id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
64e4bfbf3fdae147fa9b3948 | 7 | To demonstrate the formation of HEA alloys, we performed X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and ultraviolet photoelectron spectroscopy (UPS) analyses to determine the crystal lattice constant (XRD), the binding energy of each atom (XPS), and the work function (UPS). This allowed us to analyze structure changes occurring due to the formation of HEA. |
64e4bfbf3fdae147fa9b3948 | 8 | Spectroscopic ellipsometry was used to determine the permittivity of metals. A 200-nm-thick film was deposited on the glass substrate to eliminate the reflection effect from the substrate. To measure the permittivity over a wide bandwidth, UV-visible spectroscopic ellipsometry (from 200 nm to 1000 nm) and mid-infrared spectroscopic ellipsometry (1700 nm to 20000 nm) were used. After determining the optimal permittivity for each measurement result using the Drude-Lorenz model, the results of both UV-VIS and MIR measurements were concatenated to determine the optimal parameters using the single Drude-Lorenz model and the least-squares method. The following equation defines the used the Drude-Lorenz model for analysis in this study: |
64e4bfbf3fdae147fa9b3948 | 9 | where ε(∞) is the real part permittivity at high frequency, which is the sum of ε(∞) from both Drude and Lorenz models, ω P,D is the plasma frequency of the Drude model, Γ D is the dumping constant of the Drude model, ω P,L, j is the plasma frequency of the Lorenz model, ω 0, j is the central angular frequency in the Lorenz model, Γ L, j is the dumping constant of the Lorenz model, j is the indicator number, which is connected to the number of oscillation, 1 or 2 is used in this study. |
64e4bfbf3fdae147fa9b3948 | 10 | FDTD calculations were performed using the Ansys-Lumerical FDTD Solutions package. Palik's parameters were used for the thin gold film in the lower layer of the MIM structure. Calculations were performed using the experimentally obtained permittivity for the upper nanodisk structure. The nanodisc diameter was swept from 500 nm to 2000 nm in 100 nm increments and recalculated in 10 nm increments, especially in regions of lower reflectivity. The permittivity of Palik was used for SiO 2 , and calculations were performed for the film thicknesses of 50, 100, 150, 200, 300, and 400 nm. The thickness of the Au nanostructures was set to 50 nm under the same conditions as those used in the experiment. |
64e4bfbf3fdae147fa9b3948 | 11 | Figure shows the results of the XRD crystallographic analysis of metal films. The HEA film exhibited a single-phase FCC-type structure, which is distinct from the individual metal XRD patterns obtained at the same conditions. The lattice constants were calculated from the diffraction peaks of the (111) plane, which is the central peak of the fcc structure, and the results are shown in Table . The values of the crystal lattice constants showed excellent agreement with the average values of the five metals, and the crystal lattice structure was close to the values of palladium and platinum. This indicates that the elements were well mixed and formed a single-phase alloy. |
64e4bfbf3fdae147fa9b3948 | 12 | The XPS and UPS spectra are presented in Figures and, respectively. The broadband XPS measurements showed that the binding energies derived from almost all the atomic orbitals of the constituent atoms could be observed in the HEA alloys (Table ). The spectra with an adjusted baseline showed very similar results to the average of the spectra of the five elements (Fig. ). However, shifts or disappearances of some peaks were observed. Therefore, a narrow-band analysis was performed to detail the observed formation of the alloy. It was found that almost all peaks shifted to the lower energies. In general, the shift to lower energies is due to the acceptance of electrons from outside . In other words, the pool of free electrons in the alloy mainly give electrons as bound electrons during the alloy formation. From the UPS measurements, the work functions were approximately 4.7 eV for the HEA and the constituent metal carriers (Fig. ). The values are close to those of the constituent elements. These results suggest that HEAs were formed in the present Au Ag Cu Pd Pt system using the sputtering method, and the nanostructures' optical properties should reflect these HEAs' properties. In the following sections, we summarise the experimental and FDTD simulation results for the resonance properties of the actual metasurfaces made out of noble HEA. |
64e4bfbf3fdae147fa9b3948 | 13 | Figure shows the experimental and FDTD simulated reflection spectra of MIM metasurfaces with different metallic elements for the SiO 2 dielectric spacer of 300 nm. All the spectral data with different SiO 2 thicknesses are summarised in Figures and for the experimental and simulated results, respectively. HEA alloys are characterized by forming a low reflection condition over a wide range of wavelengths and relatively narrow resonance line widths. In contrast, Au, Ag, and Cu metasurfaces have narrow resonance line widths but low reflection only over a limited range of wavelengths. For example, SiO 2 films of 100 nm and 200 nm thicknesses show lower reflection over a broader range than 300 nm thickness, but the range is narrower than that of HEA. In contrast, Pt and Pd have a more comprehensive range of low reflections, but their resonance spectra are broad. |
64e4bfbf3fdae147fa9b3948 | 14 | Figure shows a graphical representation of the reflectance minimum, its resonance linewidth in wavenumber units, and resonance center wavenumber extracted from the experimental spectral data. In this analysis, the resonance linewidth was the difference between the maxima and minima of the first derivative of the resonance spectrum. Although the mathematical formula shows that the difference between the maxima and minima of the derivative is slightly larger than the half-width in a bell-shaped function system, this has a great advantage for the analysis since it can be performed with uniform accuracy, including the asymmetric bell-shaped experimental results. The projections of the 3-D plot (Fig. ) are plots of the relationships between the reflectance and resonance linewidth, resonance center wavenumber, resonance linewidth, and reflectance vs. resonance center wavenumber, respectively. The results show that the relationship between the resonance center wavenumber and resonance linewidth is narrower for all the metals at the lower wavenumbers. Furthermore, the resonance line widths of Au, Ag, and Cu narrowed with decreasing reflectivity, while those of HEA, Pt, and Pd widened at low reflectivity conditions. For the SiO 2 films of 100 nm and 200 nm thickness, the same tendency to narrowing of the resonance linewidth was observed as for to Drude-Lorenz model ( j = 1, 2). For Pt and HEA, data can be well fitted with a single Lorenz model ( j = 1). |
64e4bfbf3fdae147fa9b3948 | 15 | Au, Ag, and Cu, while for SiO 2 films of 300 nm and thicker, the resonance linewidth shows the same tendency as for Pt and Pd. The resonance linewidths of Pt and Pd are up to twice as large as those of Au, Ag, and Cu, whereas those of HEA are only 1.5 times larger than Au, Ag, and Cu, which is expected to reduce the absorption loss of Pd and Pt. To discuss the optical behavior of these metals in detail, their permittivities were determined over a wide range of wavelengths from UV-visible to mid-infrared by spectroscopic ellipsometry (Table ). |
64e4bfbf3fdae147fa9b3948 | 16 | Figure shows the permittivity of the six metals from UV-visible to mid-infrared. Although the data in the 1000-1700 nm range are missing, the Drude-Lorenz model obtained a smooth and continous permittivity over the entire measurement range from 200 to 2 × 10 4 nm by fitting via the least-squares method over the entire measurement range. With Lorenz model over the full measurement range, we obtained permittivity from 200 nm to 2 × 10 4 nm over this wide spectral window. Figure also shows plots of the first three terms (Eqn. 1) of the Drude-Lorenz equation used in the analysis, and Lorenz terms separately. Note that ε ∞ is a constant term that should be included in the Drude and Lorenz terms independently. However, for analytical reasons, it is difficult to separate them; therefore, it was treated as a single constant. The results show that the permittivity of HEA is closer to those of Pd and Pt than those of Au, Ag, and Cu. However, as shown in the imaginary part of the Lorentz model, the loss-related contributions are smaller than those of Pt and Pd. Table lists parameters obtained from the analysis using the Drude-Lorentz model for the experimentally measured permittivity and extracted via fit. As mentioned above, ε ∞ is We confirmed that fixing these parameters to, for example, the ideal metal permittivity did not significantly affect the analytical results. |
64e4bfbf3fdae147fa9b3948 | 17 | In general, the plasma frequency is related to carrier density. When the effective mass of electrons is constant, a larger value corresponds to the higher carrier density. HEA had the second-highest free electron density after Au, Ag, and Cu and the highest bound electron density among the six metals. Abundant free electron density is necessary to induce a considerable plasmon resonance, and the results suggest that HEA alloys are more "plasmonic" as compared with Pd and Pt. The increased density of bound electrons indicates that some abundant free electrons in Au, Ag, and Cu were converted to the bound electrons. This agrees with the XPS measurements, which show that almost all the binding energies are shifted to the lower values, suggesting that the electrons are accepted. |
64e4bfbf3fdae147fa9b3948 | 18 | The damping coefficient is the largest for HEA for any of the metals, making the material less prone to a high quality factor Q resonance as compared with a single metal. The damping constant is also closely related to the absorption loss and is strongly affected by the crystalline state of the material, e.g. strucutal defects such as grain boundaries, dislocations, stacking faults. The very sharp peaks obtained by X-ray crystal structure analysis indicate that the variation in the size of the grain boundaries is very well suppressed. Hence, a highly homogeneous structure is formed. However, considering the lattice constant of the crystals, it is difficult to conclude that a single phase is included in the HEA. Considering the lattice constants of the crystals, the atomic radius of a single atom of 140 pm, which is assumed from the lattice constant of the HEA, a formation of a single phase HEA is considered. HEA has a larger lattice constant as compared with Cu, Pt, and Pd. It is assumed that the repulsion between nuclei and elongation of bonds due to the decrease in the inter-atomic distance causes a distorted geometry. This distortion between the crystal lattice and atomic radii is thought to be responsible for the increased damping of the optical response of the electron. At the plasmon resonance, the balance between the oscillation of free electrons and the loss and damping caused by various factors are at the interplay. In the present system, we could improve absorption over a wide range of wavelengths by coincidence. It is important to continue to analyze the permittivity of different HEA systems, find a methodology to precisely control these Drude and Lorenz parameters, and to establish a unified method for material search to find a material with a resonance line width comparable to that of Au Ag Cu, but with enhanced absorption. Having obtained the complex permittivity of the newly engineered metallic material, the resonance properties were calculated using FDTD simulations. Simultaneously, the scattering cross-sections (σ sca )and absorption cross-sections (σ abs ) can also be quantitatively bluedetermined using the proposed method. |
