id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
620a3ad70c0bf0c968e8e97a | 3 | Through treatment of HCPs with sulfonating agents, heterogeneous acid catalysts combining the desirable properties of HCPs with high acid-site densities, so called sulfonated HCPs (SHCPs), are produced. Dong et al. synthesised sulfonated hypercrosslinked benzene networks containing up to 2 mmol•g -1 of SO3H for the selective dehydration of D-(-)-fructose to 5-hydromethylfurfual, reaching conversions of >90%. An analogous set of materials catalysed the selective conversion of carbohydrate substrates to ethyl levulinate in yields of up to 67%. Therein, it was speculated that multiple variables, including surface area, catalyst dose, and acid density played a critical role in the efficiency of the catalytic transformation. Sulfonated carbazole-based HCPs with impressive SO3H densities of 3.7 mmol•g -1 were developed for the esterification of a number of fatty acids to produce biofuels, reaching conversions of up to 99%. Elsewhere, a SHCP containing a phosphonium salt in its structure was used for the two-phase hydrolysis of cyclohexyl acetate to cyclohexanol, achieving conversions and selectivity of up to 93 and 95%, respectively. SHCPs have also shown promise as selective adsorbents, including for propene/propane separation, heavy metal adsorption, and water pollutants. In conventional approaches, sulfonic acid groups are introduced via the post-synthetic treatment of HCPs using sulfonating agents, commonly chlorosulfonic acid. This route involves the swelling of the HCP network in inert solvent, followed by treatment with large excesses of the sulfonating agent. Although effective, control over the degree of sulfonation and the resulting porous properties of the catalyst is lacking. Furthermore, the approach is cumbersome, requiring the complete synthesis and work up of the HCP prior to sulfonation, resulting in a synthesis time of up to 6 days. |
620a3ad70c0bf0c968e8e97a | 4 | We present a strategy for the synthesis of SHCPs in a one-pot approach, significantly reducing reaction times and reagent use compared to conventional approaches. These SHCPs allow for careful catalyst design and have superior properties to SHCPs obtained using other synthetic procedures. Hazardous organic solvents and reagents are significantly reduced or eliminated altogether, making the method simple, safer, cleaner, and more economical. To assess the potential of these SHCPs as solid acid catalysts, we report their application to the hydrolysis of cyclohexyl acetate. The properties of these finely-tuned SHCPs had a dramatic effect on their performance in this two-phase reaction, allowing for the identification of key catalyst properties for optimal catalytic performance. |
620a3ad70c0bf0c968e8e97a | 5 | We synthesised ten SHCPs from 4,4'-bis(chloromethyl)biphenyl using chlorosulfonic acid as a dual polymerisation catalyst and sulfonation agent (a general reaction scheme is shown in Figure ). Throughout the set, the ratio of chlorosulfonic acid to the aromatic monomer was varied to investigate the effect on the resulting SHCPs' textural and chemical properties. The progression from SHCP-1 to SHCP-10 follows an increasing catalyst-to-monomer ratio, precise values of which are given in Table . Briefly, 4,4'-bis(chloromethyl)biphenyl was dissolved in 1,2dichloroethane (1,2-DCE) and cooled using an ice bath, after which a solution of chlorosulfonic acid in 1,2-DCE was added to initiate polymerisation. Upon formation of a solid, the mixture was sealed and heated at 80 °C for up to 22 h. The resulting solid was then washed with methanol before drying to yield the SHCP. More details are given in the supporting information. It is worth noting that a shortened procedure in which the heating time was reduced to 3 h showed no apparent detriment to resulting SHCP properties. Polymers were typically produced in yields of >90 % when considering the incorporation of the acid catalyst as a sulfonic acid group. Networks ranged from a beige to dark brown colour with increasing catalyst to monomer ratios (Figure ). |
620a3ad70c0bf0c968e8e97a | 6 | We confirmed the successful formation of SHCPs using C cross-polarisation/magic angle spinning solid-state NMR (CP/MAS ssNMR) (Figure ). Signals at ~38 ppm are assigned to methylene bridges from newly formed crosslinks in all cases. A weak signal assigned to C-Cl at ~42 ppm can be seen in SHCP-1 and decreases in intensity through to SHCP-3, indicative of incomplete crosslinking due to low concentrations of chlorosulfonic acid. A signal at ~120 ppm is found in the shoulder of a larger peak in samples SHCP-5 -SHCP-10, assigned to the C-S bond formed during sulfonation. At lower ratios of catalyst to monomer this peak can no longer be seen due to reduced degrees of sulfonation. All SHCPs showed strong signals at ~129 and ~139 ppm, corresponding to aromatic (CAr-H) and quaternary, or substituted, aromatic carbons (CAr-R), respectively. Although difficult to quantify due to overlapping peaks, it is clear that the ratio between substituted and unsubstituted aromatic peaks changes across the polymer set. |
620a3ad70c0bf0c968e8e97a | 7 | Increasing amounts of chlorosulfonic acid in the initial formulation appear to increase the ratio of substituted to unsubstituted aromatic carbon. Considering the increase in C-S peak as well as the disappearance of the C-Cl peak with higher initial concentrations of chlorosulfonic acid, this is in line with expectation. We also confirmed successful polymer formation and sulfonation using Fourier-transform infrared spectroscopy (Figure ). There, the presence of bands assigned to both SO3H and C-S bonds emerged with increasing ratios of catalyst to monomer, confirming increasing sulfonation. The diminishing of the C-Cl band across the series was also observed, in good agreement with ssNMR. We used X-ray photoelectron spectroscopy (XPS) to gain a more in-depth understanding of SHCPs' chemical compositions (XPS derived quantitative data for all polymers is provided in Table ). The main component of the high-resolution C 1s spectra observed at a binding energy of 284.8 eV is attributed to C-C bonding, encompassing both sp2 aromatic carbon and sp3 carbon in methylene crosslinks (SHCP-10 shown in Figure ). A peak of lower intensity is observed at 286.8 eV, corresponding to C-S. A broad, low intensity π-π* shake-up feature is also observed at 291 eV. High resolution S 2p spectra showed a typical asymmetrical peak for a sulfonic acid moiety (SHCP-10 shown in Figure ). The peaks at binding energies of 168.5 eV and 169.5 eV are assigned to S 2p3/2 and 2p1/2, respectively. This peak is present throughout all sulfonated networks, confirming the vast majority of the sulfur present exists in sulfonic acid groups. Further evidence for this is provided by a consistent ratio of 1:3 for S to O across all samples, as would be expected for sulfonic acid groups. Increasing catalyst to monomer ratios led to higher concentrations of sulfonic acid sites in the networks, reaching a maximum sulfur content of 4.6 mmol•g -1 in SHCP-10, as determined by XPS. This is amongst the highest degrees of sulfonation reported for hypercrosslinked polymers, with another similarly high example reported for carbazole based networks (3.7 mmol/g, as determined by elemental analysis). We hypothesize that polymerisation occurs more rapidly than sulfonation, leading to the uniform distribution of chlorosulfonic acid throughout the HCP structure. This dispersion facilitates excellent acid densities throughout networks upon sulfonation, as it eliminates the need for the sulfonation agent to permeate into a pre-formed structure. |
620a3ad70c0bf0c968e8e97a | 8 | We employed CHNS-O elemental analysis (EA) to measure the bulk chemical composition of the networks (Table ). The EA results showed a similar trend to XPS, with sulfonation densities increasing upon an increase in catalyst-to-monomer ratio, confirming a good dispersion of acidsites both at the surface and throughout the bulk of the networks. At higher catalyst-to-monomer ratios (≥2), the EA results deviate quite significantly from XPS, with SHCP-10 measured as containing 3.8 mmol•g -1 of SO3H compared to 4.6 mmol•g -1 in XPS. This is due to water adsorption owing to the increased hydrophilicity of the networks with higher sulfonation densities. This can be observed in the S:O atomic ratios calculated from EA, which ideally should be 1:3 but is as high as 1:5.4 in SHCP-10. Furthermore, XPS cannot consider H when determining at.% of elements present, leading to a slight overestimation. Therefore, the real concentration of SO3H likely lies somewhere between the values determined by XPS and EA. |
620a3ad70c0bf0c968e8e97a | 9 | We measured N2 sorption isotherms for all SHCP polymers at 77 K to gain information about their porous properties (Figure and). The sulfonated networks showed excellent reproducibility regarding their porous properties. Networks displayed characteristics of both Type I and Type IVa isotherms, with uptake at low relative pressures signalling microporosity and significant hysteresis observed upon desorption deriving from capillary condensation, indicative here of mesoporosity/macroporosity. All SHCPs displayed H2 type hysteresis curves, indicative of broad pore size distributions with narrow pore necks. The absence of a closed hysteresis loop is common in the isotherms of HCPs and likely due to the swelling of the networks during adsorption. With increasing sulfonation density the hysteresis loops upon desorption appear less pronounced, indicating that the mesopore component of the networks decreases. This is reflected in the values for total pore volume, VTOT, and micropore volume, VMICRO (Figure , Table ), which confirmed the broad pore size distribution. Intriguingly, the micropore volume remains relatively constant with increasing degrees of sulfonation from SHCP-3 onward, permitting the careful control of micropore-to-mesopore/macropore ratio in SHCPs. |
620a3ad70c0bf0c968e8e97a | 10 | The BET specific surface area (SSABET) of all SHCPs was >530 m 2 •g -1 (Table ), with SHCP-1 displaying the lowest SSABET of 539 m 2 •g -1 . A trend can be seen across the catalyst set, with SSABET increasing dramatically to a maximum of 1059 m 2 •g -1 in SHCP-3 before gradually decreasing to 688 m 2 •g -1 in SHCP-10. A clear trade-off between porous properties and sulfonation density was observed. In SHCP-1 and SHCP-2, inefficient hypercrosslinking occurs due to the relatively low catalyst to monomer molar ratios, as evidenced by C-Cl in ssNMR and the presence of higher concentrations of residual Cl in XPS (Table ). Although SHCP-3 displays the highest SSABET of the set, some evidence of remaining C-Cl was again seen during characterisation. It is likely that the degree of sulfonation is insufficient to become significantly detrimental to the porous properties. In previous examples, the self-condensation of 4,4'bis(chloromethyl)biphenyl catalysed using triflic acid yielded HCPs with SSABET of up to 1842 m 2 •g -1 , much higher than the approach herein. However, the production of SHCP-3 requires just one third the amount of acid catalyst used therein. When further increasing the chlorosulfonic acid concentration, as in SHCP-4 to SHCP-10, the resulting SSABET worsen due to the increased sulfonation. Even with the trade-off between SSABET and acidity, these materials have excellent porous properties considering their high sulfonation density. The carbazole-based SHCP with a comparable SO3H concentration of 3.7 mmol•g -1 displayed a SSABET of just 346 m 2 •g -1 . Table . Summary of SO3H content and textural properties of SHCPs produced from various ratios of catalyst to monomer, including BET specific surface area, SSABET, volume of micropores, VMICRO, and total pore volume, VTOT. |
