id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
61bad290d6dcc2026c3a4b10 | 2 | Only a few UV-emitting species has been employed in TTA-UC to date, with 2,5diphenyloxazole (PPO) arguably gaining the most attention. Pioneering work by the Castellano group dating back to 2009 employed PPO together with biacetyl, albeit with very low efficiencies. It is only as of 2021 that a system employing PPO surpassed 10% in FUC, which was achieved by pairing PPO with a cadmium sulfide nanocrystal sensitizer decorated with 3phenanthrene carboxylic acid. The 10% limit has also been surpassed by pairing an iridium complex or a ketocoumarin derivative with 1,4-bis((triisopropylsilyl)ethynyl)naphthalene (TIPS-Naph), systems which also demonstrate low threshold excitation intensities (Ith). Other annihilators previously investigated include other naphthalene and oxazole derivatives, species from the terphenyl family, and as of recently also a biphenyl derivative with the capability of emitting light beyond 4 eV. The spin-statistical factor, f, gives the probability that an excited annihilator triplet state ultimately ends up as a singlet excited state following TTA. In annihilator species in which the second triplet excited state (T2) is energetically accessible during TTA, f takes the value of 2/5 for strongly exchange-coupled triplet pairs, which caps FUC to 20%. Annihilators yielding significantly higher FUC, such as a few based on perylene, have been shown to have f » 1 since T2 has too high energy to be populated following TTA. This classical way of approaching the spin-statistical factor has recently been questioned, suggesting that a broader range of values could be achieved and which depends on e.g., the nature of the initially formed triplet pair states. In this study, we aim to shine light on the fundamental aspects currently limiting vis-to-UV TTA-UC. A thorough and systematic investigation of both known, relatively efficient, annihilator species as well as two compounds that has not been used in this context previously has been performed. The six annihilators used here are paired with a high triplet energy thermally activated delayed fluorescence-type (TADF) sensitizer, allowing for efficient population of also highly energetic annihilator triplet states. We show that also vis-to-UV TTA-UC systems may approach the spin-statistical limit of 20%. Specifically, employing TIPS-Naph as the annihilator species yields a record-setting FUC of 16.8% (out of a 50% maximum), which is an almost 2-fold improvement on the previously best-performing vis-to-UV TTA-UC systems. High FUC are also obtained for PPO (14.0%), 2,5-diphenylfuran (PPF, 13.0%), a compound never used for vis-to-UV TTA-UC before, and for p-terphenyl (TP, 12.6%), a compound which emits much deeper in the UV. The performance of the remaining systems are also evaluated and the intrinsic properties governing the TTA-UC process are obtained and analyzed. Further, we discuss what implications these findings have and what obstacles still need to be overcome in order to improve these systems for future application in photochemical settings. |
61bad290d6dcc2026c3a4b10 | 3 | Photophysical Characterization. The annihilators under investigation herein are presented in Figure alongside their respective absorption and fluorescence spectra. PPO, TIPS-Naph, TP, and 2,5-diphenyl-1,3,4-oxadiazole (PPD) have all been used for TTA-UC previously, while PPF and 2-phenylindene (2PI), to the best of our knowledge, are demonstrated as annihilators for the first time. These compounds all emit UV light efficiently, albeit with non-unity quantum yields (Table ), but their respective first singlet and triplet excited state energies are quite different, spanning 3.5-4.0 eV (singlets) and 2.1-2.8 eV (triplets, Table ). Even though this study is primarily conducted in toluene as the solvent, the absorption spectra in Figure are measured in THF since toluene absorption interferes with the spectral shape of annihilator absorption below 290 nm. To make comparison between annihilator species as feasible as possible we chose to use only one sensitizer. While cadmium sulfide nanocrystals have previously been used to sensitize high triplet energy annihilators such as PPD, their notoriously complex photophysics, the need for additional mediating compounds, and suboptimal performance when paired with annihilators with elevated triplet energies 30 caused us to search for molecular sensitizers with the capability to sensitize all annihilators used herein. |
61bad290d6dcc2026c3a4b10 | 4 | We focused our attention on the recently emerging group of TADF sensitizers, and found that the purely organic, blue-emitting 2,3,5,6-tetra(9H-carbazol-9-yl)benzonitrile (4CzBN), developed by the Zhang group, was able to sensitize all annihilators efficiently. TADF molecules exhibit small singlet-triplet energy splittings (DES-T, typically below 0.3 eV) which results from a high degree of intramolecular charge-transfer (CT) character in the singlet and triplet excited states. In 4CzBN this is manifested by the CT absorption band with an onset at around 430 nm (Figure ), and the covalent linkage between electron donor and acceptor units further enhances the CT character. The photophysics of organic TADF compounds has been thoroughly investigated by others, and the key processes of a conventional TADF compound are depicted in the left part of Figure . Upon excitation, the first singlet excited state can decay either non-radiatively or by prompt fluorescence. Because ISC is quite strong in these purely organic molecules TADF compounds also populate their triplet state efficiently via ISC. DES-T then dictates how fast reverse ISC (rISC) proceeds, which together with the rates for non-radiative triplet decay and phosphorescence dictates the lifetime of the triplet state. The recycling of singlet and triplet states ultimately results in thermally activated delayed fluorescence (TADF) from the singlet state, typically on the microsecond timescale. 4CzBN exhibits both prompt fluorescence and TADF in toluene. The fluorescence quantum yield (FF) and lifetime (t) of the prompt (PF) component were determined from measurements in air-saturated samples, while the delayed (DF) component was readily observed in oxygen-free samples. The total FF of 4CzBN was determined as 0.64, with FPF and FDF being 0.11 and 0.53, respectively, according well with previous studies on 4CzBN. tPF showed minor susceptibility to the presence of oxygen, decreasing from 2.34 ns in oxygen-free solution to 2.22 ns upon exposure to air. The lifetime of the delayed component, tDF, is of particular importance in TTA-UC since it corresponds to the triplet lifetime. tDF of 4CzBN was determined to be 62 µs, which is sufficiently long to promote diffusion-controlled Dexter-type triplet energy transfer (TET) upon addition of an annihilator species. This relatively long lifetime is the result of a rather large DES-T of 0.28 eV, thus impeding the rate of rISC. The ISC efficiency can be estimated as 1-FPF, yielding FISC = 0.89. The most important photophysical parameters of 4CzBN are summarized in Table . and examples include using the difference between the quenched and unquenched donor total fluorescence quantum yield or delayed component lifetime. Given that the equilibrium between the singlet and triplet state in a TADF compound is perturbed upon addition of a quencher, the methods mentioned above are riddled with assumptions that are valid only for certain compounds. To ensure that the chosen method was valid for 4CzBN we performed simulations (Figure ). The results indicate that probing the changes in tDF upon quenching of 4CzBN yield excellent agreement with the true TET efficiency, as given by Equation S1E. Note that the definition for TET efficiency used herein includes the ISC event, i.e., the maximum value for FTET = FISC (Equation S1E, for a more detailed discussion see the Supporting Information, Section 2.1). |
61bad290d6dcc2026c3a4b10 | 5 | The quenching behavior was analyzed by titration series with each annihilator species, and the obtained TET rates were calculated from Equation S2. The resulting kTET are found in Table (see Figure for Stern-Volmer plots). As expected kTET are typically higher for the annihilators with lower-lying triplets (see Table for triplet energies and Figure for the phosphorescence of PPF), but fortunately endothermic TET from 4CzBN is also possible, yielding kTET on the order of 10 8 M -1 s -1 to the high-triplet energy annihilator PPD. We note that using phosphorescence spectra of rotationally flexible molecules typically underestimates the triplet energy, so the energy commonly referenced for TP (2.53 eV) is therefore likely underestimated. We choose instead a value of 2.62 eV which was obtained from quenching experiments, and which better correlates with the relatively slow TET (kTET = 4.1 ´ 10 8 M -1 s -1 ) observed from 4CzBN to TP. |
61bad290d6dcc2026c3a4b10 | 6 | With these results at hand, we investigated the TTA-UC performance of the different systems. The concentrations employed for UC measurements were 25 µM of 4CzBN and 10 mM (1 mM for 2PI and TIPS-Naph, vide infra) of the annihilator, resulting in systems with endothermic TET (i.e., TET from 4CzBN to PPD) also having FTET close to 89% (as calculated by Equation S4). Delayed UC fluorescence could be observed from all systems upon 405 nm excitation, and the UC emission spectra of TIPS-Naph and TP are presented in Figure . |
61bad290d6dcc2026c3a4b10 | 7 | The spectral shapes are marred by the secondary inner-filter effect at the high-energy end of the spectrum which is caused by the overlap of UC emission and sample absorption. This is typically an issue in especially vis-to-UV UC, even though there are examples of sensitizers with limited UV absorption, thus somewhat mitigating this issue. The low-energy band peaking at around 440 nm is residual prompt fluorescence from 4CzBN, which is an inevitable loss-channel in all these systems. Interestingly, this feature can act as an approximate internal quantum yield reference since the prompt component of 4CzBN (with FPF = 0.11) should be virtually unaffected by the addition of annihilator species. Unfortunately, sensitizer degradation during measurements (vide infra) allows only approximate FUC values to be obtained using the prompt component. Coumarin 153 (FF = 0.53) 60 was employed as an external quantum yield reference instead, ensuring high reliability when evaluating FUC. |
61bad290d6dcc2026c3a4b10 | 8 | Here, f is the spin-statistical factor, FTET is the TET efficiency (ISC included), FTTA is the TTA quantum yield, and FF the annihilator fluorescence quantum yield. Since two low-energy photons are needed to afford one highly energetic singlet, FTTA (and subsequentially FUC,g) has a theoretical maximum of 50%. The internal, or generated, UC quantum yield (referred to as FUC,g) was determined alongside that of the external quantum yield (FUC). The difference between these mainly lie in that secondary inner-filter effects are accounted for when calculating FUC,g, which affects both the spectral shape and the peak intensities (Figure ). Reabsorption is accounted for by using the output coupling yield, Fout, with FUC = FUC,g ´ Fout. A lower value for Fout indicates stronger reabsorption of UC emission by the sample. Another factor that had to be dealt with was that of sensitizer degradation. This is a common issue in vis-to-UV UC systems, and a challenge also faced by the organic light-emitting diode (OLED) community when working with TADF materials in general. Upon 405 nm cw excitation 4CzBN suffered from degradation, which manifested itself both in changes of the absorption spectrum and in loss of fluorescence over time (Figure ). When paired with an annihilator species the UC emission intensity typically went down over time, even though efficient TET attenuated the sensitizer degradation (Figure ). |
