User:Mcannos/sandbox/gas-phase organometallic chemistry

An Introduction to Gas-Phase Organometallic Chemistry


Organometallic chemistry has long been studied in solvated systems, owing to the practicality of the moderate reaction conditions. While solution-phase chemistry has classically facilitated the study of structure, bonding, and reactivity, there are limitations its scope of practice. Namely, solution-phase chemistry offers limited insight to the mechanistic details of transition metal catalysis. Since gas-phase catalysis commonly occurs at very low concentrations (relative to solution phase chemistry), the occurrence of competing side reactions becomes significant. Although this hinders the industrial practicality of gas-phase catalysis, it presents a unique opportunity to characterize the mechanistic details of organometallic reactions without solvent effects.

Experimental characterization is primarily accomplished by ion cyclotron resonance (ICR) and mass spectrometry (MS) techniques. MS provides a means to detect reaction intermediates that might be too short-lived to detect via solution-phase techniques. Thus, unprecedented insights of organometallic mechanistic details are attainable.

Electrospray Ionization Tandem Mass Spectrometer for Gas-Phase Catalytic Analysis
Electrospray ionization (ESI) originated in the paint and coating industry, and was first applied to mass spectrometry by John Bennett Fenn and co-workers. The method was used while working with proteins and peptides, which would soon award Fenn with the 2002 Nobel prize. During the electrospray ionization, intact molecular ions are transferred from dilute solutions directly to the gas phase by a process involving droplet formation, fission, and field desorption. The ions produced in the electrospray source are then transmitted through a heated capillary to aid in desolvation, then passed in front of the skimmer, where the ions can collide with background gas. The skimmer potential is an adjustable parameter that controls the extent of further fragmentation. This can create coordinately unsaturated complexes by removal of weakly bound ligands, which is useful in reactivity studies. For gas phase reaction studies, the catalysts is ionized, fragmented to dissociate lose ligands, then sent into the tandem mass spectrometer. The LnM catalyst ions pass into the first vessel, the 24-pole reaction cell, where they are either thermalized or collided with reagent gasses. The ion-molecule reactions occur under thermal conditions around 70 °C, with upwards of 105 collisions occurring in the 24-pole cell. The ions then pass through the first quadrupole parent analyzer, where reaction products can be selected and separated, before passing into the second collision cell. In the radio frequency (rf)-octopole, the ions are again allowed to collide with more gas before passing through the second quadrupole for mass-analysis. Studying reactions in the gas phase eliminates solvents, and allows for the selection of only the specific reactive species.

Epoxide Polymerization in Gas-Phase Catalysis
Ring opening polymerization (ROP) is, along with condensation polymerization and chain polymerization, one of the three paths for the synthesis of polymers that surround our lives in the 21st century. ROP is a form of chain-growth polymerization that requires the activation of a cationic precursor, which is often bound to a metal centre. Once activated, the propagation of ring opening can be either radical induced, cationic, or anionic. For cationic ring opening polymerization, the terminal end of the growing chain has a positive charge, which is attacked by an incoming monomer via a SN2 or SN1 mechanism. Polymerization is favoured by ring strain present on the terminal end of the growing polymer.

Binding of Propylene Oxide to Metal Centres in the Gas Phase

Peter Chen and colleagues studied the binding of propylene oxide to charged ligand-metal(III) complexes in the gas phase by ESI tandem mass spectrometry. Two ligands where L= tetraphenylporphyrin (TPP) and salen, were attached to metal centres where M= Al(III), Cr(III), and Co(III). The LnM(III) species were ionized and sent through the ESI tandem mass spectrometer. In the first reaction chamber, the ionized metal species, either (TPP)M+ or (salem)M+, was collided into propylene oxide gas. The resulting (L)M(PO)n+ where n=1 or 2, were selected and sent through the quadrupole into the latter reaction chamber. In the second chamber, Xenon was chosen for collision gas, and the relative binding affinities for the propylene oxide groups were determined by fragmentation analysis.

For reactions involving (TPP)M+ and PO, where M= Al(III), Cr(III), and Co(III), a great difference in activity is seen between the metals, despite similar ionic radii. When M=Co(III), (TPP)Co+ was the dominant species found with only trace amounts of (TPP)Co(PO)2, indicating Cobalt's extremely low affinity for PO. For M=Al(III), both Aluminum cations (TPP)Al(PO)n for n=1 and 2, were found, with a slight preference for n=1. Cr(III) showed a significantly higher affinity for PO, and the (TPP)Cr(PO)2 cation was much preferred.



