A Photoswitchable Olefin Metathesis Catalyst - ACS Publications

Jan 3, 2017 - Department of Chemistry, The University of Texas at Austin, Austin, Texas ... Institute for Basic Science (IBS), Ulsan 44919, Republic o...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/Organometallics

A Photoswitchable Olefin Metathesis Catalyst Aaron J. Teator,†,‡ Huiling Shao,§ Gang Lu,§ Peng Liu,*,§ and Christopher W. Bielawski*,‡,∥ †

Department of Chemistry, The University of Texas at Austin, Austin, Texas 78712, United States Center for Multidimensional Carbon Materials (CMCM), Institute for Basic Science (IBS), Ulsan 44919, Republic of Korea § Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States ∥ Department of Chemistry and Department of Energy Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan 44919, Republic of Korea ‡

S Supporting Information *

ABSTRACT: A dithienylethene-functionalized N-heterocyclic carbene-Ru(II) complex was synthesized and found to undergo a reversible photoisomerization which influenced its intrinsic catalytic activity. UV-induced ring-closure enhanced the rate of ringclosing metathesis reactions (kclosed/kopened = 1.4−1.7) and attenuated the rate of ring-opening metathesis polymerizations (kclosed/kopened = 0.56−0.66). Visible light irradiation promoted cycloreversion and restored the initial activity. The ability to switch between the isomeric states of the catalyst was also utilized to modulate the rate of ongoing olefin metathesis reactions via photoirradiation. A computational investigation revealed how steric and electronic effects separately influence the transition states adopted by each form of the catalyst and afforded activation energies that were in agreement with the relative reaction rate constants determined by experiment.

■

INTRODUCTION Historically, improvements to existing homogeneous catalysts have focused on enhancing intrinsic catalytic properties, such as activity, selectivity, and/or functional group tolerance. Thus, once chosen, a catalyst typically performs with a certain outcome for a given transformation. An ability to “switch” the performance displayed by a catalyst over the course of a reaction could facilitate access to products with uniquely tailored structures or properties that vary in response to changing environments. Indeed, such enhanced control over reaction pathways has fueled the field of switchable catalysis, whereby stimuli responsive functional groups are incorporated into known catalysts.1−6 Light represents an ideal stimulus for the modulation of chemical reactivity due to its excellent spatio-temporal resolution and the ability to achieve selective chromophore excitation in a noninvasive manner. In recent years, a variety of light-regulated catalytic methods have been disclosed, yet, despite the substantial interest, few examples have been reported that involve photomodulating the activities displayed by transition-metal-based catalysts.7−12 Thus far, a majority of the reported systems comprise cycling a reaction between “on” (i.e., active) and “off” (i.e., inactive) states by altering the steric environment around the metal center. However, restricting the © XXXX American Chemical Society

reaction such that product formation is only possible in one of the kinetic states limits versatility and may impede advanced switching functions (e.g., alternate substrate selection, ondemand chemoselective variation, etc.) As such, we sought to develop a methodology that enabled catalyst modulation by photochemically tuning the electronics of a ligand coordinated to a catalytically active metal center.3 We13−15 and others16−20 have demonstrated that the photoinduced electrocyclization of an N-heterocyclic carbene (NHC) bearing an annulated photochromic21 dithienylethene (DTE)22−28 unit significantly altered the donating ability of the ligand. For example, a Rh-based hydroboration catalyst featuring such an NHC was found to exhibit activity that was dependent on the isomeric state of the DTE unit.29 Building on those results, we envisioned substituting the prototypical NHC (i.e., SIMes30) in the Hoveyda−Grubbs second-generation catalyst for a DTE-functionalized NHC to photomodulate ringclosing metathesis (RCM) and ring-opening metathesis polymerization (ROMP) reactions31−33 via changes in NHC donicity. Herein, we report the first photoswitchable olefin metathesis catalyst34,35 that can be reversibly toggled between Received: December 10, 2016

A

DOI: 10.1021/acs.organomet.6b00913 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics two states using light and demonstrate the remote modulation of its intrinsic activity.

■

RESULTS AND DISCUSSION As shown in Scheme 1, the synthesis of the Hoveyda−Grubbstype catalyst 3o was achieved in a single step from a Scheme 1. Synthesis and Photochromism of Complex 3o

Figure 1. (a) UV−vis spectral changes of 3o in C6H6 upon UV irradiation (λ = 313 nm). (b) UV−vis spectrum of the photostationary state (PSS) reached after UV irradiation of 3o for 420 s, and the observed spectral changes of the PSS upon visible light irradiation (λ > 500 nm). The arrows point to the evolution of spectral changes over time ([3o]0 = 3.0 × 10−5 M).

