Iron phenanthrolines: A density functional theory study

Inorganica Chimica Acta 471 (2018) 391–396

Contents lists available at ScienceDirect

Inorganica Chimica Acta journal homepage: www.elsevier.com/locate/ica

Research paper

Iron phenanthrolines: A density functional theory study Karel G. von Eschwege ⇑, Jeanet Conradie ⇑ Department of Chemistry, PO Box 339, University of the Free State, Bloemfontein 9300, South Africa

a r t i c l e

i n f o

Article history: Received 30 August 2017 Received in revised form 20 November 2017 Accepted 23 November 2017 Available online 26 November 2017 Keywords: Redox indicator Redox potential Cyclic voltammetry DFT

a b s t r a c t A comprehensive DFT study on the most extensive series of 1,10-phenanthroline iron complexes, to date, is reported here. Results have relevance to fields of active research; amongst others that of metal-toligand charge transfer complexes and of redox indicators. ADF geometry optimizations at the BP86/TZP level for a series of twenty-four iron complexes with substituted phenanthrolines for which electrochemical data is available, were obtained. Visible light excitations in these MLCT complexes involve transitions from the upper three metal based HOMO’s to the lower five ligand based LUMO’s. With high accuracy calculated HOMO energies, ionization potentials and Mulliken electronegativities are linearly correlated with experimentally obtained redox data from different studies. Molecular orbital renderings and TDDFT computed oscillators are illustrated to closely predict and explain experimental data. As part of the establishment of a larger data base, also for iron phenanthrolines an experimentally vindicated basis is now presented by which its chemical properties may theoretically be ascertained before embarking on more demanding experimental procedures. Ó 2017 Elsevier B.V. All rights reserved.

1. Introduction Synthesis of 1,10-phenanthrolines was reported as far back as 1930, where Smith used the Skraup reaction to produce phenylenediamines, which in turn were converted to phenanthrolines [1]. Oxidation of phenanthroline yields dipyridyl dicarboxylic acid, and eventually dipyridyl – after decomposition of the acid substituents while heating. Together with terpyridyl, all these compounds represent a group of aromatic nitrogen heterocycles that readily complexes transition metals. The ligand N r-donor ability however is poor, but this is compensated for by these ligands’ ability to be good p-acceptors [2]. Low energy p⁄ orbitals give rise to strong visible spectrum metal-to-ligand charge transfer absorption bands and red-shifted fluorescence in its metal complexes. Therefore, as pointed out by Bencini and Lippolis in an extensive recent review article [3], these ligands are often employed during pH and energy transfer studies – particularly in dye-sensitized solar cells, but also in catalyses, materials and biochemistry. The tris-coordinated iron phenanthroline complex, otherwise known as ferroin, is probably the most well studied complex in this group, especially with regard to its function as redox indicator during photometric determinations of the divalent iron cation. Skoog ⇑ Corresponding authors. E-mail addresses: [email protected] (K.G. von Eschwege), [email protected] (J. Conradie). https://doi.org/10.1016/j.ica.2017.11.047 0020-1693/Ó 2017 Elsevier B.V. All rights reserved.

and West note that ‘‘of all the oxidation/reduction indicators, ferroin approaches most closely the ideal substance. It reacts rapidly and reversibly, its colour change is pronounced, and its solutions are stable and easily prepared” [4]. The oxidized [Fe (phen)3]3+ species is light blue, while the reduced [Fe(phen)3]2+ complex has a deep red colour. Electron withdrawing and/or donating groups substituted on the heterocyclic ring is conveniently used to tune electrochemical and spectral properties of the ligand and its complexes. Hereby the colour/redox transition potentials in aqueous 1 M H2SO4 are 1.25 V for the 5-NO2 derivative, 1.11 V for unsubstituted phenanthroline and 1.02 V for the 5-CH3 species (see substituent position numbering in Fig. 1). A recent study reported redox data for a wide range of electrochemically altered derivatives, albeit in organic (CH3CN) medium [5]. Whereas the formal reduction potential (E00 ) of the 5-NO2 [Fe (phen)3]3+ complex was observed at 0.894 V, E00 for the 5-CH3 analogue lies at 0.669 V. Regardless differences in media and electrode systems resulting in recorded potentials in the organic solvent being lower, the potential difference between the 5-NO2 and 5-CH3 species are nevertheless strikingly similar; in both 1 M H2SO4 and CH3CN media being 0.23 V. Additionally, E00 for unsubstituted terpyridyl, phenanthroline and bipyridyl complexes lie closely grouped together, within 0.038 V from each other [5]. The substituents on the aromatic pyridyl rings therefore have a larger influence on the redox potential of the complex metal than the composition of the ring system, whether phenanthroline, bipyridine or terpyridine.