Reactions involving (salen)M+ and PO had results that parallel those for L=TPP



The metals' affinities for propylene oxide were seen in the second reaction chamber, upon collision with xenon gas. For the reactions involving the loss of PO groups from (L)M(PO)2+, the order of dissociation of PO cations is (TPP)Al(PO)2+ >> (TPP)Cr(PO)2+ ≈ (salen)Cr(PO)2+. This clearly indicates the preference for Cr(III) to retain its 6 coordination state. For the study involving (TPP)M+ and (salen)M+ cations, no polymerization of propylene oxide was observed in the gas-phase; however, Eva Schön et al showed epoxide polymerization in the gas phase was possible with the use of dinuclear Cr(salen) complexes

Gas-Phase Epoxide Polymerization using Cr(salen) complexes

Linked binuclear Cr(salen) complexes are able to polymerize propylene oxide in the gas phase, unlike their mononuclear analogs. This is thought to be explained by a dicationic mechanism in which there is µ2-coordination of the alkoxide chain between the two chromium atoms. E. J. Vandenberg first proposed the idea of more than one metal centre being used in the ring opening polymerization of epoxides on the basis of clean stereochemical inversion during polymerization using aluminum-containing catalysts. Teiji Tsuruta suggested a nearly identical mechanism for the stereoselective ring opening of propylene oxide. One metal centre electrophilically activates the epoxide while the other delivers the nucleophile, known as the "flip-flop" mechanism.

Eva Schön et al. showed that certain dinuclear (salen)Cr(III) complexes that were exposed to electrospray ionization and then collided with propylene oxide in the tandem mass spectrometer underwent ring opening polymerization. Using only mononuclear (salen)Cr(III), mass adducts were found for only species where PO filled the vacant coordination sites on the metal, indicating no polymerization. When dinuclear (salen)Cr(III) complexes were subjected to gas-phase reaction with propylene oxide, mass adducts were found at signals indicating ring opening polymerization, but at varying extents. Dinuclear [(salen)Cr]2+ as seen to the right, showed up to 10 propylene oxide units bound to the dinuclear complex. Dicationic Cr species show activity in the gas phase, whereas monocationic dinuclear Cr complexes show no activity. Vandenberg's anionic "flip-flop" mechanism would predict the opposite, which suggests a cationic mechanism for epoxide ring opening polymerization in the gas phase.



The metal bound end of the growing chain is bound as a µ2-alkoxide, and the terminal end of the growing chain is an oxiranium cation where further epoxide monomers can bind. It has been shown that using a chiral salen chromium chloride catalyst (mononuclear) results in significant chain transferring, occurring as a result of back-biting. In the dinuclear polymerization, the µ2-alkoxide suppresses back-biting, leading to longer oligomer chains. The dinuclear complex also aids in the initiation step; the µ2-coordination of the first epoxide makes the oxiranium cation more electrophilic, which aids in the epoxide ring opening.

Examination of the Bonding Trends of Coinage Metal Catalysts with Application to the Ring Opening of Highly Strained Rings
There has been vast research in the past into gold catalyzed substituted cyclopropane openings in solution phase. However, recently more specific reactions have had the same ring opening effect done by heterogeneous catalysis using a gas phase cyclopropane and a solid metal-carbene catalyst. Creation of an efficient catalyst involved complexing of gold with a stereospecifically substituted N-heterocyclic carbene. Cyclopropane rings are generally good candidates for catalyzed ring opening due to the large amount of ring strain associated with small carbon rings. In 2014, Batiste and Chen examined the trends of using coinage metals (including gold) to mediate the ring opening of cis-dimethoxycyclopropane by gas phase retro-cyclopropanation

The catalytic precursors for the three coinage metals, silver (Ag), copper (Cu) and gold (Au), all contained bulky labile ligands, which readily dissociated to form the cationic metal compound required to complex with cis-dimethoxycyclopropane. The precursors chosen were Au(IMes)(CHPhPPh3)+, Ag(IMes)(N2(CH3)2PPh3)+, and Cu(IMes)2+ where IMes = 1,3-bis(2,4,6-trimethylphenyl)imidazole-2ylidene. Once the successful IMes-[M+]-[CH2[CH(OMe)]2] complex had formed, it was monitored by Collision Induced Dissociation (CID) using argon and xenon gas to deduce fragmentation patterns. The results of this analysis showed two major fragments: IMes-[M]+ (the bare cation) seen in pathways (i) and (iii), and IMes-M-[HCOCH3] (methoxymethylidene complex) seen in pathway (ii) as shown below. An interesting observation noted was that the bare cation could be retrieved both by direct dissociation as well as by internal rearrangement by H-shift of cis-dimethoxycyclopropane prior to dissociation.

