id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
65b25b5be9ebbb4db9f4e1a6 | 36 | Interestingly, thin films are among the least commonly investigated material forms for PSs. To the best of our knowledge, thin-film synthesis has only been reported for six of the 258 experimentally synthesized PSs present in the ICDS database (Fig. ). All of them are thiophosphates: BiPS 4 , Cu 3 PS 4 , Sn 2 P 2 S 6 , Li 3 PS 4 , Na 3 PS 4 , and members of the Li-Ge-P-S system . Relevant optoelectronic characterization was conducted for BiPS 4 and Cu 3 PS 4 , but these films consisted of spin-or drop-cast nanoparticles and may not be fully representative of fully coalesced polycrystalline films. The other films were grown by standard deposition techniques (thermal evaporation or pulsed laser deposition) but they were not optoelectronically characterized. |
65b25b5be9ebbb4db9f4e1a6 | 37 | Despite their very limited presence in literature, we believe thin-film samples are the most convenient option for highthroughput discovery and development of PS semiconductors for optoelectronics. One of the main reasons is that it is relatively straightforward to engineer compositional and/or temperature gradients along two independent directions of the substrate in a single growth run using physical vapor deposition methods. These gradients typically result in material property variations across the substrate, which can be quickly measured by characterization instruments equipped with automated stages. Thus, it becomes possible to collect property-versus-composition or property-versus-processparameter combinatorial libraries with a high throughput across ternary and quaternary systems alike. This combinatorial approach to materials research for optoelectronic semiconductors has already been applied to both thin-film sulfides and thin-film phosphides. In addition to their compatibility with high-throughput methods, thin films grown from controlled deposition sources are usually compact and have smooth interfaces. These features are ideal for optical and electrical measurements. Finally, standard thin-film thicknesses between tens of nm and a few μm are in the optimal range for most PV/PEC absorbers and LED emitters with reasonable properties. In fact, the large majority of solar cell technologies with meaningful efficiencies is based on thin-film PV absorbers, with silicon being a notable exception due to its unusually low light absorption coefficient. Hence, basic materials research on thin-film samples can immediately be translated into device prototypes. |
65b25b5be9ebbb4db9f4e1a6 | 38 | While this article has so far focused on the discovery of new PS compounds, thin-film combinatorial methods are also ideal for studying PS solid solutions, structural disorder, doping by impurities, and the effects of non-stoichiometry. In fact, some quaternary PSs may be more stable in the form of pseudobinary solid solutions, rather than as unique compounds. The GaP-ZnS solution is a well-documented example. Recent CVD work on atomically-thin PSs also demonstrated growth of 5 quaternary solid solutions. Ternary PSs are unlikely to form solid solutions in theory, because S and P have a different number of valence electrons, so charge neutrality is difficult to enforce with a single metallic element. Nevertheless, Li-and Cu-based PS solid solutions have been reported in certain composition ranges. Sulfur is also known to be an effective dopant in GaP.[106] |
65b25b5be9ebbb4db9f4e1a6 | 39 | Our vision for high-throughput experimental exploration of PS thin films is summarized by the cartoons in Fig. . The full synthesis suite consists of three glove box-integrated tools with access to reactive sources of P and S (Fig. ). The tools are a reactive sputter system, a reactive rapid thermal annealing furnace (RRTP), and a simple thermal evaporator. Among them, the main deposition technique is combinatorial reactive sputtering (Fig. ), an established method to create combinatorial libraries by taking advantage of elemental composition-and growth temperature gradients across the substrate. Reactive gaseous sources of S and P are rather uncommon in sputter deposition systems. Nevertheless, the few examples available in the literature have demonstrated high versatility and high material quality. For a generic quaternary PS system M1-M2-P-S (where M1 and M2 are two different metals), a desirable outcome of sputter deposition of PSs is a combinatorial library with a gradient in the M1/M2 ratio in one direction, and in the P/S ratio in the perpendicular direction (Fig. ). A M1/M2 gradient is straightforward to obtain by appropriate orientation of the M1 and M2 targets (Fig. ). Conversely, the P/S ratio gradient is difficult to achieve if both the P and the S sources are non-directional, as for the case of H 2 S and PH 3 gases. A possible solution is to employ compound targets (a metal sulfide and a metal phosphide) rather than metallic targets. However, many of such targets are not commercially available, they have a high cost, a low deposition rate, and the target stoichiometry limits the range of compositions achievable in the final film. In our opinion, a more versatile solution to obtain P/S composition gradients is to employ a directional source for either sulfur or phosphorus, or both. Shown in Fig. ) is the combination of a non-directional PH 3 source for P, and a directional evaporation source for S. Due to the very high vapor pressure of S 8 (the standard solid-state form of sulfur) and its relatively low reactivity at low temperatures, S 8 is first evaporated at a low temperature in an effusion cell. Then, it is cracked into smaller, more reactive molecules in a separate high-temperature zone of the source. By appropriate orientation of the sulfur source, the result shown in Fig. is expected, which gives rise to the desired orthogonal gradients in composition. |
65b25b5be9ebbb4db9f4e1a6 | 40 | There are two remaining potential problems with a PS film synthesis approach exclusively based on reactive sputtering. The first is linked to the volatility of S and P, even when incorporated in solid-state compounds. For many metal sulfides and phosphides, heating in a non-reactive atmosphere leads to decomposition reactions, typically resulting in preferential loss of S in sulfide films and P in phosphide films. Since temperatures significantly above room temperature are required to crystallize most inorganic materials, the decomposition temperature may be lower than the crystallization temperature for a wide range of targeted PS compounds. The solution to this problem is to ensure high partial pressures of S and P during high-temperature crystallization. Since these species are the gaseous products of decomposition reactions, their intentional addition to the gas mix pushes the decomposition reaction towards the left (i.e., towards the desired solid PS compound) due to Le Chatelier's principle. While relatively high concentrations of S and P in the gas phase could potentially be enforced in the sputter system described above, their partial pressures would still be rather low, because the total pressure in a sputter deposition process is limited to the 10 -6 -10 -5 bar range. Reactive annealing of the sputter-deposited combinatorial libraries in a separate RRTP unit with access to gaseous PH 3 and H 2 S sources and an accessible total pressure up to 1 bar can solve this problem. Another advantage of including an RRTP setup is that it is technically much easier to reach temperatures in excess of 1000 °C and avoid corrosion in an RTP furnace than in a sputter system. In fact, an ex-situ annealing step in a S-containing atmosphere is common practice in sulfide thin film synthesis as it can ensure film crystallization without the loss of sulfur. In many labs, transferring a thin film from deposition equipment to reactive annealing equipment requires exposure to ambient air. This limits the applicability of the two-step film growth process (deposition + annealing) to materials that do not react with oxygen or moisture when they are not crystallized or not fully sulfurized/phosphorized. Sacrificial capping layers can be deposited on top of air-sensitive films to slow down their reaction with air,[115] but this adds complication to the process, and at least a small fraction of the capping layer is likely to be incorporated into the final film and may affect its electronic properties. Transfer of thin-film samples between the deposition and the annealing systems in a glove box under an inert atmosphere (Fig. ) can solve these issues. |
65b25b5be9ebbb4db9f4e1a6 | 41 | The second potential problem with PS thin-film synthesis is dealing with M1-M2-P-S material systems where either M1 or M2 is a volatile metal. Such "problematic" metals are Group 1 (alkalis), Group 2 (alkaline earths), and Group 12 elements. Since volatile metals tend to sublime in vacuum at very low temperatures, they are difficult to sputter in a controlled manner and can cause contamination problems that are in some cases irreparable. To address this problem, we propose the inclusion of a very simple thermal evaporator with replaceable parts in the deposition suite (Fig. ). In this separate chamber, Group 2 and Group 12 elements (constituting, say, M1 in the M1-M2-P-S system of interest) can be deposited from metal sources, while the substrate is not intentionally heated. In a thermal evaporator, even Group 1 elements can be evaporated as pure metals by using alkali metal dispensers as sources . The volatile M1 can be evaporated before or after sputter deposition of a M2-P-S combinatorial library, and the resulting M1/M2-P-S film stack can be reactively annealed in the RRTP to aim for crystallization of the M1-M2-P-S material of interest. Due to the typically high solid-state diffusion coefficients in volatile metals, no significant kinetic barriers are expected for the diffusion of the M1 layer into the M2-P-S matrix. Glove-box integration is key for handling and transfer of the air-sensitive Group 1 and Group 2 metal thin films. |
65b25b5be9ebbb4db9f4e1a6 | 42 | A thin-film deposition suite similar to the one described in this section has been designed and is currently being installed in our laboratory. It is our hope that it will allow rapid combinatorial exploration of (almost) any ternary and quaternary PS system with any molar ratio between the desired metals, and with a very wide range of P/S ratios. Clearly, there should be effective safety measures in place before working with such a setup. Among them, the potential release of PH 3 and H 2 S gases should be prevented, reaction products should be neutralized, and materials of unknown toxicity and reactivity should be appropriately handled. |
65b25b5be9ebbb4db9f4e1a6 | 43 | We reviewed trends in composition, structure, and optoelectronic properties of inorganic phosphosulfides, aided by the existing literature and the Materials Project database. Based on this analysis, we assessed their potential as semiconductors for optoelectronic energy conversion applications. We believe that phosphosulfides deserve much closer research attention, as this material family likely contains semiconductors with strong light absorption and favorable charge transport and recombination properties. The main arguments leading to this recommendation are the following: 1) Among the known phosphosulfides, the majority are semiconductors with band gaps in the visible and threedimensional structural networks, indicating their potential to absorb relevant fractions of the solar spectrum and to ensure good electronic transport in all directions. 2) Direct-gap, low-effective-mass semiconductors are already known in ternary phosphosulfide systems (e.g., Ag 3 PS 4 , Ba 4 P 2 S), even though 3D structural networks in PSs do not always translate to low effective masses. |
65b25b5be9ebbb4db9f4e1a6 | 44 | 3) The high chemical versatility of phosphorus in phosphosulfides implies that direct-gap, low-effectivemass semiconductors can be generated by completely different combinations of frontier orbitals. Ag 3 PS 4 and Ba 4 P 2 S illustrate this concept. This is a unique design handle for phosphosulfides, and it could be exploited to engineer defect tolerance by engineering the orbital character of the bands. 4) Phosphosulfides are a vast material space, because of the wide allowed range of phosphorus oxidation states. We estimate that the number of chemically plausible phosphosulfides exceeds the corresponding number of oxysulfides by almost an order of magnitude. 5) Phosphorus-rich phosphosulfides (i.e., those with P/S > 1) have never been synthesized. Nevertheless, the few ones that have been computationally studied appear to be thermodynamically stable and with optimal optoelectronic features. 6) Quaternary phosphosulfides are often thermodynamically stable, and there are concrete examples of quaternaries with improved optoelectronic properties with respect to their ternary constituents. KAg 2 PS 4 is an illustrative case. |
65b25b5be9ebbb4db9f4e1a6 | 45 | We recommend the utilization of high-throughput experimental methods to screen a larger range of the PSs in thin-film form and exploit synergies with the modern arsenal of computational materials science, including high-throughput DFT and machine learning. In concrete terms, we propose a custom suite of thin-film growth setups centered around reactive combinatorial sputter deposition, including options for reactive annealing and incorporation of volatile metals by thermal evaporation. phase nickel monophosphosulfide for the oxygen evolution reaction Dalton Trans. 50 12870-8 |
