Light-Activated Olefin Metathesis: Catalyst Development, Synthesis, and Applications.

The most important means for tuning and improving a catalyst's properties is the delicate exchange of the ligand shell around the central metal atom. Perhaps for no other organometallic-catalyzed reaction is this statement more valid than for ruthenium-based olefin metathesis. Indeed, even the simple exchange of an oxygen atom for a sulfur atom in a chelated ruthenium benzylidene about a decade ago resulted in the development of extremely stable, photoactive catalysts. This Account presents our perspective on the development of dormant olefin metathesis catalysts that can be activated by external stimuli and, more specifically, the use of light as an attractive inducing agent.The insight gained from a deeper understanding of the properties of cis-dichlororuthenium benzylidenes opened the doorway for the systematic development of new and efficient light-activated olefin metathesis catalysts and catalytic chromatic-orthogonal synthetic schemes. Following this, ways to disrupt the ligand-to-metal bond to accelerate the isomerization process that produced the active precatalyst were actively pursued. Thus, we summarize herein the original thermal activation experiments and how they brought about the discoveries of photoactivation in the sulfur-chelated benzylidene family of catalysts. The specific wavelengths of light that were used to dissociate the sulfur-ruthenium bond allowed us to develop noncommutative catalytic chromatic-orthogonal processes and to combine other photochemical reactions with photoinduced olefin metathesis, including using external light-absorbing molecules as "sunscreens" to achieve novel selectivities. Alteration of the ligand sphere, including modifications of the N-heterocyclic carbene (NHC) ligand and the introduction of cyclic alkyl amino carbene (CAAC) ligands, produced more efficient light-induced activity and special chemical selectivity. The use of electron-rich sulfoxides and, more prominently, phosphites as the agents that induce latency widened the spectrum of light-induced olefin metathesis reactions even further by expanding the colors of light that may now be used to activate the catalysts, which can be used in applications such as stereolithography and 3D printing of tough metathesis-derived polymers.

Key References

.[1]This work discloses the first photoactive sulfur-chelated olefin metathesis catalysts.

.[2]A concept work on guiding a metathesis-based reaction sequence outcome using light only.

.[3]Discovery of the photoactivation feature of the commercially available ruthenium phosphite complex cis-

.[4]A new generation of photoactive phosphite complexes for visible-light-induced metathesis and 3D printing applications.

Introduction

Since the emergence of well-defined transition metal catalysts, olefin metathesis has evolved to become a handy paintbrush for molecular artists in industry and academia.[5] With a plethora of options to form and reshape carbon–carbon double bonds, olefin metathesis is the method of choice for the synthesis of macrocycles[6] and heavy-duty polymeric materials[7] and even for the preparation of simple commodity chemicals from oil waste products and biomass.[8] The flexible framework of ruthenium-based complexes has opened up many research avenues for the development of task-specific catalysts,[9] such as catalysts embedded with asymmetric ligands for enantioselective olefin metathesis reactions,[10] catalysts that specialize in ring-closing metathesis (RCM) of macrocyclic structures,[11] catalysts for Z-selective[12] and stereoretentive[13] olefin metathesis reactions, and fast-initiating catalysts for ring-opening metathesis polymerization (ROMP).[14] Slow-initiating or latent metathesis catalysts are an important class of initiators that require the application of an external stimulus (heat, light, mechanical force, etc.) in order to promote metathesis reactions.[15] Particularly attractive is the use of light,[16] allowing stereolithographic 3D printing[17] and the development of chromatic-orthogonal processes.[18] The evolution of light-induced olefin metathesis from our group’s perspective will be unfolded in this Account, and its historic flow is briefly depicted as a roadmap in Scheme .

Scheme 1

Evolution of Light-Induced Olefin Metathesis Catalysts and Applications by the Lemcoff Research Group (Year of Publication in Parentheses)

Strategies in the Design of Latent Catalysts

The “magic bullet” of a latent metathesis catalyst would be a catalyst that remains inactive even when mixed with the most reactive of substrates but is able to efficiently catalyze the desired reaction when it is turned on in response to a specific external stimulus. Over the years a few approaches to slow down the initiation of olefin metathesis precatalysts have been introduced, and several latent catalysts have been reported.[15a,19] The main tactic usually perused is strengthening a metal–ligand bond to lower the initiation rate.[20] In many cases, this is accomplished by forming a bidentate chelate, which hinders dissociation. Figure presents a summary of the prevailing strategies followed for the creation of latent ruthenium-based initiators.