392

K.G. von Eschwege, J. Conradie / Inorganica Chimica Acta 471 (2018) 391–396

the latter reference also gives a brief review of the very limited reports in this field. At about the same time Lever did an extensive computational study of series of Ru complexes [17]. The NPA and total Mulliken charge on non-innocent ligands in these complexes were correlated with reduction potentials of the Ru complexes. The use of experimental data from a variety of laboratories and the use of different electrodes was presented as reason for the less than optimal scatter seen in correlation graphs. However, the Mulliken charges gave neat linear relationships, as did ELUMO, when homologous series of complexes were involved [18]. Calculated Gibbs free energy is yet another quantity via which experimental redox data may closely be correlated [19]. In 1933 Koopman’s theorem pointed to the direct relationship between ionization energy and the HOMO energy of a molecule [20]. Consequently oxidation potentials would relate to HOMO energies, while LUMO energies relate to reduction potentials. In 2006 Sereda et al. remarked about the convenience of Koopman’s theorem which allows ignoring solvation of the open shell reduced species, while only characterizing the starting compound [21]. 2. Theoretical approach

Fig. 1. Fe(phen)3 structure, with ligand substitution numbering scheme indicated.

As opposed to the highest occupied molecular orbital (HOMO) which is localized on the central metal, the lowest unoccupied molecular orbital (LUMO) is delocalized over all three ligands [5]. During oxidation of [Fe(phen)3]2+ an electron is removed from the metal-based HOMO of the complex. Partially related is the electron density that gets transferred from the metal-based HOMO to the ligand-based LUMO during photochemical excitation (ca 510 nm) [5]. This characteristic results in these iron complexes being categorized as metal-to-ligand charge transfer complexes (MLCT). Low-lying p⁄ orbitals, as often seen in aromatic ligands, are commonly associated with these complexes. With the metal being at lower oxidation state, transitions occur at sufficiently low energies to fall within the visible part of the spectrum and thus overlapping with the section of the solar spectrum where maximum irradiation is measured. This is particularly true for the favoured ruthenium bipyridyl and terpyridyl solar cell dyes [6,7]. Applied research in the latter field often requires the gathering of huge amounts of experimental data. With the advent of modern quantum computational capabilities this may however be fasttracked by theoretical simulations. Up to date only limited comprehensive studies had however seen the light, especially due to the large size and consequent expense in computational time required for the molecules here under consideration. The widest range of cyclic voltammetry data obtained from literature was consequently extracted and compared to computed energies and related descriptors, employing density functional theory with the BP86 functional as part of the ADF software package. A series of twenty-four [Fe(phen)3]2+ complexes with various substituents on the phenanthroline ligands and for which cyclic voltammetry are available were geometry optimized, molecular orbital renderings obtained and TDDFT electronic oscillators calculated. Finding best correlations between quantum computational descriptors and electrochemical data had for some time been one objective of our research [8–12]. The most significant outcome has been the very close linear correlations found between computed energies of series of systematically modified compounds and their experimental reduction or oxidation peak potentials. Having started with simple electrochemically well-behaved nitrobenzenes [13], the study advanced to bidentate ligands [14,15], and single as well as bimetallic complexes [16], where