67bdca3381d2151a02ff5d03 | 0 | Switchable molecular materials are a diverse class of molecular organizations which includes single-molecule magnets (SMMs), spin crossover (SCO) complexes, single-chain magnets (SCMs), metal-to-metal electron transfer (MMET) systems, etc. SMMs are a class of paramagnetic molecular entities displaying slow dynamics in their magnetization. The intriguing complexes show substantial advantages in magneto-structural analysis due to their negligible magnetic exchange interactions and convenient structural regulation. With a major focus on Lnbased SMM, numerous mononuclear complexes incorporating -diketone ligands including dibenzoylmethane (dbm), acetylacetonate (acac), ferrocenoylacetonate, fluoro substituted βdiketone with different bidentate nitrogen donor coligand e.g. 2,2'-bipyridine (bpy), 1,10phenanthroline (phen), 2,2'-bipyrimidine (bpm) and their azacyclo-derivative have been studied to investigate Dy-based SMMs. These studies report that magnetic anisotropy of Dy(III) based complexes is affected by the ligand field strength, local symmetry around the mental centre, etc. |
67bdca3381d2151a02ff5d03 | 1 | Thus, subtle modulations of the ligand via alteration of substituents on the -diketone ligand and/or bidentate nitrogen donor coligand play a substantial impact on the magnetization relaxation dynamics of Dy complexes. Apart from these great efforts, a significant and intense development is still needed to construct a straightforward and reliable magneto-structural relationship in lanthanide-based SMMs, which will develop a better understanding of how ligand field, structural parameters, intermolecular interactions and crystallographic packing alter the dynamic behaviour of structurally analogous lanthanide-based systems. |
67bdca3381d2151a02ff5d03 | 2 | It is important to mention that most of the aforementioned bidentate nitrogen donor coligands belong to the category of α-diimine ligands (Scheme 1). In contrast, β-diimine ligands are rarely studied in compounds of Lanthanides. However, the β-diimine ligands like bik-R (R = Et, Me, etc.) (Schemes 1, right) have not been explored yet in lanthanide chemistry. The more flexible βdiimine ligands form a six-membered chelating ring with the metal centre in comparison to less flexible α-diimine ligands, which form a five-membered chelating ring. Importantly, bik-R ligands have been widely used to study various switchable molecular magnetic materials with 3d-5d transition metals in the recent past. We have lately reported the reversible switching of photo- physical properties of SCO, MMET systems using bik-R ligands where we have shown that the chelating and flexible nature along with the electronic effect of these bik-R ligands can significantly affect the switching behaviour of the associated materials. Thus, we for the first time used the -diimine ligands bik-R (R = Et, Me) in tuning the magnetic anisotropy of the Dy(III) mononuclear complexes, the visualization of their diverse electronic nature is evident from the distinct luminescence and magnetic properties discussed below. The tuning of SMM and stark sublevel-based luminescence properties together within a compound is a challenging task since both properties have opposing requirements. The slow dynamics in magnetization relaxation in an oblate top like Dy(III) is achieved by enhancing the axial symmetry and lowering the equatorial symmetry elements to enhance the axial anisotropy and the energy gap between the ±mj levels of the Kramers doublet which again reduces the charge cloud repulsion of the f-electron density and the electron density of the ligand. On the other hand, the energy difference between the mj levels of the emitter level is lowered by a distorted pseudosymmetry favouring partial admixture of orbitals of opposite parity which favours symmetry forbidden (Laporte forbidden) ff transitions and hence the luminescence (Stark level) thermometric features. In order to observe luminescence, the selection of the ligand system should be made judiciously so that the energy transfer can take place by the 'antenna effect' from the ligand to the emitter level of the Dy(III) centre alongside curbing the thermo-induced back electron transfer. The amalgamation of the luminescence thermometric property with the SMM feature is desirable as it can act as an insitu thermometer, thereby self-monitoring the temperature of the operation and giving an insight into the fundamentals of heat generation which can act as the potential solution of the increasing need for device miniaturization and multifunctionality. This property is highly desirable in modern optoelectronic devices. The speciality of Stark-level thermometry is that its dimension can be reduced to a submicrometric scale; it is contactless and remote sensing, which is essential in the realization of microdevices in the diverse fields of material science, biology and quantum technologies. Moreover, these thermometers are self-calibrated following a ratiometric approach where the anomalies due to intensity fluctuations, i.e. variation of concentration or excitation power, are nullified and thereby the requirement of any external reference is redundant. 48 Luminescent thermometric properties in Dy SMMs are majorly studied with diimine ligand (Table ); in this work, we have explored this optical thermometric property in diimine mononuclear Dy systems. |
67bdca3381d2151a02ff5d03 | 3 | All three complexes were prepared by a simple synthesis pathway (Scheme 2). The reaction of three equivalents of dbm and triethylamine with one equivalent of Dy(NO 3 ) 3 5H 2 O in methanol produced a colourless solution with a pale yellow tinge. Subsequent addition of a methanolic solution of bik-Et ligand into the above reaction mixture formed an off-white precipitate on stirring, which was dissolved in a mixture of dichloromethane and methanol (5:2). Slow evaporation of the solution yielded analytically pure crystals of 1 in good yield (Figure ). |
67bdca3381d2151a02ff5d03 | 4 | Colourless crystals of complex 2 were obtained in a similar method by using the ligand bik-Me instead of ligand bik-Et (Figure ). For complex 3, the reaction of three equivalents of acetylacetone and triethylamine with one equivalent of Dy(NO3)35H2O in methanol produced a colourless solution. The successive addition of a methanolic solution of bik-Et ligand into the above reaction mixture produced a very pale-yellow solution. After filtration, slow evaporation of the filtrate yielded analytically pure crystals of 3 (Figure ). |
67bdca3381d2151a02ff5d03 | 5 | The purity of complexes 1 -3 was confirmed by elemental analysis studies and powder X-ray diffraction measurements (Figures ). The thermal stability of all complexes was estimated by thermogravimetry analyses (TGA) from 303 K to 573 K with a scan rate of 10 K min -1 under a nitrogen atmosphere (Figure ). The TGA analyses of 1 and 2 show that both complexes are well stable up to 500 K, indicating the absence of interstitial solvent molecules which is also consistent with the single crystal X-ray structure studies (vide infra). It is important to mention that the noteworthy thermal stability might be coming from the intense chelating nature of both the dbm and bik-Et/ bik-Me ligands. Importantly, the absence of interstitial solvent molecules offers us an opportunity to explore the effect of ligand only on the magneto-structural relationship of the complexes without any interference coming from the interstitial solvent molecules, which is rarely noticed in SMM studies. In contrast, the TGA curve of complex 3 shows continuous weight loss in the range of 304 -331 K (ca. 4 %) and 340 -375 K (ca. 4 %), suggesting the loss of interstitial solvent molecules. |
67bdca3381d2151a02ff5d03 | 6 | Complexes 1 -3 were characterized fully by solid-state IR spectroscopy at room temperature (Figure ). The IR spectra display the characteristic C=O stretching vibration of the coordinated bik-Et and bik-Me ligands at around 1644 cm -1 . The (CO) of the coordinated dbm and acac ligands were observed at around 1595 and 1587 cm -1 respectively. In addition, IR spectra of all complexes show typical absorptions of the coordinated dbm, acac, bik-Et and bik-Me ligands which are slightly shifted in comparison to the free ligands. |
67bdca3381d2151a02ff5d03 | 7 | Single crystal X-ray diffraction analyses were performed on single crystals of 1 (240 K), 2 (296 K) and 3 (120 K). Crystallographic data and structural parameters are given in Table . Selected bond distances and angles are enlisted in Table The asymmetric unit of complex 3 consists of two crystallographically independent [Dy(acac)3(bik-Et)] units and four methanol as interstitial solvent molecules (Figures and). |
67bdca3381d2151a02ff5d03 | 8 | Temperature-dependent luminescence measurements provide significant fundamental insights by enabling the optical probing of the fine structure of the electronic ground state of Dy(III) ( 6 H15/2), which directly influences the magnetic performance of a single-molecule magnet (SMM). Furthermore, the temperature sensitivity of the spectral features imparts potential applications in thermal sensing. We have performed variable temperature solid-state photoluminescence of the complexes 1-3 in the temperature range of 300 K -15 K (1, 3) and 300 K -13 K (2) along with the free ligands (bik, bik* and acac) at 300 K. On excitation at 320 nm, the emission bands due to the f-f transitions are visible for complex 3 at 300 K, which shows sharp intense peaks at 484 nm and 580 nm corresponding to 4 F9/2 → 6 H15/2 and 4 F9/2 → 6 H13/2 respectively. Upon lowering the temperature to 15 K, the intensity of both the peaks increases with the appearance of very low intense hump at 545 nm ( 4 I15/2 → 6 H13/2) When the spectrum of complex 3 at 300 K was compared to the free ligands (i.e. acac and bik-Et) spectra under similar conditions (Figure ), we observed ligand-centric emissions at 434 nm and 463 nm coming from acac and 468 nm coming from bik-Et ligand is completely quenched in complex 3. This indicates that when complex 3 is excited at the intra-ligand charge transfer band at 320 nm, the coordinated ligands i.e. acac and bik-Et on the absorption of the energy, transfer the electrons from the ground state to an excited singlet state (S0 → Sn) or an excited triplet state (S0 → Tn). From there, the energy transfer takes place to the excited states of Dy(III), namely, 4 I15/2 and 4 F9/2, which act as the emitter states. For complex 1, on excitation at 320 nm, at 300 K, the PL measurement reveals a bik-Et ligand-based emission at 465 nm with several humps between evolution of peaks at 442 nm ( I15/2 → 6 H15/2), 484 nm ( 4 F9/2 → 6 H15/2), 544 nm ( 4 I15/2 → 6 H13/2) and 577 nm ( 4 F9/2 → 6 H13/2) 37, 48, 56 (Figure ). These peaks are distinct from those obtained from the free ligands dbm (432 nm, 466 nm) and bik-Et (468 nm) at 300 K (Figure ). However, the emission contribution from the ligand is not completely quenched, as evident from the broad nature of the spectra. The PL spectrum of complex 2, obtained at 300 K on excitation at 320 nm, shows a bik-Me ligand-based emission at 468 nm and shoulders at 484 nm ( 4 F9/2 → 6 H15/2) and 545 nm ( 4 I15/2 → 6 H13/2). Upon lowering the temperature to 13 K, the PL spectrum shows sharp peaks at 484 nm ( 4 F9/2 → 6 H15/2), 545 nm ( 4 I15/2 → 6 H13/2) and 578 nm ( 4 F9/2 → 6 H13/2), along with the appearance of humps at 422 nm (emission contribution coming from dbm ligand) and 440 nm ( 4 I15/2 → 6 H15/2). Similar to complex 1, the ligand-based emissions are not completely quenched in complex 2 (Figures S15). Additional variable temperature PL measurements were performed by applying an excitation wavelength of 420 nm for complexes 1 -3 (see SI for further details) (Figure and). |
67bdca3381d2151a02ff5d03 | 9 | Comparing the CIE diagram for complex 3 we see a shift in the CIE coordinates from the white region at 300 K to the yellow region at 15 K (Figure ). In complexes 1 and 2, we see a shift from the blue to violet-blue region and a shift from the greenish blue to blue region, respectively, when we increase the temperature from 15 K to 300 K (Figure ). The shifts in the colour coordinates are due to the changes in the spectral patterns of the emission profiles of the complexes on temperature variation and, hence, further reinforce the property of thermal sensitivity of these complexes. |
67bdca3381d2151a02ff5d03 | 10 | The sensitivity of the spectral features toward temperature indicates the thermal sensing capabilities of all three Dy-complexes. The thermometric parameters -luminescent intensity ratio () was estimated as the ratio of I1/I2 ( = I 1 /I 2 ) where I1 and I2 are integrated intensity of emissions, while the thermal sensitivity (Sr) and the temperature uncertainty (𝛿T) were estimated by equations 1 and 2 respectively. |
67bdca3381d2151a02ff5d03 | 11 | was assessed using the thermal sensitivity (Sr) and the temperature uncertainty (𝛿T), key figures of merit for evaluating the efficiency of optical temperature sensors. The temperature dependence of Sr values are depicted in Figures 7, S19 and S20. Sr value decreases with increasing temperature, with a maximum value of 1.09 % K -1 at 15 K and 2.98 % K -1 at 13 K for complexes 1 and 2 respectively. For complex 3, the Sr value remains nearly constant upon increasing temperature to around 200 K, then increases sharply to reach a maximum value of 0.81 % K -1 at 300 K. For complex 1, the Sr value reaches 0.725 % K -1 at 20 K and increases thereafter. applications around the region of magnetic relaxation. The temperature uncertainty (dT) in complexes 1 -3 varied from 0.5 to 0.0 units (Figure ) as identified in accordance with Equation . All the three complexes showed emission profiles sensitive to temperature with complex 2 displaying decent luminescent thermometric properties. |
