Combining modern tools in the study of physical properties, reaction processes and mechanisms

PROGRAMA DE DOCTORADO EN QUÍMICA: QUÍMICA DE SÍNTESIS, CATÁLISIS Y MATERIALES AVANZADOS TESIS DOCTORAL: COMBINING MODERN TOOLS IN THE STUDY OF PHYSICAL PROPERTIES, REACTION PROCESSES AND MECHANISMS Presentada por Marconi Nicolás Peñas de Frutos para optar al grado de Doctor por la Universidad de Valladolid Dirigida por: Prof. Dr. Pablo Espinet Rubio Dra. Camino Bartolomé Albistegui La Tesis Doctoral titulada “Combining modern tools in the study of physical properties, reaction processes and mechanism” ha sido realizada gracias al apoyo económico del Ministerio de Educación, Cultura y Deporte (contrato predoctoral de Formación de Profesorado Universitario, FPU-2015), al Ministerio de Economía y Competitividad (proyectos CTQ2013-48406-P, CTQ2014-52796-P, CTQ2016-80913-P, CTQ2017-89217-P), y de la Junta de Castilla y León (proyectos VA256U13, GR169, VA051P17, VA062G18, VA038G18 y UIC176). Quiero agradecer especialmente al resto de los coautores de las publicaciones incluidas en este trabajo, Dr. Max García-Melchor, Dra. Andrea Vélez, Dra. Verónica Conejo, Dra. Zoraida Ramiro y Dra. Estefanía Gioria, así como el trabajo desarrollado por Sara Fernández en parte de los resultados presentados y que actualmente se están organizando para su publicación. Sin todas las personas que he mencionado gran parte de la tesis no podría haberse llevado a cabo. Asimismo, se agradece la ayuda puntual del Dr. José Miguel Martín, de Sergio Ferrero y de Jonathan Martínez. Agradecimientos El camino ha sido largo y duro, pero lo hubiese sido mucho más sin la ayuda y compañía de mucha gente especial. Sirva este espacio para agradecer a las personas que han hecho posible llegar hasta aquí. En primer lugar, debo agradecer la dedicación incansable de los directores de tesis. Al Prof. Pablo Espinet por todo lo que he aprendido de él tanto de química como de la vida, pero más aún por haberme comprendido, y haberme apoyado siempre que lo he necesitado personal y profesionalmente. A la Dra. Camino Bartolomé por haberme captado para esta aventura, por su disponibilidad y haber tenido fe en mi desde el principio. Sin ellos no sería la persona que soy hoy. Al Prof. Max García-Melchor por el tiempo que pasé de estancia en su grupo en el Trinity College Dublin. Por su hospitalidad y su paciencia para enseñarme a pesar de mi torpeza. También por su inestimable colaboración en parte de este trabajo. Gracias también a Eric, a Tom y Helen por hacer de la estancia una experiencia inolvidable. A todos los profesores, especialmente a JM, por ser a la persona a la que siempre he llamado para pedir ayuda. También a Ángel y Jesús Ángel por muchos ratos juntos. A todos los amigos de verdad que he hecho en esta etapa. A Pollo, Tomaz, Bea, Sara, Wonka y Andrea. Gracias a los cafés en compañía, ir a trabajar no ha sido tan duro y sé que cuento con ellos para lo que necesite. Al grupo de amigos de la universidad y a mis amigos de siempre, Enrique, Miguel, Adri, Chechu, Javi y María Valladares. A los demás compañeros del QUIFIMA durante todos estos años, gracias a Vero, Estefanía, Sheila, Nacho, Estela, Desi, Marta, las dos Marías, Vanesa, Olmo, Ana, Jorge, Zoraida, Beto, Lucía, Sergio, Jaime y Rodri. Gracias a María que ha sido la que me ha aguantado todo este tiempo y no creo que haya sido fácil. Por su paciencia y por estar siempre ahí. Eres la mejor compañera que podría tener. A mi familia, especialmente a mis padres por su apoyo incondicional, a mi hermana y a mis abuelos. Sé que la persona a la que más ilusión le haría tener un nieto doctor me verá desde donde esté. A mi nueva familia, a Charo y a los niños. Gracias a todos Contents Resumen .......................................................................................................................... 1 Introduction .................................................................................................................... 6 Chapter I: R/R’ and R/Cl exchanges in the RhI/AuI bimetallic system. Mechanisms and opposite kinetic effect produced by extra ligand ............................................................ 13 Chapter II: Unexpected Mechanism for Aryl Exchanges in 5-coordinate Rh Species Stabilized by the Buffer effect of Cp*. Insights into the Modulation of the Cp* Group Coordination Determined by the Trans Influence of the Other Ligands ........................ 20 Chapter III: Reactivity of Fluorinated-Chalcone Phosphines Induced upon Coordination to PdCl2. E/Z Isomerization Triggers C‒F Activation.............................. 31 Chapter IV: Protection of Gold(I) Catalysts by Substoichiometric Agents. Is the Decomposition to Metallic Gold a Simple Reduction? .................................................. 36 Chapter V: The Key Role of d8···d10 and d10···d10 Interactions in the Photophysical Properties of Crystals. Packing vs DFT .......................................................................... 42 Conclusions ................................................................................................................... 50 Methods ......................................................................................................................... 53 List of Publications ....................................................................................................... 55 1 Resumen Resumen La investigación innovadora en química requiere cada vez más el uso adecuado y complementario de metodologías mixtas tanto experimentales como computacionales. Esta probablemente es la mejor línea argumental para relacionar los distintos proyectos abordados durante mi tesis doctoral. La memoria está dividida en cinco capítulos con temáticas muy distintas. Sin embargo, todos los trabajos comparten el enfoque conceptual y el interés de aplicar diferentes herramientas para entender profunda y rigurosamente los problemas químicos propuestos. Para ello es fundamental ser consciente del potencial, pero también de las limitaciones de cada una de las técnicas empleadas. Las técnicas en las que más he empleado a lo largo de esta tesis han sido la resonancia magnética nuclear (RMN), las simulaciones cinéticas con software específico (COPASI), los cálculos computacionales DFT y la difracción de Rayos X. El objetivo es describir distintas estrategias de uso sinérgico de todas ellas, sin perder de vista que las herramientas computacionales son un apoyo muy valioso de los datos experimentales, pero no deben sustituirlos. Capítulo I: Estudios de transmetalación Ar/Ar’ y Ar/Cl en el sistema bimetálico RhI/AuI Antes del inicio de este proyecto de tesis no existían estudios mecanísticos sobre reacciones de transmetalación entre complejos de rodio (I) y oro (I), interesantes para su eventual uso posterior en procesos de catálisis bimetálica. Se han estudiado mecanísticamente, combinando datos experimentales, simulaciones cinéticas y cálculos DFT, dos reacciones de intercambio entre complejos de dichos metales: intercambio Ar/Ar’ (Ar = haloarilo) e intercambio Ar/Cl. Lo más llamativo de este estudio es el desigual efecto cinético observado tras la adición de cantidades variables de ligando libre (AsPh3) en la velocidad de la transmetalación. Mientras que en el intercambio Ar/Ar’ se ve frenado por el ligando, con un orden fraccionario y negativo, la AsPh3 acelera de forma no lineal el intercambio Ar/Cl. El perfil de reacción Ar/Ar’ tiene varias etapas con estados de transición próximos en energía, produciéndose en uno de ellos la disociación de un ligando AsPh3. En el caso de la transmetalación Ar/Cl hay dos mecanismos competitivos: el primero no requiere ligando adicional, y el segundo está catalizado por AsPh3. 2 Resumen Los cálculos teóricos revelan que estas reacciones no se producen a través del mecanismo típico, si no que tienen lugar a través de procesos de inserción oxidante que conducen a intermedios con enlaces metal‒metal. Capítulo II: Mecanismo de intercambio de arilos en complejos de 16e con RhCp*. Efecto tampón electrónico del Cp* Es bien conocido que la reactividad de los complejos de 18 electrones basados en el framento RhIIICp* (Cp* = pentametilciclopentadienilo) se inicia con la disociación de un ligando formando así intermedios activos de 16 electrones. Bajo esta premisa se pretendía obtener complejos pentacoordinados estables y estudiar reacciones de transmetalación. La reacción de transmetalación con haloarilderivados de plata conduce a los complejos de fórmula [RhCp*Ar2]. El seguimiento por RMN de una disolución de [RhCp*Ar2] y [RhCp*Ar’2] confirma que se produce el intercambio Ar/Ar’. La propuesta mecanística de intercambio directo a través de un estado de transición con doble puente arilo, aceptada habitualmente para otras reacciones de transmetalación, fue descartada mediante cálculos DFT. En nuestro caso, se ha podido constatar experimental y computacionalmente que el producto de hidrólisis (μ-OH)2[RhCp*Ar]2, presente en forma de trazas, actúa como catalizador de la reacción de intercambio. Posteriormente se decidió investigar la elevada estabilidad de los compuestos pentacoordinados [RhCp*Ar2], precursores de gran variedad de diaril y monoaril derivados de 18e. El análisis estructural de todos los complejos sintetizados permitió apoyar nuestra hipótesis de que el grupo Cp* responde a las variaciones de las necesidades electrónicas del centro metálico como consecuencia de la naturaleza de los ligandos que ocupan el resto de posiciones de coordinación. Hemos denominado este fenómeno como efecto tampón electrónico del Cp*. El análisis pormenorizado de los datos estructurales nos ha permitido establecer una serie de influencia trans en un sistema octaédrico. Por último, se han estudiado los correspondientes complejos con ligandos carbonilo y cianuro (CO y CN‒) mediante cálculos NBO. Este análisis demuestra la existencia de donaciones laterales de orbitales del grupo Cp* a los orbitales vacíos de estos ligandos πaceptores. Dichas donaciones se suman a la retrodonación clásica por parte del metal y deben ser consideradas tanto en este como en otros sistemas análogos. 3 Resumen La primera parte demuestra la conveniencia de apoyar propuestas mecanísticas aparentemente obvias con experimentos y cálculos. La segunda es un buen ejemplo de las propuestas conceptuales que se pueden derivar de un análisis estructural profundo, por lo que este es un claro ejemplo de la importancia de un estudio exhaustivo de los datos obtenidos por difracción de rayos X. Capítulo III: Reactividad de los complejos con PdCl2 y ligandos PEWO coordinados como quelato El grupo de investigación en el que he desarrollado mi tesis doctoral había diseñado previamente ligandos híbridos fosfina olefina electrónicamente deficiente (PEWO) que han dado buen resultado en catálisis de tipo Negishi y como inductores de homoacoplamientos difíciles de haloarilos con paladio. Estos ligandos poseen un esqueleto de tipo chalcona fluorada (estiril-fenil-cetona) en la olefina. El objetivo ha sido sintetizar nuevos ligandos y estudiar su reactividad una vez coordinados a PdCl2. Mientras que las PEWO libres o coordinadas únicamente por el fósforo muestran configuración E en la olefina, una vez coordinadas como quelato, se observa la configuración Z como la más estable. Dicha isomerización es posible (rápida a temperatura ambiente) gracias a la deslocalización electrónica que facilita la rotación, proceso descrito para otras olefinas de tipo chalcona. La isomerización E/Z requiere la disociación de la olefina del paladio, permitiendo atrapar la poco habitual configuración Z del doble enlace. La configuración Z en la olefina es la responsable de la interesante reactividad observada en este sistema y que se inicia por un proceso de activación de un enlace C‒F, geométricamente inaccesible desde la configuración E y, por tanto, observado únicamente tras la coordinación a paladio de forma quelato. La activación C‒F y ciclación posterior conduce a la formación de complejos con ligandos híbridos fosfina-carbeno, que posteriormente experimentan procesos controlados de hidrólisis, amonólisis y oxidación con peróxidos presentes en THF envejecido, obteniéndose derivados con ligandos pincer PCO, PCN y bidentado PO respectivamente. Todos estos procesos deben ser tenidos en cuenta cuando se exploren nuevas reacciones con los ligandos tipo PEWO, ya que, si son lentos, pueden interferir con la catálisis. 