id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
64f85f2f3fdae147fa9b65df | 19 | Cubic BaSnF 4 was synthesised by mechanically milling BaF 2 and SnF 2 in a manner similar to Ref. 94 (see Methods). The X-ray diffraction pattern (Fig. ) indexes to a face-centered cubic structure from the F m 3m (225) space group, consistent with an average fluorite-structure. The X-ray pattern shows no visible peaks for the parent SnF 2 (C2/c) phase, and Energy Dispersive X-ray analysis (EDX-mapping) shows homogeneous distributions for both Sn and Ba. Quantitative analysis of the EDXmapping data gives proportions of Ba and Sn of 49.6(7) % and 50.3(7) %, respectively, which is close to the nominal 1:1 Ba:Sn stoichiometry (see Fig. for the full mapping data). As a further check on the synthesised compound, we performed electrochemical impedance spectroscopy, and obtained an ionic conductivity at 30 • C of 4.6 × 10 -6 S cm -1 , with an activation energy of 0.56 eV. This ionic conductivity is consistent with previous literature values for c-BaSnF 4 , and is > 10 3 higher that that of fluorite-structured BaF 2 , illustrating the positive effect of cation mixing on fluoride-ion transport. |
64f85f2f3fdae147fa9b65df | 20 | Our XRD data show no superstructure reflections, indicating that Ba and Sn are disordered over the cation sites. From indexing the XRD data, we obtain a cell parameter of a = 6.1945(2) Å, which is close to the value for pristine BaF 2 of a = 6.1964(2) Å . This result is somewhat unexpected, given the smaller ionic radius of Sn 2+ compared to Ba 2+ , and suggests the possibility of local distortions in the cation substructure. Düvel et al. reported similar excess-volume behaviour in Ba 1-x Ca x F 2 solid solutions , where this was proposed as a con- tributing factor to enhanced fluoride-ion transport relative to the end-members. HRTEM data provide further evidence of local deviations from an ideal fluorite-type structure; these show visible changes in inter-reticular distances (Fig. , white arrows) that indicate regions of local strain. Additional structural information is given by our Xray total-scattering PDF data. For interatomic distances between 12 Å and 50 Å, the PDF data are relatively well described by a cubic fluorite F m 3m model (R w = 20 %; see Fig. ). At short range, however (between 1 Å and 12 Å; Fig. ), this high-symmetry structural model gives a poor fit to the PDF data (R w = 32 %), indicating that the local structure of c-BaSnF 4 deviates significantly from an ideal fluorite-type structure. The cubic model fails to predict the peak observed at r = 2.08 Å and the apparent splitting at r = 3.96 Å to 4.15 Å. In other fluorides, Sn adopts short Sn-F distances (e.g., 2.28 Å in tetragonal BaSnF 4 , and as short as 2.03 Å in SnF 2 ). We therefore provisionally assign the peak at r = 2.08 Å to short Sn-F bonds, which requires that Sn or F, or both species, are displaced from their ideal fluorite positions. |
64f85f2f3fdae147fa9b65df | 21 | Fig. (b) shows the room-temperature 119 Sn Mössbauer spectrum for our c-BaSnF 4 sample. The spectrum features an asymmetric quadrupole doublet with an isomer shift of around 3 mm s -1 , characteristic of covalently bonded Sn (II), and a large quadrupole splitting parameter (∆ > 1.5 mm s -1 ), indicating that Sn exhibits a stereoactive lone pair . The experimental spectrum can be reconstructed using two quadrupole doublets with distinct isomer-shift and quadrupole-splitting parameters (see the Supporting Information for details), indicating some degree of variation in Sn-F bonding interactions and in the coordination geometry around individual tin cations. |
64f85f2f3fdae147fa9b65df | 22 | Dénès and co-workers have previously proposed a structural model for fluorite Ca 1-x Sn x F 2 , on the basis of experimental XRD and Mössbauer data similar to those reported here . In this model, the presence of a tin stereoactive lone pair causes each tin cation to be displaced towards one face of its enclosing [F8] cube, giving square-pyramidal SnF 4 E coordination with a reduced nearest-neighbour Sn-F distance. This structural model at first appears to be consistent with our XRD and Mössbauer data and hence to provide an explanation for the deviation from the ideal fluorite structure evident in the short-range PDF data described above. The position of the first peak in our PDF data, however, at r = 2.08 Å, is too short to be explained by square-pyramidal SnF 4 E coordination within an ideal cubic array of fluoride ions: the shortest possible Sn-F distance from this model is a √ 2/2 = 2.19 Å. We therefore interpret the PDF feature at r = 2.08 Å as indicative of a significant degree of distortion to the fluorine substructure away from the reference simple-cubic structure. The structural model of Dénès et al. also predicts equivalent SnF 4 E coordination for all tin cations, which is inconsistent with the apparent variation in bonding and coordination geometry for tin cations evidenced by the Mössbauer data. |
64f85f2f3fdae147fa9b65df | 23 | More detail about the local structure of c-BaSnF 4 , including the behaviour of the Sn lone pair, is provided by analysing structures obtained by quenching from an AIMD simulation. Fig. shows a (001) cross-section through the electron localisation function (ELF) , calculated for a quenched structure from our AIMD simulation. This cross-section intersects with the Wyckoff 4a positions that are occupied by cations in the perfect fluoride structure. Atoms are visible as regions of non-zero ELF density, and each chemical species, Ba, Sn, and F, has a distinct appearance. Ba are visible as bright symmetric rings that are centered approximately on the 4a positions, indicating that barium sits close to its ideal fluorite position. Sn appear as less bright rings, with a bright eccentric lobe that corresponds to the stereoactive lone pair. These lone pairs are generally oriented approximately along ⟨100⟩ directions. The Sn centers appear either to be close to the 4a positions, or, where they are displaced, the displacement appears uncorrelated with the orientation of the lone pair. |
64f85f2f3fdae147fa9b65df | 24 | We also observe ELF features due to fluoride ions, even though the (001) plane in the figure contains no tetrahedral 8c sites, and therefore should contain no fluoride ions for a perfect fluoride structure. However, we observe a number of fluoride ions occupying either octahedral or tetrahedral-edge positions, showing a high degree of fluoride-ion disorder. |
64f85f2f3fdae147fa9b65df | 25 | A more quantitative analysis of the c-BaSnF 4 structure is presented in Fig. . Fig. (a) shows the probability distributions of cation distances from their closest 4a site, P [r(M -4a)], for Ba and Sn. Both cation species are, on average, displaced from their corresponding ideal fluorite cation positions, indicating how the cation substructure is locally distorted from a perfect fcc lattice. On average, Sn is displaced further from the nearest 4a position than Ba, which is consistent with the smaller size of Sn. In general, however, the two probability distribution functions have similar shapes, indicating no qualitative difference between Ba positions and Sn positions relative to their formal crystallographic sites. The calculated Sn-F coordination number, however, rises continuously after the first step, reaching ∼ 8 at 4 Å: the average coordination environment around Sn includes four neighbouring F that occupy 8c-type positions, with four more F in some diffuse disordered arrangement at distances of 2.3 Å to 4.0 Å. Our data therefore suggest an alternative model for the Sn coordination environment, where the Sn lone pair is accommodated not by Sn being displaced significantly from the 4a position, but instead by a significant disruption of the fluoride ions on the lone-pair-adjacent face of the nominal [F8] coordination environment, which presumably reduces the mutual electrostatic repulsion expected between these fluoride ions and the proximate lone pair . Additional evidence for significant disordering of fluoride ions comes from the F-F RDF (Fig. ), which shows a very weak second peak, more typical of an amorphous glassy phase than a regular crystalline array of atoms. ever, the fluoride-ion density is highly diffuse, which is consistent with the proposal above that Sn lone pairs are associated with significant disorder in the local fluoride substructure. These fluoride-ion density data also suggest that the dynamic behaviour of the fluoride ions is strongly dependent on the identity of the nearby cation species: fluoride ions in Ba-rich regions of c-BaSnF 4 appear to be relatively immobile, while fluoride ions in Snrich regions appear to be much more mobile, and we return to this point below. |
64f85f2f3fdae147fa9b65df | 26 | To obtain another perspective on the degree of fluoride ion disorder, we project the instantaneous fluorideion positions from our AIMD simulation trajectory onto discrete tetrahedral or octahedral sites, defined by the set of Wyckoff 4a sites that define their vertices. This siteprojection gives a remarkable 1/3 of fluoride ions occupying octahedral "interstitial" sites rather than conventional tetrahedral sites-i.e., individual octahedral sites are, on average, equally likely to be occupied by fluoride ions than individual tetrahedral sites. This degree of fluoride-ion site-disorder is even greater than the "massive disorder" found in RbBiF 4 , where 1/4 of fluoride ions occupy nominally octahedral positions . Furthermore, this disorder is not simply a large number of thermally generated anion "Frenkel pairs": quenching from our AIMD simulation produces a 0 K structure with this same proportion of fluoride ions occupying octahedral sites that is 16.7 meV/atom lower in energy than the corresponding optimised structure with all fluoride ions occupying tetrahedral positions. This extreme fluorideion disorder is therefore intrinsic to c-BaSnF 4 . |
64f85f2f3fdae147fa9b65df | 27 | Fig. ) shows the probability distribution (number frequency) of tetrahedral and octahedral sites in our AIMD simulation, subclassified by the number of Ba and Sn cations that coordinate each site. These figures also show the proportion of time during the simulation, or probability, that each type of site is occupied by a fluoride ion. For the tetrahedral sites, the occupation probability depends strongly on the identity of the coordinating cations: as the number of coordinating Sn increases the probability of that site type being occupied by fluorine decreases. Comparing the limiting cases of exclusive Ba-coordination and exclusive Sn-coordination, Ba 4coordinated sites are occupied nearly 100 % of the time, while Sn 4 -coordinated sites are nearly always vacant (raw numerical data are available in the Supporting Information). In contrast, for octahedral sites, the occupation probabilty depends much less strongly on the identity of the coordinating cations; each type of octahedral site is occupied approximately 2/3 of the time, although we do observe a weak preferential occupation of octahedral sites with equal numbers of coordinating Ba and Sn. |
64f85f2f3fdae147fa9b65df | 28 | The probability of fluoride-ions occupying a given site can be interpreted as a relative free energy for that site. Fig. ) shows distributions of per-site relative free energies, calculated for each individual tetrahedral and octahedral site as ∆F site = -kT ln(P occ ), where k is the Boltzmann constant, T is the simulation temperature, and P occ is the probability of each site being occupied, calculated from the AIMD simulation. These distributions can be thought of as effective "densities-of-states" of the different tetrahedral and octahedral site types. In Fig. (c), we also show a vertical line corresponding to the point where 2/3 of all available sites are statistically occupied, assuming that sites are preferentially occupied in order of increasing relative free energy. |
64f85f2f3fdae147fa9b65df | 29 | In a conventional fluorite, the tetrahedral sites are low energy, and the octahedral "interstitial" sites are higher energy. Moving an anion from a tetrahedral site to an octahedral site increases the total system energy, and forming Frenkel pairs is therefore a thermally activated process. Fig. (c) illustrates how this conceptual model breaks down in c-BaSnF 4 , where the relative free energy of tetrahedral sites increases with increasing Sn coordination. For sites with two or more coordinating Sn, some proportion of these sites are spontaneously depopulated, with the corresponding fluoride ions instead preferentially occupying octahedral sites. For Ba 1 Sn 3 -and Sn 4coordinated tetrahedral sites, this effect is large enough that these sites are nearly fully depopulated, contributing to the high octahedral-site occupation. This behaviour is consistent with a model where Sn lone pairs repel fluoride ions from adjacent tetrahedral sites, forcing these ions to instead occupy octahedral sites. The analysis presented here also indicates that this effect is additive; the more Sn cations coordinating a given tetrahedral site, the stronger the effective repulsion, and the greater the bias to spontaneously depopulate these sites. |
64f85f2f3fdae147fa9b65df | 30 | In M ′ x M ′′ 1-x F 2 mixed-cation fluorites with no stereoactive lone pair, such as Ca x Ba 1-x F 2 , 19 F MAS NMR spectra show five distinct features corresponding to tetrahedral fluorine environments with different combinations of neighbouring cation species, i.e. FM ′ 4-x M ′′ x (x = {0, 1, 2, 3, 4}) . The 19 F MAS NMR spectrum for c-BaSnF 4 instead shows only two distinct contributions at -14 ppm and -45 ppm (Fig. ). The first of these peaks has a δ iso value close to that of BaF 2 (-14.2 ppm), where fluoride ions have occupy Ba 4coordinated tetrahedral sites. The second peak aligns with the average δ iso value of α-SnF 2 (-46 ppm) , in which fluorine is triply coordinated with short Sn-F distances . Based on these comparisons, we assign these features at -14 ppm and -45 ppm to broadly Ba-rich and Sn-rich fluorine environments, respectively. The assignment of fluorine environments into broadly two types is qualitatively consistent with the computational fluorideion density data (Fig. ), where we observe quite different fluoride-ion densities in Ba-rich versus Sn-rich regions of our simulation model. |
