Synthesis, Structure, and Applications of α-Cationic Phosphines

Aug 16, 2016 - Simultaneously, the new very low lying σ*(P–C+) orbitals increase their π-acceptor character, and as consequence, the global electr...
1 downloads 10 Views 4MB Size
Article pubs.acs.org/accounts

Synthesis, Structure, and Applications of α‑Cationic Phosphines Manuel Alcarazo* Institut für Organische und Biomolekulare Chemie, Georg-August-Universität Göttingen, Tammannstraße 2, 37077 Göttingen, Germany

CONSPECTUS: In α-cationic phosphines, at least one of the three substituents on phosphorus corresponds to a cationic (normally, but not always heteroaromatic) group, which is attached without any spacer to the phosphorus atom by a relatively inert P−C bond. This unique architecture confers to the resulting ligand strong acceptor properties, which frequently surpass those of traditional acceptor ligands such as phosphites or polyfluorinated phosphines. In addition, the fine-tuning of the stereoelectronic properties of α-cationic phosphines is also possible by judicious selection of the number and nature of the cationic groups. The opportunities offered in catalysis by α-cationic ligands arise from this ability to deplete electron density from the metals they coordinate. Thus, if in a hypothetical catalytic cycle the step that determines the rate is facilitated by an increase of the Lewis acidity at the metal center, then an acceleration of the whole process is expected by their use as ancillary ligands. Interestingly, this situation is found more frequently than one might think; many common elementary steps involved in catalytic cycles, such as reductive eliminations, coordination of substrates to metals, or attack of nucleophiles to coordinated substrates, belong to this category and are often fostered by electron poor metal centers. In this regard, our group has observed remarkable ligand acceleration effects by the employment of α-cationic phosphines in Au(I)- and Pt(II)-promoted hydroarylation and cycloisomerization reactions. These results seem to be general in π-acid catalysis when the nucleophile used is not especially electron rich because then their attack to the activated alkene or alkyne is normally rate determining. On the other hand, the use of cationic phosphines also presents drawbacks that limit their range of application. As a general rule, the reduced σ-donation from the phosphine is not compensated by the increased π-back-donation from the metal making the resulting phosphorus−metal bond weaker, and the corresponding catalysts more prone to decomposition. This can be critical when di- or tricationic ancillary ligands are used. In addition, the positively charged groups occasionally participate in undesired side reactions, with either the metal or the substrate, which are not present when their neutral congeners are used. Stimulated by both the fundamental questions regarding bonding and their valuable applications in catalysis, the chemistry of αcationic phosphines has experienced an enormous growth during the last years. This Account describes our group’s efforts and those of others to understand their coordination behavior, study their reactivity, and further develop their range of applications in catalysis.



reaction intermediates by electrospray mass spectrometry,1 the isolation of species characterized by new coordination modes,2 the employment of chiral ion-paired ligands in asymmetric catalysis,3 and the preparation of water- or ionic liquid-soluble catalysts (Figure 1).4 However, the renaissance that the field of cationic phosphines has experienced over the past few years originates

INTRODUCTION AND BACKGROUND The world of ligands is dominated by anionic and neutral species. This is not surprising considering that they have been designed to coordinate metals, and these very often behave as Lewis acids. Cationic ligands are exceptions, and when they are used, the positively charged group is mostly located at a remote position from the donating atom. This allows the modification of the physicochemical properties of the corresponding ligands (and catalysts thereof derived) without significantly altering their donor properties. Interesting applications resulting from these ionic architectures are, among others, the identification of © XXXX American Chemical Society

Received: May 29, 2016

A

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Scheme 1. Synthetic Routes for the Preparation of Monocationic Phosphines

Figure 1. Phosphines with cationic groups attached to the periphery and their applications.

more probably from the recognition of the beneficial effects derived from the incorporation of positive charges in close proximity to the donor position. Due to the strong −I inductive effect of positive charges, the σ-donor abilities of α-cationic phosphines are systematically reduced compared with those of their neutral analogues. Simultaneously, the new very low lying σ*(P−C+) orbitals increase their π-acceptor character, and as consequence, the global electron donation of these ligands to the metal is quite low. Only polyhalogenated phosphines such as PF3, P(CF3)3, or PCl3 depict similar electronic properties, but in contrast to those, the absence of phosphorus−halogen bonds makes α-cationic phosphines much more robust against moisture, and in general, easier to handle compounds. In this Account, the different methods leading to the synthesis of α-cationic phosphines, as well as their structure and reactivity, are described. In addition, theoretical studies are included to clarify how the α-donor/π-acceptor properties of these ligands modify the electronic structure of the metals they coordinate and the implications of these changes for metal catalysis.