620a3ad70c0bf0c968e8e97a | 11 | Catalyst:Monomer We analysed the thermal stability of all SHCPs using thermogravimetric analysis (TGA) in either an inert N2 atmosphere (Figure and Figure ) or in air (Figure ). Negligible weight loss is observed in SHCP-1 and SHCP-2 until >250 °C. With increasing sulfonation density the weight loss prior to and during an isothermal step at 110 °C increases dramatically, reaching a maximum of 21 wt.% for SHCP-10, indicative of a significant increase in water adsorption at high sulfonation densities. XPS measurements showed that the chemical composition of the networks remained unchanged after heating overnight at 120 °C under N2 flow, confirming degradation was not responsible for the weight loss seen in the TGA (Table ). Finally, we employed powder X-ray diffraction (PXRD) to confirm the amorphous nature of all networks (Figure ). The synthetic approach to SHCPs described herein provides a set of acidic non-soluble networks in which the sulfonation density, hydrophilicity, and textural properties can be carefully controlled. |
620a3ad70c0bf0c968e8e97a | 12 | This presents a unique opportunity for the application of such a set of materials to acid-catalysed conversions in order to unveil optimal catalyst properties. To investigate the catalytic ability of SHCPs, the organic-aqueous biphasic liquid hydrolysis of cyclohexyl acetate to cyclohexanol was selected as a model reaction (reaction scheme in Table and mechanism in Figure ). We investigated substrate conversion and selectivity across our SHCP series (Table ). Briefly, SHCP was added to a two-phase cyclohexyl acetate and water system, which was heated at 120 °C under autogenous pressure for 3 h. The reaction products were extracted using ethyl acetate and analysed using gas chromatography-mass spectrometry (GC-MS). A complete set of GC traces are shown in Figure . It is worth noting that all catalysis reactions were repeated at least four times. |
620a3ad70c0bf0c968e8e97a | 13 | Network SHCP-1 showed no measurable catalytic activity due to low acidity (S content of 0.1 mmol•g -1 ). Control experiments with non-acidic hypercrosslinked polymer equivalents confirmed that no measurable conversion was achieved without sulfonation. The more acidic SHCP-2 (S content of 0.4 mmol•g -1 ) gave substrate conversions of 31% with a high standard deviation of ± 18%, suggesting that this polymer lies at a critical concentration upon which catalytic activity depends highly on network acidity. The remaining SHCPs all achieved ≥60% conversion of cyclohexyl acetate, outperforming both the commercial resin Amberlyst-15 and homogeneous (liquid) H2SO4. Interestingly, high acid densities did not appear crucial for catalyst performance. |
620a3ad70c0bf0c968e8e97a | 14 | SHCP-3 contained just 1.15 mmol•g -1 of SO3H sites and achieved a substrate conversion of 68 ± 8%. Catalytic conversion reached a maximum of 75 ± 2% in SHCP-5 and with increasing sulfonate content (i.e. SHCP-6 to SHCP-10), catalytic conversion decreased, reaching 60 ± 8% in SHCP-10, with a SO3H concentration of 3.8 mmol•g -1 . This emphasises the importance of other catalyst properties as well as acidity for this conversion. Figure shows each polymer in the biphasic cyclohexyl acetate/cyclohexanol and water system after catalytic conversion. Networks SHCP-5 to SHCP-8 formed quasi-stable emulsions during catalytic conversion, demonstrating amphiphilicity. As the reaction is comprised of both an organic and aqueous phase, the formation of emulsions improves the dispersion of one phase throughout the other and ensures that the catalyst resides at the organic/aqueous interface. Additionally, improved mesopore/macropore content ratios in SHCP-3 -SHCP-5 may provide good phase transfer ability, while their high surface areas ensure the substrate good access to active sites, improving conversions compared to materials with reduced mesopore/macroporosity. Selectivity in all materials was >99% for cyclohexanol, with the formation of some trace quantities of cyclohexanone in many examples. |
620a3ad70c0bf0c968e8e97a | 15 | We present a rapid one-pot method for the synthesis of polymeric solid acid catalysts with carefully controlled properties using a dual polymerisation catalyst and sulfonation agent. Ten SHCPs were prepared with varying porous properties, acidity, and hydrophilicity by simply varying the initial catalyst to monomer ratio. These materials could be obtained in less than 24 h by combining the polymerisation and sulfonation steps into a one-pot procedure, significantly increasing the speed and ease of production. This had several advantages over conventional techniques, such as eliminating the need for metal-based polymerisation catalysts, dramatically reducing the amount of toxic solvents and sulfonation agents required, and hence lowering the cost of porous heterogeneous solid-acid catalysts. Comparison with equivalent materials produced via conventional routes showed that the properties obtained in SHCPs herein are superior with respect to both acidity and surface area. The SHCPs were applied to the acidcatalysed hydrolysis of cyclohexyl acetate to cyclohexanol. Conversions of up to 75% were measured, with all networks displaying excellent selectivity (>99%). Results strongly suggested that acidity is one of a number of important properties in the design of solid acid catalysts, with high acid densities even leading to reduced conversion rates. The ability to fine-tune polymer catalysts allows for the identification of key properties for further catalyst design and optimisation. |
67cae93b6dde43c908ea85ff | 0 | In the past two decades, there has been a growing interest from the materials science and nanotechnology communities in the possibility to modulate the electronic, mechanical, and chemical properties of 2D carbon-based materials by inducing changes in their atomic topology. Two main approaches have been adopted: doping and strain application. High-pressure conditions have been shown to strongly alter the topologies of multilayer graphene, leading in some cases to the formation of new 2D materials with unique electronic and mechanical properties. For example, under high pressure, few-layered graphene systems immersed in liq-uids/solids have been observed to undergo transitions into thin diamond-like structures and/or exhibit chemical doping. Recent studies based on Diamond Anvil Cell (DAC) experiments 2 have provided evidence for the formation of a new "diamodene" phase from bilayer graphene (BLG) in ice at 5-10 GPa. A chemical-induced scenario has been proposed, where the functionalization of BLG layers by water molecules (-OH and -H groups) at 5-10 GPa triggers the formation of C-C interlayer bonds. |
67cae93b6dde43c908ea85ff | 1 | The structural and electronic behavior of BLG under high pressure has also been investigated in the presence of other polar pressuretransmitting media (PTM), such as methanol and 4:1 methanol/ethanol mixtures. Nicole et al. and Forestier et al. have detected a strong piezo-doping effect on BLG layers at the interface with alcohol in the 1-10 GPa pressure range, by monitoring the evolution of the Raman G band, which is sensitive to carbon layer strain and doping. The Authors attribute the molecular origin of this phenomenon to electronic doping induced by surface adsorbates, resulting from the decrease in the methanol-BLG distance as pressure increases. |
67cae93b6dde43c908ea85ff | 2 | These findings contradict the observations of Filintoglou et al., who compressed monolayer graphene samples grown by chemical vapor deposition (CVD) on copper substrates using both a polar (4:1 methanol/ethanol) and a non-polar (fluorinert) pressure-transmitting medium (PTM). They observed that in the 0-3 GPa pressure range, the pressure slope of the G band was consistent for both PTMs, suggesting that the mechanical stress primarily determines the response of the G-band frequency to pressure, rather than pressure doping from the PTM. In this kind of experiments, numerous microscopic effects might contribute to the observed phenomena, making their atomistic interpretation non-trivial and leading to controversies in the literature. BLG is subjected to both in-plane biaxial stress induced by the substrate (generally SiO 2 or copper) and outof-plane stress from the PTM, that increases the probability of chemical functionalization and interlayer bond formation. In the present work, we focus on the latter effect, to elucidate the interplay between interface reactivity and PTM pressure. |
67cae93b6dde43c908ea85ff | 3 | Classic molecular dynamics (MD) simulations have proven to be a useful tool for gaining a deep microscopic understanding of the structural and dynamical properties of liquids at interfaces with carbon-based layered materials, in particular methanol confined between graphite surfaces. Mosaddeghi et al. revealed that at high-pressure and confinement conditions the methanol molecules organize in a hexagonal arrangement with CO bonds perpendicular to the graphite surfaces. In the absence of one of these conditions, the hexagonal arrangement is destabilized and the methanol molecules organize with their CO bonds oriented parallel to the graphite surfaces. |
67cae93b6dde43c908ea85ff | 4 | The reactivity of carbon layered materials/liquid interfaces has been instead investigated by ab initio MD simulations, able to describe the dissociation and formation of covalent bonds. For instance, Grosjean et al. have characterized the structural, dynamical, and chemical behavior of hydroxide ions (OH -) at the graphene/water interface, obtaining adsorption free energies, by coupling ab initio MD with enhanced sampling techniques. |
67cae93b6dde43c908ea85ff | 5 | In this study, we provide a first computational investigation, at the density functional theory (DFT) level, into the chemical transformations underlying the reactivity of BLG/methanol interfaces in the 0-10 GPa regime and room temperature. The system has been simulated in a suspended geometry using ab initio MD. The adoption of a suspended geometry allows us to focus on the effect of out-of-plane stress on the structural and chemical evolution of the BLG/methanol interface. |
67cae93b6dde43c908ea85ff | 6 | To establish a link between structure and reactivity we first characterize the interface structural properties under several pressure conditions. We then explore the thermodynamics and kinetics of BLG functionalization and of interlayer carbon-carbon formation by state-ofthe-art enhanced sampling techniques 23 that allow bridging the wide gap between experimental timescales and computationally accessible ones. |
