Investigation of the Redox Chemistry of Anthraquinone Derivatives

Aug 27, 2014 - ... properties such as redox windows, effect of substitution by electron donating and electron ... For instance, a complete methylation...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCA

Investigation of the Redox Chemistry of Anthraquinone Derivatives Using Density Functional Theory Jonathan E. Bachman,†,# Larry A. Curtiss,†,‡,§ and Rajeev S. Assary*,†,‡ †

Materials Science Division, ‡Joint Center for Energy Storage (JCESR), and §Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States S Supporting Information *

ABSTRACT: Application of density functional calculations to compute electrochemical properties such as redox windows, effect of substitution by electron donating and electron withdrawing groups on redox windows, and solvation free energies for ∼50 anthraquinone (AQ) derivatives are presented because of their potential as anolytes in all-organic redox flow batteries. Computations suggest that lithium ions can increase (by ∼0.4 V) the reduction potential of anthraquinone due to the lithium ion pairing by forming a Lewis base−Lewis acid complex. To design new redox active species, the substitution by electron donating groups is essential to improve the reduction window of AQ with adequate oxidative stability. For instance, a complete methylation of AQ can improve its reduction window by ∼0.4 V. The quantum chemical studies of the ∼50 AQ derivatives are used to derive a relationship that connects the computed LUMO energy and the reduction potential that can be applied as a descriptor for screening thousands of AQ derivatives. Our computations also suggest that incorporating oxy-methyl dioxolane substituents in the AQ framework can increase its interaction with nonaqueous solvent and improve its solubility. Thermochemical calculations for likely bond breaking decomposition reactions of unsubstituted AQ anions suggest that the dianions are relatively stable in the solution. These studies provide an ideal platform to perform further combined experimental and theoretical studies to understand the electrochemical reversibility and solubility of new quinone molecules as energy storage materials.

■

Aziz et al.14 have reported promising redox flow systems using organic materials such as quinone in aqueous medium, indicating the latest developments in the utilization of metal free organic energy storage materials. Organic materials such as anthraquinone, 14,15 quinoxaline,12 viologen,16 and thiophene17,18 systems (anolytes) can be coupled with a 4 V (or higher) catholyte to form an all organic redox flow battery. For instance, the reported reduction potential of anthraquinone is 2.2 V vs Li/Li+, which can be combined with 2,5-di-tert-butyl1,4-bis(2-methoxyethoxy)benzene19 (oxidation potential of 4 V) to form a redox couple of 1.8 V. Given the vast number of organic materials as potential electro-active species in organic redox flow systems, the choice of electrolyte materials is unlimited; however, identification of ideal redox pairs is challenging. This requires rapid screening using either electrochemical experiments or computations, where experimentally searching the large electrolyte space is extremely time-consuming and expensive. A viable alternative is the application of a reliable computational approach, where screening the redox properties of a large number of molecules is

INTRODUCTION Efficient and increased energy storage for consumer electronics, electric vehicles, and stationary applications is a challenging problem that requires a fundamental breakthrough to produce better electrochemical systems such as advanced metal intercalation, metal/air, metal/sulfur, and redox flow batteries.1−7 Among these options, a redox flow battery is an electrochemical device that converts the chemical energy in the electro-active materials directly to electrical energy, similar to a conventional battery and fuels cells.8,9 A key difference of flow batteries with respect to conventional batteries is the storage of electro-active materials externally in a liquid (majority) electrolyte and the introduction of these materials to the device only during operation.10 Due to the external storage of electro-active materials, the electrodes of the battery do not undergo degradation and the energy capacity can be controlled by the size of the external storage. Consequently, a redox flow battery system could approach its theoretical energy density as the system is scaled up to a point where the weight or volume of the battery is small relative to that of the stored fuel and oxidant. Among varieties of redox flow batteries, one of the promising candidates is organic redox flow batteries based on nonaqueous solvents due to its potential to operate in large electrochemical window and sustainable nature.11−13 Recently, © XXXX American Chemical Society

Received: June 18, 2014 Revised: August 21, 2014

A

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

nonaqueous solution is treated as zero, similar to what has been used by others.40 The constant 1.24 V, is included to convert the computed reduction potential with respect to a Li/Li+ reference electrode [(−4.28 V41,42) SHE (−3.04 V) Li/Li+)] Note that all redox potentials are reported vs Li/Li+ unless mentioned otherwise. For binding of second electron to the monoanion in the gas phase is thermodynamically uphill (negative electron affinity), whereas inclusion of solvation contributions favors the binding of the second electron. The negative electron affinities result in less accurate reduction potential, but in cases where experimental values are available, agreement is reasonable. It has been found that finite basis sets can give reasonable results in comparison to gas phase experimental results for gas phase temporary anions with negative electron affinities due to a cancellation of errors.38,43 The good agreement between computed second reduction potentials of AQ with the experiment44 suggests that the finite basis set and level of theory used here is giving reasonable results for these −2 anions. The SMD solvation model (water solvent) performs favorably compared with experiments30,44,45 (2.20−2.40 V with respect to Li/Li+) for computation of reduction potentials (Table S1 of the Supporting Information), and this method is used throughout the manuscript, unless mentioned otherwise. The reduction potential of AQ was also calculated in 22 solvents with dielectric constants ranging from 1.91 (nheptane) to 109.84 (formamide) (Table S2 and Figure S2 of Supporting Information). The reduction potential dramatically increases until a dielectric constant of ∼25 and then remains relatively unaffected with the increasing dielectric constant. We also note that geometry optimization in the gas phase or solvent medium does not affect the computed reduction potentials (Table S2 of the Supporting Information), whereas the latter is computationally more intensive than the former. Thus, we have used the B3LYP/6-31G(d) level of theory for optimization and the computation of free energy corrections; subsequently, we performed a single point energy calculation using a “water”-implicit solvation model to estimate solvation contribution to compute the free energy of solvation for redox potential evaluations. Additionally, to determine solvation energies of selected AQ derivatives, an acetonitrile dielectric medium is also used for comparison.