photoswitchable NHC (1). Treatment of a hexane suspension of 1, as its free carbene,15 with commercially available complex 2 at 60 °C for 4 h, followed by filtration and trituration with diethyl ether, afforded pure 3o in 95% yield as a crystalline solid (see the SI). The complex 3o exhibited a diagnostic downfield 1 H NMR signal at δ 16.7 ppm (CD2Cl2) that was indicative of a benzylidene moiety and in the expected36 range. Further evidence was provided by the observation of the 13C NMR signal assigned to the former carbene center of 1 at δ 177.0 ppm (CD2Cl2) which resonated upfield (Δδ = −42 ppm) relative to the free carbene, consistent with coordination to a metal center. A benzylidene-type resonance was also observed at δ 293.8 ppm (CD2Cl2) and found to be comparable to previously reported complexes.36−38 The UV−vis spectrum recorded for 3o in CH2Cl2 exhibited an intense absorption band centered at 298 nm which was assigned to a combination of the n → π* and π → π* transitions of the N-heterocycle and aryl rings, respectively, as well as a prominent metal-to-ligand charge-transfer (MLCT) absorption band centered at 380 nm. Exposure of this solution ([3o]0 = 3.0 × 10−5 M) to UV radiation (λ = 313 nm) resulted in a gradual change of the reaction mixture from pale green to deep blue-green. Concomitant with this color change was a decrease in the intensity of the absorption band centered at 298 nm along with the appearance of new absorption bands centered at 453 and 639 nm, consistent with the formation of an extended π-conjugated system and ring-closure to form 3c (Figure 1a). After 7 min of UV irradiation, the spectral changes subsided as a photostationary state (PSS) was reached, and reflected a 68% conversion of 3o to its ring-closed isomer 3c.39,40 The broad low-energy absorption bands diminished upon subsequent exposure to visible light (λ > 500 nm) along with decoloration of the solution, and the UV−vis spectrum of 3o was nearly completely restored (91%) after 9 min (Figure 1b). The observation of an isosbestic point at 318 nm, in the forward as well as in the reverse cyclization processes, indicated that 3o underwent photoisomerization without significant side product formation. The photoisomerization of 3o was further quantified by monitoring the chemical shift of the 1H NMR benzylidene resonance upon ring-closure to form 3c. Subjecting a C6D6 solution of 3o ([3o]0 = 1.0 × 10−3 M) in a quartz NMR tube to UV irradiation (λ = 313 nm) for 120 min resulted in a color change analogous to that observed at higher dilution, consistent with photocyclization to 3c. As shown in Figure 2, 1H NMR

Figure 2. A series of 1H NMR spectra recorded in C6D6 during the following experiment: A sample of (a) 3o was (b) exposed to UV radiation (λ = 313 nm) for 120 min and (c) subsequently exposed to visible radiation (λ > 500 nm) for 60 min ([3o]0 = 1.0 × 10−3 M).

analysis showed the growth of an upfield-shifted resonance (δ 16.66 ppm) relative to 3o (δ 16.97 ppm).41 Integration of the two signals revealed that 80% of 3o had been converted to 3c.42 Subsequent exposure of the UV-irradiated solution to visible light (λ > 500 nm) resulted in diminished intensity of the benzylidene resonance at δ 16.66 ppm concomitant with increasing intensity of the benzylidene resonance at δ 16.97 ppm. After 60 min of irradiation, the 1H NMR spectrum of the sample matched that of 3o, which suggested to us that the cycloreversion of 3c to 3o was complete.43,44 Having demonstrated that 3o undergoes reversible photocyclization with high fidelity, the ability to use the catalyst to modulate olefin metathesis transformations was explored. Initial efforts were focused on RCM, as such reactions are common activity benchmarks for metathesis-active complexes.45 In a preliminary experiment, the introduction of 3o to a C6D6 solution of diethyl diallylmalonate (4) ([4]0 = 1.4 × 10−2 M, [4]0/[3]0 = 100) resulted in the formation of the ring-closed product 5. 1H NMR spectroscopic analysis revealed that the reaction proceeded with a pseudo-first-order rate constant ko of 2.1 ± 0.2 × 10−4 s−1, and reached 74% conversion after 120 min (Figure 3a). To determine if the UV-induced cyclization alters catalyst activity, a solution of 3o ([3o]0 = 1.0 × 10−3 M) was first subjected to UV irradiation (λ = 313 nm) for 120 min to generate 3c, then utilized in the reaction described above and monitored over time. Inspection of the corresponding 1H NMR data indicated that the reaction proceeded to 85% conversion B

DOI: 10.1021/acs.organomet.6b00913 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics

Figure 3. (a) Plot of −ln[4] versus time of the RCM of 4 to disubstituted olefin 5. (b) Plot of −ln[6] versus time of the RCM of 6 to trisubstituted olefin 7. (c) Plot of −ln[4] versus time of the RCM of 4 as catalyzed by 3c. After 30 min, the reaction was irradiated with visible light (λ > 500 nm) for 20 min and monitored for an additional 40 min. (d) Plot of −ln[COD] versus time of the ROMP of COD to polybutadiene. (e) Plot of −ln[8] versus time of the ROMP of 8 to 9. (f) Plot of −ln[COD] versus time for the ROMP of COD as catalyzed by 3c. Additional COD (300 equiv rel. to 3) was added after 20 min (indicated). The reaction was then irradiated with visible light (λ > 500 nm) for 20 min and monitored for an additional 20 min. The reactions were catalyzed by either 3o (red ■) or 3c (blue ■) and were monitored over time by 1H NMR spectroscopy (C6D6). [4]0 = [6]0 = 1.4 × 10−2 M. [COD]0 = [8]0 = 4.3 × 10−2 M. [4 or 6]0/[3]0 = 100. [COD or 8]0/[3]0 = 300.