61bad290d6dcc2026c3a4b10 | 9 | To determine FUC,g a fitting procedure that accounts for reabsorption was employed, and it is explained in detail in Section 2.3 of the Supporting Information. To our delight, all systems investigated yielded relatively high FUC,g, with the system consisting of 4CzBN/TIPS-Naph in particular yielding a high value of 16.8% (out of a 50% maximum, see Figure for the UC spectrum). This value is to the best of our knowledge the highest vis-to-UV FUC,g reported to date, and an almost 2-fold improvement on the previous record. The remaining systems yielded FUC,g values ranging from 4 to 14%, and full results are presented in Figure and Table . It should be noted that the values achieved for TP (12.6%) and PPD (5.8%), which both emit from singlet states just shy of 4 eV, are multifold improvements on that previously reported, and likely results from more efficient TET and, subsequently, more efficient TTA between triplets. The external FUC measured for our specific setup yielded Fout between 0.7-0.85, resulting from significant reabsorption of the samples. In TIPS-Naph and PPF this results from very small Stokes shifts (Figure ), causing ground state annihilators to reabsorb the UC light to a larger extent than in systems with larger Stokes shifts. In PPD and TP relatively low Fout instead results from the pronounced absorption feature of 4CzBN between 300-350 nm (Figure ), which is part of the spectral region where these annihilators emit. Measurements on all annihilators were also performed in THF, typically yielding lower FUC,g and much more pronounced sample degradation (Table and Figure ). |
61bad290d6dcc2026c3a4b10 | 10 | TIPS-Naph was synthesized in accordance with a literature procedure, and during experiments a fluorescent contamination, which has not been reported previously, was discovered. As detailed in Section 2.5 of the Supporting Information, the removal of this contamination by additional cycles of recrystallization lead to a substantial increase in FUC,g. This could potentially explain why we see a much higher FUC,g (16.8%) compared to other studies using TIPS-Naph (FUC,g » 10%) in which FTET is reported to be close to unity. Our group has previously investigated the locked t-stilbene compound 5,10dihydroindeno[2,1-a]indene (I2), a highly fluorescent compound that unfortunately suffers from very low solubility in toluene. 2PI was chosen as a potentially more soluble equivalent to I2, and the solubility was indeed much higher. When samples containing 10 mM 2PI were used for TTA-UC, however, some light scattering was evident in the absorption. Additionally, the UC signal increased strongly over time during 405 nm excitation, reaching a maximum value after approximately 30 minutes (Figure ). The measured FUC,g for 10 mM 2PI was low (1.0%), which is a lower estimate given that the extended laser exposure not only causes the UC emission signal to increase but also the sensitizer to degrade. Upon lowering the 2PI concentration to 1 mM the scattering decreased significantly, indicating that the observed behavior was due to 2PI not being fully solvated at 10 mM. At 1 mM FUC,g went up to 4.4% and no signal increase was observed over time (Figure ). Probing Triplet Kinetics Using Time-Resolved Emission. To understand the differences in FUC,g between the annihilators we examined the kinetics of the UC samples. There are several important rate constants and parameters needed to properly evaluate TTA-UC systems, e.g., the annihilator triplet excited state lifetime, the triplet-triplet annihilation rate constant (kTTA), and the excitation threshold intensity (Ith). In the following section, we show that these can be determined from the same series of time-resolved UC emission measurements, thus circumventing the need for more challenging transient absorption measurements altogether. |
61bad290d6dcc2026c3a4b10 | 11 | A key factor dictating TTA-UC performance in solution is the annihilator triplet lifetime (tT). A long tT is needed to allow annihilator triplets to diffuse and encounter, resulting in the creation of emissive singlet states via TTA. tT was measured using a previously developed method where the excitation intensity (IEX) dependence on the UC emission kinetics is used (Equation ): |
61bad290d6dcc2026c3a4b10 | 12 | Here, I(t) is the time-dependent UC emission intensity, [ 3 A*] is the annihilator triplet concentration, t is time, and b is a dimensionless parameter indicating what fraction of triplets that initially decay by second-order channels, as defined by Equation . In other words, b represents a system's TTA efficiency (with a possible maximum of 100%), and FTTA may be calculated as b/2 given that these are evaluated at identical experimental conditions. |
61bad290d6dcc2026c3a4b10 | 13 | Our group has previously determined kTTA for compounds based on 9,10diphenylanthracene (DPA) using a method where both time-resolved emission and transient absorption measurements are needed. While the same method in principle is applicable to any system, the spectral overlap between the prompt fluorescence of 4CzBN and the T1 ® Tn absorption of e.g., PPO complicates matters considerably for the systems used here. A new method has instead been developed which relies solely on time-resolved emission measurements of the UC samples, thus circumventing the need for transient absorption. |
61bad290d6dcc2026c3a4b10 | 14 | Instead of using a nanosecond pulsed laser for excitation we used a 405 nm modulated continuous wave laser diode which we coupled to a pulse generator. This way we could control the exact length of the excitation pulse such that the sample emission had reached a quasi steadystate before the excitation light is turned off and the UC emission starts to decay (Figure ). |
61bad290d6dcc2026c3a4b10 | 15 | The excitation rate, kexc, is easily estimated from the sample absorbance at the excitation wavelength and the excitation power (Equation S8). Setting [ 3 A*]0 = [ 3 ASS], a simple and solvable equation system with only two unknowns, kTTA and [ 3 ASS], is obtained from Equation 3 and 5. |
61bad290d6dcc2026c3a4b10 | 16 | Consequently, it is possible to estimate kTTA using the exact same measurements that was used to determine tT ( = 1/kT) of the annihilators. An additional benefit is that it is now possible to directly relate Ith and b. Since b is evaluated at [ 3 ASS], Ith is the excitation intensity that yields b = 0.5. Ith may, thus, be estimated from only a few measurements of the UC emission decay in which the excitation intensity is varied to yield values of b slightly above and below 0.5. |
61bad290d6dcc2026c3a4b10 | 17 | Measurements of the UC decay kinetics at different IEX were performed and was followed by a global fitting procedure in which tT was fitted to a global constant value while allowing b to vary (Figure , see Supporting Information section 2.4 for more details). The results show that PPO has a lifetime of 1.3 ms, which is substantially longer than that previously reported for UC systems employing PPO, but close to that obtained from flash photolysis experiments. The longest lifetime was found for TIPS-Naph at 2.2 ms, resonating well with its impressive performance in terms of FUC,g. The remaining lifetimes span from 0.075 to 0.75 ms (Table ), which is much shorter than those often found in visible emitters based on anthracene, where lifetimes on the order of several milliseconds are common. At high IEX most annihilators still show b values relatively close to unity, indicating that the TTA pathway dominates at high IEX (Figure ). |
61bad290d6dcc2026c3a4b10 | 18 | For this purpose, all investigated systems show unsatisfactory high Ith, with values above 200 mW cm -2 (Figure ,D and Figure ). This emanates from the fact that tT is quite short in these annihilators, combined with a relatively low molar absorptivity of 4CzBN at the excitation wavelength 405 nm (e » 7000 M -1 cm -1 ). Comparison between Ith obtained by evaluation at b = 0.5 (Figure ) and the traditional evaluation of Ith obtained from fitting the steady-state intensity to slopes 1 and 2 (Figure ) yield good agreement between the methods. |
61bad290d6dcc2026c3a4b10 | 19 | The kTTA rates were determined from the same measurements as detailed in the Supporting Information, and the obtained rates are presented in Table . Interestingly, TIPS-Naph shows the lowest kTTA of the annihilators investigated here (6.2 ´ 10 8 M -1 s -1 ), PPO and PPF show similar rates of around 1.75 ´ 10 9 M -1 s -1 , while e.g., TP has an almost twice as high rate constant of 3.3 ´ 10 9 M -1 s -1 . We note that the measured value of kTTA for PPO is approximately three times lower than that reported previously. These results indicate that while the rate of the TTA event itself obviously affects the UC efficiency, it is the annihilator triplet lifetime that preferentially dictates the outcome. This is hardly surprising but worth reiterating, and great care should be given when evaluating especially the triplet lifetime of the annihilator. d Threshold excitation intensity evaluated at b = 0.5. e Rate constant for triplet-triplet annihilation (´10 9 ). f Maximum b value as defined by Equation , estimated at a laser fluence of 18 W cm -2 . g Spin-statistical factor, calculated from Eq. 1 using FTET = 0.89 and FTTA = bmax/2. |
61bad290d6dcc2026c3a4b10 | 20 | TADF Sensitizers: Drawbacks and Opportunities. As is evident from this study, using TADF compounds as the sensitizers in TTA-UC holds great promise. The most obvious advantage compared to other sensitizers yielding decent FUC in vis-to-UV TTA-UC (i.e., Ir complexes and quantum dots containing heavy metals such as Cd and Pb) is that TADF compounds are purely organic, consisting only of earth-abundant, non-toxic elements. They are, thus, well-suited for future large-scale operation, which is not the case for Ir complexes, despite possessing promising photophysical properties otherwise. Additionally, due to the OLED community's increasing interest in TADF compounds during the last decade there is a huge variety of available molecules with different energy levels and triplet excited state lifetimes, of which the latter in many cases are orders of magnitude longer than those found in e.g., Ir complexes. Making the best use of existing TADF compounds in TTA-UC schemes is, however, not straight-forward. The sought-after qualities for use in OLEDs differ significantly from what is needed in a typical sensitizer, meaning that current TADF design in many cases has gravitated towards compounds not suitable for TTA-UC. One crucial benefit in both contexts is the access to small singlet-triplet energy splittings (DES-T). In OLEDs the excited states are created by means of electricity, and the resulting distribution is dictated by spin-statistics, leading to 75% triplets and 25% singlets (Figure ). Highly efficient rISC to generate a higher fraction of emissive singlets is, thus, one of the most important properties of TADF compounds in the context of OLEDs, and is a process that is sped up in molecules with small DES-T (generally, krISC µ exp[-DES-T/kBT]). In TTA-UC a small DES-T enables larger apparent anti-Stokes shifts since the initial energy loss during the ISC event is smaller than in typical sensitizers containing heavy metals. Once the triplet has been populated it is instead beneficial if rISC is inefficient since the generated exciton should be transferred to the annihilator instead of returning to the singlet manifold. A too small DES-T might therefore inhibit efficient TET even if the annihilator concentration is kept high. An intermediate DES-T (0.1 eV < DES-T < 0.2 eV), enabling relatively large apparent anti-Stokes shifts and slow rISC simultaneously, should be favored. |
