id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
677ea1d281d2151a025203f0 | 29 | Details relating to assay components, test strip specifications, and operating procedures for particular experiments can be found in the Supplementary Information. The following is a typical protocol for 'capture-and-release' AmpliFold experiments: To a microplate well, assay components (antigen, FabHER conjugates, nanoparticles) were added and premixed (5 min). To each well, printed capture strips were added and allowed to wick (5 min). Strips were added into wells containing wash solution and allowed to wick (5 min). |
677ea1d281d2151a025203f0 | 30 | Test line intensities were analysed using ImageJ software using a procedure depicted in the Supplementary Information (Figure ). In summary, test strip images were converted into a greyscale image. A rectangular selection box of 500 pixels in height and 200 pixels in width was drawn to include the test line region of interest. Values relating to intensity per pixel -such that pixels were averaged (mean) across the width of the strips -were then generated and plotted. In mitigating against potential bias when determining the test line signal, test line intensity was assigned based on the maximum value within the region of interest. To avoid bias when determining the noise of the nitrocellulose membrane, the value of noise for strips was represented through the mean average of all pixels in the region of interest. The mean noise value was then subtracted from the test line intensity value to give a signal minus noise representation of test line signal per strip. |
677ecf656dde43c908c0b286 | 0 | Cysteine is unique among the twenty proteogenic amino acids due to the large atomic radius of its sulfur atom and the relative weakness of the corresponding S-H bond. This imbues it with remarkable nucleophilicity, facilitating spontaneous reactions even under mild conditions. The inherent nucleophilicity of a given cysteine is governed by its pK a , a value that implies the favorability of the ionization state of the thiol. Solvent-exposed cysteines have values near 8, while buried cysteines or those located in a unique protein microenvironment can range from 3 to 12. Given their variable pK a values and unique properties, cysteine residues play various functional roles in redox and nucleophilic catalysis, metal binding, environmental sensing, and structural formation. In recent years, interest has grown in targeting cysteine residues with covalent inhibitors. To overcome poor target selectivity and drug resistance, an electrophilic warhead moiety may be incorporated into a reversible submicromolar inhibitor to covalently bind a nucleophilic residue: this modification can dramatically increase therapeutic potency. Members of this class of reactive molecules are commonly referred to as targeted covalent inhibitors (TCIs) and predicting their affinity and reversibility is particularly desirable. A key step in a TCI binding and reaction landscape is the deprotonation of the cysteine thiol and the formation of the nucleophilic thiolate. Experimental exchange-rate studies have shown that the equilibrium between the protonated and deprotonated states of solvent-exposed cysteine side chains is fast (i.e., 10 • M -1 s -1 ) and that the protonation rate and pK a are well correlated. That is to say, the pK a of a particular cysteine provides the relevant information about the energy required to form the nucleophilic thiolate and, by extension, the propensity for covalent modification. |
677ecf656dde43c908c0b286 | 1 | Experimental methods for determining the pK a value of a cysteine can involve kinetic assays, spectrophotometric titrations, or NMR spectroscopy; however, in a purely computational in silico screen of potential covalent modifiers, the ability to rapidly and accurately probe the reactivity of a target cysteine under various conditions is highly desirable. Theoretical approaches motivated by the thermodynamic cycle given in Figure , present a compelling alternative to experiment and can often be seamlessly integrated alongside existing computational free energy workflows. |
677ecf656dde43c908c0b286 | 2 | Here, we consider a cysteine residue in the protein and a capped model peptide (i.e., ACE-Ala-Cys-Ala-NH 2 ) in both vacuum and water. We have the reference pK • a = 8.55 and as such can neglect the free energy of moving a (de)protonated residue from vacuum to water (Figure , left cycle). To determine the pK a we then need only consider the free energy difference of a deprotonation event in water and in the protein (Figure , right cycle). |
677ecf656dde43c908c0b286 | 3 | Figure : Complete pK a thermodynamic cycle. The horizontal arrows mark the transfer of a titratable residue between different environments: vacuum (left), water (middle), protein (right). The vertical arrows denote the free energy difference between the deprotonated and protonated form in a corresponding environment. For cysteine, the free energy associated with proton transfer from vacuum to water is known. By using this reference pK • a we need only consider the rightmost cycle: the free energy difference of a deprotonation event in water and in protein. |
677ecf656dde43c908c0b286 | 4 | Recently we demonstrated that atomistic molecular dynamics simulations paired with nonequilibrium alchemical free energy calculations are capable of accurately resolving the pK a values of aspartate, glutamate, and lysine. Here, we assess the ability of our pmx-based, nonequilibrium switching (NES) approach to calculate the pK a values of 40 cysteines and 22 histidines across a range of wildtype and mutant proteins. We compare our results to two conventional predictor methods and the previously reported performance of three MD-based approaches, chosen because it more accurately reproduces the relevant bulk properties of water compared to conventional 3-point models (i.e., TIP3P). The result of these classical simulations was a final cubic box size of L = 15 Å, containing 109 water molecules and a methylthiolate molecule. |
677ecf656dde43c908c0b286 | 5 | Classical MD simulations were performed in three types of boxes: one identical to that used in AIMD simulations (i.e., L = 15 Å) and two larger boxes: L = 30 Å and L = 60 Å (Figure ). We found no significant differences in the solvation structure when comparing the 15 Å and 30 Å boxes (Figure ). Solvation free energy calculations of the charged methythiolate molecule were performed in both a 30 Å and 60 Å box. Importantly, we found no change in the calculated free energy between the two box sizes (Figure ) which suggests the 30 Å box is sufficiently large to minimize the non-neutral simulation cell artefact associated with the calculation. GROMACS 2023 was used to run all simulations. Simulations were carried out in the NVT ensemble with a constant temperature of 298 K, maintained using a Nose-Hoover thermostat with 3 ps coupling time. In the 30 Å and 60 Å boxes, long-range electrostatic interactions were calculated using the Particle-mesh Ewald method with a real-space cut-off of 1.2 nm and grid spacing of 0.12 nm with CHARMM and a real-space cut-off of 1.0 nm and grid spacing of 0.125 nm with Amber. For CHARMM the Lennard-Jones interactions were force-switched off between 1.0 and 1.2 nm, while for Amber, a cut-off at 1.0 nm was used and a dispersion correction was applied to the energy and pressure. |
677ecf656dde43c908c0b286 | 6 | In the 15 Å box, electrostatic interaction cut-offs were 0.7 nm with a 0.12 nm spacing with CHARMM and 0.7 nm with a 0.125 nm spacing with Amber. For CHARMM the Lennard-Jones interactions were force switched off between 0.5 and 0.7 nm, while for Amber they were cut off at 0.7 nm and a dispersion correction was applied to the energy and pressure. We used the CHARMM TIP3P water model with the non-zero Lennard-Jones parameters on hydrogen atoms (i.e., mTIP3P) and plain TIP3P for the CHARMM and Amber systems, respectively. In the case of CHARMM we also assessed the impact of using the OPC water model on the solvation structure and measured free energies. |
677ecf656dde43c908c0b286 | 7 | Production simulations were 50 ns and the first 10 ns of simulation were discarded as equilibration. From the remaining 40 ns: 1) radial distribution functions were computed; and 2) 200 non-equilibrium transitions of 20 ps were generated. Free energies were computed as described in the final paragraph of the following section. |
677ecf656dde43c908c0b286 | 8 | pmx 33 was used for system setup, hybrid topology generation, and analysis. Initial protein structures were taken from the PDB database and mutations were introduced using pdbfixer. In total we consider 12 proteins and 40 cysteine residues and 22 histidines (Table ). A double system in a single box setup was used, with a 3 nm distance between the protein and peptide (ACE-Ala-X-Ala-NH 2 ); this ensured a neutral box at every step of the alchemical transformation. To ensure that the protein and peptide did not interact, a single Cα in each molecule was positionally restrained. We used the CHARMM36m (with mTIP3P ) and Amber14SB (with TIP3P 32 ) force fields, both having previously performed well for simulations involving non-equilibrium alchemical calculations. |
677ecf656dde43c908c0b286 | 9 | GROMACS 2023 was used to run all simulations. For all systems, an initial minimization was performed using the steepest descent algorithm. A constant temperature corresponding to the reference experimental setup was maintained implicitly using the leap-frog stochastic dynamics integrator 36,37 with a friction constant of γ = 0.5 ps -1 . The pressure was maintained at 1 bar using the Parrinello-Rahman barostat with a coupling time constant of 5 ps. The integration time step was set to 2 fs. Long-range electrostatic interactions were calculated using the Particle-mesh Ewald method with a real-space cut-off of 1.2 nm and grid spacing of 0.12 nm. Lennard-Jones interactions were force-switched off between 1.0 and 1.2 nm. Bonds to hydrogen atoms were constrained using the Parallel LINear Constraint Solver. Production simulations were 50 ns in length and run in quadruplicate. The first 10 ns of simulation was discarded as equilibration and from the remaining 40 ns, 400 non-equilibrium transitions of 200 ps were generated. Work values from the forward and backward transitions were collected using thermodynamic integration and these were used to estimate the corresponding free energy with Bennett's acceptance ratio as a maximum likelihood estimator relying on the Crooks Fluctuation Theorem. Bootstrapping was used to estimate the uncertainties of the free energy estimates, and these were propagated when calculating ∆∆G values. |
677ecf656dde43c908c0b286 | 10 | Based on popularity and previously documented cysteine pK a prediction performance we considered only two methods: PropK a (v3.4) and PypK a (v2.9.4). PropK a is an empirical predictor where the ∆G contributions are described by charge-charge, desolvation, and hydrogen-bonding interactions. Default settings were used when performing the calculation. PypK a uses Monte Carlo simulations to probe the various side chain states and employs DelPhi to resolve the PBE. Default settings were used, except for the salt concentration, which was set according to the experimental setup. |
