id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
675f0698085116a1337f1a3a | 1 | In the early stages of the COVID-19 pandemic, Transmembrane Protease, Serine-2 (TMPRSS2) was identified as an important host protease enabling viral entry of SARS-CoV-2 . TMPRSS2 is responsible for cleaving the SARS-CoV-2 Spike (S) protein after it has docked to its primary receptor, angiotensin converting enzyme (ACE2; Fig. ) . TMPRSS2 S protein cleavage drives conformational changes in the S protein that induce membrane fusion between the virus and host . TMPRSS2 has also been implicated in the entry of other human respiratory viruses including MERS-CoV, SARS-CoV, and HKU1 as well other viruses including para-influenza virus and Sendai virus through similar entry mechanisms . This widespread role of TMPRSS2 as a respiratory virus entry factor makes TMPRSS2 an interesting target for the development of a broadly active antiviral agent against a diverse range of respiratory viruses . |
675f0698085116a1337f1a3a | 2 | TMPRSS2 is a member of the human type II transmembrane serine protease (TTSP) family consisting of 17 proteases that modulate a plethora of homeostatic functions including but not limited to iron regulation , blood pressure regulation , hearing development , and digestion . However, the dysregulated activity of these proteases in cancer cells promotes tumorigenesis, cancer aggressiveness, and metastatic functions . TMPRSS2's roles in homeostasis are not fully understood, but TMPRSS2 knockout mice grow without any detectable defects or impact on physiological function . Furthermore, TMPRSS2 knockout mice show reduced susceptibility to viral infection and transcriptional knockdown of TMPRSS2 decreases susceptibility to SARS-CoV-2 infection . |
675f0698085116a1337f1a3a | 3 | The clinically approved small molecule protease inhibitors camostat and nafamostat have been shown to inhibit TMPRSS2 activity in vitro and have undergone drug repurposing efforts for TMPRSS2-targeted SARS-CoV-2 antiviral therapy . Camostat and nafamostat share the same electrophilic 4-guanidino benzoate core that specifically binds to Subsite 1 (S1) of trypsin-like serine proteases and then reacts with the catalytic serine residue to form an inactive enzyme adduct, shown in Fig. for TMPRSS2. Camostat and nafamostat were developed to treat pancreatitis and to be used as an anticoagulant , respectively. |
675f0698085116a1337f1a3a | 4 | In this work, we targeted the structure of TMPRSS2 for large-scale library docking of small molecules, seeking new starting points for lead discovery. We explored two distinct, structure-guided strategies to inhibit TMPRSS2 protease activity with covalent and noncovalent small molecule inhibitors (Fig. ). By docking a library of 200 million on-demand, lead-like molecules against both a homology model and the crystal structure of TMPRSS2, an ester-based covalent scaffold and a noncovalent amine scaffold were selected for further biochemical characterization. We identified a readily crystallizable form of TMPRSS2 leading to two novel cocrystal structures and leveraged a suite of biochemical and biophysical assays to validate these inhibitor molecules and understand their potency in terms of noncovalent binding and covalent reactivity. The tools established here will aid future SAR campaigns against TMPRSS2 and other human TTSPs implicated in respiratory virus infections. |
675f0698085116a1337f1a3a | 5 | A homology model of TMPRSS2 was initially constructed, since at the onset of this study, the crystal structure of TMPRSS2 had not been published. The homology model was constructed from plasma kallikrein, coagulation factor XI, TMPRSS11E, matriptase, and Coagulation factor VII and exclusively spanned the TMPRSS2 serine protease domain (Fig. ; Extended Data Fig. ). The model was found to be highly similar to the experimentally determined crystal structure of TMPRSS2 following its acylation by nafamostat (Fig. ). The TMPRSS2 homology model was docked against the ZINC20 library 42 , which includes approximately 200 million monocationic compounds, primarily sourced from the make-ondemand Enamine REAL database (Extended Data Fig. ). Of the compounds docked within the TMPRSS2 active site, 121 were initially selected for recombinant TMPRSS2 inhibition screening and a single hit ('2805) was obtained (0.8% hit rate; Extended Data Fig. ) with a phenyl benzoate ester and a primary amino group. The nafamostat and camostat TMPRSS2 inhibition mechanism involves the formation of an inactive acyl-enzyme complex (Fig. ), and since '2805 also contains an ester group, we hypothesized that '2805 may form an inactive acyl-enzyme complex (Fig. ). Using biochemical TMPRSS2 peptidase inhibition assays, '2805 was found to have a TMPRSS2 half-maximal inhibitory concentration (IC50) of 120 nM (Fig. ). As shown in the '2805:TMPRSS2 docked pose, the primary amine in '2805 likely binds the TMPRSS2 S1, similar to the guanidinium found in nafamostat and camostat. The ester linkage in '2805 is reversed with respect to the orientation of the cationic groups found in nafamostat and camostat, however, the '2805 ester group is close in proximity to the TMPRSS2 S441 residue to enable nucleophilic attack and acyl-enzyme complex formation (S441*; Fig. ). |
675f0698085116a1337f1a3a | 6 | We decided to further explore this scaffold and similar ester-containing scaffolds by searching the Enamine REAL space for related compounds using the SmallWorld search engine . Since '2805 contained a primary amine in comparison to guanidinium in camostat and nafamostat, we decided to explore a larger range of terminal amines. An initial selection of amines was built into six scaffolds (Fig. ). Scaffold 1 analogs were designed by directly substituting the phenyl ring of the benzoate with a diverse array of functional groups. |
675f0698085116a1337f1a3a | 7 | To mimic the placement of the amines in nafamostat and camostat on the phenyl of the benzoate, the alphaamino ketone in '2805 was moved from the carboxyl phenyl to the benzoate phenyl to produce Scaffold 2. Scaffolds 3 -6 were designed by introducing methyl amine, ethyl amine, 3-oxyazetidine, and guanidino groups respectively to the benzoate. |
675f0698085116a1337f1a3a | 8 | Recombinant soluble TMPRSS2 ectodomain was purified and activated for peptidase activity inhibition assays as described previously . A total of 73 analogs were tested for TMPRSS2 inhibition using the commercial fluorogenic peptide substrate Boc-QAR-AMC (Extended Data Figs , Extended Data Tables ). For Scaffold 1, the addition of a methyl group in the para-position provided the best improvement in IC50 to 36 nM ('8). Other aryl and alkyl substitutions at para-and meta-resulted in Scaffold 1 compounds with IC50s ranging from 52 to 154 nM (Extended Data Fig. ; Extended Data Table ). Modifications to the Scaffold 1 ortho position ('6 and '3) produced weaker, micromolar IC50 inhibitors. The ester reversal of Scaffold 2 that placed the alpha-amino ketone on the benzoate phenyl in also resulted in low micromolar IC50 inhibitors (Extended Data Fig. ; Extended Data Table ). Scaffolds 3 and 4 yielded 16 compounds with nanomolar IC50 values by using amino methyl or amino ethyl S1-targeting head groups, respectively (Fig. ). Interestingly, Scaffold 5 compounds that were analogs of Scaffold 4 compounds did not substantially change the TMPRSS2 IC50 despite the replacement of the Scaffold 4 primary amine with a 3oxyazetdine. Incorporation of chemical moieties from nafamostat decreased IC50 values to single digit nanomolar for Scaffold 6 inhibitors with the introduction of the guanidinium group, and Scaffolds 3 and 4 with the addition of napthyl-amidine. This trend did not hold true for the incorporation of napthyl-amidine for Scaffold 5 resulting in an inhibitor with an IC50 greater than 4 µM ('158; Extended Data Table ). Based on these data, 23 compounds (Fig. ) were prioritized for subsequent biochemical and biophysical characterization. |
675f0698085116a1337f1a3a | 9 | After the experimental TMPRSS2 crystal structure was published , we also performed a virtual screen of the experimentally determined TMPRSS2 binding pocket against the ZINC-22 library . This screen contained over 200 million monocations. The pocket was optimized using the same parameters as the TMPRSS2 homology model. In this screen, a set of 59 compounds were prioritized for testing. Six compounds in the selected pool of noncovalent compounds shared a N-[4-(aminomethyl)-3fluorophenyl]acetamide substitution which we defined as Scaffold 7 and two of six Scaffold 7 compounds, '2222 and '2602, demonstrated measurable TMPRSS2 inhibition with IC50 valules of 23 µM and 59 µM, respectively (Fig. ; Extended Data Fig. ). |
675f0698085116a1337f1a3a | 10 | The docked poses of '2222/'2602 predicted that these compounds could use their primary amines to specifically engage the TMPRSS2 S1 residues S436 and D435, and the amide bond present within these compounds was oriented away from the TMPRSS2 oxyanion hole to prevent proteolytic cleavage (Fig. ). The distal ends of '2222/'2602 were predicted to engage the TMPRSS2 S1' residues V280 and the C281-C297 disulfide (Fig. ). Interestingly, the S436 residue of TMPRSS2 is only partially conserved amongst human TTSPs and other trypsin-like serine proteases found in the blood, suggesting that '2222/'2602 could potentially exploit the S436 residue for TMPRSS2 selectivity (Extended Data Fig. ). |
675f0698085116a1337f1a3a | 11 | Extended Data Fig. shows that ligands with primary amines bind differently to TMPRSS11D and TMPRSS15, depending on the presence of the residue equivalent to S436 in TMPRSS2 (absent in TMPRSS11D but present in TMPRSS15). Thus, we identified a novel amine scaffold ('2222 and '2602) that may be able to discriminate for the presence of S436 in S1. |
675f0698085116a1337f1a3a | 12 | To accurately rank the esters based on their covalent reactivity and binding potency, TMPRSS2 acylation kinetic studies were performed with the identified hits and compared to known covalent inhibitors nafamostat, camostat, and 6-amidino-2-naphthol. In this kinetic assay, compound inhibition was interrogated against 1.6 nM recombinant TMPRSS2 (Fig. ). Real-time acylation of the TMPRSS2 active site was measured by co-addition of substrate and inhibitor to microplates containing enzyme, as performed previously for TMPRSS2 with nafamostat and camostat . A subset of the covalent inhibitor series was tested, including '157, '156, '8, and '148. Esters showed a plateau of enzyme activity over time consistent with the formation of an inactive acyl-enzyme complex (Fig. ; Extended Data Fig. ). The kinact/KI values of '157 and '156 were 0.93, and 1.6 µM -1 s -1 , respectively, and were comparable to nafamostat's kinact/KI value of 2.8 µM -1 s -1 (Fig. ; Extended Data Fig. ). However, amongst these potent ester inhibitors, only the TMPRSS2 reaction progress curves in the presence of nafamostat showed a complete plateau of TMPRSS2 peptidase activity, indicating the 4-guanidino benzoate acyl-enzyme complex was stable and did not rapidly hydrolyze across the analyzed reaction time. Less potent inhibitors '8 and '148 had kinact/KI values of 0.0045 and 0.0036 µM -1 s -1 , respectively, which were comparable to camostat's kinact/KI value of 0.085 µM -1 s -1 (Extended Data Fig. ). In contrast, noncovalent inhibitors 6amidino-2-naphthol, '2222 and '2602 showed dose-dependent depression of enzyme activity, but no plateau of enzymatic activity over time (Fig. ). Accordingly, TMPRSS2 Ki values of 1.1, 13, and 40 µM were determined for 6-amidino-2-naphthol, '2602, and '2222, respectively. |
675f0698085116a1337f1a3a | 13 | To test whether inhibitors blocked TMPRSS2 protease activity on a more physiological substrate, inhibition assays were performed using purified SARS-CoV-2 S protein as a substrate and reaction progress was monitored by SDS-PAGE and Coomassie blue staining (Fig. ). Inhibitors were pre-incubated for 30 min with TMPRSS2, then transferred to buffer containing the S protein substrate and relative S protein cleavage evaluated after 15 min. The inhibitory potencies of the ester compounds were similar to one another due to the long pre-incubation time with TMPRSS2 prior to the start of the assay, allowing complete acylation of the TMPRSS2 Ser441 residue to block activity (Fig. ). At low inhibitor concentrations, the esters did not show any detectable inhibition of dasTMPRSS2 protease activity. In contrast, the noncovalent inhibitors 6amidino-2-naphthol, '2222, and '2602 completely blocked TMPRSS2-mediated S protein cleavage at high inhibitor concentrations but showed partial TMPRSS2-mediated cleavage of the S protein as the inhibitor concentrations were lowered (Fig. ). These gel-based TMPRSS2 inhibition assays demonstrate that these compounds can disable TMPRSS2 activity towards SARS-CoV-2. The molecular docking screens predicted '2222 and '2602 were noncovalent TMPRSS2 inhibitors that form hydrogen bonds with the catalytic S441 residue. To experimentally test this prediction, we designed assays that could measure ligand binding in the presence or absence of the TMPRSS2 catalytic Ser441 residue. |