61bad290d6dcc2026c3a4b10 | 21 | Even smaller DES-T could potentially be used by invoking strategies in which the rISC process is slowed down by clever molecular design. The TET event is further limited by the amount of prompt fluorescence in systems with TADF-type sensitizers. On the contrary to what is wanted for OLED applications, the prompt fluorescence quantum yield should be as low as possible in TTA-UC settings to promote efficient TET (Figure ). |
61bad290d6dcc2026c3a4b10 | 22 | Recently, some progress in this area has been made. Wei et al. reported two new multiresonance TADF sensitizers which when paired with TIPS-Naph or a derivative thereof afforded green-to-UV TTA-UC for the first time. While a relatively modest FUC value of 3.8% is reported, they managed to reach a low Ith of 9.2 mW cm -2 . Part of the success is ascribed to the high molar extinction that was determined for these sensitizers (e >10 5 M -1 cm -1 ), enabled by limiting their structural flexibility and by including electron-deficient boron covalently bonded to the donor units. |
61bad290d6dcc2026c3a4b10 | 23 | 4CzBN possess several of the sought-after properties of a sensitizer, with weak prompt fluorescence, a long-lived delayed component, and slow rISC (Figure ). Its major drawback is the (for most systems) unnecessarily high singlet and triplet energies, which forbid excitation at wavelengths >430 nm, leading to significant energy loss during ISC and TET. Additionally, some photo instability of the UC samples was detected which was ascribed to the degradation of 4CzBN, an issue that can be alleviated by the addition of bulky substituents. Finding complementary compounds with similar characteristics to 4CzBN but with lower excited state energies will be needed to further improve green-to-UV UC, which is especially interesting for solar applications given the vast amounts of green light in the solar spectrum. Considerations on Annihilator Design. Not only could novel TADF sensitizers contribute to improved vis-to-UV TTA-UC systems, but perhaps even more crucial is the pursuit of new annihilators. Design principles that hold true for annihilators in general must obviously be upheld, such as high FF and a long tT, but for UV-emitting systems additional considerations should be taken into account. As touched upon previously many vis-to-UV TTA-UC systems suffer from low photostability, which follows from the relatively high energy of the states involved. This aspect has recently been investigated in greater detail by Murakami et al., gaining important insights as to how the energy levels of the sensitizer and annihilator affect the photostability of TTA-UC systems in solution. They observed a correlation between the main degradation pathway and the energy difference between the LUMO levels of annihilator and solvent. In our study we found no evidence of annihilator degradation during UC experiments, and we primarily ascribe the slight decrease in UC emission over time to sensitizer degradation. |
61bad290d6dcc2026c3a4b10 | 24 | Another aspect that is especially relevant for vis-to-UV TTA-UC is the exaggerated thermodynamic driving force for TTA typically found in UV emitting species. This is the case for the compounds investigated herein: [2 ´ E(T1) -E(S1)] ≥ 0.7 eV for all species, with PPD in particular having a driving force of almost 1.6 eV. If this substantial energy loss could be mitigated substantially larger apparent anti-Stokes shifts could be realized. The relative lowering of the triplet energies should perhaps be the primary focus as this would enable excitation at longer wavelengths than currently possible. A few studies have investigated substituent effects on the energetic landscape of polyacene emitters, Controlling not only the energy of T1 but also of T2 is of significance. In molecules, such as perylene and rubrene, in which the spin-statistical factor f has been determined to lie above the commonly encountered value of 2/5, the energy difference [2 ´ E(T1) -E(T2)] < 0. In perylene this difference is strongly negative, efficiently shutting down the creation of the T2 state upon TTA, causing f to approach unity. In rubrene, however, f is reported to lie around 0.6 in solution, and the creation of T2 during TTA is only slightly endothermic. A recent study by Bossanyi et al. verify that T2 is formed during TTA in rubrene, but that the energy alignment between T2 and S1 allows fast high-level rISC (HL-rISC) from T2 to S1 to occur, outcompeting non-radiative decay from T2 to T1. HL-rISC has been found also in anthracene derivatives (not DPA however) and should be considered as a potential avenue to increase f beyond 2/5. This pathway is very sensitive to the precise alignment of S1, T1, and T2 energies, and the study by Bossanyi et al. suggest that in cases where [2 ´ E(T1) -E(T2)] approaches zero f may in fact approach unity in molecules where HL-rISC occurs. Finally, from simulations the same study states that intermolecular geometry can affect f, with parallel geometries giving rise to higher values. For the annihilators used herein [2 ´ E(T1) -E(T2)] is expected to be much greater than zero. Additionally, S1 is expected to lie several hundreds of meV above T2 for most annihilators (Table ), suggesting that HL-rISC is inefficient in these molecules. Most of the investigated annihilators show an expected value of approximately 0.4, but our results also indicate that f takes a larger value than 2/5 in TIPS-Naph (0.54) but a lower value for PPD (0.22, Table ). |
61bad290d6dcc2026c3a4b10 | 25 | In this work, we show that the UC quantum yield of visible-to-UV TTA-UC systems may approach the often-encountered spin-statistical limit of 20%. We do so by pairing six different annihilators with the purely organic, high triplet energy sensitizer 4CzBN which exhibits efficient ISC and a long triplet lifetime. The results show that the TTA-UC pair 4CzBN/TIPS-Naph achieve a record-setting 16.8% upconversion quantum yield (out of a 50% maximum), and high quantum yields are reached when using PPO (14.0%), PPF (13.0%) or TP (12.6%) as annihilators as well. We also show that the same set of time-resolved emission measurements can be used to determine the annihilator triplet lifetime, the rate constant of triplet-triplet annihilation, and the threshold excitation intensity, all of which are important parameters to probe when evaluating TTA-UC systems. The importance of having long-lived annihilator triplets is reinforced, as our results show that both the TTA-UC quantum yield and the threshold excitation intensity benefits from this. Using 4CzBN as the sensitizer limits the achievable anti-Stokes shifts, and our results are discussed in the context of extending the excitation wavelength further into the visible region. The development of high-efficiency vis-to-UV TTA-UC systems will require both new sensitizer and annihilator compounds, and finding avenues to control and alter the singlet and triplet energy levels of these will be crucial in order to combine high efficiencies with e.g., excitation with green light. |
61bad290d6dcc2026c3a4b10 | 26 | Steady-state absorption spectra were recorded on a Varian-Cary 50 Bio UV-vis spectrophotometer and steady-state fluorescence measurements were carried out on a Spex Fluorolog 3 spectrofluorometer (Horiba Jobin Yvon). The prompt fluorescence lifetimes of 4CzBN were measured using a time correlated single photon counting (TCSPC) setup using PicoQuant laser diodes (405 nm) and a PMT detector (10 000 counts, 4096 channels). Steadystate upconversion fluorescence measurements were performed on a home-built system using a continuous-wave 405 nm OBIS laser (Coherent) as the excitation source. The measured maximum power output was 87.3 mW and the laser beam diameter was 0.8 mm. A linear variable neutral density (ND) filter was used to vary the laser intensity, and data were recorded using home-built LabView software. |
61bad290d6dcc2026c3a4b10 | 27 | Nanosecond time-resolved emission measurements were performed on a home-built system using a continuous-wave 405 nm OBIS laser (Coherent) coupled to a pulse generator as the excitation source. A Cornerstone 130 monochromator (Oriel Instruments) was used when measuring the transient signals at 440 nm from 4CzBN. For transient measurements of upconverted light, a 300-390 nm band-pass filter was instead used after the sample to maximize the signal intensity. The signals were collected with a 9-stage PMT coupled to a Tektronix TDS 2022 oscilloscope. The optical response time of the PMT was set to be much shorter than the measured decay. All photophysical measurements were carried out in toluene using 2 mm quartz cuvettes. All samples were prepared in a nitrogen glovebox (Innovative Technologies) with <0.7 ppm oxygen levels and sealed with air-tight cap screws and parafilm. Temperature-dependent measurements were performed using a liquid nitrogen cryostat (Oxford Instruments) connected to a temperature controller. |
61bad290d6dcc2026c3a4b10 | 28 | When discussing triplet energy transfer (TET) one usually considers only the actual TET event when evaluating the TET efficiency (FTET), i.e. the ratio between produced donor and quencher triplets. In TADF compounds the excited state equilibrium is perturbed upon addition of the quencher, and FTET will depend also on the ISC/rISC events, as shown below. |
61bad290d6dcc2026c3a4b10 | 29 | 𝑘 +-#. = 𝑘 -#. 𝑒 :∆3 +1" /?0 (S3) While the two different methods are susceptible to the exact values of different rate constants, some qualitative conclusions may be drawn. In general, using the steady-state fluorescence intensity is a good approximation for TADF compounds with small DES-T (<0.1 eV, Figure ), while using the lifetime of the delayed component as the proxy is more accurate for compounds with larger DES-T (>0.1 eV, Figure ). For 4CzBN, which has a quite large DES-T of 0.28 eV, using only the lifetime of the delayed component as the proxy is likely to yield a very good estimate of FTET. One should however always be careful before employing either method and make sure that the implied assumptions are valid for the particular system. |
61bad290d6dcc2026c3a4b10 | 30 | To evaluate the annihilator performance both the internal, or generated, upconversion quantum yield (FUC,g) and the external quantum yield (FUC) were determined. FUC,g is of most interest when investigating the intrinsic properties of the annihilators and accounts for reabsorption effects, whereas FUC is a sample-dependent metric which depicts the number of photons actually detected from each UC sample. |
61bad290d6dcc2026c3a4b10 | 31 | To determine FUC,g we used a procedure reminiscent of that previously deployed in our group. In all UC measurements 2 mm cuvettes were used and the shorter path length was directed towards the detection source to minimize reabsorption. After the UC spectrum had been collected, the emission spectrum of an optically dilute annihilator sample was fitted to match the spectral region where the UC sample shows low absorption (Figure , blue spectra). |
61bad290d6dcc2026c3a4b10 | 32 | Here, UC and r denote upconversion and reference sample, respectively. F is the integrated emission intensity, A the absorption at 405 nm, and h the refractive index of the solvent. Except for the fitting procedure no additional corrections for reabsorption were invoked, even though the UC sample absorption at the fitting wavelength was non-negligible. The presented values of FUC,g are, thus, potentially somewhat underestimated. The procedure for calculating FUC was identical, but the high-energy end of the spectrum (of which parts were typically reabsorbed) was not accounted for (Figure , red spectra). Note that the employed definition for FUC,g gives a theoretical maximum of 50%. |
