CpCo(III) Precatalysts for [2+2+2] Cycloadditions

Catalysts applied in cobalt-catalyzed cyclotrimerizations reactions in general rely on the use of Co(I) precatalysts or the in situ generation of Co(I) catalysts from Co(II) sources by reduction in the presence of steering ligands, often by addition of less noble metals. In this paper, we report the synthesis and properties of novel stable CpCo(III) complexes as precatalysts and their exemplary evaluation for application in catalytic [2+2+2] cycloadditions. The role of phosphite neutral ligands, as well as iodide and cyanide as anionic ligands, on the reactivity of the complexes was evaluated. A modified one-pot approach to the synthesis of Cp ring-functionalized Cp’Co(III) complexes was developed. The investigations demonstrated that CpCo(III) complexes can be directly applied as catalysts in catalytic cyclotrimerizations of triynes without reducing agents as additives.


Introduction
The chemistry of CpCo(III) and Cp*Co(III) complexes (Cp = cyclopentadienyl, Cp* = pentamethylcyclopentadienyl) have gained a lot of attention during the recent decade, especially also due to their role in studying C-H functionalization reactions with respect to their capability compared to the latter group 9 metals [1]. Therefore, the chemistry of the CpCo(III) fragment came into focus from the perspective of catalysis and related organometallic chemistry [2]. Novel applications of such complexes in transformations, such as transfer hydrogenations under aerobic conditions, have been studied only recently [3]. The compound CpCo(CO) 2 is the suitable precursor for the synthetic entry, especially due to the easy availability of air-stable CpCo(III) complexes like CpCoI 2 (CO) (1) by simple reaction with elemental iodine [4]. Since the early stages of half-sandwich complex chemistry the compound CpCoI 2 (PPh 3 ) was synthesized and further derivatization reactions were investigated [5], including the possible synthesis of cobaltacyclopentadienes by reaction with dilithiated butadienes [6]. Cationic CpCo(III) complexes, generated by abstraction of the two iodides from CpCoI 2 (PPh 3 ), have found to undergo orthometalation reaction with trans-azobenzene [7]. However, none of such complexes have been systematically tested as catalyst in cyclotrimerizations, neither as phosphine nor phosphite complexes [8][9][10][11][12][13][14][15]. While phosphines proved to be detrimental to the catalytic activity of the CpCo fragment, phosphites turned out to dissociate significantly easier, however, still providing sufficient stabilization of the precatalyst [16]. The presented study is concerned with the synthesis of CpCo(III)(L)(X) 2 complexes and the study of their reactivity as precatalysts in [2+2+2] cycloaddition reactions of triynes and diynes/nitriles.

Synthesis of CpCo(III) Complexes
The ligand exchange of the CO ligand in CpCoI 2 (CO) (1) for phosphite ligands is a very smooth process, which we were able to demonstrate by the ligand exchange reac-very smooth process, which we were able to demonstrate by the ligand exch shown in Scheme 1. A series of phosphites was investigated, including th tron-donating (e.g., P(Oi-Pr)3, P(OCy)3) and electron-withdrawing ((e P(OCH(CF3)2) groups. Such substitutions on 1 have been published by Bri al. or Pauson et al. using P(OMe)3 as the structurally simplest phosphite case, the exchange reactions occurred in general at room temperature and excellent yields for 2-4, 6 and 7 were obtained. An exception was observe cally highly demanding tri(2,4-di-tert-butylphenyl)phosphite, which reac gishly and did not allow to isolate pure complex 5 in any useful amount. the attempted repeated synthesis with P(OPh)3 as ligand failed, which we not substantiate with an clear explanation. The products of the successfu were simply isolated by filtration and washing. These complexes are a pounds, soluble in polar organic solvents. It is interesting to note that the 3 in complexes 2 (117.4 ppm) and 3 (118.0 ppm) are shifted around 50 ppm compared to the CpCo(CO)(phosphite) complexes and shifted still rough higher field in the case of CpCo(olefin)(phosphite) complexes [19]. Both cla ligands with low-lying * orbitals inherit ligand back-bonding capabiliti density from the cobalt center, thus allowing larger electron-density don phosphites to the metal center compared to the iodide atoms in 2 and 3. T also observed for complex 4 containing the fluorinated phosphite P[OCH(C Scheme 1. Synthesis of phosphite complexes from CpCoI2(CO) (1) and different ph For further comparison, we also synthesized complexes containing cy elucidate the influence of this pseudo halide on complex reactivity. It is kno ogen iodide, ICN, reacts smoothly with CpCo(CO)2, although the resulting c never been investigated further for catalytic purposes [20]. Again, substitu ond CO ligand for a phosphite occurred smoothly for complex 8 as was with P(Oi-Pr) 3. In addition, we realized the exchange of both iodide atoms 2 in the presence of alkali metal cyanide, yielding complex 10 [21,22]. The sy are displayed in Scheme 2. According to the cited reference for the synthesi ment of the iodides for cyanide groups in the presence of the CO ligand lik is not possible, because it would lead to the formation of undesired reac Therefore, the presented sequence of substitution (first CO exchange for ligand, then substitution of the iodide for the cyanide) is mandatory. Scheme 1. Synthesis of phosphite complexes from CpCoI 2 (CO) (1) and different phosphites.