675f0698085116a1337f1a3a | 14 | Benzamidine Sepharose resin specifically interacts with S1 of trypsin-like serine proteases (Fig. ), and we previously showed it can immobilize TMPRSS2 and be specifically eluted with buffer containing benzamidine hydrochloride . Furthermore, TMPRSS2 S441A mutant protein can also be immobilized and eluted from Benzamidine Sepharose (Extended Data Fig. ), demonstrating that the resin's interaction with TMPRSS2 does not depend on the TMPRSS2 catalytic serine residue. '2222 and '2602 competitively eluted TMPRSS2 from Benzamidine Sepharose, producing a TMPRSS2 protein band at its expected molecular weight (Fig. ). In contrast, the cathepsin inhibitor E-64D was incapable of eluting TMPRSS2 from the resin. Interestingly, TMPRSS2 S441A was not competitively eluted from Benzamidine Sepharose resin using '2222 or '2602, which suggested that these ligands could not bind the TMPRSS2 S1 binding site with high affinity in the absence of S441 (Fig. ). |
675f0698085116a1337f1a3a | 15 | We also employed differential scanning fluorimetry (DSF) to evaluate shifts in TMPRSS2 melting temperatures (ΔTms) for TMPRSS2 and TMPRSS2 S441A proteins in the presence of '2222, '2602, 6amidino-2-naphthol, nafamostat, and '157. All noncovalent ligands were tested at their maximum achievable solubility in aqueous buffer. The Tm values for TMPRSS2 and TMPRSS2 S441A in the absence of ligands were (52.1±0.1)°C and (49.5±0.3)°C, respectively. The noncovalent ligand 6-amidino-2-naphthol induced Tm shifts of (8.2±0.4)°C and (7.3±0.7)°C for TMPRSS2 and TMPRSS2 S441A, respectively, at a ligand concentration of 10 mM (Fig. ). The TMPRSS2 thermal stabilizations induced by 6-amidino-2naphthol were linear on a semi-logarithmic scale and did not show signs of saturation even at the highest ligand concentration (Fig. ). Nafamostat and '157 induced large maximum ΔTm values for TMPRSS2 at (23.6±0.2)°C and (8.7±0.5)°C, respectively, and demonstrated saturation binding in the TMPRSS2 thermal stabilization plot that indicated they achieved maximum ligand occupancies, likely due to covalent ligand binding (Extended Data Fig. ). In contrast, nafamostat and '157 incubated with TMPRSS2 S441A only induced small ΔTms of (1.50±0.07)°C and (2.1±0.1)°C, respectively, across the same ligand concentration ranges and did not exhibit saturation binding in the thermal stabilization plot (hollow symbols; Fig. , Extended Data Fig. ). The '2602 ligand induced a TMPRSS2 ΔTm of (3.2±0.2)°C at a ligand concentration of 2 mM, but solubility challenges prevented any higher '2602 concentrations being tested (green triangles Fig. ; Extended Data Fig. ). Across the tested '2602 concentration range, no saturation binding was observed and instead the '2602 plot resembled the TMPRSS2 thermal stabilization trends induced by 6amidino-2-naphthol. The '2602 ligand only showed a small TMPRSS2 S441A ΔTm of (0.85±0.07)°C at a ligand concentration of 4 mM, suggesting this TMPRSS2 residue was critical for '2602 binding. Thus, through two independent biophysical methods, we confirmed the results obtained in the molecular docking screens that '2222/'2602 noncovalently bind within the TMPRSS2 S1 site and rely on the TMPRSS2 Ser441 for binding. |
675f0698085116a1337f1a3a | 16 | There is an unmet need for effective respiratory virus antivirals and prophylactics to mitigate viral spread in the absence of available vaccines as well as reduce disease severity for at-risk patient groups. TMPRSS2 is an important human protease that recognizes viral particles and enables coronavirus and influenza virus infection, motivating multiple TMPRSS2 medicinal chemistry campaigns . However, selective TMPRSS2 inhibitors that completely avoid off-target effects on trypsin-like serine proteases, such as the coagulation pathway, have yet to be identified. In this study, we identified covalent small molecule inhibitors and noncovalent compounds with a novel amine scaffold as potential TMPRSS2 inhibitors. This collection of reagents and assays enabled us to rapidly screen for TMPRSS2 inhibition, kinetically characterize inhibitors, and determine high resolution cocrystal structures using a TMPRSS2 protein lacking its LDLR domain. |
675f0698085116a1337f1a3a | 17 | Our first series of compounds was identified through a large library docking campaign against a homology model of TMPRSS2, before the first TMPRSS2 crystal structure became available. Over 200 million compounds were virtually screened , identifying '2805 which was predicted to bind within the S1 of TMPRSS2 and had a reactive ester group. Our biochemical data confirmed that '2805 potently inhibited TMPRSS2 protease activity and likely formed an acyl-enzyme complex with TMPRSS2 following nucleophilic attack on '2805's ester using TMPRSS2's S441 residue. To improve TMPRSS2 binding affinity, we installed various modifications to '2805, producing 6 distinct ester inhibitor scaffolds. We employed a combination of kinetic TMPRSS2 inhibition assays and DSF ligand binding assays to interrogate noncovalent binding affinity and covalent reactivity of the ester scaffolds. We compared the inhibition potencies of these esters to nafamostat, a potent TMPRSS2 inhibitor that had been used in exploratory clinical trials for anti-TMPRSS2 COVID-19 therapy. The most potent TMPRSS2 inhibitors were '156 and '157 which contained the same naphthyl amidine as nafamostat. We determined the structure of the TMPRSS2 following incubation with '157 which confirmed that a 4-ethyl amino benzoate acyl-enzyme complex was formed. However, to understand the potency of these compounds, we solved the crystal structure of TMPRSS2 bound to 6-amidino-2-naphthol and showed that these compounds could potentially bind the TMPRSS2 S1 using their amidines or their S1-targeting headgroups before TMPRSS2 nucleophilically attacks their esters. To provide evidence for this amidine-mediated noncovalent binding for nafamostat, '157, and '156, we showed that these compounds could induce Tm shifts in the TMPRSS2 S441A protein, similar to 6-amidino-2-naphthol. In contrast, camostat, which lacks the napthyl amidine, could not induce Tm shifts for the TMPRSS2 S441A protein. These data explained how nafamostat more potently disables TMPRSS2 activity (and other trypsin-like serine proteases) than other related esters containing a single cationic, S1-targeting head group. However, both nafamostat and '157 were found to potently disable the activity of various trypsin-like serine proteases (Extended Data Fig. ), indicating that future chemical modifications to the ester scaffold may be necessary to confer TMPRSS2 selectivity. |
675f0698085116a1337f1a3a | 18 | A second molecular docking campaign was conducted against TMPRSS2 when its crystal structure became available. Using the Zinc-22 library , this docking screen identified hit compounds '2222 and '2602 which were predicted to bind within the TMPRSS2 S1. We did not acquire a TMPRSS2 cocrystal structure with the '2222/'2602 compounds during this study. We instead validated the molecular interactions predicted from the docking using biophysical assays and showed that the TMPRSS2 S1 is critical for binding (likely mediated through the TMPRSS2 S436 residue), and the TMPRSS2 S441 residue is necessary for Hbonding. In total, 9/17 TTSPs contain the S436 residue within S1 (Extended Data Fig. ). Furthermore, this Ser residue is uncommon amongst human plasma proteases, with only plasmin and FVIIa possessing Ser whereas thrombin, activated Protein C and FXa have an Ala residue at this position (Extended Data Fig. ). This subclassification of trypsin-like serine proteases has been noted before and has enabled some trypsin-like serine protease selectivity through substituted benzamidine small molecule inhibitors . |
675f0698085116a1337f1a3a | 19 | The '2602 compound showed moderate TMPRSS2 selectivity over related proteases thrombin, TMPRSS13, but inhibited trypsin and TMPRSS3 at similar levels to 6-amidino-2-naphthol (Extended Data Fig. ). To improve the potency and TMPRSS2 selectivity of future '2222 and '2602 derivative compounds, more TMPRSS2-specific amino acids within the substrate binding cleft could be targeted. As noted in our previous work , the TMPRSS2 Lys342 residue situated within S2 is relatively distinct amongst TTSPs and could potentially be targeted for H-bonding or salt bridges with a ligand. |
675f0698085116a1337f1a3a | 20 | We have explored two distinct chemistries that target the TMPRSS2 active site and identified a novel scaffold that noncovalently engages the protease within the S1 region. We have built a set of kinetic and biophysical assays that allow compounds of distinct mechanisms to be accurately compared. Furthermore, we have identified a highly crystallizable form of TMPRSS2 that provides high resolution co-crystal structures. The noncovalent scaffold (and potentially the ester-based scaffold) may be amenable to structure-guided medicinal chemistry campaigns that further improve TMPRSS2 binding potency and selectivity, paving the way for new antiviral therapeutics. |
675f0698085116a1337f1a3a | 21 | The sequence of the protease domain of TMPRSS2 was analyzed with BLASTp to identify the best available structural templates. The top five templates by sequence identity were selected including kallikrein (PDB ID 6O1G), matriptase (PDB ID 1EAX), coagulation factor XI (PDB ID 4TY6), coagulation factor VII (PDB ID 5PAB), and TMPRSS11E (PDB ID 2OQ5). Template identities ranged from 38-44%. |
675f0698085116a1337f1a3a | 22 | The ligand from 4TY6 was used during the modeling process to keep the binding site open and with appropriate side chain rotamers. In total, 1500 models were generated using RosettaCM . Models were filtered to identify those that maintained proper geometry of the catalytic triad Ser-His-Asp. A structural water found in 4 templates was added back into the models and minimized with Rosetta. A set of ~200 models were taken forward into docking calculations. |
675f0698085116a1337f1a3a | 23 | A set of control compounds were collected for retrospective docking analysis . The compounds camostat, nafamostat, and benzamidine were selected as they possess pan-serine protease activity. The ligands from templates 5PAB and 4TY6 were also used as positive controls as they contain geometry of ligands likely to be active at TMPRSS2. For these 5 compounds, 50 decoys were generated with matching physico-chemical properties but scrambled topologies using the DUD-EZ pipeline. Model that could dock all active compound with high score were taken forward into the next round of optimization. |
675f0698085116a1337f1a3a | 24 | Bias in scoring grid modifications were controlled by using an Extrema decoy set to ensure that overlycharged compounds weren't given artificially high scores. Models were removed that did not satisfy these criteria. Lastly, a set of models were tested in an In-stock screen with a series of post-docking filters to identify a single best model that yielded the highest number of viable compounds. |
675f0698085116a1337f1a3a | 25 | Fluorescent emissions were read every 1 to 5 mins over 2 hours. Wells containing substrate only (no inhibitor or enzyme) were treated as no enzyme controls (NEC). Wells containing enzyme and substrate (no inhibitor) were treated as positive controls. Average velocities were calculated for all wells (for points between 20 minutes and 1 hour) using Gen5 software. Percent activity for all wells were calculated and plotted in GraphPad Prism. |
675f0698085116a1337f1a3a | 26 | Upon arrival, compounds were diluted in DMSO (final concentration: 50 mM) and stored at -20ºC. Assays were run on a BioTek Synergy H4 Plate Reader using 384-well plates. Compounds were plated in a 96 well plate (clear, round bottom) in buffer (180 µL, 50 mM Tris, 150 mM NaCl, 0.02% TWEEN-20, pH= 8) at 300 µM (3X). dasTMPRSS2 protein was thawed on ice and diluted into buffer (3 mL buffer, 420 pM (3x)). The enzyme was first plated (7 µL, final concentration 140 pM) followed by compound (7 µL, final concentration 100 µM). The enzyme and compounds were incubated 20 min at RT followed by 5 min at 37ºC. After incubation, substrate Boc-QAR-AMC (7 µL, final concentration, 133 µM, ex: 340 (20), em: 441 (20)) was added to each well. Fluorescent emissions were read every 1 to 5 mins over 2 hours. Wells containing substrate only (no inhibitor or enzyme) were treated as no enzyme controls (NEC). Wells containing enzyme and substrate (no inhibitor) were treated as positive controls. Average velocities were calculated for all wells (for points between 20 minutes and 1 hour) using Gen5 software. Percent activity for all wells were calculated and plotted in GraphPad Prism. |