Density functional theory (DFT) calculations were performed using the GGA functional BP86 [22,23] with the TZP (Triple f polarized) basis set as implemented in the Amsterdam Density Functional (ADF2013 and updates) [24] as well as the B3LYP hybrid functional [25,26] with the triple-f basis set 6-311G basis set as implemented in Gaussian 09 [27]. All complexes were computed in the gas phase and spin unrestricted, spin S = 0 for Fe(II), and S = ½ for Fe(I) (reduced) and Fe(III) (oxidized). Frequency analysis did not give any imaginary frequencies, i.e. true minima. Selected DFT calculated energies and thermodynamic data are given in the Table 2. TD-DFT calculations were done with spin restricted and spin unrestricted TD-BP86 for Fe(II) and Fe(III) respectively, calculating the lowest energy 300 excitations. Geometry relaxed (adiabatic) energies of the complexes (N electron system), and the corresponding N  1 (reduced) and N + 1 (oxidized) electron systems were calculated to determine electron affinity (EA), ionization potential (IP) and Mulliken electronegativity (v), by application of the following formulas [28,29]. EA(complex) = E(reduced complex)  E(complex) IP(complex) = E(oxidized complex)  E(complex) v = (IP + EA)/2 3. Results and discussion The X-ray crystal structure of the [Fe(phen)3]2+ complex deviates only slightly from a perfect octahedral geometry. Smallest angles of less than 90°, varying from 82.1 to 83.8°, are associated with the dative covalent bonds that the two nitrogens of any single ligand form with the central metal [30]. Consequently bond angles between adjacent ligands are 90° or larger. Similar results are found in theoretically calculated structures of this study, where again, the angles formed by the bidentate bonds are 82.5°. In certain X-ray crystal structures bond angle and length variations do occur, but this is attributed to crystal packing affects. From about 150 available crystal structures (Cambridge Crystallographic Database [31]) experimental FeAN bond distances in [Fe (phen)3]2+ complexes average at 1.976 Å. CAN bonds within one particular ligand in a complex varies from 1.332 to 1.369 and CAC bonds from 1.351 to 1.436 Å [30], see Table 1. These bond distances are representative of all phenanthroline ligands. Comparison with corresponding theoretically computed bond lengths show close agreement, namely 1.983, 1.344–1.372, and 1.385– 1.436 Å. As for the oxidized [Fe(phen)3]3+ complexes, X-ray data

393

K.G. von Eschwege, J. Conradie / Inorganica Chimica Acta 471 (2018) 391–396 Table 1 Theoretical and experimental (within brackets) bond lengths in di- and tri-valent Fe (phen)3. Bonds Bond lengths/Å FeAN CAN CAC

[Fe(phen)3]2+

[Fe(phen)3]3+

1.983 (average: 1.976) 1.344–1.372 (1.332–1.369) 1.385–1.436 (1.351–1.436)

1.997 (average: 1.978) 1.343–1.372 (1.31–1.41) 1.387–1.436 (1.33–1.45)

from only ten experimental structures are available. Here the computed FeAN bond lengths are up to 0.02 Å longer than corresponding experimental distances. Variation amongst experimentally determined M–L distances in both the divalent and trivalent species are however large, being significantly affected by packing effects, counter ions, substituents, etc. Bond lengths within the ligands are, as with the divalent complexes, computationally very closely simulated. Phenanthroline and the bi- and terpyridines in iron and ruthenium complexes are traditionally well known for its role as negative charge acceptors, this being the reason for its involvement in metal to ligand charge transfer processes (MLCT) [6]. Metal d-orbitals give rise to the HOMO, while p-orbitals of the ligand carbon and nitrogen atoms accommodate the LUMO. Fig. 2 illustrates the directional charge transfer that takes place during photochemical excitation of the upper HOMO’s of the [Fe(phen)3]2+ complex, i.e. the HOMO’s are almost solely localized on the central divalent iron atom, while the LUMO’s are spread over the ligands. The top three HOMO’s and the lowest five LUMO’s are involved in the lowest energy transitions, lying above 450 nm. (Wavelengths/energies at which all these excitations occur, with corresponding oscillator strengths and weight factors are listed in Table S1 under Supporting

