a function of latitude and altitude (Fig. ). This prediction shows that catechol + OH SOA is often in a liquid state (η < 10 2 Pa s) below ~2 km altitude, a semi-solid state (10 2 < η < 10 12 Pa s) between ~2 km and ~9 km, and a glassy state (η > 10 12 Pa s) above ~9 km. The viscosity also has important implications for the cloud formation ability of BB-SOA particles, in particular, their ability to act as ice nucleating particles (Fig. ). Highly viscous, glassy SOA particles are thought to promote ice nucleation in clouds, by providing a solid surface. Our results suggest that phenolic BB-SOA could provide a previously unrecognized source of ice nucleating particles above ~6-9 km in altitude. These altitudes are regularly reached by biomass burning plumes, which can be lifted into the upper troposphere and lower stratosphere by convection. The viscosity also impacts gas-particle partitioning and hence SOA formation and growth in biomass burning plumes. Knowledge of the mixing time of organic molecules in SOA is important for predicting SOA mass and size distributions and the long-range transport of pollutants in SOA. The mixing time of organics within BB-SOA particles as a function of RH and temperature, 𝜏 mix,org (RH, 𝑇), can be calculated as, 𝜏 mix,org (RH, 𝑇) = |
63d42e751125960c4f7cc083 | 20 | where dp is the BB-SOA particle diameter and Dorg(RH, T) is the diffusion coefficient of organics as a function of RH and temperature. To calculate Dorg(RH, T), the values of 𝜂(RH, 𝑇) were substituted into the Stokes-Einstein equation, which is valid when the size of the diffusing molecule is about the same or larger than the size of the matrix molecules that it diffuses through. We estimated the hydrodynamic radius of the molecules in our catechol + OH SOA by assuming a spherical geometry and using the average molecular weight and density from Nakao et al. (2011) (139 g mol -1 and 1.4 g L -1 , respectively). For dp, we used the approximate aerosol diameter of an accumulation mode particle in the troposphere, 200 nm. The mixing times of BB-SOA calculated with Eq. 1 can exceed 1 h at altitudes > 3-7 km (Fig. ). Fast mixing of approximately 0.5-1 h, i.e., at time scales smaller than the typical modeling time step, is often assumed in chemical transport models. Our results challenge this assumption and show significantly longer times are expected for phenolic BB-SOA in many regions of the troposphere. This has important ramifications for the transport of pollutants, such as polycyclic aromatic hydrocarbons (PAHs), which are emitted from biomass burning. PAHs can partition into the condensed phase during the formation of phenolic BB-SOA, allowing them to become trapped inside SOA particles (Fig. ). Once trapped inside, the long mixing times calculated here suggest that these PAHs may be protected from oxidation by tropospheric oxidants (i.e. OH, O3, and NO3). As a result, unoxidized PAHs may be transported to remote regions in the atmosphere and impact remote ecosystems such as the Arctic. |
626731f0ef2ade18543d643b | 0 | The unprecedented times we live in have brought about a call for chemical professionals to provide solutions for instructing laboratory techniques in our rapidly changing world. The onset of the novel coronavirus or SARS-CoV-2 in 2019 (COVID-19) changed the way chemistry is taught and brought about many hurdles to conquer. In the past two years we have seen a shift from traditional learning spaces to remote solutions allowing students to be instructed in their studies without risking contact during the pandemic. This is difficult as an important part of chemical education is physically taking part in chemical experimentation. Brilliant and dedicated educators developed and employed new procedure to provide that critical part of a well-rounded education in chemical science, for example, Silverberg published an article on an approach to laboratory instruction that allows great flexibility for students who may be in or out of quarantine. It is this goal that needs meet and, it needs to be kept in mind that not every laboratory or classroom has equal access to equipment nor does every student have equal access to a stable internet connection. This switch from in person instruction to remote across the world has made that fact front and center to programs with a diverse range of students in regards to their race, ethnicity, and socio-economic conditions. The laboratory procedure outlined here was built on a well understood model in epidemiology, the susceptible-infectious-recovered (SIR) model. Students only need access to a computer capable of running Microsoft Excel, Microsoft Word, and that has access to the world wide web. This procedure seeks to be very accessible with nearly any chemistry laboratory being able to conduct it with their students. |