675f0698085116a1337f1a3a | 27 | Compounds that reduced TMPRSS2 peptidase activity by ≥ 40% were characterized in dose-response experiments. Assays were run on a BioTek Synergy H4 Plate Reader using 384-well plates. Compounds were plated in a 96 well plate (clear, round bottom) in buffer (180 µL, 50 mM Tris, 150 mM NaCl, 0.02% TWEEN-20, pH= 8) at various concentrations. TMPRSS2 protein was thawed on ice and diluted into a buffer (3 mL buffer, 420 pM (3x)). The enzyme was first plated (7 µL, final concentration 140 pM) followed by compound (7 µL, final concentration 100 µM). The enzyme and compounds were incubated 20 min at RT followed by 5 min at 37ºC. After incubation, substrate Boc-QAR-AMC (7 µL, final concentration, 133 µM, ex: 340 (20), em: 441(20)). Fluorescent emissions were read every 1 to 5 mins over 2 hours. Wells containing substrate only (no inhibitor or enzyme) were treated as no enzyme controls (NEC). Wells containing enzyme and substrate (no inhibitor) were treated as positive controls. Average velocities were calculated for all wells (for points between 10 minutes and 30 minutes) using Gen5 software. Percent activity for all wells were calculated and plotted in GraphPad Prism. IC50 curve fittings were generated by GraphPad Prism non-linear fits using the "[Inhibitor] vs. response (three parameter)" function. |
675f0698085116a1337f1a3a | 28 | The covalent acylation of the dasTMPRSS2 active site with nafamostat and other ester inhibitors was measured using reaction progress curve fitting that has been described for dasTMPRSS11D . Briefly, stocks of Boc-QAR-AMC substrate pre-mixed with various concentrations of nafamostat (or other ester compounds) were transferred to wells containing dasTMPRSS2 enzyme and fluorescence immediately read to capture the production of AMC dye and rapid arrest of dasTMPRSS2 activity by the inhibitor. |
632b27dae665bd28bf0d3701 | 0 | In recent years, there has been considerable interest in oxyhydrides in both materials science and physics due to their Mott insulating behaviour , superconductivity and mixed electronicionic conduction . The favorable electron and hydride ion transport properties together with the low mass of hydride ions and high electrode potential of the H 2 /H redox couple ( 2.3 V) make these materials particularly attractive for developing energy storage devices with high energy density. Since the synthesis of the first oxyhydride several tens of new structures have been prepared and characterized. Among these the most extensively investigated which exhibit some of the most interesting chemicophysical properties are the perovskite-structured ATiO 3 x H y (A = Ca, Sr, Ba) materials. BaTiO 3 x H y has been shown to incorporate high hydride ion concentrations up to y = 0.6, moreover, and has demonstrated high electric and hydride ion conductivity. Complete understanding of the electron and hydride ion transport mechanisms and their relation to the oxyhydride composition is crucial for rational design of materials with specifically tailored characteristics. |
632b27dae665bd28bf0d3701 | 1 | In BaTiO 3 x H y two possible mechanisms for hydride ion diffusion have been suggested. The first can be described as an oxidative diffusion, where the hydride ion transforms into a proton followed by interstitial diffusion. In the second mechanism hydride ions move through the material with the assistance of oxygen vacancies. Several investigations using complementary methods, such as isotope exchange, quantum chemical calculations , quasielastic neutron scattering (QENS) and inelastic neutron scattering (INS) have indicated that the latter mechanism most likely occurs in BaTiO 3 x H y oxyhydride and other ATiO 3 x H y pervoskites. While the majority of the studies are in agreement on the hydride ion transport mechanism in these solids, controversies still remain on the electronic structure and the associated electrical conductivity. |
632b27dae665bd28bf0d3701 | 2 | Initial electrical conductivity measurements indicated that the bulk phase of BaTiO 3 x H y is semiconducting. However, the experiment was carried out on a powder sample without sintering, which may introduce errors due to grain-boundary effects. Recently, conductivity measurements on epitaxial thin films revealed that BaTiO 3 x H y has metallic conductivity with high hydride concentrations (y > 0.2), while it is semiconducting with lower hydride compositions (y < 0.2). Liu et al. proposed that the semiconductor-like behaviour is a result of electron polaron formation. Polarons are quasiparticles composed of a localized charge carrier within a potential energy well that is self-generated by distorting the local lattice. In BaTiO 3 x H y the substitution of oxygen with hydrogen introduces an electron in the Ti 3dband which can potentially localize and form an electron polaron. Thermally activated polarons can hop to neighboring Ti ions and would display semiconductor-like conductivity in the bulk phase. However, a more recent study using INS and density functional theory (DFT) calculations showed that the extra electron is more likely to delocalize among all Ti 3d-bands generating a bandstate with metallic electrical conductivity. Several investigations on the oxyhydride with solid-state nuclear magnetic resonance (NMR) have also proposed that the electronic structure is more likely a bandstate. In contrast to other reports Misaki and co-workers suggested that the hydride site is occupied by two hydrogens. |
632b27dae665bd28bf0d3701 | 3 | Here we report a comprehensive study of hydride local environment and dynamics by employing 2 H solid-state NMR on three distinctly prepared BaTiO 3 x H y oxyhydrides with different hydride compositions. We support our experimental observations by using DFT calculations, which provide theoretical electric-field gradient tensor parameters sensitive to the local geometry. The combined results are able to conclusively differentiate the electronic structure between the polaron, single and double occupied bandstates in BaTiO 3 x H y and to provide new insights on the hydride diffusion mechanism. We anticipate that our approach will be very beneficial for the studying similar systems in the future. |
632b27dae665bd28bf0d3701 | 4 | All samples were prepared using BaTiO 3 (500 nm particle size, 99.9%, ABCR GmbH), which was dried in an oven at 200 for 14 h. The steps involving sample preparation for synthesis were performed in an Ar-filled glovebox. The oxyhydride sample containing hydrogen was obtained by mixing BaTiO 3 (BTO) with CaH 2 in molar proportions of 1:0.6. Approximately 10 g of the mixture was pressed into pellets and placed inside a sealed stainless steel ampule and heated in an evacuated silica jacket at a temperature of 600 for 48 h. The deuterated samples (BTOD NAB and BTOD CA ) were synthesized following a similar procedure, but instead using NaBD 4 ( 98%, Sigma) and CaD 2 . In both synthesis the precursors were mixed at the same BTO to reducing agent proportion as before (1:0.6). CaD 2 was prepared as described previously. The third sample (BTOD EXCH ) was obtained by first reducing BTO with CaH 2 in molar proportion of 1:2.25 under identical conditions as given above. Subsequently, hydrogen to deuterium exchange was performed by subjecting the sample to D 2 (99.9%, AGA) in a stainless-steel autoclave pressurized to 30 bar at 600 for 12h. Afterwards, the products were washed 3-4 times with 0.1 M NH4Cl/Methanol solution (BTOH, BTOD CA and BTOD EXCH ) or 0.1 M hydrochloric acid (BTOD NAB ). Finally, all samples were dried under vacuum at 120 C for ⇠ 24 h. |
632b27dae665bd28bf0d3701 | 5 | The 2 H NMR experiments were performed on a Bruker 400 MHz spectrometer operating at a Larmor frequency of 61.405 MHz with either a 2.5 mm HX, a triple resonance low-temperature (LT) 3.2 mm probe or a triple resonance 4.0 mm probe. On the 2.5 mm probe the radio frequency (RF) amplitude of the pulses was 100 kHz, which was measured by the nutation behaviour of deuterated ethanol (C 2 H 5 OD). The 2 H shifts were referenced externally to C 2 H 5 OD at 2.50 ppm. For short high-power adiabatic pulses the tanh/tan 29 pulse scheme was employed with 5 MHz sweep width and a 50 µs pulse length. |
632b27dae665bd28bf0d3701 | 6 | On the LT 3.2 mm probe the RF amplitude of all the pulses was 60 kHz, which was calibrated with the nutation behaviour of glycine-( 13 C 2 , 15 N, 2,2-d 2 ). The 2 H shifts were referenced externally to deuterated glycine at 3.70 ppm. For frequency-swept pulses the same parameters were used as on the 2.5 mm probe with the exception of the RF field amplitude. The sample temperature for the variable temperature experiments was varied from 96 to 305 K and determined by measuring the 79 Br spin-lattice relaxation time constant (T 1 ) relaxation times of a small amount of KBr (99%, Sigma) mixed with the sample. On the 4.0 mm probe the RF amplitude of all the pulses was 42 kHz which was measured by the nutation behaviour of deuterated ethanol (C 2 H 5 OD). The 2 H shifts were referenced externally to C 2 H 5 OD at 2.50 ppm. For short high-power adiabatic pulses the tanh/tan 29 pulse scheme was employed with 5 MHz sweep width and a 100 µs pulse length. The sample temperature for the variable temperature experiments was calibrated based on the 207 Pb shift in Pb(NO 3 ) 2 (99%, Sigma). Additional experimental solid-state NMR details can be found in the supplementary material. |
632b27dae665bd28bf0d3701 | 7 | In order to extract the tensor principal components along with the Euler angles relating the shift and quadrupolar tensors we fitted the 2D static adiabatic shifting d-echo spectra using SIMPSON-simulated spectra with a Levenberg-Marquardt damped least-square minimization algorithm implemented in C++. The errors of the values were evaluated using an inhouse Monte Carlo procedure, which has been described in detail previously. Density functional theory (DFT) calculations were carried out using the Quantum ESPRESSO , WIEN2k and ORCA packages. |
632b27dae665bd28bf0d3701 | 8 | Depending on the software package and the BaTiO 3 x H y electronic state either the generalized gradient approximation exchange-correlation functional of Perdew, Burke, and Ernzerhof (PBE) , PBE with the Hubbard correction or PBE0 were used for the calculations. A detailed description of the computational methods is provided in the supplementary material. |
632b27dae665bd28bf0d3701 | 9 | H and 2 H solid-state NMR Here we are interested in examining the electronic structure of barium titanium oxyhydride (BTOH) in order to understand if the additional electrons introduced in the reduction process form electron polarons (Fig. ) or delocalized bandstates (Fig. ) and 1(c)). Furthermore, we address whether the bandstate constitutes hydride sites with a single or double occupancy as shown in Fig. ) and 1(c). |
632b27dae665bd28bf0d3701 | 10 | Hydride ions are excellent candidates for probing the electronic structure of BTOH with solid-state NMR, since they are directly bonded to the Ti ions, which host the localized electron polarons or conduction electrons in the 3d-band. Usually the most abundant hydrogen isotope 1 H is studied by NMR, however, resonances of surface-adsorbed water and/or hydroxyl groups overlap with the hydride signals complicating the quantitative and qualitative analysis of the spectra (see Fig. ). The interfering background signals can be removed by isotopic substitution of 1 H with deuterium 2 H in the material (see Fig. ). Furthermore, 2 H exhibits a nuclear electric quadrupole moment that, coupled with the electric-field gradient, gives rise to the quadrupolar interaction, which is very receptive to changes in the local geometry, and gives a lineshape that is sensitive both to the structure and dynamic processes on the ms-µs timescale. The disadvantage of 2 H is the reduced sensitivity due to the lower gyromagnetic ratio. |
632b27dae665bd28bf0d3701 | 11 | To this end deuterated versions of BaTiO 3 x H y were synthesized. Two samples were prepared by reducing barium titanium oxide with deuterated NaBD 4 and CaD 2 , while the third was obtained by reducing with CaH 2 followed by hydride exchange with gaseous deuterium D 2 . We denote each of the samples as BTOD NAB , BTOD CA and BTOD EXCH , respectively. The 2 H NMR We continued by evaluating the amount of deuterium in each sample using the 2 H spectra in Fig. . The quantification was performed by relating the 2 H peak areas between BaTiO 3 x D y and a substance (glycine) with known deuterium content. The quantification results reveal that BTOD NAB has a lower hydride concentration than samples reduced with CaH 2 , which is in agreement with previous studies. The hydride shifts are tabulated in Table , along with the determined hydride content y. We notice that the shift becomes more negative with increasing hydride concentration, which is the first indication that the examined materials are in a bandstate. We discuss this topic in detail in the section 3.2. |