For further comparison, we also synthesized complexes containing cyano groups to elucidate the influence of this pseudo halide on complex reactivity. It is known that cyanogen iodide, ICN, reacts smoothly with CpCo(CO) 2 , although the resulting compound has never been investigated further for catalytic purposes [20]. Again, substitution of the second CO ligand for a phosphite occurred smoothly for complex 8 as was demonstrated with P(Oi-Pr) 3 . In addition, we realized the exchange of both iodide atoms in compound 2 in the presence of alkali metal cyanide, yielding complex 10 [21,22]. The synthesis results are displayed in Scheme 2. According to the cited reference for the synthesis [21], replacement of the iodides for cyanide groups in the presence of the CO ligand like in complex 1 is not possible, because it would lead to the formation of undesired reaction products. Therefore, the presented sequence of substitution (first CO exchange for the phosphite ligand, then substitution of the iodide for the cyanide) is mandatory.
Due to the stability of the Co(III) diiodides, we exemplarily investigated this approach for the synthesis of a functionalized Cp'CoI 2 (CO) complexes from the corresponding substituted cyclopentadiene and Co 2 (CO) 8 , while avoiding the work-up of intermediates like 12 or 13 (Scheme 3). The reaction started out from NaCp by acetyoxylation with dimethylcarbonate to give 12. Cobaltation with in situ generated [ICo(CO) 4 ] resulted in the formation of the cobalt dicarbonyl complex 13, which was directly reacted with iodine without isolation to furnish compound 14 as pure complex with 14% yield over three steps. Subsequent ligand exchange with triisopropyl phosphite furnished complex 15 with excellent yield. Attempts to synthesize the related Cp-acetylated complex led to the Due to the stability of the Co(III) diiodides, we exemplarily investigated this approach for the synthesis of a functionalized Cp'CoI2(CO) complexes from the corresponding substituted cyclopentadiene and Co2(CO)8, while avoiding the work-up of intermediates like 12 or 13 (Scheme 3). The reaction started out from NaCp by acetyoxylation with dimethylcarbonate to give 12. Cobaltation with in situ generated [ICo(CO)4] resulted in the formation of the cobalt dicarbonyl complex 13, which was directly reacted with iodine without isolation to furnish compound 14 as pure complex with 14% yield over three steps. Subsequent ligand exchange with triisopropyl phosphite furnished complex 15 with excellent yield. Attempts to synthesize the related Cp-acetylated complex led to the formation of the expected product, however, all attempts of isolating the pure Co(III)complexes after the methods reported for 14 and 15 were not met with success.

Screening on Catalytic Activity of the CpCo(III) Precatalysts
As mentioned above, we surveyed possible reducing agents in the required reduction to catalytically active CpCo(I) species. The conventional reduction protocol using zinc powder and zinc(II) iodide as additive did not work well [22]. After one hour of pretreatment of complexes 2 and 3 either in THF (50 °C) or toluene (100 °C) with 2 eq. of Zn powder and ZnI2 each, the resulting catalyst furnished the pyridine product from 1,6-heptadiyne and benzonitrile in the test reaction with either 11% (in THF) or 6% yield (in toluene) at maximum.