Figure 1

Models of latent ruthenium olefin metathesis catalysts. L = PCy3, NHC, CAAC; L1 = electron-donating group (P, N, O, S); E = O, NR, S; X, Y = anionic ligand.

Type I latent systems (Cazin catalysts)[21] combine a phosphite (P(OR)3) and a strongly σ-electron-donating ligand that are more stable in a cis geometry with respect to one another. This ligand configuration is inactive for olefin metathesis, and ligand dissociation is necessary for activation. In type II systems, a donor ligand is tethered to one of the anionic ligands. In this case the chelation decelerates both initiation and propagation,[22] and therefore, this layout is optimally used in acid-switchable systems.[23] In the case of type III systems, a chelating carbene ligand is used. This is the most common type of ligand chelation. Here, decoordination of L1 is entropically disfavored, and consequently, the initiation is slowed down.[24] Finally, type IV systems display chelation via the alkylidene and anionic ligands.[25] Another prominent strategy to impose latency is to substitute the carbene moiety with an electron-donating group (Fischer carbene, type V).[26]

The first latent chelated catalysts, 1, were reported by van der Schaaf et al. (Figure ).[27] Later, Grubbs replaced the phosphine ligand by an N-heterocyclic carbene (NHC) to produce complex 2. The 2-trans isomer would revert to the more stable 2-cis isomer, improving the latency of the precatalyst.[28] Other pioneering efforts in the area of chelating benzylidenes for latent ruthenium olefin metathesis precatalysts were reported by Slugovc, Grela, and their co-workers (precatalysts 3 and 4 and one of the first reported NHC-bearing precatalysts in a cis-dichloro configuration, 5).[29] With an original approach, Barbasiewicz et al. exploited the steric effects of the benzylidene sixth position to tune the latency in O-chelated precatalysts 6–8.[30]

Figure 2

Selected latent ruthenium metathesis catalysts bearing chelating alkylidene ligands.

Sulfur-Chelated Catalysts: Latent and Thermoswitchable

In the group’s first foray into this field, we envisioned that stronger ligation to the ruthenium center by a sulfur atom would generate a stable initiator.[31] Unexpectedly, the new sulfur-chelated catalyst (9-cis) was isolated in a cis-dichloro geometry (Scheme ). In this unusual configuration, the sulfur atom is not coordinated trans to the strongly σ-electron-donating NHC ligand; therefore, the trans influence is reduced, and the complex remains inactive at ambient temperatures.[32] The isomeric kinetic product of this reaction, 9-trans,[33] readily isomerized upon standing to the more stable cis-dichloro form (Scheme ).

Scheme 2

Synthesis of the First Sulfur-Chelated Ruthenium Benzylidene Complex, 9-cis

Scheme 3

Thermal Equilibrium between 9-trans and 9-cis

Catalyst 9-cis exhibited exceptional stability (∼20 days). In fact, heating a toluene solution of 9-cis for 3 days at 90 °C under air was required to significantly decompose the new catalyst. Its latency was first evaluated with the benchmark RCM reaction of diethyl diallylmalonate (DEDAM) (Figure ). No conversion was observed after 1 week at room temperature. RCM could be induced by heating the reaction mixture to 80 °C. The thermoswitchable behavior was demonstrated with intermittent periods of heating to 80 °C and cooling to 25 °C, and it was shown that RCM could be switched on and off even after more than 2 days.

Figure 3

RCM reaction products obtained using catalyst 9-cis. Reaction conditions: 0.1 M substrate in toluene, 90 °C, 2 days. Conversions were measured using GC–MS and 1H NMR spectroscopy.

9-cis was also tested for ROMP of dimethyl 5-norbornene-2,3-dicarboxylate.[34] As desired, no polymerization was observed at ambient temperature. Heating to 100 °C initiated polymerization, and an unexpected thermoswitchable behavior was observed, also after more than 3 days of heating–cooling cycles (Figure ).

Figure 4

Thermoswitchable behavior of 9-cis: 0.5 M dimethyl 5-norbornene-2,3-dicarboxylate in 1,2-dichloroethane with 0.5 mol % 9-cis in a sealed vessel. Monomer consumption was measured by GC–MS.

Interestingly, the polydispersity indexes (PDIs) of the polymers obtained were noticeably lower than expected from a slow-initiating catalyst. This could be explained by a “turnover-limited polymerization” mechanism, where the degree of polymerization (DP) is controlled by the turnover number of the catalytic species and not by the ratio of monomer to initiator, as is common in living polymerization mechanisms.

