Exploration of N-Ferrocenyl Substituted Thioamides: Synthesis, Properties and Reactivity Dissertation Zur Erlangung des Grades “Doktor der Naturwissenschaften” Im Promotionsfach Chemie Am Fachbereich Chemie, Pharmazie und Geowissenschaften Der Johannes Gutenberg-Universität Mainz Vorgelegt von Torben Kienz Geboren in Worms Mainz, 2016 Die vorliegende Arbeit wurde unter Betreuung von ___ in der Zeit von Juni 2011 bis Oktober 2016 am Institut für Anorganische Chemie und Analytische Chemie der Johannes Gutenberg-Universität Mainz angefertigt. Mainz, Oktober 2016 Dekan: 1. Berichterstatter: 2. Berichterstatter: Tag der mündlichen Prüfung: 11. November 2016 Ich, Torben Kienz, Matrikelnummer 2633374, versichere, dass ich meine Promotionsarbeit selbständig verfasst und keine anderen als die angegebenen schriftlichen und elektronischen Quellen, sowie andere Hilfsmittel benutzt habe. Alle Ausführungen, die anderen Schriften wörtlich oder sinngemäß entnommen wurden, habe ich kenntlich gemacht. Datum Unterschrift Abstract Electron transfer reactions are fundamental for biological processes. Since the investigation of such processes in vivo are complicated or sometimes impossible, model complexes for biological system have been developed as a valuable tool. The investigation of electron transfer processes via the use of ferrocene derivatives provides insight into the influence of the bridging unit for the rate of electron transfer. Incorporation of asymmetric bridges, such as carboxamides, provides a model for protein environments and have been studied thoroughly. The thioamide as structurally related bridging unit, however, has not yet been explored with respect to electron transfer abilities. In this study, N-Ferrocenyl substituted thioamides are easily prepared from the parent carboxamides by the use of Lawesson’s reagent (2,4-bis(p-methoxyphenyl)-1,3- dithiaphosphetane-2,4-disulfide). Remarkably, the rotational barrier of the thioamides is high enough to observe E/Z isomerism in solution. IR and NMR spectroscopy give insight into the hydrogen bonding motif of the thioamides, as well as the secondary structure. The mixed-valent compounds of dinuclear complexes are investigated by electrochemical measurements, as well as UV/Vis and EPR spectroscopy. The electronic coupling HAB is calculated and compared to the parent carboxamides. The reaction pathways of ferrocenium compounds in the presence of a non-nucleophilic base are studied by means of spin trapping techniques. The formation of carbon centered ferrocenyl radicals is reported and the location of the unpaired spin in the ferrocenyl radicals could be determined. This study gives insight into the reaction of ferrocenyl radicals and stabilizing effects on them as a plausible mode of action for ferrocene based (pro-)drugs. Besides the generation of radicals N-ferrocenyl substituted thioamides show a second reactivity upon oxidation and deprotonation. A multistep reaction sequence leads to the formation of a novel N,S-heterocycle, which initiates oligomerization. Intermediates of this sequence include paramagnetic piano stool like complexes with η1-coordinated cyclopentadienyl rings, as well as super electrophilic ferrocenyl ketenimine cations, formed by elimination of hydrogen sulfide. A reaction mechanism is proposed and supported by mass spectrometry, EPR and NMR spectroscopy and DFT calculations. This work is crucial for polymeric materials with CN-backbone and heterocyclic chemistry. Kurzzusammenfassung Für biologische Systeme ist Elektronentransfer ein fundamentaler Prozess. Da die Untersuchung solcher Prozesse in vivo kompliziert und oft unmöglich ist, wurden Modellkomplexe für biologische Systeme als nützliches Werkzeug zu deren Erforschung entwickelt. Die Untersuchung von Elektronentransfer an Ferrocenderivaten liefert Einsicht über den Einfluss der verbrückenden Gruppen auf die Geschwindigkeit des Elektronentransfers. Asymmetrisch verbrückende Gruppen, wie Carboxamide, welche ein Modell für Proteinumgebungen darstellen, wurden gründlich untersucht. Thioamide hingegen, die in ihrer geometrischen Struktur ähnlich sind, wurden bisher noch nicht im Hinblick auf ihre Elektronentransfereingenschaften erforscht. In dieser Studie wird eine einfache Synthese N-Ferrocenyl substituierte Thioamide durch den Einsatz von Lawessons Reagenz (2,4-Bis(p-methoxyphenyl)-1,3- dithiaphosphetan-2,4-disulid) aus den entsprechenden Carboxamiden berichtet. Die Rotationsbarriere von Thioamiden ist bemerkenswerterweise groß genug, um E/Z-Isomere in Lösung beobachten zu können. IR und NMR Spektroskopie geben hierbei Aufschluss über das Motiv der Wasserstoff-brückenbindungen in Thioamiden, sowie deren Sekundärstruktur. Die gemischt-valenten Verbindungen der dinuklearen Komplexe werden mittels elektrochemischer Messungen, UV/Vis-Spektroskopie und ESR-Spektroskopie untersucht. Der elektronische Kopplungsparameter HAB wird bestimmt und mit dem der entsprechenden Carboxamide verglichen. Die Reaktionswege von Ferroceniumverbindungen in Anwesenheit von nicht- nukleophilen Basen werden mittels „Spin-Trapping“ untersucht. Die Bildung von kohlenstoffzentrierten Ferrocenylradikalen wird untersucht und die Position des ungepaarten Elektrons bestimmt. Diese Forschungsarbeit gibt Aufschluss über die Reaktivität und mögliche stabilisierende Effekte von Ferrocenylradikalen. Dies könnte eine mögliche Erklärung für die Wirkungsweise von ferrocenbasierten Medikamenten sein. Bei Deprotonierung und Oxidation zeigen N-ferrocenyl substituierte Thioamide neben der Bildung von Radikalen eine zusätzliche Reaktivität. Durch eine mehrstufige Reaktionsfolge entstehen neuartige N,S-Heterocyclen, welche Oligomerisationsreaktionen einleiten können. Zwischenprodukte dieser Reaktionsfolge beinhalten paramagnetische klavierstuhlartige Komplexe mit η1-koordinierten Cyclopentadienylringen, sowie ein äußerst elektrophiles Ferrocenylketeniminkation, welches durch die Eliminierung von Schwefelwasserstoff entsteht. Es wird ein Reaktionsmechanismus postuliert und anhand von Massenspektrometrie, ESR-Spektroskopie, NMR-Spektroskopie und DFT- Rechnungen belegt. Diese Arbeit ist äußert wichtig im Hinblick auf polymere Materialen und die Chemie von heterocyclischen Verbindungen. Mein Dank gilt für die Möglichkeit an diesem komplexen Thema zu arbeiten, die wissenschaftliche Betreuung und fruchtbaren Diskussionen, sowie die Ermutigung dort weiter zu forschen, wo andere längst aufgegeben hätten. Damit das Mögliche entsteht, muss immer wieder das Unmögliche versucht werden. Hermann Hesse, Brief an Wilhelm Gundert Every member in the Brown [Ajah] seeks to produce something lasting. Research or study that will be meaningful. Other often accuse us of ignoring the world around us. They think we only look backward. Well, that is inaccurate. If we are distracted, it is because we look forward, toward those who will come. And the information, the knowledge we gather… we leave it for them. The other Ajahs worry about making today better, we yearn to make tomorrow better. Robert Jordan, The Wheel of Time, The Gathering Storm Contents Contents 1 Introduction .............................................................................................................. 1 1.1 Mixed-Valence and Electron Transfer ............................................................. 2 1.1.1 Marcus Theory – Diabatic Approach ....................................................... 2 1.1.2 Marcus-Hush Theory – Adiabatic Approach ........................................... 6 1.1.3 Mixed-valent Compounds ........................................................................ 7 1.1.4 Optical Transitions, Band Shapes and Determination of HAB ............... 10 1.1.5 Half-Wave Potential Splitting in Mixed-valent Systems ....................... 11 1.1.6 Electron Transfer in Ferrocene Compounds .......................................... 13 1.2 Ferrocene Containing Oligopeptides ............................................................. 16 1.2.1 A Structure Comparison of Amides and Thioamides ............................ 21 1.2.2 Thioamides as Isosteric Replacement .................................................... 24 1.2.3 Thioamide Containing Ferrocene Derivatives ....................................... 26 1.3 EPR Spectroscopy and Spin Trapping of Radicals ........................................ 29 1.4 Ferrocene-Containing (Pro-)Drugs ................................................................ 31 1.5 References ...................................................................................................... 36 2 Aim of Work .......................................................................................................... 47 3 Results and Discussion .......................................................................................... 49 3.1 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry ............................................................................................. 53 3.1.1 Abstract .................................................................................................. 55 3.1.2 Introduction ............................................................................................ 56 3.1.3 Results and Discussion .......................................................................... 58 3.1.4 Conclusion ............................................................................................. 72 3.1.5 Experimental Section ............................................................................. 73 3.1.6 Associated Content ................................................................................ 77 I Contents 3.1.7 Acknowledgement ................................................................................. 78 3.1.8 References .............................................................................................. 78 3.2 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene .............................................................................................. 83 3.2.1 Abstract .................................................................................................. 85 3.2.2 Introduction ............................................................................................ 86 3.2.3 Results and Discussion .......................................................................... 90 3.2.4 Conclusion ........................................................................................... 106 3.2.5 Experimental Section ........................................................................... 107 3.2.6 Associated Content .............................................................................. 108 3.2.7 Acknowledgement ............................................................................... 109 3.2.8 Reference ............................................................................................. 109 3.3 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates ................................................................................................ 115 3.3.1 Abstract ................................................................................................ 117 3.3.2 Introduction .......................................................................................... 118 3.3.3 Results and Discussion ........................................................................ 120 3.3.4 Conclusion ........................................................................................... 136 3.3.5 Experimental Section ........................................................................... 137 3.3.6 Associated Content .............................................................................. 144 3.3.7 Acknowledgement ............................................................................... 144 3.3.8 References ............................................................................................ 144 4 Summary and Outlook ......................................................................................... 149 5 Supporting Information ........................................................................................ 153 5.1 To 3.1: Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry ........................................................................................... 155 II Contents 5.2 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene ............................................................................................ 163 5.3 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates ............................................................................. 175 6 Acknowledgements .............................................................................................. 197 7 List of Publications .............................................................................................. 201 8 Curriculum Vitae ................................................................................................. 203 III Abbreviations Abbreviations β angle between Cp ring plane and substituent bond δ chemical shift in NMR spectroscopy Δ difference descriptor ε Molar extinction coefficient θ tilt between two Cp rings λ wavelength λ Marcus reorganization energy µ micro 𝜈 wavenumber χ electronegativity ω Angle along the pseudo-C5 axis in ferrocene A hyperfine coupling constant a.u. arbitrary units or atomic units Ac acetyl Ala alanine Ar aryl B magnetic field Boc tert-butyloxycarbonyl tBu tert-butyl calcd. calculated cat. catalyst CD circular dichroism Cp Cyclopentadienyl (C5H5) Cp* Pentamethyl-cyclopentdienyl (C5Me5) CV Cyclic voltammetry CW Continuous wave d doublet DCM Dichloromethane DDQ 2,3-Dichloro-5,6-dicyano-1,4-benzoquinone DEPMPO 5-(diethoxyphosphoryl)-5-methyl-1-pyrroline-N-oxide DFT density functional theory DIPEA di-iso-propyl amine IV Abbreviations DMF dimethylformamide DMSO dimethylsulfoxide DPPH 2,2-diphenyl-1-picrylhydrazyl dq doublet of quartet E energy E½ half wave potential EPR electron paramagnetic resonance ER estrogen receptor ES excited state Et ethyl ET electron transfer exp experimental F Farraday constant Fc ferrocenyl [Fe (η5-C 5-5H5)(η C5H4)] FD(-MS) field desorption (mass spectrometry) Fmoc fluorenylmethyloxycarbonyl Fn [Fe (η5-C5H4)(η 5-C5H4)] g g-factor G free energy G Gauss Gly glycine GS ground state H hour HAB electronic coupling constant hfc hyperfine coupling constant HOMO highest occupied molecular orbital I nuclear spin IHB intramolecular hydrogen bond IR infrared IVCT intervalence charge transfer K kelvin k rate of electron transfer kB Boltzmann constant V Abbreviations LUMO lowest unoccupied molecular orbital M molar m.p. melting point Me methyl min minute MNP 2-methyl-2-nitrosopropane mT millitesla ND nitrosodurene NIR near infrared nJ coupling constant (NMR spectroscopy) NMR nuclear magnetic resonance NOB nitrosobenzene NOESY nuclear Overhauser effect spectroscopy P t1 Bu tert-butylimino-tris(dimethylamino)phosphorane PBN N-tert-butyl nitrone PCET proton coupled electron transfer Ph Phenyl C6H5 PhNO nitrosobenzene pKa decadic logarithm of acid constant ppm Parts per million pt pseudo triplet q charge R gas constant r.t. room temperature RAB distance between A and B ROS reactive oxygen species S electron spin S singlet SOMO singly occupied molecular orbital SWV square wave voltammetry TFA trifluoracetic acid THF tetrahydrofuran Trp tryptophan VI Abbreviations TS transition state U potential UV/vis ultraviolet-visible V volt Val valine X-band frequency range in EPR spectroscopy (ν≈9.4 GHz) VII VIII Mixed-Valence and Electron Transfer 1 Introduction Iron is believed to make up about 32% of the earth’s elemental composition. This high amount is largely attributed to the fact, that the inner core of the earth is believed to consist entirely of an iron-nickel alloy, while the concentration in the crust of the earth is much lower.[1] Iron is the most used metal with a stock of 2200 kg per capita and therefore essential for human society.[2] But iron is not only essential for human society, but for life itself. It is found in a variety of bioinorganic compounds, for example in hemoglobin[3], myoglobin[4], cytochrome P450[5–12], in iron-sulfur clusters[13,14] and ferritin.[15,16] The desire to understand biological systems, such as the previously ones mentioned, led to synthesis and analysis of a variety of biomimetic model complexes.[17–27] In the field of study regarding biological systems a major subject is electron transfer, which is important in many biological processes.[28–33] While the theoretical foundation has been laid by Marcus and Hush[28,34–40], for practical studies, mixed-valent compounds, especially ferrocene based compunds, have shown to be a valuable tool for the study of electron transfer. This led to the exploration of symmetrically bridged mixed-valent compounds,[41– 45] while studies of asymmetrically bridged systems are still scarce.[43,46–49] Ferrocene has not only proven to be valuable in the study of electron transfer. Since its discovery in 1951[50] and its structure determination shortly thereafter[51], ferrocene has been a fertile field of research. This is attributed to its high stability, reversible redox behavior and easy access to derivatives bearing functional groups. Current research focuses amongst others on ferrocenyl containing polymers[52–54] and pharmaceutical compounds such as anticancer medication,[55,56] antimalarial medication,[57–60] vasorelaxants[61] and antibiotics[62]. With respect to antibiotic activity a poly thioamide extracted from Clostridium cellulolyticum showed remarkable activity against multiresistant staphylococci.[63] In biological systems, the thioamide linkage is very rare. Only five natural products have been found containing thioamide units and one is believed to be a product of decomposition.[64– 68] This fact is very favorable for the use of thioamide linkage in medication, because it provides greater stability against proteolytic degradation and improved absorption, distribution, metabolism and excretion properties.[69–71] 1 Introduction All this proves that iron and especially ferrocene is a fascinating field of research, providing new insights and applications, which can only be more exciting by the addition of thioamide units, opening a horn of plenty for future scientists. 1.1 Mixed-Valence and Electron Transfer Redox reactions are one of the most fundamental reactions in chemistry. These involve per definition the oxidation of one atom or molecule and therefore the reduction of another atom or molecule, in short an electron is transferred from one to the other. While redox reactions have been commonly used since even before chemistry itself was established, the research and understanding of the electron transfer process in view of mechanism and kinetics only started in the late 1940s.[37] Since electron transfer (ET) is such a fundamental process, the knowledge about ET affected a variety of research fields. In the biological field, ET contributed to clarify the nature of intraprotein electron transfer and hence reaction rates of the forward and backward reaction.[72] Future applications originating from electron transfer considerations are processes which require fast electron transfer (molecular wires)[73,74] or prevent back electron transfer reaction (molecular diodes, dye sensitized solar cells, photocatalysis).[75– 78] 1.1.1 Marcus Theory – Diabatic Approach R. A. Marcus made the first approach to quantitative description of electron transfer reactions. While the theory was first developed to describe self-exchange reactions, see Equation (1), it was found, that the theory is applicable to electron transfer reactions in general.[34,37] 𝐹𝑒2+ + 𝐹𝑒∗3+ → 𝐹𝑒3+ + 𝐹𝑒∗2+ (1) The starting point of the theory is, that redox reactions can be described as a motion of atoms on a potential energy surface (see Figure 1). The atoms included in this motion are the ones of the reactants as well as the atoms of the solvent molecules surrounding them and can be described by the nuclear coordinates.[34,37] The transition from the reactants to the products passes a saddle point on this energy surface and can only occur if the system 2 Mixed-Valence and Electron Transfer has enough energy. The same is true for all electron transfer reactions, if the electronic interaction between the two reactants is relatively weak. Figure 1. Profile of the real potential energy surface (red) and simplified profiles (black) for a redox reaction. The basic principle of the theory is, that while thermal electron transfer occurs, the coordinates of the nuclei do not change, as it is for an optical electron transfer (Frank- Condon principle). Because of energy conservation electron transfer can only occur at the saddle point of the energy surface and therefore fluctuations of all coordinates have to occur, to fulfill these criteria.[34,37] This rather complex energy diagram, depending on nuclear motion, is simplified by two approximations. One is the assumption that the energy change for motion of the reactant atoms can be approximated by a harmonic oscillator potential. The second one is a “linear response approximation”, which states, that any change in charge of the reactants (through electron transfer) leads to a proportional change in the polarization of the solvent. This results in free energy curves for the reactant Gr and products Gp, shown in Figure 2, which are simple quadratic functions of one general reaction coordinate.[37] With these simple energy diagrams in mind, the formulas resulting from Marcus theory are easily understood. 3 Introduction Figure 2. Diabatic free energy surface for electron transfer for a) degenerate states and b) nondegenerate states and the possible pathways for electron transfer (optical and thermal). The rate constant for thermal electron transfer is given by Equation (2), an Arrhenius-like equation, where G‡ is the activation barrier and A is determined by whether the electron transfer is intramolecular or intermolecular. G‡ itself is given by Equation (3). G0 is the standard free energy of the reaction and differentiates two cases for electron transfer. The first case for G0 = 0 and therefore degenerate states and the second case for G0  0 resulting in nondegenerate states. is the reorganization energy of the system. −∆𝐺‡ 𝑘 = 𝐴 𝑒𝑥𝑝 ( ) (2) 𝑘𝐵𝑇 2 𝜆 𝛥𝐺0 ∆𝐺‡ = ( ) (3) 4 𝜆 As apparent from the free energy surfaces, there are three possible electron transfer pathways. The thermal one, which was explained before, tunneling and the optical electron transfer, which will now be disclosed. For the case of degenerate states, the optical excitation energy equals the reorganization term. This leads to the understanding, that the reorganization energy equals the energy that is liberated by rearrangement of the system 4 Mixed-Valence and Electron Transfer after an optical electron transfer, to yield the new relaxed geometry. This reorganization can be divided into to two basic terms. One is the inner reorganization energy i, namely the energy arising from the displacement of the product nuclei and one is the outer reorganization energy o, namely the energy to reorient the solvent molecules surrounding the product.[37] In case of non-degenerate states the optical excitation energy equals the sum of reorganization energy and the standard free energy G0. The relation of and G0 has a direct effect on G‡, which can be visualized for three different relations, see Figure 3. For G0 = 0 the activation barrier G‡ is at its maximum (Marcus-normal region). With decreasing standard enthalpy, the activation barrier decreases until G0 = −, where G‡ is zero. For further decreasing standard enthalpy G0 > −, the activation barrier rises again. This regime is called the Marcus-inverted region. Combined with Equation (2), this means that the electrons transfer rate first increases for greater values of G0, to a point where it reaches a maximum rate and then again becomes slower, if the reaction becomes more exothermic.[37] Figure 3. Free energy curves for different values of G0. While Marcus theory in general leads to good quantitative results, a major drawback of this theory is, that it is strictly speaking only applicable to systems with very weak electronic interactions (as stated in the premises of the theory). For systems with a significant 5 Introduction electronic coupling e.g. intramolecular electron transfer reactions, a more sophisticated approach was made by Hush, who extended the Marcus-Theory towards strongly interacting systems.[37] 1.1.2 Marcus-Hush Theory – Adiabatic Approach The Marcus-Hush theory applies when electronic communication between two redox sites is greater than the thermal energy kBT. In this adiabatic approach the electronic communication is quantified by the electron coupling HAB. Coupling of the two diabatic free energy curves for the reactants and products known from the previously discussed Marcus-Theory via HAB leads to the curves depicted in Figure 4. The free energy curves show a double minimum in the adiabatic ground state and a single minimum for the excited state. The energy separation between the ground (GS) and excited state (ES) is exactly 2 HAB for both the degenerate and non-degenerate case. Figure 4. Free energy curves for adiabatic treatment for a) degenerate states and b) non-degenerate states. In the adiabatic case the term for the activation free energy G‡ is not easy to obtain, because of the fact that the electronic coupling leads to a stabilization of the minima of the curve. What basically is done is formulating a term for free energy of the transition state and the reactant minimum depending on HAB and G 0 (because the terms for the reaction coordinate are as well depending on HAB and G 0). The difference of these two energies 6 Mixed-Valence and Electron Transfer equals the free energy of activation G‡.[79] The formula for degenerate states is given in Equation (4), the formula for non-degenerate states in Equation (5). (𝜆 − 2𝐻𝐴𝐵) 2 𝛥𝐺‡ = (4) 4𝜆 2 𝜆 𝛥𝐺0 (𝛥𝐺0) 𝐻2 𝐻4 0 𝛥𝐺‡ = + + − 𝐻 + 𝐴𝐵 − 𝐴𝐵 𝛥𝐺 𝐴𝐵 (5) 4 2 4(𝜆−2𝐻 0 0𝐴𝐵) 𝜆+𝛥𝐺 (𝜆+𝛥𝐺 )4 1.1.3 Mixed-valent Compounds The term mixed-valent was for the first time used by Klotz et al. for a CuI/CuII-Komplex.[80] In general, a mixed-valent system is considered when two or more redox active sites occur in different oxidation states. One of the most famous compounds showing this behavior is the Creutz-Taube-Ion I, see Figure 5. It was first reported in 1969 and incorporates two ruthenium redox centers bridged by a pyrazine ligand and saturated by five ammonia ligands per center. The molecule has an overall charge of 5+, leading to either RuII and RuIII centers or an averaged oxidation state of 2.5.[31,41] This compound was used to intensively study electron transfer for the next thirty years. At the same time the first ferrocene containing mixed-valent compound was published by Cowan et al., biferocenium II with an overall charge of 1+ and therefore either an FeII and FeIII center or a formal oxidation state of +2.5 for both centers.[42,81] Figure 5. Creutz-Taube-Ion I and biferrocenium II In the late 1960s, Robin and Day classified mixed-valent compounds and identified three classes of mixed-valent compounds.[82] Class I, without any electronic coupling. Class II with medium electronic coupling and Class III, with strongly coupled centers. The 7 Introduction difference clarified by comparing the electronic coupling HAB is compared to the Marcus reorganization energy  as demonstrated in Table 1.[83] In Figure 6, the free energy surfaces for class I, II and III are shown. For class I no electronic coupling is observed. The energy diagram shows two distinct curves for the ground states with no crossing point and two strictly separated minima. Therefore the Marcus reorganization energy depicted is merely of theoretical kind. No electron transfer can occur optically or thermally between the two redox centers. Robin-Day-Class Electronic Coupling I 𝐻𝐴𝐵 = 0 II 0 < 2𝐻𝐴𝐵 < 𝜆 III 2𝐻𝐴𝐵 > 𝜆 Table 1. Electronic coupling for the three Robin Day classes. Class II, which features an electronic coupling smaller than the Marcus reorganization energy is described by a double minimum in the electronic ground state (if G0 is significantly smaller than , vide supra) and optical and thermal electron transfer is possible for this compounds. With increasing G0 the energy required for the optical electron transfer and the activation energy for the thermal electron transfer increases as well. If G0 is on the order of  the double minimum vanishes and only one minimum persists. A thermal electron transfer is not possible anymore and an optical excitation does not lead to a charge transfer from one center to the other, because the relaxation converges to the same state as before the optical excitation. Class III features electronic coupling HAB at least as big as , the adiabatic ground state shows a single minimum for the degenerate case. The unpaired electron is completely delocalized between the two redox centers. Optical excitations are still possible, but do not lead to charge transfer, because the excited state is delocalized over both redox centers as well, leading to no change in dipole moment. This concludes to the fact, that a description of neither optical nor thermal electron transfer is meaningful. 8 Mixed-Valence and Electron Transfer Figure 6. Diabatic and adiabatic free energy curves for class I-III with a), c) and f) degenerate and b), d), e) and g) non-degenerate states. 9 Introduction In the case of non-degenerate states the distinction between Class II and Class III is difficult, because there is still a double minimum in the ground state with a very small thermal activation barrier. This leads to the observation, that the classification of the system is dependent on the time scale applied to the system. This has an effect on the analytical method applied, for each spectroscopic method has its own time scale i.e. 10−15 s for optical transitions, 10−11 – 10−12 s for vibrational spectroscopy, 10−9 s for Mössbauer spectroscopy, 10−8 s for EPR spectroscopy, 10−3 – 10−8 s for NMR spectroscopy. Systems with time scale dependent class assignment to class II or III are therefore often summarized in a “fourth” class, namely class II/III.[84] 1.1.4 Optical Transitions, Band Shapes and Determination of HAB The optical transition of class II systems is called intervalence charge transfer band (IVCT). It is often located at small energies, ergo in the near IR in optical spectroscopy. The shape of the band is that of a Gaussian curve, as seen in Figure 25. The shape originates from Boltzmann distributed vibrational states which are occupied for ℎ𝜈 ≪ 𝑘𝐵𝑇. The most probable transition in this setup is the one from the lowest vibrational state of the ground state to the excited state with the same coordinates (and therefore a vibrational excited state). The energy of this transition is the Marcus reorganization energy  plus the standard free energy of the reaction ΔG0 and therefore directly correlated to the maximum of the IVCT band as shown in Equation (6). 𝜈 = 𝜆 + 𝛥𝐺0𝑚𝑎𝑥 (6) The optical transitions of class III systems are also often charge resonance bands (CR) and appear in the same region of the optical spectrum as the IVCT bands of Class II system. The band shape is strongly asymmetric, comparable to a Gaussian band shape with a cutoff at 2HAB, which is the lowest energy for optical transition. The maximum of the IVCT band is still related to the Marcus reorganization energy , but additionally 2HAB can directly be received from the cutoff of the spectrum.[85] 10 Mixed-Valence and Electron Transfer Figure 7. Origin of the band shape for a) class II and b) class III systems. One of the most important derivations of the Marcus-Hush theory is the possibility to extract the value of HAB through band shape analysis of the IVCT band. For class II systems HAB is given by Equation (7). In this Equation 𝜈𝑚𝑎𝑥 is the energy of the maximum of the IVCT band in cm−1, 𝜀𝑚𝑎𝑥 is the extinction coefficient at the maximum, Δ𝜈0.5 equals the bandwidth at half height in cm−1, and RAB is the charge transfer distance in Å. (𝜈 𝜀 0.5𝑚𝑎𝑥 𝑚𝑎𝑥𝛥𝜈0.5) 𝐻𝐴𝐵 = 0.0206 (7) 𝑅𝐴𝐵 For use of Equation (7) RAB is often set as the distance between the two redox sites, because the actual effective charge transfer distance is very difficult to measure experimentally (e.g. by means of Stark spectroscopy). This in general leads to an overestimation of RAB and an underestimation of H [86]AB. For strongly coupled systems i.e. class III Equation (8) is valid, where HAB is given by half the energy of the maximum of the transition band. 𝜈𝑚𝑎𝑥 𝐻𝐴𝐵 = (8) 2 1.1.5 Half-Wave Potential Splitting in Mixed-valent Systems The above mentioned thermodynamic stabilization of mixed-valent compounds has a large impact on the electrochemical analysis of these compounds. As apparent from Equation (10) the potential splitting between the first oxidation to the mixed-valent compound and 11 Introduction subsequent oxidation ∆𝐸1/2 is dependent on the comproportionation constant of the compounds. Therefore the half-wave potential splitting is directly correlated to the free energy change of the system −∆𝐺𝑐 upon oxidation by Equation (11). 𝑀1𝑀2 → 𝑀1𝑀 2+ 2 → 2 𝑀 𝑀 + 1 2 (9) 𝑛𝐹∆𝐸1/2 𝐾 = 𝑒 𝑅𝑇 (10) 𝐶 −∆𝐺𝑐 = 𝑛𝐹∆𝐸1/2 (11) The free energy change results from five contributions to the term. The statistical contribution ∆𝐺stat, which accounts for 36 mV and arises from the fact, that the mixed- valent system is statistically favored in comparison to isovalent compounds. The inductive contribution ∆𝐺ind, which arises from the fact, that a change of the oxidation state of one metal center affects the bonding ability of the bridging ligand and therefore alters the redox potential of the second metal center. A magnetic exchange contribution ∆𝐺ex, which is essentially proportional to the antiferromagnetic exchange term. The electrostatic contribution ∆𝐺el, which expresses that adding a second charge to an already charged molecule requires more energy. And last there is the resonance contribution ∆𝐺res, due to the resonance stabilization of the mixed-valent compounds. While it is obvious, that some contributions to the free energy change, and therefore to the half-wave potential splitting, are rather small and constant, others may contribute significantly for a certain system so that a great half-wave potential splitting may serve as an indicator for electronic coupling in mixed-valent systems, but cannot be used to quantify the electronic coupling by any means.[86] ∆𝐺𝑐 = ∆𝐺𝑠𝑡𝑎𝑡 + ∆𝐺𝑖𝑛𝑑 + ∆𝐺𝑒𝑥 + ∆𝐺𝑒𝑙 + ∆𝐺𝑟𝑒𝑠 (12) 12 Mixed-Valence and Electron Transfer 1.1.6 Electron Transfer in Ferrocene Compounds Ferrocene compounds have been used since the 1970s to study mixed-valent behavior and electron transfer.