677ecf656dde43c908c0b286 | 11 | Recently, Molecular Operating Environment (MOE) was used to calculate the pK a values of a large cysteine residue dataset. We compare NES with this method on the overlapping 20 residue data set. and AUEs (lower left triangle) between ∆pK a estimates were calculated for each method over the entire data set. Comparison with experiment means that the bottom row and rightmost column correspond to the overall performance. |
677ecf656dde43c908c0b286 | 12 | Figure summarizes the main findings: our NES approach performs comparably to in silico predictors, with CHARMM36m yielding an average unsigned error (AUE) of 2.92±0.35 pK as compared to 3.09 ± 0.29 pK and 2.71 ± 0.29 pK for PropK a and PypK a , respectively. This performance was also reflected in the Pearson correlation, which was 0.24 ±0.09 with CHARMM36m compared to 0.31 ± 0.14 and 0.22 ± 0.12 with PropK a and PypK a , respectively. On the 20 residue subset evaluated by MOE, MOE exhibited an improved accuracy of 1.79 ± 0.24 pK as compared to to accuracy, no method significantly exceeds a null predictor, which assumes ∆pK a = 0. |
677ecf656dde43c908c0b286 | 13 | Previous work has illustrated that an accurate determination of the pK a may require accounting for residue coupling. With respect to cysteine residues, often found at enzyme active sites, the relevance of coupling is expected to become even more pronounced. Elsewhere, we introduced a coupling formalism that improved pK a prediction accuracy; here, we apply this approach to twelve cysteine residues found near other titratable groups. Consistent with previous work, the pK a values of coupled residues were predicted with lower accuracy, when the coupling was not explicitly accounted for. Accounting for coupling, however, could in part remedy this (Figure ). |
677ecf656dde43c908c0b286 | 14 | The observed improvement was less pronounced for Amber14SB (AUE without coupling: Previous work has assessed the ability of different MD-based approaches to predict cysteine pK a values; we can compare our performance on the overlapping datasets. For 18 cysteine residues, Awoonor-Williams and Rowley, found a thermodynamic integration, replica-exchange scheme with the CHARMM36 force field gave an AUE of 1.67±0.40 pK (compared to 1.64±0.40 pK with CHARMM36m and NES). More recently, Awoonor-Williams and co-workers found that on 25 residues, a Monte Carlo, constant-pH approach paired with CHARMM36m gave an AUE of 2.42 ± 0.36 pK (compared to 1.70 ± 0.34 pK with CHARMM36m and NES). |
677ecf656dde43c908c0b286 | 15 | In the Amber family of force fields, both the thiol and thiolate sulfur atoms share the same atom type, while the partial charge assignments between the two residues differ. Thiolate sulfur has a more diffuse electron density and a larger ionic radius, characteristics that will be reflected in the Lennard-Jones parameters, particularly the σ-value. |
677ecf656dde43c908c0b286 | 16 | Simulations of methylthiolate using both classical molecular dynamics with Amber14SB and CHARMM36m, as well as ab initio molecular dynamics, revealed substantially different hydration structures (Figure ). Specifically, Amber14SB methylthiolate exhibited a radial distribution function (RDF) peak backshift of 0.5 nm compared to AIMD, suggesting a potentially erroneous hydration structure (Figure ). This observation is consistent with previous work employing different AIMD schemes and simulation setups. Given the significant discrepancy in prediction performance between CHARMM36m and Am-10 ber14SB, as well as the notable differences in RDFs, we rescaled the Lennard-Jones σ value to improve agreement with the AIMD RDF and potentially improve pK a prediction performance. |
677ecf656dde43c908c0b286 | 17 | Exploring the matrix of rescaled σ-and ϵ-values within the interval [0.5, 1.5] with 0.1 spacing, we found that a σ-value of 1.1 well reproduced the oxygen-sulfur RDF from the AIMD trajectories (Figure ). Expanding the grid to include more values yielded the same conclusion. Because adjusting the ϵ-value led to only marginal enhancements (Figure ), we refrained from unnecessarily fitting both parameters. |
677ecf656dde43c908c0b286 | 18 | We performed a similar analysis to determine an optimal value for predicting the solvation free energy of methylthiolate. Consistent with the RDF analysis, we observed that modifications to σ yielded more significant improvements than changes to ϵ (Figure ); however, achieving an accurate solvation-free energy required a σ-scaling of 1.3 (Figure ). |
677ecf656dde43c908c0b286 | 19 | With the primary goal of improving pK a prediction performance we probed the pK a values of wild type DsbA and seven mutants. Because changes in ϵ had a limited effect on solvent structure or solvation free energy, we decided to only scan σ-values on the interval [0.8, 1.5]. We observed a sigmoidal improvement in accuracy that was saturated for σ = 1.3 with an AUE of 2.88 ± 0.24 pK (Figure ) and a Pearson correlation of 0.31 ± 0.60. This accuracy was significantly increased from unscaled Amber14SB which had an AUE of 5.21 ± 0.44 pK and a correlation of 0.00 ± 0.51 on the DsbA test set. |
677ecf656dde43c908c0b286 | 20 | Using Amber14SB-1.3σ on the full dataset gave an AUE of 2.88 ± 0.35 pK and a Pearson correlation of 0.20 ± 0.11: markedly improved from the performance of plain Amber14SB, but still worse than the accuracy previously reported for aspartate, glutamate, and lysine. Using the σ-value that maximized agreement with the ab initio determined solvation structure (i.e., σ = 1.1), yielded an AUE of 3.28 ± 0.42 pK and a Pearson correlation of 0.35 ± 0.02. This shift is consistent with previous work that found a 0.5 pK improvement using a similar scaling factor for Amber99SB. We performed an identical analysis for CHARMM36m, which indicated that although a larger σ value (i.e., σ ≈ 1.15) was required to achieve an accurate experimental solvation free energy (Figure ), the default LJ parameters could effectively reproduce the RDF data (Figure ) and maximize pK a prediction accuracy on the DsbA test set (Figure ); this observation led us to leave the σ parameter untouched. Using the OPC water model (rather than mTIP3P) resulted in an identical position of the first solvation shell (Figure ) and suggested the same scaling factor was required to reproduce the experimental solvation free energy (Figure ). |
677ecf656dde43c908c0b286 | 21 | In the course of our coupling analysis, we found that histidine pK a values are predicted significantly higher with Amber14SB than with CHARMM36m. We previously observed lower accuracy for the prediction of lysine pK a s with the Amber14SB forcefield: this was traced to the partial charge difference of the backbone between the charged and uncharged lysine species. We hypothesized that this may also play a role for histidine, as here too the backbone partial charges differ between the doubly (denoted HIP) and singly protonated histidine residues (denoted HID and HIE). To further investigate, we computed 22 histidine pK a values that were taken from a full data set previously probed using equilibrium free energy calculations. These calculations were performed using both plain Amber14SB and a modified version, here called Amber14SB-H, where the partial charges of the protonated histidine backbone are those previously reported by Best et al. To account for the fact that two neutral histidine tautomers can exist with the proton present on Nδ (HID) or Nϵ (HIE), we perform two sets of free energy calculations: HIP → HID and HIP → HIE, which yield two relative free energies of deprotonation: ∆∆G δ and ∆∆G ϵ . Taking their difference gives the relative free energy of tautomer interconversion: free energy of deprotonation (Figure ): |
677ecf656dde43c908c0b286 | 22 | Compared to plain Amber14SB, using Amber14SB-H significantly reduced the AUE from 2.06± 0.33 pK to 0.64 ± 0.11 pK and increased the Pearson correlation from 0.42 ± 0.19 to 0.88 ± 0.05 (Figure ). Unlike previously observed for aspartate, glutamate, and lysine, we found a consensus estimate for CHARMM36m and Amber14SB-H resulted in an predictor that exceeded the performance of either method alone (i.e., AUE: 0.24±0.04 pK, Pearson correlation: 0.98±0.01). |
677ecf656dde43c908c0b286 | 23 | This level of accuracy exceeded that achieved using FEP+ (i.e., 0.39 pK) on the same 22 pK a dataset (Figure ). Our results suggest that NES can resolve histidine pK a values as accurately as FEP+ and further supports our previous suggestion that free energy calculations, in particular pK a calculations, with Amber14SB should employ the more recent, Best et al. partial charges. We note that this partial charge suggestion may also apply to Amber19SB, which utilizes the same backbone charges as Amber14SB. |
677ecf656dde43c908c0b286 | 24 | Traditional MM forcefields do not explicitly account for electronic polarizability; however, charge scaling has been employed to mimic this phenomenon. Sulfur is significantly more polarizable relative to oxygen and nitrogen, and we hypothesized that the markedly reduced pK a prediction performance for cysteine, as compared to aspartate and lysine, may be due to erroneously modelled electrostatic interactions between cysteine thiolate and its protein-residue neighbourhood. |
677ecf656dde43c908c0b286 | 25 | We rescaled all full unit charges (i.e., charged side chains and ions) (Figure ) on the interval [0.60, 1.00] with a 0.05 increment in CHARMM36m and recomputed the pK a values of the DsbA test set. CHARMM36m was chosen because of its higher accuracy in predicting cysteine pK a and because recent charge scaling efforts have successfully employed this force field. We observed the accuracy saturated to an AUE of 1.42 ± 0.14 pK for a scaling of 0. ). Accounting for residue coupling improved the accuracy of CHARMM36m-ECC even further, shifting the overall AUE from 1.78 ± 0.21 pK to 1.61 ± 0.21 pK (Figure ). Compared to plain CHARMM36m, CHARMM36m-ECC did not significantly improve the already strong histidine pK a prediction performance (i.e., 0.69±0.16 pK vs 0.71±0.16 pK). Charge-scaling did also not completely resolve the significant pK a underestimation observed for YopH tyrosine phosphatase (PDB: 1YPT) (Figure ). While we could exactly reproduce the relative effects of two nearby mutations (Figure ), the absolute pK a values were downshifted by ≈ 4 pK units (Figure ). |
677ecf656dde43c908c0b286 | 26 | Having scaled down the charge of cysteine, we have also increased the effective radius of the side chain atoms, in particular sulfur. Comparing the solvation structure of charge-scaled methythiolate to the AIMD simulations, we found a slight increase in the position of the first RDF peak (Figure ). Scaling the sulfur σ by 0.91 maximized overlap between with the MD and AIMD RDF curves (Figure ). |
