Curious Case of Cobaltocenium Carbaldehyde

Cobaltocenium carbaldehyde (formylcobaltocenium) hexafluoridophosphate, a long sought-after functionalized cobaltocenium salt, is accessible from cobaltocenium carboxylic acid by a three-step synthetic sequence involving (i) chlorination to the acid chloride, (ii) copper-borohydride reduction to the hydroxymethyl derivative, and (iii) Dess–Martin oxidation to the title compound. Due to the strongly electron-withdrawing cationic cobaltocenium moiety, no standard aldehyde reactivity is observed. Instead, nucleophilic addition followed by haloform-type cleavage prevails, thereby ruling out common useful aldehyde derivatization. One-electron reduction of cobaltocenium carbaldehyde hexafluoridophosphate affords the deep-blue, isolable cobaltocene carbaldehyde 19-valence-electron radical whose spin density is located fully at cobalt and not at the formyl carbon atom. 1H/13C NMR, IR, EPR spectroscopy, high-resolution mass spectrometry, cyclic voltammetry, single crystal structure analysis (XRD), and density functional theory are applied to characterize these unusual formyl-cobaltocenium/cobaltocene compounds.


■ INTRODUCTION
In the last couple of years, we have explored the chemistry of functionalized cobaltocenium salts 1 aiming at developing the rather neglected chemistry of cobaltocenium compounds further in comparison to the very well researched and studied ferrocene derivatives and materials. 2 Cobaltocenium salts are well worth to be investigated, due to their high chemical stability, fully reversible redox chemistry, high polarity with concomitant high solubility in polar solvents, and opposite electronic character (cationic charge, electron-withdrawing) to ferrocenes (neutral, electron-donating). To exploit these properties in more complex molecular compounds or materials, we need useful functionalized cobaltocenium synthons. So far, monofunctionalized cobaltocenium salts containing CO 2 H, CO 2 Cl, C�CH, N 2 + , N 3 , I, Br, Cl, NH 2 , and Se−functionalities 1 as well as O-triflate 3 proved accessible and were used by us and others in diverse applications like redox catalysis, 1b,d,e bioorganometallic chemistry, 1a,g,h,4,5 and redox-responsive macromolecules. 6 A hitherto missing and long sought-after, synthetically highly promising functionalized cobaltocenium species is cobaltocenium carbaldehyde hexafluoridophosphate whose synthesis and peculiar properties are reported in this contribution.

■ RESULTS AND DISCUSSION
Synthesis. Our first attempt to get access to a formylcobaltocenium salt dates back to 1998 when we tried to prepare pentamethylcobaltocene aldehyde by a halfsandwich capping reaction of [Cp*CoBr] 2 with formylcyclopentadienide. 7 However, this Co(II) aldehyde proved unstable and dimerized to the corresponding biscobaltocenium pinacol via a redox disproportionation of an in equilibrium present zwitterionic pentamethylcobaltocenium formyl radical anion. 7 So far, cobaltocenium carbaldehyde remains unknown, but pentamethyliridocenium carbaldehyde hexafluoridophosphate has been briefly mentioned as a hydrolysis product from the corresponding oxazoline precursor. 8 Having gained much experience in cobaltocenium chemistry in the last few years, 1 we thought that reduction of cobaltocenium carboxylic acid hexafluoridophosphate 1i by a suitable reagent might provide a feasible synthetic route to a cobaltocenium aldehyde. However, most standard reducing agents like NaBH 4 , LiAlH 4 , or iBu 2 AlH are not applicable here, due to competing nucleophilic attack of hydride at the cationic cobaltocenium moiety. 9 After screening of various reducing agents and reaction conditions, the following three-step procedure proved successful (Scheme 1). First, cobaltocenium carboxylic acid hexafluoridophosphate (1) was converted in neat thionyl chloride to its acid c h l o r i d e , 1 i f o l l o w e d b y r e d u c t i o n w i t h b i s -(triphenylphosphine)copper(I) tetrahydridoborate. 10 This is considered in standard organic chemistry a selective reducing reagent for conversion of acid chlorides to aldehydes, 10 but in our case, "over-reduction" to primary alcohol 2 was observed. In addition, and synthetically also important, no undesired nucleophilic attack of hydride at the cationic cobaltocenium moiety was evident. Hydroxymethylcobaltocenium hexafluoridophosphate (2) is a rather labile compound with limited shelf life; therefore it was just characterized by 1 H NMR spectroscopy and then immediately used for the final oxidation step. With Dess-Martin periodinane, 11 2 was oxidized to cobaltocenium carbaldehyde hexafluoridophosphate (3) in 38% yield from cobaltocenium carboxylic acid hexafluoridophosphate (1) over all three steps. Other common alcoholaldehyde oxidations like Swern, Albright-Goldman, or Pfitzner-Moffatt activated DMSO reactions proved less suitable. Overall, we see that functional group transformations of cobaltocenium compounds are different compared to those of purely organic compounds, due to the directly attached, cationic, electron-withdrawing cobaltocenium moiety.