64f85f2f3fdae147fa9b65df | 31 | Our XRD data, above, indicate that Ba and Sn are randomly distributed across the fluorite 4a cation sites, and our AIMD simulations predict a complex fluorine substructure. Both results imply that c-BaSnF 4 contains a rich variety of fluoride-ion environments, which might be expected to be observable in the experimental 19 F MAS NMR spectrum, as in other mixed-cation fluorites ;and yet we observe only two peaks. This apparent contradiction can be reconciled with our expectation of a complex fluorine substructure if we consider fluorine exchange between different sites within the host framework . Fluorine exchange between Ba-rich sites can cause individual peaks associated with different Ba-rich environments to coalesce, giving a single observed resonance. The same reasoning applies to Sn-rich environments, suggesting that they too exhibit fluorine exchange on the NMR timescale. |
64f85f2f3fdae147fa9b65df | 32 | A third type of fluorine exchange is that between Ba-rich and Sn-rich environments, which we probe using variable-temperature 19 F MAS NMR spectroscopy. Fig. shows spectra recorded at 40 • C, 65 • C and 90 • C. As the temperature increases, the relative intensity of the peak assigned to fluoride ions in Sn-rich environments also increases, from 54 % to 60 %, at the expense of the peak assigned to fluoride ions in Ba-rich environments, confirming some degree of fluoride-ion exchange between Ba-rich and Sn-rich environments. |
64f85f2f3fdae147fa9b65df | 33 | For a simple two-site exchange between Ba-rich and Sn-rich environments, increasing temperature would be expected to produce a broadening of the associated resonances before their coalescence into a single resonance with an intermediate chemical shift. We do not observe such behaviour, and instead the peaks assigned to Ba-rich and Sn-rich fluorine environments remain distinct across the investigated temperature range. This behaviour is consistent with only some fraction of fluoride ions in Ba-rich environments undergoing exchange with ions in Sn-rich sites, with this fraction gradually increasing with temperature, and with this Ba-rich-Sn-rich exchenge process being slower than exchange between different Sn-rich environments ; i.e., on the same timescale of exchange between Ba-rich and Sn-rich environments, fluoride ions in Sn-rich environments undergo exchange between several different Sn-rich environments. |
64f85f2f3fdae147fa9b65df | 34 | The observation that fluoride ion exchange between Sn-rich environments is much faster than that between Ba-rich environments or between Ba-rich and Sn-rich environments is further supported by the observation of motional narrowing of the Sn-rich peak with increasing temperature, indicating that the so-called fast-exchange regime is reached. This picture of locally inhomogeneous fluoride-ion dynamics is also qualitatively consistent with the time-average fluoride-ion density obtained from AIMD (Fig. ), where Ba-coordinated regions show highly localised fluoride-ion density, indicative of significantly less mobile ions, while Sn-coordinated regions show diffuse interconnected fluoride-ion density, suggesting more facile fluoride-ion motion between these sites. To validate this model of faster fluoride-ion motion in Sn-rich regions, we performed additional analysis of our AIMD data to calculate site-site transition frequencies for each type of tetrahedral and octahedral site. To estimate the degree to which these fluoride-ion site-site transitions contribute to long-range diffusion, rather than simple back-and-forth motion between adjacent sites, we also calculated frequencies of "non-returning" transitions; these are transitions between two sites, 1 → 2, where the next transition made by the mobile ion takes it to a third site, 1 → 2 → 3, rather than returning it to the original site, 1 → 2 → 1. |
64f85f2f3fdae147fa9b65df | 35 | The calculated site-site transition frequencies for tetrahedral and octahedral sites as a function of their Ba/Sn coordination are shown in Fig. , normalised by the proportion of time each site type is occupied-this normalisation gives transition frequencies that are equivalent to average inverse residence times; higher transition frequencies correspond to fluoride ions leaving a particular site more quickly. The calculated site-site transition frequencies for both tetrahedral and octahedral sites generally increase with increasing degree of Sn-coordination, with this effect particularly strong for the tetrahedral sites. These data from AIMD simulation, therefore, are consistent with the model inferred from the variabletemperature NMR and fluoride-ion time-average density data (Figs. 6 & 8(b)): fluoride ions in "Sn-rich" sites are, in general, more mobile than fluoride ions in "Ba-rich" sites. We also note that the site-occupation-normalised transition frequencies are much higher for Ba-rich octahedral sites than for Ba-rich tetrahedral sites, which speaks to the relative stability of tetrahedral sites compared to octahedral sites in Ba-rich regions. |
64f85f2f3fdae147fa9b65df | 36 | The 19 F MAS NMR and AIMD data presented above show that the local mobility of fluoride ions in c-BaSnF 4 is strongly dependent on the local cation composition: fluoride ions in Sn-rich environments are significantly more mobile than those in Ba-rich environments. An obvious partial explanation for this behaviour is that the stereoactive lone pairs on tin cations somehow promote the motion of fluoride-ions in adjacent tetrahedral and octahedral sites. Our calculated time-average fluorideion density (Fig. ) shows that the fluoride-ion substructure is highly diffuse in Sn-rich regions, which further suggests a possible direct interaction between the Sn lone pairs and the mobile fluoride ions. |
64f85f2f3fdae147fa9b65df | 37 | To quantify the degree of spatial correlation between the Sn lone pairs and nearby fluoride ions, we have calculated the lone-pair-fluoride-ion polar spatial distribution function g(r, θ) (Fig. ). This distribution function describes the time-average fluoride-ion coordination environment around tin, as a function of distance from the central tin cation, r, and the angle between the Sn-F vector and the lone-pair-orientation vector, θ. On the reverse side of the central tin from the lone-pair, there is a clear feature at r = 2.1 Å with maximum intensity at 135 • , i.e., the position of the tetrahedral 8c site if the lone pair is oriented towards the center of the opposite cube-face. On the lone-pair side, however, there is a distinct lack of structure, and fluoride density is instead smeared out in a broad region from r > 3 Å. This distribution function is consistent with the model proposed from inspection of the time-average fluoride-ion density plot (Fig. ): the Sn lone pair is preferentially oriented towards one face of the enclosing cubic site, and fluoride ions that would occupy the corners of this face in a perfect fluorite structure are repelled by the lone pair, which strongly disrupts the fluoride structure in the vicinity of the lone pair. |
64f85f2f3fdae147fa9b65df | 38 | Another notable feature of the lone-pair-fluoride-ion spatial distribution function is that the intense feature corresponding to fluoride ions occupying tetrahedral sites is angularly diffuse. While some of this effect can be attributed to movement of these fluoride ions within their tetrahedral sites, it would be surprising for such movement to preserve the Sn-F separation. An alternative process that provides an explanation for the angular form of this feature is that the tin lone pair is reorienting relative to the reference fluorite lattice on a simulation timescale. To quantify any lone-pair reorientation dynamics, we calculated the Sn-dipole orientational autocorrelation function ⟨μ(0) • μ(t)⟩, which describes the average change in relative orientation of the stereoactive lone pairs in time t. This autocorrelation function (Fig. ) shows a clear decay on a picosecond timescale, showing that tin lone pairs in c-BaSnF 4 undergo dynamic reorientation. |
64f85f2f3fdae147fa9b65df | 39 | The Sn-dipole orientational autocorrelation function does not decay to zero. Therefore, on average, the orien- tation of each tin lone pair is biased, with the lone pair more likely to point in one particular direction than in another. Plotting individual dipole-orientation autocorrelation functions for each lone pair (see the Supporting Information) shows that the strength of this bias varies significantly across tins, indicating that the degree of orientational bias is sensitive to the local tin environment. |
64f85f2f3fdae147fa9b65df | 40 | To probe the degree to which the local tin environment directs the orientational bias for individual tin lone pairs, we calculated, for each lone pair, the proportion of time that this lone pair points towards each face of the enclosing cubic site. Each Sn has 12 cation nearest-neighbours arranged in a cuboctahedron. For a given ⟨001⟩ vector from the central Sn, four of these cations are infront of the central Sn, and coordinate the fluoride sites on the front-face of the Sn 4a site, and four of these cations are behind the central Sn, and coordinate the fluoride sites on the back-face of the Sn 4a site-the other four neighbouring cations occupy the same {001} plane as the central Sn. Because the local fluorine environment depends on the arrangement of the nearby Sn and Ba cations (as shown above; Fig. ), we consider the numbers of Ba and Sn cations coordinating the front-face and back-face of each tin as an effective descriptor for the degree to which a particular tin has a symmetric or asymmetric local coordination environment. |
64f85f2f3fdae147fa9b65df | 41 | In Fig. (b) we show the proportion of time a lone pair points towards a given face of the enclosing cubic site, as a function of the number of nearest-neighbour Sn (out of a maximum of 4) that coordinate the frontface 8c sites and the number of nearest-neighbour Sn (again out of a maximum of 4) that coordinate the backface 8c sites, with the data presented as a heat map. Data on the diagonal where n(Sn) front = n(Sn) back correspond to lone pair orientations with symmetric frontface-back-face nearest-neighbour cation environments. These data all show relatively low values, indicating that lone pair orientations with balanced cation coordination are weakly or negligibly biased. In contrast, lone pair orientations with more front-face Sn neighbours than backface Sn neighbours show a strong bias. As a consequence, the stereoactive Sn lone pairs in c-BaSnF 4 , on average, tend to point towards other nearby tins. Clusters of Sn cations are therefore expected to have all their lone pairs preferentially oriented towards the interior of the cluster, giving a cooperative effect where these Sn lone pairs all disrupt any fluoride ion occupation of mutually coordinated tetrahedral sites. This model is consistent with the increasing tetrahedral site free energy with increasing Sn-coordination (Fig. ) and provides an explanation for the extreme disruption of the fluoride substructure in Snrich regions, as observed in the fluoride-ion time-average density (Fig. ). |
64f85f2f3fdae147fa9b65df | 42 | The timescale for lone-pair reorientation is similar to the timescale of fluoride-ion site-site transitions, which suggests possible coupling between these two kinds of dynamics. To examine whether the fluoride ion dynamics and lone pair dynamics are, in fact, coupled, we performed an additional AIMD simulation with all fluoride ions fixed at their ideal fluorite positions, and calculated the corresponding Sn-dipole orientational autocorrelation function. With the fluoride ions fixed, the lonepair orientational autocorrelation function decorrelates on a sub-picosecond timescale (Fig. ), decaying to a rotationally symmetric (unbiased) value of zero. |
64f85f2f3fdae147fa9b65df | 43 | This rapid decay of the Sn-dipole orientational autocorrelation funcion when the fluoride positions are fixed suggests that fluoride-ion dynamics and lone pair dynamics are strongly coupled. When the fluoride ions are fixed to their 8c lattice positions, the lone pair moves freely-no matter which direction it points in there is a strong lone-pair-fluoride repulsion. When the fluoride ions are free to move, however, a number of these fluoride ions move from unstable tetrahedral sites into more favourable octahedral sites, leaving vacant tetrahedral sites next to tin. The Sn lone pair preferentially orients towards these vacant sites to minimise the lone-pairfluoride-ion repulsion (Fig. ). As fluoride ions move between sites, the Sn lone pairs dynamically reorient in concert with the changing local fluoride ion configuration, giving strong coupling between the fluoride-ion dynamics and the lone-pair reorientation dynamics. |
64f85f2f3fdae147fa9b65df | 44 | To develop solid electrolytes with high ionic conductivities it is necessary to understand how the chemistry of the host framework modulates the structure and dynamics of the mobile ion species . In many families of solid electrolytes, introducing compositional or structural disorder within the host framework is an effective strategy for increasing ionic conductivity . (M, Sn)F 2 fluorites have previously been proposed to exhibit two distinct forms of hostframework disorder : cation-site-occupation disorder, where the two cationic species are distributed randomly over the available sites; and Sn-lone-pair orientational disorder, where Sn exhibits stereoactive lone pairs with random orientations. This proposed coexistence of two distinct forms of host-framework disorder makes (M, Sn)F 2 fluorites a particularly focus of study in the context of understanding the possible interplay between disorder types, and how, together, they modulate ion transport. |
64f85f2f3fdae147fa9b65df | 45 | Here, we have investigated the structure and fluorideion dynamics of the cation-disordered fluorite cubic (c-)BaSnF 4 . Rietveld refinement of XRD data confirms an average fluorite structure with {Ba,Sn} disorder (Fig. ). 119 Sn Mössbauer spectroscopy demonstrates the presence of stereoactive Sn(II) lone pairs, and totalscattering PDF data show clear deviations from the average fluorite structure at short range (Fig. ). |
64f85f2f3fdae147fa9b65df | 46 | Using ab initio molecular dynamics (AIMD) simulations, we have shown that the fluorine substructure in c-BaSnF 4 is highly inhomogeneous and depends strongly on the local cationic composition (Fig. and Fig. ). In Ba-rich regions, the fluoride ions occupy fluoritelike tetrahedrally-coordinated sites that form [F8] cubes around barium. In Sn-rich regions, in contrast, the fluoride-ion substructure is highly diffuse, with fluoride ions displaced from tetrahedral sites adjacent to tin into octahedral "interstitial" sites. |
