Light-induced olefin metathesis

Light activation is a most desirable property for catalysis control. Among the many catalytic processes that may be activated by light, olefin metathesis stands out as both academically motivating and practically useful. Starting from early tungsten heterogeneous photoinitiated metathesis, up to modern ruthenium methods based on complex photoisomerisation or indirect photoactivation, this survey of the relevant literature summarises past and present developments in the use of light to expedite olefin ring-closing, ring-opening polymerisation and cross-metathesis reactions.


Scheme 1: Light activated metathesis of trans-2-pentene.
In the Dubois paper, when either light or carbon tetrachloride was excluded, no reaction could be observed. Furthermore, this research led to the proposal of the controversial mechanism shown in Scheme 2, which includes the generation of phosgene and the proposed active species, chloropentacarbonyl tungsten (2).
Alternatively, McNelis proposed that the active species was actually dichlorotetracarbonyl tungsten and demonstrated that phosgene was not generated by illumination of 1 when oxygen was excluded [38].
Following these first communications, Krausz, Garnier and Dubois published a series of papers investigating the photoinduced olefin metathesis with complex 1 [39][40][41][42]. Their main conclusions were: a: W(CO) 5 and CO are created by photolysis of W(CO) 6 . The reactivity of W(CO) 5 was found to be solvent dependent.
b: The intensity of irradiation, and the concentrations of both olefin and catalyst had a significant effect on reaction yields. c: The RCH=CCl 2 produced in the reaction was a result of a reaction between a tungsten dichlorocarbene species and the double bond.
Research on the photo-activation of W(CO) 6 was further expanded by Harfouch et al. [43,44], Matsuda et al. [45], Szymańska-Buzar and Ziόłkowski [46,47] and Zümreoglu [48], and was first reviewed in 1988 by Szymańska-Buzar [49]. The main conclusions from this wave of research dealt with the mechanistic role of the halide additives, as well as diverse reaction conditions, such as the use of other metals (i.e., Cr and Mo) that usually led to addition type reactions instead of metathesis.
In later work, Mol et al. [50] determined the heterogeneous character of the active catalytic species obtained on irradiation of 1 in CCl 4 , supporting the previous proposal by Harfouch and coworkers. In another example of heterogeneous light induced metathesis, Shelimov and Kazansky [51] also found that silica supported molybdenum trioxide (MoO 3 /SiO 2 ) could be activated by UV irradiation under an alkane atmosphere in the metathesis of propene and 1-hexene.

Well-defined tungsten catalysed photometathesis
The first example of well-defined early transition metal complexes for photocatalysed ROMP (PROMP) 3 and 4 ( Figure 1) was published by van der Schaaf, Hafner, and Mühlebach [59].
Complexes 3 and 4 displayed reasonable thermal stability in solution (no decomposition was observed after 1 d at 80 °C), but were moisture sensitive and had to be handled under an inert atmosphere. Complex 3 slowly polymerised dicyclopentadiene (DCPD) in the dark at 60 °C, in contrast, when exposed to light even at room temperature, polymerisation was rapid and complete after 15 min. Enhanced behaviour was observed with complex 4, which boasted true latency [60] and did not show