Figure 2. Structures of 9 (R = Ph) and 10 (R = Ph; R3 = Me) in the solid state showing bond angles around P1 and selected P−C distances for comparison. Hydrogen atoms and anions were omitted for clarity.

architectures is not straightforward. For this reason, we recently implemented a more general synthetic method based on the reaction of the alternative conceivable partners; namely, secondary phosphines and Vilsmeier-type salts.8 The availability of both starting materials and the high yields of the condensation reactions make this route very reliable even in multigram scale. Since then, the repertoire of α-cationic phosphines incorporated to the ligand tool box has been truly expanded, and it now includes cyclopropenio-,9 imidazolinio-,10 pyridinio-,11 and formamidiniophosphines,8 8−11, respectively (Scheme 1, route d). Moreover, α-cationic arsines can be prepared after only small variations of this synthetic methodology.12 The structural analysis of compounds 7−11 reveals two parameters that are crucial in understanding their coordination properties. The central phosphorus atom displays in all cases a pyramidal environment (sum of angles around P1 = 300−318°, depending on the steric demand of the substituents), while all P−C(+) bonds lengths are, within experimental error, very similar to those of the other two P−C(Ph) bonds. These observations suggest that the nonshared electron pair is retained at phosphorus (Figure 2). For this reason, the coordination chemistry of cations 7−11 seems to be as rich as that of traditional phosphines; up to now, the formation of

SYNTHESIS OF α-CATIONIC PHOSPHINES

Monocationic Phosphines

A first sight analysis of the general structure of α-cationic phosphines makes evident that they could be prepared by nucleophilic attack of free carbenes to dialkyl-/diarylchlorophosphines. In fact, the availability of imidazol-2-ylidenes made imidazoliophosphines 7 the first ever reported members of the α-cationic phosphine series (Scheme 1, route a).5 This synthetic pathway is quite robust, and the use of free carbenes is not even necessary; reagents able to generate the carbene in situ, such as imidazolium 2-carboxylates or 2-(trimethylsilyl)imidazolium salts, are also excellent precursors for this transformation (Scheme 1, route b).6 Alternatively, imidazoliophosphines have also been obtained by selective N-alkylation of 2-(imidazoyl)phosphines with strong alkylating reagents (Scheme 1, route c).5b,7 Note, however, that the three protocols described are limited to the preparation of imidazoliophosphines, and their extension to the synthesis of cationic phosphines derived from other B

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Scheme 2. Synthetic Routes for the Preparation of Di- And Tricationic Phosphines

Figure 4. Structure of 17 in the solid state showing bond angles around P1 and selected P−C distances. Hydrogen atoms and anions were omitted for clarity.

salts (route b, Scheme 2a).14 In addition, double N-alkylation of bis[2-(imidazoyl)]phosphines with MeTfO is also an appropriate method to obtain these same compounds, though quite limited from a structural point of view (route c, Scheme 2a).15 Note, however, that the direct attack of free N-heterocyclic carbenes to highly electrophilic phosphorus sources such as RPCl2 (R = alkyl, aryl) or PCl3 (hypothetical route a) cannot be used as general synthetic route. This reaction is very sensitive to the steric demand of the carbene employed, and very often undesired reductive processes take place leading to the isolation of P(I)-derivatives instead of the expected substitution products.16 Alternatively, double condensation of primary phosphines with 2 equiv of the corresponding Vilsmeier salts in the presence of strong bulky bases also proceeds cleanly, affording the desired dicationic phosphines. The scope of this protocol (route d, Scheme 2a) is again superior to the other ones reported, and thus, dications containing two formamidinium, cyclopropenium, or imidazolidinium moieties have also been prepared in modest to good yields (compounds 13−15, respectively).10,17 P−C(+) bond lengths in dicationic phosphines suffer slight reduction when compared with those in monocationic analogues. This is probably a consequence of the higher partial charge at P in dicationic structures. No important change is observed in the pyramidalization of the phosphorus (sum of angles around P1 = 303−325°, depending on the steric demand of the substituents) (Figure 3). Despite these structural similarities, dicationic phosphines demonstrate significantly reduced ability to form coordination complexes. Metal derivatives are only known for ligands 13 and 14, and even in these cases, only Pt(II), Pd(II), and Au(I) complexes have been reported.10,18 Only two tricationic phosphines have been synthesized to date, 16 decorated with three imidazolium moieties and 17 bearing three cyclopropenium rests attached to the central phosphorus atom (Scheme 2b). Here again P−C(+) bond lengths are shortened compared with their less charged or neutral congeners (16, 1.818 Å; 17, 1.790 Å) (Figure 4). Interestingly, despite the three positive charges, the phosphorus center in 17 is still able to coordinate metal fragments if those can engage in gentle back-donation, such as [PtCl3]− or [PdCl3]−.19 Given the unability for 12 to form any metal derivative, it is not surprising that no coordination chemistry has been reported for 16.20

Figure 3. Structure of 13 (R = Ph) in the solid state showing bond angles around P1 and selected P−C distances. Hydrogen atoms and anions were omitted for clarity.