4 Resumen Capítulo IV: Protección de los catalizadores de AuI frente a la descomposición mediante la adición subestequiométrica de AsPh3 Los catalizadores de oro (I), generalmente denotados como [AuL]+ (L = ligando neutro), formados tras la extracción de haluro con una sal de plata son extremadamente activos, por ejemplo, en reacciones de ciclación de eninos. Sin embargo, sufren rápidamente una caída abrupta en su actividad asociada a la descomposición del catalizador, proceso cuyo mecanismo apenas ha sido explorado. Hemos demostrado experimentalmente que en el sistema [AuCl(carbeno)] la descomposición puede inhibirse completamente mediante la adición de una cantidad subestequiométrica de AsPh3 tanto en condiciones catalíticas como en ausencia de enino. Afortunadamente, la actividad catalítica se mantiene, aunque su velocidad se ve disminuida ligeramente. La protección subestequiométrica apoya el mecanismo de desproporcionación (3AuI para dar 2 Au0 y 1 AuIII) para la descomposición del catalizador y permite descartar la simple reducción como fuente del oro metálico observado. El papel del ligando libre generado por equilibrios entre las especies [AuL(W)]+ (W = ligando débilmente coordinado) es clave. Así, un mejor ligando, como PPh3, no es capaz de evitar tan eficazmente la descomposición en cantidades subestequiométricas como un peor ligando, en nuestro caso AsPh3. Otros potenciales ligandos, siempre presentes en procesos catalíticos y generalmente ignorados, como son las trazas de agua en el seno de reacción son capaces de proteger pequeñas cantidades de catalizador, lo que permite explicar la actividad residual observada tras la masiva descomposición. Capítulo V: Interacciones d8···d10 y d10···d10 factor clave en las propiedades fotofísicas de materiales cristalinos Las interacciones d8···d8 condicionan el color de los cristales en los complejos catiónicos de tipo [Rh(CNR)4]nYn (CNR = isocianuro). Asimismo, las interacciones aurofílicas d10···d10 determinan la luminiscencia observada en estado sólido en los complejos [AuAr(CNR)]. Nuestro objetivo ha sido estudiar las modificaciones que podrían surgir en estos sistemas cuando se incorpora un segundo metal. 5 Resumen El ligando xililisocianuro es capaz de dificultar el establecimiento de interacciones de tipo π-π stacking que favorecen la formación de interacciones Rh···Rh, lo que ha permitido obtener estructuras cristalinas del complejo [Rh(CNXylyl)4][Au(CN)2], donde las interacciones heterometálicas d8···d10 prevalecen por completo. Estas interacciones, energéticamente débiles y condicionadas por el empaquetamiento, son las responsables del color del cristal. Se han sintetizado complejos homometálicos del tipo [AuAr(CNPy)] (Ar = haloarilo), que presentan luminiscencia en estado sólido y mecanocromismo. El ligando 4piridilisocianuro (CNPy) es capaz de actuar como bidentado enlazándose selectivamente a la plata por la piridina y al oro por el extremo isocianuro, de tal forma que se han podido obtener los complejos trinucleares heterobimetálicos [{AuAr(CNPy)}2Ag]BF4, que muestran luminiscencia naranja en estado sólido. Los cambios en la luminiscencia en todos los casos se deben a modificaciones en las interacciones Au···Au, de forma que distancias más cortas desplazan la emisión hacia el rojo. Los datos obtenidos mediante cálculos DFT, congelando las geometrías de fragmentos representativos de las estructuras cristalinas, nos han permitido ilustrar los orbitales participantes en las transiciones electrónicas responsables de las propiedades fotofísicas observadas. 12 Introduction mechanism of a new organometallic reaction is closely related to one known for related organic reactions (see chapter III). Furthermore, the experimental ability is basic in the planning of experiments, whether synthetic or of any type. This has included work under inert atmosphere, use of low temperatures, synthesis of unreported organometallic complexes, and planning and development of catalytic and non-catalytic experiments, both in the flask and in the NMR tube. Other spectroscopic and non-spectroscopic characterisation techniques have been commonly used, such as infrared spectroscopy, especially informative in derivatives bearing carbonyl (CO) or isocyanide (CNR) ligands as well as in molecules with ketonic groups. Mass spectrometry, UV-vis and luminescence spectroscopy, including emission lifetime and quantum yield measurements, as well as powder X-ray diffraction studies have been used sporadically. In summary, the main objective of this thesis is to describe strategies to carry out mechanistic investigations or other types of challenging studies combining the information obtained through various experimental and computational tools. It has been my privilege to have the opportunity to exploit and apply personally these several types of methodology in a synergistic way. 13 CHAPTER I RhI/AuI Transmetalation Chapter I: R/R’ and R/Cl exchanges in the RhI/AuI bimetallic system. Mechanisms and opposite kinetic effect produced by extra ligand Bimetallic catalysis concerns homogeneous processes where two transition metals (TMs), or one TM and one Group 11 element (M), cooperate in a synthetic transformation (for example C‒C coupling), in which their two catalytic cycles are linked by a transmetalation step. 1 Our research group has much experience in the elucidation of mechanisms for different group-exchanges involving two transition metals (Au/Pd, 2 or Pd/Cu, 3 among others) by combining experimental (initial rate or kinetic orders) and theoretical (DFT calculations and kinetic simulations using the COPASI software) information. The typical mechanistic proposal for the R1/X (X = halide) exchange between [M]‒R1 nucleophiles and [TMR2]‒X electrophiles involves a cyclic transition-state structure with the two metals connected through mixed R1/X bridges (equation 1). The reversibility of these processes and the eventual existence of secondary transmetalations (R1/R2) can complicate the result. 4 Transmetalation reactions are important not only in cross coupling processes. They can operate whenever [M]‒X or [M]‒R compounds (with different or identical metals) coexist in solution. Surprisingly, despite their ubiquitous presence and important 1 (a) Pérez-Temprano, M. H.; Casares, J. A.; Espinet, P. Bimetallic Catalysis using Transition and Group 11 Metals: An Emerging Tool for C‒C Coupling and Other Reactions. Chem. ‒ Eur. J. 2012, 18, 1864. DOI: 10.1002/chem.201102888. (b) Pie, D. R.; Mankad, N. P. Bimetallic catalysis for C–C and C–X coupling reactions. Chem. Sci. 2017, 8, 1705. DOI: 10.1039/C6SC05556G. 2 Pérez-Temprano, M. H.; Casares, J. A.; de Lera, A. R.; Álvarez, R.; Espinet, P. Strong Metallophilic Interactions in the Palladium Arylation by Gold Aryls. Angew. Chem. Int. Ed. 2012, 51, 4917. DOI: 10.1002/anie.201108043. 3 (a) delPozo, J.; Casares, J. A.; Espinet, P. In Situ Generation of ArCu from CuF2 Makes Coupling of Bulky Aryl Silanes Feasible and Highly Efficient. Chem. ‒ Eur. J. 2016, 22, 4274. DOI: 10.1002/chem.201504435. (b) Pérez-Iglesias, M.; Lozano-Lavilla, O.; Casares, J. A. [Cu(C6Cl2F3)(tht)]4: An Extremely Efficient Catalyst for the Aryl Scrambling between Palladium Complexes. Organometallics 2019, 38, 739. DOI: 10.1021/acs.organomet.8b00885. 4 del Pozo, J.; Salas, G.; Álvarez, R.; Casares, J. A.; Espinet, P. The Negishi Catalysis: Full Study of the Complications in the Transmetalation Step and Consequences for the Coupling Products. Organometallics 2016, 35, 3604. DOI: 10.1021/acs.organomet.6b00660. 14 CHAPTER I RhI/AuI Transmetalation consequences, these exchanges have been often overlooked and their mechanisms are still little studied. Recently, our group reported the drastic mechanistic switch occurring in the Ar/X exchange between SnPh(nBu)3 and [AuIXL] complexes: The typical concerted mechanism involving Ar/X mixed bridges operates when X = Cl, whereas an oxidative addition/reductive elimination (OA/RE) pathway via an Au‒Sn bonded intermediate takes over when X = vinyl. 5 As far as we know, at the start of our thesis project no studies focussed on gold(I)/rhodium(I) transmetalation reactions had been published. 6 We decided to explore the transmetalation possibilities and the mechanisms involved with this bimetallic pair. Understanding of the mechanisms of the R/X and R1/R2 transmetalation reactions might eventually help to the design of successful processes of bimetallic catalysis. The two reactions studied were the Rf/Pf (Rf = C6F3Cl2-3,5; Pf = C6F5) haloaryl scrambling equilibrium between trans-[RhRf(CO)(AsPh3)2] (1) and [AuPf(AsPh3)] (2), and the Rf/Cl exchange between the same rhodium reactant and the chloro-gold derivative [AuCl(AsPh3)] (5) (Scheme 1). Scheme 1. Reactions studied in this work. Above: Rf/Pf exchange. Below: Rf/Cl exchange. The Vaska type rhodium complexes used in these studies display trans disposition of the AsPh3 ligands. These gold and rhodium complexes do not show appreciable dissociation equilibria in solution, and the use of the same ligand (AsPh3) for all the complexes prevents the formation of new species via neutral ligand exchange, ubiquitously present 5 Carrasco, D.; García-Melchor, M.; Casares, J. A.; Espinet, P. Dramatic mechanistic switch in Sn/AuI group exchanges: transmetalation vs. oxidative addition. Chem. Commun. 2016, 52, 4305. DOI: 10.1039/C5CC10496C. 6 For a Rh/Au transmetalation with Cp*RhIII derivatives see: Shi, Y.; Blum, S. A. Gold and Rhodium Transmetalation: Mechanistic Insights and Dual-Metal Reactivity. Organometallics 2011, 30, 1776. DOI: 10.1021/om2001316. 15 CHAPTER I RhI/AuI Transmetalation in many systems in solution. Furthermore, the haloaryl groups (Rf and Pf) allow for simple and precise monitoring of the reactions by 19F NMR. The two reactions collected in Scheme 1 are selective, and neither side-products nor intermediates could be observed. However, the thermodynamics of the two processes differ, as reflected in their equilibrium constants (Keq ≈ 1 for Rf/Pf exchange; Keq ≈ 7 × 103 for Rf/Cl exchange). The second and the most striking difference between both processes is the opposite effect on rate provoked upon the addition of free AsPh3 (Figure 1). Figure 1. Kinetic effect of different amounts of added AsPh3 observed by 19F NMR monitoring of the corresponding reaction at a certain temperature (details in each article). Above: Data for Rf/Pf exchange. Below: Data for Rf/Cl exchange, lines represent data adjusted with COPASI overlapped with experimental data (triangles). In the Rf/Pf scrambling, the deceleration in the presence of AsPh3 deserves careful analysis. The kinetic order is fractional and close to ‒0.5. This deviation to the ideal order of ‒1 points to a Rf/Pf exchange mechanism with several steps with associated transition 16 CHAPTER I RhI/AuI Transmetalation states (TSs) similar in energy. 7 In one of those steps, the dissociation of one ligand must be taking place, explaining the negative order observed for the ligand. On the other hand, the fact that only 5 mol% of AsPh3 added diminishes the rate to the half indicates that this 5 mol% of AsPh3 relative to the starting materials must be much higher percentage relative to the species involved at the step where AsPh3 dissociates. With regard to the Rf/Cl exchange reaction, the non-linear accelerating effect of AsPh3 suggests coexistence of two competitive pathways leading to the same products: one unaffected by extra ligand, and another AsPh3-catalysed. The active species of this AsPh3catalysed mechanism must be the tricoordinate [AuCl(AsPh3)2] (7), indeed detected by mass spectrometry in reaction conditions. The concentrations of the two gold complexes [AuCl(AsPh3)] (5) and [AuCl(AsPh3)2] (7) are connected by the ligand coordination equilibrium in Scheme 2. Kinetic simulations with COPASI software (lines in Figure 1 below) allow us to fit the equilibrium constant at 273 K (Keq = 170) and the activation barriers for both pathways (19.2 and 17.7 kcal mol‒1, respectively). Scheme 2. Kinetic model for COPASI fitting. With these experimental references, and taking into account that these are reversible processes, we performed DFT calculations in order to get complete mechanistic details of both exchange reactions, Rf/Pf and Rf/Cl. The mechanism proposed for the Rf/Pf exchange is a completely reversible pathway with oxidative addition and reductive elimination steps in the extremes, connected by dissociation and association AsPh3 steps flanking a central rate limiting step (rds). This rds is an isomerization process between two pentacoordinated rhodium intermediates that display strong Rh‒Au bonds (Figure 2). 