64e4bfbf3fdae147fa9b3948 | 19 | Figure shows the reflectance R, ratios of cross sections σ abs /σ sca , σ abs , and σ sca plots for the data with the lowest reflectance condition from the comprehensive FDTD simulation results. For Au, Ag, and Cu, the peaks of the σ abs and σ sca are coincident, and they are equal to each other at the reflectance minimum (σ abs /σ sca ∼ 1). In contrast, for HEA, Pt, and Pd, the σ abs and σ sca peaks appeared at separate wavenumbers, indicating that the σ abs is much larger than the σ sca at the resonance wavelength of low reflection where absorption is dominant. To examine this in more detail, the ratio of the σ abs and σ sca (σ abs /σ sca ) at each resonant wavelength is plotted against the reflectance (Fig. ). In the case of Au, Ag, and Cu, the minimum reflection appeared at σ abs /σ sca =1 in all simulation results for different SiO 2 thickness. Therefore, all data, including Au, Ag, Cu, and different SiO 2 thicknesses, can be plotted as a single Gaussian function, which is presented by the dotted gray line. HEA and Pt also follow this relationship when SiO 2 thickness is from 50 to 150 nm. However, when a thicker SiO 2 film was used, σ abs /σ sca > 1 was observed. At this condition, the absorption becomes dominant and corresponds well with the experimental reflection spectra. In the case of Pd, the shift value of σ abs /σ sca is the largest in all metals. This indicates the strong absorption of metals decreases the Q-factor of the MIM resonance. The shifted σ abs /σ sca value is similar to that of 50 nm Cr added as an adhesion layer, shown in our previous work . These results indicates that HEA with Au, Ag, Cu, Pd, and Pt realizes the "true perfect absorption" and "true perfect thermal emission." |
64e4bfbf3fdae147fa9b3948 | 20 | This study demonstrated formation of noble HEA and determined its optical permittivity with plasmon resonances mainly focused on the mid-infrared spectral range where such HEA could be used for realisation of perfect absorber and emitter. Such noble HEA was reported as having efficient catalytic properties . Therefore, this can be used as a plasmon-based efficient photo-catalyst and artificial photosynthesis. |
6581bcd09138d231610be057 | 0 | Electrochemical double layer capacitors (EDLCs) are a class of supercapacitor energy storage devices with superior power performances and longer cycle lives than batteries . The most 30 commonly studied and cheapest EDLCs contain carbon electrodes formed from disordered, graphene-like sheets that form a porous network with a distribution of pore sizes . In order to improve the energy densities of these devices closer towards those of batteries, many studies focus on varying the structure of the nanoporous carbon electrodes (4) so as to tune the carbon pore size (as measured by gas sorption). While early studies of titanium carbide-derived carbons (TiC-CDCs) with different pore sizes , as well as studies of activated carbons , reported a maximum capacitance as the carbon pore sizes were decreased towards the size of desolvated electrolyte ions , more recent studies have generated contradictory results. For example, a lack of correlation between capacitance and pore size across a collection of porous carbons with pore sizes ranging from 0.7 to 15 nm in a standard NEt4BF4 in acetonitrile (ACN) electrolyte, including 22 microporous 40 activated carbons from different precursors, TiC-CDCs, and 6 mesoporous carbons was reported . |
6581bcd09138d231610be057 | 1 | Further studies also observed a lack of correlation between the pore size and capacitance , with only a modest increase of 17% observed for pores around 0.75 nm from computational investigations . These contrasting results have led to unclear design principles for improving EDLC electrodes and suggest that an additional unknown structural variable impacts the capacitance. There is thus an urgent need for new investigations that unpick the connection (if any) between the structure of nanoporous carbons and their capacitance. |
6581bcd09138d231610be057 | 2 | Over the past decade, solid-state nuclear magnetic resonance (NMR) spectroscopy has emerged as a new probe of both the local chemical structure of EDLC electrodes, as well as their charge storage 50 mechanisms . NMR spectra of carbons saturated with electrolyte reveal separate resonances for "in-pore" ions (adsorbed in the carbon nanopores) and "ex-pore" ions (located outside the carbon pore network). The in-pore resonance appears at lower chemical shifts than the corresponding neat electrolyte due to the "ring currents" generated by the circulation of delocalized π-electrons in the aromatic carbon rings in the applied magnetic field . This effect is quantified by the Δδ value, 55 which is defined as: ∆𝛿 (𝑝𝑝𝑚) = 𝛿 )*+,-./ -𝛿 */12 /3/42.-352/ Where 𝛿 */12 /3/42.-352/ is the chemical shift of free electrolyte and 𝛿 )*+,-./ is the chemical shift of "in-pore" resonance. |
6581bcd09138d231610be057 | 3 | The magnitude of the Δδ value is a measure of the strength of the ring current effect and is therefore a powerful probe of the local structure of the nanoporous carbon, and in particular the "ordered domain size" , i.e., the average size of the graphene-like fragments that form the carbon pore walls . Our previous study showed that carbons prepared at lower synthesis temperatures have Δδ values of smaller magnitude due to their smaller ordered domain sizes and more disordered local 65 structures . The carbon pore size also impacts the magnitude of Δδ, though to a lesser extent, with carbons with smaller pore sizes giving rise to Δδ values of larger magnitude . Our previously developed lattice simulation method enables the extraction of ordered domain sizes from Δδ values by accounting for pore size effects, as well as effects from preferential adsorption of the probe molecules on the carbon surfaces . Summarizing, the Δδ value measured by NMR 70 spectroscopy is emerging as a new probe of structural disorder in nanoporous carbon structures , offering a new opportunity to connect electrode structure with capacitive performance. |
6581bcd09138d231610be057 | 4 | In this work we carry out NMR spectroscopy and capacitance measurements on a series of 20 nanoporous carbons to explore how carbon structure impacts capacitance. We find that more 75 disordered carbons have higher capacitances due to the more efficient storage of electrolyte ions in the carbon pores . Our findings go against the idea that pore size is the main structural factor in determining capacitance and will stimulate new research on the design and synthesis of improved electrodes for EDLCs. |
6581bcd09138d231610be057 | 5 | Ten commercially available nanoporous carbons from a range of suppliers were initially selected for our study (see Methods). Analysis of gas sorption data revealed that five of the studied carbons had very similar pore size distributions (Fig. , fig. ) and similar specific BET define surface areas (1694-1821 m 2 /g, table ), so these were selected to test whether factors beyond porosity 85 impact capacitance. Despite the similarity in pore size distributions, these five carbons exhibit significantly different capacitance values with a standard NEt4BF4 in acetonitrile (1 M) electrolyte, with values ranging from 138 F/g to 83 F/g (Fig. , fig. , fig. ). Measurements at different charging rates (fig. ), as well as the measurements with ionic liquid electrolytes (fig. ), revealed similar results, with ACS-PC and SC-1800 displaying the highest capacitances and PW-400 90 displaying the smallest capacitance. Together, these findings suggest that factors beyond pore size and specific surface area impact capacitance. |
6581bcd09138d231610be057 | 6 | These materials demonstrate similar capacitance values, with a range of only ~9 F/g (Fig. ), suggesting a minor impact of porosity on capacitance for these materials. Combining these results together with a wider series of nanoporous carbon samples including thermally annealed samples (see below, and Methods), no obvious correlation is observed between capacitance and average pore size (Fig. ), nor BET surface area (Fig. ), suggesting that structural features other than pore 100 size and surface area may govern the capacitance. Finally, oxygen content from X-ray photoelectron spectroscopy (XPS) measurements also did not show a clear correlation with capacitance (fig. ). |
6581bcd09138d231610be057 | 7 | As the porosity data fails to explain the wide variations in capacitance of the studied carbon series, we developed an NMR spectroscopy assay to probe local structural order and its impact on 115 capacitance (Fig. ). MAS (magic angle spinning) NMR spectra of the studied electrolyte-saturated carbons contain at least two resonances as expected (Fig. ). The left-hand resonances, with a similar chemical shift to neat electrolyte at -149 ppm, are assigned to "ex-pore" anions, while the right-hand resonances are assigned to "in-pore" anions, as in previous work . Notably, the NMR spectra of the electrolyte-saturated carbons show significant differences, particularly in 120 terms of the Δδ values (see fig. and S8 for discussion of linewidth and intensity effects). We initially hypothesized that the ion adsorption capacities measured by NMR might correlate with capacitance, but a clear correlation was not found (fig. ). |
6581bcd09138d231610be057 | 8 | Note that the measured Δδ values are consistent irrespective of the choice of the nucleus probed (fig. ; i.e., they represent nucleus independent chemical shifts. The Δδ values are therefore indicative of the different structures of the carbons, rather than any specific interactions between the carbons and the studied ions. The correlation of capacitance with Δδ was also observed for ionic liquid 130 electrolytes (fig. ). The observed correlations are striking given that the carbons were selected from six different independent suppliers and are thus very likely synthesized by a range of different processes/conditions. Since previous studies showed that Δδ values are dominated by domain size effects (rather than pore size effects) for predominantly microporous carbons , our results further suggest that carbons with smaller ordered domains give rise to higher capacitances. To test this hypothesis ACS-PC, the most disordered nanoporous carbons in our series, was thermally annealed at a range of temperatures in argon (see Methods). We hypothesized that thermal annealing would increase structural order in the carbon , leading to carbons with larger magnitude Dd values and smaller capacitances. Importantly, gas sorption results were first performed and confirmed that there were minimal changes in the carbon pore structure upon thermal 150 annealing (fig. ), while XPS measurements revealed a decrease in oxygen content upon annealing (fig. ). For higher thermal annealing temperatures, Dd values increase in magnitude as hypothesized, consistent with the formation of carbons with larger ordered domains (Fig. ). |
6581bcd09138d231610be057 | 9 | Increases in structural order observed by NMR again led to decreases in the capacitance, although less significantly than for ACS-PC, suggesting a limit of lowering the capacitance by increasing the ordered domain sizes for this connected pore system. The consistent observations on annealed carbons further support the hypothesis that it is the local structural disorder that governs the 160 capacitance, rather than pore size (Fig. vs. Fig. ), with more disordered carbons with smaller ordered domains having higher capacitance. |