Examination of the data revealed that photoirradiation diminished the rate of the reaction (ko = 1.5 ± 0.1 × 10−4 s−1) and the ratio of the respective rate constants (kc/ko = 2.1) was consistent with that measured from reactions catalyzed independently by each form of 3. Building on these results, subsequent efforts were shifted toward photoswitching ROMP reactions. In an initial test, a C6D6 solution of 1,5-cis,cis-cyclooctadiene (COD) was treated with 3o ([COD]0 = 4.3 × 10−2 M, [COD]0/[3]0 = 300) and the reaction was monitored over time using 1H NMR spectroscopy. High conversion to polybutadiene (>95%) was observed after 30 min (Figure 3d). After an initial induction period, the polymerization proceeded with pseudo-first-order kinetics and a rate constant ko of 1.7 ± 0.1 × 10−3 s−1 was measured. Conversely, when freshly prepared 3c (vide supra) was utilized in lieu of 3o, the polymerization exhibited a relatively extended induction period and proceeded with a lower rate constant kc of 1.12 ± 0.03 × 10−3 s−1 (kc/ko = 0.66). The relative activities displayed by 3o and 3c were opposite to those observed in the aforementioned RCM reactions, which indicated that the performance of the catalyst may be substratedependent. To test this hypothesis, the ROMP of a different monomer (8) was explored ([8]0 = 4.3 × 10−2 M, [8]0/[3]0 = 300). The addition of 3o to a C6D6 solution of 8 resulted in the polymerization of 8 to form 9 (Figure 3e). After an induction period, the polymerization proceeded to >90% conversion over the course of 50 min. Pseudo-first-order kinetics were observed,

over the course of 120 min with a pseudo-first-order rate constant kc of 3.0 ± 0.1 × 10−4 s−1.46 The observed rate difference (kc/ko = 1.4) reflected a difference in activity between the ring-opened and ring-closed catalysts, whereby 3c facilitated the RCM reaction at a faster rate and afforded a higher conversion over the same period of time.47 To determine if substrate sterics would influence catalytic activity, an analogous set of experiments utilizing diethyl allyl methallylmalonate (6) to generate a trisubstituted RCM product (7) were performed (Figure 3b).45 In accordance with the aforementioned conclusions, the reaction proceeded to higher conversion (95%) and at a faster rate with 3c (kc = 1.5 ± 0.1 × 10−4 s−1) when compared to an analogous reaction performed using 3o (ko = 0.9 ± 0.1 × 10−4 s−1; 85% conversion) under otherwise identical conditions. The ratio of rate constants measured using each form of the catalyst (kc/ko = 1.7) was larger than that measured for the RCM of 4.48 The modulation of the activity displayed by 3 during an ongoing RCM reaction was then explored. Given the apparent sensitivity of the catalytically active intermediate to UV radiation,49 efforts were directed toward switching an ongoing reaction that was initiated with 3c. Initially, 3c was generated in situ as described above and then injected into a C6D6 solution of 4. As summarized in Figure 3c, the corresponding RCM reaction was monitored for 30 min (kc = 3.1 ± 0.2 × 10−4 s−1), at which point the mixture was exposed to visible light (λ > 500 nm) for 20 min to promote ring-opening of the NHC ligand. C