possible on the basis of the descriptors developed from accurate quantum chemical studies. Thus, to screen thousands of molecules for the desired electrochemical window, solubility, and stability, it is essential to develop descriptors from firstprinciples and available experimental data. For instance, quantum chemical studies to understand the electrochemical windows of selected chemical families such as quinoxalines, quinones,20−24 and isoindoles25,26 are available in the literature. Such studies provide some understanding of the electrochemical windows of these molecules that can be used as a first level of screening approach to narrow down a large molecular set. Such a “genome” scale approach was found to be very efficient in discovery of materials27,28 for battery and photovoltaic applications.29 In this investigation, our central aim is to investigate the redox properties of anthraquinone (AQ) using affordable and reliable computational methods. Because AQ is a potential anolyte candidate for an all organic redox flow battery,14,30 using density functional theory, we have performed computations aimed at understanding the following key properties of AQ and its derivatives: 1. Influence of lithium salts and decomposition products of salts on reduction potentials. 2. Fine tuning of reduction or oxidation potential through introduction of various functional groups. 3. Improving the solvation energies of AQ by functional group substitution (50 derivatives). 4. Identifying the stability of the anthraquinone framework upon reduction in solution. Details of the computations are presented in the next section, and the Results and Discussion section elaborates on the computation of accurate electrochemical windows, fine-tuning of the reduction potential of AQ upon substitution, their relative solvation energies, and stabilities of anions in the solution.

■

COMPUTATIONAL DETAILS All computations presented in this paper are performed using the Gaussian 09 software.31 The B3LYP/6-31G(d) level of theory was chosen to compute the geometry and energetics of all the species and to compute free energy corrections required to evaluate the Gibbs free energy at 298 K in the gas phase. The choice of the level of theory is justified by evaluating the electron affinity, which is consistent with the results from highlevel G4MP232 (Figure S1 of Supporting Information). To approximate the solvation contributions, single point energy calculations were performed using the water solvent dielectric and the SMD33 solvation model. The solvation free energy term is added to the gas phase free energy to compute the Gibbs free energy of redox molecule in solution (Gsoln = Ggas + ESMD − Egas). We note that the details of the computation of redox potential using the thermodynamic cycle can be seen elsewhere.34−39 The reduction potential of a redox active molecule in solution can be computed using the following equation; ΔEred =

■

RESULTS AND DISCUSSION Influence of Li Salts and Salt Decomposition Products. In this section, we describe the influence of lithium salts and likely salt decomposition products on the computed first and second reduction potentials of anthraquinone (AQ). There are few experimental studies available in the literature; for example, an experimental study by Zhan et al.46 using LiTFSI/dioxolane (DOL) and dimethoxyethane (DME) solvents shows a pair of redox peaks centering at 2.33 V (2.2−2.5 V Li/Li+) with high symmetry and a peak separation of 0.30 V for AQ. Compton et al.,44 have observed two separate peaks in the absence of lithium ions in aprotic solvents at 2.01 (−1.26 V vs Ag/Ag+) and 1.47 V (−1.80 V vs Ag/Ag+), respectively. They have also demonstrated the occurrence of Li+ ion pairing with a singly reduced AQ molecule, where in the presence of lithium ions (5 mM LiClO4), a dominant first reduction peak in the region of 2.00−2.27 V (−1.0 to −1.20 vs Ag/Ag+) is observed.44 Both studies clearly indicate the influence of lithium ions in affecting the reduction processes of AQ.

−ΔGsoln − 1.24 V nF

where ΔGsoln is the difference in Gibbs free energy in solution upon reduction (Gsoln,reduced − Gsoln,neutral) in eV, n is the number of electrons transferred, and F is the Faraday constant. The change in energy of electrons when going from vacuum to B

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

To understand the effect of various lithium salts and salt decomposition products on the first (E1red) and second reduction potentials (E2red) of AQ, we have explored 21 scenarios (various salts in conjunction with explicit solvent molecules), listed as entries 1−21 in Table 1. Computed Table 1. Influence of Various Salts/Salt Decomposition Products (P) on the Computed First (E1red (V)) and Second (E2red (V)) Reduction Potentials of Anthraquinone (Shown in Entries 1−21)a complexation energies (eV) entry

salt/salt decomposition products (P)

E1red (V)

E2red (V)

ΔH

ΔG

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

no saltb PF5 BF3 (PF5)2 (BF3)2b LiBF4b LiPF6 LiClO4 LiSO3CF3 PCc + LiBF4 PCc + LiPF6 PCc + LiClO4 PCc + LiSO3CF3 (LiBF4)2b (LiPF6)2 (LiClO4)2 (LiSO3CF3)2 (PC)2 + (LiBF4)2b (PCc)2 + (LiPF6)2 (PCc)2 + (LiClO4)2 (PCc)2 + (LiSO3CF3)2 experiment (ref 44) (no Li salt) experiment (ref 30) LiTFSI/ DOL