632b27dae665bd28bf0d3701 | 12 | We investigated the local hydride environment in the three BTOD samples using state-of-the-art static solid-state NMR methods. Static solid-state NMR experiments allow the determination of the tensor parameters of the interactions affecting the nuclear spin transitions. In the present case the relevant interactions are the shift and the quadrupole interactions. The shift interaction provides information about the chemical environment, bonding geometry and delocalisation/polarization of the spin density via the hyperfine interaction, whereas the quadrupole interaction is very sensitive to the local lattice and coordination geometry and dynamics on the µs-ms time scale. The recently reported 2D adiabatic shifting d-echo experiment can separate and correlate the two interactions as shown in Fig. , which allows a very accurate evaluation of the tensor parameters. The projections onto each axis of the 2D spectrum give information about the two interaction tensors, while the 2D lineshape provides the Euler angles relating the orientation of the two tensors. |
632b27dae665bd28bf0d3701 | 13 | The adiabatic shifting d-echo experimental spectra together with the best fits are given in Fig. , while the extracted NMR parameters are summarized in Table . A qualitative comparison of the spectra reveals that the main spectral features are very sim-ilar among the three samples, except of course for the sharp peak appearing at zero frequency in the quadrupolar dimension due to the impurity (TiD 2 ) in BTOD EXCH . Interestingly, despite the anionic disorder and the presence of oxygen vacancies in these solids the spectra display very well-defined Pake patterns in the quadrupolar dimension indicating high structural fidelity of the local hydride geometry throughout the materials. Furthermore, the quadrupolar coupling constants C Q in Table are relatively small and constant at ⇠ 25 kHz (typical 2 H environments have C Q up to 200 kHz) suggesting high local symmetry as expected for an approximately cubic crystal system. The quadrupolar tensor parameters and the Euler angles (see Table ) match within error among the investigated samples, which implies that the local hydride environment and electronic structure type is identical in the three oxyhydrides. |
632b27dae665bd28bf0d3701 | 14 | Having established that the three materials exhibit the same type of electronic structure we continued by elucidating this electronic state. We began by examining the shift tensor parameters, which display considerable differences among the different samples (see Table ), which is unsurprising since variation of the isotropic shift with sample composition was already detected in the MAS spectra (see Fig. and Table ). We note that there are significant discrepancies between isotropic shift values determined under MAS and static conditions, which arise due to an inherent imprecision of fitting the broad static lineshapes. Therefore, we used only the isotropic shift values (see Table ) estimated from MAS spectra to draw conclusions about properties of the oxyhydrides. |
632b27dae665bd28bf0d3701 | 15 | The measured isotropic shifts are well outside the typical region of 1 H/ 2 H shifts (0 12 ppm). This indicates the presence of the hyperfine interaction between the nuclei and unpaired electrons, which can lead to large isotropic shifts and shift anisotropies. In the polaron state the source of unpaired electrons would be the electron polaron, while for the bandstate the conduction electrons. In both cases the main contribution to shift would arise from the Fermi contact (FC) term. In metals the NMR shift is called the Knight shift K with a Fermi contact part K FC given by: |
632b27dae665bd28bf0d3701 | 16 | where h|f k (0)| 2 i E F is the average unpaired electron density of the s-band electrons near the Fermi level and c P is the Pauli susceptibility, which is proportional to the number of states at the Fermi level. The Fermi contact shift d FC for the polaron state takes a similar form: |
632b27dae665bd28bf0d3701 | 17 | where r a b (0) is the unpaired electron density in the s-orbital and c is the susceptibility per paramagnetic centre. We notice from Equation 2 that the shift in the case of the polaron depends on the transferred spin density to the hydrogen s-orbital r a b (0), therefore this contribution is additive meaning that a hydride adjacent to two electron polarons would give double the shift compared to a hydride in the proximity of a single polaron. Therefore, we expect with increasing hydride content (potentially increasing electron polaron concentration) for a sec- For the bandstate the FC part of shift is proportional to the density of states at the Fermi level as shown in Equation ), and so depends on the number of electrons in the conduction band (3dband of Ti) and the number of charge carriers. In a bandstate configuration substitution of an oxide ion with a hydride introduces an additional electron in the Ti 3d-band, therefore an increase in the number of charge carriers is expected. A prior study has confirmed that higher hydride concentrations in the structure result in a larger number of charge carriers for BaTiO 3 x H y with y = 0.14 0.35. Thus, we attribute the isotropic shift becoming more negative with the increasing hydride content (see Table ) to the metallic character of the material. Furthermore, the negative shift is consistent with the polarization mechanism. More specifically, according to the Goodenough-Kanamori rules the conduction electrons in the t 2g -band polarize the e g -band, resulting in a negative spin density in the hydride s-band and consequently a negative contact shift. We speculate that this mechanism may not occur for the double occupied bandstate due to the different bonding geometry of the two hydrides, which allows direct overlap between the hydride s-orbitals and the t 2g -orbital giving rise to the delocalisation mechanism and a positive shift. Until now we only considered the FC term of the Knight shift, however, it also possesses orbital and spin-dipolar parts, which are orientation dependent and contribute to the shift anisotropy. The spindipolar term depends on the density of states at the Fermi level and so increases with higher hydride concentrations. This explains the directly proportional relation between shift anisotropy (see Table ) and the hydride concentration. |
632b27dae665bd28bf0d3701 | 18 | Together these data provide strong evidence that the barium titanium oxyhydrides have a bandstate electronic structure and the NMR shift corresponds to the Knight shift. Previously, a number of NMR investigations had observed negative 1 H or 2 H shifts and associated it with the metallic character of the material. Misaki et al. , in correspondence with the current study, demonstrated an essentially linear relation between the hydride content in BaTiO 3 x H y and the Knight shift. In contrast, Guo and coworkers detected different hydride isotropic shift values for samples with similar hydrogen content. This discrepancy can be explained by the formation of oxygen vacancies, which also introduce electrons in the 3d-band of Ti. Indeed, the Knight shift becomes more negative in the reported materials with increasing oxygen vacancy concentration, confirming the proposed correlation between number of conduction electrons and the NMR shift. Finally, we call attention to the lineshapes of the MAS spectra in Fig. ) and of the static shift projection in Fig. ) both of which display a skew towards shifts frequencies. This can be explained by different crystallites in the sample having distinct hydride or oxygen vacancy concentrations, which lead to a distribution of Knight shifts. |
632b27dae665bd28bf0d3701 | 19 | In the previous section we showed that the three oxyhydrides have the same type of electronic structure, in which the additional electrons introduced during the reduction form delocalized bandstates rather than electron polarons. Here we provide further evidence for the bandstate electronic structure by exploiting differences in the temperature dependence of NMR parameters between metals and polaronic states. Thereby we measured longitudinal relaxation times T 1 and Knight shifts K of a representative sample (BTOD CA ) for range of temperatures ⇠ 100 300 K. |
632b27dae665bd28bf0d3701 | 20 | In simple metals there is an important relationship between the longitudinal relaxation times T 1 and the Knight shift K, namely the Korringa relation, which states that the quantity T T 1 K 2 is constant as function of temperature T . The obtained T T 1 K 2 values against temperature are given in Fig4(a), which demonstrates that NMR parameters of 2 H in BTOD CA indeed follow the Korringa relation. Moreover, the magnetic susceptibility was determined as a function of temperature (see Fig. )), which is essentially constant in this temperature range (varies less than 10% from the mean value). This is in agreement with the character of the material as the Pauli susceptibility is temperature independent, while for the polaronic state the susceptibility is expected to change by several orders of magnitude over this tem-perature range according to a Curie-Weiss law. Together these observations provide unambiguous evidence for metallic character in the electronic structure of the studied materials. Moreover, the data indicate that the bandstate electronic structure is prevalent for different hydride contents y = 0.13 0.31 and within the temperature range of ⇠ 100 300 K. However, our conclusions are in discrepancy with the study by Bouilly et al. , who investigated the electric conductivity of epitaxial films of barium titanium and reported semiconducting behaviour at low hydride concentrations y < 0.2, while metallic character was detected with y > 0.2. Furthermore, with moderate hydride content a semiconductor-metal transition at ⇠ 200 K was observed. Previously, it was proposed that strain in epitaxial films could favor polaron formation, which would account for the discrepancies with the current study. |
632b27dae665bd28bf0d3701 | 21 | Until now all lines of evidence presented here have indicated that the electrons form delocalized bandstates which is in agreement with previous INS and solid-state NMR studies. However, we have not resolved whether the bandstate constitutes single or double occupied hydride sites, which has been subject of dispute in previous reports. In order to address this issue we computed electric-field gradient (EFG) tensors using DFT calculations on the two bandstate structures along with polaronic state for a comprehensive comparison between the different states. We performed the calculations on three different quantum chemistry programs, Quantum ESPRESSO, WIEN2k and ORCA, each treating the electron wavefunction in a distinct manner. |
632b27dae665bd28bf0d3701 | 22 | The DFT optimized structures used for computations in periodic codes (Quantum ESPRESSO and WIEN2k) are given in Fig. , while ORCA utilized the cluster models derived from these structures shown in Fig. . In the periodic DFT calculations the correct electronic structures were confirmed by calculating the electron band structure (see Fig. ), which clearly show conduction bands crossing the Fermi level for the two bandstates and a narrow band gap for the polaron. Afterwards we computed the 2 H EFG tensors and evaluated the quadrupolar parameters, which are given in Table . First we note that for all of the three DFT approaches the polaron and single occupied bandstate quadrupolar parameters are very similar and differentiation between the two experimentally would not be possible, while the double occupied bandstate 2 H quadrupolar parameters are distinct from the other two structures. All of the DFT approaches predict that the double occupied bandstate quadrupolar coupling constant is almost an order of magnitude larger than the experimental value, which indicates that the hydrogen sites host only a single hydrogen ion. However, in the single occupied cases (both bandstate and polaron) there is noticeable discrepancy between the quadrupolar coupling constants obtained with Quantum ESPRESSO and the other two methods. This is most likely due to the different treatment of the electron wavefunctions in each of the software packages. In Quantum Espresso calculations employ plane wave functions to represent the valence electron wavefunction, while the core electrons are approximated with a pseudopotential. WIEN2k |
632b27dae665bd28bf0d3701 | 23 | and ORCA treat all electrons explicitly and therefore are more accurate methods for computing properties sensitive to core electrons wavefunction, i.e. chemical shift, hyperfine and electricfield gradient tensors. As a result, with the latter DFT methods we observe an excellent match between the experiment and calculated quadrupolar parameters and so we can conclude that BTOH adopts a single occupied bandstate structure. |