Upon individual observation of fragmentation patterns for Ag, Cu, and Au, it appeared as though Ag followed only the bare cation (BC) channel, while Cu and Au followed both the bare cation channel and the carbene abstraction (CA) channel. Threshold Collision Induced Dissociation technique (TCID) was also used to determine the energy barriers of reaction pathways for all three coinage metals. As a general trend, Au followed a more complicated mechanism maintaining significantly lower activation energy for each given intermediate. Cu and Ag follow a slightly less complicated pathway with a much higher activation energy per intermediate overall.

Ag and Cu were better competitors when monitoring the energy barriers associated with the H-shift BC channel. All three coinage metals had more similar activation energies with Au still marginally favourable.

Observation of the above results made it clear that Au was the most efficient catalyst. Bond Energy Decomposition Analysis was used to discover the reasoning behind this. It was shown that Au complexes exhibit a much stronger electrostatic interaction than Ag and Cu due to its large Pauli repulsion. Both sigma (σ) and pi (π) interactions are important stabilizing factors for coinage metals with the σ interaction more important for Au. The out of plane π contribution of IMes and carbenes were considered versus the in plane π contributions as the out of plane interaction allows overlap of the empty p-orbital of the carbene to be filled.

This experiment was particularly useful as it aimed to examine the relative reactivity of coinage metal(I) complexes with cis-dimethoxycyclopropane. The results were interpreted using CID which showed two competing mechanisms. For Ag, the weak σ interaction leading to a weak π back bonding interaction caused little stabilization of the carbene moiety making the CA pathway uncompetitive. On the other hand, Au maintained a strong σ interaction and strong π back donation leading to CA as its main channel. In general, for N-heterocyclic metal carbenes in similar systems, the M-C bond strength displayed the trend Au>Cu>Ag.

Reactions of first row transition metals with unsaturated hydrocarbons in the gas phase
Gas-phase organometallic reactions with alkenes have been applied to many d-block transition elements with the emphasis on the ubiquitous first-row transition metal ions. The results for some of the more known second and third row M+ have often revealed behaviour not predicted from their lighter counterpart. For example, Cr+ reactions with unsaturated hydrocarbons are essentially inert under conditions where Mo+ is highly reactiveGibson. These metals are generally more reactive than their first row homologs, due to the greater radial extension of the valence 4d:5d orbtials compared with the valence 5s:6s orbitals.

Fe+ reactions with alkenes and alkynes

From the work of Baranov et al, ethane, propene, allene, 1,3-butadiene, and isobutene was found to ligate sequentially to Fe+. Based on the available thermochemical data, mainly standard enthalpies of formation, thermodynamics predicts that bi-molecular reactions with ethylene, propene, allene, 1,3-butadiene, and isobutene are energetically unfavorable. Rates of ligation increased as the size of the alkene increased. Behaviours of the alkynes: acetylene, propyne, and diacetylene were similar to that of the alkenes.

Sequential ligation:



L = ethylene, propene, allene, 1,3-butadiene, isobutene, acetylene, propyne, or diacetylene

Cu+ reactions with alkenes

In a study conducted by Gibson, Cu+ was found to be unreactive towards ethane, propene, and cis-2-butene. With cyclohexane and COD, however, it forms CuC4H6+. The reactivity towards COD was expected to be higher because the encounter complex is longer lived, due to more degrees of freedom that can dissipate the internal energy released during formation of the complex.

Dehydrogenation of ethane by Ti+ and V+

Bowers et al studied the dehydrogenation of ethene with V+ and Ti+ in the gas phase. V+ in the ground-state reacts with ethene to give VC2H2+ and was found to be exoergic by 1.3 kcal/mol. The insertion of V+ into the C-H bond, required spin-orbit coupled crossing, and was thought to be the result of the low reaction efficiency. Dehydrogenation with Ti+ in the ground state was found to be exothermic by 10.4 kcal/mol. Like V+ in the ground state, the reaction with Ti+ required spin-orbit coupled crossing, but exhibited a higher reaction efficiency due to stronger covalent bonds formed by Ti+.