[42,81] These studies have led to interesting results for the value of the electronic coupling depending on the type, length and electron richness of the bridge. While it would be desirable to study any mentioned effect on its own, the nature of chemical bonds prohibits strict separation between above mentioned effects. However, an attempt can be made derive as much information about single effect as possible. Comparing biferrocene II, diferrocenylmethane III, 1,2-diferrocenylethene IV, and diferrocenylacetylene V, it is evident that the type of the bridge with respect to its hybridization has a large impact, see Figure 1. While biferrocenium II shows a strong IVCT band with 𝜆max = 1800 nm, differocenylmethane III displays no IVCT band and therefore no electronic communication between the ferrocene units. Compound IV has a slightly smaller electronic coupling, derived from an IVCT band at 𝜆max = 1750 nm, compared to Figure 8. Singly oxidized biferrocene1,[43] II, diferrocenylmethane2,[43] III, 1,2-diferrocenylethene[44] IV, differocenylacetylene1,[43] V and their HAB in CH2Cl2. biferrocene II. Finally, diferrocenylacetylene V exhibits the strongest electronic coupling with an IVCT band at 𝜆max = 1560 nm. These results are of rather qualitative nature, because biferrocene II has a significantly smaller distance between the metal centers than III, IV and V and while the distance in III, IV and V is rather similar, the bridge of V is more electron rich than that of IV. Therefore, the conclusion is that permitting interaction 1 Calculated from given values for the interaction parameter α. 2 No solvent given in reference 13 Introduction between the cyclopentadienyl ligands of the two redox active centers via the orbitals of the bridge leads to strong electronic coupling. Figure 9. Singly oxidized diferrocenylpolyenes and their HAB under the assumption of trans-configuration VIn and under the assumption of cis- configuration VIIn in CH2Cl [44]2. Launay and Spangler conducted intensive research on differrocenylpolyenes with respect to distance dependency, see Figure 9.[44] The mixed-valent compounds vary in their polyene chain length from n=1–6 and are, concerning the polyene chain, considered to be trans only (all E isomers). This leaves two cases VIn and VIIn, wherein in VIn the ferrocene moieties are considered to be anti in respect to the polyene chain and in VIIn the Ferrocene moieties are considered to be syn. This affects the inter-metal distance for each compound and therefore the calculated value of HAB as shown in Equation (7). The differentiation between both cases is merely a hypothetical one, because de facto no systematic crystallographic results are available, most likely sterical influences are rendered negligible with increasing chain length and the compound is most likely flexible in solution, so an averaged UV/vis spectrum and therefor HAB is measured. However, the data clearly shows decreasing electronic coupling with increasing distance. In fact at least for the values calculated for VIn the electronic coupling decays exponentially with distance. The slope of this decay corresponds to a division in half of the value each 8 Å. 14 Mixed-Valence and Electron Transfer Figure 10. Singly oxidized 2,5-diferrocenyl heterocycles in CH 3[87]2Cl2. Electron rich and poor bridges were studied by Lang et al. with heterocyclic five-membered ring systems as a bridge connecting two ferrocenyl substituents, see Figure 10. Changing the bridge form thiophene VIII1 over furan VIII2, N-methylpyrrole VIII3 to N- phenylpyrrole VIII4, so varying the bridge from electron poor to electron rich heterocycles yields higher electronic coupling HAB. The same is true for changing the substituents on the phenyl ring of N-phenylpyrrole derivatives IXn. Starting from 4-dimethylaminophenyl substituted pyrrole IX1 to 4-ethylcarboxylphenyl substituted pyrrole IX5 HAB decreases with increasing electron deficiency of the bridge. In conclusion, generally higher electronic couplings are achieved with electron rich bridges, rather than with electron poor bridges. Concerning asymmetrically bridged ferrocene derivatives less data is available. Molina and Veciana et al. studied 1,4-bis(ferrocenyl)-2-aza-1,3-butadiene X, see Figure 11. Compared to its all-carbon analogue VI2 HAB is just half as large. It should be noted, that the comparison is not fair, due to the fact, that with asymmetric bridging the redox potentials of the ferrocenes are nomore equivalent. Heinze et al. did an intensive study of amide bridged ferrocenes diminishing the effect of different oxidation potentials by the linkage of two redox smiliar ferrocene units in derivative XII. Comparing the derivative 3 HAB recalculated for RAB = 7.084 Å, derived from single crystal X-Ray diffraction of VIII1,[45] from given HAB values for RAB = 2.1 Å (experimental effective electron transfer distance) for ease of comparison. 15 Introduction XII with its analogue with different redox potentials XI shows an almost 1.5-fold increase in HAB. The values given for this compound are derived from oxidation in THF rather than CH2Cl2 and can therefore not directly be compared to the other values given in this section, but it should be noted that the HAB value for compound XI increases to 200 cm −1 in CH2Cl2 solution, see Results and Discussion, stating that asymmetrical bridges can lead to a reasonably strong coupling, but with the advantage of facilitating a directional electron transfer, in contrast to symmetrical bridges.[88,89] Figure 11. Different asymmetrically bridged ferrocene derivatives (X,[46], XI[84], XII[47]) and their HAB in THF. 1.2 Ferrocene Containing Oligopeptides Oligopeptides containing ferrocene units in the backbone can be derived from three different ferrocene derivatives: 1,n’-diaminoferrocene XIII, 1,n’-ferrocenedicarboxylic acid XIV, and 1-amino-ferrocene-n’-carbocylic acid XV (ferrocene amino acid; Fca), shown in Figure 12. While oligopeptides containing 1,n’-diaminoferrocene XIII and 1,n’- ferrocenedicarboxylic acid XIV feature a parallel arrangement of peptide strands, ferrocene amino acid XV yield an antiparallel arrangement as common in natural peptides.[90] Due to conformational freedom of ferrocene derivatives — along the pseudo-C5 axis described by the angle between the two substituents , the angle β between the cylopentadienyl ring plane to substituent bond as well as the tilt angle between the two cyclopentadienyl rings  — stereochemical descriptors have been established, see Figure 13. 16 Ferrocene Containing Oligopeptides Figure 12. 1,n’-diaminoferrocene XIII, 1,n’-ferrocenedicarbocylic acid XIV, 1-amino-ferrocene-n’-caboxylic acid XV and their peptide strand arrangements. Rotational freedom along the pseudo-C5 axis distinguishes five isomers. The 1,1’-isomer with an angle between the two substituents of −36° <  < 36°, the 1,2’ isomer with 36° <  < 108°, the 1,3’ isomer with 108° <  < 180°, the 1,4’ isomer with −108° <  < −180°, and the 1,5’ isomer with −36° <  < −108°. The 1,2’ isomer and the 1,3’ isomer introduce P-helical chirality, while the 1,4’ isomer and 1,5’ isomer introduce M-helical chirality. The orientation of the Cp amide bonds is given with an E/Z nomenclature, where Z isomer denotes that the substituents at the amide with highest priority are facing towards the substituent at the other Cp ring. If the substituent with highest priority is facing the other direction the amide bond is considered the E isomer. Figure 13. Stereochemical descriptors of ferrocene containing peptides.[90] All three ferrocene derivatives mentioned above permit interstrand hydrogen bonding. In peptide chemistry the hydrogen bonding pattern defines the nature of the turn structure. In Figure 14 “normal” and “reverse” turn structures are shown. The turn is called “normal” if the H bond acceptor is located at the N teminus of the peptide, while it is called “reverse” if the hydrogen bond acceptor is located at the C terminus. In case of a normal turn α- helices are featured by 13-membered rings, while β- and -turns are featured by ten- and 17 Introduction seven-membered rings. For the “reverse” turn an eleven-membered ring features a reverse α-helix, an eight-membered ring a reverse β-turn and a five-membered ring a reverse - turn. Figure 14. “Normal” and “reverse” turns in peptide chemistry. The hydrogen bonding motifs of ferrocene peptides feature similar turns, so the same nomenclature can be used. Additionally some structural motifs are named as shown in Figure 15. The “Herrick” conformation[91] features two ten-membered rings and therefore induces a β-turn. The “van Staveren” conformation[92] features a seven-membered ring and therefore a -turn. Last the open conformation is named “Xu”.[93] Figure 15. Different H-bonding motifs for 1,n’-ferrcocendicarboxylic acid. Intensive research concerning the properties of amino acid conjugates of ferrocene dicarboxylic acid XIV was done by several working groups. The general findings are that the ferrocenyl (Fn) bioconjugates show a ferrocene based positive cotton effect in CD 18 Ferrocene Containing Oligopeptides spectroscopy.[94] Bioconjugates of L-amino acids show P-helicity, while bioconjugates derived from D-amino acids show M-helicity.[90] Various amino acid conjugates are known throughout the literature and their structure in the solid state and solution has been disclosed.[93,95–99] One of the most interesting findings for this kind of bioconjugates was a tryptophane containing Fn[CO-L-Trp-OMe]2, which shows self-assembly to a supramolecular nanofibrillar network. Stimuli responsive (oxidation/reduction, temperature) morphological transformations for this compound were observed, rendering it a good starting point for the development of new redox-active functional biomaterials.[100] The first bioconjugates derived from 1,1’-diaminoferrocene XIII were reported by Kraatz et al.[101] A [Boc-Ala-NH]2Fn, which shows interstrand hydrogen bonding in two ten-mem-bered rings in the solid state and preserved H bonding in solution. The work was carried on by Heinze and Rapic[102] with asymmetric conjugates of the type Boc-AA-NH- Fn-NH-Ac [AA=Gly, Ala, D-Ala, Val). It was found, that a single covalently bonded amino acid is enough to induce chiral organization of the ferrocene central unit and the majority possess P-helical conformation for L-amino acids. Bioconjugates of ferrocene amino acid XV (Fca) enable even more diverse H-bonding motifs compared to 1,1’-diaminoferrocene XIII and ferrocene dicarboxylic acid XIV.[103] Amongst others they could be realized in form of X-CO-Fca-AA-OMe (X=Me or Boc), with only a single interstrand hydrogen bond due to steric effects. For example, the Me- CO-Fca-Val-OMe derivative exhibits one NHFca•••OCAA hydrogen bond forming a nine- membered ring.[104] X-CO-AA1-Fca-AA2-OMe derivatives were explored, with two hydrogen bonds forming a 9-membered and eleven-membered ring namely NHFca•••OCAA2 and NHAA2•••COAA1 for the Me-CO-Ala 1-Fca-Ala2-OMe case. [104,105,105,106] These rings can even be selectively enlarged by the use of β-amino acids.[107] Figure 16. Zig-zag-Conformation of olgoamides of Fca, exemplary for Fmoc-Fca5-OMe XVI. 19 Introduction Figure 17 DFT optimized geometries of [Ac-Fca5-OMe]n+ XVI (n = 1-5); spin density at 0.002 a. u.[48]4 Derivatives with more than one ferrocenyl moiety in the peptide backbone have been studied since 1998. This began with experiments from Nakamura et al.[108] with Me-CO- Fca2-NHMe and was later continued by Heinze et al. [109] with Me-CO-Fca-NH-Fc. For both compounds only intermolecular hydrogen bonds are observed in the solid state. Thorough investigation of compounds X-CO-Fcan-NH-Fc (n = 1,2, X= Me, OtBu, OCH2-fluorenyl) revealed hydrogen bonding in solution. All oligopeptides form an eight-membered ring with a 1,2’ (1,5’) conformation derived from nuclear Overhauser spectroscopy.[84] This is conform to DFT calculations of the system showing that one conformer is thermodynamically favored over the others. DFT calculations also suggest, that the hydrogen bonding motif is conserved during oxidation until every single ferrocenyl moiety is oxidized. Then the open conformation becomes the energetically favored one. Experimental evidence for this is currently only obtainable by IR spectroscopy and the absence of hydrogen bonded amide groups, because the strong paramagnetic nature of the fully oxidized compound renders NMR spectroscopy useless. This very stable hydrogen bonding motif is also called zig-zag-conformation and leads to a nearly parallel alignment of the individual amide dipole moments, forming a permanent macrodipole, see Figure 16. 4 Adapted from: D. Siebler, Dissertation, Johannes Gutenberg-University, 2010. 20 Ferrocene Containing Oligopeptides DFT calculations for this system suggest mixed-valent states, in which alternating FeII and FeIII units are present, see Figure 17. The same conformation is seen in oligoamides of the kind Fmoc-Fcan-OMe (n=2-5). [47] 1.2.1 A Structure Comparison of Amides and Thioamides In contrast to amides, which have been extensively studied because of their relation to peptide conformations, thioamides have been essentially neglected. Just as the amide group the thioamide shows a nearly planar arrangement. The bond angles for both groups are essentially identical, as seen in Figure 18. The largest structural difference in structure is the significantly longer C=S bond length of 1.65 Å, exemplarily for N-methyl thioacetamide, compared to the C=O bond length of 1.23 Å of its amide analogue.[110] Another difference is the smaller electronegativity of the sulfur atom ( = 2.58)[111] compared to oxygen ( = 3.44)[111], which renders the C=S bond (Δ = 0.03) significantly less polar compared to the C=O bond (Δ = 0.89). Figure 18. DFT optimized geometries of N-methyl acetamide and N-methyl thioacetamide with their respective bond lengths in Å and bond angles in deg (italics). In both amides and thioamides an E and Z configuration is feasible, see Figure 19. In general the rotational barrier around the N-C bond is lower for amides compared to their analogue thioamides.[112] The origin of the higher rotational barrier for thioamides is the larger stabilization of the ground state. The above mentioned difference in electronegativity for oxygen and sulfur leads to very different polarities for the C=X (X = O, S) bond. The C=O bond is very polar therefore the  orbital of the double bond has a higher coefficient at the oxygen atom and a smaller one at the carbon atom. For the * orbital the situation is reversed with larger coefficients at the carbon atom. The electron density in the C=S bond 21 Introduction is evenly distributed between the carbon and sulfur atom and therefore the orbital coefficients of the  and * orbital are evenly distributed as well. For both amides and thioamides the nitrogen lone pair donates electrons into the * orbital of the C=X (X = O, S) bond, see Figure 20. In amides little stabilization is gained through electron donation from the nitrogen atom to the carbon atom, because the oxygen atom already has a high electron density. In case of thioamides the stabilization is much larger, because the charge is not only donated from the nitrogen to carbon, but is passed on to the sulfur atom, which can easily accommodate higher electron densities through its lager orbitals.[113] Figure 19. Z and E conformation for amides and thioamides. For the transition states TS1 and TS2, see Figure 20, Lauvergnat and Hiberty showed via DFT calculations, that a sp3 hybridization at the nitrogen atom is energetically favored compared to the sp2 hybridization of the ground state for formamide and thioformamide.[114] This is applicable to all amides and thioamides. During the rotation from the ground state to the transition state, the orbital overlap between the nitrogen lone pair and the *-orbital decreases essentially to zero in the transition state, due to the fact, that the orbitals are orthogonal in the TS. Comparing the ground state and the TS 1 for the amide only small change in C=O bond length is visible, while for the thioamides a significant shortening of the C=S bond length occurs. C-N bond is elongated in the transition state for both amides and thioamides, due to the cleavage of the partial double bond. This is conform to the model of stabilization of the thioamides via electron donation to the sulfur atom, and for amide to a donation to the carbon atom only. 22 Ferrocene Containing Oligopeptides Figure 20. Schematic representation of the orbitals in amides (X=O) and thioamides (X=S) in the ground state (GS) and both transition states (TS), and calculated bond length in Å. Because of the charge (𝑞) transfer from the nitrogen atom to the sulfur atom and the longer distance (𝑙) thioamides display a larger dipole moment (?⃗?) in the ground state than amides ( ?⃗? = 𝑞 ∙ 𝑙).[115] This was shown for dimethyl thioacetamide and dimethyl acetamide in their ground state (5.38 vs. 4.37 D), whereas the transition state exhibits essentially the same dipole moment for both (2.45 vs. 2.33 D).[113] This has great impact for the solvent dependency of the rotational barrier for amides and thioamides. The rotational barrier for both increases with the polarity of the solvent, due to the energy needed for reorganization of the solvent molecules. The greater change in dipole moment during rotation for the thioamides compared to amides, means that the effect of the solvent on the rotational barrier is stronger for thioamides as shown in Table 2. Galabov et al. conducted DFT calculations for the effect of electron donating and withdrawing substituents R2, see Figure 19. Electron withdrawing substituents lower the rotational barrier significantly, while electron donating substituents lead to higher rotational barriers.[116] In case of hydrogen bonding, no studies for thioamides have been done yet, but data for amides show, that a hydrogen bond at the NH group with anionic acceptors destabilizes the E isomer, while the effect of hydrogen atom donors is negligible. The study also shows, that the effect of hydrogen bonding is not reliably predictable, because chelating effects can occur and reverse the general effect.[117] The effect of substitution patterns of secondary thioamides has been thoroughly studied by Rao et al. for N-monosubstituted thioamides.[118] In general, the E/Z ratio for thioamides is determined by steric effects and therefore the larger sulfur atom compared to oxygen leads to a greater proportion of the E isomer. The E/Z ratio also increases with the size of 23 Introduction the alkyl substituent R2, see Figure 19. Increasing the size of R1 has the contrary effect, and leads to an increase of the Z-isomer. Medium  Dimethylthioacetamide Dimethylacetamide gas phase 1.00 74.1 64.1 cyclohexane 1.93 82.0 68.5 dichlormethane 9.0 91.6 75.1 acetonitrile 32.7 92.9 74.3 water 61.0 97.9 79.7 Table 2. Experimental rotational barriers of for N,N’- dimethylthioacetamide and N,N’-dimethylacetamide at 80°C.[113,119] 1.2.2 Thioamides as Isosteric Replacement Thioamides are generally considered to be isosteric replacements for amides in amino acid conjugates (e.g. peptides), but literature shows that this statement is not entirely true. Because of the previously mentioned longer C=S distance compared to the C=O distance in amides the hydrogen bonding of thioamides might introduce changes in protein folding.[120] The higher acidity of the thioamide NH proton makes it a better proton donor for hydrogen bonding compared to amides (pKa = 18.5 vs. 25.5). The larger sulfur atom on the other hand is a weaker hydrogen bond acceptor than oxygen, leading to weaker hydrogen bonds. Hence the SCN-H•••O hydrogen bond is stronger than the OCN-H•••O hydrogen bond, while the N-H•••SC hydrogen bond is weaker than the N-H•••OC hydrogen bond. [69] This was tested by Miwa et al. by preparation of a thioamide analogue of GCN4, which is a helical peptide consisting of 35 amino acids. Substitution of one amide by a thioamide at the C terminus and in the middle of the strand did not lead to any significant change in structure, as probed by circular dichroism measurements.[121] Kiefhaber et al. found contrary results. Upon substitution of an amide unit by a thioamide unit in Alanine based helical peptides at the N-terminus or in the center a highly helix destabilizing effect was observed. Studies of the incorporation of thioamide units into β-turn hairpin structures lead to the conclusion, that the turn structure is essentially unaffected if the sulfur of the thioamide group is pointing toward the exterior and hydrogen bonding of the NH group is 24 Ferrocene Containing Oligopeptides feasible. But if the thiocarbonyl sulfur is forced to participate in hydrogen bonding a perturbation of the turn structure occurs. This effect is diminished if the thioamide substitution is positioned at the terminal region. In conclusion, the effect of thioamides as isosteric replacement in peptides is highly depending on the position of incorporation, particularly if the thioamide unit participates in hydrogen bonding interactions.[122] The weak hydrogen bonding of the thiocarbonyl sulfur can also be put to good use. Beeson et al. probed the hydrogen bonding interaction of a ligand to a major histocompatibility complex class II, an antigen presenting peptide, in the main chain. Substitution of the main chain carbonyl function involved in the hydrogen bonding with sulfur lead to a 30-fold increase of ligand dissociation, underlining the crucial role of hydrogen bonding interactions.[123] Figure 21. Natural products containing thioamide units. 25 Introduction Only a single naturally occurring protein containing a thioamide group is known today. In the methyl-coenzyme M reductase XVII, see Figure 21, a thioglycin residue is incorporated close to the active site.[65,66,124] Here the thio amino acid is not engaged in hydrogen bonding forming secondary structures like α-helices or β-sheets. It is proposed that the thioamide unit acts as redox mediator in the methane-forming step and maybe undergoes a redox induced E/Z conformational facilitating the reaction. Thioamides as well as carboxamides undergo an E/Z-conformational change under irradiation. For amides the →* electronic transition is located a 200 nm and leads to photodecomposition. This decomposition is remarkably reduced in thioamides, where the →* transition is red-shifted to 270 nm. This was put to use in a semisynthetic ribonuclease S containing a thioamide unit. Upon irradiation 30% of the thioamide switched from the Z- to the E-conformer rendering the enzyme inactive. More widespread application of this method could disclose important structural influences on enzyme reactivity.[125] Thioamides compared to amides exhibit greater stability against proteolytic degradation and improved absorption, distribution, metabolism and excretion properties.[69–71] That biological systems are unfamiliar to thioamides is also represented by the fact, that besides the apo-methanobactin XVII in the methyl-coenzyme M reductase only four other natural products containing thioamide groups are known. The oxothioacetamide derivative[67] XVIII, which is believed to be a decomposition product of more complex molecules, cycasthioamide[68] XIX, which is a non proteinogenic amino acid isolated from plants, thioviridiamide[64] XX, an apoptosis inducer of bacterial origin, and closthioamide[63,126] XXI. The latter shows remarkable antibiotic activity and was isolated from the anaerobic bacterium Clostridium cellulolyticum as a secondary metabolite, see Figure 21. It was shown, that the thioamide groups are of fundamental importance in this compound, because tests with the all carboxamide compound closamide showed no significant antibacterial activity. 1.2.3 Thioamide Containing Ferrocene Derivatives All thioamide containing ferrocene derivatives known before this thesis are formally derived from ferrocene carboxylic acid. The first thioamide bearing ferrocenyl derivative was N,N-dimethyl ferrocene carbothioamide XXII and was reported by Nonoyama et 26 Ferrocene Containing Oligopeptides al.[127], see Figure 22. This derivative was used to form cyclopalladated compounds. This work was later expanded using N,N,N’,N’-tetramethylferrocene-1,1’-dicarbothioamide XXIII.[128] Figure 22. Various ferrocenyl thioamide containing ligand system and metal complexes derived from them. The essence of ferrocenyl substituted thioamides as a ligand was later picked up by Štěpnička et al.[129] who prepared 1’-(dipehylphosphino)-1-[(dimethylamino)-thio- carbonyl]-ferrocene XXIV and it was found that this compound can act as a very flexible ligand, which allows perfect preorganization via rotation of the ferrocenyl moiety and twisting of the thioamide unit around the C-C bond. The redox behavior of compound XXIV is much more complex than that of the oxo analogue, showing only irreversible oxidations, which is probably attributed to the fact that the HOMO of the thioamide compound encompasses the ferrocenyl and thioamide moiety. The coordination properties of compound XXIV was probed by reaction with copper(I) and silver(I) resulting in complexes of the type [Cu(κP,κS-XXIV)2][BF4] XXV and the dimeric complex [Ag2(ClO4)( κP,κS-µS-XXIV)2] XXVI. Zakrzewski et al. used ferrocenecarbothioamide XXVIII and its N-ethoxycarbonyl derivative XXIX as precursors for the synthesis of 2,4-diferrocenylthiazole XXX.[130] A proof of concept for the synthesis of ferrocene containing heterocycles, which are of interest as redox active markers[131,132] and for their potential biological activity.[57,59,60,133] This work was continued by the use of XXXI as a precurser for “click”-chemistry compounds containing a ferrocenyl moiety, see Scheme 1. 27 Introduction Scheme 1. Synthesis of ferrocenyl substituted heterocycles and “click” precursors. The use of thioamides as anion receptors was disclosed by Beer et al.. They synthesized ((butylamino)-thiocarbonyl)ferrocene from the oxo analogue by reaction with Lawessons’ reagent. The comparison of the two ferrocene derivatives showed, that the thioamide binds halogen anions more effectively than the carboxamide, because of it’s higher acidity and therefore better hydrogen bonding ability, providing a better NMR antenna for anion detection[134]. Scheme 2. Condensation of (dimethylaminomethyl)ferrocene XXXII and aminomethylferrocene XXXIII with aliphatic and aromatic amines. A big step for the widespread application of ferrocenyl thioamides XXXIV was done by Gasser et al.[135], see Scheme 2. A new synthetic route was developed starting from (dimethylaminomethyl)ferrocene XXXII or aminomethylferrocene XXXIII, elemental sulfur and primary alky- or arylamines or in case of morpholine a secondary amine with 28 EPR Spectroscopy and Spin Trapping of Radicals good tolerance of functional groups (Ether-, Ester-, Nitro-, Amino- and hydroxylgroups as well as halogens). This makes it one of the most straightforward synthetic routes to C- ferrocenyl thioamides. 1.3 EPR Spectroscopy and Spin Trapping of Radicals EPR spectroscopy is the analytical method of choice for detection of unpaired electrons. When applying an external electromagnetic field to a paramagnetic compound, the Zeeman effect occurs, just like in NMR spectroscopy. The parallel and antiparallel orientations of the spin are separated in energy and transition between the two states is possible by absorption of irradiation. The shift of this absorption (g-value) depends mainly on the spin- orbit interaction, which itself generally depends on the mass of the nucleus, and gives a rough estimation about the location of the spin (i.e. different elements can be distinguished). The main source of information about the environment of the spin is derived from hyperfine coupling interactions. The interactions of the electron spin with nuclear spins, leads to a distinct coupling pattern, which can allow determination of the spin location, see Figure 23. Due to the high reactivity of some radicals, especially organic radicals, and the mostly applied slow sweep technique to measure EPR spectra a detection of these intermediate species is impossible. For detection of short-lived species spin trapping is a valuable tool. Figure 23. a) Schematic representation of the Zeeman effect b) Hyperfine coupling pattern. 29 Introduction Spin trapping means that the radical species of interest undergoes an addition reaction to a diamagnetic compound (spin trap) forming a relatively long–lived spin adduct.[136] Despite the fact, that over 100 different compounds have been found to be suitable spin trapping agents, the most widespread used spin trapping agents are nitrones and nitroso compounds.[137] Both form stable nitroxide radicals, however, the nitroso spin traps can provide more information, because the radical adds directly to the nitroso nitrogen and is therefore more likely to engage in hyperfine coupling interactions with the trapped spin, see Scheme 3. Scheme 3. Spin trapping reaction of nitroso and nitrone compounds. Figure 24. Most commonly used spin trapping agents. N-tert-butyl nitrone (PBN) XXXV, 5,5-dimethyl-1-pyrrolidine-N-oxide (DMPO) XXXVI, nitrosobenze (NOB) XXXVII, nitrosodurene (ND) XXXVIII and 2-methyl-2- nitrosopropane (MNP) XXXIX. The most commonly used spin trapping agents are N-tert-butyl nitrone (PBN) XXXV, 5,5- dimethyl-1-pyrrolidine-N-oxide (DMPO) XXXVI, nitrosobenzene (NOB) XXXVII, nitrosodurene (ND) XXXVIII and 2-methyl-2-nitrosopropane (MNP) XXXIX, see Figure 24. The spin trapping of organic radicals has been intensively studied and extended tables with parameters for different spin adducts exist.[136,138,139] Even inorganic radicals can be trapped for some fragments that are isolobal to organic radicals and are expected to behave accordingly, trapping a metal centered radical in a spin adduct.[137,140–143] EPR spectra 30 Ferrocene-Containing (Pro-)Drugs derived from a purely organic radical and a manganese centered radical trapped with nitrosodurene XXXVIII are shown in Figure 25. Figure 25. Examples of a) an organic[139]5 and b) a manganese[140]6 centered radical trapped with nitrosodurene. 1.4 Ferrocene-Containing (Pro-)Drugs The exploration of ferrocene containing drugs started as early as 1978, when Brynes and coworkers found, that ferrocenyl polyamines exhibit a low but significant antitumor activity against lymphocytic leukemia P-388.[144] Since then the medicinal applications for ferrocenyl derivatives have become a widespread research field.[145] In 2000 a fluconazole XL analogue bearing a ferrocenyl moiety XLI with antifungal activity has been reported by Biot et al., see Figure 26. It showed remarkable growth inhibition for some strains of yeasts of the genus Candida. Antibacterial activity was reported for ferrocenyl-thiazoleacylhydrazones XLII against Staphylococcus aureus, Esherichia coli and Pseudomonas aeruginosa.[146] Even studies for ferrocenyl containing vasorelaxants for the treatment of hypertension showed promising results.[147] The 5 EPR spectrum adapted from L. Omelka, J. Kováčová, Magn. Reson. Chem. 1994, 32, 525– 531 with permission of John Wiley and Sons. Copyright © 1994 John Wiley and Sons. 6 EPR spectrum adapted from A. Hudson, M. F. Lappert, P. W. Lednor und B. K. Nicholson, J. Chem. Soc., Chem. Commun., 1974, 966 with permission of the Royal Society of Chemistry. Copyright © 1969, Royal Society of Chemistry 31 Introduction ferrocenyl-pyrido[2,3-d]pyrimidines XLIII studied were even more effective than the control drug rolipram XLIV. Recently a Ferrocenylquinoline XLV was tested as potential agent against leishmanial disease, which is a major problem in developing countries.[61] Figure 26. Fluconazole XL, ferrocenyl-fluconazole analogue XLI, ferrocenyl- thiazoleacylhydrazones XLII, ferrocenyl-pyrido[2,3-d]pyrimidines XLIII, rolipram XLIV and ferrocenylquinoline XLV. One of the most exciting research centers around ferroquine XLVII, a structural analogue of chloroquine XLVI and potent antimalarial drug, see Figure 27. The research on incorporation of ferrocenyl moieties into antimalarial drugs began in the mid 1990s[148], but it took until 2006 for the development of ferroquine XLVII. This compound shows greater antimalarial activity than its parent analogue chloroquine XLVI and even targets chloroquine resistant malarial strains.[149] It was shown that the intramolecular hydrogen bond of ferroquine XLII, which is absent in chloroquine XLVI has a significant effect on the biological activity, most probably due to the fact that it helps to maintain the conformation of the molecule and allows better interaction with the receptor.[150] The mechanism of action of ferroquine XLII is proposed to be a dual one. On the one hand it targets the hemozoin formation, a crystalline form of β-hematin, in which the cytotoxic α- hematin is converted and therefore rendered harmless.[151] On the other hand it is able to form reactive oxygen species (ROS) upon oxidation. This causes lipid peroxidation, 32 Ferrocene-Containing (Pro-)Drugs eventually leading to death of the malarial parasites.[152] Most recently ferroquine XLII has entered phase 2 of clinical research and shows promising results as a standalone drug and even in combination with other antimalarial drugs.[153] Figure 27. Chloroquine XLVI and ferroquine XLVII. The generation of ROS was also proposed as the mechanism of action for the antitumor activity of ferrocenium derivatives as was disclosed more than 20 years after the first report of anticancer activity for this derivatives.[154] This thesis was proven to be highly probable by studies, where the oxidative damage to DNA in MCF-7, a type of breast cancer cells, was monitored, under the addition of different ferrocenium compounds. It was also possible to spin trap reactive oxygen species via the use of DEPMPO a derivative of DMPO bearing a diethylester phosphonic acid substituent (vide supra).[155] Figure 28. Antitumor agents. The most promising ferrocene derivative in anticancer studies is ferrocifen XLIX, a derivative of tamoxifen XLVIII, which is widely used to treat ER-responsive breast cancer, see Figure 29.[156] Tamoxifen XLVIII is a Selective Estrogen Receptor Modulator and therefore blocks this receptor. Without the estrogen stimulus the proliferation of the ER- 33 Introduction responsive breast cancer cells is inhibited. One major drawback of tamoxifen XLVIII is, that it is only effective against ER+ types of cancer cells, which excludes about 40% of breast cancer tumors from treatment. Ferrocifen XLIX on the other hand shows powerful anti-proliferative effects on both ER+ and ER− cancer cell lines. Ferrocifen XLIX and a second derivative ferrociphenol L show even cytotoxic effects against various types of melanoma cells.[157] These compounds even showed higher activity than cis-platin LI for mesotheolioma cells.[56] The research of ferrocifen has recently become more widespread. New derivatives bearing two aminoalkyl chains [158] have been found to show strong antiproliferative effects on breast cancer cells. Other ferrocifen derivatives have found to induce senescence in cancer cells, an irreversible arrest in the cell-cycle prohibiting proliferation.[159] Remarkably, it was found that the anticancer activity of ferrocifen is not attributed to the generation of reactive oxygen species (ROS), as reported for ferrocenium derivates.[157,160] The mode of action for this compounds is still yet to be disclosed. Mokhir et al. synthesized a series of amino- and diaminoferrocene based prodrugs for the treatment of leukemia, see Figure 29. The Concept of the prodrugs LII, LIII and LIV is based on the cleavage of the B-C bond under oxidizing conditions, as found in cancer cells. A phenolate LVI is formed, which releases a p-quinone methide LVII , which can affect cancer cells, as well as an unstable ferrocenyl derivative LVIII, which quickly undergoes decarboxylation to the aminoferrocene derivative LIX. The aminoferrocene is either oxidized under the predominant condition and acts as ROS generator or decomposes by nucleophilic attack to Fe2+/Fe3+ ions, which can than generate ROS species leading to cell death, see Scheme 4.