626731f0ef2ade18543d643b | 1 | Modeling techniques have been in use in the field of epidemiology for centuries now. With one of the earliest examples found in its use to track the life expectancy in patients who were inoculated for smallpox. The foundation is then there to take the raw data derived from public health authorities and create models to understand and predict the direction of an ongoing health crisis. One such model is the SIR model that has its roots in a simple relationship between three possible states of being susceptible, infectious, and recovered. The transitions between these three states can be treated in the same manner of chemical kinetics where a concentration of the states exists that change in time depending on various factors. Two of those factors that will be focused heavily on in the procedure will be replacement number Rt/Re and the basic replacement number R0. The SIR model correlates the replacement number with the relative spreading speed of the virus. By adding additional parameters such as vaccination rates, SIR model can be extended to simulate the real-world data with a reasonable replacement number, especially at key points of public policy or social events. We have reported a laboratory procedure last year to analyze COVID-19 spreading in a state at the early stage with no vaccinations. This year, we added the vaccination data into the analysis and updated the SIR model to SIR-vaccinated (SIRV) model to analyze and predict the trends of COVID-spreading in the U.S. The skills taught and utilized in this procedure are useful in the physical chemistry classroom and many other places in the educational journey of an undergraduate student. The ability for students to see in real time the impact of public policy, and the statistical parameters they will derive from the model leaves a lasting impression about the usefulness and limitations of such techniques. |
626731f0ef2ade18543d643b | 2 | This laboratory procedure can be completed in two weeks (4 three-hour labs). The students are instructed to collect literature research, process data, write reports, and receive additional instruction outside of class time if needed. The SIR model and a modified version SIRV model which considers vaccination rates are introduced and practiced. Pre-laboratory instruction about the SIR model along with information about the history of mathematical modeling in the study of pandemics is necessary to explain the practical usefulness to the students. Microsoft Excel is the only computational program required to conduct this laboratory. While it could be conducted just the same with other programs that allow for data modeling, some students at the undergraduate level find themselves more familiar with Excel than other programs. The loan mortgage calculations are practiced in Excel at the beginning to warm up the students about model simulations in Excel, which has been explained in details in our previous publication. Students will also need to source and download data of COVID case numbers and vaccination rates of the region or regions they are designated to. This could be a state in the United States, it could be just the city of New York, or it could be an entire nation. The only real requirement is that the region being examined has reliable and easily accessed data on COVID case rates per day and vaccination rates. We instructed students to use a state in the United States as their region of interest for this procedure. Both COVID-19 cases and vaccination data in the U.S. are originated from U.S. Centers for Disease Control and Prevention (CDC). The COVID cases are summarized and downloaded from Github site of the New York Times. The vaccination data is downloaded from Our World in Data's State by State data on Covid-19. Both are accessed and downloaded around Jan. 27, 2022 with no further modifications except for sorting over states and time. The pretreatment of the data includes running smooth for a 7-day period from three days before to three days after a given day and filling the gaps of data by copying the available ones before them. Five days are added before the first date and assume one case during these five days. Data is checked and download again on April 14, 2022 after the analysis has been done to check the predictions. The SIR model follows a flow quite like the kinds of simple diagrams you may see drawn out in a modern chemistry textbook showing you the simplest model of chemical reaction kinetics (Figure ). In the model (Figure ), given a set of data with S, I, R in the unit number of people in a community, the two rate constants β/N and γ are the key parameters to predict the spreading trend of the disease. β refers to the average infection frequency of any given carrier and γ the frequency of removing an infected carrier. Both will have units of days -1 as the data provided by public and private sources on the spread of the virus will be most often in days. Students will be working with N refers to the total population of their assigned region. |