2.05 3.23 2.99 3.83 3.52 2.48 2.47 2.46 2.53 2.38 2.38 2.29 2.38 2.74 2.79 2.80 2.84 2.66 2.62 2.68 2.71 2.01 2.20

1.61 2.74 2.56 3.57 3.25 2.03 2.06 2.03 2.01 1.82 1.92 1.91 1.89 2.33 2.32 2.35 2.36 2.07 2.15 2.10 2.20 1.47 2.50

NA −0.48 −0.41 −0.71 −0.57 −1.72 −1.54 −1.91 −1.75 −2.58 −2.44 −2.66 −2.40 −3.17 −2.80 −3.55 −3.17 −4.92 −5.28 −5.07 −4.72 NA NA

NA −0.15 −0.05 −0.05 0.06 −1.32 −1.13 −1.56 −1.36 −1.68 −1.54 −1.79 −1.56 −2.40 −2.02 −2.78 −2.35 −3.09 −3.51 −3.33 −2.95 NA NA

Figure 1. Optimized structures of anthraquinone (AQ, entry 1 Table 1), AQ-complexed with two BF3 molecules (entry 5 in Table 1), AQcomplexed with two LiBF4 molecules (entry 14 in Table 1), and AQcomplexed with two LiBF4 and two explicit PC molecules (entry 18 in Table 1) at the B3LYP level of theory.

Three notable trends can be seen from Table 1. First, coordination with salt decomposition products such as PF5 and BF3 significantly increases the first reduction potential of AQ (entries 2, AQ:1PF5; 3, AQ:1BF3; 4, AQ:2PF5; 5, AQ:2BF3) by 0.6−1.8 V. The reason for the significant increase in the reduction potential can be explained due to the increase of electron affinity of AQ when complexed with electron withdrawing Lewis acids such as BF3 and PF5. The binding of these molecules with AQ (entries 2−5) are mildly exergonic (ΔG < 0); however, quantitative assessment of their existence and stability in the solution depends on various factors including concentration (BF3/PF5), nature of cosolvents, and temperature. Second, complexation of a salt molecule (LiBF4, LiPF6, LiClO4, or LiSO3CF3) with AQ (entries 6−10) also increases the first reduction potential by ∼0.4 V, whereas complexation of two salt molecules is shown to increase the first reduction potential by ∼0.7 V. Complexation of these salt molecules (entries 6−9) with AQ is exergonic in the range 1.1− 1.5 eV. This complexation free energy can be compared with the free energy required to free up the salt molecules from the solvent/cosolvent shells. For example, the complexation energy of LiBF4 with three propylene carbonate (PC) molecules is exergonic by 1.31 eV; therefore, the presence of some equilibrium amounts of these adducts (entries 6−9) can be expected in the solution. Entries 10−13 are correspond to the complex of AQ with one salt (LiBF4, LiPF6, LiClO4, or LiSO3CF3) and one discrete propylene molecule. The computed first reduction potential using one LiClO4 molecule and one propylene carbonate solvent molecule (entry 12, E1red (V) 2.29 V, E2red (V) 1.91 V) is in good agreement with experimentally measured reduction potentials (E1red (V) 2.37 V, E2red (V) not reported) at 5 mM LiClO4 salt concentration.44 Complexation of two salt molecules (entries 14−17) with AQ are exergonic; however, the presence of such adducts are less likely (in solvents such as PC) and depends on the concentration of salt species and the nature of the aprotic solvent medium. In general, we also note that the model

a

Computed enthalpies and free energies of complexation of salt/salt decomposition products (P) with anthraquinone are also shown. All calculations are performed at the B3LYP/6-31G(d) level of theory, and the reduction potentials are given with respect to a Li/Li+ reference electrode. The solvation energies are computed at the B3LYP level of theory using water as the dielectric medium and the SMD solvation model. bSelected structures are shown in Figure 1. c Explicit propylene carbonate (PC) molecule is included in addition to the implicit solvation model.

enthalpies (H) and free energies (G) of formation of these complexes (entries 2−21) from isolated species are also shown in Table 1. In entry 1, the model is simply an unsubstituted AQ molecule. Entry 2 is a complex between PF5 and AQ. For example, the complexation free energy (ΔG) of entry 2 is computed as the difference in free energy of the complex (GComplex) and the sum of the free energy of PF5 (GPF5) and anthraquinone (GAQ) molecule in solution. Entries 22 and 23 are from previous reported experimental measurements.44,46 The optimized structures of entries 1, 5, 14, and 18 are shown in Figure 1 to highlight salt complexation with the AQ molecule. The computed first (2.05 V) and the second reduction potential (1.61 V) of AQ (entry 1) agrees very well with the experimental values reported in the absence of lithium ions (entry 22).44 C

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 2. Computed first (E1red (V), shown in blue, left) and second (E2red (V): red, right) reduction potentials of various anthraquinone derivatives (1−44) at the B3LYP/6-31G(d) level of theory.