[161] It was found, that for the most compounds the cytotoxicity correlates with the release of iron ions, except for the most active compound LIIe, which only generates ferrocenium ions. Therefore it was concluded, that ferrocenium ions are more cytotoxic, than mere iron ions, probably due to the retarded metabolism of such compounds. The diaminoferrocene based prodrugs LIII exhibited a slightly greater antitumor effect than the parent aminoferrocene compound LII. This is most likely attributed to the formation of a second p-quinone methide LVII rather than a direct effect of the substitution pattern of ferrocene. [162] 34 Ferrocene-Containing (Pro-)Drugs Figure 29. Amino- and Diamonoferrocene based Prodrugs. A variation of the substituent at the amide nitrogen led to dramatic changes in activity. While compounds LIVa – LIVd exhibit a rather similar activity, LIVe and LIVf show decreased activity for at least one cell line type.[163] This sort of compound does not only show high activity, but also high selectivity for cancer cells, as it was found that mononuclear cells are almost unaffected when treated with the prodrugs.[161] Finally mice, which carry Leukemia, treated with LIIe, had a prolonged life span, even if the applied doses, were much lower than the determined safe maximum dose of the compounds.[163] From the facts reported above it is apparent, that ferrocenyl derivatives are valuable compounds for pharmaceutical applications. The potential of ferrocene as biologically active moiety has just started to be explored and might lead to novel drugs and prodrugs in the near future. 35 Introduction Scheme 4. Reaction mode for aminoferrocene based prodrugs LII. 1.5 References [1] Morgan, J. W.; Anders, E. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 6973–6977. [2] Metal stocks in society; T. E. Graedel, Nairobi 2010. [3] Perutz, M. F.; Rossmann, M. G.; Cullis, A. F.; Muirhead, H.; Will, G.; North, A. C. T. Nature 1960, 185, 416–422. [4] Kendrew, J. C.; Dickerson, R. E.; Strandberg, B. E.; Hart, R. G.; Davies, D. R.; Phillips, D. C.; Shore, V. C. Nature 1960, 185, 422–427. [5] Haines, D. C.; Tomchick, D. R.; Machius, M.; Peterson, J. A. Biochemistry 2001, 40, 13456–13465. [6] Li, H.; Poulos, T. L. Nat. Struct. Biol. 1997, 4, 140–146. [7] Li, H.; Poulos, T. L. Curr. Top. Med. Chem. 2004, 4, 1789–1802. [8] Poulos, T. L.; Finzel, B. C.; Howard, A. J. J. Mol. Biol. 1987, 195, 687–700. [9] Rowland, P.; Blaney, F. E.; Smyth, M. G.; Jones, Jo J.; Leydon, Vaughan R.; Oxbrow, Amanda K.; Lewis, Ceri J.; Tennant, Mike G.; Modi, Sandeep; 36 References Eggleston, Drake S.; Chenery, Richard J.; Bridges, Angela M. J. Biol. Chem. 2006, 281, 7614–7622. [10] Schlichting, I.; Berendzen, J.; Chu, K.; Stock, A. M.; Maves, S. A.; Benson, D. E.; Sweet, R. M.; Ringe, D.; Petsko, G. A.; Sligar, S. G. Science 2000, 287, 1615– 1622. [11] Schoch, G. A.; Yano, J. K.; Wester, M. R.; Griffin, Keith J.; Stout, C. David; Johnson, Eric F. J. Biol. Chem. 2004, 279, 9497–9503. [12] Scott, E. E.; White, M. A.; He, Y. A.; Johnson, Eric F.; Stout, C. David; Halpert, James R. J. Biol. Chem. 2004, 279, 27294–27301. [13] Venkateswara Rao, P.; Holm, R. H. Chem. Rev. 2004, 104, 527–559. [14] Beinert, H.; Holm, R. H.; Munck, E. Science 1997, 277, 653–659. [15] Lawson, D. M.; Artymiuk, P. J.; Yewdall, S. J.; Smith, J. M.; Livingstone, J. C.; Treffry, A.; Luzzago, A.; Levi, S.; Arosio, P.; Cesareni, G. Nature 1991, 349, 541–544. [16] Quintana, C.; Bonnet, N.; Jeantet, A. Y.; Chemelle, P. Biology of the Cell 1987, 59, 247–254. [17] Schrock, R. R. Acc. Chem. Res. 2005, 38, 955–962. [18] Schrock, R. R. Angew. Chem. 2008, 120, 5594–5605. [19] Schrock, R. R. Nat. Chem. 2011, 3, 95–96. [20] Weare, W. W.; Dai, X.; Byrnes, M. J.; Chin, Jia Min; Schrock, Richard R.; Muller, Peter Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 17099–17106. [21] Kliegman, S.; McNeill, K. Dalton. Trans. 2008, 4191–4201. [22] MacKay, B. A.; Fryzuk, M. D. Chem. Rev. 2004, 104, 385–401. [23] Leppin, J.; Forster, C.; Heinze, K. Inorg. Chem. 2014, 53, 12416–12427. [24] Berg, J. M.; Holm, R. H. J. Am. Chem. Soc. 1985, 107, 917–925. [25] Tsao, Y.-Y. P.; Fritchie, C. J.; Levy, H. A. J. Am. Chem. Soc. 1978, 100, 4089– 4095. [26] Wünsche von Leupoldt, A.; Förster, C.; Fiedler, T. J.; Bings, Nicolas H.; Heinze, Katja Eur. J. Inorg. Chem. 2013, 2013, 6079–6090. [27] Visser, S. P. de; Quesne, M. G.; Martin, B.; Comba, Peter; Ryde, Ulf Chem. Commun. 2014, 50, 262–282. [28] Marcus, R. A.; Sutin, N. BBA-Bioenergetics 1985, 811, 265–322. 37 Introduction [29] Closs, G. L.; Miller, J. R. Science 1988, 240, 440–447. [30] Gray, H. B.; Winkler, J. R. Annu Rev Biochem 1996, 65, 537–561. [31] Creutz, C.; Taube, H. J. Am. Chem. Soc. 1969, 91, 3988–3989. [32] Neidlinger, A.; Ksenofontov, V.; Heinze, K. Organometallics 2013, 32, 5955– 5965. [33] Gagliardi, C. J.; Vannucci, A. K.; Concepcion, J. J.; Chen, Zuofeng; Meyer, Thomas J. Energy Environ. Sci. 2012, 5, 7704. [34] Marcus, R. A. J. Chem. Phys. 1956, 24, 966. [35] Marcus, R. A. Discuss. Faraday Soc. 1960, 29, 21–31. [36] Marcus, R. A. J. Chem. Phys. 1965, 43, 679. [37] Marcus, R. A. Angew. Chem. Int. Ed. Engl. 1993, 32, 1111–1121. [38] Day, P.; Hush, N. S.; Clark, R. J. Philos. Trans. A Math. Phys. Eng. Sci. 2008, 366, 5–14. [39] Hush, N. S. Electrocim. Acta 1968, 13, 1005–1023. [40] Hush, N. S. Coord. Chem. Rev. 1985, 64, 135–157. [41] Creutz, C.; Taube, H. J. Am. Chem. Soc. 1973, 95, 1086–1094. [42] Cowan, D. O.; LeVanda, C.; Park, J.; Kaufman, Frank Acc. Chem. Res. 1973, 6, 1–7. [43] Delgado-Pena, F.; Talham, D. R.; Cowan, D. O. J. Organomet. Chem. 1983, 253, C43-C46. [44] Ribou, A.-C.; Launay, J.-P.; Sachtleben, M. L.; Li, Hu; Spangler, Charles W. Inorg. Chem. 1996, 35, 3735–3740. [45] Mathur, P.; Singh, A. K.; Singh, V. K.; Singh, Priti; Rahul; Mobin, Shaikh M.; Thöne, Carsten Organometallics 2005, 24, 4793–4798. [46] Lloveras, V.; Caballero, A.; Tárraga, A.; Velasco, M. Desamparados; Espinosa, Arturo; Wurst, Klaus; Evans, David J.; Vidal-Gancedo, José; Rovira, Concepció; Molina, Pedro; Veciana, Jaume Eur. J. Inorg. Chem. 2005, 2005, 2436–2450. [47] Siebler, D.; Forster, C.; Heinze, K. Dalton Trans. 2011, 40, 3558–3575. [48] Siebler, D., Dissertation, Johannes Gutenberg-University 2010. [49] Heinze, K.; Siebler, D. Z. Anorg. Allg. Chem. 2007, 633, 2223–2233. [50] Kealy, T. J.; Pauson, P. L. Nature 1951, 168, 1039–1040. [51] Fischer, E. O.; Pfab, W. Z. Naturforsch. B Chem. Sci. 1952, 7, 377–379. 38 References [52] Casado, C. M.; Cuadrado, I.; Moran, M.; Alonso, B.; Garcia, B.; Gonzalez, B.; Losada, J. Coord. Chem. Rev. 1999, 185-6, 53–79. [53] Huo, J.; Wang, L.; Yu, H.; Deng, Libo; Ding, Jianhua; Tan, Qiaohua; Liu, Qingquan; Xiao, Anguo; Ren, Guoqing J. Phys. Chem. B 2008, 112, 11490– 11497. [54] V. Mittal „Polymer Brushes“, Auflage, CRC Press 2012. [55] Plazuk, D.; Vessières, A.; Hillard, E. A.; Buriez, Olivier; Labbé, Eric; Pigeon, Pascal; Plamont, Marie-Aude; Amatore, Christian; Zakrzewski, Janusz; Jaouen, Gérard J. Med. Chem. 2009, 52, 4964–4967. [56] Zanellato, I.; Heldt, J.-M.; Vessières, A.; Jaouen, Gérard; Osella, Domenico Inorg. Chim. Acta 2009, 362, 4037–4042. [57] Biot, C.; Delhaes, L.; N'Diaye, C. M.; Maciejewski, L. A.; Camus, D.; Dive, D.; Brocard, J. S. Bioorg. Med. Chem. 1999, 7, 2843–2847. [58] Biot, C.; Delhaes, L.; Abessolo, H.; Domarle, O.; Maciejewski, L.A; Mortuaire, M.; Delcourt, P.; Deloron, P.; Camus, D.; Dive, D.; Brocard, J.S J. Organomet. Chem. 1999, 589, 59–65. [59] Biot, C.; Glorian, G.; Maciejewski, L. A.; Brocard, J. S. J. Med. Chem. 1997, 40, 3715–3718. [60] Biot, C.; Nosten, F.; Fraisse, L.; Ter-Minassian, D.; Khalife, J.; Dive, D. Parasite 2011, 18, 207–214. [61] Yousuf, M.; Mukherjee, D.; Pal, A.; Dey, Somaditya; Mandal, Supratim; Pal, Chiranjib; Adhikari, Susanta ChemMedChem. 2015, 10, 546–554. [62] Chantson, J. T.; Verga Falzacappa, M. V.; Crovella, S.; Metzler-Nolte, Nils ChemMedChem. 2006, 1, 1268–1274. [63] Lincke, T.; Behnken, S.; Ishida, K.; Roth, Martin; Hertweck, Christian Angew. Chem. 2010, 122, 2055–2057. [64] Hayakawa, Y.; Sasaki, K.; Nagai, K.; Shin-ya, Kazuo; Furihata, Kazuo J. Antibiot. 2006, 59, 6–10. [65] Ermler, U. Science 1997, 278, 1457–1462. [66] Shima, S.; Krueger, M.; Weinert, T.; Demmer, Ulrike; Kahnt, Jorg; Thauer, Rudolf K.; Ermler, Ulrich Nature 2012, 481, 98–101. 39 Introduction [67] Abas, S. A.; Hossain, M. B.; van der Helm, D.; Schmitz, Francis J.; Laney, Maureen; Cabuslay, Ronnel; Schatzman, Randall C. J. Org. Chem. 1996, 61, 2709–2712. [68] Pan, M.; Mabry, T. J.; Beale, J. M.; Mamiya, Blain M. Phytochemistry 1997, 45, 517–519. [69] Lee, H.-J.; Choi, Y.-S.; Lee, K.-B.; Park, Jeunghee; Yoon, Chang-Ju J. Phys. Chem. A 2002, 106, 7010–7017. [70] Banala, S.; Sussmuth, R. D. Chembiochem. 2010, 11, 1335–1337. [71] Reiner, A.; Wildemann, D.; Fischer, G.; Kiefhaber, Thomas J. Am. Chem. Soc. 2008, 130, 8079–8084. [72] Moser, C. C.; Keske, J. M.; Warncke, K.; Farid, R. S.; Dutton, P. L. Nature 1992, 355, 796–802. [73] MacDiarmid, A. G. Synth. Met. 2001, 125, 11–22. [74] Flores-Torres, S.; Hutchison, G. R.; Soltzberg, L. J.; Abruna, Hector D. J. Am. Chem. Soc. 2006, 128, 1513–1522. [75] Capozzi, B.; Xia, J.; Adak, O.; Dell, Emma J.; Liu, Zhen-Fei; Taylor, Jeffrey C.; Neaton, Jeffrey B.; Campos, Luis M.; Venkataraman, Latha Nat. Nanotechnol. 2015, 10, 522–527. [76] Hoffmann, N. ChemSusChem 2012, 5, 352–371. [77] Yao, C.-J.; Zheng, R.-H.; Shi, Q.; Zhong, Yu-Wu; Yao, Jiannian Chem. Commun. 2012, 48, 5680–5682. [78] Hu, K.; Robson, K. C.; Johansson, P. G.; Berlinguette, Curtis P.; Meyer, Gerald J. J. Am. Chem. Soc. 2012, 134, 8352–8355. [79] Brunschwig, B. S.; Sutin, N. Coord. Chem. Rev. 1999, 187, 233–254. [80] Klotz, I. M.; Czerlinski, G. H.; Fiess, H. A. J. Am. Chem. Soc. 1958, 80, 2920– 2923. [81] Cowan, D. O.; Kaufman, F. J. Am. Chem. Soc. 1970, 92, 219–220. [82] Robin, M. B.; Day, P. Adv. Inorg. Chem. Radiochem., 1968, 247–442. [83] Heckmann, A.; Lambert, C. Angew. Chem. Int. Ed. Engl. 2012, 51, 326–392. [84] Siebler, D.; Linseis, M.; Gasi, T.; Carrella, Luca M.; Winter, Rainer F.; Förster, Christoph; Heinze, Katja Chem. Eur. J. 2011, 17, 4540–4551. [85] Brunschwig, B. S.; Creutz, C.; Sutin, N. Chem. Soc. Rev. 2002, 31, 168–184. 40 References [86] Winter, R. F. Organometallics 2014, 33, 4517–4536. [87] Hildebrandt, A.; Lang, H. Organometallics 2013, 32, 5640–5653. [88] Kraatz, H.-B.; Bediako-Amoa, I.; Gyepi-Garbrah, S. H.; Sutherland, Todd C. J. Phys. Chem. B 2004, 108, 20164–20172. [89] Long, Y.-T.; Abu-Irhayem, E.; Kraatz, H.-B. Chemistry 2005, 11, 5186–5194. [90] Kirin, S. I.; Kraatz, H.-B.; Metzler-Nolte, N. Chem. Soc. Rev. 2006, 35, 348–354. [91] Herrick, R. S.; Jarret, R. M.; Curran, T. P.; Dragoli, Dean R.; Flaherty, Maryellen B.; Lindyberg, Susan E.; Slate, Rebecca A.; Thornton, Lisa C. Tetrahedron Lett. 1996, 37, 5289–5292. [92] van Staveren, D. R.; Weyhermüller, T.; Metzler-Nolte, N. Dalton Trans. 2003, 210. [93] Xu, Y.; Saweczko, P.; Kraatz, H.-B. J. Organomet. Chem. 2001, 637-639, 335– 342. [94] Kirin, S. I.; Wissenbach, D.; Metzler-Nolte, N. New J. Chem. 2005, 29, 1168. [95] Nomoto, A.; Moriuchi, T.; Yamazaki, S.; Ogawa, Akiya; Hirao, Toshikazu Chem. Commun. 1998, 1963–1964. [96] Moriuchi, T.; Nomoto, A.; Yoshida, K.; Ogawa, Akiya; Hirao, Toshikazu J. Am. Chem. Soc. 2001, 123, 68–75. [97] Moriuchi, T.; Nomoto, A.; Yoshida, K.; Hirao, Toshikazu J. Organomet. Chem. 1999, 589, 50–58. [98] Appoh, F. E.; Sutherland, T. C.; Kraatz, H.-B. J. Organomet. Chem. 2004, 689, 4669–4677. [99] Hatten, X. d.; Weyhermüller, T.; Metzler-Nolte, N. J. Organomet. Chem. 2004, 689, 4856–4867. [100] Adhikari, B.; Afrasiabi, R.; Kraatz, H.-B. Organometallics 2013, 32, 5899–5905. [101] Chowdhury, S.; Mahmoud, K. A.; Schatte, G.; Kraatz, Heinz-Bernhard Org. Biomol. Chem. 2005, 3, 3018–3023. [102] Djaković, S.; Siebler, D.; Semenčić, M. Č.; Heinze, Katja; Rapić, Vladimir Organometallics 2008, 27, 1447–1453. [103] Modeling of Molecular Properties; P. Comba, Weinheim, Germany 2011. [104] Semenčić, M. Č.; Siebler, D.; Heinze, K.; Rapić, Vladimir Organometallics 2009, 28, 2028–2037. 41 Introduction [105] Barisic, L.; Dropucic, M.; Rapić, V.; Pritzkow, Hans; Kirin, Srecko I.; Metzler- Nolte, Nils Chem. Commun. 2004, 2004–2005. [106] Heinze, K.; Wild, U.; Beckmann, M. Eur. J. Inorg. Chem. 2007, 2007, 617–623. [107] Semenčić, M. Č.; Heinze, K.; Förster, C.; Rapić, Vladimir Eur. J. Inorg. Chem. 2010, 2010, 1089–1097. [108] Okamura, T.-a.; Sakauye, K.; Doi, M.; Yamamoto, Hitoshi; Ueyama, Norikazu; Nakamura, Akira Bull. Chem. Soc. Jpn. 2005, 78, 1270–1278. [109] Heinze, K.; Schlenker, M. Eur. J. Inorg. Chem. 2004, 2004, 2974–2988. [110] Alemán, C. J. Phys. Chem. A 2001, 105, 6717–6723. [111] Handbook of chemistry and physics on CD-ROM; D. R. Lide, Boca Raton Fla. 2003. [112] Wiberg, K. B.; Rush, D. J. J. Org. Chem. 2002, 67, 826–830. [113] Wiberg, K. B.; Rush, D. J. J. Am. Chem. Soc. 2001, 123, 2038–2046. [114] Lauvergnat, D.; Hiberty, P. C. J. Am. Chem. Soc. 1997, 119, 9478–9482. [115] Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1995, 117, 2201–2209. [116] Galabov, B.; Ilieva, S.; Hadjieva, B.; Dinchova, Eli J. Phys. Chem. A 2003, 107, 5854–5861. [117] Deetz, M. J.; Fahey, J. E.; Smith, B. D. J. Phys. Org. Chem. 2001, 14, 463–467. [118] Rao, C. N.; Rao, K. G.; Goel, A.; Balasubramanian, D. J. Chem. Soc. A 1971, 3077–3083. [119] Wiberg, K. B.; Rablen, P. R.; Rush, D. J.; Keith, T. A. J. Am. Chem. Soc. 1995, 117, 4261–4270. [120] Artis, D. R.; Lipton, M. A. J. Am. Chem. Soc. 1998, 120, 12200–12206. [121] Miwa, J. H.; Pallivathucal, L.; Gowda, S.; Lee, Katherine E. Org. Lett. 2002, 4, 4655–4657. [122] Petersson, E. J.; Goldberg, J. M.; Wissner, R. F. Phys. Chem. Chem. Phys. 2014, 16, 6827–6837. [123] McFarland, B. J.; Katz, J. F.; Sant, A. J.; Beeson, Craig J. Mol. Biol. 2005, 350, 170–183. [124] Behling, L. A.; Hartsel, S. C.; Lewis, D. E.; DiSpirito, Alan A.; Choi, Dong W.; Masterson, Larry R.; Veglia, Gianluigi; Gallagher, Warren H. J. Am. Chem. Soc. 2008, 130, 12604–12605. 42 References [125] Wildemann, D.; Schiene-Fischer, C.; Aumuller, T.; Bachmann, Annett; Kiefhaber, Thomas; Lucke, Christian; Fischer, Gunter J. Am. Chem. Soc. 2007, 129, 4910– 4918. [126] Lincke, T.; Behnken, S.; Ishida, K.; Roth, Martin; Hertweck, Christian Angew. Chem. Int. Ed. Engl. 2010, 49, 2011–2013. [127] Nonoyama, M.; Hamamura, K. J. Organomet. Chem. 1991, 407, 271–277. [128] Hamamura, K.; Kita, M.; Nonoyama, M.; Fujita, Junnosuke J. Organomet. Chem. 1993, 463, 169–177. [129] Fernandes, T. A.; Solarova, H.; Cisarova, I.; Uhlik, Filip; Sticha, Martin; Stepnicka, Petr Dalton. Trans. 2015, 44, 3092–3108. [130] Damian Plazuk; Janusz Zakrzewski; Agnieszka Rybarczyk-Pirek; Slawomir Domagala J. Organomet. Chem. 2005, 690, 4302–4308. [131] van Staveren, D. R.; Metzler-Nolte, N. Chem. Rev. 2004, 104, 5931–5985. [132] Forrow, N. J.; Foulds, N. C.; Frew, J. E.; Law, John T. Bioconjug. Chem. 2004, 15, 137–144. [133] Marco-Contelles, J.; León, R.; Morales, E.; Villarroya, Mercedes; Garcı́a, Antonio G. Tetrahedron Lett. 2004, 45, 5203–5205. [134] Beer, P. D.; Graydon, A. R.; Johnson, A. O.; Smith, David K. Inorg. Chem. 1997, 36, 2112–2118. [135] Patra, M.; Hess, J.; Konatschnig, S.; Spingler, Bernhard; Gasser, Gilles Organometallics 2013, 32, 6098–6105. [136] Buettner, G. R. Free Radical Biol. Med 1987, 3, 259–303. [137] Rehorek, D.; Hennig, H.; Dubose, C. M.; Kemp, Terence J.; Janzen, Edward G. Free Radical Res. 2009, 10, 75–84. [138] Klauschenz, E.; Haseloff, R. F.; Volodarskii, L. B.; Blasig, Ingolf E. Free Radical Res. 2009, 20, 103–111. [139] Omelka, L.; Kováčová, J. Magn. Reson. Chem. 1994, 32, 525–531. [140] Hudson, A.; Lappert, M. F.; Lednor, P. W.; Nicholson, Brian K. J. Chem. Soc., Chem. Commun. 1974, 966. [141] Huffadine, A. S.; Peake, B. M.; Robinson, B. H.; Simpson, Jim; Dawson, Pete A. J. Organomet. Chem. 1976, 121, 391–403. [142] Benner, L. S.; Balch, A. L. J. Organomet. Chem. 1977, 134, 121–130. 43 Introduction [143] Hudson, A.; Lappert, M. F.; Nicholson, B. K. J. Chem. Soc., Dalton Trans. 1977, 551. [144] Fiorina, V. J.; Dubois, R. J.; Brynes, S. J. Med. Chem. 1978, 21, 393–395. [145] Ornelas, C. New J. Chem. 2011, 35, 1973. [146] Zhang, J. Appl. Organometal. Chem. 2008, 22, 6–11. [147] Arellano, I.; Rodríguez-Ramos, F.; González-Andrade, M.; Navarrete, Andrés; Sharma, Manju; Rosas, Noé; Sharma, Pankaj J. Heterocyclic Chem. 2016, 53, 1147–1154. [148] Wani, W. A.; Jameel, E.; Baig, U.; Mumtazuddin, Syed; Hun, Lee Ting Eur. J. Med. Chem. 2015, 101, 534–551. [149] Biot, C.; Daher, W.; Ndiaye, C. M.; Melnyk, Patricia; Pradines, Bruno; Chavain, Natascha; Pellet, Alain; Fraisse, Laurent; Pelinski, Lydie; Jarry, Christian; Brocard, Jacques; Khalife, Jamal; Forfar-Bares, Isabelle; Dive, Daniel J. Med. Chem. 2006, 49, 4707–4714. [150] Biot, C.; Chavain, N.; Dubar, F.; Pradines, Bruno; Trivelli, Xavier; Brocard, Jacques; Forfar, Isabelle; Dive, Daniel J. Organomet. Chem. 2009, 694, 845–854. [151] Pisciotta, J. M.; Sullivan, D. Parasitol. Int. 2008, 57, 89–96. [152] Chavain, N.; Vezin, H.; Dive, D.; Touati, Nadia; Paul, Jean-Francois; Buisine, Eric; Biot, Christophe Mol. Pharm. 2008, 5, 710–716. [153] Held, J.; Supan, C.; Salazar, C. L.; Tinto, Halidou; Bonkian, Léa N.; Nahum, Alain; Moulero, Bancole; Sié, Ali; Coulibaly, Boubacar; Sirima, Sodiomon B.; Siribie, Mohamadou; Otsyula, Nekoye; Otieno, Lucas; Abdallah, Ahmed M.; Kimutai, Robert; Bouyou-Akotet, Marielle; Kombila, Maryvonne; Koiwai, Kimiko; Cantalloube, Cathy; Din-Bell, Chantal; Djeriou, Elhadj; Waitumbi, John; Mordmüller, Benjamin; Ter-Minassian, Daniel; Lell, Bertrand; Kremsner, Peter G. Lancet Infect. Dis. 2015, 15, 1409–1419. [154] Osella, D.; Ferrali, M.; Zanello, P.; Laschi, Franco; Fontani, Marco; Nervi, Carlo; Cavigiolio, Giorgio Inorg. Chim. Acta 2000, 306, 42–48. [155] Tabbì, G.; Cassino, C.; Cavigiolio, G.; Colangelo, Donato; Ghiglia, Annalisa; Viano, Ilario; Osella, Domenico J. Med. Chem. 2002, 45, 5786–5796. [156] Jaouen, G.; Top, S.; Vessières, A.; Leclercq, G.; McGlinchey, Michael J. Curr. Med. Chem. 2004, 11, 2505–2517. 44 References [157] Michard, Q.; Jaouen, G.; Vessieres, A.; Bernard, B. A. J. Inorg. Biochem. 2008, 102, 1980–1985. [158] Pigeon, P.; Top, S.; Vessières, A.; Huché, Michel; Görmen, Meral; El Arbi, Mehdi; Plamont, Marie-Aude; McGlinchey, Michael J.; Jaouen, Gérard New J. Chem. 2011, 35, 2212. [159] Bruyère, C.; Mathieu, V.; Vessières, A.; Pigeon, Pascal; Top, Siden; Jaouen, Gérard; Kiss, Robert J. Inorg. Biochem. 2014, 141, 144–151. [160] Osella, D.; Mahboobi, H.; Colangelo, D.; Cavigiolio, Giorgio; Vessières, Anne; Jaouen, Gerard Inorg. Chim. Acta 2005, 358, 1993–1998. [161] Hagen, H.; Marzenell, P.; Jentzsch, E.; Wenz, Frederik; Veldwijk, Marlon R.; Mokhir, Andriy J. Med. Chem. 2012, 55, 924–934. [162] Marzenell, P.; Hagen, H.; Sellner, L.; Zenz, Thorsten; Grinyte, Ruta; Pavlov, Valeri; Daum, Steffen; Mokhir, Andriy J. Med. Chem. 2013, 56, 6935–6944. [163] Daum, S.; Chekhun, V. F.; Todor, I. N.; Lukianova, Natalia Yu; Shvets, Yulia V.; Sellner, Leopold; Putzker, Kerstin; Lewis, Joe; Zenz, Thorsten; de Graaf, Inge A M; Groothuis, Geny M M; Casini, Angela; Zozulia, Oleksii; Hampel, Frank; Mokhir, Andriy J. Med. Chem. 2015, 58, 2015–2024. 45 Introduction 46 2 Aim of Work The first part of this work focuses on N-ferrocenyl substituted thioamides in comparison to their parent carboxamide compounds to understand the effect of the oxygen to sulfur exchange in respect to the use of thioamides as isosteric replacements for carboxamides and asymmetric bridging units for redox centers in general. The facile synthetis of N- ferrocenyl substituted thioamides by means of oxygen to sulfur exchange by the use of Lawesson’s reagent (2,4-bis-(4-methoxyphenyl)-1,3,2,4-dithiadiphosphetane-2,4- disulfide) is chosen to benefit from the well-established chemistry of carboxamide bridged ferrocenes to vary substitution patterns and redox potentials. The influence of the oxygen to sulfur exchange on the E/Z or respectively cis/trans conformation of the synthesized mono- and dinuclear thioamides and their secondary structure is elucidated by NMR spectroscopy in order to determine the rotational barriers and favored structural motifs of this kind of compounds. A comparison to DFT calculations for this systems is used to validate the results and vice versa validate the DFT methods applied for future investigations. Oxidation of the dinuclear compounds to the mixed-valent species and probing by UV/Vis/NIR, EPR and NMR spectroscopy as well as electrochemical methods allows to determine the site of oxidation and the electronic coupling. This promises new insights into electron transfer reactions via asymmetrically bridged e.g facilitated or hindered electron transfer compared to parent carboxamides. The second part of this work centers on the ability of ferrocenium compounds to generate radicals under alkaline, non-nuceleophilic conditions. These ferrocenyl radicals can be envisaged to play a major role in the mode of action for ferrocenyl containing drugs and prodrugs like ferrocifen by Jaouen et al., ferroquine by Biot et al. and the aminoferrocene based prodrugs by Mokhir et al.. These short lived radicals are investigated by means of spin-trapping technique for EPR spectrsocopy providing insight in the localization of the unpaired electron. Quantitative EPR measurements are performed to elucidate the reactivity of the radicals. The last part explores a specific reactivity of N-ferrocenylthioamide under oxidizing and alkaline conditions, leading to EPR signals that cannot be explained by mere oxidation and deprotonation of the compounds. This study gives insight into the process of formation of 47 Aim of Work this open shell intermediates, as well as the follow up reactions under elimination of hydrogensulfide. Influences of ligand properties of the thioamide unit are disclosed as well. The complex mechanism proposed is supported by NMR spectroscopy, mass spectrometry, single crystal X-Ray diffraction and interpreted with the aid of DFT calculations. These study is valuable with respect to of synthesis of novel ferrocenyl containing polymers as well as ferrocenyl containing (pro)drugs. 48 3 Results and Discussion In this chapter the results are presented in the form of two publications and a submitted manuscript. The first section contains the publication “Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry”. The single crystal X-Ray diffraction analysis reported were done by Christoph Förster. The publication, which was published in Organometallics, was written by Torben Kienz (50%) and Katja Heinze (50%). All synthetic work and analytical investigations, as well as DFT calculations were performed by Torben Kienz. An easy route to N-ferrocenyl substituted thioamides using Lawesson’s reagent is reported. The synthesized mono- and dinuclear ferrocene compounds are analyzed by means of IR-, NMR-, EPR-, UV/Vis/NIR-spectroscopy and mass spectrometry. The rotational barrier of N-ferrocenyl substituted thioamides is disclosed. The dinuclear complexes exhibit hydrogen bonding leading to a distinct secondary structure. Upon oxidation the dinuclear compounds are mixed-valent species of Robin-Day class II with an electron coupling parameter HAB = 190 and 230 cm −1 . In the second section, the publication “Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene” is presented. EPR spectra of the compounds were performed by Andreas Neidlinger, expcept for the thioamide containing starting material, which was performed by Torben Kienz. The simulation of EPR spectra was done by both authors: Andreas Neidlinger (40 %) and Torben Kienz (60%). Electrochemical measurements were performed by Andreas Neidlinger. DFT calculations of the thioamide containing products were performed by Torben Kienz, while the remaining calculations were performed by Andreas Neidlinger. The publication, which was published in Organometallics, was written by Andreas Neidlinger (40 %), Torben Kienz (40%), and Katja Heinze (20 %). The generation of highly reactive carbon centered ferrocenyl radicals is reported. These radicals may provide an additional mode of action for ferrocenyl containing (pro-)drugs. 49 Results and Discussion The last section consists of the manuscript “Generation and Oligomerization of N- ferrocenyl Ketenimines via Open-shell Intermediates”. The single crystal X-Ray diffraction analysis reported were done by Christoph Förster. The preparative work and analyses, as well as DFT calculations were performed by Torben Kienz. The manuscript was written by Torben Kienz (60%) and Katja Heinze (40 %) and submitted to Organometallics. A specific reaction of N-ferrocenyl thioamide is reported. The mechanistical investigation reveals, that upon oxidation and deprotonation reactive open shell intermediates are generated, which eliminate hydrogen sulfide, generating a N-ferrocenyl ketenimine and initiating an oligomerization reaction. Furthermore the influence of the ligand properties of thioamide containing ferrocene derivatives is disclosed. To complete the work DFT calculations for the reactive open shell intermediates of the novel N-ferrocenyl ketenimnine were performed. 50 Contents 3.1 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Torben Kienz, Christoph Förster and Katja Heinze Published in: Organometallics 2014, 33, 4803–4812. [DOI: 10.1021/om500052k] http://pubs.acs.org/doi/abs/10.1021/om500052k “Adapted with permission from T. Kienz, C. Förster, K. Heinze, Organometallics 2014, 33, 4803−4812. Copyright 2014 American Chemical Society.” 3.2 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Andreas Neidlinger, Torben Kienz, Katja Heinze Published in: Organometallics 2015, 34, 5310−5320. [DOI: 10.1021/acs.organomet.5b00778] http://pubs.acs.org/doi/10.1021/acs.organomet.5b00778 “This is an unofficial adaptation of an article that appeared in an ACS publication. ACS has not endorsed the content of this adaptation or the context of its use. Copyright 2015 American Chemical Society.” 3.3 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Torben Kienz, Christoph Förster, Katja Heinze Submitted to: Organometallics 51 Results and Discussion 52 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry 3.1 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Torben Kienz, Christoph Förster, and Katja Heinze Organometallics 2014, 33, 4803–4812. 53 Results and Discussion Supporting information for this article (without Cartesian coordinates from DFT calculations) is found at pp. 155. For full supporting information refer to: http://pubs.acs.org/doi/suppl/10.1021/om500052k Adapted with permission from T. Kienz, C. Förster, K. Heinze, Organometallics 2014, 33, 4803-4812. Copyright 2014 American Chemical Society. 54 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry 3.1.1 Abstract The conformations and redox chemistry of ferrocenyl amides have been investigated in considerable depth in the last few years, while ferrocenyl thioamides have attracted less interest so far, although distinctly different conformations and reactivity patterns are expected. Monoferrocenyl amides Fc-NHC(O)CH3 (1) and 1,1’-CH3O(O)C-Fn- NHC(O)CH3 (2) and diferrocenyl amides Fc-NHC(O)-Fc (5) and Fc-NHC(O)-Fn- NHC(O)CH3 (6) are easily transformed into the corresponding thioamides (3, 4, 7, 8) by treatment with Lawesson’s reagent (2,4-bis(p-methoxyphenyl)-1,3-dithiaphosphetane-2,4- disulfide) (Fc = Fe(C5H4)(C5H5), Fn = Fe(C5H4)2). The thioamide conformations (cis/trans) in 3, 4, 7, and 8 and the hydrogen bond determined secondary structure of dithioamide 8 are elucidated by IR and NMR spectroscopy as well as by DFT calculations (B3LYP, LANL2DZ, PCM CH2Cl2) and contrasted with the corresponding amides 1, 2, 5, and 6. The electronic communication via the thioamide bridge in 7+ and 8+ in comparison to the interaction in the parent mixed-valent amides 5+ and 6+ has been probed by cyclic voltammetry, square wave voltammetry, UV/Vis spectroelectrochemistry, EPR spectroscopy, and paramagnetic NMR spectroscopy. Additional chemical reactivity of the thioamide unit has been detected by electrochemical analysis. 55 Results and Discussion 3.1.2 Introduction Although oxygen-to-sulfur isosteric substitutions have been successfully employed in organic peptide and protein chemistry, it is generally accepted that organic thioamides R- C(S)NH-R’ differ considerably from their oxo counterparts R-C(O)NH-R’ and the concept of true isosteric replacement might be too simplified. Thioamides are more acidic than amides: e.g. pKa(X=S) = 14.7 and pKa(X = O) = 21.5 for Me-C(X)NH-Ph.1 In line with this higher acidity thioamides are better hydrogen atom donors than amides in NH…X hydrogen bonds.2 On the other hand, due to the reduced electronegativity of sulfur vs. oxygen, thioamides are weaker hydrogen acceptors in XH…S hydrogen bonds.2 The better hydrogen atom donor capability of thioamides has been exploited by Beer et al. in ferrocenyl thioamide anion receptors (Chart 1, A, R = nBu).3 The macrocylic derivative B Chart 1. Ferrocenyl thioamides as anion (A) or cation receptors (B), cis/trans isomerism in N-aryl Thioamides (C), secondary structure of oligoferrocenyl oligopeptides (D), and mixed-valent oligoferrocenyl amides (E), disulfides (F), and benzothiazoles (G) derived from thioamides. 56 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry had been tested by Hall et al. as receptor for hard metal cations such as Ca2+ or Dy3+. Not too unexpectedly, these metal cations do not bind to B.4 On the other hand, nature seems to employ thio amino acids for binding of the softer metal cation Cu+.5 Thioamides are more stable towards hydrolysis by natural enzymes: e.g. by the Zn2+ containing enzyme carboxypeptidase A. Replacing the hard Zn2+ ion in the enzyme’s active site by the softer Cd2+ yields an active enzyme capable of cleaving thioamides.6 Furthermore, a fine tuning of the secondary and tertiary structure of enzymes can be achieved by replacing amides with thioamides.7 All ferrocene-containing thioamides such as A (Fc-C(S)NHR) and dithioamides such as B (1,1’-Fn(C(S)NHR)2) reported to date are formally derived from ferrocenecarboxylic acid or 1,1’-ferrocenedicarboxylic acid.3,4,8 To the best of our knowledge, R-C(S)NH-Fc thioamides derived from aminoferrocenes have not yet been reported. Interestingly, organic N-aryl substituted thioamides C are typically obtained as mixtures of cis and trans isomers (Chart 1) while N-alkyl thioamides are exclusively formed as trans isomers similarly to amides.9 The thioamide conformation of N-ferrocenyl thioamides remains to be elucidated. Oligoferrocenyl complexes with amide bridges (Chart 1, D) have been shown to form distinct secondary structures with intramolecular hydrogen bonds characterized by eight- membered rings.10 Mixed-valence cations of such oligoferrocenyl amides E feature intervalence charge transfer bands around λIVCT = 1030 – 1055 nm with electronic couplings HAB = 140 – 190 cm –1, placing these mixed-valent species E in the Robin-Day class II11 with moderate electronic interaction.10 Hence, these oligoamides are molecular wires equipped with a distinct secondary structure. As multiple positive charges are opponents to the intramolecular hydrogen bonds, full oxidation should disrupt the hydrogen bonds, giving an extended conformation10 and rendering these amide-based foldamers responsive to redox stimuli. With the special features of thioamides in mind, we were interested in the effects of the O  S substitution in N-ferrocenyl (thio)amides with respect to (thio)amide conformation (cis/trans isomers?), secondary structure (occurence of intramolecular hydrogen bonds?), electronic communication in mixed-valent systems (facilitated or hampered electron transfer?), and redox stability of the ferrocenyl thioamide unit. In the last case especially the oxidation/deprotonation of the thioamide to the corresponding disulfide RC(NR’)-S-S- (NR’)CR (Chart 1, F)12 or benzothiazole G (Chart 1)12a is of particular interest, as this 57 Results and Discussion reactivity is obviously impossible for the parent amides. The diferrocenoyl disulfide Fc- C(O)-S-S-C(O)-Fc has been reported previously.13 3.1.3 Results and Discussion Synthesis and Characterization. The mono- and diferrocenyl thioamides 3, 4, 7, and 8 shown in Scheme 1 were conveniently prepared by thionation of the corresponding amides 1, 2, 5, and 610,14 with Lawesson’s reagent (2,4-bis(p-methoxyphenyl)-1,3-dithiaphosphetane-2,4-disulfide).15 After column chromatography the products were obtained as orange (3, 4) to red powders (7, 8) in 40– 90% isolated yield. All complexes were characterized by FD or ESI mass spectrometry, NMR spectroscopy, and elemental analyses proving the desired composition and purity. Scheme 1. Synthesis of thioamides 3, 4, 7, and 8 from the corresponding amides 1, 2, 5, and 6 (atom numbering for NMR assignment). In the UV/vis spectra of the thioamides 3, 4, 7, and 8 the characteristic ferrocene absorption band in CH2Cl2 is consistently shifted to lower energy in comparison to that of the corresponding amides 1, 2, 5, and 6 (λmax = 441, 441, 445, 446 nm 10,14  λmax = 446, 444, 470, 480 nm). These shifts already indicate some electronic influence of the O  S substitution and some admixed charge-transfer character of the Fc band (see Supporting Information, Figures S1 – S4). 58 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry The IR spectra obtained as KBr disks show characteristic NH and CS(I) and CS(II) absorption bands. The CO band of the ester group in 4 at 1686 cm–1 confirms that the ester carbonyl oxygen atom is not exchanged by sulfur using Lawesson’s reagent and furthermore that the COester group acts as a hydrogen atom acceptor in the solid state. The NH stretching frequencies around 3230–3310 cm–1 are indicative of the presence of hydrogen bonds in the solid state. An analogous observation has also been made for the corresponding amides.10,14 4 shows two absorptions for the NH group at 3462 (br) cm–1 and 3288 (m) cm–1, suggesting free and hydrogen-bonded NH groups, respectively (NH…Oester). The dithioamide 8 features NH absorptions at 3375 and 3233 cm–1 corresponding to free NH groups and hydrogen-bonded (NH…S) groups, respectively. In CH2Cl2 solution the hydrogen bonds of 3 and 7 are essentially disrupted, as the NH stretching vibrations are observed around 3380–3400 cm–1. This clearly proves the intermolecular nature of the hydrogen bonds in 3 and 7 in the solid state. Ester 4 shows two NH absorption bands at 3379 and 3281 cm–1 in solution. The COester vibration of 4 is shifted to 1713 cm–1, indicative of essentially free CO groups, in addition to some hydrogen- bonded CO groups (shoulder at 1695 cm–1). Hence, ester 4 features some (intramolecular) NH…Oester hydrogen bonds also in CH2Cl2 solution resulting in a six-membered ring (vide infra). Dithioamide 8 also shows the signatures of hydrogen-bonded and free NH groups at 3165 and 3384 cm–1 in solution, suggesting the presence of an intramolecular NH…S hydrogen bond in addition to a free NH group. The diamides of 1,1’-disubstituted ferrocenes have been shown to possess intramolecular hydrogen bonds as well.10,14 The detailed conformations (cis/trans) and especially the intramolecular hydrogen bond of 8 will be disclosed in the next section by NMR and DFT methods. Conformational Analysis. N-Ferrocenyl thioamides 3 and 4 display a double set of resonances in their 1H and 13C NMR spectra, indicative of the presence of cis and trans isomers which do not interconvert at room temperature on the NMR time scale (Figure 1, Experimental Section). This finding is congruent to reported organic N-aryl thioamides.9 From NOE spectra resonances are assigned to cis and trans isomers by the decisive contact of the methyl protons H6 to the thioamide proton NH4 in the trans isomer (see Supporting Information, Figure S5). In addition to significantly different chemical shifts of the methyl protons H6, the methyl 59 Results and Discussion carbon atom C6, and the C=S carbon atom C5 the cyclopentadienyl H2 resonances of trans- 3 and trans-4 are shifted to lower field by 0.47 and 0.36 ppm in comparison to the respective cis counterparts, respectively, due to the vicinity of the sulfur atom to the H2 proton in the trans isomers. Conversely, the NH4 resonance is shifted to lower field by 0.73 and 0.37 ppm, respectively, in the cis isomers as compared to the trans isomers due to the neighboring sulfur atom. The cis:trans ratio obtained by integration of corresponding proton resonances is approximately 1 : 1 for 3 and 1 : 2.7 for 4. Heating a toluene solution of 3 to 110 °C allows observation of the coalescence of the methyl proton resonances H6 at around Tc = 373 K (Figure 2). From Tc and the chemical shifts of the CH3 proton resonances (H6) an activation barrier of ΔG‡ = 75 kJ mol–1 can hence be estimated for the cis/trans isomerization of 3, similar to the barriers determined for N,N’-dimethylthioformamide and N,N’-dimethylthioacetamide.16 Figure 1. 