Synthesis and Reactivity toward H2 of (h5- C5Me5)Rh(III) Complexes with Bulky Aminopyridinate Ligands

1 Synthesis and Reactivity toward H 2 of (η5- C5Me5)Rh(III) Complexes with Bulky Aminopyridinate Ligands Ana Zamorano, Nuria Rendón,* José E. V. Valpuesta, Eleuterio Álvarez, and Ernesto Carmona* Instituto de Investigaciones Químicas (IIQ) Departamento de Química Inorgánica and Centro de Innovación en Química Avanzada (ORFEO-CINQA) Consejo Superior de Investigaciones Científicas (CSIC) and Universidad de Sevilla Av. Américo Vespucio 49, Isla de la Cartuja, 41092 Sevilla (Spain) Fax: (+34)954460565 E-mail: [email protected] (N. R.), guzma[email protected] (E. C.) ABSTRACT. Electrophilic, cationic Rh(III) complexes of composition [(η 5 -C 5 Me 5 )Rh(Ap)] + , (1 + ), were prepared by reaction of [(η 5 -C 5 Me 5 )RhCl 2 ] 2 and LiAp (Ap = aminopyridinate ligand) followed by chloride abstraction with NaBAr F (BAr F = B[3,5-(CF 3 ) 2 C 6 H 3 ] 4 ). Reactions of 2 cations 1 + with different Lewis bases (e. g. NH 3 , 4-dimethylaminopyridine or CNXyl) led in general to mono-adducts 1·L + (L = Lewis base; Xyl = 2,6-Me 2 C 6 H 3 ) but carbon monoxide provided carbonyl-carbamoyl complexes 1·(CO) 2+ as a result of metal coordination and formal insertion of CO into the Rh—N amido bond of complexes 1 + . Arguably, the most relevant observation reported in this study stemmed from the reactions of complexes 1 + with H 2 . 1 H NMR analyses of the reactions demonstrated a H 2 -catalyzed isomerization of the aminopyridinate ligand in cations 1 + from the ordinary κ 2 -N,N´-coordination to a very uncommon, formally tridentate κ-N,η 3 -pseudo-allyl bonding mode (complexes 3 + ) following benzylic C—H activation within the xylyl substituent of the pyridinic ring of the aminopyridinate ligand. The isomerization entailed in addition H—H and N—H bond activation and mimicked previous findings with the analogous iridium complexes. However, in dissimilarity with iridium, rhodium complexes 1 + reacted stoichiometrically at 20 ºC with excess H 2 . The transformations resulted in the hydrogenation of the C 5 Me 5 and Ap ligands with concurrent reduction to Rh(I), and yielded complexes [(η 4 -C 5 Me 5 H)Rh(η 6 -ApH)] + , (2 + ), in which the pyridinic xylyl substituent is η 6 - bonded to the rhodium(I) centre. New compounds reported were characterized by microanalysis and NMR spectroscopy. Representative complexes were additionally investigated by X-ray crystallography. 3 INTRODUCTION Organometallic compounds of rhodium and iridium that contain a (η 5 -C 5 Me 5 )M(III) fragment have been extensively investigated because of their capacity to participate in a wide range of chemical transformations. 1-6 Many of these complexes can be readily prepared from corresponding (C 5 Me 5 )M halides 7 by displacement reactions that utilize a variety of inorganic, organic and organometallic reagents. A main, broad area of research that makes ample use of these complexes is the activation of C—H, H—H and other element-hydrogen bonds. Lately, our group has studied the reactivity of compounds of this type that contain a cyclometallated phosphine ligand, P^C. 8 More recently, chelating, also monanionic, cyclometallated C^N and aminopyridinate 9 (N^N) ligands have also been incorporated into (η 5 -C 5 Me 5 )Ir(III) structures. For instance, bulky aminopyridinate ligands (in shorthand notation represented from now on as Ap) have produced five-coordinate [(η 5 - C 5 Me 5 )Ir(Ap)] + complexes 10 in which the amido nitrogen atom acts as a π-donor to compensate the electronic unsaturation of the Ir(III) centre (an η 5 -cyclopentadienyl ligand is regarded to occupy three coordination sites). During studies on the reactivity of these [(η 5 -C 5 Me 5 )Ir(Ap)] + complexes we found that H 2 catalyzed a reversible isomerization of the coordinated Ap ligand, from its common κ 2 ,N,N´- coordination to a novel κ,N,η 3 -pseudo-allylic binding (structures I and II respectively in Figure 1A) in a process that entailed the reversible activation of H—H, C—H and N—H bonds. 