decomposition effects on reduction potentials of anthraquinone and the calculations presented including the predicted reduction potentials and complexation energies can be used as a guideline to interpret the experimental outcomes. Influence of Functional Groups on Reduction Potential. Understanding the electronic effect of functional groups on the redox window of AQ is essential to predict new and improved redox active molecules. Additionally, a priori knowledge of the redox behavior of these modified redox active molecules via quantum chemicals studies is valuable for the formulation of descriptors required for a large scale screening and to narrow down the synthesis targets. Here, we have computed the first (E1red) and second (E2red) reduction potentials of 44 AQ derivatives using the simple scenario, i.e., no salt molecule (entry 1 in Table 1). The

consisting of AQ with one salt molecule (entries 6−9) is sufficient enough to predict the redox windows accurately. In almost all entries, the computed first (E1red) and second (E2red) reduction potentials are separated by ∼0.4 V, which is consistent with the experimental value of 0.3 V found by Zhan et al.46 Finally, inclusion of a solvent molecule such as PC (entries 10−13, 18−21) in addition to the salt in the model calculations causes marginal reduction of computed reduction potentials compared to the salt alone (entries 6−9, 14−17) . This can be explained by the weakening of the AQ-Li+ bond in the presence of a nucleophilic solvent such as PC that enables ion pairing. This trend is likely to be true for other solvents such as dimethyl sulfoxide (DMSO), dimethylformamide (DMF), and acetonitrile (AcN). We note that, further experimental studies are required to demonstrate salt/salt D

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 3. Comparison of computed reduction (ΔE1red) and oxidation (ΔE1ox) potentials of selected anthraquinone derivatives (neutral) at the B3LYP/6-31G(d) level of theory. The entries, 1−33, are shown in Figure 2.

Figure 4. Computed first and second reduction potentials for various quinones at the B3LYP/6-31G(d) level of theory.

computed reduction potentials, together with the schematic of all the AQ derivatives are shown in Figure 2. A variety of functional groups were considered including alkyl (methyl, ethyl, tert-butyl), alkoxy, phenyl, halide, acyl, and keto, as shown in Figure 2. These functional groups are either electron withdrawing or electron donating groups, and their general effect on the redox property of AQ derivatives is dictated by this character. For example, the computed first and second reduction potentials of AQ (1 in Figure 2) are 2.05 and 1.61 V, respectively. Upon substitution of eight methyl groups (electron donating groups) on the AQ ring (species 16 in Figure 2); the computed reduction potentials fall to 1.58 and 1.16 V, respectively, for the first and second electron reduction processes. In terms of methyl substitutions, the lowering of computed reduction potentials is found to be directly proportional to the number of methyl substituents. Increasing the number of methyl substituents decreases both the first and second reduction potentials by similar amounts and the location of the methyl groups does not seem to play any role in the computed reduction potential. Further, the location of other electron donating groups including tert-butyl, ethyl, phenyl, and alkoxy substituents does not show any influence on the computed reduction potential. Electron withdrawing groups such as keto (CO), carboxylic acid (COOH), and oxy-methyl dioxolane (entry 37) each increase the reduction potential in a manner similar to that of chlorine. Additionally, AQ with alkoxy

functional groups (entries 37, 38, 40−44) do not show any significant change in the redox windows compared to unsubstituted AQ but are of importance to flow cells due to their solubility, which is discussed later in this paper. The effect of electron donating methyl and electron withdrawing chloro groups (total of 15 molecules) on both the first reduction and first oxidation potentials is shown in Figure 3. Computed oxidation potentials for all these compounds are above 4.5 V, indicating that the oxidative stability of AQ derivatives during flow battery operating conditions is not an issue. In Figure 3, the decrease of the reduction potentials of AQ derivatives upon the methyl substitution is due to a decrease in the electron affinity (proportional to the reduction potential) of the methyl substituted structure. The opposite effect was found when we use electron withdrawing groups such as chlorine (species 23− 33 in Figure 2). For example, substitution of one to eight chlorine atoms in the AQ framework gradually increases the reduction potential from 2.17 to 2.45 V for the first reduction and from 1.71 to 1.99 V for the second reduction process, respectively. Substitution using electron withdrawing molecules increases the reduction potential (Figures 2 and 3) and therefore is not recommended. The computed first reduction potential of fully methylated AQ (species 16) is 1.58 V compared to 2.05 V for AQ. This suggests that for complete methylation the redox window can be increased by about 0.47 V, which suggests these molecules are favorable as an anolytes E

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 5. (a) Computed reduction potentials (with respect to Li/Li+) of various anthraquinone (AQ) derivatives vs computed energies (eV) of lowest unoccupied molecular orbital (LUMO), (b) computed reduction potentials vs computed solvation free energies of AQ derivatives. The entries in the encircled areas I and II are alkoxy substituted AQ derivatives, respectively. All calculations are performed at the B3LYP/6-31+G(d) level of theory using the SMD solvation model. Computed solvation energies of all 44 AQ derivatives are presented in Table S3 of the Supporting Information.

in flow cells. Predictions of the increase in the redox window such as mentioned above are useful parameters to compute the energy density of a given redox couples. Investigation into the influence of core structures other than AQ was also conducted, which is shown in Figure 4. In addition to the three rings in the AQ structure, the reduction potentials of a one-ring (benzoquinone, species 45), two-ring (naphthalene-1,4-dione, species 46), four-ring (tetracene-5,12-dione, species 48), five-ring (pentacene-5,14-dione, species 49), sixring (pentacene-5,7,12,14-tetraone, species 50), and seven-ring (2,2′-(1,3-phenylene)bis(anthracene-9,10-dione), species 51) as well as a three-ring structure within the 1,2-diketo family (anthracene-2,3-dione, species 47) were computed. High reduction potentials correspond to more keto groups and fewer rings and low reduction potentials correspond to fewer keto groups and more rings. This can be explained by the relative abundance of electrons in a phenyl ring versus a keto group, where the phenyl ring is electron rich and the keto group is electron poor, with electron deficient systems having a higher reduction potential. In the case of anthracene-2,3-dione, the reduction potential is substantially increased due to the proximity of the carbonyl carbons, causing an electron deficiency on one side of the molecule and stabilizing the charged species.47 In terms of the reduction window, species such as 49, 50, and 51 are recommended due to relatively lower reduction potentials. Descriptors for Large Scale Screening. In the previous section, it is suggested that the methylation can improve the redox window AQ molecule. To assess the redox window of thousands of molecules, formulation of useful descriptors from our calculations is essential. Computations of solution phase free energies of thousands of molecules is extremely time-consuming, whereas computation of energies of the lowest unoccupied molecular orbitals (LUMO) of molecules is less demanding in terms of computational resources. Because the computed reduction potential is directly proportional to the negative of LUMO energies,34 Figure 5a, we have plotted computed reduction