Photochemical Activation

Realizing that the isomerization process promotes activation of the sulfur-chelated precatalysts allowed us to imagine more efficient pathways for stimulation. On the basis of early reports on photoactive RuCl2(DMSO)4 complexes,[35] we hypothesized that the sulfur–ruthenium bond could be dissociated by applying UV irradiation. Initial attempts with catalyst 9-cis using a medium-pressure mercury UV-A lamp (100 W, 365 nm) were unsuccessful. However, addition of a suitable chromophore on the thioether increased the photosensitivity of the complex. Thus, complexes 10-cis, 11-cis, and 12-cis were designed and synthesized (Figure ).[1,36]

Figure 5

Sulfur-chelated olefin metathesis catalysts with aromatic groups on the thioether.

The photoactivity of the precatalysts was studied by irradiating a DEDAM solution in the presence of each catalyst. Indeed, the aromatic substitution on the thioether resulted in greatly increased efficiency for photoactivation, and 10-cis provided the best results with almost 90% conversion (Figure ).

Figure 6

UV-induced RCM of DEDAM using catalysts 9-cis, 10-cis, 11-cis, and 12-cis.

The suggested mechanism for photoactivation is shown in Scheme . Irradiation of 10-cis forms an unstable 14-electron cis transition state (10-cis14e) that quickly rearranges to the more stable 14-electron trans intermediate (10-trans14e), which was then capped by the sulfur atom to afford the stable metathesis-active 16-electron trans precatalyst (10-trans).

Scheme 4

Proposed Photoisomerization Mechanism

The switchable nature of the system was demonstrated by 15 min irradiation of a tetrachloroethane solution of DEDAM with 5 mol % catalyst 11-cis (activation), followed by 5 min heating at 80 °C (deactivation). In a counterintuitive fashion, heating the active trans isomer (out of equilibrium as a result of the irradiation) accelerates thermal isomerization back to the more stable cis form and reduces the amount of active precatalyst. Thus, the metathesis reaction is promoted by irradiation and actually stopped by heating for a short time (naturally, continued heating would provide a steady supply of active trans isomer to continue the reaction). The photoinduced metathesis activity of the most efficient complexes was further evaluated with several RCM and ROMP reactions (Figure and Table ).

Figure 7

Photoinduced RCM with 10-cis and 11-cis. Reaction conditions: 5 mol % catalyst and 0.1 M substrate in CH2Cl2; UV irradiation at 365 nm for 5 h. Conversions were determined by GC–MS.

Table 1

UV-Activated ROMP with Catalysts 10-cis and 11-cisa

Conversions were calculated using GC–MS after 24 h. Conditions: 0.5 M monomer in CH2Cl2; [monomer]/[cat.] = 300. Mn and PDI values were determined by triple-detector size-exclusion chromatography; Z/E values were determined by 1H NMR spectroscopy.

Although they showed improved performances compared with 9-cis, this new generation of thermally and photochemically switchable sulfur-chelated catalysts still suffered from some drawbacks. The complexes were still not active enough to accomplish metathesis reactions of “difficult” substrates, and even relatively simple reactions required high precatalyst loadings to reach completion. Moreover, highly reactive ROMP monomers such as norbornene (NB) and dicyclopentadiene (DCPD) were polymerized by 10-cis and 11-cis at ambient temperatures and in the dark (no latency). In view of these deficiencies, new systems were sought to widen the latency gap and improve the catalytic activity. Complexes 13-cis and 14-cis were designed to be extremely dormant precatalysts by using a three-point chelation. In contrast, complex 15-cis was designed to be a more efficient catalyst in its active form because of the strong electron-withdrawing influence of the trifluoromethyl substituent (Figure ).

Figure 8

Tridentate sulfur-chelated complexes 13-cis and 14-cis and highly efficient metathesis catalyst 15-cis.

Contrary to expectations, 13-cis did not form a three-point chelate with the ruthenium metal center. Indeed, its activity did not differ by much from those of the other sulfur-chelated catalysts previously reported. On the other hand, complex 14-cis was chelated by both sulfur atoms. As expected, this complex was extremely dormant, showing latency for DCPD at ambient temperatures; however, this strategy turned out to be a double-edged sword, as its metathesis activity was very poor (only 37% conversion in ROMP of DCPD after 3 h at 110 °C).

Catalyst 15-cis is a true success story. This catalyst was latent at room temperature but displayed much-superior activity when activated (see the comparison with catalyst 9-cis in Table ), especially in photoinduced reactions.

Table 2

Comparative RCM and ROMP Activities of 9-cis and 15-cisa

Conversions were measured using GC–MS.