1H NMR spectra of (a) 3 and (b) 4 in CD2Cl2 at room temperature (atom numbering see Scheme 1). DFT calculations (B3LYP, LANL2DZ, PCM CH2Cl2) yield somewhat higher activation barriers of 101 and 110 kJ mol–1 for 3 and 4, respectively (Figure 3). The calculations show a slight preference of the trans isomers, especially for 4 which matches the experimental results. trans-4 might form an intramolecular NH…Oester hydrogen bond slightly more stable than that of cis-4 (trans-4: NH…O 1.94 Å; cis-4: NH…O 2.00 Å). Such intramolecular 60 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry NH…O hydrogen bonds giving a six-membered ring (Figure 3b) are likely present to some extend in the ensemble in solution, as already suggested by the IR data. On the other hand, intermolecular hydrogen bonds as suggested by the solid-state IR spectra might stabilize the cis isomer. Indeed, thioamide 3 crystallizes in the monoclinic space group P21/c as centrosymmetric hydrogen-bonded dimers of cis isomers (Figure 4). In contrast, its amide analogue 1 crystallizes with NH…O hydrogen-bonded chains of trans- 1 in the solid state.[14a,17] The N…S distance in cis-3 amounts to 3.39 Å. The cis thioamide unit is nearly planar but twisted relative to the Cp plane with a C10-C6-N1-C11 angle of 51°. This twist is also seen in the DFT calculation of cis-3 (44°) and of the hydrogen bonded dimer (cis-3)2 (39.0°, see Supporting Information, Figure S6). In contrast to the case for N-ferrocenyl amides 1 and 2 cis/trans isomers with high rotation barriers are found in N-ferrocenyl C-methyl thioamides 3 and 4. The N-ferrocenyl C-ferrocenyl thioamide 7, however, displays only single sets of proton and carbon resonances in the respective NMR spectra suggesting the presence of a single isomer only. DFT calculations favor the trans isomer trans-7 by 9.8 kJ mol–1 with a rather planar Cp- NHC(S)-Cp bridge, while the cis isomer cis-7 is severely distorted (see Supporting Information, Figure S7). The NOE spectrum of 7 in CD2Cl2 (see Supporting Information, Figure S8) reveals NOE contacts of the NH1 proton to H8 (strong) and H3 (medium) fully consistent with the trans configuration (DFT distances of trans-7: H1…H8 2.05 Å, H1…H3 2.47 Å; DFT distances of cis-7: H1…H8 3.97 Å, H1…H3 3.06 Å, Supporting Information, Figure S7). Figure 2. Variable temperature 1H NMR spectra of 3 in d8-toluene. 61 Results and Discussion For diferrocene dithioamide 8 several further conformational possibilities are conceivable (Figure 5). A double signal set is observed in the integral ratio 1 : 6.4 in the 1H NMR spectrum of 8. The major isomer is assigned to the trans,trans isomer (trans,trans-8) on the basis of an NOE cross peak between the methyl group H15 and the thioamide proton NH13 (trans-Cp-NHC(S)Me moiety) and two NOE contacts between thioamide NH1 and H8 (strong) and H3 (medium) similar to the trans-7 case. These data are consistent with the DFT calculated metrics of the trans,trans-8 isomer (H1…H8 2.22 Å, H1…H3 2.48 Å; H13…H15 2.15 Å). The minor isomer in solution is likely the trans,cis isomer with a terminal cis-Cp-NHC(S)Me moiety (trans,cis-8) similar to 3 and 4. Figure 3. DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated cis/trans isomerization of a) 3 and b) 4. 62 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry As dithioamide 8 features two hydrogen atom donors and two hydrogen atom acceptors, intramolecular NH…S hydrogen bonds are feasible similar to the all-amide complex 6. Indeed, amide 6 folds into a secondary structure in solution with eight-membered NH…O hydrogen-bonded rings (cf. Chart 1, D). The IR spectra of 8 suggest an intramolecular hydrogen bond in CH2Cl2 solution as well (vide supra). DFT calculations of trans,trans-8 in different folding arrangements show that – in contrast to 6 – the six-membered NH…S hydrogen-bonded rings are more favorable (Figure 5). The most stable arrangement of 8 comprises the six-membered ring and a 1,2’-conformation of the Cp rings of the disubstituted ferrocene. Figure 4. Hydrogen bonding motif of cis-3 in the solid state. In the 1H1H NOE spectrum (Figure 6) the terminal thioamide proton NH13 of 8 displays three contacts to neighboring protons, namely one to the methyl group (H15), a second one to the adjacent Cp proton H11, and –decisively– a third contact to a Cp proton adjacent to the bridging C(S) substituent H8. The latter contact is only feasible in the 1,2’-conformation with the six-membered ring (NH13…H8 = 3.00 Å) while for all other conformations (Figure 5) this distance is larger than 3.88 Å according to DFT calculations. The preference of the smaller six-membered ring in the dithioamide 8 as opposed to the eight-membered ring in the diamide 6 can be ascribed to the larger atomic radius of sulfur as compared to oxygen and the concomitantly longer bonds. To conclude this section, Me-C(S)NH-Fc thioamides can form cis and trans isomers while the trans isomer is preferred in Fc-C(S)NH-Fc 63 Results and Discussion Figure 5. DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated hydrogen- bonded isomers of trans,trans-8. Figure 6. 1H1H NOESY of 8 in CD2Cl2 at room temperature (top) and selected NOE relevant metric data of the lowest energy conformation of 8 (bottom, only relevant protons are shown). 64 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry thioamides. In addition the disubstituted ferrocene Me-C(S)NH-Fn-C(S)NH-Fc (8) preferably arranges in a six-membered ring with an intramolecular NH…S hydrogen bond. Electrochemistry. The monoferrocenyl derivatives 3 and 4 are oxidized in CH2Cl2 to the respective ferrocenium cation radicals 3+ and 4+ at E½ = –0.020 V and E½ = 0.205 V (vs. FcH/FcH +), respectively (Figure 7). The higher potential of the 4/4+ couple is due to the presence of the electron withdrawing ester substituent, as expected.10a,18 In comparison to the respective amides 1/2 the potentials of 3/4 are shifted to higher values by 85 – 100 mV (Figure 7). Obviously, the NHC(S)Me group is more electron withdrawing than the NHC(O)Me substituent. At higher potentials an irreversible wave is observed (Ep = 0.345 V and Ep = 0.510 V for 3 and 4, respectively). This oxidation (corresponding to only 0.5 electrons) can be ascribed to oxidation of the sulfur in 3+ or 4+ to give transient ferrocenium surfuryl diradicals 32+ or 42+. These doubly charged ferrocenium thioamide radical cations should be highly acidic and should easily lose a proton to give the diradicals [3-H]+ or [4-H]+, respectively (Scheme 2). In full agreement DFT calculations on trans-3n+ (n = 0–2) assign the first oxidation to the Fc/Fc+ couple and the second oxidation to the sulfur atom, as judged from Fe…Cp(centroid) distances and spin densities (see the Supporting Information, Figure S9). In agreement with the proposed increased acidity the NH bond lengths increase as the positive charge becomes greater, from 1.018, 1.019 to 1.022 Å (see Supporting Information, Figure S9). The sulfur-based radicals should be prone to disulfide formation. Hence, we propose an intermolecular radical coupling of [3-H]+ with 3+ and [4-H]+ with 4+, respectively. This suggestion fits the observed transfer of only 0.5e in the second oxidation step (Figure 7). A further proton loss might stabilize the resulting mixed-valence disulfides [3-H] +2 and [4-H] +2 (Scheme 2). A ferrocenoyl-containing disulfide Fc-C(O)- SS-C(O)-Fc had been prepared before by iodine oxidation and deprotonation of Fc-C(O)SH or by thionation of Fc-C(O)-im (im = N-imidazolide) with Lawesson’s reagent.13 Oxidation of organic thioamides by iodine, Cu2+ or quinones has been reported to yield disulfides with concomitant deprotonation.12 Hence, the proposed EEC pathway (Scheme 2) is a viable mechanism. 65 Results and Discussion Figure 7. Cyclic (bottom) and square-wave voltammograms (top) of a) 1/3 and b) 2/4 in CH2Cl n2/[ Bu4N][PF6]. Scheme 2. Suggested EEC mechanism and intermolecular disulfide formation of 3 and 4 (postulated species in brackets). The observed redox processes are basically similar for the diferrocenes 7 and 8 with the first oxidation potential of the ferrocene units shifted to higher values in comparison to the corresponding amides 5 and 6 (Figure 8). We ascribe the first two oxidation waves to the oxidations of the ferrocene units to give the mixed-valent species 7+/8+ and the all-FeIII 66 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry complexes 72+/82+, respectively. At somewhat higher potentials the irreversible sulfur oxidations are observed, yielding the reactive trications 73+/83+ (Figure 8). Similar to 3/4, deprotonation and disulfide formation might be the dominating pathway after 3-fold oxidation of 7 and 8. For dithioamide 83+ intramolecular disulfide formation might be feasible in addition to intermolecular S-S coupling. A possible reaction pathway yielding a mixed-valent disulfide ansa-ferrocene [8-2H]+ is depicted in Scheme 3. According to DFT calculations the proposed disulfide [8-2H]+ is shown to be an unstrained [5]ferrocenophane with an appended ferrocenium ion (Figure 9).20 Figure 8. Cyclic (bottom) and square-wave voltammograms (top) of a) 5/7 and b) 6/8 in CH2Cl2/[nBu4N][PF6]. The monocations 1+, 2+, 5+, and 6+ as well as 3+, 4+, 7+, and 8+ were prepared electrochemically in an optically transparent thin-layer electrode cell and the UV/vis spectra were recorded. As an example, the spectroelectrochemical oxidation of 8 to 8+ is depicted in Figure 10. A clean oxidation of the ferrocene 8 (480 nm) to the mixed-valent cation 8+ (527, 751, 1170 nm) is observed with isosbestic points at 464 and 500 nm. The typical ferrocenium absorption bands of the amides are observed at 756, 799, 759 and 799 nm (1+, 2+, 5+, and 6+; from Gaussian band shape analysis) while those of the thioamides are found at 799, 797, 800 and 751 nm (3+, 4+, 7+, and 8+; from Gaussian band shape analysis, Supporting Infor-mation, Figures S10 and S11). In addition, the mixed-valent complexes 5+, 6+, 7+, and 8+ show intervalence charge transfer bands at 1075, 961, 1150, and 1170 nm from Gaussian band shape analysis (Eop = 9300, 10400, 8695, 8547 cm –1). 67 Results and Discussion Scheme 3. Suggested EEEC mechanism and intramolecular disulfide formation of 8 (postulated species in brackets). These values show that in the thioamides the optical electron transfer occurs at lower energy. From the Hush formula, the band shape parameters of the IVCT bands (see Supporting Information, Figures S10 and S11) and an estimated Fe…Fe distance of 7.2 Å the electronic couplings are calculated as HAB = 200, 260, 190, and 230 cm –1 (± 10 cm–1) for 5+, 6+, 7+, and 8+, respectively. These moderate values assign class II to 7+ and 8+ according to the Robin–Day classification.11 The similar electronic coupling in amides and their corresponding thioamides suggests that the electron transfer is not significantly mediated by the (thio)amide bridge but rather occurs through space. Activation barriers for the thermal electron transfer in the mixed-valent cations are estimated from ΔG‡ET = λ/4 + ΔG 0/2 + (ΔG0)2/(4(λ – 2HAB) – HAB + HAB2/(λ + ΔG 0) with λ = E 0op – ΔG as 0.43, 0.40, 0.36, and 0.34 eV for 5+, 6+, 7+, and 8+, respectively (42, 39, 35, 33 kJ mol–1).11b The significantly lower barriers in the thioamides Fc-NHC(S)-FnR (7+, 8+) are largely based on the smaller electronic differences of the ferrocene/ferrocenium redox sites in thioamides. The oxidation potential of Fc-NHC(S)R is more similar to that of Fc-C(S)NHR in comparison to the very different potentials of Fc-NHC(O)R and Fc- C(O)NHR. Facilitated electron transfer in thioamides 7+ and 8+ is hence based on a more pronounced redox similarity of the redox sites (smaller ΔG0) as opposed to a stronger electronic interaction via the C(X)NH bridge (similar HAB). 68 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Figure 9. DFT calculated structure of [8-2H]+ (CH hydrogen atoms omitted for clarity) including spin density distribution (isosurface value 0.01 a.u.). Figure 10. Spectroelectrochemical oxidation of 8 to 8+ in CH2Cl n2/[ Bu4N][B(C6F5)4]. To experimentally corroborate the preferred site of oxidation the diferrocenes 7 and 8 were titrated with iodine (E½ = –0.14 V vs. ferrocene in CH3CN) 19 and paramagnetic 1H NMR spectra were recorded.10b,20a As an example, the 1H NMR spectra of 7 and increasing substoichiometric amounts of iodine are displayed in Figure 11 (see Supporting information, Figure S12 for titration of 8 with I2). Due to the typically fast intermolecular self-exchange reaction of the Fc/Fc+ couple, averaged NMR resonances are observed. Clearly, the resonances of the NH-substituted ferrocene (H3, H4, H5) are more affected (paramagnetic shift, paramagnetic broadening) than the resonances of the CS-substituted ferrocene (H8, H9, H10) demonstrating that the NH-substituted ferrocene is the site of 69 Results and Discussion primary oxidation, similar to the amide case 5/5+.10b Fully analogously, the site of first oxidation in 8/8+ is the monosubstituted ferrocene, as shown by iodine titrations (see Supporting Information, Figure S12) and DFT calculations. Figure 11. 1H NMR spectra of 7 upon titration with iodine in CD2Cl2 (atom numbering according to Scheme 1). The monocations 3+, 4+, 7+, and 8+ were additionally prepared by chemical oxidation using one equiv of AgSbF6 in a mixture of CH2Cl2 and THF (4/1). After rapid freezing to 77 K EPR spectra of the ferrocenium cation radicals were recorded.21 Nearly axial EPR spectra were obtained with gz = 3.400, gx,y = 1.860 (3+, Δg = 1.54), gz = 3.300, gx,y = 1.852 (4+, Δg = 1.45) (see Supporting Information, Figures S13 and S14) and gz = 3.300, gy = 1.881, gx = 1.845 (7+, Δg = 1.46, Figure 12a). Comparable resonances have been observed previously for [Fc-NHCOR]+ radicals in CH2Cl2 and assigned to contact ion pairs with the [SbF ] – 6 ion attached via NH…F hydrogen bonds.21b In THF solvent separated ion pairs are additionally present which feature NH…O hydrogen bonds to the THF oxygen atom. These solvated [Fc-NHCOR]+ radicals possess a smaller g anisotropy and sharper resonance lines.21b Notably, 8+ shows two signal sets with a broad resonance at gz = n.o., gy = 1.889, gx = 1.849 (50%) suggestive of a contact ion pair and a sharp prominent resonance at gz = 2.340, gy = 2.067, gx = 2.001 (Δg = 0.34, 50%) with a small g anisotropy (Figure 12b). The latter resonance suggests that the spin-carrying terminal [Fc-NHCSR]+ unit is not hydrogen- bonded to [SbF ]–6 but rather to a different acceptor (oxygen of THF or sulfur of bridging C=S). In contrast to 3+, 4+, and 7+ a bridging C=S unit is available in 8+ as an internal hydrogen atom acceptor. This results in an eight-membered-ring system in 8+ (Figure 12b). 70 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry A hexafluoroantimonate counterion that might be additionally attached to the terminal NHCSCH3 group does not dramatically influence the terminal Fc + radical (Figure 12b and Supporting Information). The terminal Fc+ seems to be influenced mainly by the donor atom coordinated to its proximal NH group (THF, counter ion, sulfur atom). Figure 12. X-band EPR spectra in CH2Cl2 and THF (4:1) at 77 K, simulations and DFT calculated spin density distributions (isosurface value 0.01 a.u.) of a) 7+ and b) 8+×[SbF6]– (bottom). Only relevant hydrogen atoms are shown. 71 Results and Discussion Interestingly, the DFT optimized lowest energy conformation of 8+ features an eight- membered ring, confirming the EPR results. This conformation of 8+ (eight-membered ring, 1,2’-conformation) is calculated to be more stable than the six-membered ring (1,2’- conformation) by 8 kJ mol–1 (see Supporting Information). Hence, the intramolecular NH…S hydrogen bond of the acidic terminal Fc+-NH group favors the eight-membered ring in 8+ as opposed to neutral 8 with a six-membered ring (Figure 5). The DFT calculations do not reveal a significant spin density on sulfur atoms in 3+, 4+, 7+, and 8+, respectively, ruling out a significant contribution of sulfur radical character in the singly oxidized complexes. In summary, the mixed-valent cation 8+ likely features a different secondary structure (eight-membered ring) different from that of its parent neutral thioamide 8 (six-membered ring) due to a changed preference of intramolecular hydrogen bonds (Scheme 3). This contrasts to the secondary structure of ferrocenyl amides which is preserved upon a single electron transfer (eight-membered ring). 3.1.4 Conclusion Oxygen to sulfur exchange in ferrocenyl amides results in distinct consequences concerning preferred conformations of neutral complexes, electronic communication in mixed-valence systems, and reactivity of oxidized species. Amides Fc-NHC(O)Me (e.g. 1, 2) prefer trans conformations of the amide, while for thioamides Fc-NHC(S)Me (3, 4) both cis and trans isomers are equally accessible. The barrier of cis/trans isomerization has been measured as 75 kJ mol–1 for 3 by variable-temperature NMR spectroscopy. The trans configuration of the central thioamide is again favored for diferrocenyl thioamides of the type Fc-NHC(S)- FnR (7, 8). Dithioamide 8 features a well-defined secondary structure in solution comprising an intramolecular NH…S hydrogen bond giving a six-membered ring in contrast to the oxo analogue 6 featuring an eight-membered ring with an intramolecular NH…O hydrogen bond. In the thioamides the Fc/Fc+ redox couples are typically shifted to higher potentials in comparison to the corresponding amides. In addition irreversible oxidations follow at even higher potential which can be ascribed to disulfide formations. In the triply charged dithioamide 83+ the disulfide formation may even occur intramolecularly and the possible 72 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry formation of an ansa-ferrocene [8-2H]+ is proposed. This proposed oxidatively induced cyclization reaction will be investigated in more detail in the future. Furthermore, the preferred secondary structure (1,2’; six-membered ring) of the dithioamide 8 transforms into an eight-membered ring with an inverted NH…S hydrogen bond motif upon oxidation to 8+. This perception will be exploited in the near future for redox stimulated conformational changes in oligothiopeptides. The mixed-valence cations 7+ and 8+ belong to the Robin-Day class II with electronic coupling similar to that of the corresponding mixed-valent amides 5+ and 6+ (HAB ≈ 220 cm–1). However, electron transfer is more rapid in thioamides 7+ and 8+, due to the diminished redox asymmetry of the C(S)NH bridge as compared to the C(O)NH bridge. In any event, the most easily oxidized site is still the monosubstituted Fc-NHCX subunit irrespective of the chalcogen atom X of the bridge. 3.1.5 Experimental Section General Procedures: All reactions were performed under an argon atmosphere unless otherwise noted. Dichloromethane was dried with CaH2 and distilled prior to use. All reagents were used as received from commercial suppliers (Acros, Sigma-Aldrich, ABCR). 1, 2, 5, and 6 were prepared according to literature procedures.10,14 NMR spectra were recorded on a Bruker DRX 400 spectrometer at 400.31 MHz (1H) and 100.07 MHz (13C{1H}). All resonances are reported in ppm vs. the solvent signal as internal standard. CD2Cl2 ( 1H: δ 5.32 ppm; 13C: δ 54.0 ppm). d8-Toluene ( 1H: δ 2.30, 7.19 ppm). IR spectra were recorded with a BioRad Excalibur FTS 3100 spectrometer as KBr disks or by using KBr cells in CH2Cl2. Electrochemical experiments were carried out on a BioLogic SP-50 voltammetric analyzer by using a platinum working electrode, a platinum wire counter electrode and a 0.01 M Ag/AgNO3 reference electrode. The measurements were carried out at a scan rate of 100 mV s−1 for cyclic voltammetry experiments and at 50 mV s−1 for square wave voltammetry experiments in 0.1 M [nBu4N][PF6] as supporting electrolyte in CH2Cl2. Potentials are referenced against the decamethylferrocene/decamethylferrocenium couple (E½ = 550 ± 5 mV vs ferrocene/ferrocenium under our experimental conditions) and are given relative to the ferrocene/ferrocenium couple. UV/vis/near-IR spectra were recorded on a Varian Cary 5000 spectrometer by using 1.0 cm cells (Hellma, suprasil). Spectroelectrochemical 73 Results and Discussion experiments were performed using a thin layer quartz glass (path length 1 mm) cell kit (GAMEC Analysentechnik, Illingen, Germany) equipped with a Pt gauze working electrode, a Pt counter electrode and a Ag/AgNO3 reference electrode (0.01 M solutions in CH2Cl2 containing 0.1 M [ nBu4N][B(C6F5)4]). CW EPR spectra (X-band; ca. 9.4 GHz; ca. 20 mM) were measured on a Miniscope MS 300 at 77 K cooled by liquid nitrogen in a finger dewar (Magnettech GmbH, Berlin, Germany). Settings were as follows: center field: 2499.01 G; modulation amplitude: 3000 mG; receiver gain: 5.0; microwave attenuation: 3 dB; sweep time: 120 s. g-values are referenced to external Mn2+ in ZnS (g = 2.118, 2.066, 2.027, 1.986, 1.946, 1.906). Simulations of EPR spectra were performed with EasySpin (v 4.0.0)22 for MatLab (R2007b). FD mass spectra were recorded on a FD Finnigan MAT95 spectrometer. ESI mass spectra were recorded on a Micromass Q-TOF-Ultima spectrometer. Melting points were determined by using a Gallenkamp MFB 595 010 capillary melting point apparatus and were not corrected. Elemental analyses were performed by the microanalytical laboratory of the chemical institutes of the University of Mainz. Density Functional Calculations. Density functional calculations were carried out with the Gaussian09/DFT series23 of programs. The B3LYP formulation of density functional theory was used employing the LANL2DZ basis set. No symmetry constraints were imposed on the molecules. The presence of energy minima of the ground states and first order saddle points was checked by analytical frequency calculations. Solvent modeling was done employing the integral equation formalism polarizable continuum model (IEFPCM, dichloromethane). The approximate free energies at 298 K were obtained through thermochemical analysis of the frequency calculation, using the thermal correction to Gibbs free energy as reported by Gaussian09. Crystal Structure Determination. Intensity data were collected with a Bruker AXS Smart1000 CCD diffractometer with an APEX II detector and an Oxford cooling system and corrected for absorption and other effects using Mo Kα radiation ( = 0.71073 Å). The diffraction frames were integrated using the SAINT package and most were corrected for absorption with MULABS.24,25 The structures were solved by direct methods and refined by the full-matrix method based on F2 using the SHELXTL software package.26,27 All non-hydrogen atoms were refined 74 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry anisotropically while the positions of all hydrogen atoms were generated with appropriate geometric constraints and allowed to ride on their respective parent carbon/nitrogen atoms with fixed isotropic thermal parameters. Crystal data: C12H13FeNS (259.15); T = 173 K; yellow plate; 0.21  0.17  0.02 mm; monoclinic; P21/c; a = 14.2351(8) Å; b = 9.4951(6) Å; c = 8.4156(5) Å; β = 105.437(2)°; V = 1096.45(11) Å3; Z = 4; F(000) = 536.0;  = 1.570 g cm–3; µ = 1.528 mm–1; 2 range = 5.22° – 55.76°; index ranges –18 ≤ h ≤ 18; –12 ≤ k ≤ 11; –11 ≤ l ≤ 10; reflections collected 15617; 2615 independent reflections; 136 parameters; maximum/minimum transmission = 0.9701/0.7397; goodness of fit on F2 0.957; largest difference peak and hole = 0.468/–0.381 e Å–3; R1(I > 2) = 0.0305; R1(all data) = 0.0465; Rw(I > 2) = 0.0709; Rw(all data) = 0.0741. CCDC-978126 contains supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Synthesis of 3. Amide 1 (298 mg, 1.23 mmol) and Lawesson’s reagent (248 mg, 0.61 mmol) were suspended in toluene (250 mL) and heated to 80 °C for 3 h. The solvent was removed under reduced pressure. After column chromatography (SiO2, 3  12 cm, CH2Cl2) 3 was obtained as orange solid (286 mg, 1.10 mmol, 90%). Rf(SiO2, CH2Cl2) = 0.26. M.p. 130°C dec. 1H NMR (400 MHz, CD2Cl2, 298 K): [cis-3] δ = 9.01 (s, 1H, NH 4), 4.36 (pt, 2H, 3JHH = 1.83 Hz, H2), 4.23 (s, 5H, H7), 4.17 (pt, 2H, 3JHH = 1.83 Hz, H 3), 2.40 ppm (s, 3H, H6); [trans-3] δ = 8.28 (s, 1H, NH4), 4.83 (pt, 2H, 3JHH = 1.80 Hz, H 2), 4.27 (s, 5H, H7), 4.13 (pt, 2H, 3JHH = 1.80 Hz, H 3), 2.59 ppm (s, 3H, H6). 13C{1H} NMR (100 MHz, CD2Cl2, 298 K): [cis-3] δ = 205.9 (C5), 94.9 (C1), 69.9 (C7), 67.1 (C3), 66.0 (C2), 30.3 ppm (C6); [trans- 3] δ = 200.4 (C5), 95.8 (C1), 70.2 (C7), 66.0 (C3), 64.8 (C2), 36.2 ppm (C6). MS(FD): m/z (%) = 259.9 (100) [M]+. IR(KBr): ?̃? = 3261 (w, NH), 1548 (b, CS(I)), 1378 (b, CS(II)) cm−1. IR(CH2Cl2): ?̃? = 3383 (m, NH), 1502 (m, CS(I)), 1367 (m, CS(II)) cm−1. UV/vis (CH2Cl2): λmax(ε) = 348 (1890), 446 nm (365 M −1 cm−1). CV (CH2Cl2, vs. FcH/FcH+): E½ = –0.020 V (rev.), Ep = 0.345 V (irrev.). Anal. calcd for C12H13FeNS (259.1): C 55.62, H 5.06, N 5.40, S 12.37; Found: C 55.29, H 4.79, N 5.38, S 12.37. Synthesis of 4. Amide ester 2 (100 mg, 0.33 mmol) and Lawesson’s reagent (66 mg, 0.17 mmol) were suspended in toluene (70 mL) and heated to 80 °C for 2.5 h. The organic phase was washed with water, saturated aqueous sodium bicarbonate solution, and brine. The organic phase 75 Results and Discussion was dried over MgSO4 and the solvent was removed under reduced pressure. After column chromatography (SiO2, 3  12 cm, ethylacetate / petroleum ether (b.p. 40 – 60 °C), (1:1)) 4 was obtained as orange-red solid (56 mg, 0.18 mmol, 53%). Rf(SiO2, ethylacetate / petroleum ether (b.p. 40 – 60°C) 1:1) = 0.68. M.p. 105.5 °C. 1H-NMR (400 MHz, CD2Cl2, 298 K): [cis-4] δ = 8.86 (s, 1H, NH4), 4.88 (pt, 2H, 3JHH = 1.97 Hz, H 8), 4.49 (pt, 2H, 3JHH = 1.97 Hz, H7), 4.37 (pt, 2H, 3J = 1.96 Hz, H2HH ), 4.21 (pt, 2H, 3JHH = 1.96 Hz, H 3), 3.86 (s, 3H, H11), 2.36 ppm (s, 3H, H6); [trans-4] δ = 8.49 (s, 1H, NH4), 4.82 (pt, 2H, 3JHH = 1.95 Hz, H8), 4.73 (pt, 2H, 3J 2HH = 1.98, H ), 4.48 (pt, 2H, 3JHH = 1.95 Hz, H 7), 4.19 (pt, 2H, 3J 3HH = 1.98, H ), 3.80 (s, 3H, H 11), 2.61 ppm (s, 3H, H6). 13C{1H} NMR (100 MHz, CD2Cl2, 298 K): [cis-4] δ = 206.1 (C5), 170.7 (C10), 94.7 (C1/9), 73.3 (C7), 72.1 (C8), 68.7 (C3), 67.5 (C2), 52.4 (C11), 30.3 ppm (C6); [trans-4] δ = 202.4 (C5), 171.6 (C10), 94.8 (C1/9), 73.3 (C7), 71.9 (C8), 68.0 (C3), 67.4 (C2), 52.4 (C11), 35.5 ppm (C6). MS(ESI): m/z (%) = 340.0 (100) [M + Na]+. IR(KBr): ?̃? = 3462 (br, NH), 3288 (m, NH), 1686 (s, CO), 1578 (b, CS(I)), 1475 (b, CS(II)), 1467 (b, CS(II)) cm−1. IR(CH2Cl2): ?̃? = 3379, 3281 (m, NH), 1713 (s, CO), 1695 (sh, CO), 1523 (m, CS(I)), 1508 (m, CS(II)), 1370 (m, CS(II)) cm−1. UV/vis (CH2Cl2): λmax(ε) = 359 (sh, 3065), 444 nm (680 M −1 cm−1). CV (CH2Cl2, vs. FcH/FcH +): E½ = 0.205 V (rev.), Ep = 0.510 V (irrev.). Anal. calcd for C14H15FeNO2S (317.2): C 53.01, H 4.77, N 4.42, S 10.11; Found: C 52.83, H 4.62, N 4.35, S 10.10. Synthesis of 7. Diferrocene 5 (300 mg, 0.73 mmol) and Lawesson’s reagent (147 mg, 0.36 mmol) were suspended in toluene (300 mL) and heated to 80 °C for 16 h. The solvent was removed under reduced pressure. After column chromatography (SiO2, 5  30 cm] with ethylacetate / petroleum ether (b.p. 40 – 60 °C) (1:9) 7 was isolated as a red solid (185 mg, 0.43 mmol, 59%). Rf(SiO2, ethylacetate / petroleum ether (b.p. 40 – 60 °C) 1:9) = 0.18. M.p. 230°C (decomp). 1H-NMR (400 MHz, CD2Cl2, 298 K): δ = 8.29 (s, 1H, NH 1), 4.94 (pt, 2H, 3JHH = 1.94, H3), 4.87 (pt, 2H, 3JHH = 1.92, H 8), 4.49 (pt, 2H, 3JHH = 1.92, H 9), 4.27 (s, 5H, H5), 4.24 (s, 5H, H10), 4.16 ppm (pt, 2H, 3J 4 13 1HH = 1.94, H ) C{ H} NMR (100 MHz, CD2Cl2, 298 K): δ = 198.2 (C6), 96.4 (C2), 85.8 (C7), 71.9 (C9), 71.3 (C10), 69.9 (C5), 69.1 (C8), 65.9 (C4), 64.8 ppm (C3). MS(FD): m/z (%) = 429.3 (100) [M]+. IR(KBr): ?̃? = 3270 (w, NH), 1542 (m, CS(I)), 1367 (b, CS(II)) cm−1. IR(CH2Cl2): ?̃? = 3397 (m, NH), 1537 (m, CS(I)), 1363 (m, CS(II)) cm−1. UV/vis (CH2Cl2): λmax(ε) = 375 (2560), 470 nm (1560 M −1 cm−1). CV (CH2Cl2, vs. FcH/FcH +): E½ = −0.065 (rev.), 0.240 (rev.) V, Ep = 0.345 V (irrev Anal. 76 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry calcd for C21H19Fe2NS (429.1): C 58.77, H 4.46, N 3.26, S 7.47; Found: C 57.90, H 4.33, N 2.98, S 7.53. Synthesis of 8. Diamide 6 (250 mg, 0.53 mmol) and Lawesson’s reagent (258 mg, 0.64 mmol) were suspended in toluene (300 mL) and heated to 80 °C for 1.5 h. The solvent was removed reduced pressure. After column chromatography (SiO2, 3  15 cm) with CH2Cl2 / petroleum ether (b.p. 40 – 60 °C) (8/2 to pure CH2Cl2) 8 was obtained as a red solid (108 mg, 0.22 mmol, 40%). Rf(SiO2, CH2Cl2 = 0.21. M.p. 175°C dec. 1H-NMR (400 MHz, CD2Cl2, 298 K). δ = 9.57 (s, 1H, NH 13), 8.45 (s, 1H, NH1), 4.96 (pt, 2H, 3J 3HH = 1.96, H ), 4.90 (pt, 2H, 3J 8HH = 1.98, H ), 4.61 (pt, 2H, 3JHH = 1.98, H 11), 4.59 (pt, 2H, 3JHH = 1.98, H9), 4.26 (s, 5H, H5), 4.23 (pt, 2H, 3J 12HH = 1.98, H ), 4.19 (pt, 2H, 3J 4HH = 1.96, H ), 2.57 ppm (3H, H15) 13C{1H} NMR (100 MHz, CD Cl , 298 K): δ = 202.6 (C142 2 ), 196.1 (C 6), 95.7 (C2), 93.6 (C10), 87.0 (C7), 73.1 (C9), 70.1 (C8), 69.9 (C5), 68.8 (C11), 68.1 (C12), 66.1 (C4), 64.8 (C3), 34.4 ppm (C15). MS(FD): m/z (%) = 502.3 (100) [M]+. IR(KBr): ν ̃ = 3375, 3233 (w, NH), 1552 (b, CS(I)), 1364 (b, CS(II)), 1350 (b, CS(II)) cm−1. IR(CH2Cl2): ?̃? = 3384, 3165 (m, NH), , 1537 (m, CS(I)), 1504 (m, CS(I)), 1367 (m, CS(II)) cm−1. UV/vis (CH2Cl2): λmax(ε) = 384 (3336), 480 nm (2141 M −1 cm−1); CV (CH2Cl2, vs. FcH/FcH +): E½ = −0.035, +0.205 V (rev.), Ep = 0.655 V (irrev.); Anal. calcd for C23H22Fe2N2S2 (502.3): C 55.00, H 4.41, N 5.58, S 12.77; Found: C 54.20, H 4.14, N 5.03, S 12.41. 3.1.6 Associated Content Supporting Information Figures, tables, CIF and .xyz files giving UV/vis spectra of 1 and 3 in CH2Cl2, UV/vis spectra of 2 and 4 in CH2Cl2; UV/vis spectra of 5 and 7 in CH2Cl2; UV/vis spectra of 6 and 8 in CH Cl ; 1H12 2 H NOESY of 3 in CD2Cl2 at room temperature, DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated (cis-3)2, DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated cis- 7 and trans-7, 1H1H NOESY of 7 in CD2Cl2 at room temperature, DFT (B3LYP, LANLDZ, PCM CH2Cl2) calculated trans-3n+ (n = 0–2), Gaussian band shape analysis of IVCT bands of 7+, Gaussian band shape analysis of IVCT bands of 8+, 1H NMR spectra of 8 upon titration with iodine in CD2Cl2; EPR spectrum of 3 in THF/CH2Cl2 (1/4) and simulation, 77 Results and Discussion EPR spectrum of 4 in THF/CH2Cl2 (1/4) and simulation, Cartesian coordinates of optimized geometries. This material is available free of charge via the Internet at http://pubs.acs.org. Author Information Corresponding Author. *E-mail: katja.heinze@uni-mainz.de Notes The authors declare no competing financial interest. 3.1.7 Acknowledgement We are grateful to Dipl.-Chem. Andreas Neidlinger and Dipl.-Chem. Anica Wünsche von Leupoldt for helpful discussions of the EPR spectra, and Nathalie Pilger and Katharina Welter for preparative assistance. 3.1.8 References [1] a) Huang, Y.; Jahreis, G.; Fischer, G.; Lücke, C. Chem. Eur. J. 2012, 18, 9841– 9848; b) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456–463. [2] a) Lee, H.-J.; Choi, Y.-S.; Lee, K.-B.; Park, J.; Yoon, C.-J. J. Phys. Chem. A 2002, 106, 7010–7017; b) Alemán, C. J. Phys. Chem. A 2001, 105, 6717–6723; c) Artis, D. R.; Lipton, M. A. J. Am. Chem. Soc. 1998, 120, 12200–12206. [3] Beer, P. D.; Graydon, A. R.; Johnson, A. O. M.; Smith, D. K. Inorg. Chem. 1997, 36, 2112–2118. [4] Hall, C. D.; Danks, I. P.; Sharpe, N. W. J. Organomet. Chem. 1990, 390, 227–235. [5] a) Kloss, F.; Pidot, S.; Goerls, H.; Friedrich, T.; Hertweck, C. Angew. Chem. 2013, 125, 10945–10948; Angew. Chem. Int. Ed. 2013, 52, 10745–10748; b) Behling, L. A.; Hartsel, S. C.; Lewis, D. E.; DiSpirito, A. A.; Choi, D. W.; Masterson, L. R.; Veglia, G.; Gallagher, W. H. J. Am. Chem. Soc. 2008, 130, 12604–12605. 78 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry [6] Mock, W. L.; Chen, J.-T.; Trang, J. W. Biochem. Biophys. Res. Commun. 1981, 102, 389–396. [7] a) Miwa, H. J.; Patel, A. K.; Vivatrat, N.; Popek, S. M.; Meyer, A. M. Org. Lett. 2001, 3, 3373–3375; b) Grabarse, W.; Mahlert, F.; Shima, S.; Thauer, R. K.; Ermler, U. J. Mol. Biol. 2000, 303, 329–344. [8] a) Patra, M.; Hess, J.; Konatschnig, S.; Spingler, B.; Gasser, G. Organometallics 2013, 32, 6098–6105; b) Sandoval-Chávez, C.; López-Cortéz, J.; Gutiérrez- Hernández, A. I.; Ortega-Alfaro, M. C.; Toscano, A.; Alvarez-Toledano, C. J. Organomet. Chem. 2009, 694, 3692–3700; c) Plażuk, D.; Zakrzewski, J.; Rybarczyk-Pirek, A.; Domogała, S. J. Organomet. Chem. 2005, 690, 4302–4308. [9] Katritzky, A. R.; Moutou, J.-L.; Yang, Z. Synthesis 1995, 1497–1505. [10] a) Heinze, K.; Siebler, D. Z. Anorg. Allg. Chem. 2007, 633, 2223–2233; b) Siebler, D.; Linseis, M.; Gasi, T.; Carrella, L. M.; Winter, R. F.; Förster, C.; Heinze, K. Chem. Eur. J. 2011, 17, 4540-4551; c) Siebler, D.; Förster, C.; Heinze, K. Dalton Trans. 2011, 40, 3558–3575; d) Heinze, K.; Hüttinger, K.; Siebler, D. In Mode- ling of Molecular Properties; Comba, P., Ed.; Wiley-VCH Weinheim, Germany, 2011; pp. 325–346. [11] a) Robin, M. B.; Day, P. Adv. Inorg. Chem. Radiochem. 1967, 10, 247–422; b) Brunschwig, B. S.; Sutin, N. Coord. Chem. Rev. 1999, 187, 233–254. [12] a) Lo, W-S.; Hu, W.-P. Lo, H.-P.; Chen, C.-Y.; Kao, C.-L.; Vandavasi, J. K.; Wang, J.-J. Org. Lett. 2010, 12, 5570–5572; b) Mollin, J.; Kasparek, F.; Odlerova, Z.; Sindelar, Z. Chem. Papers 1986, 40, 239–246. [13] a) Imrie, C.; Cook, L.; Levendis, D. C. J. Organomet. Chem. 2001, 637–639, 266– 275; b) Katada, T.; Nishida, M.; Kato, S.; Mizuta, M. J. Organomet. Chem. 1977, 129, 189–196. [14] a) Heinze, K.; Schlenker, M. Eur. J. Inorg. Chem. 2004, 2974–2988; b) Okamura, T.; Sakauye, K.; Ueyama, N.; Nakamura, A. Inorg. Chem. 1998, 37, 6731–6736; c) Barišić, L.; Rapić, V.;Kovač, V. Croat. Chem. Acta 2002, 75, 199–210. [15] a) Jesberger, M.; Davis, T. P.; Baner, L. Synthesis 2003, 1929–1958; b) Murai, T. Top. Curr. Chem. 2005, 251, 247–272. [16] Wiberg, K. B.; Rush, D. J. J. Am. Chem. Soc. 2001, 123, 2038–2046. 79 Results and Discussion [17] Okamura, T.