677ecf656dde43c908c0b286 | 27 | As a cross check we also computed solvation free energies of charge-and σ-scaled methythiolate. Within the ECC framework, absolute free energies can't be compared directly with experiment but must be adjusted to account for the scaling (see SI methods). After adjusting the values, we found that similar to unscaled CHARMM36m, a slightly larger σ scaling is required to reproduce the experimental solvation free energy (Figure ). |
677ecf656dde43c908c0b286 | 28 | As an alternative to scaling all unit charges, we charge-scaled only the probed cysteine and balanced the missing negative charge by scaling the ions in solution. Computing the pK a values on the DsbA test set revealed a similar improvement trend as observed for global charge scaling, but nevertheless yielded a slightly poorer accuracy at a 0.7 scaling (Figure ). This difference is quite small and would seem to suggest that charge-scaled interactions of the probed cysteine itself and not that of other charged species is the major determinant of improved accuracy. In certain highly charged contexts (i.e., enzyme active sites), the accuracy improvement from charge-scaling other nearby residues is likely to play a larger role. |
677ecf656dde43c908c0b286 | 29 | Given the success with CHARMM36m, we also investigated charge-scaling with Amber14SB and Amber14SB-1.3σ. Amber presents difficulties because unlike CHARMM the side chain does not carry a full integer charge and cannot be simply scaled. Instead we linearly interpolate between the protonated and deprotonated cysteine to get a charge-scaled residue. Charge scaling Amber14SB improved prediction accuracy on the DsbA test set but failed to meaningfully saturate on the interval [1.00, 0.60], while charge-scaling Amber14sb-1.3σ improved prediction accuracy which was maximized for 0.80 and appeared to degrade for further scaling (Figure ). Probing Amber14sb-1.3σ with 0.80 charge scaling on the entire dataset improved the accuracy from an AUE of 2.88 ± 0.38 pK to 2.45 ± 0.27 pK; this improvement was smaller than that observed for CHARMM36m. Accounting for coupling also pushed the accuracy slightly higher to 2.37±0.29 pK. |
677ecf656dde43c908c0b286 | 30 | Figure summarizes the main findings: our NES approach with CHARMM36m-ECC and residue coupling accounted for, significantly exceeds the performance of plain CHARMM36m and two in silico predictors, with an overall average unsigned error (AUE) of 1.61 ± 0.21 pK, compared to 2.69 ± 0.29 pK and 3.08 ± 0.29 pK for PypK a and PropK a , respectively. This stronger performance was also reflected in the Pearson correlation, which was 0.60 ± 0.09 with CHARMM36m-ECC compared to 0.22 ± 0.12 and 0.31 ± 0.14 with PypK a and PropK a , respectively. Considering predictions within a certain tolerance, CHARMM36m-ECC correctly predicts 39 ± 8% of residues within 1 pK and 73±7% within 2 pK, compared to 22±6% and 44±8% with PypK a and 15±6% and 34 ± 7% with PropK a . If the objective is only to identify reactive cysteines (i.e., ∆pK a < 0) rather than assessing their exact pK a s we can consider this as a binary classification problem; performance can then be measured using the Matthews Correlation Coefficient (MCC) . In this regard, CHARMM36m-ECC provided the best performance (Figure ) with a MCC = 0.59, compared to MCC = 0.14 and MCC = 0.27 for PypKa and PropKa, respectively. This final comparison highlights that charge scaling can significantly improve cysteine pK a calculations, pushing the accuracy well above unscaled force fields and conventional predictor methods. |
677ecf656dde43c908c0b286 | 31 | Our results highlight three force field modifications that can improve pK a prediction accuracy for cysteine and histidine: 1) increasing the vdW of the deprotonated cysteine sulfur in Amber14SB; 2) altering the backbone partial charges of doubly protonated histidine in Amber14SB; and 3) charge scaling all unit charges in CHARMM36m and Amber14SB. Our investigation is not intended to provide definitive parameters for either force field or an absolute strategy for improving relative free energy calculations, particularly pK a prediction, instead, we aim to highlight potential avenues for further investigation and development. |
677ecf656dde43c908c0b286 | 32 | On the full dataset of 40 cysteines and 22 histidines, we found the strongest performing force field, CHARMM36m-ECC, to exhibit an AUE of 1.61 ± 0.21 pK for cysteine; this accuracy exceeds conventional predictors and a null model. While increasing the vdW of sulfur and charge-scaling both improved the performance of Amber14SB in predicting cysteine pK a values, the final accuracy of 2.36 ± 0.29 pK remains lower than that of Charmm36m, suggesting further reparameterization of the residue would be required. |
677ecf656dde43c908c0b286 | 33 | In the case of histidine, we found the accuracy could be significantly improved by taking a consensus of the Amber14SB-H and CHARMM36m charge-scaled force fields which yielded an AUE of 0.24±0.04 pK and a Pearson correlation of 0.98±0.01. Even standing alone, Amber14SB-H with charge-scaling attained an accuracy of 0.50 ± 0.10 pK and correlation of 0.85 ± 0.06, while CHARMM36m with charge-scaling gave an accuracy of 0.71 ± 0.16 pK and correlation of 0.56 ± 0.15. |
677ecf656dde43c908c0b286 | 34 | We note that while a NES approach yields increased accuracy, this comes at a significant computational cost. For an initial, large-scale pK a scan, a conventional in silico predictor (e.g., PypKa, PropKa, MOE, etc.) may be the more cost effective option, which could then be followed by a more accurate MD-based approach. With the aim of identifying reactive cysteines, an initial predictor should not disqualify any true reactive cysteines, even if it means introducing a larger fraction of false positives. The number of false positives introduced should be kept minimal, otherwise the computational benefit of the predictor would be small. While all methods attain relatively high precision (i.e., > 0.90), PypKa and PropKa only attained recalls near 0.30, compared to 0.87 for CHARMM36m-ECC, which reduces the usefullness of these methods as initial screening tools. |
677ecf656dde43c908c0b286 | 35 | In summary we find MD-approaches, including NES, can resolve the pK a values of cysteine and histidine residues with accuracy that exceeds conventional methods; however, this requires modification to the underlying MD force fields. The largest accuracy improvement we observed was for charge scaling the CHARMM36m force field, a result that will likely extend to other force fields and could remedy the poorer accuracy previously observed for predicting the effect of charge-changing mutations on protein thermostability and binding affinity. More work will help determine the consequences of charge-scaling; however, this work and the recent work of others seems to suggest that charge-scaling may be a general method to enhance the accuracy of non-polarizable MM force fields with only minimal and predictable costs. 1.0 0.9 0.9 0.9 0.9 0.9 0.8 0. (b) Exploring the RDF RMSD matrix of (σ, ϵ) pairs reveals that changes to σ alter the value significantly more than changes in ϵ. For CHARMM36m, the default value of σ = 1.00 well reproduces the position of the first RDF peak calculated from ab initio molecular dynamics, while for Amber14SB a value of σ ≈ 1.10 is required. The (σ, ϵ) pair that minimizes the RMSD between the MD and AIMD RDF curves is indicated with a box. on σ when using OPC (green) or mTIP3P (blue). The difference is larger than that observed for using a smaller box (e.g. 30 Å). Slightly different solvation free energies are calculated for the default sulfur thiolate: σ = 1.00; however, both models would suggest an identical sulfur thiolate: |
665788c7418a5379b093c3f1 | 0 | Multichromophoric systems enable energy and charge transfer processes which are ubiquitous in biological complexes and organic materials. On a fundamental level, assembling several (or many) chromophoric units together results in formation of molecular exciton states -the electronically excited states of the multichromophoric system. These states may be either delocalized over or localized on individual chromophores depending on the nature of individual constituents (homooligomer vs. heterooligomer) and interplay between structural distortion and electronic coupling. Moreover, upon excitation with light, the exciton states may undergo ultrafast "dynamical" localization during excited-state dynamics. A special class of multichromophoric systems are multiphotochromic systems featuring photoswitchable molecular units combined together in a single compound bearing a potential to expand switching functionality. For example, various multiazobenzenes tailored to applications ranging from energy storage to wavelength-selective control of molecular switching have been devised to date. Also, a few theoretical, nonadiabatic dynamics studies have been performed with focus on photoreaction mechanisms, quantum yields, and excited state lifetimes. However, the exciton dynamics in systems with several azobenzene chromophores have been paid due attention only recently. Specifically, Sangiogo Gil, Persico, and Granucci used surface hopping combined with an exciton model to simulate Frenkel exciton dynamics in bisazobenzenophane and an azobiphenyl monolayer. Moreover, we studied exciton localization and exciton dynamics in (non-covalent) H-type azobenzene tetramers using a supermolecule surface hopping approach at the semiempirical configuration interaction singles (CIS) level. In particular, our simulations revealed ultrafast, sub-100 fs dynamical exciton localization after ππ * excitation of the tetramers. While CIS is a computationally and conceptually attractive approach, it misses doubly (and higher) excited states, e.g., the singlet correlated triplet pair, which may play a role in nonadiabatic relaxation. Thus, a question arises how exciton dynamics and, in particular, the localization timescale are affected by inclusion of higher excitations. |
665788c7418a5379b093c3f1 | 1 | Here, we study the effect of double excitations on the exciton dynamics of a trisazoben- At the CIS level, we observe a picture which can be expected from the exciton model: Three low-lying states originating from the monomeric nπ * state and three higher lying states corresponding to the monomeric ππ * state. The nπ * states of the ring are virtually degenerate owing to small exciton coupling, whereas an exciton splitting of ∼0.37 eV is observed between the lowest ππ * state and the two (degenerate) upper ones. |
665788c7418a5379b093c3f1 | 2 | At the CISD level, apart from the singly-excited nπ * and ππ * states, which are somewhat blue-shifted in comparison to the CIS result, doubly-excited states appear in the spectrum. Specifically, we find three (nn)(π * π * ) states slightly above the nπ * states and six (nπ)(π * π * ) states just below the brightest ππ * states. |