Guiding Chemistry with Light

The extensive work of Bochet and co-workers on the development of chromatic-orthogonal systems,[18,37] i.e., systems in which two photochemical reactions can be executed commutatively using different colors of light, and the pioneering works on molecular photoswitches by Feringa and others[38] highlighted the elegance and simplicity of guiding a reaction outcome by using only light. The unique photoswitchable activity displayed by the sulfur-chelated metathesis catalysts provided the perfect opportunity to extend the concept of chromatic orthogonality to catalytic olefin metathesis reactions, asking the question “would it be possible to toggle between two different olefin metathesis products by using only light?”

Inspired by the studies of Schmidt and Nave on regioselective RCM reactions,[39] triene 16 was designed (Scheme ).[2] Accordingly, 16 could form a five-membered-ring or a six-membered-ring product depending on steric hindrance. Thus, a method was sought to promote a process that would result in significant steric volume differences by irradiation with a light source that would not induce olefin metathesis. Indeed, while the catalysts (10-cis or 11-cis) were sensitive to UV-A light, a bulky supersilyl group could be selectively photocleaved only with UV-C.[40] Thus, when the metathesis reaction was initiated in the presence of the bulky supersilyl group (path A, 350 nm followed by 254 nm), a five-membered ring (17) was preferentially formed. Alternatively, when the metathesis reaction was triggered after the cleavage of the supersilyl group (path B, 254 nm followed by 350 nm), RCM to give a six-membered ring (20) was favored.

Scheme 5

Regioselective Chromatic-Orthogonal Ring-Closing Metathesis

The inherent chromatic orthogonality displayed between the isomerization reaction of the sulfur-chelated catalysts (photoactivation) and the removal of supersilyl ethers was further exploited in the design of a catalyst with an embedded self-destruction function, 21-cis.[41] Although destruction of an active catalyst may seem to be a waste of a precious rare-metal catalyst, this function could be of great use in stereolithography, especially in layer-by-layer 3D printing processes. “Killing” the remaining active catalyst after polymerization and before applying a new layer of monomer/catalyst mixture may help in preventing loss of spatial resolution. In this design, the photosensitive supersilyl groups were directly installed on the NHC ligand (Figure , top). Thus, olefin metathesis was induced by the precatalyst upon exposure to 350 nm light, while exposure to 254 nm light led to its rapid decomposition (Figure , bottom).

Figure 9

(top) Sulfur-chelated photoswitchable catalyst 21-cis with a chromatic-orthogonal “kill-switch”. (bottom) Polymerization of a norbornene derivative in metal molds with 21-cis: (a) irradiation at 350 nm; (b) irradiation at 254 nm followed by UV-A. Reproduced with permission from ref (41). Copyright 2016 Wiley-VCH.

Sunscreen-Enabled Photoinduced Olefin Metathesis

The regioselective chromatic-orthogonal sequence used with triene 16 disclosed an unanticipated difficulty. The removal of the supersilyl protecting group was hampered because of the presence of the ruthenium catalyst (which strongly absorbs UV-C), and it took great skill to find the right conditions in order to achieve the desired chromatic-orthogonal sequence.

One of the major drawbacks of organic photochemical reactions is their poor selectivity. Often such reactions result in the formation of several products, and it is very difficult to control the selectivity for one product over the others. Thus, this “sunscreen” effect was harnessed to develop a methodology for the selective removal of photolabile protecting groups (PPGs).[42] The method relies on the difference between the molar absorption coefficients (at similar wavelengths) of a pair of PPGs and utilizes auxiliary strongly absorbing molecules that overlap the specific absorbance of the PPGs. Thus, 2-nitrobenzyl-protected alcohols were selectively photocleaved in the presence of strongly UV-C-absorbing phenanthrene, while supersilyl ethers remained unscathed (Scheme ). The UV filter could be applied as an additive or, more conveniently, as an external solution, bypassing the need for additional purification of the reaction products.

Scheme 6

Selective Photodeprotection of a 2-Nitrobenzyl PPG in the Presence of a Supersilyl Using Phenanthrene as a Molecular UV Filter

The scope of the sunscreen method was further studied with different types of photochemical reactions. The cross-metathesis product of allylic alcohols and acrylates can further react photochemically by two competing pathways to afford different organic scaffolds: butenolides and levulinates. The first reaction path involves a trans/cis double-bond isomerization followed by an intramolecular lactonization to produce the butenolide. The key step in the alternative pathway is a 1,5-hydrogen shift and subsequent tautomerization to form the levulinate (Scheme ).