complexes with Au, Ag, Cu, Pt, Pd, Ni, Ir, and Rh have been described.13 Di- and Tricationic Phosphines

The synthetic routes already described can also be applied, after minimal modifications, to the preparation of di- and tricationic phosphines. Bisimidazolium-substituted phosphines 12 are prepared by reaction of dichlorophosphines with either imidazolium 2-carboxylates or 2-(trimethylsilyl) imidazolium C

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research



DONOR PROPERTIES OF α-CATIONIC PHOSPHINES As already mentioned in the introduction, the direct attachment of positively charged groups to the phosphorus center reduces

complexes are quite sensitive to the steric demand of L ligands, and very often their structures deviate considerably from the ideal square planar geometry. This structural modification also alters the overlap between the orbitals of the metal and CO moieties, and consequently, the values measured through IR spectroscopy are not valid for an accurate comparison among the different ligands.11,22 As an illustrative example, the X-ray structure of compound 18, where distortion from square planarity is evident (angle Cl1−Rh1−C1, 165.89°), is depicted in Figure 5. Fortunately, the Tolman electronic parameters (TEPs) can be estimated with a high degree of accuracy from DFT calculations, avoiding these shortcuts.23 Figure 6 depicts a quite complete Tolman stereoelectronic map for phosphines (orange), arsines (pink), phosphites (violet), and cationic ligands (green). Experimental TEPs are represented by red points while calculated ones are in blue. At first sight, it is clear that cationic phosphines and arsines are characterized by significantly reduced overall donor ability compared with their neutral counterparts. Interestingly, they fill the region of the stereoelectronic map located between phosphites and PF3, an area that has been empty for a long time and only now starts to be populated. The calculated TEPs for compounds 12 (R = Ph) and 17 require additional discussion. Both reveal the strong influence that the introduction of a second or even a third positive charge exerts on the electronic properties of phosphines. Specifically, dication 12, which is an air stable solid, is ranked as an even stronger acceptor than PF3. Note, however, that values calculated in this remote region of the Tolman map, although informative, do not have practical significance. In fact, no metal complex derived from this ligand has ever been isolated despite many attempts.24 Probably, the very weak nature of the hypothetical phosphorus−metal bond in these complexes is

Figure 5. Structure of 18 in the solid state showing the strong deviation of the Cl1−Rh1−C1 angle from ideal linearity. Hydrogen atoms and anions were omitted for clarity.

the σ-donor ability of the new phosphine and simultaneously enhances its π-acceptor character. The traditional method to evaluate these electronic properties and compare them in a semiquantitative manner with those of known phosphines is the measurement of the CO stretching frequencies in compounds of formula [Ni(CO)3L], where L is the phosphine under study. This technique, known as Tolman analysis, is fast and experimentally simple but requires the handling of highly toxic and volatile Ni(CO)4.21 For this reason, the use of [RhCl(CO)L2] as model complexes for these studies has been popularized during the last years. However, our experience with cationic ligands advises against this practice; [RhCl(CO)L2]

Figure 6. Tolman stereoelectronic map for phosphines. D

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

In an effort to investigate these aspects, an inspection of the frontier orbitals was carried out at the B3LYP-D3/def2-TZVP level on phosphines 7−11 (R = Ph in all cases). As expected, the introduction of a cationic group lowers the HOMO energy in relation to Ph3P, in all the compounds under study. This stabilization is relatively similar in all cases, and the HOMO associated energies lay in a relatively narrow range between −9.05 and −9.85 eV. These numbers suggest that the magnitude of the σ-donation for phosphines 7−11 is similar in all cases and not specially influenced by the nature of the charged group (Figure 7). The nearly identical charges calculated at the P atom for 7−11 corroborate this interpretation.12 Conversely, as also depicted in Figure 7, the LUMOs become gradually more localized on the cationic group and their energies diminish progressively along a much wider range, in the sequence cyclopropenium (−4.10 eV), imidazolium (−4.37 eV), formamidinium (−5.00 eV), pyridinium (−6.07 eV), and CF3-substituted pyridinium (−6.34 eV). Hence, it seems that the π-acceptor properties of α-cationic phosphines are more receptive to the nature of the substituent that bears the charge. For this reason we primarily attribute the progressive reduction of global donor properties in ligands 8, 11, 7, and 10 to an increase of their acceptor character following the same sequence.