665788c7418a5379b093c3f1 | 3 | A question arises how these states affect nonadiabatic dynamics of the multiazobenzene and, in particular, dynamics of exciton localization. To tackle this question, we performed surface hopping (SH) simulations with both CIS and CISD methods. In the case of CIS, six excited states and the ground state were included in the simulations. And in the case of CISD, 15 excited states and the ground state were accounted for (compare to Fig. ). The initial conditions (geometries and velocities) for the SH simulations were sampled from 20 ps long Langevin trajectories (at T = 300 K). The SH trajectories (100 for each system) were launched from the brightest ππ * state (see absorption spectra in Figs. S2-S4 and initial populations in Tab. S1) and propagated for 10 ps. For comparison, we simulated nonadiabatic dynamics of the monomer using either CIS or CISD and the active space of three orbitals (π, n, and π * corresponding, respectively, to HOMO-1, HOMO, and LUMO of the monomer, see Fig. ). See Methods section for further details. |
665788c7418a5379b093c3f1 | 4 | The electronic state populations are shown in Fig. . For the monomer, the S 1 state is the nπ * state, and the S 2 state is the ππ * state. In the case of trisazobenzenophane, we group the excited states as follows. For CIS, S 1 -S 3 build the nπ * manifold, and S 4 -S 6 the ππ * manifold. For CISD, S 1 -S 6 are assigned to the nπ * manifold, and S 7 -S 15 to the ππ * manifold. We note that the last group also includes the (nπ)(π * π * ) doubly excited states, but we term this group a "ππ * " manifold for simplicity (note that the (nπ)(π * π * ) states are located near the ππ * states, see Fig. ). The S 0 , nπ * , and ππ * populations were fitted using P ππ * = e -t/τ ππ * (1a) |
665788c7418a5379b093c3f1 | 5 | in eq. ( ) in contrast to our earlier definition of "fraction of transition density matrix" (FTDM). Further, to judge on exciton localization we analyze how the largest (or "highest", H) diagonal element F H changes with time. The ensemble-averaged F H as a function of time is presented in Fig. . Importantly, the CIS and CISD F H (t) curves are similar. Both show the rise of F H with a localization time constant (τ loc ) of ∼50 fs (see Tab. 1) which was obtained using an exponential fit of the form |
665788c7418a5379b093c3f1 | 6 | Specifically, we count how many trajectories have (i) S < 0.5, (ii) 0.5 < S < 1.0, (iii) 1.0 < S < 1.5, and (iv) S > 1.5. The corresponding fractions are shown in Fig. for CIS and CISD, respectively. In both cases the major fraction of trajectories corresponds to S > 1.5. Thus, on average, single excitations prevail during dynamics. At that, for CISD we also observe small but nonzero fractions with S < 1. |
665788c7418a5379b093c3f1 | 7 | Exciton localization in multichromophoric systems is often quantified with a participation number (or delocalization length): It is a scalar ranging from 1 (localization on one fragment) to 3 in our case (uniform delocalization over all three fragments). Eq. ( ) is, however, problematic in case S = 0 (or, equivalently, all F XY = 0), i.e., for purely doubly excited states. But, as we have seen (Fig. ), S = 0 is not the case for our system (for most of times, at least), and, in practice, even if F XY are small they are nonzero. Thus, PN can be calculated and it is shown in Fig. . Fitting the first picosecond of PN curves as |
665788c7418a5379b093c3f1 | 8 | we obtain τ loc ≈ 45 fs and B ≈ 1.51 for CIS, and τ loc ≈ 51 fs and B ≈ 1.25 for CISD. Thus, this analysis also yields a similar localization time (∼50 fs) for CIS and CISD, with slightly larger extent of localization for CISD (corresponding to a smaller B value). |
665788c7418a5379b093c3f1 | 9 | Finally, we computed quantum yields of the trans → cis isomerization. They are summarized in Tab. 1. At the CIS level, we find Φ ≈ 17 % and Φ ≈ 16 % for the monomer and the ring, respectively. At the CISD level, the quantum yields are higher: Φ ≈ 22 % and Φ ≈ 34 % for the monomer and the ring, respectively. Moreover, at the CISD level, the quantum yield for trisazobenzenophane is higher than that for the monomer, whereas similar quantum yields are obtained for both systems at the CIS level. Normally, if a trajectory is reactive, only one azobenzene unit undergoes isomerization (Fig. ). Some trajectories, however, exhibit upward hops from S 0 (after initial internal conversion to S 0 ) which may result in switching of the second unit (Fig. ). One trajectory with back hops (at the CISD level) demonstrates isomerization of all three azobenzene units (Fig. ). Moreover, at the CISD level, five trajectories are trapped in a state with two CNNC angles oscillating near 90-100 • (and the third one staying at ∼180 • ), see Fig. . |
665788c7418a5379b093c3f1 | 10 | Importantly, we found very similar exciton localization timescales with both CIS and CISD simulations, amounting to ∼50 fs for the studied multiazobenzene. In contrast, excited state lifetimes and isomerization quantum yields are affected much more strongly by inclusion of double excitations, with former being shorter and latter being higher at the CISD level in comparison to CIS. |
665788c7418a5379b093c3f1 | 11 | The electronic structures of studied azobenzenes were described with semiempirical configuration interaction (CI) [including either only single (S) or single and double (SD) excitations within an active orbital space] based on molecular orbitals (MO) obtained from a self-consistent field calculation with floating occupation (FO) numbers using Austin Model 1 (AM1) 36 reparameterized (r) for azobenzene. The method is thus abbreviated as The nonadiabatic dynamics were modeled using the trajectory surface hopping (SH) approach combined with the semiempirical configuration interaction method, 39,40 namely rAM1/FOMO-CI introduced above. The so-called added potential 37 (applied to each monomer) is used throughout, but not the state-specific corrections. The initial conditions (geometries and velocities) were sampled from rAM1/FOMO-CI (CIS for subsequent CIS SH, and CISD for subsequent CISD SH) ground-state Langevin trajectories propagated for 20 ps at T = 300 K with time step of 0.1 fs. 100 geometries were selected starting at 2 ps and sampling every 180 fs. The SH trajectories were propagated for 10 ps with a time step of 0.1 fs. The energy-based decoherence correction was used to remedy overcoherence of the original surface hopping algorithm. The time-dependent electronic wave function was propagated using the local diabatization scheme. For trisazobenzenophane, seven (S 0 -S 6 ) and sixteen (S 0 -S 15 ) electronic states were included in the CIS and CISD simulations, respectively. For the monomer, three electronic states (S 0 -S 2 ) were accounted for (for both |
646e0002e64f843f41b1f495 | 0 | With regard to these intriguing characteristics caused by the seven-membered ring, several groups have focused on helical molecules with seven-membered rings and revealed their intrinsic properties. Takasu et al. reported that a helical nanographene C composed of a contiguous 6-7-7-6 ring system has a large racemization barrier (E calc [B3LYP/6-31G(d,p)] = 29.2 kcal mol -1 ) (Figure ). They were able to isolate an enantiopure isomer and revealed that such heptalene-embedded helical nanographene C exhibits strong circular dichroism (CD), whereas [4]helicene with a 6-6-6-6 ring system easily undergoes racemization due to its much lower energy barrier (E calc = 4.0 kcal mol ). Thus, helical nanographene C with a heptalene skeleton can achieve configurational stablitity, which is quite rare. |
646e0002e64f843f41b1f495 | 1 | We envisaged that if a similar helical structure could be synthesized more easily, we could obtain advanced molecules, the physical properties of which could be finely tuned by introducing substituents and/or switched by applying external stimuli. Herein, we report the first example of nitrogen (N)-centered helical heptalene 1 (Figure ), where the azaheptalene skeleton can be constructed by a palladiumcatalyzed one-pot reaction. Thanks to the electron-donating ablity of the central nitrogen atom, electrophilic bromination of 1 proceeds easily to produce mono-and dibrominated azaheptalene derivatives 2 and 3, and these azaheptalenes all undergo one-electron oxidation to give the corresponding radical cation species. Furthermore, due to the contiguous 6-7-7-6 ring system, these Ndoped heptalenes gain configurational stability, so that not only photophysical properties but also chiroptical properties can be controlled by an electric potential. Pentabenzoazaheptalene 1 was synthesized by a one-pot reaction (y. 82%) via Buchwald-Hartwig amination of 2,2'-dibromobiphenyl and 9H-tribenzo[b,d,f]azepine followed by the palladium-catalyzed C-C bond formation (Scheme 1). Due to the electron-donating nitrogen atom, selective bromination of 1 proceeded cleanly. By treatment of 1 with one equivalent of NBS and three equivalents of Br 2 , monoand dibrominated derivatives 2 and 3 were obtained in 53% and 86% yields, respectively. During these reactions, overbrominated compound was not found because of the lower contribution of HOMO on the benzene ring at the center of the molecule compared to the contributions of those on both sides (Figure ). |
646e0002e64f843f41b1f495 | 2 | These azaheptalenes 1-3 were characterized by NMR spectroscopy and mass spectrometry and finally determined by single-crystal X-ray diffraction. X-ray analyses revealed that all azaheptalenes 1-3 adopt a highly distorted geometry with a dihedral angle [/° = 103.89 (11) for 1, 103.9(6) and 102.1 (7) for 2, and 100.3(5) for 3, respectively] between benzene rings on both sides (Figure ). These values were well-reproduced in optimized structures [/° = 104.0 for 1, 103.9 for 2, and 103.9 for 3, respectively] by density functional theory (DFT) calculations at the B3LYP/6-31G(d,p) level (Figure ). Since the dihedral angle of helical nanographene C is 57.51(3) °, these helical azaheptalenes with much larger angles are expected to have higher racemization barriers to enable the successful isolation of both enantiomers of these azaheptalenes (vide infra). To investigate the redox properties of these azaheptalenes 1-3, we conducted cyclic voltammetry in CH 2 Cl 2 . A reversible one-electron oxidation wave was observed for all derivatives (Figure ), whereas both parent triphenylamine and N-phenyl tribenzoazepine gave an irreversible oxidation wave (Figure ). These results indicate that double ethenylene bridges certainly contribute to the stability of radical cation species. Next, we generated radical cation salts of these azaheptalenes. In fact, by treatment of dibromoazaheptalene 3 with one equivalent of (2,4-Br 2 C 6 H 3 ) 3 N +• SbCl 6 -(Magic Green), 3 +• SbCl 6 -was isolated in 79% yield as a dark green powder (Figure ). The structure of radical cation 3 +• was determined by X-ray analysis, where the dihedral angle decreased from 100.3(5) ° to 92.95(10) ° upon oxidation (Figure ). The decrease in the dihedral angle can be accounted for by an effective delocalization of the spin and positive charge over the whole molecule. For non-and monobrominated derivatives 1 and 2, a generated radical cation was too reactive to be isolated. These results indicate that the introduction of bromine atoms into reactive sites is more effective for kinetic stabilization of radical cation species despite its electron-withdrawing properties (E 1/2 /V vs SCE: +1.39 for 1, +1.45 for 2, and +1.51 for 3). ). |