632b27dae665bd28bf0d3701 | 24 | Finally, we discuss the implications of the NMR data on the vacancy-assisted dynamics mechanisms. A QENS investigation on the hydride conduction dynamics suggested that at low temperatures ( 250 K) the hydride ion hopping occurs between the nearest oxygen vacancy located orthogonally to the hydride ion, whilst at high temperatures (> 400 K) the hydride ion jumps to the next-nearest oxygen vacancy located vertically or horizontally with respect to the hydride ion in the lattice. The study also reported that for both mechanisms the hydride ion mean residence times are of the order of 1 100 ps, which indicates that the process is in the fast exchange regime on the NMR timescale and fully averaged tensor parameters would be detected. |
632b27dae665bd28bf0d3701 | 25 | In an effort to further our understanding of the hydride dynamics we acquired additional static shifting d-echo spectra of BTOD CA (see Fig. ) at a very low (100 K) and high temperature (420 K). However, no significant changes are observed in the quadrupolar spectra (see Fig. ) nor the 2D spectra (see Fig. ), which suggest either complete absence of any dynamics or dynamic processes that occur in the slow exchange regime, or otherwise do not average the quadrupolar interaction parameters over the entire temperature range 100 420 K. Fig. (b) displays the low-temperature nearest-neighbour hopping mechanism proposed by QENS, in which the EFG tensor orientation rotates with each hydride ion jumps to the nearest oxygen vacancy and we expect to observe averaged tensor parameters experimentally. Assuming a random walk diffusion model (see Fig. ) we predict that the occupation probability of all six the coordination positions of the Ti ion to be equal. According to this model the experimentally observed EFG would be the average over the six available hydride positions. The simulated spectrum of the averaged tensor for the single occupied bandstate is given in Fig. ) showing a single sharp peak. The observed experimental spectrum is clearly distinct from the simulation and suggests that at room-temperature hydride ions do not diffuse freely through the entire material on the ps-ns timescale via the nearest-neighbour hopping mechanism. Fig. (c) shows the high-temperature next-nearest-neighbour hopping mechanism proposed by QENS, which does not lead to a change in the EFG tensor orientation with vertical or horizontal hydride ion jumps to next-nearest oxygen vacancies. As a consequence, the EFG tensor is not averaged and the quadrupolar spectrum (see Fig. ) predicted by DFT is not altered, which matches the experiment (see Fig. ). Therefore, either hydride ions predominantly hop to the next-nearest oxygen vacancy throughout the studied temperature range (100 420 K) or the low-temperature mechanism takes place on a drastically slower Interestingly, a recent DFT investigation on these materials indicated that the configuration with oxygen vacancies located in the next-nearest position to the hydride ion exhibits the lowest energy, thus is the most stable. Possibly, this ion arrangement dominates within the oxyhydride structure, therefore hydride ion diffusion occurs exclusively via the next-nearest-neighbour hopping mechanism. |
632b27dae665bd28bf0d3701 | 26 | We investigated the electronic of structure and hydride conduction dynamics of barium titanium oxyhydride using combined static and MAS solid-state NMR, DFT calculations and magnetic measurements. The dependence of the electronic properties on the oxyhydride synthesis route and hydride concentration was evaluated in three oxyhydrides. The determined quadrupolar parameters and Euler angles relating the shift and quadrupolar tensors suggest hydride ions with indistinguishable local coordination geometries for the three samples. The isotropic shift and shift anisotropy in the oxyhydride increase in magnitude with higher hydrogen concentrations, which strongly suggest that electrons form delocalized bandstates. This is further supported by the temperature dependence of the NMR parameters, which follow the Korringa relation typical for metallic systems. Together these data indicate that barium titanium oxyhydride adopts a bandstate electronic structure rather than a polaronic state and would exhibit metallic electronic conductivity, which is in agreement with previous reports. Moreover, by computing the EFG tensors using DFT we resolve the hydride occupancy in the bandstate, which implies that the single occupied bandstate is more likely to occur in the studied samples. Finally, we demonstrate that the hydride ion dynamics occur on a significantly slower timescale than reported before or else occur exclusively by hopping to nextnearest oxygen vacancies. The hydride hopping mechanism between next-nearest oxygen vacancies is shown in (c). Below the models we have the overlapped spectra obtained experimentally (in black) and by simulation of the averaged quadrupolar interaction (in red) using quadrupolar parameters from WIEN2k. |
654e54fddbd7c8b54b04ed70 | 0 | Transition metal dichalcogenides (TMDs, chemical formula MX2, M = transition metal, X = S, Se, Te) are a class of two-dimensional materials consisting of individual MX2 layers stacked by van der Waals forces. Among them, molybdenum disulfide (MoS2) has been studied extensively in a wide range of fields such as electrochemical energy storage, electrocatalysis, electrochemical water desalination, or tribology . As a consequence, a large number of synthesis approaches resulting in diverse morphologies and/or crystallite sizes are reported in literature. While exfoliation or gas phase deposition methods like chemical vapor-or atomic layer deposition can produce mono-/few-layered materials with high precision, hydrothermal methods can be employed to obtain MoS2 with varying morphology in larger quantities. For the latter, morphology, particle/crystallite size, or crystallinity can be tuned by variation of synthesis parameters. Specifically, it was demonstrated that lowering the pH of the reactant solution of the hydrothermal synthesis will yield smaller crystallite sizes. Other studies have shown that the lattice parameters of MoS2 can be directly tuned, for example, by the introduction of foreign species / pillars into the interlayer space, expanding the c-lattice parameter. This includes top-down approaches like the topotactic insertion of alkylammonium cationic pillars as demonstrated by Schöllhorn et al., however a previous chemical reduction and ion exchange are required. In summary, the structural properties of MoS2 can be tuned individually by different approaches at several length scales, from macroscopic tuning of particle or crystallite size, down to the microscopic tuning of lattice parameters like the interlayer spacing. It would be desirable to control MoS2 structure over several length scales at once in a one-pot synthesis approach, allowing to readily tailor emerging MoS2 functional properties. |
654e54fddbd7c8b54b04ed70 | 1 | electrode materials are highly dependent on the MoS2 crystallite size. The authors found that the phase transition during lithiation of bulk MoS2 is suppressed in nanostructured MoS2, leading to the emergence of solid-solution Li + intercalation with pseudocapacitive properties, that is, kinetics which are not limited by solid-state diffusion. Yao et al. |
654e54fddbd7c8b54b04ed70 | 2 | further demonstrate that, in addition to decreased crystallite sizes, also the introduction of lattice disorder can enable such pseudocapacitive intercalation properties of MoS2. Manipulating the interlayer distance, i.e., the geometry of the nanoconfined space between individual MoS2 layers, has been described to promote electrochemical ion intercalation reactions by reducing diffusion barriers and increasing available ion storage sites. The use of hydroxy pillars in a "3D porous graphene aerogel decorated with oxygen-incorporated MoS2" led to an expanded MoS2 interlayer spacing from 6.15 to 10.15 Å, resulting in increased Li + intercalation kinetics and storage capacity. Similarly, hollow nanospheres of hydroxycontaining MoS2 nanosheet subunits showed an interlayer spacing between 9.5-10.0 Å, also leading to improved kinetics and capacity of Li + intercalation. Overall, the state-of-the-art literature on MoS2 ion intercalation hosts clearly demonstrates that both, variation of particle/crystallite sizes and tuning of the interlayer spacing are viable strategies to improve ion intercalation kinetics and storage capacity. The former strategy improves kinetics by reducing diffusion distances and suppressing phase transformations, while the latter reduces diffusion barriers and increases storage sites. However, to date, no versatile synthesis approach to control both, crystallite size and interlayer spacing at once has been described. Furthermore, the understanding of MoS2 host-pillar interaction in interlayerexpanded MoS2 is insufficient and it is unclear how pillars behave during electrochemical cycling. |
654e54fddbd7c8b54b04ed70 | 3 | To address these points, this study presents a versatile, one-pot hydrothermal synthesis approach to simultaneously control MoS2 crystallite size by pH adjustment, interlayer distance by the use of alkylamine organic pillars, and interlayer composition by varying pillar concentration. Organic pillars are interchangeable allowing to widely tune MoS2 nanoconfinement chemistry via this approach. Comprehensive structural investigation reveals the resulting morphology and crystallography as a function of synthesis parameters and clarifies the interaction between MoS2 host and alkylamine pillars. Electrochemical lithium intercalation properties are conclusively linked to the MoS2 structure, with pseudocapacitive intercalation properties being amplified for smaller crystallites, larger interlayer distances and reduced pillar concentration. While we demonstrate the viability of our materials for lithium intercalation, the findings may be of broader relevance. Crystallite size and/or interlayer engineering of TMDs are considered critical for, among others, intercalation of multivalent ions, metal ion adsorption for water purification, or electrocatalytic hydrogen evolution . |
654e54fddbd7c8b54b04ed70 | 4 | Molybdenum disulfide (MoS2) samples were synthesized using a one-pot hydrothermal method, following the procedure outlined by Geng et al. 300 mg MoO3 (Alfa Aesar), 450 mg thioacetamide (Thermo Fisher Scientific), and 3 g urea (Merck KGaA) were dissolved in 50 ml deionized water and stirred for 30 minutes in a glass beaker. To control the pH of the solution, 1 M hydrochloric acid (HCl, Merck) or 1 M lithium hydroxide (LiOH, Merck) were added as needed to reach either pH 1 (Nano-MoS2) or pH 5.5 (Micro-MoS2), respectively. The pH was measured using a pH meter (Blueline 14 pH/Xylem Analytics). The solution was transferred to a 100 ml Teflon vessel, which was used for hydrothermal synthesis in an autoclave (BRHS-100, Berghof). The synthesis temperature was set at 235 °C and maintained for two hours after reaching the target temperature. The pressure in the autoclave was maintained at ca. 13 -18 bar. The reaction products were collected by centrifugation at 6000 rpm for 10 min, followed by vacuum filtration with washing in deionized water and ethanol. The filtered powder was dried at 80 °C in an oven overnight. |
654e54fddbd7c8b54b04ed70 | 5 | Furthermore, interlayer-expanded MoS2 samples were also synthesized by the one-pot hydrothermal approach. In this case, 1,6-hexanediamine (HDA), 1,8-octyldiamine (ODA) or 1,12-diaminododecane (DDDA) (all Thermo Fisher Scientific) were introduced to the precursor solution described in the previous section. The quantity of HDA/ODA/DDDA added was determined based on a molar ratio of 1:x between diamine and MoS2, where x was either 1 or 0.5. Post addition of diamine, the pH of the hydrothermal solution was adjusted to pH 1. The synthesis conditions and methodology for the powder remained consistent with those employed for the pristine materials, with the exception of the synthesis temperature, which was set at 180 °C. |
654e54fddbd7c8b54b04ed70 | 6 | Powder X-ray diffraction (XRD) patterns of the MoS2-based materials were obtained in Bragg-Brentano geometry using a Bruker D8 Advance diffractometer operating with a Cu Kα radiation source (λ = 1.5406 Å) using a 0.02° step size and a dwell time of 1 s. Powder size distribution analysis was performed using a laser diffraction particle size analyzer (Mastersizer 3000, Malvern) to measure the size of non-spherical particles by the dynamic light scattering (DLS) method. A mixture of isopropanol and deionized water (1:1 volumetric ratio) was used as a dispersant. Raman spectroscopy was performed on bulk powder samples using a Renishaw InVia confocal Raman microscope with a 532 nm excitation laser. The laser power was kept at 100 µW to prevent the oxidation of samples. At least three different spots were investigated to check the homogeneity of samples. |