Table 1. Oxidation Potentials of Neutral and Cationic Phosphinesa



APPLICATIONS IN CATALYSIS The saline nature of α-cationic phosphines substantially modifies their physical properties compared with neutral ones. Specifically, introduction of charges makes the phosphine, and metal complexes thereof, much more soluble in water or ionic liquids. Thus, when used as catalysts, they benefit from the recycling opportunities provided by these solvents. Most of the pioneering work in the area of α-cationic ligands was focused in this particular aspect, and several processes including hydroformylation of olefins,25 hydrogenation,26 hydrosilylations,27 and even cross coupling reactions28 were satisfactorily performed in typical ionic liquids such as 1-butyl-3-methylimidazolium hexafluorophosphate [BMimPF6] and analogues. Once the reactions were finished, a biphasic extraction using an organic apolar solvent was used to separate the products and recycle the catalyst.

a

Oxidation potentials were calibrated versus ferrocene/ferrocinium, Bu4NPF6 (0.1 M) in CH2Cl2. bMeasured in CH3CN. cAdapted to the same scale from ref 23.

responsible for the lack of reactivity. Similar behavior, though not so extreme, is found in the case of tricationic 17 where the formation of metal complexes is limited to [MCl3]− (M = Pd, Pt) fragments. Another parameter that is often used to rank the electronic properties of phosphines is their oxidation potential (Ep(ox)), obtained by cyclic voltametry. This measurement is carried out directly on the free phosphine, and thus, the resulting scale is independent of any issue related to the synthesis or use of metal complexes. Comparison of the Ep(ox) listed in Table 1 with TEPs makes evident that both evaluation methods reproduce the same ranking with accuracy. None of the methods discussed for the evaluation of donor abilities is able to discern the relative contributions of the σdonor and π-acceptor character in the ligands tested. This information is fundamental for our study in order to really understand the effect of positive charges in phosphines. Moreover, it might also provide further insight into another interesting experimental phenomena: not only the number of charges but also the nature of the group where the positive charge resides has an impact on the global donor capacity of αcationic phosphines (compare 7−11 in Figure 6).

π-Acid Catalysis

Even more challenging than the use of phosphines of increased solubility/polarity is the exploitation of the strong π-acceptor properties characteristic of α-cationic phosphines in catalysis. An illustrative example along these lines is our study of the Aucatalyzed cycloisomerization of 2-ethynyl-1,1′-biphenyls bearing substituents on positions 6 and 6′ into twisted phenanthrenes (Figure 8). In these substrates, the difficulties in affording the desired 6endo-dig cyclization with traditional catalysts derive from the notable torsion imposed by the unfavorable steric interactions between the internal substituents (in positions 6 and 6′), which surely impedes the approach of the aromatic ring to the activated alkyne. Having identified the attack of the aromatic ring to the alkyne as the most probable rate limiting step, we hypothesized that a stronger activation of the alkyne moiety by a more acidic metal center might overcome this limitation. This change should provide an earlier transition state for this step, and therefore, the steric factors should be less determining. If this E

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 7. Frontier orbitals for cationic phosphines.

acceptor phosphine, 14(Ph), with AuCl(Me2S) (Figure 9). After screening the reaction conditions using different catalyst loadings, a series of solvents of different polarity, and various silver salts for precatalyst activation, we were able to find optimal conditions, under which the model 2-ethynyl-2,6′dimethylbiphenyl 20 was transformed into the desired 4,5dimethylphenanthrene 21 with excellent yields and surprisingly short reactions times. The significance of this result is better highlighted in Figure 10a where the kinetic profiles for the archetypical Au catalysts, Ph 3 PAuCl (red dots) and (PhO)3PAuCl (green dots), are compared with that of 19 under identical experimental conditions. The relative initial rate constants krel of the reaction promoted by 19 versus Ph3PAuCl or (PhO)3PAuCl are 492.5 and 21.7, respectively. Moreover, the scope of the cycloisomerization was demonstrated to be quite general in terms of size and types of functional groups that could be introduced on the structure of the final phenanthrene. This allowed us to explore the utility of our protocol for the synthesis of a number of biologically active, naturally occurring polyoxygenated phenanthrenes of twisted geometry, such as, calanquinone C29 (22) and epimedoicarisoside A30 (23). Other derivatives such as coeloginin31 (24), which are easily prepared from twisted phenanthrene precursors, have also been synthesized (Figure 10b). To gain more detailed insight into the origin of this ligand acceleration effect, the mechanistic pathway for the cycloisomerization of 20 into 21 using [16(Ph)Au]+ and [Ph3PAu]+ was explored by DFT-methods.18 This analysis confirmed that for [Ph3PAu]+ the rate-determining step was, as speculated, the

Figure 8. Cycloisomerization of sterically hindered biphenyls.

Figure 9. Synthesis and structure of 19 in the solid state. Hydrogen atoms and hexafluoroantimonate anions are omitted for clarity.

statement is correct, a simple modification of the ancillary ligand by a more π-acceptor one would transform a low yielding reaction into a useful route for the synthesis of 4,5-disubstituted phenanthrenes. The most appealing ancillary ligand for this task, tricationic phosphine 17 is not able to coordinate Au; therefore we focused our attention on the catalytic activity of Au-complex 19, an air stable solid prepared by reaction of the strong πF

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 11. Relative free energy profiles (kcal/mol) for the first step of the cyclization of 2-ethynyl-2′,6-dimethylbiphenyl 20 into 4,5-dimethyl phenanthrene 21. The green and red lines represent the calculated profiles for catalysts [16(Ph)Au]+ and [Ph3PAu]+ respectively.

