id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
677ec32dfa469535b936c4d7 | 21 | The new proposed NCBs were among the most stable tested, with three structures surpassing stability of those currently in our library: R3sub-II (-117.1 kcal/mol, -490.1 kJ/mol), R2sub-IV (-120.0 kcal/mol, -502.3 kJ/mol), and R1sub-IV (-117.3 kcal/mol, -490.9 kJ/mol) (Figure ). The increased stability of these NCBs relative to the remainder of the library is promising since it shows that we can use trends in existing data to suggest and successfully design new NCBs with enhanced stability. Furthermore, the higher stability seen with the new substituents at each position supports our hypotheses that NCBs with aryl R1 substituents, electron-withdrawing R2 substituents, and bulky and/or electron-withdrawing R3 substituents are likely to be stable structures. With this new information about how substituent identity and properties at each position impact stability, we can make informed decisions to eventually suggest novel NCBs which will be more resistant to decomposition via C5-C6 saturation. |
677ec32dfa469535b936c4d7 | 22 | In addition to high thermodynamic stability, these out-of-sample NCBs also show high kinetic stability and correlation between ∆G and ∆G ‡ , as was determined with NCBs in our library (R 2 = 0.92) (Figure , Table ). We also used these novel structures to test our model's performance with out-of-sample structures. All except three of these out-of-sample NCBs showed accurate predictions (< 3 kcal/mol, 12.552 kJ/mol, prediction error), with each poor prediction occurring for structures with out-of-sample R2 groups. The accurate predictions on unseen substructures gives confidence in the predictive ability of our model. Additionally, the model tends to predict stabilities higher than the benchmark DFT calculations, especially for the three out-of-sample structures errors in stability prediction for these out-of-sample structures assume higher stability than calculated using DFT, a trend present for all except one structure. Errors tending this way reduce the chance that synthetic chemists disregard highly stable structures due to our model when choosing stability-ranked NCB candidates to characterize experimentally. |
677ec32dfa469535b936c4d7 | 23 | However, there were also some structures which showed low model accuracy (error > 3 kcal/mol, 12.552 kJ/mol), likely due to the type of descriptors utilized to make predictions and absence of sufficient structure representations in the training data. When analyzing these structures with poor model accuracy, R2sub-I (error of 13.8 kcal/mol, 57.7 kJ/mol), R2sub-II (error of 8.0 kcal/mol, 33.5 kJ/mol), and R2sub-IV (error of 10.2 kcal/mol, 42.6 kJ/mol), we found that each NCB had at least three descriptors used in the model which were outside the range seen by the initial library. |
677ec32dfa469535b936c4d7 | 24 | For example, R2sub-I and R2sub-II each had f(-) values at the R2 substituent (-0.09 and -0.03, respectively) that were outside the observed range (-0.02 to 0.02), likely causing poor stability predictions in these structures. The other NCB which had a poor prediction was R2sub-IV. Our model likely failed in making an accurate prediction of stability for R2sub-IV because all other structures within our library have a neutral carbon atom at R2, so the electronic descriptors at this atom and C3 (directly bound to R2) are beyond the scope of our trained model. |
677ec32dfa469535b936c4d7 | 25 | While the expansion of our NCB space with these out-of-sample structures showed that our model carries predictive power even with substrates it was not trained with, the model can be adapted over time through expansion of the training data, which would be recommended if a substituent chemically distinct from our current library is found to be relevant in NCB design. Nevertheless, our model tends to predict stabilities higher than the benchmark DFT calculations, especially for the three high-error out-of-sample structures. This error trend reduces the chance that a highly stable structure would fail to advance to an experimental stage in a high-throughput NCB discovery pipeline, and even though some "false positives" might appear in the top ranks, our results show that is unlikely. For example, if an experimental group chose to synthesize and characterize the top 10% of stability-ranked structures, including the out-of-sample designs, this would result in two of the fifteen candidates (rounded up), with DFT calculated ∆G values that would have excluded them from the top 10%, but that still place them in the top 20%. |
677ec32dfa469535b936c4d7 | 26 | This work introduces an NCB library which covers more chemical space than has been studied before, generating a library of 132 unique NCBs to expand the candidate space and allow us to examine substituent stability effects in more detail than was previously possible. With the limited number of experimental structures with decomposition data available, we also isolated the mechanism of decomposition of nicotinamide NCBs in a phosphate buffer, suggesting decomposition occurs in a stepwise manner where the first decomposition event (TS I) is highest in energy. Furthermore, there is a strong correlation between the thermodynamic and kinetic data, so we simplified our model of stability and used ∆G for the proton transfer step of decomposition |
677805156dde43c908304eb8 | 0 | Within the field of porous materials metal-organic frameworks (MOFs) rank among the most porous yet versatile compounds. This is due to their unique modular construction based on organic linkers and inorganic metal nodes and the effect of the often particularly large inner surface making them highly advanced adsorbents. Gas adsorption in MOFs can be accompanied by unique and often counterintuitive structural changes within the porous framework due to a complex interplay of inter-and intramolecular interactions of both, the adsorbate as well as the adsorbent. This flexibility also sets them apart from other classes of porous materials such as porous carbons or zeolites and lead to a variety of analytic methods providing deep insights in the underlying mechanisms on a molecular scale. Besides these molecular level construction and interactions, crystal size has shown to have an impact on the adsorption behaviour of MOFs as well. In many cases, the decrease in crystal size and thereby the decreased ratio of inner surface compared to the crystals outer surface leads to a stabilization of a metastable open pore phase due to a lack of nucleation sites required for the phase transition. This influences the thermodynamics and kinetics of the physisorption process and the possibly underlying flexible transitions. Various groups have found that smaller crystal size in ZIF-8 increases the gate opening penalty and therefore overall decreases the flexibility of this system. On the other hand larger crystals oftentimes show more internal strain due to a larger number of differently oriented domains within the crystal. This is also reflected by computational studies which suggest that the flexible breathing transition of MIL-53(Al) can occur via a layer-by-layer mechanism as well as one involving multiple discrete nucleation points, depending also on external pressure. This leads to altered kinetics of the phase transition as in MIL-53(Al) a slower breathing transition with increasing crystal size but also a lack of a phase transition in submicron sized crystals can be observed. The altered kinetics that come with changes in crystal size have e.g. been utilized to boost the dynamic separation of ethylene and ethane by increasing the crystal size of a ZnAtzPO4 framework. In general, crystal size and morphology needs to be acknowledged as a factor that can influence the physical properties of MOF materials but can also be a tool to design new MOF materials for specific applications. While adsorbing and desorbing gas molecules, MOFs are exposed to an adsorption/desorption stress, which may cause reversible or irreversible framework deformation and may lead to changing its crystallites surface texture, causing breaks or defects, replacing solvent molecules from the synthesis or accumulates in the pore after repeated exposure. This is often accompanied by an decrease in crystal size due to fracture of larger crystals until a certain finite value is reached, an increase in gate opening pressure and a less steep opening slope, decrease in gate closing pressure and a smaller total pore volume. These structural changes can be at least partially understood as a materials response to an adsorption-induced stress exerted from a fluid adsorbate onto its solid surface. Such phenomenon can be termed "adsorption milling". Adsorption stress is also observed in other porous materials such as porous carbons. Gor and Neimark modelled the adsorption-induced deformations in the mesoporous silica using Derjaguin-Broekhoff-de Boer theory for capillary condensation and Quenched Solid Density Functional Theory indicating advantages and disadvantages of each for them for materials with particular pore size. Coudert and Neimark developed the stress-based model for breathing crystalline porous solids. In terms of experimental methodology, the mechanical stress is often evaluated by mercury intrusion , in situ high-pressure Powder X-ray Diffraction (PXRD) experiments or in silico prediction . However, these techniques have certain limitations e.g. in case of rigid mesoporous frameworks, which are often quite sensitive to ambient conditions, irreversible transitions are observed at relatively low pressures. In such cases, more precise and sensitive technique should be developed. g -1 . A characteristic feature of this material is the utilization of a carbazole based cuboctahedral 12-conneting metal-organic polyhedron (MOP) also present in other DUT materials acting as a well-defined building block for the construction of MOFs with high porosity and a high concentration of open metal sites. Additional open metal sites are introduced by using of tritopic linker H3CBCDC (9-(4'-carboxy-[1,1'-biphenyl]-4-yl)-9H-carbazole-3,6-dicarboxylic acid) in the construction of the framework, in which carboxylate groups at the biphenyl moiety form an additional copper(II) paddle-wheel connecting four MOPs yielding a (4,12)-connected network with a four, twelve connected (ftw) topology (see Figure ). Previously we found DUT-76(Cu) to have an exceptional capacity for high-pressure ethylene storage at 298 K. |
677805156dde43c908304eb8 | 1 | In the following we use DUT-76(Cu) as a model system and propose an experimental methodology for accessing the mechanical stability of the mesoporous MOFs by applying adsorption/desorption stress upon physisorption of hydrocarbons. A physisorption of selected alkanes and alkenes at their standard boiling points in combination with in situ and ex situ PXRD analysis, total scattering and SEM imaging is used for fine tuning of adsorption/desorption stress leading to guest-dependent irreversible contraction of the structure. |
677805156dde43c908304eb8 | 2 | The procedure for the desolvation of DUT-76(Cu) showed to be critical for the materials stability and physisorption behaviour. Specifically, a fast gas release for flushing as well as a release of the excess CO2 from the autoclave and a fast heating to generate supercritical CO2 have shown to lead to a degradation of the material. For this reason, we want to give detailed information on desolvation and thermal activation. Supercritical drying of the MOF powder after synthesis was performed in a Jumbo Critical Point Dryer 13200J AB (SPI Supplies). The MOF suspended in dry acetone was therefore transferred into the glass filters of the dryer. The autoclave was then filled with liquid CO2 at 290 K and 7.5 MPa. By opening the valve at the bottom of the autoclave the remaining acetone was removed while fresh liquid CO2 was refilled from the top of the autoclave. This procedure was repeated 8 times per day over two days until no acetone traces were observed in the dry ice at the outlet. In the following step, the valve from the gas cylinder into the dryer was closed and half the liquid CO2 in the autoclave was released. Then all valves of the autoclave were closed, and the temperature increased to 313 K to convert the CO2 into supercritical state. The pressure in the autoclave was maintained at 9 MPa during heat up. After equilibration of the system at 313 K overnight, the supercritical CO2 was released through the bottom valve over 12 hours. |