-a.; Sakauye, K.; Doi, M.; Yamamoto, H.; Ueyama, N.; Nakamura, A. Bull. Chem. Soc. Jpn. 2005, 78, 1270–1278. [18] Lu, S.; Strelets, V. V.; Ryan, M. F.; Pietro, W. J.; Lever, A. B. P. Inorg. Chem. 1996, 35, 1013–1023. [19] Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910. [20] a) Adams, J. J.; Curnow, O. W.; Huttner, G.; Smail, S. J.; Turnbull, M. M. J. Organomet. Chem. 1999, 577, 44–57; b) Musgrave, R. A.; Russell, A. D.; Manners, I. Organometallics 2013, 32, 5654–5667. [21] a) Siebler, D.; Förster, C.; Gasi, T.; Heinze, K. Organometallics 2011, 30, 313– 327; b) Neidlinger, A.; Ksenofontov, V.; Heinze, K. Organometallics 2013, 32, 5955–5965. [22] Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42–55. [23] Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, 2009, Revision A.02; Gaussian, Inc., Wallingford, CT. [24] SMART Data Collection and SAINT-Plus Data Processing Software for the SMART System (various versions); Bruker Analytical X-Ray Instruments, Inc.: Madison, WI, 2000. [25] Blessing, B. Acta Cryst. 1995, A51, 33–38. [26] Sheldrick, G. Ma. SHELXTL, Version 5.1; Bruker AXS: Madison, WI, 1998. 80 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry [27] Sheldrick, G. M. SHELXL-97; University of Göttingen, Göttingen, Germany, 1997. 81 Results and Discussion 82 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene 3.2 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Andreas Neidlinger,‡ Torben Kienz,‡ and Katja Heinze* Organometallics 2015, 34, 5310–5320. 83 Results and Discussion Supporting information for this article (without Cartesian coordinates from DFT calculations) is found at pp. 163. For full supporting information refer to: http://pubs.acs.org/doi/suppl/10.1021/om500052k Adapted with permission from A. Neidlinger, T. Kienz, K. Heinze, Organometallics 2015, 34, 5310–5320. “This is an unofficial adaptation of an article that appeared in an ACS publication. ACS has not endorsed the content of this adaptation or the context of its use. Copyright. 2015 American Chemical Society.” 84 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene 3.2.1 Abstract In contrast to metal centered 17 valence electron radicals, such as [Mn(CO)5]•, ferrocenium ions [Fe(C H + + + +5 5)2] (1 ), [Fe(C5Me5)2] (2 ), [Fe(C5H5)(C5H + + 4Et)] (3 ), [Fe(C5H5)(C5H4NHC(O)Me)] + (4+) and [Fe(C5H5)(C + 5H4NHC(S)Me)] (5+) do not add to nitrosobenzene PhNO to give metal-coordinated stable nitroxyl radicals. In the presence of the strong and oxidatively stable phosphazene base tert-butylimino-tris(dimethylamino)- phosphorane the quite acidic ferrocenium ions 1+ − 5+ are deprotonated to give a pool of transient and persistent radicals with different deprotonation sites [1−Hx]• − [5−Hx]•. One rather persistent iron-centered radical [4−HN]•, deprotonated at the nitrogen atom, has been detected by rapid-freeze EPR spectroscopy at 77 K. This iron-centered radical [4−HN]• is also inert toward PhNO. The transient carbon-centered radicals [1−Hx]• − [5−Hx]• appear to rapidly abstract hydrogen atoms from the adjacent base or the solvent to regenerate the corresponding ferrocenes 1 − 5. These transient radicals are only present in trace amounts (< 1%). However, some of the transient carbon-centered radicals in the radical pool can be trapped by 1 – 1.2 equiv of PhNO, even at room temperature. The corresponding resulting stable nitroxyl radicals [6]• − [10]• were studied by EPR spectroscopy at room temperature and at 77 K. The hyperfine coupling pattern to protons close to the spin center allows to assign the site of PhNO attack in radicals [6]• − [10]•, namely at the C5H5 ring in [6]•, [9Cp]• and [10Cp]•, at a methyl group in [7]• and at the methylene group in [81]•. These studies give a deeper insight into the stability and reactivity of radicals derived from ferrocene derivatives which might also be relevant for the biological activity of high-potent antitumor and antimalaria ferrocene-based drugs and prodrugs such as ferrocifen or ferroquine. 85 Results and Discussion 3.2.2 Introduction Ferrocene and its derivatives have been found useful for a variety of applications in fundamental research, especially electron transfer, in biology and catalysis, as well as in material and pharmaceutical science.1–5 While many of these applications rely on the reversible oxidation chemistry of the ferrocenyl moiety, investigation of the reactivity and possible degradation pathways of the resulting ferrocenium cations is worthwhile. For instance, ferrocene itself displays no antitumor activity, while ferrocenium derivatives show in vitro cytotoxicity due to oxidative damage of DNA.6 This has been further exploited using amino ferrocene based selective prodrugs A by Mohkir et al. (Scheme 1). In tumor cells, ferrocenes A are oxidized to the corresponding ferrocenium ions A+. The ferrocenium cations A+, as well as their degradation products, iron(II) ions, appear to catalyze the generation of reactive oxygen species. This increases the oxidative stress in cancer cells and finally leads to apoptosis.7,8 This concept has been further explored with a variety of amino ferrocene and diamino ferrocene based prodrugs (B; Scheme 1a). The degradation products of B/B+ again act as catalysts for generating reactive oxygen species.9 In fact, first in vivo experiments show promising results for further pharmaceutical applications.10 Scheme 1. a) Aminoferrocene based prodrugs A, B; b) chloroquine C and ferroquine D; c) tamoxifen E, ferrocifen F and ferrociphenol G. 86 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene The effectiveness of the antimalarial drug chloroquine C has been drastically improved by incorporation of a ferrocenyl moiety giving the potent drug ferroquine D (Scheme 1b). This adds the generation of reactive oxygen species as a further mode of action to the drug and takes effect in all stages of the live cycle of the parasites, thus inhibiting merozoites reinvasion.11–13 While the generation of reactive oxygen species is a valuable mechanism for amino ferrocene based prodrugs A and B and ferroquine D, other antitumor agents appear to follow a different mode of action. The ferrocene derivatives of tamoxifen E, ferrocifen F and ferrociphenol G (Scheme 1c) show high levels of cytotoxicity against breast cancer cells. Ferrocifen F is even more cytotoxic than cis-platin.14–18 Both ferrocifen F and hydroxyferrocifen G cause less oxidative stress compared to tamoxifen E.19,20 Yet, in cells treated with ferrocifen derivatives a higher rate of senescence has been found. As this is unlikely to be related to oxidative stress a further mode of action is assumed showing the versatility of ferrocenium ions. Indeed, carbon-centered radicals have been proposed to be generated after oxidation and deprotonation of ferrocifens F.16 Various experiments have been reported which focus on the ability of ferrocene to form radicals itself. First hints to the existence of ferrocenyl radicals were obtained by photo- and thermolysis of ferrocenyl azide by Sutherland et al.21 -Irradiation of methylferrocene lead to the ferrocenyl methylradical.22 Later, it was found that ferrocenyl substituted radicals are stabilized due to spin delocalization to the iron atom.23 This can even be exploited to generate polymetallocenylenes24 or in stereoselective pinacol coupling reactions.25 The radical reactivity of ferrocenium ions has been demonstrated by the oxidation of ferrocenophanes with silver salts in the presence of sodium methoxide. As reported by Hisatome et al., this procedure results in the intermediate formation of a carbon-centered radical in the aliphatic bridge (Scheme 2a). Ferrocenylmethylium cations, as shown by Ashkenazi and Cais, react as biradicals with nitrosobenzene PhNO as spin trap (Scheme 2b).26 87 Results and Discussion Scheme 2. a) Oxidation and deprotonation of a [4]ferrocenophane; b) spin trapping of a ferrocenylmethylium ion with nitrosobenzene and c) spin trapping of metal centered 17 valence electron radicals. Obviously, spin trapping is a valuable tool for the investigation of transient radicals. Indeed, various manganese centered 17 valence electron radicals prepared by homolysis of carbonyl complexes could be spin trapped with nitrosodurene as spin trapping agent by Hudson and Lappert27 as well as with 2,4,6-tri(tert-butyl)nitrosobenzene by Simpson et al.28 and with the aid of nitroso-tert-butane by Benner and Balch.29 Later on, Re, Co, Fe, and Mo centered 17 valence electron radicals could be trapped with 2,3,5,6-tetramethyl-1- nitrosobenzene as spin trapping agent and identified by EPR spectroscopy (Scheme 2c).30 With the versatile and not yet fully understood reaction pathways of ferrocenium ions in mind, we were interested to generate and trap radical species of different ferrocenium ions under neutral and strongly basic conditions using the spin trapping technique. The starting iron(III) complexes ferrocenium (1+), decamethylferrocenium (2+), ethylferrocenium (3+), N-acetylaminoferrocenium (4+), and N-thioacetylaminoferrocenium (5+) were prepared by oxidation of the corresponding ferrocenes 1 – 5 with silver hexafluoroantimonate. Spin trapping of radicals was attempted with nitrosobenzene (PhNO) in the absence and presence of the strong non-nucleophilic, non-coordinating, and oxidatively stable phosphazene base P t1 Bu (tert-butylimino-tris(dimethylamino)- phosphorane (Scheme 3). In the presence of P t1 Bu, 1+ – 5+ are expected to yield the corresponding radicals or radical pools [1−H]• – [5−H]•. In order to address the question as to wether radicals are formed at all and to identify the site of the generated radicals (carbon, iron, nitrogen centered), the conceivable spin trapped nitroxide radical adducts of 1+ – 5+ and [1−H]• – [5−H]• were probed by EPR spectroscopy (Scheme 3). 88 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene While ferrocenyl nitroxide radicals have been prepared via autoxidation of ferrocenyl hydrazine and investigated by EPR spectroscopy by Forrester and Hepburn, the present approach offers a different and more general access to ferrocenyl nitroxyl radicals.31a Elschenbroich has reported the electrochemical generation and EPR spectroscopic study of isoelectronic radical anions of ferrocenyl arylketones.31b Similarly, ferrocenoylsilanes have been reduced to the corresponding radical anions by Grignard reagents.31c Scheme 3. Spin trapping reactions of ferrocenium derivatives in the absence and presence of P t1 Bu for a) 1+, b) 2+, c) 3+, and d) 4+ and 5+ investigated in this study. Relevant atom numbering is given. Possible acidic hydrogen atoms are marked in red. Reactions of ferrocenium 1+ with nitrogen bases (pyrazolide, 3,5-dimethyl pyrazolide, imidazolide, and benzotriazolide) have been reported in the literature.32 Yet, these bases are oxidized by the ferrocenium cation 1+ to the corresponding azolide radicals while 1+ is reduced to 1. The azolide radicals are suggested to attack 1+ to give N-ferrocenyl azoles 89 Results and Discussion after proton loss (Scheme 4). No ferrocenyl radicals were reported. Clearly, such a reactivity is not expected with the base P t1 Bu employed in this study. Scheme 4. Suggested mechanism of the formation of N-ferrocenylpyrazole from ferrocenium radical cations and pyrazolide anions.32 3.2.3 Results and Discussion Spin Trapping of Ferrocenium Ions. The oxidations of ferrocenes 1 – 5 to their respective cations 1+ − 5+ (Scheme 3) are performed in CH2Cl2 under inert conditions using one equivalent of silver hexafluoroantimonate AgSbF6 as oxidant as its oxidation potential [E1/2(CH2Cl2) = 650 mV vs. FcH/FcH+]33a is sufficient for this purpose [E1/2(CH2Cl2) = 0, –480, –55, –50, –20 mV for 1 – 5, respectively]33. Furthermore, the coproduct silver is easily removed by filtration. The presence of 1 – 1.2 equivalents of PhNO does not influence the EPR spectra of the ferrocenium ions. Indeed, 1+ is EPR silent both at 298 K and at 77 K as its EPR spectrum has been observed only below 20 K due to fast spin-lattice relaxation.34 No EPR resonances are observed in the presence of PhNO as well. Obviously, the iron-centered 17 valence electron radical 1+ does not add to PhNO to give the conceivable nitroxide adduct [1- PhNO]+, in contrast to what has been reported for [Mn(CO) ]•5 radicals for example (Scheme 2c). DFT calculations for 1+, PhNO, and [1-PhNO]+ account for this lack of reactivity as the formation of [1-PhNO]+ is endergonic by 49 kJ mol−1 (Supporting Information, Figure S1). Similarly, 4+ gives a nearly axial EPR resonance at 77 K in frozen solution35 which remains unchanged in the presence of PhNO. At room temperature 4+ and the 4+/PhNO mixture are EPR-silent. Obviously, iron-centered ferrocenium radical ions are unable to 90 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene react with PhNO to give the nitroxyl radicals. This behavior contrasts the reactivity of other organometallic 17 valence electron complexes (Scheme 2c).27–30 Spin Trapping of Ferrocenium Ions in the Presence of a Base. The ferrocenium ions 1+ − 5+ were treated with one equivalent of the non-nucleophilic and noncoordinating phosphazene base P t1 Bu (tert-butylimino-tris(dimethylamino)- phosphorane, (pKa (MeCN) = 26.98) 36 as a proton acceptor. To ensure that the ferrocenium salts 1+ − 5+ are unable to oxidize the employed P t1 Bu base, its redox potential has been determined by cyclic voltammetry. Indeed, P t1 Bu is irreversibly oxidized at Ep = 400 mV vs. FcH/FcH+ ([nBu4N][B(C6F5)4]/CH2Cl2), significantly higher than the redox potentials of the 1/1+ – 5/5+ redox couples (Supporting Information, Figure S2). A mechanistic scenario as shown in Scheme 4 for heterocyclic nitrogen bases is possible in principle due to the irreversible nature of the P t1 Bu / P t 1 Bu + oxidation, although it is not very likely. Hence, a simple deprotonation of 1+ − 5+ to give the radicals and radical pools [1−H]• – [5−H]• should be achieved (Scheme 3). As illustrated in Scheme 3, deprotonation of 1+ and 2+ should yield the radicals [1−H]• and [2−H]•, respectively. As 3+, 4+ and 5+ possess five chemically different protons, which might be abstracted, radical pools consisting of up to five radical species might be present. The different radicals will be designated by the location of the abstracted proton as [3−HCp]•, [3−H]•, [3−H]•, [3−H1]• and [3−H2]• for the radical pool [3−H]•, [4−HCp]•, [4−H]•, [4−H]•, [4−HN]• and [4−HMe]• for the radical pool [4−H]• and [5−HCp]•, [5−H]•, [5−H]•, [5−HN]• and [5−HMe]• for the radical pool [5−H]• (Scheme 3). Trapping of these radicals or some of these radicals by the spin trapping technique using nitrosobenzene PhNO is attempted. We will start with the simple ferrocene and decamethylferrocene derivatives [1−H]• and [2−H]• and then describe the more diverse reactivity of [3−H]• − [5−H]•. DFT calculations were employed both for the radical species [1−H]• − [5−H]• as well as for the conceivable corresponding PhNO adducts [6]• − [10]•. For the 1+/PhNO/P t1 Bu mixture an EPR triplet resonance at giso = 2.0063 with nitrogen hyperfine coupling (hfc) A(14N) = 11.1 G was recorded (Figure 1), similar to typical nitroxyl radicals.37–40 Furthermore, hyperfine couplings to hydrogen nuclei of the phenyl moiety and the substituted cyclopentadienyl ring are extracted from the simulation of the EPR resonance (Table 1). Ho, Hm, and Hp denote the ortho, meta, and para protons of the phenyl substituent, respectively, while H and H are the alpha and beta protons of the substituted Cp ring. The observed coupling pattern allows a clear assignment to the 91 Results and Discussion ferrocenyl phenyl nitroxide radical [6]• (N-oxyl-N-phenyl-ferrocenylamine). The hfc to 14N in [6]• is close to the observed hfc in the reported nitroxide radical [Fc-N(O)-tBu]• [A(14N) = 11.75 G].31a Some phenyl hydrogen hfc’s of [6]• are similar to those obtained for the isoelectronic radical anion [Fc-C(O)-Ph]•–.31b At 77 K a slightly anisotropic signal at gav ≈ 2.0071 with a large nitrogen hyperfine coupling of A( 14N) = 26 G in the high field region is recorded (Table 2; Supporting Information, Figure S4a). Hyperfine couplings to hydrogen atoms are not resolved in the frozen solution spectrum, and hence a larger linewidth was applied in the simulation instead. Obviously, the increased acidity of the positively charged ferrocenium ion 1+ 35,41 allows C-H deprotonation by P t1 Bu. The initially formed C-deprotonated zwitterionic ferrocenium species corresponds to an electronically excited state and relaxes to the carbon centered radical [1−H]• by internal electron transfer (Supporting Information, Figure S3a; Mulliken spin density at C 0.787; Mulliken spin density at Fe 0.212). The C-centered radical [1−H]• attacks the nitrogen atom of PhNO leading to the spin trapped nitroxyl radical N-oxyl-N- phenyl-ferrocenylamine [6]• (Figure 2a). DFT calculations revealed that the nitroxyl radical [6]• is lower in energy than the starting materials [1−H]• and PhNO by 208 kJ mol−1 (Supporting Information, Figure S3a), explaining the facile formation of [6]•. Figure 1. X-band EPR spectrum (top) and simulated spectrum (bottom) of [6]• (20 mM 1 in CH2Cl2) at the following experimental parameters: temperature = 298 K, field = 3358.98 G, sweep = 298.72 G, sweep time = 120 s, modulation = 1000 mG, MW attenuation = 9 db. 92 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Table 1. EPR parameters obtained by simulation of experimental spectra (298 K). radical giso A( 14N) A(1Ho) / A(1Hm) A(1Hp) A(1Hα) / A(1Hβ) / A(1H1) / Gauss pp Lorentz / G G (2 ) / G / G G G G linewidth pp (2 ) / MHz linewidth / MHz [6]• 2.0063 11.10 2.90 0.80 2.70 3.80 (2 ) 0.60 (2 ) 0.15 0.08 [7]• 2.0068 11.07 2.77 0.95 2.64 8.64 (2 ) 0.02 0.015 [81]• 2.0072 11.16 2.77 1.00 2.65 2.10 0.035 0.065 [9Cp]• 2.0072 10.90 2.77 0.90 2.64 1.20 / 1.10 0.70 (2 ) 0.06 0.02 [10Cp]• 2.0068 11.09 2.77 0.91 2.64 2.55 / 2.10 1.82 / 1.08 0.085 0.005 Table 2. EPR parameters obtained by simulation of experimental spectra (77 K). radical (mixture) g1,2,3 A( 14N) / G fraction / Gauss pp Lorentz pp % linewidth / linewidth / MHz MHz [6]• 2.0104, 2.0068, 2.0040 4.0, 4.0, 26.0 0.90 0.30 [7]• 2.0094, 2.0067, 2.0048 4.0, 4.0, 26.0 1.50 0.60 [81]• 2.0095, 2.0073, 2.0045 3.0, 3.0, 27.5 0.90 0.40 [9Cp]• 2.0105, 2.0060, 2.0045 3.5, 3.5, 28.0 17 0.30 0.40 [4−HN]• N/Aa, 1.9620, 1.9450 N/A 83 0.50 0.20 [4−Hx]•(x = α,β,Cp,Me) 2.0095, 2.0065, 2.0030 0.8 0.30 0.2 [4−HN]• N/Aa, 1.9650, 1.9400 99.2 0.1 0.1 [10Cp]• 2.0105, 2.0060, 2.0020 3.5, 3.5, 28.0 5 0.60 0.60 [11a]• 2.3100, 2.0695, 1.9990 48 0.50 0.50 [11b]• 2.2250, 2.0565, 2.0095 47 0.50 0.50 a) Too broad to be observed. 93 Results and Discussion Figure 2. DFT optimized geometries with spin densities (0.01 a.u. isosurface value) in CH2Cl2 continuum solvent as well as Lewis structures of nitroxide radicals a) [6]•, b) [7]•, and c) [81]•. Successful deprotonation of 1+ with P t1 Bu was further evidenced by cyclic voltammetry of 1 in the absence and presence of the base using tetra(n-butyl)ammonium tetrakis(pentafluorophenyl)borate ([nBu4N][B(C6F5)4]) as weakly coordinating electrolyte.42,43 Expectedly, ferrocene 1 shows a reversible one-electron redox process at a potential of E +1/2 = 0 V vs. FcH/FcH per definition (Supporting Information, Figure S5a). Addition of a stoichiometric amount of P t1 Bu renders this oxidation irreversible, due to the 94 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene deprotonation of 1+ to [1−H]• and follow up reactions of the highly reactive C-centered radical [1−H]• (Supporting Information, Figure S5b). As the oxidation process occurs at essentially the same potential (against the Ag/AgNO3 reference electrode) in the presence of P t1 Bu, deprotonation of 1 to [1−H]– by P t 1 Bu prior to oxidation is unlikely. Deprotonation of ferrocene 1, however, is achieved using alkyl lithium bases as a well- known starting point for the rich ferrocene substitution chemistry.44–47 Despite the lack of aromatic C-H atoms, decamethylferrocenium 2+ reacts with PhNO as well in the presence of P t1 Bu (Scheme 3b). Obviously, the methyl groups become sufficiently acidic upon oxidation of 2 to 2+ due to the increased electron deficiency. Hence, the C-centered radical [2−H]• is generated from 2+ by deprotonation of a CH3 group (Supporting Information, Figure S3b, Mulliken spin density at C 0.650; Mulliken spin density at Fe 0.513). Similar to [1−H]•, [2−H]• is trapped by PhNO resulting in the formation of the nitroxide radical [7]• 1-[(N-oxyl-N-phenylamino)methyl]- 1’,2,2’,3,3’,4,4’,5,5’-nonamethylferrocene (Figure 2b). The driving force for the formation of [7]• from the starting materials [2−H]• and PhNO (Supporting Information, Figure S3b) is calculated as 80 kJ mol−1 by DFT methods. At 298 K [7]• gives an EPR resonance with parameters summarized in Table 1 (Figure 3). The nitrogen and hydrogen hyperfine couplings for the PhNO moiety of [7]• are very similar to those of [6]•. The two large hydrogen hyperfine couplings of [7]• (A(1H) = 8.64 G), obtained by simulation, are assigned to the protons of the methylene group. The large hfc’s nicely fit to the reported hfc’s for the ferrocenylmethyl radical prepared from methylferrocene by -irradiation (A(1H) = 14.71 G).22 At 77 K in frozen solution the nitroxide radical [7]• displays a slightly anisotropic resonance similar to [6]• (Supporting Information, Figure S4b) with one large nitrogen hyperfine coupling component (Table 2). Again, hydrogen hyperfine couplings are not resolved under these conditions. In contrast to 1 and 2, ethylferrocene 3 has five chemically different protons resulting in five conceivable C-centered radicals [3−HCp]•, [3−H]•, [3−H]•, [3−H1]•, and [3−H2]• (Scheme 3c). Consequently, five distinct nitroxyl radicals can in principle be obtained from this radical pool by reaction with PhNO, namely [8Cp]•, [8]•, [8]•, [81]•, and [82]•. Despite the possible mixture of products, the room temperature EPR spectrum of [8]• is well resolved suggesting the presence of only a single nitroxide radical. The EPR pattern of [8]• is well reproduced by assuming the typical 14N and 1H hfc’s of the PhNO unit in addition 95 Results and Discussion to a further hfc to a single hydrogen atom (Table 1). This perfectly fits to the [81]• nitroxide radical with the PhNO substituent attached to the C1 atom of the ethyl substituent. For all other nitroxide radicals [8Cp]•, [8]•, [8]•, and [82]• more than one chemically different hydrogen atom would have been expected to display hfc’s. At 77 K a similar spectrum as for [6]• is recorded for [81]• (Table 2; Supporting Information, Figure S4c). Figure 3. X-band EPR spectrum (top) and simulated spectrum (bottom) of [7]• (25 mM 2 in CH2Cl2) at the following experimental parameters: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 250 mG, MW attenuation = 5 db. The presence of only a single product [81]• derived from [3−H1]• is straightforwardly explained by the pronounced stability of the secondary C-centered radical [3−H1]• by more than 60 kJ mol−1 with respect to all other radicals [3−HCp]•, [3−H]•, [3−H]•, and [3−H2]• according to DFT calculations (Supporting Information, Figure S6, Mulliken spin density at C1 0.757; Mulliken spin density at Fe 0.341). However, all five conceivable nitroxide radicals [8Cp]•, [8]•, [8]•, [81]•, and [82]• are quite similar in energy (Supporting Information, Figure S7). The driving force for the formation of [81]• from [3−H1]• and PhNO amounts to 81 kJ mol−1 (Supporting Information, Figure S7). Hence, the product distribution is controlled by the relative stability of the C-centered radical [3−H1]• but not by the relative stability of the product nitroxide radical [81]•. In order to gain insight into the reactivity of NH-containing ferrocenyl compounds, which are substructures of cytotoxic prodrugs A and B and antimalarial drugs D (Scheme 1),7,9–13 we investigated the reactivity of the simple oxo- and thioamides, namely N- 96 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene acetylaminoferrocenium 4+ (Scheme 3d), and N-thioacetylaminoferrocenium 5+ in the presence of the P t1 Bu base. Figure 4. X-band EPR spectrum (top) and simulated spectrum (bottom) of [81]• (25 mM 3 in CH2Cl2) at the following experimental parameters: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 250 mG, MW attenuation = 0 db. Because of the presence of five chemically different protons in 4, five distinct radicals [4−HCp]•, [4−H]•, [4−H]•, [4−HN]• and [4−HMe]• can be conceived in the radical pool [4−H]• (Scheme 3d). The same holds analogously for 5. This translates to the corresponding nitroxide radicals [9Cp]•, [9]•, [9]•, [9N]•, and [9Me]• derived from 4+ and PhNO and [10Cp]•, [10]•, [10]•, [10N]•, and [10Me]• derived from the thioanalogue 5+ and PhNO. Similar to the situation observed for [81]•, the room temperature EPR spectrum of [9]• is rather well resolved, suggesting that only one or two nitroxide radical species are present (Figure 5). The main features of the experimental EPR resonance of [9]• can be simulated by the expected hfc’s to the PhNO moiety (14N, Ho, Hm, Hp, Table 1) and coupling to four protons of 1.2 G (1H), 1.1 G (1H) and 0.70 G (2H). Further, small differences between the simulated and the experimental spectrum are associated to the presence of a second nitroxide radical with rather similar parameters. However, simulation of a full second parameter set would lead to severe over-parametrization of the simulation. Hence, we note, that the hfc’s to ferrocene protons are less well defined for [9]• than for [6]•, [7]•, and [81]•. The presence of four protons close to the radical center suggest the [9Cp]• nitroxide radical derived from [4−HCp]• as the major trapped species (Figure 6). In [9Cp]• the hfc to H is smaller than that in [6]• and furthermore, two slightly different hfc’s to chemically different H atoms are 97 Results and Discussion found for [9Cp]• (Table 1). The larger hfc to H in [6]• is easily traced back to the favorable coplanar orientation of the NO unit with the C H ring (O-N-Cipso 5 4 -C = –23.0°) as compared to the corresponding torsion angle in [9Cp]• (O-N-Cipso-C = –31.4°). The larger twist in [9Cp]• arises from an intramolecular NH(amide)…O(nitroxyl) hydrogen bond (Figure 6). Similar intramolecular hydrogen bonds have been amply observed in ferrocenyl polyamides.49–51 This hydrogen bond furthermore provides a straightforward explanation for the chemically different H/H’ protons. Figure 5. X-band EPR spectrum (top) and simulated spectrum (bottom) of [9Cp]• (25 mM 4 in CH2Cl2) at the following experimental parameters: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 1000 mG, MW attenuation = 10 db. Expectedly, in the radical pool [4−H]• the nitrogen-deprotonated radical [4−HN]• is the most stable one (Supporting Information, Figure S9). However, this radical is essentially iron-centered with a Mulliken spin density at iron of 1.248 with some small contribution of the nitrogen atom (Mulliken spin density at N 0.016). Attack of PhNO at iron is excluded based on the general lack of reactivity of ferrocenium ions towards PhNO. Attack of PhNO at the nitrogen atom of [4−HN]• to give [9N]• is calculated to be disfavored by 37 kJ mol−1 relative to the starting materials [4−HN]• and PhNO (Figure 6). A ring-slipped isomer [4−HN’]• has been calculated as well, yet its spin density is also localized at the iron atom (Mulliken spin density at Fe 1.268), precluding the reaction of [4−HN’]• with PhNO (Figure S9). The calculated ring-slipped structure of [4−HN’]• suggests a viable decomposition 98 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure 6. DFT optimized geometries with spin densities for [9x]• (x = α, β, Cp, N, Me) (0.01 a.u. isosurface value) and energies in CH2Cl2 continuum solvent for [9x]• and [10x]• (x = α, β, Cp, N, Me) as well as Lewis structures. 99 Results and Discussion pathway of radicals [4−H]• releasing the substituted cyclopentadienyl ligand as N-acetyl- 2,4-cyclopentadien-1-imine. Indeed, the formation of di(cyclopentadiene) and di(aminocyclopentadiene) has been observed in the reaction of prodrugs A with H2O2 (Scheme 1).7 In addition to these PhNO-resistant iron-centered radicals [4−HN]• and [4−HN’]• the radicals [4−HCp]•, [4−H]•, [4−H]•, and [4−HMe]• deprotonated at carbon atoms were calculated by DFT (Figure S9). The reactions of [4−HCp]•, [4−H]•, [4−H]•, and [4−HMe]• and PhNO to [9Cp]•, [9]•, [9]•, and [9Me]•, respectively, are thermodynamically feasible (Figure 6). Interestingly, [4−H]• mainly features spin density at the iron center (Mulliken spin density at Fe 1.251) and is hence considered unreactive towards PhNO, so [9]• should not be observed. In [4−HMe]• the spin density is smeared over the nitrogen, oxygen, and CH2 units of the substituent. This spin delocalization reduces the probability of PhNO attack and hence [9Me]• is not particularly favored as well. The remaining two highly reactive radicals [4−HCp]• and [4−H]•, although high in energy, might account for the observed EPR pattern with [9Cp]• derived from [4−HCp]• being the major and [9]• derived from [4−H]• being the minor species based on simple statistical arguments. This interpretation agrees with the EPR spectral data. While the radicals [4−HCp]• and [4−H]• of the radical pool can be trapped by PhNO, [4−HN]• and [4−HN’]• are inert towards PhNO. In order to possibly detect the rather persistent and PhNO-resistant Fe-centered radical [4−HN]• rapid-freeze EPR techniques have been employed. The EPR spectrum recorded at 77 K rapidly after deprotonation of 4+ in the presence of PhNO shows a characteristic broad ferrocenium-based resonance (83 %) in addition to the slightly anisotropic nitrogen-split triplet resonance of a nitroxide radical with one large hfc to 14N (17 %) (Figure 7, Table 2). The latter EPR resonance is very similar to the corresponding resonance of [6]• and is hence safely assigned to already formed nitroxide radicals [9Cp/]•. The broad ferrocenium resonance (g1,2,3 = N/A, 1.9650, 1.9400, Table 2), however, clearly differs from the EPR resonance of the ferrocenium ion 4+ (g1,2,3 = 3.3500, 1.8750, 1.7870).35 Hence, the resonance is assigned to [4−HN]•. This iron-centered radical is unable to react with PhNO itself, but equilibrates with radical species deprotonated at carbon atoms [4−Hx]• (x = , , Cp, Me). After annealing the sample to room temperature for 5 min and refreezing to 77 K, the ferrocenium resonance of [4−HN]• has vanished, and only the resonance of the nitroxide radicals [9Cp/]• remained. 100 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure 7. X-band EPR spectrum (top) and simulated spectrum (bottom) of [9Cp]• / [4−HN]• (25 mM 4 in CH2Cl2) at the following experimental parameters: temperature = 77 K, field = 3346.20 G, sweep = 499.77 G, sweep time = 90 s, modulation = 5000 mG, MW attenuation = 10 db. In the absence of PhNO, rapid-freeze EPR spectroscopy of the 4+/P t1 Bu mixture yields an EPR spectrum displaying the resonance assigned to [4−HN]• (99.2 %) (Figure S8a). This finding supports the above assignment. Furthermore, a weak, but significant slightly anisotropic resonance at g = 2.0 without resolved hyperfine couplings is detected (0.8 %) (Table 2, Figure S8a). This resonance is assigned to traces of the carbon-centered radicals [4−Hx]• (x = , , Cp, Me) of the radical pool [4−H]• (Table 2). At room temperature, this sample exhibits a transient EPR resonance without discernible hfc’s at giso = 2.0078 (Supporting Information, Figure S8b) which is assigned to the C-centered radicals [4−Hx]• (x = , , Cp, Me) as well. The proposed reactivity of 4+ in the presence of a base is summarized in Scheme 5. Deprotonation of 4+ gives the radical pool [4−H]• with [4−HN]• being the most stable and abundant radical (99.2 %). Neither 4+ nor [4−HN]• react with PhNO. Less than one percent of the observed radicals are carbon-centered radicals [4−Hx]• (x = , , Cp, Me). However, two of these, namely [4−HCp]• and [4−H]•, can be trapped by PhNO to give the respective nitroxide radicals [9Cp]• and [9]•. Quantification52 of the nitroxide radical species [9Cp]• and [9]• with respect to external calibration with DPPH reveals that indeed less than 1 % of the original ferrocenium radical 4+ is transformed into the nitroxides [9Cp]• and [9]•. One possible decomposition pathway of the radicals via the ring-slipped isomer [4−HN’]• is proposed on the basis of DFT calculations of [4−HN’]•. 101 Results and Discussion Scheme 5. Suggested radical reactivity of 4+ in the presence of a base and PhNO; EPR data (77 K) of identified intermediates given. Further radical reactivity might be hydrogen atom abstraction from the solvent or the base to give the starting material 4. Indeed, quenching the radical pool [4−H]• by hydrazine hydrate essentially quantitatively recovers ferrocene 4 as shown by 1H NMR spectroscopy. The successful deprotonation of 4+ to the radical pool [4−H]• is furthermore evidenced by cyclic voltammetry. At 298 K in CH2Cl2/[ nBu4N][B(C6F5)4] a reversible one-electron process is observed at E1/2 = −50 mV vs. FcH/FcH + and assigned to the 4/4+ couple (S10a). After addition of stoichiometric amounts of P t1 Bu, this process is replaced by an irreversible oxidation at Ep = −175 mV vs. FcH/FcH + (Figure S10b). The lower potential of the 4/4+ couple in the presence of P t1 Bu is ascribed to coordination of the base to the amide unit of 4 via a hydrogen bond. Oxidation to 4+ acidifies this NH proton, and the proton is transferred to the base giving the radicals [4−H]•, with [4−HN]• being the major product. Reduction of the rather persistent radical [4−HN]• occurs at Ep’ = −890 mV vs. FcH/FcH + 102 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene (Figure S10b). These results reflect the intimate coupling of proton transfer and electron transfer reactions in ferrocenium amides as reported in the literature.35,49–51 In order to investigate the effect of O  S exchange in the amide for the spin trapping experiments in the presence of P t1 Bu, we employed thioamide 5 as sulfur analogue of 4. 