Visible-Light-Controlled Ruthenium-Catalyzed Olefin Metathesis

Abstract

Olefin metathesis is now one of the most efficient ways to create new carbon – carbon bonds. While most efforts focused on the development of ever-more efficient catalysts, a particular attention has recently been devoted to developing latent metathesis catalysts, inactive species that need an external stimulus to become active. This furnishes an increased control over the reaction which is crucial for applications in materials science. Here, we report our work on the development of a new system to achieve visible-light-controlled metathesis by merging olefin metathesis and photoredox catalysis. The combination of a ruthenium metathesis catalyst bearing two N-heterocyclic carbenes with an oxidizing pyrylium photocatalyst affords excellent and spatial resolution using only visible light as stimulus. Applications of this system in synthesis, as well as in polymer patterning and photolithography with spatially-resolved ROMP, are described.

Graphical Abstract

Olefin metathesis is one of the most attractive and powerful tools for the creation of carbon – carbon π bonds, finding numerous applications in synthetic chemistry, fine chemical synthesis and materials science.^1,2 While most synthetic efforts have been devoted to the development of ever-more efficient catalysts, increased attention has been paid to the development of catalysts that can be activated/deactivated on demand.³ Such latent catalysts are dormant species under ambient conditions and require an external stimulus to become active. Increased control on reactions is crucial not only from an understanding viewpoint but also for applications in materials science for the production of new well-defined materials.⁴ Various stimuli have been exploited to achieve such control in metathesis reactions, including heat, light, ultrasound, acid and redox switches.⁵ Light is arguably the most convenient and attractive stimulus since it is non-invasive, can be easily manipulated and provides the opportunity for high temporal and spatial resolution (Figure 1a).⁶ As a consequence, several recent reports have described light-promoted olefin metathesis.^5c,5d,7 While these have been important developments, they are dominated by UV light with most reports describing activation rather than gating control of alkene metathesis.

Figure 1.

(a) Temporal and spatial control in catalysis. (b) Bis(NHC)-ruthenium complexes as latent catalysts using visible light (this work).

We considered that the merger of olefin metathesis with photoredox catalysis⁸ could lead to visible light control of alkene metathesis. Visible light photoredox catalysis has already proven successful for metal-free olefin metathesis polymerization via a radical mechanism.⁹ In particular, excitation of the appropriate photocatalyst by visible-light irradiation should permit the activation of a latent metathesis catalyst, most probably by inducing ligand dissociation,¹⁰ and therefore lead to the development of an on-demand metathesis system. Importantly, the use of visible light is more convenient than UV light while still providing high levels of temporal and spatial resolution. Overall, the development of such a system would open new perspectives in photolithography^11,12 and in materials science for the design of new materials,¹³ as already illustrated by the impact of recent work reported for photo-controlled, living radical polymerizations.¹⁴

At the outset of these studies, we needed a ruthenium-based complex that is inactive at ambient temperature, and identified bis-NHC ligated Ru complexes first introduced by Herrmann.¹⁵ When substituted with aromatic groups on the nitrogen atoms, these catalysts lack activity for metathesis at room temperature, most probably because of the difficult dissociation of one NHC ligand to generate the corresponding 14-electron active catalyst.^15,16 At higher temperatures, the activity of these catalysts is restored. In this regard, we surmised that the NHC dissociation event could be promoted at room temperature by using photoredox catalysis. A carefully chosen photocatalyst should be capable, after excitation upon irradiation with visible light, of activating these catalysts and therefore toggling them into their corresponding active species after dissociation of one NHC (Figure 1b).