Gas-Phase Tandem Oxidative Addition/Reductive Elimination
Reductive elimination (RE) and oxidative addition (OA) are fundamental transformations in transition metal chemistry. Most catalytic cycles operate around an OA/RE mechanistic motif, therefore warranting the extensive investigation of practical TM redox chemistry. Many examples of gas phase OA/RE exist, including the following example by Couzijn et al in 2014.



In his experiments, Couzijn uses a trimethyl platinum(IV) complex as a precursor to demonstrate the reductive elimination of either methane or ethane. Mechanistically, the RE of ethane is unsurprising. Of specific interest is the mechanism by which methane gas evolves. Couzjin proposed that the formerly eliminated ethane can form an agostic complex with 1, forming 3. Subsequently σ-bond metathesis of the ethane primes complex 4 for RE of methane. As with many other reactions, the gas-phase nature of this reaction allowed for characterization of the reactive intermediates by mass spectrometry. This facilitated the mechanistic deductions that were previously inexplicable.

C-H Activation in the Gas Phase
The search for a catalyst active in mild conditions led Arndtsen and Bergman to the introduction of cationic iridium(III) complexes. Bergman believed that the C-H activation occurred in one of two ways: 1) oxidative addition of the 16-electron Ir3+ complex to the C-H bond, forming an 18-electron Ir intermediate, followed by reductive elimination, or 2) a concerted, four-center metathesis of the M-CH, and R-H bond:



It wasn’t until 2 years later when Hinderling, Plattner and Chen offered more insight into this type of reaction using tandem mass spectrometry of ions produced by electrospray ionization. The electrospray mass spectra showed a different mechanism to be operative, involving initial elimination of methane in an intramolecular reaction, followed by addition to the hydrocarbon. This showed that neither mechanism proposed by Bergman was operative:



The search for transition-metal catalysts capable of selective insertion into non-acticated C-H of alkanes and arenes is acknowledged to be one of the most important tasks in organometallic chemistry. This knowledge of the intrinsic reactivity of cationic iridium(III) complexes with regard to C-H activation in the gas phase was utilized for the design of more efficient catalysts for years to follow.

C-C Bond Activation in the Gas Phase
C-C bond activation bears a history of being notoriously difficult owing to the strength of carbon-carbon bonds. In organic chemistry, very few techniques exist for C-C bond activation aside from direct combustion. Some examples include industrial petroleum "crackers": long chain alkanes (heavy oils) can be cleaved by protonating the C-C bond in a zeolite-superacid nano particle. In the inorganic world, C-C bond activation remains unusual, but not unheard of. The diverse reactivity of transition metals facilitates C-C bond activation in relatively mild environments. For instance, in 1985, Tsarbopoulos and Allison investigated the gas-phase chemoselective Co+ insertion into long chain primary alcohols and alkyl halides. Insertion of Co+ can occur through two possible mechanisms: insertion into the C-X (X = halide, or OH) bond; or insertion into the C-C bonds. Although one might expect C-X insertion to occur more readily due to the relatively low bond strength (compared to C-C bonds), the experimental results showed surprising results. For alcohols and alkyl halides where n<3 (n= # of carbons), Co+ inserted into the C-X bonds. However, for substrates with n>4, cobalt inserted almost exclusively into the C-C bonds. Thus, one may consider n>4 substrates to behave more like alkanes. The authors reasoned that the reaction proceeded through a (relatively stable) metallocyclic intermediate, which facilitated the C-C bond activation.



When considering the mechanism shown above, one can imagine the insertion of cobalt into a different C-C bond to form a 5-membered ring. In fact, owing to the both the diverse range of insertion points and the multiple mechanisms available, C-C bond activation of long chain alcohols produced a highly complex mixture of saturated and unsaturated products.

...
Add your username here if you would like to contribute to this wiki page:

1. poyseybear

2.Thai.j.nguyen

3. Jordan Friedmann

4. Andrew_Roberts93

Scott here: Some guidance on this. The most interesting examples here are NOT unimolecular decompositions in the gas phase (i.e. simple fragmentation reactions) nor should you include examples of solution-phase chemistry that are simply characterized by gas-phase methods (i.e. most of the work my own research group does, though http://web.uvic.ca/~mcindoe/80.pdf is a notable exception). A good place to start is looking at the work of Peter Chen (ETH Zurich) - he's done neat work including gas phase polymerizations. Follow some of that work forward (i.e. who has cited him?).