677805156dde43c908304eb8 | 3 | The autoclave was opened as the pressure was reduced to ambient pressure and the dried material directly brought into an argon filled glovebox. (see Figure ). Each further handling was carried out exclusively under inert atmosphere. Subsequently for supercritical drying the material was additionally treated by heating up to 453 K in dynamic vacuum for 24 hours to remove remaining guest molecules from the pores. |
677805156dde43c908304eb8 | 4 | Ex situ powder X-ray diffraction (PXRD) patterns were collected in transmission geometry on a STOE STADI P diffractometer, equipped with a line focus Cu X-ray tube, operated at 40 kV/30 mA, and focusing Ge (111) monochromator (λ = 0.15405 nm) and MYTHEN (DECTRIS) detector. A scan speed of 120 s/step and a detector step size of 2θ = 6° was used in the measurements. |
677805156dde43c908304eb8 | 5 | Low pressure (p < 110 kPa) volumetric adsorption experiments were carried out on a BELSORPmax instrument by Microtrac MRB and the measuring routine of BELSORP-max control software was used. The dead volume was routinely determined using helium at 298 K before each measurement as well as in between each adsorption/desorption cycle of the cycling experiments. |
677805156dde43c908304eb8 | 6 | A closed cycle helium cryostat was used for cooling. The cryostat DE-202AG was operated by a temperature controller LS-336 (LAKE SHORE) and the heat produced by the cryostat was removed from the system by a water-cooled helium compressor ARS-2HW. The sample was placed in a custom-made cell consisting of a 3 cm long rod-shaped copper cell of 1 cm diameter, sealed from the exterior with a copper dome and insulated by dynamic vacuum (p < 10 -4 kPa), and connected to the BELSORP-max adsorption instrument with a 1.5 mm copper capillary. |
677805156dde43c908304eb8 | 7 | In situ PXRD experiments on DUT-76(Cu) in parallel to adsorption and desorption of hydrocarbons at their boiling points were conducted according to previously published setup and procedure. A customized setup based on laboratory powder X-ray diffractometer Empyrean-2 (PANALYTICAL GmbH), equipped with a closed-cycle helium cryostat (ARS DE-102) and home-built X-ray transparent adsorption cell, connected to volumetric adsorption instrument BELSORP-max (Microtrac MRB) was used. The TTL trigger was used to establish the communication between BELSORP-max and Empyrean and ensure the measurement of the adsorption isotherm and PXRD pattern data collection in a fully automated mode in the predefined points of the isotherm. The parallel linear Cu Kα1 beam, obtained by using a hybrid 2xGe(220) monochromator, 4 mm mask, and primary divergence and secondary antiscatter slits with 1/4° opening, were used for data collection. A Pixcel-3D detector in 1D scanning mode (255 active channels) was used. The diffraction experiments were performed using ω-2θ scans in transmission geometry in the range of 2θ = 3-70°. SEM images of DUT-76(Cu) were taken with secondary electrons in a HITACHI SU8020 microscope using 1.0 kV acceleration voltage and 15.2-16.4 mm working distance. Sample preparation was performed in an argon atmosphere on an inert sample holder. As the sample showed degradation when being prepared directly on a sticky carbon pad powdered samples were prepared on a dry silicon wafer attached to a sample holder with a sticky carbon pad as a conductor. |
677805156dde43c908304eb8 | 8 | Residual powder was blown away with argon and a magnet to prevent contaminations on the device. The samples were then transferred into the device in an argon atmosphere on an inert sample holder. For crystal size determination at least ten images of each sample were analyzed with the ImageJ Software package. |
677805156dde43c908304eb8 | 9 | Thermal analysis (TGA) was carried out in Argon using a NETZSCH STA 409 thermal analyzer at a heating rate of 5 K•min -1 . The air sensitive MOF sample was prepared in an Ar-filled glovebox and inserted in the instrument with little exposure to ambient conditions. Total X-ray scattering data were collected at the Powder Diffraction and Total Scattering Beamline, P02.1, PETRA III, Deutsches Elektronen-Synchrotron (DESY). The X-ray wavelength was λ = 0.0207361 nm and the data were collected on a Varex XRD 4343CT (150×150 µm2 pixel size, 2880 x 2880 pixel area, CsI scintillator directly deposited on amorphous Si photodiodes). For calibration of the experimental geometry, LaB6 SRM 660c NIST powder was prepared in a soda lime glass capillary with an internal diameter of 0.8 mm. The pyFAI software was used for both calibration of the experimental geometry and for azimuthal integration. The sample-to-detector distance (SDD) was calibrated to be 300.613 mm. The area detector was placed such that the beam center was at the lower right corner of the area detector, when looking downstream from the sample position, resulting in quarter Debye-Scherrer rings collected on the detector. For the LaB6 SRM 660c NIST powder, total scattering data were collected for 300 s. The DUT-76(Cu) powder was prepared in a borosilicate capillary with an internal diameter of 0.5 mm and total scattering data were collected for 1500 s. The background contribution to the total scattering data were collected for 1500 s for an empty borosilicate capillary with an internal diameter of 0.5 mm. To take X-ray intensity fluctuations during data acquisition into account, the background data were scaled prior to subtraction from the DUT-76(Cu) data. To obtain the reduced atomic pair distribution function (PDF), the PDFgetX3 algorithm was used through the xPDFsuite software. The maximum value of momentum transfer for complete quarter rings on the area detector was Qmax,inst = 27.7 Å -1 . However, to improve the signal-to-noise ratio of the resulting PDF, the Fourier transformation was terminated at a value of Qmax = 17.9 Å -1 , reflecting the poor counting statistics at higher Q-values due to limited signal of interest from the sample. |
677805156dde43c908304eb8 | 10 | The nitrogen isotherm measured on supercritically dried DUT-76(Cu) powder shows type Ib isotherm (Figure ) with three distinct steps before reaching a plateau at about 79.9 mmol•g -1 (at p/p0 = 0.4). The steps in the isotherm are obviously related to the filling of three different cages (Figure ), available in the crystal structure of the MOF: 1) the cuboctahedral pore with 12 Å in diameter; 2) the truncated octahedral pore with 21 Å in diameter; 3) cubic pore of 27 Å in diameter. The total pore volume (TPV) in saturation reaches 2.77 cm 3 •g -1 which is in agreement with the calculated total geometric volume of 2.76 cm 3 •g -1 for this structure (see Figure ). SEM images of DUT-76(Cu) were obtained indicating octahedral and cuboctahedral crystal shape (see Figure ) and the PXRD patterns of the as made as well as the desolvated MOF were measured (see Figure ) both fit with the simulated pattern generated from the originally reported structure. The TGA graph shows a small step of 1.86 % at 373 K due to moisture taken up while transferring the sample from a glovebox to the thermal analyser and then from T = 533 -638 K a strong mass loss is observed due to degradation of the linker leaving 20.9% of the initial mass as copper(II) oxide which is in agreement with the 21.9% calculated for this material. |
677805156dde43c908304eb8 | 11 | In order to prove the stability of the DUT-76(Cu) towards adsorption/desorption stress we conducted physisorption of C1-C4 hydrocarbons at their standard boiling points over several cycles (see Figure ). We started our experiments with methane physisorption at 111 K, because the methane shows the weakest interactions with the framework among all hydrocarbons. The isotherm shape upon methane physisorption is very similar to that of nitrogen with two steps until p/p0 = 0.2 and then reaching a plateau at p/p0 = 0.3 (see Figure ). The total uptake of methane in plateau is 74.3 mmol•g -1 (at p/p0 = 0.4) corresponding to TPV of 2.81 cm 3 •g -1 . These TPV of DUT-76(Cu) differ depending on the adsorbed hydrocarbon an in-depth discussion is given below (see Figure ). The physisorption is repeatable at least over three cycles without changes although the narrow hysteresis is observed in the pressure range p/p0 = 0 -0.1, corresponding to the adsorption and desorption of the methane in the cuboctahedral pore but obviously without deteriorating the crystal structure. |
677805156dde43c908304eb8 | 12 | As DUT-76(Cu) showed an exceptional ethene uptake in high-pressure adsorption experiments at 298 K we now aimed the in-depth fundamental studies of host-guest interactions with hydrocarbons of different lengths and degrees of saturation at their boiling points. When comparing the isotherms of ethane and ethylene already a significant influence of the degree of saturation is discerned (see Figure and). An almost identical uptake of ethane (45.5 mmol•g - 1 , TPV = 2.52 cm 3 •g -1 , at p/p0 = 0.40, 185 K) and ethylene (46.6 mmol•g -1 , TPV = 2.30 cm 3 •g -1 , at p/p0 = 0.40, 169 K) is reached in the first cycle. This is 0.3 and 0.5 cm 3 g -1 lower compared to methane and indicates that 10 and 18% of the DUT-76(Cu) sample respectively is irreversibly contracting during exposure to the adsorption stress. However, the three steps in the isotherm observed for nitrogen and methane are still visible. For both gases we observe two narrow hysteretic loops in the isotherms. The first one, similar to methane appears in the first step and the second one is observed between the second and third steps in the range of p/p0 = 0.1 -0.15. In the second physisorption cycle, however, the uptake of the two gases decreases significantly. Namely, for ethane a drop in gas uptake of about 20% to 36.8 mmol•g -1 (at p/p0 = 0.40, 185 K) is observed, whereas for ethylene this drop is more significant as the maximum uptake decreases to almost 50% to 26.1 mmol•g -1 (at p/p0 = 0.40, 169 K). The changes are obviously related to the desorption of the gas from the mesopore in the first adsorption/desorption cycle indicating that the framework is exposed to a larger stress upon desorption of ethylene. The second and third cycle isotherms show reversible behaviour with no hysteresis. This observation, as well as the fact that the sample was heated to 298 K and evacuated between measurements, also shows that the lower uptake is not due to adsorbate remaining in the pores but to changes in the framework itself. |
677805156dde43c908304eb8 | 13 | Physisorption of propylene and propane shows the similar trend observed for ethylene and ethane. A higher uptake of propane (29.6 mmol•g -1 , TPV = 2.30 cm 3 •g -1 , at p/p0 = 0.40, 231 K) than propylene (19.6 mmol•g -1 , TPV = 1.35 cm 3 •g -1 , at p/p0 = 0.40, 226 K) is achieved (see Figure and) indicating that the latter expose the framework to higher adsorption stress leading to irreversible contraction of 18 % and 52 % of the MOF respectively. Interestingly the third step of the isotherm is nearly eliminated indicating nearly complete contraction of the mesopore upon the first adsorption cycle. The first cycle of the propane isotherm shows a slight hysteresis at p/p0= 0.0 -0.2 relative pressure but we can also see the desorption branch lying underneath the adsorption branch meaning that the desorption occurs at higher relative pressure than the adsorption. This can be explained by a partial decomposition of the material upon guest adsorption and thereafter an earlier gas release from this partially decomposed phase. This effect becomes stronger a host guest interactions increase and was investigated using in situ PXRD (Figure , S6, S8 and S10) as well as SEM imaging (see Figure ). In the following cycles the uptake drops to 64% then further 6% and then remains almost constant in a fourth cycle. For propylene the behaviour is almost identical but without hysteresis in the first adsorption/desorption step. The uptake drops to about 55% the second cycle and slightly decreases to 52% in a third and fourth cycle. For both gases we observe the crossing of adsorption and desorption branches within the second step of the isotherm indicating the partial decomposition of the sample. |
