Ying
Zhao‡
a,
Dan
Wu‡
bc,
Yidan
Qiao
a,
Guo-Ping
Yang
*b,
Lu-Fang
Ma
*ab and
Yao-Yu
Wang
b
aCollege of Chemistry and Chemical Engineering, Henan Key Laboratory of Function-Oriented Porous Materials, Luoyang Normal University, Luoyang 471934, P.R. China. E-mail: mazhuxp@126.com
bCollege of Chemistry and Materials Science, Northwest University, Xi'an 710127, P.R. China. E-mail: ygp@nwu.edu.cn
cShaanxi Applied Physics and Chemistry Research Institute, Xi'an 710061, P.R. China
First published on 7th March 2024
The current application of porous catalytic materials for organic synthesis is always confined to comparatively simple small substrates because of the diffusion barrier. Therefore, in this study, dimensional reduction and active site addition strategies were employed for preparing unique porous {RE9}-cluster-based rare-earth metal–organic frameworks (MOFs) {[Me2NH2]4[RE9(pddb)6(μ3-O)2(μ3-OH)12(H2O)1.5(HCO2)3]·6.5DMF·11H2O}n (MOF-RE, RE = Tb, Y, and Dy) with high-density multiple active sites. It was found that MOF-RE are rare {RE9}-based two-dimensional (2D) networks including triangular-nanoporous (1.3 nm) and triangular-microporous (0.8 nm) channels decorated by abundant Lewis acid–base sites (open RE(III) sites and Npyridine atoms) on the inner surface. As anticipated, due to the coexistence of Lewis acid–base sites, activated samples exhibited better catalytic activity (a yield of 96%, and a TON value of 768 for styrene oxide) than most previously reported 3D MOF materials for the cycloaddition of CO2 and multifarious epoxides under moderate conditions. Moreover, as a heterogeneous catalyst, MOF-Tb, has excellent catalytic performance (with a TON value of 396 for benzaldehyde) for the Knoevenagel condensation reaction of malononitrile and aldehydes with high catalytic stability and recoverability. In addition, both reactions possessed high turnover numbers and frequencies. These dimensional reduction and active site addition tactics may permit the exploitation of new nanoporous MOF catalysts based on rare-earth clusters for useful and intricate organic conversions.
Metal–organic frameworks (MOFs) containing different polynuclear metal-oxo clusters are unique porous crystalline materials with widespread promise for sensing, heterogeneous catalysis and gas storage/separation.15–21 Recent studies have exhibited that polynuclear cluster-based rare-earth (RE) MOFs possess exceedingly high stability and abundant catalytic active sites,22–24 in which the catalytic efficiency for CO2 conversion and Knoevenagel condensation could be greatly accelerated by the synergistic effect of activated metal ions (Lewis acid sites, LASs) and nucleophilic groups (Lewis base sites, LBSs).25,26 However, the accessibility of these RE-MOF materials to sterically demanding substrates is limited by expanding and stabilizing the active sites. Since the substrates cannot easily reach internal LASs, RE-MOFs perform poorly with larger substrates or complex reactions.27,28 It is essential to design functional ligands and precisely build RE-microporous MOFs with a high specific surface area and access to multiple active sites via an in situ function-oriented synthesis strategy to extract the catalytic ability of polynuclear cluster-based RE-MOFs to the greatest extent.