646e0002e64f843f41b1f495 | 3 | For neutral 3, the first band was assigned to the n-* transition, and the UV spectrum was almost identical regardless of the presence or absence of bromine atom. In the case of radical cation 3 +• , a NIR absorption band extending to 1600 nm was observed, which mainly arises from the SOMO-NHOMO transition. These results were supported by time-dependent (TD)-DFT calculations (Figure ). To elucidate the chiroptical properties of these helical azaheptalenes, we carried out chiral separation. |
646e0002e64f843f41b1f495 | 4 | All azaheptalenes 1-3 could be separated by chiral HPLC on a Daicel CHIRALPAK IG (Chart S1). The 1st fraction exhibited a minus value of optical rotation (see SI), and the absolute configuration was finally determined to be P-helicity by X-ray analysis of 3 [Flack parameter: = -0.060 (10)] (Figure ). The dihedral angle of (P)-3 is 99.78(3) °, which is almost the same as that of (rac)-3. CD spectra of (P)-and ), which are much larger than those of helicene derivatives with a similar number of fused rings. Furthermore, the chiroptical properties can be switched upon oxidation. Indeed, by treatment of (P)-and (M)-3 with one equivalent of Magic Green, (P)-and (M)-3 +• SbCl 6 -were isolated in 42% and 52% yields, respectively. These radical cations with P-and M-helicity also displayed mirrorimage CD spectra even in the NIR region. |
646e0002e64f843f41b1f495 | 5 | Thus, we performed DFT calculations and found that all azaheptalenes have much higher racemization barriers (E calc /kcal mol -1 = 50.5 for 1, 50.7, for 2, and 50.9 for 3) (Figures and) than those for helical nanographene with a 6-7-7-6 ring system (E calc /kcal mol -1 = 29.2) and N-centered [4]helicene with a 6-6-6-6 ring system (E calc /kcal mol -1 = 32.7). Thus, we concluded that the theoretically obtained barrier of ca. 50 kcal mol -1 is reasonable for these systems. In the case of radical cation 3 +• , DFT calculations [B3LYP/6-31G(d,p)] revealed a smaller racemization barrier (E calc /kcal mol -1 = 40.1 for 3 +• ) (Figure ), which can be accounted for by the decreased torsion upon oxidation. In conclusion, we designed helical azaheptalenes 1-3 with a contiguous 6-7-7-6 ring system, which could be efficiently synthesized by a palladium-catalyzed one-pot reaction. Regardless of these azaheptalenes composed of a minimal number of fused rings, its helicity exhibits sufficient configurational stablity to be separated using chiral HPLC. Moreover, thanks to the electron-donating nitorogen atom, such helical azaheptalene 3 undergoes one-electron oxidation to produce radical cation species with a chiroptical response. Notably, a larger g abs value was observed in the UV region for neutral azaheptalene 3 with P or M helicity, and its radical cation (P) or (M)-3 +• SbCl 6 -exhibited obvious CD signals extending to the NIR region. Therefore, helical azaheptalene, the chiroptical properties of which can be switched by an electric potential, may help to pave the way for the development of stimuli-responsive materials. |
67989d4e6dde43c908b0913c | 0 | In 1932, E. Zintl reported on the binary phase NaTl as the first example of a compound with formally negatively charged thallium, the so-called thallides. The formal electron transfer from the less electronegative metal sodium to the more electronegative thallium results in a diamondlike thallium substructure, which was interpreted in terms of the pseudo-element approach, introduced by W. Klemm. While located left to the so-called Zintl border between group 13 and group 14, introduced by F. Laves, text-book known NaTl is still referred to as the first Zintl phase and therefore set a milestone for the chemistry of Zintl compounds in general. NaTl can be prepared using classical high-temperature solid-state synthesis. At the same time, a stoichiometry range is given in the related binary phase diagram. In addition, E. Zintl himself also prepared NaTl in a low-temperature solution route by reducing thallium(I) iodide by sodium in liquid ammonia. While the heavier congeners of the alkali metals, potassium and cesium, can be prepared by high-temperature synthesis, the low-temperature route has not yet been reported for these heavier alkali metals. Interestingly, KTl and CsTl do not yield NaTl analogue compounds, but instead [Tl6] -octahedra are present in the unit cells of the crystal structures of the latter compounds. Applying Wade rules for these clusters, a lack of electrons can be identified compared to an octahedral closo cluster, which would afford an eightfold negative charge for the [Tl6] entity. In KTl and CsTl, (2n) skeletal electrons of [Tl6] 6-classify them as hypoelectronic. This affects the shape of the clusters, as it significantly deviates from an ideal octahedral shape. The observed compression was explained as a result of a Jahn-Teller distortion. 9, 10 The binaries KTl and CsTl crystallize in different orthorhombic space groups (KTl: Cmce; CsTl: Fddd), but both still include [Tl6] 6-clusters as anionic moiety. Recently, it was shown, that the combination of potassium and cesium in ternary Cs1-xKxTl approaches allows for the formation of pentagonal bipyramidal shaped [Tl7] 7-clusters in Cs3.45K3.55Tl7 and Cs7.29K5.71Tl13. In general, mixing alkali metals increases the variety of thallide compounds, which are not yet accessible in binary materials. Concerning the thallides in an alkali metal thallium ratio of 1:1, the absence of RbTl at ambient conditions is remarkable, especially as the remaining binaries LiTl, NaTl, KTl, and CsTl are longtime known. For RbTl, only a high-pressure phase with NaTl structure has been mentioned. The binary phase diagram of the Rb-Tl system contains RbTl3 (=Rb4Tl13), RbTl2 (=Rb15Tl27) and Rb4Tl6 (=Rb8Tl11) but does not include an equimolar compound. In addition, a further binary compound Rb49Tl109.67 was later reported, which also does not appear in the phase diagram. The phase diagram for the binary system Cs-Tl again does not show an equimolar compound, but Dong and Corbett reported on CsTl already in 1996. These examples nicely demonstrate, that the absence of a binary compound in a phase diagram does not contradict its existence. |
67989d4e6dde43c908b0913c | 1 | In this article, we report on attempts to prepare binary RbTl by high-temperature synthesis as well as low-temperature approaches in liquid ammonia. Additionally, RbTl is approximated in ternary approaches of mixed alkali metals A1-xRbxTl (A=K or Cs). The homogeneity range of K1-xRbxTl is discussed, and two new monoclinic compounds Cs1-xRbxTl (x=0.18, 0.42) are presented. |
67989d4e6dde43c908b0913c | 2 | The high-temperature solid-state syntheses were carried out in sealed tantalum ampoules from the elements in argon atmosphere. The sealed ampoules were placed in quartz glass tubes (QSIL GmbH, Ilmenau, Germany) and sealed in argon atmosphere. Different temperature programs were used after holding at 773.15 K (or 673.15 K) for 48 hours. in SI. |
67989d4e6dde43c908b0913c | 3 | Liquid ammonia was condensed on sodium and kept under argon, cooled by a dry ice/ ethanol bath. Anhydrous ammonia (5 mL) was condensed on the mixture of reactants ((1) 2 Rb + TlBF4, (2) 2 Rb + TlPF6, and (3) 2 Rb + TlBr) in three times baked-out reaction vessels and stored at 233.15 K. The approaches were prepared and stored at 233.15 K. After two to five days, ammonia was evaporated and the residue was characterized by X-ray powder diffraction (see SI, chapter 4). |
67989d4e6dde43c908b0913c | 4 | A small number of crystals was transferred into dried mineral oil. From that a suitable crystal was selected and mounted on a Rigaku SuperNova diffractometer (X-ray: Mo/Ag microfocus, AtlasS2 detector) or a Rigaku XtaLAB Synergy R DW system diffractometer (X-ray: Cu/Mo rotating anode, HyPix-Arc 150 detector) (Rigaku Polska sp. Z. o. o. UI, Wroclaw, Poland) using MiTeGen loops. All data were collected at 123 K. |
67989d4e6dde43c908b0913c | 5 | CrysAlisPro (version 171.43.105a) was used for data collection and data reduction. For the structure solution, ShelXT was used and the subsequent data refinement was carried out with ShelXL or olex2refine. Olex² was taken for visualization purposes and the software Diamond4 was chosen for the representation of the crystal structure. All atoms are depicted as ellipsoids with a 50% probability level. |
67989d4e6dde43c908b0913c | 6 | Due to the sensitivity to air and moisture, all samples were prepared in sealed capillaries (Ø0.3 mm, WJM-Glas-Müller GmbH, Berlin, Germany). The data collection was carried out on an STOE Stadi P diffractometer (STOE, Darmstadt, Germany) (Monochromatic Mo Kα1 radiation, λ=0.70926 Å) equipped with a Dectris Mythen 1 K detector. For visualization and indexation, the software WinXPOW and Jana2006 were used. |
67989d4e6dde43c908b0913c | 7 | Differential scanning calorimeter (DSC) measurement was performed using a TA Instruments Q200 analyzer. DSC analysis was carried out under a flow of nitrogen (sample purge flow: N2 40 mL/min.) The sample was sealed in a fume hood using TA Instruments Tzero hermetic aluminum pan and four heating/cooling cycles were performed from 293.15 to 593.15 K in 10 K/min rate. |
67989d4e6dde43c908b0913c | 8 | To explore the theoretical aspects of Cs1-xRbxTl and K1-xRbxTl we used the different methods namely: the projector-augmented wave method (PAW) , implemented in Vienna Ab initio simulation package (VASP) and the multiple scattering Korringa-Kohn-Rostoker (KKR) Green function method as implemented in the SPRKKR code . To see an effect of disorder the coherent potential approximation (CPA) implemented in SPRKKR code was used. In both codes, our calculation are based on Perdew, Burke, and Ernzerhof generalized gradient approximation (PBE-GGA). The calculations were performed in several successive steps. The geometry optimization and electronic structure calculation was done using VASP code. All the convergence parameters in code were checked carefully. |
67989d4e6dde43c908b0913c | 9 | In VASP code the geometries have been relaxed using the conjugate gradient method with forces estimated using the Hellman-Feynman theorem. For structure relaxation the energy cutoff of 700 eV and ISIF=3 were adopted. The self-consistencies of the ground state energies of K1-xRbxTl and Cs1-xRbxTl were obtained with energy cutoff of 320 eV. For k-point sampling, an automatic k-mesh were used for both compounds with 16 k-points in the irreducible Brillouin zone (IBZ), distributed according to a (3 × 3 × 6) and a (6 × 3 × 2) Monkhorst-Pack grid. For further density of state calculation of K1-xRbxTl one used dense k-mesh by increasing the k-points grid to (4 × 4 × 7). |