10 We therefore deemed of interest to prepare the rhodium complexes analogues and to study their reactivity toward dihydrogen. Here we report the results of this work that has allowed for the isolation and structural characterization of neutral and cationic rhodium complexes, [(η 5 - 4 C 5 Me 5 )Rh(Ap)Cl] (1·Cl), and [(η 5 -C 5 Me 5 )Rh(Ap)] + (1 + ), respectively (the latter isolated as BAr F- salts, where BAr F stands for B[(3,5-(CF 3 ) 2 C 6 H 3 ] 4 ) for the aminopyridinate groups labelled a-d in Figure 1A. Somewhat unexpectedly, and in marked contrast with the analogous iridium complexes the reaction of cations 1a + -1d + with H 2 provided only minor amounts of the isomeric complexes 3 + , with structure of type II in Figure 1. These complexes could not be isolated, but were structurally characterized in solution by NMR spectroscopy. The main products of the hydrogenation reactions were instead cationic Rh(I) complexes 2 + , with structure III (Figure 1B) that possess neutral pentamethylcyclopentadiene, C 5 Me 5 H, and aminopyridine, ApH, ligands with η 4 - and η 6 -coordination, respectively, in the latter case implicating the pyridine aryl substituent. Figure 1. A: H 2 -catalyzed isomerization of aminopyridinate ligands in [(η 5 -C 5 Me 5 )Ir(Ap)] + complexes (see reference 10). B: Main Rh(I) products resulting from the reactions of compounds [(η 5 -C 5 Me 5 )Rh(Ap)] + with H 2 . Rh complexes with structure of type II were also detected and 5 were characterized by NMR studies (see below). The four Ap ligands labelled a-d were employed for both this work and that reported in reference 10a. RESULTS AND DISCUSSION Similarly to the iridium complexes analogues 10 the targeted rhodium aminopyridinate derivatives were prepared by the low-temperature reaction of [(η 5 -C 5 Me 5 )RhCl 2 ] 2 and the corresponding lithium aminopyridinate, LiAp (Scheme 1). The reactions yielded the expected neutral [(η 5 - C 5 Me 5 )Rh(Ap)Cl] complexes, 1a·Cl-1d·Cl, that were readily converted into the desired cationic species 1a + -1d + upon treatment with NaBAr F . Scheme 1. Synthesis of neutral and cationic rhodium aminopyridinate complexes of the (η ηη η 5 - C 5 Me 5 )Rh(III) fragment. 6 The chlorides 1a·Cl-1d·Cl were isolated as reddish crystalline solids. In contrast, both solution and solid samples of the 1a + -1d + cations have a characteristic dark green, almost black, colour. This is due to ligand-to-metal charge transfer π-d electronic transitions commonly encountered in complexes of this kind, in which the amido functionality exhibits a σ- and πdonor coordination behavior. 11 The new compounds were characterized by analytical techniques and by NMR spectroscopy (see the Experimental Section and the accompanying Supporting Information, SI). In addition, two neutral complexes, namely 1a·Cl and 1b·Cl, and the cationic derivatives 1b + , 1c + and 1d + were further analyzed by X-ray crystallography. Members of these series were found to have similar structures. Therefore, only the molecular structures of 1b·Cl and 1b + are depicted in Figure 2, whereas the others can be found in the SI (Figures S1 to S5). The aminopyridinate ligand of both the neutral and the cationic complexes exhibits its classical bidentate coordination despite the strain it creates within the four-member ring. This can be seen in the small N—Rh—N bite angle of ca. 62 º found for the neutral complexes, that increases slightly in the