677805156dde43c908304eb8 | 14 | Partial irreversible contraction of the framework during adsorption/desorption cycling can also be observed in ex situ PXRD experiments, conducted on the samples before and after the cycling experiments (see Figure ). The reflection intensities gradually decrease with increasing the chain length of the hydrocarbons, used in adsorption/desorption cycling experiments, although part of the sample still retain the crystallinity. This observation is well agreed with physisorption experiments indicating the reduced gas uptake and pristine shape of the isotherm. |
677805156dde43c908304eb8 | 15 | As the gas adsorption isotherms and the ex situ PXRD patterns indicate the changes of the material, in situ PXRD measurements in parallel to the adsorption of various gases should provide a clear image of the structural transitions occurring during gas adsorption. The results are in agreement with the previous observations as the reflections show the characteristic pattern of DUT-76(Cu) but the intensity rapidly decreases in the pressure range p/p0 = 0.1 -0.19, which is when the smallest kinds of pores -the cuboctahedral MOPs -are filled and the bigger tetragonal and ultimately the cubic pores fill up (see Figure ). In the desorption branch a further decrease in reflection intensities is observed. Subtle changes in the diffraction pattern are visible namely a transient increase of a peak at 2θ = 3.7° from p/p0 = 0.0 -0.039. In the higher angles range of 2θ > 8.0° reflection intensity drops and then increases again with a minimum at p/p0 = 0.0054. However, the formation of new crystalline phases was not observed. In situ PXRD, measured in parallel to physisorption of propane and n-butane show comparable results (see Figures ). The decrease in reflection intensity can be caused by either decomposition or amorphization. |
677805156dde43c908304eb8 | 16 | Since we observed the partial decomposition of the samples upon physisorption of C2-C4 hydrocarbons and no new crystalline phases appeared in PXRD patterns, SEM imaging was applied to analyze the changes on the macroscopic scale (see Figure ). f') in which the crystal shape can still be estimated but the crystal underwent significant degradation. In case of n-butane (Figure ) and butadiene (Figure ), almost the entire initial habitus vanished and mainly fragments or strongly decomposed material is visible. These observations are in line with physisorption and PXRD results indicating the contraction of the MOF with increasing of the size of the guests. The decreasing size of the intact crystals and amount of fragmented larger crystals can be explained by a stabilization of smaller crystals due to a smaller ratio of inner and outer surface. Total pore volume evolution as a function of adsorption/desorption stress Hydrocarbon physisorption data shows a gradual decreasing amount of gas adsorption as the chain length of the adsorbate increases and as double bonds are introduced. Both effects are likely to increase the amount of adsorption/desorption stress on the framework: longer chains lead to an increase of van-der-Waals forces, and the presence of double bonds introduces the possibility for π-π interactions. Both factors lead to an increase of the adsorption enthalpy and adsorption stress, which leads to the framework contraction. The underlying phenomena and correlations are clearly discerned by plotting the total pore volumes (Figure ) determined in multiple cycles with C1 to C4 alkanes and alkenes: For methane we can observe the highest pore volume of 2.81 cm 3 •g -1 which was also observed in the nitrogen physisorption measurement. This value does not change significantly over three adsorption/desorption cycles. Meaning that for methane the energy input through the adsorption enthalpy in adsorption and capillary forces in desorption is insufficient to contract the structure, leading to a reproducible and repeatable adsorption over multiple cycles. This is reflected in the SEM images as the majority of the crystals remains intact, their faces remain plain, their geometric shape and surface do not change after multiple cycles. |
677805156dde43c908304eb8 | 17 | For the C2 hydrocarbons a lower TPV is observed indicating partial contraction of the sample during adsorption. Following the trends for the rigidification of flexible MOFs a reasonable hypothesis is that larger crystals collapse completely while smaller crystals resist the contractive adsorption stress and maintain the pore volume of the sample. For ethane the TPV is decreased to 2.52 cm 3 •g -1 and then further falls by 20 % to 2.01 cm 3 •g -1 after a third cycle indicating a transition into a less porous structure. For ethylene the TPV is further decreased to 2.30 cm 3 •g -1 and after three cycles decreases by 44 % to 1.28 cm 3 •g -1 . This result is in agreement with other findings showing a higher isosteric heat of adsorption with increasing carbon chain length due to stronger Van-der-Waals interactions and additional π-π interactions when double bonds are introduced but the chain length remains the same. This lower TPV also correlates with the mean crystal size decreasing from 8.3 µm to 7.8 µm as well as the visible changes in the crystals morphology (Figure ). |
677805156dde43c908304eb8 | 18 | When going from the C2 to the C3 hydrocarbons the TPV for propane is 2.30 cm 3 •g -1 and reduces by 42 % to 1.33 cm 3 •g -1 which is very similar to the values for ethylene. A fourth adsorption/desorption cycle was measured thus the pore volume in the third cycle differed significantly from that in the second. For propylene the TPV in the first cycle is 1.35 cm 3 •g -1 with a 49 % decrease to 0.69 cm 3 •g -1 after four cycles. Again, these values are very similar to those for n-butane where a TPV of 1.33 cm 3 •g -1 with a 45 % decrease to 0.73 cm 3 •g -1 is observed. The similarity in TPV for the pairs ethylene/propane and propylene/n-butane is noticeable and may be explained by increasing Van-der-Waals interactions with compensating the lack of π-π interactions and vice versa. For 1,3-butadiene we observe the lowest total pore volume of 0.99 cm 3 •g -1 and a decrease to 0.46 cm 3 •g -1 after four cycles. |
677805156dde43c908304eb8 | 19 | These results correlate well: the adsorption stress from different alkanes larger than methane and alkenes causes an irreversible transition into an amorphous phase but also macroscopic deformation and fragmentation of the crystals when the adsorption/desorption stress becomes too strong. This leads to a decrease in gas uptake and total pore volume. Interesting empirical observation could be derived from physisorption data, namely the adsorption of the alkene with n The observations reveal an increasing fraction of collapsed framework with increasing chain length of the probe molecule, the presence of double bonds and number of adsorption/desorption cycles. For all gases, a constant adsorption/desorption performance could be reached after the third adsorption/desorption cycle. A combination of gas physisorption, X-ray diffraction, SEM imaging and total scattering allowed to quantify the amount of the contracted sample, monitor the changes in the crystallinity, macroscopic crystal shape and confirm the local structure of crystalline and amorphized samples. The proposed technique can be used for the semi-quantitative evaluation of the mechanical stability of highly sensitive mesoporous frameworks and can be considered as an alternative to existing techniques. |
666308e7e7ccf7753a38bc74 | 0 | The operating temperature of lithium-ion batteries for electric vechiles (EVs) can have a significant impact on their performance and life span. These can range from -20°C to 60°C, but for optimal performance and minimal degradation, 15°C to 35°C is generally recommended. This temperature range can be challenging to maintain, particularly in cold or hot climates, or when the battery is under high loads during driving and fast charging. Several battery thermal management systems have been developed to regulate heat transfer and ensure that battery temperatures remain within the optimal range. Traditional battery thermal management systems involve passive air cooling or indirect liquid cooling through metal pipes. In recent years immersion cooling, also called direct liquid cooling, has emerged as a promising solution to thermal management in a variety of applications such as batteries, motors, central processing units (CPUs), data centres, and photovoltaics. This involves partially or fully submerging the hot component in a dielectric fluid, which allows direct contact between the component and the coolant without causing shorting of the electrical components. In EVs, immersion cooling allows the battery pack to be directly submerged in a cooling fluid, eliminating the need for heavy cooling pipes, improving the thermal contact between the cells and the coolant, and potentially suppressing thermal runaway. Recent studies have shown that immersion cooling can remove heat more effectively than air cooling or indirect liquid cooling approaches. When designing a coolant for immersion cooling systems, there are several key factors that should be taken into consideration. Most importantly, the fluid should be dielectric in order to prevent short circuiting, as well as non-flammable and non-corrosive. The thermophysical properties of the fluid should also be considered to enable efficient heat dissipation. For example, a high thermal conductivity and specific heat capacity is required along with a low viscosity to maximise conductive and convective heat transfer. Additionally, the interfacial thermal resistance (ITR) across the battery casing-coolant interface can impact the overall cooling efficiency. While the ITR has been investigated across the cathode-separator and separator-casing interfaces within the battery pack, the ITR at the battery case-coolant interface as relevant to immersion cooling has not yet been widely studied. The importance of accurately describing the ITR is underscored by the impact that temperature fields can have on the local battery electrochemistry. For instance, 2D electrochemicalthermal modelling works have highlighted how varying cooling strategies significantly influence the average, extreme and locations of peak cell temperature and localised current density. In this respect, thermal gradients of only 3 °C can result in a positive feedback effect, which can lead to a 300% acceleration in battery degradation. Various works have measured the anisotropic and state dependent thermal properties of LIBs with with notable observations being the significant difference between in-plane (21-40.1 W K -1 m -1 ) and through-plane thermal conductivity (0.15-1.4 W K -1 m -1 ), as well as higher uncertainty in through-plane values. Whilst device-level modelling works in the literature account for these heterogeneous behaviours to varying degrees of fidelity, ITR remains poorly described from a physical perspective in the battery field and descriptions of heat removal often uses simplified thermal conduction and convective boundary conditions with fitted parameters based on surface cell temperatures. This poses potential issues with accurate description of cell temperature fields. |
666308e7e7ccf7753a38bc74 | 1 | The ITR, which is sometimes also referred to as the Kapitza resistance, or thermal boundary resistance, is a measure of the resistance to heat flow between two dissimilar materials. It is a crucial factor in understanding heat transfer across material interfaces. The ITR results in a temperature discontinuity where the two materials meet, arising from the contrasting electronic and vibrational properties of the two materials. The ITR, or its inverse, interfacial thermal conductance (ITC), is given by the equation: |