Structural and Spectroscopic Properties. Cationic aldehyde 3 has spectral properties of an acceptor-substituted aldehyde. The formyl group is clearly evident from its strong IR band observed at 1702 cm −1 , and the hexafluoridophosphate counter ion gives rise to distinctive IR absorptions at 816 and 555 cm −1 . 1 H NMR data show a regular monosubstituted cobaltocenium moiety with the common pattern of two pseudo-triplets (δ = 6.42, 6.16 ppm) for the substituted Cp and a more intense singlet (δ = 6.05 ppm) for the unsubstituted Cp ligand. The aldehyde proton is observed at 10.18 ppm in the typical low field spectral region of aldehydes. Corresponding 13 C chemical shifts are 189.6 ppm for the aldehyde carbon and 95.0−85.8 ppm for the Cp carbons. The identity of 3 is further corroborated by high-resolution mass spectrometry in agreement of the experimental most abundant monoisotopic peak and the calculated value. 3 is a yellow, moderately air-sensitive solid with a high melting point of 223.5°C, in line with its ionic character. A single crystal structure analysis provides further unambiguous proof for the identity of 3 ( Figure 1, Supporting Information).
Reactivity. In organic chemistry, aldehydes are very useful synthons for a wide range of condensation and nucleophilic addition reactions. It is quite clear that a cationic aldehyde like 3 containing a directly attached electron-withdrawing cobaltocenium group will be much more reactive in comparison to organic aldehydes. This proved to be the case, although not in the anticipated manner. In contrast to our expectations, no successful condensation reactions (for example, Schiff base formation with aromatic primary amines or cyclocondensation with pyrrole to porphyrins) were possible; instead, degradation of the cobaltocenium aldehyde 3 under loss of the formyl group to parent, unsubstituted cobaltocenium hexafluorido-phosphate was observed. Chemically, this can be explained by a haloform-type cleavage 12 of 3 under formation of formic derivative 4 and mesoionic cobaltocenylidene 13 5 which easily gets protonated by solvent due to its high basicity (pK a calculated = 38.5) 13 (Scheme 2).
In comparison to the classical haloform reaction of chloral with aqueous alkali to give chloroform and formic acid, cobaltocenium aldehyde 3 is obviously a (much) more reactive species that is easily cleaved even by nucleophiles of medium strength.
In an alternative interpretation, this reaction might be seen as a simple nucleophilic addition−nucleofuge elimination process. However, such reactions are usually favored with substrates containing good leaving groups such as halides or triflates that have a low basicity. In comparison to the leaving group CCl 3 − (pK a of CHCl 3 = 15.5), cobaltocenylidene 5 (pK a calculated = 38.5) 13 is obviously an extremely poor leaving group.
The observed facile cleavage of cobaltocenium aldehyde 3 by nucleophiles rules out another interesting alternative reactivity of 3: the potential deprotonation of 3 by suitable strong bases to a cobalt(I) ketene complex (η 4 -C 5 H 4 �C� O)Co(η 5 -C 5 H 5 ) is clearly prevented by its high susceptibility to nucleophilic attack and subsequent cleavage as depicted in Scheme 2.
To gain some insights on the degree of electrophilic activation of the formyl carbon by the adjacent cationic cobaltocenium moiety, the group electronegativity (gEN) 14 of the cobaltoceniumyl substituent (Cc + = [C 10 H 9 Co] + ) was calculated by density functional theory and compared to those of other electron-withdrawing substituents ( Table 1) 2 ] − were reacted with 3 in the presence of nucleophiles like t BuO − or CN − to effect cleavage and cobaltocenylidene metal complex formation. According to 1 H NMR analysis, some at first promising results were obtained with copper(I) electrophiles but despite many attempts, we could not develop this reaction into a synthetically useful procedure.