five-coordinate cations 1 + (to ca. 64.5 º in 1b + ). In the neutral complexes the Rh—N py and Rh—N amido bonds have similar lengths, with the former being, as expected, 1b·Cl 1b + Figure 2. X-ray structure of complexes 1b·Cl and 1b + (thermal ellipsoids set at 30% probability and 7 anions omitted for clarity). Selected bond lengths (Angstroms) and angles (degrees) for 1b·Cl: Rh(1)−N(2) 2.113(3), Rh(1)−N(1) 2.162(3), Rh(1)−Cl(1) 2.3892(10), Rh(1)−C(22) 2.170(4), Rh(1)−C(23) 2.116(4), Rh(1)−C(24) 2.146(4), Rh(1)−C(25) 2.145(4), Rh(1)−C(26) 2.175(4), N(2)-Rh(1)- Cl(1) 88.91(9), N(1)-Rh(1)-Cl(1) 85.95(9), N(2)-Rh(1)-N(1) 62.05(12), C(1)-N(2)-Rh(1) 95.8(2), C(1)- N(1)-Rh(1) 92.3(2), N(2)-C(1)-N(1) 109.4(3). Selected bond lengths (Angstroms) and angles (degrees) for 1b + : Rh(1)−N(1) 2.135(3), Rh(1)−C(24) 2.139(4), Rh(1)−N(2) 1.981(4), Rh(1)−C(25) 2.156(4), Rh(1)−C(22) 2.134(4), Rh(1)−C(26) 2.185(4), Rh(1)−C(23) 2.129(4), N(2)-Rh(1)-N(1) 64.51(14), C(1)- N(1)-Rh(1) 90.4(2), C(1)-N(2)-Rh(1) 97.4(3), N(2)-C(1)-N(1) 106.6(3). moderately longer than the latter (ca. 2.16 and 2.11 Å, respectively). These metrics are nearly identical to those reported for the analogous iridium complexes 10 and are also comparable to corresponding values in a heterobinuclear Rh/Nd aminopyridinate complex. 12 At variance with these observations, the partial multiple character of the Rh—N amido bond in the five-coordinated cations 1 + causes this bond to have a length significantly shorter (ca. 1.98 Å) than the Rh—N py bond (2.14 Å). Once more, these structural parameters match closely those reported for somewhat related compounds with chelating monoanionic N^N ligands. 10,11a,11c,13-17 The five-coordinate Rh(III) center of cationic complexes 1 + exhibited Lewis acidity, as evidenced by the facile reaction of 1a + with ammonia, 4-dimethylaminopyridine (dmap) and 2,6- dimethylphenyl isocyanide, CNXyl (Scheme 2). The reactions occurred almost instantaneously and were accompanied by a colour change from dark green to red-orange, indicative of the loss 8 Scheme 2. Lewis acid reactivity of the five-coordinate cationic complex 1a + . of the π-component of the Rh—N amido bond. The resulting adducts 1a·L + present spectroscopic properties similar to those of the analogous iridium complexes 10 (see Experimental Section and SI). In particular, 1a·CNXyl + features ῡ(CN) at 2160 cm -1 , i. e. some 45 cm -1 higher than for the free isocyanide. As for the analogous iridium adduct, 10 this shift denotes that the CNXyl ligand behaves in this compound solely as a σ-donor. X-ray data for 1a·CNXyl + (Figure 3) are also alike those obtained for the iridium adduct analogue. 10 1a·CNXyl + Figure 3. X-ray structure of complex 1a·CNXyl + (30% ellipsoids, H atoms and anion omitted for clarity). Selected bond lengths (Angstroms) and angles (degrees): Rh(1)—N(1) 2.1573(18), Rh(1)—N(2) 2.1172(19), Rh(1)—C(36) 1.989(2), C(36)—N(3) 1.156(3), Rh(1)—C(26) 2.180(3), Rh(1)—C(27) 2.144(3), Rh(1)—C(28) 2.178(2),Rh(1)—C(29) 2.193(2), Rh(1)—C(30) 2.180(3), C(36)-Rh(1)-N(1) 85.35(8), Rh(1)-N(2)-C(1) 95.37(13), C(36)-Rh(1)-N(2) 87.45(9), Rh(1)-N(1)-C(1) 92.47(13), N(1)-Rh(1)-N(2) 61.91(7), N(1)-C(1)-N(2) 109.19(19). 9 Interestingly, acetonitrile formed a bis(adduct) in which the aminopyridinate ligand is bonded to rhodium exclusively through the N amido atom, [(η 5 -C 5 Me 5 )Rh(κ 1 -Ap)(NCMe) 2 ] + , 1a·(NCMe) 2+ (only the mono-adduct [(η 5 -C 5 Me 5 )Ir(κ 2 -Ap)(NCMe)] + was detected for iridium 10 ). The existence of a symmetry plane simplifies considerably the 1 H and 13 C{ 1 H} NMR spectra of this complex. For instance, only one septet (3.22 ppm) and two doublets (1.08 and 1.35 ppm; 3 J HH = 7.0 Hz) were recorded for the i-Pr substituents of the N amido aryl group and one singlet (2.26 ppm) for the methyl protons of the pyridine xylyl substituent. Also in dissimilarity with the analogous iridium system, complexes 1 + reacted with CO with incorporation of two molecules of CO (Scheme 3) and formation of the new compounds Scheme 3. The reaction of complexes 1 + with carbon monoxide. 