Figure 12. Selected examples of the use of α-cationic phosphines in πacid catalysis.

examples of the cyclizations optimized in our group during our research are shown in Figure 12, together with the optimal catalyst system.

Figure 10. (a) Ligand effect of the Au(I)-catalyzed cyclization of 2ethynyl biphenyls into phenanthrenes. (b) Synthesis of selected naturally occurring phenanthrenes and related structures.

Beyond π-Acid Catalysis

A priori any catalytic cycle whose global rate is controlled by an elementary step that increases the electron density at the metal center might also benefit from the reduction of the energetic span supplied by strong π-acceptor ligands. In this regard, we subsequently selected catalytic processes characterized by challenging reductive eliminations from Pd(II) centers as starting points for a new investigation. Although none of the model transformations tested up to now could be performed in a catalytic fashion, the isolation of the Pd-containing species originating from catalyst decomposition during these attempts was possible (compounds 24 and 25; Figure 13). We postulate that upon formation of the Pd(0)−cationic ligand complex, intensive back-donation from Pd(0) to the low lying σ*(P−C+) orbital induces its oxidative insertion into this bond and the formation of a Pd(II)−NHC complex. In addition, two phosphide moieties act as bridging ligands between the two Pd centers, and two Ph3P, originally from (Ph3P)4Pd, complete the metal coordination spheres. Alternative Pd(0) sources and

attack of the aryl group to the activated alkyne Int1a to form a cyclopropyl intermediate Int2a. However, judging from the computed relative free energies, this same step was no longer the one with the highest activation barrier when the strongly πacidic catalyst [16(Ph)Au]+ was employed (Figure 11). Because [16(Ph)Au]+ was able to minimize the energetic span for the complete catalytic cycle, this catalyst was the best to accelerate the reaction under study.32 Gratifyingly, the situation exemplified above does not correspond to an isolated case. In fact, it is more likely to be a generality in π-acid catalysis where the weak nature of the nucleophilic partners often used (olefins or aromatic rings) make the attack of these groups to the activated alkynes relatively difficult and, therefore, rate determining.8−12,17−20,33 For this reason, α-cationic phosphines are promising ligands to test in these transformations when classical ones fail. Selected G

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Notes

The author declares no competing financial interest. Biography Manuel Alcarazo received his Ph.D. in Chemistry from the University of Seville in 2005 under J. M. Lassaletta and R. Fernández. After that, he moved to the MPI Kohlenforschung where he worked first as a postdoctoral associate in the group of A. Fürstner and later as an independent group leader. Since 2015, he has held a Chair in Organic Chemistry at the University of Göttingen. His research interests include the chemistry of late transition metals and p-block elements, as well as their applications in homogeneous catalysis and synthesis.



ACKNOWLEDGMENTS The former and current members of my group and all scientific collaborators who contributed to this project over the last five years are warmly thanked for their dedication. In addition, the support received from the Georg-August-Universität Göttingen and the Max-Planck-Institut für Kohlenforschung during our stay there is gratefully acknowledged.



Figure 13. Synthesis and structure of 24 and 25 in the solid state. Hydrogen atoms, solvent molecules, and tetrafluoroborate anions are omitted for clarity.

cationic ligands provide similar core structures, although with different flanking ligands.22a,34 The formation of undesired 24, 25, and related species upon reaction of cationic phosphines with electron rich metal species demarcates one of the current frontiers of α-cationic phosphines as ancillary ligands.



CONCLUSION AND PERSPECTIVES Over the past several years, we have demonstrated that αcationic phosphines of different structures can be easily synthesized and effectively employed as ancillary ligands. Due to their strong acceptor properties, these ligands efficiently accept electron density from the metal centers they coordinate and, for this reason, they facilitate catalytic cycles requiring strong Lewis acidity at the metal center during the ratedetermining step. This has been efficiently exploited in our group in the framework of Au(I) and Pt(II) catalysis. αCationic arsines have also been prepared, and they have been found to demonstrate even enhanced acceptor properties. We anticipate that the intensive acceleration effects observed in π-acid catalysis by the use of α-cationic phosphines might have tremendous implications in the area of asymmetric catalysis, where catalysts able to work at lower temperatures are usually required to obtain good enantiomeric excess. The profitable employment of α-cationic phosphines beyond π-acid catalysis is another challenge still remaining in this area. We look forward to addressing both of these in the near future.