7 A similar fractional order was previously reported for a Pd/Au transmetalation reaction. See ref. 2. 17 CHAPTER I RhI/AuI Transmetalation Figure 2. Gibbs energy profile combining DFT (grey and blue lines) and simulated values (red and pink lines). The first heterometallic bonded intermediate I2 is formed through a transition state (TS1) in which the oxidative insertion 8 of the rhodium centre into the Au‒Ar bond occurs. Interestingly, the stereochemistry of this oxidative process differs from the one observed for the oxidative addition of H2 to Vaska’s derivative trans-[MCl(CO)(PPh3)2] (M = Ir, Rh) because in that case the Cl‒M‒CO axis undergoes angle bending. 9 In our case, the oxidative addition step is initiated by electronic donation from rhodium to gold. This favours the formation of the M‒M’ bond that concomitantly triggers the aryl transmetalation, as supported by NBO studies. Regarding the Rf/Cl transmetalation reaction, since the products at both sides of the equilibrium are different, the profiles of reaction are expected to be quite asymmetric. We have calculated the two mechanistic pathways, the uncatalysed (Figure 3) and ligandcatalysed (Figure 4). Both pathways start with a TS in which the oxidative insertion of Au into the Rh‒Rf bond takes place (note that in the Rf/Pf case the insertion of Rh into the Au‒Pf is the one operating). The extra AsPh3 coordinated to Au in 7 eases the aryl transference because it is able to weaken the Au‒Cl bond that is even concomitantly cleaved in the first step of ligand-catalysed pathway (Figure 4). 8 We call “oxidative insertion” to the oxidative addition giving rise to formation of a M–M’ bond, since it resembles the result of that process. 9 Vaska, L. Reversible activation of covalent molecules by transition-metal complexes. The role of the covalent molecule. Acc. Chem. Res. 1968, 1, 335. DOI: 10.1021/ar50011a003. 18 CHAPTER I RhI/AuI Transmetalation Figure 3. DFT Gibbs energy profile obtained for the uncatalysed pathway. Figure 4. DFT Gibbs energy profile obtained for the AsPh3-promoted pathway. Ligand labelled in orange is completely dissociated after the aryl transmetalation. Finally, as a main and general conclusion of this Chapter, the results presented allow us to state that even very related reactions can hide completely different mechanisms and only the combination of appropriate experimental and computational techniques offers the higher chances to ascertain a mechanistic proposal. Although the possibilities offered 19 CHAPTER I RhI/AuI Transmetalation by both approaches have been clearly demonstrated in this study, one must also be aware of their limitations. There is also a more particular conclusion: The widely extended idea that transmetalations occur through a typical transition state with mixed bridges must be reconsidered in view of the results of this specific study. Perhaps some mechanistic proposals should be revisited, at least for oxidizable metal centres, considering an oxidative insertion alternative. 20 CHAPTER II RhCp*: Hidden exchanges and buffer effect Chapter II: Unexpected Mechanism for Aryl Exchanges in 5-coordinate Rh Species Stabilized by the Buffer effect of Cp*. Insights into the Modulation of the Cp* Group Coordination Determined by the Trans Influence of the Other Ligands Pentamethylcyclopentadienyl (Cp*) rhodium(III) and iridium(III) complexes have been widely used in organometallic chemistry since the 1960’s. 10 The MIIICp* 18 electron complexes typically display piano stool geometry, and their reactions are assumed to start with ligand dissociation to give 16e electrophilic intermediates. These 16e intermediates have been proposed as the active species in many catalytic processes, particularly in some C‒H activation reactions reported very recently. 11 , 12 We aimed at studying the haloaryl transmetalation possibilities in RhIIICp* derivatives using a methodology similar to that in the previous chapter. We decided to use as precursor the dimer (µ-Cl)2[RhCp*Cl]2 (1), which has a modern synthesis considerably improved using microwave heating. 13 Transmetalation of 1 with the silver haloaryl derivatives AgAr·nNCMe (Ar = Rf, Pf) 14 afforded the deep red [RhCp*Rf2] (2) and the yellow [RhCp*Pf2(NCMe)] (3) complexes, respectively. Interestingly, the complex [RhCp*Rf2] is a perfectly stable pentacoordinated species and its molecular structure differs from the typical piano stool geometry. The analogous [RhCp*Pf2] (4), structurally equivalent to 2 (Figure 1), could be obtained heating 3 in vacuum. 10 Kang, J. W.; Maitlis, P. M. Conversion of Dewar hexamethylbenzene to pentamethylcyclopentadienylrhodium(III) chloride. J. Am. Chem Soc. 1968, 90, 3259. DOI: 10.1021/ja01014a063. 11 For a particular example of active 16e species see: Wu, Q.; Chen, Y.; Yan, M.; Lu, Y.; Sun, W.-Y.; Zhao, J. Unified synthesis of mono/bis-arylated phenols via RhIII-catalyzed dehydrogenative coupling. Chem. Sci. 2017, 8, 169. DOI: 10.1039/C6SC03169B. 12 For RhCp derivatives in C‒H activation see: Piou, T.; Rovis, T. Electronic and Steric Tuning of a Prototypical Piano Stool Complex: Rh(III) Catalysis for C–H Functionalization. Acc. Chem. Res. 2018, 51, 170. DOI: 10.1021/acs.accounts.7b00444. 13 Tönnemann, J.; Risse, J.; Grote, Z.; Scopelliti, R.; Severin, K. Efficient and Rapid Synthesis of Chlorido‐ Bridged Half‐Sandwich Complexes of Ruthenium, Rhodium, and Iridium by Microwave Heating. Eur. J. Inorg. Chem. 2013, 2013, 4558. DOI: 10.1002/ejic.201300600. 14 For other examples of silver derivatives (AgPf or [AgPf2]‒) as transmetalating agents to obtain rhodium compounds see: García, M. P.; Jiménez, M. V.; Lahoz, F. J.; Oro, L. A. Synthesis and Reactivity of Mononuclear Anionic Pentafluorophenyl Compounds of Rhodium(I) and Iridium(I). X-ray Structure of [{P(OPh)3}2(C6F5)2RhAg(PPh3)] Inorg. Chem. 1995, 34, 2153. DOI: 10.1021/ic00112a030. 21 CHAPTER II RhCp*: Hidden exchanges and buffer effect Figure 1. X-Ray structures of the 5-coordinate 4 (left) and its piano-stool precursor 3 (right). These 5-coordinate derivatives 2 and 4 can be seen as models of the active intermediates formed after ligand dissociation from 18e compounds. In fact, the haloaryl scrambling reaction giving rise to the aryl-mixed species [RhCp*RfPf] (5) is observed at room temperature (Scheme 1). The apparently obvious mechanistic proposal for aryl transmetalation, via a double aryl bridged Transition State (TS), whether syn or anti (Scheme 1), was discarded due to the high computed activation energies that this pathway requires, close to 40 kcal mol‒1. Scheme 1. Aryl scrambling equilibrium between [RhCp*Ar2] complexes. TSs initially proposed for the direct transmetalation. Since the direct transmetalation is not feasible, the aryl exchange observed needs the presence of a highly active species acting as catalyst. With careful experimental reexamination we managed to detect minute amounts (< 0.5 mol%) of hydrolysis products (µ-OH)2[RhCp*Ar] with bridging hydroxyl groups. Luckily, we could confirm that these species are able to catalyse the aryl scrambling being studied, by acting as precursors of unobservable concentrations of 16e [RhCp*Ar(OH)] monomers. Indeed, these 16e species open the door to formation of OH-bridged TSs with accessible activation energies (Figure 2). 28 CHAPTER II RhCp*: Hidden exchanges and buffer effect orbital overlapping. For example, the donation coming from the Cl– ligand in the [MCp*RfCl(CO)] complexes, is noticeably high. The analysis of several data confirmed that the typical Rh→CO backdonation represents about 80% of the total, and the other lateral donations can reach a not negligible 20%. Until now, we have analysed the Cp* moiety as a group, both for the discussion of the M‒Cp*centroid distances and for the comments on lateral donation. It is well known that its preferred coordination mode is η5 and also that it experiments easy rotation in solution. However, examination of the frozen structures, whether from X-Ray diffraction or in DFT optimizations, confirms inevitably a clear loss of symmetry in the Cp* five-membered ring, whether in the C‒C or in the Rh‒C distances. Two well understood situations of this asymmetry are represented in Figures 6 21 and 7. 22 The differences are not random and can be perfectly rationalized with the same idea of the buffer effect of Cp*, presented previously: the high polarizability of the  electron density of Cp* allows for compensation of the changes induced by the different donor abilities (trans influence) of the rest of the ligands bonded to the metal centre, and this compensation has trans directionality. In other words, the Cp* is tilted as can be observed attending to the clearly different Rh‒C distances, and the distances to Rh are smaller for the transoid position(s) to the ligands with less trans influence. This means that a compensation of electron-density has taken place, which is reflected in the C‒C distances with higher contribution of the resonance form highlighted in each case in the figures. Obviously, this tilting can be conditioned by steric hindrance issues but, in the cases selected (DFT optimized geometries without bulky groups) this steric effect is discarded. Similar tendencies are observed in the corresponding X-Ray structures. The concept of transoid influence should be applicable to other complexes. 21 For ene-allyl distortion in [RhICp*(CO)2] see: Lichtenberger, D. L.; Blevins, C. H.; Ortega, R. B. Distortions in coordinated cyclopentadienyl rings: crystal, molecular, and electronic structural analysis of (η5-pentamethylcyclopentadienyl)dicarbonylrhodium. Organometallics 1984, 3, 1614. DOI: 10.1021/om00089a003. 22 For diene distortion in [Pd(η5-Cp)(PR3)( η1-Cp)] see: Werner, H.; Kraus, H.-J.; Schubert, U.; Ackermann, K.; Hofmann, P. Strukturdynamische organometall-komplexe: III. Synthese, struktur und bindungsverhältnisse der komplexe (η5-C5H5)Pd(η1-C5H5)PR3. J. Organomet. Chem. 1983, 250, 517. DOI: 10.1016/0022-328X(83)85075-X 29 CHAPTER II RhCp*: Hidden exchanges and buffer effect Figure 6. Zenith view of the DFT optimized structure of the cation [RhCp*Rf(NCMe)2]+ with Cp* C‒C and Rh‒C distances in Å. Figure 7. Zenith view of the DFT optimized structure of the complex [RhCp*RfCl(NCXylyl)] with Cp* C‒C and Rh‒C distances in Å. Finally, the indenyl effect 23 is well known to allow for associative mechanisms through metal slippage (η5 to η3), thus avoiding high energy 20e transition states. This is equivalent to converting the indenyl metal centre in a pentacoordinated one. The remarkable capability of Cp* to minimize the electronic changes around the metal centre (buffer effect) facilitates ligand dissociation from 18e compounds to form the 16e fivecoordinate intermediates needed to produce the active species required for reactivity in most cases (Scheme 5). As a numeric example, the dissociation of the strong isocyanide ligand in [RhCp*Rf2(CNXylyl)] was computed to cost only 10.8 kcal mol‒1, while for the Cp analogue the computed value obtained is 17.5 kcal mol‒1. An appreciable effect that should be worth taking into account when selecting a system for catalytic processes, 23 Calhorda, M. J.; Romão, C. C.; Veiros, L. F. The Nature of the Indenyl Effect. Chem. ‒ Eur. J. 2002, 8, 868. DOI: 10.1002/1521-3765(20020215)8:4<868::AID-CHEM868>3.0.CO;2-I 30 CHAPTER II RhCp*: Hidden exchanges and buffer effect because that difference might be determining to quench or facilitate the subsequent reactivity. Scheme 5. Structural conversions facilitating ligand association (left: indenyl effect) or dissociation (right: buffer effect of Cp*). Two important lessons are derived from the results presented in this chapter: first, the importance to confirm even apparently obvious mechanistic proposals, because presumptions can be wrong; combination of experiments and calculations is the best way to be safe; and second, intelligent analysis of X-ray structures (or geometry optimizations) can yield much more than the routine use made of them, and help to unveil interesting electronic effects with potential consequences in the energies involved in catalyses with these complexes. 