potentials against LUMO energies (negative of LUMO in eV) of 44 molecules (Figure 2), which is shown in Figure 5a. From the plot a linear equation of the following form can be obtained: y = 0.5323x + 0.6466

where “y” is the computed reduction potential and “x” is the LUMO energy. Based on this equation, prediction of reduction potential is possible by computing the LUMO energy of the redox active molecule, which requires energy evaluation of neutral species. The linearity (R2 = 0.92) of such a prediction for the 44 AQ derivatives is shown in Figure S3 of the Supporting Information, indicating this equation captures the relationship between the LUMO energy and reduction potential. Therefore, the derived equation can be used as a descriptor to assess the first reduction potentials of thousands of AQ derivatives involving the reduction at the aromatic site. Because the second reduction potential of almost all the AQ derivatives computed here (Figure 2) are observed around ∼0.4 V from the first reduction potential, it is expected that this trend exists in other AQ derivatives. In addition to reduction windows, we have compared the solvation free energies of all the 44 AQ derivatives shown in Figure 2. The computed solvation free energies (kcal/mol) are plotted against the computed reduction potentials in Figure 5b (computed solvation free energies of 44 AQ derivatives are also shown in Table S3 of the Supporting Information). To compute the solvation energy of AQ derivatives in organic medium, here we have chosen acetonitrile as the solvent medium. In Figure 5b, two regions are highlighted as I and II, where the computed reduction potentials are relatively lower and the solvation free energies relatively higher. Relatively higher solvation free energy is one of the major parameters controlling the solubility of a solute,48 the other being the sublimation energy of the solid. The computation of the latter parameter requires calculations including molecular crystals, which is beyond the scope of the manuscript. F

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 6. Computed solvation free energies (kcal/mol) and computed first and second reduction potentials (in V vs Li/Li+, both are shown in parentheses) of novel alkoxy derivatives of AQ (entries 53 to 59) computed at the B3LYP/6-31G(d) level of theory.

The redox active molecules shown in the highlighted region of I and II in Figure 5b are AQ derivatives with alkoxy functional groups. Upon alkoxy substitution, AQ derivatives were shown to increase the solvation free energy from 11 to ∼15 kcal/mol. Higher solvation energies indicate favorable interactions of these molecules with nonaqueous solvent molecules and likely exhibit higher solubility. In Figure 6, computed solvation free energies (kcal/mol), first and second reduction potentials (in V, vs Li/Li+, in parentheses) of selected alkoxy derivatives (53−59) of AQ are shown. The molecules (53−39) are similar to entry 37 shown earlier in Figure 2 and are designed with two oxy-methyl dioxolane functional groups, intended to increase the molecule−solvent interactions (solvation energy). Compared to unsubstituted AQ (solvation free energy of 11 kcal/mol in acetonitrile medium), these alkoxy substituted AQ derivatives (entries 53−59) are computed to have higher solvation free energies (solvation free energy of ∼19 kcal/mol in acetonitrile medium), indicating that functionalization of AQ using dioxolane groups can improve the solute−organic solvent interactions. In general, identification of molecules with relatively larger solvation energies is possible with a single point solvation energy calculation, which can be used as a descriptor for a large-scale screening. Understanding the Stability of Redox Active Materials. The stability of redox active materials in the solution and at the electrode surfaces is one of the central properties that determine the longer-term stability of redox flow systems. To understand the stability of a redox family, for example, quinones, thiaphenes, etc., it is crucial to assess the stability of core structure upon oxidation and reduction process. Here, the aromatic cyclic framework of AQ can undergo tworeduction processes; therefore, we have computed the likely bond cleaving reactions of mono- and dianions of AQ. Bond cleaving reactions of the AQ monoanion is highly endergonic and unlikely to occur. In Figure 7, probable bond breaking patterns (6 rxns) and the computed thermochemistry of AQ dianion in the solution are shown. From Figure 7, rxn 1 is a C− C bond breaking, which is endothermic by 0.8 eV. The rest of the reactions are highly endothermic (by 3−6 eV) and unlikely to occur. The reaction energetics is essentially the same in the

Figure 7. Probable bond breaking/forming transformations (rxns 1− 6) of AQ dianions and the computed enthalpies (H) and free energies (G) in eV, at the B3LYP/6-31G(d) level of theory in dielectric solution.

presence of one or two lithium cations (Figure S4 of the Supporting Information). Due to the highly endergonic nature of these reactions, these selected internal degradation reactions of AQ dianion are unlikely. However, the effect of explicit solvent molecules, electrode interactions, and impurities (for water) may play a role in the overall stability of anions of the redox active materials in solution. This requires further molecular levl understanding using both experiment and G

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

■

theory. Further studies also require deriving descriptors that can predict relative stabilities of these redox active materials similar to that proposed by Wang and Dahn et al.49,50

AUTHOR INFORMATION

Corresponding Author

*R. S. Assary. E-mail: [email protected]. Tel: 630-252-3536. Fax: 630-252-9555.