Redox Chemistry. Cyclic voltammetry of cobaltocenium carbaldehyde hexafluoridophosphate (3) in acetonitrile solution referenced versus the ferrocene/ferrocenium couple ( Figure 2) revealed a first reversible reduction (E 1/2 = −0.98 V) followed by a second less intense wave (E p = −1.42 V). For comparison, unsubstituted cobaltocenium hexafluoridophosphate has a reduction potential of −1.33 V. We assign the first reduction to a reversible cobaltocenium/cobaltocene-Co(III)/ Co(II) couple and the second event to the presence of some residual cobaltocenium, proven by additional electrochemical experiments with deliberately added cobaltocenium hexafluoridophosphate that led to an increase of the second reduction observed at approximately −1.42 V.
Upon the first reduction, the color of the solution became a distinctive dark blue, at first indicative of formation of a "ketyl" radical anion. However, simple organic formyl radical anions are unstable due to their insufficient steric protection and dimerize quickly to their corresponding pinacols, whereas ketyl radical anions like the well-known benzophenone radical anion are stable and display an ink blue color which is often used in organometallic chemistry as the so-called "ketyl test" to prove

Organometallics pubs.acs.org/Organometallics
Article the absence of water and oxygen in anhydrous ether solvents. Furthermore, it is interesting to note that neutral pentamethylcobaltocene carbaldehyde containing Co(II) and five electron-donating methyl groups dimerizes to its pinacol, 6 in contrast to one-electron-reduced cobaltocenium carbaldehyde hexafluoridophosphate (3). Chemical reduction of 3 with one equivalent of potassium graphite was performed under strictly inert conditions in an argon-filled glovebox, and it proved possible to isolate 6 as a dark blue, highly air-sensitive solid (Scheme 3). This material crystallized from a pentane solution at −40°C, but unfortunately no single crystal structure analysis was possible, because the highly air-sensitive blue crystals dissolved readily at room temperature in the oil used to pick and mount a single crystal.
An X-band EPR spectrum of 6 ( Figure 3, Table 2) in frozen toluene solution at 98 K showed a well resolved 59 Co coupling (I = 7/2) indicative of the spin density localized fully on the Co(II) center but not on the formyl carbon. This finding is in line with calculated data, where analysis of the spin density indicates that the excess alpha electron is localized at the Co center (Table 2, Figure 4, calculation details in the Supporting Information). Comparable EPR spectra of cobaltocenes containing non-redox-active substituents have been reported in the literature. 16 To gain further insights into the (electronic) structure of 6, density functional theory calculations [ωB97-xd3/def2-TZVP/ CPCM (dichloromethane)] were performed. The molecular orbitals (singly occupied molecular orbital: SOMO and lowest unoccupied molecular orbital: LUMO) of 6 ( Figure 4) indicate that the SOMO is a metal centered orbital that is delocalized over the entire molecule, while the LUMO is the π* orbital of the CHO group. A Hirshfeld charge analysis shows that the neutral 6 has no significant charges except for the formyl group that has the usual partially positive carbon and partially negative oxygen (details in the Supporting Information). These results were consistent and also found for the double-hybrid B2PLYP density functional (see Supporting Information).

■ CONCLUSIONS
Cobaltocenium carbaldehyde hexafluoridophosphate, a hitherto elusive monofunctionalized cobaltocenium salt, was synthesized from cobaltocenium carboxylic acid by first converting it to its carboxylic acid chloride followed by reduction with bis(triphenylphosphine)copper(I) tetrahydridoborate to its hydroxymethyl derivative and subsequent Dess−Martin oxidation. Cobaltocenium carbaldehyde hexafluoridophosphate is an untypical aldehyde that shows no standard aldehyde reactivity: no common condensation reaction to Schiff bases were possible; instead, nucleophilic haloform-type cleavage was observed, explainable by the strongly electron-withdrawing cobaltoceniumyl moiety with a formal group electronegativity comparable to that of the pentafluorophenyl group. Cobaltocenium carbaldehyde hexafluoridophosphate is (electro)chemically reversibly reducible to neutral, very air-sensitive, 19-valence-electron formylcobaltocene whose spin-density is only located at the cobalt center but not on the formyl carbon, as shown by EPR spectroscopy and DFT calculations. Therefore, no reductive pinacol coupling is possible, in contrast to not isolable formyl-(pentamethyl)cobaltocene which is known to readily form pinacol dimers and other formyl radical follow-up products. ■ EXPERIMENTAL SECTION General Procedures. Standard organometallic methods and analytical equipment were used as published previously. 16 The  Cyclic voltammograms were recorded in a glovebox under an atmosphere of argon, using a BioLogic SP-150 potentiostat with a three-electrode setup (glassy carbon working electrode, platinum wire counter electrode, silver wire pseudo reference) and NBu 4 + PF 6 − as supporting electrolyte (0.15 M). 17b,c,17,17a Potentials were calibrated internally to the ferrocene/ferrocenium redox couple.