626731f0ef2ade18543d643b | 3 | Historically, the two rate constants in the SIR model are combined to a new parameter called the reproduction number. Three different forms of reproduction number are examined in this lab, namely the time-dependent reproduction number Rt, the time-dependent effective reproduction number Re, and the basic reproduction number, R0 (Equations 1-2). The time-dependent reproduction number, Rt, reflects the infectivity of the virus and the social interaction frequency of the society and is defined: |
626731f0ef2ade18543d643b | 4 | which is proportional to the ratio between the second-order reaction rate constant of infecting and the recovering rate. For a given variant of the virus, this value should be mainly proportional to the frequency of social interactions and virus transportation rate which are affected by various conditions such as public awareness and government regulations on social frequency, social distance, air circulation, washing hands, and wearing masks. |
626731f0ef2ade18543d643b | 5 | In Equation 2, both the growth and decrease part are first order dependent on the number of infectious people, thus, Re>1 reflects an exponential increase in the daily cases, Re=1 a steady state, and Re<1 sees an exponential decay of the number of infectious people, corresponding to the upside and the downside of the waves of infected cases each day. |
626731f0ef2ade18543d643b | 6 | The changing rate of the infectious population is a tug of war between the newly infected population and the recovered population each day. The latter being a rate constant of recovering, γ, times the total number of infectious people. The value γ is the one over the number of days before each infectious person is being removed from the infectious category, we used a fixed average value 0.2 Days -1 according to the literature for simplicity, even though it should slightly vary over different population groups and time. The SIRV model is nearly identical to the SIR model but includes the effect of vaccination on reducing the number of susceptible people (Figure ). The vaccination rate is time-dependent for COVID-19 as vaccines become more available and number of people who are willing to take the vaccine varies over time. Equations (6-10) show the relationship between the variables in the SIRV model. |
626731f0ef2ade18543d643b | 7 | All the data analysis in the lab is carried out using discrete numerical methods instead of the continuous analytical model described above. These equations are converted to discrete equations with the time interval 𝑑𝑡 ⇒ ∆𝑡 = 𝑡 2 -𝑡 1 ≡ 1 day, where the subscription 1 and 2 meaning two adjacent days, e.g. 2 is today and 1 is yesterday. Examples are provided in the following for the SIRV model. Equations for the SIR model use the same principle. |
626731f0ef2ade18543d643b | 8 | The students start the numerical calculation by guessing the first date of the first infectious person in a state, e.g., five days before the first date of report. The student can practice manual fitting by changing Rt values piecewise as explained previously by comparing the simulated new cases per day with the raw data. They can also directly calculate the real-time Rt and Re values by estimating β from Equation 15 by replacing the simulated new case with the smoothed raw data of daily case, newIraw: |
626731f0ef2ade18543d643b | 9 | Δt is held 1 day and γ is held 1/5 day -1 throughout the fitting, with it taking 5 to 7 days for an infected individual to become symptomatic according to the CDC and assuming people stay quarantine once they are aware of being infected. From Rt2, one can calculate β and Re |
626731f0ef2ade18543d643b | 10 | These values can be used to calculate S, I, Imm, and daily new cases (Equation 11-15) that day. Please see SI Excel sheets for the detailed calculations on both models. Each student was assigned a state and was instructed to use a source whose data can be verified by the CDC. Daily cases can be found by subtracting that day's positive case count by the day before giving roughly the change in S with the assumption that everyone counted in the N value will be susceptible and ignore infectious people who were not tested. A smoothed daily case count can then be generated with a moving average of the daily change in cases with 3 days before and after(Figure ). |