Information). These eight frontier orbitals are well separated from the other molecular orbitals, i.e. HOMO-4 and lower, and LUMO + 6 and higher. The three upper HOMO’s of these d6 Fe2+ complexes are localized at the central metal (the t2g set of orbitals of an octahedral ligand field), while LUMO’s 1–5 are delocalized over the ligands. LUMO and LUMO + 3 involve all three ligands while the others are spread over only two ligands. The higher LUMO + 6 and LUMO + 7 in turn are metal based (see Supporting Information Fig. S1). This eg set of orbitals is ca 2.7 eV higher in energy than the t2g set. Fig. 3 shows the BP86/TZP TDDFT computed electronic oscillators of both the [Fe(phen)3]3+ and [Fe(phen)3]2+ complexes. The oscillators are overlaid with BP86/TZP calculated spectra, which were obtained by broadening each calculated transition with a Gaussian by a half width of 50 nm. These calculated spectra reproduce the essential features of the experimental spectra [32]. The calculated [Fe(phen)3]2+ oscillators above 620 nm are largely attributed to excitation of the HOMO to all of LUMO’s 1–5, while the strong oscillators between 460 and 620 nm correspond to HOMO-1 and HOMO-2 excitations, also to the five lower LUMO’s (see corresponding data under Supporting Information, Table S1). The pale blue [Fe(phen)3]3+ species has an absorbance spectrum that is significantly blue-shifted relative to [Fe(phen)3]2+, with absorbance maxima in the UV field (experimental kmax < 350 nm [32,33]). As for the red [Fe(phen)3]2+ species, a computed wavelength of maximum absorption is observed at 546 nm, which is slightly higher than the experimental kmax value of ca 510 nm. Calculated oscillator strengths of the two complexes thus follow experimental trends [5,32]. Oscillators of the blue species are not only largely shifted out of the visible field, but are also significantly weaker as compared to the red species. This is expected, since the blue compound is at higher oxidation state, with a decrease in the amount of electrons resulting in a decrease in oscillator strength. Computed results are thus in agreement with the ThomasReiche-Kuhn sum rule which states that the integrated oscillator strength for a molecule is equal to the total number of electrons in the particular molecule [34]. In its application as redox indicator the blue colour of the oxidized species is in fact so faint that titration end points are considered going from colorless to red or vice versa. End points are readily observed when the red coloured [Fe(phen)3]2+ indicator reaches a mere 10% of total indicator concentration [4]. blue

½FeðphenÞ3 

Fig. 2. BP86/TZP frontier molecular orbitals of [Fe(phen)3]2+ involved in the transitions above 450 nm in the BP86/TZP TDDFT-based electronic absorption spectra shown in Fig. 3. The MO plots use a contour of 60 e nm3. Corresponding transition data is listed under Supporting Information.

red 3þ

½FeðphenÞ3 

2þ

ð1Þ

Fig. 3. BP86/TZP TDDFT calculated [Fe(phen)3]2+ (red spikes) and [Fe(phen)3]3+ (blue spikes) oscillators. Simulated spectra (red and blue curves respectively) are obtained by Gaussian broadening of each calculated transition, by a half width of 50 nm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

394

K.G. von Eschwege, J. Conradie / Inorganica Chimica Acta 471 (2018) 391–396

The present study involves the widest range of electrochemical data obtained from various ferroin complex studies found in published literature. A series of twenty-four [Fe(phen)3]2+ complexes were employed, with various substituents on the phenanthroline ligands, as listed in Table 2. Substitution positions are in accordance with the numbering scheme in Fig. 1. Electrochemical data of the 4,7-disubstituted phenanthroline-iron complexes were obtained from five different publications, as indicated in Table 2 [5,35–38]. Depending on the substituent, the formal reduction potential (E00 ) of the FeII/FeIII redox couple in this extended series varies by more than 1.1 V, i.e. from 0.23 (complex 21) to 0.89 V (complex 1) vs Fc/Fc+. Fig. 4 illustrates the correlation between computed HOMO energies and corresponding reduction potentials obtained in different media; the lower data set (red) from 1 M H2SO4 and the upper set (blue) from acetonitrile. Both trendlines neatly fit to the coefficient of determination, R2 = 0.93. Being representative of the same chemical series, with the exception of the amine and a few methyl derivatives, it is surprising to note that the two gradients are almost similar. The effect of varied media on correlation curves within a certain class of compounds therefore manifests as a shift in the y-axis intercept, while gradients stay largely similar. Due to such observed changes it is thus useful to publish all acquired correlations, whether from different electrode systems, electrolytes or solvent media. Hereby a comprehensive data base of correlation formula’s is gradually established. As noted when comparing data from the unsubstituted species, the organic acetonitrile medium yields a positive E00 shift of ca 0.36 V above results obtained in aqueous-acidic medium. By initially computing EHOMO for newly planned dyes or redox indicators related to this series, E00 may therefore be predicted to 93% accuracy when calculated from the equations: E00 = 0.28EHOMO–1.43, in acetonitrile, and