31 CHAPTER III Reactivity of PdCl2 PEWO-chel Complexes Chapter III: Reactivity of Fluorinated-Chalcone Phosphines Induced upon Coordination to PdCl2. E/Z Isomerization Triggers C‒F Activation Phosphines with tethered electron withdrawing olefins (PEWO) have proven to be promising ligands for palladium catalysed cross coupling processes. 24 Particularly, our group has developed a family of PEWO where the olefin is a fluorinated chalcone (Figure 1). 25 These ligands have been successfully used in Negishi catalysis and, more interestingly, they are able to promote the challenging homocoupling of C6F5 from Pd(C6F5)2 fragments as fast as the Buchwald-type tBuXPhos ligand or the extremely bulky PtBu3. 26 The EWO moiety, when coordinated, diminishes dramatically the barrier for the reductive elimination step of the PdII complexes. 27 However, the Pd0 species subsequently formed are difficult to re-oxidize. 28 For this reason, we decided to synthesize new PEWO ligands modifying either the EWO (partially fluorinated) or the PR2 moiety (R = Cy), searching for a more convenient coupling/re-oxidation trade off. The two new ligands PhPEWO-H2F2 and CyPEWO-F4 are able to induce the Pf‒Pf (Pf = C6F5) coupling from cis-[PdPf2(THF)2]. The partially fluorinated ligand behaves in an intermediate way between the more efficient PhPEWO-F4, previously reported, and the 24 See for example: (a) Luo, X.; Zhang, H.; Duan, H.; Liu, Q.; Zhu, L.; Zhang, T.; Lei, A. Superior Effect of a π-Acceptor Ligand (Phosphine-Electron-Deficient Olefin Ligand) in the Negishi Coupling Involving Alkylzinc Reagents. Org. Lett. 2007, 9, 4571. DOI: 10.1021/ol701995t. (b) Shi, W.; Luo, Y.; Luo, X.; Chao, L.; Zhang, H.; Wang, J.; Lei, A. Investigation of an Efficient Palladium-Catalyzed C(sp)-C(sp) Cross-Coupling Reaction Using Phosphine-Olefin Ligand: Application and Mechanistic Aspects. J. Am. Chem. Soc. 2008, 130, 14713. DOI: 10.1021/ja8049436. 25 (a) Gioria, E.; Martínez-Ilarduya, J. M.; García-Cuadrado, D.; Miguel, J. A.; Genov, M.; Espinet, P. Phosphines with Tethered Electron-Withdrawing Olefins as Ligands for Efficient Pd-Catalyzed Aryl-Alkyl Coupling. Organometallics 2013, 32, 4255. DOI: 10.1021/om4004303. (b) Gioria, E.; Martínez-Ilarduya, J. M.; Espinet, P. Experimental Study of the Mechanism of the Palladium-Catalyzed Aryl−Alkyl Negishi Coupling Using Hybrid Phosphine−Electron-Withdrawing Olefin Ligands. Organometallics 2014, 33, 4394. DOI: 10.1021/om5005379. 26 Gioria, E.; del Pozo, J.; Martínez-Ilarduya, J. M.; Espinet, P. Promoting Difficult Carbon-Carbon Couplings: Which Ligand Does Best? Angew. Chem., Int. Ed. 2016, 55, 13276. DOI: 10.1002/anie.201607089. 27 Prez-Rodríguez, M.; Braga, A. A. C.; García-Melchor, M.; Prez-Temprano, M. H.; Casares, J. A.; Ujaque, G.; de Lera, A. R.; lvarez, R.; Maseras, F.; Espinet, P. C‒C Reductive Elimination in Palladium Complexes, and the Role of Coupling Additives. A DFT Study Supported by Experiment. J. Am. Chem. Soc. 2009, 131, 3650. 28 Tuxworth, L. W.; Baiget, L.; Phanopoulos, A.; Metters, O. J.; Batsanov, A. S.; Fox, M. A.; Howard, J. A. K.; Dyer, P. W. Phosphine−alkene ligand-mediated alkyl−alkyl and alkyl−halide elimination processes from palladium(II). Chem. Commun. 2012, 48, 10413. DOI: 10.1039/C2CC35623F. 32 CHAPTER III Reactivity of PdCl2 PEWO-chel Complexes non-fluorinated analogue developed by Lei.24 Moreover, the CyPEWO-F4 is much faster but so less controlled in reactivity, suffering not only the coupling reaction but also other competitive processes. Figure 1. Fluorinated Phosphine-EWO hybrid ligands with the chalcone skeleton. For the coupling rate measurements, a 2:1 L:Pd ratio is required to avoid ligand sequestration by the Pd0 species. Differently, in catalytic conditions a 1:1 ratio is advisable, and the best way to achieve that is using [PdCl2(PEWO-chel)] as precatalyst. Surprisingly, for PhPEWO-F4 in the chelate palladium complex, the configuration of the olefin is Z, despite the fact that the free ligand displays E configuration whether in solution, solid state or coordinated as P-monodentate ligand.25a We tried to gain deeper insight into this olefin isomerization, and how is it feasible even at room temperature. Scheme 1 collects the sequence of intermediates formed in the synthesis of complexes [PdCl2{Z-(RPEWO-F4-chel}] (Z-3) starting from trans-[PdCl2(NCMe)2] (2) and one equivalent of ligand E-RPEWO-F4 (1; R = Ph – a –; o-tol – b –; Cy – c –). For the partially fluorinated ligand PhPEWO-H2F2, olefin dissociation equilibria are observed, supporting that the ligand with more fluorine groups has the stronger olefin-Pd interaction. Scheme 1. Synthesis of complexes [PdCl2(RPEWO-F4-chel)] 33 CHAPTER III Reactivity of PdCl2 PEWO-chel Complexes Fortunately, the chelate intermediate E-3b and its Z isomer could be isolated and fully characterised (see Figure 2 for their X-ray structures). Figure 2. X-ray structures of E-3b (left) and Z-3b (right). Either for the free PEWO ligands or the P-monodentate complexes, the E isomer is very thermodynamically favored. However, the chelate coordination to PdCl2 reduces this thermodynamic difference and even makes the Z-chelate species more stable. In other words, PdCl2 acts as a trap for this unusual configuration of the double bond. The activation barrier for isomerization from the chelate complexes must be higher than from the monodentate complexes because it requires to pay first for decoordination. Complex 3b is the only case in which the two isomers are isolable, and the availability of E-3b allowed us to carry out a kinetic study of the E/Z isomerization for the o-Tol ligand. The isomerization barriers are about 23 kcal mol‒1 in non-coordinating solvents. Overall, these data show that the isomerization barrier is small, compared to the high activation energies for E/Z isomerization in common olefins. The process is catalysed by the addition of NCMe, supporting an isomerization mechanism that requires olefin dissociation to form a monodentate intermediate. In fact, chalcone-type olefins have been reported to easily isomerize when electronwithdrawing groups are present in their skeleton due to the substantial contribution of resonance forms that reduces the activation barrier for the olefin rotation (Scheme 2). 29 , 30 29 Roque, A.; Lima, J. C.; Parola, A. J.; Pina, F. Substitution and solvent effects in the chalcones isomerization barrier of flavylium photochromic systems. Photochem. Photobiol. Sci. 2007, 6, 381. 10.1039/B612612J. 30 In general, the barrier to rotation in regular olefins is very high, and different catalytic mechanisms have been proposed. See for example: Tan, E. H. P.; Lloyd-Jones, G. C.; Harvey, J. N.; Lennox, A. J. J.; Mills, B. M. [(RCN)2PdCl2]-Catalyzed E/Z Isomerization of Alkenes: A Non-Hydride Binuclear Addition−Elimination Pathway. Angew. Chem., Int. Ed. 2011, 50, 9602. DOI: 10.1002/anie.201103947. 34 CHAPTER III Reactivity of PdCl2 PEWO-chel Complexes Scheme 2. Resonance forms responsible for the feasible E/Z isomerization and the strong electron withdrawing behaviour of these PEWO ligands when coordinated to a metal centre. For free chalcones the Z-configuration can only be reached by irradiation and it is known to trigger reversible cyclization reactions with interest in optical devices. 31 Similarly, the Z-chelate complexes [PdCl2{Z-(RPEWO-F4)] (Z-3) with R = Ph, o-Tol, Cy suffer cyclization reactions with the loss of HF leading to chelate [PdCl2(P-carbene)] (4) compounds (Figure 3). It is worth noting that their reaction rates depend strongly on the R group: they are fast at room temperature for Cy, and require higher temperatures for Ph and o-Tol. Figure 3. Cyclization reaction occurring from the Z-configuration of the PEWO-F4 chelate complexes (left). X-Ray structure of the P-carbene complex [PdCl2(PhP-carbene)] (4a) obtained. The results of the NBO calculations have allowed us to state that an effective aromatic delocalization of π electron density stabilizes these P-carbene derivatives. We also predict that C3 and C5 (Figure 3) must be electrophilic. The electrophilicity of these carbon atoms is determinant for the hydrolysis and ammonolysis reactions observed subsequently. These are totally chemoselective in the nucleophilic attack to the C5 atom, forming PCO and PCN pincer complexes. On the other hand, a formal O insertion takes place in the C3‒Pd bond, promoted with peroxides present in aged THF, 32 leading to complexes with unreported hemilabile PO ligands (Scheme 3). 31 Pina, F.; Melo, M. J.; Maestri, M.; Ballardini, R.; Balzani, V. Photochromism of 4‘-Methoxyflavylium Perchlorate. A “Write-Lock-Read-Unlock-Erase” Molecular Switching System. J. Am. Chem. Soc. 1997, 119, 5556. DOI: 10.1021/ja9704646. 32 For other oxidative processes promoted by these peroxides see: Zhang, B.; Cho, M.; Fortner, J. D.; Lee, J.; Huang, C.-H.; Hughes, J. B.; Kim, J.-H. Delineating Oxidative Processes of Aqueous C60 Preparations: Role of THF Peroxide. Environ. Sci. Technol. 2009, 43, 108. DOI: 10.1021/es8019066. 35 CHAPTER III Reactivity of PdCl2 PEWO-chel Complexes We have also made reasonable mechanistic proposals for the formation of the new complexes, which have been unambiguously characterised by X-Ray diffraction studies and other spectroscopic techniques. Scheme 3. Synthesis of PCE pincer and PO chelate complexes from [PdCl2(RP-carbene)]. In the conclusion of this chapter, it is worth remarking that the reactivity reported is initiated by coordination of the olefin moiety to PdCl2 of our chalcone-type PEWO-F4 ligands. The coordination favours the E/Z isomerization, which triggers the C‒F activation process that follows. Neither the isomerization nor the subsequent reactivity is observed without the ligand in a chelate mode. These processes should be taken as a warning for reactions using fluorinated PEWO ligands, because the reactivity observed can interfere in the results when exploring new catalytic processes, if these are not fast enough or if harsh conditions are required. At least for the Negishi catalyses reported so far by the group,25 using strong nucleophiles and room temperature, this has not happened. 36 CHAPTER IV Substoichiometric Protection of AuI Catalysts Chapter IV: Protection of Gold(I) Catalysts by Substoichiometric Agents. Is the Decomposition to Metallic Gold a Simple Reduction? Gold(I) has become a widely used catalytic metal. Its ability to promote cyclizations reactions over organic substrates is particularly prolific. 33 Interesting bimetallic catalytic processes involving AuI have been also recently reported by our research group. 34 The formation of gold(0) nanoparticles during catalysis using naked [AuL]+ species is very frequent, 35 but is usually not mentioned because the extremely fast kinetics lead to high conversions anyway with low loads of catalyst. However, this decomposition produces an important decay of the catalytic activity and also prevents the recyclability of the catalyst. Moreover, the formation of gold nanoparticles lets open the question of whether they participate in the catalysis. Au0 is apparently formed in the absence of obvious reducing agents. Interestingly, some works, not focussed in catalysis, reported disproportionation processes of 3AuI into 2Au0 + AuIII, as synthetic protocol to obtain both nanoparticles 36 or gold(III) derivatives. 37 Previous to these articles, Jones et at were able to demonstrate that [AuXL] complexes (L = picoline) disproportionate in part into gold(III) and gold(0) when 50% of the halide is extracted with a silver salt (equation 1). 