In light of the previous discussion on the standard of functional RE-MOFs as heterogeneous catalysts, this work thoroughly studied the effect of the topological structures of RE-MOFs on the catalytic activity and revealed dimensional reduction and active site addition methods to enhance catalytic reactivity through permitting unrestricted access to Lewis acid–base sites in two-dimensional (2D) MOFs (Scheme 1).29–31 These strategies can accurately predict the structures of MOFs, so as to achieve the regulation of the catalytic performance. Hence, a series of unique porous 2D RE-MOFs {[Me2NH2]4[RE9(pddb)6(μ3-O)2(μ3-OH)12(H2O)1.5(HCO2)3]·6.5DMF·11H2O}n (MOF-RE; RE = Tb, Y and Dy) with {RE9} clusters as secondary building units (SBUs) are successfully prepared from C2-symmetry V-shaped 2,6-bis(4′-carboxyl-phenyl)pyridine (H2pddb) under solvothermal conditions. Notably, MOF-RE have high densities of quantified open metal sites (OMSs) acting as LASs and uncoordinated pyridines as LBSs, all of which are exposed in open channels. Benefiting from these active sites, MOF-RE exhibits excellent catalytic efficiencies for CO2 transformation with high yields, turnover numbers (TONs), and turnover frequencies (TOFs) compared to those of many reported 3D MOF materials. Moreover, they have excellent catalytic properties for Knoevenagel condensations via the synergistic effects of the LASs and LBSs.
![]() | ||
Scheme 1 Different polynuclear cluster-based rare-earth MOFs formed by the {RE9} clusters and functional connectors. |
a R 1 = ∑||Fo| − |Fc||/∑|Fo|. b wR2 = [∑w(Fo2 − Fc2)2/∑w(Fo2)2]1/2. | |||
---|---|---|---|
Complex | MOF-Tb | MOF-Y | MOF-Dy |
Empirical formula | C117H69N6O45.5Tb9 | C117H69N6O45.5Y9 | C117H69N6O45.5Dy9 |
Formula mass | 3717.06 | 3086.97 | 3749.28 |
Crystal system | Hexagonal | Hexagonal | Hexagonal |
Space group | P63/mmc | P63/mmc | P63/mmc |
a [Å] | 22.5962(4) | 22.5233(11) | 22.6331(9) |
b [Å] | 22.5962(4) | 22.5233(11) | 22.6331(9) |
c [Å] | 22.7452(5) | 22.7551(15) | 22.6392(11) |
α [°] | 90 | 90 | 90 |
β [°] | 90 | 90 | 90 |
γ [°] | 120 | 120 | 120 |
V [Å3] | 10![]() |
9997.1(12) | 10![]() |
Z | 2 | 2 | 2 |
D calcd. [g cm−3] | 1.227 | 1.026 | 1.240 |
μ [mm−1] | 3.173 | 2.634 | 3.357 |
F [000] | 3524 | 3056 | 3542 |
θ [°] | 2.071–25.349 | 2.072–25.385 | 2.078–25.380 |
Reflections collected | 34![]() |
64![]() |
66![]() |
GOOF | 1.050 | 1.061 | 1.041 |
R , indices [I > 2σ(I)] | R 1 = 0.0261 | R 1 = 0.0419 | R 1 = 0.0270 |
wR2 = 0.0648 | wR2 = 0.1209 | wR2 = 0.0730 | |
R indices (all data) | R 1 = 0.0331 | R 1 = 0.0590 | R 1 = 0.0336 |
wR2 = 0.0680 | wR2 = 0.1303 | wR2 = 0.0766 |
The [RE9(μ3-O)2(μ3-OH)12(O2C-)12(H2O)1.5(HCO2)3] cluster has a threefold symmetry and is composed of nine RE ions arrayed in a {RE9} tricapped trigonal prism, according to detailed research of the nonanuclear cluster (Fig. S1a†). In contrast to the triangular planes of the tetragonal pyramids of the RE9 tricapped trigonal prism, which are each capped by a μ3-OH, the two triangular planes of the central RE6 trigonal prism are each capped by a μ3-O (Fig. S1b†). The six RE1 ions are each coordinated with eight O atoms: two from carboxylate moieties of two separate pddb2− connecters and four from four μ3-OH, and another two coordination sites are composed of O atoms from one HCO2− ligand and a μ3-O (O1) (Fig. S2a†). The remaining three RE2 cations are each coordinated with nine O atoms: a terminal H2O molecule, four μ3-OH and four carboxylate O from four independent pddb2− ligands (Fig. S2a†). The nonanuclear [RE9(μ3-O)2(μ3-OH)12] cluster is created by twelve μ3-OH and two μ3-O connecting nine RE elements (Fig. 1a) and is terminated through twelve carboxylates from twelve independent pddb2− linkers to generate a 12-connected [RE9(μ3-O)2(μ3-OH)12(O2C-)12] SBU. As the extension point of the [RE9(μ3-O)2(μ3-OH)12(O2C-)12] core, C atoms from the carboxylate groups of twelve independent pddb2− linkers are arranged into a hexagonal prism, which conforms to the d6R vertex diagram of a 12-connected node. There are three H2O molecules and three extra HCO2− bridging linkers coordinated with the RE ions to form a nonanuclear [RE9(μ3-O)2(μ3-OH)12(O2C-)12(H2O)1.5(HCO2)3] cluster (Fig. 1a).