Fig. 4. Linear correlation graphs of experimental formal reduction potentials (E00 ) versus corresponding BP86/TZP calculated HOMO energies (EHOMO) of the substituted iron-phenanthroline series. Derivatives are indicated according to substituent positions shown in Fig. 1. Red: redox transition potentials measured in 1 M H2SO4 [37,38]. Blue: E00 obtained in acetonitrile containing 0.1 mol dm3 tetra-n-butylammonium hexafluorophosphate as supporting electrolyte [5]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

E00 = 0.33EHOMO–2.23, in 1 M H2SO4 (aq). Comparing experimental data sets from two independent studies, Figs. 5 and 6 show additional correlations, namely of E00 with computed ionization potentials (IP), Mulliken electronegativities (v) and HOMO energies. The one set involves data published by Schmittel and co-workers in 1994–1995 on 4,7-disubstituted iron phenanthroline-iron complexes [35,36] (series 1), while the other data set includes differently substituted iron phenanthroline complexes published more recently by us (series 2) [5]. Linear correlations between calculated HOMO energies and formal reduction

Table 2 DFT calculated data of the Fe-phenanthrolines, [Fe(phen)3]2+ (arranged in order of BP86 calculated HOMO energies), and redox transition potentials (ET, in 1 M H2SO4) and formal reduction potentials (E00 ) of the FeII/FeIII redox couple. Substituent positions correspond to numbering in Fig. 1. Nr

a b c

Substituent ET (1 M H2SO4)/V vs NHE

E00 (CH3CN)/V vs Fc/Fc+

E00 (CN3CN)/V vs NHE

0.89a 0.894c 0.802c 0.686a 0.698c – 0.669c 0.585c – – 0.640c – 0.613c 0.38a 0.448a 0.452c 0.56a 0.47a 0.49a 0.45a 0.47a 0.61a 0.22a 0.23a 0.09a 0.23a 0.13a

1.55 1.554 1.462 1.346 1.358 – 1.329 1.245 – – 1.300 – 1.273 1.04 1.108 1.112 1.22 1.13 1.15 1.11 1.13 1.27 0.88 0.43 0.57 0.89 0.79

1

5-NO2

1.25b

2 3

5-Cl H

1.11b 1.06b

4 5 6 7 8 9 10 11 12 13

3-Me 5-Me 5-NH2 3,4-Me 4,7-Me 5,6-Me 3,5,7-Me 4-Me 4,7-OMe 3,4,7,8-Me

1.03b 1.02b – 0.93b 0.87b 0.97b 0.89b –

14 15 16 17 18 19 20 21 22 23 24

4,7-S-Pr 4,7-O-Ph 4,7-O-naph 4,7-O-Ph-Me 4,7-O-Ph-tBu 4,7-S-Ph-Me 4,7-morpholine 4,7-NH(CH2)3NH2 4,7-NEt2 4,7-piperazine 4,7-piperidine

Values from Refs. [35,36]. Values from Refs. [37,38]. BP86 DFT data of complexes 2, 5, 6, 9, 11 from Ref. [5]. Values from Ref. [5].

BP86 b EHOMO /eV

BP86 EA /eV

BP86 IP /eV

10.669

7.90

12.43

10.221 10.018

7.13 6.95

9.996 9.818 9.786 9.745 9.655 9.638 9.578 9.477 9.211 9.160 9.017 8.791 8.657 8.615 8.524 8.470 8.192 7.913 7.884 7.808 7.739