654e54fddbd7c8b54b04ed70 | 7 | operating at an accelerating voltage of 80 kV. For TEM sample preparation, powders were ground using a mortar. Then, the carbon-coated Formvar film of a copper TEM grid was carefully rubbed on the fine powders. The surface chemistry of the MoS2-based materials was analyzed in powder form by X-ray photoelectron spectroscopy (XPS). The powder XPS was conducted in a SPECS UHV system (FOCUS 500 equipped with monochromatic X-ray source, |
654e54fddbd7c8b54b04ed70 | 8 | PHOIBOS 150 hemispherical energy analyzer with 2D DLD detector) using the Al Kα (hν = 1486.6 eV) radiation. The spectra were collected at 200 W with a pass energy of 10 eV and a step-size of 0.1 eV in a fixed analyzer transmission mode. The collected spectra were later calibrated to the signal of C-C/C-H sp3 at 284.8 eV as a reference on CasaXPS software. |
654e54fddbd7c8b54b04ed70 | 9 | A one-pot hydrothermal synthesis strategy for MoS2 is developed that allows to simultaneously control both the crystallite size and geometric nanoconfinement environment (i.e., interlayer distance and guest species density) of the material. Pristine MoS2 is obtained following a hydrothermal process outlined by X. Geng et al., and the particle size is controlled by adjusting the pH of the precursor solution, with more acidic conditions yielding |
654e54fddbd7c8b54b04ed70 | 10 | Two MoS2 samples are synthesized from precursor solutions with controlled pH value with the goal to achieve a variation in product crystallite size. Small and large particle/crystallite sizes of the MoS2 products are obtained from hydrothermal processes with aqueous precursor solutions of pH 1 and pH 5.5, and labelled nano-MoS2 and micro-MoS2, respectively. |
654e54fddbd7c8b54b04ed70 | 11 | SEM images of the products show the formation of larger secondary particles consisting of agglomerates of flake-shaped primary particles, with several microns of secondary particle size for nano-MoS2 (Fig. ) and tens of microns for micro-MoS2 (Fig. ). The differences in primary particle sizes are revealed at higher magnification, with nano-MoS2 exhibiting lateral dimensions of the two-dimensional MoS2 layers in the range of ca. 100 nm as well as a low thickness/number of layers well below 10 nm, as compared to micro-MoS2 with lateral sizes around 500-1000 nm and a larger thickness (Fig. ). The observations from SEM are confirmed by dynamic light scattering (DLS) analysis, which can quantify the size of MoS2 secondary particle agglomerates, revealing that the average size of the nano-MoS2 |
654e54fddbd7c8b54b04ed70 | 12 | X-ray diffraction (XRD) is utilized to assign the crystal structure of the synthesized materials as hexagonal 2H-MoS2 according to PDF card #01-071-9809 (Fig. ). The interlayer distance is identified according to the (002) signal at ca. 13.9° 2θ as 6.15 Å for both nano-MoS2 and micro-MoS2 samples. A significant peak broadening of the (002) peak at 13.9° for nano-MoS2 is a further qualitative confirmation of the reduced crystalline domain size compared to micro-MoS2. Raman spectra of the two samples (Fig. ) show the two main characteristic peaks at 382 and 406 cm -1 , which correspond to E 1 2g and A1g vibrational modes of the semiconducting (2H) phase. The results confirm that the employed hydrothermal synthesis route leads to the formation of 2H-MoS2, independent of crystallite size. |
654e54fddbd7c8b54b04ed70 | 13 | To understand the influence of the crystallite size of hydrothermally synthesized MoS2 materials on electrochemical lithium intercalation behavior, the samples were investigated in a standard, lithium-containing organic electrolyte (1 M LiPF6 in EC/DMC, LP30) using cyclic voltammetry (CV). The potential range was selected from 1.0 -3.0 V vs. Li + /Li allows to study the Li + intercalation reaction according to |
654e54fddbd7c8b54b04ed70 | 14 | with a theoretical capacity of 167 mAh/g (for x=1), without occurrence of the conversion-type reaction below 1.0 V vs. Li + /Li. The CV profiles of micro-MoS2 and nano-MoS2 are presented in Fig. . At low sweep rates (up to ca. 1 mV/s), both samples reveal the presence of a redox signal centered around 1.7 V vs. Li + /Li (cathodic sweep) and 1.9 V vs. Li + /Li (anodic sweep) |
654e54fddbd7c8b54b04ed70 | 15 | indicative of the Li + (de)intercalation reaction. For micro-MoS2, it is visible that the signal is split into two separate peaks, while for nano-MoS2, the signal is significantly broadened with the peaks apparently merging and a more significant rectangular current emerging. This behavior is indicative of an increasingly pseudocapacitive Li + intercalation behavior with decreasing crystallite size in nano-MoS2, in line with previous reports. Accordingly, at higher sweep rates above 10 mV/s, the CV signature of micro-MoS2 rapidly becomes resistive, indicative of diffusion-limitations, while even at 100 mV/s a capacitor-like CV signal can be observed for nano-MoS2 in line with kinetically favorable pseudocapacitive Li + intercalation. Galvanostatic charge/discharge (GCD) tests were performed at constant specific currents between 0.02 -5 A/g (Fig. ). At the lowest rate of 20 mA/g, both MoS2 samples show a comparable first cycle anodic (delithiation) capacity of 175 and 182 mAh/g for micro-MoS2 and nano-MoS2, respectively, slightly above the theoretical capacity of 167 mAh/g for LixMoS2 with x=1. At higher rates, specifically above 1 A/g, nano-MoS2 delivers significantly higher capacities due to the reduced kinetic limitations by solid-state diffusion, confirming the CV results. Nano-MoS2 still exhibits a capacity of 90 mAh/g at a high rate of 5 A/g, which corresponds to a charge/discharge time of ca. 1 min, while micro-MoS2 only exhibits 38 mAh/g. These values are quantitatively comparable to a study of Cook at el., probing lithium intercalation in bulk versus nanostructured MoS2, where the latter was prepared by thermal sulfurization of nanostructured MoO2 in H2S. |
654e54fddbd7c8b54b04ed70 | 16 | Galvanostatic cycling stability was tested at an intermediate rate of 0.5 A/g (Fig. ). A capacity drop in the initial 15 cycles can be observed, before both samples stabilize at around 105 mAh/g. This capacity remains relatively stable, reaching around 100 mAh/g after 200 cycles. We hypothesize that the initial capacity drop can be related to the ongoing phase transformation from semiconducting 2H to metallically conducting 1T phase of MoS2 during initial lithiation cycles that has been described before. Hence a route to improve cycling stability could be utilizing fully 1T-transformed MoS2. |
654e54fddbd7c8b54b04ed70 | 17 | Building on the one-pot hydrothermal synthesis route outlined in the previous section, the precursor solution of nano-MoS2 was adapted to additionally achieve microscopic control over the MoS2 crystal structure to further improve the electrochemical performance. This was done via the addition of pillar molecules, namely 1,6-hexanediamine (HDA), which we hypothesize will expand the interlayer space of MoS2, modulating its geometric nanoconfinement environment. By using two different concentrations of HDA in the precursor solution (molar ratios of HDA/Mo of 1 and 0.5), the density of pillars is aimed to be controlled. |
654e54fddbd7c8b54b04ed70 | 18 | SEM characterization reveals comparable morphology with nano-MoS2 samples, with fewmicron sized secondary particles consisting of smaller flake-like primary particles with a lateral size in the range of 100 nm (Fig. ). Both MoS2-HDA-1 and MoS2-HDA-0.5 appear comparable to nano-MoS2, indicating that the addition of HDA does not significantly affect the resulting MoS2 morphology when the pH of the precursor solution remained at 1. Quantitative analysis via DLS confirms a comparable secondary particle size distribution with MoS2-HDA-0.5 showing slightly reduced sizes (Fig. ). |
654e54fddbd7c8b54b04ed70 | 19 | XRD analysis shows the formation of MoS2 with a shift of the (002) signal from 13.9° 2θ in nano-MoS2 to 9.0° 2θ for both MoS2-HDA-1 and MoS2-HDA-0.5 samples (Fig. ). This confirms that the addition of HDA pillars to the precursor solution leads to the formation of interlayer-expanded MoS2, with the (002) spacing expanding from 0.64 to 0.98 nm. While the Raman spectrum of nano-MoS2 revealed the formation of the semiconducting 2H phase, the spectra recorded for MoS2-HDA-1 and MoS2-HDA-0.5 samples show a distinctly different signature (Fig. ). The spectra show the characteristic vibrational Raman modes J1 at 153 cm -1 , J2 at 213 cm -1 , J3 at 329 cm -1 and A1g at 412 cm -1 . The spectra indicate the formation of a distorted 1T phase of MoS2 (1T' phase) upon the confinement of HDA pillars as guest species in the MoS2 interlayer space. The distorted 1T phase (1T') can be distinguished from the ideal 1T phase via the presence of the A1g mode at 412 cm -1 and the absence of the E2g mode at 258 cm -1 . We hypothesize that the 1T' phase forms as a consequence of the confined interlayer guest species (i.e., HDA pillars), which has been described for hydrothermally synthesized MoS2 materials with various organic guest species before. The HDA content of the samples was calculated from thermogravimetric analysis (TGA) under O2-containing atmosphere that led to a full burn-off of organic species (and concomitant MoS2-to-MoO3 transformation, Fig. ). The calculated compositions for samples MoS2-HDA-1 and MoS2-HDA-0.5 are (HDA)0.33MoS2 and (HDA)0.26MoS2, respectively, confirming that the variation of HDA concentration in the precursor solution of the hydrothermal process can control the final composition of pillared MoS2 to some extent. |
654e54fddbd7c8b54b04ed70 | 20 | studies that reported significant and chemical in-plane modification of MoS2, such as nitrogen-doping via sulfur-substitution after the addition of pillaring molecules during hydrothermal synthesis. Furthermore, the distribution of HDA pillars was analyzed via energy-dispersive X-ray spectroscopy (EDX) in scanning transmission electron microscopy mode (STEM), confirming a homogenous distribution of nitrogen throughout the sample (Fig. ). It should be noted that traces of nitrogen were also found in nano-MoS2 (at much lower concentration, see Table ), indicating small residues of thioacetamide and/or urea precursors. |
654e54fddbd7c8b54b04ed70 | 21 | The structural characterization of HDA-pillared MoS2 materials demonstrated that the materials exhibit an expanded interlayer spacing compared to nano-MoS2, while other structural properties like morphology, crystallite size and in-plane MoS2 structure remained similar. Furthermore, the concentration of HDA pillars is increased in MoS2-HDA-1 compared to MoS2-HDA-0.5. Consequently, the materials can be used to study the influence of (1) interlayer spacing (i.e., nanoconfinement geometry) and ( a low initial Coulombic efficiency (ICE, 44 %) is observed, which is correlated to the above described 2H-to-1T phase transformation in the plateau region around 1.1 V vs. Li + /Li during the first lithiation (Fig. ). In comparison, both MoS2-HDA-1 and MoS2-HDA-0.5 exhibit higher ICEs of 65 % and 67 %, respectively, in absence of the 2H-to-1T phase transformation plateau. Instead, they likely undergo a 1T'-to-1T transformation during the first lithiation (vide infra). The initial anodic (delithiation) capacities of nano-MoS2 and MoS2-HDA-1 are in the same range (182 and 178 mAh/g, respectively), while the delithiation capacity of MoS2-HDA-0.5 is increased to 199 mAh/g. Given that these values are normalized to the full composite mass of MoS2 and HDA, the capacities of HDA-pillared materials significantly exceed that of nano-MoS2 per transition metal (221 mAh/gMoS2 or Li1.38-MoS2-HDA-1; and 237 mAh/gMoS2 or Li1.48-MoS2-HDA-0.5). This is in line with previous reports of increased Li + storage in interlayer-expanded MoS2 materials that potentially combines pseudocapacitive (Faradaic) Li + intercalation with an additional electric double-layer capacitance inside the widened interlayer galleries and/or at the solid/liquid interface. Cyclic voltammograms of interlayer-expanded MoS2 electrodes show an increased capacitorlike, rectangular current profile and the absence of the broad redox peak around 1.7 -1.9 V vs. Li + /Li as compared to nano-MoS2 (Fig. ). Given the lower content of electrochemically inactive HDA-pillars in MoS2-HDA-0.5, its specific current signal has a slightly higher magnitude |
654e54fddbd7c8b54b04ed70 | 22 | 118 mAh/g at a rate of 5 A/g corresponds to 63 % of the initial capacity (5 th cycle at 20 mA/g) compared to 95 mAh/g for MoS2-HDA-1 (58 %). The results underline that not only nanoconfinement geometry, i.e., an expanded interlayer spacing is beneficial for increased kinetics and storage capacity, but also the nanoconfinement chemical composition, i.e., the number of pillars determines electrochemical performance. The decreased number of pillars in MoS2-HDA-0.5 not only reduces weight of (electrochemically) inactive component, but it is hypothesized that it is also beneficial for ion transport in the nanoconfined interlayer space. Galvanostatic cycling stability over 200 cycles at 0.5 A/g is shown in Fig. . While nano-MoS2 |