67cae93b6dde43c908ea85ff | 7 | Considering the high anisotropy in the compression both of graphite as well as of bilayer graphene we have investigated BLG immersed in methanol exposed to increasing out-of-plane pressure (i.e., the P zz component of the stress tensor) from 0.0001 to 10 GPa, while keeping the in-plane pressure to the ambient value (0.0001 GPa). Experimentally, methanol is liquid below 3.5 GPa, while it forms an amorphous solid above 3.5 GPa. The first part of our study exploits DFT-MD simulations performed on models pre-equilibrated with classical MD. The main goals are i) to provide a microscopic characterization of the BLG/methanol interfaces as a function of pres-sure, shedding light on the link between interface structure and reactivity; ii) to observe BLG functionalization and/or formation of C-C covalent inter-layer bonds (sp 2 → sp 3 transformations) at high-pressure. |
67cae93b6dde43c908ea85ff | 8 | To characterize the BLG-methanol interfaces properly taking into account time-and space molecular fluctuations, we started by defining, at each MD time step, each graphene surface as a dense 2D grid formed by the atomic positions and by the midpoint between each pair of atoms belonging to the same hexagonal ring. With this instantaneous definition of the BLG layers, the density of methanol along the perpendicular direction is estimated considering the distance between the center of mass of each molecule and the closest graphene grid point, in this way accounting for graphene corrugation. |
67cae93b6dde43c908ea85ff | 9 | The interface structure as well as the methanol density as a function of the distance from the instantaneous BLG surfaces at ambient pressure (0.0001 GPa) are reported in Fig. : the first and second interface methanol layers are clearly identified. The first peak of the methanol density profiles (L1) in the range of 0.0001-10 GPa (Fig. ) reveals a decrease in the average distance between methanol centerof-mass and graphene from 3.35 to 3.09 Å when going from 0 to 2.9 GPa, afterwards remaining stable up to 10 GPa. The thickness of the L1 layer, z decreases with increasing pressure, from z=5.6 Å at 0.0001 GPa to z=5.4 Å at 1.2 GPa, and it remains stable at z=4.5 Å for pressures > 2.9 GPa. Note that density profiles are statistically converged and practically identical for the two interfaces. |
67cae93b6dde43c908ea85ff | 10 | For all the pressures explored the interface methanol molecules are undercoordinated with respect to the bulk ones. Our calculation at ambient pressure (0.0001 GPa) reports an average number HB s for bulk methanol molecules of 1.95 in good agreement with previous DFT-MD theoretical works and slightly over esti-Figure . a) Snapshot from DFT-MD simulations of the BLG/methanol system at 0.0001 GPa (with periodic boundary conditions). The oxygen, carbon and hydrogen atoms of the methanol molecules are depicted in red, green and white, respectively. The graphene bonds are coloured in dark grey. b) Time-averaged methanol density profiles normalized with respect to the bulk methanol density. The density is plotted as a function of the distance from the instantaneous BLG surfaces. c) First peaks (L1 and L2) of time-averaged methanol density profiles simulated under different pressure conditions (from 0.0001 to 10 GPa). Each profile is normalized with respect to the corresponding bulk density. d) Radial distribution functions for the carbon inter-layer distances. mated with respect to the experimental value of 1.77. Table . The first column displays the pressure conditions. The second column indicates the number of methanol molecules in the L1 layer. Columns three, four, and five report the total hydrogen bonds, intra-layer hydrogen bonds, and inter-layer hydrogen bonds for methanol molecule, respectively. The final column presents the atomic height distribution of BLG sheets. P(GPa) N mol HB HB intra HB inter S h (Å) 10 -4 (A) 10. 7 Independently on the system pressure, the interface methanol molecules (L1) are found being involved in a strong intralayer H-Bonding (see Table HB intra ) where each methanol molecule (L1) is in average involved in the same number of acceptor and donor intramolecular H-Bonds. The methanol-methanol HB s are found either parallel or perpendicular (pointing the OH toward the liquid phase) with respect to the instantaneous BLG surface. This can be easily appreciated in Fig. where the probability for methanol molecules in the L1 layers to form HBs with a given O-O distance and given orientation with respect to the normal to the BLG layers is reported for all the systems simulated in the 0.0001-10 GPa pressure range. The analysis has been performed on both the upper and lower BLG/methanol interfaces. Our data allows to clearly identify a preferential orientation of the HBs methanol molecules in the L1 layer characterized by their O-H pointed either parallel (cosines ≈ 0) or perpendicular to the methanol bulk phase (cosines <-0.4) Interestingly, the intermolecular O-O HB's distance is found to sensibly decrease only for pressure beyond 2.9 GPa, where the Graphene-methanol distance reaches a plateau at increasing pressure (see L1 peaks positions in Fig. ). For instance, the O-O HB's distances range in the 2.5-2.9 Å interval for pressure<2.9 GPa, while they are found to fluctuate in the 2.5-2.8 Å interval for pressures >2.9 GPa. We also observe a narrowing of the OH-O orientational distribution at high pressure, that indicates a reduction of the OH-O orientational freedom. This is due to the decrease in the methanol-Graphene and methanol L1-L2 distances (see first and second peaks positions in Fig. ) from 0.0001 to 4.9 GPa that limits the OH-O dynamics. The statistical convergence of the distance and orientation probabilities reported in Fig. was assessed by following their evolution every 5 ps of dynamics and verifying their invariance with time. The differences between the upper and lower BLG/methanol interfaces are possibly due to the exploration of different minima, underlying the BLG/methanol interface organization, each one divided by free energy barriers that are difficultly crossed in the picoseconds time scale. Another factor to be considered is the slow dynamics of the C-O libration of interfacial (L1) methanol molecules. |
67cae93b6dde43c908ea85ff | 11 | Our study then continues by analyzing the electronic and structural properties of the BLG sheets. To investigate the effect of pressure on the electronic properties of bilayer Graphene (BLG), we employed Bader charge population analysis. The analysis revealed that the average charges on the BLG layers are invariant as the pressure is increased (see Table in the SI). Despite the decreasing distance between methanol and BLG layers in the 0-2.9 GPa pressure range (as illustrated in the density profile in Fig. ) we find no charge transfer between the BLG layer and the adsorbate. In addition, we find that there is no change in the intra and inter layers polarization of the BLG with increasing pressure. This is supported by the same value of average charge for the bottom and upper layer, as well as no variations in the standard deviation computed at different pressures. Such findings indicate that the experimentally observed 1 increase in the pressureinduced BLG surface charge concentration in presence of methanol cannot arise from an enhancement of the methanol adsorption in the 0-10 GPa pressure range. We have then evaluated the evolution of the carbon-carbon inter-layer distance as a function of pressure by computing the associated radial distribution functions (see first peak's positions in Fig. ). We observe a progressive decrease of the carbon-carbon interlayer distance from 3.35 Å at ambient pressure to 3.09 Å at 10 GPa. Such results could indicate a higher probability of sp 3 structures' formation in the high pressure regime (> 6.9 GPa). However, no formation of stable sp 3 diamond-like structures is observed in the 20 ps of unbiased MD simulations in the 0.0001-10 GPa pressure range. We have then quantified the corrugation of the BLG layers as a function of pressure making use of the standard deviation of the BLG sheets atomic heights distribution (S h ) sampled over all atoms and frames. In Table we report for each pressure the upper and lower BLG layer's S h . The S h average values slowly decreases from 0.0001 to 1.2 GPa before reaching a plateau between 1.2 and 10 GPa The slight decrease in the BLG sheets corrugation is coherent with the reduction of the BLG inter-layer and BLG-methanol distances in the 0.0001-1.2 GPa range, that limits the out-of-plane displacements of the BLG carbon atoms. However, the small entity of the effect and the constant S h behavior in the 2.9-10 GPa range make us rule out a possible effect of corrugation on the formation of sp 3 structures in the high pressure regime. |
67cae93b6dde43c908ea85ff | 12 | To summarize, our analysis reveals a decrease of the distance between the interfacial methanol molecules and the BLG layers and an enhancement in the molecular crowding at the interface when compressing between ambient pressure and 10 GPa. These factors would be in favor of a possible promotion of the BLG layers functionalization in response to an increase in pressure. By contrast, the strong intralayer H-Bond character of the interface methanol, with the OH group oriented parallel and outward with respect to the BLG layers could disfavour the BLG functionalization by reducing the outof-plane character of the O-H chemical bonds at the interface, necessary for the formation of sp 3 structures. |
67cae93b6dde43c908ea85ff | 13 | We have therefore deepened our study on the effect of pressure on the Graphene functionalization, by assessing the thermodynamic stability of BLG functionalized with -OCH 3 and -H groups under different pressures. In this regard we have built models of functionalized BLG/methanol interfaces with -OCH 3 and -H groups bonded to the same BLG layer. Such functionalization is the product of a methanol deprotonation followed by formation of C-O and C-H bonds with the BLG layer. Models characterized by -OCH 3 and -H groups in the ortho, meta and para configurations have been simulated by DFT-MD at 0.0001, 6.8 and 10 GPa. For the first 2-3 ps of equilibration the C-O and C-H interatomic distances between the functional groups (-H and -OCH 3 ) and the BLG layer were restrained by applying semi-parabolic potentials V wall (with constant k=15000 kJ/mol*nm). This was done in order to equilibrate the functional groups first solvation shell and avoid unwanted spurious effect (e.g. bond cleavage) due to the possible high energy of the system's initial configuration. |
67cae93b6dde43c908ea85ff | 14 | The systems were then simulated unbiased (in absence of the V wall ). At ambient pressure (0.0001 GPa), the C-H and C-OCH 3 bonds are cleaved after few tens of unbiased-MD steps and the functional groups recombine at the interface forming CH 3 OH. This was observed for all the configurations (ortho, meta and para). The BLG functionalization by -OCH 3 and -H groups do not represent a stable configuration at ambient pressure, i.e. no free-energy minimum is associated to this configuration. |