51 Similar to 4+, a radical pool [5−HCp]•, [5−H]•, [5−H]•, [5−HN]• and [5−HMe]• is established after deprotonation of 5+. (Scheme 5) Fully paralleling the energies of the corresponding radicals [4−H]• , the most stable one is the iron-centered radical [5−HN]• (Figure S9) while the ring and side chain radicals [5−HCp]•, [5−H]•, [5−H]• and [5−HMe]• are higher in energy. Figure 8. X-band EPR spectrum (top) and simulated spectrum (bottom) of [10Cp]• (25 mM 5 in CH2Cl2) at the following experimental parameters: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 1000 mG, MW attenuation = 10 db. The EPR spectrum obtained from the 5+/PhNO/P t1 Bu mixture at room temperature is very well simulated by a single set of parameters (Figure 8, Table 1). The hfc’s are consistent with the nitroxide radical [10Cp]•. Similar to [9Cp]•, hyperfine coupling to the H atoms is split due to the chemical dissimilarity of H and H‘ in [10Cp]• featuring an intramolecular NH(thioamide)…O(nitroxide) hydrogen bond (Figure 6). According to the DFT calculations this hydrogen bond is shorter in [10Cp]• (NH…O 1.885 Å) than in [9Cp]• (NH…O 1.984 Å), suggesting a stronger bond in [10Cp]• in agreement with the increased acidity of thioamides.51 The stronger intramolecular hydrogen bond might also lead to a more 103 Results and Discussion pronounced differentiation between H and H‘. In fact, even the -hydrogen atoms become chemically inequivalent and yield different hfc’s in [10Cp]• (Table 1). The larger hfc to H/H’ in [10Cp]• than in [9Cp]• might be associated with the smaller torsion angle O- N-Cipso-C = –25.1° in [10Cp]•. In order to observe iron-centered radicals, the 5+/PhNO/P t1 Bu reaction mixture was subjected to EPR spectroscopy at 77 K (Figure 9). The EPR spectrum displays three discernible resonances. One resonance originates from the spin trapped product [10Cp]• (5 %) bearing close resemblance to the resonance of the amide nitroxide radical [9Cp]• (Table 2). The other two rhombic resonances, present in nearly equal intensities (48 %; 47 %; Table 2), appear to correlate to follow-up products [11a]• and [11b]• of the initially formed radicals [5−H]•. The identity of these follow-up products is as yet unknown. In the absence of PhNO, the resonances of these follow-up products [11a]• and [11b]• are observed as well (Figure 10). This finding eliminates a reaction with PhNO as being responsible for the formation of [11a]• and [11b]•. In the absence of PhNO, the resonance of the 5+ cation51 is observed additionally (Table 2, Figure 10). Hence, [11a]• and [11b]• seem to be associated to follow-up products of the sulfur substituent [5−H]•, independent of the presence of PhNO. While it has been reported previously that the chemical reactivity and properties of ferrocenyl thioamides can differ from the chemistry of the corresponding amides,51 C- radical reactivity in the presented spin trapping reaction results in the analogous product(s) [9Cp/]• and [10Cp]•. Yet, the sulfur atom appears to open further reaction pathways yielding radicals [11a]• and [11b]•. Their investigation is beyond the scope of the present study and will be reported elsewhere. Obviously, oxidized ferrocenyl amides and thioamides 4+ and 5+ can be easily deprotonated at their respective nitrogen atoms yielding the metal centered radicals [4−HN]• (observed by rapid-freeze EPR) and [5−HN]•. These iron-centered radicals are even resistant towards reaction with PhNO. Yet, the N-deprotonated tautomers [4−HN]• and [5−HN]• can isomerize to reactive C-deprotonated species [4−Hx]• and [5−Hx]• to a small extent (x = , , Cp, Me). Some of these C-centered radicals [4−Hx]• and [5−Hx]• can be trapped by PhNO (<1%) to give the stable nitroxyl radicals [9Cp/]• and [10Cp]•. In the presence of a base even two more radical species [11a]• and [11b]• are detected for the sulfur derivative 5+ by EPR in the reaction mixture. [11a]• and [11b]• are inert towards 104 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure 9. X-band EPR spectrum (exp) and simulated spectrum (sim) of [10Cp]• consisting of three components [10Cp]•, [11a]•, and [11b]•; (5 mM 5 in CH2Cl2) at the following experimental parameters: temperature = 77 K, field = 3245.45 G, sweep = 697.96 G, sweep time = 90 s, modulation = 5000 mG, MW attenuation = 10 db. Figure 10. X-band EPR spectrum (exp) and simulated spectrum (sim) of 5+/P t + • •1 Bu consisting of three components 5 , [11a] , and [11b] ; (5 mM 5 in CH2Cl2) at the following experimental parameters: temperature = 77 K, field = 2499.01 G, sweep = 3989.53 G, sweep time = 120 s, modulation = 3000 mG, MW attenuation = 10 db. 105 Results and Discussion PhNO, similar to other ferrocenium radicals 1+–5+. It appears that Fc-NHC(X)R can be oxidized to the rather stable conjugate acid/base pair [Fc-NHC(X)R]+ / [Fc-NC(X)R]• (X = O, S). C-centered reactive radicals are then slowly formed via intramolecular proton- coupled electron transfer (PCET) or hydrogen atom transfer (Scheme 5).53–64 In the PCET reaction, a proton is transferred from a cyclopentadienyl carbon atom to the nitrogen atom, while an electron from the original CH bond is transferred to the iron(III) center. Hence, 4+/[4−HN]• and 5+/[5−HN]• might act as a reservoir for reactive C-centered radicals (maybe even under physiological conditions). This might also explain some of the enhanced biological reactivity patterns observed for amino ferrocene based drugs and prodrugs such as A, B, or D (Scheme 1). C-centered radicals have also been proposed for the active species derived from ferrocifen F (Scheme 1). Hence, a common feature of such biologically active ferrocene/ferrocenium species might be the presence of some ionizable NH/OH group close to the ferrocenium site, suggesting an intimate coupling of electron and proton transfer to generate metastable iron-centered radicals. Highly reactive carbon- centered radicals might then be formed from these species in small amounts via intramolecular proton-coupled electron transfer reactions. These might account for further biological effects. 3.2.4 Conclusion In contrast to other 17 valence electron metal centered radicals, such as [Mn(CO) ]•5 , ferrocenium ions are inert towards the reaction with the spin trapping agent nitrosobenzene. However, in the presence of a suitable base, small amounts of carbon-centered radicals are generated. Some of these reactive radicals add to nitrosobenzene giving the respective stable nitroxide radicals. EPR spectra of the corresponding stable ferrocenyl phenyl nitroxide radicals clearly reveal the position of the original radical site, namely C5H5 for ferrocene, CH3 for decamethylferrocene, and CH2 for ethylferrocene. The most acidic site in NHC(X)CH3 substituted ferrocene/ferrocenium couples (X = O, S) is the NH group and iron-centered radicals are formed according to EPR studies. Again, these iron-centered radicals do not add to nitrosobenzene. Yet, small amounts of C-deprotonated tautomers generated from the N-deprotonated, iron-centered radicals are trapped by attack of nitrosobenzene at the Cp rings. Hence, these stable and inert NHC(X)CH3 substituted ferrocene/ferrocenium couples slowly release reactive C-centered radicals and can thus be 106 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene considered as a reservoir for reactive radicals. This reactivity might also be of importance in the biological mode of action of OH/NH-substituted ferrocene-based drugs and prodrugs, such as ferrocifen, ferroquine and related compounds. 3.2.5 Experimental Section General Considerations. All reactions were performed under argon atmosphere unless otherwise noted. Dichloromethane was dried over CaH2 and distilled prior to use. Ferrocene (1) was commercially available from Acros. Decamethylferrocene (2) was used as received from ABCR. Ethylferrocene (3), P t1 Bu, 2,2-diphenyl-1-picrylhydrazyl (DPPH), and nitrosobenzene (PhNO) were commercially available from Sigma-Aldrich. N- Acetylaminoferrocene (4),65 N-thioacetylaminoferrocene (5),51 and [nBu4N][B(C6F5)4] 42,43 were prepared according to literature procedures. Filtrations from precipitated silver after oxidation were performed with syringe filters (Rotilabo-Spritzenfilter, Ø = 15 mm, pore size = 0.20 µm; Carl Roth GmbH + Co. KG, Germany). Electrochemical experiments were carried out on a BioLogic SP-50 voltammetric analyzer using a platinum working electrode, a platinum wire as counter electrode, and a 0.01 M Ag/AgNO3 electrode as reference electrode. The measurements were carried out at a scan rate of 100 mV s–1 for cyclic voltammetry experiments and for square wave voltammetry experiments unless noted otherwise using 0.1 M [nBu4N][B(C6F5)4] as supporting electrolyte and 0.001 M solution of the sample in CH2Cl2. Potentials are given relative to the ferrocene/ferrocenium couple. Referencing was achieved by addition of ferrocence or decamethylcobaltocene (E1/2 = −2.04 V vs. FcH/FcH + (CH2Cl ; [ n 2 Bu N][B(C F ) ])) to the sample. 42,43 4 6 5 4 CW EPR spectra (X-band; ca. 9.4 GHz) were measured on a Miniscope MS 300 at 298 K and at 77 K cooled by liquid nitrogen in a finger dewar (Magnettech GmbH, Berlin, Germany). Settings are given at the respective displayed spectra. g-Values are referenced to external Mn2+ in ZnS (g = 2.118, 2.066, 2.027, 1.986, 1.946, 1.906). Simulations of EPR spectra were performed with EasySpin (v 5.0.0)66 for MatLab (R2015a). For quantification measurements, EPR tubes with an internal diameter of 2.0 mm were used. The calibration curve was determined using commercially available 2,2-diphenyl-1-picrylhydrazyl (DPPH) as standard. The samples were prepared in a glove box under argon, and the EPR tubes were filled with 400 µl of the solution and sealed with Critoseal®. They were inserted 107 Results and Discussion 10.4 cm (measured at the Teflon holder) into the EPR spectrometer. Three concentrations (0.03, 0.01, and 0.005 mM) in CH2Cl2 were used for the calibration. The settings for the calibration curve and the sample EPR spectra were as follows: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 5000 mG, MW attenuation = 10 db, number of passes = 3. For an estimation of the error at the insertion of the EPR tube into the spectrometer cavity, the sample with 0.03 mM concentration was inserted, measured, reinserted, and measured three times. For the c = 0.03 mM sample a variation of 13 % between highest and lowest value of the three measurements after double integration is obtained. Baseline correction was achieved with EasySpin66 for MatLab with normalization turned off. The obtained spectra were integrated twice with Origin Pro 8.0, and the double integral values were plotted against the concentration (Figures S11 and S12). Density Functional Calculations. These were carried out with the ORCA 3.0.2 / DFT series67 of programs. For geometry optimizations and energy calculations, the B3LYP formulation of density functional theory was used employing the SV(P)68,69 basis set, the RIJCOSX approximation, approximate Second Order SCF (SOSCF),70,71 the zeroth order regular approximation (ZORA),72–74 the KDIIS algorithm, at GRIDX4. No symmetry constraints were imposed on the molecules. The presence of energy minima of the ground states was checked by numerical frequency calculations. Solvent modeling was done employing the conductor like screening model (COSMO, CH2Cl2). 75 The approximate free energies at 298 K were obtained through thermochemical analysis of the frequency calculation, using the thermal correction to Gibbs free energy as reported by ORCA 3.0.2. 3.2.6 Associated Content Supporting Information. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.organomet.5b00778. Square wave and cyclic voltammograms of P t1 Bu, 1, 1/P t 1 Bu, 4, and 4/P t 1 Bu in CH n2Cl2/[ Bu4N][B(C6F5)4] at 298 K, DFT optimized geometry and spin density in CH2Cl2 for [1-PhNO]+, [1−H]•, [2−H]•, [3−Hx]● (x = , , Cp, 1, 2), [8x]● (x = , , Cp, 1, 2), 108 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene [4−Hx]• and [5−Hx]• (x = , , Cp, N, Me), and [4−HN’]•, EPR spectra and simulations in CH2Cl2 at 77 K of [6]•, [7]•, [81]•, [4−HN]• and [4−Hx]• (x = , , Cp, Me), and at 298 K of [4−Hx]• (x = , , Cp, Me), EPR spectra of quantification experiments in CH2Cl2 at 298 K of [6]•, [7]•, [81]•, [9Cp/]•, [10Cp]•, and DPPH in 0.03 mM, 0.01 mM, and 0.005 mM concentration, and Cartesian coordinates of optimized structures (PDF) Cartesian Coordinates of all DFT optimized structures in .xyz format (XYZ) Authors Information Corresponding Author *E-mail: katja.heinze@uni-mainz.de. Notes ‡Authors contributed equally to the manuscript. The authors declare no competing financial interest. 3.2.7 Acknowledgement We are grateful to Dipl.-Chem. Christoph Kreitner for his helpful remarks concerning DFT calculations. Parts of this research were conducted using the supercomputer Mogon and advisory services offered by Johannes Gutenberg University Mainz (www.hpc.uni- mainz.de), which is a member of the AHRP and the Gauss Alliance e.V. 3.2.8 Reference [1] Special issue on ferrocene chemistry: Heinze, K.; Lang, H. Organometallics 2013, 32, 5623–5625. [2] Special issue on ferrocene chemistry: Adams, R. D. J. Organomet. Chem. 2001, 637-639, 1. [3] a) Togni, A.; Hayashi, T. Ferrocenes; Togni, A., Hayashi, T., Eds.; VCH: Weinheim, 1995; b) Stepnicka, P., Ed., Ferrocenes: Ligands, Materials and Biomolecules, VCH Weinheim, 2008. [4] Ornelas, C. New J. Chem. 2011, 35, 1973–1985. [5] Gasser, G.; Ott, I.; Metzler-Nolte, N. J. Med. Chem. 2011, 54, 3–25. 109 Results and Discussion [6] Tabbì, G.; Cassino, C.; Cavigiolio, G.; Colangelo, D.; Ghiglia, A.; Viano, I.; Osella, D. J. Med. Chem. 2002, 45, 5786–5796. [7] Hagen, H.; Marzenell, P.; Jentzsch, E.; Wenz, F.; Veldwijk, M. R.; Mokhir, A. J. Med. Chem. 2012, 55, 924–934. [8] Osella, D.; Ferrali, M.; Zanello, P.; Laschi, F.; Fontani, M.; Nervi, C.; Cavigiolio, G. Inorg. Chim. Acta 2000, 306, 42–48. [9] Marzenell, P.; Hagen, H.; Sellner, L.; Zenz, T.; Grinyte, R.; Pavlov, V.; Daum, S.; Mokhir, A. J. Med. Chem. 2013, 56, 6935–6944. [10] Daum, S.; Chekhun, V. F.; Todor, I. N.; Lukianova, N. Y.; Shvets, Y. V.; Sellner, L.; Putzker, K.; Lewis, J.; Zenz, T.; de Graaf, I. A. M.; Groothuis, G. M. M.; Casini, A.; Zozulia, O.; Hampel, F.; Mokhir, A. J. Med. Chem. 2015, 58, 2015– 2024. [11] Dubar, F.; Slomianny, C.; Khalife, J.; Dive, D.; Kalamou, H.; Guérardel, Y.; Grellier, P.; Biot, C. Angew. Chem. 2013, 125, 7844−7847; Angew. Chem. Int. Ed. 2013, 52, 7690–7693. [12] Biot, C.; Nosten, F.; Fraisse, L.; Ter-Minassian, D.; Khalife, J.; Dive, D. Parasite 2011, 18, 207–214. [13] Biot, C.; Glorian, G.; Maciejewski, L. A.; Brocard, J. S.; Domaire, O.; Blampain, G.; Millet, P.; Georges, A. J.; Abessolo, H.; Dive, D.; Lebibi, J. J. Med. Chem. 1997, 40, 3715–3718. [14] Zanellato, I.; Heldt, J.-M.; Vessières, A.; Jaouen, G.; Osella, D. Inorg. Chim. Acta 2009, 362, 4037–4042. [15] Hamels, D.; Dansette, P. M.; Hillard, E. A.; Top, S.; Vessières, A.; Herson, P.; Jaouen, G.; Mansuy, D. Angew. Chem. 2009, 121, 9288−9290; Angew. Chem. Int. Ed. 2009, 48, 9124–9126. [16] Hillard, E.; Vessières, A.; Thouin, L.; Jaouen, G.; Amatore, C. Angew. Chem. 2006, 118, 291−296; Angew. Chem. Int. Ed. 2006, 45, 285–290. [17] Jaouen, G.; Top, S.; Vessieres, A.; Leclercq, G.; McGlinchey, M. Curr. Med. Chem. 2004, 11, 2505–2517. [18] Nguyen, A.; Top, S.; Pigeon, P.; Vessières, A.; Hillard, E. a.; Plamont, M. A.; Huché, M.; Rigamonti, C.; Jaouen, G. Chem. Eur. J. 2009, 15, 684–696. 110 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene [19] Michard, Q.; Jaouen, G.; Vessieres, A.; Bernard, B. A. J. Inorg. Biochem. 2008, 102, 1980–1985. [20] Osella, D.; Mahboobi, H.; Colangelo, D.; Cavigiolio, G.; Vessières, A.; Jaouen, G. Inorg. Chim. Acta 2005, 358, 1993–1998. [21] Abramovitch, R. A.; Azogu, C. I.; Sutherland, R. G. J. Chem. Soc., Chem. Commun. 1971, 134–135. [22] Jackson, R. A.; Scarmoutsos, M.; Zarkadis, A. K. J. Chem. Soc. Perkin Trans. 1991, 2, 809–810. [23] a) Cais, M.; Ashkenazi, P.; Dane, S.; Gottlieb, J. J. Organomet. Chem. 1976, 122, 403–411; b) Creary, X. Org. Lett. 2000, 2, 2069–2072. [24] Neuse, E. W. J. Macromol. Sci. Part A - Chem. 1981, 16, 3–72. [25] Taniguchi, N.; Uemura, M. Tetrahedron Lett. 1998, 39, 5385–5388. [26] Ashkenazi, P.; Cais, M. Angew. Chem. 1972, 84, 1106−1107; Angew. Chem. Int. Ed. 1972, 11, 1027–1028. [27] Hudson, A.; Lappert, M. F.; Lednor, P. W.; Nicholson, B. K. J. Chem. Soc. D Chem. Commun. 1974, 966–967. [28] Huffadine, A. S.; Dawson, P. A.; Peake, B. M.; Robinson, B. H.; Simpson, J. J. Organomet. Chem. 1976, 121, 391–403. [29] Bennar, L. S.; Balch, A. L. J. Organomet. Chem. 1977, 134, 121–130. [30] Hudson, A.; Lappert, M. F.; Nicholson, B. K. J. Chem. Soc. Dalt. Trans. 1977, 551–554. [31] a) Forrester, A. R.; Hepburn, P.; Dunlop, R. S.; Mills, H. H. J. Chem. Soc., Chem. Commun. 1969, 698–699; b) Elschenbroich, C.; Cais, M. J. Organomet. Chem. 1969, 18, 135–143; c) Alberti, A.; Benaglia, M.; Bonini, B. F.; Fochi, M.; Macciantelli, D.; Marcaccio, Paolucci, F.; Roffia, S. J. Phys. Org. Chem. 2004, 17, 1084–1090. [32] Babin, V. N.; Belousov, Y. A.; Belousova, T. A.; Borisov, Y. A.; Gumenyuk, V. V.; Nekrasov, Y. S. Russ. Chem. Bull. Int. Ed. 2011, 60, 2081–2087. [33] a) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910; b) Fujita, E.; Gordon, B.; Hillman, M.; Nagy, A. G. J. Organomet. Chem. 1981, 218, 105–114. [34] Prins, R. Mol. Phys. 1970, 19, 603–620. 111 Results and Discussion [35] Neidlinger, A.; Ksenofontov, V.; Heinze, K. Organometallics 2013, 32, 5955– 5965. [36] Kondo, Y. In Superbases for Organic Synthesis; Ishikawa, T., Ed.; John Wiley & Sons, Ltd: Chichester, UK, 2009; pp 145–185. [37] Barclay, L. R. C.; Dust, J. M. Can. J. Chem. 1982, 60, 607–615. [38] Omelka, L.; Kováčová, J. Magn. Reson. Chem. 1994, 32, 525–531. [39] Cais, M.; Ashkenazi, P.; Dani, S.; Gottlieb, J. J. Organomet. Chem. 1977, 124, 49–58. [40] Buettner, G. R. Free Radic. Biol. Med. 1987, 3, 259–303. [41] Huesmann, H.; Förster, C.; Siebler, D.; Gasi, T.; Heinze, K. Organometallics 2012, 31, 413–427. [42] LeSuer, R. J.; Buttolph, C.; Geiger, W. E. J. Organomet. Chem. 2004, 76, 6395– 6401. [43] Camire, N.; Mueller-Westerhoff, U. T.; Geiger, W. E. J. Organomet. Chem. 2001, 639, 823–826. [44] Bildstein, B.; Malaun, M.; Kopacka, H.; Wurst, K.; Ongania, K.-H.; Opromolla, G. Organometallics 1999, 18, 4325–4336. [45] Sanders, R.; Mueller-Westerhoff, U. T. J. Organomet. Chem. 1996, 512, 219–224. [46] Guillaneux, D.; Kagan, H. B. J. Org. Chem. 1995, 60, 2502–2505. [47] Förster, C.; Heinze, K. Z. Anorg. Allg. Chem. 2015, 641, 517–520. [48] Plazuk, D.; Vessières, A.; Hillard, E. A.; Buriez, O.; Labbé, E.; Pigeon, P.; Plamont, M. A.; Amatore, C.; Zakrzewski, J.; Jaouen, G. J. Med. Chem. 2009, 52, 4964–4967. [49] a) Heinze, K.; Siebler, D. Z. Anorg. Allg. Chem. 2007, 633, 2223–2233; b) Siebler, D.; Linseis, M.; Gasi, T.; Carrella, L. M.; Winter, R. F.; Förster, C.; Heinze, K. Chem. Eur. J. 2011, 17, 4540–4551; c) Siebler, D.; Förster, C.; Heinze, K. Dalton Trans. 2011, 40, 3558–3575. [50] Heinze, K.; Hüttinger, K.; Siebler, D. in Modeling of Molecular Properties (Ed. P. Comba), Wiley-VCH, 2011. [51] Kienz, T.; Förster, C.; Heinze, K. Organometallics 2014, 33, 4803–4812. [52] Eaton, G. R.; Eaton, S. S.; Barr, D. P.; Weber, Ralph, T. Quantitative EPR; Springer Wien, New York, 2010. 112 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene [53] Dempsey, J. L.; Winkler, J. R.; Gray, H. B. Chem. Rev. 2010, 110, 7024–7039. [54] Hammes-Schiffer, S. Chem. Rev. 2010, 110, 6937–6938. [55] Hammes-Schiffer, S.; Stuchebrukhov, A. Chem. Rev. 2010, 110, 6939–6960. [56] Wenger, O. S. Chem. Eur. J. 2011, 17, 11692–11702. [57] Herzog, W.; Bronner, C.; Löffler, S.; He, B.; Kratzert, D.; Stalke, D.; Hauser, A.; Wenger, O. S. ChemPhysChem 2013, 14, 1168–1176. [58] Warren, J. J.; Tronic, T. A.; Mayer, J. M. Chem. Rev. 2010, 110, 6961–7001. [59] Weinberg, D. R.; Gagliardi, C. J.; Hull, J. F.; Murphy, C. F.; Kent, C. A.; Westlake, B. C.; Paul, A.; Ess, D. H.; Mccafferty, D. G.; Meyer, T. J. Chem. Rev. 2012, 112, 4016–4093. [60] Bonin, J.; Robert, M. Photochem. Photobiol. 2011, 87, 1190–1203. [61] Savéant, J.-M. Energy Environ. Sci. 2012, 5, 7718–7731. [62] Gagliardi, C. J.; Vannucci, A. K.; Concepcion, J. J.; Chen, Z.; Meyer, T. J. Energy Environ. Sci. 2012, 5, 7704–7717. [63] Hammarström, L.; Styring, S. Energy Environ. Sci. 2011, 4, 2379–2388. [64] Costentin, C.; Robert, M.; Savéant, J.-M. Phys. Chem. Chem. Phys. 2010, 12, 11179–11190. [65] Heinze, K.; Schlenker, M. Eur. J. Inorg. Chem. 2004, 2974–2988. [66] Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42–55. [67] Neese, F. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2012, 2, 73–78. [68] Schäfer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571–2577. [69] Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297–3305. [70] Neese, F. Chem. Phys. Lett. 2000, 325, 93–98. [71] Fischer, T. H.; Almlof, J. J. Phys. Chem. 1992, 96, 9768–9774. [72] Lenthe, E. van; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1993, 99, 4597– 4610. [73] Van Wüllen, C. J. Chem. Phys. 1998, 109, 392. [74] Pantazis, D. A.; Chen, X. Y.; Landis, C. R.; Neese, F. J. Chem. Theory Comput. 2008, 4, 908–919. [75] Sinnecker, S.; Rajendran, A.; Klamt, A.; Diedenhofen, M.; Neese, F. J. Phys. Chem. A 2006, 110, 2235–2245. 113 Results and Discussion 114 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates 3.3 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Torben Kienz, Christoph Förster, and Katja Heinze Submitted 115 Results and Discussion Supporting information for this article (without Cartesian coordinates from DFT calculations) is found at pp.175. Adapted with permission from T. Kienz, C. Förster, K. Heinze. 116 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates 3.3.1 Abstract In the presence of oxidant (Ag[SbF6]) and base, N-ferrocenyl thioamide Fc-NHC(S)CH3 (H-1; Fc = Fe(η55-C H )(η55 5 5-C5H4)) converts in an unexpected multistep reaction sequence to a novel N,S-heterocyclic ring which initiates an oligomerization reaction. Key intermediates towards the resulting complicated material are Ag6(1)6 silver clusters of the anionic N,S-chelating ligand 1–, EPR-active piano stool complexes resulting from ring- slipped cyclopentadienyl ligands as well as electrophilic N-ferrocenyl ketenimine Fc- N=C=CH2 (2) and its ferrocenium cation 2+ formed by hydrosulfide elimination. Mechanistic insight is achieved using XRD, mass spectrometric as well as EPR and NMR spectroscopic studies combined with DFT calculations. In addition to the fundamental mechanistic insight, the results could have impact for smart oligo-mers/polymers, heterocycle synthesis and controlled-release materials. 117 Results and Discussion 3.3.2 Introduction N-Ferrocenyl amides Fc-NHC(O)R (Fc = Fe(η5-C 55H5)(η -C5H4); R = CH3: Z-[H-3]) and thioamides Fc-NHC(S)R (R = CH3: E/Z-[H-1]) become acidic upon oxidation to the corresponding ferrocenium ions due to the increased positive charge close to the nitrogen atom.1-8 Vice versa, deprotonation of the amide lowers the oxidation potential of the ferrocene/ferrocenium redox couple due to the negative charge adjacent to ferrocene.7-9 In addition to the expected acidity of the NH group of ferrocenyl amides and thioamides, CH activation can also occur in the presence of a non-nucleophilic base (P1- tBu) and an oxidant (Ag[SbF6]), leading to ferrocenyl radicals (P1- tBu = tert-butylimino- tris(dimethylamino)phosphorane). These carbon-centered radicals have been trapped by nitrosobenzene (PhNO) and unambiguously identified by EPR spectroscopy as their characteristic nitroxide radicals [PhN(O)-1] and [PhN(O)-3], respectively (Scheme 1a).7 Even ferrocene Fe(C5H5)2 and decamethylferrocene Fe(C5Me5)2 display such radical reactivity giving the respective ferrocenyl phenyl nitroxide radicals Fe(C 5H5)(PhN (O))C5H4) (Scheme 1b) and Fe(C5Me5)(PhN(O)CH2)C5Me4), respectively.7 Scheme 1. Established and unexplained radical reactivity of a) N-thioacetyl amino ferrocene H-1 and N-acetyl aminoferrocene H-3 and b) ferrocene Fe(C5H5)2 in the presence of oxidant and base followed by trapping with PhNO.7 118 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Sulfur derivatives such as E/Z-[H-1]6 exhibit a further, as yet undisclosed reactivity in the presence of both base and oxidant, based on electrochemical data as well as on EPR spectroscopic findings (Figure 1).6,7 Unusual rhombic EPR resonances are observed at 77 K in frozen solution (Figure 1b). These EPR patterns differ significantly from the typical axial resonances of ferrocenium ions, such as E-[H-3]+ or E/Z-[H-1]+ (Figure 1a).6-11 In addition to the rhombic symmetry, the g anisotropy (Δg) is dramatically reduced. Yet, Δg is still much larger than expected for EPR resonances of light-atom radicals, suggesting a strong participation of the iron center. The occurrence of such resonances is specific for ferrocenyl thioamides but not for ferrocenyl amides or unsubstituted ferrocene. Obviously, the sulfur atom in place of the oxygen atom plays a major role. Furthermore, the 1:1 mixture of E/Z iso-mers E/Z-[H-1] gives rise to two different rhombic EPR resonanc-es in CH2Cl2 in a 4:1 ratio suggesting an effect of the original configuration for these two open-shell species (Figure 1b). In the presence of PhNO, the ratio of the two species changes to 1:1 pointing to subtle differences in solubility or reactivity of the radicals or precursors to the radicals derived from E-[H-1] and Z-[H-1].7 In the present mechanistic study, we will disclose the unexpected reactivity of ferrocenyl thioamide E/Z-[H-1] under oxidative and alkaline conditions. The proposed pathway encompasses reactive open-shell intermediates and finally leads to diamagnetic ferrocenyl containing oligomers with [CN]n backbone and N,S-heterocyclic head groups. Details of the suggested multistep reaction mechanism are supported by EPR and NMR spectroscopic and mass spectrometric data, by single crystal XRD analyses of intermediates and products as well as by density functional theory (DFT) calculations. Apart from the fundamental mechanistic insight, the obtained results could be relevant for smart redox-active polymers with ferrocene in the side chain12-15, for heterocycle synthesis and for ferrocene-containing pro-drugs for controlled release.16 119 Results and Discussion Figure 1. a) EPR spectrum of E/Z-[H-1]+ in CH2Cl2 at 77 K and simulated spectrum [B0 2499.81 G; sweep 3791.34 G; modulation 3000 mG; MW attenuation 3 dB; gain 5.0; sweep time 120 s; c = 20 mM] and b) EPR spectrum of E/Z-[H-1] treated with Ag[SbF6] and P1-tBu in CH2Cl2 at 77 K [B0 3245.45 G; sweep 697.96 G; modulation 5000 mG; MW attenuation 15 dB; gain 5.0; sweep time 90 s; c = 19 mM]; simulated EPR spectrum of the radical mixture (ratio 4:1) and the corresponding subspectra. 3.3.3 Results and Discussion The multifaceted interactions of E/Z-[H-1] i) with Ag[SbF6], ii) with bases and iii) with both agents will be examined and discussed in the following. Reaction of E/Z-[H-1] and Ag[SbF6]. In CH2Cl2 solution, thioamide E/Z-[H-1] exists as a 1:1 mixture of E/Z isomers. 6 E/Z-[H- 1] is conveniently oxidized to the ferrocenium cations E/Z-[H-1]+ by Ag[SbF6] in CH2Cl2 according to their respective redox potentials.6,17 The presence of ferrocenium ions is 120 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates confirmed by EPR spectroscopy with E-[H-1]+ and Z-[H-1]+ delivering indistinguishable essentially axial EPR resonances (Figure 1a; g = 3.400, 1.860, 1.780).6 In CH3CN, the oxidation potential of Ag+ is insufficient for oxidation of the ferrocenyl unit in E/Z-[H- 1].17 The ESI+ mass spectrum of this mixture in CH3CN displays peak clusters at m/z = 365.92 and m/z = 624.89 fitting to the mono- and diligated silver complex cations [Ag(H- 1)]+ and [Ag(H-1)2]+ with expected isotopic patterns of 107/109Ag, respectively (Scheme 2; Supporting, Information, Figure S1). Scheme 2. Initial steps in the reaction sequence of E/Z-[H-1] in the presence of base and oxidant (compounds in boxes are identified by spectroscopic/analytical techniques highlighted in yellow). As the oxidative power of silver cations depends on their coordination environment (solvent, potential ligands)17, the redox reaction between Ag+ and sulfur-containing H-1 and the competing κS coordination reaction to give [Ag(H-1) +2] is probably a delicate equilibrium (Scheme 2). Indeed, attempts to grow crystals of silver complexes of H-1 failed. However, a copper(I) complex has been successfully prepared from E/Z-[H-1] and tetrakis(acetonitrile)copper(I) tetrafluoroborate (Scheme 2). The copper(I) complex crystallizes with a [Cu2(H-1)6][BF4]2 stoichiometry featuring hydrogen bonded counter ions (Figure 2a). A detailed discussion of the solidstate structure and comparison with 121 Results and Discussion similar compounds18-21 can be found in the Supporting Information (Supporting Information, Table S1). NMR, IR, and conductivity studies22 in solution confirm that the [Cu2(H-1)6]2+ structural motif and the hydrogen bonded counter ions ([Cu2(H-1)6][BF4]2 1:2 contact ion pairs) is retained in CH2Cl2 solution as discussed in more detail in the Supporting Information. Figure 2. a) Molecular structure of [Cu2(H-1)6][BF4]2 derived from single crystal XRD analysis (CH protons omitted for clarity; four additional [BF −4] ions are shown to illustrate the hydrogen bonding pattern). b) Molecular structure of Ag6[µ3-(E-1)]6 derived from single crystal XRD analysis (CH protons omitted for clarity). These contact ion pairs are oxidized at E½ = 0.16 V (2 e –), 0.36 V (4 e–) and 0.73 V (2 e-), suggesting that the ferrocene units are reversibly oxidized to ferrocenium cations in two 122 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates electrochemical steps. Finally, the copper(I) centers are oxidized quasireversibly as often observed for copper(I) complexes (Supporting Information, Figures S13 and S14). The high positive charge of the terminal ferrocenium ions might be compensated for by counter ions, yet the CuI/II oxidation is not reversible on the electrochemical timescale. The UV/vis absorption bands of the ferrocenyl moieties of [Cu2(H-1)6][BF4]2 in CH2Cl2 are shifted to lower energy from 348/446 nm (H-1) to 378/456 nm (Supporting Information, Figure S15). Although silver(I) complexes of thioamides H-1 could not be crystallized and NMR spectroscopic analyses of Ag[SbF6]/H-1 mixtures are hampered due to the concomitant formation of open-shell species by oxidation processes, we assume, that, based on ESI mass spectroscopic evidence (vide supra, Supporting Information, Figure S1) and the formed copper(I) complex [Cu2(H-1)6][BF4]2, contact ion pairs of the type [Agn(H-1)m][SbF6]n should be present in CH2Cl2 solution as well. Indeed, silver(I) complexes of the redoxinert HpyS ligand display polymeric structures [Ag 23n(HpyS-κS)2n][X]n in the solid state. Hence, E/Z-[H-1] can simply coordinate to Ag[SbF6] to give [Agn(E/Z-[H-1])m][SbF6]n ion pairs or can be oxidized by Ag[SbF6] to give E/Z-[H-1][SbF6] ferrocenium salts, depending on the environment, e.g. the solvent (Scheme 2). Reaction of and E/Z-[H-1] and bases. Deprotonation of thioamide E/Z-[H-1] is facile with n-butyl lithium nBuLi, lithium dimethyl amide Li(NMe2) or phosphazene base 24 P t1- Bu (Scheme 2). In all cases, the resulting 1H NMR spectra confirm the NH deprotonation based on the disappearance of the NH proton resonances at 8.96 and 8.25 ppm of E-[H-1] and Z-[H-1], respectively (Supporting Information, Figure S16). Furthermore, all resonances of [E-1]– are significantly broadened and display lower intensity, while the resonances of [Z-1]– remain sharp. Indeed, a precipitate is observed. This precipitate might well consist of less soluble [cation]+[E-1]– salts, while [cation]+[Z-1]– is obviously better soluble in CD2Cl2 (cation = [Li(THF) +n] ; [H-P t + 1- Bu] ; Supporting Information, Figure S16). This different solubility could be due to the different coordinating ability of [E-1]– and [Z-1]–. Hence, E/Z-[H-1] can be deprotonated by bases to give salts [cation][E-1] and [cation][Z-1], respectively, with different solubility (Scheme 2). 