Scheme 7

Divergent Photochemical Reactions to Form Butenolides and Levulinates

Under regular conditions, both types of products are formed with very little selectivity. However, when phenanthrene was added to the reaction mixture, the 1,5-hydrogen shift was completely blocked, allowing selective formation of butenolides. Without phenanthrene, the selectivity could be shifted to produce the levulinate esters. Thus, a protocol for a selective photochemical divergent two-step conversion of allylic alcohols and acrylates to butenolides and levulinates was successfully accomplished, as detailed in Scheme . Moreover, by the use of this photochemical protocol, the total synthesis of the marine natural product isocladospolide B could be achieved with very high efficiency (Scheme ).

Scheme 8

All-Photochemical Divergent Syntheses of Butenolides (25a–f) and Levulinates (26a–f)

Scheme 9

Synthesis of the Marine Product Isocladospolide B Using a Photochemical Divergent Protocol

Another challenge that could be addressed using the sunscreen methodology is the cross-metathesis reaction of 2-vinylphenols. When coordinated to the ruthenium catalyst, these substrates form a stable chelate and inhibit catalysis. Thus, we adopted a sunscreen effect solution to this problem. A UV-A PPG with low absorption could “survive” photoinduced activation of a ruthenium catalyst if a suitable molecular UV filter is procured. 2-Nitrobenzyl was selected to be the protecting group for the phenols on the basis of its relatively low absorption in the UV-A region. In this case, pyrene-1-carboxaldehyde, a dye with strong absorption in the UV-A and visible regions, was chosen as the UV filter. Thus, cross-metathesis of 2-nitrobenzyl-protected vinylphenol with methyl acrylate was investigated with catalyst 15-cis and 380 nm irradiation in the presence of an external solution of pyrene-1-carboxaldehyde at different concentrations (Figure ). As can be clearly observed, in the absence of pyrene-1-carboxaldehyde, the 2-nitrobenzyl PPG was removed and catalysis was inhibited. Increasing the dye concentration resulted in improved reaction conversions up to a concentration of 0.1 M, where the “protective” effect leveled off.

Figure 10

UV-filter-enabled cross-metathesis of 2-nitrobenzylvinylphenol and methyl acrylate with different concentrations of pyrene-1-carboxaldehyde as the UV filter.

This approach was exploited in an all-photochemical synthesis of coumarins. The first step was cross-metathesis between 2-nitrobenzyl-protected vinylphenols (32) and acrylates in the presence of pyrene-1-carboxaldehyde. This was followed by removal of the external filter solution and irradiation with 254 nm light to trigger a chain of three photochemical reactions: removal of the 2-nitrobenzyl PPG, then trans to cis double-bond isomerization, and finally cyclization to furnish the coumarins (Figure ).

Figure 11

(left) Two-step photochemical synthesis of coumarins via sunscreen-enabled olefin metathesis (isolated yields are shown). (right) Reaction setup of the sunscreen-assisted cross-metathesis. The inner vial contains the olefin metathesis mixture, and the external vial contains a 0.1 M pyrene-1-carboxaldehyde solution.

Exploring the Effects of Altering the N-Heterocyclic Carbene Ligand

The fine balance between the latency and augmented catalytic activity of 15-cis was the driving force for further investigation of the potential of SCF3-type chelated systems. The variation of the steric and electronic properties of NHC ligands is well-known to have a large impact on the activity of ruthenium metathesis catalysts.[43] For example, 1,3-bis(2,6-diisopropylphenyl)imidazolidin-2-ylidene (SIPr) (Scheme ) was found to increase the thermal stability and activity of metathesis complexes for terminal olefins.[44] On the other hand, NHC ligands with smaller volumes, such as 1,3-bis(2-methylphenyl)imidazolidin-2-ylidene (SITol) (Scheme ), are known to improve the RCM of sterically demanding substrates.[45]

Scheme 10

Synthesis of SCF3-Chelated Complexes with the NHC Ligands SIPr (34-cis) and SITol (35-cis)

Thus, the effect of different NHC ligands with varying steric bulk on the activity of sulfur-chelated switchable metathesis catalysts was also studied.[46]34-cis and 35-cis, bearing a SIPr and SITol ligand, respectively, were prepared, and their activities in olefin metathesis reactions were compared to that of 15-cis (Scheme ).