61bad290d6dcc2026c3a4b10 | 33 | In this section all samples consist of 10 mM annihilator + 25 µM 4CzBN if not stated otherwise (1 mM for 2PI and TIPS-Naph). Time-resolved emission measurements were performed for each annihilator at different excitation intensities, and the measurements were fitted to normalized data using Equation S6 (Equation 2 in the main text). A global fitting procedure was used where b was allowed to vary between measurements, but tT was shared globally. |
61bad290d6dcc2026c3a4b10 | 34 | To determine kTTA the measurements from Figure (and Figure atmosphere. After adding diisopropylamine (36 mL) to the solution, the mixture was heated to 100 °C. Subsequently, triisopropylsilylacetylene (3.52 g, 19.3 mmol) was added dropwise and the resulting solution was stirred at 100 °C overnight. After cooling to room temperature, THF and diisopropylamine were removed under reduced pressure. The residue was extracted with CHCl3, dried over Na2SO4, filtered, and dried under reduced pressure. The crude product was purified by column chromatography (n-hexane) followed by recrystallization from methanol to give TIPS-Naph (69% yield) with ≈99.9% optical purity. Three further consecutive recrystallisation steps was employed to yield TIPS-Naph with an optical purity of ≈99.99% following the discovery of a fluorescent contamination. |
60c75226702a9b08e918c097 | 0 | The term efficient has gained great popularity in the chemical literature, despite the lack of an applicable and relatable definition. In this perspective, a chemical definition of efficiency is discussed building on the concept of non-wasteful resource usage. It is proposed that an efficient method, synthesis or protocol is one which requires less resources in the form of money, time and materials than comparable approaches which accomplish the same task. |
60c75226702a9b08e918c097 | 1 | Chemistry is, simply spoken, the science of making and analyzing compounds. As such, the synthetic aspect of chemistry is central to many endeavors in today's research, as the ability to synthesize a molecule is a prerequisite to its experimental analysis. Therefore, much of the work at the forefront of current research in the field is focused either on i) making novel compounds, ii) improving methodologies to access known compounds or structural motifs or iii) devising new methods for these purposes. The name of the game here is better, faster, cleaner, easier -i.e. more efficient. Indeed, the use of "efficient" as a descriptor of methods or procedures has become increasingly popular in the literature over the past two decades (Figure ). In 2019, around 10,000 papers across all major chemistry journals contained the word "efficient". Despite this popularity, a clear definition of the term efficiency in a chemical sense is missing from the literature. |
60c75226702a9b08e918c097 | 2 | In several dictionaries, efficiency is defined as the ability to obtain a product or effect an outcome in a way that does not waste resources, for example in the form of money, time and materials. Conveniently, the relevant metrics to evaluate the efficiency of methods or syntheses in this sense are already at our disposal. Monetary investment can be evaluated via the cost of reagents, time by the number of steps and reaction times and material use through a green chemistry metric of choice. However, the trend in many papers recently is to describe a method as efficient solely based on the yield of a reaction. This ignores other factors such as the origin of the starting materials (i.e. if they had to be synthesized beforehand via a lengthy route of their own) and the amount of reagents, solvents and workup materials used as well as the time it took to perform that synthesis. Clearly, focusing on just one outcome paints an incomplete and potentially misleading picture. |
60c75226702a9b08e918c097 | 3 | But what exactly does waste mean in this regard? Admittedly, some sort and amount of waste material is a natural byproduct of almost any chemical transformationbe it solvents, inorganic salts or undesired reaction products. A similar situation exists for monetary investment and time, as all syntheses require some financial effort and time to perform said transformation. As such, efficiency must be relative, since it requires some form of benchmark to compare the resource investment of a given method to. Herein, the nature of a transformation or target molecule must be considered to avoid a comparison of apples to oranges. Consequently, any method should only be compared to its peers which accomplish the same task. By this logic, methods for an entirely new type of transformation cannot be termed efficient since there is no benchmark to meet. |
60c75226702a9b08e918c097 | 4 | An efficient method, synthesis or protocol is one which requires less resources in the form of money, time and materials than comparable approaches which accomplish the same task. Money can be evaluated via the cost of reagents. Time can be assessed through the reaction time as well as the duration of workup and purification procedures. Material usage can be estimated by a green chemistry metric such as the E-factor or process mass intensity. |
60c75226702a9b08e918c097 | 5 | The community should certainly be encouraged to continue to strive for more efficient chemistry. However, a more effective use of "efficient" in the literature might be appropriate. Any method which accomplishes a given task is an effective method (i.e. it did the job), but its efficiency should be determined through comparison with other methods for the same task (i.e. how well did it do the job compared to known alternatives?). Thus, to describe a method as efficient one could benchmark this method against alternatives in the literature, for example in terms of number of steps required for the route, cost of reagents or material use. The latter can be evaluated via green chemistry metrics such as the environmental factor (E-factor) or process mass intensity (PMI), which are both simple mass-based metrics to describe the amount of waste produced for a (hypothetical) kilogram of product. Usually, the E-factor or PMI alone already provide a valuable insight into the overall wastefulness of a reaction or process, since they include all materials used for a transformation or route and correlate with the number of steps. While other green chemistry metrics may also be applied and might provide a comparable picture, consistency in the literature should be advocated and either of the metrics above are recommended for their simplicity and transparency. This, of course, does not mean that methods which are not efficient by the above definition do not have synthetic value (because certainly they do!). Other factors such as an improved substrate spectrum, milder/ambient conditions, simpler/more robust operational procedures or avoidance of hazardous/toxic reagents are all highly relevant and deserve attention from the field. |
6560bd93cf8b3c3cd707325a | 0 | Dementia is the most common disease globally . Dementia is a broad term used to describe a range of cognitive impairments and symptoms that affect a person's memory , thinking, and ability to perform daily activities . It is not a specific disease but rather a syndrome or a set of symptoms that can be caused by various underlying conditions. Dementia is not a normal part of aging, and it can have a significant impact on a person's quality of life. It is characterised by the accumulation of abnormal protein deposits in the brain, leading to the progressive deterioration of cognitive function. Vascular dementia occurs when there is damage to the blood vessels in the brain, often due to strokes or other vascular problems. Lewy bodies are abnormal protein deposits in the brain, and this type of dementia is associated with their presence . The symptoms of dementia can include memory loss, confusion, difficulty with language, impaired judgment, personality changes, and a decline in the ability to perform everyday tasks . The progression and severity of symptoms can vary widely depending on the underlying cause and individual factors. Plants that have been used for anti -dementia. In traditional herbal medicine, different plants and their active constituents have been used to treat disorders. Pistacia khinjuk is a member of well known pistachio family , special item among dry fruits, that is easily seen on the sale points in the coldest areas of Balochistan, pk on the arrival of the winter season. The fruit is multicoloured, but the dark green colour is prominent. The fruit looks like a small round ball. The selected plant is drought tolerant and it is easily grown in degraded areas. There have been numerous small molecules employed for the treatment of dementia . |
6560bd93cf8b3c3cd707325a | 1 | From the oil of the mentioned plant, another terpinen molecule is selected, terpinen-4-ol, for the target protein. The smile notation is CC1=CCC(CC1)(C(C)C)O. The molecular weight is 154 g/mol. The molecular formula is C10H18O. There is one rotatable bond, one hydrogen bond donor, and one hydrogen bond acceptor. There are eleven heavy atoms and one oxygen atom. |
6560bd93cf8b3c3cd707325a | 2 | The oil was analyzed according to the method present by using capillary gas chromatography. The oils obtained from fruit were injected to gas chromatography equipped with fused silica column. Temperature was 280°C at a rate of 4°C/min. Injector and detector (FID) temperatures were 290°C, helium gas was used as carrier. GC-MS system equipped with silica column .Oven temp 250°C .The carrier gas was helium with a linear velocity of 31.5 cm/s, split ratio 1/60, Ionization energy 70eV, scan time 1 s and mass range of 40-300 amu. The percentages of compounds were calculated by the area normalization method, without considering response factors. The components of the oils were identified by comparison of their mass spectra with those of known GC-MASS |
6560bd93cf8b3c3cd707325a | 3 | After isolation and identification of natural molecule from oil, the structure of the molecule was drawn with Avogadro 1.2 . Optimization of the geometry was carried out automatically. For oral toxicity, cytotoxicity and stress response calculations, ProTox-II was used . The protein structures of experimental enzyme acetylcholinesterase were obtained from the online protein data bank database system (1EEA). A quantitative structureactivity relationship model was used to predict the potential hERG blockage or Ether-à-go-go gene .The online system is the preferred choice for the author to determine the blood brain barrier , remove unwanted with the help of Chimera 1.13.1 . Online molecular docking server were used for study ligand enzyme binding. |
6560bd93cf8b3c3cd707325a | 4 | So on the basis of molecular investigation, alpha-eudesmol prevents breaking down acetylcholine in the brain. The ligand attaches to the arginine amino acid of acetylcholine at position 289; the oxygen of eudesmol makes a bond with NH group of arginine. As a result, an increased concentration of acetylcholine leads to increased communication between nerve cells, so alpha-eudesmol become suitable candidate for an acetylcholinesterase inhibitor, a chemical that binds to the cholinesterase and prevents it from breaking down the neurotransmitter, acetylcholine inhibitor also known as anti-cholinesterase, are molecules that prevent the breakdown of the neurotransmitter acetylcholine. |