666308e7e7ccf7753a38bc74 | 2 | The ITR can have a significant impact on heat transfer, particularly at the nanoscale. It has become an focal point of research in areas such as thermal management for microelectronics. Experimental methods to study ITR, such as time-domain thermoreflectance (TDTR) and frequency-domain thermoreflectance (FDTR), are time-consuming and expensive. Non-equilibrium molecular dynamics (NEMD) simulations have emerged as a popular approach to determining the ITR. NEMD simulations have been used to investigate the ITR at a wide range of solid-liquid interfaces, including: graphene/water, for the iron oxide/hydrocarbon interface. This can be attributed, in part, to the absence of a comprehensive force field evaluation for iron oxide/hydrocarbon interfaces. Many force fields have been tested for the thermal properties of the individual components, hydrocarbons and iron oxide. Although many iron oxide/hydrocarbon solid-liquid potentials have been proposed in the literature, none of these has been verified as suitable for calculations of the ITR through comparison to first principles methods or experiments. This is important since the strength of the interfacial interactions are directly correlated to the wettability and ITR. In this study, we use NEMD simulations to compare the performance of several force fields for their ability to reproduce the key thermophysical properties of hydrocarbons and iron oxide. We then calculate the ITR for the iron oxide/hydrocarbon interface using NEMD simulations with a range of solid-liquid interaction potentials. Finally, we determine the most suitable interaction potential by performing work of adhesion calculations, which are verified against values derived from contact angle experiments. This study provides a crucial step towards the virtual screening of dielectric fluids for the immersion cooling of EV batteries as well as cooling fluids for other applications. |
666308e7e7ccf7753a38bc74 | 3 | In this study, we use poly-α-olefin (PAO) as a dielectric cooling fluid. PAO is a suitable candidate for battery immersion cooling systems due to its low cost, low toxicity, and adequate working temperature range. We simulated the branched hydrocarbon hydrogenated 1-decene trimer, which is the main component of PAO4. The molecular structure of PAO4 is shown in Figure . |
666308e7e7ccf7753a38bc74 | 4 | The fluid regions of our systems were generated using the Materials and Process Simulations platform (MAPS) from Scienomics SARL (Paris, France). In order to identify a valid model for the fluid, a series of benchmark simulations were performed using different force fields, with each being assessed on its ability to predict the density, thermal conductivity, and viscosity of the fluid. Four force fields were selected for this investigation, three atom-istic (LOPLS-AA, CHARMM36, AMBER 54 ) and one united-atom (TraPPE-UA). MD simulations were performed in LAMMPS 56 using a velocity-Verlet integrator. For the atomistic force fields, a time step of 0.5 fs was used, while for the united-atom force field this was increased to 5 fs. Note that satisfactory energy conservation was not observed for the atomistic force fields with a timestep of 1 fs. We did not constrain the C-H bonds with the SHAKE algorithm due to the difficulties in assigning the local temperature with rigid constraints. We used a Lennard-Jones (LJ) cut-off of 12 Å for all of the force fields. The unlike LJ interactions were calculated using arithmetic mean mixing rules for CHARMM36 and AMBER, while the geometric mean (Lorentz-Berthelot) mixing rules were used for LOPLS-AA and TraPPE-UA. The long-range Coulomb interactions between the partial charges on the carbon and hydrogen atoms in the atomistic force fields were calculated using the particle-particle particle-mesh (PPPM) algorithm with a relative error in the forces of 1 To calculate the thermal conductivity, a box was created with periodic boundary conditions in all three directions, containing 400 PAO4 molecules at an initial density of 0.8 g cm -1 . The systems were energy minimised, followed by a 1 ns equilibration in the isobaricisothermal (NPT) ensemble at 300 K and a pressure of 1 atm. We used the Nosé-Hoover thermostat and barostat with temperature and pressure damping coefficients of 0.1 ps and 1.0 ps, respectively. Next, the systems underwent a 0.25 ns equilibration in the canonical (NVT) ensemble at 300 K using a Nosé-Hoover thermostat. NEMD simulations were then performed to calculate the thermal conductivity by applying a heat flux along the z axis using a boundary driven approach. Hot and cold thermostat regions were defined in the xy plane, with a thickness of 5 Å, at the centre and edges of the simulation box. Atoms falling within these regions were thermostatted to 325 K and 275 K for the hot and cold regions respectively, using the Langevin thermostat. A temperature damping coefficient of 1.0 ps was used, which ensured good energy conservation during then NEMD simulations. The NEMD simulations were performed for 5.5 ns, with the initial 0.5 ns discarded to allow the system to reach a non-equilibrium steady state. |
666308e7e7ccf7753a38bc74 | 5 | where m i and v i are the mass and velocities of a particle i within the bin and N f is the total number of degrees of freedom in the bin. Since the force fields used in this investigation were fully flexible, N f = 3N b , where N b was the number of atoms in each bin. The temperature profile was calculated every 0.5 ps and averaged across the whole simulation. ∇T , was calculated by fitting the temperature profile in the linear regions between the hot and cold thermostats. The heat flux was then calculated according to the equation: |
666308e7e7ccf7753a38bc74 | 6 | , where E hot and E cold are the cumulative energies added and subtracted from the atoms in the hot and cold regions during the NEMD simulation respectively, t is the time over which the simulation proceeded and A is the cross sectional area of the simulation box in the xy plane. ⟨ Ė⟩ was determined by the gradient of the energy exchanged in the hot and cold thermostats against time. The thermal conductivity of the fluid was then calculated using Fourier's law: |
666308e7e7ccf7753a38bc74 | 7 | Figure in the Supplementary Information illustrates a typical temperature profile along the z-axis for the PAO4 systems during the NEMD simulation. The system was divided into 200 bins along the z-axis and each point on the graph represents the temperature of one bin, averaged over the final 5 ns of the NEMD simulation. ∇T was calculated by linear fitting a 10 Å region in the centre of the hot and cold thermostats to avoid fitting the non-linear regions of the temperature profile near the thermostats. ∇T was taken as the average of the two fitted regions. The energies exchanged in the hot and cold thermostats during the simulation are shown by the red and blue lines respectively in Figure in the Supplementary |
666308e7e7ccf7753a38bc74 | 8 | Information. The sign of the energy exchanged in the hot thermostat has been flipped to enable comparison of the energy exchanged at both thermostats. By determining the gradient of the two lines and taking the average of the hot and cold thermostats, we obtained the rate of heat exchange, ⟨ Ė⟩, which can be used in Equation 3 to calculate the heat flux. |
666308e7e7ccf7753a38bc74 | 9 | Using this method to study small systems, where the length of the simulation parallel to the heat flux, l z , is comparable to the phonon mean free path (MFP), can result in the emergence of finite size effects. The MFP represents the average distance a phonon travels before colliding with other energy carriers. In small systems, incomplete phonon scattering events and boundary reflections lead to under predictions in thermal conductivity estimates, compared to 'bulk' systems. To check for possible finite size effects in our systems, the simulation box was replicated in the z axis, parallel to the heat flux, and the thermal conductivity was recalculated. No significant difference in thermal conductivity, indicating that the results were size independent and therefore a 70 × 70 × 70 Å3 simulation box was sufficiently large enough to directly obtain the bulk thermal conductivity. |
666308e7e7ccf7753a38bc74 | 10 | where η αβ is the shear viscosity for components α, β, V and T are the system volume and temperature, t is the time, P αβ is the off diagonal component of the pressure tensor and ⟨P αβ (0)P αβ (t)⟩ is the pressure autocorrelation function (PACF). There are three off diagonal components of the pressure tensor; P xy , P xz , P yz and three corresponding viscosity components; η xy , η xz , η yz . The shear viscosity, η, was taken as the average of the 3 η αβ components. |
666308e7e7ccf7753a38bc74 | 11 | The viscosity simulations were performed using a smaller box, containing 100 PAO4 molecules, since previous MD studies have demonstrated a weak system size-dependence using the Green-Kubo method. Periodic boundary conditions were applied in all directions and the systems underwent the same equilibration procedure as described for the thermal conductivity calculations. The equilibrated systems then ran in the NVT ensemble for 50 ns, during which time the integral of the PACF was calculated to obtain the viscosity, using equation 5. The viscosity was obtained by calculating the mean of the viscosity profile once it had converged. The autocorrelation was calculated with a sample rate, s, of 10 fs, correlation length, p, of 100,000 fs over a correlation time period, d = sp = 1000 ps. For an explanation of how the correlation time, d, was determined, see the Supplementary Information. |
666308e7e7ccf7753a38bc74 | 12 | In this study, we have chosen to use iron oxide surfaces to represent the oxidised, outermost layer of the steel casing of a LIB. We study hematite (α-Fe 2 O 3 ), the most abundant and thermodynamically stable iron oxide under atmospheric conditions. The crystal structure of hematite features layers of oxygen atoms that are hexagonally close-packed and distorted, separated by an iron (Fe 3+ ) double layer. Due to its high thermodynamic stability, we investigate the α-Fe 2 O 3 (0001) surface with a half-metal termination (Fe-O 3 -Fe). We consider only atomically-smooth surfaces. Previous NEMD simulation have shown that for strong solid-liquid interactions, increasing the nanoscale surface roughness has only a minor effect on the ITR. The surfaces used in this study were generated using the Atomic Simulation Environment (ASE). To identify the most suitable potential for the solid, a series of benchmark simulations were performed using different force fields, with each being assessed on its ability to predict the density and thermal conductivity of the iron oxide slab. We compared the use of Additionally, the harmonic model introduces a harmonic bond potential to describe the atomic bonds. These simulations were performed in LAMMPS 56 using a velocity-Verlet integrator and a timestep of 0.5 fs. This was short enough to ensure good energy conservation. LJ and Morse interactions were cut off at 12 Å and long range electrostatic interactions were evaluated using the particle-particle particlemesh (PPPM) solver with a relative error in the forces of 1 ×10 -5 . LJ interactions between dissimilar atoms were calculated using the geometric mean mixing rules. |
666308e7e7ccf7753a38bc74 | 13 | Initially, a 50.3 × 43.6 × 54.9 Å3 hematite surface was generated, which was then replicated in the z direction, parallel to the [100] crystallographic direction, to create additional systems of length 109.8, 164.8, 219.7 and 274.6 Å. The systems were energy minimised, followed by a 1 ns equilibration in the NPT ensemble at 300 K with an anisotropic pressure coupling of 1 atm in the z-axis, using a Nosé-Hoover thermostat and barostat with temperature and pressure damping coefficients of 100 fs and 1,000 fs respectively. Finally, the system underwent a 0.25 ns equilibration in the NVT ensemble at 300K using a Nosé-Hoover thermostat. Once the system had equilibrated, NEMD simulations were performed to calculate the thermal conductivity by applying a heat flux along the z-axis. Hot and cold thermostat regions were defined in the xy plane, with a thickness of 5 Å, at the centre and edges of the simulation box. Atoms falling within these regions were thermostatted to 325 K and 275 K for the hot and cold regions respectively, using a Langevin thermostat. Previous comparisons suggested that the Langevin thermostat is preferrable to the Nosé-Hoover thermostat for this purpose. NEMD simulations were performed for 3 ns, with the initial 0.5 ns being discarded to allow the system to reach a non-equilibrium steady state. The temperature gradient, ∇T , and heat flux, J q , were calculated using method outlined previously, allowing the size dependant thermal conductivity to be calculated. |