b

BP86 /eV

B3LYP EHOMO /eV

B3LYP G (Free Energy) /eV

10.17

11.88

3592

11.99 11.89

9.56 9.42

11.46 11.26

4357 2978

– 6.80 6.63 – – 6.51 – 6.54 7.12 6.29

– 11.66 11.27 – – 11.59 – 11.26 10.57 10.90

– 9.23 8.95 – – 9.05 – 8.9 8.84 8.59

11.08 11.09 10.55 10.83 10.74 10.93 10.67 10.98 10.14 10.45

3096 3096 3144 3214 3214 3214 3332 3096 3665 3450

6.29 6.11 6.07 5.96 5.87 5.81 5.66 5.39 5.39 5.34 5.24

10.30 10.31 9.34 10.16 9.57 10.71 9.58 8.61 9.14 10.19 9.17

8.29 8.21 7.70 8.06 7.72 8.26 7.62 7.00 7.27 7.76 7.21

10.15 10.03 9.27 9.83 9.74 9.67 9.30 8.97 9.09 8.96 8.90

6075 4816 5738 5052 5759 6990 4698 4350 4253 4814 4482

v

K.G. von Eschwege, J. Conradie / Inorganica Chimica Acta 471 (2018) 391–396

Fig. 5. Linear correlation graphs of experimental formal reduction potentials (E00 ) versus corresponding BP86 (top) and B3LYP (bottom) calculated HOMO energies (EHOMO) for the substituted iron-phenanthrolines. Black: electrochemical data from Refs. [35,36]. Blue: BP86 and electrochemical data from Ref. [5].(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

potentials for series 1 and 2 coincide almost perfectly, where BP86 gives a slightly better correlation than B3LYP, see Fig. 5. Fig. 6 shows BP86 ionization potential and Mulliken electronegativity correlations with E00 . Trend lines through data from these two independent studies are almost exactly the same, in both gradient and y-axis intercept, especially in the case of Mulliken electronegativities. Data from the most recent study (series 2, eight complexes), with a voltage spread of 0.44 V [5], represents both the highest precision and accuracy. The high precision is reflected in high corresponding R2 values of 0.95, as well as the closeness of all data to the linear line. Residual plots of the graphs in Fig. 6, where residuals are near 0, may be viewed under Supporting Information, Fig. S2–5. Interestingly, the data from studies dating earlier (series 1, fifteen complexes), having a wider voltage spread of 1.12 V, nevertheless still gave a relative high accuracy, as reflected by the linear equation being almost similar to that of series 1. However, data from this study has a low precision, as seen in a lower R2 value, i.e. the data being more scattered and residuals deviating from 0 (Figs. S2–5). Comparing correlations that involve data sets obtained by different laboratories was considered imperative, as to get an indication of the universal potential of this method towards predicting redox potentials by theoretical means. With good to excellent accuracies obtained, as pointed out here above, all three computed descriptors may successfully be used to predict redox potentials. It is however to be pointed out that the first correlation (Fig. 5), which involves HOMO energies, only requires a set of geometry optimized Fe(II) structures, without the need for any additional computations or calculations. The IP correlation (Fig. 6 top) also requires a set of optimized Fe(II) structures as well as a set of optimized Fe(III) (oxidized) structures. The v correlation (Fig. 6 bottom) requires three sets op optimized series, namely Fe(II), Fe(III)

395

Fig. 6. Linear correlation graphs of experimental formal reduction potentials (E00 ) versus corresponding calculated (top) Ionization potential (IP) and (bottom) Mulliken electronegativity (v) for the substituted iron-phenanthrolines. Black: electrochemical data from Refs. [35,36]. Blue: IP and electrochemical data from Ref. [5]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(oxidized) as well as Fe(I) (reduced) structures. Instead, both precision and accuracy, as well as simplicity and convenience of correlations with HOMO energies during estimates of oxidation potentials (LUMO energies for reduction potentials) makes EHOMO the descriptor of choice. 4. Conclusion The advent of quantum computational chemistry enables chemical scientists to simulate many aspects of chemistry to a high degree of accuracy when compared to experimental work, eg. structural and orbital geometries, charge distribution and transfer, and energies involved in different processes and levels, as here again illustrated for the relatively large series of metal to ligand charge transfer iron phenanthroline complexes. Acknowledgements This work has received support from the Norwegian Supercomputing Program (NOTUR, Grant No. NN4654K) (JC), the South African National Research Foundation (JC, KvE) and the Central Research Fund of the University of the Free State, Bloemfontein (JC, KvE). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.ica.2017.11.047.