1·(CO) 2+ that contain a terminal carbonyl (ῡ(CO) = 2070 cm -1 ) and a carbamoyl unit that is part of a five-member ring (ῡ(CO) = 1690 cm -1 , data for 1a·(CO) 2+ ). Comparable reactivity has been disclosed for ruthenium 14a-d and iridium. 14e In the 13 C{ 1 H} NMR spectrum of these compounds the Rh—CO resonance appears at about 187 ppm ( 1 J CRh ca. 75 Hz) while the Rh—C(O)N functionality can be found nearby (~188 ppm) also in the form of a doublet ( 1 J CRh of 16 M N N R R A H2 H2 B M N H N CH2 R R C H2 H2 R´ R´ N NH M HCH3 R R R´ + + + M NN CH3 H3C R R H2(cat.), 25 ºC M N NH H3C CH2 R R ++ R´ R´ I II Scheme 5. General mechanism for the interconversion of rhodium and iridium isomeric complexes with structures I and II in the presence of H 2 . support of experimental and computational studies. 10 Even if it is likely that related equilibria exist between the Rh(III) complexes 1 + and 3 + , unequivocal proof for their operation could not be gathered. Probably this is due to the formation of complexes 3 + only in minute concentrations and above all to the appearing of the Rh(I) complexes 2 + (which are actually unknown for iridium) as the main reaction products. Besides, partial decomposition of the latter to [(η 5 - 17 C 5 Me 5 ) 2 Rh 2 (µ-Cl) 3 ][BAr F ] introduced additional complexity into the reaction manifold. For rhodium, it appears likely that complexes 2 + derive from a Rh(III) hydride intermediate with structure alike B, although an alternative structure B ´ (Scheme 6) in which a δ-agostic interaction replaces the η 2 -arene binding seems also reasonable. The latter hypothesis finds strong support in the observation of a related, also cationic, hydride agostic structure derived from a (η 5 - C 5 Me 5 )Rh(III) fragment and a metallated PMeXyl 2 ligand. 8b Regardless of the precise Scheme 6. Proposed formation of the Rh(I) complexes 2a + -2d + from intermediate B or B´. 18 nature of this hydride intermediate, a reductive coupling step involving its Rh—H and Rh— C 5 Me 5 functionalities, along with a conformational change aimed to establish the Rh(I)-η 6 -arene bonding interaction, would account for the formation of complexes 2 + . It is important to note that as indicated in Schemes 5 and 6, the formation of complexes 2 + could be reverted. In this manner larger amounts of complexes 3 + (structure of type II in Scheme 5) were generated (see the Experimental Section) permitting their complete characterization by NMR studies. Thus, when CD 2 Cl 2 solutions of previously isolated samples of complexes 2 + were heated at 50 ºC in an NMR tube, the corresponding isomers 3 + were readily identified. Partial decomposition to the chloride-bridged dinuclear compound [(η 5 - C 5 Me 5 ) 2 Rh 2 (µ-Cl) 3 ][BAr F ] took also place, but under these conditions the back conversion of 3 + into the isomeric complexes 1 + could not be observed. In the absence of theoretical calculations we can offer no rigorous explanation for the lack of observation of this isomer. However, it is possible that, as for iridium, the thermodynamics favour isomers 3 + relative to 1 + , making the back 3 + -to-1 + conversion slower. This, added to the perturbation introduced in the reaction system by the Rh(I) complexes 2 + , and by their relatively easy decomposition to [(η 5 - C 5 Me 5 ) 2 Rh 2 (µ-Cl) 3 ][BAr F ], could justify our failure to observe complexes 1 + under these conditions. Another relevant piece of knowledge