38 Curiously, these results had been unnoticed in the area of gold catalysis for many years until our work. 33 (a) Dorel, R.; Echavarren, A. M. Gold(I)-Catalyzed Activation of Alkynes for the Construction of Molecular Complexity. Chem. Rev. 2015, 115, 9028. DOI: 10.1021/cr500691k (b) Modern Gold Catalyzed Synthesis; Hashmi, A. S. K.; Toste, F. D.; Eds.; Wiley-VCH: Weinheim, Germany, 2012. 34 delPozo, J.; Carrasco, D.; Pérez-Temprano, M. H.; García-Melchor, M.; Álvarez, R.; Casares, J. A.; Espinet, E. Stille Coupling Involving Bulky Groups Feasible with Gold Cocatalyst. Angew. Chem. Int. Ed. 2013, 52, 2189. DOI: 10.1002/anie.201209262. 35 Naked or monocoordinated [AuL]+ are commonly proposed as the catalytically active species formed after halide extraction over [AuXL] complexes with a silver salt. 36 Bergamini, G.; Ceroni, P.; Balzani, V.; Gingras, M.; Raimundo, J.-M.; Morandi, V.; Merli, P. G. Synthesis of small gold nanoparticles: Au(I) disproportionation catalyzed by a persulfurated coronene dendrimer. Chem. Commun. 2007, 4167. DOI: 10.1039/B708115D. 37 See for example: Rana, B. K.; Nandy, A.; Bertolasi, V.; Bielawski, C. W.; Saha, K. D.; Dinda, J. Novel Gold(I)− and Gold(III)−N-Heterocyclic Carbene Complexes: Synthesis and Evaluation of Their Anticancer Properties. Organometallics 2014, 33, 2544. DOI: 10.1021/om500118x. 38 Jones, P. G.; Ahrens, B. Gold(I) Complexes with Amine Ligands, II. Methylpyridine Complexes of Gold(I). Z. Naturforsch., B: J. Chem. Sci. 1998, 53, 653. DOI: 10.1515/znb-1998-0702. 37 CHAPTER IV Substoichiometric Protection of AuI Catalysts As far as we know, before the start of this thesis, only one work published by the group of Hammond faced the question of the decomposition of gold complexes under catalytic conditions. The authors proposed that unsaturated organic substrates might induce disproportionation and they also managed to detect AuI, Au0 and AuIII by XPS spectra from solutions of [Au(OTf)(PPh3)] with cyclohexene in chloroform. 39 In this context, we decided to study the decomposition of our [AuCl(carbene)] catalysts (carbene = NAC = Nitrogen Acyclic Carbene) successfully used in cyclization of 1,6enynes (equation 2). 40 These complexes are very active when Cl is extracted with Ag(SbF6), but formation of gold nanoparticles was clearly observed during the catalyses. The cationic [Au(AsPh3)(NAC)]+ was also active and also much slower. To our surprise we soon discovered that a stoichiometric amount of AsPh3 was not necessary to protect the catalyst from decomposition. Indeed, just a very substoichiometric amount of AsPh3 (10 mol%) was totally efficient in the protection of the catalyst, 41 quenching the decomposition while only slightly decelerating the catalysis and allowing for higher turnover numbers. As a picture is worth a thousand words, the difference between two catalytic reactions (with and without substoichiometric protection with AsPh3, after 30 minutes) is shown in Figure 1. Obviously we had to focus the work to understanding this unexpected case. And for this it was also necessary to elucidate whether the formation of Au0 nanoparticles in the catalytic processes was a disproportionation or a reduction process. 39 Kumar, M.; Jasinki, J.; Hammond, G. B.; Xu, B. Alkyne/Alkene/Allene‐Induced Disproportionation of Cationic Gold(I) Catalyst. Chem. ‒ Eur. J. 2014, 20, 3113. DOI: 10.1002/chem.201304271. 40 (a) Bartolom, C.; Ramiro, Z.; García-Cuadrado, D.; Prez-Galn, P.; Bour, C.; Raducan, M.; Echavarren, A. M.; Espinet, P. Nitrogen Acyclic Gold(I) Carbenes: Excellent and Easily Accessible Catalysts in Reactions of 1,6-Enynes. Organometallics 2010, 29, 951. DOI: 10.1021/om901026m. (b) Bartolom, C.; García-Cuadrado, D.; Ramiro, Z.; Espinet, P. Exploring the Scope of Nitrogen Acyclic Carbenes (NACs) in Gold-Catalyzed Reactions. Organometallics 2010, 29, 3589. DOI: 10.1021/om100507r. 41 Ramiro, Z.; Bartolom, C.; Espinet, P. Protection of the Gold(I) Catalyst by AsPh3 in Reactions of Enynes. Eur. J. Inorg. Chem. 2014, 2014, 5499. DOI: 10.1002/ejic.201402744. 44 CHAPTER V d8···d10 and d10···d10 Interactions It is worth noting that, in spite of not being supported by intra-unit π-π stacking, d8···d10 interactions totally prevail over their homometallic combinations. 46 We find Au···Rh distances in the range 3.0–3.5 Å, well below the sum of Rh + Au van der Waals radii (4.1–4.7 Å) but also above the sum of covalent radii (2.75 Å). Their competitive formation suggests similar overall stability for the three polymorphs. At least for the shortest distances, around 3.1 Å, significant orbital overlapping and some contribution to the stability of the crystal can be presumed. However, the main stabilizer factor for the structures is the presence of inter-unit π-interactions networks that are impossible to show in Figure 1. Despite the fact that Rh···Au interactions in our case are not determining the structure, but determined by the structure, they are the main responsible for the crystal colour. This can be qualitatively understood taking a look to the frontier molecular orbitals (MO) obtained by DFT calculations. The infinite arrangement of 2c escapes the simple analysis, but representative fragments were selected for 2a (Rh···Au dimer) and 2b (Rh···Au···Rh trimer). The X-Ray structures must be maintained and for this reason geometry optimizations were discarded (the fragments are not stable in the gas phase). Figure 2 shows the orbital diagram for 2a. Figure 2. Relevant MOs for the selected fragment of 2a, and HOMOs of the separated ions. 46 Similar cases have been reported for PtII···AuI. See for example: Stork, J. R.; Rios, D.; Pham, D.; Bicocca, V.; Olmstead, M. M.; Balch, A. L. Metal−Metal Interactions in Platinum(II)/Gold(I) or Platinum(II)/Silver(I) Salts Containing Planar Cations and Linear Anions. Inorg. Chem. 2005, 44, 3466. DOI: 10.1021/ic048333a. 45 CHAPTER V d8···d10 and d10···d10 Interactions The combination of the HOMOs from Rh (4dz2) and Au (5dz2 + 6s) leads to the bonding σ-HOMO-3 and the antibonding σ*-HOMO, being the later higher in energy and the highest occupied molecular orbital in the fragment. The LUMO has no significant contribution from gold and involves mainly π density from the isocyanide ligands.43,45 Obviously, the LUMO must be much lower in energy in the real crystal because the stacking interactions are missing in the computational model. The energy gap HOMO-LUMO determines the colour of the crystal (which is complementary to the absorption in UV-vis spectra). Comparing 2a (orange) and 2b (blue), the first must have a smaller gap. The calculations performed with the selected fragments are able to reproduce qualitatively this fact, although they differ in the absolute values due to the oversimplification of the model used for the DFT calculations. The decrease in the energy gap for 2b is the consequence of the higher destabilization of the σ*-HOMO that is the multiply antibonding combination (of at least 3 metal centres if not five) while the LUMO remains relatively unaltered compared to 2a (Figure 3). This destabilization of the HOMO has no direct influence in the stability of the crystalline structure because it is concomitantly linked with the stabilization of the bonding combination, but it does determine the colour of the crystal. Figure 3. Frontier MOs of a symmetric fragment of 2b. σ*-HOMO (left) and LUMO (right). Coming now to the gold(I) isocyanide derivatives and their silver bimetallic analogues, we decided to explore the use of 4-pyridylisocyanide, CNPy, to make selectively bimetallic Au/Ag complexes. Initially we obtained the homometallic complexes [AuAr(CNPy)] with Pf (3) and Rf (4) groups and the heterometallic [{AuAr(CNPy)}2Ag](BF4) (5 and 6 respectively), with bridging CNPy C-coordinated to Au and N-coordinated to Ag. Neither the free ligand nor the complexes display 46 CHAPTER V d8···d10 and d10···d10 Interactions luminescence in solution. However, the complexes are luminescent in the solid state, and 3 and 4 display marked mechanochromic behaviour (see Figure 4). 47 , 48 Figure 4. Emission spectra of compounds 3-6 before and after grinding. Photographs taken under irradiation. In order to rationalize the modifications in the emissions observed in Figure 4, we tried to obtain extra information about the solid structures of these species. Figure 5 gathers the X-Ray molecular structures of 3, 4 and 6. In 3, long Au···Au distances (3.8 Å) are observed and π-stacking seems to direct the packing. The replacement of Pf by Rf in 4 induces a completely different structure with a variety of weak interactions present, halogen···halogen and Au···Au (intermediate distance, 3.3 Å) interactions are added to the ubiquitous stacking. Finally, bimetallic [{AuRf(CNPy)}2Ag](BF4) (6) displays very short aurophilic distances (3.1 Å) and direct participation of the silver centre in any metallic interactions is not observed. 49 The structure of 6 clearly shows that the reaction with AgBF4 abruptly modifies the packing of the product due to the steric requirements of the anion. 47 For reviews see: (a) Sagara, Y.; Yamane, S.; Mitani, M.; Weder, C.; Kato, T. Mechanoresponsive Luminescent Molecular Assemblies: An Emerging Class of Materials. Adv. Mater. 2016, 28, 1073. DOI: 10.1002/adma.201502589. (b) Xue, P.; Ding, J.; Wang, P.; Lu, R. Recent progress in the mechanochromism of phosphorescent organic molecules and metal complexes. J. Mater. Chem. C, 2016, 4, 6688. DOI: 10.1039/C6TC01503D. 48 For one of the first examples of AuAr(CNAr’) mechanochromic species see: Ito, H.; Saito, T.; Oshima, N.; Kitamura, N.; Ishizaka, S.; Hinatsu, Y.; Wakeshima, M.; Kato, M.; Tsuge, K.; Sawamura, M. Reversible mechanochromic luminescence of (C6F5Au)2(μ-1,4-diisocyanobenzene). J. Am. Chem. Soc. 2008, 130, 10044. DOI: 10.1021/ja8019356. 49 For AuI/AgI bimetallic compounds with direct Au···Ag interactions see: Lasanta, T.; Olmos, M. E.; Laguna, A.; Lopez-de-Luzuriaga, J. M.; Naumov, P. Making the Golden Connection: Reversible Mechanochemical and Vapochemical Switching of Luminescence from Bimetallic Gold-Silver Clusters Associated through Aurophilic Interactions. J. Am. Chem. Soc. 2011, 133, 16358. DOI: 10.1021/ja206845s. 47 CHAPTER V d8···d10 and d10···d10 Interactions We should not forget the electronic implications of silver that acts as a linker of two AuRf(CNPy) moieties in a trimetallic unit with no dipolar moment. 50 Figure 5. X-Ray structures of 3, 4 and 6 with Au···Au distances in Å. Dashed lines represent examples of other weak interactions. The red-shifts in the emission moving from 3 to 4 or from 4 to 6 concur with the shortening in Au···Au distances, and are in agreement with previous works, most of them reported by Ito. 51 Similarly, the mechanochromic behaviour of [AuAr(CNPy] complexes, is a consequence of a loss of crystallinity that leads to amorphous phases with shorter aurophilic interactions. We also tried to shed light into the orbitals involved in the emission, using in this case TD-DFT calculations. The same restrictions commented above for the Rh···Au system must be applied here in order to maintain the packing. We decided to study the green emitting [AuRf(CNPy)] (4) and the orange emitting [{AuRf(CNPy)}2Ag](BF4) (6). The later must have a smaller gap because its luminescence is red-shifted. A critical point is 50 The lack of dipolar moment might preclude the typical antiparallel arrangement: Seki, T.; Kobayashi, K.; Mashimo, T.; Ito, H. A gold isocyanide complex with a pendant carboxy group: orthogonal molecular arrangements and hypsochromically shifted luminescent mechanochromism. Chem. Commun. 2018, 54, 11136. DOI: 10.1039/C8CC06490C. 51 Seki, T.; Takamatsu, Y.; Ito, H. A Screening Approach for the Discovery of Mechanochromic Gold(I) Isocyanide Complexes with Crystal-to-Crystal Phase Transitions. J. Am. Chem. Soc. 2016, 138, 6252. DOI: 10.1021/jacs.6b02409. 