6581bcd09138d231610be057 | 10 | To further test whether the structural disorder correlates with capacitance, a previously reported NMR simulation approach was applied to predict the correlation length associated with the size of 165 the ordered aromatic carbon domains (Fig. , see Methods) . This simulation approach accounts for the experimental pore size distribution and the strength of the ion-carbon interactions; the NMR chemical shifts are then modelled using model polyaromatic hydrocarbon fragments separated by distances governed by the pore size distribution (see Methods). These simulations support the idea that carbons with smaller calculated ordered domain sizes generally have higher 170 capacitances (Fig. ). X-ray pair distribution function (PDF) patterns of the studied carbons supported the findings from NMR spectroscopy (Fig. ). Comparing the X-ray PDF patterns of SC-1800 (small Δδ value, high capacitance) and PW-400 (large Δδ value, low capacitance), we find that the SC-1800 has a more rapid decay of the pairwise C-C correlations , consistent with the smaller domain sizes and/or more the disordered local structure of SC-1800. X-ray PDF results for the other carbons yielded similar results (fig. ), with a quantitative analysis of the decay rates lending additional support that structural disorder is correlated with capacitance. (fig. S17-S21, table ). Summarizing, we have found that more disordered carbons with smaller ordered domains have higher capacitances. 180 These findings may help to resolve the previous contradictory reports on the impact of carbon pore size on capacitance . A master plot of our measured Dd values including literature values for TiC-CDC materials is shown in Fig. . TiC-CDC-600 (i.e., a sample prepared at 600ºC) exhibits a large capacitance and a small Dd value, while TiC-CDC-1000 has a low capacitance and large Dd value. This is consistent with a more disordered structure for TiC-CDC-600, as is well known 185 from previous molecular simulation work and conductivity measurements that showed an increase in structural order as the synthesis temperature was increased. We therefore hypothesize that carbon disorder, rather than pore size, gave rise to previously reported "anomalous" increase in capacitance for samples prepared at low temperature (3). More generally it is possible that the impact of structural disorder has been overlooked in a range of studies on capacitance, and 190 it is clear that structural disorder should be controlled as far as possible in any study of pore size effects. |
6581bcd09138d231610be057 | 11 | Having found that more disordered carbons have higher capacitances, we considered the question of why this is the case. To explore the impact of disorder on the charge storage mechanism, charged supercapacitor electrodes were studied with ex situ MAS NMR experiments on two selected carbons with contrasting levels of disorder, namely PW-400 (ordered carbon) and SC-1800 (disordered carbon). Magic angle spinning was required to resolve the in-pore peaks due to the small Dd value 200 for SC-1800, which precluded in situ measurements. To avoid solvent evaporation during cell disassembly and rotor packing, we employed a 0.5 M PEt4BF4 in propylene carbonate (PC) electrolyte, and then used 19 The ex situ NMR measurements report on the number of in-pore cations and anions at different charging voltages (Fig. , fig. ), as well as the excess ionic charge (Fig. ). For both of the 215 studied carbons, the anion uptake increases while the cation uptake decreases with increasing applied cell voltage for the positively charged electrodes, and vice versa for the negative electrodes (Fig. ). This suggests that both carbons store charge through an ion exchange mechanism, wherein counter ions are adsorbed and co-ions are expelled from the pores for both positive and negative charging . Importantly, the more disordered carbon (SC-1800) shows a greater capacity 220 to store ions at a given voltage than the more ordered carbon (PW-400) (Fig. ). |
6581bcd09138d231610be057 | 12 | We propose that for carbons with smaller domains, the charges are more localized, giving rise to stronger interactions between ions and carbon atoms, thus leading to more efficient storage of ions (fig. ) . This capacity to store ions more efficiently leads to higher capacitance for 225 carbons with smaller domains, similar to previous computational studies on the correlation between capacitance and charge compensation per carbon . We further hypothesize that the smaller domains may be connected with a higher concentration of topological defects (edge sites, pentagonal and heptagonal rings, curvature), which were previously suggested to increase capacitive performance . Indeed, previous studies of hard carbons for sodium ion batteries found a 230 more favorable interaction of sodium ions with edge sites and defects compared to basal planes of graphene-like fragments , with the latter paper ascribing the sloping voltage seen in these systems to a pinning of the inserted electrons by the carbon defects . Finally, while previous studies suggested that carbon-ion distances decrease due to desolvation , we hypothesize that defects may drive the denser packing of ions in the carbon pores . 235 |
6581bcd09138d231610be057 | 13 | In this study we aimed to resolve the now decades-long debate on how the structure of nanoporous carbons electrodes impacts their capacitive energy storage. Electrochemistry measurements on a large series of commercial activated carbons showed no clear correlation between capacitance and pore size, nor between capacitance and specific surface area. In contrast, NMR spectroscopy 240 experiments and modelling revealed a strong correlation between capacitance and electrode structural disorder for both the commercial porous carbons as well as their thermally annealed counterparts. Carbons with smaller ordered domains have higher capacitances, which we attribute to their more efficient storage of ions in the carbon nanopores. Overall, this work reveals a previously overlooked structural factor that determines the capacitance of nanoporous carbons. Our 245 findings will guide the design and synthesis of improved electrode materials for EDLCs and will motivate fundamental research on how disorder impacts charge storage. Free-standing carbon films were made by mixing carbon powder (95 wt %) and polytetrafluoroethylene binder (5 wt %) in ethanol. The films were manually rolled into a similar thickness (approximately 0.25 mm). The films were dried in a vacuum oven at 100 ℃ for at least 24 h before use to fully remove the residual ethanol and water. |
6581bcd09138d231610be057 | 14 | N2 gas physisorption experiments were carried out using a high vacuum physisorption/chemisorption analyser (autosorb iQ from Anton Paar) at 77 K. For each carbon sample, around 25 mg of the powder was placed into a glass bulb gas cell and degassed at 120 ℃ in vacuo for 16 h before the measurement. The Brunauer-Emmett-Teller (BET) surface area was calculated from the isotherm using the BET equation and Rouquerol's consistency criteria implemented in BETSI . The average pore sizes were calculated by taking the pore size value at half of the cumulative pore volume in the pore size distribution. Pore size distributions were calculated by QSDFT based on a slit-pore model. |
6581bcd09138d231610be057 | 15 | For the powdered carbon, film electrodes were used for making electrochemical cells (film-making method stated above), while for activated carbon cloths, the cloths were directly used as electrodes. All the coin cells were two-electrode cells, and were assembled in a N2-filled glovebox. The coin cells were symmetric (i.e., both electrodes were the same carbon material). The electrodes were cut using a stainless-steel manual punching cutter (diameter "5/16" inch, Hilka Tools), and the two electrodes in each cell had identical masses within 0.2 mg. Two electrodes were placed onto a coin cell bottom casing with a larger diameter ("9/16" inch) glass fibre separator placed between the electrodes. Around 150 µL of electrolyte was added around the separator with a Pasteur pipette. After the electrolyte fully wetted the electrodes, two SS316 spacing disks and one SS316 spring were placed on top of the electrodes. The top casing was then used to close the coin cell. The coin cells were sealed with a Compact Hydraulic Coin Cell Crimper (Cambridge Energy Solutions). All coin cells parts were supplied by Cambridge Energy Solutions. |
6581bcd09138d231610be057 | 16 | All electrochemical measurements were conducted in a two-electrode configuration with a Biologic BCS-805 potentiostat. Cyclic voltammograms of the cells were first obtained at a scan rate of 10 mV/s with a fixed potential window of 0-2.5 V for at least 5 cycles. For constant charge-discharge measurements, each cell was measured under different current densities at 0.05, 0.1, 0.2, 0.5, 0.75 and 1 A/g with a potential window of 0-2.5 V for at least 3 cycles under each current density. For each type of carbon in each electrolyte system, at least two cells were made. For most carbons three repeat cells were made. The capacitance of each cell was calculated through galvanostatic charge discharge measurements from the slope of the second half of the discharge curve in the last cycle. The representative capacitance for each carbon was the average of repetitive cells with a standard deviation of less than 5 F/g. The error bars in corresponding figures represent the standard deviations between the repetitive cells. |
6581bcd09138d231610be057 | 17 | The electrochemical cells for ex situ experiments were assembled in a 1/4'' Swagelok T-shaped-cell. The electrolyte applied was tetraethylphosphonium tetrafluoroborate (PEt4BF4, >95% purity, Tokyo Chemical Industry UK), 0.5 M, in propylene carbonate (PC) (≥99.0 %, Sigma-Aldrich). Carbon electrodes were cut with a "3/16" inch puncher with mass difference of less than 0.2 mg. The symmetric Swagelok cell was assembled the same way as coin cells with two identical electrodes and an excess amount of electrolyte of 200 µL in a N2-filled glovebox. The third hole was filled by another stainless-steel plunger to seal the cell and avoid H2O and O2 from entering the cell. |
6581bcd09138d231610be057 | 18 | The charging process prior to NMR experiments was achieved by constant-potential hold. The cell voltage was held at 0 V for 1 hour before being held at a desired voltage for at least another 1 hour. The electronic charge was calculated via the integral of the current vs. time data. Following charging, the Swagelok cell was then transferred into the N2-filled glovebox and disassembled as quickly as possible (typically within 5 minutes). The positive and negative electrodes were then packed into separate 2.5 mm rotors for solid-state NMR experiments. Excess electrolyte was removed using a small amount of tissue paper, to ensure stable spinning of the sample during the NMR measurements. |
6581bcd09138d231610be057 | 19 | Each carbon sample was made into a carbon film and dried for at least 24 h at 100 °C in vacuo before being transferred to a N2-filled glovebox. The film was then cut and weighed in the glovebox. The weighed film (around 5 mg) was put into a sealed vial overnight with 1 M NEt4BF4 (ACN) to fully saturate the sample (around 150 µL) before being packed into 2.5mm rotors. The rotors were quickly packed (typically within 3 minutes) to avoid evaporation of the solvent. For each sample, the rotor was weighed before and after being packed to calculate the mass of the electrolyte added to the system. Excess amount of electrolyte was carefully removed with tissue papers. |