It is noteworthy that the synthesis of 34-cis required a long reaction time (3 weeks), as the steric congestion associated with the SIPr ligand significantly increased the energy barrier for trans–cis isomerization. With the catalysts in hand, the thermal and photochemical behavior toward a set of RCM reactions (disubstituted to tetrasubstituted) was studied. As expected, the catalyst with the bulkier NHC showed superior activity toward the RCM reaction of terminal diolefins, while the catalyst with the sterically lighter NHC promoted the photoinduced metathesis of tetrasubstituted RCM reactions in high yields. To date, 35-cis is the only photoswitchable catalyst that can efficiently promote tetrasubstituted RCM reactions (Scheme ).

Scheme 11

Photoinduced RCM of Substrates with Increasing Steric Bulk Using 15-cis, 34-cis, and 35-cis

Conversions were measured by 1H NMR spectroscopy.

In 2005 Bertrand and co-workers introduced cyclic alkyl amino carbene (CAAC) ligands as strong electron-donating ligands for transition metal catalysis.[47] The first ruthenium olefin metathesis catalysts bearing CAAC ligands were reported by Grubbs and Bertrand just 2 years later.[48] These complexes were found to be extremely stable and efficient catalysts[49] and also to prevent undesired olefin migration reactions at elevated temperatures.[50] More recently, a new generation of CAAC complexes was reported by Skowerski and Fogg showing exceptional catalytic efficiency and remarkable stability.[51] To examine the effect of CAAC ligands on sulfur-chelated complexes, four new sulfur-chelated complexes bearing CAAC ligands were prepared (Figure ).[52]

Figure 12

CAAC-bearing sulfur-chelated metathesis precatalysts.

Unexpectedly, the complexes prepared using a bulky 1,3-diisopropylphenyl CAAC ligand (36-trans and 37-trans) were isolated in the trans-dichloro configuration as the thermodynamically stable form. However, these complexes could not promote photoinduced olefin metathesis reactions and only reacted thermally with highly reactive monomers. Conversely, complexes 38-cis and 39-cis with reduced steric pressure from the CAAC ligand were found to be efficient metathesis catalysts (when activated). While complex 39-cis was inactive with light stimulation, 38-cis showed good activity with 350 and 419 nm light, completing a variety of polymerization reactions (Figure ).

Figure 13

Photocatalytic activity of 38-cis. Reaction conditions: 38-cis loading, 0.1% mol; monomer concentration, 0.5 M.

Photoactive Chelated Sulfoxide Catalysts

In an attempt to tune the reactivity of novel sulfur-chelated catalysts, our group as well as the group of Grela investigated the effect of oxidizing the sulfur atom to a sulfoxide.[53] However, given the much weaker electron donation of the oxidized sulfur, the chelated complexes obtained were more stable in the trans-dichloro configuration and thus were metathesis-active at ambient temperatures. We hypothesized that by adding a strongly electron-donating group on the chelating aryl sulfoxide, the electron density on the coordinating sulfur atom may force it to a cis-dichloro geometry to avoid the stronger trans influence from the NHC ligand.[54] Thus, styrene 40 was prepared and reacted with Grubbs’ third-generation catalyst to give 41-cis (Scheme , top).

Scheme 12

Photoactivity of 41-cis

41-cis was indeed latent toward RCM and polymerization reactions at room temperature and could be thermally activated at 80 °C in toluene. As sulfoxide ligands possess rich photochemistry,[35] the photoactivity of this latent complex was investigated. Interestingly, in toluene, the complex was activated with 254 nm light; unfortunately, the high-energy light also led to rapid decomposition of the catalyst and poor metathesis efficiency. When the solvent was replaced by dichloromethane, photoactivity was observed at 419 nm. However, despite (or maybe because of) the excellent latency of this catalyst, its catalytic activity was somewhat limited (Scheme , bottom).

Photoactivation of Ruthenium Phosphite Complexes for Olefin Metathesis

Ruthenium-based metathesis catalysts stabilized by phosphite ligands were first introduced by Cazin and co-workers in 2010.[21] By the reaction of an indenylidene complex (Ind-III) with triisopropyl phosphite, trans-Caz-1 was obtained. In boiling dichloromethane, this complex isomerized to the thermodynamically stable cis-dichloro configuration, known as cis-Caz-1 (Scheme ). cis-Caz-1 showed latency toward several olefin metathesis substrates at ambient temperature and could be switched on thermally in hot toluene (80 °C) to promote the catalysis of sterically demanding RCM reactions under air with minimal catalyst loadings.[55]