48 CHAPTER V d8···d10 and d10···d10 Interactions the election of simplified fragments that properly represent the solid-structures, and in the case of 6, silver (and also BF4) can be spared because they have no significant contribution to the frontier orbitals. Figure 6 shows the frontier orbitals obtained for fragments of six [AuRf(CNPy)] molecules taken from the X-ray structures of 4 and 6, and their computed excitation energies. The calculations reflect qualitatively the red-shift observed and also show the main reason for this shift: While the LUMOs in both cases are constituted by π-density of isocyanide ligands, the HOMOs strikingly differ. The gold contribution to the HOMOs is clearly different in both structures: it is negligible in 4 (8%) because the HOMO is located in the Rf moieties, but predominant in 6 (79%). Analogous to the Rh···Au aggregates, the shorter the M···M’ distances, the higher the destabilization of the multiple antibonding combination denoted as σ*-HOMO. The consequence of this destabilization is a smaller energy gap and, in this case, a red-shifted emission. Figure 6. HOMO-LUMO orbitals and TD-DFT excitation energies for selected fragments of 4 (left) and 6 (right). 49 CHAPTER V d8···d10 and d10···d10 Interactions The most important lesson learned from the detailed study of the different structural data, together with the information obtained from the theoretical calculations presented in this Chapter, is that the crystalline packing is a delicate balance of weak interactions, being the metalophilicity just one more. This is not in contradiction with the fact that these energetically minor M···M’ interactions can determine some interesting properties of the material, within the context of an intimate structure-properties relationship. Geometryrestricted DFT studies (at low computational cost) performed over simplified models of the crystal can offer very valuable information, even if sacrificing quantitative accuracy. 50 Conclusions Conclusions Chapter I: RhI/AuI Transmetalation In Rf/Pf exchanges between RhI and AuI complexes the fractional and negative kinetic order in AsPh3 indicates that this transmetalation proceeds via a multistep mechanism, one involving ligand dissociation. In contrast, in the Rf/Cl exchange there is an accelerating effect of AsPh3, which reveals the existence of two cooperative pathways, ligand-catalysed and uncatalysed. DFT calculations reveal the remarkable tendency for both Rf/Pf and Rf/Cl exchanges to occur via oxidative insertion with formation of heterometallic bonded intermediates. Chapter II: RhCp*: Hidden exchanges and buffer effect The direct Rf/Pf exchange between 16e [RhCp*Ar2] species is unfeasible, but proceeds via Ar/OH exchanges catalysed by the hydrolysis product (µ-OH)2[RhCp*Ar]2, present in minute concentrations. The study reveals that these exchanges are efficient and must be operating also when the two aryls are identical. The synthesis of 18e [RhCp*ArCl]2 complexes by symmetrization of [RhCp*Ar2] with [RhCp*Cl2]2 is another example of Ar/X exchange, in this case Ar/Cl. The high stability of the five-coordinate 16e [RhCp*Ar2] complexes is due to the capability of the Cp* ligand to modulate its electronic donation when needed, diminishing the electronic variations around the metal centre. We call this “buffer effect of Cp*”. The structural analysis of different RhCp* complexes allows us to stablish a series of trans influence of ligands in an octahedral environment. NBO calculations support that in these RhCp* complexes the simple concept of backdonation from metal to π-acceptor ligands needs to be redefined. In fact, nonnegligible lateral donations from the Cp* to CO or CN‒ have to be taken in consideration. Chapter III: Reactivity of PdCl2 PEWO-chel Complexes PdCl2 is able to coordinate one fluorinated PEWO ligand as chelate and traps the infrequent Z configuration of the double bond. The chalcone skeleton of these ligands facilitates the easy E/Z isomerization observed. 51 Conclusions The E/Z isomerization triggers a C‒F activation process, leading to unreported P-carbene complexes. The proposed aromaticity of these derivatives allows us to rationalize the chemoselectivity of the subsequent reactivity. These processes must be taken into account as they are a potential complication in new reactions or in catalysis using PEWO ligands. Chapter IV: Substoichiometric Protection of AuI Catalysts The substoichiometric protection by AsPh3 of the commonly used [AuL(W)]+ catalysts (W = weakly coordinating ligand) strongly supports a disproportionation mechanism for their decomposition. Stronger ligands, such as PPh3, cannot play this role because a certain concentration of free ligand is required. Other silent ligands (superstoichiometric adventitious water is one example) can do the role when the catalyst in decomposition reaches low concentration. This explains the residual activity that remains after massive catalyst decomposition. Chapter V: d8···d10 and d10···d10 Interactions The use of the ortho-substituted xylylisocyanide ligand (L) hampers π-π stacking between [RhL4]+ units preventing d8···d8 interactions and this helps for the formation in the solid state of d8···d10 interactions with the bimetallic [RhL4][Au(CN)2] complex. These metallophilic interactions, although energetically modest, determine the colour of the crystals. The red shifts in the luminescence observed when modifying the aryl group, incorporating silver or grinding, both in homometallic [AuAr(CNPy)] and in bimetallic [{AuAr(CNPy)}2Ag]BF4 complexes, are intimately related with the formation of shorter aurophilic interactions in the solids. Geometry restricted DFT and TD-DFT calculations, using representative fragments of the crystalline structures, allow to identify the orbitals involved in these photophysical properties. 52 Conclusions General conclusions The different and diverse topics dealt with in this PhD work highlight the power of combining experimental and theoretical tools when facing chemical problems, not only in the field of mechanistic studies but also in the planning and understanding of other areas, such as the production of deeply coloured or luminescent materials in our case. On the other, the easy accessibility to collections of data, harder to get at the time when some chemical concepts were defined (trans influence, σ-donation + π-backdonation…), offers an opportunity to look at them in a non-routine way and get a deeper understanding of their significance and applicability. 53 Methods Methods The methodology used in this thesis is familiar to all the researchers in chemistry at this level of exigence. Herein, I will describe only the general methods while the particular details, including the characterisation of the new compounds, are collected in the Supplementary Information of the scientific articles. The reactions were performed under N2 atmosphere. Solvents were purified according to standard procedures. 1H, 13C{1H}, 19F and 31P{1H} NMR spectra were recorded on Bruker AV–400 or Varian 500/54 Premium Shielded instruments. In the spectra registered in non-deuterated solvents, a coaxial tube containing acetone-d6 was used to maintain the 2H lock signal. Infrared spectra were recorded with Perkin–Elmer Frontier (4000–200 cm–1) equipped with an ATR accessory (Attenuated Total Reflection) for the direct recording of solid samples. The elemental analyses were performed with a Carlo Erba 1108 microanalyser (by Vigo University, Spain). Kinetic experiments were monitored by 19F or 1H NMR. The reagents, the solvent and the internal reference (when needed) were added in an NMR tube and placed into a thermostated probe in a NMR apparatus. The temperature of the sample was determined using methanol as chemical shift thermometer. Concentration-time data were then acquired from the integrals of the selected signals. Linear fittings afforded initial rates. The kinetic models were fitted to the measured concentration vs. time experimental data by nonlinear least-squares (NLLS) regression, using the Program COPASI. For the X-ray diffraction studies, the crystals were attached to a glass fiber or a micromount and transferred to an Agilent Supernova diffractometer with an Atlas CCD area detector. Data collection was performed with Mo-Kα radiation ( = 0.71073 Å) or Cu-Kα (λ = 1.54184 Å). Data integration, scaling and empirical absorption correction was carried out using the CrysAlisPro program package. Using Olex2, the structure was solved with the ShelxT, and refined with ShelxL program. Density functional theory (DFT) calculations reported in this work were carried out using the dispersion corrected hybrid functional ωB97X-D. C, P, As and H atoms were described using the double-ζ basis set 6-31G(d,p), whereas the same basis set plus diffuse functions was employed to describe the more electronegative O, N, Cl and F atoms. Transition metals were described using the effective core potential LANL2DZ including f-polarization functions. Geometry optimizations in vacuum were performed without 60 List of publications Article V (Chapter III) E−Z Isomerization of Phosphine-Olefin (PEWO‑F4) Ligands Revealed upon PdCl2 Capture: Facts and Mechanism Peñas-Defrutos, M. N.; Vélez, A.; Gioria E.; Espinet, P. Organometallics 2019, 38, 4701. DOI: 10.1021/acs.organomet.9b00679. Abstract: The PEWO phosphines R2P(o-C6H4CH=CHC(O)Ph), R2P(oC6H2F2CH=CHC(O)Ph), and R2P(o-C6F4CH=CHC(O)Ph) and their P-monodentate complexes trans-[PdCl2(P-monodentate)2] show, in solution and (when available) in the X-ray diffraction structures, an E configuration of the double bond. In contrast, the structures of [PdCl2(P-chelate)] display E and Z configurations. The E/Z isomerization of the latter requires first decoordination of the double bond, which then allows for easy rotation about the electron-deficient double bond. Thus, the E/Z equilibria exist for the free and the P-monodentate complexes as well but are not observed because they are extremely displaced toward the E isomer. Their capture in the form of [PdCl2(P-chelate)], with equilibrium constants on the order Keq ≈ 1–3, allows the two configurations to be observed and isolated. Evaluation of their ability to couple Pf–Pf from cis-[PdPf2(THF)2] (Pf = C6F5) affords values of their ΔG‡(Pf–Pf)Pd parameters confirming that higher substitution of H by F produces lower coupling barriers and a double bond that is more electron deficient when it is free and more electron withdrawing when it is coordinated. 61 List of publications Article VI (Chapter III) Reactivity of Fluorinated-Chalcone Phosphines, RPEWO‑F4, Induced by C−F Activation upon Coordination to PdCl2 Peñas-Defrutos, M. N.; Vélez, A.; Espinet, P. Organometallics 2020, 39, 841. DOI: 10.1021/acs.organomet.0c00019. Abstract: The E phosphine ligands (R = Ph, o-Tol, Cy), abbreviated as RPEWO-F4, are stable in solution, but they develop a rich reactivity on coordination to PdCl2. The chelate P-olefin coordination to PdCl2 leads eventually to a Z conformation of the fluorinatedchalcone group o-C6F4CH=CHC(O)Ph. From there, a cyclization reaction occurs involving the C=O group and activation of a F atom, yielding a strongly chelated [PdCl2(P-carbene)] complex. The carbene carbon atom in the complex displays some electrophilicity, which is expressed in hydrolysis, ammonolysis, and oxidation (with peroxide) reactions, affording PdCl2 complexes with new P,C,O-pincer, P,C,N-pincer, or P,O-chelate fluorinated ligands. The C–F activation reactions are slow in comparison to the catalysis rates when the [PdCl2(RPEWO-F4)] complexes have been used in Negishi catalyses. Consequently, the reactivity discussed here is not expected to interfere with the interpretation of the data obtained in Pd-catalytic studies or processes, at least for fast transmetalating nucleophiles. 62 List of publications Article VII (Chapter IV) Some Singular Features of Gold Catalysis: Protection of Gold(I) Catalysts by Substoichiometric Agents and Associated Phenomena Bartolomé, C.; Ramiro, Z.; Peñas-Defrutos, M. N.; Espinet, P. ACS Catal. 