6560bd93cf8b3c3cd707325a | 5 | Inflimination dementia are correlate . Cyclooxygenase-2 (C0X-2) is an enzyme that plays a key role in promoting inflammation. In contrast, when cyclooxygenase-2 activity is blocked, it results in reduction of inflammation. Cox-2 becomes active only at the site of inflammation i.e peripheral tissues. These are propitious curative molecules not only for the treatment of inflammation, evidence of higher level of inflammation when compared with those of people without dementia, On the basis of scientific result alpha eudesmol & terpinen-4-ol blocked Cox-2 activity. |
6560bd93cf8b3c3cd707325a | 6 | As of today, different approaches for selecting plant fruits with multipotent activities are being investigated for the treatment of neurodegenerative disorders. This is a relatively successful method for the identification of plants and compounds that may be exploited for therapeutic use in disorders. However, further studies should be conducted to explore the effects of essential oils from other species of the same genus, as well as their synergistic effects, to optimise their potential. Additionally, the potential use of these oils on a clinical basis should also be studied.. |
6229f5b450b62150c0f0b0be | 0 | Visible light-driven photoredox catalyzed has gained numerous attention as a powerful and energy-efficient method for chemical synthesis. The central of concept is rely on oxidation addition with specific radicals from precursors and trapped by transition metal complex such as nickel, 2 palladium, cooper 4 and so on. After elimination by reduction, the desired crosscoupling product is generated and the metal complex is reduced through the photocatalytic cycle for next process. However, the whole process is highly depended on homogeneous photoresponse catalysts, such as ruthenium, iridium or organic dyes 8 in numerous reports. It is a pity that these homogeneous photocatalysts are also accompanied by weaknesses such as expensive expenses, separation from products, unable to recycle which limits the actual industrialized application to a certain extent. |
6229f5b450b62150c0f0b0be | 1 | Based on the actual background and requirement, in recent years, some representative heterogeneous catalyst such as g-CN, 9 TiO2, 10 CdS 11 involved in photoredox process has been gradually developed. König' groups 9a reported ligand-free mpg-CN/Ni dual photoredox catalytic protocols for C-N bond formation, and explored mpg-CN/Ni-dual catalysis for C(sp 2 )-C(sp 3 ) cross-coupling reactions yielding diarylmethanes analogously. Pieber and Seeberger reported mpg-CN in combination with nickel catalysis can induce selective C-O cross-couplings of carboxylic acids with aryl halides. 9c Meanwhile, they also reported g-CN/Ni with light-mediated cross-couplings of aryl bromides with alcohols via C-O bond formation. Inspired by previous investigation, 12 we hence developed mpg-CN/Ni dual catalysis sulfone compounds synthesized which are widely found in several drug-active molecules such as Dapsone, Vismodegib and Intepirdine (Scheme 1a) via C(sp 2 )-SO2Ar bond formation. Compared with the traditional method (Scheme 1b), this kind of heterogeneous photocatalyst involved in photoredox has the potential for industrial and large-scale development due to its advantages including mild conditions, visible light-driven, base-free, high-yield outcomes and reutilization. |
6229f5b450b62150c0f0b0be | 2 | In our consideration of this approach and previous report, 14 we envisioned that aryl sulfonate salts have low redox potentials (E1/2 = -0.37 V vs. SCE in CH3CN) and could undergo single electron transfer oxidation by photogenerated hole form excited g-CN which is effective to generate sulfonyl radicals. Oxidative addition of Ni(0) species to an aryl halide delivers the Ni(II) intermediate 14 which tends to trap the sulfonyl radicals yielding a Ni(III) organometallic adduct. Subsequent reductive elimination produces the targeted C(sp 2 )-SO2Ar cross-coupling product. Finally, the electron form semiconductor surface is utilized for the reduction of the Ni(I) species to Ni(0) species through another single electron transfer to complete the whole Ni catalytic cycle (Scheme 1c). Scheme 1. Significant of Sulfones and Experimental Design With this design in mind, we began to investigate the possibility and figure out the best conditions primarily. According to Table , 4-bromotoluene 1a and sodium benzene sulfinates 2a were chosen to be the model substrate separately. To our delight, the product 1-methyl-4-(phenylsulfonyl) benzene 3a was obtained with the excellent yield of up to 82% under 45W blue led (455 nm) irradiation and 10 wt% of g-CN, (Synthesis method was shown in Support information, SI), 5 mol% NiBr2•DME and 10 mol% L1 added in DMF as the standard condition (entry 1). Then, the reaction could not happen when the absence of light, g-CN, or nickel catalyst respectively (entry 2). We chose the Ni(COD)2 as the nickel source but a modest yield was acquired (entry 3). Meanwhile, we also investigated the ligand effect to this reaction system. Firstly, when no ligand was involved, the produced 3a could not be observed at all (entry 4), and then we changed similar ligands such as L2 or L3, but an unsatisfactory result (entry 5, 6) was shown. MeCN, MeOH or acetone was selected to be solvent respectively, but trace product was detected (entry 7). Shortening reaction time to 12 h, the yield was decreased to 68% accordingly (entry 8). The reaction could not be occurred without degassing (entry 9). Under the same conditions, chlorobenzene was introduced but no results were obtained. Considering the difficulty of oxidation addition between Ni complex and chlorinated benzene, so we attempted to raise the temperature to 60 o C, in that case, 12% yield of product was observed finally, and we also chose the iodobenzene as the substrate which obtained 3a up to 78% yield (entry 10). |
6229f5b450b62150c0f0b0be | 3 | We then used the optimized reaction conditions to explore the scope of the reaction with respect to the variation of different substituted bromobenzene. As shown in Scheme 2. various aryl bromides substituted with electron-donating group including methyl 3a, 3i, 3o, methoxy group 3n, tert-butyl 3f, or phenyl 3h, Naphthalene 3r or strong electron-withdrawing group including trifluoromethyl 3c, 3j, 3p, trifluoromethoxy 3m, fluorine atom 3d, 3k, 3q, cyan 3e, 3l at different positions (o, m, p) were reacted very well, generating the corresponding products with moderate to high yields. In addition, some heterocyclic compounds like thiazole 3s, pyridine 3t, thiophene 3u and furan 3v were also tolerated successfully via this method with iodine reagent alternatives as the reagent. Scheme 3. Substrate Scope of Sulfinate salt a a Reactions were performed under the standard conditions (Table , Entry 1) and isolated yields were reported. b Aryl iodide was used for the reaction partner. |
6229f5b450b62150c0f0b0be | 4 | In order to further investigate the comprehensive scope of the reaction method, various sodium benzenesulfonates were selected to participate in model reaction condition, and the corresponding good results were obtained similarly. As shown in Scheme some of the common substituent groups like methyl 4a, 4f, 4i, methoxy group 4d, tert-butyl 4e, Naphthalene 4k or trifluoromethyl 4b, fluorine atom 4c, 4g, 4h, chlorine 4j were proved to be compatible with suitable yield. In addition, dine compound was suitable for this condition. |
6229f5b450b62150c0f0b0be | 5 | In view of dapsone is a famous bio-active molecule with the corresponding structure, we used the corresponding raw materials, through the standard explored conditions, to achieve the synthesis of this drug in large scale up to 74% yield successfully (Figure ). Motivated by this result, we also realized the synthesis of 5-HT6 receptor antagonist precursor (Figure ). Used 8-chloro-3-iodoquinoline and sodium benzenesulphinate as the corresponding substrate, under the template conditions, and 42% yield of 5b was isolated successfully which could be further converted to RVT-101 16a or Lu AE60157 16b in virtue of mature C-N coupling method 17 . What's more, considering that the numbers of cycles from stabilization were the key to measuring the practical application towards heterogeneous catalysts, we have studied the effect of g-CN though repeated to use. As shown in Figure , after each single reaction cycle, an excellent conversion rate was obtained in spit of 5 times under the same conditions and recovered g-CN still representing a thin layer appearance with TEM image attached. In summary, a dual Ni/photocatalytic C(sp 2 )-SO2Ar coupling was developed using a carbon nitride semiconductor as recyclable photocatalyst with low toxicity. The semi-heterogeneous nickel/carbon nitride catalysis is an inexpensive, sustainable alternative to homogeneous protocols. The method selectively couples a broad range of aryl bromides with sodium benzenesulfonate in good to excellent isolated yields. Dapsone synthesis was also demonstrated on a gram scale and cyclic experiments also demonstrated the great potential of g-CN in such reactions. |
658065c9e9ebbb4db933af2c | 0 | Despite being a recognized issue for decades, the fundamental origins of polymer fouling remain unclear, and mitigation strategies are primarily empirical. Moreover, researchers haven't agreed upon the fouling mechanism, fouling precursors, and the location of reactions (in the bulk liquid or at the equipment surface). Additionally, very few predictive detailed models exist for polymer fouling. Oftentimes, the fouling models are composed of pseudospecies or highly simplified chemistry, which loses the resolution to investigate chemical mechanistic details. |
658065c9e9ebbb4db933af2c | 1 | In our prior research, we took the initial step of creating a predictive multiphase detailed chemical kinetic model for polymer fouling in an industrial distillation column that separates C4 and lighter species from heavier ones ("debutanizer"), as illustrated in Fig. . That model only considered the base case scenario, which includes alkanes, alkenes, conjugated dienes, and aromatics from steam cracker effluents, and excluded any impurities or industrial additives. The model showed that under the base case scenario, the main fouling mechanism is the polymerization of reactive monomers, e.g., dienes. Moreover, the model illustrated that under the base case scenario, fouling growth is primarily caused by reactions in the thin film on the tray surface, while reactions in the bulk liquid are crucial for controlling the concentration of reactive intermediates (e.g., radicals) involved in the surface reactions. |
658065c9e9ebbb4db933af2c | 2 | That prior work suggests that predictive detailed modeling can be the key to reaching a scientific consensus on fouling fundamentals. Impurities can greatly impact the fouling chemistry. One of the most prevailing impurities is traces of dissolved oxygen in the unit fluid, which can enter the process as an impurity along with the feedstock, water used for quenching, or additives. Literature experimental investigations have shown that dissolved oxygen can affect fouling significantly, where dissolved oxygen was measured to accelerate fouling even at ppm level concentrations, while the fouling rate becomes independent of oxygen concentration near air-saturation. The effects of dissolved oxygen on fouling rates were recently measured using Quartz Crystal Microbalance with Dissipation monitoring (QCM-D), similar to the prediction by our detailed fouling model for QCM-D cell fouling. It is generally accepted that oxygen affects fouling through autoxidation. Autoxidation is a process where dissolved oxygen combines with ordinary radicals formed from reactions between unsaturated hydrocarbons, transforming into peroxyl radicals. The peroxyl radicals then propagate with additional monomers and dissolved oxygens and become polyperoxides, which eventually become oxygenated deposits. For certain monomers, the autoxidation forms peroxyl radicals that are more reactive than the ordinary radicals, accelerating the fouling process. However, the relative rates of fouling contributed by ordinary radical polymerization and autoxidation depend on the specific fouling monomers involved. |