provided by the 1 H NMR monitoring of the reactions of complexes 1 + with H 2 is the facility of formation of complexes 2 + under the conditions of Scheme 4. Complexes 1a + and 1b + differ in the nature of the N amido aryl substituent, viz 2,6-Pr 2i C 6 H 3 and 2,6-Me 2 C 6 H 3 , respectively (See Figure 1A). It was found that quantitative formation of 2a + (≥ 95% by 1 H NMR) required 6 h, whereas that 2b + needed only 1h. In 19 agreement with the isomerization mechanism represented in Scheme 5, this difference reflects, most probably, the steric hindrance exerted by the i-Pr substituents of 1a + to coordination of H 2 to form intermediate A. On the other hand, complexes 1b + , 1c + and 1d + are very similar but may be distinguished thanks to the different R´ group in the 4-position of the aryl substituents of the pyridine ring (H for 1b + , NMe 2 in 1c + and F in the case of 1d + ). As already mentioned, quantitative conversion into 2b + needed 1h, whereas for 2c + 3h were required. As for 2d + , its formation was the fastest, with a 75 % yield after only 10 min. Higher yields of 2d + could not be reached due to its decomposition to [(η 5 C 5 Me 5 ) 2 Rh 2 (µ-Cl) 3 ][BAr F ], probably facilitated by dissociation of the F-containing η 6 -arene ring. The above differences may arise from the facility with which the purported Rh(III) hydride intermediates with structure B, or B´, experience reductive coupling to the corresponding isomeric Rh(I) structures 2 + . Before closing, some comments on the divergent behavior of the iridium and rhodium complexes of composition [(η 5 -C 5 Me 5 )M(Ap)] + in their reactions with H 2 appear appropriate. Firstly, reductive coupling and reductive elimination of M—H 20,24 (or M—Me 25 ) and C 5 Me 5 or other cyclopentadienyl ligands, have been observed for various transition metals, in some cases in a reversible manner. 19,20,23,25 Secondly, the formation under mild conditions of the rhodium complexes 2 + that, as already noted, are unknown for iridium, is, in almost all probability, a consequence of the easier reduction of Rh(III) to Rh(I) in comparison with the analogous Ir(III) to Ir(I) redox change. Such difference in the relative stability of metal oxidation states, not only in the +1 and +3 ones, but also in others, are commonly encountered in the chemistry of late transition elements, particularly of rhodium and iridium. For example, during studies on cyclometallations, Leong and co-workers obtained different types of products in the course of the reactions of the Rh and Ir [(η 5 -C 5 Me 5 )MCl 2 ] 2 dimers with an anyline and a terminal alkyne, that 20 were explained by reason of the more difficult accessibility of the M(V) oxidation state for rhodium. 26 On a completely different topic, in recent studies of the double deprotonation of the bis(2-picolyl)amine ligand in Rh(I) and Ir(I) complexes, [M(bpa)(cod)] + (cod = 1,5- cyclooctadiene), Tejel, de Bruin, Ciriano and co-workers explained the observed thermodynamic differences in the products as arising from the lower stability of Ir(-1) in comparison with Rh(- 1). 17b A last, truly remarkable example that implicates the +1 and +3 oxidation states was provided by Brookhart and co-workers, who took advantage of the ease of reduction of Rh(III) relative to Ir(III) to observe and characterize by solution NMR spectroscopy a relatively longlived σ-methane complex of rhodium (I). 