id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
652668e78bab5d20551ad1b7 | 32 | Figure : top 10 cells by energy obtained from the filtering criteria reported in 4; the cells are plotted as scatter points on a voltage vs capacity plane: above, the capacity by area is reported, to allow for a quick extrapolation to an arbitrary cell geometry, below, capacity by mass of active material (sum of PEAM and NEAM) is reported, for a more straightforward read. |
652668e78bab5d20551ad1b7 | 33 | The flexibility of VOLTA allows for its implementation to novel chemical spaces, too. Let us consider a battery system powering an autonomous sensor like an internet of things (IoT) device. Such devices are meant to be wireless and with low to zero maintenance requirements, since electronics have gradually reduced their nominal voltage to about 1.2 V, to decrease dissipation and consumption, and the trend is expected to continue, and reach 0.5 V. Low voltage batteries hence become useful to power them avoiding DC-DC voltage converters. |
652668e78bab5d20551ad1b7 | 34 | VOLTA can be used to perform low-voltage battery revealing tasks by applying the desired filtering criteria to the AMs, launch the cell building step and filter out the cells with undesired voltage parameters. The remaining ones can be scored, then, by capacity. The aim is to identify materials with higher capacity, even if their overall energy density is lower, due to a lower cell operating voltage. In the present work, the electrodes were first filtered with the same criteria as in the previous section, and were then coupled in cells. In order to select cells compatible with the desired application, the cells operating at a low enough average voltage (0.6 to 1.2 V) were selected. For this case study, a conventional organic liquid electrolyte was assumed, in order to exclude the virtual cells operating outside its electrochemical stability window. The chosen voltage extremes were 1 to 4.5 V to avoid the use of AMs operating outside the thermodynamic stability window of the electrolyte. Moreover, the voltage spread was limited to 0.6 V, as summarized in Table The top 17 cells with highest area capacity are reported in Figure , in the format "PEAM vs NEAM". The interested reader can refer to Table in SI for the MP entries corresponding to the reported materials. We observe a prevalence of cells with null voltage spread, otherwise visualised as an error bar on the y axis. Let us now go through some of the AMs proposed by VOLTA. |
652668e78bab5d20551ad1b7 | 35 | Li 0-1 TiNCl appears more than once: this electrode is labelled as observed, has high theoretical specific capacity (257 mA h g -1 ) and operates with a flat voltage profile at 1.34 V, thus rendering it an ideal IoT NEAM when coupled with PEAMs that operate close to 2 V. |
652668e78bab5d20551ad1b7 | 36 | The Li 0-1 CuO 2 electrode is also worth highlighting. Its high theoretical capacity of 261 mA h g -1 scores it at the 93rd percentile by specific capacity among the filtered electrodes, and it operates at a flat 4.58 V. It can be speculated that this material's frequent appearance in the top cells could be due to the fact that the space investigated requires it to be coupled with electrodes operating close to 3 -3.5 V. As we pointed out above, the recent research on battery materials is probably biased towards high energy materials, hence it could be that the materials present in the MP database operating at 3 -3.5 V tend to also have higher theoretical capacity than the electrodes operating at lower voltages. This is consistent with what can be observed in Figure . The Figure displays a kernel density estimation of the bivariate distribution of the MP electrodes versus two variables: their theoretical specific capacity and the average operational voltage. First of all, it is interesting to visualize how skewed the distribution is towards the electrodes with both high capacity and high operational voltage, leaving the high capacity, low operational voltage space largely unexplored. This could be explained by a research bias towards high energy PEAMs, that naturally fit in the high voltage, high capacity space, and it helps in rationalizing why we observe multiple appearances of the same electrodes in Figure . It becomes clear that when seeking cells with average voltage between 0.6 and 1.2 V, the two electrodes Li 0-1 CuO 2 (as PEAM), Li 0-1 TiNCl and Li 0-1 VS 2 (as NEAM) can only be matched with other AMs falling within the highlighted regions. Such a region for the former material contains more AMs with high specific capacity, resulting in a higher probability of obtaining low-voltage cells with high capacity. This, in turn, results in having multiple highest-capacity cells containing this very electrode. The low-voltage cells obtained with Li 0-1 CuO 2 would also naturally operate at high extreme voltages, being both Li 0-1 CuO 2 and its low-voltage counter NEAMs operational around 4 V. |
652668e78bab5d20551ad1b7 | 37 | In Li 0-1 VO 2 vs Li 0-1 VS 2 the low cell voltage is purely achieved by means of inductive effect [44] on the materials' Fermi levels, the sulphide having a lower one and behaving as NEAM: an elegant solution to the low voltage problem. A similar phenomenon is highlighted in Li 0-2 V(CO 3 ) 2 vs Li 0-1 VS 2 yielding a lower cell voltage than the couple oxide-sulphide although reducing V with two electrons instead of one. |
652668e78bab5d20551ad1b7 | 38 | The cell Li 0-1 VO 2 vs Li 0-1 VS 2 is worth considering further: its theoretical cell viability is 1, indicating that three out of the four end-member phases are labelled as having been observed. The only virtual phase is the delithiated, layered PEAM VO 2 (mp-714944). An analysis of the present literature shows that the delithiation of layered (R 3m) LiVO 2 was first attempted by de Picciotto et al. in 1984, demonstrating "a shift from trigonal to near-cubic lattice constants" upon delithiation, explained by the authors as a migration of the vanadium ions, that makes it an impractical electrode material. The same process has been studied more recently by Chable et al . A possible experimental way to circumvent this issue could be the introduction of a support cation, to hinder V anions migration, according to the work of Ma et al. Such a strategy was validated in the same work with the use of LiCr x V 1-x O 2 with x ranging from 0.1 to 0.5, and the authors also point out the partial substitution of V with Ga(III) and Al(III). The improvement to cyclability is crucial, but it comes at the expenses of capacity, yielding a practical discharge capacity at the tenth cycle of about 100 mA h g -1 , to be compared to the theoretical capacity of the complete delithiation of R 3m LiVO 2 amounting to 298 mA h g -1 . |
652668e78bab5d20551ad1b7 | 39 | Figure : top 17 low-voltage virtual cells by area capacity. As in Figure , area capacity was chosen to allow extrapolation of the result to arbitrary cell geometries (above). Capacity per mass of active material (sum of both PEAM and NEAM) is also reported Figure : distribution of MP electrode entries in a specific capacity vs voltage plane; the distribution is displayed with contour lines obtained with kernel density estimation, showing the density of datapoints in the specific capacity vs voltage space. Given the AM pairing rules reported in table 5, the three AM highlighted can only be paired with AMs that operate at 0.6 and 1.2 V more (or less). For the sake of graphical simplicity, we assume Li 0-1 TiNCl and Li 0-1 VS 2 to only behave as NEAMs and Li 0-1 CuO 2 as PEAM, i.e. to seek counter-AMs only in the voltage regions above and below themselves, respectively. |
652668e78bab5d20551ad1b7 | 40 | We have presented the battery screening capabilities of VOLTA, a Python software developed in the present work. VOLTA allows for a simple, back-of-the-envelope style of screening that can applied to any ensemble of electrode entities. The core concept is to screen battery materials indirectly, allowing the researcher to focus on the desired cell properties. The software also helps the user in assessing the feasibility of the constructed cells by producing ad-hoc metrics and approximates the real-world performance of the composite electrodes by relying on the physics-based manufacturing model of the ARTISTIC project. VOLTA relies heavily on the data it is input with. A battery revealing task can be conducted with DFT-based data, like in the present work. This approach has several advantages: the data is potentially uniform, as the energetic properties can all be calculated with the same density functional, and the data collection is relatively inexpensive, especially when collecting DFT data from online databases like the MP. This in turn can result in a "cutting-edge" battery revealing task, limited only by the chemical diversity of the AM dataset. |
652668e78bab5d20551ad1b7 | 41 | On the other hand, one has to keep in mind that the theoretical properties of the AMs do not necessarily reflect those of the practical AMs, supposing they can be synthesised in the first place. A clear example is theoretical capacity, which can differ substantially from the experimental one. Moreover, the assumption of equal particle size distributions for all AMs introduced in the ARTISTIC simulation step is likely affecting the accuracy of the calculation of the electrode porosity and thickness (see bullet list in Section 3). This assumption is needed as, to the best of our knowledge, there is no way to reliably correlate the materials chemistry with their particle size distributions, which are also heavily dependent on the synthesis conditions. Once a reliable correlation were to be established, the full potential of the ARTISTIC project platform could be leveraged by simulating electrode manufacturing processes for variable particle size distributions. |
652668e78bab5d20551ad1b7 | 42 | The accuracy of VOLTA can be pushed even further by using experimental cycling data, more precisely extreme voltages and associated capacity. This would result in more realistic cell voltages and capacities, and also circumvent the need for theoretical viability metrics for the materials experimental data is available for, as they must have been synthesised to gather the cycling data in the first place. The use of observed materials only, with their respective measured voltage and capacity data would thus allow for more pragmatic battery revealing tasks. This would come to the cost of reducing the screening to less AMs, and of needing extra care in collecting the experimental data in conditions as uniform as possible across different AMs. Moreover, improving the stability metric SM to a more thorough synthesizability assessment for each virtual AM would be a valuable addition to VOLTA, too. This would allow for a more accurate assessment of the practical viability of virtual cells that might be identified as particularly desirable. |
652668e78bab5d20551ad1b7 | 43 | We should lastly point out a potential use of the algorithm for educational purposes: the very nature of the algorithm makes it a tool for exploring the links between the chemical nature of active materials and their performance as battery active materials; something a newcomer to battery science might be interested looking into. |
62bb4105f51939ae3e771354 | 0 | All tert-butyl phenylcyanoacrylates (TBCA) compounds were synthesized by Knoevenagel condensation of appropriate benzaldehydes with tert-butyl cyanoacetate, catalyzed piperidine (Scheme 1). The preparation procedure was essentially the same for all the monomers. In a typical synthesis, equimolar amounts of tert-butyl cyanoacetate and an appropriate benzaldehyde were mixed in equimolar ratio in a 20 mL vial. A few drops of piperidine were added with stirring. The product of the reaction was isolated by filtration and purified by crystallization from 2-propanol. The compounds were characterized by IR, 1 H and 13 C NMR, and elemental analysis. No stereochemical analysis of the novel compounds was performed since no stereoisomers (E or/and Z) of known configuration were available. |
62bb4105f51939ae3e771354 | 1 | Yield 96%; mp 94.1°C; Nitrogen elemental analysis showed that between 11.6 and 25.4 mol% of TBCA is present in the copolymers prepared at ST/TBCA = 3 (mol), which is indicative of relatively high reactivity of the TBCA monomers towards ST radical which is typical of alkoxy ring-substituted TBCA. Since TBCA monomers do not homopolymerize, the most likely structure of the copolymers would be isolated TBCA monomer units alternating with short ST sequences (Scheme 2). |
6216933d91a2e60d14d94b41 | 0 | Poor aqueous drug solubility, and the resulting low bioavailability and potential lack of therapeutic effect, is a major challenge in oral drug delivery. One strategy to increase solubility and dissolution rate is conversion of the crystalline drug into its amorphous form. However, the amorphous forms are thermodynamically unstable and require stabilization to avoid recrystallization during storage or after in vivo administration. The drug is therefore commonly formulated as an amorphous solid dispersion (ASD) in which it is molecularly dispersed in a polymeric network. The polymer in the ASD stabilizes the drug in the solid state and thereby inhibits recrystallization during storage. Further, it may prevent (or delay) drug precipitation upon dissolution and improve solubility. This results in fast dissolution where the subsequent maintenance of supersaturated drug concentrations in vivo can drive absorption from the gastrointestinal tract. Oral dosage forms comprising ASDs have been successfully marketed, e.g. Venclexta, which is used to treat diseases such as chronic lymphocytic leukemia, and acute myeloid leukemia, and contains venetoclax as the active pharmaceutical ingredient and polyvinyl alcohol as the enabling excipient. |
6216933d91a2e60d14d94b41 | 1 | However, despite the proven advantages of ASDs, several short-comings hinder the successful translation of more ASD-based dosage forms to the market. The drug load in stable ASDs is typically low (10 -30 wt%) as it is limited by the drug solubility in the polymer at storage temperature, this in turn means that the size of the dosage form increases when a higher drug dose is required, with negative implications for the patient. Furthermore, many polymers commonly used as ASD excipients, such as polyvinylpyrrolidone (PVP) are hygroscopic. Such excipients increase the molecular mobility in the ASDs due to the plasticizing effect of adsorbed water, and hence, can induce amorphous-amorphous phase separation and ultimately recrystallization of the drug during storage. Finally, amorphous powders exhibit poor flowability in pharmaceutical tableting equipment. As a result, ASD manufacturing requires additional processing steps, such as granulation, which increases the complexity and cost of production. |
6216933d91a2e60d14d94b41 | 2 | In situ amorphization has been recently introduced to overcome the challenges of drug load, stability and manufacturing of ASDs. In situ amorphization can overcome storage stability issues of ASDs and allows for higher drug loadings (up to 50 wt%) than what is dictated by the thermodynamic solubility limit of the drug in the polymer at storage temperature since storage times can be kept short. Amorphization and ASD formation take place in the final dosage form either directly after tablet manufacturing or before administration to the patient, e.g. at hospital pharmacies. The tablet for in situ amorphization is produced by established pharmaceutical tableting protocols such as direct compaction of a crystalline drug in a polymer mixture. The ASD is then formed in situ by exposing the tablet to a radiation source, e.g. microwave or laser. This is a time-and temperature-dependent process, whereby the radiation increases the tablet temperature and the drug dissolves in the polymer at temperatures above the glass transition temperature (Tg) of the polymer. Amorphization by microwave radiation requires excipients, such as glycerol or water in the tablets. Large amounts (approx. 20 wt%) of the enabling excipients have to be added to reach complete amorphization, which hampers the mechanical properties of the tablet. The laser-induced approach uses plasmonic nanoparticles (Ag) as the enabling excipient and only requires as low as 0.1 wt% of them in the tablets. However, the penetration depth of the laser in the tablet is limited and not uniform, making the scale-up difficult. Furthermore, the use of plasmonic nanoparticles maybe toxic upon repeated administration to patients due to their dose dependent toxicity. Thus, there is a need to identify other effective and safe in situ amorphization excipients. |
6216933d91a2e60d14d94b41 | 3 | Superparamagnetic iron oxide nanoparticles (SPIONs) could overcome the shortcomings of current in situ amorphization methods by using magnetic hyperthermia. SPIONs (Fe3O4 or γ-Fe2O3) are non-toxic, biocompatible, biodegradable and efficiently cleared from the body via the iron metabolism pathway. They are approved by the US Food and Drug Administration (FDA) as oral magnetic resonance imaging contrast agents (Ferumoxsil) and for treatment of anemia (Ferumoxytol) in patients with chronic kidney disease. SPIONs release heat locally upon exposure to an alternating magnetic field (AMF) due to relaxation losses. The heat dissipation from SPIONs on exposure to an AMF is the result of relaxation of the magnetic moment within the particle (Néel relaxation) or the rotation of the particle itself (Brownian relaxation). So far, this property of SPIONs, commonly termed magnetic hyperthermia, has been used for treatment of cancer and triggered drug delivery. The heat dissipation properties can be enhanced by careful engineering of the SPION properties, e.g., with unidirectional growth of nanoparticles, doping with metals, size optimization, and formation of nanocrystal clusters. In particular, spinel crystal nanostructures of e.g. iron oxide doped with Zn 2+ or Mn 2+ release more heat than pure iron oxides. In this work, we demonstrate for the first time the use of SPIONs to induce amorphization of a poorly aqueous soluble crystalline drug (celecoxib), in tablets for oral administration (Figure ). Celecoxib, a cyclooxygenase-2 inhibitor widely used to treat osteoarthritis and rheumatoid arthritis, was chosen as the model drug due to its low aqueous solubility and poor bioavailability. |