Ruthenium catalysed photometathesis
The first example for PROMP using ruthenium initiators was disclosed by Mühlebach et al. in 1995 [61]. Three main types of ruthenium-based precatalysts: η 6 -arene sandwich complexes, half-sandwich complexes and nitrile complexes ( Figure 2) were shown to promote the polymerisation of strained cyclic alkenes ( Figure 3) when irradiated by UV light.
Advantageously, the systems showed none to moderate activity for normal ROMP, and were not sensitive to oxygen and humidity compared to the tungsten initiators described above. Notably, heating the sandwich catalysts for more than 24 h at 50 °C in the presence of the monomers did not induce polymerisation, on the other hand, the nitrile complexes displayed low activity even at room temperature. UV irradiation at 364 nm significantly enhanced the activity of all these complexes. The activity of the compounds, including comparisons with thermally active catalysts, are summarised in Table 1.
Overall, the polymers obtained from norbornene and oxanorbornene derivatives had high molecular weights (M w > 150 kDa) and also high monomodal polydispersities (M w /M n > 2.3). Polymerisation was also found to occur if the monomer was added after irradiation of the complex. In addition, the initial rate of polymerisation was found to be linearly dependent on the irradiation time. Conversely, irradiation above 420 nm did not initiate polymerisation for most of the complexes investigated.
As in most systems of this type, activation of the precatalyst is dependent on the photochemically induced cleavage of a metal-ligand bond which leads to the active species. The mechanism for initiation of the sandwich compounds was proposed to proceed through gradual photodissociation of the arene ligand, followed by solvation of the photochemically excited molecule (Scheme 3). Disadvantages of these catalysts are the overall moderate activities achieved and that they are only soluble in polar solvents, due to their cationic character. On the other hand, the complexes are readily available and the use of aqueous solvents as the reaction media can also be envisaged as an attractive feature.
Mühlebach, Hafner and van der Schaaf [62] carried on the development of ruthenium and osmium photoactivated catalysts ( Figure 4) by adding a bulky phosphane ligand to the complex. Thus, a more active and soluble neutral species with the anionic ligands bound to the metal could be obtained. The same concept of arene displacement by UV radiation was used for the release of a p-cymene ligand to produce a more reactive catalytic species. Osmium precatalysts 11a-f did not polymerise norbornene under standard conditions. However, 5 min irradiation of a toluene solution of the complex with a 200 W Hg lamp led to active catalysts. The more active catalysts were those possessing more sterically hindered phosphane ligands. Thus, complexes with larger cone angles θ, such as 11b (θ = 160°) and 11c (θ = 170°) showed strong metathetic activity for PROMP of norbornene and dicylopentadiene in toluene solution or even in aqueous dispersions; by contrast, complexes 11a,d,e,f (θ = 130°, 145°, 120°, 150°) showed slow no reaction even after UV irradiation. The ruthenium complexes 12 showed much higher reactivity in the polymerisation of norbornene, albeit none of these complexes was completely thermally latent for this reaction. The remarkable tolerance of 12b to impurities and water, was highlighted by the fact that polymerisation can take place in water dispersions containing fillers such as SiO 2, AI(OH) 3 , or CaCO 3 up to a loading of 70 wt %. The electrical and mechanical properties of PDCPD were preserved, making this system highly interesting for novel applications. The work of Mühlebach, Hafner and van der Schaaf on photoinduced ring-opening metathesis polymerisation was reviewed in 1997 [63].
The interest in photoactivated olefin metathesis motivated Fürstner [64] to use complex 12b for photoinduced ring closing metathesis (RCM) by using regular neon light or strong daylight as a photon source instead of the UV lamps used thus far. Indeed, 12b catalysed the ring-closing metathesis reaction of diallyl tosylamide in 90% yield when illuminated by neon light. Alternatively, the commercially available dimer complex 13 ( Figure 5) mixed with PCy 3 could be irradiated to produce similar results. Some of the photoinduced RCM products obtained by neon light irradiation of dimer 13 are highlighted in Figure 6. Perhaps the main benefit of this procedure lies in its simplicity since it only requires commercially available metal complexes and commonly used lighting equipment.
The best conditions reported for this reaction were the irradiation of a toluene solution of 14 and substrate with an Hg lamp at 300 nm for 30 min at room temperature, followed by heating at 80 °C until completion of the reaction. The reaction time, compared to the non-irradiated control experiment, was reduced six fold. The various dihydrofuran rings obtained by this method are shown in Figure 7 along with their respective yields and the required reaction times.  In line with the development of ruthenium benzylidene initiators [66,67], the phosphane ligand was replaced by an N-heterocyclic carbene (NHCs) in the photoactivated precatalysts. Accordingly, Noels et al. [68,69] synthesised a range of NHC substituted ruthenium cymene complexes ( Figure 8) either by replacing the phosphane ligand in complex 12 by NHC ligands or via direct synthesis from 13. Complexes 15 and 16 were tested as photoactivated ROMP catalysts. In all cases cyclooctene was used as a standard cyclic olefin monomer for the polymerisation studies.
Illuminating 15 or 16 with intense visible light, or even with regular laboratory lighting, revealed a dramatic improvement in their ROMP activity. Results of PROMP by complex 16g (Mes ligand) and cyclooctene are summarised in Table 2. The large difference in polymer molecular weight between the dark and light reactions was not explained, even though it is slightly counterintuitive.
In order to improve the understanding of the photoactivation mechanism, complex 16g was irradiated by visible light both in the presence and absence of cyclooctene. NMR and UV spectra confirmed the release of p-cymene from the complex in the photochemical process; however, the active species and overall mechanism were not elucidated. Both saturated and unsaturated NHC ligands afforded similar results. However, blocking both ortho positions on the aromatic groups of the NHCs was crucial for the performance of the catalyst.
The ruthenium photoactivated catalytic systems described so far possessed noticeable ROMP activity at temperatures higher than room temperature even without being exposed to light; especially with the more reactive monomers such as norbornene and dicyclopentadiene. Buchmeiser argued that in order to integrate light activated precatalyst in practical applications, latency must be significant also at higher temperatures. Therefore, Buchmeiser et al. concentrated efforts towards the design of 'true' latent photoactivated ruthenium precatalysts for  ROMP [70]. These precatalysts, similar to Noels' complexes, included NHCs, and in some cases the p-cymene ligand was exchanged for phenylisonitrile and in others the chloride ligands were replaced by trifluroacetate. These transformations were expected to generate more stable, inert, precatalysts that would require external triggers to initiate the dissociation of the neutral ligand to produce the active species.
Complexes 17-20 (Scheme 5) initiated ROMP of norborn-5ene-2-ylmethanol 21 only at temperatures over 40 °C or by illumination at 172 nm. However, the reaction conditions required the removal of oxygen from the system, an encumbrance in practical applications. In addition, more reactive monomers, such as 2-norbornene, were thermally polymerised by the complexes even at room temperature. In 2008 Buchmeiser introduced the improved cationic latent phototriggered precatalysts 22 and 23 ( Figure 9) [71]. These cationic species were inactive at higher temperatures (T < 45 °C) and did not thermally initiate the polymerisation of several ROMP monomers, including the highly reactive dicyclopentadiene 25 (DCPD) (Figure 10).
Irradiation at 308 nm of complex 22 or 23 in chloroform in the presence of the monomers resulted in polymerisations with 5-99% conversion yields (Table 3). On changing the light source to a 254 nm Hg lamp improved the yields to 70-99% ( Table 3).
The proposed mechanism for the photoactivation of precatalysts 22 and 23 is displayed in Scheme 6.
Three additional new complexes for PROMP were recently published by Buchmeiser et al. [72] Although the complexes Scheme 5: Ruthenium NHC complexes for PROMP containing p-cymene and trifluroacetate (17,19) or phenylisonitrile ligands (18,20).   Figure 11) are not true latent precatalysts, only minor polymerisation occurred in the absence light (<5% at room temperature after 24 h). However, when the monomer-complex mixture was irradiated with a 254 nm UV source, polymerisation occurred with more than 60% conversion within 1 h in most cases.