67cae93b6dde43c908ea85ff | 15 | By contrast, in the high pressure regime (6.8 and 10 GPa), the functional groups stay stably bonded to the surface for all the time of the unbiased MD simulation (5 ps). This occurs in all the configurations explored except for the meta systems simulated at 6.8 GPa, where the cleavage of the C-H and C-OCH 3 bonds and consequent formation of CH 3 OH at the interface is observed after 2 ps of two indipendent unbiased simulations. This indicates that high-pressure conditions thermodynamically favor the functionalization of BLG, compared to ambient ones. The molecular origin of such stabilization likely lies in the increase of the hydrophobic character of the methanol/BLG interface when increasing pressure. The functional OCH 3 group is located in the L1 layer and its first solvation shell is composed by both L1 (76.5 %) and L2 (23.5 %) methanol molecules (see Fig. in the S.I.) pointing their CH 3 moieties toward the OCH 3 functional group. In particular, we observe a decrease from 3.8 to 3.4 Å and from 7.4 to 6 Å in the case of the Methanol Carbon-Graphene distances for respectively the L1 and L2 methanol molecules (see Carbon density profiles and snapshot in Fig. -b in the S.I.). In addition, the interface molecular crowding in the high pressure regime provides a higher number of CH 3 moieties at the interface (14.5 methanol molecules at 6.8 GPa, vs 11.4 at ambient pressure) as can be appreciated in Table This results in a stronger and complete hydrophobic solvation of the OCH 3 groups that leads to a higher thermodynamic stabilization. However the functionalization of BLG with H and OCH 3 groups comes with a considerable enthalpic cost, with values of 236 and 288 kJ/mol, for 6.8 and 10 GPa respectively. Interestingly, these values remained consistent across all configurations tested (ortho, meta, para). In the next section we will characterize the kinetics of the BLG functionalization at high pressure by evaluating the free energy barriers with coordination based metadynamics. |
67cae93b6dde43c908ea85ff | 16 | The effect of the methanol pressure on the formation of carbon-carbon interlayer bonds in absence of BLG functionalization has then been considered by assessing the thermodynamic stability of diamond-like structures when exposed to high pressure (6.8 and 10 GPa) in methanol. We have performed biased MD simulations where we have promoted the formation of interlayer C-C bonds by restraining for 2.5 ps the interlayer distance (between centers of mass) to remain below 1.7 (applying semi-parabolic potentials V wall with k= 10 8 kJ/mol*nm). After few hundreds of steps we observe the formation of diamane-like structure, a single layer of diamond characterized by C 6 units in boat configuration. We find half of the BLG carbon atoms involved in C-C interlayer bonds. Once the bias is removed, the C-C interlayer bonds are broken within few femtoseconds, with consequent increase of the interlayer distance and reformation of BLG. |
67cae93b6dde43c908ea85ff | 17 | Such lack of local stability is consistent with the absence of experimental evidences of diamond-like structure formation from multilayer Graphene when exposed to pressure in the range of 0.0001-10 GPa in anhydrous solvents. However the presence of chemical groups, such as methoxy, hydrogens or hydroxy groups, may substantially decrease the pressure required to promote covalent bonds between carbon atoms in distinct layers of BLG. In the next section we will investigate the possible interplay between the BLG functionalization and the formation of diamond-like structures by combining DFT-MD simulations with metadynamics. In particular we will explore the possible mechanisms of reactions and the associated free-energy barriers characterizing the BLG/methanol functionalization and BLG's inter-layer reactivity. |
67cae93b6dde43c908ea85ff | 18 | In this section we describe an exploration of the possible transformation pathways and asso-ciated kinetic barriers underlying the BLG reactivity in methanol, at ambient pressure, as well as 2.9 and 6.8 GPa, at room temperature. We exploited metadynamics as an acceleration technique to overcome the timescale limitations of ab initio MD. |
67cae93b6dde43c908ea85ff | 19 | We started evaluating the possible highpressure pathways leading to the formation of a para-H, -OCH 3 cis-functionalized BLG, that we have shown in the previous section to be thermodynamically stabilized at 6.8 GPa. Fig. shows the mechanism and the associated metadynamics-bias landscape at 6.8 GPa, obtained employing the coordination numbers of two Graphene carbon atoms in para configuration as reaction coordinates: one with respect to the interface hydrogen atoms (C 1 -H), and the other with respect to all interface oxygen atoms (C 2 -O). |
67cae93b6dde43c908ea85ff | 20 | The reaction proceeds trough the deprotonation of an interface methanol molecule, followed by the concerted formation of one -H and one -OCH 3 bond with Graphene in a para configuration (Panel a). Inspection of the metadynamics bias (Panel b) suggests a huge free energy barrier, possibly of ca. 1060 kJ/mol, based on the bias accumulated in the initial minimum (reactant basin) until the onset of the transformation. Note that this provides only a rough (over)estimation of the free-energy barrier: a more precise estimation of the barrier, as well as of the corresponding kinetic rate, would require i) the use of an optimal reaction coordinate, and ii) the analysis of a bias displaying steady evolution, achievable only by simulating a trajectory long-enough to display many forward and backward reactions. Both conditions are, unfortunately, impossible to meet in this specific system due to the high cost of ab initio MD on our relatively large model. |
67cae93b6dde43c908ea85ff | 21 | Given the high kinetic barrier of the mechanism explored in the last section employing ad-hoc collective variables, we further explored other possible reaction mechanisms of potentially lower barrier. To this aim, we adopted a general-purpose set of collective variables, called SPRINT, designed to discover without prejudice the possible ways a system can change the bonding network (topology). We tailored the coordinates to track changes in the short-and long-range topology of i) the first interface methanol layer (L1), identified for each system by the first peak of the corresponding density profile (see Fig. ), and ii) two large areas (16 carbon atoms each) of both the upper and lower Graphene layers (see Fig. in the S.I.). The two regions were chosen approximately on the top of each other. |
67cae93b6dde43c908ea85ff | 22 | The reaction then proceeds with the migration of the -OCH 3 group on another carbon atom of the SV vacancy by means of a pivoting motion around a C-C bond (II-III). The SV structure relaxes, undergoing a distortion which decreases the C-C distance (from 2.5 to 2.0 Å) between the two unsatured carbon atoms (labelled C u1/2 in Fig. ). This process would correspond, considering the C u1 -C u2 bond as formed, to a SV characterized by one fivemembered and one nine-membered ring. Remarkably, the same SV structure has been experimentally observed by TEM on suspendend monolayer Graphene at room pressure and temperature conditions. The system subsequently evolves (III-IV) with the hydrogenation by a methanol molecule of the C u1 atom and consequent formation of a C-H bond pointing towards the methanol.In the next step (IV-V) the shared proton is transferred to the C u2 atom with consequent formation of a -C-H bond pointing toward the bottom Graphene layer. |
67cae93b6dde43c908ea85ff | 23 | The system then heals the SV via the return of the Graphene atom displaced in step I-II to the original position, bonded to C u1 and C u2 and carrying the functional metoxy group with it. This leads to the formation of trans func-Figure . Mechanism a for the trans-functionalization at the BLG-methanol interface at 6.8 GPa observed in SPRINT metadynamics simulations. On the top, snapshots associated with the main steps of the reaction mechanism (the color code is the same as in Fig. ). |
67cae93b6dde43c908ea85ff | 24 | The reaction continues with the cleavage of the C-OCH SPRINT metadynamics reveals also the existence of a distinct pathway (mechanism b), displaying analogies with the previous one, presented in Fig. . In this case the reaction proceeds through the following main steps: the initial formation of a SV (I-II) with a dangling carbon atom pointing towards the methanol; the double-functionalization of the dangling carbon by -H and OCH 3 groups originated by different methanol molecules (II-III); the healing of the SV with release of a CH 3 O -(IV-V); H migration between the Graphene layers (V-VI) until binding to the lower layer (VI). |
67cae93b6dde43c908ea85ff | 25 | Taken all together, our simulations identify the initial formation of the SV as the bottleneck step of the a-and b-pathways, both at 2.9 and 6.8 GPa, requiring a bias of 870-1160 kJ/mol. Such bias values, even though they suggest that these transformation pathways are unlikely to be observed at room temperature, should be considered as (possibly severe) overestimates of the equilibrium free-energy barriers for the methodological reasons explained above. |
67cae93b6dde43c908ea85ff | 26 | In the attempt to obtain a more precise estimate of the free-energy barrier of the key reaction-step, i.e., the SV formation, we adopted coordination-based path collective variables metadynamics, since this approach allows sampling reaction paths between Figure . Mechanism b for the trans-functionalization at the BLG-methanol interface at 6.8 GPa observed in SPRINT metadynamics simulations. On the top, MD snapshots associated to the explored reaction mechanisms. The color code is the same as in Fig. . a specific pair of reactants and products (see the supporting information for details). SV formation in the b-mechanism (transition I-III in Fig. ) has been considered: structures I and III were employed to define the reference coordination patterns. |
67cae93b6dde43c908ea85ff | 27 | Fig. shows the reaction mechanism and associated bias landscape. According to these simulations the first step of SV formation is the hydrogenation of a Graphene atom from the deprotonation of a methanol molecule (I-II). The reaction then proceeds via the deprotonation of a second methanol molecule, followed by the formation of an unstable intermediate (III * ) featuring a SV and an out-of-plane carbon atom triple-functionalized by two -H and one -OCH 3 groups. The process terminates with the dehydrogenation of the triple-functionalized carbon atom by a methanoate molecule, leading to the product (structure III). |
67cae93b6dde43c908ea85ff | 28 | Graphene presents different types of defects as impurities, vacancies and grain boundaries which concentration may depend on growing methods. In addition, experiments report a strong increase of defect concentration on BLG surfaces when increasing the pressure after PTM solidification, 10 as well as the formation of strong strain fields leading to distribution of local folds before PTM solidification. Motivated by such findings and the results of the previous section, we have investigated the effect of singleand double vacancies on BLG functionalization at 6.8 GPa by SPRINT metadynamics. |
67cae93b6dde43c908ea85ff | 29 | SV and DV models were created by removing one or two adjacent atoms from a Graphene layer, followed by an equilibration in the presence of methanol by DFT-MD for 5 ps. The SV-model is characterized by a five-membered and a nine-membered ring, with one atom under-coordinated (indicated with C u1 ), as experimentally observed by TEM. The DV model features two five-membered and one eight-membered rings, without any undercoordinated carbon atoms. |