654e54fddbd7c8b54b04ed70 | 23 | shows a significant initial capacity decay described in the previous section 3.1.2, this initial decay is much smaller for both HDA-pillared samples. This can be explained by the absence of the 2H-to-1T transition for these samples, demonstrating the advantage of the 1T' phase formed for interlayer-expanded MoS2. When comparing MoS2-HDA-0.5 to MoS2-HDA-1, the former shows significantly improved cycling stability. This is an indication that a reduced number of pillars confined in the interlayer space may be beneficial for cycling stability. The underlying mechanism and an investigation into the optimum number of pillars will be topic of future investigation using operando techniques and post mortem cryogenic TEM analysis. |
654e54fddbd7c8b54b04ed70 | 24 | This study introduces a versatile one-pot hydrothermal synthesis strategy that enables simultaneous control of MoS2 structure over several length scales, ranging from macroscopic crystallite size to microscopic lattice parameter control. Lowering the pH of the precursor solution for hydrothermal synthesis leads to the formation of smaller crystallites. The addition of alkyldiamines like 1,6-hexanediamine (HDA) of different lengths to the same precursor solution will lead to their confinement between the forming MoS2 layers, acting as structural pillars expanding MoS2 interlayer spacing. By changing the HDA concentration, the content of structural pillars in interlayer-expanded MoS2 can be varied to some extent. The interaction between HDA pillars and MoS2 host is purely physical in absence of covalent bonding, with no significant crystallographic changes in the MoS2 in-plane structure after pillaring. |
65c9eec29138d23161ec8dcc | 0 | Every year about 90% of the fabric manufactured around the world, ends up in landfills. And only 1 % of it gets recycled. The traditional materials we use daily harm our environment and contribute to pollution and waste. Our solution to this problem was the idea of making biofabrics from alginate, a substance that is obtained from seaweed. Biofabrics represent an innovative frontier in science, blending the realms of biology and textile engineering to create sustainable and versatile materials. |
65c9eec29138d23161ec8dcc | 1 | Calcium chloride is often used as a coagulant in the production of biofabrics made from natural polymers like alginate. When we dissolved it in water and then brought it into contact with the liquid based solution, a chemical reaction called cross-linking took place. This reaction resulted in the formation of a gel-like network on top of the liquid solution, which provided structural stability and strength to the material. The gelation process helped the biofabrics maintain their shape and structure during production. |
65c9eec29138d23161ec8dcc | 2 | The period taken for drying of the biofabric also depends on how thick is the layer of the liquid solution that is casted on the canvas. Placing this cast solution in dry and sunny conditions can fasten the drying process, but too much of extra heat can cause the fabric to become brittle and lose its moisture content completely. |
67adf8e06dde43c908b6ef91 | 0 | Nanomaterials and nanoparticles (NPs), e.g. gold nanoparticles (Au NPs), express unique properties that are relevant for applications in multiple areas of research and development, e.g. in catalysis, sensing, optics, electronics or medicine. Regardless of the application considered, the full exploitation of NP properties require their controlled syntheses. Various recipes have been reported for the preparation of Au NPs and there is an increasing awareness and need for the NPs to be produced by green and sustainable methods. We recently reported a very simple synthesis of gold (Au) NPs combining a range of desirable features. The synthesis is simply achieved by adding HAuCl4 to alkaline water and mono-alcohol mixtures, and builds up on previous work using preferentially polyols such as glycerol. The method does not require the typical additives / stabilizers and/or high viscosity chemicals commonly used. Since no additives with a molar mass higher than 100 g mol -1 apart from the metal precursor is required, this approach qualifies as surfactant-free, as per a definition proposed and detailed elsewhere. This room temperature surfactant-free synthesis is surprisingly simple, can be performed using ca. 20 v.% ethanol as the source of reducing agents, i.e. a cheap, relatively low toxicity and low viscosity chemical. The method is largely detailed elsewhere and leads to size controlled NPs in the range 5-30 nm. Despite the absence of common additives, the method leads to colloidal dispersions of NPs stable for months and even years. The use of different alcohols leads to relative pros and cons summarized in Table . Our recent investigation of the effect(s) of using different alcohols for the room temperature synthesis of the surfactant-free Au NPs showed that different formation pathways are followed for different alcohols used, namely: methanol, ethanol, ethylene glycol, glycerol. The synthesis is faster in glycerol and lead to relatively small size NPs in the range 5-10 nm over a wide range of conditions (e.g. different concentrations of base and/or alcohol contents). The synthesis using glycerol is also less sensitive to the light environment. The glycerol-mediated synthesis is therefore overall more robust. However, a major drawback of the glycerol-mediated synthesis is the use of a relatively viscous solvent. The high viscosity of polyols prevent a simple and direct use of the Au NPs. Ethylene glycol is a suitable alternative to glycerol, 32 but is still a viscous solvent. Ethanol presents the main benefit to have a very low viscosity: A desirable feature to more easily process the NPs, e.g. to deposit the NPs on various surfaces to develop heterogeneous catalysts without the need for intensive and wastegenerating steps typically requiring strong acids when polyols are used. Ethanolmediated syntheses still lead to rather small ca. 10 nm Au NPs stable for years. Nevertheless, the overall kinetics of NP formation are slower and the experimental window where the synthesis can be considered successful is narrower than in the case of polyols-mediated syntheses. It is here considered that desirable features for a synthesis are to lead to small size NPs, in a relatively timely approach, through a robust process and ideally in low viscosity solvents to facilitate the use and processing of the Au NPs. As presented in Table , the use of polyols does not fulfil the requirement of low viscosity but the use of the low viscosity ethanol is more sensitive to the variations of various parameters. Note that the use of methanol is here excluded due to the high toxicity of methanol and the fact that methanol-mediated syntheses lead to larger ca. 10-30 nm NPs, whether methanol is used as sole source of reducing agent or mixed with ethanol. Here, the increased understanding of the similitudes and differences between the use of glycerol and ethanol established in our previous work leads us to investigate and develop new synthetic approaches and explore the possible benefits of using [monoalcohol + polyols] mixtures as the source of reducing agents, with a focus on high mono-alcohols content. To the best of our knowledge, the strategy to use [ethanol+glycerol] mixtures to optimize surfactant-free syntheses of Au NPs has not been reported yet. It is established that small amount of glycerol around 1-2 v.% are beneficial to achieved size control in syntheses performed with a total amount of alcohol around 20 v.%. The results pave the way to greener syntheses of Au NPs and more sustainable nanotechnologies. |
67adf8e06dde43c908b6ef91 | 1 | The alcohol-mediated synthesis has been described elsewhere. The synthesis of the Au NPs is performed by adding HAuCl4 to a mixture of water, NaOH and an alcohol. The reaction proceeds at room temperature without the need for extra chemicals to lead to colloidal dispersions of Au NPs stable for months. The synthesis is scalable. The synthesis is best performed using high purity water. The synthesis is best performed by adding HAuCl4 last from a relatively concentrated solution of HAuCl4 (at 50 mM in mQ water in this work, stored in the fridge). The base is best used from stock solution stored in plastic container (at 50 mM in mQ water in this work, stored at room temperature). The synthesis is best performed under controlled light environment. Although the synthesis does not require stirring more reproducible results are obtained using magnetic stirring (PTFE cylindrical stirrer bar, 8 x 3 mm, VWR, 442-4520) and the magnet used were cleaned with aqua regia (4:1, v:v, HCl:HNO3; to be handled and disposed of with care following the related safety procedures in place in the laboratory). |
67adf8e06dde43c908b6ef91 | 2 | The reaction was performed in disposable 1 cm wide rectangular polystyrene (PS) UVvis cuvettes (340-900 nm, VWR, 634-0675) containing a magnet (unless otherwise specified), de-dusted with a flow of compressed air prior to the addition of the chemicals. The chemicals were added from the stock solutions (NaOH: 50 mM in mQ, HAuCl4: 50 mM in mQ, ethanol as received, methanol as received, glycerol from a 60 v.% stock solution in mQ to facilitate the handling of this otherwise viscous chemical) in the order water < base < alcohol(s) < HAuCl4. HAuCl4 was added last, under stirring in a photo-box (Puluz LED portable Photo Studio, PU5060EU, 60 cm x 60 cm x 60 cm, 60 W) to control the light environment. The samples were capped with dedicated stoppers and left to react for 2 hours in the photo-box with stirring. Alternatively the samples were subjected to a 365 nm light (OFK-8000 OptimaxTM Multi-LiteTM inspection kit, Spectroline), illuminated from the top and without capping, for 2 hours, as described elsewhere, or kept under dark conditions for 2 hours (capped) using aluminum foil. The samples were then left for a day at room temperature and ambient light without stirring. The samples were then stored at room temperature in a drawer. |
67adf8e06dde43c908b6ef91 | 3 | The final concentrations before taking into account volume contraction were typically 80 v.% water and 20 v.% alcohols in total, unless otherwise specified. The alcohols were ethanol (alternatively methanol) and/or glycerol in the indicated v.% (expressed before taking into account volume contraction that can happen in alcohol-water mixture such as ethanol-water mixtures due to hydrogen bonding). The other parameters were fixed as per optimal conditions previously established for ethanolmediated syntheses, i.e. 2 mM NaOH and 0.5 mM HAuCl4 for a total volume of 2 mL (before taking into account volume contraction). A NaOH/Au molar ratio of 4 was previously established as an optimal value in light of the equation of the reduction process. An overview and summary of the samples considered and their characteristics is provided in Table . |
67adf8e06dde43c908b6ef91 | 4 | A focus for this study is given to the properties retrieved from UV-vis measurements because UV-vis spectra are very informative when it comes to plasmonic Au NPs. The characterization is completed by scanning transmission electron microscopy (STEM). To develop more sustainable research, it will not be realistic to use all characterization methods reported for Au NPs for all the samples obtained here with a relatively high throughput. Further characterization of the surfactant-free fcc Au NPs by XPS, XRD, X-ray total scattering with PDF analysis or zeta-potential measurements can be found elsewhere. |
67adf8e06dde43c908b6ef91 | 5 | Metrics. The Mie theory correlates various variables to the optical properties of Au NPs. UV-vis spectra of Au NPs results from a complex interplay between the properties of gold, nanoscale effects (such as size and shape) and interaction with the solvents and/or interactions between NPs and precursor. Due to their plasmonic properties (surface plasmon resonance, spr), several relevant parameters descriptive of the Au NPs can be retrieved: λspr, 42 A400, Aspr/A450, 42 A380/A800, all defined, detailed and summarized in Table in Supplementary Information (SI). Note that the relationships indicated are only true for certain conditions that are not always met in the experiments below (e.g. for low yields experiments or for too large NPs) but are convenient indicators of the properties of the Au NPs. The good correlation between λspr values and size retrieved by STEM is detailed elsewhere and documented in Figure . |