We turned our attention to cyclizations with the standard testing triyne 16 and were quite surprised to see that in the initial experiment with CpCoI2(CO) without any reducing agent, the cyclization product 16cycl was obtained with 38% yield after 19 h at reaction Scheme 3. Synthesis of functionalized Cp'Co diiodides 14 and 15 in an one-pot approach.

Screening on Catalytic Activity of the CpCo(III) Precatalysts
As mentioned above, we surveyed possible reducing agents in the required reduction to catalytically active CpCo(I) species. The conventional reduction protocol using zinc powder and zinc(II) iodide as additive did not work well [22]. After one hour of pretreatment of complexes 2 and 3 either in THF (50 • C) or toluene (100 • C) with 2 eq. of Zn powder and ZnI 2 each, the resulting catalyst furnished the pyridine product from 1,6-heptadiyne and benzonitrile in the test reaction with either 11% (in THF) or 6% yield (in toluene) at maximum.
We turned our attention to cyclizations with the standard testing triyne 16 and were quite surprised to see that in the initial experiment with CpCoI 2 (CO) without any reducing agent, the cyclization product 16cycl was obtained with 38% yield after 19 h at reaction temperatures as low as 75 • C ( Investigation of other ligands containing partially or completely fluorinated groups and thus being less electron-rich like in complexes 4, 6 and 7 gave inferior results ranging from 8% to 27% (Table 1, entries 4-6). Finally, we also investigated the complexes 8 and 9 with heteroleptic (I/CN) anionic groups beside the Cp and CO or phosphite ligand. They did not show any catalytic activity at either 75 or 100 • C (Table 1, entries 7-9). The biscyanide complex 10, however, showed remarkably different reactivity. While no reactivity was observed at 75 • C, raising the reaction temperature to 105 • C allowed isolation of the product 16cycl with 84% yield (Table 1, entries 10 and 11). The difference of 30 • C in reaction temperature clearly covers the range leading to the catalyst activation, while at 75 • C the complex is completely stable and unreactive. Finally, the ester-substituted analogs 14 and 15 showed both identical results for the cyclization, although less yield of 16cycl compared to the unsubstituted precatalyst 2 (Table 1, entries 2, 12 and 13 for comparison). Clearly, the substitution did not change the reactivity profile significantly. Interestingly, the yield obtained in this transformation for catalyst 2 is in the range of the most reactive CpCo(olefin)(phosphite) precatalysts for the reaction at 75 • C [19]. We further exemplarily investigated modified reaction conditions, particularly to raise the conversion of the used testing triyne 16. The complex CpCo[P(Oi-Pr) 3 ]I 2 (2) was applied as standard catalyst in these investigations ( Table 2). Application of microwave reaction conditions at 100 • C shortened the reaction time but led to lower yield for 16cycl even after 12 h reaction time compared to the reaction at 75 • C (25 and 41% vs. 55%, Table 2, entries 1 and 2). Increasing the catalyst loading to 20 mol% gave a yield of 56% (Table 2, entry 3) under microwave conditions. We also investigated the utilization of silver(I) acetate as iodide abstracting agent, often used for catalytic reactions with Cp*Co(CO)I 2 [23]. In the first experiment, utilizing 20 mol% of silver(I) acetate the cyclization product 16cycl was formed with 20% yield and the starting material was mostly reisolated ( Table 2, entry 4). Repeating the experiment with only 10 mol% of silver(I) acetate under otherwise identical conditions gave with only 5% yield of 16 an even significantly lower yield ( Table 2, entry 5). An experiment in which additional 10 mol% catalyst 2 were added after 12 h reaction time and the reaction run for additional 12 h was conducted but did not lead to an increase in yield ( Table 2, entry 6). However, beside isolation of 43% of 16cycl only 15% of 16 were recovered, giving a strong hint towards the occurrence of side reactions. Finally, addition of elemental zinc as reductant led to 53% product yield after 48 h at 75 • C, beside 8% of reisolated 16 (Table 2, entry 7). The missing amount of triyne not found in product or substrate again points towards side reaction with these catalysts. yield ( Table 2, entry 5). An experiment in which addition