64f85f2f3fdae147fa9b65df | 47 | We attribute the displacement of fluoride ions from tinadjacent tetrahedral sites into octahedral interstitial sites to the presence of the stereoactive lone pair on the tin cations. The tin cations sit relatively close to their ideal fluorite positions, and exhibit highly eccentric charge distributions that are characteristic of a stereoactive lone pair, in agreement with our 119 Sn Mössbauer data. This lone-pair charge density destabilises fluoride ions occupying adjacent tetrahedral sites, in effect pushing these fluoride ions into octahedral sites, thereby strongly disrupting the fluoride-ion substructure. This effect is clearly seen in the Sn-lone-pair-fluoride-ion polar spatial distribution function (Fig. ), where fluoride ions on the back face of Sn sites-i.e., the opposite side from the lobe of the lone pair-are well structured, while fluoride ions on the front face of Sn sites-in the direction the lone pair is oriented-are strongly repelled and highly disordered. |
64f85f2f3fdae147fa9b65df | 48 | As a consequence of this Sn-lone-pair-fluoride-ion repulsion, c-BaSnF 4 exhibits a remarkable concentration of "interstitial" fluoride ions that occupy octahedral sites. In our simulations, 1/3 of fluoride ions occupy octahedral sites, making it equally likely that, on average, octahedral sites and tetrahedral sites are occupied by fluoride ions. This level of octahedral-site occupation exceeds that of previously reported "massively disordered" fluorites, such as RbBiF 4 , where 1/4 of fluoride ions occupy octahedral sites . In c-BaSnF 4 this extreme fluoride-ion disorder is a consequence of a high relative free energy of occupation for tetrahedral sites adjacent to tin centers and an associated low relative free energy of occupation for octahedral sites (Fig. ), which is a consequence of the mutual repulsion between Sn-lone-pairs and fluoride ions in adjacent tetrahedral sites. |
64f85f2f3fdae147fa9b65df | 49 | We have also directly probed fluoride ion dynamics and the effect of cation disorder using variable-temperature 19 F MAS NMR experiments and additional analysis of our AIMD data. Our NMR data show that fluoride ions in c-BaSnF 4 can be broadly categorised as residing in either "Ba-rich" or "Sn-rich" environments, with fluoride ions in Sn-rich environments more mobile than fluoride ions in Ba-rich environments. This picture of cationenvironment-dependent fluoride-ion dynamics is corroborated by our AIMD simulations: calculated site-site transition frequencies are higher for sites with a higher proportion of coordinating tin, showing a direct relationship between the local cation configuration and local anion dynamics. |
64f85f2f3fdae147fa9b65df | 50 | Our AIMD simulations also reveal that the tin lone pairs dynamically reorient on a picosecond timescale (Fig. ). By comparing the results from our unrestricted AIMD simulations to equivalent data from simulations where the fluoride ions are fixed to their ideal fluorite positions, we have shown that orientational dynamics of the tin lone pairs is coupled to the dynamics of the nearby fluoride ions. This effect is modulated by the local cation arrangement: for tins with an asymmetric Sn/Ba nearest-neighbour configuration, the tin lone pair preferentially orients in the direction of other, nearby, tins. Hence, clusters of tin cations exhibit a cooperative effect where by the lone pairs on each tin tend to orient towards the interior of this cluster. This cooperative effect explains the dramatic disruption of the fluoride ion substructure in regions where several Sn cations are clustered together (Fig. ). |
64f85f2f3fdae147fa9b65df | 51 | The measured room-temperature ionic conductivity of c-BaSnF 4 (4.6 × 10 -3 S cm -1 ) is significantly higher than that of, for example, fluorite-structured BaF 2 , which is consistent with the general observation that within structurally-related families of solid electrolytes, hostframework disorder is often correlated with increased ionic conductivities . In other materials, this relationship between host-framework disorder and ionic conductivity has been explained as a consequence of a concomitant disordering of the mobile ion species that promotes ion transport , or of a reduction in differences in site-occupation energies between mobile-ion sites that flattens the mobile-ion potential energy surface . Our results for c-BaSnF 4 are consistent with both of these conceptual models: in Sn-rich regions of the structure, the fluoride-ion density is highly diffuse (Fig. ), indicating significant local fluoride-ion disorder-which is also evident from our calculated F-F radial distribution function, Fig. (d)-while our site-occupation analysis shows a destabilisation of Sn-coordinated tetrahedral sites and a stabilisation of octahedral sites that gives overlapping tetrahedral and octahedral site energies (Fig. ). |
64f85f2f3fdae147fa9b65df | 52 | Given that c-BaSnF 4 exhibits such a high degree of fluoride-ion disorder, it is perhaps surprising that it does not have an even higher ionic conductivity. We observe greater fluoride-ion site-disorder (1/3 of fluoride ions occupying octahedral sites) than in the mixed-valence mixed-cation fluorite RbBiF 4 (1/4 of fluoride ions occupying octahedral sites), which would seem to predict a higher ionic conductivity for c-BaSnF 4 than for RbBiF 4 . The room-temperature ionic conductivity of RbBiF 4 , however, is ×10 2 greater than that of c-BaSnF 4 . This result can be explained by recognising that fastion transport in solid electrolytes requires not only that there is a small, or non-existent, energy gap between occupied and unoccupied sites, but also that these "frontier" sites form a contiguous percolating diffusion pathway through the material . In c-BaSnF 4 , the combined effects of cation disorder and lone-pair-fluoride-ion repulsion produce a large spread in tetrahedral site energies (Fig. ), causing tetrahedral sites with either high-Ba-coordination or high-Sn-coordination to be largely unavailable for long-range fluoride-ion diffusion. Highly-Ba-coordinated tetrahedral sites (e.g., Ba 4 ) have low site energies, are nearly fully occupied, and have low site-site transition frequencies, and fluoride ions occupying these sites are therefore largely immobile. As such, clusters of barium cations are expected to obstruct long-range fluoride ion diffusion. Highly-Sn-coordinated tetrahedral sites (e.g., Sn 4 ) have a similar blocking effect on diffusion, but for the opposite reason; these sites have high site energies and are therefore rarely occupied, despite having very high site-site transition frequencies. As a result, these Sn-coordinated sites obstruct long-range fluoride ion diffusion by acting as high-energy bottlenecks. The remaining mixed-coordination tetrahedral sites (e.g., Ba 2 Sn 2 ) then form a tortuous diffusion pathway, resulting in a lower macroscopic ionic conductivity than might be expected on the basis of local site-site transition frequencies or purely from the high level of fluoride ion disorder present in the structure. |
64f85f2f3fdae147fa9b65df | 53 | The results presented here demonstrate the complex interplay between two distinct forms of host-framework disorder-cationic site-occupation disorder and lone-pair orientational disorder-and the structure and dynamics of the mobile ion species within a fluoride-ion-conducting solid electrolyte. The complex nature of these interacting effects suggests that the resulting effect on mobile-ion dynamics is likely to be highly dependent on the exact composition and structure of the solid electrolyte, and we expect further exploration of the coupling between crystallographic disorder, lone pair dynamics, and ionic conductivity in solid electrolytes to be a fertile area for future research. |
64f85f2f3fdae147fa9b65df | 54 | A complete dataset for the computational modelling and analysis described in this paper is available from the University of Bath Research Data Archive . This dataset contains inputs and outputs for all DFT calculations, plus scripts for analysis of the DFT data and for plotting Figs. . A subsidiary dataset containing only the figure-plotting scripts and relevant input data is available on GitHub . |
6351ae3bee31867fd76b6222 | 0 | The transition from fossil to renewable resources presents one of today's greatest challenges. Their steady depletion and contribution to global warming are driving factors for implementing a more sustainable circular economy which includes using biomass for chemicals and fuels production. Especially for household products that frequently contain polymers, there is a strong consumer demand for green alternatives due to environmental concerns and awareness. Many polymers possess aromatic monomers, e.g., p-xylene for polyethylene terephthalate (PET), styrene for polystyrene (PS), toluene diisocyanate for polyurethanes (PUR), derived from benzene, toluene, and xylene (BTX) of which more than 122 Mt are produced annually, also for use in solvent and fuel applications. As of today, aromatics are obtained by catalytic reforming of naphtha, a crude oil fraction. Thus, using biomass as raw material feedstock for their production, ideally conceived as a drop-in solution, opens the door to a large variety of renewably sourced products without modifying the existing infrastructure downstream. |
6351ae3bee31867fd76b6222 | 1 | Several routes to biomass-derived aromatics have been reported. Among those, one important pathway is the selective Diels-Alder cycloaddition of furanic compounds, especially furan derivatives, with dienophiles such as bio-ethylene. High costs associated with the production of furan and its derivatives from biomass are currently the biggest drawback. Another fundamental route is the depolymerization of lignin by (catalytic) fast pyrolysis which yields a "bio-oil", a complex mixture of oxygenated aromatic compounds. These require extensive purification, separation, and deoxygenation to obtain alkyl aromatics suitable for downstream processing. A more selective but less explored pathway is the self-condensation of alkyl ketones. The C-Ccoupling reaction of alkanones via robust acid/base-catalyzed condensation is an efficient way to achieve deoxygenated aromatics from existing biorefinery streams, e.g., acetone from ABE fermentation, in a single-step reaction without further need for metal-catalyzed dehydrogenation or upgrading via deoxygenation with hydrogen. Furthermore, a highly promising carbonnegative route to acetone exists with gas-phase fermentation by autotrophic acetogens. Therefore, the efficient utilization of acetone via self-condensation for the formation of the aromatic product mesitylene (1,3,5-trimethylbenzene) is the focus of this study. |
6351ae3bee31867fd76b6222 | 2 | Previously, the self-condensation of acetone is mainly studied in the gas phase under atmospheric pressure, often at high reaction temperatures above 400 °C using various solid acid catalysts such as zeolites, titania, zirconia, niobia, mesoporous aluminosilicates, and tantalum phosphate. Despite the high catalytic activity, the formation of polycondensates on the strongly acidic catalysts leads to fast catalyst deactivation due to carbonaceous deposits which remains an inherent challenge. Faba et al. showed an increase in productivity for the gas-phase conversion of acetone over a mixed catalyst bed of TiO2 and Al-MCM-41 at 250 °C but lacked the proof of long-term stability. Recently, we showed that zeolite HY is stable for the mesitylene formation in the liquid phase over several hours at 190 °C. The larger pores of zeolite Y compared to other microporous zeolites proved to be beneficial for the catalyst stability. Additionally, the liquid phase conditions allowed continuous removal of products from the catalyst bed, thus reducing carbonaceous deposits. While zeolite HY was also applicable for the aromatization of larger alkyl methyl ketones, such as 2-butanone and 2-pentanone, its overall activity was still lacking regarding a potential process development for future integration into a biorefinery. In the course of the catalyst development, combining high activity with long-term operational stability remains an ongoing problem that we sought to solve by using large pore amorphous aluminosilicates as acid catalysts. |
6351ae3bee31867fd76b6222 | 3 | Herein, we report that ASA afforded remarkable activity and selectivity for mesitylene from acetone under continuous conditions not only in liquid but also in supercritical phase. The ASA catalysts were beneficial due to their combination of larger mesopores and moderate overall acid site density of medium strength with a lower number of strong Brønsted acid sites. The optimization of reaction conditions and long-term catalyst stability are examined for maximum mesitylene space-time-yield on ASA. Assessment of the acidity-activity-relationship for this material class is performed based on the degree of silica-doping. |
6351ae3bee31867fd76b6222 | 4 | Commercial amorphous silica-alumina (denoted as ASA) were supplied by Sasol Germany GmbH and used after calcination in air at 550 °C for 6 h. HY-5 catalyst was obtained by exchanging NaY three times with an aqueous solution of NH4NO3 (1 M, 60 °C, 1 h) and subsequent calcination in air (2 K min -1 , 550 °C, 6 h). γ-Al2O3 (99.9 %) was obtained from Alfa Aesar and Evonik Aeroperl 300/30 was used as pure SiO2. |
6351ae3bee31867fd76b6222 | 5 | Catalytic studies were performed in a previously described fixed-bed reactor in upward flow configuration (Figure ). Briefly, 3 g of catalyst were placed in the isothermal zone in the center of the stainless-steel reactor (ID = 1.6 cm, length = 20 cm). The catalyst powders were pressed (8 tons), crushed, and sieved to 100-200 micron particles to avoid mass transfer limitations (Figure ). Neat acetone was fed into the reactor with an HPLC pump, and the reaction was performed at 200-300 °C and 75 bar to maintain liquid/supercritical conditions. The outlet of the reaction feed was continuously mixed with a standard solution of 1,4-dioxane (1.15 mol L -1 ) in 1-butanol downstream. The product solution was analyzed via online-gas chromatography (Shimadzu GC-2030, MEGA-5 column, 40-250 °C, 10 K min -1 , H2) equipped with an FID (Figure ). Acetone, mesityl oxide, mesitylene, isophorone, 2-butanone, and 1,3,5-triethylbenzene were calibrated with pure compounds while the other identified products were estimated via the concept of the effective carbon number combined with GC-MS. |
6351ae3bee31867fd76b6222 | 6 | Temperature-programmed desorption of NH3 (NH3-TPD) was used to measure the total amount of acid sites on the silica-alumina catalysts. For this, samples (100 mg) were dried in N2 flow at 600 °C (10 K min -1 , 3 h) and subsequently loaded with NH3 (2 vol% NH3 in N2, 20 mL min -1 ) at 140 °C. When the physisorbed NH3 was desorbed, the samples were heated to 600 °C (10 K min -1 ). The amount of chemisorbed NH3 was detected by FT-IR. |