67adf8e06dde43c908b6ef91 | 6 | For kinetics studies, Amax corresponds to the maximum of absorption and is in most cases for the dataset here Amax=Aspr. Although at the initial stages of the reaction no clear spr signal is observed Amax is representative of the overall absorption since the spectra are almost straight lines. The value √𝐴 𝑚𝑎𝑥 3 when Amax=Aspr is proportional √𝑁 Samples. UV-vis spectra were acquired using a Shimadzu UV-1800 UV/Visible scanning spectrophotometer typically between 290 and 800 nm. As baseline, a solution with the corresponding amount of mQ water and the corresponding amount of ethanol or methanol and/or glycerol to what was used in the sample were preferred (no base, no HAuCl4). All samples were measured the day after synthesis (although it is important to note that the samples are stable for months, see Section 6 in SI). The samples are prepared directly in UV-vis cuvettes that were also used for those measurements. |
67adf8e06dde43c908b6ef91 | 7 | Kinetics. For kinetics study the samples were measured for 2 hours in such a way that the sample in a UV-vis cuvettes as described above was directly placed in a UVvis Go Direct® SpectroVis® Plus spectrophotometer with a fixed scan range of 380 nm to 950 nm. It takes only ca. 1-2 seconds to acquire the whole range. However, measurements below 400 nm and above 800 nm proved to be unreliable. Due to the relatively low resolution of the measurement, the data were then filtered prior to analysis using the Signal Processing Feature of the Origins software and the Adjacent Averaging method with a Point of Windows equal to 18, see Figure . The background was a mixture of alcohol and water together with the base with the same alcohol, alcohol content, base and base concentration as the sample considered. The background acquisition and the actual measurements were under stirring in a photobox (Puluz LED portable Photo Studio, PU5060EU, 60 cm x 60 cm x 60 cm, 60 W). HAuCl4 was added last. Spectra were recorded every 30 seconds with a first spectra 15 seconds after HAuCl4 was added. |
67adf8e06dde43c908b6ef91 | 8 | A FEI Talos F200X operated at 200 kV, equipped with a high-angle annular dark-field (HAADF) detector was used for electron microscopy characterization. The asprepared colloidal dispersions were dropped on copper TEM grids (Sigma-Aldrich), placed on an absorbing filter paper. The solvent was left to evaporate. The samples were characterized by imaging at least 3 randomly selected areas at 3 different magnifications. Typically, at least 100 NPs were used to estimate the diameter of the NPs (N values reported below correspond to the number of NPs counted). The ImagJ software was used for size analysis. |
67adf8e06dde43c908b6ef91 | 9 | Upon adding HAuCl4 to an alkaline solution of 80 v.% water and 20 v.% ethanol, the solution turns grey, blue, purple and finally red, indicative of the formation of Au NPs with a characteristic spr around 520 nm in UV-vis characterization. The synthesis proceeds in a similar way when methanol, ethylene glycol or glycerol are used, with however different kinetics, which is attributed to the different redox and physicochemical properties of the different alcohols. The alkaline conditions favor the formation of alkoxides from the alcohols that play the role of reducing agent for the room temperature process . The mechanisms behind the NP formation are detailed in length elsewhere, and further discussed below. A first finding reported here is that Au NPs are also obtained when mixtures of [ethanol+glycerol] are used, as illustrated in Figure . |
67adf8e06dde43c908b6ef91 | 10 | Having established that [mono-alcohol + polyol] mixtures are suitable to obtain surfactant-free Au NPs, see Figure , we further explore the effects of using mixtures of ethanol and glycerol, with a focus on low glycerol contents. The hypothesis is that the formation of the Au NPs is sensitive to the conditions under which the very initial steps of the synthesis take place. This hypothesis is motivated by the findings from our previous study focusing on different light environment, where a hybrid approach of light controlled environment and dark conditions show that the light conditions had a strong influence mainly at the beginning of the synthesis. In other words, the use of lights with low wavelengths to obtain small size NPs was mainly key in the first minutes of the reaction. It is here hypothesized that small amount of glycerol might help to achieve size control towards smaller NPs by playing a role in the initial steps of the synthesis by favoring a faster formation of the NPs. |
67adf8e06dde43c908b6ef91 | 11 | When only ethanol is used as the source of reducing agent, and when 20 v.% ethanol is used, the NPs show a λspr value around 524 nm which correspond to 13.8 ± 3.9 nm NPs. As the amount of ethanol decreases to 19 or 18 v.% the λspr values increase to 524 nm and 526 nm respectively, corresponding to 12.1 ± 4.5 nm and 16.5 ± 4.4 nm, respectively, see Figure . The trend that the NP size increases and/or the size distribution increases when the amount of ethanol decreases is confirmed when 10 v.% of ethanol is used, for which the λspr value increases to 558 nm corresponding to 18.6 ± 12.1 nm NPs. This corresponds to a case where there is not enough reducing agent and the NP synthesis is less controlled. The poorer control over the synthesis is even more pronounced for 2 and 1 v.% ethanol where there is not enough alcohol to perform the reduction of Au III , as indicated by relatively featureless UV-vis spectra with an overall low absorption, e.g. very low A400 values compared to the other samples, see Figure and S3. As a result, no STEM characterization was performed on those samples obtained with 1 or 2 v.% ethanol. |
67adf8e06dde43c908b6ef91 | 12 | Having established that (i) ca. 20 v.% ethanol is suitable to obtain relatively small size Au NPs in a low viscosity solvent and that (ii) low amount of glycerol are suitable to induce the Au NP formation, we now turn to investigate mixtures of [ethanol+glycerol] in different relative ratio but keeping the total amount of alcohol at 20 v.%, see Figure A more detailed screening of the influence of low amount of glycerol is proposed in Figure together with repeated experiments that confirm the trend observed, see Figure -S6 and S8. These results indicate that the NP size decreases as the amount of glycerol increases to ca. 5 v.%, while the viscosity of the reaction mixture increases as the glycerol content increases. 17 It is therefore concluded that small amount of polyol in the range 1-5 v.% can help controlling the NPs size towards smaller size as the amount of glycerol increases. Importantly, this small amount of viscous polyol does not change much the overall viscosity of the samples still obtained in relatively low viscosity media containing mainly water and ethanol. |
67adf8e06dde43c908b6ef91 | 13 | To understand better the effect of such a small amount of glycerol on the control of the NP size, we performed kinetics study on the formation of the Au NPs. eye) when glycerol is used compared to the use of ethanol only. When [ethanol+glycerol] mixtures are used, and even with small amount of glycerol such as 2 v.% glycerol with 18 v.% ethanol, an overall faster appearance of a red colloidal dispersion can be observed with the naked eye compared to the case where ethanol only is used. The overall formation pathway considering various metrics is confirmed to be faster for [ethanol+glycerol] compared to the case of 20 v.% ethanol without glycerol, see Figure . (Amax) 1/3 is proportional to R(N) where N is the number of particles and R their radius. The general features of the time-resolved data resembles those reported by others, for example on study considering the citrate-mediated synthesis of Au NPs. The plot of (Aspr) 1/3 as a function of time can be dived in to three phases: (I) An initial region where (Aspr) 1/3 increases in a superlinear fashion; (II) an intermediate region where (Aspr) 1/3 increases linearly; and (III) a final region where (Aspr) 1/3 exponentially increases and levels off to its final value. For clarity, those phases (I-III) are indicatively illustrated in Figure . Those phases have been attributed to: (I) NP grow via aggregation: Coupled to SAXS measurements, it was established in previous work that the number density of the NP decreases, while the particle size increases in phase I; (II) no further aggregation occurs, and the particles grow linearly in time via a surface growth reaction; and (III) NP size grows via an autocatalytic growth stage and levels up to its final value. Here, in agreement with previous results, the formation of the Au NP using 20 v.% glycerol (no ethanol) is rather fast in the initial steps of the synthesis, followed by a slow and continuous growth, which is attributed to the redox properties of glycerol and its high viscosity, respectively. Also in agreement with previous report for the ethanol mediated synthesis (no glycerol), the phases I and II are almost merged which is attributed to the fast formation of the first NPs seeds in absence of stabilizers in the low viscosity solvent together with a slow autocatalytic growth due to the room temperature process. In the case of [ethanol+glycerol], the same phases are observed but the reaction proceeds faster than when only ethanol is used. Importantly, the time normalized data from 20 v.% ethanol and [ethanol+glycerol] with 18 v.% ethanol and 2 v.% glycerol follow the same general dynamic illustrated in Figure . This suggests that glycerol mainly plays a role to accelerate the initial steps of the synthesis. Since all conditions lead to approximatively the same final (Aspr) 1/3 values , and assuming that using different alcohols does not change too much the properties of the medium surrounding the Au NPs and hence their optical properties, and given the larger size of the resulting NPs when ethanol only is used, as detailed in Figure , it can be concluded that the number of final NPs obtained in presence of glycerol is higher than without. Therefore, the difference observed can be interpreted as follow: small amount of glycerol favors a faster formation of the initial seeds. Assuming that the classical nucleation theory applies here, this fast formation of the seeds followed by a slow growth leads to the observed smaller and more monodispersed NPs compared to the case where only ethanol is used (where the use of ethanol only is characterized by a slower nucleation process). The fast formation is observed with amount of glycerol as low as 2 v.%. |
67adf8e06dde43c908b6ef91 | 14 | Note than the different overall shape of the spectra reported in Figure , differ from those reported in our previous studies. This is attributed to the fact that in our previous study a light was directly shined on the sample during measurement and the samples were not left exposed to the light of a photobox. Since the samples are here placed in the UV-vis spectrophotometer, it can be anticipated that less light is received by the samples, which leads to slower formation of the NPs in particular when ethanol is used. The samples using glycerol or [ethanol+glycerol] are less affected by this change in experimental setup. See details in SI on the influence of light environment, e.g. Figures . This observation stresses the benefit of small amount of glycerol to improve the robustness of the synthesis, for instance to light conditions. These results stress further that controlling the initial stages of the synthesis are key to achieve finer size control. |
67adf8e06dde43c908b6ef91 | 15 | Despite the absence of added stabilizers, the Au NPs colloids are stable for weeks as detailed in Figures and, where the UV-vis spectra of the as-prepared NPs and the UV-vis of the same batch after 3 months of storage at room temperature in a drawer are almost the same (the measurements were performed without homogenization of the sample). It is worth pointing that surfactant-free NPs prepared with 20-30 v.% are even stable for years as detailed elsewhere. While the stabilization mechanism of the Au NPs is still to be fully established, electrostatic interactions seem key and the contribution of species such as acetaldehyde (oxidation product of ethanol) cannot be ruled out. Figure . UV-vis spectra for a sample prepared with [18 v.% Et + 2 v.% Gly], 24 hours after the beginning of the synthesis (D1) or after 3 months (M3), as indicated. The samples were kept at room temperature in a drawer. Et: Ethanol; Gly: Glycerol. |
67adf8e06dde43c908b6ef91 | 16 | Mixtures of alcohols such as [ethanol+glycerol] mixtures can be used as source of reducing agents for the room temperature synthesis of stable surfactant-free Au NPs in alkaline aqueous media. Based on evidence from UV-vis characterization and STEM microscopy, a benefit of this approach is that even small amount of glycerol around 2 v.% for a total amount of alcohol around 20 v.% improves the size control over the Au NPs towards lower sizes by favoring a faster nucleation. The overall strategies presented here could gain to be extended to alternative polyols such as ethylene glycol and mono-alcohol such as methanol. This strategy opens opportunity for the development of more robust and yet simple more sustainable and tractable synthetic strategies of size controlled Au NPs. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.