6216933d91a2e60d14d94b41 | 4 | Because of this, high doses are required for oral administration, which causes adverse side effects in the gastrointestinal tract. The doped SPIONs were produced by flame spray pyrolysis (FSP), a bottom-up nanomanufacturing technique with proven scalability and reproducibility. The structural, morphological, magnetic and heating properties of the FSP-made SPIONs were characterized and their cytotoxicity on human Caco-2 intestinal cells assessed. A Design of Experiments (DoE) approach was applied to systematically investigate the effects of nanoparticle and tablet composition, and AMF exposure time on the degree of drug amorphization in the tablets. |
6216933d91a2e60d14d94b41 | 5 | Nanoparticle Properties. Zinc and manganese ferrites were prepared by flame spray pyrolysis (FSP) and investigated as enabling excipients for in situ amorphization in tablets containing celecoxib. Flame process parameters were set to produce nanoparticles with primary particle sizes in the range of 12-20 nm. This size range has been reported, both experimentally and computationally, to result in maximum heating efficiency during hyperthermia. Table summarizes nanoparticle size, magnetization, and heating efficiency in an AMF of as-produced SPION powders. Pure iron oxide is included for comparison. Figure shows the X-ray diffraction (XRD) patterns of pure γ-Fe2O3 and SPION doped with Zn 2+ and Mn 2+ to form the ferrites, Zn0.5Fe2.5O4 and Mn0.5Fe2.5O4. The diffractograms correspond to spinel cubic structures, exhibiting six prominent peaks originating from the (220), (311), (400), (422), (511) and (440) crystallographic planes. Pure iron oxide forms γ-Fe2O3 (maghemite), in agreement with previous SPION synthesis in oxygen-rich flames. Spinel-ferrite structures were formed by the addition of Zn 2+ or Mn 2+ , as evidenced by a slight shift of diffraction peaks towards lower angles (shown in Figure as the dashed line at the (311) plane). The peak shift is accompanied with a lattice expansion (Table ), indicating the successful incorporation of Zn and Mn 2+ ions into the γ-Fe2O3 crystal lattice. Furthermore, no diffraction peaks corresponding to pure zinc oxide or manganese oxide were observed. The average crystallite size (dXRD) was about 14 nm for both γ-Fe2O3 and Zn0.5Fe2.5O4 and 18 nm for Mn0.5Fe2.5O4 (Figure ). The larger ionic radius of Mn 2+ (0.81 Å) compared to Fe 3+ (0.64 Å) and Zn 2+ (0.74 Å) at the tetrahedral site might explain the increase in crystal size for the manganese ferrite. Figure and 2c show representative transmission electron microscopy (TEM) images of Zn0.5Fe2.5O4 and Mn0.5Fe2.5O4 nanoparticles, respectively. The nanoparticles are polyhedrons with a narrow primary particle size distribution (Figure ) in excellent agreement with the literature. The primary particle size determined from TEM analysis (dTEM) was about 18 nm for both ferrites, in good agreement with the dXRD and thus suggesting primarily monocrystalline particles. The dTEM also falls within the targeted optimal range of SPION particle size (12-20 nm) for high heating rates. The geometric standard deviation (σg) was 1.48 and 1.52 for Zn0.5Fe2.5O4 and Mn0.5Fe2.5O4, respectively, and thus corresponds to the theoretical self-preserving size distribution that is typically attained during flame synthesis. 39 Magnetic Properties and Heating Efficiency. Figure shows the magnetization of flamemade SPIONs. The nanoparticles exhibit nearly zero hysteresis confirming their superparamagnetism. At the maximum field of 1000 mT, nearly all particles reached their saturation magnetization. Bulk maghemite has a saturation magnetization of 80 emu g -1 , which is considerably higher than the 53.7 emu g -1 measured for γ-Fe2O3 here (Table ). This can be attributed to the small particle size that increases the thickness of the surface spin disordered layer, also called the "magnetically dead layer". Mn0.5Fe2.5O4 exhibits the highest saturation magnetization (66.7 emu g -1 ) of the flame-made SPIONs here, in agreement with the literature. The dopant occupies the tetrahedral or octahedral sites in the structure of γ-Fe2O3, which affects the magnetocrystalline anisotropy and increases the overall magnetic moment of the unit cell, thereby also affecting the saturation magnetization of the SPIONs. Surprisingly, the incorporation of Zn 2+ in γ-Fe2O3 did not enhance the saturation magnetization in contrast to previous reports. This could be due to the smaller size of the zinc ferrite compared to manganese ferrite and hence the increased relative volume contribution of the magnetically dead layer to the total particle volume for the former. Moreover, the high substitution of Fe 3+ ions by Zn 2+ ions at the tetrahedral site also weakens the exchange coupling between the tetrahedral and octahedral sites, thus decreasing the total magnetic moment of the unit cell. Figure shows the heating efficiency (represented as the temperature increase ΔT after 10 seconds) of the nanoparticle powders in an AMF. The heating efficiency was directly measured on the bulk powders (in contrast to how particle suspensions are typically measured) as they were used here in their dry state in the final tablet formulations, hence the heat dissipation is primarily originating from Néel relaxation of the SPIONs. The zinc, and especially the manganese ferrite, both generated significantly higher temperatures than γ-Fe2O3, in agreement with previous reports. Heating efficiency of SPIONs is often correlated with the saturation magnetization, and this was especially obvious for Mn0.5Fe2.5O4 (Figure ). Furthermore, the heating efficiency increased with increasing magnetic field amplitude, also in agreement with previous studies. Overall, these results demonstrate the superior heating efficiency of Zn and Mn ferrites compared to γ-Fe2O3. We therefore investigated these two ferrites further for the magnetic hyperthermia-induced in situ amorphization. ). b) IR images captured at four time points (i: 0 min, ii: 0.5 min, iii: 1 min, iv: 6 min). The color scale indicates the temperature of the tablet. c) XRD diffractograms and d) DSC thermograms of tablets containing 30 wt% celecoxib and 20 wt% Mn0.5Fe2.5O4 before and after 3 and 15 minutes of AMF exposure time. |
6216933d91a2e60d14d94b41 | 6 | Magnetic Hyperthermia-Induced In Situ Drug Amorphization in Tablets. Tablets were prepared based on the formulation design space (Table ) constructed using DoE to investigate factors for magnetic hyperthermia-induced drug amorphization. Table summarizes the tablet compositions, process conditions and the observed degree of amorphization. The tablet containing 30 wt% celecoxib and 20 wt% Mn0.5Fe2.5O4 exposed to AMF for 15 min (tablet 26, Table ) showed maximum amorphous content induced by in situ amorphization, and is thus presented here in detail (Figure ). Figure shows the temperature profile of this tablet and the thermal images recorded with IR camera at four time points are shown in Figure . The tablet surface temperature rapidly increased from room temperature to 154 °C within the first three minutes of AMF exposure. The tablet temperature further increased to a maximum (Tmax) of 165 °C after 6 minutes and was maintained there for the remaining time of AMF exposure. As the SPIONs are homogenously distributed in the matrix, hyperthermia results in uniform heating of the entire tablet, as shown here by the mono-colored thermal images (Figure ). The uniform heating ensures complete drug amorphization in the tablets. This is a clear advantage of magnetic hyperthermia-induced amorphization over the previously reported laser-radiation approach. The latter can only be applied to thin tablets due to the limited penetration depth of the laser. Some of the tablets showed deformation after AMF exposure due to water evaporation and increased polymer mobility above the Tg of PVP (here Tg = 80 -85 °C). A significant decrease in viscosity of the polymer has been observed at approximately 15-25 °C above the Tg of the polymer, enabling the drug dissolution and amorphization process. Here, the onset temperature for drug amorphization was determined by differential scanning calorimetry (DSC) and was found to be 103.6, 117.7, and 117.6 °C for a drug load of 30, 40, and 50 wt% respectively (Tg + (15-25) °C depending on wt% of celecoxib). At temperatures > Tg, the polymer viscosity decreases leading to a drastic increase of molecular mobility in the network. Thus, the drug dissolves easily into the polymer and in situ amorphization is initiated. The AMF-exposed tablet in Figure significantly exceeded the onset temperature required for amorphization of celecoxib and was maintained above that temperature for an adequate time, thus ensuring an efficient in situ amorphization by hyperthermia. |
6216933d91a2e60d14d94b41 | 7 | The degree of amorphization after AMF exposure of the tablets was quantified by XRD and DSC. Figure shows the XRD diffractograms of tablet 26 as prepared and after 3 and 15 minutes of AMF exposure. The as prepared tablet displays the characteristic diffraction peaks of crystalline celecoxib. The intensity of the crystalline peaks diminished already after three minutes of AMF exposure. Complete amorphization was achieved after 15 minutes, as indicated by the "halo" characteristic of an amorphous material. Similarly, the depressed melting endotherm of crystalline celecoxib in the presence of PVP at about 140 °C in the as-prepared tablets was no longer detected by DSC after AMF exposure (Figure ). The successful amorphization is evident in the DSC thermograms from a single glass transition (Tg) of the polymer after in situ amorphization (Figure ). |
6216933d91a2e60d14d94b41 | 8 | A multiple linear regression model (Table ) was generated from the tablet data (Table ) to investigate the effects of process parameters on drug amorphization. Response modeling resulted in high R 2 and Q 2 values for degree of drug amorphization (R 2 = 0.917, Q 2 = 0.842) and Tmax (R 2 = 0.906, Q 2 = 0.865). Moreover, the predicted and observed values showed strong correlation (Figure ). It should be noted that an experiment outside the action limit of ± 4 standard deviations was identified as an outlier based on the deleted studentized residuals, and removed from the final regression model. Table . Adjusted models for degree of drug amorphization and maximum tablet temperature (Tmax). SPION content in the tablets had the strongest positive influence on tablet temperature (Table ). For all experiments, tablet temperature initially increased rapidly with exposure time until it plateaued (Table ) and Tmax was then maintained until the end of AMF exposure, which is also shown in tablet 26 (Figure ). Mn ferrites showed a much stronger influence on tablet temperature than Zn ferrites. Furthermore, Tmax showed a strong linear relationship to the extent of celecoxib amorphization (r = 0.96). Bold: factors with the most influence on the given response; b) Italics: non-significant parameters retained to conserve either the hierarchy or the predictability of the model. SPION composition had the strongest influence on the degree of drug amorphization (Table ). |
6216933d91a2e60d14d94b41 | 9 | The contour plots for Zn0.5Fe2.5O4 and Mn0.5Fe2.5O4 (Figure ) show the extent of drug amorphization as a function of AMF exposure time, drug load and SPION content. Red areas in the contour plot illustrate process conditions that achieved >90% degree of amorphization. Mn0.5Fe2.5O4 was more effective than Zn0.5Fe2.5O4 in formation of completely amorphous ASDs. This can be attributed to the better heating efficiency of Mn0.5Fe2.5O4 (Figure ), resulting in higher tablet temperatures and faster dissolution of celecoxib into the PVP network. The SPION content also positively affected the degree of drug amorphization with no significant quadratic term, indicating a linear effect of this parameter in the entire design space. However, the SPION content showed a significant interaction term with the SPION composition. In contrast, the tablet drug load had a negative influence on drug amorphization. Thus, the tablets with 50 wt% drug load showed a lower degree of amorphization compared to 30 wt% at similar exposure times and temperatures. This follows the Noyes-Whitney equation that predicts that for a higher drug load, a higher Tmax can result in complete dissolution in a solvent, here the polymer. Higher temperatures increase the saturation solubility of the drug in the polymer, lower the viscosity of the continuous (polymer) phase, and increase the diffusion coefficient according to the Stokes-Einstein equation. This in turn results in a faster dissolution rate. In fact, for tablets 27 and 28 with a high SPION concentration of 20 wt% (Table ), celecoxib was not completely amorphized, even though temperatures reached above the onset temperature of dissolution. This could have been a result of the high drug load (50 wt%) that might require longer exposure times (>15 minutes) than applied here to result in complete drug amorphization, since dissolution is a time and temperature dependent process. It should be noted that at high drug loads, the drug becomes the bulk of the tablet, resulting in incomplete dissolution of the drug into the polymer also due to insufficient polymer content. |
6216933d91a2e60d14d94b41 | 10 | The AMF exposure time was not a significant factor for drug amorphization in the selected design space; however, the model term for the duration of AMF was retained to improve the predictability of the model. The contour plot for tablets containing Zn ferrite (Figure ) indicate that the tablets that failed to reach the onset temperature (Table ) show minimal amorphization despite prolonged AMF exposure. Thus, a minimum AMF exposure time of three minutes was selected to ensure at least some degree of drug amorphization. As a result, the AMF exposure time is a non-significant factor in the current model. In future studies, the effect of time on hyperthermia-induced drug amorphization could be investigated by increasing the range and interval of the design. As previously discussed, heating efficiency of SPIONs depends strongly on the operating AMF amplitude. The maximum AMF amplitude used in this study was confined due to the instrument limitations, however using higher AMF amplitude could potentially decrease the time required for AMF exposure and the SPION content required to achieve complete amorphization. Nanoparticle Cytotoxicity on Caco-2 Cells. The most effective hyperthermia-induced drug amorphization was achieved using the Mn0.5Fe2.5O4 nanoparticles (Figure ). However, the final choice of an enabling excipient for in situ drug amorphization must also be guided by the potential toxicity of the nanomaterial. Figure shows the dose-dependent viability of human intestinal Caco-2 cells following 24 hours of exposure to the SPIONs. Cell viability after nanoparticle exposure is highly dependent on cell line and differentiation state, exposure time and concentration, and physicochemical properties of the nanoparticles. Here, cell viability was assessed using non-differentiated cells, which are more sensitive than differentiated ones. In other words, our model is highly sensitive compared to the in vivo condition, and better able to detect any toxic effects exhibited by the flame-made SPIONs. The cell viability was not affected by γ-Fe2O3 and Mn0.5Fe2.5O4 exposure at 100 μg mL -1 (corresponding to an oral dose of 25 mg SPIONs ingested with a 250 mL glass of water) and was only slightly decreased at higher concentrations. This declining trend of cell viability with increasing particle concentration (Figure ), suggests a dose-dependent toxicity for all particles. The minor adverse effect of iron oxide nanoparticles on Caco-2 cell viability has also previously been reported. In contrast, Zn0.5Fe2.5O4 dramatically reduced cell viability to 75% compared to the control already after exposure to 100 μg mL -1 and further to 63% at 200 μg mL -1 . Mn0.5Fe2.5O4 on the other hand shows significantly higher cell viability than zinc ferrite. Previous studies have also reported relatively higher cytotoxicity of zinc ferrite compared to manganese ferrite, which can be attributed to the oxidative stress induced by generation of reactive oxygen species by the former. Overall, the high heating efficiency and low cytotoxicity of the Mn0.5Fe2.5O4 nanoparticles suggest they are efficient as enabling excipients for hyperthermia-induced in situ amorphization. In Vitro Dissolution Studies. In vitro dissolution assays of celecoxib ASDs in biorelevant intestinal media under non-sink conditions have previously shown good correlation with their in vivo performance. We therefore used them to benchmark the in situ amorphized ASD. Figure shows the in vitro dissolution profiles of crystalline celecoxib, amorphous celecoxib, conventional ASD (prepared by melt quenching) and tablets containing Mn0.5Fe2.5O4, before and after magnetic hyperthermia-induced amorphization. The conventional ASD showed a considerably higher dissolution rate and maximum drug concentration (Cmax) than pure crystalline or amorphous celecoxib. This is in good agreement with previous studies demonstrating the superior dissolution characteristics of ASDs compared to pure compounds. The dissolution profile of ASDs are mainly controlled by generation and stabilization of a supersaturated solution, this is commonly referred to as the 'spring and parachute effect', and suggests that the performance of an ASD is governed by the dissolution rate and crystallization inhibition of the polymer. PVP maintains the balance between the dissolution rate enhancement and precipitation inhibition. The tablets containing Mn ferrites achieved a similar dissolution profile after AMF exposure as the conventional ASD, reaching a Cmax of 239 μg mL -1 in 15 min. This concentration was significantly higher than Cmax of conventional ASD (215 μg mL -1 ) and approximately five times higher than the Cmax of the SPION-containing tablets before AMF exposure (46 μg mL -1 ). At the end of sampling period (6 hours), the AMF-exposed tablets maintained a higher solubility compared to the conventional ASD and the crystalline celecoxib (Figure ). The presence of crystalline nanoparticles in the formulation could act as a crystalline nuclei and trigger re-crystallization, however Mn ferrites did not trigger re-crystallization, thus further affirming their suitability as an in situ amorphization enabling excipient. Overall, this confirms that the novel magnetic hyperthermia-induced amorphization produces ASDs with the same performance as conventional ASDs in terms of their dissolution behavior. |