■

CONCLUSIONS The results from the quantum chemical study of ∼50 anthraquinone (AQ) derivatives are presented in this investigation, focusing on the use of these organic molecules as potential anolytes in an all organic redox flow battery. The following conclusions are drawn from this study. 1. An AQ−lithium salt molecular complex can be used as an accurate model to simulate experimental electrochemical window of anthraquinone in the presence of lithium salt. A marginal increase (∼0.4 V) of reduction potential of AQ in the presence of Li+ (Lewis acid) occurs due to its pairing with the AQ molecule (Lewis base). 2. Calculations indicate that complete methylation (electron donating groups) of an AQ molecule can increase the reduction window by 0.47 V with adequate oxidative stability. Substitution of electron withdrawing groups such as chlorine is not recommended due to increase of reduction potential, therefore causing a likely decrease of redox window of a given redox couple. 3. On the basis of computations using 44 AQ derivatives, an expression is derived to predict reduction potentials from the computation of LUMO energy of anthraquinone derivatives. This equation can be used as a descriptor to assess the reduction potential of thousands of anthraquinone (AQ) derivatives without performing detailed free energy calculations for neutral and reduced electronic states. 4. Identification of molecules with relatively larger solvation energies is possible using an implicit solvent model calculation that estimates the contribution of solute− solvent interactions toward the solubility. Anthraquinone derivatives with higher solvation energies are found to be possible using oxy-methyl dioxolane substitutions, with similar reduction window and adequate oxidative stability to the AQ molecule. 5. Thermochemical calculations for likely bond breaking reactions of unsubstituted AQ anions suggest that the dianions are relatively stable in the solution. Future combined experimental and theoretical studies are essential to understand the electrochemical reversibility, stability, and solubility of the proposed molecules as energy storage materials in the flow applications.

■

Article

Present Address

# Department of Chemical Engineering, University of Berkeley, Berkley, CA 94720, USA.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work was supported as part of the Joint Center for Energy Storage Research, an Energy Innovation Hub funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences. J.E.B. acknowledges the Student Research Participation program for funding and the National Energy Research Scientific Computing Center for computational resources. Use of the computational resources of Center for Nanoscale Materials was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

■

REFERENCES

(1) Dunn, B.; Kamath, H.; Tarascon, J. M. Electrical Energy Storage for the Grid: A Battery of Choices. Science 2011, 334 (6058), 928− 935. (2) Bruce, P. G.; Freunberger, S. A.; Hardwick, L. J.; Tarascon, J.-M. Li-O2 and Li-S Batteries with High Energy Storage. Nat. Mater. 2011, 11, 19−29. (3) Rastler, D. Electricity Energy Storage Options: A white paper Primer on Applications, Costs and Benefits. EPRI Technical Update 1020676; ERPI: Palo Alto, CA, 2010. (4) Eyer, J.; Gorey, G. Energy Storage for the Electricity Grid: benefits and Market Potential Assessment Guide. Sandia Report SAND20100815; Sandia National Laboratory: Albuquerque, NM, 2010. (5) Yang, Z. G.; Zhang, J. L.; Kintner-Meyer, M. C. W.; Lu, X. C.; Choi, D. W.; Lemmon, J. P.; Liu, J. Electrochemical Energy Storage for Green Grid. Chem. Rev. 2011, 111, 3577−3613. (6) Poizot, P.; Dolhem, F. Clean Energy new Deal for a Sustainable World: from non-CO2 Generating Energy Sources to Greener Electrochemical Storage Devices. Energy Environ. Sci. 2011, 4, 2003− 2019. (7) Armand, M.; Tarascon, J. M. Building Better Batteries. Nature 2008, 451, 652−657. (8) Skyllas-Kazacos, M.; Chakrabarti, M. H.; Hajimolana, S. A.; Mjalli, F. S.; Saleem, M. Progress in Flow Battery Research and Development. J. Electrochem. Soc. 2011, 158, R55−R79. (9) Wang, W.; Luo, Q. T.; Li, B.; Wei, X. L.; Li, L. Y.; Yang, Z. G. Recent Progress in Redox Flow Battery Research and Development. Adv. Funct. Mater. 2013, 23, 970−986. (10) Weber, A. Z.; Mench, M. M.; Meyers, J. P.; Ross, P. N.; Gostick, J. T.; Liu, Q. H. Redox Flow Batteries: a Review. J. Appl. Electrochem. 2011, 41, 1137−1164. (11) Song, Z.; Zhou, H. Towards Sustainable and Versatile Energy Storage Devices: an Overview of Organic Electrode Materials. Energy Environ. Sci. 2013, 6, 2280−2301. (12) Brushett, F. R.; Vaughey, J. T.; Jansen, A. N. An All-Organic Non-aqueous Lithium-Ion Redox Flow Battery. Adv. Energy Mater. 2013, 2, 1390−1396. (13) Renault, S.; Gottis, S.; Barres, A.-L.; Courty, M.; Chauvet, O.; Dolhem, F.; Poizot, P. A Green Li-Organic Battery Working as a Fuel Cell in Case of Emergency. Energy Environ. Sci. 2013, 6, 2124−2133. (14) Huskinson, B.; Marshak, M. P.; Suh, C.; Er, S.; Gerhardt, M. R.; Galvin, C. J.; Chen, X.; Aspuru-Guzik, A.; Gordon, R. G.; Aziz, M. J. A Metal-free Organic-Inorganic Aqueous Flow Battery. Nature 2014, 505, 195−198.