658065c9e9ebbb4db933af2c | 3 | The threshold oxygen concentration at which autoxidation begins to dominate is not well understood for many fouling monomers, let alone for mixtures. Fouling in an industrial distillation column is a much more complicated phenomenon. The interplay between the detailed chemistry of the dissolved oxygen and the complex mixture of species in feedstock, the wide range of tray temperatures (which affects the kinetics), the convective flow, the vaporliquid mass transfer, and the phase equilibria in industrial units are not well-understood. |
658065c9e9ebbb4db933af2c | 4 | In this work, we extend the base case model for debutanizer fouling in our previous work toward oxygen chemistry ("oxygen-perturbed model"), as illustrated in Fig. . We follow a similar modeling method as presented in our previous work. A general flow sheet of this work can be found in Fig. . First, we model the separation chain downstream of the steam cracker using Aspen Plus, where we introduce dissolved oxygen to the debutanizer feedstock at an air-saturated concentration for the extreme-case scenario study. We generate the polymerization and autoxidation mechanism in the bulk liquid at each tray, condenser, and reboiler, using Reaction Mechanism Generator (RMG), a software package for automatic mechanism generation, originally developed by the MIT Green Group, and now developed as an open-source project. To obtain a mechanism with a reasonable size, we use a hard cutoff for oligomer size during mechanism generation. We want to recover the chemistry related to the truncated larger oligomers. The Anderson-Schulz-Flory (ASF) distribution is an expression that predicts the weight distribution of oligomers with respect to its number of monomer units, developed from ideal chain-growth polymerization. Using the ASF distribution, we can estimate the concentration of truncated larger oligomers based on those of the smaller oligomers in the mechanism. We modify the ASF distribution expression to accommodate the combination of ordinary free radical polymerization and autoxidation to better estimate the oligomer distribution under molecular oxygen contamination. The film growth mechanism at the surface of each tray, condenser, and reboiler is modeled using a variant of fragment-based modeling, 20 a type of lumping well-suited for modeling unselective chemistry of mixtures containing macromolecules. We update the fragment-based reaction template presented previously to include important autoxidation reactions. We simulate the oxygen-perturbed fouling model with a simulation scheme similar to our previous work using ReactionMechanismSimulator.jl (RMS), a Julia software package for simulating large chemical kinetics models developed recently by the MIT Green Group. As shown below, the new oxygen-perturbed model predicts a faster fouling rate in some trays of the column and prompts us to consider diffusion limits when simulating the fouling growth. |
658065c9e9ebbb4db933af2c | 5 | We use an Aspen model of the separation chain downstream of the steam cracker to investigate the reaction and flow conditions around which polymer fouling occurs, and study the dissolved oxygen concentrations in the distillation column of interest. The debutanizer column is downstream of a depropanizer, so the feed is a C4+ mixture. The biggest component is butadiene, but it also contains benzene, cyclopentadiene, and other compounds. We modify the Aspen model from our previous work and introduce various oxygen concentrations to the debutanizer feed, up to that it contains an air-saturated level of dissolved oxygen as an upper bound. Since we weren't able to find experimental data for oxygen solubility in butadiene, we approximate it using the saturated concentration of O 2 in benzene (1.9 mol of O 2 per m 3 of benzene) . |
658065c9e9ebbb4db933af2c | 6 | The major species concentration, dissolved oxygen concentration, and the temperature at each tray, the condenser, and the reboiler in the debutanizer are shown in Fig. , where the feed to the debutanizer contains an air-saturated level of dissolved oxygen. Aspen simulation shows that dissolved oxygen in the feedstock gets distilled out and results in a higher concentration of oxygen in the top part of the column. The O 2 concentration decreases drastically in the bottom section. Concentrations below 10 -18 mol/m 3 in the Aspen output show numerical noise and are set to zeros in the subsequent steps of the modeling. |
658065c9e9ebbb4db933af2c | 7 | For the major species, the reactions in this system are slow compared with the flows, so the reactions do not significantly affect their concentrations. Although the O 2 concentration is much smaller than some major species, its steady-state concentrations in each tray are majorly determined by the relatively fast flow, including O 2 evaporation, and are not significantly affected by reactions. : Temperature (a), major species concentrations in the bulk tray liquid (b) and tray overhead vapor (c), the residence time of the bulk tray liquid (d) and the tray overhead vapor (e), and oxygen concentration in the bulk tray liquid (f) and tray overhead vapor (g) at different locations in the debutanizer as computed using Aspen Plus. Tray 1 refers to the condenser, tray 40 refers to the reboiler, and tray 20 is the feed tray. The residence time of the liquid/overhead vapor is calculated as the liquid/vapor phase volume divided by the liquid/vapor outlet flowrate. The residence time of liquid in the reboiler is calculated using the inlet liquid flowrate, and the residence time of vapor in the condenser is calculated using the inlet vapor flowrate. |
658065c9e9ebbb4db933af2c | 8 | The modeling scheme models the fouling in the distillation column as 40 interconnected tray models. Each tray model contains four submodels corresponding to different phases: the tray overhead vapor, tray bulk liquid, the liquid absorbed in the film, and the solid film growing on the tray surface. Each phase submodel is assumed to be perfectly homogeneous. |
658065c9e9ebbb4db933af2c | 9 | The vapor phase submodel accounts for the vapor-liquid mass transfer, vapor-liquid equilibria, and the vapor convective flow from one tray to another. Due to the short residence time (Fig. (e)) and low concentration observed from the Aspen simulation, the vapor phase chemical reactions are assumed to be negligible. Details on the liquid-phase submodel and the film-phase submodel can be found in Sec. 3.1 and Sec. 3.3. |
658065c9e9ebbb4db933af2c | 10 | The liquid-phase submodel accounts for the vapor-liquid mass transfer, vapor-liquid equilibria, liquid convective flow from one tray to another, and the liquid-phase oligomerization and autoxidation chemistry in the bulk tray liquid. The vapor-liquid mass transfer fluxes are modeled using two-film theory with liquid volumetric mass transfer coefficient (k liq A) and Henry's law constant (k H ). k liq A is estimated using the temperature-dependent empirical correlation derived from solvent viscosity, solvent density, and solute diffusion coefficients. |
658065c9e9ebbb4db933af2c | 11 | The detailed oligomerization and autoxidation mechanism for the liquid-phase submodel is generated using Reaction Mechanism Generator (RMG). We seed the mechanism generation using the oligomerization mechanism in the base case debutanizer fouling model (without oxygen) for consistency. For the mechanism generation, we use 40 simulators, each corresponding to the tray liquid at each tray, condenser, and reboiler. We initialize these simulators using the temperatures, major species concentrations, volumetric flow rates, and dissolved oxygen concentrations at each tray, condenser, and reboiler in the debutanizer obtained from the Aspen simulation. |
658065c9e9ebbb4db933af2c | 12 | Most of the gas-phase thermochemistry is estimated by RMG using methods in RMGdatabase or found in RMG-database. Some of the critical species are refined using quantum chemical calculations. Similarly, most of the rate coefficients are estimates by RMG using methods in RMG-database, but some of the critical ones are refined using quantum chemical calculations. See Sec. 4 for the details on quantum chemistry calculations. |
658065c9e9ebbb4db933af2c | 13 | Liquid-phase effects are accounted for as follows. The gas-phase thermochemistry is converted to the liquid-phase thermochemistry by applying the solvation energy corrections estimated using the method of Chung et al. These corrections affect the reverse direction rates, which are calculated using the forward direction rates along with equilibrium constants computed from the solvation-corrected reaction Gibbs free energies. The solvation kinetic effects can also be important for the forward direction but are not considered in this work due to the lack of an accurate and robust estimator for solvation correction in reaction barrier heights. Diffusion-limited rates are applied based on the works of Rice et al. and Flegg et al. |
658065c9e9ebbb4db933af2c | 14 | Anderson-Schulz-Flory (ASF) distribution, which is based on ideal chain-growth polymerization, is utilized to estimate the number of large oligomers formed. The ASF theory predicts the mass fraction of chains with a length of k monomers based on α, the probability of chain elongation (Eq. 1). α can be calculated using the rate of production analysis (Eq. |
658065c9e9ebbb4db933af2c | 15 | The film-phase submodel accounts for the film growth chemistry on the tray surface. The film-phase submodel assumes an initial thin film of 10 µm on the tray surface due to imperfect cleaning. We assume that the tray liquid swells the film, so all the reactive sites on the film are in contact with the liquid-phase species absorbed in the film. The absorbed liquid-phase species reacts with the reactive site on the film to form covalent bonds to the film, resulting in film growth chemically. As the film grows thicker, the film growth can become diffusionlimited, such that only the reactive sites near the surface of the film are in contact with sufficient amounts of absorbed reactive liquid-phase species to react, which would effectively cap the growth rate as the film thickens. |
658065c9e9ebbb4db933af2c | 16 | Since the film growth chemistry revolves around the reactive sites on the film, we represent the film growth chemistry using a variant of fragment-based modeling, where the reactive sites on the film are lumped as reactive fragments participating in the liquid/solid interface reactions as reactants and/or products. In order to incorporate new types of reactive sites on film due to autoxidation chemistry, new fragments are defined. The reactive fragments considered in the oxygen-perturbed fouling model are the carbon-carbon double bond (CDB), the allylic hydrogen (AH), the allylic radical (AR), the alkyl radical (KR), the carbon peroxide (CP), the hydroperoxide (HP), the peroxyl radical (PR), and the alkoxyl radical (OR), as shown in Fig. . A representative structure is assigned to each fragment for estimating its thermochemistry using methods described in Sec. The effects of the truncated liquid-phase oligomers on fouling growth are recovered using a fragment-based modeling approach. We consider the reactions between the functional groups on the heavy tail of liquid-phase oligomers and the reactive site on the solid, instead of using the heavy-chain oligomers explicitly, to prevent an over-complicated model while maintaining interpretability. The steady-state concentration of the functional groups on the heavy tail of liquid-phase oligomers in each tray is estimated using α and w(k), which is elaborated in Sec. S2.7. |