6581bcd09138d231610be057 | 20 | NMR spectroscopy experiments were carried out with a Bruker Avance Neo spectrometer in a Bruker 2.5 mm HX double resonance probe. Measurements were carried out at a magnetic field strength of 9.4 T, corresponding to a 1 H Larmor frequency of 400.1 MHz. All spectra were acquired with a 90°pulse-acquire sequence at a sample spinning speed of 5 kHz. 5 kHz was selected as it gave resolved spectra and avoided excessive frictional heating effect and centrifugation . Recycle delays were set to be more than five times T1 for each sample to ensure that the experiments were quantitative, which are 8-12 s for the studied carbons. 19 F NMR spectra were referenced relative to neat hexafluorobenzene (C6F6) at -164.9 ppm as a secondary reference. 1 H NMR spectra were referenced relative to the CH3 resonance of ethanol (C2H6O) at 1.2 ppm as a secondary reference. 31 P NMR spectra were referenced relative to 85% phosphoric acid (H3PO4) at 0 ppm. NMR spectra were deconvoluted and fitted in DMfit values. The Δδ values were simulated using a previously developed lattice model , which describes diffusion of fluid molecules in porous carbons and predicts NMR spectra of those species. The lattice used to represent an electrolyte filled carbon particle corresponds to a collection of pore sizes for which a set of quantities have to be defined: i) a pore size, ii) a quantity of adsorbed ions and iii) a chemical shift. |
6581bcd09138d231610be057 | 21 | The quantities of adsorbed ions are calculated for each pore by integrating density profiles obtained from molecular simulations. A more detailed description of this aspect of the model is available in previous works . In the present case, molecular dynamics simulations were performed on a 1M solution of NEt4BF4 in acetonitrile. Ions and solvent molecules were all represented with an all-atom force field, considering electrostatic and van der Waals interactions. Specific parameters can be found in published works . The electrolyte, consisting in 40 ion pairs and 600 acetonitrile molecules, was simulated in contact with planar carbon electrodes. All simulations were conducted in the NVT ensemble using the LAMMPS software , a timestep of 1 fs and applying a temperature of 298K. |
6581bcd09138d231610be057 | 22 | To estimate a correlation length, Δδ values are calculated for a series of 5 to 7 simple aromatic molecules (coronene, circumcoronene, dicircumcoronene, etc.) with increasing size and compared with the Δδ values determined experimentally. This (A) 19 F NMR spectra of the commercial activated carbons and activated carbon cloths soaked with 1 M NEt4BF4 (ACN). (B) 19 F NMR spectra of the pristine ACS-PC and thermally annealed ACS-PCs soaked with 1 M NEt4BF4 (ACN). (C) 9 F NMR spectra of the pristine EL-104 and thermally annealed EL-104s soaked with 1 M NEt4BF4 (ACN). The in-pore linewidths for the different studied carbons differ, which could arise from the distributions of ions in various pore environments , the mobility of ions in the pores , and/or the different exchange rates between in-pore and ex-pore ions . Table : Table of coefficients (including standard errors) from exponential equations for each carbon, where the equation for an exponential line is y = y0 + (A * exp(R0 * x)), where A is amplitude and R0 describes the rate of decay. The exponential fittings of PW-400 and YP-50F are described by lower amplitudes and less negative rates of decay, indicating a slower decay of the pairwise C-C correlations compared to SC-1800 and ACS-PC. For YP-50F and YP-80F, similarities between the exponential fittings are consistent with the similar capacitance for the two carbons. |
66d73f08cec5d6c14204f31a | 0 | Many chiral helicenes show interesting chiroptical properties and these materials have been widely investigated in such diverse fields as bioimaging, optical data storage, optical spintronics applications, and circularly polarized organic light-emitting diodes (CP-OLEDs). Most emissive chiral helicenes are fluorescent, but for efficient OLEDs, triplet excitons need to be harvested, so helicenes that emit either phosphorescence or thermally activated delayed fluorescence (TADF) would be potentially attractive emitters. Organic multi-resonant TADF (MR-TADF) materials have attracted particular attention as emitters for OLEDs as not only can devices attain 100% internal quantum efficiency (IQE) but they also show narrowband emission and high photoluminescence quantum yield (FPL). Thus, the development of chiral MR-TADF emitters would be particularly attractive for CP-MR-TADF OLEDs. |
66d73f08cec5d6c14204f31a | 1 | Where 𝜀 $ and 𝜀 % are the molar absorptivity coefficients associated with left-and righthanded circularly polarized light, respectively. IL and IR correspond to the intensities of the emitted left-and right-handed light, respectively. gPL values of -2 or 2 indicate 100% right-or left-handed CPL. As the CPL is related to the transition between excited states and ground state of the emitter, the gPL of a chiral molecule is dependent on the magnitude of the magnetic and electric transition dipole moments and their relative orientation according to equation 3: |
66d73f08cec5d6c14204f31a | 2 | where μ and m refer to the electric transition dipole moment and magnetic transition dipole moment vectors between the emissive excited state and the ground state, respectively, and q is the angle between μ and m. For closed shell organic small molecules, including MR-TADF emitters, |μ| is roughly 300 times larger than |m|, which results in very small gPL values, typically less than 10 -2 . Given the large electric transition dipole moment, equation 3 can be simplified to equation 4: |
66d73f08cec5d6c14204f31a | 3 | It therefore becomes clear that to enhance gPL requires a combination of higher |m| and lower |μ| and a parallel arrangement between them. Unfortunately, the planar geometries of most MR-TADF materials, including MR-TADF helicenes, result in a nearly perpendicular arrangements between m and μ (Figure ), which result in low absolute values of cosq and low gPL values of < 3 × 10 -3 (Figure ). However, if two identical MR-TADF helicenes are fused into one C2-symmetric dimer molecule, the integrated μ and m of the dimeric emitter would be parallel or anti-parallel to each other (Figure ). A large cosq value of 1 would certainly contribute to a higher gPL value. Based on this strategy, in particular, Guo et al. developed |
66d73f08cec5d6c14204f31a | 4 | DiKTa3 emits at lPL of 491 nm (FWHM of 52 nm/0.26 eV) and has a moderate DEST value of 0.24 eV in toluene and is very poorly emissive. The 1 wt% doped film of DiKTa3 in PMMA exhibits TADF. The density functional theory calculated (M062X/6-31G(d,p)) chiroptical properties based on the optimized geometry predict gabs values of 2.1 × 10 -2 for the S0-S1 transition due to a cosq value of 1. However, the predicted value differs considerable from the measured one in toluene (gabs of 6.8 × 10 -4 ). This is the result of DiKTa3 adopting a conformation that is not C3-symmetric; indeed, the calculated gabs value using geometry obtained from the single crystal structure is 8. moments and the corresponding gabs value. These results may also explain why many triple helicenes show low gabs values of around 10 -3 . Results & Discussion |
66d73f08cec5d6c14204f31a | 5 | The optimized geometry of the ground state and its electronic structure of DiKTa3 were calculated using Density Functional Theory (DFT) at the PBE0/6-31G(d,p) level, starting from a structure drawn and optimized using Chem3D (Figure ). The highest occupied molecular orbital (HOMO) of DiKTa3 is located on two of the DiKTa blades while the lowest unoccupied molecular orbital (LUMO) is distributed across the whole molecule. The calculated HOMO and LUMO energies of DiKTa3 are -6.42/-2.72 eV. Both orbitals are more stabilized than the parent DiKTa (HOMO of -6.20 and LUMO of -2.23 eV) and the double helicene Hel-DiDiKTa (HOMO of -6.26 and LUMO of -2.53 eV). The excited-state energies were calculated using time-dependent DFT (TD-DFT) within the Tamm-Dancoff approximation (TDA) at the PBE0/6-31G(d,p) level. The S1 and T1 energies of DiKTa3 are calculated to be 2.85/2.43 eV, with DEST of 0.42 eV. There are six triplet states below the S1 state, which may participate in the RISC process. The SOC matrix elements (SOCME) between S1 and T1 ). These values are close to those of the parent DiKTa (1.26, 6.66, 9.34 and 6.84 cm -1 for S1-T1, S1-T2, S1-T3 and S1-T4 transitions, respectively, Figure ). It is known that TD-DFT calculations overestimate ΔEST, and so spin-component scaling second-order algebraic diagrammatic construction (SCS-(ADC)2/cc-pVDZ) calculations were undertaken as these account for higher order electronic excitations. As shown in Figure , DiKTa3 shows short range charge transfer (SRCT) character with the typical alternating increasing and decreasing electron density pattern on the center of the molecule for both S1 and T1. The S1 and T1 energies of 2.72 and 2.58 eV are much more stabilized compared to those of DiKTa (S1=3.46 eV and T1=3.20 eV). The corresponding ΔEST value is calculated to be 0.14 eV. |
66d73f08cec5d6c14204f31a | 6 | Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) of DiKTa3 in degassed DCM with 0.1 M [ n Bu4N]PF6 as the supporting electrolyte and Fc/Fc + as the internal reference (0.46 V vs. SCE) were measured to experimentally determine the HOMO and LUMO energy levels (Figure ). DiKTa3 shows a reversible reduction wave and an irreversible oxidation wave, electrochemical behavior that is consistent with that of DiKTa. The reduction potential, Ered, and oxidation potential, Eox, of DiKTa3, obtained from the peak values in the DPV are -1.02 and 1.83 V vs. SCE. Both values are cathodically shifted compared to DiKTa (Ered of -1.34 V and Eox of 1.78 V vs. SCE) and closer to those of Hel-DiDiKTa (Ered of -1.16 V and Eox of 1.81 V vs. SCE). The corresponding HOMO and LUMO energies are -6.17 and -3.32 eV with energy gap DE of 2.85 eV (Table ). |
66d73f08cec5d6c14204f31a | 7 | The ultraviolet-visible (UV-vis) absorption and photoluminescence (PL) spectra of DiKTa3 in toluene at room temperature are shown in Figure . The lowest-energy absorption band is at 450 nm and has a molar absorptivity of 1.6 × 10 4 M -1 cm -1 , values that resemble those of Hel-DiDiKTa (453 nm, 1.4 × 10 4 M -1 cm -1 ) and can be assigned to the S0-S1 transition. The redshifted band and lower molar absorptivity compared to that in DiKTa (433 nm, 2.1 × 10 4 M -1 cm -1 ) align with the trends of the calculated oscillator strength and energy of the S0-S1 transition (2.72 eV and 0.0016 for DiKTa3 and 3.46 eV and 0.20 for DiKTa). DiKTa3 emits at lPL of 491 nm, shows narrowband emission, with a FWHM of 52 nm/0.26 eV, and has a small Stokes shift of 38 nm/0.21 eV in toluene. Its emission is red-shifted compared to DiKTa (453 nm) and Hel-DiDiKTa (473 nm). The shoulder emission band peaking at 515 nm and the higher-intensity absorption band at 429 nm (2.4 × 10 4 M -1 cm -1 ) may come from the transitions between different vibrational energy levels as the PL spectra remain unchanged as a function of changing excitation wavelengths between 405-455 nm (Figure ). In addition, the energy of the emission band is insensitive to solvent polarity, which indicates that its SRCT state has significant locally-excited character (Figure ). The FPL of DiKTa3 in degassed toluene is measured to be 0.4%, the magnitude of which is mainly due to the significant non- radiative decay. This low FPL value is perhaps not surprising as that of Hel-DiDiKTa is also ≈ 1% and such low FPL is not uncommon in triple-helicene compounds. which is slightly larger than the calculation result (0.14 eV) and similar to that of single DiKTa (0.20 eV). The time-resolved PL decay in degassed toluene is shown in Figure . The prompt emission of DiKTa3 decays with lifetime τPL of 6.7 ns. No delayed emission was observed in solution because of the low FPL, large DEST, and significant non-radiative decay. The photophysical properties of DiKTa3 in doped polymethyl methacrylate (PMMA) films were then investigated. The FPL remains very low at 0.6% in 1 wt% doped film. DiKTa3 emits at lPL of 492 nm with FWHM of 54 nm/0.26 eV at 300 K (Figure ). The S1, T1 energies and DEST value in film, determined from the SS PL and Ph spectra at 77 K, are 2.64, 2.41 eV and 0.23 eV (Figure ), which are close to those measured in toluene. The prompt emission decays with a τPL of 4.0 ns (Figure ). There is a weak observed delayed