27 In summary, this contribution extends previous studies on the reactivity of cationic Ir(III) complexes of composition [(η 5 -C 5 Me 5 )Rh(Ap)] + with H 210 to the rhodium analogues (complexes 1 + ). For the two metals a dihydrogen-catalyzed isomerization of the aminopyridinate ligand from the common κ 2 -N,N´ bidentate coordination to an unusual κ-N-η 3 -pseudo-allyl bonding mode (Rh complexes 3 + ) has now been documented (Figure 1A), in a rearrangement that required H— H, C—H and N—H bond activation. However, while for iridium this was the principal chemical change observed, 10 for rhodium a stoichiometric hydrogenation of the C 5 Me 5 and aminopyridinate ligands to C 5 Me 5 H and ApH, respectively, took place, with concomitant metal reduction to Rh(I) (complexes 2 + ). This was demonstrated to be the main reaction path and occurred irreversibly at room temperature in the presence of an excess of H 2 . The dissimilar chemical behavior of the rhodium compounds relative to the iridium analogues is most likely due to more facile reduction of M(III) to M(I) for rhodium than for iridium. 21 EXPERIMENTAL SECTION General Procedures: Microanalyses were performed by the Microanalytical Service of the Instituto de Investigaciones Químicas (Sevilla, Spain). Infrared spectra were obtained from Bruker Vector 22 spectrometer. The mass spectra were obtained at the Mass Spectroscopy Service of the University of Sevilla (CITIUS). The NMR instruments were Bruker DRX-500, DRX-400 and DPX-300 spectrometers. Spectra were referenced to external SiMe 4 (δ 0 ppm) using the residual protio solvent peaks as internal standards ( 1 H NMR experiments) or the characteristic resonances of the solvent nuclei ( 13 C NMR experiments). Spectral assignments were made by routine one- and two-dimensional NMR experiments where appropriate. All manipulations were performed under dry, oxygen-free dinitrogen, following conventional Schlenk techniques. The crystal structures were determined in a Bruker-Nonius, X8Kappa diffractometer. Metal complex [Cp*RhCl 2 ] 2 , 7b as well as NaBAr F28 were prepared as previously described. The lithium salt of the aminopyridinate ligands were prepared according to published procedures. 9b The 1 H and 13 C{ 1 H} NMR spectral data for the BAr F anion (BAr F = B[3,5- (CF 3 ) 2 C 6 H 3 ] 4 ) in CD 2 Cl 2 are identical for all complexes and therefore are not repeated below. 1 H NMR: δ 7.75 (s, 8 H, o-Ar), 7.58 (s, 4 H, p-Ar). 13 C{ 1 H} NMR: δ 162.1 (q, 1 J CB = 37 Hz, ipso- Ar), 135.3 (o-Ar), 129.2 (q, 2 J CF = 31 Hz, m-Ar), 124.9 (q, 1 J CF = 273 Hz, CF 3 ), 117.8 (p-Ar). Compound 1a·Cl. A toluene solution of the corresponding LiAp (240 mg, 0.66 mmoles; 2 mL) at -50 ºC was added to a suspension of [Cp*RhCl 2 ] 2 (200 mg, 0.32 mmoles) in toluene at -50 ºC. The resulting mixture was stirred, allowed to warm to room temperature and stirred for a further period of 5 h. The solution was filtered through celite and the solvent removed under reduced pressure. 1 H NMR analysis of the crude reaction mixture showed quantitative conversion into 1a·Cl, which was crystallized from Et 2 O-hexane mixtures at -23 ºC. 1 H NMR (C 6 D 6 , 25 ºC): δ = 22 7.32, 7.23, 7.21, (d, t, d, 1 H each, 3 J HH ~ 7.5 Hz, 3 CH Dipp ), 7.06, 7.02, 6.92 (t, d, d, 1 H each, 3 J HH ~ 7.5 Hz, 3 CH Xyl ), 6.65, 5.66, 5.42 (t, d, d, 1 H each, 3 J HH ~ 7.5 Hz, 3 CH Pyr ), 4.42, 3.53 (sept, 1 H each, 3 J HH ~ 7.0 Hz, 2 CH iPr ), 2.72, 2.21 (s, 3 H each, 2 Me Xyl ), 1.43, 1.35, 1.32, 1.23 (d, 3 H each, 3 J HH ~ 7.0 Hz, 4 Me iPr ), 1.03 (s, 15 H, 5 Me Cp* ). 13 C{ 1 H} NMR (C 6 D 6 , 25 ºC): δ = 172.4, 157.4 (C q-Pyr ), 148.0, 147.9, 139.6 (C q-Dipp ), 140.9, 138.7, 137.0 (C q-Xyl ), 136.8, 107.2, 106.5 (CH Pyr ), 128.0, 126.9 (2:1, CH Xyl ), 125.4, 124.3, 124.0 (CH Dipp ), 91.6 (d, 1 J RhC = 8 Hz, C q- Cp* ), 28.3, 28.0 (CH iPr ), 26.9, 25.8, 25.0, 23.6 (Me iPr ), 22.1, 20.1 (Me Xyl ), 8.8 (Me Cp* ). Anal. Calcd (%) for C 35 H 44 ClN 2 Rh: C, 66.6; H, 7.0; N, 4.4. Found: C, 66.5; H, 7.2; N, 4.1. Compounds 1b·Cl-1d·Cl: See the Supporting Information for synthetic details and characterization data. Compound [1a]BAr F . To a solution of 1a·Cl (374 mg, 0.593 mmol) in CH 2 Cl 2 (5 mL), NaBAr F (524 mg, 0.593 mmol) in CH 2 Cl 2 (3 mL) was added. Immediately, the colour of the solution turned from red to dark green as a consecuence of the formation of the cationic complex. The resulting mixture was filtered through celite, evaporated to dryness and the residue washed with pentane, to yield quantitatively complex [1a]BAr F . 