67bdca3381d2151a02ff5d03 | 12 | The static magnetic properties of 1 -3 were examined on polycrystalline samples in the temperature range of 1.85 -300 K at 1000 and 10000 Oe (Figure and Figures ). The measured T ( is magnetic susceptibility equal to M/H per Dy(III) ion) value for complex 1 is 13.75 cm 3 mol -1 K at 280 K and for complexes 2 and 3 are 14.10 and 14.22 cm 3 mol -1 K respectively at 300 K (Figure ), which are in good agreement with the expected value for one Dy(III) (S = 5/2, L = 5, 6 H15/2, g = 4/3, T = 14.17 cm 3 mol -1 K) ion. The T values do not change significantly down to 100 K for all three complexes. Upon further lowering of the temperature, the T value shows a slow decrease to attain the lowest value of 8.24 cm 3 mol -1 K at 1.9 K for complex 1, 10.93 cm 3 mol -1 K at 2.50 K for complex 2, and 10.92 cm 3 mol -1 K at 1.85 K for complex 3, while a sudden increase in T value to 13.57 cm 3 mol -1 K between 2.5 K -1.85 K was observed for complex 2. The T vs. T measurements under a higher magnetic field of 10000 The relaxation time of the magnetization was deduced as a function of the temperature at 0 Oe for complexes 1 -3 and as a function of the applied dc-field at 8 K for complex 3 from the experimental ′ vs. ν and ′′ vs. ν data fitted to the generalized Debye model (Figures and). |
67bdca3381d2151a02ff5d03 | 13 | The magnetic relaxation is temperature-independent at low temperatures (below 5, 6, and 8 K for 1 -3, respectively), describing the significant dominance of quantum tunnelling of the magnetization (QTM) effect in that region. Above this QTM regime, magnetic relaxation is temperature-dependent and is influenced by the thermal effects. Finally, it follows a thermally activated process at high temperatures. Origin of the paramagnetic relaxation behaviour in SMM, can be defined in terms of magnetization relaxation time () by four main mechanisms: Raman, direct, thermally activated (Arrhenius or Orbach-like), and QTM processes as described in the following equations ( ) and ( ) |
67bdca3381d2151a02ff5d03 | 14 | It is important to note from the above equations that each relaxation mechanism has its own characteristic temperature (T) and field (H) dependence. For all three complexes from the temperature dependence of at 0 Oe, the presence of a clear temperature-independent regime at low temperatures indicates the contribution of a QTM process. In addition, the involvement of a Direct process is excluded in the absence of an external DC magnetic field. Adding only Raman or only Orbach processes is not sufficient. A model considering QTM, Raman and Orbach processes (equation ) is applied to describe the experimental temperature dependence of The best-fit parameters are listed in Table . Table . Best-fit parameters to reproduce the temperature dependence of the magnetic relaxation for 1 -3 using equation 5. and). When compared with Dy-dbm complexes containing -diamine ligands bpy, phen, and their azacyclo-derivative dpq and dppz, complexes 1 and 2 shows much slower relaxation (τ0 -1 ) with significantly higher spin reversal energy barrier value (Ueff). While spin dynamic behaviour of complexes 1 and 2 are very much similar with the Dy-dbm complexes containing -diamine ligands dmbipy (4,4′-dimethyl-2,2′-bipyridine) and 4,5pinenepyridyl-2-pyrazine. 26 When compared with Dy-acac complexes containing -diamine ligands like phen, tmphen, and their azacyclo-derivative dpq and dppz, complexes 3 shows slower relaxation (τ0 -1 ) with slightly higher spin reversal energy barrier value (Ueff), while analogue complexes [Dy(acac)3(dmdophen)] and [Dy(acac)3bpm] shows lower τ0 -1 and higher Ueff. In addition, complexes 3 |
67bdca3381d2151a02ff5d03 | 15 | The insight into the influence of the geometry of the coordination polyhedra and the ligand field on the magnetic properties of 1 -3 has also been evaluated using theoretical methods based on the state average complete active space self-consistent field (SA-CASSCF) calculations followed by SINGLE_ANISO analysis. The calculations were performed with ORCA 5.0 software and the computational details are described in the Experimental section. The respective molecular structures were derived from the experimental X-ray data and only the hydrogen atom positions were optimized. In the case of compound 3, there are two molecular units in the asymmetric unit with dysprosium atoms Dy1 and Dy2, herein labelled as 3a and 3b, respectively. Moreover, there is a potentially significant hydrogen bond of methanol molecules to acac anion of 3b. Previously, some of us and other authors pointed to the significance of the second coordination sphere in SMMs, therefore, the computations were also performed for the complex {[Dy(acac)3(bik- |
67bdca3381d2151a02ff5d03 | 16 | Et)]•2MeOH} of 3, labelled as 3c. The results of CASSCF calculations were also analysed with ab initio ligand field theory (AILFT), which resulted in energies of the f-orbitals of the ground state term ( 6 H15/2) spanning the energy intervals up to ~700 cm -1 as shown in Figure (left). The ligand field multiplets originating from the 6 H15/2 term formed upon applying the spin-orbit coupling, span the energy interval up to ~600 cm -1 (Figure (right)) and the respective energies are listed in Table . CASSCF calculated energy levels of Dy 3+ in complexes 1 -3 upto 25,000 cm -1 is given in Figure 72 , which reveals that the calculated energy for the excited state 4 F9/2 (±5 mj levels) is in the region 24190 -24560 cm -1 and hence energy transfer from the excited state of the ligands can easily take place via antenna effect on photoirradiation which thereby acts as the emitter level in the luminescence phenomenon. The large axial anisotropy of the respective KD states, gz >> gx, gy were found in all complexes 1 -3 (Table ). Interestingly, the easy axes roughly colinear and in 1 the magnetic easy axis of the ground state (KD1) is aligned in an antagonistic fashion to that of KD2 and KD3. Moreover, the transverse components of the g tensor of the KDs are smallest for 2 followed by 3 and 1. This indicates the presence of greater axiality in 2 followed by 3 and 1. Next, the magnetization reversal blocking barriers for 1 -3 are shown in Figure . The equivalent matrix elements of the transverse magnetic moment between the two lowest Kramers doublets (KDs) are small (less than 0.1 unit). This suggests that the temperature-assisted quantum tunnelling probability becomes relevant for the third KD, which has a value of around 170 cm -1 (244.6 K) for 1 -2, and in the case of 3, it varies between 170 (244.6 K) and 216 cm -1 (310.8 K) (Table ). These values are close to Orbach barriers determined from the analysis of AC susceptibility data. Finally, we can conclude that the inclusion of the second coordination sphere, where the two MeOH molecules are hydrogen-bonded to one acac ligand of 3, resulted in higher axiality of respective g-tensors, larger energy gaps (Table , compare 3b and 3c), and lower probabilities for the quantum tunnelling of the magnetization (Figure ). Herein, we performed an additional DFT calculation with PBE0 functional for 3b and 3c and the Mayer bond orders for Dy-O/N coordination bonds and Mulliken atomic charges for {DyN2O6} chromophore are listed in Table . The comparison of the Mayer bond orders and the Mulliken atomic charges shows that the formation of hydrogen bonds to O-donor atoms O12 and O13 of acac ligand resulted in larger atomic charges of these donor atoms, which is coupled to decrease the respective Dy-O bond orders. Thus, Dy-O12/O13 coordination bonds are more ionic in 3c. As a result, we can expect an increase of the orbital reduction factor describing the covalency effect in the ligand field theory ( (3c) > (3b)), which should lead to more efficient spin-orbit interaction (-(L)•S), and larger splitting of the energy levels. |
67bdca3381d2151a02ff5d03 | 17 | for 2). These complexes exhibited distinct emissive properties, giving a value of Sr)max at 1.09 % K -1 (15 K), 2.98 % K -1 (13 K) and 0.81 % K -1 (300 K) for complexes 1 -3, respectively. All three complexes showed zero field SMM behaviour with unique relaxation features, and Ueff is in the range of 129(8) to 257(6) K. Thus, we have successfully visualised the electronic effect of the ligands in tuning the magnetic and optical properties by varying the second coordination sphere. |
661696fc418a5379b08af472 | 0 | Despite the demonstrated potential of vegetable oil as a feedstock for renewable oleochemicals and oil-based biofuels, the low yield of oil per hectare of land by conventional oilseed crops is limiting our ability to scale production. Fermentation-derived oil may provide an avenue to improve yields to meet expected demands for oil-based biofuels and oleochemicals in a growing circular bioeconomy. While high-yielding sugar crops such as sugarcane and sweet sorghum are traditionally fermented to ethanol, they can also be fermented by oleaginous yeast to produce microbial oil (i.e., single-cell oil), mainly in the form of triacylglycerides (TAG) which are suitable for biodiesel production. Oleaginous yeasts naturally produce large amounts of oils from simple sugars when an essential nutrient such as nitrogen, phosphorous, or sulfur is limiting. Rhodosporidium toruloides (i.e., Rhodotorula toruloides), a carotenogenic basidiomycetous yeast, can natively produce oils at titers far greater than many other oleaginous microorganisms (e.g., native R. toruloides can reach 8 g•L -1 while Yarrowia lipolytica, a model organism for oleaginous yeast, natively reaches 4 g•L -1 ) and can natively grow on a wide range of sugars, including glucose, xylose, arabinose, and sucrose. A preliminary techno-economic analysis (TEA) on the potential for microbial oil production from R. toruloides oil estimated a minimum selling price of microbial oil of 4 to 6 USD•kg -1 and a carbon intensity of 7.2 to 11.6 kg CO2-eq•kg -1 from glucose. Another study using sugarcane juice as the feedstock estimated the production cost of microbial oil to be 1.2 USD•kg -1 with optimistic assumptions for capacity (48,000 MT•yr -1 of oil), glucose to oil yield (0.25 g•g -1 ), and oil recovery (95 %). Given that the selling price of soybean oil from 2011 to 2021 was 0.62 to 1.5 USD•kg -1 , microbial oil production from low-value, diluted sugars may be economically feasible with further technological development. A recent study by Bonturi et al. showed that a sugarcane-to-ethanol biorefinery that diverts the hemicellulosic fraction of bagasse towards biodiesel production using R. toruloides can have an economically feasible internal rate of return (IRR) of 14.4%. However, under the same set of assumptions on feedstock price and processing capacity, a sugarcane-to-ethanol biorefinery (without microbial oil production) would have a higher (and, thus, more favorable) IRR of 22.1% (evaluated using BioSTEAM's open-source sugarcane biorefinery model). Given the magnitude of market driven uncertainties and continuous improvements in the performance of microbial oil technologies, there is a need to better understand the sustainability implications of the research, development, and deployment landscape of microbial oil production technologies, including their potential to advance goals for financial viability and greenhouse gas emission reduction from traditional and cellulosic sugarcane biorefineries. |
661696fc418a5379b08af472 | 1 | In addition to sugar-derived microbial oil, another strategy for improving oil production is through the engineering of highly productive C4 crops (e.g., sugarcane, sweet sorghum) to accumulate vegetative oil within plant tissue. Because oleaginous yeast is limited by a maximum theoretical yield of 0.32 g•g -1 TAG from glucose, diverting carbon flux away from plant sugars directly towards oil accumulation may result in greater oil yields per hectare. Recent studies have reported on the creation of oilcane ─ a sugarcane carrying transgenic alleles designed to accumulate TAG in the vegetative tissues ─ that have been demonstrated to achieve up to 4.3 dw % TAG in the stems and 8.0 dw % TAG in the leaves (a total of 5.2 dw % TAG). A detailed TEA suggests that biorefineries could be willing to pay 3.24 [-3.91, 16.42] and 11.7 [-1.9, 32.2] USD•MT -1 more for oilcane than sugarcane with and without conversion of cellulosic biomass to ethanol, respectively. However, oilcane lines with higher oil content have demonstrated lower biomass yields, possibly due to the accumulation of cytotoxic free fatty acids (FFA), which constitute 10 wt % of the total oil. For example, oilcane lines 1566 and 1580, with stem TAG contents of 1.8% and 5.4 dw %, have dry biomass yields of 69% and 19% relative to traditional sugarcane, respectively. Although oilcane may have a greater value than sugarcane per unit mass, trade-offs in biomass yield may undercut the financial viability of oilcane cultivation. |
661696fc418a5379b08af472 | 2 | While oil-accumulating crops and microbial oil may appear as competing technologies, integrating microbial oil production within an oilcane and oil-sorghum biorefinery may result in impactful synergies. Due to the capital burden of supporting two biofuel pathways, configurations co-producing biodiesel and ethanol may not be market competitive when oilcane has an oil content below 5% (assuming no losses in biomass yield). By converting sugars to microbial oil instead of ethanol, both microbial and vegetative oils can be processed together and converted to biodiesel for better economies of scale. Lastly, the integration of microbial oil also presents opportunities for process intensification in oil recovery from both plant and microbial cells. For example, the high temperature (180°C) and pressure conditions in the liquid hot water pretreatment of bagasse aids the release of up to 70% of the remaining oil within plant cells and can potentially be leveraged to disrupt microbial cells for integrated oil recovery. Taken together, integrating microbial oil production at oilcane biorefineries may result in synergies in plant and microbial oil processing for a more sustainable production process. |