emission, with a lifetime, τd, of 487 ns (Figure ). This delayed emission increases in intensity with increasing temperature, thus we can assign this to TADF. Recognizing that DiKTa3 is chiral, we next investigated its chiroptical properties. The enantiomers of the racemic sample of DiKTa3 were firstly separated by chiral HPLC to afford ee. of 100% and 99.6% for the first and the second enantiomers, respectively (Figure ). The simulated ECD spectra of (P,P,P)-DiKTa3 in toluene based on PBE0/6-31G(d,p) and M062X/6-31G(d,p) calculations were employed be able to assign the absolute configuration by comparison with the measured ECD spectra (Figure ). Both simulated ECD spectra of 39.4, 39.5, 39.4° for each functional, respectively (Figure ). They give similar predicted magnetic transition dipole moment values |m|, small q of 33 and 3° and high cosq of 0.84 and 1, respectively. The computed electric transition dipole moment values |μ| are higher using M062X/6-31G(d,p), which leads to a lower calculated gabs value of 2.1 × 10 -2 compared to that of PBE0 (5.6 × 10 -1 ). However, both predicted values are vastly overestimated compared to the experimental result in toluene. This is likely due to the differences between the near-identical optimized torsion angles and those observed in the single crystal structure (38.4, 35.7 and 42.8°, Figure ). Therefore, the chiroptical properties of (P,P,P)-DiKTa3 were also calculated using the crystal structure geometry at the M062X/6-31G(d,p) level for comparison (Figure ). The calculated |m| is close to the calculations based on the optimized geometries. However, the differences between the three torsion angles breaks the symmetry of the molecule, resulting in an increased |μ| and smaller cosq of 0.38. The corresponding gabs value is calculated to be 8.1 × 10 -4 , which is much closer to the value measured in toluene for the compound (6.8 × 10 -4 ). |
67d06040fa469535b9dff783 | 0 | Hydrogen scrambling in organic ions is a well known phenomenon and holds special significance for biological systems . The extent of this positional exchange of hydrogen atoms within a molecule depends on the internal energy of the ions and is expected to be greater for metastable ions as compared to ions formed with higher excitation energy which predominantly dissociate into two or more fragments.For smaller molecules, the term intramolecular hydrogen migration is more commonly used to describe this phenomenon. Formation of new bonds (H -H or C -H or O -H) after cleavage of existing C-H and C-O bonds within the ionic precursors of small molecules is extensively reported . Typical examples of such processes is the formation of H + 3 fragment from small alcohols , formation of H 2 O + and H 3 O + from ethanol and formation of CH + 3 from butadiene . The ideas of a roaming H-atom and a roaming H 2 molecule have been put forth to explain the results of time-resolved studies on dissociation of organic molecules. |
67d06040fa469535b9dff783 | 1 | As the H-atoms exchange their positions within a small organic molecular ion, and roam about the ionic parent as an H-atom or an H 2 molecule, it is possible that some of them may end up getting lost leaving behind a lighter molecular ion but with the same charge. For a small doubly charged molecular ion, one would naively expect that it would dissociate into two singly charged fragments in order to reduce its potential energy, a process usually termed as Coulomb explosion (CE). The emission of one or more neutral particles from such a dication is quite unexpected as this process effectively reduces the trapping volume for the positive charge. Nonetheless, if neutral particles are lost, the fate of this daughter ion will then depend on its structure and the internal energy stored in it. If the neutral particle is an H-atom/H 2 molecule or C 2 H 2 molecule, all these species may carry some internal energy during their emission. The daughter ion may thus either stabilize or dissociate into two or more fragments. |
67d06040fa469535b9dff783 | 2 | From time-resolved studies on molecular dissociation, the time scales for a roaming H 2 molecule are determined and have been found to be of the order of few hundred femtoseconds. Such experiments usually employ pump-probe techniques to determine the yield of a fragment as a function of time delay, with the chosen fragment being the one resulting from a hydrogen migration process. Since the scrambling and migration of H-atoms is dependent on the internal energy and lifetime of the molecular precursor, these experiments sample a small section of the potential energy landscape of the molecular precursor. On the other hand excitation by ion-impact is known to be much more violent leading to a much wider distribution of internal energy in the target molecule. |
67d06040fa469535b9dff783 | 3 | Recently, the evidence for presence of E -1-cyano-1,3butadiene was reported from the dark prestellar cloud TMC-1 . Since 1,3-butadiene does not have a dipole moment, it cannot be directly observed in astrophysical environments, and the use of its cyano derivative (E -1cyano-1,3-butadiene) is suggested for indirect estimation of its concentration in this molecular cloud . As 1,3butadiene plays a crucial role in formation of the first 6-membered carbon ring molecule , this makes it one of the most important hydrocarbon molecule of the inter-stellar medium. This molecule is believed to be the precursor for formation of cyanobenzene, the simplest aromatic nitrile, which was detected in TMC-1 few years ago . |
67d06040fa469535b9dff783 | 4 | In the present article we investigate the results of an interplay between hydrogen migration, hydrogen scrambling and loss of multiple neutral fragments in the dication of the smallest polyene, 1,3-butadiene, in its most stable entgegen E form . These dications were formed by the collisional interaction of 810 keV Xe 3+ ions with neutral molecules of 1,3-butadiene. The fragmentation products resulting from an isolated dissociating dication were detected together in coincidence using a time-offlight mass spectrometer. The butadiene molecular dication formed with high internal energy and left to itself, devoid of any surroundings, is capable of emitting multiple neutral particles (H/H 2 /C 2 H 2 ) leading to formation of several daughter dications having different masses. Although, the emission of a single neutral particle (H/H 2 ) as a means of releasing internal energy stored in a molecular ion has been reported in literature for many small aliphatic hydrocarbon molecules , reports on loss of multiple neutral fragments from such ions are rare and not discussed. |
67d06040fa469535b9dff783 | 5 | Our results reveal that in a large number of cases, the dissipation of energy from parent butadiene dication starts with the emission of one or more neutral particles (H/H 2 /C 2 H 2 ) followed by a two-body breakup of the daughter dications into singly charged fragments. We observe that most of these daughter dications break into several different fragment pairs which can only result from hydrogen scrambling. The high propensity of multiple neutral emission is highlighted and discussed. Our study indicates that formation of ionic precursors can open up a rich chemistry of unsaturated carbons involving multiple structures with broad fragmentation time-scales and energetically accessible pathways for evolution . With a very high propensity for neutral particle emission, such ions can also play a significant role in contributing to the H/H 2 concentration in astrophysical environments. |
67d06040fa469535b9dff783 | 6 | Evolution of dications in the first few hundreds of femtoseconds, dynamics upto 1000 fs as well as the structural and energetic information, including the possible intra-molecular cyclization is explored. Figure depicts the structure of butadiene formed upon vertical Franck Condon double ionization from the most stable neutral form and the dissociation pathways explored in this report. |
67d06040fa469535b9dff783 | 7 | The experiment was carried out at the Low Energy Ion Beam Facility (LEIBF) of Inter-University Acceler- ator Centre (IUAC), New Delhi, India. Xe 3+ projectile ions with 810 keV energy (velocity = 0.5 a.u.) were made to interact with molecules of neutral 1,3-butadiene in an evacuated chamber under single collision conditions. The ion-molecule interaction was studied using a recoil ion momentum spectrometer equipped with a position sensitive micro-channel plate (MCP) detector and a channel electron multiplier (CEM). The ions and electrons generated during the interaction are extracted using an electric field of 400 V/cm. The time-of-flight of the recoil fragments was recoded in coincidence with an emitted electron. Using a multi-hit time-to-digital convertor, all recoil fragments corresponding to a single start trigger were collected in coincidence. More details about the experimental setup can be found in reference . |
67d06040fa469535b9dff783 | 8 | The collection efficiency of ions with identical mass by charge ratio is limited in our experiment because of dead time effects. Due to the inherent kinetic energy spread caused due to CE, it is difficult to separate ions with similar mass by charge ratio. This applies to events corresponding to cleavage of central C-C bond of the butadiene molecule when fragment ions of the kind C 2 H + n (n = 0 to 3) and C 2 H + m (m = 0 to 4) are detected in coincidence. Thus, we will not discuss the relative contributions of such fragmentation channels quantitatively. |
67d06040fa469535b9dff783 | 9 | On the theoretical aspect, DFT-based Born-Oppenheimer Molecular Dynamics (BOMD) simulations were carried out in the ground electronic state, em-ploying the multipole accelerated resolution of the identity with the Ahlrich type def2-SV(P) basis set in combination with Becke-Johnson damping dispersion corrected B3LYP functional . Excitation energies (E int ) of 0.01, 1, 2.5, 5, 10, 15, 20 and 25 eV, time step ∆t = 10 a.u. (∼ 0.25 fs), maximum simulation time t sim ≃ 1000 fs parameters were chosen. To mimic the excitation and ionization of the 1,3-butadiene target system we introduce the aforementioned amount of excitation energy randomly distributed over all the nuclear degrees of freedom of the molecule for each trajectory and extracted two outer electrons in a Franck-Condon type transition. Finally, we run 100 trajectories for E int = 0.01, 1, 2.5, 5, 20 and 25 eV and 200 trajectories for E int = 10 and 15 eV, performing in total a reasonable 1000 trajectory sampling. For all BOMD we used TURBOMOLE code with default settings . |
67d06040fa469535b9dff783 | 10 | Structure and stability of dications was explored including the inactive (non-breaking) trajectories for 0.01, 1, 2.5 5, 10, 15 and 20 eV, in practice the last steps from BOMD. Geometry optimizations have been carried out at the B3LYP-D3/6-311++G(d,p) level of theory, as well as zero point energy correction were added. All static calculations were performed using the Gaussian16 program . |
67d06040fa469535b9dff783 | 11 | The time-of-flight (TOF) spectrum for the dissociation of 1,3-butadiene (mass = 54 amu) is shown in figure where several fragments of the kind C n H + m with n = 1 to 4 and m = 0 to 6 are visible. We note that fragments originating from all possible bond cleavages appear in the TOF spectrum. Furthermore, fragments H Some of the fragments recorded in our experiment can only be formed by the process of hydrogen scrambling and/or hydrogen migration like |
67d06040fa469535b9dff783 | 12 | is a two-dimensional distribution of TOF of first detected ion with TOF of second detected ion. Figure shows the coincidence islands for two singly charged fragments of the kind CH + n (n = 0 to 3) with C 3 H + m (m = 0 to 3), which corresponds to the asymmetric breakup of the molecular dications. |