DOI: 10.1021/acs.organomet.6b00913 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics and a rate constant ko of 2.5 ± 0.2 × 10−3 s−1 was calculated. Performing the analogous experiment with freshly prepared 3c resulted in a significant decrease in the reaction rate (kc = 1.4 ± 0.1 × 10−3 s−1, kc/ko = 0.56).50 Next, an ability to “switch” the rate of an ongoing polymerization reaction was explored (Figure 3f). In order to minimize initiation effects on the polymerization rate constant determinations, a C6D6 solution of COD was polymerized in the presence of freshly prepared 3c for 15 min ([COD]0 = 4.3 × 10−2 M, [COD]0/[3]0 = 300), at which point additional COD (300 equiv rel. to 3o) was added.51 After 5 min, the mixture was exposed to visible light (λ > 500 nm) for 20 min. Comparison of the reaction rate constant measured after irradiation (ko = 1.8 ± 0.1 × 10−3 s−1) to that recorded prior to visible light exposure (kc = 1.2 ± 0.1 × 10−3 s−1) reflected an enhancement (kc/ko = 0.67) that was in agreement with previous results. Efforts were also directed toward ascertaining how the ringopened and ring-closed forms of the catalyst influence the number-average molecular weight (Mn) and distributions (Đ) of the polymers produced. In order to minimize secondary metathesis reactions, ROMP reactions initiated by 3o or 3c were quenched prior to complete monomer conversion. Initially, a mixture containing 3o and 8 ([8]0 = 4.3 × 10−2 M, [8]0/[3o]0 = 300) in C6H6 was poured into cold methanol after 5 min of reaction time and the precipitated polymer was examined by gel permeation chromatography. A relatively high molecular weight material (Mn = 141 kDa, Đ = 1.5) was obtained in 36% yield. A polymer prepared from an analogous polymerization performed with 3c was found to have a similar molecular weight (Mn = 120 kDa, Đ = 1.6) but was obtained in lower yield (24%). The yields were in agreement with the monomer conversions measured by 1H NMR spectroscopy for the ROMP of 8 as initiated by 3o (43%) or 3c (28%) over the same periods of time and under similar conditions.52 The origins of the observed rate constant differences described above were elucidated through a comprehensive computational investigation using density functional theory (DFT) (see the SI for computational details and the complete reaction energy profiles). As summarized in Figure 4a, the reaction energy profile of the RCM of a model substrate, 1,6heptadiene,53 revealed that the rate-determining step is the retro-[2 + 2] cycloaddition (TS1) to form the cyclopentene product and that the catalyst resting state is the ruthenacyclobutane intermediate 11.54,55 The calculations also indicated that the reaction employing the ring-closed ligand 1c required a slightly lower barrier for the retro-[2 + 2] cycloaddition than the analogous reaction with the ring-opened form (1o). While the computed activation energies appear to have underestimated the rate constant difference (kc,theory/ko,theory = 1.2) compared to that measured by NMR spectroscopy (kc/ko = 1.4−1.7), both theory and experiment showed that the RCM reaction was relatively more promoted by the ring-closed ligand 1c. Since the computed transition state geometries exhibit similar steric interactions between the NHC ligand and the substrate for both forms of the catalyst, the observed rate constant difference was attributed to an electronic effect. In the ring-opened form, the NHC ligand is a stronger donor13,15 and stabilizes the Ru(IV) metalacyclobutane intermediate, which results in a higher barrier for the subsequent retro-[2 + 2] cycloaddition. Similarly, as summarized in Figure 4b, the retro-[2 + 2] cycloaddition was also calculated to be the rate-determining

Figure 4. Computed activation Gibbs free energies of the (a) RCM of 1,6-heptadiene, (b) ROMP of COD, and (c) ROMP of norbornene as catalyzed by ruthenium complexed with ring-opened (1o) or ringclosed (1c) photoswitchable ligands. Calculations were performed at the M06/6-311+G(d,p)-SDD/SMD(benzene)//B3LYP/6-31G(d)SDD level of theory.

step during the ROMP of COD.54−56 In this case, the calculated activation energies indicate that the ring-opened ligand (1o) would lower the barrier required for the retro-[2 + 2] cycloaddition relative to an analogous reaction involving the ring-closed form of the ligand (1c). The calculated reactivity difference is opposite to that observed in the RCM reactions. In the ROMP of COD, steric interactions with the monomer dominate the difference in reactivity displayed by each form of the catalyst. Upon photocyclization, the NHC backbone undergoes planarization which, in turn, forces the N-mesityl substituents further into the coordination sphere of the ruthenium center. The increased steric bulk in the ring-closed form results in an unfavorable steric interaction with the D

DOI: 10.1021/acs.organomet.6b00913 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics

rate of 0.8 mL/min. Detection was performed using a Viscotek Triple Detector Array (TDA305). Absolute molecular weight and polydispersity data were calculated using the following averaged dn/dc values obtained via GPC: poly(COD), 0.1025 mL/g, and poly(8), 0.1159 mL/g. UV−vis spectra were acquired using an Agilent Cary 100 UV−vis Spectrometer in 6Q Spectrosil quartz cuvettes (Starna) with 1.0 cm path lengths and 3.5 mL nominal sample solution volumes. Beer’s law measurements were performed using 10, 20, 30, 40, and 50 μM sample concentrations. The photochemical reactions were performed in quartz cuvettes, a Spectrosil quartz NMR tube, or a quartz low-pressure vacuum (LPV) NMR tube. The irradiation source for photochemical reactions was a Newport/Oriel 66942 200−500 W Hg arc lamp instrument equipped with a 350 W Hg lamp, a Newport 6117 liquid filter, a Newport 71445 electronic safety shutter, and a Newport 71260 filter holder. The source was powered by a Newport 669910 power supply and mounted on a Newport XL48 optical rail with a Newport 13950 shielded cuvette holder placed at a distance of 8 cm from the end of the source. The irradiation wavelength for the photocyclization reactions was obtained using a 313 nm band-pass filter (Andover Corporation). A long-pass edge filter (>500 nm) (Andover Corporation) was used to introduce visible light. (4,5-Bis(2′-methyl-5′-phenylthien-3′-yl)-1,3-bis(2,4,6trimethylphenyl)-imidazolylidene)Cl2RuCH-o-OiPrC6H4 (3o). This compound was synthesized by modifying a literature procedure.58 Under an atmosphere of N2 inside a glovebox, an 8 mL Teflon-capped vial equipped with a stir bar was charged with 50 mg (0.077 mmol) of 1, 48.6 mg (0.081 mmol) of Hoveyda−Grubbs first generation catalyst, and 3.0 mL of hexanes. The resulting suspension was heated to 60 °C in the sealed vial and stirred vigorously, during which time a gradual color change from red-brown to deep green was observed. After 4 h, the reaction was cooled to room temperature and the solids were collected on a glass frit, washed with Et2O (3 × 5 mL), and dried under high vacuum to afford 71 mg (95% yield) of the desired product as a deep green solid. mp 233 °C (dec.) 1H NMR (400 MHz, CD2Cl2): δ 1.33 (d, 3JH−H = 6.1, 6H, −CH3), 1.99 (s, 6H), 2.03 (br. s, 6H), 2.44 (s, 6H), 2.61 (br. s, 6H), 4.94 (septet, 3JH−H = 6.1, 1H, −CH−), 6.88−6.90 (m, 3H), 6.95−7.02 (m, 2H), 7.10 (dd, 3J1,H−H = 7.59, 3J2,H−H = 1.66, 2H, H−Ar), 7.24−7.37 (m, 12H), 7.58−7.62 (m, 1H), 16.73 (s, 1H). (400 MHz, C6D6): δ 1.41 (d, 3JH−H = 6.1, 6H, −CH3), 1.97 (s, 6H), 2.13 (s, 6H), 2.38 (br. s, 12H), 5.51 (m, 1H), 6.37 (d, 3JH−H = 8.3, 2H, H−Ar), 6.69 (t, 3JH−H = 7.4, 2H, H−Ar), 6.94−6.98 (m, 4H), 7.03−7.07 (m, 4H), 7.09 (br. s, 2H), 7.23 (m, 2H), 7.32−7.34 (m, 4H), 16.97 (s, 1H). 13C{1H} NMR (100.6 MHz, CD2Cl2): δ 14.45, 19.86 (br.), 21.34, 21.44, 75.69, 113.45, 122.21, 122.94, 123.66, 125.65, 126.01, 127.83, 129.29, 129.48 (br.), 130.57, 134.23, 134.76 (br.), 139.02 (br.), 139.66, 140.31, 140.67, 146.03, 152.58, 176.99, 293.86 (d, J = 14.6). HRMS (ESI) for C53H52Cl2N2S2ORu: [M]+: Calcd. 968.1942, Found: 968.1922. Anal. Calcd. for C53H52Cl2N2S2ORu: C, 65.69; H, 5.41; N, 2.89; S, 6.62; Found: C, 65.96; H, 5.53; N, 2.87; S, 6.97. Photocyclized Ru-Complex (3c). Method A. Under an atmosphere of N2 inside a glovebox, a 1.0 mM solution of 3o in C6H6 or CH2Cl2 was transferred to quartz cuvettes in 4.0 mL portions, capped, removed from the glovebox, and subjected to UV radiation (λIrr = 313 nm) for 120 min. After the cuvettes were returned to the glovebox, the 4 mL portions were combined and the solvent was removed under reduced pressure to afford a mixture of 3c and 3o (relative ratio = 75:25), as determined by 1H NMR spectroscopy (CD2Cl2). Method B. Under an atmosphere of N2 inside a glovebox, a Spectrosil quartz NMR tube was charged with 0.7 mL of a 1.0 mM stock solution of 3o in C6D6. The tube was then subjected to UV radiation (λIrr = 313 nm) for 120 min, coupled with vigorous shaking every 30 min to ensure proper mixing, to afford a mixture of 3c and 3o (relative ratio = 80:20), as determined by 1H NMR spectroscopy (C6D6). 1H NMR (400 MHz, CD2Cl2): δ 1.31 (d, 3JH−H = 4.0, 6H, −CH3), 2.26 (s, 6H), 2.37 (s, 6H), 2.42 (s, 6H), 2.54 (s, 6H), 5.00 (septet, 3JH−H = 8.0, 1H, −CH−), 5.26 (br. s, 2H), 6.87−7.01 (br. m, 7H), 7.15 (br. s, 4H), 7.22 (m, 6H), 7.25−7.36 (br. m, 12H), 7.58−7.63 (m, 2H), 16.44 (s, 1H), 16.73 (s, 0.3H). 1H NMR (400 MHz, C6D6): δ 1.33 (d, 3JH−H = 6.1, 6H, −CH3), 2.25 (s, 6H), 2.56 (s, 6H), 2.60 (br. s, 12H), 5.48 (s,