ASSOCIATED CONTENT

S Supporting Information *

Assessment of basis sets and density functionals (Figure S1) to evaluate electron affinity and dielectric models (Table S1)/ dielectric mediums (Table S2 and Figure S2) to compute reduction potentials, computed solvation energies of AQ derivatives in acetonitrile solvent dielectric medium (Table S3), linearity plot of predicted reduction potentials of 44 AQ derivatives (Figure S3), energetics involved in the various bondbreaking patterns of AQ anions (Figure S4), and complete citation of ref 31 are given in the Supporting Information.This material is available free of charge via the Internet at http:// pubs.acs.org. H

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Constant and Atomic Surface Tensions. J. Phys. Chem. B 2009, 113, 6378−6396. (34) Moens, J.; Geerlings, P.; Roos, G. A Conceptual DFT Approach for the Evaluation and Interpretation of Redox Potentials. Chem. −Eur. J. 2007, 13, 8174−8184. (35) Borodin, O.; Behl, W.; Jow, T. R. Oxidative Stability and Initial Decomposition Reactions of Carbonate, Sulfone, and Alkyl PhosphateBased Electrolytes. J. Phys. Chem. C 2013, 117, 8661−8682. (36) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. Single-Ion Solvation Free Energies and the Normal Hydrogen Electrode Potential in Methanol, Acetonitrile, and Dimethyl Sulfoxide. J. Phys. Chem. B 2006, 111, 408−422. (37) Assary, R. S.; Curtiss, L. A.; Redfern, P. C.; Zhang, Z.; Amine, K. Computational Studies of Polysiloxanes: Oxidation Potentials and Decomposition Reactions. J. Phys. Chem. C 2011, 115, 12216−12223. (38) Assary, R. S.; Curtiss, L. A.; Moore, J. S. Toward a Molecular Understanding of Energetics in Li−S Batteries Using Nonaqueous Electrolytes: A High-Level Quantum Chemical Study. J. Phys. Chem. C 2014, 118, 11545−11558. (39) Phillips, K. L.; Sandler, S. I.; Chiu, P. C. A Method to Calculate the One-electron Reduction Potentials for Nitroaromatic Compounds Based on Gas-Phase Quantum Mechanics. J. Comput. Chem. 2011, 32, 226−239. (40) Guerard, J. J.; Arey, J. S. Critical Evaluation of Implicit Solvent Models for Predicting Aqueous Oxidation Potentials of Neutral Organic Compounds. J. Chem. Theory Comput. 2013, 9, 5046−5058. (41) Isse, A. A.; Gennaro, A. Absolute Potential of the Standard Hydrogen Electrode and the Problem of Interconversion of Potentials in Different Solvents. J. Phys. Chem. B 2010, 114, 7894−7899. (42) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. Aqueous Solvation Free Energies of Ions and Ion-Water Clusters Based on an Accurate Value for the Absolute Aqueous Solvation Free Energy of the Proton. J. Phys. Chem. B 2006, 110, 16066−16081. (43) Szarka, A. Z.; Curtiss, L. A.; Miller, J. R. Calculation of Temporary Anion States using Density Functional Theory. Chem. Phys. 1999, 246, 147−155. (44) Wain, A. J.; Wildgoose, G. G.; Heald, C. G. R.; Jiang, L.; Jones, T. G. J.; Compton, R. G. Electrochemical ESR and Voltammetric Studies of Lithium Ion Pairing with Electrogenerated 9,10Anthraquinone Radical Anions Either Free in Acetonitrile Solution or Covalently Bound to Multiwalled Carbon Nanotubes. J. Phys. Chem. B 2005, 109, 3971−3978. (45) Weng, W.; Barile, C. J.; Du, P.; Abouimrane, A.; Assary, R. S.; Gewirth, A. A.; Curtiss, L. A.; Amine, K. Polymer Supported Organic Catalysts for O2 Reduction in Li-O2 Batteries. Electrochim. Acta 2014, 119, 138−143. (46) Song, Z.; Zhan, H.; Zhou, Y. Anthraquinone Based Polymer as High Performance Cathode Material for Rechargeable Lithium Batteries. Chem. Commun. 2009, 4, 448−450. (47) Nokami, T.; Matsuo, T.; Inatomi, Y.; Hojo, N.; Tsukagoshi, T.; Yoshizawa, H.; Shimizu, A.; Kuramoto, H.; Komae, K.; Tsuyama, H.; Yoshida, J.-i. Polymer-Bound Pyrene-4,5,9,10-tetraone for Fast-Charge and -Discharge Lithium-Ion Batteries with High Capacity. J. Am. Chem. Soc. 2012, 134, 19694−19700. (48) Palmer, D. S.; Llinos, A.; Morao, I. a.; Day, G. M.; Goodman, J. M.; Glen, R. C.; Mitchell, J. B. O. Predicting Intrinsic Aqueous Solubility by a Thermodynamic Cycle. Mol. Pharmaceutics 2008, 5, 266−279. (49) Wang, R. L.; Dahn, J. R. Computational Estimates of Stability of Redox Shuttle Additives for Li-Ion Cells. J. Electrochem. Soc. 2006, 153, A1922−A1928. (50) Chen, J.-H.; He, L.-M.; Wang, R. L. Correlation between the Stability of Redox Shuttles in Li Ion Cells and the Reactivity Defined by the Binding Energy of Redox Shuttle Cations with Ethyl Radical. J. Electrochem. Soc. 2012, 159, A1636−A1645.