626731f0ef2ade18543d643b | 11 | Once students have collected data for their chosen state, they were instructed to find the base reproduction factor Ro by manually fit the SIR model to data from the 1 st two weeks. This was done by taking the calculated cases per day generated from the SIR model and fitting it to the cases per day from their smoothed data when minimizing the residual. |
626731f0ef2ade18543d643b | 12 | Typical R0 value of a U.S. state is found to be at ~3. Then Rt is manually adjusted in a short period of choice to minimize the residual or directly calculated from the daily cases in a state. An example is shown with the data in the state of Ohio (Figure ) assuming no vaccination is taken for the SIR model. Both the Rt and Re values are found to be oscillating around 1 along the waves of outbreak. Students were instructed to continue their calculated daily cases well past 1/25/2022 that are not available when they did the analysis using the average reproduction numbers for the last 5 days of available data to predict the trend for the following months. The models overlap well with the data obtained later giving the students some predictive power (Figure ). |
626731f0ef2ade18543d643b | 13 | The basic reproduction number of COVID is better to be extracted from the piecewise method of the first 2 weeks of data due to the relatively high sensitivity to the noise at the beginning when case numbers are small. 2 weeks of data give pretty consistent results from state to state with R0 ~ 3. The piecewise method can be applied to the full data set but is time-consuming. Thus, with proper smoothing, the real-time method is better to pull out the Rt values over time for the middle and later stages. |
626731f0ef2ade18543d643b | 14 | Rt is directly related to the second-order reaction rate constants which reflect the collision frequency of individuals and the energy barriers of the disease infection. Re is related to the quasi-first-order reaction rate constant that has been normalized to the concentration of the susceptible population in the community. The Re value as designed directly correlated with the growth and decline of the daily cases and its magnitude away from 1 reflects the exponential grow or decline rate like the interest rate or payback rate in a mortgage. However, it does not reflect the effect of social regulations as well as Rt, especially in the later stage of the spread. In the later stage, the concentration of the susceptible population has been significantly reduced due to immunity either from recovery or vaccination. Thus, Rt is a more valuable value than Re to trace the pandemic during these spreading stages. The resulting Rt from the SIRV model (Figure ) is consistent with our experience that travel and social activities are almost fully recovered around 2022 new year which brings the Rt value back to near R0 in all the randomly chosen states. Both SIR (Figure ) and SIRV model (Figure ) give the same Re values but SIR model significantly lower estimates the Rt values when vaccination is significant. |
626731f0ef2ade18543d643b | 15 | Breakthrough, natural immune, partial immune, variants of the virus such as delta and omicron, social density, transportations, cultural differences, variations of γ values, and population are all ignored during the analysis to simplify the model. However, they can be added and create more reaction pathways with modifications to the calculations, which is beyond the scope of this data analysis lab course. With information on what number of daily cases represent separate variants in the total number of daily new cases and the number of reinfections, The model can be expanded to account for the impact of reported higher rates of reinfection by the Omicron variant. The model provides the students some prediction power on the spread trend of COVID-19 (Fig. ). For example, if a mask can block 50% of the virus from spreading, then we can reduce the Rt value by 50% if everybody has put on the mask; reducing 50% of social activity further halves the Rt value as seen in the first several months of the pandemic (Figure -5 piecewise Rt). In general, the Rt value is unlikely to exceed the R0 value and must be larger or equal to zero for the same virus which set the limits on the upper and lower boundary of the spreading rate. The effect of the vaccination is readily visible in the simulation on the spread rate assuming no change in the social behaviors. There is no obvious evidence observed for the different spreading rates among the different variants such as delta and omicron, which will need data on the variants among the positive cases to be analyzed. |