123 Results and Discussion Deprotonation and Oxidation of E/Z-[H-1]. Preparative scale deprotonation of E/Z-[H-1] and subsequent oxidation with Ag+ is achieved i) by one equivalent of nBuLi in THF, removal of the solvent, oxidation with Ag[SbF + 176] in CH2Cl2 (E½ = 0.65 V in CH2Cl2 vs. FcH/FcH ) (method A) and purification by column chromatography over silica, ii) by one equivalent of Li(NMe2) in THF, removal of the solvent, oxidation with Ag[SbF6] in CH2Cl2 (method B) and purification by column chromatography over aluminium oxide or iii) by one equivalent of P1- tBu in CH2Cl2 followed by oxidation with Ag[SbF6] and filtration through a syringe filter (method C). Before chromatographic workup, the samples of these mixtures were subjected to mass spectrometric analyses (ESI+, FD+). EPR samples were prepared by oxidation with Ag[SbF6] and deprotonation with P t 1- Bu in CH2Cl2, followed by filtration (method C). From a sample prepared by method C, a few orange-red crystals separated which were analyzed by single crystal XRD analysis (Figure 2b; Supporting Information, Table S2, Figure S17). A hexameric, centrosymmetric neutral cluster Ag6[µ3-(E-1)]6 consisting of a distorted Ag6 octahedron (Ag•••Ag distances 2.886 – 3.552 Å) with six edges bridged by sulfur atoms of the six E-1– ligands is present (Supporting Information, Table S2). The E configuration further allows for a κN,κS bridging mode of the deprotonated thioamide E-1–. A similar assembly has been observed in Ag6L6 clusters with L = anion of 1-R- imidazoline-2-thione (R = Pr, iPr),25 2-hydroxynaphthalene-1-carboxaldehyde thio- semicarbazonate,26 ferrocenyl thiosemicarbazonate,27 and 2-benzimidazolethiolate.28 The C=S distances in Ag6[µ3-(E-1)]6 (1.767(9) Å) are significantly larger than that in H-1 (1.671(2) Å)6 suggesting a dominant thioiminolate character of the coordinated ligand. All Fe-C distances support neutral ferrocenes without ferrocenium character, i.e. no electron transfer from ferrocene to Ag+ has occurred (Scheme 2, Table S2). The poor solubility of Ag6(E-1)6 removes E-1– from the reaction mixture, i.e. enriches the solution in Z-1– and furthermore represses the oxidation of E-1– to the neutral radical E-1 (Scheme 2). Yet, the cluster Ag (E-1) can act as reservoir of E-1–6 6 and E-1, respectively. These events are further probed by EPR spectroscopy in solution. The EPR spectrum of a sample rapidly frozen after addition of P t1- Bu to a solution of E/Z-[H-1]+ in CH2Cl2 (from a 1:1 E/Z-[H-1] mixture) merely shows the axial resonance of the ferrocenium ions E/Z-[H-1]+ (g = 3.400, 1.860, 1.780; Figure 1a). Upon warming to room temperature for 2 minutes, the rhombic resonances at g = 2.310, 2.070, 2.000 124 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates (subspectrum 1) and g = 2.220, 2.058, 2.011 (subspectrum 2) appear in a 4:1 ratio (Figure 1b), while the axial resonance of E/Z-[H-1]+ vanished. Interestingly, no comparable rhombic EPR resonances appear in the reaction of ferrocenium amide [H-3]+ and P -t1 Bu, so this reactivity is associated with the sulfur of E/Z-[H-1]+. Figure 3. (Top) Experimental EPR spectrum of Z-[H-1]/DDQ at 77 K prepared by method D in CH3CN (black) [B0 2499.14 G; sweep 3989.53 G; modulation 3000 mG; MW attenuation 7 dB; gain 5.0; sweep time 120 s; c = 19 mM]. (Bottom) simulated EPR spectrum. In CH2Cl2, thioamide H-1 is present as 1:1 E/Z mixture, while in CH3CN the Z isomer Z- [H-1] is preferred. This opens the opportunity to monitor the fate of a single isomer. However, as Ag[SbF6] is incompetent to oxidize H-1 in CH3CN, 2,3-dichloro-5,6-dicyano- 1,4-benzoquinone (DDQ) is employed both as oxidant and base in CH3CN, both for EPR and mass spectrometric experiments (method D). The EPR spectrum obtained by method D in CH3CN (g = 2.222, 2.055, 2.008; Z isomer; Figure 3) coincides with subspectrum 2 obtained by method C in CH2Cl2 (Figure 1b). Accordingly, the EPR subspectrum 1 obtained by method C is assigned to species arising from the E isomer (Figure 1b). This EPR-active open-shell species derived from Z-[H-1] decays within 30 min at room temperature in CH3CN to EPR-silent products (Scheme 3, Figure 4a; half-life τ½ = 13 min in CH3CN). EPR experiments with a 1:1 mixture of both isomers E-[H-1] and Z-[H-1] in CH2Cl2 (method C), monitoring both radicals over time show that the EPR resonance 125 Results and Discussion derived from Z-[H-1] (g = 2.220, 2.058, 2.011; τ ½ = 3 min) decays faster than the EPR resonance derived from E-[H-1] (g = 2.310, 2.070, 2.000; τ ½ = 84 min) (Figure 4b, 4c). Scheme 3. Ketene imine formation via open-shell intermediates (compounds in boxes are identified by spectroscopic/analytical techniques highlighted in yellow). Having assigned the subspectra 1 and 2 (Figure 1b) to radicals derived from E-[H-1] and Z-[H-1], respectively, their composition and structures are not yet clear (Schemes 2 and 3). The rhombic symmetry of the EPR resonances is incompatible with simple axial ferrocenium ions such as E/Z-1• (Scheme 3) Hence, a reduction of symmetry has to be taken into account, e.g. by coordination of a nucleophile to the FeIII center in the zwitterions E-1 and Z-1. Notably, the rhombic resonances (Figures 1b, 3, 4) closely resemble those obtained for “half-open” ferrocenium ions, e.g. FeIII(η5-Cp)(η5-2,4-Me 292C5H7) or pseudo- 126 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure 4. a) Experimental EPR spectra of Z-[H-1] at 77 K treated with DDQ in CH3CN (method D) after different time intervals at room temperature [B0 2499.01 G; sweep 3989.53 G; modulation 3000 mG; MW attenuation 7 dB; gain 5.0; sweep time 120 s; c = 13 mM], b) experimental EPR spectra of E-[H-1]/Z-[H-1] at 77 K treated with Ag[SbF6] and P t1- Bu in CH2Cl2 (method C) after different time intervals at room temperature [B0 3266.74 G; sweep 999.54 G; modulation 5000 mG; MW attenuation 7 dB; gain 5.0; sweep time 120 s; c = 30 mM] and c) time traces of the relative EPR resonance intensities in CH2Cl2 including exponential fits. 127 Results and Discussion octahedral d5 low-spin [FeIII(η5-Cp*)(dppe)(CCR)]+ cations [dppe = 1,2-bis(diphenyl- phosphano)ethane], i.e. complexes with a [FeIII(η5-Cp)LL’L’’]2+ piano stool symmetry.30,31 Figure 5. DFT calculated molecular structures of model piano stool open- shell species a) [(Z-1)(E-CH C(S)=N-CH )-κS]–3 3 and b) [(E-1)(E- CH C(S)=N-CH )-κNκS]–3 3 (protons omitted for clarity). Such piano stool species can be envisaged by nucleophilic attack of anions [E/Z-1]– to the zwitterions E-1 and Z-1 (Scheme 3). DFT calculations indeed suggest the formation of species with slipped Cp rings. According to the calculations, coordination of a model nucleophile [E-CH3C(S)=N-CH ] – 3 to zwitterions E-1 or Z-1 induces a ring slippage of the substituted C5H4-NC(S)CH rings from η 5 3 to η 1. With zwitterion Z-1, the sulfur atom of the substituent of the η1-Cp ring binds to iron(III) in a chelating fashion, in addition to the κS coordination of the nucleophile ([Z-1E-1]–; Scheme 3; Figure 5a). This coordination-induced ring-slippage results in a [FeIII(η5-C5H5)(η 1-1,2-C5H4-NC(S)CH3- κS)(CH3C(S)=N-CH3-κS)] – piano stool geometry. In the E-1 isomer, the sulfur substituent is incompetent to coordinate to iron(III) for steric reasons. Hence, the incoming nucleophile [E-CH3C(S)=N-CH – 3] coordinates in a chelating κN,κS fashion, resulting in a [Fe III(η5- C H )(η15 5 -1,3-C5H4-NC(S)CH3)((CH3C(S)=N-CH3-κN,κS))] – piano stool geometry [E-1E-1]–; Scheme 3; Figure 5b). This [FeCp(C)(S)(N)] coordination also accounts for the larger g anisotropy of the EPR resonance derived from E-1• ((Δg = 0.31; Figure 1b, subspectrum 1) as compared to the EPR resonance of piano stool complex [FeCp(C)(S)(S)] derived from Z-1 (Δg = 0.21; Figure 1b, subspectrum 2). Notably, [Z-1E-1]– features a η1-1,2-C5H4-NC(S)CH3 ligand while a 1,3-substitution is favored in [E-1E-1]– according to the calculations. As a working hypothesis, the piano stool radicals [Z-1E-1]–/[E-1E- 1]– either directly transform into EPR-silent products or act as a reservoir of stabilized 128 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates open-shell species which further transform into EPR-silent products (Scheme 3). These follow-up products are probed by mass spectrometry and NMR spectroscopy. Figure 6. a) FD+ mass spectrum of a sample prepared by method D in CH3CN after cobaltocene reduction; b) experimental (black) and calculated (blue) isotope distributions and assignments of relevant peaks. Treating the reaction mixture of Z-[H-1] and DDQ in CH3CN after 30 min with cobaltocene Co(C5H5)2 to reduce the paramagnetic species to diamagnetic ones gave rise to the evolution of hydrogen sulfide. The H2S was detected by its characteristic odor as well as by precipitation of black lead sulfide PbS from the reaction of Pb(OAc)2 with the evolved gas. Obviously, sulfur is easily eliminated during the course of the reaction.16 Furthermore, oligomeric species lacking H2S are observed in the FD + mass spectrum of the reaction mixture after reduction and chromatographic workup in addition to the peak of the starting 129 Results and Discussion material [H-1]+ at m/z = 258.9 (Figure 6). The peak at m/z = 484.0 can be assigned to species with the formula C24H •+24Fe2N2S ([H-5] /[H-6]•+, see below), while the peaks at 677.2 and 902.0 correspond to species with the formulas C +36H35Fe3N3 ([(2)3+2H] , see below) and C H +48 46Fe4N4 ([(2)4+2H] , see below), respectively, completely lacking sulfur (Figure 6, Scheme 3). The (at least partial) elimination of sulfur is also confirmed by the observation of peaks with mass-to-charge ratios of 226.04 (C H +12 12FeN) and 243.04 (C12H13FeNO) in the ESI mass spectrum (method B, Figure 7). These peaks can be assigned to the N-ferrocenyl keteniminium ion [H-2]+ (or its isomeric N-ferrocenyl nitrilium ion) and ferrocenium amide Z-[H-3]+, respectively (Scheme 3). These species derive from the N-ferrocenyl ketenimine/(N-ferrocenium)yl ketenimine redox couple 2/2+ by protonation ([H-2]+) and by addition of water (Z-[H-3]+), respectively (Scheme 3). The ferrocene/ferrocenium ketenimine redox pair 2/2+ obviously results from desulfurization of H-1. We suggest, that zwitterions E-1/Z-1, formed by deprotonation and oxidation of E-[H-1]/Z-[H-1], or their corresponding piano stool complexes eliminate hydrosulfide to give the ferrocenium ketenimine 2+ (Scheme 3). Tentative transition states for the AgSH eliminations from piano stool complexes [Z-1E-1]– and [E-1E-1]– are depicted in Scheme 3. Possibly, preorganized Ag[Z-1E-1]– allows for a faster [Ag(SH)(E-1)]– elimination than Ag[Z- 1E-1] accounting for the rapid decay of the [Z-1E-1]– EPR resonance (Figure 4). After back-slippage of the η1-C 55H4-N=C=CH2 rings to a η coordination, the ferrocenium ketenimine 2+ is formed. This ferrocenium ion can be reduced to ketenimine 2. Neither deprotonation, nor oxidation alone leads to the release of sulfur from H-1 according to ESI+ mass spectra, hence HS– elimination from the zwitterions E-1/Z-1 (or their piano stool derivatives, see Scheme 3) is highly conceivable. Treating the organic thioamide PhNHC(S)CH3 with oxidants and bases does not lead to detectable HS – elimination or sulfur-deficient species as deduced from odor and ESI mass spectrometry, possibly due to the incompatible redox potentials.17 Hence, the redox-active ferrocenyl substituent is crucial for the observed reactivity. The key ferrocenyl/ferrocenium ketenimines 2/2+ are certainly highly electrophilic, similar to classical ketenimines and activated keteniminium ions.32-37 A reactive ferrocenyl ketene Fc(H)C=C=O has been trapped by the Staudinger [2+2] cycloaddition reaction with imines to give β-lactams38, and spectroscopically identified by its IR signature.39 The 130 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates sterically protected ferrocenyl tri(isopropyl)silyl ketene Fc(iPr3Si)C=C=O has been even structurally characterized.40 The C-ferrocenyl ketenimine Fc(H)C=C=NH has been prepared from acetonitrile and isolated.41 To the best of our knowledge, the N-ferrocenyl ketenimine H2C=C=NFc 2 has not yet been reported. Given that reactive ferrocenyl ketenimines form, the observed peaks at mass-to-charge ratios of 677.2 and 902.0 in the FD+ mass spectrum (Figure 6) are straightforwardly assigned to oligo ketenimines [(2)n+2H]+ (n = 3, 4) with a tentative structure depicted in Scheme 3.42 Formally, the ketene imine 2 has oligomerized, has been hydrogenated at the end groups (+2H+ +2e–) and is then oxidized to a ferrocenium cation in the mass spectrometer (–e–) (Scheme 3). The N- ferrocenyl ketenimine cation 2+ is even expected to be superelectrophilic in nature due to the cationic ferrocenium substituent.32 Indeed, the DFT calculated orbital coefficient at the central carbon atom of the ketenimine’s LUMO increases upon oxidation of 2 to 2+ (Supporting Information, Figure S18).37 Nucleophiles are indeed present in the reaction mixture, namely the anions E-1–/Z-1–, simply formed by deprotonation of E-[H-1]/Z-[H-1] (vide supra). Hence, nucleophilic attack of the ambident N/S nucleophiles E-1–/Z-1– at the superelectrophilic ketenimine 2•+ is conceivable (Scheme 4) and should lead to neutral zwitterionic species 5 (N attack) and 6 (S attack) with the composition C24H23Fe2N2S (m/z = 483). Peaks with mass-to-charge ratios m/z = 483.95 (FD+, Figure 3) and 485.06 (ESI+, Figure 4) are observed in the FD+ and ESI+ mass spectra, respectively, corresponding to protonated species [H-5]+/[H-6]+ and [H +2-5] /[H2-6]+ respectively (Scheme 4). N or S attack of E-1–/Z-1– at 2+ cannot be distinguished at this stage by mass spectrometry. Yet, steric arguments should favor N attack by the Z isomer and S attack by the E isomer (Scheme 4). Preparative scale reaction (method B) yields a brown ferrocene containing product, which consists of at least two E/Z isomers of thioamides H-5 according to 1H and 13C NMR spectroscopic assignment and FD+ mass spectrometry, but more isomers are conceivable (Supporting Information, Figures S19-S25). These products arise from the nucleophilic attack of the amide nitrogen of E/Z-1– at 2+ (N attack, Scheme 4), followed by protonation and reduction of the ferrocenium fragment (Scheme 4). A single isomer H-5 predominates, showing a characteristic 13C NMR resonance for the C=S fragment at δ = 204.3 ppm. No attempts were performed to separate the E/Z isomers of H-5. 131 Results and Discussion Figure 7. a) ESI+ mass spectrum of a sample prepared by method B in CH2Cl2. b) Experimental (black) and calculated (blue) isotope distributions and assignments of important peaks. 132 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Scheme 4. Reactivity of ketene imines (compounds in boxes are identified by spectroscopic/analytical techniques highlighted in yellow). Method A allowed isolating and identifying products arising from the nucleophilic attack of the sulfur atom of 1– at 2+ (S attack) (Scheme 4). After purification, a mixture of two charged multiferrocenyl compounds (with three and four ferrocenyl units; m/z = 945.96 and 1171.00, respectively) was obtained as their [SbF –6] salts. This suggested multiple reactions with ketenimines 2/2•+, possibly by way of anionic polymerization. However, the basic structural motif remained elusive, until single crystals of cyclo-[H3-7][SbF6]2 were obtained from the reaction mixture (Figure 8; Supporting Information, Table S3, Figure S26). cyclo-[H3-7]2+ features a novel 5-amino substituted 4,5-dihydro-1,3-thiazolium heterocycle with an amidinium substituent in the 5-position. Although the 4,5-dihydro-1,3- thiazolium ring is a rather unusual N,S heterocycle per se, the structure itself features no 133 Results and Discussion exceptional metrical parameters (Supporting Information, Table S3). The C-C and C-N distances fit to the proposed double bond characters as indicated in cyclo-[H3-7]2+ (Scheme 4). The twofold positive charge of cyclo-[H 2+3-7] is basically located at the ring nitrogen atom and at the terminal amidinium nitrogen atom. All ferrocenyl substituents display Fe-C distances consistent with neutral ferrocenyl substituents (Supporting Information, Table S3). The aminidium NH group engages in a non-classical NH•••Fe hydrogen bond to the central ferrocene with a short Fe2•••N3 distance of 3.364(5) Å forming a six-membered ring. Such non-classical bonding motifs are increasingly observed in NH containing ferrocenyl compounds with rather acidic NH groups exhibiting Fe•••N distances around 3.46 Å.43-46 Figure 8. Molecular structure of cyclo-[H3-7][SbF6]2 derived from single crystal XRD analysis (Cp protons and [SbF6]− counter ions omitted for clarity; atom numbering according to CIF file). This intriguing structural motif allows tracing back the formation of the initial zwitterionic species cyclo-[H-6], which initiates the oligomerization of ketenimine 2 via zwitterionic cyclo-[H-7] and cyclo-[H-8] (Scheme 4). Protonation of these 4,5-dihydro-1,3-thiazolium enamides to the 4,5-dihydro-1,3-thiazolium imines cyclo-[H +2-7] and cyclo-[H2-8]+ terminates the anionic oligomerization process (Scheme 4). The doubly protonated species of cyclo-[H-7] and cyclo-[H-8], namely {cyclo-[H3-7][SbF6]}+ and {cyclo-[H3-8][SbF ]}+6 have been identified in the ESI+ mass spectra as ion pairs at m/z = 945.96 (calcd. 945.97) and m/z = 1170.99 (calcd. 1171.00) with expected isotopic distribution patterns (121,123Sb), respectively (Figure 7). NMR and FD+ mass spectroscopic data further confirm the composition and structure of cyclo-[H3-7][SbF6]2 (Supporting Information, Figures S27- S33). The zwitterionic 4,5-dihydro-1,3-thiazolium heterocycle of cyclo-[H-6] is formed by a ring-closing reaction of H-6 (m/z = 484, vide supra). To the best of our knowledge, no such 134 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates architectures have been reported before. Comparable mesoionic heterocycles are the münchnones, sydnones or montrealones.47-50 Acyclic H-6 itself can be envisaged as the product of the nucleophilic attack of the sulfur atom of E-1– at ketenimine 2+ after protonation and reduction (Scheme 4). Isomerization to the E imines E,E-[H-6]/E,Z-[H-6] initiates the ring closing reaction to give cyclo-[H-6] with a stereogenic carbon center. Chain growth by multiple additions of ketenimine 2 and termination by protonation of the enamide determine the chain lengths of the resulting oligomers (Scheme 4). Oligomers cyclo-[H3-7]2+ (XRD, NMR, MS) and cyclo-[H -8]2+3 (MS, DFT) are interesting ferrocene containing, redox-active building blocks (Figure 9). Deprotonation of the amidinium NH•••Fe group can possibly re-activate the active zwitterionic 4,5-dihydro- 1,3-thiazolium enamide species for further anionic polymerization processes, including formation of block copolymers. These potential applications are beyond the scope of the current exploratory and mechanistic study. Figure 9. DFT optimized geometries of a) cyclo-[H -7]2+3 and b) cyclo- [H3-8]2+ (CH protons omitted for clarity). In total, the oxidants (Ag[SbF6], DDQ) serve only as initiators to form 2+, yet these oxidants are not required in stoichiometric amounts. The same arguments hold for the tasks of the base: the base is only required to initially form 2+ and 1–. This differs from the reported dehydrogenative formation of benzothiazoles from diaryl thioamides using 135 Results and Discussion stoichiometric amounts of oxidants and base.51 The net reaction of the N attack route (Scheme 4) is: 2 H-1  H-5 + H2S. The net reaction of the S attack route (Scheme 3) is the oligomerization of 2 with zwitterionic cyclo-[H-6] as initiator. n H-1  [H-1][2]n-1 + n-1 H2S The formation of zwitterionic cyclo-[H-6] is conceptually similar to the preparation of münchnones (mesoionic 1,3- oxazolium-5-olates) and other mesoionic heterocycles.47-50 To the best of our knowledge, the formation of zwitterionic 4,5-dihydro-1,3-thiazolium heterocycles52, however, has not yet been described. 3.3.4 Conclusion Coordination of ferrocenyl thioamide E/Z-[H-1] to Ag+ and subsequent electron transfer gives the ferrocenium ions E/Z-[H-1]+. Deprotonation of E/Z-[H-1] gives the anions Z-1– and E-1–. Reaction of E-1– with Ag[SbF6] gives a Ag6(E-1)6 cluster and some E-1 while Z-1– is only oxidized to Z-1. Both zwitterions Z-1 and E-1 can be stabilized as piano stool iron(III) complexes by coordination of E-1– but engage in further reactions to sulfur- deficient products by elimination of AgSH. The combined action of proton and electron abstraction from ferrocenyl thioamide H-1 leads to elimination of HS– and formation of the super-electrophilic ketenimine cation 2+. Interception of the superelectrophile 2+ by e–/H+ gives the keteniminium cation [H-2]+. Interception of 2+ by water results in formation of ferrocenium amide [H-3]+ according to mass spectrometric studies. In addition, mass spectrometry provides evidence for oligo ketenimines [(2)n+2H] +, likely featuring a [CN]n backbone and ferrocenyl side chains. Ferrocenium ketenimine 2+ is also trapped by the ambident nucleophile 1–, generated from H-1 and the base in the reaction mixture. Nucleophilic attack of the nitrogen atom of Z-1– at 2+ and trapping by H+/e– gives H-5 as a mixture of E/Z isomers. Nucleophilic attack of the sulfur atom of E-1– at 2+ and trapping by H+/e– gives H-6 as a mixture of E/Z isomers. Cyclization of H-6 yields the novel zwitterionic 4,5-dihydro-1,3-thiazolium 5- substituted heterocycle cyclo-[H-6]. We coin this heterocyclic substance class “4,5-dihydro mogone imines”.52 The anionic substituent of cyclo-[H-6] serves as initiator for the 136 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates oligomerization of ketenimine 2 to give the oligomers cyclo-[H-7] and cyclo-[H-8]. cyclo- [H3-7]2+ and cyclo-[H 2+3-8] are structurally characterized by XRD, NMR spectroscopic and/or mass spectrometric techniques. All available spectroscopic, analytical and theoretical data are consistent with the proposed mechanistic scheme. This reactivity adds a new facet to ketenimine chemistry.53 The proposed neutral and cationic N-ferrocenyl ketenimine intermediates 2/2+ could furthermore serve as useful monomers for ferrocene containing polymeric materials12-15 with [CN]n backbones. Furthermore, the ability to trigger SH – elimination from E,Z-[H-1] by oxidants in combination with bases could allow using N-ferrocenyl thioamides as structural motifs for H2S releasing prodrugs. 16 3.3.5 Experimental Section General Procedures All reactions were performed under an argon atmosphere unless otherwise noted. Dichloromethane and acetonitrile were dried with CaH2, petroleum ether (b.p. 40-60 °C) and diethyl ether were dried with sodium. All solvents were distilled prior to use. All reagents were used as received from commercial suppliers (Acros, Sigma-Aldrich, ABCR. NMR spectra were recorded on a Bruker DRX 400 spectrometer at 400.31 MHz (1H) and 100.07 MHz (13C{1H}) or on a Bruker Avance III 600 at 600.134 MHz (1H) and 150.919 MHz (13C{1H}). All resonances are reported in ppm vs the solvent signal as internal standard: CD 12Cl2 ( H, δ 5.32 ppm; 13C, δ 54.0 ppm).54 IR spectra were recorded with a BioRad Excalibur FTS 3100 spectrometer as KBr disks or by using KBr cells in CH2Cl2. Electrochemical experiments were carried out on a BioLogic SP-50 voltammetric analyzer by using a platinum working electrode, a platinum wire counter electrode, and a 0.01 M Ag/AgNO3 reference electrode. The measurements were carried out at a scan rate of 100 mV s−1 for cyclic voltammetry experiments and at 50 mV s−1 for square wave voltammetry experiments in 0.1 M [nBu4N][B(C6F5)4] as supporting electrolyte in CH2Cl2. Potentials are referenced against the decamethylferrocene/ decamethylferrocenium couple (E½ = 550 ± 5 mV vs ferrocene/ ferrocenium under our experimental conditions) and are given relative to the ferrocene/ferrocenium couple. UV/vis/near-IR spectra were recorded on a Varian Cary 5000 spectrometer by using 1.0 cm cells (Hellma, Suprasil). CW EPR spectra (X-band; ca. 9.4 GHz; ca. 20 mM) were measured on a Miniscope MS 300 at 77 K cooled by liquid 137 Results and Discussion nitrogen in a finger Dewar (Magnettech GmbH, Berlin, Germany). g factors are referenced to external Mn2+ in ZnS (g = 2.118, 2.066, 2.027, 1.986, 1.946, 1.906). Simulations of EPR spectra were performed with EasySpin (v 5.0.0) for MatLab (R2015a).55 Relative concentrations of EPR-active species (Fig. 1b, Fig. 4) were obtained by simulation and weighted combination of subspectra using EasySpin (v 5.0.0) for MatLab (R2015a).55 ESI mass spectra were recorded on a Micromass Q-TOF-Ultima spectrometer. FD+ mass spectra were recorded on a FD Finnigan MAT95 spectrometer or on a Thermo Scientific DFS with LIFDI upgrade (Linden CMS). Conductivities were measured with a Greisinger conductivity cell, model 6MH 3431 LFE-210 with platinum electrodes in the concentration range 10–3–10–4 M in CH2Cl2. The equivalent conductivity Λe was plotted as a function of c0.5, where c is the equivalent concentration. To determine Λ0, Λe was extrapolated to infinite dilution. Λ0 – Λe was plotted versus c 0.5 to obtain the Onsager plots. From this plot, the slopes of various electrolyte types can be easily compared (Supporting Information, Figures S9 and S10). Elemental analyses were performed by the microanalytical laboratory of the chemical institutes of the University of Mainz. Density Functional Theory Calculations. These were carried out with the ORCA 3.0.2/DFT series of programs.56 For geometry optimizations and energy calculations, the B3LYP formulation of density functional theory was used employing the SV(P) basis set57,58, the RIJCOSX approximation59,60, the zeroth order regular approximation (ZORA)61,62, and the KDIIS algorithm63, at GRID4 and employed Atom-pairwise Dispersion Correction as reported by Grimme.64-66 No symmetry constraints were imposed on the molecules. The presence of energy minima of the ground states was checked by numerical frequency calculations. Solvent modeling was done employing the conductor like screening model (COSMO CH 672Cl2). The approximate free energies at 298 K were obtained through thermochemical analysis of the frequency calculation, using the thermal correction to Gibbs free energy as reported by ORCA 3.0.2. Crystal Structure Determinations. Intensity data were collected with a Bruker AXS Smart1000 CCD diffractometer with an APEX II detector and an Oxford cooling system and corrected for absorption and other effects using Mo Kα radiation (λ = 0.71073 Å). The diffraction frames were integrated using the SAINT package, and most were corrected for absorption with MULABS68,69 The structures were solved by direct methods and refined by the full-matrix method based on 138 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates F2 using the SHELXTL software package.70,71 All non-hydrogen atoms were refined anisotropically, while the positions of all hydrogen atoms were generated with appropriate geometric constraints and allowed to ride on their respective parent carbon/nitrogen atoms with fixed isotropic thermal parameters. Crystal data for [Cu2(H-1)6][BF4]2: C76H86B2Cl8Cu2F8Fe6N6S6 (2195.27); T = 173 K; orange needle; 0.36 × 0.16 × 0.12 mm; triclinic; P1̅; a = 11.092(3) Å; b = 15.327(4) Å; c = 15.563(4) Å; α = 60.772(6)°; β = 71.684(6)°; γ = 78.348(6)°; V = 2188.3(10) Å3 ; Z = 1; F(000) = 1112; ξ = 1.666 g cm−3 ; µ = 1.891 mm−1 ; Θ range 1.60−28.05°; index ranges −14 ≤ h ≤ 13, −20 ≤ k ≤ 20, −20 ≤ l ≤ 20; 21797 reflections collected; 10513 independent reflections; parameters 541; maximum/minimum transmission 0.8048/0.5492; goodness of fit on F2 0.808; largest difference peak and hole 0.959/−0.697 e Å−3 ; R1(I > 2σ) = 0.0786; R1(all data) = 0.2343; wR2(I > 2σ) = 0.1278; wR2(all data) = 0.1684. Crystal data for Ag6(E-1)6: C72H72Ag6Fe6N6S6 (2196.03); T = 173 K; orange-red block; 0.21 × 0.16 × 0.01 mm; triclinic; P1̅; a = 9.7540(17) Å; b = 12.461(3) Å; c = 16.332(3) Å; α = 109.710(5)°; β = 106.676(4)°; γ = 90.193(4)°; V = 1779.1(6) Å3 ; Z = 1; F(000) = 1080; ξ = 2.050 g cm−3 ; µ = 3.017 mm−1 ; Θ range 1.75−27.96°; index ranges −12 ≤ h ≤ 12, −16 ≤ k ≤ 15, 0 ≤ l ≤ 21; 8515 reflections collected; 8515 independent reflections; 9 restraints; parameters 437; maximum/minimum transmission 0.7456/0.5652; goodness of fit on F2 1.016; largest difference peak and hole 0.923/−1.126 e Å−3 ; R1(I > 2σ) = 0.0561; R1(all data) = 0.1040; wR2(I > 2σ) = 0.0973; wR2(all data) = 0.1156. The measured crystal was refined as a two-component twin with the twin law (-1.001/0.001/0.003; -0.004/- 1.000/0.000; -1.042/-0.891/1.001) and an occupancy factor of 0.19421. Crystal data for cyclo-[H3-7][SbF6]2: C37H39C12F12Fe3N3SSb2 (1267.72); T = 173 K; brown plate; 0.180 × 0.150 × 0.040 mm; triclinic; P1̅; a = 9.6219(8) Å; b = 13.1000(11) Å; c = 17.2736(15) Å; α = 104.403(2)°; β = 92.447(2)°; γ = 96.711(2)°; V = 2088.6(3) Å3 ; Z = 2; F(000) = 1240; ξ = 2.016 g cm−3 ; µ = 2.559 mm−1 ; Θ range 1.761−28.007°; index ranges −12 ≤ h ≤ 12, −16 ≤ k ≤ 17, −22 ≤ l ≤ 22; 24230 reflections collected; 10047 independent reflections; parameters 563; maximum/minimum transmission 1.14058/0.86244; goodness of fit on F2 0.883; largest difference peak and hole 1.703/−1.535 e Å−3 ; R1(I > 2σ) = 0.0556; R1(all data) = 0.1114; wR2(I > 2σ) = 0.1172; wR2(all data) = 0.1329. 139 Results and Discussion CCDC-1475564, 1475563, 1484852 (cyclo-[H3-7], [Cu2(H-1)6][BF4]2), Ag6(E-1)6 contain supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Deprotonation and oxidation and of E/Z-[H-1] with n-butyl lithium and Ag[SbF6] (method A). E/Z-[H-1] (100 mg; 0.38 mmol; 1.0 eq.) was dissolved in THF (20 ml). n-Butyl lithium (2.5 m in n-hexane; 0.15 ml; 0.38 mmol; 1.0 eq.) was added. The solvent was removed under reduced pressure. The residue was dissolved in dichloromethane (30 ml) and Ag[SbF6] (132 mg; 0.38 mmol; 1.0 eq.) was added as a solid. The mixture was filtered through a syringe filter and the solvent was removed under reduced pressure. Re- dissolution in dichlormethane (40 ml), filtration through a syringe filter and removal of the solvent under reduced pressure yields a brown raw product. Deprotonation and oxidation and of E/Z-[H-1] with lithium dimethyl amide and Ag[SbF6] (method B). E/Z-[H-1] (100 mg; 0.38 mmol; 1.0 eq.) was dissolved in THF (20 ml). Lithium dimethylamide (19.6 mg; 0.38 mmol; 1.0 eq.) was added. The solvent was removed under reduced pressure. The residue was dissolved in dichloromethane (30 ml) and Ag[SbF6] (132 mg; 0.38 mmol; 1.0 eq.) was added as a solid. The mixture was filtered through a syringe filter and the solvent was removed under reduced pressure. Re-dissolution in CH2Cl2 (40 ml), filtration through a syringe filter and removal of the solvent under reduced pressure yields a brown raw product. Deprotonation and oxidation and of E/Z-[H-1] with tert-butylimino- tris(dimethylamino)-phosphorane and Ag[SbF6] (method C) for EPR spectroscopy. E/Z-[H-1] (1 mg; 0.0037 mmol; 1.0 eq.) was dissolved in dichloromethane (300 µl) and treated with Ag[SbF6] (1.32 mg; 0.0037 mmol; 1.0 eq.). tert-Butylimino- tris(dimethylamino)-phosphorane (0.94 µl; 0.0037 mmol; 1.0 eq.) was added and the sample was filtered through a syringe filter into an EPR tube. X-band EPR spectra were recorded at 77 K. 140 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Deprotonation and oxidation and of E/Z-[H-1] with tert-butylimino- tris(dimethylamino)-phosphorane and Ag[SbF6] (method C). E/Z-[H-1] (100 mg; 0.38 mmol; 1.0 eq.) was dissolved in dichloromethane (20 ml) and treated with Ag[SbF6] (132 mg; 0.38 mmol; 1.0 eq.). Tert-butylimino-tris(dimethylamino)- phosphorane (0.097 ml; 0.38 mmol; 1.0 eq.) was added. The mixture was filtered through a syringe filter and the solvent was removed under reduced pressure to yield a brown raw product. Deprotonation and oxidation and of E-[H-1] with 2,3-dichloro-5,6- dicyano-1,4-benzoquinone (DDQ) in CH3CN (method D) for mass spectrometry. E-[H-1] (100 mg; 0.38 mmol; 1.0 eq.) was dissolved in acetonitrile (20 ml) and treated with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (41 mg; 0.18 mmol; 0.5 eq.). The mixture was stirred for 30 minutes and cobaltocene (73 mg; 0.38 mmol; 1.0 eq.) was added as a solid resulting in H2S evolution. The raw product was purified by column chromatography (315 cm, SiO2, ethyl acetate) giving a brown powder. Deprotonation and oxidation and of E-[H-1] with 2,3-dichloro-5,6- dicyano-1,4-benzoquinone (DDQ) in CH3CN (method D) for EPR spectroscopy. E-[H-1] (1 mg; 0.0037 mmol; 1.0 eq.) was dissolved in acetonitrile (300 µl) and treated with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (0.41 mg; 0.0018 mmol; 0.5 eq.). X-band EPR spectra were recorded at 77 K. Preparative synthesis of H-5 (isomer mixture). Method B was applied, starting with E/Z-[H-1] (50 mg, 0.19 mmol; 1.0 eq.) The raw product was purified by column chromatography (330 cm, Al2O3 Brockmann II, CH2Cl2/PE 1:1) giving a yellow powder. Yield 32 % (13.2 mg, 0.03 mmol). 1H NMR (400 MHz, CD2Cl2, 298 K): δ = 4.59 (ddd, 1H, H β1,imine/Hβ2,imine), 4.38 (ddd, 1H, Hβ1,imine/Hβ2,imine), 4.23 (ddd, 2H, Hα1,imine/Hα2,imine), 4.21 (s, 5H, HCp,imine/HCp,amide), 4.16 (ddd, 1H, Hβ1,amide/Hβ2,amide), 4.09 (ddd, 1H, Hβ1,amide/Hβ2,amide), 4.09 (ddd, 1H, Hα1,amide/Hα2,amide), 4.07 (s, 5H, HCp,imine/HCp,amide), 4.06 (ddd, 1H, Hα1,amide/Hα2,amide), 3.64 (ddd, 1H, Hα1,amide/Hα2,amide), 2.53 (s, 3H, CH imine3 ), 2.19 (s, 3H, CH amide 3 ) ppm (major 141 Results and Discussion isomer, 80 %); δ = 2.65 (s, 3H, CH amide3 ), 2.38 (s, 3H, CH imine 3 ) ppm (one minor isomer, 12 %). 13C{1H} NMR (100 MHz, CD2Cl2, 298 K): δ = 204.3 (C=S), 151,9 (C=N), 99.4 (C i, imine), 93.8 (Ci, amide), 71.2 (Cpimine/Cpamide), 70.4 (Cβ1, imine/Cβ2, imine), 70.3 (Cpimine/Cpamide), 67.7 (Cα1, imine/Cα2, imine), 67.0 (Cα1, imine/Cα2, imine), 66.8 (Cβ1, amide/Cβ2, amide/Cα1, amide/Cα2, amide), 66.5 (Cα1, amide/Cα2, amide), 60.4 (Cβ1, imine/Cβ2, imine), 33.1 (CH amide), 24.1 (CH imine3 3 ) ppm (major isomer); δ = 149.59 (Cimine) ppm (one minor isomer). MS(FD+): m/z (%) = 484.19 ([H-5]+). Preparative synthesis of cyclo-[H3-7][SbF6]2. Method A was applied. The raw product was purified by column chromatography (330 cm, SiO2, CH2Cl2  ethyl acetate) giving a brown powder. Yield 6 % (7.2 mg, 0.008 mmol; based on cyclo-[H3-7][SbF6]2). The product also contains some cyclo-[H3-8][SbF6]2. A few crystals of cyclo-[H3-7][SbF6]2 suitable for X-Ray diffraction were obtained by evaporating a saturated dichloromethane solution of the raw product. 1H NMR (600 MHz, CD2Cl2, 298 K): δ = 7.99 (s, 1H, H NH), 4.63 (pt, 2H, Hα1/β1), 4.51 (pt, 2H, Hα2/β2), 4.47 (pt, 2H, Hα1/β1), 4.46 (s, 5H, HCp1/ Cp2/ Cp3), 4.42 (s, 5H, H Cp1/ Cp2/ Cp3), 4.37 (s, 5H, HCp1/ Cp2/ Cp3), 4.40 (pt, 2H, Hα2/β2), 4.34 (pt, 2H, Hα3/β3), 4.31 (s, 1H, HCH), 4.25 (pt, 2H, Hα3/β3), 4.10 (s, 2H, HCH2), 2.58 (s, 2H, NH-iminium-CH3), 2.58 (s, 2H, Himinium-CH3) ppm. 13C{1H} NMR (150 MHz, CD2Cl2, 298 K): δ = 160.2 (C iminium-C), 154.8 (CNH-iminium-C), 71.2 (CCp1/Cp2/Cp3), 71.1 (CCp1/Cp2/Cp3), 70.6 (CCp1/Cp2/Cp3), 70.0 (CCH2), 68.0 (Cα2/β2), 68.4 (Cα1/β1), 67.0 (Cα3/β2), 66.2 (Cα1/β3), 66.0 (Cα3/β3), 65.4 (CCH), 64.7 (Cα2/β2) ppm. MS(FD+): m/z (%) = 707.34 (35, {cyclo-[H-7]-2H}+), 932.67 (100, {cyclo-[H-8]-2H}+). Synthesis of [Cu2(H-1)6][BF4]2; bis-{µ-(κS-(E-[N-thioacetyl ferrocenylamine])) bis[di-(κS-Z-[N-thioacetyl ferrocenyl-amine])- copper(I)}. Tetrakis(acetonitrile) copper(I) tetrafluoroborate (307 mg; 0.975 mmol, 3.0 eq.) and copper powder (125.0 mg; 0.375 mmol; 1.0 eq.) were suspended in CH3CN (50 ml) and stirred at room temperature for 24 hours. E/Z-[H-1] (250 mg; 0.975 mmol; 1.0 eq.) was added to the suspension and the mixture was heated to reflux for one hour. The suspension was slowly cooled to room temperature and diethyl ether (200 ml) was added. A colorless precipitate formed and was filtered off. The solvent of the filtrate was removed under reduced pressure. An orange oil was obtained, which was dissolved in CH2Cl2 (100 ml) and precipitated by 142 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates addition of petroleum ether (bp. 40-60 °C; 100 ml). After removal of the solvent mixture under reduced pressure an orange-red powder was obtained. Single crystals were obtained from a CH2Cl2 solution of [Cu2(H-1)6][BF4]2. Yield. 31 % (93.5. mg, 0.51 mmol). 1H NMR (400 MHz, CD2Cl2, 298 K): δ = 10.38 (s, 2H, NH of E-[H-1]), 10.09 (s, 4H, NH of Z-[H- 1]), 4.94 (pt, 4H, 3JHH = 1.91 Hz, H α of E-[H-1]), 4.42 (pt, 8H, 3JHH = 1.94 Hz, H α of Z-[H- 1]), 4.35 (s, 20H, CpH of Z-[H-1]), 4.33 (s, 10H, CpH of E-[H-1]), 4.25 (pt, 4H, 3JHH = 1.91 Hz, Hβ of E-[H-1]), 4.24 (pt, 8H, 3JHH = 1.94 Hz, H β of Z-[H-1]), 2.81 (s, 6H, CH3 of E-[H-1]), 2.48 (s, 12H, CH3 of Z-[H-1]) ppm. 13C{1H} NMR (100 MHz, CD2Cl2, 298 K): δ = 200.6 (C=S of Z-[H-1]), 198.1 (C=S of E-[H-1]), 94.0 (Ci of Z-[H-1]), 93.8 (Ci of E- [H-1]), 70.7 (Cp of Z-[H-1]), 70.6 (Cp of E-[H-1]), 67.7 (Cβ of E-[H-1]), 67.3 (Cβ of Z- [H-1]), 65.7 (Cα of Z-[H-1]), 65.6 (Cα of E-[H-1]), 34.7 (CH3 of E-[H-1]), 28.6 (CH3 of Z- [H-1]) ppm. MS(ESI+): m/z (%) 321.93 (40) [Cu(H-1)]+, 580.90 (100) [Cu(H-1) +2] , 839.97 (3) [Cu(H-1)3] +, 1250.91 (100) [Cu2(H-1) BF ] + 4 4 . IR (KBr): 𝜈 = 3281 (m, NH), 3198 (w, NH), 1548 (b, CS(I)), 1384 (b, CS(II)) cm−1. IR (CH2Cl2): 𝜈 = 3361 (w, NH), 3267 (m, NH), 3209 (m, NH), 1531 (m, CS(I)) cm−1. UV/vis (CH2Cl2): λmax (ε) 256 (7430), 313 (7525), 378 (2715), 456 nm (755 M−1 cm−1). CV (CH2Cl2, vs FcH/ FcH +): E½ = 0.140 V (rev), 0.350 V (rev), Ep = 0.700 V (irrev). Λ 0 m(CH2Cl2) = 43 S cm –1 mol–1. Anal. Calcd for C72H78Fe6N6S6B2F8Cu2 (1855.66)Et2O: C, 47.30; H, 4.50; N, 4.36; S, 9.97. Found: C, 47.43; H, 4.80; N, 4.55; S, 10.25. Synthesis of Ag6(E-1)6. E/Z-[H-1] (100 mg; 0.38 mmol; 1.0 eq.) was dissolved in dichloromethane (20 ml) and treated with Ag[SbF6] (132 mg; 0.38 mmol; 1.0 eq.). tert-Butylimino-tris(dimethylamino)- phosphorane (97 µl; 0.38 mmol; 1.0 eq.) was added. The mixture was filtered through a syringe filter and the solvent slowly evaporated under inert gas conditions at standard pressure for several days giving red crystals, which were separated from the mother liquor by decantation. Yield: 3.7 % (5.2 mg, 0.0024 mmol). IR (KBr): ν ̃ = 1562 (m, CS(I)), 1357 (s, CS(II)) cm−1. 143 Results and Discussion 3.3.6 Associated Content Supporting Information Description of the XRD and solution structure of [Cu2(H-1)6][BF4]2 and spectra of the new compounds (PDF). Crystal structure data for compounds [Cu2(H-1)6][BF4]2, Ag6(E-1)6 and cyclo-[H3- 7][SbF6]2 (cif). Coordinates of DFT calculated compounds (xyz). Author Information Corresponding Author *E-mail for K.H.: katja.heinze@uni-mainz.de. Notes The authors declare no competing financial interest. 3.3.7 Acknowledgement Parts of this research were conducted using the supercomputer Mogon and advisory services offered by Johannes Gutenberg University Mainz (www.hpc.uni-mainz.de), which is a member of the AHRP and the Gauss Alliance e.V.. We thank Petra Auerbach and Dr. Mihail Mondeskhi for collecting the LIFDI mass spectra and Regine Jung-Pothmann for collection of the diffraction data. Financial support from the EU-COST Action CM1302 SIPs is gratefully acknowledged. 3.3.8 References [1] Heinze, K.; Schlenker, M.; Eur. J. Inorg. Chem. 2004, 2974–2988. [2] Heinze, K.; Schlenker, M. Eur. J. Inorg. Chem. 2005, 66–71. [3] Siebler, D.; Förster, C.; Heinze, K. Eur. J. Inorg. Chem. 2010, 523–527. [4] Siebler, D.; Linseis, M.; Gasi, T.; Carrella, L. M.; Winter, R. F.; Förster, C.; Heinze, K. Chem. Eur. J. 2011, 17, 4540–4551. [5] Siebler, D.; Förster, C.; Heinze, K. Dalton Trans. 2011, 40, 3558–3575. [6] Kienz, T.; Förster, C.; Heinze, K. Organometallics 2014, 33, 4803–4812. [7] Neidlinger, A.; Kienz, T.; Heinze, K. Organometallics 2015, 34, 5310–5320. 144 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates [8] Neidlinger, A.; Ksenofontov, V.; Heinze, K. Organometallics 2013, 32, 5955– 5965. [9] Neidlinger, A.; Förster, C.; Heinze, K. Eur. J. Inorg. Chem. 2016, 1274–1286. [10] Siebler, D.; Förster, C.; Gasi, T.; Heinze, K. Organometallics 2011, 30, 313–327. [11] Hüttinger, K.; Förster, C.; Heinze, K. Chem. Commun. 2014, 50, 4285–4288. [12] Pietschnig, R. Chem. Soc. Rev. 2016, in press (DOI: 10.1039/c6cs00196c). [13] Gao, Y.; Shreeve, J. M. J. Inorg. Organomet. Polym. Mater. 2007, 17, 19–36. [14] Abd-El-Aziz, A.; Manners, I. J. Inorg. Organomet. Polym. Mater. 