661696fc418a5379b08af472 | 3 | The objectives of this study were: (i) to characterize the sustainability of various design alternatives for microbial oil production at a sugarcane biorefinery; (ii) to determine the financial viability and carbon intensity of integrated processing of plant and microbial oils and elucidate the most salient features driving sustainability; and, critically, (iii) to establish targets for oilcane and microbial oil market competitiveness. To this end, we developed and evaluated biorefinery models in BioSTEAM -an open-source platform for the design, simulation, and evaluation of biorefineries-for processing sugarcane and oilcane to biodiesel through fermentation-derived oils with and without cellulosic biomass processing (a total of four biofuel pathways). Model calibration was supported by laboratory experiments to characterize oil recovery from copretreatment of plant and microbial cells. Each biorefinery was evaluated across a landscape of potential technological performance and market conditions to establish research and development targets for sustainable plant and microbial oil production. In particular, the minimum oilcane biomass yield for market competitiveness (as compared to traditional sugarcane) was elucidated across potential oil contents. Cradle-to-grave greenhouse gas emissions were quantified by leveraging BioSTEAM simulation results coupled with inventory data from the Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) and ecoinvent 3.7 . Three different allocation methods (displacement, economic, and energy) were employed to compare the carbon intensity (reported as 100-year global warming potential, GWP100) among the multi-product biorefinery configurations. The minimum biodiesel selling price (MBSP) was used as the main profitability indicator. The sensitivity of the MBSP and carbon intensity to feedstock oil content, product prices, and other techno-economic parameters were evaluated using Spearman's rank-order correlation coefficients, facilitated by Monte Carlo uncertainty analysis. |
661696fc418a5379b08af472 | 4 | Three biorefinery configurations for biodiesel production from plant and microbial oils were developed in this study: (i) fermentation of juice and direct cogeneration (DC) of heat and power from bagasse, (ii) integrated co-fermentation (ICF) of both juice and bagasse hydrolysate, and (iii) integrated co-fermentation and recovery (ICFR) of plant and microbial oil (Figure ). |
661696fc418a5379b08af472 | 5 | [Note: The term fermentation is used here in a way that is consistent with chemical engineering literature, which represents microbial processes -including aerobic processes-for substrate conversion to a specific product in a bioreactor.] Both the DC and ICF configurations closely follow the ethanol/biodiesel-based DC and ICF configurations from our previous work, including the mass balances related to vegetative oil recovery. However, instead of neglecting the cost of managing the vinasse or using a conventional, resource-intensive wastewater treatment process, the new DC and ICF cane-to-biodiesel configurations implement a high-rate wastewater treatment design calibrated using sugarcane and oilcane wastewaters. Both the DC and ICF configurations employ a drum drier and a screw press to mechanically recover the microbial oil, a conventional strategy for oil recovery from algae with an efficiency of 70% (Figure and S1B, respectively, in the Supporting Information, SI). The ICFR configuration sends the cell mass to be pretreated together with the bagasse and then recovers the oil by centrifugation (Figure ). The microbial oil recovery from hydrothermal pretreatment at a solids loading of 50 w/v % was experimentally determined to be 50 wt %. This value was used as the baseline microbial oil recovery for the ICFR configuration. The impact of oil recovery efficiencies and microbial oil production yields on the final yield of biodiesel were calculated and depicted as Sankey diagrams (Figure ). Details on the experimental procedure and results are available in Section S8 of the SI. All configurations were modeled in BioSTEAM ─the Biorefinery Simulation and Techno-Economic Analysis Modules─ which automates the design, simulation, and techno-economic analysis of thousands of potential scenarios. Figure 1. Simplified flowsheets of the (A) direct cogeneration (DC), (B) integrated cofermentation (ICF), and (C) integrated co-fermentation and recovery (ICFR) configurations with microbial oil production. In all configurations, either sugarcane or oilcane is crushed to release the juice, the oil in the fermentation effluent is recovered through a 3-phase decanter centrifuge, and a heat exchange network (HXN) integrates heating and cooling of process streams to decrease utility demand. In the DC configuration, the bagasse is sent directly to the boiler to produce heat and power. In the ICF and ICFR configurations, the bagasse is pretreated with liquid hot water, hydrolyzed with cellulases, and co-fermented with the juice. In the DC and ICF configurations, the oil in the cell mass is recovered mechanically with a screw press after drying. |
661696fc418a5379b08af472 | 6 | In the ICFR, the oil in the cell mass is recovered by centrifugation after cellulosic pretreatment with the bagasse. Previously published biorefinery designs co-producing ethanol and biodiesel were envisioned to process oilcane with oil contents between 5 to 15 dw % to overcome trade-offs in capital investment that stem from the equipment necessary for ethanol and biodiesel production. The biorefinery configurations presented here are alternative designs based on our previous work that eliminate ethanol production in favor of microbial oil production. |
661696fc418a5379b08af472 | 7 | We intentionally varied the oil content in oilcane for uncertainty and sensitivity analyses, specifically examining the impacts of market conditions, design decisions, and potential longterm technological performance. To characterize the implications across the spectrum of observed oil content and yields, we varied oil content uniformly from 1.8 to 5.4 % and biomass yield uniformly from 4.87 to 25.62 dry MT•ha -1 (the low yield from oilcane 1580 and the high yield from sugarcane). We modeled the effect of oil content on the overall composition in manner that was consistent with previously published work. Briefly, an energy balance approach was used to balance an increase in oil content with a decrease in sucrose content. |
661696fc418a5379b08af472 | 8 | Because oil is about 2.5 times as energy-dense as sucrose, the loss of biomass from increasing oil content was compensated by fiber. In the uncertainty analysis of oilcane lines at the current state of technology, the biomass yield and the oil, sugar, fiber, and moisture content were sampled from a truncated normal distribution with mean and standard deviation extracted from published data (Table , SI). |
661696fc418a5379b08af472 | 9 | Bioindustrial-Park GitHub repository, an official repository for complete biorefinery models and results to foster accessibility and deeper communication within the biorefinery simulation community. The repository also includes flowsheets, stream tables, utility requirements, design requirements, itemized costs, cash flow analysis, and reaction stoichiometries for each biorefinery. Detailed results of all oilcane and sugarcane biorefinery configurations are available in the Bioindustrial Park at Park/blob/master/biorefineries/cane/README.md. Note that the full repositories can be downloaded directly from Github without the need to create an account. Alternatively, the biorefinery models can be installed using the command "pip install biorefineries==2.28.0" from your terminal (Mac/Linux) or command line (Windows) with a working Python 3.10 installation. |
661696fc418a5379b08af472 | 10 | The thermodynamic property package estimates phase equilibrium using modified Raoult's law with activity coefficients estimated through Dortmund UNIFAC interaction parameters. Modified Raoult's law is suitable to estimate phase equilibria of nonideal mixtures at low to moderate molecular weights and pressures up to 10 atm. Therefore, this phase equilibrium model can accurately predict the interactions of the highly polar, low molecular weight compounds present in the biorefinery (e.g., water, methanol, glycerol). The lignocellulosic components are modeled following the pure component property values used by NREL (heat of formation, molecular weight, etc.), with the exception that liquid densities and heat capacities for lignocellulosic components are held constant at 25 ⸰ C. Pure component properties (heat capacity, density, etc.) of fluids are estimated using higher order polynomial fits to the fundamental Helmholtz equation of state (a state-of-the-art property prediction model), recommended correlations from critical reviews on thermodynamic properties, and equations from well-established public references such as the CRC handbook of chemistry and physics and the NIST Webbook. Mixture properties are estimated using a molar weighted average of the pure chemical properties. This property package has been validated against a cellulosic biorefinery model by NREL implemented in Aspen Plus and both sugarcane and oilcane biorefinery models in SuperPro. 11 |
661696fc418a5379b08af472 | 11 | Aerated bioreactors, including heat exchangers, compressors, and agitators, constitute a major capital investment and utility expenditure in a bioprocess. For example, the cost of the microbial oil bioreactor system alone has been estimated to cost 17.5 million USD at 3x the baseline fermentation productivity considered in this study and at a production capacity comparable to the DC configuration. Because the economic design of an aerated bioreactor is critical for financial viability, a full discussion on the design and costing algorithm is warranted. |
661696fc418a5379b08af472 | 12 | The production of microbial oil is an aerobic process that can be modeled as the sum of three stoichiometric reactions: oil production from sugar, respiration, and cell growth. The aerobic fermenter is assumed to have a high heat-production rate of 110 kcal per mol of O2 consumed, consistent with published literature. The nitrogen present is a complex mixture of nucleic acids, protein, and other reduced forms of nitrogen; as such, the conversion of these components to cell mass is neglected. Magnesium sulfate is provided in the bioreactor model to provide Mg and S to support cell growth. These are also excluded from the growth reaction because the Mg and S contribution to the cell mass is minimal. Any substrate not converted to cell mass or oil is assumed to be consumed for respiration. Given the yield of TAG, Yp (g•g -1 ), and the fraction of biomass produced over the total substrate not consumed for oil production (Yb; g•g -1 ), the extent of reaction for TAG production (Xp), growth (Xb), and respiration (Xr), can be calculated (Table ). |
661696fc418a5379b08af472 | 13 | The molecular formula of yeast is assumed to be CH1.61O0.56. While fed-batch and draw-fill operation modes lead to higher yields and titers compared to standard batch operation, the mode of operation in this study is limited to batch due to the high nitrogen content of the substrates. Nitrogen serves as the limiting nutrient for growth and its absence drives microbes to accumulate oil during the production phase. The C:N ratio (by weight) of the oilcane bagasse hydrolysate and juice were measured to be 40 and 54, respectively, which are far lower than the optimal for the production phase (~120 C:N ratio): as a result, continuous feeding would stop the microbes from entering the production phase. The starting C:N ratios of the substrates are closer to the optimal starting C:N ratio for batch operation (70). It may be possible to supplement with hydrolysates from other feedstocks with lower C:N ratios (e.g., wood chips or pure sucrose) to enable other modes of operation, but these were not considered in this study due to the complexity of the design and logistics. |
661696fc418a5379b08af472 | 14 | The total power consumed by the fermentation systems was minimized numerically by varying the flow rate of air provided by the compressor and the power consumed by the agitators under the constraint that the oxygen uptake rate is equal to the oxygen transfer rate at a dissolved oxygen concentration of 50% saturation. The power consumed by the agitator was 0.06 kW•m -3 under the baseline fermentation performance, which is sensible compared to the heuristic value recommended for industrial homogeneous reactions (0.3 kW•m -3 ). The overall mass transfer coefficient, kLa (lumped together with the specific interfacial area), was estimated using Van't Riet's non-viscous mass transfer correlation. The design procedure accounts for the pressure gradient across the vessel due to the liquid head using the log-mean driving force for mass transfer. The uncertainty analysis considers a range of titers, productivities, and product yields from the baseline performance on hydrolysate (13.2 g•L -1 titer, 0.17 g•L -1 •h -1 productivity, 0.132 g•g -1 yield) to an optimistic batch performance achieved on pure glucose substrate (27.4 g•L -1 titer, 0.31 g•L -1 •h -1 productivity, 0.18 g•g -1 yield) . Multi-effect evaporation of the juice and hydrolysate is used to achieve the required titer in each scenario. At the lower and upper ends of performance, Yb does not change significantly (0.39 and 0.42 g•g -1 , respectively). To achieve a continuous design space, we assumed this parameter has an average value of 0.405 g•g -1 . |
661696fc418a5379b08af472 | 15 | Assuming a yield of 0.132 g•g -1 , the overall stoichiometric reaction on a molar basis for microbial oil production from oilcane hydrolysate becomes: The maximum volume of each aerated bioreactor was 500 m 3 , a "world's largest" class that exists in small numbers. The number of vessels is computed based on the influent flow rate, reaction time, and cleaning time (3 h). |
