Here, we present a protocol for active site validation of metal-organic framework catalysts by comparing stoichiometric and catalytic carbonyl-ene reactions to find out whether a reaction takes place on the inner or outer surface of metal-organic frameworks.
Substrate size discrimination by the pore size and homogeneity of the chiral environment at the reaction sites are important issues in the validation of the reaction site in metal-organic framework (MOF)–based catalysts in an enantioselective catalytic reaction system. Therefore, a method of validating the reaction site of MOF-based catalysts is necessary to investigate this issue. Substrate size discrimination by pore size was accomplished by comparing the substrate size versus the reaction rate in two different types of carbonyl-ene reactions with two kinds of MOFs. The MOF catalysts were used to compare the performance of the two reaction types (Zn-mediated stoichiometric and Ti-catalyzed carbonyl-ene reactions) in two different media. Using the proposed method, it was observed that the entire MOF crystal participated in the reaction, and the interior of the crystal pore played an important role in exerting chiral control when the reaction was stoichiometric. Homogeneity of the chiral environment of MOF catalysts was established by the size control method for a particle used in the Zn-mediated stoichiometric reaction system. The protocol proposed for the catalytic reaction revealed that the reaction mainly occurred on the catalyst surface regardless of the substrate size, which reveals the actual reaction sites in MOF-based heterogeneous catalysts. This method for reaction site validation of MOF catalysts suggests various considerations for developing heterogeneous enantioselective MOF catalysts.
MOFs are considered a useful heterogeneous catalyst for chemical reactions. There are many different reported uses of MOFs for enantioselective catalysis1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19. Still, it has yet to be determined whether the reactions take place on the inner or outer surface of the MOFs. Recent studies have raised questions concerning the utilization of the available surface and reduced diffusion20,21,22,23. A more striking issue is that the chiral environment varies with the location of each cavity in the MOF crystal. This heterogeneity of the chiral environment implies that the stereoselectivity of the reaction product depends on the reaction site24. Thus, designing an efficient enantioselective catalyst requires identification of the location where the reaction would take place. To do so, it is necessary to ensure that the reaction occurs either only on the inner surface or only on the outer surface of the MOF while leaving the interior intact. The porous structure of MOFs and their large surface area containing chiral environment active sites can be exploited for enantioselective catalysis. For this reason, MOFs are excellent replacements of solid-supported heterogeneous catalysts25. The use of MOFs as heterogeneous catalysts needs to be reconsidered if the reaction does not occur inside them. The location of the reaction site is important, as well as the size of the cavity. In porous materials, the size of the cavity determines the substrate based on its size. There are some reports of MOF-based catalysts that overlook the cavity size issue25. Many MOF-based catalysts introduce bulky catalytic species (e.g., Ti(O-iPr)4) to the original framework structure3,8,13. There is a change in the cavity size when bulky catalytic species are adopted in the original framework structure. The reduced cavity size caused by the bulky catalytic species makes it impossible for the substrate to fully diffuse into the MOFs. Thus, discrimination of substrate size by the cavity size of the MOFs needs to be considered for these cases. The catalytic reactions by MOFs often make it difficult to support evidence of reactions taking place inside the MOF cavity. Some studies have shown that substrates larger than the MOF cavities are converted to the expected products with ease, which seems contradictory8,13. These results can be interpreted as a contact between the functional group of the substrate and catalytic site initiating the catalytic reaction. In this case, there is no need for the substrate to diffuse into the MOFs; the reaction occurs on the surface of the MOF crystals26 and the cavity size is not directly involved in the discrimination of the substrate based on its size.
To identify the reaction sites of MOFs, a known Lewis-acid promoted carbonyl-ene reaction was selected2. Using 3-methylgeranial and its congeners as substrates, four types of enantioselective carbonyl-ene reactions (Figure 1) were studied27. The reactions, which have been previously reported, were classified into two classes: a stoichiometric reaction using a Zn reagent and catalytic reactions using a Ti reagent27. The reaction of the smallest substrate requires a stoichiometric amount of Zn/KUMOF-1 (KUMOF = Korea University Metal-Organic Framework); it has been reported that this reaction takes place inside of the crystal27. Two kinds of MOFs were used in this method, Zn/KUMOF-1 for the stoichiometric reaction and Ti/KUMOF-1 for the catalytic reaction. Owing to the distinct reaction mechanisms of these two kinds of MOFs, a comparison between the reaction rate versus substrate size is possible2,28,29. The effect of particle size on the carbonyl-ene reaction with Zn/KUMOF-127 demonstrated that, as seen in the previous report, the chiral environment of the outer surface was different from the inner side of the MOF crystal24. This article demonstrates a method that determines the reaction sites by comparing the reactions of three kinds of substrates with two classes of catalysts and the effect of particle size as reported in the previous paper27.