In the structure of MOF-RE, the pddb2− ligands adopt a bidentate bridging mode to connect two adjacent nonanuclear clusters (Fig. S2b†) to form 2D anionic layers with triangular-microporous (∼0.8 nm) and triangular-nanoporous (∼1.3 nm) channels (Fig. 1b–d). The 2D infinite layers are stacked along the c axis with an average interlayer spacing of 11.37 Å (Fig. 1e and i). Notably, there are honeycomb-like hexagonal channels with a diameter of ∼1.3 nm along the c axis, which are full of protonated [Me2NH2]+ cations (Fig. 1f). The porosity computed by PLATON is ∼56.1% of the overall crystal volume after excluding free solvents. Topologically, the [RE9(μ3-O)2(μ3-OH)12(O2C)12(H2O)1.5(HCO2)3] clusters and ligands may act as 12-connected nodes and linear rods, respectively, and the whole structure can be represented as a 12-connected skeleton (Fig. 1g and h).
The latent application of MOF-Tba for CO2 and light hydrocarbon (CH4 and C2Hn) sorption (Fig. S6 and 7†)/separation (Fig. 2c and d) has been carefully examined because of its inherent perpetual porosity and channel circumstances. The results displayed that the loading capacity of C2H2 in MOF-Tba is higher than that of other gases, demonstrating the maximum interaction between C2H2 and the framework. The sorption enthalpies (Qst) were determined using the virial approach to establish the adsorption affinity between the five gases and skeleton more accurately (Fig. 2b and Fig. S8†). The Qst of MOF-Tba to C2H2 (37.6 kJ mol−1) is higher than that of CO2 (22.4 kJ mol−1), C2H4 (27.2 kJ mol−1), C2H6 (26.7 kJ mol−1), and CH4 (8.6 kJ mol−1) under zero coverage, which agrees with the measured adsorption amount. Moreover, the possibility of separating CH4 from light hydrocarbons was studied by ideal solution adsorbed theory (IAST) for binary equimolar mixtures (Fig. S9 and 10†).33 The selectivities for CO2, C2H2, C2H4 and C2H6 over CH4 at 1 bar and 298 K are 5.9, 9.8, 6.8 and 7.0, respectively (Fig. 2c), making MOF-Tba an exceptional sorbent for effectively removing CO2/C2 light hydrocarbons from natural gas.