661696fc418a5379b08af472 | 16 | Previous TEAs have assumed the price of oilcane to be the same as the price of sugarcane with identical growth and harvest requirements. However, the yield per hectare is lower for oilcane lines than for sugarcane and the price of oilcane by mass will be higher. The Sugarcane Production in the Louisiana worksheet is a decision tool for the estimation of the cost of sugarcane cultivation that incorporates key assumptions for field preparation, planting, harvesting, and other operations. This tool has been leveraged for estimating the price of energy cane by adjusting the biomass yield of plant cane and stubble crop (i.e., ratoon) rotations. A key assumption embedded in the tool and used in this study is that the cost of farming sugarcane -including field expenses and variable production expenses -per unit land is independent of biomass yield. Assuming the cost of transporting sugarcane has a linear relationship with distance, the cost of transportation per unit biomass was estimated to be proportional to the square root of the total area. In the same spirit, the carbon intensity of oilcane production was estimated by summing the contributions from cultivation and transportation using extracted impacts for cultivation and transportation of sugarcane from GREET . Details on feedstock price and carbon intensity calculations are available in Section S9 (SI). |
661696fc418a5379b08af472 | 17 | The LCA was performed following the same procedure used in our previous study. Briefly, the system boundary extends from cradle to biorefinery-gate for the operational phase and does not include the construction phase of the biorefinery. All operational inventory data originated from the biorefinery models leveraged in this study (Table ; SI). Life cycle inventory data for each raw material and ancillary input were adapted from the GREET 2020 model and the ecoinvent 3.7 life cycle inventory database. The life cycle impact assessment methodology used was the Intergovernmental Panel on Climate Change (IPCC) 2013, with a focus on GWP100 as the primary indicator (reported herein as carbon intensity). Three different allocation methods were employed to compare carbon intensity across biorefinery configurations: displacement, economic, and energy allocation (a detailed breakdown of the LCA results using each allocation method is provided in Section S4 of the SI). Because the biorefinery configurations presented here mainly produce biofuel, we focus our discussion on results using energy-based allocation of carbon intensity. The only co-product with alternative end uses is crude glycerol. While crude glycerol is produced in significant quantities, the choice of allocation factor under either economic or energy allocation (minimally 1.18% and maximally 4.53%; Tables and) does not change the conclusions of this study (Figures , and D; SI), in agreement with the low energy content (12 MJ•kg -1 ) and low price (0.1 USD•kg - 1 ) 23 of crude glycerol as compared to biodiesel (37 MJ•kg -1 and 0.50 USD•kg -1 , minimally). |
661696fc418a5379b08af472 | 18 | All capital cost correlations and parameters used in the discounted cash flow analysis follow the assumptions made in our previous study. Because all biofuel configuration pathways were found to achieve over 60% reduction in carbon intensity relative to a 2016 petroleum benchmark established by the EPA (see the Results and Discussion section), the return on investment included the value of renewable identification numbers (RINs) of biomass-based diesel (i.e., D4 RIN), and cellulosic biofuel (i.e., D3 RIN) pathways following the guidance of 75 FR 14863. As specified by 40 CFR 80.1451(b)(1)(ii)(U), the fraction of fuel (biodiesel) produced from the planted crop feedstock (sugarcane juice) received D4 RINs while the remaining fraction originating from the crop residue feedstock (bagasse) received D3 RINs. RIN credits were treated as a co-product and decreased the minimum biodiesel selling price (MBSP). |
661696fc418a5379b08af472 | 19 | Carlo simulations (Section S3; SI). Parameter distributions for oil composition, cellulosic pretreatment performance, ethanol production performance, operating days, IRR, co-product prices, and feed prices follow the same assumptions as our previous study. Fuel and electricity prices were updated to include the latest market prices from the USDA Economic Research Service, 10 the U.S. Energy Information Administration, and the U.S. EPA. Latin hypercube sampling was used to generate 5,000 shared scenarios which were used to evaluate each oilcane line and configuration. At each oil content, the competitive biomass yield was characterized across 200 scenarios. Halving the number of scenarios resulted in no significant change in uncertainty and sensitivity results. The sensitivity of the MBSP and carbon intensity to each input parameter was characterized by Spearman's rank-order correlation, a measure of monotonicity between input and output parameters. |
661696fc418a5379b08af472 | 20 | The ICFR configuration makes use of the high temperature and pressure conditions in the cellulosic pretreatment of bagasse to disrupt the cells and release microbial oil, removing the need for energy and capital-intensive drying and extrusion of the cell mass (which are required in the ICF configuration). Although integrating oil recovery constitutes a process intensification by reducing the number of unit operations, the pretreatment reactor itself is energy and capitalintensive and results in comparable costs under the same assumptions for oil recovery (Figure ). The MBSP from all configurations were sensitive to microbial oil yield and microbial oil recovery, the latter of which was dampened (i.e., made less sensitive) through the burning of unrecovered oil for heat and power cogeneration. The cellulosic configurations (i.e., ICF and ICFR) produce larger quantities of microbial oil and exhibit greater sensitivity to oil recovery than the DC configuration. In addition to oil recovery and microbial yield, the productivity and the maximum achievable titer are also key technological drivers of sustainability (Figures and; SI). At the baseline case, which represented conservative assumptions for fermentation performance which have been achieved at lab-scale using oilcane hydrolysate, Although biodiesel production from microbial oil may not be financially lucrative under these assumptions, it results in greater yields and a lower carbon intensity than biodiesel from soybean oil. Even under uncertainty in potential operating days, microbial oil yields, and oil recoveries, this result holds true for 100 % of evaluated scenarios (Figure ). Specifically, the DC and ICF yields were 1780 [1310, 2360] and 2750 [2050, 3640] L•ha -1 •yr -1 , respectively, which is 2 to 6 times greater than the yield of biodiesel from soybean oil (617 L•ha -1 •yr -1 ). soybean oil, points above the horizontal grey line achieved higher biodiesel yields and points to the left of the vertical grey line had lower carbon intensities. |
661696fc418a5379b08af472 | 21 | Across all parameters, both the economic and environmental sustainability of oilcane is most sensitive to the dry biomass yield (Figure and S6; SI). Although increasing oil content at the expense of a modest reduction in biomass yield can result in a net financial benefit, the best performing oilcane line (balancing oil content and biomass yield; Oilcane 1566) and the oilcane line with the highest oil content (Oilcane 1580) have such large reductions in dry biomass yield (relative to sugarcane) that they have similar or greater MBSPs and carbon intensities compared to sugarcane (Figure ). The reduced biomass yield of current oilcane prototypes is a result of high constitutive expression of lipogenic genes WRI and DGAT, which lead to either elevated levels of toxic free fatty acids (caused by high WRI expression) or depletion of phosphatidylglycerol needed for thylakoid membranes (caused by high DGAT expression). Higher biomass yields and lipid contents may be achieved if alternative regulatory elements are implemented to fine-tune the spatial and temporal expression of lipogenic genes. Research in the development of oil-accumulating feedstocks may direct efforts towards regulating lipogenic gene expression to sustain both high biomass yield and oil contents. In addition to oilcane, researchers are currently developing new lines of oil-accumulating sweet-sorghum 57 (up to 3 dw % oil in stems) which have not exhibited growth penalties (field trials still on-going) and oilaccumulating energycane (up to 1.5 dw % TAG in leaves) which natively have greater biomass yield than sugarcane. Once field trials are completed for these new feedstocks, future assessments may seek to evaluate their sustainability for the production of biofuel and bioproducts. |
661696fc418a5379b08af472 | 22 | Given that vegetative oil is 1.9 times as carbon-dense as glucose per unit mass, we can expect that producing biodiesel directly from an oil crop can afford reductions in biomass yield compared to a process producing biodiesel from a sugar crop via fermentation. To better inform research and development on critical targets for biomass yield as a function of oil content, we determined the competitive biomass yield under uncertainty across potential feedstock oil contents (Figure ). Under assumptions for the best theoretical performance of microbial oil production (i.e., 100% of glucose carbon converted to microbial oil, minimal fermentation costs, and a microbial oil recovery as high as plant oil recovery), a 1% increase in oilcane oil content can be accompanied by (on average) a 1.21 % reduction in biomass yield (relative to the baseline yield of sugarcane) to achieve equivalent profitability. In reality, due to the need for cell growth and respiration (not all carbon is converted to oil) and the significant costs of aerated bioreactor conditions in microbial oil production, the actual affordable loss in oilcane biomass yield can be much greater. While the biomass yield for oilcane 1566 (17.7 DMT•ha -1 •y -1 ) is 31 % lower than the sugarcane wildtype, this biomass yield is just 0.56 % lower than the 5 th percentile of the competitive biomass yield (21.6 [17.8, 22.6] DMT•ha -1 •y -1 in the ICF configuration; Figure ). |
661696fc418a5379b08af472 | 23 | Therefore, reasonable improvement in either biomass yield or oil content may enable this oilcane line to present economic advantages over sugarcane at a biorefinery producing microbial oil. On the other hand, the oilcane line with the highest oil content (oilcane 1580) has a biomass yield (4.9 DMT•ha -1 •y -1 ) that is 62.0 % lower than the 5 th percentile of the competitive yield (16.0 [12.9, 18.6] % in the ICF configuration). At such a low biomass yield, only improving the biomass yield may result in financial competitiveness. Assuming a target of engineering highly productive C4 crops with 10 dw % oil content by 2027, we propose to establish a lower limit in biomass yield to conservatively dictate whether the engineered oil crop is more financially viable than its non-engineered counterpart. A prospective "target" oilcane line could have an oil content of 10 dw % and biomass yield of 18.0 DMT•ha -1 •y -1 (the maximum competitive biomass yield in the DC configuration; Figure ). |
661696fc418a5379b08af472 | 24 | The oil content and yield of oilcane lines along with the titer, productivity, and yield of microbial oil production are the main economic bottlenecks that require improvements to achieve market competitiveness. Greater oil recoveries would also be impactful. However, it would require further research and development of new oil recovery configurations (e.g., supercritical |
661696fc418a5379b08af472 | 25 | We evaluated the impact of stepwise improvements to feedstock and fermentation technologies for the DC configuration along four key indicators of sustainability: MBSP, total capital investment (TCI), carbon intensity, and biodiesel yield (Figure ). The MBSP serves as an indicator for profitability while the TCI serves as an indirect measure of financial feasibility due to the inherent difficulties of financing capital-intensive projects with emerging technologies. We focus our discussion on the DC configuration because not only is it the most financially viable, but it also has a lower carbon intensity than the ICF configuration (Figure ; SI). Taking oilcane 1566 and the demonstrated oilcane fermentation performance as the baseline, all scenarios at the current state of technology have MBSPs significantly above the market price range (Figure ). Evaluated improvements in both fermentation technologies and feedstock oil content and yield would reduce the MBSP such that 20% of evaluated scenarios lie within the market price range of biodiesel (Figure ). Further improvements to the biomass yield (up to 25.6 DMT•ha -1 •y -1 ; equivalent yield to sugarcane) would bring down the MBSP such that 87% of evaluated scenarios fall within or below the market price range (Figure ). Therefore, assuming key improvements to oilcane biomass yield and oilcane fermentation, biorefineries with integrated plant and microbial oil processing can be market competitive at 10 dw % oil content. The TCI decreases with improvements to fermentation performance and cane oil content (Figure ) and increases with cane biomass yield under the assumption of constant land availability (as expected from greater feedstock processing capacity; Figure ). Regardless of technological improvements, the TCI remains 3 -7 times as large as the TCI of soybean biodiesel (83.9 10 6 •USD; Figures ). ). Compared to biofuel production from soybean, this outcome is equivalent to 3.0 to 3.9 as much biofuel per hectare of land and a 57 to 63% reduction in carbon intensity. If biofuel policy would incentivize the land usage and environmental benefits of integrated microbial and vegetative oil-based biofuel pathways (as compared to conventional oil seed pathways), a greater fraction of scenarios would readily fall within the market price range of biodiesel. Furthermore, these policy incentives could be implemented as tax exemptions and deductions to reduce the capital investment and, in turn, lower the economic barrier for deployment. |