666308e7e7ccf7753a38bc74 | 14 | To calculate the bulk thermal conductivity, κ ∞ , the inverse of the thermal conductivity, 1 κ , was plotted against the inverse of the system length, 1 lz . The results were fitted linearly and κ ∞ was given by the inverse of the y intercept. The thermal conductivity was also calculated in the [001] and [010] directions to investigate the anisotropy of this property in hematite. Figure in the Supplementary Information illustrates a typical temperature profile along the z-axis for the Fe 2 O 3 systems during the NEMD simulations. The system was divided into 20n bins, where n was the number of times the initial 50 Å simulation was replicated in the z axis. The points on the graph represent the temperature of each bin, averaged over the final 5 ns of the simulation. As the box length, l z , increased, so did the size of the non-linear regions by the thermostats and as such, an exclusion zone of 5n Å was left on either side of the thermostats when calculating the temperature gradient to avoid fitting in the nonlinear region. The heat flux was obtained by calculating the rate of heat exchange, ⟨ Ė⟩, at the hot and cold thermostats and the thermal conductivity was calculated using Fourier's law. In all cases investigated here, there was a < 1 % difference in the total energy exchanged at the hot and cold thermostats during NEMD simulations, indicating good energy conservation. The thermal conductivity was calculated for five systems ranging from l z = 50 to 250 Å. For all four force fields, κ was found to increase as a function of l z . To obtain the bulk thermal conductivity, κ ∞ , 1 κ was plotted against 1 lz , as displayed for each force field in Figure in the Supplementary Information. For ClayFF, INTERFACE-FF and the Morse potential, all system sizes were found to fall within the linear regime; however, for the harmonic model only the systems between 100 and 250 Å fell in the linear regime. The linear regime was fitted and extrapolated to obtain the y intercept, which was equal to κ∞ . |
666308e7e7ccf7753a38bc74 | 15 | The initial starting configuration contained a 50.29 × 43.55 × 54.92 Å3 hematite surface, generated using ASE, placed between two liquid regions, containing a total of 440 PAO4 molecules. To check for possible finite size effects, the length of the PAO4 and hematite regions were varied in the direction parallel to the heat flux and R k was computed. The results showed no size dependency and it was therefore assumed that the system described above was sufficiently large enough to calculate the "bulk" ITR. |
666308e7e7ccf7753a38bc74 | 16 | The first term defines the LJ forces, where ε ij represents the strength of the interaction, σ ij represents the distance at which the potential between particles i and j is zero and r represents the separation between the particles. The second term defines the Coulombic interaction, where C represents an energy conversion constant and q i and q j represent the charges on atoms i and j. |
666308e7e7ccf7753a38bc74 | 17 | A wide range of different LJ parameters for interactions between hematite and various fluids can be found in the literature. Since none of these have been verified for the nanoscale thermal transport, we systematically vary the LJ potential parameters for the hematite atoms, ε F e/O and σ F e/O , drawing from different sources, including; ClayFF, INTERFACE-FF, Berro et al. and Savio et al. These parameters were combined, using the geometric mean mixing rules, with the LJ parameters for the fluid, ε C/H and σ C/H , which were taken from the LOPLS-AA force field. In total, we compared eight different LJ potentials, with varying interaction strength. The Fe and O LJ and partial charge parameters used for each of the solid-liquid potentials are shown in Table . Previous comparisons have suggested that, compared to ClayFF, the stronger LJ interactions proposed by Savio et al. more accurately reproduced force-distance curves obtained from DFT for hydrocarbons adsorbed on hematite surfaces. To ensure the stability of the surfaces using the solid Morse potential, the partial charges remained constant for all of the interfacial force fields at q F e = +1.800 and q O = -1.200. This also meant that any differences in the ITR were due to changes to the LJ parameters. Note that the original INTERFACE-FF (q F e = +1.740, q O = -1.160), ClayFF (q F e = +1.575, q O = -1.050), and Berro et al. (q F e = +0.771, q O = -0.514) force fields used smaller partial charges. Tests were also performed for ClayFF using the original LJ parameters for the solid-solid interactions and partial charges. For the relatively nonpolar PAO molecules considered here, we found that the ITR was identical to that obtained with the modified partial charges within the uncertainty of the NEMD simulations. These comparisons are shown in the Supplementary Information (Figure ). The NEMD simulations to calculate the ITR were performed in LAMMPS 56 using a velocity-Verlet integrator and a timestep of 0.5 fs. This timestep was sufficiently short to ensure energy conservation during the NEMD simulations. Five independent repeats were performed for each of the eight solid-liquid potentials, with each trajectory being given its own randomly generated initial velocity seed. Each system was energy minimised, followed by a 1 ns equilibration in the NPT ensemble at 300 K with an anisotropic pressure coupling of 1 atm in the z-axis, using a Nosé-Hoover thermostat and barostat with temperature and pressure damping coefficients of 0.1 ps and 1.0 ps respectively. Finally, the systems underwent a 0.25 ns equilibration in the NVT ensemble at 300K using a Nosé-Hoover thermostat. To calculate the ITR, a heat flux was applied along the z-axis, as described in the previous sections. The hot and cold thermostats regions had a thickness of 10 Å and atoms falling within these regions were thermostatted to 325 K and 275 K respectively, using a Langevin thermostat with a temperature damping coefficient of 1.0 ps. Previous studies have highlighted the importance of appropriate thermostatting when calculating the ITR from NEMD simulations. In particular, for stochastic thermostats such as Langevin, if the coupling of the thermostat and the system is too strong (short temperature damping coefficient), then spurious temperature jumps can be created between the solid and fluid re-gions, leading to overestimated ITR. Therefore, we ensure that our choice of temperature damping coefficient was sufficiently large such that it did not influence our ITC results. |
666308e7e7ccf7753a38bc74 | 18 | The NEMD simulations were performed for 5.5 ns, with the initial 0.5 ns discarded to allow the system to reach a non-equilibrium steady state. During the NEMD simulations, the center of mass (COM) of the slab was restrained in the z-axis to prevent drift. Temperature and density profiles were calculated by dividing the system into 100 and 5,000 bins, respectively, along the z-axis. The temperature within each bin was calculated using Equation 2 and the density was calculated using equation, ρ |
666308e7e7ccf7753a38bc74 | 19 | where ⟨v α (0)v α (t)⟩ is the VACF of species α, and ω is the frequency. The VDOS for the fluid, D P AO4 (ω), was calculated for all fluid atoms falling within the first adsorption layer and the VDOS for the surface, D Fe 2 O 3 (ω), was calculated for all surface atoms within a 5 Å distance from interface. After sampling various correlation times for the VACF, we determined that a correlation time of 4 ps provided the optimal balance, yielding comprehensive vibrational information while minimizing noise in the resulting VDOS. During this time, < 5% of initial interfacial PAO4 molecules had left the region. The same systems as used for the ITR calculations were also used for the VDOS calculations. |
666308e7e7ccf7753a38bc74 | 20 | The solid-liquid work of adhesion (W SL ) defines the reversible work required per unit area to separate the interface between a solid and a liquid phase to an infinite distance. It indicates the relative strengths of adhesive and cohesive forces at the interface. Positive W SL denotes strong adhesion, leading to wetting, while negative W SL signifies poor wetting with dominant cohesive forces. Zero W SL indicates equilibrium between adhesive and cohesive forces, resulting in minimal interaction. W SL is given by the Young-Dupré equation: |
666308e7e7ccf7753a38bc74 | 21 | where γ sv , γ sl and γ lv represent the solid-vapour, solid-liquid and liquid-vapour surface tensions acting on the equilibrium point. The combination of Young's equation, Equation in the Supplementary Information, and Equation 10 yields the equation for W SL in terms of the liquid-vapour surface tension, γ lv , and the contact angle θ: |
666308e7e7ccf7753a38bc74 | 22 | A direct approach to calculating W SL involves using Equation 11 and obtaining contact angles through simulations of nanoscale droplets. However, obtaining θ can give rise to a number of complications, such as long equilibration times, 82 finite size effects and difficulty measuring contact angles in extreme or fully wetting scenarios, which introduces uncertainty into the results. |
666308e7e7ccf7753a38bc74 | 23 | Recently, alternative methodologies have been developed to enable the calculation of W SL , without requiring the simulation of nano-scale droplets. Instead these approaches rely on thermodynamic integration. The phantom wall method involves the use of a virtual piston to separate separate the solid and liquid phases, whilst the dry-surface method involves perturbing the interactions between the solid and liquid phases by introducing a coupling parameter, λ, to transition from a state where interactions are fully coupled, to a fully decoupled state. In this work we will employ the dry-surface method with damped Coulomb interactions, as proposed by Surblys et al. To calculate W SL , simulations were performed using the same system as used in previous R K calculations. It was suggested by Surblys et al. that systems with long-range Coulombic interactions require the solution of additional Poisson equations when performing dry-surface calculations. As a solution, the long-range Coulombic potential for solid-liquid interactions was replaced with a new potential which made use of damped Coulombic interactions: |
666308e7e7ccf7753a38bc74 | 24 | where q was the charge of particles i or j, erf c() is the complementary error-function, α is the damping parameter and r c is the cutoff distance of Coulombic interactions. The cutoff distance was set to a value of r c = 12 Å and the damping parameter was set to a value of α = 0.2 Å-1 . LJ and Coulombic interactions were gradually switched off, by scaling λ from λ = 1 for fully coupled solid-liquid interactions, to λ ≈ 0 for fully uncoupled solidliquid interactions. The coupling parameter was not sampled at λ = 0 to avoid numerical instability that may arise when solid-liquid interactions are entirely decoupled. Values for λ = 0 were extrapolated from the two nearest data points. The work of adhesion could be obtained using the equation: |
666308e7e7ccf7753a38bc74 | 25 | The coupling parameter, λ, was sampled at 24 equally spaced data points in the range λ ∈ (0,1]. At each point, the system was equilibrated for 0.25 ns followed by a 0.75 ns run to obtain u sl . The Hamiltonian derivatives, ∂u sl ∂λ , were computed using two distinct methods: numerical differentiation and polynomial fitting. The first method employed numerical differentiation of u sl through the central differences scheme. The second method involved fitting u sl with a second-order polynomial and subsequently obtaining its derivative. The resulting numerical and polynomial Hamiltonian derivatives were then integrated using the trapezoid rule and analytic methods, respectively, allowing W sl to be calculated using equations 16-18. |
666308e7e7ccf7753a38bc74 | 26 | Thermal conductivities calculated using each force field are presented in Table . In all cases, the thermal conductivity calculated using the atomistic force fields was overestimated, with AMBER resulting in a 68% overestimation, LOPLS-AA resulting in a 63% overestimation and CHARMM-36 resulting in a 57% overestimation with respect to the experimental value of 0.14 W K -1 m -1 at room temperature and pressure. A similar overprediction has been seen in other NEMD studies where classical, atomistic force fields have been used to calculate thermal conductivity of hydrocarbons. This may be a result of thermal transport properties not being considered during the parametrisation of these force fields. It has been suggested that this over prediction may be a systematic error relating to classical force fields failing to capture quantum effects. On the other hand, the thermal conductivity calculated using the united-atom force field, TraPPE-UA, was underpredicted by 20%. This result was consistent with literature findings that suggested that united-atom force fields are capable of predicting thermal conductivity more accurately than all-atom force fields. |