Entry 1 (Table 2) displays that when MOF-Tba (0.05 mol%) was added as a catalyst, only a small amount of product with a yield of 5% was detected within 12 h. In addition, only a slight conversion (10%) could be observed under the cocatalyst of n-Bu4NBr (1 mol%) alone, as shown in entry 2. Nevertheless, the yield was tremendously improved to 29% (entry 3) when MOF-Tba (0.05 mol%) and n-Bu4NBr (1 mol%) were concurrently introduced to the reaction, suggesting that MOF-Tba and n-Bu4NBr synergistically activated the second-order reaction of CO2 and epoxides. Increasing the temperature was used to demonstrate that one of the key factors was temperature for influencing the reaction outcome, as shown in entries 4 and 5. Furthermore, entries 6and 7 examined and listed the effect of the cocatalyst n-Bu4NBr dosage, which demonstrated that the quantity of the cocatalyst had a clear impact on the reaction rate. The yield increased to 95% when 5 mol% n-Bu4NBr cocatalyst was added. The amount of MOF-Tba was increased because the practical application of the catalyst will be severely hampered by the 12 h reaction time. Styrene oxide could be converted into 4-phenyl-1,3-dioxolan-2-one more rapidly, as described in entries 8–10. In conclusion, it was found that the ideal reaction conditions are 0.125 mol% MOF-Tba catalyst, 5 mol% n-Bu4NBr cocatalyst, 60 °C and 6 h. In addition, we performed a detailed analysis of the 1H NMR spectrum for entry 10, confirming that no by-products were generated during the reaction (Fig. S19†).
Entry | MOF (mol%) | n-Bu4NBr (mol%) | T (°C) | t (h) | Yieldb (%) |
---|---|---|---|---|---|
a Reaction conditions: solvent-free, styrene oxide (20 mmol), and CO2 (1 atm). b The product yield was determined by 1H NMR. | |||||
1 | 0.05 | 0 | 25 | 12 | 5 |
2 | 0 | 1 | 25 | 12 | 10 |
3 | 0.05 | 1 | 25 | 12 | 32 |
4 | 0.05 | 1 | 40 | 12 | 54 |
5 | 0.05 | 1 | 60 | 12 | 78 |
6 | 0.05 | 3 | 60 | 12 | 86 |
7 | 0.05 | 5 | 60 | 12 | 95 |
8 | 0.075 | 5 | 60 | 10 | 94 |
9 | 0.1 | 5 | 60 | 8 | 95 |
10 | 0.125 | 5 | 60 | 6 | 96 |
The catalytic universality of MOF-Tba was further evaluated utilizing a range of propylene oxide derivatives with distinct substituents and steric hindrance under determined ideal reaction circumstances (Table 3). The outcomes indicated that there was some regularity in how different substituents affected the yield. It can be seen from the comparison of entries 1 and 2 that the epoxy complexes with electron-withdrawing groups (–Br and –Cl) could improve the efficiency of the cycloaddition reaction, and the yield could reach more than 99%, the reason for which is that the electron-withdrawing group may decrease the electron density of ethylene oxide.35,36 In contrast, electron-donating groups had a disadvantageous impact on this process, as in entry 3.37,38 Furthermore, entries 4–6 exhibited a significant decrease in the conversion of epoxide with bulky substituents, confirming the idea that big substituents restrict the mobility of substrate molecules.39–42 Table S1† lists information about the molecular sizes of all epoxide derivatives. Additionally, the TON of MOF-Tba for styrene oxide was notable compared to most reported TON values for MOF catalysts (Table S2†), which was likely attributable to the profitable contribution of {Tb9} clusters and abundant Npyridine groups in the channels.