67cae93b6dde43c908ea85ff | 30 | All Graphene atoms first-and second-nearest neighbors to the SV and DV and all the molecules of the first methanol layer were included in the SPRINT definition, hence subject to the metadynamics bias. Three indipendent metadynamics simulations reveal functionalization of the undercoordinated C u1 atom by a methoxy group (see Fig. ). The analysis of the metadynamics bias suggests a free-energy bar- In the case of the DV, the same methoxy group functionalization of BLG is observed, but associated to a much higher barrier, the bias being about 250 kJ/mol. The strong barrier decrease in presence of a SV, compared to the pristine Graphene as well as to the DV cases, could be tentatively attributed to the presence of the under-coordinated C u1 atom. To elucidate the effect of high pressure on this transition, we simulated the same reaction under ambient pressure using metadynamics with the SPRINT collective variable. Four inde-pendent metadynamics simulations (see Fig. in the Supporting Information) reveal that, under ambient conditions, SV-BLG methanol functionalization proceeds through methanol reorientation and approach toward the BLG, followed by methanol dissociation and singlevacancy functionalization with both hydrogen and methoxy groups. The reorientation of the methanol pointing its oxygen toward the BLG SV, does not occur under high-pressure conditions, since the methanol is already in proximity of the graphene surface (see density profiles in Fig. ). |
67cae93b6dde43c908ea85ff | 31 | The analysis of the metadynamics bias in Fig. reveal free-energy barriers of approximately 150-200 kJ/mol, over twice the barrier observed under high-pressure conditions for BLGmethanol SV functionalization. The molecular origin of this significant barrier reduction in high pressure is likely due to the greater proximity of the first methanol layer to the BLG, which kinetically enhance the functionalization process (see density profiles in Fig. ). Our findings open up an interesting perspective on SV-BLG functionalization as a plausible mechanism for doping under high-pressure conditions. This mechanism could explains the observed pressure-induced doping of BLG in methanol, which begins in the early stages of compression and saturates around 2-3 GPa. This saturation is consistent with the leveling of the methanol-BLG distance, which stops decreasing beyond 2.9 GPa, as shown in our density profiles (Fig. ). |
67cae93b6dde43c908ea85ff | 32 | In Fig. , the evolution of ELF attractors for the SV and the reactive methanol atoms going from structure I to IV, from reactants to products are reported. The attractors are local maxima of the ELF and allow an identifica-tion of localized bond electrons and lone pairs. The population of each attractor, highlighted in bold in Fig. , is obtained by integrating the electron density over the associated basin, defined as the region of space from which the gradient path leads to the attractor. At the reactant state (structure I), our analysis identifies three attractors (0.8 ELF value) located in the SV: attr2 and attr3 lying in the Graphene plane close to C u2 and C u3 atoms, and attr1 located slightly out-of-plane in the proximity of C u1 . These attractors correspond to one electron pair shared between the three undercoordinated carbon atoms. The electronic pair has been identified from the corresponding maximally localized Wannier function and associated center (in green), reported in Panel B of Fig. ). It should be mentioned that, in apparent contradiction with the Wannier function analysis, the total ELF basins population is close to 3. This is due to the inclusion of a π-delocalized electron pair in the ELF integration (see Supplementary Information, Fig. ). |
67cae93b6dde43c908ea85ff | 33 | The reaction first proceeds through the nucleophilic attack of the methanol molecule to the C u1 atom (structure II). At a C•••O distance of 1.74 Å, the inter-molecular interaction is evidenced by a deformation of the ELF around the oxygen atom, away from the typical lone pairs shape (see Supplementary Information, Fig. , panel a). The approach of the oxygen doublet induces a transfer of electronic density from the C u1 atom to the C u2 and C u3 ones. This can be appreciated from a decrease of population of attr1 from 0.982 to 0.646 and a simultaneous increasing of attr2 from 0.72 to 0.94. |
67cae93b6dde43c908ea85ff | 34 | The reaction terminates with the formation of a C-C bond in the SV between C u2 and C u3 and the simultaneous release of a proton from the methoxy group to a neighboring methanol molecule (structure IV). This is confirmed by the recovery of the lone-pair shape of the ELF around the oxygen atom of the methoxy group. |
67cae93b6dde43c908ea85ff | 35 | To further characterize the C-C bond formation in the SV, we have reported the ELF profile along the direction connecting C u2 and C u3 for each reactive structure (see insets in Fig. ). This is a good estimation of the evolution of the basin separatrix of the two attractors along the reaction. The basin separatrix is a local minimum in the direction linking two attractors basins that gives useful insights into the relation between them. The high ELF value of 0.58 of the basins separatrix between attr2 and attr3 (structure I), compared to those between attr1 and attr2/attr3 (around 0.30), indicates that those two basins are specially related, emphasizing the specific under-coordinated character of C u1 at the reactant state. |
67cae93b6dde43c908ea85ff | 36 | As the C-O bond formation proceeds (structure III), their basins separatrix value increases to 0.74, close to the 0.80 value of the two attractors. No covalent bond is formed yet as the two attractors are still separated. At the product state (structure IV) where the C•••C distance is 1.85 Å, attr2 and attr3 merge into a single attractor between C u2 and C u3 , with the value of 0.88, accounting for about 2 electrons. From the ELF perspective, it is clear that C-C bond formed. However, this new attractor is slightly displaced out of the bond line and the ELF shape does not correspond to any single, double or triple typical bond shape, which is noticeably consistent with the particularly stressed geometry this bond has to fit in (see Supplementary Information, Fig. ). |
67cae93b6dde43c908ea85ff | 37 | Our study provides a detailed microscopic insight into the structure and reactivity of the bilayer graphene (BLG)/methanol interface at room temperature under varying pressure conditions. A study of the interface in terms of structural descriptors has allowed to characterize the structural evolution of interface, showing a decreasing trend in the methanol-graphene distance up to 2.9 GPa, beyond which the distance stabilizes and remains constant up to 10 GPa. Additionally, we observed a gradual reduction in the carbon-carbon interlayer distance, accompanied by increased molecular crowding for increasing pressure. Notably, a strong intralayer H-bond character of the interface methanol, with the OH group oriented parallel and outward with respect to the BLG layers at high pressure, has also been identified. |
67cae93b6dde43c908ea85ff | 38 | In a second step we investigated the thermodynamics and kinetics of the chemical transformations involving the BLG/methanol interfaces under high-pressure conditions. We find pressure to stabilize functional groups on the BLG surface. We trace back such effect to an enhancement in interfacial hydrophobicity (i.e., an increase in the local concentration of hydrophobic groups) at increasing pressure. In agreement with experiments, 1,11 our simulations confirm the absence of local stability for C-C interlayer structures in the 0-10 GPa range. |
67cae93b6dde43c908ea85ff | 39 | We have then revealed the reaction pathways leading to the BLG functionalization in methanol using metadynamics. The presence of a single vacancy on the BLG surface is found to reduce by more than 20 times the free energy barrier for -OCH3 functionalization, making the process feasible at 6.8 GPa and ambient temperature conditions. We attribute the microscopic origin of such effect to the presence of an undercoordinated carbon atom at the vacancy, that facilitates the nucleophilic attack of the metoxy group (CH 3 O -). |
67cae93b6dde43c908ea85ff | 40 | Interestingly, we found that pressure plays a key catalytic role in this reaction: by bringing the first methanol layer closer to the BLG surface, it significantly enhances the reaction process compared to ambient conditions. Our results lead to conclude that graphene defects and high-pressure conditions 10 enhance the probability of functionalization. |
67cae93b6dde43c908ea85ff | 41 | A model of bilayer Graphene (BLG) including 144 atoms has been inserted in a hexagonal simulation box with cell parameters a = b = 14.81 Å and c = 30.00 Å. The AA stacking of Graphene layers has been adopted as initial configuration since it has been proposed as a reactive stacking for the formation of bilayer carbon sp 3 structures. The simulation box was then filled with 56 methanol molecules (d CH 3 OH ≃ 0.792 g•cm -3 ) using the PACKMOL software package. Five simulation boxes have been then generated by optimizing the positions and c cell parameter at T = 0 K by imposing five different values of the stress tensor P zz component going from 0.0001 to 10 GPa, while keeping a, b fixed to ambient pressure values. In all subsequent simulations, the cell parameters are not changed anymore. The distance between BLG periodic images ranges from 24.5 Å at 0.0001 GPa to 14.65 Å at 10 GPa. The pressure, the optimized c cell parameter, the theoretical and experimental methanol densities for each BLG/methanol system are reported in Table . |
67cae93b6dde43c908ea85ff | 42 | For each pressure, the system has been pre-equilibrated with 1.5 ns classical MD in the NVT ensemble at 300 K (while keeping fixed the BLG atoms), using the OPLS-AA force field. The latter has been extensively shown to accurately describe the structural organization of bulk methanol. Table . The first column displays the pressure conditions. Columns two, three and four report the optimized c cell parameter and the theoretical and experimental methanol densities In the next step, we performed Born-Oppenheimer MD based on the DFT-PBE approximation 60 together with Grimme's D3-BJ correction for dispersion interactions. This DFT level has been shown to reproduce within a 4.3% error the lattice constants, bulk moduli, elastic constants and interlayer binding energy of both graphite and diamond in the 0-25 GPa range. We employed Goedecker-Teter-Hutter pseudopotentials and combined plane-waves (850 Ry cut-off) and DZVP-MOLOPT SR (short range) basis sets, as implemented in CP2K. The electronic wavefunctions have been converged at each step imposing a threshold for the energy difference between two SCF cycles of 10 -6 Ha. All ab initio MD simulations are carried out in the spin-restricted Kohn-Sham scheme, with a timestep of 0.5 fs and the Nosé-Hoover thermostat 66 using a time constant of 100 fs, a chain length set to 3, the 3rd order Yoshida integrator and a multiple time step set to 2. Each BLG/methanol system was simulated at 300 K for 20 ps: the first 5 ps are considered as equilibration, while structural analysis (density profile and radial distribution functions) was performed on the last 15 ps of each trajectory. Fig. in Supporting Information reports the distribution of stacking configurations explored during MD between 0.0001-10 GPa. |