626731f0ef2ade18543d643b | 16 | Extending the kinetic model from SIR to SIRV helps the students to understand the later stages of disease prevention and experience kinetic model selection. Comparing the Rt and Re values help the students to distinguish the second-order reaction and the quasi-first order reaction models for the same reaction mechanism, which echoes what they have learned in the class well. Comparison between the mortgage model, the COVID-19 kinetics, and the conventional chemical reaction kinetics allow the students to compare the units in different systems, especially on the rate constants, thus helping them to understand the purpose of such analysis and the prediction task better. The real-time Rt values correlate with the social events the best thus provides more prediction abilities, but their value is dependent on various conditions such as models, assumptions, data availability, and data accuracy. Overall, applying what they have learned in chemistry classes to a real-world problem motivates the students to learn and practice various skills. |
64f85f2f3fdae147fa9b65df | 0 | The ability of solid electrolytes to allow ion transport in the solid state makes them useful for a range of applications, including fuel cells and solid-state batteries . As a consequence, a considerable amount of research has focussed on understanding how the structure and chemical composition of particular solid electrolytes modulate their ionic conductivity . In addition to providing insight into the atomic-scale mechanisms of ionic conductivity within specific families of solid electrolytes, this body of research has also produced various "design principles"-general conceptual models that seek to explain, and predict, trends in ionic conductivity across different families of solid electrolytes . Many of these solid electrolyte design principles reflect how changes to the structure and composition of the host framework-the set of non-diffusive ions within a solid electrolyte-modulate the potential energy surface for the mobile ions, and hence influence overall ionic conductivity . |
64f85f2f3fdae147fa9b65df | 1 | One such design principle arises from the observation that solid electrolytes with some form of hostframework disorder often have significantly higher ionic conductivities than related compounds with well-ordered host frameworks . Framework-disordered solid electrolytes typically exhibit one of two classes of disorder: occupational disorder, where two or more distinct species occupy the same crystallographic positions ; or orientational disorder, where molecular or polyatomic subunits within the host framework have different disordered orientations . Orientational disorder can be static, where each polyatomic subunit has a fixed average orientation over experimentally relevant timescales , or dynamic, where the polyatomic subunits rotate and reorient . In some solid electrolytes, this reorientational dynamics of the host framework is thought to couple to the diffusive dynamics of the mobile ion species, giving rise to a so-called "paddlewheel" effect . |
64f85f2f3fdae147fa9b65df | 2 | While orientational disorder in solid electrolytes is usually discussed in the context of molecular or polyanion orientational degrees of freedom, materials that contain post-transition metals with "stereoactive" lone pairs; such as Sn or Bi; may exhibit electronic orientational disorder . These cations, when in an oxidation state two less than their formal maximum (e.g., Sn II and Bi III ), have a formal electron configuration with a filled s-orbital as their last valence shell. These s 2 states can mix with neighbouring-anion p states to form a bonding state and an antibonding state, and this antibonding state mixes with metal p states to form an asymmetric lone pair state, characterised by an eccentric (off-center with respect to the atomic nucleus)"stereoactive" charge density, that directs the cation coordination geometry and often results in distorted low-symmetry cation coordination environments . |