(15) Huskinson, B.; Nawar, S.; Gerhardt, M. R.; Aziz, M. J. Novel Quinone-Based Couples for Flow Batteries. ECS Trans. 2013, 53, 101−105. (16) Bruinink, J.; Kregting, C. G. A. The Voltammetric Behavior of Some Viologens at SnO2 Electrodes. J. Electrochem. Soc. 1978, 125, 1397−1401. (17) Henderson, J. C.; Kiya, Y.; Hutchison, G. R.; Abruna, H. c. D. 5,5′-Bis(methylthio)-2,2′-bithiophene: A Potential Cathode Electroactive Material for Energy Storage Devices. J. Phys. Chem. C 2008, 112, 3989−3997. (18) Zhou, W. D.; Hernandez-Burgos, K.; Burkhardt, S. E.; Qian, H. L.; Abruna, H. D. Synthesis and Electrochemical and Computational Analysis of Two New Families of Thiophene-Carbonyl Molecules. J. Phys. Chem. C 2013, 117, 6022−6032. (19) Zhang, L.; Zhang, Z.; Redfern, P. C.; Curtiss, L. A.; Amine, K. Molecular Engineering Towards Safer Lithium-ion Batteries: a Highly Stable and Compatible Redox Shuttle for Overcharge Protection. Energy Environ. Sci. 2012, 5, 8204−8207. (20) Johnsson Wass, J. R. T.; Ahlberg, E.; Panas, I.; Schiffrin, D. J. Quantum Chemical Modeling of the Reduction of Quinones. J. Phys. Chem. A 2006, 110, 2005−2020. (21) Namazian, M.; Coote, M. L. Accurate Calculation of Absolute One-Electron Redox Potentials of Some para-Quinone Derivatives in Acetonitrile. J. Phys. Chem. A 2007, 111, 7227−7232. (22) Namazian, M.; Almodarresieh, H. A.; Noorbala, M. R.; Zare, H. R. DFT Calculation of Electrode Potentials for Substituted Quinones in Aqueous Solution. Chem. Phys. Lett. 2004, 396, 424−428. (23) Raymond, K. S.; Grafton, A. K.; Wheeler, R. A. Calculated OneElectron Reduction Potentials and Solvation Structures for Selected pBenzoquinones in Water. J. Phys. Chem. B 1997, 101 (4), 623−631. (24) Hernández-Burgos, K.; Burkhardt, S. E.; Rodríguez-Calero, G. G.; Hennig, R. G.; Abruña, H. D. Theoretical Studies of CarbonylBased Organic Molecules for Energy Storage Applications: The Heteroatom and Substituent Effect. J. Phys. Chem. C 2014, 118, 6046− 6051. (25) Karlsson, C.; Jomstorp, E.; Stromme, M.; Sjodin, M. Computational Electrochemistry Study of 16 Isoindole-4,7-diones as Candidates for Organic Cathode Materials. J. Phys. Chem. C 2012, 116, 3793−3801. (26) Karlsson, C.; Gogoll, A.; Stromme, M.; Sjodin, M. Investigation of the Redox Chemistry of Isoindole-4,7-diones. J. Phys. Chem. C 2013, 117, 894−901. (27) Wang, L.; Maxisch, T.; Ceder, G. Oxidation Energies of Transition Metal Oxides within the GGA+U Framework. Phys. Rev. B 2006, 73, 195107−195114. (28) Jain, A.; Ong, S. P.; Hautier, G.; Chen, W.; Richards, W. D.; Dacek, S.; Cholia, S.; Gunter, D.; Skinner, D.; Ceder, G.; Persson, K. A. Commentary: The Materials Project: A Materials Genome Approach to Accelerating Materials Innovation. APL Mater. 2013, 1, 011002−011014. (29) Hachmann, J.; Olivares-Amaya, R.; Jinich, A.; Appleton, A. L.; Blood-Forsythe, M. A.; Seress, L. s. R.; Roman-Salgado, C.; Trepte, K.; Atahan-Evrenk, S.; Er, S. l. Lead Candidates for High-performance Organic Photovoltaics from High-throughput Quantum Chemistry“The Harvard Clean Energy Project. Energy Environ. Sci. 2014, 7, 698− 704. (30) Wang, W.; Xu, W.; Cosimbescu, L.; Choi, D. W.; Li, L. Y.; Yang, Z. G. Anthraquinone with Tailored Structure for a Nonaqueous Metalorganic Redox Flow Battery. Chem. Commun. 2012, 48, 6669−6671. (31) Frisch, M. J.; et al. Gaussian 09, Revision B.01; Gaussian, Inc.: Wallingford, CT, 2009 (Supporting Information for complete citation). (32) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. Gaussian-4 Theory Using Reduced Order Perturbation Theory. J. Chem. Phys. 2007, 127, 124105. (33) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal Solvation Model Based on Solute Electron Density and on a Continuum Model of the Solvent Defined by the Bulk Dielectric I

dx.doi.org/10.1021/jp5060777 | J. Phys. Chem. A XXXX, XXX, XXX−XXX