6351ae3bee31867fd76b6222 | 7 | In our previous work, zeolite HY was identified as a suitable solid acid catalyst for the efficient and stable liquid phase conversion of biomass-derived ketones to aromatics due to its larger pore size and high amount of acid sites. Catalytic activity of HY increased with temperature and at 190 °C steady-state conditions were obtained for the continuous conversion of acetone to mesitylene. However, catalyst productivity was relatively low with 3 % mesitylene yield at a weight hourly space velocity (WHSV) of 7.8 h -1 . When the reaction temperature was raised to 200 °C, the initial productivity significantly increased to about twofold but the activity steadily decreased with time-on-stream due to catalyst deactivation (Figure ). In contrast, ASA Siralox 30 is found to be very stable when converting acetone to mesitylene under similar reaction conditions (200 °C, WHSV = 7.8 h -1 ). While its activity at 200 °C is lower compared to zeolite HY, it shows a three-times increase in mesitylene yield to 4 % for 220 °C. Even at this elevated temperature, the catalyst activity remains very stable for more than 7 hours-on-stream and shows no deactivation. To understand the difference in stability of HY-5 and Siralox 30, relevant properties of the solid acid catalysts are compared in Table . The surface area of the mostly microporous zeolite HY-5 is 778 m 2 g -1 and four-times higher than the one of Siralox 30 which is completely mesoporous (Figure ). The number of surface acid sites is higher on HY-5 with 0.53 mmol g -1 compared to 0.30 mmol g -1 on Siralox 30. This results in a significantly higher activity of HY-5 at reaction temperatures below 200 °C since mesitylene activity correlates with the number of acid sites. Moreover, the strength of the acid sites on HY-5 is also greater than on Siralox 30 as evidenced by the higher temperature of the maximum NH3 desorption (Figure ). |
6351ae3bee31867fd76b6222 | 8 | Thermogravimetric analysis of the spent catalysts in air determines a total mass loss of 20 wt% for HY-5 of which strongly bound, bulky carbonaceous deposits represent more than 17 wt% (Figure ). Siralox 30, on the other hand, shows a minimal weight loss of 3.3 wt% from 200-600 °C which could be due to weakly bound deposits. Thus, the stronger and greater number of acid sites on HY-5 are not only more active, but also lead to increased carbon deposition. Those block the catalyst pores and thus decrease the accessibility of active sites, as highlighted by the strongly diminished microporous surface area of spent HY-5 (Table ). The observed slight increase in the mesoporous surface area of HY is due to carbon deposition as confirmed by TGA. In comparison, the surface area of spent Siralox 30 is almost completely retained due to its mesoporous structure. |
6351ae3bee31867fd76b6222 | 9 | Hence, at higher reaction temperatures (more than 190 °C), mass transport out of the zeolite HY pores is slow compared to the formation of bulkier molecules favored by its stronger acid sites, resulting in quick deactivation of the catalyst. However, Siralox 30 remains stable even at higher temperatures due to its better mass transport capabilities and lower number of stronger acid sites. In this way, it is possible to greatly improve the activity of the catalyst with temperature without compromising stability. Therefore, the ASA Siralox 30 is a highly interesting solid acid catalyst for the liquid-phase aromatization of acetone and requires a deeper study of optimal reaction conditions for maximum activity and efficiency in the continuous flow process. |
6351ae3bee31867fd76b6222 | 10 | Acetone conversion increases almost linearly with temperature from 17 to 63 % between 200 and 300 °C (Figure ). At harsh reaction temperatures of 280 °C and above, there are notable fluctuations in the measured acetone conversions as represented by the increased error over five measurements at steady-state conditions. The selectivity to mesitylene shows a steady incline from 11 % at 200 °C to a maximum of 38 % at 260 °C but decreases for higher reaction temperatures as acetone is primarily converted to undetected products. The selectivity to the dimer mesityl oxide decreases with reaction temperature as the consecutive aromatization to mesitylene is favored. |
6351ae3bee31867fd76b6222 | 11 | While the selectivity to linear and branched aliphatic oligomers declines with temperature, the formation of the non-aromatic cyclic trimer isophorone is not significantly affected. At 240 °C and above, the formation of mesitylene is favored and identified as the main product. Total product selectivity at 260 °C is 66 %, demonstrating that the reaction is effective even at elevated temperatures. No additional impact is observed when switching from liquid acetone to supercritical acetone above 235 °C. By varying the reaction temperature, a strong influence on the mesitylene activity is found with a maximum selectivity at 260 °C for Siralox 30. The following experiments were therefore carried out at 260 °C. |
6351ae3bee31867fd76b6222 | 12 | Besides reaction temperature, space-time-yield is an important factor for process efficiency. In order to optimize the space-time-yield of the acetone aromatization with Siralox 30, the weightto-flow ratio of catalyst mass W to acetone molar flow F is varied at 260 °C, the point of highest mesitylene selectivity. The acetone conversion and mesitylene yield increase for longer contact times which is represented by higher W/F ratios but mesitylene yield does not benefit from ratios higher than 12.5 gcat h mol -1 where it is 10.1 % (Figure ). The further increase in acetone conversion is a result of increased byproduct formation due to significantly longer catalyst contact times. The yield of mesityl oxide decreases for higher W/F ratios, indicating intermediate formation of the dimerization product and subsequent conversion to mesitylene. There was no significant effect of contact time on the formation of oligomers and isophorone. For a variation at 210 °C in liquid phase (Figure ), a constant increase in mesitylene yield to only 8 % was observed at a W/F ratio of 75. It can therefore be concluded that the increased reaction temperature leads to more efficient catalysis of the acetone condensation and thus allows for faster contact times compared to lower reaction temperatures. A reaction temperature of 260 °C and a W/F ratio of 12.5 gcat h mol -1 (corresponding to a WHSV of 4.66 h -1 ) are found as optimal conditions for the consistent formation of mesitylene with the ASA Siralox 30 in the continuous flow fixed-bed reactor. |
6351ae3bee31867fd76b6222 | 13 | The stability of the Siralox 30 catalyst was evaluated under the optimum reaction conditions. In a long-term experiment, it could be shown that the catalyst remains stable for more than 50 hourson-stream despite the comparatively harsher reaction conditions (Figure ). This clearly demonstrates that the mesoporous ASA is a superior catalyst in the acetone aromatization to mesitylene. Based on these findings, Siralox 30 was also tested for the conversion of 2-butanone under the optimized reaction conditions. The aromatic self-condensation product of 2-butanone is triethylbenzene that could be used as a potential precursor to styrene-type monomers. While Siralox 30 is stable for the conversion of this larger alkyl methyl ketone for more than 40 hourson-stream, the catalytic activity is reduced compared to the conversion of acetone (Figure ). |
6351ae3bee31867fd76b6222 | 14 | For 2-butanone, the dimers present the main product with 8 % yield whereas the yield of the aromatic triethylbenzenes is 5 %. The lower catalyst activity is owed to the lower reactivity of 2butanone and its steric hindrance due to the longer alkyl chain. The latter is also responsible that in comparison to acetone, a higher variety of isomers can be formed. The total carbon balance for the conversion of 2-butanone is therefore 80 %, whereas the total product selectivity for dimers, aromatics and trimers is 48 %. As a side product propionic acid was detected in low amounts. |
6351ae3bee31867fd76b6222 | 15 | Nevertheless, the mesoporous structure of Siralox 30 with its larger pores is advantageous for the condensation as it is suitable for larger products without deactivation. For the production of the ASA of the Siralox family, silica is added to high-purity aluminas during synthesis. Depending on the amount of silica, the silica-to-alumina ratio (SAR) is varied and thus material properties such as acidity (strength and density) and thermostability can be tuned. To assess the influence of the SAR on the self-condensation of acetone, Siralox materials from 1-70 % SiO2 were studied under the optimum reaction conditions of 260 °C and 12.5 gcat h mol -1 and presented based on the steady-state conversion over 10 hours-on-stream in Figure . With silicaloading, the mesitylene yield (from 2.8 to 10.5 %) and acetone conversion (from 19 to 42 %) increase to a maximum at 30 wt% SiO2. For a further increase in silica-loadings, the conversion and yields decrease and are virtually zero for pure silica. Especially for the initial doping with 1 wt% silica, a strong increase in activity is observed for both mesitylene yield (7.5 %) and acetone conversion (28 %). The isophorone yield decreases with silica-content. Since bases promote the condensation to isophorone, the amphoteric nature of alumina exhibiting basic sites is relevant at low silica loadings. The yields of mesityl oxide and oligomers increase up to 10 wt% silica, showing a maximum of 4.8 and 3 %, respectively. With higher silica content, both yields decrease but remain constant up to 70 wt% SiO2. Overall, an increase in mesitylene yield is accompanied by an increase in acetone conversion. This confirms that the mesitylene increase mainly stems from the consecutive reaction of mesityl oxide with acetone and subsequent aromatization. Based on the findings, a silica-content of 20-40 wt% gives the highest mesitylene yields with an optimum at 30 wt%. The trends observed in the flow reactor are supported by batch experiments at 230 °C for 3 h (see experimental details in SI and Figure ). At these longer catalyst contact times, the maximum mesitylene yield at 30 wt% is more pronounced. The low activities of pure γ-Al2O3 and pure SiO2 found under flow and batch conditions highlight the effect of the silica-doping on the catalytic activity as additional active sites are created. Indeed, very large silica-loadings of 40 wt% and higher lead to a primarily silica surface with generally fewer surface acid sites. 31 Under the optimized reaction conditions, all tested Siralox materials were stable at steady-state conditions for more than 10 hours-on-stream (Figure ). The silica-content therefore does not play a significant role for the stability of the catalysts under reaction conditions. The TGA measurements of the spent ASA from the flow reactor show that the mass loss is about 7 wt% for silica-loadings of 10-40 wt% (Figure ) after reaction at 260 °C. For very high and silicaloadings, the mass loss is lower than 7 wt% due to decreased catalyst activity. For flow conditions, ~0.5 wt% less carbonaceous deposits are found compared to batch reactions which shows the efficiency of the continuous process with constant product removal from the catalyst bed. The structural stability of the Siralox materials was assessed by XRD of the spent catalysts. The diffractograms (Figure ) show that for the majority of materials, no significant change is visible. For Siralox 20 and 40, a small reflex at 49.2° can be observed, hinting to minor formation of the γ-AlO(OH) boehmite likely due to water formed in the aldol condensation. Surprisingly, Siralox 30 does not show the formation of boehmite which supports its suitability for scale-up in a continuous flow process. |
6351ae3bee31867fd76b6222 | 16 | To understand why the Siralox materials show different activities in the aldol condensation of acetone, their nature and number of acid sites are assessed. The number of acid sites initially increases by doping γ-Al2O3 with 1 wt% silica from 0.21 to 0.49 mmol g -1 . However, the number of acid sites starts decreasing from 20 wt% SiO2 and is strongly decreased for high silica-loadings as in Siralox 70 with 0.065 mmol g -1 (Figure ). No direct correlation with mesitylene yield can be found for the total number of acid sites obtained by NH3-TPD, indicating the importance of specific active acid sites. The desorption temperature of the chemisorbed NH3 is an indicator for the strength of the probed acid sites, but in the case of the ASA, the large variety of acid sites leads to a broad desorption curve which renders it difficult to quantify or select individual sites for comparison based on their acid strength. Instead, the temperature of the desorption maximum is used for estimation as it indicates the bond strength of a majority of acid sites. Analogous to the number of acid sites, low silica-doping leads to an increase in the maximum desorption temperature from 260 °C to approximately 270 °C. For higher silica-contents, except for Siralox 40, the maximum temperature declines. Considering that the highest yields of mesitylene were found for Siralox 20-40, a slightly decreased number of acid sites and strength appears beneficial for the stable formation of mesitylene found for ASA. This could be attributed to the fact that a high density of strong acid sites leads to faster deactivation due to the favored formation of larger byproducts, as seen for zeolite HY. By comparing the dimer yield with the NH3-desorption temperature, it becomes apparent that mesityl oxide benefits from the higher density of slightly stronger acid sites. As for the dimerization two acetone molecules need to react (and possibly be activated in close vicinity), the yield of mesityl oxide is mainly dictated by the amount and strength of acid sites. Due to the structural complexity of ASA with multiple types of acid sites, results obtained by NH3-TPD are insufficient to explain the observed trends. Therefore, the nature and number of acid sites on which the reactions take place must be considered in more detail. |