propagating chain, which hinders the retro-[2 + 2] cycloaddition and thus attenuates the rate of the ROMP. Indeed, the calculations predicted that the ring-closed catalyst should facilitate the ROMP of COD at a relatively slower rate (kc,theory/ ko,theory = 0.15), although the rate difference was overestimated when compared to experimental data (kc /k o = 0.66). Conversely, examination of the reaction energy profile of the ROMP of norbornene revealed the [2 + 2]-cycloaddition to be the rate-determining step (Figure 4c).54−56 Since the steric effects of the ring-opened and ring-closed forms of the NHC ligand are similar during the [2 + 2]-cycloaddition step with norbornene, we concluded that an electronic effect controls the rate of the reaction. The NHC ligand is relatively more donating in the open form and promotes the [2 + 2]cycloaddition due to stabilization of the ruthenacyclobutane intermediate, and thus results in a faster reaction. In agreement with the experimental data, the calculations indicated that the ring-opened form of the catalyst should facilitate the polymerization of a norbornene derivative at a faster rate than its ringclosed analogue. In summary, we have developed the first olefin metathesis catalyst bearing a photochromic DTE-functionalized NHC ligand. The complex 3o was found to undergo electrocyclic isomerization to its ring-closed analogue 3c upon exposure to UV radiation, a process that was successfully reversed using visible light. The ring-opened and ring-closed forms of the catalyst were both found to be metathesis-active, and their respective activities were dictated by the isomeric state of the NHC ligand. The ring-closed 3c was found to facilitate RCM reactions at a faster rate than 3o, while the ring-opened 3o was observed to be the more active ROMP catalyst. Furthermore, the cycloreversion of 3c to form 3o was used to modulate the rates of ongoing RCM and ROMP reactions. Computed reaction energy profiles supported the observed rate constant differences and indicated an electronic effect dominated during RCM and the ROMP of 8, while steric interactions governed the ROMP of COD. Future efforts may be directed toward the development of more sterically hindered photoswitchable DTE-functionalized NHCs with relatively bulky N-substituents and/or 4-substituted thiophene units as calculations predict that such scaffolds may further enhance catalytic activity.

■

EXPERIMENTAL SECTION

General Considerations. Unless otherwise specified, reagents were purchased from commercial sources and used without further purification. The compounds 4,5-bis-(2′-methyl-5′-phenylthien-3′-yl)1,3-bis(2,4,6-trimethylphenyl)-imidazolylidene (1)15 and diethyl allyl methallylmalonate (6)57 were prepared according to literature procedures. All syntheses were performed under an inert atmosphere of nitrogen unless specified otherwise. Solvents were dried and degassed using a Vacuum Atmospheres Company solvent purification system. NMR spectra were recorded using a Bruker 400 MHz spectrometer. Chemical shifts δ (in ppm) are referenced to tetramethylsilane using the residual solvent as an internal standard (1H and 13C) or using the unified scale relative to the absolute frequency for 1H of 0.1% TMS in CDCl3 (31P). For 1H NMR: C6D6, 7.16 ppm; CD2Cl2, 5.32 ppm. For 13C NMR: C6D6, 128.06 ppm; CD2Cl2, 53.84 ppm. Coupling constants (J) are expressed in Hertz (Hz). High resolution mass spectra (HRMS) were obtained with a Waters Xevo G2-XS Q-ToF (ESI) mass spectrometer. Elemental analyses were performed with a ThermoScientific Flash 2000 Organic Elemental Analyzer. Gel permeation chromatography (GPC) was performed on a Viscotek GPCmax Solvent/Sample Module. Two fluorinated polystyrene columns (IMBMW-3078) were used in series and maintained at 35 °C. THF was used as the mobile phase at a flow E

DOI: 10.1021/acs.organomet.6b00913 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics

utilize 3c, 0.5 mL of catalyst stock solution was first added to a Spectrosil quartz NMR tube and subjected to UV irradiation (λIrr = 313 nm) for 120 min, and 0.1 mL of this mixture was subsequently injected into the tube. ROMP of COD. Under an atmosphere of N2 inside a glovebox, 43 μL (0.073 mmol) of 1,5-cis,cis-cyclooctadiene (COD) was diluted to a total volume of 1.47 mL with C6D6 to afford a 0.05 M stock solution of the monomer. A 1.0 mM stock solution of the catalyst was prepared by dissolving 2.8 mg (0.003 mmol) of 3o in 2.89 mL of C6D6. A portion (0.6 mL) of the monomer stock solution was added to an NMR tube fitted with a septum screw cap, and the tube was placed into an NMR spectrometer equilibrated at 293 K. After 5 min, the tube was ejected from the spectrometer, 0.1 mL of the catalyst stock solution was injected, and the tube was turned upside down twice and then returned to the spectrometer. The polymerization was subsequently monitored by 1H NMR spectroscopy over time. For experiments that utilize 3c, 0.5 mL of catalyst stock solution was first added to a Spectrosil quartz NMR tube and subjected to UV irradiation (λIrr = 313 nm) for 120 min, and 0.1 mL of this mixture was subsequently injected into the tube. For the photoswitching experiments, a single reaction was set up as described above and irradiated with visible light (λIrr > 500 nm) after a certain period of time. ROMP of 8. Under an atmosphere of N2 inside a glovebox, 20.2 mg (0.073 mmol) of norbornene imide 8 was diluted to a total volume of 1.47 mL with C6D6 to afford a 0.05 M stock solution of the monomer. A 1.0 mM stock solution of the catalyst was prepared by dissolving 2.8 mg (0.003 mmol) of 3o in 2.89 mL of C6D6. A portion (0.6 mL) of the monomer stock solution was added to an NMR tube fitted with a septum screw cap, and the NMR tube was placed into an NMR spectrometer equilibrated at 293 K. After 5 min, the tube was ejected from the spectrometer, 0.1 mL of catalyst stock solution was injected, and the tube was turned upside down twice and then returned to the spectrometer. The polymerization was subsequently monitored by 1H NMR spectroscopy over time. For experiments that utilize 3c, 0.5 mL of catalyst stock solution was first added to a Spectrosil quartz NMR tube and subjected to UV irradiation (λIrr = 313 nm) for 120 min, and 0.1 mL of this mixture was subsequently injected into the tube. Recovery of 3o after Photoswitching. Using a microbalance, 0.168 mg (0.001 mmol) of 1,3,5-trimethoxybenzene and 0.969 mg (0.001 mmol) of 3o were added to a dry 8 mL vial. Under an atmosphere of N2 inside a glovebox, 1.0 mL of C6D6 was added to generate a 1.0 mM stock solution containing 1,3,5-trimethoxybenzene and 3o. A Spectrosil quartz NMR tube was charged with 0.6 mL of the stock solution, and the tube was subjected to UV irradiation (λIrr = 313 nm) for 120 min, coupled with vigorous shaking every 30 min to ensure proper mixing. The tube was then subjected to visible light irradiation (λIrr > 500 nm) for 60 min while being vigorously shaken every 30 min to ensure proper mixing. The reaction was monitored over time using 1H NMR spectroscopy. Inspection of the 1H NMR data, along with integration of the resonance at δ 6.26 ppm (set to 3H) for reference, revealed a 99% recovery of 3o.