X-band EPR spectra were obtained on a Bruker Magnettech MS-5000 spectrometer equipped with a temperature controller and using J-Young style quartz glass tubes. Simulations were performed using the pepper function of the EasySpin package 18 for MatLab.
Quantum Chemical Calculations. Structures were fully optimized in their respective spin states using the range-separated density functional ωB97-xd3 19a in conjunction with the triple-zeta basis set def2-TZVP 19b in dicholoromethane, described as a polarizable continuum model with a permittivity of ε = 9.3. 19c All calculations were performed with ORCA 5.0. 3. 19d Molecular orbitals and the spin density were visualized with VMD. 19e Hydroxymethylcobaltocenium Hexafluoridophosphate (2). A Schlenk vessel was charged under an atmosphere of argon with cobaltocenium carboxylic acid hexafluoridophosphate (1) (1.04 g, 2.75 mmol, 1 equiv), and 15 mL of thionyl chloride and the mixture was refluxed overnight. SOCl 2 was evaporated under reduced pressure, and the remaining solids were dried in vacuum. The vessel was transferred into an argon-filled glovebox, and the residue was taken up in 50 mL of dry CH 2 Cl 2 . Triphenylphosphine (866 mg, 3.3 mmol, 1.2 equiv) and bis(triphenylphosphine)copper(I) tetrahydridoborate 10 (1.82 g, 3.02 mmol, 1.1 equiv) were added in one portion. Gas evolved and the mixture were stirred at ambient temperature overnight. Workup under ambient conditions: the vessel was removed from the glovebox; KPF 6 (1000 mg, 5.50 mmol; 2 equiv), 25 mL of acetone, and 10 mL of toluene were added. The solvents were reduced to about 10 mL on a rotatory evaporator, 30 mL toluene were added, and the resulting solids were filtered off and washed with Et 2 O and dissolved in CH 3 CN [some white (Ph 3 P) 3 CuCl remains]. For chromatography, approximately 10 g of neutral Al 2 O 3 was added and CH 3 CN was evaporated on a rotatory evaporator at slightly reduced pressure. A short alumina column (h = 3 cm, Ø = 4 cm) was first conditioned with a solvent mixture of CH 3 CN/Et 2 O (1:1, v/v) and then dry 2 on Al 2 O 3 was poured on top. Chromatography was performed using a solvent gradient mixture of 400 mL CH 3 CN/Et 2 O (1:1), followed by 100 mL of CH 3 CN + 5% MeOH, and lastly 2 was eluted with 150 mL of CH 3 CN + 5% MeOH. All volatiles from this fraction were evaporated; some remaining Al 2 O 3 was separated by dissolving 2 in acetone, filtered, and after evaporation 650 mg of 2 were isolated in 65% yield. 2 is quite air-sensitive and has only limited stability at room temperature; therefore, only characterization by 1 H NMR was performed: 1 H NMR (400 MHz, CD 3 CN): δ 5.70−5.68 (m, 2H, Cp subst ), 5.67 (s, 5H, Cp unsubst ), 5.61 (t, 3 J = 2.0 Hz, 2H, Cp subst ), 4.36 (d, 3 J = 5.6 Hz, 2H, CH 2 ), 3.66 (t, 3 J = 5.7 Hz, 1H, OH) ppm (see Supporting Information). 2 is immediately used after preparation in the following step.
Cobaltocene Carbaldehyde (6). In an argon-filled glovebox, 3 (58 mg, 0.16 mmol; 1 equiv) was suspended in 15 mL of dry THF and cooled to −40°C. Potassium graphite KC 8 (26 mg, 0.19 mmol; 1.2 equiv) was added, and the suspension was slowly allowed to warm to room temperature under stirring for 2 h. After filtration through a pipet filter, an ink blue solution was obtained. The solvent was removed in vacuum, and the residue was taken up in 5 mL of npentane and filtered again. 6 was crystallized from this solution as dark blue needle-shaped crystals by reversed diffusion crystallization with toluene at −40°C. Unfortunately, no single crystal structure analysis proved possible, due to dissolution of 6 at room temperature in the oil used to pick a crystal. EPR spectroscopy (frozen toluene solution, 98 K): EPR data ( Figure 3, Table 2). IR and UV−vis data are not available 20 due to the high air-sensitivity of compound 6.