6351ae3bee31867fd76b6222 | 17 | ASA exhibit a predominant amount of various Lewis acid sites (LAS) but also possess BAS in lower amounts (Figure ). The latter are created by doping the alumina with silica. With increasing silica content, the number of LAS decreases while the formation of BAS increases up to a maximum at 40 wt% silica. For silica-rich alumina of 90 wt% and above, no LAS are detected and the acidity is almost completely controlled by BAS. This is explained by the enrichment of the surface with silica which exceeds what would be expected from the bulk material composition. The majority of the material's surface at silica-loadings higher than 40 wt% is covered by silica and only contains small zones of the mixed aluminosilicate. In fact, Daniell et al. assessed the strength of the BAS and correlated it to the shift of the surface hydroxyl group Δv(OH) at ~3748 cm -1 on adsorption of CO on the Siralox materials by FTIR-spectroscopy. The aluminosilicate surface shows a lower number of BAS which are strongly enhanced in strength by the addition of silica. Based on their measurement, we found that the formation of mesitylene correlates well (R 2 = 0.78) with the measured Δv(OH) (Figure ). Accordingly, stronger BAS promote mesitylene formation under reaction conditions. For silica loadings higher 60 wt%, the total number of acid sites and the strength of BAS decrease which results in the lower activities observed for these materials. Therefore, materials with silica contents of 30-40 wt% are most beneficial for the formation of mesitylene. |
6351ae3bee31867fd76b6222 | 18 | The dimer mesityl oxide is preferably formed on ASA with low silica content which could stem from two effects: 1) mesityl oxide benefits from a higher number of LAS as observed by Panov and Fripiat and 2) a lack of BAS at low silica loadings which are required to promote the consecutive aromatization with another acetone molecule. This concludes to both LAS and BAS being beneficial for the formation of mesityl oxide and dimers, while BAS promote the consecutive formation of mesitylene. |
6351ae3bee31867fd76b6222 | 19 | The remarkable stability of Siralox 30 combined with the high total product selectivity of 66 % renders it suitable for a potential scale-up and further assessment of the industrial application in a continuous flow process. Separation of products can be readily achieved via distillation based on the differences in boiling points. A benefit of the solvent-free process is the more energy-efficient separation, as no additional solvent needs to be vaporized and separated. Furthermore, recovered acetone and mesityl oxide can be recycled and added to the reactor feed to increase total efficiency (Figure ). Aromatics of the BTX-fraction can be obtained from mesitylene via industrially established transalkyation, thus offering a completely bio-based route for their production. Generation of added value is also possible from the side products by their transformation to various important intermediates used in the polymer industry, e.g., isophorone diisocyanate from isophorone for the polyurethane production. 36 Depiction of biomass conversion and separation processes is simplified. |
6351ae3bee31867fd76b6222 | 20 | In conclusion, the continuous production of biomass-derived mesitylene from acetone can be achieved over ASA Siralox 30 under solvent-free continuous conditions in liquid as well as supercritical phase. At optimum conditions, Siralox 30 showed high stability without signs of deactivation for more than 50 hours-on-stream. Contributing to the stability is the mesoporous nature of Siralox 30 which facilitates mass transport and prevents deactivation by pore blocking from bulkier aromatic/aliphatic molecules as evidenced by the low amounts of carbonaceous deposits on the spent catalyst. Additionally, the larger pores allow the stable aromatization of 2- |
66251611418a5379b0514bea | 0 | 1.1 Background. The greenhouse gases emitted as a result of the development industries have been increasing, proved by an average temperature increase of 0.32°F (0.18 °C) per decade since 1981. From 1990 to 2015, the net greenhouse gas emissions from human activities increased by 43%, with CO 2 accounting for about three-fourths of total emissions, which increased by 51 percent ("Inventory of U.S. Greenhouse Gas Emissions and Sinks," 2023). |
66251611418a5379b0514bea | 1 | There are various types of technologies used for carbon capture, which mainly fall under three categories: post-combustion capture (a primary method used in power plants), pre-combustion carbon capture (used in industrial processes), and oxy-fuel combustion systems . Because pre-combustion capture performs better in energy efficiency than other categories due to its lower energy demand for CO 2 , and greenhouse gases are primarily direct emissions from industries. Hence, this paper will focus mainly on the methods within pre-combustion capture to store and use CO 2 . In this process, pre-combustion capture first undergoes the gasification process, in which a feedstock such as coal is oxidized in steam and oxygen/air under high temperature and pressure to form a synthesis gas. This syngas will then undergo the water-gas shift reaction (CO + H 2 O → H 2 + CO 2 ) by injecting H 2 O into the synthesis gas formed with carbon monoxide. The CO 2 is removed before combustion, so the H 2 -rich syngas are then supplied to gas turbines for combustion. This allows a more concentrated stream of CO 2 to be compressed into a liquid to be stored or reused . |
66251611418a5379b0514bea | 2 | Fuel cells consist of two electrodes, a positive electrode (cathode) and a negative electrode (anode), where catalysts are located within the electrodes, as shown in Figure . Both electrodes are made of carbon due to the material's good electrical conductivity and chemical resistance to attack by acid and alkaline electrolytes within fuel cells with low inherent cost. And an electrolyte between the middle of the electrodes that allow the transport of ions (Figure ). To further enhance the efficiency of the reaction between hydrogen and oxygen molecules, the electrodes are designed with a porous network, which ensures the uniform distribution of oxygen molecules. This allows consistent conditions across the catalyst's surface, preventing localized inefficiencies and ensuring that an electrode is made of materials that always have good electrical conductance. However, this porous network could potentially contribute negatively to the stability of the catalyst . The redox reactions that take place in a fuel cell are: |
66251611418a5379b0514bea | 3 | The redox reaction of the fuel cells first occurs as the fuel, hydrogen, is fed through the anode, where it undergoes an oxidation reaction that occurs with the hydrogen fuel (H 2 → 2H + + 2e -), the hydrogen atom is then separated into cations and electrons by the catalysts inside the anode. The cations will move from the anode to the cathode through the electrolyte. During this process, the proton will be surrounded by a proton-conducting polymer, facilitating H + transport across the electrolyte. The hydrogen ions (H + ) from the electrolyte will then be integrated into the reduction reaction to combine the electrons to form water. As water is released as a by-product of this reaction, electricity will be generated from the whole reaction . Simultaneously, the electrons from the anode will then flow through an external circuit from the anode to the cathode, producing direct current electricity. At the cathode, another catalyst causes oxygen to react with the cations and electrons flown from the anode to react in another redox reaction (O 2 + 4H + + 4e -→ 2H 2 O), forming water . This redox reaction, or else called the oxygen reduction reaction (ORR), is an essential process for energy conversion within fuel cells; molecular oxygen (O 2 ) is reduced to water due to higher current, thereby releasing energy that contributes to overall electricity generation while also generating the electrical potential that generates the electricity outputted within the fuel cell, making it the critical reaction in converting energy . Within this reaction, catalysts determine the efficiency of this reaction as it help break the bonds of the molecular oxygen (O 2 ). A better catalyst results in faster reactions, thereby improving the fuel cell's overall performance. |
66251611418a5379b0514bea | 4 | As such, the importance of electrocatalysts during this process can be examined as they can lower the activation energy of the reactions happening within the fuel cells, allowing a more efficient reaction. Because electrocatalysts separate the electrons and deliver them through the external circuit, the magnitude of this current will also be determined by the rate of the electrocatalysts at both the anode and the cathode, which is limited by the electrode with the lowest electrocatalytic rate. Hence, the rate of the electrocatalysts throughout the fuel cell will be crucial to its efficiency. Because the kinetics of the cathode is typically much slower than the anode, the catalysts of the cathode will be significant to the ORR of the fuel cells to maintain a high rate of electrocatalysts at both electrodes of the fuel cell. |
66251611418a5379b0514bea | 5 | 1.5 Aim of Paper. This paper will explore the characteristics of electrocatalysts and what they control. It will first outline the set of characteristics that build a promising electrocatalyst. Then, it will compare the main catalyst groups, platinum-group metal (PGM) catalysts, and non-PGM catalysts to better understand their different advantages and disadvantages. Lastly, it will cover what we can do in the future to make electrocatalysts more efficient in fuel cells. |
66251611418a5379b0514bea | 6 | This section will examine some essential characteristics that contribute to the efficiency of an electrocatalyst, establishing criteria for a promising electrocatalyst that will enhance the overall reaction efficiency of fuel cells from prior research and papers. This section's primary focus areas are the kinetics and durability of the catalyst. While the kinetics and other aspects of an electrocatalyst will significantly influence the performance of the catalyst, a balance must be made between better kinetics and durability and the associated cost implications. The better catalysts are more expensive within the fuel cells; however, an imbalance between the electricity generated and the amount of money invested in fuel cells will no longer make the utilization of fuel cells beneficial. Therefore, to create a better balance between cost and efficiency, this section guides how kinetics and durability are characterized in catalysts, as a more durable catalyst means less frequent replacements, thereby minimizing replacement costs. Similarly, a catalyst with better kinetics will increase the efficiency of the fuel cell, lowering the operational costs of the fuel cell. Thus, all aspects of the catalyst are intricately tied to the overall cost, knowing that only the most cost-effective catalyst will benefit our society. |
66251611418a5379b0514bea | 7 | Within the ORR, O 2 from the surrounding environment will first diffuse into the fuel cell's cathode and bind to the surface of cathode catalysts. In ORR, this bound molecular oxygen electrochemically onto the catalyst surfaces, where it lowers the energy barrier of the oxygen bond and breaks the O-O bond in the oxygen molecule. These molecules will then react with four protons and electrons to create oxygen ions. The oxygen ions then travel through the electrolyte to react with hydrogen gas at the anode that releases electricity and water as by-products, completing the ORR cycle. The efficiency of this ORR dictates the electric potential generated by the fuel cell, where higher efficiency corresponds to higher potentials. |
66251611418a5379b0514bea | 8 | ORR is a multielectron reaction characterized by several elementary steps involving different reaction intermediates, as shown in Table and. In acidic solutions, ORR occurs through two pathways: the direct four-electron transfer pathway from O 2 to H 2 O and a two-electron transfer pathway from O 2 to hydrogen peroxide (H 2 O 2 ). Alternately, there is also a two-electron pathway that transfers H 2 O 2 to water . Different reaction processes will result in a different potential under different reaction conditions. |
66251611418a5379b0514bea | 9 | There are two categories of catalysts used for ORR: noble-metal catalysts and non-noble metal catalysts, and the most known noble-metal catalysts used in this reaction are platinum catalysts and non-noble metal catalysts, such as carbon-based catalysts. A cathode catalyst is necessary to enhance the slow kinetics of ORRs in fuel cells . The catalyst stabilizes reaction intermediates, enabling smoother intermediate steps and faster reactions by reducing electron congestion. Furthermore, as the oxygen molecules enter the fuel cell's cathode, they will be absorbed by the catalyst surface, known as the electrochemical surface area (ESA). The ESA of the catalyst represents the number of sites on the catalyst where oxygen molecules are absorbed and combined to lower the energy barrier for the reaction; therefore, ESA is critical to initiate the ORR of the fuel cell. On ESA, materials like platinum will provide a lower-energy pathway that lowers the energy barrier of the oxygen bond and allows the breaking of the O-O bond in the oxygen molecule, making the ORR reaction more efficient (X. . Hence, implementing a catalyst will lower the overpotential of the reaction, enabling faster kinetics and enhancing the conversion performance between hydrogen and electricity. |
66251611418a5379b0514bea | 10 | In the current stage, platinum (Pt)-based catalysts are considered the most efficient catalyst for ORR in fuel cells for several reasons . First, highly active noble-metal catalysts such as Pt, Pd, Rh, and Au are capable of breaking the O-O bond. Because these catalysts have a d-band center that is well-aligned with the energy level associated with the oxygen bond, this alignment of energy levels facilitates the transfer of electrons that are active in ORR and weakens the oxygen bond of the oxygen molecules interacting with the catalyst surface. Also, this alignment helps reduce the activation energy required for breaking the O-O bond, which makes the overall reaction more efficient. Second, like ORR, these noble-metal catalysts also follow a 4-electron pathway, making them more suitable for ORR . Therefore, noble-metal catalysts are considered the most efficient catalysts to use within fuel cells. Because of the outstanding stability of the Pt-skin structure of the Pt-based catalysts, they stand out among these other metal catalysts. |