1. Preparation of (S)-KUMOF-1 crystals in three sizes
NOTE: Each step follows the experimental section and supplementary information of previous reports2,24,27. Three different sizes of (S)-KUMOF-1 were prepared: large (S)-KUMOF-1-(L), medium (S)-KUMOF-1-(M), and small (S)-KUMOF-1-(S) with particle sizes >100 μm, >20 μm, and <1 μm, respectively. When out of the solvent, (S)-KUMOF-1 dismantles. Therefore, the crystals should always be kept wet while in use.
2. Preparation of Zn/(S)-KUMOF-1 in three sizes
NOTE: Each step follows the experimental section and supplementary information of previous reports2,24,27.
3. Preparation of Ti/(S)-KUMOF-1 in three sizes
NOTE: Each step follows the experimental section and supplementary information of previous reports2,24,27.
4. Carbonyl-ene reaction using the prepared MOFs
NOTE: Prepare a series of substrates according to the method described in our previous report27. All three substrates are used individually in each carbonyl-ene reaction except for the particle size effect determination, in which only the smallest substrate (1a) is used27. Each step follows the experimental section and supplementary information of previous reports2,24,27.
The enantioselective carbonyl-ene reaction using the Zn reagent is stoichiometric because of the difference in the binding affinities of the alkoxy and carbonyl groups to the metal (Figure 2). For this reason, the substrates were converted into the products at the reaction site and remained there. The desired products were obtained by dismantling the crystals, as detailed in section 4 of the protocol. The results of the heterogeneous enantioselective carbonyl-ene reaction of substrates by Zn/(S)-KUMOF-1 (Table 1) showed that the smallest substrate (1a) could diffuse inside the crystal and convert to the product in a high yield (92%), proving that all reaction sites of the MOF were available. The yield and enantiomeric excess (ee) decreased as the substrate size increased, which suggests that the larger substrates could not access the reaction sites inside the MOF crystal. The largest substrate (1c) did not undergo the reaction in this system. It is plausible that the reaction channel was blocked by the corresponding reaction products in this case (Figure 3). When the size of the substrate is sufficiently small in comparison to the size of the void, additional substrates can penetrate the crystal. If the size of substrate is too large, the surface reaction site makes the first contact and directly blocks the entrance of the channel, which makes it impossible for other substrates to penetrate (Animation 1). As the reaction takes place near the surface, the ee is lower24 and the blockage of the reaction site decreases the reaction yield.
Particle size effect results (Table 2) showed that the larger crystals were better than the small crystals in utilizing the reaction sites inside the crystal, clearly demonstrating the identification of the reaction site in this system. The yields in the carbonyl-ene reaction of 1a using the three sizes of Zn/(S)-KUMOF-1 were similar, which indicates that the efficacies of the three MOFs are identical. The optical purity dramatically decreased with the decreasing size of the crystals because their surface area increased. In contrast, a larger sized crystal had much lower surface area, which allowed 1a to penetrate deeply and have better access to the inner reaction sites.
Unlike the Zn-mediated system, the Ti-catalyzed system provided more information about the events occurring at the catalytic reaction sites. The results of the heterogeneous catalytic carbonyl-ene reaction by Ti/(S)-KUMOF-1 (Table 3) revealed no discrimination by the substrate size; indeed, the effect of the substrate size on the yield was marginal. The optical purity of 2a was much lower compared to the product obtained via the Zn-mediated reaction. Most of the product was found in the reaction solution, and the amount inside the crystal was negligible. These results indicate that most reactions occurred on or beneath the surface and the products were immediately removed to the solution (Figure 4) (Animation 2). The substrate that is larger than the cavity size undergoes the reaction upon contact with the reaction site on the surface. The product dissociates quickly from the catalytic site without penetrating the crystal.