after 12 h reaction time and the reaction run for additio not lead to an increase in yield ( We extended our investigations to the cyclization which are usually less reactive than the terminally unsu the substrate triynes 17 and 18 are presented in Scheme 2, the reaction at 75 °C did not give any progress an observed. Due to this reason, the reaction temperature o 100 °C for additional time finally yielded cyclization pro We extended our investigations to the cyclization of terminal substituted triynes, which are usually less reactive than the terminally unsubstituted triynes. The results for the substrate triynes 17 and 18 are presented in Scheme 4. In both cases using precatalyst 2, the reaction at 75 • C did not give any progress and only unreacted triynes were observed. Due to this reason, the reaction temperature of the experiments were raised to 100 • C for additional time finally yielded cyclization products 17cycl and 18cycl, albeit in rather low yields. Due to the lower reactivity of the terminal substituted triynes and required higher temperatures for reaction with CpCo complexes, we investigated the reactions under microwave conditions at 140 • C. For triyne 17, very good 75% yield of 17cycl was obtained and isolated. However, triyne 18 gave only minor or no cyclization product 18cycl at all. Interestingly, while under Cond. A most of triyne 18 was recovered, at the higher reaction temperatures of Cond. B only 27% of starting material was reisolated. A possible reason would be consummation of 18 by a side reaction like polymerization, as no defined further reaction products were isolated. quired higher temperatures for reaction with CpCo complexes, we investigated the reactions under microwave conditions at 140 °C. For triyne 17, very good 75% yield of 17cycl was obtained and isolated. However, triyne 18 gave only minor or no cyclization product 18cycl at all. Interestingly, while under Cond. A most of triyne 18 was recovered, at the higher reaction temperatures of Cond. B only 27% of starting material was reisolated. A possible reason would be consummation of 18 by a side reaction like polymerization, as no defined further reaction products were isolated.  For cyclizations with CpCo complexes, the cobalt oxidation state of +1 is the common feature of the catalytically active species entering the catalytic cycle. We therefore further investigated additional methods to activate the catalyst system by facilitating intramolecular reduction from a CpCo(III) species to a CpCo(I) species by reductive elimination. For late transition metals, reductive elimination of carbon-based substituents from the metal center is a very common and important process, e.g., in C-C coupling reactions, and it has been investigated theoretically and experimentally for the reductive elimination of ethane from L 3 CoMe 2 I (L = PMe 3 ) complexes [24]. The complex CpCo(Me) 2 (PPh 3 ) has been synthesized by the reaction of CpCoI 2 (PPh 3 ) with MeMgBr and did react with alkynes afterwards under reduction and formation of cobaltacyclopentadienes [25,26]. Therefore, we attempted alkylation of the diiodide complex 2 by reaction with the Grignard reagent MeMgBr to produce the dimethylated analogue (2-Me) and set out to investigate its catalytic performance (Scheme 5). 1 H NMR analysis confirmed the alkylation success of this reaction due to the unique shift of the resonance for the methyl group to 0 ppm, however, only partial alkylation was recognizable. Reaction control by 31 P NMR spectroscopy confirmed disappearance of 2 as well and emergence of two new resonances at 160 and 172 ppm. Utilization of 10 mol% (estimated for assumed complete conversion of 2) for the catalytic cyclization of 16 furnished the expected product 16cycl, albeit with slightly lower yield than before, not providing any significant advantage. A comparable investigation of 1,6-heptadiyne and benzonitrile as standard system for pyridine formation by co-cyclo-trimerization gave mediocre 30% yield of pyridine 19, which is significantly lower compared to CpCo(olefin)(phosphite) precatalysts under identical conditions [19]. Replication of the experiment by treating partially fluorinated complex 4 with MeMgBr and direct subsequent reaction of the reaction product with triyne 16 gave cyclization product 16cycl with 31% yield (44% 16 recovered).