396

K.G. von Eschwege, J. Conradie / Inorganica Chimica Acta 471 (2018) 391–396

References [1] C.R. Smith, Skraup’s reaction applied to the phenylenediamines. Preparation of the phenanthrolines and related dipyridyls, J. Am. Chem. Soc. 52 (1930) 397– 403. [2] G. Anderegg, Pyridinderivate als Komplexbildner V. Die Metallkomplexe von 1,10-Phenanthrolin und a, a0 -Dipyridyl, Pyridinderivate als Komplexbildner VI. Reaktionsenthalpie und -entropie bei der Bildung der Metallkomplexe von 1,10-Phenanthrolin und a, a0 -Dipyridyl, Helv. Chim. Acta 46 (2397–2410) (1963) 2813–2822. [3] A. Bencini, V. Lippolis, 1,10-Phenanthroline: a versatile building block for the construction of ligands for various purposes, Coord. Chem. Rev. 254 (2010) 2096–2180. [4] D.A. Skoog, D.M. West, F.J. Holler, S.R. Crouch, Fundamentals of Analytical Chemistry, 9th Ed., Brook/Cole, 2014, pp. 502–504. [5] H. Ferreira, K.G. von Eschwege, J. Conradie, Electronic properties of Fe charge transfer complexes – a combined experimental and theoretical approach, Electrochim. Acta 216 (2016) 339–346. [6] M. Grätzel, Solar energy conversion by dye-sensitized photovoltaic cells, Inorg. Chem. 44 (2005) 6841–6851. [7] H.S. ElBatal, S. Aghazada, S.A. Al-Muhtaseb, M. Grätzel, M.K. Nazeeruddin, Quasi-solid-state Dye-sensitized solar cells based on Ru (II) polypyridine sensitizers, Energy Technol. 4 (2016) 380–384. [8] J. Conradie, Density functional theory calculations of Rh-b-diketonato complexes, J. Chem. Soc., Dalton Trans. 44 (2015) 1503–1515. [9] M.M. Conradie, J. Conradie, Electrochemical behaviour of Tris (b-diketonato) Iron (III) complexes: a DFT and experimental study, Electrochim. Acta 152 (2015) 512–519. [10] R. Freitag, J. Conradie, Electrochemical and computational chemistry study of Mn(b-diketonato)3 complexes, Electrochim. Acta 158 (2015) 418–426. [11] R. Liu, J. Conradie, Tris(b-diketonato)chromium(III) complexes: effect of the bdiketonate ligand on the redox properties, Electrochim. Acta 185 (2015) 288– 296. [12] A. Kuhn, J. Conradie, Synthesis, electrochemical and DFT study of octahedral bis(b-diketonato)-titanium(IV) complexes, Inorganic Chimica Acta 453 (2016) 247–256. [13] A. Kuhn, K.G. von Eschwege, J. Conradie, Reduction potentials of parasubstituted nitrobenzenes – a theoretical approach, J. Phys. Org. Chem. 25 (2012) 58–68. [14] A. Kuhn, K.G. von Eschwege, J. Conradie, Electrochemical and DFT-modeled reduction of Enolized 1,3-Diketones, Electrochim. Acta 56 (2011) 6211–6218. [15] K.G. von Eschwege, Oxidation resilient dithizones – synthesis, cyclic voltammetry and DFT perspectives, Polyhedron 39 (2012). 99-15. [16] K.G. von Eschwege, J. Conradie, Redox potentials of ligands and complexes – a DFT approach, SA J. Chem. 64 (2011) 203–209. [17] A.B.P. Lever, Electronic characteristics of an extensive series of ruthenium complexes with the non-innocent o-Benzoquinonediimine ligand: a pedagogical approach, Coord. Chem. Rev. 254 (2010) 1397–1405. [18] M.P. Donzello, C. Ercolani, K.M. Kadish, G. Ricciardi, A. Rosa, P.A. Stuzhin, Tetrakis(thiadiazole)porphyrazines. 5. Electrochemical and DFT/TDDFT studies of the Free-base macrocycle and its MgII, ZnII, and CuII complexes, Inorg. Chem. 46 (2007) 4145–4157. [19] T.B. Demissie, K. Ruud, J.H. Hansen, DFT as a powerful predictive tool in photoredox catalysis: redox potentials and mechanistic analysis, Organometallics 34 (2015) 4218–4228. [20] T.A. Koopmans, Ordering of wave functions and eigenenergies to the individual electrons of an atom, Physica 1 (1933) 104–113. [21] G. Sereda, J. van Heukelom, M. Koppang, S. Ramreddy, N. Collins, Effect of transannular interaction on the Redox-potentials in a series of bicyclic quinones, Beilstein J. Org. Chem. 2 (2006) 1–8.