67cae93b6dde43c908ea85ff | 43 | The BLG layers are found to fluctuate in the intermediate region between the AB and AA stacking configurations, with a preference for the AB stacking at high pressure (4.9-10 GPa). Note that sp 2 -to-sp 3 phase change in BLG has recently been reported regardless of the stacking configuration. Finally, we coupled ab initio MD simulations, via the Plumed plugin, with the metadynamics algorithm, that accelerates the escape from local free-energy minima and allows reconstructing free-energy landscapes by adding to the potential energy of the system a historydependent bias profile built as a sum of repulsive Gaussian functions, acting in a lowdimensional space of collective variables. |
67cae93b6dde43c908ea85ff | 44 | First, we performed a fast, prejudice-free exploration of the possible reactive channels at the BLG/methanol interface by applying the metadynamics bias in the space of SPRINT (social permutation invariant) collective variables. The latter, obtained by diagonalizing a smooth interatomic adjacency matrix, capture the changes in atomic-bond topology and proved effective in discovering structural transformations in many systems, including covalent carbon networks. This approach allowed us to identify several chemical reaction pathways, that were further investigated, including estimation of the associated free energy barriers and metadynamics with smooth coordination numbers as well as with path collective variables based on patterns of coordination numbers. The details about the metadynamics protocols can be found in Supporting Information. The input files of CP2K and Plumed for the simulations in this work, as well as selected atomic configurations (metastable states at 300 K and different pressures), are freely available from Plumed Nest. The Electron Localization Function (ELF) for each reactive structure was calculated using CP2K, following the same setup as the one used for the dynamics simulations. The automatic identification of ELF critical points (AUTO) was carried out using the Wigner-Seitz cell (WS) and pairs |
67cae93b6dde43c908ea85ff | 45 | where r ij (t) is an interatomic distance in a particular moment t, whereas d αβ is a typical bond distance between two atomic species. Note, that d αβ does not particularly correspond to the experimentally measured bond distance and should be viewed as a fine-tuned computational parameter. The d αβ for each kind of interatomic distances are reported in Table A critical aspect of defining the path collective coordinates s and z is selecting the set of atoms for which C iβ (t) will accurately describe the reaction progress. If the set is too small, it may not capture the entire reaction mechanism, while an overly large set may introduce excessive noise and lead to inefficient sampling of the reaction space. |
67cae93b6dde43c908ea85ff | 46 | Only two reference states were chosen to describe the reaction pathways of the SV formation at the BLG/methanol interface. Models I and III (Fig. of the Main text) were employed respectively for k equal 1 and k equal 2 at the interface. Hence, the studied reaction mechanisms are not biased with a hypothesis on the path (no intermediates assumed). |
67cae93b6dde43c908ea85ff | 47 | For both k states and reaction media, unbiased AIMD simulations were performed using the CP2K/Quickstep package 3 in combination with the PLUMED v.2.3 plug-in. of C iβ (t) defined as time averages of 10 ps long trajectories collected with the time step of 0.5 fs. Trajectories as long as that were sufficient to equilibrate the system at 300 K. We chose λ = 0.404 which satisfies λD(x i , x i+1 ) ≈ 2.3 and maps the k = 1 and k = 2 references 70 to s ≈ 1.1 and s ≈ 1.9, respectively. |
67c2447d81d2151a02832368 | 0 | Polymers are promising non-viral delivery vehicles for gene editing and therapy applications, offering a safer and more versatile alternative to viral vectors, which face challenges like immunogenicity, limited capacity, and high production costs 1,2 . Polymers bind and compact DNA or RNA cargo into nanoparticle complexes, called polyplexes, for delivery into cells . Among the most popular polymer systems is polyethyleneimine (PEI), sold commercially as jetPEI ® , which is widely used as a basic scaffold for delivery systems employed in routine transfection and gene editing applications . |
67c2447d81d2151a02832368 | 1 | Despite their potential, PEI-based systems still face significant limitations in achieving the high delivery efficiencies required for clinical use . A critical factor influencing efficiency is the formation and stability of polyplexes . Strong binding and efficient nucleic acid release must be carefully balanced because overly stable polyplexes can hinder intracellular release, while weak complexes fail to protect and transport cargo effectively into cells . Previous studies have predominantly focused on optimizing transfection efficiency only by adjusting molecular weight , nitrogen-to-phosphate (N/P) ratios , or complex formation conditions . |
67c2447d81d2151a02832368 | 2 | Other investigations have examined PEI-DNA complexation mechanisms by examining factors such as binding strength, stability, and structure. For example, Ketola et al. used fluorescence spectroscopy to reveal distinct differences in DNA cargo binding for linear and branched PEI (LPEI and BPEI, respectively). Other studies suggest that BPEI offers better DNA protection than LPEI, but lower delivery efficiency . However, these studies do not provide a molecular-level picture of the binding modes or conformational dynamics of LPEI and BPEI polyplexes. Developing this molecular-level understanding is crucial for being able to tune the interactions of polymeric carriers so that they can more robustly accommodate different types of cargo, form more stable polyplexes, offer better cargo protection, or re-lease their cargo more efficiently. |
67c2447d81d2151a02832368 | 3 | Building on this need for deeper molecular insights, recent work by Reineke and coworkers introduced versatile quinine-based polymers capable of delivering multiple cargos, including plasmid DNA, mRNA, and CRISPR/Cas9 machinery. Their studies suggest that the ability of these delivery systems to robustly accommodate and deliver these different types of cargoes is facilitated by their ability to bind them through multiple mechanisms, including electrostatic, π-stacking, and hydrogen bonding interactions . This insight opens new avenues for investigating similar phenomena in PEI-based systems, whose interactions with nucleic acid cargo have historically been assumed to be merely electrostatic in nature. Interestingly, a few computational studies have predicted that branched PEI (BPEI) might engage in additional binding modes beyond electrostatics . However, there has not been any experimental evidence to confirm this. |
67c2447d81d2151a02832368 | 4 | To address this gap, we investigated the binding modes and macromolecular structure of PEI-DNA polyplexes using infrared (IR) spectroscopy and transmission electron microscopy (TEM). IR spectroscopy probes vibrations that report on bond-specific interactions and conformational changes, while TEM provides insights into polyplex morphology. Together, these methods offer a comprehensive molecular-level perspective of PEI-DNA dynamics. |
67c2447d81d2151a02832368 | 5 | Using TEM, we first examine the morphologies of LPEI and BPEI polyplexes, followed by UV-Vis absorbance spectroscopy to analyze the complexation kinetics of these polymers to DNA. We then apply Fourier Transform IR (FTIR) spectroscopy with Multivariate Curve Resolution (MCR) analysis to differentiate between the binding modes of LPEI and BPEI polyplexes. Our comprehensive approach elucidates new insights into the molecular-level binding mechanisms of PEI systems to DNA, which we anticipate will help guide the development of more efficient and versatile PEIbased gene delivery systems. |
67c2447d81d2151a02832368 | 6 | For kinetics experiments involving UV-Vis absorption spectroscopy, we prepared stock solutions of polymer and DNA in PBS (pH 7.4) with final concentrations of 0.5 µM for DNA, 200 µM for linear PEI, and 500 µM for branched PEI. We then added approximately 1 µL aliquots of polymer stock solutions with different volumes to DNA stocks to create solutions with N/P ratios (the molar ratios of polymer amines to DNA phosphate) of 1 and 5 for use in the kinetics experiments. We maintained the pH of all samples at 7.4. Since the polymer volumes were very small, the final concentration of DNA was maintained around 0.5 µM (see the Supplementary Information for details). |
67c2447d81d2151a02832368 | 7 | We used a similar procedure and final concentrations of polymer and DNA to prepare polyplex solutions for binding affinity experiments involving UV-Vis absorption spectroscopy. By varying the volumes of polymer stock solutions added to DNA stock, we created polyplex formulations with N/P ratio ranging from 0.1-10. Samples became slightly turbid upon adding polymer stock solutions to DNA stocks, indicating polyplex formation. Additionally, we prepared solutions of only polymer, without DNA, with similar final polymer concentrations as the polyplex samples. All samples were then incubated at room temperature for 24 hours to equilibrate prior to performing binding affinity experiments. |
67c2447d81d2151a02832368 | 8 | For FTIR experiments, we prepared stock solutions of polymer and DNA in phosphate buffer (pH 7.4) with and without NaCl. For NaCl screening experiment, we used concentrations of 50, 100, 150, 200, and 250 mM. The final concentration for the DNA stock solutions were 2 mM, while the final concentration for the polymer stock solutions were 5 mM. We diluted the polymer stocks further to create solutions with different final concentrations. We then added these solutions dropwise to DNA to create polyplex samples with N/P ratios ranging between 0.25 -3. The final DNA concentration after adding polymer solutions was 1 mM. We maintained the pH of all samples at 7.4. We also prepared DNA samples without polymer at 1 mM concentration with and without NaCl to use as a reference. In addition, we prepared LPEI and BPEI poly-mer solutions as controls with a final concentration of 1.8 mM and 2.6 mM, respectively. Samples were allowed to equilibrate for one hour prior to measurements. |
67c2447d81d2151a02832368 | 9 | We obtained UV absorbance spectra using an Agilent Cary UV-Vis-NIR Spectrophotometer equipped with a tungsten halogen and deuterium arc lamp. Spectra were collected between 200 -350 nm using a scan rate of 600 nm min -1 . For kinetics experiments, spectra were collected every hour over the course of a day. |
67c2447d81d2151a02832368 | 10 | We collected IR absorption spectra using a Bruker INVE-NIO Fourier transform IR spectrometer with a concentratIR2 multiple reflection silicon ATR head purchased from Harrick Scientific. This ATR unit is designed for micro-liquid samples and has eleven internal reflections with a nominal incident angle of 30 • . We performed background measurements first and then we added 60 µL of sample on the ATR sampling area. We measured all spectra in the 1100 -1800 cm -1 range with 4 cm -1 resolution and 128 scans in transmission mode. We then converted the spectra to absorption mode using the Bruker software. |