67d06040fa469535b9dff783 | 13 | For a dication dissociating into two singly charged fragments, the corresponding coincidence island will have a slope of -1 following momentum conservation . The most prominent island on figure comes from coincidence between CH + 3 and C 3 H + 3 , an observation which is in agreement with earlier reported results . The slope of this island is -1 indicating that both these fragments originated from the same C 4 H 2+ 6 dication. Thus by considering the slope of the coincidence islands, the doubly charged precursor dication can be identified. Most of the islands seen in figure have a slope close to -1. |
67d06040fa469535b9dff783 | 14 | Thus the doubly charged precursors for these islands are dications of the type C 4 H 2+ k (k = 0 to 6) formed after emission of one or more neutral particles (H/H 2 ) by the parent dication C 4 H 2+ 6 . This is shown as an indirect asymmetric CE in figure and may be described as |
67d06040fa469535b9dff783 | 15 | Apart from emission of one or more neutral particles as H-atom(s)/H 2 molecule(s), we have also inferred emission of neutral C 2 H 2 from parent dication C 4 H 2+ 6 , from the coincidence map. We observed a coincidence island with fragment pair (H + , C 2 H + 3 ) and having a slope close to -1 indicating formation of precursor dication is C 2 H 2+ 4 . This process may be represented as |
67d06040fa469535b9dff783 | 16 | In our experiment, we recorded a total of 33 fragmentation channels where the slope of the coincidence island is close to -1. Out of these, 28 channels show emission of one or more neutral particles (H/H 2 /C 2 H 2 ) prior to the CE into two singly charged fragments. Of these fragmentation channels, table I lists the channels populated without any loss of neutrals and II lists the channels populated after the loss of one or more neutrals as H/H 2 . Furthermore, we observed a coincidence island associated with a dication having a naked carbon skeleton formed after the emission of all 6-hydrogen atoms from parent butadiene dication. This is inferred from the presence of a coincidence island containing C + with C + 3 in figure . |
67d06040fa469535b9dff783 | 17 | In our Born-Oppenheimer Molecular Dynamics (BOMD) simulations, the loss of neutral fragments appears as a major route for dissipation of excess energy for butadiene dications and it depends sensitively on the internal energy. For 25 eV of internal energy, the branching ratio of trajectories with loss of one or more neutral particles (H/H 2 ) is as high as 91% while for 10 eV this ratio is 35% as evidenced from figure . |
67d06040fa469535b9dff783 | 18 | From our theoretical data, we analyzed the cleavage of C-C bond, leading to symmetric and asymmetric twobody breakup, as a function of number of H-atoms lost by the dication. It is found that for trajectories where no H-atom is lost or only one H-atom is lost, symmetric breakup dominates. Once two or more H-atoms are |
67d06040fa469535b9dff783 | 19 | lost (as H/H 2 ) the trend reverses and cleavage of terminal C-C bond leading to asymmetric breakup starts to dominate as shown in figure . Hydrogen migration is observed to be a very efficient process in our MD simulations. In this process, the "central H-atom" can migrate to (a) the "nearest terminal C-atom" or (b) the "furthest terminal C-atom" as shown in figure . H-migration is not observed within the simulation timescale of ∼1 ps at a very low excitation energy of E int of 0.01 eV. Thus, it is not a phenomena driven purely by the charge hole. The process (a) opens at 1eV (10% of trajectories lead to this tautomer) and becomes 47% of importance for 2. |
67d06040fa469535b9dff783 | 20 | Cleavage of terminal C-C bond Other isomerizations including double H-migrations are much less prominent. The fragment pair (CH + 3 , C 3 H + 3 ) formed after hydrogen migration is the most dominant direct asymmetric CE channel in our MD simulations for 10 and 15 eV excitation energy which agrees well with our experimental results. |
67d06040fa469535b9dff783 | 21 | Our MD simulations enable us to compare time scales for the onset of hydrogen migrations, C-C bond cleavage and C-H bond cleavage as a function of the internal energy. For E int = 5 eV H-migration to "nearest terminal C-atom" is much faster than migration to the "furthest terminal C-atom". However, it is surprising that for E int > 5 eV the time for both H-migrations is almost the same. Considering a range of internal energies from 1.0 eV to 20 eV, the time-scale for the onset of these migrations range from 420 fs to 10 fs respectively. To explore the appearance of multiple fragmentation pathways originating from the same parent dication, we looked at the structures obtained from our MD simulations. Multiple different structures emerged at the end of our simulation time of upto 1 ps and these were then optimized to locate the nearest minimum energy structure. For most of the dications, formed after loss of one or more neutral particles (H/H 2 ), it is observed that the H-atoms are preferentially attached to the terminal C-atoms and not the central carbon atoms. This is due to cleavage of central C-H bond and/or emission of terminal H-atoms as H/H 2 followed by scrambling which drives the central hydrogen atoms to terminal carbon atoms. |
67d06040fa469535b9dff783 | 22 | In figure we show the experimentally recorded relative branching ratio for events collected from breakup of doubly charged daughter molecular ions formed after loss of one or more neutral particles (H/H 2 ) from the butadiene dication leading to indirect asymmetric CE. For comparison, the events collected due to the asymmetric CE of intact butadiene dication are also shown. From this figure, we note that the indirect asymmetric CE is ≈200% more likely than direct asymmetric CE. Thus an internally hot and unstable butadiene dication has a very high probability of evaporating neutral particles prior to fragmentation. The results of our MD simulations are in good agreement with the observed high propensity of neutral particle emission for butadiene dications. Such intense dehydrogenation is expected to change the abundance of hydrogen in astrophysical environments, in analogy with the case of PAH molecules . Ion molecule chemistry involving [C 4 H n ] + leading to formation of 6membered C-ring cations is a proposed route for molecular growth . Several of these fragments in neutral forms are computationally determined as precursors for 6-membered rings . |
67d06040fa469535b9dff783 | 23 | From the coincidence map shown in figure and from table II it can be seen that three different pairs of singly charged fragment ions are formed from the precursor dication [C 4 the hydrogen atoms distributed at the terminal carbons, [CH 2 CCCH 2 ] 2+ . In addition, we also obtained two close lying cyclic structures, one with 4-member carbon ring and another having a 3-member carbon ring. For C 4 H 2+ 3 dication, the lowest energy structure is once again a linear carbon chain with the hydrogen atoms distributed at the terminal carbons, [HCCCCH 2 ] 2+ and the next lowest energy structure has a cyclic 3-member carbon ring. Although we see both, cyclic and linear structures for these dications, our MD simulations shows higher distribution of linear ones. For both these dications, the formation of the highest energy structure is the most likely possibility for explaining appearance of fragment pairs (CH , is formed, several different processes become active for dissipation of excess energy of this dication. These processes include the emission of neutral particles (H/H 2 /C 2 H 2 ), hydrogen migration, and hydrogen scrambling, breaking of central or terminal C-C bonds, and breaking of C-H bonds. All these processes are sensitive to the internal energy of the dication and have different onset time scales. With ion-impact induced ionization, dications are expected to have a wide range of internal energy distribution. That this is the case, is evident from the large number of dissociation fragments observed in the experiment with involvement of all the different processes mentioned above. |
67d06040fa469535b9dff783 | 24 | Figure shows selected examples of the interplay of roaming and scrambling leading to the (a) formation of H + 3 ion, (b) cyclization, and (c) emission of H-atoms. The formation of H + 3 via capturing a proton by a roaming H 2 molecule is found to be a slow process with a minimum opening time of around 650 fs, which is close to three times larger than the time scale for the formation of H + 3 via the the same process in ethanol . On the other hand, the hydrogen scrambling can start as early as 25 fs after the formation of the dication, and the CE after emission of H-atoms can take more than 100 fs. The formation of fragments with cyclic structure can take more than 200 fs. These processes and their associated time scales are also a function of the amount of internal energy available with the precursor ion. Thus, simulation times of the order of several hundreds of femtoseconds are required to capture and address all possibilities of energy dissipation by internally hot molecular ions. |
67d06040fa469535b9dff783 | 25 | To conclude, we have addressed the two-body dissociation of the dications formed after double ionization of E -1,3-butadiene using experimental and theoretical techniques. Our study of the fragmentation of doubly charged butadiene induced by Xe 3+ impact shows the important role of multiple neutral particles (H/H 2 /C 2 H 2 ) emissions. Our results indicate that this emission of neutral particles is the first step towards relaxation for these hot dications. At the experimental time scales, the formation of an unstable and fully naked doubly charged carbon skeleton is observed, which shows complete dehydrogenation of butadiene dication prior to CE. Our theoretical results show that the probability of two-body dissociation with or without emission of neutral particles is strongly de-pendent on the internal energy of these dications. The observation of multiple dissociation channels for dications formed after the loss of neutral fragments is indicative of the involvement of multiple structures, which is possible only with hydrogen migration and hydrogen scrambling. The role of scrambling and roaming processes that are slowing down the fragmentation, prior to direct CE fragmentation dynamics is observed. These processes affect the symmetric and asymmetric bond breaking of the carbon skeleton. To get insight into such complex dynamics, where several different processes with widely varying time scales are active, simulation times of several hundreds of femtoseconds are needed. Such intense dehydrogenation and the formation of a large variety of fragment monocations are expected to significantly alter the chemistry of astrophysical environments containing this molecule. |
67c221956dde43c908e8de13 | 0 | It costs 1-2.5 billion US dollars and takes 10-15 years for a new drug to be fully developed . Drug development is divided into preclinical and clinical stages. Preclinical studies focus on understanding the mechanisms and targets of candidate drug compounds and validating efficacy and possible toxicity. Drug candidates that pass preclinical screening are routed into clinical trials with human subjects. However, with current technologies, compounds that pass preclinical testing and validation have a low probability (<10%) of being launched as a drug . This high failure rate, largely caused by the limited predictive power of preclinical models, not only drives up the cost and time for drug development, which will eventually be propagated as patients' burdens, but can also lead to unexpected tragic outcomes in clinical trials or even after a drug is marketed. A recent example is the compound BIA10-2474, which killed and disabled human subjects after comprehensive preclinical assessments . |