67989d4e6dde43c908b0913c | 10 | The energy and force convergence criteria were set at 10 -6 eV and 10 -3 eV/Å, respectively. Additionally, the phonon frequencies of binary and ternary rubidium doped KTl and CsTl were calculated using first principles phonon calculations with a finite displacement method implemented in phonopy code interfaced in VASP package. The accuracy of the phonon calculation is sensitive to various technical parameters such as supercell size, force convergence symmetry, energy convergence and atomic displacement. In present calculations, we used default values of 1.0 for force convergence symmetry, energy convergence criteria to 10 -6 eV and default atomic displacement of 0.01 Å. Detail on supercell size can be found in the SI (Table ). In phonon calculations, only the Gamma-point (Γ) was used for k-space sampling. |
67989d4e6dde43c908b0913c | 11 | First attempts for the synthesis of binary RbTl were carried out by the historical low-temperature experiments in anhydrous liquid ammonia. There, elemental rubidium and thallium(I) salts are reacted in a 2:1 (Rb: Tl(I)X (X=Br, BF4, PF6)) ratio at low temperature, which is a well-known preparation route for NaTl. The variation of the anion from Tl(I)X (X=Cl, Br, I) to weakly coordinating anions like [BF4] -or [PF6] -was also tested. All approaches resulted in elemental thallium and a rubidium salt, formed by Br -or the weakly coordinating anion (see PXRD in SI chapter 4). The low-temperature route therefore seems to be limited to NaTl. |
67989d4e6dde43c908b0913c | 12 | As alternative route, high-temperature solid-state reactions from the elements in different alkali metal to thallium ratios and subsequent quenching to 77 K in liquid nitrogen always yielded a mixture of Rb8Tl11 and Rb15Tl27 (Figure ). Depending on the excess of alkali metal, also elemental rubidium was involved. These results indicate, that binary RbTl is not accessible in experimental settings. |
67989d4e6dde43c908b0913c | 13 | Due to the fact, that the attempts to prepare binary RbTl did not succeed so far, we tried to approximate this compound by ternary approaches with potassium or cesium. Combinations of rubidium and potassium yielded in a homogeneity range of the the KTl structure type up to a rubidium proportion of 69% (see crystallographic data Table ). As side phase, K8-xRbxTl11 was always present, which also was reported for binary KTl. 9 |
67989d4e6dde43c908b0913c | 14 | The different mixtures of cesium and rubidium lead mainly to the formation of Cs8-xRbxTl11 phases and alkali metal. Contrary to the potassium-rubidium mixture, no solid solution in the CsTl structure type was found, but instead the new compounds Cs1-xRbxTl (x=0.18, 0.42), which crystallize both in the monoclinic space group C2/c could be isolated from the mixture. The compound with 42% rubidium forms in samples with a rubidium proportion of 33-50%. In contrast the one with 18% rubidium only forms with a high excess of both, cesium and rubidium (see Cmce in the KTl structure type and contain compressed [Tl6] 6-octahedra as thallium substructure. |
67989d4e6dde43c908b0913c | 15 | with rubidium being located on A1, A2, or A3 (see section 5.4). The new compounds (I) & (II) crystallize in the monoclinic space group C2/c. Although there is a group subgroup relation between C2/c and Cmce (KTl) and Fddd (CsTl), respectively, the new compounds are not related in symmetry to the binary materials. Although there is no crystallographic relation between the new monoclinic compounds and KTl or CsTl, the structural relation is given by the type of thallium cluster and the similarity in the first coordination sphere of the different alkali metal positions (see Figure and SI chapters 8 and 9). |
67989d4e6dde43c908b0913c | 16 | The asymmetric unit of (I) consists of each six symmetry-independent thallium positions and alkali metal positions, while in (II) there are nine symmetry-independent thallium and alkali metal positions, which are all located on a general Wyckoff position 8f. The compressed octahedra in the two new compounds are similar to those in CsTl 10 , in the solid solution K1-xRbxTl, Cs7.29K5.71Tl13 , A10Tl6O2(A=K, Rb) , Cs10Tl6TtO4 (Tt=Si, Ge) and Cs10Tl6SnO3 47 (see table distances and distortion degree in SI chapter 5). |
67989d4e6dde43c908b0913c | 17 | The [Tl6] 6-clusters of the compounds (I) & (II) arrange in AD hexagonal layers, which results in a distorted α-uranium packing. This is similar to KTl, whereas the clusters in CsTl pack in ADD'D'' layers, which is according to a distorted γ-plutonium packing (Figure ). 49 All nine in (II) or six symmetry-independent alkali metal positions in (I) are mixed occupied by cesium and rubidium. In the case of (I), the coordination numbers vary from 12, 14, 15 to 16 (d(A-A) ≤5.5 Å; d(A-Tl)≤4.51 Å) (see Figure ). They can be differentiated first of all in coordinations with three (A1, A2, A5, A6) and four [Tl6] 6-octahedra (A3, A4). The latter positions also show the highest cesium content (see Table ). In general, the coordination polyhedra are similar to those found in binary KTl and CsTl (see Figure , 5 and, SI chapter 6). The alkali metal positions A1, A5, and A6 are coordinated similarly to Cs3 (Wyckoff 16f) in CsTl. The ternary materials when mixing rubidium and cesium combine structural features of KTl and CsTl. As demonstrated in Figure , the coordination sphere of A2 is analogous to that of K2 (Wyckoff 8d) in KTl and Cs2 In (II) the coordination numbers of the first coordination sphere of the alkali metal positions vary from 13, 14 to 16 and especially the surroundings of A3 and A8 show differences from the ones in (I) or CsTl, which may be the reason for building a different structure type (see SI chapter 9). |
67989d4e6dde43c908b0913c | 18 | The site occupancy factors for A1 to A6 (in case of (I)) or A9 (in case of (II)) show a general trend, the higher the coordination number, the higher the cesium content. This results in the sequence for decreasing cesium content and increasing rubidium content in (I) A4>A3>A5>A2>A6>A1 (see Table ). |
67989d4e6dde43c908b0913c | 19 | To support the experimental findings, theoretical calculations have been applied in order to gain deeper insights in the structure stability of the reported compounds. The first issue addressed was the general question of the stability of binary and ternary compounds in the KTl and CsTl structure type when rubidium is involved. |
67989d4e6dde43c908b0913c | 20 | where E((K/Cs) 1-𝑥 Rb 𝑥 Tl) is the total ground state energy of (K/Cs increasing trend in the formation energy for approaching RbTl. This is due to the bigger size of rubidium as compared to potassium, whereas a higher rubidium concentration in CsTl shows a decreasing trend due to the smaller size of rubidium as compared to cesium. |
67989d4e6dde43c908b0913c | 21 | The aim of the formation energy calculations was to check the trend in both structure types with increasing rubidium concentration, however claiming the structure stability with respect to the formation energy is not reliable. Therefore, phonon frequencies were used to investigate the dynamical stability of the rubidium-doped KTl and CsTl structure as indicator. The presence of positive phonon frequencies in a material is an indicator of its dynamic stability and vice versa. In Furthermore, the phonon band structure of monoclinic structure (I) (C2/c) in Figure is an indication of stable nature and complements to the experiments (see section 4.2). In addition, calculations on the hypothetical Fddd orthorhombic unit cell setting of Cs0.562Rb0.438Tl also show no negative frequencies, despite not being observed in experiments. |
67989d4e6dde43c908b0913c | 22 | The density of states (DOS) plays a key role to understand the electronic structure of materials, to predict their macroscopic physical properties and their response to various external conditions such as temperature and pressure, etc. In this work, the DOS for three different concentrations of rubidium in the KTl structure type were calculated to observe the effect of the substitution as shown in Figure . The calculated DOS for three different concentrations show a good agreement and there is no significant difference observed between them. This clarifies that the electronic structure of K1-xRbxTl (x>0) is similar to KTl and the identifiable difference is due to the size effect or different ionic radii of rubidium. |
67989d4e6dde43c908b0913c | 23 | Figure shows the calculated total density of states (TDOS) using the supercell approach and CPA implemented in VASP and SPRKKR codes. The CPA enables the analysis of mixed occupancy of crystallographic positions without needing a fully ordered supercell model, effectively representing the statistically disordered crystal structure. In the energy range from -2.0 to 0.0 eV the increasing bandwidth from -1.8 eV to -2.0 eV and shifting of the low energy peak (around -5 eV) can therefore be attributed to the disorder effect in terms of the impact of either atomic randomness or partial order of different atoms at same site. In contrast, the electronic structure around the Fermi level is similar (more detailed description and atomic coordinates of the ordered model in chapter 10). |
67989d4e6dde43c908b0913c | 24 | To check the stability of different configurations of rubidium on the three crystallographically different alkali metal positions (see Figure ) according to K0.67Rb0.33Tl, the ground state energies for three situations were calculated (see Figure and SI chapter 11). This clearly shows that the favored position for rubidium is A1. This result is in excellent agreement with the observations of the s.o.f. from the experimental data (see Table ). |
67989d4e6dde43c908b0913c | 25 | In-depth experimental investigations did not allow for the preparation of binary RbTl. Mixed alkali metal approaches K1-xRbxTl suggest that the KTl structure type is favored up to x=0.69 while the CsTl type could not be realized in ternary approaches. In addition, in these cases the formation of lower symmetry, monoclinic compounds underlines this experimental finding. The DFT calculated formation energies were used to investigate the trend between the two phases with increasing rubidium concentration. At the same time, the phonon calculations were used as an indicator of dynamical stability in two structure types. The electronic structure calculation of rubidium in the KTl structure type suggests, that an increasing rubidium content does not affect the electronic structure around the Fermi level. The preference for rubidium on the A1 crystallographic site is observed in experiment and also independently obtained in the calculated ground state energies. The discrepancy of the theoretically stable but experimentally not observable binary RbTl might be explained by an effect of temperature. Future experiments will show, if the application of more sophisticated cooling and quenching techniques could facilitate higher concentrations of rubidium in A1-xRbxTl (A=K, Cs) compounds. |