Entry | Epoxide | Product | Yieldb (%) | TONc | TOFd (h−1) |
---|---|---|---|---|---|
a Reaction conditions: solvent-free, epoxides (20 mmol), n-Bu4NBr (5 mol%), Tb-MOF catalyst (0.125 mol%), CO2 (1 atm), 60 °C, and 6 h. b Yield was determined by 1H NMR. c TON = [product (mmol)]/[catalyst (mmol)]. d TOF = TON/time. | |||||
1 |
![]() |
![]() |
>99 | 792 | 132 |
2 |
![]() |
![]() |
>99 | 792 | 132 |
3 |
![]() |
![]() |
96 | 768 | 128 |
4 |
![]() |
![]() |
96 | 768 | 128 |
5 |
![]() |
![]() |
96 | 768 | 128 |
6 |
![]() |
![]() |
95 | 760 | 127 |
The actual organic synthesis industry depends heavily on the stability and recyclability of catalysts;43 hence, additional tests were conducted about hot leaching, recovery, and recycling of MOF-Tba. First, the recycling stability of MOF-Tba for the cyclization reaction of CO2 with styrene oxide was studied under the determined ideal reaction conditions. The conversion of styrene oxide was nearly unchanged for five cycles by the recovered catalyst MOF-Tba (Fig. S11†). Meanwhile, the PXRD pattern of the recovered MOF-Tba sample after five experiments was essentially matched with the newly formed one, showing that the MOF-Tba catalyst maintained the stability of the framework (Fig. S12†). Afterward, inductively coupled plasma (ICP) analysis was then used in leaching experiments. As a result, the recovered filtrate included just a little quantity of Tb(III) ions (0.015%), further demonstrating the stability of MOF-Tba in the organic reaction. Furthermore, a thermal filtration experiment was carried out, and the results showed that the reaction hardly happens when the catalyst is filtered out (Fig. S13†), meaning that MOF-Tba possessed a heterogeneous nature.
The probable catalytic mechanism can be deduced from prior MOF-related literature,44–47 and the distinctive structural characteristics of MOF-Tba, including high specific surface area, functional channel, and plentiful {Tb9} clusters (Fig. 3). First, the epoxide rapidly diffuses into the MOF-Tba catalyst and makes weak contact with its exposed metal sites in a confined environment. Then, the nucleophilic attack of less-obstructed carbon atoms in the epoxide by the Br− anion released by n-Bu4NBr promotes the formation of the alkylcarbonate anion. Subsequently, polarized CO2 molecules tend to undergo nucleophilic addition reactions with alkylcarbonate anions to generate alkylcarbonate salt. Finally, the ring closure behavior results in the production of cyclic carbonate and liberation of catalysts.
Initially, perfect reaction conditions were studied with benzaldehyde and malononitrile as substrates in the presence of desolvated MOF-Tba as a heterogeneous catalyst, as seen in Table 4 and Fig. S29–31.† Entry 1 shows that only trace 2-benzylidenemalono-nitrile (3%) was produced at 25 °C without the catalyst MOF-Tba, whereas when MOF-Tba (0.1 mol%) was introduced, a yield of 38% was generated within 2 h (entry 2), indicating that the reaction could scarcely be carried out without the catalyst. The conversion of the substrate increased progressively when all other factors governing the reaction were held constant, and only the catalyst dosage was increased (entries 3 and 4). At 25 °C, the yield was 85% when the MOF-Tba dosage was raised to 0.25 mol% (entry 5). The reaction substrate was virtually entirely converted as the temperature rose from 25 °C to 60 °C in parallel investigations using 0.25 mol% MOF-Tba (entries 6 and 7), demonstrating that temperature was one of the crucial elements in the condensation reaction. Additionally, the connection between the conversion with reaction time was examined under the reaction environments of 0.25 mol% MOF-Tba at 60 °C (entries 8–10). It can be seen from the above results that the ideal reaction conditions were 0.25 mol% MOF-Tba, 60 °C, and 2 h when the ratio of malononitrile to aldehyde was 2:
1.