To test our hypothesis, we first evaluated the benchmark ring closing metathesis (RCM) of diethyl diallylmalonate using RuCl₂(CHPh)(IMes)₂ and RuCl₂(CHPh)(SIMes)₂, previously reported by Fogg¹⁷ and Grubbs,¹⁶ in the presence of different photocatalysts under visible-light irradiation. After screening several photocatalysts and reaction conditions (see Supporting Information for further details), we found that a combination of RuCl₂(CHPh)(IMes)₂ (Ru₁) and 2,4,6-triphenylpyrylium tetrafluoroborate (TPPT) as photocatalyst gives the desired product in 87% yield after 4h of irradiation under blue LEDs at room temperature (Table 1, entry 9). While screening photocatalysts, we observed that only highly oxidizing ones such as acridinium and pyrylium derivatives provide some reactivity (entries 6–8), while no product is observed when switching to less oxidizing photocatalysts (entries 1–5). This is consistent with an activation mode involving oxidation of the Ru catalyst followed by dissociation of one NHC to generate the catalytically active species forming the corresponding radical cation.¹⁸ We indeed note that Ru₁ has two distinct oxidation events as identified by cyclic voltammetry, with the first occurring at +0.44 V, likely corresponding to the generation of the radical cation by a metal-centered oxidation (see SI). While all photoredox catalysts should allow oxidation to the radical cation, the dissociation event might be caused by a second oxidation occurring at one NHC ligand that would only be promoted by highly oxidizing photocatalysts and explain that traditional Ir and Ru photocatalysts are not effective (see table 1, entries 1–5).¹⁹ Importantly, no reaction is observed in the absence of ruthenium, light or photocatalyst (entries 10–12). The lack of reactivity under light without photocatalyst also rules out a mechanism solely based on photo-induced dissociation of one NHC ligand and highlights the importance of the photoredox system. Finally, the use of RuCl₂(CHPh)(SIMes)₂ (RU₂) delivers similar reactivity (entry 13). However, background reactivity is observed in the absence of light and photocatalyst (entry 14), indicating that dissociation of one NHC happens slowly at ambient temperature. RuCl₂(CHPh)(IMes)₂ (Ru₁) was chosen as it displays optimal latent behavior.

Table 1.

Reaction optimization and scope of RCM, CM, and ROCM reactions.

^[a]All optimization reactions were conducted on a 0.1 mmol scale.

^[b]Determined by ¹H NMR spectroscopy using 1,2-dibromoethane as an internal standard.

^[c]Conditions: substrate (0.2 mmol), RuCl₂(CHPh)(IM es)₂ (2 mol%), TPPT (3 mol%), CH₂Cl₂ (0.2M), rt, blue LEDs, 4h.

^[d]4 mol% of TPPT.

^[e]Left substrate (0.2 mmol), right substrate (0.4 mmol).

^[f]Left substrate (0.2 mmol), right substrate (0.6 mmol).

^[g]For additional samples, see SI.

With an efficient system in hand, we first explored its ability to promote different types of metathesis reactions. While standard metathesis reactions can be readily promoted using this photoredox catalytic system, as illustrated with representative examples in Table 1 and SI, we were more interested in interrogating ring-opening metathesis polymerization (ROMP) applications. To this end, several monomers such as norbornene derivatives 1–8,11, norbornadiene 9, 1,5-cyclooctadiene 10 and dicyclopentadiene 12 could be readily polymerized within 1h under blue LED irradiation in the presence of RuCl₂(CHPh)(IMes)₂ and TPPT (Table 2). Molecular weights (M_n) obtained after polymerization of monomers 1–4 are significantly higher than the expected values which suggests that polymerization is faster than catalyst initiation. Dispersities were found in the range of 1.63 to 1.88. Monomers 5–10 are also smoothly polymerized within an hour of irradiation but lead to insoluble polymers, which precludes GPC analysis. Finally, cross-linking monomers 11 and 12 could also be efficiently polymerized to afford complete gelation within an hour, the latter only requiring 0.01 mol% of Ru₁, 0.05 mol% of TPPT and 15 minutes of irradiation. Importantly, the latency is successfully maintained with dicyclopentadiene 12 since, in the absence of light, less than 5% polymerization is observed after 24h (5% after 3 days, 9% after a week). When stopped after 90 seconds under light, 16% polymerization is observed. The rate of polymerization under visible light can therefore be estimated to be 12,000 times faster than in the dark.

Table 2.

Scope of ROMP reactions.

^[a]Conditions: monomer (0.2 mmol), RuCl₂(CHPh)(IMes)₂ (0.5 mol%), TPPT (1 mol %), CD₂Cl₂ (0.2 M), rt, blue LEDs, 1h.

^[b]Determined by ¹H NMR spectroscopy using mesitylene as internal standard.

^[c]Determined by GPC.

^[d]Using IMes₂RuCl₂ CHPh (0.01 mol%), TPPT (0.05 mol%) for 15 min under blue LEDs.

Further experiments were conducted to examine the influence of light and to probe our ability to exert temporal and spatial controls over the reaction. First, temporal control was evaluated by conducting on/off experiments with alternating periods of irradiation and darkness for the ring closing metathesis of diethyl diallylmalonate. The ability to exert temporal control over a reaction is of great interest for the design of orthogonal multicomponent reactions, as well as for the development of new systems designed to produce new highly functionalized materials. As can be seen in Figure 2, temporal control can be achieved since maximal reactivity was obtained during irradiation whereas darkness only afforded minimal increases in yields (from 0 to 3%).