64f85f2f3fdae147fa9b65df | 3 | In many materials that contain stereoactive lone pairs, the local distortions due to these lone pairs are correlated over long length scales. These materials are long-range ordered and their structures can be determined using average crystallographic techniques such as Bragg diffraction. In other materials, however, the disortions due to stereoactive lone-pairs are uncorrelated . These materials are crystallographically disordered and average-structure crystallographic methods yield inaccurate high-symmetry structural models. Because the distortion around each cation depends on the relative orientation of the corresponding lone pair, this behaviour can be considered a form of orientational disorder, analogous to molecular or polyanionic orientational disorder as discussed above. These lone-pair effects may also be dynamic, showing fluctations or even rotations of the lone-pair charge density, mirroring the dynamic orientational disorder of "paddlewheel" materials . |
64f85f2f3fdae147fa9b65df | 4 | Here, we report a combined experimental and computational study of the fluorite-structured fluoride-ion conductor, cubic (c-)BaSnF 4 . We find that c-BaSnF 4 exhibits both site-occupational disorder, due to Ba/Sn cation mixing, and dynamic orientational disorder, due to Sn stereoactive lone pairs. The combination of these two forms of host-framework disorder within an ionconducting material makes c-BaSnF 4 a particularly interesting focus of study. We show that these two forms of disorder are coupled and that together they strongly influence the structure and dynamics of the mobile fluoride ions. The fluoride ion substructure is highly disordered, with 1/3 of fluoride ions occupying "interstitial" sites, due to lone-pair repulsion of fluoride ions in highly tin-coordinated sites. Fluoride ion dynamics are strongly dependent on the local cation environment and are cou- pled to dynamical reorientations of neighbouring Sn lone pairs. Our study provides new insight into the rich structural and dynamical behaviour of fluoride-ion conducting c-BaSnF 4 and how this arises from the unusual combination of coupled site-occupation and lone-pair-orientation host framework disorder. |
64f85f2f3fdae147fa9b65df | 5 | The fluorite structure is comprised of a face-centeredcubic cation lattice, with anions occupying all of the tetrahedral holes (Fig. ). The octahedral holes are vacant, and are usually considered as "interstitial" sites. An alternative structural description is obtained by considering the positions of cations within an anionic substructure (Fig. )): from this perspective, the anions define a simple-cubic lattice, and the cations occupy half the cubic holes, giving 8-fold MX 8 cation coordination. |
64f85f2f3fdae147fa9b65df | 6 | In conventional fluorites, such as CaF 2 , anion transport requires the presence of thermally-generated point defects-vacancies and interstitials-within the anionic substructure . Fluorites are typically anti-Frenkel disordered: some fraction of anions occupy octahedral interstitial sites, leaving an equal number of tetrahedral sites vacant . While additional interstitials and vacancies can be introduced via aliovalent doping, the intrinsic defect concentration depends on the ease with which anti-Frenkel pairs can form, which, in turn, approximately depends on the relative energies of ions occupying the tetrahedral and octahedral anion sites within the fcc cationic host framework. In simple fluorites, the anti-Frenkel-pair formation energy is usually high: for CaF 2 , anion Frenkel-pair formation energies have been calculated as 2.2 eV to 2.9 eV . As a consequence, at low-to-moderate temperatures, these materi-als have low fluoride-ion defect concentrations and corresponding low ionic conductivities . |
64f85f2f3fdae147fa9b65df | 7 | Significantly higher ionic conductivities are found in fluorite-structured materials that contain cations with stereoactive lone pairs, such as β-PbF 2 . This effect has been suggested to be a possible consequence of the high polarisability of the cation facilitating diffusion of the mobile anions , or that the negative charge of the Pb lone pairs might electrostatically destabilise adjacent fluoride ions in tetrahedral sites, thereby promoting the formation of anti-Frenkel pairs . Neutron diffraction data and AIMD simulations of β-PbF 2 , however, show no evidence for octahedral site occupation by fluoride ions, bringing into question the hypothesis that the presence of stereoactive lone pairs in fluorite-structured materials promotes Frenkel pair formation . |