666308e7e7ccf7753a38bc74 | 27 | For each force field, the viscosity was calculated at 40 °C and 100 °C to enable comparison with experimental data. The results are shown in Table . The viscosity calculated with the united-atom force field was under predicted by 84% at 40 °C and 68% at 100 °C with respect to experimentally obtained values of 16.3 and 3.7 cSt at 40 and 100 °C, respectively. Of the atomistic force fields, AMBER performed the worst, resulting in over-estimations of 204% at 40 °C and 122% at 100 °C. This is consistent with previous simulations, which suggested that AMBER resulted in crystallisation of hydrocarbons under ambient conditions. CHARMM-36 under predicted the viscosity by 48% at 40 °C and 37% at 100 °C. LOPLS-AA performed the best, yielding a 36% underprediction at 40 °C and a 12% underprediction at 100 °C. |
666308e7e7ccf7753a38bc74 | 28 | Of the force fields tested in this work, we determine LOPLS-AA to be the most suited to modelling heat transport in PAO4 fluid. Whilst all force fields predicted the fluid density well, we saw large differences between simulation and experimental results with for the thermal conductivity and viscosity calculations. Whilst TraPPE-UA made the best predictions for the fluid thermal conductivity, we observed significant deviations from the experimental viscosity at both high and low temperatures. Whilst all three atomistic force fields tested all showed similar over predictions in thermal conductivity, LOPLS-AA provided significantly better predictions for the viscosity at both high and low temperatures. Therefore, we chose to use the LOPLS-AA force field to model the fluid for the remainder of simulations in this study. |
666308e7e7ccf7753a38bc74 | 29 | [100] and [010] direction, and 12.1 W K -1 m -1 in the [001] direction. We found that using the harmonic model yielded a thermal conductivity that was overestimated by > 500%. We believe this overestimation to be a result of the harmonic bond potential, which assumes that atoms oscillate around their equilibrium positions in a perfectly harmonic manner. In |
666308e7e7ccf7753a38bc74 | 30 | Using the LOPLS-AA force field to capture fluid behavior and the Morse potential to represent surface behavior, we computed the ITR for each of the eight solid-liquid LJ potentials. Figure shows the starting configuration used for these calculations and the time-averaged temperature profiles obtained for the weakest (ClayFF ) and strongest (Savio et al. ) solid-liquid LJ potentials. The ITR, R k , was calculated using the equation: |
666308e7e7ccf7753a38bc74 | 31 | where ∆T is the temperature drop across the interface and J q is the heat flux, calculated using Equation . The temperature points near the heat source and heat sink regions and those adjacent to the interface are ignored in a linear fit to the solid and liquid temperature profiles used to calculate the ITR. For ClayFF, 45,76 ∆T = 17.9 K, while for Savio et al. ∆T = 5.9 K. for Savio et al. The value calculated for the strongest potential is similar to that reported from previous NEMD simulations of the lower limit of the ITC (R k ∼ 6 m 2 K GW -1 ) for ionic liquid-graphene interfaces. The range of ITR values is also similar to that observed experimentally (R k = 6-20 m 2 K GW -1 ) for the interface between water and chemically functionalized aluminum and gold surfaces. The ITC was found to increase linearly with ε F e-C and ε F e-H , as shown in Figure in the Supplementary Information. This linear relationship between G k and ε has been reported in previous NEMD studies for simple LJ systems and the Si/Water interface. We do not observe an exponential scaling of G k with ε F e-C and ε F e-H , as has been observed in some NEMD simulations at low interaction strengths. 98,99 We conducted further analysis to explore the temperature dependency of R k . Simulations |
666308e7e7ccf7753a38bc74 | 32 | were carried out at a range of temperatures, T = 250-500 K, where T represents the mean temperature of the hot and cold thermostats. This temperature range was selected to sample the full liquid range of PAO4. Our findings revealed that there was no significant variation in ITR across the range of temperatures tested. Detailed results of these investigations are presented in Figure in the Supplementary Information. |
666308e7e7ccf7753a38bc74 | 33 | To investigate molecular-level origins of the observed dependence of the R k on the solid-liquid interaction strength, we calculated the VDOS for the α-Fe 2 O 3 and PAO4 regions adjacent to the interface for potentials each potential. The interfacial regions used to calculate the VDOS are illustrated in Figure (a) and the resulting VDOS is shown in Figure (b). The VDOS of PAO4 showed peaks at ν = 700, 900, 1160 cm -1 , which corresponded to C-C stretching, ν = 1340 cm -1 , which corresponded to C-H bending, and ν = 3000 cm -1 , corresponding to C-H stretching. The VDOS for α-Fe 2 O 3 showed a series of peaks between 100 and 800 cm -1 , which is consistent with previous experimental studies for hematite. 101 Despite significant differences in R k with solid-liquid interaction strength, there were no observable changes in the VDOS of the α-Fe 2 O 3 or PAO4 regions, as shown in Figure . |
666308e7e7ccf7753a38bc74 | 34 | where D P AO4 and D Fe 2 O 3 are the VDOS for the PAO4 and Fe 2 O 3 regions respectively. S has been used in a number of studies to rationalise changes to the ITR of solid-liquid and solid-solid interfaces. 2,102 However, we found no significant difference in S for the different solid-liquid potentials, as shown in Figure (c). Therefore, we conclude that the variations in R k with solid-liquid interaction strength do not originate from VDOS and must be attributed to other factors. These findings were consistent with results reported by who found no correlation between R k and the VDOS overlap, S, for the water-graphene interface. |
666308e7e7ccf7753a38bc74 | 35 | We observed a significant change in the density profiles of PAO4 for the different solid-liquid potentials. In all cases, the PAO4 density profile exhibits oscillatory behaviour in the region near the interface. This behaviour arises from the organisation of PAO4 molecules into layers near the surface, due to the strong intermolecular interactions between PAO4 molecules and the Fe 2 O 3 surface. The first adsorption layer, situated nearest to the Fe 2 O 3 surface, displays the highest density and is followed by a series of peaks of diminishing density as the distance from the surface is increased. At a distance of ∼ 25 Å from the surface, the density profiles level and approach the bulk liquid density (0.8 g/cm 3 ). The layering of PAO4 molecules near the interface did not significantly affect the fluids thermal conductivity, as shown by the consistent temperature gradient observed in both the bulk and interfacial liquid regions in Figure . This finding aligns with previous NEMD simulations that examined the effect of liquid ordering on heat flow in an atomic LJ solid-liquid system. The peak density of the first adsorption layer, ρ max , was found to increase with the strength of the solid-liquid LJ potential, whilst the average density in the bulk region of the liquid layer remained unchanged, Figure |
666308e7e7ccf7753a38bc74 | 36 | The findings above highlight the importance of selecting an accurate potential to describe solid-liquid interactions when calculating R k . Due to the inherent challenges associated with the experimental determination of R k at solid-liquid interfaces, there is a lack of experimental data to use as a benchmark for these simulations. In the absence of such experimental data, we consider alternative methodologies to identify the optimal parameters for describing the solid-liquid interactions. |
666308e7e7ccf7753a38bc74 | 37 | By comparing the simulation results to experimental data for the surface wettability of hematite by PAO4, 105 it is possible to select which potential best describes solid-liquid interactions. Various different types of molecular dynamics simulations can be used to predict the wettability of solid surfaces. For example, the contact angle can be quantified through monitoring the geometry of nanodroplets or free energy-based methods. Alternatively, molecular dynamics simulations can be used to directly quantify the work of adhesion. While we tried both nanodroplet contact angle and work of adhesion simulations, we found the latter to be the most instructive. Previously, experimental values for the work of adhesion have been used to tune the solid-liquid LJ interaction strength for graphene-water systems. The solid-liquid work of adhesion was calculated for the α-Fe 2 O 3 interface for each of the selected solid-liquid potentials. The solid-liquid potential energy, u sl , and corresponding derivative, ∂u sl ∂λ , for the weakest (ClayFF ) and strongest (Savio et al. ) are shown as a function of λ in Figure . The Coulombic and LJ components of the solid-liquid potential energy, u sl(C) and u sl(LJ) , respectively, are shown separately. We observe that, as λ is scaled towards zero, the contributions of both components tend towards zero. As the strength of the solid-liquid LJ potential was increased from ClayFF to Savio et al., the LJ contribution to the work of adhesion, W sl(LJ) , increased significantly. The Coulombic contribution, W sl(C) , was much smaller and was essentially unchanged between both solid-liquid potentials. These observations were expected since the PAO4 fluid is nonpolar and the partial charges for both solid-liquid potentials were identical. The overall work of adhesion W sl was calculated using Equation . |
666308e7e7ccf7753a38bc74 | 38 | We observed a linear increase of W sl with ε F e-C and ε F e-H . Since W sl ∝ (1 + cos θ), this result was in agreement with the scaling relationship, ε ∼ (1 + cos θ), proposed by Sendner et al. 107 from molecular dynamics simulations of the water/diamond interface. This relationship has been further verified in molecular dynamics simulations of water/silicon interfaces. The ITR is plotted as a function of W sl in Figure . We observed that R k ∼ 1/W sl and therefore G k ∼ W sl , as shown in Figure in the Supplementary Information. This relationship has been established in a number of previous experimental and NEMD simulation studies. We found that W sl was overestimated for all of the solid-liquid potentials compared to the value determined from contact angle expriments by Kalin and Polajnar (57.5 mJ m -2 ). 105 Furthermore, contact angle simulations performed using these potentials found that θ was underestimated in all cases, as shown in the Supplementary Information. These results indicate that all of the solid-liquid potentials overestimate the interaction strength determined experimentally. |
666308e7e7ccf7753a38bc74 | 39 | To address the overestimation of W sl , we introduced a new solid-liquid potential, ClayFF 0.5ε O , which halved the value of ε O in ClayFF for the solid-liquid LJ interactions (ε O = 0.0777 kcal mol -1 ), whilst leaving the remaining parameters unchanged. We decided to change the ε O parameter because the ε F e parameter of ClayFF was already so small (ε F e = 9.0 × 10 -6 ) that reducing it further would likely have a negligible impact on the W sl . We then recalculated θ and W sl with this new potential. The results of these simulations are detailed To achieve this agreement with W sl from contact angle experiments, 105 we reduced the strength of the O surf -C f luid and O surf -H f luid LJ interactions, which have been carefully parameterised (at least for O f luid -C f luid and O f luid -H f luid ) in the molecular force-field. This raises the question as to whether our system is fully representative of the state of solidliquid interface in the contact angle experiments. Experimental studies have shown that surface hydration and hydroxylation can occur on hematite surfaces at relative humidities as low as 0.1 %. The relative humidity was not reported for the contact angle experiments used to calculate the work of adhesion. 105 However, since no controlled atmosphere was used, it is very likely that an interfacial hydroxyl and/or water layer was present between the hematite and the PAO4 in the contact angle experiments. The presence of these layers could screen the interactions between the Fe 2 O 3 and PAO4, reducing adhesive forces with respect |