2H), 5.51 (m, 1H), 6.34 (d, 3JH−H = 8.2, 2H, H−Ar), 6.63 (t, 3JH−H = 7.4, 2H, H−Ar), 6.89 (br. m, 6H), 6.97 (br. m, 5H), 7.07 (br. m, 4H), 7.29 (d, 3JH−H = 7.3, 4H, H−Ar), 7.33 (m, 0.9H), 16.66 (s, 1H), 16.97 (s, 0.3H). 13C{1H} NMR (100.6 MHz, CD2Cl2): δ 67.00, 209.61, 296.79 (d, J = 10.9). UV−vis (C6H6): λmax = 287 nm, λmax = 369 nm, λmax = 453 nm (ε = 17230 dm3 mol−1), λmax = 639 nm (ε = 16665 dm3 mol−1). RCM of Diethyl Diallylmalonate (4). Under an atmosphere of N2 inside a glovebox, 6.0 μL (0.025 mmol) of diethyl diallylmalonate (4) was diluted to a total volume of 1.50 mL with C6D6 to afford a 16.67 mM stock solution of the substrate. A 1.0 mM stock solution of the catalyst was prepared by dissolving 2.8 mg (0.003 mmol) of 3o in 2.89 mL of C6D6. A portion (0.6 mL) of the substrate stock solution was added to an NMR tube fitted with a septum screw cap, and the tube was placed into an NMR spectrometer equilibrated at 293 K. After 5 min, the tube was ejected from the spectrometer, 0.1 mL of the catalyst stock solution was injected, and the tube was turned upside down twice and then returned to the spectrometer. The reaction was subsequently monitored by 1H NMR spectroscopy over time. For experiments that utilize 3c, 0.5 mL of catalyst stock solution was first added to a Spectrosil quartz NMR tube and subjected to UV irradiation (λIrr = 313 nm) for 120 min, and 0.1 mL of this mixture was subsequently injected into the tube. For the photoswitching experiments, a single reaction was set up as described above and irradiated with visible light (λIrr > 500 nm) after a certain period of time. RCM of Diethyl Allyl Methallylmalonate (6). Under an atmosphere of N2 inside a glovebox, 10.6 mg (0.042 mmol) of diethyl allyl methallylmalonate (6) was diluted to a total volume of 2.50 mL with C6D6 to afford a 16.67 mM stock solution of the substrate. A 1.0 mM stock solution of the catalyst was prepared by dissolving 2.8 mg (0.003 mmol) of 3o in 2.89 mL of C6D6. A portion (0.6 mL) of the substrate stock solution was added to an NMR tube fitted with a septum screw cap, and the tube was placed into an NMR spectrometer equilibrated at 293 K. After 5 min, the tube was ejected from the spectrometer, 0.1 mL of the catalyst stock solution was injected, and the tube was turned upside down twice and then returned to the spectrometer. The reaction was subsequently monitored by 1H NMR spectroscopy over time. For experiments that utilize 3c, 0.5 mL of catalyst stock solution was first added to a Spectrosil quartz NMR tube and subjected to UV irradiation (λIrr = 313 nm) for 120 min, and 0.1 mL of this mixture was subsequently injected into the tube. RCM of Diallyl Sulfide. Under an atmosphere of N2 inside a glovebox, 4.3 μL (0.033 mmol) of diallyl sulfide was diluted to a total volume of 2.00 mL with C6D6 to afford a 16.67 mM stock solution of the substrate. A 1.0 mM stock solution of the catalyst was prepared by dissolving 2.8 mg (0.003 mmol) of 3o in 2.89 mL of C6D6. A portion (0.6 mL) of the substrate stock solution was added to an NMR tube fitted with a septum screw cap, and the tube was placed into an NMR spectrometer equilibrated at 293 K. After 5 min, the tube was ejected from the spectrometer, 0.1 mL of the catalyst stock solution was injected, and the tube was turned upside down twice and then returned to the spectrometer. The reaction was subsequently monitored by 1H NMR spectroscopy over time. For experiments that utilize 3c, 0.5 mL of catalyst stock solution was first added to a Spectrosil quartz NMR tube and subjected to UV irradiation (λIrr = 313 nm) for 120 min, and 0.1 mL of this mixture was subsequently injected into the tube. RCM of 1,6-Heptadiene. Under an atmosphere of N2 inside a glovebox, 10.8 mg (0.112 mmol) of 1,6-heptadiene was diluted to a total volume of 6.74 mL with C6D6 to afford a 16.67 mM stock solution of the substrate. A 1.0 mM stock solution of the catalyst was prepared by dissolving 2.8 mg (0.003 mmol) of 3o in 2.89 mL of C6D6. A portion (0.6 mL) of the substrate stock solution was added to an NMR tube fitted with a septum screw cap, and the tube was placed into an NMR spectrometer equilibrated at 293 K. After 5 min, the tube was ejected from the spectrometer, 0.1 mL of the catalyst stock solution was injected, and the tube was turned upside down twice and then returned to the spectrometer. The reaction was subsequently monitored by 1H NMR spectroscopy over time. For experiments that