67c2447d81d2151a02832368 | 11 | A 5 µL aliquot of sample was placed on freshly glow discharged carbon-coated grid and incubated for 2 min. Following incubation, the grid was then washed 5× with water and wicked dry with filter paper. The samples were then stained using 3 % (w/v) uranyl acetate for 2 min. Excess stain was then removed from the sample grid by wicking with filter paper. The grids were air dried in a dust free environment. TEM imaging was then performed using a JEOL 1400 electron microscope operating at 120 kV, equipped with a AMT XR611 CCD camera. |
67c2447d81d2151a02832368 | 12 | We processed all FTIR and UV-Vis spectra using custom MATLAB scripts written in-house (see SI). We performed baseline correction and background subtraction for all spectra. To compare the spectra, we additionally normalized FTIR spectra to the total integrated intensity. We used Prism Graph-Pad software to fit the kinetics and binding affinity data collected with UV-Vis spectroscopy. Multivariate curve resolution (MCR) analysis was performed using the MCR-ALS software developed by Felten et al. |
67c2447d81d2151a02832368 | 13 | To better understand how the chemical architecture of PEI influences complexation behavior to DNA cargo, we studied both LPEI and BPEI (Figure ). LPEI is a linear polymer that contains only secondary amines, while BPEI has a branched structure containing primary, secondary, and tertiary amines. These structural differences can potentially influence LPEI and BPEI complexation behavior and, consequently, their delivery mechanisms. |
67c2447d81d2151a02832368 | 14 | To investigate how these chemical structural differences affect polyplex assembly, we first assessed the morphologies of LPEI and BPEI polyplexes. Figure shows the TEM images for LPEI, BPEI, and DNA controls, as well as LPEI and BPEI polyplexes at an N/P ratio of 2. The images indicate that LPEI, BPEI, and DNA exhibit fibrous morphologies (Figure ). However, adding the PEI systems to DNA results in nanoparticles with spherical morphologies (Figure and), indicating the formation of polyplexes . Although both LPEI and BPEI form similar spherical structures, some distinct differences are observed. The images show that LPEI leads to polyplexes with less compact DNA structures, as indicated by the presence of free DNA on the periphery of the polyplexes (Figure , red arrow). In contrast, BPEI forms polyplexes with more compact DNA structures that are not exposed on the exterior of the particles (Figure ). Additionally, BPEI complexes exhibit significantly greater aggregation than LPEI complexes. |
67c2447d81d2151a02832368 | 15 | The differences in LPEI and BPEI morphologies suggest molecular-level differences in their binding and packaging mechanisms to DNA. To further investigate these differences, we first studied the kinetics of polyplex formation and the binding affinity of LPEI and BPEI polymers to DNA using UV absorbance spectroscopy. The kinetics of polyplex formation at N/P ratios of 1 and 5 were followed by monitoring the apparent absorbance changes in the DNA at 260 nm (Figure ), corresponding to the π → π * transition of the nucleobases. As shown in Figure , DNA absorbance at 260 nm decreases over time in the presence of both LPEI and BPEI, indicating DNA condensation due to complexation with these polymers . |
67c2447d81d2151a02832368 | 16 | The offsets in eqs. 2 and 3 indicate that additional processes occur at long timescales (with time constants beyond 24 h) during BPEI polyplex formation. We also investigated the binding affinities of LPEI and BPEI polymers to DNA cargo (Figure ). The fraction of DNA bound to polymer (θ ) was determined using the equation: |
67c2447d81d2151a02832368 | 17 | The corresponding binding curves (Figure ) were fit to a cooperative binding model using the Hill equation: The data was fit to a cooperative binding model using the Hill equation. The goodness-offit was assessed using the R 2 value, which is 0.99 for LPEI and 0.97 for BPEI. Two replicates were measured and the standard deviation was calculated to be less than 2% of the mean for each data point shown. |
67c2447d81d2151a02832368 | 18 | Our kinetics and binding affinity results suggest that LPEI and BPEI bind to DNA and form polyplexes through different mechanisms. The kinetics data show that both polymers exhibit a complex, multi-phasic process for binding and packaging DNA. These processes may involve initial electrostatic binding of the polymers to the cargo, followed by structural rearrangements or aggregation further condense the DNA at longer timescales. BPEI complexes, however, show a slight hyperchormism effect after approximately 10 hours, suggesting possible intercalation into the DNA nucleobases that disrupts their π-stacking ? . Additionally, DNA binding to both polymer systems saturate at approximately 80% (Figure , Table ). Both LPEI and BPEI show positive binding cooperativity to DNA with similar affinities. However, BPEI has a higher Hill coefficient (4.4) than LPEI (2.6), indicating stronger cooperativity, likely due to its branched structure, which could enable additional binding modes provided by its primary, secondary, and tertiary amines, compared to LPEI, which only has secondary amines. |
67c2447d81d2151a02832368 | 19 | Wavenumber (cm -1 ) Wavenumber (cm -1 ) FIG. . FTIR spectra of DNA in the presence of (a) LPEI and (b) BPEI in phosphate buffer, pH = 7.4. The black arrow in each set of spectra represents the spectral changes that occur from N/P ratios of 0, 0.25, 0.5, 1, 1.5, 2, 2.5, 3. All the spectra were blank subtracted, baselined, and normalized to the total integrated area. level insights into their interactions. To investigate this, we measured the FTIR spectra of free DNA, LPEI, and BPEI, as well as DNA in the presence these polymers at N/P ratios ranging from 0 -3. The spectra of LPEI and BPEI (Figure ) are relatively weak compared to DNA, showing bands around ∼1450 cm -1 (CH 2 deformation) and 1620 cm -1 (NH 2 scissoring) . contrast, DNA (Figure ) possesses bands at ∼1222 -1 (PO - 2 asymmetric stretching) and ∼1680 cm -1 (C--O stretching of thymine and guanine). Additional weak bands that derive from inplane ring stretching modes of the nucleobases appear between 1250 -1600 cm -137-41 . |
67c2447d81d2151a02832368 | 20 | Significant spectral changes are observed in the DNA bands upon the addition of LPEI and BPEI (Figure ). For example, the 1222 cm -1 band of DNA decreases in intensity and down- shifts by ∼7 cm -1 as the N/P ratio of LPEI increases (Figure ). Additionally, although they overlap partially with the ∼1450 cm -1 band of the polymers, the nucleobase ring stretching bands between 1400 -1500 cm -1 increase in intensity as the N/P ratio increases. The 1680 cm -1 band also exhibits intensity changes, although no clear trend is observed with increasing N/P ratio. |
67c2447d81d2151a02832368 | 21 | In contrast, BPEI binding (Figure ) results in complex changes in the DNA bands between ∼ 1300 -1500 cm -1 , which generally increase in intensity as a function of N/P ratio. In addition, pronounced intensity decreases for both the 1222 cm -1 and 1680 cm -1 DNA bands are observed as the N/P ratio increases. However, unlike LPEI, binding of BPEI to DNA does not result in a frequency shift of the PO - 2 asymmetric stretching band at 1222 cm -1 . |
67c2447d81d2151a02832368 | 22 | These spectral changes suggest that LPEI and BPEI bind DNA and package it into polyplexes through distinct mechanisms that involve interactions with both the phosphate backbone and nucleobases. For example, the frequency shift and intensity changes in the 1222 cm -1 band indicate that LPEI primarily binds DNA through strong electrostatic interactions with the phosphate backbone. The downshift of this band to 1215 cm -1 at high N/P ratios further suggests that LPEI binding to the phosphate backbone induces a conformational transition in DNA from B-form to Z-form . |
67c2447d81d2151a02832368 | 23 | In contrast, the absence of a frequency shift in this band for BPEI suggests weaker binding to the phosphate backbone that does not alter the backbone conformation of DNA. Instead, the significant decrease in intensity of this band suggests that the primary amines of BPEI likely form hydrogen bonds with the phosphate oxygen atoms of DNA in addition to electrostatic interactions. Additionally, the substantial intensity changes observed in the nucleobase bands at 1680 cm -1 and between 1300 -1500 cm -1 indicate that BPEI also intercalates into the DNA duplex, most likely through hydrogen bonding interactions with nucleobases. The concomitant increase in the intensity of the bands between 1300-1500 cm -1 indicates that BPEI interactions with the nucleobases disrupts their native π-stacking interactions in the DNA. |
67c2447d81d2151a02832368 | 24 | To validate our spectral interpretation, we examined LPEI and BPEI polyplex formation in the presence of NaCl. We hypothesized that Na + and Cl -could screen electrostatic interactions and thereby inhibit polymer binding to the DNA phosphate backbone. If LPEI primarily interacts with DNA via electrostatic forces, its binding should be more sensitive to salt screening effects than BPEI, which our data suggests preferentially interacts with nucleobases. |
67c2447d81d2151a02832368 | 25 | Figure and Figure show the IR spectra of DNA with LPEI and BPEI at varying N/P ratios and NaCl concentrations. The spectra of DNA in the presence of NaCl (Figure ) confirm that Na + ions do not significantly bind DNA on their own. However, when introduced to polymer-DNA solutions (Figure and), clear differences can be seen in the complexation behavior of LPEI and BPEI. |
67c2447d81d2151a02832368 | 26 | To better understand these differences, we constructed polyplex phase diagrams for LPEI and BPEI as a function of both N/P ratio and NaCl concentration. These phase diagrams help visualize the chemical conditions that favor polyplex formation. Using multivariate curve resolution-alternating leastsquares (MCR-ALS) analysis (see SI for details), we decomposed the experimental spectra of LPEI and BPEI polyplexes to extract two basis spectra, one representing polymerbound DNA and the other representing unbound DNA (Figure and). The extracted basis spectra closely match the experimental spectra of bound and unbound DNA (Figure and e), confirming the accuracy of the decomposition. Using these basis set spectra, we determined the fraction of bound and unbound DNA under different NaCl concentration conditions and polymer N/P ratios (Figure ). We accomplished this by modeling each experimental spectrum as a linear combination of the bound and unbound DNA basis spectra. From the extracted concentration profiles, we then calculated the probability of DNA binding to LPEI and BPEI (P b ) (Figure and f, Figure ): |
67c2447d81d2151a02832368 | 27 | where f b and f u are the relative fractions of bound and unbound DNA obtained from the MCR-ALS fitting analysis. The probabilities obtained from eq. 6 were then used to construct the phase diagrams for LPEI and BPEI polyplexes. The resulting phase diagrams (Figure ) show that NaCl impacts LPEI and BPEI complexation to DNA differently. NaCl significantly reduces LPEI binding to DNA by up to ∼ 70% (Figure ), depending on the N/P ratio. In contrast, it has a significantly smaller effect on BPEI to DNA, except at high NaCl concentrations (200 and 250 mM) (Figure ), where binding is reduced by ∼ 20 -30%. The significant susceptibility of LPEI binding to NaCl validates our hypothesis that it preferentially binds DNA through interactions to the phosphate backbone rather than through hydrogen bonding interactions with the nucleobases. In contrast, the insensitivity of BPEI binding to NaCl supports the notion that it interacts with DNA cargo primarily through hydrogen bonding interactions with the nucleobases. |