6216933d91a2e60d14d94b41 | 11 | This work demonstrates the ability of SPIONs to induce in situ amorphization in tablets intended for oral administration. Upon exposure to an AMF, elevated temperatures were achieved in SPION-containing tablets, leading to high amorphization and improved in vitro dissolution. The degree of amorphization was strongly linked to the maximum tablet temperature, which depended significantly on the SPION composition, SPION content and the drug load. Complete amorphization and high apparent solubility were achieved in tablets with 20 wt% Mn0.5Fe2.5O4 and 30 wt% celecoxib after 15 minutes of AMF exposure. Furthermore, unlike laser and microwaveinduced amorphization, these tablets are homogenously exposed to the AMF and use safer excipients that are less likely to cause drug degradation. The in situ amorphization can be finetuned according to the obtained DoE model and the method can be widely applied to ASDs with other polymer-drug compositions. The use of a biocompatible excipient produced by a scalable synthesis technique, along with the high drug load in ASDs produced by in situ amorphization, make the developed method a promising tool for broader use of ASDs as enabling formulations in oral drug delivery. |
6216933d91a2e60d14d94b41 | 12 | The precursor solutions were stirred for at least 1 hour at room temperature. Subsequently, the precursor was fed at 6 mL min -1 and dispersed using 3 L min -1 O2 (> 99.5%, Linde AGA Gas AB, Sweden) at constant pressure (1.6 bar). The flame was ignited by a premixed supporting flame of CH4 and O2 (> 99.5%, Linde AGA Gas AB) at flow rates of 1.5 L min -1 and 3.2 L min -1 , respectively. A 5 L min -1 O2 sheath gas was fed through the outermost sinter metal plate of the FSP burner. Gas flow rates were controlled with calibrated mass flow controllers (Bronkhorst, the Netherlands). The particles were collected on a glass fiber filter (Albert LabScience, Germany) with the aid of a Mink MM 1144 BV vacuum pump (Busch, Sweden). |
6216933d91a2e60d14d94b41 | 13 | Particle Characterization. The specific surface area was determined by nitrogen adsorption (Brunauer-Emmett-Teller; BET method) at 77 K using a Tristar II Plus (Micromeritics, USA) after degassing for at least 3 hours at 110 ℃ under a flow of nitrogen gas. X-ray diffraction (XRD) patterns were obtained at ambient temperature with a MiniFlex X-ray diffractometer (Rigaku Europe, Germany) using Cu Kα1 radiation (1.5406 Å) and at 40 kV and 15 mA. The patterns were recorded between 10° and 80° 2θ at a step size of 0.01° 2θ and a scan speed of 2.00 deg min -1 . A DTex detector was used in narrow energy gap mode to suppress iron fluorescence background from the copper radiation. XRD data were analyzed using PDXL2 software (Rigaku Europe, Germany). All patterns were normalized relative to the intensity of the peak corresponding to the (311) crystal plane. The average crystal size (dXRD) of SPIONs was calculated by Rietveld refinement analysis using the PDXL2 software. Particle morphology was visualized by transmission electron microscopy (TEM) with Tecnai™ G2 Spirit BioTwin, (Thermo Fisher/FEI, USA) operating at 80 kV. The samples were suspended in 99.5% ethanol and deposited (as a 5 µL drop) on a 200-mesh copper grid (Ted Pella, USA) coated with formvar and carbon. The primary particle size distribution was measured by counting 360 and 394 particles of Zn0.5Fe2.5O4 and Mn0.5Fe2.5O4, respectively, using ImageJ software. The particle magnetization was recorded on a vibrating sample magnetometer (Lake Shore, USA) and a Physical Property Measurement System (Quantum Design, USA). Magnetization versus magnetic field was measured in the field range ±1000 mT at a constant temperature of 25 °C. The thermal dissipation of nanoparticle powders was measured using an oscillating magnetic field apparatus (Magnetherm; NanoTherics Ltd., UK). |
6216933d91a2e60d14d94b41 | 14 | The heat dissipation from the bulk powders was measured with a fiber optic probe. Approximately 25 mg of each SPION powder was transferred to a 2 mL glass vial and placed inside a coil with nine windings. The nominal oscillation frequency was set to 588.5 kHz and the magnetic field strength to 14 mT. The heating efficiency was calculated as the rise in temperature in the initial 10 seconds. |
6216933d91a2e60d14d94b41 | 15 | Cell viability assay was performed to assess the cytotoxicity of SPIONs. Cell culture media and reagents were purchased from Thermo Fisher Scientific (USA) or Sigma-Aldrich (USA). Caco-2 cells (originally obtained from the American Type Culture Collection), passage 95-105, were maintained in Dulbecco's modified Eagle's medium, containing 10% (v/v) fetal bovine serum and 1% (v/v) nonessential amino acids. The cells were cultured in a humidified incubator (at 37 °C, 10% CO2) in 75 cm 2 tissue culture flasks. Stock suspensions of the SPIONs in Milli Q water were prepared at 10 mg mL -1 for γ-Fe2O3 and Mn0.5Fe2.5O4 and at 2 mg mL -1 for Zn0.5Fe2.5O4. The suspensions were prepared with a cup horn ultrasonicator (Sonics, USA) until no change in hydrodynamic diameter was observed. The hydrodynamic diameter was measured by dynamic light scattering (Litesizer 500, AntonPaar, Austria). The stock suspensions were diluted further with cell culture medium to achieve concentrations of SPIONs at 100, 150, and 200 μg mL -1 . |
6216933d91a2e60d14d94b41 | 16 | Caco-2 cells were plated into black opaque 96-well plates at a density of 5 x 104 cells per well in 300 µL of culture medium. The cells were allowed to attach for 24 hours before treatment. Subsequently, culture medium was replaced by 100 µL particle suspensions in six replicates per treatment and incubated for 24 hours (at 37 °C, 10% CO2). Positive controls were prepared by diluting 10% (w/v) sodium dodecyl sulfate (SDS) in water to achieve a final concentration of 0.22% (v/v) SDS in the cell culture medium. Culture medium was used as negative control. The water content in all treatments was kept at or below 2%. Finally, the viability of Caco-2 cells was evaluated with the CellTiter-Glo Luminescent Cell Viability Assay (Promega, USA). The luminescent signal from each well was determined with a plate reader (Tecan, Switzerland). A two-way analysis of variance (ANOVA) using Tukey's multiple comparisons test was used to compare the groups. Data analysis was performed using GraphPad Prism 9.0 software (La Jolla, CA, USA). p values were calculated as 0.1234 (ns), 0.0332 (*), 0.0021 (**), 0.0002 (***), and <0.0001 (****). |
6216933d91a2e60d14d94b41 | 17 | Tablet preparation and hyperthermia-induced drug amorphization. A DoE approach with a D-optimal design, resembling two, parallel central composite face (CCF; Figure )-centered fractional factorial designs, was used to study the effect of critical process parameters and their interactions on the Tmax and the degree of drug amorphization in the tablets. The experiment was set up using MODDE 12.1 (Umetrics AB, Sweden). The nanoparticle load was varied at 10, 15, and 20 wt% (with respect to the total tablet weight, Table ) and the celecoxib load at 30, 40, and 50 wt% (with respect to the polymer weight, Table ). Two SPIONs (Zn0.5Fe2.5O4 and Mn0.5Fe2.5O4) and the duration of AMF exposure (3 -15 min) were also included in the design. In total, 34 experiments were conducted (Table ), of which each CCF design had eight full factorial points, three center points and six star points. Models were fitted with multiple linear regression and adjusted by removing non-significant model terms. A value of p < 0.05 was considered significant. The surface contour plot was constructed to visualize the effect of the factors on tablet temperature and drug amorphization. |
6216933d91a2e60d14d94b41 | 18 | The tablets were prepared based on the experimental design. Physical mixtures of SPIONs, celecoxib (Mw = 381.4 g mol -1 ), PVP (Mw = 2000-3000 g mol ) and magnesium stearate (0.5 wt%) (Mw = 591.3 g mol -1 ) were prepared using mortar and pestle. Celecoxib and magnesium stearate were purchased from Fagron Nordic A/S (Denmark). Kollidon 12PF (PVP) was a kind gift from BASF (Germany). From the physical mixtures, 50 ± 2 mg flat-faced tablets (Ø 6 mm) were prepared using an instrumented single punch tablet press GTP-1 (Gamlen Instruments, UK). The tablet press was fitted with a 500 mg load cell (CT6-500-022) and used at a compression pressure of 160 MPa. The tablets and the mixtures were stored in airtight containers until further use. |
6216933d91a2e60d14d94b41 | 19 | The tablets were exposed to an alternating magnetic field (AMF) using Magnetherm (Nanotherics Ltd., UK). They were placed on a glass petri dish inside a coil with nine windings and the nominal oscillation frequency was set to 588.5 kHz and the magnetic field intensity at 14 mT. The AMF exposure time was varied according to the DoE (Table ). The increase in tablet surface temperature in the AMF was measured using an IR thermal camera (Fluke Ti480 Pro, Fluke Europe, the Netherlands). A thermal image was taken every third second for the first three minutes, and thereafter every 30 seconds. The IR images were analyzed using SmartView 4.3 (Fluke Europe, the Netherlands). |
6216933d91a2e60d14d94b41 | 20 | Tablet Characterization. Tablets exposed to AMF were gently powderized by a mortar and pestle for subsequent solid-state analysis. The water content of the pure compounds, the physical mixtures and the tablets after AMF exposure was determined with a thermogravimetric analyzer (TGA; Discovery, TA Instruments Ind., USA). The experiments were conducted under a nitrogen gas purge of 25 mL min -1 and a heating rate of 10 °C min -1 from ambient temperature to 170 °C. |
6216933d91a2e60d14d94b41 | 21 | Each experiment was conducted in duplicate (n = 2) for the pure compounds and physical mixtures, and as a single run for the powderized tablets. The weight loss (corresponding to the water content) was determined using the TA Instruments TRIOS software (version 5.1.1). Drug crystallinity of the pure substances, physical mixtures and powderized AMF-exposed tablets was analyzed with an X-Ray powder X'Pert Pro diffractometer (PANalytical, The Netherlands) using Cu Kα radiation (λ = 1.54187 Å). The diffractograms were recorded from 2θ 5° to 30° at 45 kV and 40 mA. As the superparamagnetic nanoparticles fluoresce, the pulse height distribution level was adjusted to 40-80% of the energy to compensate for the baseline drift and to obtain a higher signal-to-noise ratio. The diffractograms were analyzed using the X'Pert HighScore Plus software (version 2.2.4). Differential scanning calorimetry (DSC; Discovery, TA Instruments, USA) scans were collected for the physical mixtures and the AMF-exposed tablets. All experiments were conducted under a nitrogen gas purge of 50 mL min -1 . Using physical mixtures of 10-50 wt% celecoxib in PVP (in 10% increments), the melting enthalpy of the depressed melting was determined and fitted with a second degree polynomial function. The calibration experiments were conducted in duplicate (n = 2). 3-5 mg of each sample was weighted in Tzero aluminum pans with hermetic lids. The lid was perforated to allow water evaporation. A modulated DSC (mDSC) run was performed at a heating rate of 3 °C min -1 from -20 °C to 170 °C after an isothermal period of 2 min. The modulation had an amplitude of 1 °C min -1 at a period of 50 sec. The melting enthalpy was determined in the total heat flow signal. Each experiment was conducted once (n = 1). The sample mass and melting enthalpy were corrected for the water content determined by TGA. Using the calibration fitting, the crystallinity was calculated from the determined melting enthalpy. The relative residual crystallinity was calculated by dividing it with the crystallinity of celecoxib in the physical mixtures, i.e., before exposure to the magnetic field. Degree of drug amorphization (%) is the noncrystalline fraction calculated from the relative residual crystallinity. The melting enthalpy was determined using the TRIOS software from TA Instruments (version 5.1.1). |
6216933d91a2e60d14d94b41 | 22 | The theoretical onset were determined by DSC (Q-2000, TA Instruments, New Castle, DE, USA) as described elsewhere. The onset temperature of the heating rates were extrapolated to a heating rate of 0 °C min -1 . The extrapolated value corresponds to the respective onset temperature of the dissolution process for the three drug loadings of celecoxib in PVP. It should be noted that the determined temperatures correspond to water-free systems; and these will be lower for formulations that still contain small amounts of sorbed water due to the hygroscopicity of the polymer PVP. |
6216933d91a2e60d14d94b41 | 23 | High-performance liquid chromatography (HPLC) was conducted to quantify the amount of celecoxib in the physical mixtures and in the AMF-exposed tablets. In short, the experiments were performed using a 1260 Infinity HPLC (Agilent Technologies, Inc., USA) on a reverse-phase Luna 5U C18(2) 100 A column (length 150 mm, diameter 4.60 mm; Phenomenex Ltd., Germany). The experiments were performed at room temperature using a mobile phase consisting of 30% Milli Q water and 70% ethanol (>99.7% HPLC grade; VWR). Flow rate was set at 1 mL min -1 and a sample volume of 20 µL injected. UV detection was performed at 251 nm. Celecoxib eluted at about 2.7 min. The physical mixtures and the AMF-exposed tablets were dissolved in ethanol and filtered using a nylon syringe filter Q-max RR 25 mm with a pore size of 0.45 µm (Frisenette Aps, Denmark). Each sample was injected in triplicate (n = 3). The sample mass was corrected for the corresponding water content determined with TGA. |
6216933d91a2e60d14d94b41 | 24 | In Vitro Dissolution of AMF-exposed tablets. Fasted state simulated intestinal fluid (FaSSIF) was used as a biorelevant medium to study the in vitro dissolution and was prepared according to the manufacturer's instructions (Biorelevant.com). The mixture was equilibrated for 2 hours before use and used within 48 hours of preparation. Amorphous celecoxib and conventional ASD (30 wt% drug load) were prepared by melt quenching technique. The drug polymer ratio for the conventional ASD was selected based on the tablet resulting in maximum amorphous content induced by in situ amorphization (tablet 26, Table ). For the ASD, the drug and polymer were weighed and mixed thoroughly using mortar and pestle. This mixture was spread in a thin layer on a sheet of aluminum foil and placed in an oven at 169 °C for 5 min, after which, it was removed, cooled to room temperature, and then pulverized using mortar and pestle. This process was repeated twice for the ASD and once for amorphous celecoxib. |
6216933d91a2e60d14d94b41 | 25 | The non-sink dissolution measurements were performed in FaSSIF using the μDISS Profiler (Pion Inc., USA). Each channel was calibrated with its own standard curve prior to the experiment. For calibration, stock solutions were prepared using dimethyl sulfoxide containing either crystalline celecoxib alone or celecoxib (30 wt%) in PVP. The standard curves were established with 6-8 concentrations by sequentially adding 5 μL aliquots of the stock solution into 3 mL of FaSSIF maintained at 37 ± 1 °C and stirred at 800 rpm. The standard curve showed a linear correlation R 2 ≥ 0.99 over a range of 15-270 μg mL -1 . The dissolution and solubility of all compounds was determined in FaSSIF. The path lengths (2-20 mm) of the in situ UV probes were selected based on the drug solubility in the selected medium and the expected degree of supersaturation. When a high solubility was anticipated, the shorter probe length (5 mm) was used. |
6216933d91a2e60d14d94b41 | 26 | The medium was preheated to 37 °C prior to use in the dissolution experiment. The measurement was initiated by adding an excess of drug to the vials followed by 15 mL of preheated FaSSIF. A dose of 400 μg mL -1 of celecoxib was selected based on the recommended minimum dosage for rheumatoid arthritis and the intestinal fluid volume. Vials containing crystalline and amorphous celecoxib, and ASD (30 wt% drug load) were stirred using cross-bar magnetic stirrers at a speed of 200 rpm, whereas the vials with SPION containing powders were shaken at 300 rpm to avoid interaction with the magnetic stir bar. All samples were maintained at 37 ± 1 °C and analyzed at 1, 5, 10, 15, 20, 25, 30, 45, 60, 120, 180, 240, and 360 min with the UV-probes. Each experiment was run in triplicates except the AMF-exposed SPION containing tablets, which were run in duplicates. The second derivative spectra were used to determine the concentration of celecoxib in the medium over time. Maximum drug concentration (Cmax) and time to reach Cmax (tmax) were obtained by non-compartmental analysis the in vitro dissolution data. A One-way analysis of variance (ANOVA) using Tukey's multiple comparisons test was used to compare the Cmax of various formulations. Data analysis was performed using GraphPad Prism 9.0 software (La Jolla, CA, USA). p values were calculated as 0.1234 (ns), 0. |