REFERENCES

(1) (a) Chen, P. Electrospray ionization Tandem Mass Spectrometry in High-Throughput Screening of Homogeneous Catalysts. Angew. Chem., Int. Ed. 2003, 42, 2832−2847. (b) O’Hair, R. A. J. The 3D Quadrupole Ion Trap Mass Spectrometer as a Complete Chemical Laboratory for Fundamental Gas-Phase Studies of Metal Mediated Chemistry. Chem. Commun. 2006, 1469−1481. (c) Chisholm, D. M.; McIndoe, J. S. Charged Ligands for Catalyst Immobilisation and Analysis. Dalton Trans. 2008, 3933−3945. (2) (a) Tulchinsky, Y.; Iron, M. A.; Botoshansky, M.; Gandelman, M. Nitrenium ions as ligands for transition metals. Nat. Chem. 2011, 3, 525−531. (b) Tulchinsky, Y.; Kozuch, S.; Saha, P.; Mauda, A.; Nisnevich, G.; Botoshansky, M.; Shimon, L. J. W.; Gandelman, M. Coordination Chemistry of N-Heterocyclic Nitrenium-Based Ligands. Chem. - Eur. J. 2015, 21, 7099−7110. (3) (a) Ohmatsu, K.; Ito, M.; Kunieda, T.; Ooi, T. Ion-paired chiral ligands for asymmetric palladium catalysis. Nat. Chem. 2012, 4, 473− 477. (b) Ohmatsu, K.; Ito, M.; Kunieda, T.; Ooi, T. Exploiting the Modularity of Ion-Paired Chiral Ligands for Palladium-Catalyzed Enantioselective Allylation of Benzofuran-2 (3 H)-ones. J. Am. Chem. Soc. 2013, 135, 590−593. (4) Snelders, D. J. M.; Siegler, M. A.; von Chrzanowski, L. S.; Spek, A. L.; van Koten, G.; Gebbink, R. J. M. K. Coulombic inter-ligand repulsion effects on the Pt(II) coordination chemistry of oligocationic, ammonium-functionalized triarylphosphines. Dalton Trans. 2011, 40, 2588−2600. (5) (a) Kuhn, N.; Fahl, J.; Bläser, D.; Boese, R. Synthese und Eigenschaften von [Ph2(Carb)P]AlCl4 (Carb = 2,3-Dihydro-1,3diisopropyl-4,5-dimethylimidazol-2-yliden) − ein stabiler CarbenKomplex des dreiwertigen Phosphors. Z. Anorg. Allg. Chem. 1999, 625, 729−734. (b) For a previous synthesis of a W-complex formally derived from an imidazolium-phosphine, see: Komarov, I. V.; Kornilov, M. I.; Tolmachev, A. A.; Yurchenko, A. A.; Rusanov, E. B.; Chernega, A. N. Tungsten Pentacarbonyl as a Potential Protecting Group for Soft Lewis Base Centres in Alkylation of Multifunctional Molecules. Tetrahedron 1995, 51, 11271−11280. (6) (a) Azouri, M.; Andrieu, J.; Picquet, M.; Richard, P.; Hanquet, B.; Tkatchenko, I. Straightforward Synthesis of Donor-Stabilised Phosphenium Adducts from Imidazolium-2-carboxylate and Their Electronic Properties. Eur. J. Inorg. Chem. 2007, 2007, 4877−4883. (b) Weigand, J. J.; Feldmann, K. O.; Henne, F. D. J. Am. Chem. Soc. 2010, 132, 16321−16323. (7) (a) Tolmachev, A. A.; Yurchenko, A. A.; Merculov, A. S.; Semenova, M. G.; Zarudnitskii, E. V.; Ivanov, V. V.; Pinchuk, A. M. Phosphorylation of 1-Alkylimidazoles and 1-Alkylbenzimidazoles with

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Funding

This work was supported by the Deutsche Forschungsgemeinschaft (AL 1348/5-1) and the European Commission (ERC Starting Grant 277963). H