64f85f2f3fdae147fa9b65df | 8 | x M ′′ 1-x F 2 . The highest ionic conductivity materials in this class are those with different valence cations, such as RbBiF 4 , where cation disorder induces high levels of anion disorder . A significant increase in ionic conductivity compared to analogous single-cation fluorites is also observed for mixed-cation systems where both cations have a formal 2+ oxidation state, such as Ba 1-x Ca 1+x F 2 : again, the presence of cation disorder causes considerable fluoride disorder where fluoride ions displaced significantly from their ideal crystallographic positions to form "pseudovacancies" . Some fluorite materials exhibit both stereoactive lone pairs and cation mixing . Of particular relevance to the present study is the work of Dénès et al. on Ca 1-x Sn x F 2 (x = 0.27) . For this system, X-ray diffraction (XRD) data give a cubic fluorite average structure, consistent with a solid solution of Ca and Sn distributed randomly over the cation positions, and 119 Sn Mössbauer data show a large quadrupole doublet, characteristic of a stereoactive tin lone pair. |
64f85f2f3fdae147fa9b65df | 9 | The presence of a stereoactive tin lone pair requires an asymmetric tin coordination environment, which is inconsistent with the structural model implied by the diffraction data, in which Ca and Sn both have cubic MF 8 coordination (cf. Fig. ). To reconcile these apparently contradictory data, Dénès and coworkers proposed a structural model wherein each tin is displaced towards one face of the enclosing [F8] cube to give squarepyramidal SnF 4 E coordination, where E denotes the stereoactive lone pair, with this lone pair oriented towards the more distant [F8] cube face. These tin lone pairs then are assumed to orient randomly along each possible ⟨100⟩ direction to give an average structure with cubic symmetry, consistent with the experimental diffraction data. |
64f85f2f3fdae147fa9b65df | 10 | The structural model proposed by Dénès et al. implies that Ca 1-x Sn x F 2 exhibits both occupational cation disorder and Sn lone-pair orientational disorder, and similar behaviour might be expected in other mixed-cation fluorites where one cation species has a stereoactive lone pair . Such mixed-cation fluorites are interesting to examine in the context of understanding how these distinct but coexisting forms of host-framework disorder together modulate the structure and dynamics of the mobile anion species. Here, we focus on the structure and fluoride-ion dynamics of c-BaSnF 4 , which we consider a representative member of this family of mixed-cation fluorites. c-BaSnF 4 is also of practical interest due to its shared composition with the more widely-studied layered tetragonal phase t-BaSnF 4 , which is considered to be a prospective solid electrolyte for fluoride-ion batteries . |
64f85f2f3fdae147fa9b65df | 11 | Synthesis Cubic BaSnF 4 was synthesized via a ballmilling process, using a planetary mill (Fritsch Pulverisette 6). Precursors (SnF 2 , Sigma-Aldrich, 99 %; BaF 2 , Sigma-Aldrich 99.99 %) were dried at 150 • C under vacuum for 3 h, and stored under Ar inert atmosphere. The desired amounts of precursors were weighed and sealed in Zirconia milling jars in an argon-filled glove box, with a powder to ball ratio of 1:13. The balls were 10 mm in diameter and made out of zirconia. The precursors were then milled at 400 rotations/min for 12 hours, divided into 24 cycles. Each cycle consisted of 15 minutes of milling and 15 minutes of pause, which prevented overheating. |
64f85f2f3fdae147fa9b65df | 12 | Impedance Spectroscopy Electrochemical Impedance spectroscopy was performed on c-BaSnF 4 powder pressed into a pellet. Gold was sputtered on both sides of the pellet to guarantee good contact. A BioLogic MTZ-35 impedance analyzer was used to collect data in a frequency range of 3.5 × 10 7 Hz to 1 Hz, under Ar atmosphere. The resulting data were fitted using the equivalent circuit model proposed in Ref. . |
64f85f2f3fdae147fa9b65df | 13 | 119 Sn Mössbauer Spectroscopy A lab-made constant acceleration Halder-type spectrometer operating in transmission geometry was used to carry out the Mössbauer analyses. The spectrometer was equipped with a radioactive source of 119m Sn (370 MBq) embedded in a CaSnO 3 matrix and maintained at room-temperature. Experiments were performed with 50 mg to 70 mg of sample ([Sn] = 5 mg cm -2 to 8 mg cm -2 ) at room temperature (∼ 293 K) and 77 K using a liquid nitrogen bath cryostat. The Mössbauer hyperfine parameters (δ isomer shift, ∆ quadrupole splitting, Γ signal linewidth, G 11 Goldanskii-Karyagin factor and relative areas) were refined using the WinNormos software [97]. Isomer-shift values are reported relative to CaSnO 3 at room temperature. 