1 H NMR (CD 2 Cl 2 , 25 ºC): δ = 7.30 (m, 7 H, 1 CH Pyr + 3 CH Xyl + 3 CH Dipp ), 6.20, 5.16 (d, 1 H each, 3 J HH ~ 7.5 Hz, 2 CH Pyr ), 3.53 (sept, 2 H, 3 J HH ~ 7.0 Hz, 2 CH iPr ), 2.32 (s, 6 H, 2 Me Xyl ), 1.44, 1.14 (d, 6 H each, 3 J HH ~ 7.0 Hz, 4 Me iPr ), 1.32 (s, 15 H, 5 Me Cp* ). 13 C{ 1 H} NMR (CD 2 Cl 2 , 25 ºC): δ = 179.9, 156.5 (C q-Pyr ), 145.6, 137.0 (2:1, C q-Dipp ), 144.4, 116.3, 102.9 (CH Pyr ), 137.6, 135.7 (1:2 C, C q-Xyl ), 130.0, 128.5 (1:2, CH Xyl ), 128.0, 124.6 (1:2, CH Dipp ), 94.7 (C q-Cp* ), 28.7 (CH iPr ), 25.2, 23.8 (Me iPr ), 20.2 (Me Xyl ), 9.5 (Me Cp* ). Anal. Calcd (%) for C 67 H 56 BF 24 N 2 Rh: C, 54.2; H, 3.9; N, 1.9. Found: C, 54.2; H, 4.5; N, 1.5. Compounds [1b]BAr F -[1d]BAr F : See the supporting Information for synthetic details and characterization data. 23 Compound [1a·NH 3 ]BAr F . NH 3 (g) was bubbled through a solution of compound [1a]BAr F (100 mg, 0.071 mmol) in CH 2 Cl 2 (5 mL) for 5 min. During this period of time the colour of the solution changed from dark green to bright red. The resulting mixture was stirred for 30 minutes and the volatiles were then removed under reduced pressure. 1 H NMR analysis of the crude product revealed quantitative conversion into complex [1a·NH 3 ]BAr F . 1 H NMR (CD 2 Cl 2 , 0 ºC): δ = 7.25 (m, 7 H, 3 CH Xyl + 3 CH Dipp + 1 CH Pyr ), 6.11, 5.74 (d, 1 H each, 3 J HH ~ 7.5 Hz, 2 CH Pyr ), 3.10, 2.66 (br s, 1 H each, 2 CH iPr ), 2.44 (s, 3 H, NH 3 ), 2.20, 2.17 (s, 3 H each, 2 Me Xyl ), 1.32, 1.12, 1.02 (br s, 2:1:1:, 4 Me iPr ), 1.25 (s, 15 H, 5 Me Cp* ). 13 C{ 1 H} NMR (CD 2 Cl 2 , 0 ºC): δ = 173.6, 156.8 (C q-Pyr ), 147.8, 143.7, 136.9 (br, C q-Dipp ), 138.9, 129.4, 125.9 (CH Xyl + CH Dipp + CH Pyr ), 138.5, 137.6, 133.9 (br, C q-Xyl ), 128.4, 127.9, 125.9, 123.3 (br, CH Dipp + CH Xyl ), 110.3, 108.0 (CH Pyr ), 94.0 (d, 1 J CRh = 8.2 Hz, C q-Cp* ), 29.4, 28.3 (CH iPr ), 26.0, 25.1, 24.0, 21.4 (Me iPr ), 20.7, 20.1 (Me Xyl ), 8.7 (CH Cp* ). Anal. Calcd (%) for C 67 H 54 BF 24 N 3 Rh: C, 54.5; H, 4.0; N, 2.8. Found: C, 54.5; H, 4.0; N, 2.7. Adducts of complex 1a + with dmap, CNXyl and NCMe were prepared by a similar procedure (see SI). Compound [1a·(CO) 2 ]BAr F CO (g) was bubbled through a solution of compound [1a]BAr F (100 mg, 0.071 mmol) in CH 2 Cl 2 (5 mL) for 5 min. During this period of time the colour of the solution changed from dark green to yellow-orange. The resulting mixture was stirred for 3 h and the volatiles were then removed under reduced pressure. 1 H NMR analysis of the crude product revealed quantitative conversion into complex [1a·(CO) 2 ]BAr F , which was crystallized from CH 2 Cl 2 -Et 2 O-hexane mixtures at -23 ºC. IR (Nujol): ν(CO) 2070, ν(CO amide ) 1690 cm -1 . 1 H NMR (CD 2 Cl 2 , 25 ºC): δ = 7.78, 7.00, 6.51 (t, d, d, 3 J HH ~ 7.5 Hz, 3 CH pyr ), 7.60, 7.40 (t, d, 1:2, 3 J HH ~ 7.5 Hz, 3 CH Dipp ), 7.43, 7.31, 7.27 (m, d, d, 1 H each, 3 J HH ~ 7. 5 Hz, 3 CH Xyl ), 2.67, 2,32 24 (sept, 1 H each, 3 J HH ~ 7.0 Hz, 2 CH iPr ), 2.17, 2.16 (s, 3 H each, 2 Me Xyl ), 1.59 (s, 15 H, 5 Me Cp* ), 1.26, 1.18, 1.12, 1.04 (d, 3 H each, 3 J HH ~ 7.0 Hz, 4 Me iPr ). 