66251611418a5379b0514bea | 11 | As the catalyst plays an essential role in benefiting the kinetics of ORR, a greater surface area means a greater electrochemical surface area (ESA) of the catalyst-the surface of the catalyst where molecules' absorption occurs. This enhances the catalyst's impact on the kinetics of the reaction. A higher ESA offers more active sites for catalytic reactions. These reactions include weakening the oxygen bond, which aids in breaking the bond and subsequently facilitates electron transfer upon interaction with the catalyst's ESA . Therefore, increasing the catalyst's surface area will create more ESAs with more collisions between the oxygen molecules and the catalyst, increasing the rate of reactions. |
66251611418a5379b0514bea | 12 | Researchers have been trying to make the catalyst particles smaller to increase the surface area of the catalysts. For any given mass of catalysts, the smaller the particle, the greater the surface area. Thus, we could have catalyst particle sizes ranging from 2 nanometers to 100 micrometers. The more effort the researchers put into engineering scarce catalysts such as Pt-based catalysts, the smaller the catalyst particles will be . These notable efficiencies could be further visualized through a current-voltage polarization curve. |
66251611418a5379b0514bea | 13 | The kinetics of an electrochemical reaction involves analyzing the current-voltage polarization curves and the mass-specific activities of a fuel cell. The current-voltage polarization curve is a graphical representation that measures the kinetics of a reaction within the fuel cell with the equation P=IV (Power = Current x Voltage); the reaction's power will depend on the reaction's current and voltage. With the equilibrium potential being its maximum voltage, which occurs at 0 current, approximately 1.229 volts, the operating potential of the reaction will be lower due to inefficiencies caused by slow reaction kinetics. These inefficiencies can be characterized by the distance between the equilibrium potential and the actual current-voltage curve . Within the fuel cell, faster kinetics indicates a faster electrochemical reaction taking place at the electrode surfaces, specifically the oxygen reduction reaction. Fast kinetics will reduce the overpotential of the reaction, the additional voltage needed to drive a reaction at a certain rate. This reduction in overpotential minimizes energy wastage as heat, enhancing the efficiency of the conversion from hydrogen to electricity. Inversely, slow kinetics will increase the overpotential of the reaction, making the conversion between hydrogen and electricity less efficient . |
66251611418a5379b0514bea | 14 | When evaluating the catalyst's effectiveness in the reaction, scientists will select a specific voltage on the current-voltage polarization curve of individual catalysts and compare the current they generate on the polarization curve. This assessment, called mass-specific activity (MSA), focuses on the activity at a specific voltage. A suitable catalyst should consistently have the highest current at any given voltage, as a higher MSA means better kinetics (closer to equilibrium potential). In comparison, a lower MSA means worse kinetics (further from equilibrium potential)(G.-R. . An example could be drawn from Figure to understand the current-voltage polarization curve better. With the equilibrium potential being approximately 1.23 volts, this means the overpotential of the catalyst is the difference of voltage between the equilibrium potential (approximately 1.23 volts) and the actual voltage of the two catalysts marked by the blue and red curve. Further, upon interpretation of the diagram, it can be seen that the catalyst represented by the blue curve has higher MSA than the red curve, meaning that it has better kinetics. This can be examined by drawing a line horizontally from any voltage and measuring the current at which they are. The higher current at any given voltage means a higher MSA and, therefore, better kinetics, which is the catalyst of the blue curve. |
66251611418a5379b0514bea | 15 | Another basic measure that is also often taken to determine the efficiency of the catalysts is the half-wave potential (E 1/2 ) of a reaction, which differs by the type of catalyst for that reaction. E 1/2 describes the potential to reach half the limiting current density. The limiting value of a faradaic current -the electric current generated by the reduction or oxidation of some chemical substance at an electrode -is approached as the rate of charge transfer to an electrode increases. This can be achieved by either increasing the electric potential or decreasing the mass transfer to the electrode . The E 1/2 obtained from the linear sweep voltammetry curve through the method of cyclic voltammetry is a significant indicator of ORR performance. For E 1/2 , a higher E 1/2 corresponds to a lower overpotential to reach the specific current density, resulting in a higher ORR efficiency (Design of Efficient Bifunctional Oxygen Reduction/Evolution Electrocatalyst: Recent Advances and Perspectives -Huang -2017 -Advanced Energy Materials -Wiley Online Library, n.d.). |
66251611418a5379b0514bea | 16 | To monitor the electrochemical behavior of the catalyst and the overall reaction within the fuel cells, scientists often use strategies such as cyclic voltammetry (CV) to study the redox reactions and the electrochemical processes occurring at electrodes and the performance of the catalyst. First, CV works by linearly sweeping the working electrode's potential between two defined voltage limits. As the potential is swept, the current flowing through the working electrode will be measured. This current will then be recorded and created into a set of results that demonstrates the kinetics of the reaction and fuel cell. With the result gained, scientists can better understand the kinetics of ORR at the cathode as they examine the electrochemical behavior of catalysts at the electrode-electrolyte interface, where the measure of current is taking place during the process. By measuring the current response at different scan rates, scientists can tell the rate of charge transfer and other kinetic parameters that assess the catalyst's efficiency and the reaction's kinetics . Further, CV graphs can also be used to identify the change in the surface area of the catalysts as they undergo degradation. |
66251611418a5379b0514bea | 17 | The durability of an electrocatalyst describes how long the electrocatalyst can sustain within the whole reaction . The reason durability serves as one of the most important characteristics is that it is crucial for producing a more stable catalyst and minimizing the chances of replacing the electrocatalyst if worn out, reducing the cost and increasing the efficiency of the electrocatalyst. |
66251611418a5379b0514bea | 18 | Durability is mainly affected by the degradation mechanisms related to different catalysts -how and why the catalyst degrades. Degradation of the catalytic layer includes the degradation of Pt-based catalysts. Their degradation results in the reduction of the electrochemically active area, reducing the performance of the fuel cell and making it essential for researchers to keep in mind the mechanisms causing degradation. The degradation of a catalyst can be visualized on a polarization curve since the efficiency of the catalyst will be decreased along with the degradation of the catalyst. This means the kinetics of the catalyst will decrease while decreasing the current-voltage polarization curve by lowering the MSA of the catalyst; hence, the current-voltage polarization curve will curve down after degradation happens on a catalyst. There are mainly three main degradation mechanisms of Pt-based catalysts that are worth investigating: (1) Pt particle agglomeration and growth, (2) Pt loss and migration, and (3) Contamination of active sites caused by impurities. These degradation mechanisms will reduce the active area of the catalyst or increase contact resistance with catalyst support . The degradation mechanisms will be described in further detail in the subsequent sections. |
66251611418a5379b0514bea | 19 | First, the agglomeration and growth of Pt particles means that Pt nanoparticles often agglomerate into large particles due to their high surface energy relative to large particles. The strength of attraction between molecules depends on the substrate's surface energy, with higher surface energy meaning a stronger molecular attraction and low surface energy meaning a weaker attractive force. Therefore, as Pt nanoparticles have higher surface energy relative to large particles, they will attract each other and agglomerate into larger particles, which reduces the active area and the catalyst's efficiency. This process is called Ostwald ripening, a process through which a larger particle grows at the expense of nanoparticles. Ostwald ripening begins as smaller platinum particles dissolve through the Nafion ionomer phase. Smaller Pt nanoparticles will dissolve into the electrolyte and migrate to neighboring larger particles, releasing platinum ions into the surrounding environment. These platinum ions will then be redeposited onto larger platinum particles due to the relative molecular force between the larger and smaller particles, causing the particle size to grow. Over time, the particle agglomeration of particles will result in less surface area, contributing to the degradation of the platinum catalyst over long periods and impacting the stability and efficiency of the fuel cell ; the speed of this process depends on how much the catalyst is degrading. The time scale of this phenomenon is based on the time scale of the whole catalyst. |
66251611418a5379b0514bea | 20 | Second, Pt loss and migration happen almost simultaneously with the agglomeration of the Pt particle as when Ostwald ripening happens, where platinum particles undergo dissolution due to various factors with the part of the platinum ions (Pt 2+ ) that are not agglomerated being diffused within the fuel cell. These charged ions will be migrated through the electrolyte in the fuel cell, where they will be re-deposited. However, once the platinum ions reach the membrane, they will no longer contribute to the overall catalytic activity of the fuel cell . As the reaction within the membrane does not connect with the electrodes, the platinum ions may be immobilized within the membrane without participating in electrochemical reactions that occur at catalytic sites. Additionally, they will not contribute to proton transfer or reduction of oxygen. This affects the efficiency of electrochemical reactions such as ORR and the fuel cell. Over time, this will lead to a decline in catalytic activity and the durability of the catalyst. |
66251611418a5379b0514bea | 21 | Lastly, Pt contamination of active sites caused by impurities is also a critical degradation mechanism for Pt-based catalysts. The most common contaminant within Pt-based catalysts is CO, which is typically a by-product of reforming natural gas. A small amount of CO within the fuel cell inlet gas is absorbed into the Pt's catalytically active sites, blocking the passage of H 2 contact sites and reducing the catalyst's activity. But this also reveals that a noble-metal catalyst with good durability should also have a low CO absorption performance to maintain its long durability within the fuel cell . |
66251611418a5379b0514bea | 22 | The above degradation mechanisms are determined by the change in the electrochemically active surface area (ESA or ECA) and the activity in the oxygen reduction reaction (ORR) before and after stress testing. As a result, the material would be considered more stable if less of its electrochemical characteristics decreased after the stress testing. Stress testing mainly involves the stimulation of a catalyst over a period of time and examines how the catalyst will behave under real-world conditions . As mentioned earlier in the kinetics section, the kinetics of a catalyst will decrease over time, and this is measured mainly by looking at the MSA of the current-voltage polarization curve. For a scientist to determine the durability of a catalyst in a fuel cell, it would typically take several months or years, which is costly and time-consuming. However, the time and cost would be substantially reduced with the availability of stress testing, also known as accelerated testing. During the stress testing process in Figure , researchers cycled between high and low voltages; however, the time frames were reported as hours instead of cycles. Because the Pt-based catalyst dissolves at high voltage and platinum deposits at low voltage, cycling this process thousands of times will accelerate the degradation mechanism that happens on the platinum catalyst and visualize how fast the surface area of the platinum will decrease. As a result, researchers will comprehend the durability of the catalyst in days rather than months, reducing the cost and time to determine the durability of catalysts, illustrated in Figure , as the ECA of the platinum catalyst slowly deteriorates after hours have passed in AST. |
66251611418a5379b0514bea | 23 | is an MEA polarization curve of stress testing of a Pt-based catalyst that was demonstrated simultaneously with Figure . A clear contrast between the beginning of life (BOL) -the first sweep of the AST, and the end of life (EOL) -the last sweep of the AST could be examined. This shows the performance loss of a catalyst while undergoing accelerated stress testing within just 15 hours due to mass transport loss. Mass transport loss occurs when the rate of mass transport of a species to or from electrodes limits current production. Through AST, mass transport loss is emphasized as it accelerates the speed of the uneven distribution of reactants across the catalyst surface, exacerbating catalyst degradation . From the figure, mass transport loss is more significant in regions with high current density. However, with the performance loss being gradually greater, an apparent increase in mass transport loss down the AST can be observed. This mass transport loss continues to decrease from 3h to 6h; however, the performance of 9h in the high current density region is slightly better than that after 6h, indicating a possible reduced mass transport loss. After 15h, the performance deteriorates even more, having significant kinetic degradation on the catalyst. As such, a prominent trend can be observed as the MSA of the platinum catalyst slowly deteriorates as the time of the AST increases (X. . |