19 F Solid-State NMR Quantitative 19 F Magic Angle Spinning (MAS) NMR spectra were recorded on Bruker Avance III spectrometers operating at B 0 = 7.0 T ( 19 F Larmor frequency of 282.4 MHz), using a 1.3 mm CP-MAS probe head, and, for variable temperature experiments, using a 2.5 mm double resonance ( 1 H/ 19 F-X) CP-MAS probe and a Bruker Cooling Unit (BCU-II). The 19 F MAS spectra were recorded using a Hahn echo sequence with an interpulse delay equal to one rotor period. The 90 • pulse lengths were set to 1.25 µs (for SnF 2 and BaSnF 4 ) and 1.5 µs (BaF 2 ) and the recycle delays were set to 900 s (for SnF 2 ) and 300 s (for BaF 2 and BaSnF 4 ) using the 1.3 mm CP-MAS probe head. For the variable-temperature experiments, using the 2.5 mm CP-MAS probe head, the 90 • pulse length was set to 2 µs and the recycle delay was set to 10 s. The temperature inside the rotor was estimated from the chemical shift and spin-lattice relaxation time (T 1 ) of 79 Br in KBr powder . 19 F spectra are referenced to CFCl 3 and were fitted using the Dmfit software . |
64f85f2f3fdae147fa9b65df | 14 | Pair-Distribution Functions Pair-distribution function (PDF) measurements were performed at the 11-ID-B beamline at the Advanced Photon Source at Argonne National Laboratory. High-energy synchrotron XRD (λ = 0.2128 Å) 2D total scattering data was collected and integrated into one-dimensional diffraction data using FIT2D . The PDFgetX3 software was used to carry out Fourier transformation and correction of the PDFs . Refinements were performed using the software PDFgui . |
64f85f2f3fdae147fa9b65df | 15 | To model the equilibrium structure and dynamics of c-BaSnF 4 , we performed ab initio molecular dynamics (AIMD) using the Vienna ab initio simulation package (VASP) . We used the revised Perdew-Burke-Ernzerhof generalized gradient approximation PBEsol exchangecorrelation function . Interactions between core and valence electrons were described within the projectoraugmented-wave (PAW) method , with cores of [Kr] 4d 10 for Ba, [Kr] for Sn, and for F. We simulated a 6 × 6 × 6 supercell, starting from a cation-disordered fluorite structure with a special-quasi-random configuration of Ba and Sn over the Wyckoff 4a cation sites, that we generated using the icet package . This 6 × 6 × 6 special-quasi-random structure best approximates the Ba/Sn correlations for an infinite lattice with a fully-random arrangement of cations . |
64f85f2f3fdae147fa9b65df | 16 | Our molecular dynamics simulation used a plane-wave cutoff of 350 eV with only the gamma point used for kspace sampling, and without spin-polarisation. The simulation was performed at 600 K and used a time-step of 2 fs. Before our production run, we obtained the 600 K equilibrium volume by running a preliminary series of simulations with different cell volumes for 8 ps each, and fitting the Birch-Murnaghan equation to the resulting energy-volume dataset. The simulation was run in the N V T ensemble using a Nosé-Hoover thermostat. Thermal equilibration was performed by running a 2 ps N V E run with temperature rescaling every 50 steps. The production run was 159 ps in length. |
64f85f2f3fdae147fa9b65df | 17 | For structural analysis (calculation of radial distribution functions and cation-4a displacements) we extracted a set of "inherent" structures from our simulation trajectory by performing a conjugate-gradient geometry optimisation on configurations selected every 50 timesteps. Each inherent structure represents a local minimum on the corresponding 3N-dimensional potential energy surface. To calculate an example electron localisation function (ELF) we performed full geometry optimisations with a cutoff of 500 eV with a minimum k-point spacing of 0.25 Å-1 , with atomic positions and cell parameters relaxed until all atomic forces were less than 2 × 10 -2 eV Å-1 . To obtain tin lone pair orientations, we calculated the set of maximally-localised Wannier functions for structures sampled every 50 ps, using the Wannier90 code . The net dipole on each tin atom is calculated by associating each Wannier center with the closest ion, and then, for each tin, summing over all associated Wannier-center displacement vectors . We assume that tin polarisation is dominated by contributions from the lone pair states, and that our calculated polarisation vectors therefore characterise these lone pair orientations. |
64f85f2f3fdae147fa9b65df | 18 | Analysis of the simulation data was performed using the RevelsMD , site-analysis [119], ase [120], pymatgen , numpy , and scipy codes. The time-average fluorine density (Fig. ) was calculated using a linear combination of a conventional histogram with a triangular kernel and a force-extrapolated analogue, as described in Refs. . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.