Based on these results, the reaction sites of MOFs can either be on the outer surface or the inner side of MOFs. However, as previously reported, the chiral environment of the reaction site varies by its location. A reaction that is catalytic with MOFs should follow the method proposed in this article to determine the location of the reaction site. Therefore, if the reaction is catalytic, claims of the reaction occurring inside the channel should be reconsidered.
Figure 1: Two classes of enantioselective carbonyl-ene reactions. Lewis acid Cat I and II were used for a homogeneous model reaction in a previous report27. This figure has been reprinted with permission from Han et al.27 Please click here to view a larger version of this figure.
Figure 2: Possible mechanism of the homogeneous stoichiometric carbonyl-ene reaction. Difference of binding affinity between the alkoxy and carbonyl group to metal makes the Zn-mediated carbonyl-ene reaction stoichiometric. This figure has been reprinted with permission from Han et al.27 Please click here to view a larger version of this figure.
Figure 3: Schematic presentation of the heterogeneous stoichiometric carbonyl-ene reaction. Pink colored molecules represent the substrates while greens indicate the products attached to the reaction sites. (A) 1a is small enough to penetrate the crystal regardless of blockage by 2a. (B) 1b suffers from blockage of 2b but still diffuses into the channel. (C) 1c makes first contact with the reaction site at the surface and directly blocks the entrance of the channel by 2c, which makes it impossible for another substrate to penetrate. This figure has been reprinted with permission from Han et al.27 Please click here to view a larger version of this figure.
Figure 4: Schematic presentation of the heterogeneous catalytic carbonyl-ene reaction. Pink colored molecules represent the substrates while greens indicate the products. (A) auf (D) illustrate the steps of the reaction. Dissociation of the products from the reaction site is very fast and penetrating the crystal is not necessary. This figure has been reprinted with permission from Han et al.27 Please click here to view a larger version of this figure.
entry | substrate | t (h) | yield (%) | ee (%) |
1 | 1a | 3.5 | 92 | 50 |
2 | 1b | 3.5 | 52 | 5 |
3 | 1c | 20 | NR | NA |
Table 1: Heterogeneous stoichiometric carbonyl-ene reaction of 1. NR = no reaction, under detection limit; NA = not applicable. This table has been reprinted with permission from Han et al.27
entry | catalyst | yield (%) | ee (%) |
1 | Zn/(S)-KUMOF-1-(L) | 92 | 70 |
2 | Zn/(S)-KUMOF-1-(M) | 89 | 50 |
3 | Zn/(S)-KUMOF-1-(S) | 91 | 0 |
Table 2: Result of particle size effect determination. Only 1a was used for this determination. Entries 1–3 correspond to large, medium, and small sized particles, respectively. This table has been reprinted with permission from Han et al.27
product obtained from solution | product obtained from MOF | ||||
entry | substrate | yield (%) | ee (%) | yield (%) | ee (%) |
1 | 1a | 85 | 24 | 2.8 | NA |
2 | 1b | 89 | 7 | 0.7 | NA |
3 | 1c | 83 | 0 | 0.2 | NA |
Table 3: Heterogeneous catalytic carbonyl-ene reaction of 1. NA = not applicable. This table has been reprinted with permission from Han et al.27
Animation 1: Animated illustration of the heterogeneous stoichiometric carbonyl-ene reaction. Please click here to view this video. (Right-click to download.)
Animation 2: Animated illustration of the heterogeneous catalytic carbonyl-ene reaction. Please click here to view this video. (Right-click to download.)