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.organomet.6b00913. Additional ring-opening metathesis polymerization, X-ray crystallographic and kinetics data, 1H and 13C NMR spectra, and computational details (PDF) Cartesian coordinates of calculated structures (XYZ) Crystallographic data for 3o (CIF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (P.L.). *E-mail: [email protected] (C.W.B.). F

DOI: 10.1021/acs.organomet.6b00913 Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics ORCID

(30) SIMes = 1,3-bis(2,4,6-trimethylphenyl)-4,5-dihydroimidazol-2ylidene. (31) Vougioukalakis, G. C.; Grubbs, R. H. Chem. Rev. 2010, 110, 1746. (32) Olszewski, T. K.; Bieniek, M.; Skowerski, K.; Grela, K. Synlett 2013, 24, 903. (33) Herbert, M. B.; Grubbs, R. H. Angew. Chem., Int. Ed. 2015, 54, 5018. (34) For examples of switchable olefin metathesis catalysts that undergo photochemical initiation, see: Ben-Asuly, A.; Aharoni, A.; Diesendruck, C. E.; Vidavsky, Y.; Goldberg, I.; Straub, B. F.; Lemcoff, N. G. Organometallics 2009, 28, 4652. Diesendruck, C. E.; Iliashevsky, O.; Ben-Asuly, A.; Goldberg, I.; Lemcoff, N. G. Macromol. Symp. 2010, 293, 33. Sashuk, V.; Danylyuk, O. Chem.Eur. J. 2016, 22, 6528. (35) For recent reports of latent, including photoinitiated, olefin metathesis catalysts, see: Keitz, B. K.; Grubbs, R. H. J. Am. Chem. Soc. 2009, 131, 2038. Monsaert, S.; Vila, A. L.; Drozdzak, R.; Van Der Voort, P.; Verpoort, F. Chem. Soc. Rev. 2009, 38, 3360. Khalimon, A. Y.; Leitao, E. M.; Piers, W. E. Organometallics 2012, 31, 5634. Weitekamp, R. A.; Atwater, H. A.; Grubbs, R. H. J. Am. Chem. Soc. 2013, 135, 16817. Naumann, S.; Buchmeiser, M. R. Macromol. Rapid Commun. 2014, 35, 682. (36) Garber, S. B.; Kingsbury, J. S.; Gray, B. L.; Hoveyda, A. H. J. Am. Chem. Soc. 2000, 122, 8168. (37) Gessler, S.; Randl, S.; Blechert, S. Tetrahedron Lett. 2000, 41, 9973. (38) Berlin, J. M.; Campbell, K.; Ritter, T.; Funk, T. W.; Chlenov, A.; Grubbs, R. H. Org. Lett. 2007, 9, 1339. (39) The conversions of the forward and reverse photoinduced isomerizations were determined using the molar extinction coefficient of the absorbance of 3c at 639 nm (ε = 16665 dm3 mol−1), which was determined by a combination of 1H NMR and UV−vis spectroscopy. (40) Gauglitz, G. In Photochromism: Molecules and Systems; Durr, H., Bouas-Laurent, H., Eds.; Elsevier Science: Amsterdam, 2003; Vol. 1, p 15. (41) Consistent with the formation of 3c, 13C NMR analysis (CD2Cl2) also revealed downfield-shifted resonances that corresponded to the benzylidene (δ 296.8 ppm) and the former carbene center (δ 209.6 ppm) as well as the appearance of a new signal at δ 67.0 ppm which was assigned to the newly formed sp3 carbon nuclei adjacent to the sulfur atoms (see the SI). (42) This value represents the conversion of 3o to 3c after 120 min ([3o]0 = 1.0 × 10−3 M). The conversion of 3c to 3o via exposure of ongoing reactions to visible light was conducted at a lower catalyst concentration ([3]0 = 1.4 × 10−4 M). (43) 1H NMR analysis indicated a 99% recovery of 3o after conversion to 3c and subsequent visible light irradiation. Furthermore,