614b2e4518be851f2c2c2699 | 0 | Epigenetic drug and probe discovery continue to be relevant for a plethora of potential therapeutic applications, including cancer therapy. Probes for epigenetic targets would be key components to understand epigenome data that can be the basis to develop personalized medicine. Similarly, chemical probes targeting epigenetic processes are fundamental in basic chemical biology associated with epigenetic processes. Computational and experimental screening efforts of chemical libraries have been ongoing to identify potential hits for various epigenetic targets. In addition to the traditional general screening libraries available for high-throughput screening, chemical companies have been assembling libraries focused on epigenetic targets and the chemical structures are available in the public domain. |
614b2e4518be851f2c2c2699 | 1 | Indeed, in epigenetic drug discovery, there are recent and increased efforts to design and analyze focused libraries. By design, epigenetic-focused libraries have the potential to increase the epigenetic relevant chemical space, which has recently been revised. Depending on the library developer, the compounds in the compound data sets are selected following a multi-step procedure, typically through computational approaches (although the specific methodological details remain undisclosed to the public). Many providers emphasize the diversity of the libraries, yet it remains hidden to the user what their contents and coverage of the chemical space are. |
614b2e4518be851f2c2c2699 | 2 | To this end, validated chemoinformatic approaches provide a systematic and rigorous manner to characterize the contents, chemical diversity, and, in general, coverage of the compound libraries in chemical space. The goal of this study is to rigorously characterize the chemical content, diversity, and drug-like properties of eleven epigenetic-focused libraries containing more than 50,000 compounds in total. All data sets are available in the public domain. To the best of our knowledge, this is the first systematic chemoinformatic analysis of epigenetic-focused compound libraries. As part of the analysis, we implemented a recently introduced and validated methodology to compute and identify core structures and analog series based on retrosynthetic rules. The analog series were the basis to generate Constellation Plots as a rich representation of the chemical space that combines substructure and fingerprint representations. For each epigenetic-focused data set, we identified the most frequent core structures. We also generate novel representations of the chemical space using the concept of Extended Chemical Space Networks (eCSNs) and rapidly identify the most representative individual molecules (i.e., medoids) in a large data set: these chemical structures can be used as a "chemical structural marker" or "chemical diagnostic molecules." Herein, we discuss the most representative chemical structures of the epigenetic libraries that can be used as a criterion for comparing the chemical libraries and prioritization for follow-up studies, including computational and experimental screening. |
614b2e4518be851f2c2c2699 | 3 | To conduct the chemoinformatic characterization of the 11 epigenetic-focused data sets, we calculated properties of pharmaceutical interest; molecular scaffolds using the classic and most common method by Bemis-Murcko, and a novel approach to identify core structures and analog series automatically. We also quantify the chemical diversity and synthetic accessibility using whole-molecule structural fingerprints. To explore the distribution of the chemical libraries in chemical space, we generated two complementary and recently introduced visual representations of the chemical space, namely, Constellation Plots and eCSNs. The details of the data sets, their preparation, calculation of the properties, scaffolds, and fingerprints are described in the following sections. |
614b2e4518be851f2c2c2699 | 4 | The structure files of the focused libraries were obtained from different chemical vendors. The 11 chemical libraries are summarized in Table that briefly describes each set and the number of compounds before and after data curation. The number of compounds after data curation was 53,443, in total. The chemical structures were curated using the open-source cheminformatics toolkit RDKit, version 2021.03.2 (www.rdkit.org). Data curation was performed using an established protocol. Briefly, compounds with valence errors or any chemical element other than H, B, C, N, O, F, Si, P, S, Cl, Se, Br, and I were deleted. Stereochemistry information was removed because not all compounds in datasets have it defined. Compounds with multiple components were split, and the largest component was retained. The remaining compounds were neutralized and reionized to generate a canonical tautomer subsequently. Duplicated compounds were deleted. <Table 1 here > |
614b2e4518be851f2c2c2699 | 5 | For each chemical structure of the 11 curated libraries, six properties of pharmaceutical interest were computed with RDKit, version 2021.03.2: molecular weight (MW), number of acceptor and donor hydrogen bonds (HBA, HBD, respectively), number of rotatable bonds (RB), topological surface area (TPSA), and partition coefficient octanol/water (SlogP). These properties are associated with compound size (MW), polarity (HBA, HBD, TSPA, SlogP), and flexibility (RB). The six properties are the basis of the well-known empirical rules of Lipinski and Veber (i.e., MW≤500, HBD≤5, HBA≤10, SlogP≤5, TPSA≤140, RB≤10) that help to guide the suitability of a compound to be orally absorbed (the preferred use of drug administration in most cases). The properties are not associated directly with biological activity but are commonly used to profile compound screening libraries in drug discovery projects. |
614b2e4518be851f2c2c2699 | 6 | We used two approaches to systematically generate and analyze molecular scaffolds, exemplified in Figure : Bemis-Murcko scaffolds automatically generated with RDKit, version 2021.03.2; and core structures and analog series using open-source code we previously reported available at . < Figure here > To compute the Bemis-Murcko scaffolds the side chains are removed, as illustrated in Figure . |
614b2e4518be851f2c2c2699 | 7 | More specifically, based on the graph representation of the chemical structures of all vertices of degree one. The concept of core scaffolds and analog series is also illustrated in Figure and discussed in detail elsewhere. Briefly, core scaffolds and analog series apply a series of fragmentation rules based on retrosynthetic feasibility. Two molecules are considered analogs if the fragmentation rules can map them to the same core fragment, and that fragment is a significant part of the molecule (usually, it contains at least two-thirds of the total number of heavy atoms of each molecule). |
614b2e4518be851f2c2c2699 | 8 | The scaffold diversity was also quantified based on the number of cores and analog series, and their ratio compared with the total number of molecules in the dataset. We have previously used the number of identifiable cores and analog series as a measure of chemical diversity in a dataset. However, the 11 data sets presented here are too variable in their size. Therefore, we propose two new indicators: the fraction of molecules in the constellation plot (only analog series with at least three compounds are included there) and the average size of the analog series, measured as the number of unique compounds in the dataset divided by the number of analog series represented in it. |
614b2e4518be851f2c2c2699 | 9 | The diversity of the compound libraries was computed with the extended similarity indices as recently reported. Briefly, instead of only measuring the similarity between pairs of molecules, the extended (e.g., n-ary) indices allow us to calculate the similarity of any number of molecules simultaneously. This has two key advantages when it comes to the analysis of molecular libraries. First, the n-ary indices provide a truly global description of the correlation between the molecules in the set (as exemplified by their superior performance in estimating the compactness of a set. Second, the extended indices are dramatically more efficient, requiring only O(N) operations to calculate the similarity of N molecules (opposed to the quadratic scaling of the standard binary similarity indices). Extended similarity has been successfully used in numerous applications, including compound diversity analysis, comparison of nucleotide and protein sequences, and, more recently, analysis of molecular dynamics simulations. Herein, we used this approach as a novel and efficient manner to quantify and compare the fingerprintbased diversity of the 11 compound libraries. The extended similarity indices were computed with a fractional weight function, with various coincidence thresholds. The Python code to conduct the extended similarity indices calculations is freely available at . |
614b2e4518be851f2c2c2699 | 10 | Chemical space has been defined as the set of molecular descriptors in which molecules will be represented. Visual representation of the chemical space helps to better understand the mutual relationships between compounds in that multi-dimensional descriptor coordinates. Among the several methods available (examples of the most common include principal component analysis and t-distributed stochastic neighbor embedding), herein we used two novel representations detailed in this section. |
614b2e4518be851f2c2c2699 | 11 | Constellation plots are useful to depict the chemical space of chemical libraries containing several analog series. They are helpful at concisely depicting the structure-activity (property) relationships -SA(P)Rin a summarized representation of the data set, as analog series can be represented in fewer data points than individual compounds. Of note, only a fraction of the total data is presented in the constellation plot: the compounds forming analog series; in this case, we included only analog series consisting of at least three compounds. Recently, constellation plots have been used to describe a library of antidiabetic natural products and a collection of tubulin inhibitors. |
614b2e4518be851f2c2c2699 | 12 | The chemical space networks (CSNs), proposed and developed by Maggiora and Bajorath, start by measuring the pairwise similarity between the molecules in a data set (using a given similarity coefficient and a compound representation). Then, the molecules are represented by nodes which are connected if the similarity is larger than an established threshold. A limitation of this approach is that it is difficult to visualize networks for large compound data sets. Moreover, this approach also requires O(N 2 ) operations, so it is not well-suited to represent large sections of the chemical space. To overcome this issue, the eCSNs have been recently proposed (vide supra). This is a natural generalization of the CSNs, in which any given molecular set can be taken as a node in the network (in this study, the nodes will be the 11 libraries). Then, the relations between these nodes are established via the extended similarity calculated for the union of the corresponding libraries. This coarse-grained representation is markedly more efficient since it exploits the more favorable computational scaling of the n-ary indices. |
614b2e4518be851f2c2c2699 | 13 | The complexity of the compounds generated was estimated using the synthetic accessibility (SA) score previously published. Briefly, the SA score implemented in this study is the difference between a fragment score and a complexity penalty. The fragment score captures common structural features in a large number of already synthesized molecules (934,046 representative molecules from the PubChem). |
614b2e4518be851f2c2c2699 | 14 | Molecules are fragmented using extended connectivity fragments, and the fragment score is calculated as a sum of contributions of all fragments in the molecule divided by the number of fragments in the molecule. The fragment frequency is related to their synthetic accessibility and hence easy-to-prepare substructures are present in molecules quite often. The complexity score is calculated as the sum of ring complexity (i.e., ring bridge atoms and spiro atoms), the number of stereocenters, large rings (i.e., ring size greater than eight, molecular complexity increases), and molecule size. The SA score was calculated for all epigenetic-focused libraries. |
614b2e4518be851f2c2c2699 | 15 | Figure in the Supporting Information shows box plots and summary statistics of the distribution of the six calculated properties of pharmaceutical relevance. The profiling of the six properties indicated that, in general, all 11 compound libraries are within the Lipinski and Veber parameters. Based on this criterion, the libraries are acceptable candidate compounds for drug discovery and development programs (in particular, to be administered orally). All 11 compound data sets have a comparable distribution of the six properties, as shown in Figure . The outcome of the profiling might be anticipated since it is likely that the chemical vendors (developers) filter or consider the so-called "drug-like" properties during the assembly of the focused libraries. However, the profiling disclosed in this work is relevant and encourages the experimental screening of the 11 compound libraries for drug discovery projects. |
614b2e4518be851f2c2c2699 | 16 | The relevance of analyzing the main core scaffold of a chemical compounds, in the the context of drug discovery, is particularly relevant because the central element drives the main molecule shape, arrange the substituents in their specific positions and take part of the biological activity itself. For this reason, systematic profiling of scaffold content of synthetic organic compounds for drug discovery is of utmost relevance. |
614b2e4518be851f2c2c2699 | 17 | Generating automatically and consistently the main scaffold or core structure of large data sets can be done in several ways as recently reviewed. In general, it is desirable to generate the scaffolds rapidly, consistently, and interpretable, in particular for an organic or medicinal chemist working on chemical synthesis. As detailed in the Methods section, in this work we implemented two methodologies generating the Bemis-Murcko scaffolds and analog series based on core scaffolds (Figure ). Concerning the core scaffolds and analog series, the most frequent are shown in Figure . The obtained cores overlap very little with the 201 substructures that have been annotated as epigenetic bioactive rings in a recent publication: of the 4016 cores matching at least three molecules from any epigenetic data set, only 19 contained at least one of the epigenetic rings. This highlights the structural novelty of the studied libraries and it's potential to expand the epigenetic relevant chemical space. |
614b2e4518be851f2c2c2699 | 18 | Figure shows the percentage of unique Bemis-Murcko scaffolds in each of the 11 libraries. The analysis revealed that TocrisScreen is the set with the largest percentage of unique scaffolds (85.6 %), indicating a large scaffold diversity considering the definition of the scaffold of Bemis-Murcko. The library with the second largest percentage of unique scaffolds was Targetmol (82.3 %). Figure shows the percentage of scaffolds with a frequency of at least two per library. Clearly, ChemDiv was the data set with the largest proportion of non-unique scaffolds suggesting the lowest scaffold diversity. |
614b2e4518be851f2c2c2699 | 19 | < Figure here > To further quantify scaffold diversity based on the Bemis-Murcko scaffolds, we used cyclic system recovery curves. As documented elsewhere, based on the scaffolds count, the fraction of cyclic systems is plotted against the cumulative fraction of the database. A diagonal represents maximum scaffold diversity, i.e., each compound will have its chemical scaffold. A vertical line represents the minimum scaffold diversity (all compounds have the same scaffold). Figure shows the cyclic system recovery curves for all data sets. The curves can be further characterized by the area under the curve, AUC (maximum diversity: AUC = 0.5; minimum diversity: AUC = 1.0). Table in the Supporting Information summarizes the AUC values for the 11 data sets. The results show that ChemDiv is the least diverse, followed (AUC = 0.87) by OTAVA DNMT3b. In contrast, TocrisScreen and Targetmol are the most diverse (AUC <= 0.56). |
614b2e4518be851f2c2c2699 | 20 | The analog series analysis suggested that ChemDiv, Asinex, Life Chemicals, and OTAVA DNMT3b are the least diverse data sets, as they have a larger average of compounds per series. All the other libraries seem to be more diverse, and it is hard to point out at the least diverse from this perspective. <Table 2 here > |
614b2e4518be851f2c2c2699 | 21 | Notably, ChemDiv is the compound data set with the largest number of compounds (27,543) meaning that the larger data set is not necessarily the one with the largest structural diversity, as clearly shown here. Similar results have been obtained for other data sets. These results further emphasize the need to quantify the structural diversity. Similar conclusions regarding the relative diversity of the compound libraries were obtained with other extended similarity indices (in addition to Tanimoto) and MACCS keys, as shown in Figure in the Supporting Information. |