65cc6f89e9ebbb4db9532a7f | 0 | presents an alternative avenue for functionalization, as depicted in Figure . The recent advancement in photochemical LSF of Dha-enriched peptides has facilitated targeted and chemoselective processes such as alkylation, fluoroalkylation, acylation, and to a more limited degree, arylation. Prior arylation approaches typically demanded aryl bromides as coupling agents, necessitating complex and laborious de novo synthesis to produce a variety of functionalized partners. However, recent studies by the groups of Ritter, Procter, and Alcarazo has highlighted the utility of arylsulfonium salts. These salts enable straightforward preparation routes leading to intricate aryl electrophiles, which are instrumental for both transition metal-catalyzed cross-coupling chemistry and photochemical applications. Building on this, vinyl-sulfonium salts have recently been employed for various polar transformations. Therefore, we were intrigued by the prospect of developing a photocatalytic LSF approach for Dha-containing peptides using the highly versatile and modular arylthianthrenium salts, as illustrated in Figure . This transformation was accomplished by the single electron transfer (SET) reduction of these salts, which generates an exceptionally reactive aryl radical. This radical readily adds to the α,β-unsaturated moiety within the Dha backbone. To guarantee gentle reaction conditions and minimize reaction times-key factors for ensuring broad functional group compatibility and scalability-we also devised a continuous-flow protocol. Notably, our method facilitates the efficient ligation of peptides and their conjugation with various drug scaffolds. |
65cc6f89e9ebbb4db9532a7f | 1 | Our initial efforts in achieving regioselective arylation of Dha-derivative 2 focused on the use of arylthiantrenium salt 1 (For detailed optimization, see Supporting Information). We employed DIPEA as the stoichiometric reductant and eosin Y as a photocatalyst, with the reaction driven by visible light irradiation at a wavelength of 456 nm (Table ). The reaction proceeded efficiently (72% isolated yield) in a solvent mixture of MeCN:HFIP (7:1) at room temperature over 4 hours (Table , Entry 1). Other protic or aprotic polar solvents failed to provide satisfactory conversions (Table , Entries 2 and 3). It became evident that the choice of solvent was critical; variations in the solvent system composition, containing different ratios between MeCN and HFIP, resulted in decreased yields (Table , Entry 4). A control experiment demonstrated the essential role of DIPEA (entry 5). Other potential reductants, including various amines and dihydropyridines, were also tested but failed to effectively produce the target product 3 (Table and). Furthermore, we explored a range of photoorganocatalysts and metal-based photocatalysts, which unfortunately led to less efficient processes (Table and). Gratifyingly, the synthesis of the unnatural amino acid 3 could be smoothly performed in a continuous-flow photoreactor in a significantly reduced reaction time (tR = 20 min) with comparable efficiency (Table , Entry 10). |
65cc6f89e9ebbb4db9532a7f | 2 | Building upon the established optimal conditions for the photocatalytic arylation of Dha 2, both in batch and flow (Table and), we sought to evaluate the versatility and scope of our method. Hereto, we applied various arylthianthrenium salts 4 to the established protocol, as delineated in Scheme 1. Scaling up our model reaction using arylthiantrenium salt 1 and Dha-derivative 2 via flow technology was successful, achieving a 71% isolated yield at a 1.0 mmol scale under our standard conditions. |
65cc6f89e9ebbb4db9532a7f | 3 | We then extended our approach to include alkyl-substituted arylthianthrenium salts 4, which served as competent reaction partners into our protocol, yielding the anticipated products 7-12 in good to excellent isolated yields. Notably, the synthesis of biaryl unnatural amino acids 13-15 was achieved in excellent yields. These amino acids bear functional groups amenable to subsequent synthetic manipulations, such as condensations and transition-metal catalyzed cross-couplings. Impressively, the aryl iodide bond, typically sensitive to reduction under photocatalytic conditions, proved to be compatible with our process. The method's adaptability was further underscored by its application to heteroatom-containing arylthianthrenium salts 4, leading to an array of phenylalanine derivatives (16-24). The yields were moderate to excellent, maintaining both high chemo-and regioselectivity. Additionally, our methodology demonstrated its robust functional group tolerance by facilitating access to a diverse set of heterocycle-containing amino acids. This collection of compounds included structures with dihydrobenzofuran, benzodioxole, chromanone, xanthone, pyridine, carbazole, indoline, and quinazoline-dione, showcasing the broad applicability and robustness of the synthetic strategy. |
65cc6f89e9ebbb4db9532a7f | 4 | In addition, our methodology enabled the synthesis of diverse amino acid derivatives for peptide synthesis, incorporating orthogonal protecting groups, requiring only minor adjustments to reaction conditions. We evaluated acid-, base-and hydrogenation-labile protecting groups, including tert -butyloxycarbonyl (Boc), fluorenylmethyloxycarbonyl (Fmoc), and benzyloxycarbonyl (Cbz) resulting in the successful preparation of the desired building blocks |
65cc6f89e9ebbb4db9532a7f | 5 | The process proved to be of high value, allowing for the seamless integration of a variety of drugs onto the Dha-backbone, yielding an array of novel amino acid/drug conjugates 38. The mild conditions of our photocatalytic hydroarylation were evidenced by the complete preservation of vulnerable heterocyclic structures, such as those found in pyriproxen, bifonazole, boscalid, and benzbromarone. The successful incorporation of drug scaffolds onto amino acid backbones led us to extend our strategy to the chemical ligation of peptides. We utilized an arylthianthrenium salt derivative of phenylalanine under our standard conditions, which afforded the targeted unnatural dipeptide 47 with excellent chemo-and regioselectivity. |
65cc6f89e9ebbb4db9532a7f | 6 | We then advanced to the late-stage functionalization of more complex peptide structures 48, applying our photocatalytic conditions detailed in Scheme 3. We were able to precisely and selectively functionalize a variety of di-,tri-, penta-and hexapeptides, decorating them with a multitude of functional groups derived from the aromatic core. Notably, peptides containing residues with sensitive functionalities like thioether, phenol, and indole groups were tolerated without any interference from photocatalytic single electron transfer (SET) processes or Scheme 3. LSF of Dha-containing peptides 44 leading to conjugation and ligation. For further experimental details see the Supporting Information. |
65cc6f89e9ebbb4db9532a7f | 7 | In addition, our protocol was not hindered by the position of the Dha residue in the peptide, as the functionalization at N-, C-terminus as well as inside the peptide sequence was accomplished. Furthermore, our methodology facilitated the rapid synthesis of a tripeptide 66, which showcased an atypical linkage, demonstrating the applicability of our ligation techniques. This process also proved to be highly capable at integrating drug scaffolds with peptides, further emphasizing the remarkable functional group compatibility of our system, even within the intricate context of peptide conjugation |
65cc6f89e9ebbb4db9532a7f | 8 | In conclusion, we have successfully developed a photocatalytic hydroarylation protocol that efficiently targets the Dha-backbone utilizing versatile arylthianthrenium salts. This novel strategy has demonstrated its versatility in synthesizing a diverse array of unnatural amino acids, with the potential for straightforward scale-up under a continuous-flow regime. The remarkable functional group tolerance of our methodology enables the preservation of delicate chemical structures during synthesis, broadening the scope of compatible coupling partners. |
65ef10259138d23161364879 | 0 | The United Nations has established a set of Sustainable Development Goals (SDGs) as part of their global agenda to address various social, economic, and environmental challenges. The SDG 6 is centered around universal access to safe and affordable drinking water by 2030, with a focus on the 2.1 billion people that lack access to safely managed water globally. One of the primary issues with poor water quality is microbial contamination which can cause potential acute health hazards, e.g., gastrointestinal infections, waterborne diseases, respiratory infections, etc. To supply safe and potable drinking water, centralized treatment facilities typically remove pathogens through both physical and chemical methods. While such facilities are common in developed countries, centralized systems are costly and require extended construction periods, especially when considering distribution systems. For example, a new water distribution system in lower income countries is estimated to cost 64-268 USD•person -1 for 500-2000 households. Another estimation for the implementation of a piped water supply for a small town in Ghana is in the range of 10-14 USD•person -1 •yr -1 (national minimum wage is approximately $689 USD•yr -1 ). It is notable that these estimated costs are only for the distribution system and do not include cost for water treatment. These barriers make potable piped water out of reach for many developing countries or emerging economies where the need to disinfect water is urgent. As an alternative solution to centralized systems, water can be disinfected at the household level or point-of-use (POU), providing a potential pathway for immediate safe drinking water for off-the-grid communities. Numerous POU disinfection technologies are commercially available, ranging from conventional technologies (e.g., boiling and POU chlorination) to new technologies (e.g., ultraviolet (UV) disinfection systems). For example, solar water disinfection (SODIS) can be a relatively simple intervention for disinfection when properly utilized. A year-long study in Cameroon highlighted that SODIS provided up to a 42.5% reduction in the risk for diarrheal diseases in households that properly treated their water, but only 45.8% of all households effectively adhered to the recommended practices of SODIS. Ceramic water filters are another POU technology that can be produced with local materials and provide dual mechanisms to remove bacteria, i.e., the porous physical ceramic matrix filtration and silver nanoparticle antimicrobial coating. A randomized controlled field trial in Bolivia demonstrated the effectiveness of ceramic filters in meeting World Health Organization (WHO) drinking-water standards. While results obtained proved valuable in achieving compliance when faced with turbid challenge waters, additional research is needed. One aspect that requires attention is the maintenance of ceramic filters, as an agent-based model has shown that neglecting it can hinder their long-term sustainability, despite their relative ease of use. Overall, sustained adoption of individual POU technologies can vary between communities due to contextual and end-user factors, but inadequate clarity on how decision makers and stakeholders can navigate the different POU technologies under different contexts can limit implementation and sustained adoption. Therefore, the sustainability of numerous POU technologies needs to be simultaneously assessed while considering context-specific factors enabling engineers, agencies, and researchers to make informed decisions and select the most suitable treatment technology for a specific community. |