66251611418a5379b0514bea | 24 | Following the discussion of the criteria defining an effective catalyst, it is crucial to analyze the specific advantages of different groups of catalysts and understand the characteristics that make certain ones more valuable than others. This section will compare two primary catalyst groups used in fuel cells: Platinum Group Metals (PGM) and non-platinum Group Metals (non-PGM) catalysts, as all catalysts are built upon noble metal and non-noble metal catalysts. Therefore, examining these two groups could provide a better idea of which grouped catalysts are better, hence resulting in finding the more effective catalyst for society. This section will utilize the criteria established in the prior section to support the comparison between these groups, but with a significant emphasis on cost considerations in real-life applications. Throughout the comparison of the catalyst groups, data and statistics of individual catalysts will be referenced from research papers conducted by other researchers. This focused examination aims to provide a comprehensive understanding of the key catalysts within the PGM and non-PGM categories, considering their attributes and cost implications. |
66251611418a5379b0514bea | 25 | First, we start by comparing the efficiency of PGM and non-PGM catalysts. As they have significant differences in their electronic structure, there would be a range of differences in their efficiency. PGM catalysts such as platinum, palladium, and ruthenium demonstrate exceptional catalytic activity, especially in electrochemical reactions such as ORR. Such high catalytic activity of PGM catalysts is attributed to their unique electronic structure, primarily characterized by the d-band structure . During the ORR process, the d-orbitals will interact with oxygen electrons, facilitating subsequent electron transfer. The localized d-band center will be activated as oxygen intermediates react on the catalyst's active sites, dominating the bonding strength between the absorption species and the active sites, thereby enhancing the ORR activity. Most metal catalysts possess a d-band structure that gives them a d-state energy. A higher d-state energy for metal signifies greater catalytic reactivity, therefore explaining why PGM catalysts emerge as a more efficient catalyst for ORR . For instance, the half-wave potential (E 1/2 ) of a Pt/C catalyst during ORR under acidic conditions shows a result of 0.88V (vs.RHE) in comparison to Fe-based Single Atom Catalysts (SACs) Fe-N-C showing an E 1/2 of 0.784V in acidic electrolyte, the significant disparity highlighting the better performance of Pt/C catalysts, indicating the better in performance for PGM catalysts compared to basic non-PGM catalysts . Conversely, one of the key differences faced by non-noble metal catalysts is their lower catalytic activity compared to PGM catalysts. Because they do not exhibit the same level of catalytic activity, they could potentially impact the overall fuel cell efficiency. It has been proven that non-noble metal catalysts without specific engineering will be hard to exceed the high catalytic activity of PGM catalysts such as Pt/C catalysts. Even with stability and durability concerns, non-PGM catalysts will be disadvantaged compared to PGM catalysts due to their difference in electronic structure. |
66251611418a5379b0514bea | 26 | Moreover, in chemical research, reliable catalysts are essential, and the characteristics of PGM catalysts have established them as the most dependable choice. Subsequently, PGM catalysts have been more extensively engineered into nano-sized particles. Researchers aim to optimize these catalysts to their most efficient state to maximize reaction yields in their research. This characteristic helps them have a higher surface area, providing more active sites for catalytic reactions and, therefore, increasing the overall efficiency of the catalyst. For example, typical platinum nanoparticles can be made with sizes as small as 3nm ("Platinum Nanoparticle," 2023). However, a typical iron particle at the nanoscale is approximately 20 nm; this suggests that platinum catalysts are usually engineered into smaller-sized particles, which provides it with a more catalytic surface for reactions. It is likely that because PGM catalyst materials such as platinum are very scarce , scientists would spend more time engineering them into smaller particle sizes so that the effect of such high-activity catalysts could be maximized. On the other hand, non-PGM catalysts like iron, because they could be more easily found in nature, scientists do not spend as much effort and time to maximize their use; instead, they are simply engineered into relatively smaller sizes, resulting in larger sizes for typical iron nanoparticles. |
66251611418a5379b0514bea | 27 | Several factors contribute to this feature of PGM catalysts. Because noble metals exhibit relative inertness, they possess distinct resistance to corrosion and oxidation under various conditions, such as the different intermediates within ORR, including acidic, alkaline, and aprotic solutions. Such resistance is linked to their electronic structure and the arrangement of the outermost electrons. Further, noble metal catalysts also have a stronger metal-support interaction (MSI) with their supported material. A stronger MSI means more stabilized metal nanoparticles on the supported material; this prevents the occurrence of agglomeration of the metal particles, leaving a high dispersion of active sites . These all contribute to minimizing the process of Ostwald ripening within the fuel cells, which was one of the main factors that impacted the durability of a catalyst mentioned in the last section. |
66251611418a5379b0514bea | 28 | A point of concern with using non-PGM catalysts is that, not only do they degrade faster than PGM catalysts under harsh conditions, but they also degrade other parts of the fuel cells. This will reduce not only the durability of the catalyst but also the durability of the overall fuel cell. As mentioned before, PGM catalysts have a stronger MSI to minimize the process of Ostwald ripening; however, because non-PGM catalysts such as metal do not possess the strength of MSI that PGM catalysts such as platinum does, its process of Ostwald ripening would be more consequential. Once iron degrades, the Fe 2+ ions will spread throughout the fuel cell, degrading the fuel cell membrane and causing even more degradation beyond the catalyst's degradation, significantly affecting the durability of non-PGM catalysts . |
66251611418a5379b0514bea | 29 | The cost of noble metal catalysts is high, and due to the scarcity of noble metals, the overall cost of fuel cells is even higher (Table ). Indeed, the natural scarcity of noble metals (NMs) largely determines their high cost. Common non-noble metal catalysts, such as iron, cobalt, and nickel, are usually more abundant and less expensive than noble metals. Table shows a clear difference between the cost of noble metals and non-noble metals. For example, the price of iron per gram is $0.0001274, and the price of platinum is $28.99, which is approximately 2.273 10 7 % of iron. Though the price of non-noble metal catalysts varies, it is × still significantly cheaper than noble metal catalysts. Although these prices are subject to change, they can provide a good idea of the difference in price between noble metal catalysts and non-noble metal catalysts, which helps explain why certain industries decide to use non-noble metal catalysts as their source of a catalyst even though they lack the same level of efficiency. |
66251611418a5379b0514bea | 30 | Their abundance is another aspect of non-noble metal catalysts that makes them more cost-effective. In 2022, the total worldwide reserves of platinum group metal were estimated to be approximately 70,000 metric tons; meanwhile, the total reserves of crude iron ore worldwide were estimated to be approximately 180 billion metric tons . From this, we can also see the vast difference in scarcity of the catalyst metal groups. Because noble metal catalysts are precious metals that require much effort, time, and cost to unearth and develop, replacing them with non-noble metal catalysts will help reduce the industry's dependency on scarce and costly precious metals, further maximizing the benefit of a limited budget. However, though non-noble metal catalysts are significantly cheaper than noble metal catalysts, performance-wise, there are still many challenges regarding their catalytic activity and stability, unlike noble metal catalysts. |
66251611418a5379b0514bea | 31 | Although the development of catalysts' kinetics and durability has been improving, and more researchers are starting to pay more attention to how these characteristics could be further enhanced, there are still expectations regarding developing a catalyst with high activity, promising durability, and cost-effectiveness. Solutions to these challenges have yet to be discovered, and researchers should address these problems to keep moving forward. |
66251611418a5379b0514bea | 32 | Throughout this paper, all graphical figures regarding the kinetics of the catalysts are represented by MEA graphs. More significantly, it has brought to attention the insufficient data on the MEA tests and the lack of reliable, scalable synthesis methods that measure different catalysts. Still, more efforts are devoted to the performance of catalysts at an RDE level . However, the importance of MEA tests cannot be ignored. First, mass transfer resistance is largely eliminated in RDE measurements, but because it is one of the most critical concepts that impact the fuel cell's kinetics, it cannot be overlooked. Second, MEA tests have a triple-access (gas-liquid-solid) requirement, allowing them to minimize the utilization efficiency of the catalyst at active sites. Third, the concentration of O 2 in RDE tests is extremely low but high for MEA tests. Therefore, RDE tests have seen limitations in challenging thermal, gas, and water management issues. Lastly, MEA allows a test on a more complex fuel cell system due to its specialized equipment than that of RDE . Therefore, more investigations were focused on MEA testing so that researchers could gain more insightful results for catalysts operating through fuel cells. |
66251611418a5379b0514bea | 33 | Due to the characteristics of MEA tests, researchers have proposed that catalysts' cost-effectiveness and high durability must reach an MEA level to be deployed into technologies within the marketplace for a more reliable measure. Therefore, to enhance the library of MEA results for different catalysts, more emphasis should be placed on the amount of MEA tests reported for ORR electrocatalysts. In addition, the durability of MEA should also be enhanced since the operating temperature of MEA (60 °C-80 °C) is higher than those in RDE tests. Thus, improved investigations on MEA services will also be significant to their practical applications. |
66251611418a5379b0514bea | 34 | There are various types of support materials for catalysts, such as the various kinds of activated carbon, alumina, and silica, all of which show differing improvements in the catalyst's performance and durability. However, because ORR mainly focuses on carbon-supported catalysts, increasing the carbon support material for the catalyst is the most critical step. Because of the instability of carbon support materials for catalysts due to its porous structure that interacts with active sites and creates a blockage that facilitates the degradation of catalysts, an investigation into non-carbon support has been emphasized to achieve higher stability in MEA than other carbon-supported catalysts. Therefore, an attempt to enhance the optimal loadings for ORR is developing. Compared to conventional carbon-supported nanoparticles on Pt/C catalysts that were already established, researchers are now focusing on the development of platinum-alloy nanostructure thin film (NSTF) catalysts for ORR. These catalysts are coated around the fuel cell, within electrodes and the electrolyte membrane. These PtNi alloys have proven to be more efficient catalysts than Pt/C due to their structure with no carbon support. Thus, researchers are focusing on the development of PtNi NSTF by determining the proper loading of particles on carbon needed for the catalysts, as there is an increased surface area with increased loading. As a result of the current progression on PtNi NSTF, it shows that PtNi NSTF exhibits a more specific and mass activity compared to Pt/C catalysts, proving that NSTF catalysts present a very promising system as a cathode material in PEM fuel cells . However, more research should focus on maintaining the same kinetics as typical Pt/C catalysts while maintaining the high stability NSTF-structured catalysts provide due to no carbon support. |
66251611418a5379b0514bea | 35 | To conclude, in this paper, we evaluated the characteristics of catalysts, including kinetics and durability. They are the key characteristics of a catalyst that dictate the viability to be used by determining the efficiency and lifetime of a catalyst. Because the main point of assessment of a catalyst is to observe its kinetics and durability, this paper has focused on finding the most cost-effective catalyst. This catalyst has high kinetics relative to high durability. By comparing PGM and non-PGM catalysts, the result can be drawn that if a researcher decides to use an efficient catalyst, PGM catalysts are more efficient. However, non-PGM catalysts are more cost-effective due to their ability to be engineered into different catalysts, such as the engineering of Fe-based catalysts to Fe-N-C, which significantly increases performance. Still, non-PGM catalysts are also cheaper due to their high abundance of materials in nature from Table . However, with the development of current catalysts, many problems are yet to be addressed for future advancements. To enhance the kinetics of a catalyst, research has been done regarding the composition, structure, and location of the active sites on the catalyst where interaction between the molecules and the catalyst is happening. At the same time, to maintain the stability of the catalysts, a glance at the nanostructure thin film for catalysts is also taking place. This is beneficial to the catalysts as no carbon is required; it will increase the catalyst's stability and surface area, avoiding instabilities caused by carbon-supported structure. However, more research will be needed to employ these characteristics further. |