Scheme 13

Synthesis of the Catalyst cis-Caz-1

The resemblance of the isomerization behavior of cis-Caz-1 to that of the sulfur-chelated catalysts, combined with fact that this was not a chelated catalyst, drew our attention toward the possibility of disrupting the Ru–P bond by using light. Indeed, cis-Caz-1 was found to be readily activated under UV-A light (350 nm), becoming the first (and to date only) commercially available photoswitchable metathesis catalyst, making photoinduced olefin metathesis widely accessible.[3] The light-induced catalytic activity of cis-Caz-1 was thoroughly studied and showed great competence for a variety of ROMP, RCM, and CM reactions (Figure a). The study of ROMP of 1,5-cyclooctadiene (COD) using cis-Caz-1 disclosed an intriguing solvent effect (Figure b). When light-induced polymerization was carried out in dichloromethane, a cis-rich polymer was obtained and isolated (3:2 cis:trans ratio in the polymer). However, when the same reaction was carried out in toluene, the resulting polymer had the expected thermodynamic ratio of 4:1 in favor of trans.

Figure 14

(a) Photoinduced RCM and ROMP reactions with cis-Caz-1 at 350 nm. Dark control conversions are given in parentheses. RCM conditions: 0.1 M substrate and 1% mol catalyst in dichloromethane. ROMP conditions: 0.5 M monomer and 0.2 mol % catalyst in dichloromethane. (b) Solvent effect in photopolymerization of COD. The trans-rich polymer is obtained in toluene (top), and the cis-rich polymer is obtained in dichloromethane (bottom).

The catalytic activity of cis-Caz-1 was compared to that of the state-of-the-art photoactive sulfur-chelated metathesis precatalysts using a model RCM reaction (Figure ).

Figure 15

Comparison between sulfur-chelated photoswitchable catalysts 10-cis, 11-cis, 15-cis, and cis-Caz-1. Conversions were measured by 1H NMR spectroscopy.

When Cazin et al. tried to install phosphite ligands on Grubbs benzylidenes, only complexes in a trans-dichloro configuration were obtained.[56] As expected, these catalysts were metathesis-active at ambient temperatures. On the basis of a previous study on chelating phosphine ligands,[57] we surmised that a chelating benzylidene phosphite ligand would force a cis-dichloro configuration. Thus, 42 was synthesized and reacted with the Hoveyda–Grubbs second-generation catalyst to afford complex 43-cis (Scheme ).[3]

Scheme 14

Synthesis of 43-cis

The cis-dichloro configuration of 43 was verified by NMR spectroscopy and single-crystal X-ray diffraction. This catalyst could be activated thermally by heating to 80 °C in toluene or photochemically using a 405 nm light-emitting diode (LED) in toluene to complete several RCM and ROMP reactions; unfortunately, the strong chelation also impeded a high catalytic activity, and metathesis reactions using this catalyst required long irradiation times and high catalyst loadings (Figure ).

Figure 16

Catalytic activity of 43-cis with a 405 nm LED.

In order to further our research on photoactive phosphite-containing ruthenium alkylidenes, the effect of the phosphite ligands was studied by changing the type of phosphite ligand used. Thus, two benzyl phosphite ligands, 44 and 45, were prepared. Notably, ligand 45 was designed to provide an alternative chromatic-orthogonal kill switch. Phosphites 44 and 45 were reacted with Grubbs third-generation complex to give complexes 46-cis and 47-cis, respectively (Scheme ).[4] To our satisfaction and great surprise, both complexes adopted the latent cis-dichloro geometry.

Scheme 15

Synthesis of Complexes 46-cis and 47-cis

The photochemical activities of these catalysts were benchmarked by RCM of DEDAM (Figure ). Catalyst 46-cis showed activity at all wavelengths tested but was most efficient with UV-A and visible light (420 nm). On the other hand, catalyst 47-cis was best activated with 420 nm light and surprisingly even with blue LED light (450–470 nm), but it was much less active under UV light because of the photosensitive nature of the 2-nitrobenzyl phosphite ligand, which led to catalyst decomposition (as expected).

Figure 17

Photoactivities of 46-cis and 47-cis at different wavelengths.

The photoinduced olefin metathesis scope of 46-cis and 47-cis was studied in 420 nm light (Table ). High catalytic efficiency under visible light and good latency with minimal reaction conversion in the dark control experiments were observed.

Table 3

Photoinduced Olefin Metathesis by Complexes 46-cis and 47-cis with 420 nm Lighta

Reaction conditions: RCM: 0.1 M substrate in 0.45 mL of toluene-d8 and 0.05 mL of catalyst stock solution in CD2Cl2 were irradiated for 8 h. ROMP: 0.5 M monomer in 0.45 mL of toluene-d8 and 0.05 mL of catalyst stock solution in CD2Cl2 were irradiated for 1 h.

Dark control experiments were covered with aluminum foil.