614b2e4518be851f2c2c2699 | 22 | We were only interested in analog series having no less than three compounds, even if all of them belonged to different data sets. Afterwards, we highlighted the cores (points) represented in each database. Plots of representative epigenetic libraries are shown in Figure , while other libraries are depicted in Figure of the Supporting Information. |
614b2e4518be851f2c2c2699 | 23 | The eCSNs have been used to visualize the chemical space of 19 large data sets of organic compounds, including natural products, drugs approved for clinical use and other compound libraries, with more than 18 million molecules. As discussed in that paper, this novel representation of the chemical space based on molecular fingerprints is an efficient method to compare the structural relationship among compound libraries. Figure shows a visual representation of the chemical space of the 11 compound libraries using RDKit fingerprints. The network shows that ChemDiv and Asinex (which happen to be the least diverse libraries based on RDKit fingerprints) are at the center of the representation with several connections (similarities) with other databases. In particular, ChemDiv (identified with the number 2 in this network: ID, 2) has the most number of connections, and these connections are closer (visualized by darker linkers between the nodes) with other libraries, such as ApeXBio (ID, 0), SelleckChem (ID, 8), and Targetmol (ID, 9). < Figure here > |
614b2e4518be851f2c2c2699 | 24 | Figure shows the three most representative chemical structures per library as calculated with RDKit fingerprints. The twenty more representatives (medoids) per library (including the three in Figure ) are listed in Table in the Supporting Information. The medoids are calculated using the algorithm described recently. In short, we calculate the complementary similarity of every molecule in a library, that is, the similarity of all but the selected molecule (which we can do in O(N) for the whole set). Then, we can rank all the molecules from more (e.g., medoid-like) to less (e.g., outlier-like) representative by simply ordering them according to the increasing value of their complementary similarity. The reasoning behind this is very simple: by removing a molecule that is closely related to all of the rest we leave behind a more 'disorganized' set, which will have a lower complementary similarity. In this context, the medoids could be interpreted as chemical structural markers or "signatures" of the compounds libraries and contribute to profile the chemical contents of each data set. <Figure 7 here > |
614b2e4518be851f2c2c2699 | 25 | We also profiled the chemical libraries using a validated in silico approach to estimate the synthetic accessibility, as described in the Methods section. Results presented in Figure in the Supporting Information (box plots and summary statistics) indicated that all compound libraries have a comparable profile and, in general, it is expected that the vast majority of the libraries are synthetically accessible. Of note, all focused epigenetic libraries analyzed in this work are commercially available from the chemical providers listed in Table . |
614b2e4518be851f2c2c2699 | 26 | Profiling of the six properties of pharmaceutical relevance: MW. LogP, HBD, HBA, TSPA, and RB, revealed that all 11 compound libraries are suitable to be screened in drug discovery campaigns to identify molecules that eventually could be orally administered. It was found that, other than benzene, Nphenyl-benzenesulfonamide, 1H-indol, and N-phenylbenzamide, and 1H-benzimidazol were the most prevalent. The results of the fingerprint-based diversity indicated that SelleckChem is among the most diverse libraries. In contrast, ChemDiv and Asinex are the least diverse, relative to all other data sets. |
614b2e4518be851f2c2c2699 | 27 | Regarding the Bemis-Murcko scaffolds and analog series, ChemDiv was also the least diverse, while TocrisScreen, Targetmol, and SelleckChem were the most diverse. Taken together, based on the results of structural diversity, the most diverse library overall (TocrisScreen) should be prioritized for experimental medium-throughput screening. Interestingly, out of the 11 databases analyzed, TocrisScreen was the smallest data set (100 compounds), yet it is the most diverse. In sharp contrast, ChemDiv was the largest data set (27,543 compounds) but is the least structurally diverse. The scaffold content and analog series, analyzed in the context of rings present in currently known compounds with activity against epigenetic targets revealed that the focused libraries have a large potential to expand the epigenetic relevant chemical space. Results of the calculated synthetic accessibility showed that all compound data sets are, in general, feasible to make. For practical applications, the libraries could be acquired first by the chemical vendors but, if needed, could be synthesized in-house. We anticipate that the results of the chemoinformatic characterization discussed in this work will assist research teams in the decision-making process and prioritize what libraries move forward to experimental screening in, for example, a high-throughput screening setting. |
66d9817c12ff75c3a1765eb2 | 0 | Protein structure and function underpins all aspects of life from the structural framework of cells, enzymatic activity, cellular transport, to immune function. The production of the individual proteins is controlled through the expression of the corresponding genes. Riboswitches are non-coding RNAs mediating the regulation of genes by initiating or terminating transcription or translation. Here, gene expression is being triggered by environmental factors such as temperature or pH or through ligands like ions or small molecules. Ion or metabolite sensing riboswitches form highly selective binding pockets for their specific target. Fluoride sensing riboswitches are present in both, bacteria and archaea. Increasing F -concentrations can inhibit cell growth and become acutely toxic to cells. This antimicrobial effect has been exploited to eliminate harmful microorganisms in fermentation processes. Fluoride riboswitches are involved in cellular defence mechanisms, such as initiating the production of F -exporters. The first fluoride sensing riboswitch was found in the crcB motif of Pseudomonas syringae in 2012. In the same year, the crystal structure of the sensing domain of the fluoride sensing riboswitch from the bacterium Thermotoga petrophila in its F --bound form was reported by Ren et al. (Fig. ). The tertiary structure of the aptamer comprises two stems, a pseudoknot, and reversed Watson-Crick (A6•U38) and Hoogsteen (A40•U48) base pairs (Fig. ). The riboswitch coordinates a cluster of three Mg 2+ ions which in turn encapsulates the F -ion (Fig. ). In this way the polyanionic RNA avoids electrostatic repulsion with the bound F -. This motif composed of Mg 2+ , F -, phosphate, and water was computationally found to be stable largely based on the electrostatic interactions between F -and Mg 2+ . A recent computational study proposes a stepwise assembly of the cluster with two Mg 2+ bound in the apo form but the third Mg 2+ only being incorporated together with F -. The riboswitch is strongly selective for F -discriminating against other halides. Using solution nuclear magnetic resonance (NMR) spectroscopy, the fluoride binding riboswitch from Bacillus cereus was shown to adopt a highly similar fold. Here, stem 1 and stem 2 form already in absence of Mg 2+ and F -, whereas the formation of the pseudoknot requires presence of Mg 2+ . Further addition of F -does not result in a significant change of the tertiary structure but suppresses dynamics involving a lowly populated excited apo state responsible for transcription termination. Pulsed electron-electron double resonance (PELDOR, aka DEER for double electron-electron resonance ) spectroscopy permits measuring distances ranging from 20 Å to 100 Å and beyond between paramagnetic centres such as spin labels, paramagnetic metal ions, amino acid radicals and radical cofactors. This provides a unique opportunity for investigating conformational ensembles through their spin-spin distance distributions. Complementary to PELDOR, F electron-nuclear double resonance (ENDOR) is an emerging technique to detect distances in a shorter range of about 5 -20 Å between nitroxide spin labels and F nuclei. This has recently been expanded to distances between fluorine nuclei and other spin centres such as triarylmethyl (TAM) radicals, tyrosyl radicals, Cu II , or Gd III . In this work, we have investigated the T. petrophila riboswitch in solution by magnetic resonance methods. H NMR spectroscopy allowed monitoring the formation of base pairs through signals in the imino region of the free RNA, the Mg 2+ -bound apo riboswitch, and the F --bound holo riboswitch. We have further studied the solution-state preorganisation of the tertiary fold in the free and apo riboswitches by PELDOR and we have employed F ENDOR to investigate the structure of the binding pocket of the holo riboswitch. 4ENC) with the F -(green sphere) encapsulated by three Mg 2+ (blue spheres); (b) schematic representation of the secondary structure of the 50 nucleotide construct of the fluoride binding riboswitch used here, with the two stem structures (green and blue), the pseudoknot structure (magenta), and the reversed Watson-Crick (A6•U38) and reversed Hoogsteen (A40•U48) base pairs (salmon). Residue numbers for spin-labelling sites are indicated. The residues whose O -are involved in the coordination of the Mg 2+ are highlighted with red asterisks; (c) Graphical representation of the binding pocket of the fluoride binding riboswitch (PDB: 4ENC) with the F -(green sphere) encapsulated by three Mg 2+ (blue spheres). O -involved in the coordination are represented as red spheres. Additional ions not involved in the fluoride encapsulation and water molecules have been omitted for clarity throughout. |
66d9817c12ff75c3a1765eb2 | 1 | PELDOR and F ENDOR measurements require the presence of two or one paramagnetic spin labels, respectively. Several methods for site-directed spin labelling of nucleic acids at the base, ribose, or phosphate exist. Custom synthesised RNA sequences with phosphorothioate modification are commercially available as is the spin label precursor for labelling at the phosphate backbone (Scheme 1) making this labelling approach readily accessible. This procedure leads to the introduction of a diastereomeric pair. The spin labelled RNA is therefore expected to display both diastereomers of the phosphothiotriester. |
66d9817c12ff75c3a1765eb2 | 2 | All constructs were designed based on the crystal structure of the T. petrophila riboswitch. The labelling sites were screened using in silico labelling with MtsslSuite and MMM to predict distances between pairs of spin labels or spin labels and F -. Furthermore, ensembles of spin label rotamers can also be extracted allowing analysis of orientation selection in the anisotropic EPR spectrum. In absence of an experimental highresolution structure a similar modelling can be applied using computational structure predictions that have become available also for nucleic acids with the introduction of AlphaFold3. To select the labelling sites, following general criteria were applied: (i) nucleotides involved in the formation of the cluster, i.e. residues coordinating Mg 2+ were excluded (red asterisks in Fig. ); (ii) distances in the range between 18 -45 Å were considered for PELDOR, and distances below 20 Å for F ENDOR. |
66d9817c12ff75c3a1765eb2 | 3 | For PELDOR measurements four different constructs with two spin labels each were prepared: G29/G36 within stem 2 and the corresponding hairpin (blue), C14/C44 located in the region between the pseudoknot and stem 1 (green), A9/A49 across stem 1 and the 3' end of the RNA (magenta), and C4/C17 across the pseudoknot (red, Fig. and Fig. ). |
66d9817c12ff75c3a1765eb2 | 4 | Scheme 1. RNA labelling at the phosphorothioate modification of the backbone using either a protonated or a per-deuterated spin label. The sugar 5' to the labelling site is modified to a 2'-deoxyribose to avoid strand scission. For F ENDOR, three constructs with one spin label each were chosen to result in linearly independent distance vectors between the respective label and the F -, allowing the trilateration of the F - position within the RNA structure (Fig. ). G5 and G43 gave respective calculated mean distances (Rmodel) of 8.5 Å and 13.5 Å to the F -, well resolvable below the 20 Å threshold. For U46 however, in silico labelling displayed a bimodal distance distribution. The longer Rmodel,1 was predicted to be at about 21.5 Å, likely too long to be resolved by F ENDOR. Nevertheless, the second, shorter Rmodel,2 of 15.5 Å was within the resolvable distance limit for F ENDOR with nitroxide labels. To ensure reproducible folding of the riboswitch and infer any perturbation of folding by the spin label, we established a folding protocol (see SI section 1) using 1 H NMR spectroscopy. This allowed to identify the free aptamer in absence of Mg 2+ and F -, the Mg 2+ -bound apo riboswitch, and the F --bound holo riboswitch in analogy to results for the B. cereus fluoride riboswitch. Comparison of 1 H NMR spectra in the imino regions of the free and holo aptamer for both, unlabelled and labelled RNA showed indistinguishable signals indicating appropriate folding after labelling (Fig. ). Alternatively, isothermal titration calorimetry (ITC) could be used to investigate labelled and mutated riboswitch constructs for retained aptamer functionality. |
66d9817c12ff75c3a1765eb2 | 5 | PELDOR distance measurements using the 4-pulse DEER sequence (Scheme 2 (a)) are particularly well suited to study structural or conformational changes. To map overall structure and preorganisation, PELDOR time traces were measured for the free, apo and holo aptamers for each of the four doubly spin labelled constructs. Sample preparation and experimental conditions for PELDOR are described in SI section 1. All spectra were recorded in frozen solution at 50 K. The time traces (Fig. ) displayed differently pronounced dipolar modulations, which indicated different widths of the respective distance distributions. The time traces were analysed using Tikhonov regularisation implemented in the software DeerAnalysis (Fig. ). All obtained distance distributions, despite being broad, showed a distinct most probable distance, but did not show significant differences between the free, apo and holo aptamers for the C14/C44, G29/G36, and A9/A49 constructs. The C4/C17 construct displayed the most pronounced dipolar modulation resulting in the narrowest distance distribution with the lowest uncertainty. Here, the free and apo forms provided no difference within uncertainty but the holo form yielded a slightly reduced most probable distance. These results could be confirmed by an alternative processing with a different implementation of Tikhonov regularisation in the software DeerLab and by neuronal network analysis with the software DeerNet (Fig. ). Interestingly, all distance distributions displayed significant distance probability up to 50 Å and beyond in all three processing approaches. Comparing the experimental distance distributions with the modelling performed to select these constructs revealed substantially broader distributions in the former. Nevertheless, all modelled distance probability is fully covered by the experimental distributions. Particularly, the DeerLab analysis hinted at the presence of multiple distances corresponding to multiple conformations. This illustrated a conformational flexibility of the RNA backbone that was not reproduced by rotamer modelling on the single conformer in the crystal structure. Interestingly, for all cases but C4/C17 the most probable distances (i.e., maxima of the experimental distributions) were shorter than the maximum of the modelled distributions. This indicated the presence of more compact structural arrangements in solution than in the crystal structure. The similarity in the PELDOR data between the free, apo, and holo forms confirmed the preorganisation of stems 1 and 2 in the free RNA and their retention in the apo and holo aptamers. The shortening of the C4/C17 distance in the holo compared to the free and apo forms was consistent with an increased formation of the pseudoknot only upon addition of F -in contrast to the findings for the B. cereus fluoride riboswitch. The long distances tailing up to 50 Å and beyond hint to an equilibrium involving a disordered or unfolded form of the aptamer. This was consistent with earlier fluoride binding ITC data that had been fitted to 0.75 and 0.87 binding sites suggesting 13-25% of aptamers not binding F -and contributing very broad peaks to the distance distributions of all three forms. ) spin labelled in silico with MtsslSuite and with arrows indicating the spin-spin distance assed by PELDOR measurements. |
66d9817c12ff75c3a1765eb2 | 6 | The fluoride binding riboswitch provides a unique opportunity to measure inter-spin distances to the active site in the holo aptamer, containing an endogenous F -, by F ENDOR spectroscopy. For this, F Mims ENDOR (pulse sequence in Scheme 2b) was performed with three constructs containing a single nitroxide label at positions G5, G43, or U46 (Fig. ). F ENDOR spectra of G5 and G43 were measured at W-band (94 GHz) and four different excitation positions in the EPR spectrum (Fig. , left). The orientationally selected ENDOR spectra were weighted by the spectral intensity at the excitation position, added, and analysed (see section 1 in SI). For construct U46, spectra were measured at Q-band (34 GHz), taking advantage of superior concentration sensitivity, as the expected distance was longer (Fig. ). At Q-band, H ENDOR signals from the spin label methyl groups overlap with the F ENDOR signals (Fig. ). This was circumvented by introducing a deuterated nitroxide spin label (see section 1 in SI). Furthermore, this construct showed the slowest echo decay (i.e., slowest transverse dephasing, Fig. ) contributing to improved sensitivity. At 34 GHz 19 F ENDOR spectra were measured only at two different spectral positions of the EPR line (Fig. , right), and no significant effect of orientation selection could be observed within the achievable resolution and signal-to-noise ratio. The three F ENDOR spectra of the holo riboswitch are displayed in Fig. . For constructs G5 and G43, a well-resolved hyperfine splitting arising from the maxima of a Pake pattern was observed. From this splitting, called Tread, the perpendicular component T⊥ of the hyperfine tensor was estimated, giving first insight into the magnitude of the dipolar coupling and the corresponding inter-spin distance. Using the point-dipolar approximation the point-dipole distance (Rread) can be estimated as: 𝑇 𝑟𝑒𝑎𝑑 ≈ 𝑇 ⊥ = 𝜇 0 4𝜋ℎ ( 𝑔 𝑒 𝑔 𝑛 𝜇 𝐵 𝜇 𝑁 𝑅 𝑟𝑒𝑎𝑑 3 ) = 𝐶 𝑅 Eq. 1 |