Catalysts 2021, 11, x FOR PEER REVIEW For cyclizations with CpCo complexes, the cobalt oxidation state of +1 is the feature of the catalytically active species entering the catalytic cycle. We therefo investigated additional methods to activate the catalyst system by facilitating in ular reduction from a CpCo(III) species to a CpCo(I) species by reductive elimin late transition metals, reductive elimination of carbon-based substituents from center is a very common and important process, e.g., in C-C coupling reactions, been investigated theoretically and experimentally for the reductive elimination from L3CoMe2I (L = PMe3) complexes [24]. The complex CpCo(Me)2(PPh3) has thesized by the reaction of CpCoI2(PPh3) with MeMgBr and did react with alky wards under reduction and formation of cobaltacyclopentadienes [25,26]. Ther attempted alkylation of the diiodide complex 2 by reaction with the Grignar MeMgBr to produce the dimethylated analogue (2-Me) and set out to investigat lytic performance (Scheme 5). 1 H NMR analysis confirmed the alkylation succe reaction due to the unique shift of the resonance for the methyl group to 0 ppm, only partial alkylation was recognizable. Reaction control by 31 P NMR spectros firmed disappearance of 2 as well and emergence of two new resonances at 16 ppm. Utilization of 10 mol% (estimated for assumed complete conversion of catalytic cyclization of 16 furnished the expected product 16cycl, albeit with sligh yield than before, not providing any significant advantage. A comparable invest 1,6-heptadiyne and benzonitrile as standard system for pyridine formation by trimerization gave mediocre 30% yield of pyridine 19, which is significantly lo pared to CpCo(olefin)(phosphite) precatalysts under identical conditions [19]. R of the experiment by treating partially fluorinated complex 4 with MeMgBr a subsequent reaction of the reaction product with triyne 16 gave cyclization prod with 31% yield (44% 16 recovered). We repeated this procedure with the complex 9, containing iodide and c anionic groups (Scheme 6). Reaction with one equivalent of the Grignard reag conversion of 9, being confirmed by 1 H and 31 P NMR spectroscopy reaction cont latter case by a shift from 126.7 ppm (9) to 142.4 ppm (9-Me). Subsequent direct generated species 9-Me as catalyst gave surprising results compared to Scheme case of pyridine (19) synthesis, the yields are slightly lower, independent from th We repeated this procedure with the complex 9, containing iodide and cyanide as anionic groups (Scheme 6). Reaction with one equivalent of the Grignard reagent led to conversion of 9, being confirmed by 1 H and 31 P NMR spectroscopy reaction control, in the latter case by a shift from 126.7 ppm (9) to 142.4 ppm (9-Me). Subsequent direct use of the generated species 9-Me as catalyst gave surprising results compared to Scheme 5.
In the case of pyridine (19) synthesis, the yields are slightly lower, independent from the reaction temperature and no difference appeared for 9-Me at both temperatures (compare Schemes 5 and 6, below). The picture is different for the cyclization of 16, where the yield was basically doubled at 100 • C reaction temperature and are still significantly higher at 75 • C compared to 2-Me (Scheme 6, top). This observation is particularly interesting, as complex 9 did not show any reactivity in the cyclization of 16 before (Table 1, entries 8,9). Replacement of the iodide for the methyl group clearly increases the reactivity in the presence of cyanide as second anionic ligand, including reactivity already observed at 75 • C, when the biscyanide complex 10 was completely inactive (see Table 1, entry 10). A possible reason for this is the stronger bonding of the nitrile group vs. the methyl group, requiring more energy to induce the formation of a reactive species for the catalytic process. We assumed that the catalytic activity might arise from preceding reductio CpCo(III) complex to a CpCo(I) complex. There is evidence for this assump CpCo(III)(PPh3)-dialkyl complexes like CpCo(Me)2(PPh3), who gave cobaltacyclo enes upon reaction with alkynes under thermal conditions, implying reductive tion of the alkyl groups [25]. However, the PPh3 ligand is detrimental for the activity of the complexes and less comparable to phosphites as ligand in this set periments with 2 under addition of elemental zinc did basically show no differen pared to catalysis with 2 without additive (compare entries Tables 1 and 2). The process would thus be reductive elimination of the anionic groups. While this ca agined with complex 9-Me (reductive elimination of MeCN), the elimination fr dide complex 2 or dicyanide complex 10 are significantly less likely. We investig possibility for the latter by scavenging experiments, as the complexes are heate presence of dimethyl fumarate, leading in the case of successful