13 C{ 1 H} NMR (CD 2 Cl 2 , 25 ºC): δ = 188.4, 186.8 (d, 1 J CRh = 30 Hz, 1 J CRh = 75 Hz, Rh-CON and Rh-CO, resp.), 160.2, 158.6 (C q-Pyr ), 146.2, 146.0 (C q-Dipp ), 141.4, 123.6, 111.0 (CH Pyr ), 138.6, 136.3, 133.2 (C q-Xyl + C q-Dipp ), 137.2 (C q.Xyl ), 131.4, 125.8, 125.6 (CH Dipp ), 130.8, 129.4, 129.2 (CH Xyl ), 109.6 (d, 1 J RhC = 5 Hz, C q-Cp* ), 30.0, 29.5 (CH iPr ), 24.3, 24.1, 23.9, 23.5 (Me iPr ), 22.1, 21.3 (Me Xyl ), 9.4 (Me Cp* ). Anal. Calcd (%) for C 69 H 56 BF 24 N 2 O 2 Rh: C, 54.7; H, 3.7; N, 1.8. Found: C, 54.8; H, 4.0; N, 1.8. Related carbonyl derivatives of 1b + -1d + were prepared by an analogous procedure (see SI). Compound [2a]BAr F . In a Young NMR tube, a solution of complex [1a]BAr F (20 mg, 0.014 mmol) in CD 2 Cl 2 (0.5 mL) was treated with H 2 (1 atm). After 5 h at room temperature 1 H NMR analysis of the reaction mixture revealed transformation into complex [2a]BAr F in ≥95 % spectroscopic yield. IR (Nujol): ν(NH) 3426 cm -1 . 1 H NMR (CD 2 Cl 2 , 25 ºC): δ = 7.52, 6.57, 6.13 (t, d, d, 1 H each, 3 J HH ~ 7.5 Hz, 3 CH Pyr ), 7.37, 7.11, 6.50 (m, d, t, 1 H each, 3 J HH ~ 7.5 Hz, 3 CH η6-Xyl ), 7.27 (br d, 3 H, 3 J HH ~ 7.5 Hz, 3 CH Dipp ), 6.00 (br s, 1 H, NH), 3.18 (sept, 2 H, 3 J HH ~ 6.5 Hz, 2 CH iPr ), 3.00 (br s, 1 H, CHMe), 2.12 (br s, 6 H, 2 Me η6-Xyl ), 2.06, 1.47 (s, 6 H each, 4 Me η4-Cyclopentadiene ), 1.16 (d+d, 12 H, 3 J HH ~ 6.5 Hz, 4 Me iPr ), 0.56 (d, 3 H, 3 J HH ~ 6.5 Hz, CHMe). 13 C{ 1 H} NMR (CD 2 Cl 2 , 25 ºC): δ = 160.1, 151.1 (C q-Pyr ), 148.2, 136.3 (C q-Dipp ), 139.2, 114.7, 100.8 (CH Pyr ), 129.0, 128.0, 103.0 (CH Xyl ), 124.8, 115.2 (2:1, C q-η6-Xyl ), 102.8, 79.8 (d, 1 J CRh = 10 Hz, C q-η4-Cyclopentadiene ), 59.7 (CHMe), 29.1 (CH iPr ), 23.8 (Me iPr ), 23.2 (CHMe), 18.5 (Me η6-Xyl ), 14.0, 11.6 (Me η4-Cyclopentadiene ). HRMS (FAB): m/z calcd for C 35 H 46 N 2 Rh [M] + : 597.2716. Found: 597.2682. 25 Compound [2b]BAr F . Following the procedure described above, complex [2b]BAr F was obtained after 24 h in 85 % spectroscopic yield, and it was crystallized from CH 2 Cl 2 -pentane mixtures at -23 ºC. 1 H NMR (CD 2 Cl 2 , 25 ºC): δ = 7.44, 6.46, 6.17 (t, d, d, 1 H each, 3 J HH ~ 7.5 Hz, 3 CH Pyr ), 7.08 (s, 3 H, 3 CH Xyl ), 6.39, 6.01 (t, d, 1:2, 3 J HH ~ 6.5 Hz, 3 CH η6-Xyl ), 5.93 (br s, 1 H, NH), 2.78 (q, 1 H, 3 J HH ~ 6.5 Hz, CHMe), 2.15 (s, 6 H, 2 Me Xyl ), 1.94 (s, 6 H, 2 Me η6-Xyl ), 1.95, 1.32 (s, 6 H each, 4 Me η4-Cyclopentadiene ), 0.43 (d, 3 H, 3 J HH = 6.3 Hz, CHMe). 13 C NMR (CD 2 Cl 2 , 25 ºC): δ = 159.2, 151.5 (C q-Pyr ), 139.6, 115.3, 107.8 (CH Pyr ), 137.8, 129.7 (2:1, C q-Xyl ), 129.5 (CH Xyl ), 124.3, 115.4 (1:2, C q-η6-Xyl ), 103.3, 101.2 (s, d, 1:2, 1 J CRh = 4 Hz, CH η6-Xyl ), 103.2, 80.2 (d, 1 J CRh ~ 10 Hz, C q-η4-Cyclopentadiene ), 59.9 (CHMe), 23.8 (CHMe), 19.0 (Me η6-Xyl ), 18.9 (Me Xyl ), 14.4, 12.0 (Me -η4-Cyclopentadiene ). Anal. Calcd (%) for C 63 H 50 BF 24 N 2 Rh: C, 53.9; H, 3.6; N, 2.0. Found: C, 53.5; H, 3.8; N, 2.1. The generation of the related complexes [2c]BAr F and [2d]BAr F is detailed in the accompanying SI. Compound [3a]BAr F . Method a: In a Young NMR tube, a solution of complex [1a]BAr F (0.04 g, 0.03 mmol) in CH 2 Cl 2 (0.5 mL) was treated with H 2 (200 mol %). After 24 h at room temperature 1 H NMR analysis of the reaction mixture revealed transformation into complex [3a]BAr F in 6% spectroscopic yield. 32 Insert Table of Contents Graphic and Synopsis Here A study on the reactivity of cationic Rh(III) complexes of composition [(η 5 - C 5 Me 5 )Rh(Ap)] + (complexes 1 + ) with H 2 is reported. The main products result from a stoichiometric hydrogenation of the C 5 Me 5 and aminopyridinate ligands to C 5 Me 5 H and ApH, respectively, with concomitant metal reduction to Rh(I) (complexes 2 + ). Additionally, a dihydrogen-catalyzed isomerization of the aminopyridinate ligand from the common κ 2 -N,N´ bidentate coordination to an unusual κ-N-η 3 -pseudo-allyl bonding mode (Rh complexes 3 + ) is observed.