67c221956dde43c908e8de13 | 1 | Currently, preclinical studies rely heavily on cell culture and animal models . Cell culture in well-established containers such as flasks, Petri dishes, and well plates is simple, scalable, relatively cost-efficient, and yields fast results. Nonetheless, growing cells in a 2D monolayer on a flat surface does not provide in vivo cell-cell and cellextracellular matrix (ECM) interactions, and thus may not represent human physiology. |
67c221956dde43c908e8de13 | 2 | The ECM is a network of micro-/nano-fibers composed of glycoproteins, collagen, fibronectin, elastin, and other macromolecules . Emerging evidence suggests that in addition to acting as a nesting niche for cells, the chemical and physical properties of the ECM (e.g., stiffness, microstructures ) are key regulators of intracellular biochemical processes . On the other hand, although animal models provide 3D tissue structures, drug responses from animals can differ drastically from humans . While lab animals and humans have only slight differences in their genomes, these differences can be significantly amplified during transcriptional and post-transcriptional processes, causing major phenotypic variances , such as in hepatocyte enzyme activity . |
67c221956dde43c908e8de13 | 3 | Therefore, new preclinical models that align more closely to human in vivo responses are being actively studied. Hitherto, two technologies have mainly been investigated: 3D-cell culture-based physiological models and organs-on-a-chip models . By incorporating ECM materials such as hydrogels and/or fibrous scaffolds in human cell culture, 3D-cell culture-based physiological models enhance intercellular and cell-ECM interactions, making cell functions and activities more closely match in vivo behaviors . For instance, 3D spheroid tumor models exhibit more complex, in vivo-like responses to therapeutics vs. flat, 2D cultures . Organs-on-a-chip can connect different cell types in physiologically relevant architectures with finely tuned flows to nourish cells, remove waste, and provide physical stimuli (e.g., shear stress ) to replicate relevant microenvironments. Such systems can be used to study interactions between organs, such as between the gut and liver, which are known to impact drug toxicity . |
67c221956dde43c908e8de13 | 4 | Investigating absorption, distribution, metabolism, elimination, and toxicity (ADMET) is an essential part of drug development . Drug metabolism is mainly carried out by the liver in vivo . Understanding the metabolism of a drug candidate is critical: 1) certain drugs must be metabolized by the liver before they can be excreted, and knowing the metabolism rates is indispensable to assess drug and metabolite accumulation issues ; 2) prodrugs need to be metabolized to become active compounds ; 3) orally delivered drugs are extensively metabolized in the liver before reaching circulation, and understanding the metabolism rates guides predictions of bioavailability ; 4) the metabolism rates determine the dosage regimes of a drug . |
67c221956dde43c908e8de13 | 5 | Integrating reliable and predictive liver models in the drug development pipeline will reduce cost burdens and safety issues. Indeed, various enlightening models have recently been investigated. For instance, Ewart et al. reported a liver-on-a-chip model on a polydimethylsiloxane (PDMS) microfluidic apparatus, containing sinusoidal endothelial cells, stellate cells, and primary human hepatocytes sandwiched in Matrigel ECMs. The model showed toxicity responses to clozapine and olanzapine. A similar model has also been tested to show cross-species toxicity responses from primary hepatocytes to bosentan . However, while toxicity responses are an important component of preclinical testing, hepatocyte models that can reproduce physiological drug metabolism/clearance profiles are limited in literature. Although a few models using primary cells are reported , stringent statistical comparisons with in vivo data are not available. |
67c221956dde43c908e8de13 | 6 | Hence, we report our effort to develop a 3D hepatocyte model based on a composite nanofibrous scaffold that mimics both the microstructures of the native liver ECM and its biochemical composition. Instead of using primary human hepatocytes, which have limited availability and are difficult to trace (e.g., donor health/medical history), or hepatocyte-like cells derived from pluripotent stem cells, which are expensive and challenging to maintain , we used HepaRG hepatocytes in our model . HepaRG is an immortalized cell line that can be differentiated into human hepatocyte cells in one simple step . Numerous pieces of evidence have demonstrated the cells preserve liverspecific genes, protein expression, and functions . We found that HepaRG progenitor cells were successfully seeded and differentiated on the scaffold and exhibited typical hepatic functions. More importantly, by treating the differentiated hepatocytes with well-documented drugs that need to be metabolized before excretion, we discovered that hepatocytes on our model reproduced in vivo drug clearance rates, while cells in conventional well plates (2D culture) showed significantly slower drug clearance and reduced clearing capacities. Also, to make the model more translational-after all, end users tend to implement the simplest models, we developed modular 3D-printed holders and batch-produced scaffold inserts, which can be assembled in seconds and housed in standard well plates. The ease of use of the fibrous scaffold and its ability to modulate in vivo-like functions of HepaRG hepatocytes represents a novel and significant improvement in preclinical modeling. |
67c221956dde43c908e8de13 | 7 | Fresh porcine liver tissues were obtained from local slaughterhouses, cut into ~1 cm cubes, and put in a 2 L Erlenmeyer flask. The cubes were rinsed with doubly deionized (DDI) water then soaked in 1.5 L decellularization solution composed of 0.5% Triton X-100 (A16046.AP, Thermo Fisher, Pittsburgh, PA, USA) and 47.6 mM ammonium hydroxide (AX1303-3, MilliporeSigma, St. Louis, MO, USA) and placed on a stir plate in 4 °C, stirring at 180 rpm. The decellularization solution was changed every 24 hours for 7 days. The decellularized ECM (dECM) was then thoroughly rinsed with DDI water, patted dry with paper towels, placed in two 250 mL weigh boats (10803-170, VWR International, Radnor, PA, USA), and frozen at -80 °C, after which it was lyophilized and ground into a coarse powder. A total of 300-400 mg of coarsely powdered dECM was digested with 40 mL of 3500 U/mL pepsin (P7000, MilliporeSigma, St. Louis, MO, USA) in 0.01 M HCl solution for 24 hours at 37 °C on an orbital shaker, 60 rpm. The solution was neutralized with 0.1 M NaOH to pH 8.0 to permanently denature the pepsin. The digested dECM solution was centrifuged for 20 min at 10,000 x g, 4 °C in a high-speed centrifuge (Avanti J-E, Beckman Coulter, Brea, CA, USA), the supernatant collected in 50 mL conical tubes, lyophilized for 72 hours and stored in -20 °C. |
67c221956dde43c908e8de13 | 8 | The dECM electrospinning solution was made by combining two solutions: Solution 1 consisted of 20% w/v polycaprolactone (PCL) with an average molecular weight 80,000 (440744, MilliporeSigma, St. Louis, MO, USA) in 2,2,2-trifluoroethanol (TFE) (A10788, Thermo Fisher, Pittsburgh, PA, USA). Solution 2 consisted of 20% w/v pepsinated dECM powder in TFE + 5% acetic acid (AC42322, Fisher Scientific, Pittsburgh, PA, USA). 1.5 mL of Solution 2 was pipetted into 1.5mL of Solution 1 to produce 3.0 mL of the dECM electrospinning solution (10% w/v PCL in TFE, 10% w/v dECM in TFE + 2.5% acetic acid). Electrospinning was performed on the TL-Pro-BM electrospinning machine (Tong Li, Shenzhen, China). As shown in Figure in the Supplementary Information (SI), a 50 mm wide metal mandrel was covered by a polystyrene (PS) sheet with 14.5 mm diameter circular holes cut with a Hobby 5S laser cutter (Full Spectrum Laser, Las Vegas, NV, USA). The circular holes defined the nanofiber deposition locations. A 5 mL Leur-Lok syringe (30964, BD, MA, USA) was used to pump the electrospinning solution through a 24 G metal blunt needle placed 17 cm from the surface of the mandrel. The syringe pump was set to a flow rate of 0.25 mL/hr. |
67c221956dde43c908e8de13 | 9 | The mandrel was rotated at 300 rpm, and the voltage of the mandrel was set to -5 kV, while the needle was set to +17 kV. The electrospinning was conducted for 4 hours. The needle scanned back and forth across the width of the mandrel at a rate of 5 mm/s to deposit a uniform layer of nanofibers (Figure ). After the electrospinning was completed, the PS sheet was placed in the laser cutter to cut out circular inserts of 18.5 mm diameter (Figure and Figure ). The laser-cutting process also heat-fused the edges of the nanofiber and PS layers. |
67c221956dde43c908e8de13 | 10 | An SEM (FEI Nova NanoSEM 450, Thermo Fisher, Pittsburgh, PA, USA) was used to image the dECM inserts at high (2,000x -20,000x) magnification. An insert was mounted on a pin stub SEM mount (75210, Electron Microscopy Sciences, Hatfield, PA, USA), sputter coated with gold, and imaged on the SEM. The DiameterJ 45 plugin for ImageJ/FIJI was used to measure the diameters of the fibers in SEM images. |
67c221956dde43c908e8de13 | 11 | The Picrosirius Red Stain Kit (24901, Polysciences, Warrington, PA, USA) was used to stain collagen in the scaffold inserts. The insert was rinsed with DDI water then soaked in 0.5 mL of Solution B (Picrosirius Red F3BA stain) for 60 minutes. The insert was then rinsed with Solution C (0.1 N HCl) for 1 minute, twice, followed by rinsing in DDI water and then in 70% ethanol for 60 s. The stained inserts were then imaged or placed in a plate reader (SpectraMax i3x, Molecular Devices, San Jose, CA, USA) with the wellscanning mode to detect PSR fluorescence signals (560 nm excitation, 650 nm emission). |
67c221956dde43c908e8de13 | 12 | The CoraLite Plus 488-conjugated Collage Type I Monoclonal antibody (CL488-67288, Proteintech, Rosemont, IL, USA) was used to label the dECM inserts to visualize collagen from the extracted ECM in the nanofibers. The inserts were rinsed 3x with PBS, fixed for 10 min in 4% paraformaldehyde solution, rinsed 3x in wash buffer (PBS + 0.1% Tween 20) for 5 min, and gently rocked in blocking solution (PBS + 0.1% Tween 20 + 1% bovine serum albumin (0332-25G, VWR International, Radnor, PA, USA)) for 60 min at room temperature. The blocking solution was then replaced with 1 mL of antibody solution (1:200 dilution of antibody in blocking solution). The insert was rocked gently overnight at 4 °C, covered with foil to protect it from light. The insert was then washed 3x in wash buffer for 5 min and placed on a WillCo glass bottom dish (HBSB-5040, WillCo Wells, Amsterdam, The Netherlands) with 20 µL SlowFade Diamond Antifade Mountant (S36972, Thermo Fisher, Pittsburgh, PA, USA). |
67c221956dde43c908e8de13 | 13 | A confocal laser scanning microscope (LSM 900, Zeiss, Oberkochen, Germany) was used to image the fluorescence from PSR-stained and collagen I antibody-labeled inserts. A 561 nm laser was used to excite the PSR-stained sample with a detection band of 613-700 nm. A 488 nm laser was used to excite the CoraLite 488 fluorophore conjugated to the Collagen I antibody with a detection band of 505-550 nm. A 63x objective (oil immersion) was used for both the PSR and CoraLite 488 imaging. |
67c221956dde43c908e8de13 | 14 | The dECM insert holder assembly was composed of an insert holder (Figure The growth medium was refreshed every 2-3 days over 7 days of proliferation in a T75 flask. The flask was then rinsed twice with 5 mL PBS prior to adding 5 mL of 37 °C trypsin (25200056, Thermo Fisher, Pittsburgh, PA, USA) and incubated at 37 °C for 10 min. |