661696fc418a5379b08af472 | 26 | The integrated plant and microbial oil processing pathway developed in this study (available as installable open-source applications in Python) could be modified to produce a wide range of oleochemicals (e.g., thermoset plastics through ozonolysis and plasticizers through epoxidation) and biofuels (e.g., green diesel and sustainable aviation fuel) with lower environmental impact and land usage than soybean oil-based pathways. Biomass-based pathways are critical for achieving the Sustainable Aviation Fuel (SAF) Grand Challenge to supply 3 billion gallons of SAF per year by 2030 and replace 100% of conventional jet fuel by 2050. However, the use of feedstocks such as corn (through the alcohol-to-jet pathway) and soybean (through hydroprocessed esters and fatty acids) for SAF production is linked with high land use expansion that limits our ability to scale. The configurations for plant and microbial oil processing developed in this study may provide an avenue for even greener and more scalable biomass-based SAF. Further techno-economic and life cycle assessment studies are needed to quantify the potential benefits of plant/microbial oil pathways to oil-based biofuels and oleochemicals other than biodiesel. |
60c74984bb8c1a03413dae4d | 0 | The mechanism of natural photosynthesis can conveniently be divided into several distinct steps : i) light harvesting in antenna proteins, ii) a fast primary electron-transfer event in the reaction center (RC), iii) protein stabilization due to charge separation, and iv) subsequent steps involving the synthesis of new chemical compounds. In the first step, the photosynthetic system does not undergo any chemical transformations and simply transfers the absorbed energy to the RC. The RC contains (bacterio)chlorophyll [(B)Chl] molecule(s), which constitute the primary electron donor. For the cases of green plants, algae, and cyanobacteria, two primary electron donors are known: P700 in photosystem I (PSI) and P680 in photosystem II (PSII) . Bacterial reaction centers (BRCs) containing the electron donors P870 and P960 can be found in Rhodobacter sphaeroides and Rhodopseudomonas viridis, respectively . Note that the numbers correspond to the absorption maxima of those complexes. The primary electron donors of PSI and the BRC are pairs of inner (B)Chl a , which are called the special pair (SP) in cases of BRCs (for an overview on RC structures, see Sec. 2). In case of PSII, the primary electron donor is composed of an accessory Chl a co-factor . Being excited by the delivered energy, these primary electron donors donate an electron forming short-lived radical pairs with neighboring co-factors (for more detail, see Refs. ). In the third step, a chain of secondary electron-transfer reactions is initiated preventing the recombination of radical pairs back to neutral compounds and stabilizing the separation of charges. Thus, the absorbed energy is converted into electro-chemical potential, which is further used for synthesis of chemical compounds. |
60c74984bb8c1a03413dae4d | 1 | Although (B)Chl molecules in an inner pair are structurally very similar, they often have very different properties. The nearby protein environment could break the symmetry in the electronic structure of the pair, thus considerably changing its behavior during the electron donation. In case of a small asymmetry, this dimer could participate in electron-transfer processes as a supermolecule, whereas a strong asymmetry causes it to act as a pair of independent and weakly-coupled subunits. The degree of the electronic-structure asymmetry in inner pairs and corresponding radical cations, formed during the primary electron donation (for PSI and the BRC) or subsequent charge separation reactions (for PSII), is a frequently discussed issue in the field and received vast attention in the literature (for example, see Refs. ). In order to assess the asymmetry of the electronic ground state, chemical shifts of (B)Chl a molecules are often considered , whereas spin-density distributions are used for cation radical states of inner pairs . Solid-state Photochemically Induced Dynamic Nuclear Polarization Nuclear Magnetic Resonance (photo-CIDNP NMR) gives an access to both molecular properties. This methodology can be used to reconstruct spin-density maps of radical pairs inside RCs of plants and bacteria and can be applied to whole living cells or even entire intact plants . |
60c74984bb8c1a03413dae4d | 2 | Another strategy is to access the s-orbital contributions to the spin density from measured hyperfine coupling constants (Fermi contact term) . To that end, Electron Paramagnetic Resonance (EPR) and related methods such as Electron-Nuclear Double Resonance (ENDOR), Electron Spin Echo Envelope Modulation (ESEEM), and Electron-Nuclear-Nuclear Triple Resonance (commonly referred to as TRIPLE) can be applied to single photosynthetic pigments as well as to primary electron donors . |
60c74984bb8c1a03413dae4d | 3 | Theoretical studies are especially important for investigating the spin-density asymmetry as they provide a direct access to electronic structures of photosynthetic pigments and can be applied to obtain spin densities and hyperfine coupling constants of many shortlived open-shell intermediates produced in electron-transfer reactions. However, reliable calculations of inner pair spin-density distributions may require the consideration of the surrounding protein environment. This considerably increases the size of molecular models to be computed and restricts the choice of the electronic structure method to Kohn-Sham Density Function Theory (KS-DFT). The latter is affected by the self-interaction error (SIE) and often produces overly delocalized spin-density distributions . This effect can be especially strong for non-covalently bonded molecular systems (for examples, see Refs. ) and, therefore, is expected to be crucial for pigments in RC models. Moreover, the degree of spin delocalization may strongly depend on the exchangecorrelation (XC) functional applied . Although many different approaches were proposed in the literature to overcome this problem , the ultimate solution is not yet found and the issue remains relevant. |
60c74984bb8c1a03413dae4d | 4 | One of these proposed solutions is based on the use of the Frozen-Density Embedding (FDE) formalism and allows to effectively avoid the spin-density overdelocalization error at the intermolecular region due to an embedding potential applied. This methodology seems especially attractive due to the possibility to consider much larger molecular systems than generally accessible with KS-DFT. However, its use leads to an opposite problem of overly localized spin densities. This drawback was recently lifted with FDE-diab , which extended and generalized the previously known Electron-Transfer Frozen-Density Embedding method (FDE-ET) . The underlying idea of FDE-diab lies in the construction of charge-and spin-localized quasi-diabatic states and computing electronic couplings between them. A similar strategy was successfully applied in various electron-transfer calculations earlier and proved to be very effective. The FDE-diab methodology was validated for a number of dimeric complexes featuring different degrees of the spin-density localization and was shown to be a more reliable and robust approach for spin-density calculations than standard KS-DFT . In this work, we present FDE-diab calculations of spin densities and electronic couplings for RC models of PSI, PSII, and bacteria from Synechococcus elongatus, Thermococcus vulcanus, and Rhodobacter sphaeroides, respectively. We aim at reliable assessments of the spin-density asymmetry in inner pair radical cations and validation of previously proposed photosynthetic models. To that end, we consider a number of molecular systems ranging from dimeric complexes in vacuum to large protein and investigate roles of relative co-factor arrangements. |
60c74984bb8c1a03413dae4d | 5 | In what follows, we briefly introduce RC architectures for PSI, PSII, and bacteria in Sec. 2. The underlying theory of the theoretical approach used is described in Sec. 3. In Sec. 4, we provide details on model setup and computations conducted followed by results presented in Sec. 5. Further discussions and conclusions are given in Sec. 6. |
60c74984bb8c1a03413dae4d | 6 | In the following, we limit our consideration to molecular structures of the RCs from PSII, PSI, and bacteria characteristical to the Thermococcus vulcanus, Synechococcus elongatus, and Rhodobacter sphaeroides photosynthetic organisms, respectively. The corresponding crystal structures were taken from Protein Data Bank (PDB) entries 3WU2 , 1JB0 , and 1M3X . Similar RCs found in other organisms as well as mutants of the ones mentioned above are outside the scope of this paper and are not discussed here. |
60c74984bb8c1a03413dae4d | 7 | PSII is a protein complex located in the thylakoid membrane of oxygenic photosynthetic organisms. It is composed of two protein sub-units D1 and D2 (see Fig. ). The membrane of PSII contains 36 trans-membrane α-helices. Five of these helices belong to the proteins D1 and D2 each. Six of those belong to the proteins CP43 and CP47. CP43 and CP47 are called the core antenna proteins that bind 13 and 16 Chl a molecules, respectively. Overall 11 β-carotene units can be found in the antenna proteins. The central magnesium atom of the majority of the Chl a molecules in the antenna proteins is coordinated by a histidine (His) residue. Additionally, the tetrapyrrole macrocycle of these Chl a molecules is not exactly planar, but slightly bent out of the macrocycle plane. The proteins D1 and D2 contain the co-factors that are associated with the RC of PSII. The D1 and D2 sub-units are distinguished by the presence of the so-called Oxygen Evolving Complex (OEC) near the sub-unit D1 (not shown in Fig. ) and other structural aspects. The role of this manganese cluster is the oxidation of water and formation of oxygen . |
60c74984bb8c1a03413dae4d | 8 | The RC contains six Chl a molecules, of which four are believed to participate in the electron-transfer processes: i) the inner P D1 and P D2 co-factors composing the inner pair structurally similar to the SP in the BRC (depicted in green in Fig. As seen from Fig. , PSI (just like PSII) also contains two protein sub-units, which are nearly C 2 -symmetric. Its inner pair (shown in green) is formed by one Chl a molecule and a C-13 2 -epimer of Chl a (often referred to as Chl a ), which are denoted as P B and P A , respectively. It is well-known that P A forms several hydrogen bonds with amino acids from the surrounding protein environment, while P B does not . The two accessory pigments are Chl a molecules and called A -1A and A -1B (blue). Similar to the case of PSII, the magnesium atoms of the inner co-factors P A and P B are coordinated by His molecules, whereas the magnesium atoms of accessory pigments A -1A and A -1B are coordinated by water molecules. Instead of Phe a molecules, the RC of PSI contains two Chl a co-factors, which are referred to as A 0A and A 0B (red) and feature the central magnesium atoms being coordinated by sulfur atoms of methionine residues. In addition to the above-mentioned pigments, this RC also includes two phylloquinone co-factors A 1A and A 1B (violet) and an [4Fe-4S] cluster F x (brown). |
60c74984bb8c1a03413dae4d | 9 | In the BRC, the SP of pigments is formed by BChl a molecules, which are denoted as D A and D B (green). Similar BChl a molecules compose the pair of accessory co-factors B A and B B (blue). It is interesting to note that the central magnesium atoms of these four pigments are all coordinated by axial His molecules. In contrast to the case of PSII, the BRC features a pair of BPhe a molecules instead of Phe a, which are called H A and H B and depicted in red in Fig. . The pair of quinone molecules is represented by menaquinone Q A and ubiquinone Q B . Additionally, the BRC contains, just like in PSII, a Fe(II) ion (brown) located between the Q A and Q B co-factors. PDB entries 3WU2 , 1JB0 , and 1M3X . For the color code used, see the main text. |
60c74984bb8c1a03413dae4d | 10 | The underlying theory of the FDE-diab approach will be briefly repeated here for the specific case of two non-covalently bound molecules A and B forming a radical cation complex [A • • • B] +• . This rather simple consideration has a direct connection to CT processes often occurring between two pigments in photosynthetic systems and is sufficient enough for dimeric molecular models presented in this work. Note, however, that generalization to multiple fragments is straightforward from a mathematical point of view. In these radical cation complexes, charge and spin are delocalized to some extent between both molecules. The spin-density ratio can range from a complete delocalization (i.e., 50%/50%) to a localization at only one fragment. The former case can be found in |
60c74984bb8c1a03413dae4d | 11 | , while the latter case is characteristic for fragments with different ionization energies and large intermolecular separations. As was mentioned above, a description of the radical cation electronic structure in general and prediction of spin-density delocalizations in particular may be problematic with both the KS-DFT and FDE approaches. However, the tendency of FDE to overlocalize charge and spin density can be used to construct an approximate electronic wave function Ψ of the radical cation complex [A • • • B] +• , which provides a qualitatively correct description of the electronic structure. This can be achieved by computing two wave functions Φ 1 and Φ 2 corresponding to the charge-localized states |A +• B and |AB +• and taking the total electronic wave function Ψ as a linear combination, |