reductive elimina cesses to CpCo[P(Oi-Pr)3](dimethyl fumarate) (20) and the corresponding elim products. However, the experiments showed neither in the NMR experiment no parative scale that reductive elimination occurred from the CpCo(III) complexes 7). We assumed that the catalytic activity might arise from preceding reduction of the CpCo(III) complex to a CpCo(I) complex. There is evidence for this assumption for CpCo(III)(PPh 3 )-dialkyl complexes like CpCo(Me) 2 (PPh 3 ), who gave cobaltacyclopentadienes upon reaction with alkynes under thermal conditions, implying reductive elimination of the alkyl groups [25]. However, the PPh 3 ligand is detrimental for the catalytic activity of the complexes and less comparable to phosphites as ligand in this setting. Experiments with 2 under addition of elemental zinc did basically show no difference compared to catalysis with 2 without additive (compare entries Tables 1 and 2). The simplest process would thus be reductive elimination of the anionic groups. While this can be imagined with complex 9-Me (reductive elimination of MeCN), the elimination from diiodide complex 2 or dicyanide complex 10 are significantly less likely. We investigated this possibility for the latter by scavenging experiments, as the complexes are heated in the presence of dimethyl fumarate, leading in the case of successful reductive elimination processes to CpCo[P(Oi-Pr) 3 ](dimethyl fumarate) (20) and the corresponding elimination products. However, the experiments showed neither in the NMR experiment nor on preparative scale that reductive elimination occurred from the CpCo(III) complexes (Scheme 7). possibility for the latter by scavenging experiments, as the complexes are presence of dimethyl fumarate, leading in the case of successful reductive el cesses to CpCo[P(Oi-Pr)3](dimethyl fumarate) (20) and the correspondin products. However, the experiments showed neither in the NMR experime parative scale that reductive elimination occurred from the CpCo(III) comp 7).

Scheme 7. Attempted scavenging experiments towards reductive elimination from and 10.
We have reacted complex 2 and triyne 16 in a 1:1 ratio to prove if com verted during the reaction (Scheme 8). The 1 H and 31 P NMR spectra before ing to 100 °C showed, that complete conversion of 16 to 16cycl has occurred shows a single signal at 120 ppm before and after the reaction, correspon changed coordination environment for the triisopropyl phosphite. How spectra one cannot deduce an unchanged complex 2 after the reaction becau of signals in the 1 H NMR spectra and shifted or disappeared signals in th Cp group. On the other hand, the spectra did not contain significant traces We have reacted complex 2 and triyne 16 in a 1:1 ratio to prove if complex 2 is converted during the reaction (Scheme 8). The 1 H and 31 P NMR spectra before and after heating to 100 • C showed, that complete conversion of 16 to 16cycl has occurred. The 31 P NMR shows a single signal at 120 ppm before and after the reaction, corresponding to an unchanged coordination environment for the triisopropyl phosphite. However, from the spectra one cannot deduce an unchanged complex 2 after the reaction because of overlaps of signals in the 1 H NMR spectra and shifted or disappeared signals in the region of the Cp group. On the other hand, the spectra did not contain significant traces of byproducts or decomposition products as well and the cyclization appeared to be a rather clean process.
alysts 2021, 11, x FOR PEER REVIEW or decomposition products as well and the cyclization appeared to be cess. Finally, we investigated the reactivity of cationic CpCo(III) comp zations for comparison and to see, if they provide any cyclization activ 9). For this purpose, two eq. silver(I) tetrafluoroborate were added to t and a precipitate was formed. Under thermal as well as microwave c temperatures cyclization products 16cycl and 17cycl were formed, ho cantly lower yields compared to the reactions without AgBF4 (com Scheme 4). This result points towards a possible different reaction mec the cationic CpCo(III) complex compared to the commonly accept CpCo complexes [27]. Finally, we investigated the reactivity of cationic CpCo(III) complexes in these cyclizations for comparison and to see, if they provide any cyclization activity by itself (Scheme 9). For this purpose, two eq. silver(I) tetrafluoroborate were added to the reaction mixture and a precipitate was formed. Under thermal as well as microwave conditions at higher temperatures cyclization products 16cycl and 17cycl were formed, however with significantly lower yields compared to the reactions without AgBF 4 (compare Table 1 and Scheme 4). This result points towards a possible different reaction mechanism initiated by the cationic CpCo(III) complex compared to the commonly accepted mechanisms for CpCo complexes [27].