658065c9e9ebbb4db933af2c | 17 | When the spatial gradients in the films are considered, the number of state variables becomes even larger. The radicals in the system have a short chemical time scale, typically in the range of milliseconds. On the other hand, the residence time, representing the average time a species spends in the system, is tens of seconds at each tray. Furthermore, the overall simulation time span for the fouling model is extended to 1 year to capture long-term effects and assess the model performance over the extended period representative of industrial fouling time scales. The large number of state variables and large separation of time scales pose challenges in simulating the fouling model. |
658065c9e9ebbb4db933af2c | 18 | where N j vap/liq are the number of moles in the bulk liquid or vapor in the j th tray of the distillation column. We call the condenser Tray 1 and the reboiler as Tray 40. V j vap/liq,out are the vapor/liquid volumetric flowrate leaving the j th tray, C j vap/liq are concentrations in the bulk liquid or vapor in the j th tray, k liq A are the effective rates of distillation/condensation of each species, T j is the temperature in the j th tray, and k H are Henry's Law constants. S liq 13 is the matrix of stoichiometric coefficients for liquid+liquid reactions, r liq are the reaction rates, V liq is the volume of liquid in each tray. Here we assume all the trays are full, with a diameter of 2.5 m and a height of 0.3 m, i.e., V 2 liq = V 3 liq = ...V 39 liq = (2.5/2) 2 × π × 0.3 = 1.47m 3 . We assume the liquid volume in the condenser and the reboiler are the same, i.e., |
658065c9e9ebbb4db933af2c | 19 | As in our prior work, we neglect reactions occurring in the vapor phase and the effects of the very slow film growth chemistry on the composition of the bulk liquid in each tray of the distillation column. We solve this tightly coupled vapor-liquid system of ordinary differential equations (ODEs) using a variant of Speth-Strang operator splitting iteratively solving, |
658065c9e9ebbb4db933af2c | 20 | where the constants q j n,vap and q j n,liq for the n th time step are chosen to reduce the error in steady-state solution associated with operator splitting as discussed in Ref. 30 and in our previous work. Note that the transport equations are decoupled in species, so the transport of each species can be solved separately. Also, the reaction equation is decoupled by trays. |
658065c9e9ebbb4db933af2c | 21 | For more details on the simulation scheme, see the SI and our previous work. The presence of oxygen significantly accelerates the film growth process, allowing the film to reach a considerable thickness. As the film grows thicker, the film growth rate transitions into a diffusion-limited regime. Consequently, our previous 6 assumption that the concentration of liquid-phase species absorbed in the film equals the concentration in the bulk liquid is no longer valid. Specifically, the diffusion of dissolved oxygen from the bulk liquid into the film becomes a limiting factor in the growth process. This phenomenon can give rise to distinct regions with different dominating film growth pathways. To investigate these spatial effects on film growth chemistry, we consider axial (z) film growth, assuming the film is uniform in the other directions. There are many components to consider, as illustrated in Fig. . |
658065c9e9ebbb4db933af2c | 22 | There are reactive sites on the solid in the film, participating in liquid+solid reactions to cause film growth. There are species in the liquid absorbed in the film, participating in both liquid+liquid reactions to cause chain initiation, propagation, and termination and liquid+solid reactions to cause film growth. The gradient in species concentration in the absorbed liquid causes diffusive fluxes. As the film grows, the interface between the bulk liquid and the film is moving and more liquid and liquid-phase species are absorbed into the film. We simulate the one-dimensional film growth using the finite volume method and discretize the film in the axial direction. We refer to the control volume at the boundary near the tray surface as i = 1 and the one near the bulk liquid as i = N , where i is the index of each discretization, and N is the total number of discretizations. For boundary conditions, we assume a no-flux boundary at the tray surface, and we assume the liquid-phase species concentrations in the bulk liquid stay at the steady-state concentrations, as illustrated in Fig. . In order to handle the moving boundary, we define the control volume such that it expands when the solid mass within it increases, and it moves in space when control volumes below it grow and push it toward the bulk liquid. We derive the governing equations starting from the macroscopic conservation equation for moving control volume, 31 |
658065c9e9ebbb4db933af2c | 23 | where C is the concentration, V (t) is the time-dependent control volume, S(t) is the timedependent control surface, -→ n is the outward unit normal vector, -→ F is the diffusive flux, ω is the rate of formation per unit volume, -→ F S is the flux caused by the moving control surface. |
658065c9e9ebbb4db933af2c | 24 | Below, we derive the governing equations for solid-bound reactive sites and liquid phase species absorbed in the film, respectively, starting from Eq. 13. Assuming reactive site concentrations within a control volume are homogeneous, we can integrate Eq. 13 for the solid part to obtain the governing equation for reactive sites in control volume i of swollen film on tray j (Eq. 14). We defne the control volume with respect to the solid-bound reactive sites and masses, such that all changes in solid-bound reactive sites and masses are due to liquid+solid reactions, as illustrated in Fig. . |
658065c9e9ebbb4db933af2c | 25 | where N j,i fragm is the number of moles of reactive sites in control volume i of swollen film on the surface of tray j, S liq+solid,fragm is the rectangular 8 × 6,606 stoichiometric matrix for reactive sites (the shape is the number of fragments by the number of liquid+solid reactions), R liq+solid are the rates of the liquid+solid reactions, k f,liq+solid and K eq,liq+solid are the rate constants and the equilibrium constants for the liquid+solid reactions (see the SI of our prior work 6 for the estimation of K eq,liq+solid ), and C j,i liq∈film are the concentrations of liquid-phase species absorbed in the film in control volume i of swollen film on tray j. |
658065c9e9ebbb4db933af2c | 26 | We use the consumption of the liquid-phase species by liquid+solid reactions to track how much mass has been added to the solid part of the film. The corresponding governing equation for the mass of solid in control volume i of tray j (m j,i ) is dm j,i dt = -M T w,liq S liq+solid,liq R liq+solid (15) where S liq+solid,liq is the stoichiometric matrix for liquid-phase species absorbed in the film (with a shape of the number of liquid-phase species by the number of liquid+solid reactions), and M w,liq is the molecular weight of the liquid-phase species. The negative sign is because when a liquid-phase species gets consumed by a liquid+solid reaction to form a covalent bond with the solid, the mass of the solid increases. |
658065c9e9ebbb4db933af2c | 27 | Similarly, we can integrate Eq. 13 for the liquid part to obtain the governing equations for liquid-phase species absorbed in control volume i of swollen film on the surface of tray j (Eq. 16). As illustrated in Fig. ), the number of moles of liquid-phase species in control volume i of tray j (N j,i liq∈film ) can change due to the diffusive flux entering/leaving the control volume i at the top and bottom boundaries (AF liq ), liquid+liquid reactions (Ω liq ), liquid+solid reactions (Ω liq+solid ), translation of the control volume i due to the expansion of control volumes below it, and expansion of the control volume i due to mass accumulation. |
658065c9e9ebbb4db933af2c | 28 | Ω liq+solid = S liq+solid,liq R liq+solid (19) where C j,i liq∈film refers to the concentration of liquid-phase species in control volume i, and C j,N +1 liq∈film = C j liq,S.S. . D liq is the diffusion coefficient for liquid-phase species, i ± 1/2 refers to the top/bottom boundary of the control volume i, V j,i liq∈film is the volume of liquid part in control volume i, V j,i liq∈film is the rate of change for the volume of liquid part in control volume i. V j,i liq∈film and V j,i liq∈film are computed using |
658065c9e9ebbb4db933af2c | 29 | Film growth on the tray surface can arise from either non-radical or radical pathways. The main non-radical pathways are Diels-Alder addition reactions involving the monomers in the liquid absorbed in the film and the reactive sites on the solid, while the main radical pathways include the propagation of liquid-phase radicals with solid-bound double bonds, or solid-bound radicals with liquid-phase monomers. |
658065c9e9ebbb4db933af2c | 30 | For non-radical pathways, the influence of O 2 is marginal, as the concentration of monomers in the absorbed liquid remains similar to that in the bulk liquid, which predominantly depends on the convective flows, similar to the previously published base case (zero oxygen) scenario. However, for radical pathways, O 2 has a much larger impact because the concentration of participating radicals in the absorbed liquid is primarily governed by diffusive fluxes and liquid+liquid reactions. The liquid+liquid reactions affecting the radical concentrations in the bulk liquid include both oligomerization and autoxidation. |
658065c9e9ebbb4db933af2c | 31 | Radicals in the bulk liquid can originate via inter-tray convective flow, vapor-liquid mass transfer, or they can be generated via chain initiation reactions, as shown in Fig. . The relative significance of these sources varies with location within the distillation column. In hotter regions with low dissolved oxygen concentrations, such as the reboiler and trays near it, the dominant source of radicals remains consistent with the base case scenario studied in our previous work: 6 reverse disproportionation reactions between species with loosely bonded hydrogens (allylic hydrogens and bisallylic hydrogens) and conjugated dienes produce allylic radicals and bisallylic radicals, maintaining a relatively high radical concentration, as shown in Fig. (a) and (d). Many of these radicals have low molecular weight, and the elevated temperature in the reboiler facilitates their evaporation. They are carried upward by the convective flow of vapor to the condenser, where the temperature is low enough that they condense back into the liquid. |