66eb9a1712ff75c3a1c5a2be | 0 | The story of Clar's goblet dates back to 1972, when Erich Clar envisioned this mysterious hydrocarbon . It contains an equal number of masked and unmasked carbons, suggesting that all πelectrons should be spin-paired (Fig. ). However, it cannot be represented by a Kekulé structure without leaving unpaired electrons. In its resonance structure, two unpaired electrons are always present. According to classic valence bond theory, it is challenging to reconcile how a diradical molecule can have all its π-electrons spin-paired. The nullity (η) is equal to the difference between the numbers of the maximum set of non-adjacent sites and the remaining sites. The dash lines in the disjoint diradical structures indicate that the two radicals can only be distributed on their respective halves in all resonance structures. c, d, DFT calculated SOMO (α), SOMO (β) (c), spin density distribution (d) of Clar's goblet at um06-2x/6-31g(d,p) level of theory. Herein, SOMOs are actually the same as NMBOs. R and Ar represent substituent groups. |
66eb9a1712ff75c3a1c5a2be | 1 | Clar's goblet, with delocalized unpaired electrons in its π-conjugated sp -carbon skeleton, structurally belongs to the category of graphene fragment radicals . These graphenic radicals possess distinct optical, electronic and magnetic properties, making them highly promising for applications in materials science . Two terms, "disjoint" and "non-disjoint" have been used to describe delocalized diradicals . In non-disjoint diradicals, such as Clar's hydrocarbon (triangluene) and trimethylenemethane (TMM) , two radicals can interchange positions through resonance structures (Fig. ). From the perspective of molecular orbital, two degenerate nonbonding molecular orbitals (NBMOs) spatially overlap, leading electrons to prefer unpaired spins to minimize repulsion, in accordance with Hund's rule . This triplet ground state of non-disjoint diradicals can be readily determined by sublattice imbalance by using Lieb's theorem or Ovchinnikov's rule . On the contrary, in Clar's goblet or tetramethyleneethane (TME) , two degenerate NBMOs are disjoint (i.e., have no atoms in common. Fig. ). The two radicals are only distributed in separate halves, based on resonance structure analysis (Fig. ). Consequently, the two electrons have no spatial overlap and do not need to obey Hund's rule. Furthermore, they tend to have opposite spins, although the mechanism is still unclear and may involve dynamic spin polarization . The calculated spin density distribution clearly illustrates the polarized distribution of the up and down spins in Clar's goblet (Fig. ). This unusual spin pairing between the two spatially separated electrons in Clar's goblet represents a form of spin entanglement . Although this local spin entanglement is far from practical application, pursuing demonstration of this property remains crucial for understanding quantum phenomena at the molecular level. |
66eb9a1712ff75c3a1c5a2be | 2 | Therefore, Clar's goblet has been a long-sought synthetic target for decades, motivating extensive advancements in synthetic chemistry , molecular structure and magnetism theory . The primary challenge in the chemical synthesis of Clar's goblet is to find a feasible way to construct its topologically frustrated sp 2 -C network with atomic precision, particularly the C-C bonds within the disjoint region. In addition, the high reactivity of radicals often leads to undesired decomposition. In 2020, Fasel and Feng's teams reported the on-surface synthesis of Clar's goblet and in situ investigation . The properties obtained by on-surface chemistry are unavoidably influenced by the interactions with the gold surface. Moreover, many key properties of Clar's goblet are difficult to explore under the stringent conditions (e.g., on a metal surface, under ultrahigh vacuum) of onsurface chemistry, and thus are still hidden. Overall, the precise synthesis of Clar's goblet by wet chemistry and revealing its unusual physical properties remain a long-standing and highly important challenge. In this Article, we report on the solution-phase synthesis of Clar's goblet and deciphering its intriguing physical properties, especially the spin-entanglement between the two confined electrons. The key C-C bonds in the disjoint region are constructed by an intermolecular radicalradical coupling approach. Moreover, introducing bulky substituents at the periphery of Clar's goblet enhances its kinetical inertness, making it easier to handle during isolation and investigation. Temperature-dependent magnetic studies reveal an antiferromagnetic (AFM) coupling between the two spatially segregated spins in Clar's goblet, with an average spin-spin distance of 8.7 Å and coupling strength of -0.29 kcal/mol. This unique spin spatial segregation is further demonstrated by theoretical and experimental studies of the radical cation state of Clar's goblet. This work experimentally demonstrated what has been long expected theoretically: the two individual spins in the hydrocarbon -Clar's goblet are entangled in the ground state. |
66eb9a1712ff75c3a1c5a2be | 3 | Based on retrosynthetic analysis, the skeleton of Clar's goblet can be viewed as the fusion of two olympicenyl radicals , with two new C-C bonds formed in the disjoint region. Therefore, a straightforward synthetic strategy is to create these two connecting bonds between olympicenyl radicals or derivatives through intermolecular radical-radical coupling, which was also the idea behind Clar's attempt 1 . However, olympicenyl radicals or derivatives have zero spin density on these two bottom carbons, leading to the failure of radical-radical coupling. We thus designed a modified diradical precursor, making this strategy feasible (Fig. ). Dihydroxyl naphthanthrone (1) was the starting compound, which was further transformed into diradical precursor 2 under the oxidative condition developed by Zeng's group . The diradical precursor 2 has considerable spin density on the desired carbons, as suggested by its carbon-centered radical resonance structure. Intermolecular radical-radical coupling between two molecules of 2 and subsequent simultaneous dehydrogenation driven by aromaticity, led to the formation of the proposed intermediate 3. This dimerized intermediate 3 was further reduced into tetrahydroxyl 4 using Na2S2O4. Due to their poor solubilities, intermediates 3 and 4 were directly used in the next steps without characterization. The hydroxyl groups in 4 were converted to n-butoxy ( n BuO) groups in 5 by reacting with excess K2CO3 and 1bromobutane. The key precursor 5 with the target carbon skeleton can be fully characterized including single-crystal X-ray diffraction analysis (Supplementary Fig. ). In the final step, treatment of diketone 5 with excess lithium (triisopropylsilyl)acetylide followed by reductive dehydroxylation with H2SnCl4 resulted in the formation of target Clar's goblet -cg-1. However, the high reactivity of cg-1, with an estimated half-life time (τ1/2) of 8 hours (Supplementary Fig. ), limits further reliable characterization. We thus revised the molecular design of Clar's goblet to cg-2 by introducing bulky aryl substituents at the periphery to increase its kinetical inertness. Ir-catalyzed borylation reaction of compound 5 produced compound 6 with four pinacol boronate (Bpin) groups. Subsequent Suzuki coupling reaction allowed us to install aryl substituents at the peripheral positions, yielding the diketone precursor 7. This precursor underwent the same nucleophilic addition/reductive dehydroxylation steps to generate cg-2. The sterically hindered substituents successfully made cg-2 more inert, with a substantially extended τ1/2 of 6 days (Supplementary Fig. ), allowing for in-depth investigation. Single crystals of cg-2 were obtained by slow vapor diffusion of acetonitrile into its chlorobenzene solution in a glovebox filled with nitrogen gas. Crystallographic analysis unambiguously demonstrated the structure. The skeleton of Clar's goblet is surrounded by peripheral bulky substituents, with a head-to-tail distance of 11.3 Å (Fig. ). The presence of middle n BuO chains induces a twist in the backbone with an angle of about 37 o . Moreover, this twisted backbone imparts intrinsic chirality to cg-2, resulting in Pand Menantiomers. The packing structure shows the racemic composition (Fig. ). Interestingly, two identical enantiomers form a pseudo-molecular box, encapsulating two chlorobenzene molecules inside. The formation of this super-structure is driven by hydrogen bonding, π•••π interactions, and CH•••π interactions (Supplementary Fig. ). The distance between the top and bottom cg-2 in the super-structure is approximately 6.8 Å, indicating weak intermolecular spin-spin interactions. |
66eb9a1712ff75c3a1c5a2be | 4 | As explained in the introduction, the two electrons in Clar's goblet are spatially segregated and cannot be paired by forming a chemical bond. The long-standing question is: how do these two individual electrons interact with each other? In principle, a two-spin system can exist in a magnetically active triplet state when the spins are aligned, or in a magnetically silent singlet state when the spins are opposite. Alternatively, if spin-spin interaction is very weak, the system might behave like an almost independent doublet state. Thus, the key to addressing this question is to distinguish the spin ground state of Clar's goblet. Continuous-wave electron paramagnetic resonance (cw-EPR) measurement was first used to probe the interaction between the two spins. At 170 K, the spectrum of cg-2 in the frozen toluene-d8 solution shows resonance signals composed of two components: zero-field splitting (ZFS) from a triplet (S = 1) component and central peaks from a doublet (S = 1/2) component, which can be readily separated by simulation (Fig. , simulation parameters are attached in the Supplementary Information). A forbidden transition (|∆𝑚𝑚 s | = 2 ) signal is also observed (ms refers to magnetic quantum number). The characteristic ZFS and half-field absorption provide solid evidence for the existence of triplet state in Clar's goblet and rule out the possibility that the two spins are weakly interacted. Furthermore, ZFS parameters of |𝐷𝐷| = 114.5 MHz and |𝐸𝐸| = 15.7 MHz are obtained from simulations. Using the equation D = 1.39 × 10 4 (𝑔𝑔 𝑟𝑟 3 ⁄ ) where unpaired electrons are considered as point-dipoles , the average spin-spin distance in cg-2 is determined as 8.7 Å, further supporting the spatial segregation of the two spins (Fig. ). Additionally, the doublet component in cw-EPR spectrum is proposed to be monoradical impurities with structures similar to cg-2. The observation of doublet signals in the EPR measurement of diradicals is very common, and the origin of the monoradical impurities includes the interactions or reactions with oxygen, solvents, light or intermolecular interactions . |
66eb9a1712ff75c3a1c5a2be | 5 | To further identify the spin ground state of Clar's goblet, pulse-EPR was used because of its high signal reliability at low temperatures. The echo-detected field-swept (EDFS) spectra of cg-2 were recorded at 10 K, 50 K, and 100 K, respectively. These EDFS spectra can be accurately deconvoluted into one doublet and one triplet component with relative weights by simulations (Fig. , simulation parameters are attached in the Supplementary Information). As the temperature increases from 10 K to 100 K, the triplet component substantially increases from 0.36 to 1.70 relative to the doublet component (its weight sets to 1). This thermal populated property clearly demonstrates that the triplet state is the excited state of cg-2. To further verify the spin states of the proposed doublet and triplet components in the EDFS spectra, echo-detected nutation experiments were performed using the pulse sequency tp-T-π/2-τ-π-τ-echo (T refers to waiting time). In nutation experiments, a microwave nutation pulse with duration tp, is applied to drive the spin system in the sample into a superposition of its two ms levels of a given resonance. With increasing nutation pulse length, the spin system is cycled through all arbitrary superpositions of the two ms levels, resulting in the Rabi oscillations in the detected signal. As shown in Fig. , the Rabi oscillations were recorded at the magnetic field of 3401 G, temperature of 10 K and 100 K, respectively. After fast Fourier transformation, the oscillation data are transformed into two separate frequencies corresponding to the doublet and triplet specie. The spin states of these two species are confirmed because the ratio of the Rabi frequency for the assigned triplet species (𝛺𝛺 𝑚𝑚 ±1 ↔𝑚𝑚 0 ) to that of the doublet species |
66eb9a1712ff75c3a1c5a2be | 6 | ) is very close to the theoretical value of √2 (Eq 1 in the Supplementary Information, and Supplementary Fig. ). In addition, the intensity of the triplet species relative to the doublet species noticeably increases at 100 K compared to at 10 K (Fig. ). This observation further demonstrates that the triplet state is the thermally populated excited state of cg-2. Next, superconducting quantum interference device (SQUID) measurement was performed on freshly prepared crystalline powder of cg-2 to quantify the energy gap (∆ES-T) between the singlet ground state and the triplet excited state, more fundamentally, the coupling strength between the two spins in cg-2. The product value of the molar magnetic susceptibility and temperature (χm•T) rapidly increase from 25 K, and gradually approaches saturation after 200 K (Fig. ). This is the characteristic feature of thermally populated transition from the magnetically silent singlet state to the magnetically active triplet state. By fitting the χm•T -T curve with Bleaney-Bowers equation (Eq 3 in the Supplementary Information, the contribution from doublet impurities is treated as a constant in the fitting), ∆ES-T was determined to be -0.29 kcal/mol, which is consistent with DFTcalculated value -0.35 kcal/mol. This ∆ES-T value is considerably smaller than the energy gaps for many open-shell singlet diradicaloid systems , suggesting the coupling strength between the spatially segregated spins in cg-2 is much weaker than bonding interaction. In addition, χm•T value rapidly decreases below 25 K, indicating that the influence of intermolecular AFM coupling can no longer be ignored at extremely low temperatures. At present, the interaction between the two spatially segregated electrons in cg-2 has been clearly revealed: the two electrons can feel the existence of each other and keep their spins paired in the ground state, suggesting a spin entanglement in Clar's goblet. The singlet ground state can be thermally excited to triplet state. Theoretically calculated spin density distribution (Supplementary Fig. ) of the singlet state of cg-2 retains the same spin polarization feature as the pristine Clar's goblet, with the up and down spins separately localized on the two halves. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.