Not determined because of polymer insolubility.

47-cis was designed with a chromatic-orthogonal self-destruct switch, similar to that of 21-cis.[41] The kill switch in this case is based on the photosensitivity of the 2-nitrobenzyl phosphite ligand, and irradiation of 47-cis with UV-C light led to rapid degradation of the catalyst. However, 47-cis has additional advantages compared with 21-cis. First, the activation with visible light is more efficient, and another important advantage is the straightforward and facile synthesis of the catalyst.[58] The efficiency of the kill switch was demonstrated with an RCM reaction. Two identical reaction mixtures were exposed to irradiation at 254 nm and then at 420 nm. During the exposure to UV-C light, the control experiment was covered with aluminum foil and afforded an 86% yield of the RCM product after exposure to visible light (Figure a). On the other hand, no conversion was observed in the experiment that was exposed to the UV-C light (Figure b). The kill switch was also tested in the ROMP of cyclooctene (Figure c,d). A monomer/precatalyst mixture was placed in metal molds and subjected to UV-C followed by UV-A irradiation or just to UV-A irradiation. While the experiment that was exposed to visible light yielded a functional and elastic polymer film, the experiment that was exposed first to UV-C yielded a sticky amorphous gel.

Figure 18

“Kill switch” efficiency of catalyst 47-cis. (a, b) RCM of N,N-diallyltosylsulfonamide: (a) control experiment, in which the reaction mixture was covered during UV-C exposure; (b) the reaction mixture was exposed to UV-C, resulting in catalyst decomposition. (c) Photopolymerization of cyclooctene under irradiation at 420 and 254 nm. (d) the metal mold used for making the polymer films.

3D printing of metathesis-based polymeric materials using stereolithography is an attractive application that is making its first steps toward industrial implementation (e.g., polySpectra, a company founded by Raymond Weitekamp, specializes in additive manufacturing using olefin metathesis technology).[59] This is due to the excellent mechanical properties of metathesis-based polymers, such as poly-DCPD.[7a,7c,60] Currently, however, the most desirable monomers are often very reactive, and even highly latent catalysts fail to form processable and durable monomer/precatalyst liquid mixtures. Notably, Moore and co-workers added phosphites to mixtures of Grubbs second-generation catalyst and DCPD to create stable catalyst/monomer formulations with long “pot life” values.[7b,7d,61] Inspired by this work, we anticipated that the latent benzylidene phosphite catalysts would also form stable formulations with DCPD, as they already bear the “retarding” phosphite ligand. Indeed, mixing 46-cis (0.02 or 0.03 mol %) with DCPD resulted in a stable formulation that could be processed for more than 4 h at room temperature (and longer times at lower temperatures). These DCPD/catalyst formulations were then used for light-induced stereolithography using a digital light processing (DLP) projector (385 nm) for short exposure times (30–60 s per layer), exploiting the excellent spatial resolution obtained by the DLP projector and the short reaction times (Figure ). This demonstrated the excellent suitability of photoswitchable latent phosphite complexes for these type of applications.

Figure 19

Light-induced stereolithography using catalyst 46-cis. (a) Four layers: (i) full circle; (ii) three-fourth of a circle; (iii) half a circle; (iv) one fourth of a circle. (b) Ben-Gurion University of the Negev logo printed in two layers. The logo is used with permission from Ben-Gurion University of the Negev. (c) Pyramids consisting of three layers of decreasing size. (d) Zoom in on one pyramid from (c).

Outlook and Conclusions

This Account has surveyed our group’s activity in the design and development of latent photoswitchable metathesis catalysts during the past 12 years and their applications in photoinduced polymerizations and selective light-triggered olefin metathesis reactions. The serendipitous finding that sulfur chelation leads to latency by changing the complex’s geometry allowed advances in the area of light-induced olefin metathesis by selective disruption of the chelating bond with light. Since the first S-chelated photoswitchable catalyst report in 2009, the field of light-induced metathesis has grown significantly. Our own contributions, which have produced several novel ruthenium photoinduced catalysts, brought about new research in chromatic-orthogonal sequences, photochemical divergent syntheses, efficient 3D printing of metathesis polymers, and the use of a wide light spectrum ranging from UV-C to visible light. Future work will include studies of new activation methodologies (IR light, dual-wavelength activation) and achieving novel photoinduced selectivity by changing the anionic ligands and using new NHC ligands. Indeed, it has been a colorful journey, and this summary recounts the path taken toward developing light activation of olefin metathesis and hopefully will inspire future research toward the rainbow of opportunities that await in photochemical metathesis.