66d9817c12ff75c3a1765eb2 | 7 | with the vacuum permeability µ0, the Planck constant h, the g-values of the electron and fluorine nucleus ge and gn, respectively, the Bohr magneton µB and the nuclear magneton µN. With these constants C amounts to C = 74.52 MHz Å 3 . For G5, a coupling constant Tread = 50 ± 7 kHz was read off that corresponds (Eq. 1) to a point-dipole distance of Rread = 11.5 ± 0.5 Å. The error was estimated from the width of the maxima of the peaks. For G43, a smaller splitting from a coupling constant Tread = 30 ± 4 kHz was observed, corresponding to a point-dipole distance of 13.5 ± 0.6 Å. For U46 no splitting could be observed, however a clear peak was still detected. To ensure that this signal of U46 did not arise from free F -interacting with the nitroxide label, a control experiment was performed. F ENDOR spectra of the free aptamer labelled in position G43 and U46 were measured in the presence of F -at W-band and Qband, respectively. In absence of Mg 2+ the F -binding site cannot form and indeed no ENDOR effect has been observed (Fig. ). Thus, all F ENDOR signals observed in the holo form arise from bound F -. To extract more information on conformational distributions from the ENDOR line shape we performed a spectral analysis in two steps. First, we simulated the spectra using our fast simulation routine SimSpec (see section 1 in SI) without any prior model, just assuming a Gaussian distribution of distances centred at Rread. The centre and the width of the Gaussian distribution were manually adjusted to minimize the residual between experiment and simulated spectrum (Fig. and). The simulations are displayed in Fig. (left). The Gaussian distance distribution reproduced the line shapes of G5 and G43 very well using a broadening (full width at half maximum, FWHM) corresponding to 2.4 Å. The simulated Tsim component resulted in Rsim values as given in Fig. (right) and in Table . For U46 a series of simulations was performed for Gaussian distributions of constant width 2.4 Å but varying Rsim between 18 and 28 Å. Examination of the residuals (Fig. and Fig. ) suggested that Rsim ≳ 20 Å can generally be considered a lower boundary for the distance, when no splitting occurs. We note that this is the first detection of a F -nitroxide distance ≳ 20 Å, demonstrating an extension of the accessible distance range. Indeed, the resolution of the splitting depended on the choice of the line width parameter. Our choice of a Lorentzian line of 7 kHz (slightly larger than the power broadening of the RF pulse) is based on an ongoing investigation of intrinsic ENDOR line widths. This resulted in Rsim ≳ 23 Å. Using a larger line width of 12 kHz, resulted in a shorter limit of Rsim ≳ 21 Å (Fig. ). Moreover, a single distance above 20 Å was not sufficient for reproducing the broad base of the ENDOR spectrum. This feature could be simulated by introducing an additional, shorter distance centred at about 16 Å and in a much smaller population (only 10% of the overall distance distribution). We note that the Mims ENDOR experiment enhances the contribution of shorter distances, making such a small contribution discernible in the spectrum. In order to examine to what extent conformational distributions affect the spectra, we also performed simulations of rigid Pake patterns based on a single Rsim value convoluted with the estimated intrinsic ENDOR line width. Comparison of spectral simulations (Fig. , Fig. ) showed indeed additional broadening caused by a distance distribution between F -and spin label, which was most pronounced for the shortest distance in G5. In the second step of analysis, we estimated the contribution of the spin label to the ENDOR line width by rotamer modelling of label orientations in MtsslSuite and MMM (Fig. ), based on the available crystal structure (PDB 4ENC, see section 1 in SI). This modelling produces rotamer ensembles of spin label orientations as represented in Fig. . For each rotamer, distance and orientation were considered. The distance was taken from the midpoint of the NO-bond to the F -while the orientation was defined through the Euler angles α and β between the dipolar tensor and the spin label g-tensor, as illustrated in Fig. . The Mims ENDOR spectrum for each of up to 400 rotamers was computed in SimSpec (see section 1 in SI) and then added to generate the sum ENDOR spectrum (Fig. ). Both diastereomers were considered, however the modelling does not predict their relative weight. For G5, both diastereomers were considered in equal weight, while for G43 only one diastereomer was considered, as the second cannot be populated due to steric clashes. For U46, the weight of the diastereomer leading to the shorter distance was adjusted to reproduce the ENDOR spectrum. Starting from the construct labelled in position G5 we obtained a substantial deviation of 3 Å between the predicted Rmodel and the observed Rsim distance. However, the distance distribution width with ΔR = 2.1 Å was very close to the simulated value of 2.4 Å (Fig. ). For construct G43 we found a good agreement between the Gaussian and the predicted distance distribution. For U46 a small deviation in the long distance (Rmodel,1) could be observed while the width of this distribution contribution was in good agreement with the Gaussian. For all three constructs, we found that the width of the distance distributions predicted by the rotamer modelling, when keeping a rigid body of the RNA, was consistent with the width extracted from the model-free Gaussian distributions. This means that no substantial broadening arising from RNA backbone heterogeneity is observed in the ENDOR spectra. This is a significant result, which differed from the PELDOR findings, where the experimental distance distributions were considerably broader than the ones predicted from rotamer modelling. This is important, as it is known that rigid biomolecules (e.g., model protein GB1) allow reconciling PELDOR and RIDME data and ENDOR data all based on a single protein structure. Furthermore, considering the mean distances, we observed good agreement only for G43 and U46. At G5, the significant deviation between the experimental and modelled distances can be rationalised by crystal contacts in the 5' region (Fig. ) not being present in solution (vide infra). We conclude that the ENDOR data, although sensitive to small distance shifts of a few angstroms, do not report structural heterogeneity of the holo aptamer in the vicinity of the binding pocket. This indicates that the RNA structure around the active site is well-defined in the holo form. |
66d9817c12ff75c3a1765eb2 | 8 | The observed deviations between the modelled and experimental F ENDOR distances around G5 prompted us to trilaterate the F - position based on the experimental distances. This places the F - in a position between the pseudoknot and the unpaired bases between the pseudoknot and stem 2 (Fig. ). Here, the F --Mg 2+ cluster could not be stabilised by coordination from the phosphate backbone (Fig. ), strongly indicating a different structure or conformation around G5 well beyond the error of the F ENDOR measurement. The crystal structure shows significant intermolecular contacts (Fig. ) for the first two bases from the 5' end, suggesting a structural variation in absence of these contacts in solution. The deviation in the distance between the nitroxide spin label and the F -at G5 could be explained by a different orientation of the first five nucleotides at the 5' end. The distance between the G5 phosphate group and the F -is 6.3 Å in the crystal structure. The spin label rotamers have distances of 7 -7.5 Å between phosphate and NO group adding to a maximum NO-F distance of about 14 Å. Thus, the 12 Å distance seen in the F ENDOR measurement (Fig. ) could arise from a differently oriented phosphate group. Nevertheless, a high similarity between five crystal structures of the riboswitch in presence of various ions as well as with the NMR derived structure of the B. cereus riboswitch has been reported (Fig. ). Thus, the ENDOR and PELDOR results reveal different levels of structural heterogeneity -representing different functional dynamics -not directly obvious from crystal structures. and orientations (α, β) of all rotamers were extracted, a F ENDOR spectrum for each rotamer was simulated using SimSpec and the sum of the simulated |
66d9817c12ff75c3a1765eb2 | 9 | In this work we have shown that the free, apo and holo forms of the sensing domain of the fluoride riboswitch from T. petrophila could be reproducibly formed in vitro. However, a significant population of unfolded or disordered aptamer was present in all PELDOR samples manifesting as distance distributions tailing up to 50 Å and beyond. Nevertheless, PELDOR has revealed a preorganisation similar to the B. cereus fluoride riboswitch with the exception that also F -was required for pseudoknot formation. Modelling the widths of distance populations based on the single backbone conformation of the crystal structure matched the experimental distributions for F ENDOR but significantly underestimated the experimental PELDOR distance distributions. Thus, the combined PELDOR and F ENDOR results indicated a preorganised riboswitch structure with a more rigid F --binding site and more conformations sampled at the periphery. |
66d9817c12ff75c3a1765eb2 | 10 | For further understanding of conformational flexibility and presence of unfolded forms, investigation of the fluoride riboswitch structure in solution using, e.g., NMR or fluorescence methods is needed. Molecular dynamic simulations are complicated by the unusual cluster of Mg 2+ ions, but might lead to further insight. As a next step, extending the riboswitch from the sensing domain to the full-length expression platform will potentially allow forming the transcription terminator state in the absence of F -and observing the inherent conformational change by EPR methods, giving further insights into this intricate molecular mechanism of gene regulation. Finally, we have shown that the use of a fluorine bearing ligand allows to selectively the ligand bound form by F ENDOR in a complex mixture of bound and unbound molecules. In a more general context, similar strategies could be employed to study ligand binding of fluorinated pharmaceuticals to biological targets. F ENDOR complements PELDOR with more precise distances on a shorter length scale. Here, we first observed a F --nitroxide distance in the range of 2 nm, which completely closes the gap in the distance range accessible from pulse dipolar spectroscopies. Therefore, the combination of the two EPR-based methods, for distance measurements in frozen solution under similar sample conditions, provides valuable information on structure and conformational distributions, constituting a powerful tool for validation of structural models of complex biological systems. |
660a22a5e9ebbb4db917b00e | 0 | Carbon nanotubes (CNTs) are a one-dimensional (1-D) allotrope of carbon, made up of a sp 2 hybridized carbon lattice in the form of a cylinder. A single-wall CNT (SWCNT), the simplest version of a CNT, consists of a single, cylindrical graphene tube. In double-wall CNTs (DWCNTs) and multi-wall CNTs (MWCNTs), two or more tubes are either nested concentrically or wrapped like a scroll. SWCNTs are typically 0.4-2 nm in diameter, while MWCNTs can be much larger, with diameters of tens of nanometers, having both ends generally capped by fullerene-like domes. Similar to fullerenes and graphene, which are the analogous 0-and 2-D allotropes of carbon, the graphite-like arrangement of carbon atoms in CNTs and their shape give them remarkable and unique properties, which has resulted in the widespread use of CNTs in several fields of science and technology, including energy storage, electronics, structural composites, biomaterials, and others. These properties include a high Young's modulus (0.27-1.8 TPa) and tensile strength (11-63 GPa). CNTs are stable in vacuum up to 1225-1400 °C, though their stability is compromised in oxidizing environments (in air they decompose through oxidation at 400-550 °C). Because of their 1-D structure, CNTs have highly anisotropic mechanical, electrical, and thermal properties, which make them ideal for use in polymer, ceramic, and metal matrix composites. Additionally, their electrical properties, very high surface to volume ratio, and potential for chemical functionalization has led to their use in battery and supercapacitor components. The electrical properties of CNTs are primarily determined by the configuration of the C-C bonds relative to their macroscale orientation. CNTs can have three configurations: armchair, zigzag, and chiral. Armchair-configured CNTs, which are achiral, are electrically similar to metals, while the zigzag (also achiral) and chiral CNTs are semiconducting. In the following sections, we will discuss the trends in CNT-related research from 2003 to 2023, with a focus on identifying the emerging topics of research in this area over the last 3 years. This analysis was performed by first identifying journal and patent publications in the CAS Content Collection™ related to CNTs, which resulted in a set of roughly 270,000 documents. We identified emerging applications, materials, applications, properties, and other concepts in CNT-related research using a Natural Language Processing (NLP) based approach. This analysis identified over 39,000 concepts that appear in at least 20 documents, which were then classified, counted, grouped, and analyzed. The content of the following sections is based on the results of this analysis, in combination with substances, concepts, and other information indexed in the CAS Content Collection. For a more detailed description of this process, please see the Methods section in the Supplementary Information. |
660a22a5e9ebbb4db917b00e | 1 | The current era of CNT research began in 1991 when Sumio Iijima of Nippon Electrical Corporation's (NEC) Fundamental Research Laboratory in Tsukuba, Japan, synthesized MWCNTs using arc-discharge evaporation from a carbon anode and characterized them with transmission electron microscopy. Subsequent research by groups at NEC and International Business Machines Corporation (IBM) in 1993 produced SWCNTs by incorporating a metal catalyst into the cathode. Chronologically, CNTs were discovered between the isolation of fullerenes in 1985 and graphene in 2004. A timeline of the key events in the history of CNT research is shown in Figure . These include efforts to synthesize CNTs using new methods or into alternate form factors, understand the properties of CNTs through characterization and modeling, and evaluate and optimize their performance in various applications. Shortly after their discovery, three theoretical studies provided valuable insights on the electronic properties of SWCNTs. By calculating the electronic structures of SWCNTs, these studies predicted that individual tubes could have either metallic or semiconducting electrical properties, depending on their chirality and diameter. Furthermore, they determined that Peierls distortion, which can interfere with the electrical conductivity of 1-D metals, was not a significant effect in metal-like SWCNTs at ambient temperature. The theoretical predictions in these publications would later be confirmed by direct measurement of diameter and chirality using scanning tunneling microscopy, combined with measurement of the electrical properties of individual CNTs. Building on this work, in 1998 research groups at Delft University of Technology and IBM fabricated CNT-based field-effect transistors (FETs), placing individual SWNTs over a siliconbased gate structure between metallic electrodes, similar to a doped Si-based FET. Over the next 15 years, 20 full logic circuits and eventually a computer were built using CNTs as the active components. |