65ef10259138d23161364879 | 1 | Toward this end, technoeconomic analysis (TEA) and life cycle assessment (LCA) can serve as valuable methods for evaluating trade-offs in terms of cost and environmental impacts when comparing different POU technologies. For instance, a study conducted on POU chlorination (Aquatabs), flocculant disinfection (Procter and Gamble Purifier of Water), and ceramic filters evaluated the cost effectiveness considering costs related to startup, management, and logistics. While POU chlorination was found to be the most cost-effective method, this study was limited to one year period, which may be relatively shorter than necessary in other contexts. A recent LCA of four UV-based systems, chlorination, and trucked water delivery found chlorination to have the lowest environmental impacts over various time and scale horizons. Leveraging both TEA and LCA can help identify trade-offs between cost and environmental impacts for POU technologies. These tools together have been used in a limited way to evaluate several conventional disinfection technologies (boiling, ceramic filters, bio-sand filters, and POU chlorination). Under a specific set of assumptions, boiling and chlorination had the highest environmental impacts, while boiling was the most expensive (0.053 USD•L -1 ) and chlorination was the least expensive (0.0005 USD•L -1 ). In general, accurately comparing the relative sustainability among different studies can be difficult to due to variations in assumption, leading to different outcomes for sustainability indicators. For example, shorter studies with technology lifespans of less than one year may not consider all materials and supplies that are used in the process throughout the technology's lifetime. Considering the inherent uncertainty associated with changes over the lifetime, location, and other factors while assessing the relative sustainability of POU technologies, can help to account for the fluctuating assumptions. |
65ef10259138d23161364879 | 2 | The goal of this study is to comparably assess the relative sustainability of several readily available POU disinfection technologies. Specifically, the objectives of this work are to (i) characterize the overall cost and environmental impacts while considering necessary disinfection efficacy of these technologies and (ii) elucidate drivers for sustainability to better inform appropriate adoption in specific contexts. The technologies assessed in this study include: POU chlorination, silver nanoparticle enabled ceramic water filter (AgNP CWF), UV with mercury lamp, and UV with light-emitting diode (LED). This study leverages the quantitative sustainable design (QSD) methodology for TEA, LCA, and disinfection efficacy assessment using an open-source Python packages QSDsan (QSD for sanitation and resource recovery systems). Uncertainty was incorporated in the assumptions inputted into the models, and sensitivity analysis (via Spearman's rank correlation coefficients) was completed to identify key drivers of sustainability. The impact of water quality was evaluated by updating assumptions considering two different water compositions (surface and groundwater), and a technology adoption period ranging from 1 to 15 years was assessed. A contextual analysis was also included to reflect the implications of location-specific parameters on technology deployment in ten different communities across the world. Findings of this study are expected to offer valuable insights for decision-makers, non-profit organizations, and future research endeavors focusing on sustainable approaches to safe drinking water through POU technologies. |
65ef10259138d23161364879 | 3 | To explore trade-offs among POU technologies, we leverage the QSD methodology and software QSDsan , where integrated TEA and LCA were performed with parameters covering design, materials, energy and capital requirements, and operation and maintenance requirements (Figure ). All essential decision variables and technological parameters were derived from a comprehensive range of sources (published research, manufacturers' specifications, and guideline reports). All the Python scripts are publicly available on GitHub with a README file for instructions, and an online (i.e., without local installation of Python) Python environment capable of running the scripts in Jupyter Notebook can be accessed in web browsers through the binder link on QSDsan GitHub repository. All input assumptions are included in the Supplemental Information (SI). A 5-year technology adoption period and an average household size of 4 people were used as the baseline for all the POU technologies. The number of people per household is aligned with the average number of households in most of the countries with lower access to basic drinking water. Two types of raw waters-surface water and groundwater-were modeled with their characteristic water quality parameters (described in detail below). To standardize the disinfection efficacy of the technologies minimum of 3 log reduction was evaluated for all systems. |
65ef10259138d23161364879 | 4 | The disinfection method for POU chlorination was designed based on the use of a solution of sodium hypochlorite (NaClO). This solution is used to disinfect drinking water in households with a relatively simple setup (Figure ). The specific NaClO product used here is marketed as WaterGuard, and each bottle contains 150 mL of the NaClO solution. The treated water volume was 20 L based on the assumed container capacity. This disinfection method is designed to be relatively simple to use, where the bottle cover of the WaterGuard bottle is used to dose the NaClO solution into 20 L of raw water. An expected one WaterGuard bottle cap is a measure of a single dose of NaClO solution while two is used for a double dose. The full materials and cost inventory data for POU chlorination system are accounted for (Table ). The code of this system was designed for three influent streams, i.e., the raw water, NaClO (chlorine stream), and polyethylene (WaterGuard bottles). To keep the goal of the minimum log reduction of bacteria, the dosing of NaClO was 1.88 mg•L -1 at low turbidity (≤ 10 NTU), and a double dose of 3.75 mg•L -1 was used at higher turbidity (> 10 NTU). Algorithms were developed to capture this impact of water quality on cost and environmental impacts. |
65ef10259138d23161364879 | 5 | The CWF is coated with AgNPs such that the ceramic matrix filters for a combination of physical (through filtering) and chemical (through AgNPs) disinfection. As shown in Figure , the setup has the ceramic coated with AgNP placed over a plastic bucket that holds the filtered and treated water. Materials used to make the setup (sawdust, clay, wood for the filters and polyethylene for the plastic container) were incorporated to account for their costs and environmental impacts. The unit has one influent stream of raw water and an effluent of treated water. Here, AgNP is the main consumable as recoating will be needed after every 0.5 to 2 years. Algorithms in this unit were developed to account for the length of time before recoating the filter with AgNPs was necessary based on the quality of the water type. |
65ef10259138d23161364879 | 6 | The lifetime for AgNPs in this unit depends on the water quality. Specifically, more frequent recoating is expected for higher turbidity and hardness because these constituents have been reported to remove more AgNPs. In this analysis, a turbidity >10 NTU and/or hardness >60 mg CaCO3⋅L -1 were used as the thresholds for more frequent recoating of AgNPs. |
65ef10259138d23161364879 | 7 | for bacteria inactivation at a wavelength between 200 to 280 nm. The system in this study has two UV lamps on opposite sides with water flowing through a quartz tube to maximize light transmittance to microbes (Figure ). The materials accounted for included lamps, aluminum, polyethylene, and polyvinyl chloride. The mercury UV lamps used in this work were expected to have a lifespan of approximately 2,000 hours; however, varying lifespan has been reported by manufacturers of other mercury lamps. The unit is modelled to have two mercury lamps that use 30 W of electricity each. It is designed based on a flow of 9.46 L•min -1 with a UV dose of 215 mJ⋅mc -2 . In this unit, we incorporated the impact of water quality through turbidity on UV light transmittance, UV dose, and detention time. These factors influence the energy requirements and potential cost and environmental impacts. The UV lamp was assumed to be on for double the time in higher turbidity (>10 NTU) to account for the increased retention time in water with less UV light transmittance. The extended residence time was also accounted for in the unit's electricity demand. |
65ef10259138d23161364879 | 8 | The last POU technology in this study is UV with LEDs as the source of disinfection. These lights generally are considered more environmentally sustainable as they do not contain mercury like the lamps for traditional UV systems. This unit also allows the UV dose to be adjusted and offers design flexibility as the UV LEDs can be arranged in different formats to optimize disinfection. The design capital materials included quartz, stainless steel, aluminum, and 30 UV LEDs. The unit was designed in a flow-through system so the water is surrounded by arrays of UV LEDs separated from by a quartz material that allows adequate transmittance of UV lights for disinfection. As shown in Figure , the plan view UV LEDs are set up with 15 LEDs on each side of the unit. The system is set up such that an array of UV LEDs require 23 W of electricity. UV LEDs used in this study are estimated to have a life span of approximately 10,000 hours. Other study and manufacturers have reported higher lifetime of up to 100,000 hours, although many of these are still in a developing stage. The unit is designed to incorporate the influence of water quality similar to the unit for UV with mercury lamp. Turbidity of > 10 NTU was assigned a double retention time factor, which was used for the accounting of electricity demand and the lifetime of the lamps. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.