67c221956dde43c908e8de13 | 15 | Growth media (5 mL) was added to the flask to neutralize the trypsin, and the contents of the flask were transferred to a 15 mL conical tube and centrifuged for 5 min @ 2,000 rpm (SCL456, Southwest Science, Trenton, NJ, USA). The supernatant was decanted, and the cell pellet was resuspended in 5-7 mL of growth medium by gentle pipetting. The cells were counted using the EVE automated cell counter (NanoEntek, Seoul, South Korea), and the cell suspension was diluted to 500,000 cells / mL. For the flat culture condition, the cell suspension was pipetted into four wells of a 12-well plate (1 mL into each well). |
67c221956dde43c908e8de13 | 16 | The growth medium for the flat and dECM inserts was refreshed 6-18 hours after seeding, and then every 2-3 days over 7 days of proliferation. To start differentiation, the growth medium was replaced with 1 mL of 50% growth medium and 50% differentiation medium composed of HepaRG differentiation medium supplement with antibiotics (ADD711C, Biopredic International, Saint-Gregoire, France) in Williams E medium with GlutaMAX supplement (32551020, Thermo Fisher, Pittsburgh, PA, USA). Afterward, the medium was replaced with 1 mL differentiation medium every 2-3 days. |
67c221956dde43c908e8de13 | 17 | Urea in phenol red-free cell differentiation medium was quantified using a colorimetric Urea Assay Kit III (MAK471, MilliporeSigma, St. Louis, MO, USA). Phenol red-free differentiation medium was composed of HepaRG differentiation medium supplement with antibiotics (ADD711C, Biopredic International) in phenol red-free Williams E medium (A1217601, Thermo Fisher, Pittsburgh, PA, USA) and GlutaMAX 100X supplement (35050061, Thermo Fisher, Pittsburgh, PA, USA) diluted to 1X final concentration. An aliquot of 0.75 mL of 24-hr culture medium was collected for urea quantification following the assay kit's manual. |
67c221956dde43c908e8de13 | 18 | The MTS Assay Kit (ab197010, Abcam, Cambridge, UK) was used to measure the metabolic activity of differentiated HepaRG cells 13-43 days after differentiation. For each well of flat or dECM insert culture, the differentiation medium was replaced with 500 µL of phenol red-free differentiation medium + 50 µL MTS Reagent Solution. A blank with 20 µL MTS Reagent Solution added to 200 µL of phenol red-free differentiation medium was also prepared. The cell culture and blank were incubated at 37 °C for 2 hours, rocking at culture and blank were pipetted into a 96-well plate, and the absorbance was measured on a plate reader (SpectraMax i3x, Molecular Devices, San Jose, CA, USA) at 490 nm. |
67c221956dde43c908e8de13 | 19 | For the intrinsic clearance data, the following literature was analyzed: lidocaine , clozapine . For reports with the unit per kg, the numbers were multiplied by 70 kg, the average weight of a healthy adult. To normalize the in vivo intrinsic clearance data in the aforementioned literature to per cell, the published results that 1 g human liver tissue contains 116x10 6 hepatocytes and an adult human liver weighs 1561g on average were used. |
67c221956dde43c908e8de13 | 20 | Cells cultured on flat surfaces and on dECM inserts were rinsed in PBS, lysed in 0.3 mL RIPA buffer (J52624, Thermo Fisher, Pittsburgh, PA, USA) diluted to 1X, and collected into microtubes followed by 90 sec of sonication (2 sec on, 4 sec off) using an ultrasonic homogenizer (Pulse 150, Benchmark Scientific, South Plainfield, NJ, USA). |
67c221956dde43c908e8de13 | 21 | The lysed cell solution was then centrifuged for 12 min at 13,200 rpm at 4 °C (5415 D, Eppendorf, Hamburgh, Germany). An aliquot of 10 µL of the supernatant was used in a Bradford assay (23246, Thermo Fisher, Pittsburgh, PA, USA) to quantify the cell number (OD=595 nm). A calibration curve between known cell numbers (EVE, NanoEntek, Seoul, South Korea) and the Bradford assay signals was prepared to elucidate cell numbers in the lysed samples. |
67c221956dde43c908e8de13 | 22 | Intrinsic clearance rates (CL) were calculated from experimental data (remaining concentrations as a function of time) with the following equations . The volume of media in the experiments, which was 750 µL for lidocaine and 1000 µL for clozapine, was used for V. The CL was then normalized to per cell based on the cell counting protocol above. |
67c221956dde43c908e8de13 | 23 | We aimed to create a scaffold biologically relevant to the liver ECM in terms of microstructures and biochemical composition. Literature suggests that the native liver ECM is a fibrous network, with estimated ECM fiber diameters ranging from 100 to 500 nm . Thus, we used electrospinning to generate nanofibers similar in thickness to those in the native ECM. Electrospinning transforms a polymer solution into nano-/microfibers by pumping the solution through a syringe needle at a high voltage, forming a Taylor cone . By varying parameters such as voltage, needle diameter, flow rate, and distance to the collecting platform, the fiber size, orientation, porosity, and layer thickness can be tuned . Polycaprolactone (PCL), a biocompatible polymer, was used as the base material to make the fibers . Meanwhile, cut pieces of porcine liver, which have proven to be similar to the human liver for various biological applications , were decellularized to collect the native decellularized ECM (dECM), which was then ground, digested by pepsin and lyophilized. This dECM was added to the PCL solution before electrospinning so that the fibers would contain the native ECM components. Ideally, a larger ratio of the native dECM to PCL would be more biorelevant. However, we observed that if the native dECM exceeded the amount of PCL (weight), the dECM started to precipitate and form layers while in the electrospinning solution. Therefore, we made the solution with 10% PCL (w/v) and 10% (w/v) dECM, which remained homogenous for electrospinning. |
67c221956dde43c908e8de13 | 24 | Figure is a Scanning Electron Microscopic (SEM) view of the composite (PCL + dECM) fibers, and Figure shows the fiber diameter distribution. As shown in Figure in the Supplementary Information, fibers were deposited onto a polystyrene (PS) sheet with precut holes. After electrospinning, the sheet with fibers was laser-cut again around the outside of each hole to generate circular inserts. What remained was a PS ring with a layer of nanofibers on top, fused by the heat generated from the laser cutting. A total of 38 dECM inserts could be prepared in one electrospinning session, which enhanced efficiency and enabled experiments to be conducted on inserts from the same batch, reducing possible variations. Figure compares two inserts: PCL-only fibers (left) and PCL + dECM fibers (right). The inserts were incubated with Picrosirius red, a common dye in histology to stain collagen . After thorough rinsing, the dECM insert remained red, compared to the PCL-only fibers which were white. Because Picrosirius red is moderately fluorescent , we scanned the inserts with a plate reader and observed fluorescent signal heatmaps on the fibers with dECM (bottom panel of Figure ). Further, Figure shows fluorescent microscopic views of the composite (PCL + dECM) fibers stained with Picrosirius red, and Figure shows staining by collagen I antibodies conjugated with CoraLite 488 of the composite fibers. Moreover, Fourier Transform Infrared (FTIR) spectroscopy was conducted over the fiber layers. As shown in Figure , plain PCL fibers showed clusters of peaks between 1000-1200 cm -1 and 1400-1500 cm -1 , mainly corresponding to -C-O stretching and -C-H bending in PCL. However, with the inclusion of the dECM in the electrospinning solution, the resulting fibers showed a broad FTIR peak above 3000 cm -1 (arrow), a typical signal of -OH groups. New peaks at 1540 cm -1 and 1640 cm -1 strongly indicate secondary amide (peptide bonds; inset of Figure ). |
67c221956dde43c908e8de13 | 25 | An insert holder was created to help seed cells onto the dECM insert and house them in a 12-well plate, as detailed in Figure . Figure demonstrates the assembly of the dECM insert into the 3D-printed insert holder device, with the bottom insert holder and a top hollow lid. After a circular insert was placed in the holder, a silicone O-ring was placed onto the insert before the hollow lid was screwed into the insert holder via the printed threads. |
67c221956dde43c908e8de13 | 26 | We evaluated HepaRG cells after 13-14 days of differentiation on the fibrous dECM inserts. As shown in Figure , although the number of cells on dECM appeared to be less than on flat, the difference was not significant. There were over 200,000 cells for both conditions. Figure shows confocal microscopic views (top and sides) of hepatocytes in the scaffold. Clearly, the cells could infiltrate by ~50 µm into the fibrous layer to form a 3D culture. Hepatocytes were characterized by the ability to synthesize albumin and produce urea, two key fundamental functions of the cells . As suggested by Figures and, cells on both culture substrates showed albumin and urea production, indicating successful differentiation of HepaRG progenitors to hepatocytes. Interestingly, we did not observe significant differences in albumin production from cells on flat vs. on dECM, suggesting that some biosynthesis pathways of HepaRG hepatocytes might not be affected by the ECM. In contrast, cells on dECM produced significantly more urea than on flat, indicating that the metabolic activities of the hepatocytes might be enhanced by the ECM. Further experiments aligned with this observation, where the metabolic activity as measured by the MTS assay (Figure ) was significantly higher from cells on dECM than on flat for up to 43 days. Between Day 13 and Day 20, cells on both substrates showed increased MTS metabolism (although not significant for cells on flat), and thus, cells within this time frame were used for the following studies. |
67c221956dde43c908e8de13 | 27 | To explore the scaffold's possible role in enhancing drug metabolism/clearance by the HepaRG hepatocytes, we chose drugs that 1) need to be metabolized by the liver before excretion, with limited direct excretion, 2) are well-studied and well-documented with abundant in vivo data available in the literature. The first drug tested was lidocaine 82 , a common anesthetic, which is mainly metabolized by CYP3A4 to monoethylglycinexylidide (MEGX), with less than 10% direct excretion (Figure ). The next drug we investigated was clozapine , an antipsychotic medication, which needs to be fully metabolized before excretion (Figure ). Clozapine is primarily metabolized by CYP1A2 and CYP3A4 to norclozapine. As with lidocaine, we observed faster clearance of the drug by the hepatocytes on dECM than on flat (Figure ). The error bars for the flat controls were large because we found the cells started to die with the clozapine, but not the cells on dECM, even with the same drug concentrations. As the key enzyme for fluoxetine metabolism (Figure ). These data highlight the necessity of including permeable excretion models to couple with liver models in future studies for efficient removal of metabolites, as occurs in vivo, for certain drug clearance studies. |
67c221956dde43c908e8de13 | 28 | We demonstrated that a composite nanofibrous scaffold composed of PCL and native porcine liver ECM components enabled enhanced metabolic activities of HepaRG hepatocytes compared to conventional 2D cultures. Combined with a modular 3D-printed holder device fitting in a standard 12-well plate, this hepatic model can easily be integrated into existing cell culture workflows. With this 3D cell culture model, we tested well-documented drugs with in vivo data and found that the 3D model significantly enhanced drug clearance vs. the 2D well-plate model. More importantly, the drug clearance times and rates on our model were comparable with published in vivo data without significant differences. This 3D model thus holds promises to facilitate hepatocyte studies during preclinical drug development, especially given that it reproduced the in vivo observations with a widely accessible cell source (HepaRG), rather than primary human hepatocytes or stem cells. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.