660a22a5e9ebbb4db917b00e | 2 | During this time, researchers began using CNTs in other applications that benefit from their 1-D geometry, for example, using CNTs as field-emission electron sources. In 1998, CNT-tipped atomic force microscopy (AFM) probes were reported, demonstrating the potential of CNTs in enhancing scanning probe microscopy capabilities owing to their inherent high-aspect-ratio geometry. An innovative example of this is the use of carboxyl groups present on partially oxidized ends of the CNT AFM tips as attachment points for other chemical moieties, thereby forming AFM probes that are sensitive to the chemical functionality present on a sample surface. The ability to explore the use of CNTs in a wide variety of applications was enabled by advances in CNT synthesis in the 10 years following their discovery. In 1992, by optimizing the conditions of the arc-discharge process, particularly the helium pressure, Ebbesen and Ajayan were able to generate and isolate gram-scale quantities of MWCNTs. Subsequently, a group at Rice University led by Richard Smalley developed a process based on laser ablation of metal/graphite composite targets to produce SWCNTs with high yield. Another major development around this time was the development of chemical vapor deposition (CVD) methods to synthesize CNTs. In CVD, a carbon-containing precursor gas (benzene and acetylene in these early studies) flows over a substrate at elevated temperatures. Reactions between the precursor and surface then lead to the growth of CNTs from the surface. The degree of morphological control, adaptability, and scalability of CVD has led to it being the most common synthesis route for CNTs today. An example of the control and scalability of CNT CVD is the ability to generate aligned 'forests' of CNTs on a substrate consisting of catalytic iron nanoparticles embedded in silica. A total count of the CNT-related journal and patent publications between 1992 and 2023 is shown in Figure . Research in this field grew rapidly between 1991 and 1999, to roughly 1,000 publications per year by 2000. Between 2000 and 2013, an approximately linear, ten-fold increase in the number of journal publications is observed. Between 2013 and 2017, growth in journal publications appear to have plateaued and remain more or less constant at 10,000 publications per year, while the frequency of patent publications continued to grow, eventually leveling off at ~7,000 per year in 2018 (Figure ). Since 2016, the number of publications generally increased again, driven by diverse, emerging applications including electrochemical CO2 reduction, flexible sensors, and others that will be discussed in more detail in the "Applications of carbon nanotubes" section. Geographical distribution of publications indicates that commercial and non-commercial entities from China (CHN) appear to dominate in the field of CNT-related research with the total number of publications exceeding those from the United States (US), South Korea (KOR) and Japan (JPN) combined (Figure ). Other key countries/regions include India (IND), Iran (IRN), Germany (DEU), Russia (RUS), Taiwan (TWN), France (FRA), United Kingdom (GBR), Italy (ITA), Canada (CAN), Spain (ESP) and Brazil (BRA). The patent-to-journal ratios for these countries/regions indicates that most of them have a higher proportion of journal publications, the exceptions being South Korea and Japan which have a slightly higher proportion of patent publications. India and Iran on the other hand, have an overwhelming number of journal publications as compared to patent publications suggestive of minimal/lack of commercial efforts by organizations in those countries. |
660a22a5e9ebbb4db917b00e | 3 | Next, we identified leading organizations in CNT research by considering both the total volume of publications and average number of citations per publication, the latter parameter being a rough representation of the influence exerted by the institution in this field. These institutions are fairly well distributed geographically, with representation from the United States, Singapore (SGP), the United Kingdom, China, Japan, and South Korea (Figure ) -though 60% of the leading 15 organizations originate from USA. Standford University was identified as having the highest average number of citations per publications, with examples of highly cited publications including the use of CNTs in chemical sensors, field emitters, and FETs. Other important research organizations identified using this method include the University of Pennsylvania (with highly cited papers on topics including CNT-containing nanocomposites and CNT solubilization in water ), and Northwestern University (with highly cited papers on sorting CNTs 39 and aminofunctionalized CNTs 40 ). Our analysis indicates that the top commercial assignees for CNT-related patents include companies involved in the electronics, chemical, battery, and automotive industries (Figure ). Patents from the leading commercial assignees discuss an array of diverse subjects, distributed across a wide range of applications. For example, recently published patent applications assigned to LG Chem, the leading patent assignee, include a reactor to synthesize CNTs that uses a nozzle to feed reactant which generates a spiral flow, 41 a mixed metal/inorganic/CNT catalyst for CO2 electrochemical reduction, 42 and a composite Si/CNT anode for lithium secondary batteries. Recently published patent applications assigned to Foxconn Technology Group (also known as Hon Hai Precision Industry Co., Ltd.), the top electronics manufacturer in terms of assigned CNT patents, include a Si/CNT lithium battery anode, an electrically modulated light source based on aligned CNT films, and a CNT-based transistor structure. Geographical distribution of commercial patent assignees indicates a sizeable contribution by China (40% of the total), Japan, South Korea, and the US. Countries with smaller degrees of contribution include Germany, Taiwan, France, United Kingdom, India and Russia (Figure ). |
660a22a5e9ebbb4db917b00e | 4 | Several methods are known for synthesizing CNTs. In the arc-discharge method used in the initial experiments in the early 1990's, a high DC current is passed between carbon electrodes, leading to carbon evaporation and the formation of MWCNTs. The first demonstration of controlled SWCNT synthesis also used arc discharge, but with the inclusion of a transition metal catalyst. Later studies showed that ablating a carbon target using a high-powered laser could also be used to form CNTs. However, as shown in Figure , the most common method for synthesis of CNTs over the last two decades is based on chemical vapor deposition (CVD), wherein passage of a carboncontaining precursor gas over a substrate in a reactor allows for reaction between the precursor and the surface leading to gradual buildup of CNTs. The substrate can incorporate a catalyst, commonly nanoparticles made of Fe, Ni, Co, or combinations of these metals, which can nucleate CNT growth and allow it to proceed at lower temperature. A catalyst is generally not needed for MWCNT synthesis, but is needed to make SWCNTs, since the size and composition of the catalyst determines the initial stages of CNT growth, which determines the diameter and chirality of the final nanotube. |
660a22a5e9ebbb4db917b00e | 5 | The overwhelming prevalence of CVD can be attributed to advantages associated with it, such as scalability, industrial maturity of the technique, and adaptability and flexibility through the form factor of the reactor. An example of this adaptability is floating catalyst CVD, where catalyst particles are suspended in a stream of gas containing a hydrocarbon precursor, which decomposes and forms CNTs on the catalyst surface. Tuning the parameters of this process can generate a wide range of CNT bulk morphologies. In plasma-enhanced CVD (PE-CVD), a plasma above the substrate is used to partially dissociate precursor molecules before surface reactions take place. This allows CVD to be done at lower temperatures, which allows better control of CNT properties in some cases. A significant challenge in CNT synthesis by CVD is achieving the required selectivity for chirality, diameter, and other structural parameters for a given application. The CVD process has several parameters that can be used to control structure, including the catalyst type, catalyst morphology (thin film, supported, nanoparticle), gas precursor composition, template, and reaction temperature. Our analysis indicates that the frequency of publications associated with synthesis of CNTs began to decline around 2006. However, while the number of studies on the synthesis of CNTs in general appears to be decreasing, there are some areas of continued active research. For example, some emerging CNT applications require very precise control over chemical and electrical properties, so tuning synthesis parameters to achieve those properties continues to be an active area of research. One particular area of focus is controlling the distribution of chirality. For example, in electrical applications, high selectivity for a particular chirality is needed because the chirality controls the bandgap of a SWCNT. The most widely studied parameters to control chirality involve the catalyst, in particular varying the catalyst particle size, crystal structure, support, and melting point. |
660a22a5e9ebbb4db917b00e | 6 | Owing to the unique combination of material properties, CNTs are used in a wide variety of applications where they play critical functional roles. Figure shows the distribution of carbon nanotube applications referenced in journal and patent publications. The largest fraction of CNT publications focused on their use in composites, including polymermatrix composites (Figure ). The structure of CNT gives them strength along their longest dimension, while the length of CNT allows them to act on macroscopic length scales. In addition, the ability to functionalize the exterior of CNT either covalently or noncovalently can be used to make them compatible with a variety of materials. The electrical and thermal conductivity of CNT allows them to act as functional components and not just as structural components. Composites are used in many products on both large and small scales; their ubiquity and the useful characteristics of CNT in them can account for their large fraction in CNT applications. Some of the applications for CNT in composites include their use in batteries, in biocompatible hydrogels for heart cell growth, for use with a glycidyl polyketone to form nonconductive layers, to increase strength and wear resistance in composites with high-density polyethylene, to absorb high-frequency electromagnetic waves (CN117042427A 67 ), and as components in heatresistant aerogels (CN114275770 68 ). One caveat with the analysis of composites is that the term "composite" can refer to both a direct use of CNTs for their physical properties (i.e. structural composites where CNTs are incorporated into a matrix to provide mechanical strength, or electrical or thermal conductivity) and a more general structure type in which CNTs are combined with other components (such as their use as catalyst supports); classification as a composite thus may not exclude the use of CNTs in other applications. |
660a22a5e9ebbb4db917b00e | 7 | The largest fraction of patent publications on CNT applications was for use in batteries (Figure ). The length of CNT and their conductivity makes them useful as molecular wires, and their strength could allow them to reduce the expansion and contraction of battery materials during charging/discharging, which leads to component separation and battery failure, or to form a barrier to dendrites, preventing short circuits and their associated hazards. These characteristics motivate the use of CNT in batteries such as (with ZnO and carbon felt) in lithium-ion (Li-ion) battery electrodes resistant to lithium dendrite formation, and materials such as graphene networks, NiCoO2, MnO2, nitrogen-doped carbon, and cobalt-iron metal-organic frameworks in biocompatible and hybrid zinc or in zinc-air batteries. One feature of note is that battery applications represent a much higher fraction of patent publications compared to journal publications (more journal publications are devoted to CNT-containing sensors than CNTcontaining batteries). Many of the patents for batteries originate from China or South Korea; a push to transition from internal combustion to electrical power for transportation, and the large consumer electronic industries in those nations likely motivate battery development and commercialization. For recyclable batteries, laboratory-scale battery manufacturing and testing can differ significantly from the methods required to produce batteries on larger scales, making the type of research performed for journal publication difficult to directly implement on large production scale, perhaps decreasing commercial interest in CNT exploratory research for batteries. The ratio of patent-to-journal publications in the use of CNT in batteries (Figure ) is consistent with other battery research; for example, in Li-ion battery recycling, patent publications exceeded journal publications by nearly a factor of Sensors incorporating CNT are a significant area of research, comparable to CNT/Polymer composites (Figure ). The strength and aspect ratio of CNT makes them able to transmit physical forces, while their electrical conductivity makes them useful at converting and transmitting stimuli. In addition, CNTs have a large specific surface area, increasing their susceptibility to external changes. A variety of sensors using CNT have been assembled, such as a pressure sensor using conductive polymers, CNT, and poly(dimethylsiloxane) (PDMS) (CN2023116972905A 75 ), CNT supported on polyurethane for use as a triboelectric generator, movement sensor, or chemosensor, with superhydrophobic surface as a flexible airflow sensor, in concert with nickel triphenylene complexes as a nonbiological glucose sensor, and as nanoresistors to detect disinfectants. MWCNT have been used in combination with samarium to detect solvent vapors, with polyaniline (PANI) to form ammonia-responsive materials, with epoxy resin to form a strain-sensing conductive material, and with polyurethane and polypyrrole to form a bendable pressure sensor 83 (in the latter sensor, the use of components with differing compressibilities was critical to its effectiveness). The fraction of patent publications using CNT for sensors was smaller than the corresponding fraction of journal publications, making up 14% of journal publications but only 7% of patent publications (Figure ). One possible explanation is that sensor fabrication on laboratory scale may require more labor to assemble or more expensive materials to be readily transferable to large scale; alternatively, laboratory-conceived sensors may either be proofs of concept for sensor design or may utilize target stimuli less amenable for further development. |
660a22a5e9ebbb4db917b00e | 8 | Other applications of CNT make up significant fractions of the publication record. Biomedical uses of CNT make up 12% of journal publications, though only 4% of patent publications. CNTs have been used as a reinforcing material for drug-delivering stents made with biocompatible polymers (WO2014186532A1 84 ), as a conductive material with poly (3,4-ethylenedioxythiophene) (PEDOT) used as a coating for nerve catheters (CN2023116983475A ), as an antibacterial and structural material for bone implants. CNT have also been used as delivery vehicles for drug and gene moieties. While CNTs are readily functionalized, allowing their surface chemistry and charge to be altered appropriately for bioavailability, they can also exhibit toxicity depending on their purities, sizes, shapes, and whether they are integrated into macroscale forms, and may require dispersion or solubility in water; the properties necessary for the desired activities may lead to side effects, making them less commercializable. The safety of CNTs generally is an active area of research and policy debate. Besides sensors, CNTs have also been incorporated into other electronic materials accounting for 8% and 9% of journal and patent publications, respectively. As noted earlier, the ability to functionalize CNTs (with the purpose of modifying their electronic and chemical properties) is important. The orientation and morphologies of CNT can also be manipulated to achieve the desired device properties, but the manipulation of these properties is often manual and may require significant optimization to be produced on larger scale. The relative amounts of patent and journal publications indicate that the benefits of CNT design likely outweigh impediments to their commercial uses. Some of their potential uses include as components in FETs (WO2022040565 ) for instrument amplifiers, with bamboo-derived fibers, phenolic resins, and cobalt in a layered electrode material, and in a conductive fabric. Environmental remediation is a significant application for CNT, with roughly 6% of journal publications and 4% of patent publications discussing it (Figure ). The high surface area and conductivity of CNT are important properties, as they make CNT available for absorption and for mediating interactions with pollutants. The ability to modify CNT allows them to be tailored for other co-catalysts or for reaction with pollutants. For example, lanthanum-and iron-modified CNT have been disclosed for environmental use (CN2021113842883A 96 ), while a lanthanumgadolinium-iron oxide-perovskite reinforced with CNT showed improved photocatalysis of the degradation of phenol red over the CNT-free perovskite. A CNT membrane was used to activate MnO2 for oxidative degradation of water pollutants. Journal articles describing the use of CNT in environmental remediation are significantly more numerous than patent publications (Figure ); the disparity may be due to the lack of economic incentive in environmental remediation which makes incremental improvements or improvements with more expensive materials difficult to implement. Even though SWCNT can conduct electrons well and have high surface area, enabling improved reactivity and utility, their expense may make their use in environmental remediation difficult because the increased costs cannot be recovered by increased prices. |
660a22a5e9ebbb4db917b00e | 9 | Besides batteries, the use of CNTs in adhesives is the only other application with a high patentto-journal ratio (Figure ), which suggests considerable commercial interest in this area. The shape of CNTs could allow them to become entangled; alternatively, functionalized nanotubes could react with surfaces with their strength making them strongly adherent. For example, CNT were used to reinforce PDMS dry adhesive pads; the CNT improve the stability of the pads and diffuse concentration of charge at the PDMS surface which leads to surface deformation and loss of conformity. Alternatively, a poly(ethyleneimine), CNT and epoxy resin material was used as an adhesive. Here, the nanotubes improve the strength/mass ratio of the material, while adhesion depends on the fraction of poly(ethyleneimine) used. Selected patents describe the use of CNT for strengthening concrete (RU2669835 101 ) and the use of a trans-stilbene/maleimide polymermodified CNT to make an adhesive with improved thermal conductivity (CN2023116970357A 102 ). |