666308e7e7ccf7753a38bc74 | 40 | The three atomistic force fields overestimated the thermal conductivity by a similar amount (∼ 60%), while the united-atom force field under predicted the thermal conductivity by 20%. From viscosity calculations, we saw significant under predictions at high and low temperatures with CHARMM-36 and TraPPE-UA, and over predictions using AMBER. On the other hand, LOPLS-AA was found to be in good agreement with experimental data and was therefore chosen as the force field that best captures the behaviour of PAO4. |
666308e7e7ccf7753a38bc74 | 41 | For the hematite surface, we benchmarked four different force fields and again observed that, whilst all force fields could accurately predict the density, the predicted thermal conductivity varied significantly from experimental results. The most accurate force field we tested was the Morse potential. The thermal conductivity calculated using this model was in excellent agreement with experimental data obtained in all three crystallographic directions. |
666308e7e7ccf7753a38bc74 | 42 | To rationalise this behaviour we calculated the VDOS for the hematite and PAO4 regions adjacent to the interface as well as the overlap integral of the two spectra. We observed no significant changes in the vibrational spectra for the hematite or PAO4 regions or the overlap integral as the strength of the interactions were increased. However, we observed that a significant increase in the first adsorption layer density of the fluid as the strength of the solid-liquid interactions was increased. We therefore conclude that the decrease in ITR arises, not from changes in the vibrational spectra of the two components, but from the increased density of PAO4 molecules close to the surface as the solid-liquid interaction strength was increased. |
666308e7e7ccf7753a38bc74 | 43 | We observed that, for all of the solid-liquid potentials tested, the work of adhesion was over predicted compared to the value derived from contact angle experiments. To reconcile our simulation results with the experiments, we had to reduce the strength of the O surf -C f luid and O surf -H f luid interactions. This lead to a much larger ITR than for the other potentials. |
666308e7e7ccf7753a38bc74 | 44 | These findings indicate that a more detailed representation of the solid-liquid interface might be needed to accurately reproduce experimental properties. We propose that the inclusion of a hydroxyl and/or water layer at between hematite and PAO4 may provide a more accurate model of the interface in the experiments. Further investigation in future studies is necessary to increase our understanding and improve the accuracy of molecular simulations of the hematite/PAO4 interface. The methodology proposed here could also be readily extended to study immersion cooling fluids for other applications, such as CPUs, data centres, and photovoltaics. |
67780e9e6dde43c90831f476 | 0 | Photocatalytic hydrogen evolution (HER) has garnered significant attention in recent years as a promising method for sustainable hydrogen production. Among the various strategies employed, single-atom photocatalysis stands out as an innovative approach . In single-atom photocatalysis, metal atoms are isolated and anchored on a support material, allowing for the maximum exposure of active sites . This unique configuration facilitates highly efficient light absorption and electron transfer, making single-atom photocatalysts (SACs) particularly valuable in the HER process. Their ability to function effectively under ambient conditions, coupled with the possibility of tuning their electronic properties through careful material design, has positioned SACs as a critical component in advancing hydrogen production technologies . Their importance lies in the potential for high catalytic efficiency with low loading of precious metals, thus reducing the overall cost of photocatalytic systems. |
67780e9e6dde43c90831f476 | 1 | Organic photocatalysis has emerged as a versatile and cost-effective alternative to traditional inorganic photocatalysts in the field of HER . Organic photocatalysts, particularly small organic molecules and conjugated polymers, are attractive due to their tunable electronic properties, which can be engineered to match the light absorption spectra of the solar spectrum . The presence of conjugated π-electrons in organic materials facilitates efficient electron transfer upon light excitation, making them suitable for catalyzing the hydrogen evolution reaction . Moreover, organic photocatalysts are often light, flexible, and capable of being processed into thin films or coatings, offering additional advantages in terms of scalability and ease of integration into renewable energy systems . This flexibility, combined with their generally lower toxicity compared to metal-based catalysts, makes organic photocatalysts an appealing choice for largescale HER applications. |
67780e9e6dde43c90831f476 | 2 | The properties of organic photocatalysts play a crucial role in their effectiveness for HER. The absorption of visible light, due to their narrow bandgaps, allows organic materials to utilize a broader spectrum of sunlight, increasing their potential efficiency. Additionally, low-cost synthesis methods, along with the ability to modify their structure at the molecular level, enable fine-tuning of their electronic properties for enhanced photocatalytic activity . The charge carrier dynamics, such as efficient charge separation and transfer, are optimized through the careful design of organic materials, allowing for better hydrogen production rates . Furthermore, heteroatom doping, such as introducing nitrogen or sulfur into the organic structure, has been shown to significantly improve the electronic properties and the stability of the photocatalysts, enabling long-term hydrogen evolution without significant degradation . |
67780e9e6dde43c90831f476 | 3 | Despite their promising properties, organic photocatalysts face several challenges that hinder their efficiency in HER . One major issue is the low stability of many organic materials under harsh photocatalytic conditions, particularly in aqueous environments . These materials are susceptible to photocorrosion, which can reduce their lifetime and effectiveness over time. Additionally, while organic photocatalysts are capable of absorbing sunlight efficiently, poor charge separation often occurs due to the relatively low conductivity of many organic materials. This leads to the rapid recombination of photogenerated charge carriers, which reduces the overall efficiency of the HER process . The recombination of electrons before they can participate in the hydrogen evolution reaction results in a significant loss of available energy, making it one of the most critical challenges to overcome for enhancing the performance of organic photocatalysts in HER. |
67780e9e6dde43c90831f476 | 4 | The phenomenon of electron recombination is particularly detrimental to photocatalytic hydrogen evolution efficiency. When a photocatalyst absorbs light, it generates electron-hole pairs ; however, if these charge carriers recombine before they can be utilized in the HER process, the catalytic efficiency is significantly reduced. Electron recombination occurs when the excited electrons and holes return to their original states, releasing energy in the form of heat or light instead of driving the desired reaction. In organic photocatalysts, poor charge separation and slow electron transfer rates exacerbate this issue . Additionally, the lack of efficient electron transport networks in organic materials can hinder the migration of charge carriers to the catalytic sites, increasing the likelihood of recombination. Tackling this challenge requires the development of advanced materials that can not only promote effective charge separation but also facilitate rapid charge transport to the reaction sites, thereby improving the efficiency of HER processes. |
67780e9e6dde43c90831f476 | 5 | Here, we present the findings of an efficient HER photocatalytic model with limited electron recombination and stable, comparable optical properties. The model demonstrates a robust ability to maintain long-term catalytic activity, even under harsh reaction conditions, by minimizing charge carrier loss and enhancing charge separation efficiency. Furthermore, it is expected that the photocatalyst would exhibit excellent light absorption across a wide spectrum, ensuring optimal utilization while maintaining structural integrity throughout the hydrogen evolution process. This anticipated combination of high efficiency and stability positions the model as a promising candidate for scalable, sustainable hydrogen production. |
67780e9e6dde43c90831f476 | 6 | Phenalene is noteworthy for its propensity to generate stable anions, cations, and radicals. Quantum mechanical analyses have demonstrated that the anion, cation, and radical of phenalene possess equivalent π-electron delocalization energies . The properties have been studied and reported to exhibit remarkable physical properties. Notably, phenalene has been observed to generate a relatively stable anion, cation, and radical . Quantum mechanical studies have provided a satisfactory explanation for the observed stability of these molecules. A comprehensive report on the structural properties noted that the bond distances and their changes in the naphthalene-like fragment are similar to those of phenalene. However, changes in the C-C bond distances involving sp3-hybridized carbon atoms exhibit significant variation depending on the specific location of hydrogen addition. The disruption of aromaticity leads to the elongation of C1-C9 (1.52 Å) and C1-C2 bonds (1.50 Å), both of which correspond to a single C-C bond. This disruption also results in the localization of single and double bonds to C3-C3 (1.47 Å) and C2-C3 (1.34 Å), respectively. Furthermore, a variant with the elongation of the bonds connecting the sp3hybridized C2 is analogous (1.47 Å for C2-C1 and C2-C3), while C1-C9a and C3-C3a correspond to aromatic bonds with distances of 1.40 Å . |
67780e9e6dde43c90831f476 | 7 | Phenalene displays distinctive electrical characteristics attributable to its intrinsic structure, which is distinguished by a delocalized π-electron system . This property renders it a promising candidate for applications in organic electronics. It readily forms stable anions, cations, and radicals, with the negative charge predominantly localized on the central carbon atom in its anionic form. This property signifies its capacity to readily accept electrons and potentially conduct electricity when appropriately doped or modified. The next paragraphs will dwell more on the attractive properties of phenalene. : The delocalized π-electron system in phenalene allows it to efficiently accept electrons, making it an excellent electron acceptor molecule with potential applications in electronic and optoelectronic devices. |
67780e9e6dde43c90831f476 | 8 | Radical stability : Phenalene can form a remarkably stable radical species due to its aromatic πsystem, which effectively delocalizes unpaired electrons. This stability enhances its role in charge transport mechanisms and makes it a promising candidate for organic semiconductors and redoxactive materials. Potential for conductivity: While phenalene is not inherently highly conductive, reports has pointed out its potential for conductivity . Its electrical properties can be significantly improved through chemical doping. Introducing suitable dopants can inject additional electrons or holes into the system, creating charge carriers that enable high conductivity. This tunability makes phenalene a versatile material for use in organic electronic devices. π-conjugation: The extensive π-conjugation across the phenalene molecule facilitates efficient electron delocalization, reducing energy loss and enhancing charge mobility. This property is crucial for its performance in conductive applications, as it allows for seamless charge transport over long molecular distances, contributing to its potential as a material for next-generation electronic and energy devices." |
67780e9e6dde43c90831f476 | 9 | Phenalene demonstrates a high degree of absorption in the visible light spectrum due to its conjugated pielectron system, resulting in a distinctive deep red coloration . Its optical properties are predominantly characterized by its propensity to readily form a stable radical anion (phenalenyl radical) with a delocalized negative charge on the central carbon atom. This radical anion exerts a substantial influence on the absorption and emission characteristics of the phenalene. From the optical perspective, the attractive properties are given below. |