658065c9e9ebbb4db933af2c | 32 | can further convert into hydroperoxides or peroxides. The downward liquid convective flow transports these hydroperoxides and peroxides to the bottom section of the column. In this region, where the temperature is comparatively higher, a fraction of the hydroperoxides and peroxides decompose, turning into peroxyl and alkoxyl radicals and becoming the source of trace oxygen compound contamination despite negligible dissolved molecular oxygen, as shown in Fig. . In all trays, the radical reaction timescales are shorter than the residence time, so most radicals react. However, radical slows into and out of trays are often significant relative to the slow net radical formation reactions. In the liquid-phase system, in addition to the Diels-Alder (DA) polymerization and radical polymerization in the base case scenario, the presence of dissolved oxygen gives rise to alkylperoxyl propagation chain, as illustrated in Fig. . For conciseness, the illustration omits the substituent groups of the monomers. For radical polymerization, monomers can be non-conjugated alkenes as well as dienes. The growing chain can either be alkyl or alkylperoxyl. |
658065c9e9ebbb4db933af2c | 33 | For DA polymerization, the elongation probability of an oligomer was found to be negligible in our previous work. For radical polymerization, the elongation probability depends on various types of radical consumption rates, fluxes, and reactions. The equation approximating the elongation probability for carbon-center growing chains (α RC. ), accounting for the effects of dissolved oxygen and oxygen chemistry, is as follows: |
658065c9e9ebbb4db933af2c | 34 | α RC. = r RC.,p r RC.,p + r RC.,f + r RC.,t r RC.,p = r RC. add + r RC.+O 2 r RC.,f = r RC. Habs r RC.,t = r RC. recomb + r RC. disprop + r RC. cyc + F out, RC. + F evap, RC. (26) where r RC. stands for the flux of carbon-center radical consumption by a certain reaction type in moles per second, indicated by its subscript (radical addition 'RC. add'; hydrogen atom abstraction 'RC. Habs'; radical recombination 'RC. recomb'; radical disproportionation 'RC. disprop'; combination with oxygen 'RC.+O 2 '; cyclic ether formation 'RC. cyc'), F out, RC. is the flux of carbon-center radicals leaving via the liquid outlet flow, and F evap, RC. is the evaporation flux of carbon-center radicals. Note that hydrogen atom abstraction terminates the growing chain and initiates a new chain. |
658065c9e9ebbb4db933af2c | 35 | where r ROO. stands for the peroxyl radical consumption flux by a certain reaction type in moles per second, indicated by its subscript (radical addition 'ROO. add'; hydrogen atom abstraction 'ROO. Habs'; radical recombination 'ROO. recomb'; radical disproportionation 'ROO. disprop'; concerted HO 2 elimination from peroxyl radical 'ROO. eli'), F out, ROO. is the total out flux of peroxyl radicals, and F evap, ROO. is the total evaporation flux of peroxyl radicals. |
658065c9e9ebbb4db933af2c | 36 | where r RC.+O 2 r RC. add +r RC.+O 2 represents the probability of a carbon-center growing chain transitioning into an oxygen-center growing chain during elongation, while 1 -α ROO. represents the probability of an oxygen-center growing chain not elongating. Fig. (a) to (d) show the computed values of α RC. , α ROO. , α R. , and the corresponding ASF distribution for α R. in the liquid-phase system. In Fig. (a), α RC. is represented by two components: the elongation probability through monomer addition (α RC. add ) and the elongation probability by combining with dissolved oxygen to form an oxygen-center propagation chain. It can be observed that in the colder section, where the oxygen concentration is relatively higher, the carbon-center propagation chains have a greater likelihood of elongating by combining with dissolved oxygen (Fig. ). This transformation leads to the formation of peroxyl propagation chains with higher elongation probabilities (Fig. ). |
658065c9e9ebbb4db933af2c | 37 | The ASF distribution, shown in Fig. , demonstrates that the presence of oxygen allows for the growth of larger and heavier oligomers. These oligomers may become heavy enough to physically deposit onto the tray surface if their concentration gets high enough, but the film growth attributed to monomer-involving pathways still dominates, so we did not attempt to model that process. |
658065c9e9ebbb4db933af2c | 38 | Sensitivities of chain elongation probability shown in Figs. 8 to different dissolved oxygen concentrations can be found in Fig. . The elongation probability of carbon-center radicals turning into peroxyl radicals increases as the dissolved oxygen concentration increases for the colder trays. These peroxyl propagation chains have a larger probability to add to monomers compared to the carbon-center radicals, leading to the increase in the overall elongation. |
658065c9e9ebbb4db933af2c | 39 | Reactions forming covalent bonds between liquid-phase species absorbed in the film and reactive sites on the solid part of the film result in film growth. Understanding the key chemical reactions contributing to film growth is crucial for developing effective mitigation strategies. Film growth can occur through both non-radical and radical pathways. In our previous work, we demonstrated the significant role of radical pathways in the base case where no oxygen was present. |
658065c9e9ebbb4db933af2c | 40 | In this section, we investigate the major liquid+solid reactions that contribute to film growth in the presence of air-saturated dissolved oxygen. We analyze one-dimensional film growth simulation snapshots over a one-year period. Our goal is to gain insights into the impact of oxygen on film growth and explore potential strategies to mitigate fouling in the debutanizer system. Fig. shows the rate of film growth through important reactions at selected trays at the initial time and 1 year. The dominating film growth chemistry changes over time. With the higher concentration of dissolved oxygen in the colder section of the debutanizer, the film growth is initially dominated by oxygen-driven fast-growth pathways. |
658065c9e9ebbb4db933af2c | 41 | There are two possible pathways in which dissolved oxygen can accelerate film growth in the cold section. The first one is through the solid-bound radical pathways. Under an anaerobic condition, there are two types of solid-bound radical reactive sites: the allylic radical reactive site (AR) and the alkyl radical reactive site (KR). With dissolved oxygen, there is enough dissolved oxygen within the film initially. Some of these carbon-center solidbound radicals rapidly combine with an oxygen molecule and convert into peroxyl radical reactive site (PR). All of these solid-bound radicals can convert to one another through propagation reactions with conjugated diene monomers. The monomer concentrations within the film stay similar to those in the bulk liquid. The relatively high concentration of dissolved oxygen and less availability of loosely bonded allylic hydrogens in the colder trays (since the dominant species, butadiene, does not have any very weak C-H bonds) generally promote this film growth cycle. However, our quantum calculations show that PR propagates with conjugated diene monomers at similar k's to AR (Fig. ). Additionally, the rate of production analysis shows that the amount of growth caused by a PR (Fig. ) is similar to that of AR (Fig. ). Considering these, the fast growth caused by dissolved oxygen is not through the solid-bound radical pathways. |
658065c9e9ebbb4db933af2c | 42 | The second one is through the liquid-phase radical pathways. Both the liquid-phase carbon-center and oxygen-center (mostly peroxyl) radicals react with solid-bound carboncarbon double-bond reactive sites (CDB) to cause film growth. There are few solid-bound conjugated diene reactive sites (CD) as they are brought in by the slow propagation of bisallylic radicals, so the propagation of liquid-phase radicals with CD is not significant. |
658065c9e9ebbb4db933af2c | 43 | Our quantum calculations show that these peroxyl radicals react with CDB at a much faster rate (ranging from 1 to 3 orders of magnitude depending on the radical structure, as shown in Fig. ) compared to their corresponding allylic radicals. The rate of production analysis shows that the amount of growth caused by a liquid-phase oxygen-center radical (Fig. ) is several orders of magnitude more than that by a liquid-phase carbon-center radical (Fig. ). Moreover, due to the reason explained in Sec. 5.1, under an aerobic condition, many of these liquid-phase radicals are relatively large and have many carbon-carbon double bonds and peroxides on their backbone. This suggests that peroxyl radicals also bring in more reactive sites onto the film as they react. In short, the dissolved oxygen accelerates the film growth through the liquid-phase radical pathways, by making the liquid-phase radicals more reactive towards the solid. |
658065c9e9ebbb4db933af2c | 44 | As the film grows thicker, the diffusion of dissolved oxygen from the bulk liquid into the film becomes a controlling factor in the growth process. This gives rise to two distinct regions within the film, as illustrated in Fig. . One is the oxygen-driven fast growth region, located near the tray surface where the concentration of dissolved oxygen is sufficiently high, thereby enabling fast oxygen-involved film growth pathways to prevail. The other region is characterized by slower anaerobic growth similar to the situation studied in the previous work. Due to the low concentration of dissolved oxygen in the hotter section of the distillation column, film growth is initially governed by carbon-center radical pathways. As the film grows thicker, the flow of radicals from the bulk into the film becomes diffusion-limited, causing the accumulation of solid-bound radical reactive sites to decrease. Note that the loss of solid-bound radical reactive sites is not affected by film thickness, because the dominant consumption pathways of solid-bound radical reactive sites are via the hydrogen atom abstraction with a loosely bounded hydrogen on a monomer, while the monomer concentration in the film remains about the bulk liquid and is not limited by diffusion. Consequently, the film growth via the Diels-Alder pathways is not affected by diffusion and thus becomes competitive with the carbon-center radical pathways. In the reboiler, the elevated temperature and the higher concentration of cyclopentadiene cause the Diels-Alder pathways to dominate even at early times. |
658065c9e9ebbb4db933af2c | 45 | This suggests that additives such as antioxidants applied in the colder section of the debutanizer have the potential to mitigate film growth at early times, while inhibitors targeting ordinary radical polymerization may be required at a later stage. Inhibitors targeting ordinary radical polymerization may be required in the hotter section of the debutanizer. |
658065c9e9ebbb4db933af2c | 46 | Figure : Illustration of diffusion-limited film growth when oxygen is present in the bulk liquid at the initial time (t 0 ) and 1 year (t f ). As dissolved oxygen is quickly consumed, two film growth regions are created: an oxygen-activated growth region near the tray surface, and an anaerobic growth region with depleted oxygen. |
658065c9e9ebbb4db933af2c | 47 | There are many possible sources of dissolved oxygen and oxygenates in a steam-cracking plant. Traces of dissolved oxygen can enter a distillation column along with the feedstock or the water used for quenching or any additives. The concentration of dissolved oxygen in the feedstock of the distillation column plays a significant role in the fouling growth rate. To investigate this sensitivity, we sample the dissolved oxygen concentration in the feedstock by a factor to its saturated concentration: 1 (feedstock contains saturated oxygen), 10 -1 , 10 -2 , 10 -3 (parts per million level), 10 -4 , 10 -5 , 10 -6 (parts per billion level), and 0 (no oxygen in the feedstock). The system is simulated for 1 year (representative industrial fouling 32 timescale). The results of this sensitivity study are shown in Fig. . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.