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Phosphorus (III) Halides in the Presence of Bases. Heteroat. Chem. 1999, 10, 585−597. (b) Debono, N.; Canac, Y.; Duhayon, C.; Chauvin, R. An Atropo-Stereogenic Diphosphane Ligand with a Proximal Cationic Charge: Specific Catalytic Properties of a Palladium Complex Thereof. Eur. J. Inorg. Chem. 2008, 2008, 2991−2999. (c) Canac, Y.; Maaliki, C.; Abdellah, I.; Chauvin, R. Carbeniophosphanes and their carbon → phosphorus → metal ternary complexes. New J. Chem. 2012, 36, 17−27. (8) Alcarazo, M. α-Cationic Phosphines: Synthesis and Applications. Chem. - Eur. J. 2014, 20, 7868−7877. (9) Petuškova, J.; Bruns, H.; Alcarazo, M. CyclopropenylylideneStabilized Diaryl and Dialkyl Phosphenium Cations: Applications in Homogeneous Gold Catalysis. Angew. Chem., Int. Ed. 2011, 50, 3799− 3802. (10) Haldón, E.; Kozma, Á .; Tinnermann, H.; Gu, L.; Goddard, R.; Alcarazo, M. Synthesis and reactivity of α-cationic phosphines: the effect of imidazolinium and amidinium substituents. Dalton Trans. 2016, 45, 1872−1876. (11) Tinnermann, H.; Wille, C.; Alcarazo, M. Synthesis, Structure, and Applications of Pyridiniophosphines. Angew. Chem., Int. Ed. 2014, 53, 8732−8736. (12) Dube, J. W.; Zheng, Y.; Thiel, W.; Alcarazo, M. α-Cationic Arsines: Synthesis, Structure, Reactivity, and Applications. J. Am. Chem. Soc. 2016, 138, 6869−6877. (13) For selected examples, see: (a) Kuhn, N.; Göhner, M.; Henkel, G. {Ph2(Carb)P}MCl3 (M = Pd, Pt; Carb = 2,3-dihydro-1,3diisopropyl-4,5dimethylimidazol-2-ylidene) − a Novel Cationic Phosphane Ligand. Z. Anorg. Allg. Chem. 1999, 625, 1415−1416. (b) Andrieu, J.; Azouri, M.; Richard. First donor stabilizedphosphenium rhodium complexes. Inorg. Chem. Commun. 2008, 11, 1401−1404. (c) Digard, E.; Andrieu, J.; Cattey, H. First donor stabilized-phosphenium copper(I) complexes. Inorg. Chem. Commun. 2012, 25, 39−42. (d) Abdellah, I.; Lepetit, C.; Canac, Y.; Duhayon, C.; Chauvin, R. Imidazoliophosphines are True N-Heterocyclic Carbene (NHC)−Phosphenium Adducts. Chem. - Eur. J. 2010, 16, 13095− 13108. (e) Abdellah, I.; Canac, Y.; Mboyi, C. D.; Duhayon, C.; Chauvin, R. Copper(I) complexes of chelating imidazolo- and imidazolio-diphosphines. J. Organomet. Chem. 2015, 776, 149−152. (14) Azuori, M.; Andrieu, J.; Picquet, M.; Cattey, H. Synthesis of New Cationic Donor-Stabilized Phosphenium Adducts and Their Unexpected P-Substituent Exchange Reactions. Inorg. Chem. 2009, 48, 1236−1242. (15) Maaliki, C.; Canac, Y.; Lepetit, C.; Duhayon, C.; Chauvin, R. Poxidation of gem-dicationic phosphines. RSC Adv. 2013, 3, 20391− 20398. (16) Ellis, B. D.; Dyker, C. A.; Decken, A.; Macdonald, C. L. B. The synthesis, characterisation and electronic structure of N-heterocyclic carbene adducts of PI cations. Chem. Commun. 2005, 1965−1967. (17) Mehler, G.; Linowski, P.; Carreras, J.; Zannardi, A.; Dube, J. W.; Alcarazo, M. Bis(cyclopropenium)phosphines: Synthesis, Reactivity and Applications. Chem. Eur. J. 2016, accepted for publication. (18) Carreras, J.; Gopakumar, G.; Gu, L.; Gimeno, A.; Linowski, P.; Petuškova, J.; Alcarazo, M.; Thiel, W. Polycationic Ligands in Gold Catalysis: Synthesis and Applications of Extremely π-Acidic Catalysts. J. Am. Chem. Soc. 2013, 135, 18815−18823. (19) Petuškova, J.; Patil, M.; Holle, S.; Lehmann, C. W.; Thiel, W.; Alcarazo, M. Synthesis, Structure, and Reactivity of Carbene-Stabilized Phosphorus(III)-Centered Trications [L3P]3+. J. Am. Chem. Soc. 2011, 133, 20758−20760. (20) Henne, F. D.; Dickschat, A. T.; Hennersdorf, F.; Feldmann, K. O.; Weigand, J. J. Synthesis of Selected Cationic Pnictanes [LnPnX3−n]n+ (L = Imidazolium-2-yl; Pn = P, As; n = 1−3) and Replacement Reactions with Pseudohalogens. Inorg. Chem. 2015, 54, 6849−6861. (21) Tolman, C. A. Steric Effects of Phosphorus Ligands in Organometallic Chemistry and Homogeneous Catalysis. Chem. Rev. 1977, 77, 313−348. (22) (a) Kozma, Á .; Deden, T.; Carreras, J.; Wille, C.; Petuškova, J.; Rust, J.; Alcarazo, M. Coordination Chemistry of Cyclopropenylidene-