2016, 6, 6537. DOI: 10.1021/acscatal.6b01825. Abstract: This study deals with two striking phenomena: the complete protection against decomposition of hypothetically monocoordinated AuI intermediates [AuL]Y (L = strongly coordinating ligand; Y– = poorly coordinating anion) by addition of small substoichiometric amounts (5 mol % relative to Au) of not strongly coordinating ligands (e.g., AsPh3) and the fact that, in contrast, strongly coordinating ligands cannot provide this substoichiometric protection. The two phenomena are explained considering that (i) the existence of real monocoordinated [AuL]Y is negligible in condensed phases and the kinetically efficient existing species are dicoordinated [AuL(W)]Y (W = any very weakly coordinating ligand existing in solution, including OH2, the solvent, or the Y– anion) and (ii) these [AuL(W)]Y intermediates give rise to decomposition by a disproportionation mechanism, via polynuclear intermediates formed by associative oligomerization with release of some W ligands. It is also shown that very small concentrations of [AuL(W)]Y are still catalytically efficient and can be stabilized by overstoichiometric adventitious water, so that full decomposition of the catalyst is hardly reached, although eventually the stabilized concentration can be kinetically inefficient for the catalysis. These results suggest that, in cases of gold catalysis requiring the use of a significant quantity of gold catalyst, the turnover numbers can be increased or the concentration of gold catalyst widely reduced, using substoichiometric protection properly tuned to the case. 63 List of publications Article VIII (Chapter V) d8···d10 RhI···AuI interactions in Rh 2,6-xylylisocyanide complexes with [Au(CN)2]‒: bond analysis and crystal effects Conejo-Rodríguez, V.; Peñas-Defrutos, M. N.; Espinet, P. Chem. Commun. 2019, 55, 5005. DOI: 10.1039/c9cc01377f. Abstract: The well-known [RhL4]n(anion)n structures, with RhI⋯RhI d8⋯d8 interactions, are replaced by others with RhI⋯AuI d8⋯d10 interactions such as [{RhL4}{Au(CN)2}] (L = 2,6-xylylisocyanide) or [{RhL4}{Au(CN)2}{RhL4}{Au2(CN)3}·4(CHCl3)]∞ when the anion is [Au(CN)2]−. Orbital (Rh⋯Au), coulombic, and inter-unit π–π aryl stacking interactions stabilize these crystal structures. 64 List of publications Article IX (Chapter V) 4-Pyridylisocyanide gold(I) and gold(I)-plus-silver(I) luminescent and mechanochromic materials: the silver role Conejo-Rodríguez, V.; Peñas-Defrutos, M. N.; Espinet, P. Dalton Trans. 2019, 48, 10412. DOI: 10.1039/c9dt01618j. Abstract: Crystallographic and DFT examination of the metalloligands [AuAr(CNPy)] (Ar = C6F5 (1), C6F3Cl2-3,5 (2)) and their silver complexes [Ag[AuAr(CNPy)]2](BF4) (3 and 4) support that the marked luminescence red-shifts observed on moving from 1 to 2, from 1,2 to 3,4, or upon grinding, are not caused by electronic differences (either by changing the aryls C6F5/C6F3Cl2, or by N coordination to silver), nor by non-existent Au···Ag interactions. They are always due to structural changes disturbing stronger π–π stackings in order to allow for shorter Au···Au interactions. COMMUNICATION 1 Accelerating Effect of AsPh3 in the RhI/AuI Transmetalation. Reversible Tricoordination of AuI Eases its Oxidative Insertion in the Rh‒C Bond Marconi N. Peñas-Defrutos,[a] Camino Bartolomé,* [a] Max García-Melchor,*[b] and Pablo Espinet*[a] Abstract: The Ar/Cl (Ar = C6Cl2F3-3,5) exchange between [AuClL] (L = AsPh3) and trans-[RhAr(CO)L2] is accelerated by addition of an excess of L. By combining experimental data, DFT simulations and microkinetic modeling, we demonstrate the existence of two competitive pathways. The addition of L has a catalytic effect and opens a mechanism involving the coordination of the additional L giving a tricoordinate [AuClL2] species, which transmetalates faster compared to the linear analogue [AuClL]. Instead of involving double-bridged transition states, both mechanisms involve the initial oxidative insertion of Au into the Rh‒C bond to afford intermediates with strong metal-metal bonds. Bimetallic catalysis 1 refers to homogeneous processes in which two transition metals (TM), or one TM and one group-11 element (M), 2 cooperate in a synthetic transformation (often C−C coupling) and their actions are linked by a transmetalation step. The understanding of transmetalation mechanisms in this recent area may help in the design of efficient catalytic systems (see for instance the gold assisted Stille type couplings for bulky aryls). 3 , 4 The cooperativity of PdII/AuI and PdII/CuI pairs (e.g. the Sonigashira reaction) has been reasonably studied in the past. The lack of mechanistic studies on the promising RhI/AuI dyad prompted us to investigate their transmetalation possibilities, and very recently we reported the aryl scrambling between the Vaska-type complex trans-[RhRf(CO)(AsPh3)2] (Rf = C6F3Cl23,5) and [AuPf(AsPh3)] (Pf = C6F5). 5 We discovered that, instead of the classical transmetalation mechanism featuring a double aryl bridged transition state, the aryl exchange occurred unexpectedly via oxidative insertion of rhodium into the Au‒C bond (Figure 1). 6 The fluorinated-aryls facilitate 19F NMR monitoring, providing valuable kinetic information to elucidate the transmetalation mechanism. 7 In that study, the unconventional fractional and negative kinetic dependence upon free ligand addition (AsPh3 slows down that reaction) revealed an entangled mechanism. Only combining kinetics, density functional theory (DFT) calculations, and microkinetic modelling, could the complete pathway be established as a reversible process involving oxidative insertion, ligand dissociation and isomerization steps, laying very close in energy. 8 Figure 1. A) Reversible aryl exchange between trans-[RhRf(CO)(AsPh3)2] and [AuPf(AsPh3)]. B) Ligand effect on the initial transmetalation rate. C) Transition state leading to Rh insertion into the Au‒CPf bond. An octahedral intermediate with a Rh‒Au bond is formed. Ph groups in the arsine are omitted for clarity. The most frequently desired exchange in catalysis is R for X (R = organic group; X = halide), in order to bring the R group to the catalytic transition metal. It is worth noting that the mechanistic differences between R/R’ and R/X exchanges, even when using the same metal combination, can be striking. For instance, for AuI/SnIV exchanges the double bridges mechanism was found in the Ph/Cl exchange, whereas the Ph/vinyl transmetalation followed an oxidative addition/reductive elimination (OA/RE) pathway via an intermediate with an Au–Sn bond. 9 Here we report a thorough thermodynamic, kinetic, and DFT investigation of the Rf/Cl exchange in (Eq. 1), where the squareplanar complex trans-[RhRf(CO)(AsPh3)2] (1) and the linear complex [AuCl(AsPh3)] (2) lead to the products trans- [RhCl(CO)(AsPh3)2] (3) and [AuRf(AsPh3)] (4). The aryl Rf is chosen for simplicity of its 19F NMR signals, and the Rh complex because the high preference of CO for the trans position to Rf or Cl avoids the formation of other isomers. Au and Rh complexes with the same neutral ligand (AsPh3) are used to prevent multiplication of signals due exclusively to ligand exchanges. (1) In the case studied here, the exchange equilibrium is highly displaced to formation of [AuRf(AsPh3)] (4) and trans- [RhCl(CO)(AsPh3)2] (3). 10 The exchange in stoichiometric conditions (1:1 molar ratio) was monitored in CD2Cl2 in both senses until steady concentrations of the species were reached [a] Mr. Marconi N. Peñas-Defrutos, Dr. Camino Bartolomé, and Prof. Dr. Pablo Espinet. IU CINQUIMA/Química Inorgánica, Facultad de Ciencias, Universidad de Valladolid, 47071-Valladolid (Spain). E-mails: cami[email protected].es and espi[email protected].es [b] Prof. Max García-Melchor School of Chemistry, CRANN and AMBER Research Centres, Trinity College Dublin, College Green, Dublin 2, Ireland. E-mail: [email protected]e Supporting information for this article is given via a link at the end of the document. COMMUNICATION 2 (24 h at room temperature), affording Keq = 7×103 (ΔG0 = ‒5.2 kcal mol‒1). This equilibrium constant corresponds to 99% conversion (Figure SX), which would allow, for a simpler kinetic fitting, handling the exchange as an irreversible process with sufficiently good approximation. The reaction 1+2 in CD2Cl2 has an appropriate rate for very precise 19F NMR monitoring at 273 K, and the initial rates method can be applied. 11 The integrals of well-separated Fortho signals of species 1 and 4 were monitored. Neither by-products nor reaction intermediates were detected. Least squares adjustment of the data (Figure 2, orange line) yielded an initial reaction rate r0 = 2.11×10−7 mol L−1 s−1, corresponding to ΔG‡273K = 19.3 kcal mol−1 for the overall process (for details, see kinetic section in the SI). This Rh to Au Rf/Cl exchange is remarkably faster than the Rf/Pf scrambling studied before.5 Experiments using 2:1 or 1:2 Au:Rh ratios, doubled the rate of the stoichiometric reaction, confirming first order reaction kinetics in both reactants. Finally, in striking contrast with the Rf/Pf exchange, which showed high deceleration with small percentages of AsPh3 added, the addition of free AsPh3 to Eq. 1 accelerates the transmetalation (Figure 2). Figure 2. Concentration vs time data (triangles) obtained by monitoring the formation of the product [AuRf(AsPh3)] (4) by means of 19F NMR in CD2Cl2 at 273 K with different amounts of free ligand added. Initial concentrations of the reactants: [Rh]0 = [Au]0 = 1.0 × 10‒2 mol×L‒1. Lines represent data adjusted with COPASI software. As shown in Figure 2, even substoichometric amounts of ligand result in a noticeable increase in the reaction rate. More specifically, the addition of 10 mol% AsPh3 relative to gold concentration accelerates the transmetalation rate by ca. 60% (blue line), while a four-fold increase in the initial rate compared to the reaction in the absence of additional AsPh3 (orange line) is observed with 50 mol% of ligand (violet line). The reaction rate increases further when moving to superstoichometric amounts of ligand, although the observed correlation is not linear, i.e. the experiment with 500 mol% AsPh3 is only ca. 60% faster than with 50 mol% (see Table SX for details). Hence, the transmetalation rate seems to level off at high ligand concentrations as we can see comparing grey and green lines. These experiments support that, different to the case of Rf/Pf exchange, there is no AsPh3 dissociation involved. On the contrary they suggest the existence of two competing reaction pathways: one that has no participation of the added AsPh3 ligand (Pathway A), and a second one with a lower activation energy, in which AsPh3 participates (Pathway B). For the latter a reasonable proposal is that some added arsine coordinates to gold, makes it more prone to oxidation, and is released after the transmetalation. Consequently, it is in fact an AsPh3 catalyzed process. Both mechanisms lead to formation of the same species and are connected by a coordination equilibrium constant. In order to achieve a fine adjustment of the experimental data we have carried out kinetic simulations with COPASI software. 