Figure 2.

Temporal control over the RCM of diethyl diallylmalonate and corresponding proposed mechanism.

A series of experiments (described in the SI) lead us to suggest the following mechanism for the on/off behavior enabled by photoredox catalysis and light irradiation.^9a It is commonly accepted that Ru catalysts mediate olefin metathesis via a coordinatively unsaturated Ru(II) intermediate such as II (Figure 2). Given that only highly oxidizing excited state photocatalysts provide appreciable yield (Table 1), we propose that ligand loss occurs at ambient temperature from an oxidized Ru intermediate, potentially at the IMes moiety to give active metathesis catalyst II and reduced pyrylium V as well as VI. The latter two can combine to form VII, analogous to adducts reported lacking substitution at the 4 position.²⁰ Release of the IMes provides a pool of free ligand which can coordinate IV and arrest catalysis.²¹

We also interrogated our ability to exert spatial control over metathesis with this system, due to the potential applications in materials science with polymer patterning, 3D printing and photolithography. In this regard, the development of a system controlled by visible light appears especially attractive and convenient. To this end, dicyclopentadiene 12, and some other monomers,²² were first irradiated with visible light (blue Kessil lamp, 40 W) in the presence of RuCl₂(CHPh)(IMes)₂ and TPPT through different photomasks in order to produce macroscopic polymers with controlled geometric patterns. Removal of the masks and unreacted monomers nicely affords the corresponding patterned polymers in short irradiation times (15–60 minutes) and with minimal bleeding in the unexposed areas (Figure 3a–c).²³ Interestingly, the thickness of these patterned polymers can be easily controlled by tuning the irradiation time (see SI). Finally, an important feature of this system is its practicality and user-friendliness. While the excited state TPPT* is modestly sensitive to oxygen, the photomask patterning experiments can be performed with minimal precautions of placing the monomer/catalyst mixture under a blanket of inert gas.

Figure 3.

Polymer patterning and photolithographic olefin metathesis polymerization (PLOMP) using visible light.

Higher resolutions are required in order to apply this visible-light-controlled system for applications in PhotoLithographic Olefin Metathesis Polymerization (PLOMP). While most photolithographic techniques are based on the use of high resolution photomasks, an attractive alternative is the use of high resolution light sources, such as lasers, which should provide a straightforward way to reach pinpoint resolution and find new applications in photolithography.²⁴ As proof of concept, we could successfully induce similar patterning from dicyclopentadiene solutions using a simple blue laser pointer (200 mV). In these cases, the patterns are directly and conveniently “drawn” from the bulk solution in a few minutes, either manually (Figure 3d,e) or using an orbital shaker providing constant movement (Figure 3f).

The two afore-described techniques allow the convenient fabrication of macroscopic patterned polymers through spatially-resolved ROMP promoted by visible light and without the need for grafting of the monomers. As for microscopic patterning, we also demonstrate the efficiency of our system for PLOMP applications.¹² Although photolithography is now a commonly used technique in microfabrication, such systems based on olefin metathesis are still rare. To this end, 1 cm × 1 cm silicon wafers were first pre-functionalized with a norbornene unit to ensure grafting of the growing polymer onto the surface.^12a Those pre-functionalized silicon wafers were then used as support to perform the spatially-resolved polymerization of norbornadiene on a microscale by simply irradiating a solution of the monomer, RuCl₂(CHPh)(IMes)₂ and TPPT in dichloromethane with a regular blue LED light bulb (blue Kessil lamp, 40 W) through high resolution photomasks (Figure 3g).

After developing in dichloromethane, patterns of poly(norbomadiene) with resolutions down to 30–40 microns could be successfully printed over silicon wafers within 10 minutes of irradiation (Figure 3h,i). These results are accomplished using readily available visible light sources, and are complementary to Fourkas’ positive photoresist,^12a as well as Grubbs’ negative photoresist system,^12b both of which use UV light, while giving access to similar resolutions.

In conclusion, we have reported the development of an efficient and user-friendly system for olefin metathesis controlled by visible light. The combination of a latent bis(NHC)-ruthenium complex and an oxidizing pyrylium photocatalyst efficiently promotes olefin metathesis under visible light, providing high levels of temporal and spatial resolution. In particular, applications in polymer patterning and photolithographic olefin metathesis polymerization are also demonstrated. Further studies to broaden the scope and applications of this system, as well as to gain more insights into its mechanism, are currently ongoing in our group.