67c2447d81d2151a02832368 | 28 | Taken together, we present a binding model for LPEI and BPEI polyplexes in Figure . Our model suggests that LPEI binds electrostatically to the phosphate backbone of DNA with a weaker binding to the nucleobases (Figure , step 1). This binding causes DNA structural rearrangements from Bform to Z-form and condensation (Figure , step 2). The addition of NaCl inhibits the binding of LPEI to the phosphate backbone, thereby making binding to the DNA nucleobases preferable (Figure10, step 3). In contrast, BPEI binds DNA primarily through hydrogen bonding and weak electrostatic interactions to the phosphate backbone, as well as hydrogen bonding interactions with the nucleobases (Figure , step a). This multi-modal binding causes efficient condensation of the DNA, which subsequently contributes to the disruption of Developing effective polymers that can deliver multiple types of cargo is essential for advancing polymer-based gene therapies. This can be achieved by designing polymers with multiple binding mechanisms while optimizing cargo release. PEI-based delivery systems are among these and can be optimized to efficiently deliver plasmid DNA, mRNA, and CRISPR-Cas9 technology . PEI is traditionally thought to interact with nucleic acids through electrostatic forces, but our findings show that multiple binding modes play a key role in DNA condensation. Recognizing and using these distinct binding mechanisms provides a foundation for engineering more versatile PEI formulations that improve DNA release, accommodation of other cargo types, and enhance gene delivery. |
668d80145101a2ffa8e358c3 | 0 | We adapted literature salt metathesis protocols for the synthesis of the dinuclear Sm(II) complex [Sm{P(SiMe3)2}{μ-P(SiMe3)2}3Sm(THF)3] from [SmI2(THF)2] and 2 eq. K{P(SiMe3)2} in THF to prepare 1-Ln and 2-Yb from parent [LnI2(THF)2] (Ln = Sm, Eu, Yb) and 2 or 3 eq. of K{P(SiMe3)2} in diethyl ether (Scheme 1). Reactions were initiated at -78 °C and reaction mixtures were allowed to stir for 1 hr at this temperature before allowing to warm to room temperature briefly, before volatiles were removed in vacuo and the products extracted into pentanes. Under these conditions recrystallization from saturated solutions reproducibly gave poor isolated yields of 1-Ln (ca. 15%); 2-Yb was only characterized by single crystal XRD (see below). These 'ate' complexes were the only Ln-containing complexes that we were able to identify in the reaction mixtures, though if either excess pyridine or stoichiometric 18-crown-6 is added before work-up then the respective solvated adducts 3-Ln and 4-Ln form, with concomitant loss of K{P(SiMe3)2} incomplete combustion arising from carbide formation, which is a common feature for highly air-and moisture-sensitive complexes. For 3-Ln the hydrogen and nitrogen values obtained are consistent with the loss of one bound pyridine upon drying samples in vacuo, with low carbon values again assigned to carbide formation; the low carbon and hydrogen values observed in elemental analysis results of [Yb(PPh2)2(THF)4] was previously assigned to the partial loss of THF in vacuo. The ATR-IR spectra of 1-Ln, 3-Ln and 4-Ln show largely overlapping absorption bands for each separate family, indicating that the bulk samples show similar features in the solid state. |
668d80145101a2ffa8e358c3 | 1 | The solid state structures of 1-Ln, 2-Yb, 3-Ln and 4-Ln were determined by single crystal XRD (see Figure for depictions of 1-Eu, 2-Yb, 3-Eu and 4-Eu; as other congeners of 1-Ln, 3-Ln and 4-Ln are isostructural they are depicted in the SI Figures ). Selected bond distances and angles for all complexes are presented in Tables , and crystallographic parameters are compiled in the SI Tables . |
668d80145101a2ffa8e358c3 | 2 | The solvent-free Yb(II) 'ate' complex 2-Yb (Figure and Table ) is a 1D coordination polymer in the solid state that is held together by electrostatic interactions between Yb and K cations, and Pʹʹ anions. The structure can be viewed as a block copolymer with alternating {KYb(Pʹʹ)3} and {K2(Pʹʹ)2} units, or alternatively as {K2Yb(Pʹʹ)4} fragments bridged by {K(Pʹʹ)}. ). The solid-state structures of 1-Ln do not appear to be fully maintained in d6benzene solutions, with three broad signals observed in the H NMR spectra of each complex due to a combination of paramagnetic broadening and dynamic aggregation processes. These resonances did not give reliable integrations for silyl groups or the α-and β-H of THF and so these spectra could not be interpreted; similarly, no signals arising from 1-Ln could be assigned in their C{ 1 H}, Si DEPT90 and 31 P{ 1 H} NMR spectra. |
668d80145101a2ffa8e358c3 | 3 | Although 2-Yb is diamagnetic this complex was obtained as a mixture together with other Yb(II)-containing products, thus the NMR spectra could also not be fully assigned. At 298 K the 1 H.NMR spectrum of a d8-toluene solution of 2-Yb showed the presence of THF, indicating that THF-solvated Yb or K complexes were present in this mixture, thus the broad resonance at 0.65 ppm that is likely due to the SiMe3 groups could not be reliably integrated. Similarly multiple signals were observed in the 31 P{ 1 H} NMR spectrum of this mixture at 298 K, with only a signal at -236.9 ppm that could be assigned to K{P(SiMe3)2} with reasonable confidence by comparison with an authentic sample. We do not observe a signal with well-resolved 1 JPYb coupling for unambiguous characterization of 171 Yb-bound P nuclei, though the broad signal at δP = -219.0 ppm is the predominant feature by approximate integration and is likely a Yb(II)-containing complex. We investigated the dynamic behavior of this mixture by performing variable temperature resolved with heating or cooling. The corresponding ). The 1 H and 13 C{ 1 H} NMR spectra of 3-Yb and 4-Yb were essentially unremarkable, save that the molecular symmetry in solution is higher than in the solid state, and that in the 13 C{ 1 H} NMR spectra the silyl group resonances are virtual triplets due to splitting by strongly coupled 100% abundant I = ½ Complex 1 H (δ) C{ H} (δ) Si DEPT90 (δ) |
668d80145101a2ffa8e358c3 | 4 | As the electronic absorption features of Ln(II) complexes are sensitive to both the identity of the Ln(II) ion and the coordination environment, 1 the electronic transitions of 2 mM toluene solutions of 1-Ln, 3-Ln and 4-Ln were studied by UV/Vis/NIR absorption spectroscopy (see Figure for compiled spectra; individual spectra are available in the SI Figures ). Within the visible region for Ln(II) complexes, strong broad absorptions are observed as a result of the formal spin-and Laporte-allowed f-d transitions from the stabilization of 5d-orbitals (c.f. Ln(III) complexes). The intraconfigurational f-f transitions, which are typically Laporte-forbidden are often not observed for Ln(II) complexes due to the intense, broad f-d features. It is known that the Sm(II) ion gives rise to complexes which exhibit intense colors as a result of the multiple spin-and orbital-allowed f-d transitions within the visible region. Toluene solutions of dinuclear 1-Sm are dark green and display a broad and intense absorption feature containing multiple shoulders which spans across the visible region (Figure ), with λmax = 22,730 cm -1 (440 nm, ε = 1575 M -1 cm -1 ). This complex has the highest absorption coefficient amongst the family of complexes reported herein, likely as a result of the presence of two Sm(II) centers. |
668d80145101a2ffa8e358c3 | 5 | The absorption band in the visible region is more intense than the UV feature, which is typically assigned to a charge transfer process. Toluene solutions of mononuclear 3-Sm and 4-Sm are also dark green and exhibit similar broad intense absorptions to 1-Sm, with these bands also spanning the visible and UV regions. Complex 3-Sm shows two absorption maxima at λmax = 22,530 cm -1 |
668d80145101a2ffa8e358c3 | 6 | for individual spectra). The Sm(II) analogs of these complexes were not studied as the intense absorption features spanning the visible region fully quenches emission from f-d charge transfer electronic transitions through non-radiative decay. The intra 4f luminescent transitions of Ln(III) complexes are well-understood and do not vary significantly with the ligand field due to the radially contracted 4f orbitals. By contrast, the energies of the f-d luminescent transitions of Ln(II) complexes tend to vary to a much greater extent and are more vibrationally broadened. Here, the local crystal field and solvent stabilization effects in the excited state, in combination with the relatively more stabilized d-orbitals, enable other competitive nonradiative relaxation pathways. |
668d80145101a2ffa8e358c3 | 7 | We observed vibrationally broadened emission bands for all solutions studied, which are characteristic for f-d transitions; the luminescence profile of 3-Eu is asymmetric and very broad, spanning most of the visible spectrum. Following excitation between 24,390-50,000 cm -1 (200-410 nm) for 1-Eu and 3-Eu the emission maxima occur in the green region (em = 18,020 cm -1 , 555 nm) of the electromagnetic spectrum and for 4-Eu and 4-Yb these are in the blue region (em = 21,050 cm -1 , 475 nm); |
668d80145101a2ffa8e358c3 | 8 | the spectra of 3-Yb are very weak and could not be interrogated further. As the emission spectra observed do not vary with changes in excitation wavelength it can be inferred that each complex has one excited state in common. These emission bands are best compared with those seen for the Ln(II) complexes [K(2.2.2cryptand)][Ln{N(Si t BuMe2)2}3] (Ln = Eu, em = 18,520 cm -1 , 540 nm, green; Ln = Yb, em = 15,380 cm - 1 , 650 nm, red) and [Eu{N(Si i Pr3)2}2] (em = 16,950 cm -1 , 590 nm, yellow-orange). The excitation spectra of all complexes studied are asymmetric and exhibit some features that match those in the electronic absorption spectra with higher energy maxima than literature complexes (1-Eu: = 31,750 cm -1 , 315 nm; |
668d80145101a2ffa8e358c3 | 9 | 3-Eu: = 34,480 cm -1 , 290 nm; 4-Eu and 4-Yb: = 24,690 cm -1 , 405 nm). The excitation spectra additionally exhibit vibrationally broadened bands at lower energies and no higher energy electronic excitations are observed that are similar to those seen for [K(2.2.2.cryptand)][Ln{N(Si t BuMe2)2}3], where two features are clearly resolved (Ln = Eu, 30,300 and 35,710 cm -1 , 280 and 330 nm; Yb, 27,780 and 31,250 cm -1 , 320 and 360 nm). The data obtained for all complexes are in accord with emission being due to the initial population of a higher energy charge transfer state followed by the deactivation of an f-d excited state. The luminescence lifetimes of 1-Eu and 3-Eu recorded at the respective emission maxima following 26,670 cm -1 (375 nm) excitation were fitted satisfactorily to biexponential decay models, whereas |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.