temperatures cyclization products 16cycl and 17cycl were formed, ho cantly lower yields compared to the reactions without AgBF4 (com Scheme 4). This result points towards a possible different reaction mec the cationic CpCo(III) complex compared to the commonly accept CpCo complexes [27]. Scheme 9. Cyclizations of 16 and 17 using precatalyst 2 and AgBF4 as additiv Cationic CpCo(III) complexes can also be synthesized from the d by treatment with a strong alkylation reagent [18,21]. Reaction of 10 nium tetrafluoroborate furnished the dicationic salt 21 with exce (Scheme 10). The reactivity screening with complex 21 was again undertaken f (Scheme 11). Cationic CpCo(III) complexes can also be synthesized from the dicyano complex 10 by treatment with a strong alkylation reagent [18,21]. Reaction of 10 with trimethyloxonium tetrafluoroborate furnished the dicationic salt 21 with excellent isolated yield (Scheme 10). Scheme 4). This result points towards a possible different reaction mec the cationic CpCo(III) complex compared to the commonly accepte CpCo complexes [27]. Cationic CpCo(III) complexes can also be synthesized from the d by treatment with a strong alkylation reagent [18,21]. Reaction of 10 nium tetrafluoroborate furnished the dicationic salt 21 with excel (Scheme 10). The reactivity screening with complex 21 was again undertaken f (Scheme 11). The reactivity screening with complex 21 was again undertaken for triynes 16 and 17 (Scheme 11). The cationic precatalyst 21 did not give any reactivity at 75 °C with either triyne 16 or 17 and starting material as well as catalyst were recovered. At 105 °C terminally unsubstituted triyne 16 gave 57% yield of 16cycl and most of unreacted 16 was recovered, pointing towards a clean conversion without by-product formation. With triyne 17 only negligible cyclization reactivity was observed, while again most of the starting material was recovered. The reactions appeared to be quite clean because no significant amount of starting material was consumed for undesired reactions, dividing this reaction from the combination of complex 2/AgBF4 (Scheme 9). Overall, the isolated complex 21 is less reactive compared to the in situ generated cationic complex, as demonstrated by the reactivity towards 16 at 75 °C reaction temperature. The complex shows no significant reactivity for the cyclization of internal triyne 17, which is more difficult to transform compared to triyne 16. Future investigations will be directed to the elucidation of the reactivity of such cationic CpCo(III) complexes at higher reaction temperatures and the general differences in the mode of action in cyclizations of such CpCo(III) precatalysts compared to the common CpCo(I) precatalysts. The cationic precatalyst 21 did not give any reactivity at 75 • C with either triyne 16 or 17 and starting material as well as catalyst were recovered. At 105 • C terminally unsubstituted triyne 16 gave 57% yield of 16cycl and most of unreacted 16 was recovered, pointing towards a clean conversion without by-product formation. With triyne 17 only negligible cyclization reactivity was observed, while again most of the starting material was recovered. The reactions appeared to be quite clean because no significant amount of starting material was consumed for undesired reactions, dividing this reaction from the combination of complex 2/AgBF 4 (Scheme 9). Overall, the isolated complex 21 is less reactive compared to the in situ generated cationic complex, as demonstrated by the reactivity towards 16 at 75 • C reaction temperature. The complex shows no significant reactivity for the cyclization of internal triyne 17, which is more difficult to transform compared to triyne 16. Future investigations will be directed to the elucidation of the reactivity of such cationic CpCo(III) complexes at higher reaction temperatures and the general differences in the mode of action in cyclizations of such CpCo(III) precatalysts compared to the common CpCo(I) precatalysts.