12 The kinetic model summarized in Scheme 1, fits well the experimental data (lines in Figure 2) both in the absence and presence of additional concentrations of AsPh3 ligand (see SI for details). Scheme 1. Kinetic model for COPASI fitting. Without extra ligand only pathway A is operating, with a COPASI adjusted activation energy barrier value of ΔGA‡ = 19.2 kcal mol-1. A crucial aspect to be modelled upon addition of AsPh3 is the equilibrium constant to form the active [AuCl(AsPh3)2], which determines the amount tricoordinate species formed depending on the of arsine added. The saturation observed when large amounts of ligand are added (5 and 20 eq, Figure 2) suggests substantial contribution of Pathway B, and the adjustment leads to ΔGB‡ = 17.7 kcal×mol‒1 (see SI for details). We conducted a mechanistic investigation of the two competing reaction pathways by means of density functional theory (DFT) calculations at the wb97xd level (see SI for details). The Gibbs energy profile obtained for the uncatalyzed Pathway A is shown in Figure 3. Figure 3. Gibbs energy diagram (in kcal mol–1) for the transmetalation reaction between 1 and 2 in CH2Cl2 (Pathway A). COMMUNICATION 3 The first step involves the interaction of complexes 1 and 2 to yield a weakly bound Van der Waals complex (I1), which lies 3.8 kcal mol–1 above the separated reactants. From this intermediate, the oxidative insertion of the Au center into the Rh‒C bond takes place via the transition state TS1, requiring an overall activation energy of 18.2 kcal mol–1. This step results in the formation of the reaction intermediate I2, wherein both metal centers display square planar geometries characteristic of AuIII and RhI complexes. Subsequently, the transmetalation of the Cl group occurs in a stepwise process which involves two transition states (TS2 and TS3) of relatively low energy. In particular, TS2 entails the formation of an Au–Cl–Rh bridge with a relative barrier of only 1.6 kcal mol–1 leading to the intermediate I3, which features substantial labilization of the AsPh3 coordinated to gold. Finally, the Au–Cl bond cleavage takes place via TS3 to afford the intermediate I4. This last intermediate is analogous to I1 but with the reaction products trans-[RhCl(CO)(AsPh3)2] (3) and [AuRf(AsPh3)] (4) formed. The rate determining step corresponds to the oxidative insertion (TS1) with an activation energy that matches very well with the experimental kinetic studies (18.2 vs 19.2 kcal×mol‒1, respectively.) To gain a better understanding of the oxidative insertion step, in the following we analyze the optimized structures of TS1 and I2 (Figure 4) in detail. Figure 4. Balls and sticks optimized structures of TS1 (left) and I2 (right) with selected bond distances (Å). Ph groups of the AsPh3 ligands are omitted for clarity. Sum of covalent radii: Au-Rh = 2.78 Å; Au-C = 2.09 Å; Rh-C = 2.15 Å; Au-As = 2.55 Å; Au-Cl = 2.38 Å. 13 Interestingly, while the Rf group acts as a bridging group between both metals in TS1 (with typical M‒Cipso distances ca. 2.3 Å), the Cl behaves as a mere spectator and remains coordinated to gold. Hence, the sequential formation of one bridge with strong Au‒Rh interaction (2.828 Å in TS1) results to be more feasible than the exchange via mixed doubled bridges, which precludes this type of metallophilic interactions. Similarly, the resulting intermediate, I2, features an Au–Rh bond with an intermetallic distance comparable to that found by X-Ray diffraction in an heterobimetallic complex, (i.e. 2.673 and 2.690 Å, respectively). 14 Moreover, a noticeably elongation of the Au‒ As is observed in I2, which can be attributed to the very large trans influence of the Au‒Rh bond.5 Coming back to the AsPh3 accelerating effect, it must be the consequence of the existence of an uncovered associative process. We also note that the 19F NMR spectra recorded either from trans-[RhRf(CO)(AsPh3)2] (1) or [AuRf(AsPh3)] (4) did not show any noticeable change in the chemical shifts upon the addition of a large excess of free AsPh3 ligand, which rules out these species as the origin of the ligand dependence. Despite the fact that gold(I) tendency to linear coordination is much higher than that of other M(I) coinage metals, there are some examples in the literature of trigonal planar or pseudotetrahedral AuI complexes. 15 For example, the reaction of [AuCl(PPh3)] with an excess of PPh3 in polar solvents has been shown to form [AuCl(PPh3)n] with n = 2 or 3. 16 Notably, the XRay structure of the four-coordinate [AuCl(PPh3)3] shows a clear deviation from tetrahedral geometry with a very long Au···Cl distance (2.71 Å vs 2.28 Å in the linear [AuCl(PPh3)]) which resembles an ion pair. Similar multiple coordination equilibria might also occur with an excess of AsPh3. In fact, [AuX(AsPh3)m] structures (X = halide; m = 1, 2, 3,) have been reported, featuring a progressive elongation of the Au‒X bond upon coordination of additional ligands. 17 Even the formation of the tetrahedral complex [Au(AsPh3)4]+ can be achieved with a noncoordinating counteranion such as BF4‒. 18 In our case, the formation of [AuCl(AsPh3)2] under reaction conditions (CH2Cl2) was confirmed by mass spectrometry, even in the presence of substoichoimetric amounts of added ligand (exact mass of Au(AsPh3)n for n = 2 was observed, while the aggregate with n = 3 was not. See SI for details). These findings are in agreement with DFT calculations, which predict a Gibbs reaction energy of –1.1 kcal×mol−1 in CH2Cl2 for the formation of [AuCl(AsPh3)2] (5) from 2 and AsPh3, pointing to an equilibrium between these species. With this knowledge, we set out to model the transmetalation mechanism for the AsPh3-promoted Rf–Cl exchange (Pathway B) starting from the three-coordinate Au complex 5 and the Rh complex 1. The Gibbs energy profile obtained for this exchange is shown in Figure 5. Figure 5. Gibbs energy diagram (in kcal mol–1) for the ligand-promoted transmetalation between 1 and 5 (Pathway B) in CH2Cl2. Similarly, to the uncatalyzed mechanism (Figure 3), the first step in the AsPh3-catalized pathway involves the formation of the Van COMMUNICATION 4 der Waals complex between 1 and 5, followed by the oxidative insertion of Au into the Rh‒CRf bond through TS1*. Importantly, this transition state represents the rate determining step of the process, requiring an activation energy barrier of 18.8 kcal mol−1 and features the Rf ligand as the only bridging group and a rather short Au‒Rh distance (i.e. 2.820 Å), despite the bulky environment created by the four AsPh3 ligands. Furthermore, the additional AsPh3 ligand in TS1* (highlighted in orange in Figure 5) remains coordinated to Au, while the Cl group is completely dissociated (highlighted in green) and interacting with several CH groups from the AsPh3 ligands (see SI for the structures of TSs and intermediates with relevant distances). Hence, both TS1* and the resulting intermediate I2* are better described as ion pair species. The latter also displays a noticeably elongation of the Au‒As bond in trans to Rh, which undergoes complete dissociation with the concomitant formation of the Rh‒Cl bond via TS2*. This last step requires a relative barrier of only 2.0 kcal mol−1 and leads to the intermediate I3*, wherein the transmetalated products 3 and 4 are weakly interacting and the additional AsPh3 ligand has been regenerated to participate in the next catalytic cycle. The computed values for the overall activation energy barriers for both pathways are practically the same (18.2 vs 18.8 kcal mol‒1). Energy differences in the order of 1 kcal mol‒1 are excellent for computations (specially involving ion pair species, in which the SMD solvent corrections are less accurate 19 ). Although in this case this happens to be unfortunate, since experimentally the AsPh3-catalyzed Pathway B is somewhat lower in energy than the uncatalyzed Pathway A, this contradiction (possible within the limits of calculation accuracy) does not question the quality of the DFT results and the value and correctness of the structures observed computationally along the pathway. In conclusion, the RhIAr/AuICl transmetalation reaction studied in this work does not involve the prototypical transition states with mixed double bridges. Instead, this exchange occurs via oxidative insertion of Au into the Rh‒C bond through transition states and reaction intermediates which display short Au‒Rh distances. This marked tendency for both Ar/Ar’,5 and Ar/Cl exchanges to occur via heterometallic bonded species might make to reconsider existing mechanistic proposal for other transmetalation reactions in related systems, made in the lack of experimental data and calculations. Moreover, the accelerating effect of AsPh3 reveals the existence of cooperative pathways, being the ligand promoted the most feasible one. The fact that gold tricoordination eases its oxidative insertion is closely related with several examples reported recently where oxidative addition probes to be much more feasible in bent (by chelate ligands) AuI species compared to their linear analogues. 20 This work, therefore, paves the way for the development of new ligand promoted processes involving gold(I) complexes, not only with chelate ligands but also via reversible ligand association equilibria. Acknowledgements The authors thank the financial support from the Spanish MINECO (projects CTQ2016-80913-P and CTQ2017-89217-P) and the Junta de Castilla y León (projects VA051P17 and VA062G18). The DJEI/DES/SFI/HEA Irish Centre for High-End Computing (ICHEC) is also acknowledged for the provision of computational facilities and support. M. N. P.-D. gratefully acknowledges the Spanish MECD for a FPU scholarship. Keywords: bimetallic catalysis • oxidative addition • gold • rhodium • transmetalation mechanism • density functional theory • microkinetic modeling [ 1 ] a) M. H. Pérez-Temprano, J. A. Casares, P. Espinet, Chem. –Eur. J. 2012, 18, 1864–1884; b) D. R. Pye, N. P. Mankad, Chem. Sci. 2017, 8, 1705–1718. [ 2 ] a) Y. Shi, S. A. Blum, Organometallics, 2011, 30, 1776–1779; b) J. delPozo, J. A. Casares, P. Espinet, Chem. – Eur. J. 2016, 22, 4274– 4284; c) R. J. Oeschger, P. Chen, J. Am. Chem. Soc. 2017, 139, 1069– 1072; d) M. Oi, R. Takita, J. Kanazawa, A. Muranaka, C. Wang, M. Uchiyama, Chem. Sci. 2019, 10, 6107–6112. [ 3 ] For detailed mechanistic study of a Pd/Au transmetalation see: M. H. Pérez-Temprano, J. A. Casares, A. R. de Lera, R. Álvarez, P. Espinet, Angew. Chem. Int. Ed. 2012, 51, 4917–4920. [ 4 ] For the Pd/Au bimetallic application see: J. delPozo, D. Carrasco, M. H. Pérez-Temprano, M. García-Melchor, R. Álvarez, J. A. Casares, P. Espinet, Angew. Chem. Int. Ed. 2013, 52, 2189–2193. [ 5 ] M. N. Peñas-Defrutos, C. Bartolomé, M. García-Melchor, P. Espinet, Angew. Chem. Int. Ed. 2019, 58, 3501–3505. [ 6 ] By “oxidative insertion” we mean an oxidative addition where, being two metals involved in the redox process with formation of an M-M' bond, each metal changes only in one unit its formal oxidation number. [ 7 ] For transmetalation reactions employing haloaryl groups see: a) A. L. Casado, P. Espinet, Organometallics, 1998, 17, 3677–3683. b) M. N. Peñas-Defrutos, C. Bartolomé, M. García-Melchor, P. Espinet, Chem. Commun. 2018, 54, 984–987. c) M. Pérez-Iglesias, O. Lozano-Lavilla, J. A. Casares, Organometallics, 2019, 38, 739–742. [ 8 ] Note that the reversibility of transmetalations is often source of complications, such as formation of homocoupling products: J. del Pozo, G. Salas, R. lvarez, J. A. Casares, P. Espinet, Organometallics, 2016, 35, 3604–3611; and references therein. [ 9 ] D. Carrasco, M. García-Melchor, J. A. Casares, P. Espinet, Chem. Commun. 2016, 52, 4305–4308. [ 10 ] Obviously only 1 and 4 are observed by 19F NMR (1H NMR is not informative in our case). C6F3Cl3-1,3,5 is used as internal reference. [ 11 ] S. Kozuch, S. Shaik, Acc. Chem. Res. 2011, 44, 101–110. [ 12 ] S. Hoops, S. Sahle, R. Gauges, C. Lee, J. Pahle, N. Simus, M. Singhal, L. Xu, P. Mendes, U. Kummer, Bioinformatics, 2006, 22, 3067–3074. [ 13 ] B. Cordero, V. Gómez, A. E. Platero-Prats, M. Revés, J. Echeverría, E. Cremades, F. Barragán, S. Alvarez, Dalton Trans. 2008, 2832–2838. [ 14 ] C. Bianchini, C. J. Elsevier, J. M. Ernsting, M. Peruzzini, F. Zanobini, Inorg. Chem. 1995, 34, 84‒92. [ 15 ] (a) M. C. Gimeno, A. Laguna, Chem. Rev. 1997, 97, 511‒522. (b) M. A. Carvajal, J. J. Novoa, S. Alvarez, J. Am. Chem. Soc. 2004, 126, 1465– 1477. [ 16 ] [AuCl(PPh3)2] has been thoroughly studied due to its photophysical interest: (a) M. Hoshino, H. Uekusa, S. Sonoda, T. Otsuka, Y. Kaizu, Dalton Trans. 2009, 3085‒3091. (b) M. Hoshino, H. Uekusa, S. Ishii, T. Otsuka, Y. Kaizu, Y. Ozawa, K. Toriumi, Inorg. Chem. 2010, 49, 7257‒ 7265. For [AuCl(PPh3)3] see: (c) P. G. Jones, G. M. Sheldrick, J. A. Muir, M. M. Muir, L. B. Pulgar, J. Chem. Soc., Dalton Trans. 1982, 2123‒2125. COMMUNICATION 5 [ 17 ] For [AuCl(AsPh3)] see: B. Weissbart, L. J. Larson, M. M. Olmstead, C. P. Nash, D. S. Tinti, Inorg. Chem. 1995, 34, 393–395. For [AuX(AsPh3)m] (m = 2, 3 see: (b) G. A. Bowmaker, P. C. Healy, A. N. Sobolev, A. H. White, Aust. J. Chem. 2019. DOI: 10.1071/CH19340. (c) P. C. Healy, A. H. White, W. T. Robinson, 2017, CCDC number 1526150 (Private Communication). [ 18 ] U. M. Tripathi, A. Bauer, H. Schmidbaur, J. Chem. Soc. Dalton Trans. 1997, 2865‒2888. [ 19 ] A. V. Marenich; C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B 2009, 113, 6378‒6396. [ 20 ] (a) M. Joost, A. Amgoune, D. Bourissou, Angew. Chem. Int. Ed. 2015, 54, 15022–15045. (b) M. Joost, A. Zeineddine, L. Estévez, S. MalletLadeira, K. Miqueu, A. Amgoune, D. Bourissou, J. Am. Chem. Soc. 2014, 136, 14654–14657. (c) Y. Yang, L. Eberle, F. F. Mulks, J. F. Wunsch, M. Zimmer, F. Rominger, M. Rudolph, A. S. K. Hashmi, J. Am. Chem. Soc. 2019, 141, 17414–17420.