67780e9e6dde43c90831f476 | 10 | Intense color: Phenalene exhibits a striking deep red hue due to its strong absorption in the visible region, particularly at longer wavelengths . This vivid coloration is a result of its extensively conjugated π-electron system, which lowers the energy gap between the HOMO and LUMO, allowing efficient absorption of visible light and contributing to its optical richness. |
67780e9e6dde43c90831f476 | 11 | A defining characteristic of phenalene is its remarkable ability to readily form a stable phenalenyl radical anion . This stability arises from the delocalization of unpaired electrons over the conjugated system, making the radical anion a key player in the unique electronic and optical properties of phenalene. Such radical species have potential applications in charge transport and redox-active systems. |
67780e9e6dde43c90831f476 | 12 | Absorption spectra: The absorption spectrum of phenalene is characterized by a broad band in the visible region, with a peak typically in the range of 500-600 nm , depending on environmental factors such as the solvent, pH, and specific substituents. This broad and intense absorption makes phenalene and its derivatives attractive for optoelectronic and light-harvesting applications. |
67780e9e6dde43c90831f476 | 13 | Fluorescence: While pristine phenalene may exhibit weak fluorescence due to non-radiative decay pathways, its derivatives, particularly those with strategically positioned electron-donating or electron-withdrawing groups, can demonstrate strong and tunable fluorescence. These derivatives emit light across a range of wavelengths, depending on the substituents and their electronic effects, making them useful in sensing, imaging, and light-emitting applications. |
67780e9e6dde43c90831f476 | 14 | Electrochromism: Phenalene's ability to easily form a radical anion under applied potential gives it notable electrochromic behavior . This means that its color can shift dramatically upon electrical stimulation, driven by changes in its electronic structure. This property, combined with its stability and optical tunability, positions phenalene as a promising material for electrochromic devices, such as smart windows, display panels, and adaptive camouflage technologies. |
67780e9e6dde43c90831f476 | 15 | Platinum is widely recognized as an outstanding catalyst for the Hydrogen Evolution Reaction (HER) due to its unparalleled catalytic properties . Its optimal hydrogen binding energy ensures efficient hydrogen adsorption and desorption, enabling high-performance hydrogen production with minimal overpotential. This makes platinum one of the most active and reliable catalysts for HER. However, its high cost and scarcity pose significant challenges, limiting its large-scale application and driving the search for alternative, cost-effective materials. |
67780e9e6dde43c90831f476 | 16 | Optimal hydrogen binding energy: The remarkable HER performance of platinum is attributed to its ideal hydrogen adsorption energy. It strikes a perfect balance between binding hydrogen atoms strongly enough to enable efficient electron transfer while ensuring easy desorption of hydrogen gas, a critical step for sustained catalytic performance. Versatility in different environments: Platinum is highly adaptable, functioning effectively as a catalyst in both acidic and alkaline media. This versatility allows its use across a wide range of HER systems, including those integrated into electrolyzers, fuel cells, and renewable energy technologies. |
67780e9e6dde43c90831f476 | 17 | These qualities underscore platinum's significance as a benchmark HER catalyst, but its limitations in cost and availability continue to motivate ongoing research into alternative materials with comparable activity and greater sustainability. A single atom photocatalyst born of these two has the potential of pushing boundaries for hydrogen evolution without charge recombination barriers. |
67780e9e6dde43c90831f476 | 18 | DMol³ was used to calculate the geometry optimization of the system, employing the gradient descent method as the primary algorithm for minimizing the total energy of the system . Geometry optimization aims to find the equilibrium configuration of a molecular or solid-state system, where the total potential energy is minimized, and the forces on all atoms approach zero. The gradient descent method iteratively updates the atomic positions along the direction of the steepest descent of the energy gradient. Mathematically, the position of an atom at step n +1 is updated based on the gradient of the potential energy E with respect to its coordinates, expressed as in equation ( ): |
67780e9e6dde43c90831f476 | 19 | In the gradient descent method, convergence is achieved when the magnitude of the forces, |∇E|, falls below a predefined threshold, indicating that the atoms are near their equilibrium positions. To improve convergence efficiency, advanced implementations often include dynamic adjustments of the step size α to balance between stability and rapid convergence . For systems with many degrees of freedom, the potential energy surface (PES) can exhibit local minima, and the algorithm is designed to avoid being trapped in these through careful initialization and adaptive corrections. The total potential energy is evaluated using density functional theory (DFT), and the forces (equation 2) are derived as the negative gradient of this energy, F=-∇E. |
67780e9e6dde43c90831f476 | 20 | The Adsorption Locator in Materials Studio calculates the rigid adsorption energy to evaluate how molecules interact with surfaces without considering structural relaxations. The rigid adsorption energy (equation 3), is defined as the energy difference between the total energy of the combined adsorbate-surface system and the sum of the energies of the isolated surface and adsorbate. This is expressed as: |
67780e9e6dde43c90831f476 | 21 | Using the MATLAB console to calculate overpotential is highly convenient due to its robust computational capabilities and ease of handling complex equations. MATLAB provides built-in functions for solving electrochemical models, plotting current-voltage curves, and performing numerical optimization, making it ideal for analyzing overpotential. Its interactive environment allows for quick parameter adjustments and real-time visualization of results, streamlining the iterative process of evaluating and optimizing electrochemical systems. |
67780e9e6dde43c90831f476 | 22 | The bond angles Pt4-C3-C2 (120.625°) and Pt4-C5-C10 (120.624°), which were absent in the pristine phenalene structure, are now present in the Pt-doped phenalene. These newly formed bond angles indicate the significant influence of platinum doping on the molecular geometry, altering the bond network and introducing new structural features not observed in the undoped system . |
67780e9e6dde43c90831f476 | 23 | Absorption Spectra: The prominent peak at 0.233 (Figure ) oscillation evident in the (440-500)nm wavelength range underscores the substantial chromophoric activity of platinum-doped phenalene, attributable to robust π-π* and metal-to-ligand charge transfer (MLCT) transitions. This resonance indicates that platinum dopants considerably augment the light-harvesting capabilities of the phenalene system, thereby optimizing its interaction with visible light. The introduction of platinum into the molecular framework is likely to strengthen localized surface plasmon resonance (LSPR) effects, thereby amplifying the material's ability to generate photoinduced charge carriers. This property is critical for photophysical applications, as the strong absorption in this range underpins its utility in catalytic systems, particularly in hydrogen evolution reactions (HER). |
67780e9e6dde43c90831f476 | 24 | In contrast, the weak oscillation of 0.001 observed in the near-infrared (980 nm to 1000 nm) range indicates a subtle but measurable interaction with lower-energy photons, which may arise from extended vibrational overtones or weak long-wavelength resonance effects introduced by the platinum doping. While not as visible, the interaction in this spectral range could still contribute to catalytic activity. Such resonance could support electronic communication within the system, albeit indirectly. The combined optical features of platinum-doped phenalene impact charge recombination dynamics and HER efficiency. The strong visible light absorption ensures the generation of abundant charge carriers, while the presence of platinum facilitates rapid charge separation via its catalytic active sites, thereby minimizing charge recombination and enhancing the material's quantum efficiency. This ensures that the majority of the photogenerated electrons are utilized in the HER pathway. |
67780e9e6dde43c90831f476 | 25 | The optical characteristics of platinum-doped phenalene impact charge separation, which is important for HER efficiency. The material's catalytic platinum active sites enable rapid separation of charge pairs and minimize charge recombination, increasing the quantum efficiency of the material. The photogenerated electrons are primarily used in the HER process due to this enhanced efficiency. Consequently, the material's tailored optical and electronic properties position it as a highly efficient photocatalyst for sustainable hydrogen production. |
67780e9e6dde43c90831f476 | 26 | The HOMO-LUMO energy difference (Table ) for platinum-doped phenalene (Figure ), ranging from -0.409 eV to -0.285 eV, signifies a remarkably narrow electronic band gap, which has profound implications for its optical and optoelectronic properties . This small gap suggests the material exhibits quasi-metallic behavior, facilitating low-energy electronic transitions and enhancing absorption in the near-infrared to visible spectral regions, making it highly suitable for photonic applications such as photodetectors and solar harvesting. Additionally, the reduced energy gap increases the polarizability of the system, potentially amplifying its nonlinear optical (NLO) responses, which are crucial for frequency conversion and all-optical switching. Furthermore, the ease of electronic excitation and charge transfer implied by this narrow gap could bolster the material's plasmonic behavior, enhancing localized surface plasmon resonance (LSPR) effects desirable in sensing and catalytic applications. The doping of phenalene with platinum not only narrows the band gap but also introduces tunability, allowing for precision tailoring of its optical absorption and emission profiles. Collectively, these properties make platinum-doped pyrene a promising candidate for advanced optical and optoelectronic applications. |
67780e9e6dde43c90831f476 | 27 | Electron Distribution: The electron density distribution of Pt-doped phenalene (Figure ) reveals the strategic exposure of platinum within the molecular framework. The design aimed at optimizing the configuration where platinum atoms are situated at accessible sites. This placement ensures minimal steric hindrance and maximum interaction with the surrounding environment. Additionally, the electronic structure of the material may facilitate charge delocalization by caging the reactant, further enhancing the accessibility of platinum for chemical interactions. Such structural and electronic attributes ensure that the platinum center is accessible within the matrix and are effectively presented for catalytic and electrochemical engagement. |
67780e9e6dde43c90831f476 | 28 | This strategic exposure of platinum significantly enhances the catalytic photoelectrochemical performance of the material. The accessible platinum centers facilitate efficient electron transfer and adsorption of reactant molecules, critical for catalytic reactions like hydrogen evolution or other redox processes. The improved electron density near the platinum sites minimizes overpotentials, thereby increasing the catalytic efficiency and selectivity. Furthermore, the enhanced exposure reduces mass transport limitations, allowing reactants to interact with the active sites more readily. Collectively, this design not only boosts the catalytic activity but also improves the durability and reusability of the material in photoelectrochemical applications, making it a highly efficient and sustainable catalyst. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.