660a22a5e9ebbb4db917b00e | 10 | The co-occurrence of SWCNTs versus MWCNTs for different applications is shown in the form of a Sankey graph (Figure ). The choices of which type of CNT to use in different circumstances depends on the necessary properties and their costs. SWCNT can be obtained as both semiconducting and metallic CNT. Conducting MWCNT have larger thermal and electrical conductivities than SWCNT 103 and may have a narrower range of conductivities. SWCNTs, however, have more predictable structures, making them more easily and predictably functionalized. In addition, SWCNTs have a Raman absorption at 532 nm and near-IR absorptions between 1000 and 1700 nm; 105 the near-IR absorption is in a region at which tissue is most permeable, making them amenable to use in biological sensing. The more predictable structures of SWCNT may allow them to be more predictably and easily functionalized by covalent and noncovalent methods. SWCNTs are roughly 1 nm in diameter, while MWCNT can have diameters between 2 and 100 nm. Purified SWCNT with reduced metal content can be obtained for 2000 USD/kg 106 though a 99% SWCNT content grade with <0.1% metal content is available, and MWCNT can be obtained for 100 USD/kg. As shown in Figure , for most of the applications of CNT, MWCNTs are predominantly used, likely because they are significantly less expensive than SWCNTs. Publications discussing solar cells, hydrogen storage, and actuators, however, use SWCNT and MWCNT at similar frequencies. SWCNTs are also used fairly extensively in biomedical applications, where their effective absorption and fluorescence properties at near-IR wavelengths make them amenable to tissue penetration and potential use in living things. SWCNT are likely to have more predictable functionalization behavior and toxicities, important for biocompatibility. |
660a22a5e9ebbb4db917b00e | 11 | Sensors and electronics use SWCNT less frequently than MWCNT but at significant frequencies. Both SWCNT and MWCNT can be electrically conductive, but SWCNT can also be semiconducting; electronics needing semiconducting CNT thus are likely to need SWCNT. SWCNTs are likely to have more predictable electronic behavior, while the presence of a single structure instead of multiple structures allows them to be physically manipulated and aligned, necessary for use in field-effect transistors and other small device components. On nanometer scale, SWCNTs are likely the only sufficiently predictable CNT that can be used, and on small scale or for proof of principle, known CNTs are important for understanding the behavior of the devices made using them. The single walls of SWCNT makes the paths of electrons through the CNT more responsive to external stimuli and thus more effective as sensors. The availability of SWCNT with defined size and electronic MWCNT can be used in gas sensors, however, 103 and so their reduced cost may make them useful alternatives to SWCNT in nonbiological sensing. Transistors and photonics use SWCNT almost exclusively. Publications discussing composites show the use of SWCNT and MWCNT at similar rates. Many of the functions that are needed for composites can be performed both with SWCNT and MWCNT; both have similar Young's moduli and both can be thermally and electrically conductive. The choice of SWCNT or MWCNT may depend mainly on the fraction of nanotubes that can be tolerated in the composite. 111 In Figure , the landscape of emerging trends in the field of CNTs was mapped by identifying concepts that were found the most frequently in literature from 2020-2022, or that increased significantly in usage in the literature during that time using a NLP-based method. These emerging concepts were grouped into three major areas: properties, materials, and applications with further sub-categorization within each area. Within properties, two notable concepts that were frequently found in CNT literature were thickness and thermal conductivity. Thickness appears to primarily occur in publications wherein CNTs are incorporated in microwave absorbing and electromagnetic interference (EMI) shielding composites, to describe the optimal or tested physical thickness of the shielding material. Thermal conductivity is commonly used in the context of using CNTs to improve the thermal properties of composite materials, and for that reason there is no dominant application that results in the high use of this term. Recently, these composite materials have included phase change materials (PCMs), which are used to store energy in the form of latent heat. In these materials, CNTs are incorporated with a bulk phase change material, such as paraffin, where the high thermal conductivity of the CNTs reduces the amount of time needed for the phase change material to take up and release thermal energy. A major advantage to using CNTs for this purpose is that they can improve thermal conductivity in phase change materials significantly at relatively low concentrations, as shown in a recent study where adding 5 weight % CNTs increased the thermal conductivity of a polyurethane PCM by 2.3 times. In addition, CNTs are used in solar water evaporation devices which convert seawater to freshwater, 118 where low thermal conductivity is desirable to prevent heat loss through conduction. In these devices, CNTs are used primarily to take advantage of their high photothermal conversion efficiency. Recently, it has also been demonstrated that aligned CNTs can be incorporated into self-healing polymer matrices, providing high thermal conductivity to intrinsically self-healing materials; 119 both are critical properties in thermal interface materials used in high-power and dynamic devices. |
660a22a5e9ebbb4db917b00e | 12 | The properties associated with the fastest growth in CNT-related publications are brittleness and exergy. Prominent examples of the use of the term brittleness include reinforcing materials with CNTs to reduce their intrinsic brittleness, including hydrogels (primarily for use in sensors), poly(lactic acid) (PLA) 122 (enabling the use of this biodegradable polymer for mechanically demanding applications), epoxy resins, 123, 124 conjugated microporous polymer membranes, 125 and silica aerogels. Other examples of brittleness relate to the use of CNTs as replacements for brittle, traditional materials used in battery components and other applications. Brittleness also appears in the context of strain sensors which combine brittle materials and CNTs in composites, where cracking of the brittle material in response to strain improves the piezoresistive response, resulting in a higher gauge factor. |
660a22a5e9ebbb4db917b00e | 13 | The majority of references where the concept of exergy is referenced involve the use of CNTs as thermal conductivity additives in nanofluids used for heat transfer, specifically in photovoltaic/thermal (PVT), solar thermal, 134 and combined PVT/PCM 135 applications, where exergy efficiency is used as a measure of the performance of these fluids. |
660a22a5e9ebbb4db917b00e | 14 | In the "Materials" branch, terms with an especially high growth rate (>1.5 fold increase in publications over 2020-2022) include poly(butylene adipate-co-terephthalate) (PBAT), ciprofloxacin, biochar, cellulose nanofibers, and MXene nanosheets. This branch includes both materials that are often used in combination with CNTs as well as materials that are detected or removed using CNTs. |
660a22a5e9ebbb4db917b00e | 15 | PBAT is a biodegradable polymer which has been recently used to make electromagnetic shielding materials in combination with CNTs. Because of its mechanical strength, PBAT has also been combined with PLA by melt-blending to make PBAT/PLA/CNT composites. Ciprofloxacin is a commonly used antibiotic which is difficult to remove via conventional wastewater treatment methods, and has been detected at significant levels in the environment. It is toxic to certain aquatic organisms, 143 and can interact with bacteria in the environment, contributing to a rise in antibiotic resistant pathogens and impacting wastewater treatment processes. CNTs can be used to remove ciprofloxacin from water through adsorption and catalytic processes. When used as adsorbents for wastewater, CNTs can be modified with magnetic nanoparticles such as ferrites 145 or Fe2O3, 146 allowing them to be separated downstream using a magnetic field. Examples of catalytic removal of ciprofloxacin include those where CNTs perform a catalytic function (through the generation of OH radicals 147 ), and where CNTs are combined with other catalytic materials, or are used to support oxidizing compounds in ultrasound-assisted oxidation processes. The co-occurrence of biochar with CNTs seems to be driven primarily by studies which compare the performance of the two materials for their ability to adsorb or otherwise treat environmental pollutants, reflecting the growing importance of this application. To a lesser extent, biochar and CNTs have been combined in composites for heavy metal adsorption and energy storage; biochar can also be used as a precursor for CNT production. Cellulose nanofibers are similar in size and shape to CNTs, and have recently been combined with CNTs in a variety of applications. These include conductive, hydrophilic aerogels made using a mix of cellulose nanofibers and CNT, 158, 159 hydrogels made using a mixture of 2,2,6,6tetramethylpiperidine-1-oxyl (TEMPO)-oxidized cellulose nanofibers and CNTs. Cellulose nanofibers and CNTs can also be combined in composite electrodes, where 3D networks of the two materials are used as a support for active materials. The rapid growth of publications involving both MXenes and CNTs is not unexpected, since MXenes have been a rapidly growing field of research since their discovery in 2011. They have been tested in a wide range of applications, including energy, catalysis, and biomedicine. Sensors has been a relatively recent application area for MXenes, with 2017 having been identified as their first significant use in this field. A recent example is a piezoresistive strain sensor based on a composite thermoplastic polyurethane (TPU)/CNT foam with a layer of MXene (Ti3C2Tx) deposited on its surface. The deformation and cracking in both the foam and MXene layers results in a change in resistance per unit strain and is much higher than the sum of analogous sensors made using only a TPU/CNT foam, or a TPU foam/MXene sensor without CNTs. In this example, the CNTs and MXenes are not combined on a microscopic level, but their complementary properties lead to a synergistic effect. |
660a22a5e9ebbb4db917b00e | 16 | CNTs can also be combined with MXene nanosheets on a nanoscale level, making 2-D/1-D composites composed of alternating layers of individual MXene sheets/CNTs, for applications including battery electrodes and separators, 167 supercapacitors, 168 EMI shielding, 169 water splitting, 170 and sensors. In these materials, CNTs are used for a variety of reasons, including to control the spacing between MXene sheets and to prevent sheets from restacking (to improve ion transport in battery electrodes, for example), to improve mechanical properties such as tensile strength, 166, 174 electrical conductivity, 170 and electromagnetic shielding and absorption. CNTs can also be used as a strategy for embedding metallic nanoparticles between MXene sheets. Turning to the "Applications" branch, energy storage is the most active area in terms of total number of publications, with battery research driving much of that growth, as shown in Figure . Other notable application areas with a large number of publications in 2020-2022 include sensors, environmental, and biomedical (Supplementary Figure ). Fewer total publications, but more rapid growth, is seen in electronics and energy conversion applications. Full time trends of journal and patent documents in these areas are shown in Figure . |
660a22a5e9ebbb4db917b00e | 17 | The breakdown of CNT-related publications for different battery chemistries is shown in Figure . Li-ion batteries dominate, but Zn-ion batteries represent the highest growth. In Zn-ion batteries, prominent examples of the use of CNTs are in the cathodes of aqueous Zn-ion batteries, in combination with KV3O8•0.75H2O (KVO), 177 ZnMn2O4 (ZMO), 178 M phase VO2, 179 NaxV2O5•nH2O(NVO), 180 and V2O5. In these cases, the role of the CNT is to form a high surface area, mechanically resilient, flexible, and electron-conductive scaffold that supports these materials. The goal of improving these properties is to improve the rate capability and capacity retention after repeated cycling. CNTs are also used to enable Zn metal anodes, either as a way to control Zn deposition, preventing the formation of dendrites, 182 or as a scaffold for Zn nanosheets. Research in triboelectric nanogenerators and thermoelectric generators are two other emerging applications of CNTs (represented as "energy conversion in Figure ). Triboelectric nanogenerators are devices which can generate electrical energy from many different types of motion, including human motion, or the motion of plant leaves in the wind. The origin of this electrical energy is charge separation that takes place during the mechanical interaction of two surfaces. In this application, CNTs are used to increase the surface charge density, 186-188 thereby increasing the output power of the device. CNT-containing triboelectric nanogenerators have many possible applications, including self-powered sensors, or providing power to wearable devices. Hydrogen storage shows a significantly different trend compared to the other energy topics, with generally flat or negative growth since 2010. There appear to be multiple reasons for this. One reason is the difficulty in obtaining reproducible and accurate quantitative measurements of hydrogen storage capacity, the other being that it is now thought that pure CNTs have a fairly low gravimetric storage capacity for hydrogen at room temperature, 192 below the initial US Department of Energy goal set in 2010 of 6.5 wt% for automotive H2 storage applications (later reduced in 2015 to 5.5 weight %). However, work continues on the use of modified CNTs for hydrogen storage, for example CNTs loaded with metals or heteroatoms such as N. A final application to highlight in Figure is the use of CNTs in flexible sensors and electronics. CNTs are used here primarily for their electrical properties, which can impart piezoresistive, piezocapacititve, 195 and other properties useful for sensor applications. This can involve incorporating CNTs into matrices, including hydrogels, or combining them with polymers (for example, in a polyurethane/CNT yarn with hierarchical structure 197 ). Flexible sensors can also be fabricated by printing CNT-containing ink onto flexible substrates. The major emerging application in these cases is monitoring human motion, either through wearable sensors or electronic skin, 199 as shown in Figure . The other prominent biomedical application of CNTs is drug delivery. However, publication trends in this area (Figure ) indicate this is a relatively mature field of research. |
660a22a5e9ebbb4db917b00e | 18 | As shown in Figure , another emerging application for CNTs is in the electrochemical reduction of CO2. This can be used as an approach for carbon capture that simultaneously generates a variety of useful chemical products. CNTs can be used in several ways related to catalyzing this process, including as a scaffold for supporting single-molecule metal based catalysts, covalent organic frameworks (COFs), and pipet-like bismuth nanorods. CNTs, particularly when they are doped with nitrogen or other atoms, can also catalyze CO2 reduction, acting as metal-free electrocatalysts. To better understand the use of CNTs in composites, we have broken down the types of composites by the matrix material (left), application (center), and the most commonly used matrix substances (extreme right) and showed their co-occurrences in a Sankey graph (Figure ). This data reflects the common use of polymers in CNT composites, for a variety of applications. Among these, the use of conductive polymers including PANI, polypyrrole, and others in energy storage and sensor applications is especially prominent. Examples include the use of a PANI/SWCTN composite in Al-ion batteries, 209 a combination of polypyrrole and CNTs to enable the use of black phosphorus in Li-ion battery anodes, 210 and the use of CNTs and polypyrrole in aqueous Zn-ion battery cathodes. A similar breakdown of matrix material (left), application (center), and commonly used matrix substances (extreme right) with respect to patent publications can be found in the Supplementary Information. |
660a22a5e9ebbb4db917b00e | 19 | When CNTs are used as additives to improve the mechanical, thermal, or electrical properties of a polymer, a key to effective incorporation of the CNTs is preventing aggregation of the CNTs and maximizing their interaction with the polymer matrix, which can be done through functionalization of the CNTs. This analysis also highlights the frequent use of metal-organic frameworks (MOFs) in energy storage applications. A recent example of this is the use of a cerium-based MOF / CNT composite separator in a Li-S battery. Here, the CNT structure provides electrical conductivity and is an effective physical barrier for polysulfide shuttling, while the MOF catalyzes the breakdown of larger polysulfides into Li2S. In another example, MWCNTs were used as a template for the growth of NiCo-based MOFs, improving their performance as supercapacitor electrodes. 215 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.