Stabilized Phosphenium Cations: Synthesis and Reactivity of Pd and Pt Complexes. Chem. - Eur. J. 2014, 20, 2208−2214. (b) Fürstner, A.; Alcarazo, M.; Krause, K.; Lehmann, A. C. Effective Modulation of the Donor Properties of N-Heterocyclic Carbene Ligands by “ThroughSpace” Communication within a Planar Chiral Scaffold. J. Am. Chem. Soc. 2007, 129, 12676−12677. (c) César, V.; Misal Castro, L. C.; Dombray, T.; Sortais, J.; Darcel, C.; Labat, S.; Miqueu, K.; Sotiropoulos, J.; Brousses, R.; Lugan, N.; Lavigne, G. (Cyclopentadienyl)iron(II) Complexes of N-Heterocyclic Carbenes Bearing a Malonate or Imidate Backbone: Synthesis, Structure, and Catalytic Potential in Hydrosilylation. Organometallics 2013, 32, 4643−4655. (23) Gusev, D. G. Donor Properties of a Series of Two-Electron Ligands. Organometallics 2009, 28, 763−770. (24) Maaliki, C.; Lepetit, C.; Canac, Y.; Bijani, C.; Duhayon, C.; Chauvin, R. On the P-Coordinating Limit of NHC−Phosphenium Cations toward RhI Centers. Chem. - Eur. J. 2012, 18, 7705−7714. (25) (a) Brasse, C. C.; Englert, U.; Salzer, A.; Waffenschmidt, H.; Wasserschied, P. Ionic Phosphine Ligands with Cobaltocenium Backbone: Novel Ligands for the Highly Selective, Biphasic, Rhodium-Catalyzed Hydroformylation of 1-Octene in Ionic Liquids. Organometallics 2000, 19, 3818−3823. (b) Brauer, D. J.; Kottsieper, K. W.; Liek, G.; Stelzer, O.; Waffenschmidt, H.; Wasserscheid, P. Phosphines with 2-imidazolium and para-phenyl-2-imidazolium moieties-synthesis and application in two-phase catalysis. J. Organomet. Chem. 2001, 630, 177−184. (26) Saleh, S.; Fayad, E.; Azouri, M.; Hierso, J. C.; Andrieu, J.; Picquet, M. Donor-Stabilized Phosphenium Adducts as New Efficient and Immobilizing Ligands in Palladium-Catalyzed Alkynylation and Platinum-Catalyzed Hydrogenation in Ionic Liquids. Adv. Synth. Catal. 2009, 351, 1621−1628. (27) Li, J.; Peng, J.; Bai, Y.; Zhang, G.; Lai, G.; Li, X. Phosphines with 2-imidazolium ligands enhance the catalytic activity and selectivity of rhodium complexes for hydrosilylation reactions. J. Organomet. Chem. 2010, 695, 431−436. (28) Sirieix, J.; Oßberger, M.; Betzemeier, B.; Knochel, P. Palladium Catalyzed Cross-Couplings of Organozincs in Ionic Liquids. Synlett 2000, 11, 1613−1615. (29) Lee, C. L.; Chang, F. R.; Yen, M. H.; Yu, D.; Liu, Y. H.; Bastow, K. F.; Morris-Natschke, S. L.; Wu, Y. C.; Lee, K. H. Cytotoxic Phenanthrenequinones and 9,10-Dihydrophenanthrenes from Calanthe arisanensis. J. Nat. Prod. 2009, 72, 210−213. (30) Li, W. K.; Pan, J. Q.; Lü, M. J.; Zhang, R. Y.; Xiao, P. G. A 9,10dihydrophenanthrene derivate from Epimedium koreanum. Phytochemistry 1995, 39, 231−233. (31) Majumder, P.; Bandyopadhyay, D.; Joardar, S. Coelogin and coeloginin: two novel 9,10-dihydrophenanthrene derivatives from the orchid Coelogyne cristata. J. Chem. Soc., Perkin Trans. 1 1982, 1131− 1136. (32) Kozuch, S.; Shaik, S. How to Conceptualize Catalytic Cycles? The Energetic Span Model. Acc. Chem. Res. 2011, 44, 101−110. (33) For general reviews on π-acid catalysis, see: (a) Dorel, R.; Echavarren, A. M. Gold(I)-Catalyzed Activation of Alkynes for the Construction of Molecular Complexity. Chem. Rev. 2015, 115, 9028− 9072. (b) Fürstner, A.; Davies, P. W. Catalytic Carbophilic Activation: Catalysis by Platinum and Gold π Acids. Angew. Chem., Int. Ed. 2007, 46, 3410−3449. (c) Hashmi, A. S. K.; Hutchings, G. J. Gold Catalysis. Angew. Chem., Int. Ed. 2006, 45, 7896−7936. (34) Kozma, Á .; Petuškova, J.; Lehmann, C. W.; Alcarazo, M. Synthesis, structure and reactivity of cyclopropenyl-1-ylidene stabilized S(II), Se(II) and Te(II) mono- and dications. Chem. Commun. 2013, 49, 4145−4147. CCDC 1473951 contains the supplementary crystallographic data for compound 24.

I

DOI: 10.1021/acs.accounts.6b00262 Acc. Chem. Res. XXXX, XXX, XXX−XXX