60c74984bb8c1a03413dae4d | 12 | with linear combination coefficients a and b. This basis of two quasi-diabatic wave functions Φ 1 and Φ 2 can then be used to evaluate the elements of overlap and Hamilton matrices to subsequently solve a generalized eigenvalue problem. The resulting eigenvector elements a and b define the total wave function Ψ and, thus, give an access to calculations of various molecular properties of the radical cation complex [A • • • B] +• . In the following sections, we briefly describe the FDE-diab methodology and show how accurate electronic couplings, long-range excitations, and spin densities can be calculated from the total wave function. |
60c74984bb8c1a03413dae4d | 13 | ] with positive charges localized at the molecules A and B, respectively. In these calculations, sets of molecular orbitals (MOs) ψ I,σ i ( r ) of separate subsystems A, B, A +• , and B +• and their corresponding energies I,σ i are obtained from the so-called Kohn-Sham equations with constrained electronic density (KSCED) , |
60c74984bb8c1a03413dae4d | 14 | where -∇ 2 2 is the one-electron kinetic energy operator, while υ I,σ KS ( r ) and υ I,σ emb ( r ) are the one-electron KS and embedding potential, respectively. The superscript I denotes the subsystem under study and σ = α or β are used as spin labels. The charge-localized states Φ 1 and Φ 2 are then constructed as direct products , |
60c74984bb8c1a03413dae4d | 15 | The use of this monomer basis set is advantageous for large molecular systems as it allows for considerable savings in the overall computational cost. It should be noted that the KS-like orbitals forming Slater determinants of the charge-localized states Φ 1 and Φ 2 , are in general not orthogonal, which leads to non-orthogonality of these states. The overlaps for the latter are given by ] |
60c74984bb8c1a03413dae4d | 16 | needs to be solved in the basis of two charge-localized states {Φ 1 , Φ 2 }. Here, H and S are the 2×2 matrices of the Hamiltonian and overlap, respectively, whereas c is a column vector of the linear combination coefficients. The elements S nm of the overlap matrix are given according to Eq. ( ), whereas the elements of the Hamiltonian are calculated approximately as functionals of the transition electronic densities ρ nm ( r ) = ρ α nm ( r ) + ρ β nm ( r ), scaled by the overlap S nm , |
60c74984bb8c1a03413dae4d | 17 | This approximation is invoked to obtain both diagonal and off-diagonal matrix elements in a consistent way using approximate exchange-correlation (XC) functionals (for more details, see Ref. ). The α-and β-components of the electronic transition density ρ nm ( r ) can be obtained as eigenvalues of the one-electron σ-density operator ρσ , |
60c74984bb8c1a03413dae4d | 18 | and solve the generalized eigenvalue problem from Eq. ( ). This gives us energies E 0 and E 1 of the resulting ground (Ψ 0 ) and excited (Ψ 1 ) state with the corresponding sets of linear combination coefficients {a 0 , b 0 } and {a 1 , b 1 }. |
60c74984bb8c1a03413dae4d | 19 | FDE-ET introduces an additional approximation to the generalized eigenvalue problem described above enabling the use of an analytical expression for electronic couplings. This can be achieved by i) assuming that S 11 = S 22 , ii) dividing the off-diagonal element S 12 by √ S 11 S 22 (to obtain S 12 ), and iii) omitting the term S nn on the right-hand side of Eq. ( ) for the case of diagonal elements such that |
60c74984bb8c1a03413dae4d | 20 | In our previous study in Ref. , we outlined in very detail how the spin density ρ α-β ( r ) can be obtained from the total wave function Ψ. The idea behind this is rather simple and is based on the direct evaluations of the α-and β-density contributions as eigenvalues of ρσ , |
60c74984bb8c1a03413dae4d | 21 | from Eq. ( ). This rather trivial fact has, however, a very large importance as it allows for DFT-type calculations of various molecular properties that are derived from the spin density, e.g. hyperfine coupling constants, with potentially higher quality than those from conventional KS-DFT. The higher quality of FDE-diab spin densities compared to those of conventional KS-DFT was already validated in previous work . Whether this also holds for spin density derived properties remains to be tested. |
60c74984bb8c1a03413dae4d | 22 | Molecular models for PSI, PSII, and the BRC were extracted from crystal structures of Synechococcus elongatus (PDB entry 1JB0), Thermococcus vulcanus (PDB en-try 3WU2), and Rhodobacter sphaeroides (PDB entry 1M3X), respectively, obtained from the Protein Data Bank . Missing hydrogen atoms of photosynthetic co-factors were added with the Reduce program , whereas the hydrogen atoms of the protein environment were added using CHARMM22 topology files . To saturate cut protein bonds, neutral groups -C(O)CH 3 and -NH 2 were applied as N-and C-termini, respectively. In these capping groups, bond distances were initially set to experimentally determined average values characteristic for small organic compounds . Two types of models were considered in this work: i) models, in which the long hydrocarbon tails R of Chl molecules (see Fig. ) were left unchanged and ii) models with the tail R truncated and replaced by a -CH 3 group. In case of dimeric molecular models such as the inner pair, sets of modified molecular structures with different orientations of the fragments were created. Thus, the initial crystal structures were adjusted by making mutual vertical and horizontal shifts of chlorophyll molecules. For this purpose, coordinates of truncated chlorophyll rings (without hydrogens and substituents at aromatic cycles) were fitted by a pair of parallel planes minimizing vertical distances to the atoms. Then, vectors of perpendicular and parallel shifts were determined. For the latter case, we chose vectors lying within the parallel planes and oriented along the Mg→C-5 directions of the corresponding molecules. In the following, we denote distances between these parallel planes as R ⊥ , while vertical and horizontal shifts are referred to as ∆R ⊥ and ∆R , respectively. Note that in case of horizontal shifts ∆R , coordinates of both aromatic macro-cycles are modified simultaneously by the corresponsing translation vectors of lengths ∆R . Binding pocket models of PSI, PSII, and BRC considered in Sec. 5.3 were created by specifying radii of 4.0 Å around each atom of the inner pair. All surrounding co-factors, water molecules, and amino acid residues with at least one atom within these radii were included explicitly into the models. Addition of missing hydrogens and cut bond saturation were conducted in a similar fashion as for smaller models considered in this work (see above). Amino acid residues were considered to be neutral and protonated in the created binding pocket models. Chl a BChl a To distinguish the created inner pair models with and without phytyl tails, we introduce label "t" for truncation. For example, [tP D1 • • • tP D2 ] +• should be understood as the truncated complex of P D1 and P D2 , while its non-truncated counterpart is denoted as Created molecular structures for neutral dimeric models in vacuum were optimized in two subsequent steps using the Orca program package. At first, the positions of hydrogens were relaxed keeping all other atomic coordinates fixed. Secondly, the structure optimization was carried out for the set of intramolecular bond distances. Two atoms A and B were considered as bound if the distance between them is in the interval of r(A)+r(B)±t, where r(A) and r(B) are covalent radii of the corresponding atoms and t = 0.5 Å is a tolerance parameter. This procedure ensured that structural parameters are partly optimized, while the relative orientation of fragments is still very close to that in the original crystal structure. Note that the second step does not imply any explicit constraints to bond and dihedral angles. In fact, they can slightly change during the optimization procedure. In both steps, the BP86 XC functional and the def2-TZVP basis set were applied. To reduce the computational cost, the resolution-ofthe-identity approximation in conjunction with the auxiliary Coulomb-fitting def2-TZV/J basis was enabled. Dispersion interactions were taken into account with the D3BJ correction with Becke-Johnson damping . "Loose" convergence criteria were applied for inner pair models. These optimized molecular structures of neutral complexes were then used in subsequent single-point calculations of the corresponding radicals. |
60c74984bb8c1a03413dae4d | 23 | For geometry optimizations of binding pockets as described in Sec. 5.3, the DFTB3 method within the Ams-Dftb module from the Adf 2019 package was used. The Third-Order Parametrization for Organic and Biological Systems (3ob) parameters from the corresponding Slater-Koster file were used. Dispersion interactions were taken into account with the D3BJ correction with Becke-Johnson damping . The optimization was carried out in two consecutive steps. In the first step, the coordinates corresponding to hydrogen atoms were optimized, while all other nuclear positions were kept fix. In the second step, only the nuclear coordinates corresponding to the inner pair were optimized. This procedure ensures to keep the relative arrangement of environment residues and co-factors unchanged, while the inner pair coordinates are relaxed in the presence of the surrounding pocket. Default convergence criteria were applied for the optimization of hydrogen positions, whereas "loose" criteria were used in the second step of geometry optimization. |
60c74984bb8c1a03413dae4d | 24 | For FDE-diab and FDE-ET computations, locally modified versions of the Adf program and the PyAdf scripting framework were used. All calculations employed the TZP basis set from the Adf program library and the PW91 XC functional with the conjoint kinetic-energy functional PW91k . Mutual relaxations of subsystem densities were taken into account applying freeze-and-thaw cycles . Three such cycles were found sufficient to obtain accurate electronic densities (see also Ref. ). |
60c74984bb8c1a03413dae4d | 25 | In FDE-diab and FDE-ET calculations of dimeric systems such as inner pair models, KSlike MOs of both molecules were used for the construction of diabatic states. For binding pocket models, MOs of only inner pair Chls were explicitly included into the FDE-diab step, while interactions with MOs of environment molecules were taken into account by means of orbital polarization during preceding freeze-and-thaw cycles. |
60c74984bb8c1a03413dae4d | 26 | During this procedure, grid points were assigned to particular nuclei forming overlapping atomic basins. Spin-density values at grid points were extracted with the Densf program from the Adf package and then integrated over the generated basins giving atomic spin-density populations. The reported spin-density delocalization ratios were obtained by summing up these atomic contributions for the corresponding molecules. The error introduced by numerical integration was found to be below 5.0×10 -4 a.u. in all cases studied. For comparison of this integration scheme with standard Bader and Mulliken population analysis, see Sec. S3 in the Supporting Information (SI). |
60c74984bb8c1a03413dae4d | 27 | In Sec. 5.1 we provide a detailed explanation of spin densities obtained within the two-state model for a better understanding of the following result sections. In Sec. 5.2, physically motivated dimeric molecular models for the electronic donors of PSI, PSII, and BRCs are calculated in vacuum. The influence of the protein environment is then considered in Sec. 5.3. |
60c74984bb8c1a03413dae4d | 28 | The first important difference to be mentioned is the spatial gap between the co-factors composing the inner pair. If we define this gap through the distance R ⊥ between parallel planes going through the aromatic macrocycles of the co-factors (for a more robust definition of these planes, see Sec. 4), then the largest separation between inner (B)Chl molecules can be found in the crystal structure of PSI and is equal to about 3.9 Å. |
60c74984bb8c1a03413dae4d | 29 | A slightly smaller value of 3.7 Å can be found in case of PSII . The SP of the BRC features the smallest distance R ⊥ = 3.1 Å between the co-factors . Therefore, one might expect that the interaction energy between the co-factors is the largest for the It is also interesting to note that the BChl a macrocycles in the crystal structure of BRC are essentially planar, while a strong out-of-plane bent is present for both PSI and PSII inner pairs. These spatial arrangements of co-factors are largely governed by interactions with the surrounding protein pockets, which are often arranged in an asymmetrical way around the constituting inner pair molecules and interact differently with those. Thus, the inner co-factors of PSII interact strongly with the surrounding protein, whereas P A and P B are reported to be essentially undisturbed . |
60c74984bb8c1a03413dae4d | 30 | These experimental measurements predicted 82% of the unpaired spin to be localized at the P D1 monomer, while the 18% left were assigned to P D2 . The authors of Ref. conducted a 13 C photo-CIDNP MAS NMR experiment assigning most of the measured signals to a single Chl a co-factor. The observed spin-density distribution showed a pronounced asymmetry (shift towards rings III and V) within Chl a compared to the results obtained in solution (largest contribution at ring II) . This asymmetry was later confirmed in a subsequent experiment . Finally, 15 N photo-CIDNP MAS NMR measurements carried out in Ref. detected signals originating from an axial histidine molecule leading to the formulation of the hinge-type model for the primary electron donor of PSII. Therefore, according to the experimental results the inner pair of PSII is a weakly-coupled pair with the largest spin-density contribution localized at the P D1 co-factor. The remaining spin-density is probably localized at the axial His molecule or shared between His and P D2 . |