After the synthesis of (S)-KUMOF-1, crystals in some vials seem to be powdery and are not appropriate for use in catalysis. Therefore, proper crystals of (S)-KUMOF-1 need to be selected. The yield of (S)-KUMOF-1 is calculated using only those vials in which it was successfully synthesized. When withdrawn from the solvent, (S)-KUMOF-1 dismantles. Therefore, the crystals should always be kept wet. For this reason, weighing of intact (S)-KUMOF-1 crystals dipped into the solvent is difficult. The amount of (S)-KUMOF-1 needs to be determined before its use in catalysis. By synthesizing (S)-KUMOF-1 on a massive scale and abandoning some crystal samples in the vial, a statistical calculation of the yield was possible. The yield was calculated by weighing perfectly dried samples per vial. Samples were selected randomly from the population of synthesized (S)-KUMOF-1 averaged by the number of vials. This method gave a statistically defined amount of (S)-KUMOF-1 in one vial. The amount of (S)-KUMOF-1 required for catalysis was prepared by collecting the crystals in vials (e.g., the required amount of (S)-KUMOF-1 = amount of (S)-KUMOF-1 per vial x number of vials). Subdivision of (S)-KUMOF-1 floating in the solvent is incorrect for matching the calculated equivalents of substrates; the amount of substrate used is calculated from the calculated amount of (S)-KUMOF-1. The crystal structure and characteristics have been reported previously2,27.
Homogeneous carbonyl-ene reactions with Zn and Ti catalysts have been performed previously to prove that there is no discrimination by substrate size in the homogeneous reaction. At this point, the influence of the substrate size on the reaction efficiency can be neglected as the same as in the heterogenous reaction. The temperature required for the carbonyl-ene reaction using Ti/(S)-KUMOF-1 is 0 °C. Owing to the smashing problems of crystals, all reactions should be performed by shaking and not stirring. However, a low temperature shaking incubator chamber was not available. Instead, a polystyrene foam icebox was used. A stainless steel wire test tube rack was installed in the icebox and tightly sealed reaction vials were fixed into the rack. Water was poured to ~1 cm height into the icebox, and ice cubes were added. The lid-covered icebox was put on the shaker and fixed with adhesive tape. New ice cubes were added to replace the melting ice. For the carbonyl-ene reaction using Zn/(S)-KUMOF-1, the reaction vial was kept in a cryogenic cooling bath (dry ice with acetone) before adding the substrate to the solution. After adding the substrate, the reaction vial was moved to the icebox described above.
Additional well-marked data for the reaction site validation of Zn/(S)-KUMOF-1 and Ti/(S)-KUMOF-1 used in the carbonyl-ene reaction can be visualized by Two-Photon Microscopy (TPM) measurements27. Characterization of (S)-KUMOF-1 crystals by TPM has been previously reported. To gauge the cavity size of newly synthesized MOFs, TPM measurements with various size of dyes are available30.
The authors have nothing to disclose.
This work was supported by a National Research Foundation of Korea (NRF) Basic Science Research Program NRF-2019R1A2C4070584 and the Science Research Center NRF-2016R1A5A1009405 funded by the Korea government (MSIP). S. Kim was supported by NRF Global Ph.D. Fellowship (NRF-2018H1A2A1062013).
Acetone | Daejung | 1009-4110 | |
Analytical Balance | Sartorius | CP224S | |
Copper(II) nitrate trihydrate | Sigma Aldrich | 61194 | |
Dichloromethane | Daejung | 3030-4465 | |
Dimethyl zinc | Acros | 377241000 | |
Ethyl acetate | Daejung | 4016-4410 | |
Filter paper | Whatman | WF1-0900 | |
Methanol | Daejung | 5558-4410 | |
Microwave synthesizer | CEM | Discover SP | |
Microwave synthesizer 10 mL Vessel Accessory Kit | CEM | 909050 | |
N,N-Diethylformamide | TCI | D0506 | |
N,N-Dimethylaniline | TCI | D0665 | |
n-Hexane | Daejung | 4081-4410 | |
Normject All plastic syringe 5 mL luer tip 100/pk | Normject | A5 | |
Pasteur Pipette 150 mm | Hilgenberg | HG.3150101 | |
PTFE tape | KDY | TP-75 | |
Rotary Evaporator | Eyela | 243239 | |
Shaker | DAIHAN Scientific | DH.WSO04010 | |
Silica gel 60 (230-400 mesh) | Merck | 109385 | |
Synthetic Oven | Eyela | NDO-600ND | |
Titanium isopropoxide | Sigma Aldrich | 87560 | |
Vial (20 mL) | SamooKurex | SCV2660 | |
Vial (5 mL) | SamooKurex | SCV1545 |