[22] A.D. Becke, Density-functional exchange-energy approximation with correct asymptotic behavior, Phys. Rev. A 38 (1988) 3098–3100. [23] J.P. Perdew, Density-functional approximation for the correlation energy of the inhomogeneous electron gas, Phys. Rev. B33 (1986) 8822–8824. Erratum: J.P. Perdew, Phys. Rev. B 33, 1986, pp. 8822–8824. [24] The ADF program system was obtained from Scientific Computing and Modeling, Amsterdam (http://www.scm.com/). For a description of the methods used in ADF, see: G. te Velde, F.M. Bickelhaupt, E.J. Baerends, C.F. Guerra, S.J.A. van Gisbergen, J.G. Snijders, T.J. Ziegler, Chemistry with ADF, J. Comp. Chem., 22, 2001, 931. [25] A.D. Becke, Density-functional thermochemistry. III. The role of exact exchange, J. Chem. Phys. 98 (1993) 5648–5652. [26] C. Lee, W. Yang, R.G. Parr, Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density, Phys. Rev. B. 37 (1988) 785– 789. [27] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery (Jr), J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision D.01, Gaussian Inc., Wallingford CT, 2010. [28] R.S. Mulliken, A new electroaffinity scale; together with data on valence states and on valence ionization potentials and electron affinities, J. Chem. Phys. 2 (1934) 782. [29] M.V. Putz, N. Russo, E. Sicilia, About the mulliken electronegativity in DFT, Theor. Chem. Acc. 114 (2005) 38–45. [30] I. Potocnak, L. Vahovska, P. Herich, Low-dimensional compounds containing cyanide groups. XXV. Synthesis, spectroscopic properties and crystal structures of two Ionic Iron(II) complexes with tricyanomethanide: Tris (1,10-Phenanthroline-Κ(2)N, N’)Iron(II) Bis(Tricyanomethanide) and Tris (2,2’-Bipyridine-Κ(2)N, N’)Iron(II) Bis(Tricyanomethanide) Sesquihydrate. Acta Crystallogr., Sect. C: Cryst, Struct. Chem. 70 (2014) 432–436. [31] Cambridge Structural Database (CSD), Version 5.38, May 2017 update, Cambridge, UK, 2017. [32] L. Kozak, P. Niedzielski, W. Wachowiak, The tandem analytical method of flow injection diode array spectrophotometry and flame atomic absorption spectrometry (FI-DAD(Vis)-FAAS) in Iron speciation studies using 1,10Phenanthroline complexes, Microchem. J. 110 (2013) 54–60. [33] J. Aihara, Reduced HOMOLUMO Gap as an index of kinetic stability for polycyclic aromatic hydrocarbons, J. Phys. Chem. A 103 (1999) 7487–7495. [34] L. Zheng, N.F. Polizzi, A.R. Dave, A. Migliore, D.N. Beratan, Where is the electronic oscillator strength? Mapping oscillator strength across molecular absorption spectra, J. Phys. Chem. A 120 (2016) 1933–1943. [35] M. Schmittel, H. Ammon, C. Wöhrle, Tris(1,10-phenanthroline)iron(II) complexes – broad variation of the redox potential by 4,7-Substitution at the phenanthroline ligands, Chemische Berichte 128 (1995) 845–850. [36] M. Levis, U. Lüning, M. Müller, M. Schmittel, C. Wöhrle, Tris(l,10phenanthroIine)iron(II) Complexes. Influence of 4,7-Donor Substitution on the Redox Potential, in: Zeitschriftt für Naturforschung B 49 (1994) 675–682. [37] F.W. Brandt, G.F. Smith, Polysubstituted 1,10-phenanthrolines and bipyridines as multiple range redox indicators, Anal. Chem. 21 (1949) 1313–1319. [38] G.F. Smith, F.P. Richter, Substituted 1,10-phenanthroline ferrous complex oxidation-reduction indicators potential determinations as a function of acid concentration, Ind. Eng. Chem. Anal. Ed. 16 (1944) 580–581.