60c74984bb8c1a03413dae4d | 31 | Early EPR and ENDOR measurements of PSI reported a fully symmetrical spindensity distribution between both monomers in P700. However, more recent work proposed a large range of asymmetric spin-density distributions ranging from 75%/25% to about 91%/9% in favor of P B and even completely localized at P B distributions. Thus, ESEEM experiments by Davis et al. proposed that the inner pair in spinach PSI is either a dimer with the spin-density ratio from 75%/25% to 80%/20% or a single monomer. The ratio of 75%/25% was confirmed by Rigby et al. using ENDOR and special triple (ST) spectroscopy. Later, 2D-ESEEM measurements for 15 N-labeled P700 and ESEEM analysis of single crystals of PSI reported spin-density distributions of about 90%/10% and 87.5%/12.5%, respectively. In an extensive study by Mac et al. including ESEEM analysis, isotopic substitution, and numerical simulations of the ESEEM data, it was shown that the spin density of P700 + is completely localized at one Chl a molecule, which interacts with the surrounding protein environment via axial ligation or hydrogen bonds. The complete localization was also reported by Lubitz and co-workers employing ENDOR spectroscopy and site-directed mutation on P700. |
60c74984bb8c1a03413dae4d | 32 | According to this work, the spin density is localized at the P B co-factor. Later, theoretical calculations by Plato et al. confirmed the asymmetric spin-density distribution in the inner pair of PSI. The ratio was found to be 75%/25% in favor of P B . It was also proposed that this asymmetry is not caused by the structural difference of isolated P B (Chl a) and P A (Chl a ), but by the interaction of the co-factors with the protein environment. This assumption was verified by calculating molecular models including a different number of amino acid residues from the surrounding protein environment. This showed an increasing degree of the spin-density asymmetry for larger molecular models. In case of photo-CIDNP MAS NMR experiments, measured 13 For the BRC of Rhodobacter viridis, an asymmetric spin-density distribution of about 66%/33% in favor of the D A co-factor was reported in Ref. . In case of Rhodobacter sphaeroides, fully-assigned EPR and ENDOR spectra were not presented for a long time. |
60c74984bb8c1a03413dae4d | 33 | were also confirmed by ESEEM measurements conducted for the primary electron donors P870 from Rhodobacter sphaeroides and P960 from Rhodopseudomonas viridis . A somewhat different spin-density ratio of 60%/40% was reported later in Ref. based on theoretical KS-DFT computations. Photo-CIDNP MAS NMR measurements carried out for the selectively 13 C-isotope labeled Rhodobacter sphaeroides WT predicted the spin density to be strongly delocalized over the two BChl a molecules of the SP . |
60c74984bb8c1a03413dae4d | 34 | Interestingly, in this work some small amount of the spin density was also detected at an accessory BChl a molecule. It is, however, generally accepted that no spin-density contributions are present at the axial His molecules as the opposite was not observed in experimental investigations conducted . Hence, the BRC SP in the radical cation state is a strongly coupled, slightly asymmetric pair of molecules with the largest contribution localized at the D A co-factor. |
60c74984bb8c1a03413dae4d | 35 | , where A is P A , P D1 or D A , while B is P B , P D2 or D B . The case of truncated radical cation models in the electronic ground state is summarized in Fig. , while the results for their non-truncated counterparts are given in Fig. . As can be seen from Fig. (top), inner pairs of PSI, PSII, and BRC show very different spindensity distributions. At ∆R ⊥ = 0 Å, the ratios are equal to 96%/4%, 80%/20%, and 72%/28% for the inner pairs in PSI, PSII, and the BRC, respectively. Hence, FDE-diab predicts most of the spin-density to be localized at the P A , P D1 , and D A co-factors of the corresponding pairs. The results obtained for PSII and the BRC are in a very good agreement with experimental measurements , where 82% and 66% of the spin-density were found to be localized at P D1 and D A , respectively. In case of PSI, however, most of the spin density is expected to be localized at P B , while the calculated ratio shows the opposite result. This can probably be explained by the inability of FDE-diab to reproduce the correct order of nearly degenerate electronic states in PSI, which are separated by only 0.14 eV. The spin-density ratio in the first excited state of the PSI inner pair is equal to 4%/96% (see Fig. ) and, thus, shows the anticipated strong localization at P B . Another reason for the exchanged order of electronic states could be the complete neglect of co-factor-protein interactions in the presented FDE-diab calculations. The results obtained for non-truncated models are generally very similar (see Fig. ) to those |
60c74984bb8c1a03413dae4d | 36 | In cases of PSII and the BRC, the effect of truncation is stronger and leads to changes in the spin-density ratios of about 4%. The resulting equilibrium distance ratios of the PSII and BRC are equal to 76%/24% and 68%/32%, respectively. Despite the observed discrepancy in the electronic level ordering for the inner pair of PSI, the obtained results |
60c74984bb8c1a03413dae4d | 37 | As expected, at larger intermolecular separations [see Fig. (top)] spin-density distributions become more and more localized. The complete localization is observed at displacements ∆R ⊥ larger than 1.0 Å. This behavior is caused by the absolute magnitudes of the electronic coupling decreasing exponentially with the intermolecular separation. Interestingly, the calculated spin-density ratios for PSI, PSII, and BRC show a qualitatively different behavior at shorter distances. The inner pairs of PSI and PSII have more localized distributions at ∆R ⊥ = -0.5 Å than at the equilibrium distances. |
60c74984bb8c1a03413dae4d | 38 | Contrary to that, the spin-density of the BRC reaches a ratio of 62%/38%. This, however, is not observed for the non-truncated molecular models, which all feature a larger degree of delocalization at shorter intermolecular distances (see Fig. ). It was expected that the overlap of the inner pair co-factor densities and, as the result, the spin-density delocalization degree could both be increased by mutual parallel displacement of aromatic macro-cycles. In case of the BRC, this effect can, indeed, be observed for both truncated and non-truncated models as seen from Fig. |
60c74984bb8c1a03413dae4d | 39 | In these models, all residues within 4.0 Å from the corresponding inner pair were calcu- This redistribution leads to a better agreement with the experimental ENDOR results predicting 82% of spin being localized at P D1 . Based on these calculations and on experimental measurements reported in Ref. , it can be assumed that the remaining spin-density distribution is probably localized at P D2 rather than at the axial His. In order to verify this assumption, larger molecular models, which explicitly include two histidine molecules coordinated at the inner pair Chls, need to be calculated with FDEdiab. However, the current implementation of the FDE-diab approach does not allow us to calculate couplings between more than two diabatic states and, therefore, prevents us from conducting these calculations. Comparing experimental photo-CIDNP MAS NMR results and theoretical calculations of protein binding pockets in Figs. ) and d), respectively, it can be seen that most of measured signals are successfully reproduced with FDE-diab. |
60c74984bb8c1a03413dae4d | 40 | However, somewhat small spin populations are calculated for carbons C-3, C-7, and C-13, while only negligible contributions are obtained at C-10 and C-15. This does not directly contradict the photo-CIDNP MAS NMR results depicted in Fig. ) as relative spin populations were not provided in this experimental work . However, it does not support the conslusions drawn by the authors of Ref. . Based on the measured signals, it was proposed in that reference that most of spin density is shifted towards ring C of Chl a and the spin distribution, thereby, appears to be strongly asymmetric. |
60c74984bb8c1a03413dae4d | 41 | The calculated spin-density distributions for the inner pair in the binding pocket do not show a strong degree of asymmetry compared to the single Chl a molecule in vacuum as can be seen from Fig. ) and d). This, however, can be a consequence of an insufficiently large size of the binding pocket considered as well as of a complete neglect of axial His MOs in diabatic state construction. These aspects will be investigated in future work, where larger multi-state molecular models will be considered. |
60c74984bb8c1a03413dae4d | 42 | The Chl a spin-density populations in the monomer and in the [P A • • • P B ] +• dimer are qualitatively different (see Fig. ). The spin population of the monomer is very similar to those calculated earlier for the chlorin and truncated Chl a molecules in Ref. . These spin densities were largely dominated by the highest occupied molecular orbital (HOMO), transforming according to the irreducible representation of point group 2 A 2 in case of chlorin and of point group 2 A for the truncated Chl a molecule . The distribution in Fig. ) is, therefore, closely related to this single 2 A 2 -like MO. Contrary to that, the spin population in Fig. ) combines features of both HOMO ( 2 A 2 -like) and HOMO-1 ( 2 B 1like) MOs of chlorin and is, therefore, strongly mixed. This mixing can be explained by the fact that the molecular structure of [P A • • • P B ] +• was only partially relaxed, while a full structure optimization was carried out for the monomer. This assumption is supported by experimental results in Fig. ) as well as by FDE-diab calculations for a fully optimized inner pair in Fig. ). In both cases, the spin density is found to be similar to that for the monomer. The mixing observed for the dimer in vacuum is, therefore, artificial and is, indeed, caused by the limitations of the structure-optimization protocol applied. |
60c74984bb8c1a03413dae4d | 43 | Comparing results for the inner pair in vacuum and in the protein environment as seen in Figs. ) and d), a redistribution of the spin density from P A to P B in the latter case can be observed. The resulting spin-density ratio in the binding pocket is equal to 24%/76% in favor of P B . This strong localization at the P B co-factor is in good agreement with experimental EPR and ENDOR measurements , where 75% to 100% of spin density was found at the P B co-factor. The calculated ratio in the binding pocket also agrees very well (deviations within 1%) with previous theoretical computations reported in Ref. and supports the assumption that asymmetry in the inner pair of PSI is induced by the surrounding protein environment. As can be seen from Figs. ). This observation cannot be validated by comparison with experimental results as, to the best of our knowledge, spin population data is currently unavailable in literature for the BRC. However, the surrounding protein environment does influence the resulting spin-density ratio. Thus, the ratio changes from 68%/32% for the dimer in vacuum to 93%/7% for the SP in the protein binding pocket, showing much stronger localization at the D A co-factor in the latter case. This result is rather surprising taking into account that ENDOR and TRIPLE experiments as well as previous KS-DFT calculations predict 66% and 60% of spin density to be localized at D A , respectively. However, it agrees with photo-CIDNP experiments reporting a strong localization of the spin density in the SP of Rhodobacter sphaeroides . A definite conclusion on the correct spin-density localization in the SP of the BRC cannot be drawn on the basis of the presented results and would require much larger molecular In this work, we presented an application of the recently developed FDE-diab approach to RC models from PSII, PSI, and purple This metodology allowed us to avoid the consequences of the DFT overdelocalization error in the intermolecular regime and to reliably calculate spin distributions and electronic couplings for a number of RC models. |
60c74984bb8c1a03413dae4d | 44 | The spin-density ratios calculated for the isolated inner pair were equal to about 80%/20% and 76%/24% in favor of the P D1 co-factor for the trucated and non-truncated models, respectively. Inclusion of the nearby protein environment led to even more delocalized distribitions with the spin-density ratio of 65%/35% in favor of P D1 . These results are in good agreement with the experimental ENDOR measurements predicting 82% of spin to be localized on P D1 . Taking into account large electronic couplings between the inner pair co-factors, we can assume that the remaining spin density is probably localized at the P D2 molecule. This, however, contradicts the hinge-type model for the RC in PSII suggested in Ref. and suggests the delocalization of the spin-density distribution between P D1 and the axial His molecule. Calculations of spin-density ratios between two axial His molecules and the inner pair require the use of MOs of all four molecules in the diabatic-states construction procedure, and are not possible in terms of two-state models. This requires a generalization of the FDE-diab approach to multiple electronic states. Although most of experimentally measured spin-density populations of the P D1 co-factor in PSII were successfully reproduced by us in FDE-diab calculations of large protein binding pockets, the proposed strong asymmetry in the spin-density pattern was also not observed. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.