Entry | MOF-Tba (mol%) | Time (h) | T (°C) | Yieldb (%) |
---|---|---|---|---|
a Reaction conditions: malononitrile (20 mmol), benzaldehyde (10 mmol). b The product yield was determined by 1H NMR. | ||||
1 | 0 | 2 | 25 | 3 |
2 | 0.1 | 2 | 25 | 38 |
3 | 0.15 | 2 | 25 | 58 |
4 | 0.2 | 2 | 25 | 76 |
5 | 0.25 | 2 | 25 | 85 |
6 | 0.25 | 2 | 40 | 93 |
7 | 0.25 | 2 | 60 | 99 |
8 | 0.25 | 1.5 | 60 | 95 |
9 | 0.25 | 1 | 60 | 88 |
10 | 0.25 | 0.5 | 60 | 65 |
Based on aforementioned discoveries, we chose several aldehyde derivatives with various substituents and steric hindrance (Table S3†) to confirm the MOF-Tba catalyst's suitability for the Knoevenagel condensation reaction, and the outcomes are displayed in Table 5 and Fig. S32–38.†
Entry | Substrate | Product | Yieldb (%) | TONc | TOFd (h−1) |
---|---|---|---|---|---|
a Reaction conditions: aldehyde derivatives (10 mmol), malononitrile (20 mmol), catalyst MOF-Tba (0.25 mol%), 2 h, 60 °C. b The yield was calculated by 1H NMR. c TON = [product (mmol)]/[catalyst (mmol)]. d TOF = TON/time. | |||||
1 |
![]() |
![]() |
>99 | 396 | 198 |
2 |
![]() |
![]() |
>99 | 396 | 198 |
3 |
![]() |
![]() |
>99 | 396 | 198 |
4 |
![]() |
![]() |
>99 | 396 | 198 |
5 |
![]() |
![]() |
98 | 384 | 192 |
6 |
![]() |
![]() |
94 | 372 | 186 |
7 |
![]() |
![]() |
87 | 348 | 174 |
Entries 2–4 show that the conversion efficiency of benzaldehyde with electron-drawing groups (–F, –Br, and –NO2) surpassed 99%, whereas the catalytic yield was slightly decreased due to the presence of the electron-donating groups (–CH3 and –CH2CH3) (entries 5 and 6), indicating that the electron-donating moieties greatly inhibit the Knoevenagel condensation reaction. Additionally, the conversion efficiency clearly reduced as the molecular size and steric hindrance of the substrate increased (entry 7). Remarkably, benzaldehyde had a TON value of 396, which was much higher than the majority of previously documented MOF catalysts (Table S4†).
The best experimental conditions were used to study the stability and recyclability of MOF-Tba. The utilized MOF-Tba catalyst was recovered and repeatedly cleaned with DMF after each reaction. Over 97% of 2-benzylidenemalono-nitrile was produced after five repetitions of the process using MOF-Tba, which retained excellent catalytic activity (Fig. S26†). Furthermore, the PXRD peaks of gathered MOF-Tba demonstrated that the host framework remained unchanged, suggesting the great stability of the microporous heterogeneous catalyst (Fig. S27†). Following the catalytic recycling experiment, leached homogenous Tb(III) was monitored by ICP analysis as well. The probability of leaching metal ions from the MOF-Tba network during Knoevenagel condensation was ruled out when trace Tb(III) of 0.018% was found in the filtrate. Under optimal reaction circumstances, a heat filtration test was conducted to confirm the heterogeneous nature of MOF-Tba. The solid catalyst was filtered out after the reaction had been going on for 0.5 hours. Since the conversion rate barely altered (Fig. S28†), the heterogeneous nature of MOF-Tba was further supported.
Fig. 4 implies a likely catalytic reaction mechanism based on relevant published studies53–59 and structural characteristics of MOF-Tba. First, the carbonyl oxygen of the aldehyde group made a weak interaction with the exposed Tb(III) sites of MOF-Tba, converting its carbon atom into a positive carbon center and initiating the reaction. Simultaneously, Npyridine atoms as the LBSs caused the carbonyl carbon of malononitrile to polarize into a negative center. Second, a covalent bond was formed between two carbon atoms with opposing electric charges, resulting in an imine intermediate. Finally, the recombination of intramolecular electrons combined with the release of H2O molecules and the used catalyst resulted in the formation of the product benzylidenemalononitrile.
Footnotes |
† Electronic supplementary information (ESI) available: Experimental details and supporting figures. CCDC 2216930–2216932. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3qi02527f |
‡ These authors contributed equally to this work. |
This journal is © the Partner Organisations 2024 |