30-32 (
A desirable enhancement in phototriggered catalysis is the generation of photoswitchable systems. Thus, a specific reaction may be turned on by one type of stimulus (heat, light), and turned off by another. We have recently developed latent sulfur chelated Hoveyda-Grubbs type complexes (Figure 12), as thermo-switchable catalysts for RCM and ROMP [73][74][75][76][77].  The thermal latency of the sulfur chelated complexes make them ideal candidates for photoswitchable applications; especially in light of the well documented photodissociation of ruthenium sulfur ligands.
Therefore, complexes 33a, e, f, g were irradiated at 365 nm in the presence of RCM and ROMP substrates [78]. A summary of the results is presented in Table 4 and Table 5.

Scheme 7:
Proposed mechanism for the photoactivation of sulfur-chelated ruthenium benzylidene.   The discovery that light irradiation induces photoisomerisation of the cis-dichloro complexes led to the proposed mechanism shown in Scheme 7. Thus, photoactivation of sulfur-chelated ruthenium benzylidene complexes was found to depend on the generation of the active trans-dichloro isomer via 14-electron intermediates.
The design of a photoswitchable system was based on the fact that the latent isomer (cis-dichloro) was thermodynamically more stable than its active counterpart (trans-dichloro). Thus, illumination with UV light generates an active isomer, but a short heating period regenerates the inactive isomer. The switchable nature of the system was demonstrated by 15 min irradiation of a tetrachloroethane solution of diethyldiallyl malonate with 5 mol % of catalyst 33e (activation), followed by 5 min heating at 80 °C (deactivation). Thus, whilst heating may be initially perceived as counterintuitive, it may be used to regenerate the latent species instead of activating it.

Indirect metathesis photoactivation
An alternative approach for photoinitiated metathesis is indirect activation.
Grubbs et al. [79] demonstrated the use of photoacid generators (PAG) 34 and 35 ( Figure 13) for the sub-300 nm UV activation of metathesis precatalysts 36 and 37 (Scheme 8). Thus, an acid sensitive olefin metathesis catalyst can be photoactivated by using a PAG in a tandem approach. Clearly, all other acids must be excluded from the reaction mixtures for the procedure to be effective. Acid sensitive complexes 36 and 37 were synthesised starting from the Grubbs type catalysts 38 and 39 and thallium hexamethylacetylacetonate (Scheme 8). Several RCM substrates were closed as shown in Table 6 by the PAG induced metathesis. Complexes 36 or 37 and PAG 34 also polymerised typical ROMP monomers in excellent conversion (Table 7).
A dramatic loss of activity was observed when the chloride anion of PAG 31 was replaced by the non-nucleophilic nonaflate ion, implying that the chloride plays a crucial role in  the photoactivation of this system. A cunning trapping experiment using the isopropoxy aromatic derivative 40 (Scheme 9) provided a better understanding of the activation mechanism and led to the proposal of the well-known 14-electron complex 41 as the actual active species.
Another excellent example for nondirect activation of metathesis is the light-triggerable liquid-filled solid microcapsules (MCs) presented by Fréchet et al. in 2009 [80]. A solution of 39 in toluene inside macrocapsules ( Figure 14) can stand for weeks in neat DCPD without any appreciable reaction. However, near-IR laser bursting of the MCs causes gelling due to polymerisation within min.

Conclusion
Initially, the field of photoinduced olefin metathesis was mainly based on W(CO) 6 chemistry. However, slowly over time functional group tolerant ruthenium applications emerged and these now dominate the field. Either by indirect or direct methods, the activation of ruthenium olefin metathesis catalysts may lead the way to novel applications in the areas of photolithography [81,82], roll-to-roll coating [69], 3D-printing, and self-healing [83] procedures in polymers. The use of photoswitchable catalysts may also have important safety advantages for industrial processes. As frequently seen in the successful field of olefin metathesis, the use of light to activate and direct these reactions certainly has a bright future ahead.