2005, 15, 157– 195. [15] Gallei, M.; Tockner, S.; Klein, R.; Rehahn, M. Macromol. Rapid Commun. 2010, 31, 889–896. [16] Zheng, Y.; Yu, B.; Ji, K.; Pan, Z.; Chittavong, V.; Wang, B. Angew. Chem. 2016, 128, 4590–4594; Angew. Chem. Int. Ed. 2016, 55, 4514–4518. [17] Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877−910. [18] Kloss, F.; Pidot, S.; Goerls, H.; Friedrich, T.; Hertweck, C. Angew. Chem. 2013, 125, 10945–10948; Angew. Chem. Int. Ed. 2013, 52, 10745–10748. [19] Constable, E. C.; Raithby, P. R.; J. Chem. Soc., Dalton Trans. 1987, 2281–2283. [20] Kokkou, S. C.; Fortier, S.; Rentzeperis, P. J.; Karagiannidis, P. Acta Cryst. 1983, C39, 178-180. [21] Lobana, T. S.; Castineiras, A. Polyhedron 2002, 21, 1603–1611. [22] Feltham, R. D.; Hayter, R. G. J. Chem. Soc. 1964, 4587–459. [23] Su, W.; Hong, M.; Weng, J.; Liang, Y.; Zhao, Y.; Cao, R.; Thou, Z.; Chan, A. S. C. Inorg. Chim. Acta 2002, 331, 8–15. [24] Schwesinger, R.; Hasenfratz, C.; Schlemper, H.; Walz, L.; Peters, E.-M., Peters, K.; von Schnering, H.G. Angew. Chem. 1993, 105, 1420–1422; Angew. Chem. Int. Ed. 1993, 32, 1361–1363. [25] García-Vázquez, J. A.; Sousa-Pedrares, A.; Carabel, M.; Romero, J.; Sousa, A. Polyhedron 2005, 24, 2043–2054. [26] Sun, Q.; Chai, L.; Liu, H.; Wang, J. Acta Cryst. 2013, E69, m50–m51. [27] Ashfield, L. J.; Cowley, A. R.; Dilworth, J. R.; Donnelly, P. S. Inorg. Chem. 2004, 43, 4121–4123. 145 Results and Discussion [28] Yue, C.; Yan, C.; Feng, R.; Wu, M.; Chen, L.; Jiang, F.; Hong, M. Inorg. Chem. 2009, 48, 2873–2879. [29] Elschenbroich, C.; Bilger, E.; Ernst, R. D.; Wilson, D. R.; Kralik, M. S. Organometallics 1985, 4, 2068-2070. [30] Costuas, K.; Paul, F.; Toupet, L.; Halet, J.-F.; Lapinte, C. Organometallics 2004, 23, 2053–2068. [31] Lohan, M.; Justaud, F.; Lang, H.; Lapinte, C. Organometallics 2012, 31, 3565– 3574. [32] Lecomte, M.; Evano, G. Angew. Chem. 2016, 128, 4623–4627; Angew. Chem. Int. Ed. 2016, 55, 4547–4551. [33] Madelaine, C.; Valerio, V.; Maulide, N. Chem. Asian J. 2011, 6, 2224–2239. [34] Evano, G.; Coste, A.; Jouvin, K. Angew. Chem. 2010, 122, 2902–2921; Angew. Chem. Int. Ed. 2010, 49, 2840–2859. [35] Yasui, H.; Yorimitsu, H.; Oshima, K. Chem. Lett. 2008, 37, 40–41. [36] Kanemura, S.; Kondoh, A.; Yasui, H.; Yorimitsu, H.; Oshima, K. Bull. Chem. Soc. Jpn. 2008, 81, 506–514. [37] Ding, W.; Fang, D. J. Org. Chem. 2001, 66, 6673–6678. [38] Bonini, B. F.; Femoni, C.; Comes-Franchini, M.; Fochi, M.; Mazzanti, G.; Ricci, A.; Varchi, G. Synlett 2001, 1092–1096. [39] Aguilar-Aguilar, C.; Allen, A. D.; Cabrera, E.; Federov, A. V.; Fu, N.; Henry- Riyad, H.; Kobayashi, S.; Leuninger, J.; Tidwell, T. T. J. Org. Chem. 2005, 70, 9556–9561. [40] Fu, N.; Allen, A. D.; Tidwell, T. T. ARKIVOC 2007, iii, 143–150. [41] Dewey, F. Tetrahedron Lett. 1968, 4207–4208. [42] Tidwell, T. T. Ketenes II, 2006, John Wiley & Sons, Inc., Hoboken, New Jersey. [43] Förster, C.; Veit, P.; Ksenofontov, V.; Heinze, K. Chem. Commun. 2015, 51, 1514–1516. [44] Veit, P.; Förster, C.; Seibert, S.; Heinze, K. Z. Anorg. Allg. Chem. 2015, 641, 2083–2092. [45] Veit, C.; Prantl, E.; Förster, C.; Heinze, K. Organometallics 2016, 35, 249–257. [46] Hanauer, K.; Pham, M. T.; Förster, C.; Heinze, K. Eur. J. Inorg. Chem. in press (DOI: 10.1002/elic.201600918]). 146 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates [47] Reissig, H.-U.; Zimmer, R. Angew. Chem. 2014, 37, 9708-9710; Angew. Chem. Int. Ed. 2014, 53, 9708-9710. [48] Huisgen, R.; Gotthardt, H.; Bayer, H. O.; Schaefer, F. C. Angew. Chem. 1964, 76, 185-186; Angew. Chem. Int. Ed. 1964, 3, 136-137. [49] Greco, C.; V.; Nyberg, W. H.; Cheng, C. C. J. Med. Chem. 1962, 5, 861–865. [50] Cyr, D. J. S.; Arndtsen, B. A. J. Am. Chem. Soc. 2007, 129, 12366–12367. [51] Wang, H.; Wang, L.; Shang, J.; Li, X.; Wang, H.; Guia, J.; Lei, A. Chem. Commun. 2012, 48, 76–78. [52] In analogy to the so-called “münchnones”, “sydnones” and “montrealones” (mesoionic O,N-, O,N,N- and N,P-five-membered heterocycles) explored and exploited by research groups in Munich (München), Sydney and Montreal, we suggest the name “mogones” (Latin Mogontiacum for Mainz) for mesoionic 1,3- thiazolium-5-enolates and “4,5-dihydro mogone imines” for 4,5-dihydro-1,3- thiazolium-5-enamides such as cyclo-[H-6], respectively. [53] M. Alajarin, M. Marin-Luna, A. Vidal, Eur. J. Org. Chem. 2012, 5637–5653. [54] Fulmer, G. R.; Miller, A. J. M.; Sherden, N. H.; Gottlieb, H. E.; Nudelman, A.; Stoltz, B. M.; Bercaw, J. E.; Goldberg, K. I. Organometallics 2010, 29, 2176– 2179. [55] Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42–55. [56] Neese, F. WIREs Comput Mol Sci, 2012, 2, 73–78. [57] Schäfer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571. [58] Schäfer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829. [59] Neese, F.; Wennmohs, F.; Hansen, A.; Becker, U. Chem. Phys. 2009, 356, 98– 109. [60] Izsák, R.; Neese, F. J. Chem. Phys. 2011, 135, 144105. [61] van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1993, 99, 4597. [62] van Wüllen, C. J. Chem. Phys. 1998, 109, 392. [63] Kollmar, C. J. Chem. Phys. 1996, 105, 8204. [64] Grimme, S. J. Comput. Chem. 2006, 27, 1787–1799. [65] Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys. 2010, 132, 154104. [66] Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput. Chem. 2011, 32, 1456–1465. 147 Results and Discussion [67] Sinnecker, S.; Rajendran, A.; Klamt, A.; Diedenhofen, M.; Neese, F. J. Phys. Chem. A 2006, 110, 2235–2245. [68] SMART Data Collection and SAINT-Plus Data Processing Software for the SMART System (various versions); Bruker Analytical X-Ray Instruments, Inc., Madison, WI, 2000. [69] Blessing, R. H. Acta Crystallogr., Sect. A 1995, 51, 33–38. [70] Sheldrick, G. M. SHELXTL, version 5.1, Bruker AXS, Madison, WI, 1998. [71] Sheldrick, G. M. SHELXL-97, University of Göttingen, Germany, 1997. 148 4 Summary and Outlook The successful preparation and analyses of mono und dinuclear N-ferrocenyl amides is presented and a comparison to parent carboxamides is made to gain insight on the direct effect of oxygen → sulfur exchange. The synthesis reported with Lawessons reagent allows easy access to a variety of N-ferrocenyl thioamides. The use of the ferrocenyl carboxamides as starting materials allows variation of the substitution pattern of the thioamides through the well-established chemistry of the starting materials. Unlike the carboxamides the thioamides feature a high enough rotational barrier (75 kJ mol−1) to be observed E/Z- isomers in solution at room temperature. Similar to carboxamides oxidation of the thioamides leads to the expected ferrocenium compounds. However, in contrast to the carboxamides, a second oxidation process at higher potentials is observed, most likely attributed to the oxidation of the electron rich sulfur atom of the thioamide unit. Dinuclear compounds are transformed into the mixed-valent compounds upon single oxidation. These mixed-valent compounds could be assigned to Robin-Day class II. Compared to the parent carboxamides the electronic coupling is almost the same for the thioamides, but the shift of the IVCT band to lower energies suggests facilitated thermal electron transfer, due to a smaller activation barrier originating from the more similar redox potentials of the ferrocenyl moieties. The dinuclear compound featuring two thioamide units features a NH•••S hydrogen bond in solution forming a six-membered ring in 1,2’-conformation, contrary to the eight- membered ring of the parent carboxamide. Upon oxidation the secondary structure changes to an eight-membered ring in 1,2’-conformation, according to DFT calculations. Alltogether N-ferrocenyl thioamides are promising building blocks for oligomers and polymers, featuring a distinct secondary structure in solution. The conformational change of the hydrogen bonding motif from an eight-membered ring of the parent carboxamides to the six-membered ring in the thioamides opens new applications as turn mimetics. Especially the conformational change from a six-membered ring to an eight-membered ring upon oxidiation can attribute to research of redox triggered peptidmimetika for use in pharmaceutical applications. Pharmaceutical applications of ferrocene have been explored in various compounds including ferrocifen, ferroquine, aminoferrocene based (pro-)drugs and anti leishmanial, antibacterial, antifungal drugs and vasorelaxants containing ferrocen. One possible mode 149 Summary and Outlook of action for ferrocene based (pro-)drugs centers around the oxidation of the ferrocenyl moiety to ferrocenium. In this work the formation of ferrocenyl radicals of various ferrocene derivatives under the addition of base is reported. The addition of nitrosobenenze as spin trapping agent allowed the detection of highly reactive ferrocenyl radicals and facile determination of the radical location. Aminoacetyl and thioaminoacetyl substitution of the ferrocenyl moiety lead to stabilization of the oxidized and deprotonated compounds as ferrocenium iminolate and ferrocenium thioiminolate, while the unsubstituted and alkyl substituted ferrocene derivatives undergo rapid back reaction by hydrogen abstraction from the base. These iron centered radicals do not react with nitrosobenzene, however even these stabilized ferrocene derivatives can be analyzed by the spin trapping technique, showing that the ferrocenium iminolate and thioiminolate serve as reservoir for the highly reactive carbon centered radicals. This radical reservoir may contribute to the activity of amino ferrocene based (pro-)drugs. This work clearly shows the benefit of the spin trapping technique in order to disclose the mode of action for certain (pro-)drugs. Thorough investigation may lead to new insight and development of more specifically targeting pharmaceutical compounds. Apart from the reported radical activity for N-thioacetylamino substituted ferrocene an additional reaction pathway upon oxidation and deprotonation of N-ferrocenyl substituted thioamides is reported in this work. A crucial factor for this additional pathway is the ability of sulfur to act as a ligand, which is shown by the formation of copper(I) and silver(I) complexes of N-ferrocenyl substituted thioamides and thioiminolates. This enables the ferrocenium thioiminolates to stabilize themselves as piano stool like complexes by coordination of the adjacent thioiminolate sulfur, as well as another thioiminolate functional group from a second molecule to the iron nucleus of the ferrocenium moiety. The substituted cyclopentadienyl therefore undergoes a change in hapticity from η5 to η1. Finally, elimination of hydrosulfide from the molecule and backshift of the substituted cyclopentadienyl ring to η5 hapticity leads to the formation of super electrophilic ketenimine cations. These react rapidly by either oligomerization or are attacked by a thioiminolate. The second reaction pathway leads to the formation of a novel zwitterionic 4,5-dihydro-1,3-thiazolium 5-substituted heterocycle, the “4,5-dihydro mogone imines”. The reaction of N-ferrocenyl thioamides upon oxidation and deprotonation by loss of sulfur may lead to novel (pro-)drugs with a mode of action based on the release of H2S. 150 The proposed ketenimine may provide interesting oligomers and polymers with a [CN]n backbone and ferrocenyl side chains. Finally, possible applications of the novel hetrocyclic zwitterionic “4,5-dihydro mogone imines” cannot be foreseen and have yet to be unveiled. 151 Summary and Outlook 152 5 Supporting Information Contents To 3.1 Impact of OS Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Torben Kienz, Christoph Förster and Katja Heinze Published in: Organometallics 2014, 33, 4803–4812. [DOI: 10.1021/om500052k] http://pubs.acs.org/doi/abs/10.1021/om500052k “Adapted with permission from T. Kienz, C. Förster, K. Heinze, Organometallics 2014, 33, 4803–4812. Copyright 2014 American Chemical Society.” To 3.2 Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Andreas Neidlinger, Torben Kienz and Katja Heinze Published in: Organometallics 2015, 34, 5310–5320. [DOI: 10.1021/acs.organomet.5b00778] http://pubs.acs.org/doi/10.1021/acs.organomet.5b00778 “This is an unofficial adaptation of an article that appeared in an ACS publication. ACS has not endorsed the content of this adaptation or the context of its use. Copyright 2015 American Chemical Society.” To 3.3 Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open- shell Intermediates Torben Kienz, Christoph Förster and Katja Heinze Submitted to: Organometallics 153 Supporting Information 154 To 3.1: Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry 5.1 To 3.1: Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Torben Kienz, Christoph Förster, and Katja Heinze Organometallics 2014, 33, 4803–4812. Adapted with permission from T. Kienz, C. Förster, K. Heinze, Organometallics 2014, 33, 4803–4812. Copyright 2014 American Chemical Society. 155 Supporting Information Figure S1. UV/Vis spectra of 1 and 3 in CH2Cl2. Figure S2. UV/Vis spectra of 2 and 4 in CH2Cl2. 156 To 3.1: Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Figure S3. UV/Vis spectra of 5 and 7 in CH2Cl2. Figure S4. UV/Vis spectra of 6 and 8 in CH2Cl2. 157 Supporting Information Figure S5. 1H1H NOESY of 3 in CD2Cl2 at room temperature. Figure S6. DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated (cis-3)2. 158 To 3.1: Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Figure S7. DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated cis-7 and trans-7. Figure S8. 1H1H NOESY of 7 in CD2Cl2. at room temperature. 159 Supporting Information Figure S9. DFT (B3LYP, LANL2DZ, PCM CH2Cl2) calculated trans-3n (n = 0, 1, 2). Figure S9. Gaussian band shape analysis of IVCT bands of 7+. Figure S10. Gaussian band shape analysis of IVCT bands of 8+. 160 To 3.1: Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry Figure S11. 1H NMR spectra of 8 upon titration with iodine in CD2Cl2. Figure S12. EPR spectrum of 3 in THF/CH2Cl2 (1:4) and simulation. 161 Supporting Information Figure S13. EPR spectra of 4 in THF/CH2Cl2 (1:4) and simulation. 162 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene 5.2 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Andreas Neidlinger, Torben Kienz, and Katja Heinze Organometallics 2015, 34, 5310–5320. Adapted with permission from A. Neidlinger, T. Kienz, K. Heinze, Organometallics 2015, 34, 5310–5320. “This is an unofficial adaptation of an article that appeared in an ACS publication. ACS has not endorsed the content of this adaptation or the context of its use. Copyright 2015 American Chemical Society. Copyright 2014 American Chemical Society.” 163 Supporting Information Figure S1: DFT optimized geometry with spin density (0.01 a.u. isosurface value) in CH2Cl2 continuum solvent for iron spin-trapped [1-PhNO]+. Figure S2: (top) Square wave and (bottom) cyclic voltammogram of P t1 Bu in CH2Cl2 containing [nBu4N][B(C6F5)4] as supporting electrolyte at 298 K. 164 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure S3: DFT optimized geometries with spin densities (0.01 a.u. isosurface value) in CH2Cl2 continuum solvent for a) ferrocenyl radical [1−H]• and b) decamethylferrocenyl radical [2−H]•. 165 Supporting Information Figure S4: X-band EPR spectrum (top) and simulated spectrum (bottom) of a) [6]•, b) [7]•, and c) [81]• (25 mM starting ferrocene in CH2Cl2) at the following experimental parameters: a) and b) temperature = 77 K, field = 3346.20 G, sweep = 499.77 G, sweep time = 90 s, modulation = 5000 mG, MW attenuation = 10 db and c) temperature = 77 K; field = 3346.20 G, sweep = 499.77 G, sweep time = 90 s, modulation = 1000 mG, MW attenuation = 5 db. 166 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure S5: a) (top) Square wave and (bottom) cyclic voltammogram of 1 and b) cyclic voltammograms of a 1/P t1 Bu mixture. All measurements performed in CH2Cl2 containing [nBu4N][B(C6F5)4] as supporting electrolyte at 298 K. 167 Supporting Information Figure S6: DFT optimized geometries with spin densities (0.01 a.u. isosurface value), Lewis structures, and energies in CH2Cl2 continuum solvent for radicals [3−Hx]● (x = α, β, Cp, 1, 2). 168 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure S7: DFT optimized geometries with spin densities (0.01 a.u. isosurface value) and energies in CH2Cl2 continuum solvent as well as Lewis structures for nitroxide radicals [8x]● (x = α, β, Cp, 1, 2). 169 Supporting Information Figure S8: X-band EPR spectrum (top) and simulated spectrum (bottom) of a mixture of [4−HN]• and [4−Hx]• (x = , , Cp, Me) (5 mM 4+ + P t1 Bu in CH2Cl2) at the following experimental parameters: a) temperature = 77 K, field = 3346.20 G, sweep = 499.77 G, sweep time = 90 s, modulation = 5000 mG, MW attenuation = 10 db and b) temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 1000 mG, MW attenuation = 10 db. 170 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure S9: DFT optimized geometries with spin densities for [4−Hx]• (x = , , Cp, N, Me) (0.01 a.u. isosurface value) and energies in CH2Cl2 continuum solvent for [4−Hx]•, [5−Hx]• (x = , , Cp, N, Me), and [4−HN’]• as well as Lewis structures. 171 Supporting Information Figure S10: a) (top) Square wave and (bottom) cyclic voltammogram of 4 and b) cyclic voltammograms of a 4/P t1 Bu mixture; red, dashed curve represents a scan excluding the irreversible oxidation. All measurements performed in CH2Cl2 containing [nBu4N][B(C6F5)4] as supporting electrolyte at 298 K. Figure S11: Experimental, baseline corrected X-band EPR spectra and values of double integration of EPR spectra of (a) [6]•, (b) [7]•, (c) [81]•, d) [9x]• (x = β, Cp), and e) [10Cp]• (5 mM starting ferrocene in CH2Cl2) recorded at the following experimental parameters: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 1000 mG, MW attenuation = 10 db. 172 To 3.2: Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene Figure S12: Experimental, baseline corrected X-band EPR spectra and values of double integration of EPR spectra of (a) 0.03 mM, (b) 0.01 mM, and (c) 0.005 mM DPPH solutions in CH2Cl2 recorded at the following experimental parameters: temperature = 298 K, field = 3346.20 G, sweep = 94.79 G, sweep time = 90 s, modulation = 1000 mG, MW attenuation = 10 db. (d) Linear regression of double integral values against concentration. 173 Supporting Information 174 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates 5.3 To 3.3: Generation and Oligomerization of N- Ferrocenyl Ketenimines via Open-shell Intermediates Torben Kienz, Christoph Förster and Katja Heinze* 175 Supporting Information Discussion of XRD of [Cu2(H-1)6][BF4]2: Copper : thioamide ratios of 1:3 have also been observed in the copper(I) complex with the biologically relevant hexathioamide closthioamide (CTA) [Cu2(CTA)](ClO4) , 1 2 in [Cu2(HpyS)6]Cl2, 2 and in [Cu(HpyS)3](NO3)3 (HpyS = 2(1H)-pyridinethione). [Cu2(H- 1) ]2+6 features two copper centers bridged by two κS coordinating E-[H-1] ligands forming a centrosymmetric Cu2S2 diamond core. Four terminal κS coordinating Z-[H-1] ligands complete the coordination sphere of the two distorted tetrahedrally coordinated copper(I) centers (for bond angles see Table S1). The Cu•••Cu distance amounts to 3.410 Å, which is quite long compared to dinuclear copper(I) complexes [Cu2(HpyS) ] 2+ 6 with HpyS bridging ligands sharing the same Cu2S2 core motif (Cu•••Cu 2.907(3) Å). 2 The elongation can be attributed to the six sterically demanding ferrocenyl moieties surrounding the Cu2S2 core. The Cu-S distances are similar to those reported for [Cu 2+2(HpyS)6] with smaller thioamide ligands (2.284(2) – 2.498(3) Å and Table S1).2 Expectedly, the C=S distances increase from 1.671(2) Å in E-[H-1] to 1.689(8) – 1.707(8) Å in the copper(I) complex. Multiple NH•••S and NH•••F hydrogen bonds are formed in the solid state (Figure 2a). The NH group of the bridging E-[H-1] ligand forms a hydrogen bond to the S atom of an adjacent terminally coordinated Z-[H-1] ligand (Table S1, Figure 2a). An analogous NH•••S hydrogen bonding motif has been found in [Cu (HpyS) ]2+,22 6 while in [Cu2Br2(µ- HpyS)2(PTol3) 2+ 2] the NH•••Br hydrogen bond from the bridging thioamides includes the coordinated bromide ligands.4 The NH groups of the four terminally coordinated Z-[H-1] ligands form hydrogen bonds to the [BF –4] counter ions (Table S1, Figure 2a). Similarly, [Cu2(HpyS)6]Cl2 forms NH•••Cl hydrogen bonds to its counter ions. 2 As all four terminally coordinated Z-[H-1] ligands form hydrogen bonds to the counter ions, the anions bridge the [Cu2(H-1) ]2+6 dications to give a two-dimensional polymer (Supporting Information, Figure S2). 176 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Table S1. Selected bond lengths (Å) and bond angles (°) of [Cu2(H-1)6][BF4]2. Cu1-S1 2.290(3) N1•••F2 2.933(8) Cu1-S2 2.309(2) N1•••F4’’’ 3.253(9) Cu1-S2’ 2.612(2) N3•••F1’’ 2.936(8) Cu1-S3 2.305(2) N3•••F3’’ 3.352(9) S1-C11 1.689(8) N2•••S3 3.304(7) S2-C23 1.700(9) Cu1•••Cu1’ 3.410(2) S3-C35 1.707(8) Cu1-S2-Cu1’ 87.5(8) S1-Cu1-S2 108.26(9) S2-Cu1-S3 121.97(9) S1-Cu1-S2’ 112.09(9) S2’-Cu1-S3 94.92(8) S1-Cu1-S3 121.34(9) S2-Cu1-S2’ 92.50(8) Solution structure of [Cu2(H-1)6][BF4]2: NMR spectra of [Cu2(H-1)6][BF4]2 in CD2Cl2 show two sets of resonances for coordinated H-1 in a 2:1 ratio for the Z and E isomers, respectively (Supporting Information, Figures S3 – S8). As all terminal Z-[H-1] ligands with and without a hydrogen bond to the bridging E-[H-1] ligands appear as a single set of resonances on the NMR time scale, the NH•••S hydrogen bonds flip rapidly on the 1H NMR time scale between the terminal C=S acceptor sites by rotation round the C=S bond of the bridging ligand (Figure 2a). Apart from the expected intraligand NOE contacts of the NH, CpH and CH3 groups, several interligand NOE contacts are observed, most notably from the NH group of the terminal Z-[1-H] ligands to the CH3 groups of the bridging E-[H-1] ligands. This suggests that the [Cu2(H- 1)6]2+ core remains essentially intact in solution. Furthermore, EXSY signals between the NH groups of the E and Z ligands are observed suggesting chemical exchange. In principle, the bridging and terminal ligands can exchange their positions by a dissociation – isomerization - association mechanism, by a pseudorotation mechanism involving isomerization or by intra- or intermolecular proton transfer processes. As the barrier for E/Z isomerization of H-1 5 amounts to 75 kJ mol–1, the former processes should be rather slow and hence, proton transfer might account for this exchange process. Conductivity measurements in CH2Cl2 suggest that the dimeric structure is retained and even the coordination of the counter ions is highly preserved. The limiting conductivity and 177 Supporting Information the slope of the Onsager plot of the formal 1:2 electrolyte [Cu2(H-1)6][BF4]2 in CH2Cl2 are even below those of 1:1 electrolytes such as [nBu4N][B(C6F5)4] (Supporting Information, Figures S9 and S10).6 This supports the view, that [Cu2(H-1)6][BF4]2 essentially exists as hydrogen-bonded contact ion pair in CH2Cl2 solution. Solid state and solution IR spectra (Supporting Information, Figures S11 and S12) further corroborate the ion pair formation. As KBr disk, [Cu2(H-1)6][BF4]2 displays absorption bands for hydrogen bonded NH groups at 3281 cm–1 and 3198 cm–1 which can be assigned to NH•••F and NH•••S hydrogen bonds, respectively, based on their relative intensity. In CH2Cl2 solution these bands are still observed at 3267 cm–1 and 3209 cm–1, respectively. Additionally, a band at 3361 cm–1 appears due to the presence of two free NH groups in the neutral [Cu2(H-1)6][BF4]2 contact ion pair. This value correlates well to the NH stretching vibration of H-1 itself (3383 cm–1).5 [1] Kloss, F.; Pidot, S.; Goerls, H.; Friedrich, T.; Hertweck, C. Angew. Chem. 2013, 125, 10945–10948; Angew. Chem. Int. Ed. 2013, 52, 10745–10748. [2] Constable, E. C.; Raithby, P. R.; J. Chem. Soc., Dalton Trans. 1987, 2281–2283. [3] Kokkou, S. C.; Fortier, S.; Rentzeperis, P. J.; Karagiannidis, P. Acta Cryst. 1983, C39, 178-180. [4] Lobana, T. S.; Castineiras, A. Polyhedron, 2002, 21, 1603–1611. [5] Kienz, T.; Förster, C.; Heinze, K. Organometallics 2014, 33, 4803–4812. [6] Feltham, R. D.; Hayter, R. G. J. Chem. Soc. 1964, 4587–459. 178 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Table S2. Selected bond lengths (Å) and bond angles (°) of Ag6[µ3-(E-1)]6. Ag1•••Ag2 2.8863(11) Ag1-N1 2.300(7) Ag1•••Ag3 3.0314(10) Ag2-N2 2.308(7) Ag1•••Ag2’ 3.5516(12) Ag3-N3 2.325(6) Ag1•••Ag3’ 3.1023(10) S1-C11 1.768(8) Ag2•••Ag3 2.9446(10) S2-C23 1.764(9) Ag2•••Ag3’ 3.2510(10) S3-C35 1.768(8) Ag2-S1 2.490(2) Ag3’-S1 2.472(2) Ag2-S1-Ag3’ 81.9(7) Ag1-S2 2.470(2) Ag1-S2-Ag3 75.6(6) Ag3-S2 2.475(2) Ag1-S3-Ag2’ 88.8(7) Ag2-S3 2.544(2) Ag1’-S3 2.531(2) Table S3 Selected bond lengths (Å) and bond angles (°) of cyclo-[H3-7][SbF6]2. N1-C11 1.308(7) C13-N1-C11-S1 –5.9(7) N1-C13 1.479(7) Fe2•••N3 3.364(5) N2-C25 1.347(7) C11-N1-C13-C14 18.9(8) N2-C14 1.478(8) C11-N1-C6-C10 52.9(9) N3-C25 1.298(7) C25-N3-C27-C31 –53.7(8) N3-C27 1.430(7) C25-N2-C15-C19 131.7(6) 179 Supporting Information Figure S1. ESI+ mass spectrum of H-1 and Ag[SbF6] in CH3CN. Figure S2. Thermal ellipsoid representation of [Cu2(H-1)6][BF4]2 (ellipsoids at 50% probability level) derived by single crystal XRD analysis (protons omitted for clarity). 180 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S3. 1H NMR spectrum of [Cu2(H-1)6][BF4]2 in CD2Cl2. Figure S4. 13C{1H} NMR spectrum of [Cu2(H-1)6][BF4]2 in CD2Cl2. 181 Supporting Information Figure S5. 1H13C HSQC spectrum of [Cu2(H-1)6][BF4]2 in CD2Cl2. Figure S6. 1H13C HMBC spectrum of [Cu2(H-1)6][BF4]2 in CD2Cl2. 182 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S7. 1H1H COSY spectrum of [Cu2(H-1)6][BF4]2 in CD2Cl2. Figure S8. 1H1H NOESY spectrum of [Cu2(H-1)6][BF4]2 in CD2Cl2. 183 Supporting Information Figure S9. Equivalent conductivities of [Cu2(H-1)6][BF4]2, ferrocene, [nBu4N][B(C6F5)4] and [CoII(bpy)3][B(C6F5)4]2 in CH2Cl2. Figure S10. Onsager plots of [Cu2(H-1)6][BF4]2, ferrocene, [nBu4N][B(C6F5)4] and [CoII(bpy)3][B(C6F5)4]2 in CH2Cl2. 184 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S11. Infrared spectrum of [Cu2(H-1)6][BF4]2 in KBr. Figure S12. Infrared spectrum of [Cu2(H-1)6][BF4]2 in CH2Cl2. 185 Supporting Information Figure S13. Square wave voltammogram of [Cu2(H-1)6][BF4]2 at 100 mV s-1 with 3 eq. of decamethylferrocene in CH2Cl2/[nBu4N][B(C6F5)4]. Figure S14. Cyclic voltammogram of [Cu2(H-1)6][BF4]2 at 100 mV s−1 in CH Cl /[n2 2 Bu4N][B(C6F5)4]. 186 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S15. UV/Vis/NIR spectrum of [Cu2(H-1)6][BF4]2 in CH2Cl2. Figure S16. 1H NMR spectra of H-1 after addition of 1 eq. P t1 Bu, nBuLi or Li(NMe2) in THF, evaporation and dissolution in CD2Cl2. 187 Supporting Information Figure S17. Thermal ellipsoid representation of Ag6(E-1)6 (ellipsoids at 50% probability level) derived by single crystal XRD analysis. Figure S18. Frontier orbitals (DFT, B3LYP, def2-SVP, ZORA, D3) of a) N- ferrocenyl ketenimine 2 (HOMO and LUMO at 0.1 a.u.) and b) N-ferrocenyl ketenimine cation 2+ (HOMO, SOMO and LUMO+2 at 0.1 a.u.). Cp protons omitted for clarity. 188 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S19. 1H NMR spectrum of H-5 in CD2Cl2. Figure S20. 13C{1H} NMR spectrum of H-5 in CD2Cl2. 189 Supporting Information Figure S21. 1H13C HSQC spectrum of H-5 in CD2Cl2. Figure S22. 1H13C HMBC spectrum of H-5 in CD2Cl2. 190 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S23. 1H1H COSY spectrum of H-5 in CD2Cl2. Figure S24. 1H1H NOESY spectrum of H-5 in CD2Cl2. 191 Supporting Information Figure S25. FD+ mass spectrum of H-5 in CH2Cl2. Figure S26. Thermal ellipsoid representation of cyclo-[H3-7][SbF6]2 (ellipsoids at 50% probability level) derived by single crystal XRD analysis. 192 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S27. 1H NMR spectrum of cyclo-[H3-7][SbF6]2 in CD2Cl2. Figure S28. 13C{1H} NMR spectrum of cyclo-[H3-7][SbF6]2 in CD2Cl2. 193 Supporting Information Figure S29. 1H13C HSQC spectrum of cyclo-[H3-7][SbF6]2 in CD2Cl2. Figure S30. 1H13C HMBC spectrum of cyclo-[H3-7][SbF6]2 in CD2Cl2. 194 To 3.3: Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates Figure S31. 1H1H COSY spectrum of cyclo-[H3-7][SbF6]2 in CD2Cl2. Figure S32. 1H1H NOESY spectrum of cyclo-[H3-7][SbF6]2 in CD2Cl2. 195 Supporting Information Figure S33. FD+ mass spectrum of cyclo-[H3-7][SbF6]2 in CH2Cl2. 196 6 Acknowledgements Many thanks to… … ___ for the opportunity to explore this subjected in depth and the scientific discussions. … ___ for solving even the most complicated crystal structures. … ___ aka ___ for being one of the founding members of the „ferrocene special forces“, for many helpful discussion, substance donations, for the crackling bed sheets during our conference visits and providing the right words to express myself like “Uffresche!”. … ___ aka ___ for his advice under all circumstances, for giving me the insight, that molybdenum is the answer to everything, except if the answer can be beer or metal, than choose beer or metal, and the ability to keep me sane with his professionally engineered, nearly indeterminable sarcasm. … ___ aka ___ for the very helpful discussions and advice and just his being there, to ensure I was never in the wrong room, and therefore keep on improving. … ___ for introducing me to Bouldering; for a lot of fun times and showing me that skill surpasses strength by all means. … ___ for introducing me to EPR spectroscopy, cheerful conversation and the best interpretation of an overweight elf, I have ever seen. … ___ for lightening up my mood and bearing all my bad moods without choice as my office neighbor. … ___ for substance donations and providing a whole lot of funny conversation. 197 Acknowledgements … ___ aka ___, for substance donations, questioning my understanding of everything research related and forcing me to get it right and for showing me, that there are still truehearted people in the world … ___ for bearing my bursts of rage, which must have threatened him to death, being my second and nearest office neighbor. … ___ for her positive attitude and her ability to tie the group together. … ___ for enduring my taste in music and being a real “Panzermensch” in case of patience. … ___ for being a constant reminder to keep calm. … ___ for always providing a critical point of view. … ___ aka ___ for being am member of the “ferrocene special forces” and sacrificing his next years to this cause to ensure that the ferrocene chemistry is in good hands. … ___ aka ___ for providing comical relief in stressful situations. … ___ for his introduction to the work with our cluster and constant helpful advice for this work. … ___ for doing all the things that kept our work going. … all current and former members of ___, on whose results, experience and suffering this work is built on, for the great working atmosphere. … my research students ___, ___, ___, ___, ___ and ___ for their help in synthesis and analytics on which parts of this work are based on. 198 … my bachelor student ___ for restoring my hope in the next generation of chemists due to her excellent work. … ___ for constant cake donations to keep the spirit up, insightful conversations and all the laughter we permanently, entirely, nearly indefatigably shared. … ___ aka ___ for keeping my spirit up, when I needed it and being there, so I did not feel alone. … the ___ for having a reason to flee the canteen food and have some not work-related talk (sometimes). … all my friends, who are not directly related to this work, but had to bear everything the creation of this work made me do, but still are my friends. … ___ for providing distractions, when desperately needed. … my whole family for their constant support to follow my dreams. … ___ for NMR and LIFDI support even at the most uncivilized hours and his ability to shimm faster than his own shadow. … ___ for the XRD sample preparation and data collection and for getting the best reflexes out of the worst crystals … ___ and ___ for mass spectrometric measurements. … Chemikalienlager aka ___ for providing materials as fast as possible. … the ”Zentrum for Qualitätssicherung und –entwicklung” of the Johannes Gutenberg University for financial support for conference visits. 199 Acknowledgements … microanalytical laboratory of the chemical institutes of the University of Mainz for the measurements of elemental analyses. 200 7 List of Publications Publications T. Kienz, C. Förster, K. Heinze: “Generation and Oligomerization of N-Ferrocenyl Ketenimines via Open-shell Intermediates”, submitted A. Neidlinger, T. Kienz, K. Heinze: “Spin Trapping of Carbon-Centered Ferrocenyl Radicals with Nitrosobenzene”, Organometallics 2015, 34, 5310-4320. [DOI: 10.1021/acs.organomet.5b00778] T. Kienz, C. Förster, K.Heinze: „Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry”, Organometallics 2014, 33, 4803-4812. [DOI: 10.1021/om500052k] Oral Presentations T. Kienz, K. Heinze: “Conformational and Electronic Effects of Oxygen→Sulfur Exchange in Ferrocene Oligopeptides”, oral presentation at the 12th Ferrocene Colloquium, 17. – 19. February 2014, Innsbruck, Austria. T. Kienz, K. Heinze: “Conformational and Electronic Effects of Oxygen→Sulfur Exchange in Ferrocene Oligopeptides”, short-oral presentation at the 11th Ferrocene Colloquium, 06. – 08. February 2013, Hannover, Germany. Poster Presentations T. Kienz, K. Heinze: „Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry”, poster presented at “17. Vortragstagung der Wöhler-Vereinigung”, 24. – 26. September 2014, Saarbrücken, Germany. T. Kienz, K. Heinze: „Impact of O→S Exchange in Ferrocenyl Amides on Structure and Redoxchemistry”, poster presented at Electrochemistry 2014, 22. – 24. September 2014, Mainz, Germany. 201 List of Publications T. Kienz, K. Heinze: “Conformational and Electronic Effects of Oxygen→Sulfur Exchange in Ferrocene Oligopeptides”, poster presented at the 11th Ferrocene Colloquium, 06. – 08. February 2013, Hannover, Germany. 202 8 Curriculum Vitae Torben Kienz Date of birth 16. Oktober 1985 Place of birth Worms Professional Experience 06/2011 – present Research Associate Johannes Gutenberg-University Mainz  Organization of lab courses  Conception of research projects  Communication with supplying companies  Training of technicians Education 06/2011 – 11/2016 Doctoral Thesis in the group of ___ Johannes Gutenberg-University Mainz: „Exploration of N-Ferrocenyl Substituted Thioamides: Synthesis, Properties and Reactivity“  Development of new synthetic strategies for organometallic compounds  Practical Experience in analysis methods (NMR spectroscopy, IR spectroscopy, UV/vis spectroscopy, fluorescence spectroscopy, EPR spectroscopy, electrochemical analyses)  Performance of DFT calculation for interpretation of research results (Gaussian 09, Orca)  Representation of the working group during international conferences by poster oral presentations 10/2005 – 05/2011 Academic studies in Chemistry Johannes Gutenberg-University Mainz  Inorganic Chemistry 203 Curriculum Vitae  Organic Chemistry  Physical Chemistry  Macromolecular Chemistry Diploma Thesis in the group of ___: „Synthesis and Charakterisation of N,N‘,N‘‘-Triaryl Substituted Guanidines“ 08/1996 – 03/2005 Abitur Eleonoren Gymnasium Worms Skills Languages German: First language English: Business fluent Computing Microsoft Excel, Powerpoint, Outlook, Word Office specific Chemcraft, Chemdraw, Gaussian 09, software LaTeX, Mercury, MestReNova, Orca, Origin, Photoshop, Mainz, 07 October 2016 204