Synthesis of Complexes
Attention: Due to evolution of CO gas the reactions comprising CO ligand exchange need be performed in a well-ventilated fume hood! Synthesis of CpCoI 2 (CO) (1) [4]: CpCo(CO) 2 (3.0 g, 16.6 mmol) was dissolved in methanol and solid iodine (4.23 g, 16.6 mmol, 1 eq.) added in small portions, while evolution of CO gas was taking place. The color of the solution changed from red to black. The reaction mixture was stirred for additional two hours, then the methanol was removed in vacuo and the remaining black solid was dissolved in CH 2 Cl 2 and filtrated over a filter frit tubing. The residual was further washed with CH 2 Cl 2 and the filtrates concentrated under reduced pressure. The obtained solid product was washed three times with Et 2 O and allowed to dry in air (black solid, 6.42 g, 95%). The spectroscopic data are in accordance with the literature.
Synthesis of CpCoICN[P(Oi-Pr) 3 ] (9): Following the GP1, P(Oi-Pr) 3 (341 mg, 1.64 mmol, 1 eq.) was dissolved in CH 2 Cl 2 (10 mL) and added dropwise to a solution of CpCoICN(CO) (8, 0.5 g, 1.64 mmol) in CH 2 Cl 2 (10 mL), during which the evolution of gas was observed. The reaction mixture was stirred for 20 h at 25 • C. Afterwards, the solvent was evaporated, and the residue suspended in n-hexane, stirred for a short period of time and filtrated under air. The blackish residue was dried under vacuo (yield: 0.745 g, 94%  (14): To a solution of dicobalt octacarbonyl (6.838 g, 20 mmol) in THF (40 mL) solid iodine was added in small portions (5.076 g, 20 mmol), during which the evolution of gas was observed. The mixture was stirred for 2 h and then added dropwise a solution of freshly made sodium cyclopentadienylmethylcarboxylate (20 mmol) via a syringe during which the evolution of gas was observed. The color from the suspension changed from green to brown. It was stirred for 19 h at room temperature, before the solvent was removed under reduced pressure. The residue was dissolved in methanol (40 mL) and then solid iodine added in small portions (5.076 g, 20 mmol). The resulting brown suspension was bubbling heavily and stirred for additional 22 h. The solvent was removed under reduced pressure, the residue dissolved in dichloromethane (80 mL) and filtrated over a tube frit. The solvent was removed again and the residue filtrated over dry silica gel with THF as the eluent (yield: 1.31 g, 14%). 1 H NMR (300 MHz, CDCl3): δ = 5.38 (t, 2H), 5.18 (t, 2H), 3.74 (s, 3H) ppm. 13 (15): Following the synthesis protocol for complex 2, P(Oi-Pr) 3 (121 mg, 0.582 mmol, 1 eq.) was dissolved in CH 2 Cl 2 (10 mL) and added dropwise to a solution of C 5 H 4 C(O)OCH 3 CoI 2 (CO) (14, 0.27 g, 0.582 mmol) in CH 2 Cl 2 (10 mL), during which the evolution of gas was observed. The reaction mixture was stirred for 20 h at 25 • C. Afterwards, the solvent was removed under reduced pressure and the residue was filtrated over dry silica gel with n-hexane. The collected black solid was dried thoroughly under reduced pressure (yield: 0.325 g, 88%). 1 (Table 1) The CpCo(III) precatalyst (1-10, 14, 15 or CpCo(CO) 2 , with 10 mol% catalyst loading with regard to triyne) was added to a solution of triyne 16 (1 eq.) in toluene and was stirred at the given reaction temperature for a specific time. After cooling to room temperature, the solvent was removed and purified by column chromatography to give the product, which was identified by NMR.
Taken together the experimental results demonstrated the ability of such CpCo(III) complexes to be suitable catalysts for cyclotrimerization reactions even without reduction. The conducted experiments also provided evidence to exclude simple reductive elimination of the anionic ligands to yield active CpCo(I) catalysts, which are known to mediate [2+2+2] cycloaddition reactions. Testing the reactions under reducing conditions in the presence of zinc at 75 • C gave only a very slightly different yield. Isolated CpCo(I) precatalysts are usually more active for the cyclization of nitriles and diynes, which was not found in our presented study, while on the other hand the yields received for cyclization of triynes are lower than those for the most active catalysts observed in the presented study [16]. Therefore, the observed discrepancy of reactivities might be indeed credited to CpCo(III) complexes as active species, and detailed studies are in progress in our laboratory to confirm the oxidation state of the active species.